Citation
Habitat- and Density-Mediated Influences on Snook Ecology

Material Information

Title:
Habitat- and Density-Mediated Influences on Snook Ecology Lessons Learned from Manipulative Release Experiments with Hatchery-Reared Juvenile Snook
Creator:
Brennan, Nathan
Place of Publication:
[Gainesville, Fla.]
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (183 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Fisheries and Aquatic Sciences
Committee Chair:
Lindberg, William J.
Committee Co-Chair:
Leber, Kenneth Miles
Committee Members:
Frazer, Tom K.
Allen, Micheal S.
Osenberg, Craig W.
Graduation Date:
8/9/2008

Subjects

Subjects / Keywords:
Animal cannibalism ( jstor )
Creeks ( jstor )
Fish ( jstor )
Fisheries ( jstor )
Human cannibalism ( jstor )
Juveniles ( jstor )
Mortality ( jstor )
Predation ( jstor )
Predators ( jstor )
Shorelines ( jstor )
Fisheries and Aquatic Sciences -- Dissertations, Academic -- UF
cannibalism, centropomus, creeks, cwt, density, depletion, ecology, elastomer, enhancement, estuary, growth, habitat, juvenile, predation, snook, stocking, tags
City of Sarasota ( local )
Genre:
Electronic Thesis or Dissertation
born-digital ( sobekcm )
Fisheries and Aquatic Sciences thesis, Ph.D.

Notes

Abstract:
Common snook Centropomus undecimalis stocks in Florida, USA are experiencing increasing pressure from the recreational fishing industry and habitat loss. This dissertation investigates components of snook stock enhancement research, as a potential supplementary management tool. As juvenile stages are typically used for stocking, I examined influences of habitat and ambient snook density on post-release fates, and the potential for cannibalism in juvenile stages. Chapter 2 examines post-release growth, survival, and dispersal of juveniles stocked into rearing habitats, and their numerical and biomass contributions to adult stocks. Growth was inversely related to recapture ratios and release sites producing low growth had the highest numerical contributions to adult stocks, but highest biomass contributions were from sites producing intermediate growth rates. In Chapter 3, manipulations of release magnitude of hatchery-reared juveniles showed tidal creeks with high augmentation levels (n=2) resulted in overall post-release abundance increases (10 months post-release) but imperceptible changes in low augmentation creeks (n=2) of wild and hatchery conspecifics. Disproportionate post-release loss (85%) of hatchery-reared snook occurred in one high augmentation creek but was not detected in the other. No loss of wild conspecifics was detected in any of the creeks. Snook are cannibalistic and show evidence for this behavior intensifying during early stages. In Chapter 4, I examined ontogenetic differences in cannibalism potential and its role in controlling abundance. Morphological comparisons (anal spines, mouth gapes, and confusion colorations [eyespots]) showed that predatory defense and compensatory feeding adaptations decreased with size. Examination of existing research on snook diet coupled with model simulations showed that inter- and intra-cohort cannibalism intensified in nursery habitats and during winter when older age groups were present. Overall this work used released snook to quantify tradeoffs in production to adult stocks based on characteristics of snook rearing habitats. Within an age-1 cohort, stocking increased overall abundance but sometimes with high post-release mortality in hatchery snook. Intra- and inter-cohort cannibalism was focused on age-0 snook in rearing habitats, and could be influential on inter-annual recruitment dynamics. These findings provide valuable input for stakeholders evaluating snook stock enhancement potential in Florida. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2008.
Local:
Adviser: Lindberg, William J.
Local:
Co-adviser: Leber, Kenneth Miles.
Statement of Responsibility:
by Nathan Brennan.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Brennan, Nathan. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
664028250 ( OCLC )
Classification:
LD1780 2008 ( lcc )

Downloads

This item has the following downloads:

brennan_n.pdf

brennan_n_Page_095.txt

brennan_n_Page_068.txt

brennan_n_Page_136.txt

brennan_n_Page_090.txt

brennan_n_Page_097.txt

brennan_n_Page_116.txt

brennan_n_Page_078.txt

brennan_n_Page_065.txt

brennan_n_Page_101.txt

brennan_n_Page_160.txt

brennan_n_Page_151.txt

brennan_n_Page_133.txt

brennan_n_Page_042.txt

brennan_n_Page_176.txt

brennan_n_Page_022.txt

brennan_n_Page_114.txt

brennan_n_Page_031.txt

brennan_n_Page_128.txt

brennan_n_Page_102.txt

brennan_n_Page_010.txt

brennan_n_Page_079.txt

brennan_n_Page_103.txt

brennan_n_Page_108.txt

brennan_n_Page_107.txt

brennan_n_Page_021.txt

brennan_n_Page_106.txt

brennan_n_Page_044.txt

brennan_n_Page_146.txt

brennan_n_Page_104.txt

brennan_n_Page_170.txt

brennan_n_Page_020.txt

brennan_n_Page_086.txt

brennan_n_Page_041.txt

brennan_n_Page_126.txt

brennan_n_Page_094.txt

brennan_n_Page_008.txt

brennan_n_Page_132.txt

brennan_n_Page_169.txt

brennan_n_Page_073.txt

brennan_n_Page_121.txt

brennan_n_Page_040.txt

brennan_n_Page_092.txt

brennan_n_Page_053.txt

brennan_n_Page_109.txt

brennan_n_Page_048.txt

brennan_n_Page_130.txt

brennan_n_Page_043.txt

brennan_n_Page_181.txt

brennan_n_Page_017.txt

brennan_n_Page_105.txt

brennan_n_Page_159.txt

brennan_n_Page_064.txt

brennan_n_Page_061.txt

brennan_n_Page_120.txt

brennan_n_Page_183.txt

brennan_n_Page_100.txt

brennan_n_Page_029.txt

brennan_n_Page_019.txt

brennan_n_Page_150.txt

brennan_n_Page_123.txt

brennan_n_Page_067.txt

brennan_n_Page_117.txt

brennan_n_Page_134.txt

brennan_n_Page_180.txt

brennan_n_Page_003.txt

brennan_n_Page_069.txt

brennan_n_Page_071.txt

brennan_n_Page_015.txt

brennan_n_Page_156.txt

brennan_n_Page_096.txt

brennan_n_Page_127.txt

brennan_n_Page_142.txt

brennan_n_Page_033.txt

brennan_n_Page_018.txt

brennan_n_Page_161.txt

brennan_n_Page_056.txt

brennan_n_Page_030.txt

brennan_n_Page_088.txt

brennan_n_Page_025.txt

brennan_n_Page_045.txt

brennan_n_Page_129.txt

brennan_n_Page_163.txt

brennan_n_Page_125.txt

brennan_n_Page_077.txt

brennan_n_Page_084.txt

brennan_n_Page_063.txt

brennan_n_Page_162.txt

brennan_n_Page_154.txt

brennan_n_Page_035.txt

brennan_n_Page_050.txt

brennan_n_Page_011.txt

brennan_n_Page_070.txt

brennan_n_Page_085.txt

brennan_n_Page_004.txt

brennan_n_Page_099.txt

brennan_n_Page_062.txt

brennan_n_Page_034.txt

brennan_n_Page_145.txt

brennan_n_Page_049.txt

brennan_n_Page_005.txt

brennan_n_Page_039.txt

brennan_n_pdf.txt

brennan_n_Page_138.txt

brennan_n_Page_081.txt

brennan_n_Page_037.txt

brennan_n_Page_110.txt

brennan_n_Page_157.txt

brennan_n_Page_164.txt

brennan_n_Page_174.txt

brennan_n_Page_055.txt

brennan_n_Page_007.txt

brennan_n_Page_038.txt

brennan_n_Page_139.txt

brennan_n_Page_023.txt

brennan_n_Page_028.txt

brennan_n_Page_177.txt

brennan_n_Page_027.txt

brennan_n_Page_131.txt

brennan_n_Page_006.txt

brennan_n_Page_076.txt

brennan_n_Page_002.txt

brennan_n_Page_141.txt

brennan_n_Page_091.txt

brennan_n_Page_026.txt

brennan_n_Page_111.txt

brennan_n_Page_137.txt

brennan_n_Page_115.txt

brennan_n_Page_175.txt

brennan_n_Page_155.txt

brennan_n_Page_122.txt

brennan_n_Page_158.txt

brennan_n_Page_014.txt

brennan_n_Page_178.txt

brennan_n_Page_075.txt

brennan_n_Page_032.txt

brennan_n_Page_119.txt

brennan_n_Page_143.txt

brennan_n_Page_153.txt

brennan_n_Page_172.txt

brennan_n_Page_059.txt

brennan_n_Page_098.txt

brennan_n_Page_167.txt

brennan_n_Page_171.txt

brennan_n_Page_087.txt

brennan_n_Page_124.txt

brennan_n_Page_052.txt

brennan_n_Page_182.txt

brennan_n_Page_066.txt

brennan_n_Page_058.txt

brennan_n_Page_060.txt

brennan_n_Page_024.txt

brennan_n_Page_009.txt

brennan_n_Page_051.txt

brennan_n_Page_080.txt

brennan_n_Page_093.txt

brennan_n_Page_173.txt

brennan_n_Page_047.txt

brennan_n_Page_001.txt

brennan_n_Page_118.txt

brennan_n_Page_082.txt

brennan_n_Page_089.txt

brennan_n_Page_057.txt

brennan_n_Page_135.txt

brennan_n_Page_140.txt

brennan_n_Page_054.txt

brennan_n_Page_144.txt

brennan_n_Page_113.txt

brennan_n_Page_016.txt

brennan_n_Page_012.txt

brennan_n_Page_179.txt

brennan_n_Page_149.txt

brennan_n_Page_165.txt

brennan_n_Page_013.txt

brennan_n_Page_168.txt

brennan_n_Page_046.txt

brennan_n_Page_148.txt

brennan_n_Page_072.txt

brennan_n_Page_083.txt

brennan_n_Page_112.txt

brennan_n_Page_152.txt

brennan_n_Page_166.txt

brennan_n_Page_074.txt

brennan_n_Page_036.txt

brennan_n_Page_147.txt


Full Text





CHAPTER 4
CRYPTIC CANNIBALISM IN SIZE-STRUCTURED SNOOK POPULATIONS

Introduction

Many species experience elevated mortality during specific ontogenetic stages, often

highest during early life stages (Smith and Fretwell, 1974; Jakobsson and Eriksson, 2000; Einum

et al., 2006). In fish, this especially occurs during larval stages (Houde 1987; Doherty and

Fowler, 1994; Lorenzen, 1996). While many processes contribute to early-stage mortality,

predation can be a primary contributor (Mol, 1996; Duffy et al., 1997; Taylor and Danila, 2005;

Titelman and Hansson, 2006). Nonetheless, high mortality can also manifest in later stages with

reduced refuge or increased foraging risk (e.g. Mooij et al., 1996; Bystroim, 2004).

Cannibalism, common among piscivorous fish, can be a significant contributor to

mortality during these early life stages (e.g. in Eurasian perch (Persson et al., 2000), cod

[Nuenfeldt and Koster 2000], cape hake [Pillar and Wilkinson, 1995], pollock [Bailey, 1989;

Wespestad et al., 2000], silver hake [Garrison and Link, 2000], and walleye [Liao et al., 2004]).

In controlled high density settings (i.e. aquaculture) cannibalism is commonly documented

(Baras et al., 2003; Chang and Lio 2003; Tucker, 2003; Correa and Cerqueira, 2007), yet

documentation of cannibalism in the wild is less common. This may be attributable to

inappropriate collection times or sample sizes (Bowen and Allanson, 1982; Johnson and

Dropkin, 1993), difficulties in identifying digested prey items with fast gut evacuation rates

(Kennedy, 1969), sampling logistics,(Rindorf, 2004) or high variation in inter-annual recruitment

patterns if cannibalism is density dependent (Nuenfeldt and Koster, 2000). In cod (Uzars and

Plikshs, 2000), for example, cannibalism rates are density dependent, with elevated cannibalism

during high recruitment, but negligible otherwise.










cannibalistic (Tucker, 2003; Adams and Wolfe, 2006), and age-0 and age-1 snook share nursery

habitat for a prolonged period and increases in abundance of age-1 snook may have negative

consequences on young-of-year snook populations through cannibalism and competition for

refuge and food resources. Cyclic patterns of abundance due to cannibalism are common

(Frankiewicz et al., 1999; Sanderson et al., 1999; Fromentin et al., 2000; Claessen et al., 2000;

Persson and de Roos, 2006). Snook populations in Florida show strong evidence for patterns of

alternating fluctuations of abundance between age-0 and age-1 year classes (Brennan,

unpublished data) as do a cousin of snook, barramundi Lates calcifer in Australian estuaries

(Walters and Martell, 2004; Griffin, unpublished data) due to intra- and inter-cohort cannibalism

and competition. Release programs that elevate age-1 abundance could intensify such effects

and a cautionary approach is warranted. Nonetheless, the potential for elevating stocks that are

below capacity is demonstrated by this study and its potential over the long-term must be

evaluated.

The presence of small (~ 60 mm FL) snook in the creeks during fall reflects continued

use of creek habitat by young snook throughout the year. Dramatic declines in abundance of

age-1 snook in the creeks during late summer early fall (~ 200-325 mm FL at this time),

probably represents an ontogenetic habitat shift of the age-1 snook. This could possibly be due

to (1) a reduction in predation threat as snook attained larger sizes (Koczaj a et al., 2005), (2)

superior foraging habitat and improved prey availability and preferences outside the creeks

(Ruzycki and Wurtsbaugh, 1999), (3) better water quality outside the creeks (Brennan and Leber,

unpublished data), or (4) combinations of the above reflecting tradeoffs with size-dependant

predation threat, and improvements in growth due to better foraging opportunities (e.g., Werner

and Gilliam, 1984; Walters and Juannes, 1993; Olsen et al., 1995).










Jobling, M., Gwyther, D., Grove, D.J., 1977. Some effects of temperature, meal size and body
weight on gastric evacuation time in the dab Limanda limanda (L). J. Fish Biol. 10, 291-
298.

Jokikokko, E., Kallio-Nyberg, I., Saloniemi, I., Jutila, E., 2006. The survival of semi-wild, wild
and hatchery-reared Atlantic salmon smelts of the Simojoki River in the Baltic Sea. J.
Fish Biol. 68 (2), 430-442.

Johnson, J.H., Dropkin, D.S., 1993. Diel variation in diet composition of a riverine fish
community. Hydrobiologia 271, 149-158.

Jonsson, N., Jonsson, B., Hansen, L.P., 2003. The marine survival and growth of wild and
hatchery-reared Atlantic salmon. J. App. Ecol. 40 (5), 900-911.

Jones, G.P., 1997. Relationships between recruitment and postrecruitment processes in lagoonal
populations of two coral reef Eishes. J. Exp. Mar. Biol. Ecol. 213 (2), 23 1-246.

Jorstad, K., 2004. Genetic studies in marine stock enhancement in Norway. In: Leber, K.M.,
Kitada, S., Blankenship, H.L., Svsand, T. (Eds.), Stock enhancement and sea ranching.
Blackwell Publishing, Oxford, UK, pp. 339-352.

Juanes, F., 2003. Allometry of cannibalism in piscivorous Hishes. Can. J. Fish. Aquat. Sci. 60,
594-602.

Kaeriyama, M., 1999. Hatchery programmes and stock management of Salmonid populations in
Japan. In Howell, B.R., Moksness, E., Svisand, T., (Eds.), Stock Enhancement and Sea
Ranching (Ed. B.R. Howell, E. Moksness, and T.), Blackwell Scientifie Publications,
Oxford, UK, pp. 153-167.

Keeley, E.R., 2001. Demographic responses to food and space competition by juvenile steelhead
trout. Ecol. 82 (5), 1247-1259.

Kellison, G.T., Eggleston, D.B., Tanaka, M., 2002. Density-dependent predation and
implications for stock enhancement with Japanese flounder. J. Fish. Biol., 60 (4), 968-
980.

Kellison, G.T., Eggleston, D. B., 2004. Coupling ecology and economy: modeling optimal
release scenarios for summer flounder (Pal(Iid irulip dentatus) stock enhancement. Fish.
Bull. 102 (1), 78-93.

Kemp, W.M, Sampou, P.A., Garber, J., Tuttle, J., Boynton, W.R., 1992. Seasonal depletion of
oxygen from bottom waters of Chesapeake Bay roles of benthic and planktonic
respiration and physical exchange processes. Mar. Ecol. Prog. Ser. 85 (1-2), 137-152.

Kennedy, C.R. 1969. Tubifieid oligochaetes as food of dace. J. Fish Biol. 1, 1 1-15.









in abundance from pre- to post-release (using either summer or winter as post-release values;

p=0.42, p=0.45, R2=0.09-0.33).

Creek Density

The treatment creeks used in this study were quite different in their ability to

accommodate augmented snook densities. The analysis of covariance model showed no

significant differences in pre-release snook density (p=0.51, 3 df, Figure 3-6), however, after

releases occurred, WB showed a significant increase in overall density (over twice the pre-

release level) that was maintained through the winter (p=0.034, 3-6). In BC (also a high

augmentation treatment), the expected augmented levels were not sustained (Figure 3-6), and I

estimated that 64-85% of the released cultured fish were lost by the time of the first post-release

sample. Loss within wild snook populations was not detected over this period, however. Creeks

that received low levels of augmentation showed no significant changes in overall juvenile snook

density throughout the study (Figure 3-6). Chi-square analysis of the observed and expected

densities in all four creeks reflected the initial loss of hatchery fish in BC by showing a

significant difference between observed and expected post-release densities (p<0.001, Chi-

square= 306.82, 3 df). Thus, responses of juvenile snook density to various levels of

augmentation were creek specific.

Discussion

Experimental releases of high-densities of age-1 hatchery-reared snook in two estuarine

tributaries elevated total age-1 abundance above pre-release levels throughout the study. At the

same time, I found no detectable evidence of suppressed abundance of wild age-1 snook in these

habitats. Surplus production capacity for age-1 snook existed at these particular experimental

times and locations, although inter-cohort or community-wide responses were unclear. This is

not surprising as others have found population sizes below habitat productive capacities (e.g.,









the stocked snook remained by first winter, but in the other creek only 0.15-0.36 remained by

first winter). Negligible losses of wild conspecifics were observed in all stocked creeks

however. These results were also positively related to ambient density. Sites with high ambient

density of wild juveniles also had high post-release survival rates of hatchery juveniles, and sites

with low ambient densities resulted in low survival rates of the stocked fish. This suggests that

while growth responses may have been density-dependent, stocking of age-1 snook (on this order

of magnitude) may not have resulted in density-dependent mortality within the cohort, but might

be controlled by density-independent effects such as rearing habitat quality or refuge availability.

Stocking at higher magnitudes may have very different results and a cautionary approach is

warranted. Inter-annual variation in recruitment of snook, its competitors, prey, or predators

could all influence the dynamics of density-dependent responses (Chapter 4 examines the

potential influence of intercohort cannibalism, see below). Aside from further potential

influence of variability in hatchery-based components (such as variation in health of the stocked

snook, or complications with release logistics [Congleton et al., 2000; Sulikowski et al., 2006]),

stocking the same habitats and release magnitudes could still cause very different results, just

due to year-to-year effects. Nonetheless, increases in experience with this work (e.g. increases in

the number of experimental years, additional stocking levels, and across habitats) can increase

confidence in our expectations of such activities, however. Coupling this with experimental

modeling scenarios of stock enhancement (see below), can further add to our understanding of

these issues.

An important consideration for stocking piscivorous and cannibalistic fishes, such as

snook, is the potential influence of intra- and inter-cohort cannibalism and competition as I

examined in Chapter 4. With snook, this potential is further intensified because, as in northern










Steneck, R.S., 2006. Is the american lobster, Homarus amnericanus, overfished? A review of
overfishing with an ecologically based perspective. Bull. Mar. Sci 78 (3), 607-632.

Stottrup, J.G., Sparrevohn, C.R., Modin, J., Lehmann, K., 2002. The use of releases of reared
fish to enhance natural populations A case study on turbot Psetta maxima (Linne, 1758).
Fish. Res. 59 (1-2), 161-180.

Story, M., Gruder, E.W., 1936. Mortality of Eishes due to cold at Sanibel Island, Florida.
Ecology 18(1), 10-26.

Sulikowski, J.A., Fairchild, E.A., Rennels, N., Howell, W.H., Tsang, P.C.W., 2006. The effects
of transport density on cortisol levels in juvenile winter flounder, Pseudopleuronectes
amnericanus. J. World Aqua. Soc. 37 (1), 107-112.

Svisand, T., 1998. Cod Enhancement Studies in Norway Background and Results with
Emphasis on Releases in the Period 1983-1990. Bull. Mar. Sci. 62,313-324.

Svisand, T., Kristiansen, T.S., 1990. Enhancement studies of coastal cod in western Norway,
part IV. Mortality of reared cod after release. J. Con. Cons. Int. Explor. Mer 47, 30-39.

Svisand, T., Kristiansen T.S., Pedersen, T, Salvanes, A.G., Engelsen, V.R., Naevdal, G.,
Nardtvedt, M., 2000. The enhancement of cod stocks. Fish and Fish. 1, 173 205.

Taylor, D.L., Danila, D.J., 2005. Predation on winter flounder (Pseudopleuronectes amnericanus)
eggs by the sand shrimp (Cranzgon septemspinosa). Can. J. Fish. Aquat. Sci. 62 (7),
1611-1625.

Taylor, R.G, Grier, H.J., Whittington, J.A., 1998. Spawning rhythms of common snook in
Florida. J. Fish Biol. 53 (3), 502-520.

Taylor, R.G., Whittington, J.A., Grier, H.J., Crabtree, R.E., 2000. Age, growth, maturation, and
protandric sex reversal in common snook, Centropomus undecimalis, from the east and
west coasts of South Florida. Fish. Bull. 98 (3), 612-624.

Taylor, M.D., Palmer, P.J., Fielder, D.S., Suthers, I.M., 2005. Responsible estuarine finish
stock enhancement: an Australian perspective. J. Fish. Biol. 67 (2), 299-331

Teknomo, K., 2006. Bootstrap Sampling Tutorial. http:.//people.revoledu. com/kard/
tutorial/b ootstrap/

Titelman J., Hansson L.J., 2006. Feeding rates of the j ellyfish Aurelia aurita on fish larvae.
Mar. Biol.149 (2), 297-306.

Tolley, S.G., Dohner, E.T., Peebles, E.B., 1987. Occurrence of larval snook, Centropomus
undecimalis (Bloch), in Naples Bay, Florida. Florida. Sci. 50 (1), 34-38.











Liao, H.S., Pierce, C.L., Larscheid, J.G., 2004. Consumption dynamics of the adult piscivorous
fish community in Spirit Lake, Iowa. North American Journal of Fisheries Management
24 (3), 890-902.

Linehan, J.E., Gregory, R.S., Schneider, D.C., 2001. Predation risk of age-0 cod (Galdus)
relative to depth and substrate in coastal waters. J. Exp. Mar. Biol. Ecol. 263, 25-44.

Linton, T. L., Richards, W. L. 1965. Young common snook on the coast of Gerogia. Quart. J.
Florida Acad. Sci. 28 (2), 185-189.

Lipcius, R.N., Vanengel, W.A., 1990. Blue-crab population-dynamics in Chesapeake Bay -
variation in abundance (York River, 1972-1988) and stock-recruitment functions. Bull.
Mar. Sci. 46 (1), 180-194.

Lochet, A., Lambert, P., Lepage, M., Rochard, E., 2004. Growth comparison between wild and
hatchery-reared juvenile European sturgeons Acipenser sturio (Acipenseridae) during
their stay in the Gironde estuary (France). CYBIUM 28 (1), 91-98 Suppl. S.

Longhurst, A., 2006. The sustainability myth. Fish. Res. 81, 107-112.

Lotze, H.K., Lenihan, H.S., Bourque, B.J., Bradbury, R.H., Cooke, R.G., Kay, M.C., Kidwell,
S.M., Kirby, M.X., Peterson, C.H., Jackson, J.B.C., 2006. Depletion, degradation, and
recovery potential of estuaries and coastal seas. Science 312 (5781), 1806-1809.

Lorenz, J.J., 1999. The response of Eishes to physicochemical changes in the mangroves of
northeast Florida Bay. Estuaries 22 (2B), 500-517.

Lorenz, J.J., Sarafy, J.E., 2006. Subtroprical wetland fish assemblages and changing salinity
regimes: Implications for everglades restoration. Hydrobiologia 569, 401-422

Lorenzen, K., 1996. The relationship between body weight and natural mortality in fish: a
comparison of natural ecosystems and aquaculture. J. Fish Biol. 49, 627-647.

Lorenzen, K., 2000. Allometry of natural mortality as a basis for assessing optimal release size
in fish-stocking programmes. Can. J. Fish. Aquat. Sci. 57, 2374-2381.

Lorenzen, K., 2005. Population dynamics and potential of fisheries stock enhancement: practical
theory for assessment and policy analysis. Philos. T. Roy. Soc. B. 260, 171-189.

Lorenzen, K., 2006. Population management in fisheries enhancement: gaining key information
from release experiments through use of a size-dependent mortality model. Fish. Res. 80,
19-27.

Lorenzen, K., 2008. Understanding and managing fisheries enhancement systems. Rev. Fish.
Sci. 16 (1-3), 10-23.













Shore
1 36 m



&3 m


Shore


Figure 3-2.


Schematic of various depletion-removal methods including (A) "two-bank
depletion", (B) "single-bank depletion", and (C) "two-bank extended" depletion.
Dark lines indicate block nets, lighter dashed lines indicate locations of seine
efforts, and arrows indicate the direction of the seine hauls. The numbers in
brackets represent the sequence and location of seine hauls (hauls 1 through 9).









effects on wild stocks (Hilborn, 2004; Walters and Martell, 2004). As a result, experiments are

critical to assess the potential feasibility of stock enhancement prior to implementation of a

large-scale program (Blankenship and Leber 1995; Hillborn, 1998; Walters and Martell, 2004).

Cannibalism is one factor that may lead to unintended effects of stock enhancement.

In snook, cannibalism has been documented in aquaculture (Serfling, 1998; Tucker,

2003) and is most intense during early life stages in hatcheries (Tucker, 2003; N. Brennan,

personal observation). In the wild, snook juveniles actively prey on small Eishes and crustaceans

(Fore and Schmidt 1973; Gillmore et al. 1983; McMichael et al. 1989; Sazima 2002), and adults

are highly piscivorous (Blewitt et al., 2006), yet cannibalism is only rarely documented (e.g.,

Adams and Wolfe, 2006; N. Brennan, unpublished data). Therefore, cannibalism could be

perceived as non-signifieant factor influencing population size. However, even low observed

rates of cannibalism or the threat of cannibalism could strongly influence population dynamics

when considering competition for limited refuge in the presence of predators, and prolonged

seasonal threats of cannibalism. Identifying cannibalism from dietary studies in the Hield may be

a relatively rare event compared to more common prey choices of a particular predator of various

ages, especially when older age classes are not always occupying the same habitats as juvenile

conspecifies. A single incidence of cannibalism results in mortality of a prey individual and

potentially the predator (if the cannibal also chokes and suffocates). Over time (throughout a

season or year) cannibalism focused on a year class of juveniles could act as a strong force in

structuring abundance of the year class (Claesson et al., 2000; Neuenfeldt and Koster, 2000;

Persson et al., 2000), especially during years of high abundance of particular year classes.

In many species, ontogenetic changes in morphology, and behavior suggests adaptations

in earlier stages for coping with smaller absolute body size in the presence of predation and









hatchery-reared fish were included three creeks showed abundance increases compared to April

(BC =1.20x, WB =2.26x, NC =0.81Ix, SC = 2.61x that of pre-release estimates). In June,

abundance estimates of hatchery-reared fish were 120% larger than wild juvenile abundance in

WB and 40% of the wild abundance in BC (but 97% in July), 29% of NC wild abundance and

6% of SC wild abundance (control creeks) (Figure 3-5a, Figure 3-5b).

Juvenile snook generally emigrated from all four creeks by fall, then returned in winter

(Figure 3-4). As summer progressed, in-creek abundance of age-1 juveniles steadily declined,

and only young-of-the-year snook (from 2002 summer spawns, < 100 mm FL) numerically

dominated snook catches. For example, in early summer, hatchery-reared snook abundance was

37% of the number released in BC, ~100% in WB, 85% in NC, and 75% in SC, respectively. By

fall, large declines in hatchery-reared fish were evident in all creeks; 6%, 6%, 17%, and 0% of

the original numbers released in BC, WB, NC, and SC, respectively (Figure 3-4). Wild age-1

snook exhibited similar patterns of large, late-summer declines in abundance in all creeks. After

the first significant drop in temperature (below 22oC), both wild and hatchery-reared fish

abundance returned to substantial levels comparable to early-summer abundance; winter

abundance of wild snook was 120% the peak summer abundance in BC, 134% in WB, 96% in

NC, and 89% in SC. Winter abundance of hatchery-reared snook was estimated at 54% of the

summer abundance in BC, 78% in WB, 69% in NC, and 258% in SC (Figure 3-4).

Power analysis showed that, given the variability in the catch data and the small number

of replicates used in the study, power was low (0.3 0.8) for detecting a 100% difference in

population means. Nonetheless, repeated measures analysis showed a significant time and

treatment interaction effect for pre- and post-release abundance (p=0.024, F = 10.78, df = 2, 3).

Regression analysis showed no significant relationship between augmentation level and change









able to tolerate hypoxic conditions than are larger snook and hypoxic sites may serve as an

important physiological refuge from aquatic predators. Other attributes of the habitats (i.e.

gentle bank slopes, soft sediments, low salinity, and high temperatures) were also typically found

at these sites and other studies characterizing juvenile habitats corroborate this finding (Gilmore

et al., 1983; McMichael et al., 1989; Peters et al. 1998a). Summer time water quality at island

habitats correlated with higher dissolved oxygen levels, higher salinity, pH, and loosely

correlated with steeper banks. Arguably, the low recapture ratios at these sites might reflect

higher predation risk and mortality because more potential predators can tolerate these

conditions. High predation risk could cause size-dependent survival where only the largest and

fastest growing snook survive beyond the vulnerable size range. If this were true, however lower

recapture rates of small SAR snook should have been observed at these sites. I found evidence

that would support this in plots of recapture lengths according to release habitats (Figure 2-7) but

not in plots of growth rates and SAR, or the canonical correlation analysis (Figure 2-8, 2-13).

Within a 42 d window after release, larger SAR individuals showed patterns of higher

mean dispersal distances compared with smaller individuals, and SAR was loosely correlated

with first summer and winter dispersal rates. Other studies have shown that relative predation

risk and dispersal associated with body size (e.g. Werner and Gilliam, 1984; Walters and

Juannes, 1993; Armstrong et al., 1997). Smaller individuals, in general, are at higher risk from

predation than large individuals, especially outside of refugia, and movement is comparatively

reduced. Lorenzen (2000) found a negative, linear relationship between mortality of released

fish and length. In contrast, large fish benefit from improved foraging opportunities therefore

increased growth rates. In this study, patterns of increasing dispersal distances with body size is

only suggestive of lower predation risk in large fish.











400


375


350


325


300


275


250


225


200


175


150


ISLANDS


TIDY


Figure 2-7.


Size-at-recapture for snook released in different habitats. Data are from
recaptures collected 6-10 months after releases in April 1998, and 1999. Sites
from left to right are arranged in a North to South order.









or no) and was fit to linear regression models through maximum likelihood methods using

logistic regression techniques (CATMOD procedure in SAS, [SAS, 1999]). I used Bonferroni

adjustments on treatment comparisons to reduce the likelihood of making a Type I error.

Approach 3: Cannibalism Intensity in Wild Snook Populations

Literature Review: I evaluated cannibalism intensity in wild common snook using

published and unpublished literature, (Table 4-1). Published literature was obtained from the

Web of Science database using the search criteria "Centropomus undecimalis" and "diet or

feeding or cannibalism". I gained further literature by using citations of other work from the

reference lists provided in these manuscripts. Unpublshed material was obtained from work in

Sarasota from the snook stock enhancement research program. Whenever possible, I extracted

data on the size and numbers of the snook examined, collection year, season, and region of data

collection, habitat descriptions of where samples were taken, the percent of the samples

containing food items in the stomach, and the number of snook reported cannibalized.

Field Collections: I also assessed cannibalism rates in the field, by conducting my own

investigations in Sarasota, Florida with snook collected for Mote Marine Laboratory's Stock

Enhancement Research Program from 2001- 2005 (Table 4-1). All collections occurred in

estuarine tidal creeks, an important rearing habitat for age-0 snook. Snook were collected with

bag seines during (1) June 2001 (2) February May 2004, (3) July-August 2004, and (4) May-

August 2005. All snook were measured and stomach contents were extracted using a non-lethal

lavage technique (Adams and Wolfe, 2006), immediately preserved in 10% buffered formalin,

and returned to the laboratory for identification. Stomach contents were visually identified to the

lowest possible taxonomic level. For more digested stomach contents, I used any conspicuous

body parts for identification such as carapaces, rostrums, fish jaws and otoliths. Identification of

stomach contents beyond this were categorized as "unidentified".











Hilborn, R, Walters, C.J., 1992. Quantitative Fisheries Stock Assessment: Choice, Dynamics
and Uncertainty. Chapman and Hall, New York, NY.

Honma, A., 1993. Aquaculture in Japan. Japan FAO Association, Tokyo.

Hossain MAR, Tanaka M, Masuda R., 2002. Predator-prey interaction between hatchery-reared
Japanese flounder juvenile, Pat(Iid irulip olivaceus, and sandy shore crab, Matuta lunaris:
daily rhythms, anti-predator conditioning and starvation. J. Exp. Mar. Biol. Ecol 267 (1),
1-14.

Houde, E.D., 1987. Fish early life dynamics and recruitment variability. In: Houde, E.D., (Ed),
10th Annual Larval Fish Conference. American Fisheries Society, Symposium 2,
Bethesda, MA, pp. 17-29.

Hurst, T.P., Conover, D.O., 1998. Winter mortality of young-of-the-year Hudson River striped
bass (Moon sxails) size-dependent patterns and effects on recruitment. Can. J. Fish.
Aquat. Sci. 55, 2709-2709.

Hutchings, J.A., 2000. Collapse and recovery of marine fishes. Nature 406, 882-885.

Immamura, K., 1999. The organization and development of sea farming in Japan. In Howell,
B.R., Moksness, E., Svisand, T., (Eds.), Stock Enhancement and Sea Ranching (Ed. B.R.
Howell, E. Moksness, and T.), Blackwell Scientific Publications, Oxford, UK, pp. 91-
101.

Islam, M.S., Haque, M., 2004. The mangrove-based coastal and nearshore fisheries of
Bangladesh: ecology, exploitation and management. Rev. Fish Biol. Fish. 14 (2), 153-
180.

Jackson, J.B.C., Kirby, M.X., Berger, W.H., Bjorndal, K.A., Botsford, L.W., Bourque, B.J.,
Bradbury, R.H., Cooke, R., Erlandson, J., Estes, J.A., Hughes, T.P., Kidwell, S., Lange,
C.B., Lenihan, H.S., Pandolfi, J.M., Peterson, C.H., Steneck, R.S., Tegner, M.J., Warner,
R.R., 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science
293 (5530), 629-638.

Jakobsson, A., Eriksson, O., 2000. A comparative study of seed number, seed size, seedling size
and recruitment in grassland plants. Oikos 88 (3), 494-502.

Jefferts, K.B., Bergman, P.K., Fiscus, H.F., 1963. A coded wire identification system for macro-
organisms. Nature 198, 460-462.

Jenkins, W.E., Denson, M.R., Bridgham, C.B., Collins, M.R., Smith, T.I.J., 2002. Retention of
oxytetracycline-induced marks on Sagittae of red drum. N. Am. J. Fish. Manage. 22 (2),
590-594.









Researchers found that snook acclimated in predator-free enclosures located at release sites had

about twice the survival rates compared to those released without acclimation (Brennan et al.,

2006).

Ongoing sampling of the localized snook population has documented survival to

adulthood of the released hatchery-reared juveniles. Since 1999, individuals from release

cohorts have been regularly documented in annual wild snook spawning aggregations. Both

male and female hatchery-reared snook have been recaptured with mature gonads, and some

recaptured individuals have outgrown the harvestable size slot (27-33 inches total length [TL, or

686-838 mm TL]) (Leber and Brennan, unpublished data). Releasing juvenile snook into a

variety of habitats has provided valuable understanding of essential fish habitat (as defined in the

1996 amendments to the Magnuson-Stephens Act) for juvenile snook (Chapter 2). Quantitative

sampling in juvenile and adult snook habitats over the years also allowed researchers to measure

relative contribution rates of various juvenile rearing habitats to adult demes, and documenting

associated migratory patterns from juvenile to adulthood (Chapter 2). Mark-recapture data from

hatchery-reared and wild snook using a variety of tag technology has provided important insight

into short and long-term dispersal patterns (Chapter 2 and 3; Pine et al, unpublished data). Using

hatchery-reared juvenile snook to experimentally manipulate ambient juvenile abundance in tidal

creeks researchers were able to examine how density-dependent responses within a cohort varied

between habitats and across seasons (Chapter 3).

Stomach content examination of wild and hatchery-reared snook from field collections

has also provided valuable insight into snook dietary preferences, stomach content evacuation

rates Brennan and Heinlein, unpublished data), and diel feeding periodicity (Rock et al.,

unpublished data). With evidence for density-dependent survival in juvenile snook, the









individuals may have negative consequences for snook stocks on both local and regional levels.

During winter, however, mature snook showed consistently high Eidelity to rearing habitats

(especially in tidal creeks) demonstrated through consistent use of these sites each winter as they

sought thermal refugia (Figure 2-16). Understanding these snook behaviors and life history

patterns are important for understanding how production and quality of rearing habitats can

influence localized demes and fisheries.

Recapture ratios showed some difficulties when used as indices of survival because of

uncertainties in post-release distribution of the juvenile snook. Our short-term sampling showed

significantly higher recapture ratios from fish released at upstream and midstream habitats than

those released at barrier islands. These results may have been biased in several ways: (1)

Sampling took place at the release sites. Thus, for each release creek system proportionally more

sampling effort occurred within the creeks compared to barrier islands (about three times as

much effort in creek habitat). (2) Barrier island habitats represent a larger and more "open"

habitat than within-stream habitats. Even if snook dispersed randomly, our sampling regimen

would recapture fewer snook released at barrier islands, than within the creeks, because the

creeks were more restrictive, with less overall dispersal area available. (3) Within the creeks, a

general downstream movement of snook may have resulted in disproportionately more

recaptures of snook from "upstream" release locations compared to stream mouth release sites.

Very few juvenile hatchery snook were recaptured outside of the creeks, but this could have been

due to less sampling effort in these habitats. A generalized downstream movement pattern was

not found from our recaptured snook, however. Often, snook released downstream were

captured upstream and midstream, and snook released at the midstream habitats were found in









the enclosures, appeared to have died many days prior to examination (about 5d or more).

Mortality appears to have been handling-induced (93% from trials on November 12 and 19,

2003) and this is supported by evidence that the stocked snook died soon after stocking (highly

decayed remains). Increased replication of these treatments, the use of controls, and examination

of stomach contents to confirm cannibalistic behavior in the enclosures is needed to clarify

relative size-specific cannibalism potential in snook.

Snook diets revealed a surprisingly low rate of cannibalism (7 cases recorded of 6,768

stomachs examined, 0.10%) in populations over a wide variety of locations. Assuming these

rates are reflective of typical snook diet, the average number of snook processed for stomach

contents in a single study (mean=3 59) was just over 1/3 the number needed to identify a single

cannibalism instance. As snook cannibalism was targeted at and within the age-0 cohort, sample

location and timing could strongly bias cannibalism rates. Samples collected in age-0 habitats,

especially nursery habitats (see Beck et al., 2001), are therefore much more likely to contain

evidence of cannibalism. Within a given year, cannibalism rates are also likely to vary over

seasons due to variations in intra- and inter-cohort abundance and interactions of size-specific

habitat use (e.g. Bystroim et al., 2003). With snook in northern end of their temperature tolerance

range, older age classes of snook often over-winter in rearing habitats such as tidal creeks and

backwater habitats because of warmer water (Marcinkiewicz, 2007; Chapter 2) In this

scenario, over-winter habitats coincide with snook nursery habitats and cannibalism may be

intensified.

Inter-annual timing of sample collection could also bias results as cannibalism rates are

strongly linked to snook density. Samples collected during years of low juvenile recruitment

will be less likely to contain evidence for cannibalism, while cannibalism is much more likely









Recent progress has also occurred through understanding the broader potential for stock

enhancement (Taylor et al., 2005; Garaway et al., 2006; Lorenzen, 2006; Lorenzen, 2008). This

involves an integrated analysis of biological, economic, and social criteria for suitable candidate

species for stock enhancement activities. Much of this work directly incorporates stakeholder' s

priorities and needs that are primarily responsible for driving most stock enhancement activities.

Models are needed to evaluate the candidacy of species targeted for stock enhancement

considering what is known about life history traits and other parameters of the stock (e.g.

fecundity, recruitment dynamics, density-dependent growth and survival), cost effectiveness of

hatchery production, fishery economics and logistics, how it integrates with other forms of

fishery management, and intuitional goals and priorities (Lorenzen, 2008). These types of

proactive evaluation measures are needed prior to commitments of large investments in stock

enhancement programs (Hilborn, 1998; Walters and Martell, 2004; Lorenzen, 2008). This has

been the exception rather than the norm for most stock enhancement programs however (Bell et

al., 2006; Bell et al., 2008). The inclusion of stakeholders in the process (e.g. Garaway et al.,

2006; Uki et al., 2006; Becker et al., 2008; Tomiyama et al., 2008; ), however, is a positive step

towards understanding and directing stock enhancement and other fishery management programs

toward adaptive management approaches (Walters, 1997; Perry et al., 1999; Bell et al., 2008).

Less understood and less demonstrated, yet key areas of uncertainty in this field, include

density-dependent effects, genetic effects, and economic feasibility. Numerical responses to

stocking in wild and hatchery-sourced animals (e.g. intra- and inter-specific competition or

cannibalism) and interactive effects within and across generations are rare (Leber et al., 1995),

and remain a high priority (Leber et al., 2004; Walters and Martell, 2004; Lorenzen, 2005).

Studies examining the effects of stock enhancement programs on inter-annual and spatial










De Innocentiis, S., Longobardi, A., Marino, G., 2008. Molecular tools in a marine restocking
program for the endangered dusky grouper, Epinephelus marginatus. Rev. Fish. Sci. 16
(1-3), 269-277.

Doherty, P. J., and A. Fowler. 1994. An empirical test of recruitment limitation in a coral reef
Eish. Science 263, 935-939.

Doherty, P.J., Dufour, V., Galzin, R., Hixon, M.A., Meekan, M.G., Planes, S., 2004. High
mortality during settlement is a population bottleneck for a tropical surgeonfish. Ecol. 85
(9), 2422-2428.

Dorsey, L.G., 2004. Retention of coded wire tags by age-0 Muskellunge. J. Freshwater Ecol. 19
(2), 333-337.

dos Santos, J., Jobling, M., 1991. Factors affecting gastric evacuation in Cod, Gadus morhua-L.,
fed single-meals of natural prey. J. Fish Biol. 3 8, 697-713.

Duffy, J.T., Epifanio, C.E., Fuiman, L.A., 1997. Mortality rates imposed by three scyphozoans
on red drum (Sciaenops ocellatus Linnaeus) larvae in field enclosures. J. Exp. Mar. Biol.
Ecol. 212 (1), 123-131.

Dumont, H.J., A. Jawahar, A., 2004. Stage-specific cannibalism and spontaneous cyst hatching
in the freshwater fairy shrimp Streptocephalus proboscideus Frauenfeld. Hydrobiologia,
524, 103-113.

Einum, S., Sundt-Hansen, L., Nislow, K.H., 2006. The partitioning of density-dependent
dispersal, growth and survival throughout ontogeny in a highly fecund organism. Oikos
113 (3), 489-496.

Fairchild, E.A., Rennels, N., Howell, W.H., 2008. Predators are attracted to acclimation cages
used for winter flounder stock enhancement. Rev. Fish. Sci. 16 (1-3), 262-268.

FAO, 2004. The state of world fisheries and aquaculture: 2004. FAO, Rome (also available at
ftp://ftp.fao.org/docrep/fao/007/y56 00e/y50r0pf

Faunce, C.H., Serafy, J.E., 2006. Mangroves as fish habitat: 50 years of field studies. Mar.
Ecol.-Prog. Ser. 318, 1-18.

Fleming, I.A., Petersson, E., 2001. The ability of released, hatchery salmonids to breed and
contribute to the natural productivity of wild populations. Nordic J. Freshw. Res. 75 (5),
71-98.

Fore, P.L. Schmidt, T.W., 1973. Biology of juvenile and adult snook, Centropomus undecimalis,
in the Ten Thousand Islands, Florida. U.S. Environmental Protection Agency,
Surveillance and Analysis Division. Publication number EPA 904/9-74-002, Chapter 16.
Region IV, Atlanta, GA.











Table 2-5. Snook captures organized by sampling type and locations. These data were used
to verify results from short-term sampling and monitor recruitment to adult
habitats
Year Beaches and Passes Intercoastal Red Tide Tournaments Release sites
1998 .112 37 518
1999 349 603 48 1,904
2000 393 102 .191 1,071
2001 867 1,103 128 2,838
2002 431 .195 1,886
2003 489 13 155 2,587
2004 354 208 .309 3,293
2005 ..350 195 888
Totals 2,883 1,038 1,453 1,258 14,985
Grand Total 21,617










Dispersal: Generally mean dispersal distances increased over seasons and years (Figure

2-16). Snook recaptured during summer and fall had the highest mean dispersal distances, and

were rarely recaptured in tidal creeks during these seasons, although sampling occurred regularly

there. Mean dispersal distances during winter did not increase across years however, and

showed high rearing site fidelity each winter (Figure 2-16). Snook tended to recruit to adult

spawning habitat nearest their rearing habitats.

Mulitvariate analysis with long-term responses: The second principle component

analysis comparing key features of release sites and short-term post-release responses with long-

term responses (Table 2-7) showed the first three components explained 80% of the standardized

variance (36%, 28%, 16% respectively). Examining a plot of the first two components (Figure

2-17), age-1 summer CPUE, age-1 winter recapture ratio, and long-term recapture ratios (from

dispersers and those captured at beaches and passes) were all within the upper left quadrant.

Short- and long-term growth patterns were closely related and generally inversely related to

recapture ratios.

Contribution to adult stocks: Least squares procedures predicted recapture ratios as a

linear function of loge (In) transformed first winter growth rates (giron il, mm/d) as:

Recapture ratio = -1. 9596* Ln (giron "\ il 0.2578 (2-3)

The correlation coefficient for this function was 0.59. Predictably, plots of numbers of hatchery-

released snook recruiting to adult stocks according to age and rearing habitat (expressed as short-

term growth rates), shows highest contributions from rearing habitats with lowest growth rates

and highest post-release survival (Figure 2-18a). Recruitment of stocked juveniles from rearing

habitats to age-2 adult stocks (Nage-2, to second summer) were expressed as:

Nage-2 = -286 (In(growth))+ 81 (2-4)









in understanding stock enhancement potential, empirical research such as this study, can aid in

resolving these issues. In this study I found that rearing habitats where snook were stocked

influenced post-release growth and recapture ratios at first winter and these traits were

perpetuated for years after releases. Generally, post-release growth was inversely related to

recapture ratios and tied to characteristics of the rearing habitats. Sites with highest recapture

ratios and low growth rates (often tidal creeks), generally contained soft sediment with high

organic content and low dissolved oxygen levels in the water. These sites were also correlated

with high ambient densities of juvenile snook. Sites with low recapture ratios and high growth

rates generally had high salinity and pH, firmer sediment, and low ambient densities of juvenile

snook (often barrier island habitat). I tied first winter recapture ratios from this study to real

survival rates of stocked snook at first winter from another study (Brennan et al., 2008, Chapter

3) and used long-term growth data to proj ect contributions of the various rearing habitats to adult

stocks. Sites with highest recapture ratios and low growth made the highest numerical

contributions to adult stocks, but biomass contributions were among the lowest. Hatchery-reared

juveniles stocked at sites that produced intermediate growth rates and recapture ratios, made the

highest biomass contributions to adults stocks. Follow up work is needed to understand

interactive implications between stocking magnitude (Brennan et al., 2008, Chapter 3) and

responses in hatchery and wild snook (across age classes). This work provides important

information for resource managers considering the productivity of various rearing habitats and

has both fishery and habitat management implications. Follow up models examining stock

recruitment and fisheries dynamics under various enhancement scenarios (e.g. Lorenzen, 2006)

can be coupled with results from this study. Together, these can be powerful tools that provide

essential information for responsible and adaptive management of these valuable stocks.









variation in recruitment across community and ecosystems that aligns with current ecosystem-

based management programs (e.g. Pauley et al., 2000) are also a high priority. Genetic effects

and consequences of releasing hatchery-reared animals on wild populations have been predicted

(e.g. Ward, 2006; De Innocentiis et al., 2008; Tringali et al., 2008a) and serve as useful

guidelines for stock enhancement programs. Application of these guidelines, however, remains a

challenge and adds to the difficulties in achieving cost-effective programs (Lorenzen, 2008).

Preliminary economic analyses generally predict that high costs of hatchery production and

inflexible social implications of hatchery production often preclude economic success (Hilborn,

1998; Walters and Martell, 2004; Bell et al., 2006; Lorenzen, 2008). Optimizing cost-

effectiveness of stocking programs by improving post-release survival (Leber et al., 2005;

Brennan et al., 2006) are positive steps toward this, and creative methods for production may

also be necessary to make stock enhancement programs feasible (e.g. Japanese scallops

Patinopecten yessoensis are produced in hanging cages in situ, bypassing the need for hatcheries

[Uki, 2006]). As with all forms of fisheries management programs, challenges lie in how stock

enhancement programs can adjust to adaptive management measures based on dynamic feedback

from biological, economic and social sources (Walters, 1997; Walters and Martell, 2004; Bell et

al., 2006; Lorenzen, 2008).

The common snook Centropomus undecimalis, is one of the focal species of stock

enhancement research aimed at supplementing Florida' s existing stock management programs.

Common snook are an estuarine-dependent catadromous fish found in tropical and subtropical

habitats of the eastern coasts of the Americas, including Florida. Along with substantial

alteration of snook habitat (Bruger and Haddad, 1986), the increasing popularity of the

recreational fishery has caused concern for managers and stakeholders. Florida' s snook










Table 4-5. Details of actual cannibalism occurrences found in snook stomachs collected in
the field.


Source


Sample
date
December-03
November-03
March-03
January-03
January-03
December-03
February-04


Cannibalistic snook

length age
278 SL age-1
313 SL age-1
312 SL age-1
591 SL age-2+
610 SL age-2+
651 SL age-2+
145 FL age-0


Cannibalized Prey

Length Wt. (g) % pred. length
48 0.7 17.3
77 3.4 24.6
113 11.5 36.2
117 12.9 19.8
121 14.4 19.8
141 23.3 21.7
64 1.5 50.0


Adams and Wolfe 2006
Adams and Wolfe 2006
Adams and Wolfe 2006
Adams and Wolfe 2006
Adams and Wolfe 2006
Adams and Wolfe 2006
Brennan




















































Figure 3-1.


Map of experimental study sites along the coasts of Sarasota and Manatee
Counties, Florida. Dots represent locations where collections with the
standardized seine occurred.









when they can be found in high concentrations in coastal passes and along beaches during

summer months (Taylor et al., 1998). Two year old snook have also been documented with

mature gonads (Taylor et al., 1998) and snook 150-400 mm SL (typical age two sizes) have been

documented to move to seagrass habitats during spring and early summer (Gilmore et al., 1983;

Brennan, unpublished data). Therefore I used data from annual collections at beaches (barrier

island coastlines along the Gulf of Mexico), barrier island passes, and intercoastal seagrass

habitats. Sampling at beaches, passes and seagrass habitats was conducted with seines (137 m

long x 2.4 m deep, with 5 cm multifilament nylon mesh), cast nets (4 m radius, 5 cm

multifilament mesh) and trammel nets (137 m long x 2.4 m deep, 3 cm inner nylon mesh, and 30

cm outer nylon mesh). We also performed standardized sampling efforts at randomly selected

coastal sites (mainland and barrier islands, about 1.6 kilometers of shoreline).

In September 2001 and July 2005, red tide outbreaks caused mass mortality of snook

along beaches and passes. During these events, we scanned individual carcasses for CWTs. I

also used data from snook captured in an annual angling tournament specifically created and

designed to evaluate contribution rates of hatchery-released snook to localized adult stocks.

These tournaments were conducted from 1998- 2005, and anglers were directed to fish between

Anna Maria Island and Venice Inlet (Figure 2-1, Table 2-5). Any snook, captured by an angler,

was measured (FL mm) and scanned for CWTs. Tagged snook were kept for CWT extraction,

untagged snook were released alive.

For analysis of long-term data, I removed migrational bias from recapture ratio

comparisons by only using recaptures from beaches and passes (adult snook spawning habitat)

and those captured along shoreline habitat at least 1 km away from any given release site

(because sampling at locations near release sites could result in higher recapture ratios of snook









instantaneous natural mortality rates (Muw) in a typical population growth function, were used to

calculate abundance of the age group i (yages) Over time (t, number of days/365):

yage-lt = age-l, to e jh' t) (4-1)

Weight-specific instantaneous natural mortality rates extrapolated from Lorenzen (1996) were

calculated as a power function of the initial mean weight (W) of individuals in the age-i group on

November 1:

Ms = Mr Wb(4-2)

where, Mtu is the annual natural mortality at unit weight and b is a scaling parameter (Lorenzen,

1996). Because age-1 and age-2+ age groups of snook are typically not fully recruited to the

tidal creeks (winter habitat) in November, I used the relative abundance in January (typical peak

abundance for all age groups in the creeks) from samples collected in tidal creeks from 1998 -

2005 (Figure 4-4) to calculate the older age group's abundance based on the age-0 January

projected abundances. Then using equations 4-1 and 4-2 (and weight-based instantaneous

mortality rates of these age groups [Lorenzen, 1996]) I back calculated the age-1 and age-2+

abundance to November 1 (t-0) to obtain Page-7,t0. During the simulation period, the model

calculated daily abundance estimates for each age group that were weighted according age-

specific residence patterns (p ,t = probability of Page-1,t in the creek at time t) across the time

period (see below, Table 4-2, Figure 4-3).

For the model period, I assumed cannibalism mortality was proportional to natural

mortality of the age-0 cohort. I used field data from snook collections that occurred from 11

November 14 May (i.e. winter and spring data from Adams and Wolfe, [2006], and Brennan;

see Table 4-1) to obtain an estimate of the proportion of the creek population cannibalizing

snook (31,, i.e. number of age i snook observed with cannibalized prey in their stomachs/total









the field and actual rates are probably under-represented. In a laboratory study, casual visual

examination of differently sized soft-bodied fish prey (snook and Gobiidae) in the stomachs of

snook held in aquaria over time have shown these prey were generally identifiable to species

until about 70% of original prey mass remained (Heinlein and Brennan, unpublished data). In

the same study, gobies weighing 2g (about 50 mm TL) at consumption were approximately 77%

their consumption weight after 4h in the stomach, but 0.5g (about 30 mm TL) gobies were

unidentifiable after 4h (17% 24% their original weight remained) (Heinlein and Brennan,

unpublished data).

Diet samples are also known to over-represent large-bodied prey items (Mann and Orr,

1969; Gannon, 1976). In the laboratory, 15g snook prey (mean size=122 mm FL) were more

than 93% their original weights 4h after consumption, and only 50% digested by 18 h after

consumption by larger snook (Heinlein and Brennan, unpublished data). These data align with

predictions that larger-bodied prey are comparatively over-represented in diet studies (Mann and

Orr, 1969; Gannon, 1976) and thus reported cannibalism rates of larger prey are probably closer

to actual rates. Snook also demonstrate peak feeding activity during crepuscular and nocturnal

periods (Rock et al., unpublished data), and diet samples collected during daylight often contain

highly digested prey items that are indistinguishable. All studies reported here sampled snook

stomachs during daylight hours. Crepuscular and nocturnal sampling may have revealed higher

rates of cannibalism. Further study could identify the relative influence of these factors on

evacuation rates and cannibalism rates for common snook.

