234 



Fishery Bulletin 89(2). 1991 



Ovaries of the three species of Spratelloides and A. 

 zosterophora all contained only a single size-group of 

 developing oocytes, which suggests that they spawn 

 all the eggs in the ovaries at once. Whether a female 

 develops another batch of eggs after spawning was not 

 determined. However, given the high proportion of 

 female Spratelloides spawning at any time, it seems 

 probable that each female produces more than,one 

 batch of eggs. Most apogonids are mouth-brooders 

 (Thresher 1984), and other species have been found 

 with eggs in their buccal cavity that are at more than 

 one stage of development, which suggests that they are 

 multiple spawners (Thresher 1982). 



Flexibility in reproduction is well documented among 

 a range of animals and usually involves a direct physi- 

 ological response to nutrient level or some associated 

 environmental cue (Giesel 1976). Extended breeding 

 seasons are a common phenomenon among tropical 

 fishes, particularly coral reef fish (Munro et al. 1973, 

 Russell et al. 1977, Johannes 1978, Lowe-McConnell 

 1979, Walsh 1987). The present work, and other studies 

 of Encrasicholina, Spratelloides, and apogonid repro- 

 duction (Leary et al. 1975; Russell et al. 1977; Dalzell 

 and Wankowski 1980; Dalzell 1985, 1987ab; Conand 

 1985; McCarthy 1985; Clarke 1987), found that the 

 spawning season of these fish is protracted, with some 

 individuals in the population spawning at any time dur- 

 ing the year. There are also periods when most of the 

 population is spawning. If reproduction were controlled 

 only by endogenous cycles and females spawned as 

 soon as they were physiologically capable, then as the 

 proportion of larger fish increased there should be a 

 greater proportion of spawning fish in the population. 

 Examination of our data failed to find any significant 

 relationship between the proportion of larger fish and 

 the proportion spawning for any species at any site. 

 Non-parametric tests also showed that the frequency 

 of major spawnings was not random. Hence it is unlike- 

 ly that spawning periodicities shown by baitfish are a 

 result of intrinsic mechanisms. The timing and inten- 

 sity of periods of increased spawning activity appear 

 to be highly variable, linked to exogenous stimuli and 

 with no endogenous rhythms. 



Of the several hypotheses put forward to explain the 

 timing of fish reproduction, one of the more widely ac- 

 cepted is the "match-mismatch" hypothesis of Cushing 

 (1967). This proposes that if the timing of reproduc- 

 tion coincides with peaks in the plankton cycle, larval 

 survival is enhanced. Timing of peak plankton produc- 

 tion does not necessarily occur at exactly the same time 

 each year, so if fish are serial spawners they can ex- 

 ploit the plankton cycle well (e.g., California sardine) 

 by spawning during any of the months of spring or sum- 

 mer when food becomes superabundant (Cushing 1975). 

 Much of the data to support this hypothesis comes from 



temperate waters, but data on the reproduction of 

 tropical clupeoids from open oceans also support this 

 hypothesis (Longhurst 1971, Roy et al. 1989). 



Johannes (1978) proposed that coastal tropical fish 

 tend to spawn at times and locations that will reduce 

 predation on larvae by transporting them out of the 

 adult habitat, while enabling them to return to suitable 

 areas for postlarval settlement. At many sites, periods 

 of major spawning coincide with low winds and full 

 moon. The baitfish species studied here spawn in 

 lagoons fringed by coral reefs, where water currents 

 are negligible. Both adults and larvae are pelagic and 

 live in the deeper waters of the lagoon. When the lar- 

 vae are abundant, adults prey on them (Milton et al. 

 1990). If deeper waters of the lagoon are the favored 

 habitat, then according to Johannes' hypothesis (1978) 

 adults should spawn when local flushing is greatest 

 (spring tides) but regional flushing is least (low winds) 

 to ensure larvae develop in the same area. 



Although baitfish spawn throughout the year, with 

 no consistent seasonal pattern, there appears to be 

 some regularity in their spawning. At each site, cer- 

 tain proximate factors were significant for more than 

 one species. This suggests that if these fish are re- 

 sponding to exogenous spawning stimuli, their influ- 

 ence must be interposed by local hydrography and 

 topography. At Vona Vona, spawning by five of the 

 six species examined was correlated with moon phase. 

 Both Encrasicholina and two Spratelloides species 

 spawned around the full moon, while S. lewisi spawned 

 around the new moon. Both Encrasicholina species 

 also spawned when wind strength was reduced. The 

 spawning around phases of the moon has often been 

 assumed to be related to increased tidal exchange dur- 

 ing spring tides (e.g., Johannes 1978, Walsh 1987). 

 However, in this study greater tidal range at any site 

 was not directly correlated with moon phase (Table 3). 

 Vona Vona is an enclosed lagoon with many islands and 

 patch reefs. In this habitat, spawning around the full 

 or new moon does not appear to be related to the poten- 

 tial to flush eggs and larvae to more favorable habitats 

 by tidal water movement (Johannes 1978). Possibly, 

 spawning around the full moon when the wind is re- 

 duced may increase spawning success by increasing 

 visibility at night (during spawning) and reducing the 

 dispersal of eggs, and hence increasing the chances of 

 fertilization. This hypothesis would also explain the 

 strong negative relationship between spawning by A. 

 zosterophora and tidal range at this site. 



Moon phase was the single variable most often cor- 

 related with spawning in this study, both for five spe- 

 cies at one site (Vona Vona) and for one species at dif- 

 ferent sites (S. delicatulus). However, our data do not 

 indicate a strong relationship with moon phase for 

 either Encrasicholina species at Munda or Tulagi. This 



