PLANT HYBRIDIZATION BEFORE MENDEL 279 



work of the various breeders of the cereals, Garton Brothers, 

 Rimpau, Farrer, Vilmorin, and others. The statement follows : 



"Sachs remarks that Kolreuter, the first man to produce plant hybrids, 

 covered the ground so completely that subsequent investigators have 

 added little to his results." 



The comment is then made : 



"But quantitative investigations have been too seldom undertaken. It 

 seems to me that they are not unimportant." (p. 93.) 



The quantitative results in question follow in very accurately 

 arranged detail, in 14 tables, covering the quantitative distribu- 

 tion of the types of the second generation. The characters of 

 the heads involved are, long and short, bearded and beardless ; 

 velvet (pubescent) chaff, and glabrous chaff ; brown-colored and 

 light-colored chaff. The investigator had no conception at the 

 outset, as had Mendel, of consciously crossing contrasting charac- 

 ter-pairs as such, and unfortunately did not take note of the 

 fact of dominance in the case of the bearded-beardless, and 

 pubescent-glabrous crosses. This was unquestionably due to the 

 fact that length of head, the most salient character, did not show 

 Fj dominance for long X short crosses, but intermediacy. 



In all the tables, the numbers which reproduce the characters 

 of the first generation are printed in heavy type, so that there is 

 statistical evidence of the dominance of characters involved, al- 

 though no reference is made to it as such. 



The individual columns give, in exactness and detail, the dis- 

 tribution of the plants in classes, according to the head-characters, 

 but there is no summary of the proportionate numbers of these 

 types. With the total available data obtained, it would have been 

 possible for Spillman to have not only verified F^ dominance for 

 beardlessness over beardedness, pubescent chaff over glabrous 

 chaff, and brown pubescence over light pubescence, but also to 

 have determined the ratios of the distribution of those characters 

 in the second generation. A few of the numerical results follow, 

 summarized from some of the tables. 



The data comprising Spillman's results are given in fourteen tables 

 (pp. 94-98 of the memoir). The principal data from these tables which 

 may be taken as examples of his Mendelian ratios, are those dealing 

 with the inheritance of length of spike, awns, pubescence of the glumes, 

 and color of glumes. In all of the tables, the progeny are classified in 

 percentages, first, as to long, semi-long and short (head-length charac- 



