HEREDITY 537 



parents is " almost the limit of the reduction of variability, even 

 if the whole back ancestry be selected." He remarks, of course, 

 that the new variability is from the new type, not the unselected 

 type ; but, he adds, " continuous selection does not indefinitely 

 modify variability r , however much it shifts the type." * 



The principal function of selection, therefore, is to alter the 

 type, not to reduce variability, and the facts here cited show 

 the inherent impossibility of " fixing" the type in the sense 

 that individuals will not depart much from it. But, on the 

 other hand, the same principle assures us that, however much 

 we improve by shifting the type, there always remains sufficient 

 variability for still further selection, and as long as variability 

 remains there is hope and possibility for still further improve- 

 ment. We may therefore fix the type by unchanging standards 

 of selection, in the sense that it will remain stationary and not 

 shift, but we cannot " fix " it in the sense of reducing to any 

 great extent the proportion of individuals that will deviate 

 from it. 2 



SECTION XVI POWER OF SELECTION TO PERMA- 

 NENTLY MODIFY TYPES BY THE ESTABLISH- 

 MENT OF BREEDS 



Though selection cannot greatly reduce variability, it is yet 

 immensely powerful in shifting the type, as has been shown, 

 and, if long continued, in so establishing the new type that it 

 will breed true thereafter without selection, 3 as will now be 

 shown. 



This will necessitate a variety of assumptions as to the 

 ancestry back of the parent, according as our knowledge of its 

 character is much or little, and according as it may be assumed 



1 Pearson, Grammar of Science, pp. 458, 472-485. (Italics are mine.) 



2 Ibid. pp. 481-485. 



3 " Without selection" here means absolute freedom from the influence of all 

 laws but those of chance. In practice we never realize these conditions, so that 

 it is always necessary to use enough systematic selection to offset the effects of 

 that degree of natural selection which is found to be always at work in nature 

 everywhere. What is meant is, that by continued selection we soon reach a 

 point at which the inherent variability of the race is powerless of itself to alter 

 the type. 





