An assumption inherent in the use of microsphere methodology for measuring regional blood flow is that microspheres are removed from the circulation by impacting in arterial vessels of approximately their own diameter. We investigated the in vivo relationship between the diameter of varying size microspheres and the calibre of trapping microvessels within the hamster cheek pouch. Intracardiac injection of 8-10, 15 +/- 5, 25 +/- 5 micron carbonized microspheres with diversion of the cardiac output into the bracheocephalic trunk provided direct, in vivo visualization of the impaction sites of these various size microspheres within the cheek pouch microvasculature. Fifteen micron microspheres usually impacted at the orifice or neck of small arterioles and protruded into the lumen of the parent vessel. Eight to ten micron microspheres lodged in vessels with a mean diameter of 11.5 +/- SD 3.4 micron indicating that they usually impacted primarily in terminal arterioles.