EVIDENCE OF POSTCAPTURE INGESTION BY 

 MIDWATER FISHES IN TRAWL NETS 



The ingestion of food items by midwater fishes in 

 trawl nets, if it occurs at appreciable levels, may 

 pose serious bias problems for dietary studies 

 based on stomach content analyses. In a recent 

 discussion of "net feeding," Hopkins and Baird 

 (1977) reviewed the available evidence and found 

 that while it may occur to some degree, net feeding 

 is probably not extensive. In an earlier field study, 

 Hopkins and Baird ( 1975) used side-by-side nets 

 that provided captured fishes with different levels 

 of exposure to captured zooplankton. On one side 

 the fish were allowed to enter the cod end of the net 

 and mingle with the zooplankton concentrated 

 there. In the adjacent net fishes were excluded 

 from the cod end by an 11 mm mesh bag at its 

 mouth. Their results from 19 intraspecific com- 

 parisons of 700 myctophid and gonostomatid 

 fishes showed little significant data that indicated 

 net feeding. 



All of the evidence to date, both for and against 

 postcapture ingestion, has been indirect. This is 

 because there was no sure way to determine 

 whether a food item had been ingested in the net. 

 The following study was conducted in order to 

 provide a more direct investigation of stomach 

 content contamination. 



Methods 



Experiments were conducted by introducing 

 bogus food items into the cod end of a net before 

 launching it, and then examining the stomach 

 contents of captured fishes after recovery. Eleven 

 such hauls were made with Tucker-type midwater 

 trawls in deep water off southern California (Ta- 

 ble 1). The nets had a main scoop of 6 mm nylon 

 mesh and a rear section of 0.333 mm plankton 

 netting. The 9 m^ net utilized an enclosed, bag- 

 type cod end (Baker et al. 1973) on two hauls ( 10, 

 11) and a rigid closing cod end (Childress et al. 

 1978) on three hauls (7, 8, 9). Both of these cod 

 ends are of the flow- through variety and allow the 

 passage of water out the rear. The 2.3 m^ net had a 

 rigid, nonclosing, 3.7 1 plastic jug cod end that 

 restricted flow. 



Prior to launching the trawl, approximately 100 

 ml (or about 3,000 pieces) of artificial prey were 

 placed in the cod end. In all cases, the amount of 

 bogus prey introduced was much less than the 

 eventual catch of similarly sized zooplankton in 



Table l. — Trawling data for the ingestion experiments at 

 Santa Barbara (SB) and San Clemente (SC) Basins, and off 

 Guadalupe Island (GI), Calif. 



Date 

 1977 



Mouth 



opening Day/ 

 Location (m^) night 



Depth 



(m) 



Duration 



(min) 



the cod end. The material consisted of rubber band 

 fragments and bits of filter paper, between 2 and 

 15 mm in greatest dimension. Their individual 

 volumes ranged between 0.5 and 60 mm^, which 

 falls within the size range of natural prey items. 

 Upon recovery, the cod end samples were pre- 

 served initially in 10% formaldehyde then trans- 

 ferred to 50% isopropanol. In the laboratory, fish 

 stomachs were removed from the body cavity be- 

 fore being opened for examination. Only material 

 from intact stomachs was counted, material found 

 in the mouth and esophagus was not recorded. 

 Data from haul 23 are biased toward larger indi- 

 viduals because the smallest specimens in the 

 catch were not examined. Percent net feeding rep- 

 resents the relative number of individuals of any 

 species which had ingested at least one bogus prey 

 item. It is necessarily a conservative representa- 

 tion because zooplankton from the cod end may 

 also have been ingested after capture but could not 

 be distinguished from naturally ingested prey. 



Results 



A total of 1,211 specimens were examined, rep- 

 resenting 15 midwater fish species. Fifty-nine in- 

 dividuals (5% of the total) from 10 species were 

 found with artificial prey in their stomachs (Table 

 2). Most of the bogus prey ingested (92%) were 

 small (0.5-6 mm'*) and only four fish had swal- 

 lowed artificial items >12 mm^. Generally, the 

 average number per stomach was low ( Table 2 ) but 

 a few fish had their stomachs packed with arti- 

 ficial prey. Only 5 of the 59 fishes containing bogus 

 prey had stomachs which were otherwise empty; 

 all others also contained zooplankton, some por- 

 tion of which may have been ingested in the cod 

 end. 



Notable differences in net feeding occurred both 

 interspecifically and intraspecifically. The two 



nSHERY BULLETIN; VOL. 77, NO, 3, 1980, 



713 



