Citation
Effects of introduced peacock cichlids Cichla ocellaris on native largemouth bass Micropterus salmoides in southeast Florida

Material Information

Title:
Effects of introduced peacock cichlids Cichla ocellaris on native largemouth bass Micropterus salmoides in southeast Florida
Creator:
Hill, Jeffrey E., 1966-
Publication Date:
Language:
English
Physical Description:
viii, 155 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Canals ( jstor )
Crayfish ( jstor )
Ecology ( jstor )
Fish ( jstor )
Fisheries ( jstor )
Freshwater bass ( jstor )
Introduced species ( jstor )
Peafowl ( jstor )
Predators ( jstor )
Species ( jstor )
Dissertations, Academic -- Fisheries and Aquatic Sciences -- UF ( lcsh )
Fisheries and Aquatic Sciences thesis, Ph.D ( lcsh )
Miami metropolitan area ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 2003.
Bibliography:
Includes bibliographical references.
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Jeffrey E. Hill.

Record Information

Source Institution:
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:
030589380 ( ALEPH )
83217140 ( OCLC )

Downloads

This item has the following downloads:


Full Text











EFFECTS OF INTRODUCED PEACOCK CICHLIDS CICHLA OCELLARIS
ON NATIVE LARGEMOUTH BASS MICROPTERUS SALMOIDES
IN SOUTHEAST FLORIDA










By

JEFFREY E. HILL


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


2003
















ACKNOWLEDGMENTS


Many people provided tremendous assistance in the completion of this project. All deserve my acknowledgment. I apologize beforehand to anyone inadvertently omitted.

The primary acknowledgment goes to my wife, Susan, for her unfailing support throughout my entire graduate career. I thank her for sacrificing in order for me to fulfill our shared goal. I especially thank her for putting up with my obsession for fish.

I am grateful for the constant support of my family-my father, mother, and sister (Baker, Jacqueline, and Kim). They shared the dream of my doctorate-I am thankful to have fulfilled our collective aspiration.

I greatly appreciate the guidance and support of my doctoral committee-Drs.

Charles E. Cichra (Chair), Carter R. Gilbert, William J. Lindberg, Leo G. Nico, and Craig W. Osenberg. It was a great pleasure to work for Dr. Cichra-the experience I received in extension, research, and teaching, along with strong mentorship and his friendship, were instrumental in my professional development. Carter Gilbert has been a tremendous influence and is one of my real "fish heroes". I especially thank Carter for staying involved in my graduate education after his retirement-that meant a lot to me. Leo Nico's remarkable field experience with nonindigenous fishes and south Florida systems was invaluable. The many discussions (and some disagreements) over nonindigenous fishes with Leo helped shape my scientific philosophy. I appreciate the support and


ii









encouragement that Bill Lindberg always gave. Our discussions inside and outside of the classroom about ecology and the philosophy of science (e.g., tautology!) helped open my eyes to the wider world of science. I thank Craig Osenberg for agreeing to help out on the committee of someone studying a crazy system like the south Florida canals. I especially appreciate his assistance with preparing for my qualifying exam and his extraordinary input during the revision process. I have learned a great deal from each committee member and thank them for being a part of my graduate education.

A special acknowledgment is due to Jeff Sowards for all of the hard work in the field. Thirty-six hour "days" can be rough, but he never complained. Jeff was instrumental in the field work, in the lab, and as an "ideas man" to keep the project doable. I could not have done it without his help. Also, Sharon Fitz-Coy was a tremendous help in, among other things, collecting prey for experiments and identifying insects in fish stomach contents. She was always willing to help.

Special thanks are due to Paul Shafland (Florida Fish and Wildlife Conservation Commission; FWC) for a great deal of assistance, facilitation, literature, and advice during this project. Paul's largest contributions, however, came during many discussions of scientific philosophy regarding nonindigenous fish.

Craig Watson and Roy Yanong (Tropical Aquaculture Laboratory) provided

exceptional support throughout the process. They made the tribulations more bearable.

Debbie Pouder facilitated all work done at the Sam Mitchell Aquaculture

Demonstration Farm (SMADF) in Blountstown, Florida. I appreciate her help and that of the staff, especially Randall Kent.


iii









Ken Portier of the Department of Statistics at the University of Florida helped with the design and analysis of experiments in Chapter 5.

I appreciate the hospitality of Don and Judy Bemecker who provided a great place to stay and many delicious meals during field work.

Many others are due thanks for their contributions. At the University of Florida, Department of Fisheries and Aquatic Sciences, Mike Allen, Mary Cichra, Doug Colle, Ruth Francis-Floyd, Tom Glancy, Scott Graves, Robert Leonard, Allen Riggs, Beth Sargent, Troy Thompson, Larry Tolbert, and Dan VanGenechten helped with various aspects of the project. Kelly Jacoby drew the maps. Susan Morgan greatly facilitated the graduation process. At the Florida Caribbean Science Center, U.S. Geological Survey, Gary Hill, Howard Jelks, and Bill Stranghoener assisted with facilities and equipment. Frank Morello, Jerry Krummrich, and Fred Cross (FWC) issued scientific collecting permits; Ted Hoehn (FWC) kindly provided access and instruction to the FWC fish database; Eric Nagid (FWC) helped answer questions about use and content of the database.

I gratefully acknowledge the financial and facilities assistance of the Department of Fisheries and Aquatic Sciences, the Sam Mitchell Aquaculture Demonstration Farm, and the Tropical Aquaculture Laboratory of the University of Florida. Financial assistance was further provided by a Dean's Fellowship for Graduate Research from the Institute of Food and Agricultural Sciences, University of Florida, the Florida Department of Agriculture and Consumer Services, by Ryan Kelley Memorial Scholarships from the International Women's Fishing Association, and the Roger Rottmann Memorial Scholarship from the Florida Chapter, American Fisheries Society.


iv















TABLE OF CONTENTS
page

ACKNOW LEDGM ENTS...................................................................................................ii

A B S T R A C T ......................................................................................................................v ii

CHAPTER

1 INTRODUCTION................................................................................................1

2 STUDY SYSTEM .................................................................................................11

S tu dy Sp ecies.........................................................................................................1 1
Canal System.................................................................................................... 14
Study Sites......................................................................................................... 18

3 A QUALITATIVE RISK ASSESSMENT OF PEACOCK CICHLIDS
CICHLA OCELLARIS AND SOME IMPLICATIONS FOR THE
M ETHODOLOGY........................................................................................... 24

Introduction....................................................................................................... 24
Background Information.................................................................................... 27
Qualitative Risk Assessment............................................................................. 36
Discussion......................................................................................................... 43

4 ONTOGENETIC DIETARY SHIFTS AND DIETARY OVERLAP OF A
NONINDIGENOUS FISH AND AN ECOMORPHOLOGICALLY
SIM ILAR NATIVE FISH ................................................................................. 49

Introduction....................................................................................................... 49
M eth o d s..................................................................................................................5 1
R e su lts....................................................................................................................5 6
Discussion......................................................................................................... 68

5 EXPERIMENTAL PREY HANDLING AND SELECTIVITY OF TWO
MORPHOLOGICALLY SIMILAR PREDATORY FISHESNATIVE LARGEMOUTH BASS AND INTRODUCED PEACOCK
CICHLIDS......................................................................................................... 79


V









Introduction ............................................................................................................79
M ethods..................................................................................................................82
Results....................................................................................................................90
Discussion......................................................................................................... 99

6 THE INTRODUCTION OF PEACOCK CICHLIDS INTO SOUTHEAST
FLORIDA: POPULATION-LEVEL CONSEQUENCES FOR
LA RG EM O UTH BA SS......................................................................................109

Introduction..........................................................................................................109
M ethods................................................................................................................112
Results..................................................................................................................114
D iscussion............................................................................................................119

7 SU M M A RY .........................................................................................................127

REFEREN CES................................................................................................................133

BIO G RA PH ICA L SK ETCH ...........................................................................................154


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

EFFECTS OF INTRODUCED PEACOCK CICHLIDS CICHLA OCELLARIS
ON NATIVE LARGEMOUTH BASS MICROPTERUS SALMOIDES IN SOUTHEAST FLORIDA

By

Jeffrey E. Hill

December 2003


Chair: Charles E. Cichra
Major Department: Fisheries and Aquatic Sciences

The predatory peacock cichlid Cichla ocellaris, native to South America, was introduced into southeast Florida to enhance recreational fishing and to provide biological control over nonindigenous fishes. This action was controversial and there is concern over the effects of the introduction on native species, especially the ecomorphologically similar largemouth bass Micropterus salmoides. I conducted a formal qualitative risk assessment to evaluate the risks associated with the peacock cichlid introduction and to test predictions of the algorithms. The estimate of risk for peacock cichlids was Low. The evaluation elucidated shortcomings in the methodology (e.g., emphasis on a single output) and I made recommendations for its improvement. Part of the concern over the peacock cichlid introduction is their similarity to largemouth bass. I tested and found support for the hypothesis that morphological similarity equals similarity in prey handling ability and prey selection. I documented substantial overlap in


vii









trophic morphology and dietary ontogeny between juveniles and found similar prey handling ability and prey selection by adults in tank experiments. These results suggest that largemouth bass and peacock cichlids have similar effects on the prey base. However, largemouth bass had exclusive use of highly defensive prey (i.e., crayfish). Largemouth bass populations might be expected to have declined, given evidence of direct negative effects of peacock cichlids. I investigated this question by comparing largemouth bass populations in southeast Florida canals with populations in lakes and streams (statewide and in south Florida). Additionally, I compared canals with peacock cichlids to canals without peacock cichlids. The data did not support a hypothesis of declines in largemouth bass abundance due to peacock cichlids. These results question common assumptions concerning the effects of nonindigenous species-1) that direct effects of predation and competition are the predominant effects and 2) the narrow view that invaded communities are organismic (i.e., Clementsian)-and challenge the uncritical acceptance of the premise that nonindigenous fishes will invariably cause ecological harm. My results suggest that other factors (e.g., indirect effects) may be important and that invasion biology should develop a pluralistic theory that incorporates Gleasonian concepts (i.e., individualistic communities).


viii















CHAPTER 1
INTRODUCTION


Introduced species are considered threats to natural ecosystems and biodiversity worldwide (Lodge 1993). For example, Miller et al. (1989) reported that nonindigenous species were a factor in the extinction of 27 North American fishes (68% of extinctions). Similarly, Lassuy (1995) stated that nonindigenous species were factors in listing 48 imperiled freshwater fishes under the Endangered Species Act in the USA (70% of listed fishes). Although ecological damage due to nonindigenous species is difficult to estimate, economic costs in the billions of dollars are cited for aquatic and terrestrial systems in the USA (OTA 1993). Indeed, the overwhelming consensus is that species introductions may result in deleterious effects (e.g., lowered abundance of native species) within the receiving system (Magnuson 1976; Courtenay 1993, 1995; Minns and Cooley 2000).

Nonindigenous fishes have a long history in the USA and have entered this country through several intentional and unintentional pathways (Fuller et al. 1999). Many fishes have high economic and social value and have been purposefully stocked to provide fisheries benefits. For example, centrarchids (e.g., Lepomis and Micropterus), percids (e.g., Sander [formerly Stizostedioni), salmonids (e.g., Oncorhynchus and Salmo), and other fishes have been transferred to many regions to enhance recreational fisheries (Fuller et al. 1999; Heidinger 1999). Purposeful introductions for recreational fishing are perhaps the most studied cases of interaction between native and


I









nonindigenous fishes. Nevertheless, thorough empirical evaluations are few (but see Huckins et al. 2000). Such research is costly and may be controversial (e.g., Churchill et al. 2002). Introduced salmonids have been intensely scrutinized, yet there is wide variation in their reported effects on native salmonids (Fausch 1988). In a review of the effects of salmonid introductions, Fausch (1988) found that some authors documented little or no effect of introduced salmonids, yet other authors concluded that such fishes constitute a serious threat to native species (see also Fuller et al. 1999; Crawford 2001). Similarly, data for non-salmonid introductions are often mixed and inconclusive (Taylor et al. 1984; Shafland 1996a; Courtenay 1997; Fuller et al. 1999). Regardless of the potential or realized benefits, the possible dangers have led to many calls for a cessation of purposeful fish introductions (Magnuson 1976; Courtenay and Robins 1973, 1989; Courtenay 1995; Minns and Cooley 2000). These papers point out the numerous theoretical negative impacts of fish introductions and that some introduced fishes have proven to be notable and costly pests (e.g., the unintentionally introduced sea lamprey Petromyzon marinus in the Great Lakes; Mills et al. 1994).

Given increasing transfers of fishes outside their natural ranges (Nico and Fuller 1999), there has been significant scientific and public interest in predicting the effects of species invasions on native aquatic faunas (Lodge 1993; OTA 1993; Moyle and Light 1996a; Kolar and Lodge 2002). Indeed, inadequate predictive ability in this field has frustrated scientists and resource managers. This situation is partly due to the surprisingly few empirical evaluations of the effects of nonindigenous fishes (Taylor et al. 1984; Moyle et al. 1986; Fuller et al. 1999). In fact, much of the literature on harmful effects is anecdotal. Furthermore, a lack of a truly pluralistic ecological theory of bio-









invasion may be a factor hampering the advance of this field (Schoener 1987; Ross 1991; see also Lodge 1993; Williamson 1996; Shrader-Frechette 2001; Sakai et al. 2001).

General ecological theory provides a framework for a variety of predictions of the effects of species introductions. For example, there is a rich literature concerning the effects of species introductions and removals in temperate lakes on lake productivity and food web structure (e.g., Carpenter et al. 1985; Carpenter et al. 1987; Mittelbach et al. 1995). Some researchers have used ecological theory to predict and subsequently document negative effects of species introductions (e.g., Huckins et al. 2000); others have predicted and then demonstrated effects that were positive for aquatic resources (e.g., Carpenter et al. 1985; Carpenter et al. 1987). Therefore, ecological theory is not biased toward an invariable prediction of negative effects.

Despite this rich body of ecological theory, many scientists and resource managers interested in the ecological effects of nonindigenous species assume that invading species have only net negative effects on native species (see reviews in Stauffer 1984; Taylor et al. 1984; Moyle et al. 1986; Li and Moyle 1999; but see Williamson 1996). The current emphasis on negative effects of introduced species are founded on two principles: 1) the organismic concept of equilibrial, stable, and competitivelystructured communities (Clements 1936) and 2) that direct effects of invasive species (via competition and predation) are far more important than direct or indirect positive effects that might negate or overwhelm the direct negative interactions. The expectation that competition will play a key role was influenced by Gause's articulation of the competitive exclusion principle (Gause 1934; Hardin 1960), the Hutchinsonian concept of niche (Hutchinson 1957), Elton's (1958) work on species invasions, and MacArthur






4


and colleague's papers on competition and island biogeography (e.g., MacArthur and Levins 1967; MacArthur and Wilson 1967). Such theory (e.g., island biogeography) predicts that any species addition must lead to a species extinction (MacArthur and Wilson 1967). Under this view, the observed increase in invasions (e.g., Nico and Fuller 1999) is certain to lead to increased rates of extinction (or extirpation) of native species (and previously established nonindigenous species)-a prediction leading to considerable concern for conservation biologists and resource managers over a loss of biodiversity. Furthermore, the native species most at risk of extinction should be those species that are most similar to invading species with respect to resource use (Hardin 1960; MacArthur and Levins 1967; Diamond 1975; but see Resetarits 1995). Because similarity in resource use is often associated with similarity in morphology (Wainwright 1988, 1996; Wainwright and Richard 1995), species of similar morphology are expected to be the most sensitive to invading species. Moreover, proponents of this view of species invasions aggressively challenge any notion of "empty niches" on semantic and theoretical grounds (e.g., Herbold and Moyle 1986; Courtenay 1995), practically precluding the possibility that species introductions can take place without any deleterious effects for the receiving system.

Nonindigenous fishes may adversely affect native species directly or indirectly via competition, predation, habitat alteration, reproductive inhibition, and shared pathogens (Taylor et al. 1984; Moyle et al. 1986). Indeed, many conclusions of impacts of nonindigenous fishes assume that these mechanisms of negative interaction will directly translate into reductions in the density of native species. If native fish populations are low or are perceived to have declined coincidentally with the introduction






5


of a nonindigenous fish, then evidence for the existence of a putative mechanism is often assumed to be proof that the mechanism is a contributing factor. This apparent causeand-effect scenario may be correct for some cases of species introductions. Nevertheless, the existence of a negative interaction between the two species may not be important to the native species at the population level if, for example, the total strength of the interaction (per capita effect x population size) is low, if other factors (e.g., abiotic events) are strong, or if density-dependence at later life stages leads to compensatory responses (e.g., Vonesh and De la Cruz 2002). As a result, data are required that directly link the interactions between introduced and native species and the observed effects on population density.

Despite these views that invasions should lead to reductions in abundance or

extinctions of native species, there is growing evidence that new species can be added to native communities with little noticeable effect (Moyle and Light 1996a; Williamson 1996; Gido and Brown 1999). Indeed, perhaps as many as 80-90% of all established nonindigenous species fall into this category (Williamson 1996). For example, Moyle and Light (I 996a) noted several examples of invading fishes integrating into freshwater and marine communities without obvious changes in native species.

A partial explanation for the failure of many scientists and resource managers to expect observed cases of integration of invading species into communities is a conscious or tacit adherence to the Clementsian view of organismic communities (Clements 1936). On the other hand, if communities are considered loosely-structured assemblages of species that result from probabilistic processes, that are transient in time and space, and are dependent on the characteristics of individual species (individualistic concept;






6


Gleason 1926), then integration of invaders into local assemblages is a possibility. Clearly, there is a great deal of evidence that plant and animal associations have changed over time (e.g., Graham et al. 1996), implying that communities may change in species composition and are not fixed in structure. Exchanges of species during numerous faunal interchanges throughout the Neogene period (i.e., last 25 million years) have often led to enrichment of faunas rather than species replacements, suggesting that many communities are not saturated with species (Vermeij 1991). Moreover, Rosenzweig (2001) argued on theoretical grounds that species introductions often permanently increase local biodiversity. Likewise, Gido and Brown (1999) reviewed fish community data for 125 drainages in North America and concluded that fish introductions generally increased species richness and that this pattern suggests that North American fish communities are not saturated with species. McIntosh (1995) reviewed the controversy associated with the competing paradigms of Clements and Gleason and noted that, although debate is ongoing, many animal ecologists had joined with plant ecologists in viewing communities as transient, individualistic assemblages of species.

Although it is a mistake to assume that biotic factors are not important in

individualistic communities, strong community structuring by biotic factors is a defining characteristic of the organismic concept of community (McIntosh 1995). The emphasis of the organismic concept is apparent in much of the literature on nonindigenous fishes. For example, even though abiotic factors are frequently considered when evaluating risks of establishment of nonindigenous fishes (Baltz and Moyle 1993; Moyle and Light 1996a, b; RAM Committee 1998), most studies have considered biotic interactions to be preeminent once a nonindigenous species becomes established (Taylor et al. 1984;






7


Courtenay 1993; but see Moyle and Light 1996b). However, in his review of fish introductions and stream fish assemblage structure, Ross (1991) pointed out that the establishment of a nonindigenous species with no apparent effect on the native assemblage suggests strong abiotic effects within the community.

Another important reason for the general prediction of harm to native species

following a species introduction is the prominent emphasis on the direct negative effects of competition and predation. Both mechanisms can have strong negative effects on species abundance (Garman and Neilson 1982; McComas and Drenner 1982; Crowder and Binkowski 1983; Kaufman 1992; Mills et al. 1994). However, indirect effects can be both facilitative and powerful. For example, Lawlor (1979) reported that 30-40% of interactions in eight bird communities were positive. Miller (1994) estimated the magnitude of direct and indirect interactions of five weedy plant species. He found that although direct negative effects were large, indirect effects were generally positive, relatively strong, and reduced the influence of competition. Additionally, Stone and Roberts (1991) argued that a high proportion (i.e., 20-40%) of species interactions must be advantageous within the context of a community and that "hypercompetitive communities" (i.e., when every species suffers from the presence of every other species) must be rare. In some cases, positive interactions between putative competitors can overwhelm direct negative effects and result in increases in population abundance (i.e., facilitation) (e.g., Miller 1994).

The exotic peacock cichlid Cichla ocellaris (Cichlidae) was intentionally

introduced into southeast Florida canals in 1984-1987 to provide biological control of other exotic fishes and to create a novel sport fishery (Shafland 1995). This relatively






8


large, predatory species is ecomorphologically similar to the native largemouth bass Micropterus salmoides (Centrarchidae) (Norton and Brainerd 1993) that already occurred in these habitats. In their presentation of empirically-derived "rules" of biological invasions, Moyle and Light (1996a) concluded that the piscivore trophic group is one of the most likely to become successful invaders and to subsequently alter the fish assemblages of receiving systems. The ecomorphological similarity between these two species is also of concern given the potential for resource competition. Indeed, there has been reported substantial overlap in size-specific gape width (Hill 1998), diet (Shafland 1995; 1999b), and habitat use (Lilyestrom and Churchill 1996) between largemouth bass and peacock cichlids. Additionally, there has been speculation and concern regarding possible deleterious effects of peacock cichlids on native fishes in Florida (Courtenay and Robins 1989; OTA 1993; Courtenay 1994; Cox 1999; see also Lachner et al. 1970). Indeed, Cox (1999) reviewed species introductions into North America and Hawaii and listed the peacock cichlid as an "invasive exotic" species in Florida, a designation requiring actual or probable harm resulting from its introduction.

To evaluate the actual effect of peacock cichlids on native largemouth bass, I first considered the potential direct and indirect effects of the introduction on largemouth bass. Using these theoretical effects and additional biological information, I conducted a qualitative risk assessment (RAM Committee 1998) to provide a framework to formalize the predictions of peacock cichlid effects and to evaluate this methodology using peacock cichlids as a test case (Chapter 3).

Given the natural history of this system, predation by peacock cichlids on small largemouth bass and competition for food between similar size classes of peacock






9


cichlids and largemouth bass were the most likely interactions; in the absence of mitigating factors, both would lead to a decline in largemouth bass population density in response to an introduction of peacock cichlids. Other potential mechanisms commonly cited in the literature such as spawning site competition, habitat alteration, hybridization, and introduction of disease (Taylor et al. 1984; Li and Moyle 1999) were considered but discarded as unlikely. Although both species are substrate spawners, there is relatively little temporal overlap in spawning activity (Shafland 1999b). Unlike common carp Cyprinus carpio, grass carp Ctenopharyngodon idella, or some tilapia species (Oreochromis spp., Tilapia spp.) (Taylor et al. 1984; Fuller et al. 1999), peacock cichlids are not known to alter habitat by destroying aquatic macrophytes, by disturbing sediments, or by any other mechanism. The possibility of hybridization is extremely remote because peacock cichlids are phylogenetically distinct from largemouth bass (or any other native Florida fish) at the familial level. Additionally, samples of peacock cichlids were screened for known pathogens by personnel of the US Fish and Wildlife Service prior to release (Shafland 1995). On the other hand, a diet study conducted in Florida confirmed the use of largemouth bass as prey by peacock cichlids in Florida (Shafland 1999b). However, the potential for food competition had received little investigation (but see Shafland 1995, 1999b; Hill 1998). In the present study, I assessed food competition indirectly, largely through overlap in food resource use. Some work on trophic overlap had been conducted with subadult and adult largemouth bass and peacock cichlids; Hill (1998) had previously investigated their similarity in gape size and Shafland (1999b) had compared diets of both species in a southeast Florida canal. I therefore determined overlap in trophic morphology (i.e., morphological ability to handle prey) and






10


dietary overlap from stomach contents using small juveniles (i.e., up to about 160 mm total length). Additionally, I investigated patterns of prey selectivity of subadult and adult largemouth bass and peacock cichlids in two experimental settings. Although overlap in resource use is not a direct proxy for competition (Colwell and Futuyma 1971; Sale 1979; Schoener 1982), shared use of a resource is generally considered to be a prerequisite for exploitation competition (Moermond 1979; Sale 1979; but see Rosenzweig 1981 for situations where competition is an important interaction but overlap in resource use is essentially zero; Chapter 4). Additionally, resource overlap is often used to justify conclusions of interspecific competition in evaluations of species introductions.

Another objective for the present study was to test the hypothesis that peacock cichlids have had large negative effects on largemouth bass populations in southeast Florida. Based on the expectation that predation and food competition lead to reductions in largemouth bass density, I predicted that south Florida canals would have reduced abundance of largemouth bass relative to other types of Florida systems (i.e., lakes and streams) and that canals with peacock cichlids would have lower abundance of largemouth bass than canals lacking peacock cichlids. Therefore, I used population data for largemouth bass from south Florida canals and other Florida systems to test these hypotheses. An additional objective was to determine if the present study supported the prevailing view among many invasion biologists based on Clementsian (i.e., organisimic) communities or a more pluralistic alternative that incorporates Gleasonian (i.e., individualistic) communities.















CHAPTER 2
STUDY SYSTEM

Study Species

The largemouth bass Micropterus salmoides (Centrarchidae) is native to much of eastern North America, including southern Florida where it is considered a distinct subspecies, the Florida largemouth bass Micropterus salmoides floridanus (Lee 1980). The biology and ecology of largemouth bass are well studied (e.g., Heidinger 1975; Carlander 1977). Largemouth bass is a habitat generalist, with small individuals occurring in or near cover and larger individuals occupying a wide range of cover types and open water habitats (Keast et al. 1978; Mesing and Wicker 1986; Annett et al. 1996; Annett 1998; pers. obs.). It is predatory, with a general ontogenetic dietary progression from zooplankton to insects and small crustaceans to decapod crustaceans and fish (Chew 1974). Adults are considered primarily piscivorous (Chew 1974; Carlander 1977; Cailteux et al. 1996). Nests are constructed by males in relatively shallow water, often near cover. Largemouth bass in southern Florida spawn in winter to early spring (peak in March in canals) over a relatively short time period (Shafland 1999b). This results in a more or less distinct yearly cohort that appears in spring and declines in numbers through the year (pers. obs.).

Members of the genus Cichla (Cichlidae) are relatively large, predatory fishes that superficially resemble the North American black basses (i.e., Micropterus). Cichla are native to most major Atlantic drainages of tropical South America. In English, this genus


11






12


is known as peacock cichlids or peacock bass. Cichla comprise a group of five described and perhaps six or more undescribed species (Machado-Allison 1971; Kullander 1986; Kullander and Nijssen 1989). To date, the introduced form established in Florida has been designated as the peacock cichlid C. ocellaris, a species considered native to northeastern South America, although there remains some uncertainty about the alpha taxonomy of the genus (see Kullander and Nijssen 1989).

During 1984-1987, fingerling peacock cichlids were intentionally introduced into freshwater canals in Broward and Miami-Dade counties of southeast Florida by the Florida Game and Fresh Water Fish Commission-currently Florida Fish and Wildlife Conservation Commission; FWC (Shafland 1995). The introduction of this piscivore was planned to provide a biological control on the large populations of prey fish, primarily of exotic origin, to correct an unbalanced fishery (sensu Swingle 1950) and to supplement recreational fisheries that were principally based on largemouth bass (Shafland 1995).

Peacock cichlids spawn in nests constructed in relatively shallow water

throughout the warmer months (Zaret 1980a). Therefore, unlike largemouth bass, cohorts of age-0 peacock cichlids arise throughout much of the year, although they appear later than the single largemouth bass cohort. Peacock cichlids are habitat generalists and occur syntopically with largemouth bass at all life stages (pers. obs.).

Largemouth bass and peacock cichlids are not naturally sympatric. Nevertheless, these species co-occur in Hawaii and Puerto Rico due to the introduction of both species and in Florida due to the introduction of peacock cichlids. Both species are important sport fish in all three regions. Devick (1980) reported that peacock cichlids in Hawaiian reservoirs complemented largemouth bass as a sport fish due to differing spawning






13


seasons. Water withdrawals from the reservoirs often led to loss of nests or fry of substrate spawning fishes (e.g., largemouth bass and peacock cichlids). The extended breeding period of peacock cichlids relative to largemouth bass often allowed its successful reproduction and subsequent recruitment into the sport fishery, in effect carrying the sport fishery over periodic years of largemouth bass recruitment failure (Devick 1972; Devick 1980). Lilyestrom and Churchill (1996) found both species to coexist in Puerto Rican reservoirs with little evidence of competition based on slight differences in habitat use and diet and major differences in diel activity periods and spawning seasons. Shafland (1995, 1999a, b, c) noted overlap in diet but concluded that peacock cichlids had not harmed largemouth bass populations in Florida canals. Also of note is the fact that two or more species of peacock cichlids co-occur in many South American river systems (Jepsen et al. 1997; Winemiller et al. 1997; L. G. Nico, USGS, pers. comm.).

Largemouth bass and peacock cichlids are ecomorphologically similar based on gape width, general morphology, feeding mechanics, and diet (Norton and Brainerd 1993; Hill 1998). Given this similarity in ecology, a few experimental studies have explored possible competitive interactions between largemouth bass and peacock cichlids. Swingle (1967) concluded that peacock cichlids were less efficient predators of tilapia (probably redbelly tilapia Tilapia zillii but listed as T. melanopleura) and fathead minnows Pimephales promelas than largemouth bass under a similar stocking rate and conditions in an Alabama pond. Lilyestrom et al. (1994) stocked age-0 largemouth bass (mean TL = 48.8 1.0 mm) and peacock cichlids (mean TL = 18.1 1.1 mm) into experimental ponds in Puerto Rico and found that largemouth bass reduced growth and






14


survival of peacock cichlids but that largemouth bass grew faster and survived better with peacock cichlids. As pointed out by the authors, the results might have been strongly influenced by the disparity in initial sizes of the two species, although these were chosen to reflect the size relationship early in the peacock cichlid spawning season.

Although peacock cichlids have been introduced into several regions, there is little evidence of deleterious effects except in Lake Gatun, Panama (Zaret and Paine 1973) and in Venezuelan reservoirs (cited in Winemiller et al. 1997). The accidental introduction of peacock cichlids into Lake Gatun, Panama, and the subsequent environmental harm, is one of the most cited examples of the danger of fish introductions. Native fishes in the massive reservoir were reportedly devastated by this novel predator and local extirpations occurred (Zaret and Paine 1973). The predation effects apparently cascaded throughout the food web and resulted in significant changes in zooplankton, insect, and bird populations (Zaret and Paine 1973). Less frequently cited is the report of Welcomme (1988) mentioning the rebound of native fish populations in subsequent years.

Canal System

Southeast Florida is drained by a vast network of artificial canals. Much of the

canal network was constructed during the early to mid 20th Century (mainly from 1913 to 1969), primarily to lower the local water table to create agricultural land, to facilitate urban development, and to protect these lands from flooding (VanArman et al. 1984). Flood control is largely achieved by redirecting water from the Kissimmee River-Lake Okeechobee-Greater Everglades system generally south and east to the coast. Historically, most of the flow of this system was to the south and southwest. Although






15


some reaches are channelized streams, most of the canal system is entirely new, permanent, deep-water habitat-a freshwater habitat type that is naturally scarce in southern Florida (Loftus and Kushlan 1987). Many canals have substantial connection with the surficial aquifer, especially in eastern and southern portions of the region. The canal system is operated and maintained by the South Florida Water Management District (SFWMD) which controls flow and water routing with a combination of pump stations and water control gate structures.

Although canals are found in much of southern Florida, peacock cichlids are

largely confined to systems along the southeast coast, particularly the eastern portions of Broward and Miami-Dade counties (about 530 km of canal; Shafland and Stanford 1999). This species is also found in some southern Palm Beach County canals (total range of about 650 km of canals; Shafland and Stanford 1999). In Florida, peacock cichlids generally inhabit canals that flow through suburban and urban areas, including the densely-populated Ft. Lauderdale and Miami metropolitan areas. Some are found in rural canals (i.e., pass through agricultural lands) in western portions of the peacock cichlid range. Peacock cichlid distribution is limited by cool winter temperatures (Shafland 1995; i.e., lower lethal temperature of 15' C; Swingle 1967; Guest et al. 1980). Its range expands slightly north (i.e., into Palm Beach County) and west during years with mild winters.

Canal morphometry is characterized as box-cut, with vertical walls cut deeply (often > 3-5 m) into the substrate. In much of the region, the substrate is porous coral rock. A narrow shelf (usually 1-2 m wide), often with submersed vegetation, occurs along the banks and may be the only littoral zone for the canal. Smaller, lateral canals






16


are found in many systems. These are often shallower and have more stable aquatic macrophyte communities (pers. obs.).

Water chemistry can vary widely both temporally and spatially throughout the canal system (Shafland 1999a; pers. obs.). Intermittent releases of water from Lake Okeechobee or the Water Conservation Areas south of the lake greatly increase flow and alter water physico-chemical parameters (e.g., alkalinity, color, dissolved oxygen, pH, temperature, and turbidity). These changes can occur rapidly (i.e., within minutes or hours), especially in larger canals that are used as main conduits for de-watering. In particular, water releases can stress fishes by inducing hypoxic conditions. Tidal influences are minimal due to salinity control gates near the coast. These structures only open to release fresh water, although euryhaline fishes may swim up the current and enter the freshwater canal system.

The primary purpose of the canals is drainage and therefore SFWMD maintains an extensive program to remove aquatic macrophytes and overhanging terrestrial vegetation. Many canals support dense stands of aquatic macrophytes, often exotic invasives (e.g., hydrilla Hydrilla verticillata), that inhibit flow. The SFWMD uses biological (i.e., grass carp Ctenopharyngodon idella), chemical, and mechanical control methods, but mechanical harvesting is the most common. Mechanical harvesting occurs throughout the year and may be frequent in main canals during warmer months. Coverage of aquatic macrophytes exceeding 90% (percent areal coverage, PAC) can be reduced to < 5% in a single day by mechanical harvesters (pers. obs.). Such drastic changes in habitat structural complexity likely have dramatic effects on invertebrate and fish populations, including largemouth bass and peacock cichlids. Aquatic macrophytes






17


serve as important habitat and as a refuge from predation for many invertebrates and small fishes (Savino and Stein 1982; Werner et al. 1983; Gilinsky 1984; Chick and McIvor 1997; Jacobsen and Perrow 1998; Burks et al. 2001). Individual fishes, including age-0 largemouth bass and peacock cichlids, are also removed directly within the harvested vegetation (L. G. Nico, USGS, pers. comm.; pers. obs.). Electrofishing catchper-unit-effort (CPUE) for age-0 individuals is significantly lower immediately after harvesting activities (unpubl. data).

The peacock cichlid is one of about 25 exotic fish species established in Florida (Fuller et al. 1999; Hill 2002), with about 18 of these established in southeast Florida canals (Loftus and Kushlan 1987; Fuller et al. 1999; Loftus 2000; Hill 2002). In the canals, nonindigenous fishes, particularly cichlids, may dominate the fish assemblage in numbers and biomass (Courtenay et al. 1974; Shafland 1999a). The native fish fauna is relatively depauperate, with about 43 freshwater and 13 euryhaline species (Loftus and Kushlan 1987; Loftus 2000; Table 1). The native freshwater species are affiliated mostly with North American temperate families and are relatively recent invaders due to the short geologic time that southern Florida has been exposed above sea-level (Gilbert 1987).

The general perception in the literature is that native fishes in south Florida urban canals are subjected to numerous stressors (e.g., canal morphometry, variable and often harsh water physico-chemistry conditions, frequent aquatic macrophyte manipulation, and presumably strong biotic interactions with exotic fishes) that result in reduced population abundances (Loftus and Kushlan 1987; Courtenay 1997; Annett 1998). For example, several authors have suggested that cichlids exert strong negative effects on






18


native fishes (Buntz and Manooch 1968; Hogg 1976; Noble and Germany 1986; Courtenay and Robins 1989; Courtenay 1997). On the other hand, freshwater canals that traverse marsh areas serve as important refugia for large-bodied fishes (native and nonindigenous) during low water periods and generally enhance populations of such fishes (Loftus and Kushlan 1987); these canals are found mostly west of the distribution of peacock cichlids.

Study Sites

Most field work for this dissertation was conducted in Cutler Drain Canal (CIOOC) and Snapper Creek Canal (C-2) in the Biscayne Bay Drainage, Miami-Dade County, Florida (Figures 1 and 2). Collections of juvenile largemouth bass and peacock cichlids were made from the S-119 water control structure (250 38.58' N, 80' 20.31' W) to about 1.6 km upstream of the structure in Cutler Drain Canal and from the Don Shula Expressway bridge (downstream of the boat ramp on Snapper Creek Drive [25' 42.00' N, 800 21.25' W]) upstream up to, and including, the two lateral canals at the crossing of SW 99th Avenue in Snapper Creek Canal. These two canals were chosen because they are within the center of the range of peacock cichlids in southeast Florida, are relatively large canals, are accessible by boat, are in reasonably safe locations for personnel and equipment, and have different characteristic water chemistry and clarity.

Cutler Drain Canal (Figure 2A) in the Howard area is the smaller of the two canals and lacks large lateral canals in the study area. Under most conditions, water clarity is relatively high, flow is low, and groundwater is the major water source. Cutler Drain Canal is used as a main flow conduit only during infrequent high water events. I






19


0


25 km


Figure 1. Location of select canals in Miami-Dade County, Florida. Snapper Creek Canal and Cutler Drain Canal were study canals. Inset map shows Florida. Map redrawn from Shafland et al. 2001.

observed a substantial change in aquatic macrophyte species composition during the study period. In 1999, hydrilla Hydrilla verticillata, fanwort Cabomba caroliniana, and


Miami-Dade
County Miam.
n Miami
C
C.,Snap Creek Canal
Cutler Drain anal



Everglades National
Park Hom stead














0 25 Miles








20


A


BN





A




Snapper
Creek Al
Drive



SW 72 St |_ _'


Boat


SW 88 St Ramp


50 1 Mile


co U)0 1 2 km


Figure 2. (A) Cutler Drain Canal (C-IOOC) and (B) Snapper Creek Canal (C-2), MiamiDade County, Florida. Maps redrawn from Hidalgo 1997.


SW_112_St






0)






The Falls
c) Shopping How-ard Dr Center






S-119





0 1 Mile


0 1 2 km






21


bacopa Bacopa sp. were the main species, with hydrilla being dominant. By 2001, Asian hygrophila Hygrophila polysperma became the dominant macrophyte, with bacopa also being relatively common. Concurrently, numerous large grass carp were found in the area, presumably stocked by the SFWMD for aquatic macrophyte control. Hydrilla is a highly selected food of grass carp (Shireman and Smith 1983), a possible factor in the species replacement. Also, Cutler Drain Canal had low overall macrophyte density during 1999 due to macrophyte removal. Qualitatively, abundance of both native and exotic fish was relatively high (see Table I for a list of species).

Snapper Creek Canal (Figure 2B) is a major canal conduit in southeast Florida. The study area encompassed a large, mainstem canal and two large lateral canals. Water flow is frequently high in the main canal and water characteristics are most influenced by surface water. Snapper Creek Canal receives most of its water from the Tamiami Canal system, a significant drainage from the Water Conservation Areas and, ultimately, Lake Okeechobee from the north. The water is highly colored and occasionally turbid. Water clarity is usually greater in the lateral canals. Aquatic macrophytes are generally abundant in Snapper Creek Canal, especially in the laterals. Hydrilla dominated the main canal and cabomba in the laterals. Significant amounts of muskgrass Chara sp., pond weed Potamogeton sp., and vallisneria Vallisneria americana also occurred. Snapper Creek Canal has one of the most diverse fish faunas of the canal system (Table 1) and fish abundance was often high, especially in the lateral canals.







22


Table 1.- Fish species collected from two freshwater canals in southeast Florida in 19972001. A "*" indicates a nonindigenous species. A "?" indicates a tentative identification. Cutler Drain Canal had 31 species (21 native and 10 nonindigenous). Snapper Creek Canal had 40 species (22 native and 18 nonindigenous).



Scientific Name Common Name Occurrence

Cutler Drain Snapper Creek


Amiidae
Amia calva


Anguillidae
Anguilla rostrata

Atherinidae
Labidesthes sicculus

Catostomidae
Erimyzon sucetta

Centrarchidae
Enneacanthus gloriosus
Lepomis gulosus
Lepomis macrochirus Lepomis microlophus
Lepomis punctatus
Micropterus salmoides

Centropomidae
Centropomus undecimalis

Cichlidae
Astronotus ocellatus*
Cichla ocellaris*
Cichlasoma bimaculatum*
Cichlasoma citrinellum*
Cichlasoma managuense*
Cichlasoma urophthalmus*
Geophagus surinamensis*
Hemichromis letourneauxi*
Heros severus*
Oreochromis aureus*
Oreochromis mossambicus*
Tilapia mariae*

Clariidae
Clarias batrachus*

Clupeidae
Dorosoma cepedianum


Bowfin


American eel


Brook silverside


Lake chubsucker


Bluespotted sunfish Warmouth Bluegill Redear sunfish Spotted sunfish Largemouth bass


Common snook


Oscar
Peacock cichlid Black acara Midas cichlid Jaguar guapote Mayan cichlid Redstriped eartheater Jewel cichlid Green severum Blue tilapia Mozambique tilapia Spotted tilapia


Walking catfish


Gizzard shad X


x


x


x


x


x


x


x
x
x
x
x
x


x


x
x
x
x
x
x
x
x
x
x
X
X


x
x
x
x
x
x


x


x
x
x
x

x

x


x


x


x







23


Table 1. Continued.


Scientific Name Common Name Occurrence

Cutler Drain Snapper Creek


Cyprinidae
Ctenopharyngodon idella*
Cyprinus carpio*
Notemigonus crysoleucas

Cyprinodontidae
Lucania goodei

Elopidae
Elops saurus

Fundulidae
Fundulus chrysotus

Ictaluridae
Ameiurus natalis
Ameiurus nebulosus

Lepisosteidae
Lepisosteus platyrhinchus


Grass carp Common carp (koi) Golden shiner Bluefin killifish Ladyfish Golden topminnow Yellow bullhead Brown bullhead Florida gar


Loricariidae
Ancistrus sp.*
Hypostomus sp.*
Pterygoplichthys multiradiatus*


Mugilidae
Mugil cephalus


Percidae
Etheostoma fusiforme

Poeciliidae
Gambusia holbrooki Heterandria formosa
Poecilia latipinna


Striped mullet


Swamp darter


Eastern mosquitofish Least killifish Sailfin molly


x
x
x


x


x

x


x


x


x


x
x


x


x


x
x


x


x
x
x


x


x


x
x
x


x


x


x


x

x















CHAPTER 3
A QUALITATIVE RISK ASSESSMENT OF PEACOCK CICHLIDS
CICHLA OCELLARIS AND SOME IMPLICATIONS FOR THE METHODOLOGY


Introduction

Understanding and predicting the effects of introduced species are of interest for conservation biology and community ecology. Moreover, resource managers and regulatory authorities need such information. Unfortunately, prediction has proven elusive because of uncertainty due to the complexity of interacting factors in ecological systems (Baltz 1991; Ross 1991; FAO 1996). Furthermore, the outcome of some species introductions has been both unpredicted and undesired (i.e., Frankenstein Effect; Moyle et al. 1986).

There have been many methods used to predict the effect of species introductions. Quantitative modeling methods (e.g., Kolar and Lodge 2002) can be used, but these are data-intensive, require ecosystem-, taxon-, and stage-specific information, and have received little testing. As an alternative to quantitative models, qualitative methods also exist to estimate the effects of nonindigenous species (e.g., loop analysis; Levins 1974 [cited in Li et al. 2000]; Li and Moyle 1981; Li et al. 2000). Less formal methods include soliciting the opinions of experts who have extensive knowledge of the species and systems involved. Major drawbacks of this method are subjectivity and non-repeatability (Ambrose et al. 1996). A slightly better approach might involve a more formal method to integrate expert opinions into a risk assessment. Indeed, the Generic Nonindigenous


24






25


Aquatic Organism Risk Analysis Review Process (i.e., RAM methodology) has been developed by the federal Aquatic Nuisance Species Task Force as a method to predict risks associated with newly established species, species proposed for intentional introduction, and individual introduction pathways (RAM Committee 1998). This methodology uses information on species life history and ecology, characteristics of its native range and the region of potential introduction, and vectors of introduction to assign qualitative probabilities that the focal species will colonize open waters, successfully reproduce and become established, and adversely affect native species or economic activity (RAM Committee 1998).

The RAM methodology has been recently used by the U.S. Fish and Wildlife Service to synthesize information and to guide decision-making regarding proposals to list certain fishes as Injurious Wildlife Species (e.g., black carp Mylopharyngodon piceus; Nico et al. 2001). There have been only three formal applications of the RAM methodology for fish and all have been assessments for specific organisms or groups of organisms (i.e., not for introduction pathways)-black carp (Nico et al. 2001), snakeheads (Family Channidae) (W. R. Courtenay, Jr. and J. D. Williams, USGS, in prep.), and Asian swamp eel Monopterus albus (L. G. Nico, USGS, in prep.). Another risk assessment using this methodology is planned to include bighead carp Hypopthalmichthys nobilis and silver carp H. molitrix (J. D. Williams, USGS, pers. comm.). Of these risk assessments, only the black carp risk assessment was conducted prior to the occurrence of reproducing populations of the focal species (or one or more members of the family for snakeheads) in open waters within the USA. Furthermore, black carp was considered the test case for the use of the RAM methodology, partly









because there was still time to apply risk management strategies if warranted (Nico et al. 2001).

The black carp risk assessment tested the RAM methodology as a vehicle for risk assessment (i.e., would the method work as envisioned and produce an estimate of risk and certainty?). The black carp risk assessment demonstrated that the methodology could achieve these goals. On the other hand, there has been no testing of the predictions of the risk algorithm relative to any test case. In order to test these predictions, a number of species or groups that are already established should be chosen to represent a variety of nonindigenous species scenarios (e.g., various taxa, geographic distributions, and socioeconomic importance). Risk assessments could then be conducted and the results of the risk assessments (i.e., risk predictions) could be compared to the observed outcomes (up until that time). This process could provide insight into biases or grossly incorrect predictions based on the risk algorithm and suggest improvements in the process. Indeed, changes in the risk assessment methodology were envisioned by the authors who acknowledged the existence of uncertainty in the risk assessment process (RAM Committee 1998).

My objective is to begin this process of testing the RAM methodology using the peacock cichlid Cichla ocellaris (Cichlidae) as a test case. The emphasis of this assessment will be Florida, the only region of the continental USA with established Cichla (see Cichla introductions and Limiting factors on Cichla distribution, below). This species was chosen because 1) it has been introduced into the USA (i.e., Florida) as a sport fish and for biological control (i.e., has socio-economic importance), 2) it is predatory and ecomorphologically similar to native species (i.e., is of potential ecological






27


concern), 3) unlike the previously assessed fishes, it is limited to a rather narrow potential geographic range in the USA, 4) the species is not a new introduction (i.e., was introduced into Florida in 1984), and 5) it has been the subject of fisheries research in Florida and elsewhere. I will provide a brief overview of the biology, ecology, and history of introductions of Cichla to provide a background for the risk assessment process. I will then describe the RAM methodology and assign estimates of risk and certainty to the various rating elements. Lastly, I will discuss the predictions of the risk assessment algorithm relative to the data on observed effects of Cichla in Florida and suggest some modifications to the methodology.

Background Information

Cichla Taxonomy and Life History

The genus Cichla is composed of five valid nominal species (i.e., C. intermedia Machado-Allison, C. monoculus Spix, C. ocellaris Schneider, C. orinocensis Humboldt, and C. temensis Humboldt) and perhaps another six undescribed species (Kullander and Nijssen 1989). There has been considerable confusion over the alpha taxonomy of Cichla (Lowe-McConnell 1969; Kullander 1986; Kullander and Nijssen 1989; Shafland 1995); the name C. ocellaris (i.e., peacock cichlid) has been applied to most Cichla that have been introduced outside of South America.

