By use of the H2 clearance technique, blood flow was measured in the sciatic nerve of healthy, anaesthetized Sprague-Dawley rats at rest, during inferior vena cava occlusion and during 5-hydroxytryptamine infusion. The purpose was to clarify the mechanisms underlying the biexponential curves which are commonly obtained using this technique. 2. An analysis of the frequency distribution of rate constants of 270 nerve and thirty-three arterial samples indicated that H2 clearance rates cluster below 20 ml min-1 100 g-1 and between 70 and 100 ml min-1 100 g-1. This suggests that at least two compartments are present. 3. The contribution of diffusion was studied by recording H2 clearance immediately following circulatory arrest. Slow clearance rates (median = 2.4 ml min-1 100 g-1) were observed, indicating that diffusion is not likely to contribute significantly to nutritive flow under most situations. 4. The contribution of arteriovenous shunts to H2 clearance was assessed by determining H2 clearance during inferior vena cava occlusion and the infusion of 5-hydroxytryptamine. Both manoeuvres caused abolition of, or a significant reduction in the weight of, the fast component which indicates that this compartment is closely related to arteriovenous shunts in nerve. 5. By use of a multi-compartmental model, it was shown that H2 clearance should follow a multi-exponential course, where the weights of the components reflect the relative volumes of each compartment and the exponents represent the relative flow (i.e. flow per unit volume) in each compartment. 6. By use of other mathematical models, estimates were made for the clearance rates attributable to polarographic oxidation of H2 at the tip of the microelectrode (0.2 ml min-1 100 g-1) and to diffusion to air (2 ml min-1 100 g-1). The latter estimate is very close to the measured value of 2.4 ml min-1 100 g-1. 7. These findings indicate that it is possible to separately assess nutritive and non-nutritive flow by application of biexponential analysis to H2 clearance curves. The data suggest that the fast component of a H2 clearance curve is closely associated with arteriovenous shunts, while the slower component is likely to represent capillary flow. Processes such as diffusion to air or oxidation of H2 by the electrode are very slow and therefore are unlikely to distort the assessment of blood flow by using this technique.
