605 
and their Mycorhiza . 
There is a well-marked endodermis in the cells of which no fungus is 
apparent. This encloses a phloem of elongate, thin-walled cells, which are 
usually in a good state of preservation. The transverse walls are slightly 
oblique. 
The rootlets are usually diarch. This point was not known to William- 
son when he described them originally, but the following observations can 
be seen in his manuscript catalogue of slides in the British Museum. ‘ There 
is obvious evidence that the vascular bundle originated in a bipolar manner 
and may be centripetal in its further development ; this fact was not 
observed when Memoir 5 was written.’ The diarch origin is very well 
shown in some of the younger rootlets in Mr. Watson’s slides, two groups 
of protoxylem being visible separated by a few thin-walled elements not yet 
lignified. At the same time triarch rootlets are not uncommon. I have 
not seen any tetrarch or pentarch rootlets , though in full-grown roots with 
secondary thickening this condition is to be met with not infrequently. 
It may have been that only the primary roots had the larger numbers 
of protoxylem-groups, and the chance of finding these is, of course, not 
great. In longitudinal section the tracheids are seen to be spiral in the 
protoxylem. The rest of the centripetal wood has scalariform tracheids 
with oblique ends. The tracheids of secondary growth are of typical 
Cordaitean structure with bordered pits. 
In slide A 281 I have been fortunate enough to come across a longitu- 
dinal section of a root apex (PI. XLVI, P"ig. 3). There is every reason to 
believe that this belongs to Amyelon , though there is no actual connexion. 
There area number of unmistakable young rootlets near ; these are the only 
well-preserved roots on the slide, and it agrees with them in every way that 
I have been able to compare them. 
A well-marked root-cap, triangular in shape, and of presumably corky 
cells (measured across which, the root is about -4 mm. in diameter), protects 
the meristematic zone. Unfortunately the section is not absolutely median, 
so I am unable to trace the full extent of this tissue. The few cells that are 
visible are quite small (8 \P) and very thin walled ; the dermatogen, periblem, 
and calyptrogen are not distinguishable as is usual in Gymnosperms. 
Petrifaction appears to have occurred at a period of rest, for behind this 
region rapid differentiation occurs, the elements derived from the periblem 
and pleurome being easily distinguishable. Only one spiral tracheid is to 
be seen against the latter cells, which otherwise are elongate and parenchy- 
matous. The fungus is absent from the growing region. About 1 mm. 
from the apex (the bit of root is but 1*5 mm. long) traces of hyphae are to 
be seen running longitudinally in the inner cortex. Infection of the rootlets 
appears to have been from behind forwards as they elongated. 
The rootlets appear to have had a typical endogenous origin. Before 
secondary thickening began in the roots, before periderm had begun to 
