172 SECTIONAL ADDRESSES. 



oolomiug matter authocyaiiin in the petals of plaiits siicli as ilie .Stock 

 and Sweet Pea. Our proof that two iactorti (at least) are here involved 

 is obtained when we find that two true breeding forms devoid of colour 

 yield coloured offspring when mated together. In this case the two 

 complementary factors are can-ied, one by each of the two crossed 

 forms. When both factors meet in the one individual, colom- is 

 developed. We have in such cases the solution of the famihar, but 

 previously unexplained, phenomenon of Reversion. Confirmatory evi- 

 dence is afforded when among the offspring of such cross-bred indi- 

 viduals we find the simple 3 to 1 ratio of the one-factor difference 

 replaced by a ratio of 9 to 7. Similarly we deduce from a ratio of 

 27 to 37 that three factors are concerned, from a ratio of 81 to 175 

 four factors, and so on. The occunence of these higher ratios proves 

 that the hereditaiy process follows the same course whatever the 

 number of factors controlliTig the character in question. 



And here I may pause to dwell for a moment upon a point of which 

 it is well that we should remind ourselves from time to time, since, 

 though tacitly recognised, it finds no explicit expression in our ordinary 

 representation of genetic relations. The method of factorial analysis 

 based on the results of inter-breeding enables us to ascertain the least 

 possible number of genetic factors concerned in controlling a particular 

 somatic character, but what the total of such factors actually is we 

 cannot tell, since our only criteiiou is the number by which the forais 

 we employ are found to differ. How many may be common to these 

 forms remains unknown. In illusti'ation I may take the case of sur- 

 face character in the genera Lychnis and Mailhiohi. In L. vesperthia 

 the type forn: is hairy; in the variety glabra, recessive to the type, 

 hairs are entirely lacking. Here all glabrous individuals have so far 

 pi'oved to be similar in constitution, and when bred with the type give a 

 3 to 1 ratio in F„.' We speak of Hairiness in this case, therefore, as 

 being a one-factor character. In the case of Matthiola incana v. 

 glabra, of which many strains are in cultivation, it so happened that 

 the commercial material originally employed in these investigations 

 contained all the factors since identified as present in the type and 

 essential to the manifestation of hairiness except one. Hence it 

 appeared at first that here also hairiness must be controlled, as in 

 Lychnis, by a single factor. But farther experiment revealed the fact 

 that though tlie total number of factors contained in these glabrous 

 forms was the same, the respective factorial combinations wei'e not 

 identical. By inter-breeding these and other strains obtained later, 

 hairy F^ cross-breds were produced giving ratios in F, which proved 

 that at least four distinct factors are concerned.^ Whereas, then, the 

 glabrous appearance in Lychnis always indicates the loss (if for con- 

 venience we may so represent the nature of the recessive condition) 

 of one and the same factor, analysis in the Stock shows that the 

 glabrous condition results if any factor out of a growp of four is repre- 

 sented by its recessive allelomorph. Hence w'e describe hairiness in 

 the latter case as a four-factor character. 



1 ItepoTt to the Evolution Committef, Iloyal Society, i. , 1902. 



2 Proc. Roy. Hoc. B, vol. 85, 1912. 



