FlSHKKVBrLLF.TI\;V()l., h:1N().3 



mental fish responded to laboratory conditions by 

 reducing batch fecundity somewhat, but spawning 

 more frequently, and thereby producing about twice 

 the number of eggs as in nature (Table 1). The daily 

 rate of egg production was 24-47 eggs/g female body 

 weight per d in the laboratory, but averaged about 

 14 eggs/g female body weight per d in the field. 



At the termination of the experiment, four of the 

 eight females in the outdoor pools were spent, three 

 contained only recruitment eggs, and one had both 

 recruitment and maturing eggs. Female A died of 

 unknown causes after its last spawning on 22 July. 

 Female C also died (9 June) before cessation of 

 spawning by jumping out of the tank. 



The total weight specific egg production for the 

 experimental fish was generally within the range of 

 total eggs available prior to the beginning of the 

 breeding season. The one exception was female A 

 which produced about twice the total number of eggs 

 that a fish of its size should have had available at the 

 beginning of the spawning season (see Figure 3). 

 Hence, under certain laboratory conditions, females 

 may be capable of producing new oocytes from 

 oogonia during the breeding season, as recruitment 

 eggs become depleted. These laboratory observa- 

 tions show that the reproductive patterns of egg 

 maturation and spawning which are highly synchro- 

 nized with and influenced by environmental factors 

 in the field, easily become disrupted when individuals 

 are removed from their natural habitat. 



CONCLUSIONS 



This study indicates that annual fecundity in 

 Menidia menidia, and perhaps certain other fishes, 

 can be estimated from the difference between total 

 number of eggs (recruitment plus maturing) prior to 

 the breeding season and recruitment egg retention 

 near the end of the breeding season. Dividing the 

 estimated total number of eggs shed per female by 

 mean batch fecundity provided an estimate of 

 spawning frequency. The accuracy of this value for 

 spawning frequency was tested and found to agree 

 closely with the spawning fre(]uency inferred from 

 direct field observations of breeding fish. Previous 

 estimates of the fecundity of M menidia were about 

 3-10 times less than that rejwrted here because 

 spawning frequency was not determined (Bayliff 

 1950; Jessop 1983). The studies of Hunter and his 

 coworkers on northern anchovy, Engraulis mordax 

 (Hunter and Goldberg 1980; Hunter and Macewicz 

 1980; Hunter and Leong 1981), and DeMartini and 

 Fountain (1981) on queenfish, Seriphus politus, have 

 amply demonstrated that estimates of annual fecun- 



dity can be in error by over an order of magnitude 

 when multiple spawning is ignored. 



The estimation of fecundity from the difference be- 

 tween total prespawning fecundity and recruitment 

 egg retention is dependent on the assumption that 

 new oocytes are not simultaneously produced from 

 oogonia and added to the reservoir of recruitment 

 eggs as mature eggs are spawned. Agreement be- 

 tween predicted and observed spawning frequency 

 suggests that this may be true in M. menidia. Many 

 more recruitment eggs were present in ovaries at 

 the beginning of the spawning season than were ac- 

 tually spawned in nature. Evidently, the recruitment 

 egg pool is largely formed before the breeding 

 season in Menidia, as is believed for some other 

 seasonal spawners (Tokarz 1978; Jones 1978; 

 Baggerman 1980). However, the generality of this 

 pattern in other multiple spawning temperate or 

 tropical fishes is not clear. Clark (1925) noted that 

 the relative abundance of mature, intermediate, and 

 immature eggs in Leuresthes tenuis was relatively 

 constant during the breeding season and concluded 

 from this that new oocytes must be continuously pro- 

 duced to replenish those spawned. Taylor and 

 DiMichele (1980) reached a similar conclusion based 

 on the relative abundance of different developmental 

 stages of oocytes during the semilunar spawning 

 cycle of Fundulus heteroclitiis. However, analyses 

 based on relative proportions do not take into ac- 

 count that gonad weight (GSI) generally declines as 

 the season progresses (e.g., Fig. 2) and that number 

 of eggs in the most advanced mode is not necessarily 

 constant during the breeding season. Comparison of 

 the relative abundance of egg sizes from sections of 

 an ovary may not reflect changes in absolute 

 number. For example, the relative abundance of 

 recruitment eggs in M. menidia during 1977 was 

 0.88 on 6 May, 0.78 on 6 June, 0.76 on 22 June, and 

 0.79 during 6-13 July. Hence, the relative proportion 

 of recruitment eggs did not consistently decline dur- 

 ing the breeding season even though the absolute 

 number of eggs declined by a factor of 2.4. In any 

 event, too little is known about patterns of oocyte 

 growth in fishes to recommend that the annual 

 fecundity of multiple spawners can generally be 

 determined by monitoring the decline in the standing 

 stock of ova as was done here. For instance, in 

 tropical species that breed most of the year recruit- 

 ment eggs may be produced continuously. Whenever 

 possible, the results of several different api)roaches 

 to estimating fecundity should be compared. 



The results of the laboratory study demonstrated 

 that M. menidia is physiologically capable of spawn- 

 ing much more frequently and over a shorter interval 



338 



