ESTIMATION OF STRENGTH 167 



was an inheritable characteristic. By some lucky 

 accident Red Fife was chosen as one of the wheats 

 from which hybrids were to be raised, for we did 

 not know at the time that it would retain its 

 quality under our conditions. The problem on 

 which so much turns is one which presents many 

 difficulties to the plant-breeder. In the first place 

 we have no methods of determining with absolute 

 certainty whether a wheat is strong or not except 

 by milling and baking it. Such tests involve the 

 use of large quantities of grain. Baking trials 

 are not at the best extremely accurate, and with 

 unknown flours several trials have often to be made 

 in order to hit off the best methods of handling 

 the doughs. As a result one has to be prepared 

 to sacrifice about half a hundredweight of grain 

 in order to secure a result which is accurate within 

 one or two per cent. This degree of precision is, 

 as we have frequently found by the addition of 

 duplicates in a series of trials, attained by a skilled 

 baker familiar with the intricacies of testing work. 

 Such an amount of grain is, however, beyond the 

 resources of a plant-breeder in the earlier stages 

 of his work, as for instance in the F 2 generations 

 where he requires to know the strength of the grain 

 of individual plants, though with good management 

 it can be obtained in the F 4 generations. In 

 practice, as for instance when purchasing imported 

 wheats, strength is largely judged by the appear- 

 ance of the sample. The points to which attention 



