36 A. G. Tansley. 
the Cycads the latter is represented merely by a number of proto- 
xylems from which the cambium starts its activity. 
The lower members of the series can be paralleled among the 
primitive Ferns. We have already seen how a solid mesarch stele 
like that of Gleichenia might easily be derived from a simple endarch 
stele like that of Botryopteris, by enlargement and independent 
decurrency of protoxylems from the leaf-traces. The primary wood 
of Heterangium is almost exactly like that of Gleichenia, except that 
the peripheral groups of xylem corresponding with the leaf-traces 
tend to be more individualised and the leaf-traces themselves pass 
off as isodiametric instead of as arched strands. In Lyginodendron 
the primary wood is confined to the leaf-traces and compensation- 
strands, and the same is true of Osmunda. But while in the former 
plant a considerable amount of secondary xylem was formed, the Fern 
never acquired this habit, and consequently, like all plants which 
have no secondary thickening is entirely dependent on the putting 
out of adventitious roots from the stem. This of course is the real 
cause of the strict limitations of the Fern type of structure. 
Directly continuous secondary thickening has been initiated in both 
the ascending and descending axis, the erect stem is able to grow 
indefinitely in height, to branch and increase its transpiring surface, 
because it is able pari passu to increase its absorptive surface and 
its conducting channels. The subsequent history of the evolution 
of the vascular system of the stem is mainly a history of the 
perfecting of the mechanism of secondary thickening and of the 
increasing differentiation of the secondary tissue. 
The petiolar strand, at first fern-like ( Heterangium , Lygino¬ 
dendron) separates into bundles in the higher types, but all along it 
lags behind the cauline system in progress and is still mesarch in 
the Cycads. This is in striking contrast to the state of things we 
have found everywhere in the Ferns, in which the leaf-trace always 
leads, and the cauline cylinder follows. The reason of this contrast 
appears to be that in the Cycadofilicinean series the dominant 
influence is adaptation to secondary thickening of the axis, the 
condition of the production of an unlimited number of leaves, rather 
than elaboration of the individual leaf, and though the Cycads retain 
the fern-like habit, and the typical fern-like leaf-trace curve, leaf- 
reduction is found in Poroxylon, Cordailes, and the Angiosperms. 
Those Cycadofilices which stand apart from the main series 
may all be regarded as more or less futile experiments in trying to 
graft secondary thickening on to the more complicated forms of 
