608 
Telopea Vol. 6(4): 1996 
Karroochloa 
Karroochloa includes four species, in two very distinctive groups. Within an African 
context there is no possibility of confusing these two species groups with any other 
grasses, as they are the only members of Ri/tidospenua s.l. in Africa, but it is more 
difficult finding a synapomorphy combining them in a global context, and the only 
morphological synapomorphy found is a rounded sinus base. In addition, 
Karroochloa is the only diploid group in Rytidosperma s.l., the rest of the group being 
tetraploid or with higher ploidy levels. The four species are very variable in terms 
of lemma indumentum. 
Clayton and Renvoize (1986) suggested inclusion of Karroochloa, as well as 
Merxmuellera, in Rytidosperma s.l. Merxmuellera is very variable, and most likely 
contains several disparate elements which are not closely related (see also comments 
in Tomlinson 1985, Ellis 1982, Barker & Ellis 1991). There is no indication from the 
leaf anatomy for a relationship with Rytidosperma, and the caryopsis shape is quite 
different (Barker 1994). Although the lemma indumentum of several species of 
Merxmuellera is tufted, these tufts do not form regular transverse rows as is typical 
of most species of Rytidosperma. Karroochloa is indeed very close to Rytidosperma s.l., 
and the only character by which it can be separated is the chromosome number of 
2)1 = 12. However, its inclusion in the Rytidosperma s.l. clade would appear to be 
unlikely, and would need to be investigated further. We suggest that placing the two 
Karroochloa species within Rytidosperma may be erroneous, and the result of 
convergence between the two genera. The chromosome numbers indicate that 
Karroochloa should occupy a basal position relative to the rest of the clade, as it is 
highly unlikely that there could be a reversal of a diploid number of 24 to 12. We 
would like to see Karroochloa to be the sister group to the 2n = 24 Rytidosperma clade. 
Nonetheless, the most parsimonious solution suggests that the diploid doubling 
either occurred three times, or that the reversal did in fact occur. Forcing Karroochloa 
to a basal position, followed by a single doubling in the chromosome number, results 
in an increase in tree length from 193 to 196 steps. 
Rytidosperma s.s. 
The most distinctive subclade, containing Monostachya, Erythranthera and Pyrrhanthera, 
as well as about half of Rytidosperma sensu Zotov, is defined by the presence of 
disarticulating leaf blades and narrow metaxylem vessels. Zotov (1963) recognised 
part of the group as Notodanthatiia sect. Buchamnia Zotov, defined by the folded, 
glabrous leaves, red anthers, narrow paleas and short lemma lobes, and separated 
other parts of the clade as the segregate genera Erythranthera and Pyrrhanthera. Although 
the leaf indumentum, palea shape and anther colour characters have exceptions among 
the Australian and New Guinean taxa, this remains a rather distinct group of small 
plants with tough, folded, usually deciduous leaves, and with the upper row of lemma 
hair tufts weakly defined, often more or less reduced. The species are numerous and 
common in New Zealand; in Australia and New Guinea the plants are found in 
montane or alpine habitats. Zotov, and Connor and Edgar (1979), who largely agreed 
with Zotov's infrageneric classification of Rytidosperma, did not consider Monostachya, 
En/thranthera and Pyrrhanthera to be part of this clade. The South American species of 
Rytidosperma all belong to this clade, too. 
Monostachya was originally described for R. oreoboloides) Jacobs (1982) also included 
R. craigii, R. montis-wilhelmii and R. nardifolia, and argued for the separation of 
Monostachya from Rytidosperma. Veldkamp (1980), while arguing for the inclusion of 
Monostachya in Rytidosperma, listed (a) the cushion growth-form, (b) inflorescence 
reduced to one spikelet, (c) the shortly bidentate and aristate inflorescences and 
