HOBSON and CHESS: MOVEMENTS OF PRANESUS PING U IS 



off the point at the western end of the bay, and 

 then progressing steadily eastward along the 

 shore. Thus, in the morning the silverside 

 reversed the route they had taken offshore the 

 night before. The vanguard of the school arrived 

 at the diurnal schooling site on each of the 

 three mornings at 20, 18, and 16 min before 

 sunrise, respectively. Because these observa- 

 tions, unlike those of the evening, were made 

 from the same vantage point, they are directly 

 comparable. 



For a while after arriving inshore, the silver- 

 side remained in a compact school but pro- 

 gressively strung themselves out along the 

 beach until the midday situation prevailed. 



P. piiigiiis is another example of a fish that 

 assembles in relatively inactive schools in shal- 

 low water during the day and then migrates to 

 feed elsewhere at night. This phenomenon is 

 known in certain clupeids, pomadasyids, caran- 

 gids, sciaenids, and lutjanids in the Gulf of 

 California (Hobson, 1965, 1968), and has also 

 been reported for many species in the tropical 

 Atlantic (Starck and Davis, 1966). Obviously it 

 is a widespread activity pattern among shore 

 fishes in tropical seas. 



FEEDING HABITS 



The inactivity of silverside (>35 mm long) 

 that hover in their inshore schools during the 

 day indicates that they do not feed at this time. 

 To obtain more data on this question, specimens 

 were collected for gut-content analysis three 

 times during the day (19 March). The first collec- 

 tion was made early in the day (0700 h); the 

 second collection, during midday (1030 h); and 

 the third collection, late in the day (1600 h). Ten 

 specimens from each collection were examined. 



All 10 specimens (61 to 73 mm long) examined 

 from the early collection had material in their 

 guts, which on the average were about 20% full. 

 Composition of these gut contents, ranked as 

 the mean percent (by volume) of each item in all 

 silverside specimens, were as follows: hyperiid 

 ami)hipods, 49.6% ; caridean shrimp larvae, 

 6.7% ; calanoidcopepods,6.2% ; ostracods, 5.5% ; 

 crab megalops, 2.3% ; mollusk veliger larvae, 

 1.0% ;polychaetes,0.5% ;foraminiferans(tretom- 

 phalus stage), 0.4% ; and unidentified, well- 



digested material 27.8% . Digestion had damaged 

 all of this material so that identifications could 

 be made only to major groups. 



Of the 10 specimens (61 to 70 mm long) 

 examined from the midday collection, three 

 were empty; the guts of the remaining seven 

 were on the average about 15% full, but diges- 

 tion had reduced this material to a soup of 

 unidentifiable fragments. 



Finally, the guts of all 10 specimens (58 to 

 64 mm long) from the collection made late in the 

 day were empty. 



The data clearly demonstrate that the silver- 

 side in these inshore schools do not feed during 

 the day. But even in those individuals taken dur- 

 ing early morning, digestion had damaged the 

 gut contents beyond the point where a precise 

 analysis could be made. Being small themselves, 

 the silverside take tiny prey, mostly under 2 

 mm long, whose identifying features do not 

 resist digestion well. 



To obtain material for a detailed analysis of 

 food habits, we went 200 to 800 m offshoi'e from 

 Arniel Island between 0400 and 0500 h on 23 

 and 24 March. Here, using a dip net and flash- 

 light from a moving skiff, we collected 64 speci- 

 mens between 35 and 69 mm long, which were 

 immediately placed into a 10% Formalin^ solu- 

 tion. With this procedure, we obtained fish that 

 were actively feeding, and whose gut contents 

 (Table 1) included fresh material. In Table 1, the 

 items are ranked by a "ranking index," which 

 is computed by multiplying the mean percent 

 (by volume) of each item in the gut contents of 

 all silverside specimens bj^ the ratio of silverside 

 containing that item to total silverside sampled. 

 Thus, for Hyperia boigaleusis found in silver- 

 side collected on 23 March, the ranking index is 

 46.71 X 30/31 - 45.20. To increase the pre- 

 cision of the ranking index, the mean percent of 

 each item in the gut contents is carried to two 

 decimal places. The actual observed values for 

 individual fish were estimated to the nearest 

 percent, except where the estimate was <1.00, 

 in which case the estimate was to > or <0.50% . 

 For purpose of calculating the mean percent. 



' Reference to trade names does not imply endorsement 

 by the National Marine Fisheries Service, NOAA. 



781 



