582 PRACTICAL PROBLEMS 



after birth, with its attendant losses. In all theory we would 

 prevent the birth of unprofitable individuals, and we succeed 

 nearly in proportion as we are skillful in selection. 



Selection results in absolute increase in quality, not merely in 

 an elevation of the average by eliminating the lower values. 

 We have been told that selection results only in raising the 

 average by cutting off the lower values, but that the upper 

 values are not influenced thereby. This is clearly an error, as 

 will be seen by a reference to any systematic breeding experi- 

 ments and especially to the tables giving the results of selecting 

 corn for high or low protein or high or low oil. 1 Here it is seen 

 that, in the progress of selection, by the use of successively 

 increasing standards, new and JiigJicr values constantly appeared. 

 Not only that, but the principle is still operative after ten years 

 of selection, and the coefficient of variability is not, in most cases, 

 growing less? In general it may be said that the result of 

 systematic selection is to shift the type but not greatly to 

 reduce variability, and when applied to a number of characters 

 at the same time it very clearly and very rapidly defines the 

 type of the strain or breed. 



In breeding the beet for sugar, the cow for milk, the horse 

 for speed, or any animal or plant for any definite quality, there 

 is every reason for believing that we have succeeded in pro- 

 ducing a higher order of excellence than ever arose spontaneously 

 in the race while in a state of nature ; that is to say, we have 

 done more than to raise the average, we have elevated the 

 upper limits. 



The upper limits of improvement. Manifestly this increase of 

 quality cannot go on indefinitely. We cannot breed the horse 

 to be as large as the elephant ; or, if we could, there will be an 

 upper limit somewhere. What will set these limits is an interest- 

 ing question. In some cases, no doubt, the limit would be fixed 

 by purely mechanical principles, 3 in others by physiological 

 restrictions, such as the size of the heart and the labor of 



1 See pages 494 and 496. 



2 Ibid. 



8 For example, there is a mechanical limit to the length of leg, or to the sixe 

 of udder. 



