64 On Mendels Categories 



with 40 Ol- .30 hairs per s(]uare ceiitiinctre is obviously and unquestionably hairy ; 

 it is just as clearly intermcdiate between a plant with 1100 haii-s per centimetre 

 and a glabrous plant. Without some proof that the mean nuniber of hairs per 

 Unit aroa was not less in the offspring of a cross between a glabrous and a hairy 

 form than it was in the hairy parent, the result described cannot be distinguished 

 from that most probable on any theory of "blended" inheritance. 



It is most unfortunate that the degree of hairiness of each individual was not 

 recorded, beoanso the great variability of the parental races would have given an 

 excellent opportunity of donionstrating the trutli of the statemeiit that pure 

 domiuants "are not merely like, but identical with the pure parents." We might 

 at least have hoped to Icarn whcther a Li/clniis diui-ivi with 40 hairs per centi- 

 metre, and a L. vesperiina with 1100 or 1200, produce plants with ditlercnt degrees 

 of hairiness when crossed with the same glabrous variety, or whether the presence 

 or absence of gland hairs on the leaves of the hairy parent has any effect upon 

 this "charactcr of leaf surface" in the offspring I 



Unless all the hairy parents used, of either race, were equally hairy, the first 

 cross-bred generation must either have contained individuals of different degrees 

 of hairiness, or the idcntity between pure dominants and their parents disappears. 

 On the other hand, if the cross-breds and their parents differed in this character 

 we have no way of distinguishing "dominance" from the result of "blended" 

 inheritance, until we abandon the Mendelian categories, and adopt a rational way 

 of measuring hairiness. The ab.sence of glabrous cross-breds will not help, until 

 we know the variability of the hairy race, and the ch.iracter of the hairy parent; 

 for the chance of obtaining a glabrous individual aniong 1000 plants of any ordinary 

 race of either species is admittedly sniall, and tiie variability of a few faniilies, in 

 each of which one parent is of fixed character, will on any hypothesis be less 

 than the variability of the race, and the chance that sucii a series of families would 

 contain a glabrous individual, on any theory of inheritance, cannot be estimated 

 without information which is at present not available. 



The difficulties whicli arise from iniperfect description when we consider the 

 question of dominance in the first cross-bred generation are equally formidable in 

 the case of later generations. The category hairy is so wide that it is impossible 

 to judge how the individuals included in it resemble or differ from their parents. 

 The total resulta of the various crosses recorded are in better agreement with 

 Mendel's Statements than is usuai, but the published data do not afford material 

 for discussing the question how far any particular theory of inheritance can be 

 successfully applied to them. 



It is deeply to be regretted that so man}' iiitercsting e.xperiments, involving so 

 much time and labour, should be recorded in a form which makes it impossible to 

 understand the results actually obtained, and so gives rise to misconceptions both 

 in the ininds of the recorders and in others. Such justification as there may prove 

 to be for classifying the form of inheritance exhibited by the hairs of Lychnis with 



