ABSTRACT
Jacqueline A. Jenkins. FEEDING ECOLOGY OF JUVENILE FISHES FROM
THE SOUTHEAST UNITED STATES CONTINENTAL SHELF. (Under the
direction of Dr. Lisa Clough). Department of Biology, July 2005.
The main objective of this research was to describe the feeding ecology of
eight juvenile fishes from the southeast US continental shelf: Stenotomus sp.,
Centropristis ocyurus, Serranus phoebe. Diplectrum formosum, Rypticus
maculatus, Stephanolepis hispidus, Serraniculus pumilio, and Serranus baldwini.
Diet studies abound for adult fishes, but are much more limited for juvenile
stages, despite well documented ontogenetic change in diet. Diets were
quantified by determining frequency of occurrence, numerical abundance,
specific volume and the index of relative importance of prey items. Cumulative
prey curves were constructed to determine if an adequate number of fish were
analyzed, and Schoener’s index of dietary overlap was analyzed. Trophic guilds
were delineated using multivariate classification and ordination. The eight
juvenile fish species examined in this study appear to be opportunistic feeders
capable of utilizing a wide variety of prey items. Stenotomus sp. and
Stephanolepis hispidus fed predominantly on small, planktonic copepods.
Serranid and Grammistid species fed largely on amphipods, with a preference for
larger-sized prey items. Ten feeding guilds based upon habitat and prey size
were established. Prey behavior, morphological differences, and differences in
habitat utilization, are hypothesized to contribute to diet differences. Results
from this study provide dietary data on fish species that have not previously been
studied, and present the first data on trophic guild organization within the juvenile
fish community on the southeast US continental shelf. This information is
necessary for an ecosystem-based approach to fisheries management and the
conservation of marine biodiversity on the southeast United States continental
shelf.
FEEDING ECOLOGY OF JUVENILE FISHES FROM THE SOUTHEAST
UNITED STATES CONTINENTAL SHELF
A Thesis
Presented to
The Faculty of the Department of Biology
East Carolina University
In Partial Fulfillment
Of the Requirements for the Degree
Master of Science in Biology
By
Jacqueline A. Jenkins
July 2005
FEEDING ECOLOGY OF JUVENILE FISHES FROM THE SOUTHEAST
UNITED STATES CONTINENTAL SHELF
By
Jacqueline A. Jenkins
APPROVED BY:
CHAIR OF THE DEPARTMENT OF BIOLOGY /O /
Ronald J. Newton, Ph.D.
DEAN OF THE GRADUATE SCHOOL
Paul Tschetter, Ph.D.
DEDICATION
To Mom and Dad
I will forever be your Little Trout
ACKNOWLEDGEMENTS
First and foremost I want to thank my co-advisors, Dr. Lisa Clough and Dr. Jon
Hare, for their advice and support. As well, thanks to Dr. Jeff Buckel and Dr.
Terry West for their assistance with many details in the writing process. Thanks
so much to K. Marancik and H. Walsh for their help with Gray’s Reef information
and statistical assistance. Fellow office-mates G. Bath-Martin and M. Wuenschel
provided much appreciated guidance and humor during the whole process.
Thanks to E. Laban for help with the Optimus Imaging System. Thanks to Dr. G.
Kalmus for residence support, and Dr. K. O’Brien for statistical assistance. Many
friends have made the past two and a half years memorable: B. Degan, E.
Jugovich, J. Morley, N. Bacheler, S. Daley, K. Marancik, H. Walsh, M.
Wuenschel, C. Taylor, and Maple. I don’t know where I would be without you.
Finally, I want to thank my family - Edward, Ada, and Mike Jenkins - who have
instilled in me a stubborn independence that I can’t seem to shake. You have
made me strong, and I love and miss you. I’ll be home soon.
TABLE OF CONTENTS
LIST OF TABLES vii
LIST OF FIGURES viii
CHAPTER 1: INTRODUCTION 1
Fish Trophic Ecology 1
Ecosystem-Based Management 6
Gray’s Reef National Marine Sanctuary 8
Fish Species 10
Objectives 16
CHAPTER 2: MANUSCRIPT 17
Introduction 17
Materials and Methods 20
Sampling Methods 20
Analysis of Stomach Contents 20
Analysis of Food Habits 22
Diet Overlap 24
Ontogenetic Diet Variations 24
Trophic Guild Analyses 26
Results 27
Food Habits 28
Stenotomus sp 28
Centropristis ocyurus 28
Serranas phoebe 29
Diplectrum formosum 30
Rypticus maculatus 30
Stephanolepis hispidus 31
Serraniculus pumilio and Serranas baldwini 32
Schoener’s Index of Dietary Overlap 32
Ontogenetic Diet Variations 32
Stenotomus sp 33
Centropristis ocyurus 34
Serranas phoebe 36
Trophic Guild Analyses 37
Discussion 41
Food Habits 41
Dietary Overlap 48
Ontogenetic Diet Variations 49
Trophic Guild Analyses 50
CHAPTER 3: DISCUSSION 101
Food Habits 101
Dietary Overlap 112
Ontogenetic Diet Variations 114
Trophic Guild Analyses 118
Conclusions 122
REFERENCES 126
LIST OF TABLES
2-1 : Fish species, total number of specimens included in diet analyses, and
station locations where fish were collected 57
2-2: Prey category list and number of stomachs containing prey in
all fish species 58
2-3: Frequency of occurrence (%F), contribution by numbers (%N),
percent specific volume (%SV), and percent index of relative importance
(%IRI) of the different prey categories to the overall diet of Stenotomus
sp., Centropristis ocyurus, and Serranus phoebe 59
2-4: Frequency of occurrence (%F), contribution by numbers (%N), percent
specific volume (%SV), and percent index of relative importance (%IRI)
of the different prey categories to the overall diet of Diplectrum
formosum, Rypticus maculatus, and Stephanolepis hispidus 60
2-5: Observed values of Schoener’s index of diet overlap between size
species of juvenile fishes 61
2-6: Observed values of Schoener’s index of diet overlap for size classes of
Stenotomus sp., Centropristis ocyurus, and Serranus phoebe 62
LIST OF FIGURES
2-1 : A map of the study area and the individual sample sites and station
numbers used for sampling 63
2-2: Size frequency distributions of Stenotomus sp., Centropristis ocyurus,
Serranus phoebe, and Diplectrum formosum 64
2-3: Size frequency distributions of Rypticus maculatus, Stephanolepis
hispidus, Serraniculus pumilio, and Serranus baldwini 65
2-4: Randomized cumulative prey curves for Stenotomus sp., Centropristis
ocyurus, Serranus phoebe, Rypticus maculatus, Stephanolepis hispidus,
and Diplectrum formosum 66
2-5: An index of relative importance (IRl) diagram of the prey categories and
their numerical abundance (%N), specific volume (%SV), and frequency
of occurrence (%F) values for Stenotomus sp 67
2-6: An index of relative importance (IRl) diagram of the prey categories and
their numerical abundance (%N), specific volume (%SV), and frequency
of occurrence (%F) values for Centropristis ocyurus 68
2-7: An index of relative importance (IRl) diagram of the prey categories and
their numerical abundance (%N), specific volume (%SV), and frequency
of occurrence (%F) values for Serranus phoebe 69
2-8: An index of relative importance (IRl) diagram of the prey categories and
their numerical abundance (%N), specific volume (%SV), and frequency
of occurrence (%F) values for Diplectrum formosum 70
2-9: An index of relative importance (IRl) diagram of the prey categories and
their numerical abundance (%N), specific volume (%SV), and frequency
of occurrence (%F) values for Rypticus maculatus 71
2-10: An index of relative importance (IRl) diagram of the prey categories and
their numerical abundance (%N), specific volume (%SV), and frequency
of occurrence (%F) values for Stephanolepis hispidus 72
2-11 : Frequency histograms showing size (standard length, mm) of fish
analyzed for Stenotomus sp., Centropristis ocyurus, and Serranus
phoebe 73
XI
2-12: Range and median values for standards lengths of Stenotomus sp. in
each size class 74
2-13: Cumulative prey curves for four size classes of Stenotomus sp 75
2-14: Mean percentage composition by average specific volume for the major
prey categories of the diet of four size classes of Stenotomus sp 76
2-15: Range and median values for standard lengths of Centropristis
ocyurus in each size class 77
2-16: Cumulative prey curves for three size classes of Centropristis ocyurus..78
2-17: Mean percentage composition by average specific volume for the major
prey categories of the diet of three size classes of Centropristis
ocyurus 79
2-18: Range and median values for standard lengths of Serranus phoebe in
each size class 80
2-19: Cumulative prey curves for two size classes of Serranus phoebe 81
2-20: Mean percentage composition by average specific volume for the major
prey categories of the diet of two size classes of Serranus phoebe 82
2-21 : Hierarchical agglomerative cluster dendrogram of three groups and ten
feeding guilds 83
2-22: Mean proportional composition by volume of major prey categories in
guilds 1A, IB, and 1C 84
2-23: Mean proportional composition by volume of major prey categories in
guilds 2A, and 2C 85
2-24: Mean proportional composition by volume of major prey categories in
guilds 2D, 2E, and 2F 86
2-25: Non-metric multidimensional scaling plot with superimposed groups 87
2-26: Non-metric multidimensional scaling plot with superimposed guilds 88
2-27: Range and median values for standard lengths of fish within each
feeding guild 89
2-28: Range and median values for standard lengths of Stenotomus sp.,
Centropristis ocyurus, and Serranas phoebe within each guild 90
2-29: Range and median values for standard lengths of Rypticus maculatus,
Diplectrum formosum, and Stephanolepis hispidus 91
2-30: Multidimensional scaling plot of dietary data with superimposed size
class for Stenotomus sp 92
2-31 : Multidimensional scaling plot of dietary data with superimposed size
class for Centropristis ocyurus 93
2-32: Multidimensional scaling plot of dietary data with superimposed size
class for Serranas phoebe 94
2-33: Multidimensional scaling plot of dietary data with superimposed station
number for Stenotomus sp 95
2-34: Multidimensional scaling plot of dietary data with superimposed station
number for Centropristis ocyurus 96
2-35: Multidimensional scaling plot of dietary data with superimposed station
number for Serranas phoebe 97
2-36: Multidimensional scaling plot of dietary data with superimposed station
number for Diplectrum formosum 98
2-37: Multidimensional scaling plot of dietary data with superimposed station
number for Rypticus maculatus 99
2-38: Multidimensional scaling plot of dietary data with superimposed station
number for Stephanolepis hispidus 100
3-1 : A map of the study area and the individual sample sites and station
numbers used for sampling 125
CHAPTER 1
INTRODUCTION
Fish Trophic Ecology
Marine fishes occupy virtually every possible heterotrophic niche, from
detritivores such as striped mullet (Mugil cephalus) to carnivorous piscivores,
such as bluefin tuna {Thunnus thynnus) (Gerking 1994). However, there is much
flexibility in the trophic ecology of individual species, and different life stages may
occupy different trophic levels (Werner and Gilliam 1984). Fish feeding studies
can be used to infer competition between different species (Buckel 2002),
competition between individuals of the same species (Persson 1983a, b),
specialization in feeding habits (Holbrook et al. 1992), and size selectivity of prey
resources (Bethea et al. 2004). Many diet studies concentrate on adult stages
(Marancik and Hare 2005), in part due to their ease of capture and/or their
economic importance. However, size-specific shifts in food types can occur at
much smaller sizes, as fish go through a number of changes during development
from the larval to adult stage, including changes in overall body morphology,
feeding style, gape size, visual acuity, and alterations to the digestive system
(Persson et al. 2000). For instance, while growing only 120 mm in length, the
pinfish {Lagodon rhomboïdes) progresses from carnivore to herbivore in five well-
ordered stages (Stoner 1980; Stoner et al. 1984). As another example, early
post-settlement Scarus species feed on harpacticoid copepods, while larger
2
juveniles feed exclusively on algae and sand (Bellwood 1988). Additionally, fish
may have different competitors and predators at different life history stages,
which alter inter- and intra-specific interactions (Piet et al. 1999). As a result,
adult abundance and performance in size-structured populations may be limited
(Osenberg et al. 1992). By examining the trophic ecology and competitive
interactions of juvenile fishes, we may be able to explain important drivers of
adult abundance and community structure.
Three indices used in trophic analyses are the numerical index (%N),
frequency of occurrence (%F), and the gravimetric index based on volume (%V).
Percent numerical abundance is advantageous when prey items are easily
identified, but should not be used in isolation as an index of dietary importance,
since it overemphasizes the importance of small prey items taken in large
numbers (Hyslop 1980). However, %N does provide information on feeding
behavior (Macdonald and Green 1983). Frequency of occurrence provides a
qualitative picture of the food spectrum (Hyslop 1980), and represents
population-wide food habits (Cortes 1997), but does not include information on
bulk of the prey categories. Volumetric measurements provide information on
the bulk, and therefore the nutritional value of a prey category. However, to
determine the relative importance of a category by bulk, the volume must be
related to stomach capacity or fish size (Hyslop 1980).
The index of relative importance (IRI) is a combination of %N, %F, and
%V, and therefore cancels out biases of the individual components (Pinkas et al.
3
1971; Cortes 1997). As suggested by Cortes (1997), IRI expressed as a
percentage makes comparisons easier among food types. The graphical
representation of %IRI combines all three measures and allows each separate
food-importance measure to be visualized. Percent numerical abundance is
plotted on the positive y-axis, % specific volume on the negative y-axis, and %
frequency of occurrence on the x-axis (Cailliet et al. 1986).
The characterization of sample size in dietary studies is advocated to
facilitate statistical comparisons between fish species. Ferry and Cailliet (1997)
reviewed over 200 fish feeding papers, and found that none of them utilized any
technique for determining if an adequate number of samples had been collected
to precisely describe diet prior to performing comparisons among species. They
recommend cumulative prey curves, which are based on the fact that as sample
sizes increase, variation (and species richness) of prey tends to decrease, and
thus the curve reaches an asymptote as new prey types are being introduced
into the diet only rarely. Cortes (1997) also advocates the use of cumulative prey
curves to facilitate comparisons and promote consistency of methodological
approaches in feeding studies. However, to date there is no objective,
quantitative method to determine if the curve has reached an asymptote. To
quantitatively evaluate if cumulative prey curves have reached an asymptote in
this study, dietary data are modeled using an asymptotic function frequently used
in fish growth studies. In this way, a quantitative, objective analysis of the
asymptotic relationship in cumulative prey curves can be conducted.
4
Many indices of diet overlap have been developed to facilitate
comparisons among species, including Morisita’s measure (Morisita 1959),
Horn’s index (Horn 1966), Levins index (1968), and Schoener’s index (1970).
Comparative studies of diet overlap indices (Hurlbert 1978; Cailliet and Berry
1979; Wallace 1981) found that generally all resulted in similar conclusions about
the degree of overlap. Yet, Wallace (1981) suggested that Schoener’s index is
the least objectionable index available when resource-availability data are
absent. Schoener’s index of dietary overlap can provide insight into possible
competitive interactions among fish species.
Trophic analyses can also provide information on feeding groups, or
guilds, which feed on the same general class of organisms in a similar way, and
may provide further insight into possible competitive interactions (Root 1967;
Simberloff et al. 1991). One advantage of guild analyses is that fish can be
grouped solely on the basis of stomach content similarity, without regard to
taxonomy. A multivariate classification technique used in guild analyses is
clustering, which ignores any preconceived classification offish stomachs, such
as species or size of fish, and only looks at the gut contents of each individual
(Crow 1979). A common technique used to delineate dietary guilds is
agglomerative hierarchical clustering (Garrison and Link 2000; Baldó et al. 2002),
which begins with the calculation of distances of each individual fish to all other
fish, depending on the amount of each prey category found in their stomachs. All
fish begin alone in groups of size one, and groups that are close together with
5
regard to diet are merged (Manley 2005). A clustered group offish could then be
considered a feeding guild and the species and size classes offish in that guild
could then be identified (Crow 1979). The Bray-Curtis coefficient is a common
similarity measure used in ecological studies (Platell et al. 1999; Szedimayer et
al. 2000; Linke et al. 2001 ; Baldo et al. 2002). It satisfies two important criteria in
regards to dietary data. First, its value is unchanged by inclusion or exclusion of
a species that is jointly absent from two samples. Secondly, inclusion or
exclusion of a third sample, C, in the data array makes no difference to the
similarity between samples A and B (Clarke et al. 2001).
An additional multivariate technique used in dietary analyses is ordination,
which produces a small number of variables to describe the relationship between
groups of objects (Manley 2005). Non-metric multidimensional scaling constructs
a map or configuration of each individual fish, in a specified number of
dimensions from the Bray-Curtis similarity matrix. The rank ordered distances
between fish on the ordination attempt to match the corresponding ranked
similarities in diet: nearby fish consume very similar prey items, and fish which
are far apart have few prey items in common or the same prey at very different
levels of abundance (Clarke et al. 2001). The MDS algorithm is an iterative
procedure that attempts to refine the positions of the points until they satisfy, as
closely as possible, the rank similarity matrix. A non-parametric (monotonie)
regression line is fitted to a scatter plot of original distances against similarities.
The goodness of fit between the configuration distances and the similarities is
6
measured by Kruskal’s stress formula (Manley 2005). The lower the stress
value, the less scatter about the monotonie regression line. The combination of
clustering and ordination analyses can be a very effective way of evaluating the
adequacy and mutual consistency of both representations (Clarke et al. 2001).
Cluster analysis and multidimensional scaling may be useful tools in determining
resource use of juvenile fishes on the southeast United States continental shelf.
Ecosystem-based Management
Ecosystem-based management, an approach recently adopted by the
National Oceanic and Atmospheric Association and state fishery managers
(Busch et al. 2003), moves away from traditional single species management
towards incorporating multiple species into management, as well as ecological
processes that affect those species (Link 2002). To successfully implement an
ecosystem-based management approach, a number of tasks are recommended
(NOAA 1999; Busch et al. 2003) including descriptions of habitat needs of
different life-history stages to delineate geographic areas of an ecosystem, and
biological indicators such as trophic structure and species diversity to
characterize the condition of specific ecosystems.
A number of concepts have been proposed to protect habitat that is
integral to the development and survival of a species during certain life stages.
Essential fish habitat (EFH) is defined as those waters and substrate necessary
to fish for spawning, breeding, feeding, or growth to maturity (SAFMC 1998), and
7
delineation of EFH is required for all fisheries that are federally managed.
Habitat Areas of Particular Concern are subsets of EFH which are rare,
particularly susceptible to human-induced degradation, or especially ecologically
important, and generally include high value intertidal and estuarine sites, and
offshore areas of high habitat value or vertical reef (SAFMC 1998).
Essential fish habitat and HAPC are important monitoring and
conservation tools for economically valuable and non-target marine fish species,
as well as their prey. On the southeast US continental shelf, approximately 30%
of the bottom is estimated to be rocky-reef habitat (Parker et al. 1983; SEAMAP-
SA 2001). The remaining area consists of unconsolidated sediments of
moderate to low relief. Live bottoms have been characterized as areas which
contain biological assemblages consisting of such sessile invertebrates as sea
fans, sea whips, hydroids, anemones, ascidians, sponges, bryozoans, and corals
living upon and attached to naturally occurring hard or rocky formations with
rough, broken, or smooth topography (Wenner 1983; GRNMS 2005). These live-
bottom habitats provide refuge to many invertebrates (Wenner et al. 1983), such
as amphipods, and therefore a possibly important food resource to post-
settlement stage fishes (Caine 1991 ). Juvenile habitat use of many marine fish
species on the southeast United States continental shelf is lacking (Lindeman et
al. 2000). A better understanding of the habitats that serve as nurseries for
marine species will improve conservation and management of these areas (Beck
et al. 2001).
8
Gray’s Reef National Marine Sanctuary
One geographically defined Habitat Area of Particular Concern in the
South Atlantic is Gray’s Reef National Marine Sanctuary (GRNMS) (GRNMS
2005). Gray’s Reef National Marine Sanctuary is located 17.5 nautical miles off
the coast of Sapelo Island, Georgia, on the southeastern US continental shelf,
and covers nearly 17 nautical square miles of live-bottom (NRC 2001). Managed
by the Office of National Marine Sanctuaries of the National Oceanographic and
Atmospheric Administration, Gray’s Reef was designated a National Marine
Sanctuary in January of 1981. The purpose of the sanctuary was not to manage
commercially important fisheries, but to protect and manage the rocky-reef
habitat and associated biological communities found within its boundaries (FR
1981 ). Current management objectives of Gray’s Reef NMS include the
conservation of live-bottom resources and fishery habitats, and to gain better
understanding of live-bottoms and their role as an ecosystem (GRNMS 2005).
