BRITISH ASSOCIATION. 307 



failure of a parental type to reappear in its completeness after a 

 cross — the merino sheep or the fantail pigeon, for example. These 

 exceptions may still be plausibly ascribed to the interference of a 

 multitude of factors, a suggestion not easy to disprove ; though it 

 seems to me equally likely that segregation has been in reality 

 imperfect. Of the descent of quantitative characters we still know 

 practically nothing. These and hosts of difficult cases remain 

 almost untouched. In particular the discovery of E. Baur, and the 

 evidence of Winkler in regard to his " graft hybrids," both showing 

 that the sub-epidermal layer of a plant — the layer from which the 

 germ-cells are derived — may bear exclusively the characters of a part 

 only of the soma, give hints of curious complications, and suggest 

 that in plants at least the interrelations between soma and gamete 

 may be far less simple than we have supposed. Nevertheless, speak- 

 ing generally, we see nothing to indicate that qualitative characters 

 descend, whether in plants or animals, according to systems which 

 are incapable of factorial representation. 



The body of evidence accumulated by this method of analysis is 

 now very large, and is still growing fast by the labours of many 

 workers. Progress is also beginning along many novel and curious 

 lines. The details are too technical for inclusion here. Suffice it to 

 say that not only have we proof that segregation affects a vast range 

 of characteristics, but in the course of our analysis phenomena of most 

 unexpected kinds have been encountered. Some of these things 

 twenty years ago must have seemed inconceivable. For example, 

 the two sets of sex organs, male and female, of the same plant may 

 not be carrying the same characteristics ; in some animals characteris- 

 tics, quite independent of sex, may be distributed solely or predomin- 

 antly to one sex ; in certain species the male may be breeding true to 

 its own type, while the female is permanently mongrel, throwing off 

 eggs of a distinct variety in addition to those of its own type ; 

 characteristics, essentially independent, may be associated in special 

 combinations which are largely retained in the next generation, so that 

 among the grandchildren there is numerical preponderance of those 

 combinations which existed in the grandparents — a discovery which 

 introduces us to a new phenomenon of polarity in the organism. 



We are accustomed to the fact that the fertilised egg has a 

 polarity, a front and hind end for example ; but we have now to 

 recognise that it, or the primitive germinal cells formed from it, may 

 have another polarity shown in the groupings of the parental elements. 

 I am entirely sceptical as to the occurrence of segregation solely in 

 the maturation of the germ-cells, :|c preferring at present to regard it 

 as a special case of that patch-work condition we see in so many 

 plants. These mosaics may break up, emitting bud-sports at various 

 cell-divisions, and I suspect that the great regularity seen in the F 2 

 ratios of the cereals, for example, is a consequence of very late 

 segregation, whereas the excessive irregularity found in other cases 



* The fact that in certain plants the male and female organs respectively 

 carry distinct factors may be quoted as almost decisively negativing the 

 suggestion that segregation is confined to the reduction division. 



