Î358 WILLIS: MORPHOLOGY OP THS PODOSTEMACEÆ 
apex, which very rapidly widens to its full size, the endo- 
genous secondary shoots may be seen developing, with a 
branch of the central vascular strand going to each. The 
apex is continually branching, so that the thallus assumes 
the general form shown in PL XXII., fig. 3. 
In cross section the thallus when in young vegetative con- 
dition shows a slightly thickened rib in the middle, and 
long thin lateral wings turned down at the extreme margin 
(PI. XXII., fig. 7). The vascular strand is rather below the 
centre, as usual. It is shown on a larger scale in fig. 8, and 
has two well-marked groups of tissue, as described by 
Warming, more or less separated by prosenchymatous tissue 
between. Its subsequent growth, and the meristematic 
activity that goes on, are closely similar to what occurs in 
D. elongata, and have been described in part by Warming, 
whose accounts of the peculiar bundles of the very dorsi ven- 
tral types of Podostemaceæ appear to have been somewhat 
overlooked by anatomists. 
Growth in thickness also goes on in the thallus itself out- 
side the bundle. Instead of being, as in D. elongata, almost 
equal throughout the thallus tissues, it is practically confined 
to the tissue above and beloAV the bundles, and in these there 
are more tangential than vertical divisions, especially on the 
upper side. PL XXII., fig. 9, shows a transverse section at 
an early stage, a little to the right of the bundle. The upper 
cells are mainly dividing tangentially, while the lower are 
also dividing vertically, the result being an upeurving of the 
thallus. It is in this way that the shape of the foot arises 
and similar causes produce the curvatures and irregularities 
which are so marked in these Fucus-like thalli. The lateral 
wings of the thallus do not take part in this growth in 
thickness, and thus a stout central rib is gradually formed, 
just as in Fucus itself. In transverse section near the base 
of an old thallus the result of this continual tangential 
division is often very striking, as shown in Warming’s figure 
(42, II., PI. XI., fig. 22), where there are long, vertical, 
parallel rows of cells not unlike a very deep palisade layer 
