We have examined the diameter response of rat femoral artery segments in the presence and absence of endothelium to changes in flow rate. The segments were isolated, mounted on microcannulae, maintained at 37 degrees C, and perfused at 90 mmHg with Tyrode's solution. The external arterial diameter was measured using video-microscopy. The mean control diameter was 741+/-22 microm (mean+/-SEM,n=7). The arteries were preconstricted to 75+/-1% of the control diameter with a superfusion of 1 microM norepinephrine (NE). Endothelial function was verified by perfusion of 1 micro;M acetylcholine (ACh). Two different flow protocols were employed: step changes in flow (n=7) and low-frequency sinusoidal flow changes (0.01Hz<f<0.05Hz,n=3). The endothelium was then removed by perfusion of the artery with 0.3% CHAPS (n=7) and the flow protocol repeated. All arteries contracted with increasing flow. For rapid step changes in flow, contractions and relaxations were modelled with an exponential fit where characteristic time constants were 34+/-2 and 40+/-3s, respectively (functional endothelium) and 35+/-2 and 37+/-2s, respectively (non-functional endothelium). The time constants in the presence and absence of endothelium were not significantly different (paired t-test, p>0.05). Sinusoidal flow oscillations resulted in sinusoidal diameter oscillations, whose amplitude and phase lag were inversely proportional to the frequency of the flow oscillations. A first-order low-pass filter, with a time constant of 28+/-3 and 30+/-5s for arteries with and without functional endothelium, respectively, was used to describe the relation between oscillatory flow and diameter. The response of the rat femoral arteries to changes in flow was not found to be different whether the endothelium was intact or removed.