CLARKE: DIETS OF FOURTEEN SPECIES OF MESOPELAGIC FISHES 



majority of prey types eaten by the fishes consid- 

 ered here. Certain types, e.g., large (>10 mm) 

 euphausiids, shrimps, or fish larvae, must cer- 

 tainly be able to avoid the bongo nets; con- 

 sequently, high ASV's associated with such types 

 must be considered as doubtful. 



Uncertainty associated with the estimated 

 densities from the plankton tows probably limits 

 interpretation more than any other factor. Since 

 only one or two pairs of zooplankton samples were 

 available from each depth, the sampling error as- 

 sociated with estimated densities cannot be 

 specified. Overall, the between tow, between net, 

 and between aliquot differences in counts of abun- 

 dant types indicated that the densities and there- 

 fore the ASV's are probably accurate to within a 

 factor of 0.5-2 X of the values given. Thus small 

 differences in ASV's cannot be considered real. 

 Absurdly high values of ASV frequently resulted 

 for prey types that were very rare or absent in the 

 plankton samples. Such types were frequently 

 large forms that may have been "rare" because of 

 net avoidance, and even for those that were truly 

 rare, the potential sampling error was probably 

 large due to insufficient volumes sampled. Con- 

 sequently, after inspection of the data, all values 

 >1.0 m^ were lumped together. 



RESULTS 



A total of 14 species of fishes comprising 51 size- 

 depth-species categories (Table 2) yielded suffi- 

 cient data to merit presentation and discussion. 

 Although most prey items were identified to genus 

 or species and all were measured to the nearest 0.1 

 mm, certain prey were grouped by higher taxa or 

 size ranges for presentation of prey densities (Ta- 

 ble 3) and to avoid dealing with low numbers in 

 calculations of ASV's. 



In the individual species accounts below, an at- 

 tempt is made to summarize the major points in 

 the tabulated data. For these purposes and sub- 

 sequently throughout the paper, "microzooplank- 

 ton" are operationally defined as those prey types 

 too small ( <1.0 mm) to have been accurately sam- 

 pled by the plankton tows and thus those whose 

 ASV's are overestimated. The remaining prey 

 types or "macrozooplankton" are considered by 

 species or as small (1.0-1.5 mm), medium (1.5-3.0 

 mm), or large (^3.0 mm). For each category of 

 fishes considered, the number of macrozooplank- 

 ton prey types and their frequencies in the diet are 



grouped by ASV values in 0.1 m'' increments be- 

 tween and 1.0 m^ (Table 2). 



LampanyctNs steinhecki (Table 4) 



The data for L. steinhecki are the most extensive 

 of all species considered. Large numbers of at least 

 two size classes were taken at each of the four 

 depths sampled, and, in spite of the rather low 

 numbers of prey per fish, the numbers identified 

 for most categories were relatively high. The 18-25 

 mm fish from 90 and 110 m and 36-45 mm fish from 

 170 m may have included some individuals caught 

 in transit above the towing depth. 



Microplankton were of minor importance in the 

 diets of all but the smallest size groups considered. 

 Small macrozooplankton were eaten infrequently 

 and had low ASV's for all sizes of fish. The most 

 frequently taken prey were euphausiids and 

 medium to large copepods. The ASV's for these and 

 other large prey were usually relatively high. 

 Candacia longimana was most consistent in this 

 respect. The ASV's for Pleuromamma xiphias at 

 90 m were markedly lower than at the other 

 depths as were those for Euphausia spp. at 70 m. 

 Neither of these exceptions appeared to result 

 from differences in importance in the diet. 

 Pleuromamma xiphias was extremely abundant 

 at 90 m (Table 3), and this, combined with the 

 lower numbers of prey per fish at this depth, 

 caused most of the reduction in ASV Euphausia 

 spp. were extremely abundant at 70 m (Table 3); 

 most were E. tenera, a species eaten infrequently. 

 As a coni^equence of these and similar differences 

 between depths, there was no clear trend or consis- 

 tency to the distribution of ASV's of the different 

 prey types. Most types and most items had low 

 ASV's at 90 m, ASV's were more nearly evenly 

 distributed at 70 and 110 m, and the majority of 

 prey had high ASV's at 170 m (Table 2). 



Lampanyctus nohilis (Table 5) 



Lam.panyctus nobilis was taken from three 

 depths; with the possible exception of the smallest 

 size group from 110 m, the data were unlikely to 

 have been seriously affected by catches in transit 

 to and from towing depth. 



The diet of L. nohilis was generally similar to 

 that of L. steinhecki but with a greater frequency 

 of large prey. Microzooplankton were hardly eaten 

 (Table 2), and ASV's for the few types of small 

 macrozooplankton were very low. The most fre- 



623 



