158 PAUL WEISS 



Technically it might be easier to start with post-embryonic stages, par- 

 ticularly regeneration. For instance, the fact that transplants of mature 

 cartilage may induce new cartilage formation lends itself to a much more 

 detailed analysis than has as yet been undertaken. Nageotte {2'j^ re- 

 ported that the implantation of alcohol-fixed cartilage into the rabbit's 

 ear stimulates the formation of new cartilage in the adjacent connective 

 tissue. In preliminary experiments (50) in which I transplanted pieces 

 of salamander skeleton which had been devitalized by quick-freezing 

 and drying, I found that cells of the larval host built a new shell of 

 cartilage onto the dead model (Plate II) . Since new cartilage has formed 

 only in direct continuity with the graft, we must conclude that the influ- 

 ence in question is transmitted only by contact and not by diffusion. It 

 seems that objects of this kind would be very favorable for a detailed 

 physical and chemical analysis of communicable differentiation. 



In terms of a molecular population concept, this type of inductive 

 effect could be described as follows. The exposed surface of the cartilage 

 contains certain cartilage-specific key compounds. Any potentially chon- 

 drogenic cell would contain similar key species in its interior. On making 

 contact with cartilage, these key species would become concentrated and 

 oriented along the surface by virtue of their affinity to the cartilage. 

 After saturating the surface, they would then direct the subsequent 

 chemical transformations in the cell interior into a cartilage-specific 

 course. Thereupon the cell can, in its turn, act on the next one, and so 

 forth. In the case of cartilage, the selective effect would be exerted by 

 the secreted ground substance rather than the cell body itself, but in other 

 instances it would pass directly from cell to cell. This concept has been 

 developed in greater detail in an earlier publication {$1). Its main thesis 

 is that adjacent cells influence each other's development by means of 

 selective attraction among specifically configurated molecules, key-lock 

 fashion. An alternative assumption would be an actual transfer of 

 specific master units from cell to cell, in the manner of a spreading 

 infection. Present evidence is insufficient for a final decision. 



The second category to be illustrated here is "heteroinduction," in 

 which one tissue induces another tissue, with which it has secondarily 

 come in contact, to differentiate in a direction different from that of 

 the inducing part. Inductions of neural plate by chorda-mesoderm and 

 of lens by retina are familiar examples. According to our hypothesis the 

 mechanism might be as follows. The cells of the inductive layer have a 

 certain specific surface organization, that is, they possess already a 

 surface layer of specific key molecules in relatively stable configuration 



