ANATOMY 647 
supported by the structure of the. individual plant when young. In both 
cases a prevalence of a cylindrical protostelic state, with comparatively 
slight disturbance of the axial system on departure of the leaf-traces, would 
be expected if the shoot were primitively strobiloid. Further, the leaf-trace 
would be relatively simple. 
Leaving aside for the moment the Marattiaceae, which are anatomi- 
cally a peculiarly specialised series in themselves, a comparison of the 
early types of Ferns points clearly to origin from a protostelic state with 
a leaf-trace consisting of a single strand, which comes off with but slight 
local disturbance from the periphery of the stele. This, with certain 
variants, is the typical condition in the Botryopterideae: a near approach 
to it is found also in the earliest Osmundaceae, though those of later 
epochs depart from the simple type by elaboration of the stele, as 
described above (p. 539). The close correspondence of the Hymeno- 
phyllaceae with certain of the Botryopterideae stamps their structure as 
relatively primitive also, though it shows some variants upon the simple 
protostelic state. Lygodiwm also, recognised as the most primitive genus 
of the Schizaeaceae, is protostelic, and the same is the case with the 
simpler species of Gleschenta; in fact, those early stocks of Ferns which 
are recognised by comparison of other characters, as well as by their 
geological history, as forming the phyletic basis of the Leptosporangiate 
series, show the protostelic structure, or a condition very little removed 
from it. 
The Ferns thus mentioned are all included in the Simplices, except the 
Hymenophyllaceae. These are exceptional among the Gradatae in showing 
a protostelic structure of the axis: most of the Gradatae have a more 
elaborate stem-structure, which may be held to be derivative from the pro- 
tostele, just as the basipetal sorus is probably derivative from the type of 
the Simplices. The probable steps towards a solenostelic state are illustrated 
in Lindsaya and Dennstaedta, and suggested also in Gletchenia; but the 
solenostelic structure is typically seen in Dzpteris and Loxsoma, as well as 
in the Dennstaedtiinae. Here at each leaf-insertion the vascular tube opens 
by a foliar gap. Where the internodes are long and the gap itself short, 
as in the rhizomic species, the structure is easily intelligible. It is but a 
slight step to the dictyostelic type, as seen in Ferns with short axis and 
overlapping leaf-gaps: the transition is illustrated in the Dennstaedtia- 
Davallia series, and has probably occurred also in the Alsophila-Cyathea 
series, and elsewhere. It seems probable that the progression from a 
- protostelic to a solenostelic or dictyostelic state has been effected -in 
several distinct phyletic lines, while the dictyostelic, with or without internal 
accessory strands, is the most elaborate system of all. 
It usually accompanies an advanced soral condition: that this is, how- 
ever, no obligatory parallelism is shown by the comparison of Aatonia with 
Dipteris. The latter retains a simple solenostelic structure of the axis, 
though its sori have progressed to the condition of the Mixtae: the 
