Citation
Influence of changes in freshwater flow on the use of mangrove prop root habitat by fishes

Material Information

Title:
Influence of changes in freshwater flow on the use of mangrove prop root habitat by fishes
Creator:
Ley, Janet A., 1951-
Publication Date:
Language:
English
Physical Description:
ix, 172 leaves : ill., maps ; 29 cm.

Subjects

Subjects / Keywords:
Ecology ( jstor )
Estuaries ( jstor )
Fish ( jstor )
Forage ( jstor )
Fresh water ( jstor )
Juveniles ( jstor )
Predators ( jstor )
Salinity ( jstor )
Species ( jstor )
Water depth ( jstor )
Dissertations, Academic -- Environmental Engineering Sciences -- UF
Environmental Engineering Sciences thesis Ph. D
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1992.
Bibliography:
Includes bibliographical references (leaves 161-171)
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Janet A. Ley.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
027843704 ( ALEPH )
26625367 ( OCLC )

Downloads

This item has the following downloads:


Full Text












INFLUENCE OF CHANGES IN FRESHWATER FLOW ON THE USE OF
MANGROVE PROP ROOT HABITAT BY FISHES














By

JANET A. LEY








A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA


1992













UNIVERSITY O FLORIDA �N188















ACKNOWLEDGEMENTS

I would like to extend my thanks to all my committee members and my Department Chairman, Joseph Delfino. They advised and supported me throughout my field work and writing. Clay Montague introduced me to science and Florida Bay, and encouraged me to pursue my interests. Carole McIvor gave me guidance, scientific insight, and maintained faith in my abilities under all circumstances. I am extremely grateful to Bill Seaman for his efforts in obtaining Sea Grant support for a substantial portion of the project. In his Wetlands Ecology class, Ronnie Best introduced me to working in mud and swamps, features which later became a major part of my life. He also allowed me to live in the Center's Winnebago for two years, permitting me to operate on an intense and flexible schedule in Key Largo. Frank Nordlie expressed constant interest and encouraged me in my work. My visits with Nick Funicelli always included valuable personal and professional insights.

I am extremely grateful for the efforts of Dan Haunert of the South Florida Water Management District. He believed in the benefits of this research to Florida Bay and aggressively oversaw the process of obtaining funding. In




ii









addition, my conversations with him gave me renewed enthusiasm and perspective. Dewey Worth, Dan's colleague, followed through with continued encouragement and support in later phases of the project.

Laura Flynn and Luke Hasty were my very competent field assistants, providing good humor and constructive suggestions. They remained enthusiastic in every circumstance, from diving with major unknown creatures, to measuring 2-day old fish in 95 degree heat, to snorkeling in double hoods and wetsuits.

Jacque Stevens, Harriett McCurdy and my brother, Fred Ledtke, Jr., were my most faithful volunteers. Jacque helped me tether over 200 fish and her ideas were invaluable. Fred devoted his hard earned vacations to his older sister's unusual effort.

I would also like to express my gratitude to the staff of Everglades National Park. At the Key Largo Ranger Station, Dave and Louise King, Linda Cramer, and Dave Viscera included me as part of their small neighborhood during my 2 year residency. I am grateful for their support and rescues during boat break-downs. From the South Florida Research Center, Mike Robblee allowed me to use Park boats, provided insights concerning my research questions and encouraged my efforts. DeWitt Smith also gave me encouragement and perceptive advice. Bill Loftus' help in




iii










fish taxonomy was invaluable. Katy Kuss was both a great adviser and a good friend.

Gordon Thayer, National Marine Fisheries Service,

advised me on the use of enclosure nets. In addition, other visiting scientists shared ideas with me, including Paul Carlson, Florida DNR; Jay and Rita Zieman and Jim Fourqurean, University of Virginia; and Dave Porter, University of Georgia. I am also grateful to the scientists of the National Audubon Society in Tavernier, Florida, who treated me as an adjunct staff member, especially George Powell, Mike Ross and Jerry Lorenz.

In Gainesville, Ken Portier helped in the initial study design and last phases of analysis. In the bulk of the analysis effort, Steve Linda advised me on handling a very large data set. Hans Gottgens, my officemate, was constantly patient and extremely helpful in offering computer, scientific and personal advice.

Most importantly, I would like thank those who

encouraged my pursuit of this degree as a personal goal and supported me throughout the process. These are my parents who nurtured my spirit of independence and appreciation of nature, and Darlene Kalada, my best friend, on whose support and encouragement I could always depend.







iv
















TABLE OF CONTENTS
Page

ACKNOWLEDGEMENTS....................................... ii

ABSTRACT ........................................................ Viii

CHAPTERS
1. GENERAL INTRODUCTION.............................. 1
Estuarine Fish Ecology .......................... 1
Mangrove Fish Ecology........................... 3
The Florida Bay Ecosystem ....................... 5
Problem Definition .............................. 9
Objectives.... .................................... 12
Study Area.................................... 13
Fish Community Sampling Design .................. 14

2. FISH DENSITIES AND ASSEMBLAGE PATTERNS IN
MANGROVE HABITATS: COMPARISONS
ACROSS SALINITY GRADIENTS ....................... 17
Materials and Methods ........................... 19
Results ..................................... 32
Discussion......... ........................ ...... 84

3. FISH COMMUNITIES IN FLORIDA BAY MANGROVE
SHORELINE HABITATS: RELATIONS WITH PHYSICAL
PARAMETERS AND COVER ............................. 98
Materials and Methods ........................... 99
Results .... . ...................... ........ 109
Discussion..................................... . 114

4. FOOD HABITS OF MANGROVE FISHES:
A COMPARISON ACROSS SALINITY GRADIENTS .......... 120 Materials and Methods ........................... 121
Results ........... ...... ..................... 123
Discussion...................................... 129

5. PREDATOR ENCOUNTER RATES ON SMALL BENTHIC
FISH ACROSS A SALINITY GRADIENT ................. 134
Materials and Methods ........................... 137
Results......................................... 143
Discussion...................................... 149







v










6. IMPLICATIONS AND CONCLUSIONS ...................... 154
Implications for Mangrove Fish Ecology........... 154
Implications for Estuarine Fish Ecology:
the Nursery-ground Hypothesis................. 157
Management Implications ......................... 158

LITERATURE CITED....................................... 161

BIOGRAPHICAL SKETCH.................................... 172













































vi














Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

INFLUENCE OF CHANGES IN FRESHWATER FLOW ON THE USE OF
MANGROVE PROP ROOT HABITAT BY FISHES By Janet A. Ley

May 1992

Chairperson: Clay L. Montague Cochairperson: Carole C. McIvor Major Department: Environmental Engineering Sciences

The hypothesis that seasonal changes in freshwater inflow (indicated by salinity) influence habitat use by fishes was tested in northeastern Florida Bay, extreme south Florida. Fishes were sampled monthly for 13 months using visual censuses and enclosure nets.

Of the 305,589 individuals observed, 91% were estuarine residents, numerically dominated by engraulids, atherinids and cyprinodontids. Occasional marine and freshwater visitors comprised 2% of the individuals, and estuarine transients, 8%. No young-of-the-year estuarine transients were observed.

Salinity ranged between 0.0 to 58 parts per thousand (ppt) upstream, 19.5 to 54 ppt midstream, and 30 to 50 ppt downstream. The 77 species were grouped for analysis:






vii









small benthic, small water column, and larger fishes. Abundances of larger fishes were consistently lower upstream (0.15 fish/square meter (m2)), than mid- (0.65 fish/m2), or downstream (0.55 fish/m2). Species of larger fishes numbered fewer upstream (11), than midstream (15), and downstream (22). Benthic and water column fish abundances did not vary along the gradient. Temporally, fish distribution was uncorrelated with salinity.

Development of mangrove habitat and submerged aquatic vegetation (SAV) were reduced upstream. Fish diets shifted to other foods upstream. Thus, where seasonal changes in freshwater inflow were greater (i.e. upstream), species and numbers of larger fishes were lower, possibly due to salinity conditions, food availability and habitat development.

To determine if lower salinity conditions alone led to reduced predation, prey fishes were tethered along the gradient. Predator encounter rates were not different over the salinity range tested, but were 50% lower at the most remote sites. This was perhaps a function of accessibility of the sites to roving predators.

Water management strategies to increase mangrove

development and SAV are recommended research priorities. However, severe ecotonal differences between Bay and ocean waters, coupled with limited circulation and significant




viii









predation may inhibit recruitment and survival of postlarval fishes from offshore. An unbroken continuum of good habitat from outer to upper reaches may be necessary if northeastern Florida Bay is to function as a prime nursery area for estuarine transient fishes.











































ix














CHAPTER 1
GENERAL INTRODUCTION


The goal of improved management of surface waters to benefit estuarine fish populations in Florida Bay provided the incentive for this research. Before objectives and strategies can be established toward this goal, however, a better understanding of how freshwater inflow influences fish communities in mangrove estuaries overall is needed. This involves aspects of fish ecology of both estuarine and mangrove ecosystems.

Estuarine Fish Ecology

Ecologists often divide estuarine fishes into three groups: estuarine residents (complete their entire life cycle in the estuary), estuarine transients (spawn offshore, their young use the estuary as a nursery), and occasional marine visitors (usually adults) (Day et al. 1989). Resident and transient species tend to be widespread, but marine visitors are usually restricted to the higher salinity zones of the lower estuary (Weinstein 1979). At certain times of the year densities may increase dramatically as influxes of transient juveniles enter estuarine systems. Some species tend to migrate to upstream-most habitats upon initially entering the estuary;



1









they then may disperse to lower reaches as they grow larger (Weinstein 1979, Rogers et al. 1984, Loneragan et al. 1990).

The prominence of transient juvenile fish and

crustaceans led to the application of the term "nurseryground" to many estuaries (Gunter 1961, McHugh 1967, Weinstein 1979). A major role of freshwater discharge in such systems may be to increase food availability for fishes by transporting nutrients which stimulate primary production and by increasing detrital transport and processing (Odum et al. 1982). Freshwater inflow may also improve the chance for survival of juvenile fish in estuaries by reducing salinity levels below the limits tolerable by stenohaline marine predators (Gunter 1961, 1967).

Browder and Moore (1981) offered a comprehensive

nursery ground hypothesis linking several of these concepts. They split habitat factors into those that are relatively stable (e.g. shoreline edge, bottom type) and those that are movable (e.g. salinity, food resources). Favorable habitat for particular juveniles consists of combinations of these factors that promote growth. According to their theory, the inflow of freshwater acts to position an area of favorable moveable habitat relative to important stationary habitat. Thus, for any estuary there is a rate of freshwater flow sufficiently high to push the band of potentially favorable moveable features beyond estuarine boundaries into open waters, perhaps eliminating favorable habitat entirely. Likewise, for every estuary, there is a rate of freshwater






3

flow so low that the band of favorable salinities retreats upstream where stationary features may be unfavorable. The ideal situation with regard to freshwater inflow is one that maximizes the area of favorable habitat within the estuary over the peak period of nursery use. This hypothesis seems particularly applicable for analyses of fish ecology in mangrove-dominated estuaries.

Mangrove Fish Ecology

In tropical and subtropical areas of the world,

mangroves are dominant shoreline features. Mangrove-derived detritus forms a food base for fish occupying mangrove ecosystems (Odum 1971). Mangrove shorelines may also provide cover for fishes (Thayer et al. 1987a). However, few studies have documented aspects of the direct use of mangrove habitats by fishes probably because monitoring fishes within the complex tangle of roots and branches is extremely difficult. Efforts have only recently focused on obtaining quantitative data on habitat use (i.e. Thayer et al. 1987a, Sheridan 1991, Morton 1990, Robertson & Duke 1987). Strong linkages between mangroves and adjacent habitats may exist. For example, diel habitat shifts occur in both non-tidal (Thayer et al. 1987a) and tidal systems (Morton 1990, Robertson & Duke 1987). Shifts from other habitats to mangroves occur during the life history of some species such as gray snapper (Lutjanus griseus) (Starck & Schroeder 1971).






4


The degree of selection among mangrove habitats by fishes has not been determined. Due to variation within mangrove forests, however, preferences are likely to be displayed. In south Florida, for example, three species of mangrove trees occur: red mangroves (Rhizophora mangle), black mangroves (Avicennia germinans) and white mangroves (Laguncularia racemosa) (Odum et al. 1982). These trees vary greatly in type of submerged features and potential cover for fishes. For example, red mangroves provide proproots; these are strong, woody structures that tend to extend from the mid-tree trunk downward to the substrate. Black mangroves, in contrast, tend to support a bed of pneumatophores that are pencil-like structures that grow upward from the substrate to several centimeters.

Variation in degree of exposure to flushing also contributes to differences among mangrove forests and, hence, may influence habitat use by fishes. While the fringing mangroves along the shoreline are regularly flooded and thus accessible, more interior basin forests are irregularly inundated and thus occasionally available (Odum et al. 1982).

Within fringing shorelines, higher flushing rates

contribute to greater mangrove habitat development (e.g. taller trees, more leaf production), which, in turn, is likely to generate more massive submerged structure for cover. Furthermore, detritus-based food resources are likely to be more abundant near highly productive mangroves.






5


Most mangrove-dominated estuaries contain examples of all these habitats and may thus provide a diversity of conditions for use by fishes.

The Florida Bay Ecosystem

Florida Bay is a large (1,500 km2), mangrove/seagrass dominated estuary located in extreme south Florida. The majority of the Bay is not subject to tidal influence. Wind-driven water movements can, however, raise or lower Bay water levels rapidly. Sustained strong easterly winds can literally blow water out of Florida Bay into the open Gulf of Mexico. In similar fashion, winds from the north can accelerate the introduction of mainland drainage into the northern part of the Bay, and winds from the west can move the water into the northeastern corner of the Bay (Ginsberg 1956).

Internal circulation is restricted due to several features of the Bay. The interior contains over 300 mangrove-fringed and overwash islands. On the east, U. S. Highway 1 separates Florida Bay and Barnes Sound with only one pass and two culverts providing water exchange. On the west, it is separated from the open waters of the Gulf of Mexico by a series of mud banks that are at least 2.0 km wide and are often exposed (Holmquist et al. 1989b) (Figure 1-1). The lower Bay is separated from the thermally stable and constant flow of the Gulfsteam by a series of limestone islands known as the Florida Keys. Several major passes occur through the Keys in the western Bay, but in the























Figure 1-1. Regional maps of the study area showing features of the upstream drainage basins and Florida Bay. a. Boundaries of drainage basins and tributaries to northeastern Florida Bay. (Source: Schomer & Drew 1982). b. Florida Bay showing the extensive mudbank system (stiple pattern). Arrows indicated passes to the Atlantic Ocean and Barnes Sound. (Source: Holmquist et al. 1989b)







7

a. \US Highway 41











Taylor 0--111 canal Slough
.Basin














b.* Study Area C jewash Creek





Adam's Cut Mud Banks ( Florida Bay




Tavemier Creek 10km.

Long Key









northeastern portion, only one pass exists, a man-made cut through Key Largo to the ocean side (Adam's Cut) (Figure 11). On the northern boundary of the Bay are the Florida Everglades.

The Florida Bay area is subject to an annual water

deficit with evaporation exceeding total rainfall (Tabb et al. 1962). Annual rainfall in northeastern Florida Bay ranges from 1600 mm on the mainland at Homestead to 1200 mm on the south at Key Largo. The climate of subtropical south Florida is characterized by a relatively long and severe dry season (November through April) and a wet season (May through October) (Schomer & Drew 1982).

Sea level becomes relatively high on an annual basis from August to December reaching a maximum of about 15 cm above the annual average in October (Ginsberg 1956, Provost 1973, Holmquist et al. 1989b). By late November or early December, Bay level recedes to the annual average, which probably accelerates the drainage of freshwater into the Bay from the mainland. At this time, the zone of reduced salinity may extend farther south and southeast into midand downstream Florida Bay areas.

The major source of freshwater flow into Florida Bay is from a series of approximately 20 creeks and Taylor River, which carry surface water from the Taylor Slough/C-lll drainage area into the Bay. This system is smaller than the









Shark River Slough, a separate system which extends from Lake Okeechobee southward toward the Gulf of Mexico and drains most of the Everglades.

These overall features contribute to several

environmental and biological patterns. Gradients in environmental variables occur in Florida Bay, from southwest to northeast. These gradients include amount of water exchange, sediment depth, and seagrass standing crop (Zieman et al. 1989). The area northeast of the central line of mud banks is characterized by very restricted circulation and no tidal influence (Schomer & Drew 1982). A thin sediment veneer covers the basin bedrock in the northeast Bay, deepening towards the southwest. In addition, seagrass density and productivity decreases dramatically from southwest to northeast (Zeiman et al. 1989).

Problem Definition

Water management decisions in the eastern Everglades

have potentially impacted Florida Bay through changes in the timing and quantity of freshwater discharge. Under predrained conditions, in this area, surface freshwater moved over grassy marl prairies that were seasonally flooded (Schomer & Drew 1982). A complex network of streams, bordered by mangroves and other shrubs carried freshwater inflow to receiving waters downstream in a manner that was presumably both gradual and dispersed.

Beginning in the early 1900s, construction began on an extensive system of canals and ditches throughout much of






10


the Everglades system. The effects of these canals may have included the overall reduction in the amount of freshwater storage in the system (T. MacVicar, South Florida Water Management District, personal communication). In addition, after Everglades drainage, drier conditions may have occurred more frequently in the prairies and sloughs, with greater contrast between wet season and dry. Overland flow of freshwater entering the downstream estuaries under these altered conditions has probably been more rapid and less spatially dispersed.

After entering upstream portions of Florida Bay, freshwater moves through extensive mangrove wetlands consisting of shallow swamp lands, creeks, ponds and bays, eventually reaching the open portions of Florida Bay. Changes in these brackish and marine receiving waters attributed to managed freshwater inflow may have included alteration of the annual salinity pattern, which led to unnatural cycles of both reduced and hypersaline conditions. Historical salinity data for Everglades waters, however, is lacking for the period prior to initiation of drainage.

Evidence of the ecological effects of drainage on the downstream estuary has been discerned from National Audubon Society studies showing declining populations of estuarine wading birds (e.g. spoonbills). Changes in hydroperiod have been hypothetically linked to a reduced fish and shellfish prey-base for the birds (National Audubon Society, unpublished data, 1989). Furthermore, recent decreases in









sportfish populations have been linked to hypersalinity stress for certain sportfish in Everglades National Park (Rutherford et al. 1989). The Everglades estuaries are also critical habitat areas for other endangered aquatic species (e.g. American crocodile) that rely on the same forage base as do birds and sportfish (SFWMD 1989).

Thus, groups concerned about these problems spurred

South Florida Water Management District (SFWMD) officials to take action that would return more natural drainage patterns to the estuarine areas of the Everglades. Some of these actions have focused on the C-1ll Canal/Taylor Slough watershed which includes agricultural lands in a large drainage basin east of the Park. The downstream leg of the canal runs northwest to southeast, passes under U. S. Highway 1 and continues southward outside of Florida Bay, to Barnes Sound (Figure 1-1). In low flow periods, the canal has functioned like a dike by preventing overland flow of freshwater from reaching both the downstream prairies and the approximately seven small creeks tributary to northeastern Florida Bay. In high flow conditions, water still flows through, sometimes sending slugs of freshwater into northeastern Florida Bay. Local topographic conditions tend to direct more freshwater toward U.S. Highway 1 than toward the west (Tabb et al. 1967). Thus, under these management conditions, the historic salinity regime is likely to have been altered. Changes in hydroperiod have probably resulted in more severe hypersaline conditions and







12

sudden salinity changes of great magnitude, especially in the eastern part of Florida Bay.

In the mid-1980s engineering alterations created

several cutouts, each 20 meters wide, in the south bank of C-111 canal. The cutouts were intended to restore the more dispersed and gradual pattern of freshwater inflow to northeastern Florida Bay. Furthermore, an earthen plug was installed to block the C-111 outfall to Barnes Sound except on extreme floods when SFWMD can release water by opening it with draglines. The result of these alterations was to provide more flexible management of freshwater flow to northeastern Florida Bay. The question remaining is how to utilize this flexibility to improve ecological conditions.

Objectives

Fish and Salinity

The first study objective was to determine the extent to which species composition and abundance were influenced by salinity variability in the northeastern Florida Bay study area. Because of direct and indirect salinity influences, more variable fish abundances and distinct community differences were expected at the upstream locations over an annual cycle that included both wet and dry seasonal differences in freshwater inflow. The eastern portion of the study area was also expected to be distinctly more variable than the western portion because of the influence of the C-111 Canal.







13


Habitat Features

Because salinity is not the only feature of the habitat that varies along the complex environmental gradient within the study area, it was also necessary to consider other features of the fixed and moveable habitat (Browder & Moore 1981) as potentially influencing fish community structure. The second study objective was to determine important habitat features that influence the abundance and species composition of mangrove fish communities and compare these features across the salinity gradient. Fixed habitat structural features such as mangrove tree height and prop root density, environmental features such as water temperature, and fish diet and predation, were expected to influence the differences among fish assemblages across environmental gradients.

Study Area

The 250 km2 study area, located in extreme northeastern Florida Bay, consists of a series of shallow bays and ponds (less than 1.0 m in depth) bordered by mangroves. The upstream portion of the area is subject to freshwater inflow from seven mangrove-lined tributaries originating in the Taylor Slough/C-lll drainage basin.

In this region of Florida Bay, rapid ecological changes can take place when salinity variations occur suddenly, as at the start of the rainy season (Montague et al. 1989). Because tidal influences are almost negligible in the northeastern Florida Bay area, salinity changes are caused






14

by the variations in rainfall and subsequent freshwater flowing south through the tributaries, and variations in wind speed and direction. The rate and degree of salinity change are relatively unpredictable and can be rapid (hours) or slow (days) depending on changes in the weather.

Fish Community Sampling Design

To monitor fish community changes across the dynamic salinity gradient in northeastern Florida Bay, a balanced two-way analysis of variance (ANOVA) design was used, with two systems, each composed of three salinity regimes (Figure 1-2). Generally, upstream locations included one of the creeks which carries freshwater from the Taylor Slough/C-111 Basin, an interior bay downstream from the creek but still measurably affected by freshwater inflow, and an outer bay much less affected by freshwater inflow but more by marine influences. Specifically, the locations were as follows: upstream sites were located in Highway Creek and Long Sound in the eastern system and Snook Creek and Joe Bay in the western system (Figure 1-2); midstream sites were located in Little Blackwater Sound in the eastern system and the Trout Cove area in the western system; downstream sites were located in Blackwater Sound in the eastern system and Buttonwood Sound in the western system.

























Figure 1-2. Map of the Florida Bay study area.








16



















S Snook Creek .... Joe Bay


Trout
Cove






Blackwater Sound




Study Ates




Buttonwoodg fS. Sound
2. 25 km Sg: . Key Largo Ranger Statio















CHAPTER 2
FISH DENSITIES AND
ASSEMBLAGE PATTERNS IN MANGROVE HABITATS:
COMPARISONS ACROSS SALINITY GRADIENTS



Fishes tolerate salinities within a range of

survivability (Moyle & Cech 1988). If suitable conditions are not available within their environment, fish will experience stress, as evidenced by metabolic inefficiency and, in extreme cases, death (Moyle & Cech 1989). In general, fewer species of all faunal taxa are able to tolerate conditions in zones with salinity conditions typical of the upper estuary (Remane & Schlieper 1971). This may explain the occurence of lower numbers of fish species that occupy such areas (Deaton & Greenberg 1986).

As an alternative strategy to permanent occupancy and metabolic adjustment, fishes can shift habitats when salinity levels generate stress (Moser & Gerry 1989). The occurence of a salinity gradient in the estuary provides the opportunity for fish to exploit different habitats and thereby avoid unsuitable salinities by movement (Weinstein 1979). By stimulating such movements, salinity conditions may contribute to spatial and temporal fluctuations in species composition and abundances.



17






18


Contrary to what might be expected based on such

physiological factors, however, abundance peaks for many estuaries occur in conjunction with the initiation of the period of maximum freshwater inflow, when salinity levels drop dramatically. Estuarine transient juveniles may constitute most of the individuals during these peak periods (Yanez-Arancibia et al. 1980, Bell et al. 1984, Pinto 1987). In some species, juvenile fishes may be capable of exploiting salinities at lower levels than adults (Gunter 1967, Moser & Gerry 1989). However, in environments with more stable salinities, estuarine transient juveniles can also be abundant (Little et al. 1988, Robertson & Duke 1990b). Thus, the role of seasonal changes in salinity on fish communities requires further exploration.

Based on investigations conducted near and within the northeastern Florida Bay study area, at the initiation of the rainy season (June), changes in salinity were expected to occur, expanding the zone of low salinity further downstream (Ginsburg 1956, Tabb et al. 1962, Lindall et al. 1973, Thayer et al. 1987). This zone of lower salinity was expected to persist after the end of the rainy season, as freshwater from the eastern Everglades gradually drained into Florida Bay.

Toward the goal of understanding the influence of

freshwater inflow on fishes, the objective of this portion of the study was to identify spatial and temporal patterns in fish assemblages across the salinity gradient and thereby







19


test the following hypotheses. First, temporal changes in fish densities were expected to occur in conjunction with salinity changes. Secondly, in areas where salinities were more variable, numbers of species of fishes were expected to be lower than more stable areas. Thirdly, a community of fishes including estuarine transient juveniles was expected to occur in the study area. Finally, relatively lower densities were expected to occur in the upstream locations as an function of variable salinity conditions.


Materials and Methods

Pilot Study

A six-month pilot study was conducted to determine the most effective methods for quantitatively sampling fishes in mangrove prop root habitat throughout northeastern Florida Bay. Absence of tidal exchange in the study area was an important factor in selecting methods. Both collecting and observational methods were explored. Collecting gear selected for preliminary testing included minnow traps, Caribbean fish traps, gill nets, pull-up nets and enclosure nets with rotenone. The two visual census methods tested were: 1) direct recording of fishes observed with mask and snorkel on underwater data sheets and, 2) underwater video taping. Two complementary methods were selected from those tested in an attempt to sample the entire fish community. These two methods were enclosure nets and direct visual observation.






20


Design of the Main Study

The climate of subtropical Florida is characterized by a relatively long and severe dry season (November through April) and a wet season (May through October). Thus, the sampling schedule included monthly sampling for a one year period to encompass the influence of changes triggered by seasonal climatic conditions.

To monitor fish community changes across the dynamic salinity gradient of northeastern Florida Bay, a balanced sampling design suitable for analysis of variance was used. The design consisted of two systems, each having three locations along the salinity gradient (Figure 2-1 and Table 2-1). Based on the pilot study, this geographic design was to encompass three regimes of salinity variability within each system:

Upstream: low mean / high variation; Midstream: mid mean / mid variation; Downstream: high mean / low variation. Enclosure Nets

From the pilot study, one collecting method proved to be superior to the others, both in terms of sampling the breadth of species at the sites and providing a quantitative sample of fish density. This was the enclosure net first used by Thayer et al. (1987) to sample mangrove shoreline fishes in western and central Florida Bay. This method was selected for targeting small benthic and water column fish in particular.







21









Highway C


: Snook Creek
Joe Bay g





Cove


Blackwater Sound







1 0 Snorkel Sites


0 2.25 km Sound








Figure 2-1. Northeastern Florida Bay study area with sampling stations indicated.










Table 2-1. List of stations used in sampling fish in mangroves using enclosure nets and visual census surveys.

Enclosure Net Visual Census Stations Stations
System Gradient General Location Number Name Number Name West Upstream Joe Bay 1 east 1 Snook Creek Pond 1 2 mid 2 Snook Creek Pond 2 3 west 3 Snook Creek Pond 3
4 west Joe Bay* Midstream Trout Cove 4 northeast 5 southeast Trout Cove
5 mid 6 Tern Key 6 southeast 7 Deer Key* 8 Duck Key
Downstream Buttonwood 7 northeast 9 Buttonwood Point Sound 8 mid 10 Whaleback Key 9 southwest 11 Unnamed Key 12 Key Largo Ranger Station East Upstream Highway Creek 10 east 13 Shark Pond 11 island 14 Highway Creek Big Island 12 west 15 Critter Pond* 16 northeast Long Sound* Midstream Little Blackwater 13 east 17 northeast L. Blackwater Sd.* Sound 14 mid 18 northwest L. Blackwater Sd.* 15 west 19 L. Blackwater Sd. island 20 south L. Blackwater Sd. Downstream Blackwater Sound 16 near Gilberts 21 trestle 17 mid 22 Gilberts 18 far 23 hydrostation 24 Bush Point
* Stations that were dropped from the analysis due to missing data as a result of numerous poor visibility days primarily in 1990.






23


Sites. Because enclosure net stations were located in open bays, sites were selected that were protected from prevailing direct winds. Such siting prevented the wind from pulling the bottom of the net off the substrate. Sites were further chosen to have between 20 and 100 cm mean water depth at the outer prop root edge. This criterion was intended to provide some uniformity among the sites in terms of volume of water enclosed in the net. At each site, a natural berm consisting of packed detritus approximately 15 cm high and 30 cm wide occurred along the landward edge. This berm was exposed at high tide and provided a bank beyond which fish could not escape when rotenone was applied within the net (see below).

Procedures. At the start of the study, three sites were selected in each of the six general locations. A maximum of three enclosure nets could be deployed at each location by two persons in a day if the nets were deployed no more than approximately 1.0 km from each other. Environmental measurements taken during each collection included water depth, salinity, temperature, wind speed, wind direction, and air temperature. Salinity and water temperature were measured with a calibrated electronic instrument (YSI Model 33 S-C-T meter). For salinities above 35 ppt, a calibrated hand-held refractometer was used.

For each net, two 30 cm wide paths were cut

perpendicular to the shoreline through the mangrove fringe back to the berm. The paths were cleared of bottom roots






24


and overhanging branches so that a person could walk up the path carrying one end of the 30 m long net.

The same site was sampled repeatedly throughout the

study unless "stress" was observed in submerged vegetation. For example, sites with clay sediments supporting seagrasses had suffered some visible damage (e.g. grass trampling) from the sampling procedure by the fourth month at three stations. To maintain consistency in types of habitat encompassed by the nets, at these three sites, one new path was cut so that an unimpacted site could be sampled adjacent to the old one. These minor site changes were taken into account in later calculations of net area sampled.

On the day of sampling, a 6.0 mm mesh nylon seine was deployed by two people who carried it, scrolled around two wooden dowels, to the mid-point between the two prepared paths. Starting 10 m from the edge, they waded in opposite directions parallel to the edge, unrolling the net, and then walked toward the mangroves and up the paths. The dowel end was pounded into the sediment at the landward end of the path and the lead line was pressed down into the sediments all around the bottom edge. The top edge of the net was hung over several PVC poles to prevent fish from jumping over the net. All three nets were set in similar fashion (Figure 2-2).

Liquid rotenone was then applied within the enclosed area to a final concentration of 5 mg L-1. Fish that immediately began to surface were collected using hand nets










25











... .... ..











..........


'"''"..................
....... ..... . . ... .........
........... .. ... X: ...... ' . r

.. . ... ....... ~~ ~ ~r~
....... .........ts ~::~~::~ ~ ~ ~ ~~ ...




""' " "' 30 mr














Fi ur 2 2. En lo ur n t ll st at on in ic ti g e dimensions.r- '~~~~.-






26


for 30 to 45 minutes by two persons. After repeating the process at the other two sites, and allowing the rotenone to dissipate (approximately 3 - 4 hours), a snorkeler retrieved sunken fish and a wader collected along the berm edge from within each enclosure. Since rotenone effectiveness is reduced with decreasing temperature (Neilson & Johnson 1983), in colder months, floating fish were also collected after leaving the nets up overnight.

Fish and invertebrates collected were initially placed on ice and then frozen. Later they were identified to species and measured to total and standard (or carapace) length.

Efficiency tests. Fish recovery efficiency was tested at least once at each location using a mark-recapture method and normal net procedures. To collect test fish, several minnow traps were placed inside the area to be enclosed on the day before net deployment. After the net was in place, the minnow traps were removed, cleared and the fish placed into buckets. Fish were measured, marked by fin-clipping, and returned to the enclosed net area in minimal time. At least 30 fish were used per net in the test. Visual Census

Based on the pilot study, direct recording of census

data on underwater paper was identified as the better visual method tested. Quality of video tape was inconsistent and often low under the variable turbidity conditions encountered.






27

Study Sites. For visual censusing, site selection criteria included adequate depth (20 to 100 cm), red mangrove dominance, and wind-protection. At 70 m in length, however, each transect encompassed a range of depths and other physical characteristics. Four sites were randomly selected for permanent sampling stations within each of the six general locations. These sites were chosen in the same general locations as the net sites but were more widespread than the net stations (Figure 2-1).

Procedures. To prepare permanent census stations,

mangrove edge transects were designated with flagging tape every 10 m along the 70 m edge. Physico-chemical variables were recorded in conjunction with each visual census (e.g. air temperature, salinity, wind conditions). Other variables measured included: 1) water depth was recorded at a permanent stake located in each transect; 2) abundance of submerged aquatic vegetation adjacent to the transects was noted using a scale of zero to three (abundant); and, 3) range of visibility was determined by using a white PVC pole set vertically into the mud and measuring the horizontal distance at which the pole became visible as one snorkeled towards it. If the visibility was less than 100 cm, the census was rescheduled. If visibility was initially poor, three attempts (on subsequent days) were made to conduct surveys. However, in some months it was impossible to conduct a visual census at particular sites.







28

To conduct the census, a snorkeler approached the

flagged edge and remained stationary under each flag for 30 seconds, an adequate period for recording observable fish. On underwater data sheets, they recorded the species, numbers and estimated sizes of fish observed. The census surveys were conducted by myself and one assistant. Each census consisted of four complete swims of each transect.

Efficiency tests. Efficiency tests were conducted for the visual census technique using a mark-recapture method. Large fish (18-25 cm) were caught using hook and line and smaller fish (less than 18 cm) were collected with minnow traps. Tags made of plastic waterproof tape in various colors and labeled with a unique code were used. Fishing line was securely fastened to the tag and the line was sewn with a small sewing needle through the flesh just under a fish's dorsal fin (for gray snapper and larger fish) or through the lower jaw (for smaller fish).

To conduct a test, block nets were used to enclose an area of mangrove shoreline of adequate size to accommodate at least two snorkeling stations. Tagged fish (six to nine larger fish and ten smaller fish) were placed inside the enclosure for several hours to allow them to acclimatize to the habitat. Snorkelers then carefully entered the enclosed areas and conducted a visual census of the site by recording the species, tag color and code of each fish they observed. Three such tests were conducted in the summer of 1990, at three separate stations.






29


Definitions

As presented in Chapter 1, three categories of residency are recognized by estuarine ecologists: residents, transient juveniles, and occasional visitors. In analyzing the community in northeastern Florida Bay, several sources were consulted for life history information on individual species to designate each by residency (e.g. Odum & Heald 1972, Lee et al. 1980, Yanez-Arancibia 1980, Robins et al. 1986). Without conducting specific gonad analysis to determine maturity (e.g. Robertson & Duke 1990b), unequivocal distinction between juveniles and adults in the transient category were not possible. In addition, life history information is sketchy for all species except for certain killifish. Thus, these designations are approximate and serve for discussion purposes only. Such designations were not used in statistical analyses of the fish community.

For purposes of detailed analysis, all fish were

assigned to one of three groups of species based on size, behavior and primary portion of the mangrove habitat occupied during the day. Forage fish were considered those species whose members were generally less than 15 cm in size. Two groups of forage fish occupied different portions of the mangrove habitat: benthic and water column. Benthic forage fish live in close association with the substrate and include such species as gobies, killifish and mojarras. Water column forage fish are exclusively schooling fishes, that occupy the upper water column habitats, including






30

anchovies and silversides. The third fish group, large roving fish, are generally greater than 15 cm in size and occupy both the bottom and water column locations. This group included such species as snook, tarpon, snappers, catfish, grunts and barracuda. Density determinations

To obtain densities for each enclosure net sample, abundance was divided by the area the net encompassed. Areas enclosed ranged from 72 to 196 m2 (mean = 119, sd 33.2).

For visual census samples, measurements of horizontal secchi distance and fringe width were used as radii in calculating the area observed (half the area of a circle). Because it was impossible to see and accurately identify small fish at a great distance, maximum radius values of

2.0, 3.0, or 4.0 m were applied to benthic, water column and large roving fish respectively. Thus, the areas sampled in the visual censuses ranged from 1.6 and 25.0 m2 at each of the eight stations along a 70 m long transect. Analysis Methods

Temporal patterns in density. To initially inspect the data for patterns, salinity and density by fish group were graphed. Temporal patterns were examined graphically and quantitatively. For each of the three fish groups, and the three species that were most abundant within each group,







31

correlations were calculated between monthly averages of these fish densities and corresponding values for salinity, water temperature, and water depth.

Spatial patterns in density. To discern spatial

patterns in distribution for each fish group and the top 3 species within each group, repeated measures analyses of variance (ANOVA) with multiple comparison tests were used (SAS GLM procedure). All density data was effectively normalized by log-transformation. In the initial ANOVA model, density of fish was the dependent variable and gradient position (up-, mid- and downstream) and system (east and west) were the independent variables. To explore the relative density patterns among the general locations, a second ANOVA with general location as the independent variable was conducted. Further analysis was conducted to determine if spatial variation in certain environmental parameters might indicate why the densities varied among the stations. The average values of fish density, water depth, and salinity for each station (n = 18 stations) were calculated. Additionally, the amount of salinity variation over time at a particular station was also calculated by determining the standard deviation of salinity. Correlations between average fish densities and the means for these parameters were then determined.

Community patterns. For comparisons among general locations, an index of species richness (Odum 1983) was calculated with total number of species as the numerator and






32

log-transformed abundances as the denominator. Actual fish assemblage patterns were compared to gradient positions using cluster analysis. Data for each station, date and species were used to form matrices of stations based on similarity values (SAS CLUSTER procedure). An average linkage method was used to join clusters of stations. The resulting dendrograms were compared with the gradient positions. Those stations that were placed in a group other than the correct up-, mid- or downstream position, were denoted as misclassified. A second analysis was conducted on the log-transformed densities at each station in order to more thoroughly explore the data.

Results

Tests of Recovery Efficiency

Results of recovery efficiency tests for both the enclosure net and visual census techniques measured the number of fish sampled out of the total that were at a site (Table 2-2). However, no estimate is available for either method for sample accuracy, i.e. for how many fish escaped the area as the net was being deployed or the observer approached the area.

Enclosure net efficiencies. In all tests spanning up-, mid-, and downstream locations, 492 fish were marked. Of 14 total species, 60% of the fish used in the tests were goldspotted killifish (Floridichthys carpio). An average of 18% of all fish were recovered in the initial dip-net collections. By adding the same-day snorkeling procedure,












Table 2-2. Efficiency test results for enclosure nets and visual census sampling obtained by mark-recapture tests.



Method Number Size Classes Number Mean Standard Deviation
of Trials Total Length of Fish Percent Percent Efficiency (centimeters) Tagged Efficiency Enclosure net 18 2.5 to 7.5 467 36 38 18 7.5 to 15 25 68 31

1 15 to 25 6 100 Not applicable Visual census 3 4 to 7 31 29 2 3 18 to 25 22 86 12






34


efficiency increased by 7%. The total mean recovery rate was increased to 37% by leaving the nets up overnight and collecting the next day.

Overall, a greater percentage of larger fish were recovered than smaller (Table 2-2). Of six large fish (Lutjanus griseus) that were tagged, all were recovered after rotenone application. Twice as many mid-sized as small fish were recovered.

Visual census efficiencies. Individual test results for small fish (all Floridichthys carpio) ranged from 25 to 27% efficiency for the visual censusing method (Table 2-2). For large fish (all Lutjanus griseus), results ranged from 78 to 100%. Several tagged fish were observed more than one time during the four swims along the transect. Thus, when analyzing the data for each sample, to prevent counting the same fish more than once, after recording the first swim, only unique species and size classes of fish were added to the dataset for the second, third and fourth swims.

These efficiency analyses were intended to identify trends in fish recovery rates. Due to wide ranges in the test results, subsequent data analyses were not corrected for efficiencies.

Overall Abundance

Results of the visual census differed from enclosure net sampling results (Table 2-3). Enclosure net sampling resulted in the collection of 82,633 fish from 59 species and 29 families. The greatest abundance was collected at











Table 2-3. Number of fish collected using enclosure nets and observed during visual censuses. Explanation of group/residency given at end of table.


General Locations
Group/
Residency Total Up-west Mid-west Down-west Up-east Mid-east Down-east Family Species
Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual

Carcharhinidae (requium shark)
LR/r Carcharhinus leucas 2 1 1 2
Orectolobidae (nurseshark)
LR/r Ginglymostoma cirratum 4 4
Dasyatidae (stingray)
LR/r Dasyatis sabina 1 2 1 2 2
Elopidae (tarpon)
LR/o Megalops atlanticus 6 6
Anguillidae (freshwater eel)
LR/o Anguilla rostrata 1 1
Clupeidae (herring)
WC/o Opisthonema oglinum 3 1 1 2 7 WC/o Clupeid (species unk) 11 3 151 185 WC/o Harengula jaguana 3 000 3 00
Engraulidae (anchovy)
WC/r Anchoa mitchelli 335 41 40 2 100 454 2 17.770 4 18,805 143 WC/r Anchoa cayorum 5 5 Ariidae (sea catfish)
LR/r Arius fells 4 1 551 2 2 3 9 554
Batrachoididae (toadfish)
BF/r Opsanus beta 38 59 134 23 275 529 BF/r Porichthys plectrodon 1 1
Gobiesocidae (clingfish)
BF/r Gobiesox strumosus 1 1
Bythitidae (viviparous brotula)
BF/r Ogilbia cayorum 17 17 BF/r Gunterichthys longipenis 1 , 1 continued

tn











Table 2-3, continued

General Locations
Group/
Residency
Up-west Mid-west Down-west Up-east Mid-east Down-east Total Family Species
Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual

Belonidae (needlefish)
LR/r Strongylura notate 177 53 198 407 223 478 195 95 78 219 90 601 959 1,853
LR/r Strongylura timucu 1 1
Cyprinodonditae (killifish)
BF/r Floridichthyscarpio 1,792 678 5.011 1,294 2,183 1,095 2.127 1,473 1,056 227 849 1,475 13,018 8,240 BF/r Lucania parva 964 2,480 924 617 1,779 1.821 1,576 3,310 389 78 4,005 3,079 10,237 11,374 BF/r Cyprinodon variegatus 221 768 144 88 79 1 308 957 351 21 1,124 1.812 BF/r Fundulus grandis 309 252 99 15 97 7 65 31 147 76 793 305 BF/r Fundulus confluentus 98 63 91 16 47 40 199 30 481 103
BF/r Adinia xenica 25 21 46 BF/r Fundulus similis 2 1 10 13 BF/r Lucania goodei 1 1 BF/r Rivulus marmoratus 1 1
Poeciliidae (livebearers)
BF/r Poecilia latipinna 1,483 2,820 4,395 1,254 1,006 2.440 257 1,535 2,981 149 898 544 11,000 8,742 WC/r Gambusia sp. 26 407 325 937 293 2,921 31 148 583 210 649 449 1,907 5,132 WC/r Belonesox belizanus 4 1 2 2 8 1
Atherinidae (silverside)
WCIr Atherinomoruestipes 397 450 630 31,200 7,078 25,187 1.010 89 4,352 2,848 58,491 11,042 120,750 WC/r Atherinidae (genus unk) 900 14,581 887 272 1,681 2,369 0 20,690 WC/r Menidia sp. 2,242 543 175 3,093 357 256 1,071 1,815 325 200 178 3,222 4,348 9.129
WC/r Membras martinica 1 1
Syngnathidae (pipefish)
BF/r Syngnathus scovelli 43 83 19 12 8 53 80 213 83
BF/r Syngnathus florida 1 8 11 20 BF/r Micrognathus criniger 6 8 BF/r Hippocampus erectus 2 2 4 BFIr Hippocampus zoeterae 1 2 3 continued




01










Table 2-3, continued
General Locations
Group/
Residency
Up-west Mid-west Down-west Up-east Mid-east Down-east Total Family Species
Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual

Centropomidae (snook)
LR/o Centropomus undecimalis 2 1 2 25 1 3 24 5 53
Echeneididiae (remoras)
LR/o Echenes naucrates 1 I
Carangidae (jack)
LR/o Trachinotus goodei 1 I LR/o Naucrates ductor 2 2 LR/o Carangidae (ep unk) 4 4 LR/o Caranx hippos 1 6 1 1 2 11
LR/o Carangidae uv.) 2 2 LR/o Trachinotus falcatus 1 1
Lutjanidae (snapper)
LARIt Lutjanus jocu 23 12 12 27 74 LR/ta Lutjanus griseus 13 74 1 4,737 41 4,756 10 11 2 1,157 29 7,726 96 18,461 LR/t Lutjanus apodus 1 209 1 129 49 80 2 456
Gerreidae (mojarras)
BF/r Eucinostomus ap 4 2 13 19 BF/r Eucinostornusharengulus 349 32 212 128 67 448 396 135 505 458 68 882 1.597 2,081 BF/r Eugerres plumierl 427 302 37 2 88 212 152 12 704 528 BF/r Eucinostomus gula 2 1 241 1,289 104 1,739 1 30 184 33 2,350 410 5,564 BF/r Gerres cinereus 39 7 26 34 164 18 3 116 143 52 900 2086 1,238
Haemulidae (grunts)
LRAt Haemulidae (p unk) 1 1 LR/t Haemulon parral 26 102 11 139 LR/ta Haemulon sciurus 16 1,112 50 1,584 2,768
Sparidae (porgies)
LR/o Lagodon rhomboides 2 20 6 28 LRo Archosargus rhomboidalie 52 52 LR/ta Archosrgus probatocephalus 41 114 8 163 continued

-4










Table 2-3, continued

General Locations
Group/
Residency
Up-west Mid-west Down-west Up-east Mid-east Down-east Total Family Species
Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual Nets Visual


Lepisosteidae (gar)
LR/o Lepisosteusplatyrhincus 1I I
Centrarchidae (sunfish)
BF/r Lepomis macrochirus 7 1 20 1 27
Cichlidae (cichlid)
BF/r Cichlaesoma urophthalmus 85 387 7 977 90 1,062 484 BF/r Tilapia mariae 3 3 5 3 8
Ephippidae (spadefish)
LR/o Chaetodipterus faber 3 15 18
Scaridae (parrotfish)
LR/o Sparisoma radians 2 1 3 LR/o Scaridae (sp unk) 3 1 4
Lobotidae (tripletail)
LRio Lobotes surinamensis 2 1 3
Mugilidae (mullet)
LR/o Mugil cephalus 11 90 6 12 45 1 3 10 1 1211 29 1361
LR/o Mugil curema 1 22 1 24 LR/o Mugil liza 1 1 5 0 7
Sphyraenidae (barracuda)
LR/ta Sphyraena barracuda 12 1 17 196 18 185 3 1 29 90 25 352 104 825 Bleniidae (combtooth blenny)
BF/r Charsmodee saburrae 6 3 1 2 2 14
Gobiidae (goby)
BF/r Microgoblus gulosus 1,011 675 92 125 33 1,848 81 110 6 11 9 2,007 804
BF/r Goblosome robustum 235 239 1 59 534
BF/r Lophogoblus cyprinoides 151 18 1 18 44 30 3 1 200 68
BF/r Gobiosoma bosci 124 1 125 continued









Table 2-3, continued

General Locations
Group/
Residency
Up-west Mid-west Down-west Up-east Mid-east Down-east Total Family Species
Nets visual Nets visual Nets visual Nets Visual Nets Visual Nets visual Nets visual


Acanthuridae surgeonfishh)
LR/o Acanthurus chirurgus 9
Balistidae (leatherjackets)
LR/o Aluterus scriptus 1 Soleidae (sole)
BF/r Trinectes maculatus 7 1 28 36 Tetraodontidae (puffer)
LR/o Sphoeroides spengleri 1 2 Diodontidae (spiny puffer)
LRIo Chilomryterus schopfi 1 1 LR/o Diodontidae (species unk) 1 14 15



Total 10,624 11,185 12,723 00,275 13,712 44,703 9,547 11,301 25,115 9,280 10.912 886,212 82.833 222,960 No. Species 37 27 22 27 32 33 32 26 36 18 35 31 50 51 No. Samples 36 35 36 44 36 50 36 33 36 23 36 48 216 233

* Groups: BF = Benthic forage fish WC = Water column forage fish LR = Large roving fish

/Residency: r = resident
o = occasional visitors t = estuarine transient juvenile ta = estuarine transient juvenile (also present as adults)






40

Little Blackwater Sound, the midstream-east location. The greatest number of species, however, was found in samples from Joe Bay, the upstream-west location.

Visual census sampling resulted in observation of

222,960 fish from 51 species and 31 families (Table 2-3). Greatest abundance and greatest number of species were observed in samples taken in Blackwater Sound.

Samples obtained by the two methods differed in

relative abundance and numbers of species within these three fish groups (Table 2-4). For example, many more species of benthic forage fish were collected in the enclosure nets

(33) than were observed in the visual census (16). In contrast, many more large roving fish species were sampled in the visual censuses (29) than in the enclosure nets (17). Temporal Patterns in Density by Fish Group

Benthic forage fish. In Figure 2-3, one can compare changes in salinity with changes in density from the enclosure net sampling; however, no consistent patterns emerge. Great density variations occur independently of salinity changes. Salinity varied widely over the study period at the upstream/east (0.0-39.0 ppt) and upstream/west (13.0-58.0 ppt) locations. Salinity also ranged widely at the midstream/east location (19.5-50.0 ppt). However, at the other three locations (downstream/west, downstream/east and midstream/west), salinity remained high (29.8 to 54.0 ppt) throughout the study. Not only was the period of low salinity longer in the upstream/east location, but also, a







41







Table 2-4. Summary of abundances and number of species by method of collection and fish group.


Fish Methods
Group

Parameter Enclosure Visual Nets Census


Benthic Forage Total 45,458 39,476
Fish No. Species 33 16


Water Column Total 35,926 156,610 Forage Fish No. Species 9 6


Large Roving Total 1,249 26,874
Fish No. Species 17 29



All Fish Total 82,633 222,960 No. Species 59 51








Benthic Forage Fish

Density & Salinity vs. Month
Enclosure Nets
West East 60
Up Up
45

6 30

3 15




_Q 9-/ \ 45 Q) (D 60 60



S30
3i15 )











M0y89 May90 May89 May90


Figure 2-3. Density of benthic forage fish collected with enclosure nets in mangrove habitats (histogram) and corresponding salinity measurements (lines). Error bars indicate standard deviation among the three enclosure nets deployed in each general location.
location. h






43


substantial decrease (from 35.0 ppt to 10.4 ppt) was evident in June 1990, that did not occur in stations sampled in the upstream/west location (which became increasingly hypersaline).

None of these salinity changes correspond with patterns observed for fish densities. Density of benthic forage fish peaked in winter months at four of the six general locations (Figure 2-3). The highest density collection (13.6 fish m2) was at the mid-Trout Cove station in winter 1989; lowest density occurred at the mid-Little Blackwater Sound in the June 1989 (0.12 fish m-2).

Water column forage fish. In Figure 2-4, one can

compare changes in salinity with changes in density from the enclosure net sampling; again, however, no consistent patterns emerge. Density of water column fish was highly variable and the graphs illustrate no consistent seasonal patterns. In general, either very low or very high densities of these schooling fishes were collected. The highest density collection (25.3 fish m-2) occurred at midLittle Blackwater Sound in September 1989. No water column forage fish were collected in several samples. As with the benthic forage fish, these density fluctuations were also not related to the seasonal fluctuations in salinity.

Large roving fish. In Figure 2-5, changes in salinity can be compared with changes in density for this group from the visual census sampling; again, however, no consistent temporal patterns emerge. In the upstream/west location,








Water Column Forage Fish


Density & Salinity vs. Month

Enclosure Nets
West East 16 60
Up Up
12 45
--*.
8 *' ' 30


4 .115

I o - -_* _'u
16 60
S Mid Mid 22.7 16.4

E 12 45 -o


3 0
Cr
4 15 1

(n

0) 16 60
Down 19.9 Down I

12 - 45


8 " * 30


15


0 sh m1 di J. ..... 05
May89 May90 May89 May90

Figure 2-4. Density of water column forage fish collected with enclosure nets in mangrove habitats (histogram) and corresponding salinity measurements (lines). Error bars indicate standard deviation among the three enclosure nets deployed in each general location.








Large Roving Fish

Density & Salinity vs. Season Visual Census
West East
1.6 60
Up Up
1.2
40

0.8

0.4

0.0 M 0 0 1.6 60 U
Mid 2.2
Q) (DMid
E 1.2 40 40

S 0.8
20 C
T 0.4 20
C 0
Q) 0.0 0

C 1.6 60
. Down Down
1.2
40
0.8

0.4

0.0 0
Spr89 Fall Spr90O Spr89 Fall Spr90

Figure 2-5. Density of large roving fish sampled by visual census in mangrove habitats (histograms) and corresponding salinity measurements (lines). Error bars indicate standard deviation among the visual censuses.







46

the visual census stations were notably fresher (1.4-50.0 ppt) than the corresponding net stations (13.0-58.0 ppt) (Figures 2-3 & 2-5). In this location, the visual census stations were located within the creek itself while the net stations were located in Joe Bay immediately at the creek mouth (Figure 2-1). These seemingly slight differences in location may have contributed to observed differences in salinity patterns for the visual and net stations.

For the visual census, the salinity patterns in both the eastern and western upstream locations were similar to each other except in spring 1990, when freshwater entered the upstream/east location reducing salinity to 30.0 ppt and the upstream/west location became strongly hypersaline (47.3 ppt) (Figure 2-5). The midstream/east location was overall more variable than the midstream/west location. At the mid/west and downstream visual census locations, salinity patterns were uniformly high.

Density of large roving fish ranged from zero at the uppermost upstream/west location in the spring, fall and winter of 1989, to 2.3 fish m-2 at Duck Key (midstream/west) in winter. From the graphs, it appears that changes in density of this group were independent of salinity changes (Figure 2-5).

Temporal correlations. No significant correlations between temporal changes in salinity or water depth, and temporal changes in fish densities were found (Table 2-5). However, changes in water temperature were correlated with








Table 2-5. Correlations between fish density for each month averaged over the stations and salinity, water temperature and water depth. Data for large roving fish were obtained by visual censuses. Data for benthic and water column forage fish were obtained with enclosure nets. All density data were converted to logarithms (log x + 1) prior to calculations. Significant (p < 0.05) correlations are underlined.


Species or Salinity Water depth Water Temperature
Category


correlation p-value correlation p-value correlation p-value


Benthic Forage +0.17 0.5769 +0.05 0.8721 -0.80 0.0010
Fish

Water Column +0.06 0.8437 +0.57 0.4350 -0.04 0.8986 Forage Fish

Large Roving +0.31 0.2996 -0.35 0.7145 -0.76 0.0028 Fish






48

density of benthic forage fish and large roving fish. In both cases, lower abundances occurred at higher water temperatures.

Spatial Patterns in Density by Fish Group

Spatial patterns in fish density varied among the fish groups (Figure 2-6). From these graphs, one can see that only the larger roving fish group seems to vary consistently along the salinity gradient, with much lower densities at the upstream locations.

Benthic forage fish analysis of variance. Results of the repeated measures ANOVA's by fish group differed among the fish groups (Tables 2-6 and 2-7). Neither gradient position nor system were important determinants of variation in densities among the stations for the benthic forage fish group (Table 2-6). Although densities tended to vary significantly from one general location to another, these variations were not systematic along the salinity gradient, as indicated by the significant interaction between gradient and system.

The mid/west general location had significantly greater densities than the other midstream location (Table 2-7, Figure 2-6). Other locations were intermediate and not significantly different from these two.

Water column forage fish analysis of variance. Again, although densities tended to vary significantly from one general location to another, these variations were not systematic along the salinity gradient, as indicated by the



























Figure 2-6. Mean density of fish by general location for each fish group. Error bars illustrate the magnitude of the standard deviation in density over all the months. Samples of benthic and water column forage fish taken with enclosure nets. Large roving fish were sampled with visual methods.








50






Density by General Location


Benthic Forage Fish

7 6

E 5


3

~ 2

1

0




Water Column Forage Fish

7

6
04 E 3
0 4 cL
3

2


0




1.20 Large Roving Fish

1.00


E 0.80
Cl
0) 0.60

S0.40

0.20 0.00
Up Mid Down Up Mid Down
West East









Table 2-6. Repeated measures analysis of variance with densities of fish as dependent variables and gradient and system as independent variables. Benthic and water column forage fish were collected using enclosure nets. Large roving fish were sampled using visual census techniques. Data were transformed to logarithms prior to performing calculations.

Source Benthic forage fish Water Column Forage Fish Large Roving Fish df* F p df* F p df* F p Between Stations:

Among gradient positions 2/12 0.16 0.8511 2/12 4.17 0.0421 2/12 6.30 0.0135 Among systems 1/12 3.39 0.0904 1/12 0.38 0.5514 1/12 0.22 0.6460 Gradient X System 2/12 7.56 0.0075 2/12 8.39 0.0052 2/12 0.65 0.5393 Within stations

Among months 10/120 9.83 0.0001 10/120 5.16 0.0001 4/48 2.17 0.0863 Month X Gradient 20/120 1.15 0.3131 20/120 4.92 0.0001 8/48 1.29 0.2707 Month X System 10/120 1.48 0.1627 10/120 1.5 0.1486 4/48 1.56 0.2015 Month X Gradient X System 20/120 1.07 0.3942 20/120 4.61 0.0001 4/48 0.97 0.4707


Multiple comparisons among Location Sign. Location Location Sign. Location Location Sign. Location means for all months: greater greater greater than than than Gradient positions No differences Down > Up Mid & Down > Up Systems No differences No differences No differences
*Source degrees of freedom / error degrees of freedom

('1









Table 2-7. Repeated measures analysis of variance with densities of fish as dependent variables and general locations as independent variables. Benthic and water column forage fish were collected using enclosure nets. Large roving fish were sampled using visual census techniques. Data were transformed to logarithms prior to performing calculations.
Source Benthic forage fish Water Column Forage Fish Large Roving Fish

df* F p df* F p df* F p Between Stations:

Among general locations 5/12 3.77 0.0277 5/12 5.1 0.0097 5/12 2.94 0.0586 Within stations

Among months 10/120 9.53 0.0001 10/120 5.16 0.0001 4/48 2.17 0.0863 Month X General locations 50/120 1.18 0.2360 50/120 4.11 0.0001 20/48 1.16 0.3249


Multiple comparisons among Location Sign. Location Location Sign. Location Location Sign. Location means for all months: greater greater greater than than than

2 > 5 5 & 3 > 4 & 2 No differences (others (others
inter- intermediate) mediate)
*Source degrees of freedom / error degrees of freedom

** General locations:
1 = Joe Bay, upstream/west 4 = Highway Creek, upstream/east
2 = Trout Cove, midstream/west 5 = Little Blackwater Sound, midstream/east
3 = Buttonwood Sound, downstream/west 6 = Blackwater Sound, downstream/east

(31






53

significant interaction between gradient and system for density of water column fish (Table 2-6). Little Blackwater Sound (mid/east) and Buttonwood Sound (down/west) had significantly greater densities than Highway Creek (up/east) and Trout Cove (mid/west) (Table 2-7, Figure 2-6).

Large roving fish analysis of variance. In contrast to the other two groups, for large roving fish, a clear effect of gradient position on fish density occurred. Fish in this group were significantly less abundant at the upstream gradient locations than mid- or downstream (Table 2-6). No general locations varied significantly from the others (Table 2-7).

Spatial correlations. To analyze spatial trends, correlations between mean fish densities and salinity, salinity variation, and water depth were determined (Table 2-8). A significant correlation between average density of large roving fish and station salinity was found; lower densities occurred at stations with lower mean salinity levels and greater temporal variability. In addition, both water column forage fish and large roving fish were significantly more abundant at stations with deeper water. Temporal Patterns in Density by Species

As indicated by the correlations between mean densities and salinity, water depth and water temperature, temporal patterns differed among the species (Table 2-9). No significant correlations were found between salinity changes from month to month and densities for any species. Temporal








Table 2-8. Correlations between fish density for each station (n=18) averaged over the months and salinity, temporal standard deviation of salinity and water depth. Density data were converted to logarithms (log x + 1) prior to calculations. Significant (p<0.05) correlations are underlined. Benthic and Water column forage fish were collected with enclosure nets. Large roving fish were sampled using visual techniques.


Category Salinity Temporal Standard Water Depth Deviation of Salinity

correlation p-value correlation p-value correlation p-value


Benthic Forage -0.08 0.7540 -0.02 0.9303 -0.12 0.6463
Fish

Water Column +0.41 0.0884 -0.38 0.1233 +0.48 0.0438 Forage Fish

Large Roving +0.56 0.0150 -0.54 0.0208 +0.54 0.0213 Fish












V-I








Table 2-9. Correlations between fish densities for each month (n=13) averaged over the stations and salinity, water temperature and water depth. Density data were converted to logarithms (log x + 1) prior to calculations. Significant (p<0.05) correlations are underlined.


Species Method Salinity Water depth Water Temperature


correlation p-value correlation p-value correlation p-value Floridichthys carpio Nets +0.15 0.6117 +0.26 0.3681 -0.50 0.0788 Lucania parva Nets -0.06 0.8420 -0.62 0.0237 -0.69 0.0084 Poecilia latipinna Nets +0.49 0.0888 +0.09 0.7756 -0.72 0.0056 Anchoa mitchelli Nets +0.04 0.8961 +0.27 0.3620 +0.24 0.4227 Menidia spp. Nets +0.18 0.5638 -0.33 0.2765 +0.51 0.0740 Atherinomorus stipes Nets -0.14 0.6449 +0.56 0.0470 -0.14 0.6259 Lutjanus griseus Visual +0.47 0.1039 -0.43 0.1439 -0.92 0.0001 Strongylura notata Visual -0.35 0.2382 +0.76 0.0026 -0.56 0.0482 Haemulon sciurus Visual -0.04 0.8931 +0.56 0.0475 +0.62 0.0233





oi







56


patterns in water temperature were significantly correlated with densities of several species. Lucania parva, Poecilia latipinna, Lutjanus griseus, and Strongylura notata were less abundant when the water temperatures were higher. In contrast, greater abundances of Haemulon sciurus were observed in warmer months. Periods of higher water levels in the study area (e.g. late fall) corresponded to periods when greater densities of Atherinomorus stipes, Strongylura notata, and Haemulon sciurus were collected. In contrast, Lucania parva was in greater abundance during low water periods.

Spatial Patterns in Density by Species

Analyses of variance. Density patterns varied for the top three species of benthic forage fish (Figure 2-7). Results of repeated measures analyses of variance also differed among these species (Tables 2-10 and 2-11). Poecilia latipinna was more abundant at the midstream locations, particularly Trout Cove (mid/west). Distributions of Floridichthys carpio and Lucania parva were not significantly influenced by gradient position or system. However, Floridichthys carpio was more abundant at Trout Cove than all other locations and Lucania parva was most abundant at Blackwater Sound (mid/east).

The top three water column forage fish species differed in spatial distribution (Figure 2-8). Repeated measures ANOVA results also varied among these species (Table 2-12 and 2-13). Distribution of the silversides differed























Figure 2-7. Mean density of fish by general location for the three most abundant species in the benthic forage fish group. Error bars illustrate the magnitude of the standard deviation in density over all the months.













Density by General Location

Top 3 Benthic Forage Fish Species West Enclosure Nets East
2.5 - Up

2.0 1.5

1.0 0.5 a, 0.0 E 2.5 Mid Mid Q 2.0 L.
S1.5

1) 1.0

0.5 0.0 r-
V) 2.5
Down Down
2.0 1.5 1.0 0.5

0.0
Floridichthys Lucania Poecilia Floridichthys Lucania Poecilia
carpio parva latipinna corpio parva latipinna


Ln









Table 2-10. Repeated measures analysis of variance with density of benthic forage fish as the dependent variable and gradient and system as the independent variables. Samples were taken within mangrove habitats using enclosure nets. Densities were transformed to logarithms prior to calculations.


Floridichthys carpio Lucania parva Poecilia latipinna

Source df* F p df* F p df* F p Between Stations:

Among gradient positions 2/12 3.29 0.0726 2/12 12.96 0.0010 2/12 5.96 0.0160 Among systems 1/12 33.49 0.0001 1/12 2.23 0.1610 1/12 6.00 0.0306 Gradient X System 2/12 21.21 0.0001 2/12 5.09 0.0251 2/12 1.71 0.2229 Within Stations:

Among months 10/120 7.43 0.0001 10/120 7.70 0.0001 10/120 5.73 0.0002 Month X Gradient 20/120 3.62 0.0001 20/120 1.99 0.0317 20/120 2.73 0.0002 Month X System 10/120 1.43 0.1777 10/120 1.54 0.1712 10/120 1.28 0.2827 Month X System X Gradient 20/120 2.14 0.0072 20/120 1.45 0.1561 20/120 0.93 0.5113


Multiple comparisons among Location Sign. LocationLocation Sign. Location Location Sign. Location means for all months: greater greater greater than than than Gradient positions No differences Down > Up & Mid Mid > Up & down

Systems West > East No differences West > East
*Source degrees of freedom / error degrees of freedom

('1
%D0









Table 2-11. Repeated measures analysis of variance with density of benthic forage fish as the dependent variable and general location as the independent variables. Samples were taken within mangrove habitats using enclosure nets. Densities were transformed to logarithms prior to calculations.

Floridichthys carpio Lucania parva Poecilia latipinna


Source df* F p df* F p df* F p


Among general locations 5/12 16.50 0.0001 5/12 7.67 0.0019 5/12 4.26 0.0184 Among months 10/120 7.43 0.0001 10/120 7.70 0.0001 10/120 5.73 0.0002 Month X general location 50/120 2.59 0.0001 50/120 1.68 0.0314 50/120 1.72 0.0426


Multiple comparisons General Sign. General General Sign. General General Sign. General
among means for location gr. location location gr. location location gr. location
all months ** than ** than ** than


2 > All 6 > All 2 > All


3&4 > 5&6
(others
intermediate)


* source degrees of freedom / error degrees of freedom

** General locations:
1 = Joe Bay, upstream/west 4 = Highway Creek, upstream/east
2 = Trout Cove, midstream/west 5 = Little Blackwater, midstream/east 3 = Buttonwood Sound, downstream/west 6 = Blackwater Sound, downstream/east

























Figure 2-8. Mean density of fish by general location for the three most abundant species in the water column forage fish group. Error bars illustrate the magnitude of the standard deviation in density over all the months.










Density by General Location
Top 3 Water Column Forage Fish Species East Enclosure Nets West
4
Up Up



2






(D Mid Mid
L
3

(1) 2



aC 1
0

4

Down Down








Anchoa Atherinomorus Menidia Anchoo Atherinomorus Menidia mitchelli stipes spp. mitchelli stipes spp.









Table 2-12. Repeated measures analysis of variance with density of water column forage fish as dependent variables and gradient and system as independent variables. Samples taken within mangrove habitats using enclosure nets. Densities were transformed to logarithms prior to calculations.


Anchoa mitchelli Atherinomorus stipes Henidia spp.

Source df* F p df* F p df* F p Between Stations:

Among gradient positions 2/12 10.95 0.0020 2/12 116.61 0.0001 2/12 13.44 0.0009 Among systems 1/12 13.41 0.0031 1/12 20.32 0.0007 1/12 1.62 0.2272 Gradient X System 2/12 11.32 0.0017 2/12 2.81 0.1001 2/12 0.51 0.6136 Within Stations:

Among months 10/120 4.73 0.0060 10/120 8.54 0.0001 10/120 3.05 0.0065 Month X Gradient 20/120 4.95 0.0007 20/120 5.46 0.0001 20/120 2.62 0.0033 Month X System 10/120 4.95 0.0047 10/120 2.46 0.0389 10/120 4.44 0.0006 Month X System X Gradient 20/120 4.43 0.0015 20/120 3.00 0.0031 20/120 2.70 0.0025


Multiple comparisons among Location Sign. Location Location Sign. Location Location Sign. Location means for all months: greater greater greater than than than Gradient positions Mid > Up & down Down > Up & Mid Up > Mid & down


Systems East > West West > East No diffs > East

*Source degrees of freedom / error degrees of freedom


a%









Table 2-13. Repeated measures analysis of variance with density of water column forage fish as the dependent variable and general location as the independent variable. Samples were taken in mangrove habitats with enclosure nets. Density data were transformed to logarithms for calculations.

Menidia spp. Atherinomorus stipes Anchoa mitchelli


Source df* F p df* F p df* F p


Among general locations 5/12 5.90 0.0056 5/12 51.83 0.0001 5/12 11.59 0.0003 Among months 10/120 3.05 0.0065 10/120 8.54 0.0001 10/120 4.73 0.0060 Month X general location 50/120 3.02 0.0001 50/120 3.68 0.0001 50/120 4.74 0.0001


Multiple comparisons General Sign. General General Sign. General General Sign. General
among means for location gr. location location gr. location location gr. location
all months ** than ** than ** than


1 > All 3 > All 5 > All except
4 6 > All
others





* source degrees of freedom / error degrees of freedom

** General locations:
1 = Joe Bay, upstream/west 4 = Highway Creek, upstream/east
2 = Trout Cove, midstream/west 5 = Little Blackwater, midstream/east 3 = Buttonwood Sound, downstream/west 6 = Blackwater Sound, downstream/east






65


significantly among the gradient positions. Atherinomorus stipes, the hardhead silverside, was more abundant downstream; Menidia spp. was more abundant upstream. Individuals of both Menidia beryllina and Menidia peninsulae were collected. The distribution of these species overlaps in northeastern Florida Bay, and distinctive characters are extremely difficult to confirm (C. Gilbert, personal communication). Thus, Menidia spp. has been used in this study to designate these species. Anchoa mitchelli, although not influenced by gradient or system, was significantly more abundant at Little Blackwater Sound (mid/east) than at the other general locations.

Patterns varied in spatial distributions for the top

three species of large roving fish (Figure 2-9). For these species, repeated measures ANOVA results also varied (Table 2-14 and 2-15). Among these species, Haemulon sciurus was never present upstream. Lutjanus griseus was significantly less abundant up- than mid- or downstream. In contrast, Strongylura notata had significantly greater densities upstream/east.

Correlations. To further analyze spatial trends for

these nine species, correlations between mean densities and salinity, salinity variation, and water depth were determined and are presented in Table 2-16. Densities of Menidia spp. were greater at locations with lower mean salinities and greater variation. In contrast, densities of Atherinomorus stipes were greater at locations with higher





























Figure 2-9. Mean density of fish by general location for the three most abundant species in the large roving fish group. Error bars illustrate the magnitude of the standard deviation in density over all the months.












Density by General Location Top 3 Large Roving Fish Species West Visual Census East

0.6 Up Up


0.4 0.2

0 . 0.00 o.o
4. 0.00
0)
E 0.0 Mid 1.30 Mid 0.95



3 0.4 CU)
c

L 0.2 C 0.0


0.6 Down Down 1,2


0.4 0.2


0.0
Haemulon Lutjanus Strongylura Haemulon Lutjanus Strongylura
sciurus grnseus notato sciurus grseus notata









Table 2-14. Repeated measures analysis of variance with density of large roving fish as dependent variables and gradient and system as independent variables. Samples taken by visual census methods along mangrove shoreline habitats. Densities transformed to logarithms prior to calculations.

Haemulon sciurus Lutjanus griseus Strongylura notata
df* F p df* F p df* F p Source

Between Stations:

Among gradient positions 2/12 2.36 0.1367 2/12 5.67 0.0185 2/12 3.52 0.0626 Among systems 1/12 0.13 0.7278 1/12 0.20 0.6654 1/12 4.24 0.0618 Gradient X System 2/12 0.04 0.9650 2/12 0.48 0.6317 2/12 3.32 0.0714 Within Stations:

Among seasons 4/48 0.29 0.8794 4/48 9.22 0.0001 4/48 2.97 0.0416 Season X Gradient 8/48 0.73 0.6659 8/48 2.64 0.0309 8/48 2.95 0.0168 Season X System 4/48 1.97 0.1167 4/48 2.59 0.0667 4/48 0.76 0.5292 Season X System X Gradient 4/48 1.10 0.3794 4/48 1.51 0.2032 4/48 0.43 0.8603 Multiple comparisons among Location Sign. Location Location Sign. Location Location Sign. Location means for all seasons: greater greater greater than than than Gradient positions No differences Mid & > Up No differences Down
Systems No differences No differences No differences

*Source degrees of freedom / error degrees of freedom


CO









Table 2-15. Repeated measures analysis of variance with density of large roving fish as the dependent variable and general location as the independent variable. Samples were taken by visual census along mangrove shoreline habitats. Densities were transformed to logarithms prior to calculations.

Haemulon sciurus Lutjanus griseus Strongylura notata
df* F p df* F p df* F p Source


Among general locations 5/12 1.03 0.4442 5/12 2.26 0.0765 5/12 3.63 0.0311

Among seasons 4/48 0.29 0.8794 4/48 9.22 0.0001 4/48 2.97 0.0416

Season X general location 20/48 1.31 0.2190 20/48 1.96 0.0480 20/48 1.51 0.1489

Multiple comparisons General Sign. General General Sign. General General Sign. General
among means for location greater location location greater location location greater location
all seasons ** than ** ** than ** ** than **
General locations No differences No differences 4 > 1 & 3 (others
interm)
* source degrees of freedom / error degrees of freedom

** general locations:
1 = Joe Bay, upstream/west 4 = Highway Creek, upstream/east
2 = Trout Cove, midstream/west 5 = Little Blackwater, midstream/east 3 = Buttonwood Sound, downstream/west 6 = Blackwater Sound, downstream/east









%Di









Table 2-16. Correlations between density for each station (n=18) averaged over all months and salinity, temporal standard deviation of salinity, and water depth. Density data were converted to logarithms (log x + 1) prior to calculations. Significant (p<0.05) correlations are underlined.


Species Method Salinity Temporal Standard Water Depth Deviation of Salinity


correlation p-value correlation p-value correlation p-value Floridichthys carpio Nets +0.07 0.7655 -0.12 0.6446 -0.60 0.0079 Lucania parva Nets +0.13 0.6096 -0.44 0.0665 +0.62 0.0064 Poecilia latipinna Nets +0.40 0.0935 -0.28 0.2429 -0.20 0.4165 Anchoa mitchelli Nets -0.03 0.9127 +0.14 0.5879 +0.05 0.8530 Menidia spp. Nets -0.52 0.0253 +0.69 0.0013 -0.01 0.9634 Atherinomorus stipes Nets +0.60 0.0087 -0.72 0.0007 +0.45 0.0626 Lutjanus griseus Visual +0.58 0.0108 -0.58 0.0123 +0.53 0.0226 Strongylura notata Visual -0.11 0.6595 -0.06 0.8234 +0.22 0.3765 Haemulon sciurus Visual +0.37 0.1300 -0.38 0.1226 +0.26 0.2942






-J






71


mean salinities and less variation (Table 2-16). Lutjanus griseus also appeared to avoid the lower salinity, more variable areas. Floridichthys carpio was significantly more abundant at shallower stations in the study area. Species correlated with deeper waters were Lucania parva, Lutjanus griseus and possibly, Atherinomorus stipes. Community Patterns

Species richness. Species richness differed among the fish groups and for all fish combined (Table 2-17). A total of 305,589 fish from 77 species was sampled using both the visual census and enclosure net methods combined. Midstream locations in both the east and west systems had the lowest species richness of the three gradient positions. Among the fish groups, benthic forage fish were also least species rich at midstream locations. Water column forage fish had distinctly lower species richness at Little Blackwater Sound, with only four species but great abundances. For large roving fish, upstream and midstream locations were lowest. For number of species alone, large roving fish species followed a clear gradient from upstream (10 and 11 species) to midstream (13 and 14 species) to downstream (20 and 22 species).

Cluster analysis. Results of two cluster analyses for species collected with enclosure nets were illustrated using dendrograms (Figure 2-10). The initial analysis classified stations based on presence of species. Most of the upstream









Table 2-17. Species richness index for fish sampled with enclosure nets and visual census. Species Richness Index = (Number of species - 1) / log (Total individuals) (Odum 1983).




General Benthic Forage Fish Water Column Forage Fish Large Roving Fish All Fish Location
Total Number Index Total Number Index Total Number Index Total Number Index
Indiv. of Indiv. of Indiv. of Indiv. of
Species Species Species Species




Up-west 15,951 24 5.5 5,408 6 1.3 450 11 3.8 21,809 64 14.5 Mid-west 16,049 14 3.1 51,052 4 0.6 5,897 15 3.7 72,998 49 9.9 Down-west 13,442 21 4.8 37,083 6 1.1 7,890 20 4.9 58,415 65 13.4

Up-east 15,716 21 4.8 4,805 6 1.4 327 10 3.6 20,848 58 13.2 Mid-east 7,434 19 4.6 25,222 7 1.4 1,739 13 3.7 34,395 54 11.7 Down-east 16,342 22 5.0 68,966 6 1.0 11,816 22 5.2 97,124 66 13.0



All 84,934 33 6.5 192,536 9 1.5 28,123 35 7.6 305,589 77 13.9























Figure 2-10. Cluster analysis dendrograms based on species collected using enclosure nets. Stations that grouped differently than actual gradient positions are designated as misclassified.
a. Presence of each species (each species weighted equally) was used in one
analysis and,
b. Density of each species (fish per square meter) was used in the other.










Enclosure Nets Presence Density








4


2
% S






0
gggggorlt 10 3 11 12 15 13 14(D() 17 16 18 7 8 4- 1 8 10 11 3 17 18 18 87 lo=ns Misclassified Up Mid & Down Mid Up Down (8 = Misclassified






75


stations, including one in Joe Bay and all in Highway Creek, grouped separately from those located mid- and downstream. In a second analysis based on densities of each species, all but one upstream station clustered separately from the midand downstream locations.

For the visual census, cluster analysis results were also graphed using dendrograms (Figure 2-11). Based on presence of species, all the upstream stations formed one cluster. Based on densities of all species, three of the five upstream stations clustered together. These were the three that were most upstream.

For each cluster group defined by the cluster analysis, the most common or dominant sets of species were identified. Three species that commonly occurred at all stations in the study area were: goldspotted killifish (Floridichthys carpio), rainwater killifish (Lucania parva), and redfin needlefish (Strongylura notata). Species that were very common in upstream stations included the inland silverside (Menidia spp.), clown goby (Microgobius gulosus), tidewater mojarra (Eucinostomus harengulus), striped mojarra (Eugerres plumieri), and Mayan cichlid (Cichlasoma urophthalmus). Downstream species commonly included hardhead silverside (Atherinomorus stipes), gray snapper (Lutjanus griseus), silver jenny mojarra (Eucinostomus gula), great barracuda (Sphyraena barracuda), blue-striped grunt (Haemulon sciurus), and snook (Centropomus undecimalis). Midstream
























Figure 2-11. Cluster analysis dendrograms based on species sampled by visual census methods. Stations that grouped differently than actual gradient positions are designated as misclassified.
a. Presence of each species (each species weighted equally) was used in one
analysis and,
b. Density of each species (fish per square meter) was used in the other.










Visual Census


Presence Density





4-L
4 4


0 0
I I o




Stations ' " =' * " " ' ' ' s r .13 1, 14 20 9 10 19 11 2) 3) 8 23 24 5 22 12 21
I I .... I ' ' ...... I~
Up Mid & Down Up Mid & Down

(X)I- Misclassified







78


stations had mixtures of both upstream and downstream species excepting the Mayan cichlid, which was only common upstream.

Residency. Abundance and numbers of species by

residency category varied among the general locations (Table 2-18). All 33 benthic forage fish species sampled were permanent estuarine residents. Among the water column forage fish, seven of the nine species observed were residents, while two (Clupeids) were occasional visitors. Large roving fish were represented in all three residency categories. Six species of large roving fish were permanent estuarine residents including needlefish, catfish, bull sharks, and stingrays. The vast majority of large roving fish were transient juveniles, however.

Frequency distribution by size varied for 6 transient species of large roving fish (Figures 2-12 and 2-13). Clearly, the mangrove habitats of northeastern Florida Bay were nurseries for Sphyraena barracuda most of which occurred in juvenile sizes (Figure 2-12). Snook, however, did not occur in juvenile size classes. Although adult-sized Lutjanus griseus appeared to share the habitat with largersized juveniles, since no gray snapper sampled was smaller than 7.5 cm, one can assume that young-of-the-year juveniles occur outside the mangrove habitats sampled in this study. Habitat use patterns similar to Lutjanus griseus were observed for Lutjanus apodus, Haemulon sciurus, and









Table 2-18. Comparison of abundance and number of species by residency, fish group, and general location. Both sampling methods were combined for the table.
Residency General Locations
Fish Group Up-west Mid-west Down-west Up-east Mid-east Down-east All

Residents: complete entire life cycle in the study area

Benthic forage fish abundance 15951 16049 13442 15712 6965 16329 84915 Benthic forage fish species 24 14 21 21 18 22 33

Water column forage fish abundance 4505 36460 36195 4532 23536 65843 171071 Water column forage fish species 5 4 5 5 6 5 7
Large roving fish abundance 236 603 1255 293 299 701 3387 Large roving fish species 3 1 4 2 2 4 6

Transient Juveniles: juvenile offspring of species that spawn offshore

Benthic forage fish abundance
Benthic forage fish species

Water column forage fish abundance
Water column forage fish species

Large roving fish abundance 101 5268 6470 25 1396 9832 23092 Large roving fish species 3 8 7 2 6 8 9

Occasional Visitors: marine and freshwater adults that occupy the study area

Benthic forage fish abundance
Benthic forage fish species
Water column forage fish abundance 3 1 1 2 603 610 Water column forage fish species 1 1 1 1 1 2
Large roving fish abundance 113 25 162 9 42 1263 1618 Large roving fish species 5 5 9 6 5 9 19
































Figure 2-12. Length-frequency histograms based on visual census data. Adult size given whenever the information was available from the literature. a. Sphyraena barracuda (great barracuda), and b. Centropomus undecimalis (snook).








81











Abundance by Size Class




Sphyraena
350
barracuda 280 210 140 S70 o- o
0

) Centropomus 40 - undecimalis



30



20

a) N
10 Cl)



O
0 15 30 45 60 75 90 105 120 135 150

Total length (cm)




























Figure 2-13. Length-frequency histograms based on visual census data. Adult size given whenever the information was available from the literature for a. Lutjanus griseus (gray snapper), b. Lutjanus apodus (schoolmaster) c. Haemulon sciurus (blue-striped grunt) d. Archosargus probatocephalus (sheepshead).







83

Abundance by Size Class



1E4 Lutjanus "
griseus
8000 6000 4000 2000

0
Lutjanus
300 - apodus


200


100 C)
0
c 1500 Haernulon





500


0
Archosargus
6o - probatocephalus


40


20 0
0 10 20 30 40 50 60 70 Total length (cm)







84


Archosargus probatocephalus, although adequate life history information is not available for these species to determine size at maturity (Figure 2-13).

Overall, the community was dominated by residents in both numbers of individuals (91%) and numbers of species (60%). Occasional visitors accounted for 28% of the species but less than 1.0% of the individuals. Estuarine transients (juveniles and adults) comprised 12% of the species and 8% of the abundance.

The distribution by residency varied among the

locations (Figure 2-14). In terms of abundance, permanent residents dominated at all six general locations. The trend was most pronounced upstream. Transient individuals were somewhat more prevalent at mid- and downstream locations. Although less than 2.0% of the abundance, occasional visitors comprised 15 to 21% of the number of species overall.

Discussion

Methods

Enclosure net methodology. Using the enclosure net method, direct sampling of fish occupying submerged red mangrove shorelines was possible. The fine mesh net captured the small forage fish that numerically dominated the habitats. Inevitable escapes, especially by large roving fish, probably occurred, however, as persons deploying the net waded up to the sites. The recovery efficiencies of fishes encircled in the nets were comparable























Figure 2-14. Comparison of residency among the general locations based on percent of total abundance. Data used in calculations were taken with enclosure nets (for benthic forage fish and water column forage fish) and by visual census for large roving fish. a. Abundance of fish in each category. b. Number of species in each category.







86



Residency


Occasional Transient Resident Abundance
100 80


c 60 Q)
L
O) 40


20 0

Number of Species 100


80


C 60 Q) 40


20



Up Mid Down Up Mid Down

West East







87


to those found by other investigators who targeted small fish (less than about 7.0 cm in total length) in vegetated, shallow areas using (Weinstein & Davis 1980). Although rotenone required cautious handling, fish immediately began re-occupying sites after nets were removed. Rapid rotenone degradation in the relatively high temperatures which prevailed in these waters (Nielson and Johnson 1983), and dilution due to the turnover of water at the relatively open sites prevented cumulative adverse rotenone effects.

Other factors could have affected the variable

efficiency of recovery from individual nets. Turbidity from disturbed sediments reduced the ability of the collectors to capture fish with dip nets. In addition, variation in winds onto the sites may have caused rotenone effectiveness to vary.

Visual census methodology. The visual census

methodology had several advantages for sampling fish in mangrove habitats. The speed and flexibility of the method permitted sampling in a broad range of habitats that could not be sampled with the net, including mangrove locations with greater depths and a wider fringe. The visual method was non-destructive, thus allowing evaluation of the persistence of use by repeated observations of the same stations and fishes. The problem of net avoidance was eliminated with this method. In addition, large fish (e.g. tarpon) that would normally escape nets and trawls could be surveyed. As in studies of streams and coral reefs in which







88

the visual method has increasingly been used, it has possible further value for use in studies on behavior, species interactions, and microhabitat use in mangrove habitats.