Understanding the use of these habitats within and beyond Gray’s Reef NMS
boundaries will provide data towards broader spatial scale protection of important
habitats for juvenile fishes. Clarifying the resource use of different life-history
stages, and its potential effects on species diversity and ecosystem condition, is
necessary to meet the management objectives of GRNMS, and to successfully
implement an ecosystem-based approach to fisheries management.
The National Marine Sanctuary Program, in cooperation with the National
Centers for Coastal Ccean Science (NCCCS), initiated a project in April of 2000
9
to conduct site characterization at Gray’s Reef National Marine Sanctuary. One
of the project’s objectives was to determine the importance of non-reef habitats
to juvenile stages of reef fishes and to evaluate the linkages between non-reef
and reef habitats. Previous studies have shown that juvenile reef fishes utilize
open sand habitats as settlement and nursery areas (Lindeman et al. 2000;
Dahlgren et al. 2001; Walsh et al. in review); however, the ecological importance
of non-reef habitats has not been quantified. The purpose of this study is to
analyze the feeding habits of eight juvenile fishes on the southeast United States
continental shelf, and attempt to quantify the trophic dependence of juvenile fish
on reef and non-reef food resources. The findings from this study will provide a
more complete description of the habitat required to support fish species in the
GRNMS area.
Biodiversity offish species in the GRNMS area is considered high
compared to other habitat types on the southeastern coast of the US, however,
community structure changes seasonally and interannually. Sedberry and Van
Dolah (1984) found that species richness was higher in summer than in winter,
and diversity values were higher than those reported on the northeast US
continental shelf. This high diversity was attributed to an increase in species
richness (41 species in 264 trawl collections over four seasons on the northeast
US continental shelf compared to 128 species in 83 collections over two seasons
on the southeast US continental shelf).
10
Visual surveys by divers at GRNMS identified 89 fish species (Parker et
al. 1994), while bottom trawl surveys, beam trawl surveys, fish traps, point
counts, and bongo collections added another 92 species, for a total of 181
species reported from the vicinity of GRNMS (Hare et al. in review).
Ichthyoplankton collections indicated that 51 of these species are spawning in
the vicinity of GRNMS and approximately a third of these are reef fish species
(Marancik et al. 2005). A number of reef fish species settle to non-reef habitats,
including species from the family Serranidae, such as Centropristis ocyurus and
Diplectrum formosum (Walsh et al. in review). Serranid species, as well as
species from the families Grammistidae, Sparidae, and Monacanthidae, will be
included in this study. In analyzing the prey resources in the diet, habitat choices
by juvenile fish in the GRNMS area may be ascertained.
Fish Species
The juvenile stages of eight fish species were chosen for diet analyses
due to their abundance in the vicinity of Gray’s Reef NMS (Hare et al. in review),
in addition to a lack of trophic studies on these species on the southeast United
States continental shelf. These include Stenotomus sp. (Sparidae), Centropristis
ocyurus (Serranidae), Serranus phoebe (Serranidae), Diplectrum formosum
(Serranidae), Rypticus maculatus (Grammistidae), Serraniculus pumilio
(Serranidae), Serranus baldwini (Serranidae), and Stephanolepis hispidus
(Monacanthidae),
11
Sparids live in warm-temperate and tropical waters, with approximately
120 species worldwide. Stenotomus sp. included in this study has several
synonyms: S. chrysops, S. caprinus, and S. aculeatus; however, the taxanomic
status of the different described species is uncertain (Powell and Greene 2002).
Here, Stenotomus sp. will be used. These fish are deep-bodied, usually silvery,
with a large eye and small mouth. On the east coast of the US, sparids range
from as far north as Nova Scotia, to as far south as Florida. On the northeast
United States shelf S. chrysops appears to make extensive migrations and a few
fish tagged off New England have been caught south of Cape Harteras, NO
(NMFS 1999). Movements of Stenotomus sp. on the southeast US shelf have
not been examined. In the Gulf of Mexico, young-of-the-year S. caprinus
recruited in inshore waters during the spring and moved towards deeper water as
they grew (Geoghegan and Chittenden 1982). Diet studies in the Gulf of Mexico
found that S. caprinus teá predominantly on small polychaetes, crustaceans,
echinoderms, sipunculids, and nemerteans (Henwood et al. 1978, 23-145 mm
SL; Sheridan and Trimm 1983, 25-74 mm SL). Stenotomus sp. (identified by the
authors as S. chrysops, 53-170 mm SL) from shelf and reef habitats in Onslow
Bay fed mainly on cyclopoid copepods, eucarids, polychaetes and bivalves
(Lindquist et al. 1994). Diet studies for Stenotomus < 53 mm from the southeast
US continental shelf are lacking.
Most serranids live in close association to the bottom, and are typically
sedentary. The five serranid species included in this study have similar
12
geographic distributions and diets. Centropristis ocyurus, the bank sea bass,
ranges from Cape Hattaras, NC southward to the Yucatan Banks in the Gulf of
Mexico. Movements of this species are not documented. Sedberry et al. (1998)
have shown that tagged black sea bass {Centropristis striata), a similar species,
are capable of moving as far as 167 km from Gray’s Reef to St. Augustine,
Florida. In the Gulf of Mexico, C. ocyurus are rare in waters less than 30-35
meters deep (Bullock and Smith 1991), and occupy areas of high relief (Sedberry
and Van Dolah 1984), as well as sand substrate near the edges of outcroppings
(Parker and Ross 1986). Bullock and Smith (1991) analyzed the diet of twenty-
seven specimens (11-200 mm) from the Gulf of Mexico and found that xanthid
crabs were the most frequently found (29.6% F) and the most numerous (43.6%
N) prey items eaten.
The tattler, Serranus phoebe, occurs from South Carolina to Florida, in the
Gulf of Mexico, southward to northern South America (Bullock and Smith 1991).
Off North Carolina, S. phoebe is known to prefer soft substrates surrounding
reefs, similar to C. ocyurus (Parker and Ross 1986). Bullock and Smith (1991)
concluded that adult Serranus phoebe and Centropristis ocyurus occurred
offshore, in depths >55 meters, but juvenile specimens were found in waters as
shallow as 26 meters and 6 meters, respectively. Diet studies for this species
are lacking on the southeast US continental shelf. In the Gulf of Mexico, four
specimens ranging in size from 133 to 137 mm standard length consumed crabs,
shrimps, polychaetes, amphipods and copepods (Bullock and Smith 1991).
13
Diplectrum formosum, the sand perch, ranges from North Carolina to the
Bahamas and in the northern Gulf of Mexico to Uruguay (Bortone 1971b). It is
primarily an inshore species, occurring at depths <37 meters (Bullock and Smith
1991), however has been known to occur in depths greater than 50 meters
(Bortone 1971b). Beam trawls for the monitoring project initiated by the National
Marine Sanctuary Program demonstrated that distinct cross-shelf zonation exists
for serranines, with Centropristis striata (black sea bass) settling inshore,
Centropristis philadelphica (rock sea bass) and Centropristis ocyurus settling
offshore, and Diplectrum formosum settling across the southeast US continental
shelf (Walsh et al. in review). Diplectrum formosum are presumed to be limited
by low temperatures in the Gulf of Mexico, and move offshore to deeper water
during the winter months (Bortone 1971b). Off the coast of Guyana they feed on
bottom-dwelling crustaceans and peneid shrimp (size not reported; Lowe 1962).
Bortone (1971b) analyzed 151 specimens (21-223 mm SL) from the Gulf of
Mexico and found they consumed decapods (67.0% F), amphipods (11.1% F)
and a variety offish species, including Scorpaeniforms (3.9% F), Perciforms
(3.9% F), and Pleuronectiformes (2.6% F).
The pygmy sea bass, Serraniculus pumilio, is the smallest serranid and
ranges from North Carolina to Florida, throughout the Gulf of Mexico, southward
to Venezuela (Bullock and Smith 1991). blastings (1973) reported that inshore
populations of S. pumilio move offshore to deeper water during winter in the Gulf
of Mexico, and advanced juveniles and adults do not have a particularly
14
restricted home range. Two individuals (50 mm and 31 mm SL) ate crabs and
shrimp, respectively, in the Gulf of Mexico (Bullock and Smith 1991). In
Choctawhatchee Bay, Florida, twenty-nine specimens (15.8 to 54.1 mm SL)
frequently ate a large number of amphipods (38.2 % N, 58.6% F), however
shrimp and crabs comprised a large volume of the diet and may be the most
important food items when present (Hastings 1973).
Serranas baldwini, the lantern sea bass, has a warmer affiliation,
compared to the other species, ranging from south Florida to Brazil (Humann and
Deloach 2002). The depth distribution of this species ranges from waters edge
to about 240 feet, and it occurs on or around small piles of shell and coral debris
(Robins and Starck 1961). It is a benthic species, rarely straying far from the
cover afforded by rock and shell fragments. Three individuals collected off of the
US Virgin Islands (30-41 mm SL) contained two-third by volume caridean shrimp
and one-third Labridae fishes (Robins and Starck 1961). There are currently no
diet studies of this species on the southeast US continental shelf.
Grammistids have been described as a separate family from Serranidae
(Grammistidae) due to the production of grammistin, a skin toxin thought to deter
predation (Randall et al. 1971). They are known to occur from Cape Hatteras,
NC, to the western Gulf of Mexico (Guimaráes 1999). Like Diplectrum
formosum, Rypticus maculatus is primarily an inshore species, occurring at
depths <37 meters (Bullock and Smith 1991). Courtenay (1967) stated that R.
maculatus shows a preference for cooler and deeper waters and a sandy
15
substrate rather than calcareous marl or mud in depths ranging from 15 to 50
fathoms. Six R. maculatus ranging in size from 34 to 192 mm consumed shrimp,
crabs, and one bronze cardinalfish {Astrapogon alutus) in the eastern Gulf of
Mexico. Currently there are no diet studies of this species from the southeast US
continental shelf.
Monacanthids are small to medium-sized fishes that live in warm-
temperate and tropical waters. The planehead filefish, Stephanolepis hispidus,
ranges as far north as Nova Scotia, is common along the southeast United
States shelf, and is found as far south as Brazil (Bigelow and Schroeder 1953).
Early life-history stages are pelagic in offshore waters (Berry and Vogele 1959),
while late juveniles and adults typically associate with benthic structure (Lindquist
et al. 1989). Stephanolepis hispidus ranging in size from 6.5 to 81.5 mm SL
comprised 78.5% of all species associated with pelagic Sargassum spp. habitats
on the continental shelf of the southeastern US (Settle 1993). In the northeast
continental shelf demersal ecosystem, S. hispidus (5-19 cm) consumed
polychaetes predominantly, followed by cephalopoda and echinoderms (Bowman
et al. 2000). Currently, there are no diet studies of this species on the southeast
United States continental shelf demersal ecosystem.
Ross (1986) concluded that when marine fishes co-occur in the same
general area, food and then habitat are the resources that are most commonly
partitioned among those species. Also, the potential for competition is high for
species that are morphologically similar and occupy similar habitats (Brodeur et
16
al. 1984). Due to morphological similarities, as well as overlaps in spatial
distributions during the juvenile stage, the eight species examined here may
compete for resources.
Objectives
The first objective of this study is to describe the diet and dietary overlap
of 8 juvenile fish species. Second, ontogenetic diet comparisons between
Stenotomus sp., Centropristis ocyurus and Serranus phoebe will be conducted to
determine if there are shifts in food preference in these species with age. Finally,
classification and ordination analyses will be used to determine the possible
occurrence of feeding guilds in 8 juvenile fishes on the southeast United States
continental shelf, including the GRNMS area. The results of this study will
provide basic knowledge of juvenile feeding relationships that is lacking for these
species on the southeast US continental shelf (Marancik and Hare 2005) and
necessary for an ecosystem-based approach to fisheries management and the
conservation of marine biodiversity
CHAPTER 2
MANUSCRIPT
introduction
Understanding resource use of early life-stages is essential in clarifying
inter- and intra-specific interactions that ultimately may explain important drivers
of adult abundance and community structure (Piet et al. 1999). Fish feeding
studies can be used to infer competition between different species (Buckel 2002),
competition between individuals of the same species (Persson 1983a,b),
specialization in feeding habits (Holbrook 1992), and size selectivity of prey
resources (Bethea 2004). With the adoption of an ecosystem-based
management approach (Busch et al. 2003), emphasis is placed upon
incorporating multiple species, size-structured changes in resource use, as well
as ecological processes to predict changes in population dynamics and
abundances (Osenberg 1992).
Marine fishes occupy virtually every possible heterotrophic niche, from
detritivores such as striped mullet (Mugil cephalus) to carnivorous piscivores,
such as bluefin tuna {Thunnus thynnus){Gerk\ng 1994). Additionally, there is
much flexibility in the trophic ecology of individual species, and different life
stages often occupy different trophic levels (Werner and Gilliam 1984). Many
diet studies concentrate on adult stages (Marancik and Hare 2005), in part due to
their ease of capture and/or their economic importance. However, size-specific
18
shifts in food types can occur at much smaller sizes, as fish go through a number
of changes during development from the larval to adult stage, including changes
in overall body morphology, feeding style, gape size, visual acuity, and
alterations to the digestive system (Persson et al. 2000). For instance, while
growing only 120 mm in length, the pinfish {Lagodon rhomboïdes) progresses
from carnivore to herbivore in five well-ordered stages (Stoner 1980; Stoner et al.
1984). As another example, early post-settlement Scarus species feed on
harpacticoid copepods, while larger juveniles feed exclusively on algae and sand
(Bellwood 1988). Additionally, fish may have different competitors and predators
at different life history stages, which alter inter- and intra-specific interactions
(Piet et al. 1999). As a result, adult abundance and performance in size-
structured populations may be limited (Osenberg et al. 1992). By examining the
trophic ecology and competitive interactions of juvenile fishes, we may be able to
explain important drivers of adult abundance and community structure.
Another important goal of ecosystem-based approaches to fishery
management is the protection of habitat integral to the development and survival
of a species during certain life stages. This is currently being implemented
through such concepts as Essential Fish Habitat (EFH), Habitat Areas of
Particular Concern (HAPC), and marine protected areas (MPA’s) (SAFMC 1998;
Busch et al. 2003; GRNMS 2005). Essential fish habitat (EFH) is defined as
those waters and substrate necessary to fish for spawning, breeding, feeding, or
growth to maturity (SAFMC 1998). Habitat Areas of Particular Concern are
19
subsets of EFH which are rare, particularly susceptible to human-induced
degradation, or especially ecologically important, and generally include high
value intertidal and estuarine sites, and offshore areas of high habitat value or
vertical reef (SAFMC 1998).
A geographically defined Habitat Area of Particular Concern on the
southeast US continental shelf is Gray’s Reef National Marine Sanctuary (NMS)
(GRNMS 2005). The goals of Gray’s Reef NMS include conserving biodiversity
and habitat, protecting critical life history stages, and monitoring and maintaining
ecological interactions (NRC 2001). Currently, information on predator-prey
interactions of juvenile fishes on the southeast US continental shelf, including
Gray’s Reef NMS, is lacking for a number of species, including species from the
families Serranidae, Grammistidae, Sparidae and Monacanthidae.
The purposes of this study are to analyze the feeding habits of eight
juvenile fishes (Table 2-1), describe shifts in food preferences through ontogeny
in three species, and delineate groups of fishes that feed on similar organisms
(‘feeding guilds’) through the use of multivariate classification and ordination.
The results of this study will provide basic knowledge of juvenile feeding
relationships that is lacking for these species on the southeast US continental
shelf (Marancik and Hare 2005) and necessary for an ecosystem-based
approach to fisheries management, the management of habitat areas of
particular concern, and the conservation of marine biodiversity.
20
Materials and Methods
Sampling Methods
Fish and invertebrate samples were collected during April 2002 from 32
sites along three cross-shelf transects off the coast of Georgia (Figure 2-1 ). At
each site, three replicate 5-minute tows were made on the bottom with a 2 m
beam trawl with 3 mm mesh cod end. Tows were made only at night. Samples
were preserved in 5% formalin-seawater and transferred to 95% ethanol after 48
hours. Fish and invertebrates were separated, and the fish measured, weighed
and identified to species.
Analysis of Stomach Contents
Juveniles of eight fish species were included in the diet analysis (Table 2-
1). All the individuals of the eight species were included in diet analyses, unless
the number offish captured at a given station was large; in this instance 30
individuals were sub-sampled for analyses. Individual fish were blotted dry,
standard length (SL) and total length (TL) measured to the nearest 1 mm, and
body weight measured to the nearest 0.01 g. To determine if diet changed with
ontogeny, three of the eight fish species were separated into size classes and
diet composition analyzed on each size class (see below). One individual of
21
Serranas baldwini and two specimens of Serraniculus pumilio were collected and
were included simply for descriptive purposes.
The esophagus, cut behind the cardiac sphincter, and stomach, ending at
the pyloric sphincter, were removed and weighed to the nearest 0.0001 g.
Contents were removed and preserved in 95% ethanol and the empty esophagus
and stomach were re-weighed. Contents were examined microscopically, and
identified to the lowest practical category. Due to the amount of digestion, it was
impossible to identify prey items to species. Prey categories that were counted
but not measured volumetrically include coral, copepod egg sacs, fish scales,
nematodes, sand grains, shell fragments and unknown (Table 2-2). These items
were not included in dietary comparisons.
Cumulative prey curves were constructed to determine if an adequate
number offish had been collected to characterize diet (sensu Ferry et al. 1996,
Gelsleichter et al. 1999). For each species, the order in which the guts which
contained food were analyzed was randomized ten times, and the mean number
of prey categories found consecutively in the guts plotted against the number of
stomachs examined. To quantitatively evaluate if cumulative prey curves have
reached an asymptote, dietary data are modeled using an asymptotic function fit
with non-linear regression techniques using the Statistical Package for the Social
Sciences (SPSS) 12.0. The asymptotic function is:
Pn = Pco(1-e''^"’)
22
where Pn is the number of prey categories found in a sample size of n fish, P«, is
the asymptotic number of prey categories, and k is the rate at which new prey
categories are added as more fish are added to the sample. Nonlinear
regressions using the mean number of new prey categories from 10
randomizations with increasing sample size provide estimates of and k. If the
actual number of prey categories was within the confidence interval of the
estimated P*, we concluded that a satisfactory sample size was analyzed.
Analysis of Food Habits
Diet composition was analyzed using three indices described by Hyslop
(1980): the numerical index (%N), frequency of occurrence (%F), and the
gravimetric index based on specific volume (%SV):
%Nspfif,ifis = number of individuals of prey category X 100
total number of prey individuals from all categories
%SVspecies = specific volume of prey category X 100
total specific volume of all prey categories
%Fspecies = number of guts containing prey category X 100
number of guts containing food
23
Volume was measured using an adaptation of the method described by
Hellewell and Abel (1971). Each prey category from each specimen was placed
on a microscope slide and depressed to a known depth. The surface area of the
depressed prey was measured using an Optimus 8.0 Imaging System. Volume
was calculated by multiplying the surface area by the known depth (Szedimayer
et al. 2004; Hyslop 1980; Hellewell and Abel 1971). Specific volumes for prey
categories were calculated by dividing prey volume by individual fish weight
(mm^/fish wt g). The index of relative importance (IRI) was calculated, since it
may provide a more accurate description of dietary importance and seems to
cancel out biases in the individual components (Cortes 1997):
IRI = (% numerical abundance + % specific volume) X % frequency of
occurrence
To determine % IRI, all prey category IRI’s for one fish species were summed,
and the individual prey categories were calculated as a percentage of the sum.
Total percent specific volume was used in the determination of both IRI and
%IRI. The most important prey categories based on %IRI values comprising
95% of the total were chosen for graphical display. Categories comprising less
than 5 %IRI combined were not included in the %IRI graph.
24
Diet Overlap
To help explain community structure and to clarify possible competitive
interactions, dietary overlap was measured between each species and between
size classes within each species. Schoener’s (1970) index of niche overlap is
the most commonly used diet overlap index (Wallace 1981), and calculated as:
0!= 1 -0.5 {L] pij-pik\)
where a = the degree of overlap, p,j = the proportion of the / th resource (in this
case, prey category) used by species j, and Pik = the proportion of the / th prey
category used by species k. Index values range from 0 to 1 ; they approach 0 for
species that share no prey categories and approach 1 for species pairs that have
nearly identical prey utilizations. Values exceeding 0.6 are assumed to represent
‘biologically significant’ overlap in resource use (Wallace 1981; Bacheler et al.
