WorsdelL — The Structure and Origin of the Cycadaceae. 147 
pletely lost their inner inverted secondary tissues ; this is, possibly, also the 
case with M. Solmsii , var. lignosa . That the primitive type possessed two 
or more solenosteles of secondary origin we do not know. It is possible 
that these outer secondary zones may have been collateral in structure 
from the beginning, but this, in the light of the structures presently to be 
described in some Cycads, is not probable ; yet even if it were so, these 
zones would still have to be regarded as morphologically equivalent to the 
outer or dorsal half of a solenostele, and, therefore, these £ extrafascicular * 
collateral cylinders are morphologically derived from solenosteles , owing to 
the fact that the primary ring possesses the structure of this latter ; and all 
the rings are to be regarded as homologous. 
Let us consider now the case of the Cycadaceae. In that most 
primitive part of the axis of the plant (see supra), viz. the cotyledonary node, 
Gregg (in Cycas Seemanni ), I myself (in C. 
revoluta , Encephalartos horridus, Macrozamia 
Denisonii, Bowenia spectabilis ), and Matte (in 
Cycas siamensis ), observed and described a 
polystelic structure controlling either the whole 
or part of the first-formed secondary cylinder, 
i. e. the first ‘ extrafascicular ’ ring. Imme- 
diately outside the very sinuous contour of the 
root-stele Matte figures a number of concentric 
structures or steles, some of which are in the 
act of separating off from the central stele 
(Fig. 12) ; this will, perhaps, account for the 
imperfect continuity of contour of some of the 
steles. As in the cases of Cycas revoluta and Bowenia observed by me, 
some of the steles have a thicker layer of secondary wood on their inner 
than on their outer side. This ancestral stelar structure, wherever it 
occurs, becomes replaced, higher tip in the axis, where modern tendencies 
hold supreme sway, by a continuous , collaterally-constructed cylinder ; if 
traced downwards into the root these steles merge into the central cylinder. 
In the mature plant all trace of these structures in the cotyledonary node 
would probably be lost. 
In Encephalartos horridus I observed £ on the dorsal side of, and closely 
abutting on the large [concentric] strand [of the first “ extrafascicular ” 
cylinder] . . . two other smaller ones of the ordinary collateral structure 
and orientation. At another level two strands are seen, one on each side of 
the large one, each with a considerable amount of secondary thickening, 
one of which has inverted orientation 1 5 Further out in the cortex are 
scattered other much smaller bundles with inverted orientation. Another 
group of three bundles was observed, with their xylems mutually directed 
towards each other? These smaller strands I regard as the first sporadic 
Fig. 12. Cycas siamensis : 
transverse section of vascular sys- 
tem of seedling-axis in region of 
cotyledonary node, showing outer 
ring of steles (diagrammatic and 
reversed) (after Matte). 
