ADAPTIVENESS AND PURPOSIVENESS 327 



controlled by the past to hold fast to that which is good. 

 It is a very interesting fact that some monsters have been 

 experimentally produced by disharmonious mongrel fertilisa- 

 tion of egg-cells. 



(3) It is admitted that one of the characteristics of ^ Na- 

 ture's workshop ' is the number of automatic arrangements. 

 In making a machine an artificer literally selects; in estab- 

 lishing a breed of animals Man literally selects ; but Natural 

 Selection is a metaphorical term, — the sifting is very largely 

 automatic. The survivors survive automatically in virtue of 

 the possession of certain advantageous qualities; the elim- 

 inated disappear automatically because of the absence of 

 certain advantageous qualities or the presence of others that 

 are fatal. But this is not the whole truth. 



The selection that occurs is not haphazard; it bears some 

 relation to the previously established external systematisa- 

 tion which we call the web of life, just as social criticism 

 which makes it difficult for the unreliable to get on is not 

 haphazard, but bears some relation to previously established 

 traditions and standards. The elimination in either case 

 is remote from fortuity or capriciousness. It always has, 

 of course, an immediate reference to the present and not to 

 the future; but the present has been determined by a past 

 selection of the fit and embodies that selection in an ob- 

 jective sieve of great subtlety. Since the sieve is a sys- 

 tematisation of fitness, it tends to sift towards fitness in 

 the future as well as in the present — unless, indeed, the 

 conditions of the future should greatly change. 



Let us repeat this argument. There is in each case a 

 line of evolution that pays ; it has been reached by past varia- 

 tions ; new variations that are congruent with the past are 

 on the whole most likely to appear and to catch on; there 