2004).
Ontogenetic Diet Variations
Ontogenetic diet comparisons were made between size classes within
Stenotomus sp., Centropristis ocyurus, and Serranas phoebe. Since volumetric
data best represent the relative importance of any particular dietary category.
25
ontogenetic comparisons were described using the average percentage by
specific volume instead of percentage of specific volume, as the latter places
increased emphasis on fish that consume one or two prey items with large
volumes (Wallace 1981; Platell et al. 1999). The average of specific volume
percentage (ASV) was calculated as the sum of proportion by specific volume of
an individual prey category in each stomach divided by the number of stomachs
examined (Platell et al. 1999; Hyslop 1980). Cumulative prey curves were also
constructed to determine if enough guts were analyzed to describe the feeding
habits of each size group. In each case, the order in which prey were analyzed
was randomized five times and a mean number of new prey categories plotted
against the number of guts analyzed. Presence of an asymptote was assessed
using the method described above.
Size categories were chosen on the basis of optimizing a sample size of at
least 14 individuals for all three species. Sixty-seven specimens of Stenotomus
sp. were grouped into four size classes: 0-30 mm SL, 31 - 34 mm SL, 35 - 40
mm SL, and > 41 mm SL. Forty-seven specimens of C. ocyurus were grouped
into 3 size classes: 0-25 mm SL, 26 - 34 mm SL, and > 35 mm SL. Finally, 33
specimens of S. phoebe were grouped into two size classes: < 20 mm SL and >
20 mm SL.
26
Trophic Guild Analyses
Classification and ordination were performed to determine the occurrence of
feeding guilds within the juvenile fishes analyzed. Average percent specific
volumes of prey categories that occurred in at least 10 stomachs (Table 2-2)
were double root transformed and analyzed using the PRIMER 5.0 package
(Clarke et al. 2001 ). The ‘other crustacean’ prey category and S. pumilio and S.
baldwini \Nere not included in classification or ordination. Similarities of average
percent specific volume of prey categories between each pair of individual fish
were calculated using the Bray-Curtis coefficient, and agglomerative hierarchical
clustering using group-average linkage based on Bray-Curtis similarities was
used to define trophic groups. Clusters with a similarity of ^0 % were arbitrarily
chosen to represent feeding guilds (Luczkovich et al. 2002).
Non-metric multidimensional scaling (MDS) was used in conjunction with
cluster analysis to allow any relationship between groupings to be more
informatively displayed (Clarke et al. 2001 ). Bray-Curtis similarities were used to
construct ordination plots with the PRIMER 5.0 package. Differences in size of
each species among guilds were analyzed with a Kruskal-Wallis test (P< 0.05) to
further examine ontogenetic patterns in feeding guild membership. Additionally,
ontogenetic size classes for Stenotomus sp., C. ocyurus, and S. phoebe were
superimposed on the ordination plot to compare overlap with those values
determined using Schoener’s index. Finally, separate ordination plots were
27
constructed to visually inspect spatial patterns of feeding habits within each
species.
Results
The gut contents of 323 juvenile fish were analyzed, of which 135 (41.8%)
were empty (Figures 2-2 & 2-3). Diplectrum formosum had the highest average
standard length (59.75 mm ± 20.88 mm SE) and weight (20.55 g ± 16.58 g), but
this was due to two fish that were greater than 100 mm (103 mm and 185 mm),
and weighed 22.46 g and 135.07 g, respectively. Due to low a sample size for
this species, these fish were included in the diet analyses. Rypticus maculatus
had the next largest standard length (46.05 mm ± 1.71 mm SE). Stephanolepis
hispidus had the next largest average weight (2.11 g ± 1.87 g SE). The species
with the lowest average standard length (20.58 mm ± 1.53 mm SE) and weight
(0.26 g ± 0.06 g) was Serranus phoebe.
Thirty prey categories were defined, seven of which were considered other
categories. Other prey categories, which were counted but not measured
volumetrically, were coral, copepod egg sacs, fish scales, nematodes, sand
grains, shell fragments and unknown (Table 2-2). Nematodes were most
frequent (83.7% F) and numerous (82.0% N) in Stenotomus sp., but incidence of
parasitic nematodes are common in S. caprinus, and should not be considered
food items (Henwood 1978).
28
Food Habits
Stenotomus sp. A total of 136 juvenile Stenotomus sp. were analyzed,
of which 67 (49.3%) had ‘food’ categories present in the gut and were included in
the diet analyses. The cumulative prey curve estimated with the asymptotic
function established a Poo of 18.1 prey categories, with a confidence interval of
between 17.8 and 18.4. From these estimates, the maximum number offish
needed to accurately describe the diet is between 44 and 73 individuals. The
number of fish at which the curve begins an asymptotic relationship is 55
individuals (Figure 2-4). Therefore, a sufficient number of fish were analyzed to
characterize their diet. Eighteen prey categories were consumed (Table 2-3).
Small calanoid copepods (prosome length < 2 mm) were the most numerically
abundant (63.0% N), most frequently found (76.1% F), and had the highest
specific volume (65.6% SV). The percent index of relative importance for small
calanoid copepods was 73.9%. Cyclopoid copepods (17.9% IRI), large calanoid
copepods (prosome length > 2 mm) (3.19% IRI), and cladocearns (3.19% IRI)
were also important in the diet (Figure 2-5).
Centropristis ocyurus. Seventy C. ocyurus were analyzed, of which
47 (67%) had ‘food’ categories present in the gut and were included in the diet
analyses. The cumulative prey curve estimated with the asymptotic function
established a Pœ of 19.8 prey categories, with a confidence interval of between
29
19.3 and 20.3. From these estimates, the maximum number of fish needed to
accurately describe the diet is between 38 and 63 individuals. The number of
fish at which the curve begins an asymptotic relationship is 49 individuals (Figure
2-4). Forty-seven individuals were included in this study. Nineteen prey
categories were eaten (Table 2-3). Caprellid amphipods were the most
numerically abundant (30.5% N) and most frequently found prey category (57.4%
F). The percent index of relative importance for this prey category was 40.9%.
Gammarid amphipods were the next most important prey category (22.3% IRl),
followed by other crustaceans (11.1% IRl), polychaetes (7.8% IRl) and large
calanoid copepods (6.1% IRl) (Figure 2-6).
Serranas phoebe. A total of forty-three S. phoebe were analyzed, of
which 33 (77%) had ‘food’ categories present in the gut and were included in the
diet analyses. The cumulative prey curve estimated with the asymptotic function
established a Poo of 14.4 prey categories, with a confidence interval of between
13.3 and 15.6. From these estimates, the maximum number of fish needed to
accurately describe the diet is between 19 and 65 individuals. The number of
fish at which the curve begins an asymptotic relationship is 33 individuals (Figure
2-4). Therefore, a sufficient number offish were analyzed to characterize their
diet. A total of 13 prey categories were found (Table 2-3), of which other
crustaceans were the most important (43.6% IRl), followed by large calanoid
copepods (27.5% IRl), and mysids (12.3% IRl). Other crustaceans were the
30
most frequently found prey category (63.6% F) and had the highest specific
volume (38.2% SV), however, large calanoid copepods were the most
numerically abundant prey category (22.8% N) (Figure 2-7).
Diplectrum formosum. A total of eight individuals out of 23 analyzed
had ‘food’ categories present in the gut and were included in the diet analyses.
The cumulative prey curve estimated with the asymptotic function established a
Poo of 7.3 prey categories, with a confidence interval of between 5.3 and 9.2.
From these estimates, the estimated number of prey categories never reaches 6
prey categories; therefore the maximum number offish needed to accurately
describe the diet cannot be determined from the 8 individuals examined (Figure
2-4). As a result, an insufficient number offish were analyzed to characterize
their diet. A total six prey categories were consumed (Table 2-4), of which large
calanoid copepods dominated in number (50% N) and frequency (50% F), and
shrimp dominated in specific volume (58.2% SV). Large calanoid copepods
(48.5% IRI) and shrimp (28.1% IRI) were the top two most important prey
categories, followed by fish (11.7% IRI), other crustaceans (8.1% IRI), caprellid
amphipods (2.4% IRI) and crabs (1.2% IRI) (Figure 2-8).
Rypticus maculatus. Nineteen of 24 R. maculatus had food
categories present in the gut and were included in diet analyses. The cumulative
prey curve estimated with the asymptotic function established a Pa, of 5 prey
31
categories, with a confidence interval of between 4.9 and 5.2. From these
estimates, the maximum number offish needed to accurately describe the diet is
between 6 and 14 individuals. The number of fish at which the curve begins an
asymptotic relationship is 9 individuals (Figure 2-4). Therefore, a sufficient
number of fish were analyzed to characterize their diet. Gammarid amphipods
were the most frequently found prey category (84.2% F) and the most
numerically abundant (78.6% N), while crabs had the highest specific volume
(56.3% SV) of the five prey categories consumed by the species (Table 2-4).
Gammarid amphipods also had the highest index of relative importance of
87.5%, while cyclopoid copepods had the lowest (0.8% IRI) (Figure 2-9).
Stephanolepis hispidus. Eight prey categories were consumed by 12
out of 23 fish analyzed (Table 2-4). The cumulative prey curve estimated with
the asymptotic function established a Poo of 8.1 prey categories, with a confidence
interval of between 7.7 and 8.5. From these estimates, the maximum number of
fish needed to accurately describe the diet is between 5 and 16 individuals. The
number of fish at which the curve begins an asymptotic relationship is 8
individuals (Figure 2-4). Therefore, a sufficient number of fish were analyzed to
characterize their diet. Cyclopoid copepods had the highest index of relative
importance (69.6% IRI) due to the highest numerical abundance (71.8% N), the
highest frequency of occurrence (66.7% F), and the highest specific volume
32
(35.6% SV). Polychaetes and eggs had the lowest IRI’s at 0.2% and 0.1%,
respectively (Figure 2-10).
Serraniculus pumilio and Serranus baldwini. Two individuals of
Serraniculus pumilio and one individual of Serranus baldwini \Nere captured
(Figure 2-3). Of the two S. pumilio individuals, one was empty, while the other
consumed a crab. Serranus baldwini ate one small calanoid in addition to other
unidentifiable crustaceans. Due to the low sample sizes, these species were not
included in diet comparisons with the other 6 juvenile fish species.
Schoener’s Index ofDietary Overlap
Based on the Schoener’s index, there was no ‘biologically significant’
overlap between any of the fish species (Table 2-5). The highest overlap was
found between S. phoebe and S. hispidas (0.56), followed by S. phoebe and D.
formosum (0.48), and S. phoebe and C. ocyurus (0.42). The lowest dietary
overlap was between R. maculatus and D. formosum (0.14).
Ontogenetic Diet Variations
Three species were analyzed for ontogenetic diet variations: Stenotomus
sp., Centropristis ocyurus, and Serranus phoebe (Figure 2-11).
33
Stenotomus sp. Four sub-groups (0 - 30 mm SL, 31 - 34 mm SL, 35
- 40 mm SL, and > 41 mm SL) were selected for analysis as separate size
classes (Figure 2-12). The cumulative prey curve estimated with the asymptotic
function established a Poo of 14.7 prey categories for the 0-30 mm SL size class,
with a confidence interval of between 13.8 and 15.6. From these estimates, the
maximum number offish needed to accurately describe the diet is between 5 and
19 individuals. The number of fish at which the curve begins an asymptotic
relationship is 23 individuals (Figure 2-13). For the 31 - 34 mm SL size class,
the Poo confidence interval was between 11.6 and 12.9, with an asymptote
beginning at 13 individuals. From 35-40 mm SL, the asymptote begins at 16
individuals, with a confidence interval between 13.1 and 17.7 prey categories. In
the largest size class, the cumulative prey curve estimated with the asymptotic
function established a Poo of 15.4 prey categories, with a confidence interval of
between 13.3 and 17.5. However, the estimated number of prey categories
never reaches 14 prey categories, therefore the maximum number offish needed
to accurately describe the diet cannot be determined from the 14 individuals
examined in the > 41 mm SL size class. From these findings, a sufficient number
offish were analyzed in the two middle size classes, but not in the smallest and
largest size classes.
The dominant prey category by average specific volume (ASV) in fish of
standard length 0-30 mm was small calanoid copepods (37.3% ASV), followed
by cyclopoid copepods (17.6% ASV), other crustaceans (16.4% ASV), eggs
34
(7.4% ASV), and caprellid amphipods (6.3% ASV) (Figure 2-14). With growth,
fish of 31 - 34 mm standard length shifted to include a higher proportion of
cladocearns (9.5% ASV), large calanoid copepods (8.5% ASV), and gammarid
amphipods (6.3% ASV) in their diet. Compared to the smaller size class,
Stenotomus sp. 31 - 34 mm standard length did not consume polychaetes and
clams (Figure 2-14). From 35-40 mm standard length, small calanoid
copepods are still the most volumetrically abundant prey category (48.3% ASV),
followed by harpacticoid copepods (11.6% ASV), large calanoid copepods
(10.0% ASV), and cyclopoid copepods (8.7% ASV) (Figure 2-14). In the largest
size class (>41 mm SL), cyclopoid and small calanoid copepods have similar
average specific volumes (24.8% ASV and 23.2% ASV, respectively), followed
by gammarid amphipods (15.2% ASV), larvaceans (12.3% ASV), and other
crustaceans (9.9% ASV)(Figure 2-14).
Schoener’s index of dietary overlap reveals biologically significant overlap
between fish of length 0-30 mm SL with the three larger size classes, and
between the 31 - 34 mm SL and 35 - 40 mm SL size classes (Table 2-6). Size
classes 31 - 34 mm SL and >41 mm SL, and 35 - 40 mm SL and >41 mm SL
did not have biologically significant dietary overlap.
Centropristis ocyurus. Three sub-groups (0 - 25 mm SL, 26 - 30 mm
SL, and > 31 mm SL) of C. ocyurus were analyzed as separate size classes
(Figure 2-15). The cumulative prey curve estimated with the asymptotic function
35
established a PooOf 12.9 prey categories for the 0-25 mm SL size class, with a
confidence interval of between 11.7 and 14.1. From these estimates, the
maximum number offish needed to accurately describe the diet is between 7 and
34 individuals. The number of fish at which the curve begins an asymptotic
relationship is 13 individuals (Figure 2-16). For the 26 - 30 mm SL size class,
the Pco confidence interval was between 14.9 and 17.2, with an asymptote
beginning at 18 individuals. For fish > 31 mm SL, the asymptote begins at 13
individuals, with a confidence interval between 12.4 and 15.7 prey categories.
However, the estimated number of prey categories never reaches 13 prey
categories, therefore the maximum number of fish needed to accurately describe
the diet cannot be determined from the 14 individuals examined in the > 31 mm
SL size class. From these findings, a sufficient number offish were analyzed in
the two smaller size classes, but not in the largest size class.
By average specific volume, gammarid amphipods (23.3% ASV) and
cyclopoid copepods (20.4% ASV) dominated the diet offish in the 0-25 mm
size class, followed by caprellid amphipods (15.7% ASV), polychaetes (14.6%
ASV), and large calanoid copepods (7.3% ASV). Small C. ocyurus also
consumed other crustaceans (7.1% ASV), and mysids (4.5% ASV) (Figure 2-17).
With growth, caprellid amphipods (20.4% ASV) and polychaetes (16.8% ASV)
become most volumetrically important and large calanoid (4.4% ASV) and
cyclopoid copepods (1.7% ASV) become less volumetrically important in the
middle size class (Figure 2-17). Large C. ocyurus consumed caprellid
36
amphipods (32.4% ASV), ostracods (12.9% ASV), other crustaceans (8.6%
ASV), and crabs (8.6% ASV) out of 13 prey categories eaten (Figure 2-17).
Dietary overlap between the small and middle size classes, and between
the middle and large size classes was biologically significant, although overlap
between the small and large size classes was not biologically significant (Table
2-6).
Serranus phoebe. The sample of Serranus phoebe had the smallest
average standard length (20.58 mm ± 1.53 mm SE) and was separated into two
size classes for ontogenetic diet comparisons: < 20 mm SL and > 20 mm SL
(Figure 2-18). The cumulative prey curve estimated with the asymptotic function
established a Poo of 7.7 prey categories for the < 20 mm SL size class, with a
confidence interval of between 7.5 and 8.0. From these estimates, the maximum
number of fish needed to accurately describe the diet is between 8 and 31
individuals. The number offish at which the curve begins an asymptotic
relationship is 12 individuals (Figure 2-19). For the > 20 mm SL size class, the
Poo confidence interval was between 11.6 and 15.2, with an asymptote beginning
at 14 individuals. From these findings, a sufficient number offish were analyzed
in both size classes.
The number of prey categories eaten by S. phoebe increases from 8 in the
small size class to 12 in the large size class (Figure 2-20). Other crustaceans
have the highest average specific volume in both size classes (41.7% ASV and
37
30.0% ASV, in small and large size classes, respectively). In addition to other
crustaceans, smaller fish consume large calanoid copepods (24.6% ASV),
caprellid amphipods (10.7% ASV), mysids (8.7% ASV), and gammarid
amphipods (7.2% ASV) (Figure 2-20). For those fish greater and equal to 20 mm
standard length, other crustaceans are followed in volumetric importance by
mysids (28.9% ASV), large calanoid copepods (15.4% ASV), gammarid
amphipods (9.8 % ASV), fish (6.3% ASV), and caprellid amphipods (3.8 % ASV)
(Figure 2-20). Size classes exhibit biologically significant overlap (Table 2-6).
Trophic Guild Analyses
Agglomerative hierarchical clustering of 170 individual fish established 10
‘feeding guilds’ at the 40% similarity level (Figure 2-21 ). These guilds were
broadly categorized into three groups (1,2, and 3), emphasizing similarities in
habitat and size of prey consumed. Within group 1, three guilds (1A, IB, and 1C)
reflect utilization of large, benthic prey types. Guild 1A included S. phoebe, R.
maculatus, and C. ocyurus. Mysids accounted for 55.4% of the diets of these
individuals. Gammarid amphipods and large calanoid copepods accounted for
16.5% and 3.5% of their diets, respectively (Figure 2-22). In order of abundance,
R. maculatus, S. phoebe, Stenotomus sp., and C. ocyurus were included in guild
IB. These fish consumed primarily gammarid amphipods with an average
proportion of 79.8% (Figure 2-22). Centropristis ocyurus, as well as one
individual Stenotomus sp. and S. phoebe, ate polychaetes in guild 1C. These
38
individuals also consumed gammarid amphipods (17.5%), large calanoid
copepods (5.7%), and caprellid amphipods (4.0%). Cyclopoid copepods,
ostracods, eggs and harpacticoid copepods were eaten in very low proportions
(Figure 2-22).
Within group 2, there were 6 guilds of planktivores and epibenthic
predators that utilized smaller sized prey compared to group 1 (Figure 2-23).
Caprellid amphipods accounted for 47.8% of the diets of the individuals in guild
2A, however, they also ate large calanoid copepods (9.9%), gammarid
amphipods (9.1%), cyclopoid copepods (4.3%), mysids (3.7%), and ostracods
(2.6%). Cladocearns, polychaetes, harpacticoid copepods, and small calanoid
copepods were eaten in very low proportions (Figure 2-23). Guild 2B, containing
one S. phoebe, was separated from 2A due to the small volume of caprellid
amphipods found in its stomach (1.2%)(Figure 2-23). Large calanoid copepods
dominated the diet offish in guild 2C (Figure 2-23), with an average proportion of
54.1%, while guilds 2A and 2B consumed primarily caprellid amphipods. Guilds
2A and 2B contained 5 of the 6 fish species analyzed. Rypticus maculatus were
not included.
Stenotomus sp. dominated guild 2D, due to the large proportion of small
calanoid copepods in their diet (53.9%). One individual of C. ocyurus and S.
phoebe also ate small calanoid copepods. Fish in guild 2D also ate cyclopoid
copepods (17.9%), large calanoid copepods (7.9%), and cladocearns (6.0%).
Eggs, harpacticoid copepods, gammarid and caprellid amphipods, mysids, and
39
ostracods were also eaten in very small proportions (Figure 2-24). Guild 2E ate
cyclopoid copepods (64.1%) and eggs (6.8%), and did not include individuals of
D. formosum and R. maculatus (Figure 2-24). Two Stenotomus sp. comprised
guild 2F. Harpacticoid copepods were the only prey items these fish consumed
and accounted for 98.8% of their diets (Figure 2-24). Group 3 was comprised of
an individual C. ocyurus that consumed one small ostracod and no other prey.
