FISHERY BULLETIN: VOL. 81, NO. 3 



night lights as far as 10 km offshore in water hun- 

 dreds of meters deep. 



Before the appearance of the gold spot herring, the 

 dominant pelagic planktivorous fishes in Kaneohe 

 Bay were the Hawaiian anchovy or nehu (which is the 

 main source of bait for the local skipjack tuna fishery) 

 and the iao, Pranesus insularum (Atherinidae). 

 Although the nehu and gold spot herring cooccurred 

 in almost all purse seine catches and adult herring oc- 

 casionally eat larval or juvenile nehu (see below), 

 there was no evidence that the herring's appearance 

 substantially affected the nehu population. The 

 nighttime areas of highest abundance for the two 

 species appear to be slightly different within the bay, 

 and nehu by day prefer more turbid, brackish areas 

 than do the herring. The iao population, however, ap- 

 pears to have been affected markedly by the herring. 

 Prior to the appearance of herring, iao were regularly 

 present in large numbers over shallow reef flats dur- 

 ing the day, but since have been nearly completely 

 replaced by herring in the same situations and are 

 much less frequently seen. 



Qualitative examination of stomach contents of 

 gold spot herring showed that they eat a wide variety 

 of zooplankton and indicated that they feed primarily 

 but not exclusively at night. Fish < 3 mm SL ( 1 1 6 ex- 

 amined) had eaten small (< 1.0mm) copepods almost 

 exclusively. Those 30-80 mm SL (31 examined) also 

 ate copepods but included larger zooplankton such 

 as decapod zoeae, the pelagic shrimp Lucifer chacei, 

 mysids, and small fish larvae. Adult herring (243 ex- 

 amined) ate copepods less frequently and larger 

 zooplankton more frequently than did juveniles and 

 also took considerably larger prey such as chaeto- 

 gnaths, polychaetes, shrimp, and fish (Pranesus in- 

 sularum 23-25 mm SL and Stolephorus purpureus 

 6-3 1 mm SL). In one late afternoon sample of adults, 

 the stomachs were mostly packed with what ap- 

 peared to be planulae. Usually, however, both full- 

 ness and composition of prey in individual fish were 

 variable even with the same sample; some fish con- 

 tained several types of prey while others were mainly 

 full of a single type. As with Marichamy's (1970) 

 study of Herklotsichthys punctatus, we found no 

 evidence that any sizes of H. quadrimaculatus eat 

 phytoplankton. Overall, fresh prey was more fre- 

 quent in fish caught at night and empty stomachs 

 more so during the day, but fish in both conditions 

 were found in almost every sample examined. 



Gold spot herring apparently spawn mostly or en- 

 tirely outside Kaneohe Bay. No eggs or larvae were 

 found in any plankton tows taken in the bay while 

 adult sardines were present; these tows included 

 eight oblique tows taken during the peak spawning 



season in the same areas where adults had been 

 collected. The smallest fish collected or observed 

 were 17 mm SL and in the process of transforming 

 from larvae to juveniles. 



There were other movements of gold spot herring 

 both within the bay and between the bay and exposed 

 areas. Adult herring on several occasions vanished 

 from all areas of the bay for varying periods and then 

 reappeared with no obvious relation to any environ- 

 mental factor. What appeared to be the same schools 

 of juveniles were often observed in the same place for 

 several days in a row, but before there was any ob- 

 servable change in size composition, the schools 

 vanished — often to be replaced by another school of 

 obviously different- sized fish. Similarly, the size 

 composition of juveniles from purse seine catches 

 showed no coherent seasonal trend. 



Between 17 and ca. 30 mm SL, body depth of the 

 transforming herring obviously increased relatively 

 more rapidly than the standard length. Matsuura 

 (1975) showed that in Sardinella brasiliensis the os- 

 sification of ventral scales is not completed until a 

 similar size. For 50 gold spot herring 17-29 mm SL, 

 the relationship between SL (mm) and dw, (g), as 

 determined by linear least squares regression on the 

 logarithms, was 



dw, = 1.691 X 10" 9 SL 5 - 200 (r 2 = 0.94). 



The relationships calculated from data for 60 

 juveniles 31-79 mm SL and 157 adults 80-128 mm 

 SL were, respectively: 



dw, = 8.335 X 10" 7 SL 3377 (r 2 = 0.98), 



dw t = 3.462 X 10~ fi SL 3044 (r 2 = 0.90). 



These differed significantly (analysis of covariance, 

 P< 0.01) and indicated progressively more nearly 

 isometric growth with increasing size. The equations 

 for somatic weight vs. standard length differed little 

 from those for total dry weight. Somatic weights of 

 adult females tended to be lower than those of similar- 

 sized males, but the difference was not significant. 

 Wet weights of 108 fish 42-121 mm SL were related 

 to standard length by 



ww = 9.168 X 10" 6 SL 3121 (r 2 = 0.99). 



The relationship (based on linear least squares 

 regression) between ww and dw t for 123 fish 42-121 

 mm SL was 



dw, = 0.053 + 0.278 ww (r 2 = 0.98). 



590 



