Arginine vasopressin messenger RNA is axonally transported in the rat hypothalamo-neurohypophysial system [for review see Mohr et al. (1993) In Vasopressin (eds Gross P., Richter D. and Robertson C. L.), pp. 119-129, John Libbey Eurotext]. Upon chronic dehydration (2% saline-feeding for seven days), vasopressin messenger RNA within this axonal compartment is dramatically increased and appears aggregated in a selected subset of axonal swellings confined to the median eminence and posterior pituitary. In this study, we analysed the axonal distribution of the vasopressin messenger RNA within the hypothalamo-neurohypophysial tracts of control and saline-fed animals, and compared this distribution to that of the vasopressin peptide. Our data further support a selective aggregation of the vasopressin messenger RNA in a subset of distal axonal swellings and/or terminals of the median eminence and posterior pituitary. The selective aggregation is observed not only in saline-fed animals, but also in control animals. Although the osmotic stimulus dramatically enhances the axonal transport of vasopressin messenger RNA, the consequent general distribution pattern of the messenger RNA in the hypothalamo-neurohypophysial system is not changed. However, the physiological perturbation does increase the number of vasopressin messenger RNA-containing swellings within the median eminence and the posterior pituitary. In both saline-fed and control animals, the level of messenger RNA label within individual swellings appeared roughly similar to that found in the perikaryal cytoplasm of extra-hypothalamic vasopressinergic neurons. A detailed comparison of the axonal compartmentalization of vasopressin messenger RNA and vasopressin peptide demonstrates that the axonal distribution of vasopressin messenger RNA does not precisely overlap that of vasopressin peptide along the hypothalamo-neurohypophysial tract. In seven-day saline-fed animals, the majority of the messenger RNA-containing swellings of the median eminence also contain detectable vasopressin peptide; however in the same animals, nearly all the messenger RNA-containing swellings of the posterior pituitary appear devoid of vasopressin peptide. Therefore, our work strongly suggests that at least in the posterior pituitary, the vasopressin messenger RNA might be selectively targeted and aggregated in a selected subset of axonal swellings containing little if any vasopressin, and hence very few neurosecretory granules. Given this evidence that vasopressin messenger RNA and neuropeptide are differentially compartmentalized in axons of magnocellular neurons, we propose that vasopressin messenger RNA and peptide probably rely on different intracellular transport systems with respect to packaging, transport and/or aggregation within these selected axonal locations.