CHAP. IX.] NATUBAL SELECTION. 287 







of Species (p. 104) he says, &quot;Until reading an article in 

 the North British Eeview (18G7), I did not appre- and admlt . 

 ciate how rarely single variations, whether slight or ted errors- 

 strongly marked, could be perpetuated.&quot; Again: he was 

 formerly &quot;inclined to lay much stress on the principle of 

 protection, as accounting for the less bright colours of female 

 birds &quot; ( Descent of Man, vol. i. p. 198) ; but now he speaks 

 as if the correctness of his old conception of such colours 

 being due to protection was unlikely. &quot; Is it probable,&quot; he 

 asks, &quot; that the head of the female chaffinch, the crimson on 

 the breast of the female bullfinch, the green of the female 

 chaffinch, the crest of the female golden-crested wren, have 

 all been rendered less bright by the slow process of selection 

 for the sake of protection ? I cannot think so &quot; [the italics 

 are mine]. (Vol. ii. p. 176.) 



Once more Mr. Darwin shows us (vol. i. p. 125) how he has 

 been over-hasty in attributing the development of certain 

 structures to reversion. He remarks, &quot; In my * Variations of 

 Animals under Domestication (vol. ii. p. 57) I attributed 

 the not very rare cases of supernumerary mammas in women 

 to reversion.&quot; &quot; But Professor Preyer states that mammse 

 erraticee have been known to occur on the back ; so that the 

 force of my argument is greatly weakened or perhaps quite 

 destroyed.&quot; 



Finally, we have a postscript in the second volume of the 

 Descent of Man which contains an avowal more remark 

 able than even what has been cited. He therein declares : 



&quot; I have fallen into serious and unfortunate error, in relation to the 

 sexual differences of animals, in attempting to explain what seemed to 

 me a singular coincidence in the late period of life at which the neces 

 sary variations have arisen in many cases, and the late period at which 

 sexual selection acts. The explanation given is wholly erroneous, as I 

 have discovered by working out an illustration in figures.&quot; 



It would be idle to dissemble, and disingenuous not to 

 declare, the amount of distrust with which such consequent 



. need of care- 



repeated over-hasty conclusions and erroneous cal- fui criticism. 



culations should properly inspire his readers. When then 



