191 2] On the Distribution of Potassium in Renal Cells 395 



distributed by the knife in cutting, for, if so, the heavy precipitates would 

 appear in more or less parallel lines at right angles to the direction of 

 cutting and in much coarser aggregations, as was the case in some 

 preparations in which diffusion had occurred before precipitation. Nor 

 could they ha\ e been due to a penetration of the reagent into the lumina 

 and intertubular spaces fi'-st, then into the cells. The character of the 

 section excludes this possibility. The reagent comes in direct contact 

 with a cross section of a tubule from lOft to 20/i thick. It will penetrate 

 directly at all points and not first to the interspaces, then to the base- 

 ment membrane and so diffuse throughout a tubule of 18^ radius. 



It might be urged that the reagent, penetrating the tissue at some 

 points quicker than at others, precipitates the potassium with which it 

 first comes in contact at these spots. Then the rest of the potassium in 

 the cytoplasm, from its great diffusibility, would immediately gather and 

 precipitate at the centres thus formed. This might be a real danger 

 in organs limited by a cuticle or membrane, but in these prepara- 

 tions it could not obtain. The conditions are all against such a result, 

 for a frozen cross-section dropped into the reagent while still frozen 

 thaws equally and uniformly throughout and permits of the penetra- 

 tion of the reagent at all points of the surface at the same time. If this 

 tendency toward diffusion to the point of initial precipitation operated, 

 one would expect to find a heavy surface precipitate over the membranes, 

 for these, being firmer, stand out as it were and first meet the reagent. 

 Here would be the first precipitation, and to this would diffuse, not only 

 the potassium in the cytoplasm, but that from the deeper portions of 

 the membranes themselves. This would be apparent immediately on 

 adjusting the microscope so as to focus the deeper portions of the section 

 with little or no precipitate in contrast with the heavy peripheral and 

 central surface precipitates thus supposed to occur. If only the surface 

 layers were affected by this first diffusion, then, as the section thaws, 

 the reagent would penetrate the cytoplasm more quickly than the 

 membrane, and a heavy deposit in the deeper portion of the cytoplasm 

 would result. 



But none of tnese conditions obtain, actually the opposite being 

 the case. Where diffusion has taken place on the surface before contact 

 with the reagent, it is found that the tendency is for a uniform distribu- 

 tion over the entire surface of the section. The same applies to the 

 deeper portions as well, the tendency being to a uniform distribution 

 throughout. In Fig. 21a is shown a tubule in which a large measure of 

 surface diffusion has taken place, this being a drawing of the surface 

 view only, while Fig. 216 is of the same tubule observed at a deeper level 

 of the section. What, however, may be considered the strongest 



