CNIDARIA AS THE ONLY COELENTERETA 69 



reptilian characteristics) which finally led by way of all tran- 

 sitions to the new type, the birds; but instead there was an 

 accession, or an admixture, of new bird characteristics to the 

 already existing reptilian characteristics, as this can be seen in 

 the remains oiArchaeopteryx. De Beer beUeves that in the latter 

 the "reptilian-characters" are still prevalent. In such a case the 

 old characteristics can be preserved, yet they are naturally out- 

 weighed by the new, progressive, and speciaHzed character- 

 istics. It has been possible to show that other classes of Ver- 

 tebrata, too, had probably developed along a similar divergent 

 and mosaic-like way. 



It has been possible to find this type of evolution in con- 

 nection with Vertebrata because their fossil remains which 

 have been found go back to a more or less recent past when 

 this process of transformation had been taking place. The 

 fortunate case of the Archaeopteryx has been particularly 

 welcome. Yet unfortunately no such intermediate forms 

 belonging to invertebrate types have been found, nor is there 

 a prospect that they will ever be found; this is particularly true 

 for the lowest, and therefore phylogenetically oldest, inverte- 

 brates whose manifold and numerous types had been deve- 

 loped even before the beginning of the era of the formation of 

 the fossil-bearing earth strata. In this naturally disadvantageous 

 situation the researchers in invertebrates have to use methods 

 that are less reliable. Above all, it is a priori quite probable that 

 Watson's rule has a general validity as we cannot see the reason 

 why it should be limited to Vertebrata only. Secondly, if 

 Watson's rule can be considered valid at all, then it must con- 

 sequently be expected that even if the mosaic-like appearance 

 of characteristics be most manifest during the phase of transi- 

 tion—because here the characteristics of ascendants and of 

 their descendants appear most concentrated in individual spe- 

 cies, in this case in Archaeopterjx—^^^ have nevertheless to 

 expect the same characteristics and properties of the two sub- 

 sequent types to appear, even if somewhat "thinned down," 

 in the parent as well as in the descendant species. It is 

 6 



