72 



improvijStg the quality of wheat. 



Another effect of such selection, as indicated by the foregoing 

 results, would be to increase the yield of grain from each plant 

 when grown under the conditions that obtained in these experi- 

 ments. What the effect would be upon the yield under ordinary 

 field conditions these experiments do not indicate. 



On the other hand, selection based upon percentage of proteid 

 nitrogen alone would not result in securing plants of greatest yield 

 when raised under these conditions. It would, moreover, not result 

 in obtaining plants producing the greatest amount of proteid nitro- 

 gen, nor even of kernels containing the largest quantity of proteid 



nitrogen. 

 Table 12.- 



-Summary of analyses of plants, arranged according to grams of proteid nitrogen 

 in average kernel. Crop of 1903. 



Range of proteid nitrogen in 

 average kernel (gram). 



Below 00030 .. . 

 0.00030 to 0.00040 

 0.00040 to 00050 

 0.00050 too 00060 

 0.00060 to 0(X)70 

 0.00070 to 0.(K)080 

 0.00080 to 0.00090 

 0.00090 to 0.00100 

 0.00100 and over. 



Proteid 



nitrogen 



in average 



kernel 



(gram). 



Num- 

 ber of 

 analy- 

 ses. 



Number 

 of ker- 

 nels on 

 plant. 



Weight (in grams) 

 of— 



0. 0002.509 

 . 0003602 

 . 0004537 

 . 0005406 

 . 0006409 

 . 0007430 

 . 0008538 

 . 0009588 

 .0011578 



14 



42 



80 



116 



59 



24 



9 



1 



11 



257.9 

 266.7 

 409.2 

 341.5 

 310. 3 

 204.9 

 189.1 

 591.0 

 244.9 



Kernels 

 on plant. 



3. 9190 

 4. 6742 

 7. 5309 

 6. 7159 

 6. 7257 

 4.5158 

 4.2480 

 14. 6802 

 6. 6082 



Average 

 kernel. 



0.01.364 

 .01628 

 .01811 

 .01908 

 .02137 

 .02110 

 .02334 

 . 02484 

 .02875 



Percent- 

 age of 

 I)roteid 

 nitrogen 

 in ker- 

 nels. 



1.96 

 2.31 

 2.54 

 2.86 

 3.07 

 3.66 

 3.79 

 3.86 

 4.62 



Proteid 

 rutrogeri 

 in ker-' 

 nels on 

 plant 

 (gram). 



0. 06531 

 . 09644 

 .18644 

 .18440 

 . 19805 

 . 15318 

 .15944 

 . 56666 

 . 27980 



It will be shown later that the determination of gliadin-plus-glutenin 

 nitrogen is a safer guide to the bread-making value of wheat than is 

 a determination of proteid nitrogen, but whether selection should bB 

 based upon the percentage of nitrogen or the total production of 

 nitrogen by the plant, or upon the amount contained in the average 

 kernel, is a question that can not be solved except by trial under field 

 conditions. 



wSome results of experiments with light and with heavy seed con- 

 ducted on large field plots for several years may throw some light 

 on this subject, and are given herewith. 



YIELD OF NITROGEN PER ACRE. 



It is important to know whether the absolute amount of nitro- 

 gen per acre of grain raised is greater in light or in heavy wheat. 



If the absolute amount of nitrogen per acre is less in light than 

 in heavy wheat the supposition would be justifiable that the kernels 

 were immature or had been prematurely checked in their develop- 

 ment. On the other hand, if the amount of nitrogen per acre is 

 greater in the light wheat it would be reasonable to suppose that, as 

 both had been raised under the same conditions, the light wheat had, 

 in part at least, come from plants that possessed greater ability to 

 acquire and elaborate nitrogenous material. 



