Taylor and Gervasi: Feeding habits and dietary overlap of age-0 flounder 
177 
HAmphipod 
^ Polychaete 
^ Bivalve 
^Copepod 
1 1 Isopod 
1 Other 
100 
£ 
Total length (mm) 
Small 
Small-medium 
Medium-large Large 
B 
Amphipod 
^ Mysid 
^ Fish 
^ Copepod 
1 1 Sand shrimp 
im Other 
100 
£ 
Total length (mm) 
Small 
Small- 
medium 
Medium-large 
Large 
Figure S 
Contributions of prey taxa to the diets of (A) winter flounder (Pseu- 
dopleuronectes americanus) and (B) summer flounder (Paralichthys 
dentatus) as a function of “preserved” total length (i.e., measured in 
the laboratory after specimens were preserved in 70% ethanol). Diet 
is expressed as an index of relative importance (%1RP, Eq. 1), and 
horizontal brackets represent distinct (size-based) dietary groups 
determined from hierarchical cluster analyses and similarity profil¬ 
ing (Fig. 3). Winter and summer flounder were collected from the 
Seekonk and Taunton Rivers during 2009-2015. 
Moreover, in the Taunton River, winter 
flounder consumed a high proportion of 
crab megalope in July and August-Sep- 
tember, especially in the middle and up¬ 
per portions of the river (TR1-TR3, July 
and August-September %IRI for crab: 
1.6-39.4%). Spatial differences in the diet 
of winter flounder from the Taunton River 
were attributed also to the disproportion¬ 
ate contribution of bivalves and anthurid 
isopods in the upper reaches of the river 
(TRl and TR2 %IRI: bivalve=7.4-26.4%; 
isopod=3.0-12.4%), and polychaetes were 
dominant at TR4, particularly from mid- 
to late summer (TR4, July and August- 
September %IRI for polychaetes, mostly 
unidentified, phyllodocid, and Polydora 
spp., was 52.0-77.4%) (Fig. 6C). In con¬ 
trast, bivalves and idoteid isopods were 
important prey in the lower reaches of 
the Seekonk River (SR3 and SR4 %IRI: 
bivalves=9.1-27.6%; isopod=2.3-6.3%), 
whereas chironomid larvae were relatively 
unique to SR2 (SR2 %IRI for chironomids: 
1.2-20.0%) (Fig. 6A). Amphipods were 
a broadly used prey resource by winter 
flounder, but no discernible spatiotempo- 
ral patterns in their dietary contribution 
were evident (Fig. 6). 
Summer flounder feeding habits in 
each river varied temporally (2-way PER- 
MANOVA for month: Seekonk, pseudo- 
F=5.39, P<0.001; Taunton, pseudo-F'=3.86, 
P<0.002), but dietary differences across 
sites were evident only in the Taunton Riv¬ 
er (2-way PERMANOVA for site: Seekonk, 
pseudO”P=1.90, P=0.08; Taunton, pseudo- 
P=3.97, P<0.002) (Fig. 7). Further, in each 
instance, the site-month interaction effect 
was not significant (2-way PERMANOVA: 
sitexmonth, pseudo-P=0.50-0.64, P=0.79- 
0.96). The first and second axes of the 
PCO biplots were correlated most with 
month and site, respectively, and account¬ 
ed for 55.4-66.5% and 14.7-24.3% of the 
total variation in diet of summer flounder 
in the Seekonk and Taunton Rivers (Fig. 
7, B and D). 
The significant temporal variation 
in diet of summer flounder was attrib¬ 
uted to the initial contribution of mysid 
shrimp and copepods in May-June and 
subsequent dietary shifts toward amphipods in later 
months (from May to August-September, %IRI de¬ 
creased for mysid shrimp from 54.8% to 16.2% and 
for copepods from 10.9% to 0.0% and increased for 
amphipods from 15.6% to 50.0%) (Fig. 7). The dietary 
contributions of other important prey taxa, includ¬ 
ing sand shrimp and fish, varied inconsistently across 
months. Further, summer flounder collected from the 
upper reaches of the Taunton River consumed more 
amphipods and fish than conspecifics from southerly 
locations (TR1-TR4 %/P/: amphipods=47.3% versus 
0.0%; fish=8.9% versus 0.0%) (Fig. 7, C and D). Sand 
shrimp were also overwhelmingly dominant at TR4 
in July (%/P/=97.5%); however, dietary resolution for 
this site-month interaction was confounded by small 
samples sizes (?ic=3). 
