152 



slender, and tapering; and as the seed bead is pretty heavy it sways back- 

 wards and forwards in the wind, and in storms is very apt to "lodge." 

 Selections made from a rather limited number of canes in which this 

 tendency was partially eliminated gave individuals which were great 

 improvements upon the typical cane of the variety. This shows how a 

 fault may be gradually eradicated by selection of desirable variations. 

 Again, from a cross of the Links Hybrid with the Early Orange, in- 

 dividuals were obtained which retained most of the desirable qualities 

 of the former, with its typical seed head, and engrafted upon the stout, 

 stocky canes of the latter. This shows the breeding out of an undesir- 

 able quality by crossing. The photographs which show the canes ob- 

 tained by these two diiferent methods contrasted with typical canes of 

 the variety, illustrate very graphically the possible progress that can 

 be made in two generations in the improvement of a variety in form. 

 Faults in form are so readily seen, that it is much more easy to eradi- 

 cate them by selection than faults of composition, which can only be 

 ascertained by chemical analysis. Only such canes, then, should be 

 taken for analysis as show not only freedom from general faults, but 

 also a tendency towards elimination of faults of the variety to which it 

 belongs. 



It may be as well to insert here a caution as to the use of crossing. 

 It has been shown that sorghum is extremely variable, and this fact is a 

 sort of guaranty that by continued selection improved varieties can be 

 created, for variation makes selection passible, and selection makes im- 

 provement possible, but care should be exercised in making use of this 

 tendency. There is a well founded prejudice against u mixed " varieties 

 of sorghum. The most worthless men, animals, and plants are those 

 which belong to heterogeneous and indiscriminately mixed races. Bad 

 qualities are transmitted as well as good. The most of the crosses 

 grown at this station were worthless. Darwin says, " A variety may be 

 variable, but a distinct and improved race will not be formed without 

 selection." After the desired degree of variation in the variety has 

 been obtained selection should be based upon uniformity rather than 

 variability, in order that the qualities may become fixed and stable. 



The most careful and rigid precautions should be taken against acci- 

 dental crossing, none being permitted that is not artificially controlled 

 by methods well known to horticulturists. It would probably be well 

 to prevent cross fertilization even in plots of the same variety. Inten- 

 sive cultivation has yet to be tried on the sorghum plant, and perhaps 

 where there is already so great a range of variability, there is greater 

 prospect of improvement by selection and self-fertilization than by 

 crossing. It would certainly be best in crosses, and probably best in 

 varieties, to plant single plots from but one seed head. 



In selecting the seed from the best individuals by analysis of the 

 juice, not only the percentage of sugar, but also the purity and the 

 percentage of glucose must be considered. This problem is rendered 



