Vascular System of Pteris. 13 
the most part independent of the outer primitive system, but, as 
Stenzel shewed, having occasional connexions with it. It is note¬ 
worthy that these connexions, according to Stenzel, are mainly, 
though not exclusively, 1 lateral, i.e. in much the same position as the 
connexion between the internal cylinders and the solenostele in 
Pteris incisa var. integrifolia (compare Figs. 16 and 50, with Figs. 6 
and 7). Figs. 36—57 represent selected sections from a series 
through two successive nodes of the rhizome diagrammatically 
represented in Fig. 35. The petiolar system is in this case consi¬ 
derably simpler than that of the node represented in Figs. 14—34, 
and hence its resemblance to Pteris incisa var. integrifolia is much 
less marked, but it will be seen that the relations of the petiolar and 
rhizomic strands is identical with that already described. These 
relations in fact appear to be perfectly constant. 
It is difficult to resist the conclusion that the vascular system 
of Pteris aquilina has passed through a stage like that now repre¬ 
sented by Pteris incisa var. integrifolia on the way from its mono- 
cyclic ancestor. It does not of course follow that this intermediate 
stage was solenostelic; dictyostely may have arisen at any point in 
the series. 
A few words must be devoted to the relation of the insertion 
of the leaf-trace to the branching of the rhizome. Two distinct 
types of branching exist in this species, first, ordinary dichotomy, 
in which the vascular system of the rhizome typically divides into two 
approximately equal parts, 2 and which has no necessary relation to 
the origin of a petiole ; and secondly, the lateral origin of a shoot- 
bud from the posterior side of the base of a petiole (Figs. 13 and 
26—33), in which the vascular system of the bud is derived entirely 
from that of the petiole. It is, however, very common for the 
weaker branch of the dichotomy to bear a leaf on the inner side 
more or less close to the fork of the rhizome. If the insertion of 
this leaf is sufficiently removed from the fork, the petiolar strands 
arise from the rhizome branch-system in the ordinary way. This 
seems to be Mettenius’ second case, referred to above, but why he 
should consider the stretch from the base of the petiole to the fork 
as belonging to the petiole is not obvious. If the insertion of the 
leaf is closer to the angle, the strand forming the lateral fold of the 
1 We have not seen the dorsiventral connexions of the inner 
and outer strands figured by Stenzel (’61. Taf. V., Figs. 8 
and 9), but it appears that they occur in larger rhizomes 
than those we have worked with. 
* The division is however very often far from being equal, see 
below. 
