None of the female fish sampled on the May spawn- 

 ing peaks had undeveloped gonads (individual gonad 

 indices were all > 10%), whereas many had 

 undeveloped gonads on the July peaks (38% of the 

 sample for gonad weights had individual indices 

 < 5%, and 25% of the sample tested for spawning 

 were not ripe). Second, gonad size of fish with de- 

 veloped gonads declined during the season. Mean 

 gonad indices of females with developed ovaries of 

 the May peaks were significantly greater than those 

 of the July peaks (15% versus 9%, respectively) 

 (Student-Newman-Keuls test of arcsine transformed 

 data, P < 0.05). 



Discussion 



In Fundulus heteroclitus and F. conjluentus, 

 regulation of the annual reproductive cycle is ap- 

 parently dependent on changing combinations of 

 photoperiod and temperature (Harrington 1959; Day 

 and Taylor 1982). The decline in gonad index during 

 the season in the present study is consistent with a 

 similar decline observed by Kneib and Stiven (1978) 

 and Taylor et al. (1979) and suggests that, along with 

 physical factors, energy availability may be limiting 

 reproductive output late in the season. Weisberg 

 (1981) found that supplemental food increased the 

 gonad indices of F. heteroclitus in Delaware salt 

 marshes. Histological evidence shows that primary 

 oocytes in F. heteroclitus can undergo complete 

 development in one biweekly spawning cycle (Taylor 

 and DiMichele 1980), so lag time for recrudescence 

 of spent gonads does not account for reduced repro- 

 ductive activity late in the season. Funduhis grandis 

 reproduces throughout the summer in some areas of 

 the gulf coast (Greeley and MacGregor 1983), but 

 shows bimodal reproductive activity in spring and 

 fall with no reproduction in midsummer in shallow 

 ponds (Waas and Strawn 1983). 



In addition to F. heteroclitus (Taylor et al. 1979; 

 Taylor and DiMichele 1980), semilunar spawning 

 cycles occur in some populations of F. grandis 

 (Greeley and MacGregor 1983), F. similis (Greeley 

 1982), F. majalis (Tedesco et al."), and probably inF. 

 notatus (inferred from observations of spawning 

 behavior in Carranza and Winn 1954). Waas and 

 Strawn (1983) measured a weak lunar, but not a 

 semilunar cycle in populations of F. grandis in two 

 nontidal ponds and a tidal creek with lunar tides fre- 

 quently overridden by wind effects. The semilunar 



^Tedesco, M., A. H. Hines, and L. A. Wiechert. 1983. Semi- 

 lunar gonadal cycles mFunduliis maja/i-s (Pisces: Cyprinodontidae). 

 Technical Report of Smithsonian Environmental Research Center, 

 P.O. Box 28, Edgewater, MD 21037. 



spawning cycle in F. grandis has been induced in the 

 laboratory during their reproductive season by inter- 

 action of photoperiod and tidal changes in water 

 temperature shifting on a lunar cycle (Miller et al. 

 1981). Daily disturbances (netting) of F. grandis also 

 apparently induces circadian gonadal responses 

 (Weld and Meier 1982). Tidal rhythms in water tur- 

 bulence and hydrostatic pressure have been postu- 

 lated for lunar and semilunar reproductive rhythms 

 in fish (Korringa 1947; Ross 1983) and for semilunar 

 hatching in intertidal chironomid insects (Newmann 

 1978). Middaugh (1981) suggested that the biweekly 

 coincidence of a high tide at the time of sunrise 

 and/or a lunar cue may serve as a synchronizer for 

 spawning of Menidia menidia in a South Carolina 

 estuary. In contrast to M. menidia, which showed a 

 precise reproductive response to diurnal tidal and 

 lighting schedules (Middaugh and Takita 1983), M. 

 peninsulae exhibited a variable and labile response in 

 which a combination of semidiurnal interruptions of 

 current and diel light cues was optimal at inducing 

 spawning synchrony (Middaugh and Hemmer 1984). 

 However, this synchrony in Menidia spp. had a diel 

 period, and the laboratory experiments were not run 

 long enough to test for lunar rhythms. Semilunar 

 rhythms of larval hatching are well documented for 

 several species of intertidal and estuarine crabs (see 

 Christy 1982; Forward et al. 1982). Hatching in the 

 mud crab, Rhithropanopeus harrisii, followed a 

 semilunar and diurnal rhythm in populations from an 

 estuary with semidiurnal and lunar tidal cycles, 

 whereas hatching was not associated with lunar 

 rhythms or tidal levels in an estuary with aperiodic 

 tides (Forward et al. 1982). Moveover, Forward et 

 al. (1982) induced circatidal rhythms in larval release 

 in crabs from an estuary with irregular tides by 

 transplanting them to an estuary with semidiurnal 

 and lunar tides. However, Saigusa (1980) showed 

 that the larval hatching cycle of a semiterrestrial 

 crab Sesarma haematocheir is entrained directly by a 

 moonlight cycle. 



In the present study, in an area without a lunar 

 periodicity in the tidal cycle, F. heteroclitus had a 

 semilunar reproductive cycle which lagged the new 

 and full moons by 3.5 d. Although the study area had 

 tidal fluctuations with an approximately semidiurnal 

 period, the diel timing of high tides with respect to 

 photoperiod also did not appear to cue the semilunar 

 reproductive cycle. Changes in currents associated 

 with the diurnal tidal cycle may serve as a synchro- 

 nizing cue for the time of day of spawning, as it does 

 for Menidia menidia and M. peninsulae. However, 

 while identity of the cue remains unknown, the pres- 

 ent study indicates that tidally mediated factors such 



470 