In conclusion, juvenile snook show many morphological adaptations for predation

defense such as elongated anal spines, proportionally deeper body height, and self mimicry in

dorsal fin spot, and also show greater cannibalism potential with proportionally wider mouth















12 14-0.8
-0.7
9 421-0.

1~ 0.5


-0.4 ~
ICPUE
-m- GROVVTH 0.3 eb

-0.2 ,


-0.1O


South North Hudson Bowlees Tidy Barrier
Creek Creek Bayou Creek Peninsula Islands


Catch per unit effort (CPUE) of wild and hatchery snook less than 300 mm fork
length according to release habitat and associated mean growth rates (mm/d)
through the first winter after release. Error bars are 95% confidence intervals.
Numbers above bars are the numbers of snook recaptured.


1


0.5


0 -






Figure 2-12.










km from their release sites (Table 2-9). For snook released in BC, PC, NC, and SC, 30.5%

(n=262) were recaptured at other release sites within the first year after release, but no clear

patterns of upstream or downstream movement were apparent. For the same period, only nine

(of 262 recaptured, 3.4%) were captured outside of their respective creek within the first year

after release. For the barrier island release sites, only one recapture (of 16 total, 6.3%) was

found away from its release site. The farthest migration documented within the first year after

release was from a fish released at TIDY that emigrated to the upstream section of NC, 24 km

away .

Generally, first winter mean growth rates showed strong patterns of differentiation

according to release site (see below for statistical analysis). Highest mean growth rates were

found in snook released at barrier island sites (0.67-0.78 mm/d) followed by BC (0.67 mm/d),

TIDY (0.63 mm/d), HB (0.52 mm/d), NC (0.44 mm/d), SC (0.44 mm/d), and PC (0.41 mm/d)

(Figure 2-6). Sizes-at-recapture during the first winter showed similar patterns with largest sizes

from snook released at islands, TIDY, and BC, and smallest sizes at recapture from NC and SC

(Figure 2-7). There was no apparent relationship between SAR and first winter growth rates,

however (Figure 2-8, and see below).

Notable differences were found in patterns of snook size structures among the release

sites and sub-sites throughout the seasons (Figure 2-9). Juvenile snook were common

throughout the year in upstream, mid-stream, and stream mouth sub-habitats of tidal creeks, but

rare at barrier islands except during summer and fall months (170-300 mm FL, Figure 2-9).

Large snook (>3 50 mm FL) predominated frequency distributions at barrier islands throughout

the year, but were most abundant here during summer and especially fall months. During winter

and spring, relative abundance of large snook increased in the creeks particularly at stream










Burke, B.J., Rice, J.A., 2002. A linked foraging and bioenergetics model for southern flounder.
Trans. Am. Fish. Soc. 131 (1), 120-131.

Bystriim, P., Persson, L., Wahlstrom, E., Westman, E., 2003. Size- and density-dependent
habitat use in predators: consequences for habitat shifts in young fish. J. Anim. Ecol. 72
(1), 156-168.

Bystriim, P., Andersson, J., Persson, L., De Roos, A.M., 2004. Size-dependent resource
limitation and foraging-predation risk trade-offs: growth and habitat use in young arctic
char. Oikos 104, 109-121.

Bystriim, P., 2006. Recruitment pulses induce cannibalistic giants in Arctic char. J. Anim. Ecol.
75 (2), 434-444.

Campbell, P., 1997. Population projections: States 1995-2025. U.S. Census Bureau.
http://www. census. gov/prod/2/pop/p2 5/p2 5- 11 3 1.pdf.

Carline, R.F., 2006. Regulation of an unexploited brown trout population in Spruce creek,
Pennsylvania. Trans. Am. Fish. Soc., 135 (4), 943-954.

Carr, S.D., Hench, J.L., Luettich, R.A., Forward, R.B., Tankersly, R.A., 2005. Spatial patterns
in the ovigerous Callinectes sapidus spawning migration: results from a coupled
behavioral-physical model. Mar. Ecol. Prog. Ser. 294, 213-226.

Carscadden, J.E., Frank, K.T., Leggett, W.C., 2001. Ecosystem changes and the effects on
capelin (Mallotus villosus), a major forage species. Can. J. Fish. Aquat. Sci. 58 (1), 73-
85.

Caselle, J.E., Wamner, R.R., 1996. Variability in recruitment of coral reef fishes: The importance
of habitat at two spatial scales. Ecology 77 (8), 2488-2504.

Chambers, R.C., Leggett, W.C., 1996. Maternal influences on variation in egg sizes in temperate
marine fishes. Am. Zool. 36, 180-196.

Chang, E. Y. Y., Leo, I.C., 2003. Sibling cannibalism of young red drum, Sciaenops ocellatus,
in relation to size disparity and metabolic rates. Eny. Biol. Fish. 68 (4), 407-415.

Chavez, H., 1963. Contribucion al conocimiento de la biologia de los robalos, chucumite y
constantino (Centropomus spp.) del estado de Veracruz (pisc., Centrop.). Ciencia, Mex.
22, 141-161.

Chen, Y., Jackson, D.A., Harvey, H.H., 1992. A comparison of von Bertalanffy and polynomial
functions in modeling fish growth data. Can. J. Fish. Aquat. Sci. 49, 1228-1235.

Chipps, S.R., Wahl, D.H., 2008. Bioenergetics modeling in the 21st century: Reviewing new
insights and revisiting old constraints. Trans. Am. Fish. Soc. 137 (1), 298-313.










Lang, G.M., Brodeur, R.D., Napp, J.M., Schabetsberger, R., 2000. Variation in groundfish
predation on juvenile walleye pollock relative to hydrographic structure near the Pribilof
Islands, Alaska. ICES J. Mar. Sci. 57 (2), 265-271.

Lau, S.R., Shafland, P.L., 1982. Larval development of snook, Centropomus undecimalis
(Bloch) (Pices: Centropomidae). Copeia 1982, 618-627.

Leber, K.M., 1995. Signifieance of Eish size at release on enhancement of striped mullet Hisheries
in Hawaii. J. World Aquacult. Soc. 26, 143 153.

Leber, K.M., Brennan, N.P., Arce, S.M., 1998. Recruitment patterns of cultured juvenile Pacific
threadfin, Polydactylus sexfilis (Polynemidae), released along sandy marine shores in
Hawaii. Bull. Mar. Sci. 62 (2), 389-408.

Leber, K.M., 1999. Rationale for an experimental approach to stock enhancement. In Howell,
B.R., Moksness, E., Svisand, T., (Eds.), Stock Enhancement and Sea Ranching (Ed. B.R.
Howell, E. Moksness, and T.), Blackwell Scientifie Publications, Oxford, UK, pp. 63-75.

Leber, K.M., 2004. Marine Stock Enhancement in the USA: Status, trends and needs. In: Leber,
K.M., Kitada, S., Blankenship, H.L., Svsand, T. (Eds.), Stock enhancement and sea
ranching. Blackwell Publishing, Oxford, UK, pp. 11-24.

Leber, K.M., Brennan, N.P., Arce, S.M., 1995. Marine enhancement with striped mullet: are
hatchery releases replenishing or displacing wild stocks? In: Schramm Jr., H.L., Piper,
R.G. (Eds.), Uses and effects of cultured fishes in aquatic ecosystems. American
Fisheries Society Symposium 15. Bethesda, MD, pp.376-387.

Leber, K.M., Arce, S.M., Sterritt, D.A., Brennan, N.P., 1996. Marine stock enhancement
potential in nursery habitats of striped mullet, Meugil cephahts, in Hawaii. Fish. Bull. 94,
452-471.

Leber, K.M, Blankenship, H.L., Arce, S.M., Brennan, N.P., 1997. The influence of release
season on size-dependent survival of striped mullet, Meugil cephahts, in a Hawaiian
estuary. Fish. Bull. 95 (2), 267-279.

Leber, K.M., Kitada, S., Blankenship, L., Svisand, T., 2004. Stock Enhancement and Sea
Ranching: Developments, Pitfalls and Opportunities. Blackwell Publishing, Oxford,
United Kingdom.

Leber, K.M., Cantrell, R.N., Leung, P.S., 2005. Optimizing cost-effectiveness of size at release
in stock enhancement programs. N. Am. J. Fish. Manage. 25, 1596-1608.

Lenore, C., Greenberg, A., Eaton, A., 1998. Standard methods for the examination of water and
wastewater. 1998. Section 2540. American Public Health Association, American
Waterworks Association, Water Environment Federation. United Book Press, Inc.,
Baltimore, Maryland.









This study, conducted in Sarasota and Manatee counties of southwestern Florida, USA,

explores the influence of settlement habitat (via releasing hatchery-reared juvenile snook into

various habitats) on growth, survival, and dispersal and recruitment to adult demes of common

snook Centropomus undecimalis. Common snook (herein referred to as "snook") are a tropical

to sub-tropical estuarine-dependent piscivorous fish found along the Atlantic Coast of the

Americas. In Florida, snook stocks are subj ected to intense recreational fishing pressure (Muller

and Taylor, 2005) and have suffered dramatic coastal habitat alteration (Bruger and Haddad,

1986). Snook are also sensitive to cold waters (Storey and Guder 1936; Marshall 1958; Shafland

and Foote 1983) and Florida represents the northern limit of their latitudinal range (Marshall,

1958). As a result, Florida's snook populations undergo annual temperature-dependent range

expansions and contractions. During mild winters, snook larval are dispersed north of the 29th

parallel and successfully survive (e.g. Linton and Richards, 1965), but during harsher winters,

cold water (<14 o C) generally causes snook mortality north of this (Storey and Gruder, 1936;

Shafland and Foote, 1983). In Sarasota and Manatee counties, Florida (latitude range = 270 7' to

270 28' N) snook sometimes suffer mortality during cold winters, but local populations find

thermal refuge in warmer tidal creeks (Marcinkiewicz, 2007; N. Brennan, unpublished data).

Snook spawn in high salinity (>28 %o) seawater in inlets and tidal passes of estuaries, mouths of

rivers and canals, and off sandy beaches primarily from May August (Marshal, 1958; Volpe,

1959; Tucker and Campbell, 1988; Taylor et al., 1998). Hatching occurs within 17-18h after

spawning (Lau and Shafland, 1982; Tucker, 2003), and within two weeks recruitment to

nearshore habitats is thought to occur by larvae taking advantage of a two-layered estuarine

circulation system (Norcross and Shaw, 1984; Tolley et al., 1987) and vertical positioning within

the water column (Peters et al., 1998b). Post-larval snook are found mostly along vegetated

















(n=90)


3.5

3

2.5

2

1 5

1

0.5

0


(n=84)


(n= 52)


(n= 15)






Barrier Islands


Upstream


Midstream


Stream Mouth


Figure 2-3.


Mean recapture ratios (number recaptured/number released x100) of snook from
1998 and 1999 releases, by sub-subsystems. Means are calculated from replicate
release years, stream systems and release lots. Data only include recaptures
within one year after release. Numbers in brackets are the numbers of individual
snook recaptures for each particular release microhabitat. Error bars are 95%
confidence limits of the means.










mouth habitats compared to in-creek abundance during summer and fall. Among creeks, length

frequencies of snook were very similar with respect to sub-habitats.

Multivariate analyses with short-term responses: The principle component analysis of

release site variables and short-term post-release responses showed that the first two principle

components explained 27% and 21% of the standardized variance and 67% was explained by the

first three principle components. A plot of the first two principle components (Figure 2-10)

shows that recapture ratios and age-1 densities were inversely related to growth rates at first

winter (Figure 2-10, Figure 2-11, Figure 2-12). High CPUE and recapture ratios were positively

related to sediment softness and moisture content but negatively related to growth, salinity, pH,

dissolved oxygen, and mangrove coverage (Figure 2-10, Figure 2-11). Release sites with high

CPUE and recapture ratios, but low growth included NCU, NCM, and NCL. Release sites with

low CPUE, low recapture ratios, but high growth, salinity, dissolved oxygen and mangrove

coverage included SCA, NCA, TIDYI, and TIDY. Among our release sites, bank slopes

(steepness) were inversely related to vegetative coverage, sediment organic, and temperature.

Considering bank steepness and vegetative coverage, BCU had steep banks and low vegetative

coverage, but PCL, PCA, and NCL had more gradual bank slopes and more vegetative shoreline

coverage.

The canonical correspondence analysis showed that mean SAR according to release site,

was most correlated to summer (0.77) and winter (0.73) recapture lengths, but not well correlated

with mean dispersal rates during summer (0.33) or winter (0.28), or first winter growth (0.08), or

recapture ratios (-0.33). A plot of the canonical variables for SAR and the response variables

(Figure 2-13) reflects the strongest correlations between SAR and recapture lengths but much

less correlation between the other responses. While canonical variables were not calculated for










Kennedy, T.B., Vinyard, G.L., 1997. Drift ecology of western catostomid larvae with emphasis
on Warner suckers, Catostomus warnerensis (Teleostei). Eny. Biol. Fish. 49 (2), 187-
195.

Kennedy, S.B., Tucker, J.W., Neidig, C.L., Vermeer, G.K., Cooper, V.R., Jarrell, J.L., Sennett,
D.G., 1998. Bacterial management strategies for stock enhancement of warmwater
marine fish: a case study with common snook (Centropomus undecimalis). Bull. Mar.
Sci. 62 (2), 573-588.

Kitada, S., 1999. Effectiveness of Japan's stock enhancement programmes: current perspectives.
In Howell, B.R., Moksness, E., Svisand, T., (Eds.), Stock Enhancement and Sea
Ranching (Ed. B.R. Howell, E. Moksness, and T.), Blackwell Scientific Publications,
Oxford, UK, pp. 103-131.


Kitchell, J.F., Stewart, D.J., Weininger, D., 1977. Applications of a bioenergetics model to
yellow perch (Perca flavescens) and walleye (Stizostedion vitreum vitreum). J. Fish. Res.
Board Can. 34: 1922-1934.

Kocovsky, P.M., Carline, R.F., 2001. Dynamics of the unexploited walleye population of
Pymatuning Sanctuary, Pennsylvania, 1997-1998. N. Am. J. Fish. Manage. 21 (1), 178-
187.

Koczaja, C., McCall, L., Fitch, E., Glorioso, B., Hanna, C., Kyzar, J., Niemiller, M., Spiess, J.,
Tolley, A., Wyckoff, R., Mullen, D., 2005. Size-specific habitat segregation and
intraspecific interactions in banded sculpin (Cottus carolinae). Southeast Nat. 4 (2), 207-
218.

Kostow, K.E., Zhou, S.J., 2006. The effect of an introduced summer steelhead hatchery stock on
the productivity of a wild winter steelhead population. Trans. Am. Fish. Soc. 35 (3), 825-
841.

Krause, J., Loader, S.P., McDermott, J., Ruxton, G.D., 1998. Refuge use by fish as a function of
body length-related metabolic expenditure and predation risks. Proc. Roy. Soc. Lond.
Ser. B: Biol. Sci. 265 (1413), 2373-2379.

Kristiansen T.S., Svisand, T., 1990. Enhancement studies of coastal cod in western Norway.
Part III. Interrelationships between reared and indigenous cod in a nearly land-locked
fjord. J. Con. Int. Explor. Mer 47, 23-29.

Krivan ,V., Vrkoc, I., 2000. Patch choice under predation hazard. Theor. Popul. Biol. 58 (4),
329-340.

Landeau, L., Terborgh, J., 1986. Oddity and the 'confusion effect' in predation. Anim. Behav.,
34, 1372-1380.











Bohlin, T., Sundstrom, L.F., Johnsson, J.I., Hojesjo, J., Pettersson, J., 2002. Density-dependent
growth in brown trout: effects of introducing wild and hatchery fish. J. Anim. Ecol., 71
(4), 683-692.

Botsford, L.W., Castilla, J.C., Peterson, C.H., 1997. The management of Eisheries and marine
ecosystems. Science 277 (5325), 509-515.

Bowen, S.H., Allanson, B.R., 1982. Behavioral and trophic plasticity of juvenile Tilapia
mossamica in utilization of the unstable littoral habitat. Eny. Biol. Fish. 7, 357-362.

Bravington, M.V., Ward, R.D., 2004. Microsatellite DNA markers: evaluating their potential for
estimating the proportion of hatchery-reared offspring in a stock enhancement
programme. Mol. Ecol. 13 (5), 1287-1297.

Brennan, N.P., 1997. Establishment and management of a multi-state, coded-wire tag,
paddlefish database. Masters thesis. Cooperative Fisheries Research Unit. Tennessee
Technological University, Cookeville, Tennessee. 38505.

Brennan, N.P., Leber, K.M., Blankenship, H.L., Ransier, J.M., DeBruler Jr., R., 2005. An
evaluation of coded wire and elastomer tag performance in juvenile common snook under
Hield and laboratory conditions. N. Am. J. Fish. Manage. 25, 437-445.

Brennan, N.P., Darcy, M.C., Leber, K.M., 2006. Predator-free enclosures improve post-release
survival of stocked common snook. J. Exp. Mar. Biol. Ecol. 335 (2), 302-311.

Brennan, N.P., Leber, K.M., Blackburn, B.R., 2007. Use of coded-wire and visible implant
elastomer tags for marine stock enhancement with juvenile red snapper Lutjanus
canspechanus. Fish. Res. 83 (1), 90-97.

Brennan, N.P., Walters, C.J., Leber, K.M., 2008. Manipulations of stocking magnitude:
addressing density-dependence in a juvenile cohort of common snook Centroponaus
undecintalis. Rev. Fish. Sci. 16 (1-3), 215-227.

Brooke, M.D., 1998. Ecological factors influencing the occurrence of 'flash marks' in wading
birds. Funct. Ecol., 12 (3), 339-346.

Bruger, G.E., 1980. Preliminary analyses of snook, Centroponaus undecimalis, population
dynamics in the Naples-Marco Island region of southwest Florida. Fla. Sci. 43 (suppl. 1),
25.

Bruger, G.E., Haddad, K.D., 1986. Management of tarpon, bonefish, and snook in Florida.
Stroud, R.H. (Ed.). Multi-jurisdictional management of marine fishes. National
Coalition for Marine Conservation, Inc., Savannah, Georgia, 53-57.









efficiency and extrapolating mean CPUE to creek-wide shoreline distance I estimated in-creek

juvenile populations to be 375 juveniles in BC, 476 in WB, 2018 in NC, and 639 in SC (Table

3-1).

Tagging and Release of Hatchery-Reared Snook

Hatchery-reared snook ranged from 84 270 mm FL at the time of tagging (mean length

= 177.5 mm FL +/- 2.95 SE) and samples collected in April 2002 of wild juvenile snook of the

same age were 79-219 mm FL (mean= 155 +/- 2.43 SE). We tagged and released a total of

2,372 hatchery-reared juvenile snook (Table 3), and six days after tagging, CWT retention in the

snook averaged 99.5% (from 14 groups with an average of 84 fish / group), and VIE retention

was 100%. Total transport time to the release sites was ca. 2 5 hours for all release groups and

releases occurred between 1230 and 1900 hrs.

Post-Release Evaluation

Overall, from 1 June 2002 30 March 2003, we performed 556 standard seine hauls,

captured 3261 snook and captured, tagged, and released 1588 juvenile snook (Table 3-2). Of

these, 275 were hatchery-reared snook recaptures, and 155 were wild snook recaptures. I found

little evidence of inter-creek movement; of 155 color-coded wild snook recaptures, only one was

recaptured in a creek other than its release creek (about 0.6% found in other creeks) (Table 3-2).

Because all hatchery-reared snook were originally tagged and released with a common VIE

color, I were not able to obtain estimates of inter-creek migration from first-time recaptures.

Only 11 hatchery-reared snook were recaptured twice, and all of these were captured in the same

creek as the first recapture event.

Samples from June 2002 resulted in a total of 410 captured snook, 294 of which were

juveniles. Relative abundance of wild juvenile snook in June was 0.86x, 1.87x, 0.63x, and 2.45x

the April abundance for BC, WB, NC and SC, respectively (Figure 3-4). However, when















9


8


7


6



4


2
1y


200%





150%





100%





50%


Bowlees Whitaker North South
Creek Bayou Creek Creek


Bowlees Whitaker North South
Creek Bayou Creek Creek


Figure 3-5.


Results from post-release samples collected in June 2002 showing catch-per-unit
effort (graph a) and relative abundance (graph b) of hatchery snook (white bars)
to wild juvenile (grey bars) abundance (set at 100%).














-25 A


-20


-15 ~


-10


-5


-0




0


12000

10000

8000

6000

4000

2000

0










12000100 I


V.


O T~


1997 1998 1999


- LL
_

6 d, -


-0.075


-0.05


-0.025 1


0


6000 t


- oClc


4000

2000

0


1997 1998 1~1999


Figure 2-2.


Release numbers (grey bars) and post-release results showing percent hatchery
snook found in total snook catch (graph A) and recapture ratios (number
recaptured/number released) adjusted for effort (from 21 m seine, graph B). The
chronological representation on the horizontal axis is not to scale.










Morphological characteristics (Andraso, 1997; Sillett and Foster, 2000; Magnhagen and

Heibo, 2004; Palma and Stenek, 2001) and behavioral traits (e.g., schooling or ontogenetic

habitat shifts) can affect predation risk (Gallego and Heath, 1994; Wennhage, 2000). Many of

the traits are adaptations that minimize predation risk, such as body depth, spines, or coloration,

and can vary allometrically. Understanding how these traits change with body size can give

insights into life stages most vulnerable to cannibalism and predation.

Cannibalism is common to many species and may provide nutritional and energetic

benefits, reduced competition for food and refuge elimination of potential conspecific predators,

and reduction in foraging risk for juveniles (see Claesson et al., 2004 for review). Cannibalism

risks include choking and suffocating (commonly evidenced in hatcheries), bodily harm from

handling of prey, incorporation of conspecific pathogens, increased predation risk by reducing

school size, and reduction in inclusive fitness from consuming kin. Nonetheless, the prevalence

of cannibalism in many organisms suggests a net evolutionary advantage.

In Florida, aspects of stock abundance, productivity and regulating mechanisms have

been of particular concern for fishery managers of the common snook Centropomus undecimalis

("snook" herein). A coastal, warm-water species, inhabiting estuaries and river systems

(Marshall, 1958; Fore and Schmidt, 1973), snook are subjected to intense recreational fishing

pressure in Florida; stocks are considered over-fished, and recruitment has declined over the past

20 years (Muller and Taylor, 2005). Widespread habitat loss (Bruger and Haddad, 1986), winter

freezes (Storey and Guder, 1936; Marshall, 1958), and red tide blooms also threaten snook

populations in Florida. These factors have led fishery managers to investigate the feasibility of

using stock enhancement to augment snook populations in Florida (Brennan et al., 2005;

Brennan et al., 2006). Despite the potential benefits, stock enhancement can have deleterious































































Schematic of model process predicting cannibalism-related mortality by various

age groups of snook. Shaded boxes represent model components where simulated
variability occurred.


Field data

Length frequency data

s mlin i ust
habitats in November.


Length frequency data and
CPUE from standardized _
sampling in nurses
habitats in Januat

Marine species




Alark-recapture data for
age snook in and around


(Chapter 2 and 3, and
Brennan, unpublished data)

Number of age, snook
stomachs examined

Number of
cannibalism ocurrences


i I __ ^ I


1 un1 1 1ru1 n 111 l1 11

.I l. rs.. 1.r111 iI i a 1 11


I I


Mean weight of bnac fae
age,, cohort in November cor dn Nove r
(n =5000)


Relative abundance
of age-1 or age-2+
Snook in November
(Eqn. 4-4)





Population growth
function
(Eqn. 4-1)m


Number of snook
cannibalized by
age, cohort


Proportion of Proportion of

goertgna age,dmortality
cannibalized cannibalism
by age snook by age, snook


Repeat with 10,000
independent simulations


I


Figure 4-3.











Table 4-2. Numbers of snook processed for stomach contents and observed instances of
cannibalism from 2003 and 2004 winter data in nursery habitats (i.e. from Adams
and Wolfe, 2006, and Sarasota data, 2004). Totals and percent are also shown
according to age group.
Age group Stomach examination Nov.-Dec. February March April May Total Percent
age-0 Number examined 7 160 502 74 96 838 42.13
Number of cannibals 0 1 0 0 0 1
age-1 Number examined 135 110 320 50 62 678 34.07
Number of cannibals 2 0 1 0 0 3
age-2+ Number examined 19 87 257 61 50 474 23.80
Number of cannibals 3 0 0 0 0 3
Total 161 357 1,079 185 208 1,990 100.00









Increased replication may have reduced variation in the study and results should be verified with

examinations of stomach contents and the use of controls.

As a further consideration of snook cannibalism potential, I also examined published and

unpublished data to measure rates of cannibalism detected in wild snook of different life stages.

All cases of cannibalism found were directed at young-of-year snook. These rates were used in a

model predicting cannibalism potential in a localized snook stock in a tidal creek habitat through

a winter. As prey items in a fish' s stomach generally represents a single day's or night' s worth

of feeding activity, the collective examination of snook stomach contents, showed the sample

sizes of most studies (hundreds) were grossly insufficient for detecting cannibalism instances in

snook (where thousands needed). Also, because cannibalism is generally directed at very small

snook, it would be extremely difficult to detect this in the Hield because of the high evacuation

rates of small prey. Field based examination of stomach contents are generally biased towards

detecting large prey items, but negatively biased against detecting small prey items (Kennedy,

1969. Ignoring this, and just examining observed cannibalism rates in the field, my model

showed that a population of juveniles consuming just a few conspecifics throughout a season can

collectively and significantly influence a stock size, however. My model (Chapter 4) predicted

that collective cannibalism by the age groups could contribute substantially to the age-0 natural

mortality rates and cannibalism rates were density-dependent. High cannibalism rates were

predicted when strong year classes of age-0 snook were produced, but low during years of low

recruitment of age-0 snook. This work has strong implications for stock enhancement programs

seeking to numerically supplement particular cohorts of snook.

Rigorous experimental monitoring from standardized sampling efforts over time is integral

to understanding basic progress of the stocked organisms and hatchery-wild fish interactions.









enhancement work in Sarasota, are useful for providing necessary input into modeling

enhancement scenarios and feasibility in Florida.









all release sites (because of missing response data), the canonical variables for TIDY, SCL, and

HBL showed strongest correlation to SAR, but snook released in North Creek showed no

associations with SAR (Figure 2-13). Just examining SAR and dispersal distances however,

snook released in the largest size class (and larger size-at-recapture) showed patterns of greater

dispersal distances compared to the smaller size classes (Figure 2-14).

Long-Term Assessment

General: From August 1, 1998 August 31, 2005 sampling at beaches, passes, and

barrier island habitats and sampling at release sites from April 1, 2000 to August 31, 2005

resulted in an additional 18,344 snook captured (Table 2-5). Of these, we collected 2,883 snook

from beaches and passes, 1,038 from random sampling at barrier island sites, 1,453 snook from

red tide collections, 1,063 from local snook tournaments, and 11,907 snook from sampling at

release sites. Of the snook released in 1998 and 1999 (at upstream, midstream, stream mouth,

and along island shorelines), 174 were recaptured after over 1 yr (0.95% of total snook captured).

Most long-term recaptures were captured at their release sites (n=132), but 42 had "dispersed"

(i.e., moved >1 km). Overall, the largest recaptured hatchery snook was 816 mm FL (34 inch

total length) and weighed 5,492 g (12.1 pounds). The longest time at liberty was 2,399 days

(6.57 years).

Long-term Growth: I obtained age estimates of 514 hatchery-reared snook at large for 6

mos. or longer (maximum age 7.39 years). Of these, four release habitats had over 40 recaptured

snook (5-7 age cohorts, n=59 recaptures from BC, and n=249 from NC) and I used these to

generate age-at-length von Bertalanffy growth models (Table 2-8). Fastest growth was

predicted for snook released at barrier island habitats (Ling = 824), followed by snook released in

Bowlees Creek (Ling = 704), South Creek ((Ling = 657), and North Creek (Ling = 552) (Figure 2-

15).









Orth and Maughan, 1982; Conder and Annear, 1987). Regardless of a habitat' s productive

capacity, stocks (such as snook) at low levels, with high fecundity, at the edge of their range, and

exposed to high environmental variation, typically demonstrate highly variable recruitment

(Myers, 2001), and thus can occur at densities below capacity.

This study examines an important assumption about recruitment processes that density-

dependent survival of wild fish may be reduced by the addition of hatchery-reared fish. In

salmonid populations, evidence exists for density-dependent responses in wild stocks due to

stocking hatchery-reared fish (Nickelson et al., 1986; Nickelson, 2003; Kostow and Zhou, 2006).

Density-dependent growth responses were not measured in this study, but other work in the same

creeks (Brennan and Leber, unpublished data), showed suppressed growth in tributaries with

high densities of age-1 snook. I found no evidence of high augmentation treatment effects on

wild conspecifie density, although an effect may have been expressed through initial loss of

hatchery-reared fish in BC where the productive capacity may have been substantially exceeded.

Abundance patterns of wild age-1 snook throughout the study, however, followed similar

patterns in all creeks.

After emigration from the creeks in fall, there was movement back into the creeks of age-

1 wild snook in winter, resulting in abundances slightly higher than those in summer. Given the

low observed rates of inter-creek movement (0.6% mixing between experimental creeks), the

additional abundance in creeks of wild age-1 snook during winter (observed in all creeks) was

probably due to immigration of wild snook from other sources, emphasizing their importance as

thermal refuge habitat. Furthermore, hatchery-reared snook demonstrated the same seasonal

movement pattern and (after initial loss of Esh in one treatment creek) early-summer abundances

were very similar to winter abundance in both treatment and control creeks. Declines in









Table 4-3. Linear regression statistics for various morphometric comparisons
Comparison N slope L95 U95 R2
Ln(VA-AN) Ln(FL) 31 1.001 0.957 1.045 0.987
Ln(SL) Ln(FL) 94 1.000 0.997 1.002 0.999
Ln(BD) Ln(SL) 94 0.901 0.892 0.915 0.998
Ln(VG) Ln(SL) 135 0.815 0.792 0.831 0.979
Ln(AS) Ln(SL) 332 0.664 0.646 0.682 0.940
Note: vertical gape (VG), body depth (BD, anal spine (AS), throat to anus (TRH-AN), standard
length (SL), and fork length (FL). Prior to regression, all values were loge transformed (Ln)
Lower (L95) and upper (U95) 95% confidence limits of the slopes are given.










4-1 Summary of studies examined regarding common snook diets............... .................13

4-2 Numbers of snook processed for stomach contents and observed instances of
cannibalism from winter data in nursery habitats (i.e. from Adams and Wolfe, 2006,
and Sarasota data, 2004) ........._.._.._ ....__. ...............136..

4-3 Linear regression statistics for various morphometric comparisons: vertical gape
(VG), body depth (BD, anal spine (AS), throat to anus (TRH-AN), standard length
(SL), and fork length (FL). ............. ...............137....

4-4 Number of replicate experimental treatments (N) of field enclosure trials examining
cannibalism according to predator size, the inclusion of additional prey (bait), snook
prey density, and two prey size categories (expressed as a percent of predator size) .....138

4-5 Details of actual cannibalism occurrences found in snook stomachs collected
in the field. ............. ...............139....









Experimental trials were conducted in Sarasota County, Florida, in twenty field

enclosures; ten of these were located in an inland creek estuary (North Creek lagoon), and ten

located along the intra-coastal side of a barrier island (Mote site). Enclosures were 3 m x 3.6 m

rectangular, and 2 m high constructed with 3 mm vinyl mesh in a box shape with four sides and a

bottom sewn together using dental floss (Figure 4-2). Enclosures were attached at the corners to

2.4 m long steel re-bars (20 mm diameter) implanted in the substrate. Tree branches were placed

in each enclosure to provide additional refuge habitat. The tops of the enclosures were covered

with 3 cm square plastic bird mesh.

A single snook predator and selected snook prey were stocked simultaneously into each

enclosure and held for one week. Predator sizes were categorized into "small" (74-163 mm FL),

and "large" (218-332 mm FL) size classes. Snook prey, were approximately 50% (42-56%), and

30% (28-42%) the length of the predator. To test the relative effects of prey density on

cannibalism rates, half of the trials were stocked with one snook prey (about 0. 1 fish per m2), and

half with ten snook prey (about one fish per m2). Within each of the above experimental

combinations, half of the enclosures were stocked with additional prey species (captured nearby).

One week after stocking, animals in the enclosures were harvested by removing the cable

ties from the re-bars and lifting the sides and bottoms out of the water. At harvest, all cages were

carefully searched for live and dead snook. A cannibalism incidence was assumed to have

occurred if prey snook were missing at harvest. Trials where the predator was missing or dead at

harvest were not included in the analysis. No stomach contents were collected. After harvest, a

new set of predator/prey combinations were randomly stocked into the enclosures until all trial

combinations were completed. Trials were conducted in October and November, 2003, prior to

onset of cooler winter temperatures. Cannibalism incidence was recorded in a binary format (yes









Manatee Counties, on the west coast of Florida, USA (Figure 3-1). These creeks are tidally

influenced and water levels typically fluctuate by about 0.75 m daily. Salinities vary with tidal

influence and seasonal rains (wet season during summer). Water temperatures during summer

(June September) typically range from ca 28 oC 34 oC, and winter (December March)

temperatures range from ca 12oC 22oC. Dissolved oxygen levels during summer are low

compared to winter (typically <5 mg/1 versus >7 mg/1). All creeks are partially influenced by

anthropogenic shoreline alterations, and sections of BC, WB, and SC have been dredged to some

extent for maritime navigation within the last 20 years (Table 3-1). North Creek remains

relatively unaltered, although upstream reaches are surrounded by a golf course development and

shoreline residential properties. Densities of wild age-1 snook (~150-300 mm fork length [FL])

during early summer were about 0.5 3.0 fish/30 m of shoreline (Table 3-1).

Experimental Design and Sampling Methods

I defined "age-1 snook" or "juvenile snook" as offspring from the 2001 spawning year.

By April 2002, these snook were approximately 10 (+/-2) months old. Experimental releases of

hatchery-reared juveniles occurred in May, 2002 and our study continued until June 2003 when

the juveniles were about 24 months old. I imposed two levels of manipulation: 1) a large

addition of hatchery-reared juveniles equal to 100% of the wild juvenile abundance (in creeks

BC and WB), and 2) a smaller addition of hatchery-reared fish equal to 10% of the wild stock (in

NC and SC). Quantitative sampling was performed in all creeks: once before releases (to

estimate pre-release age-1 abundance), and six times after releases through the following spring

(approximately in June, August, October, December 2002, and February, June 2003, Table 3-2)

to monitor the effects of the releases on in-creek abundance of juvenile snook.

To estimate in-creek juvenile snook population size, I used a "leap-frog" systematic

sampling design where a standardized bag seine (73 m long x 3 m deep with 1 cm nylon









My weight specific instantaneous natural mortality rate

NC North Creek

N ct total in-creek juvenile snook abundance at time t

PC Phillippi Creek

pco;;;;on the conditional maximum likelihood estimate for binomial sampling for a
all depletion removal trials

p It proportion of age i snook in a given tidal creek at time t

p;;;a the conditional maximum likelihood estimate for binomial sampling for a
particular depletion removal trial

s seined area of a creek

SAR Size-at-release (FL, mm)

SC South Creek

SL standard length (mm)

snook common snook (Centropomus undecimalis)

t time in days

to vonBertalanffy parameter, theoretical time when fish length = 0

TIDY Tidy Peninsula, also known as Tidy Island

TIDYI Barrier island counter part to TIDY

TL total length (mm)

TO

U Upstream habitat

W snook body weight (or mass, in g)

WB Whitaker Bayou

WtC, typical weight of individual prey that age group i would cannibalize

Wt, the weight of an individual in September from cohort i

VA-AN the distance from the ventral gill arch (hypohyal bone) to the anus



































Figure 4-2.


Photograph of enclosures where cannibalism trials occurred. Steel re-bars
supported the material attached at the corners. Bird mesh prevented birds (seen
on first enclosure) from preying on the enclosed snook from above.









Inter-annual variation in water temperatures could change the timing and breadth of peak

abundance, however.

Bioenergetics approaches (e.g. Kitchell et al., 1977; Rice and Cochran, 1984; Rose et al.,

1999; Burke and Rice, 2002) are becoming increasingly popular in fisheries applications (Chips

and Wahl, 2008) and can be useful because they address physiological and activity-level

responses to environmental variation. Nonetheless, Hield data has consistently not well correlated

with model predictions (Ney, 1993; Bajer et al., 2004; Selch and Chips, 2007; Chips and Wahl,

2008). Often, prey availability, predator-prey encounter rates, and feeding rates are primary

sources of uncertainty (Ney, 1993; Walters and Martell, 2004; Bajer et al., 2004; Chips and Wah,

2008), and are difficult to measure in the Hield (Walters and Martell, 2004). With snook, these

and estimates of food conversion efficiencies, activity levels, and physiological responses to

abiotic variation (e.g. temperature, salinity, and oxygen interactive effects on physiological

responses) are not well quantified. Therefore, to minimize unpredictability, in this model I used

a simplified approach that bypassed the need for predicting many of the above bioenergetic

variables. Much of this model was based on data from Hield examinations of snook stomachs

during winter and spring.

While environmental conditions have a strong influence on gastric evacuation rates (e.g.

temperature and salinity, or interactive effects [Vinagre et al., 2007]), the cannibalism model

only focused on Hield collections during winter and spring and thus reduced these effects. Prey

size (mass) has also been shown to be directly related to residence times in digestive tracts

(Jobling et al., 1977; dos Santos and Jobling, 1995; Andersen, 1998). Kennedy (1969) suggested

that small-sized prey items were often underrepresented in diet studies, due to faster gastric

evacuation rates. For snook, this could strongly bias observations of age-0 cannibalism rates in










Mooij, W.M., Van Densen, W.L.T., Lammens, E.H.R.R., 1996. Formation of year-class strength
in the bream population in the shallow eutrophic Lake Tjeukemeer. J. Fish Biol. 48 (1),
30-39.

Muller, R.G., Taylor, R.G., 2005. The 2005 stock assessment update of common snook,
Centropomus undecimalis. Florida Fish and Wildlife Conservation Commission, Florida
Marine Research Institute. Available: www.floridamarine. org

Mumby, P.J., Edwards, A.J., Arias-Gonzalez, J.E., Lindeman, K.C., Blackwell, P.G., Gall, A.,
Gorczynska, M.I., Harborne, A.R., Pescod, C.L., Renken, H., Wabnitz, C.C.C.,
Llewellyn, G., 2004. Mangroves enhance the biomass of coral reef Eish communities in
the Caribbean. Nature 427 (6974), 533-536.

Munro, J.L., Bell, J.D., 1997. Enhancement of marine fisheries resources. Rev. Fish. Sci. 5, 185-
222.

Mushiake, K., Muroga, K., 2004. Fish health management in seed production. In: Leber, K.M.,
Kitada, S., Blankenship, H.L., Svsand, T. (Eds.), Stock enhancement and sea ranching.
Blackwell Publishing, Oxford, UK, pp. 119-129.

Myers, R.A., 2001. Stock and recruitment: generalizations about maximum reproductive rate,
density dependence, and variability using meta-analytic approaches. ICES J. Mar. Sci.
58 (5), 937-951.

Neill, S.R.St. J, Cullen, J.M., 1974. Experiments on whether schooling by their prey affects the
hunting behavior of cephalopods and fish predators. J. Zool. Lond. 172, 549-569.

Nemeth, R.S., 1998. The effect of natural variation in substrate architecture on the survival of
juvenile bicolor damselfish. Eny. Biol. Fish. 53 (2), 129-141.

Neuenfeldt, S., Koster, F.W., 2000. Trophodynamic control on recruitment success in Baltic
cod: the influence of cannibalism. ICES J. Mar. Sci. 57, 300-309.

Ney, J.J., 1993. Bioenergetics modeling today growing pains on the cutting edge. Trans. Am.
Fish. Soc. 122(5), 736-748.

Nickelson, T., 2003. The influence of hatchery coho salmon (Oncorhynchus kisutch) on the
productivity of wild coho salmon populations in Oregon coastal basins. Can. J. Fish.
Aquat. Sci. 60 (9), 1050-1056.

Nickelson, T.E., Solazzi, M.F., Johnson, S.I., 1986. Use of hatchery coho salmon
(Oncorhynchus kisutch) presmolts to rebuild wild populations in Oregon coastal streams.
Can. J. Fish. Aquat. Sci. 43 (12), 2443-2449.

Norcross, B.L., 1984. Oceanic and estuarine transport of Esh eggs and larvae a review. Trans.
Am. Fish. Soc. 113 (2), 153-165.










Tomiyama, T., Watanabe, M., Fujita, T., 2008. Community-based stock enhancement and
fisheries management of the Japanese flounder in Fukushima, Japan. Rev. Fish. Sci. 16
(1-3) 146-153.

Travis, J., Coleman, F., Grimes, C., Conover, D., Bert, T., Tringali, M., 1998. Critically
assessing stock enhancement: an introduction to the Mote symposium. Bull. Mar. Sci. 62
(2), 305-311.

Tringali, M.D., Leber, K.M., Halstead, W.G., McMichael, R., O'Hop, J., Winner, B., Cody, R.,
Young, C., Neidig, C.L., Wolfe, C., Forstchen, A., Barbieri, L., 2008a. Marine stock
enhancement in Florida: a multi-disciplinary, stakeholder-supported, accountability-based
approach. Rev. Fish. Sci. 16 (1-3), 51-57.

Tringali, M.D., Seyoum, S., Wallace, E.M., Higham, M., Taylor, R.G., Trotter, A.A.,
Whittington, J.A., 2008b. Limits to the use of contemporary genetic analyses in
delineating biological populations for restocking and stock enhancement. Rev. Fish. Sci.
16 (1-3), 111-116.

Tsukamoto, K., Kuwada, H., Hirokawa, J., Oya, M., Sekiya, S., Fujimoto, H., Imaizumi, K.,
1989. Size-dependent mortality of red sea bream, Pagrus major, juveniles released with
fluorescent otolith tags in News Bay. J. Fish Biol. 35 (Supp. A), 56-69.

Tsukamoto, K., Masuda, S., Endo, M., Otake, T., 1996. Behavioral characteristic of the ayu,
Plecoglossus altivelis, as predictive indices for stocking effectiveness in rivers. Nippon
Suisan Gakkaishi, 56, 1177-1186.

Tucker, J., 2003. Snook culture. World Aqua. 43 (4), 42-47.

Tucker, J.W., Campbell, S.W., 1988. Spawning season of common snook along the east central
Florida coast. Fla. Sci. 51 (1), 1-6.

Tupper, M. Boutilier, R.G., 1997. Effects of habitat on settlement, growth, predation risk and
survival of a temperate reef Eish. Mar. Ecol. Prog. Ser. 151 (1-3), 225-236.

Uki, N. 2006. Stock enhancement of the Japanese scallop Patinopecten yessoensis in Hokkaido.
Fish. Res. 80, 62-66.

Uzars, D., Plikshs, M., 2000. Cod (Gadus morhua L.) cannibalism in the Central Baltic:
interannual variability and influence of recruit abundance and distribution. ICES J. Mar.
Sci. 57 (2), 324-329.

Victor, B.C., 1983. Recruitment and population dynamics of a coral reef Eish. Science 219,
419-420.

Victor, B.C., 1986. Larval settlement and juvenile mortality in recruitment-limited coral reef
Eish population. Ecol. Monogr. 56, 145 160.









Midstream creek habitats (sub-system=M), are influenced by tidal flux and creek outflow, and

waters were generally less saline. Mangrove, live oak (Quercus virginiana), Juncus sp., and

Brazilian pepper (Schinus terebinthifolius) predominated shoreline vegetation. Water depths in

NC and SC (primarily non-dredged habitat) were generally <1.0 m, and at PC and BC (dredged

habitat) were about 0.5 2.0 m. Substrate at the midstream sites was primarily mud, clay and

sand mixtures. Low salinity waters (generally 0-15 %o) typified upstream tributary habitats (sub-

system = U), but these sites were still influenced by tidal flux. Shoreline vegetation consisted of

some mangrove, but more live oak, leather fern (Acrostichum aureum), Juncus sp., and Brazilian

pepper. Water depths were generally <1 m (except for BC; about 1 2 m deep) and substrate

was primarily mud, sand, and some clay.

Hatchery Fish Production, Tagging and Release

Wild adult brood stock snook were captured annually from May September with large

seines (140-m to 160-m long, 3-m deep, 3-cm multifilament nylon mesh) along barrier island

passes and beaches where snook aggregated for spawning activities each new and full moon.

Snook were held in floating net pens, then individuals were sexed, and eggs and milt were

collected by gently squeezing the abdominal area. Snook eggs were fertilized with milt in 250

ml beakers, transferred to pre-filtered sea water in oxygenated plastic bags, and transported to

2,200 1 recirculating seawater tanks (salinity = 28-32 %o) at the laboratory. Tank salinity was

gradually lowered to 3-6 %o by 30 d after hatching. After this, juvenile snook were sorted by size

on a regular basis to prevent cannibalism and reared until tagging (Main, unpublished data).

Snook were tagged according to release site, release lot (within-year replication over

time), and SAR (small 80-129 mm FL, medium 130-179 mm FL, and large 180-270 mm FL).

Tagging occurred in April 1998 and 1999 (about 8-11 mo. after hatching for each respective

release year). Individual snook were tagged with single-length coded-wire tags (CWT, as in









LIST OF ABBREVIATIONS

a area covered by a standardized depletion net

fl number of snook captured

6 number of snook remaining in a standardized seine sampled area

Yage-l age Specific estimate of the number of snook in a given cohort

31, field cannibalism rate, i.e. observed number of age i snook with
cannibalized prey in their stomachs / total number of age i snook stomachs
examined

AS anal spine length (mm)

b scaling parameter of My function

BC Bowlees Creek

BD body depth (mm), measured at the widest part of the body

c mortality due to cannibalism (number of individuals)

CPUE catch-per-unit effort,

CWT coded-wire tag

Dc density (number per 30 m of shoreline) of juvenile snook in the creeks

D.O. dissolved oxygen (mg/1)

FL fork length (mm)

HB Hudson Bayou

He total creek shoreline distance

I Barrier island habitat

k vonBertalanffy growth coefficient

L Lower stream mouth habitat

Linf vonBertalanffy parameter, asymptotic length

M Midstream habitat

M~u annual natural mortality at unit weight












Tagging and Release of Hatchery-Reared Snook ................. ............... ......... ...96
Post-Release Evaluation .............. ...............96....

Creek Density .............. ...............98....
Discussion ................. ...............98.................


4 CRYPTIC CANNIBALISM IN SIZE-STRUCTURED SNOOK POPULATIONS ...........112


Introducti on ................. ...............112................
M materials and M ethods ................. .............. ... ......... ............11

Generalized Snook Life History in a Tidal Creek ................. ............................116

Approach 1: Phenotypic Characteristics ................. ...............116........... ...
Approach 2: Field Enclosure Trials ................... .. .. ......... ...............117.
Approach 3: Cannibalism Intensity in Wild Snook Populations ................. ...............119
Re sults................... .. ....... ...............123......

Phenotypic Characteristics .............. ...............123....
Field Enclosure Trials................ .......... ... ... ........2

Cannibalism Meta-Analysis and Model Results .............. ...............124....
Discussion ................. ...............126................


5 CONCLU SION................ ..............15


LIST OF REFERENCES ................. ...............159................


BIOGRAPHICAL SKETCH ................. ...............183......... ......













A SHORT-TERM RRATIOS (< 1 YEAR)


o LO NG-TE RM R RATIOS (1 -8 YEARS)


r-2 =0.7228


r2 = 0.9298


1 1.5 2 2.5


Mean Age-1 CPUE in winter
(at release habitats)
Linear regression models using mean catch-per-unit effort during winter of age-1
snook to predict short- (< 1 year post-release) and long-term (>=1 year post-
release) recapture ratios (rrate; number recaptured/number released x 100).


Figure 2-5.









HABITAT- AND DENSITY-MEDIATED INFLUENCES ON SNOOK ECOLOGY: LESSONS
LEARNED FROM MANIPULATIVE RELEASE EXPERIMENTS WITH HATCHERY-
REARED JUVENILE SNOOK




















By

NATHAN PAUL BRENNAN


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

2008










creeks received high augmentation release treatments, and two creeks received low augmentation

treatments. Abundance was measured before and through the first year after the experimental

releases. This study focused on augmentation effects within a cohort of juvenile snook, but

future work should also examine inter-cohort predation and mortality and how inter-annual

natural variation on ambient snook abundance could influence similarly staged release

experiments.

As snook are piscivorous and have shown evidence for density-dependent growth and

mortality, in Chapter 4, I examined the potential influence of cannibalism in controlling juvenile

snook abundance. I examined published and unpublished literature on snook diets, and

supplemented this with my own examinations of Sarasota snook stomach contents. Then I used

these data to provide input into a model I developed to understand the role of cannibalism in

snook abundance dynamics and to provide insight into conditions that promote cannibalism.

Overall, this work provided important insight into understanding the potential and feasibility of

stock enhancement with common snook stocks in Florida, and provides a better understanding of

snook ecology, recruitment patterns, and characteristics of essential fish habitat for snook.









Post-Release Evaluation

Post-release creek-wide standard sampling efforts were performed in each creek six times

after the release (approximately during June, August, October, December 2002, February 2003

and June 2003, Table 2). I identified year classes of wild snook according to length-at-capture

based on (1) length frequency graphs of captured snook, (2) corresponding sizes of cultured

snook, and (3) tag-recapture data on known young-of-year snook. For hatchery-reared snook I

adjusted catch by accounting for expected tag loss over time (e.g., VIE tag loss after 6 months

was 3%, [Brennan et al., 2005]). All captured snook were counted, measured, and checked for

the presence of CWTs with magnetic tag detectors, and visually examined for VIE tags.

Untagged wild snook and first-time recaptured hatchery-reared snook were tagged with VIE tags

in the caudal fins (using a unique color for each creek) and CWTs. For subsequent recaptures,

snook were re-tagged with a VIE of a specific color (BC = orange, WB = green, NC = yellow,

SC = pink) and implant location (caudal or anal fin) to identify the creek in which it was

captured, and the number of times recaptured.

Results

Gear Efficiency

Overall we performed 21 depletion-removal trials. Of these, juvenile snook were

captured in 12 trials and these data were used to generate estimates of seine efficiency. The

maximum likelihood estimate of pcommon value was 0.79 (90% CI = 0.75 0.83, 3 for likelihood

profile for the estimate).

In April 2002, our "leap-frog" sampling regime resulted in 117 standardized seine hauls

performed throughout the four experimental creeks. Our samples represented about 4 10% of

the total shoreline habitat in these creeks (Table 3-1, 3-1). We captured 505 snook and of these

183 were estimated as age-1 juveniles (2001 cohort, ~ 10 months old). After adjusting for seine









ACKNOWLEDGMENTS

I will always be thankful and grateful for my wife, children, my mother and father, and

family members who showed their continued patience, support, and encouragement throughout

this venture. I gratefully acknowledge the contributions of my graduate committee Ken Leber

and Bill Lindberg (co-chairs), Carl Walters, Craig Osenberg, Mike Allen, and Tom Frazer for

their overall direction in experimental design, scientific approach, and editorial assistance.

Throughout, Ken has also provided constant support, direction, and encouragement which

allowed me to pursue independent avenues of research in a larger stock enhancement program.

I thank Carl for his additional assistance with field work, data analysis, and overall direction.

Bill Lindberg also provided extra direction, time, and encouragement during this endeavor.

Mote Marine Laboratory financially supported my graduate work and I commend Mote

administration for supporting the educational pursuits of their employees. Much gratitude is due

to R. DeBruler, D. Wilson, V. Mooney, A. Warner, B. Blackburn, G. Wahl, C. Ondercin, A.

Read, S. Andersen, and many interns and volunteers in the Stock Enhancement Program at Mote

Marine Laboratory. I thank those at Mote' s Center for Aquaculture Research and Development

for rearing snook used in this study. I thank L. Blankenship (Northwest Marine Technology,

Olympia, WA) and staff at the Stock Enhancement Research Facility, Florida Fish and Wildlife

Conservation Commission (FWC) for assistance during tagging and release activities. I thank P.

Hull, D. Doherty, and M. McLeod for assistance during snook spawning trips and snook

releases.

Funding for this work was provided by William Mote for initiating the snook hatchery, the

Florida Fish and Wildlife Conservation Commission, and the National Oceanic and Atmospheric

funded research consortium-the Science Consortium for Ocean Replenishment (SCORE).









abundance of hatchery-reared snook from summer to winter (after the initial high post-release

loss), presumably reflected mortality rates in resident wild age-1 snook.

The use of a Eixed-station and systematic sampling program was appropriate in our study

because it incorporated a representative spectrum of population and habitat size/quality gradients

in the creeks (by attempting to sample every third 30 m section of shoreline habitat, see Hillborn

and Walters, 1992). Increases in abundance from April to June (as seen in WB and SC) may

reflect improvements in the efficiency and sampling ability of Hield workers. Such

improvements in sampling efficiency or catchability with experience are not uncommon (e.g.,

Walters and Maguire, 1996). In our experimental creeks, it is doubtful that abundance of wild

age-1 snook actually increased (via immigration) prior to large-scale emigration of juveniles

from the creeks in the fall. Furthermore, samples from NC and BC showed expected declines in

abundance. Logistical considerations and cost restricted replication and sampling effort for this

study. The low experimental power meant that the probability of a Type-II error was high, yet

repeated measures detected a significant time~treatment interactive effect on population means

and thus treatment-effect size was high. Furthermore, sampling over time showed consistent

patterns in all of the creeks (e.g., abundance declines in fall, and increases in winter), and further

supports our confidence in our estimates of abundance.

Inter-annual variation could have a strong effect on the results of similarly staged release

experiments performed at other times (through variations in carrying capacity, recruitment, and

subsequent competition and predation; see Walters and Martell, 2004; Brennan, unpublished

data). Common snook stocks in Florida have characteristics of high recruitment variability

(Myers, 2001) and, even over a 30 km stretch of coastal habitat, evidence exists for strong

variation in intra-annual recruitment of juveniles (Brennan, unpublished data). As snook are also










M.P., 2001. The identification, conservation, and management of estuarine and marine
nurseries for fish and invertebrates. Bioscience 51 (8), 633-641.

Becker, P., Barringer, C., Marelli, D.C., 2008. Thirty years of sea ranching Manila clams
(Venerupis philippinarunt): Successful techniques and lessons learned. Rev. Fish. Sci. 16
(1-3), 44-50.

Belk, M.C., 1995. Variation in growth and age at maturity in bluegill sunfish genetic or
environmental-effects J. Fi sh Biol. 47 (2), 23 7-247.

Belk, M.C., 1998. Predator-induced delayed maturity in bluegill sunfish (Leponsis nzacrochirus):
Variation among populations. Oecologia 113 (2), 203-209.

Bell, J.D., Bartley, D.M., Lorenzen, K., Loneragan, N.R., 2006. Restocking and stock
enhancement of coastal fisheries: potential, problems and progress. Fish Res. 80 (1), 1-8.

Bell, J.D., Leber, K.M., Blankenship, H.L., Loneragan, N.R., Masuda, R., 2008. A new era for
restocking, stock enhancement and sea ranching of coastal fisheries resources. Rev. Fish.
Sci. 16 (1-3), 1-9.

Bennett, J.P., 2006. Using telemetry to assess natural and fishing mortality of common snook in
Sarasota Bay, Florida. Masters thesis. Fisheries and Aquatic Sciences, University of
Florida.



Berejikian, B.A., Smith, R.J.F., Tezak, E.P., Schroder, S.L., Knudsen, C.M., 1999. Chemical
alarm signals and complex hatchery rearing habitats affect antipredator behavior and
survival of chinook salmon (Oncorhynchus tshan/ytscha) juveniles. Can. J. Fish. Aquat.
Sci. 56 (5), 830-838.

Berkeley, S.A., Chapman, C., Sogard, S.M., 2004. Maternal age as a determinant of larval
growth and survival in a marine fish, Seba~stes nzelan2ops. Ecology 85, 1258-1264.

Blankenship, H.L., 1990. Effects of time and fish size on coded wire tag loss from Chinook and
Coho salmon. In: Parker, N.C., Giorgi, A.E., Heidinger, R.C., Jester Jr., D.B., Prince,
E.D., Winans, G.A. (Eds.), Fish marking techniques. American Fisheries Society
Symposium 7. Bethesda, MD, pp. 237-243.

Blankenship, H.L., Leber, K.M., 1995. A Responsible Approach to Marine Stock Enhancement.
In: Schramm Jr., H.L., Piper, R.G. (Eds.), Uses and effects of cultured fishes in aquatic
ecosystems. American Fisheries Society Symposium 15. Bethesda, MD, pp.376-387.

