58 



SCIENCE 



[Vol. LY, No. 1412 



record. In a spirit of paradox even this lias 

 been questioned. It has been asked how do 

 you know for instance that there were no 

 mammals in palaeozoic times? May there not 

 have been mammals somewhere on the earth 

 though no vestige of them has come down to 

 us? We may feel confident there were no 

 mammals then, but are we sure? In very an- 

 cient rocks most of the gi-eat orders of ani- 

 mals are represented. The absence of the 

 others might by no great stress of imagina- 

 tion be ascribed to accidental circumstances. 

 Happily however there is one example of 

 which we can be sui'e. There were no Angio- 

 sperms — that is to say "higher plants" with 

 protected seeds — in the carboniferous epoch. 

 ' Of that age we have abundant remains of a 

 world wide and rich flora. The Angiosperms 

 are cosmopolitan. By theii- means of dispersal 

 they must immediately have become so. Their 

 remains are very readily preserved. If they 

 had been in existence on the earth in carboni- 

 ferous times they must have been present with 

 the carboniferous plants, and must have been 

 preserved with them. Hence we may be sure 

 that they did appear on the earth since those 

 times. We are not certain, using certain in 

 the strict sense, that the Angiosperms are the 

 lineal descendants of the carboniferous plants, 

 but it is very much easier to believe that they 

 are than that they are not. 



Where is the difficulty? If the Angiosperms 

 came from the carboniferous flora why may we 

 not believe the old comfortable theory in the 

 old way ? WeU so we may if by belief we mean 

 faith, the substance, the foundation of 

 things hoped for, the evidence of things not 

 seen. In dim outline evolution is evident 

 enough. Prom the facts it is a conclusion 

 which inevitably follows. But that particular 

 and essential bit of the theory of evolution 

 which is concerned with the origin and natm'e 

 of species remains utterly mysterious. We no 

 longer feel as we used to do, that the process 

 of variation, now contemporaneously occur- 

 ring, is the beginning of a work which needs 

 merely the element of time for its completion; 

 for even time can not complete that which has 

 not vet begun. The conclusion in which we 



were brought up, that species are a product of 

 a summation of variations ignored the chief 

 attribute of species first pointed out by John 

 Eay that the product of their crosses is fre- 

 quently sterile in greater or less degree. Hux- 

 ley, very early in the debate pointed out this 

 grave defect in the evidence, but before breed- 

 ing researches had been made on a large scale 

 no one felt the objection to be serious. Ex- 

 tended work might be trusted to supply the 

 deficiency. It has not done so, and the signi- 

 ficance of the negative evidence can no longer 

 be denied. 



When Darwin discussed the problem of in- 

 ter-specific sterility in the "Origin of Species" 

 this aspect of the matter seems to have es- 

 caped him. He is at gi-eat pains to prove that 

 inter-specific crosses are not always sterile, 

 and he shows that crosses between forms which 

 pass for distinct species may produce hybrids 

 which range from complete fertility to com- 

 plete sterility. The fertile hybrids he claims 

 in support of his argument. If species arose 

 from a common origin, clearly they should 

 not always give sterile hybrids. So Darwin 

 is concerned to prove that such hybrids are 

 by no means always sterile, which to lis is a 

 commonplace of everyday experience. If 

 species have a common origin, where did 

 they pick up the ingredients which produce 

 this sexual incompatibility? Almost cer- 

 tainly it is a variation in which something 

 has been added. We have come to see that 

 variations can very commonly — I do not say 

 always — be distinguished as positive and nega- 

 tive. The validity of this distinction has been 

 doubted, especially by the Drosophila workers. 

 Nevertheless in application to a very large 

 range of characters, I am satisfied that the 

 distinction holds, and that in analysis it is a 

 useful aid. Now we have no difficulty in find- 

 ing evidence of variation by loss. Examples 

 abound, but variation by addition are rarities, 

 even if there are any which must be so ac- 

 counted. The variations to which inter- 

 specific sterility is due are obviously varia- 

 tions in which something is apparently added 

 to the stock of ingredients. It is one of the 

 common experiences of the breeder that when 



