The development of eosinophilia has an interesting biological specificity, in that the numbers of eosinophils can increase from a normal paucity to very high levels, without a concomitant increase in other leukocytes. This suggests a unique controlling mechanism. Although IL-3, GM-CSF, and IL-5 all induce eosinophil differentiation, IL-5 is the only identified cytokine specific for the eosinophil lineage. This together with the fact that IL-5 can be detected in the serum of mice with eosinophilia suggests that it may be a critical controlling factor. The mechanism of this control is not yet clear. It appears that a high proportion of T-cell clones produce IL-5, even when isolated using antigens that do not induce eosinophilia. If IL-5 is produced in response to all antigens, what is the basis for the biological specificity of eosinophilia? It is possible that the production of lymphokines by isolated T-cell clones does not reflect the true biological situation, because there may be a selection process in the isolation of T cells that is not yet understood. In addition, IL-5 appears not to stimulate the production of precursors committed to the eosinophil lineage, at least in vitro. Thus, it would seem likely that other controlling mechanisms exist to generate the precursors. Attempts to study this stage of eosinophil production are difficult to interpret. The experiments are technically difficult, and conflicting results have been reported. Interleukin-3 and GM-CSF appear to have this activity in vitro, but it is not clear how these factors could be active in the production of eosinophil precursors without also inducing increases in other leukocyte types, particularly neutrophils. Another intriguing aspect of the biology of IL-5 is the activity on other cell types, particularly B cells in the mouse. This activity as a murine BCGF precedes the identification of IL-5 as an eosinophil differentiation factor, and has been studied in considerable detail for many years. However, if IL-5 is important in the production of antibody, why are eosinophils not seen in all immune responses? An important insight into this problem has come from the observation that human IL-5 has no analogous activity on human B cells. If IL-5 is not active on human B cells, it raises questions about the biological role of IL-5 as a BCGF in the mouse.(ABSTRACT TRUNCATED AT 400 WORDS)