Blest, A.D., 1957. The function of eyespot patterns in the Lipidotera. Behaviour 1 1, 209-256.

Blewett, D.A., Hensley, R.A., Stevens, P.W., 2006. Feeding habits of common snook,
Centroponaus undecimalis, in Charlotte Harbor, Florida. Gulf Carib. Res. 18, 1-13.













Table 3-2. Mark-recapture results for the four experimental creeks organized according to

sample month
Sample Creek Effort Captured, tagged, and released Recaptures Migrants
Month (No. seines) Wild Hatchery (recaptures) Hatchery 2x Wild Wild 2x N Origin
June Bowlees Creek 27 16 12 0 1 0 0
2002 Whitaker Bayou 16 41 65 0 0 0 0
North Creek 45 91 23 0 8 0 0
South Creek 26 62 4 0 2 0 0
July-August Bowlees Creek 28 22 22 13 0 0
2002 Whitaker Bayou 16 13 12 0 4 0 0
North Creek 46 79 23 113 0 0
South Creek 26 33 2 0 7 0 0
Aug-Sept. Bowlees Creek 28 20 4 2 0 0 0
2002 Whitaker Bayou 16 6 4 0 1 0 0
North Creek 45 124 14 1 5 0 0
South Creek 29 19 10 2 10
Oct. -Dec. Bowlees Creek 26 72 7 0 4 0 0
2002 Whitaker Bayou 16 37 0 0 2 0 0
North Creek 34 317 5 1 14 0 0
South Creek 0 0 0 0 0 0 0
Jan.-March Bowlees Creek 28 4 5 0 7 0 1 Whitaker Bayou
2003 Whitaker Bayou 16 36 45 4 11 0 0
North Creek 46 239 18 1 51 10
South Creek 26 17 8 0 11 0 0
June-July Bowlees Creek 28 14 10 1 0 0
2003 Whitaker Bayou 16 6 0 0 2 0 0
North Creek 45 38 0 0 6 0 0
South Creek 27 7 0 0 0 0 0
1313 275 11 155 2 1

Note: "Hatchery 2x" and "Wild 2x" represent snook that were recaptured in two subsequent

sampling events. Data are only for the juvenile snook cohort (age-1).


























1.5 Mean recapture rate



1-


Intercoastal
Islands


Tidy
Island


Bowlees
Creek


Hudson
Bayou


Phillippi North
Creek Creek


South
Creek


Figure 2-4.


Recapture ratios (number recaptured/number released x 100) of snook by release
habitat. Data are from recaptures collected within the first year after release.
Mean recapture ratio is indicated by the horizontal line.


































Figure 4-6.


Photograph of a 70 mm (TL) and a 178 mm (TL) common snook. Black spot at
tip of dorsal fin is obvious in smaller snook, but spot is nearly absent in larger
snook. Photograph by L. Mitchell.










Seitz, R.D., Lipcius, R.N., Seebo, M.S., 2005. Food availability and growth of the blue crab in
seagrass and unvegetated nurseries of Chesapeake Bay. J. Exp. Mar. Biol. Ecol. 319 (1-
2), 57-68 Sp. Iss. SI

Sekino, M., Saitoh, K., Yamada, T., Yamashita, Y., .2005. Genetic tagging of released Japanese
flounder (Paid~iL hibys olivaceus) based on polymorphic DNA markers. Aquaculture
244(1-4), 49-61.

Selch, T.M., Chipps, S.R., 2007. The cost of capturing prey: measuring largemouth bass
(M~icropterus salmoides) foraging activity using glycolytic enzymes (lactate
dehydrogenase). Can. J. Fish. Aquat. Sci. 64 (12), 1761-1769.

Serfling, S.A., 1998. Breeding and culture of snook, Centropomus undecimalis, in a closed-
cycle, controlled environmental culture system. World Aquaculture Society.
Aquaculture '98 Book of Abstracts. P. 482. February 15-19, 1998. Las Vegas, USA.

Shacklee, J.B., Bentzen, P., 1998. Genetic identification of stocks of marine fish and shellfish.
Bull. Mar. Sci. 62, 589 -621.

Shafland, P.L, Foote, K.J.,1983. A lower lethal temperature for fingerling snook, Centropomus
undecimalis. Northeast Gulf Sci. 6 (2),175-177.

Shimizu, D., Sakiyama, K., Sakakura, Y., Takatani, T., Takahashi, Y., 2007. Predation
differences between wild and hatchery-reared tiger puffer Taxkifugu rubripes juveniles in
a salt pond mesocosm. Nippon Suisan Gakkaishi 73 (3), 461-469.

Sillett, K.B., Foster, S.A., 2000. Ontogenetic niche shifts in two populations of juvenile
threespine stickleback, Gasterosteus aculeatus, that differ in pelvic spine morphology.
Oikos 91 (3), 468-476.

Smith, P.E., 1995. Development of the population biology of the Pacific hake, M~erluccius
products. CA Coop. Ocean. Fish. Invest., 36, 144-152.

Smith, C.C., Fretwell, S.D., 1974. Optimal balance between size and number of offspring. Am.
Nat. 108 (962), 499-506.

Sogard, S.M. 1992. Variability in growth-rates of juvenile fishes in different estuarine habitats.
Mar. Ecol. Prog. Ser. 85 (1-2), 35 -53.

Solazzi, M.F., Nickelson, T.E., Johnson, S.L., 1991. Survival, contribution, and return of
hatchery coho salmon (Oncorhynchus kisutch) released into freshwater, estuarine, and
marine environments. Can. J. Fish. Aquat. Sci. 48, 248-253.

Steele, D. H., Andersen, R., Green, J.M., 1992. The managed commercial annihilation of
northern cod. Newf. Stud., 8 (1), 34-68.










gapes in smaller snook. Field enclosure trials indicated that smaller snook had greater

cannibalism tendencies than larger snook, and increased snook densities increased cannibalism

rates. Snook diet studies to date have shown sparse evidence of cannibalism, however,

cannibalism rates are influenced by density-dependent functions, and an interaction between

foraging arenas (nursery habitats versus adult habitats) and seasonal effects on a population's

movement. All cases of cannibalism recorded have shown that age-0 snook are the target prey

group for intracohort and intercohort cannibalism and this process could have a mediating

influence on recruitment dynamics. This work underlines the necessity for multifaceted annual

recruitment monitoring programs and could contribute to modeling stock enhancement viability

assessments and overall snook management programs.













Table 2-7. Variables used for the principle component analysis examining both short-term

(days-at-large <= 365) and long-term (days-at-large > 365) post-release responses
of hatchery-reared snook stocked in 1998 and 1999 according to release site
character stick s
Variable SCU SC~M SCL SCA NCU NCM NCL NCA PCU PCM PCL PCA HBM HBL HBA BCU BC~M BCL BCA tidy TIA
Salinity (pt) 3 3 26 2 33 7 34 9 6 5 13 7 30 3 32 6 0 3 8 2 30 0 32 1 34 1 33 8 35 8 32 1 33 3 32 9 36 0 334 35 1
Dissolvedoxygen (mg/1) 36 44 51 74 62 58 59 97 71 64 56 52 27 34 71 41 63 58 71 56 85
Temperature (celslus) 28 8 31 2 29 6 30 8 30 8 29 2 32 6 31 9 31 8 31 5 32 5 33 4 28 8 28 5 27 7 30 8 30 7 30 1 30 2 28 8 30 8
pH 79 79 81 83 79 71 80 84 78 79 80 80 81 81 84 82 84 84 83 80 83
Substrate softness (cm) 267 223 153 87 96277 343 10 26 41 13 23 32 10 7 22 63 19 17 12 18
Bank slope (lse/run) 0 25 0 20 0 15 0 30 0 18 0 14 0 09 0 17 0 26 0 21 0 11 0 14 0 34 0 24 0 24 0 31 0 15 0 20 0 38 0 16 0 15
Sediment organics(% 4 78 8 04 632 7 35 4 5 896 11 28 4 13 7 38 4 96 9 15 23 72 8 25 8 12 5 14 4 59 16 08 4 32 5 22 2 05 1 23
Semidmen moisture (%) 43 48 50 23 43 32 48 74 49 05 60 47 66 3 37 73 49 84 43 84 48 02 74 65 49 9 51 3 43 21 39 23 55 26 38 34 46 81 35 69 28 58
Shoreline vegetation (% 100 97 100 100 100 100 100 100 100 100 100 83 50 37 100 27 32 73 48 90 86
Shoreline mangroves (% 13 100 100 100 24 43 97 100 0 37 93 83 42 28 100 8 24 73 48 83 86
Age-1sumerCPUE 0 040 0 077 0 086 0 077 0 299 0 269 0 118 0 028 0 000 0 151 0 148 0 019 0 083 0 000 0 071 0 294 0 154 0 231 0 026 0 063 0 000
Age-1 witer CPUE 0308 0 361 0 241 0 454 0 465 0 126 0 417 0 167 0 083 0 181 0 472 0 296 0 149 0 339 0 111 0 105 0 167
1st vvnter growth rate (mm/d) 0 34 0 47 0 74 0 44 0 41 0 52 0 71 0 29 0 79 0 52 0 69 0 67 0 67 0 63 0 62 0 68
1st alnter Rratio 0 0 210 0 743 0 359 1 625 1713 0 966 0 265 0 000 0 000 0 083 0 166 0 000 0 000 1 120 0 976 0 272 0 110 0 000 0 123 0 400
R ratio, long-term dispersers 0 0 105 0 425 0 179 0 217 0 214 0429 0 000 0 000 0 083 0 165 0 331 0 000 0 288 0 560 0 488 0 408 0 439 0 178 0 084 0 000
R ratio, winter habitats (long-term) 0 0 943 0 212 0 000 0 758 0 535 1 502 0 177 0 000 0 000 0 083 0 166 0 000 0 288 0 280 0 366 0 136 0 329 0 000 0 252 0 000
R ratio, summer habitats (long-term 0 0 000 0 106 0 090 0 000 0 214 0 322 0 000 0 000 0 000 0 083 0 000 0 000 0 000 0 560 0 488 0 272 0 329 0 178 0 056 0 000
Linf(FLmm) 657 657 657 824 595 595 595 824 824 704 704 704 824
3 yr weight (g) 665 1020 679 856 885 672 988 1231 951 1290 685
Note: Catch per unit effort (CPUE) of age-1 snook was measured according to 21 long standard
seine hauls. Long-term growth was expressed as Line (von Bertalanffy parameter), and weight of

age-3 snook recaptures (days-at-large = 366-720) and recapture ratios (R ratio, number
recaptured/number released*100, standardized for effort). Recapture ratios over the long-term
were organized according to those recaptured in winter habitats, those recaptured at least 1 kmn
from their release sites disperserss), and in adult snook summer habitats (beaches, passes, and

seagrass habitats).

















































64









BIOGRAPHICAL SKETCH

Nathan Paul Brennan was born in Hartford, Connecticut, USA in 1965. At the age of 2, he

moved with his family to the remote highlands of Papua New Guinea, in the Engan Province

where his father worked in the Lutheran Mission. After attending elementary school at the

Highlands Lutheran School, Nathan moved with his family to the capital city, Port Moresby.

There he attended Port Moresby International High School for three years. In 1981, the

Brennans moved to Kailua, Hawaii, USA, located on the island of Oahu where he graduated

from Kailua High School in 1984. In May, 1989, he received the degree of Bachelors of Arts in

Zoology from the University of Hawaii at Manoa.

From 1990 -1995, Nathan worked at the Oceanic Institute located at Makapu'u studying

fisheries and aquaculture under the Stock Enhancement Program. In August 1995, he was

recruited to Cookeville, Tennessee for graduate work in a master' s program at the Biology

Department of Tennessee Technological University (TTU) studying Paddlefish stocks in the

Mississippi drainage basin and earned a Master' s degree in Biology in December 1997. By

March 1998, he moved to Sarasota gaining employment at Mote Marine Laboratory. While

maintaining his position at Mote, Nathan started his doctoral program at the University of

Florida in September 2002, examining juvenile snook ecology using stock enhancement as his

primary experimental tool. Nathan currently lives in Sarasota with his wife and two children.














Table 2-1. Numbers of released snook in 1998 and 1999 according to creek, sub habitats

within creeks, and release size category. "Large" refers to snook 180-270 mm FL,
"medium" 131-179 mm FL, "small" 180-129 mm FL
Year Season Estuarine Lot SAR Release sub-habitat Total
Tributary U M L I
1998 Spring North Creek 1 Large 100 0 100 150 350
1 Medium 0 100 0 150 250
1 Small 308 310 310 300 1228
Subtotal 408 410 410 600 1828
South Creek 98-1 Large 100 100 100 150 450
98-1 Medium 100 0 99 146 345
98-1 Small 210 300 210 300 1020
Subtotal 410 400 409 596 1815
Phillippi Creek 98-1 Large 300 150 299 0 749
98-1 Medium 300 450 300 400 1450
98-1 Small 610 610 610 204 2034
Subtotal 1210 1210 1209 604 4233
Bowlees Creek 98-1 Large 300 300 300 601 1501
Tidy Island 98-1 Large 500 500
Medium 2000 0 2000
Lot 98-1 Total 2328 2320 2328 2901 11877


Subtotal 258 259 259 258 1034
South Creek 99-1 Large 87 87 87 87 348
99-1 Medium 86 105 86 86 363
99-1 Small 85 87 87 87 346
Subtotal 258 279 260 260 1057
Bowlees Creek 99-1 Large 87 87 87 85 346
99-1 Medium 87 87 88 86 348
99-1 Small 87 87 89 87 350
Subtotal 261 261 264 258 1044
Hudson Bayou 99-1 Large 87 97 87 87 358
99-1 Medium 90 87 87 86 350
99-1 Small 87 0 0 0 87
Subtotal 264 184 174 173 795
Tidy Island 99-1 Large .345 0 345
99-1 Medium .85 0 85
99-1 Small .260 0 260
Subtotals .. 690 0 690


Large
Medium
Small

Large
Medium
Small


Subtotal 242 275 273 259 1049
Bowlees Creek 99-2 Large 86 0 173 87 346
99-2 Medium 86 87 87 87 347
99-2 Small 87 87 87 88 349
Subtotal 259 174 347 262 1042
Hudson Bayou 99-2 Large 86 85 87 97 355
99-2 Medium 86 87 86 87 346
99-2 Small 78 0 0 0 78
Subtotal 250 172 173 184 779
Tidy Island 99-2 Large .408 0 408
99-2 Medium .470 0 470
Subtotal .. 878 0 878


87 88 86
87 86 87
84 85 86


1041 983 1647
85 87 87
85 87 87
87 91 89
257 265 263
78 87 101
78 86 86
86 102 86


1008 886 1934
2049 1869 3581
4377 4189 5909


87 348
85 345
86 341


949 4620
90 349
89 348
94 361
273 1058
86 352
86 336
87 361


978 4806
1927 9426
4828 21303


1999 Spring North Creek 99-1
99-1
99-1


Large
Medium
Small


Lot 99-1 Total
North Creek 99-2
99-2
99-2
Subtotal
South Creek 99-2
99-2
99-2


Lot 99-2 Total
1999 Tot1
Grand Total









occurred in one high augmentation creek but was not detected in the other. No loss of wild

conspecifies was detected in any of the creeks.

Snook are cannibalistic and show evidence for this behavior intensifying during early

stages. In Chapter 4, I examined ontogenetic differences in cannibalism potential and its role in

controlling abundance. Morphological comparisons (anal spines, mouth gapes, and confusion

colorations eyespotss]) showed that predatory defense and compensatory feeding adaptations

decreased with size. Examination of existing research on snook diet coupled with model

simulations showed that inter- and intra-cohort cannibalism intensified in nursery habitats and

during winter when older age groups were present.

Overall this work used released snook to quantify tradeoffs in production to adult stocks

based on characteristics of snook rearing habitats. Within an age-1 cohort, stocking increased

overall abundance but sometimes with high post-release mortality in hatchery snook. Intra- and

inter-cohort cannibalism was focused on age-0 snook in rearing habitats, and could be influential

on inter-annual recruitment dynamics. These Eindings provide valuable input for stakeholders

evaluating snook stock enhancement potential in Florida.









the upstream, midstream and stream mouth habitats. Most often, snook released in a particular

habitat were also recaptured there.

Our long-term sampling program was designed to validate short-term samples at the

release sites. Because snook disperse to adult spawning habitat during summer months, samples

from these habitats should have provided the least biased data for recapture ratio analysis

independent of juvenile release habitat. In this study, of 6,437 snook sampled along beaches and

barrier island passes (where snook were found spawning, Table 3), only 21 snook (0.33%) were

recaptured from these habitats. The low percentage of hatchery snook at large made it difficult

to obtain a large enough sample size. Long-term samples (snook recaptures from any location

that were at sea for over one year) resulted in similar data as the short-term data set with

significantly higher recaptures of snook from midstream and stream mouth habitats compared to

barrier island habitats. Because most of these snook were recaptured in the creeks, geographical

biases may have again influenced these data. These data correlate well with other Eindings of

high site fidelity in snook. Volpe (1959) and Bruger (1980) conducted snook dispersal studies

using tag returns from anglers and found that 70-79% of snook recaptures were within 8-9 km of

the tag and release site. In these studies angler catches were not particularly linked to the tag and

release sites. This correlates well with our Eindings of 89% of our recaptures within 5 km of the

release sites. The higher percentage in this study could be because our sampling stations were

centered around release sites and less random than angler's capture sites.

Conclusion: Florida' s snook stocks are subj ected to increasing Eishing pressure (Muller

and Taylor, 2005) and coastal rearing habitats have been subj ected to widespread ongoing habitat

modification (Bruger and Haddad, 1986; Lorenz, 1999; Davis et al., 2005; Lorenz and Sarafy,

2006). Considering these resource management problems, coupled with considerable uncertainty
































































57









confuse hawk predators (Powell, 1982). Fish in schools and shoals engage in anti-predator

behavior, and group size has been inversely related to predator success (Neill and Cullen, 1974;

Milinski and Curio, 1975; Major, 1978; Landeau and Terborgh, 1986). Many of these species

are commonly marked with a prominent body spot (e.g. Muglidae, Clupeidae Mullidae,

Gerreidae, Sciaenidae, Carangidae, Sapridae) that may aid in the "confusion effect" by doubling

the number of "eyes" seen by a predator. Eye spots, common in juveniles, reflect ontogenetic

changes in predation risk as the eye spots are lost in later life stages (e.g. snook, red snapper,

butterfly fish, pork fish, sea bass, pork fish, hog fish, wrasses, and cardinal fish). These

ontogenetic variations in morphology of snook even within the age-0 year class (30 175 mm

FL) may be adaptations to compensate for small body size in the presence of predation threat and

competition (see Werner and Gilliam, 1983; Walters and Martell, 2004; Harrell and Gibb,

2006).

Results of cannibalism behavior in the field enclosure trials reflected morphological

differences in small- and larger-sized snook juveniles. Proportional prey sizes tested in the field

enclosures were below the maximum gape and length consumption capabilities of the cannibals

(82-62% predator size), and are probably more realistic. While means were not significantly

different between experimental groups (except for a significant increase in cannibalism rates

when prey were stocked at high density), mean cannibalism rates of small cannibals were

consistently higher than larger cannibals for all the various experimental treatments (Figure 4-8).

Performing the enclosure studies in the field gave the experiment a more realistic setting for

inducing cannibalistic behavior although actual cannibalism rates here were probably lower than

if treatments had been conducted in laboratory tanks and higher if done at natural prey densities.

Several logistical complications occurred in this study. Prey and predator snook found dead in










predation threat for juveniles. Furthermore, relative abundance of wild juvenile snook could also

be used as an indicator of suitable habitat for released fish, although density-dependent effects

should be considered. During winter, snook released in the creeks would be exposed to higher

cannibalism risk due to high snook densities in the creeks at this time (see Adams and Wolfe,

2006; Brennan, unpublished data). These fish would also be only about six months old and

smaller than spring-released snook. Over-winter mortality (e.g. Hurst and Conover 1998) is also

a major concern with snook because of their susceptibility to cold temperatures. For stock

enhancement, these concerns must also be weighed against rearing costs to determine economic

feasibility.

Overall, hatchery-reared snook showed high site fidelity among creeks (or release

habitats) and even for specific locales within creeks (or release sub-habitats). Within the first

year after release, juvenile snook had emigration rates of only 3-6% outside of the release

habitats. This agrees with an earlier study examining recaptures of juvenile snook released in

NC (1% emigration from the creek, Brennan et al., 2006). During subsequent summers and fall

(i.e. spawning months, Taylor et al., 1998), however, mean distances between release and

recapture sites (i.e. dispersal distances) continued to increase with. Young mature snook tended

mostly to disperse to spawning habitat and grass flats nearest their rearing habitats. Older and

larger snook showed increasing dispersal distances during summer and fall. The increasing

dispersal tendencies with increasing age have important implications for stock management.

While young mature snook disperse to local spawning sites, larger snook dispersed more widely.

Because the larger snook are more fecund (Chambers and Leggett, 1996; Berkeley et al., 2004;

Longhurst, 2006; Francis et al., 2007), these fish may have a greater influence on snook

recruitment dynamics and genetic patterns. Widespread and intense harvest of these larger









and percent shoreline composed of mangrove vegetation. For standardization purposes, I took

samples at perpendicular distances away from the mean high tide elevation indicated on the

bank. I estimated mean high tide elevations on the bank as the upper limit of barnacle, limpet, or

algae distribution on shoreline structure. Shoreline slopes (rise/run) were calculated using

vertical measurements to the bottom at Im perpendicular from the bank starting at the mean high

tide elevation. Substrate softness was measured 2 m perpendicular from the bank's mean high

tide elevation. To do this a 0.75 cm2 x 1.5m cylindrical fiberglass rod was held in a vertical

position, resting on the substrate. Then, a 4.1 kg standard weight was placed on the top-side of

the rod and the subsequent penetration distance (cm) of the rod into the sediment was measured.

Substrate core samples (3.8 cm diameter x 10 cm depth) were also collected 2 m perpendicular

from the banks mean high tide elevation at each site. Substrate cores were analyzed for percent

organic material and moisture (by weight; Lenore et al., 1998). I also visually estimated the

percent coverage of macro-vegetation (e.g. trees, grasses, shrubs) along an approximately 100 m

stretch of the shoreline centered on the collection site. Similarly, for the same stretch of given

shoreline, I also estimated the percent coverage of mangrove vegetation (i.e. Rhizophora mangel,

Avicennia germinans, or Laguncularia racemosa).

From each core sample, I removed an approximately 15 ml sub sample after thorough

mixing to obtain a homogeneous mixture. Each subsample was placed in an aluminum pan,

weighed (mg), and then dried at 105oC for 24 H. After drying, subsamples were reweighed (to

obtain a percent moisture content), then ashed at 550oC and weighed again to obtain the

percentage of organic material in the sample.

I also collected water quality measurements (temperature [oC], salinity (%o), dissolved

oxygen [mg/1], and pH) with a YSI 556 handheld multi-parameter model (YSI incorporated,










4-5 Graphical comparisons of snook morphometrics (in mm) .............. .....................144

4-6 Photograph of a 70 mm (TL) and a 178 mm (TL) common snook. Black spot at tip
of dorsal fin is obvious in smaller snook, but spot is nearly absent in larger snook.............. 145

4-7 Graphical results from field enclosure trials describing mean cannibalism rates of
small and large snook according to various prey treatments ................. ............... .....146

4-8 Plot of body lengths (fork length FL) of snook cannibals and cannibalized prey ...........147

4-10 Monte-Carlo simulation results (n=10,000 iterations) showing proportions of age-0
mortality due to age-specific cannibalism .............. ...............148....

4-11 Monte-Carlo simulation results (n=10,000 iterations) showing the proportion of age-
0 abundance (at the beginning of the model period) cannibalized by the various age
groups of snook ........._.._.. ...._.. ...............149....









Pre-Release April 2002 Sampling

To estimate pre-release juvenile snook abundance, we performed creek-wide standard

seining efforts in each creek prior to stocking hatchery-reared snook (Figure 3-1). Captured

snook were also marked with externally visible implant elastomer (VIE) tags implanted in the

caudal Eins (see Brennan et al., 2005) with VIE colors specific to creeks to (1) aid in determining

site fidelity and inter-creek migration rates, and (2) identify snook recaptured multiple times.

Tagging and Release

Hatchery-reared snook, Fl progeny from wild parental stocks, were hatched in the

summer of 2001 and reared until spring 2002 when tagging and release activities occurred in

May, 2002. Each released hatchery-reared snook was tagged with (1) a coded-wire tag (CWT)

to identify it's associated release information (i.e., release creek, release date, size-at-release),

and (2) two red VIE tags implanted in the caudal fin to externally identify it as a hatchery-reared

snook (as in Brennan et al., 2005). Tagged juveniles were then returned to their tanks and held

for one week to recover from the tagging process. Salinities in the rearing tanks ranged from 3 -

6 ppt and water temperature from 26-30 oC.

On the day of release (six days after tagging), Eish from each release group were

harvested and checked for tag presence. Snook were transported in tanks with brackish water by

truck and boat, and stocked directly into 22 m3 predator-free acclimation enclosures and held for

3 d (a protocol to improve post-release survival rates, Brennan et al., 2006). Enclosures were

located along shorelines, mostly vegetated with mangrove (Rhizophora mangle), oaks (Quercus

spp.) palms (Sabal palmetto), and Brazilian pepper (Schinus terebinthifolius), and at some

locations near boat docks and riprap. After 3d of acclimation, snook were allowed to swim

freely from the enclosures. Release numbers were proportional (either 100% or 10%) to pre-

release abundance estimates of juvenile snook in each experimental creek.


























Table 4-1. Summary of studies examined regarding common snook diets. 'Juvenile habitat' refers to whether samples were

collected in juvenile snook habitat. The numbers of individual snook reported from the various age groups are based


on body length information.

Reference study Sample Region Season Habitat description Juvnelle Sizerange No ofstomach's Stomach's Percent Obsenred Estimatedsampleage Known
dates collected habitat? (mm) Sampled w/contents w/contents canmibals Age-0 Age-1 Age-2+ sizes(num)

Marshall, 1958 1954-1955 IRL, West Coast, East Coast FL 9 Mangrove shorelines No 230-851 FL 128 61 47 7 0 0 0 128 128


Total 6768 3241 617 7 2604 1608 1735 5947


Harrington and Harrington, 1961

Chavez, 1963

Linton and Richards, 1965

Austin and Austm. 1971


Indian River Lagoon, FL

Veracmz. Mexico

1963-1964 Georgia

1967-1968 WesternPuerto Rco


summer, fall Marshes Yes <100 FL

all seasons Mangrove shroelines Yes 47-341 TL

9 ~estuarine marsh No 24-75 SL

all seasons Mangrove-lined Yes 23-87 TL
canals and river banks


0 172




0 62

0 38


Rojas, 1975

McMichaels et al, 1989


1972-1973 Laguna de Terrmnos, Campeche, Mexico fall, winter

1984-1985 Tampa Bay, FL all seasons


Mangrove wetlands Yes luvemlles mature

Grass flats, beaches, Yes 11-346 SL
backwatersrivers


327 207 633 3

480 427 89 0 0 395


Gillmore et al,1983

Gillmore et al,1983
Gillmore et al,1983

Luskovitch et al, 1995

Fore and Schmidt, 1973

Fore and Schmidt, 1973

Allaume et al 1997


1974-1980 Indian River Lagoon, FL

1955-1956 Indian River Lagoon, FL

1974-1980 Indian River Lagoon, FL

1985-1986 St Lucle County, FL

1971-1973 Everglades, FL

1971-1973 Everglades, FL

1992-1994 Puerto Rico





Feb-Mar 3003 Charolette Harbor, FL

Nov 03-Feb 04 Charolette Harbor, FL

2000-2002 Charolette Harbor, FL


all seasons Freshwater Yes 11-156 SL

all seasons Marshes Yes ca 20-275 SL

all seasons Seagrass Yes 15-600 SL

summerfall,winter Mangrove impoundment Yes 5-119 SL

summer, fall Tidal creeks, canals Yes 14-196 TL

summer, fall Mangrove shorelmnes Yes 224-1020 TL

spring, summer,fall Rivers, freshwater and Yes <300 SL
saltwater lagoons,
backwaters


0 99

0 441

0 21 300

0 258

0 183

0 0 0 271

0 91


Adams and Wolfe, 2006

Adams and Wolfe, 2006

Blewett et al, 2005



Brennan, unpublished data

Brennan, unpublished data

Brennan, unpublished data

Brennan, unpublished data


winter

wnter

all seasons


Estuarine ponds, creeks Yes 141-680

Estuarine ponds, creeks Yes 190-651

Coastal mangrove No 300-882 SL
and seagrass


153 94 61 4 3 43 25 42

127 3 0 150 19

694 432 62 2 0 0 0 694



114 90 79 0 0 0 114 0


2001 Sarasota. FL


summer Tidal creeks

summer Tidal creeks

winter, spring Tidal creeks, canals

summer Tidal creeks


Yes 100-250 FL

Yes 100-735

Yes 80-965 FL

Yes 103-466 FL


114


Sarasota, FL

Sarasota, FL

Sarasota, FL


0 12 248 93

1 789 513 408

0 0 258 80















-----age-2
NC/SC -age-3
.. .age-4
(N age-2 = 316) -age-5
1 ~ age-6
ISLANDS
(N age- = 171)


c 400


C 300


200


100


0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75

Growth rate (mm/d) in rearing habitat


0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75

Growth rate (mm/d) in rearing habitat


Figure 2-18.


Numerical (graph A) and biomass (graph B) proj sections to adult stages of juvenile
hatchery-reared snook released in various rearing habitats with characteristic
growth rates by first winter. Projections assume equal stocking levels (1000 age-
1 snook per rearing habitat). Arrows indicate projections according to rearing
habitats observed in Sarasota and Manatee Counties (NC/SC = North Creek or
South Creek, HB=Hudson Bayou, TIDY=Tidy Peninsula, BC= Bowlees Creek,
ISLANDS=barrier island habitat). Shaded dots in B highlight the differences in
biomass production of age-4 snook between NC/SC and ISLANDS.









I used repeated measures analysis to compare pre- and post-release abundance for

treatment and control creeks. If the April 2002 abundance estimate of wild snook juveniles in a

particular creek (pre-release) was less than its early summer (post-release June or July 2002

samples) abundance estimate, I used early summer abundance of age-1 wild snook as our pre-

release estimate (I assumed that no new recruitment of wild age-1 snook occurred; age-1 snook

typically begin to decline in abundance in creeks by mid-summer [Brennan and Leber,

unpublished data; Gilmore et al., 1983]). Since large-scale emigration of age-1 snook occurred

from mid-summer through fall, I only used early-summer and winter abundance estimates for

post-release repeated measures analysis. I also used a linear regression to model augmentation

level (expressed as a percentage of standing wild stock) with change in abundance after stocking

(also expressed as a percentage) to test if a significant relationship existed between augmentation

level and change in abundance.

Density (number per 30 m of shoreline) of juvenile snook in the creeks (Dc) at time t was

calculated by dividing total in-creek juvenile snook abundance N- cr by total creek shoreline

distance (Hc) as follows:

Dc, = N / He (3 -7)

I used data from pre- and post-release sampling, to compare observed and expected densities

within the creeks. To generate expected density, I assumed that numbers of stocked fish were

completely additive to wild juvenile numbers, then used a Chi-Square test to compare differences

in observed and expected age-1 densities in the creeks after releases occurred. I also used an

analysis of covariance procedure to compare in-creek changes in density (Dc,) of juvenile snook

before and after releases (using pre-release density, peak summer density, and peak winter

density).














12




10


6 1 3
v 6


W2 8

6l -- -- -
9 -



17 14 1
2 -- -- -- ---- ---



37 1
525
15322921 2
71 3 3


S S FW S S FWS S FW S S FW S S FW S S FW
SAGE-1 AGE-2 IIAGE-3 IIAGE-4 IIAGE-5 IIAGE-6


Figure 2-16. Mean dispersal distances from rearing habitats (minimum distance from release
site to recapture site) of snook recaptured within seasons (spring, summer, fall
winter) and by snook age (years). Data are from snook released in the spring of
1998, and 1999. Numbers of individuals recaptured within each season are shown
above bars. Error bars are standard error.









water ("U"), (2) midstream reaches in brackish water habitats ("M"), (3) lower reaches in higher

salinity waters near the creek mouth ("L" [lower site]), and (4) along barrier island high salinity

habitats ("I") nearest the respective tidal creek (Table 2-1, Table 2-2, Figure 2-1). Experimental

releases in sub-systems were repeated across the following release systems (or tidal creeks):

South Creek (SC), North Creek (NC), Phillippi Creek (PC), Hudson Bayou (HB), and Bowlees

Creek (BC), (Table 2-1, Figure 2-1). Releases also occurred at Tidy Peninsula (TIDY) a high

salinity mainland mangrove-lined peninsula and a nearby barrier island.

Study Habitat and Timeline

The experimental area encompassed about 30 km of coastline located in Sarasota and

Manatee Counties, on the west coast of Florida, USA (Figure 2-1). Stocking occurred during

April 1998 and 1999 (Table 2-1). In 1998 we stocked the four sub-habitats of four tidal creeks

SC, NC, PC, and BC, plus two subhabitats of TIDY (Table 2-1). In 1999 we stocked the same

sites as in 1998, but replaced PC with HB, a high salinity developed bayou and its adj acent

barrier island habitat (Figure 2-1), and stocked only one sub-system at TIDY. Releases in 1999

were replicated with a second pulse of releases at the same sites two weeks later (Table 2-1).

Barrier island release habitat (sub-system=I) were characterized by high salinity water

(typically 30-34 %o), with mangroves vegetating the shorelines (i.e. Rhizophora nzangel,

Avicennia gernzinans, and Laguncularia racentosa). Water depths along the shoreline were

generally 1 2.5 m at low tide (tidal variation at all sites was approximately 0.6 m), with

substrates of sand, shell, and mud. At the tidal creek mouths (sub-system=L) salinities were

typically more variable due to seasonal rain and tidal flux, but generally around 25 32 %o.

Here, shoreline vegetation was also primarily mangroves. Water depths at the mouth of NC

(non-dredged habitat) were generally < 0.5 1.0 m, but 1.0 2.0 m at the other sites (dredged).

Substrate was comprised primarily of mud, sand, and live oyster bars (Cra~ssostera virginica).
















Mean = 0.189 +/- 0.003 Model sensitivity
-- -- -- age-0 field cannib. Rate 94.760 %
Age-0 FL 20.754 %
~Mu 14.910 %
b 8.527 %
Age-0 residence 2.838 %
~Age-1 FL 1.307 %
Age-2+ relative abundance 0.868 %


I


Model sensitivity
Age-1 field cannib. Rate 88.462%
Age-1 relative abundance 39.981 %
Age-0 FL 12.916 %
Mu 9.375 %
b 4.582 %
Age-2+ FL 1.888 %
Age-0 residence 1.770 %
Age-2+ residence 1.374 %
Age-1 FL 0.277 %
Age-1 residence 0.258 %


Model sensitivity
Age-2+ field cannib. rate 82.781 %
-------Age-2+ relative abundance 46.233 %
Age-0 FL 18.197 %
Mu 9.964 %
b 6.636 %
Age-0 residence 2.457 %
Age-2+ FL 1.164 %
Age-0 field cannib. rate 0.926 %
Age-1 FL 0.634 %
Age-1 relative abundance 0.518 %


700
600
500

400
300


0 0 0 0 0 0 00dIC) \D0 0


2500


Mean = 0.133 +/- 0.008


2000 -


1500

a 1000

500


0 0 0 0 0 0 00dIC \D0 0 -


800


600


400


200


-


Mean = 0.070 +/- 0.002


0 0 0 0 0 0 0 0 C 0 0 +

Age-0 Cannibalism mortality/ Age-0 natural mortality




Figure 4-10. Monte-Carlo simulation results (n=10,000 iterations) showing proportions of age-
0 mortality due to age-specific cannibalism: (A) age-0 snook, (B) age-1 snook,
and (C) adult snook (>= age-2 snook). Results from each sensitivity analysis are
inset in each graph. Respective means and associated 95% confidence limits of
the mean are included.











Robbins, R.K., 1980. The lycaenid "false head" hypothesis: historical review and quantitative
analysis. J. Lepid. Soc. 34, 194-208.

Rojas, J.C., 1975. Contribucion al conocimiento de la biologia de los robalos Centropomus
undecimalis y C. Poeyi en la Laguna de Terminos, Campeche, Mexico. Biol. Inst.
Oceanogr. Univ. Oriente 14 (1), 51-70.

Ross, S.W., 2003. The relative value of different estuarine nursery areas in North Carolina for
transient juvenile marine fishes. Fish. Bull. 101 (2), 384 404.

Roussel, J.M., Bardonnet, A., 1999. Ontogeny of diel pattern of stream-margin habitat use by
emerging brown trout, Salmo trutta, in experimental channels: influence of food and
predator presence. Eny. Biol. Fish. 56 (1-2), 253-262.

Rozas, L.P., Odum, W.E., 1988. Occupation of submerged aquatic vegetation by fishes: testing
the roles of food and refuge. Oecologia 77, 101-106.

Russell, D.J., Rimmer, M.A., McDougall, A.J., Kistle, S.E., Johnston, W.L., 2004. Stock
enhancement of barramundi, Lates calcarifer (Bloch), in a coastal river system in
Northern Australia: stocking strategies, survival and benefit-cost. In: Leber, K.M.,
Kitada, S., Blankenship, H.L., Svsand, T. (Eds.), Stock enhancement and sea ranching.
Blackwell Publishing, Oxford, UK, pp. 490 500.

Ruzycki, J.R., Wurtsbaugh, W.A., 1999. Ontogenetic habitat shifts of juvenile Bear Lake
sculpin. Trans. Am. Fish. Soc. 128 (6), 1201-1212.

Sanchez-Lamadrid, A., 2002. Stock enhancement of gilthead sea bream (Sparus aurata, L.):
assessment of season, Eish size and place of release in SW Spanish coast. Aquaculture
210 (1-4), 187-202.

Sanderson, B.L., Hrabik, T.R., Magnuson, J.J., Post, D.M., 1999. Cyclic dynamics of a yellow
perch (Perca flavescens) population in an oligotrophic lake: evidence for the role of
intraspecific interactions. Can. J. Fish. Aquat. Sci. 56 (9), 1534-1542.

SAS Institute, 1998. SAS/Stat Users Guide. Version 6: volumes 1 and 2. Cary, North Carolina.

Savino, J.F., Stein, R.A., 1982. Predator-prey interaction between largemouth bass and bluegills
as influenced by simulated, submersed vegetation. Trans. Am. Fish. Soc. 111 (3), 255-
266.

Sazima, I., 2002. Juvenile snooks (Centropomidae) as mimics of mojarras (Gerreidae), with a
review of aggressive mimicry in Hishes. Eny. Biol. Fish. 65, 37-45.











Table 2-8. von Bertalanffy growth parameters for snook recaptures (n) sourced from
different release habitats. Length and age data only included recaptures of
hatchery snook at-large for 180d or longer.
Linf k tO n
South Creek 657 0.250 -0.486 46
North Creek 552 0.550 0.470 249
Bowlees Creek 704 0.408 -0.502 60
Island Habitat 824 0.164 -1.650 40










Age-Structured Cannibalism Model: Evidence of cannibalism in field collections of

snook (Adams and Wolfe, 2006; N. Brennan, unpublished data) and other fishes (Lang et al.,

2000; Persson et al., 2000; Vik et al., 2001) have shown that predation efforts are typically

focused age-0 individuals. As such, I developed a simplified age-structured forecast model to

estimate predation mortality from age-0, age-1, and adult ("age-2+") snook age groups on an

age-0 year class in Sarasota, Florida. I extracted cannibalism rates (number of cannibals

observed in stomachs/ total stomachs examined) from studies that were conducted in conditions

where cannibalism was most likely to occur. In Sarasota, juvenile and adult snook over-winter

in tidal creeks (Brennan et al., 2008; Chapter 2 and 3), and emigrate back to summer habitat

through the spring (Gilmore et al., 1983, Brennan, unpublished data). In November, snook from

the age-0 cohort are generally large enough to be captured by our standard seines (1 cm mesh,

snook mean lengths about 100 mm FL). As such, I modeled age-0 cannibalism from November

1 May 15 of a typical year in a tidal creek.

I used a series of variables in a cohort-based Monte Carlo simulation model to predict

how cannibalism mortality from each of the above age categories decreased the original

abundance of the age-0 year class while over-wintering in a typical rearing habitat. Within the

age groups and across all ages, I also measured the contribution of cannibalism mortality to the

total natural mortality for the period. Model calculations were based on real data collected in

Southwestern Florida in rearing habitats of age-0 snook (tidal creek). The model simulation time

frame was November 1 May 15 of a hypothetical year. For the model, the number of age-0

snook in the hypothetical creek on November 1 was set to 5000. Length frequency data from

Sarasota snook collections in tidal creeks from 1998- 2005 (Brennan, unpublished data) were

used to obtain typical size estimates of each of the age categories on November 1. Size-specific









varied the parameter, "mean days in creek" around a normalized distribution with SD = 2. For

age-1 and age-2+ snook, I assumed they immigrated to the creek from outside habitats to over-

winter, and then gradually dispersed from the creeks as the season warmed. I used a normal

density distribution function, and the model selected from normally distributed parameters

(mean=50, SD = 5 for age-1 snook, and mean= 50, SD = 2 for age2+ snook).

(4) Relative abundance of age-1 snook (compared to the age-0 cohort' s abundance) varied within

a uniform distribution (0. 10 1.51 for age-1, 0.05 0.70 for age-2+).

(5) Estimates of cannibalism rates (h,) were based on a normalized distribution of the mean field

rate (overall number of cannibals found/number stomachs examined) with a variance of 1 SD of

the mean. I generated variance around these rates through bootstrap random sampling of the data

set (individual snook of age group i expressed as 0 [no cannibalism] or 1 [cannibal]). I used

Microsoft Excel for bootstrapping procedures (Teknomo, 2006) where the original data set was

randomly sampled to generate 100 data sets (n=1,000 per set). l used sensitivity analysis to

assess the influence of the above mentioned parameter estimates on the results. The model

performed 10,000 iterations using Microsoft Excel with the Crystal Ball add-in (Werckmand et

al., 1998).

Results

Phenotypic Characteristics

Mouth gape, body depth, and anal spine length showed negative allometric scaling with

body length. Mean slopes from linear relationships of the log transformed variables above were

0.664-0.901 (Table 4-3, and Figure 4-5) with the anal spine (slope=0.664) showing strong

allometric scaling. Isometric scaling was found for log transformed relationships of SL and FL,

and VA-AN and FL, however. The dark spot located at the tip of the first dorsal fin was









Based 1000 hatchery-reared juveniles stocked in rearing habitats during early summer and the

above assumptions, Creeks such as NC and SC (with first winter growth rates of 0.44 mm/d),

were predicted to produce about 316 age-2 snook while a barrier island release site stocked at the

same level would produce about 171 age-2 snook (Figure 2-18a). Under these conditions, and

not accounting for fishing mortality, the numbers of age-4 snook produced by NC or SC would

be about the same as the number of age-2 snook produced by an island habitat stocked at the

same level. Total biomass estimates of hatchery-released snook recruiting to adult stocks based

on rearing habitat growth rates showed a hump-shaped relationship (Figure 2-18b). Rearing

habitats with intermediate first winter growth rates of stocked snook produced the highest

biomass of adult snook. Biomass production of age-2 hatchery snook from rearing habitats as

NC or SC with slow growth rates produced about the same biomass of age-2 snook as did island

sites where growth rates were much higher (Figure 2-18b). By age-4, however, about 60% more

biomass was predicted for NC or SC sources compared to island sources.



Discussion

In this study I used stock enhancement to understand important life history characteristics

and habitat interactions of juvenile snook stocked into various rearing habitats and examined

contribution rates of the stocked juveniles to adult stages. Due to the vulnerability of Florida' s

snook stocks to overfishing (Muller and Taylor, 2005) and widespread modification of coastal

rearing habitats (Bruger and Haddad, 1986; Lorenz, 1999; Davis et al., 2005b; Lorenz and

Sarafy, 2006) investigations of stock enhancement potential is warranted. Notwithstanding, the

controversial nature of this measure coupled with the status of snook stocks in Florida (Muller

and Taylor, 2005) emphasizes the need for a cautionary approach for stock enhancement

research (Blankenship and Leber, 1995; Walters and Martell, 2004, Lorenzen, 2008). This









performance in catfish Claria~s gariepinus of various sizes (1 11-923 mm TL) (Wassenbergh et

al., 2006), absolute suction feeding performance clearly increased with Hish size. Small animals

have competitive disadvantages because of their smaller body size (Werner and Gilliam, 1984).

However, when catfish suction performance was scaled according to predator size and relative

prey size, small fish clearly outperformed the larger fish (Wassenbergh et al., 2006). This work

strengthens arguments by Harrel and Gibb (2006) that young animals of smaller sizes, have

compensatory adaptations to minimize this difference. These compensatory adaptations would

be especially important in fish that cannibalize within a cohort. Cannibalized prey are typically

proportionally larger than other prey items (Juanes, 2003), and cannibalism is often associated

with low alternative prey availability or high competition for prey forcing cannibalism among

similarly sized individuals (Persson et al., 2000; Walters and Martell, 2004).

Snook coloration suggests high predation threat among juveniles. Sazima (2002)

observed that juvenile snook generally resemble moj arra (Eucinostomus gula) in overall body

shape and have a prominent black spot at the tip of the dorsal fin. He theorized that snook use

aggressive mimicry and school with mojarra to take advantage of unsuspecting prey. I suggest

that in addition to aggressive mimicry, this spot (prominent in snook smaller than 125 mm, but

generally much less distinct in snook larger than 175 mm FL) (see 4-6) is a form of self mimicry,

an adaptation to confuse predators, causing them to strike at an apparent "eye spot" in a

peripheral, less-essential, body location (Blest, 1957; Wickler, 1968; Hailman, 1977; Robbins,

1980). The presence of an additional mark or spot on the body (in addition to the eyes) is

common in schooling species and probably adds to the predator "confusion effect" (Welty, 1934;

Landeau and Terborgh, 1986). Some birds have 'flash marks' in their plumage possibly as an

adaptation to confuse predators (Brooke, 1998); weasel's black-tipped tails are thought to












Table 3-1. Results from standardized sampling in April 2002 and physical attributes of
experimental creeks (shoreline distance and percent altered habitat). Population
estimates are of age-1 snook found in each creek.
Expenmental Number of Number of Total Sampled Sampled shorelme % Altered Mean CPE Adlusted Variance Calculated Calculated
creek seme hauls shores sampled shorelme (m) shorelme (m) (%of total) shorelme mean CPE age-1 abundance max abundance
Bowlees Creek 28 38 14051 1159 8 25 65 0 64 0 81 2 07 375 1,715
Whitaker Bayou 18 26 7650 793 10 37 14 1 50 1 90 9 30 476 1,913
North Creek 45 66 25189 2013 7 99 2 1 93 2 44 13 82 2,018 8,003
South Creek 26 34 24262 1037 4 27 7 0 63 0 80 2 07 639 2,904
Totals 117 164 71153 5002 7 03 3,508 14,535










Wennhage, H., 2000. Vulnerability of settling plaice Pleuronectes platessa to predation: effects
of developmental stage and alternative prey. Mar. Ecol. Prog. Ser. 203, 289-299.

Werckmand, C., Hardy, T., Wainwright, E., 1998. Crystal Ball Users Manual. Decisioneering,
Colorado.

Werner, E.E., Gilliam, J.F., Hall, D.J., Mittelbach, G.G., 1983. An experimental test of the
effects of predation risk on habitat use in fish. Ecology 64, 1540-1548.

Werner, E.E., Gilliam, J.F., 1984. The ontogenetic niche and species interactions in size-
structured populations. An. Rev. Ecol. Systemat. 15, 393-425.

Wespestad, V.G., Fritz, L.W., Ingraham, W.J., Megrey, B.A., 2000. On relationships between
cannibalism, climate variability, physical transport, and recruitment success of Bering
Sea walleye pollock (Theragra chalcogramnma). ICES J. Mar. Sci. 57 (2), 272-278.

Wickler, W., 1968. Mimicry in plants and animals. McGraw-Hill, New York.

Yamamura, O., Yabuki, K., Shida, O., Watanabe, K., Honda, S., 2001. Spring cannibalism on 1
year walleye Pollock in the Doto area, northern Japan: is it density-dependent? J. Fish.
Biol., 59 (3), 645-656.

Yamashita, Y., Nagahora, S., Yamada, H., Kitagawa, K., 1994. Effects of release size on
survival and growth of Japanese Flounder Pat idiL Irhy s olivaceus in coastal waters off
Iwate Prefecture, northeastern Japan. Mar. Ecol. Prog. Ser. 105, 269-276.

Yoder, J.M., Marschall, E.A., Swanson, D.A., 2004. The cost of dispersal: predation as a
function of movement and site familiarity in ruffed grouse. Behav. Ecol. 15 (3), 469-476.

Zohar, Y., Hines, A.H., Zmora, O., Johnson, E.G., Lipcius, R.N., Seitz, R.D., Eggleston, D.B.,
Place, A.R., Schott, E.J., Stubblefield, J.D., Chung, J.S., 2008. The Chesapeake Bay blue
crab (Callinectes sapidus): A multidisciplinary approach to responsible stock
replenishment. Rev. Fish. Res. 16 (1-3), 24-34.









multifilament mesh) was hauled at roughly every third 30 m section of shoreline throughout each

creek (Figure 3-1, Table 3-1). Each haul approximately sampled a 30 m x 21 m rectangle (630

m2). For creek sections less than 21 m wide, we sampled 30 m of creek length but incorporated

both shorelines into the sample. The seine, loaded on a 4 m kayak, was deployed by hauling

about 21 m of the leading end of the seine towards the shoreline. Once there, a second person

held the outside "corner" of the block and a third person pulled the kayak parallel to the bank (or

against the opposite shoreline) for 30 m then turned and deployed the remaining 21 m of net

towards the shoreline. Once against the bank, both persons moved toward each other along the

shoreline and retrieved the remainder of the seine. To avoid overestimating snook abundance, I

divided snook catch by two for samples collected from two-bank sites because of the increased

shoreline habitat. Seines were not hauled in narrow areas (i.e. less than 3 m across) such as

mosquito ditches, but these shorelines were assumed to be valid snook habitat and were used to

calculate total snook abundance.

Gear Efficiency and Population Estimation

Prior to abundance estimates, we first conducted depletion-removal trials in January and

February, 2002 to measure gear efficiency. I used a modified depletion-removal population

estimate (Hillborn and Walters, 1992) within a defined area to predict single-pass seine

efficiency (Figure 3-2). Three types of depletion-removal methods were used depending on

habitat type. In creek areas narrower than 21 m across (i.e., two stream banks, "two-bank

depletion"), two nets (21-24 m long, 0.9 m deep, 1 cm multifilament mesh), about 42 m apart,

blocked off upstream and downstream borders of the sampling area (Figure 3-2a). We used the

standard seine described above, to sample ~ 30 m within this zone, allowing ~ 6 m from the

seine perimeters to the block nets measured along the shoreline to allow for escapement (as

might occur in a single pass seine). After this, we seined the entire area with the above seine two





0.50


recapture ratio at
Age-1 beaches and passes O BCU
summer density *
recapture ratio
of long-term
dispersers


bank slope
eAge-3weight




short-term
ICL
growth
salinity
*pH


substrate softness


BCM
OO B


Age-1 w inter CPUE

1st w inter recapture ratio


*sediment organic


0.00


NCM O sediment *
moisture


*long-term w inter
recapture ratio
O NCL
SCL O


*temperature


*dissolved oxygenONC


shoreline vegetation


mangrove cover


0


-0.50 C


0.00


0.50


Principle component 1 (36%)


Figure 2-17.


Plot of the first two principle components comparing long-term responses
(recaptures at-large for over 1 year, black dots) of the released fish, short-term
responses (from recaptures at-large less than 1 yr., also black dots), and
characteristics of the release sites (shaded dots). The percentages in brackets on
the axes represent the percent of the overall variation explained by each respective
principle component.












700
600
500
400
300
200
100
0


O n8 OO O


0 100 200 300 400 500
Predator length (FL mm)


Figure 4-8.


Plot of body lengths (fork length FL) of snook cannibals and cannibalized prey.
Data originally expressed as standard length were converted to fork length (FL).