Direct observational sampling also has disadvantages. Somewhat surprisingly, many fish were attracted to the snorkelers (Dibble 1991). Fishes such as snook, tarpon, gray snapper, cichlids, bluegill, and killifish, frequently came within a few inches of the snorkeler's clipboard, presumably out of curiosity. As verified in the snorkeling efficiency tests, this attraction led to double-counting of individuals on the multiple swims.

A major disadvantage of the method was its sensitivity to reduced visibility conditions. Although the fishes approached more closely in low visibility situations (1.0 to

2.0 m), the uncertainty level in identification of species was often increased. Overall, those species that tended to remain at the bottom or far back in the fringe were surveyed less accurately in deeper and low visibility sites. This problem was a particular disadvantage for surveying the benthic forage fish at deeper stations. Even in depths of only 1.0 m, the use of SCUBA equipment might be an advantage if visibility range is less than water depth since it would permit better sampling of the bottom. Some small, rare fishes and shy cryptic forms were less accurately sampled with the visual method. Finally, the occurrence of large







89

dangerous predators may preclude visual sampling in certain areas. Sites in deep creeks were abandoned based on sitings of large sharks and alligators.

Combined methods Using two methods to complement each other in the same habitats and regions had great advantages. The two methods increased the range of mangrove shorelines that could be surveyed. However, when using two methods, interpretation problems can arise when neither is 100% efficient. If the results do not agree, it is difficult to determine if the discrepancies are due to differences in the fish sampled or due to differences in the efficiency of the methods. In this study, the two methods targeted different size groups of fishes. Overall, benefits of using the two methods clearly out-weighed the disadvantages. Fish and Salinity

Salinity. As confirmed in this study, salinity

conditions in the area vary from year-to-year (Ginsburg 1956). The east and west systems differed greatly during the drought conditions that prevailed during the main study year. They were very similar, however, during the pilot study year when rainfall was locally more plentiful. Under low rainfall periods, the C-1ll Canal may effectively block most freshwater from flowing into the western system by routing it towards the east. During high flow periods, however, when more freshwater is available for distribution, the east and west systems appear to have similar salinity patterns.







90

Historically, such annual differences due to local rainfall variations may have been more moderate. Before freshwater wetlands were extensively drained, freshwater probably gradually seeped into the study area from the greater Taylor Slough drainage basin, resulting in more dispersed distribution patterns and prolonged periods of lower salinity levels.

Temporal Patterns. The first hypothesis proved to be incorrect: none of the temporal patterns in density of any fish group or species examined was attributable to changes in salinity. In the study area, therefore, the fish do not seem to react to salinity changes by short-term movement in and out of the general locations (regions approximately 12 km2 in area). However, patterns for all species collected were not individually analyzed. Thus, some short-term relationships may be identifiable on further examination of the data.

Temporal patterns were related to temperature, however. For benthic forage and large roving fish, as temperature increased, density decreased. Similarly, Thayer et al. (1987) and Tabb et al. (1962) also found greater densities in western and central Florida Bay in the late fall and winter when temperatures were cooler. Temporal patterns for the current study are not typical for estuarine fish populations; usually peaks occur when freshwater inflow is greatest (Gunter 1967, Weinstein 1979, Yanez-Arancibia et al. 1980, Rogers et al. 1984, Stoner 1986, Flores-Verdugo






91

1990). Judging from the salinity data, this would have been late summer for northeastern Florida Bay in the study year.

One possible explanation for this unusual condition may be that in the summer, high temperatures combined with low circulation to create a stressful environment for fish in the Bay (Moyle & Cech 1988). This is supported by reports by fishermen of fish kills in Florida Bay during those hot summer months (M. Robblee, Everglades National Park, personal communication).

Additionally, the density of the major component of the large roving fish group, Lutjanus griseus, probably accounted for much of the temperature related trend in large roving fish densities overall. The larger individuals of this species, migrate offshore in the summer, when spawning occurs, and return in the winter (Starck & Schroeder 1971, Rutherford et al. 1989). This migration may account for the reduced densities of large roving fish in the summer.

Spatial patterns. The density of fish decreases from west to east in Florida Bay (Sogard et al. 1987, 1989a). This trend appears to continue into the northeastern Bay (this study, Funicelli et al. 1986). Using an almost identical method of sampling (enclosure nets), the mean fish density found in western and central Florida Bay mangroves by Thayer et al. (1987) was 8.0 fish m-2, compared to 3.3 fish m-2 found with nets for the northeastern Florida Bay area.




Full Text
119
may be due to the selection of habitats based on general
features (i.e. mangrove over seagrass habitats).
Although density of large roving fishes as a group were
correlated with the physical features, some species that
were categorized as "roving in the current study may not
actually roam among the mangrove habitats. Individuals of
these species appear to persist at particular locations for
long periods of time (based on limited observations of
tagged fish). Thus, while some large fishes (such as
mullet, catfish, barracuda and needlefish) may truly be
wanderers and display no discrimination among mangrove
shoreline habitats, other species (gray snapper and blue-
striped grunts) may maintain more permanent residency at
certain locations and display definite habitat preferences.


31
correlations were calculated between monthly averages of
these fish densities and corresponding values for salinity,
water temperature, and water depth.
Spatial patterns in density. To discern spatial
patterns in distribution for each fish group and the top 3
species within each group, repeated measures analyses of
variance (ANOVA) with multiple comparison tests were used
(SAS GLM procedure). All density data was effectively
normalized by log-transformation. In the initial ANOVA
model, density of fish was the dependent variable and
gradient position (up-, mid- and downstream) and system
(east and west) were the independent variables. To explore
the relative density patterns among the general locations, a
second ANOVA with general location as the independent
variable was conducted. Further analysis was conducted to
determine if spatial variation in certain environmental
parameters might indicate why the densities varied among the
stations. The average values of fish density, water depth,
and salinity for each station (n = 18 stations) were
calculated. Additionally, the amount of salinity variation
over time at a particular station was also calculated by
determining the standard deviation of salinity.
Correlations between average fish densities and the means
for these parameters were then determined.
Community patterns. For comparisons among general
locations, an index of species richness (Odum 1983) was
calculated with total number of species as the numerator and


102
Grossman et al. 1991). Principal components of the habitat
variables were calculated for both the visual census and the
enclosure net data sets using the SAS correlation matrix and
varimax rotation methods (Afifi & Clark 1984, Smith & Duke
1987). Each component was interpreted by examining
correlations between the original variables and the derived
components.
SAS FACTOR was then used to calculate individual
factors from the principal components for each observation
in the original data set. These uncorrelated factors were
used as the independent variables in a multiple linear
regression with log-transformed fish densities as the
dependant variables. Fish density data were divided into 3
groups of species for the analysis based on their size,
mobility and position in the mangrove habitat. These
categories were benthic forage fish, water column forage
fish, and large roving fish as defined in Chapter 2. All
species were also analyzed separately. Only principal
components with eigenvalues greater than the proportion of
the variance in the data that could be explained by an
individual variable (i.e. those > 1.0) were used in the
regressions (Grossman et al. 1991).
The slope of the regression line describes the nature
of the mangrove habitat and fish relationship. The variance
in fish density explained by each component (R2) indicates
the strength of this relationship and also provides a basis


Table 5-1. Summary of studies testing predation hypotheses in aquatic habitats
using tethering techniques.
Taxa Tethered
Hypothesis
Tested
Decapod
Crustaceans
Brittlestars
Fish
Mangrove Leaves
and Propagules
Comparison of predator
encounter rates among
macro-habitats
Heck & Wilson 1987
Wilson 1989
Aronson 1989
Shulman 1985
Mclvor & Odum 1988
THIS STUDY
Smith 1987
Prey vulnerability in
and out of vegetated
micro-habitats
Barshaw & Able 1990a
Barshaw & Able 1990b
Hay et al. 1989
Heck & Thoman 1981
Herrnkind & Butler 1986
Wilson 1989
Wilson et al. 1987
Wilson et al. 1990
Rozas & Odum 1988
Absolute rate of
predation
Robertson 1987
Smith 1987
Modification of prey
behavior by the
presence of predators
Power & Matthews 1983
Phillips & Swears 1979
Preference by predators
for a particular prey
Aronson 1988
Smith 1987




160
longer captured many fish, their use was abandoned. This
evidence indicates that the Ruppia community was probably
supporting a greater fish population in the area than
observed at any time during the course of the regular study.
Thus, northeastern Florida Bay, in the drought year
recorded in this study, was unusual in comparison to other
tropical, subtropical and warm temperate estuaries. Water
management efforts may be needed to restore sustained low
salinity periods, thereby inducing greater submerged aquatic
vegetation development and greater influx of estuarine
transient juveniles upstream into the more protected
habitats.


predation may inhibit recruitment and survival of post-
larval fishes from offshore. An unbroken continuum of good
habitat from outer to upper reaches may be
northeastern Florida Bay is to function as
area for estuarine transient fishes.
necessary if
a prime nursery
ix


71
mean salinities and less variation (Table 2-16). Lutjanus
griseus also appeared to avoid the lower salinity, more
variable areas. Floridichthys carpi was significantly more
abundant at shallower stations in the study area. Species
correlated with deeper waters were Lucania parva, Lutjanus
griseus and possibly, Atherinomorus stipes.
Community Patterns
Species richness. Species richness differed among the
fish groups and for all fish combined (Table 2-17). A total
of 305,589 fish from 77 species was sampled using both the
visual census and enclosure net methods combined. Midstream
locations in both the east and west systems had the lowest
species richness of the three gradient positions. Among the
fish groups, benthic forage fish were also least species
rich at midstream locations. Water column forage fish had
distinctly lower species richness at Little Blackwater
Sound, with only four species but great abundances. For
large roving fish, upstream and midstream locations were
lowest. For number of species alone, large roving fish
species followed a clear gradient from upstream (10 and 11
species) to midstream (13 and 14 species) to downstream (20
and 22 species).
Cluster analysis. Results of two cluster analyses for
species collected with enclosure nets were illustrated using
dendrograms (Figure 2-10). The initial analysis classified
stations based on presence of species. Most of the upstream


97
rates. This scenario may serve as an alternative or
complementary hypothesis to salinity intolerance, and lack
of access from distant passes, in explaining the lower
overall abundances of large roving fish observed in mangrove
habitats upstream.


84
Archosargus probatocephalus, although adequate life history
information is not available for these species to determine
size at maturity (Figure 2-13).
Overall, the community was dominated by residents in
both numbers of individuals (91%) and numbers of species
(60%). Occasional visitors accounted for 28% of the species
but less than 1.0% of the individuals. Estuarine transients
(juveniles and adults) comprised 12% of the species and 8%
of the abundance.
The distribution by residency varied among the
locations (Figure 2-14). In terms of abundance, permanent
residents dominated at all six general locations. The trend
was most pronounced upstream. Transient individuals were
somewhat more prevalent at mid- and downstream locations.
Although less than 2.0% of the abundance, occasional
visitors comprised 15 to 21% of the number of species
overall.
Discussion
Methods
Enclosure net methodology. Using the enclosure net
method, direct sampling of fish occupying submerged red
mangrove shorelines was possible. The fine mesh net
captured the small forage fish that numerically dominated
the habitats. Inevitable escapes, especially by large
roving fish, probably occurred, however, as persons
deploying the net waded up to the sites. The recovery
efficiencies of fishes encircled in the nets were comparable


Figure 2-14. Comparison of residency among the general
locations based on percent of total abundance. Data used in
calculations were taken with enclosure nets (for benthic
forage fish and water column forage fish) and by visual
census for large roving fish.
a. Abundance of fish in each category.
b. Number of species in each category.


Fish per m2 Fish per m2
Density by General Location
1.20
Large Roving Fish
CM
1.00
0.80
a) 0.60
CO
0.40
0.20
0.00
.ll.ll
Up Mid Down Up Mid Down
West East


19
test the following hypotheses. First, temporal changes in
fish densities were expected to occur in conjunction with
salinity changes. Secondly, in areas where salinities were
more variable, numbers of species of fishes were expected to
be lower than more stable areas. Thirdly, a community of
fishes including estuarine transient juveniles was expected
to occur in the study area. Finally, relatively lower
densities were expected to occur in the upstream locations
as an function of variable salinity conditions.
Materials and Methods
Pilot Study
A six-month pilot study was conducted to determine the
most effective methods for guantitatively sampling fishes in
mangrove prop root habitat throughout northeastern Florida
Bay. Absence of tidal exchange in the study area was an
important factor in selecting methods. Both collecting and
observational methods were explored. Collecting gear
selected for preliminary testing included minnow traps,
Caribbean fish traps, gill nets, pull-up nets and enclosure
nets with rotenone. The two visual census methods tested
were: 1) direct recording of fishes observed with mask and
snorkel on underwater data sheets and, 2) underwater video
taping. Two complementary methods were selected from those
tested in an attempt to sample the entire fish community.
These two methods were enclosure nets and direct visual
observation.


5
Most mangrove-dominated estuaries contain examples of all
these habitats and may thus provide a diversity of
conditions for use by fishes.
The Florida Bay Ecosystem
Florida Bay is a large (1,500 km2), mangrove/seagrass
dominated estuary located in extreme south Florida. The
majority of the Bay is not subject to tidal influence.
Wind-driven water movements can, however, raise or lower Bay
water levels rapidly. Sustained strong easterly winds can
literally blow water out of Florida Bay into the open Gulf
of Mexico. In similar fashion, winds from the north can
accelerate the introduction of mainland drainage into the
northern part of the Bay, and winds from the west can move
the water into the northeastern corner of the Bay (Ginsberg
1956) .
Internal circulation is restricted due to several
features of the Bay. The interior contains over 300
mangrove-fringed and overwash islands. On the east, U. S.
Highway 1 separates Florida Bay and Barnes Sound with only
one pass and two culverts providing water exchange. On the
west, it is separated from the open waters of the Gulf of
Mexico by a series of mud banks that are at least 2.0 km
wide and are often exposed (Holmquist et al. 1989b) (Figure
1-1). The lower Bay is separated from the thermally stable
and constant flow of the Gulfsteam by a series of limestone
islands known as the Florida Keys. Several major passes
occur through the Keys in the western Bay, but in the


Moyle, P. B. and T. J. Cech, Jr. 1988. Fishes: an
introduction to Ichthyology. Prentice Hall, Englewood
Cliffs, New Jersey, 559 pages.
Nielsen, L. A. and D. L. Johnson 1983. Fisheries
Techniques. American Fisheries Society, Bethesda, Maryland.
468 pages.
Nordlie, F. G., and S. J. Walsh 1989. Adaptive radiation in
osmotic regulatory patterns among three species of
Cyprinodontids (Teleostei: Atherinomorpha). Physiological
Zoology 62(6):1203-1218.
Odum, E. P. 1983. Basic Ecology. CBS College Publishing,
New York, New York. 613 pages.
Odum, W. E. 1971. Pathways of energy flow in a south
Florida estuary. Ph.D. dissertation, University of Miami,
Coral Gables, Florida. 162 pages.
Odum, W. E. and E. J. Heald 1972. Trophic analysis of an
estuarine mangrove community. Bulletin of Marine Science
22(3): 671-738.
Odum, W. E., C. C. Mclvor, and T. J. Smith, III. 1982. The
ecology of the mangroves of south Florida: a community
profile. U. S. Fish and Wildlife service, Office of
Biological Services, Washington, D. C. FWS/OBS-81/24. 144
pages.
Phillips, R. R. and S. B. Swears 1979. Social hierarchy,
shelter use, and avoidance of predatory toadfish (Opsanus
tau) by the striped blenny (Chasmodes bosquianus). Animal
Behavior 27: 1113-1121.
Pinto, L. 1987. Environmental factors influencing the
occurance of juvenile fishes in the mangroves of Pagbilao,
Phillipines. Hydrobiologia 150:283-301.
Power, P. E. and W. J. Matthews 1983. Algae-grazing minnows
(Campostoma anomalum), piscivorous bass (Micropterus spp.),
and the distribution of attached algae in a small prairie-
margin stream. Oecologia (Berlin) 60: 328-332.
Provost, M. W. 1973. Mean high water mark and use of
tidelands in Florida. Florida Scientist 36(1). 50-66.
Reid, R. K. 1954. An ecological study of the Gulf of Mexico
fishes in the vicinity of Cedar Key, Florida. Bulletin of
Marine Science 4(1): 1-94.
Remane, A. and C. Schlieper 1971. Biology of brackish
water. Wiley Interscience Division, New York, New York.


123
percent of the total number of fish examined (Hynes 1950).
In addition, the mean percent composition for each item was
calculated by obtaining an average value for all specimens
of a species. The results of both analyses are presented to
give a complete picture of the relative dietary importance
of the items consumed (Hyslop 1980).
Two sets of multivariate analyses of variance (MANOVA)
were performed for each species with the major items found
in the gut as dependent variables. All data were
transformed using an arcsine square-root function prior to
these calculations (Kleinbaum & Kupper 1978). The first set
of MANOVAs addressed spatial variation. Gradient (upstream,
midstream, downstream), system (east, west) and interaction
of gradient and system, were used as independent variables.
The second MANOVA looked at temporal variation. Season was
the independent variable. Only major food items, defined as
those which occurred in at least 20% of the specimens or
exceeded an average of 4% in composition, were included in
the MANOVAs.
Results
Shared Resources
Of the 24 items (counting fish as one and excluding
unrecognizable), 8 occurred in all 6 species (Table 4-1).
Of these, amphipods were the most ubiquitously consumed,
present in at least 5% of the specimens of all species.



PAGE 1

INFLUENCE OF CHANGES IN FRESHWATER FLOW ON THE USE OF MANGROVE PROP ROOT HABITAT BY FISHES By JANET A. LEY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1992 'UNIVERSITY OF FLORIDA LIBRARIES

PAGE 2

ACKNOWLEDGEMENTS I would like to extend my thanks to all my committee members and my Department Chairman, Joseph Delfino. They advised and supported me throughout my field work and writing. Clay Montague introduced me to science and Florida Bay, and encouraged me to pursue my interests. Carole Mclvor gave me guidance, scientific insight, and maintained faith in my abilities under all circumstances. I am extremely grateful to Bill Seaman for his efforts in obtaining Sea Grant support for a substantial portion of the project. In his Wetlands Ecology class, Ronnie Best introduced me to working in mud and swamps, features which later became a major part of my life. He also allowed me to live in the Center's Winnebago for two years, permitting me to operate on an intense and flexible schedule in Key Largo. Frank Nordlie expressed constant interest and encouraged me in my work. My visits with Nick Funicelli always included valuable personal and professional insights. I am extremely grateful for the efforts of Dan Haunert of the South Florida Water Management District. He believed in the benefits of this research to Florida Bay and aggressively oversaw the process of obtaining funding. In ii

PAGE 3

addition, my conversations with him gave me renewed enthusiasm and perspective. Dewey Worth, Dan's colleague, followed through with continued encouragement and support in later phases of the project. Laura Flynn and Luke Hasty were my very competent field assistants, providing good humor and constructive suggestions. They remained enthusiastic in every circumstance, from diving with major unknown creatures, to measuring 2-day old fish in 95 degree heat, to snorkeling in double hoods and wetsuits. Jacgue Stevens, Harriett McCurdy and my brother, Fred Ledtke, Jr., were my most faithful volunteers. Jacgue helped me tether over 200 fish and her ideas were invaluable. Fred devoted his hard earned vacations to his older sister's unusual effort. I would also like to express my gratitude to the staff of Everglades National Park. At the Key Largo Ranger Station, Dave and Louise King, Linda Cramer, and Dave Viscera included me as part of their small neighborhood during my 2 year residency. I am grateful for their support and rescues during boat break-downs. From the South Florida Research Center, Mike Robblee allowed me to use Park boats, provided insights concerning my research guestions and encouraged my efforts. DeWitt Smith also gave me encouragement and perceptive advice. Bill Loftus' help in iii

PAGE 4

fish taxonomy was invaluable. Katy Kuss was both a great adviser and a good friend. Gordon Thayer, National Marine Fisheries Service, advised me on the use of enclosure nets. In addition, other visiting scientists shared ideas with me, including Paul Carlson, Florida DNR; Jay and Rita Zieman and Jim Fourgurean, University of Virginia; and Dave Porter, University of Georgia. I am also grateful to the scientists of the National Audubon Society in Tavernier, Florida, who treated me as an adjunct staff member, especially George Powell, Mike Ross and Jerry Lorenz. In Gainesville, Ken Portier helped in the initial study design and last phases of analysis. In the bulk of the analysis effort, Steve Linda advised me on handling a very large data set. Hans Gottgens, my officemate, was constantly patient and extremely helpful in offering computer, scientific and personal advice. Most importantly, I would like thank those who encouraged my pursuit of this degree as a personal goal and supported me throughout the process. These are my parents who nurtured my spirit of independence and appreciation of nature, and Darlene Kalada, my best friend, on whose support and encouragement I could always depend. iv

PAGE 5

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS ii ABSTRACT viii CHAPTERS 1. GENERAL INTRODUCTION 1 Estuarine Fish Ecology 1 Mangrove Fish Ecology 3 The Florida Bay Ecosystem 5 Problem Definition 9 Objectives 12 Study Area 13 Fish Community Sampling Design 14 2. FISH DENSITIES AND ASSEMBLAGE PATTERNS IN MANGROVE HABITATS: COMPARISONS ACROSS SALINITY GRADIENTS 17 Materials and Methods 19 Results 32 Discussion 84 3. FISH COMMUNITIES IN FLORIDA BAY MANGROVE SHORELINE HABITATS: RELATIONS WITH PHYSICAL PARAMETERS AND COVER 98 Materials and Methods 99 Results 109 Discussion 114 4. FOOD HABITS OF MANGROVE FISHES: A COMPARISON ACROSS SALINITY GRADIENTS 120 Materials and Methods 121 Results 123 Discussion 129 5. PREDATOR ENCOUNTER RATES ON SMALL BENTHIC FISH ACROSS A SALINITY GRADIENT 134 Materials and Methods 137 Results 143 Discussion 149 v

PAGE 6

6. IMPLICATIONS AND CONCLUSIONS 154 Implications for Mangrove Fish Ecology 154 Implications for Estuarine Fish Ecology: the Nursery-ground Hypothesis 157 Management Implications 158 LITERATURE CITED 161 BIOGRAPHICAL SKETCH 172 vi

PAGE 7

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy INFLUENCE OF CHANGES IN FRESHWATER FLOW ON THE USE OF MANGROVE PROP ROOT HABITAT BY FISHES By Janet A. Ley May 1992 Chairperson: Clay L. Montague Cochairperson: Carole C. Mclvor Major Department: Environmental Engineering Sciences The hypothesis that seasonal changes in freshwater inflow (indicated by salinity) influence habitat use by fishes was tested in northeastern Florida Bay, extreme south Florida. Fishes were sampled monthly for 13 months using visual censuses and enclosure nets. Of the 305,589 individuals observed, 91% were estuarine residents, numerically dominated by engraulids, atherinids and cyprinodontids. Occasional marine and freshwater visitors comprised 2% of the individuals, and estuarine transients, 8%. No young-of-the-year estuarine transients were observed. Salinity ranged between 0.0 to 58 parts per thousand (PPt) upstream, 19.5 to 54 ppt midstream, and 30 to 50 ppt downstream. The 77 species were grouped for analysis: vii

PAGE 8

small benthic, small water column, and larger fishes. Abundances of larger fishes were consistently lower upstream (0.15 fish/square meter (m 2 )), than mid(0.65 fish/m 2 ) , or downstream (0.55 fish/m 2 ). Species of larger fishes numbered fewer upstream (11) , than midstream (15) , and downstream (22) . Benthic and water column fish abundances did not vary along the gradient. Temporally, fish distribution was uncorrelated with salinity. Development of mangrove habitat and submerged aquatic vegetation (SAV) were reduced upstream. Fish diets shifted to other foods upstream. Thus, where seasonal changes in freshwater inflow were greater (i.e. upstream), species and numbers of larger fishes were lower, possibly due to salinity conditions, food availability and habitat development. To determine if lower salinity conditions alone led to reduced predation, prey fishes were tethered along the gradient. Predator encounter rates were not different over the salinity range tested, but were 50% lower at the most remote sites. This was perhaps a function of accessibility of the sites to roving predators. Water management strategies to increase mangrove development and SAV are recommended research priorities. However, severe ecotonal differences between Bay and ocean waters, coupled with limited circulation and significant viii

PAGE 9

predation may inhibit recruitment and survival of postlarval fishes from offshore. An unbroken continuum of good habitat from outer to upper reaches may be necessary if northeastern Florida Bay is to function as a prime nursery area for estuarine transient fishes. ix

PAGE 10

CHAPTER 1 GENERAL INTRODUCTION The goal of improved management of surface waters to benefit estuarine fish populations in Florida Bay provided the incentive for this research. Before objectives and strategies can be established toward this goal, however, a better understanding of how freshwater inflow influences fish communities in mangrove estuaries overall is needed. This involves aspects of fish ecology of both estuarine and mangrove ecosystems. Estuarine Fish Ecology Ecologists often divide estuarine fishes into three groups: estuarine residents (complete their entire life cycle in the estuary) , estuarine transients (spawn offshore, their young use the estuary as a nursery) , and occasional marine visitors (usually adults) (Day et al. 1989). Resident and transient species tend to be widespread, but marine visitors are usually restricted to the higher salinity zones of the lower estuary (Weinstein 1979) . At certain times of the year densities may increase dramatically as influxes of transient juveniles enter estuarine systems. Some species tend to migrate to upstream-most habitats upon initially entering the estuary; l

PAGE 11

2 they then may disperse to lower reaches as they grow larger (Weinstein 1979, Rogers et al. 1984, Loneragan et al. 1990). The prominence of transient juvenile fish and crustaceans led to the application of the term "nurseryground" to many estuaries (Gunter 1961, McHugh 1967, Weinstein 1979) . A major role of freshwater discharge in such systems may be to increase food availability for fishes by transporting nutrients which stimulate primary production and by increasing detrital transport and processing (Odum et al. 1982). Freshwater inflow may also improve the chance for survival of juvenile fish in estuaries by reducing salinity levels below the limits tolerable by stenohaline marine predators (Gunter 1961, 1967). Browder and Moore (1981) offered a comprehensive nursery ground hypothesis linking several of these concepts. They split habitat factors into those that are relatively stable (e.g. shoreline edge, bottom type) and those that are movable (e.g. salinity, food resources) . Favorable habitat for particular juveniles consists of combinations of these factors that promote growth. According to their theory, the inflow of freshwater acts to position an area of favorable moveable habitat relative to important stationary habitat. Thus, for any estuary there is a rate of freshwater flow sufficiently high to push the band of potentially favorable moveable features beyond estuarine boundaries into open waters, perhaps eliminating favorable habitat entirely. Likewise, for every estuary, there is a rate of freshwater

PAGE 12

flow so low that the band of favorable salinities retreats upstream where stationary features may be unfavorable. The ideal situation with regard to freshwater inflow is one that maximizes the area of favorable habitat within the estuary over the peak period of nursery use. This hypothesis seems particularly applicable for analyses of fish ecology in mangrove-dominated estuaries. Mangrove Fish Ecology In tropical and subtropical areas of the world, mangroves are dominant shoreline features. Mangrove-derived detritus forms a food base for fish occupying mangrove ecosystems (Odum 1971) . Mangrove shorelines may also provide cover for fishes (Thayer et al. 1987a) . However, few studies have documented aspects of the direct use of mangrove habitats by fishes probably because monitoring fishes within the complex tangle of roots and branches is extremely difficult. Efforts have only recently focused on obtaining quantitative data on habitat use (i.e. Thayer et al. 1987a, Sheridan 1991, Morton 1990, Robertson & Duke 1987) . Strong linkages between mangroves and adjacent habitats may exist. For example, diel habitat shifts occur in both non-tidal (Thayer et al. 1987a) and tidal systems (Morton 1990, Robertson & Duke 1987) . Shifts from other habitats to mangroves occur during the life history of some species such as gray snapper (Lutjanus griseus) (Starck & Schroeder 1971) .

PAGE 13

4 The degree of selection among mangrove habitats by fishes has not been determined. Due to variation within mangrove forests, however, preferences are likely to be displayed. In south Florida, for example, three species of mangrove trees occur: red mangroves (Rhizophora mangle) , black mangroves (Avicennia germinans) and white mangroves (Laguncularia racemosa) (Odum et al. 1982) . These trees vary greatly in type of submerged features and potential cover for fishes. For example, red mangroves provide proproots; these are strong, woody structures that tend to extend from the mid-tree trunk downward to the substrate. Black mangroves, in contrast, tend to support a bed of pneumatophores that are pencil-like structures that grow upward from the substrate to several centimeters. Variation in degree of exposure to flushing also contributes to differences among mangrove forests and, hence, may influence habitat use by fishes. While the fringing mangroves along the shoreline are regularly flooded and thus accessible, more interior basin forests are irregularly inundated and thus occasionally available (Odum et al. 1982) . Within fringing shorelines, higher flushing rates contribute to greater mangrove habitat development (e.g. taller trees, more leaf production), which, in turn, is likely to generate more massive submerged structure for cover. Furthermore, detritus-based food resources are likely to be more abundant near highly productive mangroves.

PAGE 14

5 Most mangrove-dominated estuaries contain examples of all these habitats and may thus provide a diversity of conditions for use by fishes. The Florida Bay Ecosystem Florida Bay is a large (1,500 km 2 ), mangrove/ seagrass dominated estuary located in extreme south Florida. The majority of the Bay is not subject to tidal influence. Wind-driven water movements can, however, raise or lower Bay water levels rapidly. Sustained strong easterly winds can literally blow water out of Florida Bay into the open Gulf of Mexico. In similar fashion, winds from the north can accelerate the introduction of mainland drainage into the northern part of the Bay, and winds from the west can move the water into the northeastern corner of the Bay (Ginsberg 1956) . Internal circulation is restricted due to several features of the Bay. The interior contains over 300 mangrove-fringed and overwash islands. On the east, U. S. Highway 1 separates Florida Bay and Barnes Sound with only one pass and two culverts providing water exchange. On the west, it is separated from the open waters of the Gulf of Mexico by a series of mud banks that are at least 2 . 0 km wide and are often exposed (Holmguist et al. 1989b) (Figure 1-1) . The lower Bay is separated from the thermally stable and constant flow of the Gulfsteam by a series of limestone islands known as the Florida Keys. Several major passes occur through the Keys in the western Bay, but in the

PAGE 15

Figure 1-1. Regional maps of the study area showing features of the upstream drainage basins and Florida Bay. a. Boundaries of drainage basins and tributaries to northeastern Florida Bay. (Source: Schomer & Drew 1982) . b. Florida Bay showing the extensive mudbank system (stiple pattern) . Arrows indicated passes to the Atlantic Ocean and Barnes Sound. (Source: Holmquist et al. 1989b)

PAGE 16

a. US Highway 41 Long Key

PAGE 17

8 northeastern portion, only one pass exists, a man-made cut through Key Largo to the ocean side (Adam's Cut) (Figure 11) . On the northern boundary of the Bay are the Florida Everglades. The Florida Bay area is subject to an annual water deficit with evaporation exceeding total rainfall (Tabb et al. 1962) . Annual rainfall in northeastern Florida Bay ranges from 1600 mm on the mainland at Homestead to 1200 mm on the south at Key Largo. The climate of subtropical south Florida is characterized by a relatively long and severe dry season (November through April) and a wet season (May through October) (Schomer & Drew 1982) . Sea level becomes relatively high on an annual basis from August to December reaching a maximum of about 15 cm above the annual average in October (Ginsberg 1956, Provost 1973, Holmquist et al. 1989b). By late November or early December, Bay level recedes to the annual average, which probably accelerates the drainage of freshwater into the Bay from the mainland. At this time, the zone of reduced salinity may extend farther south and southeast into midand downstream Florida Bay areas. The major source of freshwater flow into Florida Bay is from a series of approximately 2 0 creeks and Taylor River, which carry surface water from the Taylor Slough/C-111 drainage area into the Bay. This system is smaller than the

PAGE 18

9 Shark River Slough, a separate system which extends from Lake Okeechobee southward toward the Gulf of Mexico and drains most of the Everglades. These overall features contribute to several environmental and biological patterns. Gradients in environmental variables occur in Florida Bay, from southwest to northeast. These gradients include amount of water exchange, sediment depth, and seagrass standing crop (Zieman et al. 1989). The area northeast of the central line of mud banks is characterized by very restricted circulation and no tidal influence (Schomer & Drew 1982) . A thin sediment veneer covers the basin bedrock in the northeast Bay, deepening towards the southwest. In addition, seagrass density and productivity decreases dramatically from southwest to northeast (Zeiman et al. 1989). Problem Definition Water management decisions in the eastern Everglades have potentially impacted Florida Bay through changes in the timing and guantity of freshwater discharge. Under predrained conditions, in this area, surface freshwater moved over grassy marl prairies that were seasonally flooded (Schomer & Drew 1982). A complex network of streams, bordered by mangroves and other shrubs carried freshwater inflow to receiving waters downstream in a manner that was presumably both gradual and dispersed. Beginning in the early 1900s, construction began on an extensive system of canals and ditches throughout much of

PAGE 19

the Everglades system. The effects of these canals may have included the overall reduction in the amount of freshwater storage in the system (T. MacVicar, South Florida Water Management District, personal communication) . In addition, after Everglades drainage, drier conditions may have occurred more frequently in the prairies and sloughs, with greater contrast between wet season and dry. Overland flow of freshwater entering the downstream estuaries under these altered conditions has probably been more rapid and less spatially dispersed. After entering upstream portions of Florida Bay, freshwater moves through extensive mangrove wetlands consisting of shallow swamp lands, creeks, ponds and bays, eventually reaching the open portions of Florida Bay. Changes in these brackish and marine receiving waters attributed to managed freshwater inflow may have included alteration of the annual salinity pattern, which led to unnatural cycles of both reduced and hypersaline conditions. Historical salinity data for Everglades waters, however, is lacking for the period prior to initiation of drainage. Evidence of the ecological effects of drainage on the downstream estuary has been discerned from National Audubon Society studies showing declining populations of estuarine wading birds (e.g. spoonbills) . Changes in hydroperiod have been hypothetically linked to a reduced fish and shellfish prey-base for the birds (National Audubon Society, unpublished data, 1989) . Furthermore, recent decreases in

PAGE 20

11 sportfish populations have been linked to hypersalinity stress for certain sportfish in Everglades National Park (Rutherford et al. 1989) . The Everglades estuaries are also critical habitat areas for other endangered aquatic species (e.g. American crocodile) that rely on the same forage base as do birds and sportfish (SFWMD 1989) . Thus, groups concerned about these problems spurred South Florida Water Management District (SFWMD) officials to take action that would return more natural drainage patterns to the estuarine areas of the Everglades. Some of these actions have focused on the C-lll Canal/Taylor Slough watershed which includes agricultural lands in a large drainage basin east of the Park. The downstream leg of the canal runs northwest to southeast, passes under U. S. Highway 1 and continues southward outside of Florida Bay, to Barnes Sound (Figure 1-1) . In low flow periods, the canal has functioned like a dike by preventing overland flow of freshwater from reaching both the downstream prairies and the approximately seven small creeks tributary to northeastern Florida Bay. In high flow conditions, water still flows through, sometimes sending slugs of freshwater into northeastern Florida Bay. Local topographic conditions tend to direct more freshwater toward U.S. Highway 1 than toward the west (Tabb et al. 1967). Thus, under these management conditions, the historic salinity regime is likely to have been altered. Changes in hydroperiod have probably resulted in more severe hypersaline conditions and

PAGE 21

12 sudden salinity changes of great magnitude, especially in the eastern part of Florida Bay. In the mid-1980s engineering alterations created several cutouts, each 20 meters wide, in the south bank of C-lll canal. The cutouts were intended to restore the more dispersed and gradual pattern of freshwater inflow to northeastern Florida Bay. Furthermore, an earthen plug was installed to block the C-lll outfall to Barnes Sound except on extreme floods when SFWMD can release water by opening it with draglines. The result of these alterations was to provide more flexible management of freshwater flow to northeastern Florida Bay. The guestion remaining is how to utilize this flexibility to improve ecological conditions. Objectives Fish and Salinity The first study objective was to determine the extent to which species composition and abundance were influenced by salinity variability in the northeastern Florida Bay study area. Because of direct and indirect salinity influences, more variable fish abundances and distinct community differences were expected at the upstream locations over an annual cycle that included both wet and dry seasonal differences in freshwater inflow. The eastern portion of the study area was also expected to be distinctly more variable than the western portion because of the influence of the C-lll Canal.

PAGE 22

Habitat Features Because salinity is not the only feature of the habitat that varies along the complex environmental gradient within the study area, it was also necessary to consider other features of the fixed and moveable habitat (Browder & Moore 1981) as potentially influencing fish community structure. The second study objective was to determine important habitat features that influence the abundance and species composition of mangrove fish communities and compare these features across the salinity gradient. Fixed habitat structural features such as mangrove tree height and prop root density, environmental features such as water temperature, and fish diet and predation, were expected to influence the differences among fish assemblages across environmental gradients. Study Area The 250 knr study area, located in extreme northeastern Florida Bay, consists of a series of shallow bays and ponds (less than 1.0 m in depth) bordered by mangroves. The upstream portion of the area is subject to freshwater inflow from seven mangrove-lined tributaries originating in the Taylor Slough/C-111 drainage basin. In this region of Florida Bay, rapid ecological changes can take place when salinity variations occur suddenly, as at the start of the rainy season (Montague et al. 1989). Because tidal influences are almost negligible in the northeastern Florida Bay area, salinity changes are caused

PAGE 23

by the variations in rainfall and subsequent freshwater flowing south through the tributaries, and variations in wind speed and direction. The rate and degree of salinity change are relatively unpredictable and can be rapid (hours) or slow (days) depending on changes in the weather. Fish Community Sampling Design To monitor fish community changes across the dynamic salinity gradient in northeastern Florida Bay, a balanced two-way analysis of variance (ANOVA) design was used, with two systems, each composed of three salinity regimes (Figure 1-2) . Generally, upstream locations included one of the creeks which carries freshwater from the Taylor Slough/C-111 Basin, an interior bay downstream from the creek but still measurably affected by freshwater inflow, and an outer bay much less affected by freshwater inflow but more by marine influences. Specifically, the locations were as follows: upstream sites were located in Highway Creek and Long Sound in the eastern system and Snook Creek and Joe Bay in the western system (Figure 1-2); midstream sites were located in Little Blackwater Sound in the eastern system and the Trout Cove area in the western system; downstream sites were located in Blackwater Sound in the eastern system and Buttonwood Sound in the western system.

PAGE 24

Figure 1-2. Map of the Florida Bay study area.

PAGE 26

CHAPTER 2 FISH DENSITIES AND ASSEMBLAGE PATTERNS IN MANGROVE HABITATS: COMPARISONS ACROSS SALINITY GRADIENTS Fishes tolerate salinities within a range of survivability (Moyle & Cech 1988) . If suitable conditions are not available within their environment, fish will experience stress, as evidenced by metabolic inefficiency and, in extreme cases, death (Moyle & Cech 1989) . In general, fewer species of all faunal taxa are able to tolerate conditions in zones with salinity conditions typical of the upper estuary (Remane & Schlieper 1971) . This may explain the occurence of lower numbers of fish species that occupy such areas (Deaton & Greenberg 1986) . As an alternative strategy to permanent occupancy and metabolic adjustment, fishes can shift habitats when salinity levels generate stress (Moser & Gerry 1989) . The occurence of a salinity gradient in the estuary provides the opportunity for fish to exploit different habitats and thereby avoid unsuitable salinities by movement (Weinstein 1979) . By stimulating such movements, salinity conditions may contribute to spatial and temporal fluctuations in species composition and abundances. 17

PAGE 27

18 Contrary to what might be expected based on such physiological factors, however, abundance peaks for many estuaries occur in conjunction with the initiation of the period of maximum freshwater inflow, when salinity levels drop dramatically. Estuarine transient juveniles may constitute most of the individuals during these peak periods (Yanez-Arancibia et al. 1980, Bell et al. 1984, Pinto 1987). In some species, juvenile fishes may be capable of exploiting salinities at lower levels than adults (Gunter 1967, Moser & Gerry 1989). However, in environments with more stable salinities, estuarine transient juveniles can also be abundant (Little et al. 1988, Robertson & Duke 1990b) . Thus, the role of seasonal changes in salinity on fish communities reguires further exploration. Based on investigations conducted near and within the northeastern Florida Bay study area, at the initiation of the rainy season (June) , changes in salinity were expected to occur, expanding the zone of low salinity further downstream (Ginsburg 1956, Tabb et al. 1962, Lindall et al. 1973, Thayer et al. 1987). This zone of lower salinity was expected to persist after the end of the rainy season, as freshwater from the eastern Everglades gradually drained into Florida Bay. Toward the goal of understanding the influence of freshwater inflow on fishes, the objective of this portion of the study was to identify spatial and temporal patterns in fish assemblages across the salinity gradient and thereby

PAGE 28

19 test the following hypotheses. First, temporal changes in fish densities were expected to occur in conjunction with salinity changes. Secondly, in areas where salinities were more variable, numbers of species of fishes were expected to be lower than more stable areas. Thirdly, a community of fishes including estuarine transient juveniles was expected to occur in the study area. Finally, relatively lower densities were expected to occur in the upstream locations as an function of variable salinity conditions. Materials and Methods Pilot Study A six-month pilot study was conducted to determine the most effective methods for guantitatively sampling fishes in mangrove prop root habitat throughout northeastern Florida Bay. Absence of tidal exchange in the study area was an important factor in selecting methods. Both collecting and observational methods were explored. Collecting gear selected for preliminary testing included minnow traps, Caribbean fish traps, gill nets, pull-up nets and enclosure nets with rotenone. The two visual census methods tested were: 1) direct recording of fishes observed with mask and snorkel on underwater data sheets and, 2) underwater video taping. Two complementary methods were selected from those tested in an attempt to sample the entire fish community. These two methods were enclosure nets and direct visual observation.

PAGE 29

Design of the Main Study The climate of subtropical Florida is characterized by a relatively long and severe dry season (November through April) and a wet season (May through October) . Thus, the sampling schedule included monthly sampling for a one year period to encompass the influence of changes triggered by seasonal climatic conditions. To monitor fish community changes across the dynamic salinity gradient of northeastern Florida Bay, a balanced sampling design suitable for analysis of variance was used. The design consisted of two systems, each having three locations along the salinity gradient (Figure 2-1 and Table 2-1) . Based on the pilot study, this geographic design was to encompass three regimes of salinity variability within each system: Upstream: low mean / high variation; Midstream: mid mean / mid variation; Downstream: high mean / low variation. Enclosure Nets From the pilot study, one collecting method proved to be superior to the others, both in terms of sampling the breadth of species at the sites and providing a quantitative sample of fish density. This was the enclosure net first used by Thayer et al. (1987) to sample mangrove shoreline fishes in western and central Florida Bay. This method was selected for targeting small benthic and water column fish in particular.

PAGE 30

Figure 2-1. Northeastern Florida Bay study area with sampling stations indicated.

PAGE 31

22 tn c o fl p (fl P co tn 3 CD C ID U 10 3 tn -h > M CD H CN ro 73 "0 C C O 0 04 0< M M M Q) 0) 0) 0)0)0) h h h O t) o J< M O O O OOO C C C co co co OJ > o o p 3 0 _ U X H C P u CO 0) fl) H 0) P 3 O CO * li « ST B -h i Q >i 0) O 3 D >i CO >i US 0) r** O 73 aS 0) C 0 -H -p as P CO In 5 H 3 73 C o en o< c * K u n) O X! 0> CO H (fl H >1 0 D6 73 c ta rH tn M 0> •H Q) OJ U Q >i (fl 3 -C tn -H * C 3 0 CO O P tn as CD £ -p u o c <-h (N n in vo roo ui o h in m ^ m vo * K 73 "0 CO CO c fl) H w I— 1 CD CD CQ -P P •H *> 73 U o CO OQ £ x: -P -p h >-i 0 0 C c r00 H rH tH 73 W !-l 0) -P (fl 0) rH O P (fl en rH CD CQ Vh p a o CQ -H P P fl) 01 -P X) (0 rH O •H P O 73 >i 03 x: p 3 O ra O H M l<1 c •H CQ CQ •H £ O 4J 0) 3 73 CQ CQ >i rH fl) C fl) • o 0) O-i .c at P H E C O "H h MH >i H X) J 0) P ft fl) a e o -H P P 73 ft a* ca p s a* n) 3 XI -P 0) z CD 1-1 3 tn o rH o c w tn c O -p OS P CO 3 2 P P ca 2 E 3 roo o\ O rH CN ro ^* in CO p u OJ XI rH •H O (0 OJ c \o rao 73 •H s >1 fl) OQ i to OJ rH P tn ft D P tn OJ S CD P. -P tn 73 •H OJ u 4J tn c 3 o a x 8 (fl I I 0) rH P tn r^ P tn (fl w rH OJ p fl) u rH OQ P P •H rH 0) rH P CO 73 •H 2 73 C 3 0 CO p 0) p fl) o fl) rH OQ CD tH P tn c 3 o a p •H P (fl rC P rH CQ XI C -H O CO •H -H P > fl) P P CO 0 * a

PAGE 32

Sites . Because enclosure net stations were located in open bays, sites were selected that were protected from prevailing direct winds. Such siting prevented the wind from pulling the bottom of the net off the substrate. Sites were further chosen to have between 20 and 100 cm mean water depth at the outer prop root edge. This criterion was intended to provide some uniformity among the sites in terms of volume of water enclosed in the net. At each site, a natural berm consisting of packed detritus approximately 15 cm high and 30 cm wide occurred along the landward edge. This berm was exposed at high tide and provided a bank beyond which fish could not escape when rotenone was applied within the net (see below) . Procedures. At the start of the study, three sites were selected in each of the six general locations. A maximum of three enclosure nets could be deployed at each location by two persons in a day if the nets were deployed no more than approximately 1.0 km from each other. Environmental measurements taken during each collection included water depth, salinity, temperature, wind speed, wind direction, and air temperature. Salinity and water temperature were measured with a calibrated electronic instrument (YSI Model 33 S-C-T meter) . For salinities above 35 ppt, a calibrated hand-held ref ractometer was used. For each net, two 30 cm wide paths were cut perpendicular to the shoreline through the mangrove fringe back to the berm. The paths were cleared of bottom roots

PAGE 33

24 and overhanging branches so that a person could walk up the path carrying one end of the 30 m long net. The same site was sampled repeatedly throughout the study unless "stress" was observed in submerged vegetation. For example, sites with clay sediments supporting seagrasses had suffered some visible damage (e.g. grass trampling) from the sampling procedure by the fourth month at three stations. To maintain consistency in types of habitat encompassed by the nets, at these three sites, one new path was cut so that an unimpacted site could be sampled adjacent to the old one. These minor site changes were taken into account in later calculations of net area sampled. On the day of sampling, a 6.0 mm mesh nylon seine was deployed by two people who carried it, scrolled around two wooden dowels, to the mid-point between the two prepared paths. Starting 10 m from the edge, they waded in opposite directions parallel to the edge, unrolling the net, and then walked toward the mangroves and up the paths. The dowel end was pounded into the sediment at the landward end of the path and the lead line was pressed down into the sediments all around the bottom edge. The top edge of the net was hung over several PVC poles to prevent fish from jumping over the net. All three nets were set in similar fashion (Figure 2-2) . Liguid rotenone was then applied within the enclosed area to a final concentration of 5 mg L" 1 . Fish that immediately began to surface were collected using hand nets

PAGE 34

25 Figure 2-2. Enclosure net illustration indicating net dimensions .

PAGE 35

for 30 to 45 minutes by two persons. After repeating the process at the other two sites, and allowing the rotenone to dissipate (approximately 3-4 hours) , a snorkeler retrieved sunken fish and a wader collected along the berm edge from within each enclosure. Since rotenone effectiveness is reduced with decreasing temperature (Neilson & Johnson 1983) , in colder months, floating fish were also collected after leaving the nets up overnight. Fish and invertebrates collected were initially placed on ice and then frozen. Later they were identified to species and measured to total and standard (or carapace) length. Efficiency tests . Fish recovery efficiency was tested at least once at each location using a mark-recapture method and normal net procedures. To collect test fish, several minnow traps were placed inside the area to be enclosed on the day before net deployment. After the net was in place, the minnow traps were removed, cleared and the fish placed into buckets. Fish were measured, marked by fin-clipping, and returned to the enclosed net area in minimal time. At least 30 fish were used per net in the test. Visual Census Based on the pilot study, direct recording of census data on underwater paper was identified as the better visual method tested. Quality of video tape was inconsistent and often low under the variable turbidity conditions encountered.

PAGE 36

27 Study Sites . For visual censusing, site selection criteria included adequate depth (20 to 100 cm) , red mangrove dominance, and wind-protection. At 70 m in length, however, each transect encompassed a range of depths and other physical characteristics. Four sites were randomly selected for permanent sampling stations within each of the six general locations. These sites were chosen in the same general locations as the net sites but were more widespread than the net stations (Figure 2-1) . Procedures . To prepare permanent census stations, mangrove edge transects were designated with flagging tape every 10 m along the 70 m edge. Physico-chemical variables were recorded in conjunction with each visual census (e.g. air temperature, salinity, wind conditions) . Other variables measured included: 1) water depth was recorded at a permanent stake located in each transect; 2) abundance of submerged aquatic vegetation adjacent to the transects was noted using a scale of zero to three (abundant) ; and, 3) range of visibility was determined by using a white PVC pole set vertically into the mud and measuring the horizontal distance at which the pole became visible as one snorkeled towards it. If the visibility was less than 100 cm, the census was rescheduled. If visibility was initially poor, three attempts (on subsequent days) were made to conduct surveys. However, in some months it was impossible to conduct a visual census at particular sites.

PAGE 37

28 To conduct the census, a snorkeler approached the flagged edge and remained stationary under each flag for 30 seconds, an adeguate period for recording observable fish. On underwater data sheets, they recorded the species, numbers and estimated sizes of fish observed. The census surveys were conducted by myself and one assistant. Each census consisted of four complete swims of each transect. Efficiency tests . Efficiency tests were conducted for the visual census technigue using a mark-recapture method. Large fish (18-25 cm) were caught using hook and line and smaller fish (less than 18 cm) were collected with minnow traps. Tags made of plastic waterproof tape in various colors and labeled with a unigue code were used. Fishing line was securely fastened to the tag and the line was sewn with a small sewing needle through the flesh just under a fish's dorsal fin (for gray snapper and larger fish) or through the lower jaw (for smaller fish). To conduct a test, block nets were used to enclose an area of mangrove shoreline of adeguate size to accommodate at least two snorkeling stations. Tagged fish (six to nine larger fish and ten smaller fish) were placed inside the enclosure for several hours to allow them to acclimatize to the habitat. Snorkelers then carefully entered the enclosed areas and conducted a visual census of the site by recording the species, tag color and code of each fish they observed. Three such tests were conducted in the summer of 1990, at three separate stations.

PAGE 38

Definitions As presented in Chapter 1, three categories of residency are recognized by estuarine ecologists: residents, transient juveniles, and occasional visitors. In analyzing the community in northeastern Florida Bay, several sources were consulted for life history information on individual species to designate each by residency (e.g. Odum & Heald 1972, Lee et al. 1980, Yanez-Arancibia 1980, Robins et al. 1986). Without conducting specific gonad analysis to determine maturity (e.g. Robertson & Duke 1990b), uneguivocal distinction between juveniles and adults in the transient category were not possible. In addition, life history information is sketchy for all species except for certain killifish. Thus, these designations are approximate and serve for discussion purposes only. Such designations were not used in statistical analyses of the fish community. For purposes of detailed analysis, all fish were assigned to one of three groups of species based on size, behavior and primary portion of the mangrove habitat occupied during the day. Forage fish were considered those species whose members were generally less than 15 cm in size. Two groups of forage fish occupied different portions of the mangrove habitat: benthic and water column. Benthic forage fish live in close association with the substrate and include such species as gobies, killifish and mojarras. Water column forage fish are exclusively schooling fishes, that occupy the upper water column habitats, including

PAGE 39

30 anchovies and silversides. The third fish group, large roving fish, are generally greater than 15 cm in size and occupy both the bottom and water column locations. This group included such species as snook, tarpon, snappers, catfish, grunts and barracuda. Density determinations To obtain densities for each enclosure net sample, abundance was divided by the area the net encompassed. Areas enclosed ranged from 72 to 196 m 2 (mean = 119, sd = 33.2) . For visual census samples, measurements of horizontal secchi distance and fringe width were used as radii in calculating the area observed (half the area of a circle) . Because it was impossible to see and accurately identify small fish at a great distance, maximum radius values of 2.0, 3.0, or 4.0 m were applied to benthic, water column and large roving fish respectively. Thus, the areas sampled in the visual censuses ranged from 1.6 and 25.0 m 2 at each of the eight stations along a 70 m long transect. Analysis Methods Temporal patter ns in density . To initially inspect the data for patterns, salinity and density by fish group were graphed. Temporal patterns were examined graphically and guantitatively. For each of the three fish groups, and the three species that were most abundant within each group,

PAGE 40

31 correlations were calculated between monthly averages of these fish densities and corresponding values for salinity, water temperature, and water depth. Spatial patterns in density . To discern spatial patterns in distribution for each fish group and the top 3 species within each group, repeated measures analyses of variance (ANOVA) with multiple comparison tests were used (SAS GLM procedure) . All density data was effectively normalized by log-transformation. In the initial ANOVA model, density of fish was the dependent variable and gradient position (up-, midand downstream) and system (east and west) were the independent variables. To explore the relative density patterns among the general locations, a second ANOVA with general location as the independent variable was conducted. Further analysis was conducted to determine if spatial variation in certain environmental parameters might indicate why the densities varied among the stations. The average values of fish density, water depth, and salinity for each station (n = 18 stations) were calculated. Additionally, the amount of salinity variation over time at a particular station was also calculated by determining the standard deviation of salinity. Correlations between average fish densities and the means for these parameters were then determined. Community patterns. For comparisons among general locations, an index of species richness (Odum 1983) was calculated with total number of species as the numerator and

PAGE 41

32 log-transformed abundances as the denominator. Actual fish assemblage patterns were compared to gradient positions using cluster analysis. Data for each station, date and species were used to form matrices of stations based on similarity values (SAS CLUSTER procedure) . An average linkage method was used to join clusters of stations. The resulting dendrograms were compared with the gradient positions. Those stations that were placed in a group other than the correct up-, midor downstream position, were denoted as misclassif ied. A second analysis was conducted on the log-transformed densities at each station in order to more thoroughly explore the data. Results Tests of Recovery Efficiency Results of recovery efficiency tests for both the enclosure net and visual census technigues measured the number of fish sampled out of the total that were at a site (Table 2-2) . However, no estimate is available for either method for sample accuracy, i.e. for how many fish escaped the area as the net was being deployed or the observer approached the area. Enclosure net ef ficiencies . In all tests spanning up-, mid-, and downstream locations, 492 fish were marked. Of 14 total species, 60% of the fish used in the tests were goldspotted killifish (Floridichthys carpio) . An average of 18% of all fish were recovered in the initial dip-net collections. By adding the same-day snorkeling procedure,

PAGE 42

M 3 W c Q) U 03 A3 n -p 0) c u o rH o c • -p o a) to a) p u 3 -p w a Q) (0 ^ U a> i -p no •P >h >i B o C >i a) xj •H O TS •H Q) >1 o c 0) H o H (0 a) -h a> o o 33 Jh "H a> 4h w u X! B 3 2 X! DO H &H 4n O T3 0) CP id Eh W 0) W 03 id rH U 0 N -H W X -p CP c 3 rH DO U 9 P
PAGE 43

34 efficiency increased by 7%. The total mean recovery rate was increased to 37% by leaving the nets up overnight and collecting the next day. Overall, a greater percentage of larger fish were recovered than smaller (Table 2-2) . Of six large fish (Lutjanus griseus) that were tagged, all were recovered after rotenone application. Twice as many mid-sized as small fish were recovered. Visual census efficiencies . Individual test results for small fish (all Floridichthys carpio) ranged from 25 to 27% efficiency for the visual censusing method (Table 2-2) . For large fish (all Lutjanus griseus) , results ranged from 78 to 100%. Several tagged fish were observed more than one time during the four swims along the transect. Thus, when analyzing the data for each sample, to prevent counting the same fish more than once, after recording the first swim, only unique species and size classes of fish were added to the dataset for the second, third and fourth swims. These efficiency analyses were intended to identify trends in fish recovery rates. Due to wide ranges in the test results, subsequent data analyses were not corrected for efficiencies. Overall Abundance Results of the visual census differed from enclosure net sampling results (Table 2-3). Enclosure net sampling resulted in the collection of 82,633 fish from 59 species and 29 families. The greatest abundance was collected at

PAGE 44

cm (d rH l-> CD CQ 4-1 X5 O o T3 "0 a d a) (0 4-> CQ 4-> CD a a Q) > CD •H U en CQ >l o u 1—1 a u CD H CQ en CD (_l •H \ CQ a o s-i cd en -!-> o 4-4 ai o < — i H a O o u •H -P X! o « CD i c S o a CO TO CD I g 2 to TO a CL 3 5 i c § Q 1 I 3 i g 2 c5 s i a. 5 I rm u> in a a> r» fCN w CM O If) CM CO cn, 8 s s o CD "O 8 i QC0 E f i I 8 IT 3 0) CD i i 11 a u I 1 « *1 | f j 1 _ -Q O 5 CD to 3 a.
PAGE 45

36 I CO O 3 0> a ? 2! IS 2 rS >^ Q O <\i oj co S5 <<0" »-" « s » n r JCO CM a 00 O CM ir~ 5 m o" as So ° & " o * sis si si * CM O O 3 CO O CO cm to £= o CO c O Q) C H a o u I CN 3 CO re a> i c 5 o Q CD z CD CD g 2 to » CO 8 i • Q. Q) 3 Z 5 i c o I i •a I 1 *— * Co" 0> IT) «CO Q * Q CM (V CO n 0> 8 1 8S N CO CO 5 » «lO CO II O CM CO o CO CM CM I a K W O) (J) N (0 00 |^ 0ft CS| 01 § 8 O CM K K CO 00 co * » oa cm 00 N O CO CM CO CO CM tK O 5 K <0 If) , & 00 *• cm «0 ; « I 1? 5 O CD y~ I 8 Q. CD 8 R | 8 a 8 c© n K cm N N j r i a e cu t ff $ S 8 " w CMCO <0 CO CM 5 CM cm" & 00 3 Q. 2 $ o cc 8 CO E CO >. LL O c o "O 1 S § 5 s I 1 i CO m 3 i .§ 0 = & 1 I i ill o I ? c| & S 5 5 S.1 " 5 8 CS CO s 1 1 J l) H. H ^ l[ 6 co S s e J k 1 1 cc co (3 CQ i O O m 5 § 3 C * 1 1 S I c 5 -5 c co q> c • £ £ J I ^ 5 <0 CO 5 i ? 1 1 ills i 1 f i p cc e> os 5 c p e f 11 to «0 3 3 c» * o u

PAGE 46

37 p CN Total Nets Visual 5 S3 1 1 2 4 2 11 2 1 74 96 18,461 2 456 19 1.597 2,081 704 528 410 5.564 266 1,236 1 139 2,768 52 163 General Locations Down-east Nets Visual X •» >-e\J^ K Co 0> OCuCMQQ >^ CO Co cm ^g* -g-^g »-gj K." c\j W Mid-east Nets Visual s S t " 9 S 8 8 CO iN r eg in eg o CD 5 i i F Up-east Nets Visual s * $ Sj •» O CO CO CO 1 » Down -west Nets Visual 41 4,75*7 1 129 67 446 104 1,739 34 /«4 20 52 114 Mid-west Nets Visual 2 23 1 4,737 209 212 r^s 37 2 241 /„2«p 7 26 26 16 2 41 Up-west Nets Visual * r a § * 28 S N 8 CO V Group/ Residency Family Species Centropomidae (snook) LR/o Centropomus undecimalis Echeneididiae (remoras) LR/o Echeneis naucrates Carangidae (jack) LR/o Trachinotus goodei LR/o Naucrates ductor LR/o Carangidae (sp unk) LR/o Caranx hippos LR/o Carangidae (juv.) LR/o Trachinotus falcatus Lutjanidae (snapper) LR/t Lutjanus jocu LR/ta Lutjanus griseus LR/t Lutjanus apodus Gerreidae (mojarras) BF/r Eucinostomus sp BF/r Eucinostomus harengulus BF/r Eugerres plumieri B F/r Eucinostomus gula BF/r Gerres cinereus Haemulidae (grunts) LR/t Haemulidae (sp unk) LR/t Haemulon parrai LR/ta Haemulon sciurus Sparidae (porgies) LR/o Lagodon rhomboides LR/o Archosargus rhomboidalis LRAa Archosargus probatocephalus

PAGE 47

38 0 CD General Locations Total Nets Visual 1 1 27 1 .062 484 3 8 18 3 4 3 29 1361 24 7 104 825 14 2,997 804 534 200 66 125 Down -east Nets Visual Cg « «**o in cm *-»*-•CM r35 Mid-east Nets Visual OCMIO 0> CM O 1O T* CM CM iH T— Up-east Nets Visual at co cm Down -west Nets Visual to * CO v> 11) CO ^ CO <0 co a ID f CM Mid-west Nets Visual K cm 8 CM CO fCM *~ at Up-west Nets Visual 1 7 85 387 3 3 11 90 1 12 7 6 1,011 675 235 151 18 124 Group/ Residency Family Species Lepisosteidae (gar) LR/o Lepisosteus platyrhincus Centrarchidae (sunfish) BF/r Lepomis macrochirus Cichlidae (cichlid) BF/r Cichlasoma urophthalmus BF/r Tilapia mariae Ephippidae (spadefish) LR/o Chaetodipterus faber Scaridae (parrotfish) LR/o Spar i soma radians LR/o Scaridae (sp unk) Lobotidae (tripletail) LR/o Lobotes surinamensis Mugilidae (mullet) LR/o Mug// cephalus LR/o Mug/V curema LR/o Mop;/ //za Sphyraenidae (barracuda) LRAa Sphyraena barracuda Bleniidae (combtooth blenny) BF/r Chasm odes saburrae Gobiidae (goby) BF/r Microgobius gulosus BF/r Gobiosoma robustum BF/r Lophogobius cyprinoides BF/r Gobiosoma bosci

PAGE 48

CO o 1 CO c o CO 8 CO i_ 0) c a> O 0 a H 4-1 a o u n i Eh 13 I 8 i I £ o ® Q Z 1 CO T5 a. 13 1 z I c/5 i ~ o °> Q Z is z CM W) S 8 I I 5 s a. a> 3 Z 8 i o II o §. .5 -s. § 3 ? I o S. u £ » I 5 ® Q "o ^ « to 3 S < 5 O £ S s s i I M H < m w •c 111 IIS 2 i £ 6 « co if) TCM •i s ? CM CO CO g 8 S 9 oo m ft f. «m o>* 2S8 CM § ^ ^ N « ID JDS 0» 1^ co S 8 CM 8 m cm CM CM C\J 8 55 8 o" if CD CO CQ 43 0) 03 0) •H fd,C 4-1 5-1 CQ O-H 01 4-1 cm 0) rd a u g o 3 O 4-1 a a CD CD > > 3 3 CQ 4-1 4-> Mac O CD CD U -H -H H CQ CQ CQ a H rd rd > ^4 M 4-) 4-> H (T) CD CD U U M 4-> C! G a •H a O-H •H X! »-i CD CD •H Ch u 4J H •H fd d cd id m £ J CD U CQ CQ u O CD CD ii ii ii II II II II m j (h O 4-1 fd 4-> • • >i u G CD a T3 a H CQ o CD H U

PAGE 49

40 Little Blackwater Sound, the midstream-east location. The greatest number of species, however, was found in samples from Joe Bay, the upstream-west location. Visual census sampling resulted in observation of 222,960 fish from 51 species and 31 families (Table 2-3). Greatest abundance and greatest number of species were observed in samples taken in Blackwater Sound. Samples obtained by the two methods differed in relative abundance and numbers of species within these three fish groups (Table 2-4) . For example, many more species of benthic forage fish were collected in the enclosure nets (33) than were observed in the visual census (16) . In contrast, many more large roving fish species were sampled in the visual censuses (29) than in the enclosure nets (17) . Temporal Patterns in Density by Fish Group Benthic forage fish . In Figure 2-3, one can compare changes in salinity with changes in density from the enclosure net sampling; however, no consistent patterns emerge. Great density variations occur independently of salinity changes. Salinity varied widely over the study period at the upstream/east (0.0-39.0 ppt) and upstream/west (13.0-58.0 ppt) locations. Salinity also ranged widely at the midstream/east location (19.5-50.0 ppt). However, at the other three locations ( downstream/ west , downstream/east and midstream/west), salinity remained high (29.8 to 54.0 ppt) throughout the study. Not only was the period of low salinity longer in the upstream/east location, but also, a

PAGE 50

41 Table 2-4. Summary of abundances and number of species by method of collection and fish group. Fi sh Group Methods Parameter Enclosure Nets Visual Census Benthic Forage Fish Total No. Species 45,458 33 39,476 16 Water Column Forage Fish Total No. Species 35, 926 9 156, 610 6 Large Roving Fish Total No. Species 1,249 17 26, 874 29 All Fish Total No. Species 82, 633 59 222,960 51

PAGE 51

42 Parts per Thousand J9}9LU 9JDHbS J9d L| SI J 0) > O H u <0 CP w ^ ChO is id c CP C M •H O £ u tO k P w 0) c o (0 a) © — • 3 (D Q) M C >i O-H O HHH O ^ ft C 0) a> to *d •p A C 01 •h g a> ft 0) c 3 u o a> to MO a> to P p e tj •H (0 h TJ O <0 Q P T3 to C •H (0 • -P n to (N 10 Q) P -P
PAGE 52

43 substantial decrease (from 35.0 ppt to 10.4 ppt) was evident in June 1990, that did not occur in stations sampled in the upstream/west location (which became increasingly hypersaline) . None of these salinity changes correspond with patterns observed for fish densities. Density of benthic forage fish peaked in winter months at four of the six general locations (Figure 2-3). The highest density collection (13.6 fish m" 2 ) was at the mid-Trout Cove station in winter 1989; lowest density occurred at the mid-Little Blackwater Sound in the June 1989 (0.12 fish m~ 2 ) . Water column forage fish . In Figure 2-4, one can compare changes in salinity with changes in density from the enclosure net sampling; again, however, no consistent patterns emerge. Density of water column fish was highly variable and the graphs illustrate no consistent seasonal patterns. In general, either very low or very high densities of these schooling fishes were collected. The highest density collection (25.3 fish m" 2 ) occurred at midLittle Blackwater Sound in September 1989. No water column forage fish were collected in several samples. As with the benthic forage fish, these density fluctuations were also not related to the seasonal fluctuations in salinity. Large roving fish. In Figure 2-5, changes in salinity can be compared with changes in density for this group from the visual census sampling; again, however, no consistent temporal patterns emerge. In the upstream/west location,

PAGE 53

44

PAGE 54

Parts per Thousand J9}9iu ojonbs J9d qsij i CM c -h 0) 3

PAGE 55

the visual census stations were notably fresher (1.4-50.0 ppt) than the corresponding net stations (13.0-58.0 ppt) (Figures 2-3 & 2-5) . In this location, the visual census stations were located within the creek itself while the net stations were located in Joe Bay immediately at the creek mouth (Figure 2-1) . These seemingly slight differences in location may have contributed to observed differences in salinity patterns for the visual and net stations. For the visual census, the salinity patterns in both the eastern and western upstream locations were similar to each other except in spring 1990, when freshwater entered the upstream/east location reducing salinity to 30.0 ppt and the upstream/west location became strongly hypersaline (47.3 ppt) (Figure 2-5) . The midstream/east location was overall more variable than the midstream/west location. At the mid/west and downstream visual census locations, salinity patterns were uniformly high. Density of large roving fish ranged from zero at the uppermost upstream/west location in the spring, fall and winter of 1989, to 2.3 fish m~ 2 at Duck Key (midstream/west) in winter. From the graphs, it appears that changes in density of this group were independent of salinity changes (Figure 2-5) . Temporal correlations. No significant correlations between temporal changes in salinity or water depth, and temporal changes in fish densities were found (Table 2-5) . However, changes in water temperature were correlated with

PAGE 56

47 >i P u •H CD 0 Ul > •H C 0 th de TS c CD CD i— i CP (0 (0 p < u (0 CD Q o > (0 Ul 43 x! p -p c o a t3 £ J-I o a) (0 -p (0 p ft Q CD ui c o •H p (0 P H J-I o <+h a) ra jh u Ul o O +J T3 C aS >i H P CD (C 3 -H U) C -P a> H JH • O -0 H Q) Jh C ft-H M i ft X) B — Jh + c X CD XI Ul •H H -H Q) 0) (0 as ft > o 1 Jh P CD c •H £! Ul (0 (0 -H Ul <*H P H CD Jh M O U in T3 CP C C <0 u B o •p cm a} -P a) ui > o u CD Cn M (0 43 J-i a> Ul (0 jh •H tn J-i rH C o o O O -P c a> at ft X! CD U (0 43 O Eh -P 1 p •H c •H rH CO J-I O n H o a> cp CD P 0) (0 ft u w CD rH (0 > ft c o •H P (0 H CD JH o u CD rH > I ft C o H p id rH CD J-i Jh O U CD rH (0 > I ft c o •H p (0 H CD Jh Jh o o o vo CO H CO CN O CP o O CO o • • • o o o o + CD CP (0 Jh o fa o •H 43 P C CD cq o o I 10 o • o + 43 U) H fa C B 3 rH o u M CD P id 5 x 10 •H fa CD CP (0 o fa H o in CM in r» rH CO • • • o o o in m o m n • • • o o o + + 1 VO vo r» in CO (N • • o o O t o + CP C 2 & (X ui •H CD fa CP u (0 3

PAGE 57

48 density of benthic forage fish and large roving fish. In both cases, lower abundances occurred at higher water temperatures . Spatial Patterns in Density by Fish Group Spatial patterns in fish density varied among the fish groups (Figure 2-6). From these graphs, one can see that only the larger roving fish group seems to vary consistently along the salinity gradient, with much lower densities at the upstream locations. Benthic forage fish analysis of variance . Results of the repeated measures ANOVA's by fish group differed among the fish groups (Tables 2-6 and 2-7). Neither gradient position nor system were important determinants of variation in densities among the stations for the benthic forage fish group (Table 2-6) . Although densities tended to vary significantly from one general location to another, these variations were not systematic along the salinity gradient, as indicated by the significant interaction between gradient and system. The mid/west general location had significantly greater densities than the other midstream location (Table 2-7, Figure 2-6) . Other locations were intermediate and not significantly different from these two. Water column forage fish analysis of variance . Again, although densities tended to vary significantly from one general location to another, these variations were not systematic along the salinity gradient, as indicated by the

PAGE 58

Figure 2-6. Mean density of fish by general location for each fish group. Error bars illustrate the magnitude of the standard deviation in density over all the months. Samples of benthic and water column forage fish taken with enclosure nets. Large roving fish were sampled with visual methods.

PAGE 59

Density by General Location Benthic Forage Fish CN E a. Water Column Forage Fish CM E i_ v Cl JZ CO CN 1.20 1.00 0.80 Large Roving Fish a> o.6o V) 0.40 0.20 0.00 Ji.ll Up Mid Down Up Mid Down West East

PAGE 60

•p c c Q) g •O 3 C rH CD O o 0) TJ U -P C -p O -H rH U rd CP o U rH (0 J o p . T3 (0 1 rH (0 c (0 oi 0) Ul rd 0) e T3 a) -p id 0) 01 (0 e p 01 >i 01 c rd p c 0) •H TS (0 u rH 01 o rH u c 0) CP c H 01 T3 i p o 01 c m rH -P ai rH rd p (0 a m; -a c rd I CM 01 01 C XI (0 •H rH rd > ai CP rd rH O <4-l U -P rd rH rH (0 3 3 O 01 rH •H (0 > o 01 •H fa Cn C -H > O K oj CP VH rd en •H fa 0 cn to u o fa c 1 ^1 0 u u CD HJ rd s J3 tn •H HH o O VO O CN CN • • O O o O o o O o CN in r» cn vO t> CM VO rH CN IT) cn • • • cO O o CN rH rH o CN CN CN 00 00 00 00 rH rH \ rH CN rH CN 00 cn in vo oo 00 C^ CN in rH rH m cn tH VO • • • • rH • o oo in o O o o CN CN CN CN CN CN CN rH \ rH "»s rH H rH rH rH CN rH CN ^ O o o O rH CN rH CN rH in rH rH r> CN rH o r> O CO CN in o O tH ID CO o o O n rH rn • O o o o o o O VO o\ ID in CO r> rH en in co rH O • O rn r> rH rH rH O o o O CN CN CN CN CN CN CN rH \ rH "V. rH \ rH rH rH rH CN rH CN \ o \ o o \ o rH CN rH CN n o H u <0 u to u < 0 H ID i w H 01 TS P 0) u £ rd cn -p X to -P tH >i 01 w c O w >1 M 0] 0 01 C S X X Cn ai q •H •H cn x; c -o •c C -p p 0 rd •w 0 c c 1 U •H 0 0 o Di 1 s X E 1 w -p c OJ H T3 rd M O x: -p c o X c 0 •H -P rd U 0 rH . OJ C HJ C« rd •H 0) CO V< Cn C o iH -P rd O 0 C o •H +J rd O 0 M M . a) C HJ Cn tfl H OJ CO u c 0 H -p rd CJ O C 0 H -P rd O 0 rJ . dj C HJ Cn rd •H OJ CO U Cn C o •H -P rd O 0 rH Cn e o co en c s: o HJ CO c •H 0 DnrH 6 rH 0 id o u OJ 0 rH DH Oh •H CO HJ C rH rd 3 0) X s C o Q US 73 •H X Dh D C 0 Q cn OJ u c OJ M OJ 1*H HH •H T3 0 z cn a o •H +J •H 01 0 ft -p c 0) •H -D rd u o 01 01 o c OJ u OJ MH HH •H o z 01 CD CJ c OJ h CD tH VM -H D 0 Z 01 S OJ -p 01 >1 CO

PAGE 61

-p c o TJ C C 1 a. o CP c H 3 CP C S T) OH h a o w m e •H ft s: -P O -H M (0 > TJ C -p -p i c o o m a) to 0) in n to -o o -P •H 01 >i o a) c u H W 0) Q) (C (0 c (0 to c o •H -p (0 TJ H +J to (0 3 Q to 0) M W O 0) (0 O -P to CD e TJ CD P (0 u a) a> 3 h h & rtj rH -H ^ o c a> o xs o Doom a. m -p a) « TJ i CN C (0 £5 (0 10 a) O (0 (0 X3 to en c > o K 01 Cn p. cd •J tn -h h | cd p. o fa g i »— I 0 o p 0) p cd x; to H Cn (0 P O ft ft fa ft fa * TJ 00 ID O 0> CN CN rH "V. in CN rH in cn c o -H p CO O o cd p CD c CD 0> Cn C O cn to oo o CTi CN r> to rH H CN rH CX) 'J' CO o CN rH to rH H rH LCI in o O CN CN CN rH H H "«» in O O rH tn r» rH o ro to CN o m o o CN O o o m tn cr> O CN CO o CN o in CO c 0 •H P 0) U 0 rH to iH tj cd 0 )H •H CO 0 4J 43 C m P i +j c o CO 0 6 X Ej H Cn 43 *J c P -u 0 c H 0 1 c 0 •H P cd u o M P. . 0) C P Cn cd H w c 0 •H p cd o 0 1-3 c 0 •H p cd o 0 ij u • OJ C P CP cd •h a) CO u c 0 •H p cd o 0 J c 0 *H p cd u o • OJ C +J Cn cd •h cu CO M Cn c o -H p cd o 0 h3 Cn c o CO 0) C JC p c o e 0 CO -H cd tt rH e rH 0 cd o Ih a> o rH H cd CD S -t J4 x: +j o CP +J cfl •H -H rH ffi ^ CQ fl II II 4 in to p to CD u p CO ft 3 p to OJ 3 CD u p CO c id o 0) TJ p. p m tj TJ C •H 3 S 0 CO >1 rH (fl 03 rH
PAGE 62

53 significant interaction between gradient and system for density of water column fish (Table 2-6) . Little Blackwater Sound (mid/east) and Buttonwood Sound (down/west) had significantly greater densities than Highway Creek (up/east) and Trout Cove (mid/west) (Table 2-7, Figure 2-6). Large roving fish analysis of variance . In contrast to the other two groups, for large roving fish, a clear effect of gradient position on fish density occurred. Fish in this group were significantly less abundant at the upstream gradient locations than midor downstream (Table 2-6) . No general locations varied significantly from the others (Table 2-7) . Spatial correlations . To analyze spatial trends, correlations between mean fish densities and salinity, salinity variation, and water depth were determined (Table 2-8) . A significant correlation between average density of large roving fish and station salinity was found; lower densities occurred at stations with lower mean salinity levels and greater temporal variability. In addition, both water column forage fish and large roving fish were significantly more abundant at stations with deeper water. Temporal Patterns in Density by Species As indicated by the correlations between mean densities and salinity, water depth and water temperature, temporal patterns differed among the species (Table 2-9) . No significant correlations were found between salinity changes from month to month and densities for any species. Temporal

PAGE 63

3 5 U XS OJ a> -p t7> Id (d 5 rH a) T3 > c (d (0 — >1 00 +j H -H II c C -H (0 c O Xi •H 0) P -H 3 u u oj H -P id (0 * — H O ft C o •H P id p oi XI o id a) o 4-1 (0 (0 4-1 O c o H -P (0 •H > a) X3 O T3 P C tn c r| > o r-l . oj «j tn rH >h O O id •H -H (J ^ 43 c • — 0) 10 H « 4J OJ + C TS rH >i Id P TJ C (0 •p oi oi c a) T3 43 (0 ^ •H O oj -p c oj § -p >i 0J -P C •H c o -H P (0 -a h c oj td r-i (0 o 43 u -p c o e 00 I 0J O (3 O ~ 0 m o 0J • u o OJ V > tt (0 p -p (0 c T3 Id o P 4H •H -rH 01 c c tr« OJ -H Q W 0 u (0 o rH o c 0J 43 -p • •H 0) S OJ 0J -H p c O 43 Q) O rH 0J o o 43 IT 01 C •H -H 4H (0 a) CP T3 (0 0J U rH o a 1 p •H c -H Id to u Id T3 B (0 -P 4H oo o rH C id o rH -H O -P a (0 S -H 0J > Eh 0J a >i -P •H c H rH (0 w >1 U o CP a) -p U 0J rH > I c o •H •P (0 rH 0J Sh Vh O O 0J rH (0 > I c o •H -p (0 rH 0) rH u o o at rH (0 > c o P (0 rH (1) rH r^ O 0 00 ro n H ^« CM o O • • • o o o CM H • O I CO o + o ID CO CO o OJ Cn (0 U o o •H 43 P C o « 0J CP id 43 01 •H fa

PAGE 64

55 T3 a> aS U > OS C o H P o P T3 c o u S a) o (0 •P (0 C (0 u o •H rH o 05 a> •H P H a c a> a) M 3 -P (0 M 0) c o -H 6 •o -p S X! *h w a) p •H rH rH QJ T3 C d) rH •H 4J fl> (C 4-1 (0 0) () >iO p p c •H P 0) m c o •H P (0 ui as u c w P rH > c o rH CN ro rH oo 00 in CN in O 00 ro • 10 ro o CN rCN O CN n 3 > o o o Q VC t — i V_J o 1 • P id o o o o Q Q o h at c H -rH I P o vr* CN CN VO CN M Id in VO CN in r-| VJ« in VO rH M o o o o o o o O 0) i 1 + + 1 I + P h « o k o • rH rVO o in o Oi vO in r| oo ro in CN VO rro CN vO CN VO ^* o x: id > CO o rro CN o rH o o pt i b> 1 — J \j r^ O o O o o c 0 u •H ai P VO CN CFV ro vO ro vO VO p M iu CN tn (— > CN CO in in Id * Q) o o Q Q Q VJ o (4 M f H> H> | + 1 H> + 0 O ro oo rH oo CJ\ vn CN rH o O o o o d d O o >l i P •H c M •r) r\ o rH "•1 (0 P in vO CTt oo rin 01 id rH O <* o rH rH ro o rH a) o O o o o O o O o fa + 1 + + + 1 + 1 i o o XI rH rH rH 0 01 0) 00 CD at tn id id id si JJ HJ 4J 4J 4J 3 3 3 P 1 Q) 0) ID 0) vV 0) CO 09 1 SB 2 2 a •H •r) > •> > o In "H 5* rv u« hJ H u H U 0] t) 01 0 H ~l to 3 HJ 3 (0 01 a, • O in 0) U •M 1) 01 01 q 3 •H S» HJ • 3 •H •H u *: •g »H M O a) HJ -H HJ §• 0 o> to a *! Q, •H 6 3 w O IB 6 o 0] •H H 0 •c •H m •r-> 0 6 O 0 Q> 0 c: J? HJ >H i rH 3 0 4) +j 3 4J >H Cv, 5 *: «»: to

PAGE 65

56 patterns in water temperature were significantly correlated with densities of several species. Lucania parva, Poecilia latipinna , Lutjanus griseus , and Strongylura notata were less abundant when the water temperatures were higher. In contrast, greater abundances of Haemulon sciurus were observed in warmer months. Periods of higher water levels in the study area (e.g. late fall) corresponded to periods when greater densities of Atherinomorus stipes, Strongylura notata, and Haemulon sciurus were collected. In contrast, Lucania parva was in greater abundance during low water periods . Spatial Patterns in Density by Species Analyses of variance . Density patterns varied for the top three species of benthic forage fish (Figure 2-7) . Results of repeated measures analyses of variance also differed among these species (Tables 2-10 and 2-11) . Poecilia latipinna was more abundant at the midstream locations, particularly Trout Cove (mid/west). Distributions of Floridichthys carpio and Lucania parva were not significantly influenced by gradient position or system. However, Floridichthys carpio was more abundant at Trout Cove than all other locations and Lucania parva was most abundant at Blackwater Sound (mid/east) . The top three water column forage fish species differed in spatial distribution (Figure 2-8) . Repeated measures ANOVA results also varied among these species (Table 2-12 and 2-13). Distribution of the silversides differed

PAGE 66

p c rC TS C 3 4-1 O CD P 03 -H C P o e 0) i X5 £! W -H •H 4-1 1 p H (/] c CD T3 C (0 CD a CD > O >i P H U] c 0) CD TS (0 rs c rC P m

PAGE 67

58 0) QJ9^3lu ajDnbs J3d qsij

PAGE 68

59 o •H X. P c 0) 4-1 o P -P (0 CP c •H (0 S w >i 0) P P (0 -H •h xi oi c gr TJ c a> 0 d) B C rH 01 (0 (0 0 -H p 4-1 0) 0) 01 a) c H x< id id (0 Q) rH c •H > P id (0 0) a) 01 > a) u OJ rH a> M P G e 01 03 (0 01 rC TS a) 0) C H s 0) p a -rH Tj 01 05 d) TJ 0) C P OJ (0 a) X! a 0) cC a p H 0) M « 01 01 rC > p XI to 0) •H 4-1 TJ P C a> 01 c c o u C 3 0) cr a o (0 0) rH rl TJ O o c c p (0 rH u rH (0 H 4-1 -H Q) O O ON CN CN r~ ro c Dwn Cm vo o CN O o CN H 0 rH ro CN o o 00 H •H TJ p CO o • O • CN • o • o • CN • in • -P CO ca »— ' 0) <0 r* »>i o o O o o o o Loc w "H -ii VD o rrj ro m w CO U <0 01 Q [ — CN I/) 10 * CN » — \ w A A » n) •H CN rH CN o rH o CN o rH O CN Loc X O o rH rH rCN rH .on TJ Cm rH rH W O rH rH •H o io CN o ro t~ in •rt X o rH O o O rH rH P rd • • t « • m *o o O o o o o o u 0 i3 Cm D <0 Q, CO a\ o CT> ciin M CN o cy» H *" c 0 0) IH o o o o on f e * CN CN CN CN CN CN CN -H 4H 4-1 o iH rH rH rH rH iH rH p -H »»» V. — . CD 0 TJ CN rH CN O rH o CN o rH o CN Loc a No VO rH rH rH rH rCN c 0 o •H Cm CN o O O o r~ r•H p ro O o o rO P n i carj o o o O o H O CO (0 • o • o • o • o • o • O • o OC J ui ON rH ro CN CO Sign, reater i-nan CO 0) •c Cm CN <• CN 43 «* rH • • • • • o A *; ro rH rn rH CN c • 0 ro CN •H u Tj c (0 S 0 o o o o 0 4-1 a CN CN CN CN CN CN CN -H , 0 -A Q) Cm cfl 0 4J (11 w H CO TJ § Cm i >i O O +J to (0 en a 0 w 01 4H C m jh >i -p to 0 0) 01 e • q o> 01 c s X X X rH m •H CD » o> OJ I? Cm C TJ P 0 • •H •H s: J3 •h n) CO CO u c c TJ c, c -p +j P Q) U >1 a 0 o CO 40 0 c a c rH 6 o w 0 0) 1 U H 0 0 0 3 w oq o 1 X X e o TJ OJ OJ u 4-1 4H 0 (0 OJ OJ h OJ TJ IH 0 u u 0 i OJ OJ u 14-1 0) HJ OJ u 01 OJ TJ 0) u in p 0 0] *