The results of non-metric multidimensional scaling parallel those of
clustering. Groups 1 and 2 formed groups in the upper section of the plot, with
moderate distinction from group 3 in the bottom section of the plot (Figure 2-25).
When the 10 feeding guilds from the cluster analysis are superimposed on the
ordination plot (Figure 2-26), there is a moderate distinction between guilds, but
there is also a large amount of overlap. A stress value of 0.14 gives a useful 2-
dimensional picture, but too much reliance should not be placed on the detail of
the plot (Clarke 2001). This moderately high stress value indicates some
difficulty in displaying relationships between the 170 fish in two dimensions.
Since changes in diet with growth are of interest, differences in standard
length among guilds were analyzed. Overall, there is a significant difference in
standard length between guilds (Kruskal-Wallis, P < 0.05) (Figure 2-27).
Stenotomus sp., C. ocyurus, S. phoebe, D. formosum, and R. maculatus did not
differ significantly in size among guilds (Figures 2-28 & 2-29). Stephanolepis
hispidus, however, which occurred in guilds 2A, 2C, and 2E, were significantly
smaller in guild 2E (Figure 2-29). This guild consumed primarily cyclopoid
40
copepods, while guilds 2A and 2C ate caprellid amphipods and large calanoid
copepods, respectively.
Size classes of Stenotomus sp., C. ocyurus, and S. phoebe superimposed
on the ordination plot show considerable overlap between size classes in all
three species (Figures 2-30, 2-31, & 2-32), however, stress values of >0.10 give
useful 2-dimensional pictures, but too much reliance should not be placed on the
detail of the plots. These plots give credence to the overlap values computed
with Schoener’s index, however, comparisons are uncertain since different
values were used to compute overlap using Schoener’s index versus non-metric
multidimensional scaling.
Stations where fish were captured were superimposed on individual
species ordination plots to determine the spatial orientation of each individual fish
(Figures 2-33 to 2-38). Stenotomus sp. were collected from stations 2.1 and 35
(Figure 2-1 ), and therefore these stations are shown as two groups on the
ordination plot (Figure 2-33). When compared to the guild ordination, the
individuals captured from these two stations ate primarily small calanoid
copepods (Figure 2-26). Similar relationships between guild membership and
station location were not found in the other 6 species (Figures 2-34 to 2-38).
41
Discussion
This study represents the first attempt at describing the diets and diet
overlap of several juvenile fishes on the southeast US continental shelf. For
juvenile Serranus phoebe, Stephanolepis hispidas, and Rypticus macuiatus this
represents the first quantitative information on foraging ecology on the southeast
US continental shelf. Additionally, the dietary guild data presented here are the
first for these species of juvenile fish in the southeast United States.
Food Habits
Stenotomus sp. fed predominantly on copepod species, most importantly
small calanoid copepods. These findings differ from previous studies.
Michelman (1988) found that juvenile S. chrysops (10.5 cm to 12.5 cm) in
Narragansett Bay, Rl, fed mainly on polychaetes and crustaceans, especially the
mysid Neomysis americana. Henwood et al. (1978) analyzed S. caprinas in the
Gulf of Mexico that ranged in size from 23 to 145 mm and they most frequently
consumed polychaetes, harpacticoid copepods, caridean shrimps, fish remains
and unidentified amphipods. Cyclopoid copepods, eucarids, polychaetes, and
bivalves were eaten by scup ranging from 53 to 173 mm SL in Onslow Bay, NO
(Lindquist et al. 1994).
Caprellid and gammarid amphipods, other crustaceans, polychaetes, large
calanoid and cyclopoid copepods comprised 92.1% IRI of C. ocyaras. The diet
42
of C. ocyurus described in this study is similar to the diet of C. striata (black sea
bass) from live bottom habitats in the South Atlantic Bight (Sedberry 1988).
Nineteen small black sea bass (50 - 100 mm SL) fed selectively on motile
epifaunal amphipods, including caprellid species Caprella equilibra, C. penantis
and Phtisica marina. The two latter species were selected as prey on deeper
reefs (23 - 37 m) but not on inner shelf reefs (16-22 m). For larger black sea
bass (> 100 mm SL), decapods and fishes contributed the greatest volume of
prey. Wenner et al. (1983) demonstrated that live bottom habitats have a highly
diverse and abundant assemblage of invertebrates. They found that annelids
were the most numerically dominant (64.7% N), followed by amphipods (17.4%
N) in suction and grab samples collected on the southeast US continental shelf.
Caprellid amphipods were higher in abundance in Centropristis striata stomachs
than in benthic samples (Sedberry 1988).
Following other crustaceans (43.6% IRI), large calanoid copepods and
mysids comprised approximately 40% IRI of the diet of S. phoebe, followed by
caprellid (9.0% IRI) and gammarid (8.3% IRI) amphipods. Diet differed for six
specimens collected off of Florida, Mexico and South America, which consumed
over 90% N shrimp, with the remaining 10% consisting of crabs and bivalves
(Robins and Starck 1961). Fish were larger in that study, with standard lengths
ranging from 34 to 154 mm, and this may account for differences in prey choice.
Diplectrum formosum ate large benthic and epibenthic crustaceans,
including large calanoid copepods, shrimp, fish, caprellid amphipods, and crabs.
43
These findings were similar to previous studies. Seventeen D. formosum (SL
range 48- 196 mm) collected from the Gulf of Mexico consumed xanthid and
portunid crabs followed by shrimp, mysids and polychaetes (Bullock and Smith
1991). Bortone (1971) analyzed 154 specimens (SL range 21 - 223 mm) that
consumed shrimp (56.5% F), crabs (37.7% F), and amphipods (11.1% F), as well
as a variety offish (Perciformes, Scorpaeniformes, Pleuronectiformes, and
Tetradontiformes). Bortone (1971) and others (Wenner et al. 1983) found that D.
formosum has a diurnal pattern, with most activity taking place during daylight
hours. This may be why such a high percentage of specimens in this study were
empty (65%).
Serraniculus pumilio, Serranus baldwini, and the grammistid, Rypticus
maculatus, were also collected from station 43. The pygmy sea bass, S. pumilio,
is known to occur over sand or shell bottoms near irregularities, such as coral or
rock outcrops, but apparently does not have a restricted home range (Hastings
1973). In the Gulf of Mexico, S. pumilio is a territorial species that feeds
indiscriminately on small crustaceans, such as amphipods (42.2% N), isopods
(11.1% N), shrimp (14.2% N) and crabs (12.2% N) (Hastings 1973). The lantern
sea bass, S. baldwini, is a tropical species that is rarely found north of Palm
Beach, Florida (Robins and Starck 1961). In shallow water, S. baldwini prefers
Thalassia beds, whereas prefers shell and coral fragments in deeper waters.
Due to the small numbers of these species captured, a complete diet study is not
possible.
44
Gammahd amphipods were the most important prey category consumed
by R. maculatus, with a %IRI of 87.5%. This species is highly cryptic, preferably
inhabits rocky- bottomed areas, and feeds primarily at night (Guimaráes 1999).
In the Gulf of Mexico, thirty-three R. maculatus with a mean fork length of 197.42
mm consumed xanthid crabs (65.9% IRI), followed by penaeidae shrimp (11.2%
IRI), Gobiidae (9.5% IRI) and pisces (9.5% IRI) (Nelson and Bortone 1996).
Bullock and Smith (1991 ) also reported that R. maculatus feeds heavily on
shrimps and crabs. Courtenay (1967) stated that R. maculatus has invaded
waters north of the tropical areas inhabited by other western Atlantic soapfishes
and has thus almost entirely eliminated competition from relatives. However,
Courtney does not mention the presence of similarly related Serranid species,
and possible interactions. This study represents the first description of feeding
ecology for this species on the southeast US continental shelf.
Stephanolepis hispidus, the planehead filefish, consumed cyclopoid
copepods, other crustaceans, and caprellid amphipods. Studies on S. hispidus
from the southern Atlantic Ocean have been limited to distribution and
abundance studies in floating Sargassum communities (Settle 1993; Stachowicz
and Lindquist 1997), and age and growth analyses (Rogers et al. 2001; Mancera-
Rodríguez et al. 2004). Diet studies conducted on the northeast continental shelf
have found that adult S. hispidus (5-19 cm) consumed polychaetes
predominantly, followed by cephalopoda and echinoderms (Bowman et al. 2000).
In the Gulf of Mexico, unidentified items (56.9% IRI), Ectoprocta (21.1% IRI),
45
xanthid crabs (8.6% IRl), and polychaetes (6.0% IRl) were the most important
prey consumed by S. hispidus (reported as Monacanthus hispidus) with a mean
fork length of 247.75 mm (Nelson and Bortone 1996). Juvenile S. hispidus (21 -
40 mm SL) collected from Apalachee Bay, FL, fed primarily on gammarid
amphipods, only rarely taking copepods (Clements and Livingston 1983).
Observed differences in diet of species between this study and previous
studies may be due to a number of reasons. Temporal and spatial differences in
feeding preferences, prey availability and behavior, and morphological
differences may affect diet choices. For instance, Flenwood et al. (1978) and
Bowman et al. (1987) concluded that most S. caprinus and S. chrysops feed
during the day. Stenotomus sp. for this study were collected from 2000 hours to
500 hours. If these fish were feeding during daylight hours, soft-bodied
organisms would be digested relatively quickly, and therefore less importance
would be placed on them as a dietary category. The high index of relative
importance for copepods may be an artifact of slow digestion time for organisms
with exoskeletons. Temporal differences in feeding may have caused diet
differences between C. ocyurus and S. phoebe as well, such that the other
crustacean category in S. phoebe comprised amphipods that were eaten earlier
than those consumed by C. ocyurus.
Vertical and horizontal differences in prey behavior can have drastic
effects on availability to predators. Vertical migration of small calanoid copepods
from the depths into the water column during the night may result in more of
46
these organisms being eaten by Stenotomus sp., since these fish tend to feed
higher in the water column compared to the other fish species in this study
(Sedberry 1983). Copepods are also able to form dense swarms that may attract
schooling fishes and increase the amount of these organisms eaten (Hamner
1979). Similarly, prey items may be available at different times. For example,
the demersal behavior of amphipod species may cause spatial disparities in the
number of gammarid amphipods on the sea floor (Alldredge and King 1985).
During the night, large demersal specimens (> 2 mm) moved up 1 to 2 meters off
the bottom, and were hypothesized to take advantage of stronger currents to
disperse them into more favorable feeding and breeding habitats.
Morphological differences between predator species may also affect diet
choices. Morphological adaptations for feeding include alterations in body form,
mouth shape and position, the presence/absence of marginal and pharyngeal
teeth, gillrakers and the length of the alimentary canal (Wootton 1990). Serranas
phoebe was, on average, 10 mm smaller than C. ocyurus, and this size
difference may influence the prey eaten. Following other crustaceans, large
calanoid copepods were the next most important prey category for S. phoebe.
Calanoid copepods average no larger than 5 mm in length (Ruppert et al. 2004).
They are smaller than caprellid and gammarid amphipods, which range in size
from 5 to 15 mm in length (Ruppert et al. 2004). Stenotomus sp. and
Stephanolepis hispidus have smaller mouths than the other six species, and fed
on smaller prey items, such as calanoid and cyclopoid copepods.
47
Cumulative prey curves are recommended as an a priori method in
determining the adequacy of sample sizes in diet studies (Ferry and Cailliett
1996; Cortes 1997). However, to date there is no objective, quantitative method
to determine if the curve has reached an asymptote. In this study, an asymptotic
function fit with non-linear regression techniques was used to model the
asymptotic relationship of dietary data. An advantage of this is a quantitative,
objective method to assess if a curve has reached an asymptote. Conversely,
the data inputted into the model, as well as the initial estimates, will greatly affect
the outcome. In other words, validity of the model will depend greatly on the
maximum number of prey categories, and the rate of change in the number of
cumulative new prey categories found in each individual. The maximum number
of prey categories is dependent upon the level of identification of the prey items.
For instance, if prey were grouped into less specific categories (crustaceans,
fish, etc), the maximum number of prey categories would be lower, and therefore
fewer specimens would need to be analyzed before an asymptotic relationship
occurred. As well, the rate of change in the number of prey categories depends
upon a number of factors, including spatial orientation, temporal behavior,
swarming, cryptic morphologies, and the habitats in which the predators and prey
exist.
Cortes (1997) stresses the importance of standardizing methods of dietary
analyses. It must be stressed that the numbers of sufficient sample sizes for
each of the species in this study will most likely be different for a follow-up study
48
during a different time (daytime), and/or at different stations. However, in
agreement with Cortes (1997), to facilitate comparisons among studies, it is
advantageous to utilize the same methods of analyses.
Dietary Overlap
No biologically significant overlap values were found among the species
examined. Several reasons may account for the lack of biologically significant
overlap exhibited by the fishes in this study. Optimal foraging theory predicts that
as food becomes scarce, predators will take a wide variety of food and similar
predators occupying the same habitat will converge in diet (Pyke et al. 1977;
Sedberry 1983; Mittelbach and Osenberg 1994). Alternatively, it has been
hypothesized that as food density decreases, predators living in similar habitats
specialize in eating certain prey items, consequently decreasing dietary overlap.
In this case, high diet overlap would be expected if food were abundant.
The results from this study suggest that juvenile fishes are not selective in
their feeding, in that each species ate a variety of prey items. It is hypothesized
that the abundance of microhabitats provided by rocky-reefs and hard-bottom
areas promote a high diversity of invertebrates, and therefore an increase in prey
for juvenile fishes. To determine if the lack of overlap is due to food shortage or
food abundance, additional information on prey availability and microhabitat is
necessary. Furthermore, diet overlap relationships change considerably with
season and fish size.
49
Both Schoener’s index and Bray-Curtis similarities used in multivariate
analyses provide insight into dietary overlap. Schoener’s index describes
overlap between species, while the Bray-Curtis similarities describe overlap
between individuals. Untransformed total specific volumes of each prey category
were used in the Schoener’s index, while double root transformed average
specific volumes were used for multivariate analyses. In this study, there was no
biologically significant overlap found by the Schoener’s index between any
species, yet 10 feeding guilds were delineated easily at the 40% level using
multivariate analyses. This finding suggests that analyzing dietary overlap
between each individual fish gives different results than overlap values computed
between species, supporting the occurrence of stage structured differences in
resource use.
Ontogenetic Diet Variations
As demonstrated, Stenotomus sp., C. ocyurus, and S. phoebe go through
ontogenetic changes in diet during the juvenile life-history stage. Size of prey
and number of prey categories increase with growth for all three species.
Morphological differences between size classes may explain these shifts. In
other words, as fish grow they are more capable of capturing and ingesting larger
prey items. However, larger predators consistently include small-bodied prey in
their diet, possibly representing profitable foraging behaviors when size-
50
dependent probabilities of encounter and capture are combined with handling
costs of prey (Scharf et al. 2000).
Schoener’s index indicated that dietary overlap among size classes within
each species was biologically significant for most pairs. This was expected since
intraspecific interactions are often more intense by many orders of magnitude
than are interspecific interactions (Schroder and Rosenzweig 1975). It has been
argued that individuals within the same population are likely to be most similar in
their resource requirements and so are potentially intense competitors (Wootton
1990). As a result, growth of smaller individuals may be suppressed by the
presence of larger individuals (Dou et al. 2004). Superimposition of designated
size classes on the ordination plot for each species showed considerable overlap
for all sizes of Stenotomus sp., C. ocyurus, and S. phoebe. Different values
were used for each method, therefore it is difficult to compare overlap values.
To determine if these shifts in food preferences have effects on
community structure and dynamics, further examination of ontogenetic changes
in diet across temporal and spatial axes are necessary to fully understand
juvenile resource use on the southeast US continental shelf.
Trophic Guild Analyses
The term guild has been used in many terrestrial (eg. Pianka 1980),
freshwater (eg. Lappalainen and Kjellman 1998), and marine ecosystem studies
(eg. Garrison and Link 2000) since it was first used by Root (1967) to define
51
feeding guilds of blue-gray gnatcatchers. Root (1967) defined a guild as a group
of species that exploit the same class of environmental resources in the same
way. He further described the term as not limited by taxonomic boundaries (Root
1986). Guilds can be based on several factors or combinations of factors,
including diet, morphology, and behavior (Nelson and Bortone 1996). In this
study, guilds were delineated using dietary data, regardless of morphology or
behavior of the fish species. The ten dietary guilds reflect similarity in the
utilization of specific prey categories. Within each guild, only one prey category
usually accounted for > 50% offish diets, with 1 to 9 prey categories contributing
< 50 % of the diet. On average, larger fish in guilds 1A and 1B ate larger, benthic
prey items (mysids and gammarid amphipods), compared to smaller fish eating
smaller epibenthic and planktivorous prey (i.e. caprellid amphipods and
copepods) in the remaining 8 guilds. However, within species there was no
difference in size between guilds, with S. hispidus as an exception.
Trophic analyses from the southeast US continental shelf are lacking for
juvenile fishes. Nelson and Bortone (1996) analyzed the trophic structure among
artificial reef fishes in the northern Gulf of Mexico, and included C. ocyurus, R.
maculatus and S. hispidus (reported as Monacanthus hispidus), as well as 22
other species, in their study. Using a Bray-Curtis cluster analysis dendrogram,
C. ocyurus (mean fork length 172 mm) and R. macuiatus (mean fork length 197
mm) were grouped in the lower structure crustacean predator guild, because
xanthid crabs had the highest index of relative importance for both species.
52
Unfortunately, M. hispidus (mean fork length 248 mm) was not included in the
cluster analysis, since < 5 fish were analyzed. Comparatively, this study focused
on fish of smaller sizes, and clustered individuals rather than species, yet found
Centropristis ocyurus and Rypticus maculatus feed on larger, benthic
crustaceans.
For the purposes of this study, guilds were defined based on food
resources, which have been found to be important characteristics in structuring
communities (Scharf et al. 2000). If a guild is defined based on a resource or
characteristic that is relatively unimportant in structuring the community, a
prediction of community change based on that guild system will likely have little
relevance to community dynamics (Austen et al. 1994). However, choosing what
level of resource is ecologically relevant is a subjective process. For instance, 10
feeding guilds with a ^0% similarity were designated in this study; yet, these
guilds may not describe an organizing characteristic within the community. The
measure of dietary similarity using the Bray-Curtis coefficient provided a resource
matrix in which species with similar values were placed in the same guilds.
Pianka (1980) suggested that guilds were areas of “intense interspecific
competition with strong interactions within guilds, but weak interactions between
members of different guilds.” Nonetheless, the degree of competition among
species in this study cannot be interpreted from the resource matrix alone. As
discussed previously, additional information on prey availability and microhabitat
is necessary to understand competitive interactions. As well, to test the
53
generality of guild definitions is to study the changes in guild member biomass or
abundance through time or in various systems (Orth 1980).
Austen et al. (1994) reviewed the importance of the guild concept in
fisheries management and indicated guild management would be effective if
statistically delineated guilds based on key resources were used. They also
suggest the development of guilds that function as “super species” - units that
respond to environmental change (including fishing pressure) in a more
predictable manner than any individual member species. Information about the
functioning of the fish community can be obtained at a higher level of
organization by developing guilds based on similar responses of guild members
to environmental change (Skagen et al. 1991). Consequently, the use of guilds
to characterize the condition of specific ecosystems meets the requirements of
ecosystem-based management and the conservation of biodiversity on the
southeast US continental shelf.
In summary, the eight juvenile fish species examined in this study appear
to be opportunistic feeders capable of utilizing a wide variety of prey items.
Stenotomus sp. and Stephanolepis hispidas feed predominantly on small,
planktonic copepods from the families Calanoida and Cyclopoida. The five
Serranid species and one Grammistid feed largely on gammarid and caprellid
amphipods, with a preference for larger-sized prey items. Prey selection may be
explained by habitat choices. It may be that Stenotomus sp. and S. hispidas
spend more time higher in the water column, while the six other species occupy
54
habitats on the benthos. Although diets and presumed feeding habitat differed,
all fish were caught on the bottom. Differences in vertical microhabitat use may
contribute to differences in diet. Conversely, morphological differences may
influence the size of prey consumed. Fish species with larger mouths are able to
ingest larger prey. This increase in prey size was suggested with the analysis of
ontogenetic size classes in Stenotomus sp., C. ocyurus, and S. phoebe. Prey
availability is an important component to diet studies. Extensive invertebrate
collections from microhabitats were not made in conjunction with fish samples,
therefore prey availability data were lacking in this study.