The peacock cichlid is a relatively large (up to about 650 mm total length [TL] and 5.4 kg) species native to Atlantic drainages of northeast South America (Kullander and Nijssen 1989). Cichla superficially resemble the North American black bass of the genus Micropterus (Norton and Brainerd 1993).






28


Peacock cichlids are predatory and adults are largely piscivorous (LoweMcConnell 1969; Zaret 1980a; Shafland 1995, 1999b; Jepsen et al. 1997; Winemiller et al. 1997). Juveniles progress through ontogenetic dietary stages of feeding on zooplankton, insects, and decapod crustaceans prior to feeding on fishes (Zaret 1980a; Chapter 4).

Peacock cichlids form pairs during spawning. Although Cichla probably only spawn once per year in South American rivers (Jepsen et al. 1999), introduced populations in lakes, reservoirs, and canals spawn throughout the warmer months (LoweMcConnell 1969; Zaret 1980a; Jepsen et al. 1999; Shafland 1999b). Both parents guard a nest (cleaned substrate of wood, rock, or sand) where up to 10,000 eggs are laid (Zaret 1980a). The parents further protect the free-swimming fly for up to eight to 10 weeks (Zaret 1980). Fry occupy open water areas until they disperse from the guarded schools into littoral vegetation (Schroder and Zaret 1979; Zaret 1980a). Sexual maturity often is attained within one year at lengths of about 240-290 mm TL (Fontenele 1950; Shafland 1995). Subadults and adults are habitat generalists (pers. obs.). Cichla Introductions

Cichla are important food and sport fish and have been introduced into several

tropical and subtropical regions (Zaret 1980a; Welcomme 1981, 1988; Lever 1996; Fuller et al. 1999). Cichla have been introduced into experimental ponds in Alabama (Swingle 1967), Florida (Ogilvie 1966), and Georgia (Ogilvie 1966) and into power plant cooling reservoirs in Texas (Shafland 1995; Fuller et al. 1999). These introductions did not result in established populations. Cichla have been intentionally stocked by fisheries agencies






29


and are established in open waters of Florida, Hawaii, and Puerto Rico (Erdman 1984; Shafland 1995; Fuller et al. 1999).

In Hawaii and Puerto Rico, there were no native freshwater fish faunas to interact with introduced Cichla. Peacock cichlids were introduced as sport fish into reservoirs along with other nonindigenous species such as centrarchids (Lepomis spp. and largemouth bass Micropterus salmoides) and cichlids (Tilapia and Oreochromis spp.) (Erdman 1984; Devick 1980; Lilyestrom et al. 1994). Although the dynamics of other species were not reported, peacock cichlids and largemouth bass were not considered to significantly compete in Hawaii (Devick 1980) or Puerto Rico (Lilyestrom et al. 1994; Lilyestrom and Churchill 1996).

Peacock cichlids were intentionally introduced into freshwater canals in southeast Florida in 1984-1987 (Shafland 1995). The primary reasons for introducing peacock cichlids into Florida were 1) unbalanced prey-to-predator biomass ratios (sensu Swingle 1950) resulting from the high abundance of illegally introduced cichlids, particularly spotted tilapia Tilapia mariae, and 2) to supplement recreational fisheries that were based on largemouth bass (Shafland 1995). Studies evaluating the effects of the peacock cichlid introduction have not found evidence of negative effects on native fish assemblages, including largemouth bass (Shafland 1999a, c; Chapter 6), or on fishery values (Shafland and Stanford 1999). The Florida peacock cichlid fishery is estimated at US$ 8 million annually (Shafland and Stanford 1999).

On the other hand, the introduction of Cichla into Lake Gatun, Panama, is one of the most frequently cited examples of the danger of fish introductions. Cichla accidentally introduced into the Chagres River expanded downstream into the Panama






30


Canal reservoir in the 1960s (Zaret and Paine 1973). Once in the reservoir, Cichla populations expanded and this novel predator apparently decimated populations of small native fishes (Zaret and Paine 1973). The introduction also correlated with changes in zooplankton, insect, and bird populations (Zaret and Paine 1973). However, native fish populations may have rebounded since their initial reduction (Welcomme 1988). Limiting Factors on Cichla Distribution

Cichla are freshwater fishes with an upper lethal salinity of about 18 ppt

(Shafland and Hilton 1986). This factor would limit the ability of peacock cichlids to disperse via river mouths, estuaries, and salt marshes. Additionally, Cichla are true tropical fishes and cannot survive cold temperatures. The lower lethal temperature for Cichla in laboratory studies was about 150 C (Swingle 1967; Guest et al. 1980). Indeed, low winter temperatures resulted in the elimination of reproducing populations of purposefully introduced Cichla in some Texas power plant cooling reservoirs (Fuller et al. 1999). Experimental stocks of Cichla also have succumbed to cool temperatures in Florida (Ogilvie 1966; Shafland 1995; pers. obs.). Salinity and temperature limit the distribution of Cichla in Florida to freshwater canals and borrow lakes in the southeast portion of the state (Shafland 1995; see also Shafland and Pestrak 1982). Moreover, winter temperatures would make the establishment of Cichla in other portions of the continental USA unlikely in the absence of reliable thermal refuges. The possible limiting effects of water chemistry parameters (e.g., pH, hardness) have not been investigated. However, C. ocellaris appears to be adaptable (i.e., survives and reproduces) to a wide range of water conditions (pers. obs.). Additionally, unlike some established nonindigenous fishes in Florida (e.g., walking catfish Clarias batrachus),






31


Cichla are not capable of breathing air or crossing land barriers. However, because they are highly-favored sport fish, humans are likely to illegally transport Cichla to new waterbodies.

Ecological Interactions

Theoretically, peacock cichlids could have several ecological interactions with other species based on the dynamics of predation and competition. These interactions could have negative, positive, or mixed consequences for native species.

Cichla are predatory and can consume small native fishes, including juvenile

largemouth bass (Hill 1998). Indeed, in Florida, fish comprise 92% by number of the diet of peacock cichlids (Shafland 1999b). The effect of predation by Cichla could be strong-this was implicated in the observed changes in native fish abundance in Lake Gatun, Panama (Zaret and Paine 1973). Furthermore, there are many other examples of significant changes in native fish abundance following the introduction of piscivorous fish (e.g., Kaufman 1992; Mills et al. 1994; see also Taylor et al. 1984). In addition to the direct effect of predation on native fishes, increased predation risk due to Cichla could force vulnerable prey (including juvenile largemouth bass) to refuge and potentially compete with each other for food within the limited refuge area (Mittelbach 1984, 1988; Osenberg et al. 1994). These direct and indirect effects of predation could reduce the abundance of largemouth bass and other native fishes.

Nevertheless, predation by peacock cichlids on abundant nonindigenous cichlids (e.g., spotted tilapia), should have benefits for native fishes. Indeed, prey fishes can compete with early life stages of piscivores (e.g., largemouth bass) (Osenberg et al. 1994; Olson et al. 1995) and many prey fishes also are nest or larval fish predators (Bain and






32


Helfrich 1983; Popiel et al. 1996; Trexler et al. 2000). As a result, the introduction of a nonindigenous fish species that reduces the density of competitors or egg predators might actually lead to increased population sizes of native fish. Interestingly, peacock cichlids were introduced, in part, to reduce numbers of small nonindigenous fishes (Shafland 1995). Thus this possible positive effect of peacock cichlids was an implicit part of the original management goal. Some correlative data support this outcome within the Florida canal system (Shafland 1995, 1999a). For example, the introduction of peacock cichlids into Black Creek Canal was associated with a decline in the mean prey-to-predator biomass ratio from 21.1 (1983-1988) to 10.0 (1988-1991) and a 36% decline in the biomass and 55% decline in numerical abundance of spotted tilapia (the most abundant exotic species) from 1983-1988 compared to 1988-1993 (Shafland 1999a). Although suggestive, there are no data to help evaluate if these trends (i.e., decreased prey-topredator biomass ratios and reduced abundance of nonindigenous spotted tilapia) were caused by peacock cichlids.

The ontogenetic dietary shifts of peacock cichlids could lead to competition with various life stages of native fishes for zooplankton, insects, decapod crustaceans, or fish (Chapter 4). This mechanism could lead to reductions in native fish abundance following the introduction of Cichla. For example, peacock cichlids and largemouth bass overlap in diet (Shafland 1999b; Chapter 4; L. G. Nico, USGS, and J. E. Hill unpubl. data), which might indicate the potential for competition mediated through reductions in prey fish abundance, shifts in community composition, or changes in prey habitat use or behavior (Soluk and Collins 1988; He and Kitchell 1990; Hambright et al. 1991; Soluk 1993;






33


Persson et al. 1996). Nevertheless, the indirect effects of peacock cichlid predation on native fishes have not been investigated.

Peacock cichlids also might facilitate largemouth bass (or other native predators) by altering the behavior of their prey. For example, largemouth bass and peacock cichlids have different hunting strategies and activity periods. Peacock cichlids often rove in small schools and rely on speed rather than ambush as a hunting strategy (Erdman 1969; Devick 1972; Shafland 1995). Moreover, peacock cichlids are diurnal predators with different primary activity periods than largemouth bass (Lowe-McConnell 1969; Zaret 1980a; Lilyestrom and Churchill 1996). As a result of these differences, individuals of prey species may be presented with conflicting behavioral choices regarding activity patterns (e.g., foraging and refuging), perhaps enhancing the hunting success of one or both predators (Soluk and Collins 1988; Soluk 1993). This possibility has not been studied for this system.

Peacock cichlids might enhance populations of native predators by serving as prey themselves. For example, largemouth bass prey on juvenile peacock cichlids (Shafland 1999b). If largemouth bass dynamics are limited by food for the larger, piscivorous life stages, then the introduction of a new food source (i.e., peacock cichlids) could benefit largemouth bass. Although Florida canals often have high prey biomass, perhaps suggesting that food limitation is not very important, much of this potential food is morphologically unavailable to largemouth bass. Gape size (i.e., throat width) limits the maximum size of prey for largemouth bass (Lawrence 1958; Hambright 1991) and many canal prey species rapidly grow out of vulnerable sizes (Shafland et al. 1985; Hill 1998; L. G. Nico, USGS, unpubl. data). By breeding throughout the warm months






34


(Shafland 1999b), peacock cichlids provide numerous, appropriately-sized offspring for largemouth bass predation throughout the year (Chapter 4; pers. obs.). Small peacock cichlids are relatively abundant (Shafland 1999a, c; Chapter 4), are vulnerable (i.e., small, elongate, and with weak spines; Hill 1998) to a wide range of sizes of largemouth bass, and comprise 9% (by number) of prey consumed by largemouth bass in Tamiami Canal (Shafland 1999b).

Spawning site competition via territorial aggression has been postulated as a

significant negative centrarchid-cichlid interaction (Hogg 1976; Taylor et al. 1984; Noble and Germany 1986). However, evidence of this mechanism is largely anecdotal or correlative and no quantitative investigation has demonstrated spawning site competition for these fishes. Centrarchids would be the most likely native group to experience spawning site competition from peacock cichlids (or other cichlids) due to general similarities in reproductive behavior (Heidinger 1975; Zaret 1980a; Annett et al. 1996). However, in a study of the effect of blue tilapia Oreochromis aureus on largemouth bass reproduction in experimental ponds, non-nesting (i.e., all female) blue tilapia had a similar effect on largemouth bass reproduction as did all male or mixed sex treatments (i.e., nest-building and territorial) (Shafland and Pestrak 1983). This result suggests that spawning site competition is not the mechanism responsible for observed reductions of largemouth bass recruitment that occurred in some lakes with introduced blue tilapia (see Taylor et al. 1984; Germany and Noble 1986). Additionally, there is relatively small overlap in spawning times of largemouth bass and peacock cichlids (Shafland 1999b; pers. obs.) and little agonistic behavior has been observed between the two species (P. L. Shafland. FWC, pers. comm.; Chapter 5; pers. obs.). However, there is temporal, and






35


perhaps spatial, overlap in spawning activity between peacock cichlids and other centrarchids (e.g., sunfish Lepomis spp.) (pers. obs.).

Of course, negative and positive effects of peacock cichlids could both exist.

Moreover, these effects may not be spatially or temporally consistent. Depending upon the direction and intensity of these effects, the outcome for native fishes could be mixed (e.g., stage or context dependent), or even cancel. For example, peacock cichlids could enhance juvenile largemouth bass growth by reducing numbers of competing fishes while simultaneously decreasing adult largemouth bass growth by depleting numbers of available prey. This type of effect is documented for bluegill Lepomis macrochirus and largemouth bass in temperate lakes (Olson et al. 1995). Nevertheless, additional research is required to document mixed effects in this system. Fishing Effects

The influence of peacock cichlids on sport fishing could have mixed effects for the other major canal sport fish-largemouth bass. Combined recreational fishing values for largemouth bass and peacock cichlids (the most commonly targeted and valuable sport fish) in southeast Florida canals exceed US$ 13 million annually (Shafland and Stanford 1999). Fishing tactics for both species are similar, although some important differences occur (Hidalgo 1997). These differences (e.g., fishing time of day and lure choice) are exploited by anglers specifically targeting one or the other species (Hidalgo 1997). Given the high angler effort and increasing popularity of peacock cichlids in southeast Florida canals (Shafland and Stanford 1999), exploitation of largemouth bass could be increasing. This could result from by-catch of anglers targeting peacock cichlids and by the promotion of canal fishing resulting from the attention that peacock






36


cichlids have generated. On the other hand, the introduction of peacock cichlids also might reduce exploitation mortality of largemouth bass. Because of the differences in fishing for the two species, peacock cichlid fishing may actually redirect angler effort away from largemouth bass and therefore reduce largemouth bass total mortality. This interesting question has not been investigated.

Qualitative Risk Assessment

The potential of nonindigenous species to be introduced, become established, and become invasive can be qualitatively modeled given information concerning introduction pathways and the biology and ecology of the organism in question (RAM Committee 1998). There are two main sections of the risk model: 1) Probability of Establishment and 2) Consequences of Establishment. An overall Organism Risk Potential (ORP) is the final output of the model (i.e., low, medium, or high) and is calculated based on the risk estimates of both sections. Within these sections of the risk model there are seven rating elements, each assigned a qualitative estimate of risk (i.e., low, medium, and high) and an estimate of certainty (i.e., very certain, reasonably certain, moderately certain, reasonably uncertain, and very uncertain). Four of the seven elements address the Probability of Establishment and three elements attempt to predict the Consequences of Establishment. Rating Elements of Risk Model-Probability of Establishment

1) Estimate probability of the exotic organism being on, with, or in the

pathway. The peacock cichlid is established in southeast Florida (Shafland 1995) and is thus in the United States Pathway. Various members of the genus Cichla are occasionally imported in the aquarium industry (pers. obs.). State and commonwealth






37


agencies have imported or exchanged peacock cichlids for fishery introductions (Maciolek 1984; Shafland 1995; Fuller et al. 1999). High Very Certain

2) Estimate probability of the organism surviving in transit. The peacock cichlid has survived transit on numerous occasions. High Very Certain

3) Estimate probability of the organism successfully colonizing and maintaining a population where introduced. The peacock cichlid has been reproducing in southeast Florida since the mid-1980s and is considered established in the state (Shafland 1995). Cichla have been intentionally introduced and established reproducing populations in Hawaii and Puerto Rico in reservoirs (Fuller et al. 1999). Nevertheless, peacock cichlids have been unsuccessful at establishing persistent populations in Texas power plant cooling reservoirs and possibly in southern Florida in the 1960s due to cool temperatures (Fuller et al. 1999). A congener, the speckled peacock cichlid C. temensis, was intentionally introduced alongside C. ocellaris yet failed to become established in southeast Florida (Shafland 1995). Because of thermal constraints, it is unlikely that Cichla could become established in the continental USA outside of southeastern Florida (see Limiting factors on Cichla distribution, above). High

- Very Certain

4) Estimate probability of the organism to spread beyond the colonized area. The initial introduction area of the peacock cichlid nearly encompasses its present range in Florida. Peacock cichlids have successfully colonized adjacent canals and have been illegally or naturally transplanted into artificial lakes and ponds in southeast Florida. Limiting factors for peacock cichlids include a lower lethal temperature of about 150 C (Swingle 1967; Guest et al. 1980) and an upper salinity tolerance of about 18 ppt (Guest






38


et al. 1980; Shafland 1995). Relatively warm winters allow peacock cichlids to temporarily colonize canals slightly north and west of the targeted area of introduction (Shafland 1995; Loftus 2000; L. G. Nico, USGS, pers. comm.). Nevertheless, given the inability of peacock cichlids to survive cool winter temperatures or salt water, there is little possibility for significant expansion outside of the present range in southeast Florida. Additionally, due to cool winter temperatures, it is unlikely that Cichla could become established in other areas of the continental USA. Low Very Certain Rating Elements of Risk Model-Consequences of Establishment

5) Estimate economic impact if established. The peacock cichlid could negatively affect the largemouth bass recreational fishery in southeast Florida via predation and competition for food (see Ecological interactions, above). Moreover, peacock cichlid fishing may have negative effects on largemouth bass populations (see Fishing effects, above). Also, peacock cichlids theoretically could enhance largemouth bass populations via ecological and fisheries mechanisms (see Ecological interactions and Fishing effects, above). However, studies investigating the effects of peacock cichlids on largemouth bass have not reported significant reductions in largemouth bass abundance (Shafland 1999a, c; see also Chapter 6). The largemouth bass fishery in the area occupied by peacock cichlids has been estimated at US$5 million annually (see Shafland and Stanford 1999).

Major objectives for introducing peacock cichlids into Florida were to create a novel sport fishery and to supplement recreational fisheries for largemouth bass (Shafland 1995). These objectives have been met and the recreational fishery has been estimated at US$8 million annually (combined fishery of largemouth bass and peacock






39


cichlids of US$13 million) (Shafland and Stanford 1999). Therefore, the overall economic effect has been positive for the region of introduction. The economic risk is estimated to be low due to the sustained values of largemouth bass fishing where peacock cichlids have been present since 1984-1987 (Shafland and Stanford 1999) and the relatively small potential effect on statewide fishery values (i.e., the small geographic area occupied-portions of three south Florida counties) regardless of any possible negative impacts on the largemouth bass fishery. Additionally, the similarity of largemouth bass and peacock cichlid fishing (i.e., similar equipment and techniques; Hidalgo 1997) suggests that largemouth bass anglers, guides, and tackle shops could adjust to any losses to the largemouth bass fishery by switching to peacock cichlids. Low

- Very Certain

6) Estimate environmental impact if established. There is a potential for

peacock cichlids to adversely affect native fishes by predation and competition for food (see Ecological interactions, above). Predatory fishes can exert top-down effects that may alter trophic relations and energy flow in aquatic systems (Zaret and Paine 1973; Carpenter et al. 1987; Northcote 1988; Mittelbach et al. 1995). Nevertheless, it is also possible that peacock cichlids may positively affect native fishes by reducing numbers of other exotic fishes and by serving as prey themselves. Current data indicate the operation of mechanisms by which peacock cichlids negatively effect native species (e.g., predation), but also suggest that peacock cichlids have not significantly reduced native fish populations in southeast Florida (Shafland 1995, 1999a, c; Chapter 6). Furthermore, peacock cichlids are confined to artificial water bodies within a small geographic area that lacks highly specialized or endemic native fishes. Based on existing information, the






40


risk of environmental harm is relatively low, but risks may increase if conditions within the system significantly change. One possible change that could increase the risk of environmental harm from peacock cichlids is climatic change. An increase in mean low temperatures in southern Florida might lead to an expansion of the range of peacock cichlids. Nevertheless, peacock cichlids are less tolerant of cool temperatures than are many established tropical fishes in Florida (Shafland and Pestrak 1982; Shafland 1995) and experience winterkills even within the present thermal refuge of the canal system, (P. L. Shafland, FWC, pers. comm.), a region with mean monthly low air temperatures in January (Florida's coldest month) of> 18.30 C (Shafland and Pestrak 1982). Low Moderately Certain

7) Estimate impact from social and/or political influences. Peacock cichlids have had fairly broad approval and support from within the angling community before and since the introduction in the mid-1980s. The recreational fishery is popular and supports important socio-economic activity (Shafland and Stanford 1999). A few anglers, environmental groups, and some scientists, particularly ecologists, express concern over the peacock cichlid. Anglers against the introduction are concerned with possible effects on the largemouth bass fishery. Environmental groups and many ecologists are against the introduction of any non-native species, their concerns generally based on the theoretical or potential effects of nonindigenous species on native species (e.g., Courtenay and Robins 1989; Minns and Cooley 2000). An additional concern for environmental groups and governmental agencies is the possibility of peacock cichlids invading the nearby Everglades National Park. Given the widespread acceptance and






41


support of peacock cichlids in south Florida, the overall effect is estimated to be positive. Low Moderately Certain

Organism Risk Potential (ORP)

The Organism Risk Potential (ORP) has two components: 1) the Probability of Establishment and 2) the Consequences of Establishment. The Probability of Establishment is the lowest ranking amongst the first four ratings elements of the risk model (i.e., within pathway, entry potential, colonization potential, and spread potential). Given the fact that peacock cichlids are established in Florida, intuitively this probability should be rated as "High". Nevertheless, because the potential for spread is rated as "Low", the overall Probability of Establishment from the model is rated as "Low" (Figure 3). The fourth rating element (i.e., spread potential) was estimated as "Low" because the area of introduction (i.e., the area of canals where peacock cichlids were stocked), the intended target area of the introduction, and the potential maximum range of colonization for peacock cichlids in the continental USA (i.e., Florida) are very similar (i.e., about 650 km of canals; Shafland and Stanford 1999; see Limiting factors on Cichla distribution, above).

The Consequences of Establishment rating is defined as the highest of the

probabilities assigned to the economic and environmental impacts. Perceived impact (i.e., social and political) does not influence the assessment of Consequences of Establishment unless the economic and environmental risk estimates are both "Low". All three elements were estimated to be "Low" for the potential for negative consequences. Therefore, the Consequences of Establishment are estimated to be "Low" (Figure 3).






42


Probability Organism Entry Colonization Spread
of = within Potential Potential Potential = Low
Establishment Pathway (High) (High) (Low)
(High)




Consequences
of = Economic Environmental Perceived = Low
Establishment (Low) (Low) (Low)





Probability of Consequences of
ORP Risk = Establishment Establishment = Low
(Low) (Low)



Figure 3. Organism Risk Potential (ORP) for peacock cichlid Cichla ocellaris within the continental USA based on the RAM methodology (RAM Committee 1998), given a "Low" Probability of Establishment. The Probability of Establishment is the lowest rating among the four rating elements (i.e., within pathway, entry potential, colonization potential, and spread potential). The Consequences of Establishment is the highest rating between economic and environmental ratings. The ORP is the average of Probability of Establishment and Consequences of Establishment ratings. An ORP of "Low" is defined as "acceptable risk organism(s) of little concern (does not justify mitigation)" (RAM Committee 1998).

The ORP rating is the average of the two risk potentials assigned for the

Probability of Establishment and the Consequences of the Establishment, with the

average rounded up (i.e., a conservative estimation). Therefore the ORP for peacock

cichlids introduced into the continental USA is "Low" (i.e., "Low" Probability of

Establishment and Consequences of Establishment) (Figure 3). Conversely, had the risk

of spread rating element been disregarded, then the Probability of Establishment would

have been assigned a rating of "High" and the ORP would have been "Medium" (i.e.,






43


average of "High" Probability of Establishment and "Low" Probability of Consequences). An ORP of "Medium" is considered unacceptable risk and would require mitigation (RAM Committee 1998).

Discussion

Risk assessments attempt to predict effects of nonindigenous species and are used as a basis for economic, environmental, and policy decisions. As a result, it is important to evaluate the methods used to conduct risk assessments to determine if risk assessments yield accurate predictions and thus are useful tools.

Although the peacock cichlid is well-established in southeast Florida (Shafland 1995), the RAM method algorithm resulted in an estimate of the Probability of Establishment as "Low". This somewhat nonsensical outcome of the model (i.e., rating the probability that an established species would become established as low) was due to the similarity in the geographic area of the initial introduction and the small potential maximum range in the continental USA (i.e., there is little chance of range expansion). Previous risk assessments using this methodology have been conducted on fishes with large potential ranges in the USA (i.e., black carp, Asian swamp eel, and snakeheads). Therefore, the situation with peacock cichlids may represent a relatively unusual caseone not adequately anticipated in the methodology. Perhaps any apparent conflict over the assignment of a "Low" risk of establishment is semantic. The Probability of Establishment label could be modified to Probability of Establishment and Spread to avoid confusion in such cases (see also Kolar and Lodge 2002). Regardless of the label, the essence of this risk element (i.e., Probability of Establishment) is the potential for an organism to access the region of interest, and then survive, reproduce, and spread once






44


there. Additionally, this issue suggests that each of rating elements should be considered separately as well as in combination, and that any method of risk assessment requires some level of flexibility in interpretation, factors acknowledged by the authors of the methodology (RAM Committee 1998).

The qualitative risk assessment methodology combines estimates of risk for two very different processes into a single estimate of the risks posed by a nonindigenous species (i.e., establishment and consequences). The assignment of the average risk rating between these two processes as the overall risk potential that this organism represents (i.e., ORP) can be misleading. An organism may have a low probability of establishment but a high potential for adverse effects and have the same Medium ORP rating as a species with a high potential for establishment but a very low probability of negative consequences. A Medium ORP is defined as an unacceptable risk, justifying mitigation (RAM Committee 1998). A logical argument would be to place more emphasis on the potentially harmful species even though there is a relatively low chance of becoming established because of the greater danger posed by a successful invasion. Indeed, it might be a waste of governmental resources and an undue economic burden to attempt to exclude species that have low potential for environmental or economic harm regardless of their perceived ability to colonize. It may be more appropriate to consider the two processes independently and do away with or otherwise de-emphasize the ORP because the ORP rating seems to lose information rather than simply condense it.

Two alternative ORP estimates for peacock cichlids (i.e., "Low" [Figure 1] and "Medium") were presented in this risk assessment. The estimate of "Low" was based on a strict interpretation of the RAM method algorithm. The estimate of "Medium" risk






45


logically could be argued based on the fact that peacock cichlids are established in the continental USA (i.e., should have a "High" Probability of Establishment). An ORP result of "Medium" risk would place the peacock cichlid into the category of unacceptable risk (RAM Committee 1998). However, it seems illogical to require mitigation against a socio-economically important sport fish when there is no evidence that it has harmed native fish species in the nearly two decades since it has been intentionally introduced (Shafland 1995, 1999a,c; Shafland et al. 2001; Chapter 6).

The RAM algorithm can only return a "Medium" or "High" ORP if a

nonindigenous species is established or has a medium or higher probability of becoming established (RAM Committee 1998). This presupposes that if established, any nonindigenous organism will have an unacceptable risk of negative impacts. This inherent bias prejudices a supposedly objective process to lead to exclusion or restriction of any species that may become established. If this strict precautionary principle (FAO 1995) is the foundation for the RAM methodology, then the entire assessment of risk hinges on the estimate of the ability of the organism to become established. The section concerning the possible consequences to the environment or economy become of no importance, except perhaps in defining regulatory, eradication, or control priorities. For example, the upcoming risk assessment of bighead carp and silver carp must conclude that these species represent an unacceptable risk and require mitigation (i.e., the ORP must be either "Medium" or "High") due to the fact that both already are established in the Mississippi River (and could spread to other regions), regardless of the assessor's assignment of risk associated with the presence of these species. Therefore, the outcome of the risk assessment for these species is largely predetermined.






46


Regardless of any inherent bias, the predictive ability of any risk assessment tool can only be as good as the information put into the model. Although it provides a meaningful way of approaching the problem and attempts to interject much-needed objectivity into the field, the method still leaves substantial room for personal interpretation and subjectivity (i.e., "uncertainty of the assessor"; RAM Committee 1998). For example, everyone will not agree with my estimates of risk and certainty for peacock cichlids or other species. Some disagreements will be over interpretation of data, others over philosophy. In particular, some ecologists would rate the environmental risk of peacock cichlids as high due to demonstrated predation and the potential for competition affecting native species regardless of existing data. Furthermore, data are lacking for many non-fish components of southeast Florida's aquatic systems, increasing uncertainty. Although empirical rules of bio-invasion have been proposed that seek to provide a framework for prediction based on past invasions (Moyle and Light 1996a), their universality seems doubtful and the performance of many nonindigenous species has been unpredicted (Moyle et al. 1986). Indeed, in any risk assessment process, a major impediment is uncertainty associated with the organisms (RAM Committee 1998). In the case of peacock cichlids, in contrast to other fishes assessed by this methodology, there are fairly substantial data on introduced populations within the region of concern (i.e., continental USA) (Shafland 1995, 1999a, b, c; Hill 1998; Shafland and Stanford 1999; Shafland et al. 2001; Chapter 4, 5, and 6), thereby increasing the estimates of certainty (RAM Committee 1998).

In the RAM methodology guide, it is argued that the actual ratings assigned to individual elements are not as important as the transparent presentation of information






47


upon which the ratings are based (RAM Committee 1998). Although this is an important point, the fact that the algorithm produces a single term, the ORP, as a guide to decision making (i.e., acceptability of risk and requirement of mitigation) places a distinct emphasis on the ORP as the end product of the risk assessment. This factor, coupled with the previously mentioned bias relative to any species capable of becoming established (i.e., a de facto assumption that all established nonindigenous species represent an unacceptable risk), somewhat limit the utility of the process. Indeed, it could be argued that this process merely adds a pseudo-quantitative shroud to the subjective method of expert opinion. On the other hand, the process does provide a vehicle for gathering information on the focal species and forces a qualitative evaluation of various categories of risk.

In conclusion, it is clear that the peacock cichlid test case pointed out problems associated with the RAM methodology. Established species with small potential ranges yield a nonsensical result for the Probability of Establishment. This could be largely addressed by changing the label to Probability of Establishment and Spread. The combining of probabilities of a species becoming established with subsequent consequences of the establishment into an ORP output places too much emphasis on a single value and loses important information in the process. This could be addressed by reducing the prominence of the ORP (or eliminating it altogether) and retaining the Probability of Establishment and Spread along with the Consequences of Establishment as the primary outputs of the RAM model. Regulators and managers should have clear objectives in mind prior to commissioning a risk assessment and should determine if the threat of successful introduction alone is the primary concern or if negative effects of the






48


nonindigenous species are more (or equally) important. Additionally, the built-in assumption of the RAM model that all nonindigenous species capable of establishing persistent populations represent an unacceptable risk should be acknowledged. The RAM guide states that a risk assessment cannot be used to determine an acceptable level of risk, yet the ORP ratings are defined explicitly in terms of acceptable and unacceptable risk (RAM Committee 1998). The level of acceptable risk is not a question for the assessor, but one for society, and definitions should be changed to remove conclusions of acceptability. Also, conclusions of the need for mitigation are not needed, although suggestions for mitigation (or eradication or control) may be appropriate within the risk assessment. In essence, the definitions of "Low", "Medium", and "High" risk should be self-explanatory and interpreted within the context of the stakeholder environment. Lastly, additional test cases should be assessed to determine if other biases occur in the methodology.















CHAPTER 4
ONTOGENETIC DIETARY SHIFTS AND DIETARY OVERLAP
OF A NONINDIGENOUS FISH AND AN ECOMORPHOLOGICALLY SIMILAR NATIVE FISH


Introduction

Changes in resource use during ontogeny have considerable influence on the dynamics of size-structured populations (Werner and Gilliam 1984). For piscivorous fishes, ontogenetic diet shifts can be dramatic. For example, the largemouth bass Micropterus salmoides typically progresses through relatively discrete stages of feeding on zooplankton, insects, and decapod crustaceans before becoming largely piscivorous (Keast and Eadie 1985; Olson 1996; but see Schramm and Maceina 1986; Cailteux et al. 1996; Shafland 1999b; Huskey and Turnigan 2001 for studies reporting continued and substantial consumption of decapod crustaceans in addition to fish). These dietary shifts are size-related and dependent upon the relative size of the predator's gape and the prey's body depth or girth (Hambright 1991). For gape-limited species, growth-related increases in gape size allow for capture of larger, more energetically profitable prey (e.g., initiation of piscivory; Pasch 1975; Ludsin and DeVries 1997). Indeed, individual growth is enhanced by switching to a fish diet (Pasch 1975; Timmons et al. 1980; Olson 1996). Moreover, increased growth also leads to increased survival of age-0 largemouth bass because mortality is strongly size-dependent (Gutreuter and Anderson 1985; Ludsin and DeVries 1997).


49






50


Trophic morphology, including gape width, is considered to be an indicator of a fish's ability to handle prey (Wainwright and Richard 1995; Wainwright 1996). Indeed, numerous studies have investigated gape width and its relationship to prey use in largemouth bass (e.g., Lawrence 1958; Hambright 1991; Hambright et al. 1991). Given this link between trophic morphology and prey use, species that are morphologically similar might be expected to exhibit similar ontogenetic dietary shifts and have considerable overlap in diet. An ecological implication of this is that if prey are limiting and are depleted by predators, then competition will occur between species that significantly overlap in diet.

The peacock cichlid Cichla ocellaris (Cichlidae) was intentionally introduced into the native range of largemouth bass (Centrarchidae) in southeast Florida in 1984-1987 (Shafland 1995). Norton and Brainerd (1993) described these species as ecomorphologically similar based on morphology, diet, and feeding mechanics. Moreover, Hill (1998) found broad similarity in length-specific gape width between these species. Although the ontogenetic dietary patterns of largemouth bass diet are well known (Keast and Eadie 1985; Olson 1996), little information is published on the diet of juvenile Cichla (Lowe-McConnell 1969; Zaret 1980a; Lilyestrom and Churchill 1996; Shafland 1999b). Nevertheless, available information and the similarity in morphology of these species imply that largemouth bass and peacock cichlids may have similar diets through ontogeny. Moreover, preliminary observations of canals in southeast Florida suggested that juvenile peacock cichlids were abundant and used similar habitats as juvenile largemouth bass (pers. obs.).






51


Studies documenting ontogenetic dietary stages or dietary overlap of naturally cooccurring fishes are common (e.g., Nico and Taphorn 1988; Winemiller 1989, 1991; Murie 1995; Scott and Angermeier 1998). Additionally, much research has investigated shared resource use of introduced and native fishes, including species that are morphologically similar (e.g., Mathur 1977; Parrish and Margraf 1990; Matthews et al. 1992; Huckins 1997; Huckins et al. 2000; Declerck et al. 2002; Sutton and Ney 2002; Stoffels and Humphries 2003). However, studies are rare that compare dietary ontogeny and estimate the overlap in diet of a nonindigenous fish with an ecomorphologically similar, but phylogenetically distant, native fish.

This study investigated the hypothesis that an introduced fish will have an

ontogenetic dietary progression similar to that of a native species of similar morphology, regardless of their phylogenetic relations. Additionally, this study documented the temporal co-occurrence of juvenile largemouth bass and peacock cichlids of similar size and ability to handle prey, and estimated the extent to which the two species overlapped in prey use in two southeast Florida canals.

Methods

Collections for this study were made in Cutler Drain Canal (C-IOOC) and Snapper Creek Canal (C-2), Biscayne Bay Drainage, Miami-Dade County, Florida. Both canals occur in urban and suburban areas of the Greater Miami Metropolitan Area. Cutler Drain Canal averages about 25 m wide and 3-4 m deep. It had highly variable aquatic macrophyte species composition and abundance due to biological and mechanical weed control methods used by the South Florida Water Management District (SFWMD). In sampled reaches, aquatic macrophyte coverage ranged from about 2% percent areal






52


coverage (PAC) to > 75% PAC. Asian hygrophila Hygrophila polysperma was the dominant submersed macrophyte. Water clarity was generally > 3 m of Secchi depth. Collections were made from the S-119 water control structure (250 38.58' N, 800 20.3 W) upstream for about 1.6 km. Snapper Creek Canal is a large, important conduit for water drainage. Collections were made from the Don Shula Expressway bridge (downstream of the boat ramp on Snapper Creek Drive [25' 42.00' N, 80' 21.25' W]), upstream along the main canal and in two adjacent lateral canals accessed near the crossing of SW 991h Avenue-a total of about 1.8 km. In sampled reaches, aquatic macrophyte abundance ranged from < 5% PAC to > 50% PAC in the main canal and from < 20% PAC to near 100% PAC in the laterals. The main canal had highly variable aquatic macrophyte abundance, with hydrilla Hydrilla verticillata being the dominant species. The lateral canals had relatively little mechanical harvesting of weeds by the SFWMD, and therefore had relatively stable aquatic macrophyte abundance. Hydrilla and fanwort Cabomba caroliniana were the dominant species in the lateral canals. Muskgrass Chara sp., a macroalgae, also was common in the southern lateral canal. Snapper Creek Canal had highly colored and occasionally turbid water (Secchi depth usually 1-2 in).

Four collections were made in Snapper Creek Canal in 1999 (July, September, October, and November) and two collections were made in both canals in 2001 (August and October). Juvenile largemouth bass and peacock cichlids were collected by daytime boat electrofishing. An attempt was made to collect all individuals < 150 mm TL until target numbers were reached, 25 (in 1999) or 50 (in 2001). Fish were euthanized and






53


preserved in formalin. Formalin was injected into the abdominal cavity of specimens collected in 2001.

In the laboratory, specimens were transferred to ethanol. All fish were measured for maximum total length (TL), and a representative sample covering the collected size range of largemouth bass and peacock cichlids was measured for gape width (GW). Gape width measurements were made as external mouth width with the mouth closed to approximate intercleithral distance (Lawrence 1958; Hambright 1991; Hill 1998). Wainwright (1996) determined that intercleithral distance is the limiting dimension for swallowing prey in centrarchids. External mouth width has an approximate 1:1 relationship to intercleithral distance in largemouth bass (Lawrence 1958). Additionally, Hill (1998) evaluated this method for peacock cichlids and found that peacock cichlid prey maxed out at about 96.5% of predicted gape size (maximum sizes ranged from 88104% with a SD of 5.6%).

Visceral masses of all fish were removed, stomachs were opened with forceps, and prey items were removed, identified, and counted with the aid of a dissecting microscope. Only items occurring in the stomach portion of gastro-intestinal tract were examined and used for diet analyses. All prey measurements and predator GW measurements were made with dial calipers. Predator total-length measurements were made with a measuring board. Fish prey were measured for maximum standard length

(SL). Decapod crustaceans in the fish stomachs were found with abdomens folded underneath the cephalothorax and were therefore measured for folded body length (FBL; i.e., tip of rostrum to outside of bend in abdomen). Maximum body depth (BD) was measured for each prey if the specimen was intact. Decapod crustacean BD was






54


measured with the abdomen folded underneath the cephalothorax. An estimate of the volume of prey in the stomachs was obtained by water displacement using either a graduated cylinder or a pipette. A qualitative estimate of stomach fullness also was recorded. The values ranged from 1 to 4, with 1 being empty, 2 having some food but less than 25% of the estimated stomach capacity, 3 having 25-75% of stomach capacity filled, and 4 having > 75% of estimated stomach capacity.

Juvenile largemouth bass and peacock cichlids were grouped into 10-mm TL size classes. Estimates of length-specific gape width were used to construct histograms of gape size distribution for each sampling date to determine the extent of overlap in morphological ability of juvenile largemouth bass and peacock cichlids to handle prey. Estimates of gape width were obtained by regressing gape width measurements on total length (PROC GLM; SAS Institute, Cary, North Carolina). Analysis of covariance (ANCOVA) was used to test for differences in slopes and intercepts of developed regressions for largemouth bass and peacock cichlids (PROC GLM; SAS Institute, Cary, North Carolina). All statistical analyses were conducted at a significance level of a =

0.05 unless otherwise noted.

The stomach contents of juvenile largemouth bass and peacock cichlids were

described by frequency of occurrence, numerical abundance, and volume for all prey, as well as prey size for fish and decapod crustacean prey. A predator was defined as a piscivore when fish prey were found in at least 60% of predator stomachs that contained any food (Bettoli et al. 1992; Cailteux et al. 1996). Prey sizes (i.e., prey BD and prey length of fish and decapod crustacean prey) were regressed on predator total length and the resulting relations were compared with ANCOVA (PROC GLM; SAS Institute, Cary,






55


North Carolina). Also, relative prey size was estimated in two ways-prey body depth divided by predator gape width, and prey length (SL for fishes and FBL for decapod crustaceans) divided by predator total length, multiplied by 100 to obtain percentages.

Estimates of stomach fullness were compared between species within sampling dates by the Wilcoxon rank sum test (PROC NPAR1 WAY; SAS Institute, Cary, North Carolina) and within species and among sampling dates by the Kruskal-Wallace test (PROC NPARI WAY; SAS Institute, Cary, North Carolina). A significant KruskalWallace test was followed by Dunn's non-parametric multiple comparison test (Hollander and Wolfe 1973). Due to the conservative nature of Dunn's test, a significance level of a = 0.10 was used to distinguish mean rank differences among sampling dates (Hollander and Wolfe 1973).

Dietary overlap values between predators within sampling date were estimated using Schoener's Index (Schoener 1970):

C, = 1 0.5 ( pi pyi),

where C., is the index value, pxi is the proportion of food i in the stomach contents of species x, and p, is the proportion of food i in the stomach contents of species y. This index ranges from 0 (no overlap) to 1 (complete overlap). I defined proportions using either numerical abundances or prey volumes. Although subjective, values of dietary overlap > 0.6 have been considered as significant dietary overlap (Zaret and Rand 1971; Mathur 1977). Schoener's index of dietary overlap has been recommended for use when resource availability data are lacking (Hurlbert 1978; Wallace 1981).






56


Results

A total of 548 fish were examined for stomach contents (247 largemouth bass and 301 peacock cichlids) (Figure 4). Of these specimens, 216 largemouth bass and 200 peacock cichlids had food in their stomachs. The results for stomach contents include only those individuals with food in the stomach.


30
U Largemouth Bass

25 0 Peacock Cichlid


20


15
0.
OL 10


5





Total Length Group (mm)


Figure 4. Length-frequency for juvenile largemouth bass (N = 247) and peacock cichlids (N = 301) electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade, County, Florida.


Prey items were sorted into eight prey categories-zooplankton, amphipods,

insects, decapod crustaceans, fish, larval fish/eggs, anurans, and organic material. The zooplankton in the stomachs consisted mostly of calanoid copepods and cladocerans. The insects were mostly mayfly larvae (Ephemeroptera), chironomids (Diptera), and water boatmen (Corixidae; Hemiptera). However, odonate larvae (Odonata), water






57


striders (Gerridae; Hemiptera), and some terrestrial insects were present. A few terrestrial insects were found in largemouth bass stomachs, but most were aquatic insects. No terrestrial insects were found in peacock cichlid stomachs. All decapod crustaceans found in peacock cichlid stomachs and most in largemouth bass stomachs were grass shrimp Palaemonetes sp. Crayfish Procambarus sp. also were found in largemouth bass stomachs. Swamp darter Etheostoma fusiforme and, to a lesser extent, bluefin killifish Lucania goodei, were the dominant fish prey; poeciliids (i.e., Gambusia holbrooki and Poecilia latipinna), bullhead catfish Ameiurus sp., loricariid catfish, cichlids (including peacock cichlids), and centrarchids also were consumed. Larval fish in the stomachs were yolk-sac fry and were not identified to species. Organic materials included welldigested animals, detritus, and plant material. Ontogenetic Dietary Shifts

For largemouth bass, insects, and to a lesser extent, zooplankton and amphipods dominated frequency of occurrence for small individuals (i.e., < 70 mm TL) (Figure 5A). Zooplankton became less important and dropped out of the diet after the 80 mm TL size class. The longest largemouth bass with zooplankton was 86 mm TL. Amphipods declined in importance and were not found in fish larger than 130 mm TL. Insects remained frequent in the diet over the range of largemouth bass length examined in this study (< 170 mm TL), but were replaced by fish as the most frequently occurring prey category in specimens of 120 mm TL and larger. Decapod crustaceans occurred in only 13.7% of largemouth bass < 90 mm TL, but were commonly consumed (in 40.2% of stomachs) by largemouth bass > 90 mm TL. The smallest largemouth bass that consumed a fish was 67 mm TL. An increasing fraction of largemouth bass consumed







58


100 A

80




40
LL
20

01



Largemouth Bass Total Length Group (mm)





100

80 B




40
LL
260



20
CC






Peacock Cichlid Total Length Group (mm)
---- Zooplankton -0- Amphipod -- -A -- Insect m Decapod Fish


Figure 5. Frequency of occurrence (i.e., percent of predators with prey category in stomach contents) of main prey categories in juvenile (A) largemouth bass (N = 216) and
(B) peacock cichlids (N = 200) electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida. The horizontal line at 60 % corresponds to the frequency of fish in the diet recommended by Bettoli et al. (1992) as the minimum frequency for categorizing a predator as piscivorous.






59


fish as largemouth bass size increased. Nevertheless, the frequency of occurrence of fish did not exceed 60% (i.e., piscivory as defined by Bettoli et al. 1992) until the 120 mm TL group.

Peacock cichlids exhibited a similar pattern of diet change, although there were important differences, particularly regarding the extent of piscivory (Figure 5B). The smallest size class of peacock cichlids (i.e., 30-50 mm TL group) ate zooplankton at a high frequency. Zooplankton was frequent or moderately frequent in the diets of peacock cichlids up to the 100 mm TL group. The largest peacock cichlid with zooplankton in its stomach was 105 mm TL. Insects were eaten by a high frequency of peacock cichlids in the 60-80 mm TL groups, but rapidly declined in percent occurrence and dropped out of the diet by the 110 mm TL group. Decapod crustaceans were moderately frequent in the stomachs of peacock cichlids > 70 mm TL, and of high frequency in fish > 140 mm TL. The smallest peacock cichlid with a fish in its stomach was 54 mm TL. Fish were found in a low to moderate percent of peacock cichlid stomachs in the 30-80 mm TL groups, but rapidly increased in frequency with increasing peacock cichlid size. Piscivory (sensu Bettoli et al. 1992) was attained by peacock cichlids > 100 mm TL.

By number, diets of small largemouth bass (i.e., < 70 mm TL) were dominated by zooplankton (Figure 6A). Additionally, zooplankton made up about half of the number of prey for largemouth bass in the 70 mm TL group. Amphipods made up only a small percentage by number of the prey of largemouth bass. Insect numbers were important in the 70 mm TL group and dominated the 80-130 mm TL groups. Moderate numbers of insects also were eaten by the largest size classes of largemouth bass. Decapod crustaceans were of relatively minor importance by number for largemouth bass > 90 mm






60


TL. Fish were of low frequency by number in fish > 90 mm TL except for the 140+ mm TL groups where fish were dominant.


A 100%.

80%.
E
z 60%.

40%.

20%
(.
no/.


-


I


I


I


* b h ~ & & I I I h ** ~ i


rov~


I U U I U
N N?


Largermouth Bass Total Length Group (mm)


B 100%.

80%.0
E
z 60%40%20% 0%


7


12


-~ . .... .


,~ / /


9 V


RP oP -$ 4411 N


+ b"


Peacock Cichlid Total Length Group (mm)

El Zooplankton R Amphipod 0 Insect n Decapod U Fish

Figure 6. Percent contribution to diet by number of main prey categories in juvenile (A) largemouth bass (N = 216) and (B) peacock cichlids (N = 200) electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida.






