154 PAUL WEISS 



take place all at once throughout the cell. We must assume that it starts 

 from a localized area (in our concept, the surface) and spreads at a 

 measurable rate. Therefore if the cell were returned to its original 

 condition A during the early phases of this process, it would still con- 

 tain sufficient amounts of all the species it had before to resume its 

 original state. To the observer this would appear as reversible modula- 

 tion. But continued residence under condition B would permit con- 

 tinued progress of the changes indicated, until eventually the cell will 

 have lost a major segment of its earlier constituents. It would thus have 

 undergone an irreversible change of its constitution, i.e. true differentia- 

 tion. The lability of differentiation in its incipient stages is a matter of 

 record. 



If a cell that has assumed character B then moves on to a site C (or if 

 its neighbor cells or the intercellular milieu changes to a condition C) , 

 its surface population will again become reconditioned and then initiate 

 a new chain of processes, marking a further step of differentiation, the 

 earliest phase of which would again be labile. A cell transferred from 

 A to C directly may or may not respond in the same fashion as when 

 coming by way of B, depending on how decisively its molecular popula- 

 tion has been altered during the intermediary stay at B. 



We are portraying "differentiation" as a chain of events consisting 

 of alternate physical regrouping and chemical alteration of the molecu- 

 lar populations, the latter phenomenon involving the emergence of novel 

 species of compounds. It can be seen that according to this concept the 

 potentialities for differentiation are strictly limited by the initial chemi- 

 cal endowment of the cell and that they become further restricted as the 

 cell passes through the various phases A, B, C. What we conventionally 

 call "loss of potency" is therefore merely the counterpart of the positive 

 increase in definite chemical specialization incurred by the "differentiat- 

 ing" segments of the molecular population. We need assume no separate 

 inhibitory agents. 



This concept contains also, in principle, the answer to the problem of 

 divergent cytodifferentiation, that is, of how two initially identical cells 

 may become the source of two qualitatively different strains. Let us 

 assume that of two cells with identical molecular populations, including 

 key species a, ^, y, S, e, etc., one is exposed to an environment E, and the 

 other to a different environment F. Given some agitation and sufficient 

 mobility of the cell contents, we may further assume that in each cell a 

 selected fraction of key species will be concentrated at the surface, 

 namely, those species which best conform to the adjacent medium; let 



