16 BEEEDIXG DEOUGHT-BESISTAXT FOEAGE PLAXTS. 



secure comparable results of yields per plant from each progeny row, 

 as the end plants, because of their favored position, make a larger 

 growth. The purpose is to secure accurate comparative yields of all 

 the progeny rows. 



Later in the season, when the seed is ripe, the superior individuals 

 are selected as mother plants to furnish seed for planting the following 

 season and thus continue the work of selection. 



After the superior plants have been selected the bulk of the plants 

 in the row are harvested, dried in shocks, weighed, and thrashed. 

 Since a record is kept of the number of plants harvested, an accurate 

 estimate of the producing power per plant of each row is easily made. 

 The yields of the progenies grown at Beiiefourche during the season 

 of 1909, which are in the fourth generation of selection, are presented 

 in Table I. 



UNIFORMITY OF PLAXTS EN THE PROGENY ROWS. 



Breeding work with a plant like alfalfa has the special advantage 

 that one is able to compare living plants belonging to different genera- 

 tions of selection. Alfalfa being perennial, the mother plants can be 

 retained in their original places in the breeding nursery for com- 

 parison with their progeny. Thus, the degree in which the progeny 

 has inherited the desirable characters of the mother plant can be 

 checked by direct comparison. In general, there has been great uni- 

 formity in the rows although they are the progeny of plants that 

 were selected without any precaution to insure close pollination. 

 As shown in Table I. in 29 out of 36 progeny rows harvested sepa- 

 rately, in which the plants "off type'" were discarded, over 80 per 

 cent of the plants in each row were harvested as uniform in type. 

 Some prominent types may be noted, as E-2, in which the plants 

 were very erect, rather slender, and only moderately branched, and 

 had dark-purple flowers. This is a rather distinct, easily recognizable 

 form and it will be noted that 84 per cent of the plants in this row 

 conformed to the type. B-l is another distinct type; the plants are 

 tall, coarse, slightly branched, and woody, with very light purple 

 flowers fading to white. Of the plants in this row 92 per cent were 

 typical. In some progeny rows the variation in type of plant has 

 been great, but in general the uniformity is close enough to show that 

 this method of simple selection without isolation can give valuable 

 results in breeding alfalfa. 



a Heretofore the plants have not been inclosed with, screens to insure self-pollination; 

 but it is the plan in future work to inclose a number of plants and pollinate them by 

 hand and thus get a comparison of the uniformity of progeny of screened plants and 

 those which are exposed in the normal way to the chance of cross-pollination by 

 insects. These screens will be placed over the plants at the beginning of the bios. 

 soming period. Hitherto the only distinction made with superior plants has been to 

 harvest them separately at the time the seed matured. 

 196 



