CLARKE: FEEDING HABITS OF STOMIATOID FISHES 



spp.) and amphipods were not further identified, 

 and other prey types were identified only to 

 major taxa. Prey types of the same genus or of 

 similar size, pigmentation, etc. were often 

 lumped for convenience of presentation of re- 

 sults. Densities of zooplankton (Table 2) were 

 calculated from the counts (corrected for any 

 subsampling) and estimated volumes filtered; 

 however, since these are based on so few samples, 

 they can be considered as only rough estimates of 

 prey abundance at the depths where the fishes 

 were caught. Furthermore, the densities of types 

 under 1.0 mm long are underestimated due to 

 mesh escapement; for most of these, densities are 

 probably about 4-5 times higher than estimated 

 from the samples (Clarke 1980). 



The nekton-eating species were much less 

 abundant than the planktivores, and their feed- 

 ing incidence and number of prey per fish were 

 lower; consequently, in order to gather as much 

 data as possible I examined specimens from a 

 wide variety of trawl samples taken between 

 1969 and 1978. These included both horizontal 

 and oblique samples in the upper 1,200 m— 

 mostly either above 350 m at night or deeper dur- 

 ing the day. Almost all were taken with a 3 m 

 Isaacs-Kidd trawl towed at ca. 2 m/s. The termi- 

 nal section was of fine (333 yum) plankton mesh 

 for about two-thirds of the samples. For a few 

 rare species I also took material from collections 

 with a 5 m Isaacs-Kidd, a 3 m Tucker, or a 2/3 

 Cobb pelagic trawl. Data from the more abun- 

 dant fishes were grouped by arbitrary size 

 classes or by time of capture. For the latter, 

 "day" included all tows started and completed 

 between sunrise and sunset plus a few dusk tows 

 which were completed after sunset but fished at 

 or near the day depths of the fishes. Similarly, 

 "night" included tows taken wholly between sun- 

 set and sunrise plus a few dawn tows that fished 

 at or near night depths of the fishes. 



Specimens were identified, measured, and 

 classified into one of four categories: Damaged — 

 the stomach ruptured or lost during capture, 

 Empty — stomach completely empty or with only 

 a trace of unidentifiable remains, Remains — 

 prey completely disintegrated but identifiable to 

 major taxon, Intact — prey in one piece or a few 

 large pieces. Sizes (standard length of fishes, 

 length without telson of crustaceans, and mantle 

 length of squids) of all intact prey items were re- 

 corded. Depending upon size of the item and de- 

 gree of digestion, the accuracy of these measure- 

 ments was an estimated ±1-5 mm. Relative 



length of the prey items, as percentage of stan- 

 dard length of the predators, was used for pre- 

 sentation. Intact crustaceans and many of their 

 remains could be identified to genus or species, 

 but only a fraction of the intact fishes could be 

 unquestionably identified. Where a fish prey 

 could not be identified positively, a probable 

 identification could be often given based on a 

 process of elimination. Because of their photo- 

 phores, myctophids and some stomiatoids could 

 be identified as such at more advanced stages of 

 digestion than other fishes; in most cases where 

 an item was clearly not a myctophid, it was in 

 good enough condition to be more precisely iden- 

 tified. 



There was little evidence of postcapture inges- 

 tion of large items. A few very fresh items, i.e., 

 those without a coating of stomach mucus or with 

 the limbs not flattened against the body, were not 

 counted. Most of these items were still partly in 

 the esophagus and were usually types not found 

 as digested remains in the same predator spe- 

 cies, e.g., a euphausiid in an otherwise piscivo- 

 rous species. As with the planktivores, there was 

 no evidence of postcapture regurgitation of prey 

 by nekton-eating species. 



Most of the nekton-eating species proved to eat 

 only small nekton (prey >10 mm long). Zooplank- 

 ton (usually copepods) were very rarely found in 

 their stomachs— always in near-perfect condi- 

 tion and never as digested remains. Certain spe- 

 cies of these fishes, however, appeared to eat both 

 small and large prey, and zooplankton were rou- 

 tinely found in their stomachs. In spite of the fact 

 that many specimens of these species were taken 

 by trawls with a fine mesh terminal section, 

 there was little evidence that the data were 

 biased by postcapture ingestion. As with the 

 strictly planktivorous species, the types of prey 

 found intact included only a narrow range of the 

 types collected in the cod end of the trawl rather 

 than a mixture as would be expected from indis- 

 criminate ingestion in the net, and digested re- 

 mains of the same types of prey were also re- 

 corded for these species. Finally, as will be 

 shown below, the incidence of small items in the 

 stomachs decreased with size of the fish; this 

 would not be expected if these species were for 

 some reason prone to ingestion after capture. 



(At towing speeds of less than ca. 1.5 m/s, post- 

 capture ingestion of both large and small items 

 appears to be a serious problem. During the 

 course of this study, I examined specimens from 

 several tows taken at 1.0-1.5 m/s. Zooplankton— 



289 



