Selection 379 



crop from the most productive rows, or the remaining kernels 

 of the ears that produced the most desirable rows are sown the 

 following year. This procedure is repeated over a period of 

 years. In these experiments, the desirable types were deter- 

 mined by chemical analyses of some of the fruits of certain ears^ 

 and the remaining fruits of the extreme ears were planted. 

 Curves of oil and protein content show that great progress is 

 made during the first year, that after a few years progress be- 

 comes slower, and that finally a horizontal line results showing 

 that after a w^iile no further progress is made. After ten gen- 

 erations the average crop of the high protein line had 14.26 per 

 cent protein and of the low line 8.64 per cent. The original 

 line was slightly under 11 per cent, which shows that consid- 

 erable progress was made in selecting in each direction. After 

 the first ten years, however, practically no further gain was 

 made even when selection was continued as before. 



Results were slightly different when selecting for oil. Starting 

 w^ith a strain that showed about 4.6 per cent oil, the low oil 

 strain reached 2.66 per cent after ten generations, declined to 

 about 2 per cent after seventeen generations, and thereafter 

 remained essentially at that level. The high oil strain reached 

 7.37 per cent after ten generations, was close to 8 at the seven- 

 teenth generation, and continued to increase slightly after that. 



East's explanation for the results obtained in this selection 

 experiment assumes the original stock was a mixed race con- 

 taining a number of different biotypes which were rapidly iso- 

 lated by selection. After this isolation was complete, which 

 came about when the strains became essentially homozygous, 

 selection accomplished nothing. The ear-to-row method of 

 breeding appeared to be a very excellent method when it was 

 introduced, but it has very decided limitations. M. T. Jenkins 

 in the 1936 Yearbook of Agriculture, for example, points out 

 that it has been effective over a period of a few years in increas- 

 ing the yield of plants that are relatively unselected to begin 

 with but that there was no evidence that this method had any 

 cumulative effect. 



An interesting development of the maternal-line method was 

 reported for alfalfa by Fryer in 1939 at the University of 

 Alberta in Canada. This method involves the use of four-year 

 cycles. During the first year about 80 progenies containing ap- 



