The blood-nerve movement of a small molecular weight non-electrolyte was studied in control and galactose-fed rats by measuring the permeability-surface area (PSA) product of the blood-nerve interface to [14C]mannitol in sciatic nerve using an in vivo injection method. PSA products were measured after 9 to 11 months of feeding control rats a diet containing 0% galactose and galactose-intoxicated rats a diet containing 40% galactose. Nerves of the galactose-fed group were hydrated as reflected by a significant increase in nerve water content and wet weight to dry weight ratio (both P < 0.05). Compared to controls, PSA products were increased by 51% (P < 0.01) in galactose-fed animals when referenced to nerve dry weight (13.59 +/- 2.90 x 10(-5) ml/s/g dry wt. versus 8.99 +/- 1.59 x 10(-5) ml/s/g dry wt.; mean +/- S.D.; galactose vs. control, respectively) or by 30% (P < 0.001) when referenced to nerve length (2.43 +/- 0.43 x 10(-5) ml/s/mm vs. 1.87 +/- 0.48 x 10(-5) ml/s/mm) but not when referenced to nerve wet weight. It is suggested that in galactose intoxication, where endoneurial volume changes reflect increases in nerve water content, PSA products are best normalized to dry weight or length, which are not affected by volume changes. Normalized to dry weight, the blood-nerve barrier surface area (i.e. vessels and perineurium) was determined by morphometric methods to be increased by 34% in the galactose-intoxicated group.(ABSTRACT TRUNCATED AT 250 WORDS)