PAGE 69

P c o 0) 01 •H -a p -C c CD P CD c c (X CD CD CD a pu ur 0 p rH 0 (0 CD rH p O 0 -H c rH £2 0 (0 rrj (0 p -H rH rH rH (0 c CD 0 > 0) > -H 0 rH 4-1 u 0 TS c C 0) w (0 (0 £ -H e 05 CD c 0 rH p CD p c M 0) CD (0 TS (0 rs P CD CD c e e CD CD rH a U 0 73 Q CN] m co o CN n o C O o O 5 • • • ft o o o •H •U H 0 * 4H 0 ON o o o o o o o CO rrr> rH o O CN CN CN rH rH rH in o o rH in ft o o o o o o o o o * HH T3 o n in in CN rH o o CN CN CN rH rH rH in o O rH in s: c c U 0 0 c 0 1 X G • 0> U •H 0> u K •H tji w c • 01 u •H 01 co tn c 0 u 01 0 XT S
PAGE 70

-p c n c o Si (0 -P w o e 0) 0) M A 4-> 0) -p 0) 4-> w 3 0) H jC -h 4-> CO ^ S~( O (0 c u o o •H 4-> U u [fl 4-1 c o e (0 u Q) C CP W • >i Ci4-> 3 -H O (A m c Si S3 w •H i 4-> H m c 0) T3 C 10 0) s 0) CP (0 M o c e 3 C •H C o •H 4-> (0 H > n H ^ O (0 O TJ C 03 4-> W 0) S3 4-1 H -H u c H fa

PAGE 71

62 J8}91/\| ajDnbs Jad qsij

PAGE 72

i-t -P P C i a) -p p •H Ul Ul >i C Ul a) c X3 (0 •P •H +J > c a) a> -h O TJ C (0 (0 M •H tJl M (0 T3 > C (0 (0 (0 >i-H c > fd .p w c 0) 0 3 c (0 a) (0 ft a> q> S T( -o ui a) P C -H (0 CP o cn O P 73 0) B C o p u W (0 c a> to H Vh ft -P E to Q ft & H 73 H § to CD Si H u to to O E o q •H M 0) J5 •H -H -H CD g U •H 6 « o g * TJ ft * i*h TJ * MH 73 CT> CN ID in CO lO in O c~ m CO o CN O CN rH o o o O o CN o o o o d O o o o o o CN rH m CN o in o lO • • • • to rH o m CN CN rH Q Q Q Q CN CN CN CN CN CN CN rH rH rH rH r— ( 4 \ \ \ CN rH CN I— 1 rH CN rH CN rH rrH rH rH CTi rH O o o O O CO CO O o o O o n O O o rH o o o O • o o o o o o o H CN rH o 1*1 CO in "3" o • • • • • o CN CO Lfi CN CO rH CN rH o o o o CN CN CN CN CN CN CN rH rH rH rH rH rH rH "N. "•». \ "*V CN rH CN o o O O rH CN rH CN O rH r~ o r» r~ m CN CO rH o *t rH O O o o o O O O O o o o O O • o O o o o o O in rH CN CO m in CO CO rCJ» • • • • • • o co rH rH rH rH o o o o CN CN CN CN CN CN CN rH rH rH rH rH rH rH \ >v "•«, CN rH CN O o O o rH CN rH CN 01 c o -p c t o -rH a) 0) +J a) e CO •H to TJ p p * •r| 0) IB m to -p X i >i CO cfl ys •p St on 0 w w C & to c 8 X X 0) 9 (1) q 0 0) & •rl H D> u c c a 1 CO e o •a 0) CD (4 <4H 01 d) (1) M en a u o u u oj e o T3 0) i l»H o 01 ^ en ai a CD o u 0 W *

PAGE 73

CD jq CD p ur w 10. (0 lo c o 0 c •H irH > -H P -p •H rH 10 rrj W (0 P rH C M (0 CD 0 p o TJ c •H rH Q) XI Xj Cn (0 o P X! •H 0 rH > c Q) 0 (0 > 4H 0 0 CD CO C rH CT> (0 £1 C 1 •H H M > (0 C (0 > -H CP o 0 nt en rH 10 H 3 co (0 a) B T3 CD P (0 cd ai rH X! (0 •H rH (0 > o i p -p w c c a) a> -d a c 0) a • a> w •d -p C CD •rH C 0) U •H 6 to o D H to 01 3 >H O E o q •H 0) *! ft ft tn to H TJ H rj a * MH TJ ro O O O in CN — , in O lO O O ro ro (N o o o o CM o in ft o o o o o o o o o * HH TJ ft * t«H ro t 00 CO in • rH oo ro IT) o o CN CN CN rH rH rH *») ID "»». o O i-H in >D in rH ID <0 o O o o O o o • o d o o in CM o o in ro ro O O CN CM CM rH rH rH in o o 0) o u 3 0 w c tn 0 C •rl 0 P •H n) P O rti 0 o rH 0 rH rH C c -p 0 c 1 0 1 s c • Ol IH •rl 0< c • •rl 01 co c • D> U •r| Ol CO 0) C o u co o •r| *H >*H 0 U) ai i id Q) r 3HA< J3 +J O 0>+) (fl •r) -r) rH I JO II II II m vo 4-1 tn 0 3 0 0 u 0) « 3 tn 0) 3 (U rl g -P > id CO 73 0) TJ C U 3 P g O CO CO ft 3 (V > * O >iO id CO -P 3 0) O 0 rl Eh II II II rH CM CO

PAGE 74

significantly among the gradient positions. Atherinomorus stipes, the hardhead silverside, was more abundant downstream; Menidia spp. was more abundant upstream. Individuals of both Menidia beryllina and Menidia peninsulae were collected. The distribution of these species overlaps in northeastern Florida Bay, and distinctive characters are extremely difficult to confirm (C. Gilbert, personal communication) . Thus, Menidia spp. has been used in this study to designate these species. Anchoa mitchelli , although not influenced by gradient or system, was significantly more abundant at Little Blackwater Sound (mid/east) than at the other general locations. Patterns varied in spatial distributions for the top three species of large roving fish (Figure 2-9) . For these species, repeated measures ANOVA results also varied (Table 2-14 and 2-15) . Among these species, Haemulon sciurus was never present upstream. Lutjanus griseus was significantly less abundant upthan midor downstream. In contrast, Strongylura notata had significantly greater densities upstream/ east . Correlations . To further analyze spatial trends for these nine species, correlations between mean densities and salinity, salinity variation, and water depth were determined and are presented in Table 2-16. Densities of Menidia spp. were greater at locations with lower mean salinities and greater variation. In contrast, densities of Atherinomorus stipes were greater at locations with higher

PAGE 75

-p c CO a >i o xi p w -p c o e p xi H 0] 4-1

1 p H W B d) -0 c Id i P H Ul c 0) Cr>T> C H C > -H o p 0) 0) T3 X! P 73 P C (0 •H T3 C 0) ITS •H W Cr> o) a) W 4->

PAGE 77

d) -p ^ 0) (0 o >1 p •H 03 c o TS P -H tH o w •H 01 >1 rH (0 c 03 03 (0 (U B T3 C 0) 0) c •H rH 0> M O rs c •H w (0 S o -p to >i to T3 C (0 -p c 0) -H T3 (0 u & w c 03 c o H p (d rH 3 D id o c o O -P (0 w o p e 03 3 c (0 03 (1) rH X) (0 H *H ro > o •H u a 03 I p •H M (0 tr> o o p 10 w •a •H > 0) p u >i o p c a) T3 Tf p (U CP O vD rrH CN kO O o o o o in CD • o o o o o o o CN CM in ro in CN ro CP o\ r» • • • ro ro CN CN d o CN CN CN rH rH rH CN rH CN oo CO 00 CO 00 in rrH CP r> CN CO in rH o O ro rH vo co o ro \o o o > 00 •* r-. 00 o CP r*r <£) rin in cp 1*1 CM r» yo rH t» rH r 00 rH CO O o O o O O o <£> CO CP CO C~ o CO rH O CM CP rH CN O o O o rH rH CN CM CN iH rH rH CM rH CM 00 00 00 00 •=f 00 CO o u D o to to o H 40 IB +J to q H 0 to TS 4J e to •H c rd to 1 4J m to O to en U cd >i to OJ X X en c to CP CD q c c •H CP 0 o C -D c to to 0 cd 0 rd rd M •H d) 1 to X c o to rd Q) w c 0 •rH -p rd O O h) U . Q) C 4J cp id •rH
PAGE 78

o rH ttl rH OO CN O o O n CTi rH o CN ro rH O rH 00 00 O CN B 0 CO •H c -P 0 rd •rl 0 P 0 CD rH U 0 rH rH rd ^ rH CD CD rd c C 14 0 CD CD CO cn c rd CD 0) X Cn a) c cn cn 0 C c CD 0 0 CD1 CD w c rH 0 Id -H a> co c o a rH 0 IT) -r) M 4J d) (I) C O CD O O rH C rH 0 ft) -r| d) id c o (U o O rH u • a> _ C +J c cn id •r| CD CO 5h Cn c o r| 4-> CD id c o CD O C5 rH 0 U HI 0 H MH cn U c id cn 0 ft C cn s rd rd 0 CD CD u s cn CD Cn rH rH c rH ft 0 id •rl -P rH 3 s CO u "J! MH o tn (D CD In Cn CD "0 M 0 u u cn c o •rl -P rd O O Cn rH CD XI rH -P 4-> co m id id CD CD \ — o> el }| id CD CD aj u u e co co o c h s e o a CD U 4J CO ft U 3 CD TJ 4J id X 3 CD ^ CO 0) o U rd Sh O rH (D to 4J id CD 0> 4J id •r) -rl rH B rH c 3 O rd 0 CO II II II in vo I CD U 4J cn ft 3 4J 01 s CD >H 4-1 cn c S 0 CD V U 4-> v CO TJ "0 C •H 3 E O co CD T3 > O id CD U O CD >-) a CD II CP 0 c 0 4-> 4J 3 rH CN CO

PAGE 79

70 P u cu > rs o u T3 0) CD -P cn (0 • (0 > -p M > c «J (0 o 00 >i H +J II -H c c * »H c o •H p (0 4J w o id CD M o 1 M •P (C 73 (0 (0 0) T3 U O -P O -H rH a 01 c CD T! c a) CD p 0) X! W c o •H p (0 H 0) rH rH o u c -p X T3 0 tr> c o -h rH rH — >H a> w e p •H >H (0 tP O (0 M o o tt -p T3 C i P •H c -H rH W 73 C rC w £ p c o g 73 CD P U CD > C o o in CD >H CD > rC P (0 73 >1 P •H cn c CD Q o V p c (0 o -H t*H H c CP -H t/1 40 ft 0) a 14 Q) -p m s >1 4-> H C •H rd 2.2 IS Q 4-1 -H c H rH rd W O x; 4-1 CU m cu •h o cu ft w 0) 3 rH rd > l ft C o -r4 rd rH a) 1H u 0 u a) 3 rH ft) > I ft c o •rH 4-> n) rH 0) V4 r4 0 O Q) 3 rH rd > I ft C 0 -H 4-> (0 rH CU U u o u ^* in o to in o ro O o O o o O o o O 1 1 1 + + 1 I 1 1 in to in r> ro t-~ co in O in ro CN in CO o CM o to o a\ rH CN o tH in ro r» to o cn O o o to rH o o o o O o o o O rro o ro CN o CO tH r*> o rH O in to in tH ro • • • • • o o o O O o o o O + + + 1 l + + 1 + CD CD cn CD CD CD 4-1 4-> 4-1 4J 4-) 4-> CU CD CJ > > o H (d 0 tn •d 4J o "H •o •H O H C rd • > 4U rrj 0 td H u ft •H ra B rd •H H rd c: "1 O rd 0 0 0) 3 0 a. 5 CD Q) ft rd H 40 40 CD rd CD CD 3 40 3 CU 0 >H CD CD 3 3 •H •H -4 £ rd 0 i 0 >H CD CD 6 3 rd o CD "H 3 0 H •H c; cS M o rH rd 3 "H QJ '1*1 0 E c; 40 >H Q) 40 3 40 rd *: «< •Hto

PAGE 80

71 mean salinities and less variation (Table 2-16) . Lutjanus griseus also appeared to avoid the lower salinity, more variable areas. Floridichthys carpio was significantly more abundant at shallower stations in the study area. Species correlated with deeper waters were Lucania parva, Lutjanus griseus and possibly, Atherinomorus stipes. Community Patterns Species richness . Species richness differed among the fish groups and for all fish combined (Table 2-17). A total of 305,589 fish from 77 species was sampled using both the visual census and enclosure net methods combined. Midstream locations in both the east and west systems had the lowest species richness of the three gradient positions. Among the fish groups, benthic forage fish were also least species rich at midstream locations. Water column forage fish had distinctly lower species richness at Little Blackwater Sound, with only four species but great abundances. For large roving fish, upstream and midstream locations were lowest. For number of species alone, large roving fish species followed a clear gradient from upstream (10 and 11 species) to midstream (13 and 14 species) to downstream (20 and 22 species) . Cluster analysis. Results of two cluster analyses for species collected with enclosure nets were illustrated using dendrograms (Figure 2-10) . The initial analysis classified stations based on presence of species. Most of the upstream

PAGE 81

Tl 1-1 IVJ UJ i) PI C I m fli (T\ > U vJ QJ A. rv-tl UJ /rt UJ ru X) UJ ... j •r*1 •>> •rH, II II O M r" v .a __j "rn 0) rH 0 w 0) W a w c Cn u ft) ft tn -H fa a) en rain o fa c £ 3 rH 0 o (-1 p id * J5 cn •H fa a) Cn rfl M 0 fa u H •P c 0) PQ oj c tn 0) rH 0 (1) ft w •P 0 M H c 3 2 tn 0) •H O ai ft CO « -H P J 0 c a) C CO tU u, -H 1 ° ft ft 2 CO rB P 0 C (1) 3 2 0) a> •H o U» UJ UI CO in 0> O cn rH cn CN oo 00 cn oo n rH in fa rH CM co o * r> in CN f> in CM ro CM O ro 00 cn VO (N vO • CO rn i" n in rrH in o o ro CN in rH rH CN rH rH CN ro O r~ O r> cn VO ro tn cn cn CN ro rH (N in 00 CN in ro CN 10 cn rH in rH 00 00 KD o in • • in ro in vO iH rH cn CN ro CN rH CN CN rH CN ro •H cn CN in «* Ol O vj * « m vo ro vo CN H O ^ f> ro in r> vo s | tn p l c 0 Q P tn to tt) i ft D -P tn rd 0) I c 3 0 Q t ro cn GO

PAGE 82

UJ r" >-« H oi vy H \J ' UJ c n H ._j o. ™ ni rn *H r< l-< 1 \ i-> +-* rr* /II 0) 0 -H rn UJ 01 w Jj Tl I—* 0 flJ jj 0) Li Uj iH CP >— i >> H M rrt vj 0) UJ +^ , — | W a) o , — | •H flj ffl Q {J) (1) c rr \j ffl rv w UJ r»— < 0) 1) 01 -i-< ** (J ' 01 OJ C UJ w 01 ffl Li *> > 01 iH (1 1 UJ ITI [»J tfi U_l UJ <-* rn cr TO vj rn w (It UJ M-J rn UJ 01 Li CL m 0) U (0 rn UJ CD M-t 01 -P i-P "H •H OJ -H w u H ffj 01 00 a) 0 C C (0 a) id oh a •H 0 w u -P en ,c u -p -p g 0 01 W a> f0 d W 73 0) H (0 , 0 C «3 C W +J H CO) >i-H 1 (J &II1H Ul
PAGE 83

74

PAGE 84

75 stations, including one in Joe Bay and all in Highway Creek, grouped separately from those located midand downstream. In a second analysis based on densities of each species, all but one upstream station clustered separately from the midand downstream locations. For the visual census, cluster analysis results were also graphed using dendrograms (Figure 2-11) . Based on presence of species, all the upstream stations formed one cluster. Based on densities of all species, three of the five upstream stations clustered together. These were the three that were most upstream. For each cluster group defined by the cluster analysis, the most common or dominant sets of species were identified. Three species that commonly occurred at all stations in the study area were: goldspotted killifish (Floridichthys carpio) , rainwater killifish (Lucania parva) , and redfin needlefish (Strongylura notata) . Species that were very common in upstream stations included the inland silverside (Menidia spp.), clown goby (Microgobius gulosus) , tidewater mojarra (Eucinostomus harengulus) , striped mojarra (Eugerres plumieri) , and Mayan cichlid (Cichlasoma urophthalmus) . Downstream species commonly included hardhead silverside (Atherinomorus stipes) , gray snapper (Lutjanus griseus) , silver jenny mojarra (Eucinostomus gula) , great barracuda (Sphyraena barracuda) , blue-striped grunt (Haemulon sciurus) , and snook (Centropomus undecimalis) . Midstream

PAGE 85

10 c o H 4-> o (0 -H c 3 01 o • 01 0 •H Q< c CD > •H >i C T3 o xi q) (D Ul CD CD <0 1— 1 >-l 01 & £7> (TJ c c H 01 (0 73 >, CD 01 -J-> Ul i 4J CD r— 4J T3 -P CD d) g -H p r— t 01 (1) a) (0 CD X! (11 T3 Q. 0 • 0) •H 14-1 01 -P a) •H (0 H 01 01 X! 4-1 0) 0) >1.p H -H 0) rH 01 0 •H (0 01 01 O) 0 c c Id & CD JJ OH 01 a •H 0 01 M 4-> o) x: B (0 u 01 w 0) 3 10 13 HI U • O (0 i h x: to c oi +j H+J C 41 >i-H I 0) CP 01 rH 01 n gH mm c 01 U C d) a> oi a> a<
PAGE 86

77

PAGE 87

stations had mixtures of both upstream and downstream species excepting the Mayan cichlid, which was only common upstream. Residency . Abundance and numbers of species by residency category varied among the general locations (Table 2-18) . All 33 benthic forage fish species sampled were permanent estuarine residents. Among the water column forage fish, seven of the nine species observed were residents, while two (Clupeids) were occasional visitors. Large roving fish were represented in all three residency categories. Six species of large roving fish were permanent estuarine residents including needlefish, catfish, bull sharks, and stingrays. The vast majority of large roving fish were transient juveniles, however. Freguency distribution by size varied for 6 transient species of large roving fish (Figures 2-12 and 2-13) . Clearly, the mangrove habitats of northeastern Florida Bay were nurseries for Sphyraena barracuda most of which occurred in juvenile sizes (Figure 2-12) . Snook, however, did not occur in juvenile size classes. Although adult-sized Lutjanus griseus appeared to share the habitat with largersized juveniles, since no gray snapper sampled was smaller than 7.5 cm, one can assume that young-of-the-year juveniles occur outside the mangrove habitats sampled in this study. Habitat use patterns similar to Lutjanus griseus were observed for Lutjanus apodus , Haemulon sciurus , and

PAGE 88

0 T3 c a) i vO rH ro 00 CP o ro rH ro 00 rrH cp CN ro in rH •* CN CN «* O CO 00 f» V0 LT) rH VD in oo VO VD cn CN vO rH ro a> cp in CN vo ro CN CN rH CN in ro CN rH CN ro in CN in •tf rH CN rH in m in CN CP in rH CN ro VO rH rH ro CP o ro rH rH VO o O *t vO VO vO rH ro rH in in vO ro LO CN CP m CN in rH (1) OJ OJ o tn u CO u CO c OJ c OJ c a> CO •H cfl rl cfl •H "D CJ V O *o O C OJ c OJ c 3 ft 3 ft p & X) CQ X) CO X! m CO cfl rd x: 0) tn X! XI CO CO •H -iH MH MH ai a) & o> n) « M U 0 o MH >H u o •rH -rH x: xi p p c c OJ CD 03 CQ •H -rH MH IIH ai oj o> o> id id M U o o MH MH x; x: CO CO H -rH mh m 0> 0> c c c e rH O CJ u u 01 (U -p -p id id 5 » OJ OJ H tH id id 1-3 hH0) o c; to Mh >»H o it) ft tn ft) u to Q) •H 0 OJ ft to >*H o c; •-H ft tn ^ o OJ "H H OJ s> 3 OJ -H •H 0) 3 h> •u d id M U O 0 MH MH o o -H -A XI X! V +J C C OJ OJ CQ OQ •r| .r4

p -p id « CN Ol o ro CN CN 00 ro CO cn 10 10 OA ro in cn CN O rU3 oo oo vO CN in rH ro o OJ OJ ai CJ tn 0 CO o CO c OJ c a> c aj cfl -H id •rl id •rH n u V O V O c Q) c OJ c OJ ft a ft a ft XI to XI CO XI CO cfl id « x; X! tn CO x: x: co ca •r| -r| o> c c -r) > > 0 o u u aj cu o> CP u u id id 4 rJ id 0) cfl 3 0) -SS u ft 3 o o o H *H •o w m cu OJ CP CP id id u u 0 0 •4H >U V •rl x; p c OJ OQ CQ O CN CO OA rH rH vO vO ro rH ro 0> O iO VD CN CN cn in o» vo CM CP vO in in CN ro rH ro in a) o c id V c 3 XI cfl OJ CJ c id T) c 3 XI id tn OJ •rl CJ OJ ft 01 x; x; 01 m •r| r| MH VM d) OJ CP CP Cfl cfl -J 0 0 MH MH c c £ £ 3 3 rH rH 0 O 0 O u P. OJ OJ p p cfl id 2 2 -c x; co co •rH -rH MH VH CP CP c c •rH -rH > > 0 o u u at cu o* o> u u id id (-} rJ

PAGE 89

Figure 2-12. Lengthfrequency histograms based on visual census data. Adult size given whenever the information was available from the literature. a. Sphyraena barracuda (great barracuda) , and b. Centropomus undecimalis (snook) .

PAGE 90

81 Abundance by Size Class 0 15 30 45 60 75 90 105 120 135 150 Total length (cm)

PAGE 91

Figure 2-13. Lengthfrequency histograms based on visual census data. Adult size given whenever the information was available from the literature for a. Lutjanus griseus (gray snapper) , b. Lutjanus apodus (schoolmaster) c. Haemulon sciurus (blue-striped grunt) d. Archosargus probatocephalus (sheepshead) .

PAGE 92

Abundance by Size Class 83 0 10 20 30 40 50 60 70 Total length (cm)

PAGE 93

Archosargus probatocephalus , although adequate life history information is not available for these species to determine size at maturity (Figure 2-13) . Overall, the community was dominated by residents in both numbers of individuals (91%) and numbers of species (60%) . Occasional visitors accounted for 28% of the species but less than 1.0% of the individuals. Estuarine transients (juveniles and adults) comprised 12% of the species and 8% of the abundance. The distribution by residency varied among the locations (Figure 2-14) . In terms of abundance, permanent residents dominated at all six general locations. The trend was most pronounced upstream. Transient individuals were somewhat more prevalent at midand downstream locations. Although less than 2.0% of the abundance, occasional visitors comprised 15 to 21% of the number of species overall. Discussion Methods Enclosure net methodology . Using the enclosure net method, direct sampling of fish occupying submerged red mangrove shorelines was possible. The fine mesh net captured the small forage fish that numerically dominated the habitats. Inevitable escapes, especially by large roving fish, probably occurred, however, as persons deploying the net waded up to the sites. The recovery efficiencies of fishes encircled in the nets were comparable

PAGE 94

Figure 2-14. Comparison of residency among the general locations based on percent of total abundance. Data used calculations were taken with enclosure nets (for benthic forage fish and water column forage fish) and by visual census for large roving fish. a. Abundance of fish in each category. b. Number of species in each category.

PAGE 95

86 Residency Occasional Transient Resident Number of Species 100 F F=i i 1 i =i i 1 Up Mid Down Up Mid Down West East

PAGE 96

to those found by other investigators who targeted small fish (less than about 7.0 cm in total length) in vegetated, shallow areas using (Weinstein & Davis 1980) . Although rotenone reguired cautious handling, fish immediately began re-occupying sites after nets were removed. Rapid rotenone degradation in the relatively high temperatures which prevailed in these waters (Nielson and Johnson 1983) , and dilution due to the turnover of water at the relatively open sites prevented cumulative adverse rotenone effects. Other factors could have affected the variable efficiency of recovery from individual nets. Turbidity from disturbed sediments reduced the ability of the collectors to capture fish with dip nets. In addition, variation in winds onto the sites may have caused rotenone effectiveness to vary. Visual census methodology . The visual census methodology had several advantages for sampling fish in mangrove habitats. The speed and flexibility of the method permitted sampling in a broad range of habitats that could not be sampled with the net, including mangrove locations with greater depths and a wider fringe. The visual method was non-destructive, thus allowing evaluation of the persistence of use by repeated observations of the same stations and fishes. The problem of net avoidance was eliminated with this method. In addition, large fish (e.g. tarpon) that would normally escape nets and trawls could be surveyed. As in studies of streams and coral reefs in which

PAGE 97

the visual method has increasingly been used, it has possible further value for use in studies on behavior, species interactions, and microhabitat use in mangrove habitats. Direct observational sampling also has disadvantages. Somewhat surprisingly, many fish were attracted to the snorkelers (Dibble 1991) . Fishes such as snook, tarpon, gray snapper, cichlids, bluegill, and killifish, frequently came within a few inches of the snorkeler's clipboard, presumably out of curiosity. As verified in the snorkeling efficiency tests, this attraction led to double-counting of individuals on the multiple swims. A major disadvantage of the method was its sensitivity to reduced visibility conditions. Although the fishes approached more closely in low visibility situations (1.0 to 2.0 m) , the uncertainty level in identification of species was often increased. Overall, those species that tended to remain at the bottom or far back in the fringe were surveyed less accurately in deeper and low visibility sites. This problem was a particular disadvantage for surveying the benthic forage fish at deeper stations. Even in depths of only 1.0 m, the use of SCUBA equipment might be an advantage if visibility range is less than water depth since it would permit better sampling of the bottom. Some small, rare fishes and shy cryptic forms were less accurately sampled with the visual method. Finally, the occurrence of large

PAGE 98

89 dangerous predators may preclude visual sampling in certain areas. Sites in deep creeks were abandoned based on sitings of large sharks and alligators. Combined methods Using two methods to complement each other in the same habitats and regions had great advantages. The two methods increased the range of mangrove shorelines that could be surveyed. However, when using two methods, interpretation problems can arise when neither is 100% efficient. If the results do not agree, it is difficult to determine if the discrepancies are due to differences in the fish sampled or due to differences in the efficiency of the methods. In this study, the two methods targeted different size groups of fishes. Overall, benefits of using the two methods clearly out -weighed the disadvantages. Fish and Salinity Salinity . As confirmed in this study, salinity conditions in the area vary from year-to-year (Ginsburg 1956) . The east and west systems differed greatly during the drought conditions that prevailed during the main study year. They were very similar, however, during the pilot study year when rainfall was locally more plentiful. Under low rainfall periods, the C-lll Canal may effectively block most freshwater from flowing into the western system by routing it towards the east. During high flow periods, however, when more freshwater is available for distribution, the east and west systems appear to have similar salinity patterns.

PAGE 99

90 Historically, such annual differences due to local rainfall variations may have been more moderate. Before freshwater wetlands were extensively drained, freshwater probably gradually seeped into the study area from the greater Taylor Slough drainage basin, resulting in more dispersed distribution patterns and prolonged periods of lower salinity levels. Temporal Patterns . The first hypothesis proved to be incorrect: none of the temporal patterns in density of any fish group or species examined was attributable to changes in salinity. In the study area, therefore, the fish do not seem to react to salinity changes by short-term movement in and out of the general locations (regions approximately 12 km in area) . However, patterns for all species collected were not individually analyzed. Thus, some short-term relationships may be identifiable on further examination of the data. Temporal patterns were related to temperature, however. For benthic forage and large roving fish, as temperature increased, density decreased. Similarly, Thayer et al. (1987) and Tabb et al. (1962) also found greater densities in western and central Florida Bay in the late fall and winter when temperatures were cooler. Temporal patterns for the current study are not typical for estuarine fish populations; usually peaks occur when freshwater inflow is greatest (Gunter 1967, Weinstein 1979, Yanez-Arancibia et al. 1980, Rogers et al. 1984, Stoner 1986, Flores-Verdugo

PAGE 100

91 1990) . Judging from the salinity data, this would have been late summer for northeastern Florida Bay in the study year. One possible explanation for this unusual condition may be that in the summer, high temperatures combined with low circulation to create a stressful environment for fish in the Bay (Moyle & Cech 1988) . This is supported by reports by fishermen of fish kills in Florida Bay during those hot summer months (M. Robblee, Everglades National Park, personal communication) . Additionally, the density of the major component of the large roving fish group, Lutjanus griseus , probably accounted for much of the temperature related trend in large roving fish densities overall. The larger individuals of this species, migrate offshore in the summer, when spawning occurs, and return in the winter (Starck & Schroeder 1971, Rutherford et al. 1989). This migration may account for the reduced densities of large roving fish in the summer. Spatial patterns . The density of fish decreases from west to east in Florida Bay (Sogard et al. 1987, 1989a). This trend appears to continue into the northeastern Bay (this study, Funicelli et al. 1986). Using an almost identical method of sampling (enclosure nets) , the mean fish density found in western and central Florida Bay mangroves by Thayer et al. (1987) was 8.0 fish m" 2 , compared to 3 . 3 fish m~ 2 found with nets for the northeastern Florida Bay area.

PAGE 101

9 As indicated by comparisons from upto downstream, salinity regime did not affect the overall density of benthic and water column forage fish. Furthermore, in the current study, no differences were found between the eastern and western systems. Similarly, Thayer et al. (1987) found no effect due to gradient for mangrove fishes collected with enclosure nets. The sites sampled by Thayer et al. (1987) , were located in central and western Florida Bay — from downstream near the Keys, to upstream in Whitewater Bay and Coot Bay — and their collections were dominated by small forage fish. Thus, salinity regime may have little influence on densities of forage fish species throughout Florida Bay. The individual species of benthic forage fish were likewise distributed widely and not systematically along the gradient. Although small benthic fishes have certain other life history characteristics that may explain the widespread distributions observed (Sogard et al. 1987) , most species are notably euryhaline (Robins et al. 1986, Nordlie & Walsh 1989) . Thus, they are good colonizers of all types of habitats found in Florida Bay. In general, high variance was evident in the density estimates for water column forage fish. This was due, in large part, to their schooling behavior. With the nets, a school was either collected (and thus the abundances were great) or not collected (and thus the numbers were zero) . Individual species of water column forage fish, however, do

PAGE 102

93 appear to be systematically distributed along the salinity gradient. Although only one rough silverside (Membras martinica) , was collected in the current study, they were abundant at central and western mangrove sites (Thayer et al. 1987) . Abundant Atherinomorus stipes were collected by Thayer et al. (1987) at his most downstream locations in central Florida Bay (Crane Key and Captains Key) . This corresponded to the very abundant collection of this species at the downstream-most locations in northeastern Florida Bay (Blackwater Sound and Buttonwood Sound) . Since salinity regime correlated with the distribution of Atherinomorus stipes and Menidia spp., relative densities of the species in the family Atherinidae may be indicative of salinity conditions . The density of large roving fish was dramatically lower upstream than midor downstream. Among the large roving fish species, however, the influence of salinity regime on spatial distributions was mixed. Blue-striped grunts and gray snapper were less abundant upstream, but redfin needlefish were more abundant at the upstream/east location than elsewhere. Thus, some species may be limited by the conditions that occur upstream, while others tend to thrive there. Community patterns . Although no systematic pattern of distribution occurred along the gradient for benthic and water column forage fish, for the large roving fish, greater numbers of species occurred downstream. In many other

PAGE 103

estuarine studies greater numbers of species have been found downstream (Weinstein 1979, Yanez-Arancibia 1980, Rogers et al. 1984, Sogard et al. 1987 & 1989b, Thayer & Chester 1989, Lonaragen et al. 1990) . The species usually responsible for the greater downstream richness are adult members of the marine-visitor group of fishes. Residency . Overall, the dominance over the entire study area by permanent residents (91% of abundance) was unusual even for tropical estuarine systems (Yanez-Arancibia et al. 1980, Davis 1988, Morton 1990) . Only three species (13 individuals, all adults) were members of the reef community (Acanthurus chirurgus, Aluterus scriptus , Sparisoma radians) (Jaap 1984) . The islands of the Florida Keys may inhibit connection of the fish community in Florida Bay with that of the extensive reef tract adjacent to Florida Bay (Sogard et al. 1987) . The mudbanks in the central and western Bay may further inhibit travel into northeastern Florida Bay from the Gulf of Mexico. Perhaps more significantly, the moderating influence of the thermally stable Gulfstream water masses do not enter the Bay. When temperatures reach extreme low (or high) levels, those reef species that have entered Florida Bay may be forced to migrate or be killed. For example, a doctorfish (Acanthurus chirurgus) that was observed every month at a downstream/west station from May through December, disappeared once temperatures began to drop.

PAGE 104

95 Sciaenids (drums) and bothids (flounder) , major nursery species in other estuaries in the southeastern United States, Gulf of Mexico and the Caribbean (e.g. Roessler 1970, Lindall et al. 1973, Weinstein 1979, Yanez-Arancibia 1980, Stoner 1988, Sheridan 1991), were not collected at all in the present study. Since spawning takes place in the Gulf of Mexico, young juveniles may not survive the journey from distant passes into northeastern Florida Bay due to lack of tidal exchange and little circulation in the central Bay (Sogard et al. 1987). Mangroves in northeastern Florida Bay clearly are nursery grounds, however, for several species of estuarine transients, all of which are popular sportfish: gray snapper, schoolmaster, blue-striped grunt, sheepshead and great barracuda. They also support adult snook and tarpon, species that are known to use similar mangrove habitats as nursery areas elsewhere (Gilmore et al. 1983, Seaman & Collins 1983) . A common life history pattern occurs among these fishes: recruitment from offshore as post-larvae, settlement and growth in inshore areas, and movement back offshore or to deeper water as they attain larger size classes (DeSylva 1963, Starck & Schroeder 1971, Jennings 1985, Robins et al. 1986) . Since these species comprise a major portion of the large roving fish group, aspects of this basic life history pattern may indicate why they were less abundant as a group in the upstream locations.

PAGE 105

Submerged aquatic vegetation (SAV) , such as algae and seagrasses, may provide cover for newly settled forms of estuarine transient juveniles at a scale compatible with their size. As the fishes grow, however, the SAV becomes less likely to provide adequate cover for later stage juveniles, and they may seek larger forms of structure for shelter. A habitat expansion of this type was identified for juvenile gray snapper by Starck & Schroeder (1971) . While smaller snappers dwell in seagrass beds, larger juvenile snappers congregate near mangroves and other brush during the day and return to feed in seagrass beds at night. The smallest gray snapper individuals found in mangroves in the current study were 7.5 cm, the same size indicated by Starck & Schroeder (1971) at which snappers expand their habitat use. Seagrass beds are generally poorly developed in northeastern Florida Bay (Zieman et al. 1989) . Abundance of SAV is temporally variable and SAV is often absent altogether upstream (Montague et al. 1988) . Without SAV, young fishes may not have an adequate intermediate habitat between the planktonic and mangrove stages in which to settle. In addition, if larger juveniles (over 7.5 cm) do wander upstream, they may find inadequate food resources; many of the benthic invertebrates that they consume live epiphytically on SAV and may not occur in adequate abundance levels without SAV (Montague et al. 1989) . Thus, lack of SAV may result in both reduced fish recruitment and growth

PAGE 106

97 rates. This scenario may serve as an alternative or complementary hypothesis to salinity intolerance, and lack of access from distant passes, in explaining the lower overall abundances of large roving fish observed in mangrove habitats upstream.

PAGE 107

CHAPTER 3 FISH COMMUNITIES IN FLORIDA BAY MANGROVE SHORELINE HABITATS: RELATIONS WITH PHYSICAL PARAMETERS AND COVER Specific features of mangrove habitats may contribute to the structure of fish communities in predictable patterns. In the only study to specifically address this question to date, limited support for this concept was found: certain species seemed to prefer mangrove habitats to more open sites (Sheridan 1991) . Other investigators have identified patterns between fish community structure and development of vegetative structure in seagrass beds (e.g. Sogard et al. 1987, Thayer & Chester 1989), kelp beds (Ebeling & Laur 1985) , and littoral zone plants in ponds (Werner et al. 1983) . In addition, the "rugosity" and vertical structure of coral reefs have also been identified as factors in increased densities of some species of fish (Luckhurst & Luckhurst 1978) . One role of structure in aquatic habitats is to protect vulnerable prey fishes from predators (Werner et al. 1983, Ebeling & Laur 1985) . However, since food resources can become exhausted or be of lower quality in vegetated habitats, the safest refuge is not always the location that 98

PAGE 108

99 vulnerable fishes choose (Werner et al. 1983, Schmitt & Holbrook 1985) . In addition, since recruitment in aquatic systems is largely based on widely-dispersed larvae, the occupation of particular sites may be based on chance vacation of living space by a previous occupant and the largely unpredictable occurrence of available recruits from the plankton (Sale 1980, Sutherland 1980, Sale & Douglas 1984) . Thus, the prediction of habitat use is a complex problem involving recruitment, species interactions, resource availability, and random influences. If fish are not randomly distributed within the mangrove shoreline habitats of northeastern Florida Bay, it may be possible to identify features correlated with fish densities. Density can be used as a quantitative approximation of habitat quality (Sogard & Able 1991) . Thus, the objective of this portion of the study was to analyze density data and habitat information in northeastern Florida Bay to determine any habitat preferences among the fish found in the mangrove shoreline. Materials and Methods For both the visual census and enclosure net methods, fishes and salinity were monitored at each station repeatedly over the period of May 1989, through May 1990 as described in Chapter 2. These stations were located across a gradient from upstream near sources of freshwater inflow to downstream (Chapter 2). A total of 328,960 fish

PAGE 109

100 were censused or collected in enclosure nets and visual samples. Due to the nature of the sampling methods, the values for habitat structural variables were determined using slightly different procedures for the two fish data sets. Each of the 17 visual census sites analyzed consisted of 8 substations. At each substation, a transect perpendicular to the shoreline was designated. Since corresponding fish censuses were conducted at each visual census substation, all 13 6 substations were used in the analysis. In each of the 18 enclosure net stations, 3 transects were designated within the area repeatedly enclosed by the net. Since corresponding fish data were available for each net site as a whole, mean values for the 3 transects within each of the 18 stations were used in the analysis. For all 190 transects, data were collected using a 1.0 m frame. Starting at the shoreline edge, the frame was placed every three meters outward along the transect to 6 m off the mangrove fringe. The following data were collected using this overall method: Water depth: Within each frame, four measurements were taken, one in each quadrant using a meter stick. Fringe width: Along each transect the distance from the shoreline to the outermost mangrove fringe was measured. Tree height: Within each frame, four measurements were taken, one in each quadrant using marked poles or by visual estimation for trees taller than 2 m.

PAGE 110

101 Tree cover . The percent of canopy coverage over the frame as observed from the surface of the water was estimated. Prop root size and density . Calipers were used to measure the diameters of red mangrove prop roots. Observations were divided into two groups: those less than and those greater than 2 cm in diameter. Numbers in each category were counted within each frame. Few prop roots less than 2 cm occurred downstream. To avoid biasing the analysis, only the data for prop roots greater than 2 cm in diameter were used. Submerged aquatic vegetation . Within each frame, the total volume of seagrass, algae and detrital material was determined separately by placing a meter stick in each guadrant and noting the height of each component above the substrate. The percent areal coverage of each component was also estimated for each guadrant. These values were summed for the analysis. Salinity mean : The average value for all months at each station was determined from measurements taken with a ref ractometer . Salinity variation : The standard deviation for each station was calculated from the concurrent measurements . Analysis Datasets containing 136 visual census substations and the 18 enclosure net stations were analyzed separately. For comparison, means and standard deviations were determined for each variable. To select the most appropriate method of analysis, correlations among the variables were also calculated. Multi-colinearity among several of the variables was revealed in the correlation analysis. Thus, principal components analysis, a multivariate technigue designed to discern factors that have generated interdependence within a data set was used (Afifi & Clark 1984, Robblee 1987,

PAGE 111

102 Grossman et al. 1991) . Principal components of the habitat variables were calculated for both the visual census and the enclosure net data sets using the SAS correlation matrix and varimax rotation methods (Afifi & Clark 1984, Smith & Duke 1987) . Each component was interpreted by examining correlations between the original variables and the derived components . SAS FACTOR was then used to calculate individual factors from the principal components for each observation in the original data set. These uncorrelated factors were used as the independent variables in a multiple linear regression with log-transformed fish densities as the dependant variables. Fish density data were divided into 3 groups of species for the analysis based on their size, mobility and position in the mangrove habitat. These categories were benthic forage fish, water column forage fish, and large roving fish as defined in Chapter 2. All species were also analyzed separately. Only principal components with eigenvalues greater than the proportion of the variance in the data that could be explained by an individual variable (i.e. those > 1.0) were used in the regressions (Grossman et al. 1991). The slope of the regression line describes the nature of the mangrove habitat and fish relationship. The variance in fish density explained by each component (R 2 ) indicates the strength of this relationship and also provides a basis

PAGE 112

103 for comparisons of mangrove habitat/ fish relationships between fish groups and species (Robblee 1987) . Results Relationships Among the Habitat Variables Correlations among habitat variables . For visual censuses and enclosure net data, mean values and correlations among the habitat variables differed (Tables 31 and 3-2) . The visual census sites had greater water depths and tree heights but less cover than the enclosure net sites. This difference was probably due to the broader range of mangrove habitats covered in the visual census. The greatest correlation between habitat variables was found for submerged aquatic vegetation and salinity, with greater volumes occurring where salinity means were higher (Table 32) . Salinity means were inversely correlated with salinity variation, indicating that where salinity mean was higher, salinity variation was less. Underlying factors among habitat variables . Results of the principal components analysis and correlations of those components with each habitat variable differed for the enclosure nets (Table 3-3) and for visual the censuses (Table 3-4) . The first 3 principal components explained 86.2% of the variance in the original habitat data associated with enclosure nets, and 79.7% of the variance in the enclosure net data. For both data sets, the first principal component explained 52% of the variation.

PAGE 113

104 (0 u (fl >1 0H w c o •H p (0 p w C -P (1) -p o a (fl c o -H P (0 rH 0 rJ y o u n c (0 0) c 0) u in o H o c 0) CO tfl c o •rl -p -H > (U TS (0 c p c s I 0) rH (0 0) p .c p H T3 P c 0) E C o rJ •rl > c QJ U (D > O u d) >i P H c •rl rH (fl W H c rH (0 w c rvO o rH 0 00 CM o • o p P 00 ro • • (fl ft rH o a 1 (fl > 1 s e ft 2 8 0) £ Oi +J C T3 -.H -iH h h 2 £5 a> ° 5 -o CM fO O ro vO CM ro o oo ro in CM VO CM CM rH oo in O in cm oo ^ ro cm in rH in 00 ro oo <« oo ro O O o o> o cm in oo in d d i 0) > c I P Q) X! m c o H P (fl rH p. 3 O 00 CM CM O VO CM rH o in in o • • • • rH O O O I I o cm oo in cm o * H vo rO rH in ro in rH O O o vo r~ cm ^ in O O ^ H H PI o «o o * * rH O O O O I O vO ro ro CM in ro O H ro VO 00 CM CM O vO rH vo «3" ro O O O O O O I or-inmcMCMr-cM OVOrOOOrHCMrHrO OrHvOCMrO'S'^tro o o o o o o r. p ft 0> V u p Ifl p 0 Q) x; Q) 0* C 01 -•H Q) u u c o •rl p a) -p 0) o> c oi id > 0) a o a) >1 0> +J H -H dl c e -ri XJ rH 3 IS W CO c o •rl p (fl -rl u (fl > p -rl c •-I rH (fl

PAGE 114

105 rH u -H Ul Ul >i c X3 0 0. •H •P Q) (0 Xh -P -P CO D 1 Ul C 0 03 6 c CD >H VO >H M 0 H o 01 T3 C -P (0 Ul -P C •H 0 -H -P TS (0 0) •H P > rG •H T3 U O T3 Ul M Ul (0 (0 T5 C Ul (0 01 •P rH Ul £1 (0 H t/T M c rd (0 > a rH rd -p • c CN g m c 0 rH H > (0 c Eh 0) >1 p -H C -H w >1 P -H C -H rH W P. > T3 0) Cn P. 9 6 X) W a o u fa c o -H +J rd p a) cn a) > -p o 0 u p a) * 0) CP g Vl -rl O Eh 0) X. 0) £ o> -P c -a •H -H u z fa P. X! 01 +J _ P ft E cd •0 V u CO X> c c rd cd 01 P £ CO O CN O ro O O • • rH O I o cn o H vo o cn o o o i cn a) X) rd •H !H rd > C 0) 0) P 0 XI c 0 •rH P rd rH 0) Cn U § XI 3 w c rd Q) 6 >< P •H C •H rH rd W a o •H p rd H u rd > >i P •H C H rH rd W

PAGE 115

106 CO in o o in o o O rH o o o o o rHiHiniHTttMcna* CMOOCNrHOCMCM OOOOOCOOO oooooooo CO o o rH O H rH co O rH vDcoOrHcOfOvOco OOrHrHOrHOin oo'o'ododo tN O O co o o CM Ol Ol • • • O Ol f» 00 CO [»* in co CM rH 00 CM CTi CTi r» oo VO rH CO VO CO CM O O rH CO oooooooo I i VO O O CO O rH CO CM O o O rH CO r00 vO VO *t O CM CN 00 o o o o cm rai rH N * (•) r/ rH CM rH CM O O O a\ o o VO rH rH CO VO 00 o o CM O O ri o ai H CT> H rH CO VO rH 00 0 CM c/> r~ o co co H O O rH OTV oooooooo II II CO CO rf CO CO a\ o CM CO O C* o o co ^ vo ro oo oo vo ra\ oo en rH CO O ^" rH rH o o o o o o I -p c a) G o Ck e o o id a -H U c •rH u vO O O 00 CO OK in oo cm i-t 0"> CN rH |o 3 •rH lH rd > rd c •rH rH rH O VO 00 rH VO in oo in to en rrH rH rH rH CO in 00 oooooooo I i i i i i I <7\ CO CM o o •rH rH o in vO rH Ov rji VO VO iH CO CO o rH VO rH M> 0 CTi o ^l" rH rH CN rH « tN CN Q) o o o o o O O in in x; 1 p Tj c i T) MH 0 c Q) 0 •»H G -rH C p O Ifl H rd •rH rH rd ~u p P a rH Q) rd a Di c •rH a) <#> X w s: Q) rd P. 0) rd Q) m 0 p TJ x; 0) e > 3 0 > OJ •rH a "rH cn P TJ rH c -rH O p , IT) rd P C rd TJS ai 0 en p P > -rH rd rd rH g m u p •rH a) 0 J3 rH rH 3 > 0 rd u in u 3 rd rd w U o &< W w w

PAGE 116

03 c o rH -H (0 -P O (0 •H +J M 0) >i x: o) a d n CD c X! 0) -P O <4H O w -H 03 >i H (0 B (0 (0 (fl H > VO rn Q) X! 33 03 P C 0) X! C -P O -H a-* O TJ O fl 0) rH X! ro H id p c a) i c o rH -H > c 0) c (0 c (0 XI p rH a) 4J (0 (13 & 03 C o •rH -P (0 rH 0) rH U o u -P c d) c o § 0 o rH fl ft •H O C •H CO CO o o o o o o in CD o • • • o o o o rH rH o o ro o o rH CD t • • • o rH CT. cr. rH o o fl o o r• • • o m cr. ro o o CO o o in • • • O in CM ON CN O o m o o in ON • • • o CO CD r-M 8 X! CT. O o 4-1 CT. o o r* o rSS • • -P o o •H rH r3 CD Q) rH X! ro o o (0 cr> o o -H ro ^ • • • (fl rH rCT. > rH rH (fl C -H CP rH o o •H 00 o o U rH ro ro 0 . CM CN T3 CH c 0) 0 H C (fl •H ^-^ rH (fl IH ft rH — X ft d) 0 X 01 OJ 0) 0 9 U > •H rH c •H 0 P (fl (fl -P c (fl > H (fl rH c M rH •H £ (fl u •H > 0 w CJ O CN CN iH CD CO 00 CN O o o CO o o O o o rH rH o O o o 1 o o 1 o o oo ro r~ r« CD o o O CO r« o o o o o rH) o ro o o 1 1 o o Q o o 00 ro rH 00 o <* ro CT> CN CD rH CN en in rH rH o CN CN rH o o o o o o o o 1 1 CTv CN H ao rH ro t> •>* rCN cr. vO CN 00 rH O CN rH rH rH o o o o o o o 1 1 rr~ o rH CO in o a\ CO in rH CD ro O Cr. o o o i-H o 1 o 1 o o i o 1 o o i r-CD o CD o oo in vo CN P> CD CD CD rH cr. o rH rH rH rH o o o O O o O 00Ht>O*i , vO*l , CM vOvOOOvOvOOrovo OiHrHiHOOOOl oooooooo co n co o in vo in en «* ro CN O 00 o o o o o c 0 •H +J (fl +J 01 Cn vo o CN CN o o CM o> o O o 1 o 1 V ft 0) o u 03 -p (fl 4-) 0 H aj C" c •H u -p JS Cn •H 0) JS CI) 01 u Cn 4J •H C •H u OJ E £ 3 (fl CO CO c o •H (fl H iH (fl > >. 4J •H c •H rH (fl w

PAGE 117

108 Using the correlation between the original habitat variables and each component, each component was interpreted in ecological terms. Correlations greater than 0.60 were considered in the interpretation (Afifi & Clark 1984) . For the enclosure net data, the first component described gradients of water depth and percent canopy cover. The second component associated with the net data described the salinity regime, in particular, salinity variation. As salinity variation increased, all other habitat variables decreased, especially submerged aquatic vegetation and tree cover. The third component was associated with fringe width. In combination, these components, describe a gradient of mangrove habitat development, together with salinity regime. For the visual census data, the first component was most strongly correlated with salinity regime, water depth and submerged aquatic vegetation. Tree cover was the only strongly correlated variable associated with the second component, and tree height, with the third. Together with salinity regime, these components combined describe a gradient of total habitat development including both mangroves and submerged aquatic vegetation. Relationships Between Habitat Factors and Fish Densities The first 6 factors were included as independent variables in the multiple regression analysis. Densities of fish in the three fish groups were dependent variables (Table 3-5) . No significant relationships were found for

PAGE 118

Tj Q) C -H COrn a x w lite H o ti id n) o H > 3 tr P W (0 CQ C > g •H w 0) 0 T3 p rH •0 C c GO H 01 • -H o in fi P U I o c 0) n SI 0) ft a) U) 1 Ph X! ft o o 0) o c 0) u d) 14-1 0) >H _ in cd o i> r> t> CM CM (N H IT) CM CN cs • • • • • • ^. HO a, rH rH O O CM CD CD CO CM 0> fO H ^ 00 O _ 0> CO CM VO CO Tift 0'*HO'*'J oooddd co. OJ u 3 O co rin r> r~ in cm in vo cm cm o o o o o o o o o o o o iii i rH cm co * in io uooooo Oh eu Hi h, cu oj oi o CM O K ro O ft t> rH ^ O VO VO co CO vo cn ro co in in in O co co co cm ro OOhOKIH ft O H ro r» cm in • ••••• o o o o o o cm h o tcm o in h to oj a> _5 H W o ,« J c Q) o c 0) M OJ MH 0) M Oh ft CO OJ O 1H 3 0 CO w co ft o to C 0) id (U o u 1 10 rH OlCH HUB) X Q H g C 0) •rl +J V4 td M-l S CD ft TJ •H ^ CO fO C* rH CM H cm rro cm vo ro K — VO O CM O CM VO Oh rH rH rH O rH rH O O O -H O O o o o t o o o o o o o o o o o o o o MN 01 O H in o\ ro in t-» vo cn O O O rH O O • • • o • • o o o • o o o rH cm ro >* in vo 0 O O O O O 01 0i Oh Oh 0) Oh 00 CO CM O o o ft o o o o O ro ^ in ro CN vo r~ to OJ OJ « K

PAGE 119

benthic forage fish, however (Table 3-5) . As a group, these small fishes (less than 15 cm) did not appear to select habitats based on the parameters measured in this study. A significant regression was derived for density of water column forage fish, but only the first principal component was a significant source of variation (Table 3-5) . Thus, the greater mangrove canopy coverage and water depth, the greater the abundance of water column forage fish. In contrast to the other two fish groups, for the density of large roving fish, all factors were significant sources of variation except prop root density (Table 3-5) . Thus, sites with greater development of mangroves and submerged aguatic vegetation (SAV) , and with higher, less variable salinity had greater densities of large roving fish. Of the 77 total species collected, 14 had significant (p<.0001) regressions on the 6 factors (Table 3-6). Nine of these species were benthic forage fish and 3 were water column forage fish. The greatest amount of variation explained was 38.8%. This value was derived for densities of Opsanus beta, the Gulf toadfish, which was most abundant at high salinity sites with greater development of both mangrove and SAV. Poecilia latipinna (sailfin molly) was found most abundantly where mangrove prop roots were more dense and the width of the fringe was greater. One of the most abundant fishes, Floridichthys carpio (gold-spotted killif ish) , was more prevalent in shallow sites with more

PAGE 120

Table 3-6. Summary of multiple regression results with 6 factors derived from the principal components analysis. Only species for which p-values were < 0.0001 are presented in the table. (Abbreviations as per Table 3-5) Family/ species Abundance Adjusted R2 Preference (p<.0001) P(R2) Engraulidae (anchovies) Anrhna mitrhplli /ntittt) 18 605 0.174 I o. / Dallaunululuae ^loaoiisnj Opsanus beta (nets) 529 0.388 deep water/ 9.0 dense canopy high salinity 14.9 abundant SAV 8.6 tall mangroves 4.5 Cyprinodontidae (killifish) Floridichthys carpio (nets) 13,018 0.278 shallow water/ 15.6 open canopy short mangroves 7.8 Lucania pan/a (nets) 10,237 0.350 deep water/ 10.6 dense canopy high salinity 4.9 narrow fringe 8.3 abundant SAV 5.5 tall mangroves 5.2 Poecilidae (livebearers) Poecilia latipinna (nets) 1 1 ,000 0.156 wide fringe 4.9 o.o Gambusia sp. (nets) 1,907 0.171 high salinity 9.9 Atherinidae (silversides) Atherinomorus stipes (nets) 11,042 0.211 high salinity 7.7 abundant SAV 9.2 Menidia sp.(nets) 4,348 0.159 low salinity 12.9 Lutjanidae (snappers) Lutjanus griseus (nets) 18,461 0.161 high salinity/ 5.3 abundant SAV tall mangroves 1.0 wide fringe 4.6 deep water 4.4 Gerreidae (mojarra) Eugenes plumieh (nets) 704 0.171 low salinity 14.8 Haemulidae (grunts) Haemulon sciurus (nets) 2,768 0.103 high salinity/ 1.6 abundant SAV tall mangroves 3.7 deep water 4.8 Gobiidae (gobies) Gobiosoma robustum (nets) 534 0.137 low salinity 10.6 Lophogobius cyphnoides (nets) 200 0.321 low salinity 25.7 Microgobius gulosus (nets) 2,997 0.335 low salinity 19.9 narrow fringe 11.3

PAGE 121

112 open mangrove tree canopy cover. Anchoa mitchelli was also more abundant in open locations that were low in volume of SAV. For several species, salinity appeared to over-ride all other variables in importance. Salinity accounted for over 10% of the variation in Gambusia sp. densities; this species appeared to prefer higher salinity sites in the study area. Salinity also explained from 10 to 25% of the variation in densities of Menidia spp. , Eugerres plumieri, Gobiosoma robustum, Lophogobius cyprinoides , and Microgobius gulosus; these species were more abundant where salinities were lower. Of all the variables, salinity was a significant source of variation for 11 of the 14 species (Table 3-7) . None of the other variables came close to this level of apparent importance. Seven benthic forage fish species that were abundantly collected were not significantly correlated with the factors. Among these were the 2 Fundulus species and Cyprinodon variegatus as well as the 3 most abundant mo j arras: Eucinostomus gula, Eucinostomus harengulus , and Gerres cinereus. These species are probably very flexible in habitat selection. Among the species of large roving fish in the study, only the densities of gray snappers and blue-striped grunts were significantly correlated with the variables measured.

PAGE 122

113 0) rH a> fd rH H rH ftf (0 (It •H > M V4 (It viz (0 c 4-4 > Q) > rH •H Li (0 Cn U c •H 0) •H p M O o dt 0) UJ X! a) frl vl/ a) o c P •H 0 -H 0 c -P a) -H (0 a P 05 r-l •H <1) rH X (0 c p 3 (0 •H S _C .5 -g fa 3 >1 P H C .-I (0 W ft "J 8 S fa p U 43 0) P P ft 1 P. O Oi P rrj CO O o in in cn r» o o + o o I o o + o + + o o o o + o o o + o o + CD o> « u 0 IM c 0> u E c 0 3 -H rH > 0 0 u 0 M H 43 P. c p u (1) <« ns CO 2= IH in oo CN CO O I O I o + o + o H & 0 co m b r" 43 U s H TJ H o IB 5 o 3 m co cn ro in in CN CN in co cn ro in ro r> i-i oo m in o m cn t> in + o o o o o o o + o CO CN rH H H h + + + + I O I + 01 H Ol CN ^ CN 0+ 0+ II II +0 +I++ 2 £ £ V CN CN 1+ oo oo oo +o oooo oo +o oo oo oo oooo o o o o o o o o o o o o o + o o o o co 0 CO TJ H§ s •H CO U -P P. 0 •H Q> CO a, *H 0. •H 3 5S 3 H E XI 0 Ol P 3 O in CN VO ai CN >J3 in 0 in n u vo 0) * 4q P 0 CD 4J f -g to 3 E 44 Q) 3 10 3 4: CO OJ •w O CD ft CO Oi c o co to CO CO p o O 3 C 14 C rH CD CD ft £ N (*> # <(P

PAGE 123

114 Densities of other abundant species such as Arius felis, Mugil cephalus, Sphyraena barracuda and Strongylura notata were not correlated with any of the measured variables. For the regressions with individual groups of fish and species, the range of percent of variance explained by the combination of variables was low overall (10.3 to 38.8%). However, the regressions were significant for 18% of all the species, and approximately one-half of the species that were collected in great abundances (i.e. over 100 individuals) . Discussion Sites with a combination of lower mean salinity and high salinity variation had lower levels of all the other habitat development variables indicating reduced habitat development at such locations. This finding is consistent with previous results for benthic community development in northeastern Florida Bay (Montague et al. 1989). An example of a well-developed mangrove shoreline is illustrated in Figure 3-1. In variable salinity conditions, such mangrove habitats are less likely to occur. Among the species preferring sites with greater mangrove habitat development are the snappers, grunts, toadfish, rainwater killifish, and sailfin mollies. Other species may utilize mangrove habitats on a less discriminating basis and tend to occupy all mangrove habitats. Such species include snook, barracuda and sheepshead. This assemblage of fishes is enhanced when mangrove shorelines occur and especially where

PAGE 124

u a) p id to X3 +J M O C c •H -P id -p •H X3 03 £ > O n Cn C 03 e a> a o rH > a) T3 i 0) O c o H -p 03 M -P to 3 I 0) P CP >i 03 PQ 03 •a •H P. O

PAGE 125

116

PAGE 126

117 such habitats are well-developed. Activities which directly destroy mangrove shorelines or degrade the guality of such habitats have negative consequences for these important species of sport fish. Although no significant relationship was found among the habitat variables and benthic forage fish as a whole, certain species did appear to discriminate on the basis of the measured variables. Lucania parva, Opsanus beta, and Poecilia latipinna, three abundant species collected, appeared to select sites with greater mangrove development. In contrast, Floridichthys carpio chose more open sites. This killifish seemed to prefer shallower locations, a trend also noted for individuals living in seagrass habitats (Sogard et al. 1987). In addition, Opsanus beta and Lucania parva, chose sites with more abundant SAV. Similarly, Sogard et al. (1987) collected greater abundances of these species in seagrass bed sites with greater vegetation densities. In the current study, densities of water column forage fish as a group were greatest where water depths were greater and mangrove canopy more completely blocked the daylight from reaching the submerged habitat. These species use schooling as a possible defense mechanism against predators. Thus, they may use greater volumes of water to increase school size and prefer sites shaded by mangrove tree canopy for additional cover.

PAGE 127

118 While these factors were important for the water column forage fish group as a whole, other features of the habitat, seemed to segregate the individual species. The most abundant species in this group, Anchoa mitchelli , was less abundant at sites with greater volumes of SAV. This species was extremely dense at only one location in the study area (Little Blackwater Sound) . Although not a variable included in this analysis, turbidity may have been the more attractive habitat feature for the bay anchovy at this site. Of all the significant regressions for the abundant species, the range of percentage variation in fish density explained by habitat variables in this study (10.3-38.8%) was only slightly lower than that found by Sogard et al. (1987) in seagrass beds (24.8-42.7%). Thus, fish may select particular habitats based on salinity and physical features in Florida Bay but other factors (e.g. foraging reguirements, species interactions) are also undoubtedly important. The present findings do not differ greatly from longterm observations of fishes on small patches of coral, in which habitat attributes other than overall size were of little value in predicting the structure of fish assemblages (Sale & Douglas 1984). Many species of reef fish may therefore select habitats based on overall parameters (i.e. large coral reefs vs. very small patches) and not detailed features. The low magnitude of variance explained by the habitat variables measured in seagrass beds and mangroves

PAGE 128

119 may be due to the selection of habitats based on general features (i.e. mangrove over seagrass habitats) . Although density of large roving fishes as a group were correlated with the physical features, some species that were categorized as "roving" in the current study may not actually roam among the mangrove habitats. Individuals of these species appear to persist at particular locations for long periods of time (based on limited observations of tagged fish) . Thus, while some large fishes (such as mullet, catfish, barracuda and needlefish) may truly be wanderers and display no discrimination among mangrove shoreline habitats, other species (gray snapper and bluestriped grunts) may maintain more permanent residency at certain locations and display definite habitat preferences.

PAGE 129

CHAPTER 4 FOOD HABITS OF MANGROVE FISHES: A COMPARISON ACROSS SALINITY GRADIENTS Diets have been studied for fishes associated with whole estuaries (Darnell 1961), mangroves (Odum 1971, Beumer 1978, Salini et al. 1990), salt marshes (Harrington & Harrington 1961, Rozas & LaSalle 1990) , seagrass beds (Livingston 1982) , lakes (Werner & Hall 1983) , and streams (McNeely 1987) . One successful dietary strategy identified in extremely variable habitats is omnivory, or the consumption of a wide variety of prey organisms including both plant and animal material (Darnell 1961, Harrington & Harrington 1961) . Opportunism, or an ability to exploit alternative foods depending on availability, is also a valuable dietary strategy for survival in variable environments (Odum & Heald 1972, Livingston 1982, Salini et al. 1990) . Trophic systems in aguatic habitats are often characterized by shared common resources among the various species of fishes (Harrington & Harrington 1961, Livingston 1982) . In estuaries, seasonality of freshwater inflow increases habitat variability upstream relative to more 120

PAGE 130

121 stable downstream locations (Rogers et al. 1984, Moyle & Cech 1988) . In northeastern Florida Bay, biomass of submerged aquatic vegetation is much lower and highly variable upstream than down (Montague et al. 1989) . Abundances of polychaetes, crustaceans and other benthic invertebrates are highly correlated with plant biomass in this area: 80% of the epifauna live among the blades of seagrass and algae. For estuarine fishes, these animals are among the more heavily exploited food items (Darnell 1961, Livingston 1982) . In light of these conditions, fish diets are likely to display patterns along the gradient of salinity variation in the northeastern Florida Bay study area, due to fish foraging habits and variation in prey base. Seasonal variations in diets may also be expected. Toward the overall goal of identifying the influence of variation in freshwater inflow on habitat use, the objective of this chapter is to identify dietary components and make comparisons among the more ubiquitous and abundant species. Breadth and variability of diets should reflect environmental conditions in the more variable versus stable habitats, and over the seasons. Materials and Methods Fish were collected in mangrove shorelines located up-, midand downstream in two systems (east and west) in northeastern Florida Bay (Chapter 2) . Species were selected for the analysis of food habits because they ranged across

PAGE 131

122 the entire area and occurred consistently over the study period. For Lutjanus griseus, Sphyraena barracuda and Fundulus grandis , all samples collected were analyzed. Due to great abundances, for Floridichthys carpio, Strongylura notata and Eucinostomus harengulus , smallest individuals (< 3 cm) were eliminated and subsamples were selected from the remainder. Fish were not divided into size or age classes for this analysis, however. Gut analysis of the 6 selected species was contracted out to Mote Marine Laboratory. For the laboratory analysis, 36 taxonomic levels (e.g. family) were selected for consistency with other estuarine studies of fish food habits (Brook 1977, Beumer 1978, Livingston 1982). The analysis chosen follows the "points method" of Hynes (1950) . For each fish (n= 1,222), total length was recorded and the stomach was extracted. In order to identify variability among individuals, no stomachs were pooled. The material found in each stomach was distributed to a standardized level within a gridded petri dish (Hellawell & Abel 1971) . Each stomach containing food was considered to be uniformly full (Starck & Schroeder 1971) . Using a dissecting microscope, percent composition of each food category for each specimen was calculated by estimating the area covered by the material on the grid (Neilson & Johnson 1983) . Analysis To determine frequency of occurrence, the number of fish in which each food item occurred was listed as a

PAGE 132

123 percent of the total number of fish examined (Hynes 1950) . In addition, the mean percent composition for each item was calculated by obtaining an average value for all specimens of a species. The results of both analyses are presented to give a complete picture of the relative dietary importance of the items consumed (Hyslop 1980) . Two sets of multivariate analyses of variance (MANOVA) were performed for each species with the major items found in the gut as dependent variables. All data were transformed using an arcsine sguare-root function prior to these calculations (Kleinbaum & Kupper 1978) . The first set of MANOVAs addressed spatial variation. Gradient (upstream, midstream, downstream) , system (east, west) and interaction of gradient and system, were used as independent variables. The second MANOVA looked at temporal variation. Season was the independent variable. Only major food items, defined as those which occurred in at least 20% of the specimens or exceeded an average of 4% in composition, were included in the MANOVAs. Results Shared Resources Of the 24 items (counting fish as one and excluding unrecognizable) , 8 occurred in all 6 species (Table 4-1) . Of these, amphipods were the most ubiquitously consumed, present in at least 5% of the specimens of all species.

PAGE 133

124 9 £ o> E co o o O 5 * c\j*-KO)KC\|C})K Cm It) o CO >— >— »— no > COOCJ--OOOOC0 oo dri t^o 6 6 d s K Co CO c\i w cb N 00 CM co co o CM yO d d 5?° CM O co d CO O CM W) c\j K >» K CM •* 6 s Cm cm CM * I Is I 21 II §5 « a. !s I E -5 c i? «» a. E o O I a. & it" II .5 E as ^ # Q)OQ)*— KtJjCoO) (oddtidddK COOOCMCMOOCO dddddddci HOO'-ttO*'Kddco'*^d»^co did 6 r » d 6 » «oot~ci*-cd CM CM sootgsooc cdddcsicdddcM <0C00»Olf)OC0»Kd oj d K d 1 >»• CM COOCMprvOCMCO GDOT-OT-O'-d CMO^»CMOCOOO « !) » t, O tl !; «qoiOr-owtM (AdddcModd o«o«o co d ^ d d d »» O O O o o yd d d d d d d n d d d o o o o o d d d d d o co cs 00 co if) dKdK* d cm d to os q co o co d d d d d to cm co to o d a> d co d q co p in o d d d d d co cm co to o d cm' d oo d cm CM O 0) o d d d ^ d 1 Q. i _ Q. f| o (0 as if o o if > c c to CO — cd cd co -a a £ ? co co iSiilS i) o t» d-r-dcMcjodddcb OOOCOIQIOOOOCm ddo-H)oiddd.£>,»« O'c >LUCQO(!)mO 3 O CO o odd o o o odd o o o odd o o o odd p p co co o o d d cm' d d *rO odd cm cm o co cd d r-. o o o CO o co to co at a' d CO If) o odd cm cm o cm' cm d i-CO o odd CD CC V I « o * CD D. . CO > 2 CD 3 §to o> o *. d co o d d co o K d *o d d o o 55° ro cm d (o o 3°' IO o CM d ^ CO Co Co K K o >f> c^ ifi CO O) TO I--' d O) cm co o o d co o n »o cr> d Co Q) CO O I*. d CM >f to If) Co K S ». CM CO 2 CO O) o CO 2 K CO CM 2 CM CO O co o co d «i a » ifi 5 oj CM "Cf b 8 I C 2: CD 3 = CO CD ; < £ t= coo > to II i <<» 51 CO 1 O 0) 2^ CD N IS Q. c E > 5 1,5 o o M » w a CO iCO E „ o "i? 8 £ f c c 3 1 I i cS & 03 E CO £ S J= CO CO o !^ ic % o co CO CO CD o o I — 1 — frt

PAGE 134

125 Other items exploited to varying degrees by all 6 species included isopods, shrimp, nematodes, eggs, fish parts, adult terrestrial insects and algae. Breadth of Diets Excluding unidentifiable material and counting all prey fish as a single food item, of the 6 species, Floridichthys carpio and Eucinostomus harengulus foraged on the widest variety of items (20 and 18 out of 24 total, respectively) (Table 4-1) . Only 2 items, however, were consumed in mean guantities exceeding 4% of total diet: amphipods and algae. Thus, although a wide variety of items were utilized by these species, a degree of specialization was apparent. In contrast, Fundulus grandis, Lutjanus griseus , and Strongylura notata, not only consumed a wide range of items overall, but several items (5,6 & 4 respectively) were consumed in mean guantities exceeding 4%. Isopods and fish were major items commonly found in all 3 of these omnivorous species. Sphyraena barracuda was the most specialized of the species analyzed, consuming mostly fish. For this species, the benthic forage fishes in the family Cyprinodontidae were consumed with particular freguency, even though the most abundant fishes in the study area were Engraulids and Atherinids (Chapter 2) . Thus, in the study area, among the 6 species, 3 feeding strategies were evident. In one strategy, used by Floridichthys carpio and Eucinostomus harengulus, many items

PAGE 135

126 were consumed, but only 2 in great abundance. In a second strategy, used by Fundulus grandis, Lutjanus griseus, Strongylura notata, many items were also consumed, with several (4 to 6) items ingested in substantial abundances. In the third strategy, used by Sphyraena barracuda, fewer items were consumed overall, with only one in significant quantity. Diet Variability The MANOVA results for all species, as presented in Table 4-2, indicate the degree to which location (i.e. gradient, system) was a significant source of variation in fish diets. For Floridichthys carpio and Eucinostomus harengulus , gradient position was a significant source of variation. In both species, more animal food (i.e. copepods, ostracods, nematodes) was obtained downstream than upand midstream, while more algae was consumed upstream. The only other influence of gradient occurred for Lutjanus griseus, which consumed significantly more crabs upstream than midor downstream. For Strongylura notata, Fundulus grandis and Sphyraena barracuda, diets differed among the locations, but not systematically along the salinity variation gradient. Seasonal influences on the fish diets are indicated by the second MANOVA results presented in Table 4-3. For Floridichthys carpio and Eucinostomus harengulus, algae was more important in the spring. For Lutjanus griseus, crabs, an important dietary item, were found abundantly in diets

PAGE 136

127 I P co M >i o CO
in (0 +j o -p TJ 0) CO 0) u O 4H Q) CO •H C 4-1 O C -P o c o H p -H CO o o CO CO in a i 0) CO 3 c CP(fl «J 0) 45 (0 (0 o 1 a) Eh Q CO > 2 5 II S3 5 1 So "5 TJ il © E s » 2=* • -2 I i li tj -a -a G 1 1 »C flfl >8 3 3 3 -2 AAA C C C I g I TJ TJ TJ = 1 A C c A la a 3 ft-. c~ rr~ co co co co co co m io in in m m o o o o o o 6 d 6 o o o IAOOCO' OB © © : § a « i a. ic © O. CO aS E 5 <0 Qco E oS © g a> E a co o o c £ co OHIO O O O E I >. « c _ c I E E 5 B 'T3 id b j O) CO CD I 11 2 >, 2 CO CO CD I 3 c c ,3 CO 23333 to to to to to to 2 2 2 2 2 2 coojeococo §§§§§ ttfcttt 555555 KKKKS 55 55 S 35 » 55 m 55 5) 55 S 555555555555 m m m m m SO) O) O CO CO C\J t— CD CO Nm 00 q in to»-cor»(o onosri-ouiton* co 00 co in *o o o co to r; tco to to oooi^oo r-qqtjT-io co « q 00 00000 000000 600000 00000 tt1o O CO IO 1O CM CM i-i-MOrO i-i-BC2 3 -o CO rl !| i a> c t I 8 CO CO 088 O O CM OO-666 TJ 5 9 « a. o. _ E o » to .£ * CO CM CM MOO) 3 * CM rt r-. 1c CD to ® o = c c O CD CD CO CO CD £ -a c a s c .9-S Q. 3 .9£ Q. CD C O. g C E 2 »2 E E J» E g S-g a> « o.S he « co .2 c» co -E ;== a) «Uie a -no) 0 f~ in 9J» t-~ 1o S» cm m co »o 8 co m r» cm o> m co CO CM O odd So tO CO © (0 TCM •» CM c-inio ^6^ > to > rO t-' o rt o ti— CM E CD i »rl — CD E CD tj £ S , 2 CO CO CO f E CD f E 11 « 3 » C 3 •a > 1 h I! is CO -Q s 0 •o 0) Q) M-) O m
PAGE 137

128 3 : & Ih C fo Con summ •rH 01 01 O c i-t rd rd me c rd on St ii CP 01 id c rd -H oj p CM N. CO us 01 p con /alue 0 26 0 05 ? id p •H c li. u p 01 • 4J 0 •H 01 ft iM c 01 0 fd o -P -H ft 01 c 4-1 rd a) OJ H r\ 5 s OJ p U u 05 on p pe OJ pern ariar. u p rH c§> "845 £ 01 ft *! P 0J •H f 1 O p • P 2a. § 8 CP (0 c rH c U o 0 3 0 O p •H B 5 S P -P o n 01 ed ia ke 3 6 01 4-1 01 rd p 0 e 1 CL •H > Q c 01 0 rd 0 OJ p •H T3 m • TS •H $ T3 OJ £« T— 0 0) 01 fl a E CD 0 0 O 4-1 3 8 e 3 o 01 TO o OJ 4-1 •H u ar P if of • 0 c co 73 0 0) u •H H CO 01 0) fd O) ft o CD OJ u (0 5. CO i rH (0 rd X! dich io A p 3 01 O rd rd fd rd Eh Q 01 OJ II E i c iSo So CO o b CD « p e I 5S 1 a 1c a> cl a E 3 CN CO o o 6 co S 5 co 1 S s 5 8 Uj j; 1 I c 5 1 Cl sills §§§§§§ mil SCO CN C\| C\J CO CO CO CO ^hr« co co oo io co lo o> o> co o 00 i00 CM * CO >O CO CO N A ta a O v O CO CO a> rr P o o o o o o o o b 6 o onctf N i-sgocn V'-COOM OJCOrtr-TO IO O CM £ CO CO »CM oq--o» cs^>-»io oooooo ooooo O tCO IO CN CD CO CO CO f O) O N T CD O LO LO tf rCD Scores co lo cm ^co 3»#»-r.t-c» noaas *— ' o o co *— OHrdbo tcj ^ cd b Or^dd : i ; a, CB * cc S 9 "8 a. Cl-Q E S CO OJ CM 6 -o Q. .2 -O «cg 9 .S ?. tl 5 .9--S D. • C .Cl » -c o P 5i x: co Q. a ce o _ Cl O c m s O) CL ra >» CD O i CD cr <» CO O o o C\J 6 1 ii 3 co C 3 Ii if il v a) M u o 0 0) ai o c e a) 0 u TJ 0) 0) 4-1 0) 4-1 4-i -d o c rd CO u Q) *H 0) 4-1 V< -H 0> C ai cn •O -H 0) T3 0) -p o •H •a c •H a 4-1 -H o p an n 0) IT) 0) h ^ +J C a) o V o H 0 OJ c O 4-1 X M * « *

PAGE 138

129 all year except in winter, when they were seldom consumed. For both Fundulus grandis and Strongylura notata, adult insects, a prevalent item, were greater in diets in the spring. Discussion Shared Resources As commonly occurs in aguatic habitats, diets overlapped among the 6 species as resources were often shared (Harrington & Harrington 1961, Livingston 1982, Odum 1983) . This overlap occurred in all areas and throughout the study period. Diet overlap can become particularly evident when one resource attains a periodic peak of abundance. In other aguatic habitats, for example, populations of penaeid shrimp (Salini et al. 1990) or larval insects (Harrington & Harrington 1961) increase under certain conditions, and opportunistic fish take advantage of the resource abundance. A similar seasonal increase in exploitation of a particular food resource was found in the current study for adult insects exploited by Fundulus grandis and Strongylura notata . In western Florida Bay, gray snapper and spotted sea trout (Cynoscion nebulosus) diets overlapped when peak abundances of pink shrimp (Penaeus duorarum) occurred in November (Hettler 1989). Although a similar pattern might have been expected in the current study, none became evident. Migrating juvenile pink shrimp, while a major

PAGE 139

130 portion of the epibenthic fauna in western Florida Bay, have lower densities in the interior and eastern Bay (Holmquist et al. 1989a) . Diet Breadth and Variability Three types of feeding strategies were identified among the 6 species based on diet breadth and degree of opportunism. Firstly, Floridichthys carpio and Eucinostomus harengulus appeared to rely strongly on the plasticity of their diets, feeding on one basic resource consistently, but also consuming smaller quantities of many other resources. Although it was not separated from other materials in the unrecognizable category, these species probably consumed detritus. A product of breakdown of dead plants, this material contains a "coating" of bacteria and fungi of nutritional value (Heald et al. 1974). Primary consumers of detritus include amphipods, shrimp, crabs, and certain fishes. Among the 6 species analyzed in this study, a major portion of the diet of Floridichthys carpio was probably composed of detritus; in the North River, the diet of this killifish was 21% detrital material (Odum 1971) . In that study, Eucinostomus harengulus was considered a secondary consumer of this material, with only 6% of the gut contents directly composed of detritus (Odum 1971) . The second major strategy was employed by Fundulus grandis, Lutjanus griseus and Strongylura notata. Rather than switching among a very wide variety of items, they consumed about 5 items consistently and abundantly. Thus,

PAGE 140

131 these fishes tended to rely more strongly on omnivory, or constant foraging for these several items, rather than opportunism. The diets of Fundulus grandis probably also included detritus (Rozas & LaSalle 1990) . However, the 3 other species analyzed in this study were apparently not direct consumers of detritus (deSylva 1963, Odum 1971, Starck & Schroeder 1971, Brook 1977, Thayer et al. 1987a). The third major strategy was exemplified by the piscivore, Sphyraena barracuda. One resource, fish, was consistently and effectively targeted by the barracuda. These modes of feeding are consistent with other estuarine investigations. Two extreme modes of feeding were identified in Lake Ponchartrain, for example, with detritivore/omnivores, such as mullet, on one end of the spectrum, and piscivore/specialists, such as gar and jacks, on the other end (Darnell 1961) . Similar extremes were observed in a red mangrove/ saltmarsh habitat in east Florida, with killifish and snook at opposite poles (Harrington & Harrington 1961) . In these examples, intermediate strategies incorporate the consumption of small benthic invertebrates in diets with greater and lesser portions of detritus and fish. In the current study, the detritivore was probably most strongly represented by Floridichthys carpio, with the barracuda at the opposite extreme. Intermediate species include the mojarra, Fundulus grandis, Lutjanus griseus and Strongylura notata, in order of increasing piscivory.

PAGE 141

132 Influence of the Salinity Gradient Diets of three species were significantly influenced by location along the salinity gradient. Eucinostomus harengulus and Floridichthys carpio diets included more algae upstream and more copepods, ostracods, and nematodes downstream. This finding tends to be consistent with the reduced levels of benthic invertebrate populations found upstream in a previous investigation in northeastern Florida Bay (Montague et al. 1989). Plant materials, other than seeds, are generally less concentrated energy sources than animal sources (Odum 1983) . Lutjanus griseus consumed more crabs upthan downstream. Although all the parts were not identified to species in the current study, most were probably from mud crabs of the species Rhithropanopeus harrisi. This brackish water crab was abundantly collected in minnow traps placed in mangrove shorelines upstream, but was never collected downstream (unpublished data) . It occurred abundantly in the gray snapper stomachs in the North River (Odum 1971) . Rhithropanopeus harrisi is, thus, an important forage item for snappers that appears to occur abundantly in brackish water in the study area. As mean salinity rose from 16 to over 30 ppt in northeastern Florida Bay, these crabs were virtually eliminated from seagrass beds over a period of four years (Holmguist et al. 1989a) . Crabs in general may not be readily assimilated as food by fishes due to a higher percentage of exoskeleton

PAGE 142

133 (Weisberg & Lotrich 1982) . In an experiment, Fundulus heteroclitus , fed a diet of only fiddler crabs (Uca pugnax) lost weight (Weisberg & Lotrich 1982) . In a salt marsh where fiddler crabs comprised 17 times more food volume than any other item for Fundulus grandis , the killifish may have compensated for the crabs' low caloric value by consuming their prey in very large guantities (Rozas & LaSalle 1990) . Since two items consumed more abundantly upstream, algae and crabs, were lower guality energy sources for fish growth, these locations may be somewhat lower guality habitats than midand downstream in terms of obtainable foods for these species. The fishes that forage upstream may compensate, however, by foraging on greater guantities of the lower guality items that are available.

PAGE 143

CHAPTER 5 PREDATION RATES ON SMALL BENTHIC FISH ACROSS A SALINITY GRADIENT The value of estuaries as nursery areas for fishes that have wider distributions as adults has been confirmed in many studies (e.g. Gunter 1938, Reid 1954, Carter et al. 1973, Blaber & Blaber 1980, Yanez-Arancibia et al. 1980, Bell et al. 1984, Blaber et al. 1985, Blaber et al. 1989, Pinto 1987, Robertson and Duke 1987). An influx of juveniles to an estuary usually coincides with the season of highest freshwater discharge, when salinity levels drop. As juveniles develop in the estuary, they tend to migrate from fresher upstream areas to more saline downstream habitats (Weinstein 1979, Rogers et al. 1984). Based on observations such as these, one of the paradigms of estuarine ecology has developed: that estuarine salinity conditions contribute to the survival of juvenile fish because stenohaline marine predators are precluded from entering portions of the estuary having lower, more variable conditions of salinity (Gunter 1961, Austin 1971, Browder & Moore 1981, Odum et al. 1982) . Despite its widespread acceptance, this hypothesis has not been tested before now. 134

PAGE 144

135 Predation is one of the most complex of the species interactions studied by ecologists. In comparing levels of predation intensity among habitats, not only are relative abundances of predators and prey important factors, but the potential rates of predation must also be estimated (Kitching 1983) . These rates depend on behavior of predator and prey and characteristics of the habitat. Some behavioral and microhabitat aspects have been modeled by observing individual components of predator/prey interactions in the laboratory (e.g. Holling 1966; Barshaw and Able 1990a) . Other habitat related factors, however, must be measured in the field, and are thus, more difficult to determine. The tethering technigue, a useful field method, has increasingly been used by ecologists studying the effects of different habitat features on predator/prey interactions. Briefly, the tethering technigue involves affixing a line to a prey organism so that evidence of predation can be determined from its condition after a period of time. These investigations are usually accompanied by complementary laboratory studies or censuses of predators and prey. A summary of tethering studies is presented in Table 5-1. In this study, the tethering technigue has been used to compare predator encounter rates across a gradient of salinity conditions. The guestion of interest is, are these rates lower in the more variable upstream locations relative to midand downstream where conditions remain more saline?

