West.— A Contribution to the Study of the Marattiaceae . 391 
ring of bundles at the time being, no strands from the second cylinder 
contributing directly to the leaf-trace. 
The year 1902 was marked by a great advance in our knowledge of the 
stelar system in this genus, when Farmer and Hill ( 29 ) published a full 
account of the arrangement and development of the vascular tissues in 
Angiopteris evecta\ this paper was illustrated by means of a series of wax 
models representing the stelar skeleton of this plant at various stages in its 
development. In this way, a most accurate description of the successive 
stages in the elaboration of the complicated vascular system of the adult 
plant was obtained. A gradual transition was traced from the solid axile 
rod of vascular tissue (= protostele) of the very young sporeling to a hollow 
cylinder or siphonostele, and with perforations corresponding to foliar gaps 
enclosing a core of pith, which they regarded as distinct from the now 
tubular stele. Sooner or later, the gap above one leaf fails to be repaired 
until after the exit of the traces of the next leaf ; in this way a typical 
dictyostele is produced. Meanwhile commissural strands are differentiated 
across the central parenchyma and serve to connect the opposite sides 
of the stelar cylinder. ‘ Finally ( 1 . c., pp. 377-8) the siphonostele opens out 
to a considerable width ( 1 . c., PI. XVI, Fig. 6), whilst the axile commissures 
assume an ever-growing importance forming a sort of sympodial column. 
. . . The leaf-traces also become more complex, and anastomoses take place 
at irregular intervals with the strands which can still be recognized as 
the relics of the original siphonostele, as well as with one another. Irregu- 
larities also begin to become apparent as to the relative height at which the 
two members of the leaf-traces become freed from the plexus of tissue, and 
a stage is thus reached at which the vascular skeleton appears to consist of 
a stout axile strand surrounded by upwardly diverging zones of steles which 
ultimately pass out above to the leaves. The complexity and obscurity is 
primarily due to the commissural strands which connect up the margins of 
siphonostelic foliar gaps, and the whole arrangement is to be correlated 
with the presence of the bulky parenchyma of the stem.’ According to 
these observers, the roots of Angiopteris sometimes unite with the more 
central strands, though far more commonly with those peripherally situated, 
but Shove ( 1 . c., p. 506) states that the majority of the roots originate from 
the inner zones, although a few arise from the outer ones, usually at the 
points where the strands anastomose. 
In his monograph on the 4 Eusporangiatae ’, Campbell (20, p. 200) 
states that ‘ in the early stages Angiopteris appears to agree closely with 
the other Marattiaceae in the development of its vascular system, but the 
single central stele without leaf-gaps is retained much longer than in the 
other genera, and it also becomes much larger and has a better-developed 
xylem, and the open dictyostele, formed from the anastomosing of the early 
single leaf-traces, characteristic of Danaea and Marattia , is not present ’. 
Thus it is seen that the anatomy of this genus has already been 