Cumulative prey curves are necessary to determine the adequacy of
sample sizes in any diet study. The number of individuals needed for this study
ranged from 8 to 55, however these computations will depend upon the level of
identification of the prey, as well as temporal and spatial characteristics of diet
samples. It is recommended that an objective, quantitative method be developed
to assess the asymptotic relationship of cumulative prey curves. In this study an
objective, quantitative method to determine an asymptotic relationship in
cumulative prey curves was used.
Ten feeding guilds were established within the juvenile fish community on
the southeast US continental shelf based upon habitat and prey size. Three
guilds made up the larger, benthic predators, while six guilds made up the small
epibenthic and planktonic predators. It is important to stress that these guilds will
change due to many factors, including differences in when and where predators
55
are eating, and temporal and spatial differences in prey availability. As well,
other predator species not included in this study may add other guilds and/or
increase the number of individuals within each guild. Identifying trophic guilds is
a useful first step for defining groups of functionally similar species, and is a
potentially valuable tool to simplify interspecies interactions on the southeast US
continental shelf.
To successfully implement an ecosystem-based management approach, a
number of tasks are recommended (NOAA 1999; Busch et al. 2003) including
descriptions of habitat needs of different life-history stages to delineate
geographic areas of an ecosystem, and biological indicators such as trophic
structure and species diversity to characterize the condition of specific
ecosystems. The results from this study provide dietary data on fish species that
have not previously been studied, and present the first data on trophic guild
organization within the juvenile fish community on the southeast US continental
shelf. Specifically, the results of this study show that both pelagic and benthic
prey items are important food resources of the juvenile fish community, and
therefore provide important bentho-pelagic links in this ecosystem. As well,
size-structured populations do exist in the juvenile fish community on the
southeast US continental shelf. These findings are important for the classification
of essential fish habitats in this area, are necessary for an ecosystem-based
approach to fisheries management, and the conservation of marine biodiversity
56
in the vicinity of Gray’s Reef National Marine Sanctuary and on the southeast US
continental shelf.
Table 2-1 Fish species, total number of specimens included in diet analyses and station locations where fish were
collected. See Figure 2-1 for station locations.
Family Species Common Name N Station Location
Sparidae Stenotomus sp. Unidentified porgy 136 2.1-2.4,23,25, 29,35, 37,38
Serranidae Centropristis ocyurus Bank sea bass 70 38, 41,42, 43
Serranidae Serranus phoebe Tattler 43 38, 39, 43, 44
Serranidae Diplectrum formosum Sand perch 24 2.3, 23, 37, 38, 43
Grammistidae Rypticus maculatus Whitespotted soapfish 24 43
Monacanthidae Stephanolepis hispidus Planehead filefish 23 36, 40, 43
Serranidae Serraniculus pumilio Pygmy sea bass 2 43
Serranidae Serranus baldwini Lantern bass 1 43
Total 323
cn
-vi
58
Table 2-2 Prey category list and number of stomachs containing prey in all fish
species. Y = included in guild analyses, N = not included in guild analyses.
Prey Category Number Stomachs Included in
^measured volumetrically Containing Category Guild Analyses
Egg 23 Y
Hydrozoa 1 N
Polychaeta 15 Y
Mollusca
Bivalvia 7 N
Chelicerata
Acari 1 N
Crustacea
Unknown crustaceans 53 N
Cladocera 19 Y
Cirripedia 2 N
Ostracoda 11 Y
Copepoda
Calanoida
Small (prosome < 2 mm) 57 Y
Large (prosome >2 mm) 65 Y
Cyclopoida 71 Y
Harpacticoida 41 Y
Amphipoda
Caprellidea 52 Y
Gammaridea 60 Y
Hyperiidea 3 N
Tanaidacea 3 N
Mysidacea 26 Y
Decapoda
Shrimp 3 N
Crab 8 N
Hermit Crab 1 N
Teleostei
Fish 6 N
Chordata
Appendicularia (Larvacea) 7 N
Other Category Number Stomachs Included in
not measured volumetrically Containing Category Guild Analyses
Coral 2 N
Copepod Egg Sac 4 N
Fish Scales 34 N
Nematoda 124 N
Sand Grains 86 N
Shell Fragments 1 N
Unknown 44 N
Table 2-3 Frequency of occurrence (%F), contribution by numbers (%N), percent specific volume (%SV), and
percent index of relative importance (%IRI) of the different prey categories to the overall diet of Stenotomus sp.,
Centropristis ocyurus, and Serranus phoebe.
Stenotomus so. C. ocyurus S. phoebe
Rrey Category %F %N %SV %IRI %F %N %SV %IRI %F %N %SV ®/olRI
Egg 31.3 0.8 1.1 0.5 2.1 2.1 0.0 < 0.1 —
Hydrozoa 1.5 < 0.1 0.1 < 0.1 — — — — ___ ___ ___ ___
Polychaeta 3.0 < 0.1 0.1 < 0.1 23.4 6.0 15.1 7.8 3.0 0.7 1.6 < 0.1
Mollusca
Bivalvia ('clam') 3.0 < 0.1 < 0.1 < 0.1 2.1 0.4 0.3 < 0.1 3.0 0.7 0.1 < 0.1
Chelicerata
Acari ('water mite') 1.5 < 0.1 < 0.1 < 0.1 — ___ — —- — —
Crustacea
Other crustacean 9.0 0.1 .9 0.1 51.9 6.0 7.6 1 1.1 63.6 15.4 38.2 43.6
Cladocera 23.9 8.1 7.8 2.9 2.1 0.4 < 0.1 < 0.1 — — — —
Cirripedía 3.0 0.1 0.1 < 0.1 ___ — — — — — —
Ostracoda 4.5 0.1 < 0.1 <0.1 14.9 3.0 0.8 0.9 3.0 0.7 0.1 <0.1
Copepoda
Calanoida
Small (< 2 mm) 76.1 63.0 65.6 73.9 2.1 1.3 0.1 < 0.1 15.2 7.4 1.0 1.6
Large ( s2 mm) 41.8 3.6 6.5 3.2 27.7 9.9 4.0 6.1 51.5 22.8 18.9 27.5
Cyclopoida 71.6 19.5 13.7 17.9 21.3 8.6 3.1 3.9 9.1 13.2 2.1 1.8
Harpacticoida 46.5 2.2 0.5 1.0 19.1 5.6 0.4 1.8 3.0 0.7 < 0.1 < 0.1
Amphipoda
Caprellidea 16.4 0.7 1.1 0.2 57.4 30.5 14.8 40.9 27.3 1 1.0 9.0 7.0
Gammaridea 20.9 0.6 0.6 0.2 44.7 17.2 14.5 22.3 27.3 8.1 8.3 5.7
Hyperiidea 1.5 < 0.1 < 0.1 < 0.1 4.3 0.9 2.6 0.2 — —
Tanaidacea — — — 4.3 0.9 0.5 < 0.1 3.0 0.7 0.1 < 0.1
Mysidacea 7.5 0.1 0.5 < 0.1 17.0 3.9 4.9 2.3 27.3 16.2 19.0 12.3
Decapoda
Shrimp — — — — 2.1 0.4 18.0 0.6 — — — —
Crab — — — 8.5 1 .7 1 .4 1 .8 — -— —
Hermit Crab — ... 2.1 0.4 0.6 < 0.1 ... —- —-
Teleostei
Fish — ___ ___ — 4.3 0.9 .2 0.1 6.1 2.2 1 .7 0.3
Chordata
Append icularia 10.4 1 o 0.3 0.1 — — — —
TOTAL 100 100 100 100 100 100 100 100 100
Other Category
Copepod Egg Sac 3.0 0.6 — ... — - - — — — — —
Coral — — — 2.9 1.9 — — — — — —
Empty 8.1 — — — 14.3 — — — 18.6 — — —
Fish Scales 22.2 4.2 — — 4.3 3.7 “ — — — —
Nematoda 83.7 82.0 — — 8.6 7.5 — — 4.7 32.0 — —
Sand Grains 33.3 10.2 — — 45.7 67.3 — — 16.3 68.0 — —
Shell Fragments 0.7 0.1 — ”— — — — — — — "
Unknown 19.3 3.0 ?" 25.7 19.6 — — ?? ??
Total # Stomachs 136 70 43
Total # Stomachs w/ Food 67 47 33
% Empty 50.7 32.9 23.3
Stomach Weight (mean ± SD) 0.0134 g ± 0.0405 g 0.01 13 g ± 0.0097 g 0.0067 g ± 0.0106 g cn
Standard Length (mean ± SD) 33.8456 mm ± 9.8326 mm 27 1 143 mm ± 7.8769 mm 19.8372 mm ± 8.5466 CD
Table 2-4 Frequency of occurrence (%F), contribution by numbers (%N), percent specific volume (%SV), and
percent index of relative importance (%IRI) of the different prey categories to the overall diet of Diplectrum
formosum, Rypticus maculatus, and Stephanolepis hispidus.
D. formosum R. maculatus S. hispidus
Prey Category %F %N %sv %IRI %F %N %SV %IRI %F %N %SV %IRI
Egg — — — — — — — — 8.3 1.0 0.3 0.1
Hydrozoa — — — — — — — — — — — —
Polychaeta — — — — — — — — 8.3 1.0 1.1 0.2
Mollusca
Bivalvia ('clam') — — — — — — — — 25.0 3.9 0.6 1.1
Chelicerata
Acari ('water mite') — — — — — — — — — —
Crustacea
Other crustacean 25.0 12.5 7.8 8.1 21.1 4.8 1.6 1.2 50.0 5.8 19.9 12.5
Cladocera — — — — — — — — 16.7 5.8 4.1 1.6
Cirripedia — — — — — — — — — — — —
Ostracoda — — — — — — — — — -— — —
Copepoda
Calanoida
Small (< 2 mm) — — — — — — — — — — —
Large ( S:2 mm) 50.0 50.0 11.0 48.5 — — — — 25.0 3.9 7.3 2.7
Cyclopoida — — — — 10.5 8.3 0.2 0.8 66.7 71.8 35.6 69.6
Harpacticoida — — — — — — — — — — — —
Amphipoda
Caprellidea 12.5 6.3 6.0 2.4 — — — — 33.3 6.8 31.1 12.3
Gammaridea — — — — 84.2 78.6 36.6 87.5 — — — —
Hypehidea — — — — — — — — — — — —
Tanaidacea — — — — — — — — — — — —
Mysidacea — — — — 21.1 4.8 5.3 1.9 — — — —
Decapoda
Shrimp 25.0 12.5 58.2 28.1 — — — — — — — —
Crab 12.5 6.3 < 0.1 1.2 15.8 3.6 56.3 8.5 — — — —
Hermit Crab — — — — — — — — — — — —
Teleostei
Fish 25.0 12.5 17.0 1 1 .7 — — — — — — — —
Chordata
Append icula ría — — — — — — — — — — — —
TOTAL 100 100 100 100 100 100 100 100 100
Other Category
Copepod Egg Sac — — — — — — - - — — - —
Coral — — — ~ ~ ~ — “ “ “ —
Empty 66.7 — — — 16.7 — — — 43.5 — — ~
Fish Scales — — — — 4.2 4.2 — — — — — —
Nematoda 4.2 100 — — — — — — 8.7 100 — —
Sand Grains — — — — 8.3 95.8 — — — — ~ —
Shell Fragments — — — — — — — ~ — — — “
Unknown — — — — — —
Total # Stomachs 24 24 23
Total # Stomachs w/ Food 8 19 12
% Empty 66.7 20.8 47.8
Stomach Weight (mean ± SD) 0.2116 g ± 0.7986 g 0.0431 g ± 0.0190 g 0.0314 g ± 0.1184 g O)
Standard Length (mean ± SD) 33.0833 mm ± 43.0631 mm 45.9167 mm ± 7.30049 mm 20 1304 mm ± 15.5893 mm o
Table 2-5 Observed values of diet overlap between six species of juvenile fishes. Values marked with # are
considered ‘biologically significant.’
Species C. ocyurus S.phoebe D. formosum R. maculatus S. hispidus
Stenotomus sp. 0.33 0.30 0.18 0.19 0.41
C. ocyurus 0.42 0.33 0.33 0.48
S. phoebe 0.48 0.34 0.56
D. formosum 0.14 0.40
R. maculatus 0.19
Table 2-6 Observed values of Schoener’s index of dietary overlap for size classes of Stenotomus sp., Centropristis
ocyurus, and Serranus phoebe. Values marked with # are considered ‘biologically significant.’
Stenotomus sp.
Size Class 31-34 mm 35-40 mm >41 mm
17-30 mm 0.65" 0.63" 0.64"
31-34 mm 0.78" 0.57
35-40 mm 0.51
Centropristis ocyurus
Size Class 26-30 mm >31 mm
11-25 mm 0.67" 0.53
26-30 mm 0.62"
Serranus phoebe
Size Class >20 mm
< 20 mm 0.67"
05
K>
81°0’W 80°0’W
Figure 2-1 Map of study area and the individual sample sites and station numbers used for sampling. Four
stations were located around Gray’s Reef National Marine Sanctuary. O)
CO
64
Stenotomus sp.
60 -1
50 ' I
N = 136
? = 69
? = 67
0-9 10-19 20-29 30-39 40-49 50-59 60-69 70-79 80-89 250
Standard Length (mm)
Centropristis ocyurus
Serranus phoebe
N = 23
_ ? = 15
? II 00
- H .... B
^^ 1
0-9 10-19 20-29 30-39 40-49 50-59 60-69 70-79 80-89 ^0
Standard Length (mm)
Figure 2-2 Size frequency distributions of Stenotomus sp., Centropristis ocyurus,
Serranus phoebe, and Diplectrum formosum. ? = empty stomachs. ? =
stomachs containing prey.
Rypticus maculatus
12 1 N = 24
? = 5
? = 19
0-9 10-19 20-29 30-39 40-49 50-59 60-69 70-79 80-89 ^0
Standard Length (mm)
Stephanolepis hispidus
N = 23
? = 11
? = 12
10-19 20-29 30-39 40-49 50-59 60-69 70-79 80-89 à90
Standard Length (mm)
serranicuius pumifio
Figure 2-3 Size frequency distributions of Rypticus maculatus, Stephanolepis
hispidus, Serranicuius pumilio, and Serranus baldwini. ? = empty stomachs.
= stomachs containing prey.
66
Stenotomus sp. Centropristis ocyurus
Number of Guts Analyzed
Figure 2-4 Randomized cumulative prey curves for Stenotomus sp.,
Centropristis ocyurus, Serranus phoebe, Rypticus maculatus, Stephanolepis
hispidus, and Diplectrum formosum. Mean values are plotted. Error bars
represent ± standard deviation. Red line represents estimated curve. Red
arrows represent the number offish at which the curve begins an asymptotic
relationship.
67
Stenotomus sp.
Calanoida - small I I Harpacticoida
Cyclopoida
Calanoida - large Caprellidea
Cladocera I I Gammandea
Figure 2-5 An Index of Relative Importance (IRI) diagram of the prey categories
and their numerical (%N), specific volumetric (%SV), and frequency of
occurrence (%F) values for Stenotomus sp. Prey items are ranked in order of
percent IRI.
68
Centropristis ocyurus
40 -1
% F
?H Caprellidea ?? Polychaeta ^?1 Mysidacea 1 1 Ostracoda
1 1 Gammaridea 1 1 Calanoida - large 1 1 Harpacticoida Shrimp
WÊÊÊl Other Crustacean ?? Cyclopolda 1 1 Crab
Figure 2-6 An Index of Relative Importance (IRI) diagram of the prey categories
and their numerical (%N), specific volumetric (%SV), and frequency of
occurrence (%F) values for Centropristis ocyurus. Prey items are ranked in order
of percent IRI.
69
Serranus phoebe
H
10 %
% F
?? other Crustacean Caprellldea Calanoida - small
I I Calanoida - large Gammaridea Fish
Mysidacea Cyclopoida
Figure 2-7 An Index of Relative Importance (IRI) diagram of the prey categories
and their numerical (%N), specific volumetric (%SV), and frequency of
occurrence (%F) values for Serranus phoebe. Prey items are ranked in order of
percent IRI.
70
Diplectrum formosum
60
40 -
%N
20 -
0 -
20 -
%SV 40 -
60 -
80 -
Calanoida - large Fish Caprellidea
Shrimp Other Crustacean Crab
Figure 2-8 An Index of Relative Importance (IRI) diagram of the prey categories
and their numerical (%N), specific volumetric (%SV), and frequency of
occurrence (%F) values for Diplectrum formosum. Prey items are ranked in
order of percent IRI.
71
Rypticus maculatus
100
80
60
% N
40
20
0
20
% SV
10%
40
% F
60
Gammaridea
Crab
Mysidacea
Other Crustacean
Cyclopoida
Figure 2-9 An Index of Relative Importance (IRI) diagram of the prey categories
and their numerical (%N), specific volumetric (%SV), and frequency of
occurrence (%F) values for Rypticus maculatus. Prey items are ranked in order
of percent IRI.
72
Stephanolepis hispidus
Figure 2-10 An Index of Relative Importance (IRI) diagram of the prey categories
and their numerical (%N), specific volumetric (%SV), and frequency of
occurrence (%F) values for Stephanolepis hispidus. Prey items are ranked in
order of percent IRI.
73
Stenotomus sp.
50 -
>41 mm
Standard Length (mm)
Frequency Centropristis ocyurus30252015 H10 15Frequency 0 0-25 mm 26 - 30 mm >31 mmStandard Length (mm)Serranas phoebe30 n25 -20 -Frequency
!
0 4-
< 20 mm >20 mm
Standard Length (mm)
Figure 2-11 Frequency histograms showing sizes (standard length, mm) offish
analyzed of Stenotomus sp., Centropristis ocyurus, and Serranus phoebe in each
size class. ? = empty stomachs. ? = stomachs containing prey.
74
Size Class
Figure 2-12 Range and median values for standard lengths of Stenotomus sp. in
each size class. Whisker bars represent and 90^*^ percentiles.
75
Stenotomus sp.
Figure 2-13 Cumulative prey curves for four size classes of Stenotomus sp.: A)
0-30 mm SL, B) 31 - 34 mm SL, C) 35 - 40 mm SL, D) >41 mm SL. Red
arrows represent the number offish at which the curve begins an asymptotic
relationship.
76
0-30 mm 31 -34 mm
35 - 40 mm >41 mm
? Acari ? Append icularia ? BIvalvia ? Calanoida large
? Calanoida small ? Caprellidea ? Cirripedla ? Cladocera
? Cyclopolda ? Egg ? Gammahdea ? Harpacticolda
? Hydrozoa ? Hyperildea ? Mysidacea ? Ostracoda
? Other Crustacean ? Polychaeta
Figure 2-14 Mean percentage composition by average specific volume for the
major prey categories of the diet of four sequential size classes of Stenotomus
sp.
77
50-
40-
O)
5 30-
(O
?c
c
<0
(O
20-
10-
0-25 mm 26 - 30 mm >31 mm
Size Class
Figure 2-15 Range and median values for standard lengths of Centropristis
ocyurus in each size class. Whisker bars represent and 90*'^ percentiles.
Red dot represents one outlier.
78
Centropristis ocyurus
No. Guts Analyzed
Figure 2-16 Cumulative prey curves for three size classes of Centropristis
ocyurus: A) 0 - 25 mm SL, B) 26 -30 mm SL, C) >31 mm SL. Red arrows
represent the number offish at which the curve begins an asymptotic
relationship.
79
0-25 mm 26 - 30 mm
>30 mm
? Bivalvia ? Calanoida large B Calanoida small
? Caprellidea ? Cladocera ? Crab
? Cyclopoida ? Egg B Fish
? Gammaridea ? Harpacticoida ? Hermit Crab
? Hyperiidea B Mysidacea B Ostracoda
? Other Crustacean B Polychaeta B Shrimp
? Tanaidacea
Figure 2-17 Mean percentage composition by specific volume for the major prey
categories of the diet of three sequential size classes of Centropristis ocyurus.
80
s(Ltmaendgmatrhd)
Figure 2-18 Range and median values for standard lengths of Serranas phoebe
in each size class. Whiskers represent 10‘^ and 90‘^ percentiles.
81
Serranas phoebe
(/)
o
C5>
o
o5
o
>.
Q_
d
No, Guts Analyzed
Figure 2-19 Cumulative prey curves for tvyo size classes of Serranas phoebe: A)
< 20 mm SL, B) >20 mm SL. Red arrows represent the number offish at which
the curve begins an asymptotic relationship.
