by Rosanorr (1866 p. 70), but as shown by Sorms (1881, p. 24), they are really 
hairs or hair-producing cells. They are easily recognizable after the hair has been 
shed, showing a scar left by the latter; I propose to name them trichocytes or 
hair-cells. 
The sporangia are always divided by one or three transversal walls, dividing 
the cells into two or four spore-cells. Vertical divisions I have only found as rare 
exceptions in Lilhothamnion Sonderi (fig. 137 E, F). The tetrasporie sporangia are 
first divided by a transversal wall, but the formation of this wall proceeds slowly 
from the periphery towards the centre, and the formation of the two following 
walls has frequently begun before that of the first is completed. In Corallina offi- 
cinalis I found that the primary nucleus of the young sporangia divides into four 
nuclei which arrange themselves in a longitudinal row in the middle of the spor- 
angia, and that these rest for a long time in this stage before the divisions, which 
take place almost simultaneously (fig. 197). Also in Epilithon membranaceum the 
three divisions are almost simultaneous (fig. 152, comp. otherwise fig. 134). The di- 
viding wall is shown in figs. 131 B and 142. 
The number of spores is constant in most of the species, either 4 or 2; but in 
some species both disporic and tetrasporic sporangia are met with. One of these 
species is Lithothamnion leve which, however, in the Danish waters has only been 
found with disporic sporangia. The above mentioned fact that the divisions of the 
tetrasporic sporangia are, at least in several cases observed, almost simultaneous, 
makes it improbable that the disporic sporangia can here be interpreted as unripe, 
not fully divided. It is an incontestable fact that some species may, according to 
circumstances, have tetrasporic or disporic sporangia. This I have also found to 
be the case in L. Lenormandi, in which only tetrasporic sporangia were previously 
known. In Melobesia Fosliei also, and in M. minutula, both kinds of sporangia would 
seem to occur. 
In material fixed in Juel’s liquid the protoplasm of the tetraspores showed a 
foamy structure. The central part containing the nucleus was brighter and distinctly 
marked off from the outer (figs. 132, 142 B and Plate III fig. 1). 
The antheridia present considerable differences as to their position and devel- 
opment. In Lithothamnion Lenormandi they have a similar position to that pre- 
viously described in L. polymorphum, being produced on the surface of great bushes 
extending from the periphery towards the centre of the conceptacle (Plate III fig. 2). 
If this structure is to be found in all the species of the genus, we have here an 
important generic character. In Epilithon membranaceum, referred by some authors 
to the genus Lithothamnion, the antheridia are, as shown by GuiGnarD (Rev. gén. 
de Bot. I. 1889, p. 182) seriate in short filaments clothing the bottom of the con- 
ceptacle, and in the other genera the antheridia (spermatangia) are also placed on 
the bottom of the conceptacle, being produced as outgrowths from a layer of small 
cells, but they are not seriate. The antheridia are more or less lengthened, short 
cylindrical or upwards somewhat thickened and more or less curved. In Melobesia 
