tion (n = 12, range = 0.5-5.0 days). These observa- 

 tions show 1) that females ovulate in the absence 

 of a male, but 2) delay their ovulations under those 

 conditions; and 3) such eggs disappear from the 

 brood pouches a few days after ovulation. 



The introduction of a male to a female any time 

 during the approximately 3-d period that females 

 could delay their ovulations stimulated ovulation. 

 Copulations generally occurred within 5 minutes 

 after males were introduced to females 0, 1, or 2 

 days past their molts, and all ovulations occurred 

 within 3 hours of copulation. There was no signif- 

 icant difference in the frequencies of copulations and 

 ovulations among females isolated for different 

 lengths of time (Table 3; xl = 0.008, P > 0.05 and 

 xl = 0.712, P > 0.05 for copulations and ovulations, 

 respectively). 



In contrast, the broods of females who copulated 

 and ovulated and 1 day after their molts devel- 

 oped significantly more often than did the broods 

 of 2-d postmolt females (Table 3;xi = 3.61, P< 

 0.05). This suggests that unfertilized eggs aged past 

 2 days have reduced viability. 



Table 3.— Frequencies of copulations, ovulations, and viable 

 broods of isolated females introduced to males 0, 1, and 2 days 

 past the females' molts. 



Discussion 



The present study has shown that males and 

 females remain apart until towards the end of the 

 female intermolt period, when amplexus is begun. 

 This continues until the female molts, when copula- 

 tion, followed by ovulation occtirs. Females repeated 

 this cycle in the laboratory until they died. Some 

 females produced three broods in succession, with- 

 out a rest period. In contrast, females in the field 

 in Nova Scotia produce only one brood (Strong 

 1978), and in the northern Baltic most have one 

 brood (although some may have two (Salemaa 

 1979)). 



It is impossible to say whether the difference in 

 the number of broods produced in the field and in 

 the laboratory is caused by environmental and/or 

 genetic differences. In support of the first explana- 

 tion, a recent symposium has demonstrated the im- 



portance of photoperiod in regulating the timing of 

 reproductive activities of marine animals (Marcus 

 1986), and, specifically, in the number of broods of 

 some peracarids (Steele and Steele 1986). The photo- 

 period of Nova Scotia may limit the number of 

 broods there. In support of the second explanation, 

 Healy and O'Neill (1984) noted that two populations 

 of /. granulosum produced different numbers of 

 broods in Ireland and in Britain, although there was 

 no significant difference in temperature or latitude 

 at the two locations. Further, while Tinturier- 

 Hamelin (1963) found no reproductive barriers 

 among /. baltica from different parts of Europe that 

 were hybridized in the laboratory, she did conclude 

 that the subspecies were genetically distinct. Only 

 laboratory culture, under identical conditions, of 

 representatives from different geographic localities 

 will reveal whether the difference in the number of 

 broods is genetic. 



The present observations also reveal that /. baltica 

 females spend relatively little of their intermolt 

 periods in heterosexual pairs (14%) compared with 

 other peracarids (the mean of seven other peracarids 

 was 32% (Borowsky 1986)). Strong (1978) reported 

 that a short amplexus period was correlated with 

 a high risk of fish predation in the amphipod 

 Hyallela azteca and suggested that couples are more 

 visible than single individuals. This may be the case 

 here as well. Fish and predaceous shrimp are abun- 

 dant in the community from which individuals for 

 the present study were collected (D. Franz pers. 

 commun.^). 



Both sexes cast off the posterior parts of their exo- 

 skeleta before the anterior parts. Since the female's 

 anterior portion includes most of the brood pouch, 

 the new marsupium is not completely exposed un- 

 til both parts are shed. Males did not copulate with 

 females that had not molted the anterior cast. The 

 sequence of molts and copulation makes sense, 

 because the present study shows that copulations 

 stimulate ovulations. If copulation occurred before 

 the anterior portion of the pouch were shed, the new 

 eggs might be secreted into the old brood pouch, and 

 then discarded along with the anterior molt. 



The present study reveals some flexibility in the 

 timing of key reproductive events. Females could 

 copulate until about 36 hours after they cast off their 

 anterior molts without incurring a significant reduc- 

 tion in fecundity. However, if copulation did not 

 occur by about 3 days after their molts, females 



'D. Franz, Department of Biology, Brooklyn College, City 

 University of Nevi^ York, Bedford Avenue and Avenue H, Brook- 

 lyn, NY 11210, pers. commun. August 1986. 



379 



