FROM THE COAL-FORMATION OF NOVA SCOTIA. 841 



and still more fully in that on Sigillaria and its allies in 1870*, 

 several distinct types of Sigillarioid trees ; though whether we can, as 

 suggested in those papers, separate those with the Glaihraria and 

 Favidaria styles of markings from the other Sigillariw, is still doubt- 

 ful. The French authors above cited regard their S. elegans and 

 S. spimdosa, which are of the Favularia type, as true Sigillaria?, and 

 hold that their wood}' cylinder, with its fibres in radial series and 

 with medullary rays and radiating bundles proceeding from the inner 

 cylinder, allies these trees with the gymnospermous exogens. Wil- 

 liamson regards his Sigillaria? of the Diploxylon type of structure as 

 probably cryptogamous and allied to Lepidodendron, though main- 

 taining that the structure of these stems is truly exogenous. There 

 can scarcely be any doubt that the higher type of Sigillaria, which 

 I described in 1870, and which, I think, represents the ordinary 

 coarsely-ribbed species of the type of my S. Brownii, are allied to 

 gymnosperms. Prof. Newberry and the writer have adduced strong 

 circumstantial evidence to show that Sigillaria? produced the fruits 

 known as Trigonocarpa, found so constantly with their remains. 

 Goldenberg, on the other hand, has figured a sort of strobile as at- 

 tached to Sigillaria. Williamson has figured fruit-scars, which he 

 regards as attachments of cones. I have figuredf well-preserved 

 fruit-scars of two species which cannot have borne strobiles, but may 

 very probably have borne Trigonocarpa or racemes of such fruits. 

 These facts, I think, taken along with those of structure, tend to show 

 that there may be included in the genus Sigillaria, as originally 

 founded on the markings of the surface, species widely differing in 

 organization, and of both gymospermous and acrogenous rank. This 

 conclusion is further confirmed by the fact, which I have long ago 

 amply demonstrated in my papers on the structures and mode of 

 accumulation of coal, that in the great coal-beds tissues of gymno- 

 spermous character, but distinct from those of Conifers, exist to an 

 enormous amount, while no other trees are found in connexion with 

 these beds to which such tissues can be referred except the Sigillaria?. 



Should this view be finally established, these trees will present an 

 interesting link of connexion between the gymnosperms and the 

 higher cryptogams. They connect the Lepidodendra with the Cycads 

 and Conifers in the gradations of exogenous structure seen in their 

 wood and bark, and also in the remarkable transitions which they 

 exhibit between woody tissues of the discigerous type and those sca- 

 lariform tissues which, though resembling scalariform vessels pro- 

 perly so called, yet in these plants are evidently arranged in the 

 manner of woody fibres, and take the place of these in the construc- 

 tion of the stem. 



The tendency of investigation of late has been to convey the im- 

 pression that the Sigillarioid and Lepidodendroid trees of the coal- 

 formation were of one somewhat uniform and monotonous type. On 

 the other hand, the great number of species of these trees indicated 



* Quart. Journ. Geol. Soc. vol. xxvii. (1871) p. 147. 



t Quart. Journ. Geol. Soc. toI. xxii. Beport on Fossil Plants of the Lower 

 Carboniferous: 1873. 



