FISHERY BULLETIN: VOL. 87, NO. 3. 1989 



spawning nehu fluctuates throughout the year, 

 but that there is no tendency for e.g., Mai'ch 

 spawners to be larger (or smaller) than Septem- 

 ber spawners. Egg size has been shown to be 

 both positively and negatively affected by food 

 supply or ration (Bagenal 1969; Hislop et al. 

 1978). The abundance of macrozooplankton, 

 upon which adult nehu feed, does not appear to 

 change systematically with season in Kaneohe 

 Bay (Hirota and Szyper 1976). Daily ration or 

 the fraction available for reproduction could, 

 however, change seasonally due to differences 

 in temperature or day length, but relevant 

 studies to consider this possibihty have not been 

 done. 



Imai and Tanaka (1987) demonstrated that egg 

 size of Engraulis japonica responds more di- 

 rectly to temperature changes, and several 

 potential mechanisms have been proposed. 

 Daoulas and Economou (1986) suggested that egg 

 size would be inversely correlated with tempera- 

 ture if oocyte differentiation rates increased rel- 

 ative to oocyte growth rates with increasing 

 temperature. Tanasichuk and Ware (1987) hy- 

 pothesized that gonadotropin secretion rates in- 

 crease with temperature and cause a decrease in 

 preovulatory atresia. For a given effort per 

 batch, this would result in more, but smaller 

 eggs at ovulation. Nehu fecundity is higher in 

 the summer when egg size is lowest, and the 

 increase in the summer is only partly accounted 

 for by higher effort per batch (Clarke 1987). If 

 effort per batch is controlled by some other fac- 

 tor, either of the above hypotheses could apply 

 to nehu. 



Several studies have hypothesized that both 

 within- and between-stock differences in egg size 

 are adaptive and related to minimizing mortality 

 owing to starvation of early larvae. Bla.xter and 

 Hempel (1963) showed that larvae from large 

 herring eggs were probably better able to sur- 

 vive situations with low food abundance. Ware 

 (1977) showed that egg size in Scomber sco^ti- 

 brus was positively correlated with the size of 

 food items available for newly hatched larvae. 

 There is, however, no evidence that nehu larvae 

 encounter marked seasonal changes in food 

 abundance or size. Kaneohe Bay is highly pro- 

 ductive all year; Hirota and Szyper (1976) found 

 no seasonal trends in total microzooplankton 

 abundance. My own unpubhshed data indicate 

 that concentrations and sizes of small copepod 

 nauplii, the dominant prey of first-feeding nehu 

 larvae (Burdick 1969), are similar throughout 

 the year. 



Ware (1975) and Tanasichuk and Ware (1987) 

 have shown that a mechanism resulting in a de- 

 crease in egg size with increasing temperature 

 would be selectively advantageous if egg and 

 larval mortality rates were inversely related to 

 egg size and if incubation times were inversely 

 related to temperature (and not to egg size). 

 Small nehu eggs and larvae would probably be 

 subject to higher mortality owing to predation 

 than would larger eggs and larvae at all times of 

 the year, regardless of the apparent lack of vari- 

 ability in larval food supply. Furthermore, 

 Yamashita's (1951) results indicate that differ- 

 ences in incubation time are caused by tempera- 

 ture and not egg size. Yamashita incubated nehu 

 eggs from a single sample from Kaneohe Bay at 

 different temperatures. Allowing for the fact 

 that all his eggs had abeady spent a few hours at 

 ambient temperature, the change in incubation 

 time with temperature was similar to that evi- 

 dent from the seasonal data of the present study. 

 Similar studies on other engi-aulids (King et al. 

 1978; Lasker 1964) have also shown negative 

 correlations between incubation time and tem- 

 perature that were presumably independent of 

 egg size. Thus Ware's (1975) basic assumptions 

 and hypothesis about factors selecting for egg 

 size-temperature relationships seem to be ap- 

 plicable to nehu. 



The duration of the period between hatching 

 and first feeding (HF) was also negatively corre- 

 lated with temperature. There were, however, 

 seasonal differences in the ratio of posthatch to 

 prehatch embryonic development time. HF was 

 1.60, 1.78, 2.02, and 2.14 times incubation time 

 for September, December, March, and June, 

 respectively (using HF = 47 hours for June). 

 This indicates either that the effects of tempera- 

 ture on HF are qualitatively different from those 

 on incubation, or that some other factor also 

 affects posthatch development rate. For a range 

 of temperatures that included those observed in 

 this study, Houde (1974) showed that increasing 

 temperature caused decreases in the period be- 

 tween hatching and eye pigmentation of larval 

 Anchoa mitckelli. To the extent that seasonal 

 differences in HF of nehu are similarly caused by 

 temperature differences, this would tend to aug- 

 ment any selective advantage for larger eggs 

 and larvae during the winter. 



Even though Ware's (1975) hypothesis could 

 potentially apply to nehu, the point in the diel 

 cycle at which larval feeding becomes possible 

 could have more important consequences to lar- 

 val survival than differences in either egg size or 



598 



