146 HEREDITY [AP. 



Boveri [3] to believe that the subsequent develop- 

 ment of the cells of the embryo depends on the 

 distribution of the chromosomes in the abnormal 

 divisions consequent on double fertilisation. And 

 Herbst [16] has obtained sea-urchin larvae made by 

 crossing distinct species, which on one side of the 

 body resemble one parent, and on the other side 

 the other parent. He shows that these differences 

 are connected with differences in the size of the 

 nuclei of the two sides, and that probably the part 

 with maternal characters contains only maternal 

 nuclear substance, while the part showing the 

 paternal character has nuclei derived from both 

 parents. 



But probably the best evidence for regarding the 

 chromosomes as bearing the essential determinants 

 for hereditary characters is provided by the behaviour 

 of the chromosomes themselves in the maturation 

 divisions of the germ-cells. It has been pointed out 

 that at these nuclear divisions the chromosome 

 number is halved, and restored to the full number 

 again in the next generation by the union of two 

 germ-cells each bearing the half-number. Now it 

 has been found in certain cases that the chromosomes 

 are not all alike, but differ among themselves in size 

 and shape, and when this is so it can be seen that 

 the nucleus just before maturation contains two of 

 each kind. If the different kinds of chromosomes 