82
< 20 >20
? Bivalvia ? Calanoida large ? Calanoida small ? Caprellidea
? Cyclopoida ? Fish ? Gammaridea ? Flarpacticoida
? Mysidacea ? Ostracoda ? Other Crustacean ? Polychaeta
? Tanaidacea
Figure 2-20 Mean percentage connposition by specific volume for the major prey
categories of the diet of two sequential size classes of Serranus phoebe.
represents ^0% similarity.
C»
w
84
Guild 1A
Guild 1B
0.90 -
0.80 H
o 0.70 -
Û 0.60 -
o
c 0.50 ^
o
r 0.40 H
o
Q.
O 0.30 -
0. 0.20 H
0.10 J
0.00 4-
Gammaridea Ostracoda Cyclopoida Harpacticoida
Guild 1C
Figure 2-22 Mean proportional composition by volume of major prey categories
in Guilds 1A, 1B, and 1C. Bars represent average prey content within guilds.
Four prey items in guild 1A accounted for 76% of predator diets, fours prey items
in guild 1B accounted for 81 % of predator diets, and eight prey items in guild 1C
accounted for 91 % of predator diets.
85
Guild 2A
0.60 n
.£ 0.50 ^
Q
*5 0.40 ^
o 0.30 -
?-E
o 0.20 -
£ 0.10 -
0.00 -
,# # .r-r
O' ç/ cf
&
Guild 2C
0.60 n
Figure 2-23 Mean proportional composition by volume of major prey categories
in Guilds 2A and 2C. Bars represent average prey content within guilds. Ten
prey items in guild 2A accounted for 82% of predator diets, and seven prey items
in guild 2C accounted for 67% of predator diets.
86
Guild 2D
0.60 -
Guild 2E
Guild 2F
0.70 n
0.60 1 l-OO n
0.50
¡5 0.80 -
c 0.40 °
o 1 0.60 -
c H -V O
o
Q. J r 0.40 J
o o?
Û. 0.10 J o 0.20 -k.
Û.
0.00 0.00 ^
Figure 2-24 Mean proportional composition by volume of major prey categories
in Guilds 2D, 2E, and 2F. Bars represent average prey content within guilds.
Ten prey items in guild 2D accounted for 91% of predator diets, two prey items in
guild 2E accounted for 71% of predator diets, and two prey items in guild 2F
accounted for 100% of predator diets.
87
Stress: 0.14
A = Group 1 ? = Group 2 D = Group 3
Figure 2-25 Non-metric multidimensional scaling plot with superimposed groups.
88
Stress: 0.14
3
^ 1^ M ^
1C
1A ? 2C
A IB 2D
1C ? 2E
• 2A 2F
V 2B X 3
Figure 2-26 Non-metric multidimensional scaling plot with superimposed feeding
guilds.
89
60-
50-
E
E
40-
O)
c
0}
?§30-
c
(Q
CO
20-
10-
Figure 2-27 Range and median values for standard lengths of fish within each
feeding guild. Whisker bars represent 10‘^and 90*^ percentiles. Red dots
represent outliers.
90
60-
50-
r 40-
5 30-
1C 2A 2C 2D 2E 2F
Guild
1 1 1 1 1 r T T
1A IB 1C 2A 2B 2C 2D 2E
Guild
Figure 2-28 Range and median values for standard lengths of Stenotomus sp.
(A), Centropristis ocyurus (B), and Serranas phoebe (C) within each guild.
Whisker bars represent 10*^ and percentiles. Red dots represent outliers.
91
Guild
Guild
Guild
Figure 2-29 Range and median values for standard lengths of R. maculatus (A),
D. formosum (B), and S. hispidus (C) within each guild. Whisker bars represent -
10*'^ and 90‘^ percentiles. Red dots represent outliers.
92
? 0-30 mm ? 35 - 40 mm
? 31 - 34 mm A > 41 mm
Figure 2-30 Multidimensional scaling plot of dietary data with superimposed size
class for Stenotomus sp.
93
I I 0 — 25 mm 26 - 30 mm > 31 mm
Figure 2-31 Multidimensional scaling plot of dietary data with superimposed size
class for Centropristis ocyurus.
94
Figure 2-32 Multidimensional scaling plot of dietary data with superimposed size
class for Serranus phoebe.
95
A 2.1 29
? 2.2 X 35
? 2.4 37
? 23 V 38
• 25
Figure 2-33 Multidimensional scaling plot of dietary data with superimposed
station number for Stenotomus sp.
96
A 38 ? 42
T 41 ^43
Figure 2-34 Multidimensional scaling plot of dietary data with superimposed
station number for Centropristis ocyurus.
97
A 38 ? 43 ? 44
Figure 2-35 Multidimensional scaling plot of dietary data with superimposed
station number for Serranas phoebe.
98
? 38 A 43
Figure 2-36 Multidimensional scaling plot of dietary data with superimposed
station number for Diplectrum formosum.
99
A 43
Figure 2-37 Multidimensional scaling plot of dietary data with superimposed
station number for Rypticus maculatus.
100
? 36 A 40
Figure 2-38 Multidimensional scaling plot of dietary data with superimposed
station number for Stephanolepis hispidus.
CHAPTER 3
DISCUSSION
This study represents the first attempt at describing the diets and diet
overlap of several juvenile fishes on the southeast US continental shelf. For
juvenile Serranus phoebe, Stephanolepis hispidas, and Rypticus macuiatus this
represents the first quantitative information on foraging ecology on the southeast
US continental shelf. Additionally, the dietary guild data presented here are the
first for these species of juvenile fish in the southeast United States.
Food Habits
Stenotomus sp. fed predominantly on copepod species, most importantly
small calanoid copepods. These findings differ from previous studies.
Michelman (1988) found that juvenile S. chrysops (10.5 cm to 12.5 cm) in
Narragansett Bay, Rl, fed mainly on polychaetes and crustaceans, especially the
mysid Neomysis americana. Henwood et al. (1978) analyzed S. caprinas in the
Gulf of Mexico that ranged in size from 23 to 145 mm and they most frequently
consumed polychaetes, harpacticoid copepods, caridean shrimps, fish remains
and unidentified amphipods. Cyclopoid copepods, eucarids, polychaetes, and
bivalves were eaten by scup ranging from 53 to 173 mm SL in Onslow Bay, NC
(Lindquist et al. 1994). Age-0 classes of longspine porgy (S. caprinas-, SL range
102
25-74 mm) consumed polychaetes (33.0% dry weight), shrimps (6.7% dry
weight), and copepods (4.0% dry weight), along the Texas coast in depths 18 to
44 m. Fish collected from deeper areas (45-73 m; SL range 40-51 mm) ate
polychaetes (60.9% dry weight), gastropods (6.4% dry weight), and copepods
(4.9% dry weight) (Sheridan and Trimm 1983). Stenotomus chrysops (SL range
51 - 100 mm) collected from the Middle Atlantic Bight consumed hyperiid and
gammarid amphipods, polychaetes, and copepods (Sedberry 1983).
Caprellid and gammarid amphipods, other crustaceans, polychaetes, large
calanoid and cyclopoid copepods comprised 92.1% IRI of C. ocyurus. In the Gulf
of Mexico, C. ocyurus fed predominantly on xanthid crabs (29.6% F),
polychaetes (11.1% F), and caridean shrimp (7.4% F) (Bullock and Smith 1991).
However, the twenty-seven individuals included in that study ranged in size from
11 to 200 mm SL. This difference in size may contribute to diet differences.
Caprellid and gammarid amphipods range in size from 5 to 15mm, and are, on
average, smaller in size than shrimps and crabs (Ruppert et al. 2004). Caprellid
amphipods are chosen by fish on the basis of size, behavior, and visual contrast
with the substratum (Caine 1989), and are adapted to clinging to complex
substrata such as seaweeds, bryzoans and hydroids (Caine 1989, 1998; Ruppert
et al. 2004). As well, gammarid amphipods are chosen prey items of many
juvenile fishes (Caine 1980, 1983). Most are benthic and usually remain closely
associated with the bottom, despite being capable of swimming (Ruppert et al.
2004).
103
The diet of C. ocyurus described in this study is similar to the diet of C.
striata (black sea bass) from live bottom habitats in the South Atlantic Bight
(Sedberry 1988). Nineteen small black sea bass (50 - 100 mm SL) fed
selectively on motile epifaunal amphipods, including caprellid species Caprella
equilibra, C. penantis and Phtisica marina. The two latter species were selected
as prey on deeper reefs (23 - 37 m) but not on inner shelf reefs (16-22 m). For
larger black sea bass (> 100 mm SL), decapods and fishes contributed the
greatest volume of prey. Wenner et al. (1983) demonstrated that live bottom
habitats have a highly diverse and abundant assemblage of invertebrates. They
found that annelids were the most numerically dominant (64.7% N), followed by
amphipods (17.4% N) in suction and grab samples collected on the southeast US
continental shelf. Caprellid amphipods were higher in abundance in Centropristis
striata stomachs than in benthic samples (Sedberry 1988).
Following other crustaceans (43.6% IRI), large calanoid copepods and
mysids comprised approximately 40% IRI of the diet of S. phoebe, followed by
caprellid (9.0% IRI) and gammarid (8.3% IRI) amphipods. Diet differed for six
specimens collected off of Florida, Mexico and South America, which consumed
over 90% N shrimp, with the remaining 10% consisting of crabs and bivalves
(Robins and Starck 1961). Fish were larger in that study, with standard lengths
ranging from 34 to 154 mm, and this may account for differences in prey choice.
Diplectrum formosum ate large benthic and epibenthic crustaceans,
including large calanoid copepods, shrimp, fish, caprellid amphipods, and crabs.
104
These findings were similar to previous studies. Seventeen D. formosum (SL
range 48- 196 mm) collected from the Gulf of Mexico consumed xanthid and
portunid crabs followed by shrimp, mysids and polychaetes (Bullock and Smith
1991). Bortone (1971) analyzed 154 specimens (SL range 21 - 223 mm) that
consumed shrimp (56.5% F), crabs (37.7% F), and amphipods (11.1% F), as well
as a variety offish (Perciformes, Scorpaeniformes, Pleuronectiformes, and
Tetradontiformes). Bortone (1971) and others (Wenner et al. 1983) found that D.
formosum has a diurnal pattern, with most activity taking place during daylight
hours. This may be why such a high percentage of specimens in this study were
empty (65%).
Serraniculus pumilio, Serranus baldwini, and the grammistid, Rypticus
maculatus, were also collected from station 43. The pygmy sea bass, S. pumilio,
is known to occur over sand or shell bottoms near irregularities, such as coral or
rock outcrops, but apparently does not have a restricted home range (Hastings
1973). In the Gulf of Mexico, S. pumilio is a territorial species that feeds
indiscriminately on small crustaceans, such as amphipods (42.2% N), isopods
(11.1% N), shrimp (14.2% N) and crabs (12.2% N) (Hastings 1973). The lantern
sea bass, S. baldwini, is a tropical species that is rarely found north of Palm
Beach, Florida (Robins and Starck 1961). In shallow water, S. baldwini prefers
Thalassia beds, whereas prefers shell and coral fragments in deeper waters.
Due to the small numbers of these species captured, a complete diet study is not
possible.
105
Gammarid amphipods were the most important prey category consumed
by R. maculatus, with an IRI of 87.5%. This species is highly cryptic, preferably
inhabits rocky- bottomed areas, and feeds primarily at night (Guimaráes 1999).
In the Gulf of Mexico, thirty-three R. maculatus with a mean fork length of 197.42
mm consumed xanthid crabs (65.9% IRI), followed by penaeidae shrimp (11.2%
IRI), Gobiidae (9.5% IRI) and pisces (9.5% IRI) (Nelson and Bortone 1996).
Bullock and Smith (1991) also reported that R. maculatus feeds heavily on
shrimps and crabs. Courtenay (1967) stated that R. maculatus has invaded
waters north of the tropical areas inhabited by other western Atlantic soapfishes
and has thus almost entirely eliminated competition from relatives. However,
Courtney does not mention the presence of similarly related Serranid species,
and possible interactions. This study represents the first description of feeding
ecology for this species on the southeast US continental shelf.
Stephanolepis hispidus, the planehead filefish, consumed cyclopoid
copepods, other crustaceans, and caprellid amphipods. Previous studies on
similar Monacanthid species have found different diet compositions, including the
preference for herbivorous food sources. For instance, three species of
Monacanthids (SL range 18-195 mm) from Posidonia australis seagrass habitats
in New South Wales consumed considerable amounts of seagrass and algae
(Bell et al. 1978). However, small specimens (< 6.5 cm) of Monacanthus
tomentosus collected from seagrass habitats in Indonesia consumed
polychaetes, crustaceans and molluscs, and contained no plant material in their
106
stomachs (Peristiwady and Geistdoerfer 1991). Juvenile specimens of
Monacanthus chinensis (SL range 18-63 mm) ate algae and sea grass (26.8%
and 26.5% V, respectively), but also consumed polychaetes (15.5% V) and
gastropods (9.7% V) (Bell et al. 1978). A similar study in Botany Bay, New South
Wales, discovered that juvenile M. chinensis (SL <100 mm) in Posidonia
habitats ate gammarid amphipods and seagrass (45.9% and 13.8% wet weight,
respectively) (Conacher et al. 1979).
Studies on S. hispidus from the southern Atlantic Ocean have been
limited to distribution and abundance studies in floating Sargassum communities
(Settle 1993; Stachowicz and Lindquist 1997), and age and growth analyses
(Rogers et al. 2001 ; Mancera-Rodriguez et al. 2004). Diet studies conducted on
the northeast continental shelf have found that adult S. hispidus (5-19 cm)
consumed polychaetes predominantly, followed by cephalopods and
echinoderms (Bowman et al. 2000). In the Gulf of Mexico, unidentified items
(56.9% IRI), Ectoprocta (21.1% IRl), xanthid crabs (8.6% IRI), and polychaetes
(6.0% IRI) were the most important prey consumed by S. hispidus (reported as
Monacanthus hispidus) with a mean fork length of 247.75 mm. Juvenile S.
hispidus (21 -40 mm SL) collected from Apalachee Bay, FL, fed primarily on
gammarid amphipods, only rarely taking copepods (Clements and Livingston
1983).
Observed differences in diet of species between this study and previous
studies may be due to a number of reasons. Temporal and spatial differences in
107
feeding preferences, prey availability and behavior, and morphological
differences may affect diet choices. For instance, Henwood et al. (1978) and
Bowman et al. (1987) concluded that most S. caprinus and S. chrysops feed
during the day. Stenotomus sp. for this study were collected from 2000 hours to
500 hours. If these fish were feeding during daylight hours, soft-bodied
organisms would be digested relatively quickly, and therefore less importance
would be placed on them as a dietary category. The high index of relative
importance for copepods may be an artifact of slow digestion time for organisms
with exoskeletons. For example, Sutela and Huusko (1997) found that
copepodid carapaces were identified in the guts of vendace {Coregonus albula)
after 72 hours. Generally, the greater the percentage of chitin or shell in a food
item, the longer it remains recognizable (Macdonald and Waiwood 1982).
Temporal and spatial differences in feeding habits may have caused diet
differences between other species as well. According to Ross (1986), when
marine fishes co-occur in the same general area, food and then habitat are the
resources that are most commonly partitioned among those species. Many
serranid species overlap both spatially and temporally during the juvenile stage.
In the Gulf of Mexico, C. ocyurus and S. phoebe were captured from 55 m and
greater in depth. In this study, the 80 serranids captured from stations 38 and 43
(45 meters in depth) were C. ocyurus and S. phoebe. Prey categories consumed
by these species were somewhat similar, but differed according to relative
importance. Temporal differences in feeding may have caused diet differences
108
between these two species, such that the other crustacean category in S.
phoebe comprised amphipods that were eaten earlier than those consumed by
C. ocyurus.
Additionally, even though these species were collected from the same
stations, segregation of these species at a smaller spatial scale may influence
their diet choices. For instance, Caine (1983, 1991 ) reports that fish recognize
certain caprellid habitats as feeding stations. Pinfish swam between visible
gorgonian corals in Thalassia seagrass beds, ignoring other possible feedings
sites or non-emergent corals. Subsequent diet analyses confirmed that the
caprellid amphipod, Caprella penantis, was the only member of the epibotic
community that was consumed (Caine 1983). Centropristis ocyurus and S.
phoebe may also recognize certain microhabitats as beneficial feeding habitats
over other areas, and therefore segregate themselves according to differences in
prey abundance and preference.
Vertical and horizontal differences in prey behavior can have drastic
effects on availability to predators. Diel vertical migration is the movement of
copepods and other zooplankton from greater depths during the day to shallower
depths at night (Lampert 1989; Hays et al. 1994). It is hypothesized to be an
adaptive response to visual predators, in that moving to greater depths during the
day decreases the chances of being seen and captured (Zaret and Suffern
1976). Vertical migration of small calanoid copepods from the depths into the
water column during the night may result in more of these organisms being eaten
109
by Stenotomus sp., since these fish tend to feed higher in the water column
compared to the other fish species in this study (Sedberry 1983). Copepods are
also able to form dense swarms that may attract schooling fishes and increase
the amount of these organisms eaten (Hamner 1979). For instance, 47 of 67
Stenotomus sp. were collected from two inshore stations - 2.1 and 35 (Figure 3-
1). These individuals consumed an average of 36 and 84 small calanoid
copepods, respectively. Prey availability at these stations may differ from other
areas.
Similarly, prey items may be available at different times. For example, the
demersal behavior of amphipod species may cause spatial disparities in the
number of gammarid amphipods on the sea floor (Alldredge and King 1985).
During the night, large demersal specimens (> 2 mm) moved up 1 to 2 meters off
the bottom, and were hypothesized to take advantage of stronger currents to
disperse them into more favorable feeding and breeding habitats.
Prey biodiversity on a horizontal scale can also affect feeding habits of
predators. Bradshaw et al. (2003) studied the effect of upright sessile epifauna,
specifically hydroid colonies, on biodiversity and community composition in the
Irish Sea. They found that hydroid colonies increase both the diversity and
abundance of benthic fauna, including amphipods and juvenile bivalves that
physically attach themselves to the hydroids, polychaetes, caprellid and
gammarid amphipods that live amongst the upright structure of the hydroids, and
mobile taxa that shelter at the base of the hydroid. The presence of hydroid
110
colonies, and other upright structures that attach to the hard-bottom areas of the
southeast US continental shelf, may provide habitat to many epibenthic species,
and therefore alter the diet of those fish that spend their time feeding in rocky-
reef and live-bottom habitats.
Morphological differences between predator species may also affect diet
choices. Morphological adaptations for feeding include alterations in body form,
mouth shape and position, the presence/absence of marginal and pharyngeal
teeth, gillrakers and the length of the alimentary canal (Wootton 1990). Serranus
phoebe was, on average, 10 mm smaller than C. ocyurus, and this size
difference may influence the prey eaten. Following other crustaceans, large
calanoid copepods were the next most important prey category for S. phoebe.
Calanoid copepods average no larger than 5 mm in length (Ruppert et al. 2004).
They are smaller than caprellid and gammarid amphipods, which range in size
from 5 to 15 mm in length (Ruppert et al. 2004). Stenotomus sp. and
Stephanolepis hispidus have smaller mouths than the other six species, and fed
on smaller prey items, such as calanoid and cyclopoid copepods.
Cumulative prey curves are recommended as an a priori method in
determining the adequacy of sample sizes in diet studies (Ferry and Cailliett
1997; Cortes 1997). However, to date there is no objective, quantitative method
to determine if the curve has reached an asymptote. In this study, an asymptotic
function fit with non-linear regression techniques was used to model the
asymptotic relationship of dietary data. An advantage of this is a quantitative.
111
objective method to assess if a curve has reached an asymptote. Conversely,
the data inputted into the model, as well as the initial estimates, will greatly affect
the outcome. In other words, validity of the model will depend greatly on the
maximum number of prey categories, and the rate at which new prey categories
are added as more fish are added to the sample. The maximum number of prey
categories is dependent upon the level of identification of the prey items. For
instance, if prey were grouped into less specific categories (crustaceans, fish,
etc), the maximum number of prey categories would be lower, and therefore
fewer specimens would need to be analyzed before an asymptotic relationship
occurred. As well, the rate of change in the number of prey categories depends
upon a number of factors, including spatial orientation, temporal behavior,
swarming, cryptic morphologies, and the habitats in which the predators and prey
exist. In this study, the estimated values were not independent of the original
data inputted into the asymptotic function; therefore these findings are
preliminary and need further investigation.