PAGE 145

136 m > «) 0 u D> C id x d) 00 rH r0> rrD oo rH 00 oo CP cr> id rH 6 rH H ft 0 0 -C i ^ P •P V P •H u •H •H e a> a) rH rH <3 13 i£> CO 0> 03 3 id " C r4 e P H a) u oo oo tn u u id id ca CO t3 id •* 33 Q) as -p 9 Pfl 13 73 C •H B u u 0) c 0 tn rH -r-t ro oo cm o> en id id P P H u Id H 3 O •H •U Q) U 0 id C a, a> >H
PAGE 146

137 This experiment actually integrates several of the steps involved in predation that were described by Holling (1966) . First, measurements indicate whether or not predators occur in the locations under comparison. Secondly, the tests indicate how well the predators can perceive the prey, given the conditions at the site (e.g. turbidity levels) . Thirdly, the ability and propensity to consume particular prey species relative to size and palatability is indicated. To add to the understanding of the influence of freshwater inflow on fish assemblages, the objective of this chapter was to compare predator encounter rates upand downstream during the rainy season. If the paradigm is true, fewer tethered prey should be consumed upstream than downstream, as the marine predators are excluded from these locations due to the lower, more variable salinity conditions. Materials and Methods Preliminary Tests To evaluate the effectiveness of the tethering technigue for specific prey fishes and conditions in the study area, preliminary tests were conducted within an enclosure formed from two 30 m seine nets that excluded all potential predators. For these tests, fish were tethered by sewing one end of a 1.0 m length of 8 lb test monofilament

PAGE 147

138 fishing line through the lower jaw of a small fish (4-10 cm). The other end was looped over a 1.25 cm diameter polyvinylchloride (PVC) pole that had been driven into the substrate. Sixteen fish were tethered inside each enclosure and checked at 3 , 6 and 24 hour intervals. Although all fish survived and remained securely tethered for three hours, in these preliminary tests, one (Eucinostomus gula) died after six hours and several other fishes died after 24 hours (Table 5-2) . In addition, fish that died at the bottom were guickly attacked by scavengers (e.g. crabs and gastropods) , possibly interfering with interpretation of test results. Since the study objective was to determine rates of prey encounters with predators, not scavengers, these results prompted further investigation into alternative tether and stake designs. After several prototypes, an Lshaped stake was made by joining two 1.0 m long PVC pipes with an elbow (Figure 5-1) . A hole was drilled at the free end for attaching the tethered fish (as above) and the other end was driven into the sediment. When deployed, the top bar of the L-shaped stake was above the surface of the water. To ensure that predators and not scavengers were responsible for removing the prey, the tethered fish were forced to remain above the substrate by adjusting the depth to which the pole was driven into the substrate.

PAGE 148

139 OJ u Ifl 0 • • H CD 0 c 0) • Q) > or c 73 0) & W <« c 0 -H x: U] to -p 0) >i H 42 0 0J a s w w "H •H 73 4-1 •H • Is 73 73 0 i XI 73 0) 6 o 0 V to H g b to a R 3 a M 3 0 CN U d> p ai 3 0 ID H CD P 4-1 < tn u 3 0 OJ p 4-1 en c tn (D X •o o> c •r) 0) (0 •rt s T3 n) a> a a) > c •H tn tn £ •a M 0 Q Q) > M 4J Q) to 4 Q) 6 H 3 C * 00 rH 00 01 73 O & O H •P 01 01 CD C 8 0) c a p • d> 01 OJ TS •H 01 c •H TJ C 3 O 01 M OJ 0> c Q) > nS U 01 XI >l CD XI •rt 4-1 TJ 0) OJ rH E 73 3 d) tn d) c C 0 o P c JG -H CP (1) •H XI c u Q) > o C 73 3 d) 0 -H 73 73 >i 73 r1 (fl H X!
PAGE 149

Figure 5-1. Illustration and dimensions of the tethering systems used in this study as deployed near mangrove edges.

PAGE 150

141 Tethering Experiment Small fish were tethered at far/up-, mid/up-, midand downstream locations in two systems (west and east) in northeastern Florida Bay (Figure 5-2) . Tethered species included those that are consumed by dominant members of the predator guild in the study area, as indicated by the results of the food habits portion of this overall study (Chapter 4) . Small fish make up over 25% of the diets of Strongylura notata (redf in needlefish) , Lutjanus griseus (gray snapper) and Sphyraena barracuda (great barracuda) . Trials were conducted on two dates at each of the eight locations during mid-summer 1990. For each trial, salinity, horizontal secchi distance, and water depth were recorded. Fish to be used in each trial were collected by setting out several minnow traps near each site on the day before a test. Small fish (4 to 10 cm total length) from five species were used in the 16 trials: killifish (Floridichthys carpio, Cyprinodon variegatus, Fundulus grandis, and Fundulus confluentus) and crested gobies (Lophogobius cyprinoides) . In each trial, ten to sixteen fish were tethered 10 m apart and about 2.0 m from the mangrove edge (Figure 5-1) . Sites with water depths of about 50 cm were selected for each fish. The total time each prey was tethered ranged from 3 to 3.5 hours. For approximately 1.5 hours, while observers were within 10 to 100 m of the tethered stakes, they recorded when possible the type, size, and number of predators that approached or

PAGE 151

142 Figure 5-2. Locations of study sites in northeastern Florida Bay.

PAGE 152

143 attacked tethered fish. Other potential predators in the vicinity were also noted. At the end of the test period, each stake was examined. Predation was assumed to have occurred if a fish was missing, severely damaged or if a predator was tethered. Analysis For each test, the percent of prey that were subjects of predation was used as the dependent variable for statistical analyses. Analysis of variance (SAS GLM) was used to determine if predator encounter rates differed due to gradient position (i.e. far/up-, mid/up-, midand downstream) or system (east, west) . To make multiple comparisons, specific F-tests were used to contrast means for each pair of gradient positions. SAS GLM with Student Newman-Keuls multiple comparison tests were also used to determine if tethering different species created an extraneous source of variation. In addition, correlations were calculated between the predator encounter rates and corresponding ranges of visibility (horizontal secchi distance) , salinity, and water depth. ANOVA was also used to determine if differences in salinity, visibility and water depth were due to gradient position. Results Of 235 tethered fish used, a mean of 83.5% were subjects of predation after the 3 to 3.5 hour period (Table 5-3) . Predation rates averaged 90% in mid/up-, midand

PAGE 153

144 C •H 0} <1> P (0 i p w >1 CP w c •H O (0 -P e o o w rH (0 •H P rH Q) P C o o c o u Cn C (0 B r> P 3 Cn d o p (0 4-1 o p c o g O U (w 01 c o •H P •H W O 0) o (0 cn cn TS (0 W M o a) 0) u X! p o 4-1 -o 0) p c e o u 4H 73 OJ P O 73 C O O p « s c Ifl i to H O £ C o (d a> in p — I c •H fH (0 w p c ai u 0, p ai oj e EH 3 S3 * p c OJ •H T3 n) o * e 0) p ai >i w "ftNcioiinojooco vovOLnin^vomin r-coooocNrororo oo o oinoocNOtn't oooooooooo ro HHHHMHlOtO OJCNHHOO^"0» CN inincnr-oooo roocninoooo CNCOCN'3'CNOOm H (N H in ^ ^ ^ ro cNr-2co222cn lOOtNinicmioin rro *t rin cn oo cn ro r~ oo vo cn co ro CO invorovovoioin „ rH rH rH rH rH rH rH rHCNrHCNrHCNHCN CN HHMNflnfl^ rHrHCNCNrOrO'*'* CNCNCNCNCNCNCNCN c (0 OJ s p 01 ro M II ai a> u u g +J io tn oj ft ft u 3 3 -P P --^ \ 01 0) )h "O TJ 0J (d -r) S S CN P •H n ii ii ii rH cn ro p •• c oi oj g "H OJ V P n) CD U >i o w * * *

PAGE 154

145 downstream locations contrasted to 55% for those far/upstream locations (Figure 5-3) . Salinity, range of visibility, and water depth also varied across the gradient (Figure 5-3) . The far /upstream predation rate was significantly lower than the rates for the mid/up-, mid-, and downstream locations, but there were no significant differences between east and west systems (Table 5-4) . Rates of predation were not significantly different among the species (df=4, F=1.98, p=0. 1296) . Neither salinity, secchi distance nor water depth were significantly correlated with observed predator encounter rates (Table 5-5) . These factors did vary significantly among the locations, however. Salinity means were not significantly different between the far/upand mid/upstream locations (p<0.3454, df=l) , nor between the midand downstream locations (p<0.3921, df = 1) . However, the subgroup formed by the far /upand mid/upstream stations had a significantly lower mean salinity than the subgroup formed by the midand downstream locations (p<0.0001 for each pair of contrasts). Range of visibility was significantly greater downstream than at the other locations (p<0.005 for each pair of contrasts) . Water depth, however, was significantly greater at the downstream/east location (p<0.05 for each pair of contrasts) .

PAGE 155

146 Predator Encounter Rates Mean and Standard Deviation 100 80 c CD O 60 CD _ 40 20 4-0 Missing Fish L a. cl CD E o 100 80 60 +0 Salinity Visibility Water Depth 1 1 Far Mid Mid Down Up Up Figure 5-3. Mean predation rates for small (<10 cm TL) benthic forage fish tethered for 3 hours across a salinity gradient in northeastern Florida Bay. Error bars indicate the standard deviations of test rates resulting from 4 tests in each of the 4 gradient positions.

PAGE 156

147 ui p m c •P W O >i-P W tp (0 c O -P M C io a) o (0 (0 Cp T-i £ C T3 -P (0 T3 (0 4-t Cp o a) oi H oi o P 01 4-t O Ul •H 01 >1 rH (d c << 01 H 01 0) p o a (o K > a) u tp o •P a -p o 1 rH u o I 3 — W XS 0) 3 H rO > I 01 3 > n CN O n ro rH CN <£> O o ro • • O o O o 10 CN CN o rH CN • • CN J w a o X -p c 0) •H X) cd (H 0 B •p oi >i w -p a 0) •H x> ft) IH a X E -p 0) >. W B o u 01 01 Sh 4H m o m 0) ai u Cn 0) XI M 0 in >H M E 0 T3 0) 0) Sh Mh O W 01 0) U Cn 0) •a 0) Xi 0 £ IH -p 01 0 a ii ai u •P 01 XI 0) >H -P 01 ft 3 X) •H X 0) >H 4J 0) a 3 \ Sh crj h +J G 0) •rH XJ CD Sh (9

PAGE 157

148 Table 5-5. Correlations of predator encounter rates with environmental variables. Data are percentage of fish missing (arcsine transformed) after being tethered for three hours adjacent to mangrove edges in 16 trials. Pearson correlation coefficients and probability values are indicated. Variable Test Results Correlation Value (R) p-value Mean Water -0.1463 0.5885 Depth Salinity 0.3688 0.1597 Secchi 0.1308 0.6293 Distance

PAGE 158

149 Thus, although the predator encounter rate was greater at the two far/upstream locations, salinity, visibility, and water depth were not unusual at these locations relative to the others. Salinity, the parameter of most interest, was lower at both the far/upand mid/uplocations, but the rate of predation was lower far /upstream, and higher at the mid/upstream location. Potential predators that were observed approaching tethered fish at all the locations included redfin needlefish, gray snapper, and barracuda (Table 5-6) . Observed only in the far/upstream locations were bull shark juveniles, gar, alligators, and turtles. A total of 28 actual predation events were recorded in which the attacking predator could be identified (Table 56) . Most of these events (22) were due to needlefish (Strongylura notata) , half of which became tethered themselves. Once they had swallowed the prey, the needlefish were unable to cut the fishing line with the teeth on their elongated jaw, but they usually were eventually able to work the prey and themselves free of the tether (e.g. by leaping) . Discussion Predator encounter rates averaged more than 50%, even in the most upstream locations. Thus, piscivorous predators are ubiguitous along the estuarine gradient and were effective at consuming tethered prey throughout the system.

PAGE 159

150 Q) P (0 u u 0) P c o o c CU X5 P •H •H o o P CD co o >i p •H c H 0 •H > 0 X5 P c •H P O tt O CO in CD H £* (0 Eh c H p CO rH o p P. (0 0) T3 a o) x P 0) Q, * P c 0) •H~ d) T) 43 IT) S M o ID > H i -P •H c -H o > *1 o c 0 •H p T3 C H 01 0) > 0) W J3 o 01 d) U 0) w m CN ID CN rH I O -an m rH m *. H rH rH CN rH CN CN CN rH rH rH O CN in o O o in o O in n 1 r> in in in l» in CN r~ in l in rI in 1 in 1 in in I in 1 O i-i rH rH t rH 00 + o Vfi rH O O + + O rH O rH CO + + O O 10 O O H + o m CN CN CN U M Q id U 0) js id c CO X! 0> C C -rl 0 H 3 0) 0) u 1-1 o p c 2 o> rl -0 CD CD ^ Oh Oh id a -r( o> id u a jj CO 01 I >i 0) id a U rH o< XI id ai 73 rH 3 -P u u id 73 •H rl 0 rH id 3 c H -P o u u u id Oi 0 CD XI C +J X! •h id P ft 01 C 4 Oi'H 11 0) Id rH d) H C rH H 0> co id oi 01 01 •H u 01 +J <0 01 0 H 01
H co >H 01 3 Q. g co U •H Id o H 01 O 0] •Q E 01 3 01 01 (D
PAGE 160

151 In all study locations, these predators included euryhaline species (e.g. snook, needlefish, gray snapper, crevalle jacks, juvenile barracuda) . Upstream, they were joined by freshwater species such as gar that can forage in brackish conditions. Similarly, lemon sharks are primarily marine predators that were also observed upstream. A relatively straightforward hypothesis can be given to explain the finding that a great number (90%) of the tethered fish were subject to predation in the midand downstream locations. Greater predator encounter rates are associated with proximity to some type of structurally massive habitat such as a rocky breakwater (Aronson 1989) or coral reef (Shulman 1985) . Similarly, findings in Chapter 3 indicate that where mangrove habitat is more developed (i.e. greater tree height, fringe width, canopy cover) greater abundances of large roving fish occur; this fish group includes many of the predators occurring in the study area. In addition, habitat development is significantly greater in midand downstream locations. Thus, the greater encounter rates found further downstream may be explained by the attraction of large roving predators to the structure associated with well-developed mangrove habitats. While attraction to highly developed mangrove habitats may explain why predation rates were so high at midand downstream locations, it does not account for the high rate of predation that was also found at the mid/upstream locations. In previous chapters, findings indicate that in

PAGE 161

152 upstream locations overall, large roving fish densities are lower (Chapter 2) . Conditions upstream overall may be inhospitable for large roving fish, preventing them from permanently residing in these locations (Chapters 3 and 4) . In addition, salinity regime in both the far/upand mid/upstream locations, was low and variable in the current study. Based on these findings, one would have expected the experiments to indicate lower predator encounter rates at both the far /upand mid/upstream locations. However, in actuality, tethered fish suffered lower rates of predation far/upstream than mid/upstream. Thus, the abundance of predators was probably equivalent at mid/up-, mid-, and downstream locations, but less abundant a far /upstream locations. The predator encounter rates, therefore, do not appear to simply be functions of salinity regime or mangrove habitat development alone. One hypothesis that could explain the greater rates mid/upstream in comparison to far/upstream, could be that a significant number of the stenohaline predators primarily residing in marine habitats also temporarily forage at the edge of their primary range (e.g. Weinstein 1979) . The mid/upstream locations would be at the edge of this range for marine predators. An analogous situation may occur for freshwater predators at the other side of the marine/ freshwater interface. To explain the lack of predators far /upstream, perhaps other characteristics of the these locations (besides

PAGE 162

153 salinity regime and mangrove habitat development) make them safer havens for small benthic fishes. In contrast to all the other locations, to forage far /upstream, a predator would have to negotiate a series of sinuous channels and interspersed ponds. Shallow shoals occur at pond/creek intersections that are sometimes only a few centimeters deep. This complex system seems likely to prevent access by casual foragers and predators above certain size limits. Based on these results, the ecological paradigm may thus be gualified. Large predators may be prevented from permanently occupying upstream locations by low and variable salinity conditions, and the complexity of sinuous channels may prevent them from foraging in far/upstream locations. A safe haven for small benthic fish thus occurs in complex habitats at the marine/ freshwater interface. These findings tend to support the hypothesis suggested by Browder & Moore (1981) : ideal juvenile fish habitat may occur where the variable salinity conditions overlap areas of such habitat complexity. They tend to refute, however, the hypothesis that small fishes are protected from predation in estuaries by lower salinities which tend to exclude stenohaline marine predators.

PAGE 163

CHAPTER 6 IMPLICATIONS AND CONCLUSIONS Implications for Mangrove Fish Ecology Fish and Mangrove Shorelines Mangrove shorelines fulfill the food and cover habitat requirements for many fishes. Among the functional roles of inter-tidal mangroves as habitats for fish, the best established is that of a temporary feeding location for a wide range of species (Robertson & Duke 1990a, Blaber et al. 1985, Morton 1990). These species include small schooling planktivores, benthic forage fish, and large piscivores. They spend periods of low tide in deep open water or shallow ponds and feed in the mangroves when the tide inundates the forest (Davis 1988) . Since northeastern Florida Bay is non-tidal, the mangrove habitats are unlike those in many other areas. In Florida Bay permanently inundated mangrove habitats are believed to provide young fish with refuge from larger predators (Thayer et al. 1987a). This role was supported in the current study, particularly for larger juveniles of the 154

PAGE 164

155 estuarine transient species, Lutjanus griseus and Haemulon sciurus. Of the habitats within Florida Bay, these species prefer mangroves during the day (Thayer et al. 1987a & 1987b) . An important aspect of habitat use by these species is their migration away from the mangrove shorelines at night, to feed in nearby seagrass beds (Starck & Schroeder 1971, Sogard et al. 1989c) . Thus, in both tidal and nontidal habitats, linkages between mangrove shorelines and other habitats may be critical for diel behavior patterns. Besides snappers and grunts, other species that may similarly rely on both mangroves and seagrass beds include snook, sheepshead, barracuda and nurse sharks. A common life history pattern occurs among these fishes: recruitment from offshore as postlarvae, settlement and growth in inshore habitats, and movement back offshore or to deeper water as they attain larger size classes (Starck & Schroeder 1971, DeSylva 1963, Jennings 1985). These fishes tend to use seagrass beds when they are smaller and move to mangroves when they attain larger juvenile sizes. Thus, shallow water habitats including mangroves and seagrass beds, may be linked to one another through such behavioral and life history patterns (Odum et al. 1982, Parrish 1989) . Mangroves provide cover and food resources that are very different from adjacent habitats dominated by submerged aguatic vegetation (SAV) . However, both types of habitats appear to be necessary to support certain fish species. Prime locations for supporting these species may

PAGE 165

156 occur where both mangrove development and SAV are greater. Reduced SAV may thus account for the reduced abundance of large roving fish upstream. Selection Among Mangrove Habitats Mangrove shorelines vary in functional value depending on the degree of development and habitat needs of the fishes. In freshwater streams, a gradient model has been proposed (Schlosser 1987) that appears to apply egually well in northeastern Florida Bay. This model identifies a gradient from upstream areas (environmentally unstable, shallower, lesser habitat development) to downstream areas (stable, deeper, greater development) . Greater habitat development is linked with the occurrence of more species and larger piscivorous individuals (Schlosser 1987) . For smaller fishes, upstream areas provide refugia from larger piscivores that are more abundant downstream. In northeastern Florida Bay, the mangrove habitats ranged from less developed upstream (small trees, narrow fringe, shallow water, high environmental variability) , to more developed downstream (tall trees, deeper water, environmentally more stable) . Among the large roving species, greater abundances of gray snappers and grunts were associated with more developed mangrove habitats in the current study. In addition, numbers of species of large roving fish were more abundant downstream than upstream. Thus, in terms of large roving fish, the model applies well to northeastern Florida Bay.

PAGE 166

157 For smaller fishes, the implications of the model are that upstream habitats are occupied by colonizing fishes. Competition for food resources limits the abundances of these fishes more than predation. Downstream, predation is a more powerful force in the community structure. Thus, upstream areas should be relatively advantageous for small fishes in both freshwater streams and estuarine habitats. Small fishes, however, were not more abundant upstream in northeastern Florida Bay. Young-of-the-year estuarine transient juveniles may represent the colonizers described in the model. However, these fishes were absent in the current study. Thus, there appears to be a missing component of the fish community in the study area, particularly in the upstream fish assemblages in the current study . Implications For Estuarine Fish Ecology: The Nursery-ground Hypothesis The absence of young-of-the-year juveniles of estuarine transient fishes in the study area was also a significant departure from the results one would have expected based on widely accepted theories in estuarine ecology. In northeastern Florida Bay, this condition was not unique to the mangroves; both seagrass and mangrove habitats in eastern and central Florida Bay also have low populations of very young transients (Sogard et al. 1987, 1989as, Thayer et al. 1987a).

PAGE 167

158 The potential pool of post-larval estuarine transient species from the Atlantic Ocean and Gulf of Mexico includes sciaenids, lutjanids, haemulids, centropomids and elopids. Presumably, currents carry potential recruits to the northeastern Florida Bay area. Exchange with these sources is limited in northeastern Florida Bay due to the presence of the Keys and western mudbanks. Internal circulation in Florida Bay is also weak due to the many islands and lack of tides. If post-larval transients do enter the northeastern Bay, they may meet with significant predation pressure due not only to the occurrence of piscivores but also to a reduction in the cover afforded by seagrass beds, which tend to become less developed in the eastern Bay. These conditions indicate that the chances of post-larval forms reaching the upstream locations in the study area may be small. Such conditions may not be unusual in estuaries, however . Habitat conditions documented in this study for the upstream locations included almost no SAV and reduced mangrove habitat development. However, shallow ponds and sinuous creeks upstream may effectively reduce predator encounter rates. Thus, a key factor in improving the use of upstream habitats by estuarine transient juveniles appears to be the presence of a persistent abundance of SAV. Management Implications Northeastern Florida Bay may have historically supported more estuarine transient juveniles and greater

PAGE 168

159 densities of fish than was observed in the current study. Higher, abruptly changing salinity conditions may somehow inhibit the development of lush communities of submerged aquatic vegetation that provide cover for small fish and benthic invertebrates (Montague et al. 1989) . Sustained lower salinity periods may promote growth of lush seagrass (Ruppia maritima) or algal (Chara, Batophora) communities (Tabb et al. 1961, Montague et al. 1989) . During more saline periods, less dense growth of Halodule wrightii communities may develop, if any vegetation grows at all. As salinity changes with seasons, these communities may alternate. Although the regular study was conducted during a period of very low rainfall and can only serve to provide information on the ecosystem under low freshwater inflow conditions, the pilot study took place at a time of higher rainfall and high freshwater inflow conditions (summer 1988) . In the pilot study, from October 1988 through March 1989, in upstream habitats, extremely dense Ruppia and algal communities were observed. With experimental traps and gill nets, great numbers of fishes were collected including gray snappers, jacks, catfish and cichlids. A small mangrove island and its surrounding waters in the Joe Bay study area were heavily used by white pelicans (fish eating birds) . However, during the regular study (May 1989 May 1990) , very little Ruppia was observed. The pelican island was evidently never used by birds and because the traps no

PAGE 169

160 longer captured many fish, their use was abandoned. This evidence indicates that the Ruppia community was probably supporting a greater fish population in the area than observed at any time during the course of the regular study. Thus, northeastern Florida Bay, in the drought year recorded in this study, was unusual in comparison to other tropical, subtropical and warm temperate estuaries. Water management efforts may be needed to restore sustained low salinity periods, thereby inducing greater submerged aguatic vegetation development and greater influx of estuarine transient juveniles upstream into the more protected habitats.

PAGE 170

LITERATURE CITED Afifi, A. A. and V. Clark 1984. Computer-aided multivariate analysis. Wadsworth Inc., Belmont, California. 458 pages. Aronson, R. B. 1988. Palatability of five Caribbean ophiuroids. Bulletin of Marine Science 43(1): 93-97. Aronson, R. B. 1989. Brittlestar beds: low-predation anachronisms in the British Isles. Ecology 70(4): 856-865. Austin, H. M. 1971. A survey of the ichthyofauna of the mangroves of western Puerto Rico during December, 1967August, 1968. Caribbean Journal of Science 11(1-2): 27-39. Baelde, P. 1990. Differences in the structures of fish assemblages in Thalassia testudinum beds in Guadeloupe, French West Indies, and their ecological significance. Marine Biology 105: 163-173. Barshaw, D. E. 1990. Tethering as a technigue for assessing predation rates in different habitats: an evaluation using juvenile lobsters Homarus americanus. Fishery Bulletin 88: 415-417. Barshaw, D. E. and K. W. Able 1990. Deep burial as a refuge for lady crabs Ovalipes ocellatus: comparisons with blue crabs Callinectes sapidus. Marine Ecology Progress Series 66: 75-79. Bell, J. D., D. A. Pollard, J. J. Burchmore, B. C. Pease, and M. J. Middleton 1984. Structure and function of a fish community in a temperate tidal mangrove creek in Botany Bay, New South Wales. Australian Journal of Marine and Freshwater Resources 35:33-46. Beumer, J. P. 1978. Feeding ecology of four fishes from a mangrove creek in north Queensland, Australia. Journal of Fish Biology 12: 475-490. Blaber, S. J. M. and T. G. Blaber 1980. Factors affecting the distribution of juvenile estuarine and inshore fish. Journal of Fish Biology 17: 143-162 161

PAGE 171

162 Blaber, S. J. M. , D. T. Brewer, and J. P. Salini 1989. Species composition and biomasses of fishes in different habitats of a tropical northern Australian estuary: their occurrence in the adjoining seas and estuarine dependence. Estuarine, Coastal and Shelf Science 29:509-531. Blaber, S. J. M. , J. W. Young and M. C. Dunning 1985. Community structure and zoogeographic affinities of the coastal fishes of the Dampier Region of north-western Australia. Australian Journal of Marine and Freshwater Research 36:247-266. Brook, I. M. 1977. Trophic relationships in a seagrass community (Thalassia testudinum) , in Card Sound, Florida. Fish diets in relation to macrobenthic and cryptic faunal abundance. Transactions of the American Fisheries Society 106(3): 219-229. Browder, J. A. and D. Moore 1981. A new approach to determining the guantitative relationships between fishery production and the flow of freshwater to estuaries. In R. Cross and D. Williams (Eds.) Proceedings of the National Symposium on Freshwater inflow to estuaries. U. S. Fish and Wildlife Service, Office of Biological Services, Volume I: 403-430. Carter, L. A. Burns, T. R. Cavinder, K. R. Dugger, P. L. Fore, D. B. Hicks, H. L. Revells, T. W. Schmidt 1973. Ecosystems analysis of the Big Cypress Swamp and estuaries. U. S. Environmental Protection Agency, Atlanta, Georgia. Chester, A. J. and G. W. Thayer 1990. Distribution of spotted seatrout (Cynoscion nebulosus) and gray snapper (Lutjanus griseus) juveniles in seagrass habitats of western Florida Bay. Bulletin of Marine Science 46(2): 345-357. Croker, R. A. 1960. Growth and food of the gray snapper, Lutjanus griseus, in Everglades National Park. Transactions of the American Fisheries Society 91(4): 375-378. Darnell, R. M. 1961. Trophic spectrum of an estuarine community, base on studies of Lake Ponchartrain, Louisiana. Ecology 42(3): 553-568. Day, J. W. Jr., C. A. S. Hall, W. M. Kemp, A. YanezArancibia 1989. Estuarine Ecology, John Wiley & Sons, New York, New York. 558 pages. Davis, T. L. O. 1988. Temporal changes in the fish fauna entering a tidal swamp system in tropical Australia. Environmental Biology of Fishes 21(3): 161-172.

PAGE 172

163 Deaton, L. E. and M. J. Greenberg 1986. There is no horohalinicum. Estuaries. 9(1): 20-30. DeSylva, D. P. 1963. Systematics and life history of the great barracuda. Studies in Tropical Oceanography, Institute of Marine Science, University of Miami, Miami, Florida. 179 pages. Dibble, E. 1991. A comparison of diving and rotenone methods for determining relative abundance of fish. Transactions of the American Fisheries Society (In press) . Ebeling, A. W. and D. R. Laur 1985. The influence of plant cover on surf perch abundance at an offshore temperate reef. Environmental Biology of Fishes 12(3): 169-179. Flores-Verdugo, F. J., J. W. Day, L. Mee and R. BrisenoDuenas 1988. Phytoplankton production and seasonal biomass variation of seagrass, Ruppia maritima , in a tropical Mexican lagoon with an ephemeral inlet. Estuaries 11(1): 51-56. Flores-Verdugo, F. , F. Gonzalez-Farias, 0. Ramirez-Flores 1990. Mangrove ecology, aguatic primary productivity, and fish community dynamics in the Teacapan-Agua Brava Lagoon estuarine system (Mexican Pacific). Estuaries 13(2): 219230. Funicelli, N. A., H. E. Bryant, M. R. Dewey, G. M. Ludwig, D. A. Meineke, L. J. Mengel, and J. E. Skjeveland 1986. Movements, relative importance, and standing stock of important sport and commercial species in Everglades National Park, Florida. U. S. Fish and Wildlife Service, Gainesville, Florida. 186 p. Gilmore, R. G. , C. J. Donohoe, D. W. Cooke 1983. Observations on the distribution and biology of east-central Florida populations of the common snook, Centropomus undecimalis . Florida Scientist Special Supplement 45(3/4) :313-336. Ginsburg, R. N. 1956. Environmental relationships of grain size and constituent particles in some south Florida carbonate sediments. Bulletin of the American Association of Petroleum Geologists. 40(10): 2384-2427. Grossman, G. D., D. M. Nickerson, and M. C. Freeman 1991. Principal component analysis of assemblage structure data: utility of tests based on eigenvalues. Ecology 72(1): 341-347. Gunter, G. 1938. Seasonal variations in abundance of certain estuarine and marine fishes in Louisiana, with particular reference to life histories. Ecological

PAGE 173

164 Monographs 8(3): 314-336. Gunter, G. 1961. Some relations of estuarine organisms to salinity. Limnology and Oceanography VI(2): 182-190. Gunter, G. 1967. Some relationships of estuaries to the fisheries of the Gulf of Mexico, in G. H. Lauff ed. Estuaries, American Association for the Advancement of Science, Washington, D. C. pages 621-638. Harrington, R. W. , and E. S. Harrington 1961. Food selection among fishes invading a high subtropical salt marsh: from onset of flooding through the progress of a mosguito brood. Ecology 42(4): 646-666. Hay, M. E., J. R. Pawlik, J. E. Duffy, and W. Fenical 1989. Seaweed-herbivore-predator interactions: host-plant specialization reduces predation on small herbivores. Oecologia 81: 418-427. Heald, R. J., W. E. Odum and D. C. Tabb 1974. Mangroves in the estuarine food chain. in P. J. Gleason (Ed.) Environments of south Florida: present and past. Memoir 2 Miami Geological Society, Miami, Florida 182-189. Heck, Jr., K. L. and T. A. Thoman 1981. Experiments on predator-prey interactions in vegetated aguatic habitats. Journal of Experimental Marine Biology and Ecology 53: 125134. Heck, Jr., K. L. and K. A. Wilson 1987. Predation rates on decapod crustaceans in latitudinally separated seagrass communities: a study of spatial and temporal variation using tethering technigues. Journal of Experimental Marine Biology and Ecology 107: 87-100. Hellawell, J. M. and R. Abel 1971. A rapid volumetric method for the analysis of the food of fishes. Journal of Fish Biology 3:29-37 Herrnkind, W. F. and M. J. Butler IV 1986. Factors regulating postlarval settlement and juvenile microhabitat use by spiny lobsters Panulirus argus . Marine EcologyProgress Series 34: 23-20. Hettler, Jr., W. F. 1989. Food habits of juveniles of spotted seatrout and gray snapper in western Florida Bay. Bulletin of Marine Science 44(1): 155-162. Holling, C. S. 1966. The functional response of invertebrate predator to prey density. Mem. Entomo. Soc. Canada, No. 48, 86 p. Holmguist, J. G. , G. V. N. Powell and S. M. Sogard 1989a.

PAGE 174

Decapod and stomatopod communities of seagrass-covered mudbanks in Florida Bay: interand intra-bank heterogeneity with special reference to isolated subenvironments. Bulletin of Marine Science 44(1): 251-262. Holmguist, J. G., G. V. N. Powell and S. M. Sogard 1989b. Sediment, water level and water temperature characteristics of Florida Bay's grass-covered mudbanks. Bulletin of Marine Science 44(1): 348-364. Hynes, H. B. N. 1950. The food of fresh-water sticklebacks (Gasterosteus aculeatus and Pygosteus pungitius) , with a review of methods used in studies of the food of fishes. Journal of Animal Ecology 19: 36-58. Hyslop, E. J. 1980. Stomach contents analysis — a review of methods and their application. Journal of Fish Biology 17:411-429. Jaap, W. C. 1984. The ecology of the south Florida coral reefs: a community profile. U. S. Fish and Wildlife Service. FWS/OBS 82/08. 138 pages. Jennings, C. A. 1985. Species profiles: life histories and environmental reguirements of coastal fishes and invertebrates (Gulf of Mexico) — sheepshead. U. S. Fish and Wildlife Service. Slidell, Louisiana. 10 p. Kitching, R.L. 1983. Systems ecology. University of Queensland Press, St. Lucia, Queensland, Australia, 280 pages. Kleinbaum, D. G. and L. L. Kupper 1978. Applied regression analysis and other multivariate methods. Duxbury Press, Belmont, California 486 pages. Lee, D. S., C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D. E. McAllister, and J. R. Stauffer, Jr. 1980. Atlas of North American freshwater fishes, North Carolina State Museum of Natural History. Raleigh, North Carolina. 867 pages. Lindall, W. N. Jr., J. R. Hall, W. A. Fable, and L. A. Collins 1973. A survey of fishes and commercial invertebrates of the nearshore and estuarine zone between Cape Romano and Cape Sable, Florida. National Marine Fisheries Service, Panama City, Florida, 62 p.

PAGE 175

166 Little, M. C. f P. J. Reay, S. J. Grove, 1988. The fish community of an East African mangrove creek. Journal of Fish Biology 32:729-747. Livingston, R. J. 1982. Trophic organization of fishes in a coastal seagrass system. Marine Ecology Progress Series 7: 1-12. Loneragan, N. R. and I. C. Potter 1990. Factors influencing community structure and distribution of different life-cycle categories of fishes in shallow waters of a large Australian estuary. Marine Biology 106: 25-37. Luckhurst, B. E. and K. Luckhurst 1978. Analysis of the influence of substrate variables on coral reef fish communities. Marine Biology 49: 317-323. McHugh, J. L. 1967. Estuarine nekton, in G. H. Lauff (Ed.) Estuaries. American Association for the Advancement of Science, Washington, D. C. pages 581-620. Mclvor, C. C. and W. E. Odum 1988. Food, predation risk, and microhabitat selection in a marsh fish assemblage. Ecology 69(5): 1341-1351. McNeely, D. L. 1987. Niche relations within an Ozark cyprinid assemblage. Environmental Biology of Fishes 18 (3) : 195-208. Miller, J. M. and M. L. Dunn 1980. Feeding strategies and patterns of movement in juvenile estuarine fishes. in Estuarine Perspectives. Academic Press, Inc. New York, New York. 437-448 pages. Montague, C. L. , R. D. Bartelson, and J. A. Ley 1989. Assessment of benthic communities along salinity gradients in northeast Florida Bay, University of Florida, Gainesville, for the South Florida Water Management District, West Palm Beach, Florida. Morton, R. M. 1990. Community structure, density, and standing crop of fishes in a subtropical Australian mangrove area. Marine Biology 105: 385-394. Moser, M. L. , L. R. Gerry 1989. Differential effects of salinity changes on two estuarine fishes, Leisostomus xanthurus and Micropogonius undulatus. Estuaries 12(1): 3541.

PAGE 176

167 Moyle, P. B. and T. J. Cech, Jr. 1988. Fishes: an introduction to Ichthyology. Prentice Hall, Englewood Cliffs, New Jersey, 559 pages. Nielsen, L. A. and D. L. Johnson 1983. Fisheries Technigues. American Fisheries Society, Bethesda, Maryland. 468 pages. Nordlie, F. G. , and S. J. Walsh 1989. Adaptive radiation in osmotic regulatory patterns among three species of Cyprinodontids (Teleostei: Atherinomorpha) . Physiological Zoology 62 (6) : 1203-1218. Odum, E. P. 1983. Basic Ecology. CBS College Publishing, New York, New York. 613 pages. Odum, W. E. 1971. Pathways of energy flow in a south Florida estuary. Ph.D. dissertation, University of Miami, Coral Gables, Florida. 162 pages. Odum, W. E. and E. J. Heald 1972. Trophic analysis of an estuarine mangrove community. Bulletin of Marine Science 22(3): 671-738. Odum, W. E., C. C. Mclvor, and T. J. Smith, III. 1982. The ecology of the mangroves of south Florida: a community profile. U. S. Fish and Wildlife service, Office of Biological Services, Washington, D. C. FWS/OBS-81/24 . 144 pages. Phillips, R. R. and S. B. Swears 1979. Social hierarchy, shelter use, and avoidance of predatory toadfish (Opsanus tau) by the striped blenny (Chasmodes bosquianus) . Animal Behavior 27: 1113-1121. Pinto, L. 1987. Environmental factors influencing the occurance of juvenile fishes in the mangroves of Pagbilao, Phillipines. Hydrobiologia 150:283-301. Power, P. E. and W. J. Matthews 1983. Algae-grazing minnows (Campostoma anomalum) , piscivorous bass (Micropterus spp.), and the distribution of attached algae in a small prairiemargin stream. Oecologia (Berlin) 60: 328-332. Provost, M. W. 1973. Mean high water mark and use of tidelands in Florida. Florida Scientist 36(1). 50-66. Reid, R. K. 1954. An ecological study of the Gulf of Mexico fishes in the vicinity of Cedar Key, Florida. Bulletin of Marine Science 4(1): 1-94. Remane, A. and C. Schlieper 1971. Biology of brackish water. Wiley Interscience Division, New York, New York.

PAGE 177

168 Robblee, M. B. 1987. The spatial distribution of the nocturnal fish fauna of a tropical seagrass feeding ground. Ph.D. Dissertation, University of Virginia, Charlottesville, Virginia. 144 pages. Robertson, A. I. and N. C. Duke 1987. Mangroves as nursery sites: comparisons of the abundance and species compostion of fish and crustaceans in mangroves and other nearshore habitats in tropocal Australia. Marine Biology 96: 193-205. Robertson, A. I. and N. C. Duke 1990a. Mangrove fishcommunities in tropical Queensland, Australia: spatial and temporal patterns in densities, biomass and community structure. Marine Biology 104: 369-379. Robertson, A. I. and N. C. Duke 1990b. Recruitment, growth and residence time of fishes in a tropical Australian mangrove system. Estuarine, Coastal and Shelf Science 31: 723-743. Robins, C. R. , G. C. Ray, J. Douglass, 1986. A field guide to Atlantic Coast fishes of North America. Houghton Mifflin Company, Boston. 355 pages. Rogers, T. E. Target, S. B. Van Sant 1984. Fish-nursery use in Georgia salt marsh estuaries: the influence of springtime freshwater conditions. Transactions of the American Fisheries Society 113: 595-606. Rozas, L. P. and M. W. LaSalle 1990. A comparison of diets of Gulf killifish, Fundulus grandis Baird and Girard, entering and leaving a Mississippi brackish marsh. Estuaries 13 (3) : 332-336. Rozas, L. P. and W. E. Odum 1988. Occupation of submerged aguatic vegetation by fishes: testing the roles of food and refuge. Oecologia 77: 101-106. Rutherford, E. S., J. T. Tilmant, E. B. Thue and T. W. Schmidt, 1989. Fishery harvest and population dynamics of spotted seatrout, Cynoscion nebulosus, in Florida Bay and adjacent waters. Bulletin of Marine Science 44 (1) : 108-125 . Sale, P. F. 1980. Assemblages of fish on patch reefs — predictable or unpredictable? Environmental Biology of Fishes 5(3): 243-247. Sale, P. F. and W. A. Douglas 1984. Temporal variability in the community structure of fish on coral patch reefs and the relation of community structure to reef structure. Ecology 65(2): 409-422. Salini, J. P., S.J.M. Blaber and D. T. Brewer 1990. Diets of piscivorous fishes in a tropical Australian estuary, with

PAGE 178

169 special reference to predation on penaeid prawns. Marine Biology 105: 363-374. Schlosser, I. J. 1987. A conceptual framework for fish communities in small warmwater streams. In: Matthews, W. J. and D. C. Heins (Eds.) Community and evololutionary ecology of North American stream fishes. Oklahoma University Press, Norman, Oklahoma, pages 17-26. Schmitt, R. J. and S. J. Holbrook 1985. Patch selection by juvenile black surfperch (Embiocidae) under variable risk: interactive influence of food quality and structural complexity. Journal of Experimental Marine Biology and Ecology 85: 269-285. Schomer, N. S. and R. D. Drew 1982. An ecological characterization of the lower Everglades, Florida Bay, and the Florida Keys. U. S. Fish and Wildlife Service, Office of Biological Services, Washington, D. C. 24688. Seaman, W. Jr. and M. Collins 1983. Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (South Florida) — snook. U. S. Fish and Wildlife Service. Slidell, Louisiana. 16 pages. Sheridan, P. F. 1991. Comparative habitat utilization by estuarine macrofauna within the mangrove ecosystem of Rookery Bay, Florida. Bulletin of Marine Science 49(3). Shulman, M. J. 1985. Recruitment of coral reef fishes: effects of distribution of predators and shelter. Ecology 66(3): 1056-1066. Smith III, T. J. 1987. Seed predation in relation to tree dominance and distribution in mangrove forests. Ecology 68(2): 266-273. Smith, III T. J. and N. C. Duke 1987. Physical determinants of inter-estuary variation in mangrove species richness around the tropical coastline of Australia. Journal of Biogeography 14: 9-19. Sogard, S. M. and K. W. Able 1991. A comparison of eelgrass, sea lettuce macroalgae, and marsh creeks as habitats for epibenthic fishes and crustaceans. Estuarine, Coastal and Shelf Science 33: 501519. Sogard, S. M. , G. V. N. Powell, and J. G. Holmquist 1987. Epibenthic fish communities on Florida Bay banks: relations with physical parameters and seagrass cover. Marine Ecology Progress Series 40:25-39. Sogard, S. M. , G. V. N. Powell, and J. G. Holmquist 1989a. Spatial distribution and trends in abundance of fishes

PAGE 179

170 residing in seagrass meadows on Florida Bay mudbanks. Bulletin of Marine Science 44(1): 179-199. Sogard, S. M. , G. V. N. Powell, and J. G. Holmguist 1989b. Utilization by fishes of shallow, seagrass-covered banks in Florida Bay: 1. species composition and spatial heterogeneity. Environmental Biology of Fishes 24(1): 5365. Sogard, S. M. , G. V. N. Powell, and J. G. Holmguist 1989c. Utilization by fishes of shallow, seagrass-covered banks in Florida Bay: 2. diel and tidal patterns. Environmental Biology of Fishes 24(2): 81-92. South Florida Water Management District 1989. Surface water improvement and management (SWIM) plan for the Everglades: vol. II. planning and implementation. West Palm Beach, Florida Starck, W. A. II, and R. E. Schroeder, 1971. Investigations on the gray snapper, Lutjanus griseus. University of Miami Press, Coral Gables, Florida 150 pages. Stoner, A. W. 1986. Community structure of the demersal fish species of Laguna Joyuda, Puerto Rico, Estuaries 9(2): 142-152. Sutherland, J. P. 1980. Dynamics of the epibenthic community on roots of the mangrove Rhizophora mangle, at Bahia de Buche, Venezuela. Marine Biology 58, 75-84. Tabb, D. C. , T. R. Alexander, T. M. Thomas, N. Maynard 1967. The physical, biological and geological character of the area south of C-lll Canal in extreme southeast Everglades National Park, University of Miami, Coral Gables. Tabb, D. C. , D. L. Dubrow, and R. B. Manning 1962. The ecology of northern Florida Bay and adjacent estuaries. Florida Board of Conservation Technical Series no. 39, Tallahassee. Thayer, G. W. and A. J. Chester 1989. Distribution and abundance of fishes among basin and channel habitats in Florida Bay. Bulletin of Marine Science 44: 200-219. Thayer, G. W. , D. R. Colby and W. F. Hettler, Jr., 1987a. Utilization of the red mangrove prop root habitat by fishes in south Florida. Marine Ecology Progress Series 35:25-38. Thayer, G. W. , W. . F. Hettler, A. J. Chester, D. R. Colby and P. J. McElhaney 1987b. Distribution and abundance of fish communities among sected estuarine and marine habitats in Everglades National Park. South Florida Research Center Report SFRC-87/02. 166 pages.

PAGE 180

171 Van Engle, W. A. and E. B. Joseph 1967. Characteristics of coastal and estuarine nursery grounds as natural communities. Unpublished report to U.S. Fish and Wildlife Service, 43 pages. Available from: Virginia Institute of Marine Science, Glouchester Point, Virginia. Weinstein, M. P. 1979. Shallow marsh habitats as primary nurseries for fishes and shellfish, Cape Fear River, North Carolina. Fishery Bulletin 77(2): 339-357. Weinstein, M. P. and R. W. Davis 1980. Collection efficiency of seine and rotenone samples from tidal creeks, Cape Fear, N. C, Estuaries 3(2): 98-105. Weisberg, S. B. and V. A. Lotrich 1982. Ingestion, egestion, excretion, growth, and conversion efficiency for the mummichog, Fundulus heteroclitus . Journal of Experimental Marine Biology and Ecology 62: 237-249. Werner, E., J. F. Gilliam, D. J. Hall and G. G. Mittelbach 1983. An experimental test of the effects of predation risk on habitat use by fish. Ecology 64: 1540-1548. Wilson, K. A. 1989. Ecology of mangrove crabs: predation, physical factors and refuges. Bulletin of Marine Science 44(1): 263-273. Wilson, K. A., K. W. Able, K. L. Heck 1990. Predation rates on juvenile crabs in estuarine nursery habitats: evidence for the importance of macroalgae (Ulva lactuca) . Marine Ecology-Progress Series 58: 243-251. Wilson, K. A., K. L. Heck and K. W. Able 1987. Juvenile blue crab, Callinectes sapidus , survival: an evaluation of eelgrass, Zoster a marina, as refuge. Fishery Bulletin 85(1): 53-58. Yanez-Arancibia, A., F. A. Linares, and J. W. Day, Jr. 1980. Fish community structure and function in Terminos Lagoon, a tropical estuary in the southern Gulf of Mexico. Estuarine Perspectives, Academic Press, Inc. 465-482. Zieman, J. C, J. W. Fourgurean and R. L. Iverson 1989. Distribution, abundance and productivity of seagrasses and macroalgae in Florida Bay. Bulletin of Marine Science 44(1): 292-311.

PAGE 181

BIOGRAPHICAL SKETCH Janet Ann Ledtke Ley was born 11 June, 1951, in Detroit, Michigan, to Frederick G. and Helen M. Ledtke. Janet graduated from Rochester High School, Rochester, Michigan, in 1969. She received her Bachelor of Science degree in resource development at Michigan State University, East Lansing, Michigan, in 1973. Janet devoted ten years to environmental planning for the Pinellas County government, in Clearwater, Florida, from 1974 through 1984. While working as a planner, in 1979, she earned her Master of Science degree at the University of South Florida. Her thesis was entitled "Exploring Transfer of Development Rights," a concept that was later incorporated into her work on Pinellas County's plan for environmental protection of wetland ecosystems. Janet worked as a consultant for the Tampa office of Dames & Moore during 1985. In 1986, Janet enrolled in the Ph.D. program in systems ecology, in the College of Environmental Engineering Sciences, at the University of Florida, Gainesville, Florida. In May 1992, Janet received her Ph.D., and hopes to continue to work in ecosystems research. 172

PAGE 182

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the deqree of Doctor of Philosophy. _ Aj. Montaque// Chairman Associate Professor of Environmental Enqineerinq Sciences I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the deqree of Doctor of Philosophy. n Carole C. Mclvor, Cochairman Assistant Professor of Forest Resources and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the deqree of Doctor of Philosophy. Professor of Zooloqy I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the deqree of Doctor of Philosophy. G. Ronnie Best Scientist Environmental Enqineerinq Sciences

PAGE 183

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the deqree of Doctor of Philosophy. William Seaman, Jr. Associate Professor of ^ Forest Resources and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the deqree of Doctor of Philosophy. Nicholas Funicelli Assistant Professor of Forest Resources and Conservation This dissertation was submitted to the Graduate Faculty of the Colleqe of Enqineerinq and to the Graduate School and was accepted as partial fulfillment of the requirements for the deqree of Doctor of Philosophy. May 1992 Q

CHAPTER 4
FOOD HABITS OF MANGROVE FISHES:
A COMPARISON ACROSS SALINITY GRADIENTS
Diets have been studied for fishes associated with
whole estuaries (Darnell 1961), mangroves (Odum 1971, Beumer
1978, Salini et al. 1990), salt marshes (Harrington &
Harrington 1961, Rozas & LaSalle 1990), seagrass beds
(Livingston 1982), lakes (Werner & Hall 1983), and streams
(McNeely 1987). One successful dietary strategy identified
in extremely variable habitats is omnivory, or the
consumption of a wide variety of prey organisms including
both plant and animal material (Darnell 1961, Harrington &
Harrington 1961). Opportunism, or an ability to exploit
alternative foods depending on availability, is also a
valuable dietary strategy for survival in variable
environments (Odum & Heald 1972, Livingston 1982, Salini et
al. 1990). Trophic systems in aquatic habitats are often
characterized by shared common resources among the various
species of fishes (Harrington & Harrington 1961, Livingston
1982) .
In estuaries, seasonality of freshwater inflow
increases habitat variability upstream relative to more
120


65
significantly among the gradient positions. Atherinomorus
stipes, the hardhead silverside, was more abundant
downstream; Menidia spp. was more abundant upstream.
Individuals of both Menidia beryllina and Menidia peninsulae
were collected. The distribution of these species overlaps
in northeastern Florida Bay, and distinctive characters are
extremely difficult to confirm (C. Gilbert, personal
communication). Thus, Menidia spp. has been used in this
study to designate these species. Anchoa mitchelli,
although not influenced by gradient or system, was
significantly more abundant at Little Blackwater Sound
(mid/east) than at the other general locations.
Patterns varied in spatial distributions for the top
three species of large roving fish (Figure 2-9). For these
species, repeated measures ANOVA results also varied (Table
2-14 and 2-15). Among these species, Haemulon sciurus was
never present upstream. Lutjanus griseus was significantly
less abundant up- than mid- or downstream. In contrast,
Strongylura notata had significantly greater densities
upstream/east.
Correlations. To further analyze spatial trends for
these nine species, correlations between mean densities and
salinity, salinity variation, and water depth were
determined and are presented in Table 2-16. Densities of
Menidia spp. were greater at locations with lower mean
salinities and greater variation. In contrast, densities of
Atherinomorus stipes were greater at locations with higher


1.6
West
I
CD
-4-^
CD
E
0)
L-
o
D
CT
CO
L_
0)
CL
1~
V)
L_
1.2
0.8
0.4
0.0
1.6
1.2
0.8
0.4
0.0
Large Roving Fish
Density Sc Salinity vs. Season
Visual Census
East
60
40
20
0
60
40
20
0
Figure 2-5. Density of large roving fish sampled by visual census in
mangrove habitats (histograms) and corresponding salinity measurements
(lines). Error bars indicate standard deviation among the visual censuses.
Ui
Parts per Thousand


151
In all study locations, these predators included
euryhaline species (e.g. snook, needlefish, gray snapper,
crevalle jacks, juvenile barracuda). Upstream, they were
joined by freshwater species such as gar that can forage in
brackish conditions. Similarly, lemon sharks are primarily
marine predators that were also observed upstream.
A relatively straightforward hypothesis can be given to
explain the finding that a great number (90%) of the
tethered fish were subject to predation in the mid- and
downstream locations. Greater predator encounter rates are
associated with proximity to some type of structurally
massive habitat such as a rocky breakwater (Aronson 1989) or
coral reef (Shulman 1985) Similarly, findings in Chapter 3
indicate that where mangrove habitat is more developed (i.e.
greater tree height, fringe width, canopy cover) greater
abundances of large roving fish occur; this fish group
includes many of the predators occurring in the study area.
In addition, habitat development is significantly greater in
mid- and downstream locations. Thus, the greater encounter
rates found further downstream may be explained by the
attraction of large roving predators to the structure
associated with well-developed mangrove habitats.
While attraction to highly developed mangrove habitats
may explain why predation rates were so high at mid- and
downstream locations, it does not account for the high rate
of predation that was also found at the mid/upstream
locations. In previous chapters, findings indicate that in


Figure 2-11. Cluster analysis dendrograms based on species sampled by visual
census methods. Stations that grouped differently than actual gradient positions
are designated as misclassified.
a. Presence of each species (each species weighted equally) was used in one
analysis and,
b. Density of each species (fish per square meter) was used in the other.


fish taxonomy was invaluable. Katy Kuss was both a great
adviser and a good friend.
Gordon Thayer, National Marine Fisheries Service,
advised me on the use of enclosure nets. In addition, other
visiting scientists shared ideas with me, including Paul
Carlson, Florida DNR; Jay and Rita Zieman and Jim
Fourqurean, University of Virginia; and Dave Porter,
University of Georgia. I am also grateful to the scientists
of the National Audubon Society in Tavernier, Florida, who
treated me as an adjunct staff member, especially George
Powell, Mike Ross and Jerry Lorenz.
In Gainesville, Ken Portier helped in the initial study
design and last phases of analysis. In the bulk of the
analysis effort, Steve Linda advised me on handling a very
large data set. Hans Gottgens, my officemate, was
constantly patient and extremely helpful in offering
computer, scientific and personal advice.
Most importantly, I would like thank those who
encouraged my pursuit of this degree as a personal goal and
supported me throughout the process. These are my parents
who nurtured my spirit of independence and appreciation of
nature, and Darlene Kalada, my best friend, on whose support
and encouragement I could always depend.
iv


INFLUENCE OF CHANGES IN FRESHWATER FLOW ON THE USE OF
MANGROVE PROP ROOT HABITAT BY FISHES
By
JANET A. LEY
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1992
imiVERSin OF ROMBft,1WKKS


146
Predator Encounter Rates
Mean and Standard Deviation
100
-t-1 80
C
CD
O 60
L_
0)
Q_ 40
20
40
Salinity
CL
Cl
to
cu
E
o
100
80
60
40
Missing Fish
L
Visibility
Water Depth
1
Far
Up
Mid
Up
Mid Down
Figure 5-3. Mean predation rates for small (<10 cm TL)
benthic forage fish tethered for 3 hours across a salinity
gradient in northeastern Florida Bay. Error bars indicate
the standard deviations of test rates resulting from 4 tests
in each of the 4 gradient positions.


Table 2-3, continued
Group/
Residency
Family Species
General Locations
Up-west
Nets Visual
Mid-west
Nets Visual
Down-west
Nets Visual
Up-east
Nets Visual
Mid-east
Nets Visual
Down-
Nets
east
Visual
Total
Nets
Visual
Lepisosteldae (gar)
LR/o
Lepisosteus platyrhincus
1
1
Centrarchidae (sunfish)
BF/r
Lepomis macrochirus
7
1
20
1
27
Cichlidae (cichlid)
BF/r
Cichlasoma urophthalmus
85
387
7
977
90
1,062
484
BF/r
Tilapia mariae
3
3
5
3
8
Ephippidae (spadefish)
LR/o
Chaetodipterus faber
3
15
18
Scaridae (parrotfish)
LR/o
Sparisoma radians
2
1
3
LR/o
Scaridae (sp unk)
3
1
4
Lobotidae (tripletail)
LR/o
Lobotes surinamensis
2
1
3
Mugilidae (mullet)
LR/o
Mugil cephalus
11
90
6 12
45
1
3
10
1
1211
29
1361
LR/o
Mugil curema
1
22
1
24
LR/o
Mugil liza
1
1
5
0
7
Sphyraenidae (barracuda)
LR/ta
Sphyraena barracuda
12
1
17 196
18 185
3
1
29 90
25
352
104
825
Bleniidae (combtooth blenny)
BF/r
Chasmodes saburrae
6
3
1
2
2
14
Gobiidae (goby)
BF/r
Microgobius gulosus
1,011
675
92
125 33
1,648
81
110 6
11
9
2,997
804
BF/r
Gobiosoma robu stum
235
239
1
59
534
BF/r
Lophogobius cyprinoides
151
18
1
18
44
30 3
1
200
66
BF/r
Gobiosoma bosci
124
1
125
continued
OJ
00


residing in seagrass meadows on Florida Bay mudbanks.
Bulletin of Marine Science 44(1): 179-199.
Sogard, S. M., G. V. N. Powell, and J. G. Holmquist 1989b.
Utilization by fishes of shallow, seagrass-covered banks in
Florida Bay: 1. species composition and spatial
heterogeneity. Environmental Biology of Fishes 24(1): 53-
65.
Sogard, S. M., G. V. N. Powell, and J. G. Holmquist 1989c.
Utilization by fishes of shallow, seagrass-covered banks in
Florida Bay: 2. diel and tidal patterns. Environmental
Biology of Fishes 24(2): 81-92.
South Florida Water Management District 1989. Surface water
improvement and management (SWIM) plan for the Everglades:
vol. II. planning and implementation. West Palm Beach,
Florida
Starck, W. A. II, and R. E. Schroeder, 1971. Investigations
on the gray snapper, Lutjanus griseus. University of Miami
Press, Coral Gables, Florida 150 pages.
Stoner, A. W. 1986. Community structure of the demersal
fish species of Laguna Joyuda, Puerto Rico, Estuaries 9(2):
142-152.
Sutherland, J. P. 1980. Dynamics of the epibenthic
community on roots of the mangrove Rhizophora mangle, at
Bahia de Buche, Venezuela. Marine Biology 58, 75-84.
Tabb, D. C. T. R. Alexander, T. M. Thomas, N. Maynard
1967. The physical, biological and geological character of
the area south of C-lll Canal in extreme southeast
Everglades National Park, University of Miami, Coral Gables.
Tabb, D. C., D. L. Dubrow, and R. B. Manning 1962. The
ecology of northern Florida Bay and adjacent estuaries.
Florida Board of Conservation Technical Series no. 39,
Tallahassee.
Thayer, G. W. and A. J. Chester 1989. Distribution and
abundance of fishes among basin and channel habitats in
Florida Bay. Bulletin of Marine Science 44: 200-219.
Thayer, G. W., D. R. Colby and W. F. Hettler, Jr., 1987a.
Utilization of the red mangrove prop root habitat by fishes
in south Florida. Marine Ecology Progress Series 35:25-38.
Thayer, G. W., W.. F. Hettler, A. J. Chester, D. R. Colby
and P. J. McElhaney 1987b. Distribution and abundance of
fish communities among sected estuarine and marine habitats
in Everglades National Park. South Florida Research Center
Report SFRC-87/02. 166 pages.


special reference to predation on penaeid prawns. Marine
Biology 105: 363-374.
Schlosser, I. J. 1987. A conceptual framework for fish
communities in small warmwater streams. In: Matthews, W. J.
and D. C. Heins (Eds.) Community and evololutionary ecology
of North American stream fishes. Oklahoma University Press,
Norman, Oklahoma, pages 17-26.
Schmitt, R. J. and S. J. Holbrook 1985. Patch selection by
juvenile black surfperch (Embiocidae) under variable risk:
interactive influence of food quality and structural
complexity. Journal of Experimental Marine Biology and
Ecology 85: 269-285.
Schomer, N. S. and R. D. Drew 1982. An ecological
characterization of the lower Everglades, Florida Bay, and
the Florida Keys. U. S. Fish and Wildlife Service, Office
of Biological Services, Washington, D. C. 24688.
Seaman, W. Jr. and M. Collins 1983. Species profiles: life
histories and environmental requirements of coastal fishes
and invertebrates (South Florida) snook. U. S. Fish and
Wildlife Service. Slidell, Louisiana. 16 pages.
Sheridan, P. F. 1991. Comparative habitat utilization by
estuarine macrofauna within the mangrove ecosystem of
Rookery Bay, Florida. Bulletin of Marine Science 49(3).
Shulman, M. J. 1985. Recruitment of coral reef fishes:
effects of distribution of predators and shelter. Ecology
66(3): 1056-1066.
Smith III, T. J. 1987. Seed predation in relation to tree
dominance and distribution in mangrove forests. Ecology
68(2): 266-273.
Smith, III T. J. and N. C. Duke 1987. Physical determinants
of inter-estuary variation in mangrove species richness
around the tropical coastline of Australia. Journal of
Biogeography 14: 9-19.
Sogard, S. M. and K. W. Able 1991. A comparison of
eelgrass, sea lettuce macroalgae, and marsh creeks as
habitats for epibenthic fishes and crustaceans. Estuarine,
Coastal and Shelf Science 33: 501- 519.
Sogard, S. M., G. V. N. Powell, and J. G. Holmquist 1987.
Epibenthic fish communities on Florida Bay banks: relations
with physical parameters and seagrass cover. Marine Ecology
Progress Series 40:25-39.
Sogard, S. M., G. V. N. Powell, and J. G. Holmquist 1989a.
Spatial distribution and trends in abundance of fishes


Table 5-6. Species of predators associated with encounter rate
experiments. Observers in the boat or water recorded the following
predators taking, or in the vicinity of, the tethered fish.
Species
Predation
Events
Observed
Estimated
Abundance
In Vicinity
of Tethers
Estimated
Size
cm.
Gradient*
Where
Observed
Negaprion brevirostris
lemon shark
3
6
70-125
2,3
Carcharihinus leucas
bull shark
1
75
1
Strongulura notata
redfin needlefish
22
1000+
15-32
1,2,3,4
Centropomus undecimalis
snook
10+
45-75
2,3,4
Epinephelus itajara
jewfish
1
75
4
Caranx hippos
crevalle jack
2
30+
45-50
1,4
Lutjanus griseus
gray snapper
100+
15-40
1,2,3,4
Haemulon sciurus
blue-striped grunt
100+
15-40
4
Lepisosteus platyrhincus
Florida gar
16
45-55
1
Sphyraena barracuda
great barracuda
1
30+
15-70
1,2,3,4
Trionyx ferox
snapping turtle
1
45
1
Alligator mississippiensis
alligator
2
180-200
1
Butoroides striatus
green heron
1
1
*
Gradient: 1= Far/Upstream, 2=Mid/Upstream, 3=Midstream, 4=Downstream


Enclosure Nets
e
5
4
E 3
¡5
2
1
0
Jtatlons
Presence
Density
Up Mid & Down
= Misclassified


157
For smaller fishes, the implications of the model are
that upstream habitats are occupied by colonizing fishes.
Competition for food resources limits the abundances of
these fishes more than predation. Downstream, predation is
a more powerful force in the community structure. Thus,
upstream areas should be relatively advantageous for small
fishes in both freshwater streams and estuarine habitats.
Small fishes, however, were not more abundant upstream
in northeastern Florida Bay. Young-of-the-year estuarine
transient juveniles may represent the colonizers described
in the model. However, these fishes were absent in the
current study. Thus, there appears to be a missing
component of the fish community in the study area,
particularly in the upstream fish assemblages in the current
study.
Implications For Estuarine Fish Ecology:
The Nursery-ground Hypothesis
The absence of young-of-the-year juveniles of estuarine
transient fishes in the study area was also a significant
departure from the results one would have expected based on
widely accepted theories in estuarine ecology. In
northeastern Florida Bay, this condition was not unique to
the mangroves; both seagrass and mangrove habitats in
eastern and central Florida Bay also have low populations of
very young transients (Sogard et al. 1987, 1989as, Thayer et
al. 1987a).


135
Predation is one of the most complex of the species
interactions studied by ecologists. In comparing levels of
predation intensity among habitats, not only are relative
abundances of predators and prey important factors, but the
potential rates of predation must also be estimated
(Kitching 1983). These rates depend on behavior of predator
and prey and characteristics of the habitat. Some
behavioral and microhabitat aspects have been modeled by
observing individual components of predator/prey
interactions in the laboratory (e.g. Holling 1966; Barshaw
and Able 1990a). Other habitat related factors, however,
must be measured in the field, and are thus, more difficult
to determine.
The tethering technique, a useful field method, has
increasingly been used by ecologists studying the effects of
different habitat features on predator/prey interactions.
Briefly, the tethering technique involves affixing a line to
a prey organism so that evidence of predation can be
determined from its condition after a period of time. These
investigations are usually accompanied by complementary
laboratory studies or censuses of predators and prey. A
summary of tethering studies is presented in Table 5-1.
In this study, the tethering technique has been used to
compare predator encounter rates across a gradient of
salinity conditions. The question of interest is, are these
rates lower in the more variable upstream locations relative
to mid- and downstream where conditions remain more saline?


1E4
8000
6000
4000
2000
0
300
200
100
0
1500
1000
500
0
60
40
20
0
Abundance by Size Class
Lutjanus
griseus
Lutjanus
J 1 L
Haemulon
10 20 30 40 50 60 70
Total length (cm)


Table 2-3, continued
Group/
Residency
Family Species
General Locations
Total
Nets
Visual
Up-west
Nets Visual
Mid-west
Nets Visual
Down-
Nets
west
Visual
Up-east
Nets Visual
Mid-east
Nets Visual
Down-
Nets
east
Visual
Belonidae (needlefish)
LR/r
Strongylura notata
177
S3
196
407
223
478
195
95
78
219
90
601
959
1,853
LR/r
Strongylura timucu
1
1
Cyprinodonditae (killifish)
BF/r
Floridichthys carpi
1,792
676
5,011
1,294
2,183
1,095
2,127
1,473
1,056
227
849
1.475
13,018
6,240
BF/r
Lucania parva
964
2,469
924
617
1,779
1.821
1,576
3,310
389
78
4,605
3,079
10,237
11,374
BF/r
Cyprinodon variegatus
221
768
144
86
79
1
308
957
351
21
1,124
1,812
BF/r
Fund ulus grandis
309
252
99
15
97
7
65
31
147
76
793
305
BF/r
Fundulus confuentus
98
63
91
16
47
40
199
30
481
103
BF/r
Minia xenica
25
21
46
BF/r
Fundulus simiiis
2
1
10
13
BF/r
Lucania goodei
1
1
BF/r
Rivulus marmoratus
1
1
Poecillidae (livebearers)
BF/r
Poecilia iatipinna
1,463
2.820
4,395
1,254
1,006
2,440
257
1,535
2,981
149
898
544
11,000
8,742
WC/r
Gambusia sp.
26
467
325
937
293
2,921
31
148
583
210
649
449
1,907
5,132
WC/r
Beionesox belizanus
4
1
2
2
8
1
Atherinidae (sllverside)
WC/r
Atherinomorus stipes
397
450
630
31,260
7,078
25,187
1,010
89
4,352
2,848
58,491
11,042
120,750
WC/r
Atherinidae (genus unk)
900
14,581
887
272
1,681
2,369
0
20,690
WC/r
Menidia sp.
2,242
543
175
3,093
357
256
1,071
1,815
325
200
178
3,222
4,348
9,129
WC/r
Membras martinica
1
1
Syngnathidae (pipefish)
BF/r
Syngnathus scovelli
43
83
19
12
6
53
80
213
83
BF/r
Syngnathus floridae
1
8
11
20
BF/r
Micrognathus criniger
6
6
BF/r
Hippocampus erectus
2
2
4
BF/r
Hippocampus zosterae
1
2
3
continued


Table 2-1. List of stations used in sampling fish in mangroves
using enclosure nets and visual census surveys.
System
Gradient
General
Location
Enclosure Net
Stations
Number Name
Number
Visual Census Stations
Name
West
Upstream
Joe Bay
1
east
1
Snook Creek Pond 1
2
mid
2
Snook Creek Pond 2
3
west
3
Snook Creek Pond 3
4
west Joe Bay*
Midstream
Trout Cove
4
northeast
5
southeast Trout Cove
5
mid
6
Tern Key
6
southeast
7
Deer Key*
8
Duck Key
Downstream
Buttonwood
7
northeast
9
Buttonwood Point
Sound
8
mid
10
Whaleback Key
9
southwest
11
Unnamed Key
12
Key Largo Ranger Station
East
Upstream
Highway Creek
10
east
13
Shark Pond
11
island
14
Highway Creek Big Island
12
west
15
Critter Pond*
16
northeast Long Sound*
Midstream
Little Blackwater
13
east
17
northeast L. Blackwater Sd.*
Sound
14
mid
18
northwest L. Blackwater Sd.*
15
west
19
L. Blackwater Sd. island
20
south L. Blackwater Sd.
Downstream
Blackwater Sound
16
near Gilberts
21
trestle
17
mid
22
Gilberts
18
far
23
hydrostation
24
Bush Point
* Stations that were dropped from the analysis due to missing data as a result of numerous
poor visibility days primarily in 1990.
to
to


Figure 2-1. Northeastern Florida Bay study area with
sampling stations indicated.


126
were consumed, but only 2 in great abundance. In a second
strategy, used by Fundulus granis, Lutjanus griseus,
Strongylura notata, many items were also consumed, with
several (4 to 6) items ingested in substantial abundances.
In the third strategy, used by Sphyraena barracuda, fewer
items were consumed overall, with only one in significant
quantity.
Diet Variability
The MANOVA results for all species, as presented in
Table 4-2, indicate the degree to which location (i.e.
gradient, system) was a significant source of variation in
fish diets. For Floridichthys carpi and Eucinostomus
harengulus, gradient position was a significant source of
variation. In both species, more animal food (i.e.
copepods, ostracods, nematodes) was obtained downstream than
up- and midstream, while more algae was consumed upstream.
The only other influence of gradient occurred for Lutjanus
griseus, which consumed significantly more crabs upstream
than mid- or downstream. For Strongylura notata, Fundulus
grandis and Sphyraena barracuda, diets differed among the
locations, but not systematically along the salinity
variation gradient.
Seasonal influences on the fish diets are indicated by
the second MANOVA results presented in Table 4-3. For
Floridichthys carpi and Eucinostomus harengulus, algae was
more important in the spring. For Lutjanus griseus, crabs,
an important dietary item, were found abundantly in diets


small benthic, small water column, and larger fishes.
Abundances of larger fishes were consistently lower upstream
(0.15 fish/square meter (nr)), than mid- (0.65 fish/m ), or
downstream (0.55 fish/m2). Species of larger fishes
numbered fewer upstream (11), than midstream (15), and
downstream (22). Benthic and water column fish abundances
did not vary along the gradient. Temporally, fish
distribution was uncorrelated with salinity.
Development of mangrove habitat and submerged aquatic
vegetation (SAV) were reduced upstream. Fish diets shifted
to other foods upstream. Thus, where seasonal changes in
freshwater inflow were greater (i.e. upstream), species and
numbers of larger fishes were lower, possibly due to
salinity conditions, food availability and habitat
development.
To determine if lower salinity conditions alone led to
reduced predation, prey fishes were tethered along the
gradient. Predator encounter rates were not different over
the salinity range tested, but were 50% lower at the most
remote sites. This was perhaps a function of accessibility
of the sites to roving predators.
Water management strategies to increase mangrove
development and SAV are recommended research priorities.
However, severe ecotonal differences between Bay and ocean
waters, coupled with limited circulation and significant
viii


93
appear to be systematically distributed along the salinity
gradient. Although only one rough silverside (Membras
martinica), was collected in the current study, they were
abundant at central and western mangrove sites (Thayer et
al. 1987). Abundant Atherinomorus stipes were collected by
Thayer et al. (1987) at his most downstream locations in
central Florida Bay (Crane Key and Captains Key). This
corresponded to the very abundant collection of this species
at the downstream-most locations in northeastern Florida Bay
(Blackwater Sound and Buttonwood Sound). Since salinity
regime correlated with the distribution of Atherinomorus
stipes and Menidia spp., relative densities of the species
in the family Atherinidae may be indicative of salinity
conditions.
The density of large roving fish was dramatically lower
upstream than mid- or downstream. Among the large roving
fish species, however, the influence of salinity regime on
spatial distributions was mixed. Blue-striped grunts and
gray snapper were less abundant upstream, but redfin
needlefish were more abundant at the upstream/east location
than elsewhere. Thus, some species may be limited by the
conditions that occur upstream, while others tend to thrive
there.
Community patterns. Although no systematic pattern of
distribution occurred along the gradient for benthic and
water column forage fish, for the large roving fish, greater
numbers of species occurred downstream. In many other


8
northeastern portion, only one pass exists, a man-made cut
through Key Largo to the ocean side (Adam's Cut) (Figure 1-
1). On the northern boundary of the Bay are the Florida
Everglades.
The Florida Bay area is subject to an annual water
deficit with evaporation exceeding total rainfall (Tabb et
al. 1962). Annual rainfall in northeastern Florida Bay
ranges from 1600 mm on the mainland at Homestead to 1200 mm
on the south at Key Largo. The climate of subtropical south
Florida is characterized by a relatively long and severe dry
season (November through April) and a wet season (May
through October) (Schomer & Drew 1982).
Sea level becomes relatively high on an annual basis
from August to December reaching a maximum of about 15 cm
above the annual average in October (Ginsberg 1956, Provost
1973, Holmquist et al. 1989b). By late November or early
December, Bay level recedes to the annual average, which
probably accelerates the drainage of freshwater into the Bay
from the mainland. At this time, the zone of reduced
salinity may extend farther south and southeast into mid-
and downstream Florida Bay areas.
The major source of freshwater flow into Florida Bay is
from a series of approximately 20 creeks and Taylor River,
which carry surface water from the Taylor Slough/C-111
drainage area into the Bay. This system is smaller than the


Abundance by Size Class
o
c
o
~o
c
Z3
_Q
<
0 15 30 45 60 75 90 105 120 135 150
Total length (cm)


116


155
estuarine transient species, Lutjanus griseus and Haemulon
sciurus. Of the habitats within Florida Bay, these species
prefer mangroves during the day (Thayer et al. 1987a &
1987b). An important aspect of habitat use by these species
is their migration away from the mangrove shorelines at
night, to feed in nearby seagrass beds (Starck & Schroeder
1971, Sogard et al. 1989c). Thus, in both tidal and non-
tidal habitats, linkages between mangrove shorelines and
other habitats may be critical for diel behavior patterns.
Besides snappers and grunts, other species that may
similarly rely on both mangroves and seagrass beds include
snook, sheepshead, barracuda and nurse sharks. A common
life history pattern occurs among these fishes: recruitment
from offshore as post-larvae, settlement and growth in
inshore habitats, and movement back offshore or to deeper
water as they attain larger size classes (Starck & Schroeder
1971, DeSylva 1963, Jennings 1985). These fishes tend to
use seagrass beds when they are smaller and move to
mangroves when they attain larger juvenile sizes.
Thus, shallow water habitats including mangroves and
seagrass beds, may be linked to one another through such
behavioral and life history patterns (Odum et al. 1982,
Parrish 1989). Mangroves provide cover and food resources
that are very different from adjacent habitats dominated by
submerged aquatic vegetation (SAV). However, both types of
habitats appear to be necessary to support certain fish
species. Prime locations for supporting these species may


29
Definitions
As presented in Chapter 1, three categories of
residency are recognized by estuarine ecologists:
residents, transient juveniles, and occasional visitors. In
analyzing the community in northeastern Florida Bay, several
sources were consulted for life history information on
individual species to designate each by residency (e.g. Odum
& Heald 1972, Lee et al. 1980, Yanez-Arancibia 1980, Robins
et al. 1986). Without conducting specific gonad analysis to
determine maturity (e.g. Robertson & Duke 1990b),
unequivocal distinction between juveniles and adults in the
transient category were not possible. In addition, life
history information is sketchy for all species except for
certain killifish. Thus, these designations are approximate
and serve for discussion purposes only. Such designations
were not used in statistical analyses of the fish community.
For purposes of detailed analysis, all fish were
assigned to one of three groups of species based on size,
behavior and primary portion of the mangrove habitat
occupied during the day. Forage fish were considered those
species whose members were generally less than 15 cm in
size. Two groups of forage fish occupied different portions
of the mangrove habitat: benthic and water column. Benthic
forage fish live in close association with the substrate and
include such species as gobies, killifish and mojarras.
Water column forage fish are exclusively schooling fishes,
that occupy the upper water column habitats, including


Table 2-8. Correlations between fish density for each station (n=18) averaged
over the months and salinity, temporal standard deviation of salinity and water depth.
Density data were converted to logarithms (log x + 1) prior to calculations.
Significant (p<0.05) correlations are underlined. Benthic and Water
column forage fish were collected with enclosure nets. Large roving fish
were sampled using visual techniques.
Category
Salinity
correlation p-value
Temporal Standard
Deviation of Salinity
correlation p-value
Water Depth
correlation p-value
Benthic Forage
Fish
-0.08
0.7540
-0.02
0.9303
-0.12
0.6463
Water Column
Forage Fish
+0.41
0.0884
-0.38
0.1233
+0.48
0.0438
Large Roving
Fish
+0.56
0.0150
-0.54
0.0208
+0.54
0.0213
U1
t*


Table 2-9. Correlations between fish densities for each month (n=13) averaged
over the stations and salinity, water temperature and water depth.
Density data were converted to logarithms (log x + 1) prior to calculations.
Significant (p<0.05) correlations are underlined.
Species
Method
Salinity
correlation p-value
Water depth
correlation p-value
Water Temperature
correlation p-value
Floridichthys carpi
Nets
+0.15
0.6117
+0.26
0.3681
-0.50
0.0788
Lucania parva
Nets
-0.06
0.8420
-0.62
0.0237
-0.69
0.0084
Poecilia latipinna
Nets
+0.49
0.0888
+0.09
0.7756
-0.72
0.0056
Anchoa mitchelli
Nets
+0.04
0.8961
+0.27
0.3620
+0.24
0.4227
Menidia spp.
Nets
+0.18
0.5638
-0.33
0.2765
+0.51
0.0740
Atherinomorus stipes
Nets
-0.14
0.6449
+0.56
0.0470
-0.14
0.6259
Lutjanus griseus
Visual
+0.47
0.1039
-0.43
0.1439
-0.92
0.0001
Strongylura notata
Visual
-0.35
0.2382
+0.76
0.0026
-0.56
0.0482
Haemulon sciurus
Visual
-0.04
0.8931
+0.56
0.0475
+0.62
0.0233


2
they then may disperse to lower reaches as they grow larger
(Weinstein 1979, Rogers et al. 1984, Loneragan et al. 1990).
The prominence of transient juvenile fish and
crustaceans led to the application of the term "nursery-
ground" to many estuaries (Gunter 1961, McHugh 1967,
Weinstein 1979). A major role of freshwater discharge in
such systems may be to increase food availability for fishes
by transporting nutrients which stimulate primary production
and by increasing detrital transport and processing (Odum et
al. 1982). Freshwater inflow may also improve the chance
for survival of juvenile fish in estuaries by reducing
salinity levels below the limits tolerable by stenohaline
marine predators (Gunter 1961, 1967).
Browder and Moore (1981) offered a comprehensive
nursery ground hypothesis linking several of these concepts.
They split habitat factors into those that are relatively
stable (e.g. shoreline edge, bottom type) and those that are
movable (e.g. salinity, food resources). Favorable habitat
for particular juveniles consists of combinations of these
factors that promote growth. According to their theory, the
inflow of freshwater acts to position an area of favorable
moveable habitat relative to important stationary habitat.
Thus, for any estuary there is a rate of freshwater flow
sufficiently high to push the band of potentially favorable
moveable features beyond estuarine boundaries into open
waters, perhaps eliminating favorable habitat entirely.
Likewise, for every estuary, there is a rate of freshwater


Figure 2-13. Length-frequency histograms based on visual
census data. Adult size given whenever the information was
available from the literature for
a. Lutjanus griseus (gray snapper),
b. Lutjanus apodus (schoolmaster)
c. Haemulon sciurus (blue-striped grunt)
d. Archosargus probatocephalus (sheepshead).


110
benthic forage fish, however (Table 3-5). As a group, these
small fishes (less than 15 cm) did not appear to select
habitats based on the parameters measured in this study.
A significant regression was derived for density of
water column forage fish, but only the first principal
component was a significant source of variation (Table 3-5).
Thus, the greater mangrove canopy coverage and water depth,
the greater the abundance of water column forage fish.
In contrast to the other two fish groups, for the
density of large roving fish, all factors were significant
sources of variation except prop root density (Table 3-5).
Thus, sites with greater development of mangroves and
submerged aquatic vegetation (SAV), and with higher, less
variable salinity had greater densities of large roving
fish.
Of the 77 total species collected, 14 had significant
(p<.0001) regressions on the 6 factors (Table 3-6). Nine of
these species were benthic forage fish and 3 were water
column forage fish. The greatest amount of variation
explained was 38.8%. This value was derived for densities
of Opsanus beta, the Gulf toadfish, which was most abundant
at high salinity sites with greater development of both
mangrove and SAV. Poecilia latipinna (sailfin molly) was
found most abundantly where mangrove prop roots were more
dense and the width of the fringe was greater. One of the
most abundant fishes, Floridichthys carpi (gold-spotted
killifish), was more prevalent in shallow sites with more


ACKNOWLEDGEMENTS
I would like to extend my thanks to all my committee
members and my Department Chairman, Joseph Delfino. They
advised and supported me throughout my field work and
writing. Clay Montague introduced me to science and Florida
Bay, and encouraged me to pursue my interests. Carole
Mclvor gave me guidance, scientific insight, and maintained
faith in my abilities under all circumstances. I am
extremely grateful to Bill Seaman for his efforts in
obtaining Sea Grant support for a substantial portion of the
project. In his Wetlands Ecology class, Ronnie Best
introduced me to working in mud and swamps, features which
later became a major part of my life. He also allowed me to
live in the Center's Winnebago for two years, permitting me
to operate on an intense and flexible schedule in Key Largo.
Frank Nordlie expressed constant interest and encouraged me
in my work. My visits with Nick Funicelli always included
valuable personal and professional insights.
I am extremely grateful for the efforts of Dan Haunert
of the South Florida Water Management District. He believed
in the benefits of this research to Florida Bay and
aggressively oversaw the process of obtaining funding. In
ii


164
Monographs 8(3): 314-336.
Gunter, G. 1961. Some relations of estuarine organisms to
salinity. Limnology and Oceanography VI(2): 182-190.
Gunter, G. 1967. Some relationships of estuaries to the
fisheries of the Gulf of Mexico, in G. H. Lauff ed.
Estuaries, American Association for the Advancement of
Science, Washington, D. C. pages 621-638.
Harrington, R. W., and E. S. Harrington 1961. Food
selection among fishes invading a high subtropical salt
marsh: from onset of flooding through the progress of a
mosquito brood. Ecology 42(4): 646-666.
Hay, M. E., J. R. Pawlik, J. E. Duffy, and W. Fenical 1989.
Seaweed-herbivore-predator interactions: host-plant
specialization reduces predation on small herbivores.
Oecologia 81: 418-427.
Heald, R. J., W. E. Odum and D. C. Tabb 1974. Mangroves in
the estuarine food chain. in P. J. Gleason (Ed.)
Environments of south Florida: present and past. Memoir 2
Miami Geological Society, Miami, Florida 182-189.
Heck, Jr., K. L. and T. A. Thoman 1981. Experiments on
predator-prey interactions in vegetated aquatic habitats.
Journal of Experimental Marine Biology and Ecology 53: 125-
134.
Heck, Jr., K. L. and K. A. Wilson 1987. Predation rates on
decapod crustaceans in latitudinally separated seagrass
communities: a study of spatial and temporal variation using
tethering techniques. Journal of Experimental Marine
Biology and Ecology 107: 87-100.
Hellawell, J. M. and R. Abel 1971. A rapid volumetric
method for the analysis of the food of fishes. Journal of
Fish Biology 3:29-37
Herrnkind, W. F. and M. J. Butler IV 1986. Factors
regulating postlarval settlement and juvenile microhabitat
use by spiny lobsters Panulirus argus. Marine Ecology-
Progress Series 34: 23-20.
Hettler, Jr., W. F. 1989. Food habits of juveniles of
spotted seatrout and gray snapper in western Florida Bay.
Bulletin of Marine Science 44(1): 155-162.
Holling, C. S. 1966. The functional response of
invertebrate predator to prey density. Mem. Entorno. Soc.
Canada, No. 48, 86 p.
Holmquist, J. G., G. V. N. Powell and S. M. Sogard 1989a.


