Wheat Investigations. 17 



Minnesota wheats were retained. These were grown in the sea- 

 son of 1918 in plots ranging from one two-hundredth to one- 

 fortieth acre in area. A further selection in the fall of 1918 

 reduced the number of strains to 26 originally selected from 

 Aroostook wheats and to 5 Minnesota lines. A complete chem- 

 ical analysis of 37 lines and baking tests of 31 wheat lines were 

 made in the spring of 1919. Using the baking test as the final 

 index of strength, out of the 31 tested strains 12 Aroostook lines 

 and 4 Minnesota lines were retained and propagated in 1919 in. 

 one-twentieth acre plots. 



Analysis of Data, 

 physical characteristics of the grain. 



The most striking feature in the physical appearance of the 

 grain from the Aroostook grown lines is the size and weight 

 of the kernels. Their size is larger than that of spring wheat 

 of the Northwest and their plump, well rounded shape, distin- 

 guishes them from the wheats of the prairie regions. The size 

 and shape relations of the grain of the Aroostook pure lines are 

 illustrated in Figures 5 and 6, which represent kernels of four 

 pure lines each belonging to a distinct wheat variety. As a re- 

 sult of these size and shape relations the weight of the kernels 

 is correspondingly high. In Table 2 is given the weight per 

 1000 kernels of each of the pure lines grown in 1918. The lines 

 are grouped according to the variety from which they originated 

 and their rank in respect to kernel weight. 



It may be well to point out that the determination of the 

 kernel weight based on 1000 kernels expresses very exact rela- 

 tions repeated determinations showing the experimental error 

 to be either negligible or nil. Therefore* the differences in the 

 iooo-kernel-weight of these lines are more significant than 

 they might appear at first sight. From Table 2 it will be 

 noted that the average weight per 1000 kernels for all lines is 

 quite high — 35.314 grams. The average weights per 1000 ker- 

 nels for the varieties do not show great differences though they 

 very well reflect varietal means around which the strains with- 

 in the varieties are grouped. The individual strains show a 

 considerable deviation from the average of the varieties from 



