GABRIEL and PEARCY: FEEDING SELECTIVITY OF DOVER SOLE 



The mean depth of a prey taxon within the sed- 

 iment was significantly related to fish size at 

 which frequency of that prey increased (Figure 6, 

 r = 0.542, P <0.05). Prey which were important to 

 large fish were usually found deeper in the sedi- 

 ment. Prey occurring frequently in small fish were 

 found near the sediment surface. Prey which oc- 

 curred at statistically equal frequencies for all- 

 sized fish were generally found within the top 4 cm 

 of the core sample, e.g., Lumhrineris latreilli, 

 which was consumed in large numbers by fish of 

 all sizes. Prey which decreased in occurrence in 

 larger fish were often concentrated near the sur- 

 face, with a mean depth distribution of 2 cm. Al- 

 though the relationship between depth of prey in 

 sediment and index of body prey was not signifi- 

 cant (Figure 7) (r = 0.220), few large-bodied prin- 

 cipal prey taxa had a mean depth in the sediment 

 <4 cm. 



At SGIO, few prey taxa changed significantly in 

 frequency of occurrence over the range offish sizes 

 sampled. The mean depth of a taxon within the 

 sediment was rarely >4 cm, and usually <3 cm, 

 regardless of prey body size. The depth range of all 



7t- 



6 - 



E 

 u 



X 4 



I- 



Q. 



liJ 



O 



liJ 



,22 



± 



15 20 25 30 35 



FISH LENGTH (cm) 



40 



-~ 3 



E 



o 



CL 

 Q 



< 

 UJ 



6 - 



7- 



FlGURE 6. — Mean depth of prey in the sediment vs. fish length at 

 which frequency of prey increased significantly. Numbers desig- 

 nate prey taxa (see Table 3). 



1 2 3 4 5 6 



INDEX OF BODY SIZE (mg/cm) 



Figure 7. — Mean depth of prey in the sediment vs. index of prey 

 body size. Numbers designate prey taxa i see Table 3. Nonprinci- 

 pal prey: 38 = Arlcidea neosuecica, 39 = Laonice cirrata, 40 = 

 Pherusa papillata ,41 = Poly dorasocialis, 42 = Spiochaetopterus 

 costarum, 43 = Haploscoloplos elongatus). 



invertebrates at this station was generally shal- 

 lower than at SG29. 



Prey Abundance Patterns 



Prey abundance also varied with location. Al- 

 though the total density of individuals per square 

 meter was slightly higher at SGIO than SG29, the 

 density of polychaetes, a preferred taxon, aver- 

 aged 886 individuals/m^ at SG29 and only 397 

 individuals/m=^ at SGIO (Table 7). Most principal 

 polychaete taxa were found in lower densities at 

 SGIO than SG29. The density of pelecypod mol- 

 luscs, a negatively selected (avoided) taxon, was 

 several times higher at SGIO than SG29, 861 and 

 297 individuals/m^, respectively. 



DISCUSSION 



How similar is the Dover sole to the hypothetical 



759 



