50 BULLETIN 698, U. S. DEPARTMENT OF AGRICULTUEE. ; 



5 originally under experiment. The total number of plats grown in ! 

 the nine years has been 105. The results are shown in Tables XVI ! 

 and XVII, and the best strains are compared with others of the milo- \ 

 durra group in Table XVIII. Beghming in 1910, an attempt was' 

 made to isolate both black-glumed and white-glumed selections from j 

 some of the selections which were not homozygous for glume color. ' 



From a study of the data in Table XVI it appears that the selec- : 

 tions of 198, 237, and 240 are early in maturing and the selection \ 

 No. 240-6 especially so. The average duration of the growing period \ 

 lies between 100 and 110 days in fairl}^ normal seasons. In 1914. , 

 when all varieties of grain sorghums matured quickly. No. 240-6 | 

 matured in 87 days, and none of the six selections required more than \ 

 97 days to mature. In 1915, when all grain sorghums were late, these i 

 selections required only 118 days on the average, or a little less time] 

 than milo. ] 



The germination of these hybrid selections has been good, on the ' 

 average. Only once in the nine years has the plant space exceeded j 

 15 inches. In five of these years the plant space has averaged less j 

 than 10 inches and in one year less than 5 inches. In the very dry ' 

 spring of 1910 the germination was poor and the average space was i 

 25.6 inches. In 1915, however, with its very wet spring, the average ' 

 spacing was only 10.4 and 15.2 inches for the two subgroups, respec- j 

 tively. These selections do not tiller as freely as their kafir parent, ! 

 but a little more freely than the durra parent. Only once, in 1910, ; 

 have half of the stalks been suckers and only three times have as | 

 many as 30 per cent been suckers. The lowest record is 19.2 per ' 

 cent, in 1911. The average height has been 4 feet. j 



The yields of all varieties are given in Table XVI and the yields \ 

 of the best strains or races in Table XVII. No fair comparison can \ 

 be made by considering the average acre yield of all the selections i 

 in all the years, because these are widely varying hybrid forms i 

 instead of pure-line selections from a mass variety. It is necessary, \ 

 therefore, to determine the average acre yield of each of such selections ■ 

 as have been grown throughout the 9-year period. There are only ; 

 three of them, but there are two others that were separated from ' 

 two of the original three after the experiment had run two years. 

 There is still another, No. 240-6, which has been grown during the 

 last six years. 



As grown during 1908 and 1909, both black-glumed and white- 

 glumed forms occurred in the selections of No. 198. In the autumn 

 of 1909 selections were made for glume color, and in 1910 head rows 

 of each color were sown. The yields recorded in 1908 and 1909 are 

 from the mixed lots, and hence are identical m both sections of the 

 table. The subsequent yields are those given by the different 

 selections. 



