494 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



only 7*21 per cent., and the crop average was 8'64. Seeds with as 

 low a percentage as 7*04 occur m the tables. 



As regards the high oil race (3) the selected seeds had 5'39 per cent, 

 (crop average 4*70), and by 1906 the selected seeds had risen to a per- 

 centage of 7'86, whilst the crop averaged 7'37 per cent. Seeds with 

 8'59 per cent, of oil occur in the tables. 



So also with the low oil race (4), the 1896 seeds had a percentage 

 of 4'03 (crop 4'70), but in 1906 the average of the crop was 2*66 per 

 cent., and the seeds sown had only 2'20 per cent. In one case" 

 recorded in the tables the oil percentage amounted to only 1'60 per cent. 



It will be seen, therefore, that in individual cases the two protein 

 races had diverged enormously — 17'67 per cent., as compared with 

 7'04 per cent. So also as regards the oil races. Seeds with 8'59 per 

 cent, occurred in the high oil, and with 1'60 per cent, in the low oil 

 strain. As regards the averages the result has been in all cases 

 satisfactory. 



In order to eliminate the effects of soil, climate, and cultivation, 

 mixed protein and mixed oil plots were arranged, in which seeds of 

 the different races were grown on the same " hill. " The results showed 

 that both protein and oil contents are influenced directly by the seed 

 sown. 



Other experiments have been carried out (5 years) to test differences 

 in ash content of stem, leaf, and grain in these four strains. It was 

 found that there was slightly more ash in the high protein and high 

 oil grains as compared respectively with the low protein and low oil 

 races. 



The amount of protein in the upper stalk and leaves is always 

 higher in the high protein race. The high oil strain has more protein 

 in the grain. The phosphorus contents are also higher in both plant and 

 grain of the high protein race, and in the grain only of the high oil race. 

 The difference in oil contents between the high and low protein races 

 was very slight at first (4'52, as compared with 4'35), but in 1906 the 

 high protein had 5'28 per cent, oil and the low protein only 3*86 

 per cent. 



As regards the average yields of the crops the results in 1906 

 were as follows: Illinois high protein gave 65' 1 bushels of shelled corn 

 to the acre. Low protein yielded 73'2 ; high oil, 66*3 ; and low oil, 83*2. 

 Two unselected standard varieties yielded (" Silvermine ") 75" 7 and 

 (Leaming) 87'9 bushels to the acre. 



Throughout the four years the yields of these races seem to have 

 been, as a rule, lower than those of the best standard varieties. Yet, 

 as regards the amount of protein produced per acre, 65*1 bushels at 

 14"26 per cent, protein means 520 pounds of protein to the acre, which 

 is a higher yield than the Silvermine," of which 75'7 bushels will 

 only give 463 pounds to the acre. 



In-breeding may have had some effect in diminishing the yield, for 

 each of these four varieties is descended from at most three ears. The 

 " high protein " race is descended from a single ear. 



