1896. LAMARCK AND LYELL. 91 



it seems to be, on the whole, rather fortunate than otherwise that 

 evidence of the inheritance of their effects is so hard to find. 



Now is there any ground for the beHef that the case is any 

 different with intellectual and moral improvement ? All known 

 mental phenomena have structural correlatives, and are capable of 

 development and improvement only so far as structural adjustments 

 for bringing about this development and improvement already exist. 

 Capacity for individual improvement under the normal conditions of 

 life is an adaptation ; the most wonderful and admirable of adapta- 

 tions ; and the beneficial influence of the " Lamarckian factors," so 

 far as this influence is beneficial, is not an explanation, but a fact 

 which itself calls for explanation. 



Is there any evidence that the influence of the environment is 

 inherently beneficial ? If there is no such evidence we must believe 

 that all its effects, except the effects which are already deducible from 

 adaptive structure, must be hap-hazard. 



When we remember how narrow the range of adjustment of each 

 organism is, it must be clear that the probability that hap-hazard 

 effects will be injurious or neutral rather than beneficial is prodigious. 

 Even if they are inherited, the effects of the " Lamarckian principles" 

 cannot cumulate in adaptation, except as an accident, which is so 

 improbable that we are justified in doubting whether it has given rise 

 to any specified adaptation until the possibility of a better explanation 

 has been rigorously excluded. 



While I find a difficulty in reconciling all Romanes' pubHshed 

 opinions on this subject with each other, he seems to hold, p. 153, 

 that " These Lamarckian principles are direct causes of determinate 

 variation in adaptive lines," although it is clear that this must be 

 proved before we can agree with him that " variation in these lines 

 being cumulative, the result is that natural selection is in large part 

 presented with the raw material of its manufacture — special material 

 of the particular kind required, as distinguished from promiscuous 

 material of all kinds." 



Some fifteen years ago I published a book in which I sought to 

 prove that we have in sexual reproduction a mechanism, produced by 

 selection, for the purpose of causing changes in parts which need 

 change, and that this mechanism expedites selection. My book 

 found the oblivion which it no doubt deserved, but I suspect that it 

 may have led, in some circuitous way, to the enrolment of my name 

 in the list of Neo-Lamarckians, although I explicitly stated that I did 

 not believe that these changes in parts which need change are 

 adaptive ; and I still believe, as I did then, that the " Lamarckian 

 principles" must be proved to be adaptive, antecedently to selection, 

 before we can seriously consider them as factors in adaptive 

 modification. 



There is, in Romanes' book, one sentence,— only one so far as I 

 have discovered, — in which he seems to admit that this is not only 



H 2 



