240 REPRODUCTION 



variation in leaf and flower combinations: yellow flowers and 

 short leaflets, yellow flowers and long leaflets, blue flowers and 

 short leaflets, blue flowers and long leaflets. The offspring of 

 D^ will all have smooth stems and terminal flowers, but they also 

 will show variations with respect to flower color and form of 

 leaflets, as in the case of Dg. 



Necessity of Pedigree Cultures. — Our diagrams will illus- 

 trate for us how it is that pedigree cultures are a necessity in 

 arriving as soon as possible at a strain of any particular type 

 capable of coming true to seed. Suppose that one in looking 

 over his field finds plants of the types shown in the D row of 

 our figure, and wishes to propagate the sort shown in Dg and D^. 

 To be certain that no intermixture comes in from the other 

 varieties, he ties bags over the flowers of his chosen plants so 

 that all foreign pollen is excluded. Self-pollination having taken 

 place, he saves the seed from all his chosen plants and puts 

 them together in one package. The following year on planting 

 these seeds, all from plants seemingly alike and self-pollinated, 

 he is surprised to find the four types K, L, M, and N appearing 

 in his seedbed, the impurity having come, as we can see, from 

 D^ plants, which, so far as he could know at the time, were the 

 same as the D^ plants. But now, if in harvesting the seeds he 

 had put those from each plant into a package by themselves, 

 and had planted them in separate plots, the pure and the impure 

 strains would have made themselves at once manifest. 



Since the parents of the hybrid we are now discussing had four 

 pairs of contrasting characters, we would expect, in accordance 

 with the formula previously given, sixteen varieties of offspring 

 of the hybrid, each showing a different combination of char- 

 acters from any of the others. These are shown in Fig. 135. 

 A great number of seeds would have to be produced to give these 

 various chance combinations opportunity to be made. More- 

 over, in the long run all of these varieties would be produced cap- 

 able of coming true to seed. This would happen when each 

 was the product of a homozygote fertilized egg cell, as is the case 

 in D J and D, of Fig. 134. This possibility is relied on by Burbank 



