SEASONAL-VARIATION IN THE FLUX OF MICROPLANKTON AND RADIOLARIAN ASSEMBLAGE COMPOSITIONS IN THE NORTHEASTERN TROPICAL ATLANTIC AT 2,195 M

Fluxes of silicoflagellates, the dinoflagellate Actiniscus sp., polycy stine and phaeodarian radiolarians, tintinnids, ciliate(?) cysts, and pelagic molluscs were estimated for 13; sediment trap samples from the northeastern tropical Atlantic (20 degrees 5 5.3'N, 19 degrees 44.5'W ) at 2,195 m between 22 March 1988 and 8 March 1989 (site CB1). Each s ample integrated the flux over 27 d, and polycystines were identified to species in all samples. Polycystines had the highest fluxes. For ph ytoplankters, our estimates are lower than most reported data, and for polycystines and tintinnids the values are among the highest ever rec orded. Temporal variations in the fluxes of the heterotrophic organism s counted generally were in good agreement with total mass flux, sugge sting fairly tight couplings with primary production at the surface. F luxes of tintinnids were more variable through time and better associa ted with variations in total mass flux than those of the slower reprod ucing radiolarians. We identified 145 polycystine taxa. Species compos itions changed little throughout the year and did not vary with change s in total mass flux. Comparison of our data with a similar survey of sediment trap samples retrieved between 1 March 1989 and 16 March 1990 from 853 m at the nearby GBN3 site showed significant differences in the fluxes of the groups and in the percentages of many polycystine sp ecies. All groups (except silicoflagellates) had higher output rates a t CB1, and proportions of several polycystines associated with colder or more productive environments also were higher at CB1. Conversely, G BN3 yielded higher proportions of various radiolarians characteristic of warmer, more oligotrophic waters. Because temperatures below simila r to 70 m are higher at CB1 than at GBN3, different productivity level s, rather than different surface temperatures, may be important in str ucturing the specific differences recorded.