WILLIAM ROUX'S THEORY 169 



then in responding to that stimulation which becomes 

 necessary to the cell's life. The activity of a cell, of 

 a tissue or of an organ determines, therefore, the shape 

 and degree of development of that cell, tissue or 

 organ. This is universally admitted as far as the 

 anatomical appearance of organs is concerned. It is 

 true also, however, of their histological structure. 

 Roux gave as evidence the structure of the spongy 

 tissue of the long bones. It was noticed long ago 

 that the lamella? of this substance were so disposed as 

 actually best to withstand the stresses brought to bear 

 on the bones. Natural selection could not have 

 brought about this disposition which is undoubtedly 

 useful. If, at the very beginning, when the bone 

 lamellae were orientated in every possible direction, 

 some of them had assumed a useful orientation, this 

 slight variation would not have been sufficient to influ- 

 ence natural selection. If we suppose, on the other 

 hand, that the variation affected at once a large num- 

 ber of lamella? (and this could no longer be described 

 as the usual fluctuating variation), this large number 

 of lamella must have constituted the absolute major- 

 ity in order to insure usefulness. In this case we can- 

 not see why the transformation did not cease then 

 and there instead of continuing until all the lamellae 

 had been orientated similarly, which was superfluous. 

 Natural selection could not favor the development of 

 a useful structure beyond what is necessary; only 



