[PENHALLOW] OSMUNDITES SKIDEGATENSIS 5 
way as to give a columnar or basaltic aspect to the specimens (Plate L., 
fig. 1), while they also contribute to a very ready and often extended 
removal of external parts whereby, under the influence of even slight 
mechanical effects, the size of the specimen may be readily reduced. 
It is, therefore, possible to account for the differences in diameter pre- 
sented by some of the specimens, more particularly when it is recalled 
that some of the smallest had undoubtedly been exposed for some time 
to the action of salt water, as shown by the presence of barnacles. 
In nearly all the specimens, the stem enlarges gradually upward. 
As shown in the photograph (Plate I., fig. 1), a stem having a total 
length of 9°5 em. is 2°5 cm. broad at the base and 3°5 cm. broad at the 
upper end. In two specimens no such variation was observed, from 
which the conclusion was drawn that the tapering specimens represent 
the basal portions of stems, while the latter were derived from a higher 
position where the diameter was more uniform. This is in harmony 
with the view already stated, that the specimens as now found belonged 
to stems of far greater length than any of the single fragments. These 
views are also sustained by evidence derived from comparison with the 
rhizomes of related existing species. 
In nearly all the specimens the surface is made up of a series of 
longitudinal grooves and ridges, with no evidence of cortication; but in 
two cases there were found prominent areas of glistening, coaly matter 
suggestive of the former presence of a structure allied to a cortex. A 
brief examination, however, was sufficient to disclose the fact that the 
coaly matter had been derived from structures similar to those consti- 
tuting the surface ridges, and this conclusion, taken in connection with 
evidence derived from comparison with existing species, led to the final 
conclusion that no real cortex could have been present in the original 
plant, and that the coaly matter represents the decayed bases of the 
outermost portions of the stipes. 
The longitudinal ridges to which reference has already been made, 
and which have been described as of a dull red colour, are often several 
centimetres long. They are normally flattened radially into a some- 
what broadly lenticular transverse section, and are generally about 5 mm. 
broad (Plate I., fig. 1) and fig. 1. Where most numerous, they are 
separated by very slight intervals in such a way as to suggest over- 
lapping in the original plant. It is, moreover, seen that they arise at 
different levels on the central axis, and in such a way that the newer 
ones are continually overlapped by the older, thus bringing about the 
upward enlargement already noted. (Plate I., fig. 1.) It is also 
observed that these ridges vary in size and number on opposite sides 
of the stem. Thus on one side, as shown in Plate I., fig. 1, they are 
