Patterns of Lunar Settlement and Early Recruitment in Caribbean Reef Fishes at Panama

Abstract

Abstract: Lunar patterns of settlement of 15 Caribbean reef fishes (Stegastes diencaeus, Stegastes dorsopunicans, Stegastes leucostictus, Stegastes parititus, Stegastes planifrons, Stegastes variabilis, Abudefduf saxatilis, Acanthurus bahianus, Acanthurus chirurgus, Acanthurus coeruleus, Thalassoma bifasciatum, Chaetodon capistratus, Pomacanthus paru, Holacanthus ciliaris, Canthigaster rostrata) were assessed from daily collections of newly arrived fishes from each of two small (A = 63 and B = 28 m-2) patch reefs at Punta de San Blas (Panama) during two consecutive 2 yr periods (1984, 1985 and 1986, 1987). In 12 species settlement was lunar-cyclic, with a broad peak of activity around the new moon in 11 cases, and at first quarter in 1 species. However, 3 of these "lunar-cyclic" species displayed intermittent semilunar periodicity in settlement, with peaks around the quarter moons. One other species had a semilunar cycle of settlement with peaks at both quarter moons. Two species apparently lacked lunar settlement cycles. In some, but not all, pairs of congeners: (a) monthly variation in settlement intensity was positively correlated, (b) the duration of settlement pulses in the same month consistently differed, and (c) the timing of settlement pulses in the same month often differed (although one species did not consistently arrive before the other). Recruitment of juvenile fish that survived to the end of the lunar cycle in which they settled was monitored monthly, concurrently with settlement, at a number of large sites scattered up to 2 km apart. The levels of variation in the amount of settlement and recruitment each month differed in only 2 of 17 cases, with settlement variation exceeding recruitment variation in only one of these. The intensity of recruitment was positively correlated with the intensity of settlement in 16 of 17 cases. In 15 of those 16 cases, R-2 values for linear regressions of recruitment on settlement exceeded R-2 values for curvilinear regressions that would indicate either increasing or decreasing recruitment success with increasing intensity of settlement. Thus it appears that variation in recruitment can be used to estimate variation in settlement, and that recruitment success may be density-independent.