250 Journal of Agricultural Research voi. vi, no. & 



examined the smallest number of sterile spikelets was found on a beard- 

 less variety and the largest number on a bearded variety. 



(3) Early seeding seems to increase the percentage of sterile spike- 

 lets on each spike. Wheat seeded very late had the smallest percentage 

 of sterile spikelets. 



(4) The application of nitrogen alone as a fertilizer produced the lowest 

 percentage of sterile spikelets. Phosphoric acid singly gave the highest 

 percentage of sterile spikelets, while potash was intermediate as to the 

 percentage of sterile spikelets. Where two elements of fertilizers were 

 combined, phosphoric acid and potash gave the highest percentage of 

 sterile spikelets, with nitrogen and phosphoric acid next and nitrogen 

 and potash last. In every instance the check or untreated plots 

 gave a lower percentage of sterile spikelets than those treated with a 

 complete fertilizer. 



(5) There is a distinct correlation between the length of spike as 

 expressed by the number of spikelets and the number of sterile spikelets. 

 As the number of spikelets per spike increases (in other words, the length 

 of spike), the number of sterile spikelets becomes greater. That is, 

 varieties with the shorter spikes tend toward a smaller number of sterile 

 spikelets than the varieties with the longer spikes. However, the per- 

 centage of sterile spikelets per spike may be greater among the varieties 

 with the shorter spikes, as was shown to be the case where spikes of vary- 

 ing lengths within a single variety were examined. 



(6) There is only a very slight correlation between the percentage of 

 sterile spikelets and the number of tillers to each plant. 



(7) The yield of grain per plant is correlated to a fair degree with a 

 low percentage of sterile spikelets. 



(8) The weight of the kernel or quality of grain is correlated to a con- 

 siderable degree with a low percentage of sterile spikelets. 



(9) The yield of grain per spike, the length of spike, and the length of 

 culm are strongly correlated with a low percentage of sterile spikelets. 



(10) There is a slight correlation between the average number of 

 spikelets per spike and a low percentage of sterile spikelets. 



