174 HEREDITARY CHARACTERS 



under ordinary circumstances, breed true. The first gene- 

 ration had rose-coloured flowers with red streaks. The 

 hybrids, however, produced many different varieties, some 

 white, some red, some rose, some yellow, and some yellowish. 

 All these colours occurred pure or with various streaks. 1 

 Similar results have been obtained by crossing varieties of 

 sweet peas. 



If the colour characters are represented by entities, as is 

 suggested so strongly by the alternative distribution, it seems 

 obvious that there must be a good many entities all repre- 

 senting colour, but which need not necessarily represent the 

 same colour. Where a particular colour has been the object 

 of man's selection in a domesticated race, it would seem 

 probable that although his selection may have reached a 

 point where the majority of the colour entities represent the 

 desired colour, it can but rarely happen that the selection 

 has been so stringent and so long continued, that all the 

 colour entities are the same. Take the case of a cross be- 

 tween a grey and an albino rabbit, the albino being derived 

 from a black variety. From this cross, black, white, and 

 grey rabbits can be bred out. This appears to show that 

 the black in the albino remained discrete, although the 

 albino character concealed it. In the first cross the pro- 

 geny are grey, but the albino and black characters are also 

 present. Assume 20 as a convenient hypothetical number 

 of entities representing individual variations in colour ; A 

 representing albinism, and B black. Now even if the albino 

 character is represented by only one of these entities, if it 

 be subjected to selection it will in time completely oust 

 all the other 19 entities which are not selected. But it is 

 very unlikely that in an artificial breed of rabbits the colour 

 entities can have been so selected that they are all of one 

 kind. In this particular case the black character is present. 

 Suppose that in this rabbit the colour entities are 10A+5B 

 +5X ; X representing all the other colour entities, which 



1 Correns, C., "Uber Bastardierungsversuche rait Mirabilis sippen," Ber, 

 Dcutsch. Bot. Ges., xx., 1902, 



