[SIR J. W.DAWSON] ON THE GENUS LEPIDOPHLOIOS 71 



but cannot be a new Sigillaria. This wjis named L. j^arvus. In point 

 of fact the Carboniferous forests contained many species of trees belong- 

 ing to an advanced type of acrogenous structure, and so nearly allied 

 that it is difficult, if not impossible, to distinguish them when in the state 

 of imperfectly preserved fragments. The difficulties of the palseobotanist 

 are increased by the presence of numerous species of Sigillaria, which, 

 while for the most part distinguishable from all the Lepidodendreœ, yet 

 in some of their genera approach them very nearly both in external 

 markings of the stem and its internal structure. On the other hand, there 

 are plants included among the Sigillariae which there is at least reason 

 to suspect belong to a higher type, akin in structure to the modern 

 cycads. In a paper on the affinities of these plants, published in the 

 Journal of the Geological Society in 1871, I suggested the following 

 scheme of their affinities, placing the Sigillariae as a group on the con- 

 fines of the Acrogens and Gymnosperms. After the lapse of twenty-six 

 years, and in view of the progress of discovery in the meantime, this 

 scheme must require some modification ; but we may well make its 

 amendment a basis for discussing the present aspects of the question. 



Cycadaceœ. Coniferœ. 



Dadoxylon. 

 Favularia ? Palteoxylon. 



Ormoxylon. 

 Dictyoxylon. 



Sigillaria. 



Ehytidolepis. Calamodendron. 



Favularia ? Calamopitus. 



Clathraria. Eornia. 



Syringodendron. Calamités. 



Lepidophloîos. Equisetaceœ. 

 Lei^idodendron. 

 Lycopodiaceœ. 



In this the SigillaritT? are regarded as a central generalized group, 

 from which, in regard to structure and affinities, various genera radiate 

 towards Cycads and Conifers on the one hand, and Lycopods and Equiset- 

 ums on the other. 



The Sigillarian structure is based on that of a remarkable axis show- 

 ing structure which I had at that time found in an erect trunk at the 

 South Joggins, and of which the details are figured in the paper above 

 referred to. Though I have since found a still better preserved axis of 

 difierent type, to be referred to in the sequel, I still hold that my 

 original specimen represents one, and that the more advanced, Sigill- 



