1866.] DAWSON— THE COAL FLORA. 367 



coals and in the interior of erect trees, showing that they were 

 produced in the coal-swamps. The structures of these fruits 

 show that they are phgenogamous and probably gymnospermous. 

 Now the only plants known to us in the Coal formation, whose 

 structures entitle them to this rank, are the Conifers, Sigillariae, 

 and Calamodendra. All the others were in structure allied to 

 cryptogams, and the fructification of most of them is known. 

 But the Conifers were too infrequent in the Carboniferous swamps 

 to have afforded numerous species of Carpolites ; and, as I shall 

 presently show, the Calamodendra were very closely allied to 

 Sigillariae, if not members of that family. Unless, therefore, 

 these fruits belonged to Sigillaria. they must have been produced 

 by some other trees of the coal-swamps, which, though very 

 abundant and of numerous species, are as yet quite unknown to 

 us. Some of the Trigonocarpa have been claimed for Conifers, 

 and their resemblance to the fruits of Salisbury a gives counten- 

 ance to this claim ; but the Conifers of the Coal period are much 

 too few to afford more than a fraction of the species. One species 

 of Ehabdocarpus has been attributed by Geinitz to the genus 

 Nseggerathia ; but the leaves which he assigns to it are very like 

 those of Sigillaria elegans, and may belong to some allied species. 

 With regard to the mode of attachment of these fruits, I have 

 shown that one species, Trigonocarpum racemosum of the 

 Devonian strata,* was borne on a rhachis in the manner of a 

 loose spike, and I am convinced that some of the groups of inflor- 

 escence named Antholithes are simply young Rhabdocarpi or 

 Trigonocarpa borne in a pinnate manner on a broad rhachis and 

 subtended by a few scales. Such spikes may be regarded as 

 corresponding to a leaf with fruits borne on the edges, in the 

 manner of the female flower of Cycas ; and I believe with Golden- 

 berg that these were borne in verticils at intervals on the stem. 

 In this case it is possible that the ^strobiles described by that 

 author may be male organs of fructification containing, not spores, 

 but pollen. In conclusion, I would observe that I would not 

 doubt the possibility that some of the fruits known as Cardio- 

 carpa may have belonged to sigillarioid trees. I am aware that 

 some so-called Cardiocarpa are spore-cases of Lepidodendron ; 

 but there are others which are manifestly winged nutlets allied to 



* ' Flora of the Devonian Period,' Quart. Journ. Geol. Soc., vol. viii. 

 p. 324. 