Table 2-16. Correlations between density for each station (n=18) averaged over
all months and salinity, temporal standard deviation of salinity, and water depth.
Density data were converted to logarithms (log x + 1) prior to calculations.
Significant (p<0.05) correlations are underlined.
Species
Method
Salinity
Temporal
Deviation
Standard
of Salinity
Water
Depth
correlation
p-value
correlation
p-value
correlation
p-value
Floridichthys carpi
Nets
+0.07
0.7655
-0.12
0.6446
-0.60
0.0079
Lucania parva
Nets
+0.13
0.6096
-0.44
0.0665
+0.62
0.0064
Poecilia latipinna
Nets
+0.40
0.0935
-0.28
0.2429
-0.20
0.4165
Anchoa mitchelli
Nets
-0.03
0.9127
+0.14
0.5879
+0.05
0.8530
Menidia spp.
Nets
-0.52
0.0253
+0.69
0.0013
-0.01
0.9634
Atherinomorus stipes
Nets
+0.60
0.0087
-0.72
0.0007
+0.45
0.0626
Lutjanus griseus
Visual
+0.58
0.0108
-0.58
0.0123
+0.53
0.0226
Strongylura notata
Visual
-0.11
0.6595
-0.06
0.8234
+0.22
0.3765
Haemulon sciurus
Visual
+0.37
0.1300
-0.38
0.1226
+0.26
0.2942


78
stations had mixtures of both upstream and downstream
species excepting the Mayan cichlid, which was only common
upstream.
Residency. Abundance and numbers of species by
residency category varied among the general locations (Table
2-18). All 33 benthic forage fish species sampled were
permanent estuarine residents. Among the water column
forage fish, seven of the nine species observed were
residents, while two (Clupeids) were occasional visitors.
Large roving fish were represented in all three residency
categories. Six species of large roving fish were permanent
estuarine residents including needlefish, catfish, bull
sharks, and stingrays. The vast majority of large roving
fish were transient juveniles, however.
Frequency distribution by size varied for 6 transient
species of large roving fish (Figures 2-12 and 2-13).
Clearly, the mangrove habitats of northeastern Florida Bay
were nurseries for Sphyraena barracuda most of which
occurred in juvenile sizes (Figure 2-12). Snook, however,
did not occur in juvenile size classes. Although adult-sized
Lutjanus griseus appeared to share the habitat with larger-
sized juveniles, since no gray snapper sampled was smaller
than 7.5 cm, one can assume that young-of-the-year juveniles
occur outside the mangrove habitats sampled in this study.
Habitat use patterns similar to Lutjanus griseus were
observed for Lutjanus apodus, Haemulon sciurus, and


a.
Study Ana\ 0
\
US Highway 41
C-111 Canal
Study Area
L12 km|


56
patterns in water temperature were significantly correlated
with densities of several species. Lucania parva, Poecilia
latipinna, Lutjanus griseus, and Strongylura notata were
less abundant when the water temperatures were higher. In
contrast, greater abundances of Haemulon sciurus were
observed in warmer months. Periods of higher water levels
in the study area (e.g. late fall) corresponded to periods
when greater densities of Atherinomorus stipes, Strongylura
notata, and Haemulon sciurus were collected. In contrast,
Lucania parva was in greater abundance during low water
periods.
Spatial Patterns in Density by Species
Analyses of variance. Density patterns varied for the
top three species of benthic forage fish (Figure 2-7).
Results of repeated measures analyses of variance also
differed among these species (Tables 2-10 and 2-11).
Poecilia latipinna was more abundant at the midstream
locations, particularly Trout Cove (mid/west).
Distributions of Floridichthys carpi and Lucania parva were
not significantly influenced by gradient position or system.
However, Floridichthys carpi was more abundant at Trout
Cove than all other locations and Lucania parva was most
abundant at Blackwater Sound (mid/east).
The top three water column forage fish species differed
in spatial distribution (Figure 2-8). Repeated measures
ANOVA results also varied among these species (Table 2-12
and 2-13). Distribution of the silversides differed


89
dangerous predators may preclude visual sampling in certain
areas. Sites in deep creeks were abandoned based on sitings
of large sharks and alligators.
Combined methods Using two methods to complement each
other in the same habitats and regions had great advantages.
The two methods increased the range of mangrove shorelines
that could be surveyed. However, when using two methods,
interpretation problems can arise when neither is 100%
efficient. If the results do not agree, it is difficult to
determine if the discrepancies are due to differences in the
fish sampled or due to differences in the efficiency of the
methods. In this study, the two methods targeted different
size groups of fishes. Overall, benefits of using the two
methods clearly out-weighed the disadvantages.
Fish and Salinity
Salinity. As confirmed in this study, salinity
conditions in the area vary from year-to-year (Ginsburg
1956). The east and west systems differed greatly during
the drought conditions that prevailed during the main study
year. They were very similar, however, during the pilot
study year when rainfall was locally more plentiful. Under
low rainfall periods, the C-lll Canal may effectively block
most freshwater from flowing into the western system by
routing it towards the east. During high flow periods,
however, when more freshwater is available for distribution,
the east and west systems appear to have similar salinity
patterns.


90
Historically, such annual differences due to local
rainfall variations may have been more moderate. Before
freshwater wetlands were extensively drained, freshwater
probably gradually seeped into the study area from the
greater Taylor Slough drainage basin, resulting in more
dispersed distribution patterns and prolonged periods of
lower salinity levels.
Temporal Patterns. The first hypothesis proved to be
incorrect: none of the temporal patterns in density of any
fish group or species examined was attributable to changes
in salinity. In the study area, therefore, the fish do not
seem to react to salinity changes by short-term movement in
and out of the general locations (regions approximately 12
km in area). However, patterns for all species collected
were not individually analyzed. Thus, some short-term
relationships may be identifiable on further examination of
the data.
Temporal patterns were related to temperature, however.
For benthic forage and large roving fish, as temperature
increased, density decreased. Similarly, Thayer et al.
(1987) and Tabb et al. (1962) also found greater densities
in western and central Florida Bay in the late fall and
winter when temperatures were cooler. Temporal patterns for
the current study are not typical for estuarine fish
populations; usually peaks occur when freshwater inflow is
greatest (Gunter 1967, Weinstein 1979, Yanez-Arancibia et
al. 1980, Rogers et al. 1984, Stoner 1986, Flores-Verdugo


Figure 1-1. Regional maps of the study area showing
features of the upstream drainage basins and Florida Bay.
a. Boundaries of drainage basins and tributaries to
northeastern Florida Bay. (Source: Schomer & Drew 1982).
b. Florida Bay showing the extensive mudbank system (stiple
pattern). Arrows indicated passes to the Atlantic Ocean and
Barnes Sound. (Source: Holmguist et al. 1989b)


Figure 2
habitats
indicate
location
I
CD
-t1
CD
E
CD
i_
O
D
CT
cn
(D
Q_
_C
CO
L_
Benthic Forage Fish
Density & Salinity vs. Month
Enclosure Nets
West
East
T>
Q
-i
H-
(/)
X)
0)
ir
o
c
in
a
3
Q.
May89
May90 Moy89
May90
3. Density of benthic forage fish collected with enclosure nets in mangrove
(histogram) and corresponding salinity measurements (lines). Error bars
standard deviation among the three enclosure nets deployed in each general
fO


Little, M. C., P. J. Reay, S. J. Grove, 1988. The fish
community of an East African mangrove creek. Journal of
Fish Biology 32:729-747.
Livingston, R. J. 1982. Trophic organization of fishes in a
coastal seagrass system. Marine Ecology Progress Series 7:
1-12.
Loneragan, N. R. and I. C. Potter 1990. Factors influencing
community structure and distribution of different life-cycle
categories of fishes in shallow waters of a large Australian
estuary. Marine Biology 106: 25-37.
Luckhurst, B. E. and K. Luckhurst 1978. Analysis of the
influence of substrate variables on coral reef fish
communities. Marine Biology 49: 317-323.
McHugh, J. L. 1967. Estuarine nekton, in G. H. Lauff (Ed.)
Estuaries. American Association for the Advancement of
Science, Washington, D. C. pages 581-620.
Mclvor, C. C. and W. E. Odum 1988. Food, predation risk,
and microhabitat selection in a marsh fish assemblage.
Ecology 69(5): 1341-1351.
McNeely, D. L. 1987. Niche relations within an Ozark
cyprinid assemblage. Environmental Biology of Fishes
18(3):195-208.
Miller, J. M. and M. L. Dunn 1980. Feeding strategies and
patterns of movement in juvenile estuarine fishes. in
Estuarine Perspectives. Academic Press, Inc. New York, New
York. 437-448 pages.
Montague, C. L., R. D. Bartelson, and J. A. Ley 1989.
Assessment of benthic communities along salinity gradients
in northeast Florida Bay, University of Florida,
Gainesville, for the South Florida Water Management
District, West Palm Beach, Florida.
Morton, R. M. 1990. Community structure, density, and
standing crop of fishes in a subtropical Australian mangrove
area. Marine Biology 105: 385-394.
Moser, M. L., L. R. Gerry 1989. Differential effects of
salinity changes on two estuarine fishes, Leisostomus
xanthurus and Micropogonius undulatus. Estuaries 12(1):35-
41.


Table 2-2. Efficiency test results for enclosure nets and visual census
sampling obtained by mark-recapture tests.
Method
Number
of Trials
Size Classes
Total Length
(centimeters)
Number
of Fish
Tagged
Mean
Percent
Efficiency
Standard Deviation
Percent Efficiency
Enclosure net
18
2.5 to 7.5
467
36
38
18
7.5 to 15
25
68
31
1
15 to 25
6
100
Not applicable
Visual census
3
4 to 7
31
29
2
3
18 to 25
22
86
12


LITERATURE CITED
Afifi, A. A. and V. Clark 1984. Computer-aided multivariate
analysis. Wadsworth Inc., Belmont, California. 458 pages.
Aronson, R. B. 1988. Palatability of five Caribbean
ophiuroids. Bulletin of Marine Science 43(1): 93-97.
Aronson, R. B. 1989. Brittlestar beds: low-predation
anachronisms in the British Isles. Ecology 70(4): 856-865.
Austin, H. M. 1971. A survey of the ichthyofauna of the
mangroves of western Puerto Rico during December, 1967-
August, 1968. Caribbean Journal of Science 11(1-2): 27-39.
Baelde, P. 1990. Differences in the structures of fish
assemblages in Thalassia testudinum beds in Guadeloupe,
French West Indies, and their ecological significance.
Marine Biology 105: 163-173.
Barshaw, D. E. 1990. Tethering as a technique for assessing
predation rates in different habitats: an evaluation using
juvenile lobsters Homarus americanus. Fishery Bulletin 88:
415-417.
Barshaw, D. E. and K. W. Able 1990. Deep burial as a refuge
for lady crabs Ovalipes ocellatus: comparisons with blue
crabs Callinectes sapidus. Marine Ecology Progress Series
66: 75-79.
Bell, J. D., D. A. Pollard, J. J. Burchmore, B. C. Pease,
and M. J. Middleton 1984. Structure and function of a fish
community in a temperate tidal mangrove creek in Botany Bay,
New South Wales. Australian Journal of Marine and Freshwater
Resources 35:33-46.
Beumer, J. P. 1978. Feeding ecology of four fishes from a
mangrove creek in north Queensland, Australia. Journal of
Fish Biology 12: 475-490.
Blaber, S. J. M. and T. G. Blaber 1980. Factors affecting
the distribution of juvenile estuarine and inshore fish.
Journal of Fish Biology 17: 143-162
161


96
Submerged aquatic vegetation (SAV), such as algae and
seagrasses, may provide cover for newly settled forms of
estuarine transient juveniles at a scale compatible with
their size. As the fishes grow, however, the SAV becomes
less likely to provide adequate cover for later stage
juveniles, and they may seek larger forms of structure for
shelter. A habitat expansion of this type was identified
for juvenile gray snapper by Starck & Schroeder (1971).
While smaller snappers dwell in seagrass beds, larger
juvenile snappers congregate near mangroves and other brush
during the day and return to feed in seagrass beds at night.
The smallest gray snapper individuals found in mangroves in
the current study were 7.5 cm, the same size indicated by
Starck & Schroeder (1971) at which snappers expand their
habitat use.
Seagrass beds are generally poorly developed in
northeastern Florida Bay (Zieman et al. 1989). Abundance of
SAV is temporally variable and SAV is often absent
altogether upstream (Montague et al. 1988). Without SAV,
young fishes may not have an adequate intermediate habitat
between the planktonic and mangrove stages in which to
settle. In addition, if larger juveniles (over 7.5 cm) do
wander upstream, they may find inadequate food resources;
many of the benthic invertebrates that they consume live
epiphytically on SAV and may not occur in adequate abundance
levels without SAV (Montague et al. 1989). Thus, lack of
SAV may result in both reduced fish recruitment and growth


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EAGUB5JH4_9X0B4W INGEST_TIME 2015-02-17T20:26:11Z PACKAGE AA00028768_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


91
1990). Judging from the salinity data, this would have been
late summer for northeastern Florida Bay in the study year.
One possible explanation for this unusual condition may
be that in the summer, high temperatures combined with low
circulation to create a stressful environment for fish in
the Bay (Moyle & Cech 1988). This is supported by reports
by fishermen of fish kills in Florida Bay during those hot
summer months (M. Robblee, Everglades National Park,
personal communication).
Additionally, the density of the major component of the
large roving fish group, Lutjanus griseus, probably
accounted for much of the temperature related trend in large
roving fish densities overall. The larger individuals of
this species, migrate offshore in the summer, when spawning
occurs, and return in the winter (Starck & Schroeder 1971,
Rutherford et al. 1989). This migration may account for the
reduced densities of large roving fish in the summer.
Spatial patterns. The density of fish decreases from
west to east in Florida Bay (Sogard et al. 1987, 1989a).
This trend appears to continue into the northeastern Bay
(this study, Funicelli et al. 1986). Using an almost
identical method of sampling (enclosure nets), the mean fish
density found in western and central Florida Bay mangroves
by Thayer et al. (1987) was 8.0 fish m-2, compared to 3.3
fish m 2 found with nets for the northeastern Florida Bay
area.


CHAPTER 3
FISH COMMUNITIES IN
FLORIDA BAY MANGROVE SHORELINE HABITATS:
RELATIONS WITH PHYSICAL PARAMETERS AND COVER
Specific features of mangrove habitats may contribute
to the structure of fish communities in predictable
patterns. In the only study to specifically address this
question to date, limited support for this concept was
found: certain species seemed to prefer mangrove habitats to
more open sites (Sheridan 1991). Other investigators have
identified patterns between fish community structure and
development of vegetative structure in seagrass beds (e.g.
Sogard et al. 1987, Thayer & Chester 1989), kelp beds
(Ebeling & Laur 1985), and littoral zone plants in ponds
(Werner et al. 1983). In addition, the "rugosity" and
vertical structure of coral reefs have also been identified
as factors in increased densities of some species of fish
(Luckhurst & Luckhurst 1978).
One role of structure in aquatic habitats is to protect
vulnerable prey fishes from predators (Werner et al. 1983,
Ebeling & Laur 1985). However, since food resources can
become exhausted or be of lower quality in vegetated
habitats, the safest refuge is not always the location that
98


94
estuarine studies greater numbers of species have been found
downstream (Weinstein 1979, Yanez-Arancibia 1980, Rogers et
al. 1984, Sogard et al. 1987 & 1989b, Thayer & Chester 1989,
Lonaragen et al. 1990). The species usually responsible for
the greater downstream richness are adult members of the
marine-visitor group of fishes.
Residency. Overall, the dominance over the entire
study area by permanent residents (91% of abundance) was
unusual even for tropical estuarine systems (Yanez-Arancibia
et al. 1980, Davis 1988, Morton 1990). Only three species
(13 individuals, all adults) were members of the reef
community (Acanthurus chirurgus, Aluterus scriptus,
Sparisoma radians) (Jaap 1984). The islands of the Florida
Keys may inhibit connection of the fish community in Florida
Bay with that of the extensive reef tract adjacent to
Florida Bay (Sogard et al. 1987). The mudbanks in the
central and western Bay may further inhibit travel into
northeastern Florida Bay from the Gulf of Mexico. Perhaps
more significantly, the moderating influence of the
thermally stable Gulfstream water masses do not enter the
Bay. When temperatures reach extreme low (or high) levels,
those reef species that have entered Florida Bay may be
forced to migrate or be killed. For example, a doctorfish
(Acanthurus chirurgus) that was observed every month at a
downstream/west station from May through December,
disappeared once temperatures began to drop.


Table 5-4. Analysis of sixteen predator encounter rate tests.
Data used are percentage of fish missing (arcsine-transformed)
after being tethered for three hours adjacent to mangrove edges.
ANOVA was used to test the hypothesis that gradient and system
were significant sources of variation. Specific F-tests to contrast
means for each pair of gradient positions were used to make multiple
comparisons.
Source of
Variation
F-value
p-value
df*
Contrast
Summary**
MODEL
4.06
0.0332
7/8
Gradient
6.94
0.0129
3/8
1 < 2,3 & 4
System
1.12
0.3206
1/8
No differences
Gradient
2.22
0.1634
3/8
No differences
X
System
Model degrees of freedom / Error degrees of freedom
* *
Gradient: 1 = Far/upstream; 2 = Mid/upstream; 3 = Midstream; 4 = Downstream


Table 2-3, continued
Group/
Residency
Family Species
General Locations
Total
Nets
Visual
Up-west
Nets Visual
Mid-west
Nets Visual
Down-
Nets
west
Visual
Up-east
Nets Visual
Mid-east
Nets Visual
Down-
Nets
east
Visual
Centropomidae (snook)
LFVo
Centropomus undecimalis
2
1
2
25
1
3
24
5
53
Echeneididlae (remoras)
LR/o
Echeneis naucrates
1
1
Carangidae (Jack)
LR/o
Trachinotus goodei
1
1
LR/o
Naucrates ductor
2
2
LR/o
Carangidae (sp unk)
4
4
LR/o
Caranx hippos
1
6
1
1
2
11
LR/o
Carangidae (juv.)
2
2
LR/o
Trachinotus falcatus
1
1
Lutjanidae (snapper)
LR/t
Lutjanus jocu
23
12
12
27
74
LR/ta
Lutjanus griseus
13
74
1
4,737
41
4,756
10
11
2
1,157
29
7,726
96
18,461
LR/t
Lutjanus apodus
1
209
1
129
49
69
2
456
Gerreidae (mojarras)
BF/r
Eucinostomus sp
4
2
13
19
BF/r
Eucino8tomu8 harengulus
349
32
212
128
67
446
396
135
505
458
68
882
1,597
2,081
BF/r
Eug erres plumieri
427
302
37
2
88
212
152
12
704
528
BF/r
Eucinostomus guia
2
1
241
1,289
104
1,739
1
30
184
33
2,350
410
5,564
BF/r
Gerres cinereus
39
7
26
34
164
18
3
116
143
52
900
266
1,236
Haemulidae (grunts)
LR/t
Haemuiidae (sp unk)
1
1
LR/t
Haemulon parrai
26
102
11
139
LR/ta
Haemuion sciurus
16
1,112
56
1,584
2,768
Sparldae (porgies)
LR/o
Lagodon rhomboides
2
20
6
28
LR/o
Archosargus rhomboidalis
52
52
LR/ta
Archosargus probatocephalus
41
114
8
163
continued


14
by the variations in rainfall and subsequent freshwater
flowing south through the tributaries, and variations in
wind speed and direction. The rate and degree of salinity
change are relatively unpredictable and can be rapid (hours)
or slow (days) depending on changes in the weather.
Fish Community Sampling Design
To monitor fish community changes across the dynamic
salinity gradient in northeastern Florida Bay, a balanced
two-way analysis of variance (ANOVA) design was used, with
two systems, each composed of three salinity regimes (Figure
1-2). Generally, upstream locations included one of the
creeks which carries freshwater from the Taylor Slough/C-111
Basin, an interior bay downstream from the creek but still
measurably affected by freshwater inflow, and an outer bay
much less affected by freshwater inflow but more by marine
influences. Specifically, the locations were as follows:
upstream sites were located in Highway Creek and Long Sound
in the eastern system and Snook Creek and Joe Bay in the
western system (Figure 1-2) ; midstream sites were located
in Little Blackwater Sound in the eastern system and the
Trout Cove area in the western system; downstream sites
were located in Blackwater Sound in the eastern system and
Buttonwood Sound in the western system.


28
To conduct the census, a snorkeler approached the
flagged edge and remained stationary under each flag for 30
seconds, an adeguate period for recording observable fish.
On underwater data sheets, they recorded the species,
numbers and estimated sizes of fish observed. The census
surveys were conducted by myself and one assistant. Each
census consisted of four complete swims of each transect.
Efficiency tests. Efficiency tests were conducted for
the visual census technique using a mark-recapture method.
Large fish (18-25 cm) were caught using hook and line and
smaller fish (less than 18 cm) were collected with minnow
traps. Tags made of plastic waterproof tape in various
colors and labeled with a unique code were used. Fishing
line was securely fastened to the tag and the line was sewn
with a small sewing needle through the flesh just under a
fish's dorsal fin (for gray snapper and larger fish) or
through the lower jaw (for smaller fish).
To conduct a test, block nets were used to enclose an
area of mangrove shoreline of adequate size to accommodate
at least two snorkeling stations. Tagged fish (six to nine
larger fish and ten smaller fish) were placed inside the
enclosure for several hours to allow them to acclimatize to
the habitat. Snorkelers then carefully entered the enclosed
areas and conducted a visual census of the site by recording
the species, tag color and code of each fish they observed.
Three such tests were conducted in the summer of 1990, at
three separate stations.


Table 2-11. Repeated measures analysis of variance with density of benthic
forage fish as the dependent variable and general location as the independent
variables. Samples were taken within mangrove habitats using enclosure nets.
Densities were transformed to logarithms prior to calculations.
Floridichthys
carpi
Lucania parva
Poecilia latipinna
Source
df*
F
P
df*
F
P
df*
F p
Among general locations
5/12
16.50
0.0001
5/12
7.67
0.0019
5/12
4.26 0.0184
Among months
10/120
7.43
0.0001
10/120
7.70
0.0001
10/120
5.73 0.0002
Month X general location
50/120
2.59
0.0001
50/120
1.68
0.0314
50/120
1.72 0.0426
Multiple comparisons
among means for
all months
General
location
* *
Sign.
gr.
than
General
location
General
location
* *
Sign.
gr.
than
General
location
General
location
* *
Sign. General
gr. location
than
2
>
All
6
>
All
2
> All
3 & 4
>
5 & 6
(others
inter
mediate)
* source degrees of freedom / error degrees of freedom
** General locations:
1 = Joe Bay, upstream/west 4 = Highway Creek, upstream/east
2 = Trout Cove, midstream/west 5 = Little Blackwater, midstream/east
3 = Buttonwood Sound, downstream/west 6 = Blackwater Sound, downstream/east
cr>
o


Table 3-7. Summary of the qualitative influence of the original variables
on density of fish by group and species. If the value of the given variable
is increased, density will either increase (+), decrease (-), or there
will be no effect (0).
Category/
Species
Tree
height
SAV
Water
depth
Prop
roots
Tree
cover
Salinity
mean
Salinity
variation
Fringe
width
Percent
correlates
Benthic forage
0
0
0
0
0
0
0
0
0
Water column forage
0
0
+
0
+
0
0
0
25
Large roving
+
+
0
0
+
+

+
75
Floridichthys carpi
0
0
_
0
.
0
0
0
25
Lucania parva
+
+
+
0
+
+
-
-
88
Poecilia latipinna
0
0
0
+
0
0
0
+
25
Gambusia sp.
+
0
0
0
0
+
-
0
38
Eugerres plumieri
0
0
0
0
0
_
+
0
25
Gobiosoma robustum
0
0
0
0
0
-
+
0
25
Lophogobius cyprinoides
0
0
0
0
0
_
+
0
25
Microgobius guio sus
0
0
0
0
0
-
+
-
38
Opsanus beta
+
+
+
0
+
+
_
0
75
Anchoa mitchelli
0
-
0
0
0
0
0
0
13
Atherinomorus stipes
0
+
0
0
0
+
_
0
38
Menidia sp.
0
0
0
0
0
-
+
0
25
Lutjanus griseus
+
+
+
0
0
+

+
75
Haemulon sciurus
+
0
+
0
0
+
-
0
50
Trends among species:
% plus
35
29
29
6
24
41
29
18
% minus
0
6
6
0
6
29
41
12
% zeros
65
65
65
94
71
29
29
71


Table 2-14. Repeated measures analysis of variance with density of large
roving fish as dependent variables and gradient and system as independent
variables. Samples taken by visual census methods along mangrove shoreline
habitats. Densities transformed to logarithms prior to calculations.
Haemulon sclurus
Lutjanus grlseus
Strongylura notata
Source
df*
F
P
df*
F
P
df*
F
P
Between Stations:
Among gradient positions
2/12
2.36
0.1367
2/12
5.67
0.0185
2/12
3.52
0.0626
Among systems
1/12
0.13
0.7278
1/12
0.20
0.6654
1/12
4.24
0.0618
Gradient X System
2/12
0.04
0.9650
2/12
0.48
0.6317
2/12
3.32
0.0714
Within Stations:
Among seasons
4/48
0.29
0.8794
4/48
9.22
0.0001
4/48
2.97
0.0416
Season X Gradient
8/48
0.73
0.6659
8/48
2.64
0.0309
8/48
2.95
0.0168
Season X System
4/48
1.97
0.1167
4/48
2.59
0.0667
4/48
0.76
0.5292
Season X System X Gradient
4/48
1.10
0.3794
4/48
1.51
0.2032
4/48
0.43
0.8603
Multiple comparisons among
means for all seasons:
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Gradient positions
No differences
Mid &
Down
>
Up
No differences
Systems
No differences
No differences
No differences
Source degrees of freedom / error degrees of freedom


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
INFLUENCE OF CHANGES IN FRESHWATER FLOW ON THE USE OF
MANGROVE PROP ROOT HABITAT BY FISHES
By Janet A. Ley
May 1992
Chairperson: Clay L. Montague
Cochairperson: Carole C. Mclvor
Major Department: Environmental Engineering Sciences
The hypothesis that seasonal changes in freshwater
inflow (indicated by salinity) influence habitat use by
fishes was tested in northeastern Florida Bay, extreme south
Florida. Fishes were sampled monthly for 13 months using
visual censuses and enclosure nets.
Of the 305,589 individuals observed, 91% were estuarine
residents, numerically dominated by engraulids, atherinids
and cyprinodontids. Occasional marine and freshwater
visitors comprised 2% of the individuals, and estuarine
transients, 8%. No young-of-the-year estuarine transients
were observed.
Salinity ranged between 0.0 to 58 parts per thousand
(ppt) upstream, 19.5 to 54 ppt midstream, and 30 to 50 ppt
downstream. The 77 species were grouped for analysis:
vii


CHAPTER 5
PREDATION RATES ON SMALL BENTHIC FISH
ACROSS A SALINITY GRADIENT
The value of estuaries as nursery areas for fishes that
have wider distributions as adults has been confirmed in
many studies (e.g. Gunter 1938, Reid 1954, Carter et al.
1973, Blaber & Blaber 1980, Yanez-Arancibia et al. 1980,
Bell et al. 1984, Blaber et al. 1985, Blaber et al. 1989,
Pinto 1987, Robertson and Duke 1987). An influx of
juveniles to an estuary usually coincides with the season of
highest freshwater discharge, when salinity levels drop. As
juveniles develop in the estuary, they tend to migrate from
fresher upstream areas to more saline downstream habitats
(Weinstein 1979, Rogers et al. 1984). Based on observations
such as these, one of the paradigms of estuarine ecology has
developed: that estuarine salinity conditions contribute to
the survival of juvenile fish because stenohaline marine
predators are precluded from entering portions of the
estuary having lower, more variable conditions of salinity
(Gunter 1961, Austin 1971, Browder & Moore 1981, Odum et al.
1982). Despite its widespread acceptance, this hypothesis
has not been tested before now.
134


Table 2-3, continued
Group/
Residency
Family Species
General Locations
Total
Nets
Visual
Up-west
Nets Visual
Mid-west
Nets Visual
Down-west
Nets Visual
Up-east
Nets Visual
Mid-east
Nets Visual
Down-
Nets
east
Visual
Acanthuridae (surgeonfish)
LR/o
Acanthurus chirurgus
9
9
Balistldae (leatherjackets)
LR/o
Aluterus scriptus
1
1
Soleidae (sole)
BF/r
Trinectes maculatus
7
1
28
36
Tetraodontidae (puffer)
LR/o
Sphoeroides spengleri
1
1
2
Diodontidae (spiny puffer)
LR/o
Chilomycterus schoepf
1
1
LR/o
Diodontidae (species unk)
1
14
15
Total
10,624
11,185
12,723
60.275
13,712 44,703
9,547
11,301
25,115
9,280
10,912
86,212
82,633
222,960
No. Species
37
27
22
27
32 33
32
26
36
18
35
31
59
51
No. Samples
36
35
36
44
36 50
36
33
36
23
36
48
216
233
* Groups: BF = Benthic forage fish
WC = Water column forage fish
LR = Large roving fish
/Residency: r
o
t
ta
resident
occasional visitors
estuarine transient juvenile
estuarine transient juvenile (also present as adults)


122
the entire area and occurred consistently over the study
period. For Lutjanus griseus, Sphyraena barracuda and
Fundulus grandis, all samples collected were analyzed. Due
to great abundances, for Floridichthys carpi, Strongylura
notata and Eucinostomus harengulus, smallest individuals (<
3 cm) were eliminated and subsamples were selected from the
remainder. Fish were not divided into size or age classes
for this analysis, however.
Gut analysis of the 6 selected species was contracted
out to Mote Marine Laboratory. For the laboratory analysis,
36 taxonomic levels (e.g. family) were selected for
consistency with other estuarine studies of fish food habits
(Brook 1977, Beumer 1978, Livingston 1982). The analysis
chosen follows the "points method" of Hynes (1950). For
each fish (n= 1,222), total length was recorded and the
stomach was extracted. In order to identify variability
among individuals, no stomachs were pooled. The material
found in each stomach was distributed to a standardized
level within a gridded petri dish (Hellawell & Abel 1971).
Each stomach containing food was considered to be uniformly
full (Starck & Schroeder 1971). Using a dissecting
microscope, percent composition of each food category for
each specimen was calculated by estimating the area covered
by the material on the grid (Neilson & Johnson 1983) .
Analysis
To determine frequency of occurrence, the number of
fish in which each food item occurred was listed as a


Figure 2-8. Mean density of fish by general location for the three most abundant
species in the water column forage fish group. Error bars illustrate the
magnitude of the standard deviation in density over all the months.


TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS ii
ABSTRACT viii
CHAPTERS
1. GENERAL INTRODUCTION 1
Estuarine Fish Ecology 1
Mangrove Fish Ecology 3
The Florida Bay Ecosystem 5
Problem Definition 9
Objectives 12
Study Area 13
Fish Community Sampling Design 14
2. FISH DENSITIES AND ASSEMBLAGE PATTERNS IN
MANGROVE HABITATS: COMPARISONS
ACROSS SALINITY GRADIENTS 17
Materials and Methods 19
Results 32
Discussion 84
3. FISH COMMUNITIES IN FLORIDA BAY MANGROVE
SHORELINE HABITATS: RELATIONS WITH PHYSICAL
PARAMETERS AND COVER 98
Materials and Methods 99
Results 109
Discussion 114
4. FOOD HABITS OF MANGROVE FISHES:
A COMPARISON ACROSS SALINITY GRADIENTS 120
Materials and Methods 121
Results 123
Discussion 129
5. PREDATOR ENCOUNTER RATES ON SMALL BENTHIC
FISH ACROSS A SALINITY GRADIENT 134
Materials and Methods 137
Results 143
Discussion 149
v


Fish per square meter
Density by General Location
Top 3 Large Roving Fish Species
Census East
- Up


159
densities of fish than was observed in the current study.
Higher, abruptly changing salinity conditions may somehow
inhibit the development of lush communities of submerged
aguatic vegetation that provide cover for small fish and
benthic invertebrates (Montague et al. 1989). Sustained
lower salinity periods may promote growth of lush seagrass
(Ruppia martima) or algal (Chara, Batophora) communities
(Tabb et al. 1961, Montague et al. 1989). During more
saline periods, less dense growth of Halodule wrightii
communities may develop, if any vegetation grows at all. As
salinity changes with seasons, these communities may
alternate.
Although the regular study was conducted during a
period of very low rainfall and can only serve to provide
information on the ecosystem under low freshwater inflow
conditions, the pilot study took place at a time of higher
rainfall and high freshwater inflow conditions (summer
1988) In the pilot study, from October 1988 through March
1989, in upstream habitats, extremely dense Ruppia and algal
communities were observed. With experimental traps and gill
nets, great numbers of fishes were collected including gray
snappers, jacks, catfish and cichlids. A small mangrove
island and its surrounding waters in the Joe Bay study area
were heavily used by white pelicans (fish eating birds).
However, during the regular study (May 1989 May 1990),
very little Ruppia was observed. The pelican island was
evidently never used by birds and because the traps no


24
and overhanging branches so that a person could walk up the
path carrying one end of the 30 m long net.
The same site was sampled repeatedly throughout the
study unless "stress" was observed in submerged vegetation.
For example, sites with clay sediments supporting seagrasses
had suffered some visible damage (e.g. grass trampling) from
the sampling procedure by the fourth month at three
stations. To maintain consistency in types of habitat
encompassed by the nets, at these three sites, one new path
was cut so that an unimpacted site could be sampled adjacent
to the old one. These minor site changes were taken into
account in later calculations of net area sampled.
On the day of sampling, a 6.0 mm mesh nylon seine was
deployed by two people who carried it, scrolled around two
wooden dowels, to the mid-point between the two prepared
paths. Starting 10 m from the edge, they waded in opposite
directions parallel to the edge, unrolling the net, and then
walked toward the mangroves and up the paths. The dowel end
was pounded into the sediment at the landward end of the
path and the lead line was pressed down into the sediments
all around the bottom edge. The top edge of the net was
hung over several PVC poles to prevent fish from jumping
over the net. All three nets were set in similar fashion
(Figure 2-2).
Liguid rotenone was then applied within the enclosed
area to a final concentration of 5 mg L1. Fish that
immediately began to surface were collected using hand nets


Table 2-13. Repeated measures analysis of variance with density of
water column forage fish as the dependent variable and general location as the
independent variable. Samples were taken in mangrove habitats with enclosure
nets. Density data were transformed to logarithms for calculations.
Menidia spp.
Atherinomorus
stipes
Anchoa
mitchelli
Source
df*
F
P
df*
F
P
df*
F p
Among general locations
5/12
5.90
0.0056
5/12
51.83
0.0001
5/12
11.59 0.0003
Among months
10/120
3.05
0.0065
10/120
8.54
0.0001
10/120
4.73 0.0060
Month X general location
50/120
3.02
0.0001
50/120
3.68
0.0001
50/120
4.74 0.0001
Multiple comparisons
General
Sign.
General
General
Sign.
General
General
Sign. General
among means for
location
gr.
location
location
gr.
location
location
gr. location
all months
*
than
**
than
**
than
1
>
All
3
>
All
5
> All
except
4
6
>
All
others
* source degrees of freedom / error degrees of freedom
** General locations:
1 = Joe Bay, upstream/west
2 = Trout Cove, midstream/west
3 = Buttonwood Sound, downstream/west
4 = Highway Creek, upstream/east
5 = Little Blackwater, midstream/east
6 = Blackwater Sound, downstream/east


25
Figure 2-2. Enclosure net illustration indicating net
dimensions.


112
open mangrove tree canopy cover. Anchoa mitchelli was also
more abundant in open locations that were low in volume of
SAV.
For several species, salinity appeared to over-ride all
other variables in importance. Salinity accounted for over
10% of the variation in Gambusia sp. densities; this species
appeared to prefer higher salinity sites in the study area.
Salinity also explained from 10 to 25% of the variation in
densities of Menidia spp., Eugerres plumieri, Gobiosoma
robustum, Lophogobius cyprinoides, and Microgobius gulosus;
these species were more abundant where salinities were
lower. Of all the variables, salinity was a significant
source of variation for 11 of the 14 species (Table 3-7).
None of the other variables came close to this level of
apparent importance.
Seven benthic forage fish species that were abundantly
collected were not significantly correlated with the
factors. Among these were the 2 Fundulus species and
Cyprinodon variegatus as well as the 3 most abundant
mojarras: Eucinostomus gula, Eucinostomus harengulus, and
Gerres cinereus. These species are probably very flexible
in habitat selection.
Among the species of large roving fish in the study,
only the densities of gray snappers and blue-striped grunts
were significantly correlated with the variables measured.


125
Other items exploited to varying degrees by all 6 species
included isopods, shrimp, nematodes, eggs, fish parts, adult
terrestrial insects and algae.
Breadth of Diets
Excluding unidentifiable material and counting all prey
fish as a single food item, of the 6 species, Floridichthys
carpi and Eucinostomus harengulus foraged on the widest
variety of items (20 and 18 out of 24 total, respectively)
(Table 4-1). Only 2 items, however, were consumed in mean
guantities exceeding 4% of total diet: amphipods and algae.
Thus, although a wide variety of items were utilized by
these species, a degree of specialization was apparent.
In contrast, Fundulus granis, Lutjanus griseus, and
Strongylura notata, not only consumed a wide range of items
overall, but several items (5,6 & 4 respectively) were
consumed in mean quantities exceeding 4%. Isopods and fish
were major items commonly found in all 3 of these omnivorous
species.
Sphyraena barracuda was the most specialized of the
species analyzed, consuming mostly fish. For this species,
the benthic forage fishes in the family Cyprinodontidae were
consumed with particular frequency, even though the most
abundant fishes in the study area were Engraulids and
Atherinids (Chapter 2).
Thus, in the study area, among the 6 species, 3 feeding
strategies were evident. In one strategy, used by
Floridichthys carpi and Eucinostomus harengulus, many items


118
While these factors were important for the water column
forage fish group as a whole, other features of the habitat
seemed to segregate the individual species. The most
abundant species in this group, Anchoa mitchelli, was less
abundant at sites with greater volumes of SAV. This species
was extremely dense at only one location in the study area
(Little Blackwater Sound). Although not a variable
included in this analysis, turbidity may have been the more
attractive habitat feature for the bay anchovy at this site.
Of all the significant regressions for the abundant
species, the range of percentage variation in fish density
explained by habitat variables in this study (10.3-38.8%)
was only slightly lower than that found by Sogard et al.
(1987) in seagrass beds (24.8-42.7%). Thus, fish may select
particular habitats based on salinity and physical features
in Florida Bay but other factors (e.g. foraging
requirements, species interactions) are also undoubtedly
important.
The present findings do not differ greatly from long
term observations of fishes on small patches of coral, in
which habitat attributes other than overall size were of
little value in predicting the structure of fish assemblages
(Sale & Douglas 1984). Many species of reef fish may
therefore select habitats based on overall parameters (i.e.
large coral reefs vs. very small patches) and not detailed
features. The low magnitude of variance explained by the
habitat variables measured in seagrass beds and mangroves


99
vulnerable fishes choose (Werner et al. 1983, Schmitt &
Holbrook 1985). In addition, since recruitment in aquatic
systems is largely based on widely-dispersed larvae, the
occupation of particular sites may be based on chance
vacation of living space by a previous occupant and the
largely unpredictable occurrence of available recruits from
the plankton (Sale 1980, Sutherland 1980, Sale & Douglas
1984). Thus, the prediction of habitat use is a complex
problem involving recruitment, species interactions,
resource availability, and random influences.
If fish are not randomly distributed within the
mangrove shoreline habitats of northeastern Florida Bay, it
may be possible to identify features correlated with fish
densities. Density can be used as a quantitative
approximation of habitat quality (Sogard & Able 1991).
Thus, the objective of this portion of the study was to
analyze density data and habitat information in northeastern
Florida Bay to determine any habitat preferences among the
fish found in the mangrove shoreline.
Materials and Methods
For both the visual census and enclosure net methods,
fishes and salinity were monitored at each station
repeatedly over the period of May 1989, through May 1990
as described in Chapter 2. These stations were located
across a gradient from upstream near sources of freshwater
inflow to downstream (Chapter 2). A total of 328,960 fish


Robblee, M. B. 1987. The spatial distribution of the
nocturnal fish fauna of a tropical seagrass feeding ground.
Ph.D. Dissertation, University of Virginia, Charlottesville,
Virginia. 144 pages.
Robertson, A. I. and N. C. Duke 1987. Mangroves as nursery
sites: comparisons of the abundance and species compostion
of fish and crustaceans in mangroves and other nearshore
habitats in tropocal Australia. Marine Biology 96: 193-205.
Robertson, A. I. and N. C. Duke 1990a. Mangrove fish-
communities in tropical Queensland, Australia: spatial and
temporal patterns in densities, biomass and community
structure. Marine Biology 104: 369-379.
Robertson, A. I. and N. C. Duke 1990b. Recruitment, growth
and residence time of fishes in a tropical Australian
mangrove system. Estuarine, Coastal and Shelf Science 31:
723-743.
Robins, C. R., G. C. Ray, J. Douglass, 1986. A field guide
to Atlantic Coast fishes of North America. Houghton Mifflin
Company, Boston. 355 pages.
Rogers, T. E. Target, S. B. Van Sant 1984. Fish-nursery use
in Georgia salt marsh estuaries: the influence of springtime
freshwater conditions. Transactions of the American
Fisheries Society 113: 595-606.
Rozas, L. P. and M. W. LaSalle 1990. A comparison of diets
of Gulf killifish, Fundulus grandis Baird and Girard,
entering and leaving a Mississippi brackish marsh.
Estuaries 13(3):332-336.
Rozas, L. P. and W. E. Odum 1988. Occupation of submerged
aquatic vegetation by fishes: testing the roles of food and
refuge. Oecologia 77: 101-106.
Rutherford, E. S., J. T. Tilmant, E. B. Thue and T. W.
Schmidt, 1989. Fishery harvest and population dynamics of
spotted seatrout, Cynoscion nebulosus, in Florida Bay and
adjacent waters. Bulletin of Marine Science 44(1):108-125.
Sale, P. F. 1980. Assemblages of fish on patch reefs
predictable or unpredictable? Environmental Biology of
Fishes 5(3): 243-247.
Sale, P. F. and W. A. Douglas 1984. Temporal variability in
the community structure of fish on coral patch reefs and the
relation of community structure to reef structure. Ecology
65(2): 409-422.
Salini, J.P., S.J.M. Blaber and D. T. Brewer 1990. Diets of
piscivorous fishes in a tropical Australian estuary, with


3
flow so low that the band of favorable salinities retreats
upstream where stationary features may be unfavorable. The
ideal situation with regard to freshwater inflow is one that
maximizes the area of favorable habitat within the estuary
over the peak period of nursery use. This hypothesis seems
particularly applicable for analyses of fish ecology in
mangrove-dominated estuaries.
Mangrove Fish Ecology
In tropical and subtropical areas of the world,
mangroves are dominant shoreline features. Mangrove-derived
detritus forms a food base for fish occupying mangrove
ecosystems (Odum 1971). Mangrove shorelines may also
provide cover for fishes (Thayer et al. 1987a). However,
few studies have documented aspects of the direct use of
mangrove habitats by fishes probably because monitoring
fishes within the complex tangle of roots and branches is
extremely difficult. Efforts have only recently focused on
obtaining quantitative data on habitat use (i.e. Thayer et
al. 1987a, Sheridan 1991, Morton 1990, Robertson & Duke
1987). Strong linkages between mangroves and adjacent
habitats may exist. For example, diel habitat shifts occur
in both non-tidal (Thayer et al. 1987a) and tidal systems
(Morton 1990, Robertson & Duke 1987). Shifts from other
habitats to mangroves occur during the life history of some
species such as gray snapper (Lutjanus griseus) (Starck &
Schroeder 1971).


9
Shark River Slough, a separate system which extends from
Lake Okeechobee southward toward the Gulf of Mexico and
drains most of the Everglades.
These overall features contribute to several
environmental and biological patterns. Gradients in
environmental variables occur in Florida Bay, from southwest
to northeast. These gradients include amount of water
exchange, sediment depth, and seagrass standing crop (Zieman
et al. 1989). The area northeast of the central line of mud
banks is characterized by very restricted circulation and no
tidal influence (Schomer & Drew 1982). A thin sediment
veneer covers the basin bedrock in the northeast Bay,
deepening towards the southwest. In addition, seagrass
density and productivity decreases dramatically from
southwest to northeast (Zeiman et al. 1989).
Problem Definition
Water management decisions in the eastern Everglades
have potentially impacted Florida Bay through changes in the
timing and quantity of freshwater discharge. Under pre
drained conditions, in this area, surface freshwater moved
over grassy marl prairies that were seasonally flooded
(Schomer & Drew 1982). A complex network of streams,
bordered by mangroves and other shrubs carried freshwater
inflow to receiving waters downstream in a manner that was
presumably both gradual and dispersed.
Beginning in the early 1900s, construction began on an
extensive system of canals and ditches throughout much of


88
the visual method has increasingly been used, it has
possible further value for use in studies on behavior,
species interactions, and microhabitat use in mangrove
habitats.
Direct observational sampling also has disadvantages.
Somewhat surprisingly, many fish were attracted to the
snorkelers (Dibble 1991). Fishes such as snook, tarpon,
gray snapper, cichlids, bluegill, and killifish, frequently
came within a few inches of the snorkeler's clipboard,
presumably out of curiosity. As verified in the snorkeling
efficiency tests, this attraction led to double-counting of
individuals on the multiple swims.
A major disadvantage of the method was its sensitivity
to reduced visibility conditions. Although the fishes
approached more closely in low visibility situations (1.0 to
2.0 m), the uncertainty level in identification of species
was often increased. Overall, those species that tended to
remain at the bottom or far back in the fringe were surveyed
less accurately in deeper and low visibility sites. This
problem was a particular disadvantage for surveying the
benthic forage fish at deeper stations. Even in depths of
only 1.0 m, the use of SCUBA equipment might be an advantage
if visibility range is less than water depth since it would
permit better sampling of the bottom. Some small, rare
fishes and shy cryptic forms were less accurately sampled
with the visual method. Finally, the occurrence of large


30
anchovies and silversides. The third fish group, large
roving fish, are generally greater than 15 cm in size and
occupy both the bottom and water column locations. This
group included such species as snook, tarpon, snappers,
catfish, grunts and barracuda.
Density determinations
To obtain densities for each enclosure net sample,
abundance was divided by the area the net encompassed.
Areas enclosed ranged from 72 to 196 m (mean = 119, sd =
33.2) .
For visual census samples, measurements of horizontal
secchi distance and fringe width were used as radii in
calculating the area observed (half the area of a circle).
Because it was impossible to see and accurately identify
small fish at a great distance, maximum radius values of
2.0, 3.0, or 4.0 m were applied to benthic, water column and
large roving fish respectively. Thus, the areas sampled in
the visual censuses ranged from 1.6 and 25.0 m2 at each of
the eight stations along a 70 m long transect.
Analysis Methods
Temporal patterns in density. To initially inspect the
data for patterns, salinity and density by fish group were
graphed. Temporal patterns were examined graphically and
guantitatively. For each of the three fish groups, and the
three species that were most abundant within each group,


Table 4-1. Food items found in the stomachs of fish. Data include: % Comp, the
mean % composition; % Freq, the % of specimens in which the given items occurred.
Species
Food Item 8
Eucinostomus
harengulus
% Comp. % Freq
Floridichthys
carpi
% Comp. % Freq
Fundulus
grandis
% Comp. % Freq
Lutjanus
griseus
% Comp. % Freq
Sphyraena
barracuda
% Comp. % Freq
Strongyiura
notata
% Comp. % Freq
Average
% Comp. % Freq
Crustaceans
Amphipods
25.2
76.2
9.8
47.6
8.7
27.0
4.0
19.5
2.0
7.8
0.3
6.9
8.3
33.2
Cladocerans
0.0
0.0
0.0
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
Copepods
0.9
14.4
1.2
29.9
0.0
0.8
0.0
0.0
0.1
0.0
0.0
0.9
0.4
10.7
Crabs
0.3
1.1
0.0
0.0
2.8
6.5
12.8
23.4
1.4
3.1
0.2
3.1
2.9
3.9
Isopods
2.1
12.2
1.7
7.5
8.6
21.0
7.3
27.3
5.4
1.6
0.2
10.7
4.2
12.7
Mysids
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.9
0.0
0.2
Ostracods
0.2
13.8
1.2
30.8
0.0
1.2
0.0
0.0
0.0
1.6
0.0
0.3
0.2
10.9
Shrimp
0.2
2.8
0.3
2.1
2.6
6.5
4.4
19.5
4.5
3.1
2.3
7.9
2.4
5.7
Non-chitinous invertebrates
Holothuroideans
0.2
0.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
Nematodes
0.2
12.2
0.3
19.2
0.7
8.9
0.1
7.8
0.0
4.7
0.0
1.9
0.2
10.1
Nudibranches
0.0
0.6
0.0
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.2
Polychaetes
1.9
18.8
0.5
6.3
0.3
2.4
1.3
7.8
0.0
0.0
0.0
0.6
0.7
5.6
Sipunculids
0.0
0.0
0.0
0.0
0.0
0.0
0.9
1.3
0.0
0.0
0.0
0.0
0.2
0.1
Eggs
0.2
3.3
2.6
13.5
1.3
13.3
0.0
6.5
0.3
7.8
0.1
1.3
0.7
8.0
Fish
Clupeidae
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.6
0.0
0.0
0.0
0.1
Engraulidae
0.0
0.0
0.0
0.0
0.1
0.4
0.0
0.0
1.0
6.3
2.8
1.6
0.7
0.8
Belonidae
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.4
3.1
2.4
0.6
0.5
0.3
Cyprinodontidae
0.0
0.0
0.0
0.0
1.1
1.2
1.0
1.3
2.4
26.6
17.0
3.8
3.6
2.7
Atherinidae
0.0
0.6
0.0
0.0
0.0
0.0
4.5
6.5
2.0
4.7
3.3
2.8
1.6
1.5
Syngnathidae
0.0
0.0
0.0
0.0
0.0
0.0
0.9
2.6
0.0
0.0
0.0
0.0
0.2
0.2
Gerreidae
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
1.6
0.8
0.3
0.1
0.2
Blenniidae
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.3
0.0
0.4
0.3
0.1
0.1
Gobiidae
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.6
1.6
0.4
0.6
0.2
0.2
unidentified fish
0.1
2.8
0.0
0.9
5.8
20.2
11.2
44.2
8.9
48.4
20.4
31.8
7.7
18.3
Insects
Ants
0.0
0.0
0.0
0.0
1.4
2.4
0.0
0.0
4.4
0.0
0.0
8.5
1.0
2.7
Insect larvae
0.9
6.6
0.2
0.3
1.9
0.8
0.0
0.0
2.7
0.0
0.0
0.3
1.0
1.3
Terrestrial adult insects
0.8
4.4
0.0
3.6
4.8
19.8
0.0
1.3
2.9
3.1
1.2
9.1
1.6
8.3
Mollusks
Bilvalves
0.1
2.2
0.6
9.3
0.7
3.2
0.1
2.6
0.0
0.0
0.0
0.0
0.2
3.7
Gastropods
0.8
2.2
0.5
9.6
5.0
3.2
0.1
0.0
0.0
0.0
0.0
0.6
1.1
3.8
Veliger larvae
0.0
0.0
0.0
0.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.2
Plants
Algae
6.3
55.8
7.1
67.4
2.5
20.6
2.7
26.0
0.1
7.8
0.6
1.9
3.2
33.4
Seeds
0.0
0.0
0.0
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
Protozoans
Foramifera
0.2
15.5
0.9
45.8
0.0
2.4
0.0
1.3
0.0
0.0
0.0
0.9
0.2
15.6
Unrecognizable
50.4
84.0
68.4
92.8
37.2
76.6
22.7
64.9
14.5
48.4
13.0
46.2
34.4
72.0
Empty
9.4
-
5.7
-
12.5
-
14.3
-
28.1
-
35.8
-
17.2
Total fish analyzed
181
334
248
77
64
318
1222
Total # of items (fish=1, unrec-O)
18
20
17
13
10
17
24
Sizes of fish analyzed (min-max.cm) 4.1 to 13.2
3.0 to 7.7
4.2 to 13.8
12.5 to 37.7 8.0 to 50.5
9.5 to 42.0
124


117
such habitats are well-developed. Activities which directly
destroy mangrove shorelines or degrade the quality of such
habitats have negative consequences for these important
species of sport fish.
Although no significant relationship was found among
the habitat variables and benthic forage fish as a whole,
certain species did appear to discriminate on the basis of
the measured variables. Lucania parva, Opsanus beta, and
Poecilia latipinna, three abundant species collected,
appeared to select sites with greater mangrove development.
In contrast, Floridichthys carpi chose more open sites.
This killifish seemed to prefer shallower locations, a trend
also noted for individuals living in seagrass habitats
(Sogard et al. 1987). In addition, Opsanus beta and Lucania
parva, chose sites with more abundant SAV. Similarly,
Sogard et al. (1987) collected greater abundances of these
species in seagrass bed sites with greater vegetation
densities.
In the current study, densities of water column forage
fish as a group were greatest where water depths were
greater and mangrove canopy more completely blocked the
daylight from reaching the submerged habitat. These species
use schooling as a possible defense mechanism against
predators. Thus, they may use greater volumes of water to
increase school size and prefer sites shaded by mangrove
tree canopy for additional cover.


34
efficiency increased by 7%. The total mean recovery rate
was increased to 37% by leaving the nets up overnight and
collecting the next day.
Overall, a greater percentage of larger fish were
recovered than smaller (Table 2-2). Of six large fish
(Lutjanus griseus) that were tagged, all were recovered
after rotenone application. Twice as many mid-sized as
small fish were recovered.
Visual census efficiencies. Individual test results for
small fish (all Floridichthys carpi) ranged from 25 to 27%
efficiency for the visual censusing method (Table 2-2). For
large fish (all Lutjanus griseus), results ranged from 78 to
100%. Several tagged fish were observed more than one time
during the four swims along the transect. Thus, when
analyzing the data for each sample, to prevent counting the
same fish more than once, after recording the first swim,
only unique species and size classes of fish were added to
the dataset for the second, third and fourth swims.
These efficiency analyses were intended to identify
trends in fish recovery rates. Due to wide ranges in the
test results, subsequent data analyses were not corrected
for efficiencies.
Overall Abundance
Results of the visual census differed from enclosure
net sampling results (Table 2-3). Enclosure net sampling
resulted in the collection of 82,633 fish from 59 species
and 29 families. The greatest abundance was collected at


32
log-transformed abundances as the denominator. Actual fish
assemblage patterns were compared to gradient positions
using cluster analysis. Data for each station, date and
species were used to form matrices of stations based on
similarity values (SAS CLUSTER procedure). An average
linkage method was used to join clusters of stations. The
resulting dendrograms were compared with the gradient
positions. Those stations that were placed in a group other
than the correct up-, mid- or downstream position, were
denoted as misclassified. A second analysis was conducted
on the log-transformed densities at each station in order to
more thoroughly explore the data.
Results
Tests of Recovery Efficiency
Results of recovery efficiency tests for both the
enclosure net and visual census technigues measured the
number of fish sampled out of the total that were at a site
(Table 2-2). However, no estimate is available for either
method for sample accuracy, i.e. for how many fish escaped
the area as the net was being deployed or the observer
approached the area.
Enclosure net efficiencies. In all tests spanning up-,
mid-, and downstream locations, 492 fish were marked. Of 14
total species, 60% of the fish used in the tests were
goldspotted killifish (Floridichthys carpi). An average of
18% of all fish were recovered in the initial dip-net
collections. By adding the same-day snorkeling procedure,


162
Blaber, S. J. M., D. T. Brewer, and J. P. Salini 1989.
Species composition and biomasses of fishes in different
habitats of a tropical northern Australian estuary: their
occurrence in the adjoining seas and estuarine dependence.
Estuarine, Coastal and Shelf Science 29:509-531.
Blaber, S. J. M., J. W. Young and M. C. Dunning 1985.
Community structure and zoogeographic affinities of the
coastal fishes of the Dampier Region of north-western
Australia. Australian Journal of Marine and Freshwater
Research 36:247-266.
Brook, I. M. 1977. Trophic relationships in a seagrass
community (Thalassia testudinum), in Card Sound, Florida.
Fish diets in relation to macrobenthic and cryptic faunal
abundance. Transactions of the American Fisheries Society
106(3): 219-229.
Browder, J. A. and D. Moore 1981. A new approach to
determining the quantitative relationships between fishery
production and the flow of freshwater to estuaries. In R.
Cross and D. Williams (Eds.) Proceedings of the National
Symposium on Freshwater inflow to estuaries. U. S. Fish and
Wildlife Service, Office of Biological Services, Volume I:
403-430.
Carter, L. A. Burns, T. R. Cavinder, K. R. Dugger, P. L.
Fore, D. B. Hicks, H. L. Revells, T. W. Schmidt 1973.
Ecosystems analysis of the Big Cypress Swamp and estuaries.
U. S. Environmental Protection Agency, Atlanta, Georgia.
Chester, A. J. and G. W. Thayer 1990. Distribution of
spotted seatrout (Cynoscion nebulosus) and gray snapper
(Lutjanus griseus) juveniles in seagrass habitats of western
Florida Bay. Bulletin of Marine Science 46(2): 345-357.
Croker, R. A. 1960. Growth and food of the gray snapper,
Lutjanus griseus, in Everglades National Park. Transactions
of the American Fisheries Society 91(4): 375-378.
Darnell, R. M. 1961. Trophic spectrum of an estuarine
community, base on studies of Lake Ponchartrain, Louisiana.
Ecology 42(3): 553-568.
Day, J. W. Jr., C. A. S. Hall, W. M. Kemp, A. Yanez-
Arancibia 1989. Estuarine Ecology, John Wiley & Sons, New
York, New York. 558 pages.
Davis, T. L. 0. 1988. Temporal changes in the fish fauna
entering a tidal swamp system in tropical Australia.
Environmental Biology of Fishes 21(3): 161-172.


23
Sites. Because enclosure net stations were located in
open bays, sites were selected that were protected from
prevailing direct winds. Such siting prevented the wind
from pulling the bottom of the net off the substrate. Sites
were further chosen to have between 20 and 100 cm mean water
depth at the outer prop root edge. This criterion was
intended to provide some uniformity among the sites in terms
of volume of water enclosed in the net. At each site, a
natural berm consisting of packed detritus approximately 15
cm high and 30 cm wide occurred along the landward edge.
This berm was exposed at high tide and provided a bank
beyond which fish could not escape when rotenone was applied
within the net (see below).
Procedures. At the start of the study, three sites
were selected in each of the six general locations. A
maximum of three enclosure nets could be deployed at each
location by two persons in a day if the nets were deployed
no more than approximately 1.0 km from each other.
Environmental measurements taken during each collection
included water depth, salinity, temperature, wind speed,
wind direction, and air temperature. Salinity and water
temperature were measured with a calibrated electronic
instrument (YSI Model 33 S-C-T meter). For salinities above
35 ppt, a calibrated hand-held refractometer was used.
For each net, two 30 cm wide paths were cut
perpendicular to the shoreline through the mangrove fringe
back to the berm. The paths were cleared of bottom roots


Table 2-5. Correlations between fish density for each month averaged over
the stations and salinity, water temperature and water depth. Data
for large roving fish were obtained by visual censuses. Data for benthic
and water column forage fish were obtained with enclosure nets. All density
data were converted to logarithms (log x + 1) prior to calculations.
Significant (p < 0.05) correlations are underlined.
Species or
Category
Salinity
Water
depth
Water Temperature
correlation
p-value
correlation
p-value
correlation
p-value
Benthic Forage
Fish
+0.17
0.5769
+0.05
0.8721
-0.80
0.0010
Water Column
Forage Fish
+0.06
0.8437
+0.57
0.4350
-0.04
0.8986
Large Roving
Fish
+0.31
0.2996
-0.35
0.7145
-0.76
0.0028
4^
-J


121
stable downstream locations (Rogers et al. 1984, Moyle &
Cech 1988). In northeastern Florida Bay, biomass of
submerged aquatic vegetation is much lower and highly
variable upstream than down (Montague et al. 1989) .
Abundances of polychaetes, crustaceans and other benthic
invertebrates are highly correlated with plant biomass in
this area: 80% of the epifauna live among the blades of
seagrass and algae. For estuarine fishes, these animals are
among the more heavily exploited food items (Darnell 1961,
Livingston 1982).
In light of these conditions, fish diets are likely to
display patterns along the gradient of salinity variation in
the northeastern Florida Bay study area, due to fish
foraging habits and variation in prey base. Seasonal
variations in diets may also be expected. Toward the
overall goal of identifying the influence of variation in
freshwater inflow on habitat use, the objective of this
chapter is to identify dietary components and make
comparisons among the more ubiquitous and abundant species.
Breadth and variability of diets should reflect
environmental conditions in the more variable versus stable
habitats, and over the seasons.
Materials and Methods
Fish were collected in mangrove shorelines located up-,
mid- and downstream in two systems (east and west) in
northeastern Florida Bay (Chapter 2). Species were selected
for the analysis of food habits because they ranged across


95
Sciaenids (drums) and bothids (flounder), major nursery
species in other estuaries in the southeastern United
States, Gulf of Mexico and the Caribbean (e.g. Roessler
1970, Lindall et al. 1973, Weinstein 1979, Yanez-Arancibia
1980, Stoner 1988, Sheridan 1991), were not collected at all
in the present study. Since spawning takes place in the
Gulf of Mexico, young juveniles may not survive the journey
from distant passes into northeastern Florida Bay due to
lack of tidal exchange and little circulation in the central
Bay (Sogard et al. 1987).
Mangroves in northeastern Florida Bay clearly are
nursery grounds, however, for several species of estuarine
transients, all of which are popular sportfish: gray
snapper, schoolmaster, blue-striped grunt, sheepshead and
great barracuda. They also support adult snook and tarpon,
species that are known to use similar mangrove habitats as
nursery areas elsewhere (Gilmore et al. 1983, Seaman &
Collins 1983).
A common life history pattern occurs among these
fishes: recruitment from offshore as post-larvae, settlement
and growth in inshore areas, and movement back offshore or
to deeper water as they attain larger size classes (DeSylva
1963, Starck & Schroeder 1971, Jennings 1985, Robins et al.
1986). Since these species comprise a major portion of the
large roving fish group, aspects of this basic life history
pattern may indicate why they were less abundant as a group
in the upstream locations.


Table 2-15. Repeated measures analysis of variance with density of
large roving fish as the dependent variable and general location as the independent
variable. Samples were taken by visual census along mangrove shoreline
habitats. Densities were transformed to logarithms prior to calculations.
Haemulon sciurus
Lutj anus griseus
Strongylura notata
Source
Among general locations
Among seasons
Season X general location
df F p
5/12 1.03 0.4442
4/48 0.29 0.8794
20/48 1.31 0.2190
df F p
5/12 2.26 0.0765
4/48 9.22 0.0001
20/48 1.96 0.0480
df F p
5/12 3.63 0.0311
4/48 2.97 0.0416
20/48 1.51 0.1489
Multiple comparisons
among means for
all seasons
General Sign. General
location greater location
** than **
General Sign. General
location greater location
** than **
General Sign,
location greater
** than
General
location
* *
General locations
No differences
No differences
1 & 3
(others
interm)
* source degrees of freedom / error degrees of freedom
** general locations:
1 = Joe Bay, upstream/west
2 = Trout Cove, midstream/west
3 = Buttonwood Sound, downstream/west
4 = Highway Creek, upstream/east
5 = Little Blackwater, midstream/east
6 = Blackwater Sound, downstream/east


40
Little Blackwater Sound, the midstream-east location. The
greatest number of species, however, was found in samples
from Joe Bay, the upstream-west location.
Visual census sampling resulted in observation of
222,960 fish from 51 species and 31 families (Table 2-3).
Greatest abundance and greatest number of species were
observed in samples taken in Blackwater Sound.
Samples obtained by the two methods differed in
relative abundance and numbers of species within these three
fish groups (Table 2-4). For example, many more species of
benthic forage fish were collected in the enclosure nets
(33) than were observed in the visual census (16). In
contrast, many more large roving fish species were sampled
in the visual censuses (29) than in the enclosure nets (17).
Temporal Patterns in Density by Fish Group
Benthic forage fish. In Figure 2-3, one can compare
changes in salinity with changes in density from the
enclosure net sampling; however, no consistent patterns
emerge. Great density variations occur independently of
salinity changes. Salinity varied widely over the study
period at the upstream/east (0.0-39.0 ppt) and upstream/west
(13.0-58.0 ppt) locations. Salinity also ranged widely at
the midstream/east location (19.5-50.0 ppt). However, at
the other three locations (downstream/west, downstream/east
and midstream/west), salinity remained high (29.8 to 54.0
ppt) throughout the study. Not only was the period of low
salinity longer in the upstream/east location, but also, a


143
attacked tethered fish. Other potential predators in the
vicinity were also noted. At the end of the test period,
each stake was examined. Predation was assumed to have
occurred if a fish was missing, severely damaged or if a
predator was tethered.
Analysis
For each test, the percent of prey that were subjects
of predation was used as the dependent variable for
statistical analyses. Analysis of variance (SAS GLM) was
used to determine if predator encounter rates differed due
to gradient position (i.e. far/up-, mid/up-, mid- and
downstream) or system (east, west). To make multiple
comparisons, specific F-tests were used to contrast means
for each pair of gradient positions. SAS GLM with Student-
Newman-Keuls multiple comparison tests were also used to
determine if tethering different species created an
extraneous source of variation.
In addition, correlations were calculated between the
predator encounter rates and corresponding ranges of
visibility (horizontal secchi distance), salinity, and water
depth. ANOVA was also used to determine if differences in
salinity, visibility and water depth were due to gradient
position.
Results
Of 235 tethered fish used, a mean of 83.5% were
subjects of predation after the 3 to 3.5 hour period (Table
5-3). Predation rates averaged 90% in mid/up-, mid- and


Figure
Florida
-1.
Bay.
Illustration of a well-developed mangrove habitat in northeastern


100
were censused or collected in enclosure nets and visual
samples. Due to the nature of the sampling methods, the
values for habitat structural variables were determined
using slightly different procedures for the two fish data
sets. Each of the 17 visual census sites analyzed consisted
of 8 substations. At each substation, a transect
perpendicular to the shoreline was designated. Since
corresponding fish censuses were conducted at each visual
census substation, all 136 substations were used in the
analysis. In each of the 18 enclosure net stations, 3
transects were designated within the area repeatedly
enclosed by the net. Since corresponding fish data were
available for each net site as a whole, mean values for the
3 transects within each of the 18 stations were used in the
analysis.
For all 190 transects, data were collected using a 1.0
m frame. Starting at the shoreline edge, the frame was
placed every three meters outward along the transect to 6 m
off the mangrove fringe. The following data were collected
using this overall method:
Water depth: Within each frame, four measurements were
taken, one in each quadrant using a meter stick.
Fringe width: Along each transect the distance from
the shoreline to the outermost mangrove fringe was
measured.
Tree height: Within each frame, four measurements were
taken, one in each quadrant using marked
poles or by visual estimation for trees taller
than 2 m.


Table 2-4. Summary of abundances and number of
species by method of collection and fish group.
Fish
Group
Methods
Parameter
Enclosure
Nets
Visual
Census
Benthic Forage
Fish
Total
No. Species
45,458
33
39,476
16
Water Column
Forage Fish
Total
No. Species
35,926
9
156,610
6
Large Roving
Fish
Total
No. Species
1,249
17
26,874
29
All Fish
Total
No. Species
82,633
59
222,960
51


Table 3-3. Results of the principal components analysis of the physical
and environmental variables associated with the 18 enclosure net stations
Correlations greater than 0.6 are underlined.
Principal
component
1
2
3
4
5
6
7
8
Eigenvalue
4.197
1.586
1.112
0.369
0.336
0.232
0.113
0.055
% Variance explained
52.460
19.830
13.900
4.610
4.200
2.900
1.400
0.100
Cumulative %
52.460
72.290
86.190
90.810
95.010
97.900
99.310
100.000
variance explained
Correlations (r) of
the original variables
with the
PC's
Water depth
0.931
-0.151
0.029
0.192
0.080
0.248
0.066
0.021
Fringe width
0.039
-0.181
0.930
0.074
0.261
0.153
0.063
0.001
Tree height
0.400
-0.150
0.183
-0.008
0.263
0.837
0.120
0.005
Prop roots
0.145
-0.176
0.364
-0.139
0.847
0.262
0.101
0.021
Submerged vegetation
0.116
-0.308
0.076
0.932
-0.122
-0.019
-0.043
0.014
Salinity mean
0.291
-0.591
0.497
0.373
0.247
0.039
0.163
0.302
Salinity variation
-0.169
0.876
-0.180
-0.343
-0.139
-0.167
-0.076
0.029
Tree cover
0.613
-0.239
0.198
-0.121
0.241
0.338
0.583
0.029
106


CHAPTER 2
FISH DENSITIES AND
ASSEMBLAGE PATTERNS IN MANGROVE HABITATS:
COMPARISONS ACROSS SALINITY GRADIENTS
Fishes tolerate salinities within a range of
survivability (Moyle & Cech 1988). If suitable conditions
are not available within their environment, fish will
experience stress, as evidenced by metabolic inefficiency
and, in extreme cases, death (Moyle & Cech 1989). In
general, fewer species of all faunal taxa are able to
tolerate conditions in zones with salinity conditions
typical of the upper estuary (Remane & Schlieper 1971).
This may explain the occurence of lower numbers of fish
species that occupy such areas (Deaton & Greenberg 1986).
As an alternative strategy to permanent occupancy and
metabolic adjustment, fishes can shift habitats when
salinity levels generate stress (Moser & Gerry 1989). The
occurence of a salinity gradient in the estuary provides the
opportunity for fish to exploit different habitats and
thereby avoid unsuitable salinities by movement (Weinstein
1979). By stimulating such movements, salinity conditions
may contribute to spatial and temporal fluctuations in
species composition and abundances.
17


Figure 2-10. Cluster analysis dendrograms based on species collected using
enclosure nets. Stations that grouped differently than actual gradient positions
are designated as misclassified.
a. Presence of each species (each species weighted equally) was used in one
analysis and,
b. Density of each species (fish per square meter) was used in the other.


Table 3-2. Means, standard deviations and correlations among the physical/
environmental variables associated with the 136 visual census stations.
Water
depth
cm
Fringe
width
m
Tree
height
cm
Prop
roots
n
Submerged
vegetation
cm3
Salinity
mean
ppt
Salinity
variation
ppt
Tree
cover
%
Mean
71.27
8.26
252.08
13.04
25.77
33.99
9.97
49.59
Standard deviation
20.69
4.82
102.61
11.82
10.40
8.17
5.01
12.00
Correlations (r) between variables
Water depth
1.000
Fringe width
0.537
1.000
Tree height
0.391
0.435
1.000
Prop roots
-0.187
-0.223
-0.146
1.000
Submerged vegetation
0.632
0.599
0.472
-0.146
1.000
Salinity mean
0.633
0.575
0.449
-0.103
0.937
1.000
Salinity variation
-0.598
-0.530
-0.438
0.139
-0.862
-0.914
1.000
Tree cover
0.172
0.317
0.364
-0.339
0.137
0.069
-0.032
1.000


20
Pesian of the Main Study
The climate of subtropical Florida is characterized by
a relatively long and severe dry season (November through
April) and a wet season (May through October). Thus, the
sampling schedule included monthly sampling for a one year
period to encompass the influence of changes triggered by
seasonal climatic conditions.
To monitor fish community changes across the dynamic
salinity gradient of northeastern Florida Bay, a balanced
sampling design suitable for analysis of variance was used.
The design consisted of two systems, each having three
locations along the salinity gradient (Figure 2-1 and Table
2-1). Based on the pilot study, this geographic design was
to encompass three regimes of salinity variability within
each system:
Upstream: low mean / high variation;
Midstream: mid mean / mid variation;
Downstream: high mean / low variation.
Enclosure Nets
From the pilot study, one collecting method proved to
be superior to the others, both in terms of sampling the
breadth of species at the sites and providing a quantitative
sample of fish density. This was the enclosure net first
used by Thayer et al. (1987) to sample mangrove shoreline
fishes in western and central Florida Bay. This method was
selected for targeting small benthic and water column fish
in particular.


Ill
Table 3-6. Summary of multiple regression results with 6
factors derived from the principal components analysis.
Only species for which p-values were < 0.0001 are presented
in the table. (Abbreviations as per Table 3-5)
Family/
Abundance
Adjusted Preference
P(R2)
species
R2 (pc.0001)
Engraulidae (anchovies)
Anchoa mitcheM (nets)
18,605
0.174 sparse SAV
13.7
Batrachoididae (toadfish)
Opsanus beta (nets)
529
0.388 deep water/
9.0
dense canopy
high salinity
14.9
abundant SAV
8.6
tall mangroves
4.5
Cyprinodontidae (killifish)
Floridichthys carpi (nets)
13,018
0.278 shallow water/
15.6
open canopy
short mangroves
7.8
Lucania parva (nets)
10,237
0.350 deep water/
10.6
dense canopy
high salinity
4.9
narrow fringe
8.3
abundant SAV
5.5
tall mangroves
5.2
Poecilidae (livebearers)
Poecilia latipinna (nets)
11,000
0.156 wide fringe
4.9
dense prop roots
5.3
Gambusia sp. (nets)
1,907
0.171 high salinity
9.9
Atherinidae (silversides)
Atherinomorus stipes (nets)
11,042
0.211 high salinity
7.7
abundant SAV
9.2
Menidia sp.(nets)
4,348
0.159 lowsalinity
12.9
Lutjanidae (snappers)
Lutjanus griseus (nets)
18,461
0.161 high salinity/
5.3
abundant SAV
tall mangroves
1.0
wide fringe
4.6
deep water
4.4
Gerreidae (mojarra)
Eugerres plumieri (nets)
704
0.171 lowsalinity
14.8
Haemulidae (grunts)
Haemuion sciurus (nets)
2,768
0.103 high salinity/
1.6
abundant SAV
tall mangroves
3.7
deep water
4.8
Gobiidae (gobies)
Gobiosoma robu stum (nets)
534
0.137 lowsalinity
10.6
Lophogobius cyprinoides (nets)
200
0.321 low salinity
25.7
Microgobius gulosus (nets)
2,997
0.335 low salinity
19.9
narrow fringe
11.3


131
these fishes tended to rely more strongly on omnivory, or
constant foraging for these several items, rather than
opportunism. The diets of Fundulus granis probably also
included detritus (Rozas & LaSalle 1990). However, the 3
other species analyzed in this study were apparently not
direct consumers of detritus (deSylva 1963, Odum 1971,
Starck & Schroeder 1971, Brook 1977, Thayer et al. 1987a).
The third major strategy was exemplified by the
piscivore, Sphyraena barracuda. One resource, fish, was
consistently and effectively targeted by the barracuda.
These modes of feeding are consistent with other
estuarine investigations. Two extreme modes of feeding were
identified in Lake Ponchartrain, for example, with
detritivore/omnivores, such as mullet, on one end of the
spectrum, and piscivore/specialists, such as gar and jacks,
on the other end (Darnell 1961). Similar extremes were
observed in a red mangrove/saltmarsh habitat in east
Florida, with killifish and snook at opposite poles
(Harrington & Harrington 1961). In these examples,
intermediate strategies incorporate the consumption of small
benthic invertebrates in diets with greater and lesser
portions of detritus and fish. In the current study, the
detritivore was probably most strongly represented by
Floridichthys carpi, with the barracuda at the opposite
extreme. Intermediate species include the mojarra, Fundulus
granis, Lutjanus griseus and Strongylura notata, in order
of increasing piscivory.


BIOGRAPHICAL SKETCH
Janet Ann Ledtke Ley was born 11 June, 1951, in
Detroit, Michigan, to Frederick G. and Helen M. Ledtke.
Janet graduated from Rochester High School, Rochester,
Michigan, in 1969. She received her Bachelor of Science
degree in resource development at Michigan State University,
East Lansing, Michigan, in 1973.
Janet devoted ten years to environmental planning for
the Pinellas County government, in Clearwater, Florida, from
1974 through 1984. While working as a planner, in 1979, she
earned her Master of Science degree at the University of
South Florida. Her thesis was entitled "Exploring Transfer
of Development Rights," a concept that was later
incorporated into her work on Pinellas County's plan for
environmental protection of wetland ecosystems. Janet
worked as a consultant for the Tampa office of Dames & Moore
during 1985.
In 1986, Janet enrolled in the Ph.D. program in systems
ecology, in the College of Environmental Engineering
Sciences, at the University of Florida, Gainesville,
Florida. In May 1992, Janet received her Ph.D., and hopes
to continue to work in ecosystems research.
172


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
^ llU0AA~ X
William Seaman, \Tr. /
Associate Professor of
Forest Resources and Conservation
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Nicholas Funicelli
Assistant Professor of
Forest Resources and Conservation
This dissertation was submitted to the Graduate Faculty
of the College of Engineering and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.
May 1992
Winfred M. Phillips
Dean, College of Engineering
Madelyn M. Lockhart
Dean, Graduate School


10
the Everglades system. The effects of these canals may have
included the overall reduction in the amount of freshwater
storage in the system (T. MacVicar, South Florida Water
Management District, personal communication). In addition,
after Everglades drainage, drier conditions may have
occurred more frequently in the prairies and sloughs, with
greater contrast between wet season and dry. Overland flow
of freshwater entering the downstream estuaries under these
altered conditions has probably been more rapid and less
spatially dispersed.
After entering upstream portions of Florida Bay,
freshwater moves through extensive mangrove wetlands
consisting of shallow swamp lands, creeks, ponds and bays,
eventually reaching the open portions of Florida Bay.
Changes in these brackish and marine receiving waters
attributed to managed freshwater inflow may have included
alteration of the annual salinity pattern, which led to
unnatural cycles of both reduced and hypersaline conditions.
Historical salinity data for Everglades waters, however, is
lacking for the period prior to initiation of drainage.
Evidence of the ecological effects of drainage on the
downstream estuary has been discerned from National Audubon
Society studies showing declining populations of estuarine
wading birds (e.g. spoonbills). Changes in hydroperiod have
been hypothetically linked to a reduced fish and shellfish
prey-base for the birds (National Audubon Society,
unpublished data, 1989) Furthermore, recent decreases in


158
The potential pool of post-larval estuarine transient
species from the Atlantic Ocean and Gulf of Mexico includes
sciaenids, lutjanids, haemulids, centropomids and elopids.
Presumably, currents carry potential recruits to the
northeastern Florida Bay area. Exchange with these sources
is limited in northeastern Florida Bay due to the presence
of the Keys and western mudbanks. Internal circulation in
Florida Bay is also weak due to the many islands and lack of
tides. If post-larval transients do enter the northeastern
Bay, they may meet with significant predation pressure due
not only to the occurrence of piscivores but also to a
reduction in the cover afforded by seagrass beds, which tend
to become less developed in the eastern Bay. These
conditions indicate that the chances of post-larval forms
reaching the upstream locations in the study area may be
small. Such conditions may not be unusual in estuaries,
however.
Habitat conditions documented in this study for the
upstream locations included almost no SAV and reduced
mangrove habitat development. However, shallow ponds and
sinuous creeks upstream may effectively reduce predator
encounter rates. Thus, a key factor in improving the use of
upstream habitats by estuarine transient juveniles appears
to be the presence of a persistent abundance of SAV.
Management Implications
Northeastern Florida Bay may have historically
supported more estuarine transient juveniles and greater


46
the visual census stations were notably fresher (1.4-50.0
ppt) than the corresponding net stations (13.0-58.0 ppt)
(Figures 2-3 & 2-5). In this location, the visual census
stations were located within the creek itself while the net
stations were located in Joe Bay immediately at the creek
mouth (Figure 2-1). These seemingly slight differences in
location may have contributed to observed differences in
salinity patterns for the visual and net stations.
For the visual census, the salinity patterns in both
the eastern and western upstream locations were similar to
each other except in spring 1990, when freshwater entered
the upstream/east location reducing salinity to 30.0 ppt and
the upstream/west location became strongly hypersaline (47.3
ppt) (Figure 2-5). The midstream/east location was overall
more variable than the midstream/west location. At the
mid/west and downstream visual census locations, salinity
patterns were uniformly high.
Density of large roving fish ranged from zero at the
uppermost upstream/west location in the spring, fall and
winter of 1989, to 2.3 fish m-2 at Duck Key
(midstream/west) in winter. From the graphs, it appears
that changes in density of this group were independent of
salinity changes (Figure 2-5) .
Temporal correlations. No significant correlations
between temporal changes in salinity or water depth, and
temporal changes in fish densities were found (Table 2-5).
However, changes in water temperature were correlated with


Water Column Forage Fish
Density & Salinity vs. Month
Enclosure Nets
O
o
(/)
O
CD
ir
o
c/)
Q
D
CL
Figure 2-4. Density of water column forage fish collected with enclosure nets in
mangrove habitats (histogram) and corresponding salinity measurements (lines). Error
bars indicate standard deviation among the three enclosure nets deployed in each
general location.


Table 5-3. Results of predator encounter trials comparing rates in
four gradient positions from down to far upstream in two systems.
Fish were tethered for three hours adjacent to mangrove edges.
Trials were conducted from June 20, 1990 to August 3, 1990.
System*
Gradient**
Test
Number
Results
Total Percent
Fish Missing
Salinity
Secchi
distance
Mean Water
Depth
1
1
1
16
25
22.5
1.0
64
1
1
2
10
70
23.5
1.5
62
1
2
1
12
100
22.9
1.0
53
1
2
2
15
87
14.7
1.0
52
1
3
1
16
100
52.0
2.2
45
1
3
2
15
100
50.0
1.0
62
1
4
1
16
100
50.0
5.5
58
1
4
2
15
93
45.0
3.4
58
2
1
1
13
77
7.3
2.0
47
2
1
2
15
53
16.0
2.0
48
2
2
1
16
94
29.9
1.0
48
2
2
2
13
100
15.5
1.0
50
2
3
1
16
88
50.0
0.8
62
2
3
2
16
69
48.0
0.8
73
2
4
1
16
93
45.0
4.0
93
2
4
2
15
87
45.0
9.0
93
Total
235
Mean
83.5
33.4
2.3
60.8
Systems: 1 = West, 2 = East
Gradient: 1 = Far/upstream
2 = Mid/upstream
3 = Midstream
4 = Downstream


4
The degree of selection among mangrove habitats by
fishes has not been determined. Due to variation within
mangrove forests, however, preferences are likely to be
displayed. In south Florida, for example, three species of
mangrove trees occur: red mangroves (Rhizophora mangle),
black mangroves (Avicennia germinans) and white mangroves
(Laguncularia racemosa) (Odum et al. 1982). These trees
vary greatly in type of submerged features and potential
cover for fishes. For example, red mangroves provide prop-
roots; these are strong, woody structures that tend to
extend from the mid-tree trunk downward to the substrate.
Black mangroves, in contrast, tend to support a bed of
pneumatophores that are pencil-like structures that grow
upward from the substrate to several centimeters.
Variation in degree of exposure to flushing also
contributes to differences among mangrove forests and,
hence, may influence habitat use by fishes. While the
fringing mangroves along the shoreline are regularly flooded
and thus accessible, more interior basin forests are
irregularly inundated and thus occasionally available (Odum
et al. 1982).
Within fringing shorelines, higher flushing rates
contribute to greater mangrove habitat development (e.g.
taller trees, more leaf production), which, in turn, is
likely to generate more massive submerged structure for
cover. Furthermore, detritus-based food resources are
likely to be more abundant near highly productive mangroves.


