PRIMARY STRUCTURE OF CERTAIN PALAEOZOIC STEMS. 



337 



of the outgoing leaf-traces. Sometimes, as shown in diagram 3, as many as three 

 successive traces, passing out through the wood, are seen in the same transverse section. 

 In the lower part of its course, as we have seen, the leaf- trace bundle passes very 

 gradually outwards, diverging but little from the vertical direction. When it has once 

 fairly entered the wood, however, it curves more rapidly (owing to the growth in thickness 

 of this zone), so as to assume a more nearly horizontal course, and is consequently cut 

 in an approximately transverse plane when intersected by a tangential section of the 

 wood ; such a section, from the Williamson specimen, is represented in PI. IV. fig. 7. 

 From the course of these strands, as described above, there can be no doubt that 



Diagram 4 (K. 540 c ). 



they represent the leaf-traces, or rather their xylem-constituent, passing out to leaves 

 with a 2/5 phyllotaxis. On a superficial view, it might perhaps be supposed that the 

 larger strands belonged to branches, but more careful observation shows every grade of 

 transition between the larger and smaller strands, and proves their identical nature 

 (see diagrams 1-4, and figs. 1 and 2). It is evident that the bundle, as it approached 

 its point of exit from the pith, increased rapidly in size, attaining its full dimensions 

 where it began to pass outwards. A similar increase in size, though perhaps less 

 striking, occurs in the outgoing strands of Lyginodendron * and Poroxylon.f 



The structure of the primary xylem-strands is most obvious where they attain their 



* Williamson and Scott, loc. tit., PL 21, fig. 1. 



t Bertrand et Eenault, " Recherches sur les Poroxylons," Arch,. Bot. du Nord 

 tigs. 198, 199, etc. 



la France, 3 me Annee, 1886, 



