The influence of osmotic conditions on the gradient of capillary permeability disclosed by the distribution into mammalian muscle of vital dyes has been tested experimentally. The percentage of circulating blood proteins was increased in rabbits by the injection of compatible plasma, or of compatible serum concentrate obtained by means of a new method of ultrafiltration which has proved both rapid and effective. It was found that when this had been done and the circulatory conditions had stabilized themselves, the gradient of capillary permeability still existed, though its effective extent was less than under normal circumstances. When the percentage of circulating blood proteins was reduced, on the other hand, by repeated bleedings with return of the cells, either as such or suspended in the protein-free fluid obtained by dialysis of serum, the extent of the gradient of capillary permeability was broadened and dye passed out into the tissue more readily than usual from the capillary as a whole. In contrast to these findings injection of a very hypertonic dextrose solution during the period when dye was escaping had no perceptible effect on the gradient of capillary permeability.
The observed phenomena cannot be explained by a flow of dye-stained fluid into or out of the blood vessels. The gradient of capillary permeability exists independently of osmotic conditions, though its extent can be markedly influenced by altering the amount of circulating blood proteins. A considerable proportion of the dye used to study the gradient is adsorbed upon these proteins, as subsidiary experiments have shown. This happening provides a sufficient cause for the differences observed in the extent of the gradient when the percentage of proteins is increased or diminished.
The evidence like that of previous papers indicates that the cause for the gradient is to be found in a structural differentiation along the capillary, such that the barrier offered by its wall progressively diminishes on the way to the venule. Most current estimates of the effective osmotic pressure of the blood proteins fail to take into account the existence of local differences in permeability along the capillary.