214 The Laxo of Ancedral Ilcrcditn 



one of eight childreti light bliic (colour of patenial graiidfather). Father black 

 and mothcr dark brown, onc child out of threc hazel (the colour of paternal 

 grandmother). Father and niother very dark brown or black, two childrcu blue 

 and one black. Father and mother very dark brown or black, two out of four 

 children blue (colour of niaterual grandmother). It is clear that noither end 

 of the eye-colour scale givcs a recessive character wliich breeds true to itself 

 Now notc what we get if we attempt to select for several generations paying 

 attention tu ancestry. Father and mothcr blue, all four grandparents blue, and 

 the fivc recorded grcat-gnuidparents blue, all four children blue. Parents and 

 grandparents all blue, all eight children blue. Father and mother light grej', 

 and the four grandparents all light grey or blue, all seven children grcy. Father 

 and mothcr grey and all the knovvn ancestry of both grey or blue, all four 

 children grey or blue. Father and mother both dark brown, all known ancestry 

 brown for both, out of nine children seven dark brown, one brown and one light 

 brown. Father and mother hazel, all ancestry to grcat-grandparents hazel witli 

 the exception of a blue-eyed grandfather, all fivc children hazel. 



Lastly let us look at crosses between blue and dark eyes. Father light blue 

 and mothcr dark brown, three children blue, four children dark brown. Father 

 dark brown and mother blue, two children blue and four brown-eyed. Father 

 dark brown and mother blue, three brown and six blue-eyed children ; here all 

 mother 's known ancestry was blue. If blue werc "dominant" such results are 

 impossible unless we suppose evory dark brown pareiit above recorded was a 

 hybrid. If blue be "recessive" it ouglit to breed true, but we have seen that 

 it does not. But as a matter of fact we cau pick out cases from the record 

 of one blue and tme ver}' dark ])arent giving almost every conccivable result, 

 e.g. father light blue and mother black, four children blue and one light grey. 

 Father brown and mother blue, eight children ranging from brown to dark brown. 

 In all cases an examination of ancestry throws light, if it does not fully explain in 

 each case what h;is taken place. It is clear from these cases that parentage 

 will not enable us to predict offspring definitely. But what is also equally clear is 

 that if all the ancestry be the same the otfspring may differ among themselves 

 considerably, e.g. by the presence or total absence of black pigmentation. Hence 

 we reach our second conclusion : 



{h) That a knowledge of the whole ancentrij while it certainly limits tlie ränge 

 of Variation does not absolutely define the character of the offspring. 



(2) Before wo leave the cases above it is just worth reiterating that nothing 

 corresponding to Mendels principles appears in these characters for horscs, dogs, 

 or men. If black or grey coat-colinir in liorses were " recessive," when two blacks 

 were matcd we should cxpect only hiack offspring, but black can disappear for a 

 generation or even two and then reappear. Or, take a case like that of a grey 

 horse Viscount, where grey remained dominant for three generations only to 

 di.sappcar before the chestnut of the mare Blue Stocking in the Viscount and Blue- 

 Stocking filly Miss Johanna l It is the same with cvcry coat-colour taken, its 



