Nr. 1] SUBAÉRIAL ALGAE FROM SOUTH AFRICA 27 
with our knowledge of other species of the genus in this regard, 
wnere the shape of the cells may be very varying — not only in the 
proportions of length and breadth. but also as to the shape of the 
cells on the whole — this character does not seem to be a suitable 
basis for the distinetion of species. Thus, both cylindric and 
inflated cells are to be found together in the same species, not in 
one and the same sample only, but even in the very same filament. 
Other authors attach great importance to the breadth of the cells, 
Ir. abietina being recorded to have narrower cells than Tr. aurea, 
but the range of variation of both species with regard to this is very 
large, and as the limits of both species not only are drawn very 
differently by the various authors, but even over-lap (Tr. aurea 
8—30 u and Tr. abietina 4—10 yu), is does not seem possible, also 
with our experiences from other species in the genus — to base a 
distinction between the two species on this character either. 
I have really been looking in vain for å good systematic charac- 
ter on which might be based a distinetion between these two species. 
HARroT, indeed, says: I. c€. p. 52 on Tr. abietina: «Tr. aureae formis 
gracilioribus adeo proxima ut non certe aliquando distinguatur». 
Neither the shape, size nor arrangement of the sporangies seem fo 
give safe hold with regard to a distinction between the species men- 
tioned, as both of them, as known, are subjected to considerable 
variations, and the limitations also in this respect must be made 
very wide. 
The accompanying plates VIT and VIII will give an idea of the 
appearance of the specimens found by me. The cells are eylindrie, 
or slightly inflated, 6—10, usually 7—9 u broad. 1—2 times, usually 
14% times as long as broad. In some samples, however, are to be 
found broader filaments, to 17 4 broad (sample no. 294). They form 
long, straight, or slightly curved filaments, more or less branched 
in various ways. The branches are sometimes spreading, at nearly 
right angles, or more appressed. The cells of the branches de not 
differ perceptibly either in thickness or in length from those of the 
main filament. The terminal cells are never acute at the top, but 
senerally rather obtusely rounded. The membrane is comparatively 
thin and colourless, glabrous, or sometimes finely erenulate. Å real 
reticulation is not to be found on the membrane, however. The 
gametangies are usually spherical, rarely slightly ellipsoid, or at 
times ovoid, 15—25 u in diameter. Their membrane is thin and 
glabrous, or crenulate. Specimens with crenulate membranes, 
however, may at times have glabrous gametangies. The game- 
tangies are terminal or lateral, borne directly on the vegetative 
cells, and are to be found both on the main filament and on the 
branches, singly or in series. At times, two gametangies may arise 
from one vegetative cell. (Pl. VIII, fig. 250 and 252). The game- 
tangies are opened by a hole, in the membrane, through which the 