53
significant interaction between gradient and system for
density of water column fish (Table 2-6). Little Blackwater
Sound (mid/east) and Buttonwood Sound (down/west) had
significantly greater densities than Highway Creek (up/east)
and Trout Cove (mid/west) (Table 2-7, Figure 2-6).
Large roving fish analysis of variance. In contrast to
the other two groups, for large roving fish, a clear effect
of gradient position on fish density occurred. Fish in this
group were significantly less abundant at the upstream
gradient locations than mid- or downstream (Table 2-6). No
general locations varied significantly from the others
(Table 2-7).
Spatial correlations. To analyze spatial trends,
correlations between mean fish densities and salinity,
salinity variation, and water depth were determined (Table
2-8). A significant correlation between average density of
large roving fish and station salinity was found; lower
densities occurred at stations with lower mean salinity
levels and greater temporal variability. In addition, both
water column forage fish and large roving fish were
significantly more abundant at stations with deeper water.
Temporal Patterns in Density by Species
As indicated by the correlations between mean densities
and salinity, water depth and water temperature, temporal
patterns differed among the species (Table 2-9). No
significant correlations were found between salinity changes
from month to month and densities for any species. Temporal


Figure 3-5. Results of principal components regression. B, least squares
regression coefficient; p, significance level; R2, percent variance explained
by the entire model; df, degrees of freedom; SS, sums of squares;
P(R2), percent variance explained by an individual principal component.
BENTHIC FORAGE FISH
(Net Data)
Source
df
SS p
R2
Source
B
P
P(R2)
Preference (p<-001)
Regression
6
2.647 0.1031
0.049
PCI
-0.057
0.0992
1.25
Residual
209
51.601
PC2
-0.027
0.4338
0.28
PC3
-0.054
0.1218
1.10
PC4
0.065
0.0643
1.57
PC5
0.027
0.4382
0.27
PC 6
-0.027
0.4409
0.27
WATER COLUMN
FORAGE
FISH (Net Data)
Source
df
SS p
R2
Source
B
P
P(R2)
Preference (p<.001)
Regression
6
6.826 0.0142
0.073
PCI
0.152
0.0009
5.08
Deep water/dense canopy
Residual
209
87.084
PC2
-0.071
0.1083
1.15
PC3
0.040
0.3788
0.35
PC4
-0.017
0.7085
0.06
PC5
0.052
0.2525
0.58
PC6
0.030
0.5135
0.19
LARGE ROVING
FISH (Visual Data)
Source
df
SS p
R2
Source
B
P
P(R2)
Preference (p<.001)
Regression
6
44.100 0.0001
0.188
PCI
0.097
0.0001
6.48
High salinity/dense SAV
Residual
1617
235.130
PC2
0.032
0.0001
0.73
Dense canopy
PC3
0.059
0.0001
2.37
Tall trees
PC4
-0.0170
0.0410
0.21
PC5
0.061
0.0001
2.62
Wide fringe
PC6
0.095
0.0001
6.31
Deep water
109


INFLUENCE OF CHANGES IN FRESHWATER FLOW ON THE USE OF
MANGROVE PROP ROOT HABITAT BY FISHES
By
JANET A. LEY
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1992
imiVERSin OF ROMBft,1WKKS

ACKNOWLEDGEMENTS
I would like to extend my thanks to all my committee
members and my Department Chairman, Joseph Delfino. They
advised and supported me throughout my field work and
writing. Clay Montague introduced me to science and Florida
Bay, and encouraged me to pursue my interests. Carole
Mclvor gave me guidance, scientific insight, and maintained
faith in my abilities under all circumstances. I am
extremely grateful to Bill Seaman for his efforts in
obtaining Sea Grant support for a substantial portion of the
project. In his Wetlands Ecology class, Ronnie Best
introduced me to working in mud and swamps, features which
later became a major part of my life. He also allowed me to
live in the Center's Winnebago for two years, permitting me
to operate on an intense and flexible schedule in Key Largo.
Frank Nordlie expressed constant interest and encouraged me
in my work. My visits with Nick Funicelli always included
valuable personal and professional insights.
I am extremely grateful for the efforts of Dan Haunert
of the South Florida Water Management District. He believed
in the benefits of this research to Florida Bay and
aggressively oversaw the process of obtaining funding. In
ii

addition, my conversations with him gave me renewed
enthusiasm and perspective. Dewey Worth, Dan's colleague,
followed through with continued encouragement and support in
later phases of the project.
Laura Flynn and Luke Hasty were my very competent field
assistants, providing good humor and constructive
suggestions. They remained enthusiastic in every
circumstance, from diving with major unknown creatures, to
measuring 2-day old fish in 95 degree heat, to snorkeling in
double hoods and wetsuits.
Jacque Stevens, Harriett McCurdy and my brother, Fred
Ledtke, Jr., were my most faithful volunteers. Jacque
helped me tether over 200 fish and her ideas were
invaluable. Fred devoted his hard earned vacations to his
older sister's unusual effort.
I would also like to express my gratitude to the staff
of Everglades National Park. At the Key Largo Ranger
Station, Dave and Louise King, Linda Cramer, and Dave
Viscera included me as part of their small neighborhood
during my 2 year residency. I am grateful for their support
and rescues during boat break-downs. From the South Florida
Research Center, Mike Robblee allowed me to use Park boats,
provided insights concerning my research questions and
encouraged my efforts. DeWitt Smith also gave me
encouragement and perceptive advice. Bill Loftus' help in
iii

fish taxonomy was invaluable. Katy Kuss was both a great
adviser and a good friend.
Gordon Thayer, National Marine Fisheries Service,
advised me on the use of enclosure nets. In addition, other
visiting scientists shared ideas with me, including Paul
Carlson, Florida DNR; Jay and Rita Zieman and Jim
Fourqurean, University of Virginia; and Dave Porter,
University of Georgia. I am also grateful to the scientists
of the National Audubon Society in Tavernier, Florida, who
treated me as an adjunct staff member, especially George
Powell, Mike Ross and Jerry Lorenz.
In Gainesville, Ken Portier helped in the initial study
design and last phases of analysis. In the bulk of the
analysis effort, Steve Linda advised me on handling a very
large data set. Hans Gottgens, my officemate, was
constantly patient and extremely helpful in offering
computer, scientific and personal advice.
Most importantly, I would like thank those who
encouraged my pursuit of this degree as a personal goal and
supported me throughout the process. These are my parents
who nurtured my spirit of independence and appreciation of
nature, and Darlene Kalada, my best friend, on whose support
and encouragement I could always depend.
iv

TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS ii
ABSTRACT viii
CHAPTERS
1. GENERAL INTRODUCTION 1
Estuarine Fish Ecology 1
Mangrove Fish Ecology 3
The Florida Bay Ecosystem 5
Problem Definition 9
Objectives 12
Study Area 13
Fish Community Sampling Design 14
2. FISH DENSITIES AND ASSEMBLAGE PATTERNS IN
MANGROVE HABITATS: COMPARISONS
ACROSS SALINITY GRADIENTS 17
Materials and Methods 19
Results 32
Discussion 84
3. FISH COMMUNITIES IN FLORIDA BAY MANGROVE
SHORELINE HABITATS: RELATIONS WITH PHYSICAL
PARAMETERS AND COVER 98
Materials and Methods 99
Results 109
Discussion 114
4. FOOD HABITS OF MANGROVE FISHES:
A COMPARISON ACROSS SALINITY GRADIENTS 120
Materials and Methods 121
Results 123
Discussion 129
5. PREDATOR ENCOUNTER RATES ON SMALL BENTHIC
FISH ACROSS A SALINITY GRADIENT 134
Materials and Methods 137
Results 143
Discussion 149
v

6. IMPLICATIONS AND CONCLUSIONS 154
Implications for Mangrove Fish Ecology 154
Implications for Estuarine Fish Ecology:
the Nursery-ground Hypothesis 157
Management Implications 158
LITERATURE CITED 161
BIOGRAPHICAL SKETCH 172
vi

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
INFLUENCE OF CHANGES IN FRESHWATER FLOW ON THE USE OF
MANGROVE PROP ROOT HABITAT BY FISHES
By Janet A. Ley
May 1992
Chairperson: Clay L. Montague
Cochairperson: Carole C. Mclvor
Major Department: Environmental Engineering Sciences
The hypothesis that seasonal changes in freshwater
inflow (indicated by salinity) influence habitat use by
fishes was tested in northeastern Florida Bay, extreme south
Florida. Fishes were sampled monthly for 13 months using
visual censuses and enclosure nets.
Of the 305,589 individuals observed, 91% were estuarine
residents, numerically dominated by engraulids, atherinids
and cyprinodontids. Occasional marine and freshwater
visitors comprised 2% of the individuals, and estuarine
transients, 8%. No young-of-the-year estuarine transients
were observed.
Salinity ranged between 0.0 to 58 parts per thousand
(ppt) upstream, 19.5 to 54 ppt midstream, and 30 to 50 ppt
downstream. The 77 species were grouped for analysis:
vii

small benthic, small water column, and larger fishes.
Abundances of larger fishes were consistently lower upstream
(0.15 fish/square meter (nr)), than mid- (0.65 fish/m ), or
downstream (0.55 fish/m2). Species of larger fishes
numbered fewer upstream (11), than midstream (15), and
downstream (22). Benthic and water column fish abundances
did not vary along the gradient. Temporally, fish
distribution was uncorrelated with salinity.
Development of mangrove habitat and submerged aquatic
vegetation (SAV) were reduced upstream. Fish diets shifted
to other foods upstream. Thus, where seasonal changes in
freshwater inflow were greater (i.e. upstream), species and
numbers of larger fishes were lower, possibly due to
salinity conditions, food availability and habitat
development.
To determine if lower salinity conditions alone led to
reduced predation, prey fishes were tethered along the
gradient. Predator encounter rates were not different over
the salinity range tested, but were 50% lower at the most
remote sites. This was perhaps a function of accessibility
of the sites to roving predators.
Water management strategies to increase mangrove
development and SAV are recommended research priorities.
However, severe ecotonal differences between Bay and ocean
waters, coupled with limited circulation and significant
viii

predation may inhibit recruitment and survival of post-
larval fishes from offshore. An unbroken continuum of good
habitat from outer to upper reaches may be
northeastern Florida Bay is to function as
area for estuarine transient fishes.
necessary if
a prime nursery
ix

CHAPTER 1
GENERAL INTRODUCTION
The goal of improved management of surface waters to
benefit estuarine fish populations in Florida Bay provided
the incentive for this research. Before objectives and
strategies can be established toward this goal, however, a
better understanding of how freshwater inflow influences
fish communities in mangrove estuaries overall is needed.
This involves aspects of fish ecology of both estuarine and
mangrove ecosystems.
Estuarine Fish Ecology
Ecologists often divide estuarine fishes into three
groups: estuarine residents (complete their entire life
cycle in the estuary), estuarine transients (spawn offshore,
their young use the estuary as a nursery), and occasional
marine visitors (usually adults) (Day et al. 1989) .
Resident and transient species tend to be widespread, but
marine visitors are usually restricted to the higher
salinity zones of the lower estuary (Weinstein 1979) At
certain times of the year densities may increase
dramatically as influxes of transient juveniles enter
estuarine systems. Some species tend to migrate to
upstream-most habitats upon initially entering the estuary;
1

2
they then may disperse to lower reaches as they grow larger
(Weinstein 1979, Rogers et al. 1984, Loneragan et al. 1990).
The prominence of transient juvenile fish and
crustaceans led to the application of the term "nursery-
ground" to many estuaries (Gunter 1961, McHugh 1967,
Weinstein 1979). A major role of freshwater discharge in
such systems may be to increase food availability for fishes
by transporting nutrients which stimulate primary production
and by increasing detrital transport and processing (Odum et
al. 1982). Freshwater inflow may also improve the chance
for survival of juvenile fish in estuaries by reducing
salinity levels below the limits tolerable by stenohaline
marine predators (Gunter 1961, 1967).
Browder and Moore (1981) offered a comprehensive
nursery ground hypothesis linking several of these concepts.
They split habitat factors into those that are relatively
stable (e.g. shoreline edge, bottom type) and those that are
movable (e.g. salinity, food resources). Favorable habitat
for particular juveniles consists of combinations of these
factors that promote growth. According to their theory, the
inflow of freshwater acts to position an area of favorable
moveable habitat relative to important stationary habitat.
Thus, for any estuary there is a rate of freshwater flow
sufficiently high to push the band of potentially favorable
moveable features beyond estuarine boundaries into open
waters, perhaps eliminating favorable habitat entirely.
Likewise, for every estuary, there is a rate of freshwater

3
flow so low that the band of favorable salinities retreats
upstream where stationary features may be unfavorable. The
ideal situation with regard to freshwater inflow is one that
maximizes the area of favorable habitat within the estuary
over the peak period of nursery use. This hypothesis seems
particularly applicable for analyses of fish ecology in
mangrove-dominated estuaries.
Mangrove Fish Ecology
In tropical and subtropical areas of the world,
mangroves are dominant shoreline features. Mangrove-derived
detritus forms a food base for fish occupying mangrove
ecosystems (Odum 1971). Mangrove shorelines may also
provide cover for fishes (Thayer et al. 1987a). However,
few studies have documented aspects of the direct use of
mangrove habitats by fishes probably because monitoring
fishes within the complex tangle of roots and branches is
extremely difficult. Efforts have only recently focused on
obtaining quantitative data on habitat use (i.e. Thayer et
al. 1987a, Sheridan 1991, Morton 1990, Robertson & Duke
1987). Strong linkages between mangroves and adjacent
habitats may exist. For example, diel habitat shifts occur
in both non-tidal (Thayer et al. 1987a) and tidal systems
(Morton 1990, Robertson & Duke 1987). Shifts from other
habitats to mangroves occur during the life history of some
species such as gray snapper (Lutjanus griseus) (Starck &
Schroeder 1971).

4
The degree of selection among mangrove habitats by
fishes has not been determined. Due to variation within
mangrove forests, however, preferences are likely to be
displayed. In south Florida, for example, three species of
mangrove trees occur: red mangroves (Rhizophora mangle),
black mangroves (Avicennia germinans) and white mangroves
(Laguncularia racemosa) (Odum et al. 1982). These trees
vary greatly in type of submerged features and potential
cover for fishes. For example, red mangroves provide prop-
roots; these are strong, woody structures that tend to
extend from the mid-tree trunk downward to the substrate.
Black mangroves, in contrast, tend to support a bed of
pneumatophores that are pencil-like structures that grow
upward from the substrate to several centimeters.
Variation in degree of exposure to flushing also
contributes to differences among mangrove forests and,
hence, may influence habitat use by fishes. While the
fringing mangroves along the shoreline are regularly flooded
and thus accessible, more interior basin forests are
irregularly inundated and thus occasionally available (Odum
et al. 1982).
Within fringing shorelines, higher flushing rates
contribute to greater mangrove habitat development (e.g.
taller trees, more leaf production), which, in turn, is
likely to generate more massive submerged structure for
cover. Furthermore, detritus-based food resources are
likely to be more abundant near highly productive mangroves.

5
Most mangrove-dominated estuaries contain examples of all
these habitats and may thus provide a diversity of
conditions for use by fishes.
The Florida Bay Ecosystem
Florida Bay is a large (1,500 km2), mangrove/seagrass
dominated estuary located in extreme south Florida. The
majority of the Bay is not subject to tidal influence.
Wind-driven water movements can, however, raise or lower Bay
water levels rapidly. Sustained strong easterly winds can
literally blow water out of Florida Bay into the open Gulf
of Mexico. In similar fashion, winds from the north can
accelerate the introduction of mainland drainage into the
northern part of the Bay, and winds from the west can move
the water into the northeastern corner of the Bay (Ginsberg
1956) .
Internal circulation is restricted due to several
features of the Bay. The interior contains over 300
mangrove-fringed and overwash islands. On the east, U. S.
Highway 1 separates Florida Bay and Barnes Sound with only
one pass and two culverts providing water exchange. On the
west, it is separated from the open waters of the Gulf of
Mexico by a series of mud banks that are at least 2.0 km
wide and are often exposed (Holmquist et al. 1989b) (Figure
1-1). The lower Bay is separated from the thermally stable
and constant flow of the Gulfsteam by a series of limestone
islands known as the Florida Keys. Several major passes
occur through the Keys in the western Bay, but in the

Figure 1-1. Regional maps of the study area showing
features of the upstream drainage basins and Florida Bay.
a. Boundaries of drainage basins and tributaries to
northeastern Florida Bay. (Source: Schomer & Drew 1982).
b. Florida Bay showing the extensive mudbank system (stiple
pattern). Arrows indicated passes to the Atlantic Ocean and
Barnes Sound. (Source: Holmguist et al. 1989b)

a.
Study Ana\ 0
\
US Highway 41
C-111 Canal
Study Area
L12 km|

8
northeastern portion, only one pass exists, a man-made cut
through Key Largo to the ocean side (Adam's Cut) (Figure 1-
1). On the northern boundary of the Bay are the Florida
Everglades.
The Florida Bay area is subject to an annual water
deficit with evaporation exceeding total rainfall (Tabb et
al. 1962). Annual rainfall in northeastern Florida Bay
ranges from 1600 mm on the mainland at Homestead to 1200 mm
on the south at Key Largo. The climate of subtropical south
Florida is characterized by a relatively long and severe dry
season (November through April) and a wet season (May
through October) (Schomer & Drew 1982).
Sea level becomes relatively high on an annual basis
from August to December reaching a maximum of about 15 cm
above the annual average in October (Ginsberg 1956, Provost
1973, Holmquist et al. 1989b). By late November or early
December, Bay level recedes to the annual average, which
probably accelerates the drainage of freshwater into the Bay
from the mainland. At this time, the zone of reduced
salinity may extend farther south and southeast into mid-
and downstream Florida Bay areas.
The major source of freshwater flow into Florida Bay is
from a series of approximately 20 creeks and Taylor River,
which carry surface water from the Taylor Slough/C-111
drainage area into the Bay. This system is smaller than the

9
Shark River Slough, a separate system which extends from
Lake Okeechobee southward toward the Gulf of Mexico and
drains most of the Everglades.
These overall features contribute to several
environmental and biological patterns. Gradients in
environmental variables occur in Florida Bay, from southwest
to northeast. These gradients include amount of water
exchange, sediment depth, and seagrass standing crop (Zieman
et al. 1989). The area northeast of the central line of mud
banks is characterized by very restricted circulation and no
tidal influence (Schomer & Drew 1982). A thin sediment
veneer covers the basin bedrock in the northeast Bay,
deepening towards the southwest. In addition, seagrass
density and productivity decreases dramatically from
southwest to northeast (Zeiman et al. 1989).
Problem Definition
Water management decisions in the eastern Everglades
have potentially impacted Florida Bay through changes in the
timing and quantity of freshwater discharge. Under pre
drained conditions, in this area, surface freshwater moved
over grassy marl prairies that were seasonally flooded
(Schomer & Drew 1982). A complex network of streams,
bordered by mangroves and other shrubs carried freshwater
inflow to receiving waters downstream in a manner that was
presumably both gradual and dispersed.
Beginning in the early 1900s, construction began on an
extensive system of canals and ditches throughout much of

10
the Everglades system. The effects of these canals may have
included the overall reduction in the amount of freshwater
storage in the system (T. MacVicar, South Florida Water
Management District, personal communication). In addition,
after Everglades drainage, drier conditions may have
occurred more frequently in the prairies and sloughs, with
greater contrast between wet season and dry. Overland flow
of freshwater entering the downstream estuaries under these
altered conditions has probably been more rapid and less
spatially dispersed.
After entering upstream portions of Florida Bay,
freshwater moves through extensive mangrove wetlands
consisting of shallow swamp lands, creeks, ponds and bays,
eventually reaching the open portions of Florida Bay.
Changes in these brackish and marine receiving waters
attributed to managed freshwater inflow may have included
alteration of the annual salinity pattern, which led to
unnatural cycles of both reduced and hypersaline conditions.
Historical salinity data for Everglades waters, however, is
lacking for the period prior to initiation of drainage.
Evidence of the ecological effects of drainage on the
downstream estuary has been discerned from National Audubon
Society studies showing declining populations of estuarine
wading birds (e.g. spoonbills). Changes in hydroperiod have
been hypothetically linked to a reduced fish and shellfish
prey-base for the birds (National Audubon Society,
unpublished data, 1989) Furthermore, recent decreases in

11
sportfish populations have been linked to hypersalinity
stress for certain sportfish in Everglades National Park
(Rutherford et al. 1989). The Everglades estuaries are also
critical habitat areas for other endangered aquatic species
(e.g. American crocodile) that rely on the same forage base
as do birds and sportfish (SFWMD 1989).
Thus, groups concerned about these problems spurred
South Florida Water Management District (SFWMD) officials to
take action that would return more natural drainage patterns
to the estuarine areas of the Everglades. Some of these
actions have focused on the C-lll Canal/Taylor Slough
watershed which includes agricultural lands in a large
drainage basin east of the Park. The downstream leg of the
canal runs northwest to southeast, passes under U. S.
Highway 1 and continues southward outside of Florida Bay, to
Barnes Sound (Figure 1-1). In low flow periods, the canal
has functioned like a dike by preventing overland flow of
freshwater from reaching both the downstream prairies and
the approximately seven small creeks tributary to
northeastern Florida Bay. In high flow conditions, water
still flows through, sometimes sending slugs of freshwater
into northeastern Florida Bay. Local topographic conditions
tend to direct more freshwater toward U.S. Highway 1 than
toward the west (Tabb et al. 1967). Thus, under these
management conditions, the historic salinity regime is
likely to have been altered. Changes in hydroperiod have
probably resulted in more severe hypersaline conditions and

12
sudden salinity changes of great magnitude, especially in
the eastern part of Florida Bay.
In the mid-1980s engineering alterations created
several cutouts, each 20 meters wide, in the south bank of
C-lll canal. The cutouts were intended to restore the more
dispersed and gradual pattern of freshwater inflow to
northeastern Florida Bay. Furthermore, an earthen plug was
installed to block the C-lll outfall to Barnes Sound except
on extreme floods when SFWMD can release water by opening it
with draglines. The result of these alterations was to
provide more flexible management of freshwater flow to
northeastern Florida Bay. The guestion remaining is how to
utilize this flexibility to improve ecological conditions.
Objectives
Fish and Salinity
The first study objective was to determine the extent
to which species composition and abundance were influenced
by salinity variability in the northeastern Florida Bay
study area. Because of direct and indirect salinity
influences, more variable fish abundances and distinct
community differences were expected at the upstream
locations over an annual cycle that included both wet and
dry seasonal differences in freshwater inflow. The eastern
portion of the study area was also expected to be distinctly
more variable than the western portion because of the
influence of the C-lll Canal.

13
Habitat Features
Because salinity is not the only feature of the habitat
that varies along the complex environmental gradient within
the study area, it was also necessary to consider other
features of the fixed and moveable habitat (Browder & Moore
1981) as potentially influencing fish community structure.
The second study objective was to determine important
habitat features that influence the abundance and species
composition of mangrove fish communities and compare these
features across the salinity gradient. Fixed habitat
structural features such as mangrove tree height and prop
root density, environmental features such as water
temperature, and fish diet and predation, were expected to
influence the differences among fish assemblages across
environmental gradients.
Study Area
The 250 km2 study area, located in extreme northeastern
t
Florida Bay, consists of a series of shallow bays and ponds
(less than 1.0 m in depth) bordered by mangroves. The
upstream portion of the area is subject to freshwater inflow
from seven mangrove-lined tributaries originating in the
Taylor Slough/C-111 drainage basin.
In this region of Florida Bay, rapid ecological changes
can take place when salinity variations occur suddenly, as
at the start of the rainy season (Montague et al. 1989).
Because tidal influences are almost negligible in the
northeastern Florida Bay area, salinity changes are caused

14
by the variations in rainfall and subsequent freshwater
flowing south through the tributaries, and variations in
wind speed and direction. The rate and degree of salinity
change are relatively unpredictable and can be rapid (hours)
or slow (days) depending on changes in the weather.
Fish Community Sampling Design
To monitor fish community changes across the dynamic
salinity gradient in northeastern Florida Bay, a balanced
two-way analysis of variance (ANOVA) design was used, with
two systems, each composed of three salinity regimes (Figure
1-2). Generally, upstream locations included one of the
creeks which carries freshwater from the Taylor Slough/C-111
Basin, an interior bay downstream from the creek but still
measurably affected by freshwater inflow, and an outer bay
much less affected by freshwater inflow but more by marine
influences. Specifically, the locations were as follows:
upstream sites were located in Highway Creek and Long Sound
in the eastern system and Snook Creek and Joe Bay in the
western system (Figure 1-2) ; midstream sites were located
in Little Blackwater Sound in the eastern system and the
Trout Cove area in the western system; downstream sites
were located in Blackwater Sound in the eastern system and
Buttonwood Sound in the western system.

Figure 1-2. Map of the Florida Bay study area.


CHAPTER 2
FISH DENSITIES AND
ASSEMBLAGE PATTERNS IN MANGROVE HABITATS:
COMPARISONS ACROSS SALINITY GRADIENTS
Fishes tolerate salinities within a range of
survivability (Moyle & Cech 1988). If suitable conditions
are not available within their environment, fish will
experience stress, as evidenced by metabolic inefficiency
and, in extreme cases, death (Moyle & Cech 1989). In
general, fewer species of all faunal taxa are able to
tolerate conditions in zones with salinity conditions
typical of the upper estuary (Remane & Schlieper 1971).
This may explain the occurence of lower numbers of fish
species that occupy such areas (Deaton & Greenberg 1986).
As an alternative strategy to permanent occupancy and
metabolic adjustment, fishes can shift habitats when
salinity levels generate stress (Moser & Gerry 1989). The
occurence of a salinity gradient in the estuary provides the
opportunity for fish to exploit different habitats and
thereby avoid unsuitable salinities by movement (Weinstein
1979). By stimulating such movements, salinity conditions
may contribute to spatial and temporal fluctuations in
species composition and abundances.
17

18
Contrary to what might be expected based on such
physiological factors, however, abundance peaks for many
estuaries occur in conjunction with the initiation of the
period of maximum freshwater inflow, when salinity levels
drop dramatically. Estuarine transient juveniles may
constitute most of the individuals during these peak periods
(Yanez-Arancibia et al. 1980, Bell et al. 1984, Pinto 1987).
In some species, juvenile fishes may be capable of
exploiting salinities at lower levels than adults (Gunter
1967, Moser & Gerry 1989). However, in environments with
more stable salinities, estuarine transient juveniles can
also be abundant (Little et al. 1988, Robertson & Duke
1990b). Thus, the role of seasonal changes in salinity on
fish communities reguires further exploration.
Based on investigations conducted near and within the
northeastern Florida Bay study area, at the initiation of
the rainy season (June), changes in salinity were expected
to occur, expanding the zone of low salinity further
downstream (Ginsburg 1956, Tabb et al. 1962, Lindall et al.
1973, Thayer et al. 1987). This zone of lower salinity was
expected to persist after the end of the rainy season, as
freshwater from the eastern Everglades gradually drained
into Florida Bay.
Toward the goal of understanding the influence of
freshwater inflow on fishes, the objective of this portion
of the study was to identify spatial and temporal patterns
in fish assemblages across the salinity gradient and thereby

19
test the following hypotheses. First, temporal changes in
fish densities were expected to occur in conjunction with
salinity changes. Secondly, in areas where salinities were
more variable, numbers of species of fishes were expected to
be lower than more stable areas. Thirdly, a community of
fishes including estuarine transient juveniles was expected
to occur in the study area. Finally, relatively lower
densities were expected to occur in the upstream locations
as an function of variable salinity conditions.
Materials and Methods
Pilot Study
A six-month pilot study was conducted to determine the
most effective methods for guantitatively sampling fishes in
mangrove prop root habitat throughout northeastern Florida
Bay. Absence of tidal exchange in the study area was an
important factor in selecting methods. Both collecting and
observational methods were explored. Collecting gear
selected for preliminary testing included minnow traps,
Caribbean fish traps, gill nets, pull-up nets and enclosure
nets with rotenone. The two visual census methods tested
were: 1) direct recording of fishes observed with mask and
snorkel on underwater data sheets and, 2) underwater video
taping. Two complementary methods were selected from those
tested in an attempt to sample the entire fish community.
These two methods were enclosure nets and direct visual
observation.

20
Pesian of the Main Study
The climate of subtropical Florida is characterized by
a relatively long and severe dry season (November through
April) and a wet season (May through October). Thus, the
sampling schedule included monthly sampling for a one year
period to encompass the influence of changes triggered by
seasonal climatic conditions.
To monitor fish community changes across the dynamic
salinity gradient of northeastern Florida Bay, a balanced
sampling design suitable for analysis of variance was used.
The design consisted of two systems, each having three
locations along the salinity gradient (Figure 2-1 and Table
2-1). Based on the pilot study, this geographic design was
to encompass three regimes of salinity variability within
each system:
Upstream: low mean / high variation;
Midstream: mid mean / mid variation;
Downstream: high mean / low variation.
Enclosure Nets
From the pilot study, one collecting method proved to
be superior to the others, both in terms of sampling the
breadth of species at the sites and providing a quantitative
sample of fish density. This was the enclosure net first
used by Thayer et al. (1987) to sample mangrove shoreline
fishes in western and central Florida Bay. This method was
selected for targeting small benthic and water column fish
in particular.

Figure 2-1. Northeastern Florida Bay study area with
sampling stations indicated.

Table 2-1. List of stations used in sampling fish in mangroves
using enclosure nets and visual census surveys.
System
Gradient
General
Location
Enclosure Net
Stations
Number Name
Number
Visual Census Stations
Name
West
Upstream
Joe Bay
1
east
1
Snook Creek Pond 1
2
mid
2
Snook Creek Pond 2
3
west
3
Snook Creek Pond 3
4
west Joe Bay*
Midstream
Trout Cove
4
northeast
5
southeast Trout Cove
5
mid
6
Tern Key
6
southeast
7
Deer Key*
8
Duck Key
Downstream
Buttonwood
7
northeast
9
Buttonwood Point
Sound
8
mid
10
Whaleback Key
9
southwest
11
Unnamed Key
12
Key Largo Ranger Station
East
Upstream
Highway Creek
10
east
13
Shark Pond
11
island
14
Highway Creek Big Island
12
west
15
Critter Pond*
16
northeast Long Sound*
Midstream
Little Blackwater
13
east
17
northeast L. Blackwater Sd.*
Sound
14
mid
18
northwest L. Blackwater Sd.*
15
west
19
L. Blackwater Sd. island
20
south L. Blackwater Sd.
Downstream
Blackwater Sound
16
near Gilberts
21
trestle
17
mid
22
Gilberts
18
far
23
hydrostation
24
Bush Point
* Stations that were dropped from the analysis due to missing data as a result of numerous
poor visibility days primarily in 1990.
to
to

23
Sites. Because enclosure net stations were located in
open bays, sites were selected that were protected from
prevailing direct winds. Such siting prevented the wind
from pulling the bottom of the net off the substrate. Sites
were further chosen to have between 20 and 100 cm mean water
depth at the outer prop root edge. This criterion was
intended to provide some uniformity among the sites in terms
of volume of water enclosed in the net. At each site, a
natural berm consisting of packed detritus approximately 15
cm high and 30 cm wide occurred along the landward edge.
This berm was exposed at high tide and provided a bank
beyond which fish could not escape when rotenone was applied
within the net (see below).
Procedures. At the start of the study, three sites
were selected in each of the six general locations. A
maximum of three enclosure nets could be deployed at each
location by two persons in a day if the nets were deployed
no more than approximately 1.0 km from each other.
Environmental measurements taken during each collection
included water depth, salinity, temperature, wind speed,
wind direction, and air temperature. Salinity and water
temperature were measured with a calibrated electronic
instrument (YSI Model 33 S-C-T meter). For salinities above
35 ppt, a calibrated hand-held refractometer was used.
For each net, two 30 cm wide paths were cut
perpendicular to the shoreline through the mangrove fringe
back to the berm. The paths were cleared of bottom roots

24
and overhanging branches so that a person could walk up the
path carrying one end of the 30 m long net.
The same site was sampled repeatedly throughout the
study unless "stress" was observed in submerged vegetation.
For example, sites with clay sediments supporting seagrasses
had suffered some visible damage (e.g. grass trampling) from
the sampling procedure by the fourth month at three
stations. To maintain consistency in types of habitat
encompassed by the nets, at these three sites, one new path
was cut so that an unimpacted site could be sampled adjacent
to the old one. These minor site changes were taken into
account in later calculations of net area sampled.
On the day of sampling, a 6.0 mm mesh nylon seine was
deployed by two people who carried it, scrolled around two
wooden dowels, to the mid-point between the two prepared
paths. Starting 10 m from the edge, they waded in opposite
directions parallel to the edge, unrolling the net, and then
walked toward the mangroves and up the paths. The dowel end
was pounded into the sediment at the landward end of the
path and the lead line was pressed down into the sediments
all around the bottom edge. The top edge of the net was
hung over several PVC poles to prevent fish from jumping
over the net. All three nets were set in similar fashion
(Figure 2-2).
Liguid rotenone was then applied within the enclosed
area to a final concentration of 5 mg L1. Fish that
immediately began to surface were collected using hand nets

25
Figure 2-2. Enclosure net illustration indicating net
dimensions.

26
for 30 to 45 minutes by two persons. After repeating the
process at the other two sites, and allowing the rotenone to
dissipate (approximately 3-4 hours), a snorkeler retrieved
sunken fish and a wader collected along the berm edge from
within each enclosure. Since rotenone effectiveness is
reduced with decreasing temperature (Neilson & Johnson
1983), in colder months, floating fish were also collected
after leaving the nets up overnight.
Fish and invertebrates collected were initially placed
on ice and then frozen. Later they were identified to
species and measured to total and standard (or carapace)
length.
Efficiency tests. Fish recovery efficiency was tested
at least once at each location using a mark-recapture method
and normal net procedures. To collect test fish, several
minnow traps were placed inside the area to be enclosed on
the day before net deployment. After the net was in place,
the minnow traps were removed, cleared and the fish placed
into buckets. Fish were measured, marked by fin-clipping,
and returned to the enclosed net area in minimal time. At
least 30 fish were used per net in the test.
Visual Census
Based on the pilot study, direct recording of census
data on underwater paper was identified as the better visual
method tested. Quality of video tape was inconsistent and
often low under the variable turbidity conditions
encountered.

27
Study Sites. For visual censusing, site selection
criteria included adequate depth (20 to 100 cm), red
mangrove dominance, and wind-protection. At 70 m in length,
however, each transect encompassed a range of depths and
other physical characteristics. Four sites were randomly
selected for permanent sampling stations within each of the
six general locations. These sites were chosen in the same
general locations as the net sites but were more widespread
than the net stations (Figure 2-1).
Procedures. To prepare permanent census stations,
mangrove edge transects were designated with flagging tape
every 10 m along the 70 m edge. Physico-chemical variables
were recorded in conjunction with each visual census (e.g.
air temperature, salinity, wind conditions). Other
variables measured included: 1) water depth was recorded at
a permanent stake located in each transect; 2) abundance of
submerged aquatic vegetation adjacent to the transects was
noted using a scale of zero to three (abundant); and, 3)
range of visibility was determined by using a white PVC pole
set vertically into the mud and measuring the horizontal
distance at which the pole became visible as one snorkeled
towards it. If the visibility was less than 100 cm, the
census was rescheduled. If visibility was initially poor,
three attempts (on subsequent days) were made to conduct
surveys. However, in some months it was impossible to
conduct a visual census at particular sites.

28
To conduct the census, a snorkeler approached the
flagged edge and remained stationary under each flag for 30
seconds, an adeguate period for recording observable fish.
On underwater data sheets, they recorded the species,
numbers and estimated sizes of fish observed. The census
surveys were conducted by myself and one assistant. Each
census consisted of four complete swims of each transect.
Efficiency tests. Efficiency tests were conducted for
the visual census technique using a mark-recapture method.
Large fish (18-25 cm) were caught using hook and line and
smaller fish (less than 18 cm) were collected with minnow
traps. Tags made of plastic waterproof tape in various
colors and labeled with a unique code were used. Fishing
line was securely fastened to the tag and the line was sewn
with a small sewing needle through the flesh just under a
fish's dorsal fin (for gray snapper and larger fish) or
through the lower jaw (for smaller fish).
To conduct a test, block nets were used to enclose an
area of mangrove shoreline of adequate size to accommodate
at least two snorkeling stations. Tagged fish (six to nine
larger fish and ten smaller fish) were placed inside the
enclosure for several hours to allow them to acclimatize to
the habitat. Snorkelers then carefully entered the enclosed
areas and conducted a visual census of the site by recording
the species, tag color and code of each fish they observed.
Three such tests were conducted in the summer of 1990, at
three separate stations.

29
Definitions
As presented in Chapter 1, three categories of
residency are recognized by estuarine ecologists:
residents, transient juveniles, and occasional visitors. In
analyzing the community in northeastern Florida Bay, several
sources were consulted for life history information on
individual species to designate each by residency (e.g. Odum
& Heald 1972, Lee et al. 1980, Yanez-Arancibia 1980, Robins
et al. 1986). Without conducting specific gonad analysis to
determine maturity (e.g. Robertson & Duke 1990b),
unequivocal distinction between juveniles and adults in the
transient category were not possible. In addition, life
history information is sketchy for all species except for
certain killifish. Thus, these designations are approximate
and serve for discussion purposes only. Such designations
were not used in statistical analyses of the fish community.
For purposes of detailed analysis, all fish were
assigned to one of three groups of species based on size,
behavior and primary portion of the mangrove habitat
occupied during the day. Forage fish were considered those
species whose members were generally less than 15 cm in
size. Two groups of forage fish occupied different portions
of the mangrove habitat: benthic and water column. Benthic
forage fish live in close association with the substrate and
include such species as gobies, killifish and mojarras.
Water column forage fish are exclusively schooling fishes,
that occupy the upper water column habitats, including

30
anchovies and silversides. The third fish group, large
roving fish, are generally greater than 15 cm in size and
occupy both the bottom and water column locations. This
group included such species as snook, tarpon, snappers,
catfish, grunts and barracuda.
Density determinations
To obtain densities for each enclosure net sample,
abundance was divided by the area the net encompassed.
Areas enclosed ranged from 72 to 196 m (mean = 119, sd =
33.2) .
For visual census samples, measurements of horizontal
secchi distance and fringe width were used as radii in
calculating the area observed (half the area of a circle).
Because it was impossible to see and accurately identify
small fish at a great distance, maximum radius values of
2.0, 3.0, or 4.0 m were applied to benthic, water column and
large roving fish respectively. Thus, the areas sampled in
the visual censuses ranged from 1.6 and 25.0 m2 at each of
the eight stations along a 70 m long transect.
Analysis Methods
Temporal patterns in density. To initially inspect the
data for patterns, salinity and density by fish group were
graphed. Temporal patterns were examined graphically and
guantitatively. For each of the three fish groups, and the
three species that were most abundant within each group,

31
correlations were calculated between monthly averages of
these fish densities and corresponding values for salinity,
water temperature, and water depth.
Spatial patterns in density. To discern spatial
patterns in distribution for each fish group and the top 3
species within each group, repeated measures analyses of
variance (ANOVA) with multiple comparison tests were used
(SAS GLM procedure). All density data was effectively
normalized by log-transformation. In the initial ANOVA
model, density of fish was the dependent variable and
gradient position (up-, mid- and downstream) and system
(east and west) were the independent variables. To explore
the relative density patterns among the general locations, a
second ANOVA with general location as the independent
variable was conducted. Further analysis was conducted to
determine if spatial variation in certain environmental
parameters might indicate why the densities varied among the
stations. The average values of fish density, water depth,
and salinity for each station (n = 18 stations) were
calculated. Additionally, the amount of salinity variation
over time at a particular station was also calculated by
determining the standard deviation of salinity.
Correlations between average fish densities and the means
for these parameters were then determined.
Community patterns. For comparisons among general
locations, an index of species richness (Odum 1983) was
calculated with total number of species as the numerator and

32
log-transformed abundances as the denominator. Actual fish
assemblage patterns were compared to gradient positions
using cluster analysis. Data for each station, date and
species were used to form matrices of stations based on
similarity values (SAS CLUSTER procedure). An average
linkage method was used to join clusters of stations. The
resulting dendrograms were compared with the gradient
positions. Those stations that were placed in a group other
than the correct up-, mid- or downstream position, were
denoted as misclassified. A second analysis was conducted
on the log-transformed densities at each station in order to
more thoroughly explore the data.
Results
Tests of Recovery Efficiency
Results of recovery efficiency tests for both the
enclosure net and visual census technigues measured the
number of fish sampled out of the total that were at a site
(Table 2-2). However, no estimate is available for either
method for sample accuracy, i.e. for how many fish escaped
the area as the net was being deployed or the observer
approached the area.
Enclosure net efficiencies. In all tests spanning up-,
mid-, and downstream locations, 492 fish were marked. Of 14
total species, 60% of the fish used in the tests were
goldspotted killifish (Floridichthys carpi). An average of
18% of all fish were recovered in the initial dip-net
collections. By adding the same-day snorkeling procedure,

Table 2-2. Efficiency test results for enclosure nets and visual census
sampling obtained by mark-recapture tests.
Method
Number
of Trials
Size Classes
Total Length
(centimeters)
Number
of Fish
Tagged
Mean
Percent
Efficiency
Standard Deviation
Percent Efficiency
Enclosure net
18
2.5 to 7.5
467
36
38
18
7.5 to 15
25
68
31
1
15 to 25
6
100
Not applicable
Visual census
3
4 to 7
31
29
2
3
18 to 25
22
86
12

34
efficiency increased by 7%. The total mean recovery rate
was increased to 37% by leaving the nets up overnight and
collecting the next day.
Overall, a greater percentage of larger fish were
recovered than smaller (Table 2-2). Of six large fish
(Lutjanus griseus) that were tagged, all were recovered
after rotenone application. Twice as many mid-sized as
small fish were recovered.
Visual census efficiencies. Individual test results for
small fish (all Floridichthys carpi) ranged from 25 to 27%
efficiency for the visual censusing method (Table 2-2). For
large fish (all Lutjanus griseus), results ranged from 78 to
100%. Several tagged fish were observed more than one time
during the four swims along the transect. Thus, when
analyzing the data for each sample, to prevent counting the
same fish more than once, after recording the first swim,
only unique species and size classes of fish were added to
the dataset for the second, third and fourth swims.
These efficiency analyses were intended to identify
trends in fish recovery rates. Due to wide ranges in the
test results, subsequent data analyses were not corrected
for efficiencies.
Overall Abundance
Results of the visual census differed from enclosure
net sampling results (Table 2-3). Enclosure net sampling
resulted in the collection of 82,633 fish from 59 species
and 29 families. The greatest abundance was collected at

Table 2-3. Number of fish collected using enclosure nets and observed during
visual censuses. Explanation of group/residency given at end of table.
Group/
Residency
Family Species
General Locations
Total
Nets Visual
Up-west
Nets Visual
Mid-west
Nets Visual
Down-west
Nets Visual
Up-east
Nets Visual
Mid-east
Nets Visual
Down-east
Nets Visual
2
1
1 2
4
4
1
2 1
2 2
6
6
1
1
3
1
1
2
7
11
3
151
165
3 600
3 600
335 41
40
2 100
454 2
17,770
4
18,605 143
5
5
4
1 551
2
2 3
9 554
38
59
134
23
275
529
1
1
1
1
17
17
1
1
Carcharhinidae (requium shark)
LR/r Carcharhinus laucas
Orectolobidae (nurseshark)
LR/r Ginglymostoma cirratum
Dasyatidae (stingray)
LR/r Dasyatis sabina
Elopidae (tarpon)
LR/o Megalops atlanticus
Anguillidae (freshwater eel)
LR/o Anguilla rostrata
Clupeidae (herring)
WC/o Opisthonema oglinum
WC/o Clupaid (species unk)
WC/o Harengula jaguana
Engraulidae (anchovy)
WC/r Anchoa mitchelll
WC/r Anchoa cayorum
Ariidae (sea catfish)
LR/r Arius felis
Batrachoididae (toadfish)
BF/r Opsanus beta
BF/r Porichthys plectrodon
Gobiesocidae (clingfish)
BF/r Gobiesox strumosus
Bythitidae (viviparous brotula)
BF/r Ogilbia cayorum
BF/r Gunterichthys long ¡penis
continued

Table 2-3, continued
Group/
Residency
Family Species
General Locations
Total
Nets
Visual
Up-west
Nets Visual
Mid-west
Nets Visual
Down-
Nets
west
Visual
Up-east
Nets Visual
Mid-east
Nets Visual
Down-
Nets
east
Visual
Belonidae (needlefish)
LR/r
Strongylura notata
177
S3
196
407
223
478
195
95
78
219
90
601
959
1,853
LR/r
Strongylura timucu
1
1
Cyprinodonditae (killifish)
BF/r
Floridichthys carpi
1,792
676
5,011
1,294
2,183
1,095
2,127
1,473
1,056
227
849
1.475
13,018
6,240
BF/r
Lucania parva
964
2,469
924
617
1,779
1.821
1,576
3,310
389
78
4,605
3,079
10,237
11,374
BF/r
Cyprinodon variegatus
221
768
144
86
79
1
308
957
351
21
1,124
1,812
BF/r
Fund ulus grandis
309
252
99
15
97
7
65
31
147
76
793
305
BF/r
Fundulus confuentus
98
63
91
16
47
40
199
30
481
103
BF/r
Minia xenica
25
21
46
BF/r
Fundulus simiiis
2
1
10
13
BF/r
Lucania goodei
1
1
BF/r
Rivulus marmoratus
1
1
Poecillidae (livebearers)
BF/r
Poecilia iatipinna
1,463
2.820
4,395
1,254
1,006
2,440
257
1,535
2,981
149
898
544
11,000
8,742
WC/r
Gambusia sp.
26
467
325
937
293
2,921
31
148
583
210
649
449
1,907
5,132
WC/r
Beionesox belizanus
4
1
2
2
8
1
Atherinidae (sllverside)
WC/r
Atherinomorus stipes
397
450
630
31,260
7,078
25,187
1,010
89
4,352
2,848
58,491
11,042
120,750
WC/r
Atherinidae (genus unk)
900
14,581
887
272
1,681
2,369
0
20,690
WC/r
Menidia sp.
2,242
543
175
3,093
357
256
1,071
1,815
325
200
178
3,222
4,348
9,129
WC/r
Membras martinica
1
1
Syngnathidae (pipefish)
BF/r
Syngnathus scovelli
43
83
19
12
6
53
80
213
83
BF/r
Syngnathus floridae
1
8
11
20
BF/r
Micrognathus criniger
6
6
BF/r
Hippocampus erectus
2
2
4
BF/r
Hippocampus zosterae
1
2
3
continued

Table 2-3, continued
Group/
Residency
Family Species
General Locations
Total
Nets
Visual
Up-west
Nets Visual
Mid-west
Nets Visual
Down-
Nets
west
Visual
Up-east
Nets Visual
Mid-east
Nets Visual
Down-
Nets
east
Visual
Centropomidae (snook)
LFVo
Centropomus undecimalis
2
1
2
25
1
3
24
5
53
Echeneididlae (remoras)
LR/o
Echeneis naucrates
1
1
Carangidae (Jack)
LR/o
Trachinotus goodei
1
1
LR/o
Naucrates ductor
2
2
LR/o
Carangidae (sp unk)
4
4
LR/o
Caranx hippos
1
6
1
1
2
11
LR/o
Carangidae (juv.)
2
2
LR/o
Trachinotus falcatus
1
1
Lutjanidae (snapper)
LR/t
Lutjanus jocu
23
12
12
27
74
LR/ta
Lutjanus griseus
13
74
1
4,737
41
4,756
10
11
2
1,157
29
7,726
96
18,461
LR/t
Lutjanus apodus
1
209
1
129
49
69
2
456
Gerreidae (mojarras)
BF/r
Eucinostomus sp
4
2
13
19
BF/r
Eucino8tomu8 harengulus
349
32
212
128
67
446
396
135
505
458
68
882
1,597
2,081
BF/r
Eug erres plumieri
427
302
37
2
88
212
152
12
704
528
BF/r
Eucinostomus guia
2
1
241
1,289
104
1,739
1
30
184
33
2,350
410
5,564
BF/r
Gerres cinereus
39
7
26
34
164
18
3
116
143
52
900
266
1,236
Haemulidae (grunts)
LR/t
Haemuiidae (sp unk)
1
1
LR/t
Haemulon parrai
26
102
11
139
LR/ta
Haemuion sciurus
16
1,112
56
1,584
2,768
Sparldae (porgies)
LR/o
Lagodon rhomboides
2
20
6
28
LR/o
Archosargus rhomboidalis
52
52
LR/ta
Archosargus probatocephalus
41
114
8
163
continued

Table 2-3, continued
Group/
Residency
Family Species
General Locations
Up-west
Nets Visual
Mid-west
Nets Visual
Down-west
Nets Visual
Up-east
Nets Visual
Mid-east
Nets Visual
Down-
Nets
east
Visual
Total
Nets
Visual
Lepisosteldae (gar)
LR/o
Lepisosteus platyrhincus
1
1
Centrarchidae (sunfish)
BF/r
Lepomis macrochirus
7
1
20
1
27
Cichlidae (cichlid)
BF/r
Cichlasoma urophthalmus
85
387
7
977
90
1,062
484
BF/r
Tilapia mariae
3
3
5
3
8
Ephippidae (spadefish)
LR/o
Chaetodipterus faber
3
15
18
Scaridae (parrotfish)
LR/o
Sparisoma radians
2
1
3
LR/o
Scaridae (sp unk)
3
1
4
Lobotidae (tripletail)
LR/o
Lobotes surinamensis
2
1
3
Mugilidae (mullet)
LR/o
Mugil cephalus
11
90
6 12
45
1
3
10
1
1211
29
1361
LR/o
Mugil curema
1
22
1
24
LR/o
Mugil liza
1
1
5
0
7
Sphyraenidae (barracuda)
LR/ta
Sphyraena barracuda
12
1
17 196
18 185
3
1
29 90
25
352
104
825
Bleniidae (combtooth blenny)
BF/r
Chasmodes saburrae
6
3
1
2
2
14
Gobiidae (goby)
BF/r
Microgobius gulosus
1,011
675
92
125 33
1,648
81
110 6
11
9
2,997
804
BF/r
Gobiosoma robu stum
235
239
1
59
534
BF/r
Lophogobius cyprinoides
151
18
1
18
44
30 3
1
200
66
BF/r
Gobiosoma bosci
124
1
125
continued
OJ
00

Table 2-3, continued
Group/
Residency
Family Species
General Locations
Total
Nets
Visual
Up-west
Nets Visual
Mid-west
Nets Visual
Down-west
Nets Visual
Up-east
Nets Visual
Mid-east
Nets Visual
Down-
Nets
east
Visual
Acanthuridae (surgeonfish)
LR/o
Acanthurus chirurgus
9
9
Balistldae (leatherjackets)
LR/o
Aluterus scriptus
1
1
Soleidae (sole)
BF/r
Trinectes maculatus
7
1
28
36
Tetraodontidae (puffer)
LR/o
Sphoeroides spengleri
1
1
2
Diodontidae (spiny puffer)
LR/o
Chilomycterus schoepf
1
1
LR/o
Diodontidae (species unk)
1
14
15
Total
10,624
11,185
12,723
60.275
13,712 44,703
9,547
11,301
25,115
9,280
10,912
86,212
82,633
222,960
No. Species
37
27
22
27
32 33
32
26
36
18
35
31
59
51
No. Samples
36
35
36
44
36 50
36
33
36
23
36
48
216
233
* Groups: BF = Benthic forage fish
WC = Water column forage fish
LR = Large roving fish
/Residency: r
o
t
ta
resident
occasional visitors
estuarine transient juvenile
estuarine transient juvenile (also present as adults)

40
Little Blackwater Sound, the midstream-east location. The
greatest number of species, however, was found in samples
from Joe Bay, the upstream-west location.
Visual census sampling resulted in observation of
222,960 fish from 51 species and 31 families (Table 2-3).
Greatest abundance and greatest number of species were
observed in samples taken in Blackwater Sound.
Samples obtained by the two methods differed in
relative abundance and numbers of species within these three
fish groups (Table 2-4). For example, many more species of
benthic forage fish were collected in the enclosure nets
(33) than were observed in the visual census (16). In
contrast, many more large roving fish species were sampled
in the visual censuses (29) than in the enclosure nets (17).
Temporal Patterns in Density by Fish Group
Benthic forage fish. In Figure 2-3, one can compare
changes in salinity with changes in density from the
enclosure net sampling; however, no consistent patterns
emerge. Great density variations occur independently of
salinity changes. Salinity varied widely over the study
period at the upstream/east (0.0-39.0 ppt) and upstream/west
(13.0-58.0 ppt) locations. Salinity also ranged widely at
the midstream/east location (19.5-50.0 ppt). However, at
the other three locations (downstream/west, downstream/east
and midstream/west), salinity remained high (29.8 to 54.0
ppt) throughout the study. Not only was the period of low
salinity longer in the upstream/east location, but also, a

Table 2-4. Summary of abundances and number of
species by method of collection and fish group.
Fish
Group
Methods
Parameter
Enclosure
Nets
Visual
Census
Benthic Forage
Fish
Total
No. Species
45,458
33
39,476
16
Water Column
Forage Fish
Total
No. Species
35,926
9
156,610
6
Large Roving
Fish
Total
No. Species
1,249
17
26,874
29
All Fish
Total
No. Species
82,633
59
222,960
51

Figure 2
habitats
indicate
location
I
CD
-t1
CD
E
CD
i_
O
D
CT
cn
(D
Q_
_C
CO
L_
Benthic Forage Fish
Density & Salinity vs. Month
Enclosure Nets
West
East
T>
Q
-i
H-
(/)
X)
0)
ir
o
c
in
a
3
Q.
May89
May90 Moy89
May90
3. Density of benthic forage fish collected with enclosure nets in mangrove
(histogram) and corresponding salinity measurements (lines). Error bars
standard deviation among the three enclosure nets deployed in each general
fO

43
substantial decrease (from 35.0 ppt to 10.4 ppt) was evident
in June 1990, that did not occur in stations sampled in the
upstream/west location (which became increasingly
hypersaline).
None of these salinity changes correspond with patterns
observed for fish densities. Density of benthic forage fish
peaked in winter months at four of the six general locations
(Figure 2-3). The highest density collection (13.6 fish m-
O ...
) was at the mid-Trout Cove station in winter 1989; lowest
density occurred at the mid-Little Blackwater Sound in the
June 1989 (0.12 fish m-2) .
Water column forage fish. In Figure 2-4, one can
compare changes in salinity with changes in density from the
enclosure net sampling; again, however, no consistent
patterns emerge. Density of water column fish was highly
variable and the graphs illustrate no consistent seasonal
patterns. In general, either very low or very high
densities of these schooling fishes were collected. The
highest density collection (25.3 fish m-2) occurred at mid-
Little Blackwater Sound in September 1989. No water column
forage fish were collected in several samples. As with the
benthic forage fish, these density fluctuations were also
not related to the seasonal fluctuations in salinity.
Large roving fish. In Figure 2-5, changes in salinity
can be compared with changes in density for this group from
the visual census sampling; again, however, no consistent
temporal patterns emerge. In the upstream/west location,

Water Column Forage Fish
Density & Salinity vs. Month
Enclosure Nets
O
o
(/)
O
CD
ir
o
c/)
Q
D
CL
Figure 2-4. Density of water column forage fish collected with enclosure nets in
mangrove habitats (histogram) and corresponding salinity measurements (lines). Error
bars indicate standard deviation among the three enclosure nets deployed in each
general location.

1.6
West
I
CD
-4-^
CD
E
0)
L-
o
D
CT
CO
L_
0)
CL
1~
V)
L_
1.2
0.8
0.4
0.0
1.6
1.2
0.8
0.4
0.0
Large Roving Fish
Density Sc Salinity vs. Season
Visual Census
East
60
40
20
0
60
40
20
0
Figure 2-5. Density of large roving fish sampled by visual census in
mangrove habitats (histograms) and corresponding salinity measurements
(lines). Error bars indicate standard deviation among the visual censuses.
Ui
Parts per Thousand

46
the visual census stations were notably fresher (1.4-50.0
ppt) than the corresponding net stations (13.0-58.0 ppt)
(Figures 2-3 & 2-5). In this location, the visual census
stations were located within the creek itself while the net
stations were located in Joe Bay immediately at the creek
mouth (Figure 2-1). These seemingly slight differences in
location may have contributed to observed differences in
salinity patterns for the visual and net stations.
For the visual census, the salinity patterns in both
the eastern and western upstream locations were similar to
each other except in spring 1990, when freshwater entered
the upstream/east location reducing salinity to 30.0 ppt and
the upstream/west location became strongly hypersaline (47.3
ppt) (Figure 2-5). The midstream/east location was overall
more variable than the midstream/west location. At the
mid/west and downstream visual census locations, salinity
patterns were uniformly high.
Density of large roving fish ranged from zero at the
uppermost upstream/west location in the spring, fall and
winter of 1989, to 2.3 fish m-2 at Duck Key
(midstream/west) in winter. From the graphs, it appears
that changes in density of this group were independent of
salinity changes (Figure 2-5) .
Temporal correlations. No significant correlations
between temporal changes in salinity or water depth, and
temporal changes in fish densities were found (Table 2-5).
However, changes in water temperature were correlated with

Table 2-5. Correlations between fish density for each month averaged over
the stations and salinity, water temperature and water depth. Data
for large roving fish were obtained by visual censuses. Data for benthic
and water column forage fish were obtained with enclosure nets. All density
data were converted to logarithms (log x + 1) prior to calculations.
Significant (p < 0.05) correlations are underlined.
Species or
Category
Salinity
Water
depth
Water Temperature
correlation
p-value
correlation
p-value
correlation
p-value
Benthic Forage
Fish
+0.17
0.5769
+0.05
0.8721
-0.80
0.0010
Water Column
Forage Fish
+0.06
0.8437
+0.57
0.4350
-0.04
0.8986
Large Roving
Fish
+0.31
0.2996
-0.35
0.7145
-0.76
0.0028
4^
-J

48
density of benthic forage fish and large roving fish. In
both cases, lower abundances occurred at higher water
temperatures.
Spatial Patterns in Density by Fish Group
Spatial patterns in fish density varied among the fish
groups (Figure 2-6). From these graphs, one can see that
only the larger roving fish group seems to vary consistently
along the salinity gradient, with much lower densities at
the upstream locations.
Benthic forage fish analysis of variance. Results of
the repeated measures ANOVA's by fish group differed among
the fish groups (Tables 2-6 and 2-7). Neither gradient
position nor system were important determinants of variation
in densities among the stations for the benthic forage fish
group (Table 2-6). Although densities tended to vary
significantly from one general location to another, these
variations were not systematic along the salinity gradient,
as indicated by the significant interaction between gradient
and system.
The mid/west general location had significantly greater
densities than the other midstream location (Table 2-7,
Figure 2-6). Other locations were intermediate and not
significantly different from these two.
Water column forage fish analysis of variance. Again,
although densities tended to vary significantly from one
general location to another, these variations were not
systematic along the salinity gradient, as indicated by the

Figure 2-6. Mean density of fish by general location for
each fish group. Error bars illustrate the magnitude of the
standard deviation in density over all the months. Samples
of benthic and water column forage fish taken with enclosure
nets. Large roving fish were sampled with visual methods.

Fish per m2 Fish per m2
Density by General Location
1.20
Large Roving Fish
CM
1.00
0.80
a) 0.60
CO
0.40
0.20
0.00
.ll.ll
Up Mid Down Up Mid Down
West East

Table 2-6. Repeated measures analysis of variance with densities of fish as dependent
variables and gradient and system as independent variables. Benthic and water column
forage fish were collected using enclosure nets. Large roving fish were sampled using
visual census techniques. Data were transformed to logarithms prior to performing
calculations.
Source
Benthic forage
fish
Water Column Forage Fish
Large
Roving
Fish
df*
F
P
df*
F
P
df*
F
P
Between Stations:
Among gradient positions
2/12
0.16
0.8511
2/12
4.17
0.0421
2/12
6.30
0.0135
Among systems
1/12
3.39
0.0904
1/12
0.38
0.5514
1/12
0.22
0.6460
Gradient X System
2/12
7.56
0.0075
2/12
8.39
0.0052
2/12
0.65
0.5393
Within stations
Among months
10/120
9.83
0.0001
10/120
5.16
0.0001
4/48
2.17
0.0863
Month X Gradient
20/120
1.15
0.3131
20/120
4.92
0.0001
8/48
1.29
0.2707
Month X System
10/120
1.48
0.1627
10/120
1.5
0.1486
4/48
1.56
0.2015
Month X Gradient X System
20/120
1.07
0.3942
20/120
4.61
0.0001
4/48
0.97
0.4707
Multiple comparisons among
means for all months:
Location
Sign. Location
greater
than
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Gradient positions
No differences
Down
>
Up
Mid & Down
>
Up
Systems
No differences
No differences
No differences
Source degrees of freedom / error degrees of freedom

Table 2-7. Repeated measures analysis of variance with densities of fish as dependent
variables and general locations as independent variables. Benthic and water column
forage fish were collected using enclosure nets. Large roving fish were sampled using
visual census techniques. Data were transformed to logarithms prior to performing
calculations.
Source
Benthic forage fish
Water Column Forage Fish
Large
Roving Fish
df*
F
P
df*
F
P
df*
F p
Between Stations:
Among general locations
5/12
3.77
0.0277
5/12
5.1
0.0097
5/12
2.94 0.0586
Within stations
Among months
10/120
9.53
0.0001
10/120
5.16
0.0001
4/48
2.17 0.0863
Month X General locations
50/120
1.18
0.2360
50/120
4.11
0.0001
20/48
1.16 0.3249
Multiple comparisons among
means for all months:
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Location Sign. Location
greater
than
2
>
5
(others
inter
mediate)
5 & 3
>
4 & 2
(others
inter
mediate)
No differences
Source degrees of freedom / error degrees of freedom
** General locations:
1 = Joe Bay, upstream/west
2 = Trout Cove, midstream/west
3 = Buttonwood Sound, downstream/west
4 = Highway Creek, upstream/east
5 = Little Blackwater Sound, midstream/east
6 = Blackwater Sound, downstream/east
to

53
significant interaction between gradient and system for
density of water column fish (Table 2-6). Little Blackwater
Sound (mid/east) and Buttonwood Sound (down/west) had
significantly greater densities than Highway Creek (up/east)
and Trout Cove (mid/west) (Table 2-7, Figure 2-6).
Large roving fish analysis of variance. In contrast to
the other two groups, for large roving fish, a clear effect
of gradient position on fish density occurred. Fish in this
group were significantly less abundant at the upstream
gradient locations than mid- or downstream (Table 2-6). No
general locations varied significantly from the others
(Table 2-7).
Spatial correlations. To analyze spatial trends,
correlations between mean fish densities and salinity,
salinity variation, and water depth were determined (Table
2-8). A significant correlation between average density of
large roving fish and station salinity was found; lower
densities occurred at stations with lower mean salinity
levels and greater temporal variability. In addition, both
water column forage fish and large roving fish were
significantly more abundant at stations with deeper water.
Temporal Patterns in Density by Species
As indicated by the correlations between mean densities
and salinity, water depth and water temperature, temporal
patterns differed among the species (Table 2-9). No
significant correlations were found between salinity changes
from month to month and densities for any species. Temporal

Table 2-8. Correlations between fish density for each station (n=18) averaged
over the months and salinity, temporal standard deviation of salinity and water depth.
Density data were converted to logarithms (log x + 1) prior to calculations.
Significant (p<0.05) correlations are underlined. Benthic and Water
column forage fish were collected with enclosure nets. Large roving fish
were sampled using visual techniques.
Category
Salinity
correlation p-value
Temporal Standard
Deviation of Salinity
correlation p-value
Water Depth
correlation p-value
Benthic Forage
Fish
-0.08
0.7540
-0.02
0.9303
-0.12
0.6463
Water Column
Forage Fish
+0.41
0.0884
-0.38
0.1233
+0.48
0.0438
Large Roving
Fish
+0.56
0.0150
-0.54
0.0208
+0.54
0.0213
U1
t*

Table 2-9. Correlations between fish densities for each month (n=13) averaged
over the stations and salinity, water temperature and water depth.
Density data were converted to logarithms (log x + 1) prior to calculations.
Significant (p<0.05) correlations are underlined.
Species
Method
Salinity
correlation p-value
Water depth
correlation p-value
Water Temperature
correlation p-value
Floridichthys carpi
Nets
+0.15
0.6117
+0.26
0.3681
-0.50
0.0788
Lucania parva
Nets
-0.06
0.8420
-0.62
0.0237
-0.69
0.0084
Poecilia latipinna
Nets
+0.49
0.0888
+0.09
0.7756
-0.72
0.0056
Anchoa mitchelli
Nets
+0.04
0.8961
+0.27
0.3620
+0.24
0.4227
Menidia spp.
Nets
+0.18
0.5638
-0.33
0.2765
+0.51
0.0740
Atherinomorus stipes
Nets
-0.14
0.6449
+0.56
0.0470
-0.14
0.6259
Lutjanus griseus
Visual
+0.47
0.1039
-0.43
0.1439
-0.92
0.0001
Strongylura notata
Visual
-0.35
0.2382
+0.76
0.0026
-0.56
0.0482
Haemulon sciurus
Visual
-0.04
0.8931
+0.56
0.0475
+0.62
0.0233

56
patterns in water temperature were significantly correlated
with densities of several species. Lucania parva, Poecilia
latipinna, Lutjanus griseus, and Strongylura notata were
less abundant when the water temperatures were higher. In
contrast, greater abundances of Haemulon sciurus were
observed in warmer months. Periods of higher water levels
in the study area (e.g. late fall) corresponded to periods
when greater densities of Atherinomorus stipes, Strongylura
notata, and Haemulon sciurus were collected. In contrast,
Lucania parva was in greater abundance during low water
periods.
Spatial Patterns in Density by Species
Analyses of variance. Density patterns varied for the
top three species of benthic forage fish (Figure 2-7).
Results of repeated measures analyses of variance also
differed among these species (Tables 2-10 and 2-11).
Poecilia latipinna was more abundant at the midstream
locations, particularly Trout Cove (mid/west).
Distributions of Floridichthys carpi and Lucania parva were
not significantly influenced by gradient position or system.
However, Floridichthys carpi was more abundant at Trout
Cove than all other locations and Lucania parva was most
abundant at Blackwater Sound (mid/east).
The top three water column forage fish species differed
in spatial distribution (Figure 2-8). Repeated measures
ANOVA results also varied among these species (Table 2-12
and 2-13). Distribution of the silversides differed

Figure 2-7. Mean density of fish by general location for the three most abundant
species in the benthic forage fish group. Error bars illustrate the magnitude of
the standard deviation in density over all the months.

Fish per square meter
Density by General Location
Top 3 Benthic Forage Fish Species
West Enclosure Nets East
2.5
2.0
1.5
1.0
0.5
0.0
2.5
2.0
1.5
1.0
0.5
0.0
2.5
2.0
1.5
1.0
0.5
0.0
Floridichthys Lucarna Poecilia Floridichthys Lucania Poecilia
carpi parva latipinna carpi parva latipinna
(ji
oo

Table 2-10. Repeated measures analysis of variance with density of benthic
forage fish as the dependent variable and gradient and system as the
independent variables. Samples were taken within mangrove habitats using
enclosure nets. Densities were transformed to logarithms prior to
calculations.
Floridichthys
carpi
Lucania parva
Poecilia latipinna
Source
df *
F
p
df*
F
P
df*
F
P
Between Stations:
Among gradient positions
2/12
3.29
0.0726
2/12
12.96
0.0010
2/12
5.96
0.0160
Among systems
1/12
33.49
0.0001
1/12
2.23
0.1610
1/12
6.00
0.0306
Gradient X System
2/12
21.21
0.0001
2/12
5.09
0.0251
2/12
1.71
0.2229
Within Stations:
Among months
10/120
7.43
0.0001
10/120
7.70
0.0001
10/120
5.73
0.0002
Month X Gradient
20/120
3.62
0.0001
20/120
1.99
0.0317
20/120
2.73
0.0002
Month X System
10/120
1.43
0.1777
10/120
1.54
0.1712
10/120
1.28
0.2827
Month X System X Gradient
20/120
2.14
0.0072
20/120
1.45
0.1561
20/120
0.93
0.5113
Multiple comparisons among
means for all months:
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Gradient positions
No differences
Down
>
Up & Mid
Mid
>
Up & down
Systems
West
>
East
No differences
West
>
East
Source degrees of freedom / error degrees of freedom
(ji
vo

Table 2-11. Repeated measures analysis of variance with density of benthic
forage fish as the dependent variable and general location as the independent
variables. Samples were taken within mangrove habitats using enclosure nets.
Densities were transformed to logarithms prior to calculations.
Floridichthys
carpi
Lucania parva
Poecilia latipinna
Source
df*
F
P
df*
F
P
df*
F p
Among general locations
5/12
16.50
0.0001
5/12
7.67
0.0019
5/12
4.26 0.0184
Among months
10/120
7.43
0.0001
10/120
7.70
0.0001
10/120
5.73 0.0002
Month X general location
50/120
2.59
0.0001
50/120
1.68
0.0314
50/120
1.72 0.0426
Multiple comparisons
among means for
all months
General
location
* *
Sign.
gr.
than
General
location
General
location
* *
Sign.
gr.
than
General
location
General
location
* *
Sign. General
gr. location
than
2
>
All
6
>
All
2
> All
3 & 4
>
5 & 6
(others
inter
mediate)
* source degrees of freedom / error degrees of freedom
** General locations:
1 = Joe Bay, upstream/west 4 = Highway Creek, upstream/east
2 = Trout Cove, midstream/west 5 = Little Blackwater, midstream/east
3 = Buttonwood Sound, downstream/west 6 = Blackwater Sound, downstream/east
cr>
o

Figure 2-8. Mean density of fish by general location for the three most abundant
species in the water column forage fish group. Error bars illustrate the
magnitude of the standard deviation in density over all the months.

4
3
2
1
O
4
3
2
1
O
4
3
2
1
O
Density by General Location
East
Top 3 Water Column Forage Fish Species
Enclosure Nets
West
Mid
i
Down
Anchoa Atherinomorus Menidia
mitchelli stipes spp.
Down
j
Anchoa Atherinomorus Menidia
mitchelli stipes spp.
o>
NJ

Table 2-12. Repeated measures analysis of variance with density of water
column forage fish as dependent variables and gradient and system as independent
variables. Samples taken within mangrove habitats using enclosure nets.
Densities were transformed to logarithms prior to calculations.
Anchoa mitchelli
Atherinomorus
stipes
Menidia spp.
Source
df*
F
P
df*
F
P
df*
F
P
Between Stations:
Among gradient positions
2/12
10.95
0.0020
2/12
116.61
0.0001
2/12
13.44
0.0009
Among systems
1/12
13.41
0.0031
1/12
20.32
0.0007
1/12
1.62
0.2272
Gradient X System
2/12
11.32
0.0017
2/12
2.81
0.1001
2/12
0.51
0.6136
Within Stations:
Among months
10/120
4.73
0.0060
10/120
8.54
0.0001
10/120
3.05
0.0065
Month X Gradient
20/120
4.95
0.0007
20/120
5.46
0.0001
20/120
2.62
0.0033
Month X System
10/120
4.95
0.0047
10/120
2.46
0.0389
10/120
4.44
0.0006
Month X System X Gradient
20/120
4.43
0.0015
20/120
3.00
0.0031
20/120
2.70
0.0025
Multiple comparisons among
means for all months:
Location Sign. Location
greater
than
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Gradient positions
Mid
>
Up & down
Down
>
Up & Mid
Up
>
Mid & down
Systems
East
>
West
West
>
East
No diffs
>
East
Source degrees of freedom / error degrees of freedom
o\
u>

Table 2-13. Repeated measures analysis of variance with density of
water column forage fish as the dependent variable and general location as the
independent variable. Samples were taken in mangrove habitats with enclosure
nets. Density data were transformed to logarithms for calculations.
Menidia spp.
Atherinomorus
stipes
Anchoa
mitchelli
Source
df*
F
P
df*
F
P
df*
F p
Among general locations
5/12
5.90
0.0056
5/12
51.83
0.0001
5/12
11.59 0.0003
Among months
10/120
3.05
0.0065
10/120
8.54
0.0001
10/120
4.73 0.0060
Month X general location
50/120
3.02
0.0001
50/120
3.68
0.0001
50/120
4.74 0.0001
Multiple comparisons
General
Sign.
General
General
Sign.
General
General
Sign. General
among means for
location
gr.
location
location
gr.
location
location
gr. location
all months
*
than
**
than
**
than
1
>
All
3
>
All
5
> All
except
4
6
>
All
others
* source degrees of freedom / error degrees of freedom
** General locations:
1 = Joe Bay, upstream/west
2 = Trout Cove, midstream/west
3 = Buttonwood Sound, downstream/west
4 = Highway Creek, upstream/east
5 = Little Blackwater, midstream/east
6 = Blackwater Sound, downstream/east

65
significantly among the gradient positions. Atherinomorus
stipes, the hardhead silverside, was more abundant
downstream; Menidia spp. was more abundant upstream.
Individuals of both Menidia beryllina and Menidia peninsulae
were collected. The distribution of these species overlaps
in northeastern Florida Bay, and distinctive characters are
extremely difficult to confirm (C. Gilbert, personal
communication). Thus, Menidia spp. has been used in this
study to designate these species. Anchoa mitchelli,
although not influenced by gradient or system, was
significantly more abundant at Little Blackwater Sound
(mid/east) than at the other general locations.
Patterns varied in spatial distributions for the top
three species of large roving fish (Figure 2-9). For these
species, repeated measures ANOVA results also varied (Table
2-14 and 2-15). Among these species, Haemulon sciurus was
never present upstream. Lutjanus griseus was significantly
less abundant up- than mid- or downstream. In contrast,
Strongylura notata had significantly greater densities
upstream/east.
Correlations. To further analyze spatial trends for
these nine species, correlations between mean densities and
salinity, salinity variation, and water depth were
determined and are presented in Table 2-16. Densities of
Menidia spp. were greater at locations with lower mean
salinities and greater variation. In contrast, densities of
Atherinomorus stipes were greater at locations with higher

Figure 2-9. Mean density of fish by general location for the three most abundant
species in the large roving fish group. Error bars illustrate the magnitude of
the standard deviation in density over all the months.

Fish per square meter
Density by General Location
Top 3 Large Roving Fish Species
Census East
- Up

Table 2-14. Repeated measures analysis of variance with density of large
roving fish as dependent variables and gradient and system as independent
variables. Samples taken by visual census methods along mangrove shoreline
habitats. Densities transformed to logarithms prior to calculations.
Haemulon sclurus
Lutjanus grlseus
Strongylura notata
Source
df*
F
P
df*
F
P
df*
F
P
Between Stations:
Among gradient positions
2/12
2.36
0.1367
2/12
5.67
0.0185
2/12
3.52
0.0626
Among systems
1/12
0.13
0.7278
1/12
0.20
0.6654
1/12
4.24
0.0618
Gradient X System
2/12
0.04
0.9650
2/12
0.48
0.6317
2/12
3.32
0.0714
Within Stations:
Among seasons
4/48
0.29
0.8794
4/48
9.22
0.0001
4/48
2.97
0.0416
Season X Gradient
8/48
0.73
0.6659
8/48
2.64
0.0309
8/48
2.95
0.0168
Season X System
4/48
1.97
0.1167
4/48
2.59
0.0667
4/48
0.76
0.5292
Season X System X Gradient
4/48
1.10
0.3794
4/48
1.51
0.2032
4/48
0.43
0.8603
Multiple comparisons among
means for all seasons:
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Gradient positions
No differences
Mid &
Down
>
Up
No differences
Systems
No differences
No differences
No differences
Source degrees of freedom / error degrees of freedom

Table 2-15. Repeated measures analysis of variance with density of
large roving fish as the dependent variable and general location as the independent
variable. Samples were taken by visual census along mangrove shoreline
habitats. Densities were transformed to logarithms prior to calculations.
Haemulon sciurus
Lutj anus griseus
Strongylura notata
Source
Among general locations
Among seasons
Season X general location
df F p
5/12 1.03 0.4442
4/48 0.29 0.8794
20/48 1.31 0.2190
df F p
5/12 2.26 0.0765
4/48 9.22 0.0001
20/48 1.96 0.0480
df F p
5/12 3.63 0.0311
4/48 2.97 0.0416
20/48 1.51 0.1489
Multiple comparisons
among means for
all seasons
General Sign. General
location greater location
** than **
General Sign. General
location greater location
** than **
General Sign,
location greater
** than
General
location
* *
General locations
No differences
No differences
1 & 3
(others
interm)
* source degrees of freedom / error degrees of freedom
** general locations:
1 = Joe Bay, upstream/west
2 = Trout Cove, midstream/west
3 = Buttonwood Sound, downstream/west
4 = Highway Creek, upstream/east
5 = Little Blackwater, midstream/east
6 = Blackwater Sound, downstream/east

Table 2-16. Correlations between density for each station (n=18) averaged over
all months and salinity, temporal standard deviation of salinity, and water depth.
Density data were converted to logarithms (log x + 1) prior to calculations.
Significant (p<0.05) correlations are underlined.
Species
Method
Salinity
Temporal
Deviation
Standard
of Salinity
Water
Depth
correlation
p-value
correlation
p-value
correlation
p-value
Floridichthys carpi
Nets
+0.07
0.7655
-0.12
0.6446
-0.60
0.0079
Lucania parva
Nets
+0.13
0.6096
-0.44
0.0665
+0.62
0.0064
Poecilia latipinna
Nets
+0.40
0.0935
-0.28
0.2429
-0.20
0.4165
Anchoa mitchelli
Nets
-0.03
0.9127
+0.14
0.5879
+0.05
0.8530
Menidia spp.
Nets
-0.52
0.0253
+0.69
0.0013
-0.01
0.9634
Atherinomorus stipes
Nets
+0.60
0.0087
-0.72
0.0007
+0.45
0.0626
Lutjanus griseus
Visual
+0.58
0.0108
-0.58
0.0123
+0.53
0.0226
Strongylura notata
Visual
-0.11
0.6595
-0.06
0.8234
+0.22
0.3765
Haemulon sciurus
Visual
+0.37
0.1300
-0.38
0.1226
+0.26
0.2942

71
mean salinities and less variation (Table 2-16). Lutjanus
griseus also appeared to avoid the lower salinity, more
variable areas. Floridichthys carpi was significantly more
abundant at shallower stations in the study area. Species
correlated with deeper waters were Lucania parva, Lutjanus
griseus and possibly, Atherinomorus stipes.
Community Patterns
Species richness. Species richness differed among the
fish groups and for all fish combined (Table 2-17). A total
of 305,589 fish from 77 species was sampled using both the
visual census and enclosure net methods combined. Midstream
locations in both the east and west systems had the lowest
species richness of the three gradient positions. Among the
fish groups, benthic forage fish were also least species
rich at midstream locations. Water column forage fish had
distinctly lower species richness at Little Blackwater
Sound, with only four species but great abundances. For
large roving fish, upstream and midstream locations were
lowest. For number of species alone, large roving fish
species followed a clear gradient from upstream (10 and 11
species) to midstream (13 and 14 species) to downstream (20
and 22 species).
Cluster analysis. Results of two cluster analyses for
species collected with enclosure nets were illustrated using
dendrograms (Figure 2-10). The initial analysis classified
stations based on presence of species. Most of the upstream

Table 2-17. Species richness index for fish sampled with enclosure nets and
visual census. Species Richness Index =
(Number of species 1) / log (Total individuals) (Odum 1983).
General
Location
Benthic Forage
Fish
Water Column Forage
Fish
Large
Roving
Fish
All Fish
Total
Number Index
Total
Number Index
Total
Number
Index
Total
Number
Index
Indiv.
of
Indiv.
of
Indiv.
of
Indiv.
of
Species
Species
Species
Species
Up-west
15,951
24
5.5
5,408
6
1.3
450
11
3.8
21,809
64
14.5
Mid-west
16,049
14
3.1
51,052
4
0.6
5,897
15
3.7
72,998
49
9.9
Down-west
13,442
21
4.8
37,083
6
1.1
7,890
20
4.9
58,415
65
13.4
Up-east
15,716
21
4.8
4,805
6
1.4
327
10
3.6
20,848
58
13.2
Mid-east
7,434
19
4.6
25,222
7
1.4
1,739
13
3.7
34,395
54
11.7
Down-east
16,342
22
5.0
68,966
6
1.0
11,816
22
5.2
97,124
66
13.0
All
84,934
33
6.5
192,536
9
1.5
28,123
35
7.6
305,589
77
13.9
to

Figure 2-10. Cluster analysis dendrograms based on species collected using
enclosure nets. Stations that grouped differently than actual gradient positions
are designated as misclassified.
a. Presence of each species (each species weighted equally) was used in one
analysis and,
b. Density of each species (fish per square meter) was used in the other.

Enclosure Nets
e
5
4
E 3
¡5
2
1
0
Jtatlons
Presence
Density
Up Mid & Down
= Misclassified

75
stations, including one in Joe Bay and all in Highway Creek,
grouped separately from those located mid- and downstream.
In a second analysis based on densities of each species, all
but one upstream station clustered separately from the mid-
and downstream locations.
For the visual census, cluster analysis results were
also graphed using dendrograms (Figure 2-11). Based on
presence of species, all the upstream stations formed one
cluster. Based on densities of all species, three of the
five upstream stations clustered together. These were the
three that were most upstream.
For each cluster group defined by the cluster analysis,
the most common or dominant sets of species were identified.
Three species that commonly occurred at all stations in the
study area were: goldspotted killifish (Floridichthys
carpi), rainwater killifish (Lucania parva), and redfin
needlefish (Strongylura notata). Species that were very
common in upstream stations included the inland silverside
(Menidia spp.), clown goby (Microgobius gulosus), tidewater
mojarra (Eucinostomus harengulus), striped mojarra (Eugerres
plumieri), and Mayan cichlid (Cichlasoma urophthalmus).
Downstream species commonly included hardhead silverside
(Atherinomorus stipes), gray snapper (Lutjanus griseus),
silver jenny mojarra (Eucinostomus gula), great barracuda
(Sphyraena barracuda), blue-striped grunt (Haemulon
sciurus), and snook (Centropomus undecimalis). Midstream

Figure 2-11. Cluster analysis dendrograms based on species sampled by visual
census methods. Stations that grouped differently than actual gradient positions
are designated as misclassified.
a. Presence of each species (each species weighted equally) was used in one
analysis and,
b. Density of each species (fish per square meter) was used in the other.

Distance
Visual Census
Presence
Up 1 Mid & Down
(X) Misdassified
Density
Mid & Down

78
stations had mixtures of both upstream and downstream
species excepting the Mayan cichlid, which was only common
upstream.
Residency. Abundance and numbers of species by
residency category varied among the general locations (Table
2-18). All 33 benthic forage fish species sampled were
permanent estuarine residents. Among the water column
forage fish, seven of the nine species observed were
residents, while two (Clupeids) were occasional visitors.
Large roving fish were represented in all three residency
categories. Six species of large roving fish were permanent
estuarine residents including needlefish, catfish, bull
sharks, and stingrays. The vast majority of large roving
fish were transient juveniles, however.
Frequency distribution by size varied for 6 transient
species of large roving fish (Figures 2-12 and 2-13).
Clearly, the mangrove habitats of northeastern Florida Bay
were nurseries for Sphyraena barracuda most of which
occurred in juvenile sizes (Figure 2-12). Snook, however,
did not occur in juvenile size classes. Although adult-sized
Lutjanus griseus appeared to share the habitat with larger-
sized juveniles, since no gray snapper sampled was smaller
than 7.5 cm, one can assume that young-of-the-year juveniles
occur outside the mangrove habitats sampled in this study.
Habitat use patterns similar to Lutjanus griseus were
observed for Lutjanus apodus, Haemulon sciurus, and

Table 2-18. Comparison of abundance and number of species by residency,
fish group, and general location. Both sampling methods were combined
for the table.
Residency
Fish Group
General
Up-west Mid-west Down-west Up
Locations
-east Mid-east
Down-east
All
Residents: complete entire
life cycle
in the study
area
Benthic forage fish
abundance
15951
16049
13442 15712
6965
16329
84915
Benthic forage fish
species
24
14
21
21
18
22
33
Water column forage fish
abundance
4505
36460
36195
4532
23536
65843
171071
Water column forage fish
species
5
4
5
5
6
5
7
Large roving fish
abundance
236
603
1255
293
299
701
3387
Large roving fish
species
3
1
4
2
2
4
6
Transient Juveniles: juvenile offspring of species
that spawn
offshore
Benthic forage fish
abundance
Benthic forage fish
species
Water column forage fish
abundance
Water column forage fish
species
Large roving fish
abundance
101
5268
6470
25
1396
9832
23092
Large roving fish
species
3
8
7
2
6
8
9
Occasional Visitors: marine
and freshwater adults
that occupy
the study
area
Benthic forage fish
abundance
Benthic forage fish
species
Water column forage fish
abundance
3
1
1
2
603
610
Water column forage fish
species
1
1
1
1
1
2
Large roving fish
abundance
113
25
162
9
42
1263
1618
Large roving fish
species
5
5
9
6
5
9
19
-j
VO

Figure 2-12. Length-frequency histograms based on visual
census data. Adult size given whenever the information was
available from the literature.
a. Sphyraena barracuda (great barracuda), and
b. Centropomus undecimalis (snook).

Abundance by Size Class
o
c
o
~o
c
Z3
_Q
<
0 15 30 45 60 75 90 105 120 135 150
Total length (cm)

Figure 2-13. Length-frequency histograms based on visual
census data. Adult size given whenever the information was
available from the literature for
a. Lutjanus griseus (gray snapper),
b. Lutjanus apodus (schoolmaster)
c. Haemulon sciurus (blue-striped grunt)
d. Archosargus probatocephalus (sheepshead).

1E4
8000
6000
4000
2000
0
300
200
100
0
1500
1000
500
0
60
40
20
0
Abundance by Size Class
Lutjanus
griseus
Lutjanus
J 1 L
Haemulon
10 20 30 40 50 60 70
Total length (cm)

84
Archosargus probatocephalus, although adequate life history
information is not available for these species to determine
size at maturity (Figure 2-13).
Overall, the community was dominated by residents in
both numbers of individuals (91%) and numbers of species
(60%). Occasional visitors accounted for 28% of the species
but less than 1.0% of the individuals. Estuarine transients
(juveniles and adults) comprised 12% of the species and 8%
of the abundance.
The distribution by residency varied among the
locations (Figure 2-14). In terms of abundance, permanent
residents dominated at all six general locations. The trend
was most pronounced upstream. Transient individuals were
somewhat more prevalent at mid- and downstream locations.
Although less than 2.0% of the abundance, occasional
visitors comprised 15 to 21% of the number of species
overall.
Discussion
Methods
Enclosure net methodology. Using the enclosure net
method, direct sampling of fish occupying submerged red
mangrove shorelines was possible. The fine mesh net
captured the small forage fish that numerically dominated
the habitats. Inevitable escapes, especially by large
roving fish, probably occurred, however, as persons
deploying the net waded up to the sites. The recovery
efficiencies of fishes encircled in the nets were comparable

Figure 2-14. Comparison of residency among the general
locations based on percent of total abundance. Data used in
calculations were taken with enclosure nets (for benthic
forage fish and water column forage fish) and by visual
census for large roving fish.
a. Abundance of fish in each category.
b. Number of species in each category.

