302 Sykes . — Anatomy and Histology of Macrocystis pyrifera and 
arrangement of the pits in their walls, suggest that their function may- 
be the conduction of water. 
A much more definite boundary can be recognized between the cortex 
and medulla at this stage, at least in a transverse section, but this appearance 
is partly due to the fact that the ingrowing hyphae now originate only 
from the innermost cortical layer. From this layer they arise in large 
numbers and run more or less radially towards the centre; they often 
branch, and are thus to a large extent responsible for the increase in size of 
the medulla. 
From six to nine sieve-tubes are present in each radial row, and the 
innermost element of each row is generally already quite filled with callus. 
The primary pith filaments are more widely separated than in the younger 
stages, and in many cases their lumina are choked by callus. 
Stage v. The oldest stem which I examined was about half an inch in 
diameter (cf. Fig. 45, PI. XXI). 
From this stage nothing is gained to add to the description of the 
cortex already given in the last stage, beyond the fact that it is now eight 
or nine times broader than the medulla. 
Not more than nine rows of sieve-tubes, the two or three innermost 
rows of which are now completely obliterated by callus, were found in any 
radial row ; it would seem probable, therefore, that no new ones have been 
formed since the original primary inner cortex was transformed to the 
sieve-tube layer. On the other hand, young sieve-tubes composed of quite 
short cells without callus are still present, and in longitudinal section 
are often difficult to distinguish from the innermost cortical cells, though 
in cross section they are sharply differentiated from the latter by the 
much greater thickness of their walls. The innermost cortical cells have 
always a ring of pits on their transverse walls, while the young sieve-tubes 
have definite sieve-plates, and it is not easy to imagine how an adult wall 
with its ring of pits (such as that figured in Figs. 1 and 2, PI. XIX) could give 
rise to a sieve-plate. Certainly no evidence of such a transition is forth- 
coming. In order to prove definitely that young sieve-tubes do not continue 
to be formed slowly from the inner secondary cortex throughout the life 
of the plant, it would be necessary to investigate some considerably older 
material. On the whole, however, from the evidence forthcoming, it seems 
probable that the increase in size of the medulla in the oldest specimens is 
due partly to the great increase in diameter of the elements composing 
it, but still more to the large number and repeated branching of the hyphae 
anastomosing among them. 
In this old material hyphae which contain callus are occasionally met 
with. Some of these have taken a longitudinal course, and have very 
much the appearance of extremely narrow sieve-tubes. They are composed 
of elongated cells, swollen af the cross walls, and have dense contents ; 
