202 ESSAYS ON WHEAT 



In discussing the chewing test, Dr. Saunders says: 

 " The variations in the quality of the ghiten observed in 

 different plants of the same pedigree are often very great, 

 and it is not claiming too much to say that this simple and 

 apparently inefficient method of testing enables one to se- 

 lect with a fair degree of certainty a few of the best plants 

 each season, and this enormously reduces the labor which 

 would otherwise be involved in multiplying a large num- 

 ber of strains, most of which would ultimately be rejected. 

 That this crude method of testing is infallible, one cannot 

 maintain. The writer has shown, however, by actual bak- 

 ing trials that it possesses sufficient accuracy to be worthy 

 of the very serious attention of all wheat-breeders, although 

 an investigator may have to perform the test several hun- 

 dreds of times before acquiring any proficiency in it. It is 

 not really a very difficult matter to judge in this way, as 

 a rule with fair accuracy, both the gluten strength and 

 the color of the flour which would be produced from the 

 wheat in question ; and, if time were taken to weigh the 

 wheat used and the globule of gluten produced, it would 

 certainly be possible to form a rough estimate of the pro- 

 portion of gluten which the wheat would yield. The 

 writer does not know whether this method of judging the 

 quality of individual plants has been used by other investi- 

 gators or not, but he would strongly recommend it as im- 

 peratively necessary for any one attempting to breed wheat 

 for high quality. It requires some patience and a fairly 

 good set of teeth, but these two attributes can be consid- 

 ered essential to all breeders of wheat. A study of the 

 later sections of this bulletin will make it clear to any one 

 that the usual observations on the color and hardness of 

 the kernels are almost useless for estimating flour strength 

 in breeding new wheats. The chewing test is certainly of 

 great value although it should always be confirmed by 



