TESTS IN TEXAS. 41 



In the Corsicana test four of the hi<!;hest yielding crosses are pro- 

 geny of four of the highest yielding female parents. Among the 

 crosses showing the greatest percwitago increase over the better 

 parent three are identical with tlu'ee of the highest yielding crosses, 

 two are progeny of the high-yielding female parents, and two are 

 progeny of the poor-yielding female parents. Of the poorest yield- 

 ing crosses, four are progeny of four of the poorest yielding female 

 parents. Out of the crosses showing the greatest percentage of 

 decrease from the better parent, one is the progeny of the poorest 

 yielding female parent, and two are the progeny of two of the liigh- 

 est yiekUng female parents. Two are identical with two of the 

 poorest yielding crosses. 



It is evident in these tests that in general the most productive 

 crosses have come from the most productive female parents; that 

 the crosses showing the greatest percentage of increase over the 

 better parents are in most instances identical with the highest yield- 

 ing crosses. Likewise, in most cases, the poorest yielding crosses 

 are from the poorest yielding female parents, as are also most of the 

 crosses that show the greatest percentage of decrease in yield. There 

 is no instance in which one of the highest yielding crosses springs 

 from one of the poorest yielding female parents. 



There are, however, both in the test at Waco and at Corsicana, 

 two crosses (Gorham Yellow X Chisholm and McCuUough X Chisholm) 

 which stand liigh in the list of crosses showing greatest percentage 

 increase over the better parent; they rank among the crosses in 

 actual production per stalk sixth and eighth, respectively, at Waco, 

 while at Corsicana the rank is seventh and eleventh. The female 

 parents of these crosses are among the poorest yielding varieties at 

 both places. These two comparisons are the most striking among 

 the crosses in regard to increase over the parent varieties, but their 

 ■ actual yield is less than the best producing original varieties in the 

 test, and there would be no incentive to grow them commercially. 



These results would indicate the most promising method of secur- 

 ing increased yields to be by crossing the liighest yielding varieties, 

 but such a method would necessitate the previous testing of the 

 varieties to ascertain the highest yielding, and the crosses could not 

 then with certainty be depended upon to give increased yields, as is 

 seen in the case of the cross Surcropper X Chisholm, wliich is less 

 productive than Surcropper at both Sherman and Corsicana and but 

 httle more productive at Waco. Selection 137 X Cliisholm is less 

 productive than Selection 137 (the better parent) at Sherman, and 

 less than Cliisholm (the better parent) at Waco and at Corsicana. The 

 crosses Gourd Seed X Chisholm and Selection 136 X Cliisholm w^ere 



218 



