PRIMARY STRUCTURE OF CERTAIN PALEOZOIC STEMS. 333 



of nearly 3 cm. in the former, and about 2 cm. in the latter, but these dimensions are 

 of no significance, as the wood is manifestly incomplete. In neither specimen is anything 

 beyond the wood preserved. 



In Mr Kidston's specimen the pith is complete, though somewhat contracted (see 

 fig. 1) ; consequently all the xylem-strands can be recognised, and their course traced 

 in successive transverse sections. In the Williamson stem the pith has perished, and 

 the smaller xylem-strands are obscure ; for details of the w T ood and larger primary 

 strands, however, this specimen is rather superior to the other. Fig. 1, from one of the 

 transverse sections of Mr Kidston's specimen, gives a good idea of the general structure. 

 The parenchymatous pith, which, owing to shrinkage, has partly separated from the sur- 

 rounding wood, has, in itself, a very uniform structure ; the peripheral cells are narrower 

 and have rather thinner walls than those towards the centre (compare the longitudinal 

 section in fig. 3) ; some of the larger elements are filled with dense carbonaceous contents, 

 which may indicate that they had a secretory function during life. Around the pith a 

 number of xylem-strands are disposed, forming an irregular ring. Eight of these strands 

 are wholly or partially embedded in the pith ; a ninth strand (B), much the largest of 

 all, is passing out into the zone of wood. It is a constant rule, holding good for all the 

 sections of both specimens, that the outgoing bundles are those which attain the maximum 

 dimensions (cf. phot. 2, from the Williamson specimen). A, the next largest strand, 

 is in close contact with the secondary wood, and will be the next to pass out above, as 

 is shown by the comparison of successive transverse sections. Most of the smaller 

 strands are actually embedded in the pith, and are separated from the inner edge of the 

 wood by several layers of parenchyma (cf. the longitudinal section, fig. 3). We have 

 here an approach to a condition which we shall find existing, in a much more marked 

 degree, in Pitys antiqua. 



There are in all eight transverse sections of Mr Kidston's specimen. They were cut 

 at different times, and I have no record of their order, but have been able to deter- 

 mine it with practical certainty by careful comparison of the peculiarities of the 

 individual sections as to detailed structure, position and form of cracks, exact state 

 of preservation, etc. The succession of the sections having been thus ascertained, 

 it became possible to determine the course of the xylem-strands. The camera lucida 

 diagrams in the text, 1 to 4, prepared for me by Mr L. A. Boodle, F.L.S., are taken 

 from a series of four consecutive sections, sufficient to fix the essential points in the 

 distribution of the strands. The series is from above downwards. The top section 

 of the four (diagram l) shows three principal bundles, markedly larger than the rest. 

 One of them, C, is still far out in the secondary wood ; another, B, of equal or even 

 greater size, has just reached, with its inner edge, the periphery of the pith ; the third, 

 A, which is much smaller, though still far exceeding the other circum-medullary strands 

 in size, has already entered the pith. These three leaf-traces may be taken as fixing 

 the position of the three orthostichies, A, B and C, on which the leaves supplied by 

 these traces were inserted. A, being cut lowest down in its course, belongs to the 