61


Zooplankton was an important prey category by number for peacock cichlids up through the 100 mm TL size class (Figure 6B). Amphipods were relatively unimportant by number for peacock cichlids. Insects contributed most to prey numbers for peacock cichlids in the 80 mm TL group, but were relatively important by number for fish of 60110 mm TL. Decapod crustaceans were of low to moderate numbers in peacock cichlids of 70-140 mm TL, but were about 35% of the prey numbers for fish in the 140+ mm TL group. Fish were important by number for peacock cichlids beginning at the 100 mm TL group and were dominant for peacock cichlids > 110 mm TL.

Zooplankton and amphipods contributed relatively little to the volume of stomach contents of largemouth bass in this study, except for the 30-50 mm TL group (Figure 7A). Insects made up over half of the volume of this size class (i.e., 30-50 mm TL group) and generally made up at least moderate amounts of the prey volume for largemouth bass in the 60-100 mm TL groups and again in the 130 mm TL group. The decapod crustacean category was an important constituent of prey volume for largemouth bass in the 70-130 mm TL groups. Fish prey made up a small to moderate amount of the prey volume for largemouth bass in the 60-100 mm TL groups but became dominant in all larger size classes.

Zooplankton made up a substantial percent of the prey volume only for peacock cichlids 30-50 mm TL (Figure 7B). Insects contributed a relatively small amount to prey volume for peacock cichlids in the 30-50, 70, and 80 mm TL groups, and made up the largest amount of prey volume in the 60 mm TL group. Decapod crustaceans were dominant by volume in the 70 mm TL group but were of moderate to low volume in






62


larger peacock cichlids. Fish were important volumetrically for peacock cichlids 30-80 mm TL, and were dominant for larger peacock cichlids.


A 100%

80%
E

> 60%

40%

2j 20

0%



Largemouth Bass Total Length Group (mm)


E
B01


.2
0
> 6

4

0) 2
a.


0% 0% 0% 0% 0% 0%1// x


A% RP CP "N" N ll? NP M


Peacock Cichlid Total Length Group (mm)

I Zooplankton & Other R Amphipod El Insect 22Decapod U Fish

Figure 7. Volume contribution to diet of prey categories in juvenile (A) largemouth bass (N = 216) and (B) peacock cichlids (N = 200) electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida. The "Other" category includes larval fish/fish eggs, organic materials, and anurans.


rolzI






63


The Scaling of Gape Width with Fish Length

The relations between gape width and total length for both predators were: Largemouth bass GW = 0.113 TL 2.21,

2
r = 0.965, N = 87, TL range = 58-175 mm; and Peacock cichlid GW = 0.0997 TL 1.570, r2 = 0.968, N = 110, TL range = 39-165 mm (Figure 8). The ANCOVA revealed that the regression slopes were different (p < 0.0001), with largemouth bass having a slightly larger length-specific gape width. Nevertheless, there was substantial overlap of gape measurements, especially for fish below 100 mm TL (Figure 8). The distribution of gape sizes for both predators also overlapped considerably during the sampling period (Figure 9).


20 18 16

-14

E 12

10.



6

4

2

0
0


A Largemouth bass A
0 Peacock cichlid

A

A?'







0 0M


50


100
Total Length (mm)


150


200


Figure 8. Length-specific gape width measurements from juvenile largemouth bass (N = 87) and peacock cichlids (N = 110) electrofished in southeast Florida canals.











64


July 1999 Snapper Creek Canal


LMB = 18 PC = 13


B LMB IDPCI


n ni.4 fl.


2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 September 1999- Snapper Creek Canal


August 2001 Snapper Creek Canal


70
VLMB = 50 60 PC = 50
S0



30

4) 20




2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17


October 2001 Snapper Creek Canal


LMB = 19 PC = 30









SJ I .L~ l. I


70. 60.

50

U40. 30.

20. S10


2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 October 1999- Snapper Creek Canal


LMB = 8 PC = 31


L D. I


nfl I


70, 60.

50.

040.

30.

20. L
10.


2 3 4 5 6 7 6 9 10 11 12 13 14 15 16 17 November 1999- Snapper Creek Canal


LMB = 41 PC = 50










ll I LiI


2 3 4 5 6 7 0 0 10 il 12 13 14 15 16 17


August 2001 Cutler Drain Canal


LMB = 50 PC = 47


ill,


E


2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 October 2001 Cutler Drain Canal


LMB = 10 PC = 30








H o n I Ii


70.

-60.

50,

4

30. D20
LL 0


2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Gape Width Group (mm)


LMB = 50 PC = 50 n


LI


II-.L


I


2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Gape Width Group (mm)


Figure 9. Gape width frequency for juvenile largemouth bass (LMB) and peacock cichlids (PC) electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida. Sample sizes are shown for each sampling event.


70 60

50



30

20
'4
10


70.

60.

50. 2 40

30.


LL1.


70.

60. s0. U 40.
C
o 30. 20


70.

60.

50.

U 40.

30

20.
1 0






65


Prey Size and Stomach Fullness

Prey size increased with predator size for fish and decapod crustacean prey

(largemouth bass, r = 0.426, p = 0.0015; peacock cichlid, r = 0.490, p < 0.000 1) (Figure 10). Moreover, there was relatively little difference in prey size between largemouth bass and peacock cichlids on either a length-specific or gape-specific basis. Indeed, ANCOVA detected no differences in either the slope (p = 0.608) or intercept (p = 0.165) values of the prey length-predator length regressions. The largest estimates of prey length relative to predator length were 34.4% for largemouth bass and 35.4% for peacock cichlids. Prey body depth also was significantly correlated with largemouth bass total length (r = 0.437, p = 0.0024) and peacock cichlid total length (r = 0.271, p = 0.0037). The ANCOVA revealed no significant difference in slope between species (p = 0.177), but did detect a significant difference in intercept values (p = 0.0013), with largemouth bass having a higher value. Highest estimates of prey body depth relative to predator gape size were 101.4% for largemouth bass and 95.0% for peacock cichlids.

Stomach fullness estimates (Table 2) for the July, September, and October 1999 Snapper Creek Canal samples were not significantly different between predator species. However, juvenile largemouth bass had fuller stomachs than did peacock cichlids in the November 1999 sample (Wilcoxon; p = 0.0163). In 2001, largemouth bass stomachs were significantly fuller than peacock cichlid stomachs in all samples (Wilcoxon; p < 0.0001 for all four sampling date by canal combinations). There were no significant differences in stomach fullness for largemouth bass among sampling events (KruskalWallace, p = 0.0541). In contrast, stomach fullness did vary significantly for peacock cichlids (K-W, p < 0.0001). Dunn's test demonstrated that peacock cichlid stomachs







66


50 A
A

40 A Largemouth bass
4 o Peacock cichlid


30


10
>4 20OCP00 El]q AA


3 0A 0 0
0OA 0 13~

0 ~ : A A

10 40AA 0A0 A


01 0
0 20 40 60 80 100 120 140 160 180
Predator Total Length (mm)



18B 16

14- Largemouth Bass
14

12 Peacock Cichlid


10










Predator Total Length (mm)


Figure 10. (A) Prey length (mm) of decapod crustacean (folded body length) and fish (standard length) prey eaten by juvenile largemouth bass and peacock cichlids electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida. Results include 53 prey from largemouth bass and 123 prey from peacock cichlids. (B) Body depth (mm) of decapod crustacean and fish prey eaten by juvenile largemouth bass and peacock cichlids electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida. Results include 53 prey from largemouth bass and 113 prey from peacock cichlids. The diagonal lines indicate estimated maximum gape for both predators based on the lengthspecific gape width regressions developed in the present study.






67


Table 2. Percent (%) of juvenile largemouth bass and peacock cichlid stomachs within each stomach fullness category (i.e., 1 = empty, 2 = < 25% of estimated stomach capacity, 3 = 25-75% of stomach capacity, and 4 => 75% of stomach capacity) for each sampling event. Specimens were electrofished from Snapper Creek Canal (SC) and Cutler Drain Canal (CD), Miami-Dade County, Florida.


Largemouth Bass Peacock Cichlid
Stomach Fullness Stomach Fullness
Sampling Event Canal 1 2 3 4 1 2 3 4
July 1999 SC 33.3 11.1 44.4 11.1 23.1 15.4 38.5 23.1
September 1999 SC 20.0 5.0 45.0 30.0 6.7 13.3 50.0 30.0
October 1999 SC 12.5 50.0 12.5 25.0 16.1 29.0 25.8 29.0
November 1999 SC 22.2 0 22.2 55.6 24.1 41.4 27.6 6.9
August 2001 SC 8.0 26.0 32.0 34.0 33.3 41.2 17.7 7.8
October 2001 SC 12.2 14.6 51.2 22.0 29.4 56.9 9.8 3.9
August 2001 CD 12.2 38.8 34.7 14.3 50.0 37.5 6.2 6.3
October 2001 CD 6.1 49.0 28.6 16.3 56.0 40.0 4.0 0


were significantly fuller in the 1999 samples than in either canal in 2001, except that the

mean rank for Snapper Creek Canal in August 2001 was not separable from the mean

rank for Snapper Creek Canal in November 1999. Peacock cichlids in the October 2001

Cutler Drain Canal sample had the smallest mean rank stomach fullness of all the

peacock cichlid samples.

Diet Overlap

Diets of juvenile largemouth bass and peacock cichlids overlapped on all dates

(Table 3). Of the eight samples collected, four samples had a Schoener's Index value >

0.60 for prey numerical abundance. In particular, largemouth bass and peacock cichlids

overlapped considerably in volume. Five of eight overlap values exceeded 0.60 and the

other three values exceeded 0.50.






68


Table 3. Schoener's Index values for dietary overlap by number and volume between juvenile largemouth bass and peacock cichlids electrofished from two southeast Florida canals (Miami-Dade County). A "*" indicates large overlap (i.e., Cxy > 0.60). CD = Cutler Drain Canal; SC = Snapper Creek Canal.


Schoener's Index Values
Sampling Event Canal Numerical Occurrence Volume
July 1999 SC 0.102 0.957 *
September 1999 SC 0.658 0.836 *
October 1999 SC 0.499 0.522
November 1999 SC 0.012 0.873 *
August 2001 SC 0.920 0.576
October 2001 SC 0.551 0.654 *
August 2001 CD 0.932 0.586
October 2001 CD 0.895 0.785 *


Discussion

The results of the present study broadly agree with earlier investigations of diets of juvenile largemouth bass in Florida (Chew 1974; Cailteux et al. 1996; Huskey and Turnigan 2001), although there were some important differences. Because prior studies of juvenile largemouth bass diets were conducted in lakes or impoundments in central or north-central Florida (versus southeast Florida canals), differences in stomach contents could be the result of a suite of factors, including habitat type, latitude, or the presence of peacock cichlids.

Zooplankton was eaten by larger juvenile largemouth bass in the southeast Florida canals than was documented for most other Florida systems. Consumption of zooplankton in the present study occurred in largemouth bass in the 30-80 mm TL groups. Chew (1974) reported zooplankton from the stomachs of largemouth bass in the 31-75 mm TL range from two Florida lakes, but in the 76-296 mm TL range only in Lake Weir. Neither Cailteux et al. (1996) nor Huskey and Turnigan (2001) found






69


zooplankton in fish of the size ranges reported in the present study (i.e., 58 mm TL and larger).

The present study documented high consumption of insects by juvenile

largemouth bass in the canals relative to the previous studies. Insects were important prey in Lake Weir, but not in Lake Hollinsworth (Chew 1974). Likewise, insects were relatively important prey for juvenile largemouth bass in vegetated and unvegetated north-central Florida lakes (Cailteux et al. 1996), but were a comparatively small part of the diet of fish in east-central Florida lakes and impoundments (Huskey and Turnigan 2001).

Consumption of decapod crustaceans varied widely in the studies. The data for canal largemouth bass showed that decapod crustaceans were major prey for fish of 90 mm TL and larger. Chew (1974) found that decapod crustaceans were important in the diet of juvenile largemouth bass in Lake Weir, but were absent from stomach contents from Lake Hollinsworth. Cailteux et al. (1996) found that decapod crustacean were important prey in vegetated lakes but were of only minor importance in unvegetated lakes. Huskey and Turnigan (2001) reported substantial use of decapod crustaceans by largemouth bass of the sizes included in the present study.

Although nearly all size classes of largemouth bass in the present study included individuals with fish prey in their stomachs, only largemouth bass > 120 mm TL surpassed 60% frequency of occurrence, the designation of a "true" piscivore as defined by Bettoli et al. (1992; see also Cailteux et al. 1996). The study canals generally had dense beds of aquatic macrophytes along the shoreline and often throughout the deeper water column as well. Cailteux et al. (1996) found that largemouth bass in unvegetated






70


lakes became fully piscivorous (i.e., > 60% frequency of occurrence of fish in diet) by 60 mm TL, but did not become fully piscivorous until > 120 mm TL in vegetated lakes. Similar results were documented by Bettoli et al. (1992) for heavily-vegetated Lake Conroe, Texas, where largemouth bass did not become fully piscivorous until > 140 mm TL. Predatory fishes are less efficient at capturing fish prey in structurally complex habitats (Savino and Stein 1982; Hayse and Wissing 1996). Moreover, largemouth bass has been found to increase diet breadth (i.e., feeding on invertebrates and fish) in structurally complex habitats (Anderson 1984). Therefore, the late switch of juvenile largemouth bass to piscivory in the southeast Florida canals may be related to the abundant submersed vegetation.

Very little information has been published on the diet of juvenile peacock

cichlids. Lowe-McConnell (1969) reported palaemonid shrimp in 80-130 mm TL (?; type of length measurement is uncertain but assumed to be TL) peacock cichlids (N = 4) in Guyana. The reference is unclear, but fish also may have been eaten by peacock cichlids < 160 mm TL. Additionally, the author found zooplankton and filamentous algae in 40 mm TL (?) peacock cichlids. Again, the reference is unclear, but the sample size of peacock cichlids < 160 mm TL was not likely to have exceeded 15 or 20 individuals. In his investigation of an introduced population occurring in Lake Gatun, Panama, Zaret (1 980a) examined stomach contents of seven Cichla specimens (either C. ocellaris or a congener) ranging in size from about 55 to 90 mm SL (or about 70 to 112 mm TL using the conversion equation of Hill [1998]). The stomachs contained terrestrial insects, palaemonid shrimp, and fish. In Puerto Rico, introduced peacock cichlids < 150 mm TL ate fish and aquatic insects (Lilyestrom and Churchill 1996). In the only






71


published study from Florida, Shafland (1999b) documented fish of several species, but no non-fish prey, from peacock cichlids 100-149 mm TL in Tamiami Canal. The present study was in general agreement with previous studies, but documented consumption of a much wider variety of prey.

The present study documented considerable overlap in gape width between

juvenile largemouth bass and peacock cichlids of similar sizes. This finding agrees with Hill (1998), who reported broad similarity in gape size of these species over a wide range of lengths (i.e., largemouth bass 17-423 mm TL; peacock cichlids 50-430 mm TL). Given the extent of temporal and spatial overlap in total length distributions found in this study, juvenile largemouth bass and peacock cichlids of similar size and morphological prey-handling ability co-occurred in both canals. Morphological factors, coupled with syntopy, implied that juvenile largemouth bass and peacock cichlids had a high potential for undergoing similar dietary shifts and therefore sharing common food resources.

As suggested by morphology, the ontogenetic prey-use pattern of largemouth bass and peacock cichlids was similar (i.e., zooplankton, followed by insects, then decapod crustaceans and fish). The main dietary differences were in a slightly broader range of prey eaten by largemouth bass (e.g., inclusion of crayfish in addition to grass shrimp in the decapod crustacean category and inclusion of anurans in the diet) and in the timing (i.e., total length) of shifts in prey proportions. For example, peacock cichlids consumed zooplankton at larger sizes than did largemouth bass (i.e., through the 100 mm TL group versus the 80 mm TL group for largemouth bass). Furthermore, amphipods and insects







72


100. 80. 6040. 204 0.


Figure 11. Frequency of occurrence (% of predator stomachs with the prey type) for largemouth bass (N = 113; 100-349 mm TL) and peacock cichlids (N = 156; 100-348 mm TL) sampled from Tamiami Canal, Miami-Dade County, Florida. Data from Shafland 1999b.


E
M



.


100%. 80%q 60%. 40% m 20%4


Largemouth Bass


Ej nsect
P Decapod
N Fish








Peacock Cichlid


Figure 12. Percent (%) by number of prey for largemouth bass (N = 113; 100-349 mm TL) and peacock cichlids (N = 156; 100-348 mm TL) sampled from Tamiami Canal, Miami-Dade County, Florida. Data from Shafland 1999b.



were generally more important for largemouth bass than peacock cichlids, although certain peacock cichlid size classes included these in high numbers and frequency. Stomach contents data from another southeast Florida canal also show a similar pattern of


0Dnsect E2 Decapod
0 Fish


Peacock Cichlid


U 0L


Largemouth Bass


W__






73


consumption of decapod crustaceans and fish by largemouth bass and peacock cichlids of 100 to about 350 mm TL (Shafland 1999b; Figures 11 and 12).

Initiation of piscivory is an important ontogenetic event for predatory fishes

(Aggus and Elliott 1975; Pasch 1975; Ludsin and DeVries 1997). Fish were included in the peacock cichlid diet at a generally higher frequency, number, and volume than in the largemouth bass diet at all predator sizes and peacock cichlids may have started to eat fish at a shorter total length than did largemouth bass. As a result, peacock cichlids became piscivorous (sensu Bettoli et al. 1992) before largemouth bass (i.e., by the 100 mm TL group versus the 120 mm TL group). These trends were even more pronounced based on numeric and volumetric proportions of fish in the diets of juvenile predators (e.g., for predator size classes > 100 mm TL, largemouth bass averaged [mean 1 SE] 56.0 11.9% fish by volume versus 88.1 1.7% for peacock cichlids).

The percentage of empty stomachs was high for peacock cichlids and much lower for largemouth bass. This translated into significantly higher stomach fullness estimates for largemouth bass on several sampling dates. These data agree with the results of small and large tank studies, which documented lower feeding rates for peacock cichlids than for largemouth bass (Chapter 5). Relatively low daily rations for a congener, Cichla monoculus, also have been reported (Rabelo and Araujo-Lima 2002). Surprisingly, however, juvenile peacock cichlids may grow faster than juvenile largemouth bass. There was substantial overlap in the total lengths of juveniles throughout the sampling period. Because largemouth bass spawn earlier than peacock cichlids (Shafland 1999b), the overlap in length frequencies in the present study suggests faster growth of age-0 peacock cichlids. Unfortunately, there are no studies available that directly measured






74


growth rates of juvenile largemouth bass and peacock cichlids. If lower stomach fullness are due to lower feeding rates, then these results suggest that peacock cichlids are more efficient than largemouth bass (i.e., that despite lower consumption rates they are able to grow faster). Because a shift to a fish diet has been linked to increases in growth rates in piscivorous fishes (Pasch 1975; Timmons et al. 1980; Olson 1996), early and substantial consumption of fish prey by juvenile peacock cichlids may have resulted in their apparent faster growth rates relative to largemouth bass.

Not only were the stomach contents of juvenile largemouth bass and peacock cichlids similar in taxonomic identity, they also were similar in relative prey size. For both species, increasing prey size was weakly correlated with increased predator length. The only statistically significant difference in prey size between the species being marginally larger (i.e., deeper-bodied) prey in largemouth bass. This result may be due to the fact that a largemouth bass has a slightly larger gape width than a peacock cichlid of the same total length.

Competition for food can be an important interaction for juvenile largemouth

bass. For example, sunfishes (i.e., Lepomis spp.) and juvenile largemouth bass have been shown to compete for food in the littoral zones of temperate lakes, possibly limiting largemouth bass recruitment (Osenberg et al. 1994; Olson et al. 1995). Likewise, growth and survival of largemouth bass juveniles is density-dependent (Olson et al. 1995; Garvey et al. 2000). The presumed mechanism is intraspecific food competition. The results of these studies suggest that food may be limiting for juvenile largemouth bass under some circumstances and that the presence of fishes with similar food requirements and habitat use may be detrimental to largemouth bass abundance.






75


Species with similar ecological requirements which consume resources from the same habitat should have higher competition coefficients than a species pair which feeds in different habitats or consumes different resources within a habitat (Rosenzweig 1981). Juvenile largemouth bass and peacock cichlids overlapped considerably in diet and such large overlap index estimates have been considered significant (Zaret and Rand 1971; Mathur 1977). Nevertheless, diet overlap indices by themselves do not necessarily demonstrate competition (Colwell and Futuyma 1971; Hulbert 1978). For example, high overlap can occur without competition when resources are abundant. Rock-dwelling cichlids (i.e., "mbuna"; Cichlidae) in Lake Malawi, Africa, exhibit extreme trophic differentiation based on differences in tooth structure, mouth morphology (e.g., size and orientation), and pharyngeal jaw apparatus (Fryer and Iles 1972). Normally, most of these species are herbivorous, grazing on aufwuchs (i.e., bio-cover of algae and bacteria) in specific microhabitats (Ribbink 1991). Nevertheless, during times of high zooplankton abundance, mbuna will abandon their trophic specialties and feed on superabundant zooplankton (McKaye and Marsh 1983; Ribbink et al. 1983). Dietary overlap is very high during these periods, but competition is not. On the other hand, Rosenzweig (1981) modeled two-species competition and concluded that a competition coefficient of zero (implying little or no resource overlap) was possible, even between two strong competitors, due to competitor-induced habitat shifts. Indeed, low overlap in diet could be the consequence of competition. Diets of native lake trout Salvelinus namaycush significantly differ from diets of introduced smallmouth bass Micropterus dolomieu and rock bass Ambloplites rupestris in Canadian lakes (Vander Zanden et al. 1999). In these lakes, smallmouth bass and rock bass eat littoral prey fish and lake trout feed primarily on






76


zooplankton and include fish prey as only a small part of the diet. However, in similar lakes without introduced smallmouth bass or rock bass, lake trout diets contain a considerably higher percentage of littoral prey fish. These results imply a competitioninduced diet shift.

Indeed, high dietary overlap values have been used both to refute and to

demonstrate the existence of interspecific competition in fishes (Colwell and Futuyma 1971; Hulbert 1978). In some studies, high dietary overlap between two fishes is interpreted as evidence that food resources are not limiting (i.e., therefore competition is not occurring) or to explain observed interspecific habitat or activity period differences (e.g., Zaret and Rand 1971; Keast 1978, 1985; Keast et a]. 1978; George and Hadley 1979; Winemiller 1989; Hirst and DeVries 1994). Much of this research is based on morphologically similar species, feeding guilds, or entire species assemblages that are native to the study area, apparently have coevolved, and presumably have reacted to past competition through character displacement or resource partitioning. In other studies, particularly investigations of nonindigenous fishes, dietary overlap is used as evidence of food competition or of the high potential for food competition to occur (e.g., Mathur 1977; Parrish and Margraf 1990; Zale and Gregory 1990; Scoppettone 1993; Ogle et al. 1995; Declerck et al. 2002).

Dietary data from southeast Florida canals show high overlap, and previous studies in temperate lakes demonstrate that largemouth bass are food-limited. Unfortunately, there are no good tests of food limitation in southeast Florida canals, although there are studies of the abundances of small fishes upon which largemouth bass and peacock cichlids feed. During the past two decades (i.e., since the introduction of






77


peacock cichlids in 1984) qualitative and quantitative surveys of fishes in the canals where largemouth bass and peacock cichlids co-occur have found high numbers of small fishes, both native and nonindigenous (boat electrofishing surveys: L. G. Nico, USGS, unpubl. data; pers. obs.; blocknet sampling: Shafland et al. 1985; Shafland 1999a; Shafland et al. 2001). These high densities suggest that small fish availability does not limit predators such as largemouth bass and peacock cichlids. However, dynamics of invertebrate prey, which may influence early growth patterns of largemouth bass and peacock cichlids, have not been investigated in these canals. Finally, if food was limiting, then largemouth bass population densities should have declined following the introduction of peacock cichlids, a potential competitor. Nevertheless, available data do not suggest that largemouth bass populations have declined during the 20-year period following the introduction of peacock cichlids (Shafland 1995, 1999a, c; Shafland and Stanford 1999; Shafland et al. 2001; Chapter 6).

Fishes that are piscivorous as adults typically undergo marked changes in diet related to increases in body length and mouth size (Keast 1985, Keast and Eadie 1985; Olson 1996; Mittlebach and Persson 1998; present study). Knowledge of the extent to which ecomorphological similarity is predictive of the timing (i.e., predator size) and prey progression of such shifts can be important in evaluating the risks associated with fish introductions, whether planned or illegal. The results of the present study suggest that ecomorphological similarity to a native fish is predictive of dietary ontogeny and dietary overlap for nonindigenous species. This implies that such information can be used for initial assessments of risk prior to fish introductions in the absence of diet data for early life history stages of native or introduced populations of the nonindigenous






78


species. Studies of nonindigenous fishes typically concentrate on adult stages. However, careful attention also must be paid to early life stages and the possible interactions of such stages with native fishes for those species that differ in habitat and food use through ontogeny (Courtenay 1995). Given the ecomorphological convergence of many cichlids and centrarchids (Norton and Brainerd 1993), using similarity to predict food resource use may be particularly valuable when cichlids invade systems where centrarchids are native (e.g., Florida) or when centrarchids invade systems in Africa, Central America, or South America.















CHAPTER 5
EXPERIMENTAL PREY HANDLING AND SELECTIVITY OF TWO
MORPHOLOGICALLY SIMILAR PREDATORY FISHESNATIVE LARGEMOUTH BASS AND INTRODUCED PEACOCK CICHLIDS Introduction

The morphology of predatory fishes is known to have a profound effect on their prey use. For example, gape limitation is an important constraint for many predatory fishes. The size of the mouth, pharyngeal, or intercleithral opening (i.e., gape size) places limits on maximum prey size and influences the energetic return from submaximal prey (Lawrence 1958; Werner 1974, 1977, 1979; Zaret 1980b; Wainwright 1988; Wainwright and Richard 1995). Furthermore, patterns of prey selection frequently scale with patterns of gape size (Wainwright and Richard 1995; Mittelbach and Persson 1998). The constraint of gape limitation often results in an ontogenetic dietary progression characterized by the initial consumption of zooplankton, then of intermediate-sized invertebrates, and finally by fish (Keast 1985; Wainwright and Richard 1995).

Given this relationship of gape size to prey use, a reasonable prediction is that an introduced species with gape size (and other features of morphology) similar to a native species will also exhibit similarities in prey handling ability and prey selection. Additionally, such similarities suggest that the nonindigenous species will exert comparable pressures on prey species. These predictions are of general concern following a fish introduction because they imply the potential for food competition


79






80


between the invader and any ecomorphologically similar native species. Indeed, competition is the most often cited mechanism whereby introduced fishes negatively effect native fishes (Moyle et al. 1986).

The native largemouth bass (Centrarchidae) is an economically important

predatory fish in southern Florida (Shafland and Stanford 1999). The peacock cichlid Cichla ocellaris (Cichlidae), a species native to South America, was legally introduced into freshwater canals in southeast Florida in 1984-1987 by the Florida Game and Fresh Water Fish Commission (currently Florida Fish and Wildlife Conservation Commission) (Shafland 1995). The predatory peacock cichlid was stocked to reduce the abundance of other nonindigenous fishes, particularly spotted tilapia Tilapia mariae, and to augment recreational fisheries that primarily targeted largemouth bass (Shafland 1995). Largemouth bass and peacock cichlids are ecomorphologically similar based on appearance, body size, gape size, feeding mechanics, and diet (Norton and Brainerd 1993; Hill 1998; Chapter 4). Dietary overlap of juveniles in southeast Florida canals is generally large (Chapter 4) and adults of both species are regarded as piscivorous (Carlander 1977; Shafland 1995, 1999b). Additionally, experimental studies of largemouth bass prey selection have demonstrated that fish are often selected at a higher rate than non-fish prey (e.g., Lewis et al. 1961). There was considerable overlap in diets of subadult and adult largemouth bass and peacock cichlids in Tamiami Canal (Shafland 1999b). Nevertheless, largemouth bass consumed substantial amounts of non-fish prey (54 % by number; primarily decapod crustaceans, but also insects) in addition to fish whereas co-occurring peacock cichlids were almost exclusively piscivorous (Shafland 1999b).






81


These results suggest that ecomorphological similarity can predict coarse patterns but is not precise enough to predict patterns of a finer scale. Alternately, some other mechanism (e.g., differential encounter rates mediated by habitat, predator behavior, or competition) may be operating to modify the prey use patterns of one or both species. Prey selection by predators results from the interaction of differential encounter rates with prey, probabilities that a predator will attack a prey item if encountered, and probabilities that an attack results in consumption (O'Brien 1979; Osenberg and Mittelbach 1989). Encounter rate is fundamental to the process because it sets the initial conditions from which the remaining predation components operate (Osenberg and Mittelbach 1989). For example, rarely encountered prey may represent a small portion of a predator's diet even if the product of attack and success probabilities approaches unity (i.e., 1.0). Encounter rates, especially under field conditions, are difficult to estimate (Osenberg and Mittelbach 1989). Moreover, structurally complex habitat can greatly reduce encounter rates and feeding efficiency of predators (Savino and Stein 1982; Anderson 1984; Hayse and Wissing 1996). Competition between predators also can reduce encounter rates with prey by prey depletion, by modifying prey behavior in response to predation risk, or by direct interference with predator feeding behavior (Hobson 1979; Mittelbach 1984; Fausch et al. 1997; Turner et al. 1999). Alternately, species-specific hunting behavior of predators may influence the rate of encounter with different prey types (Cooper et al. 1985; Magnhagen 1986; Grant and Noakes 1987). However, by creating situations where encounter rates are similar between two predators (e.g., by making encounters almost certain), attack and success probabilities can be estimated and comparisons can be made.






82


I conducted a series of tank experiments with largemouth bass and peacock cichlids to test predictions concerning prey use by similar nonindigenous and native fishes (i.e., similar morphology equals similar prey handling ability, prey selection, and thus potential effects on prey base). I used estimates of prey handling times to test for differences in the ability of both predators to handle prey. For example, prey handling and the resulting information on prey profitability are important for predictions of prey selection based on optimal diet models (Emlen 1966; MacArthur and Pianka 1966). This experiment also provided information on the probability of attack and successful consumption of prey for both predators within the experimental system. This information was used to formulate and test predictions of prey selection for largemouth bass and peacock cichlids given choices among prey types.

Communities often have multiple species of predators (Sih et al. 1998) and a

number of studies have investigated the equivalence of multiple predators in their effect on prey assemblages (Harris 1995; Morin 1995; Kurzava and Morin 1998; Baber and Babbitt 2003). The introduction of peacock cichlids into southeast Florida canals represents the addition of a predator to a predator assemblage formerly dominated by a single species, the largemouth bass (Chapter 2). Therefore, I tested assemblages made up of varying densities and percent composition of largemouth bass and peacock cichlids for predator assemblage effects on prey selection and prey mortality.

Methods

Experiment 1 Prey Handling

I conducted a laboratory experiment to test the relative ability of largemouth bass and peacock cichlids to capture and consume fish and non-fish prey. This experiment






83


was conducted in the University of Florida Demonstration Aquaculture Recirculating System (see Hill and Cichra in prep. for a complete description). This indoor system has

1.1 -m diameter by 0.6-m high, circular fiberglass tanks connected to biofilters. Tank interiors were light blue and had single airstones and 50-mm central standpipes. Water temperature was maintained between 25.5" and 270 C and the light:dark cycle was 13 h: 11 h. Eight largemouth bass were electrofished from Lake Alice, Alachua County, Florida and eight peacock cichlids were electrofished from Cutler Drain Canal, MiamiDade County, Florida. The largemouth bass averaged 322 mm total length (TL) (255408 mm TL) and the peacock cichlids averaged 340 mm TL (251-407 mm TL). Prey were collected from ponds at the Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville, Florida.

Prey handling times, attack probability, and probability of success given an attack were estimated for single prey items presented to individual predators. Prey used in the experiment were of three representative prey types-1) a soft-rayed fish (i.e., sailfin molly Poecilia latipinna), 2) a spiny-finned fish (i.e., bluegill Lepomis macrochirus), and 3) a decapod crustacean (i.e., crayfish [a mix of Procambarus fallax and P. paeninsulanus]). Each prey type is present in southeast Florida. Prey sizes were chosen to encompass a range of about 20-100% of predator gape size (i.e., prey body depth divided by predator gape width and multiplied by 100) based on relations for bluegill, largemouth bass, and peacock cichlid found in Hill (1998) and for crayfish and sailfin molly developed in the present study. Length-specific body depth relations for crayfish and sailfin molly were developed from fresh specimens (PROC GLM; SAS Institute, Cary, North Carolina):






84


Crayfish: BD = 0.339 TL 2.841; r2 = 0.90 Sailfin Molly: BD = 0.318 TL 3.890; r2 = 0.98 where BD = maximum body depth. Handling time was measured as the total time from initial prey contact to the cessation of all feeding movements of mouth and opercula, including time spent recapturing released prey. In the small tanks, initial searching and pursuit times were essentially zero. Prey mass was estimated using length-mass regressions developed from fresh specimens (PROC GLM; SAS Institute, Cary, North Carolina).

Bluegill: LogioWT = 2.790 LogioTL 4.525; r = 0.95 Crayfish: LogioWT = 3.498 LogioTL 5.472; r2 = 0.97

Sailfin Molly: LogioWT = 3.388 LogioTL 5.436; r2 = 0.99

Energy values for prey were taken from the literature-I 160 cal/g for centrarchids (Miranda and Muncy 1989), 1100 cal/g for other fish (Pope et al. 2001), and 750 cal/g for crayfish (Pope et al. 2001) for wet weights.

I incorporated effects of prey size and predator size by dividing prey body depth by predator gape (and multiplied by 100), which I called relative prey size, and then categorized relative prey size into 10% size classes. Two-way analysis of variance (ANOVA) was used to detect differences in mean prey profitability across relative prey size classes for sailfin molly and bluegill prey using predator species and relative prey size class as main effects (PROC GLM; SAS Institute, Cary, North Carolina). Nonsignificant interaction terms (i.e., predator x size class) were dropped from the model (bluegill interaction F5,54 = 1.21, p = 0.319; sailfin molly interaction F3,55 = 1.55, p =

0.212). One-way ANOVA was used for crayfish prey because peacock cichlids did not






85


eat crayfish. ANOVAs were followed by the least squares means multiple comparison procedure. ANOVA (followed by Tukey's HSD multiple comparison procedure) was also used to test for the effect of prey size and type on the probability of successful consumption of prey for both predators given that an attack had taken place (PROC GLM; SAS Institute, Cary, North Carolina). Two-sample t-tests were conducted to investigate differences between predators in the probability of successfully consuming each prey type (PROC TTEST; SAS Institute, Cary, North Carolina). Probabilities were arcsine-square root transformed prior to analysis. All tests were conducted at a Type I error rate (a) of 0.05.

Experiment 2 Prey Selection by Individual Predators

Prey selection patterns were determined empirically in the same laboratory tanks that were used in Experiment 1. A single predator per tank was tested and eight largemouth bass and eight peacock cichlids were tested. Eight largemouth bass were electrofished from Lake Alice, Alachua County, Florida and eight peacock cichlids were electrofished from Cutler Drain Canal, Miami-Dade County, Florida. The largemouth bass averaged 322 mm total length (TL) (255-408 mm TL) and the peacock cichlids averaged 340 mm TL (251-407 mm TL). Prey were collected from ponds at the Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville, Florida. Predators were fasted the day prior to each experiment. Predators not being tested were fed crayfish, eastern mosquitofish Gambusia holbrooki, sailfin mollies, bluegills, and redear sunfish Lepomis microlophus. Prey species selectivity was tested using three representative prey types-a soft-rayed prey (i.e., sailfin molly), a spinyfinned prey (i.e., bluegill), and a decapod crustacean (i.e., crayfish). Five of each prey






86


type were presented simultaneously to each predator (i.e., total of 15 prey per trial). Prey body depths were 40-60% of predator gape size. Remaining prey were retrieved and counted after one day (i.e., about 24 h).

Prey selectivities for largemouth bass and peacock cichlids feeding on fish and crayfish prey were calculated using the Manly-Chesson index (Manly 1974; Chesson 1978, 1983), accounting for prey depletion during the experiment (i.e., Case 2; Chesson 1983):

ln((n0 -rO)/njO) i~ l,... m

Iln((njo -r )njo)
j=1

where m is the number of prey categories, rio is the number of prey type i in the diet and nio is the number of prey i in the environment. Values of a range from 0 to 1, with 0 being avoidance, 1 being complete preference, and 1/m being random selection. Oneway ANOVA (followed by the Tukey HSD multiple comparison procedure) was used to test the hypothesis of random prey selection for each predator (i.e., all prey categories equally selected) (PROC GLM; SAS Institute, Cary, North Carolina). Two-sample t-tests were used to determine if prey selectivity (i.e., a) differed between largemouth bass and peacock cichlids for sailfin molly and bluegill prey (PROC TTEST; SAS Institute, Cary, North Carolina). The Wilcoxon signed rank test was used to test for differences in mean numbers of prey eaten per predator between largemouth bass and peacock cichlids (PROC NPAR1 WAY; SAS Institute, Cary, North Carolina). A significance level of a =

0.05 was used for all statistical analyses.

Data on probability of attack and probability of successful consumption from Experiment 1 were used to predict prey selection for largemouth bass and peacock






87


cichlids in Experiment 2, assuming that all available prey were encountered (given the small size of the tanks), and subject to the constraint that the total number of prey eaten was equal to that observed. Chi-square tests were used to test the hypothesis that the predicted diet was not different from the observed diet for each predator species (PROC FREQ; SAS Institute, Cary, North Carolina).

To determine if decapod crustaceans are acceptable prey for subadult and adult peacock cichlids given the lack of consumption of crayfish in the experiments, I stocked 10 crayfish into tanks with individual peacock cichlids and offered no alternative prey for five days. This experiment was conducted at the Florida-Caribbean Science Center, U. S. Geological Survey, Gainesville, Florida in indoor, 1.2-m diameter by 0.6-m high, circular fiberglass tanks. Tank interior color was light blue and the interior was bare except for a single airstone and a central 10-cm standpipe. There was a constant, slow flow of well water through each tank and water temperature was maintained at about 26.70 C by the use of two 300-w heaters. Peacock cichlids were collected by electrofishing from Cutler Drain Canal and Snapper Creek Canal in Miami-Dade County, Florida and crayfish were trapped from ponds at the Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville, Florida. A total of five peacock cichlids ranging from 196 to 355 mm TL were tested. Crayfish body depths were smaller than predator gape width (Hill 1998).

Experiment 3 Prey Selection by Predator Assemblages

I conducted an experiment at the Sam Mitchell Aquaculture Demonstration Farm (SMADF), University of Florida, Blountstown, Florida, to test for differences in the effect of predation by assemblages of largemouth bass and peacock cichlids on prey






88


assemblages consisting of three representative prey types. The experimental tank system consisted of 6-m diameter by 1.2-m high circular fiberglass tanks with a central 10-cm, screened standpipe, several 15-cm airstones around the tank perimeter, and a constant flow of pond water from a single, 0.4-ha, aerated pond. Tank interior color was offwhite. Water in each tank was maintained at a depth of about 45 cm and had a temperature of 26 290 C during the experiment. Morning dissolved oxygen readings in the tanks exceeded 8.0 mg/L. Water clarity (i.e., vertical secchi depth) in the pond exceeded pond depth (> 2.4 m) and the mean value in the tanks for color was 8.82 Pt-Co (SD = 1.29) and for turbidity was 0.32 NTU (SD = 0.18). The tanks were outdoors under a 20% shade cloth covering.

Largemouth bass and peacock cichlids were used as predators and golden shiners Notemigonus crysoleucas, redear sunfish, and crayfish were used as prey. Largemouth bass were angled from ponds at SMADF and the Department of Fisheries and Aquatic Sciences (FAS), University of Florida, Gainesville, Florida. Peacock cichlids were electrofished from Cutler Drain Canal, Miami-Dade County, Florida. The largemouth bass averaged 340 mm TL (299-376 mm TL) and the peacock cichlids averaged 324 mm TL (290-365 mm TL). Golden shiners and redear sunfish were purchased from bait and fingerling suppliers. Crayfish were trapped from ponds at FAS. It is unlikely that any prey had experience with large predatory fish. Prey species were chosen to represent distinct prey types-1) an elongate, soft-rayed fish (i.e., golden shiner), 2) laterallycompressed, spiny-finned fish (i.e., redear sunfish), and 3) a decapod crustacean (i.e., crayfish). These species are present in southeast Florida canals. Prey sizes used were smaller than predator gape size according to length-specific gape width relations from






89


Hill (1998). Prey were chosen to be similar in body depth (mean BD = 16-18.5 mm), the prey dimension of most importance to gape-limited predators (Lawrence 1958; Hambright 1991). Mean prey length and weight were compared by one-way ANOVA followed by the Tukey HSD multiple comparison test (PROC GLM; SAS Institute, Cary, North Carolina). Golden shiners were longer than crayfish and redear sunfish (ANOVA, F2,78 = 69.31, p < 0.0001) and heavier than redear sunfish (ANOVA, F2,78 = 38.03, p < 0.0001). Crayfish and redear sunfish were of similar length, but crayfish were heavier than golden shiners and redear sunfish (ANOVA, F2,78 = 38.03, p < 0.0001).

Nine predator treatments were tested with an equal initial food resource level of 50 each of golden shiners, redear sunfish, and crayfish. Predator density was four, six, or eight predators (i.e., low, medium and high predator density). Within each density level, the species composition of predators was 100% largemouth bass, 100% peacock cichlid, or 50% of each species, giving nine treatment combinations (i.e., 3 predator densities x 3 species combinations = 9). A tenth treatment, zero predators, was used to estimate losses not due to predation, but was not included in the analyses because no losses were observed in this treatment. Treatments were randomly assigned to tanks and each treatment was run three times. Trials were run for three days. Tanks were inspected at least twice daily for prey mortality. Missing prey were not replaced. At the end of the trial, the tanks were drained and prey were removed and counted. Predator densities, initial prey numbers, and trial duration were chosen to provide a range of resource availability from abundant (i.e., the most selected prey would not be eradicated) to scarce (i.e., highly selected prey will be eradicated and alternative prey should be exploited).






90


Prey selectivity values for golden shiners, redear sunfish, and crayfish were

estimated for each predator density x combination treatment using the Manly-Chesson index (i.e., 6 Case 2, with depletion) (Chesson 1983). Values were arcsine-square roottransformed prior to analysis. Two-way ANOVA, followed by the least squares means multiple comparison procedure was used to test for the effects of predator assemblage characteristics (i.e., predator density, combination, and density x combination interaction) on mean a values (PROC GLM; SAS Institute, Cary, North Carolina). Non-significant interactions were excluded from the model and ANOVA was repeated with only the main effects.

In addition to examining prey selectivity (which is a measure of the relative mortality imposed on prey types), I also compared the absolute mortality imposed on prey. I used two-way ANOVA to test the effect of predator density and combination on the logio-transformed odds of a predator eating a prey during the trial versus leaving it alive at its end, followed by the least squares means multiple comparison procedure (PROC GLM; SAS Institute, Cary, North Carolina). Non-significant interactions (i.e., predator density x predator combination) were excluded from the model and ANOVA was repeated with only the main effects. One-way ANOVA was used to test for differences in mean per capita consumption of crayfish by largemouth bass (PROC GLM; SAS Institute, Cary, North Carolina). Type I error rate (a) was set at 0.05.

Results

Experiment 1 Prey Handling

In general, handling times increased with prey size but varied between largemouth bass and peacock cichlids and among prey types; however, there was considerable






91


variation in these relationships leading to pronounced variation in prey profitability. Sailfin mollies were more energetically favorable prey for peacock cichlids than for largemouth bass (ANOVA, F1,58= 16.48, p = 0.0001; Figure 13A). Nevertheless, sailfin mollies were of high value relative to other prey for largemouth bass (Figures 13A and 13B). Additionally, the relative size of sailfin mollies did not have a significant effect on their profitability (ANOVA, F4,58 = 0.43, p = 0.785). Contrary to sailfin molly prey, profitability of bluegill prey did not depend on predator species (ANOVA, F1,59= 2.33, p = 0.1322), but was influenced by relative prey size (ANOVA, F5.59= 8.14, p < 0.0001; Figure 13B). The smallest size classes of bluegill (i.e., 40 and 50%) were more profitable than were larger bluegills. Only largemouth bass ate crayfish during this experiment. The relative size of crayfish prey did not have a significant effect on crayfish profitability for largemouth bass (ANOVA, F3,43 = 2.50, p = 0.072; Figure 13B); however, crayfish in the larger than 70 % of gape size were offered to largemouth bass but were not eaten.

Largemouth bass and peacock cichlids always attacked fish prey (i.e., sailfin mollies and bluegills) placed into the experimental tanks (Table 4). Largemouth bass usually attacked crayfish, but peacock cichlids never attacked crayfish. Additionally, successful consumption of attacked prey varied among prey types. Largemouth bass were more successful at consuming fish than crayfish (ANOVA, F2,178 = 9.33, p = 0.0001) and peacock cichlids were more successful with sailfin mollies than with bluegills (ANOVA, F1,48 = 7.68, p = 0.008). Prey size did not influence predator success (given that an attack occurred) for largemouth bass eating sailfin mollies (ANOVA, F5,42 = 0.95, p = 0.460) or bluegills (ANOVA, F5,54 = 0.35, p = 0.877) or for peacock cichlids







92


40
Relative


50 60 70 80
Prey Size (Prey BD/Predator GW 100)


40 50 60 70 80
Relative Prey Size (Prey BD/Predator GW 100)


>90


>90


Figure 13. Estimated mean profitability ( I SE) to largemouth bass (LMB) and peacock cichlids (PC) of sailfin molly (SFM) (A), bluegill (BG) (B), crayfish (CF) (B) versus relative prey size. BD = body depth; GW = gape width.


800070006000I
5000

4000C 3000Q 20001000

0.


A


<40


B


2000



150 ~100'CL

500

C
(U


-A- LMB-BG

-0-- LMB-CF

--- PC-BG


U.


