1. An acute in vivo hindlimb skeletal muscle preparation was developed in anaesthetized sheep in order to facilitate studies of the effects of altered blood flow states on drug kinetics in skeletal muscle. 2. A continuous index of blood flow was recorded via ultrasonic Doppler probes on the femoral artery and vein. Skeletal muscle effluent blood was sampled via a catheter in the femoral vein proximal to the probe. Low- and high-blood flow states were achieved by direct femoral artery infusion of adrenaline (0.002-0.006 mg/min) or magnesium (0.4-1 mmol/min), which produced mean (+/-SD) stable flow states of 25+/-12 and 185+/-56% (both n = 5) of baseline, respectively. The correlation coefficients between arterial and venous Doppler frequency shifts in five sheep during and after adrenaline infusion were 0.96 (indicating these vessels probably supplied and drained common tissue). 3. The venous Doppler frequency shifts were calibrated against timed collections of the femoral vein outflow to provide estimates of the low and normal blood flow states (mean flows of 2.8+/-1.7 and 9.3+/-5.7mL/100 g per min; both n = 3) and against an indirect Kety-Schmidt method during low- and high-blood flow states (mean flows of 2.2+/-1.3 and 18.2+/-7.0 mL/100 g per min; both n = 5). There was a generally good agreement between the two methods. 4. The tissue was neither hypoxic nor acidotic in the low- or high-flow states and altering the flow produced no change in muscle oxygen consumption, suggesting the flow changes were largely due to changes in resistance vessel tone. 5. Postmortem femoral artery dye injection and dissection of stained tissues showed the artery supplied 657+/-96 g (n = 5) of skeletal muscle and 42+/-20 g (n = 5) of other tissues. 6. It is concluded that the method is suitable for sampling predominantly muscle effluent blood at low- or high-blood flow states according to experimental requirements.