610 



Fishery Bulletin 92(3). 1994 



Table 1 



Habitat characteristics (means) of three sampling sites in the New York Bight apex for the 18 months before (A) and 

 21 months after (B) sewage sludge disposal cessation; benthic biomass (wet wt) does not include the biomass of two 

 large bivalves, Pitar morrhuanus and Arctica islandica. 



Stations 



R2 



NYli 



NY11 



Depth (m) 

 Sediment 



grain size (phi)' 

 TOC (%dry wt)' 

 Cr (ppm dry wt) 2 



Bottom water' 



min. dissolved oxygen (mg/L) 

 temp, range CO 



Benthos 



biomass (g/m-) 

 species (n )/grab 



29 



:il 



29 



4.2 



5.7 J 



4.3 5.8-* 



-3-17 for all sites 5 



5.3 



<5.5 4 



218.2 

 30 



202.8 

 33 



63.4 

 19 



28.2 

 33 



66.6 

 41 



73.2 



45 



' Packer et at, in press. 



2 Zdanowicz et al. in press. 



3 Arlen. L, A. Draxler. and R. Bruno. Hydrographic observations in the bottom water of the New York Bight at the "12 mile" dumpsite: 1983-1990. 

 Unpubl. manuscr. 



4 Dissolved oxygen content of <3 mg/L was recorded in September, 1989. 



5 Summer levels in 1988 were 2-3°C lower than the long-term mean. 



The mean proportional volume (percent of the to- 

 tal volume) of a given prey item in the stomach was 

 compared with the mean proportional biomass of that 

 prey in the benthic macrofauna, and the relation- 

 ship between these proportions was used to deter- 

 mine whether selective predation was evident. Al- 

 though this approach to examining predator-prey as- 

 sociations differs from the traditional use of prey nu- 

 merical abundance in stomachs and the environment, 

 it is nonetheless realistic and useful because 1) prey 

 volume can be a more precise dietary variable com- 

 pared with uncertain enumeration of prey that are 

 easily fragmented and for which only parts are 

 present and 2) volume and biomass are approxi- 

 mately equivalent for most common prey taxa (see 

 below). However, amphipods were also considered 

 numerically because they are usually eaten whole 

 and their exoskeletons are resistant to digestion and 

 thus allow a reasonably accurate assessment of the 

 number of individuals eaten. 



The stomachs of 3,556 adult winter flounder, 18- 

 30 cm total length, examined from the three study 

 sites had identifiable food in them. To examine over- 

 all predator-prey relationships, stomach content and 

 benthic data from the entire study period were pooled 

 for each station because there were only minor 

 changes in dominant prey and benthic species asso- 



ciated with sludge disposal cessation (Reid et al., in 

 press; Steimle, in press). The use of pooled stomach 

 content data to estimate prey preference by a preda- 

 tor population is recommended by some authors, e.g. 

 Rachlin et al. ( 1987). Changes in predator-prey rela- 

 tionships related to cessation of sludge disposal are 

 considered separately. Any seasonal or annual vari- 

 ability was assumed to be distributed equally within 

 the pooled data as there were no gaps in collections. 

 Log transformation of Shorigin's forage ratio index 

 (K=rtlpt), adapted from a numerical approach for prey 

 volume and biomass, was used to estimate prey se- 

 lectivity (Berg, 1979): K = Log I0 irt/pt), where rt = pro- 

 portion of prey in the diet estimated by contribution to 

 stomach volume and pi = proportion of prey in the 

 benthic biomass. Positive K values suggest a degree of 

 selective predation. Near-zero K values suggested that 

 predation is directly proportional to abundance. Nega- 

 tive K values suggest underutilization or avoidance of 

 a potential prey relative to its availability. 



The use of volume and biomass to calculate K is 

 reasonable because 1 niL or cm :i of prey volume is 

 considered approximately equivalent to 1 g of macro- 

 fauna wet weight (Bowman, 1986). We partially veri- 

 fied this assumption by determining the mean vol- 

 ume to wet weight ratios for a number of individuals 

 for a range of common prey taxa. The ratios for an 