Percent Percent
86
Residency
Occasional Transient Resident
Number of Species
100 F i= =i
Up Mid Down Up Mid Down
West East

87
to those found by other investigators who targeted small
fish (less than about 7.0 cm in total length) in vegetated,
shallow areas using (Weinstein & Davis 1980). Although
rotenone reguired cautious handling, fish immediately began
re-occupying sites after nets were removed. Rapid rotenone
degradation in the relatively high temperatures which
prevailed in these waters (Nielson and Johnson 1983), and
dilution due to the turnover of water at the relatively open
sites prevented cumulative adverse rotenone effects.
Other factors could have affected the variable
efficiency of recovery from individual nets. Turbidity from
disturbed sediments reduced the ability of the collectors to
capture fish with dip nets. In addition, variation in winds
onto the sites may have caused rotenone effectiveness to
vary.
Visual census methodology. The visual census
methodology had several advantages for sampling fish in
mangrove habitats. The speed and flexibility of the method
permitted sampling in a broad range of habitats that could
not be sampled with the net, including mangrove locations
with greater depths and a wider fringe. The visual method
was non-destructive, thus allowing evaluation of the
persistence of use by repeated observations of the same
stations and fishes. The problem of net avoidance was
eliminated with this method. In addition, large fish (e.g.
tarpon) that would normally escape nets and trawls could be
surveyed. As in studies of streams and coral reefs in which

88
the visual method has increasingly been used, it has
possible further value for use in studies on behavior,
species interactions, and microhabitat use in mangrove
habitats.
Direct observational sampling also has disadvantages.
Somewhat surprisingly, many fish were attracted to the
snorkelers (Dibble 1991). Fishes such as snook, tarpon,
gray snapper, cichlids, bluegill, and killifish, frequently
came within a few inches of the snorkeler's clipboard,
presumably out of curiosity. As verified in the snorkeling
efficiency tests, this attraction led to double-counting of
individuals on the multiple swims.
A major disadvantage of the method was its sensitivity
to reduced visibility conditions. Although the fishes
approached more closely in low visibility situations (1.0 to
2.0 m), the uncertainty level in identification of species
was often increased. Overall, those species that tended to
remain at the bottom or far back in the fringe were surveyed
less accurately in deeper and low visibility sites. This
problem was a particular disadvantage for surveying the
benthic forage fish at deeper stations. Even in depths of
only 1.0 m, the use of SCUBA equipment might be an advantage
if visibility range is less than water depth since it would
permit better sampling of the bottom. Some small, rare
fishes and shy cryptic forms were less accurately sampled
with the visual method. Finally, the occurrence of large

89
dangerous predators may preclude visual sampling in certain
areas. Sites in deep creeks were abandoned based on sitings
of large sharks and alligators.
Combined methods Using two methods to complement each
other in the same habitats and regions had great advantages.
The two methods increased the range of mangrove shorelines
that could be surveyed. However, when using two methods,
interpretation problems can arise when neither is 100%
efficient. If the results do not agree, it is difficult to
determine if the discrepancies are due to differences in the
fish sampled or due to differences in the efficiency of the
methods. In this study, the two methods targeted different
size groups of fishes. Overall, benefits of using the two
methods clearly out-weighed the disadvantages.
Fish and Salinity
Salinity. As confirmed in this study, salinity
conditions in the area vary from year-to-year (Ginsburg
1956). The east and west systems differed greatly during
the drought conditions that prevailed during the main study
year. They were very similar, however, during the pilot
study year when rainfall was locally more plentiful. Under
low rainfall periods, the C-lll Canal may effectively block
most freshwater from flowing into the western system by
routing it towards the east. During high flow periods,
however, when more freshwater is available for distribution,
the east and west systems appear to have similar salinity
patterns.

90
Historically, such annual differences due to local
rainfall variations may have been more moderate. Before
freshwater wetlands were extensively drained, freshwater
probably gradually seeped into the study area from the
greater Taylor Slough drainage basin, resulting in more
dispersed distribution patterns and prolonged periods of
lower salinity levels.
Temporal Patterns. The first hypothesis proved to be
incorrect: none of the temporal patterns in density of any
fish group or species examined was attributable to changes
in salinity. In the study area, therefore, the fish do not
seem to react to salinity changes by short-term movement in
and out of the general locations (regions approximately 12
km in area). However, patterns for all species collected
were not individually analyzed. Thus, some short-term
relationships may be identifiable on further examination of
the data.
Temporal patterns were related to temperature, however.
For benthic forage and large roving fish, as temperature
increased, density decreased. Similarly, Thayer et al.
(1987) and Tabb et al. (1962) also found greater densities
in western and central Florida Bay in the late fall and
winter when temperatures were cooler. Temporal patterns for
the current study are not typical for estuarine fish
populations; usually peaks occur when freshwater inflow is
greatest (Gunter 1967, Weinstein 1979, Yanez-Arancibia et
al. 1980, Rogers et al. 1984, Stoner 1986, Flores-Verdugo

91
1990). Judging from the salinity data, this would have been
late summer for northeastern Florida Bay in the study year.
One possible explanation for this unusual condition may
be that in the summer, high temperatures combined with low
circulation to create a stressful environment for fish in
the Bay (Moyle & Cech 1988). This is supported by reports
by fishermen of fish kills in Florida Bay during those hot
summer months (M. Robblee, Everglades National Park,
personal communication).
Additionally, the density of the major component of the
large roving fish group, Lutjanus griseus, probably
accounted for much of the temperature related trend in large
roving fish densities overall. The larger individuals of
this species, migrate offshore in the summer, when spawning
occurs, and return in the winter (Starck & Schroeder 1971,
Rutherford et al. 1989). This migration may account for the
reduced densities of large roving fish in the summer.
Spatial patterns. The density of fish decreases from
west to east in Florida Bay (Sogard et al. 1987, 1989a).
This trend appears to continue into the northeastern Bay
(this study, Funicelli et al. 1986). Using an almost
identical method of sampling (enclosure nets), the mean fish
density found in western and central Florida Bay mangroves
by Thayer et al. (1987) was 8.0 fish m-2, compared to 3.3
fish m 2 found with nets for the northeastern Florida Bay
area.

92
As indicated by comparisons from up- to downstream,
salinity regime did not affect the overall density of
benthic and water column forage fish. Furthermore, in the
current study, no differences were found between the eastern
and western systems. Similarly, Thayer et al. (1987) found
no effect due to gradient for mangrove fishes collected with
enclosure nets. The sites sampled by Thayer et al. (1987),
were located in central and western Florida Bay from
downstream near the Keys, to upstream in Whitewater Bay and
Coot Bay and their collections were dominated by small
forage fish. Thus, salinity regime may have little
influence on densities of forage fish species throughout
Florida Bay.
The individual species of benthic forage fish were
likewise distributed widely and not systematically along the
gradient. Although small benthic fishes have certain other
life history characteristics that may explain the widespread
distributions observed (Sogard et al. 1987) most species
are notably euryhaline (Robins et al. 1986, Nordlie & Walsh
1989). Thus, they are good colonizers of all types of
habitats found in Florida Bay.
In general, high variance was evident in the density
estimates for water column forage fish. This was due, in
large part, to their schooling behavior. With the nets, a
school was either collected (and thus the abundances were
great) or not collected (and thus the numbers were zero).
Individual species of water column forage fish, however, do

93
appear to be systematically distributed along the salinity
gradient. Although only one rough silverside (Membras
martinica), was collected in the current study, they were
abundant at central and western mangrove sites (Thayer et
al. 1987). Abundant Atherinomorus stipes were collected by
Thayer et al. (1987) at his most downstream locations in
central Florida Bay (Crane Key and Captains Key). This
corresponded to the very abundant collection of this species
at the downstream-most locations in northeastern Florida Bay
(Blackwater Sound and Buttonwood Sound). Since salinity
regime correlated with the distribution of Atherinomorus
stipes and Menidia spp., relative densities of the species
in the family Atherinidae may be indicative of salinity
conditions.
The density of large roving fish was dramatically lower
upstream than mid- or downstream. Among the large roving
fish species, however, the influence of salinity regime on
spatial distributions was mixed. Blue-striped grunts and
gray snapper were less abundant upstream, but redfin
needlefish were more abundant at the upstream/east location
than elsewhere. Thus, some species may be limited by the
conditions that occur upstream, while others tend to thrive
there.
Community patterns. Although no systematic pattern of
distribution occurred along the gradient for benthic and
water column forage fish, for the large roving fish, greater
numbers of species occurred downstream. In many other

94
estuarine studies greater numbers of species have been found
downstream (Weinstein 1979, Yanez-Arancibia 1980, Rogers et
al. 1984, Sogard et al. 1987 & 1989b, Thayer & Chester 1989,
Lonaragen et al. 1990). The species usually responsible for
the greater downstream richness are adult members of the
marine-visitor group of fishes.
Residency. Overall, the dominance over the entire
study area by permanent residents (91% of abundance) was
unusual even for tropical estuarine systems (Yanez-Arancibia
et al. 1980, Davis 1988, Morton 1990). Only three species
(13 individuals, all adults) were members of the reef
community (Acanthurus chirurgus, Aluterus scriptus,
Sparisoma radians) (Jaap 1984). The islands of the Florida
Keys may inhibit connection of the fish community in Florida
Bay with that of the extensive reef tract adjacent to
Florida Bay (Sogard et al. 1987). The mudbanks in the
central and western Bay may further inhibit travel into
northeastern Florida Bay from the Gulf of Mexico. Perhaps
more significantly, the moderating influence of the
thermally stable Gulfstream water masses do not enter the
Bay. When temperatures reach extreme low (or high) levels,
those reef species that have entered Florida Bay may be
forced to migrate or be killed. For example, a doctorfish
(Acanthurus chirurgus) that was observed every month at a
downstream/west station from May through December,
disappeared once temperatures began to drop.

95
Sciaenids (drums) and bothids (flounder), major nursery
species in other estuaries in the southeastern United
States, Gulf of Mexico and the Caribbean (e.g. Roessler
1970, Lindall et al. 1973, Weinstein 1979, Yanez-Arancibia
1980, Stoner 1988, Sheridan 1991), were not collected at all
in the present study. Since spawning takes place in the
Gulf of Mexico, young juveniles may not survive the journey
from distant passes into northeastern Florida Bay due to
lack of tidal exchange and little circulation in the central
Bay (Sogard et al. 1987).
Mangroves in northeastern Florida Bay clearly are
nursery grounds, however, for several species of estuarine
transients, all of which are popular sportfish: gray
snapper, schoolmaster, blue-striped grunt, sheepshead and
great barracuda. They also support adult snook and tarpon,
species that are known to use similar mangrove habitats as
nursery areas elsewhere (Gilmore et al. 1983, Seaman &
Collins 1983).
A common life history pattern occurs among these
fishes: recruitment from offshore as post-larvae, settlement
and growth in inshore areas, and movement back offshore or
to deeper water as they attain larger size classes (DeSylva
1963, Starck & Schroeder 1971, Jennings 1985, Robins et al.
1986). Since these species comprise a major portion of the
large roving fish group, aspects of this basic life history
pattern may indicate why they were less abundant as a group
in the upstream locations.

96
Submerged aquatic vegetation (SAV), such as algae and
seagrasses, may provide cover for newly settled forms of
estuarine transient juveniles at a scale compatible with
their size. As the fishes grow, however, the SAV becomes
less likely to provide adequate cover for later stage
juveniles, and they may seek larger forms of structure for
shelter. A habitat expansion of this type was identified
for juvenile gray snapper by Starck & Schroeder (1971).
While smaller snappers dwell in seagrass beds, larger
juvenile snappers congregate near mangroves and other brush
during the day and return to feed in seagrass beds at night.
The smallest gray snapper individuals found in mangroves in
the current study were 7.5 cm, the same size indicated by
Starck & Schroeder (1971) at which snappers expand their
habitat use.
Seagrass beds are generally poorly developed in
northeastern Florida Bay (Zieman et al. 1989). Abundance of
SAV is temporally variable and SAV is often absent
altogether upstream (Montague et al. 1988). Without SAV,
young fishes may not have an adequate intermediate habitat
between the planktonic and mangrove stages in which to
settle. In addition, if larger juveniles (over 7.5 cm) do
wander upstream, they may find inadequate food resources;
many of the benthic invertebrates that they consume live
epiphytically on SAV and may not occur in adequate abundance
levels without SAV (Montague et al. 1989). Thus, lack of
SAV may result in both reduced fish recruitment and growth

97
rates. This scenario may serve as an alternative or
complementary hypothesis to salinity intolerance, and lack
of access from distant passes, in explaining the lower
overall abundances of large roving fish observed in mangrove
habitats upstream.

CHAPTER 3
FISH COMMUNITIES IN
FLORIDA BAY MANGROVE SHORELINE HABITATS:
RELATIONS WITH PHYSICAL PARAMETERS AND COVER
Specific features of mangrove habitats may contribute
to the structure of fish communities in predictable
patterns. In the only study to specifically address this
question to date, limited support for this concept was
found: certain species seemed to prefer mangrove habitats to
more open sites (Sheridan 1991). Other investigators have
identified patterns between fish community structure and
development of vegetative structure in seagrass beds (e.g.
Sogard et al. 1987, Thayer & Chester 1989), kelp beds
(Ebeling & Laur 1985), and littoral zone plants in ponds
(Werner et al. 1983). In addition, the "rugosity" and
vertical structure of coral reefs have also been identified
as factors in increased densities of some species of fish
(Luckhurst & Luckhurst 1978).
One role of structure in aquatic habitats is to protect
vulnerable prey fishes from predators (Werner et al. 1983,
Ebeling & Laur 1985). However, since food resources can
become exhausted or be of lower quality in vegetated
habitats, the safest refuge is not always the location that
98

99
vulnerable fishes choose (Werner et al. 1983, Schmitt &
Holbrook 1985). In addition, since recruitment in aquatic
systems is largely based on widely-dispersed larvae, the
occupation of particular sites may be based on chance
vacation of living space by a previous occupant and the
largely unpredictable occurrence of available recruits from
the plankton (Sale 1980, Sutherland 1980, Sale & Douglas
1984). Thus, the prediction of habitat use is a complex
problem involving recruitment, species interactions,
resource availability, and random influences.
If fish are not randomly distributed within the
mangrove shoreline habitats of northeastern Florida Bay, it
may be possible to identify features correlated with fish
densities. Density can be used as a quantitative
approximation of habitat quality (Sogard & Able 1991).
Thus, the objective of this portion of the study was to
analyze density data and habitat information in northeastern
Florida Bay to determine any habitat preferences among the
fish found in the mangrove shoreline.
Materials and Methods
For both the visual census and enclosure net methods,
fishes and salinity were monitored at each station
repeatedly over the period of May 1989, through May 1990
as described in Chapter 2. These stations were located
across a gradient from upstream near sources of freshwater
inflow to downstream (Chapter 2). A total of 328,960 fish

100
were censused or collected in enclosure nets and visual
samples. Due to the nature of the sampling methods, the
values for habitat structural variables were determined
using slightly different procedures for the two fish data
sets. Each of the 17 visual census sites analyzed consisted
of 8 substations. At each substation, a transect
perpendicular to the shoreline was designated. Since
corresponding fish censuses were conducted at each visual
census substation, all 136 substations were used in the
analysis. In each of the 18 enclosure net stations, 3
transects were designated within the area repeatedly
enclosed by the net. Since corresponding fish data were
available for each net site as a whole, mean values for the
3 transects within each of the 18 stations were used in the
analysis.
For all 190 transects, data were collected using a 1.0
m frame. Starting at the shoreline edge, the frame was
placed every three meters outward along the transect to 6 m
off the mangrove fringe. The following data were collected
using this overall method:
Water depth: Within each frame, four measurements were
taken, one in each quadrant using a meter stick.
Fringe width: Along each transect the distance from
the shoreline to the outermost mangrove fringe was
measured.
Tree height: Within each frame, four measurements were
taken, one in each quadrant using marked
poles or by visual estimation for trees taller
than 2 m.

101
Tree cover. The percent of canopy coverage over the
frame as observed from the surface of the water
was estimated.
Prop root size and density. Calipers were used to
measure the diameters of red mangrove prop roots.
Observations were divided into two groups: those
less than and those greater than 2 cm in diameter.
Numbers in each category were counted within each
frame. Few prop roots less than 2 cm occurred
downstream. To avoid biasing the analysis, only
the data for prop roots greater than 2 cm in
diameter were used.
Submerged aquatic vegetation. Within each frame, the
total volume of seagrass, algae and detrital
material was determined separately by placing a
meter stick in each quadrant and noting the height
of each component above the substrate. The
percent areal coverage of each component was also
estimated for each quadrant. These values were
summed for the analysis.
Salinity mean: The average value for all months at
each station was determined from measurements
taken with a refractometer.
Salinity variation: The standard deviation for each
station was calculated from the concurrent
measurements.
Analysis
Datasets containing 136 visual census substations and
the 18 enclosure net stations were analyzed separately. For
comparison, means and standard deviations were determined
for each variable. To select the most appropriate method of
analysis, correlations among the variables were also
calculated.
Multi-colinearity among several of the variables was
revealed in the correlation analysis. Thus, principal
components analysis, a multivariate technique designed to
discern factors that have generated interdependence within a
data set was used (Afifi & Clark 1984, Robblee 1987,

102
Grossman et al. 1991). Principal components of the habitat
variables were calculated for both the visual census and the
enclosure net data sets using the SAS correlation matrix and
varimax rotation methods (Afifi & Clark 1984, Smith & Duke
1987). Each component was interpreted by examining
correlations between the original variables and the derived
components.
SAS FACTOR was then used to calculate individual
factors from the principal components for each observation
in the original data set. These uncorrelated factors were
used as the independent variables in a multiple linear
regression with log-transformed fish densities as the
dependant variables. Fish density data were divided into 3
groups of species for the analysis based on their size,
mobility and position in the mangrove habitat. These
categories were benthic forage fish, water column forage
fish, and large roving fish as defined in Chapter 2. All
species were also analyzed separately. Only principal
components with eigenvalues greater than the proportion of
the variance in the data that could be explained by an
individual variable (i.e. those > 1.0) were used in the
regressions (Grossman et al. 1991).
The slope of the regression line describes the nature
of the mangrove habitat and fish relationship. The variance
in fish density explained by each component (R2) indicates
the strength of this relationship and also provides a basis

103
for comparisons of mangrove habitat/fish relationships
between fish groups and species (Robblee 1987).
Results
Relationships Among the Habitat Variables
Correlations among habitat variables. For visual
censuses and enclosure net data, mean values and
correlations among the habitat variables differed (Tables 3-
1 and 3-2). The visual census sites had greater water
depths and tree heights but less cover than the enclosure
net sites. This difference was probably due to the broader
range of mangrove habitats covered in the visual census.
The greatest correlation between habitat variables was found
for submerged aquatic vegetation and salinity, with greater
volumes occurring where salinity means were higher (Table 3-
2). Salinity means were inversely correlated with salinity
variation, indicating that where salinity mean was higher,
salinity variation was less.
Underlying factors among habitat variables. Results of
the principal components analysis and correlations of those
components with each habitat variable differed for the
enclosure nets (Table 3-3) and for visual the censuses
(Table 3-4). The first 3 principal components explained
86.2% of the variance in the original habitat data
associated with enclosure nets, and 79.7% of the variance in
the enclosure net data. For both data sets, the first
principal component explained 52% of the variation.

Table 3-1. Means, standard deviations, and correlations among the physical/
environmental variables associated with the 18 enclosure net stations.
Water
depth
cm
Fringe
width
m
Tree
height
cm
Prop
roots
n
Submerged
Vegetation
cm3
Salinity
mean
ppt
Salinity
variation
ppt
Tree cover
%
Mean
48.53
8.88
112.49
10.82
22.30
34.10
8.87
67.02
Standard deviation
10.33
3.43
55.38
5.51
6.58
7.79
3.26
14.33
Correlations (r)
between
variables
Water depth
1.000
Fringe width
0.167
1.000
Tree height
0.635
0.416
1.000
Prop roots
0.285
0.633
0.606
1.000
Submerged vegetation
0.319
0.163
0.047
-0.142
1.000
Salinity mean
0.492
0.689
0.412
0.518
0.568
1.000
Salinity variation
-0.417
-0.425
-0.414
-0.365
-0.599
-0.829
1.000
Tree cover
0.732
0.393
0.735
0.572
-0.012
0.550
-0.441
1.000

Table 3-2. Means, standard deviations and correlations among the physical/
environmental variables associated with the 136 visual census stations.
Water
depth
cm
Fringe
width
m
Tree
height
cm
Prop
roots
n
Submerged
vegetation
cm3
Salinity
mean
ppt
Salinity
variation
ppt
Tree
cover
%
Mean
71.27
8.26
252.08
13.04
25.77
33.99
9.97
49.59
Standard deviation
20.69
4.82
102.61
11.82
10.40
8.17
5.01
12.00
Correlations (r) between variables
Water depth
1.000
Fringe width
0.537
1.000
Tree height
0.391
0.435
1.000
Prop roots
-0.187
-0.223
-0.146
1.000
Submerged vegetation
0.632
0.599
0.472
-0.146
1.000
Salinity mean
0.633
0.575
0.449
-0.103
0.937
1.000
Salinity variation
-0.598
-0.530
-0.438
0.139
-0.862
-0.914
1.000
Tree cover
0.172
0.317
0.364
-0.339
0.137
0.069
-0.032
1.000

Table 3-3. Results of the principal components analysis of the physical
and environmental variables associated with the 18 enclosure net stations
Correlations greater than 0.6 are underlined.
Principal
component
1
2
3
4
5
6
7
8
Eigenvalue
4.197
1.586
1.112
0.369
0.336
0.232
0.113
0.055
% Variance explained
52.460
19.830
13.900
4.610
4.200
2.900
1.400
0.100
Cumulative %
52.460
72.290
86.190
90.810
95.010
97.900
99.310
100.000
variance explained
Correlations (r) of
the original variables
with the
PC's
Water depth
0.931
-0.151
0.029
0.192
0.080
0.248
0.066
0.021
Fringe width
0.039
-0.181
0.930
0.074
0.261
0.153
0.063
0.001
Tree height
0.400
-0.150
0.183
-0.008
0.263
0.837
0.120
0.005
Prop roots
0.145
-0.176
0.364
-0.139
0.847
0.262
0.101
0.021
Submerged vegetation
0.116
-0.308
0.076
0.932
-0.122
-0.019
-0.043
0.014
Salinity mean
0.291
-0.591
0.497
0.373
0.247
0.039
0.163
0.302
Salinity variation
-0.169
0.876
-0.180
-0.343
-0.139
-0.167
-0.076
0.029
Tree cover
0.613
-0.239
0.198
-0.121
0.241
0.338
0.583
0.029
106

Table 3-4. Results of the principal components analysis of the physical
and environmental variables associated with the 136 visual census stations
Correlations greater than 0.6 are underlined.
Principal
component
1
2
3
4
5
6
7
8
Eigenvalue
4.181
1.393
0.799
0.552
0.463
0.431
0.131
0.500
% Variance explained
52.300
17.400
10.000
6.900
5.800
5.400
1.600
0.600
Cumulative %
52.300
69.700
79.700
86.600
92.500
97.700
99.400
100.000
variance explained
Correlations (r) of the original variables with the
PC's
Water depth
0.408
0.068
0.141
-0.082
0.201
0.873
-0.010
0.002
Fringe width
0.354
0.161
0.167
-0.097
0.879
0.198
-0.008
0.002
Tree height
0.263
0.187
0.926
-0.047
0.146
0.123
-0.007
0.002
Prop roots
-0.054
-0.160
-0.040
0.980
-0.072
-0.061
0.003
-0.001
Submerged vegetation
0.896
0.064
0.176
-0.051
0.221
0.218
0.167
0.186
Salinity mean
0.926
0.006
0.160
-0.018
0.198
0.220
0.077
-0.148
Salinity variation
-0.900
0.034
-0.168
0.065
-0.161
-0.194
0.306
0.008
Tree cover
-0.002
0.962
0.165
-0.170
0.123
0.053
0.004
0.002
107

108
Using the correlation between the original habitat
variables and each component, each component was interpreted
in ecological terms. Correlations greater than 0.60 were
considered in the interpretation (Afifi & Clark 1984) For
the enclosure net data, the first component described
gradients of water depth and percent canopy cover. The
second component associated with the net data described the
salinity regime, in particular, salinity variation. As
salinity variation increased, all other habitat variables
decreased, especially submerged aquatic vegetation and tree
cover. The third component was associated with fringe
width. In combination, these components, describe a
gradient of mangrove habitat development, together with
salinity regime.
For the visual census data, the first component was
most strongly correlated with salinity regime, water depth
and submerged aquatic vegetation. Tree cover was the only
strongly correlated variable associated with the second
component, and tree height, with the third. Together with
salinity regime, these components combined describe a
gradient of total habitat development including both
mangroves and submerged aquatic vegetation.
Relationships Between Habitat Factors and Fish Densities
The first 6 factors were included as independent
variables in the multiple regression analysis. Densities of
fish in the three fish groups were dependent variables
(Table 3-5). No significant relationships were found for

Figure 3-5. Results of principal components regression. B, least squares
regression coefficient; p, significance level; R2, percent variance explained
by the entire model; df, degrees of freedom; SS, sums of squares;
P(R2), percent variance explained by an individual principal component.
BENTHIC FORAGE FISH
(Net Data)
Source
df
SS p
R2
Source
B
P
P(R2)
Preference (p<-001)
Regression
6
2.647 0.1031
0.049
PCI
-0.057
0.0992
1.25
Residual
209
51.601
PC2
-0.027
0.4338
0.28
PC3
-0.054
0.1218
1.10
PC4
0.065
0.0643
1.57
PC5
0.027
0.4382
0.27
PC 6
-0.027
0.4409
0.27
WATER COLUMN
FORAGE
FISH (Net Data)
Source
df
SS p
R2
Source
B
P
P(R2)
Preference (p<.001)
Regression
6
6.826 0.0142
0.073
PCI
0.152
0.0009
5.08
Deep water/dense canopy
Residual
209
87.084
PC2
-0.071
0.1083
1.15
PC3
0.040
0.3788
0.35
PC4
-0.017
0.7085
0.06
PC5
0.052
0.2525
0.58
PC6
0.030
0.5135
0.19
LARGE ROVING
FISH (Visual Data)
Source
df
SS p
R2
Source
B
P
P(R2)
Preference (p<.001)
Regression
6
44.100 0.0001
0.188
PCI
0.097
0.0001
6.48
High salinity/dense SAV
Residual
1617
235.130
PC2
0.032
0.0001
0.73
Dense canopy
PC3
0.059
0.0001
2.37
Tall trees
PC4
-0.0170
0.0410
0.21
PC5
0.061
0.0001
2.62
Wide fringe
PC6
0.095
0.0001
6.31
Deep water
109

110
benthic forage fish, however (Table 3-5). As a group, these
small fishes (less than 15 cm) did not appear to select
habitats based on the parameters measured in this study.
A significant regression was derived for density of
water column forage fish, but only the first principal
component was a significant source of variation (Table 3-5).
Thus, the greater mangrove canopy coverage and water depth,
the greater the abundance of water column forage fish.
In contrast to the other two fish groups, for the
density of large roving fish, all factors were significant
sources of variation except prop root density (Table 3-5).
Thus, sites with greater development of mangroves and
submerged aquatic vegetation (SAV), and with higher, less
variable salinity had greater densities of large roving
fish.
Of the 77 total species collected, 14 had significant
(p<.0001) regressions on the 6 factors (Table 3-6). Nine of
these species were benthic forage fish and 3 were water
column forage fish. The greatest amount of variation
explained was 38.8%. This value was derived for densities
of Opsanus beta, the Gulf toadfish, which was most abundant
at high salinity sites with greater development of both
mangrove and SAV. Poecilia latipinna (sailfin molly) was
found most abundantly where mangrove prop roots were more
dense and the width of the fringe was greater. One of the
most abundant fishes, Floridichthys carpi (gold-spotted
killifish), was more prevalent in shallow sites with more

Ill
Table 3-6. Summary of multiple regression results with 6
factors derived from the principal components analysis.
Only species for which p-values were < 0.0001 are presented
in the table. (Abbreviations as per Table 3-5)
Family/
Abundance
Adjusted Preference
P(R2)
species
R2 (pc.0001)
Engraulidae (anchovies)
Anchoa mitcheM (nets)
18,605
0.174 sparse SAV
13.7
Batrachoididae (toadfish)
Opsanus beta (nets)
529
0.388 deep water/
9.0
dense canopy
high salinity
14.9
abundant SAV
8.6
tall mangroves
4.5
Cyprinodontidae (killifish)
Floridichthys carpi (nets)
13,018
0.278 shallow water/
15.6
open canopy
short mangroves
7.8
Lucania parva (nets)
10,237
0.350 deep water/
10.6
dense canopy
high salinity
4.9
narrow fringe
8.3
abundant SAV
5.5
tall mangroves
5.2
Poecilidae (livebearers)
Poecilia latipinna (nets)
11,000
0.156 wide fringe
4.9
dense prop roots
5.3
Gambusia sp. (nets)
1,907
0.171 high salinity
9.9
Atherinidae (silversides)
Atherinomorus stipes (nets)
11,042
0.211 high salinity
7.7
abundant SAV
9.2
Menidia sp.(nets)
4,348
0.159 lowsalinity
12.9
Lutjanidae (snappers)
Lutjanus griseus (nets)
18,461
0.161 high salinity/
5.3
abundant SAV
tall mangroves
1.0
wide fringe
4.6
deep water
4.4
Gerreidae (mojarra)
Eugerres plumieri (nets)
704
0.171 lowsalinity
14.8
Haemulidae (grunts)
Haemuion sciurus (nets)
2,768
0.103 high salinity/
1.6
abundant SAV
tall mangroves
3.7
deep water
4.8
Gobiidae (gobies)
Gobiosoma robu stum (nets)
534
0.137 lowsalinity
10.6
Lophogobius cyprinoides (nets)
200
0.321 low salinity
25.7
Microgobius gulosus (nets)
2,997
0.335 low salinity
19.9
narrow fringe
11.3

112
open mangrove tree canopy cover. Anchoa mitchelli was also
more abundant in open locations that were low in volume of
SAV.
For several species, salinity appeared to over-ride all
other variables in importance. Salinity accounted for over
10% of the variation in Gambusia sp. densities; this species
appeared to prefer higher salinity sites in the study area.
Salinity also explained from 10 to 25% of the variation in
densities of Menidia spp., Eugerres plumieri, Gobiosoma
robustum, Lophogobius cyprinoides, and Microgobius gulosus;
these species were more abundant where salinities were
lower. Of all the variables, salinity was a significant
source of variation for 11 of the 14 species (Table 3-7).
None of the other variables came close to this level of
apparent importance.
Seven benthic forage fish species that were abundantly
collected were not significantly correlated with the
factors. Among these were the 2 Fundulus species and
Cyprinodon variegatus as well as the 3 most abundant
mojarras: Eucinostomus gula, Eucinostomus harengulus, and
Gerres cinereus. These species are probably very flexible
in habitat selection.
Among the species of large roving fish in the study,
only the densities of gray snappers and blue-striped grunts
were significantly correlated with the variables measured.

Table 3-7. Summary of the qualitative influence of the original variables
on density of fish by group and species. If the value of the given variable
is increased, density will either increase (+), decrease (-), or there
will be no effect (0).
Category/
Species
Tree
height
SAV
Water
depth
Prop
roots
Tree
cover
Salinity
mean
Salinity
variation
Fringe
width
Percent
correlates
Benthic forage
0
0
0
0
0
0
0
0
0
Water column forage
0
0
+
0
+
0
0
0
25
Large roving
+
+
0
0
+
+

+
75
Floridichthys carpi
0
0
_
0
.
0
0
0
25
Lucania parva
+
+
+
0
+
+
-
-
88
Poecilia latipinna
0
0
0
+
0
0
0
+
25
Gambusia sp.
+
0
0
0
0
+
-
0
38
Eugerres plumieri
0
0
0
0
0
_
+
0
25
Gobiosoma robustum
0
0
0
0
0
-
+
0
25
Lophogobius cyprinoides
0
0
0
0
0
_
+
0
25
Microgobius guio sus
0
0
0
0
0
-
+
-
38
Opsanus beta
+
+
+
0
+
+
_
0
75
Anchoa mitchelli
0
-
0
0
0
0
0
0
13
Atherinomorus stipes
0
+
0
0
0
+
_
0
38
Menidia sp.
0
0
0
0
0
-
+
0
25
Lutjanus griseus
+
+
+
0
0
+

+
75
Haemulon sciurus
+
0
+
0
0
+
-
0
50
Trends among species:
% plus
35
29
29
6
24
41
29
18
% minus
0
6
6
0
6
29
41
12
% zeros
65
65
65
94
71
29
29
71

114
Densities of other abundant species such as Arius felis,
Mugil cephalus, Sphyraena barracuda and Strongylura notata
were not correlated with any of the measured variables.
For the regressions with individual groups of fish and
species, the range of percent of variance explained by the
combination of variables was low overall (10.3 to 38.8%).
However, the regressions were significant for 18% of all the
species, and approximately one-half of the species that were
collected in great abundances (i.e. over 100 individuals).
Discussion
Sites with a combination of lower mean salinity and
high salinity variation had lower levels of all the other
habitat development variables indicating reduced habitat
development at such locations. This finding is consistent
with previous results for benthic community development in
northeastern Florida Bay (Montague et al. 1989).
An example of a well-developed mangrove shoreline is
illustrated in Figure 3-1. In variable salinity conditions,
such mangrove habitats are less likely to occur. Among the
species preferring sites with greater mangrove habitat
development are the snappers, grunts, toadfish, rainwater
killifish, and sailfin mollies. Other species may utilize
mangrove habitats on a less discriminating basis and tend to
occupy all mangrove habitats. Such species include snook,
barracuda and sheepshead. This assemblage of fishes is
enhanced when mangrove shorelines occur and especially where

Figure
Florida
-1.
Bay.
Illustration of a well-developed mangrove habitat in northeastern

116

117
such habitats are well-developed. Activities which directly
destroy mangrove shorelines or degrade the quality of such
habitats have negative consequences for these important
species of sport fish.
Although no significant relationship was found among
the habitat variables and benthic forage fish as a whole,
certain species did appear to discriminate on the basis of
the measured variables. Lucania parva, Opsanus beta, and
Poecilia latipinna, three abundant species collected,
appeared to select sites with greater mangrove development.
In contrast, Floridichthys carpi chose more open sites.
This killifish seemed to prefer shallower locations, a trend
also noted for individuals living in seagrass habitats
(Sogard et al. 1987). In addition, Opsanus beta and Lucania
parva, chose sites with more abundant SAV. Similarly,
Sogard et al. (1987) collected greater abundances of these
species in seagrass bed sites with greater vegetation
densities.
In the current study, densities of water column forage
fish as a group were greatest where water depths were
greater and mangrove canopy more completely blocked the
daylight from reaching the submerged habitat. These species
use schooling as a possible defense mechanism against
predators. Thus, they may use greater volumes of water to
increase school size and prefer sites shaded by mangrove
tree canopy for additional cover.

118
While these factors were important for the water column
forage fish group as a whole, other features of the habitat
seemed to segregate the individual species. The most
abundant species in this group, Anchoa mitchelli, was less
abundant at sites with greater volumes of SAV. This species
was extremely dense at only one location in the study area
(Little Blackwater Sound). Although not a variable
included in this analysis, turbidity may have been the more
attractive habitat feature for the bay anchovy at this site.
Of all the significant regressions for the abundant
species, the range of percentage variation in fish density
explained by habitat variables in this study (10.3-38.8%)
was only slightly lower than that found by Sogard et al.
(1987) in seagrass beds (24.8-42.7%). Thus, fish may select
particular habitats based on salinity and physical features
in Florida Bay but other factors (e.g. foraging
requirements, species interactions) are also undoubtedly
important.
The present findings do not differ greatly from long
term observations of fishes on small patches of coral, in
which habitat attributes other than overall size were of
little value in predicting the structure of fish assemblages
(Sale & Douglas 1984). Many species of reef fish may
therefore select habitats based on overall parameters (i.e.
large coral reefs vs. very small patches) and not detailed
features. The low magnitude of variance explained by the
habitat variables measured in seagrass beds and mangroves

119
may be due to the selection of habitats based on general
features (i.e. mangrove over seagrass habitats).
Although density of large roving fishes as a group were
correlated with the physical features, some species that
were categorized as "roving in the current study may not
actually roam among the mangrove habitats. Individuals of
these species appear to persist at particular locations for
long periods of time (based on limited observations of
tagged fish). Thus, while some large fishes (such as
mullet, catfish, barracuda and needlefish) may truly be
wanderers and display no discrimination among mangrove
shoreline habitats, other species (gray snapper and blue-
striped grunts) may maintain more permanent residency at
certain locations and display definite habitat preferences.

CHAPTER 4
FOOD HABITS OF MANGROVE FISHES:
A COMPARISON ACROSS SALINITY GRADIENTS
Diets have been studied for fishes associated with
whole estuaries (Darnell 1961), mangroves (Odum 1971, Beumer
1978, Salini et al. 1990), salt marshes (Harrington &
Harrington 1961, Rozas & LaSalle 1990), seagrass beds
(Livingston 1982), lakes (Werner & Hall 1983), and streams
(McNeely 1987). One successful dietary strategy identified
in extremely variable habitats is omnivory, or the
consumption of a wide variety of prey organisms including
both plant and animal material (Darnell 1961, Harrington &
Harrington 1961). Opportunism, or an ability to exploit
alternative foods depending on availability, is also a
valuable dietary strategy for survival in variable
environments (Odum & Heald 1972, Livingston 1982, Salini et
al. 1990). Trophic systems in aquatic habitats are often
characterized by shared common resources among the various
species of fishes (Harrington & Harrington 1961, Livingston
1982) .
In estuaries, seasonality of freshwater inflow
increases habitat variability upstream relative to more
120

121
stable downstream locations (Rogers et al. 1984, Moyle &
Cech 1988). In northeastern Florida Bay, biomass of
submerged aquatic vegetation is much lower and highly
variable upstream than down (Montague et al. 1989) .
Abundances of polychaetes, crustaceans and other benthic
invertebrates are highly correlated with plant biomass in
this area: 80% of the epifauna live among the blades of
seagrass and algae. For estuarine fishes, these animals are
among the more heavily exploited food items (Darnell 1961,
Livingston 1982).
In light of these conditions, fish diets are likely to
display patterns along the gradient of salinity variation in
the northeastern Florida Bay study area, due to fish
foraging habits and variation in prey base. Seasonal
variations in diets may also be expected. Toward the
overall goal of identifying the influence of variation in
freshwater inflow on habitat use, the objective of this
chapter is to identify dietary components and make
comparisons among the more ubiquitous and abundant species.
Breadth and variability of diets should reflect
environmental conditions in the more variable versus stable
habitats, and over the seasons.
Materials and Methods
Fish were collected in mangrove shorelines located up-,
mid- and downstream in two systems (east and west) in
northeastern Florida Bay (Chapter 2). Species were selected
for the analysis of food habits because they ranged across

122
the entire area and occurred consistently over the study
period. For Lutjanus griseus, Sphyraena barracuda and
Fundulus grandis, all samples collected were analyzed. Due
to great abundances, for Floridichthys carpi, Strongylura
notata and Eucinostomus harengulus, smallest individuals (<
3 cm) were eliminated and subsamples were selected from the
remainder. Fish were not divided into size or age classes
for this analysis, however.
Gut analysis of the 6 selected species was contracted
out to Mote Marine Laboratory. For the laboratory analysis,
36 taxonomic levels (e.g. family) were selected for
consistency with other estuarine studies of fish food habits
(Brook 1977, Beumer 1978, Livingston 1982). The analysis
chosen follows the "points method" of Hynes (1950). For
each fish (n= 1,222), total length was recorded and the
stomach was extracted. In order to identify variability
among individuals, no stomachs were pooled. The material
found in each stomach was distributed to a standardized
level within a gridded petri dish (Hellawell & Abel 1971).
Each stomach containing food was considered to be uniformly
full (Starck & Schroeder 1971). Using a dissecting
microscope, percent composition of each food category for
each specimen was calculated by estimating the area covered
by the material on the grid (Neilson & Johnson 1983) .
Analysis
To determine frequency of occurrence, the number of
fish in which each food item occurred was listed as a

123
percent of the total number of fish examined (Hynes 1950).
In addition, the mean percent composition for each item was
calculated by obtaining an average value for all specimens
of a species. The results of both analyses are presented to
give a complete picture of the relative dietary importance
of the items consumed (Hyslop 1980).
Two sets of multivariate analyses of variance (MANOVA)
were performed for each species with the major items found
in the gut as dependent variables. All data were
transformed using an arcsine square-root function prior to
these calculations (Kleinbaum & Kupper 1978). The first set
of MANOVAs addressed spatial variation. Gradient (upstream,
midstream, downstream), system (east, west) and interaction
of gradient and system, were used as independent variables.
The second MANOVA looked at temporal variation. Season was
the independent variable. Only major food items, defined as
those which occurred in at least 20% of the specimens or
exceeded an average of 4% in composition, were included in
the MANOVAs.
Results
Shared Resources
Of the 24 items (counting fish as one and excluding
unrecognizable), 8 occurred in all 6 species (Table 4-1).
Of these, amphipods were the most ubiquitously consumed,
present in at least 5% of the specimens of all species.

Table 4-1. Food items found in the stomachs of fish. Data include: % Comp, the
mean % composition; % Freq, the % of specimens in which the given items occurred.
Species
Food Item 8
Eucinostomus
harengulus
% Comp. % Freq
Floridichthys
carpi
% Comp. % Freq
Fundulus
grandis
% Comp. % Freq
Lutjanus
griseus
% Comp. % Freq
Sphyraena
barracuda
% Comp. % Freq
Strongyiura
notata
% Comp. % Freq
Average
% Comp. % Freq
Crustaceans
Amphipods
25.2
76.2
9.8
47.6
8.7
27.0
4.0
19.5
2.0
7.8
0.3
6.9
8.3
33.2
Cladocerans
0.0
0.0
0.0
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
Copepods
0.9
14.4
1.2
29.9
0.0
0.8
0.0
0.0
0.1
0.0
0.0
0.9
0.4
10.7
Crabs
0.3
1.1
0.0
0.0
2.8
6.5
12.8
23.4
1.4
3.1
0.2
3.1
2.9
3.9
Isopods
2.1
12.2
1.7
7.5
8.6
21.0
7.3
27.3
5.4
1.6
0.2
10.7
4.2
12.7
Mysids
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.9
0.0
0.2
Ostracods
0.2
13.8
1.2
30.8
0.0
1.2
0.0
0.0
0.0
1.6
0.0
0.3
0.2
10.9
Shrimp
0.2
2.8
0.3
2.1
2.6
6.5
4.4
19.5
4.5
3.1
2.3
7.9
2.4
5.7
Non-chitinous invertebrates
Holothuroideans
0.2
0.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
Nematodes
0.2
12.2
0.3
19.2
0.7
8.9
0.1
7.8
0.0
4.7
0.0
1.9
0.2
10.1
Nudibranches
0.0
0.6
0.0
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.2
Polychaetes
1.9
18.8
0.5
6.3
0.3
2.4
1.3
7.8
0.0
0.0
0.0
0.6
0.7
5.6
Sipunculids
0.0
0.0
0.0
0.0
0.0
0.0
0.9
1.3
0.0
0.0
0.0
0.0
0.2
0.1
Eggs
0.2
3.3
2.6
13.5
1.3
13.3
0.0
6.5
0.3
7.8
0.1
1.3
0.7
8.0
Fish
Clupeidae
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.6
0.0
0.0
0.0
0.1
Engraulidae
0.0
0.0
0.0
0.0
0.1
0.4
0.0
0.0
1.0
6.3
2.8
1.6
0.7
0.8
Belonidae
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.4
3.1
2.4
0.6
0.5
0.3
Cyprinodontidae
0.0
0.0
0.0
0.0
1.1
1.2
1.0
1.3
2.4
26.6
17.0
3.8
3.6
2.7
Atherinidae
0.0
0.6
0.0
0.0
0.0
0.0
4.5
6.5
2.0
4.7
3.3
2.8
1.6
1.5
Syngnathidae
0.0
0.0
0.0
0.0
0.0
0.0
0.9
2.6
0.0
0.0
0.0
0.0
0.2
0.2
Gerreidae
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
1.6
0.8
0.3
0.1
0.2
Blenniidae
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.3
0.0
0.4
0.3
0.1
0.1
Gobiidae
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.6
1.6
0.4
0.6
0.2
0.2
unidentified fish
0.1
2.8
0.0
0.9
5.8
20.2
11.2
44.2
8.9
48.4
20.4
31.8
7.7
18.3
Insects
Ants
0.0
0.0
0.0
0.0
1.4
2.4
0.0
0.0
4.4
0.0
0.0
8.5
1.0
2.7
Insect larvae
0.9
6.6
0.2
0.3
1.9
0.8
0.0
0.0
2.7
0.0
0.0
0.3
1.0
1.3
Terrestrial adult insects
0.8
4.4
0.0
3.6
4.8
19.8
0.0
1.3
2.9
3.1
1.2
9.1
1.6
8.3
Mollusks
Bilvalves
0.1
2.2
0.6
9.3
0.7
3.2
0.1
2.6
0.0
0.0
0.0
0.0
0.2
3.7
Gastropods
0.8
2.2
0.5
9.6
5.0
3.2
0.1
0.0
0.0
0.0
0.0
0.6
1.1
3.8
Veliger larvae
0.0
0.0
0.0
0.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.2
Plants
Algae
6.3
55.8
7.1
67.4
2.5
20.6
2.7
26.0
0.1
7.8
0.6
1.9
3.2
33.4
Seeds
0.0
0.0
0.0
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
Protozoans
Foramifera
0.2
15.5
0.9
45.8
0.0
2.4
0.0
1.3
0.0
0.0
0.0
0.9
0.2
15.6
Unrecognizable
50.4
84.0
68.4
92.8
37.2
76.6
22.7
64.9
14.5
48.4
13.0
46.2
34.4
72.0
Empty
9.4
-
5.7
-
12.5
-
14.3
-
28.1
-
35.8
-
17.2
Total fish analyzed
181
334
248
77
64
318
1222
Total # of items (fish=1, unrec-O)
18
20
17
13
10
17
24
Sizes of fish analyzed (min-max.cm) 4.1 to 13.2
3.0 to 7.7
4.2 to 13.8
12.5 to 37.7 8.0 to 50.5
9.5 to 42.0
124

125
Other items exploited to varying degrees by all 6 species
included isopods, shrimp, nematodes, eggs, fish parts, adult
terrestrial insects and algae.
Breadth of Diets
Excluding unidentifiable material and counting all prey
fish as a single food item, of the 6 species, Floridichthys
carpi and Eucinostomus harengulus foraged on the widest
variety of items (20 and 18 out of 24 total, respectively)
(Table 4-1). Only 2 items, however, were consumed in mean
guantities exceeding 4% of total diet: amphipods and algae.
Thus, although a wide variety of items were utilized by
these species, a degree of specialization was apparent.
In contrast, Fundulus granis, Lutjanus griseus, and
Strongylura notata, not only consumed a wide range of items
overall, but several items (5,6 & 4 respectively) were
consumed in mean quantities exceeding 4%. Isopods and fish
were major items commonly found in all 3 of these omnivorous
species.
Sphyraena barracuda was the most specialized of the
species analyzed, consuming mostly fish. For this species,
the benthic forage fishes in the family Cyprinodontidae were
consumed with particular frequency, even though the most
abundant fishes in the study area were Engraulids and
Atherinids (Chapter 2).
Thus, in the study area, among the 6 species, 3 feeding
strategies were evident. In one strategy, used by
Floridichthys carpi and Eucinostomus harengulus, many items

126
were consumed, but only 2 in great abundance. In a second
strategy, used by Fundulus granis, Lutjanus griseus,
Strongylura notata, many items were also consumed, with
several (4 to 6) items ingested in substantial abundances.
In the third strategy, used by Sphyraena barracuda, fewer
items were consumed overall, with only one in significant
quantity.
Diet Variability
The MANOVA results for all species, as presented in
Table 4-2, indicate the degree to which location (i.e.
gradient, system) was a significant source of variation in
fish diets. For Floridichthys carpi and Eucinostomus
harengulus, gradient position was a significant source of
variation. In both species, more animal food (i.e.
copepods, ostracods, nematodes) was obtained downstream than
up- and midstream, while more algae was consumed upstream.
The only other influence of gradient occurred for Lutjanus
griseus, which consumed significantly more crabs upstream
than mid- or downstream. For Strongylura notata, Fundulus
grandis and Sphyraena barracuda, diets differed among the
locations, but not systematically along the salinity
variation gradient.
Seasonal influences on the fish diets are indicated by
the second MANOVA results presented in Table 4-3. For
Floridichthys carpi and Eucinostomus harengulus, algae was
more important in the spring. For Lutjanus griseus, crabs,
an important dietary item, were found abundantly in diets

Table 4-2. Comparison of food habits among gradient positions and systems.
Data used are percent composition of stomach contents transformed using an arcsin
square-root function. MANOVA was used to test the hypothesis that gradient and system
were significant sources of diet variation. Specific F-tests to contrast means for
each pair of gradient positions were used to make multiple comparisions.
Species
Independ.
variable
Manova results
(Wilks lambda)
Dependent
variables
Univariate results
Contrast
summary"
F-value
p-value
F-value
p-value
df'
Floridichthys
gradient
7.72
0.0001
amphipods
5.59
0.0001
5/317
carpi
system
2 52
0.0251
copepods
9.06
0.0001
5/317
down > up & mid
gradient* system
6.73
0.0001
ostracods
10.01
0.0001
5/317
down > up & mid
nematodes
4.37
0.0007
5/317
down > up & mid
forams
2.43
0.0352
5/317
algae
14.41
0.0001
5/317
up > mid & down
Eucinostomus
gradient
6.59
0.0001
amphipods
5.88
0.0001
5/169
down > up & mid
harengulus
system
6.00
0.0030
algae
4.61
0.0006
5/169
up > mid & down
gradient'system
1.81
0.1263
Fundulus
gradient
1.12
0.3473
amphipods
1.84
0.1054
5/243
grandis
system
0.54
0.7452
isopods
1.71
0.1328
5/243
gradient'system
1.52
0.1292
fish (unidentified)
1.06
0.3822
5/243
insects (terrestrial adults)
0.64
0.6716
5/243
algae
0.72
0.6122
5/243
Lutjanus
gradient
4.93
0.0001
amphipods
3.09
0.0139
5/76
griseus
system
0.70
0.6473
crabs
15.29
0.0001
5/76
up > mid & down
gradientsystem
1.65
0.0859
isopods
1.34
0.2568
5/76
shrimp
0.91
0.4806
5/76
fishs (unidentified)
2.52
0.0376
5/76
algae
2.39
0.0465
5/76
Strongylura
gradient
0.58
0.8567
amphipods
1.51
0.1868
5/310
notata
system
1.68
0.1241
isopods
1.95
0.0859
5/310
gradient'system
1.76
0.0510
shrimp
2.75
0.0189
5/310
ants
0.27
0.9300
5/310
insects (terrestrial adults)
1.62
0.1548
5/310
fish (unidentified)
0.71
0.6153
5/310
Sphyraena
gradient
1.06
0.3950
amphipods
1.23
0.3062
5/58
barracuda
system
1.35
0.2578
eggs
1.24
0.3011
5/58
gradient'system
2.07
0.0325
Cyprinodontids
3.97
0.0036
5/58
fish (unidentified)
0.55
0.7373
5/58
algae
1.17
0.3367
5/58
Model degrees of freedom / error degrees of freedom
** If no contrast is indicated, either an interaction or no significant differences occurred.
127

Table 4-3. Comparison of food habits among the seasons.
Data used are percent composition of stomach contents transformed using an arcsin
sguare-root function. MANOVA was used to test the hypothesis that season
was a significant source of diet variation. Specific F-tests to contrast means for
each pair of seasons were used to make multiple comparisions.
Species
Independ.
variable
Manova results
(Wllks lambda)
Dependent
variables
Univariate results
Contrast
summary**
F-value
p-value
F-value
p-value
df*
Floridichthys
season
1.95
0.0101
amphipods
1.32
0.2672
3/319
carpi
copepods
2.52
0.0581
3/319
O8tracods
3.28
0.0213
3/319
all>8um
nematodes
1.91
0.3122
3/319
forams
0.85
0.4672
3/319
algae
4.20
0.0062
3/319
wln&spr>sum&fall
Eucinostomus
season
3.34
0.0032
amphipods
0.61
0.6070
3/171
harengulus
algae
6.09
0.0006
3/171
spr>all
Fundulus
season
1.33
0.1745
amphipods
1.60
0.1896
3/245
grandis
isopods
0.17
0.9163
3/245
fish (unidentified)
0.05
0.9862
3/245
insects (terrestrial adults)
3.11
0.0272
3/245
spr>all
algae
1.80
0.1482
3/245
Lutjanus
season
1.25
0.2232
amphipods
0.82
0.4858
3/73
griseus
crabs
3.56
0.0183
3/73
all>win
isopods
1.23
0.3056
3/73
shrimp
0.73
0.5395
3/73
fishs (unidentified)
0.13
0.9391
3/73
algae
0.84
0.4757
3/73
Strongylura
season
3.03
0.0001
amphipods
1.59
0.1907
3/312
notata
isopods
4.40
0.0048
3/312
win>all
shrimp
2.17
0.0914
3/312
ants
1.74
0.1589
3/312
insects (terrestrial adults)
6.16
0.0004
3/312
spr>all
fish (unidentified)
0.93
0.4287
3/312
Sphyraena
season
1.21
0.2691
amphipods
0.55
0.6491
3/60
barracuda
eggs
1 65
0.1864
3/60
Cyprinodontids
1.94
0.1333
3/60
fish (unidentified)
0.97
0.4110
3/60
algae
0.76
0.5215
3/60
Model degrees of freedom / error degrees of freedom
** If no contrast is indicated, no significant differences occurred.
128

129
all year except in winter, when they were seldom consumed.
For both Fundulus granis and Strongylura notata, adult
insects, a prevalent item, were greater in diets in the
spring.
Discussion
Shared Resources
As commonly occurs in aguatic habitats, diets
overlapped among the 6 species as resources were often
shared (Harrington & Harrington 1961, Livingston 1982, Odum
1983). This overlap occurred in all areas and throughout
the study period.
Diet overlap can become particularly evident when one
resource attains a periodic peak of abundance. In other
aguatic habitats, for example, populations of penaeid shrimp
(Salini et al. 1990) or larval insects (Harrington &
Harrington 1961) increase under certain conditions, and
opportunistic fish take advantage of the resource abundance.
A similar seasonal increase in exploitation of a particular
food resource was found in the current study for adult
insects exploited by Fundulus granis and Strongylura
notata.
In western Florida Bay, gray snapper and spotted sea
trout (Cynoscion nebulosus) diets overlapped when peak
abundances of pink shrimp (Penaeus duorarum) occurred in
November (Hettler 1989). Although a similar pattern might
have been expected in the current study, none became
evident. Migrating juvenile pink shrimp, while a major

130
portion of the epibenthic fauna in western Florida Bay, have
lower densities in the interior and eastern Bay (Holmquist
et al. 1989a).
Diet Breadth and Variability
Three types of feeding strategies were identified among
the 6 species based on diet breadth and degree of
opportunism. Firstly, Floridichthys carpi and Eucinostomus
harengulus appeared to rely strongly on the plasticity of
their diets, feeding on one basic resource consistently, but
also consuming smaller quantities of many other resources.
Although it was not separated from other materials in the
unrecognizable category, these species probably consumed
detritus. A product of breakdown of dead plants, this
material contains a "coating" of bacteria and fungi of
nutritional value (Heald et al. 1974). Primary consumers of
detritus include amphipods, shrimp, crabs, and certain
fishes. Among the 6 species analyzed in this study, a major
portion of the diet of Floridichthys carpi was probably
composed of detritus; in the North River, the diet of this
killifish was 21% detrital material (Odum 1971). In that
study, Eucinostomus harengulus was considered a secondary
consumer of this material, with only 6% of the gut contents
directly composed of detritus (Odum 1971).
The second major strategy was employed by Fundulus
granis, Lutjanus griseus and Strongylura notata. Rather
than switching among a very wide variety of items, they
consumed about 5 items consistently and abundantly. Thus,

131
these fishes tended to rely more strongly on omnivory, or
constant foraging for these several items, rather than
opportunism. The diets of Fundulus granis probably also
included detritus (Rozas & LaSalle 1990). However, the 3
other species analyzed in this study were apparently not
direct consumers of detritus (deSylva 1963, Odum 1971,
Starck & Schroeder 1971, Brook 1977, Thayer et al. 1987a).
The third major strategy was exemplified by the
piscivore, Sphyraena barracuda. One resource, fish, was
consistently and effectively targeted by the barracuda.
These modes of feeding are consistent with other
estuarine investigations. Two extreme modes of feeding were
identified in Lake Ponchartrain, for example, with
detritivore/omnivores, such as mullet, on one end of the
spectrum, and piscivore/specialists, such as gar and jacks,
on the other end (Darnell 1961). Similar extremes were
observed in a red mangrove/saltmarsh habitat in east
Florida, with killifish and snook at opposite poles
(Harrington & Harrington 1961). In these examples,
intermediate strategies incorporate the consumption of small
benthic invertebrates in diets with greater and lesser
portions of detritus and fish. In the current study, the
detritivore was probably most strongly represented by
Floridichthys carpi, with the barracuda at the opposite
extreme. Intermediate species include the mojarra, Fundulus
granis, Lutjanus griseus and Strongylura notata, in order
of increasing piscivory.

132
Influence of the Salinity Gradient
Diets of three species were significantly influenced by
location along the salinity gradient. Eucinostomus
harengulus and Floridichthys carpi diets included more
algae upstream and more copepods, ostracods, and nematodes
downstream. This finding tends to be consistent with the
reduced levels of benthic invertebrate populations found
upstream in a previous investigation in northeastern Florida
Bay (Montague et al. 1989). Plant materials, other than
seeds, are generally less concentrated energy sources than
animal sources (Odum 1983) .
Lutjanus griseus consumed more crabs up- than
downstream. Although all the parts were not identified to
species in the current study, most were probably from mud
crabs of the species Rhithropanopeus harrisi. This brackish
water crab was abundantly collected in minnow traps placed
in mangrove shorelines upstream, but was never collected
downstream (unpublished data). It occurred abundantly in
the gray snapper stomachs in the North River (Odum 1971).
Rhithropanopeus harrisi is, thus, an important forage item
for snappers that appears to occur abundantly in brackish
water in the study area. As mean salinity rose from 16 to
over 30 ppt in northeastern Florida Bay, these crabs were
virtually eliminated from seagrass beds over a period of
four years (Holmguist et al. 1989a).
Crabs in general may not be readily assimilated as food
by fishes due to a higher percentage of exoskeleton

133
(Weisberg & Lotrich 1982). In an experiment, Fundulus
heteroclitus, fed a diet of only fiddler crabs (Uca pugnax)
lost weight (Weisberg & Lotrich 1982). In a salt marsh
where fiddler crabs comprised 17 times more food volume than
any other item for Fundulus granis, the killifish may have
compensated for the crabs' low caloric value by consuming
their prey in very large quantities (Rozas & LaSalle 1990).
Since two items consumed more abundantly upstream,
algae and crabs, were lower quality energy sources for fish
growth, these locations may be somewhat lower quality
habitats than mid- and downstream in terms of obtainable
foods for these species. The fishes that forage upstream
may compensate, however, by foraging on greater quantities
of the lower quality items that are available.

CHAPTER 5
PREDATION RATES ON SMALL BENTHIC FISH
ACROSS A SALINITY GRADIENT
The value of estuaries as nursery areas for fishes that
have wider distributions as adults has been confirmed in
many studies (e.g. Gunter 1938, Reid 1954, Carter et al.
1973, Blaber & Blaber 1980, Yanez-Arancibia et al. 1980,
Bell et al. 1984, Blaber et al. 1985, Blaber et al. 1989,
Pinto 1987, Robertson and Duke 1987). An influx of
juveniles to an estuary usually coincides with the season of
highest freshwater discharge, when salinity levels drop. As
juveniles develop in the estuary, they tend to migrate from
fresher upstream areas to more saline downstream habitats
(Weinstein 1979, Rogers et al. 1984). Based on observations
such as these, one of the paradigms of estuarine ecology has
developed: that estuarine salinity conditions contribute to
the survival of juvenile fish because stenohaline marine
predators are precluded from entering portions of the
estuary having lower, more variable conditions of salinity
(Gunter 1961, Austin 1971, Browder & Moore 1981, Odum et al.
1982). Despite its widespread acceptance, this hypothesis
has not been tested before now.
134

135
Predation is one of the most complex of the species
interactions studied by ecologists. In comparing levels of
predation intensity among habitats, not only are relative
abundances of predators and prey important factors, but the
potential rates of predation must also be estimated
(Kitching 1983). These rates depend on behavior of predator
and prey and characteristics of the habitat. Some
behavioral and microhabitat aspects have been modeled by
observing individual components of predator/prey
interactions in the laboratory (e.g. Holling 1966; Barshaw
and Able 1990a). Other habitat related factors, however,
must be measured in the field, and are thus, more difficult
to determine.
The tethering technique, a useful field method, has
increasingly been used by ecologists studying the effects of
different habitat features on predator/prey interactions.
Briefly, the tethering technique involves affixing a line to
a prey organism so that evidence of predation can be
determined from its condition after a period of time. These
investigations are usually accompanied by complementary
laboratory studies or censuses of predators and prey. A
summary of tethering studies is presented in Table 5-1.
In this study, the tethering technique has been used to
compare predator encounter rates across a gradient of
salinity conditions. The question of interest is, are these
rates lower in the more variable upstream locations relative
to mid- and downstream where conditions remain more saline?

Table 5-1. Summary of studies testing predation hypotheses in aquatic habitats
using tethering techniques.
Taxa Tethered
Hypothesis
Tested
Decapod
Crustaceans
Brittlestars
Fish
Mangrove Leaves
and Propagules
Comparison of predator
encounter rates among
macro-habitats
Heck & Wilson 1987
Wilson 1989
Aronson 1989
Shulman 1985
Mclvor & Odum 1988
THIS STUDY
Smith 1987
Prey vulnerability in
and out of vegetated
micro-habitats
Barshaw & Able 1990a
Barshaw & Able 1990b
Hay et al. 1989
Heck & Thoman 1981
Herrnkind & Butler 1986
Wilson 1989
Wilson et al. 1987
Wilson et al. 1990
Rozas & Odum 1988
Absolute rate of
predation
Robertson 1987
Smith 1987
Modification of prey
behavior by the
presence of predators
Power & Matthews 1983
Phillips & Swears 1979
Preference by predators
for a particular prey
Aronson 1988
Smith 1987

137
This experiment actually integrates several of the
steps involved in predation that were described by Holling
(1966). First, measurements indicate whether or not
predators occur in the locations under comparison.
Secondly, the tests indicate how well the predators can
perceive the prey, given the conditions at the site (e.g.
turbidity levels). Thirdly, the ability and propensity to
consume particular prey species relative to size and
palatability is indicated.
To add to the understanding of the influence of
freshwater inflow on fish assemblages, the objective of this
chapter was to compare predator encounter rates up- and
downstream during the rainy season. If the paradigm is
true, fewer tethered prey should be consumed upstream than
downstream, as the marine predators are excluded from these
locations due to the lower, more variable salinity
conditions.
Materials and Methods
Preliminary Tests
To evaluate the effectiveness of the tethering
technigue for specific prey fishes and conditions in the
study area, preliminary tests were conducted within an
enclosure formed from two 30 m seine nets that excluded all
potential predators. For these tests, fish were tethered by
sewing one end of a 1.0 m length of 8 lb test monofilament

138
fishing line through the lower jaw of a small fish (4-10
cm). The other end was looped over a 1.25 cm diameter
polyvinylchloride (PVC) pole that had been driven into the
substrate. Sixteen fish were tethered inside each enclosure
and checked at 3, 6 and 24 hour intervals. Although all
fish survived and remained securely tethered for three
hours, in these preliminary tests, one (Eucinostomus gula)
died after six hours and several other fishes died after 24
hours (Table 5-2). In addition, fish that died at the
bottom were quickly attacked by scavengers (e.g. crabs and
gastropods), possibly interfering with interpretation of
test results.
Since the study objective was to determine rates of
prey encounters with predators, not scavengers, these
results prompted further investigation into alternative
tether and stake designs. After several prototypes, an L-
shaped stake was made by joining two 1.0 m long PVC pipes
with an elbow (Figure 5-1). A hole was drilled at the free
end for attaching the tethered fish (as above) and the other
end was driven into the sediment. When deployed, the top
bar of the L-shaped stake was above the surface of the
water. To ensure that predators and not scavengers were
responsible for removing the prey, the tethered fish were
forced to remain above the substrate by adjusting the depth
to which the pole was driven into the substrate.

Table 5-2. Results of tests using tethered fish within an enclosure
formed by two block nets at Buttonwood Sound. Species used were:
Fundulus granis, Eucinostomus gula, and Floridichthys carpi.
Test
number
After 3
hours
After 6
hours
Live
After 24
Dead
hours
Missing
Live
Dead
Missing
Live
Dead
Missing
1
16
0
0
15
1
0
8
0
8*
2
16
0
0
16
0
0
14
0
2
3
16
0
0
16
0
0
3
12**
1
* Net partially down overnight; needlefish found inside net
** Fish that had died were being consumed by scavengers (e.g. anemones, gastropods,
and crabs)

140
Figure 5 1. Illustration and dimensions of the tethering
systems used in this study as deployed near mangrove edges.

141
Tethering Experiment
Small fish were tethered at far/up-, mid/up-, mid- and
downstream locations in two systems (west and east) in
northeastern Florida Bay (Figure 5-2). Tethered species
included those that are consumed by dominant members of the
predator guild in the study area, as indicated by the
results of the food habits portion of this overall study
(Chapter 4). Small fish make up over 25% of the diets of
Strongylura notata (redfin needlefish), Lutjanus griseus
(gray snapper) and Sphyraena barracuda (great barracuda).
Trials were conducted on two dates at each of the eight
locations during mid-summer 1990. For each trial, salinity,
horizontal secchi distance, and water depth were recorded.
Fish to be used in each trial were collected by setting out
several minnow traps near each site on the day before a
test. Small fish (4 to 10 cm total length) from five
species were used in the 16 trials: killifish
(Floridichthys carpi, Cyprinodon variegatus, Fundulus
grandis, and Fundulus confluentus) and crested gobies
(Lophogobius cyprinoides). In each trial, ten to sixteen
fish were tethered 10 m apart and about 2.0 m from the
mangrove edge (Figure 5-1) Sites with water depths of
about 50 cm were selected for each fish. The total time
each prey was tethered ranged from 3 to 3.5 hours. For
approximately 1.5 hours, while observers were within 10 to
100 m of the tethered stakes, they recorded when possible
the type, size, and number of predators that approached or

142
Far/Up
Far/Up
Mid/Up
Mid/Up
O Snorkel Sites
* Enclosure Net Sites
Tethering stations
Y Key
Ranger Station
*
Figure 5-2. Locations of study sites in northeastern
Florida Bay.

143
attacked tethered fish. Other potential predators in the
vicinity were also noted. At the end of the test period,
each stake was examined. Predation was assumed to have
occurred if a fish was missing, severely damaged or if a
predator was tethered.
Analysis
For each test, the percent of prey that were subjects
of predation was used as the dependent variable for
statistical analyses. Analysis of variance (SAS GLM) was
used to determine if predator encounter rates differed due
to gradient position (i.e. far/up-, mid/up-, mid- and
downstream) or system (east, west). To make multiple
comparisons, specific F-tests were used to contrast means
for each pair of gradient positions. SAS GLM with Student-
Newman-Keuls multiple comparison tests were also used to
determine if tethering different species created an
extraneous source of variation.
In addition, correlations were calculated between the
predator encounter rates and corresponding ranges of
visibility (horizontal secchi distance), salinity, and water
depth. ANOVA was also used to determine if differences in
salinity, visibility and water depth were due to gradient
position.
Results
Of 235 tethered fish used, a mean of 83.5% were
subjects of predation after the 3 to 3.5 hour period (Table
5-3). Predation rates averaged 90% in mid/up-, mid- and

Table 5-3. Results of predator encounter trials comparing rates in
four gradient positions from down to far upstream in two systems.
Fish were tethered for three hours adjacent to mangrove edges.
Trials were conducted from June 20, 1990 to August 3, 1990.
System*
Gradient**
Test
Number
Results
Total Percent
Fish Missing
Salinity
Secchi
distance
Mean Water
Depth
1
1
1
16
25
22.5
1.0
64
1
1
2
10
70
23.5
1.5
62
1
2
1
12
100
22.9
1.0
53
1
2
2
15
87
14.7
1.0
52
1
3
1
16
100
52.0
2.2
45
1
3
2
15
100
50.0
1.0
62
1
4
1
16
100
50.0
5.5
58
1
4
2
15
93
45.0
3.4
58
2
1
1
13
77
7.3
2.0
47
2
1
2
15
53
16.0
2.0
48
2
2
1
16
94
29.9
1.0
48
2
2
2
13
100
15.5
1.0
50
2
3
1
16
88
50.0
0.8
62
2
3
2
16
69
48.0
0.8
73
2
4
1
16
93
45.0
4.0
93
2
4
2
15
87
45.0
9.0
93
Total
235
Mean
83.5
33.4
2.3
60.8
Systems: 1 = West, 2 = East
Gradient: 1 = Far/upstream
2 = Mid/upstream
3 = Midstream
4 = Downstream

145
downstream locations contrasted to 55% for those
far/upstream locations (Figure 5-3). Salinity, range of
visibility, and water depth also varied across the gradient
(Figure 5-3).
The far/upstream predation rate was significantly lower
than the rates for the mid/up-, mid-, and downstream
locations, but there were no significant differences between
east and west systems (Table 5-4). Rates of predation were
not significantly different among the species (df=4, F=1.98,
p=0.1296).
Neither salinity, secchi distance nor water depth were
significantly correlated with observed predator encounter
rates (Table 5-5). These factors did vary significantly
among the locations, however.
Salinity means were not significantly different between
the far/up- and mid/upstream locations (p<0.3454, df=l), nor
between the mid- and downstream locations (p<0.3921, df =
1). However, the subgroup formed by the far/up- and
mid/upstream stations had a significantly lower mean
salinity than the subgroup formed by the mid- and downstream
locations (p<0.0001 for each pair of contrasts).
Range of visibility was significantly greater
downstream than at the other locations (p<0.005 for each
pair of contrasts). Water depth, however, was significantly
greater at the downstream/east location (p<0.05 for each
pair of contrasts).

146
Predator Encounter Rates
Mean and Standard Deviation
100
-t-1 80
C
CD
O 60
L_
0)
Q_ 40
20
40
Salinity
CL
Cl
to
cu
E
o
100
80
60
40
Missing Fish
L
Visibility
Water Depth
1
Far
Up
Mid
Up
Mid Down
Figure 5-3. Mean predation rates for small (<10 cm TL)
benthic forage fish tethered for 3 hours across a salinity
gradient in northeastern Florida Bay. Error bars indicate
the standard deviations of test rates resulting from 4 tests
in each of the 4 gradient positions.

Table 5-4. Analysis of sixteen predator encounter rate tests.
Data used are percentage of fish missing (arcsine-transformed)
after being tethered for three hours adjacent to mangrove edges.
ANOVA was used to test the hypothesis that gradient and system
were significant sources of variation. Specific F-tests to contrast
means for each pair of gradient positions were used to make multiple
comparisons.
Source of
Variation
F-value
p-value
df*
Contrast
Summary**
MODEL
4.06
0.0332
7/8
Gradient
6.94
0.0129
3/8
1 < 2,3 & 4
System
1.12
0.3206
1/8
No differences
Gradient
2.22
0.1634
3/8
No differences
X
System
Model degrees of freedom / Error degrees of freedom
* *
Gradient: 1 = Far/upstream; 2 = Mid/upstream; 3 = Midstream; 4 = Downstream

148
Table 5-5. Correlations of predator encounter rates
with environmental variables. Data are percentage of
fish missing (arcsine transformed) after being tethered
for three hours adjacent to mangrove edges in 16 trials.
Pearson correlation coefficients and probability
values are indicated.
Variable
Test
Results
Correlation
Value (R)
p-value
Mean Water
-0.1463
0.5885
Depth
Salinity
0.3688
0.1597
Secchi
0.1308
0.6293
Distance

149
Thus, although the predator encounter rate was greater
at the two far/upstream locations, salinity, visibility, and
water depth were not unusual at these locations relative to
the others. Salinity, the parameter of most interest, was
lower at both the far/up- and mid/up- locations, but the
rate of predation was lower far/upstream, and higher at the
mid/upstream location.
Potential predators that were observed approaching
tethered fish at all the locations included redfin
needlefish, gray snapper, and barracuda (Table 5-6).
Observed only in the far/upstream locations were bull shark
juveniles, gar, alligators, and turtles.
A total of 28 actual predation events were recorded in
which the attacking predator could be identified (Table 5-
6). Most of these events (22) were due to needlefish
(Strongylura notata), half of which became tethered
themselves. Once they had swallowed the prey, the
needlefish were unable to cut the fishing line with the
teeth on their elongated jaw, but they usually were
eventually able to work the prey and themselves free of the
tether (e.g. by leaping).
Discussion
Predator encounter rates averaged more than 50%, even
in the most upstream locations. Thus, piscivorous predators
are ubiquitous along the estuarine gradient and were
effective at consuming tethered prey throughout the system.

Table 5-6. Species of predators associated with encounter rate
experiments. Observers in the boat or water recorded the following
predators taking, or in the vicinity of, the tethered fish.
Species
Predation
Events
Observed
Estimated
Abundance
In Vicinity
of Tethers
Estimated
Size
cm.
Gradient*
Where
Observed
Negaprion brevirostris
lemon shark
3
6
70-125
2,3
Carcharihinus leucas
bull shark
1
75
1
Strongulura notata
redfin needlefish
22
1000+
15-32
1,2,3,4
Centropomus undecimalis
snook
10+
45-75
2,3,4
Epinephelus itajara
jewfish
1
75
4
Caranx hippos
crevalle jack
2
30+
45-50
1,4
Lutjanus griseus
gray snapper
100+
15-40
1,2,3,4
Haemulon sciurus
blue-striped grunt
100+
15-40
4
Lepisosteus platyrhincus
Florida gar
16
45-55
1
Sphyraena barracuda
great barracuda
1
30+
15-70
1,2,3,4
Trionyx ferox
snapping turtle
1
45
1
Alligator mississippiensis
alligator
2
180-200
1
Butoroides striatus
green heron
1
1
*
Gradient: 1= Far/Upstream, 2=Mid/Upstream, 3=Midstream, 4=Downstream

151
In all study locations, these predators included
euryhaline species (e.g. snook, needlefish, gray snapper,
crevalle jacks, juvenile barracuda). Upstream, they were
joined by freshwater species such as gar that can forage in
brackish conditions. Similarly, lemon sharks are primarily
marine predators that were also observed upstream.
A relatively straightforward hypothesis can be given to
explain the finding that a great number (90%) of the
tethered fish were subject to predation in the mid- and
downstream locations. Greater predator encounter rates are
associated with proximity to some type of structurally
massive habitat such as a rocky breakwater (Aronson 1989) or
coral reef (Shulman 1985) Similarly, findings in Chapter 3
indicate that where mangrove habitat is more developed (i.e.
greater tree height, fringe width, canopy cover) greater
abundances of large roving fish occur; this fish group
includes many of the predators occurring in the study area.
In addition, habitat development is significantly greater in
mid- and downstream locations. Thus, the greater encounter
rates found further downstream may be explained by the
attraction of large roving predators to the structure
associated with well-developed mangrove habitats.
While attraction to highly developed mangrove habitats
may explain why predation rates were so high at mid- and
downstream locations, it does not account for the high rate
of predation that was also found at the mid/upstream
locations. In previous chapters, findings indicate that in

152
upstream locations overall, large roving fish densities are
lower (Chapter 2). Conditions upstream overall may be
inhospitable for large roving fish, preventing them from
permanently residing in these locations (Chapters 3 and 4).
In addition, salinity regime in both the far/up- and
mid/upstream locations, was low and variable in the current
study. Based on these findings, one would have expected the
experiments to indicate lower predator encounter rates at
both the far/up- and mid/upstream locations.
However, in actuality, tethered fish suffered lower
rates of predation far/upstream than mid/upstream. Thus,
the abundance of predators was probably equivalent at
mid/up-, mid-, and downstream locations, but less abundant a
far/upstream locations. The predator encounter rates,
therefore, do not appear to simply be functions of salinity
regime or mangrove habitat development alone.
One hypothesis that could explain the greater rates
mid/upstream in comparison to far/upstream, could be that a
significant number of the stenohaline predators primarily
residing in marine habitats also temporarily forage at the
edge of their primary range (e.g. Weinstein 1979). The
mid/upstream locations would be at the edge of this range
for marine predators. An analogous situation may occur for
freshwater predators at the other side of the
marine/freshwater interface.
To explain the lack of predators far/upstream, perhaps
other characteristics of the these locations (besides

153
salinity regime and mangrove habitat development) make them
safer havens for small benthic fishes. In contrast to all
the other locations, to forage far/upstream, a predator
would have to negotiate a series of sinuous channels and
interspersed ponds. Shallow shoals occur at pond/creek
intersections that are sometimes only a few centimeters
deep. This complex system seems likely to prevent access by
casual foragers and predators above certain size limits.
Based on these results, the ecological paradigm may
thus be qualified. Large predators may be prevented from
permanently occupying upstream locations by low and variable
salinity conditions, and the complexity of sinuous channels
may prevent them from foraging in far/upstream locations. A
safe haven for small benthic fish thus occurs in complex
habitats at the marine/freshwater interface. These findings
tend to support the hypothesis suggested by Browder & Moore
(1981): ideal juvenile fish habitat may occur where the
variable salinity conditions overlap areas of such habitat
complexity. They tend to refute, however, the hypothesis
that small fishes are protected from predation in estuaries
by lower salinities which tend to exclude stenohaline marine
predators.

CHAPTER 6
IMPLICATIONS AND CONCLUSIONS
Implications for Mangrove Fish Ecology
Fish and Mangrove Shorelines
Mangrove shorelines fulfill the food and cover habitat
requirements for many fishes. Among the functional roles of
inter-tidal mangroves as habitats for fish, the best
established is that of a temporary feeding location for a
wide range of species (Robertson & Duke 1990a, Blaber et al.
1985, Morton 1990). These species include small schooling
planktivores, benthic forage fish, and large piscivores.
They spend periods of low tide in deep open water or shallow
ponds and feed in the mangroves when the tide inundates the
forest (Davis 1988).
Since northeastern Florida Bay is non-tidal, the
mangrove habitats are unlike those in many other areas. In
Florida Bay permanently inundated mangrove habitats are
believed to provide young fish with refuge from larger
predators (Thayer et al. 1987a). This role was supported in
the current study, particularly for larger juveniles of the
154

155
estuarine transient species, Lutjanus griseus and Haemulon
sciurus. Of the habitats within Florida Bay, these species
prefer mangroves during the day (Thayer et al. 1987a &
1987b). An important aspect of habitat use by these species
is their migration away from the mangrove shorelines at
night, to feed in nearby seagrass beds (Starck & Schroeder
1971, Sogard et al. 1989c). Thus, in both tidal and non-
tidal habitats, linkages between mangrove shorelines and
other habitats may be critical for diel behavior patterns.
Besides snappers and grunts, other species that may
similarly rely on both mangroves and seagrass beds include
snook, sheepshead, barracuda and nurse sharks. A common
life history pattern occurs among these fishes: recruitment
from offshore as post-larvae, settlement and growth in
inshore habitats, and movement back offshore or to deeper
water as they attain larger size classes (Starck & Schroeder
1971, DeSylva 1963, Jennings 1985). These fishes tend to
use seagrass beds when they are smaller and move to
mangroves when they attain larger juvenile sizes.
Thus, shallow water habitats including mangroves and
seagrass beds, may be linked to one another through such
behavioral and life history patterns (Odum et al. 1982,
Parrish 1989). Mangroves provide cover and food resources
that are very different from adjacent habitats dominated by
submerged aquatic vegetation (SAV). However, both types of
habitats appear to be necessary to support certain fish
species. Prime locations for supporting these species may

156
occur where both mangrove development and SAV are greater.
Reduced SAV may thus account for the reduced abundance of
large roving fish upstream.
Selection Among Mangrove Habitats
Mangrove shorelines vary in functional value depending
on the degree of development and habitat needs of the
fishes. In freshwater streams, a gradient model has been
proposed (Schlosser 1987) that appears to apply equally well
in northeastern Florida Bay. This model identifies a
gradient from upstream areas (environmentally unstable,
shallower, lesser habitat development) to downstream areas
(stable, deeper, greater development). Greater habitat
development is linked with the occurrence of more species
and larger piscivorous individuals (Schlosser 1987) For
smaller fishes, upstream areas provide refugia from larger
piscivores that are more abundant downstream.
In northeastern Florida Bay, the mangrove habitats
ranged from less developed upstream (small trees, narrow
fringe, shallow water, high environmental variability), to
more developed downstream (tall trees, deeper water,
environmentally more stable). Among the large roving
species, greater abundances of gray snappers and grunts were
associated with more developed mangrove habitats in the
current study. In addition, numbers of species of large
roving fish were more abundant downstream than upstream.
Thus, in terms of large roving fish, the model applies well
to northeastern Florida Bay.

157
For smaller fishes, the implications of the model are
that upstream habitats are occupied by colonizing fishes.
Competition for food resources limits the abundances of
these fishes more than predation. Downstream, predation is
a more powerful force in the community structure. Thus,
upstream areas should be relatively advantageous for small
fishes in both freshwater streams and estuarine habitats.
Small fishes, however, were not more abundant upstream
in northeastern Florida Bay. Young-of-the-year estuarine
transient juveniles may represent the colonizers described
in the model. However, these fishes were absent in the
current study. Thus, there appears to be a missing
component of the fish community in the study area,
particularly in the upstream fish assemblages in the current
study.
Implications For Estuarine Fish Ecology:
The Nursery-ground Hypothesis
The absence of young-of-the-year juveniles of estuarine
transient fishes in the study area was also a significant
departure from the results one would have expected based on
widely accepted theories in estuarine ecology. In
northeastern Florida Bay, this condition was not unique to
the mangroves; both seagrass and mangrove habitats in
eastern and central Florida Bay also have low populations of
very young transients (Sogard et al. 1987, 1989as, Thayer et
al. 1987a).

158
The potential pool of post-larval estuarine transient
species from the Atlantic Ocean and Gulf of Mexico includes
sciaenids, lutjanids, haemulids, centropomids and elopids.
Presumably, currents carry potential recruits to the
northeastern Florida Bay area. Exchange with these sources
is limited in northeastern Florida Bay due to the presence
of the Keys and western mudbanks. Internal circulation in
Florida Bay is also weak due to the many islands and lack of
tides. If post-larval transients do enter the northeastern
Bay, they may meet with significant predation pressure due
not only to the occurrence of piscivores but also to a
reduction in the cover afforded by seagrass beds, which tend
to become less developed in the eastern Bay. These
conditions indicate that the chances of post-larval forms
reaching the upstream locations in the study area may be
small. Such conditions may not be unusual in estuaries,
however.
Habitat conditions documented in this study for the
upstream locations included almost no SAV and reduced
mangrove habitat development. However, shallow ponds and
sinuous creeks upstream may effectively reduce predator
encounter rates. Thus, a key factor in improving the use of
upstream habitats by estuarine transient juveniles appears
to be the presence of a persistent abundance of SAV.
Management Implications
Northeastern Florida Bay may have historically
supported more estuarine transient juveniles and greater

159
densities of fish than was observed in the current study.
Higher, abruptly changing salinity conditions may somehow
inhibit the development of lush communities of submerged
aguatic vegetation that provide cover for small fish and
benthic invertebrates (Montague et al. 1989). Sustained
lower salinity periods may promote growth of lush seagrass
(Ruppia martima) or algal (Chara, Batophora) communities
(Tabb et al. 1961, Montague et al. 1989). During more
saline periods, less dense growth of Halodule wrightii
communities may develop, if any vegetation grows at all. As
salinity changes with seasons, these communities may
alternate.
Although the regular study was conducted during a
period of very low rainfall and can only serve to provide
information on the ecosystem under low freshwater inflow
conditions, the pilot study took place at a time of higher
rainfall and high freshwater inflow conditions (summer
1988) In the pilot study, from October 1988 through March
1989, in upstream habitats, extremely dense Ruppia and algal
communities were observed. With experimental traps and gill
nets, great numbers of fishes were collected including gray
snappers, jacks, catfish and cichlids. A small mangrove
island and its surrounding waters in the Joe Bay study area
were heavily used by white pelicans (fish eating birds).
However, during the regular study (May 1989 May 1990),
very little Ruppia was observed. The pelican island was
evidently never used by birds and because the traps no

160
longer captured many fish, their use was abandoned. This
evidence indicates that the Ruppia community was probably
supporting a greater fish population in the area than
observed at any time during the course of the regular study.
Thus, northeastern Florida Bay, in the drought year
recorded in this study, was unusual in comparison to other
tropical, subtropical and warm temperate estuaries. Water
management efforts may be needed to restore sustained low
salinity periods, thereby inducing greater submerged aquatic
vegetation development and greater influx of estuarine
transient juveniles upstream into the more protected
habitats.