Cortes (1997) stresses the importance of standardizing methods of dietary
analyses. It must be stressed that the numbers of sufficient sample sizes for
each of the species in this study will most likely be different for a follow-up study
during a different time (daytime), and/or at different stations. However, in
agreement with Cortes (1997), to facilitate comparisons among studies, it is
advantageous to utilize the same methods of analyses.
112
Dietary Overlap
No biologically significant overlap values were found among the species
examined. This finding differs from overlap values between the same species in
different areas. For example, the highest resource overlap in an artificial-reef fish
community in the northern Gulf of Mexico (Nelson and Bortone 1996) occurred
between R. maculatus and C. ocyurus (Morisita’s modified index; 0.754).
Conversely, in this study C. ocyurus and R. maculatus had a low Schooner’s
dietary overlap value of 0.33. On the outer continental shelf of the Middle
Atlantic Bight, S. chrysops showed the greatest similarity to silver hake
{Merluccius bilinearis) and red hake {Urophycis chuss), due to the dominance of
small planktonic crustaceans in their diet (Sedberry 1983). According to
Sedberry (1989), southern porgy (S. aculeatus) has low overlap in diet with
whitebone porgy {Calamus leucosteus) on hard bottom habitats in the
southeastern US.
Several reasons may account for the lack of biologically significant overlap
exhibited by the fishes in this study. Optimal foraging theory predicts that as food
becomes scarce, predators will take a wide variety of food and similar predators
occupying the same habitat will converge in diet (Pyke et al. 1977; Sedberry
1983). Alternatively, it has been hypothesized that as food density decreases,
predators living in similar habitats specialize in eating certain prey items,
consequently decreasing dietary overlap. In this case, high diet overlap would be
113
expected if food were abundant. The results from this study suggest that juvenile
fishes are not selective in their feeding, in that each species ate a variety of prey
items. It is hypothesized that the abundance of microhabitats provided by rocky-
reefs and hard-bottom areas promote a high diversity of invertebrates, and
therefore an increase in prey for juvenile fishes. To determine if the lack of
overlap is due to food shortage or food abundance, additional information on
prey availability and microhabitat is necessary. Furthermore, diet overlap
relationships change considerably with season and fish size.
Both Schooner’s index and Bray-Curtis similarities used in multivariate
analyses provide insight into dietary overlap. Schooner’s index describes
overlap between species, while the Bray-Curtis similarities describe overlap
between individuals. Untransformed total specific volumes of each prey category
were used in the Schooner’s index, while double root transformed average
specific volumes were used for multivariate analyses. In this study, there was no
biologically significant overlap found by the Schooner’s index between any
species, yet 10 feeding guilds were delineated easily at the 40% level using
multivariate analyses. This finding suggests that analyzing dietary overlap
between each individual fish gives different results than overlap values computed
between species, supporting the occurrence of stage structured differences in
resource use.
114
Ontogenetic Diet Variations
Cumulative prey curves for each size class of Stenotomus sp. show that
fewer individuals are needed to describe dietary habits in the two middle size
classes compared to the smallest and largest size classes. This indicates that
the diet varied ontogenetically. Small calanoid copepods were the most
important prey category up until 41 mm standard length, when cyclopoid
copepods were the most important prey. With growth, there was a shift to
include larger prey items in the diet, including large calanoid copepods and
gammarid amphipods. This finding is well established in diet literature (Bowen
1983; Wootton 1990). Henwood et al. (1978) found the only differences in food
items of large and small S. caprinus individuals were in the size of prey. This
finding is apparent in this study, in that large calanoid copepods and gammarid
amphipods are included in the diet with growth. The highest volume of gammarid
amphipods occur in largest size category (>41 mm SL). This is similar to
Sedberry’s (1988) finding that S. chrysops from 51 to 100 mm SL consumed
hypehid and gammarid amphipods. Changes in morphology during growth allow
a fish to pursue, engulf and consume larger and/or faster prey items (Wootton
1990).
Cumulative prey curves for C. ocyurus show similar results to Stenotomus
sp. A sufficient number offish were analyzed in the two smaller size classes, but
115
not in the largest size class. This indicates that as C. ocyurus increase in size,
larger sample sizes are necessary to describe the diet accurately.
Amphipods are important prey items for all size classes of C. ocyurus. In
the smallest size class, gammarid amphipods had the highest average percent
specific volume, but with growth, caprellid amphipods become important. Similar
conclusions were found for C. striata, the black sea bass, captured in the same
area (Sedberry 1988). In Sedberry’s smallest size class (50 - 100 mm SL),
amphipods dominated in frequency, number and volume. Unfortunately,
amphipod identification to suborder (Caprellidea, Gammaridea, Hyperiidea) was
not reported for each size class. For all sizes, a caprellid amphipod {Caprella
equilibria) and a corophoid {Erichthonius brasiliensis) dominated in the diet of C.
striata (Sedberry 1988). With growth, C. ocyurus also eat a higher diversity of
prey compared to the smaller size classes, by including such items as clams,
crabs, fish, and cladoceara in their diet. The spotted sand bass, Paralabrax
maculatofasciatus, in Baja California, also had higher prey richness in the largest
size class (Ferry et al. 1997). Conflicting results were found in the dusky grouper
{Epinephelus marginatus) in the western Mediterranean. With growth, groupers
satisfied their nutritional requirements through the capture of large prey, and not
by the ingestion of a greater number of prey items (Linde et al. 2004).
Cumulative prey curves for S. phoebe again show a similar relationship as
found in Stenotomus sp., and C. ocyurus. In the smallest size class, twelve
individuals are adequate to describe the diet; however this increases to fourteen
116
individuals in the largest size class. Again, this indicates that as S. phoebe
increase in size, larger sample sizes are necessary to describe the diet.
Serranus phoebe also exhibits diet shifts with growth. Prey category
diversity increases from eight to twelve prey categories from the small to large
size range, and S. phoebe begin to include larger prey items in their diet, such as
fish, polychaetes, and clams. In the largest prey category, mysids are the most
volumetrically important prey category after other crustaceans. Bullock and
Smith (1991) analyzed only four specimens of S. phoebe and found crabs and
shrimps in the stomachs. One larger specimen (137 mm) contained two codiets.
Serranus phoebe > 20 mm also ate fish prey in this study.
Studies of ontogenetic changes in other Serranid species have found
similar results. Juvenile dusky groupers feed predominantly on small
crustaceans, such as amphipods, crabs and isopods (Linde et al. 2004).
Serranus subligarius (belted sandfish) less than 40 mm SL from the Gulf of
Mexico ate gammarid amphipods (27.4% W) and natantia decapods (22.2% W),
but with growth (> 40 mm SL) consumed reptantia (29.4% W) and natantia
(19.3% W) decapods, and fish (19.6% W) (Hastings and Bortone 1980). Ferry
et al. (1997) suggested that smaller spotted sand bass are less capable of
catching more mobile prey, such as mysids.
As demonstrated, Stenotomus sp., C. ocyurus, and S. phoebe go through
ontogenetic changes in diet during the juvenile life-history stage. Size of prey
and number of prey categories increase with growth for all three species.
117
Morphological differences between size classes may explain these shifts. In
other words, as fish grow they are more capable of capturing and ingesting larger
prey items. However, larger predators consistently include small-bodies prey in
their diet, possibly representing profitable foraging behaviors when size-
dependent probabilities of encounter and capture are combined with handling
costs of prey (Scharf et al. 2000).
Dietary overlap among size classes within each species was biologically
significant for most pairs. This was expected since intraspecific interactions are
often more intense by many orders of magnitude than are interspecific
interactions (Schroder and Rosenzweig 1975). It has been argued that
individuals within the same population are likely to be most similar in their
resource requirements and so are potentially intense competitors (Wootton
1990). As a result, growth of smaller individuals may be suppressed by the
presence of larger individuals (Dou et al. 2004). Superimposition of designated
size classes on the ordination plot for each species showed considerable overlap
for all sizes of Stenotomus sp., C. ocyurus, and S. phoebe. Different values
were used for each method, therefore it is difficult to compare overlap values.
To determine if these shifts in food preferences have effects on
community structure and dynamics, further examination of ontogenetic changes
in diet across temporal and spatial axes are necessary to fully understand
juvenile resource use on the southeast US continental shelf.
118
Trophic Guild Analyses
The term guild has been used in many terrestrial (eg. Pianka 1980),
freshwater (eg. Lappalainen and Kjellman 1998), and marine ecosystem studies
(eg. Garrison and Link 2000) since it was first used by Root (1967) to define
feeding guilds of blue-gray gnatcatchers. Root (1967) defined a guild as a group
of species that exploit the same class of environmental resources in the same
way. He further described the term as not limited by taxonomic boundaries (Root
1986). Guilds can be based on several factors or combinations of factors,
including diet, morphology, and behavior (Nelson and Bortone 1996). In this
study, guilds were delineated using dietary data, regardless of morphology or
behavior of the fish species. The ten dietary guilds reflect similarity in the
utilization of specific prey categories. Within each guild, only one prey category
usually accounted for > 50% offish diets, with 1 to 9 prey categories contributing
< 50 % of the diet. On average, larger fish in guilds 1A and IB ate larger, benthic
prey items (mysids and gammarid amphipods), compared to smaller fish eating
smaller epibenthic and planktivorous prey (i.e. caprellid amphipods and
copepods) in the remaining 8 guilds. However, within species there was no
difference in size between guilds, with S. hispidus as an exception.
Trophic analyses from the southeast US continental shelf are lacking for
juvenile fishes. Nelson and Bortone (1996) analyzed the trophic structure among
artificial reef fishes in the northern Gulf of Mexico, and included C. ocyurus, R.
119
maculatus and S. hispidus (reported as Monacanthus hispidus), as well as 22
other species, in their study. Using a Bray-Curtis cluster analysis dendrogram,
C. ocyurus (mean fork length 172 mm) and R. maculatus (mean fork length 197
mm) were grouped in the lower structure crustacean predator guild, because
xanthid crabs had the highest index of relative importance for both species.
Unfortunately, M. hispidus (mean fork length 248 mm) was not included in the
cluster analysis, since < 5 fish were analyzed. Comparatively, this study focused
on fish of smaller sizes, and clustered individuals rather than species, yet found
Centropristis ocyurus and Rypticus maculatus feed on larger, benthic
crustaceans.
For the purposes of this study, guilds were defined based on food
resources, which have been found to be important characteristics in structuring
communities (Scharf et al. 2000). If a guild is defined based on a resource or
characteristic that is relatively unimportant in structuring the community, a
prediction of community change based on that guild system will likely have little
relevance to community dynamics (Austen et al. 1994). However, choosing what
level of resource is ecologically relevant is a subjective process. For instance, 10
feeding guilds with a ^0% similarity were designated in this study; yet, these
guilds may not describe an organizing characteristic within the community. The
measure of dietary similarity using the Bray-Curtis coefficient provided a resource
matrix in which species with similar values were placed in the same guilds.
Pianka (1980) suggested that guilds were areas of “intense interspecific
120
competition with strong interactions within guilds, but weak interactions between
members of different guilds.” Nonetheless, the degree of competition among
species in this study cannot be interpreted from the resource matrix alone. As
discussed previously, additional information on prey availability and microhabitat
is necessary to explain competitive interactions. As well, to test the generality of
guild definitions is to study the changes in guild member biomass or abundance
through time or in various systems (Orth 1980).
Specifically, monitoring juvenile fish distribution and abundance on the
southeast US continental shelf in multiple microhabitats through time will help
clarify competitive interactions. In addition, laboratory studies can help
determine possible types of competition. Interactive segregation occurs when the
use of resources differ when species are sympatric although are similar when
they are allopatric. This may be caused by interference competition, where
access to a resource is reduced when the behavior of the competitor interferes
with the ability of and individual to acquire that resource (Wootton 1990). For
example, field observations and laboratory experiments have suggested that
interactive segregation plays a part in the coexistence of juvenile salmonids in
British Columbia, Canada (Hartman 1965).
Austen et al. (1994) reviewed the importance of the guild concept in
fisheries management and indicated guild management would be effective if
statistically delineated guilds based on key resources were used. They also
suggest the development of guilds that function as “super species” - units that
121
respond to environmental change (including fishing pressure) in a more
predictable manner than any individual member species. Information about the
functioning of the fish community can be obtained at a higher level of
organization by developing guilds based on similar responses of guild members
to environmental change (Skagen et al. 1991). Consequently, the use of guilds
to characterize the condition of specific ecosystems meets the requirements of
ecosystem-based management and the conservation of biodiversity on the
southeast US continental shelf.
122
Conclusions
The eight juvenile fish species examined in this study appear to be
opportunistic feeders capable of utilizing a wide variety of prey items.
Stenotomus sp. and Stephanolepis hispidus feed predominantly on small,
planktonic copepods from the families Calanoida and Cyclopoida. The five
Serranid species and one Grammistid feed largely on gammarid and caprellid
amphipods, with a preference for larger-sized prey items. Prey selection may be
explained by habitat choices. It may be that Stenotomus sp. and S. hispidus
spend more time higher in the water column, while the six other species occupy
habitats on the benthos. Although diets and presumed feeding habitat differed,
all fish were caught on the bottom. Differences in vertical microhabitat use may
contribute to differences in diet. Conversely, morphological differences may
influence the size of prey consumed. Fish species with larger mouths are able to
ingest larger prey. This increase in prey size was suggested with the analysis of
ontogenetic size classes in Stenotomus sp., C. ocyurus, and S. phoebe. Prey
availability is an important component to diet studies. Extensive invertebrate
collections from microhabitats were not made in conjunction with fish samples,
therefore prey availability data were lacking in this study.
Cumulative prey curves are necessary to determine the adequacy of
sample sizes in any diet study. The number of individuals needed for this study
ranged from 8 to 55, however these computations will depend upon the level of
123
identification of the prey, as well as temporal and spatial characteristics of diet
samples. It is recommended that an objective, quantitative method be developed
to assess the asymptotic relationship of cumulative prey curves. In this study the
an asymptotic function fit with non-linear regression techniques was used,
providing an objective, quantitative method to determine an asymptotic
relationship in cumulative prey curves.
Ten feeding guilds were established within the juvenile fish community on
the southeast US continental shelf based upon habitat and prey size. Three
guilds made up the larger, benthic predators, while six guilds made up the small
epibenthic and planktonic predators. It is important to stress that these guilds will
change due to many factors, including differences in when and where predators
are eating, and temporal and spatial differences in prey availability. As well,
other predator species not included in this study may add other guilds and/or
increase the number of individuals within each guild. Identifying trophic guilds is
a useful first step for defining groups of functionally similar species, and is a
potentially valuable tool to simplify interspecies interactions on the southeast US
continental shelf.
To successfully implement an ecosystem-based management approach, a
number of tasks are recommended (NOAA 1999; Busch et al. 2003) including
descriptions of habitat needs of different life-history stages to delineate
geographic areas of an ecosystem, and biological indicators such as trophic
structure and species diversity to characterize the condition of specific
124
ecosystems. The results from this study provide dietary data on fish species that
have not previously been studied, and present the first data on trophic guild
organization within the juvenile fish community on the southeast US continental
shelf. Specifically, the results of this study show that both pelagic and benthic
prey items are important food resources of the juvenile fish community, and
therefore provide important bentho-pelagic links in this ecosystem. As well,
size-structured populations do exist in the juvenile fish community on the
southeast US continental shelf. These findings are important for the classification
of essential fish habitats in this area, are necessary for an ecosystem-based
approach to fisheries management, and the conservation of marine biodiversity
in the vicinity of Gray’s Reef National Marine Sanctuary and on the southeast US
continental shelf.
SI O'W SO'y'W
I
Figure 3-1 Map of study area and the individual sample sites and station numbers used for sampling. Four
stations were located around Gray’s Reef National Marine Sanctuary. 125
REFERENCES
Alldredge, AL and King JM. 1985. The distance demersal zooplankton migrate
above the benthos: implications for predation. Marine Biology 84:253-260
Austen, DJ; Bayley, PB; Menzel, BW. 1994. Importance of the guild concept to
fisheries research and management. Fisheries 19(6): 12-20
Bacheler, NM; Neal, JW; Noble, RL. 2004. Diet overlap between native bigmouth
sleepers {Gobiomorus dormitory) and introduced predatory fishes in a
Puerto Rico reservoir. Ecology of Freshwater Fish 13:111-118
Baldo, F and Drake, P. 2002. A multivariate approach to the feeding habits of
small fish in the Guadalquivir Estuary. Journal of Fish Biology 61 (Suppl.
A):21-32
Beck, MW; Heck, KL Jr; Able, KW; Childers, DL; Eggleston, DB; Gillanders, BM;
Halpern, B; Hays, CG; Hoshino, K; Minello, TJ; Orth, RJ; Sheridan, PF;
Weinstein, MP. 2001. The identification, conservation, and management
of estuarine and marine nurseries for fish and invertebrates. BioScience
51(8):633-641
Bell, JD; Burchmore, JJ; Pollard, DA. 1978. Feeding ecology of three sympatric
species of leatherjackets (Pisces: Monacanthidae) from a Posidonia
seagrass habitat in New South Wales. Australian Journal of Marine and
Freshwater Research 29:631-643
Bellwood, DR. 1988. Ontogenetic changes in the diet of early post-settlement
Scarus species (Pisces: Scaridae). Journal of Fish Biology 33:213-219
Berry, FH and Vogele, LE. 1959. Filefishes (Monacanthidae) of the western
North Atlantic Fish and Wildlife Service Fishery Bulletin US 181:61-109
Bethea, DM; Buckel, JA; Carlson, JK. 2004. Foraging ecology of the early life
stages of four sympatric shark species. Marine Ecology Progress Series
268:245-264
Bigelow, HB and Schroeder, WC. 1953. Fishes of the Gulf of Maine. North
Atlantic Fish and Wildlife Service Fishery Bulletin US 53. 577 p
Bortone, SA. 1971. Studies on the biology of the sand perch. Diplectrum
formosum (Perciformes: Serranidae). Florida Department of Natural
Resources Marine Research Laboratory Technical Series 65:1-27
127
Bowen, SH. 1983. Quantitative description of the diet. In Fisheries Techniques.
Edited by LA Nielsen and DL Johnson. The American Fisheries Society,
Bethesda, Maryland. Pp 325-336
Bowman, RE; Azarovitz, TR; Howard, ES; Hayden, BP. 1987. Food and
distribution of juveniles of seventeen northwest Atlantic fish species, 1973-
1976. National Oceanic and Atmospheric Administration Technical
Memorandum NMFS-F/NEC-45. 57p
Bowman, RE; Stillwell, CE; Michaels, WL; Grosslein, MD. 2000. Food of
Northwest Atlantic fishes and two common species of squid. United States
Department of Commerce, National Oceanic and Atmospheric
Association, NMFS. 138p
Bradshaw, C; Collins, P; Brand, AR. 2003. To what extent does upright sessile
epifauna affect benthic biodiversity and community composition? Marine
Biology 143:783-791
Brodeur, RD and Pearcy, WG. 1984. Food habits and dietary overlap of some
shelf rockfishes (genus Sebastes) from the northeastern Pacific ocean.
Fishery Bulletin 82(2):269-293
Browman, HI and Stergiou, K. 2004. Perspectives on ecosystem-based
approaches to the management of marine resources. Marine Ecology
Progress Series 274:269-303
Buckel, JA and McKown, KA. 2002. Competition between juvenile striped bass
and bluefish: resource partitioning and growth rate. Marine Ecology
Progress Series 234:191-204
Bullock, LH and Smith, GB. 1991. Seabasses (Pisces: Serranidae). Memoirs of
the Hourglass Cruises. Florida Marine Research Institute. Department of
Natural Resources, St. Petersburg, FL (USA) 8(2):243 p
Busch, WDN; Brown, BL; Mayer, GF. 2003. Strategic guidance for implementing
an ecosystem-based approach to fisheries management. United States
Department of Commerce, National Oceanic and Atmospheric
Administration, NMFS, Silver Spring, MD 62p
Cailliet, GM and Barry, JP. 1979. Comparison of food array overlap measures
useful in fish feeding habit analysis. In Fish Food Habits Studies:
Proceedings of the 2nd Pacific Northwest Technical Workshop. Edited by
SJ Lipovsky and CA Simenstad. Washington Sea Grant publication.
128
Washington Sea Grant Program, University of Washington, Seattle, Wash.
PP67-79.
Cailliet, GM; Love, MS; Ebeling, AW. 1986. Fishes: a field and laboratory manual
on their structure, identification and natural history. Wadsworth Publishing
Co., Belmont, California
Caine, EA. 1983. Community interactions of Caprella penantis Leach (Crustacea:
Amphipoda) on sea whips. Journal of Crustacean Biology 3(4):497-504
Caine, EA. 1989. Caprellid amphipod behavior and predatory strikes by fish.
