YIELD OF NITKOGEN PER ACRE. 



73 



To afford information on this point analyses were made of crops 

 grown from light and from heavy seed. Records of the yields of the 

 plots were kept in each case so that the actual amount of proteid 

 nitrogen contained in an acre of each kind of wheat can be calculated. 

 The number of grams of proteid nitrogen in 1,000 kernels of each seed 

 and crop sample is also stated. The first samples separated, Nos. 78 

 and 79 of the Turkish Red variety and 80 and 81 of the Big Frame 

 variety, were taken from seed that had never before been treated 

 in this wa3^ When planted they produced the crops indicated in- 

 Table 13 by 78b, 79b, 80b, and 81b, respectiveh^. Each of these 

 crops was then separated into two portions, of which the light portion 

 of the light wheat was retained for analyzmg and planting, and the 

 heavy portion of the heavj' wheat likewise retained. Thus No. 383 

 is the light portion of No. 78b, and No. 384 is the hea^n,^ portion of 

 No. 79b. 



The accuracy of the records of relative yields of light and heavy 

 seed harvested in 1902 being open to suspicion, samples of the same 

 seed were sown again in the autumn of 1902 and harvested in 1903. 

 The results from this test are stated at the bottom of the table under 

 the heading ''Check experiment." 



These experiments are to be understood as duplicating those of 

 1902, which, as regards the relative peld of light and heavj* wheat, 

 should be accurate, although tried ill 1903. The difference between 

 this check experiment and the regular one of 1903 is that in the 

 check experiment the seed of the crop of 1901 was used, while in the 

 regular experiment m 1903 the seed of the crop of 1902 was used. 



Table 13. — Crops grown from light and from heavy seed for four years. 



SEED. 



