Thirty years ago physiologists' ideas about blood circulation in the kidney were based on what was known of its structure. The branching arterial system supplied the glomerular capillaries, which recombined to form efferent arterioles which fed a second capillary system around the tubules. Accordingly, the glomerular capillary pressure would be higher and the peritubular capillary pressure lower than pressures in the capillaries elsewhere in the body, with consequent pressing out of ultrafiltrate in the glomeruli and the sucking back from the tubules of most of it by colloid osmotic action of plasma proteins. Then methods of measuring blood flow were developed, including direct measurements in isolated perfused kidneys, flow-meters in anesthetized animals and later the indirect clearance methods which can be used in unanesthetized animals and man. Experiments spoiled the simple and satisfying theory. The blood flow did not increase proportionately or more than proportionately with pressure as it does