The availability of suitably characterized dispersed islet cell preparations may assist in studies of islet function. Since freshly dispersed adult rat islet cells failed to respond appropriately to secretagogues (no alteration in insulin, glucagon, or somatostatin release after glucose change; modest response to IBMX), these cells were established in primary monolayer culture. We then tested the hypothesis that islet function is at least partially determined by islet structure. B cells which had attached to Petri dishes during a culture period of four days were well preserved at the ultrastructural level, with mitochondria clustered at the cell face attached to the Petri dish and secretory granules concentrated towards the portion of the cell facing the medium. Since it was not possible to estimate cellular hormone content or hormone release as a function of the number of specific types of cells, fractional rates of release and hormone content ratios were compared with those for intact islets maintained in culture in parallel. Whereas the ratio of somatostatin:insulin content was similar for islets and cells (approximately 0.7:100), the dispersed cell population appeared depleted in glucagon (glucagon:insulin ratios being 17:100 for islets and 4:100 for cells) reflecting either degranulation or relative loss of A cells. In contrast to the lack of responsiveness seen with freshly dispersed islet cells, the cultured cells released insulin in response to glucose and glucose plus IBMX in a fashion comparable to that seen with cultured islets. Proinsulin biosynthesis (incorporation of [3H] leucine) was higher in cultured cells than islets. Somatostatin release was lower from dispersed cells than from islets while the opposite was true for glucagon.