rONO\ KK I'AITKKNS IN KKdNDI TV I )K ATLANTIC SIlAKkSlDK 



than normally occurs in the field. The reasons for the 

 higher spawning frequency an(i cumulative egg pro- 

 duction for fish in captivity are probably several. 

 Fecundity may have been increased because ration 

 size was unlimited. Fecundity is dependent on the 

 food supply in many species (Wootton 1979). In the 

 stickleback, Gasterosteus aculeatus, (Wootton 1977) 

 and the convict cichlid, Cichlasoma nigrofasciatum, 

 (Townsend and Wootton 1984) experimental studies 

 have demonstrated that the number of spawnings 

 was positively related to food ration and the interval 

 betu'een spawnings was inversely related to ration. 



In my experiments on M, nienidia, spawning fre- 

 quency may also have been increased beyond that in 

 nature due to the continuous availability of appro- 

 priate spawning substrates and lack of tidal spawn- 

 ing cues in the laboratory. Conover and Kynard 

 (1984) noted that both marine and freshwater 

 populations of Menidia spp. tend to spawn during 

 midmorning, and speculated that spawning in nature 

 may be restricted by the fact that suitable spawning 

 substrates are covered by high tide during midmorn- 

 ing only at fortnightly intervals. Correspondingly, a 

 lacustrine population of M. beryllina spawns daily at 

 midmorning (Hubbs 1976). Hence, in the laboratory 

 where tidal cues are removed, spawning substrates 

 are continuously available, and food is abundant, M. 

 menidia responded by spawning more frequently. 

 The high egg production of female A also suggests 

 that if the supply of recruitment eggs is exhausted, 

 new recruitment eggs can be formed. It is clear that 

 estimates of fecundity in natural populations of 

 multiple spawners based on laboratory studies alone 

 should be interpreted with caution. 



Many aspects of the fecundity and spawning 

 periodicity of M. menidia are paralleled in a west 

 coast atherinid, Leuresthes tenuis. The California 

 grunion has a well-known semilunar spawning cycle 

 (Walker 1952). Clark (1925) conducted a detailed 

 study of egg diameter frequencies in ovaries of L. 

 tenuis and concluded that each female spawns once 

 about every 15 d. Batch fecundity was very similar 

 to that reported here for M. menidia. Although 

 Clark measured batch fecundity in only a few indivi- 

 duals, a 118 mm grunion contained 1,613 ova. I 

 calculate that a 118 mm Atlantic silverside would be 

 expected to have 1,704 ripening eggs during the mid- 

 dle of the breeding season. Clark also found reten- 

 tion of recruitment eggs at the end of the breeding 

 season and presented histological evidence that re- 

 tained eggs were being resorbed. 



Based on my estimate of the average annual fecun- 

 dity of M. menidia (893 ± 197 eggs/g ovary-free 

 body weight) and the wet weight of ripe eggs (0.8 



g/1,000 eggs), an Atlantic silverside produces nearly 

 0.7 of its body weight in eggs during the breeding 

 season in nature. In the laboratory, females pro- 

 duced 1.1-2.7 times their body weight in eggs. 

 Studies of other multi[)le spawners have yielded 

 similar results. DeMartini and Fountain (1981) esti- 

 mated that the queenfish could spawn about 114% of 

 its body weight in a year. Experiments on several 

 species of cyprinids indicate that they are capable of 

 spawning 0.7 to 6.8 times the volume of the female in 

 eggs, at least in the laboratory ((lale and Gale 1977; 

 (iale and Buynak 1978, 1982; Gale 1983). Hubbs 

 (1976) estimated that a freshwater population of 

 Menidia beryllina spawned 6-8 times female weight 

 in eggs, although his assumption that each female 

 spawns daily throughout the length of the breeding 

 season needs further documentation. 



Subseasonal trends in batch fecundity among 

 multiple spawners have been examined by few inves- 

 tigators. If trends in batch fecundity within the 

 breeding season are the adaptive result of natural 

 selection, then periods of maximum batch fecundity 

 should reflect the period when the probability of off- 

 spring survival is greatest. On the other hand, trends 

 in batch fecundity could simply result from varying 

 food conditions for adults. Three general relation- 

 ships between the batch fecundity and the time of 

 the breeding season have emerged from field studies 

 with which I am familiar. These include 1) constant 

 batch fecundity during the breeding season (Fig. 6, 

 curve A), 2) a concave downward relation between 

 batch fecundity and the breeding season (Fig. 6, 

 curve B), and 3) a constant decline in batch fecundity 

 during the breeding season (Fig. 6, curve C). Con- 

 stant fecundity (curve A) might be expected where 

 the optimal environmental conditions for reproduc- 

 tion and offspring survival are constant or vary un- 

 predictably during the breeding season. This pattern 

 has been found in the queenfish (DeMartini and 

 Fountain 1981), a pelagic spawner of the western 

 North American coast where aperiodic upwelling 

 events produce unpredictable variations in plankton 

 productivity and potential larval survival (Lasker 

 1978). When seasonal environmental conditions 

 change in a predictable manner, there may be an op- 

 timal period for reproduction that occurs at roughly 

 the same time each year, and batch fecundity would 

 be expected to be maximal at that time (curve B). In 

 M. menidia, the relation between batch fecundity 

 and the breeding season was concave downward, 

 suggesting that reproductive success is maximal dur- 

 ing the middle of the breeding season. There is some 

 independent evidence to support this hypothesis. 

 Winter mortality during the offshore migration is 



339 



