June 14, 1907] 



SCIENCE 



943 



cence, the brown seeds, the impaired fertility 

 of the original hybrid plant, have remained 

 unchanged. 



Such behavior is by no means rare among 

 blend-hybrids in other genera than Viola. 

 Willow hybrids, for example, are said to pro- 

 duce stable offspring always like the parent 

 hybrid. Dr. MacDougal states' that "more 

 than a thousand such fixed hybrids or hybrid 

 species are known." In fact, de Vries is of 

 the opinion that this procedure takes place in 

 all hybrids between pure species — that is, 

 species that differ from each oth&c by no 

 * varietal ' character. In such hybrids the dif- 

 ferences of the parents are so thoroughly 

 blended, that they do not disunite in the germ- 

 cells, as in hybrids governed by the law of 

 Mendel, and therefore the offspring simply 

 repeat the form of the parent hybrid. 



One further observation is of interest — a 

 tendency in certain individual seedlings to re- 

 cover from the marked impairment of fer- 

 tility that characterizes nearly all violet hy- 

 brids. It is well known that partial or com- 

 plete sterility is usually found in a hybrid, 

 when the parent forms differ from each other 

 in several or many characters; but that when 

 the differences are few, especially when only 

 one or two (as often between a species and its 

 variety), there is seldom any loss of fertility. 

 We are further taught by Mendel's law, that 

 when the parents of the first cross differ in 

 more than one character, the majority of the 

 offspring will be hybrid in fewer characters 

 than the parent; in fact, if the offspring be 

 sufficiently numerous, there will be found a 

 certain percentage' of forms that have no hy- 

 brid character, but various combinations of 

 the pure characters of the two parents; and 



^ ' Heredity, and the Origin of Species,' p. 8. 



' The average ratio of such forms to the whole 

 number of offspring is expressed by the fraction 

 1^ where n is the number of differences between 

 the parents of the first cross; this fraction also 

 expresses the proportion of the- forms that ex- 

 actly resemble the parent hybrid; the remainder, 

 or the oflfspring with reduced hybrid characters, 

 will be, respectively, 0, J, |, f, J|, etc., when the 

 differences of the grandparents are, respectively, 

 1, 2, 3, 4, 5, etc. 



such forms, though often new, will prove 

 stable when reproduced by self-fertilized seed. 

 With this diminution, or entire loss of hy- 

 bridity, we should natiirally expect a partial 

 or total recovery from the impairment of fer- 

 tility produced in the first-cross. At any rate, 

 it is an observed fact that many violet seed- 

 lings, whose hybrid parents produced seed 

 from only about one tenth of their ovules, are 

 themselves normally fertile. 



We have, then, in our blue stemless violets 

 a rather large group of closely allied species 

 that freely interbreed, producing ' blend- 

 hybrids,' that is, hybrids in which the differ- 

 ences of the parents appear in a compromis* 

 form. While it may occur in Viola, there has 

 been observed no instance of what Mendel calls 

 ' dominance ' — the appearance in the hybrid 

 in full force of a character of one of the 

 parents to the suppression of the contrasting 

 character of the other parent. Even differ- 

 ences in respect to pubescence, or in respect 

 to color of capsules or of seeds, are in violet 

 hybrids represented in an intermediate condi- 

 tion, though these differences in other genera 

 usually give rise to dominance. 



Nevertheless, in the behavior of their off- 

 spring many violet hybrids obey the Mendelian 

 law of segregation. The compromise effected 

 in the sporophyte stage between the conflict- 

 ing characters of the parentage, is annulled 

 when the plant passes into the gametophyta 

 stage; the germ-cells are for the most part 

 pure, and the offspring heterogeneous, con- 

 sisting of reversions to the original species, of 

 new and stable forms, and of various hybrid 

 forms. This is by no means a novel phe- 

 nomenon in the history of hybridism. Pro- 

 fessor Castle says," " dominance is purely a 

 secondary matter; it may, or may not, occur 

 along with segregation." Professor Bateson 

 says,° "the applicability of the Mendelian 

 hypothesis has, intrinsically, nothing to do 

 with the question of the inheritance being 

 blended or alternative." Numerous instances 

 might be cited; but it may be questioned, 

 if in any other group of plants the phenome- 



°Mark Anniversary Volume, p. 383. 



" ' Mendel's Principles of Heredity,' Smithsonian 

 Report, 1902, p. 574. 



