Polycycly. 
235 
central strand with an adaxial member of the internal ring, then of 
the strands of this ring with the adaxial members of the outermost 
ring. This takes place so gradually that a number of the internal 
strands continue their course above the pulvinus right up into 
the region of the rachis. 
Angiopteris evectn, as is well known, exhibits, in its huge fleshy 
stem and leaf-bases, the most complicated type of vascular system 
known, not only in Marattiaceae, but in any Fern, living or extinct. 
Nevertheless it is quite clearly an elaboration of the simpler 
Marattiaceous type along the lines we have already traced. In 
place of the two or three concentric dictyostelie cylinders of 
Marattia, large stems of Angiopteris may show as many as nine 
distinct concentric circles of strands as seen in transverse section 
(Fig, 78). 
The adult stem, comparatively narrow at the base, rapidly 
increases in size towards the apex. In the specimen examined by 
Miss Shove (’01) the stem had a length of 25 cm., and was 8 cm. 
broad at the top, tapering to 1 cm. at the base. The dorsiventral 
structure of the base was well-marked, the lower surface being 
destitute of leaves but thickly covered with roots for a distance of 
8 cm. from the basal extremity. At the apex the stem was nearly 
radial in structure, correlated with the diminished production of 
roots. The meshes of the vascular network are much larger on 
the ventral root-bearing, than on the dorsal leaf-bearing side. 
“ The roots originate, a few in the outer, the majority in the inner 
zones ” of vascular strands, some actually arising from the strands 
situated near the upper surface of the stem, penetrating most of its 
thickness and emerging on the lower surface. At the base of the 
stem the vascular system is simple, increasing in complexity 
upwards, in exactly in the manner described in the other genera, 
till a number of concentric, highly dissected dictyostelie cylinders 
are formed. Owing to the shape of the stem each of these 
“ cylinders ” has the general form of an inverted cone, as pointed 
out by Mettenius (’65). As each leaf-trace, consisting of several 
strands of the outermost cylinder, passes off, a segment of the 
second cylinder moves out to fill the gap, its strands anastomosing 
right and left with those of the outer cylinder bordering the leaf- 
gap (Fig. 79). Owing to the crowding of the leaf-insertions, 
however, a given set of strands remains in the outer cylinder for a 
very short distance only, the compensating strands which have 
filled any given gap almost immediately contributing to the traces 
