372 A CENTURY OF PROGRESS IN THE NATURAL SCIENCES 



equivalent order, viz., the Cycadophyta, Coniferophyta, and the Chlamydosper- 

 mophyta, while Eothmaler has Gymnospermophytina (pteridosperms and cy- 

 cads), Stachyospermophytina (including the ephedras), and Chlamydospermo- 

 phytina (including the bennettites). Johansen agrees in principle with Arnold, 

 but replaces the Gymnospermae by an even greater number of phyla or divisions 

 — the Pteridospermophyta, Cycadophyta, Ginkgophyta, Coniferophyta, and the 

 Ephedrophyta (chlamydosperms). The phylum Ginkgophyta contains the orders 

 Cordaitales and Ginkgoales, which in Arnold's system were included in the Coni- 

 ferophyta. Emberger deals with the gymnosperms in another way. He stresses 

 the different kinds of "seeds," in the first place differentiating between Prae- 

 phanerogamae and Phanerogamae. The former are regarded as a major group, 

 intermediate between the vascular cryptogams and the phanerogams. His clas- 

 sification appears to be intentionally horizontal, and consequently nonphylo- 

 genetic. The pteridosperms in the wide sense, and the cycads, cordaites, and 

 ginkgoes, are said to be praephanerogams, while the conifers, taxads, bennettites 

 and chlamydosperms — being true gymnosperms — are referred to the phanero- 

 gams. In this way the cordaites and ginkgoes have been far removed from the 

 conifers, with which they are otherwise usually considered to be rather closely 

 related. And whatever the natural position of the megaphyllous bennettites 

 might be, this group has little to do with the microphyllous conifers. Whenever 

 a subdivision named Cycadophyta (Cycadopsida, Cycadomorphae, Phyllosper- 

 minae) of the gymnosperms (or of the Pteropsida in the wide sense, the phanero- 

 gams, and the mesoeormophji:es, respectively) has been proposed in recent years, 

 both the Cycadinae (Cycadales) and the Bennettitinae (Bennettitales, Cycadeoi- 

 dales) have, with one exception (Eothmaler), been referred to this. Whether 

 the relationship between these two groups is really as close as thus often assumed 

 is a question, which can hardly be settled without a satisfactory knowledge of 

 the nature and early evolution of the bennettitinean gynoecium. Disregarding 

 the chlamydosperms (gnetophytes or ephedrophytes), Gaussen's system has two 

 subdivisions of the Gymnospermae, viz., the above-mentioned Cycadophyta and 

 Coniferophyta. The latter corresponds in the main to the Stachyospermae of 

 Sahni. Besides the cycads and the bennettites, the Cycadophyta sensu Gaussen 

 include the pteridosperms in the wide sense, and Salmi's Pentoxylaceae. A simi- 

 lar arrangement is adopted in the systems of Lam, Arnold, Chadefaud, Nemejc 

 and Eothmaler, although Chadefaud's Cycadomorphae and Nemejc's Cycado- 

 phyta also include the chlamydosperms, and Eothmaler refers the bennettites 

 to the latter group. Takhtadjan has divided the gymnosperms into four sub- 

 classes, viz., the Pteridosperminae, the Phyllosperminae, the Stachyosperminae, 

 and the Chlamydosperminae. Finally, in the system of Pulle the gymnosperm 

 classes (or orders) have not been brought together into subdivisions of higher 

 rank, and the pteridosperms are kept separate from them. 



The consensus on the phylogenetic classification of the gjnnnosperms is thus 

 by no means complete. Eeasons for this are not difficult to find. The paleo- 

 botanical evidence, after all, is still in many cases inadequate to justify a deci- 

 sion between one view and the other. Additional information is required, par- 

 ticularly on the paleozoic and mesozoic pteridosperms, the early bennettites 

 and the Pentoxylaceae, and on many other plants of mesozoic age. It is more- 

 over often difficult to distinguish primitive from advanced characters, and to 



