38 
Transactions of the Royal Society of South Africa. 
Ill Scample 827 there occurred numerous Cosmaria diverging more or less 
markedly from 0. suhprotumiclum. proper, as described and figured for 
instance in Messrs. West's monograph. It has not proved possible 
to discriminate between these different Cosmaria and, although the extreme 
forms are sufficiently distinct from C. subiwotitmidum, they are connected 
with it by a whole series of transitional forms (Fig. 15). All of these forms 
agree to a more or less marked extent in certain respects which constitute 
points of difference from the typical G. sulprotumidum. These common 
features are as follows. 
Special granules are not developed in the central part of the semicell, 
although now and again slightly larger ones are found in this region. This 
feature is connected with a weaker differentiation of the central tumour — a 
fact which is very obvious in the side- and end-views (Fig. 15). In a few 
cases there appears to be a semicircular bare area between the marginal and 
central granules of the semicell. as seen in front view — an approximation to 
the condition found in the type — although the central granules are not 
larger than the others. As a rule, however, the entire semicell is uniformly 
granulated, the granules being arranged in more or less well-marked 
concentric series and often also in radiating series, although the arrangement 
is always more or less irregular in the centre of the semicell. In all the 
specimens, lastly, the apex appears smooth and truncate, an occasional 
appearance of undulation (as in Fig. 15, YI and VIII) being due to the 
presence of three to four granules just within the margin. 
It may be noticed that W. West (Fresh w. Alg. W. Ireland, Jourii. Linn. 
Soc, Bot., xxix, 1891, p. 157, PI. XXIV, fig. 21) has described a form of 
C. subiyrotumidmn with more scattered granules and weakly differentiated 
central granules {cf. also Messrs. West's monograph, loc. cit., p. 232, 
PL LXXXVI, fig. 22), and such a form helps to link up the specimens here 
described with the typical C. suhprotumidum ; it would be interesting to 
know the character of the side- and end-views of this form, since they 
would presumably show reduction of the central tumour. The form 
described by Schmidle as C. occultum Avould differ from all these in the fact 
that granules are altogether wanting in the middle of the semicells, and it is 
significant that Schmidle's rather poor figure (Engler's Bot. Jalirb,, xxxii, 
1902, Tab. I, fig. 25) shows a relatively slight tumour in end- and 
side-views. 
The specimens in sample 327 that come nearest to C. subprotumidum are 
shown in Fig. 15, VIII, IX, and X (IX having the same sort of end- and 
side-views as VI). Such specimens are generally of rather small dimensions: 
long., 21-22 /X (rarely as much as 27 /x, or even 30 /x, the greatest length 
noted) ; lat., 18-19 /x (rarely up to 25 /x) ; lat. isthm., 5-6 /x; crass., 14-15 /x. 
They agree more or less with the type in the shape of the semicells (especially 
Fig. 15, VIII and X), in the rectangular basal angles, and in the fact that 
