RELATIVE STABILITY OF LIVING MATTER 



297 



favorable and adverse, yielding but slowly, and sometimes not 

 at all, to modifying influences, and often suffering extinction 

 when a slight modification would have resulted in preservation. 

 Thus the oaks and the tulip tree have come down to us from 

 remote ages practically unchanged, and the elephant is with us 

 yet, substantially the same as he has been for probably thou- 

 sands of years. 



And yet there is constant variation, in these as in more flexi- 

 ble species. They have not freed themselves from variability, 

 even though the species as a whole has come to be remarkably 

 constant. Indeed, the more the question is studied, the more 

 evident it becomes that a great deal that passes for variability 

 is merely individual fluctuation around a practically stationary 

 point, not necessarily involving actual change in type. That 

 is to say, few individuals exactly reproduce the type of the 

 species, however fixed it may have become ; most of them 

 depart slightly this way or that, making a great show of varia- 

 tion, so that we seem to be in the midst of bewildering differ- 

 ences, even though the type is practically unchanged. In cases 

 of this sort deviations represent not so much departures from 

 type as individual approximations to a general average. 



Here is ground very deceptive to the breeder. Generally 

 speaking, variation denotes flexibility of organization, and there- 

 fore possibility of improvement, but the breeder must not assume 

 that great variation denotes large possibility for improvement. 

 Fundamentally it denotes quite as much an inherent failure to 

 assume a distinct type ; and often a lesser deviation, repre- 

 senting a true departure from type, affords a far more favorable 

 basis for improvement than do those deviations that after all 

 are merely fluctuations about a center that has a strong tend- 

 ency to remain fixed. 1 



1 Pearson believes that the extent of variability cannot be reduced more than 

 about 1 1 per cent, however rigid the selection. He does not claim that the type 

 cannot be shifted more than that amount, but that, however much it may be 

 shifted, there is still variability about the new center, and that this variability is 

 at least 89 per cent of the original variability of the race. See Pearson, Grammar 

 of Science, pp. 481-485 ; also chapter on " Selection." This subject will be fully 

 studied in a succeeding chapter entitled " Type and Variability." 



