Nov. 1. 1919 Yellow-Berry in Hard Winter Wheat , 157 



plant or in the same head, indicates that yellow-berry is actually different 

 from ordinary soft wheat. 



Roberts and Freeman (6) suggested that heredity is a strong factor in 

 determining the occurrence of yellow-berry in wheat and that pure 

 varieties could probably be isolated that would produce little or no 

 yellow-berry. To establish the correctness or incorrectness of this view 

 a large number of pure strains of winter wheat were examined by the 

 writer, and the percentage by weight of yellow-berry was determined in 

 each. 



The method pursued was as follows: From each strain of wheat, 100 

 cc. of grain were taken and weighed. The yellow-berry kernels were then 

 separated and weighed. The starchy spots in the kernel almost invariably 

 begin to appear around the germ, or embryo — that is to say, at the lower 

 end of the kernel as it stands in the glumes — and spread from there 

 upward. In no case do the starchy spots begin to appear at the brush or 

 tip end of the kernel. The area of the starchy spots may vary from 

 minute dots to the entire grain. Since the opaque, starchy spots in a 

 flinty, translucent wheat kernel may be large or small, the separation 

 of the yellow-berry kernels must be made according to an arbitrary 

 standard. It was decided to include as yeUow-berry all kernels of which 

 one-half or more of the grain was opaque and starchy. The starchy 

 kernels w^ere separated on this basis and weighed. The flinty kernels — 

 those showing no opaque spots at all — were also separated and weighed, 

 and the residue, if any, was designated as "neutral grains." 



The separating and weighing of the kernels was done by two persons, 

 designated in the table by their initials, " L" and "A," who by long expe- 

 rience became very expert in making the analyses of the samples- By 

 having a part of the samples which were analyzed by one checked by the 

 other it was found that very little difference resulted from the different 

 individual judgments of "L," who did the earlier, and "A," who did the 

 later work. It is therefore concluded that the percentage of error due 

 to the personal equation is negligible, and that it is completely over- 

 shadowed by the positive differences in the samples themselves. 



In all, 164 lots of wheat were studied, of which yy were pure strains and 

 87 were checks or controls. The pure strains were grown in single rows 

 alternately with the controls. All the rows were of the same length, 

 66 feet, and stood 8 inches apart. The variety used for the control rows 

 was not a pure line but was, nevertheless, an unusually pure race of 

 Kharkov, a standard variety long and successfully grown here. All 

 the rows, whether of pure-line wheats or controls, contained 250 grains 

 each, planted equidistant in the rows. The wheat was all grown in the 

 same field, which was divided lengthwise into blocks, and each block 

 into plots separated by narrow alleyways. The plots were all 100 by 100 

 feet in size. In Table I the rows are grouped according to dates of 

 harvesting in 1908. The control row following each pure strain is given 