< 40


e i


9


+M LMB-SFM

-[a- PC-SFM




Full Text

PAGE 1

EFFECTS OF INTRODUCED PEACOCK CICHLIDS CICHLA OCELLARIS ON NATIVE LARGEMOUTH BASS MICROPTERUS SALMOIDES IN SOUTHEAST FLORIDA By JEFFREY E. HILL 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 2003

PAGE 2

ACKNOWLEDGMENTS Many people provided tremendous assistance in the completion of this project. All deserve my acknowledgment. I apologize beforehand to anyone inadvertently omitted. The primary acknowledgment goes to my wife, Susan, for her unfailing support throughout my entire graduate career. I thank her for sacrificing in order for me to fulfill our shared goal. 1 especially thank her for putting up with my obsession for fish. I am gratefial for the constant support of my family — ^my father, mother, and sister (Baker, Jacqueline, and Kim). They shared the dream of my doctorate — I am thankfial to have fulfilled our collective aspiration. I greatly appreciate the guidance and support of my doctoral committee — Drs. Charles E. Cichra (Chair), Carter R. Gilbert, William J. Lindberg, Leo G. Nico, and Craig W. Osenberg. It was a great pleasure to work for Dr. Cichra — the experience I received in extension, research, and teaching, along with strong mentorship and his fiiendship, were instrumental in my professional development. Carter Gilbert has been a tremendous influence and is one of my real "fish heroes". I especially thank Carter for staying involved in my graduate education after his retirement — that meant a lot to me. Leo Nico's remarkable field experience with nonindigenous fishes and south Florida systems was invaluable. The many discussions (and some disagreements) over nonindigenous fishes with Leo helped shape my scientific philosophy. I appreciate the support and ii

PAGE 3

encouragement that Bill Lindberg always gave. Our discussions inside and outside of the classroom about ecology and the philosophy of science (e.g., tautology!) helped open my eyes to the wider world of science. I thank Craig Osenberg for agreeing to help out on the committee of someone studying a crazy system like the south Florida canals. I especially appreciate his assistance with preparing for my qualifying exam and his extraordinary input during the revision process. I have learned a great deal from each committee member and thank them for being a part of my graduate education. A special acknowledgment is due to Jeff Sowards for all of the hard work in the field. Thirty-six hour "days" can be rough, but he never complained. Jeff was instrumental in the field work, in the lab, and as an "ideas man" to keep the project doable. I could not have done it without his help. Also, Sharon Fitz-Coy was a tremendous help in, among other things, collecting prey for experiments and identifying insects in fish stomach contents. She was always willing to help. Special thanks are due to Paul Shafland (Florida Fish and Wildlife Conservation Commission; FWC) for a great deal of assistance, facilitation, literature, and advice during this project. Paul's largest contributions, however, came during many discussions of scientific philosophy regarding nonindigenous fish. Craig Watson and Roy Yanong (Tropical Aquaculture Laboratory) provided exceptional support throughout the process. They made the tribulations more bearable. Debbie Pouder facilitated all work done at the Sam Mitchell Aquaculture Demonstration Farm (SMADF) in Blountstown, Florida. I appreciate her help and that of the staff, especially Randall Kent. iii

PAGE 4

Ken Portier of the Department of Statistics at the University of Florida helped with the design and analysis of experiments in Chapter 5. I appreciate the hospitality of Don and Judy Bemecker who provided a great place to stay and many delicious meals during field work. Many others are due thanks for their contributions. At the University of Florida, Department of Fisheries and Aquatic Sciences, Mike Allen, Mary Cichra, Doug CoUe, Ruth Francis-Floyd, Tom Glancy, Scott Graves, Robert Leonard, Allen Riggs, Beth Sargent, Troy Thompson, Larry Tolbert, and Dan VanGenechten helped with various aspects of the project. Kelly Jacoby drew the maps. Susan Morgan greatly facilitated the graduation process. At the Florida Caribbean Science Center, U.S. Geological Survey, Gary Hill, Howard Jelks, and Bill Stranghoener assisted with facilities and equipment. Frank Morello, Jerry Krummrich, and Fred Cross (FWC) issued scientific collecting permits; Ted Hoehn (FWC) kindly provided access and instruction to the FWC fish database; Eric Nagid (FWC) helped answer questions about use and content of the database. I gratefully acknowledge the financial and facilities assistance of the Department of Fisheries and Aquatic Sciences, the Sam Mitchell Aquaculture Demonstration Farm, and the Tropical Aquaculture Laboratory of the University of Florida. Financial assistance was further provided by a Dean's Fellowship for Graduate Research from the Institute of Food and Agricultural Sciences, University of Florida, the Florida Department of Agriculture and Consumer Services, by Ryan Kelley Memorial Scholarships from the International Women's Fishing Association, and the Roger Rottmann Memorial Scholarship from the Florida Chapter, American Fisheries Society. iv

PAGE 5

TABLE OF CONTENTS gage ACKNOWLEDGMENTS ii ABSTRACT vii CHAPTER 1 INTRODUCTION 1 2 STUDY SYSTEM 11 Study Species 11 Canal System 14 Study Sites 18 3 A QUALITATIVE RISK ASSESSMENT OF PEACOCK CICHLIDS CICHLA OCELLARIS AND SOME IMPLICATIONS FOR THE METHODOLOGY 24 Introduction 24 Background Information 27 Qualitative Risk Assessment 36 Discussion 43 4 ONTOGENETIC DIETARY SHIFTS AND DIETARY OVERLAP OF A NONINDIGENOUS FISH AND AN ECOMORPHOLOGICALLY SIMILAR NATIVE FISH 49 Introduction 49 Methods 51 Results 56 Discussion 68 5 EXPERIMENTAL PREY HANDLING AND SELECTIVITY OF TWO MORPHOLOGICALLY SIMILAR PREDATORY FISHESNATIVE LARGEMOUTH BASS AND INTRODUCED PEACOCK CICHLIDS 79 V

PAGE 6

Introduction 79 Methods 82 Results 90 Discussion 99 6 THE INTRODUCTION OF PEACOCK CICHLIDS INTO SOUTHEAST FLORIDA: POPULATION-LEVEL CONSEQUENCES FOR LARGEMOUTH BASS 109 Introduction 1 09 Methods 112 Results 114 Discussion 119 7 SUMMARY 127 REFERENCES 133 BIOGRAPHICAL SKETCH 154 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 EFFECTS OF INTRODUCED PEACOCK CICHLIDS CICHLA OCELLARIS ON NATIVE LARGEMOUTH BASS MICROPTERUS SALMOIDES IN SOUTHEAST FLORIDA By Jeffrey E. Hill December 2003 Chair: Charles E. Cichra Major Department: Fisheries and Aquatic Sciences The predatory peacock cichlid Cichla ocellaris native to South America, was introduced into southeast Florida to enhance recreational fishing and to provide biological control over nonindigenous fishes. This action was controversial and there is concern over the effects of the introduction on native species, especially the ecomorphologically similar largemouth bass Micropterus salmoides I conducted a formal qualitative risk assessment to evaluate the risks associated with the peacock cichlid introducfion and to test predictions of the algorithms. The esfimate of risk for peacock cichlids was Low. The evaluation elucidated shortcomings in the methodology (e.g., emphasis on a single output) and I made recommendations for its improvement. Part of the concern over the peacock cichlid introduction is their similarity to largemouth bass. I tested and found support for the hypothesis that morphological similarity equals similarity in prey handling ability and prey selection. I documented substantial overlap in vii

PAGE 8

trophic morphology and dietary ontogeny between juveniles and found similar prey handling ability and prey selection by adults in tank experiments. These results suggest that largemouth bass and peacock cichlids have similar effects on the prey base. However, largemouth bass had exclusive use of highly defensive prey (i.e., crayfish). Largemouth bass populations might be expected to have declined, given evidence of direct negative effects of peacock cichlids. I investigated this question by comparing largemouth bass populations in southeast Florida canals with populations in lakes and streams (statewide and in south Florida). Additionally, I compared canals with peacock cichlids to canals without peacock cichlids. The data did not support a hypothesis of declines in largemouth bass abundance due to peacock cichlids. These results question common assumptions concerning the effects of nonindigenous species — 1) that direct effects of predation and competition are the predominant effects and 2) the narrow view that invaded communities are organismic (i.e., Clementsian) — and challenge the uncritical acceptance of the premise that nonindigenous fishes will invariably cause ecological harm. My results suggest that other factors (e.g., indirect effects) may be important and that invasion biology should develop a pluralistic theory that incorporates Gleasonian concepts (i.e., individualistic communities). viii

PAGE 9

CHAPTER 1 INTRODUCTION Introduced species are considered threats to natural ecosystems and biodiversity worldwide (Lodge 1993). For example, Miller et al. (1989) reported that nonindigenous species were a factor in the extinction of 27 North American fishes (68% of extinctions). Similarly, Lassuy (1995) stated that nonindigenous species were factors in listing 48 imperiled freshwater fishes under the Endangered Species Act in the USA (70% of listed fishes). Although ecological damage due to nonindigenous species is difficult to esfimate, economic costs in the billions of dollars are cited for aquafic and terrestrial systems in the USA (OTA 1993). Indeed, the overwhelming consensus is that species introductions may result in deleterious effects (e.g., lowered abundance of native species) within the receiving system (Magnuson 1976; Courtenay 1993, 1995; Minns and Cooley 2000). Nonindigenous fishes have a long history in the USA and have entered this country through several intentional and unintentional pathways (Fuller et al. 1999). Many fishes have high economic and social value and have been purposefially stocked to provide fisheries benefits. For example, centrarchids (e.g., Lepomis and Micropterus ). percids (e.g., Sander [formerly Sfizostedion] ). salmonids (e.g., Oncorhynchus and Salmo ), and other fishes have been transferred to many regions to enhance recreational fisheries (Fuller et al. 1999; Heidinger 1999). Purposeftil introducfions for recreafional fishing are perhaps the most studied cases of interaction between native and 1

PAGE 10

2 nonindigenous fishes. Nevertheless, thorough empirical evaluations are few (but see Huckins et al. 2000). Such research is costly and may be controversial (e.g., Churchill et al. 2002). Introduced salmonids have been intensely scrutinized, yet there is wide variation in their reported effects on native salmonids (Fausch 1988). In a review of the effects of salmonid introductions, Fausch (1988) found that some authors documented little or no effect of introduced salmonids, yet other authors concluded that such fishes constitute a serious threat to native species (see also Fuller et al. 1999; Crawford 2001). Similarly, data for non-salmonid introductions are often mixed and inconclusive (Taylor et al. 1984; Shafland 1996a; Courtenay 1997; Fuller et al. 1999). Regardless of the potential or realized benefits, the possible dangers have led to many calls for a cessation of purposeful fish introductions (Magnuson 1976; Courtenay and Robins 1973, 1989; Courtenay 1995; Minns and Cooley 2000). These papers point out the numerous theoretical negative impacts of fish introductions and that some introduced fishes have proven to be notable and costly pests (e.g., the unintenfionally introduced sea lamprey Petromyzon marinus in the Great Lakes; Mills et al. 1994). Given increasing transfers of fishes outside their natural ranges (Nico and Fuller 1 999), there has been significant scientific and public interest in predicting the effects of species invasions on native aquatic faunas (Lodge 1993; OTA 1993; Moyle and Light 1996a; Kolar and Lodge 2002). Indeed, inadequate predicfive ability in this field has frustrated scientists and resource managers. This situafion is partly due to the surprisingly few empirical evaluafions of the effects of nonindigenous fishes (Taylor et al. 1984; Moyle et al. 1986; Fuller et al. 1999). In fact, much of the literature on harmfiil effects is anecdotal. Furthermore, a lack of a truly pluralistic ecological theory of bio-

PAGE 11

3 invasion may be a factor hampering the advance of this field (Schoener 1987; Ross 1991 ; see also Lodge 1993; Williamson 1996; Shrader-Frechette 2001; Sakai et al. 2001). General ecological theory provides a fi-amework for a variety of predictions of the effects of species introductions. For example, there is a rich literature concerning the effects of species introductions and removals in temperate lakes on lake productivity and food web structure (e.g., Carpenter et al. 1985; Carpenter et al. 1987; Mittelbach et al. 1995) Some researchers have used ecological theory to predict and subsequently document negative effects of species introductions (e.g., Huckins et al. 2000); others have predicted and then demonstrated effects that were positive for aquatic resources (e.g., Carpenter et al. 1985; Carpenter et al. 1987). Therefore, ecological theory is not biased toward an invariable prediction of negative effects. Despite this rich body of ecological theory, many scientists and resource managers interested in the ecological effects of nonindigenous species assume that invading species have only net negative effects on native species (see reviews in Stauffer 1984; Taylor et al. 1984; Moyle et al. 1986; Li and Moyle 1999; but see Williamson 1996) The current emphasis on negative effects of introduced species are founded on two principles: 1) the organismic concept of equilibrial, stable, and competitivelystructured communities (Clements 1936) and 2) that direct effects of invasive species (via competition and predation) are far more important than direct or indirect positive effects that might negate or overwhelm the direct negative interactions. The expectation that competition will play a key role was influenced by Cause's articulation of the competitive exclusion principle (Cause 1 934; Hardin 1 960), the Hutchinsonian concept of niche (Hutchinson 1957), Elton's (1958) work on species invasions, and MacArthur

PAGE 12

4 and colleague's papers on competition and island biogeography (e.g., MacArthur and Levins 1967; MacArthur and Wilson 1967). Such theory (e.g., island biogeography) predicts that any species addition must lead to a species extinction (MacArthur and Wilson 1967). Under this view, the observed increase in invasions (e.g., Nico and Fuller 1 999) is certain to lead to increased rates of extinction (or extirpation) of native species (and previously established nonindigenous species) — a prediction leading to considerable concern for conservation biologists and resource managers over a loss of biodiversity. Furthermore, the native species most at risk of extinction should be those species that are most similar to invading species with respect to resource use (Hardin 1960; MacArthur and Levins 1967; Diamond 1975; but see Resetarits 1995). Because similarity in resource use is often associated with similarity in morphology (Wainwright 1988, 1996; Wainwright and Richard 1995), species of similar morphology are expected to be the most sensitive to invading species. Moreover, proponents of this view of species invasions aggressively challenge any notion of "empty niches" on semantic and theoretical grounds (e.g., Herbold and Moyle 1986; Courtenay 1995), practically precluding the possibility that species introductions can take place without any deleterious effects for the receiving system. Nonindigenous fishes may adversely affect native species directly or indirectly via competition, predation, habitat alteration, reproductive inhibition, and shared pathogens (Taylor et al. 1984; Moyle et al. 1986). Indeed, many conclusions of impacts of nonindigenous fishes assume that these mechanisms of negative interaction will directly translate into reductions in the density of native species. If native fish populations are low or are perceived to have declined coincidentally with the introduction

PAGE 13

of a nonindigenous fish, then evidence for the existence of a putative mechanism is often assumed to be proof that the mechanism is a contributing factor. This apparent causeand-effect scenario may be correct for some cases of species introductions. Nevertheless, the existence of a negative interaction between the two species may not be important to the native species at the population level if, for example, the total strength of the interaction (per capita effect x population size) is low, if other factors (e.g., abiotic events) are strong, or if density-dependence at later life stages leads to compensatory responses (e.g., Vonesh and De la Cruz 2002). As a result, data are required that directly link the interactions between introduced and native species and the observed effects on population density. Despite these views that invasions should lead to reductions in abundance or extinctions of native species, there is growing evidence that new species can be added to native communities with little noticeable effect (Moyle and Light 1 996a; Williamson 1996; Gido and Brown 1999). Indeed, perhaps as many as 80-90% of all established nonindigenous species fall into this category (Williamson 1996). For example, Moyle and Light (1996a) noted several examples of invading fishes integrating into Ireshwater and marine communities without obvious changes in native species. A partial explanation for the failure of many scientists and resource managers to expect observed cases of integration of invading species into communities is a conscious or tacit adherence to the Clementsian view of organismic communities (Clements 1936). On the other hand, if communities are considered loosely-structured assemblages of species that result from probabilistic processes, that are transient in time and space, and are dependent on the characteristics of individual species (individualistic concept;

PAGE 14

Gleason 1926), then integration of invaders into local assemblages is a possibility. Clearly, there is a great deal of evidence that plant and animal associations have changed over time (e.g., Graham et al. 1996), implying that communities may change in species composition and are not fixed in structure. Exchanges of species during numerous faunal interchanges throughout the Neogene period (i.e., last 25 million years) have often led to enrichment of faunas rather than species replacements, suggesting that many communities are not saturated with species (Vermeij 1991). Moreover, Rosenzweig (2001) argued on theoretical grounds that species introductions often permanently increase local biodiversity. Likewise, Gido and Brown (1999) reviewed fish community data for 125 drainages in North America and concluded that fish introductions generally increased species richness and that this pattern suggests that North American fish communities are not saturated with species. Mcintosh (1995) reviewed the controversy associated with the competing paradigms of Clements and Gleason and noted that, although debate is ongoing, many animal ecologists had joined with plant ecologists in viewing communities as transient, individualistic assemblages of species. Although it is a mistake to assume that biotic factors are not important in individualistic communities, strong community structuring by biotic factors is a defining characteristic of the organismic concept of community (Mcintosh 1 995). The emphasis of the organismic concept is apparent in much of the literature on nonindigenous fishes. For example, even though abiotic factors are frequently considered when evaluating risks of establishment of nonindigenous fishes (Baltz and Moyle 1993; Moyle and Light 1996a, b; RAM Committee 1998), most studies have considered biotic interacfions to be preeminent once a nonindigenous species becomes established (Taylor et al. 1984;

PAGE 15

7 Courtenay 1993; but see Moyle and Light 1996b). However, in his review of fish introductions and stream fish assemblage structure, Ross (1991) pointed out that the establishment of a nonindigenous species with no apparent effect on the native assemblage suggests strong abiotic effects within the community. Another important reason for the general prediction of harm to native species following a species introduction is the prominent emphasis on the direct negative effects of competition and predation. Both mechanisms can have strong negative effects on species abundance (Garman and Neilson 1982; McComas and Drenner 1982; Crowder and Binkowski 1983; Kaufinan 1992; Mills et al. 1994). However, indirect effects can be both facilitative and powerfiil. For example, Lawlor (1979) reported that 30-40% of interactions in eight bird communities were positive. Miller (1994) estimated the magnitude of direct and indirect interactions of five weedy plant species. He found that although direct negative effects were large, indirect effects were generally positive, relatively strong, and reduced the influence of competition. Additionally, Stone and Roberts (1991) argued that a high proportion (i.e., 20-40%) of species interactions must be advantageous within the context of a community and that "hypercompetitive communities" (i.e., when every species suffers ft-om the presence of every other species) must be rare. In some cases, positive interactions between putative competitors can overwhelm direct negafive effects and result in increases in populafion abundance (i.e., facilitation) (e.g.. Miller 1994). The exotic peacock cichlid Cichla ocellaris (Cichlidae) was intenfionally introduced into southeast Florida canals in 1984-1987 to provide biological control of other exotic fishes and to create a novel sport fishery (Shafland 1995). This relatively

PAGE 16

8 large, predatory species is ecomorphologically similar to the native largemouth bass Micropterus salmoides (Centrarchidae) (Norton and Brainerd 1993) that already occurred in these habitats. In their presentation of empirically-derived "rules" of biological invasions, Moyle and Light (1996a) concluded that the piscivore trophic group is one of the most likely to become successful invaders and to subsequently alter the fish assemblages of receiving systems. The ecomorphological similarity between these two species is also of concern given the potential for resource competition. Indeed, there has been reported substantial overlap in size-specific gape width (Hill 1998), diet (Shafland 1995; 1999b), and habitat use (Lilyestrom and Churchill 1996) between largemouth bass and peacock cichlids. Additionally, there has been speculation and concern regarding possible deleterious effects of peacock cichlids on native fishes in Florida (Courtenay and Robins 1989; OTA 1993; Courtenay 1994; Cox 1999; see also Lachner et al. 1970). Indeed, Cox (1999) reviewed species introductions into North America and Hawaii and listed the peacock cichlid as an "invasive exotic" species in Florida, a designation requiring actual or probable harm resulting from its introduction. To evaluate the actual effect of peacock cichlids on native largemouth bass, I first considered the potential direct and indirect effects of the introduction on largemouth bass. Using these theoretical effects and additional biological information, I conducted a qualitative risk assessment (RAM Committee 1 998) to provide a framework to formalize the predictions of peacock cichlid effects and to evaluate this methodology using peacock cichlids as a test case (Chapter 3). Given the natural history of this system, predation by peacock cichlids on small largemouth bass and competition for food between similar size classes of peacock

PAGE 17

9 cichlids and largemouth bass were the most likely interactions; in the absence of mitigating factors, both would lead to a decline in largemouth bass population density in response to an introduction of peacock cichlids. Other potential mechanisms commonly cited in the literature such as spawning site competition, habitat alteration, hybridization, and introduction of disease (Taylor et al. 1984; Li and Moyle 1999) were considered but discarded as unlikely. Although both species are substrate spawners, there is relatively little temporal overlap in spawning activity (Shafland 1999b). Unlike common carp Cyprinus carpio grass carp Ctenopharyngodon idella or some tilapia species ( Oreochromis spp., Tilapia spp.) (Taylor et al. 1984; Fuller et al. 1999), peacock cichlids are not known to alter habitat by destroying aquatic macrophytes, by disturbing sediments, or by any other mechanism. The possibility of hybridization is extremely remote because peacock cichlids are phylogenetically distinct from largemouth bass (or any other native Florida fish) at the familial level. Additionally, samples of peacock cichlids were screened for known pathogens by personnel of the US Fish and Wildlife Service prior to release (Shafland 1 995). On the other hand, a diet study conducted in Florida confirmed the use of largemouth bass as prey by peacock cichlids in Florida (Shafland 1999b). However, the potential for food competition had received little investigation (but see Shafland 1995, 1999b; Hill 1998). In the present study, I assessed food competition indirectly, largely through overlap in food resource use. Some work on trophic overlap had been conducted with subadult and adult largemouth bass and peacock cichlids; Hill (1998) had previously investigated their similarity in gape size and Shafland (1 999b) had compared diets of both species in a southeast Florida canal. I therefore determined overlap in trophic morphology (i.e., morphological ability to handle prey) and

PAGE 18

10 dietary overlap from stomach contents using small juveniles (i.e., up to about 160 mm total length). Additionally, I investigated patterns of prey selectivity of subadult and adult largemouth bass and peacock cichlids in two experimental settings. Although overlap in resource use is not a direct proxy for competition (Col well and Futuyma 1971 ; Sale 1979; Schoener 1982), shared use of a resource is generally considered to be a prerequisite for exploitation competition (Moermond 1979; Sale 1979; but see Rosenzweig 1981 for situations where competition is an important interaction but overlap in resource use is essentially zero; Chapter 4). Additionally, resource overlap is often used to justify conclusions of interspecific competition in evaluations of species introductions. Another objective for the present study was to test the hypothesis that peacock cichlids have had large negative effects on largemouth bass populations in southeast Florida. Based on the expectation that predation and food competition lead to reductions in largemouth bass density, I predicted that south Florida canals would have reduced abundance of largemouth bass relative to other types of Florida systems (i.e., lakes and streams) and that canals with peacock cichlids would have lower abundance of largemouth bass than canals lacking peacock cichlids. Therefore, I used population data for largemouth bass from south Florida canals and other Florida systems to test these hypotheses. An additional objective was to determine if the present study supported the prevailing view among many invasion biologists based on Clementsian (i.e., organisimic) communities or a more pluralistic alternative that incorporates Gleasonian (i.e., individualistic) communities.

PAGE 19

CHAPTER 2 STUDY SYSTEM Study Species The largemouth bass Micropterus salmoides (Centrarchidae) is native to much of eastern North America, including southern Florida where it is considered a distinct subspecies, the Florida largemouth bass Micropterus salmoides floridanus (Lee 1980). The biology and ecology of largemouth bass are well studied (e.g., Heidinger 1975; Carlander 1 977). Largemouth bass is a habitat generalist, with small individuals occurring in or near cover and larger individuals occupying a wide range of cover types and open water habitats (Keast et al. 1978; Mesing and Wicker 1986; Armett et al. 1996; Annett 1998; pers. obs.). It is predatory, with a general ontogenetic dietary progression from zooplankton to insects and small crustaceans to decapod crustaceans and fish (Chew 1974). Adults are considered primarily piscivorous (Chew 1974; Carlander 1977; Cailteux et al. 1996). Nests are constructed by males in relatively shallow water, often near cover. Largemouth bass in southern Florida spawn in winter to early spring (peak in March in canals) over a relatively short time period (Shafland 1 999b). This results in a more or less distinct yearly cohort that appears in spring and declines in numbers through the year (pers. obs.). Members of the genus Cichla (Cichlidae) are relatively large, predatory fishes that superficially resemble the North American black basses (i.e., Micropterus ). Cichla are native to most major Atiantic drainages of tropical South America. In English, this genus 11

PAGE 20

12 is known as peacock cichlids or peacock bass. Cichla comprise a group of five described and perhaps six or more undescribed species (Machado-Allison 1971; Kullander 1986; Kullander and Nijssen 1989). To date, the introduced form established in Florida has been designated as the peacock cichHd C. ocellaris a species considered native to northeastern South America, although there remains some uncertainty about the alpha taxonomy of the genus (see Kullander and Nijssen 1989). During 1984-1987, fmgerling peacock cichlids were intentionally introduced into freshwater canals in Broward and Miami-Dade counties of southeast Florida by the Florida Game and Fresh Water Fish Commission — currently Florida Fish and Wildlife Conservation Commission; FWC (Shafland 1 995). The introduction of this piscivore was planned to provide a biological control on the large populations of prey fish, primarily of exotic origin, to correct an unbalanced fishery (sensu Swingle 1950) and to supplement recreational fisheries that were principally based on largemouth bass (Shafland 1995). Peacock cichlids spawn in nests constructed in relatively shallow water throughout the warmer months (Zaret 1 980a). Therefore, unlike largemouth bass, cohorts of age-0 peacock cichlids arise throughout much of the year, although they appear later than the single largemouth bass cohort. Peacock cichlids are habitat generalists and occur syntopically with largemouth bass at all life stages (pers. obs.). Largemouth bass and peacock cichlids are not naturally sympatric. Nevertheless, these species co-occur in Hawaii and Puerto Rico due to the introducfion of both species and in Florida due to the introducfion of peacock cichlids. Both species are important sport fish in all three regions. Devick (1980) reported that peacock cichlids in Hawaiian reservoirs complemented largemouth bass as a sport fish due to differing spawning

PAGE 21

13 seasons. Water withdrawals from the reservoirs often led to loss of nests or fry of substrate spawning fishes (e.g., largemouth bass and peacock cichlids). The extended breeding period of peacock cichlids relative to largemouth bass often allowed its successftil reproduction and subsequent recruitment into the sport fishery, in effect carrying the sport fishery over periodic years of largemouth bass recruitment failure (Devick 1972; Devick 1980). Lilyestrom and Churchill (1996) found both species to coexist in Puerto Rican reservoirs with little evidence of competition based on slight differences in habitat use and diet and major differences in diel activity periods and spawning seasons. Shafland (1995, 1999a, b, c) noted overlap in diet but concluded that peacock cichlids had not harmed largemouth bass populations in Florida canals. Also of note is the fact that two or more species of peacock cichlids co-occur in many South American river systems (Jepsen et al. 1997; Winemiller et al. 1997; L. G. Nico, USGS, pers. comm.). Largemouth bass and peacock cichlids are ecomorphologically similar based on gape width, general morphology, feeding mechanics, and diet (Norton and Brainerd 1993; Hill 1998). Given this similarity in ecology, a few experimental studies have explored possible competitive interactions between largemouth bass and peacock cichlids. Swingle (1967) concluded that peacock cichlids were less efficient predators of tilapia (probably redbelly tilapia Tilapia zillii but listed as T. melanopleura ) and fathead minnows Pimephales promelas than largemouth bass under a similar stocking rate and conditions in an Alabama pond. Lilyestrom et al. (1994) stocked age-0 largemouth bass (mean TL = 48.8 1 .0 mm) and peacock cichlids (mean TL = 1 8. 1 1 1 mm) into experimental ponds in Puerto Rico and found that largemouth bass reduced growth and

PAGE 22

14 survival of peacock cichlids but that largemouth bass grew faster and survived better with peacock cichlids. As pointed out by the authors, the results might have been strongly influenced by the disparity in initial sizes of the two species, although these were chosen to reflect the size relationship early in the peacock cichlid spawning season. Although peacock cichlids have been introduced into several regions, there is little evidence of deleterious effects except in Lake Gatun, Panama (Zaret and Paine 1973) and in Venezuelan reservoirs (cited in Winemiller et al. 1997). The accidental introduction of peacock cichlids into Lake Gatun, Panama, and the subsequent environmental harm, is one of the most cited examples of the danger of fish introductions. Native fishes in the massive reservoir were reportedly devastated by this novel predator and local extirpations occurred (Zaret and Paine 1973). The predation effects apparently cascaded throughout the food web and resulted in significant changes in zooplankton, insect, and bird populafions (Zaret and Paine 1973). Less frequently cited is the report of Welcomme (1988) menfioning the rebound of nafive fish populations in subsequent years. Canal System Southeast Florida is drained by a vast network of artificial canals. Much of the canal network was constructed during the early to mid 20"' Century (mainly fi-om 1913 to 1969), primarily to lower the local water table to create agricultural land, to facilitate urban development, and to protect these lands from flooding (VanArman et al. 1984). Flood control is largely achieved by redirecting water fi-om the Kissimmee River-Lake Okeechobee-Greater Everglades system generally south and east to the coast. Historically, most of the flow of this system was to the south and southwest. Although

PAGE 23

15 some reaches are channelized streams, most of the canal system is entirely new, permanent, deep-water habitat — a freshwater habitat type that is naturally scarce in southern Florida (Loftus and Kushlan 1987). Many canals have substantial connection with the surficial aquifer, especially in eastern and southern portions of the region. The canal system is operated and maintained by the South Florida Water Management District (SFWMD) which controls flow and water routing with a combination of pump stations and water control gate structures. Although canals are found in much of southern Florida, peacock cichlids are largely confined to systems along the southeast coast, particularly the eastern portions of Broward and Miami-Dade counties (about 530 km of canal; Shafland and Stanford 1999). This species is also found in some southern Palm Beach County canals (total range of about 650 km of canals; Shafland and Stanford 1999). In Florida, peacock cichlids generally inhabit canals that flow through suburban and urban areas, including the densely-populated Ft. Lauderdale and Miami metropolitan areas. Some are found in rural canals (i.e., pass through agricultural lands) in western portions of the peacock cichlid range. Peacock cichlid distribution is limited by cool winter temperatures (Shafland 1995; i.e., lower lethal temperature of 15 C; Swingle 1967; Guest et al. 1980). Its range expands slightly north (i.e., into Palm Beach County) and west during years with mild winters. Canal morphometry is characterized as box-cut, with vertical walls cut deeply (often > 3-5 m) into the substrate. In much of the region, the substrate is porous coral rock. A narrow shelf (usually 1-2 m wide), often with submersed vegetation, occurs along the banks and may be the only littoral zone for the canal. Smaller, lateral canals

PAGE 24

16 are found in many systems. These are often shallower and have more stable aquatic macrophyte communities (pers. obs.). Water chemistry can vary widely both temporally and spatially throughout the canal system (Shafland 1999a; pers. obs.). Intermittent releases of water from Lake Okeechobee or the Water Conservation Areas south of the lake greatly increase flow and alter water physico-chemical parameters (e.g., alkalinity, color, dissolved oxygen, pH, temperature, and turbidity). These changes can occur rapidly (i.e., within minutes or hours), especially in larger canals that are used as main conduits for de-watering. In particular, water releases can stress fishes by inducing hypoxic conditions. Tidal influences are minimal due to salinity control gates near the coast. These structures only open to release fresh water, although euryhaline fishes may swim up the current and enter the freshwater canal system. The primary purpose of the canals is drainage and therefore SFWMD maintains an extensive program to remove aquatic macrophytes and overhanging terrestrial vegetation. Many canals support dense stands of aquatic macrophytes, often exotic invasives (e.g., hydrilla Hydrilla verticillata ), that inhibit flow. The SFWMD uses biological (i.e., grass carp Ctenopharvngodon idella ), chemical, and mechanical control methods, but mechanical harvesting is the most common. Mechanical harvesting occurs throughout the year and may be frequent in main canals during warmer months. Coverage of aquatic macrophytes exceeding 90% (percent areal coverage, PAC) can be reduced to < 5% in a single day by mechanical harvesters (pers. obs.). Such drastic changes in habitat structural complexity likely have dramatic effects on invertebrate and fish populations, including largemouth bass and peacock cichlids. Aquatic macrophytes

PAGE 25

17 serve as important habitat and as a refuge from predation for many invertebrates and small fishes (Savino and Stein 1982; Werner et al. 1983; Gilinsky 1984; Chick and Mclvor 1997; Jacobsen and Perrow 1998; Burks et al. 2001). Individual fishes, including age-0 largemouth bass and peacock cichlids, are also removed directly within the harvested vegetation (L. G. Nico, USGS, pers. comm.; pers. obs.). Electrofishing catchper-unit-effort (CPUE) for age-0 individuals is significantly lower immediately after harvesting activities (unpubl. data). The peacock cichlid is one of about 25 exotic fish species established in Florida (Fuller et al. 1999; Hill 2002), with about 18 of these established in southeast Florida canals (Loftus and Kushlan 1987; Fuller et al. 1999; Loftus 2000; Hill 2002). In the canals, nonindigenous fishes, particularly cichlids, may dominate the fish assemblage in numbers and biomass (Courtenay et al. 1974; Shafland 1999a). The native fish fauna is relatively depauperate, with about 43 freshwater and 1 3 euryhaline species (Loftus and Kushlan 1987; Loftus 2000; Table 1). The native freshwater species are affiliated mostly with North American temperate families and are relatively recent invaders due to the short geologic time that southern Florida has been exposed above sea-level (Gilbert 1987). The general perception in the literature is that native fishes in south Florida urban canals are subjected to numerous stressors (e.g., canal morphometry, variable and often harsh water physico-chemistry conditions, frequent aquatic macrophyte manipulation, and presumably strong biotic interactions with exotic fishes) that result in reduced population abundances (Loftus and Kushlan 1987; Courtenay 1997; Annett 1998). For example, several authors have suggested that cichlids exert strong negative effects on

PAGE 26

18 native fishes (Buntz and Manooch 1968; Hogg 1976; Noble and Germany 1986; Courtenay and Robins 1989; Courtenay 1997). On the other hand, freshwater canals that traverse marsh areas serve as important refligia for large-bodied fishes (native and nonindigenous) during low water periods and generally enhance populations of such fishes (Loftus and Kushlan 1987); these canals are found mostly west of the distribution of peacock cichlids. Study Sites Most field work for this dissertafion was conducted in Cutler Drain Canal (C1 OOC) and Snapper Creek Canal (C-2) in the Biscayne Bay Drainage, Miami-Dade County, Florida (Figures 1 and 2). Collections of juvenile largemouth bass and peacock cichlids were made from the S-1 19 water control structure (25 38.58' N, 80 20.31 W) to about 1 .6 km upstream of the structure in Cutler Drain Canal and from the Don Shula Expressway bridge (downstream of the boat ramp on Snapper Creek Drive [25 42.00' N, 80 21 .25' W]) upstream up to, and including, the two lateral canals at the crossing of SW 99"* Avenue in Snapper Creek Canal. These two canals were chosen because they are within the center of the range of peacock cichlids in southeast Florida, are relatively large canals, are accessible by boat, are in reasonably safe locations for personnel and equipment, and have different characteristic water chemistry and clarity. Cutler Drain Canal (Figure 2A) in the Howard area is the smaller of the two canals and lacks large lateral canals in the study area. Under most conditions, water clarity is relatively high, flow is low, and groundwater is the major water source. Cutler Drain Canal is used as a main flow conduit only during infrequent high water events. I

PAGE 27

19 I Miami-Dade County N A Cutler Drain Creek Canal Everglades National Park 25 km Figure 1 Location of select canals in Miami-Dade County, Florida. Snapper Creek Canal and Cutler Drain Canal were study canals. Inset map shows Florida. Map redrawn from Shafland et al. 2001. observed a substantial change in aquatic macrophyte species composition during the study period. In 1999, hydrilla Hydrilla verticillata fanwort Cabomba caroliniana and

PAGE 28

20 Figure 2. (A) Cutler Drain Canal (C-IOOC) and (B) Snapper Creek Canal (C-2), MiamiDade County, Florida. Maps redrawn from Hidalgo 1997.

PAGE 29

21 bacopa Bacopa sp. were the main species, with hydrilla being dominant. By 2001 Asian hygrophila Hygrophila polysperma became the dominant macrophyte, with bacopa also being relatively common. Concurrently, numerous large grass carp were found in the area, presumably stocked by the SFWMD for aquatic macrophyte control. Hydrilla is a highly selected food of grass carp (Shireman and Smith 1983), a possible factor in the species replacement. Also, Cutler Drain Canal had low overall macrophyte density during 1999 due to macrophyte removal. Qualitatively, abundance of both native and exotic fish was relatively high (see Table 1 for a list of species). Snapper Creek Canal (Figure 2B) is a major canal conduit in southeast Florida. The study area encompassed a large, mainstem canal and two large lateral canals. Water flow is frequently high in the main canal and water characteristics are most influenced by surface water. Snapper Creek Canal receives most of its water fi-om the Tamiami Canal system, a significant drainage from the Water Conservation Areas and, ultimately, Lake Okeechobee from the north. The water is highly colored and occasionally turbid. Water clarity is usually greater in the lateral canals. Aquatic macrophytes are generally abundant in Snapper Creek Canal, especially in the laterals. Hydrilla dominated the main canal and cabomba in the laterals. Significant amounts of muskgrass Chara sp., pond weed Potamogeton sp., and vallisneria Vallisneria americana also occurred. Snapper Creek Canal has one of the most diverse fish faunas of the canal system (Table 1) and fish abundance was often high, especially in the lateral canals.

PAGE 30

22 Table 1.Fish species collected from two freshwater canals in southeast Florida in 19972001 A "*" indicates a nonindigenous species. A "?" indicates a tentative identification. Cutler Drain Canal had 3 1 species (21 native and 1 0 nonindigenous). Snapper Creek Canal had 40 species (22 native and 1 8 nonindigenous). Scientific Name Common Name Occurrence Cutler Drain Snapper Creek Amiidae Amia calva Bowfin X Anguillidae Anguilla rostrata Atherinidae Labidesthes sicculus American eel Brook silverside X Catostomidae Erimyzon sucetta Lake chubsucker Centrarchidae Enneacanthus gloriosus Lepomis gulosus Lepomis macrochirus Lepomis microlophus Lepomis punctatus Micropterus salmoides Centropomidae Centropomus undecimalis Cichlidae Astronotus ocellatus Cichla ocellaris Cichlasoma bimaculatum* Cichlasoma citrinellum Cichlasoma managuense* Cichlasoma urophthalmus Geophagus surinamensis Hemichromis letoumeauxi Heros severus Oreochromis aureus Oreochromis mossambicus Tilapia mariae Clariidae Clarias batrachus* Bluespotted sunfish Warmoutii Bluegill Redear sunfish Spotted sunfish Largemouth bass Common snook Oscar Peacock cichlid Black acara Midas cichlid Jaguar guapote Mayan cichlid Redstriped eartheater Jewel cichlid Green severum Blue tilapia Mozambique tilapia Spotted tilapia Walking catfish X X X X X X X X X X X X X X X X X X X X X X X X X X X X? X X Clupeidae Dorosoma cepedianum Gizzard shad X

PAGE 31

23 Table 1 Continued. Scientific Name Common Name Occurrence Cutler Drain Snapper Creek Cyprinidae Ctenopharyngodon idella Cvprinus carpio Notemigonus crysoleucas Cyprinodontidae Lucania goodei Grass carp Common carp (koi) Golden shiner Elopidae Elops saurus Fundulidae Fundulus chrysotus Ictaluridae Ameiurus natalis Ameiurus nebulosus Lepisosteidae Lepisosteus platyrhinchus Loricariidae Ancistrus sp.* Hypostomus sp.* Pterygoplichthys multiradiatus Mugilidae Mugil cephalus Percidae Etheostoma fiisiforme Poeciliidae Gambusia holbrooki Heterandria formosa Poecilia latipinna Bluefin killifish Ladyfish Golden topminnow Yellow bullhead Brown bullhead Florida gar Striped mullet Swamp darter Eastern mosquitofish Least killifish Sailfin molly X X X X X X X X X X X X X X X X X X X

PAGE 32

CHAPTER 3 A QUALITATIVE RISK ASSESSMENT OF PEACOCK CICHLIDS CICHLA OCELLARIS AND SOME IMPLICATIONS FOR THE METHODOLOGY Introduction Understanding and predicting the effects of introduced species are of interest for conservation biology and community ecology. Moreover, resource managers and regulatory authorities need such information. Unfortunately, prediction has proven elusive because of uncertainty due to the complexity of interacting factors in ecological systems (Baltz 1991 ; Ross 1991 ; FAO 1996). Furthermore, the outcome of some species introductions has been both unpredicted and undesired (i.e., Frankenstein Effect; Moyle etal. 1986). There have been many methods used to predict the effect of species introductions. Quantitative modeling methods (e.g., Kolar and Lodge 2002) can be used, but these are data-intensive, require ecosystem-, taxon-, and stage-specific information, and have received little testing. As an alternative to quantitative models, qualitative methods also exist to estimate the effects of nonindigenous species (e.g., loop analysis; Levins 1974 [cited in Li et al. 2000]; Li and Moyle 1981; Li et al. 2000). Less formal methods include soliciting the opinions of experts who have extensive knowledge of the species and systems involved. Major drawbacks of this method are subjectivity and non-repeatability (Ambrose et al. 1996). A slightly better approach might involve a more formal method to integrate expert opinions into a risk assessment. Indeed, the Generic Nonindigenous 24

PAGE 33

25 Aquatic Organism Risk Analysis Review Process (i.e., RAM methodology) has been developed by the federal Aquatic Nuisance Species Task Force as a method to predict risks associated with newly established species, species proposed for intentional introduction, and individual introduction pathways (RAM Committee 1998). This methodology uses information on species life history and ecology, characteristics of its native range and the region of potential introduction, and vectors of introduction to assign qualitative probabilities that the focal species will colonize open waters, successfully reproduce and become established, and adversely affect native species or economic activity (RAM Committee 1998). The RAM methodology has been recently used by the U.S. Fish and Wildlife Service to synthesize information and to guide decision-making regarding proposals to list certain fishes as Injurious Wildlife Species (e.g., black carp Mylopharyngodon piceus ; Nico et al. 2001 ). There have been only three formal applications of the RAM methodology for fish and all have been assessments for specific organisms or groups of organisms (i.e., not for introduction pathways) — black carp (Nico et al. 2001), snakeheads (Family Channidae) (W. R. Courtenay, Jr. and J. D. Williams, USGS, in prep.), and Asian swamp eel Monopterus albus (L. G. Nico, USGS, in prep.). Another risk assessment using this methodology is planned to include bighead carp Hypopthalmichthys nobilis and silver carp H. molitrix (J. D. Williams, USGS, pers. comm.). Of these risk assessments, only the black carp risk assessment was conducted prior to the occurrence of reproducing populations of the focal species (or one or more members of the family for snakeheads) in open waters within the USA. Furthermore, black carp was considered the test case for the use of the RAM methodology, partly

PAGE 34

26 because there was still time to apply risk management strategies if warranted (Nico et al. 2001). The black carp risk assessment tested the RAM methodology as a vehicle for risk assessment (i.e., would the method work as envisioned and produce an estimate of risk and certainty?). The black carp risk assessment demonstrated that the methodology could achieve these goals. On the other hand, there has been no testing of the predictions of the risk algorithm relative to any test case. In order to test these predictions, a number of species or groups that are already established should be chosen to represent a variety of nonindigenous species scenarios (e.g., various taxa, geographic distributions, and socioeconomic importance). Risk assessments could then be conducted and the results of the risk assessments (i.e., risk predictions) could be compared to the observed outcomes (up until that time). This process could provide insight into biases or grossly incorrect predictions based on the risk algorithm and suggest improvements in the process. Indeed, changes in the risk assessment methodology were envisioned by the authors who acknowledged the existence of uncertainty in the risk assessment process (RAM Committee 1998). My objective is to begin this process of testing the RAM methodology using the peacock cichlid Cichla ocellaris (Cichlidae) as a test case. The emphasis of this assessment will be Florida, the only region of the continental USA with established Cichla (see Cichla introductions and Limiting factors on Cichla distribution, below). This species was chosen because 1 ) it has been introduced into the USA (i.e., Florida) as a sport fish and for biological control (i.e., has socio-economic importance), 2) it is predatory and ecomorphologically similar to native species (i.e., is of potential ecological

PAGE 35

27 concern), 3) unlike the previously assessed fishes, it is limited to a rather narrow potential geographic range in the USA, 4) the species is not a new introduction (i.e., was introduced into Florida in 1984), and 5) it has been the subject of fisheries research in Florida and elsewhere. I will provide a brief overview of the biology, ecology, and history of introductions of Cichla to provide a background for the risk assessment process. I will then describe the RAM methodology and assign estimates of risk and certainty to the various rating elements. Lastly, I will discuss the predictions of the risk assessment algorithm relative to the data on observed effects of Cichla in Florida and suggest some modifications to the methodology. Background Information Cichla Taxonomy and Life History The genus Cichla is composed of five valid nominal species (i.e., C. intermedia Machado-Allison, C. monoculus Spix, C. ocellaris Schneider, C. orinocensis Humboldt, and C. temensis Humboldt) and perhaps another six undescribed species (KuUander and Nijssen 1989). There has been considerable confusion over the alpha taxonomy of Cichla (Lowe-McConnell 1969; Kullander 1986; Kullander and Nijssen 1989; Shafland 1995); the name C. ocellaris (i.e., peacock cichlid) has been applied to most Cichla that have been introduced outside of South America. The peacock cichlid is a relatively large (up to about 650 mm total length [TL] and 5.4 kg) species native to Atlantic drainages of northeast South America (Kullander and Nijssen 1989). Cichla superficially resemble the North American black bass of the genus Micropterus (Norton and Brainerd 1993).

