26 president's address, 



but with the remembrance of it, and will show itself in the familiar snarl 

 of the angry dog. This movement is now, I presume, hereditary in dogs, 

 and is so strongly inherited by ourselves (from simian ancestors) that 

 a lifting of the corner of the upper lip is a recognised signal of adverse 

 feeling. Is it really conceivable that the original snarl is due to that 

 unspecialised stimulus we call pain, whereas the inherited snarl is due to 

 fortuitous upsets of the determinants in the germ-cell 1 



I am well aware that many other objections may be advanced against 

 the views I advocate. To take a single instance, there are many cases 

 where we should expect somatic inheritance, but where we look in vain 

 for it. This difficulty, and others equally important, must for the present 

 be passed over. Nor shall I say anything more as to the possible means 

 of communication between the soma and the germ-cells. To me it seems 

 conceivable that some such telegraphy is possible. But I shall hardly 

 wonder if a majority of my hearers decide that the available evidence in 

 its favour is both weak and fantastic. Nor can I wonder that, apart 

 from the problem of mechanism, the existence of somatic inheritance is 

 denied for want of evidence. But I must once more insist that, accord- 

 ing to the mnemic hypothesis, somatic inheritance lies at the root of all 

 evolution. Life is a gigantic experiment which the opposing schools 

 interpret in opposite ways. I hope that in this dispute both sides will 

 seek out and welcome decisive results. My own conviction in favour of 

 somatic inheritance rests primarily on the automatic element in ontogeny. 

 It seems to me certain that in development we have an actual instance of 

 habit. If this is so, somatic inheritance must be a vera causa. Nor does 

 it seem impossible that memory should rule the plasmic link which connects 

 successive generations — the true miracle of the camel passing through the 

 eye of a needle— since, as I have tried to show, the reactions of living 

 things to their surroundings exhibit in the plainest way the universal 

 presence of a mnemic factor. 



We may fix our eyes on phylogeny and regard the living world as a 

 great chain of forms, each of which has learned something of which its 

 predecessors were ignorant ; or we may attend rather to ontogeny, where 

 the lessons learned become in part automatic. But we must remember 

 that the distinction between phylogeny and ontogeny is an artificial one, 

 and that routine and acquisition are blended in life.^ 



The great engine of natural selection is taunted nowadays, as it was 

 fifty years ago, with being merely a negative power. I venture to think 

 that the mnemic hypothesis of evolution makes the positive value of 

 natural selection more obvious. If evolution is a process of drilling 

 organisms into habits, the elimination of those that cannot learn is an 



' This subject is dealt with in a very interesting manner in Professor James 

 Ward's forthcoming lectures on the Realm of Eoids. Also in his article on 

 Mechanism and Morals in the Hihlert Journal, October 1905, p. 92 ; and in his 

 article on Psychology in the JEwyclopa>dia Britannica, 1886, vol. xx. p. 44. 