Full Text

PAGE 1

HABITATAND DENSITY-MEDIATED INFLUE NCES ON SNOOK ECOLOGY: LESSONS LEARNED FROM MANIPULATIVE RELE ASE EXPERIMENTS WITH HATCHERYREARED JUVENILE SNOOK By NATHAN PAUL BRENNAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 Nathan Paul Brennan 2

PAGE 3

To my father who encouraged and believed in me, who nurtured my intellectual curiosity, provided perspective, and strengthened me, making this milestone possible. 3

PAGE 4

ACKNOWLEDGMENTS I will always be thankful and grateful for my wife, children, my mother and father, and family members who showed their continued patience, support, and encouragement throughout this venture. I gratefully acknowledge the contributions of my graduate committee Ken Leber and Bill Lindberg (co-chairs), Carl Walters, Craig Osenberg, Mike Allen, and Tom Frazer for their overall direction in experi mental design, scientific appro ach, and editorial assistance. Throughout, Ken has also provided constant support, direction, and encouragement which allowed me to pursue independent avenues of res earch in a larger stock enhancement program. I thank Carl for his additional as sistance with field work, data an alysis, and overall direction. Bill Lindberg also provided extr a direction, time, and encouragement during this endeavor. Mote Marine Laboratory fina ncially supported my graduate work and I commend Mote administration for supporting the educational pursuits of their employees. Much gratitude is due to R. DeBruler, D. Wilson, V. Mooney, A. Warn er, B. Blackburn, G. Wahl, C. Ondercin, A. Read, S. Andersen, and many interns and volunteers in the Stock Enhancement Program at Mote Marine Laboratory. I thank those at Motes Ce nter for Aquaculture Research and Development for rearing snook used in this study. I thank L. Blankenship (North west Marine Technology, Olympia, WA) and staff at the Stock Enhancemen t Research Facility, Fl orida Fish and Wildlife Conservation Commission (FWC) fo r assistance during tagging and release activities. I thank P. Hull, D. Doherty, and M. McLeod for assist ance during snook spawning trips and snook releases. Funding for this work was provided by William Mote for initiating the snook hatchery, the Florida Fish and Wildlife Cons ervation Commission, and the Na tional Oceanic and Atmospheric funded research consortiumthe Science Consor tium for Ocean Replenishment (SCORE). 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES.........................................................................................................................9 ABSTRACT...................................................................................................................................15 CHAPTER 1 INTRODUCTION................................................................................................................. .17 2 INFLUENCE OF REARING HABITAT ON GROWTH, SURVIVAL, DISPERSAL, AND RELATIVE STOC K CONTRIBUTION OF COMMON SNOOK Centropomus undecimalis ........................................................................................26 Introduction................................................................................................................... ..........26 Methods..................................................................................................................................29 Experimental Design.......................................................................................................29 Study Habitat and Timeline.............................................................................................30 Hatchery Fish Production, Tagging and Release............................................................31 Collection of Environmental Variables at Release Sites.................................................32 Short-Term Collections...................................................................................................34 Long-Term Collections...................................................................................................34 Sample Processing and Data Analysis............................................................................36 Results.....................................................................................................................................39 Short-Term collections at Release Sites..........................................................................40 Long-Term Assessment...................................................................................................43 Discussion...............................................................................................................................45 3 MANIPULATIONS OF STOCKING MAGNITUDE: ADDRESSING DENSITY-DEPENDENCE IN A JUVENILE COHORT OF COMMON SNOOK Centropomus undecimalis ........................................................................................85 Introduction................................................................................................................... ..........85 Methods..................................................................................................................................88 Study Area.......................................................................................................................88 Experimental Design and Sampling Methods.................................................................89 Gear Efficiency and Population Estimation....................................................................90 Pre-Release April 2002 Sampling...................................................................................94 Tagging and Release........................................................................................................94 Post-Release Evaluation..................................................................................................95 Results.....................................................................................................................................95 Gear Efficiency................................................................................................................95 5

PAGE 6

Tagging and Release of Hatchery-Reared Snook............................................................96 Post-Release Evaluation..................................................................................................96 Creek Density..................................................................................................................98 Discussion...............................................................................................................................98 4 CRYPTIC CANNIBALISM IN SIZESTRUCTURED SNOOK POPULATIONS...........112 Introduction................................................................................................................... ........112 Materials and Methods.........................................................................................................116 Generalized Snook Life Hi story in a Tidal Creek.........................................................116 Approach 1: Phenotypic Characteristics.......................................................................116 Approach 2: Field Enclosure Trials...............................................................................117 Approach 3: Cannibalism Intensity in Wild Snook Populations..................................119 Results...................................................................................................................................123 Phenotypic Characteristics............................................................................................123 Field Enclosure Trials....................................................................................................124 Cannibalism Meta-Analysis and Model Results...........................................................124 Discussion.............................................................................................................................126 5 CONCLUSION................................................................................................................... ..150 LIST OF REFERENCES.............................................................................................................159 BIOGRAPHICAL SKETCH.......................................................................................................183 6

PAGE 7

LIST OF TABLES Table page 2-1 Numbers of released snook in 1998 and 1999 according to creek, sub habitats within creeks, and release size category.......................................................................................58 2-2 Nomenclature for classification of release si tes, release systems, release sub-systems and release habitats........................................................................................................... .59 2-3 Numbers of snook released according to release system and year. 1997-1 and 1997-2 represent two release expe riments performed in 1997..........................................60 2-4 Release sites, dates and sampling schedule at the release sites.........................................61 2-5 Snook captures organized by sampling type and locations. These data were used to verify results from short-term sampling a nd monitor recruitment to adult habitats..........62 2-6 Variables used for the principle component analysis examining characteristics of the release sites and associated observed short-term post-release responses (first summer and winter after releas e) of hatchery-reared snook age-1 snook.......................................63 2-7 Variables used for the principle component analysis examining both short(days-atlarge <= 365) and long-term (days-at-lar ge > 365) post-release responses of hatchery-reared snook stocked in 1998 and 1999 according to release site characteristics.....................................................................................................................64 2-8 von Bertalanffy growth parameters for snook recaptures (n) sourced from different release habitats. Length and age data only included recaptures of hatchery snook atlarge for 180d or longer.....................................................................................................65 2-9 Numbers of snook recaptured at various distances (km) away from their respective release sites. Data are only for snook recaptu red within the first year after releases.......66 3-1 Results from standardized sampling in April 2002 and physical attributes of experimental creeks (shoreline di stance and percent altered habitat)..............................104 3-2 Mark-recapture results for the four expe rimental creeks organized according to sample month. Hatchery 2x and Wild 2x represent snook that were recaptured in two subsequent sampling events..................................................................................105 3-3 Numbers of hatchery-reared snook released at the four experimental tidal creeks. Abundance estimates by creek for the age1 wild snook before releases (April 2002) are provided with the expect ed and observed (in June 2002) percentage of hatchery snook in post-release collections.....................................................................................106 7

PAGE 8

4-1 Summary of studies examined regarding common snook diets.......................................135 4-2 Numbers of snook processed for stomach contents and observed instances of cannibalism from winter data in nursery habitats (i.e. from Adams and Wolfe, 2006, and Sarasota data, 2004)..................................................................................................136 4-3 Linear regression statistics for various morphometric comparisons: vertical gape (VG), body depth (BD, anal spine (AS), thro at to anus (TRH-AN), standard length (SL), and fork length (FL)...............................................................................................137 4-4 Number of replicate experimental treatments (N) of field enclosure trials examining cannibalism according to predator size, the inclusion of additional prey (bait), snook prey density, and two prey size categories (e xpressed as a percent of predator size).....138 4-5 Details of actual cannibalism occurrences found in snook stomachs collected in the field........................................................................................................................139 8

PAGE 9

LIST OF FIGURES Figure page 2-1 Map of experimental release sites located off the southwest coast of Florida USA. Release creeks are denoted, and release sub-habitats (U= upstream, M=midstream, L=lower stream, and I=barrier island habitat) are enclosed in transparent shaded ovals ......................................................................................................................... .....67 2-2 Release numbers (grey bars) and post-release results showing percent hatchery snook found in total snook catch (graph A) and recapture ratios (number recaptured/number released) adjusted fo r effort (from 21 m seine, graph B)...................68 2-3 Mean recapture ratios (number recapture d/number released x100) of snook from 1998 and 1999 releases, by sub-subsystems......................................................................69 2-4 Recapture ratios (number recaptured/numb er released x 100) of snook by release habitat. Data are from recaptures collected within the first year after release. Mean recapture ratio is indicated by the horizontal line..............................................................70 2-5 Linear regression models using mean catch -per-unit effort during winter of age-1 snook to predict short(< 1 year post-releas e) and long-term (>=1 year post-release) recapture ratios (rratio; number recaptured/number released x 100).................................71 2-6 Mean growth rates from fish released in di fferent habitats and s ub-habitats. Data are from recaptures collected 6-10 months after releases in 1998, and 1999..........................72 2-7 Size-at-recapture for snook released in different habitats. Data are from recaptures collected 6-10 months after re leases in April 1998, and 1999...........................................73 2-8 Mean size-at-release and growth rates of hatchery-reared snook recaptured during the first winter after release (6-10 mo.).............................................................................74 2-9 Length frequency distributions of wild a nd hatchery snook at microhabitat release sites for each season.......................................................................................................... .75 2-10 Plot of first two principle components for s hort-term data related to characteristics of the release sites.............................................................................................................. ....76 2-11 Plot of growth rate versus recapture ra tio (number recaptured/number released*100) from standardized seine collections du ring the first winter after release...........................77 2-12 Catch per unit effort (CPUE) of wild and hatchery snook less than 300 mm fork length according to release habitat and asso ciated mean growth rates (mm/d) through the first winter after release................................................................................................7 8 2-13 Plot of the canonical variab les (triangles) for size-at-release (SAR mm FL) and associated response variables. Releas e sites are plotted as shaded dots...........................79 9

PAGE 10

2-14 Mean dispersal distances (km) organized by size-at-recapture (RecapFL, graph A), and size-at-release (SAR, mm FL, graph B)......................................................................80 2-15 Plot of length and age data from hatchery snook recaptures and associated vonBertalanffy models.......................................................................................................81 2-16 Mean dispersal distances from rearing hab itats (minimum distance from release site to recapture site) of snook recaptured with in seasons (spring, summer, fall winter) and by snook age (years)....................................................................................................82 2-17 Plot of the first two principle components comparing long-term responses (recaptures at-large for over 1 year, black dots) of the released fish, short-term responses (from recaptures at-large le ss than 1 yr., also black dots), and characteristics of the rel ease sites (shaded dots)................................................................83 2-18 Numerical and biomass projections to adult stages of juvenile hatchery-reared snook released in various rearing ha bitats with characteristic gr owth rates by first winter.........84 3-1 Map of experimental study sites along the coasts of Sarasota and Manatee Counties, Florida 107 3-2 Schematic of various depletion-removal me thods including (a) t wo-bank depletion, (b) single-bank depletion, and (c ) two-bank extended depletion..............................108 3-3 Graphical representation of individual capture probabilities (p) of snook from various depletion-removal methods (circles, triangles and squares, graph a) and the likelihood profile for pcommon (graph b)...........................................................................109 3-4 Abundance estimates of wild (grey area) a nd hatchery (white area) age-1 snook over time in treatment creeks receiving high a ugmentation levels (top graphs) and low augmentation levels (bottom graphs)...............................................................................110 3-5 Results from post-release samples collect ed in June 2002 showing catch-per-unit effort (graph a) and relative abundance (g raph b) of hatchery snook (white bars) to wild juvenile (grey bars) abundance (set at 100%)..........................................................111 4-1 Map of Sarasota and Manatee counti es where snook collections and stock enhancement research has been performed. Arrows indicate the tidal creeks where snook were captured and subsequently examined for stomach contents.........................140 4-2 Photograph of enclosures where cannibalism trials occurred..........................................141 4-3 Schematic of model process predicting cannibalism-related mortality by the various age groups of snook.........................................................................................................142 4-4 Annual relative abundance of age-1 and adult snook standardized around age-0 abundance........................................................................................................................143 10

PAGE 11

11 4-5 Graphical comparisons of s nook morphometrics (in mm)..............................................144 4-6 Photograph of a 70 mm (TL) and a 178 mm (TL) common snook. Black spot at tip of dorsal fin is obvious in smaller snook, but sp ot is nearly absent in larger snook.............. 145 4-7 Graphical results from field enclosure tr ials describing mean cannibalism rates of small and large snook according to various prey treatments...........................................146 4-8 Plot of body lengths (fork length FL) of snook cannibals and cannibalized prey...........147 4-10 Monte-Carlo simulation results (n=10,000 ite rations) showing proportions of age-0 mortality due to age-specific cannibalism.......................................................................148 4-11 Monte-Carlo simulation results (n=10,000 iterations) showing the proportion of age0 abundance (at the beginning of the model period) cannibalized by the various age groups of snook................................................................................................................149

PAGE 12

LIST OF ABBREVIATIONS area covered by a standardized depletion net number of snook captured number of snook remaining in a standardized seine sampled area age-i age specific estimate of the num ber of snook in a given cohort i field cannibalism rate, i.e. observed number of age i snook with cannibalized prey in their st omachs / total number of age i snook stomachs examined AS anal spine length (mm) b scaling parameter of MW function BC Bowlees Creek BD body depth (mm), measured at the widest part of the body c mortality due to cannibalism (number of individuals) CPUE catch-per-unit effort, CWT coded-wire tag Dc density (number per 30 m of shoreline) of juvenile snook in the creeks D.O. dissolved oxygen (mg/l) FL fork length (mm) HB Hudson Bayou HC total creek shoreline distance I Barrier island habitat k vonBertalanffy growth coefficient L Lower stream mouth habitat Linf vonBertalanffy parameter, asymptotic length M Midstream habitat Mu annual natural mortality at unit weight 12

PAGE 13

MW weight specific instantane ous natural mortality rate NC North Creek N Ct total in-creek juvenile snook abundance at time t PC Phillippi Creek pcommon the conditional maximum likelihood estimate for binomial sampling for a all depletion removal trials p it proportion of age i snook in a given tidal creek at time t pmax the conditional maximum likelihood estimate for binomial sampling for a particular depletion removal trial s seined area of a creek SAR Size-at-release (FL, mm) SC South Creek SL standard length (mm) snook common snook ( Centropomus undecimalis ) t time in days t0 vonBertalanffy parameter, theoreti cal time when fish length = 0 TIDY Tidy Peninsula, also known as Tidy Island TIDY_I Barrier island counter part to TIDY TL total length (mm) T0 U Upstream habitat W snook body weight (or mass, in g) WB Whitaker Bayou WtCi typical weight of indivi dual prey that age group i would cannibalize Wti the weight of an individual in September from cohort i VA-AN the distance from the ventral gi ll arch (hypohyal bone ) to the anus 13

PAGE 14

VG vertical gape, defined as the vertic al distance between the anterior tip of the premaxlilla and the anterior tip of the dentary bone VIE visible implant elastomer 14

PAGE 15

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy HABITATAND DENSITY-MEDIATED INFLUENCES ON SNOOK ECOLOGY: LESSONS LEARNED FROM MANIPULATIVE RELEASE EXPERIMENTSWITH HATCHERYREARED JUVENILE SNOOK By Nathan Paul Brennan August 2008 Chair: William J. Lindberg Cochair: Kenneth M. Leber Major: Fisheries and Aquatic Sciences Common snook Centropomus undecimalis stocks in Florida, USA are experiencing increasing pressure from the recreational fishi ng industry and habitat lo ss. This dissertation investigates components of snook stock enhanc ement research, as a potential supplementary management tool. As juvenile stages are typica lly used for stocking, I examined influences of habitat and ambient snook density on post-release fates, and the potential for cannibalism in juvenile stages. Chapter 2 examines post-release growth, surviv al, and dispersal of j uveniles stocked into rearing habitats, and their numer ical and biomass contributions to adult stocks. Growth was inversely related to recapture ratios and rele ase sites producing low growth had the highest numerical contributions to adul t stocks, but highest biomass contributions were from sites producing intermediate growth rates. In Chapter 3, manipulations of release magnit ude of hatchery-reared juveniles showed tidal creeks with high augmentation levels (n=2) result ed in overall post-release abundance increases (10 months post-release) but imperceptible chan ges in low augmentation creeks (n=2) of wild and hatchery conspecifics. Disproportionate post-release loss (85%) of hatchery-reared snook 15

PAGE 16

16 occurred in one high augmentation creek but was not detected in the other. No loss of wild conspecifics was detected in any of the creeks. Snook are cannibalistic and show evidence fo r this behavior intensifying during early stages. In Chapter 4, I examin ed ontogenetic differences in canni balism potential and its role in controlling abundance. Morphologi cal comparisons (anal spines, mouth gapes, and confusion colorations [eyespots]) showed that predatory defense and compensatory feeding adaptations decreased with size. Examination of existi ng research on snook diet coupled with model simulations showed that intera nd intra-cohort cannibalism intens ified in nursery habitats and during winter when older ag e groups were present. Overall this work used released snook to qua ntify tradeoffs in production to adult stocks based on characteristics of snook rearing habitats. Within an age-1 cohort, stocking increased overall abundance but sometimes with high post-release mortality in hatchery snook. Intraand inter-cohort cannibalism was focuse d on age-0 snook in rearing habita ts, and could be influential on inter-annual recruitment dynamics. These findings provide valuable input for stakeholders evaluating snook stock enhancement potential in Florida.

PAGE 17

CHAPTER 1 INTRODUCTION Global declines in coastal habi tat quality, ecological diversity, and fishery stocks continue to concern stakeholders and resource managers Anthropogenic activities are thought to be the primary causative agents (i.e. over harvest and habitat degradation, Lotze et al., 2006), and efforts to reverse or slow this trend have b een largely unsuccessful. Accurate knowledge of ecosystem-wide function and processes, and spec ies-specific ecologica l knowledge must be coupled with responsive and adaptive correctiv e actions to restore ecosystem health. Florida has one of the highest rates of huma n population increase in the United States (Campbell, 1997). With a high coast to land-area ratio, much associated activity and development is concentrated al ong Floridas coast, which can in fluence the coastal ecosystems (Bruger and Haddad, 1986, Lorenz, 1999; Lorenz and Sarafy, 2006). Unfortunately, shallow coastal systems also are some of the most diverse and productive systems, making them particularly important for conservation and manageme nt that target critical biological habitats and associated resources. Many valu able fishery stocks in these sa me areas are also the focus of increasingly popular recreational fisheries. Floridas USD 5.4 billion recreational fishery is easily the largest in the USA (American Sportfishing Association, 2004), and is known internationally as a hot spot fo r recreational fishing. Traditiona l management measures are used to control harvest and harvest laws are continuall y adjusted to maintain resource levels in the face of the recreational fishing popularity. Alongside traditional management practices, Fl oridas managers and stakeholders have directed efforts at stock enhancement to aid in restoring depleted stocks. Stock enhancement seeks to augment the natural supply of juvenile s to overcome recruitment limitation by releasing animals produced in aquaculture se ttings into the wild (Bell et al., 2008). This approach is 17

PAGE 18

becoming increasingly popular among stakeholde rs and managers worldwide (Honma, 1993; Bell et al., 2006; Bell et al., 2008) probably owing to its seemi ngly direct application (Travis et al., 1998). Unfortunately, direct evidence of augmentation effects in localized demes, populations, and ecological communities is spar se (Hilborn, 1998; Travis et al., 1998; Leber 1999; Bell et al., 2006). While economically important stocks e xperience high harvest rates and are declining (FAO, 2004), and many stocks ha ve collapsed from overfishing (Steele, 1992; Hutchings, 2000), responsive changes in communi ty structure (Carscadde n et al., 2001; Steneck, 2006) and ecosystem process (Pauley et al., 1998; Jackson et al., 2001) along with alteration and loss of critical habitats (Vitousek et al., 1997) has added considerable uncertainty toward the feasibility of stock enhancement programs. Asid e from biological impacts, broader measures of the overall feasibility of stock enhancement programs that weigh economic, social, and institutional influences have been few (Garaw ay et al., 2006; Lorenzen, 2008) and add to the difficulties and uncertainties of this approach. Nonetheless, advances in culture technology of marine animals since the 1970s, coupled with a more focused scientific approach has rapidly progressed stock enhancement research (Blankenship and Leber, 1995; Munroe and Bell, 1997; Immamura, 1999; Leber, 1999). Major progress has occurred through improving fitne ss of hatchery-reared animals, tagging and subsequent understanding of releas e techniques and recruitment of stocked animals to harvested and spawning populations, genetic implications for stock enhancement, and understanding biological, economic and social criteria necessa ry for stock enhancement programs to be feasible. Improvements in fitness of ha tchery-reared animals has provided substantial progress in the field through an understanding of behavioral defici encies of hatchery-reared animals (Olla 18

PAGE 19

and Davis, 1998; Hossain et al., 2002; Masuda, 2004; Davis et al., 2004), and hatchery selection effects (Berejikian et al., 1999) Proactive steps toward ma naging fish health and disease transmission from hatchery to wild stocks are al so important contributions in this field (Kennedy et al., 1998; Mushiaki and Muroga, 2004). Tags (or marks) are crucial elements in pr oviding informative feedback necessary for evaluating the efficacy of stock enhancement programs (Blankenship and Leber, 1995; Walters and Martell, 2004). Typically, st ock enhancement programs require tags to be applied to many juvenile fishes with minimal harm yet have high information content, low associated costs, and have high retention rates throughout the lifetime of the organism (Blankenship and Leber, 1995). A variety of tag types have been used to m onitor stocked individuals including genetic tags (Bravington et al., 2004; Jorstad, 2004; Sekino et al., 2005), chemical and thermal marks (Volk et al., 1999; Jenkins 2002), internal tags (Davis et al., 2004; Brennan et al., 2005), electronic archival tags (e.g. Prentice et al. 1990), and external tags (Stoettr up, 2002). The use of tags along with the application of controlled expe rimentation and assessment of enhanced stocks has provided quantitative feedback to substantia lly progress the field (L eber et al., 2004). Specific improvements in stocking strategies include determining op timal size-at-release (Tsukamoto, 1989; Yamashita et al., 1994; Leber, 1995), release season, and size-season interactions (Leber et al., 1997; Leber et al., 1998; SanchezLamadrid, 2002; Gwak et al., 2003), and release habitat (Solazzi et al. 1991; Russell et al., 2004; Davi s et al., 2005; Andersen et al., 2005) and acclimation to the rel ease sites (Brennan et al., 20 06; Fairchild, 2008). Eventual recruitment of stocked animals to adulthood and to fishery landings has also been documented (Svsand et al., 1990; Kitada et al., 1992; Leber and Arce, 1996; K aeriyama, 1999; Zohar et al., 2008). 19

PAGE 20

Recent progress has also occurred through understanding the broader potential for stock enhancement (Taylor et al., 2005; Garaway et al ., 2006; Lorenzen, 2006; Lorenzen, 2008). This involves an integrated analysis of biological, economic, and social criteria for suitable candidate species for stock enhancement activities. Much of this work directly in corporates stakeholders priorities and needs that are pr imarily responsible for driving mo st stock enhancement activities. Models are needed to evaluate the candidacy of species targeted for stock enhancement considering what is known about life history tr aits and other parameters of the stock (e.g. fecundity, recruitment dynamics, de nsity-dependent growth and surv ival), cost effectiveness of hatchery production, fishery economics and logist ics, how it integrates with other forms of fishery management, and intuitional goals and priorities (Lorenzen, 2008). These types of proactive evaluation measures are needed prior to commitments of large investments in stock enhancement programs (Hilborn, 1998; Walters a nd Martell, 2004; Lorenzen, 2008). This has been the exception rather than the norm for mo st stock enhancement programs however (Bell et al., 2006; Bell et al., 2008). The inclusion of st akeholders in the pro cess (e.g. Garaway et al., 2006; Uki et al., 2006; Becker et al., 2008; Tomiyama et al., 2008; ), however, is a positive step towards understanding and direc ting stock enhancement and othe r fishery management programs toward adaptive management approaches (Walte rs, 1997; Perry et al., 1999; Bell et al., 2008). Less understood and less demonstrated, yet key ar eas of uncertainty in this field, include density-dependent effects, genetic effects, a nd economic feasibility. Numerical responses to stocking in wild and hatchery-sourced animal s (e.g. intraand inter-specific competition or cannibalism) and interactive eff ects within and across generations are rare (Leber et al., 1995), and remain a high priority (Leber et al., 2004; Walters and Martell, 2004 ; Lorenzen, 2005). Studies examining the effects of stock enhancement programs on inter-annual and spatial 20

PAGE 21

variation in recruitment across community and eco systems that aligns with current ecosystembased management programs (e.g. Pauley et al., 2000) are also a high priority. Genetic effects and consequences of releasing hatchery-reared animals on wild populations have been predicted (e.g. Ward, 2006; De Innocentiis et al., 2008; Tr ingali et al., 2008a) and serve as useful guidelines for stock enhancement programs. App lication of these guidelines, however, remains a challenge and adds to the difficulties in achieving cost-effective programs (Lorenzen, 2008). Preliminary economic analyses generally predic t that high costs of hatchery production and inflexible social implications of hatchery production often preclude economic success (Hilborn, 1998; Walters and Martell, 2004; Bell et al., 2006; Lorenzen, 2008). Optimizing costeffectiveness of stocking programs by improving post-release survival (Leber et al., 2005; Brennan et al., 2006) are positive steps toward this, and creative methods for production may also be necessary to make stock enhancemen t programs feasible (e.g. Japanese scallops Patinopecten yessoensis are produced in hanging cages in situ bypassing the need for hatcheries [Uki, 2006]). As with all forms of fisheries ma nagement programs, challenges lie in how stock enhancement programs can adjust to adaptive ma nagement measures based on dynamic feedback from biological, economic and social sources (Wa lters, 1997; Walters and Ma rtell, 2004; Bell et al., 2006; Lorenzen, 2008). The common snook Centropomus undecimalis is one of the focal species of stock enhancement research aimed at supplementing Flor idas existing stock management programs. Common snook are an estuarine-dependent catadrom ous fish found in tropical and subtropical habitats of the eastern coasts of the Ameri cas, including Florida. Along with substantial alteration of snook habitat (B ruger and Haddad, 1986), the in creasing popularity of the recreational fishery has caused concern for managers and stakeholders. Floridas snook 21

PAGE 22

populations are considered overf ished and stocks are below management goals (Muller and Taylor, 2005). Despite increa singly restrictive fishery regulations on common snook, annual fishing mortality rates have steadily increased over the last 20 years, and recruitment has generally declined (Muller and Taylor, 2005). Hi gh fishing pressure, coupled with relatively few spawner-sized females in the a dult population, has made manage ment of snook stocks difficult (Muller and Taylor, 2005). Furthermore, snook a ssociate with shoreline habitat (Marshall, 1958; Gilmore et al., 1983, McMichael et al., 1989; Pete rs et al., 1998a) and thus depend on coastal waters that may be subjected to intense anth ropogenic influences. Collectively, overfishing and habitat loss have caused a gene ral decline in the popul ation although the relative influence of these is unclear (Mulle r and Taylor, 2005). In 1985, the Florida Fish and Wildlife C onservation Commission (FWC) and Mote Marine Laboratory (MML) ini tiated a long-term stock enhan cement research program with common snook. This was followed a decade later by advancements in snook husbandry technology that facilitated the production of su fficient numbers of hatchery-reared snook juveniles to support experimental releases at a reasonable scale (thousands of individuals). Based in Sarasota, Florida, snook stock enhancement research has prioritized a responsible approach (sensu Blankenship a nd Leber, 1995) coupled with intensive field evaluation to provide experimental feedback. Since 1996, over 52,000 juvenile hatchery-reared snook have been released through a variety of e xperimental manipulations. All released snook were tagged with coded-wire tags (CWT) to iden tify experimental releas e groups from juvenile through adult stages of the rel eased snook (Brennan et al., 2005). Visible implant elastomer (VIE) tags were also used to pr ovide external identification of hatchery-reared snook to aid in fishery returns and provide benign evaluation of basic release groups (B rennan et al., 2005). 22

PAGE 23

Researchers found that snook acclimated in predato r-free enclosures located at release sites had about twice the survival rates co mpared to those released wit hout acclimation (Brennan et al., 2006). Ongoing sampling of the localized snook popul ation has documented survival to adulthood of the released hatchery-reared juveniles. Since 1999, individuals from release cohorts have been regularly documented in annu al wild snook spawning aggregations. Both male and female hatchery-reared snook have been recaptured with mature gonads, and some recaptured individuals have outgrown the harvestable size slot (27-33 inches total length [TL, or 686-838 mm TL]) (Leber and Brennan, unpublished data ). Releasing juvenile snook into a variety of habitats has provided valuable understandi ng of essential fish habitat (as defined in the 1996 amendments to the Magnuson-Stephens Act) fo r juvenile snook (Chapter 2). Quantitative sampling in juvenile and adult snook habitats ove r the years also allowed researchers to measure relative contribution rates of various juvenile rearing habitats to adult demes, and documenting associated migratory patterns from juvenile to adulthood (Chapter 2). Mark-recapture data from hatchery-reared and wild snook us ing a variety of tag technology has provided important insight into short and long-term dispersal patterns (Chapter 2 and 3; Pine et al, unpublished data). Using hatchery-reared juvenile snook to experimentally manipulate ambi ent juvenile abundance in tidal creeks researchers were able to examine how dens ity-dependent responses within a cohort varied between habitats and across seasons (Chapter 3). Stomach content examination of wild and ha tchery-reared snook from field collections has also provided valuable insight into snook dietary preferences, stomach content evacuation rates Brennan and Heinlein, unpublished data), and diel feeding periodicity (Rock et al., unpublished data). With evidence for densit y-dependent survival in juvenile snook, the 23

PAGE 24

potential for cannibalism was also examined using laboratory and field st udies (Chapter 4). Associated research has also measured natural and fishing mortality rates in slot-sized snook (Bennett, 2006), and provided an important unders tanding of seasonal and annual variability in microhabitat use by adult snook (Marcinkiewicz, 2007). This dissertation represents an integral part of the FWC/MML project and is focused on understanding snook life history and ecology, especia lly in juvenile stages characteristics of snook rearing habitats, and productive capacities of these habitats in response to variations in inter-annual recruitment fluctuat ions. More specifically, in Chapter 2, using theoretical perspectives of size-structured growth, survival, and movement, and rigorous monitoring of the stocked fish I examined evidence for habitat-spec ific tradeoffs in post-release growth, survival, and dispersal of the stocked fish. I related th ese post-release responses to characteristics and conditions at the rearing habitats Long-term monitoring of the st ocked fish for years after the releases showed that patterns in growth and surviv al established within the first year after release were correlated to characteristics of the reari ng habitats, and continued to be reflected in the stocked fish as they matured. Coupling this with related work on post-rele ase survival (Chapter 3), I was able to predict numerical and biomass contributions of sp ecific rearing habitats to adult stages, then examine differences and tradeoffs be tween growth and survival of the stocked snook based on rearing habitat characteristics. In Chapter 3, I examined the potential for de nsity-dependent mortalit y in juvenile snook cohorts. I used releases of j uvenile hatchery-reared snook to ar tificially manipulate localized recruitment and abundance across diff erent habitats. Depletion-rem oval trials were first used to calculate the efficiency of a st andardized seine for sampling, th en, coupled with an intensive standardized sampling program, in-creek abundan ce of juvenile snook was measured. Two tidal 24

PAGE 25

creeks received high augmentati on release treatments, and two creeks received low augmentation treatments. Abundance was measured before and through the first year after the experimental releases. This study focused on augmentation e ffects within a cohort of juvenile snook, but future work should also examine inter-cohort predation and mortalit y and how inter-annual natural variation on ambient snook abundance co uld influence similarly staged release experiments. As snook are piscivorous and have shown ev idence for density-dep endent growth and mortality, in Chapter 4, I examined the potential influence of cannibalism in controlling juvenile snook abundance. I examined published and unpublished literature on snook diets, and supplemented this with my own examinations of Sarasota snook stomach contents. Then I used these data to provide input into a model I de veloped to understand the role of cannibalism in snook abundance dynamics and to provide insight into conditions that promote cannibalism. Overall, this work provided impor tant insight into understanding th e potential and feasibility of stock enhancement with common snook stocks in Florida, and provides a better understanding of snook ecology, recruitment patterns, and characteri stics of essential fish habitat for snook. 25

PAGE 26

CHAPTER 2 INFLUENCE OF REARING HABITAT ON GROWTH, SURVIVAL, DISPERSAL, AND RELATIVE STOCK CONTRIBUTION OF COMMON SNOOK Centropomus undecimalis Introduction In fish, individuals often make ontogenetic habitat shifts over their lifetime to maximize growth and overall fitness (Wer ner and Gilliam, 1984). For example, catadromous fishes spawn in marine waters, but rear in neashore and of ten lower-salinity environments making them ideal organisms for investigating ecological and physiol ogical consequences of habitat shifts and the evolutionary traits acquired to deal with those co nsequences. In juvenile stages, predation threat is primarily linked to body size (Werner and Gill iam, 1984) and is a primary motivational driver behind behavioral tradeoffs between spending time in refuge habitat or ri sky foraging to grow (Werner and Gilliam, 1984; Walters and Martell, 2004) Foraging arena theory (Walters and Juanes, 1993; Walters and Marte ll, 2004) states that predation refugia are the principal microhabita ts occupied by young recr uits and that fish abundance is strongly mediated by food availability within and near refuge, with higher mortality in foragers that move out of refugia. Variation in food res ources, forager density, predation threat within and outside of re fuges, and interactions among thes e can strongly influence growth and mortality rates (e.g. Connel and Jones, 1991; Tupper and Boutilier, 1997; Walters and Kitchell, 2001; Bystrom et al., 2004 ). Hi gh forager density can reduce localized food availability, and increase competition, foraging time and subsequent predation risk (Walters and Juanes 1993, Walters and Martell 2004). As a result, increased density leads to both reduced growth and increased mortality (e.g. Love et al 1991, Persson and Elkov, 1995, Jones 1997, Madon et al. 2001, Seitz et al. 2005 ). Foraging arena theory pred icts greater su rvival for postlarvae and juveniles that recruit to refuge hab itats with higher forage availability and reduced predation threat, than in habita ts with lower food availability and greater predation threat. 26

PAGE 27

This study, conducted in Sarasota and Manatee counties of southwestern Florida, USA, explores the influence of settle ment habitat (via releasing hatc hery-reared juvenile snook into various habitats) on growth, survival, and disper sal and recruitment to adult demes of common snook Centropomus undecimalis. Common snook (herein referre d to as snook) are a tropical to sub-tropical estuarine-dependent piscivorous fish found along the A tlantic Coast of the Americas. In Florida, snook stocks are subjected to intense recreational fi shing pressure (Muller and Taylor, 2005) and have suffered dramatic coastal habitat alteration (Bruger and Haddad, 1986). Snook are also sensitive to cold waters (Storey and Guder 1936; Ma rshall 1958; Shafland and Foote 1983) and Florida represents the northern limit of their latit udinal range (Marshall, 1958). As a result, Floridas snook populations undergo annual temperature-dependent range expansions and contractions. During mild winters, snook larval are dispersed north of the 29th parallel and successfully surviv e (e.g. Linton and Richards, 1965) but during harsher winters, cold water (<14 C) generally causes snook mort ality north of this (Storey and Gruder, 1936; Shafland and Foote, 1983). In Sarasota and Manat ee counties, Florida (latitu de range = 27 7 to 27 28 N) snook sometimes suffe r mortality during cold winter s, but local populations find thermal refuge in warmer tidal creeks (Marci nkiewicz, 2007; N. Brennan, unpublished data). Snook spawn in high salinity (>28 ) seawater in inlets and tidal passes of estuaries, mouths of rivers and canals, and off sandy beaches primar ily from May August (Marshal, 1958; Volpe, 1959; Tucker and Campbell, 1988; Taylor et al., 1998). Hatching occurs within 17-18h after spawning (Lau and Shafland, 1982; Tucker, 2003) and within two w eeks recruitment to nearshore habitats is thought to occur by larvae taking advant age of a two-layered estuarine circulation system (Norcross and Shaw, 1984; Tolle y et al., 1987) and vertical positioning within the water column (Peters et al., 1998b). Po st-larval snook are found mostly along vegetated 27

PAGE 28

shorelines of shallow brackish tidal creeks, can als and lagoons in both low(riverine) and highsalinity (mangrove swamp and salt marsh) habitats (Tolley et al., 1987; Pe ters et al., 1998a). Juvenile snook rearing habitats are diverse, but include estuarine streams and canals, barrier island and mainland marshes. Juveniles occur in salinities betw een 0 and 40 along moderately sloping shorelines, in basin depths of 1 m or less, mud or sand substrate, and various forms of shoreline vegetation and structure (H arrington and Harrington, 1961; Fore and Schmidt, 1973; Gilmore et al., 1983; McMich ael et al., 1989; Peters et al., 1998a). In Sarasota and Manatee counties, snook rearing habitats are located in tidal creeks and estuarine backwater habitats. During winter, snook of all sizes are abundant in these habitats (especially tidal creeks) as they probably serve as thermal refugia (N. Brennan, unpublished data). During winter, McMichael et al. (1989) sampled at fixed stations within age-0 snook rearing habitats in Tampa Bay, have evidence for small-scale repositioning of the cohort during cold winter months. After winter, Gilmore et al. (1983) hypothesized that adolescent sn ook from east-central Florida populations (150 400 mm SL, mean = 240 mm SL) undergo an ontogenetic habitat shift to seagrass beds. Adult snook are also known to disperse to seagrass beds estuarine inlets, and beaches by late spring and early summer (Marshall, 1958; Gilmore et al., 1983; Taylor et al., 1998). In this study, I tested the following hypothe ses: (1) Snook juveniles remain in their settlement habitat until maturation. (2) Snook that disperse out of se ttlement habitats return to those same habitats, suggesting either homing behavior or short disper sal distances to nearby habitats. (3) Mangrove habitats along shorelines of barrier islands afford as good a foraging and refuge habitat for juvenile (late age-0, age-1) snook as do mangrove habitats in tidal creeks. (4) 28

PAGE 29

Relative contribution of juven ile habitats to adult abundance is proportional to relative abundances of young snook cohorts in those habitats. Using hatchery-reared snook as proxies for w ild snook (about 8-11 m o. old at release), I tested the hypotheses by comparing growth, surv ival, and dispersal patterns in the rearing habitats (i.e. release habitats). I also examined how these early-on patterns in the various rearing habitats were reflected in long-te rm recruitment to adult stocks. The release of hatchery-reared fish into the wild can be a useful approach to test ecological hypothe ses (Leber et al., 1995; Leber, 1999; Miller and Walters, 2004; Loren zen, 2005). Although hatchery fish generally suffer greater initial post-release mortality (wit hin weeks) than wild fish (Lorenzen, 2000; Fleming and Petersson, 2001; Jonsson et al., 2003; Brennan et al., 2008), by contrast no consistent differences are reported for growth (S vsand et al., 2000; Bohlin et al., 2002; Jonsson et al., 2003; Lochet et al., 2004; Dahl et al., 2006). In this st udy, I assumed that the relative growth, survival, and recruitment patterns, established early on in rearing habitats, was reflective of wild juvenile snook found in th e same locales, and reflective of eventual recruitment patterns to adult stocks. Methods Experimental Design I used a replicated release-re capture experiment to test th e effects of release site, and size-at-release (SAR), on growth, survival, dispersal, and recr uitment to adult populations. Hatchery-reared juvenile snook (80-270 mm fork length [FL], 8-11 months old) were tagged and released into various coastal habitats in Saraso ta and Manatee counties, in southwest Florida, USA (Table 2-1). Release sites were categorized into sub-syst ems along a salinity gradient of typical tidal creek (system) (Table 2-2): (1) upper reaches of the creek with typically low-salinity 29

PAGE 30

water (U), (2) midstream reaches in brackish wa ter habitats (M), (3) lower reaches in higher salinity waters near the creek mouth (L [lower site]), and (4) along barr ier island high salinity habitats (I) nearest the respective tidal creek (T able 2-1, Table 2-2, Figure 2-1). Experimental releases in sub-systems were repeated across the following release syst ems (or tidal creeks): South Creek (SC), North Creek (NC), Phillippi Creek (PC), Hudson Bayou (HB), and Bowlees Creek (BC), (Table 2-1, Figure 21). Releases also occurred at Tidy Peninsula (TIDY) a high salinity mainland mangrove-lined peninsula and a nearby barrier island. Study Habitat and Timeline The experimental area encompassed about 30 km of coastline located in Sarasota and Manatee Counties, on the west co ast of Florida, USA (Figure 21). Stocking occurred during April 1998 and 1999 (Table 2-1). In 1998 we stocked the four sub-habitats of four tidal creeks SC, NC, PC, and BC, plus two s ubhabitats of TIDY (Table 2-1). In 1999 we stocked the same sites as in 1998, but replaced PC with HB, a high salinity developed bayou and its adjacent barrier island habitat (Figure 21), and stocked only one sub-system at TIDY. Releases in 1999 were replicated with a second pulse of releases at the same sites two weeks later (Table 2-1). Barrier island release habitat (sub-system=I) were characterized by high salinity water (typically 30-34 ), with mangroves vegetating the shorelines (i.e. Rhizophora mangel Avicennia germinans, and Laguncularia racemosa). Water depths along the shoreline were generally 1 2.5 m at low tide ( tidal variation at all sites was approximately 0.6 m), with substrates of sand, shell, and mud. At the tid al creek mouths (sub-system=L) salinities were typically more variable due to seasonal ra in and tidal flux, but generally around 25 32 Here, shoreline vegetation was also primarily ma ngroves. Water depths at the mouth of NC (non-dredged habitat) were generally < 0.5 1.0 m, but 1.0 2.0 m at the other sites (dredged). Substrate was comprised primarily of mud, sand, and live oyster bars ( Crassostera virginica ). 30

PAGE 31

Midstream creek habitats (sub-system=M), are in fluenced by tidal flux and creek outflow, and waters were generally less saline. Mangrove, live oak ( Quercus virginiana ), Juncus sp., and Brazilian pepper ( Schinus terebinthifolius ) predominated shoreline vegetation. Water depths in NC and SC (primarily non-dredged habitat) were generally <1.0 m, and at PC and BC (dredged habitat) were about 0.5 2.0 m. Substrate at th e midstream sites was primarily mud, clay and sand mixtures. Low salinity waters (generally 0-15 ) typified upstream tr ibutary habitats (subsystem = U), but these sites were still influenced by tidal flux. Shoreline vegetation consisted of some mangrove, but more live oak, leather fern (Acrostichum aureum ), Juncus sp., and Brazilian pepper. Water depths were generally <1 m (e xcept for BC; about 1 2 m deep) and substrate was primarily mud, sand, and some clay. Hatchery Fish Production, Tagging and Release Wild adult brood stock snook were captured annually from May September with large seines (140-m to 160-m long, 3-m deep, 3-cm multifilament nylon mesh) along barrier island passes and beaches where snook aggregated for spawning activities each new and full moon. Snook were held in floating net pens, then individuals were sexed, and eggs and milt were collected by gently squeezing the abdominal area Snook eggs were fertilized with milt in 250 ml beakers, transferred to pre-filtered sea wa ter in oxygenated plastic ba gs, and transported to 2,200 l recirculating seawater tanks (salinity = 28-32 ) at the laboratory. Tank salinity was gradually lowered to 3-6 by 30 d after hatching. After this, juvenile sn ook were sorted by size on a regular basis to prevent cannibalism and r eared until tagging (Main, unpublished data). Snook were tagged according to release site, release lot (within-y ear replication over time), and SAR (small 80-129 mm FL, medium 130-179 mm FL, and large 180-270 mm FL). Tagging occurred in April 1998 and 1999 (about 811 mo. after hatching for each respective release year). Individual snook were tagged with single-length coded-wire tags (CWT, as in 31

PAGE 32

Brennan et al., 2005) to identify in formation about release batches (e.g. release site, release date, and size class, but no unique information on indivi duals however). These snook were also fitted with visible implant elastomer marks as a simp le external identifier of hatchery-reared snook (Brennan et al., 2005). During ta gging, sub-samples of snook from each release group were taken to obtain mean lengths and weights accord ing to release batch. After tagging, juveniles were held in tanks with recirculating brackish water for 1-24 d until release. On each release date, snook were taken from their tanks, checked for tag retention, and then loaded into transport tanks. Transport to the release sites took between 2 h depending upon access conditions at the release sites. At mo st upstream and midstream release sites, snook were transported in aerated 150 l coolers, or 800 l hauling tanks by truck, whereas at the stream mouth and island habitats snook were transpor ted by boat in 800 1200 l tanks. If necessary, transport salinities we re gradually adjusted en route to match release site salinity. Snook were either released directly from transport tanks, from 75 l containers, or 3.5 m long canoes partially filled with water into shorelin e habitat. At each release site, snook were stocked along 100-300 m of shoreline during daylight hours, between 0900 1400. The number of fish stocked varied among size-classes and sites, but was generally between 80 and 300 (fish per date, site, and size class: Table 2-1). Stocking al so occurred in 1997 and 2000-2004 in similar creeks as part of other stock enhancement studies (Table 2-3). Collection of Environmental Variables at Release Sites Aside from the experimentally imposed habi tat classifications, I collected environmental variables at each release site to quantify differences among the rel ease habitats aid in predicting growth and survival rates of j uvenile snook. I measured the fo llowing habitat parameters from three collection sites at each release site: bank slope, substr ate softness, percent organic composition and percent moisture of the substrat e, percent shoreline composed of vegetation, 32

PAGE 33

and percent shoreline co mposed of mangrove vegetation. Fo r standardization purposes, I took samples at perpendicular distances away from the mean high tide elevation indicated on the bank. I estimated mean high tide elevations on the bank as the upper limit of barnacle, limpet, or algae distribution on shoreline structure. S horeline slopes (rise/run) were calculated using vertical measurements to the bottom at 1m perp endicular from the bank starting at the mean high tide elevation. Substrate softness was measured 2 m perpendicular from the banks mean high tide elevation. To do this a 0.75 cm2 x 1.5m cylindrical fiberglass rod was held in a vertical position, resting on the substrate. Then, a 4.1 kg standard weight was placed on the top-side of the rod and the subsequent penetration distance (cm) of the rod in to the sediment was measured. Substrate core samples (3.8 cm di ameter x 10 cm depth) were al so collected 2 m perpendicular from the banks mean high tide elevation at each s ite. Substrate cores were analyzed for percent organic material and moisture (by weight; Lenore et al., 1998). I also visually estimated the percent coverage of macro-vegetation (e.g. trees grasses, shrubs) along an approximately 100 m stretch of the shoreline centered on the collection site. Similarl y, for the same stretch of given shoreline, I also estimated the percen t coverage of mangrove vegetation (i.e. Rhizophora mangel, Avicennia germinans, or Laguncularia racemosa ). From each core sample, I removed an approximately 15 ml subsample after thorough mixing to obtain a homogeneous mixture. Each subsample was placed in an aluminum pan, weighed (mg), and then dried at 105oC for 24 H. After drying, subsamples were reweighed (to obtain a percent moisture content), then ashed at 550oC and weighed again to obtain the percentage of organic mate rial in the sample. I also collected water quality measurements (temperature [0C], salinity (), dissolved oxygen [mg/l], and pH) with a YSI 556 handheld multi-parameter model (YSI incorporated, 33

PAGE 34

Yellow Springs, Ohio, USA) at a minimum of three seine haul locations within each sub-habitat. These were collected during standardized samp ling events within the first month after each release. For these sites, wa ter quality measurements were ta ken both at the surface and bottom of the water column. Short-Term Collections After the 1998 releases, we performed standardized sampling efforts in May, June, July and November 1998, and in January and March 1999 (Table 2-4). After April 1999 releases, we sampled in May, June, July, August, November 1999, and January 2000 (Table 2-4). From May October 1998, a standardized sampling effort ( one unit) consisted of nine hauls with a 21.3 x 1.8 m bag seine (1cm nylon mesh) and 18 casts with a 2.4m radius cast net (1 cm monofilament mesh) at each release site. During November 1998, we modified the sampling effort and gear at M, L, and I sub-systems to increase tag returns. At these sub-systems we performed six hauls with a 45.7m x 3m bag seine (1cm nylon mesh) and three hauls with a 21.3 x 1.8 m (1cm nylon mesh) bag seine. We did not sample with a cast ne t. Because of the narrowness of the streams at the upstream sites (sub-system=U), use of the 45.7m x 3m bag seine was impossible, so we continued sampling with our earlie r sampling regimen of cast nets (18 casts) and a 21 m seine (9 hauls). Because of this change in methods, I on ly used 21m standard seine data for comparisons between upstream and other sub-systems, and only co mpared recapture ratios from data collected during winters for year-toyear comparisons. Long-Term Collections Recaptures over the long-term (d efined as recaptures at large < one year) were collected from 1999 2006 to validate patterns observed in the rearing habitats (i.e. growth and recapture ratios), and to document relative contribution rates of stocked juveniles in the various rearing habitats to adult stocks. Generally, snook do not recruit to adult spaw ning habitat until age 3, 34

PAGE 35

when they can be found in high concentrations in coastal passes a nd along beaches during summer months (Taylor et al., 1998). Two year old snook have also been documented with mature gonads (Taylor et al., 1998) and snook 150400 mm SL (typical age two sizes) have been documented to move to seagrass habitats duri ng spring and early summer (Gilmore et al., 1983; Brennan, unpublished data). Therefore I used data from annual collections at beaches (barrier island coastlines along the Gulf of Mexico), ba rrier island passes, and intercoastal seagrass habitats. Sampling at beaches, passes and seag rass habitats was conducted with seines (137 m long x 2.4 m deep, with 5 cm multifilament nylon mesh), cast nets (4 m radius, 5 cm multifilament mesh) and trammel nets (137 m long x 2.4 m deep, 3 cm inner nylon mesh, and 30 cm outer nylon mesh). We also performed standardized sampling efforts at randomly selected coastal sites (mainland and barrier islands about 1.6 kilometers of shoreline). In September 2001 and July 2005, red tide outbreaks caused mass mortality of snook along beaches and passes. During these events, we scanned individual carcasses for CWTs. I also used data from snook captured in an a nnual angling tournament specifically created and designed to evaluate contributi on rates of hatchery-released s nook to localized adult stocks. These tournaments were conducted from 19982005, and anglers were directed to fish between Anna Maria Island and Venice Inlet (Figure 2-1, Table 2-5). Any snook, captured by an angler, was measured (FL mm) and scanned for CWTs. Tagged snook were kept for CWT extraction, untagged snook were released alive. For analysis of long-term data, I removed migrational bias from recapture ratio comparisons by only using recaptures from beaches and passes (adult snook spawning habitat) and those captured along shoreline habitat at least 1 km away from any given release site (because sampling at locations near release sites could result in higher recapture ratios of snook 35

PAGE 36

released nearby). Hatchery snook captured within the same creek as their release creek were not included in the analysis. I considered recaptured snook that had met these criteria as dispersed and were used in the re capture ratio analysis. Sample Processing and Data Analysis All captured snook were counted, measured, a nd checked for the pres ence of CWTs with magnetic CWT detectors (Northwest Marine Technology Inc., Seattle WA). Untagged snook were released, while tagged hatchery-reared sn ook were preserved on ice and returned to the laboratory where lengths and wei ghts were taken. Coded-wire ta gs were extracted from tissue and read using a binocular micros cope (at 60x). Each tag was read twice by independent readers to verify tag codes. Recaptured fishs tag c odes were linked to releas e data through a releaserecapture database system using Microsoft Ex cel, Microsoft Access, and SAS (Brennan, 1997). Recapture ratios for ea ch release group (i ) were calculated as: Recapture ratioi = (number recapturedi /tag retentioni) / number releasedi (2-1) where tag retentioni is the proportional tag rete ntion of fish from group ( i ). I assumed CWT retention estimates at the time of release to be valid after release thereafter because other studies have shown that loss of CWTs implanted in muscle tissue was insignificant after 30 d (Blankenship, 1990; Dorsey 2004; Meeuwig et al ., 2007; Brennan et al ., 2007), and in our studies with snook (Brennan et al., 2005), CWT re tention estimates one year after tagging, were not significantly different from initial CWT rete ntion estimates taken 1-30d after tagging. Data on short term growth rates (mm/d) we re collected from snook recaptured during the first winter after release (6-10 m onths after release). To calcula te growth rates, I subtracted mean SAR (FL mm, based on tag codes) from in dividual recapture lengt hs (FL mm) divided by days at liberty. Because of th e short-time period betw een release and recapt ure dates, I assumed that a linear relationship (slope = mm/d) was adequate to describe growth during this period. 36

PAGE 37

To examine growth over the long-term, I fit von Bertalanffy growth models to length and age data of snook released in different release ha bitats (e.g. BC, NC, or islands, Table 2-2) with least squares regression techniques using Microsoft Excel and th e solver add-in. Age of recaptured snook was estimated by adding 300 d to the days since release, because snook were about 10 months old at the time of release. For th is analysis I only used data for recaptures more than 180 d since release (to ensure that potential habita t influences on growth had occurred) from release experiments in 1997-2004 (Table 2-3 for release information). Dispersal distances were cal culated as the minimum in-water distance (km) between release and recapture site s within the confines of the water body, measured to the nearest 0.1 km. Because snook spawn on an annual basis (general ly during summers), I was able to use length frequency distributions to differentiate betw een age-0 (or young-of-year snook), age-1 (those spawned the summer prior to the age-0 group, same age class as the released snook), and older snook. I then calculated catch-p er-unit-effort (CPUE) of th e age-1 cohort captured during standardized sampling at the release sites during th e first summer and winter after each release. Age-1 CPUE was calculated separately for each standard seine (21 m and 45 m). Multivariate analyses: To understand how various habita t characteristics were related to the selected release sites and how these relate d to short-term responses in the released snook, I performed a principle component analysis also using SAS procedures (SAS, 1999). The PCA was based on the correlation matrix of environmental variables collected at the release sites (see above section), estimates of mean CPUE for age1 snook collected the time of release (May-July, summer) and during the first winter after the release (November), and first winter mean growth rates and recapture ratios of th e released fish (summarized in Table 2-6). To quantify the relationships between release va riables (SAR and release site) and response variables (size at 37

PAGE 38

recapture, mean dispersal distances from release sites [summer and winter separately], firstwinter growth, and first-winter recapture rati os, I performed a canonical correspondence analysis using SAS (SAS, 1999). For this, recapture data were organized according to mean SAR (FL mm) and release site. I also conducted a second principle component analysis to examine how components of the rearing habitats and short-term patterns in grow th and recapture ratios were related to long-term response pa tterns in the released snook (l ong-recapture ratios and growth variables, Table 2-7). Predicting contributions of rearing habitats to adult stocks : Because first winter growth and recapture ratios were strongly correlat ed with long-term growth and recapture ratios, I used these data as the basis for predicting contributions of rearing habitats to adults stocks. I used a linear function of loge (ln) transformed first winter growth rates (growth mm/d) to predict loge transformed first winter recapture ratios usin g least squares procedures and the solver add-in feature in Microsoft Excel. While recapture ratio s were only reflective of relative survival rates for any particular release group, I used data from a subsequent st udy (Chapter 3; Brennan et al., 2008) conducted in 2002 in many of the same habitats to provide estimates of real survival rates of the released snook. In Brennan et al., (2008, Chapter 3), abundance of age-1 snook was estimated prior to, and subseque ntly through the first year afte r release. While post-release survival is a function of ambi ent density and stocking magnitude (among other things), I used estimates of survival to first winter from creek s that received similar stocking magnitudes (about 1000-2000 stocked) as the releases during 1998 an d 1999. Therefore, the high augmentation creeks from the 2002 releases (BC and Whitaker Ba you [WB]) were similar in magnitude to the 1998 and 1999 release years. Of these two creeks WB had very little loss of hatchery snook by the first winter (about 0.8 5 of the snook released in May survived to first winter) and I used this 38

PAGE 39

rate as a best case survival scen ario for released fish. An im portant difference in 2002, however, was that we used an in situ acclimation release technique which has been shown to approximately double the survival rates of stocke d snook (Brennan et al., 2006). Therefore, to standardize with recapture ratios from 1998 and 1999 releases (non acclimated releases) I used half this rate (i.e. 0.42 for surviv al to first winter) as a best case scenario. Highest recapture ratios from 1998 and 1999 were set at 0.42 and others were down-scaled accordingly. After obtaining abundance estimates at first winter, I used 0.9 as the proportion survived from first winter to second summer, and 0.7 for subs equent annual survival rates. To project biomass contributions of juvenile snook released in specifi c rearing habitats to adult stocks in subsequent summ ers, I first plotted growth rate s at first winter versus the associated Linf parameter (FL mm) from von Bertalanffy growth functions specific to release habitats (Table 2-8). Subsequent parameters (k and t0) for predicted von Bertalanffy growth functions were calculated as linear relationships to Linf. After this, I calculated length (FL mm) at age for 2 5 summers after the release. Lengths (FL mm) were converted to body weight (Wt, g) using the function, Wt = 0.000009*(FL)2.9892 (2-2) with parameters fit from length and weight data from Sarasota s nook recaptures (Brennan, unpublished data, FL =176-816 mm, n=738) by least squares procedures using Microsoft Excel and the solver add-in. Individual weight pred ictions for the respective summers were multiplied by numbers-at-age estimates to generate biomass estimates. For this analysis I ignored differential fishing mortality normally subjected to snook of these sizes recruiting to the fishery. Results In April 1998 and again in April 1999 a total of 21,392 juvenile snook were tagged and released (Table 2-1). Mean tag retention in the laboratory by rele ase year and lot varied between 39

PAGE 40

83.3% and 97.2%. From April 27, 1998 to August 5, 2005, we sampled 21,304 snook along shorelines and estuaries from Venice Inlet to Anna Maria Island, in Sarasota and Manatee counties, representing about 40 km of coast. Of these, 480 were hatchery snook recaptures from snook released in 1998-1999: 437 contained CWTs that were de coded, 18 CWTs were lost during processing (3.76%), 7 recaptures contained CW Ts but were released alive after recapture, and 18 snook had only VIE tags. Of the 437 tags recaptured and decoded from this study, 202 were from the April 1998 releases and 235 from 1999. Recapture data described below are only from deciphered tag codes. Short-Term collections at Release Sites General : Sampling at release sites from Ap ril 27, 1998 April 30, 2000 yielded 3,085 snook. Of these 262 (8.49% of th e snook catch) were hatchery re ared tagged snook: 93 were from releases in April 1998 (0.78%), and 169 fr om April 1999 (1.78%). Ov erall during the first month after release, ha tchery snook from each respective release represented 25.4 27.4% of the snook catch, but gradually declined to 0.0 9.0% 6-8 months later. Recapture ratios also declined over time but were be tween 0.01% 0.05% per 21m seine haul within the first month after release (Figure 2-2). Highest mean recapture ratios were from fish released in upstream and midstream subsystems) (Figure 2-3). Fish released in North Creek had recapture ratios about twice as high as any other tributary in both 1998 and 1999 even though sampling effort was similar among all sites (Figure 2-4). Also mean recapture ratios of snook released in creeks were higher than those of snook released along islands. Recapture ratios were genera lly correlated with age-1 snook abundance (CPUE 21 m seine) (Figure 2-5). Overall, released snook showed high fidelity to their particular release site. Generally, 60% of the recaptured snook were recaptured at thei r release site and 89% were no farther than 5 40

PAGE 41

km from their release sites (Table 2-9). Fo r snook released in BC, PC, NC, and SC, 30.5% ( n=262) were recaptured at other release sites with in the first year afte r release, but no clear patterns of upstream or downstream movement we re apparent. For the same period, only nine (of 262 recaptured, 3.4%) were captured outside of their respective creek within the first year after release. For the barrier island release si tes, only one recapture (of 16 total, 6.3%) was found away from its release site. The farthest migration documented within the first year after release was from a fish released at TIDY that emigrated to the upstream section of NC, 24 km away. Generally, first winter mean growth rate s showed strong pattern s of differentiation according to release site (see belo w for statistical analysis). Highest mean growth rates were found in snook released at barrier island si tes (0.67-0.78 mm/d) followed by BC (0.67 mm/d), TIDY (0.63 mm/d), HB (0.52 mm/d), NC (0.44 mm/ d), SC (0.44 mm/d), and PC (0.41 mm/d) (Figure 2-6). Sizes-at-recapture during the first winter showed similar patterns with largest sizes from snook released at islands, TIDY, and BC, and smallest sizes at recapture from NC and SC (Figure 2-7). There was no appa rent relationship between SAR a nd first winter growth rates, however (Figure 2-8, and see below). Notable differences were found in patterns of snook size structures among the release sites and sub-sites throughout the seasons (Figure 2-9). Juvenile snook were common throughout the year in upstream, mi d-stream, and stream mouth sub-habitats of tidal creeks, but rare at barrier islands exce pt during summer and fall mont hs (170-300 mm FL, Figure 2-9). Large snook (>350 mm FL) predominated frequenc y distributions at ba rrier islands throughout the year, but were most abundant here during summe r and especially fall months. During winter and spring, relative abundance of large snook increased in the creeks par ticularly at stream 41

PAGE 42

mouth habitats compared to in-creek abundance during summer and fall. Among creeks, length frequencies of snook were very similar with respect to sub-habitats. Multivariate analyses with short-term responses: The principle component analysis of release site variables an d short-term post-release responses s howed that the first two principle components explained 27% and 21% of the standa rdized variance and 67% was explained by the first three principle components. A plot of the first two principle components (Figure 2-10) shows that recapture ratios and age-1 densities were inversely related to growth rates at first winter (Figure 2-10, Figure 2-11, Figure 2-12). High CPUE and recapture ratios were positively related to sediment softness and moisture conten t but negatively related to growth, salinity, pH, dissolved oxygen, and mangrove coverage (Figure 2-10, Figure 2-11). Release sites with high CPUE and recapture ratios, but low growth includ ed NCU, NCM, and NCL. Release sites with low CPUE, low recapture ratios, but high growth, salinity, dissolved oxygen and mangrove coverage included SCA, NCA, TIDY_I, and TIDY. Among our release sites, bank slopes (steepness) were inversely related to vegetative coverage, sediment organics, and temperature. Considering bank steepness and vegetative coverage, BCU had steep banks and low vegetative coverage, but PCL, PCA, and NCL had more gradual bank slopes and more vegetative shoreline coverage. The canonical correspondence analysis showed that mean SAR according to release site, was most correlated to summer (0.77) and winter (0.73) recapture lengths, but not well correlated with mean dispersal rates during summer (0.33) or winter (0.28), or first wi nter growth (0.08), or recapture ratios (-0.33). A plot of the canoni cal variables for SAR and the response variables (Figure 2-13) reflects the strongest correlations between SAR and recapture lengths but much less correlation between the other responses. While canonical vari ables were not calculated for 42

PAGE 43

all release sites (becau se of missing response data), the ca nonical variables for TIDY, SCL, and HBL showed strongest correlation to SAR, but snook released in North Creek showed no associations with SAR (Figure 2-13). Just examining SAR a nd dispersal distances however, snook released in the largest size class (and larger size-at-recaptur e) showed patterns of greater dispersal distances compared to the smaller size classes (Figure 2-14). Long-Term Assessment General: From August 1, 1998 August 31, 2005 sampling at beaches, passes, and barrier island habitats and sampling at re lease sites from April 1, 2000 to August 31, 2005 resulted in an additional 18,344 snook captured (Tab le 2-5). Of these, we collected 2,883 snook from beaches and passes, 1,038 from random sampling at barrier island sites, 1,453 snook from red tide collections, 1,063 from local snook tournaments, and 11,907 snook from sampling at release sites. Of the snook released in 1998 and 1999 (at upstr eam, midstream, stream mouth, and along island shorelines), 174 were recaptured af ter over 1 yr (0.95% of total snook captured). Most long-term recaptures were captured at their release sites (n=132), but 42 had dispersed (i.e., moved >1 km). Overall, the largest re captured hatchery snook was 816 mm FL (34 inch total length) and weighed 5,492 g (12.1 pounds). The longest time at liberty was 2,399 days (6.57 years). Long-term Growth: I obtained age estimates of 514 ha tchery-reared snook at large for 6 mos. or longer (maximum age 7.39 years). Of these, four release habitats had over 40 recaptured snook (5-7 age cohorts, n=59 recaptures from BC, and n=249 from NC) and I used these to generate age-at-length von Bertalanffy growth models (Table 2-8). Fastest growth was predicted for snook released at barrier island habitats (Linf = 824), followed by snook released in Bowlees Creek (Linf = 704), South Creek ((Linf = 657), and North Creek (Linf = 552) (Figure 215). 43

PAGE 44

Dispersal: Generally mean dispersal distances incr eased over seasons and years (Figure 2-16). Snook recaptured during summer and fall had the highest mean dispersal distances, and were rarely recaptured in tidal creeks during these seasons, although sampling occurred regularly there. Mean dispersal distances during wint er did not increase across years however, and showed high rearing site fidelity each winter (F igure 2-16). Snook tended to recruit to adult spawning habitat nearest their rearing habitats. Mulitvariate analysis with long-term responses: The second principle component analysis comparing key features of release sites and short-term post-release responses with longterm responses (Table 2-7) showed the first th ree components explained 80% of the standardized variance (36%, 28%, 16% respectively). Examini ng a plot of the first two components (Figure 2-17), age-1 summer CPUE, age-1 winter recapture ratio, and l ong-term recapture ratios (from dispersers and those captured at beaches and pass es) were all within the upper left quadrant. Shortand long-term growth patterns were closely re lated and generally i nversely related to recapture ratios. Contribution to adult stocks: Least squares procedures predicted recapture ratios as a linear function of loge (ln) transformed first winter growth rates ( growth mm/d) as: Recapture ratio = -1.9596* Ln (growth) 0.2578 (2-3) The correlation coefficient for this function was 0.59. Predictably, plots of numbers of hatcheryreleased snook recruiting to adult stocks according to age and rearing habitat (expressed as shortterm growth rates), shows highest contributions from rearing hab itats with lowest growth rates and highest post-release survival (Figure 2-18a). Recr uitment of stocked juveniles from rearing habitats to age-2 adult stocks ( Nage-2, to second summer) were expressed as: Nage-2 = -286 (ln(growth)) + 81 (2-4) 44

PAGE 45

Based 1000 hatchery-reared juvenile s stocked in rearing habitats during early summer and the above assumptions, Creeks such as NC and SC (with first winter growth rates of 0.44 mm/d), were predicted to produce about 316 age-2 snook while a barrier island release site stocked at the same level would produce about 171 age-2 snook (Figure 2-18a). Under these conditions, and not accounting for fishing mortality, the numbers of age-4 snook produced by NC or SC would be about the same as the number of age-2 s nook produced by an island habitat stocked at the same level. Total biomass estimates of hatchery -released snook recruiting to adult stocks based on rearing habitat growth rates showed a humpshaped relationship (Figure 2-18b). Rearing habitats with intermediate first winter grow th rates of stocked snook produced the highest biomass of adult snook. Biomass production of ag e-2 hatchery snook from rearing habitats as NC or SC with slow growth rates produced abou t the same biomass of age-2 snook as did island sites where growth rates were much higher (Fig ure 2-18b). By age-4, however, about 60% more biomass was predicted for NC or SC s ources compared to island sources. Discussion In this study I used stock enhancement to understand important life hi story characteristics and habitat interactions of juvenile snook stocke d into various rearing habitats and examined contribution rates of the stocked j uveniles to adult stages. Due to the vulnerability of Floridas snook stocks to overfishing (Muller and Taylor, 2005) and widespread modification of coastal rearing habitats (Bruger and Haddad, 1986; Lo renz, 1999; Davis et al., 2005b; Lorenz and Sarafy, 2006) investigations of stock enhancemen t potential is warranted. Notwithstanding, the controversial nature of this measure coupled with the status of snook stoc ks in Florida (Muller and Taylor, 2005) emphasizes the need for a cautionary approach for stock enhancement research (Blankenship and Leber, 1995; Walte rs and Martell, 2004, Lorenzen, 2008). This 45

PAGE 46

study, operated on an experimental level, and co upled with extensive post-release evaluation, provided valuable information on juvenile snook life history characteristics tied to various characteristics of rearing habitats and eventual contribution of these fish to adults stocks. Releasing juvenile snook into varied habitats showed evidence of tradeoffs between growth and survival. Sites where snook growth rates were highest also had comparatively low recapture ratios, suggesting low survival. Conve rsely, suppressed growth was exhibited at sites with high recapture ratios (high survival) and high overall densities (CPUE) of juvenile snook. Generally, habitats with higher ambient densities of wild juvenile snook correlated with higher survival of hatchery-reared snook, but at the cost of density-dependent growth. This suggests superior refuge habitat and overall lower pred ation threat, but also high competition for prey items among the juveniles. Beverton-Holt mode ls predict density-depen dent mortality with increasing population size. In fish, adjustme nts in growth responses usually occur under predation threat as a probable tradeoff against mortality risk (Werner et al., 1983; Belk, 1995; Belk, 1998; Krause et al ., 1998). In this study, the differen tial post-release gr owth rates and recapture ratios probably reflect variations in refuge quality, predation threat, and competitive interactions. The inverse relationship between growth rate s and recapture ratios of stocked juvenile snook was among the strongest qua ntitative correlations I found. Relationships among release habitat features and snook production were more complicated, however, a nd reflects the diversity of juvenile snook habitat (Gilmor e et al., 1983; McMichael et al., 1989; Peters et al., 1998a) and Florida estuaries (Faunce and Sarafy, 2006). Because mangrove or vegetative shoreline coverage was not well correlated with recapture ra tios or growth rates, other features such as sediment softness, and organic content are mo re predictive of snook pr oduction in Sarasota 46

PAGE 47

rearing habitats. The inverse relationship be tween bank slope and vegetative coverage probably reflects common anthropogenic alte rations in the form of dredgi ng and vegetative removal. Rearing habitat production and stock enhancement considerations: Rearing habitats with characteristic differences in attr ibutes of early-on post-re lease responses of the stocked juveniles showed remarkable differences in numerical and biomass contributions to adult stages. Beck et al. (2001) de fined a nursery habitat as those that produce a grea ter number of individuals per unit area that recr uit to adult populati ons than the average production of other habitats where juveniles occur. Others have presen ted theoretical implications for numerical contributions of nursery habitats to adult stocks (e.g. Ross et al 2003; Dahlgren et al., 2006), but field studies demonstrating this are rare. In fi sh that attain large body sizes, are highly fecund, and have variable growth and survival rates (s uch as snook, Taylor et al ., 1998), differences in numerical and biomass contributions could be substantial. In Sarasota, the model showed that equal levels of hatchery-reared juveniles stocked into NC and SC (sites with low growth rates) produced more age-2 individuals per unit area than other habitats but contributions to age-2 biomass production (i.e. potential egg production) were among the lowest and similar to barrier island release sites that produced snook with high growth rates, but had low recapture ratios. Biomass production per unit area was hump shaped (Figure 2-18b) and maximized at release sites with intermediate growth rates. As age of the recruiting hatchery snook increased, the maximum biomass began to shift towards rearing ha bitats with lower growth rates again. Using this work, resource managers can assess the rela tive production value (in terms of snook) of various rearing habitat. Quantifi cation of total habitat area is an important next step to evaluate overall production according to habitat type. 47

PAGE 48

While growth rates and survival are func tions of production, I i gnored other important considerations that should not be overlooked. As snook attain larger si zes, fishing mortality increases and places a disproportionate pressure on larger-sized snook (Muller and Taylor, 2005). This pressure must be accounted for. While adult biomass is an index of potential recruitment, snook are protandric hermaphrodite s (Taylor et al., 1998) and further complicate stock recruitment models (Muller and Taylor, 2005), it remains an important consideration. Snook rearing habitats must also be quantified if total produc tion by habitat type is to be considered (as outlined in Dahlgr en et al., 2006). Quantification of rearing habitats might be accomplished with rough habitat measures based on GIS mapping systems, but some measurements (e.g. measuring organic content in sediments) would clearl y require more effort. Coupling post-release survival estimates fr om snook released in similar creeks in 2002 (Brennan et al., 2008; Chapter 3) to this st udy are not without comp lication. Post-release survival and growth are, in part, functions of ambient density of snook j uveniles in the rearing habitats, refuge quality, stocki ng magnitude, predator and prey abundance, encounter rates, and feeding rates. Within release sites, I assumed that ambient densities of juvenile snook were generally the same during the two studies. This might be reasonable because I observed similar post-release growth responses across the years (e.g. North Creek has had relatively high recapture rates and relatively low grow th rates across the years 1997, 1998, 1999, 2000, 2001, Brennan, unpublished data) when the same habitats were stocked. Within a given year, however, snook recruitment can vary across rearing habitats so re lative measures across the years at a particular rearing habitat ma y not be representative of recru itment in other rearing habitats. In 2005 red tide blooms (Karenia brevis) in Sarasota Bay may have caused disproportionately lower recruitment strength in nearby rearing habitats (Brennan unpublished data). While its 48

PAGE 49

difficult account for, physical changes in shorelin e habitats (e.g. installation of docks, seawalls, associated dredging activities, growth or rem oval of vegetation, or pulses of organics from upstream sources) may also influence year-to-year results. Increases in experience with this work (e.g. increases in the numbe r of experimental years, addi tional stocking levels, and across habitats) can increase th e confidence in our expectations of such activities, but should be coupled with experimental stock enhancement modeling scenarios (Lorenzen, 2005; Lorenzen, 2008). Strong localized differences in post-release growth and recapture ratios were found at release sites as close as 1-km apart. Thes e habitat-specific patterns, esta blished early-on, continued to be found in snook recaptures for six years after the release. Similar growth and abundance patterns are seen on a broader scale between snook stocks found along the east and west coasts of Florida. East coast stocks are generally faster growing, attain larger sizes, but are less numerous than stocks from the Gulf coast (Muller and Taylor, 2005). This emphasizes how various features of rearing habitat can influen ce juvenile snook production to adult stocks. Although it may be coincidental, in our study area, tidal creeks with high est growth rates but lowest survival were also highly anthropogenically impacted (s horeline habitats were highly altered through seawalls, channelization, and creeks were routinely dredged). Habitats along Floridas east coast are also c onsidered more developed (Muller and Taylor, 2005). Research is warranted to further understand the potential effects of differi ng anthropogenic development on valuable coastal habitats and their influence on stock-recruitment dynamics for snook stocks and other important coastal stocks. Because wild age-1 snook density was st rongly correlated to hatchery-reared snook recapture ratios and growth rates, these relationships probably hold true for wild snook, although survival estimates of wild conspecifics are probably higher. Hatche ry-released animals are 49

PAGE 50

generally more vulnerable to mortality compared to wild conspecifics of the same sizes (Mason et al., 1967; Lorenzen 2006; J okikokko et al., 2006; Shimizu et al., 2007; Fritts et al., 2007). This has been attributed to a variety of behavi oral and other deficienci es due to rearing in psychosensory-deprived environments such as ha tcheries (Olla et al., 1998, Berejikian et al., 1999). Experimental releases of hatchery-reared snook in 2002 (B rennan et al.; 2008, Chapter 3) showed disproportionately high mortality within the first month af ter release compared to wild conspecifics. In an earlier study (Brennan et al., 2006), snook acclimated in predator-free enclosures for 3d had survival rates 1.92x higher than naive snook stocki ng directly into the wild. Therefore, highest mortal ity of stocked age-1 snook apparent ly occurs soon after release. Nonetheless, even if subsequent survival rates of wild and hatchery snook are similar, first winter recapture ratios would continue to reflect the in itial post-release mortality. Study is warranted to distinguish between real surv ival rates of wild and hatchery snook over time. Other: Water quality at release sites contai ning sediments with high percent moisture and percent organic material was generally hypoxic with relativel y low salinity during summer. Generally these habitat types were found in sections of the tidal creeks (especially NC and SC). Aquatic habitats containing sediment with high loads of organic material typically experience hypoxic conditions due to oxygen consumption by microbial activity often most pronounced during summer (e.g. Kemp et al., 1992; MacPherson et al., 2007). While survival rates and overall juvenile snook density were generally hi gher at these sites, snook growth rates were suppressed. Growth responses were likely influenced by density-dependent competition for limited food resources (as above), but growth would likely be further suppressed by higher physiological costs associated with tolerance to hypoxic conditions. This supports physiological work (Peterson and Gilmore, 1991) demonstra ting that small snook (< 150 mm SL) are better 50

PAGE 51

able to tolerate hypoxic conditions than are larger snook and hypoxic sites may serve as an important physiological refuge from aquatic predat ors. Other attributes of the habitats (i.e. gentle bank slopes, soft sediment s, low salinity, and high temperat ures) were also typically found at these sites and other studies characterizing juvenile habitats corroborate this finding (Gilmore et al., 1983; McMichael et al., 1989; Peters et al. 1998a). Su mmer time water quality at island habitats correlated w ith higher dissolved oxygen levels, hi gher salinity, pH, and loosely correlated with steeper banks. Arguably, the low recapture ratios at these sites might reflect higher predation risk and mortality because mo re potential predators can tolerate these conditions. High predation risk could cause size-dependent survival where only the largest and fastest growing snook survive beyond the vulnerable si ze range. If this were true, however lower recapture rates of small SAR snook should have been observed at these sites. I found evidence that would support this in plots of recapture leng ths according to release habitats (Figure 2-7) but not in plots of growth rates and SA R, or the canonical correlation anal ysis (Figure 2-8, 2-13). Within a 42 d window after release, larger SAR individuals showed patterns of higher mean dispersal distances compared with smalle r individuals, and SAR was loosely correlated with first summer and winter dispersal rates. Other studies have shown that relative predation risk and dispersal associated with body size (e.g. Werner and Gilliam, 1984; Walters and Juannes, 1993; Armstrong et al., 1997). Smaller i ndividuals, in general, ar e at higher risk from predation than large individuals, especially outside of refugia, and movement is comparatively reduced. Lorenzen (2000) found a negative, linear relationship be tween mortality of released fish and length. In contrast, large fish benefit from improved foragi ng opportunities therefore increased growth rates. In th is study, patterns of in creasing dispersal dist ances with body size is only suggestive of lower predation risk in large fish. 51

PAGE 52

Size structure analysis from snook collected at the release sub-habitats year-round demonstrates the importance of tidal creeks as rearing habitats (or nur sery habitats), and important winter refuge for snook of all ages. Ge nerally, tidal creeks served as nursery habitats for snook throughout the year. During winter, tid al creeks experienced increases in adult snook abundance, with corresponding declines in a bundance at barrier isla nd habitats (highest abundance during summer and fall). Comparatively high winter abundance of snook in the creeks and associated temperature profiles suggests that creek habitat serves as a thermal refuge for snook (Marcinkiewicz, 2007). Furthermore, di ssolved oxygen levels in tidal creeks during summer aregenerally low (e.g. Kemp et al., 1992; MacPherson et al., 2007) compared to other times of the year, but dissolved oxygen levels al ong barrier islands are ge nerally high year round. These patterns imply that juvenile and adult snoo ks spatial distribution mi ght also be controlled by abiotic conditions. In the wint er, large-scale aggregation occurs in the creeks where water is warmer. During summer and fall months, tidal creek waters become hypoxic, and large snook are more common along beaches, passes (spawning habitats), and grass flats. The year-round presence of juveniles in tidal creeks (even during summer hypoxic conditions) coupled with recapture ratios suggests their potential im portance as refugia from predation. Important considerations for release experime nts can also be gained from natural size structures of snook in various habitats and acros s seasons. Snook are pisc ivorous (McMichael et al., 1989; Aliume et al., 1997) and cannibalistic (Tucker, 2003; Ad ams and Wolfe, 2006; Chapter 4). Resultantly, abundance of la rger snook could serve as a pr oxy for juvenile snook predation threat. With large snook present, juvenile s nook are at high cannibalism risk. Based on size structures, juvenile snook at ba rrier island habitat would be subjected to high predation threat year round. Low abundance of adults in cree ks during summer sugge sts lower associated 52

PAGE 53

predation threat for juveniles. Furthermore, rela tive abundance of wild ju venile snook could also be used as an indicator of suitable habitat fo r released fish, although density-dependent effects should be considered. During winter, snook releas ed in the creeks would be exposed to higher cannibalism risk due to high snook densities in the creeks at this time (see Adams and Wolfe, 2006; Brennan, unpublished data). These fish woul d also be only about six months old and smaller than spring-released snook. Over-winte r mortality (e.g. Hurst and Conover 1998) is also a major concern with snook because of their sus ceptibility to cold temperatures. For stock enhancement, these concerns must also be weig hed against rearing costs to determine economic feasibility. Overall, hatchery-reared snook showed hi gh site fidelity among creeks (or release habitats) and even for specific local es within creeks (or release sub-habitats). Within the first year after release, juvenile snook had emigration rates of only 3-6% outside of the release habitats. This agrees with an earlier study examining recaptures of juvenile snook released in NC (1% emigration from the creek, Brennan et al., 2006). During subsequent summers and fall (i.e. spawning months, Taylor et al., 1998), however, mean distances between release and recapture sites (i.e. dispersal distances) continued to increase with. Young mature snook tended mostly to disperse to spawning habitat and grass flats nearest their rearing habitats. Older and larger snook showed increasing dispersal dist ances during summer and fall. The increasing dispersal tendencies with increasing age have important implications for stock management. While young mature snook disperse to local spawning sites, larger snook dispersed more widely. Because the larger snook are more fecund (Chambers and Leggett, 1996; Berkeley et al., 2004; Longhurst, 2006; Francis et al., 2007), these fi sh may have a greater influence on snook recruitment dynamics and genetic patterns. Wide spread and intense harvest of these larger 53

PAGE 54

individuals may have negative consequences for snook stocks on both local and regional levels. During winter, however, mature snook showed cons istently high fidelity to rearing habitats (especially in tidal creeks) demonstrated through c onsistent use of these s ites each winter as they sought thermal refugia (Figure 2-16). Unde rstanding these snook beha viors and life history patterns are important for unders tanding how production and quality of rearing habitats can influence localized demes and fisheries. Recapture ratios showed some difficulties when used as in dices of survival because of uncertainties in post-release dist ribution of the juvenile snook. Our short-term sampling showed significantly higher recapture ratios from fish released at upstream and midstream habitats than those released at barrier islands These results may have been biased in several ways: (1) Sampling took place at the release sites. Thus, fo r each release creek syst em proportionally more sampling effort occurred within the creeks compar ed to barrier islands (about three times as much effort in creek habitat). (2) Barrier isla nd habitats represent a larger and more open habitat than within-stream ha bitats. Even if snook disperse d randomly, our sampling regimen would recapture fewer snook released at barrier islands, than within the creeks, because the creeks were more restrictive, with less overall dispersal area available. (3) Within the creeks, a general downstream movement of snook may ha ve resulted in disproportionately more recaptures of snook from upstream release locatio ns compared to stream mouth release sites. Very few juvenile hatchery snook were recaptured out side of the creeks, but this could have been due to less sampling effort in these habitats. A generalized downstream movement pattern was not found from our recaptured snook, however Often, snook released downstream were captured upstream and midstream, and snook released at the midstream habitats were found in 54

PAGE 55

the upstream, midstream and stream mouth habitats. Most often, snook released in a particular habitat were also recaptured there. Our long-term sampling program was designed to validate short-term samples at the release sites. Because snook disperse to adult spawning habitat during summer months, samples from these habitats should have provided the l east biased data for recapture ratio analysis independent of juvenile release habitat. In this study, of 6,437 snook sampled along beaches and barrier island passes (where snook were found sp awning, Table 3), only 21 snook (0.33%) were recaptured from these habitats. The low percentage of hatchery snook at large made it difficult to obtain a large enough sample size. Long-term samples (snook recaptures from any location that were at sea for over one y ear) resulted in similar data as the short-term data set with significantly higher recaptures of snook from midstream and stream mouth habitats compared to barrier island habitats. Because most of these snook were recaptu red in the creeks, geographical biases may have again influenced these data. These data correlate well with other findings of high site fidelity in snook. Volpe (1959) a nd Bruger (1980) conducted snook dispersal studies using tag returns from anglers a nd found that 70-79% of snook recap tures were within 8-9 km of the tag and release site. In these studies angler catches were not particularly linked to the tag and release sites. This correlates well with our fi ndings of 89% of our recaptu res within 5 km of the release sites. The higher percen tage in this study could be beca use our sampling stations were centered around release site s and less random than anglers capture sites. Conclusion: Floridas snook stocks are subjected to increasing fishing pressure (Muller and Taylor, 2005) and coastal rearin g habitats have been subjected to widespread ongoing habitat modification (Bruger and Haddad, 1986; Lorenz, 1999; Davis et al., 2005; Lorenz and Sarafy, 2006). Considering these resource management probl ems, coupled with considerable uncertainty 55

PAGE 56

in understanding stock enhancement potential, empi rical research such as this study, can aid in resolving these issues. In this study I found that rearing habitats where snook were stocked influenced post-release growth and recapture ratios at first winter and these traits were perpetuated for years after releases. Generally, post-release growth was inversely related to recapture ratios and tied to characteristics of th e rearing habitats. Sites with highest recapture ratios and low growth rates (often tidal creeks) generally contained soft sediment with high organic content and low dissolved oxygen levels in the water. Th ese sites were also correlated with high ambient densities of j uvenile snook. Sites with low recapture ratios and high growth rates generally had high salinity a nd pH, firmer sediment, and low ambient densities of juvenile snook (often barrier island habitat). I tied first winter recapture ratios from this study to real survival rates of stocked snook at first winter from another study (Brennan et al., 2008, Chapter 3) and used long-term growth data to project cont ributions of the various r earing habitats to adult stocks. Sites with highest re capture ratios and low growth made the highest numerical contributions to adult stocks, but biomass contributions were among the lowest. Hatchery-reared juveniles stocked at sites that produced intermedia te growth rates and reca pture ratios, made the highest biomass contributions to adults stocks. Follow up work is needed to understand interactive implications between stocking magnitude (Brennan et al., 2008, Chapter 3) and responses in hatchery and wild snook (across age classes). This work provides important information for resource managers considering th e productivity of various rearing habitats and has both fishery and habitat management impli cations. Follow up models examining stock recruitment and fisheries dynamics under various enhancement scenarios (e.g. Lorenzen, 2006) can be coupled with results from this study. T ogether, these can be powerful tools that provide essential information for responsible and adaptive management of these valuable stocks. 56

PAGE 57

57

PAGE 58

Table 2-1. Numbers of rele ased snook in 1998 and 1999 according to creek, sub habitats within creeks, and releas e size category. "Large" re fers to snook 180-270 mm FL, medium 131-179 mm FL, small 180-129 mm FL YearSeasonEstuarineLotSARRelease sub-habitat Total Tributary UMLI 1998SpringNorth Creek1Large 1000100150350 1Medium 01000150250 1Small 3083103103001228 Subtotal 4084104106001828 South Creek98-1Large 100100100150450 98-1Medium100099146345 98-1Small 2103002103001020 Subtotal 4104004095961815 Phillippi Creek98-1Large 3001502990749 98-1Medium3004503004001450 98-1Small 6106106102042034 Subtotal 1210121012096044233 Bowlees Creek98-1Large 3003003006011501 Tidy Island98-1Large ..0500500 Medium ..200002000 Lot 98-1 Total 232823202328290111877 1999SpringNorth Creek99-1Large 87888687348 99-1Medium 87868785345 99-1Small84858686 Subtotal 2582592592581034 South Creek99-1Large 87878787348 99-1Medium 861058686363 99-1Small85878787 Subtotal 2582792602601057 Bowlees Creek99-1Large 87878785346 99-1Medium 87878886348 99-1Small87878987 Subtotal 2612612642581044 Hudson Bayou99-1Large 87978787358 99-1Medium 90878786350 99-1Small 8700087 Subtotal 264184174173795 Tidy Island99-1Large ..3450345 99-1Medium ..85085 99-1Small ..2600260 Subtotals ..6900690 Lot 99-1 Total 104198316479494620 North Creek99-2Large 85878790349 99-2Medium 85878789348 99-2Small87918994 Subtotal 2572652632731058 South Creek99-2Large 788710186352 99-2Medium 78868686336 99-2Small 861028687361 Subtotal 2422752732591049 Bowlees Creek99-2Large 86017387346 99-2Medium 86878787347 99-2Small87878788 Subtotal 2591743472621042 Hudson Bayou99-2Large 86858797355 99-2Medium 86878687346 99-2Small 7800078 Subtotal 250172173184779 Tidy Island99-2Large ..4080408 99-2Medium ..4700470 Subtotal ..8780878 Lot 99-2 Total 100888619349784806 1999 Total 20491869358119279426 Grand Total 437741895909482821303 341 346 350 361 349 58

PAGE 59

Table 2-2. Nomenclature for cl assification of release sites, release systems, release subsystems and release habitats ReleaseReleaseRel easeRelease Habitat Site SystemSub-systemHabitat Type SC_USouth U South CreekTidal Creek SC_MSouth M South CreekTidal Creek SC_LSouth L South CreekTidal Creek SC_ISouth I Island Barrier Island NC_UNorth U North CreekTidal Creek NC_MNorth M North CreekTidal Creek NC_LNorth L North CreekTidal Creek NC_INorth I Island Barrier Island PC_UPhillippiU Phillippi CreekTidal Creek PC_MPhillippiM Phillippi CreekTidal Creek PC_LPhillippiL Phillippi CreekTidal Creek PC_IPhillippiI Island Barrier Island HB_UHudsonU Hudson BayouTidal Creek HB_MHudsonM Hudson BayouTidal Creek HB_LHudsonL Hudson BayouTidal Creek HB_IHudsonI Island Barrier Island BC_UBowleesU Bowlees CreekTidal Creek BC_MBowleesM Bowlees CreekTidal Creek BC_LBowleesL Bowlees CreekTidal Creek BC_IBowleesI Island Barrier Island TIDYTidy TIDY Tidy PeninsulaMainland Peninsula TIDY_ITidy I Island Barrier Island 59

PAGE 60

Table 2-3. Numbers of snook released accordin g to release system and year. 1997-1 and 1997-2 represent two release e xperiments performed in 1997 Release SystemRelease Years 1997-11997-2199819992000200120022004Totals Intercoastal Islands1,6671,7952,9011,92700008, Tidy Peninsula002,0001,56800003, Bowlees Creek 48909001,566008893,8417,685 Hudson Bayou0001,21700001, Phillippi Creek 1,3421,7683,6290 00006, North Creek 1,0098911,2281,5614,1831,63443643311,375 South Creek 08971,2191,58708241283695,024 Totals 4,5075,35111,8779,4264,1832,4581,4534,64343,898 290 568 217 739 60

PAGE 61

Table 2-4. Release sites, dates and sampling sche dule at the release sites. Dates highlighted in grey represent the original sampling regimen; all other dates represent the modified sampling regimen LocationRelease dates and sample months: Months after 1998 release: Months after 1999 release: Release1237911Release123479 South CreekApr-98May-98Jun-98Jul-98Nov-98Jan-99Mar-99Apr-99May-99Jun-99Jul-99Aug-99Nov-99Jan-00 North CreekApr-98May-98Jun-98Jul-98Nov-98Jan-99Mar-99Apr-99May-99Jun-99Jul-99Aug-99Nov-99Jan-00 Phillippi CreekApr-98May-98Jun-98Jul-98Nov-98Jan-99Mar-99Apr-99May-99Jun-99Jul-99Aug-99Nov-99Jan-00 Bowlees CreekApr-98May-98Jun-98Jul-98Nov-98Jan-99Mar-99 Hudson Bayou Apr-99May-99Jun-99Jul-99Aug-99Nov-99Jan-00 Tidy IslandApr-98May-98Jun-98Jul-98Nov-98Jan-99Mar-99Apr-99May-99Jun-99Jul-99Aug-99Nov-99Jan-00 61

PAGE 62

Table 2-5. Snook captures organized by sampling t ype and locations. These data were used to verify results from short-term sampling and monitor recruitment to adult habitats YearBeaches and PassesIntercoastalRed TideTournamentsRelease sites 1998 112 37 518 1999 349 603 481,904 2000 393 102 1911,071 2001 867 .1,103 1282,838 2002 431 1951,886 2003 489 13 1552,587 2004 354 208 3093,293 2005 350 195 888 Totals 2,8831,0381,4531,25814,985 Grand Total 21,617 62

PAGE 63

Table 2-6. Variables used for the principle co mponent analysis examining characteristics of the release sites and associated observed short-term post-release responses (first summer and winter after release) of ha tchery-reared snook age-1 snook stocked in 1998 and 1999 Release site SCUSCMSCLSCANCUNCMNCLNCAPCUPCMPCLPCAHBMHBLHBABCUBCMBCLBCATIDYTIDYA Age-1 summer CPUE0.0400.0770.0860.0770.2990.2690.1180.0280.0000.1510.1480.0190.0830.0000.0710.2940.1540.2310.0260.0630.000 Age-1 winter CPUE0.0000.0830.0000.2500.1410.2560.4580.1250.0000.2780.0000.0000.0000.3330.3810.0280.0670.0420.1850.000 Growth (mm/d) 0.340.470.740.440.410.520.71..0.290.79.0.520.690.670.670.63.0.620.68 R_ratio 0.0000.2100.7430.3591.6251.7130.9660.2650.0000.0000.0830.1660.0000.0001.1200.9760.2720.1100.0000.1230.400 Salinity (ppt) 3.326.233.734.96.513.730.332.60.38.230.032.134.133.835.832.133.332.936.033.435.13 Dissolved oxygen (mg/l)3.64.45.17.46.25.85.99.77.16.45.65.22.73.47.14.16.35.87.15.68.54 Temperature (celsius)28.831.229.630.830.829.232.631.931.831.532.533.428.828.527.730.830.730.130.228.830.84 pH 7.97.98.18.37.97.18.08.47.87.98.08.08.18.18.48.28.48.48.38.08.30 Substrate softness (cm)272215910283410264113233210722631917121 Bank slope (rise/run)0.250.200.150.300.180.140.090.170.260.210.110.140.340.240.240.310.150.200.380.160.15 Sediment organics (%)4.788.046.327.354.508.9611.284.137.384.969.1523.728.258.125.144.5916.084.325.222.051.23 Semidmen moisture (%)43.4850.2343.3248.7449.0560.4766.3037.7349.8443.8448.0274.6549.9051.3043.2139.2355.2638.3446.8135.6928.58 Shoreline vegetation (%)10097100100100100100100100100100835037100273273489086 Shoreline mangroves (%)13100100100244397100037938342281008247348838 8 6 Note: Catch per unit effort (CPUE) of age-1 snook was measured according to 21 long standard seine hauls. Growth and recapture ratio s (number recaptured/number released 100, standardized for effort) were from snook r ecaptured during first winter after release. 63

PAGE 64

Table 2-7. Variables used for the principle component analysis examining both short-term (days-at-large <= 365) and long-term (day s-at-large > 365) post-release responses of hatchery-reared snook stocked in 1998 and 1999 according to release site characteristics Variable SCUSCMSCLSCANCUNCMNCLNCAPCUPCMPCLPCAHBMHBLHBABCUBCMBCLBCAtidyTIA Salinity (ppt) 3.326.233.734.96.513.730.332.60.38. 230.032.134.133.835.832.133.332.936.033.435.1 Dissolved oxygen (mg/l) 3.64.45.17.46.25.85.99.77.16.45.65.22.73.47.14.16.35.87.15.68.5 Temperature (celsius) 28.831.229.630.830.829.232.631.9 31.831.532.533.428.828.527. 730.830.730.130.228.830.8 pH 7.97.98.18.37.97.18.08.47.87.98.08.08.18.18.48.28.48.48.38.08.3 Substrate softness (cm) 26.722.315.38.79.627.734.3102641132332107226319171218 Bank slope (rise/run) 0.250.200.150.300.180.140.090.170 .260.210.110.140.340.240.240.310.150.200.380.160.15 Sediment organics (%) 4.788.046.327.354.58.9611.284.13 7.384.969.1523.728.258.125.144.5916.084.325.222.051.23 Semidmen moisture (%) 43.4850.2343.3248.7449.0560.4766.337. 7349.8443.8448.0274.6549.951.3 43.2139.2355.2638.3446.8135.6928.58 Shoreline vegetation (%) 10097100100100100100100100100100835037100273273489086 Shoreline mangroves (%) 131001001002443971000379383422810082473488386 Age-1 summer CPUE 0.0400.0770.0860.0770.2990.2690.1180.0280. 0000.1510.1480.0190.0830.0000.0710.2940.1540.2310.0260.0630.000 Age-1 winter CPUE .0.3080.3610.241.0.4540.4650.126.0.4170.1670.0830.1810.472.0. 2960.1490.3390.1110.1050.167 1st winter growth rate (mm/d) 0.340.470.740.440.410.520.71.. 0.290.79. 0.520.690.670.670.63. 0.620.68 1st winter R ratio 00.2100.7430.3591.6251.7130.9660.2650.0000.0000.0830.1660.0000.0001.1200.976 0.2720.1100.0000.1230.400 R ratio, long-term dispersers 00.1050.4250.1790.2170.2140.4290.0000.0000.0830.1650.3310.0000.2880.5600.488 0.4080.4390.1780.0840.000 R ratio, winter habitats (long-term) 00.9430.2120.0000.7580.5351.5020.1770.0000.0000.0830.1660.0000.2880.2800.366 0.1360.3290.0000.2520.000 R ratio, summer habitats (long-term 00.0000.1060.0900.0000.2140.3220.0000.0000.0000.0830.0000.0000.0000.5600.488 0.2720.3290.1780.0560.000 Linf (FL mm) 657657657824595595595824...... 824704704704824.. 3 yr weight (g) .. 665.1020679856885.. 672... 98812319511290. 685. Note: Catch per unit effort (CPUE) of age-1 snook was measured according to 21 long standard seine hauls. Long-term gr owth was expressed as Linf (von Bertalanffy parameter), and weight of age-3 snook recaptures (days-at-large = 366720) and recapture ratios (R ratio, number recaptured/number released*100, standardized for effort). Recapture ratios over the long-term were organized according to those recaptured in wi nter habitats, those recaptured at least 1 km from their release sites (dispersers), and in a dult snook summer habitats (beaches, passes, and seagrass habitats). 64

PAGE 65

Table 2-8. von Bertalanffy growth paramete rs for snook recaptures (n) sourced from different release habitats Length and age data only included recaptures of hatchery snook at-large for 180d or longer. Linfkt0n South Creek 6570.250-0.48646 North Creek 5520.5500.470249 Bowlees Creek7040.408-0.50260 Island Habitat8240.164-1.65040 65

PAGE 66

Table 2-9. Numbers of snook recaptured at various distances (km) away from their respective release sites. Data are only for snook recaptured within the firs t year after releases. Release Site Release HabitatRelease sub-systemNumber recaptured by distance from release site (km) 00.1 to 55.1 to 1010.0 to 20> 20 BCU B o w l e e s C r e e kU p s t r e a m 1 42 0121 BCM B o w l e e s C r e e kM i d s t r e a m 41 0130 BCL B o w l e e s C r e e kS t r e a m M o u t h 1 24711 PCU P h i l l i p p i C r e e kU p s t r e a m 00000 PCM P h i l l i p p i C r e e kM i d s t r e a m 01200 PCL P h i l l i p p i C r e e kS t r e a m M o u t h 1 62240 NCU N o r t h C r e e kU p s t r e a m 1 43 6010 NCM N o r t h C r e e kM i d s t r e a m 5 11 9110 NCL N o r t h C r e e kS t r e a m M o u t h 5 12 7560 SCU S o u t h C r e e kU p s t r e a m 2 51000 SCM S o u t h C r e e kM i d s t r e a m 3 14100 SCL S o u t h C r e e kS t r e a m M o u t h 79100 HBU H u d s o n B a y o uU p s t r e a m 06000 HBM H u d s o n B a y o uM i d s t r e a m 47000 HBL H u d s o n B a y o uS t r e a m M o u t h 55100 TIDY T i d y P e n i n s u l aM a i n l a n d I s l a n d 3 55221 TIDYI T i d yI s l a n d 11000 BCI B u t t o n w o o d H a r b o rI s l a n d 11501 HBI L i d o L a g o o nI s l a n d 52000 PCA Coconut BayouIsland 30300 NCA B i r d K e y sI s l a n d 1 00100 SCA S o u t h C r e e kI s l a n d 61010 Total 295 161 33 21 4 Cumulative percent 57%89%95%99%100% 66

PAGE 67

67Figure 2-1. Map of experimental release si tes located off the southwest coast of Flor ida USA. Release cr eeks are denoted and release sub-habitats (U= upstream, M=midstream, L=lower stream, and I=barrier island habitat) are enclosed in transparent shaded ovals. So u t h C r e e k I N T E R C O A S T A L W A T E R W A YC a s e y K e y METERS 0 500 1000 27o10.75 N U L M IM A I N L A N D C O AS T 2500 m B i g Pa s sGulf of MexicoLido KeyS i e s ta K e yP h i l l i p p i C r e e kHudson Bayou I L U M IU M LRoberts BayI n t e r c o a s t a l W a t e rw a y 1200 M 1200 M Longboat KeyT i d y I s l an d L o n g b o a t P a s s Gulf of MexicoA n n a M a r i a I s l a n dI nt e r c o a s t a l W a t e r w a y Sarasota Bay I 2200 M Longboat KeySarasota BayGulf of Mexico Bowlees Creek I L M U 800 MB a r r i e r I s l a n dNorth Creek Bird KeysGulf of Mexico I L M UIn te r c o a s t a l W a t er w ay 27o 12 58.55 N 82o 29 56.21 W 27o 16 17.74 N 82o 34 2.48 W 82o 39 8.98 W 27o 27 31.23 N 27o 24 47.81 N 82o 34 43.95 W 82o 29 9.05 W

PAGE 68

68 Figure 2-2. Release numbers (grey bars) and po st-release results showing percent hatchery snook found in total snook catch (gra ph A) and recapture ratios (number recaptured/number released) adjusted for effort (from 21 m seine, graph B). The chronological representati on on the horizontal axis is not to scale. 0 2000 4000 6000 8000 10000 12000r-97 t-97 v-97 c-97 an-98 eb-98 r-98 r-98 y-98 un-98 Jul-98 v-98 n-99 r-99 r-99 y-99 un-99 Jul-99 ug-99 v-99 an-00 Release Numbe r Ap Oc No De J F Ma Ap Ma J No Ja Ma Ap Ma J A No J0 5 10 15 20 25% hatchery snoo k 0 2000 4000 6000 8000 10000 12000Apr-97 Oct-97 Nov-97 Dec-97 Jan-98 Feb-98 Mar-98 Apr-98 May-98 Jun-98 Jul-98 Nov-98 Jan-99 Mar-99 Apr-99 May-99 Jun-99 Jul-99 Aug-99 Nov-99 Jan-00 Release Numb e 0 0.025 0.05 0.075 0.1Recapture rat e Release number Release number Apr Jan Dec Nov Oct Mar Apr May Jun Jul Nov Jan Mar Feb May Jun Jul Apr Aug Nov Jan1997 1998 1999 Apr Jan Dec Nov Oct Mar Apr May Jun Jul Nov Jan Mar Feb May Jun Jul Apr Aug Nov Jan1997 1998 1999 B % Hatchery snook Recapture rateA

PAGE 69

0 0.5 1 1.5 2 2.5 3 3.5 4Upstream Barrier Islands Midstream Stream Mouth (n=90) (n=84) (n=52) (n=15)Recapture rate Figure 2-3. Mean recapture ratios (number r ecaptured/number released x100) of snook from 1998 and 1999 releases, by sub-subsystems. Means are calculated from replicate release years, stream systems and rel ease lots. Data only include recaptures within one year after release. Numbers in brackets are the nu mbers of individual snook recaptures for each particular releas e microhabitat. Error bars are 95% confidence limits of the means. 69

PAGE 70

0 0.5 1 1.5 2 2.5 3 3.5 4 IslandsTidyBowleesHudsonPhillippiNorthSouth Intercoastal Islands Tidy Island Bowlees Creek Hudson Bayou Phillippi Creek North Creek South Creek Mean recapture rateRecapture rate Figure 2-4. Recapture ratios (number recapture d/number released x 100) of snook by release habitat. Data are from recaptures collected within the first year after release. Mean recapture ratio is indicat ed by the horizontal line. 70

PAGE 71

Figure 2-5. Linear regression models using mean catch-per-unit effort during winter of age-1 snook to predict short(< 1 year post-re lease) and long-term (>=1 year postrelease) recapture ratios (rrate; num ber recaptured/number released x 100). r2 = 0.7228 r2 = 0.9298 0 0.5 1 1.5 2 2.5 3 3.5 4 00 511 522 5 MEAN AGE-1 CPUE IN WINTER (in release habitats) 3 SHORT-TERM RRATIOS (< 1 YEAR) LONG-TERM RRATIOS (1-8 YEARS) Recapture ratio Mean Age-1 CPUE in winter (at release habitats) 71

PAGE 72

Figure 2-6. Mean growth rates from fish released in different habitats and sub-habitats. Data are from recaptures collected 6-10 months after releases in 1998, and 1999. Sites from left to right are arranged in a Nort h to South order. Error bars are 95% confidence intervals. Numbers above bars are number of recaptures. 0.35 0.45 0.55 0.65 0.75 0.85TIDYITI BC HBLIDOKEYPC PCA NCNCA SC SCA TIDYI TIDY BC HB NCA HBA PC PCA NC SCA SCGrowth rate (mm/d) 2 14 12 1 4 3 1 4 4 42 9 72

PAGE 73

150 175 200 225 250 275 300 325 350 375 400 012345Size at recapture (Fork length, mm)6 ISLANDSTIDYBCNC SCFigure 2-7. Size-at-recapture for snook released in different habitats. Data are from recaptures collected 6-10 months after releases in April 1998, and 1999. Sites from left to right are arranged in a North to South order. 73

PAGE 74

Figure 2-8. Mean size-at-release and growth ra tes of hatchery-reared snook recaptured during the first winter after release (6-10 mo.). 0 0.2 0.4 0.6 0.8 1 1.2 050100150200250 Mean size-at-release (FL, mm)Growth rate (mm/d) Mean size-at-release (FL, mm) Growth rate (mm/d) 74

PAGE 75

75 Figure 2-9. Length frequency distributions of wild and hatchery snook at microhabitat release sites for each season. Hatchery snook bars (red) are stacked above wild snook bars (blue). UPSTREAM 0 20 40 60 50150250350450550650750850 MIDSTREAM 0 20 40 60 50150250350450550650750850 STREAM MOUTH 0 20 40 60 50150250350450550650750850 INTERCOASTAL ISLANDS 0 20 40 60 50150250350450550650750850SPRING UPSTREAM 0 20 40 60 50150250350450550650750850 MIDSTREAM 0 20 40 60 50150250350450550650750850 STR E AM MOUT H 0 20 40 60 50150250350450550650750850 INTERCOASTAL ISLANDS 0 20 40 60 50150250350450550650750850SUMMER UPSTREAM0 20 40 6050150250350450550650750850 MIDSTREAM 0 20 40 60 50150250350450550650750850 STREAM MOUT H 0 20 40 60 50150250350450550650750850 INTERCOASTAL ISLANDS 0 20 40 60 50150250350450550650750850FALL U PSTR EA M 0 20 40 60 50150250350450550650750850 MID STR EA M 0 20 40 60 50150250350450550650750850 STRE AM MOU T H 0 20 40 60 50150250350450550650750850 INTE R COASTAL ISLAND S 0 20 40 60 50150250350450550650750850WINTER Fork Length (mm)Number of Snook

PAGE 76

76 Figure 2-10. Plot of first two pr inciple components for short-term data related to characteristics of the release sites. Release sites are represented by shaded dots, release site characteristics are represented as vectors. ot of first two pr inciple components for short-term data related to characteristics of the release sites. Release sites are represented by shaded dots, release site characteristics are represented as vectors. NCA TIDYA PCA NCM SCA BCM BCL PCL TIDY NCU NCL0 0.5 0.5)Priom2 ()S e d i m e n t o r g a n i c sT e m p e r a t u r eC P U E w in t e rM a ng r o v e c o v e rD i s s o l v e d o x y g e n ( m g / l )B a n k s l o p eGrowthSe di m e n t s o f t n e s sC P U E s u m m e rVegetative coverS e d i m e n t m o i s t u reS a lin it yR e c a p tu r e r a ti oinc ple c pone nt 21%BCU-0.5 -0.5 0Principle component 1 (27%

PAGE 77

y = -1.9596Ln(x) 0.2578 R2 = 0.59230 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0.40.450.50.550.60.650.70.750.8Growth rate (mm/d)Recapture ratio Figure 2-11. Plot of growth rate versus recapture ratio (number recaptured/number released*100) from standardized seine co llections during the first winter after release. The linear model (inset) shows the best fit of loge transformed growth rates to recapture ratios with an a ssociated correlation coefficient. 77

PAGE 78

0 0.5 1 1.5 2 2.5 3 SC NC HB BC TIISLANDSCPUE0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9Mean growth rate (mm/d ) CPUE GROWTH South Creek North Creek Hudson Bayou Bowlees Creek Tidy Peninsula Barrier Islands CPUE Mean growth rate (mm/d) 9 1 12 14 15 42 Figure 2-12. Catch per unit effort (CPUE) of wild and hatchery snook less than 300 mm fork length according to release habitat and associated mean growth rates (mm/d) through the first winter after release. Error bars are 95% confidence intervals. Numbers above bars are the numbers of snook recaptured. 78

PAGE 79

BCM NCM SCM SAR WINDISP GROWTH WINFL RRATIO SUMDISP SUMFL BCU HBL NCL NCU TIDY-1.5 -1 -0.5 0 0.5 1 1.5 -1 -0.5 0 0.5 1 Variable 1Variable 2Figure 2-13. Plot of the canonica l variables (triangles) for size -at-release (SAR mm FL) and associated response variables incl uding first winter re capture ratio (RRATIO) growth to first winter (GROWTH, mm/d), disp ersal rates during summer (SUMDISP) and winter (WINDISP), and size-at-recapture (mm FL) during summer (SUM FL) and winter (WINFL). Release sites are plotted as shaded dots. 79

PAGE 80

0 0.050.10.150.20.250.30.35 80-129 130-179 180-229Distance Moved (km) 0 0.050.10.150.20.250.3 87-129 130-179 180-225Distance Moved (km)RecapFL(mm) SAR (FL mm) Distance moved (km) Distance moved (km) A BFigure 2-14. Mean dispersal distances (km) organized by size-at-recapture (RecapFL, graph A), and size-at-release (SAR, mm FL, graph B) Error bars are standard error. 80

PAGE 81

0 100 200 300 400 500 600 700 800 12345678Age (years)Length (FL) BC model NC Model BC obs NC obs NC length=594(1-e-0.553(age-0.478)) BC length=704(1-e-0.409(age-0.059)) NC length=594(1-e-0.553(age-0.478)) BC length=704(1-e-0.409(age-0.059)) Age (years) Fork length (mm) Figure 2-15. Plot of length and age data from hatchery snook recaptures. Data include recaptures from NC (shaded dots) and BC (black triangles). Von Bertalanffy length-at-age predictions are shown for bot h creeks, with growth equations shown in box. 81

PAGE 82

0 2 4 6 8 10 120.250.50.7511.251.51.7522.252.52.7533.253.53.7544.254.54.7555.255.55.756 SSFWSSFWSSFWSSFWSSFWSSFW SSFWSSFWSSFWSSFWSSFWSSFWAGE-1AGE-2AGE-3 AGE-5 AGE-4 MEAN DISPERSAL DISTANCE (KM) AGE-6125 32 52 71 37 17 14 29 10 25 8 24 6 14 9 5 6 8 33 1 1 3 2 Figure 2-16. Mean dispersal distances from rearing habitats (minimum distance from release site to recapture site) of snook recaptu red within seasons (spring, summer, fall winter) and by snook age (years). Data ar e from snook released in the spring of 1998, and 1999. Numbers of individuals re captured within each season are shown above bars. Error bars are standard error. 82

PAGE 83

NCL BCU BCL NCA Age-1 winter CPUE 1st winter recapture ratio sediment moisture L_inf short-term growth pH salinity Age-3 weight bank slope dissolved oxygen temperature recapture ratio at beaches and passes recapture ratio of long-term dispersers substrate softness Age-1 summer density sediment organics shoreline vegetation long-term winter recapture ratio NCM mangrove cover SCL BCM -0.50 0.00 0.50 -0.50 0.00 0.50Principle component 1 (36%)Principle component 2 (28%)Figure 2-17. Plot of the first two principl e components comparing long-term responses (recaptures at-large for over 1 year, black dots) of the released fish, short-term responses (from recaptures at-large le ss than 1 yr., also black dots), and characteristics of the releas e sites (shaded dots). The percentages in brackets on the axes represent the per cent of the overall variation explained by each respective principle component. 83

PAGE 84

0 100 200 300 400 500 0.350.40.450.50.550.60.650.70.75growth rates in rearing habitatsNumber of snook age-2 age-3 age-4 age-5 age-6 0 25 50 75 100 125 150 175 200 0.350.40.450.50.550.60.650.70.75growth rates in rearing habitatsBiomass (k g age-2 age-3 age-4 age-5 age-6 Growth rate (mm/d) in rearing habitat Number of snook Biomass (kg) Growth rate (mm/d) in rearing habitat A BNC/SC BC TIDYISLANDS HB NC/SC (N age-2 = 316) ISLANDS (N age-2 = 171) Figure 2-18. Numerical (graph A) and biomass (graph B) projections to adu lt stages of juvenile hatchery-reared snook released in various rearing habitats with characteristic growth rates by first winter. Projections assume equal stocking levels (1000 age1 snook per rearing habitat). Arrows i ndicate projections according to rearing habitats observed in Sarasota and Mana tee Counties (NC/SC = North Creek or South Creek, HB=Hudson Bayou, TIDY=Ti dy Peninsula, BC= Bowlees Creek, ISLANDS=barrier island habitat). Shaded dots in B highlight the differences in biomass production of age-4 snook betw een NC/SC and ISLANDS. 84

PAGE 85

CHAPTER 3 MANIPULATIONS OF STOCKING MAGNITUDE: ADDRESSING DENSITYDEPENDENCE IN A JUVENILE COHORT OF COMMON SNOOK Centropomus undecimalis Introduction Overfishing and habitat alteration are c onsidered to be the primary anthropogenic disturbances to coastal ecosystems (Hallegrae ff, 1993; Jackson, 2001; Vitousek et al., 1997). Worldwide demands on ocean fishery resour ces since the early 1990s have exceeded 100 million tons of harvest and fishery experts pred ict global marine catch has approached its upper limit (Botsford et al., 1997). Furthermore, histor ical data indicate that long-term overfishing has been a primary contributor to major structural and functional changes in coastal ecosystems (Jackson et al., 2001). The primary goal of sustai nable fisheries has evidently not been widely achieved, considering the number of overfished populations and indi rect effects of fisheries on ecosystems (e.g. bycatch) (Steele et al., 1992; Bots ford et al., 1997; Pauly et al. 1998; Hamilton and Haedrich, 1999). Anthropogenic habitat cha nge and pollution, although difficult to quantify, also pose substantial threats to fishery stocks worldwide. For many marine species, juvenile nursery habitats are associated with coas tlines where development and pollution are concentrated, resulting in inhere ntly reduced ability of fisherie s stocks to recover (Bruger and Haddad, 1986; Islam and Haque, 2004; Mumby et al., 2004). In Florida, common snook Centropomus undecimalis (snook) are a coastal, warmwater fish whose populations are a major con cern for fishery managers because of their ecological and economic value. Snook are valued as one of the top marine sport fishes in Florida and contribute to an annual USD 5.4 billon salt water recreational fishing industry in Florida alone (American Sport Fishing Association, 2004). Despite incr easingly restrictive fishery regulations on common snook, these populations are considered overfished and below management goals (Muller and Taylor, 2005). A nnual fishing mortality rates have steadily 85

PAGE 86

increased over the last 20 years, and recruitm ent has generally declined (Muller and Taylor, 2005). High fishing pressure, coupled with relatively few spawner-sized females in the adult population, has made management of snook stoc ks difficult (Muller and Taylor, 2005). Furthermore, snook associate w ith shoreline habitat (Marsha ll, 1958; Gilmore et al., 1983, McMichael et al., 1989; Peters et al., 1998a) and thus depend on coastal waters that may be subjected to intense anthropogeni c influences. Collectively, over fishing and habitat loss have caused a general decline in the population although the relative influence of these is unclear (Muller and Taylor, 2005). In Florida, snook spawn primarily during summ er (June August) in high salinity (>28 ppt) seawater in the inlets and tidal passes of es tuaries, mouths of rive rs and canals, and along sandy beaches (Marshal, 1958; Volpe, 1959; Taylor et al., 1998). Post-larval snook recruit mostly to vegetated shallow brackish tidal creeks canals and lagoons in both low(riverine) and high-salinity (mangrove swamp and salt marsh) habitats (Peters et al., 1998a), with moderately sloping shorelines and basin depths of 1 m or less, with mud or sand substrate. Age-0 cohorts are thought to generally remain in their rearing hab itats through the winter and early spring (Gilmore et al., 1983; McMichael et al., 1989). In Sarasota and Ma natee counties, these habitats are found in tidal creeks and estuarine backwate rs, and can also serve as thermal refuge for snook of all sizes during winter (N. Brennan, unpublished data). Gilmore et al. (1983) hypothesized that after winter, adolescent snook (150 400 mm standard length [SL], mean = 240 mm SL, from east-central Florida populations) undergo an ontogenetic habitat shift to seagrass beds. Adult snook are also known to disperse to seagrass beds estuarine inlets, and beaches by late spring and early summer (Marshall, 1958; Gilmore et al., 1983; Taylor et al., 1998). 86

PAGE 87

Faced with rapid human population growth in Florida and the limitation of traditional management tools, managers have investigated the potential to augment overfished populations with supply-side approaches such as stocki ng hatchery-reared juvenile snook. However, evidence for stock enhancement programs actually accomplishing stock management goals has been sparse, and such programs can be ineff ective and even deleterious to wild stocks (Nickelson, 2003; Walters and Martell, 2004; Kosteau and Zhou, 2006). Therefore, stock enhancement programs might best be operated on an experimental basis with rigorous scientific evaluation prior to full-scale a cceptance as a management tool (Blankenship and Leber, 1995; Leber, 1999; Hilborn, 2004; Walte rs and Martell, 2004). Experi ments with stock enhancement can provide valuable insight into population dyna mics, behavior, growth, and survival responses to habitat quality, for example (Miller and Walters, 2004). The efficacy of stock enhancement should depend on the magnitude of density-dependent processes. With snook, a cannibalistic and pisc ivorous species, the degree to which habitat availability and quality, or recruitment limitati on, influence stock sizes remains unclear. Many species suffer elevated and density-dependent mortality during specific ontogenetic stages, typically during early life stag es (Houde, 1987; Bystrm et al., 2004; Doherty et al., 2004). In fish, highest mortality rates o ccur during the la rval stages (Houde, 1987; Lorenzen, 1996). In some cases however, density-dependent mortality is high in later stages and overall recruitment rates are restricted by these survival bottlenecks For example, Bystrm et al. (2004) showed that while age-0 char were not affected by density-dependent pr ocesses, age-1 char underwent highly density-dependent ontogene tic feeding shifts and became exposed to high rates of predation. With snook stocks, relative contributions of various life stages to overall survival rates remain unclear. Loss of habitat availabi lity can inherently re duce a stocks production 87

PAGE 88

capacity, yet high density-dependent mortality ca n still occur in remaining habitats from competition and predation. An examination of population responses to manipulative stocking experiments over a variety of hab itats would aid in determining th e timing and extent of densitydependent effects. Furthermore, if snook stocks are overfished, then experimental and manipulative stocking may be necessary to elic it density-dependent responses in augmented populations. In this study, I released juve nile (age-1) hatchery-reared snook to manipulate localized age-1 recruitment and identify potential resultant density-dependent mortality effects. Mortality responses in the age-0 cohort from age-1 canniba lism or competition were not addressed in this study because primary recruitment pulses of age-0 snook occurred from June September, well after the experimental releases occurred (in May), and relate d abundance responses would be confounded with natural variations in creek-specific recruitment. Specifically, I aimed to determine if snook nursery habitats are naturall y filled to capacity, i.e., if strongly densitydependent survival would preclude increasing th e abundance of age-1 snook leading to increased abundances of older snook. To accomplish this I (1) estimated wild age-1 snook abundance in four estuarine tributarie s, (2) released hatchery-reared age-1 snook to increase th e total localized abundance of age-1 snook by either 100% (hi gh augmentation; n=2 creeks) or 10% (low augmentation; n=2 creeks); (3) determined if ha tchery-released snook were subjected to higher mortality rates than wild age-1 s nook, and (4) determined if overal l loss of age-1 snook in creeks with high augmentation was higher than those in creeks with low augmentation. Methods Study Area Experimental releases were made in four estuarine creeks (Bowlees Creek [BC], Whitaker Bayou [WB], North Creek [NC], and South Creek [SC] ) located in Sarasota and 88

PAGE 89

Manatee Counties, on the west co ast of Florida, USA (Figure 31). These creeks are tidally influenced and water levels typically fluctuate by about 0.75 m daily. Salini ties vary with tidal influence and seasonal rains (wet season duri ng summer). Water temperatures during summer (June September) typically range from ca 28 oC 34 oC, and winter (December March) temperatures range from ca 12oC 22oC. Dissolved oxygen levels during summer are low compared to winter (typically <5 mg/l versus >7 mg/l). All creeks are partially influenced by anthropogenic shoreline alterations, and sections of BC, WB, and SC have been dredged to some extent for maritime navigation within the last 20 years (Table 3-1). North Creek remains relatively unaltered, although upstream reaches are surrounded by a golf course development and shoreline residential properties. Densities of wild age-1 snook (~150300 mm fork length [FL]) during early summer were about 0.5 3.0 fi sh/30 m of shoreline (Table 3-1). Experimental Design and Sampling Methods I defined age-1 snook or juvenile snook as offspring from the 2001 spawning year. By April 2002, these snook were approximately 10 (+/2) months old. Experimental releases of hatchery-reared juveniles occurred in May, 2002 and our study continued until June 2003 when the juveniles were about 24 months old. I impos ed two levels of mani pulation: 1) a large addition of hatchery-reared juveniles equal to 100% of the wild juvenile abundance (in creeks BC and WB), and 2) a smaller addition of hatchery-re ared fish equal to 10% of the wild stock (in NC and SC). Quantitative sampling was performed in all creeks: once before releases (to estimate pre-release age-1 abundance), and six ti mes after releases thr ough the following spring (approximately in June, August, October, D ecember 2002, and February, June 2003, Table 3-2) to monitor the effects of th e releases on in-creek abundan ce of juvenile snook. To estimate in-creek juvenile snook population size, I used a leap -frog systematic sampling design where a standardized bag se ine (73 m long x 3 m deep with 1 cm nylon 89

PAGE 90

multifilament mesh) was hauled at roughly every third 30 m section of shoreline throughout each creek (Figure 3-1, Table 3-1). Each haul approximately samp led a 30 m x 21 m rectangle (630 m2). For creek sections less than 21 m wide, we sampled 30 m of creek length but incorporated both shorelines into the sample. The seine, loaded on a 4 m kayak, was deployed by hauling about 21 m of the leading end of the seine toward s the shoreline. Once there, a second person held the outside corner of the block and a thir d person pulled the kayak parallel to the bank (or against the opposite shoreline) for 30 m then turned and deployed the remaining 21 m of net towards the shoreline. Once against the bank, bo th persons moved toward each other along the shoreline and retrieved the remainder of the seine. To avoid overestimating snook abundance, I divided snook catch by two for samples collected from two-bank sites because of the increased shoreline habitat. Seines were not hauled in narrow areas (i.e. less th an 3 m across) such as mosquito ditches, but these shorelines were assu med to be valid snook habitat and were used to calculate total snook abundance. Gear Efficiency and Population Estimation Prior to abundance estimates, we first conduc ted depletion-removal trials in January and February, 2002 to measure gear efficiency. I used a modified depl etion-removal population estimate (Hillborn and Walters, 1992) within a defined area to predict single-pass seine efficiency (Figure 3-2). Thr ee types of depletion-removal methods were used depending on habitat type. In creek areas narrower than 21 m across (i.e., two stream banks, two-bank depletion), two nets (21-24 m long, 0.9 m deep, 1 cm multifilament mesh), about 42 m apart, blocked off upstream and downstream borders of the sampling area (Figure 3-2a). We used the standard seine described above, to sample ~ 30 m within this zone, al lowing ~ 6 m from the seine perimeters to the block nets measured along the shoreline to allow for escapement (as might occur in a single pass seine) After this, we seined the enti re area with the above seine two 90

PAGE 91

more times, for a total of three seine hauls (mos t snook were captured within the first seine, so three seine passes sufficiently described the de pletion rate, see belo w). We performed 10 separate two bank depletion experiments. We also performed single-bank depletion along open shorelines (Figure 3-2b), and used the same method if additional shorelines were farther than 21 m from the sampled shoreline. A 45 m long block net (1 cm multifilament mesh) was deployed parallel to the shoreline, ~ 21 m away from shore, and two 21 m seines (same sized mesh, attached at each end of the deployed 45 m long seine) were deployed toward the shore to enclose a rectangular area of ~ 945 m2. Once block nets were in pl ace, a standardized seine was hauled along the inside perimeter of the bloc ked area. The standard net was stretched 30 m along the inside perimeter of the block net, then each end was pulled toward shore. I allowed ~ 4 m of space between the ends of the block to allow for escapemen t (Figure 3-2b). Again, second and third seine hauls sampled the entirety of th e blocked area. For each seine haul, captured snook were counted and measured (FL, mm), th en placed in holding containers with aerated creek water until the depletion trial was comple te. We performed eight separate open bank depletion trials. We also performed two-bank extended depletion experiments where we blocked off ~ 116 m of creek habitat (Figure 3-2c). With a st andard seine, we performed three seine hauls sequentially upstream to the upstream block net. Th is process continued for a total of nine seine hauls in the blocked section. We performed this method in two separate areas. Using data from the three-pass depletion removal trials, I calculated the maximum likelihood of the number of snook remaining () in the net area after the last depletion pass i (which captured i juvenile snook) assuming binomially distributed catch for each seine pass, 91

PAGE 92

with equal capture probability fo r each pass. I then estimated the total abundance of snook in the net area ( ): = i + (3-1) From this I calculated the number present at the time of each pass ( 1s, 2s, 3s ) and generated an estimate of the capture probability per seine haul ( pmax) as the conditional maximum likelihood estimate for binomial sampling: p = i / ( 0s) (3-2) I also calculated a pcommon value for all depletion trials j : pcommon = i j / i j (3-3) where i and j index seine hauls and depletion trials respectively. Estimates of age-1 snook (estimated by length frequency analysis, see below) abundance in a seined area at time t ( t) were calculated as: t = t / pcommon (3-4) Where t is the catch of juvenile snook in the seine area at time t. Adjusted mean catch from all areas s in creek C at time t ( tC) were then extrapolated to total creek shoreline habitat, Hc (obtained from aerial photographs [1cm=24 m], expressed in 30 m units) to obtain an in-creek population (Ct) of juvenile snook: N N Ct = tC Hc (3-5) with variance: V(Ct) = AC 2 V( tC) (3-6) N Therefore, the variance estimate uses the maxi mum likelihood estimate of catch efficiency (without associated variability) bu t incorporates variation around th e mean catch in each creek at time t and assumes creek area HC is constant. 92

PAGE 93

I used repeated measures analysis to compare preand post-release abundance for treatment and control creeks. If the April 2002 abundance estimate of wild snook juveniles in a particular creek (pre-release) was less than it s early summer (post-re lease June or July 2002 samples) abundance estimate, I used early su mmer abundance of age-1 wild snook as our prerelease estimate (I assumed that no new recruitm ent of wild age-1 snook occurred; age-1 snook typically begin to decline in abundance in creeks by midsummer [Brennan and Leber, unpublished data; Gilmore et al., 1983]). Since large-scale emigration of age-1 snook occurred from mid-summer through fall, I only used early -summer and winter abundance estimates for post-release repeated measures analysis. I also used a linear regression to model augmentation level (expressed as a percentage of standing wild stock) with ch ange in abundance after stocking (also expressed as a percentage) to test if a significant relationship existed between augmentation level and change in abundance. Density (number per 30 m of shorelin e) of juvenile snook in the creeks ( Dc ) at time t was calculated by dividing total in-c reek juvenile snook abundance Ct by total creek shoreline distance (HC) as follows: N Dct = Ct / HC (3-7) N I used data from preand post-release sampling, to compare observed and expected densities within the creeks. To generate expected dens ity, I assumed that numbers of stocked fish were completely additive to wild juvenile numbers, then used a Chi-Square test to compare differences in observed and expected age-1 densities in the creeks after releases occurred. I also used an analysis of covariance procedure to co mpare in-creek changes in density (DCt) of juvenile snook before and after releases (using pre-release density, peak summer density, and peak winter density). 93

PAGE 94

Pre-Release April 2002 Sampling To estimate pre-release juvenile snook a bundance, we performed creek-wide standard seining efforts in each creek prior to stocking hatchery-reare d snook (Figure 3-1). Captured snook were also marked with externally visible implant elastomer (VIE) tags implanted in the caudal fins (see Brennan et al., 2005 ) with VIE colors specific to creeks to (1) aid in determining site fidelity and inte r-creek migration rates, and (2) id entify snook recaptured multiple times. Tagging and Release Hatchery-reared snook, F1 progeny from wild parental stocks, were hatched in the summer of 2001 and reared until spring 2002 when tagging and release ac tivities occurred in May, 2002. Each released hatchery-reared snook was tagged with (1) a coded-wire tag (CWT) to identify its associated release information (i .e., release creek, release date, size-at-release), and (2) two red VIE tags implanted in the caudal fin to externally identify it as a hatchery-reared snook (as in Brennan et al., 2005). Tagged juveniles were then returned to their tanks and held for one week to recover from the tagging process. Salinities in the rearing tanks ranged from 3 6 ppt and water temperature from 26-30 C. On the day of release (six days after ta gging), fish from each release group were harvested and checked for tag presence. Snook were transported in tanks with brackish water by truck and boat, and stocked directly into 22 m3 predator-free acclimation enclosures and held for 3 d (a protocol to improve post-release survival rates, Brennan et al., 2006 ). Enclosures were located along shorelines, mostly vegetated with mangrove (Rhizophora mangle ), oaks ( Quercus spp.) palms ( Sabal palmetto ), and Brazilian pepper ( Schinus terebinthifolius ), and at some locations near boat docks and riprap. After 3d of acclimation, snook were allowed to swim freely from the enclosures. Release numbers we re proportional (either 100% or 10%) to prerelease abundance estimates of juvenile snook in each experimental creek. 94

PAGE 95

Post-Release Evaluation Post-release creek-wide standard sampling effo rts were performed in each creek six times after the release (app roximately during June, August, October, December 2002, February 2003 and June 2003, Table 2). I identified year classe s of wild snook accordin g to length-at-capture based on (1) length frequency graphs of captu red snook, (2) correspondin g sizes of cultured snook, and (3) tag-recapture data on known youngof-year snook. For hatchery-reared snook I adjusted catch by accounting for expected tag lo ss over time (e.g., VIE tag loss after 6 months was 3%, [Brennan et al., 2005]). All captured sn ook were counted, measured, and checked for the presence of CWTs with magnetic tag detectors, and visually examined for VIE tags. Untagged wild snook and first-time recaptured hatchery-reared snook were tagged with VIE tags in the caudal fins (using a unique color for each creek) and CWTs. For subsequent recaptures, snook were re-tagged with a VIE of a specific color (BC = orange, WB = green, NC = yellow, SC = pink) and implant location (caudal or anal fin) to iden tify the creek in which it was captured, and the number of times recaptured. Results Gear Efficiency Overall we performed 21 depletion-removal trials. Of these, juvenile snook were captured in 12 trials and these data were used to generate estimates of seine efficiency. The maximum likelihood estimate of pcommon value was 0.79 (90% CI = 0.75 0.83, 3 for likelihood profile for the estimate). In April 2002, our leap-frog sampling regime resulted in 117 standardized seine hauls performed throughout the four experimental creeks Our samples represen ted about 4 10% of the total shoreline habitat in th ese creeks (Table 3-1, 3-1). We captured 505 snook and of these 183 were estimated as age-1 juveniles (2001 cohort, ~ 10 months old). After adjusting for seine 95

PAGE 96

efficiency and extrapolating mean CPUE to creek-wide shoreline distance I estimated in-creek juvenile populations to be 375 juveniles in BC, 476 in WB, 2018 in NC, and 639 in SC (Table 3-1). Tagging and Release of Hatchery-Reared Snook Hatchery-reared snook ranged from 84 270 mm FL at the time of tagging (mean length = 177.5 mm FL +/2.95 SE) and samples collected in April 2002 of wild j uvenile snook of the same age were 79-219 mm FL (mean= 155 +/2.43 SE). We tagged and re leased a total of 2,372 hatchery-reared juvenile snook (Table 3), and six days af ter tagging, CWT retention in the snook averaged 99.5% (from 14 groups with an aver age of 84 fish / group), and VIE retention was 100%. Total transport time to the release si tes was ca. 2 5 hours for all release groups and releases occurred between 1230 and 1900 hrs. Post-Release Evaluation Overall, from 1 June 2002 30 March 2003, we performed 556 standard seine hauls, captured 3261 snook and captured, tagged, and releas ed 1588 juvenile snook (Table 3-2). Of these, 275 were hatchery-reared snook recaptures, and 155 were wild snook recaptures. I found little evidence of inter-creek movement; of 155 co lor-coded wild snook recaptures, only one was recaptured in a creek other than its release creek (about 0.6% found in other creeks) (Table 3-2). Because all hatchery-r eared snook were originally tagged and released with a common VIE color, I were not able to obtain estimates of inter-creek migrati on from first-time recaptures. Only 11 hatchery-reared snook were recaptured twice, and all of these were captured in the same creek as the first recapture event. Samples from June 2002 resulted in a total of 410 captured snook, 294 of which were juveniles. Relative abundance of wild juvenile snook in June was 0.86x, 1.87x, 0.63x, and 2.45x the April abundance for BC, WB, NC and SC, re spectively (Figure 3-4). However, when 96

PAGE 97

hatchery-reared fish were included three creek s showed abundance increases compared to April (BC =1.20x, WB =2.26x, NC =0.81x, SC = 2.61x that of pre-releas e estimates). In June, abundance estimates of hatchery-reared fish were 120% larger than wild juvenile abundance in WB and 40% of the wild abundanc e in BC (but 97% in July), 29% of NC wild abundance and 6% of SC wild abundance (control cr eeks) (Figure 3-5a, Figure 3-5b). Juvenile snook generally emigrated from all four creeks by fall, then returned in winter (Figure 3-4). As summer progressed, in-creek abundance of age-1 juveniles steadily declined, and only young-of-the-year snook (from 2002 summer spawns, < 100 mm FL) numerically dominated snook catches. For example, in earl y summer, hatchery-reared snook abundance was 37% of the number released in BC, ~100% in WB 85% in NC, and 75% in SC, respectively. By fall, large declines in hatchery-reared fish were evident in all creeks; 6%, 6%, 17%, and 0% of the original numbers released in BC, WB, NC, and SC, respectiv ely (Figure 3-4). Wild age-1 snook exhibited similar patterns of large, late-summer declines in abundance in all creeks. After the first significant drop in temperature (below 22oC), both wild and hatchery-reared fish abundance returned to substantial levels co mparable to early-sum mer abundance; winter abundance of wild snook was 120% the peak summer abundance in BC, 134% in WB, 96% in NC, and 89% in SC. Winter abundance of hatc hery-reared snook was estimated at 54% of the summer abundance in BC, 78% in WB, 69% in NC, and 258% in SC (Figure 3-4). Power analysis showed that, given the variab ility in the catch data and the small number of replicates used in the study, power was low (0.3 0.8) for de tecting a 100% difference in population means. Nonetheless, repeated measur es analysis showed a significant time and treatment interaction effect for preand postrelease abundance (p=0.024, F = 10.78, df = 2, 3). Regression analysis showed no significant relati onship between augmentation level and change 97

PAGE 98

in abundance from preto post-release (using eith er summer or winter as post-release values; p=0.42, p=0.45, R2=0.09-0.33). Creek Density The treatment creeks used in this study were quite different in their ability to accommodate augmented snook densities. The analysis of covariance model showed no significant differences in prerelease snook density (p=0.51, 3 df Figure 3-6), however, after releases occurred, WB showed a significant in crease in overall density (over twice the prerelease level) that was maintained through th e winter (p=0.034, 3-6). In BC (also a high augmentation treatment), the expected augmented levels were not sustained (Figure 3-6), and I estimated that 64-85% of the released cultured fish were lost by the time of the first post-release sample. Loss within wild snook populations was not detected over this period, however. Creeks that received low levels of augmentation showed no significant changes in overall juvenile snook density throughout the study (Figure 3-6). Chi-s quare analysis of the observed and expected densities in all four creeks reflected the in itial loss of hatchery fish in BC by showing a significant difference between observed and expected post-release densities (p<0.001, Chisquare= 306.82, 3 df). Thus, responses of j uvenile snook density to various levels of augmentation were creek specific. Discussion Experimental releases of high-densities of age-1 hatchery-reared snook in two estuarine tributaries elevated total age-1 abundance above pre-release levels throughout the study. At the same time, I found no detectable evidence of suppressed abundance of wild age-1 snook in these habitats. Surplus production cap acity for age-1 snook existed at these particular experimental times and locations, although inte r-cohort or community-wide responses were unclear. This is not surprising as others have found population sizes below habitat productive capacities (e.g., 98

PAGE 99

Orth and Maughan, 1982; Conder and Annear, 1987). Regardless of a habitats productive capacity, stocks (such as snook) at low levels, with high fecundity, at the edge of their range, and exposed to high environmental variation, typica lly demonstrate highly variable recruitment (Myers, 2001), and thus can occur at densities below capacity. This study examines an important assumption about recruitment processes that densitydependent survival of wild fish may be reduced by the addition of hatchery-reared fish. In salmonid populations, evidence exists for density-dependent respons es in wild stocks due to stocking hatchery-reared fish (Nickelson et al., 1986; Nickelson, 2003; Kostow and Zhou, 2006). Density-dependent growth responses were not meas ured in this study, but ot her work in the same creeks (Brennan and Leber, unpublished data), sh owed suppressed growth in tributaries with high densities of age-1 snook. I found no eviden ce of high augmentation treatment effects on wild conspecific density, although an effect may have been e xpressed through initial loss of hatchery-reared fish in BC where the productive capacity may have been substantially exceeded. Abundance patterns of wild age-1 snook thr oughout the study, however, followed similar patterns in all creeks. After emigration from the creeks in fall, ther e was movement back into the creeks of age1 wild snook in winter, resulting in abundances slightly higher th an those in summer. Given the low observed rates of inter-creek movement (0.6% mixing between experimental creeks), the additional abundance in creeks of wild age-1 sno ok during winter (observed in all creeks) was probably due to immigration of wild snook from other sources, emphasizing their importance as thermal refuge habitat. Furthermore, hatche ry-reared snook demonstrated the same seasonal movement pattern and (after ini tial loss of fish in one treatment creek) early-summer abundances were very similar to winter abundance in both treatment and control creeks. Declines in 99

PAGE 100

abundance of hatchery-reared snook from summer to winter (after the in itial high post-release loss), presumably reflected mortality rates in resident wild age-1 snook. The use of a fixed-station and systematic sampling program was appropriate in our study because it incorporated a repres entative spectrum of population a nd habitat size/quality gradients in the creeks (by attempting to sample every thir d 30 m section of shoreline habitat, see Hillborn and Walters, 1992). Increases in abundance from April to June (as seen in WB and SC) may reflect improvements in the efficiency and sampling ability of field workers. Such improvements in sampling efficiency or catchability with experience are not uncommon (e.g., Walters and Maguire, 1996). In our experimental creeks, it is doubtful that abundance of wild age-1 snook actually increas ed (via immigration) prior to la rge-scale emigration of juveniles from the creeks in the fall. Furthermore, samp les from NC and BC showed expected declines in abundance. Logistical consideratio ns and cost restricted replica tion and sampling effort for this study. The low experimental power meant that the probability of a Type-II error was high, yet repeated measures detected a significant time*t reatment interactive effect on population means and thus treatment-effect size was high. Furt hermore, sampling over time showed consistent patterns in all of the creeks (e.g., abundance declines in fall, and in creases in winter), and further supports our confidence in our estimates of abundance. Inter-annual variation could have a strong eff ect on the results of si milarly staged release experiments performed at other times (through vari ations in carrying capac ity, recruitment, and subsequent competition and predation; see Walters and Martell, 2004; Brennan, unpublished data). Common snook stocks in Florida have ch aracteristics of high r ecruitment variability (Myers, 2001) and, even over a 30 km stretch of coastal habitat, evidence exists for strong variation in intra-annual recruitm ent of juveniles (Brennan, unpublis hed data). As snook are also 100

PAGE 101

cannibalistic (Tucker, 2003; Adams and Wolfe, 2006), and age-0 and age-1 snook share nursery habitat for a prolonged period a nd increases in abunda nce of age-1 snook may have negative consequences on young-of-year snook populations through cannibalism and competition for refuge and food resources. Cyclic pattern s of abundance due to cannibalism are common (Frankiewicz et al., 1999; Sanderson et al., 1999; Fromentin et al., 2000; Claessen et al., 2000; Persson and de Roos, 2006). S nook populations in Florida show strong evidence for patterns of alternating fluctuations of abundance between age-0 and ag e-1 year classes (Brennan, unpublished data) as do a c ousin of snook, barramundi Lates calcifer in Australian estuaries (Walters and Martell, 2004; Griffin, unpublished data) due to intraand inter-cohort cannibalism and competition. Release programs that elevate age-1 abundance could intensify such effects and a cautionary approach is warranted. Nonethele ss, the potential for elev ating stocks that are below capacity is demonstrated by this study and its potential over the long-term must be evaluated. The presence of small (~ 60 mm FL) snook in the creeks during fall reflects continued use of creek habitat by young snook throughout the year. Dramatic declines in abundance of age-1 snook in the creeks during late summer early fall (~ 200-325 mm FL at this time), probably represents an ontogenetic habitat shift of the age-1 snook. This could possibly be due to (1) a reduction in predation threat as snook attained larger sizes (Koczaja et al., 2005), (2) superior foraging habitat and improved prey av ailability and preferences outside the creeks (Ruzycki and Wurtsbaugh, 1999), (3) better water qua lity outside the creeks (Brennan and Leber, unpublished data), or (4) combinations of the a bove reflecting tradeoffs with size-dependant predation threat, and improvements in growth du e to better foraging oppo rtunities (e.g., Werner and Gilliam, 1984; Walters and Juannes, 1993; Olsen et al., 1995). 101

PAGE 102

Although our results indicated that stocki ng was additive in both high augmentation treatments (BC and WB), depressed survival of the hatchery-reared fish occurred in one creek (65-85% loss). While density-i ndependent causes cannot be ruled out, it may indicate that the productive capacity in BC was exceeded, and may re flect variation in the productive potential of various juvenile snook habitats. Pre-release dens ity of age-1 snook in BC was lower than in the other creeks, yet density in itsel f was not predictive of the habi tats potential to accommodate more snook. In BC, shoreline habitat alteration in the form of seawalls, and steeply dredged shorelines has been extensive, and ~ 65% of the total shoreline ha bitat has been altered (Table 31). Shoreline alteration in th e other creeks was much less (ab out 14% in WB, <2% in NC, and 7% in SC). Age-1 snook densities in NC and SC were similar, however, post-release densities in WB were ~ 8 times higher than BC densities and four times that of NC and SC. While this suggests an effect on a habitat s productive capacity, va riation in recruitment, anthropogenic development, and low sample size confound such a comparison. High juvenile snook abundance was common in habitat characterized by deep mud, overhanging vegetation, gently sloping banks, and low current (< 0.1 m/s or stagnant) (Chapter 2). In our experience, this type of habitat is scarce in highly devel oped creeks with dredged bottoms, and shorelines modified with seawall channelization and boat docks (e.g., BC). These altered habitats often attract predaceous fish po ssibly due to improved water exchange (from less flow restrictions) or dock lights that attract bait fishes. Such conditions may be costly to snook juveniles through loss of predation refuge and increases in predator abundance. Further research is needed to quantify the influence of anthr opogenic alterations on juvenile snook abundance to identify potential loss of production. 102

PAGE 103

In this study, juvenile snook nurseries acco mmodated abundance increases that persisted over time. Clearly, the use of f our experimental tributaries duri ng a particular year to conduct manipulative tests on recruitment, albeit logistically difficult, is insufficient to capture a broad picture of recruitment dynamics and the long-te rm productive capacity of various systems. Future studies should also focus on smaller and younge r size classes to iden tify potential survival bottlenecks (e.g., Doherty et al., 2004), and effects of multiple year classes sharing important refugia, where density-dependent growth and mortality may be strongly influential. Understanding community-wide implic ations and predator-prey interactive effects (e.g. Walters and Kitchell, 2001; Persson et al., 2007), and in terannual variation in re cruitment and habitat capacity, are important considerations that may best be addressed by coupling ecological models with empirical experimentation. Stock enhancem ent programs should consider strategies that maximize their cost-effectiveness while minimizing threats to wild stocks. Empirical studies such as this, however, are im portant first steps toward understanding these dynamics and can provide baseline information for dynamic age-stru ctured stock recruitment models, often lacking data for juvenile life stages. 103

PAGE 104

Table 3-1. Results from standardized sampli ng in April 2002 and physical attributes of experimental creeks (shoreline distance and percent altered habitat). Population estimates are of age-1 snook found in each creek. ExperimentalNumber of Number of Total SampledSampled shor eline% AlteredMean CPEAdjustedVarianceCalculated Calculated creek seine haulsshores sampledshoreline (m)shoreline (m)(% of total)shoreline mean CPE age-1 abundancemax abundance Bowlees Creek 28 38140511159 8.25650.640.812.07 375 1,715 Whitaker Bayou 18 267650793 10.37141.501.909.30 476 1,913 North Creek 45 66251892013 7.9921.932.4413.82 2,018 8,003 South Creek 26 34242621037 4.2770.630.802.07 639 2,904 Totals 117 164711535002 7.03 3,508 14,535 104

PAGE 105

Table 3-2. Mark-recapture results for the four experimental creeks organized according to sample month SampleCreek EffortCaptured, tagged, and releasedRecaptures Migrants Month (No. seines)Wild Hatchery (recaptures)Hatchery 2xWildWild 2xNOrigin JuneBowlees Creek 2716 12 010 0 2002Whitaker Bayou 1641 65 000 0 North Creek 4591 23 080 0 South Creek 2662 4 020 0 July-AugustBowlees Creek 2822 22 130 0 2002Whitaker Bayou 1613 12 040 0 North Creek 4679 23 1130 0 South Creek 2633 2 070 0 Aug-Sept.Bowlees Creek 2820 4 200 0 2002Whitaker Bayou 166 4 010 0 North Creek 45124 14 150 0 South Creek 2919 1 021 0 Oct. Dec.Bowlees Creek 2672 7 040 0 2002Whitaker Bayou 1637 0 020 0 North Creek 34317 5 1140 0 South Creek 0 0 00000 Jan.-MarchBowlees Creek 284 5 070 1Whitaker Bayou 2003Whitaker Bayou 1636 45 4110 0 North Creek 46239 18 1511 0 South Creek 2617 8 0110 0 June-JulyBowlees Creek 2814 1 010 0 2003Whitaker Bayou 166 0 020 0 North Creek 4538 0 060 0 South Creek 277 0 000 0 1313 275 111552 1 Note: Hatchery 2x and Wild 2x represent snook that were recaptured in two subsequent sampling events. Data are only for th e juvenile snook cohort (age-1). 105

PAGE 106

Table 3-3. Numbers of hatchery -reared snook released at four experimental release sites Release dateRelease groupSize classPre-releaseExpectedJun-02smallmediumlargejumboTotalswild N% cultured% culturedMay 20, 2002 Bowlees Creek 106 454169160889 37570.3429.03 Whitaker Bayou1475141821821025 47668.2853.27 North Creek1341786460436 201817.7723.66 South Creek 40521818128 63916.696.25 Totals 32111984334202372 3508 Note: Snook mean lengths were 138 mm FL for the small size class, 160 mm FL for medium, 188 mm FL for lar ge, and 218 mm FL for jumbo Abundance estimates by creek for the age-1 wild snook before releases (April 2002) are provided with the expected and observed (in June 2002) percen tage of hatchery snook in post-release collections. 106

PAGE 107

Tampa Bay Tampa Bay SOUTH CREEK I N T ER C O A ST A L W AT ER W A YB A R R I E R IS LA N D METERS0 500 1000 SARASOTA BAY BOWLEES CREEK 0METERS 250500 H W Y U S 4 1 SARASOTA BAY H W Y U S 4 1 WHITAKER BAYOU METERS 0250 1000 M a i n l a n d C o a s t METERS NORTH CREEK 0 500I N T E R C O A S T A L W A T E R W A Y 82o299.05 W 27o 12 58.55 N 82o 29 56.21 W 82o 33 7.00 W 27o 21 12.71 N 82o 34 43.95 W 27o 24 47.81 N 82o 29 9.05 W M A I N L A ND C O A S TH W Y U S 4 1H W Y U S 4 1 Figure 3-1. Map of experimental study sites along the coasts of Sarasota and Manatee Counties, Florida. Dots represent locations where collections with the standardized seine occurred. 107

PAGE 108

116 m Block Net Block Net 21 m (1,4,7) (2,5,8) (3,6,9) 30 m 30 m 36 m Block Net Block Net 21 m 30 m 36 m Block Net Block Net 21 m Block Net (Standard Seine) (Standard Seine)Shore Shore Open Water Shore Shore Shore A 30 m B C Figure 3-2. Schematic of various depleti on-removal methods including (A) two-bank depletion, (B) single-bank depletion, and (C) two-bank extended depletion. Dark lines indicate block nets, lighter da shed lines indicate locations of seine efforts, and arrows indicate the directi on of the seine hauls. The numbers in brackets represent the sequence and location of seine hauls (hauls 1 through 9). 108

PAGE 109

0.00 0.20 0.40 0.60 0.80 1.00 012345678910111213141516171819 "two bank" "single bank" "two-bank extended" Depletion IDProbability of capture (p) 0 1E-12 2E-12 3E-12 4E-12 5E-12 6E-12 7E-12 0.650.690.730.770.810.850.890.93 LikelihoodProbability of capture AB Figure 3-3. Graphical representa tion of individual capture pr obabilities (p) of snook from various depletion-removal methods (circles triangles and squares, graph A) and the likelihood profile for pcommon (graph B). 109

PAGE 110

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 0 500 1000 1500 2000 2500 3000 0 200 400 600 800 1000 1200 1400 Bowlees Creek Whitaker Bayou North Creek JunAugOctDecFebApr 2002 2003 JulSepNovJanMarMay May Jun South CreekJunAugOctDecFebApr 2002 2003 JulSepNovJanMarMayJun JunAugOctDecFebApr 2002 JulSepNovJanMarMay May Jun 2003 JunAugOctDecFebApr 2002 JulSepNovJanMarMay May Jun 2003Low Augmentation Creeks High Augmentation CreeksNumber of snook Number of snook Post-release loss wild hatchery hatchery wild hatchery hatchery wild hatchery hatchery wild hatchery hatchery Figure 3-4. Abundance estimates of wild (grey area) and hatchery (w hite area) age-1 snook over time in treatment creeks receiving high augmentation levels (top graphs) and low augmentation levels (bottom graphs ). Error bars are from abundance estimates obtained from 90% confidence intervals around pcommon values. Arrows indicate the timing of hatche ry-reared snook releases. General increase in wild abundance from summer to winter indicates immigration from other sources. 110

PAGE 111

0 1 2 3 4 5 6 7 8 9 Bowlees Creek Whitaker Bayou North Creek South Creek 0% 50% 100% 150% 200% Bowlees Creek Whitaker Bayou North Creek South Creek Bowlees Creek Whitaker Bayou North Creek South Creek Bowlees Creek Whitaker Bayou North Creek South CreekRelative Abundance Catch per effortabFigure 3-5. Results from post-release samples collected in June 2002 showing catch-per-unit effort (graph a) and relative abundance (graph b) of hatchery snook (white bars) to wild juvenile (grey bars ) abundance (set at 100%). 111

PAGE 112

CHAPTER 4 CRYPTIC CANNIBALISM IN SIZESTRUCTURED SNOOK POPULATIONS Introduction Many species experience elevated mortality during specific ontogenetic stages, often highest during early life stages (Smith and Fretwell, 1974; Jakobsson and Eriksson, 2000; Einum et al., 2006). In fish, this especially occurs during larval stages (Houde 1987; Doherty and Fowler, 1994; Lorenzen, 1996). While many processes contribute to early-stage mortality, predation can be a primary contributor (Mol, 199 6; Duffy et al., 1997; Taylor and Danila, 2005; Titelman and Hansson, 2006). Nonetheless, high mort ality can also manifest in later stages with reduced refuge or increased foraging risk (e.g. Mooij et al., 1996; Bystrm, 2004). Cannibalism, common among piscivorous fis h, can be a significant contributor to mortality during these early life stages (e.g. in Eurasian perch (Persson et al., 2000), cod [Nuenfeldt and Koster 2000], cape hake [Pilla r and Wilkinson, 1995], pollock [Bailey, 1989; Wespestad et al., 2000], silver hake [Garrison and Link, 2000], and walleye [Liao et al., 2004]). In controlled high density settings (i.e. a quaculture) cannibalism is commonly documented (Baras et al., 2003; Chang and Lio 2003; Tucker, 2003; Correa and Cerqueira, 2007), yet documentation of cannibalism in the wild is less common. This may be attributable to inappropriate collection times or sample sizes (Bowen and Allanson, 1982; Johnson and Dropkin, 1993), difficulties in identi fying digested prey items w ith fast gut evacuation rates (Kennedy, 1969), sampling logistics,(Rindorf, 2004) or high variation in in ter-annual recruitment patterns if cannibalism is density dependent (N uenfeldt and Koster, 200 0). In cod (Uzars and Plikshs, 2000), for example, cannibalism rates are density dependent, with elevated cannibalism during high recruitment, but ne gligible otherwise. 112

PAGE 113

Morphological characteristics (Andraso, 1997; Sillett and Foster, 2000; Magnhagen and Heibo, 2004; Palma and Stenek, 2001) and behavi oral traits (e.g., school ing or ontogenetic habitat shifts) can affect pred ation risk (Gallego and Heath, 1994; Wennhage, 2000). Many of the traits are adaptations that minimize predation risk, such as body depth, spines, or coloration, and can vary allometrically. Understanding how these traits change with body size can give insights into life st ages most vulnerable to ca nnibalism and predation. Cannibalism is common to many species a nd may provide nutri tional and energetic benefits, reduced competition for food and refuge elimination of potential conspecific predators, and reduction in foraging risk for juveniles (see Claesson et al., 2004 for review). Cannibalism risks include choking and suffocating (commonly ev idenced in hatcheries), bodily harm from handling of prey, incorporation of conspecific pathogens, increas ed predation risk by reducing school size, and reduction in incl usive fitness from consuming ki n. Nonetheless, the prevalence of cannibalism in many organisms suggest s a net evolutionary advantage. In Florida, aspects of stock abundance, productivity and regulating mechanisms have been of particular concern for fi shery managers of the common snook Centropomus undecimalis (snook herein). A coastal, warm-water sp ecies, inhabiting estuar ies and river systems (Marshall, 1958; Fore and Schmidt, 1973), snook are subjected to intense recreational fishing pressure in Florida; stocks are considered over-fished, and recruitm ent has declined over the past 20 years (Muller and Taylor, 2005) Widespread habitat loss (Bruger and Haddad, 1986), winter freezes (Storey and Guder, 1936; Marshall, 1958), and red tide blooms also threaten snook populations in Florida. These fact ors have led fishery managers to investigate the feasibility of using stock enhancement to augment snook populat ions in Florida (Brennan et al., 2005; Brennan et al., 2006). Despite the potential benefits, stock enhancement can have deleterious 113

PAGE 114

effects on wild stocks (Hilborn, 2004; Walters and Martell, 2004). As a result, experiments are critical to assess the potential feasibility of stock enhancement prior to implementation of a large-scale program (Blankenship and Leber 199 5; Hillborn, 1998; Walters and Martell, 2004). Cannibalism is one factor that may lead to unintended effects of stock enhancement. In snook, cannibalism has been documented in aquaculture (Serfling, 1998; Tucker, 2003) and is most intense during early life stages in hatcheries (Tucker, 2003; N. Brennan, personal observation). In the wild, snook juvenile s actively prey on small fishes and crustaceans (Fore and Schmidt 1973; Gillmore et al. 1983; Mc Michael et al. 1989; Sazima 2002), and adults are highly piscivorous (Blewitt et al., 2006), yet cannibalism is only rarely documented (e.g., Adams and Wolfe, 2006; N. Brennan, unpublished da ta). Therefore, cannibalism could be perceived as non-significant fact or influencing populat ion size. However, even low observed rates of cannibalism or the threat of cannibalism could strongly influence population dynamics when considering competition for limited refuge in the presence of predators, and prolonged seasonal threats of cannibalism. Identifying cannibalism from dietar y studies in the field may be a relatively rare event compared to more common prey choices of a particul ar predator of various ages, especially when older age classes are not always occupying the same habitats as juvenile conspecifics. A single incidence of cannibalism re sults in mortality of a prey individual and potentially the predator (if the cannibal also chokes and suffocates). Over time (throughout a season or year) cannibalism focused on a year cla ss of juveniles could act as a strong force in structuring abundance of the year class (Claes son et al., 2000; Neuenf eldt and Koster, 2000; Persson et al., 2000), especially during years of high abundance of particular year classes. In many species, ontogenetic changes in mo rphology, and behavior suggests adaptations in earlier stages for coping with smaller absolute body size in the presence of predation and 114

PAGE 115

competitive threats from larger animals shari ng the same habitat (Werner and Gilliam, 1984; Mondor and Roitberg, 2002; Harrel and Gibb, 2006). Such adaptations during early ontogenetic stages can include proportionally larger gapes, predator avoida nce capabilities (Harrel and Gibb, 2006), and a disproportionate tendency to canniba lize conspecifics (Bystrm, 2006). Evidence found in the life history of snook suggests st rong interactive potent ial between various ontogenetic stages. In Florida, snook have a protracted spawning season (Taylor et al., 1998) and subsequent recruitment period to early-stage rearing habitats (McMichael et al., 1989; Peters et al., 1998a). Here, early and late season cohorts share rearing habita t and the potential for competition and cannibalism is high. Evidence for competition and predation within a year class has been documented in other species (e .g. Olson et al., 1995; Miranda and Hubbard, 1994). With snook, the potential for interspeci fic cannibalism is also high because during winter, adult and juvenile share the same habita t (Adams and Wolfe, 2006; also see Chapter 2). Other work (Chapter 3) has also indicated that juvenile snook abundance is density-dependent and may be limited by hab itat availability. In this study I evaluate the ontogenetic potential for snook cannibalism in laboratory and field settings using an allometric approach, a nd estimate the incidence and impact of cannibalism in wild snook stocks. My hypotheses were: (1) ontogenetic differe nces in morphology and coloration reflect high predation risk and cannibalism potential in small snook, (2) the potential for intra-cohort cannibalism decreases with increasing predator size, and (3) cannibalism on age0 snook cohorts in the wild, although infrequently observed, can have a strong influence on population size. I used three approaches to addr ess these hypotheses; (i) quantitative assessment of allometric variation in phenotypic characteristics of snook, (ii) a field experiment to quantify cannibalism rates of different sized predatory sn ook, (iii) a review of published and unpublished 115

PAGE 116

literature on snook diet, and (iv) the development of an age-structured model parameterized with data from i-iii that simulates cannibalism on age-0 snook in a tidal creek. Materials and Methods Generalized Snook Life History in a Tidal Creek Snook in Florida are known to primarily sp awn from May September (Taylor et al., 1998). Spawning grounds include barrier island pa sses and along beaches (Taylor et al., 1998). Post-larvae recruit to a variety of nearshore habitat (Peters et al., 1998a) including estuarine tidal creeks and backwaters through the summer and fa ll (McMichael et al., 1989; Peters et al., 1998b). By November age-0 cohorts can weigh approximately 12 g and remain in these rearing habitats through the winter until the following spring and summer when they undergo an ontogenetic habitat shift to a va riety of habitats including seag rass beds (Gilmore et al., 1983) and beaches (Taylor et al., 1998). By the following winter, at the onset of the first significant cold snap (when water temp eratures drop below about 22oC), yearling snook (about age-1.5) and adults move into winter habitats for thermal refuge (often back to their rearing habitats such as creeks and backwaters where age-0 snook are rear ing; creek temperatur es are often about 5oC warmer than non-creek habitat, Chapter 3). After temperatures warm again in the spring, adolescent and adult snook move back to b eaches, passes and grass flats (Brennan, unpublished data). Approach 1: Phenotypic Characteristics To test hypothesis (1), I examined di fferences in morphological and coloration characteristics of 256 hatchery-reared snook ( 39-395 mm Fork Length [FL]) and 170 wild snook (88-903 mm FL). Hatchery snook (F-1 generation, eggs collected from wild snook) were reared at Mote Marine Laboratory, Center for Aquacu lture Research and Development, located in Sarasota, Florida, USA. Wild snook were collected in creeks and estu aries of Sarasota and 116

PAGE 117

Manatee Counties (Figure 4-1). Measurements in cluded total length (TL) FL, standard length (SL), vertical gape (VG, defined as the vert ical distance between the anterior tip of the premaxlilla and the anterior tip of the dentar y bone, as in Luczkovich et al. 1995), body depth (BD, measured at the widest part of the body, as in Luczkovich et al. 1995), anal spine (AS) length, and distance from the vent ral gill arch (hypohyal bone) to the anus (VA-AN). I also examined allometric changes in the first dorsal fin markings, as juvenile snook have a dominant black spot at the tip of the dorsal fin. Morphometrics were compared graphically against body length (i.e. VG, AS, VA-AN, BD). Because growth is an exponential phenomenon, data were loge transformed and fit to linear regression models with l east squares procedures (with 95% confidence intervals). Therefore, a slope parameter of 0 meant that changes in compared variables were independent, sl opes =1 showed an isometric scaling and slopes around this showed negative (the response variable scale pr oportionally larger with decreasing body length) or positive allometry (the response variable scale proportionally larger with increasing body length). Approach 2: Field Enclosure Trials The second goal of this study was to test th e hypothesis that canniba lism rates are higher for smaller-sized cannibals compared to larger ones. I examined this in field enclosures using a factorial experimental design in which I varied: (1) predator size (small vs. large), (2) relative prey size (30% vs. 50% of the pr edator size), (3) prey density (1 vs. 10 prey per enclosure), (4) the presence of other (non-snook) prey species (yes or no). Th e response variable was binary: the occurrence of cannibalism. Each experimental combination was replicated three times, and the entire design was blocked at tw o field locations, yielding a total of 96 trials (2 predator size classes x 2 relative prey sizes x 2 prey densities x 2 alternate prey x 2 locations x 3 replicates). 117

PAGE 118

Experimental trials were conducted in Sa rasota County, Florida, in twenty field enclosures; ten of these were located in an in land creek estuary (North Creek lagoon), and ten located along the intra-coastal side of a barrier island (Mote site ). Enclosures were 3 m x 3.6 m rectangular, and 2 m high constructed with 3 mm vinyl mesh in a box shape with four sides and a bottom sewn together using dental floss (Figure 4-2) Enclosures were attached at the corners to 2.4 m long steel re-bars (20 mm diameter) implanted in the substrate. Tree branches were placed in each enclosure to provide additional refuge hab itat. The tops of the enclosures were covered with 3 cm square plastic bird mesh. A single snook predator and selected snook prey were stoc ked simultaneously into each enclosure and held for one week. Predator sizes were categorized into small (74-163 mm FL), and large (218-332 mm FL) size classes. Snook prey, were approximately 50% (42-56%), and 30% (28-42%) the length of the pr edator. To test the relative effects of prey density on cannibalism rates, half of the trials were stoc ked with one snook prey (about 0.1 fish per m2), and half with ten snook prey (about one fish per m2). Within each of the above experimental combinations, half of the enclosures were stocked with additional prey species (captured nearby). One week after stocking, animals in the enclosures were harvested by removing the cable ties from the re-bars and lifting the sides and bottoms out of the water. At harvest, all cages were carefully searched for live and dead snook. A cannibalism incidence was assumed to have occurred if prey snook were missing at harvest. Trials where the pr edator was missing or dead at harvest were not included in the an alysis. No stomach contents we re collected. After harvest, a new set of predator/prey combinations were random ly stocked into the encl osures until all trial combinations were completed. Trials were c onducted in October and November, 2003, prior to onset of cooler winter te mperatures. Cannibalism incidence was recorded in a binary format (yes 118

PAGE 119

or no) and was fit to linear regression m odels through maximum likelihood methods using logistic regression techniques (C ATMOD procedure in SAS, [SAS, 1999]). I used Bonferroni adjustments on treatment comparisons to reduce the likelihood of making a Type I error. Approach 3: Cannibalism Intensity in Wild Snook Populations Literature Review: I evaluated cannibalism intensity in wild common snook using published and unpublished literature, (Table 4-1). Published literature was obtained from the Web of Science database using the search criteria Centropomus und ecimalis and diet or feeding or cannibalism. I gained further litera ture by using citations of other work from the reference lists provided in these manuscripts. U npublshed material was obtained from work in Sarasota from the snook stock enhancement research program. Whenever possible, I extracted data on the size and numbers of the snook examin ed, collection year, season, and region of data collection, habitat descriptions of where samp les were taken, the percent of the samples containing food items in the stomach, and th e number of snook repor ted cannibalized. Field Collections: I also assessed cannibalism rates in the field, by conducting my own investigations in Sarasota, Fl orida with snook collected for Mo te Marine Laboratorys Stock Enhancement Research Program from 20012005 (T able 4-1). All collections occurred in estuarine tidal creeks, an important rearing habitat for ag e-0 snook. Snook were collected with bag seines during (1) June 2001 (2) February May 2004, (3) July-August 2004, and (4) MayAugust 2005. All snook were measured and stomach contents were extracted using a non-lethal lavage technique (Adams and Wolfe, 2006), immediately preserved in 10% buffered formalin, and returned to the laboratory for identification. Stomach contents were visually identified to the lowest possible taxonomic level. For more dige sted stomach contents, I used any conspicuous body parts for identification such as carapaces, rostrums, fish jaws and otoliths. Identification of stomach contents beyond this were categorized as unidentified. 119

PAGE 120

Age-Structured Cannibalism Model: Evidence of cannibalism in field collections of snook (Adams and Wolfe, 2006; N. Brennan, unpublis hed data) and other fishes (Lang et al., 2000; Persson et al., 2000; Vik et al., 2001) have shown that predation efforts are typically focused age-0 individuals. As such, I develope d a simplified age-struct ured forecast model to estimate predation mortality from age-0, age-1, and adult (age-2+) snook age groups on an age-0 year class in Sarasota, Florida. I extracted cannibalism rates (number of cannibals observed in stomachs/ total stomachs examined) from studies that were conducted in conditions where cannibalism was most likely to occur. In Sarasota, juvenile a nd adult snook over-winter in tidal creeks (Brennan et al., 2008; Chapter 2 and 3), and emig rate back to summer habitat through the spring (Gilmore et al., 1983, Brennan, unpublished data). In November, snook from the age-0 cohort are generally large enough to be captured by our standard seines (1 cm mesh, snook mean lengths about 100 mm FL). As such, I modeled age-0 cannibalism from November 1 May 15 of a typical y ear in a tidal creek. I used a series of variable s in a cohort-based Monte Carlo simulation model to predict how cannibalism mortality from each of the above age categories decreased the original abundance of the age-0 year class while over-wintering in a typical rearing hab itat. Within the age groups and across all ages, I also measured the contribution of cannibalism mortality to the total natural mortality for the pe riod. Model calculations were ba sed on real data collected in Southwestern Florida in rearing habitats of age-0 snook (tidal cr eek). The model simulation time frame was November 1 May 15 of a hypothetical year. For the model, the number of age-0 snook in the hypothetical creek on November 1 wa s set to 5000. Length frequency data from Sarasota snook collections in tidal creeks from 19982005 (Brennan, unpublished data) were used to obtain typical size estimates of each of the age categories on November 1. Size-specific 120

PAGE 121

instantaneous natural mortality rates ( MW) in a typical population growth function, were used to calculate abundance of the age group i ( age-i) over time (t number of days/365): age-it = age-i, t0 e (M W t) (4-1) Weight-specific instantaneous natural mortality rates extrapolated from Lorenzen (1996) were calculated as a power function of the initial mean weight (W ) of individuals in the age -i group on November 1: MW = Mu Wb (4-2) where, Mu is the annual natural mort ality at unit weight and b is a scaling parameter (Lorenzen, 1996). Because age-1 and age-2+ age groups of snook are typically not fully recruited to the tidal creeks (winter habitat) in November, I used the relative abundance in January (typical peak abundance for all age groups in the creeks) from samples collected in tidal creeks from 1998 2005 (Figure 4-4) to calculate the older age groups abundance based on the age-0 January projected abundances. Then using equations 41 and 4-2 (and weight -based instantaneous mortality rates of these age groups [Lorenzen, 19 96]) I back calculated the age-1 and age-2+ abundance to November 1 (t =0) to obtain age-i,t0. During the simulation period, the model calculated daily abundance estimates for each age group that were weighted according agespecific residence patterns ( p i,t = probability of age-i,t in the creek at time t ) across the time period (see below, Table 4-2, Figure 4-3). For the model period, I assumed cannibalism mortality was proportional to natural mortality of the age-0 cohort. I used field data from snook collections that occurred from 11 November 14 May (i.e. winter and spring data from Adams and Wolfe, [2006], and Brennan; see Table 4-1) to obtain an estimate of th e proportion of the creek population cannibalizing snook ( i, i.e. number of age i snook observed with cannibalized prey in their stomachs/total 121

PAGE 122

number of age i snook stomachs examined). I then calculated daily cannib alism mortality (c ) in the creek by age group i as: ci,t = age-i,t p i,t i (4-3) Total cannibalism ( ci,t ) within and across age groups wa s then calculated to give an estimate of the total number of age-0 snook cannibalized during the model period. Sources of variation in th e model (Table 4-2) were: (1) Initial size (FL, mm) of each age-i group in November. These data were based on length frequency histograms of snook captured in Sa rasota during November (1998 2005, snook stock enhancement program standardized sampling [B rennan, unpublished data]). The model sampled from a normalized distribution, based on the mean length and standard de viation obtained from the field. After this, I used a power exponential weight-length model to convert lengths (FL, mm) to weights ( W in g): Wti = 0.000009 FL (3.0104) (4-4) (2) Instantaneous mortality rates. The model se lected from normalized distributions of the parameter estimates Mu and b (equation 4-2, from Lorenzen ( 1996) with associated standard deviations. (3) Creek residence patterns ( p, expressed as the probability of being in the creek) were calculated differently for each age group based on the best understanding of wild snook abundance in the creeks in Sarasota. This was obtained from (a) length frequency distributions from seasonal sampling in creeks and nearby ha bitats (Chapter 2, 3, Brennan, unpublished data), and (b) recapture data of tagged hatchery-reared and w ild snook (Chapter 2, 3). For age-0 snook, I assumed that they gradually dispersed from th e creeks as the winter and spring progressed, and used an inverse cumulative normal distribution function (mean days in creek = 120, SD = 30). I 122

PAGE 123

varied the parameter, mean days in creek aro und a normalized distribution with SD = 2. For age-1 and age-2+ snook, I assumed they immigrated to the creek from outside habitats to overwinter, and then gradually dispersed from the cr eeks as the season warmed. I used a normal density distribution function, and the model sele cted from normally distributed parameters (mean=50, SD = 5 for age-1 snook, and mean= 50, SD = 2 for age2+ snook). (4) Relative abundance of age-1 snook (compared to the age-0 cohorts abundance) varied within a uniform distribution (0.10 1.51 for ag e-1, 0.05 0.70 for age-2+). (5) Estimates of cannibalism rates ( i) were based on a normalized di stribution of the mean field rate (overall number of canniba ls found/number stomachs examined) with a variance of 1 SD of the mean. I generated variance around these ra tes through bootstrap random sampling of the data set (individual snook of age group i expressed as 0 [no cannibalism] or 1 [cannibal]). I used Microsoft Excel for bootstrapping procedures (Teknomo, 2006) wher e the original data set was randomly sampled to generate 100 data sets (n=1 ,000 per set). I used sensitivity analysis to assess the influence of the above mentioned pa rameter estimates on th e results. The model performed 10,000 iterations using Microsoft Excel with the Crystal Ball add-in (Werckmand et al., 1998). Results Phenotypic Characteristics Mouth gape, body depth, and anal spine length showed negative allometric scaling with body length. Mean slopes from linear relationships of the log transformed variables above were 0.664-0.901 (Table 4-3, and Figure 4-5) with th e anal spine (slope=0.664) showing strong allometric scaling. Isometric scaling was found for log transfor med relationships of SL and FL, and VA-AN and FL, however. The dark spot lo cated at the tip of th e first dorsal fin was 123

PAGE 124

prominent in snook juveniles (30 125 mm FL) but in snook generally la rger than 175 mm FL pigmentation was much reduced and only outli ned the tip of the fin (Figure 4-6). Field Enclosure Trials One hundred one enclosure trials were attemp ted, but of these 27 contained combinations of dead (23) or missing predators (15) or at le ast one dead prey snook (16). Trials with dead prey or predator snook occurred mostly (25) dur ing the experimental trial weeks of November 12, 2003 (12), or November 19, 2003 (13). Th ere was no pattern of the abovementioned mortality between the sites (North Creek lagoon = 11, Mote site = 14). Therefore, 74 trials were used in the analysis, yielding 3-5 replicates per treatment (Table 4-4). The logistic regression model showed a goodness of fit of 0.72 indicating a reasonable fit to the mode but after correcting for Type I errors, no treatments or interactions between factors were significant (Figure 4-8). The overall frequency of cannibalism was 0.307 +/0.057 (SE). Cannibalism Meta-Analysis and Model Results I found 17 studies (published and unpublishe d) describing diets of common snook (Table 4-1). Of these, 11 were conducted in Florida, 1 in Georgia, 2 in Mexico, 1 in Puerto Rico, 1 in the Caribbean, and 1 in Venezuela. Although there was overlap, 13 studies collected samples during summer, 12 during fall, 9 during winter, and 7 during spring. Overall 6,768 snook stomachs were examined for dietary contents an d the reported percentage of snook containing food in their stomachs ranged from 21-89% with a mean of 62% +/5.1 SE. Using knowledge of fish size and collection season, I estimat ed that 2,604 age-0 snook, 1,603 age-1 snook, and 1,753 age-2+ snook were examined for stomach cont ents (Table 4-1). In Sarasota, 2,515 snook were examined for stomach contents, and of these, I estimated 801 age-0 snook, 1,133 age-1, and 518 adult snook. Of these 1,341 snook contained f ood contents in the stomachs (53.3 %). Two studies (Adams and Wolfe, 2006, and Brennan, this work) specifically examined stomach 124

PAGE 125

contents in juvenile habitats during winter (w hen juvenile habitat is shared with older age groups). In these studies, 1,990 snook were examined for stomach contents (Table 4-2), of these 1,405 (70.6%) contained food items Only two the studies reported cannibalism for a total of 7 instances from 6,768 stomachs examined (0.10%). Six (reported by Adams and Wolfe, 2006) were found in nursery habitat in Charlotte Harbor, FL during wint er months (Table 4-5). In Sarasota, I found one instance of cannibalism from creek collections in February 2004 (Table 4-5). Based on body length and collection time, all cannibalized snook were estimated as age-0 (mean length = 93.6 mm SL, range 48-141 mm, mean weight = 8.8 g). Field data confirmed that smaller size snook cannibalized proportionally larger snook prey (Figure 4-9). Model Results After 10,000 Monte Carlo simulations, using all relevant data, the model predicted that cannibalism contributed to 18.9%, 13.3%, and 7% of the accrued age-0 mortality due to predation from age-0, age-1, and age-2+ snook resp ectively (Figure 4-10. Considering initial abundance of the age-0 cohort at the start of the model period (n=5,000, November 1), the model predicted that by May 15, 0.114 of age-0 cohor t had been cannibalized by age-0 snook, 0.174 from age-1 snook, and 0.192 from age-2+ snook. Analysis of the models sensitivity to pa rameter variation, showed that results were strongly controlled by variations in the estimates of field cannibalism rates. More specifically, variations in field cannibalism ra tes within the age-0 cohort contri buted to approximately 95% of the estimates of cannibalism contribution to the age-0s total natural mortality (Figure 4-10). For the same measurement, field based cannibalism rates for age-1 snook had an 88% influence on size-specific results, and age-2+ field estimates of cannibalism rates had a 83% influence on the 125

PAGE 126

results. For all age groups, variations in fiel d cannibalism rates had more than a 99% influence on predictions on the proportion of the initial cohort consumed by cannibalism (Figure 4-11). Aside from field cannibalism rates, variation in th e mean size of the age-0 cohort in November (a density-dependent characteristic) was the second most influential (contributing to 21% for age-0 results) or third (contributing 13% for age-1, a nd 18% for age-2+ results) most influential when predicting contribution of cannibalism to natural mortality (strongly associated Mu and b were also highly ranked). Variations in relati ve abundance had a 40% influence on estimates of the proportion of age-1 cannibalism to total age-0 mortality, and a 46% influence on the estimated proportion of age-2+ cannibalism mortal ity on total age-0 mortality. Variation in creek residence patterns cont ributed to 0.63 2.84% of the estimates of the proportions cannibalism contributions to natural mortality or the proportion of the initial age-0 cohort succumbed to cannibalism. Discussion Juvenile snook show many mo rphological features that are possible adaptations to reduce predation threat, and evidence for potentia l cannibalism during juvenile stages. The proportionally longer anal spine in smaller snook may aid in deterring predators, and even physically block small fish from being swallowe d. In rearing tanks with age-0 snook, I have commonly observed snook with missing body scales fr om the anal spine to the tail, but scales anterior to this were not missing; evidence that th e fish had been held in the mouth of a larger conspecific but not swallowed, most likely due to the anal spine. Snook body depth was proportionally deeper in smaller snook and might also deter similarly-sized predators (including cannibals) from swallowing a snook. Small snook al so exhibit proportionally larger gapes than larger snook allowing a proportionall y greater prey size range for th e smaller fish (see 4-7, 4-9), and thus potentially allowing highe r cannibalism rates. In a study comparing suction feeding 126

PAGE 127

performance in catfish Clarias gariepinus of various sizes (111923 mm TL) (Wassenbergh et al., 2006), absolute suction feeding performance clea rly increased with fish size. Small animals have competitive disadvantages because of thei r smaller body size (Werner and Gilliam, 1984). However, when catfish suction performance was scaled according to predator size and relative prey size, small fish clearly outperformed the la rger fish (Wassenbergh et al., 2006). This work strengthens arguments by Harrel and Gibb (2006) that young animals of smaller sizes, have compensatory adaptations to minimize this difference. These compensatory adaptations would be especially important in fish that cannibalize within a cohort. Cannibalized prey are typically proportionally larger than other prey items (Juanes, 2003), and ca nnibalism is often associated with low alternative prey availability or high competition for prey forcing cannibalism among similarly sized individuals (Persson et al., 2000; Walters and Martell, 2004). Snook coloration suggests high predation threat among juveniles. Sazima (2002) observed that juvenile snook generally resemble mojarra ( Eucinostomus gula ) in overall body shape and have a prominent black spot at the ti p of the dorsal fin. He theorized that snook use aggressive mimicry and school with mojarra to take advantage of unsuspecting prey. I suggest that in addition to aggressive mimicry, this sp ot (prominent in snook smaller than 125 mm, but generally much less distinct in snook larger than 175 mm FL) (see 4-6) is a form of self mimicry, an adaptation to confuse predators, causing them to strike at an appa rent eye spot in a peripheral, less-essential, body location (Blest, 1957; Wickler, 1968; Hailman, 1977; Robbins, 1980). The presence of an additional mark or spot on the body (in add ition to the eyes) is common in schooling species and probably adds to the predator confusio n effect (Welty, 1934; Landeau and Terborgh, 1986). Some birds have flash marks in their plumage possibly as an adaptation to confuse predator s (Brooke, 1998); weasels black-tipped tails are thought to 127

PAGE 128

confuse hawk predators (Powell, 1982). Fish in schools and shoals engage in anti-predator behavior, and group size has been inversely re lated to predator success (Neill and Cullen, 1974; Milinski and Curio, 1975; Major, 1978; Landeau and Terborgh, 1986). Many of these species are commonly marked with a prominent body spot (e.g. Muglidae, Clupeidae Mullidae, Gerreidae, Sciaenidae, Carangidae, Sapridae) th at may aid in the conf usion effect by doubling the number of eyes seen by a predator. Eye sp ots, common in juven iles, reflect ontogenetic changes in predation risk as the eye spots are lost in later life stages (e.g. snook, red snapper, butterfly fish, pork fish, sea bass, pork fish, hog fish, wrasses, and car dinal fish). These ontogenetic variations in morphology of snook even within the age-0 year class (30 175 mm FL) may be adaptations to compensate for small body size in the presence of predation threat and competition (see Werner and Gilliam, 1983; Walters and Martell, 2004; Harrell and Gibb, 2006). Results of cannibalism behavior in the field enclosure trials reflected morphological differences in smalland larger-s ized snook juveniles. Proportional prey sizes tested in the field enclosures were below the maximum gape and length consumption capabilities of the cannibals (82-62% predator size), and ar e probably more realistic. While means were not significantly different between experimental groups (except fo r a significant increase in cannibalism rates when prey were stocked at high density), m ean cannibalism rates of small cannibals were consistently higher than larger cannibals for all th e various experimental treat ments (Figure 4-8). Performing the enclosure studies in the field gave the experiment a more realistic setting for inducing cannibalistic behavior although actual cannibalism rates here were probably lower than if treatments had been conducted in laboratory tanks and higher if done at natural prey densities. Several logistical complications occurred in th is study. Prey and pred ator snook found dead in 128

PAGE 129

the enclosures, appeared to have died many days prior to examination (about 5d or more). Mortality appears to have been handling-indu ced (93% from trials on November 12 and 19, 2003) and this is supported by evidence that the stocked snook died soon after stocking (highly decayed remains). Increased replication of these treatments, the use of controls, and examination of stomach contents to confirm cannibalistic behavior in the encl osures is needed to clarify relative size-specific cannibali sm potential in snook. Snook diets revealed a surprisi ngly low rate of cannibalism (7 cases recorded of 6,768 stomachs examined, 0.10%) in populations over a wide variety of locations. Assuming these rates are reflective of typical snook diet, the average number of snook processed for stomach contents in a single study (mean=359) was just ov er 1/3 the number needed to identify a single cannibalism instance. As snook cannibalism was targ eted at and within the age-0 cohort, sample location and timing could strongly bias cannibalism rates. Samples collec ted in age-0 habitats, especially nursery habitats (see Beck et al., 200 1), are therefore much more likely to contain evidence of cannibalism. Within a given year, ca nnibalism rates are also likely to vary over seasons due to variations in in traand inter-cohort abundance and interactions of size-specific habitat use (e.g. Bystrm et al., 2003). With snook in northern end of their temperature tolerance range, older age classes of snook of ten over-winter in rearing habita ts such as tidal creeks and backwater habitats because of warmer water (M arcinkiewicz, 2007; Chapter 2) In this scenario, over-winter habitats coincide with snook nursery habitats and cannibalism may be intensified. Inter-annual timing of sample collection could also bias results as cannibalism rates are strongly linked to snook density. Samples collec ted during years of low juvenile recruitment will be less likely to contain evidence for canni balism, while cannibalism is much more likely 129

PAGE 130

during years of high juvenile r ecruitment (Henderson and Corps, 1997; Uzars and Plikshs, 2000; Walters and Martell, 2004). Adams and Wo lfe (2006) found exceptionally high rates of cannibalism (6 of 280 snook stomach examined). Their study met two primary requirements for a greater probability of detecting cannibalism : (1) collections were made in putative snook nursery habitat and (2) during winter months when snook of all ages are concentrated in these nursery habitats. In Sarasota, my samples met both of these criteria also, however only one instance of cannibalism was f ound out of 1,710 snook stomachs. Both studies used the same stomach lavage apparatus and techniques. Thes e differences might be explained by differences in juvenile density, a function of inter-annual and spatial recru itment variation. The samples were collected during different y ears and in different regions. S nook habitats in Sarasota have shown large differences in juvenile snook recrui tment in coastal nursery habitats separated by only 19-24 km (Brennan, unpublished data). The di et samples collected in Sarasota were primarily collected during the winter of 2003-2004, a relatively w eak recruitment year for the age-0 snook, but a strong year for age-1 snook (Figure 4-4, relative abundance of age-1 and age0 snook in Sarasota). Model results showed that collective cannibalism by the age groups could contribute substantially to the age-0 natural mortality rates. Because field-based cannibalism rates overwhelmingly influenced the model output, care ful consideration must be given to collection parameters such as sample number, sizes of the predators examined (and associated prey sizes), sampling timing, and inter-annual influences in recruitment and ambient density of the various age groups. In this model, I used a density-dep endent relationship with instantaneous mortality rates linked to body weight (Lonenzen, 1996). Body weight is inversely proportional to density of competitors (Chapter 2, Keeley, 2001) and thus this function accounted for, in part, density130

PAGE 131

dependent competition and predation rates. Further expansion of this model should include a more direct influence of density-dependent func tions, however. In the m odel, the age-specific cannibalism results are, in some ways, independe nt of each other. Additionally, cannibalisminduced mortality could theoretically exceed natu ral mortality rates because the model does not relate ambient age-specific densities in the fiel d (where cannibalism rates are a reflection of the ambient densities) to modeled age-specific densities. Variations in the simulated relative ratio of older age groups to age-0 abundance in January also had a strong influence on model results (40% for output predictions of age-1 cannibalism contributions to age-0 natural mortal ity, and 46% influencing predictions of age-2+ cannibalism contributions to age-0 natural mortal ity). This strong influence of older cohort abundance on age-0 snook cannibalism is reflected in the saw tooth pattern of relative abundance of the two age classes (Figure 4-4). Similar patterns have been observed in other species (Barramundi [Walters and Martell, 2004] Pacific Hake [Smith, 1995], Japanese flounder [Kellison et al., 2002], brown trout [Vik et al ., 2001; Nordwall et al., 2001; Carline, 2006], walleye [Kocovsky and Carline, 2001], walleye Pollock [Yamamura et al., 2001], Sacramento pikeminnow [Gard, 2005], European sea bass [H enderson and Corps, 1997], vendace [Helminen and Sarvala, 1994], freshwater fa iry shrimp [Dumont and Jawahar, 2004], blue crab [Lipcius and Vanengel, 1990]. This suggests that inter-cohort canni balism could have a mediating influence on subsequent strong year classes. Adjustments to other parameters such as re sidency patterns (and associated variation) could also be important in determining canniba lism intensity. Modeled variation around the age-specific creek residence patt erns was small (e.g. age-1 reside nce in the creeks peaked at d50 with only a standard deviation of 5d) and were ba sed on migratory responses to thermal changes. 131

PAGE 132

Inter-annual variation in water temperatures could change the timing and breadth of peak abundance, however. Bioenergetics approaches (e.g. Kitchell et al ., 1977; Rice and Cochran, 1984; Rose et al., 1999; Burke and Rice, 2002) are be coming increasingly popular in fisheries applications (Chips and Wahl, 2008) and can be useful because th ey address physiological and activity-level responses to environmental variati on. Nonetheless, field data has consistently not well correlated with model predictions (Ney, 1993; Bajer et al ., 2004; Selch and Chips, 2007; Chips and Wahl, 2008). Often, prey availability, predator-prey en counter rates, and feed ing rates are primary sources of uncertainty (Ney, 1993; Walters and Martell, 2004; Ba jer et al., 2004; Chips and Wah, 2008), and are difficult to measure in the field (W alters and Martell, 2004). With snook, these and estimates of food conversion efficiencies, ac tivity levels, and physiological responses to abiotic variation (e.g. temperat ure, salinity, and oxygen inter active effects on physiological responses) are not well quantified. Therefore, to minimize unpredictability, in this model I used a simplified approach that bypassed the need fo r predicting many of th e above bioenergetic variables. Much of this model was based on data from field examinations of snook stomachs during winter and spring. While environmental conditions have a str ong influence on gastric evacuation rates (e.g. temperature and salinity, or interactive effects [Vinagre et al., 2007]), the cannibalism model only focused on field collections du ring winter and spring and thus reduced these effects. Prey size (mass) has also been shown to be directly related to residence times in digestive tracts (Jobling et al., 1977; dos Santos and Jobling, 1995; Andersen, 1998). Kennedy (1969) suggested that small-sized prey items were often underrepresented in diet studies, due to faster gastric evacuation rates. For snook, this could strongly bias observations of age-0 cannibalism rates in 132

PAGE 133

the field and actual rates are probably under-represented. In a laboratory study, casual visual examination of differently sized soft-bodied fi sh prey (snook and Gobiidae) in the stomachs of snook held in aquaria over time have shown thes e prey were generally identifiable to species until about 70% of original prey mass remained (Heinlein and Brennan, unpublished data). In the same study, gobies weighing 2g (about 50 mm TL) at consumption were approximately 77% their consumption weight after 4h in the stomach, but 0.5g (about 30 mm TL) gobies were unidentifiable after 4h (17% 24% their original weight re mained) (Heinlein and Brennan, unpublished data). Diet samples are also known to over-repres ent large-bodied prey items (Mann and Orr, 1969; Gannon, 1976). In the laboratory, 15g sno ok prey (mean size=122 mm FL) were more than 93% their original wei ghts 4h after consumption, and only 50% digested by 18 h after consumption by larger snook (Heinlein and Brenna n, unpublished data). These data align with predictions that larger-bodied pr ey are comparatively over-represen ted in diet studies (Mann and Orr, 1969; Gannon, 1976) and thus repor ted cannibalism rates of larger prey are probably closer to actual rates. Snook also demonstrate peak feeding activity during cr epuscular and nocturnal periods (Rock et al., unpublished da ta), and diet samples collected during daylight often contain highly digested prey items that are indistinguishable. All stud ies reported here sampled snook stomachs during daylight hours. Crepuscular an d nocturnal sampling may have revealed higher rates of cannibalism. Further study could identif y the relative influence of these factors on evacuation rates and cannibalism rates for common snook. In conclusion, juvenile snook show many morphological adaptations for predation defense such as elongated anal spines, proportio nally deeper body height, and self mimicry in dorsal fin spot, and also show greater cannibalism potential with proportionally wider mouth 133

PAGE 134

gapes in smaller snook. Field enclosure tria ls indicated that smaller snook had greater cannibalism tendencies than larg er snook, and increased snook de nsities increased cannibalism rates. Snook diet studies to date have s hown sparse evidence of cannibalism, however, cannibalism rates are influenced by density-dependent functions and an interaction between foraging arenas (nursery habitats versus adu lt habitats) and seasonal effects on a populations movement. All cases of cannibalis m recorded have shown that ag e-0 snook are the target prey group for intracohort and intercohort cannibalism and this process could have a mediating influence on recruitment dynamics. This work underlines the necessity for multifaceted annual recruitment monitoring programs and could contri bute to modeling stock enhancement viability assessments and overall snook management programs. 134

PAGE 135

135Reference study Sample Region Season Habitat description JuvnelieSize rangeNo. of stomach'sStomach'sPercentObservedEstimated sample ageKnown dates collected habitat?(mm) Sampled w/contentsw/ contentscannibalsAge-0Age-1Age-2+sizes (num.) Marshall, 1958 1954-1955IRL, West Coast, East Coast FL ? Mangrove shorelinesNo230-851 FL 128 6147.7000128128 Harrington and Harrington, 1961 Indian River Lag oon, FL summer, fallMarshes Yes<100 FL 172 ? ?017200172 Chavez, 1963 Veracruz, Mexico all seasons Mangrove shroelinesYes47-341 TL 60 5185.00??? 0 Linton and Richards, 1965 1963-1964Georgia ? estuarine marsh No24-75 SL 62 2438.706200 62 Austin and Austin, 1971 1967-1968Western Puerto Rico all seasons Mangrove-lined Yes23-87 TL 38 821.003800 38 canals and river banks Rojas, 1975 1972-1973Laguna de Terminos, Campeche, Mexicofall, winter Mangrove wetlandsYesjuveniles mature 32720763.30??? 0 McMichaels et al., 1989 1984-1985Tampa Bay, FL all seasons Grass flats, beaches,Yes11-346 SL 48042789.00395??395 backwaters, rivers Gillmore et al., 1983 1974-1980Indian River Lag oon, FL all seasons Freshwater Yes11-156 SL 99 6089.0 09900 99 Gillmore et al., 1983 1955-1956Indian River Lag oon, FL all seasons Marshes Yes ca. 20275 SL 441 ? ?044100441 Gillmore et al., 1983 1974-1980Indian River Lag oon, FL all seasons Seagrass Yes15-600 SL 321 ? ?0213000321 Luskovitch et al., 1995 1985-1986St. Lucie County, FL summer, fall, winterMangrove impoundmentYes5-119 SL 258 ? ?025800258 Fore and Schmidt, 1973 1971-1973Everglades, FL summer, fallTidal creeks, canalsYes14-196 TL 18314277.5018300183 Fore and Schmidt, 1973 1971-1973Everglades, FL summer, fallMangrove shorelinesYes224-1020 TL 27112746.8000271271 Aliaume et al., 1997 1992-1994Puerto Rico spring, summer, fallRivers, freshwater and Yes<300 SL 43926861.0091?? 91 saltwater lagoons, backwaters Adams and Wolfe, 2006 Feb-Mar 3003Charolette Harbor, FL winter Estuarine ponds, creeksYes141-680 153 9461.43432542110 Adams and Wolfe, 2006 Nov 03-Feb 04Charolette Harbor, FL winter Estuarine ponds, creeksYes190-651 127 3015019169 Blewett et al., 2005 2000-2002Charolette Harbor, FL all seasons Coastal mangrove No300-882 SL 69443262.2000694694 and seagrass Brennan, unpublished data 2001 Sarasota, FL summer Tidal creeks Yes100-250 FL 114 9079.0001140114 Brennan, unpublished data 2004 Sarasota, FL summer Tidal creeks Yes100-735 35316245.901224893353 Brennan, unpublished data 2004 Sarasota, FL winter, springTidal creeks, canalsYes80-965 FL 171087951.417895134081710 Brennan, unpublished data 2005 Sarasota, FL summer Tidal creeks Yes103-466 FL 33821068.10025880338 Total 6768324161.772604160817355947 Table 4-1. Summary of studies examined regarding common snook diets. Juveni le habitat refers to whether samples were collected in juvenile snook ha bitat. The numbers of individual snook repo rted from the various age groups are based on body length information.

PAGE 136

136 Age groupStomach examinationNov.-Dec.FebruaryMarchAprilMayTotalPercent age-0 Number examined 7160502749683842.13 Number of cannibals010001 age-1 Number examined135110320506267834.07 Number of cannibals201003 age-2+ Number examined 1987257615047423.80 Number of cannibals300003 Total 1613571,0791852081,990100.00Table 4-2. Numbers of snook processed for st omach contents and observed instances of cannibalism from 2003 and 2004 winter data in nursery habitats (i.e. from Adams and Wolfe, 2006, and Sarasota data, 2004). Totals and percents are also shown according to age group.

PAGE 137

Table 4-3. Linear regressi on statistics for various morphometric comparisons Comparison NslopeL95U95 R2Ln(VA-AN) Ln(FL)311.0010.9571.0450.987 Ln(SL) Ln(FL) 941.0000.9971.0020.999 Ln(BD) Ln(SL) 940.9010.8920.9150.998 Ln(VG) Ln(SL) 1350.8150.7920.8310.979 Ln(AS) Ln(SL) 3320.6640.6460.6820.940 Note: vertical gape (VG), body depth (BD, anal sp ine (AS), throat to anus (TRH-AN), standard length (SL), and fork length (FL). Prior to regr ession, all values were loge transformed (Ln) Lower (L95) and upper (U95) 95% confid ence limits of the slopes are given. 137

PAGE 138

Table 4-4. Number of replicate experimental treatments (N) of field enclosure trials examining cannibalism according to predator size, the inclusion of additional prey (bait), snook prey density, and two prey si ze categories (expresse d as a percent of predator size) Experimental trials N Predator sizeBaitPrey densityPrey size Large Y 1303 Large Y 1505 Large Y 10304 Large Y 10504Subtotal16 Large N 1305 Large N 1505 Large N 10304 Large N 10505 Subtotal 19 Small Y 1305 Small Y 1505 Small Y 10304 Small Y 10505 Subtotal 19 Small N 1305 Small N 1505 Small N 10305 Small N 10505Subtotal20 Total 74 138

PAGE 139

Table 4-5. Details of actual cannibalism occurrences found in snook stomachs collected in the field. Source SampleCannibalistic snookCannibalized Prey date lengthage LengthWt. (g)% pred. length Adams and Wolfe 2006December-03278 SLage-1 480.717.3 Adams and Wolfe 2006November-03313 SLage-1 773.424.6 Adams and Wolfe 2006March-03312 SLage-1 11311.536.2 Adams and Wolfe 2006January-03591 SLage-2+ 11712.919.8 Adams and Wolfe 2006January-03610 SLage-2+ 12114.419.8 Adams and Wolfe 2006December-03651 SLage-2+ 14123.321.7 Brennan February-04145 FLage-0 641.550.0 139

PAGE 140

Sarasota County Manatee CountyFigure 4-1. Map of Sarasota and Manatee counties where snook co llections and stock enhancement research has been performe d. Arrows indicate the tidal creeks where snook were captured and subsequen tly examined for stomach contents. Map of Florida is offset. 140

PAGE 141

Figure 4-2. Photograph of enclosures where cannibalism tria ls occurred. Steel re-bars supported the material attached at the corners. Bird mesh prevented birds (seen on first enclosure) from preying on the enclosed snook from above. 141

PAGE 142

Number of ageisnook stomachs examined Number of cannibalism occurrences of ageisnook Number of ageisnook stomachs examined Number of cannibalism occurrences of ageisnook Simulated proportion of cannibalized meals Simulated relative abundance in January: (Age-1, Age-2+) Simulated individual mean length.(FL, mm) of age0snook in November Simulated weight-based instantaneous natural mortality rate ( zi) ( Eqn. 4-2 ) Population growth function ( Eqn. 4-1 ) Number of snook cannibalized by ageicohort Proportion of original age0cohort cannibalized by ageisnook Repeat with 10,000 independent simulations Length frequency data and CPUE from standardized sampling in nursery habitats in January Creek residence schedule of ageisnook Mark-recapture data for ageisnook in and around tidal creeks (wild and hatchery-reared recaptures) (Chapter 2 and 3, and Brennan, unpublished data) Length frequency data from standardized sampling in nursery habitats in November. Marine species mortality rates (Lorenzen, 1996) Mean weight of age0cohort in November (Eqn. 4-4) Relative abundance of age-1 or age-2+ Snook in November (Eqn. 4-4) Abundance of age0cohort in November (n = 5000) Proportion of age0mortality due to cannibalism by ageisnook Field dataNovember 1 May 15 Figure 4-3. Schematic of model process predicting cannibalism-related mortality by various age groups of snook. Shaded boxes represent model components where simulated variability occurred. 142

PAGE 143

0 0.5 1 1.5 2 2.5 3 19981999200020012002200320042005 age 0 age-1 adultsRelative AbundanceCollection Year Figure 4-4. Annual relative abundance of age1 and adult snook standardized around age-0 abundance. Data are from annual samples collected in Sarasota during January March of each year. 143

PAGE 144

0.1 0.3 0.5 0.7 0.9 1.1 1.3050100150200250300350400 VG/SL BD/SL (Throat-AN)/SL FL/SL 0 0.05 0.1 0.15 0.2 0.25 0.3 01002003004005006007008009001000Fork length (mm)AS/FL hatchery snook wild snook hatchery snook wild snook Standard Length (mm) Fork Length (mm) Ratio AS/FLA BFigure 4-5. Graphical comparisons of snook morphometrics (in mm). The ratio of anal spine length (AS) to fork length (graph A) declines as a power exponential model with increasing body length. Graph B shows the ratios VG/SL (i.e. VG=vertical gape, SL=standard length), BD/SL, (i.e. BD=body depth), and (throat-AN)/SL (i.e. distance from throat to anus) plotte d against standard length. 144

PAGE 145

Figure 4-6. Photograph of a 70 mm (TL) and a 178 mm (TL) common snook. Black spot at tip of dorsal fin is obvious in smaller snook, but spot is nearly absent in larger snook. Photograph by L. Mitchell. 145

PAGE 146

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7SMALL (74-163mm FL)LARGE (220-332mm FL) 1 PREY 10 PREYMean cannibalism ratePredator size 0 0.1 0.2 0.3 0.4 0.5 0.6SMALL (74-163mm FL) LARGE (220-332mm FL) BAIT NO BAIT 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7SMALL (74-163mm FL) LARGE (220-332mm FL) 30% 50%A B C Figure 4-7. Graphical results from field enclos ure trials describing mean cannibalism rates of small and large snook according to various prey treatments: (A) proportional prey size, (B) snook prey density, and (C) additional bait effect. Error bars are standard error. 146

PAGE 147

Figure 4-8. Plot of body lengths (fork length FL) of snook canni bals and cannibalized prey. Data originally expressed as standard le ngth were converted to fork length (FL). 0 100 200 300 400 500 600 700 0100200300400500600700 Predator length (FL mm)Prey Size (FL mm) Predator length (FL mm) Prey length (FL mm) 147

PAGE 148

0 500 1000 1500 2000 25000.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0 100 200 300 400 500 600 7000.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0 200 400 600 8000.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 Mean = 0.189 +/-0.003 Mean = 0.133 +/-0.008 Mean = 0.070 +/-0.002Age-0 Cannibalism mortality/ Age-0 natural mortalityFrequency age-0 field cannib. Rate 94.760 % Age-0 FL 20.754 % Mu 14.910 % b 8.527 % Age-0 residence 2.838 % Age-1 FL 1.307 % Age-2+ relative abundance0.868 % Age-2+ FL 0.392 % Age-2 residence 0.309 % Age-1 residence 0.308 %Model sensitivity Age-1 field cannib. Rate 88.462 % Age-1 relative abundance 39.981 % Age-0 FL 12.916 % Mu 9.375 % b 4.582 % Age-2+ FL 1.888 % Age-0 residence 1.770 % Age-2+ residence 1.374 % Age-1 FL 0.277 % Age-1 residence 0.258 %Model sensitivity Age-2+ field cannib. rate82.781 % Age-2+ relative abundance46.233 % Age-0 FL 18.197 % Mu 9.964 % b 6.636 % Age-0 residence 2.457 % Age-2+ FL 1.164 % Age-0 field cannib. rate0.926 % Age-1 FL 0.634 % Age-1 relative abundance0.518 %Model sensitivity A B C Figure 4-10. Monte-Carlo simulation results (n=10,000 iterations) showin g proportions of age0 mortality due to age-specific cannibalism: (A) age-0 snook, (B) age-1 snook, and (C) adult snook (>= age-2 snook). Re sults from each sensitivity analysis are inset in each graph. Respective means a nd associated 95% confidence limits of the mean are included. 148

PAGE 149

0 100 200 300 400 500 600 7000.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0 200 400 600 800 1000 12000.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0 100 200 300 400 500 600 700 8000.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 Mean = 0.114 +/-0.002 Mean = 0.174 +/-0.008 Mean = 0.192 +/-0.006 Age-0 field cannib. rate 99.305 % Age-0 FL 8.222 % Mu 4.477 % b 3.172 % Age-0 residence 2.460 % Age-2+ residence 1.352 % Age-1 FL 1.332 % Age-2+ relative abundance0.874 % Age-2 field cannib. Rate0.532 % Age-1 relative abundance0.518 %Model sensitivity Number cannibalized/ original Age-0 abundanceFrequency A Age-1 field cannib. rate 99.964 % Age-1 residence 2.100% Age-0 residence 1.798 % Age-2 residence 1.731 % Age-0 FL 1.629 % Mu 1.199 % Age-2 field cannib. rate1.189 % Age-2 FL 0.836 % b 0.796 % Age-1 FL 0.694 %Model sensitivity Age-2 field cannib. rate 99.949 % Age-2+ residence 2.753 % b 1.818 % Age-2+ FL 1.284 % Mu 1.118 % Age-0 field cannib. rate0.644 % Age-0 residence 0.629 % Age-0 FL 0.582 % Age-1 relative abundance0.361 % Age-1 residence 0.162 %Model sensitivity B C Figure 4-11. Monte-Carlo simulation results (n=10,000 iterations) showing proportions of the original age-0 abundance (at the beginni ng of the model period) cannibalized by the various age groups of snook: (A) ag e-0 snook, (B) age-1 snook, and (C) adult snook (>= age-2 snook). Respective means and associated 95% confidence limits of the mean are included. 149

PAGE 150

CHAPTER 5 CONCLUSION In this dissertation I used stock enhancemen t to understand, test, and reveal mechanisms behind juvenile snook life history, stock dynamics and important ecological and environmental interactions. The practice of stocking fish fo r management purposes has occurred for centuries, yet only within the past 20 years has clear applic ation of a scientific approach occurred (Leber, 1999). With this approach, much progress has occurred including unde rstanding behavioral differences between hatchery and wild conspecifi cs (Olla and Davis, 1989; Hossain et al., 2002), managing fish health and disease transmission from hatchery to wild st ocks (Kennedy et al., 1998; Mushiaki and Muroga, 2004), tagging and marking of released fish (Blankenship, 1990; Bravington et al., 2004; Davis et al., 2004), release technique s and stocking strategies (Tsukamoto, 1989; Yamashita et al., 1994; Leber, 1995; Brennan et al., 2006; Fairchild, 2008), and documenting recruitment to fishery landi ngs (Svsand et al., 1990; Kitada et al., 1999; Kaeriyama, 1999; Zohar et al., 2008). Nonetheless, considerable controversy remain s in this field largely due to unresolved critical uncertainties (Leber et al., 2004; Lore nzen, 2005). Because stake holders are increasingly promoting the use of stock enhancement technology as a fishery management tool, (Walters and Martell, 2004; Hilborn, 2004; Le ber et al., 2004; Lorenzen, 2 008) the need for understanding these uncertainties are becoming increasingly impor tant. Evaluating ecosystem wide and intraand inter-cohort density-dependent responses to stocking is a top priority (Leber et al., 2004; Walters and Martel, 2004). Understanding genetic consequences of hatchery-reared individuals interbreeding with w ild stocks is also a high prio rity (Bell et al., 2008). 150

PAGE 151

In this dissertation, I coupled the release of hatchery-reared juvenile snook with rigorous follow up evaluation, to identify fundamental ecol ogical interactions, a nd important components of snooks life history (Leber, 1999; Miller and Walters, 2004); especi ally during juvenile stages. Having thousands of organisms tagged and experi mentally situated in situ is difficult to accomplish with wild animals (e.g. Leber et al., 1 995). Post-release responses of hatchery-reared snook stocked into various habitats allowed me to make quantitative correl ations with measured habitat characteristics. It also allowed me to make comparisons between post-release recapture ratios and growth rates over time and show how patterns established in rearing habitats can have long-term consequences on production to adult stocks. Monitoring la rge numbers of tagged stocked fish over time also allowed me to identify when and where ontogenetic habitat shifts were occurring (Chapter 2 and 3), and examine fide lity rates in relation to rearing habitats for years after the releases (Chapter 2). I was able to pr edict numerical and biom ass contributions of the released cohorts to adults st ages (Chapter 2). In Chapter 3, I addressed the cr itical issue of density dependent survival and used variations in release magnitudes to examine this potential in post-release survival within a juvenile c ohort of wild and hatchery-reared snook. In Chapter 2, I examined how stocking vari ous rearing habitats in fluenced post-release responses in the hatchery fish, and coupled with other studies (Chapter 3) I was able to project numerical and biomass contribu tions of hatchery snook to a dult stocks. I identified key measures of snook microhabitat, and identified tradeoffs in production to adult stages based on these features. I found a consistent inverse relationship between growth rates and recapture ratios. These responses were co rrelated to release habitats wi th high juvenile snook densities, soft sediments with high organic content, and waters of low salinity, and hypoxic conditions. Release habitats with higher salinity, firmer sediment, a nd higher dissolved oxygen, and low 151

PAGE 152

ambient densities of age-1 snook were correlated with high post-release growth rates, but low recapture ratios. Over time, sites with low growth rates, but high juvenile densities showed the highest numerical contributions of hatchery reared snook to adults stocks, but biomass contributions were among the lowest. Snook stoc ked into rearing habitats with intermediate growth rates had the highest contribu tions to adult biomass however. Others have presented theoretical implicati ons for numerical contributions of nursery habitats to adult stocks (e.g. Beck et al., 2001; Dahlgren et al., 2006), but field studies demonstrating this are rare. Differences in numerical and biomass production could be substantial for species like snook that attain large body sizes, are hi ghly fecund, and have variable growth and survival rate s. Using this work, resource managers can assess the relative production value (in terms of snook) according to ch aracteristics of rearing habitats. Habitats along estuarine shorelines were the focus of this study, but additional work including representative habitats of a pr oduction system is important to ev aluating how habitats contribute to stocks on a broader scale. Quantification of to tal habitat area is also an important next step to evaluate overall production accord ing to habitat type. In Chapter 3 (Brennan et al., 2008), I examined potential density-dependent mortality of hatchery-reared juvenile snook w ith manipulations in stocking magnitude. I varied release magnitudes among four tidal creeks to evaluate the productive capacities of the habitats and determine potential mortality effects on wild and hatchery-reared conspecifics. Creeks receiving high augmentation treatments showed overall increas es in post-release densities, whereas no overall changes were detected in age-1 snook densities in low augmentation creeks. Loss of hatchery-reared snook was remarkably high in one creek (BC) that receiv ed a high augmentation treatment, but loss of wild conspecifics was not de tected. This is consistent with earlier studies 152

PAGE 153

(Chapter 2) where hatchery-reared juveniles, stocked into BC, had lower recapture ratios compared to snook released at other sites. In BC, the overall increase in post-release density, accompanied by high loss rates of hatchery-rear ed snook suggests the productive capacity may have been exceeded, but alternatively may be due to density independent losses, or that it was not conducive to survival of hatchery fish. Work in Chapter 3 examined an important assumption about recruitment processes that density-dependent survival of wild fish may be reduced by th e addition of hatchery-reared fish. In Salmonid populations, evidence exists for density -dependent responses in wild stocks due to stocking hatchery-reared fish (Nickelson et al., 1986; Nickelson, 2003; Kostow and Zhou, 2006). Density-dependent growth responses were not meas ured in this study, but ot her work in the same creeks (Chapter 2), showed suppres sed growth in tributaries with high densities of age-1 snook. I found no evidence of high augmen tation treatment effects on w ild conspecific density, although an effect may have been expre ssed through initial loss of hatchery-reared fi sh in BC where the productive capacity may have been substantia lly exceeded. Abundance pa tterns of wild age-1 snook throughout the study, however, followed similar patterns in all creeks. Clearly post-release surv ival and growth are, in part, f unctions of ambient density of snook juveniles in the rearing habitats, refuge quality, stocking magnitude, predator and prey abundance, encounter rates, and feeding rates. Release magnit udes in 1998 and 1999 (Chapter 2) (hundreds to approximately a thou sand, by release habitat) were similar to stocking magnitudes in the 2002 study (Brennan et al., 2008; Chapter 3), although acclimation techniques were not used in 1998 and 1999 (which adds other pote ntially important logistical complications, (Brennan et al., 2006; Fairchild et al., 2008). Two creeks receiving high augmentation levels in the 2002 study showed very different responses in hatchery snook survival (in one creek 0.85 of 153

PAGE 154

the stocked snook remained by first winter, bu t in the other creek only 0.15-0.36 remained by first winter). Negligible losse s of wild conspecifics were observed in all stocked creeks however. These results were also positively rela ted to ambient density. Sites with high ambient density of wild juveniles also had high post-releas e survival rates of hatchery juveniles, and sites with low ambient densities resulted in low survival rates of the stocked fish. This suggests that while growth responses may have been densitydependent, stocking of age-1 snook (on this order of magnitude) may not have resulted in density-dependent mortality within the cohort, but might be controlled by density-inde pendent effects such as rearing hab itat quality or refuge availability. Stocking at higher magnitudes may have very different results and a cautionary approach is warranted. Inter-annual variation in recruitmen t of snook, its competitors, prey, or predators could all influence the dynamics of density-d ependent responses (Chapter 4 examines the potential influence of intercohort cannibalism, see below). Aside from further potential influence of variability in hatc hery-based components (such as va riation in health of the stocked snook, or complications with rel ease logistics [Congleton et al., 2000; Sulikowski et al., 2006]), stocking the same habitats and release magnitudes could still cause very different results, just due to year-to-year effects. N onetheless, increases in experience with this work (e.g. increases in the number of experimental years, additional stocking levels, and across habitats) can increase confidence in our expectations of such activitie s, however. Coupling th is with experimental modeling scenarios of stock enhancement (see be low), can further add to our understanding of these issues. An important consideration for stocking piscivorous and cannibalistic fishes, such as snook, is the potential influence of intraa nd inter-cohort cannibalism and competition as I examined in Chapter 4. With snook, this potentia l is further intensified because, as in northern 154

PAGE 155

latitudes such as Sarasota, age-0 and older snook share nursery habitat for a prolonged period, Increases in abundance of age-1 snook, due to stocking or naturally, may have negative consequences on young-of-year snook populations through cannibalism and competition for refuge and food resources. Release programs th at elevate age-1 abundan ce could intensify such effects and a cautionary approach is warranted. Nonetheless, the potential for elevating stocks that are below capacity is demonstrated by this st udy and its potential over the long-term must be evaluated. In chapter 4, I firs t examined phenotypic characteris tics among snook of different sizes (39-903 mm FL) to find evidence for variation in cannibalism potential across snook sizes. Smaller snook had proportionally larger mouth gapes (able to consume proportionally larger prey), were deeper bodied a nd had proportionally longer anal spines. These morphological features make predation more difficult, and are likely a reflection of hi gher predation threat among smaller-sized individuals. I also found that snook less than 125 mm FL exhibited a conspicuous dorsal fin spot, but was absent in snook larger than 175 mm FL. I suggested that this spot may be a form of self-mimicry, an adaptation to confuse predators, causing them to strike at an apparent eye spot in a peripheral, less-essent ial, body location. Ontogenetic loss of this spot may reflect decreases in preda tion threat with larger body sizes. I also used field enclosure trials to measur e relative cannibalism rates among differently sized snook. Manipulating predat or size, relative prey size, pr ey density, and the presence of additional prey options, I found higher (although not statistically significant) mean rates of cannibalism in small snook than larger snook,. Small snook had higher cannibalism rates on proportionally larger prey compared to large snook predators, the presence of additional prey species reduced mean cannibalism rates, but again, means were not significantly different. 155

PAGE 156

Increased replication may have reduced variation in the study and results should be verified with examinations of stomach contents and the use of controls. As a further consideration of snook cannibali sm potential, I also examined published and unpublished data to measure rates of cannibalism dete cted in wild snook of different life stages. All cases of cannibalism found were directed at young-of-year snook. These rates were used in a model predicting cannibalism potential in a localized snook stock in a tidal creek habitat through a winter. As prey items in a fishs stomach ge nerally represents a singl e days or nights worth of feeding activity, the collective examination of snook stom ach contents, showed the sample sizes of most studies (hundreds) were grossly in sufficient for detecting cannibalism instances in snook (where thousands needed). Also, because ca nnibalism is generally directed at very small snook, it would be extremely difficult to detect this in the field because of the high evacuation rates of small prey. Field based examination of stomach contents are generally biased towards detecting large prey items, but negatively bias ed against detecting small prey items (Kennedy, 1969. Ignoring this, and just examining observed cannibalism rates in the field, my model showed that a population of juveniles consuming just a few conspecifics throughout a season can collectively and significantly infl uence a stock size, however. My model (Chapter 4) predicted that collective cannibalism by the age groups coul d contribute substantiall y to the age-0 natural mortality rates and cannibalism rates were de nsity-dependent. High cannibalism rates were predicted when strong year classes of age-0 s nook were produced, but low during years of low recruitment of age-0 snook. This work has stro ng implications for stock enhancement programs seeking to numerically supplement particular cohorts of snook. Rigorous experimental monitoring from standard ized sampling efforts over time is integral to understanding basic progress of the stocked or ganisms and hatchery-wild fish interactions. 156

PAGE 157

Work in this dissertation has highlighted the importance of this Long-term sampling programs, especially in juvenile habitats, should be a top priority fo r management agencies tasked with performing stock assessments and understandin g basic biology and ecology of the target organism while keeping a broad focus on the ecos ystem as a whole (Walters and Martell, 2004). Our ongoing standardized sampling program has re vealed important findings on juvenile snook feeding ecology (Rock et al., unpubli shed data; Brennan and Heinle in, unpublished data). Multiyear sampling programs (10 years and longer) dire cted at multiple life stages are extremely powerful for understanding inter-ann ual recruitment and recruitmen t across geographical areas. This dissertation revealed how strong differenc es in inter-annual recruitment could result in dramatic differences in density-dependent mortal ity (i.e. cannibalism) and serves as an example of this. In other work, I identified strong w ithin-year differences in recruitment strength in habitats only kilometers away, during years when red tide blooms were especially prevalent in northern sites of our study ar ea. Multi-year sampling programs allow for adaptive management, a necessary approach in the face of environmental and social variation. Investing this towards stocks subjected to high harvest pr essures and mortality is certainly warranted. A clear next step will be modeling the cost-effectiveness and social benefits involving stakeholders and using various enhancement s cenarios (Lorenzen, 2006; Lorenzen, 2008) of a snook stock enhancement management program in Fl orida. Work in this dissertation provides essential input into such m odels. Whereas snook aquacultur e technology has not yet been developed enough to support large-scale enhanc ement, modeling the costs, benefits, and alternative management approaches should pr ecede any large-scale application of stock enhancement (Hilborn, 1998; Walters and Martell, 2004; Lorenzen, 2008). As snook stocks are fairly localized and non-migratory (Tringali et al., in press[a]), the results of the stock 157

PAGE 158

enhancement work in Sarasota, are useful for providing necessary input into modeling enhancement scenarios and feas ibility in Florida. 158

PAGE 159

159 LIST OF REFERENCES Adams, A.J., Wolfe, R.K., 2006. Cannibalism of juveniles by adult Centropomus undecimalis. Gulf of Mex. Sci. 24 (1-2), 11-13. Adams. A.J., Wolfe, R.K., Pine W.E., Thornton, B.L., 2006. Efficacy of PIT tags and an autonomous antenna system to study juvenile li fe stage of an estuarine-dependent fish. Estuaries Coasts 29 (2), 311-317. Aliaume C, Zerbi, A., Miller, J.M., 1997. Nursery habitat and diet of juvenile Centropomus species in Puerto Rico estuaries. Gulf of Mex. Sci. 15 (2), 77-87. American Sport Fishing Association, 2004. Sport fishing in America: Values of our traditional pastime. Available: www.asafishing.org/asa/ Andersen, N.G., 1998. The effect of meal size on gastric evacuation in whiting. J. Fish. Biol. 2 (4), 743-755. Andersen, A.K., Schou, J., Sparrebohn, C.R., Nicola jsen, H., Stottrup, J.G., 2005. The quality of release habitat for reared juvenile flounder, Platichthys flesus with respect to salinity and depth. Fish. Manage. Ecol. 12 (3), 211-219. Andraso, G.M., 1997. A comparison of startle response in two morphs of the brook stickleback ( Culaea inconstans): Further evidence for a trade-o ff between defensive morphology and swimming ability. Evol. Ecol. 11 (1), 83-90. Armstrong, J.D., Braithwaite, V.A., Huntingford, F.A ., 1997. Spatial strategies of wild Atlantic salmon parr: exploration and settlement in unfamiliar areas. J.Anim. Ecol. 66, 203-211. Austin, H., Austin, S., 1971. The feeding habits of some juvenile marine fishes from the mangroves in western Puerto Rico. Carib. J. Sci., 11(3-4), 171-178. Bailey, K.M., 1989. Interaction between the verti cal distribution of juvenile walleye Pollock, Theragra chalcogramma in the eastern Bering Sea and cannibalism. Mar. Ecol. Prog. Ser. 53, 205-213. Bajer, P.G., Whitledge, G.W., Hayward, R.S ., 2004. Widespread consumption-dependent systematic error in fish bioenergetics models and its impli cations. Can. J. Fish. Aquat. Sci. 61 (11), 2158-2167. Baras, E., Kestemont, P., Melard, C., 2003. Effe ct of stocking density on the dynamics of cannibalism in sibling larvae of Perca fluviatilis under controlled conditions. Aquacult. 219 (1-4), 241-255. Beck, M.W., Heck, K.L. Jr., Able, K.W., Child ers, D.L., Eggleston, D.B., Gillanders, B.M., Halpern, B, Hays, C.G., Hoshino, K., Minello, T.J., Orth, R.J., Sheridan, P.F., Weinstein,

PAGE 160

M.P., 2001. The identification, conservation, an d management of estuarine and marine nurseries for fish and invertebra tes. Bioscience 51 (8), 633-641. Becker, P., Barringer, C., Marelli, D.C., 2008. Thirty years of sea ranching Manila clams ( Venerupis philippinarum ): Successful techniques and lessons learned. Rev. Fish. Sci. 16 (1-3), 44-50. Belk, M.C., 1995. Variation in growth and age at maturity in bluegill sunfish genetic or environmental-effects. J. Fish Biol. 47 (2), 237-247. Belk, M.C., 1998. Predator-induced dela yed maturity in bluegill sunfish ( Lepomis macrochirus ): Variation among populations. Oecologia 113 (2), 203-209. Bell, J.D., Bartley, D.M., Lorenzen, K., L oneragan, N.R., 2006. Restocking and stock enhancement of coastal fisheries: potential, problems and progress. Fish Res. 80 (1), 1-8. Bell, J.D., Leber, K.M., Blankenship, H.L., Loneragan, N.R., Masuda, R., 2008. A new era for restocking, stock enhancement and sea ranching of coastal fisheries resources. Rev. Fish. Sci. 16 (1-3), 1-9. Bennett, J.P., 2006. Using telemetry to assess na tural and fishing mortality of common snook in Sarasota Bay, Florida. Masters thesis. Fi sheries and Aquatic Sciences, University of Florida. Berejikian, B.A., Smith, R.J.F., Tezak, E.P., Schroder, S.L., Knudsen, C.M., 1999. Chemical alarm signals and complex hatchery rearing ha bitats affect antipredator behavior and survival of chinook salmon (Oncorhynchus tshawytscha ) juveniles. Can. J. Fish. Aquat. Sci. 56 (5), 830-838. Berkeley, S.A., Chapman, C., Sogard, S.M., 2004. Maternal age as a de terminant of larval growth and survival in a marine fish, Sebastes melanops. Ecology 85, 1258-1264. Blankenship, H.L., 1990. Effects of time and fi sh size on coded wire ta g loss from Chinook and Coho salmon. In: Parker, N.C., Giorgi, A.E., Heidinger, R.C., Jester Jr., D.B., Prince, E.D., Winans, G.A. (Eds.), Fish marking techniques. American Fisheries Society Symposium 7. Bethesda, MD, pp. 237-243. Blankenship, H.L., Leber, K.M., 1995. A Responsible Approach to Marine Stock Enhancement. In: Schramm Jr., H.L., Piper, R.G. (Eds.), Us es and effects of cultured fishes in aquatic ecosystems. American Fisheries Society Symposium 15. Bethesda, MD, pp.376-387. Blest, A.D., 1957. The function of eyespot patte rns in the Lipidotera. Behaviour 11, 209. Blewett, D.A., Hensley, R.A., Stevens, P. W., 2006. Feeding habits of common snook, Centropomus undecimalis in Charlotte Harbor, Florid a. Gulf Carib. Res. 18, 1-13. 160

PAGE 161

Bohlin, T., Sundstrom, L.F., Johnsson, J.I., Hoje sjo, J., Pettersson, J., 2002. Density-dependent growth in brown trout: effects of introducing wild and hatchery fish. J. Anim. Ecol., 71 (4), 683-692. Botsford, L.W., Castilla, J.C., Peterson, C.H., 1997. The management of fisheries and marine ecosystems. Science 277 (5325), 509-515. Bowen, S.H., Allanson, B.R., 1982. Behavioral and trophic plasti city of juvenile Tilapia mossamica in utilization of the unstable l ittoral habitat. Env. Biol. Fish. 7, 357-362. Bravington, M.V., Ward, R.D., 2004. Microsatellite DNA markers: evaluating their potential for estimating the proportion of hatchery-rear ed offspring in a stock enhancement programme. Mol. Ecol. 13 (5), 1287-1297. Brennan, N.P., 1997. Establishment and management of a multi-state, coded-wire tag, paddlefish database. Masters thesis. Cooperative Fisherie s Research Unit. Tennessee Technological University, Cookeville, Tennessee. 38505. Brennan, N.P., Leber, K.M., Blankenship, H.L., Ransier, J.M., DeBruler Jr., R., 2005. An evaluation of coded wire and elastomer ta g performance in juvenile common snook under field and laboratory conditions. N. Am. J. Fish. Manage. 25, 437-445. Brennan, N.P., Darcy, M.C., Leber, K.M., 2006. Pr edator-free enclosures improve post-release survival of stocked common snook. J. E xp. Mar. Biol. Ecol. 335 (2), 302-311. Brennan, N.P., Leber, K.M., Blackburn, B.R., 2007. Use of coded-wire and visible implant elastomer tags for marine stock enha ncement with juvenile red snapper Lutjanus campechanus Fish. Res. 83 (1), 90-97. Brennan, N.P., Walters, C.J., Leber, K.M., 2008. Manipulations of stocking magnitude: addressing density-dependence in a juvenile cohort of common snook Centropomus undecimalis. Rev. Fish. Sci. 16 (1-3), 215-227. Brooke, M.D., 1998. Ecological fact ors influencing the occurrence of flash marks in wading birds. Funct. Ecol., 12 (3), 339-346. Bruger, G.E., 1980. Preliminary analyses of snook, Centropomus undecimalis population dynamics in the Naples-Marco Island region of southwest Florida. Fl a. Sci. 43 (suppl.1), 25. Bruger, G.E., Haddad, K.D., 1986. Management of tarpon, bonefish, and snook in Florida. Stroud, R.H. (Ed.). Multi-jurisdictional ma nagement of marine fishes. National Coalition for Marine Conservati on, Inc., Savannah, Georgia, 53-57. 161

PAGE 162

Burke, B.J., Rice, J.A., 2002. A linked foraging a nd bioenergetics model for southern flounder. Trans. Am. Fish. Soc. 131 (1), 120-131. Bystrm, P., Persson, L., Wahlstrom, E., We stman, E., 2003. Sizeand density-dependent habitat use in predators: consequences for habitat shifts in young fish. J. Anim. Ecol. 72 (1), 156-168. Bystrm, P., Andersson, J., Persson, L., De Roos, A.M., 2004. Size-dependent resource limitation and foraging-predation risk trade-offs: growth and habitat use in young arctic char. Oikos 104, 109-121. Bystrm, P., 2006. Recruitment pulses induce cannibali stic giants in Arctic char. J. Anim. Ecol. 75 (2), 434-444. Campbell, P., 1997. Population projections: States 1995-2025. U.S. Census Bureau. http://www.census.gov/prod/2/pop/p25/p25-1131.pdf Carline, R.F., 2006. Regulation of an unexpl oited brown trout population in Spruce creek, Pennsylvania. Trans. Am. Fish. Soc., 135 (4), 943-954. Carr, S.D., Hench, J.L., Luettich, R.A., Forward, R.B., Tankersly, R.A., 2005. Spatial patterns in the ovigerous Callinectes sapidus spawning migration: results from a coupled behavioral-physical model. Mar. Ecol. Prog. Ser. 294, 213-226. Carscadden, J.E., Frank, K.T., Leggett, W.C., 2001. Ecosystem changes and the effects on capelin ( Mallotus villosus ), a major forage species. Can. J. Fish. Aquat. Sci. 58 (1), 7385. Caselle, J.E., Warner, R.R., 1996. Va riability in recruitm ent of coral reef fishes: The importance of habitat at two spatial s cales. Ecology 77 (8), 2488-2504. Chambers, R.C., Leggett, W.C., 1996. Maternal infl uences on variation in egg sizes in temperate marine fishes. Am. Zool. 36, 180-196. Chang, E. Y. Y., Leo, I.C., 2003. Sibling cannibalism of young red drum, Sciaenops ocellatus in relation to size disparity and metabolic rates. Env. Biol. Fish. 68 (4), 407-415. Chavez, H., 1963. Contribucion al conocimiento de la biologia de los robalos, chucumite y constantino ( Centropomus spp.) del estado de Veracruz (p isc., Centrop.). Ciencia, Mex. 22, 141-161. Chen, Y., Jackson, D.A., Harvey, H.H., 1992. A comparison of von Bertalanffy and polynomial functions in modeling fish growth data Can. J. Fish. Aquat. Sci. 49, 1228-1235. Chipps, S.R., Wahl, D.H., 2008. Bioenergetics modeling in the 21st cen tury: Reviewing new insights and revisiting old c onstraints. Trans. Am. Fish. Soc. 137 (1), 298-313. 162

