iioo Thompson . — The Anatomy and 
With regard to the more recent Cycads they possess one feature of 
resemblance lacking in the Bennettitales, namely, the double leaf-trace whose 
two strands were inserted separately in the stele. But as the double leaf- 
trace is common to so many groups of Gymnosperms it might be used with 
equal force to connect any of them with the Gnetales. As for the fact that 
the two strands are inserted separately in the stele, we have seen that this 
is a variable feature, even the ancient Gymnosperms showing both a condi- 
tion of separation at the stele ( Poroxylon ), and one in which the strands 
unite before entering ( Lyginodendron ). Miss Sykes, in her study of 
Welwitschia , brought forward other points of resemblance to Cycads, 
namely, the occurrence of centripetal wood in the peduncle and of inversely 
oriented and concentric bundles in the reproductive axis. A preliminary 
study of the reproductive organs of Ephedra has failed to reveal any of 
these features. On the whole it may be said that the differences between 
Ephedra and the modern Cycads are almost as strongly marked as between 
Ephedra and the ancient Bennettitales. Moreover, if the Gnetales have been 
derived from Cycadalean stock, most botanists would agree from the 
evidence of floral organization that it must have been from the Bennettitales, 
so that any advantages of the modern over the ancient members of that 
alliance should not have weight. 
With the growth in favour of the idea of Cycadalean affinity the older 
view of the Coniferous relationship of the Gnetales is being supplanted. 
The present study has led the writer to conclude that the older view had 
much more in its favour than the more modern one. Every one of the points 
enumerated as opposed to the Cycadalean relationship may be used in an 
argument for the Coniferous relationship. To emphasize only the strongest 
points one should mention the arrangement and structure of pits of the 
tracheides, bars of Sanio, tertiary spirals, trabeculae, primitive uniseriate and 
lignified rays, lack of centripetal and development of centrifugal wood in 
the leaf-trace, and the structure of the vascular bundles of the leaves. These 
points of remarkable resemblance do not prove a descent from any modern 
group of Conifers. Indeed, such a descent seems difficult to establish. 
Opposed to it is the very generalized character of the secondary wood com- 
bining many characters found in the various groups of Conifers, such as 
Araucarian and Abietinean pitting, bars of Sanio, tertiary spirals, trabeculae, 
resin plates, and wood parenchyma. These points would rather indicate an 
affinity with that Abietinean-Araucarinean stock which it is now generally 
conceded gave rise to the modern Conifers. Of course it is possible that 
the Gnetales represent a line of development paralleling to a considerable 
extent that of the Conifers. But, however opinions may differ as to 
the exact point of origin of the Gnetalean line, there is a large mass 
of evidence to indicate an affinity with the base of the Coniferous line of 
descent. 
