CHAP, xvn METAPHYSICS OF NATURAL SELECTION 199 



speaking so one assumes, what is habitually assumed, 

 and never proved, that variation is casual, i.e. non- 

 advantageous (in itself and on the average). If it 

 turned out that variation moved even in part along 

 the lines of evolutionary change, then Darwinism or 

 even Hyper-Darwinism might warrant the hope of 

 rapid progress. Hence it is extraordinarily difficult 

 to bring to the test of experiment the questions 

 between the Lamarckians and the Weismann school. 

 One glides into the habit of thinking that it is mainly 

 a question of pace. And yet quick pace, if it were 

 proved, might not be a presumption in favour of 

 Lamarckian use-inheritance. It might only point to 

 a neglected element in Darwinism, to the necessity of 

 regarding variation per se as telic not casual. 



We do not mean here to affirm that variation must 

 be advantageous, or even that it must proceed along 

 definite lines. We merely claim that such possibili- 

 ties should not be forgotten. The questions are 

 questions of fact, and further evidence is required. 

 Causeless variations are inconceivable things ; in that 

 view, presumably, all will agree. But, just as little 

 as the evolutionist would waste time over a hypothesis 

 which involved surrendering the causal law, so little 

 would others consent to trifle with a great question 

 by framing the hypothesis of variations perversely 

 opposed to the specific type. Still, within limits, we 

 might conceive of " casual " variations, if variability 

 worked along one of several fixed ' possible directions, 

 while the reasons why it chose one track rather than 

 another were highly obscure. 



Let us take an illustration. Every house of two or 

 more storeys must include a staircase. The stair may 



