EXPERIMENT 8TATI0X IfEPOHT. 449 



was planted with seeds selected for having the least amount of 

 color, while another lot, showing the largest percentage of rwl, was 

 planted in a plot a long distance from the former. From the crop 

 secured from each of these plots a second selection, in manner as 

 above stated, Avas made, and as a result it was stated in the last 

 annual report that "the light seed from the light seeds are nearly 

 as dark as those from the second generation of selected dark seeds." 

 The present year a block of plants has been grown from the light- 

 est seeds that are descendants of the seeds selected for their large 

 percentage of light color, and another from similarly selected 

 dark seeds. A sample of the product for the latter is shown at 3, 

 ill Plate XII. A similar sample of the light seeds is shown at 5, 

 while the original unselected 'Mackson Wonder" is placed Ijetween 

 the selected seeds at 4. 



It is seen that selection has removed much of the white in the 

 case of the series of dark seeds, while the opposite is not so evident, 

 namely, the seeds after three years of selection for whiteness are 

 nearly as much colored as the regular stock of the variety. 



Crosses of the '* Jackson W^onder" Lima. 



The 'Mackson Wonder" lima has been grown in the Gardens for 

 some years in parallel breeding rows with other varieties in the 

 hope that it would cross in the ordinary course of nature — the 

 working force not being sufficient to carry out the plan by artificial 

 methods. Two rows, six inches apart, one of '^Jackson Wonder" 

 and one of the sort to be bred upon, were planted, isolated from 

 all other lima beans, and seed from the row with white seeds was. 

 planted in considerable quantity the next season. In this way 

 plants with ^'marked" seeds were obtained last year (1905) with 

 several kinds of lima beans. From the crop coming from these 

 strij>ed or mottled beans — ^and frequently they were of a solid 

 purple color — certain of the crosses shown in Plate XIIT. have 

 been selected. 



Sample seeds of the cross of ^'Jackson Wonder" upon "Burpee" 

 are shown in the upiper rows ; ten seeds of the male parent are at 

 the left (1) and the same number of the mother at the right (5), 

 the piles between representing the crosses — ^three in this instance — 

 each pile in all cases being the full product of a single plant. The 



29 



