FISHERY BULLETIN: VOL. 70. NO. I 



only euphausiids were reported in a previous 

 study on the diet of S. similis (Renfro and 

 Pearcy, 1966). 



A large percentage of shrimp of all sampling 

 categories contained mixtures of granular and 

 fibrous debris. The unidentifiable state of this 

 material may not be entirely the result of di- 

 gestive processes, but of the decomposed nature 

 of the material at the time of its ingestion. Ser- 

 gestes similis, like S. lucens (Omori, 1969), may 

 scavenge decomposing dead material in addition 

 to taking living prey. Another possibility is that 

 this material represents the remains of prey 

 stomach contents, e.g., those of euphausiids. 



There is a tendency for day-net overall means 

 to be less than those of night-net samples (Table 

 3). This and the trend toward less advanced 

 digestion in night-net samples suggest more in- 

 tense feeding activity at night, as was reported 

 for S. lucens (Omori, 1969) . The lesser average 

 number of calanoids per foregut in day-net 

 samples may be attributed primarily to the no- 

 tably fewer Calanus in that category (Table 4). 



Although the median values are not statisti- 

 cally different, day and night differences in over- 

 all mean numbers of the two calanoid species 

 most frequently occurring in net samples, Me- 

 tridia pacifica and Calamis pacificus, are notable 

 in that they agree with differences in the vertical 

 distributions of these two species. In the Cal- 

 ifornia Current region south of lat 33° N, Cal- 

 anus pacificus usually concentrates in and near 

 the thermocline during the day and disperses 

 throughout the mixed layer at night. M. pa- 

 cifica, on the other hand, occurs in the vicinity 

 of the thermocline at night and disperses down- 

 ward during the day (Fleminger, unpublished 

 data). Studies have shown S. similis to be con- 

 centrated between the surface and 200 m at night 

 and between 250 and 500 m during the day 

 (Barham, 1957; Pearcy and Forss, 1966). The 

 diurnal vertical range of M. pacifica, then, seems 

 to correspond more closely with that of S. similis 

 than does the diurnal range of C. pacificus. Ser- 

 gestes similis probably has access to quantities 

 of M. pacifica during both day and night. Most 

 likely, S. similis encounters and feeds upon con- 

 centrations of C. pacificus primarily at night 

 after the sergestid has ascended to shallower 



depths. However, without knowledge of di- 

 gestive rates, these considerations are specu- 

 lative. 



The large numbers of fish scales in net samples 

 and their complete absence in fish-stomach sam- 

 ples strongly suggest that feeding in the net 

 has occurred. Although the source of these scales 

 cannot be ascertained, it seems probable that 

 they are the highly deciduous scales of lantern- 

 fish captured in the net with the shrimp. The 

 significantly greater numbers of euphausiids and 

 chaetognaths and the higher diversity of cala- 

 noids in net samples may also be indicative of 

 feeding after capture. Many preserved net- 

 caught S. similis have been observed by one of 

 us (Judkins) to have fish scales, chaetognaths, 

 and small crustaceans packed into their mouth- 

 parts and sometimes gripped in their mandibles. 



Special conditions under which fish-caught 

 shrimp might have been feeding before capture 

 (related perhaps to time or depth) may also have 

 contributed to the observed disparities between 

 net and fish samples. Albacore are thought to 

 feed primarily during daylight hours and prob- 

 ably most intensively in the early morning and 

 early evening (Iversen, 1962). Time of capture 

 by trolling of the albacore that we examined is 

 available for only about half of the specimens. 

 Of these about half were taken in morning day- 

 light hours and the remainder were from the 

 late afternoon and early evening. If all of the 

 fish-caught shrimp were captured during the 

 day, the relatively small quantities of identifiable 

 components might reflect less intensive feeding 

 by the shrimp at the time they were ingested 

 by the albacore. 



The generally smaller size of fish-caught spec- 

 imens is probably not a factor; small net-caught 

 shrimp (carapace length less than 10 mm) con- 

 tained numbers of fish scales, euphausiids, and 

 chaetognaths proportionally as high as larger 

 net-caught individuals. Differences amongst 

 sergestid specimens in the length of time spent 

 in an albacore stomach also appear to be a neg- 

 ligible factor. Foregut contents of nearly in- 

 tact fish-caught shrimp do not differ appreciably 

 from those of extensively decomposed shrimp. 



In general, the various samples of foreguts 

 from adult S. similis that were analyzed provide 



222 



