THE CAUSE OF VARIATION 89 



in cells in some degree to supply the answer. The proximate 

 cause of variation lies in the differential division of the un- 

 fertilised germ-cell. Those who have traced the process of 

 fission under the microscope, describe it as an elaborate 

 arrangement for securing an equal division of every part of the 

 mother-cell between the two daughter-cells. Their descriptions 

 naturally relate only to what they see, but it is difficult to doubt 

 that that process is carried out through every minute part of the 

 cell far beyond the range of vision. It is reasonable to assume 

 that the seen process extends to the unseen, and in no other 

 way can we imagine the powers of the mother-cell to be 

 transferred to both the daughters. The immense complexity 

 of the germ-cell of one of the higher animals is, I suppose, 

 generally assumed. At all events it seems to me a necessary 

 assumption. The brain of an ant has been called the most 

 wonderful piece of matter in existence, but it is nothing in 

 comparison with this speck of comparable size, of which a part 

 only, if the whole be fertilised and provided with proper food 

 and environment, develops into the brain of a man. When the 

 unfertilised cell divides into two daughter-cells, it is as if every 

 part of something far more complicated than a watch were 

 divided each into two equal and similarly shaped parts, and 

 the whole were put together again to make two watches of 

 smaller size. There must be many thousands of factors, each 

 with its special function to perform after fertilisation, should 

 this occur, in the growth of the animal or plant, and every one 

 must be equally divided between the two daughter-cells. 

 Being so numerous, these factors must be almost inconceivably 

 small, yet so much smaller are chemical atoms and molecules 

 that immense numbers may enter into the composition of each, 

 whether they are grouped, as is probable, into intermediate 

 units or not. I have said that on a fission each factor must be 

 equally divided, but here is the point, the division of the mole- 

 cules of each factor between the two daughter-cells can never 

 be more than approximately equal, or at least the chances are 

 immensely against such an event, and the division is therefore 

 differential. Let us assume what is extremely improbable, but 

 will equally well serve the purpose of argument, that a certain 

 very small factor of a cell about to divide consists not of millions, 

 but of so few as a hundred molecules of one very complex 

 chemical compound, and of nothing else. On a division the 