PAGE 36

28 Peacock cichlids are predatory and adults are largely piscivorous (LoweMcConnell 1969; Zaret 1980a; Shafland 1995, 1999b; Jepsen et al. 1997; Winemiller et al. 1997). Juveniles progress through ontogenetic dietary stages of feeding on zooplankton, insects, and decapod crustaceans prior to feeding on fishes (Zaret 1980a; Chapter 4). Peacock cichlids form pairs during spawning. Although Cichla probably only spawn once per year in South American rivers (Jepsen et al. 1999), introduced populations in lakes, reservoirs, and canals spawn throughout the warmer months (LoweMcConnell 1969; Zaret 1980a; Jepsen et al. 1999; Shafland 1999b). Both parents guard a nest (cleaned substrate of wood, rock, or sand) where up to 1 0,000 eggs are laid (Zaret 1980a). The parents further protect the free-swimming fry for up to eight to 10 weeks (Zaret 1980). Fry occupy open water areas until they disperse from the guarded schools into littoral vegetation (Schroder and Zaret 1979; Zaret 1980a). Sexual maturity often is attained within one year at lengths of about 240-290 mm TL (Fontenele 1950; Shafland 1995). Subaduhs and adults are habitat generalists (pers. obs.). Cichla Introductions Cichla are important food and sport fish and have been introduced into several tropical and subtropical regions (Zaret 1980a; Welcomme 1981, 1988; Lever 1996; Fuller et al. 1999). Cichla have been introduced into experimental ponds in Alabama (Swingle 1967), Florida (Ogilvie 1966), and Georgia (Ogilvie 1966) and into power plant cooling reservoirs in Texas (Shafland 1995; Fuller et al. 1999). These introductions did not result in established populations. Cichla have been intentionally stocked by fisheries agencies

PAGE 37

29 and are established in open waters of Florida, Hawaii, and Puerto Rico (Erdman 1984; Shafland 1995; Fuller et al. 1999). In Hawaii and Puerto Rico, there were no native freshwater fish faunas to interact with introduced Cichla Peacock cichlids were introduced as sport fish into reservoirs along with other nonindigenous species such as centrarchids ( Lepomis spp. and largemouth bass Micropterus salmoides ) and cichlids ( Tilapia and Oreochromis spp.) (Erdman 1984; Devick 1980; Lilyestrom et al. 1994). Although the dynamics of other species were not reported, peacock cichlids and largemouth bass were not considered to significantly compete in Hawaii (Devick 1980) or Puerto Rico (Lilyestrom et al. 1994; Lilyestrom and Churchill 1996). Peacock cichlids were intentionally introduced into freshwater canals in southeast Florida in 1984-1987 (Shafland 1995). The primary reasons for introducing peacock cichlids into Florida were 1 ) unbalanced prey-to-predator biomass ratios (sensu Swingle 1950) resulting from the high abundance of illegally introduced cichlids, particularly spotted filapia Tilapia mariae, and 2) to supplement recreational fisheries that were based on largemouth bass (Shafland 1 995). Studies evaluating the effects of the peacock cichlid introducfion have not found evidence of negative effects on nafive fish assemblages, including largemouth bass (Shafland 1999a, c; Chapter 6), or on fishery values (Shafland and Stanford 1999). The Florida peacock cichlid fishery is esfimated at US$ 8 million annually (Shafland and Stanford 1999). On the other hand, the introducfion of Cichla into Lake Gatun, Panama, is one of the most frequently cited examples of the danger of fish introducfions. Cichla accidentally introduced into the Chagres River expanded downstream into the Panama

PAGE 38

30 Canal reservoir in the 1960s (Zaret and Paine 1973). Once in the reservoir, Cichla populations expanded and this novel predator apparently decimated populations of small native fishes (Zaret and Paine 1973). The introduction also correlated with changes in zooplankton, insect, and bird populations (Zaret and Paine 1973). However, native fish populations may have rebounded since their initial reduction (Welcomme 1988). Limiting Factors on Cichla Distribution Cichla are fi-eshwater fishes with an upper lethal salinity of about 1 8 ppt (Shafland and Hilton 1986). This factor would limit the ability of peacock cichlids to disperse via river mouths, estuaries, and salt marshes. Additionally, Cichla are true tropical fishes and cannot survive cold temperatures. The lower lethal temperature for Cichla in laboratory studies was about 15 C (Swingle 1967; Guest et al. 1980). Indeed, low winter temperatures resulted in the elimination of reproducing populations of purposefially introduced Cichla in some Texas power plant cooling reservoirs (Fuller et al. 1999). Experimental stocks of Cichla also have succumbed to cool temperatures in Florida (Ogilvie 1966; Shafland 1995; pers. obs.). Salinity and temperature limit the distribution of Cichla in Florida to fi-eshwater canals and borrow lakes in the southeast portion of the state (Shafland 1995; see also Shafland and Pestrak 1982). Moreover, winter temperatures would make the establishment of Cichla in other portions of the continental USA unlikely in the absence of reliable thermal refuges. The possible limiting effects of water chemistry parameters (e.g., pH, hardness) have not been investigated. However, C. ocellaris appears to be adaptable (i.e., survives and reproduces) to a wide range of water conditions (pers. obs.). Additionally, unlike some established nonindigenous fishes in Florida (e.g., walking catfish Clarias batrachus),

PAGE 39

31 Cichla are not capable of breathing air or crossing land barriers. However, because they are highly-favored sport fish, humans are likely to illegally transport Cichla to new waterbodies. Ecological Interactions Theoretically, peacock cichlids could have several ecological interactions with other species based on the dynamics of predation and competition. These interactions could have negative, positive, or mixed consequences for native species. Cichla are predatory and can consume small native fishes, including juvenile largemouth bass (Hill 1998). Indeed, in Florida, fish comprise 92% by number of the diet of peacock cichlids (Shafland 1 999b). The effect of predation by Cichla could be strong — this was implicated in the observed changes in native fish abundance in Lake Gatun, Panama (Zaret and Paine 1973). Furthermore, there are many other examples of significant changes in native fish abundance following the introduction of piscivorous fish (e.g., Kaufman 1992; Mills et al. 1994; see also Taylor et al. 1984). In addition to the direct effect of predation on native fishes, increased predation risk due to Cichla could force vulnerable prey (including juvenile largemouth bass) to refuge and potentially compete with each other for food within the limited refuge area (Mittelbach 1984, 1988; Osenberg et al. 1994). These direct and indirect effects of predation could reduce the abundance of largemouth bass and other native fishes. Nevertheless, predation by peacock cichlids on abundant nonindigenous cichlids (e.g., spotted tilapia), should have benefits for native fishes. Indeed, prey fishes can compete with early life stages of piscivores (e.g., largemouth bass) (Osenberg et al. 1 994; Olson et al. 1995) and many prey fishes also are nest or larval fish predators (Bain and

PAGE 40

32 Helfrich 1983; Popiel et al. 1996; Trexler et al. 2000). As a result, the introduction of a nonindigenous fish species that reduces the density of competitors or egg predators might actually lead to increased population sizes of native fish. Interestingly, peacock cichlids were introduced, in part, to reduce numbers of small nonindigenous fishes (Shafland 1995). Thus this possible positive effect of peacock cichlids was an implicit part of the original management goal. Some correlative data support this outcome within the Florida canal system (Shafland 1995, 1999a). For example, the introduction of peacock cichlids into Black Creek Canal was associated with a decline in the mean prey-to-predator biomass ratio fi-om 21.1 (1983-1988) to 10.0 (1988-1991) and a 36% decline in the biomass and 55% decline in numerical abundance of spotted tilapia (the most abundant exofic species) fi-om 1983-1988 compared to 1988-1993 (Shafland 1999a). Although suggesfive, there are no data to help evaluate if these trends (i.e., decreased prey-topredator biomass ratios and reduced abundance of nonindigenous spotted tilapia) were caused by peacock cichlids. The ontogenetic dietary shifts of peacock cichlids could lead to competition with various life stages of native fishes for zooplankton, insects, decapod crustaceans, or fish (Chapter 4). This mechanism could lead to reductions in native fish abundance following the introduction of Cichla For example, peacock cichlids and largemouth bass overlap in diet (Shafland 1999b; Chapter 4; L. G. Nico, USGS, and J. E. Hill unpubl. data), which might indicate the potential for competition mediated through reductions in prey fish abundance, shifts in community composition, or changes in prey habitat use or behavior (Soluk and Collins 1988; He and Kitchell 1990; Hambright et al. 1991; Soluk 1993;

PAGE 41

33 Persson et al. 1 996). Nevertheless, the indirect effects of peacock cichhd predation on native fishes have not been investigated. Peacock cichhds also might facilitate largemouth bass (or other native predators) by altering the behavior of their prey. For example, largemouth bass and peacock cichlids have different hunting strategies and activity periods. Peacock cichlids often rove in small schools and rely on speed rather than ambush as a hunting strategy (Erdman 1969; Devick 1972; Shafland 1995). Moreover, peacock cichlids are diurnal predators with different primary activity periods than largemouth bass (Lowe-McConnell 1969; Zaret 1980a; Lilyestrom and Churchill 1996). As a result of these differences, individuals of prey species may be presented with conflicting behavioral choices regarding activity patterns (e.g., foraging and refiiging), perhaps enhancing the hunting success of one or both predators (Soluk and Collins 1988; Soluk 1993). This possibility has not been studied for this system. Peacock cichlids might enhance populations of native predators by serving as prey themselves. For example, largemouth bass prey on juvenile peacock cichlids (Shafland 1999b). If largemouth bass dynamics are limited by food for the larger, piscivorous life stages, then the introduction of a new food source (i.e., peacock cichlids) could benefit largemouth bass. Although Florida canals often have high prey biomass, perhaps suggesting that food limitation is not very important, much of this potential food is morphologically unavailable to largemouth bass. Gape size (i.e., throat width) limits the maximum size of prey for largemouth bass (Lawrence 1 958; Hambright 1 99 1 ) and many canal prey species rapidly grow out of vulnerable sizes (Shafland et al. 1985; Hill 1998; L. G. Nico, USGS, unpubl. data). By breeding throughout the warm months

PAGE 42

34 (Shafland 1 999b), peacock cichlids provide numerous, appropriately-sized offspring for largemouth bass predation throughout the year (Chapter 4; pers. obs.). Small peacock cichlids are relatively abundant (Shafland 1999a, c; Chapter 4), are vulnerable (i.e., small, elongate, and with weak spines; Hill 1998) to a wide range of sizes of largemouth bass, and comprise 9% (by number) of prey consumed by largemouth bass in Tamiami Canal (Shafland 1999b). Spawning site competition via territorial aggression has been postulated as a significant negative centrarchid-cichlid interaction (Hogg 1976; Taylor et al. 1984; Noble and Germany 1986). However, evidence of this mechanism is largely anecdotal or correlative and no quantitative investigation has demonstrated spawning site competition for these fishes. Centrarchids would be the most likely native group to experience spawning site competition from peacock cichlids (or other cichlids) due to general similarities in reproductive behavior (Heidinger 1975; Zaret 1980a; Annett et al. 1996). However, in a study of the effect of blue tilapia Oreochromis aureus on largemouth bass reproduction in experimental ponds, non-nesting (i.e., all female) blue tilapia had a similar effect on largemouth bass reproduction as did all male or mixed sex treatments (i.e., nest-building and territorial) (Shafland and Pestrak 1983). This result suggests that spawning site competition is not the mechanism responsible for observed reductions of largemouth bass recruitment that occurred in some lakes with introduced blue tilapia (see Taylor et al. 1984; Germany and Noble 1986). Additionally, there is relatively small overlap in spawning times of largemouth bass and peacock cichlids (Shafland 1999b; pers. obs.) and little agonistic behavior has been observed between the two species (P. L. Shafland. FWC, pers. comm.; Chapter 5; pers. obs.). However, there is temporal, and

PAGE 43

35 perhaps spatial, overlap in spawning activity between peacock cichlids and other centrarchids (e.g., sunfish Lepomis spp.) (pers. obs.). Of course, negative and positive effects of peacock cichlids could both exist. Moreover, these effects may not be spatially or temporally consistent. Depending upon the direction and intensity of these effects, the outcome for native fishes could be mixed (e.g., stage or context dependent), or even cancel. For example, peacock cichlids could enhance juvenile largemouth bass growth by reducing numbers of competing fishes while simultaneously decreasing adult largemouth bass growth by depleting numbers of available prey. This type of effect is documented for bluegill Lepomis macrochirus and largemouth bass in temperate lakes (Olson et al. 1995). Nevertheless, additional research is required to document mixed effects in this system. Fishing Effects The influence of peacock cichlids on sport fishing could have mixed effects for the other major canal sport fish — largemouth bass. Combined recreational fishing values for largemouth bass and peacock cichlids (the most commonly targeted and valuable sport fish) in southeast Florida canals exceed US$ 13 million annually (Shafland and Stanford 1999). Fishing tactics for both species are similar, although some important differences occur (Hidalgo 1997). These differences (e.g., fishing time of day and lure choice) are exploited by anglers specifically targeting one or the other species (Hidalgo 1 997). Given the high angler effort and increasing popularity of peacock cichlids in southeast Florida canals (Shafland and Stanford 1999), exploitation of largemouth bass could be increasing. This could result fi-om by-catch of anglers targeting peacock cichlids and by the promotion of canal fishing resulting from the attention that peacock

PAGE 44

36 cichlids have generated. On the other hand, the introduction of peacock cichHds also might reduce exploitation mortality of largemouth bass. Because of the differences in fishing for the two species, peacock cichlid fishing may actually redirect angler effort away from largemouth bass and therefore reduce largemouth bass total mortality. This interesting question has not been investigated. Qualitative Risk Assessment The potential of nonindigenous species to be introduced, become established, and become invasive can be qualitatively modeled given information concerning introduction pathways and the biology and ecology of the organism in question (RAM Committee 1998). There are two main sections of the risk model: 1) Probability of Establishment and 2) Consequences of Establishment. An overall Organism Risk Potential (ORP) is the final output of the model (i.e., low, medium, or high) and is calculated based on the risk estimates of both sections. Within these sections of the risk model there are seven rating elements, each assigned a qualitative estimate of risk (i.e., low, medium, and high) and an estimate of certainty (i.e., very certain, reasonably certain, moderately certain, reasonably uncertain, and very uncertain). Four of the seven elements address the Probability of Establishment and three elements attempt to predict the Consequences of Establishment. Rating Elements of Risk Model — Probability of Establishment 1) Estimate probability of the exotic organism being on, with, or in the pathway. The peacock cichlid is established in southeast Florida (Shafland 1995) and is thus in the United States Pathway. Various members of the genus Cichla are occasionally imported in the aquarium industry (pers. obs.). State and commonwealth

PAGE 45

37 agencies have imported or exchanged peacock cichHds for fishery introductions (Maciolek 1984; Shafland 1995; Fuller et al. 1999). High Very Certain 2) Estimate probability of the organism surviving in transit. The peacock cichlid has survived transit on numerous occasions. High Very Certain 3) Estimate probability of the organism successfully colonizing and maintaining a population where introduced. The peacock cichlid has been reproducing in southeast Florida since the mid-1980s and is considered established in the state (Shafland 1 995). Cichla have been intentionally introduced and established reproducing populations in Hawaii and Puerto Rico in reservoirs (Fuller et al. 1999). Nevertheless, peacock cichlids have been unsuccessful at establishing persistent populations in Texas power plant cooling reservoirs and possibly in southern Florida in the 1960s due to cool temperatures (Fuller et al. 1999). A congener, the speckled peacock cichlid C. temensis was intentionally introduced alongside C. ocellaris yet failed to become established in southeast Florida (Shafland 1995). Because of thermal constraints, it is unlikely that Cichla could become established in the continental USA outside of southeastern Florida (see Limiting factors on Cichla distribution, above). High Very Certain 4) Estimate probability of the organism to spread beyond the colonized area. The initial introduction area of the peacock cichlid nearly encompasses its present range in Florida. Peacock cichlids have successfully colonized adjacent canals and have been illegally or naturally transplanted into artificial lakes and ponds in southeast Florida. Limifing factors for peacock cichlids include a lower lethal temperature of about 1 5 C (Swingle 1967; Guest et al. 1980) and an upper salinity tolerance of about 18 ppt (Guest

PAGE 46

38 et al. 1980; Shafland 1995). Relatively warm winters allow peacock cichlids to temporarily colonize canals slightly north and west of the targeted area of introduction (Shafland 1995; Loftus 2000; L. G. Nico, USGS, pers. comm.). Nevertheless, given the inability of peacock cichlids to survive cool winter temperatures or salt water, there is little possibility for significant expansion outside of the present range in southeast Florida. Additionally, due to cool winter temperatures, it is unlikely that Cichla could become established in other areas of the continental USA. Low Very Certain Rating Elements of Risic Model — Consequences of Establishment 5) Estimate economic impact if established. The peacock cichlid could negatively affect the largemouth bass recreational fishery in southeast Florida via predation and competition for food (see Ecological interactions, above). Moreover, peacock cichlid fishing may have negafive effects on largemouth bass populafions (see Fishing effects, above). Also, peacock cichlids theoretically could enhance largemouth bass populations via ecological and fisheries mechanisms (see Ecological interacfions and Fishing effects, above). However, studies investigating the effects of peacock cichlids on largemouth bass have not reported significant reductions in largemouth bass abundance (Shafland 1999a, c; see also Chapter 6). The largemouth bass fishery in the area occupied by peacock cichlids has been estimated at US$5 million annually (see Shafland and Stanford 1999). Major objectives for introducing peacock cichlids into Florida were to create a novel sport fishery and to supplement recreational fisheries for largemouth bass (Shafland 1995). These objectives have been met and the recreational fishery has been estimated at US$8 million annually (combined fishery of largemouth bass and peacock

PAGE 47

39 cichlids of US$13 million) (Shafland and Stanford 1999). Therefore, the overall economic effect has been positive for the region of introduction. The economic risk is estimated to be low due to the sustained values of largemouth bass fishing where peacock cichlids have been present since 1984-1987 (Shafland and Stanford 1999) and the relatively small potential effect on statewide fishery values (i.e., the small geographic area occupied — portions of three south Florida counties) regardless of any possible negative impacts on the largemouth bass fishery. Additionally, the similarity of largemouth bass and peacock cichlid fishing (i.e., similar equipment and techniques; Hidalgo 1997) suggests that largemouth bass anglers, guides, and tackle shops could adjust to any losses to the largemouth bass fishery by switching to peacock cichlids. Low Very Certain 6) Estimate environmental impact if established. There is a potential for peacock cichlids to adversely affect native fishes by predation and competition for food (see Ecological interactions, above). Predatory fishes can exert top-down effects that may alter trophic relations and energy flow in aquatic systems (Zaret and Paine 1973; Carpenter et al. 1987; Northcote 1988; Mittelbach et al. 1995). Nevertheless, it is also possible that peacock cichlids may positively affect native fishes by reducing numbers of other exotic fishes and by serving as prey themselves. Current data indicate the operation of mechanisms by which peacock cichlids negatively effect native species (e.g., predation), but also suggest that peacock cichlids have not significantly reduced native fish populations in southeast Florida (Shafland 1995, 1999a, c; Chapter 6). Furthermore, peacock cichlids are confined to artificial water bodies within a small geographic area that lacks highly specialized or endemic nafive fishes. Based on exisfing information, the

PAGE 48

40 risk of environmental harm is relatively low, but risks may increase if conditions within the system significantly change. One possible change that could increase the risk of environmental harm from peacock cichlids is climatic change. An increase in mean low temperatures in southern Florida might lead to an expansion of the range of peacock cichlids. Nevertheless, peacock cichlids are less tolerant of cool temperatures than are many established tropical fishes in Florida (Shafland and Pestrak 1982; Shafland 1995) and experience winterkills even within the present thermal refuge of the canal system, (P. L. Shafland, FWC, pers. comm.), a region with mean monthly low air temperatures in January (Florida's coldest month) of > 1 8.3 C (Shafland and Pestrak 1 982). LowModerately Certain 7) Estimate impact from social and/or political influences. Peacock cichlids have had fairly broad approval and support from within the angling community before and since the introduction in the mid-1980s. The recreational fishery is popular and supports important socio-economic activity (Shafland and Stanford 1999). A few anglers, environmental groups, and some scientists, particularly ecologists, express concern over the peacock cichlid. Anglers against the introduction are concerned with possible effects on the largemouth bass fishery. Environmental groups and many ecologists are against the introduction of any non-native species, their concerns generally based on the theoretical or potential effects of nonindigenous species on native species (e.g., Courtenay and Robins 1989; Minns and Cooley 2000). An additional concem for environmental groups and governmental agencies is the possibility of peacock cichlids invading the nearby Everglades National Park. Given the widespread acceptance and

PAGE 49

41 support of peacock cichlids in south Florida, the overall effect is estimated to be positive. Low Moderately Certain Organism Risk Potential (ORP) The Organism Risk Potential (ORP) has two components: 1) the Probability of Establishment and 2) the Consequences of Establishment. The Probability of Establishment is the lowest ranking amongst the first four ratings elements of the risk model (i.e., within pathway, entry potential, colonization potential, and spread potential). Given the fact that peacock cichlids are established in Florida, intuitively this probability should be rated as "High". Nevertheless, because the potential for spread is rated as "Low", the overall Probability of Establishment from the model is rated as "Low" (Figure 3). The fourth rating element (i.e., spread potential) was estimated as "Low" because the area of introduction (i.e., the area of canals where peacock cichlids were stocked), the intended target area of the introduction, and the potential maximum range of colonization for peacock cichlids in the continental USA (i.e., Florida) are very similar (i.e., about 650 km of canals; Shafland and Stanford 1 999; see Limiting factors on Cichla distribution, above). The Consequences of Establishment rating is defined as the highest of the probabilities assigned to the economic and environmental impacts. Perceived impact (i.e., social and polifical) does not influence the assessment of Consequences of Establishment unless the economic and environmental risk estimates are both "Low". All three elements were estimated to be "Low" for the potential for negative consequences. Therefore, the Consequences of Establishment are estimated to be "Low" (Figure 3).

PAGE 50

42 Probability Organism Entry Colonization Spread of within Potential Potential Potential = Low Establishment Pathway (High) (High) (High) (Low) Consequences of Economic Environmental Perceived = Low Establishment (Low) (Low) (Low) Probability of Consequences of ORP Risk Establishment Establishment = Low (Low) (Low) Figure 3. Organism Risk Potential (ORP) for peacock cichlid Cichla ocellaris within the continental USA based on the RAM methodology (RAM Committee 1998), given a "Low" Probability of Establishment. The Probability of Establishment is the lowest rating among the four rating elements (i.e., within pathway, entry potential, colonization potential, and spread potential). The Consequences of Establishment is the highest rating between economic and environmental ratings. The ORP is the average of Probability of Establishment and Consequences of Establishment ratings. An ORP of "Low" is defined as "acceptable risk organism(s) of little concern (does not justify mitigation)" (RAM Committee 1998). The ORP rating is the average of the two risk potentials assigned for the Probability of Establishment and the Consequences of the Establishment, with the average rounded up (i.e., a conservative estimation). Therefore the ORP for peacock cichlids introduced into the continental USA is "Low" (i.e., "Low" Probability of Establishment and Consequences of Establishment) (Figure 3). Conversely, had the risk of spread rating element been disregarded, then the Probability of Establishment would have been assigned a rating of "High" and the ORP would have been "Medium" (i.e.,

PAGE 51

43 average of "High" Probability of Establishment and "Low" Probability of Consequences). An ORP of "Medium" is considered unacceptable risk and would require mitigation (RAM Committee 1998). Discussion Risk assessments attempt to predict effects of nonindigenous species and are used as a basis for economic, environmental, and policy decisions. As a result, it is important to evaluate the methods used to conduct risk assessments to determine if risk assessments yield accurate predictions and thus are useful tools. Although the peacock cichlid is well-established in southeast Florida (Shafland 1995), the RAM method algorithm resulted in an estimate of the Probability of Establishment as "Low". This somewhat nonsensical outcome of the model (i.e., rating the probability that an established species would become established as low) was due to the similarity in the geographic area of the initial introduction and the small potential maximum range in the continental USA (i.e., there is little chance of range expansion). Previous risk assessments using this methodology have been conducted on fishes with large potential ranges in the USA (i.e., black carp, Asian swamp eel, and snakeheads). Therefore, the situation with peacock cichlids may represent a relatively unusual case — one not adequately anticipated in the methodology. Perhaps any apparent conflict over the assignment of a "Low" risk of establishment is semantic. The Probability of Establishment label could be modified to Probability of Establishment and Spread to avoid confusion in such cases (see also Kolar and Lodge 2002). Regardless of the label, the essence of this risk element (i.e., Probability of Establishment) is the potential for an organism to access the region of interest, and then survive, reproduce, and spread once

PAGE 52

44 there. Additionally, this issue suggests that each of rating elements should be considered separately as well as in combination, and that any method of risk assessment requires some level of flexibility in interpretation, factors acknowledged by the authors of the methodology (RAM Committee 1998). The qualitative risk assessment methodology combines estimates of risk for two very different processes into a single estimate of the risks posed by a nonindigenous species (i.e., establishment and consequences). The assignment of the average risk rating between these two processes as the overall risk potential that this organism represents (i.e., ORP) can be misleading. An organism may have a low probability of establishment but a high potential for adverse effects and have the same Medium ORP rating as a species with a high potential for establishment but a very low probability of negative consequences. A Medium ORP is defined as an unacceptable risk, justifying mitigation (RAM Committee 1998). A logical argument would be to place more emphasis on the potentially harmfial species even though there is a relatively low chance of becoming established because of the greater danger posed by a successful invasion. Indeed, it might be a waste of governmental resources and an undue economic burden to attempt to exclude species that have low potential for environmental or economic harm regardless of their perceived ability to colonize. It may be more appropriate to consider the two processes independently and do away with or otherwise de-emphasize the ORP because the ORP rating seems to lose information rather than simply condense it. Two alternative ORP estimates for peacock cichlids (i.e., "Low" [Figure 1] and "Medium") were presented in this risk assessment. The estimate of "Low" was based on a strict interpretation of the RAM method algorithm. The estimate of "Medium" risk

PAGE 53

45 logically could be argued based on the fact that peacock cichlids are established in the continental USA (i.e., should have a "High" Probability of Establishment). An ORP result of "Medium" risk would place the peacock cichlid into the category of unacceptable risk (RAM Committee 1 998). However, it seems illogical to require mitigation against a socio-economically important sport fish when there is no evidence that it has harmed native fish species in the nearly two decades since it has been intentionally introduced (Shafland 1995, 1999a,c; Shafland et al. 2001; Chapter 6). The RAM algorithm can only return a "Medium" or "High" ORP if a nonindigenous species is established or has a medium or higher probability of becoming established (RAM Committee 1 998). This presupposes that if established, any nonindigenous organism will have an unacceptable risk of negative impacts. This inherent bias prejudices a supposedly objective process to lead to exclusion or restriction of any species that may become established. If this strict precautionary principle (FAO 1 995) is the foundation for the RAM methodology, then the entire assessment of risk hinges on the estimate of the ability of the organism to become established. The section concerning the possible consequences to the environment or economy become of no importance, except perhaps in defining regulatory, eradication, or control priorities. For example, the upcoming risk assessment of bighead carp and silver carp must conclude that these species represent an unacceptable risk and require mitigation (i.e., the ORP must be either "Medium" or "High") due to the fact that both already are established in the Mississippi River (and could spread to other regions), regardless of the assessor's assignment of risk associated with the presence of these species. Therefore, the outcome of the risk assessment for these species is largely predetermined.

PAGE 54

46 Regardless of any inherent bias, the predictive abihty of any risk assessment tool can only be as good as the information put into the model. Although it provides a meaningful way of approaching the problem and attempts to interject much-needed objectivity into the field, the method still leaves substantial room for personal interpretation and subjectivity (i.e., "uncertainty of the assessor"; RAM Committee 1998). For example, everyone will not agree with my estimates of risk and certainty for peacock cichlids or other species. Some disagreements will be over interpretation of data, others over philosophy. In particular, some ecologists would rate the environmental risk of peacock cichlids as high due to demonstrated predation and the potential for competition affecting native species regardless of existing data. Furthermore, data are lacking for many non-fish components of southeast Florida's aquatic systems, increasing uncertainty. Although empirical rules of bio-invasion have been proposed that seek to provide a framework for prediction based on past invasions (Moyle and Light 1996a), their universality seems doubtful and the performance of many nonindigenous species has been unpredicted (Moyle et al. 1986). Indeed, in any risk assessment process, a major impediment is uncertainty associated with the organisms (RAM Committee 1 998). In the case of peacock cichlids, in contrast to other fishes assessed by this methodology, there are fairiy substantial data on introduced populations within the region of concern (i.e., continental USA) (Shafland 1995, 1999a, b, c; Hill 1998; Shafland and Stanford 1999; Shafland et al. 2001; Chapter 4, 5, and 6), thereby increasing the estimates of certainty (RAM Committee 1998). In the RAM methodology guide, it is argued that the actual ratings assigned to individual elements are not as important as the transparent presentation of information

PAGE 55

47 upon which the ratings are based (RAM Committee 1 998). AUhough this is an important point, the fact that the algorithm produces a single term, the ORP, as a guide to decision making (i.e., acceptability of risk and requirement of mitigation) places a distinct emphasis on the ORP as the end product of the risk assessment. This factor, coupled with the previously mentioned bias relative to any species capable of becoming established (i.e., a de facto assumption that all established nonindigenous species represent an unacceptable risk), somewhat limit the utility of the process. Indeed, it could be argued that this process merely adds a pseudo-quantitative shroud to the subjective method of expert opinion. On the other hand, the process does provide a vehicle for gathering information on the focal species and forces a qualitative evaluation of various categories of risk. In conclusion, it is clear that the peacock cichlid test case pointed out problems associated with the RAM methodology. Established species with small potential ranges yield a nonsensical result for the Probability of Establishment. This could be largely addressed by changing the label to Probability of Establishment and Spread. The combining of probabilities of a species becoming established with subsequent consequences of the establishment into an ORP output places too much emphasis on a single value and loses important information in the process. This could be addressed by reducing the prominence of the ORP (or eliminating it altogether) and retaining the Probability of Establishment and Spread along with the Consequences of Establishment as the primary outputs of the RAM model. Regulators and managers should have clear objectives in mind prior to commissioning a risk assessment and should determine if the threat of successful introduction alone is the primary concern or if negative effects of the

PAGE 56

48 nonindigenous species are more (or equally) important. Additionally, the built-in assumption of the RAM model that all nonindigenous species capable of establishing persistent populations represent an unacceptable risk should be acknowledged. The RAM guide states that a risk assessment cannot be used to determine an acceptable level of risk, yet the ORP ratings are defined explicitly in terms of acceptable and unacceptable risk (RAM Committee 1998). The level of acceptable risk is not a question for the assessor, but one for society, and definitions should be changed to remove conclusions of acceptability. Also, conclusions of the need for mitigation are not needed, although suggestions for mitigation (or eradication or control) may be appropriate within the risk assessment. In essence, the definitions of "Low", "Medium", and "High" risk should be self-explanatory and interpreted within the context of the stakeholder environment. Lastly, additional test cases should be assessed to determine if other biases occur in the methodology.

PAGE 57

CHAPTER 4 ONTOGENETIC DIETARY SHIFTS AND DIETARY OVERLAP OF A NONINDIGENOUS FISH AND AN ECOMORPHOLOGICALLY SIMILAR NATIVE FISH Introduction Changes in resource use during ontogeny have considerable influence on the dynamics of size-structured populations (Werner and Gilliam 1984). For piscivorous fishes, ontogenetic diet shifts can be dramatic. For example, the largemouth bass Micropterus salmoides typically progresses through relatively discrete stages of feeding on zooplankton, insects, and decapod crustaceans before becoming largely piscivorous (Keast and Eadie 1985; Olson 1996; but see Schramm and Maceina 1986; Cailteux et al. 1996; Shafland 1999b; Huskey and Tumigan 2001 for studies reporting continued and substantial consumption of decapod crustaceans in addition to fish). These dietary shifts are size-related and dependent upon the relative size of the predator's gape and the prey's body depth or girth (Hambright 1991). For gape-limited species, growth-related increases in gape size allow for capture of larger, more energetically profitable prey (e.g., inifiation of piscivory; Pasch 1975; Ludsin and DeVries 1997). Indeed, individual growth is enhanced by switching to a fish diet (Pasch 1975; Timmons et al. 1980; Olson 1996). Moreover, increased growth also leads to increased survival of age-0 largemouth bass because mortality is strongly size-dependent (Gutreuter and Anderson 1985; Ludsin and DeVries 1997). 49

PAGE 58

50 Trophic morphology, including gape width, is considered to be an indicator of a fish's ability to handle prey (Wainwright and Richard 1995; Wainwright 1996). Indeed, numerous studies have investigated gape width and its relationship to prey use in largemouth bass (e.g., Lawrence 1958; Hambright 1991; Hambright et al. 1991). Given this link between trophic morphology and prey use, species that are morphologically similar might be expected to exhibit similar ontogenetic dietary shifts and have considerable overlap in diet. An ecological implication of this is that if prey are limiting and are depleted by predators, then competition will occur between species that significantly overlap in diet. The peacock cichlid Cichla ocellaris (Cichlidae) was intentionally introduced into the native range of largemouth bass (Centrarchidae) in southeast Florida in 1984-1987 (Shafland 1995). Norton and Brainerd (1993) described these species as ecomorphologically similar based on morphology, diet, and feeding mechanics. Moreover, Hill (1 998) found broad similarity in length-specific gape width between these species. Although the ontogenetic dietary patterns of largemouth bass diet are well known (Keast and Eadie 1985; Olson 1996), little information is published on the diet of juvenile Cichla (Lowe-McConnell 1969; Zaret 1980a; Lilyestrom and Churchill 1996; Shafland 1999b). Nevertheless, available information and the similarity in morphology of these species imply that largemouth bass and peacock cichlids may have similar diets through ontogeny. Moreover, preliminary observations of canals in southeast Florida suggested that juvenile peacock cichlids were abundant and used similar habitats as juvenile largemouth bass (pers. obs.).

PAGE 59

51 Studies documenting ontogenetic dietary stages or dietary overlap of naturally cooccurring fishes are common (e.g., Nico and Taphom 1988; Winemiller 1989, 1991; Murie 1995; Scott and Angermeier 1998). Additionally, much research has investigated shared resource use of introduced and native fishes, including species that are morphologically similar (e.g., Mathur 1977; Parrish and Margraf 1990; Matthews et al. 1992; Huckins 1997; Huckins et al. 2000; Declerck et al. 2002; Sutton and Ney 2002; Stoffels and Humphries 2003). However, studies are rare that compare dietary ontogeny and estimate the overlap in diet of a nonindigenous fish with an ecomorphologically similar, but phylogenetically distant, native fish. This study investigated the hypothesis that an introduced fish will have an ontogenetic dietary progression similar to that of a native species of similar morphology, regardless of their phylogenetic relations. Additionally, this study documented the temporal co-occurrence of juvenile largemouth bass and peacock cichlids of similar size and ability to handle prey, and estimated the extent to which the two species overlapped in prey use in two southeast Florida canals. Methods Collections for this study were made in Cutler Drain Canal (C-IOOC) and Snapper Creek Canal (C-2), Biscayne Bay Drainage, Miami-Dade County, Florida. Both canals occur in urban and suburban areas of the Greater Miami Metropolitan Area. Cutler Drain Canal averages about 25 m wide and 3^ m deep. It had highly variable aquatic macrophyte species composition and abundance due to biological and mechanical weed control methods used by the South Florida Water Management District (SFWMD). In sampled reaches, aquatic macrophyte coverage ranged from about 2% percent areal

PAGE 60

52 coverage (PAC) to > 75% PAC. Asian hygrophila Hygrophila polysperma was the dominant submersed macrophyte. Water clarity was generally > 3 m of Secchi depth. Collections were made from the S-1 19 water control structure (25 38.58' N, 80 20.31' W) upstream for about 1 .6 km. Snapper Creek Canal is a large, important conduit for water drainage. Collections were made from the Don Shula Expressway bridge (downstream of the boat ramp on Snapper Creek Drive [25 42.00' N, 80 21.25' W]), upstream along the main canal and in two adjacent lateral canals accessed near the crossing of SW 99 Avenue — a total of about 1 .8 km. In sampled reaches, aquatic macrophyte abundance ranged from < 5% PAC to > 50% PAC in the main canal and from < 20%) PAC to near 100%) PAC in the laterals. The main canal had highly variable aquatic macrophyte abundance, with hydrilla Hydrilla verticillata being the dominant species. The lateral canals had relatively little mechanical harvesting of weeds by the SFWMD, and therefore had relatively stable aquatic macrophyte abundance. Hydrilla and fanwort Cabomba caroliniana were the dominant species in the lateral canals. Muskgrass Chara sp., a macroalgae, also was common in the southern lateral canal. Snapper Creek Canal had highly colored and occasionally turbid water (Secchi depth usually 1-2 m). Four collections were made in Snapper Creek Canal in 1 999 (July, September, October, and November) and two collections were made in both canals in 2001 (August and October). Juvenile largemouth bass and peacock cichlids were collected by daytime boat electrofishing. An attempt was made to collect all individuals < 150 mm TL until target numbers were reached, 25 (in 1999) or 50 (in 2001). Fish were euthanized and

PAGE 61

53 preserved in formalin. Formalin was injected into the abdominal cavity of specimens collected in 2001. In the laboratory, specimens were transferred to ethanol. All fish were measured for maximum total length (TL), and a representative sample covering the collected size range of largemouth bass and peacock cichlids was measured for gape width (GW). Gape width measurements were made as external mouth width with the mouth closed to approximate intercleithral distance (Lawrence 1958; Hambright 1991; Hill 1998). Wainwright (1996) determined that intercleithral distance is the limiting dimension for swallowing prey in centrarchids. External mouth width has an approximate 1:1 relationship to intercleithral distance in largemouth bass (Lawrence 1958). Additionally, Hill (1998) evaluated this method for peacock cichlids and found that peacock cichlid prey maxed out at about 96.5% of predicted gape size (maximum sizes ranged from 88104% with a SD of 5.6%). Visceral masses of all fish were removed, stomachs were opened with forceps, and prey items were removed, identified, and counted with the aid of a dissecting microscope. Only items occurring in the stomach portion of gastro-intestinal tract were examined and used for diet analyses. All prey measurements and predator GW measurements were made with dial calipers. Predator total-length measurements were made with a measuring board. Fish prey were measured for maximum standard length (SL). Decapod crustaceans in the fish stomachs were found with abdomens folded underneath the cephalothorax and were therefore measured for folded body length (FBL; i.e., dp of rostrum to outside of bend in abdomen). Maximum body depth (BD) was measured for each prey if the specimen was intact. Decapod crustacean BD was

PAGE 62

54 measured with the abdomen folded underneath the cephalothorax. An estimate of the volume of prey in the stomachs was obtained by water displacement using either a graduated cylinder or a pipette. A qualitative estimate of stomach fullness also was recorded. The values ranged from 1 to 4, with 1 being empty, 2 having some food but less than 25% of the estimated stomach capacity, 3 having 25-75% of stomach capacity filled, and 4 having > 75% of estimated stomach capacity. Juvenile largemouth bass and peacock cichlids were grouped into 10-mm TL size classes. Estimates of length-specific gape width were used to construct histograms of gape size distribution for each sampling date to determine the extent of overlap in morphological ability of juvenile largemouth bass and peacock cichlids to handle prey. Estimates of gape width were obtained by regressing gape width measurements on total length (PROC GLM; SAS Institute, Gary, North Carolina). Analysis of covariance (ANCOVA) was used to test for differences in slopes and intercepts of developed regressions for largemouth bass and peacock cichlids (PROC GLM; SAS Institute, Gary, North Carolina). All statistical analyses were conducted at a significance level of a = 0.05 unless otherwise noted. The stomach contents of juvenile largemouth bass and peacock cichlids were described by frequency of occurrence, numerical abundance, and volume for all prey, as well as prey size for fish and decapod crustacean prey. A predator was defined as a piscivore when fish prey were found in at least 60% of predator stomachs that contained any food (Bettoli et al. 1992; Gailteux et al. 1996). Prey sizes (i.e., prey BD and prey length offish and decapod crustacean prey) were regressed on predator total length and the resulting relations were compared with ANCOVA (PROG GLM; SAS Institute, Gary,

PAGE 63

55 North Carolina). Also, relative prey size was estimated in two ways — prey body depth divided by predator gape width, and prey length (SL for fishes and FBL for decapod crustaceans) divided by predator total length, multiplied by 100 to obtain percentages. Estimates of stomach fullness were compared between species within sampling dates by the Wilcoxon rank sum test (PROC NPARIWAY; SAS Institute, Gary, North Carolina) and within species and among sampling dates by the Kruskal-Wallace test (PROC NPARIWAY; SAS Institute, Gary, North Carolina). A significant KruskalWallace test was followed by Dunn's non-parametric multiple comparison test (Hollander and Wolfe 1973). Due to the conservative nature of Dunn's test, a significance level of a = 0.10 was used to distinguish mean rank differences among sampling dates (Hollander and Wolfe 1973). Dietary overlap values between predators within sampling date were estimated using Schoener's Index (Schoener 1970): Cxy= 1 -0.5(X \Pxi-Pyi\), where C;^^ is the index value, p^i is the proportion of food / in the stomach contents of species x, and p^, is the proportion of food i in the stomach contents of species y. This index ranges from 0 (no overlap) to 1 (complete overlap). 1 defined proportions using either numerical abundances or prey volumes. Although subjective, values of dietary overlap > 0.6 have been considered as significant dietary overlap (Zaret and Rand 1971; Mathur 1 977). Schoener's index of dietary overlap has been recommended for use when resource availability data are lacking (Hurlbert 1978; Wallace 1981).

PAGE 64

56 Results A total of 548 fish were examined for stomach contents (247 largemouth bass and 301 peacock cichlids) (Figure 4). Of these specimens, 216 largemouth bass and 200 peacock cichlids had food in their stomachs. The results for stomach contents include only those individuals with food in the stomach. Largemouth Bass Peacock Cichlid ^ <^ \^
PAGE 65

57 striders (Gerridae; Hemiptera), and some terrestrial insects were present. A few terrestrial insects were found in largemouth bass stomachs, but most were aquatic insects. No terrestrial insects were found in peacock cichlid stomachs. All decapod crustaceans found in peacock cichlid stomachs and most in largemouth bass stomachs were grass shrimp Palaemonetes sp. Crayfish Procambarus sp. also were found in largemouth bass stomachs. Swamp darter Etheo stoma fusiforme and, to a lesser extent, bluefm killifish Lucania goodei were the dominant fish prey; poeciliids (i.e., Gambusia holbrooki and Poecilia latipinna ), bullhead catfish Ameiurus sp., loricariid catfish, cichlids (including peacock cichlids), and centrarchids also were consumed. Larval fish in the stomachs were yolk-sac fry and were not identified to species. Organic materials included welldigested animals, detritus, and plant material. Ontogenetic Dietary Shifts For largemouth bass, insects, and to a lesser extent, zooplankton and amphipods dominated frequency of occurrence for small individuals (i.e., < 70 mm TL) (Figure 5A). Zooplankton became less important and dropped out of the diet after the 80 mm TL size class. The longest largemouth bass with zooplankton was 86 mm TL. Amphipods declined in importance and were not found in fish larger than 130 mm TL. Insects remained frequent in the diet over the range of largemouth bass length examined in this study (< 1 70 mm TL), but were replaced by fish as the most frequently occurring prey category in specimens of 120 mm TL and larger. Decapod crustaceans occurred in only 13.7% of largemouth bass < 90 mm TL, but were commonly consumed (in 40.2% of stomachs) by largemouth bass > 90 mm TL. The smallest largemouth bass that consumed a fish was 67 mm TL. An increasing fraction of largemouth bass consumed

PAGE 66

58 Peacock Cichlid Total Length Group (mm) Zooplankton — o— Amphipod A Insect — — Decapod Rsh Figure 5. Frequency of occurrence (i.e., percent of predators with prey category in stomach contents) of main prey categories in juvenile (A) largemouth bass (N = 216) and (B) peacock cichlids (N = 200) electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida. The horizontal line at 60 % corresponds to the frequency of fish in the diet recommended by Bettoli et al. (1992) as the minimum frequency for categorizing a predator as piscivorous.

PAGE 67

59 fish as largemouth bass size increased. Nevertheless, the frequency of occurrence offish did not exceed 60% (i.e., piscivory as defined by Bettoli et al. 1992) until the 120 mm TL group. Peacock cichlids exhibited a similar pattern of diet change, although there were important differences, particularly regarding the extent of piscivory (Figure 5B). The smallest size class of peacock cichlids (i.e., 30-50 mm TL group) ate zooplankton at a high frequency. Zooplankton was frequent or moderately frequent in the diets of peacock cichlids up to the 100 mm TL group. The largest peacock cichlid with zooplankton in its stomach was 105 mm TL. Insects were eaten by a high frequency of peacock cichlids in the 60-80 mm TL groups, but rapidly declined in percent occurrence and dropped out of the diet by the 1 1 0 mm TL group. Decapod crustaceans were moderately frequent in the stomachs of peacock cichlids > 70 mm TL, and of high frequency in fish > 140 mm TL. The smallest peacock cichlid with a fish in its stomach was 54 mm TL. Fish were found in a low to moderate percent of peacock cichlid stomachs in the 30-80 mm TL groups, but rapidly increased in frequency with increasing peacock cichlid size. Piscivory (sensu Bettoli et al. 1992) was attained by peacock cichlids > 100 mm TL. By number, diets of small largemouth bass (i.e., < 70 mm TL) were dominated by zooplankton (Figure 6A). Additionally, zooplankton made up about half of the number of prey for largemouth bass in the 70 mm TL group. Amphipods made up only a small percentage by number of the prey of largemouth bass. Insect numbers were important in the 70 mm TL group and dominated the 80-1 30 mm TL groups. Moderate numbers of insects also were eaten by the largest size classes of largemouth bass. Decapod crustaceans were of relatively minor importance by number for largemouth bass > 90 mm

PAGE 68

60 TL. Fish were of low frequency by number in fish > 90 mm TL except for the 140+ mm TL groups where fish were dominant. A 100%^ 80%' E 3 z 60%>. •4-1 40%c 0) 20%' Q. 0%. SS5 I i Largemouth Bass Total Length Group (mm) Peacock Cichlid Total Length Group (mm) Zooplankton SAmphipod m Insect Q Decapod Fish Figure 6. Percent contribution to diet by number of main prey categories in juvenile (A) largemouth bass (N = 216) and (B) peacock cichlids (N = 200) electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida.

PAGE 69

61 Zooplankton was an important prey category by number for peacock cichlids up through the 1 00 mm TL size class (Figure 6B). Amphipods were relatively unimportant by number for peacock cichlids. Insects contributed most to prey numbers for peacock cichlids in the 80 mm TL group, but were relatively important by number for fish of 60110 mm TL. Decapod crustaceans were of low to moderate numbers in peacock cichlids of 70-140 mm TL, but were about 35% of the prey numbers for fish in the 140+ mm TL group. Fish were important by number for peacock cichlids beginning at the 1 00 mm TL group and were dominant for peacock cichlids > 1 1 0 mm TL. Zooplankton and amphipods contributed relatively little to the volume of stomach contents of largemouth bass in this study, except for the 30-50 mm TL group (Figure 7A). Insects made up over half of the volume of this size class (i.e., 30-50 mm TL group) and generally made up at least moderate amounts of the prey volume for largemouth bass in the 60-100 mm TL groups and again in the 130 mm TL group. The decapod crustacean category was an important constituent of prey volume for largemouth bass in the 70-130 mm TL groups. Fish prey made up a small to moderate amount of the prey volume for largemouth bass in the 60-100 mm TL groups but became dominant in all larger size classes. Zooplankton made up a substantial percent of the prey volume only for peacock cichlids 30-50 mm TL (Figure 7B). Insects contributed a relatively small amount to prey volume for peacock cichlids in the 30-50, 70, and 80 mm TL groups, and made up the largest amount of prey volume in the 60 mm TL group. Decapod crustaceans were dominant by volume in the 70 mm TL group but were of moderate to low volume in

PAGE 70

62 larger peacock cichlids. Fish were important volumetrically for peacock cichlids 30-80 mm TL, and were dominant for larger peacock cichlids. Largemouth Bass Total Length Group (mm) Peacock Cichlid Total Length Group (mm) Zooplankton 0 Other SAmphipod Ulnsect 0 Decapod Fish Figure 7. Volume contribution to diet of prey categories in juvenile (A) largemouth bass (N = 216) and (B) peacock cichlids (N 200) electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida. The "Other" category includes larval fish/fish eggs, organic materials, and anurans.

PAGE 71

63 The Scaling of Gape Width with Fish Length The relations between gape width and total length for both predators were: Largemouth bass GW = 0. 11 3 TL 2.2 1 r^ = 0.965, N = 87, TL range = 58-175 mm; and Peacock cichlid GW = 0.0997 TL 1 .570, ^ = 0.968, N = 1 10, TL range = 39-165 mm (Figure 8). The ANCOVA revealed that the regression slopes were different (p < 0.0001), with largemouth bass having a slightly larger length-specific gape width. Nevertheless, there was substantial overlap of gape measurements, especially for fish below 1 00 mm TL (Figure 8). The distribution of gape sizes for both predators also overlapped considerably during the sampling period (Figure Largemouth bass Peacock cichlid 14E E 12" 6" 0 0 50 100 150 200 Total Length (mm) Figure 8. Length-specific gape width measurements from juvenile largemouth bass (N = 87) and peacock cichlids (N = 110) electrofished in southeast Florida canals.

PAGE 72

64 July 1999 Snapper Creek Canal LMB = 18 PC = 13 LMB PC L li 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 August 2001 Snapper Creek Canal aLMB = 50 PC = 50 ml 9 10 11 12 13 14 15 16 17 September 1999Snapper Creek Canal >. U AO I 30 20 LMB = 19 PC = 30 4UU iUJL 2 3 4 5 6 7 S g 10 11 12 13 14 15 16 17 October 1999Snapper Creek Canal >< U 40 LMB = 8 PC = 31 i n n 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 October 2001 Snapper Creek Canal >• U 40 c P 20 LMB PC = = 41 50 nfl. IImiJi 2 3 4 5 6 7 8 0 10 11 12 13 14 15 16 17 August 2001 Cutler Drain Canal 50 U 40 C 30 CT 0) 20 LMB = SO PC = 47 i.hl.l 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 November 1999Snapper Creek Canal 5 B 7 8 9 10 11 12 13 14 15 16 17 Gape Width Group (mm) October 2001 Cutler Drain Canal ^ 60 ^50 >> U 40 C 30 D" Q) 20 LMB = 50 PC = 50 , ,1,1 5 6 7 8 9 10 11 12 13 14 15 16 17 Gape Width Group (mm) Figure 9. Gape width frequency for juvenile largemouth bass (LMB) and peacock cichlids (PC) electro fished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida. Sample sizes are shown for each sampling event.

