446 DR. ERIC DRABBLE ON THE 
apical meristematic procambial strand, but by the appr oximation of a central free strand 
which finally closes the gap in the arc, so completing the circle. The xylem-bundles of 
Fig. 2. 
the strand are three in number and radiate from a common centre alternating with three 
phloem-groups. When fusion with the incurved are takes place one of these phloem- 
groups (a) disappears, while the two xylem-bundles, originally orientated towards the 
exterior of the root, rotate internally as seen in text-fig. 2. 
Another interesting case is shown in text-figs. 3, 4, 5, and 6. In fig. 3 a complete 
cylinder has been established somewhat externally to the general ring of ares of fibrous 
tissue. Followed distally this ring opens out internally, and at the same time one of the 
fibrous wedges bounding the gap between the two arcs just internal to the ring under 
consideration breaks free and passes internally as a medullary strand, its vascular 
elements being reduced to a single central vessel (text-figs. 4 & 5). At the same time 
Fig. 3. Fig. 4. 
Ses CH m 
WWE, ao 
one edge of the internally opening ring establishes connection with the edge of the are 
from which the medullary strand has become separated (text-fig. 6). This method of 
formation of a medullary strand is widely different from the usual method and has not 
been observed in any other case, the medullary strands being normally the reduced 
distal prolongations of the internal cylinders of the base of the root. 
It has already been mentioned that generally the whole series of changes culminating 
in the formation of a normal complete central cylinder takes place in the intracauline 
portion of the root. This occurs, for instance, in Latania borbonica, and Pl. 51. fig. 102 
shows the vascular cylinder of this species dissected out. The upper portion, which is 
enclosed in the cortex of the stem, is lobed, while the lower portion belonging to the 
external root is reduced to the normal condition. 
. Pl. 51. figs. 103 & 104 are from external portions of roots of Kentia. In fig. 104 
the upper end shows a series of free strands which fuse below, resulting in ii lobed 
cylinder. 7 
_A very curious appearance is presented in Pl. 51, fig. 105. The root appeared to 
dichotomize, and on dissection of the vascular cylinder this idea was strengthened as the 
partially separated strands are continuous, some into one section of the dichotomy, 
the rest into the other. Were there any trace of a blind ending of a principal root, the 
