Because Prostaglandin E(2) (PGE(2)) and dibutiryl cyclic AMP (dbcAMP) modulate the production and effects of haemopoietic cytokines in allergy, we examined their ability to modulate responses of myeloid progenitors to GM-CSF, and of eosinophil precursors to IL-5. 2. The ability of PGE(2), dbcAMP, rolipram, forskolin, dbcGMP and PGD(2), to modulate the responses to GM-CSF and IL-5 in colony formation (progenitor) and eosinophil differentiation (precursor) assays using bone-marrow from nonsensitized or from intranasally-challenged, ovalbumin-sensitized mice of five strains was studied. 3. PGE(2) (10(-7) M) inhibited GM-CSF-stimulated colony formation in bone-marrow from BP-2 mice. This effect was duplicated by dbcAMP (0.3 - 1x10(-6) M), Rolipram (10(-5) M) and forskolin (3x10(-5) M), but not Prostaglandin D(2) (10(-6) M). Inhibition affected similarly all myeloid colony types. Progenitors from sensitized and challenged BP-2 mice were also inhibited by PGE(2) and cyclic AMP. PGE(2) inhibited progenitors from C57BL/10, CBA/J and A/J, but not BALB/c mice. However, BALB/c progenitors were sensitive to dbcAMP and Forskolin (10(-4) M). In contrast, in precursor assays, PGE(2) (10(-7) - 10(-9) M) blocked responses to IL-5 in bone-marrow from BP-2 and BALB/c mice, either naïve or sensitized and challenged, to a similar extent. PGD(2) (10(-6) M) was ineffective, as was PGE(2) (10(-7) M), if added after 48 h of culture. 4. In conclusion, PGE(2) inhibits the responses of bone-marrow myeloid progenitors to GM-CSF and of eosinophil precursors to IL-5, in naïve or ovalbumin sensitized and challenged mice. These effects are duplicated by cyclic AMP-elevating agents. In the BALB/c strain, the resistance of progenitors, but not precursors, to PGE(2) inhibition, indicates these developmental stages are separate targets for PGE(2) modulation.