PAGE 73

65 Prey Size and Stomach Fullness Prey size increased with predator size for fish and decapod crustacean prey (largemouth bass, r = 0.426, p = 0.0015; peacock cichlid, r = 0.490, p < 0.0001) (Figure 10). Moreover, there was relatively little difference in prey size between largemouth bass and peacock cichlids on either a length-specific or gape-specific basis. Indeed, ANCOVA detected no differences in either the slope (p = 0.608) or intercept (p = 0.165) values of the prey length-predator length regressions. The largest estimates of prey length relative to predator length were 34.4% for largemouth bass and 35.4% for peacock cichlids. Prey body depth also was significantly correlated with largemouth bass total length (r = 0.437, p = 0.0024) and peacock cichlid total length (r = 0.271, p = 0.0037). The ANCOVA revealed no significant difference in slope between species (p = 0.177), but did detect a significant difference in intercept values (p = 0.0013), with largemouth bass having a higher value. Highest estimates of prey body depth relative to predator gape size were 101.4% for largemouth bass and 95.0% for peacock cichlids. Stomach fullness estimates (Table 2) for the July, September, and October 1999 Snapper Creek Canal samples were not significantly different between predator species. However, juvenile largemouth bass had fuller stomachs than did peacock cichlids in the November 1999 sample (Wilcoxon; p = 0.0163). In 2001, largemouth bass stomachs were significantly fiiller than peacock cichlid stomachs in all samples (Wilcoxon; p < 0.0001 for all four sampling date by canal combinafions). There were no significant differences in stomach fullness for largemouth bass among sampling events (KruskalWallace, p = 0.0541). In contrast, stomach fullness did vary significantly for peacock cichlids (K-W, p < 0.0001). Dunn's test demonstrated that peacock cichlid stomachs

PAGE 74

66 50 40 >=30 I' 5 I 20 10 A Largemouth bass Peacock cichlid 'a CD A a a A ft A i A A — I— 40 — I— 60 — I— 80 I 100 I 120 I 140 I 160 —I 180 20 Predator Total Length (mm) B itiiiiii 0 20 40 60 80 100 120 140 160 180 Predator Total Length (mm) Figure 10. (A) Prey length (mm) of decapod crustacean (folded body length) and fish (standard length) prey eaten by juvenile largemouth bass and peacock cichlids electrofished from Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida. Results include 53 prey fi-om largemouth bass and 123 prey from peacock cichlids. (B) Body depth (mm) of decapod crustacean and fish prey eaten by juvenile largemouth bass and peacock cichlids electrofished fi-om Snapper Creek Canal (1999 and 2001) and Cutler Drain Canal (2001), Miami-Dade County, Florida. Results include 53 prey from largemouth bass and 1 1 3 prey fi-om peacock cichlids. The diagonal lines indicate estimated maximum gape for both predators based on the lengthspecific gape width regressions developed in the present study.

PAGE 75

67 Table 2. Percent (%) of juvenile largemouth bass and peacock cichlid stomachs within each stomach fullness category (i.e., 1 = empty, 2 = < 25% of estimated stomach capacity, 3 = 25-75% of stomach capacity, and 4 = > 75% of stomach capacity) for each sampling event. Specimens were electrofished from Snapper Creek Canal (SC) and Cutler Drain Canal (CD), Miami-Dade County, Florida. Largemouth Bass Peacock Cichlid Stomach Fullness Stomach Fullness Sampling Event Canal 1 2 3 4 1 2 3 4 July 1999 SC 33.3 11.1 44.4 11.1 23.1 15.4 38.5 23.1 September 1999 SC 20.0 5.0 45.0 30.0 6.7 13.3 50.0 30.0 October 1999 SC 12.5 50.0 12.5 25.0 16.1 29.0 25.8 29.0 November 1999 SC 22.2 0 22.2 55.6 24.1 41.4 27.6 6.9 August 2001 SC 8.0 26.0 32.0 34.0 33.3 41.2 17.7 7.8 October 2001 SC 12.2 14.6 51.2 22.0 29.4 56.9 9.8 3.9 August 2001 CD 12.2 38.8 34.7 14.3 50.0 37.5 6.2 6.3 October 2001 CD 6.1 49.0 28.6 16.3 56.0 40.0 4.0 0 were significantly fuller in the 1999 samples than in either canal in 2001, except that the mean rank for Snapper Creek Canal in August 2001 was not separable from the mean rank for Snapper Creek Canal in November 1999. Peacock cichlids in the October 2001 Cutler Drain Canal sample had the smallest mean rank stomach fullness of all the peacock cichlid samples. Diet Overlap Diets of juvenile largemouth bass and peacock cichlids overlapped on all dates (Table 3). Of the eight samples collected, four samples had a Schoener's Index value > 0.60 for prey numerical abundance. In particular, largemouth bass and peacock cichlids overlapped considerably in volume. Five of eight overlap values exceeded 0.60 and the other three values exceeded 0.50.

PAGE 76

68 Table 3. Schoener's Index values for dietary overlap by number and volume between juvenile largemouth bass and peacock cichlids electrofished from two southeast Florida canals (Miami-Dade County). A "*" indicates large overlap (i.e., Cxy > 0.60). CD = Cutler Drain Canal; SC = Snapper Creek Canal. Schoener's Index Values Sampling Event Canal Numerical Occurrence Volume July 1999 SC 0.102 0.957 September 1999 SC 0.658 0.836 October 1999 SC 0.499 0.522 November 1999 SC 0.012 0.873 August 2001 SC 0.920 0.576 October 2001 SC 0.551 0.654 August 2001 CD 0.932 0.586 October 2001 CD 0.895 0.785 Discussion The results of the present study broadly agree with earlier investigations of diets of juvenile largemouth bass in Florida (Chew 1974; Cailteux et al. 1996; Huskey and Tumigan 2001), although there were some important differences. Because prior studies of juvenile largemouth bass diets were conducted in lakes or impoundments in central or north-central Florida (versus southeast Florida canals), differences in stomach contents could be the result of a suite of factors, including habitat type, latitude, or the presence of peacock cichlids. Zooplankton was eaten by larger juvenile largemouth bass in the southeast Florida canals than was documented for most other Florida systems. Consumption of zooplankton in the present study occurred in largemouth bass in the 30-80 mm TL groups. Chew (1974) reported zooplankton from the stomachs of largemouth bass in the 3 1-75 mm TL range from two Florida lakes, but in the 76-296 mm TL range only in Lake Weir. Neither Cailteux et al. (1996) nor Huskey and Tumigan (2001) found

PAGE 77

69 zooplankton in fish of the size ranges reported in the present study (i.e., 58 mm TL and larger). The present study documented high consumption of insects by juvenile largemouth bass in the canals relative to the previous studies. Insects were important prey in Lake Weir, but not in Lake Hollinsworth (Chew 1974). Likewise, insects were relatively important prey for juvenile largemouth bass in vegetated and unvegetated north-central Florida lakes (Cailteux et al. 1996), but were a comparatively small part of the diet offish in east-central Florida lakes and impoundments (Huskey and Tumigan 2001). Consumption of decapod crustaceans varied widely in the studies. The data for canal largemouth bass showed that decapod crustaceans were major prey for fish of 90 mm TL and larger. Chew (1 974) found that decapod crustaceans were important in the diet of juvenile largemouth bass in Lake Weir, but were absent fi-om stomach contents fi-om Lake Hollinsworth. Cailteux et al. (1996) found that decapod crustacean were important prey in vegetated lakes but were of only minor importance in unvegetated lakes. Huskey and Tumigan (2001) reported substantial use of decapod crustaceans by largemouth bass of the sizes included in the present study. Although nearly all size classes of largemouth bass in the present study included individuals with fish prey in their stomachs, only largemouth bass > 120 mm TL surpassed 60% frequency of occurrence, the designation of a "true" piscivore as defined by Bettoli et al. (1992; see also Cailteux et al. 1996). The study canals generally had dense beds of aquatic macrophytes along the shoreline and often throughout the deeper water column as well. Cailteux et al. (1996) found that largemouth bass in unvegetated

PAGE 78

70 lakes became ftilly piscivorous (i.e., > 60% frequency of occurrence of fish in diet) by 60 mm TL, but did not become fiilly piscivorous until > 120 mm TL in vegetated lakes. Similar results were documented by Bettoli et al. (1992) for heavilyvegetated Lake Conroe, Texas, where largemouth bass did not become fijlly piscivorous until > 140 mm TL. Predatory fishes are less efficient at capturing fish prey in structurally complex habitats (Savino and Stein 1982; Hayse and Wissing 1996). Moreover, largemouth bass has been found to increase diet breadth (i.e., feeding on invertebrates and fish) in structurally complex habitats (Anderson 1984). Therefore, the late switch of juvenile largemouth bass to piscivory in the southeast Florida canals may be related to the abundant submersed vegetation. Very little informafion has been published on the diet of juvenile peacock cichlids. Lowe-McConnell (1969) reported palaemonid shrimp in 80-130 mm TL (?; type of length measurement is uncertain but assumed to be TL) peacock cichlids (N = 4) in Guyana. The reference is unclear, but fish also may have been eaten by peacock cichlids < 1 60 mm TL. Additionally, the author found zooplankton and filamentous algae in 40 mm TL (?) peacock cichlids. Again, the reference is unclear, but the sample size of peacock cichlids < 160 mm TL was not likely to have exceeded 15 or 20 individuals. In his investigation of an introduced population occurring in Lake Gatun, Panama, Zaret (1980a) examined stomach contents of seven Cichla specimens (either C. ocellaris or a congener) ranging in size fi-om about 55 to 90 mm SL (or about 70 to 1 12 mm TL using the conversion equation of Hill [1998]). The stomachs contained terrestrial insects, palaemonid shrimp, and fish. In Puerto Rico, introduced peacock cichlids < 1 50 mm TL ate fish and aquatic insects (Lilyestrom and Churchill 1996). In the only

PAGE 79

71 published study from Florida, Shafland (1999b) documented fish of several species, but no non-fish prey, from peacock cichlids 100-149 mm TL in Tamiami Canal. The present study was in general agreement with previous studies, but documented consumption of a much wider variety of prey. The present study documented considerable overlap in gape width between juvenile largemouth bass and peacock cichlids of similar sizes. This finding agrees with Hill (1998), who reported broad similarity in gape size of these species over a wide range of lengths (i.e., largemouth bass \ mm TL; peacock cichlids 50-430 mm TL). Given the extent of temporal and spatial overlap in total length distributions found in this study, juvenile largemouth bass and peacock cichlids of similar size and morphological prey-handling ability co-occurred in both canals. Morphological factors, coupled with syntopy, implied that juvenile largemouth bass and peacock cichlids had a high potential for undergoing similar dietary shifts and therefore sharing common food resources. As suggested by morphology, the ontogenetic prey-use pattern of largemouth bass and peacock cichlids was similar (i.e., zooplankton, followed by insects, then decapod crustaceans and fish). The main dietary differences were in a slightly broader range of prey eaten by largemouth bass (e.g., inclusion of crayfish in addition to grass shrimp in the decapod crustacean category and inclusion of anurans in the diet) and in the timing (i.e., total length) of shifts in prey proportions. For example, peacock cichlids consumed zooplankton at larger sizes than did largemouth bass (i.e., through the 1 00 mm TL group versus the 80 mm TL group for largemouth bass). Furthermore, amphipods and insects

PAGE 80

72 100 1 Largemouth Bass Peacock Cichlid Figure 1 1 Frequency of occurrence (% of predator stomachs with the prey type) for largemouth bass (N = 1 13; 100-349 mm TL) and peacock cichlids (N = 156; 100-348 mm TL) sampled from Tamiami Canal, Miami-Dade County, Florida. Data from Shafland 1999b. E] Insect B Decapod Fish Largemouth Bass Peacock Cichlid Figure 12. Percent (%) by number of prey for largemouth bass (N = 1 13; 100-349 mm TL) and peacock cichlids (N = 156; 100-348 mm TL) sampled from Tamiami Canal, Miami-Dade County, Florida. Data from Shafland 1999b. were generally more important for largemouth bass than peacock cichlids, although certain peacock cichlid size classes included these in high numbers and frequency. Stomach contents data from another southeast Florida canal also show a similar pattern of

PAGE 81

73 consumption of decapod crustaceans and fish by largemouth bass and peacock cichlids of 100 to about 350 mm TL (Shafland 1999b; Figures 1 1 and 12). Initiation of piscivory is an important ontogenetic event for predatory fishes (Aggus and Elliott 1975; Pasch 1975; Ludsin and DeVries 1997). Fish were included in the peacock cichlid diet at a generally higher frequency, number, and volume than in the largemouth bass diet at all predator sizes and peacock cichlids may have started to eat fish at a shorter total length than did largemouth bass. As a result, peacock cichlids became piscivorous (sensu Bettoli et al. 1992) before largemouth bass (i.e., by the 100 mm TL group versus the 120 mm TL group). These trends were even more pronounced based on numeric and volumetric proportions of fish in the diets of juvenile predators (e.g., for predator size classes > 100 mm TL, largemouth bass averaged [mean 1 SE] 56.0 1 1 .9% fish by volume versus 88.1 1.7% for peacock cichlids). The percentage of empty stomachs was high for peacock cichlids and much lower for largemouth bass. This translated into significantly higher stomach fiallness estimates for largemouth bass on several sampling dates. These data agree with the results of small and large tank studies, which documented lower feeding rates for peacock cichlids than for largemouth bass (Chapter 5). Relatively low daily rations for a congener, Cichla monoculus also have been reported (Rabelo and Araujo-Lima 2002). Surprisingly, however, juvenile peacock cichlids may grow faster than juvenile largemouth bass. There was substantial overlap in the total lengths of juveniles throughout the sampling period. Because largemouth bass spawn earlier than peacock cichlids (Shafland 1999b), the overlap in length frequencies in the present study suggests faster growth of age-0 peacock cichlids. Unfortunately, there are no studies available that directly measured

PAGE 82

74 growth rates of juvenile largemouth bass and peacock cichlids. If lower stomach fullness are due to lower feeding rates, then these results suggest that peacock cichlids are more efficient than largemouth bass (i.e., that despite lower consumption rates they are able to grow faster). Because a shift to a fish diet has been linked to increases in growth rates in piscivorous fishes (Pasch 1975; Timmons et al. 1980; Olson 1996), early and substantial consumption of fish prey by juvenile peacock cichlids may have resulted in their apparent faster growth rates relative to largemouth bass. Not only were the stomach contents of juvenile largemouth bass and peacock cichlids similar in taxonomic identity, they also were similar in relative prey size. For both species, increasing prey size was weakly correlated with increased predator length. The only statistically significant difference in prey size between the species being marginally larger (i.e., deeper-bodied) prey in largemouth bass. This result may be due to the fact that a largemouth bass has a slightly larger gape width than a peacock cichlid of the same total length. Competition for food can be an important interaction for juvenile largemouth bass. For example, sunfishes (i.e., Lepomis spp.) and juvenile largemouth bass have been shown to compete for food in the littoral zones of temperate lakes, possibly limiting largemouth bass recruitment (Osenberg et al. 1994; Olson et al. 1995). Likewise, growth and survival of largemouth bass juveniles is density-dependent (Olson et al. 1995; Garvey et al. 2000). The presumed mechanism is intraspecific food competition. The results of these studies suggest that food may be limiting for juvenile largemouth bass under some circumstances and that the presence of fishes with similar food requirements and habitat use may be detrimental to largemouth bass abundance.

PAGE 83

75 Species with similar ecological requirements which consume resources from the same habitat should have higher competition coefficients than a species pair which feeds in different habitats or consumes different resources within a habitat (Rosenzweig 1981). Juvenile largemouth bass and peacock cichlids overlapped considerably in diet and such large overlap index estimates have been considered significant (Zaret and Rand 1971; Mathur 1977). Nevertheless, diet overlap indices by themselves do not necessarily demonstrate competition (Colwell and Futuyma 1971; Hulbert 1978). For example, high overlap can occur without competition when resources are abundant. Rock-dwelling cichlids (i.e., "mbuna"; Cichlidae) in Lake Malawi, Africa, exhibit extreme trophic differentiation based on differences in tooth structure, mouth morphology (e.g., size and orientation), and pharyngeal jaw apparatus (Fryer and lies 1972). Normally, most of these species are herbivorous, grazing on aufwuchs (i.e., bio-cover of algae and bacteria) in specific microhabitats (Ribbink 1991). Nevertheless, during times of high zooplankton abundance, mbuna will abandon their trophic specialties and feed on superabundant zooplankton (McKaye and Marsh 1983; Ribbink et al. 1983). Dietary overlap is very high during these periods, but competition is not. On the other hand, Rosenzweig (1981) modeled two-species competition and concluded that a competition coefficient of zero (implying little or no resource overlap) was possible, even between two strong competitors, due to competitor-induced habitat shifts. Indeed, low overlap in diet could be the consequence of competition. Diets of native lake trout Salvelinus namaycush significantly differ fi-om diets of introduced smallmouth bass Micropterus dolomieu and rock bass Ambloplites rupestris in Canadian lakes (Vander Zanden et al. 1999). In these lakes, smallmouth bass and rock bass eat littoral prey fish and lake trout feed primarily on

PAGE 84

76 zooplankton and include fish prey as only a small part of the diet. However, in similar lakes without introduced smallmouth bass or rock bass, lake trout diets contain a considerably higher percentage of littoral prey fish. These results imply a competitioninduced diet shift. Indeed, high dietary overlap values have been used both to refute and to demonstrate the existence of interspecific competition in fishes (Colwell and Futuyma 1971; Hulbert 1978). In some studies, high dietary overlap between two fishes is interpreted as evidence that food resources are not limiting (i.e., therefore competition is not occurring) or to explain observed interspecific habitat or activity period differences (e.g., Zaret and Rand 1971; Keast 1978, 1985; Keast et al. 1978; George and Hadley 1979; Winemiller 1989; Hirst and DeVries 1994). Much of this research is based on morphologically similar species, feeding guilds, or entire species assemblages that are native to the study area, apparently have coevolved, and presumably have reacted to past competition through character displacement or resource partitioning. In other studies, particularly investigations of nonindigenous fishes, dietary overlap is used as evidence of food competition or of the high potential for food competition to occur (e.g., Mathur 1977; Parrish and Margraf 1990; Zale and Gregory 1990; Scoppettone 1993; Ogle et al. 1995;Declercket al. 2002). Dietary data trom southeast Florida canals show high overlap, and previous studies in temperate lakes demonstrate that largemouth bass are food-limited. Unfortunately, there are no good tests of food limitation in southeast Florida canals, although there are studies of the abundances of small fishes upon which largemouth bass and peacock cichlids feed. During the past two decades (i.e., since the introduction of

PAGE 85

77 peacock cichlids in 1 984) qualitative and quantitative surveys of fishes in the canals where largemouth bass and peacock cichlids co-occur have found high numbers of small fishes, both native and nonindigenous (boat electrofishing surveys: L. G. Nico, USGS, unpubl. data; pers. obs.; blocknet sampling: Shafland et al. 1985; Shafland 1999a; Shafland et al. 2001). These high densities suggest that small fish availability does not limit predators such as largemouth bass and peacock cichlids. However, dynamics of invertebrate prey, which may influence early growth patterns of largemouth bass and peacock cichlids, have not been investigated in these canals. Finally, if food was limiting, then largemouth bass population densities should have declined following the introduction of peacock cichlids, a potential competitor. Nevertheless, available data do not suggest that largemouth bass populations have declined during the 20-year period following the introduction of peacock cichlids (Shafland 1995, 1999a, c; Shafland and Stanford 1999; Shafland et al. 2001; Chapter 6). Fishes that are piscivorous as adults typically undergo marked changes in diet related to increases in body length and mouth size (Keast 1985, Keast and Eadie 1985; Olson 1996; Mittlebach and Persson 1998; present study). Knowledge of the extent to which ecomorphological similarity is predictive of the timing (i.e., predator size) and prey progression of such shifts can be important in evaluating the risks associated with fish introductions, whether planned or illegal. The results of the present study suggest that ecomorphological similarity to a native fish is predictive of dietary ontogeny and dietary overlap for nonindigenous species. This implies that such information can be used for initial assessments of risk prior to fish introductions in the absence of diet data for early life history stages of native or introduced populations of the nonindigenous

PAGE 86

78 species. Studies of nonindigenous fishes typically concentrate on adult stages. However, careful attention also must be paid to early life stages and the possible interactions of such stages with native fishes for those species that differ in habitat and food use through ontogeny (Courtenay 1995). Given the ecomorphological convergence of many cichlids and centrarchids (Norton and Brainerd 1993), using similarity to predict food resource use may be particularly valuable when cichlids invade systems where centrarchids are native (e.g., Florida) or when centrarchids invade systems in Africa, Central America, or South America.

PAGE 87

CHAPTER 5 EXPERIMENTAL PREY HANDLING AND SELECTIVITY OF TWO MORPHOLOGICALLY SIMILAR PREDATORY FISHESNATIVE LARGEMOUTH BASS AND INTRODUCED PEACOCK CICHLIDS Introduction The morphology of predatory fishes is known to have a profound effect on their prey use. For example, gape limitation is an important constraint for many predatory fishes. The size of the mouth, pharyngeal, or intercleithral opening (i.e., gape size) places limits on maximum prey size and influences the energetic return from submaximal prey (Lawrence 1958; Werner 1974, 1977, 1979; Zaret 1980b; Wainwright 1988; Wainwright and Richard 1995). Furthermore, patterns of prey selection frequently scale with patterns of gape size (Wainwright and Richard 1995; Mittelbach and Persson 1998). The constraint of gape limitation often results in an ontogenetic dietary progression characterized by the initial consumption of zooplankton, then of intermediate-sized invertebrates, and finally by fish (Keast 1985; Wainwright and Richard 1995). Given this relationship of gape size to prey use, a reasonable prediction is that an introduced species with gape size (and other features of morphology) similar to a native species will also exhibit similarifies in prey handling ability and prey selection. Additionally, such similarities suggest that the nonindigenous species will exert comparable pressures on prey species. These predictions are of general concern following a fish introduction because they imply the potential for food competition 79

PAGE 88

80 between the invader and any ecomorphologically similar native species. Indeed, competition is the most often cited mechanism whereby introduced fishes negatively effect native fishes (Moyle et al. 1986). The native largemouth bass (Centrarchidae) is an economically important predatory fish in southern Florida (Shafland and Stanford 1999). The peacock cichhd Cichla ocellaris (Cichlidae), a species native to South America, was legally introduced into freshwater canals in southeast Florida in 1984-1987 by the Florida Game and Fresh Water Fish Commission (currently Florida Fish and Wildlife Conservation Commission) (Shafland 1995). The predatory peacock cichlid was stocked to reduce the abundance of other nonindigenous fishes, particularly spotted tilapia Tilapia mariae and to augment recreational fisheries that primarily targeted largemouth bass (Shafland 1995). Largemouth bass and peacock cichlids are ecomorphologically similar based on appearance, body size, gape size, feeding mechanics, and diet (Norton and Brainerd 1993; Hill 1998; Chapter 4). Dietary overlap of juveniles in southeast Florida canals is generally large (Chapter 4) and adults of both species are regarded as piscivorous (Carlander 1977; Shafland 1995, 1999b). Additionally, experimental studies of largemouth bass prey selection have demonstrated that fish are often selected at a higher rate than non-fish prey (e.g., Lewis et al. 1961). There was considerable overlap in diets of subadult and adult largemouth bass and peacock cichlids in Tamiami Canal (Shafland 1999b). Nevertheless, largemouth bass consumed substantial amounts of non-fish prey (54 % by number; primarily decapod crustaceans, but also insects) in addition to fish whereas co-occurring peacock cichlids were almost exclusively piscivorous (Shafland 1999b).

PAGE 89

81 These results suggest that ecomorphological similarity can predict coarse patterns but is not precise enough to predict patterns of a finer scale. Alternately, some other mechanism (e.g., differential encounter rates mediated by habitat, predator behavior, or competition) may be operating to modify the prey use patterns of one or both species. Prey selection by predators results from the interaction of differential encounter rates with prey, probabilities that a predator will attack a prey item if encountered, and probabilities that an attack results in consumption (O'Brien 1979; Osenberg and Mittelbach 1989). Encounter rate is fundamental to the process because it sets the initial conditions from which the remaining predation components operate (Osenberg and Mittelbach 1 989). For example, rarely encountered prey may represent a small portion of a predator's diet even if the product of attack and success probabilities approaches unity (i.e., 1.0). Encounter rates, especially under field conditions, are difficult to estimate (Osenberg and Mittelbach 1989). Moreover, structurally complex habitat can greatly reduce encounter rates and feeding efficiency of predators (Savino and Stein 1982; Anderson 1984; Hayse and Wissing 1996). Compefifion between predators also can reduce encounter rates with prey by prey depletion, by modifying prey behavior in response to predation risk, or by direct interference with predator feeding behavior (Hobson 1979; Mittelbach 1984; Fausch et al. 1997; Turner et al. 1999). Alternately, species-specific hunting behavior of predators may influence the rate of encounter with different prey types (Cooper et al. 1985; Magnhagen 1986; Grant and Noakes 1987). However, by creating situations where encounter rates are similar between two predators (e.g., by making encounters almost certain), attack and success probabilities can be estimated and comparisons can be made.

PAGE 90

82 I conducted a series of tank experiments with largemouth bass and peacock cichlids to test predictions concerning prey use by similar nonindigenous and native fishes (i.e., similar morphology equals similar prey handling ability, prey selection, and thus potential effects on prey base). I used estimates of prey handling times to test for differences in the ability of both predators to handle prey. For example, prey handling and the resulting information on prey profitability are important for predictions of prey selection based on optimal diet models (Emlen 1966; MacArthur and Pianka 1966). This experiment also provided information on the probability of attack and successful consumption of prey for both predators within the experimental system. This information was used to formulate and test predictions of prey selection for largemouth bass and peacock cichlids given choices among prey types. Communities often have multiple species of predators (Sih et al. 1998) and a number of studies have investigated the equivalence of multiple predators in their effect on prey assemblages (Harris 1995; Morin 1995; Kurzava and Morin 1998; Baber and Babbitt 2003). The introduction of peacock cichlids into southeast Florida canals represents the addition of a predator to a predator assemblage formerly dominated by a single species, the largemouth bass (Chapter 2). Therefore, I tested assemblages made up of varying densities and percent composition of largemouth bass and peacock cichlids for predator assemblage effects on prey selection and prey mortality. Methods Experiment 1 Prey Handling I conducted a laboratory experiment to test the relative ability of largemouth bass and peacock cichlids to capture and consume fish and non-fish prey. This experiment

PAGE 91

83 was conducted in the University of Florida Demonstration Aquaculture Recirculating System (see Hill and Cichra in prep, for a complete description). This indoor system has 1.1 -m diameter by 0.6-m high, circular fiberglass tanks connected to biofilters. Tank interiors were light blue and had single airstones and 50-mm central standpipes. Water temperature was maintained between 25.5 and 27 C and the light:dark cycle was 13 h:ll h. Eight largemouth bass were electrofished from Lake Alice, Alachua County, Florida and eight peacock cichlids were electrofished from Cutler Drain Canal, MiamiDade County, Florida. The largemouth bass averaged 322 mm total length (TL) (255408 mm TL) and the peacock cichlids averaged 340 mm TL (251^07 mm TL). Prey were collected from ponds at the Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville, Florida. Prey handling times, attack probability, and probability of success given an attack were estimated for single prey items presented to individual predators. Prey used in the experiment were of three representative prey types — 1) a soft-rayed fish (i.e., sailfin molly Poecilia latipinna ), 2) a spiny-finned fish (i.e., bluegill Lepomis macrochirus ), and 3) a decapod crustacean (i.e., crayfish [a mix of Procambarus fallax and P. paeninsulanus]) Each prey type is present in southeast Florida. Prey sizes were chosen to encompass a range of about 20-100% of predator gape size (i.e., prey body depth divided by predator gape width and multiplied by 1 00) based on relafions for bluegill, largemouth bass, and peacock cichlid found in Hill (1998) and for crayfish and sailfin molly developed in the present study. Length-specific body depth relations for crayfish and sailfin molly were developed from fresh specimens (PROC GLM; SAS Institute, Cary, North Carolina):

PAGE 92

84 Crayfish: BD = 0.339 TL 2.841 ; = 0.90 Sailfin Molly: BD = 0.3 1 8 TL 3.890; = 0.98 where BD = maximum body depth. Handling time was measured as the total time from initial prey contact to the cessation of all feeding movements of mouth and opercula, including time spent recapturing released prey. In the small tanks, initial searching and pursuit times were essentially zero. Prey mass was estimated using length-mass regressions developed from fresh specimens (PROC GLM; SAS Institute, Gary, North Carolina). Bluegill: LogioWT = 2.790 LogioTL 4.525; r^ = 0.95 Crayfish: LogioWT = 3.498 LogioTL 5.472; ^ = 0.97 Sailfin Molly: LogioWT = 3.388 LogioTL 5.436; r^ = 0.99 Energy values for prey were taken from the literature — 1 160 cal/g for centrarchids (Miranda and Muncy 1989), 1 100 cal/g for other fish (Pope et al. 2001), and 750 cal/g for crayfish (Pope et al. 2001) for wet weights. I incorporated effects of prey size and predator size by dividing prey body depth by predator gape (and multiplied by 100), which I called relative prey size, and then categorized relative prey size into 10% size classes. Two-way analysis of variance (ANOVA) was used to detect differences in mean prey profitability across relative prey size classes for sailfin molly and bluegill prey using predator species and relative prey size class as main effects (PROC GLM; SAS Institute, Gary, North Carolina). Nonsignificant interacfion terms (i.e., predator x size class) were dropped from the model (bluegill interacfion F5^54 = 1 .2 1 p = 0.3 1 9; sailfin molly interacfion Fs.ss = 1 .55, p = 0.212). One-way ANOVA was used for crayfish prey because peacock cichlids did not

PAGE 93

85 eat crayfish. ANOVAs were followed by the least squares means multiple comparison procedure. ANOVA (followed by Tukey's HSD multiple comparison procedure) was also used to test for the effect of prey size and type on the probability of successful consumption of prey for both predators given that an attack had taken place (PROC GLM; SAS Institute, Gary, North Carolina). Two-sample t-tests were conducted to investigate differences between predators in the probability of successfully consuming each prey type (PROC TTEST; SAS Institute, Cary, North Carolina). Probabilities were arcsine-square root transformed prior to analysis. All tests were conducted at a Type I error rate (a) of 0.05. Experiment 2 Prey Selection by Individual Predators Prey selection patterns were determined empirically in the same laboratory tanks that were used in Experiment 1 A single predator per tank was tested and eight largemouth bass and eight peacock cichlids were tested. Eight largemouth bass were electrofished from Lake Alice, Alachua County, Florida and eight peacock cichlids were electrofished from Cutler Drain Canal, Miami-Dade County, Florida. The largemouth bass averaged 322 mm total length (TL) (255^08 mm TL) and the peacock cichlids averaged 340 mm TL (251^07 mm TL). Prey were collected from ponds at the Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville, Florida. Predators were fasted the day prior to each experiment. Predators not being tested were fed crayfish, eastern mosquitofish Gambusia holbrooki sailfin mollies, bluegills, and redear sunfish Lepomis microlophus Prey species selectivity was tested using three representative prey types — a soft-rayed prey (i.e., sailfin molly), a spinyfinned prey (i.e., bluegill), and a decapod crustacean (i.e., crayfish). Five of each prey

PAGE 94

86 type were presented simultaneously to each predator (i.e., total of 1 5 prey per trial). Prey body depths were 40-60% of predator gape size. Remaining prey were retrieved and counted after one day (i.e., about 24 h). Prey selectivities for largemouth bass and peacock cichlids feeding on fish and crayfish prey were calculated using the Manly-Chesson index (Manly 1974; Chesson 1978, 1983), accounting for prey depletion during the experiment (i.e., Case 2; Chesson 1983): ln((n,o-r,o)/,.o) ^, = 1^...^^ 7=1 where m is the number of prey categories, r/o is the number of prey type i in the diet and n,o is the number of prey / in the environment. Values of a range from 0 to 1 with 0 being avoidance, 1 being complete preference, and \lm being random selection. Oneway ANOVA (followed by the Tukey HSD multiple comparison procedure) was used to test the hypothesis of random prey selection for each predator (i.e., all prey categories equally selected) (PROC GLM; SAS Institute, Cary, North Carolina). Two-sample t-tests were used to determine if prey selectivity (i.e., a) differed between largemouth bass and peacock cichlids for sailfin molly and bluegill prey (PROC TTEST; SAS Institute, Cary, North Carolina). The Wilcoxon signed rank test was used to test for differences in mean numbers of prey eaten per predator between largemouth bass and peacock cichlids (PROC NPARl WAY; SAS Institute, Cary, North Carolina). A significance level of a = 0.05 was used for all statistical analyses. Data on probability of attack and probability of successfiil consumption from Experiment 1 were used to predict prey selection for largemouth bass and peacock

PAGE 95

87 cichlids in Experiment 2, assuming that all available prey were encountered (given the small size of the tanks), and subject to the constraint that the total number of prey eaten was equal to that observed. Chi-square tests were used to test the hypothesis that the predicted diet was not different from the observed diet for each predator species (PROC FREQ; SAS Institute, Gary, North Carolina). To determine if decapod crustaceans are acceptable prey for subadult and adult peacock cichlids given the lack of consumption of crayfish in the experiments, I stocked 1 0 crayfish into tanks with individual peacock cichlids and offered no alternative prey for five days. This experiment was conducted at the Florida-Caribbean Science Center, U. S. Geological Survey, Gainesville, Florida in indoor, 1 .2-m diameter by 0.6-m high, circular fiberglass tanks. Tank interior color was light blue and the interior was bare except for a single airstone and a central 1 0-cm standpipe. There was a constant, slow flow of well water through each tank and water temperature was maintained at about 26.7 C by the use of two 300-w heaters. Peacock cichlids were collected by electrofishing from Cutler Drain Canal and Snapper Creek Canal in Miami-Dade County, Florida and crayfish were trapped from ponds at the Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville, Florida. A total of five peacock cichlids ranging from 196 to 355 mm TL were tested. Crayfish body depths were smaller than predator gape width (Hill 1998). Experiment 3 Prey Selection by Predator Assemblages I conducted an experiment at the Sam Mitchell Aquaculture Demonstration Farm (SMADF), University of Florida, Blountstown, Florida, to test for differences in the effect of predafion by assemblages of largemouth bass and peacock cichlids on prey

PAGE 96

88 assemblages consisting of three representative prey types. The experimental tank system consisted of 6-m diameter by 1 .2-m high circular fiberglass tanks with a central 10-cm, screened standpipe, several 15-cm airstones around the tank perimeter, and a constant flow of pond water from a single, 0.4-ha, aerated pond. Tank interior color was offwhite. Water in each tank was maintained at a depth of about 45 cm and had a temperature of 26 29 C during the experiment. Morning dissolved oxygen readings in the tanks exceeded 8.0 mg/L. Water clarity (i.e., vertical secchi depth) in the pond exceeded pond depth (> 2.4 m) and the mean value in the tanks for color was 8.82 Pt-Co (SD = 1.29) and for turbidity was 0.32 NTU (SD = 0.18). The tanks were outdoors under a 20% shade cloth covering. Largemouth bass and peacock cichlids were used as predators and golden shiners Notemigonus crysoleucas redear sunfish, and crayfish were used as prey. Largemouth bass were angled from ponds at SMADF and the Department of Fisheries and Aquatic Sciences (FAS), University of Florida, Gainesville, Florida. Peacock cichlids were electrofished from Cutler Drain Canal, Miami-Dade County, Florida. The largemouth bass averaged 340 mm TL (299-376 mm TL) and the peacock cichlids averaged 324 mm TL (290-365 mm TL). Golden shiners and redear sunfish were purchased from bait and fingerling suppliers. Crayfish were trapped from ponds at FAS. It is unlikely that any prey had experience with large predatory fish. Prey species were chosen to represent distinct prey types — 1) an elongate, soft-rayed fish (i.e., golden shiner), 2) laterallycompressed, spiny-finned fish (i.e., redear sunfish), and 3) a decapod crustacean (i.e., crayfish). These species are present in southeast Florida canals. Prey sizes used were smaller than predator gape size according to length-specific gape width relations from

PAGE 97

89 Hill (1998). Prey were chosen to be similar in body depth (mean BD = 16-18.5 mm), the prey dimension of most importance to gape-limited predators (Lawrence 1958; Hambright 1991). Mean prey length and weight were compared by one-way ANOVA followed by the Tukey HSD multiple comparison test (PROC GLM; SAS Institute, Gary, North Carolina). Golden shiners were longer than crayfish and redear sunfish (ANOVA, F2,78 = 69.31, p < 0.0001) and heavier than redear sunfish (ANOVA, F2,7g = 38.03, p < 0.0001). Crayfish and redear sunfish were of similar length, but crayfish were heavier than golden shiners and redear sunfish (ANOVA, F2,78 = 38.03, p < 0.0001). Nine predator treatments were tested with an equal initial food resource level of 50 each of golden shiners, redear sunfish, and crayfish. Predator density was four, six, or eight predators (i.e., low, medium and high predator density). Within each density level, the species composition of predators was 100% largemouth bass, 100% peacock cichlid, or 50% of each species, giving nine treatment combinafions (i.e., 3 predator densities x 3 species combinations = 9). A tenth treatment, zero predators, was used to estimate losses not due to predation, but was not included in the analyses because no losses were observed in this treatment. Treatments were randomly assigned to tanks and each treatment was run three times. Trials were run for three days. Tanks were inspected at least twice daily for prey mortality. Missing prey were not replaced. At the end of the trial, the tanks were drained and prey were removed and counted. Predator densities, initial prey numbers, and trial duration were chosen to provide a range of resource availability fi-om abundant (i.e., the most selected prey would not be eradicated) to scarce (i.e., highly selected prey will be eradicated and alternative prey should be exploited).

PAGE 98

90 Prey selectivity values for golden shiners, redear sunfish, and crayfish were estimated for each predator density x combination treatment using the Manly-Chesson index (i.e., a Case 2, with depletion) (Chesson 1983). Values were arcsine-square roottransformed prior to analysis. Two-way ANOVA, followed by the least squares means multiple comparison procedure was used to test for the effects of predator assemblage characteristics (i.e., predator density, combination, and density x combination interaction) on mean a values (PROC GLM; SAS Institute, Gary, North Carolina). Non-significant interactions were excluded fi^om the model and ANOVA was repeated with only the main effects. In addition to examining prey selectivity (which is a measure of the relative mortality imposed on prey types), I also compared the absolute mortality imposed on prey. I used two-way ANOVA to test the effect of predator density and combination on the logio-transformed odds of a predator eating a prey during the trial versus leaving it alive at its end, followed by the least squares means multiple comparison procedure (PROC GLM; SAS Institute, Gary, North Carolina). Non-significant interactions (i.e., predator density x predator combination) were excluded fi'om the model and ANOVA was repeated with only the main effects. One-way ANOVA was used to test for differences in mean per capita consumption of crayfish by largemouth bass (PROC GLM; SAS Institute, Gary, North Carolina). Type I error rate (a) was set at 0.05. Results Experiment 1 Prey Handling In general, handling times increased with prey size but varied between largemouth bass and peacock cichlids and among prey types; however, there was considerable

PAGE 99

91 variation in these relationships leading to pronounced variation in prey profitability. Sailfin mollies were more energetically favorable prey for peacock cichlids than for largemouth bass (ANOVA, Fi,58= 16.48, p = 0.0001; Figure 13A). Nevertheless, sailfin mollies were of high value relative to other prey for largemouth bass (Figures 13A and 13B). Addifionally, the relafive size of sailfin mollies did not have a significant effect on their profitability (ANOVA, F4,58 = 0.43, p = 0.785). Contrary to sailfin molly prey, profitability of bluegill prey did not depend on predator species (ANOVA, Fi,59= 2.33, p = 0.1322), but was influenced by relafive prey size (ANOVA, F5,59= 8.14, p < 0.0001; Figure 13B). The smallest size classes of bluegill (i.e., 40 and 50%) were more profitable than were larger bluegills. Only largemouth bass ate crayfish during this experiment. The relative size of crayfish prey did not have a significant effect on crayfish profitability for largemouth bass (ANOVA, F3,43 = 2.50, p = 0.072; Figure 13B); however, crayfish in the larger than 70 % of gape size were offered to largemouth bass but were not eaten. Largemouth bass and peacock cichlids always attacked fish prey (i.e., sailfin mollies and bluegills) placed into the experimental tanks (Table 4). Largemouth bass usually attacked crayfish, but peacock cichlids never attacked crayfish. Addifionally, successful consumption of attacked prey varied among prey types. Largemouth bass were more successfial at consuming fish than crayfish (ANOVA, F2,i78 = 9.33, p = 0.0001) and peacock cichlids were more successful with sailfin mollies than with bluegills (ANOVA, F|,48 = 7.68, p = 0.008). Prey size did not influence predator success (given that an attack occurred) for largemouth bass eating sailfin mollies (ANOVA, F5,42 = 0.95, p = 0.460) or bluegills (ANOVA, F5,54 = 0.35, p = 0.877) or for peacock cichlids

PAGE 100

92 Figure 13. Estimated mean profitability ( 1 SE) to largemouth bass (LMB) and peacock cichlids (PC) of sailfm molly (SFM) (A), bluegill (BG) (B), crayfish (CF) (B) versus relative prey size. BD = body depth; GW = gape width.

PAGE 101

93 eating sailfin mollies (all attacked sailfin mollies were successfiilly consumed) or bluegills (ANOVA, F5,23 = 0.42, p 0.832). However, largemouth bass had lower success after an attack for the largest size class of crayfish eaten (i.e., 60%) compared to the smallest crayfish size class (i.e., 30%) (ANOVA, F3,69 = 2.94, p = 0.039). Comparing the predators to each other, peacock cichlids were more successfial in consuming sailfin mollies than were largemouth bass (t-test, to.os, 47 = -2.34, p = 0.024), but there was no significant difference between predators for bluegills (t-test, to.05, gv = 0.09, p = 0.928). Because only largemouth bass ate crayfish, it may be inferred that largemouth bass were more successful in their consumption of crayfish prey. Table 4. The probability of attack [Prob(A)], probability of successful consumption given an attack [Prob(SC)], and the product of those probabilities [Prob(AxSC)] for individual largemouth bass (N = 8) and peacock cichlids (N = 8) sequentially offered three prey types. No Prob(SC) was estimated for peacock cichlids feeding on crayfish because there were no attacks (i.e., ND = no data). Largemouth Bass Peacock Cichlid Prob(A) Prob(SC) Prob(AxSC) Prob(A) Prob(SC) Prob(AxSC) Sailfin Molly 1.0 0.896 0.896 1.0 1.0 1.0 Bluegill 1.0 0.738 0.738 1.0 0.704 0.704 Crayfish 0.884 0.548 0.484 0.0 ND 0.0

PAGE 102

94 Experiment 2 Prey Selection by Individual Predators Prey selection by both largemouth bass (ANOVA, 2,21 = 32.61, p < 0.0001) and peacock cichlids (ANOVA, 2,21 = 22.10, p < 0.0001) in the small tanks was non-random. Prey fish species were selected at a high frequency by predators relative to crayfish (Figure 14 and Table 5). Bluegills and sailfin mollies were not selected at a significantly different rate by largemouth bass; however, peacock cichlids had higher selectivity index values for sailfin mollies than for bluegills. T-tests revealed no difference between selectivity estimates for bluegill (to.05, i4 = 1 -57, p = 0.138), but peacock cichlids had higher mean selectivity for sailfin molly prey than did largemouth bass (to.05,14 = -2.24, p = 0.042). Largemouth bass ate significantly more prey per individual than did peacock cichlids (Wilcoxon, z = 1 .981, p = 0.048). 1^ 0.8 • .e 0.6 4 u CO 0.2 Sailfin Molly i Bluegill Largemouth Bass A Peacock Cichlid Crayfish Figure 14. Observed mean prey selectivity values (a) ( 1 SE) for individual largemouth bass (N = 8) and peacock cichlids (N = 8) preying on sailfin mollies, bluegills, and crayfish in small tanks. The horizontal dashed line indicates random prey selecfion (i.e., a = \/m = 0.333, where m = number of prey categories).

PAGE 103

95 Prey selection by largemouth bass differed from the predicted pattern based on attack and success probabilities estimated in Experiment 1 {"i = 7.28, df 2, p = 0.026; Table 5). In particular, crayfish were underrepresented in the observed diet and both fish prey were overrepresented. The diet change was in the direction of the more profitable prey (i.e., fish > crayfish), suggesting that attack probabilities decreased for crayfish in the presence of more profitable prey (i.e., fish). On the other hand, prey selecfion by peacock cichlids did not significantly deviate from the predicted values (x^ = 0.245, df = l,p = 0.621). Table 5. Predicted and observed prey selection for largemouth bass (N = 8) and peacock cichlids (N = 8) simultaneously presented with three prey types. Predicted diets were derived from estimates of the product of the probability of attack and the probability of successfial consumption given an attack obtained in Experiment 1 (Table 3). % = percent of prey; N = number of prey (based on observed consumpfion of 57 prey by largemouth bass and 35 prey by peacock cichlids). Largemouth Bass Predicted Diet Observed Diet Peacock Cichlid Predicted Diet Observed Diet % N % N N % N Sailfin 42 Molly 24 51 29 59 21 66 23 Bluegill 35 Crayfish 23 20 13 44 25 41 14 34 12 0

PAGE 104

96 No crayfish were eaten by peacock cichlids in the prey handhng or prey selectivity trials. Unlike largemouth bass, peacock cichlids also did not eat crayfish during regular feedings (i.e., non-experimental), but readily ate offered fish prey. Indeed, crayfish were not acceptable prey for most of the peacock cichlids, even in the absence of alternative prey. Of the 50 crayfish offered to peacock cichlids that had no alternative prey for five days, only one (i.e., 2%) was eaten. All five experimental peacock cichlids subsequently ate several fish prey on the day after the experiment concluded. Experiment 3 Prey Selection by Predator Assemblages There was no significant effect of predator assemblage on prey selectivity in the present study (Figure 1 5). Predator density had no significant effect on prey selectivity (golden shiner ANOVA, 2,22 = 1 .2 1 p = 0.3 1 6; redear sunfish ANOVA, 2,22 = 0.97, p = 0.394; crayfish ANOVA, 2^2 = 0.74, p = 0.488). Moreover, ANOVA failed to detect a significant effect of assemblage species composition on prey selectivity for golden shiners {2,22 = 1-97, p = 0.163), redear sunfish {2,22 = 2.43, p = 0.1 1 1), or crayfish (2,22 = 3.06, p = 0.067). In contrast, predator density and species composition did effect the absolute mortality rates of prey (Figure 1 6). Largemouth bass, peacock cichlids, and the mixed combination had similar odds of eating redear sunfish (ANOVA, 2,22 = 1 -90, p = 0. 1 74), but not golden shiners (ANOVA, 2,22 = 4.80, p = 0.019) or crayfish (ANOVA, 2^2 = 3.70, p = 0.041 ). In both cases, largemouth bass had greater odds of eating these prey than did peacock cichlids. The 50% predator combinations were intermediate between, and not statistically different from, the single species combinations.

