Microenvironmental Influences in Cytodifferentiation 225 



rudiments of secretory tubules appear (Fig. 7). These facts, together 

 with others previously gathered on kidney development in the intact 

 organism, demonstrate that epithelial tubule formation by nephro- 

 genic mesenchyme is dependent upon an outside contribution. Nor- 

 mally this is assumed to be provided by the ureteric bud; experi- 

 mentally it can also be supplied by dorsal spinal cord. 



For reasons of convenience, it is the combination of dorsal spinal 

 cord and metanephrogenic mesenchyme which has received most at- 

 tention. When a fragment of cord is included among fragments of 

 metanephrogenic mesenchyme, secretory tubule rudiments appear in 

 the mesenchyme after 30 to 40 hours of incubation. It is important 

 to note that these rudiments do not appear generally in the mesen- 

 chyme but only in immediate association with the surface of the spinal 

 cord. I say "immediate association" now; when the observation was 

 first made the relationship suggested "■contact.''' I note this to empha- 

 size that the system under consideration shows the usual character- 

 istic of intimacy in embryonic induction. 



Having such an inductive system under reasonable control in vitro, 

 the next question is the one of separability of components. It turned 

 out that if dorsal spinal cord and metanephrogenic mesenchyme were 

 cultured "back to back" on a highly porous membrane filter ( milli- 

 pore), secretory tubule rudiments appeared in the metanephrogenic 

 mesenchyme immediately opposite the spinal cord (Fig. 8). I say 

 "immediately opposite" to emphasize again that the tubules do not 

 appear just anywhere in the mesenchyme; they form as close to the 

 cord as they can under the circumstances, separated only by the thick- 

 ness of the membrane filter. The filter used in the early experiments 

 was approximately 20 microns thick, with average pore diameter in 

 the vicinity of Yo micron. By examining sections with the optical 

 microscope, particularly under phase conditions, it was possible to 

 discern materials penetrating into the filter. Electron microscopy 

 showed that some of this material, at least, was cytoplasmic — so one 

 could not say with certainty with this filter that a 20-micron separa- 

 tion between cell surfaces was achieved. Subsequent experiments, 

 however, with filters of lesser porosity (in the vicinity of 0.1 micron), 

 showed that cytoplasm-free separation of at least 20 microns can be 

 achieved without blocking the induction, though the intensity is 

 reduced. 



This last point, the reduction of intensity at 20 microns by mem- 

 branes of porosity in the vicinity of 0.1 micron, raises a red flag to 

 easy assumptions that separability implies free diffusion and indefi- 

 nite mobility on the part of the active materials. The caution is 



