28 
A. S. Marsh. 
to be the same as that described above. There were also present 
in one series of sections the bundles of a formerly produced cone. 
These were given off from points along one half of the vascular ring, 
and gathered together to form a fan-like mass which passed through 
a gap in the opposite half of the bundle-ring. This is exactly 
similar to that described for Stangeria by Solms Laubach (8). 
Conclusion. 
The principal theoretical conclusion arrived at from this work 
is that from the structure of the petiolar bundles of Stangeria, and 
presumably of other Cycads, a close relationship can be argued 
between the modern Cycadales and the fossil Cycadofilicales. 
Fig. 10.—Longitudinal section of tracheids of petiole. I, centrifugal ; 
II, centripetal. 
The “ mesarch ” foliar bundles of one type are to he derived 
from a truly mesarch vascular strand with secondary centrifugal 
wood. Such a type is found in the stems of such forms as Lygino- 
pteris and from this we evolve the modern structure by dropping 
out the lignification of the primary centrifugal xylem, leaving two 
distinct systems separated in space—the centripetal and the secondary 
centrifugal. The evidence for this is the successional series at the 
leaf-base (where, as Sinnott (7) has pointed out in his work on fern 
leaf-traces, primitive structures are known to persist) leading at a 
slightly higher level not only to normal forms, but also to bundles 
with no centrifugal lignified elements and to others with a persistent 
chain of elements joining centrifugal and centrifugal. This probably 
recapitulates a stage in evolution when both primary and secondary 
centrifugal wood were greatly but equally reduced, but at a still 
higher level in the petiole we have the secondary centrifugal re¬ 
asserting itself and increasing in amount. As long as even a midrib 
