438 Darwin, and after Darwin. 



popular (and very erroneous) view of organic evolution as 

 a process which is always and everywhere bound to promote 

 the specialization of organic types only then ought we to 

 see any real difficulty in the absence of generalized types 

 preceding these existing types. Of course we may wonder 

 why still lower down in the geological series we do not 

 meet with more generalized (or ancestral) types; but this 

 is the difficulty number 3, which we now proceed to 

 examine. 



Concerning the other two difficulties, then, the only possible 

 way of meeting that as to the absence of any parent forms 

 lower down in the geological series is by falling back as in 

 the analogous case of animals upon the imperfection of the 

 geological record. Although it is certainly remarkable that 

 we should not encounter any forms serving to connect the 

 Dicotyledonous plants of the Chalk with the lower forms of 

 the underlying Wealden, we must again remember that diffi- 

 culties thus depending on the absence of any corroborative 

 record, are by no means equivalent to what would have 

 arisen in the presence of an adverse record such, for in- 

 stance, as would have been exhibited had the floras of the 

 Wealden and the Chalk been inverted. But, as the case 

 actually stands, the mere fact that Dicotyledonous plants, 

 where they first occur, are found to have been already differ- 

 entiated into their three main divisions, is in itself sufficient 

 evidence, on the general theory of evolution, that there must 

 be a break in the record as hitherto known between the 

 Wealden and the Chalk. Nor is it easy to see how the op- 

 ponents of this theory can prove their negative by furnishing 

 evidence' to the contrary. And although such might justly be 

 deemed an unfair way of putting the matter, were this the only 

 case where the geological record is in evidence, it is not so 

 when we remember that there are numberless other cases 

 where the geological record does testify to connecting links in 

 a most satisfactory manner. For in view of this consideration 