LITERATURE CITED
Afifi, A. A. and V. Clark 1984. Computer-aided multivariate
analysis. Wadsworth Inc., Belmont, California. 458 pages.
Aronson, R. B. 1988. Palatability of five Caribbean
ophiuroids. Bulletin of Marine Science 43(1): 93-97.
Aronson, R. B. 1989. Brittlestar beds: low-predation
anachronisms in the British Isles. Ecology 70(4): 856-865.
Austin, H. M. 1971. A survey of the ichthyofauna of the
mangroves of western Puerto Rico during December, 1967-
August, 1968. Caribbean Journal of Science 11(1-2): 27-39.
Baelde, P. 1990. Differences in the structures of fish
assemblages in Thalassia testudinum beds in Guadeloupe,
French West Indies, and their ecological significance.
Marine Biology 105: 163-173.
Barshaw, D. E. 1990. Tethering as a technique for assessing
predation rates in different habitats: an evaluation using
juvenile lobsters Homarus americanus. Fishery Bulletin 88:
415-417.
Barshaw, D. E. and K. W. Able 1990. Deep burial as a refuge
for lady crabs Ovalipes ocellatus: comparisons with blue
crabs Callinectes sapidus. Marine Ecology Progress Series
66: 75-79.
Bell, J. D., D. A. Pollard, J. J. Burchmore, B. C. Pease,
and M. J. Middleton 1984. Structure and function of a fish
community in a temperate tidal mangrove creek in Botany Bay,
New South Wales. Australian Journal of Marine and Freshwater
Resources 35:33-46.
Beumer, J. P. 1978. Feeding ecology of four fishes from a
mangrove creek in north Queensland, Australia. Journal of
Fish Biology 12: 475-490.
Blaber, S. J. M. and T. G. Blaber 1980. Factors affecting
the distribution of juvenile estuarine and inshore fish.
Journal of Fish Biology 17: 143-162
161

162
Blaber, S. J. M., D. T. Brewer, and J. P. Salini 1989.
Species composition and biomasses of fishes in different
habitats of a tropical northern Australian estuary: their
occurrence in the adjoining seas and estuarine dependence.
Estuarine, Coastal and Shelf Science 29:509-531.
Blaber, S. J. M., J. W. Young and M. C. Dunning 1985.
Community structure and zoogeographic affinities of the
coastal fishes of the Dampier Region of north-western
Australia. Australian Journal of Marine and Freshwater
Research 36:247-266.
Brook, I. M. 1977. Trophic relationships in a seagrass
community (Thalassia testudinum), in Card Sound, Florida.
Fish diets in relation to macrobenthic and cryptic faunal
abundance. Transactions of the American Fisheries Society
106(3): 219-229.
Browder, J. A. and D. Moore 1981. A new approach to
determining the quantitative relationships between fishery
production and the flow of freshwater to estuaries. In R.
Cross and D. Williams (Eds.) Proceedings of the National
Symposium on Freshwater inflow to estuaries. U. S. Fish and
Wildlife Service, Office of Biological Services, Volume I:
403-430.
Carter, L. A. Burns, T. R. Cavinder, K. R. Dugger, P. L.
Fore, D. B. Hicks, H. L. Revells, T. W. Schmidt 1973.
Ecosystems analysis of the Big Cypress Swamp and estuaries.
U. S. Environmental Protection Agency, Atlanta, Georgia.
Chester, A. J. and G. W. Thayer 1990. Distribution of
spotted seatrout (Cynoscion nebulosus) and gray snapper
(Lutjanus griseus) juveniles in seagrass habitats of western
Florida Bay. Bulletin of Marine Science 46(2): 345-357.
Croker, R. A. 1960. Growth and food of the gray snapper,
Lutjanus griseus, in Everglades National Park. Transactions
of the American Fisheries Society 91(4): 375-378.
Darnell, R. M. 1961. Trophic spectrum of an estuarine
community, base on studies of Lake Ponchartrain, Louisiana.
Ecology 42(3): 553-568.
Day, J. W. Jr., C. A. S. Hall, W. M. Kemp, A. Yanez-
Arancibia 1989. Estuarine Ecology, John Wiley & Sons, New
York, New York. 558 pages.
Davis, T. L. 0. 1988. Temporal changes in the fish fauna
entering a tidal swamp system in tropical Australia.
Environmental Biology of Fishes 21(3): 161-172.

163
Deaton, L. E. and M. J. Greenberg 1986. There is no
horohalinicum. Estuaries. 9(1): 20-30.
DeSylva, D. P. 1963. Systematics and life history of the
great barracuda. Studies in Tropical Oceanography,
Institute of Marine Science, University of Miami, Miami,
Florida. 179 pages.
Dibble, E. 1991. A comparison of diving and rotenone
methods for determining relative abundance of fish.
Transactions of the American Fisheries Society (In press).
Ebeling, A. W. and D. R. Laur 1985. The influence of plant
cover on surfperch abundance at an offshore temperate reef.
Environmental Biology of Fishes 12(3): 169-179.
Flores-Verdugo, F. J., J. W. Day, L. Mee and R. Briseno-
Duenas 1988. Phytoplankton production and seasonal biomass
variation of seagrass, Ruppia maritima, in a tropical
Mexican lagoon with an ephemeral inlet. Estuaries 11(1):
51-56.
Flores-Verdugo, F., F. Gonzalez-Farias, O. Ramirez-Flores
1990. Mangrove ecology, aguatic primary productivity, and
fish community dynamics in the Teacapan-Aqua Brava Lagoon
estuarine system (Mexican Pacific). Estuaries 13(2): 219-
230.
Funicelli, N. A., H. E. Bryant, M. R. Dewey, G. M. Ludwig,
D. A. Meineke, L. J. Mengel, and J. E. Skjeveland 1986.
Movements, relative importance, and standing stock of
important sport and commercial species in Everglades
National Park, Florida. U. S. Fish and Wildlife Service,
Gainesville, Florida. 186 p.
Gilmore, R. G., C. J. Donohoe, D. W. Cooke 1983.
Observations on the distribution and biology of east-central
Florida populations of the common snook, Centropomus
undecimalis. Florida Scientist Special Supplement
45(3/4):313-336.
Ginsburg, R. N. 1956. Environmental relationships of grain
size and constituent particles in some south Florida
carbonate sediments. Bulletin of the American Association
of Petroleum Geologists. 40(10): 2384-2427.
Grossman, G. D., D. M. Nickerson, and M. C. Freeman 1991.
Principal component analysis of assemblage structure data:
utility of tests based on eigenvalues. Ecology 72(1):
341-347.
Gunter, G. 1938. Seasonal variations in abundance of
certain estuarine and marine fishes in Louisiana, with
particular reference to life histories. Ecological

164
Monographs 8(3): 314-336.
Gunter, G. 1961. Some relations of estuarine organisms to
salinity. Limnology and Oceanography VI(2): 182-190.
Gunter, G. 1967. Some relationships of estuaries to the
fisheries of the Gulf of Mexico, in G. H. Lauff ed.
Estuaries, American Association for the Advancement of
Science, Washington, D. C. pages 621-638.
Harrington, R. W., and E. S. Harrington 1961. Food
selection among fishes invading a high subtropical salt
marsh: from onset of flooding through the progress of a
mosquito brood. Ecology 42(4): 646-666.
Hay, M. E., J. R. Pawlik, J. E. Duffy, and W. Fenical 1989.
Seaweed-herbivore-predator interactions: host-plant
specialization reduces predation on small herbivores.
Oecologia 81: 418-427.
Heald, R. J., W. E. Odum and D. C. Tabb 1974. Mangroves in
the estuarine food chain. in P. J. Gleason (Ed.)
Environments of south Florida: present and past. Memoir 2
Miami Geological Society, Miami, Florida 182-189.
Heck, Jr., K. L. and T. A. Thoman 1981. Experiments on
predator-prey interactions in vegetated aquatic habitats.
Journal of Experimental Marine Biology and Ecology 53: 125-
134.
Heck, Jr., K. L. and K. A. Wilson 1987. Predation rates on
decapod crustaceans in latitudinally separated seagrass
communities: a study of spatial and temporal variation using
tethering techniques. Journal of Experimental Marine
Biology and Ecology 107: 87-100.
Hellawell, J. M. and R. Abel 1971. A rapid volumetric
method for the analysis of the food of fishes. Journal of
Fish Biology 3:29-37
Herrnkind, W. F. and M. J. Butler IV 1986. Factors
regulating postlarval settlement and juvenile microhabitat
use by spiny lobsters Panulirus argus. Marine Ecology-
Progress Series 34: 23-20.
Hettler, Jr., W. F. 1989. Food habits of juveniles of
spotted seatrout and gray snapper in western Florida Bay.
Bulletin of Marine Science 44(1): 155-162.
Holling, C. S. 1966. The functional response of
invertebrate predator to prey density. Mem. Entorno. Soc.
Canada, No. 48, 86 p.
Holmquist, J. G., G. V. N. Powell and S. M. Sogard 1989a.

165
Decapod and stomatopod communities of seagrass-covered
mudbanks in Florida Bay: inter- and intra-bank heterogeneity
with special reference to isolated subenvironments.
Bulletin of Marine Science 44(1): 251-262.
Holmguist, J. G., G. V. N. Powell and S. M. Sogard 1989b.
Sediment, water level and water temperature characteristics
of Florida Bay's grass-covered mudbanks. Bulletin of Marine
Science 44(1): 348-364.
Hynes, H. B. N. 1950. The food of fresh-water sticklebacks
(Gasterosteus aculeatus and Pygosteus pungitius), with a
review of methods used in studies of the food of fishes.
Journal of Animal Ecology 19: 36-58.
Hyslop, E. J. 1980. Stomach contents analysisa review of
methods and their application. Journal of Fish Biology
17:411-429.
Jaap, W. C. 1984. The ecology of the south Florida coral
reefs: a community profile. U. S. Fish and Wildlife Service.
FWS/OBS 82/08. 138 pages.
Jennings, C. A. 1985. Species profiles: life histories and
environmental reguirements of coastal fishes and
invertebrates (Gulf of Mexico) sheepshead. U. S. Fish
and Wildlife Service. Slidell, Louisiana. 10 p.
Kitching, R.L. 1983. Systems ecology. University of
Queensland Press, St. Lucia, Queensland, Australia, 280
pages.
Kleinbaum, D. G. and L. L. Kupper 1978. Applied regression
analysis and other multivariate methods. Duxbury Press,
Belmont, California 486 pages.
Lee, D. S., C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D.
E. McAllister, and J. R. Stauffer, Jr. 1980. Atlas of North
American freshwater fishes, North Carolina State Museum of
Natural History. Raleigh, North Carolina. 867 pages.
Lindall, W. N. Jr., J. R. Hall, W. A. Fable, and L. A.
Collins 1973. A survey of fishes and commercial
invertebrates of the nearshore and estuarine zone between
Cape Romano and Cape Sable, Florida. National Marine
Fisheries Service, Panama City, Florida, 62 p.

Little, M. C., P. J. Reay, S. J. Grove, 1988. The fish
community of an East African mangrove creek. Journal of
Fish Biology 32:729-747.
Livingston, R. J. 1982. Trophic organization of fishes in a
coastal seagrass system. Marine Ecology Progress Series 7:
1-12.
Loneragan, N. R. and I. C. Potter 1990. Factors influencing
community structure and distribution of different life-cycle
categories of fishes in shallow waters of a large Australian
estuary. Marine Biology 106: 25-37.
Luckhurst, B. E. and K. Luckhurst 1978. Analysis of the
influence of substrate variables on coral reef fish
communities. Marine Biology 49: 317-323.
McHugh, J. L. 1967. Estuarine nekton, in G. H. Lauff (Ed.)
Estuaries. American Association for the Advancement of
Science, Washington, D. C. pages 581-620.
Mclvor, C. C. and W. E. Odum 1988. Food, predation risk,
and microhabitat selection in a marsh fish assemblage.
Ecology 69(5): 1341-1351.
McNeely, D. L. 1987. Niche relations within an Ozark
cyprinid assemblage. Environmental Biology of Fishes
18(3):195-208.
Miller, J. M. and M. L. Dunn 1980. Feeding strategies and
patterns of movement in juvenile estuarine fishes. in
Estuarine Perspectives. Academic Press, Inc. New York, New
York. 437-448 pages.
Montague, C. L., R. D. Bartelson, and J. A. Ley 1989.
Assessment of benthic communities along salinity gradients
in northeast Florida Bay, University of Florida,
Gainesville, for the South Florida Water Management
District, West Palm Beach, Florida.
Morton, R. M. 1990. Community structure, density, and
standing crop of fishes in a subtropical Australian mangrove
area. Marine Biology 105: 385-394.
Moser, M. L., L. R. Gerry 1989. Differential effects of
salinity changes on two estuarine fishes, Leisostomus
xanthurus and Micropogonius undulatus. Estuaries 12(1):35-
41.

Moyle, P. B. and T. J. Cech, Jr. 1988. Fishes: an
introduction to Ichthyology. Prentice Hall, Englewood
Cliffs, New Jersey, 559 pages.
Nielsen, L. A. and D. L. Johnson 1983. Fisheries
Techniques. American Fisheries Society, Bethesda, Maryland.
468 pages.
Nordlie, F. G., and S. J. Walsh 1989. Adaptive radiation in
osmotic regulatory patterns among three species of
Cyprinodontids (Teleostei: Atherinomorpha). Physiological
Zoology 62(6):1203-1218.
Odum, E. P. 1983. Basic Ecology. CBS College Publishing,
New York, New York. 613 pages.
Odum, W. E. 1971. Pathways of energy flow in a south
Florida estuary. Ph.D. dissertation, University of Miami,
Coral Gables, Florida. 162 pages.
Odum, W. E. and E. J. Heald 1972. Trophic analysis of an
estuarine mangrove community. Bulletin of Marine Science
22(3): 671-738.
Odum, W. E., C. C. Mclvor, and T. J. Smith, III. 1982. The
ecology of the mangroves of south Florida: a community
profile. U. S. Fish and Wildlife service, Office of
Biological Services, Washington, D. C. FWS/OBS-81/24. 144
pages.
Phillips, R. R. and S. B. Swears 1979. Social hierarchy,
shelter use, and avoidance of predatory toadfish (Opsanus
tau) by the striped blenny (Chasmodes bosquianus). Animal
Behavior 27: 1113-1121.
Pinto, L. 1987. Environmental factors influencing the
occurance of juvenile fishes in the mangroves of Pagbilao,
Phillipines. Hydrobiologia 150:283-301.
Power, P. E. and W. J. Matthews 1983. Algae-grazing minnows
(Campostoma anomalum), piscivorous bass (Micropterus spp.),
and the distribution of attached algae in a small prairie-
margin stream. Oecologia (Berlin) 60: 328-332.
Provost, M. W. 1973. Mean high water mark and use of
tidelands in Florida. Florida Scientist 36(1). 50-66.
Reid, R. K. 1954. An ecological study of the Gulf of Mexico
fishes in the vicinity of Cedar Key, Florida. Bulletin of
Marine Science 4(1): 1-94.
Remane, A. and C. Schlieper 1971. Biology of brackish
water. Wiley Interscience Division, New York, New York.

Robblee, M. B. 1987. The spatial distribution of the
nocturnal fish fauna of a tropical seagrass feeding ground.
Ph.D. Dissertation, University of Virginia, Charlottesville,
Virginia. 144 pages.
Robertson, A. I. and N. C. Duke 1987. Mangroves as nursery
sites: comparisons of the abundance and species compostion
of fish and crustaceans in mangroves and other nearshore
habitats in tropocal Australia. Marine Biology 96: 193-205.
Robertson, A. I. and N. C. Duke 1990a. Mangrove fish-
communities in tropical Queensland, Australia: spatial and
temporal patterns in densities, biomass and community
structure. Marine Biology 104: 369-379.
Robertson, A. I. and N. C. Duke 1990b. Recruitment, growth
and residence time of fishes in a tropical Australian
mangrove system. Estuarine, Coastal and Shelf Science 31:
723-743.
Robins, C. R., G. C. Ray, J. Douglass, 1986. A field guide
to Atlantic Coast fishes of North America. Houghton Mifflin
Company, Boston. 355 pages.
Rogers, T. E. Target, S. B. Van Sant 1984. Fish-nursery use
in Georgia salt marsh estuaries: the influence of springtime
freshwater conditions. Transactions of the American
Fisheries Society 113: 595-606.
Rozas, L. P. and M. W. LaSalle 1990. A comparison of diets
of Gulf killifish, Fundulus grandis Baird and Girard,
entering and leaving a Mississippi brackish marsh.
Estuaries 13(3):332-336.
Rozas, L. P. and W. E. Odum 1988. Occupation of submerged
aquatic vegetation by fishes: testing the roles of food and
refuge. Oecologia 77: 101-106.
Rutherford, E. S., J. T. Tilmant, E. B. Thue and T. W.
Schmidt, 1989. Fishery harvest and population dynamics of
spotted seatrout, Cynoscion nebulosus, in Florida Bay and
adjacent waters. Bulletin of Marine Science 44(1):108-125.
Sale, P. F. 1980. Assemblages of fish on patch reefs
predictable or unpredictable? Environmental Biology of
Fishes 5(3): 243-247.
Sale, P. F. and W. A. Douglas 1984. Temporal variability in
the community structure of fish on coral patch reefs and the
relation of community structure to reef structure. Ecology
65(2): 409-422.
Salini, J.P., S.J.M. Blaber and D. T. Brewer 1990. Diets of
piscivorous fishes in a tropical Australian estuary, with

special reference to predation on penaeid prawns. Marine
Biology 105: 363-374.
Schlosser, I. J. 1987. A conceptual framework for fish
communities in small warmwater streams. In: Matthews, W. J.
and D. C. Heins (Eds.) Community and evololutionary ecology
of North American stream fishes. Oklahoma University Press,
Norman, Oklahoma, pages 17-26.
Schmitt, R. J. and S. J. Holbrook 1985. Patch selection by
juvenile black surfperch (Embiocidae) under variable risk:
interactive influence of food quality and structural
complexity. Journal of Experimental Marine Biology and
Ecology 85: 269-285.
Schomer, N. S. and R. D. Drew 1982. An ecological
characterization of the lower Everglades, Florida Bay, and
the Florida Keys. U. S. Fish and Wildlife Service, Office
of Biological Services, Washington, D. C. 24688.
Seaman, W. Jr. and M. Collins 1983. Species profiles: life
histories and environmental requirements of coastal fishes
and invertebrates (South Florida) snook. U. S. Fish and
Wildlife Service. Slidell, Louisiana. 16 pages.
Sheridan, P. F. 1991. Comparative habitat utilization by
estuarine macrofauna within the mangrove ecosystem of
Rookery Bay, Florida. Bulletin of Marine Science 49(3).
Shulman, M. J. 1985. Recruitment of coral reef fishes:
effects of distribution of predators and shelter. Ecology
66(3): 1056-1066.
Smith III, T. J. 1987. Seed predation in relation to tree
dominance and distribution in mangrove forests. Ecology
68(2): 266-273.
Smith, III T. J. and N. C. Duke 1987. Physical determinants
of inter-estuary variation in mangrove species richness
around the tropical coastline of Australia. Journal of
Biogeography 14: 9-19.
Sogard, S. M. and K. W. Able 1991. A comparison of
eelgrass, sea lettuce macroalgae, and marsh creeks as
habitats for epibenthic fishes and crustaceans. Estuarine,
Coastal and Shelf Science 33: 501- 519.
Sogard, S. M., G. V. N. Powell, and J. G. Holmquist 1987.
Epibenthic fish communities on Florida Bay banks: relations
with physical parameters and seagrass cover. Marine Ecology
Progress Series 40:25-39.
Sogard, S. M., G. V. N. Powell, and J. G. Holmquist 1989a.
Spatial distribution and trends in abundance of fishes

residing in seagrass meadows on Florida Bay mudbanks.
Bulletin of Marine Science 44(1): 179-199.
Sogard, S. M., G. V. N. Powell, and J. G. Holmquist 1989b.
Utilization by fishes of shallow, seagrass-covered banks in
Florida Bay: 1. species composition and spatial
heterogeneity. Environmental Biology of Fishes 24(1): 53-
65.
Sogard, S. M., G. V. N. Powell, and J. G. Holmquist 1989c.
Utilization by fishes of shallow, seagrass-covered banks in
Florida Bay: 2. diel and tidal patterns. Environmental
Biology of Fishes 24(2): 81-92.
South Florida Water Management District 1989. Surface water
improvement and management (SWIM) plan for the Everglades:
vol. II. planning and implementation. West Palm Beach,
Florida
Starck, W. A. II, and R. E. Schroeder, 1971. Investigations
on the gray snapper, Lutjanus griseus. University of Miami
Press, Coral Gables, Florida 150 pages.
Stoner, A. W. 1986. Community structure of the demersal
fish species of Laguna Joyuda, Puerto Rico, Estuaries 9(2):
142-152.
Sutherland, J. P. 1980. Dynamics of the epibenthic
community on roots of the mangrove Rhizophora mangle, at
Bahia de Buche, Venezuela. Marine Biology 58, 75-84.
Tabb, D. C. T. R. Alexander, T. M. Thomas, N. Maynard
1967. The physical, biological and geological character of
the area south of C-lll Canal in extreme southeast
Everglades National Park, University of Miami, Coral Gables.
Tabb, D. C., D. L. Dubrow, and R. B. Manning 1962. The
ecology of northern Florida Bay and adjacent estuaries.
Florida Board of Conservation Technical Series no. 39,
Tallahassee.
Thayer, G. W. and A. J. Chester 1989. Distribution and
abundance of fishes among basin and channel habitats in
Florida Bay. Bulletin of Marine Science 44: 200-219.
Thayer, G. W., D. R. Colby and W. F. Hettler, Jr., 1987a.
Utilization of the red mangrove prop root habitat by fishes
in south Florida. Marine Ecology Progress Series 35:25-38.
Thayer, G. W., W.. F. Hettler, A. J. Chester, D. R. Colby
and P. J. McElhaney 1987b. Distribution and abundance of
fish communities among sected estuarine and marine habitats
in Everglades National Park. South Florida Research Center
Report SFRC-87/02. 166 pages.

171
Van Engle, W. A. and E. B. Joseph 1967. Characteristics of
coastal and estuarine nursery grounds as natural
communities. Unpublished report to U.S. Fish and Wildlife
Service, 43 pages. Available from: Virginia Institute of
Marine Science, Glouchester Point, Virginia.
Weinstein, M. P. 1979. Shallow marsh habitats as primary
nurseries for fishes and shellfish, Cape Fear River, North
Carolina. Fishery Bulletin 77(2): 339-357.
Weinstein, M. P. and R. W. Davis 1980. Collection
efficiency of seine and rotenone samples from tidal creeks,
Cape Fear, N. C., Estuaries 3(2):98-105.
Weisberg, S. B. and V. A. Lotrich 1982. Ingestion,
egestion, excretion, growth, and conversion efficiency for
the mummichog, Fundulus heteroclitus. Journal of
Experimental Marine Biology and Ecology 62: 237-249.
Werner, E., J. F. Gilliam, D. J. Hall and G. G. Mittelbach
1983. An experimental test of the effects of predation risk
on habitat use by fish. Ecology 64: 1540-1548.
Wilson, K. A. 1989. Ecology of mangrove crabs: predation,
physical factors and refuges. Bulletin of Marine Science
44(1): 263-273.
Wilson, K. A., K. W. Able, K. L. Heck 1990. Predation rates
on juvenile crabs in estuarine nursery habitats: evidence
for the importance of macroalgae (Ulva lactuca). Marine
Ecology-Progress Series 58: 243-251.
Wilson, K. A., K. L. Heck and K. W. Able 1987. Juvenile
blue crab, Callinectes sapidus, survival: an evaluation of
eelgrass, Zostera marina, as refuge. Fishery Bulletin
85(1): 53-58.
Yanez-Arancibia, A., F. A. Linares, and J. W. Day, Jr. 1980.
Fish community structure and function in Trminos Lagoon, a
tropical estuary in the southern Gulf of Mexico. Estuarine
Perspectives, Academic Press, Inc. 465-482.
Zieman, J. C., J. W. Fourqurean and R. L. Iverson 1989.
Distribution, abundance and productivity of seagrasses and
macroalgae in Florida Bay. Bulletin of Marine Science
44(1): 292-311.

BIOGRAPHICAL SKETCH
Janet Ann Ledtke Ley was born 11 June, 1951, in
Detroit, Michigan, to Frederick G. and Helen M. Ledtke.
Janet graduated from Rochester High School, Rochester,
Michigan, in 1969. She received her Bachelor of Science
degree in resource development at Michigan State University,
East Lansing, Michigan, in 1973.
Janet devoted ten years to environmental planning for
the Pinellas County government, in Clearwater, Florida, from
1974 through 1984. While working as a planner, in 1979, she
earned her Master of Science degree at the University of
South Florida. Her thesis was entitled "Exploring Transfer
of Development Rights," a concept that was later
incorporated into her work on Pinellas County's plan for
environmental protection of wetland ecosystems. Janet
worked as a consultant for the Tampa office of Dames & Moore
during 1985.
In 1986, Janet enrolled in the Ph.D. program in systems
ecology, in the College of Environmental Engineering
Sciences, at the University of Florida, Gainesville,
Florida. In May 1992, Janet received her Ph.D., and hopes
to continue to work in ecosystems research.
172

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Clay/L. Montague^/ Chairman
Associate Professor of
Environmental Engineering Sciences
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
i-S- C rv^\'
Carole C. Mclvor, Cochairman
Assistant Professor of
Forest Resources and Conservation
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Professor of Zoology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Scientist
Environmental Engineering Sciences

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
^ llU0AA~ X
William Seaman, \Tr. /
Associate Professor of
Forest Resources and Conservation
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Nicholas Funicelli
Assistant Professor of
Forest Resources and Conservation
This dissertation was submitted to the Graduate Faculty
of the College of Engineering and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.
May 1992
Winfred M. Phillips
Dean, College of Engineering
Madelyn M. Lockhart
Dean, Graduate School



Distance
Visual Census
Presence
Up 1 Mid & Down
(X) Misdassified
Density
Mid & Down


Figure 2-12. Length-frequency histograms based on visual
census data. Adult size given whenever the information was
available from the literature.
a. Sphyraena barracuda (great barracuda), and
b. Centropomus undecimalis (snook).


138
fishing line through the lower jaw of a small fish (4-10
cm). The other end was looped over a 1.25 cm diameter
polyvinylchloride (PVC) pole that had been driven into the
substrate. Sixteen fish were tethered inside each enclosure
and checked at 3, 6 and 24 hour intervals. Although all
fish survived and remained securely tethered for three
hours, in these preliminary tests, one (Eucinostomus gula)
died after six hours and several other fishes died after 24
hours (Table 5-2). In addition, fish that died at the
bottom were quickly attacked by scavengers (e.g. crabs and
gastropods), possibly interfering with interpretation of
test results.
Since the study objective was to determine rates of
prey encounters with predators, not scavengers, these
results prompted further investigation into alternative
tether and stake designs. After several prototypes, an L-
shaped stake was made by joining two 1.0 m long PVC pipes
with an elbow (Figure 5-1). A hole was drilled at the free
end for attaching the tethered fish (as above) and the other
end was driven into the sediment. When deployed, the top
bar of the L-shaped stake was above the surface of the
water. To ensure that predators and not scavengers were
responsible for removing the prey, the tethered fish were
forced to remain above the substrate by adjusting the depth
to which the pole was driven into the substrate.


165
Decapod and stomatopod communities of seagrass-covered
mudbanks in Florida Bay: inter- and intra-bank heterogeneity
with special reference to isolated subenvironments.
Bulletin of Marine Science 44(1): 251-262.
Holmguist, J. G., G. V. N. Powell and S. M. Sogard 1989b.
Sediment, water level and water temperature characteristics
of Florida Bay's grass-covered mudbanks. Bulletin of Marine
Science 44(1): 348-364.
Hynes, H. B. N. 1950. The food of fresh-water sticklebacks
(Gasterosteus aculeatus and Pygosteus pungitius), with a
review of methods used in studies of the food of fishes.
Journal of Animal Ecology 19: 36-58.
Hyslop, E. J. 1980. Stomach contents analysisa review of
methods and their application. Journal of Fish Biology
17:411-429.
Jaap, W. C. 1984. The ecology of the south Florida coral
reefs: a community profile. U. S. Fish and Wildlife Service.
FWS/OBS 82/08. 138 pages.
Jennings, C. A. 1985. Species profiles: life histories and
environmental reguirements of coastal fishes and
invertebrates (Gulf of Mexico) sheepshead. U. S. Fish
and Wildlife Service. Slidell, Louisiana. 10 p.
Kitching, R.L. 1983. Systems ecology. University of
Queensland Press, St. Lucia, Queensland, Australia, 280
pages.
Kleinbaum, D. G. and L. L. Kupper 1978. Applied regression
analysis and other multivariate methods. Duxbury Press,
Belmont, California 486 pages.
Lee, D. S., C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D.
E. McAllister, and J. R. Stauffer, Jr. 1980. Atlas of North
American freshwater fishes, North Carolina State Museum of
Natural History. Raleigh, North Carolina. 867 pages.
Lindall, W. N. Jr., J. R. Hall, W. A. Fable, and L. A.
Collins 1973. A survey of fishes and commercial
invertebrates of the nearshore and estuarine zone between
Cape Romano and Cape Sable, Florida. National Marine
Fisheries Service, Panama City, Florida, 62 p.


Table 2-10. Repeated measures analysis of variance with density of benthic
forage fish as the dependent variable and gradient and system as the
independent variables. Samples were taken within mangrove habitats using
enclosure nets. Densities were transformed to logarithms prior to
calculations.
Floridichthys
carpi
Lucania parva
Poecilia latipinna
Source
df *
F
p
df*
F
P
df*
F
P
Between Stations:
Among gradient positions
2/12
3.29
0.0726
2/12
12.96
0.0010
2/12
5.96
0.0160
Among systems
1/12
33.49
0.0001
1/12
2.23
0.1610
1/12
6.00
0.0306
Gradient X System
2/12
21.21
0.0001
2/12
5.09
0.0251
2/12
1.71
0.2229
Within Stations:
Among months
10/120
7.43
0.0001
10/120
7.70
0.0001
10/120
5.73
0.0002
Month X Gradient
20/120
3.62
0.0001
20/120
1.99
0.0317
20/120
2.73
0.0002
Month X System
10/120
1.43
0.1777
10/120
1.54
0.1712
10/120
1.28
0.2827
Month X System X Gradient
20/120
2.14
0.0072
20/120
1.45
0.1561
20/120
0.93
0.5113
Multiple comparisons among
means for all months:
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Gradient positions
No differences
Down
>
Up & Mid
Mid
>
Up & down
Systems
West
>
East
No differences
West
>
East
Source degrees of freedom / error degrees of freedom
(ji
vo


Table 4-3. Comparison of food habits among the seasons.
Data used are percent composition of stomach contents transformed using an arcsin
sguare-root function. MANOVA was used to test the hypothesis that season
was a significant source of diet variation. Specific F-tests to contrast means for
each pair of seasons were used to make multiple comparisions.
Species
Independ.
variable
Manova results
(Wllks lambda)
Dependent
variables
Univariate results
Contrast
summary**
F-value
p-value
F-value
p-value
df*
Floridichthys
season
1.95
0.0101
amphipods
1.32
0.2672
3/319
carpi
copepods
2.52
0.0581
3/319
O8tracods
3.28
0.0213
3/319
all>8um
nematodes
1.91
0.3122
3/319
forams
0.85
0.4672
3/319
algae
4.20
0.0062
3/319
wln&spr>sum&fall
Eucinostomus
season
3.34
0.0032
amphipods
0.61
0.6070
3/171
harengulus
algae
6.09
0.0006
3/171
spr>all
Fundulus
season
1.33
0.1745
amphipods
1.60
0.1896
3/245
grandis
isopods
0.17
0.9163
3/245
fish (unidentified)
0.05
0.9862
3/245
insects (terrestrial adults)
3.11
0.0272
3/245
spr>all
algae
1.80
0.1482
3/245
Lutjanus
season
1.25
0.2232
amphipods
0.82
0.4858
3/73
griseus
crabs
3.56
0.0183
3/73
all>win
isopods
1.23
0.3056
3/73
shrimp
0.73
0.5395
3/73
fishs (unidentified)
0.13
0.9391
3/73
algae
0.84
0.4757
3/73
Strongylura
season
3.03
0.0001
amphipods
1.59
0.1907
3/312
notata
isopods
4.40
0.0048
3/312
win>all
shrimp
2.17
0.0914
3/312
ants
1.74
0.1589
3/312
insects (terrestrial adults)
6.16
0.0004
3/312
spr>all
fish (unidentified)
0.93
0.4287
3/312
Sphyraena
season
1.21
0.2691
amphipods
0.55
0.6491
3/60
barracuda
eggs
1 65
0.1864
3/60
Cyprinodontids
1.94
0.1333
3/60
fish (unidentified)
0.97
0.4110
3/60
algae
0.76
0.5215
3/60
Model degrees of freedom / error degrees of freedom
** If no contrast is indicated, no significant differences occurred.
128


43
substantial decrease (from 35.0 ppt to 10.4 ppt) was evident
in June 1990, that did not occur in stations sampled in the
upstream/west location (which became increasingly
hypersaline).
None of these salinity changes correspond with patterns
observed for fish densities. Density of benthic forage fish
peaked in winter months at four of the six general locations
(Figure 2-3). The highest density collection (13.6 fish m-
O ...
) was at the mid-Trout Cove station in winter 1989; lowest
density occurred at the mid-Little Blackwater Sound in the
June 1989 (0.12 fish m-2) .
Water column forage fish. In Figure 2-4, one can
compare changes in salinity with changes in density from the
enclosure net sampling; again, however, no consistent
patterns emerge. Density of water column fish was highly
variable and the graphs illustrate no consistent seasonal
patterns. In general, either very low or very high
densities of these schooling fishes were collected. The
highest density collection (25.3 fish m-2) occurred at mid-
Little Blackwater Sound in September 1989. No water column
forage fish were collected in several samples. As with the
benthic forage fish, these density fluctuations were also
not related to the seasonal fluctuations in salinity.
Large roving fish. In Figure 2-5, changes in salinity
can be compared with changes in density for this group from
the visual census sampling; again, however, no consistent
temporal patterns emerge. In the upstream/west location,


11
sportfish populations have been linked to hypersalinity
stress for certain sportfish in Everglades National Park
(Rutherford et al. 1989). The Everglades estuaries are also
critical habitat areas for other endangered aquatic species
(e.g. American crocodile) that rely on the same forage base
as do birds and sportfish (SFWMD 1989).
Thus, groups concerned about these problems spurred
South Florida Water Management District (SFWMD) officials to
take action that would return more natural drainage patterns
to the estuarine areas of the Everglades. Some of these
actions have focused on the C-lll Canal/Taylor Slough
watershed which includes agricultural lands in a large
drainage basin east of the Park. The downstream leg of the
canal runs northwest to southeast, passes under U. S.
Highway 1 and continues southward outside of Florida Bay, to
Barnes Sound (Figure 1-1). In low flow periods, the canal
has functioned like a dike by preventing overland flow of
freshwater from reaching both the downstream prairies and
the approximately seven small creeks tributary to
northeastern Florida Bay. In high flow conditions, water
still flows through, sometimes sending slugs of freshwater
into northeastern Florida Bay. Local topographic conditions
tend to direct more freshwater toward U.S. Highway 1 than
toward the west (Tabb et al. 1967). Thus, under these
management conditions, the historic salinity regime is
likely to have been altered. Changes in hydroperiod have
probably resulted in more severe hypersaline conditions and


148
Table 5-5. Correlations of predator encounter rates
with environmental variables. Data are percentage of
fish missing (arcsine transformed) after being tethered
for three hours adjacent to mangrove edges in 16 trials.
Pearson correlation coefficients and probability
values are indicated.
Variable
Test
Results
Correlation
Value (R)
p-value
Mean Water
-0.1463
0.5885
Depth
Salinity
0.3688
0.1597
Secchi
0.1308
0.6293
Distance


CHAPTER 6
IMPLICATIONS AND CONCLUSIONS
Implications for Mangrove Fish Ecology
Fish and Mangrove Shorelines
Mangrove shorelines fulfill the food and cover habitat
requirements for many fishes. Among the functional roles of
inter-tidal mangroves as habitats for fish, the best
established is that of a temporary feeding location for a
wide range of species (Robertson & Duke 1990a, Blaber et al.
1985, Morton 1990). These species include small schooling
planktivores, benthic forage fish, and large piscivores.
They spend periods of low tide in deep open water or shallow
ponds and feed in the mangroves when the tide inundates the
forest (Davis 1988).
Since northeastern Florida Bay is non-tidal, the
mangrove habitats are unlike those in many other areas. In
Florida Bay permanently inundated mangrove habitats are
believed to provide young fish with refuge from larger
predators (Thayer et al. 1987a). This role was supported in
the current study, particularly for larger juveniles of the
154


Table 5-2. Results of tests using tethered fish within an enclosure
formed by two block nets at Buttonwood Sound. Species used were:
Fundulus granis, Eucinostomus gula, and Floridichthys carpi.
Test
number
After 3
hours
After 6
hours
Live
After 24
Dead
hours
Missing
Live
Dead
Missing
Live
Dead
Missing
1
16
0
0
15
1
0
8
0
8*
2
16
0
0
16
0
0
14
0
2
3
16
0
0
16
0
0
3
12**
1
* Net partially down overnight; needlefish found inside net
** Fish that had died were being consumed by scavengers (e.g. anemones, gastropods,
and crabs)


149
Thus, although the predator encounter rate was greater
at the two far/upstream locations, salinity, visibility, and
water depth were not unusual at these locations relative to
the others. Salinity, the parameter of most interest, was
lower at both the far/up- and mid/up- locations, but the
rate of predation was lower far/upstream, and higher at the
mid/upstream location.
Potential predators that were observed approaching
tethered fish at all the locations included redfin
needlefish, gray snapper, and barracuda (Table 5-6).
Observed only in the far/upstream locations were bull shark
juveniles, gar, alligators, and turtles.
A total of 28 actual predation events were recorded in
which the attacking predator could be identified (Table 5-
6). Most of these events (22) were due to needlefish
(Strongylura notata), half of which became tethered
themselves. Once they had swallowed the prey, the
needlefish were unable to cut the fishing line with the
teeth on their elongated jaw, but they usually were
eventually able to work the prey and themselves free of the
tether (e.g. by leaping).
Discussion
Predator encounter rates averaged more than 50%, even
in the most upstream locations. Thus, piscivorous predators
are ubiquitous along the estuarine gradient and were
effective at consuming tethered prey throughout the system.


133
(Weisberg & Lotrich 1982). In an experiment, Fundulus
heteroclitus, fed a diet of only fiddler crabs (Uca pugnax)
lost weight (Weisberg & Lotrich 1982). In a salt marsh
where fiddler crabs comprised 17 times more food volume than
any other item for Fundulus granis, the killifish may have
compensated for the crabs' low caloric value by consuming
their prey in very large quantities (Rozas & LaSalle 1990).
Since two items consumed more abundantly upstream,
algae and crabs, were lower quality energy sources for fish
growth, these locations may be somewhat lower quality
habitats than mid- and downstream in terms of obtainable
foods for these species. The fishes that forage upstream
may compensate, however, by foraging on greater quantities
of the lower quality items that are available.


103
for comparisons of mangrove habitat/fish relationships
between fish groups and species (Robblee 1987).
Results
Relationships Among the Habitat Variables
Correlations among habitat variables. For visual
censuses and enclosure net data, mean values and
correlations among the habitat variables differed (Tables 3-
1 and 3-2). The visual census sites had greater water
depths and tree heights but less cover than the enclosure
net sites. This difference was probably due to the broader
range of mangrove habitats covered in the visual census.
The greatest correlation between habitat variables was found
for submerged aquatic vegetation and salinity, with greater
volumes occurring where salinity means were higher (Table 3-
2). Salinity means were inversely correlated with salinity
variation, indicating that where salinity mean was higher,
salinity variation was less.
Underlying factors among habitat variables. Results of
the principal components analysis and correlations of those
components with each habitat variable differed for the
enclosure nets (Table 3-3) and for visual the censuses
(Table 3-4). The first 3 principal components explained
86.2% of the variance in the original habitat data
associated with enclosure nets, and 79.7% of the variance in
the enclosure net data. For both data sets, the first
principal component explained 52% of the variation.


Table 2-6. Repeated measures analysis of variance with densities of fish as dependent
variables and gradient and system as independent variables. Benthic and water column
forage fish were collected using enclosure nets. Large roving fish were sampled using
visual census techniques. Data were transformed to logarithms prior to performing
calculations.
Source
Benthic forage
fish
Water Column Forage Fish
Large
Roving
Fish
df*
F
P
df*
F
P
df*
F
P
Between Stations:
Among gradient positions
2/12
0.16
0.8511
2/12
4.17
0.0421
2/12
6.30
0.0135
Among systems
1/12
3.39
0.0904
1/12
0.38
0.5514
1/12
0.22
0.6460
Gradient X System
2/12
7.56
0.0075
2/12
8.39
0.0052
2/12
0.65
0.5393
Within stations
Among months
10/120
9.83
0.0001
10/120
5.16
0.0001
4/48
2.17
0.0863
Month X Gradient
20/120
1.15
0.3131
20/120
4.92
0.0001
8/48
1.29
0.2707
Month X System
10/120
1.48
0.1627
10/120
1.5
0.1486
4/48
1.56
0.2015
Month X Gradient X System
20/120
1.07
0.3942
20/120
4.61
0.0001
4/48
0.97
0.4707
Multiple comparisons among
means for all months:
Location
Sign. Location
greater
than
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Gradient positions
No differences
Down
>
Up
Mid & Down
>
Up
Systems
No differences
No differences
No differences
Source degrees of freedom / error degrees of freedom


Table 4-2. Comparison of food habits among gradient positions and systems.
Data used are percent composition of stomach contents transformed using an arcsin
square-root function. MANOVA was used to test the hypothesis that gradient and system
were significant sources of diet variation. Specific F-tests to contrast means for
each pair of gradient positions were used to make multiple comparisions.
Species
Independ.
variable
Manova results
(Wilks lambda)
Dependent
variables
Univariate results
Contrast
summary"
F-value
p-value
F-value
p-value
df'
Floridichthys
gradient
7.72
0.0001
amphipods
5.59
0.0001
5/317
carpi
system
2 52
0.0251
copepods
9.06
0.0001
5/317
down > up & mid
gradient* system
6.73
0.0001
ostracods
10.01
0.0001
5/317
down > up & mid
nematodes
4.37
0.0007
5/317
down > up & mid
forams
2.43
0.0352
5/317
algae
14.41
0.0001
5/317
up > mid & down
Eucinostomus
gradient
6.59
0.0001
amphipods
5.88
0.0001
5/169
down > up & mid
harengulus
system
6.00
0.0030
algae
4.61
0.0006
5/169
up > mid & down
gradient'system
1.81
0.1263
Fundulus
gradient
1.12
0.3473
amphipods
1.84
0.1054
5/243
grandis
system
0.54
0.7452
isopods
1.71
0.1328
5/243
gradient'system
1.52
0.1292
fish (unidentified)
1.06
0.3822
5/243
insects (terrestrial adults)
0.64
0.6716
5/243
algae
0.72
0.6122
5/243
Lutjanus
gradient
4.93
0.0001
amphipods
3.09
0.0139
5/76
griseus
system
0.70
0.6473
crabs
15.29
0.0001
5/76
up > mid & down
gradientsystem
1.65
0.0859
isopods
1.34
0.2568
5/76
shrimp
0.91
0.4806
5/76
fishs (unidentified)
2.52
0.0376
5/76
algae
2.39
0.0465
5/76
Strongylura
gradient
0.58
0.8567
amphipods
1.51
0.1868
5/310
notata
system
1.68
0.1241
isopods
1.95
0.0859
5/310
gradient'system
1.76
0.0510
shrimp
2.75
0.0189
5/310
ants
0.27
0.9300
5/310
insects (terrestrial adults)
1.62
0.1548
5/310
fish (unidentified)
0.71
0.6153
5/310
Sphyraena
gradient
1.06
0.3950
amphipods
1.23
0.3062
5/58
barracuda
system
1.35
0.2578
eggs
1.24
0.3011
5/58
gradient'system
2.07
0.0325
Cyprinodontids
3.97
0.0036
5/58
fish (unidentified)
0.55
0.7373
5/58
algae
1.17
0.3367
5/58
Model degrees of freedom / error degrees of freedom
** If no contrast is indicated, either an interaction or no significant differences occurred.
127


132
Influence of the Salinity Gradient
Diets of three species were significantly influenced by
location along the salinity gradient. Eucinostomus
harengulus and Floridichthys carpi diets included more
algae upstream and more copepods, ostracods, and nematodes
downstream. This finding tends to be consistent with the
reduced levels of benthic invertebrate populations found
upstream in a previous investigation in northeastern Florida
Bay (Montague et al. 1989). Plant materials, other than
seeds, are generally less concentrated energy sources than
animal sources (Odum 1983) .
Lutjanus griseus consumed more crabs up- than
downstream. Although all the parts were not identified to
species in the current study, most were probably from mud
crabs of the species Rhithropanopeus harrisi. This brackish
water crab was abundantly collected in minnow traps placed
in mangrove shorelines upstream, but was never collected
downstream (unpublished data). It occurred abundantly in
the gray snapper stomachs in the North River (Odum 1971).
Rhithropanopeus harrisi is, thus, an important forage item
for snappers that appears to occur abundantly in brackish
water in the study area. As mean salinity rose from 16 to
over 30 ppt in northeastern Florida Bay, these crabs were
virtually eliminated from seagrass beds over a period of
four years (Holmguist et al. 1989a).
Crabs in general may not be readily assimilated as food
by fishes due to a higher percentage of exoskeleton


Figure 2-9. Mean density of fish by general location for the three most abundant
species in the large roving fish group. Error bars illustrate the magnitude of
the standard deviation in density over all the months.


Percent Percent
86
Residency
Occasional Transient Resident
Number of Species
100 F i= =i
Up Mid Down Up Mid Down
West East


101
Tree cover. The percent of canopy coverage over the
frame as observed from the surface of the water
was estimated.
Prop root size and density. Calipers were used to
measure the diameters of red mangrove prop roots.
Observations were divided into two groups: those
less than and those greater than 2 cm in diameter.
Numbers in each category were counted within each
frame. Few prop roots less than 2 cm occurred
downstream. To avoid biasing the analysis, only
the data for prop roots greater than 2 cm in
diameter were used.
Submerged aquatic vegetation. Within each frame, the
total volume of seagrass, algae and detrital
material was determined separately by placing a
meter stick in each quadrant and noting the height
of each component above the substrate. The
percent areal coverage of each component was also
estimated for each quadrant. These values were
summed for the analysis.
Salinity mean: The average value for all months at
each station was determined from measurements
taken with a refractometer.
Salinity variation: The standard deviation for each
station was calculated from the concurrent
measurements.
Analysis
Datasets containing 136 visual census substations and
the 18 enclosure net stations were analyzed separately. For
comparison, means and standard deviations were determined
for each variable. To select the most appropriate method of
analysis, correlations among the variables were also
calculated.
Multi-colinearity among several of the variables was
revealed in the correlation analysis. Thus, principal
components analysis, a multivariate technique designed to
discern factors that have generated interdependence within a
data set was used (Afifi & Clark 1984, Robblee 1987,


Table 2-12. Repeated measures analysis of variance with density of water
column forage fish as dependent variables and gradient and system as independent
variables. Samples taken within mangrove habitats using enclosure nets.
Densities were transformed to logarithms prior to calculations.
Anchoa mitchelli
Atherinomorus
stipes
Menidia spp.
Source
df*
F
P
df*
F
P
df*
F
P
Between Stations:
Among gradient positions
2/12
10.95
0.0020
2/12
116.61
0.0001
2/12
13.44
0.0009
Among systems
1/12
13.41
0.0031
1/12
20.32
0.0007
1/12
1.62
0.2272
Gradient X System
2/12
11.32
0.0017
2/12
2.81
0.1001
2/12
0.51
0.6136
Within Stations:
Among months
10/120
4.73
0.0060
10/120
8.54
0.0001
10/120
3.05
0.0065
Month X Gradient
20/120
4.95
0.0007
20/120
5.46
0.0001
20/120
2.62
0.0033
Month X System
10/120
4.95
0.0047
10/120
2.46
0.0389
10/120
4.44
0.0006
Month X System X Gradient
20/120
4.43
0.0015
20/120
3.00
0.0031
20/120
2.70
0.0025
Multiple comparisons among
means for all months:
Location Sign. Location
greater
than
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Gradient positions
Mid
>
Up & down
Down
>
Up & Mid
Up
>
Mid & down
Systems
East
>
West
West
>
East
No diffs
>
East
Source degrees of freedom / error degrees of freedom
o\
u>


Fish per square meter
Density by General Location
Top 3 Benthic Forage Fish Species
West Enclosure Nets East
2.5
2.0
1.5
1.0
0.5
0.0
2.5
2.0
1.5
1.0
0.5
0.0
2.5
2.0
1.5
1.0
0.5
0.0
Floridichthys Lucarna Poecilia Floridichthys Lucania Poecilia
carpi parva latipinna carpi parva latipinna
(ji
oo


87
to those found by other investigators who targeted small
fish (less than about 7.0 cm in total length) in vegetated,
shallow areas using (Weinstein & Davis 1980). Although
rotenone reguired cautious handling, fish immediately began
re-occupying sites after nets were removed. Rapid rotenone
degradation in the relatively high temperatures which
prevailed in these waters (Nielson and Johnson 1983), and
dilution due to the turnover of water at the relatively open
sites prevented cumulative adverse rotenone effects.
Other factors could have affected the variable
efficiency of recovery from individual nets. Turbidity from
disturbed sediments reduced the ability of the collectors to
capture fish with dip nets. In addition, variation in winds
onto the sites may have caused rotenone effectiveness to
vary.
Visual census methodology. The visual census
methodology had several advantages for sampling fish in
mangrove habitats. The speed and flexibility of the method
permitted sampling in a broad range of habitats that could
not be sampled with the net, including mangrove locations
with greater depths and a wider fringe. The visual method
was non-destructive, thus allowing evaluation of the
persistence of use by repeated observations of the same
stations and fishes. The problem of net avoidance was
eliminated with this method. In addition, large fish (e.g.
tarpon) that would normally escape nets and trawls could be
surveyed. As in studies of streams and coral reefs in which


137
This experiment actually integrates several of the
steps involved in predation that were described by Holling
(1966). First, measurements indicate whether or not
predators occur in the locations under comparison.
Secondly, the tests indicate how well the predators can
perceive the prey, given the conditions at the site (e.g.
turbidity levels). Thirdly, the ability and propensity to
consume particular prey species relative to size and
palatability is indicated.
To add to the understanding of the influence of
freshwater inflow on fish assemblages, the objective of this
chapter was to compare predator encounter rates up- and
downstream during the rainy season. If the paradigm is
true, fewer tethered prey should be consumed upstream than
downstream, as the marine predators are excluded from these
locations due to the lower, more variable salinity
conditions.
Materials and Methods
Preliminary Tests
To evaluate the effectiveness of the tethering
technigue for specific prey fishes and conditions in the
study area, preliminary tests were conducted within an
enclosure formed from two 30 m seine nets that excluded all
potential predators. For these tests, fish were tethered by
sewing one end of a 1.0 m length of 8 lb test monofilament


Figure 1-2. Map of the Florida Bay study area.


171
Van Engle, W. A. and E. B. Joseph 1967. Characteristics of
coastal and estuarine nursery grounds as natural
communities. Unpublished report to U.S. Fish and Wildlife
Service, 43 pages. Available from: Virginia Institute of
Marine Science, Glouchester Point, Virginia.
Weinstein, M. P. 1979. Shallow marsh habitats as primary
nurseries for fishes and shellfish, Cape Fear River, North
Carolina. Fishery Bulletin 77(2): 339-357.
Weinstein, M. P. and R. W. Davis 1980. Collection
efficiency of seine and rotenone samples from tidal creeks,
Cape Fear, N. C., Estuaries 3(2):98-105.
Weisberg, S. B. and V. A. Lotrich 1982. Ingestion,
egestion, excretion, growth, and conversion efficiency for
the mummichog, Fundulus heteroclitus. Journal of
Experimental Marine Biology and Ecology 62: 237-249.
Werner, E., J. F. Gilliam, D. J. Hall and G. G. Mittelbach
1983. An experimental test of the effects of predation risk
on habitat use by fish. Ecology 64: 1540-1548.
Wilson, K. A. 1989. Ecology of mangrove crabs: predation,
physical factors and refuges. Bulletin of Marine Science
44(1): 263-273.
Wilson, K. A., K. W. Able, K. L. Heck 1990. Predation rates
on juvenile crabs in estuarine nursery habitats: evidence
for the importance of macroalgae (Ulva lactuca). Marine
Ecology-Progress Series 58: 243-251.
Wilson, K. A., K. L. Heck and K. W. Able 1987. Juvenile
blue crab, Callinectes sapidus, survival: an evaluation of
eelgrass, Zostera marina, as refuge. Fishery Bulletin
85(1): 53-58.
Yanez-Arancibia, A., F. A. Linares, and J. W. Day, Jr. 1980.
Fish community structure and function in Trminos Lagoon, a
tropical estuary in the southern Gulf of Mexico. Estuarine
Perspectives, Academic Press, Inc. 465-482.
Zieman, J. C., J. W. Fourqurean and R. L. Iverson 1989.
Distribution, abundance and productivity of seagrasses and
macroalgae in Florida Bay. Bulletin of Marine Science
44(1): 292-311.


CHAPTER 1
GENERAL INTRODUCTION
The goal of improved management of surface waters to
benefit estuarine fish populations in Florida Bay provided
the incentive for this research. Before objectives and
strategies can be established toward this goal, however, a
better understanding of how freshwater inflow influences
fish communities in mangrove estuaries overall is needed.
This involves aspects of fish ecology of both estuarine and
mangrove ecosystems.
Estuarine Fish Ecology
Ecologists often divide estuarine fishes into three
groups: estuarine residents (complete their entire life
cycle in the estuary), estuarine transients (spawn offshore,
their young use the estuary as a nursery), and occasional
marine visitors (usually adults) (Day et al. 1989) .
Resident and transient species tend to be widespread, but
marine visitors are usually restricted to the higher
salinity zones of the lower estuary (Weinstein 1979) At
certain times of the year densities may increase
dramatically as influxes of transient juveniles enter
estuarine systems. Some species tend to migrate to
upstream-most habitats upon initially entering the estuary;
1


27
Study Sites. For visual censusing, site selection
criteria included adequate depth (20 to 100 cm), red
mangrove dominance, and wind-protection. At 70 m in length,
however, each transect encompassed a range of depths and
other physical characteristics. Four sites were randomly
selected for permanent sampling stations within each of the
six general locations. These sites were chosen in the same
general locations as the net sites but were more widespread
than the net stations (Figure 2-1).
Procedures. To prepare permanent census stations,
mangrove edge transects were designated with flagging tape
every 10 m along the 70 m edge. Physico-chemical variables
were recorded in conjunction with each visual census (e.g.
air temperature, salinity, wind conditions). Other
variables measured included: 1) water depth was recorded at
a permanent stake located in each transect; 2) abundance of
submerged aquatic vegetation adjacent to the transects was
noted using a scale of zero to three (abundant); and, 3)
range of visibility was determined by using a white PVC pole
set vertically into the mud and measuring the horizontal
distance at which the pole became visible as one snorkeled
towards it. If the visibility was less than 100 cm, the
census was rescheduled. If visibility was initially poor,
three attempts (on subsequent days) were made to conduct
surveys. However, in some months it was impossible to
conduct a visual census at particular sites.


140
Figure 5 1. Illustration and dimensions of the tethering
systems used in this study as deployed near mangrove edges.


12
sudden salinity changes of great magnitude, especially in
the eastern part of Florida Bay.
In the mid-1980s engineering alterations created
several cutouts, each 20 meters wide, in the south bank of
C-lll canal. The cutouts were intended to restore the more
dispersed and gradual pattern of freshwater inflow to
northeastern Florida Bay. Furthermore, an earthen plug was
installed to block the C-lll outfall to Barnes Sound except
on extreme floods when SFWMD can release water by opening it
with draglines. The result of these alterations was to
provide more flexible management of freshwater flow to
northeastern Florida Bay. The guestion remaining is how to
utilize this flexibility to improve ecological conditions.
Objectives
Fish and Salinity
The first study objective was to determine the extent
to which species composition and abundance were influenced
by salinity variability in the northeastern Florida Bay
study area. Because of direct and indirect salinity
influences, more variable fish abundances and distinct
community differences were expected at the upstream
locations over an annual cycle that included both wet and
dry seasonal differences in freshwater inflow. The eastern
portion of the study area was also expected to be distinctly
more variable than the western portion because of the
influence of the C-lll Canal.


129
all year except in winter, when they were seldom consumed.
For both Fundulus granis and Strongylura notata, adult
insects, a prevalent item, were greater in diets in the
spring.
Discussion
Shared Resources
As commonly occurs in aguatic habitats, diets
overlapped among the 6 species as resources were often
shared (Harrington & Harrington 1961, Livingston 1982, Odum
1983). This overlap occurred in all areas and throughout
the study period.
Diet overlap can become particularly evident when one
resource attains a periodic peak of abundance. In other
aguatic habitats, for example, populations of penaeid shrimp
(Salini et al. 1990) or larval insects (Harrington &
Harrington 1961) increase under certain conditions, and
opportunistic fish take advantage of the resource abundance.
A similar seasonal increase in exploitation of a particular
food resource was found in the current study for adult
insects exploited by Fundulus granis and Strongylura
notata.
In western Florida Bay, gray snapper and spotted sea
trout (Cynoscion nebulosus) diets overlapped when peak
abundances of pink shrimp (Penaeus duorarum) occurred in
November (Hettler 1989). Although a similar pattern might
have been expected in the current study, none became
evident. Migrating juvenile pink shrimp, while a major


Table 2-18. Comparison of abundance and number of species by residency,
fish group, and general location. Both sampling methods were combined
for the table.
Residency
Fish Group
General
Up-west Mid-west Down-west Up
Locations
-east Mid-east
Down-east
All
Residents: complete entire
life cycle
in the study
area
Benthic forage fish
abundance
15951
16049
13442 15712
6965
16329
84915
Benthic forage fish
species
24
14
21
21
18
22
33
Water column forage fish
abundance
4505
36460
36195
4532
23536
65843
171071
Water column forage fish
species
5
4
5
5
6
5
7
Large roving fish
abundance
236
603
1255
293
299
701
3387
Large roving fish
species
3
1
4
2
2
4
6
Transient Juveniles: juvenile offspring of species
that spawn
offshore
Benthic forage fish
abundance
Benthic forage fish
species
Water column forage fish
abundance
Water column forage fish
species
Large roving fish
abundance
101
5268
6470
25
1396
9832
23092
Large roving fish
species
3
8
7
2
6
8
9
Occasional Visitors: marine
and freshwater adults
that occupy
the study
area
Benthic forage fish
abundance
Benthic forage fish
species
Water column forage fish
abundance
3
1
1
2
603
610
Water column forage fish
species
1
1
1
1
1
2
Large roving fish
abundance
113
25
162
9
42
1263
1618
Large roving fish
species
5
5
9
6
5
9
19
-j
VO


Table 3-4. Results of the principal components analysis of the physical
and environmental variables associated with the 136 visual census stations
Correlations greater than 0.6 are underlined.
Principal
component
1
2
3
4
5
6
7
8
Eigenvalue
4.181
1.393
0.799
0.552
0.463
0.431
0.131
0.500
% Variance explained
52.300
17.400
10.000
6.900
5.800
5.400
1.600
0.600
Cumulative %
52.300
69.700
79.700
86.600
92.500
97.700
99.400
100.000
variance explained
Correlations (r) of the original variables with the
PC's
Water depth
0.408
0.068
0.141
-0.082
0.201
0.873
-0.010
0.002
Fringe width
0.354
0.161
0.167
-0.097
0.879
0.198
-0.008
0.002
Tree height
0.263
0.187
0.926
-0.047
0.146
0.123
-0.007
0.002
Prop roots
-0.054
-0.160
-0.040
0.980
-0.072
-0.061
0.003
-0.001
Submerged vegetation
0.896
0.064
0.176
-0.051
0.221
0.218
0.167
0.186
Salinity mean
0.926
0.006
0.160
-0.018
0.198
0.220
0.077
-0.148
Salinity variation
-0.900
0.034
-0.168
0.065
-0.161
-0.194
0.306
0.008
Tree cover
-0.002
0.962
0.165
-0.170
0.123
0.053
0.004
0.002
107


addition, my conversations with him gave me renewed
enthusiasm and perspective. Dewey Worth, Dan's colleague,
followed through with continued encouragement and support in
later phases of the project.
Laura Flynn and Luke Hasty were my very competent field
assistants, providing good humor and constructive
suggestions. They remained enthusiastic in every
circumstance, from diving with major unknown creatures, to
measuring 2-day old fish in 95 degree heat, to snorkeling in
double hoods and wetsuits.
Jacque Stevens, Harriett McCurdy and my brother, Fred
Ledtke, Jr., were my most faithful volunteers. Jacque
helped me tether over 200 fish and her ideas were
invaluable. Fred devoted his hard earned vacations to his
older sister's unusual effort.
I would also like to express my gratitude to the staff
of Everglades National Park. At the Key Largo Ranger
Station, Dave and Louise King, Linda Cramer, and Dave
Viscera included me as part of their small neighborhood
during my 2 year residency. I am grateful for their support
and rescues during boat break-downs. From the South Florida
Research Center, Mike Robblee allowed me to use Park boats,
provided insights concerning my research questions and
encouraged my efforts. DeWitt Smith also gave me
encouragement and perceptive advice. Bill Loftus' help in
iii


Figure 2-6. Mean density of fish by general location for
each fish group. Error bars illustrate the magnitude of the
standard deviation in density over all the months. Samples
of benthic and water column forage fish taken with enclosure
nets. Large roving fish were sampled with visual methods.


145
downstream locations contrasted to 55% for those
far/upstream locations (Figure 5-3). Salinity, range of
visibility, and water depth also varied across the gradient
(Figure 5-3).
The far/upstream predation rate was significantly lower
than the rates for the mid/up-, mid-, and downstream
locations, but there were no significant differences between
east and west systems (Table 5-4). Rates of predation were
not significantly different among the species (df=4, F=1.98,
p=0.1296).
Neither salinity, secchi distance nor water depth were
significantly correlated with observed predator encounter
rates (Table 5-5). These factors did vary significantly
among the locations, however.
Salinity means were not significantly different between
the far/up- and mid/upstream locations (p<0.3454, df=l), nor
between the mid- and downstream locations (p<0.3921, df =
1). However, the subgroup formed by the far/up- and
mid/upstream stations had a significantly lower mean
salinity than the subgroup formed by the mid- and downstream
locations (p<0.0001 for each pair of contrasts).
Range of visibility was significantly greater
downstream than at the other locations (p<0.005 for each
pair of contrasts). Water depth, however, was significantly
greater at the downstream/east location (p<0.05 for each
pair of contrasts).


Figure 2-7. Mean density of fish by general location for the three most abundant
species in the benthic forage fish group. Error bars illustrate the magnitude of
the standard deviation in density over all the months.


163
Deaton, L. E. and M. J. Greenberg 1986. There is no
horohalinicum. Estuaries. 9(1): 20-30.
DeSylva, D. P. 1963. Systematics and life history of the
great barracuda. Studies in Tropical Oceanography,
Institute of Marine Science, University of Miami, Miami,
Florida. 179 pages.
Dibble, E. 1991. A comparison of diving and rotenone
methods for determining relative abundance of fish.
Transactions of the American Fisheries Society (In press).
Ebeling, A. W. and D. R. Laur 1985. The influence of plant
cover on surfperch abundance at an offshore temperate reef.
Environmental Biology of Fishes 12(3): 169-179.
Flores-Verdugo, F. J., J. W. Day, L. Mee and R. Briseno-
Duenas 1988. Phytoplankton production and seasonal biomass
variation of seagrass, Ruppia maritima, in a tropical
Mexican lagoon with an ephemeral inlet. Estuaries 11(1):
51-56.
Flores-Verdugo, F., F. Gonzalez-Farias, O. Ramirez-Flores
1990. Mangrove ecology, aguatic primary productivity, and
fish community dynamics in the Teacapan-Aqua Brava Lagoon
estuarine system (Mexican Pacific). Estuaries 13(2): 219-
230.
Funicelli, N. A., H. E. Bryant, M. R. Dewey, G. M. Ludwig,
D. A. Meineke, L. J. Mengel, and J. E. Skjeveland 1986.
Movements, relative importance, and standing stock of
important sport and commercial species in Everglades
National Park, Florida. U. S. Fish and Wildlife Service,
Gainesville, Florida. 186 p.
Gilmore, R. G., C. J. Donohoe, D. W. Cooke 1983.
Observations on the distribution and biology of east-central
Florida populations of the common snook, Centropomus
undecimalis. Florida Scientist Special Supplement
45(3/4):313-336.
Ginsburg, R. N. 1956. Environmental relationships of grain
size and constituent particles in some south Florida
carbonate sediments. Bulletin of the American Association
of Petroleum Geologists. 40(10): 2384-2427.
Grossman, G. D., D. M. Nickerson, and M. C. Freeman 1991.
Principal component analysis of assemblage structure data:
utility of tests based on eigenvalues. Ecology 72(1):
341-347.
Gunter, G. 1938. Seasonal variations in abundance of
certain estuarine and marine fishes in Louisiana, with
particular reference to life histories. Ecological


152
upstream locations overall, large roving fish densities are
lower (Chapter 2). Conditions upstream overall may be
inhospitable for large roving fish, preventing them from
permanently residing in these locations (Chapters 3 and 4).
In addition, salinity regime in both the far/up- and
mid/upstream locations, was low and variable in the current
study. Based on these findings, one would have expected the
experiments to indicate lower predator encounter rates at
both the far/up- and mid/upstream locations.
However, in actuality, tethered fish suffered lower
rates of predation far/upstream than mid/upstream. Thus,
the abundance of predators was probably equivalent at
mid/up-, mid-, and downstream locations, but less abundant a
far/upstream locations. The predator encounter rates,
therefore, do not appear to simply be functions of salinity
regime or mangrove habitat development alone.
One hypothesis that could explain the greater rates
mid/upstream in comparison to far/upstream, could be that a
significant number of the stenohaline predators primarily
residing in marine habitats also temporarily forage at the
edge of their primary range (e.g. Weinstein 1979). The
mid/upstream locations would be at the edge of this range
for marine predators. An analogous situation may occur for
freshwater predators at the other side of the
marine/freshwater interface.
To explain the lack of predators far/upstream, perhaps
other characteristics of the these locations (besides


130
portion of the epibenthic fauna in western Florida Bay, have
lower densities in the interior and eastern Bay (Holmquist
et al. 1989a).
Diet Breadth and Variability
Three types of feeding strategies were identified among
the 6 species based on diet breadth and degree of
opportunism. Firstly, Floridichthys carpi and Eucinostomus
harengulus appeared to rely strongly on the plasticity of
their diets, feeding on one basic resource consistently, but
also consuming smaller quantities of many other resources.
Although it was not separated from other materials in the
unrecognizable category, these species probably consumed
detritus. A product of breakdown of dead plants, this
material contains a "coating" of bacteria and fungi of
nutritional value (Heald et al. 1974). Primary consumers of
detritus include amphipods, shrimp, crabs, and certain
fishes. Among the 6 species analyzed in this study, a major
portion of the diet of Floridichthys carpi was probably
composed of detritus; in the North River, the diet of this
killifish was 21% detrital material (Odum 1971). In that
study, Eucinostomus harengulus was considered a secondary
consumer of this material, with only 6% of the gut contents
directly composed of detritus (Odum 1971).
The second major strategy was employed by Fundulus
granis, Lutjanus griseus and Strongylura notata. Rather
than switching among a very wide variety of items, they
consumed about 5 items consistently and abundantly. Thus,


153
salinity regime and mangrove habitat development) make them
safer havens for small benthic fishes. In contrast to all
the other locations, to forage far/upstream, a predator
would have to negotiate a series of sinuous channels and
interspersed ponds. Shallow shoals occur at pond/creek
intersections that are sometimes only a few centimeters
deep. This complex system seems likely to prevent access by
casual foragers and predators above certain size limits.
Based on these results, the ecological paradigm may
thus be qualified. Large predators may be prevented from
permanently occupying upstream locations by low and variable
salinity conditions, and the complexity of sinuous channels
may prevent them from foraging in far/upstream locations. A
safe haven for small benthic fish thus occurs in complex
habitats at the marine/freshwater interface. These findings
tend to support the hypothesis suggested by Browder & Moore
(1981): ideal juvenile fish habitat may occur where the
variable salinity conditions overlap areas of such habitat
complexity. They tend to refute, however, the hypothesis
that small fishes are protected from predation in estuaries
by lower salinities which tend to exclude stenohaline marine
predators.


6. IMPLICATIONS AND CONCLUSIONS 154
Implications for Mangrove Fish Ecology 154
Implications for Estuarine Fish Ecology:
the Nursery-ground Hypothesis 157
Management Implications 158
LITERATURE CITED 161
BIOGRAPHICAL SKETCH 172
vi


108
Using the correlation between the original habitat
variables and each component, each component was interpreted
in ecological terms. Correlations greater than 0.60 were
considered in the interpretation (Afifi & Clark 1984) For
the enclosure net data, the first component described
gradients of water depth and percent canopy cover. The
second component associated with the net data described the
salinity regime, in particular, salinity variation. As
salinity variation increased, all other habitat variables
decreased, especially submerged aquatic vegetation and tree
cover. The third component was associated with fringe
width. In combination, these components, describe a
gradient of mangrove habitat development, together with
salinity regime.
For the visual census data, the first component was
most strongly correlated with salinity regime, water depth
and submerged aquatic vegetation. Tree cover was the only
strongly correlated variable associated with the second
component, and tree height, with the third. Together with
salinity regime, these components combined describe a
gradient of total habitat development including both
mangroves and submerged aquatic vegetation.
Relationships Between Habitat Factors and Fish Densities
The first 6 factors were included as independent
variables in the multiple regression analysis. Densities of
fish in the three fish groups were dependent variables
(Table 3-5). No significant relationships were found for


75
stations, including one in Joe Bay and all in Highway Creek,
grouped separately from those located mid- and downstream.
In a second analysis based on densities of each species, all
but one upstream station clustered separately from the mid-
and downstream locations.
For the visual census, cluster analysis results were
also graphed using dendrograms (Figure 2-11). Based on
presence of species, all the upstream stations formed one
cluster. Based on densities of all species, three of the
five upstream stations clustered together. These were the
three that were most upstream.
For each cluster group defined by the cluster analysis,
the most common or dominant sets of species were identified.
Three species that commonly occurred at all stations in the
study area were: goldspotted killifish (Floridichthys
carpi), rainwater killifish (Lucania parva), and redfin
needlefish (Strongylura notata). Species that were very
common in upstream stations included the inland silverside
(Menidia spp.), clown goby (Microgobius gulosus), tidewater
mojarra (Eucinostomus harengulus), striped mojarra (Eugerres
plumieri), and Mayan cichlid (Cichlasoma urophthalmus).
Downstream species commonly included hardhead silverside
(Atherinomorus stipes), gray snapper (Lutjanus griseus),
silver jenny mojarra (Eucinostomus gula), great barracuda
(Sphyraena barracuda), blue-striped grunt (Haemulon
sciurus), and snook (Centropomus undecimalis). Midstream


114
Densities of other abundant species such as Arius felis,
Mugil cephalus, Sphyraena barracuda and Strongylura notata
were not correlated with any of the measured variables.
For the regressions with individual groups of fish and
species, the range of percent of variance explained by the
combination of variables was low overall (10.3 to 38.8%).
However, the regressions were significant for 18% of all the
species, and approximately one-half of the species that were
collected in great abundances (i.e. over 100 individuals).
Discussion
Sites with a combination of lower mean salinity and
high salinity variation had lower levels of all the other
habitat development variables indicating reduced habitat
development at such locations. This finding is consistent
with previous results for benthic community development in
northeastern Florida Bay (Montague et al. 1989).
An example of a well-developed mangrove shoreline is
illustrated in Figure 3-1. In variable salinity conditions,
such mangrove habitats are less likely to occur. Among the
species preferring sites with greater mangrove habitat
development are the snappers, grunts, toadfish, rainwater
killifish, and sailfin mollies. Other species may utilize
mangrove habitats on a less discriminating basis and tend to
occupy all mangrove habitats. Such species include snook,
barracuda and sheepshead. This assemblage of fishes is
enhanced when mangrove shorelines occur and especially where


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Clay/L. Montague^/ Chairman
Associate Professor of
Environmental Engineering Sciences
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
i-S- C rv^\'
Carole C. Mclvor, Cochairman
Assistant Professor of
Forest Resources and Conservation
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Professor of Zoology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Scientist
Environmental Engineering Sciences


26
for 30 to 45 minutes by two persons. After repeating the
process at the other two sites, and allowing the rotenone to
dissipate (approximately 3-4 hours), a snorkeler retrieved
sunken fish and a wader collected along the berm edge from
within each enclosure. Since rotenone effectiveness is
reduced with decreasing temperature (Neilson & Johnson
1983), in colder months, floating fish were also collected
after leaving the nets up overnight.
Fish and invertebrates collected were initially placed
on ice and then frozen. Later they were identified to
species and measured to total and standard (or carapace)
length.
Efficiency tests. Fish recovery efficiency was tested
at least once at each location using a mark-recapture method
and normal net procedures. To collect test fish, several
minnow traps were placed inside the area to be enclosed on
the day before net deployment. After the net was in place,
the minnow traps were removed, cleared and the fish placed
into buckets. Fish were measured, marked by fin-clipping,
and returned to the enclosed net area in minimal time. At
least 30 fish were used per net in the test.
Visual Census
Based on the pilot study, direct recording of census
data on underwater paper was identified as the better visual
method tested. Quality of video tape was inconsistent and
often low under the variable turbidity conditions
encountered.


18
Contrary to what might be expected based on such
physiological factors, however, abundance peaks for many
estuaries occur in conjunction with the initiation of the
period of maximum freshwater inflow, when salinity levels
drop dramatically. Estuarine transient juveniles may
constitute most of the individuals during these peak periods
(Yanez-Arancibia et al. 1980, Bell et al. 1984, Pinto 1987).
In some species, juvenile fishes may be capable of
exploiting salinities at lower levels than adults (Gunter
1967, Moser & Gerry 1989). However, in environments with
more stable salinities, estuarine transient juveniles can
also be abundant (Little et al. 1988, Robertson & Duke
1990b). Thus, the role of seasonal changes in salinity on
fish communities reguires further exploration.
Based on investigations conducted near and within the
northeastern Florida Bay study area, at the initiation of
the rainy season (June), changes in salinity were expected
to occur, expanding the zone of low salinity further
downstream (Ginsburg 1956, Tabb et al. 1962, Lindall et al.
1973, Thayer et al. 1987). This zone of lower salinity was
expected to persist after the end of the rainy season, as
freshwater from the eastern Everglades gradually drained
into Florida Bay.
Toward the goal of understanding the influence of
freshwater inflow on fishes, the objective of this portion
of the study was to identify spatial and temporal patterns
in fish assemblages across the salinity gradient and thereby


4
3
2
1
O
4
3
2
1
O
4
3
2
1
O
Density by General Location
East
Top 3 Water Column Forage Fish Species
Enclosure Nets
West
Mid
i
Down
Anchoa Atherinomorus Menidia
mitchelli stipes spp.
Down
j
Anchoa Atherinomorus Menidia
mitchelli stipes spp.
o>
NJ


156
occur where both mangrove development and SAV are greater.
Reduced SAV may thus account for the reduced abundance of
large roving fish upstream.
Selection Among Mangrove Habitats
Mangrove shorelines vary in functional value depending
on the degree of development and habitat needs of the
fishes. In freshwater streams, a gradient model has been
proposed (Schlosser 1987) that appears to apply equally well
in northeastern Florida Bay. This model identifies a
gradient from upstream areas (environmentally unstable,
shallower, lesser habitat development) to downstream areas
(stable, deeper, greater development). Greater habitat
development is linked with the occurrence of more species
and larger piscivorous individuals (Schlosser 1987) For
smaller fishes, upstream areas provide refugia from larger
piscivores that are more abundant downstream.
In northeastern Florida Bay, the mangrove habitats
ranged from less developed upstream (small trees, narrow
fringe, shallow water, high environmental variability), to
more developed downstream (tall trees, deeper water,
environmentally more stable). Among the large roving
species, greater abundances of gray snappers and grunts were
associated with more developed mangrove habitats in the
current study. In addition, numbers of species of large
roving fish were more abundant downstream than upstream.
Thus, in terms of large roving fish, the model applies well
to northeastern Florida Bay.


Table 3-1. Means, standard deviations, and correlations among the physical/
environmental variables associated with the 18 enclosure net stations.
Water
depth
cm
Fringe
width
m
Tree
height
cm
Prop
roots
n
Submerged
Vegetation
cm3
Salinity
mean
ppt
Salinity
variation
ppt
Tree cover
%
Mean
48.53
8.88
112.49
10.82
22.30
34.10
8.87
67.02
Standard deviation
10.33
3.43
55.38
5.51
6.58
7.79
3.26
14.33
Correlations (r)
between
variables
Water depth
1.000
Fringe width
0.167
1.000
Tree height
0.635
0.416
1.000
Prop roots
0.285
0.633
0.606
1.000
Submerged vegetation
0.319
0.163
0.047
-0.142
1.000
Salinity mean
0.492
0.689
0.412
0.518
0.568
1.000
Salinity variation
-0.417
-0.425
-0.414
-0.365
-0.599
-0.829
1.000
Tree cover
0.732
0.393
0.735
0.572
-0.012
0.550
-0.441
1.000


92
As indicated by comparisons from up- to downstream,
salinity regime did not affect the overall density of
benthic and water column forage fish. Furthermore, in the
current study, no differences were found between the eastern
and western systems. Similarly, Thayer et al. (1987) found
no effect due to gradient for mangrove fishes collected with
enclosure nets. The sites sampled by Thayer et al. (1987),
were located in central and western Florida Bay from
downstream near the Keys, to upstream in Whitewater Bay and
Coot Bay and their collections were dominated by small
forage fish. Thus, salinity regime may have little
influence on densities of forage fish species throughout
Florida Bay.
The individual species of benthic forage fish were
likewise distributed widely and not systematically along the
gradient. Although small benthic fishes have certain other
life history characteristics that may explain the widespread
distributions observed (Sogard et al. 1987) most species
are notably euryhaline (Robins et al. 1986, Nordlie & Walsh
1989). Thus, they are good colonizers of all types of
habitats found in Florida Bay.
In general, high variance was evident in the density
estimates for water column forage fish. This was due, in
large part, to their schooling behavior. With the nets, a
school was either collected (and thus the abundances were
great) or not collected (and thus the numbers were zero).
Individual species of water column forage fish, however, do


13
Habitat Features
Because salinity is not the only feature of the habitat
that varies along the complex environmental gradient within
the study area, it was also necessary to consider other
features of the fixed and moveable habitat (Browder & Moore
1981) as potentially influencing fish community structure.
The second study objective was to determine important
habitat features that influence the abundance and species
composition of mangrove fish communities and compare these
features across the salinity gradient. Fixed habitat
structural features such as mangrove tree height and prop
root density, environmental features such as water
temperature, and fish diet and predation, were expected to
influence the differences among fish assemblages across
environmental gradients.
Study Area
The 250 km2 study area, located in extreme northeastern
t
Florida Bay, consists of a series of shallow bays and ponds
(less than 1.0 m in depth) bordered by mangroves. The
upstream portion of the area is subject to freshwater inflow
from seven mangrove-lined tributaries originating in the
Taylor Slough/C-111 drainage basin.
In this region of Florida Bay, rapid ecological changes
can take place when salinity variations occur suddenly, as
at the start of the rainy season (Montague et al. 1989).
Because tidal influences are almost negligible in the
northeastern Florida Bay area, salinity changes are caused


Table 2-3. Number of fish collected using enclosure nets and observed during
visual censuses. Explanation of group/residency given at end of table.
Group/
Residency
Family Species
General Locations
Total
Nets Visual
Up-west
Nets Visual
Mid-west
Nets Visual
Down-west
Nets Visual
Up-east
Nets Visual
Mid-east
Nets Visual
Down-east
Nets Visual
2
1
1 2
4
4
1
2 1
2 2
6
6
1
1
3
1
1
2
7
11
3
151
165
3 600
3 600
335 41
40
2 100
454 2
17,770
4
18,605 143
5
5
4
1 551
2
2 3
9 554
38
59
134
23
275
529
1
1
1
1
17
17
1
1
Carcharhinidae (requium shark)
LR/r Carcharhinus laucas
Orectolobidae (nurseshark)
LR/r Ginglymostoma cirratum
Dasyatidae (stingray)
LR/r Dasyatis sabina
Elopidae (tarpon)
LR/o Megalops atlanticus
Anguillidae (freshwater eel)
LR/o Anguilla rostrata
Clupeidae (herring)
WC/o Opisthonema oglinum
WC/o Clupaid (species unk)
WC/o Harengula jaguana
Engraulidae (anchovy)
WC/r Anchoa mitchelll
WC/r Anchoa cayorum
Ariidae (sea catfish)
LR/r Arius felis
Batrachoididae (toadfish)
BF/r Opsanus beta
BF/r Porichthys plectrodon
Gobiesocidae (clingfish)
BF/r Gobiesox strumosus
Bythitidae (viviparous brotula)
BF/r Ogilbia cayorum
BF/r Gunterichthys long ¡penis
continued


Table 2-17. Species richness index for fish sampled with enclosure nets and
visual census. Species Richness Index =
(Number of species 1) / log (Total individuals) (Odum 1983).
General
Location
Benthic Forage
Fish
Water Column Forage
Fish
Large
Roving
Fish
All Fish
Total
Number Index
Total
Number Index
Total
Number
Index
Total
Number
Index
Indiv.
of
Indiv.
of
Indiv.
of
Indiv.
of
Species
Species
Species
Species
Up-west
15,951
24
5.5
5,408
6
1.3
450
11
3.8
21,809
64
14.5
Mid-west
16,049
14
3.1
51,052
4
0.6
5,897
15
3.7
72,998
49
9.9
Down-west
13,442
21
4.8
37,083
6
1.1
7,890
20
4.9
58,415
65
13.4
Up-east
15,716
21
4.8
4,805
6
1.4
327
10
3.6
20,848
58
13.2
Mid-east
7,434
19
4.6
25,222
7
1.4
1,739
13
3.7
34,395
54
11.7
Down-east
16,342
22
5.0
68,966
6
1.0
11,816
22
5.2
97,124
66
13.0
All
84,934
33
6.5
192,536
9
1.5
28,123
35
7.6
305,589
77
13.9
to


48
density of benthic forage fish and large roving fish. In
both cases, lower abundances occurred at higher water
temperatures.
Spatial Patterns in Density by Fish Group
Spatial patterns in fish density varied among the fish
groups (Figure 2-6). From these graphs, one can see that
only the larger roving fish group seems to vary consistently
along the salinity gradient, with much lower densities at
the upstream locations.
Benthic forage fish analysis of variance. Results of
the repeated measures ANOVA's by fish group differed among
the fish groups (Tables 2-6 and 2-7). Neither gradient
position nor system were important determinants of variation
in densities among the stations for the benthic forage fish
group (Table 2-6). Although densities tended to vary
significantly from one general location to another, these
variations were not systematic along the salinity gradient,
as indicated by the significant interaction between gradient
and system.
The mid/west general location had significantly greater
densities than the other midstream location (Table 2-7,
Figure 2-6). Other locations were intermediate and not
significantly different from these two.
Water column forage fish analysis of variance. Again,
although densities tended to vary significantly from one
general location to another, these variations were not
systematic along the salinity gradient, as indicated by the


Table 2-7. Repeated measures analysis of variance with densities of fish as dependent
variables and general locations as independent variables. Benthic and water column
forage fish were collected using enclosure nets. Large roving fish were sampled using
visual census techniques. Data were transformed to logarithms prior to performing
calculations.
Source
Benthic forage fish
Water Column Forage Fish
Large
Roving Fish
df*
F
P
df*
F
P
df*
F p
Between Stations:
Among general locations
5/12
3.77
0.0277
5/12
5.1
0.0097
5/12
2.94 0.0586
Within stations
Among months
10/120
9.53
0.0001
10/120
5.16
0.0001
4/48
2.17 0.0863
Month X General locations
50/120
1.18
0.2360
50/120
4.11
0.0001
20/48
1.16 0.3249
Multiple comparisons among
means for all months:
Location
Sign.
greater
than
Location
Location
Sign.
greater
than
Location
Location Sign. Location
greater
than
2
>
5
(others
inter
mediate)
5 & 3
>
4 & 2
(others
inter
mediate)
No differences
Source degrees of freedom / error degrees of freedom
** General locations:
1 = Joe Bay, upstream/west
2 = Trout Cove, midstream/west
3 = Buttonwood Sound, downstream/west
4 = Highway Creek, upstream/east
5 = Little Blackwater Sound, midstream/east
6 = Blackwater Sound, downstream/east
to


142
Far/Up
Far/Up
Mid/Up
Mid/Up
O Snorkel Sites
* Enclosure Net Sites
Tethering stations
Y Key
Ranger Station
*
Figure 5-2. Locations of study sites in northeastern
Florida Bay.


141
Tethering Experiment
Small fish were tethered at far/up-, mid/up-, mid- and
downstream locations in two systems (west and east) in
northeastern Florida Bay (Figure 5-2). Tethered species
included those that are consumed by dominant members of the
predator guild in the study area, as indicated by the
results of the food habits portion of this overall study
(Chapter 4). Small fish make up over 25% of the diets of
Strongylura notata (redfin needlefish), Lutjanus griseus
(gray snapper) and Sphyraena barracuda (great barracuda).
Trials were conducted on two dates at each of the eight
locations during mid-summer 1990. For each trial, salinity,
horizontal secchi distance, and water depth were recorded.
Fish to be used in each trial were collected by setting out
several minnow traps near each site on the day before a
test. Small fish (4 to 10 cm total length) from five
species were used in the 16 trials: killifish
(Floridichthys carpi, Cyprinodon variegatus, Fundulus
grandis, and Fundulus confluentus) and crested gobies
(Lophogobius cyprinoides). In each trial, ten to sixteen
fish were tethered 10 m apart and about 2.0 m from the
mangrove edge (Figure 5-1) Sites with water depths of
about 50 cm were selected for each fish. The total time
each prey was tethered ranged from 3 to 3.5 hours. For
approximately 1.5 hours, while observers were within 10 to
100 m of the tethered stakes, they recorded when possible
the type, size, and number of predators that approached or