McFAKl.ANI) KT Al,.; KKCRUITMKNT PATTERNS IN FRKNCll CKUNTS 



familial Anisotremus virginicus, of known age, sup- 

 plied by Martin Moe). Thus, to each "otolith age" ( = 

 total increments counted) 2 d were added to establish 

 the "absolute" daily age of an individual fish. 



This method was applied to postlarval French 

 grunts, collected throughout the year in the vicinity 

 of the 20 discrete census sites. On most census days 

 individual PL-l's were collected with a fine mesh net 

 and fixed and preserved in 95% ethanol. Each post- 

 larva was measured (SL) and the otoliths were 

 removed and placed in immersion oil. The number of 

 growth increments was counted and corrected ( -i- 2 

 d) to the actual date of fertilization. 



Tides 



A tidal gauge (NOAA, Ocean Survey #9751224) 

 operates at the West Indies Laboratory dock, about 

 0.5 km from our Tague Bay study site. Hourly tidal 

 heights for the year 1980 were obtained from 

 NOAA. Missing, due to malfunction of the gauge, 

 are records from 26 June to 9 September 1980. 

 These missing values were approximated from calcu- 

 lated tidal data for San Juan, Puerto Rico. The phase 

 of the tides in Puerto Rico matched closely the tides 

 at Tague Bay (comparisons of dates before and after 

 the missing records), but the actual excursion of the 

 tides was less at St. Croix than calculated for Puerto 

 Rico. 



RESULTS 



Tague Bay Sandhill Study Site 



During 1979, 27 censuses were executed during 

 300 possible days of sampling (25 February to 21 

 December). The mean interval between censuses 

 was 1 1 .2 d ± 6.6 ( 1 SD), the intervals ranging from 4 

 to 23 d. Five population peaks were recorded, with 

 the number of PL-l's counted between peaks often 

 declining to < 100 individuals. The 1979 census 

 clearly indicated that settlement was represented by 

 a series of pulses, but the sampling intervals were 

 too long to resolve periodicities of much less than 1 

 mo. Therefore, in 1980 the sampling resolution was 

 improved by increasing the number of censuses to 47 

 over a possible 325 sampling days (6 February to 26 

 December 1980); mean sampling interval was 7.0 d 

 ±7.7, the shortest interval being 1 d and the longest 

 interval 49 d (i.e., the first interval). Eleven peaks, of 

 which 8 are distinct, establish that the postlarvae 

 settle in pulses, with the population numbers on the 

 bay bottom often decreasing to between pulses 

 (Fig. 1). Although the pulses in population numbers 



1950 



Figure 1.- Numbers of recently settled postlarval grunts observed 

 over 800 m of bottom on the sandhill site area of Tague Bay, St. 

 Croix, V.I. 



are suggestive of a semilunar periodicity, the peaks 

 do not coincide more closely with either quarter 

 moons or with the new and full moons than do the 

 population declines. 



Individual Barrier Reef Study Sites 



The mean sampling interval on the 20 individual 

 census sites from 6 May to 27 December 1980 (possi- 

 ble 236 sampling days) was 1.37 d ± 0.7 (1 SD). This 

 shorter sampling interval vividly reveals the period- 

 icity in the appearance of PL-l's on the sites (Fig. 2). 

 Coincidence between these peaks and the peaks for 

 the sandhill study sites is quite good (compare 

 Figures 1 and 2) and indicates that the timing of the 

 settlement of PL-l's is general over the entire area. 

 Also, as in the sandhill area, the numbers of PL-l's 

 declined to between most pulses. Furthermore, the 

 pooled data for the 20 sites indicate a bimodal 

 seasonal influx of PL-l's, one in late spring and a sec- 

 ond increase in fall, as seen also in the sandhill study 

 site (compare Figures. 1 and 2). The summer settle- 

 ment pulses involved about one-third the number of 

 individuals associated with the bimodal peaks. 

 Winter settlement was very low; only small numbers 

 of PL-l's were observed during January and 

 February 1981. 



Age and Reproductive Activity 



A total of 2,353 postlarvae from 141 collections 

 spread over 391 d were captured between 20 Febru- 

 ary 1980 and 16 March 1981 (Fig. 3). From each of 

 these 141 samples 15 PL-l's were measured and 

 aged, or fewer if the collected samples contained 



415 



