64 APPLICATION OF PRINCIPLES OF HEREDITY TO BREEDING. 



which the Primus berry is descended was unhke any of the others. 

 But this individual has reproduced true to seed from the beginning. 



Kosenberg has shown that in a similar case in the genus Hieracium 

 the first-generation hybrids, which are not alike, have different 

 numbers of chromosomes. Presumably, in such cases, the chromo- 

 somes of the two species crossed differ so much in habit that they 

 can not function together properly, and some of them are lost in the 

 early cell divisions of the young hybrid. In some individuals, one 

 chromosome is lost, in others other chromosomes are lost; so that, 

 if the supposition here made is true, the young first-generation 

 hybrids do not have the same complement of hereditary characters, 

 which would account for their not being alike. 



The fact that these hybrids do not split up in Mendelian fashion 

 in the second and later generations suggests that the corresponding 

 chromosomes in the two sets of chromosomes brought together in 

 these wide crosses are so unlike that they are not drawn together 

 to form bivalents in the reduction division. That is, the reduction 

 division does not occur in the mother cells which produce ovules 

 and pollen. If this should prove to be true, then either of two alter- 

 native courses of events would give hybrids which reproduce true 

 to seed without Mendelian splitting. 



(1) Seed may be produced parthenogenetically, without the in- 

 tervention of a reduction division and subsequent fertilization, and 

 this asexual production of seed may continue from generation to 

 generation. 



(2) Without a reduction division in the first-generation hybrid, 

 a pollen nucleus may unite with an ovule nucleus, thus giving a 

 nucleus having two sets of chromosomes like the one possessed by 

 the first-generation hybrid. The set of chromosomes in the first- 

 generation hybrid is presumably composed of chromosomes part of 

 which are from one species and part from the other. The double 

 nucleus thus formed will have two sets of chromosomes, every one 

 in one set having an exact duplicate in the other. After this, that 

 is, in the second and later generations, reproduction would take place 

 in the usual manner, without any Mendelian splitting, for the two 

 halves of each bivalent formed in the reduction division would be 

 exactly alike. 



It would probably be possible, by cytological study of these con- 

 stant hybrids, to determine whether the absence of Mendelian split- 

 ting is due to either of the causes above suggested. 



MUTATIONS. 



The term ''mutation" has been used and is now used in so many 

 senses that a great deal of confusion has arisen in consequence. 

 Prof. Hugo De Vries, of Holland, after investigating some hundreds 



165 