PAGE 163

Claessen, D., de Roos, A.M., Persson, L ., 2000. Dwarfs and giants: Cannibalism and competition in size-structured populations. Am. Nat. 155 (2), 219-237. Claessen, D., de Roos, A.M., Persson, L., 2004. Population dynamic theory of size-dependent cannibalism. Proc. Roy. Soc. Lond. Se r. B: Biol. Sci. 271 (1537), 333-340. Clarke, C.W., Levy, D.A., 1988. Diel vertical migrations by juvenile sockeye salmon and the antipredation window. Am. Nat. 131, 271 290. Conder, A.L., Annear, T.C., 1987. Test of wei ghted usable area estimates derived from a PHABSIM model for instream flow studies on tr out streams. N. Am. J. Fish. Manage 7, 339-350. Congleton, J.L., LaVoie, W.J., Schreck, C.B., Da vis, L.E., 2000. Stress indices in migrating juvenile Chinook salmon and stee lhead of wild and hatchery origin before and after barge transportation. Trans. Am. Fish. Soc. 129 (4), 946-961. Connell, S.D., Jones, G.P., 1991. The influence of habitat complexity on post-recruitment processes in a temperate reef fish populati on. J. Exp. Mar. Biol. Ecol., 151 (2), 271-294. Correa, C.F., Cerqueira, V.R., 2007. Effects of stocking density and size distribution on growth, survival and cannibalism in juvenile fat snook ( Centropomus parallelus Poey). Aquacult. Res. 38 (15), 1627-1634. Dahl, J., Pettersson, E., Dannewitz, J., Jarvi, T ., Lof, A.C., 2006. No difference in survival, growth and morphology between offspring of wi ld-born, hatchery and hybrid brown trout ( Salmo trutta ). Ecol. Freshw. Fish 15 (4), 388-397. Dahlgren, C.P., Kellison, G.T., Adams, A.J., Gillanders, B.M., Kendall, M.S., Layman, C.A., Ley, J.A., Nagelkerken, I., Serafy, J.E., 2006. Marine nurseries and effective juvenile habitats: concepts and ap plications. Mar. Ecol. Prog. Ser. 312, 291-295. Davis, J.L.D., Young-Williams, A.C., Aguilar, R. Carswell, B.L., Goodison, M.R., Hines, A.H., Kramer, M.A., Zohar, Y., Zmora, O., 2004. Differences between hatchery-raised and wild blue crabs: implications for stock enha ncement potential. Trans. Am. Fish. Soc. 133 (1), 1-16. Davis, J.L.D., Young-Williams, A.C., Hines, A. H., Zohar, Y., 2005. Assessing the potential for stock enhancement in the case of the Chesapeake Bay blue crab ( Callinectes sapidus ). Can. J. Fish. Aquat. Sci. 62 (1), 109-122. Davis, S.M., Childers, D.L., Lorenz, J.J., Wanless, H.R ., Hopkins, T.E ., 2005. A conceptual model of ecological interacti ons in the mangrove estuaries of the Florida Everglades. Wetlands, 25 (4), 832-842. 163

PAGE 164

De Innocentiis, S., Longobardi, A., Marino, G ., 2008. Molecular tools in a marine restocking program for the endangered dusky grouper, Epinephelus marginatus. Rev. Fish. Sci. 16 (1-3), 269-277. Doherty, P. J., and A. Fowler. 1994. An empirica l test of recruitment lim itation in a coral reef fish. Science 263, 935-939. Doherty, P.J., Dufour, V., Ga lzin, R., Hixon, M.A., Meekan, M.G., Planes, S., 2004. High mortality during settlement is a population bottleneck for a tr opical surgeonfish. Ecol. 85 (9), 2422-2428. Dorsey, L.G., 2004. Retention of coded wire tags by age-0 Muskellunge. J. Freshwater Ecol. 19 (2), 333-337. dos Santos, J., Jobling, M., 1991. Factor s affecting gastric evacuation in Cod, Gadus morhua-L ., fed single-meals of natural pr ey. J. Fish Biol. 38, 697-713. Duffy, J.T., Epifanio, C.E., Fuiman, L.A., 1997. Mortality rates imposed by three scyphozoans on red drum ( Sciaenops ocellatus Linnaeus ) larvae in field enclosur es. J. Exp. Mar. Biol. Ecol. 212 (1), 123-131. Dumont, H.J., A. Jawahar, A., 2004. Stage-specific cannibalism and spontaneous cyst hatching in the freshwater fairy shrimp Streptocephalus proboscideus Frauenfeld. Hydrobiologia, 524, 103-113. Einum, S., Sundt-Hansen, L., Nislow, K.H., 2006. The partitioning of density-dependent dispersal, growth and survival throughout ontogeny in a highly fecund organism. Oikos 113 (3), 489-496. Fairchild, E.A., Rennels, N., Howe ll, W.H., 2008. Predators are a ttracted to acclimation cages used for winter flounder stock enhancement. Rev. Fish. Sci. 16 (1-3), 262-268. FAO, 2004. The state of world fisheries and aqua culture: 2004. FAO, Rome (also available at ftp://ftp.fao.org/docrep/fao/007/y5600e/y5600r00.pdf ) Faunce, C.H., Serafy, J.E., 2006. Mangroves as fish habitat: 50 years of field studies. Mar. Ecol.-Prog. Ser. 318, 1-18. Fleming, I.A., Petersson, E., 2001. The ability of released, hatchery salmonids to breed and contribute to the natural producti vity of wild populations. Nord ic J. Freshw. Res. 75 (5), 71-98. Fore, P.L. Schmidt, T.W., 1973. Bi ology of juvenile and adult snook, Centropomus undecimalis, in the Ten Thousand Islands, Florida. U.S. Environmental Protection Agency, Surveillance and Analysis Division. Pub lication number EPA 904/9-74-002, Chapter 16. Region IV, Atlanta, GA. 164

PAGE 165

Francis, R.C., Hixon, M.A., Clarke, M.E., Murawski, S.A., Ralston, S.A., 2007. Ten Commandments for ecosystem-based fisheries scientists. Fisheries 32 (5), 217-233. Frankiewicz, P., Dabrowski, K., Martyniak, A., Zalewski, M., 1999. Cannibalism as a regulatory force of pikeperch, Stizostedion lucioperca (L.), population dynamics in the lowland Sulejow reservoir (Central Pola nd). Hydrobiologia 408, 47-55. Fretwell, S.D., 1972. Populations in a Seasonal Environment. Princeton University Press, Princeton, New Jersey. Fritts, A.L., Scott, J.L., Pearsons, T.N., 2007. The effects of domestication on the relative vulnerability of hatchery and w ild origin spring Chinook salmon (Oncorhynchus tshawytscha ) to predation. Can. J. Fish. Aquat. Sci. 64 (5), 813-818. Fromentin, J.M., Gjosaeter, J., Bjornstad, O.N., Stenseth, N.C., 2000. Biological processes and environmental factors regulating the dyna mics of the Norwegian Skagerrak cod populations since 1919. ICES J. Mar. Sci. 57 (2), 330-338. Furuta, S., 1996. Predation on juvenile Japanese flounder ( Paralichthys olivaceus ) by diurnal piscivorous fish: field obse rvations and laboratory experi ments. In Watanabe, Y., Yamashita, Y., Oozeki, Y. (Eds.), Survival St rategies in Early Life Stages of Marine Resources. A. A. Balkema, Rotterdam, pp. 285-294. Gallego, A., Heath, M.R., 1994. The development of schooling behavior in Atlantic herring Clupea harengus. J. Fish Biol. 45 (4), 569-588. Gannon, J.E., 1976. The effects of differential di gestion rates of zooplankton by alewife, Alosa pseudoharengus on determinations of selective feeding. Trans. Am. Fish. Soc., 105, 8995. Garaway, C.J., Arthur, R.I., Chamsingh, B., Ho mekingkeo, P., Lorenzen, K., Saengvilaikham, B., Sidavong, K., 2006. A social science pers pective on stock enhancement outcomes: lessons learned from inland fisheries in southern Lao PDR. Fish. Res. 80, 3745. Gard, M.F., 2005. Ontogenetic microhabitat shifts in Sacramento pikeminnow, Ptychocheilus grandis: reducing intraspecifi c predation. Aquat. Ecol., 39 (2), 229-235. Garrison, L.P., Link, J.S., 2000. Diets of five hake species in the northeast United States continental shelf ecosystem. Mar. Ecol. Prog. Ser. 204, 243-255. Gibson, R.N., Pihl, L., Burrows, M.T., Modin, J., Wennhage, H., Nickell, L.A., 1998. Diel movements of juvenile plaice Pleuronectes platessa in relation to predators, competitors, food availability and abiotic factors on a microtidal nursery ground. Mar. Ecol. Prog. Ser. 165, 145-159. 165

PAGE 166

Gilmore, R.G., Bullock, L.H., F. H. Berry. 1978 Hypothermal mortality in marine fishes on south-central Florida. NE Gulf Sci. 2 (2), 77-97. Gilmore, R.G., Donohoe, C.J., Cooke, D.W., 1983. Observations on the distribution and biology of east-central Florida popul ations of the common snook, Centropomus undecimalis ( Bloch). Fl. Sci. 46, 313-336. Grant, S.M., Brown, J.A., 1998. Nearshore settleme nt and localized populations of Atlantic cod ( Gadus morhua) in shallow coastal waters of Newfoundland. Can. J. Fish. Aquat. Sci. 55 (6), 1317-1327. Gwak W.S., Tanaka, Y., Tominaga, O., Tsusak i, T., Tanaka, M., 2003. Field evaluation by RNA/DNA ratios on post-releas e nutritional status of rel eased and wild Japanese flounder Paralichthys olivaceus juveniles. J. Exp. Mar. Biol. Ecol. 293 (1), 107-124. Hailman, T.P., 1977. Optical signals: animal communication and light. Indiana University Press, Bloomington. Hallegraeff, G.M., 1993. A review of harmful alga l blooms and their apparent global increase. Phycologia 32 (2), 79-99. Harrell, A., Gibb, A.C., 2006. Ontogeny of performa nce in vertebrates. Phys. Biochem. Zoo., 79 (1), 1-6. Helminen, H., Sarvala, J., 1994. Population regulation of vendace ( Coregonus albula) in lake Pyhajarvi, southwest Finland. J. Fish. Biol., 45 (3), 387-400. Hamilton, L.C., Haedrich, R.H., 1999. Ecol ogical and population changes in fishing communities of the North Atlantic. Arc. Polar Res., 18 (2), 383-388. Hampton, S.E., 2004. Habitat overlap of enemies: temporal patterns and the role of spatial complexity. Oecologia 138 (3), 475-484. Harrington, R.W., Harrington, E.S., 1961. Food selection among fishes invading a high subtropical marsh: from onset of flooding th rough the progress of a mosquito brood. Ecology 42, 646-666. Henderson, P.A., Corps, M., 1997. The role of temperature and cannibalism in interannual recruitment variation of bass in British waters. J. Fish. Biol., 50 (2), 280-295. Hilborn, R., 1998. The economic performance of ma rine stock enhancement projects. Bull. Mar. Sci. 62 (2), 661-674. Hilborn, R., 2004. Population management in stoc k enhancement and sea ranching. In: Leber, K.M., Kitada, S., Blankenship, H.L., Svsand, T. (Eds.), Stock enhancement and sea ranching. Blackwell Publishing, Oxford, UK, pp. 201. 166

PAGE 167

Hilborn, R, Walters, C.J., 1992. Quantitative Fisheries Stock Assessm ent: Choice, Dynamics and Uncertainty. Chapman and Hall, New York, NY. Honma, A., 1993. Aquaculture in Ja pan. Japan FAO Association, Tokyo. Hossain MAR, Tanaka M, Masuda R., 2002. Pred ator-prey interaction be tween hatchery-reared Japanese flounder juvenile, Paralichthys olivaceus and sandy shore crab, Matuta lunaris: daily rhythms, anti-predator c onditioning and starvation. J. Exp. Mar. Biol. Ecol 267 (1), 1-14. Houde, E.D., 1987. Fish early life dynamics and r ecruitment variability. In: Houde, E.D., (Ed), 10th Annual Larval Fish Conference. Am erican Fisheries Society, Symposium 2, Bethesda, MA, pp. 17-29. Hurst, T.P., Conover, D.O., 1998. Winter mortal ity of young-of-the-year Hudson River striped bass ( Morone saxatilis ): size-dependent patter ns and effects on recruitment. Can. J. Fish. Aquat. Sci. 55, 2709-2709. Hutchings, J.A., 2000. Collapse and recovery of marine fishes. Nature 406, 882-885. Immamura, K., 1999. The organization and developm ent of sea farming in Japan. In Howell, B.R., Moksness, E., Svsand, T., (Eds.), Stock Enhancement and Sea Ranching (Ed. B.R. Howell, E. Moksness, and T.), Blackwell Sc ientific Publications, Oxford, UK, pp. 91101. Islam, M.S., Haque, M., 2004. The mangrove-bas ed coastal and nearshore fisheries of Bangladesh: ecology, exploitation and manage ment. Rev. Fish Biol. Fish. 14 (2), 153180. Jackson, J.B.C., Kirby, M.X., Berger, W.H., Bjor ndal, K.A., Botsford, L.W., Bourque, B.J., Bradbury, R.H., Cooke, R., Erlandson, J., Estes, J.A., Hughes, T.P., Kidwell, S., Lange, C.B., Lenihan, H.S., Pandolfi, J.M., Peterson, C.H., Steneck, R.S., Tegner, M.J., Warner, R.R., 2001. Historical overfishi ng and the recent collapse of coastal ecosystems. Science 293 (5530), 629-638. Jakobsson, A., Eriksson, O., 2000. A comparative st udy of seed number, seed size, seedling size and recruitment in grassland plants. Oikos 88 (3), 494-502. Jefferts, K.B., Bergman, P.K., Fiscus, H.F., 1963. A coded wire identification system for macroorganisms. Nature 198, 460-462. Jenkins, W.E., Denson, M.R., Bridgham, C.B., Co llins, M.R., Smith, T.I.J., 2002. Retention of oxytetracycline-induced marks on Sagittae of re d drum. N. Am. J. Fish. Manage. 22 (2), 590-594. 167

PAGE 168

Jobling, M., Gwyther, D., Grove, D.J., 1977. Some effects of temperature, meal size and body weight on gastric evacuation time in the dab Limanda limanda (L) J. Fish Biol. 10, 291298. Jokikokko, E., Kallio-Nyberg, I., Saloniemi, I., Jutil a, E., 2006. The survival of semi-wild, wild and hatchery-reared Atlantic salmon smolts of the Simojoki River in the Baltic Sea. J. Fish Biol. 68 (2), 430-442. Johnson, J.H., Dropkin, D.S., 1993. Diel variation in diet composition of a riverine fish community. Hydrobiologia 271, 149-158. Jonsson, N., Jonsson, B., Hansen, L.P., 2003. The marine survival and growth of wild and hatchery-reared Atlan tic salmon. J. App. Ecol. 40 (5), 900-911. Jones, G.P., 1997. Relationships between recrui tment and postrecruitment processes in lagoonal populations of two coral reef fishes. J. Exp. Mar. Biol. Ecol. 213 (2), 231-246. Jorstad, K., 2004. Genetic studies in marine stoc k enhancement in Norway. In: Leber, K.M., Kitada, S., Blankenship, H.L., Svsand, T. (E ds.), Stock enhancement and sea ranching. Blackwell Publishing, Oxford, UK, pp. 339-352. Juanes, F., 2003. Allometry of cannibalism in pi scivorous fishes. Can. J. Fish. Aquat. Sci. 60, 594-602. Kaeriyama, M., 1999. Hatchery programmes and stock management of Salmonid populations in Japan. In Howell, B.R., Moksness, E., Svsand, T., (Eds.), Stock Enhancement and Sea Ranching (Ed. B.R. Howell, E. Moksness, a nd T.), Blackwell Scientific Publications, Oxford, UK, pp. 153-167. Keeley, E.R., 2001. Demographic responses to food and space competition by juvenile steelhead trout. Ecol. 82 (5), 1247-1259. Kellison, G.T., Eggleston, D.B., Tanaka, M., 2002. Density-dependent predation and implications for stock enhancement with Japa nese flounder. J. Fish. Biol., 60 (4), 968980. Kellison, G.T., Eggleston, D. B., 2004. Coupling ecology and economy: modeling optimal release scenarios for summer flounder ( Paralichthys dentatus ) stock enhancement. Fish. Bull. 102 (1), 78-93. Kemp, W.M, Sampou, P.A., Garber, J., Tuttle, J., Boynton, W.R., 1992. Seasonal depletion of oxygen from bottom waters of Chesapeake Bay roles of benthic and planktonic respiration and physical exchange processes. Mar. Ecol. Prog. Ser. 85 (1-2), 137-152. Kennedy, C.R. 1969. Tubificid oligochaetes as food of dace. J. Fish Biol. 1, 11-15. 168

PAGE 169

Kennedy, T.B., Vinyard, G.L., 1997. Drift ecology of western catostomid larvae with emphasis on Warner suckers, Catostomus warnerensis (Teleostei). Env. Biol. Fish. 49 (2), 187195. Kennedy, S.B., Tucker, J.W., Neidig, C.L., Vermeer, G.K., Cooper, V.R., Jarrell, J.L., Sennett, D.G., 1998. Bacterial management strategi es for stock enhancement of warmwater marine fish: a case study with common snook (Centropomus undecimalis). Bull. Mar. Sci. 62 (2), 573-588. Kitada, S., 1999. Effectiveness of Japans stock enhancement programmes: current perspectives. In Howell, B.R., Moksness, E., Svsand, T., (Eds.), Stock Enhancement and Sea Ranching (Ed. B.R. Howell, E. Moksness, a nd T.), Blackwell Scientific Publications, Oxford, UK, pp. 103-131. Kitchell, J.F., Stewart, D.J., Weininger, D., 1977. Applications of a bioenergetics model to yellow perch ( Perca flavescens ) and walleye ( Stizostedion vitreum vitreum ). J. Fish. Res. Board Can. 34: 1922-1934. Kocovsky, P.M., Carline, R.F., 2001. Dynamics of the unexploited walleye population of Pymatuning Sanctuary, Pennsylvania, 1997-1998. N. Am. J. Fish. Manage. 21 (1), 178187. Koczaja, C., McCall, L., Fitch, E., Glorioso, B., Hanna, C., Kyzar, J., Niemiller, M., Spiess, J., Tolley, A., Wyckoff, R., Mullen, D., 2005. Size-specific habitat segregation and intraspecific interacti ons in banded sculpin (Cottus carolinae ). Southeast Nat. 4 (2), 207218. Kostow, K.E., Zhou, S.J., 2006. The effect of an introduced summer steelhead hatchery stock on the productivity of a wild winter steelhead population. Trans. Am. Fish. Soc. 35 (3), 825841. Krause, J., Loader, S.P., McDermott, J., Ruxton, G.D., 1998. Refuge use by fish as a function of body length-related metabolic expenditure and predation risks. Proc. Roy. Soc. Lond. Ser. B: Biol. Sci. 265 (1413), 2373-2379. Kristiansen T.S., Svsand, T., 1990. Enhancement studies of coastal cod in western Norway. Part III. Interrelationships between reared and indigenous cod in a nearly land-locked fjord. J. Con. Int. Explor. Mer 47, 23-29. Krivan ,V., Vrkoc, I., 2000. Patch choice under pr edation hazard. Theor. Popul. Biol. 58 (4), 329-340. Landeau, L., Terborgh, J., 1986. Oddity and the confusion effect in predation. Anim. Behav., 34, 1372-1380. 169

PAGE 170

Lang, G.M., Brodeur, R.D., Napp, J.M., Schabe tsberger, R., 2000. Variation in groundfish predation on juvenile walleye pollock relative to hydrographic structure near the Pribilof Islands, Alaska. ICES J. Mar. Sci. 57 (2), 265-271. Lau, S.R., Shafland, P.L., 1982. Larval development of snook, Centropomus undecimalis (Bloch) (Pices: Centropomidae). Copeia 1982, 618-627. Leber, K.M., 1995. Significance of fish size at rel ease on enhancement of striped mullet fisheries in Hawaii. J. World Aquacult. Soc. 26, 143 153. Leber, K.M., Brennan, N.P., Arce, S.M., 1998. Recruitment patterns of cultu red juvenile Pacific threadfin, Polydactylus sexfilis (Polynemidae) released along sandy marine shores in Hawaii. Bull. Mar. Sci. 62 (2), 389-408. Leber, K.M., 1999. Rationale for an experimental approach to stock enhancement. In Howell, B.R., Moksness, E., Svsand, T., (Eds.), Stock Enhancement and Sea Ranching (Ed. B.R. Howell, E. Moksness, and T.), Blackwell Scient ific Publications, Oxford, UK, pp. 63-75. Leber, K.M., 2004. Marine Stock Enhancement in the USA: Status, trends and needs. In: Leber, K.M., Kitada, S., Blankenship, H.L., Svsand, T. (Eds.), Stock enhancement and sea ranching. Blackwell Publishing, Oxford, UK, pp. 11-24. Leber, K.M., Brennan, N.P., Arce, S.M., 1995. Marine enhancement with striped mullet: are hatchery releases replenishing or displacing wi ld stocks? In: Schram m Jr., H.L., Piper, R.G. (Eds.), Uses and effects of cultured fishes in aquatic ecosystems. American Fisheries Society Symposium 15. Bethesda, MD, pp.376-387. Leber, K.M., Arce, S.M., Sterritt, D.A., Brennan, N.P., 1996. Marine stock enhancement potential in nursery habita ts of striped mullet, Mugil cephalus in Hawaii. Fish. Bull. 94, 452-471. Leber, K.M, Blankenship, H.L., Arce, S.M., Brennan, N.P., 1997. The influence of release season on size-dependent surv ival of striped mullet, Mugil cephalus in a Hawaiian estuary. Fish. Bull. 95 (2), 267-279. Leber, K.M., Kitada, S., Blankenship, L., Svsand, T., 2004. Stock Enhancement and Sea Ranching: Developments, Pitfalls and Oppor tunities. Blackwell Publishing, Oxford, United Kingdom. Leber, K.M., Cantrell, R.N., Leung, P.S., 2005. Optim izing cost-effectiveness of size at release in stock enhancement programs. N. Am. J. Fish. Manage. 25, 1596-1608. Lenore, C., Greenberg, A., Eaton, A., 1998. Standa rd methods for the examination of water and wastewater. 1998. Section 2540. American Public Health Association, American Waterworks Association, Wa ter Environment Federation. United Book Press, Inc., Baltimore, Maryland. 170

PAGE 171

Liao, H.S., Pierce, C.L., Larscheid, J.G., 2004. Consumption dynamics of the adult piscivorous fish community in Spirit Lake, Iowa. North American Journal of Fisheries Management 24 (3), 890-902. Linehan, J.E., Gregory, R.S., Schneider, D.C., 2001. Predation risk of age-0 cod ( Gadus ) relative to depth and substrate in coastal waters. J. Exp. Mar. Biol. Ecol. 263, 25-44. Linton, T. L., Richards, W. L. 1965. Young comm on snook on the coast of Gerogia. Quart. J. Florida Acad. Sci. 28 (2), 185-189. Lipcius, R.N., Vanengel, W.A., 1990. Blue -crab population-dynamics in Chesapeake Bay variation in abundance (York River, 1972-1988) and stock-recr uitment functions. Bull. Mar. Sci. 46 (1), 180-194. Lochet, A., Lambert, P., Lepage, M., Rochard, E., 2004. Growth comparison between wild and hatchery-reared juvenile European sturgeons Acipenser sturio (Acipenseridae) during their stay in the Gironde estuary (F rance). CYBIUM 28 (1), 91-98 Suppl. S. Longhurst, A., 2006. The sustainabili ty myth. Fish. Res. 81, 107-112. Lotze, H.K., Lenihan, H.S., Bourque, B.J., Br adbury, R.H., Cooke, R.G., Kay, M.C., Kidwell, S.M., Kirby, M.X., Peterson, C.H., Jackson, J.B.C., 2006. Depletion, degradation, and recovery potential of estuaries and co astal seas. Science 312 (5781), 1806-1809. Lorenz, J.J., 1999. The response of fishes to physicochemical changes in the mangroves of northeast Florida Bay. Estuar ies 22 (2B), 500-517. Lorenz, J.J., Sarafy, J.E., 2006. Subtroprical wetland fish a ssemblages and changing salinity regimes: Implications for everglades restoration Hydrobiologia 569, 401-422 Lorenzen, K., 1996. The relationship between body weight and natural mortality in fish: a comparison of natural ecosystems and aquaculture. J. Fish Biol. 49, 627-647. Lorenzen, K., 2000. Allometry of natural mortalit y as a basis for assessing optimal release size in fish-stocking programmes. Ca n. J. Fish. Aquat. Sci. 57, 2374-2381. Lorenzen, K., 2005. Population dynamics and potentia l of fisheries stock enhancement: practical theory for assessment and policy analysis. Philos. T. Roy. Soc. B. 260, 171-189. Lorenzen, K., 2006. Population management in fisheries enhancement: gaining key information from release experiments through use of a size-dependent mortality model. Fish. Res. 80, 19-27. Lorenzen, K., 2008. Understanding and managing fisheries enhancement systems. Rev. Fish. Sci. 16 (1-3), 10-23. 171

PAGE 172

Lorenzen, K, Enberg, K., 2002. Density-dependent growth as a key mechanism in the regulation of fish populations: evidence from amongpopulation comparisons. Proc. Roy. Soc. Lond. Ser. B: Biol. Sci. 269 (1486), 49-54. Love, M. S., Carr, M.H., Haldorson, L.I., 1991. The ecology of substrate-associated juveniles of the genus Sebastes. Env. Biol. Fish. 30 (1-2), 225-243. Luczkovich, J. J., Norton, S.F., Gilmore Jr., R.G ., 1995. The influence of oral anatomy on prey selection during the ontoge ny of two percoid fishes, Lagodon rhomboids and Centropomus undecimalis Env. Biol. Fish. 44, 79-95. MacPherson, T.A., Cahoon, L.B., Mallin, M.A., 2007. Water column oxygen demand and sediment oxygen flux: patterns of oxygen deplet ion in tidal creeks. Hydrobiologia 586, 235-248. Madon, S.P., Williams, G.D., West, J.M., Zedler, J.B., 2001. The importance of marsh access to growth of the Ca lifornia killifish, Fundulus parvipinnis evaluated through bioenergetics modeling. Ecol. Model. 136 (2-3), 149-165. Magnhagen, C., Heibo, E., 2004. Growth in length and in body depth in young-of-the-year perch with different predation risk. J. Fish Biol. 64 (3), 612-624. Major, P.F., 1978. Predator-prey inte ractions in two schooling fishes, Caranx ignobilis and Stolephorus purpureus Anim. Behav., 27, 855-865. Mann, R.H. K., and D.R. Orr. 1969. A preliminary st udy of the feeding relationships of fish in a hard-water and a soft-water stream in southern England. J. Fish Biol., 1, 31-44. Marcinkiewicz, L., 2007. Examining seasonal m ovement and habitat use patterns of adult common snook. Masters thesis. Fisheries and A quatic Sciences, Univers ity of Florida. Marshall, A.N., 1958. A survey of the snook fishery of Florida, with st udies of the biology of the principal species, Centropomus undecimalis (Bloch). Florida Board of Conservation, Marine Research Laboratory Technical Series No. 22. Mason, J.W., Brynilson, O.M., Degurse, P.E., 1967. Co mparative survival of wild and domestic strains of brook trout in streams. Trans. Am. Fish. Soc. 96, 313-319. Masuda, R., 2004. Behavioral approaches to fish stock enhancement: a practical review. In: Leber, K.M., Kitada, S., Blankenship, H.L ., Svsand, T. (Eds.), Stock enhancement and sea ranching. Blackwell Publis hing, Oxford, UK, pp. 83-90. Mather, J.A., Odor, R.K., 1991. Fo raging strategies and predation risk shape the natural-history of juvenile octopus-vulgaris Bull. Mar. Sci. 49 (1-2), 256-269. 172

PAGE 173

Maynard, D.J., Flagg, T.A., Mahnken, C.V.W., Schroder, S.L. 1996a. Natural rearing technologies for increasing postrelease survival of hatchery -reared salmon. Bulletin of National Research Institute, Supp. 2, 71-77. Maynard, D.J, McDowell, G.C., Tezak, E.P., Flagg, T.A., 1996b. Effects of diets supplemented with live food on the foraging be havior of cultured fall chinook salmon. Prog. Fish Cult. 58, 187-191. McMichael Jr., R.H., Peters, K., Parsons, G.R., 1989. Early life history of the snook Centropomus undecimalis in Tampa Bay, Florida. Northeast Gulf Sci. 10, 113-125. Meeuwig, M.H., Puls, A.L., Bayer, J.M., 2007. Su rvival and tag retenti on of Pacific lamprey larvae and macrophthalmia marked with codedwire tags. N. Am. J. Fish. Manage. 27 (1), 96-102. Metcalfe, N.B., Fraser, N.H.C., Burns, M.D., 1998. State dependent shifts between nocturnal and diurnal activity in salm on. Proc. Roy. Soc. Lond. Ser. B: Biol. Sci. 265 (1405), 1503-1507. Metcalfe, N.B., Steele, G.I., 2001. Changing nutritional status causes a shift in the balance of nocturnal to diurnal activity in European Minnows. Funct. Ecol. 15 (3), 304-309 Milinski, M., Curio, E., 1975. Experiments on the eco-ethology of predator-prey relations.. Verh. Dtsch. Zool. Ges. no. 68, pp. 81-89. Miller, J.M., Walters, C.J., 2004. Experimental ec ological tests with stocked marine fish. In: Leber, K.M., Kitada, S., Blankenship, H.L ., Svsand, T. (Eds.), Stock enhancement and sea ranching. Blackwell Publis hing, Oxford, UK, pp. 142-152. Miranda, L.E., Hubbard, W.D., 1994 Length-dependent winter survival and lipid-composition of age-0 largemouth bass in bay springs reser voir, Mississippi. Trans. Am. Fish. Soc. 123 (1), 80-87. Mittelbach, G. G. 1986. Predator-mediated habitat use: some consequences for species interactions. Env. Biol. Fish. 16, 159 169. Mol, J.H., 1996. Impact of predation on early st ages of the armoured catfish Hoplosternum thoracatum (Siluriformes-Callichthyidae) and implications for th e syntopic occurrence with other related cat fishes in a neotropical multi-pre dator swamp. Oecologia 107 (3), 395-410. Mondor, E.B., Roitberg, B.D., 2002. Pea aphid, Acyrthosiphon pisum cornicle ontogeny as an adaptation to differential predation risk. Can. J. Zoo.-Rev. Can. Zoo. 80 (12), 2131-2136. 173

PAGE 174

Mooij, W.M., Van Densen, W.L.T., Lammens, E.H. R.R., 1996. Formation of year-class strength in the bream population in the shallow eutrophic Lake Tjeukemeer. J. Fish Biol. 48 (1), 30-39. Muller, R.G., Taylor, R.G., 2005. The 2005 stock assessment update of common snook, Centropomus undecimalis Florida Fish and Wildlife Conservation Commission, Florida Marine Research Institute. Available: www.floridamarine.org Mumby, P.J., Edwards, A.J., Arias-Gonzalez, J. E., Lindeman, K.C., Blackwell, P.G., Gall, A., Gorczynska, M.I., Harborne, A.R., Pe scod, C.L., Renken, H., Wabnitz, C.C.C., Llewellyn, G., 2004. Mangroves enhance the biom ass of coral reef fish communities in the Caribbean. Nature 427 (6974), 533-536. Munro, J.L., Bell, J.D., 1997. Enhancement of ma rine fisheries resources. Rev. Fish. Sci. 5, 185222. Mushiake, K., Muroga, K., 2004. Fish health management in seed production. In: Leber, K.M., Kitada, S., Blankenship, H.L., Svsand, T. (E ds.), Stock enhancement and sea ranching. Blackwell Publishing, Oxford, UK, pp. 119-129. Myers, R.A., 2001. Stock and recruitment: generalizations about maximum reproductive rate, density dependence, and variability using meta -analytic approaches. ICES J. Mar. Sci. 58 (5), 937-951. Neill, S.R.St. J, Cullen, J.M., 1974. Experiments on whether schooling by their prey affects the hunting behavior of cephalopods and fish predators. J. Zool. Lond. 172, 549-569. Nemeth, R.S., 1998. The effect of natural variati on in substrate architect ure on the survival of juvenile bicolor damselfish. Env. Biol. Fish. 53 (2), 129-141. Neuenfeldt, S., Koster, F.W ., 2000. Trophodynamic control on recruitment success in Baltic cod: the influence of cannibalism ICES J. Mar. Sci. 57, 300-309. Ney, J.J., 1993. Bioenergetics modeling today gr owing pains on the cutting edge. Trans. Am. Fish. Soc. 122(5), 736-748. Nickelson, T., 2003. The influenc e of hatchery coho salmon ( Oncorhynchus kisutch) on the productivity of wild coho salmon populations in Oregon coastal basins. Can. J. Fish. Aquat. Sci. 60 (9), 1050-1056. Nickelson, T.E., Solazzi, M.F., Johnson, S.I., 1986. Use of hatchery coho salmon ( Oncorhynchus kisutch) presmolts to rebuild wild populations in Ore gon coastal streams. Can. J. Fish. Aquat. Sci. 43 (12), 2443-2449. Norcross, B.L., 1984. Oceanic and estuarine transpor t of fish eggs and larvae a review. Trans. Am. Fish. Soc. 113 (2), 153-165. 174

PAGE 175

Nordwall, F., Naslund, I., Degerman, E., 2001. Intercohort competition effects on survival, movement, and growth of brown trout ( Salmo trutta ) in Sweedish streams. Can. J. Fish. Aquat. Sci., 58 (11), 2298-2308. Olla, B.L., Davis, M.W., 1989. The role of le arning and stress in predator avoidance of hatchery-reared coho salmon (Onchorhynchus kisutch) smolts in seawater. Aquacult. 76, 209-214. Olla, B.L., Davis, M.W., Ryer, C.H., 1998. Understanding how the hatchery environment represses or promotes the development of behavi oral survival skills. Bull. Mar. Sci. 62, 531-550. Olson, M.H., Mittelbach, G.G., Osenberg, C.W., 1995. Competition between predator and prey: Resource-based mechanisms and implications for stage-structured dynamics. Ecology 76 (6), 1758-1771. Orth, D.J., Maughan, O.E., 1982. Evaluation of the incremental methodology for recommending instream flows for fishes. Trans. Am. Fish. Soc. 111 (4), 413-445. Palma, A.T., Steneck, R.S., 2001. Does variable co loration in juvenile ma rine crabs reduce risk of visual predation? Ecology 82 (10), 2961-2967. Pauly, D., Christensen, V., Dalsgaard, J., Froese, R., Torres Jr., F., 1998. Fishing down marine food webs. Science 279 (6), 860-863. Pauly, D., Christensen, V., Walters, C., 2000. Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impact of fisheries ICES J. Mar. Sci. 57 (3), 697-706. Prez-Pinzn, M.A, Lutz, P.L., 1991. Activity rela ted cost of osmoregulation in the juvenile snook ( Centropomus undecimalis). Bull. Mar. Sci. 48 (1), 56 66. Pergent-Martini C, Pergent G., 2000. Marine phaner ogams as a tool in the evaluation of marine trace-metal contamination: an example from the Mediterranean. Int. J. Env. Poll. 13 (16), 126-147. Perry, R.I., Walters, C.J., Boutillier, J.A., 1999. A framework for providing scientific advice for the management of new and developing inverteb rate fisheries. Rev. Fish. Biol. Fisher. 9 (2), 125-150. Persson. L., 1993. Predator-mediated competition in prey refuges the importance of habitat dependent prey resources. Oikos 68 (1), 12-22. Persson, L, Eklov, P., 1995. Prey refuges affecti ng interactions between piscivorous perch and juvenile perch and roach. Ecology 76 (1), 70-81. 175

PAGE 176

Persson, L., Bystrom, P., Wahlstrom, E., 2000. Cannibalism and competition in Eurasian perch: population dynamics of an ontogenetic omnivore. Ecol 81 (4), 1058-1071. Persson, L., de Roos, A.M., 2006. Size-structured interactions and the dynamics of aquatic systems. Pol. J. Ecol. 54 (4), 621-632. Persson, L., de Roos, A.M., Bystrm, P., 2007. State-dependent invasi on windows for prey in size-structured predator-prey systems: whole lake experiments. J. Anim. Ecol. 76 (1), 94-104. Peters, K.M., Matheson, Jr., R.E., Taylor, R.G ., 1998a. Reproduction and early life history of common snook, Centropomus undecimalis (Bloch), in Florida. Bull. Mar. Sci. 62 (2), 509-529. Peters, K.M., Giordano, S.D., Adams, J.M., Fe nwick, R.W., McWilliams, R.W., Berill, N.J., 1998b. Factors influencing the distribution of larvae of the common snook, Centropomus undecimalis (Bloch), in Southwestern Fl orida. Fla. Mar. Res. Publ., 53, 33 pp. Peterson, M.S., Gilmore Jr., R.G., 1991. Eco-physiology of juvenile snook Centropomus undecimalis (Bloch): life history implications. Bull. Mar. Sci. 48 (1), 46 57. Pillar, S.C., Wilkinson, I.S., 1995. The diet of Cape hake Merluccius capensis on the south coast of South Africa. S. Afr. J. Mar. Sci. 15, 225-239. Powell, R.A., 1982. Evolution of black-tipped tails in weasels: predator confusion. Am. Nat. 119, 126-131. Prentice, E.F., Flagg, T.A., McCutcheon, C.S., 1990. Feasibility of using implantable passive integrated transponder (PIT) tags in salm onids. In: Parker, N.C., Giorgi, A.E., Heidinger, R.C., Jester Jr., D.B., Prince, E.D., Winans, G.A. (Eds.), Fish marking techniques. American Fisheries Soci ety Symposium 7. Bethesda, MD, pp.317-322. Railsback, S.F., Stauffer, H.B., Harvey, B.C., 200 5. What can habitat pr eference models tell us? Tests using a virtual trout popula tion. Ecol. App. 13 (6), 1580-1594. Raimondi, P.T., 1990. Patterns, mechanisms, and cons equences of variability in settlement and recruitment on an intertidal barnacle. Ecol. Monog. 60, 283. Rice, J.A., Cochran, P.A., 1984. Independent evaluation of a bioenergetics model for largemouth bass. Ecol. 65 (3), 732-739. Rose, K.A., Cowan, J.H., Clark, M.E., Houde, E.D., Wang, S.B., 1999. An individual-based model of bay anchovy population dynamics in the mesohaline region of Chesapeake Bay. Mar. Ecol. Prog. Ser. 185, 113-132. Rindorf, A., 2004. Estimation of evacuation rate s in the field. J. Fish. Biol. 65, 261-284. 176

PAGE 177

Robbins, R.K., 1980. The lycaenid false head hypot hesis: historical re view and quantitative analysis. J. Lepid. Soc. 34, 194-208. Rojas, J.C., 1975. Contribucion al conocim iento de la biologia de los robalos Centropomus undecimalis y C. Poeyi en la Laguna de Terminos, Ca mpeche, Mexico. Biol. Inst. Oceanogr. Univ. Oriente 14 (1), 51-70. Ross, S.W., 2003. The relative value of different estuarine nursery areas in North Carolina for transient juvenile marine fishes Fish. Bull. 101 (2), 384 404. Roussel, J.M., Bardonnet, A., 1999. Ontogeny of diel pattern of stream-margin habitat use by emerging brown trout, Salmo trutta in experimental channels: influence of food and predator presence. Env. Bi ol. Fish. 56 (1-2), 253-262. Rozas, L.P., Odum, W.E., 1988. Occupation of submerged aquatic vegetation by fishes: testing the roles of food and ref uge. Oecologia 77, 101-106. Russell, D.J., Rimmer, M.A., McDougall, A.J ., Kistle, S.E., Johns ton, W.L., 2004. Stock enhancement of barramundi, Lates calcarifer (Bloch), in a coastal river system in Northern Australia: stocking strategies, survival and benefit-cost. In: Leber, K.M., Kitada, S., Blankenship, H.L., Svsand, T. (E ds.), Stock enhancement and sea ranching. Blackwell Publishing, Oxford, UK, pp. 490 500. Ruzycki, J.R., Wurtsbaugh, W.A., 1999. Ontogeneti c habitat shifts of juvenile Bear Lake sculpin. Trans. Am. Fish. Soc. 128 (6), 1201-1212. Sanchez-Lamadrid, A., 2002. Stock e nhancement of gilthead sea bream ( Sparus aurata, L .): assessment of season, fish size and place of re lease in SW Spanish coast. Aquaculture 210 (1-4), 187-202. Sanderson, B.L., Hrabik, T.R., Magnuson, J.J., Post, D.M., 1999. Cyclic dynamics of a yellow perch ( Perca flavescens ) population in an oligotrophic la ke: evidence for the role of intraspecific interactions. Can. J. Fish. Aquat. Sci. 56 (9), 1534-1542. SAS Institute, 1998. SAS/Stat Users Guide. Versi on 6: volumes 1 and 2. Cary, North Carolina. Savino, J.F., Stein, R.A., 1982. Predator-prey inter action between largemouth bass and bluegills as influenced by simulated, submersed vegetation. Trans. Am. Fish. Soc. 111 (3), 255266. Sazima, I., 2002. Juvenile snooks (Centropomidae) as mimics of mojarras (Gerreidae), with a review of aggressive mimicry in fishes. Env. Biol. Fish. 65, 37-45. 177

PAGE 178

Seitz, R.D., Lipcius, R.N., Seebo, M.S., 2005. Food av ailability and growth of the blue crab in seagrass and unvegetated nurseries of Chesap eake Bay. J. Exp. Mar. Biol. Ecol. 319 (12), 57-68 Sp. Iss. SI Sekino, M., Saitoh, K., Yamada, T., Yamashita, Y., .2005. Genetic tagging of released Japanese flounder ( Paralichthys olivaceus ) based on polymorphic DNA markers. Aquaculture 244(1-4), 49-61. Selch, T.M., Chipps, S.R., 2007. The cost of capturing prey: measuring largemouth bass ( Micropterus salmoides ) foraging activity using glyc olytic enzymes (lactate dehydrogenase). Can. J. Fish. Aquat. Sci. 64 (12), 1761-1769. Serfling, S.A., 1998. Breedi ng and culture of snook, Centropomus undecimalis in a closedcycle, controlled environmental culture system. World Aquaculture Society. Aquaculture Book of Abstracts. P. 482. February 15-19, 1998. Las Vegas, USA. Shacklee, J.B., Bentzen, P., 1998. Ge netic identification of stocks of marine fish and shellfish. Bull. Mar. Sci. 62, 589 621. Shafland, P.L, Foote, K.J.,1983. A lower lethal temperature for fingerling snook, Centropomus undecimalis. Northeast Gulf Sci. 6 (2),175-177. Shimizu, D., Sakiyama, K., Sakakura, Y., Ta katani, T., Takahashi, Y., 2007. Predation differences between wild and hatchery-reared tiger puffer Takifugu rubripes juveniles in a salt pond mesocosm. Nippon Suisan Gakkaishi 73 (3), 461-469. Sillett, K.B., Foster, S.A., 2000. Ontogenetic ni che shifts in two populations of juvenile threespine stickleback, Gasterosteus aculeatus that differ in pelvic spine morphology. Oikos 91 (3), 468-476. Smith, P.E., 1995. Development of the population biology of the Pacific hake, Merluccius productus. CA Coop. Ocean. Fish. Invest., 36, 144-152. Smith, C.C., Fretwell, S.D., 1974. Optimal balance between size and number of offspring. Am. Nat. 108 (962), 499-506. Sogard, S.M. 1992. Variability in growth-rates of juvenile fishes in different estuarine habitats. Mar. Ecol. Prog. Ser. 85 (1-2), 35 53. Solazzi, M.F., Nickelson, T.E., Johnson, S.L., 1991. Survival, contribution, and return of hatchery coho salmon ( Oncorhynchus kisutch ) released into freshwater, estuarine, and marine environments. Can. J. Fish. Aquat. Sci. 48, 248-253. Steele, D. H., Andersen, R., Green, J.M., 1992. The managed commercial annihilation of northern cod. Newf. Stud., 8 (1), 34-68. 178

PAGE 179

Steneck, R.S., 2006. Is the american lobster, Homarus americanus, overfished? A review of overfishing with an ecological ly based perspective. Bull. Mar. Sci 78 (3), 607-632. Stottrup, J.G., Sparrevohn, C.R., Modin, J., Lehma nn, K., 2002. The use of releases of reared fish to enhance natural populat ions A case study on turbot Psetta maxima (Linne, 1758). Fish. Res. 59 (1-2), 161-180. Story, M., Gruder, E.W., 1936. Mortality of fishes due to cold at Sani bel Island, Florida. Ecology 18(1), 10-26. Sulikowski, J.A., Fairchild, E.A., Rennels, N., Howell, W.H., Tsang, P.C.W., 2006. The effects of transport density on cortisol leve ls in juvenile winter flounder, Pseudopleuronectes americanus J. World Aqua. Soc. 37 (1), 107-112. Svsand, T., 1998. Cod Enhancement Studies in Norway Background and Results with Emphasis on Releases in the Peri od 1983-1990. Bull. Mar. Sci. 62,313-324. Svsand, T., Kristiansen, T.S., 1990. Enhancemen t studies of coastal cod in western Norway, part IV. Mortality of reared cod after rele ase. J. Con. Cons. Int. Explor. Mer 47, 30-39. Svsand, T., Kristiansen T.S., Pedersen, T, Salvanes, A.G., Engelsen, V.R., Nvdal, G., Ndtvedt, M., 2000. The enhancement of cod stocks. Fish and Fish. 1, 173 205. Taylor, D.L., Danila, D.J., 2005. Predation on winter flounder ( Pseudopleuronectes americanus) eggs by the sand shrimp ( Crangon septemspinosa ). Can. J. Fish. Aquat. Sci. 62 (7), 1611-1625. Taylor, R.G, Grier, H.J., Whittington, J.A ., 1998. Spawning rhythms of common snook in Florida. J. Fish Biol. 53 (3), 502-520. Taylor, R.G., Whittington, J.A., Grier, H.J., Crab tree, R.E., 2000. Age, growth, maturation, and protandric sex reversal in common snook, Centropomus undecimalis from the east and west coasts of South Florida. Fish. Bull. 98 (3), 612-624. Taylor, M.D., Palmer, P.J., Fiel der, D.S., Suthers, I.M., 2005. Responsible estuarine finfish stock enhancement: an Australian perspe ctive. J. Fish. Biol. 67 (2), 299-331 Teknomo, K., 2006. Bootstrap Sampling Tutori al. http://people.re voledu.com/kard/ tutorial/bootstrap/ Titelman J., Hansson L.J., 2006. Feeding rates of the jellyfish Aurelia aurita on fish larvae. Mar. Biol.149 (2), 297-306. Tolley, S.G., Dohner, E.T., Peebles, E. B., 1987. Occurrence of larval snook, Centropomus undecimalis (Bloch), in Naples Bay, Florida. Florida. Sci. 50 (1), 34-38. 179

PAGE 180

Tomiyama, T., Watanabe, M., Fujita, T., 2008. Community-based stock enhancement and fisheries management of the Japanese flounder in Fukushima, Japan. Rev. Fish. Sci. 16 (1-3) 146-153. Travis, J., Coleman, F., Grimes, C., Conover, D., Bert, T., Tringali, M., 1998. Critically assessing stock enhancement: an introduction to the Mote symposium. Bull. Mar. Sci. 62 (2), 305-311. Tringali, M.D., Leber, K.M., Halstead, W.G., McMichael, R., OHop, J., Winner, B., Cody, R., Young, C., Neidig, C.L., Wolfe, C., Forstchen, A., Barbieri, L., 2008a. Marine stock enhancement in Florida: a multi-disciplinary, stakeholder-supported, accountability-based approach. Rev. Fish. Sci. 16 (1-3), 51-57. Tringali, M.D., Seyoum, S., Wallace, E.M., Higham, M., Taylor, R.G., Trotter, A.A., Whittington, J.A., 2008b. Limits to the use of contemporary genetic analyses in delineating biological populations for restocking and stock enhancement. Rev. Fish. Sci. 16 (1-3), 111-116. Tsukamoto, K., Kuwada, H., Hirokawa, J., Oya, M., Sekiya, S., Fujimoto, H., Imaizumi, K., 1989. Size-dependent mortal ity of red sea bream, Pagrus major juveniles released with fluorescent otolith tags in News Ba y. J. Fish Biol. 35 (Supp. A), 56-69. Tsukamoto, K., Masuda, S., Endo, M., Otake, T., 1996. Behavioral characteristic of the ayu, Plecoglossus altivelis as predictive indices for stocking effectiveness in rivers. Nippon Suisan Gakkaishi, 56, 1177-1186. Tucker, J., 2003. Snook culture. World Aqua. 43 (4), 42-47. Tucker, J.W., Campbell, S.W., 1988. Spawning season of common snook along the east central Florida coast. Fla. Sci. 51 (1), 1-6. Tupper, M. Boutilier, R.G., 1997. Effects of hab itat on settlement, growt h, predation risk and survival of a temperate reef fish. Mar. Ecol. Prog. Ser. 151 (1-3), 225-236. Uki, N. 2006. Stock enhancement of the Japanese scallop Patinopecten yessoensis in Hokkaido. Fish. Res. 80, 62-66. Uzars, D., Plikshs, M., 2000. Cod ( Gadus morhua L. ) cannibalism in the Central Baltic: interannual variability a nd influence of recruit abundance and distribution. ICES J. Mar. Sci. 57 (2), 324-329. Victor, B.C., 1983. Recruitment and population dynamics of a coral reef fish. Science 219, 419. Victor, B.C., 1986. Larval settlement and juvenile mortality in recruitment-limited coral reef fish population. Ecol. Monogr. 56, 145 160. 180

PAGE 181

Vik, J.O., Borgstrm, R., Skaala, ., 2001. Canni balism governing mortality of juvneile brown trout, Salmo trutta in a regulated stream. Regul. Ri vers: Res. Manage., 17 (4-5), 583594. Vinagre,C., Maia, A., Cabral, H.N., 2007. Effect of temperature and salinity on the gastric evacuation of juvenile sole Solea solea and Solea senegalensis. J. App. Ichth. 23 (3), 240-245. Vitousek, P.M., Aber, J.D., Howarth, R.W., Likens, G.E., Matson, P.A., Schindler, D.W., Schlesinger, W.H., Tilman, D.G., 1997. Human alteration of the global nitrogen cycle: Sources and consequences. Ecol. App. 7 (3), 737-750. Volk, E.C., Schroder, S.L., Grimm, J.J., 1999. Ot olith thermal marking. Fish. Res. 43 (1-3), 205-219. Volpe, A.V., 1959. Aspects of the biology of the common snook, Centropomus undecimalis (Bloch), of southwest Florida. Florida State Board of Conservation, Technical Series Number 31. Walters, C.J., 1997. Challenges in the management of riparian and coastal ecosystems. Cons. Ecol. 1(2), http://www.consecol.orgt/vol1/iss2/art1 Walters, C.J., Juanes, F., 1993. Recruitment limitations as a consequence of natural selection for use of restricted feeding habitats and predat ion risk taking by juvenile fishes. Can. J. Fish. Aquat. Sci. 50, 2058 2070. Walters, C.J., Maguire, J., 1996. Lessons for stock assessment from the northern cod collapse. Rev. Fish Biol. Fisher. 6, 125-137. Walters, C.J., Kitchell, J.F., 2001. Cultivation/de pensation effects on juvenile survival and recruitment: implications for the theory of fishing. Can. J. Fish. Aquat. Sci. 58, 39-50. Walters, C. J., Martell, S.J.D., 2004. Fisheries ecology and management. Princeton University Press, New Jersey, USA. Ward, R.D., 2006. The importance of identifying spatial population structure in restocking and stock enhancement programmes. Fish. Res. 80, 9-18. Wassenbergh, S.van, Aerts, P., Herrel, A., 2006. S caling of suction feeding performance in the catfish Clarias gariepinus Phys.Biochem. Zoo. 79 (1), 43-56. Weiss, S., Schmutz, S., 1999. Performance of ha tchery-reared brown trou t and their effects on wild fish in two small Austrian streams. Trans. Am. Fish. Soc. 128, 302-316. Welty, J.C., 1934. Experiments in group behavi or of fishes. Physiol. Zool. 7, 85-128. 181

PAGE 182

Wennhage, H., 2000. Vulnerab ility of settling plaice Pleuronectes platessa to predation: effects of developmental stage a nd alternative prey. Mar. Ecol. Prog. Ser. 203, 289-299. Werckmand, C., Hardy, T., Wainwright, E., 1998. Crystal Ball Users Manual. Decisioneering, Colorado. Werner, E.E., Gilliam, J.F., Hall, D.J., Mittelba ch, G.G., 1983. An experimental test of the effects of predation risk on ha bitat use in fish. Ecology 64, 1540. Werner, E.E., Gilliam, J.F., 1984. The ontogenetic niche and species interactions in sizestructured populations. A n. Rev. Ecol. Systemat. 15, 393-425. Wespestad, V.G., Fritz, L.W., Ingraham, W.J., Megrey, B.A., 2000. On relationships between cannibalism, climate variabil ity, physical transport, and recruitment success of Bering Sea walleye pollock ( Theragra chalcogramma ). ICES J. Mar. Sci. 57 (2), 272-278. Wickler, W., 1968. Mimicry in plants and animals. McGraw-Hill, New York. Yamamura, O., Yabuki, K., Shida, O., Watana be, K., Honda, S., 2001. Spring cannibalism on 1 year walleye Pollock in the Doto area, norther n Japan: is it density-d ependent? J. Fish. Biol., 59 (3), 645-656. Yamashita, Y., Nagahora, S., Yamada, H., K itagawa, K., 1994. Effects of release size on survival and growth of Japanese Flounder Paralichthys olivaceus in coastal waters off Iwate Prefecture, northeastern Japa n. Mar. Ecol. Prog. Ser. 105, 269-276. Yoder, J.M., Marschall, E.A., Swanson, D.A., 2004. The cost of dispersal: predation as a function of movement and site familiarity in ruffed grouse. Behav. Ecol. 15 (3), 469-476. Zohar, Y., Hines, A.H., Zmora, O., Johnson, E.G ., Lipcius, R.N., Seitz, R.D., Eggleston, D.B., Place, A.R., Schott, E.J., Stubblefield, J.D., Chung, J.S., 2008. The Chesapeake Bay blue crab (Callinectes sapidus): A multidiscipl inary approach to responsible stock replenishment. Rev. Fish. Res. 16 (1-3), 24-34. 182

PAGE 183

183 BIOGRAPHICAL SKETCH Nathan Paul Brennan was born in Hartford, Connecticut, USA in 1965. At the age of 2, he moved with his family to the remote highlands of Papua New Guinea, in the Engan Province where his father worked in the Lutheran Mission. After attending elementary school at the Highlands Lutheran School, Nathan moved with his family to the capita l city, Port Moresby. There he attended Port Moresby Internationa l High School for three years. In 1981, the Brennans moved to Kailua, Hawaii, USA, locat ed on the island of Oahu where he graduated from Kailua High School in 1984. In May, 1989, he r eceived the degree of Bachelors of Arts in Zoology from the University of Hawaii at Manoa. From 1990 -1995, Nathan worked at the Oceanic Institute located at Makapuu studying fisheries and aquaculture under the Stock E nhancement Program. In August 1995, he was recruited to Cookeville, Tennessee for graduate work in a masters program at the Biology Department of Tennessee Technological Univer sity (TTU) studying Paddlefish stocks in the Mississippi drainage basin and earned a Mast ers degree in Biology in December 1997. By March 1998, he moved to Sarasota gaining employment at Mote Mari ne Laboratory. While maintaining his position at Mote, Nathan starte d his doctoral program at the University of Florida in September 2002, examining juvenile snook ecology using stock enhancement as his primary experimental tool. Nathan currently lives in Sarasota with his wife and two children.