FACTORS REGULATING THE DISTRIBUTION AND POPULATION DYNAMICS OF rrHE AMPHIPOD GAMMARU..5 PALUSTRIS

Ahstract. Seasonal changes in the intertidal distribution, abundance and population dynamics of the epifaunal amphipod C]lllllllillrus palustris were studied in salt marshes bordering 2 estuarine rivers flowing into the Chesapeake Bay. The amphipod populations inhabiting the Patuxent River study site showed abundance peaks during the spring and autumn, and nlajor declines in density during the summer and winter. Populations at 2 study sites in the Rhode River showed only 1 peak of abundance during the late spring and early summer which was followed by a decline in density throughout the renlainder of the year. In both rivers, the low amphipod densities observed during the winter corresponded with a subtidal nligration. Migrations did not account for the low nunlbers observed at other tinles and an examination of the life cycle of this species as well as an egg-ratio analysis of the populations indicated that low densities during the sunlmer (Patuxent) or late summer and fall (Rhode) were not due solely to life-cycle events. The tolerance of (l. pailistris to 3 environnlental paranleters was tested in the laboratory. These included low-salinity, heat/desiccation and freezing-stress experinlents. Comparisons of the results with observed fluctuations of these variables at the study sites demonstrated that only freezing stress would probabl)' cause significant mortality in intertidal populations of this species. Insufficient tolerance to this stress was postulated as the reason for the observed distributional shift to subtidal areas during the winter. Amphipod distribution within the intertidal zone at other times of the year was highly correlated with Spartina density. Substratum preference experiments indicated that this was due to a strong behavioral preference by this species for Spar/ina culms. Intraspecific and interspecific competition for food were tested by an analysis of 3 reproductive indices: the estinlated birth rate as calculated by the egg-ratio method, the average brood size and the average brood size/ovigerous 9. The former 2 indices declined during the early summer as a result of natural adult fenlale mortality and a decreased proportion of ovigerous to nonovigerous . The average brood size/ovigerous 9 did not decline significantly throughout the reproductive period indicating that food limitation did not induce the observed sunlnler decline in amphipod abundance. Intraspecific competition for space was tested in the laboratory by crowding and competitivedisplacement experiments. The results indicated that competition for space was not directly responsible for the summer decrease in amphipod density but did influence anlphipod distribution when Spartina culms were a limited resource. In this situation, G. palus/ris was capable of intraspecific displacement and evidence is presented which indicates that adults are able to displace juveniles from the preferred substratum. Interspecific competition was not examined experimentally because most of the associated fauna inhabiting the marsh beds were infaunal species. Laboratory predation experiments showed that 3 species, Fundulus heteroclitus, Rhithropanopells harrisii and Pale0l110netes pugio could potentially regulate G. pailistris densities. Furthermore, F. heteroclitus predation decreased significantly with increased Spartina density and was strongly size selective for large amphipods. Rhithropanopells harrisii and P, pugio predation was not significantly affected by Spartina density when amphipod abundance was low and predation by both species was significant with high anlphipod and culm density, similar to that observed at the Rhode River sites. Although adult R. harrisii showed some indication of size selective predation, neither species was strongly size selective for large G. pailistris when Spartina culnls were dense. Field caging experinlents conlbined with estimates of predator density indicated that Fundulus predation was the primary source of nl0rtality in the Patuxent River population during the summer. The results of caging experinlents in the Rhode River were inconclusive but large increases in predator density correlated with major declines in amphipod abundance. Therefore, predation is postulated as the major source of mortality during the summer and fall at the Rhode River sites. Thus, both environmental and biological factors regulate these amphipod populations with the former important during the winter only and the latter important at other times of the year.