PAGE 105

97 0.8 • •f 0.6 '.5 O 0) 0) 0.4(0 0.2 Golden Shiner A Redear Sunfish • Crayfish 5
PAGE 106

98 100 2 80 60 • \ 40. Q. 20 • A ^ $ • A A 4 ^ Predator Treatment N
PAGE 107

99 Discussion One of the implicit and seldom tested assumptions of ecological studies is the premise that morphological similarity between species leads to similarity in food resource use (Harris 1995; Wainwright and Richard 1995) and therefore increased likelihood that two species will compete. This assumption is often invoked in many discussions of introduced species (i.e., introduced and native species that are similar in morphology are likely to compete for resources [e.g., Matthews et al. 1992; Huckins et al.2000]). Overall, the results of this study support the hypotheses that morphologically similar species have similar general patterns of prey handling and prey selection. Similar prey handling ability of largemouth bass and peacock cichlids was most evident with bluegills, a species representative of spiny-finned, laterally-compressed fish prey. Southeast Florida canals have several abundant centrarchid (e.g., redear sunfish and spotted sunfish Lepomis punctatus ) and cichlid (e.g., jewel cichlid Hemichromis letoumeauxi and spotted tilapia) prey species that are similar to bluegill in morphology (Hill 1998; Shafland 1999a) and that are commonly consumed by largemouth bass and peacock cichlids (Shafland 1999b; L. G. Nico, USGS, and J. E. Hill, unpubl. data). This category includes spotted tilapia, an abundant nonindigenous fish that was the primary target of biological control efforts that were part of the management plan that resulted in the peacock cichlid introduction (Shafland 1995). Although there was support for the assumption of similar prey use by similar predators, potentially important differences were evident between largemouth bass and peacock cichlids in prey profitability and in the probability of successfully consuming (i.e., probability of attack x probability of success given an attack) encountered sailfin

PAGE 108

100 mollies and crayfish. Peacock cichlids were better than largemouth bass at preying on sailfin mollies (i.e., they incurred lower handling times and had higher capture success). However, sailfin mollies also were the most profitable prey for largemouth bass and the significance of this advantage for peacock cichlids is unknown. In contrast, peacock cichlids rarely (or never) consumed crayfish, providing largemouth bass with exclusive use of this resource. The ability of largemouth bass to exploit non-fish prey could be an important refuge against potential competifion with introduced peacock cichlids. Hypothefically, the lack of crayfish in peacock cichlid stomachs sampled from canals (e.g., Shafland 1999b) could be due to low (or zero) encounter rates between crayfish and peacock cichlids. For example, peacock cichlids are diurnal predators (e.g., Lilyestrom and Churchill 1996) and crayfish are crepuscular and nocturnal (Pennak 1989), thus limiting potential encounters. However, my results show that even when crayfish are present and encountered, peacock cichlids will not consume them. It is possible that crayfish are extremely difficult for peacock cichlids to subdue and eat or that peacock cichlids do not normally recognize crayfish as potential prey. These hypotheses are not distinguishable given the data gathered in this study. However, peacock cichlids (mostiy juveniles) in Florida regularly consume another decapod crustacean, the grass shrimp Palaemonetes sp. (Shafland 1999b; Chapter 4; L. G. Nico, USGS, and J. E. Hill, unpubl. data). Nevertheless, grass shrimp, in contrast to crayfish, lack strong chelae, possibly making them more vulnerable to peacock cichlids. Additionally, studies of various Cichla species outside of Florida have documented consumption of decapod crustacean prey (again mostly Palaemonetes spp.; LoweMcConnell 1969; Zaret 1980a; Jepsen et al. 1997), although adult Cichla are strongly

PAGE 109

101 piscivorous (Lowe-McConnell 1969; Lilyestrom and Churchill 1996; Jepsen et al. 1997; Winemiller et al. 1997). Morphologically similar species are often expected to have similar effects on the prey base (e.g., Hambright et al. 1 991 ; Harris 1 995). However, a disagreement between predicted diets based on experimental data (present study) and published diets of largemouth bass and peacock cichlids in southeast Florida canals (Shafland 1 999b) implies that this assumption may be too simplistic (i.e., other factors may influence prey use and result in differing observed diets in similar species). Fish prey were more profitable and readily captured and consumed by largemouth bass than were crayfish. Additionally, the present study documented avoidance of crayfish (i.e., acp l/wj) when largemouth bass were presented with alternative prey (i.e., fish). Conversely, Shafland (1999b) found that largemouth bass in a southeast Florida canal ate 45% (by number) decapod crustaceans, mostly fi-eshwater prawns Macrobrachium sp., but also crayfish. The inconsistency between experimental and field data may be due to the relationship of encounter rates between largemouth bass and their potential fish and non-fish prey (i.e., encounter rates with fish are low enough that largemouth bass include non-fish prey in their diet). Encounter rates for fish versus non-fish prey could be modified by competition, predator hunting behavior, prey anti-predator behavior, or habitat structural complexity. Additional research is needed to test these alternative (but not mutually exclusive) hypotheses. The heavy use of non-fish prey by largemouth bass in the canals raises concern over the possible effects of food competition with peacock cichlids, the other prominent predator in the system. Indeed, a shift in diet (i.e., induced resource partitioning) is a

PAGE 110

102 prediction of competition theory (Sale 1979). Both species regularly consume fish (Shafland 1995, 1999b; Chapter 4; L. G. Nico, USGS, and J. E. Hill, unpubl. data) and could potentially deplete fish prey to the point of forcing a shift in diet. However, the high abundance of prey fish (numerical and biomass) in many southeast Florida canals (Shafland et al. 1985; Shafland 1995, 1999a; L. G. Nico, USGS, unpubl. data; pers. obs.) suggests that fish prey depletion is not likely. Additionally, due to the interconnecfivity of the canal system, large-bodied fishes such as largemouth bass could move to areas with higher prey abundance if fish prey become locally depleted. The presence of peacock cichlids also might reduce the encounter rate of largemouth bass with fish prey through interference competition. Nevertheless, Experiment 3 documented largemouth bass effectively preying upon fish (and selecting fish at a much higher rate than crayfish) in the presence of peacock cichlids, suggesting that interference is not the causal mechanism. Importantly, population data also suggest that competition with peacock cichlids has not led to significant reductions in largemouth bass abundance in Florida canals (Shafland 1999a, c; Shafland et al. 2001 ; Chapter 6). Characterisfics of the predator, prey, and habitat influence prey use and might be responsible for the heavy reliance of largemouth bass on crayfish prey. For example, predator and prey behavior can modify encounter rates (Cooper et al. 1985; Magnhagen 1986; Grant and Noakes 1987; Winemiller and Taylor 1987). In canals, largemouth bass hunting behavior may increase encounters with crayfish by hunting where crayfish occur (i.e., in structurally complex habitats) and when they are active (i.e., low-light conditions). Moreover, prey fish may take refuge fi-om predators in structurally complex habitats. Habitat structural complexity (e.g., aquatic macrophytes) reduces hunting

PAGE 111

103 efficiency of predators and can lead to reduced encounters with fish relative to non-fish prey (Savino and Stein 1982; Anderson 1984; Hayse and Wissing 1996). Several studies have documented significant consumption of decapod crustaceans by largemouth bass in vegetated (i.e., structurally complex) systems in Florida (Chew 1974; Schramm and Maceina 1986; Cailteux et al. 1996; Huskey and Tumigan 2001; Wheeler and Allen 2003). Cailteux et al. (1996) compared the diets of largemouth bass in vegetated and non-vegetated lakes in central Florida and found that diets were dominated by decapod crustaceans in vegetated lakes but by fish in non-vegetated lakes. They suggested that the structural complexity of aquatic macrophytes was responsible for the diet difference. The southeast Florida canals often have extensive and dense beds of aquafic macrophytes (Chapter 2), potentially an important factor in the observed pattern of significant use of non-fish prey by largemouth bass. The effect of macrophytes on largemouth bass diets in southeast Florida canals could be determined by comparing largemouth bass stomach contents fi-om areas with macrophytes with those from areas without macrophytes (i.e., harvested areas; Chapter 2). The present study generally agrees with previous studies that have found differences in prey vulnerability that are correlated to prey defenses (e.g., Webb 1986; Hoyle and Keast 1 987). The three prey types used in the experiments were chosen to present the predators with prey of differing morphology. Crayfish possess formidable morphological defenses — large chelae and a strong exoskeleton. Bluegills are relafively deep-bodied and possess spines in the dorsal and anal fins. However, sailfin mollies have few morphological attributes that could be considered defensive against subadult and adult largemouth bass or peacock cichlids. Such defenses do not provide absolute

PAGE 112

104 predation refuges for prey (except when body depth exceeds mouth size for gape-hmited predators [e.g., Lawrence 1958; Hambright et al. 1991]), but often increase handling times and the probability of escape (Gillen et al. 1981; Webb 1986; Wahl and Stein 1988; Nilsson and Bronmark 2000). Gape-limited predators typically swallow fish prey headfirst and decapod crustacean prey tail-first to avoid interference fi-om spines or claws (e.g., Lawrence 1 958). This often requires the predator to reorient prey prior to swallowing. To accomplish this, the predator may forcibly release the prey and then rapidly suck the item back into its mouth. Furthermore, prey with defensive morphology may struggle ft-ee from the predator's mouth. In either case, released prey may escape on its own (e.g., reftige in cover) or other predators may interfere and steal the prey item (i.e., kleptoparasifism) or facilitate its escape (e.g., Nilsson and Bronmark 2000). Therefore, attacking large and defensive prey has a potential energetic penalty of increased handling time and potential loss of prey due to escape and therefore defensive prey may be selected by predators at a lower rate than less defensive prey. For largemouth bass, crayfish were less likely than fish to be attacked and more likely to escape if attacked. Largemouth bass also selected crayfish at a low rate relative to fish prey. However, the spines and deep body of bluegills did not confer an advantage over the relatively defenseless sailfm molly if largemouth bass were predators. The lack of difference differences in selection between sailfin mollies and bluegill for largemouth bass may be due to the relatively small size of bluegill prey in the experiment. Larger bluegills, with their more developed spines and deeper bodies, were less profitable prey than were small bluegills (Experiment 1 ) and thus bluegill size may have limited the effecdveness of their morphological defenses. However, bluegills were less profitable

PAGE 113

105 prey than were sailfin mollies for peacock cichlids, possibly due to the morphological defenses of bluegills. This translated into higher selection values for sailfin mollies in Experiment 2. This pattern, including a general refusal by peacock cichlids to eat highly defensive crayfish, suggests that prey morphological defenses may be relatively more effective against peacock cichlids than largemouth bass. In the present study, largemouth bass ate more prey per capita than did peacock cichlids. This is consistent with lower stomach fullness estimates for juvenile peacock cichlids collected in Florida canals (Chapter 4). Similarly, relatively low daily rations have been reported for a peacock cichlid congener, Cichla monoculus (Rabelo and Araujo-Lima 2002). In the present study, predator length was not a factor because individuals were chosen to obtain similar mean and extreme lengths. Predator mass is a potential factor, but largemouth bass are only slightly heavier for their length than are peacock cichlids (unpubl. data). It is possible that largemouth bass are under less stress in the confines of an experimental tank than are peacock cichlids and therefore feed better (see Schreck et al. 1997). Peacock cichlids in tanks are easily disturbed whereas largemouth bass quickly become accustomed to the presence of people near their tanks (pers. obs.). Additionally, peacock cichlids seem to be prone to disease (e.g., Ichthyophthirius multifiliis and opportunisfic bacteria) and shedding of fin epithelium when first collected from canals and placed into a tank (pers. obs.). The outbreak of disease in fishes is often stress-related (Wedemeyer 1 997) and confinement stress correlates with epithelial loss in fishes (e.g., striped bass Morone saxatilis ; Noga et al. 1998). However, the peacock cichlids used in the experiments responded well to appropriate prophylactic and veterinary treatments and began feeding consistently within

PAGE 114

106 one or two days. Another possibility is that largemouth bass may have eaten more prey simply because they were better at capturing it. However, this hypothesis is not supported by the estimates of the probability of attack and success obtained in Experiment 1 which showed that peacock cichlids were equally or more successful in prey capture (Table 4). Conversely, peacock cichlids may have lower energy expenditures than largemouth bass, maintaining body mass and growth with fewer resources than an equivalent largemouth bass. This would likely translate into the ability of peacock cichlids to sustain positive population growth at lower resource levels than would be possible for largemouth bass. Given this case, peacock cichlids would be favored during any resource crunch period. Nevertheless, additional data (e.g., bioenergetics modeling, growth studies) are needed to test this hypothesis. A relative lack of effects of predator assemblage characteristics on prey selection suggested similar effects of both predators on experimental prey assemblages. Similarity of predators has been found in several experimental studies. For example, salamanders ( Notophthalmus viridescens and Ambystoma opacum ) (Morin 1995) and a salamander (Notophthalmus viridescens) and a fish (banded sunfish Enneacanthus obesus) (Kurzava and Morin 1998) had similar predatory effects on tadpole assemblages. Baber and Babbitt (2003) also found fiinctional similarity among three fish species (eastern mosquitofish Gambusia holbrooki flagfish Jordanella floridae, and golden topminnow Fundulus chrysotus ) preying on tadpoles; nevertheless, they found a significant difference between the effect of these predators and the morphologically and behaviorally different walking catfish Clarias batrachus On the other hand, largemouth bass and peacock cichlids, although morphologically similar, expressed some important

PAGE 115

107 differences in their effects on prey. The previously mentioned greater individual prey consumption of largemouth bass resulted in a greater effect of largemouth bass on the odds of golden shiner and crayfish death. As found in Experiment 2, the main difference between the predators was in the consumption of crayfish. In conclusion, the present study provides some support for the assumptions that predators with similar morphology will have similar ability to handle prey and that this ability will lead to similar patterns of prey selection. Therefore, using trophic groupings based on morphology may be a valid methodology for some ecological studies (Briand and Cohen 1984; Harris 1995; but see Paine 1992; Yodzis and Winemiller 1999 for problems with this approach). Moreover, similarity in morphology with native species could be predictive of the diets of nonindigenous species and therefore the potential for dietary overlap and food competition. However, the predictions of experimental data (i.e., that largemouth bass and peacock cichlids would have similar, fish-dominated diets) were not congruent with published field data on largemouth bass diets in canals (i.e., comparable proportions of fish and non-fish prey) and implied that the expectation that morphological similarity will yield similar effects on prey in the field may be too simplistic for complex systems. Thus, the incorporation of ecological interactions, species-specific behavior, and habitat influences may be required for better predictive ability. For example, some altemafive hypotheses were presented in this discussion that may explain the differences in observed diets (i.e., in canals) in the context of prey encounter rates, namely competition, predator or prey behavior, and habitat characteristics. Of these altemafives, the presence of aquatic macrophytes (i.e., habitat

PAGE 116

108 structural complexity) was perhaps the simplest explanation, although the present study did not specifically attempt to refute one or more of these hypotheses. Additionally, the present study suggested that prey morphological defenses (e.g., spines, chelae, and hard exoskeleton) may be relatively more effective against peacock cichlids than largemouth bass (in terms of influencing prey selection) but that peacock cichlids may be more efficient users of food resources than are largemouth bass. Although peacock cichlids might be compefifively superior to largemouth bass if peacock cichlids are indeed more efficient predators; the ability of largemouth bass to utilize nonfish prey (along with apparently high prey abundance) would represent a competition refiige fi-om peacock cichlids. Indeed, population data imply that competition has not had a significant negative effect on largemouth bass abundance (Shafland 1999a, c; Shafland etal. 2001; Chapter 6).

PAGE 117

CHAPTER 6 THE INTRODUCTION OF PEACOCK CICHLIDS INTO SOUTHEAST FLORIDA: POPULATION-LEVEL CONSEQUENCES FOR LARGEMOUTH BASS Introduction The introduction of predatory fishes outside their natural ranges often causes substantial concern for biologists and fishery managers. Nonindigenous predatory fishes are considered likely to cause negative changes in receiving systems due to predation or competition effects on native species (Taylor et al. 1984; Moyle et al. 1986; Moyle and Light 1996a). Recent introductions of predatory snakeheads (Channidae) and Asian swamp eel Monopterus albus into North America, as well as the northern pike Esox lucius eradication efforts in California, have captured the attention of the media and public. Introductions of predatory Nile perch Lates niloticus into Lake Victoria, Africa (e.g., Hughes 1986; Kauftnan 1992; Goldschmidt et al. 1993), and peacock cichlids Cichla sp. into Lake Gatun, Panama (Zaret and Paine 1973), are frequently cited to demonstrate the danger that predatory fishes pose when introduced into new regions. The predatory South American peacock cichlid Cichla ocellaris was intentionally introduced into freshwater canals of southeast Florida during 1984-1987 by the Florida Game and Fresh Water Fish Commission (currently Florida Fish and Wildlife Conservation Commission [FWC]) (Shafland 1995). The purposes of the introduction were to control excessive prey fish biomass, primarily resulting from the abundance of the exotic spotted tilapia Tilapia mariae and to supplement urban recreational fisheries 109

PAGE 118

110 (Shafland 1995). The largemouth bass Micropterus salmoides an ecomorphologically similar native species (Norton and Brainerd 1993), was already present in these canals. Both species currently support valuable recreational fisheries in southeast Florida (Shafland and Stanford 1999). There is continuing speculation that peacock cichlids negatively affect largemouth bass populations (Courtenay and Robins 1989; OTA 1993; Cox 1999; see also Lachner et al. 1970). Indeed, there is evidence that peacock cichlids exert some negative pressures on largemouth bass. Peacock cichlids consume small largemouth bass (Shafland 1999b; unpublished data). For example, Shafland (1999b) documented that largemouth bass occurred in 10% of 156 peacock cichlid stomachs from Tamiami Canal and made up 6% by number of prey. Peacock cichlids and largemouth bass also have similar feeding mechanics (Norton and Brainerd 1993), length-specific gape size (Hill 1998; Chapter 4), and dietary patterns (Lilyestrom and Churchill 1996; Shafland 1999b; Chapter 4, 5). Therefore, peacock cichlids might negatively effect largemouth bass populations via resource competition. Conversely, existing data suggest that peacock cichlids have not affected largemouth bass populations in southeast Florida (Shafland 1995, 1999a, b, c; Shafland and Stanford 1999; Shafland et al. 2001). Additionally, there is no evidence for deleterious effects of peacock cichlids on largemouth bass populations in other regions where peacock cichlids and largemouth bass have been introduced (i.e., Hawaii and Puerto Rico) (Devick 1980; Lilyestrom et al. 1994; Lilyestrom and Churchill 1996). Courtenay (1997) has argued that the absence of long-term, quantitative data prior to fish introductions explains the substantial lack of documentation of negative effects of introduced fishes, especially in Florida. There are no data for southeast Florida canals

PAGE 119

Ill covering the period prior to the initial exotic fish invasions (i.e., 1950s and 1960s; Hill 2002), and relatively few data for the period before peacock cichlids were introduced. Therefore, it is difficult to compare largemouth bass populations before and after the peacock cichlid introduction, in sites with and without peacock cichlids, a powerful approach that could have been used to conduct a more conclusive assessment (e.g., Huckins et al. 2000). However, an analysis using the pre-introduction data on largemouth bass densities does not support the hypothesis of negative effects of peacock cichlids on largemouth bass populations in southeast Florida (Shafland 1999a). Underlying questions about effects of peacock cichlids on largemouth bass is the assumption that south Florida urban canals are favorable habitat for largemouth bass and native fishes in general. If these canals represent marginal habitat (e.g., supporting only sink populations; Armett 1998) then potential effects of peacock cichlids may be relatively unimportant for the regional dynamics of largemouth bass. Based on canal morphometry, variable water physico-chemical parameters, frequent changes in aquatic macrophyte abundance due to control efforts, and presumably strong negative biotic interactions with numerous species of abundant exotic fishes, several authors have argued that urban canals are poor habitat for native fishes (Hogg 1976; Loftus and Kushlan 1987; Courtenay 1997; Annett 1998.). Annett (1998) suggested that the canals represent population sinks for largemouth bass, given a lack of observed spawning within the canals. Conversely, the perception of marginal largemouth bass populations was not supported by my own qualitative observations consisting of hundreds of hours of field work (e.g., electrofishing) and observations in the canal system. For example, it is clear that largemouth bass spawn and successfully recruit in southeast Florida canals (Chapter

PAGE 120

112 4; pers. obs.). Indeed, I was surprised by the range of sizes and the abundance of largemouth bass in these canals. I used available data (i.e., FWC database; FWC 2001) to test the hypotheses that 1) largemouth bass population densities in the canals were similar to densities found in other Florida systems (i.e., Florida lakes and streams) and 2) largemouth bass population densities in canals have been negatively effected by the introduction of peacock cichlids. Shafland (1999a) provided data on Black Creek Canal before, during, and after the introduction of peacock cichlids (see also Shafland et al. 1985; Shafland et al. 2001). The present statistical analysis is the first to compare canals with statewide data for other types of systems (i.e., lakes and streams) and to compare canals containing peacock cichlids with canals that lack peacock cichlids. If canal populations of largemouth bass are similar to other systems in Florida, then this result would question the common assumption that urban canals in south Florida are poor largemouth bass habitat. Moreover, a significant reduction in abundance of largemouth bass due to peacock cichlids would be unlikely. Methods I used the FWC database (FWC 2001) which provides data from statewide fish populafion sampling conducted by FWC personnel from 1 956 to 2000 (mostly in the 1980s and 1990s). The FWC database contains blocknet data and electrofishing data. The blocknet data set contained 790 observations (i.e., one or more blocknets [using either rotenone or detonator cord] at a site) for 71 water bodies and the electrofishing data set contained 934 observadons (i.e., one or more boat electrofishing transects at a site) for 300 Florida water bodies. A type I error rate of a = 0.05 was used for all analyses. I used

PAGE 121

113 a type II error rate of (3 0.20 to estimate detectable effect sizes (i.e., A) for tests that did not yield significant differences (Ott 1993). The present study focused on harvestable largemouth bass (i.e., > 254 mm total length [TL]) because electrofishing is not the best method for estimating the abundance of age-0 fish (i.e., age-0 largemouth bass may not be fully recruited to the gear) (Reynolds 1996) and harvestable fish are of immediate fishery importance. Indeed, if harvestable-size largemouth bass abundance is maintained over time, recruitment is confirmed in the system. Most comparisons in the present study used electrofishing catch-per-unit-effort (CPUE) data. This method is appropriate for monitoring population abundance of harvestable-size fish (Reyonlds 1996) and is a better method than is the use of blocknets (Tate et al. 2003). I combined all blocknet or electrofishing observations to create a mean value if a water body had more than one observation; regional and statewide means were obtained from these water body means (PROC MEANS; SAS Institute, Gary, North Carolina). I used estimates of numerical abundance (fish/ha) and standing crop (kg/ha) from blocknets as well as numerical (fish/min) and biomass (g fish/min) electrofishing CPUE for the analysis comparing canals to other Florida systems. Differences among statewide blocknet or electrofishing CPUE data for water body types (i.e., Florida canals, lakes, and streams) were determined by one-way analysis of variance (ANOVA), using the general linear models procedure (PROC GLM) followed by the Tukey HSD multiple comparison procedure in SAS (SAS Institute, Cary, North Carolina). I used electrofishing CPUE (i.e., numerical and biomass) to compare largemouth bass populations in south Florida canals with south Florida lakes and to compare canals

PAGE 122

114 with and without peacock cichlids. I considered water bodies to be located in south Florida if they were in or south of Charlotte, Glades, and Martin counties and Lake Okeechobee. Differences in electrofishing CPUE data for south Florida canals versus south Florida lakes and for canals with and canals without peacock cichlids were tested by two-sample t-tests following a test for equality of variance (PROC TTEST; SAS Institute, Gary, North Garolina). Results Based on electrofishing data, the density and biomass of harvestable largemouth bass varied among different types of water bodies sampled throughout Florida, with GPUE approximately one and a half to two times higher in lakes than in canals and streams (Figures 18 and 19). Harvestable largemouth bass density and biomass based on blocknet data showed a slightly different pattern, with largemouth bass being approximately 1 0-fold more abundant in canals and lakes compared to streams (Figures 20 and 21). These comparisons based on statewide samples potentially confound differences among types of water bodies with latitudinal effects (i.e., canals are more common in southern Florida). I therefore compared south Florida canals with south Florida lakes (no south Florida streams were included in the database), but found no significant differences at the a priori a = 0.05 level (Figures 22 and 23). The analysis of electrofishing GPUE data also failed to detect a significant difference between canals with and canals without peacock cichlids (Figures 24 and 25).

PAGE 123

115 0.6* Florida Canals Florida Lakes Florida Streams Figure 18. Mean ( SE) electrofishing CPUE (fish/min) for harvestable largemouth bass (> 254 mm TL) in Florida canals (N = 43), lakes (N = 1 14), and streams (N = 21). Mean CPUE for Florida lakes was significantly greater than Florida canals and streams (ANOVA, F2,i75 = 8.90, p = 0.0002). Data were obtained from FWC 2001 Figure 19. Mean ( SE) electrofishing CPUE (biomass; g/min) for harvestable largemouth bass (> 254 mm TL) in Florida canals (N = 43), lakes (N = 114), and streams (N = 21). Mean CPUE for Florida lakes was significantly greater than Florida canals and streams (ANOVA, F2,i75= 6.52, p = 0.0019). Data were obtained from FWC 2001.

PAGE 124

116 Florida Canals Florida Lakes Florida Streams Figure 20. Mean ( SE) number (fish/ha) of harvestable largemouth bass (> 254 mm TL) in Florida canals (N = 1 8), lakes (N = 39), and streams (N = 14) based on blocknet data. Mean CPUE for Florida canals and lakes were significantly greater than Florida streams (ANOVA, F2,68 = 5.57, p = 0.0057). Data were obtained from FWC 2001. Florida Canals Florida Lakes Florida Streams Figure 21. Mean ( SB) biomass (kg/ha) of harvestable largemouth bass (> 254 mm TL) in Florida canals (N = 1 8), lakes (N = 49), and streams (N = 14) based on blocknet data. Mean CPUE for Florida canals and lakes were significantly greater than Florida streams (ANOVA, F2,68 = 4.79, p = 0.01 13). Data were obtained fi-om FWC 2001.

PAGE 125

117 Figure 22. Mean ( SE) electrofishing CPUE (fish/min) for harvestable largemouth bass (> 254 mm TL) in south Florida canals (N = 39) and lakes (N = 1 1 ). Means were not significantly different (t-test, to.o5,4g= -1-83, p = 0.0734; A = 0.038 fish/min). Data were obtained from FWC 2001. Figure 23. Mean ( SE) electrofishing CPUE (biomass; g/min) for harvestable largemouth bass (> 254 mm TL) in south Florida canals (N = 39) and lakes (N = 11). Means were not significantly different (t-test, to.o5,48 = -0.99, p = 0.3262; A = 24.6 g/min). Data were obtained fi-om FWC 2001

PAGE 126

118 0.4 Canals with Peacock Cichlids Canals without Peacock Cichllds Figure 24. Mean ( SE) electrofishing CPUE (fish/min) for harvestable largemouth bass (> 254 mm TL) in southeast Florida canals with (N = 27) and without (N = 13) peacock cichlids. Means were not significantly different (t-test, to.os, 38 = 1 -33, p = 0. 1 92; A = 0. 1 69 fish/min). Data were obtained from FWC 200 1 250 Canals with Peacock Cichllds Canals without Peacock Cichllds Figure 25. Mean ( SE) electrofishing CPUE (biomass; g/min) ( 1 SE) for harvestable largemouth bass (> 254 mm TL) in southeast Florida canals with (N = 27) and without (N = 13) peacock cichlids. Means were not significantly different (t-test, to.o5,38 = 0.57, p = 0.575; A = 128.5 g/min). Data were obtained from FWC 2001

PAGE 127

119 Discussion The similarity of largemouth bass populations in Florida canals to other types of aquatic systems in Florida on both a statewide and regional (i.e., south Florida) basis is somewhat surprising given the general impression from the literature that urban canals are poor habitat for native fishes (e.g., Hogg 1976; Loftus and Kushlan 1987; Courtenay 1997; Annett 1998). In contrast, these results imply that canals are not poor habitats overall for largemouth bass. In particular, there is little evidence that canal populations of largemouth bass are sink populations (Annett 1998). Indeed, largemouth bass successftilly spawn and recruit within the canal system, based on the presence of nesting fish, free-swimming fiy, and small juveniles (Chapter 4; pers. obs.). Shafland and colleagues have presented additional data that support the hypothesis that canals in southeast Florida are not marginal habitats. Mean electrofishing CPUE (fish/min) of harvestable largemouth bass in canals with peacock cichlids are comparable to the mean statewide value used by FWC (Figure 26; see Shafland et al. 2001). Indeed, electrofishing data for several canals showed higher CPUE estimates for harvestable largemouth bass (i.e., fish/min) than the statewide average (Shafland et al. 2001). Recreational sport fishing for largemouth bass in southeast Florida, where peacock cichlids also occur, has an estimated annual economic value of US$5 million (Shafland and Stanford 1 999). Fishing tournaments in the canals yield catch rates that may exceed tournaments held in famous fishing waters (e.g., Lake Okeechobee) (Shafland et al. 2001). Recreational angler creel data collected for two large canals in Miami-Dade County rank them both in the top 10 statewide for angler effort directed at largemouth bass and in angler success. Black Creek Canal was number one (182

PAGE 128

120 hr/ha/yr) and Tamiami Canal was number five (48 hr/ha/yr) in angler effort directed at largemouth bass (Shafland and Stanford 1999). Tamiami Canal had the highest angler success rate for largemouth bass (0.41 fish/hr) and Black Creek Canal was tenth (0.30 fish/hr) (Shafland and Stanford 1999). Clearly, persistent largemouth bass populations exist in the canals regardless of negative abiotic or biotic influences. 1986-1989 1986-1999 1997-1999 Statewide Average Figure 26. Mean electrofishing CPUE (fish/min) for harvestable largemouth bass in south Florida canals that have peacock cichlids. The data for 1986-1989 include 23 canals; the data for 1986-1999 include 34 canals; and the data for 1997-1999 include 32 canals. The data and the value for statewide average were obtained from Shafland et al. 2001. Although electrofishing and blocknet data showed different trends in the statewide habitat comparison, canal populations of largemouth bass were not distinguishable from another Florida system (i.e., either lakes or streams) in both cases. Tate et al. (2003) noted differences between electrofishing and blocknet data for largemouth bass populations in two vegetated lakes in north Florida. It is therefore not surprising to have somewhat conflicting trends in the larger and more heterogeneous

PAGE 129

121 FWC database that encompasses the entire state. Also, it is possible that the physical nature of the canal systems in Florida, being intermediate between lakes and streams, may be reflected in this inconsistent outcome. In contrast to the statewide electrofishing data, no significant differences were found between south Florida canals and lakes. Both tests (i.e., fish/min and g/min) could detect differences of about 10% with reasonable power (i.e., 1 P = 0.80) (Figures 22 and 23). These results suggest that regional influences may be important for largemouth bass and that south Florida canals are not marginal habitat. Introductions of piscivores, especially species similar to native fishes, are generally predicted to result in negative effects for the native species (Taylor et al. 1984; Moyle et al. 1986; Moyle and Light 1996a). Hypothetically, peacock cichlids might negatively influence an ecomorphologically similar native species such as the largemouth bass by predation or food competition. However, the results of the present study did not support the hypothesis that the introduction of peacock cichlids has caused major reductions in largemouth bass abundance. Largemouth bass are successftil in south Florida canals despite the presence of many nonindigenous species, including peacock cichlids. It is plausible that the density of largemouth bass would be even greater if not for the possible deleterious effects of peacock cichlids. However, the available data provide no support for this hypothesis; largemouth bass abundance and biomass was not significantly different in canals with and without peacock cichlids. However, an assumption of this comparison is that these canals differ only in the presence or absence of peacock cichlids. This assumption is not valid. The geographic distribution of these two types of canals is adjacent, but non-

PAGE 130

122 overlapping. Although still within the same south Florida counties, canals without peacock cichlids are north and west of those with peacock cichlids. Therefore, geographic effects might confound peacock cichlid effects. In their fisheries survey of south Florida canals, Shafland et al. (2001) reported a number of differences between the two types. Peacock cichlids are found in canals that are often more urban in character (but there is wide variation), have a more box-cut morphometry, are deeper, have narrower littoral zones, and have less shoreline vegetation than the canals lacking peacock cichlids. Canals in the area occupied by peacock cichlids also have greater abundance and variety of nonindigenous fishes. Interestingly, many of these factors suggest that canals lacking peacock cichlids should support higher densities of largemouth bass than canals with peacock cichlids, independent of the actual presence or effects of peacock cichlids. Indeed, Shafland et al. (2001) commented that the canals lacking peacock cichlids were better habitat for sunfish Lepomis spp. However, the analysis failed to distinguish between mean electro fishing CPUE of harvestable largemouth bass for these systems. On the other hand, given reasonably high power (i.e., 1 P = 0.80), the detectable effect sizes of the analyses were > 50% difference in abundance (i.e., fish/min) (Figure 24) and > 70% difference in biomass (i.e., g/min) (Figure 25), rather large differences. Therefore, further investigation may be warranted concerning the relationship of largemouth bass populations in the two canal types. The results of the present study suggest that peacock cichlids have not had major negative effects on largemouth bass populations. However, the present study does not provide data investigating possible changes in the abundance of largemouth bass in the canals since the peacock cichlid introduction. Appropriate data were lacking in the FWC

PAGE 131

123 database to conduct such an analysis (FWC 2001). However, Shafland et al. (2001) summarized largemouth bass population data from blocknets for the pre-introduction period (i.e., 1980-1984) and a transitional period when peacock cichlids had been introduced but were not yet considered established (i.e., 1984-1987), as well as data for canals with peacock cichlids and canals lacking peacock cichlids for the period of 19801997 (i.e., covering four years of pre-introduction, three years of transitional period, and ten years of post-introduction data) (Figure 27). These data include both harvestable and sub-harvestable largemouth bass. Although error estimates were not presented, these data do not support a hypothesis of reduction in largemouth bass abundance in response to effects of peacock cichlids. Figure 27. Mean biomass (kg/ha) of largemouth bass in south Florida canals based on blocknet data. The pre-introduction period includes data from 1980-1984 (N = 22 canals). The transitional period includes data from 1984-1987 (N = 22 canals), the time period when peacock cichlids were introduced but not yet established. Canals with peacock cichlids includes data from 1980-1997 (N = 28 canals). Canals lacking peacock cichlids includes data from 1980-1997 (N = 1 1 canals). Data obtained from Shafland et al. 2001.

PAGE 132

124 Despite anecdotal reports and theoretical predictions suggesting negative effects of peacock cichlids in Florida (Courtenay and Robins 1989; OTA 1993; Cox 1999; see also Lachner et al. 1970), no published data support the hypothesis that peacock cichlids have significantly reduced largemouth bass populations. For example, a major governmental report on nonindigenous species cited a non-peer reviewed environmental magazine (i.e., Florida Environments; McClelland 1992) as the source for a conclusion that peacock cichlids were "slowly reducing populations of indigenous bass and bream" (p. 263; OTA 1993). The cited article (i.e., McClelland 1992) presented no data, only a speculative statement not attributed to any source. Similarly, Cox (1999) listed the peacock cichlid in Florida as an "invasive exotic", a designation that strongly implies that the peacock cichlid is a harmful species. However, only a brief mention of peacock cichlid predation on native fishes and invertebrates was given as the justification (Cox 1999). The lack of detectable effect may be due to many factors. The interaction strength of these mechanisms (i.e., predation and competition), and thus the magnitude of the effect, may be low. Additionally, abiotic factors or other biotic interactions, including intraspecific interactions, may be preeminent. Such dynamics may keep the abundance of predatory fishes below the level where competitive interactions between them become important. For example, Zom and Seelbach (1995) concluded that abiotic or biotic processes, especially prior abiotic events that reduce recruitment, were responsible for the lack of a positive linear relationship between predatory fish biomass and available fish habitat in warm water streams. Furthermore, a nonindigenous fish

PAGE 133

125 could integrate into the existing fish assemblage under non-equilibrium conditions, if the system is not saturated with species, or if there are underexploited resources. The common knowledge, often repeated in the literature, that south Florida canals are uniformly poor habitat for native fishes and that nonindigenous fishes therein have significantly reduced native species abundance was not supported by the data for largemouth bass. Other recent studies also have quesfioned the percepfion of across-theboard reduced native fish populations in canals, as well as the extent of negative effects that nonindigenous fishes exert in south Florida. Recent fisheries studies conducted by the FWC in south Florida canals document the persistence and abundance of native fishes, including largemouth bass, bluegill Lepomis macrochirus and redear sunfish Lepomis microlophus (Shafland 1999a, c; Shafland and Stanford 1999; Shafland et al. 2001). Indeed, Shafland (1996) argued that few data document negative effects of nonindigenous fishes in south Florida and urged a scientific evaluation of data to replace the anecdotal reports that form the usual basis of discussion on this topic. Moreover, in an analysis of the most comprehensive and long-term dataset for south Florida, covering the greater Everglades system and adjacent canals, Trexler et al. (2000) concluded that nonindigenous fishes had not significantly affected native fishes, especially in the less disturbed habitats. Clearly, the interactions of native and nonindigenous fishes in this region require a great deal of additional research. One priority is the documentation of the influence of nonindigenous fishes on the aquatic systems of south Florida (Shafland 1996; Courtenay 1997). Of particular interest, however, is the fact that nearly all of the available information on the effects of nonindigenous fishes in south Florida is contrary to the

PAGE 134

126 current paradigm of invasion biology. Much of the ecological theory adopted by the field of invasion biology (see Vermeij 1996; Williamson 1996), particularly with respect to freshwater systems, is based on island biogeography and tightly-structured, stable communities (sensu Clements 1936) (Stauffer 1984; Li and Moyle 1999). The overwhelming consensus based on this theoretical background is that species introductions lead to significant negative consequences for native species and ecosystems (Chapter 1 ). Finding expected results merely reinforces existing dogma. It is with the study of systems that yield unexpected and contrary results that this field will advance and produce a mature, pluralistic, and ultimately more useful theory.

PAGE 135

CHAPTER 7 SUMMARY Florida has numerous established and reproducing nonindigenous fishes (Fuller et al. 1999; Hill 2002). Such nonindigenous fishes represent an important challenge to fishery managers and environmental regulators in Florida and throughout the USA (Magnuson 1976; Hill 2002). Clearly, some introduced fishes in the USA have become invasive; yet others have had relatively little detectable effect on their receiving system (Moyle and Light 1996a). The intentional introduction of the peacock cichlid Cichla ocellaris into the native range of the ecomorphologically similar largemouth bass Micropterus salmoides in southeast Florida was made to increase recreational fishing opportunities and to exert biological control over other nonindigenous fishes (Shafland 1995). Nevertheless, purposeful introductions are controversial and concern has been expressed over the possible negative effects of peacock cichlids in southeast Florida (Courtenay and Robins 1989; OTA 1993; Courtenay 1994; Cox 1999). Indeed, introduced piscivores are considered likely to cause deleterious effects in their receiving systems (Moyle and Light 1996a). Moreover, morphologically similar species should use similar resources and thus be more prone to compete with one another. Concern over the potential effects of nonindigenous fishes has prompted increasing regulation, including the listing or proposed listing of additional fish species as injurious wildlife by the United States Fish and Wildlife Service (USFWS) (e.g., 127

PAGE 136

128 snakeheads Channa spp. and black carp Mvlopharyngodon piceus ; Nico et al. 2001). Risk assessments are used to provide information and assist in decision-making regarding possible regulation of nonindigenous species. The Generic Nonindigenous Aquatic Organism Risk Analysis Review Process (i.e., RAM methodology; RAM Committee 1998) is being used by the USFWS as their risk assessment method. The black carp risk assessment was a test case for the application of the methodology (Nico et al. 2001); however, there have been no previous tests of the predictions of the RAM method algorithms. My application of the RAM method using Cichla as a test case elucidated shortcomings of the method (Chapter 3). In particular, I pointed out that the precautionary principle (FAO 1995) is implicit in the RAM algorithm (i.e., the presupposition that any nonindigenous species represents an unacceptable risk if established or if establishment is of at least medium probability). Given this approach, the risk assessment outcome depends solely on the estimate of the Probability of Establishment (i.e., disregards the risks estimated under the Consequences of Establishment). Thus, any species of medium or greater Probability of Establishment has at least a medium Overall Risk Potential (ORP) (i.e., unacceptable risk requiring mitigation; RAM Committee 1998). Furthermore, the RAM methodology confounds the role of the risk assessment (i.e., to estimate risk) with that of society (i.e., to determine the acceptable level of risk). Additionally, 1 recommended that the ORP be deemphasized in favor of retaining the Probability of Establishment (and Spread) and the Consequences of Establishment as the primary model outputs. Part of the concern over the introduction of peacock cichlids into southeast Florida stems from the hypothesis that similarity in morphology leads to similarity in

PAGE 137

129 resource use (and therefore potential competition). 1 examined this issue for juvenile largemouth bass and peacock cichlids by describing their gape width relations under the assumption that trophic morphology equates to prey handling ability (Wainwright and Richard 1995), by comparing ontogenetic shifts in diet, and by estimating the degree of overlap in prey use between the two species (Chapter 4). Furthermore, I tested this hypothesis (i.e., similar morphology equals similar resource use) by documenting prey handling ability and prey selectivity of subadult and adult largemouth bass and peacock cichlids under experimental conditions (Chapter 5). I found considerable evidence for similarity in prey use by largemouth bass and peacock cichlids and thus support for the hypothesis of similar morphology equals similar resource use. Juveniles had similar gape width relations and ontogenetic dietary shifts, and overlapped considerably in diet. Larger individuals had similar prey handling ability and prey selection patterns. Additionally, the results suggested that largemouth bass and peacock cichlids have similar effects on the prey base. Nevertheless, I documented several important differences. In particular, the results suggested that largemouth bass are more capable of consuming highly defensive prey (e.g., crayfish) and that peacock cichlids are better at feeding on relatively defenseless prey (e.g., sailfm molly). Indeed, although largemouth bass generally avoided crayfish prey if alternative, more profitable prey (i.e., fish) were present, largemouth bass had exclusive use of crayfish as a food resource. Furthermore, largemouth bass consistently consumed more prey than did peacock cichlids of equal size. Given the overall similarity in prey handling ability and prey selection under experimental conditions (Chapter 5), the differences in diets between largemouth bass

PAGE 138

130 and peacock cichlids in Tamiami Canal were surprising (i.e., non-fish prey making up about five times more [as a percentage by number] of the largemouth bass diet compared to the peacock cichlid diet; Shafland 1999b). This difference is suggesfive of compefifion-induced resource partifioning (Sale 1979). Nevertheless, given the apparently high abundance of fish prey in southeast Florida canals, a simpler explanation might be differential encounter rates between prey types for largemouth bass due to habitat structural complexity (i.e., aquatic macrophytes) leading to a broadening of diet (Anderson 1984; Cailteux et al. 1996). This question would be a potentially fhiitful area for additional research. Although some authors have suggested that peacock cichlids have had negative effects on largemouth bass and other native fishes (Courtenay and Robins 1989; OTA 1993; Cox 1999; see also Lachner et al. 1970), no population data have been presented that suggest any such effects. A partial explanation might be the relative lack of preintroduction data, a common concern in evaluations of the effects of fish introductions (Courtenay 1997). Shafland (1999a) presented preand post-introduction data for Black Creek Canal and concluded that peacock cichlids had not significantly reduced any native fish population. I used the Florida Fish and Wildlife Conservation Commission database (FWC 2001) to further investigate this question by comparing populations of largemouth bass in canals with populations in Florida lakes and streams (Chapter 6). Because of potential latitudinal effects, I also compared canals to south Florida lakes. Additionally, I compared largemouth bass populations in canals that contain peacock cichlids to canals that lack peacock cichlids. My results suggest that canals are not marginal largemouth

PAGE 139

131 bass habitat and do not support a hypothesis of a reduction in largemouth bass abundance due to peacock cichlids. Although the direct effects of predation and food competition by introduced peacock cichlids hypothetically should have negative effects on largemouth bass populations, existing data do not suggest reductions in largemouth bass populations (Shafland 1999a; Shafland et al. 2001; Chapter 6). Observations suggest that food resources (at least fish prey) may not be limiting for these predators. Additionally, indirect mechanisms, perhaps mediated through prey or human exploitation, may have facilitative effects that are as yet unknown (see Stone and Roberts 1991). Indeed, the potential for positive indirect effects of species introductions is an intriguing area for fiiture research. Such effects may well be responsible for many instances when nonindigenous species have little apparent effect on receiving communities. It can be argued that the predominant theoretical view of invasion biology is based upon the Clementsian concept of organismic communities (Clements 1936), namely that communities are stable, highly-structured, biotically-controlled, and persistent in time and space (Chapter 1). Under such a concept, species introductions are significant perturbations with effects rippling throughout entire communities. Of particular concern is the prediction of increased rates of extinction concurrent with increased rates of invasion (MacArthur and Wilson 1 967). On the other hand, many species introductions seem to occur with little or no detectable negative effect on other species (Williamson 1 996). Such integration is extremely difficult to accommodate under the prevailing view within invasion biology (see Stauffer 1 984; Vermij 1996; Williamson 1996; Li and Moyle 1999). Most explanatory arguments attack the perceived

PAGE 140

132 flaws of the studies that fail to document negative effects, point out a lack of preintroduction data, or claim that sufficient time has not passed for the negative effects to surface. Indeed, a common prediction is that a species that is not a problem today will become a problem eventually. Nevertheless, the results of this dissertation call into question the uncritical acceptance of the presupposition that nonindigenous fishes will invariably cause ecological harm. Instead of adapting invasion biology in such a way as to accommodate contrary results, perhaps invasion biology should be expanded to include these results into a more pluralistic and comprehensive whole. For example, integration of invading species into receiving systems becomes a theoretical possibility if communities can be viewed as loosely-structured, primarily abiotically-controlled, stochastic, and transient in time and space (i.e., individualistic concept; Gleason 1926). Given the considerable evidence that many North American fish communities are not saturated with species (Gido and Brown 1999), that species introductions frequently result in permanent increases in species richness (Vermeij 1991; Gido and Brown 1999; Rosenzweig 2001), and that community structure often is transient (Graham et al. 1996), it seems necessary to incorporate Gleasonian theory into invasion biology.

