DIFFERENTIAL GROWTH I7I 



liferation^" could be explained by the penetration of the central fluid to 

 the deeper layers as a result of the break in the internal limiting mem- 

 brane. This is purely conjectural, but capable of an experimental test. 



Next, what determines the further fate of the cells proliferated in 

 the germinal layer? Of the two daughter cells of a germinal cell, either 

 (a) both stay in the germinal layer, or (b) both move away, or (c) one 

 stays and the other moves off. The same alternatives recur at subsequent 

 divisions. Continuation of procedure (a) would lead to exponential 

 increase of the proliferative source with no cells left for terminal differ- 

 entiation, while procedure {b) would rapidly deplete the proliferative 

 source. The actual growth of the eye at this phase is therefore deter- 

 mined by the relative incidence of events (a), (b), and (c). But again, 

 factual information on this point is practically wholly lacking. In anuran 

 amphibians the number of dividing cells in the retina has been reported 

 to remain remarkably constant (i6), which would mean that all divi- 

 sions follow scheme (c). But this is definitely not true of the chick, for 

 instance, in which the number of proliferating cells increases with the 

 expansion of the retina. ^^ Here the rate of growth is therefore largely 

 dependent on the statistical apportionment of the newly formed cells 

 according to ((7), (b), or (c). This in turn seems to resolve itself into 

 a matter of the relative orientation assumed by the daughter cells,' either 

 during mitosis or immediately after, tangential orientation leading to 

 (a), radial orientation to (b) or (c). A plausible hypothesis is the fol- 

 lowing. 



It has been demonstrated that dividing cells separate in the direction 

 of maximum cell elongation, which corresponds to the direction of 

 maximum tension. Hertwig's rules of cleavage are an illustration of 

 this principle. Assuming that retinal cells follow the same principle, we 

 may conclude that distribution of type (a) will result whenever the lo- 

 cally prevailing tension is tangential, while types (b) and (c) will re- 

 sult from radial tension or lateral pressure. According to this concept 

 the proliferating layer can expand only as long as the retina is under 

 tangential tension. The source of such force is found in the distension 

 of the eyeball by the growing lens and the vitreous body. Retina, lens, 

 and vitreous body would thereby form a self-regulating system operat- 

 ing as follows. 



Accumulation of vitreous fluid would stretch the retina, producing a 



1° See for instance the compensatory proliferation of cells throughout the remaining 

 half of an embryonic hindbrain after removal of one half (12). 



11 Recently confirmed by unpublished investigations in this laboratory by Mrs. 

 Eleanor Gould. 



