360 DR D. H. SCOTT ON THE 



closely with that of other Dadoxylons in which primary wood has not yet been 

 distinguished. 



The pentagonal pith, double leaf -traces, small and few primary xylem-strands, and 

 dense wood, with almost wholly uniseriate medullary rays, may serve to characterise the 

 species. 



The importance of Dadoxylon Spenceri lies in its being, on the one hand, a typical 

 Dadoxylon, with the type of secondary wood which we know to have belonged to 

 Cordaites, while, on the other hand, it shows, in a reduced form, primary xylem 

 comparable to that of Lyginodendron or Poroxylon. It suggests, perhaps more 

 strongly than any of the other species described, a truly Gymnospermous stem, which 

 may well have belonged to one of the Cordaiteae, but which still retains the last relics 

 of the primary wood-structure characteristic of the Poroxyleae and Lyginodendreae. 



Summary and Conclusions. 



The principal result of the present investigation has been to show that in a number 

 of stems of Palaeozoic age (most of them from the Lower, but one from the Upper, 

 Carboniferous) with secondary wood of the well-known Dadoxylon structure, distinct, 

 usually mesarch, strands of primary xylem, forming the downward continuation of 

 the leaf-traces, were present around the pith. Thus the anatomical structure, of 

 which we may take Lyginodendron oldhamium as the type, proves to have been 

 widely distributed among Palaeozoic plants, and to have extended to stems which, 

 on the basis of other characters, would have been referred with some probability to the 

 Cordaiteae. 



The stems examined appear to range themselves naturally in three groups, as 

 indicated by the generic names employed. 



The Calamopitys group is characterised by the relatively large dimensions and 

 distinct mesarch structure of those primary xylem-strands which are about to pass out 

 from the pith, while the same strands, lower down in their course, are reduced in size, 

 and in some cases assume endarch structure, owing to the dying out of the centripetal 

 wood. A single strand passed out from the pith to form the leaf-trace. The pith is 

 solid, with no trace of discoid structure ; it is very variable in size ; in C. fascicularis 

 and in some species of C. Saturni, it is remarkably small (1-3 mm. in diameter). 

 The secondary wood has the typical Dadoxylon structure ; the medullary rays are in 

 many cases one, or at most two cells in thickness (C. fascicularis, C. beinertiana ; some 

 specimens of C. annularis) ; in C. Saturni pluriseriate rays appear to prevail. 



The characters of the cortex and leaf-bases are known in C. Saturni and C. 

 annularis, but not as yet in the species occurring in Britain. The repeated sub- 

 division of the leaf- trace in passing through the cortex is one of the most important 

 characters exhibited by the Thuringian specimens. The reference of C. fascicularis and 