PAGE 141

REFERENCES Aggus, L. R., and G. V. Elliot. 1975. Effects of cover and food on year-class strength of largemouth bass. pp. 317-322 in H. Clepper, editor. Black bass biology and management. Sport Fishing Institute, Washington, D. C. Ambrose, R. F., R. J. Schmitt, and C. W. Osenberg. 1996. Predicted and observed environmental impacts: can we foretell ecological change? pp. 345-369 in R. J. Schmitt and C. W. Osenberg, editors. Detecting ecological impacts: concepts and applications in coastal habitats. Academic Press, San Diego, California. Anderson, O. 1984. Optimal foraging by largemouth bass in structured environments. Ecology 65:851-861. Annett, C. A. 1998. Hunting behavior of Florida largemouth bass, Micropterus salmoides floridanus in a channelized river. Environmental Biology of Fishes 53:75-87. Annett, C, J. Hunt, and E. D. Dibble. 1996. The compleat bass: habitat use patterns of all stages of the life cycle of largemouth bass. pp. 306-314 in Multidimensional approaches to reservoir fisheries management, American Fisheries Society Special Publication 16, Bethesda, Maryland. Baber, M. J., and K. J. Babbitt. 2003. The relative impacts of native and introduced predatory fish on a temporary wetland tadpole assemblage. Oecologia 136:289295. Bain, M. B., and L. A. Helfiich. 1983. Role of male parental care in survival of larval bluegills Lepomis macrochirus Transactions of the American Fisheries Society 112:47-52. Baltz, D. M. 1991 Introduced fishes in marine systems and inland seas. Biological Conservation 56:151-177. Baltz, D. M., and P. B. Moyle. 1993. Invasion resistance to introduced species by a native assemblage of California stream fishes. Ecological Applications 3:246-255. Bettoli, P. W., M. J. Maceina, R. L. Noble, and R. K. Betsill. 1992. Piscivory in largemouth bass as a flinction of aquafic vegetation abundance. North American Journal of Fisheries Management 12:509-516. 133

PAGE 142

134 Briand, F., and J. E. Cohen. 1984. Community food webs have scale-invariant structure. Nature 307:264-266. Buntz, J., and C. S. Manooch, III. 1968. Tilapia aurea Steindachner, a rapidly spreading exotic in south central Florida. Proceedings of the Annual Conference Southeastern Association of Game and Fish Commissioners 22:495-501. Burks, R. L., E. Jeppesen, and D. M. Lodge. 2001 Littoral zone structures as Daphnia refiigia against fish predators. Limnology and Oceanography 46:230-237. Cailteux, R. L., W. F. Porak, S. Crawford, and L. L. Connor. 1996. Differences in largemouth bass food habits and growth in vegetated and unvegetated northcentral Florida lakes. Proceedings of the Annual Conference Southeastern Association of Fish and Wildhfe Agencies 50:201-21 1. Cariander, K. D. 1977. Handbook of freshwater fishery biology, volume 2. Iowa State University Press, Ames, Iowa. Carpenter, S. R., J. F. Kitchell, and J. R. Hodgson. 1985. Cascading trophic interactions and lake productivity. Bioscience 35:634-639. Carpenter, S. R., J. F. Kitchell, J. R. Hodgson, P.A. Cochran, J. J. Elser, M. M. Elser, D. M. Lodge, D. Kretchmer, X. He, and C. N. von Ende. 1987. Regulation of lake primary productivity by food web structure. Ecology 68:1863-1867. Chesson, J. 1978. Measuring preference in selective predation. Ecology 59:21 1-215. Chesson, J. 1983. The estimation and analysis of preference and its relationship to foraging models. Ecology 64:1297-1304. Chew, R. L. 1974. Early life history of the Florida largemouth bass. Fishery Bulletin 7, Florida Game and Fresh Water Fish Commission, Tallahassee, Florida. Chick, J. H., and C. C. Mclvor. 1997. Habitat selection by three littoral zone fishes: effects of predation pressure, plant density, and macrophyte type. Ecology of Freshwater Fish 6:27-35. Churchill, T. N., P. W. Bettoli, D. C. Peterson, W. C. Reeves, and B. Hodge. 2002. Angler conflicts in fisheries management: a case study of the striped bass controversy at Norris Reservoir, Tennessee. Fisheries 27(2): 1019. Clements, F. C. 1936. Nature and structure of the climax. Journal of Ecology 24:252-284. Colwell, R. K., and D. J. Futuyma. 1971. On the measurement of niche breadth and overlap. Ecology 52:567-576.

PAGE 143

135 Cooper, S. D., D. W. Smith, and J. R. Bence. 1985. Prey selection by freshwater predators with different foraging strategies. Canadian Journal of Fisheries and Aquatic Sciences 42:1720-1732. Courtenay, W. R., Jr. 1993. Biological pollution through fish introductions, pp. 35-61 in B. N. McKnight, editor. Biological pollution: the control and impact of invasive exotic species. Indiana Academy of Science, Indianapolis, Indiana. Courtenay, W. R., Jr. 1994. Non-indigenous fishes in Florida, pp. 57-63 in D. C. Schmitz and T. C. Brown, editors. An assessment of invasive non-indigenous species in Florida's public lands. Technical Report No. TSS-94-100. Florida Department of Environmental Protection, Tallahassee, Florida. Courtenay, W. R., Jr. 1995. The case for caution with fish introductions. American Fisheries Society Symposium 15:413-424. Courtenay, W. R., Jr. 1997. Nonindigenous fishes, pp. 109-122 in D. S. Simberloff, D. C. Schmitz, and T. C. Brown, editors. Strangers in paradise. Island Press, Washington, D. C. Courtenay, W. R., Jr., and C. R. Robins. 1973. Exotic aquatic organisms in Florida with emphasis on fishes: a review and recommendations. Transactions of the American Fisheries Society 102:1-12. Courtenay, W. R., Jr., and C. R. Robins. 1989. Fish introductions: good management, mismanagement, or no management? Reviews in Aquatic Sciences 1 :1 59-1 72. Courtenay, W. R., Jr., H. F. Sahlman, W. W. Miley, II, and D. J. Herrema. 1974. Exotic fishes in tresh and brackish waters of Florida. Biological Conservation 6:292-302. Cox, G. W. 1999. Alien species in North America and Hawaii: impacts on natural ecosystems. Island Press, Washington, D. C. Crawford, S. S. 2001. Salmonine introductions to the Laurentian Great Lakes: an historical review and evaluation of ecological effects. Canadian Special Publication of Fisheries and Aquatic Sciences 132:1-205. Crowder, L. B., and F. P. Binkowski. 1983. Foraging behaviors and the interaction of alewife, Alosa pseudoharengus and bloater, Coregonus hoyi Environmental Biology of Fishes 8:105-113. Declerck, S., G. Louette, T. De Bie, and L. De Meester. 2002. Patterns of diet overlap between populations of non-indigenous and native fishes in shallow ponds. Journal of Fish Biology 61 :1 1 82-1 197.

PAGE 144

136 Devick, W. S. 1972. Life history study of the tucunare, Cichla ocellaris Project F-4-R17. Hawaii Department of Land and Natural Resources, Honolulu, Hawaii. Devick, W. S. 1980. Studies on the relationship between tucunare and largemouth bass. Project F-14-R-3. Hawaii Department of Land and Natural Resources, Honolulu, Hawaii. Diamond, J. M. 1975. Assembly of species communities, pp. 342-444. in M. L. Cody and J. M. Diamond, editors. Ecology and evolution of communities. Harvard University Press, Cambridge, Massachusetts. Elton, C. S. 1958. The ecology of invasions by animals and plants. Methuen and Co., Inc., London, England. Emlen, J. M. 1966. The role of time and energy in food preference. American Naturalist 100:611-617. Erdman, D. S. 1969. Culture and stocking of peacock bass. Project F-1-17. Puerto Rico Department of Natural Resources, San Juan, Puerto Rico. Erdman, D. S. 1984. Exotic fishes of Puerto Rico. pp. 162-176. in W. R. Courtenay, Jr. and J. R. Stauffer, Jr., editors. Distribution, biology, and management of exotic fishes. Johns Hopkins University Press, Baltimore, Maryland. FAO. 1995. Precautionary approach to fisheries. Part 1 : Guidelines on the precautionary approach to capture fisheries and species introductions. FAO Fisheries Technical Paper 350, part 1. FAO, Rome, Italy. FAO. 1996. Precautionary approach to fisheries. Part 2: Scientific papers. FAO Fisheries Technical Paper 350, part 2. FAO, Rome, Italy. Fausch, K. D. 1988. Tests of competition between native and introduced salmonids in streams: what have we learned? Canadian Journal of Fisheries and Aquatic Sciences 45:2238-2246. Fausch, K. D., S. Nakano, and S. Kitano. 1997. Experimentally induced foraging mode shift by sympatric charrs in a Japanese mountain stream. Behavioral Ecology 8:414-420. Fontenele, O. 1 950. Contribuciao para o conhecimento da biologia dos tucunares (Actinopterygii, Cichlidae), em captiveiro, aparalho de reproducao, habitos de desova e incubacao. Revista Brasiliera de Biologia 10:503-519. Fryer, G., and T. D. lies. 1972. The cichlid fishes of the Great Lakes of Africa. TFH Publications, Neptune City, New Jersey.

PAGE 145

137 Fuller, P. L., L. G. Nico, and J. D. Williams. 1999. Nonindigenous fishes introduced into inland waters of the United States. American Fisheries Society, Special Publication 27, Bethesda, Maryland. FWC (Florida Fish and Wildlife Conservation Commission). 2001 Freshwater fishes of Florida database and GIS data layers. Office of Environmental Services, Tallahassee, Florida. Carman, G. C, and L. A. Neilson. 1982. Piscivory by stocked brown trout, Salmo trutta, and its impact on the nongame fish community of Bottom Creek, Virginia. Canadian Journal of Fisheries and Aquatic Sciences 39:862-869. Garvey, J. E., R. A. Wright, R. A. Stein, and K. H. Ferry. 2000. Evaluating how localand regional-scale processes interact to regulate growth of age-0 largemouth bass. Transactions of the American Fisheries Society 129:1044-1059. Cause, G. F. 1934. The struggle for existence. Williams and Wilkins, Baltimore, Maryland. http://www.ggause.com/Contgau.htm. George, E. L., and W. F. Hadley. 1979. Food and habitat partitioning between rock bass ( Ambloplites rupestris ) and smallmouth bass ( Micropterus dolomieui ) young of the year. Transactions of the American Fisheries Society 108:253-261. Gido, K. B., and J. H. Brown. 1999. Invasion of North American drainages by alien fish species. Freshwater Biology 42:387-399. Gilbert, C. R. 1987. Zoogeography of the freshwater fish fauna of southern Georgia and peninsular Florida. Brimleyana 13:25-54. Gilinsky, E. 1984. The role of fish predation and spatial heterogeneity in determining benthic community structure. Ecology 65:455-468. Gillen, A. L., R. A. Stein, and R. F. Carline. 1981. Predation by pellet-reared tiger muskellunge on minnows and bluegills in experimental systems. Transactions of the American Fisheries Society 1 10:197-209. Gleason, H. A. 1926. The individualistic concept of the plant association. American Midland NaturaUst 21 :92-l 10. Goldschmidt, T., F. Witte, and J. Wanink. 1993. Cascading effects of the introduced Nile perch on the detritivorous/phytoplanktivorous species in the sublittoral areas of Lake Victoria. Conservation Biology 7:686-700. Graham, R. W., plus 19 coauthors. 1996. Spatial response of mammals to late Quaternary environmental fluctuations. Science 272:1601-1606.

PAGE 146

138 Grant, J. W. A., and D. L. G. Noakes. 1987. Movers and stayers: foraging tactics of young-of-the-year brook charr, Salvelinus fontinalis Journal of Animal Ecology 56:1001-1014. Guest, W. C, B. W. Lyons, and G. Garza. 1980. Effects of temperature on survival of peacock bass fmgerlings. Proceedings of the Annual Conference Southeastern Association of Fish and Wildhfe Agencies. 33:620-627. Gutreuter, S. J., and R. O. Anderson. 1985. Importance of body size to the recruitment process in largemouth bass populations. Transactions of the American Fisheries Society 114:317-327. Hambright, K. D. 1991 Experimental analysis of prey selection by largemouth bass: role of predator mouth width and prey body depth. Transactions of the American Fisheries Society 120:500-508. Hambright, K. D., R. W. Drenner, S. R. McComas, and N. G. Hairston, Jr. 1991. Gape-limited piscivores, planktivore size refuges, and the trophic cascade hypothesis. Archiv fur Hydrobiologie 121:389-404. Hardin, G. 1960. The competitive exclusion principle. Science 131:1292-1297. Harris, P. M. 1995. Are autecologically similar species also functionally similar? A test in pond communities. Ecology 76:544-552. Hayse, J. W., and T. E. Wissing. 1996. Effects of stem density of artificial vegetation on abundance and growth of age-0 bluegills and predation by largemouth bass. Transactions of the American Fisheries Society 1 125:422-433. He, X., and J. F. Kitchell. 1990. Direct and indirect effects of predation on a fish community: a whole lake experiment. Transactions of the American Fisheries Society 119:825-835. Heidinger, R. C. 1975. Life history and biology of the largemouth bass. pp. 1 1-20. in R. H. Stroud and H. Clepper, editors. Black bass biology and management. Sport Fishing Insfitute, Washington D. C. Heidinger, R. C. 1999. Stocking for sport fisheries enhancement, pp. 375-401 in C. C. Kohler and W. A. Hubert, editors. Inland fisheries management in North America, 2"'' edition. American Fisheries Society, Bethesda, Maryland. Herbold, B., and P. B. Moyle. 1986. Introduced species and vacant niches. American Naturahst 128:751-760. Hidalgo, C. 1997. South Florida's peacock bass. Catfish Books, Inc., Tamarac, Florida.

PAGE 147

139 Hill, J. E. 1998. Estimate of gape limitation on forage size for the peacock cichlid, Cichla ocellaris an exotic fish established in Florida. Master's thesis. University of Florida, Gainesville, Florida. Hill, J. E. 2002. Exotic fishes in Florida. LakeLines, North American Lake Management Society. 22(l):39-43. Hirst, S. C, and D. R. DeVries. 1994. Assessing the potential for direct feeding interactions among larval black bass and larval shad in two southeastern reservoirs. Transactions of the American Fisheries Society 123:173-181. Hobson, E. S. 1979. Interactions between piscivorous fishes and their prey. pp. 231-242 in H. Clepper, editor. Predator-prey systems in fisheries management. Sport Fishing Institute, Washington, D. C. Hogg, R. G. 1976. Ecology of fishes of the family Cichlidae introduced into the fresh waters of Dade County, Florida. Doctoral dissertation. University of Miami, Coral Gables, Florida. Hollander, M., and D. A. Wolfe. 1973. Nonparametric statistical methods. John Wiley and Sons, New York. Hoyle, J. A., and A. Keast. 1987. The effect of prey morphology and size on handling time in a piscivore, the largemouth bass ( Micropterus salmoides ). Canadian Journal of Zoology 65:1972-1977. Huckins, C. J. F. 1997. Functional linkages among morphology, feeding performance, diet, and competitive ability in moUuscivorous sinfish. Ecology 78:2401-2414. Huckins, C. J. F., C. W. Osenberg, and G. G. Mittelbach. 2000. Species introductions and their ecological consequences: an example with congeneric sunfish. Ecological Applications 10:612-625. Hughes, N. F. 1986. Changes in the feeding biology of the Nile perch, Lates niloticus (Pisces: Centropomidae), in Lake Victoria, East Africa, since its introduction in 1 960 and its impact on the native fish community of the Nyanza Gulf Journal of Fish Biology 29:541-548. Hurlbert, S. H. 1978. The measurement of niche overlap and some relatives. Ecology 59:67-77. Huskey, S. H., and R. G. Tumigan. 2001. Variation in prey-resource utilization and oral jaw gape between two populations of largemouth bass, Micropterus salmoides Environmental Biology of Fishes 61:185-194.

PAGE 148

140 Hutchinson, G. E. 1957. Concluding remarks. Cold Spring Harbor Symposium on Quantitative Biology 22:415-427. Jacobson, L., and M. B. Perrow. 1998. Predation risk from piscivorous fish influencing the diel use of macrophytes by planktivorous fish in experimental ponds. Ecology of Freshwater Fish 7:78-86. Jepsen, D. B., K. O. Winemiller, and D. C. Taphom. 1997. Temporal patterns of resource partitioning among Cichla species in a Venezuelan blackwater river. Journal of Fish Biology 51:1085-1108. Jepsen, D. B., K. O. Winemiller, D. C. Taphom, and D. Rodriguez Olarte. 1999. Age structure and growth of peacock cichlids from rivers and reservoirs of Venezuela. Journal of Fish Biology 55:433-450. Kaufrnan, L. 1992. Catastrophic change in species-rich freshwater ecosystems: the lessons of Lake Victoria. Bioscience 42:846-858. Keast, A. 1978. Trophic and spatial interrelationships in the fish species of an Ontario temperate lake. Environmental Biology of Fishes 3:7-31. Keast, A. 1985. The piscivore feeding guild of fishes in small freshwater ecosystems. Environmental Biology of Fishes 12:1 19-129. Keast, A., and J. M. Eadie. 1985. Growth depensation in year-0 largemouth bass: the influence of diet. Transactions of the American Fisheries Society 1 14:204-213. Keast, A., J. Harper, and D. TumbuU. 1978. Nearshore fish habitat ufilization and species associations in Lake Opinicon Ontario Canada. Environmental Biology of Fishes 3:173-184. Kolar, C. S., and D. M. Lodge. 2002. Ecological predictions and risk assessment for alien fishes in North America. Science 298:1233-1236. Kullander, S. O. 1986. Cichlid fishes of the Amazon River drainage of Peru. Swedish Museum of Natural History, Stockholm, Sweden. Kullander, S. O., and H. Nijssen. 1989. The cichlids of Surinam (Teleostei: Labroidei). E. J. Brill, Leiden, The Netherlands. Kurzava, L. M., and P. J. Morin. 1998. Tests of ftinctional equivalence: complementary roles of salamanders and fish in community organization. Ecology 79:477-489. Lachner, E. A., C. R. Robins, and W. R. Courtenay, Jr. 1970. Exotic fishes and other aquatic organisms introduced into North America. Smithsonian Contributions to Zoology, Number 59. Smithsonian Insfitution Press, Washington, D.C.

PAGE 149

141 Lassuy, D. R. 1995. Introduced species as a factor in extinction and endangerment of native fish species. American Fisheries Society Symposium 15:391-396. Lawlor, L. R. 1979. Direct and indirect effects of n-species competition. Oecologia 43:355-364. Lawrence, J. M. 1958. Estimated sizes of various forage fishes largemouth bass can swallow. Proceedings of the Annual Conference Southeastern Association of Game and Fish Commissioners 1 1:220-225. Lee, D. S. 1980. Micropterus salmoides (Lacepede) largemouth bass. p. 608. in D. S. Lee, C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D. E. McAllister, and J. R. Stauffer, Jr., editors. Atlas of North American freshwater fishes. North Carolina State Museum of Natural History, Raleigh, North Carolina. Lever, C. 1996. Naturalized fishes of the world. Academic Press, San Diego, California. Levins, R. 1974. The qualitafive analysis of partially-specified systems. Annals of the New York Academy of Science 231:123-138. Lewis, W. M., G. E. Gunning, E. Lyles, and W. L. Bridges. 1961. Food choice of largemouth bass as a function of availability and vulnerability of food items. Transactions of the American Fisheries Society 90:277-280. Li, H. W., and P. B. Moyle. 1981. Ecological analysis of species introductions into aquatic systems. Transactions of the American Fisheries Society 1 10:772-782. Li, H. W., and P. B. Moyle. 1999. Management of introduced fishes, pp. 345-374 in C. C. Kohler and W. A. Hubert, editors. Inland fisheries management in North America, 2"'^ edition. American Fisheries Society, Bethesda, Maryland. Li, H. W., P. A. Rossingnol, and G. Castillo. 2000. Risk analysis of species introductions: insights fi-om qualitafive modeling, pp. 431-447 in R. Claudi and J. H. Leach, editors. Nonindigenous fi"eshwater organisms vectors, biology, and impacts. Lewis Publishers, Boca Raton, Florida. Lilyestrom, C. G., P. Quinones, and G. Oliveras. 1994. Peacock and largemouth bass competifion in La Plata Reservoir. Federal Aid in Sportfish Restorafion report F30. Puerto Rico Department of Natural and Environmental Resources, San Juan, Puerto Rico. Lilyestrom, C. G., and T. N. Churchill. 1996. Diet and movement of largemouth bass and butterfly peacocks in La Plata Reservoir, Puerto Rico. Proceedings of the Annual Conference Southeastern Association of Fish and Wildlife Agencies 50:192-200.

PAGE 150

142 Lodge, D. M. 1993. Biological invasions: lessons for ecology. Trends in Ecology and Evolution 8:133-137. Loftus, W. F. 2000. Inventory of fishes of Everglades National Park. Florida Scientist 63:27-47. Loftus, W. F., and J. A. Kushlan. 1987. Freshwater fishes of southern Florida. Bulletin of the Florida State Museum, Biological Sciences 31:147-344. Lowe-McConnell, R. H. 1969. The cichlid fishes of Guyana, South America, with notes on their ecology and breeding behaviour. Zoological Journal of the Linnaean Society 48:255-302. Ludsin, S. A., and D. R. DeVries. 1997. First-year recruitment of largemouth bass: the interdependency of early life stages. Ecological Applications 7:1024-1038. MacArthur, R., and R. Levins. 1967. The limiting similarity, convergence, and divergence of coexisting species. American Naturalist 101:377-385. MacArthur, R. H., and E. R. Pianka. 1966. On the optimal use of a patchy environment. American Naturalist 100:603-609. MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography. Princeton University Press, Princeton, New Jersey. Machado-AUison, A. (R. L. Chew, translator). 1971. Contribution al conocimiento de la taxonomia del genero Cichla (Perciformes: Cichlidae) en Venezuela parte 1. Acta Biologica Venezuela 7:459-497. Maciolek, J. A. 1984. Exotic fishes in Hawaii and other islands of Oceania, pp. 131-161 in W. R. Courtenay, Jr. and J. R. Stauffer, Jr., editors. Distribution, biology, and management of exotic fishes. Johns Hopkins University Press, Baltimore, Maryland. Magnhagen, C. 1986. Activity differences influencing food selection in the marine fish Pomatoschistus microps Canadian Journal of Fisheries and Aquatic Sciences 43:223-227. Magnuson, J. J. 1976. Managing with exotics a game of chance? Transactions of the American Fisheries Society 105:1-9. Manly, B. F. J. 1974. A model for certain types of selection experiments. Biometrics 30:281-294.

PAGE 151

143 Matthews, W. J., F. P. Gelwick, and J. J. Hoover. 1992. Food and habitat use by juveniles of Micropterus and Morone in a southwestern reservoir. Transactions of the American Fisheries Society 121 :54-66. Mathur, D. 1977. Food habits and competitive relationships of the bandfin shiner in Halawakee Creek, Alabama. The American Midland Naturalist 97:89-100. McClelland, M. 1992. Invasion of the biosnatchers. Florida Environments 6(2):5, 13. McComas, S. R., and R. W. Drenner. 1982. Species replacement in a reservoir fish community: silverside feeding mechanics and competition. Canadian Journal of Fisheries and Aquatic Sciences 39:815-821. Mcintosh, R. P. 1995. H. A. Gleason's 'individualistic concept' and theory of animal communities: a continuing controversy. Biological Reviews 70:317-357. McKaye, K. R., and A. C. Marsh. 1983. Food switching by two specialized algaescraping cichlid fishes in Lake Malawi, Africa. Oecologia 56:245-248. Mesing, C. L., and A. M. Wicker. 1986. Home range, spawning migrations, and homing of radio-tagged Florida largemouth bass Micropterus salmoides in two central Florida USA lakes. Transactions of the American Fisheries Society 1 15:286-295. Miller, R. R., J. D. Williams, and J. E. Williams. 1989. Extinctions of North American fishes during the past century. Fisheries 14(6):22-38. Miller, T. E. 1994. Direct and indirect species interactions in an eariy old-field plant community. American Naturalist 143:1007-1025. Mills, E. L., J. H. Leach, J. T. Cariton, and C. L. Secor. 1994. Exotic species and the integrity of the Great Lakes: lessons from the past. Bioscience 44:666-676. Minns, C. K., and J. M. Cooley. 2000. Intentional introductions: are the incalculable risks worth it? pp. 57-59 in R. Claudi and J. H. Leach, editors. Nonindigenous freshwater organisms vectors, biology, and impacts. Lewis Publishers, Boca Raton, Florida. Miranda, L. E., and R. J. Muncy. 1989. Bioenergetic values of shads and sunfishes as prey for largemouth bass. Proceedings of the Annual Conference Southeastern Association of Fish and Wildlife Agencies 43:153-163. Mittelbach, G. G. 1984. Predation and resource partitioning in two sunfishes (Centrarchidae). Ecology 65:499-513. ach, G. G. 1988. Competition among reftiging sunfishes and effects offish density on littoral zone invertebrates. Ecology 69:614-623.

PAGE 152

144 Mittelbach, G. G., and L. Persson. 1998. The ontogeny of piscivory and its ecological consequences. Canadian Journal of Fisheries and Aquatic Sciences 55:1454-1465. Mittelbach, G. G., A. M. Turner, D. J. Hall, J. E. Rettig, and C. W. Osenberg. 1995. Perturbation and resilience: a long-term, whole-lake study of predator extinction and reintroduction. Ecology 76:2347-2360. Moermond, T. C. 1979. Resource partitioning: a dynamic competitive balance, pp. 303309 in H. Clepper, editor. Predator-prey systems in fisheries management. Sport Fishing Institute, Washington, D. C. Morin, P. J. 1995. Functional redundancy, non-additive interactions, and supply-side dynamics in experimental pond communities. Ecology 76:133-149. Moyle, P. B., H. W. Li, and B. A. Barton. 1986. The Frankenstein Effect: impact of introduced fishes on native fishes in North America, pp. 415-426 in R. H. Stroud, editor. Fish culture in fisheries management. American Fisheries Society, Bethesda, Maryland. Moyle, P. B., and T. Light. 1996a. Biological invasions of fresh water: empirical rules and assembly theory. Biological Conservation 78:149-161. Moyle, P. B., and T. Light. 1996b. Fish invasions in California: do abiotic factors determine success? Ecology 77:1666-1670. Murie, D. J. 1995. Comparative feeding ecology of two sympatric rockfish congeners, Sebastes caurinus (copper rockfish) and S. maliger (quillback rockfish). Marine Biology 124:341-353. Nico, L. G., and P. L. Fuller. 1999. Spatial and temporal patterns of nonindigenous fish introductions in the United States. Fisheries 24(1): 16-27. Nico, L. G., and D. C. Taphom. 1988. Food habits of piranhas in the low llanos of Venezuela. Biotropica 20:3 1 1 -32 1 Nico, L. G., J. D. Williams, and J. J. Herod. 2001 Black carp ( Mylopharvneodon piceus): a biological synopsis and updated risk assessment. Report to the Risk Assessment and Management Committee of the Aquatic Nuisance Species Task Force. USGS, Gainesville, Florida. Nilsson, P. A., and C. Bronmark. 2000. Prey vulnerability to a gape-size limited predator: behavioural and morphological impacts on northern pike piscivory. Oikos 88:539546.

PAGE 153

145 Noble, R. L., and R. D. Germany. 1986. Changes in fish populations of Trinidad Lake, Texas, in response to abundance of blue tilapia. pp. 455-461 in R. H. Stroud, editor. Fish culture in fisheries management. American Fisheries Society, Bethesda, Maryland. Noga, E. J., S. Botts, M. -S. Yang, and R. Avtalion. 1998. Acute stress causes skin ulceration in striped bass and hybrid bass ( Morone ). Veterinary Pathology 35:102-107. Northcote, T. G. 1988. Fish in the structure and ftinction of fi-eshwater ecosystems: a "top-down" view. Canadian Journal of Fisheries and Aquatic Sciences 45-361379. Norton, S. F., and E. L. Brainerd. 1993. Convergence in the feeding mechanics of ecomorphologically similar species in the Centrarchidae and Cichlidae. Journal of Experimental Biology 176:1 1-29. O'Brien, W. J. 1979. The predator-prey interaction of planktivorous fish and zooplankton. American Scientist 67:572-581. Ogle, D. H., J. H. Selgeby, R. M. Newman, and M. G. Henry. 1995. Diet and feeding periodicity of ruffe in the St. Louis River Estuary, Lake Superior. Transactions of the American Fisheries Society 124:356-369. Ogilvie, V. E. 1966. Preliminary report— peacock bass, revised. Florida Game and Fresh Water Fish Commission, West Palm Beach, Florida. Olson, M. H. 1996. Ontogenefic niche shifts in largemouth bass: variability and consequences for first-year growth. Ecology 77:179-190. Olson, M. H., G. G. Mittelbach, and C. W. Osenberg. 1995. Competifion between predator and prey: resource-based mechanisms and implications for stagestructured dynamics. Ecology 76:1758-1771. Osenberg, C. W., and G. G. Mittelbach. 1989. The effects of body size on the predatorprey interaction between pumpkinseed sunfish and gastropods. Ecological Monographs 59:405-432. Osenberg, C. W., M. H. Olson, and G. G. Mittelbach. 1994. Stage structure in fishes: resource producfivity and competition gradients, pp. 151-170 in D. J. Stouder, K. C. Fresh, and R. J. Felder, editors. Theory and application in fish feeding ecology University of South Carolina Press, Columbia, South Carolina. OTA (Office of Technology Assessment), U. S. Congress. 1993. Harmftil non-indigenous species in the United States, OTA-F-565. U. S. Government Printing Office Washington, D.C.

PAGE 154

146 Ott, L. 1993. An introduction to statistical methods and data analysis, 4"' edition. Wadsworth Publishing Company, Belmont, California. Paine, R. T. 1992. Food-web analysis through field measurement of per capita interaction strength. Nature 355:73-75. Parrish, D. L., and F. J. Margraf 1990. Interactions between white perch ( Morone americana) and yellow perch (Perca flavescens) in Lake Erie as determined from feeding and growth. Canadian Journal of Fisheries and Aquatic Sciences 4717791787. Pasch, R. W. 1975. Some relationships between food habits and growth of largemouth bass in Lake Blackshear, Georgia. Proceedings of the Annual Conference Southeastern Association of Fish and Wildlife Agencies 28:307-321. Pennak, R. W. 1989. Fresh-water invertebrates of the United States: Protozoa to Mollusca, 3"' edition. John Wiley & Sons, Inc., New York. Persson, L. 1993. Predator-mediated competition in prey refuges: the importance of habitat dependent prey resources. Oikos 68:12-22. Persson, L., J. Andersson, E. Wahlstrom, and P. Eklov. 1996. Size-specific interactions in lake systems: predator gape limitation and prey growth rate and mortality. Ecology 77:900-911. Pope, K. L., M. L. Brown, W. G. Duffy, and P. H. Michaletz. 2001. A caloric-based evaluafion of diet indices for largemouth bass. Environmental Biology of Fishes 61:329-339. Popiel, S. A., A. Perez-Fuentetaja, D. J. McQueen, and N. C. Collins. 1996. Determinants of nesfing success in the pumpkinseed ( Lepomis gibbosus ): a comparison of two populafions under different risks from predation. Copeia 1996:649-656. Rabelo, H., and C. A. R. M. Araujo-Lima. 2002. A dieta e o consumo diario de alimento de Cichla monoculus na Amazonia Central. Acta Amazonica 32:707-724. Resetarits, W. J., Jr. 1995. Limifing similarity and the intensity of competifive effects on the mottled sculpin, Cottus bairdi in experimental stream communities Oecologia 104:31-38. Reynolds, J. B. 1996. Electrofishing. pp. 221-253 in B. R. Murphy and D. W. WilHs, editors. Fisheries techniques, 2"'' edition. American Fisheries Society, Bethes'da, Maryland.

PAGE 155

147 Ribbink, A. J. 1991 Distribution and ecology of the cichlids of the African Great Lakes, pp. 36-59 in M. H. A. Keenleyside, editor. Cichlid fishes: behaviour, ecology, and evolution. Chapman and Hall, London, England. Ribbink, A. J., B. A. Marsh, A. C. Marsh, A. C. Ribbink, and B. J. Sharp. 1983. A preliminary survey of the cichlid fishes of rocky habitats in Lake Malawi. South African Journal of Zoology 18:149-310. Risk Assessment and Management (RAM) Committee. 1998. Generic nonindigenous aquatic organisms risk analysis review process (for estimating risk associated with the introduction of nonindigenous aquatic organisms and how to manage that risk). Report to the Aquatic Nuisance Species Task Force. U. S. Government Printing Office: 1998-693-132/62087 Region No. 10. Rosenzweig, M. L. 1981. A theory of habitat selection. Ecology 62:327-335. Rosenzweig, M. L. 2001. The four questions: what does the introduction of exotic species do to diversity? Evolutionary Ecology Research 3:361-367. Ross, S. T. 1991. Mechanisms structuring stream fish assemblages: are there lessons from introduced species? Environmental Biology of Fishes 30:359-368. Sakai, A. K., F. W. Allendorf, J. S. Holt, D. M. Lodge, J. Molofsky, K. A. With, S. Baughman, R. J. Cabin, J. E. Cohen, N. C. Ellstrand, D. E. McCauley, P. O'Neil, I. M. Parker, J. N. Thompson, and S. G. Weller. 2001. The population biology of invasive species. Annual Review of Ecology and Systematics 32:305-332. Sale, P. F. 1979. Habitat partitioning and competition in fish communities, pp. 323-331 in H. Clepper, editor. Predator-prey systems in fisheries management. Sport Fishing Institute, Washington, D. C. Savino, J. F., and R. A. Stein. 1982. Predator-prey interacfion between largemouth bass and bluegills as influenced by simulated, submersed vegetation. Transactions of the American Fisheries Society 1 1 1 :255-266. Schoener, T. W. 1970. Nonsynchronous spatial overlap of lizards in patchy habitats. Ecology 51:408-418: Schoener, T. W. 1982. The controversy over interspecific compedfion. American Scienfist 70:586-595. Schoener, T. W. 1987. Axes of controversy in community ecology, pp. 8-16 in W. J. Matthews and D. C. Heins, editors. Community an evolufionary ecology of North American stream fishes. University of Oklahoma Press, Norman, Oklahoma.

PAGE 156

148 Schramm, H. A., Jr., and M. J. Maceina. 1986. Distribution and diet of Suwannee bass and largemouth bass in the lower Santa Fe River, Florida. Environmental Biology of Fishes 15:221-228. Schreck, C. B., B. L. Olla, and M. W. Davis. 1997. Behavioral responses to stress, pp. 145-170 in G. K. Iwama, A. D. Pickering, J. P. Sumpter, and C. B. Schreck, editors. Fish stress and health in aquaculture. Cambridge University Press, Cambridge, UK. Schroder, S. L., and T. M. Zaret. 1979. The adaptive significance of color patterns in Cichla ocellaris Copeia 1979:43-47. Scoppettone, G. G. 1993. Interactions between native and normative fishes of the Upper Muddy River, Nevada. Transactions of the American Fisheries Society 122-599608. Scott, M. C, and P. L. Angermeier. 1998. Resource use by two sympatric black basses in impounded and riverine sections of the New River, Virginia. North American Journal of Fisheries Management 18:221-235. Shafland, P. L. 1995. Introduction and establishment of a successful butterfly peacock fishery in southeast Florida canals. American Fisheries Society Symposium 15:443-451. Shafland, P. L. 1996. Exotic fish assessments: an alternative view. Reviews in Fishery Science 4:123-132. Shafland, P. L. 1999a. The introduced butterfly peacock ( Cichla ocellaris ) in Florida. I. Fish community analyses. Reviews in Fishery Science 7:71-94. Shafland, P. L. 1999b. The introduced butterfly peacock ( Cichla ocellaris ) in Florida. 11. Food and reproductive biology. Reviews in Fishery Science 7:95-1 13. Shafland, P. L. 1 999c. The introduced butterfly peacock ( Cichla ocellaris ) in Florida. III. Length distribution analyses. Reviews in Fishery Science 7:1 15-126. Shafland, P. L., K. B. Gestring, and M. S. Stanford. 2001. The fisheries of metropolitan southeast Florida canals with special reference to exotic fishes. Third Annual Progress Report, Southeast Florida Canal Sportfishery Studies 1997-2000. Florida Fish and Wildlife Conservation Commission, Boca Raton, Florida. Shafland, P. L., and B. D. Hilton. 1986. Introduction and evaluation of peacock bass (Cichla spp.) in southeast Florida canals. Project No. 6170. Florida Game and Fresh Water Fish Commission, Tallahassee, Florida.

PAGE 157

149 Shafland, P. L., B. D. Hilton, and R. J. Metzger. 1985. Fishes of Black Creek Canal: completion report for study XII (1981-1 985). Florida Game and Fresh Water Fish Commission, Boca Raton, Florida. Shafland, P. L., and J. M. Pestrak. 1982. Lower lethal temperatures for fourteen nonnative fishes in Florida. Environmental Biology of Fishes 7:149-156. Shafland, P. L., and J. M. Pestrak. 1983. Suppression of largemouth bass production by blue tilapia in ponds. Proceedings of the Annual Conference Southeastern Association of Fish and Wildlife Agencies 37:441-446. Shafland, P. L., and M. S. Stanford. 1999. The introduced butterfly peacock ( Cichla ocellaris ) in Florida. IV. Socioeconomic analyses. Reviews in Fishery Science 7:127-135. Shireman, J. V., and C. R. Smith. 1983. Synopsis of the biological data on the grass carp Ctenopharvngodon idella (Cuvier and Valenciennes, 1 844). FAO Fisheries Synopsis No. 135. Rome, Italy. Shrader-Frechette, K. 2001. Non-indigenous species and ecological explanation. Biology and Philosophy 16:507-519. Sih, A., G. Englund, and D. Wooster. 1998. Emergent impacts of multiple predators on prey. Trends in Ecology and Evolution 13:350-355. Soluk, D. A. 1993. Multiple predator effects: predicting combined functional response of stream fish and invertebrate predators. Ecology 74:219-225. Soluk, D. A., and N. C. Collins. 1988. Synergistic interacfions between fish and invertebrate predators: facilitation and interference among stream predators Oikos 52:94-100. Stauffer, J. R., Jr. 1984. Colonization theory relative to introduced populations, pp. 8-21 in W. R. Courtenay, Jr. and J. R. Stauffer, Jr., editors. Distribution, biology, and management of exotic fishes. Johns Hopkins University Press, Baltimore, Maryland. Stoffels, R. J., and P. Humphries. 2003. Ontogenetic variation in the diurnal food and habitat associations of an endemic and an exotic fish in floodplain ponds: consequences for niche partifioning. Environmental Biology of Fishes 66-293305. Stone, L., and A. Roberts. 1991 Conditions for a species to gain advantage from the presence of competitors. Ecology 72:1964-1972.

PAGE 158

150 Sutton, T. M., and J. J. Ney. 2002. Trophic resource overlap between age-0 striped bass and largemouth bass in Smith Mountain Lake, Virginia. North American Journal of Fisheries Management 22:1250-1259. Swingle, H. S. 1950. Relationships and dynamics of balanced and unbalanced fish populations. Alabama Agricultural Experiment Station Bulletin No. 274. Auburn, Alabama. Swingle, H. S. 1 967. Temperature tolerance of the peacock bass and a pond test of its value as a piscivorous species. Proceedings of the Annual Conference Southeastern Association of Game and Fish Commissioners 20:297-299. Tate, W. B., M. S. Allen, R. A. Myers, and J. R. Estes. 2003. Comparison of electrofishing and rotenone for sampling largemouth bass in vegetated areas of two Florida lakes. North American Journal of Fisheries Management 23:181-188. Taylor, J. N., W. R. Courtenay, Jr., and J. A. McCann. 1984. Known impacts of exotic fishes in the continental United States, pp. 322-373 in W. R. Courtenay, Jr. and J. R. Stauffer, Jr., editors. Distribution, biology, and management of exotic fishes. Johns Hopkins University Press, Balfimore, Maryland. Timmons, T. J., W. L. Shelton, and W. D. Davies. 1980. Differenfial growth of largemouth bass in West Point Reservoir, Alabama-Georgia. Transactions of the American Fisheries Society 109:176-186. Trexler, J. C, W. F. Loftus, F. Jordan, J. J. Lorenz, J. H. Chick, and R. M. Kobza. 2000. Empirical assessment of fish introductions in a subtropical wetland: an evaluation of contrasting views. Biological Invasions 2:265-277. Turner, A. M., S. A. Fetterolf, and R. J. Bemot. 1999. Predator identity and consumer behavior: differential effects of fish and crayfish on the habitat use of a freshwater snail. Oecologia 118:242-247. VanArman, J., D. Nealon, S. Bums, B. Jones, L. Smith, T. MacVicar, M. Yansura, A. Frederico, J. Bucca, M. Knapp, and P. Gleason. 1984. South Florida Water Management District, pp. 138-157 in E. A. Femald and D. J. Patton, editors. Water resources atlas of Florida. Florida State University, Tallahassee, Florida. Vander Zanden, M. J., J. M. Casselman, and J. B. Rasmussen. 1999. Stable isotope evidence for the food web consequences of species invasions in lakes. Nature 410:464-467. Vermeij, G. J. 1991. When biotas meet: understanding biotic interchange Science 253:1099-1104. Vermeij, G. J. 1996. An agenda for invasion biology. Biological Conservation 78:3-9.

PAGE 159

151 Vonesh, J. R., and O. De la Cruz. 2002. Complex life cycles and density dependence: assessing the contribution of egg mortality to amphibian declines. Oecologia 133:325-333. Wahl, D. H., and R. A. Stein. 1988. Selective predation by three esocids: the role of prey behavior and morphology. Transactions of the American Fisheries Society 117:142-151. Wainwright, P. C. 1988. Morphology and ecology: functional basis of feeding constraints in Caribbean labrid fishes. Ecology 69:635-645. Wainwright, P. C. 1 996. Ecological explanation through functional morphology: the feeding biology of sunfishes. Ecology 77:1336-1343. Wainwright, P. C, and B. A. Richard. 1995. Predicting patterns of prey use from morphology of fishes. Environmental Biology of Fishes 44:97-1 13. Wallace, R. K., Jr. 1981. An assessment of diet-overlap indexes. Transactions of the American Fisheries Society 1 10:72-76. Webb, P. W. 1986. Effect of body form and response threshold on the vulnerability of four species of teleost prey attacked by largemouth bass ( Micropterus sahnoides). Canadian Journal of Fisheries and Aquatic Sciences 43:763-771. Wedemeyer, G. A. 1997. Effects of rearing conditions on the health and physiological quality of fish in intensive culture, pp. 35-71 in G. K. Iwama, A. D. Pickering, J. P. Sumpter, and C. B. Schreck, editors. Fish stress and health in aquaculture. Cambridge University Press, Cambridge, UK. Welcomme, R. L. 1981 Register of international transfers of inland fish species. FAO Fisheries Technical Paper No. 213., Rome, Italy. Welcomme, R. L. 1988. International introductions of inland aquatic species. FAO Fisheries Technical Paper No. 294. Rome, Italy. Werner, E. E. 1974. The fish size, prey size, handling time relation in several sunfishes and some implications. Journal of the Fisheries Research Board of Canada 3 1 • 1531-1536. Werner, E. E. 1 977. Species packing and niche complementarity in three sunfishes. American Naturalist 1 1 1 :553-578. Werner, E. E. 1979. Niche partitioning by food size in fish communities, pp. 3 1 1-322 in H. Clepper, editor. Predator-prey systems in fisheries management. Sport Fishing Institute, Washington, D.C.

PAGE 160

152 Werner, E. E., and J. F. Gilliam. 1984. The ontogenetic niche and species interactions in size-structured populations. Annual Review of Ecology and Systematics 15:393426. Werner, E. E., J. F. Gilliam, D. J. Hall, and G. G. Mittelbach. 1983. An experimental test of the effects of predation risk on habitat use in fish. Ecology 64:1540-1548. Wheeler, A. P., and M. S. Allen. 2003. Habitat and diet partitioning between shoal bass and largemouth bass in the Chipola River, Florida. Transactions of the American Fisheries Society 132:438-449. Williamson, M. 1996. Biological invasions. Chapman and Hall, London, England. Winemiller, K. O. 1989. Ontogenetic diet shifts and resource partitioning among piscivorous fishes in the Venezuelan ilanos. Environmental Biology of Fishes 26:177-199. Winemiller, K. O. 1991. Comparative ecology of Serranochromis species (Teleostei: Cichlidae) in the upper Zambezi River floodplain. Journal of Fish Biology 39:617-639. Winemiller, K. O., D. C. Taphom, and A. Barbarino-Duque. 1997. Ecology of Cichla (Cichlidae) in two blackwater rivers of southern Venezuela. Copeia 1997-690696. Winemiller, K. O., and D. H. Taylor. 1987. Predatory behavior and competition among laboratory-housed largemouth and smallmouth bass. American Midland Naturalist 117:148-166. Yodzis, P., and K. O. Winemiller. 1999. In search of operational trophospecies in a tropical aquatic food web. Oikos 87:327-340. Zale, A. v., and R. W. Gregory. 1990. Food selection by early life history stages of blue tilapia, Oreochromis aureus in Lake George, Florida: overlap with sympatric shad larvae. Florida Scientist 53:123-129. Zaret, T. M. 1980a. Life history and growth relationships of Cichla ocellaris a predatory South American cichlid. Biotropica 12:144-157. Zaret, T. M. 1980b. Predation and freshwater communities. Yale University Press, New Haven, Connecticut. Zaret, T. M., and R. T. Paine. 1973. Species introduction in a tropical lake. Science 182:449-455.

PAGE 161

153 Zaret, T. M., and A. S. Rand. 1971. Competition in tropical stream fishes: support for the competitive exclusion principle. Ecology 52:336-342. Zom, T. G., and P. W. Seelbach. 1995. The relation between habitat availability and the short-term carrying capacity of a stream reach for smallmouth bass. North American Journal of Fisheries Management 15:773-783.

PAGE 162

BIOGRAPHICAL SKETCH Jeffrey E. Hill was bom in 1 966 in Florence, Alabama, to Baker E. and Jacqueline C. Hill. He became interested in science and the outdoors at an early age — his father took him hunting, fishing, and catching minnows in the creek; his mother and siblings read dinosaur books to him; Myrtle Alexander, a local biology teacher, taught him about reptiles and mollusks. By high school, Jeffrey knew he wanted to pursue a career in ichthyology or fish ecology. He maintains a lifelong love of aquarium keeping, fishing, hunting, and the wonder of the outdoors. Jeffrey graduated from Bradshaw High School, Florence, Alabama, listed in the "Mark of Excellence" (top 5% of graduating class) and delivered the Farewell Speech for the Class of 1985. He attended the University of Florida from 1985 to 1988 in the Undergraduate Honors Program as a National Merit Scholar. Jeffrey married Susan Bemecker in 1989 and completed his Bachelor of Science degree at the University of North Alabama in 1 991 with a major in biology and a minor in geography. He has worked in Florida in the private sector in mariculture research and owned and operated a tropical fish farm. Jeffrey earned a Master of Science degree in 1 998 in the Department of Fisheries and Aquatic Sciences, University of Florida. His thesis was titled "Estimate of gape limitation on forage size for the peacock cichlid, Cichla ocellaris an exotic fish established in Florida." He entered the doctoral program in the Department of Fisheries and Aquafic Sciences in 1998 as a Dean's Fellow. Jeffrey's graduate awards include the 154

PAGE 163

155 Roger Rottman Memorial Scholarship from the Florida Chapter of the American Fisheries Society and the Department of Fisheries and Aquatic Sciences Graduate Student of the Year Award. After graduation, Jeffrey will serve as associate research faculty at the Tropical Aquaculture Laboratory of the University of Florida in Ruskin, Florida. He plans to continue his career in research, teaching, and extension.

PAGE 164

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. Charles E. Cichra, Chair Professor of Fisheries and Aquatic 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. William J. Lindb^^ Associate Professor of Fisheries anc Aquatic Sciences I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is flilly adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Carter R. Gilbert / Professor Emeritus 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. Leo ( \. Nico ^ ^Research Biologist, U.S. Geological Survey

PAGE 165

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. Craig W. Osenberg Professor of Zoology This dissertation was submitted to the Graduate Faculty of the College of Agricultural and Life Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December 2003 Dean, Graduate School


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 EHKFAX4ZZ_01E24W INGEST_TIME 2015-04-22T21:11:26Z PACKAGE AA00030056_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES