WorsdelL — The S 'true hire and Origin of the Cycadaceae . 151 
large bundle of Lyginodendron i with its several protoxylem-groups situated 
in a mesarch position, really, in my opinion, corresponds to, and is the 
equivalent of, a number of collateral mesarch bundles united together by 
means of their centripetal xylems, their phloems likewise fusing and sur- 
rounding the whole. If we imagine each of these primary centres separated, 
each constituting a distinct bundle and at the same time retaining its 
position, they would be seen to roughly exhibit the horse-shoe-shaped 
grouping so characteristic of most Cycads (Fig. 15). This grouping most 
closely resembles that of the bundles in the petiole of Stangeria ; and 
this is interesting, for the foliage of Stangeria is the most Fern-like of all 
Cycadian leaves (Fig. 16). The extreme case of subdivision of such a Rachio- 
ptcris- strand is seen in the petioles of Medullosa (. Myeloxylon ) and certain 
species of Encephalartos (e. g. E. cycadifolins ) where an enormous number of 
separate collateral bundles exist, orientated in a variety of directions. But 
in E. horridus we note a 
case which, in my opinion, 
represents a distinct tran- 
sition to the monodesmic 
condition of Lyginoden- 
dron ; for here the vascular 
system represents a modi- 
fication of the X2 in which 
the bundles are fusing to- 
gether in groups in a radial 
direction, i. e. towards the 
centre of the organ, to form 
small, usually- imperfectly, 
concentric strands. We 
have only to imagine this phenomenon proceeding still further in order 
to obtain a petiolar vascular system consisting of one or two large 
concentric bundles as in Lyguiodendron. In the case of Stangeria we 
see lateral fusion between the bundles taking place, which, of course, would 
also be necessary in order to build up the primitive large concentric strand ; 
so that I am inclined to regard all these examples of fusion as reversions 
to an ancestral condition (Fig. 16). 
If the view here set forth be adopted the two sets of petiolar vascular 
systems represented by Cycads on the one hand and Lyginodendron 
and Heterangium on the other, become easily related and harmonized. 
Ever since I first observed them, some nine years ago now, I have always 
regarded the mesarch collateral bundles of the Cycadean petiole as a 
remnant of a former concentric strand possessing a central mass of 
centripetal xylem in the common, far-back ancestors of the Cycads 
and the Medulloseae. I regard the single polyarch bundle of Lygino - 
Fig. 16. Stangeria paradoxa : diagrammatic transverse 
section of vascular system of petiole, showing lateral fusion 
of bundles (after Matte), 
