5i2 STUDIES IN FOSSIL BOTANY 



be regarded as provisional, until the investigations 

 now in progress are completed. The main features 

 of the structure are the same as in the species already 

 described ; the primary strands of centripetal wood, 

 the double bundles passing out from the pith, the 

 cortical structure and the axillary branching are 

 obvious points of similarity. The subdivision of the 

 leaf-trace pairs, first into four and then into six 

 strands, as they pass through the cortex, is particularly 

 well shown, as are also the small axillary buds (Fig. 

 184). The leaves were evidently much more crowded 

 on the stem than in the French type ; the secondary 

 wood is denser, with smaller tracheides and narrower 

 rays ; the pith is discoid, like that of Cordaites ; the 

 phloem also has some peculiarities ; it is thus probable 

 that the distinctions existing may prove to be of generic 

 value. In any case, the discovery of the Poroxylon type 

 of stem at an horizon considerably lower than that of 

 the species originally described, is on,e of much interest. 

 Until recently we were quite in the dark as to the 

 fructification of Poroxylon. On grounds of association, 

 however, M. Grand'Eury 1 now attributes the seeds 

 known as Rkabdocarpus , Brongniart, to these plants, 

 with which he also finds " floral axes, bearing large 

 and long male and female buds, without bracts." 

 Rkabdocarpus, as limited by Brongniart, is a genus of 

 seeds with bilateral symmetry, characterised by the 

 presence of numerous fibrous strands in the sarcotesta. 

 The seeds are of the same general type as those of 

 the Cordaiteae, described below (see p. 5 46). Fig. 185 

 shows a seed of this genus in transverse section, 



1 See his paper above cited, " Sur les Rhabdocarpus" etc. 



