218 



The Journal of Heredity 



nine and fifty-one per cent of central 

 stem, which was taken to be sufficient 

 evidence that this was a high percentage 

 stem producer. It was then decided to 

 try selection in the opposite direction 

 by choosing seed heads which contained 

 large central stems. The season of 

 1920 was a "stemmy season." as ob- 

 served above, and these stemmy heads 

 produced a crop that had ninety-six 

 per cent of stem. The next year's 

 crop contained fifty-nine per cent. A 

 similar selection for stem had been 

 started a year earlier in strain 45 and 

 it produced fifty-seven and eighty-three 

 per cent of stem in the years 1919 and 

 1920 respectively. All progeny of 

 original plant No. 4 have shown a ten- 

 dency toward high per cent of stem. 

 Strain 45, however, produced some 

 substrains that were fairly low in stem 

 and others that were high, indicating 

 that it was heterozygous for stem con- 

 tent, while strain 4 seemed to be rela- 

 tively homozygous. 



It seems, therefore, that the open 

 pollinated heads selected in 191 7 were 

 largely homozygous. Subsequent head- 

 row cultures made possible the sorting 

 out of those which were most nearly 

 homozygous for low stem content, and 

 at the same time of desirable quality 

 in other respects. 



The conception of the broom corn 

 crop, gained from this experiment is 

 that for breeding purposes it should be 

 classed between the self-fertilized crops 

 and the open-pollinated ones. Prac- 

 tical breeding, then, would begin as 

 usual with a study of available types 

 and varieties. From the best varieties 

 a considerable number of desirable 

 heads would be selected and given a 

 l^rogeny test in head-rows. After one 

 or two years in head-row tests the 

 better strains could be composited and 

 increased ; or a single strain might be 

 increased for commercial planting. Bag- 

 ging of heads is easily done and is 

 probably worth while to insure self- 

 pollination. There is no a])parent re- 



duction in vigor due to selfing, which 

 is to be expected since it is shown that 

 a large amount of natural selfing oc- 

 curs. Mass selection without a progeny 

 test in head-rows would accomplish 

 little, or at best would make but slow 

 progress in broom corn, for it has been 

 shown that many good quality heads 

 produce a mediocre or low-grade prog- 

 eny. The head-row test detects these 

 and makes it possible to distinguish 

 between strains whose parents looked 

 alike but were genetically different, as 

 proved by their ofi^spring. 



It seems to be clear that central stem 

 in broom corn is an inherited character, 

 for different strains produce it in dif- 

 fering amounts ; yet none has been 

 found that produces a progeny free 

 from this undesirable feature. Possibly 

 multiple factors are concerned ; in 

 which case crossing of strains with low 

 percentage of stems, followed by obser- 

 vation of numerous head-rows, should 

 discover the desired stemless individ- 

 uals. 



That production of central stem is 

 closely tied up with the physiological 

 condition of the plant is indicated by 

 our observations. When for any reason 

 ■ the head fails to emerge from the leaf 

 sheath, or is retarded in its development 

 by accident, a branch usually starts 

 from one or more nodes of the stalk. 

 These branches usually attain a length 

 of several feet and produce a small, 

 worthless brush that invariably has a 

 central stem. Stem is always present 

 in these branch heads, even on plants 

 of which the partially developed main 

 head was free from central stem. It 

 may be that the forces which cause or 

 allow stem to 1)C produced in the heads 

 of these branches are the same as those 

 involved in the seasonal effects noted 

 above. But further experiment would 

 be required to make this point clear. 



The method mentioned above of 

 eliminating seasonal influence by grow- 

 ing two or more generations in the 

 same season a])])eals to the writer as 



