PALEOBOTANY- — BERRY. 341 



Lower Cretaceous Cycadeoids and Williamsonias, we have only the 

 remains of foliage as a clue to relationships, and practically no in- 

 formation regarding anatomy or fructifications; the latter we infer 

 from the analogy of known forms were highly variable. 



Various attempts at a natural, or at least a logical, classification 

 of the Cycadophytes have been attempted. These range from re- 

 garding the phylum as consisting of two subordinate groups: the 

 Cycadales, and the Mesozoic Cycadeoideas and Williamsonias as a 

 second order — the Bennettitales, to those recognizing such illy 

 founded groups as the Xilssoniales, based upon cuticular characters. 

 Without wishing to depreciate the importance of any available fea- 

 tures in the study of fossil forms, all of which are valuable, or of 

 failing in recognition of the praiseworthy work of Nathorst, Thomas, 

 Bancroft, and others upon cuticles, it may be suggested that cuticles 

 can scarcely be regarded as affording ordinal criteria. In the case 

 of the proposal of such a group as the Nilssoniales, the variation in 

 detail of these features and their similarity to what obtains rather 

 uniformly in other groups of xerophytic gymnosperms, shows con- 

 clusively that they represent convergence due to habitat. A tendency, 

 especially among botanists, to regard the cycadeoids as the dominant 

 and progressive fossil type and to overestimate the degree of rela- 

 tionship between them and their contemporaries, the Williamson- 

 iales, is also to be deprecated. 



It must be recognized, entirely aside from the degree of converg- 

 ence or divergence of fructification morphology, that the cycado- 

 phyte stock as it was derived from the pteridosperms was a relatively 

 slender stemmed branching type, corresponding to the more primi- 

 tive Williamsonia type as exemplified by the Triassic Wielandiella. 

 Nothing is known of the anatomy or fructifications of many of the 

 so-called Williamsonia genera such as Pterophyllum, Plagiozamites, 

 Anomozamites, etc. Analogy would lead to the view that the floral 

 and stem morphology had not become stereotyped in this order and ex- 

 hibited considerable diversity, sufficient to include Podozamites on 

 the one hand, and Williamsonia on the other. It would seem that 

 the most natural arrangement of the phylum is one that regards the 

 main line of descent from which the more specialized or reduced 

 branches were given off at different times as a separate order, the 

 Williamsoniales, which was the long existent and cycadophyte 

 plexus, from which two other orders, the Cycadales and the Cyca- 

 deoidales. were derived. 



Quite naturally the Williamsoniales that are nearest the points of 

 origin of the Cycadeoidales, or Cycadales, will show great similari- 

 ties in their fructifications, but any arrangement that throws these 

 points of contact into one order and recognizes such quasi groups as 

 136(550°— 20 23 



