JEFFREY: EVOLUTION BY HYBRIDIZATION 301 



ously be most easily estimated and varies in proportion from a small 

 percent to complete sterility. The sterility resulting from hybrid- 

 ization should not be confused, as sometimes happens, with sterility 

 arising from purely physiological causes. For example the common 

 horse radish, Liliiim bulbiferum and L. candidum, under ordinary 

 conditions do not set seed, by reason of the fact that the assimilates 

 are too strongly determined to the subterranean parts to permit of 

 the necessary materials being set free for the formation of seeds. It 

 has been found however that the girdling the top of the subterranean 

 stem in the horse radish or cutting off the flowering axis in the case of 

 the lilies, brings about the formation of normal seeds. Similarly very 

 marked climatic change or subjection to starvation or other extremely 

 unfavorable physiological conditions results in the degeneracy of 

 reproductive as well as other parts. Conditions like these are, how- 

 ever, very easily distinguished from the sterility normally resulting 

 from hybridization. 



Sterility in hybrids is of particular interest from the genetical 

 standpoint because it more or less completely upsets the expectancy of 

 Mendelian ratios in cultures of the offspring of species hybrids. This 

 is doubtless one of the causes why the Mendelians have in general 

 manifested so little interest in the genetical study of hybrids between 

 natural species. Obviously however if the crossing of species in nature 

 is a common cause of the origin of new species this part of the evolu- 

 tionary field cannot be safely neglected if we are to reach any broad 

 and permanently valid conclusions as the modus operandi of the origin 

 of species. 



Another feature in the organization of hybrids is the frequent in- 

 crease in the typical generic chromosome number as a consequence of 

 crossing. For example we find in the much crossed oriental species of 

 Chrysanthemum a number of chromosomes in the gametic nuclear 

 divisions varying from 9 (the normal) to 18, 27, 36, and even 45, in 

 other words, two, three, four and five times the normal gametic 

 number. Similarly in another compositaceous genus Dahlia we find 

 in the species D. coronata sixteen chromosomata in gametophytic 

 divisions while in the hybrids between D. variabilis and D. coccinea 

 thirty-two chromosomes have been enumerated. One further example 

 will point the situation. In the monotypic Liriodendron and in certain 

 species of Magnolia, nineteen gametophytic chromosomes have been 

 counted, while in M. soulangeana (suspected of hybrid origin) as well 

 as M. Yulan and M. grandiflora twice that number or more of chromo- 

 somes have been observed. If we contrast the situation in these 

 examples with that presented by the genera Pinus and Lilium, which 

 are not at all prone to hybridism, we note a curious contrast. In the 



