Under slight lymphatic stasis (tilting the body 15 degrees) we measured the arrival of locally injected 131I-albumin in the plasma pool. From 30 min to 90 min after the injection the return rate was zero, i.e. local back transport in the two tissues studied, muscle and subcutaneous fat, is very small. Compared with a suggested steady state total 131I-albumin clearance of 1.7%/hour in the horizontal body position, we conclude that a maximum of 1% of the interstitial albumin can have a local transendothelial escape, i.e. can be handled by passive forces, such as diffusion and pinocytosis. Since passive flux is proportional to the concentration, and since the interstitial albumin concentration is about half the plasma concentration, then also diffusion and/or pinocytosis of albumin from the plasma is negligible in the resting normal human. We suggest that filtration through large leaks is the main mechanism for transendothelial protein transport.