Journal of Experimental Marine Biology and Ecology 126:173-180
Caine, EA. 1991. Caprellid amphipods: fast food for the reproductively active.
Journal of Experimental Marine Biology and Ecology 148:27-33
Caine, EA. 1998. First case of caprellid amphipod-hydrozoan mutualism. Journal
of Crustacean Biology 18(2):317-320
Clarke, KR and Warwick, RM. 2001. Change in marine communities: an
approach to statistical analysis and interpretation. 2^^ edition. PRIMER-E:
Plymouth
Clements, WH and Livingston, RJ. 1983. Overlap and pollution-induced
variability in the feeding habits of filefish (Pisces: Monacanthidae) from
Apalachee Bay, Florida. Copeia 2:331-338
Conacher, MJ; Lanzing, WJR; Larkum, AWD. 1979. Ecology of Bontany Bay. II.
Aspects of the feeding ecology of the fanbellied leatherjacket.
Monacanthus chinensis (Pisces: Monacanthidae), in Posidonia australia
seagrass beds in Quibray Bay, Botany Bay, New South Wales. Australian
Journal of Marine and Freshwater Research 27:117-127
Cortes, E. 1997. A critical review of methods of studying fish feeding based on
analysis of stomach contents: application to elasmobranch fishes.
Canadian Journal of Fisheries and Aquatic Sciences 54:726-738
Courtenay, WR Jr. 1967. Atlantic fishes of the genus Rypticus (Grammistidae).
Proceedings of the Academy of Natural Sciences of Philadelphia
119(6):241-293
Crow, M. 1979. A technique of guild analysis. In Gutshop '78. Fish food habits
studies. Proceedings of the second Pacific Northwest Technical
Workshop, held Maple Valley, WA (USA), 10-13 October, 1978.
129
Published by: Washington Sea Grant; Seattle, WA (USA)., Mar 1979., p.
97-98
Dahlgren, CP and Eggleston, DB. 2001. Spatio-temporal variability in
abundance, size and microhabitat associations of early juvenile Nassau
grouper Epinephelus striatus in an off-reef nursery system. Marine
Ecology Progress Series 217:145-156
Dou, SZ; Masuda, R; Tanaka, M; Tsukamoto, K. 2004. Size hierarchies affecting
the social interactions and growth of juvenile Japanese flounder,
Paralichthys olivaceus. Aquaculture 233:237-249
Ferry, LA and Cailliet, GM. 1996. Sample size and diet analysis: are we
characterizing and comparing diet properly? In Feeding ecology and
nutrition in fish: proceedings of the symposium on the feeding ecology and
nutrition in fish. International Congress on the Biology of Fishes, San
Francisco, California. 14-18 July 1996. Edited by D MacKinlay and K
Shearer. American Fisheries Society, San Francisco, Calif. Pp. 71-80
Ferry, LA; Clark, SL; Cailliet, GM. 1997. Food habits of spotted sand bass
{Paralabrax maculatofasciatus, Serranidae) from Bahia De Los Angeles,
Baja California. Bulletin of Southern California Academy of Science
96(1):1-21
Federal Register (FR). 1981. - Final Designation Document - Designation of the
Gray’s Reef National Marine Sanctuary. 46 FR 7943-7944
Garrison, LP and Link, JS. 2000. Dietary guild structure of the fish community in
the northeast United States continental shelf ecosystem. Marine Ecology
Progress Series 202:231-240
Geoghegan, P; Chittenden, ME Jr. 1982. Reproduction, movements, and
population dynamics of the longspine porgy, Stenotomus caprinus.
Fishery Bulletin 80(3):523-540
Gerking, SD. 1994. Feeding ecology offish. Aoademic Press, San Diego. 414
P
Gelsleichter, J; Muscik, JA; Nichols S. 1999. Food habits of the smooth dogfish,
Mustelus canis, dusky shark, Carcharhinus obscuras, Atlantic sharpnose
shark, Rhizoprionodon terraenovae, and the sand tiger. Carcharías taurus,
from the northwest Atlantic Ocean. Environmental Biology of Fishes
54:205-217
130
Gray’s Reef National Marine Sanctuary (GRNMS). 2005. About Gray’s Reef
National Marine Sanctuary, www.qravsreef.nos.noaa.qov
Guimaráes, RZP. 1999. Revision, phylogeny and comments on biogeography of
soapfishes of the genus Rypticus (Teleostei: Serranidae). Bulletin of
Marine Science 65(2):337-379
Hamner, WM and Carleton, JH. 1979. Copepod swarms: attributes and role in
coral reef ecosystems. Limnology and Oceanography 24(1); 1-14
Hare, JA; Walsh, HJ; Marancik, KE. in review. Fish fauna of Gray’s Reef National
Marine Sanctuary and the implications for place-based management.
Hartman, G. 1965. The role of behaviour in the ecology and interaction of
underyearling coho salmon {Oncorhynchus kisutch) and steelhead trout
{Salmo gairdneri). Journal of the Fisheries Research Board of Canada
22:1035-1081
Hastings, RW. 1973. Biology of the pygmy sea bass, Serraniculus pumilio
(Pisces: Serranidae). Fishery Bulletin 71(1): 235-242
Hastings, PA and Bortone, SA. 1980. Observations on the life history of the
belted sandfish, Serranas subligarius (Serranidae). Environmental Biology
of Fishes 5(4):365-374
Hays, GO; Proctor, CA; John, AWG; Warner, AJ. 1994. Interspecific differences
in the diel vertical migration of marine copepods: the implications of size,
color, and morphology. Limnology and Oceanography 39(7):1621-1629
Hellewell, JM and Abel, R. 1971. A rapid volumetric method for the analysis of
the food of fishes. Journal of Fish Biology 3:29-37
Henwood, T; Johnson, P; Heard, R. 1978. Feeding habits and food of the
longspined porgy, Stenotomus caprinas Bean. Northeast Gulf Science
2(2): 133-137
Holbrook, SJ and Schmitt, RJ. 1992. Causes and consequences of dietary
specialization in surfperches: patch choice and intraspecific competition.
Ecology 73(2):402-412
Horn, HS. 1966. Measurement of “overlap” in comparative ecological studies.
The American Naturalist 100(914):419-424
131
Humann, P and Deloach, N. 2002. Reef fish identification: Florida, Caribbean,
Bahamas. New World Publications, 3'"'^ edition. 512p
Hurlbert, SH. 1978. The measurement of niche overlap and some relatives.
Ecology 59(1):67-77
Hyslop, EJ. 1980. Stomach contents analysis - a review of methods and their
application. Journal of Fish Biology 17:411-429
Lamport, W. 1989. The adaptive significance of diel vertical migration of
zooplankton. Functional Ecology 3(1 ):21-27
Levins, R. 1968. Evolution in changing environments: some theoretical
explorations. Princeton University Press, Princeton, New Jersey. 120p
Linde, M; Grau, AM; Riera, F; Massutí-Pascual, E. 2004. Analysis of trophic
ontogeny in Epinephelus marginatus (Serranidae). Cybium 28(1):27-35
Lindeman, KC; Pugliese, R; Waugh, GT; Ault, JS. 2000. Developmental patterns
within a multispecies reef fishery: management applications for essential
fish habitats and protected areas. Bulletin of Marine Science 66(3):929-
956
Lindquist, DG; Clavijo, IE; Gaboon, LB; Bolden, SK; Burk, SW. 1989. Quantitative
diver visual surveys of innershelf natural and artificial reefs in Onslow Bay,
NC: preliminary results for 1988 and 1989. In Diving for Science. Edited by
M. Lang and W Jaap. American Academy of Underwater Scientists, St.
Petersburg, Florida, pp 219-227
Lindquist, DG; Gaboon, LB; Glavijo, IE; Posey, MH; Bolden, SK; Pike, LA; Burk,
SW; Gardullo, PA. 1994. Reef fish stomach contents and prey abundance
on reef and sand substrata associated with adjacent artificial and natural
reefs in Onslow Bay, North Garolina. Bulletin of Marine Science 55(2-
3):308-318
Link, JS. 2002. What does ecosystem-based fisheries management mean?
Fisheries 27(4):18-21
Linke, TE; Platell, ME; Potter, 1C. 2001. Factors influencing the partitioning of
food resources among six fish species in a large embayment with
juxtaposing bare sand and seagrass habitats. Journal of Experimental
Marine Biology and Ecology 266:193-217
132
Lowe, RH. 1962. The fishes of the British Guiana continental shelf, Atlantic coast
of South America, with notes on their natural history. The Journal of the
Linnean Society 44(301 ):669-700
Luczkovich, JJ; Ward, GP; Johnson, JC; Christian, DB; Neckles, H; Rizzo, WM.
2002. Determining the trophic guilds of fishes and macroinvertebrates in a
seagrass food web. Estuaries 25(6A):1143-1163
Macdonald, JS and Green, RH. 1983. Redundancy of variables used to describe
importance of prey species in fish diets. Canadian Journal of Fisheries
and Aquatic Sciences 40:635-637
Macdonald, JS; Waiwood, KG; Green, RH. 1982. Rates of digestion of different
prey in Atlantic cod {Gadus morhua), ocean pout {Macrozoarces
americanus), winter flounder {Pseudopleuronectes americanus), and
American plaice {Hippoglossoides platessoides). Canadian Journal of
Fisheries and Aquatic Sciences 39(5):651-659
Mancera-Rodriguez, NJ and Castro-Hernández, JJ. 2004. Age and growth of
Stephanolepis hispidus (Linnaeus, 1766) (Pisces: Moncanthidae), in the
Canary Islands area. Fisheries Research 66:381-386
Manly, BFJ. 2005. Multivariate statistical methods: a primer. Chapman and
Hall/CRC, 3^'^ edition. 214p
Marancik, KE and Hare, JA. 2005. An annotated bibliography of diet studies on
fish of the southeast United States and Gray’s Reef National Marine
Sanctuary. Marine Sanctuaries Conservation Series MDS-05-01. US
Department of Commerce, National Oceanic and Atmospheric
Administration, Marine Sanctuaries Division, Silver Spring, MD. 56p
Marancik, KE; Clough, LM; Hare, JA. 2005. Cross-shelf and seasonal variation in
larval fish assemblages on the southeast United States continental shelf
off the coast of Georgia. Fishery Bulletin 103:108-129
Michelman, MS. 1988. The biology of juvenile scup {Stenotomus chrysops (L.)) in
Narragansett Bay, Rl: food habits, metabolic rate and growth rate. M.S.
thesis, Univ. of Rhode Island, Kingston, Rl. 106 p.
Mittelbach, GG and Osenberg, CW. 1994. Using foraging theory to study trophic
interactions. In Theory and Application in Fish Feeding Ecology, edited by
DJ Stouder, KL Fresh, and RJ Feller. Pp 45-59
133
Morisita, M. 1959. Measuring of interspecific association and sinnilarity between
communities. Memoirs of the Faculty of Science. Kyushu University.
Series 3:64-80
Nelson, BD and Bortone, SA. 1996. Feeding guilds among artificial-reef fishes in
the northern Gulf of Mexico. Gulf of Mexico Science 2:66-80
National Marine Fisheries Service (NMFS). 1999. Essential fish habitat: new
marine fish habitat conservation mandate for federal agencies. National
Marine Fisheries Service, Habitat Conservation Division, Southeast
Regional Office, 9721 Executive Center Drive North, St. Petersburg, FL
33702
NOAA 1999. Ecosystem-based fishery management. A report to Congress by the
Ecosystem Principles Advisory Panel. U.S. Department of Commerce,
National Oceanic and Atmospheric Administration/National Marine
Fisheries Service, 54p
National Research Council (NRC). 2001. Marine Protected Areas: Tools for
Sustaining Ocean Ecosystems. Report of the National Academy Press,
Washington, DC. 272p
Osenberg, CW; Mittelbach, GG; Wainwright, PC. 1992. Two-stage life histories in
fish: the interaction between juvenile competition and adult performance.
Ecology 73(1 ):255-267
Ouzts, AC and Szedimayer, ST. 2003. Diel feeding patterns of red snapper on
artificial reefs in the north-central Gulf of Mexico. Transactions of the
American Fisheries Society 132:1186-1193
Parker, RO and Ross, SW. 1986. Observing reef fishes from submersibles off
North Carolina. Northeast Gulf Science 8(1):31-49
Parker, RO; Colby, DR; Willis, TD. 1983. Estimated amount of reef habitat on a
portion of the US South Atlantic and Gulf of Mexico continental shelf.
Bulletin of Marine Science 33:935-940
Parker, ROP; Chester, AJ; Nelson, RS. 1994. A video transect method for
estimating reef fish abundance, composition, and habitat utilization at
Gray’s Reef National Marine Sanctuary, Georgia. Fishery Bulletin 92:787-
799
134
Pehstiwady, T and Geistdoerfer, P. 1991. Biological aspects oi Monacanthus
tomentosus (Monacanthidae) in the seagrass beds of Kotania Bay, West
Seram, Moluccas, Indonesia. Marine Biology 109:135-139
Persson, L. 1983a. Food consumption and competition between age classes in a
perch Perea fluviatilis population in a shallow eutrophic lake. Oikos
40:197-207
Persson, L. 1983b. Food consumption and the significance of detritus and algae
to intraspecific competition in roach Rutilus rutilus in a shallow eutrophic
lake. Oikos 41:118-125
Persson, L; Bystrom P; Wahlstrom, E; Nijiunsing, A; Rosema, S. 2000. Resource
limitation during early ontogeny: constraints induced by growth capacity in
larval and juvenile fish. Oecologia 122:459-469
Pianka, ER. 1980. Guild structure in desert lizards. Oikos 35:194-201
Piet, GJ; Pet, JS; Guruge, WAHP; Vijverberg, J; Van Densen, WLT. 1999.
Resource partitioning along three niche dimensions in a size-structured
tropical fish assemblage. Canadian Journal of Fisheries and Aquatic
Sciences 56:1241-1254
Pinkas, L; Oliphant, MS; Iverson, ILK. 1971. Food habits of albacore, bluefin
tuna, and bonito in California waters. California Department of Fish and
Game. Fishery Bulletin 152. 105p
Platell, ME and Potter, IE. 1999. Partitioning of habitat and prey by abundant and
similar-sized species of the Triglidae and Pempherididae (Teleostei) in
coastal waters. Estuarine, Coastal and Shelf Science 48:235-252
Powell, AB; Greene, MD. 2002. Preliminary guide to the identification of the
early life history stages of sparid fishes of the western Central North
Atlantic. Miami # 480. 20p.
Pyke, GH; Pulliam, HR; Charnov, EL. 1977. Optimal foraging: a selective review
of theory and tests. Quarterly Review of Biology 52:137-154
Randall, JE; Aida, K; Hibiya, T; Mitsuura, N; Kamiya, H. 1971. Grammistin, the
skin toxin of soapfishes, and its significance in the classification of the
Grammistidae. Publications of the Seto Marine Biological Laboratory.
19(2/3):157-190
135
Robins, CR and Starck, WA. 1961. Materials for a revision of Serranas and
related fish genera. Proceedings of the Academy of Natural Sciences of
Philadelphia 113(11):259-314
Rogers, JS; Hare, JA; Lindquist, DG. 2001. Otolith record of age, growth, and
ontogeny in larval and pelagic juvenile Stephanolepis hispidas (Pisces:
Monacanthidae). Marine Biology 138:945-953
Root, RB. 1967. The niche exploitation pattern of the blue-gray gnatcatcher.
Ecological Monographs 37(4):317-350
Ross, ST. 1986. Resource partitioning in fish assemblages: a review of field
studies. Copeia 1968:352-388
Ruppert, EE; Fox, RS; Barnes, RD. 2004. Invertebrate Zology: A Functional
Evolutionary Approach. Brooks Cole, 7‘^ edition, 1008p
South Atlantic Fisheries Management Council (SAFMC). 1998. Habitat plan for
the South Atlantic region: essential fish habitat requirements for fishery
management plans of the South Atlantic Fishery Management Council.
South Atlantic Fishery Management Council. Charleston, South Carolina.
449p
Scharf, FS; Juanes, F; Rountree, RA. 2000. Predator size - prey size
relationships of marine fish predators: interspecific variation and effects of
ontogeny and body size on trophic-niche breadth. Marine Ecology
Progress Series 208:229-248
Schoener, TW. 1970. Nonsynchronous spatial overlap of lizards in patchy
habitats. Ecology 51 (3):408-418
Southeast Area Monitoring and Assessment Program - South Atlantic
(SEAMAP-SA). 2001. Distribution of bottom habitats on the continental
shelf from North Carolina through the Florida Keys. SEAMAP-SA Bottom
Mapping Workgroup, Atlantic States Marine Fisheries Commission,
Washington, DC. 166p
Sedberry, GR. 1983. Food habits and trophic relationships of a community of
fishes on the outer continental shelf. NOAA National Fisheries Service
Technical Report United States Department of Commerce National
Oceanic and Atmospheric Association, National Marine Fisheries Service
773:1-56
136
Sedberry, GR. 1988. Food and feeding of black sea bass, Centropristis striata, in
live bottom habitats in the South Atlantic Bight. The Journal of the Elisha
Mitchell Scientific Society 104(2):35-50
Sedberry, GR. 1989. Feeding habits of whitebone porgy. Calamus leucosteus
(Teleostei: Sparidae), associated with hard bottom reefs off the
southeastern United States. Fishery Bulletin 87:935-944
Sedberry, GR and Van Dolah, RF. 1984. Demersal fish assemblages associated
with hard bottom habitat in the South Atlantic Bight of the USA.
Environmental Biology of Fishes 11(4):241-258
Sedberry, GR; McGovern, JC; Barans, CA. 1998. A comparison offish
populations in Gray’s Reef National Marine Sanctuary to similar habitats
off the southeastern US: implications for reef fish and sanctuary
management. Proceedings of the 50“^ Gulf and Caribbean Fisheries
Institute Pp 452-481
Settle, L. 1993. Spatial and temporal variability in the distribution and abundance
of larval and juvenile fishes associated with pelagic Sargassum. MSc
thesis. University of North Carolina, Wilmington
Sheridan, PF and Trimm, DL. 1983. Summer foods of Texas coastal fishes
relative to age and habitat. Fishery Bulletin 81(3):643-647
Simberloff, D and Dayan, T. 1991. The guild concept and the structure of
ecological communities. Annual Review of Ecological Systems 22:115-
143
Stachowicz, JJ and Lindquist, N. 1997. Chemical defense among hydroids on
pelagic Sargassum: predator deterrence and absorption of solar UV
radiation by secondary metabolites. Marine Ecology Progress Series
155:115-126.
Stoner, AW. 1980. Feeding ecology of Lagodon rhomboids (Pisces: Sparidae):
variation and functional responses. Fishery Bulletin 78(2):337-352
Stoner, AW and Livingston, RJ. 1984. Ontogenetic patterns in diet and feeding
morphology in sympatric sparid fishes from seagrass meadows. Copeia
1:174-187
Sutela, T and Huusko, A. 1997. Varying resistance of zooplankton prey to
digestion: implications for quantifying larval fish diets. Transactions of the
American Fisheries Society 129:545-551
137
Szedimayer, ST and Lee, JD. 2004. Diet shifts of juvenile red snapper {Lutjanus
campechanus) with changes in habitat and fish size. Fishery Bulletin
102:366-375
Wallace, RK. 1981. An assessment of diet-overlap indexes. Transactions of the
American Fisheries Society 110:72-76
Walsh, HJ: Marancik, KE; Flare, JA. (submitted). Juvenile fish assemblages of
unconsolidated sediments on the southeast United States continental
shelf off the coast of Georgia, USA. Fishery Bulletin
Wenner, CA. 1983. Species associations and day-night variability of trawl-
caught fishes from the inshore sponge-coral habitat. South Atlantic Bight.
Fishery Bulletin 81:537-552
Wenner, EL; Knott, DM; Van Dolah, RF; Burrell, VG Jr. 1983. Invertebrate
communities associated with hard bottom habitats in the South Atlantic
Bight. Estuarine, Coastal and Shelf Science 17:143-158
Werner, EE and Gilliam, JF. 1984. The ontogenetic niche and species
interactions in size-structured populations. Annual Review of Ecological
Systems 15:393-425
Woofton, RJ. 1990. Ecology of teleost fishes. Chapman and Hall, London, UK
Zaret, TM and Suffern, JS. 1976. Vertical migration in zooplankton as a predator
avoidance mechanism. Limnology and Oceanography 21(6):804-813