272 
W. O. Howarth. 
of these seedlings are all intravaginal. From these observations 
it would appear that the intravaginal branch is the more primitive 
type, and that development has proceeded towards extravaginal 
branches accompanied by a change in the direction of growth, a 
reduction in the size of the prophyll, and a reduction in the 
succeeding leaves to scales by partial or complete loss of laminae. 
The third subvariety, glnucescens, in respect of the caespitose 
habit approximates more to tenuifolia. Transference to garden soil 
affects both these forms by encouraging them to produce longer 
stolons, but even under these conditions they remain quite distinct 
from grandiflora. 
Fig. 4. Subvar. tenuifolia. Seedlings showing transition front true 
intravaginal to true extravaginal branch. 
2. The Vegetative Shoot. 
All branches develop a shoot which remains purely vegetative 
during the first season’s growth. The extent of development 
depends upon whether it is laid down early or late in the season. 
The usual number of leaves on a shoot is four or five. They are 
produced according to a regular alternation but become twisted 
out of position by a torsion of the axis. From the axils of these 
“ radical ” leaves branches arise ; extravaginal from the older ones, 
intravaginal from the younger. Each shoot thus becomes the 
centre of a rapidly increasing branch system. The intravaginal 
branches form a compact turf; the extravaginal ones spread out in all 
directions. All become new centres of growth, thus tending 
to a raising of the general level of the turf. When, as frequently 
happens, the turf becomes covered with loose soil, debris or 
alluvium, not only do the new growths extend upwards but the 
