underestimated, ranked second to Crangon 

 septemspinosa in the percent volume of stomach 

 contents of tomcod from Montsweag Bay, Maine. 



Feeding intensity showed significant differ- 

 ences between seasons by analysis of variance 

 using arc-sine transformed If values 

 (^3,190 = 11.9; P<0.001). A Student Newman- 

 Keulls test showed that /, was greatest during 

 fall, and spring values were greater than winter 

 and summer, which were similar (P<0.05) (Table 

 3). Percentage of empty stomachs was highest 

 during winter, least during fall and spring. Feed- 

 ing intensity, then, was greatest both prior to and 

 subsequent to spawning, when, presumably, 

 energy requirements were greatest. A similar 

 seasonal cycle was described for juveniles (Grabe 

 1978). 



A shift in importance of primary prey, from C. 

 septemspinosa to copepods, occurred from fall to 

 winter. A similar shift from the larger prey to 

 smaller prey was noted for juveniles (Grabe 

 1978), and it was suggested that constriction of 

 the alimentary canal by maturing gonads 

 (Schaner and Sherman 1960) was a factor. To 

 clarify this shift, predation on the primary 

 species (Gammarus spp.) and large (C. sep- 

 temspinosa) and small (copepods) secondary prey 

 were examined for the period November 1974 

 through February 1975 (November and De- 

 cember fish were young-of-the-year; data sum- 

 marized in Grabe 1978). Gammarus spp. were 

 important throughout this period, especially on 4 



Table 3. — Index of fullness' (for subsamples) and percentage of 

 empty stomachs of age 1 and 2 Atlantic tomcod from 

 Haverstraw Bay. Hudson River, 1973-76. 



December (Table 4). Crangon septemspinosa was 

 important only during November and copepods 

 were important during January and February. 

 Since gonad production was generally greatest 

 November through December and coefficient of 

 maturity peaks during November for males and 

 January for females (Orange and Rockland 

 Utilities, Inc.^), the observed shift in prey selec- 

 tion corresponded well with gonad maturation. 

 Causation has yet to be determined and small 

 sample sizes may not depict the situation accu- 

 rately. 



Tomcod are occasionally piscivorous (Alexan- 

 der 1971; Scott and Grossman 1973; Nittel see 

 footnote 1). Five fish species, including eggs, lar- 

 vae, and juvenile tomcod were identified as prey 

 and were most important during the fall (Table 

 2). Cannibalism occurred at low levels during 

 winter and spring. Cannibalism has been re- 

 ported in other fishes, e.g., A/osa pseudoharengus 

 (Rhodes et al. 1974) and Stizostedion v. vitreum 

 (Chevalier 1973) and may be a factor affecting 

 recruitment. 



Acknowledgments 



Appreciation is extended to Orange and Rock- 

 land Utilities, Inc. for permission to analyze the 

 fish, to Lawler, Matusky & Skelly Engineers for 

 laboratory facilities, and to R. E. Schmidt for re- 

 viewing an earlier draft of the manuscript. 



Literature Cited 

 Alexander, L. C. 



1971. Feeding chronology and food habits of the tomcod 

 iMicrogadus tomcod Walbaum) and winter flounder 

 [Pseudopleuronectes amencanus Walbaum) in 

 Montsweag Bay (Sheepscot River) Maine. M.S. Thesis. 

 Univ. Maine, Orono, 36 p. 



^ Dry weight of stomach contents - lO'asapercentageofvyet weight of fish 

 ^Based on total number of fish analyzed: see Table t 



'Orange and Rockland Utilities. Inc 1977. Bowline Point 

 Generating Station: Near-field effects of once-through cooling 

 system operation on Hudson River biota. 



Table 4. — Changes in the importance oiCrangon septemspinosa (CS), copepods (Cop), and Gammarus spp. (Gamm) in the 

 diet of Atlantic tomcod from Haverstraw Bay. Hudson River during the period November 1974 through February 1975. 



'Number of occurrences/lolal number of fish 



^Number of prey item a /total number of prey 



^Geometric mean of (percent occurrence • percent composition) 



1005 



