LAROCHE: TROPHIC PATTERNS AMONG LARVAE OF SCULPINS (COTTIDAE) 



shown to increase feeding success in early larvae 

 (Blaxter 1965; Hunter 1980), as well as by pro- 

 longing the time before larvae become totally 

 dependent on an exogenous food supply which is 

 limited in abundance and distribution (Laurence 

 and Rogers 1976). Potential or actual energy 

 deficits prior to or coincident with the initiation 

 of feeding were found to occur in the tautog, Tau- 

 toga onitis (Laurence 1973), and Pacific sardine, 

 Sardinops sagax (Lasker 1962), both relatively 

 fecund species having planktonic eggs and lar- 

 vae. Yet, deficits based on the amount of yolk 

 absorbed did not occur until after feeding had 

 begun in rainbow trout, Salmo gairdneri; brown 

 trout, Salmo trutta; bluegill, Lepomis macro- 

 chirus', and largemouth bass, Micropterus sal- 

 moides, all of which have relatively low fecun- 

 dity, and demersal eggs and larvae (Laurence 

 1973). Although energy budgets have not been 

 described for them, it is likely that cottids have 

 yolk-utilization characteristics similar to this 

 latter group of fishes. 



The relatively higher incidence of empty guts 

 among the almost exclusively piscivorous larvae 

 of H. americanus may also be related to trophic 

 energetics. Since fish larvae probably supply 

 more energy per unit effort as prey than do crus- 

 taceans because of larger size and lack of indi- 

 gestible exoskeletons, ingestion of single, large, 

 highly digestible prey would reduce the need for 

 continuous foraging. Larvae of H. viliosus in the 

 Japan Sea, like those of its congener in the Gulf of 

 Maine, also feed on other fish larvae (Okiyama 

 and Sando 1976). Most reports of piscivorous lar- 

 vae have come from observations of laboratory- 

 reared larvae, particularly scombroid fishes 

 (Beyer 1980; Hunter 1980). 



The larval diets of four of five species of cot- 

 tids in the Damariscotta River estuary were dis- 

 tinctly different, despite similarities in mouth 

 size and morphology and the absence of obvious 

 specialized morphological adaptations for feed- 

 ing. Last (1978a) also observed differences in lar- 

 val diets among pleuronectiform fishes in the 

 North Sea and concluded that this represented a 

 mechanism whereby direct competition for food 

 is avoided. There are many theoretical and prac- 

 tical difficulties in assessing the significance of 

 competition as a cause of dietary divergence in 

 fishes (Zaret and Rand 1971; Weatherly 1972), 

 yet existence of competition among tropical 

 stream fishes has been inferred from observed 

 reductions in diet overlap during periods of low- 

 ered food supply (Zaret and Rand 1971). Com- 



parisons of dietary overlap and food supply 

 among the larvae of Myoxocephalus and Triglops 

 species during winter and early spring indicated 

 a similar pattern. The lowest values of percent 

 diet overlap, i.e., occurrence of the most dissimi- 

 lar diets, among five of six possible pairs of spe- 

 cies combinations occurred in January and Feb- 

 ruary, and coincided with lowest prey abundance 

 as indicated by the lower mean plankton volume 

 (an approximate measure of food supply) of the 

 two periods compared, 16.8 ml (n = 22) vs. 31.0 

 ml {n = 25) in March. 



Percentage diet overlap between M. aenaeus 

 and M. octodecemspinosus, whose larvae were 

 the most abundant in the estuary (Laroche 1980) 

 and who had the most similar food habits, re- 

 mained relatively high and constant through 

 winter and early spring. Some mechanism other 

 than dietary shifts, i.e., changes in dominant 

 prey, may act to reduce competition between lar- 

 vae of these two species. Although both species 

 were more abundant 1.5 m above the bottom 

 than in the upper 1.5 m during winter and spring 

 months, M. aenaeus larvae were 5 times more 

 abundant while M. octodecemspinosus were only 

 1.7 times more abundant (Laroche 1980). Such 

 relative differences in vertical distribution re- 

 sulted in spatial separation of most M. aenaeus 

 and M. octodecemspinosus larvae. If competition 

 is indeed an important factor in diet determina- 

 tion among cottid larvae, this spatial separation 

 may explain the high degree of dietary overlap 

 between these two species. 



Density-independent food exploitation (Hyatt 

 1979) by fish larvae has not been demonstrated 

 by quantitative comparisons of prey abundance 

 in larval guts to their abundance in the plankton. 

 Blaxter (1965) cited numerous examples, mostly 

 qualitative, of apparent selection among herring 

 larvae for specific taxa of copepod and diatom 

 prey over other prey which were more abundant 

 in the plankton. Qualitative comparisons of the 

 composition and relative abundance of the zoo- 

 plankton present during winter and spring 

 months of 1972 in the Damariscotta River estu- 

 ary (Lee 1974) and prey found in larval cottids 

 from winter and spring 1973 yielded indirect 

 evidence of density-independent foraging among 

 these larvae. Although zooplankton species com- 

 position and abundance may vary from year to 

 year, relative constancy has been demonstrated 

 in larval diets of at least two of the five species of 

 cottids in the estuary. Townsend (1981) observed 

 that the principal prey of M. octodecemspinosus 



837 



