265 
Archangiopteris Henry i and other Marattiaceae. 
in turn disappears by fusing in a similar manner with the outermost ring, 
or typical curve. All this takes place so slowly that a number of the 
internal strands continue their course from the pulvinus far up into the 
region of the rachis. According to Bertrand and Cornaille ( 1 . c.), the same 
holds true for the internal strands in the petiole of Angiopteris . 
It is known also that the vascular system of the petiole of Helmintho- 
stachys is characterized by the presence of an internal strand, and in the 
light of these facts it becomes necessary to determine the exact manner in 
which it arises. In this plant the leaf-trace departs from the stem stele as 
a single circular or elliptic strand with a mesarch protoxylem. In the free 
petiole the protoxylem becomes endarch, the endodermis disappears, and 
the leaf-trace divides up into several strands which arrange themselves along 
a curve almost completely closed adaxially. Then in all the petioles 
examined I was astonished to find that the terminal strand of one arm 
of the curve, and of one arm only , sinks inwards towards the centre of the 
petiole, and turns round so that its protoxylem faces adaxially. This 
internal strand has, therefore, a different origin from those in the basal 
pulvinus of the Marattiaceae, and can only be compared with the internal 
strands in the upper part of the petiole of Kaulfussia and Archangiopteris , 
the surprising feature about it being that it represents the adaxial terminal 
strand of one arm only of the curve. 
To return again to Archangiopteris ; the manner in which the pinnae 
obtain their vascular supply from the rachis is worthy of especial notice. 
The terminal strands of the petiolar curve are always separate from each 
other in the neighbourhood of the insertion of the pinnae, and each pinna 
receives two strands from the rachis. One of them is comparatively large, 
and comes from the side of the curve (Fig. 17, p. tr.) ; the other is small, 
and comes from the extremity of the incurved terminal strand on the same 
side. This small strand travels right across the central ground-tissue 
within the curve to fuse with the adaxial end of the large strand 
(Fig. 17, st.). In my examples the small strand never became quite 
free from the terminal strand, but moved along in contact with its inner 
surface. 
In the petiole of Kaulfussia , also, a strand passes across the internal 
ground-tissue from the extremity of the incurved terminal strand of the 
curve to join those that depart from its sides into the branch. Again, 
Bertrand and Cornaille (1. c.) describe a similar procedure in the petioles of 
Marattia fraxinea and Angiopteris. They find that the primary branches 
are supplied with vascular strands from the internal system (which is 
continued from the basal pulvinus far up into the rachis) in addition to 
those they obtain from the outer ring, and this is also confirmed in the 
Marattia that I myself examined. 
I believe, with these authors, that this method of vascular supply to the 
