FISHERY BULLETIN: VOL. 83. NO. 3 



however, PL-l's settled in spring-summer or in 

 summer-fall. These differences in the pattern of 

 settlement between the census sites are highly signi- 

 ficant (P « 0.001, "distribution-free" nonparametric 

 test; Hollander and Wolfe 1973, p. 139-146). The 

 patterns were so consistent, however, that we were 

 soon able to predict with considerable success not 

 only when, but which sites would attract PL-l's. We 

 can provide no correlations with depth or area of a 

 site, nor with the species of corals and other major 

 residents (fishes, sea urchins, etc.), as to why some 

 sites consistently "attracted" postlarval grunts and 

 others varied, even though it is known that post- 

 larval settlement can be partly governed by priority 

 effects on reefs (Shulman et al. 1983). To understand 

 the characteristics of a site that make it highly 

 "attractive" to settling postlarval grunts will require 

 sophisticated field manipulations. 



DISCUSSION 



The most conspicuous features concerning the 

 recruitment of French grunts from the plankton are 

 the rhythms in settlement and fertilization dates 

 (Figs. 2-4). The continuing and short 15-d periods of 

 these rhythms over an entire year contrast with 

 other coral reef fishes where settlement, although 

 rhythmic, is concentrated mostly around a particular 

 period of the year (Williams and Sale 1981). Settle- 

 ment in reef fishes is often episodic; it may show a 

 lunar periodicity (Johannes 1978), it may be monthly 

 but not coupled to a particular phase of the moon 

 (Williams 1983 -mixed guild of pomacentrids), or a 

 rhythm may not be conspicuous (Victor 1982, 1983 - 

 Thalassoma bifasciatum). 



The short period between fertilization and settle- 

 ment implies that pelagic existence in French grunts 

 is of short duration (about 2 wk; Brothers and 

 McFarland 1981). To estimate the days spent in the 

 plankton, the average age for the smallest 100 

 PL-l's was calculated from the sample of 1,478 fish 

 that were used to estimate the actual day of fertiliza- 

 tion [mean age = 15.7 d ± 2.1 (SD). range = 13-20; 

 mean length (SL mm) = 6.9 ± 0.67 (SD), range = 

 5.9-8.5]. If each of these fish had settled from the 

 plankton over the previous day, then pelagic exist- 

 ence (about 15 d) agrees with the periodicity of ferti- 

 lization and settlement (Table 3, Fig. 4). In contrast, 

 the pelagic existence of most other coral reef fishes 

 investigated exceeds 15 d (Randall 1961; Johannes 

 1978; Sale 1980; Bariow 1981). This has been veri- 

 fied by results for age at settlement as determined 

 from otoliths (Victor 1982, 1983; Brothers et al. 

 1983; Brothers and Thresher in press and unpublish- 



ed; Thresher and Brothers in press). Most of these 

 fishes settled at various ages: Thalassoma bifascia- 

 tum, 40-72 d; Paragobiodon melanosoma, 39-47 d; 

 Gobiodon sp., 27-38 d; several unidentified scarids, 

 34-58 d; and lab rids, 21-56 d. Direct aging of new 

 recruits or otolith counts to presumed settlement 

 marks rarely yield ages as low as found for the 

 French grunt. Examples of species with pelagic 

 phases of 20 d or less include the angel fish, Holacan- 

 thus paru (Brothers and Thresher in press); several 

 damselfishes, Glyphidodontops rollandi, Pomacen- 

 fyits amboinensis, P. popei, and P. wardii; the blenny 

 Petroscirtes mitratus; the nemipterid Scolopsis 

 dubiosus (Brothers et al. 1983); and the goby Gobio- 

 soma prochilos (Brothers, unpubl. data). These may 

 be exceptional cases; all or at least most of these 

 species represent families characterized by having 

 larval durations moderately to considerably longer 

 than found in grunts. At the family level, therefore, 

 only the haemulids, and perhaps the lutjanids, pro- 

 vide evidence of a short pelagic existence. In support 

 of this conclusion is the lack of haemulids or lutjanids 

 amongst the larval fishes collected offshore through- 

 out the Caribbean (Richards 1981, footnote 8). Are 

 haemulids programmed for short larval lives? We 

 cannot be sure, but out of the 1,478 recently settled 

 French grunts we have aged, not one exceeded 20 d 

 from fertilization. Such a developmental process 

 would be crucial in their survival; if they do not drift 

 over suitable substrates on which to settle after 2 

 wk, they would perish offshore. 



Our observations of a strong semilunar periodicity 

 in French grunt recruitment, coupled with what 

 seems to be a relatively fixed or invariant larval 

 duration, could be the result of a number of different 

 combinations of spawning and survivorship. Because 

 we can only determine fertilization dates for indivi- 

 duals that have successfully recruited, we cannot be 

 certain whether the apparent periodicity in spawn- 

 ing is an accurate representation of the temporal pat- 

 tern of grunt reproduction. At the other extreme, it 

 may be the result of relatively continuous spawning 

 activity, the products of which survive differentially 

 with respect to semilunar environmental variables. 



Surprisingly, for such common fishes, little is 

 known about spawning in haemulids. They produce 

 pelagic eggs and larvae (Breder and Rosen 1966; 

 Saksena and Richards 1975). Recently, paired 

 spawning has been described for Hapalogenys 

 mucronatus in aquaria (Suzuki et al. 1983). On six 



*W. Richards, Southeast Fisheries Center, National Marine 

 Fisheries Service. NOAA. 75 Virjcinia Beach Drive, Miami. P^L 

 33149-1099, pers. commun. October 1980. 



422 



