Campfield and Houde: Ichthyoplankton community structure and comparative trophodynamics 
3 
Processing of samples 
In the laboratory, fish larvae and juveniles were identi- 
fied to species and enumerated. Total lengths (up to 100 
individuals of each species per sample) were measured to 
the nearest 0.1 mm. Zooplankton concentrations (no./L) 
were estimated for several taxonomic categories, includ- 
ing copepodites and adults of the calanoid copepods 
Eurytemora affinis and Acartia so., cyclopoid copepods, 
copepod nauplii, invertebrate eggs, rotifers, and the cla- 
doceran Bosmina longirostris. Lengths of zooplankters 
were measured under microscope to the nearest 0.1 mm 
with an ocular micrometer. 
Ichthyoplankton distributions, abundances, and 
assemblages 
Larval concentrations (no./m 3 ) at each sampling site 
were calculated from numbers per tow and measured 
tow volumes. Total larval abundances in river segments 
representative of each sampling site were estimated by 
expanding larval concentrations at sites to abundances 
in documented river-segment volumes (Cox et al. 3 ). 
Abundances were analyzed with respect to hydrographic 
gradients. Species richness and Shannon-Wiener diver- 
sity (H’) (Magurran, 1988) were calculated to compare 
ichthyoplankton assemblages among river regions. Habi- 
tat use and spatial overlap were examined in principal 
components analysis (PCA) of catch per unit of effort 
(CPUE, no. of larvae/tow) and hydrographic data to 
identify and describe associations (Miller, 2002). 
Ichthyoplankton concentrations were analyzed for sta- 
tions grouped into three designated regions within the 
estuary: freshwater, salt front, and oligohaline. Fresh- 
water stations had no detectable salinity. Salt front 
stations had salinities of 0. 2-1.0 and oligohaline sta- 
tions had salinities >1.0. In each survey, a region typi- 
cally included 3-4 stations (Fig. 1). Zooplankton and 
hydrographic variables were included in multivariate 
analyses to evaluate these factors with respect to lar- 
val concentrations of alewife, striped bass, white perch, 
and naked goby. In statistical analysis, the normality 
assumption generally was met by applying log 10 (x+l) 
transformations to larval fish and zooplankton concen- 
trations. Multiple regressions with larval concentration 
as the dependent variable were run on combinations of 
independent variables in a stepwise procedure (forward- 
backward) to select variables for inclusion as descriptors 
of larval concentration. Independent variables consid- 
ered were temperature, salinity, conductivity, dissolved 
oxygen, pH, river flow, salt front location, and the con- 
centrations of copepod nauplii, calanoid copepods, and 
Bosmina. The probability threshold for including or 
removing variables was 0.10. The quality of model fits 
was assessed by using Akaike’s information criterion 
(AIC). The model with lowest AIC value was retained 
for each larval taxon (Kleinbaum et al., 1998). 
Diet composition 
Digestive tracts of larval alewife, striped bass, white 
perch, and naked goby from the three designated regions 
were analyzed to determine feeding incidence, kinds 
of prey, prey numbers, prey sizes, prey selection, and 
dietary overlap. Larvae were grouped into 1-mm (naked 
goby) or 2-mm (other taxa) length intervals for diet 
analyses. In prey selection and dietary overlap compari- 
sons, larvae were analyzed in broader length classes (<10 
mm or >10 mm total length [TL] ), except for naked goby 
which was analyzed as a single class (5-12 mm TL). All 
prey in larval guts were identified, enumerated, and 
measured with an ocular micrometer. Maximum lengths 
or widths of each prey taxon were recorded from each 
larval gut. 
Diets of 633 larvae were analyzed (135 alewife; 165 
striped bass; 200 white perch; 133 naked goby). Diets 
and zooplankton concentrations were evaluated from 
sampling sites throughout the estuarine transition zone 
(Fig. 1), and from the salt front and intersection of the 
2.0 isohaline with bottom. Trophic niche breadth (S), a 
measure of variability in sizes of prey in the larval diet 
(Pearre, 1986) was defined as the standard deviation of 
mean logarithmic (log e ) prey size. An ontogenetic index 
(0 £ ) (Fuiman et al., 1998) was calculated to character- 
ize and compare feeding with respect to developmental 
state of larvae. Mean prey size and niche breadth esti- 
mates were regressed on larval length and ontogenetic 
state to examine patterns in the sizes of prey consumed. 
Selection of prey types and sizes in larval diets was 
evaluated by applying Strauss’s (1979) prey selection 
index: 
L = r i -p i , (1) 
where r,- = the proportion of zooplankton prey type i in 
larval guts; 
p t = its proportion in the environment; and 
L can range from -1.0 to +1.0. 
Positive values indicate prey preference and negative 
values avoidance. The standard error of L was estimated 
(Strauss, 1982) and Atests were conducted to determine 
whether L differed significantly (P<0.05) from 0. 
The importance of prey types and sizes in larval diets 
was evaluated with a relative importance index (George 
and Hadley, 1979). For each fish species, the importance 
(A) of prey type a, for length class i was computed with 
the following equation: 
A i = % frequency of occurrence 
+ % total number + % total weight', 
3 Cox, A. M., P. A. Waltz, and P. G. Robertson. 1980. Patux- 
ent River Program. Bathymetric investigation of the Patuxent 
River system, p. 11-83. Maryland Department of Natural 
Resources, Water Resources Administration, Annapolis, MD. 
A,,„ 
V a = 1 
( 2 ) 
