700 Yendo .— The Development of 
in the transitional region and finally unite together at the base of the 
blade. 
Fig. 24, PL LIV, is drawn with the aid of Abbe’s reflecting camera from 
a cross-section at the transitional region of a frond about 12 mm. in total 
length. The three ribs are clearly shown. The essential part of the ribs 
is built up of the subcortical cells, which are much more increased in 
number than in the other parts. The epidermal layer is a single sheet 
of cells, now anticlinally elongated. 
The figure shows at the same time how the medulla is generated. 
The two layers of the precortical cells are separated one from the other 
by the intervening space filled with a gelatinous matrix. The latter sub¬ 
stance is undoubtedly secreted from the precortical cells. As a consequence, 
the cell-ends facing the gelatinous matrix become convex by their own 
turgescence. The hyphal cells are now generated from some of the pre¬ 
cortical cells on the side towards the space. At the time when the hyphae 
are just formed, they contain, as in the cortical cells, a small number of 
chromoplasts. It is, however, in the later stages that the hyphae fill up the 
medulla as a compactly interwoven filamentous tissue. 
When a blade with the hyphae just formed is observed from the 
surface under a low power, they are seen as delicate veinlets traversing 
longitudinally the polystromatic area (Fig. 15 , PI. LIII). Their course has 
been very difficult to trace, owing, no doubt, to the fact that they run freely 
through the gelatinous matrix. This was also the case in Undaria . In 
Laminaria the two precortical layers remained in close contact for a con¬ 
siderable time after the hyphal cells had been well developed (Figs. 44-6, 
PI. LV). More precise accounts of the formation of hyphae will be therefore 
given in a later chapter. It is noteworthy, however, that in some specimens 
of Costaria measuring 10 cm. or more, the upper part of the four-layered 
area remained without the hyphal cells. 
Soon after the formation of the three ribs, a row of tiny bullations begin 
to attract our attention on both sides of the midrib (Fig. 10, PL LIII). The 
lamina continues its stipo-frondal growth, adding new bullae successively 
below. The bullation of Costaria Turneri is not a simple over-extension 
of the lamina at certain points. From the beginning it is an elevation and 
depression occurring regularly and alternately, and parallel to one another. 
They are slightly transversely stretched at right angles to the adjacent ribs. 
At a somewhat more advanced stage an end of the elevation—depression on 
the other surface—bends at right angles and parallel to the rib. The 
depression adjacent to the elevation also bends in like manner, but at the 
opposite end and in the opposite direction to fit the former. The process 
will be more easily understood by a glance at the illustration, than by 
imitating an enunciation by Euclid (Figs. 11, 12). Briefly speaking, 
the bullation reminds us of a sort of carved fretwork. These bullations 
