42 
SOUTH AFRICAN SPECIES OF 
details) is characterised by 4-merous flowers, in which the calyx is shorter 
than the corolla, and which have free petals, 4-ovuled carpels, cuneate 
(truncate at the top) and rather large nectar-scales. This genus includes 
Or. natans, Thunb. and Cr. inanis, Thunb. It was kept up by Harvey 
(l.c., 328) who corrected an unfortunate mistake of the authors (previously 
committed by De Candolle) and rightly stated that the ovules are solitary in 
the carpels. Ecklon and Zeyher further established the genus Sarcolipes 
(La, anno 1837, 290) which has 5-merous flowers, multi-ovuled carpels and 
linear nectar scales. To this belongs only the species named later by 
Harvey Cr. Sarcolipes, Harv. Amongst their genus Petrogeton (l.c., 391), 
which includes the plants placed by Harvey under his Crenato-lobatae , we 
find their Petrogeton alpinum placed by Harvey under Bulliarda (l.c., 330). 
It is almost certainly identical with Cr. umbellata, Thunb. 
Their genus Thisantha (l.c., 302) includes glomerata, L. and allied species 
with 5-merous flowers and 1 — 2-ovuled ovaries. They erroneously assume 
that these species lack nectar-scales. 
Lastly (as far as our part of the subject is concerned), they established 
the genus Tetraphyle (l.c., anno 1837, 292) which includes a great mixture, 
as it is chiefly founded on the habit induced by the 4-ranked imbricate 
leaves. They all have 5-merous flowers, but as some of the most important 
characters, such as multi-ovuled carpels, do not apply to some of the species 
here placed, we will not concern ourselves further with the genus. 
It will be seen that amongst writers on Crassulaceae there has usually 
been a tendency to put a number of small-flowered forms, with more or less 
separate petals and an androecium isomerous with the corolla, either into 
separate genera or into special groups in the genus Crassula. The former 
course has clearly been a failure. It was based on differences in the number 
of floral parts, the number of ovules in the carpels, the shape of the nectar- 
scales, but while these characters may be here and there usefully employed 
in placing nearly allied species into groups they are useless for generic 
distinction, as they may separate species which are otherwise closely allied 
and moreover they are sometimes not constant in one and the same species. 
I, therefore, have placed them all into the genus Crassula under the 
sectional name of Tillaeoideae 1 . 
I include in them the following from Harvey and Sonder’s Flora Capensis: 
the genera Helophytum, Bulliarda, Crassula sect. Lycopodioides, sect. Glome- 
ratae, sect. Filipedes (excl. Cr. centauroides and brachypetala ) and a number 
of species published after the second volume of the Flora Capensis. 
The DISTRIBUTION of the section is very interesting and parallelled by 
other groups of plants well represented in South Africa. The majority 
1 Bentham and Hooker in Gen. PI. ii, 656 had placed most of them under Tillaea, L., but 
they state “Inter Crassulam et Tillaeam nullum est divisum bonum.” 
