of the Fossil Plants of the genus Sigillaria. 279 



Considering the great number of leaf-scars observable on the stems of 

 Sigillaria, it cannot but excite our surprise that, as yet, we have no posi- 

 tive evidence as to the kind of foliage peculiar to this genus. When 

 Brongniart, from external characters alone, maintained that Sigillaria 

 was the ancient representative of existing arborescent Ferns, there was 

 no difficulty in supposing that some of the numerous Fern leaves with 

 which the coal-formation abounds, constituted its foliage ; but since he 

 has discovered that the internal structure of this plant appoximates it to 

 the Cycadeous Gymnosperms, we have in this discovery an insuperable 

 difficult}'' in the way of such a supposition, at least, if we are to consider 

 it as utterly impossible that there could exist, during any previous 

 organic period, a tree having leaves resembling those of a Fern, and a 

 stem combining the structure of the latter with that of a Cycas. 



The following observations, I trust, will show how far we are sup- 

 ported in the view that a tree of this kind vegetated under the form of 

 Sipillaria, 



In following up the inquiry now entered upon, it will be necessary, in 

 the first instance, to take a brief survey of the various leaves which have 

 been found in deposits of the carboniferous epoch, in order to ascertain 

 which kind possesses the strongest claims to be considered as having be- 

 longed to Sigillaria, 



Passing over such remains as Lepidopht/llum, Flabellaria, Annularia^ 

 Sphenophyllum, Nceggcrathia, and a few others, from appearing to have but 

 a very remote connection with Sigillariaj or which are well known to 

 belong to other plants, we will at once pass on to the more likely forms, 

 as Sphenopteris, Pecopteris, Neuropteris, Odontopteris, ^chizopteri?, and 

 Cyclopteris. Even this list may be considerably reduced, perhaps, to the 

 first three genera, on account of the remainder, from their scarcity, not 

 appearing to have belonged to stems possessing such irrefragable evi- 

 dences, as to the profusion of their leaves, as the fossil under considera- 

 tion. Taking this for granted, our brief survey will consequently be 

 confined to the genera Sphmopteri?^ Pecopteris, and Neuropteris. 



Of the first genus, it is said that about twenty-eight species occur in 

 the coal-formation. It is generally considered to be a true Fern, al- 

 though it must be confessed there are some points connected with two 

 or more species, rather in favour of a contrary opinion. As stated in 

 the *' Vegetaux Fossiles," the recent forms which resemble Sphenopteris 

 are Gi/mnogramma, Asjjlcnium, Darea, Cheilanthes, Adiantum, Lindsea^ 

 Woodsia, Dicksonia, Davallia, Trtchomanes, Anemia, Hi/menophylhim, and 

 Botrychium. 



If it is the case that some species of Sphenopteris (e. g. 'J^. crithmifoUa) 

 are not Ferns, — are we justified in supposing that they have belonged to 

 Sigillaria ? I think not ; for, instead of standing in this relation, these 

 doubtful remains appear to have belonged to an herbaceous plant. This 

 suspicion arises from having observed some specimens appearing as if 

 roots were attached to them. I cannot speak positively on this point ; 



