It is clear from the extant literature that various fish groups face chronic osmoregulatory problems that depend on the surrounding salinity. Their physiologic and hormonal responses are largely those seen in the mammals, but their terrestrial descendants have lost osmoregulatory structures such as gills and rectal glands and depend primarily on renal function. A data base is now emerging that strongly suggests that a putative atriopeptin plays a role in osmoregulation in fishes. This conclusion is supported by the fact that heterologous AP produces relevant physiologic responses (e.g. natriuresis, vasodilation, stimulation, or inhibition of Na+ secretion by intestine, gills, and rectal gland) in both teleosts and elasmobranchs. Moreover, cardiac and brain extracts from fish can produce similar effects in both fishes and mammals, and these tissues from various fish groups contain immunoreactive AP, as does plasma. Both physiologic and immunologic evidence suggests that the ventricle may be a significant source of AP in fishes, contrary to the situation in mammals. Finally, osmotic perturbations result in a change in plasma and tissue APir levels. The finding that plasma APir levels increase in sea water, and that heterologous AP stimulates salt secretion by the teleost gill and shark rectal gland, and inhibits salt uptake by the teleost intestine, suggests that AP may primarily play a role in salt, rather than fluid, secretion in fishes. The fact that in mammals AP inhibits prolactin secretion, but is itself stimulated by cortisol, supports this conclusion, since prolactin is generally considered to be the dominant osmoregulatory hormone in freshwater fishes, and cortisol serves this function in marine fishes. In addition, if AP inhibits brain AVT release in fishes, as it apparently inhibits vasopressin release in mammals, this also would be adaptive in marine fishes since AVT in fishes is diuretic, rather than antidiuretic. Interactions between AP and these hormones (prolactin, cortisol, and ATV) have not been studied in fishes to date, but these theoretical interactions do lend support to the hypothesis that AP may function primarily in salt homeostasis in fishes. At least one potential hormonal interaction counters this argument, however. Atriopeptin is known to inhibit the production and effects of angiotensin II in mammals, and since this hormone is apparently dipsogenic in fishes, it may play a critical role in osmoregulation in sea water. Finally, it is of some historical interest that in Keys' (67) original description of the eel heart-gill perfusion system in 1931 he commented that gill resistance remained constant for hours only if the heart itself was perfused.(ABSTRACT TRUNCATED AT 400 WORDS)