122 Salmon. — A Monograph of 
Kew Herbarium), in which the lower leaves are broader and 
less finely acuminate than usual, and show the rather firm 
areolation with thickened cell-walls of the Madagascan plant. 
As regards the limb of the leaf it may be noted that 
although this is sometimes continued to the extreme apex 
of the leaf in the Madagascan plant, it is in such cases always 
narrow, and I have not seen in the Madagascan plant the 
broad limb at the apex of the leaf that is found in some 
examples of .S', erythrodontus from South America. This 
character is, however, clearly one of little importance syste- 
matically, since on the same stem of some plants of S', ery- 
throdontus leaves may be found in which the limb vanishes 
below the apex, whilst other leaves are bordered to the apex. 
On one stem of ‘ S. HildebrandtiV from Muller’s herbarium 
two upper leaves occurred in which, presumably from exposure 
to unusual conditions, the cells at the apex of the leaf became 
suddenly, for several rows, on one side of the nerve only, 
prosenchymatous and thick-walled, so that the apex of the 
leaf in that region was composed wholly of prosenchymatous 
cells like those of the ‘ limb.’ This was all the more remark- 
able since in this form of the Madagascan plant the limb as 
a rule ceases at some distance below the leaf-apex. The 
suggestion may perhaps be made that a correlation exists 
between certain conditions of growth and the shape of the 
leaf-cells; if so, we might find an explanation in climatic or 
other ecological conditions of the fact that in some plants of 
5. erythrodontus the limb is composed of prosenchymatous 
cells continued to the apex of the leaf, whilst in others the 
limb ceases, and all the cells become parenchymatous. 
It is perhaps worthy of note that 5. erythrodontus type is 
slightly variable in the size of the leaf-cells ; and although the 
areolation is never, so far as I have seen, quite so lax as in 
the Madagascan plant, on the other hand South American 
examples occur in which the cells are much narrower than usual. 
The occurrence of the South American S', erythrodontus 
in Madagascar under the form of the variety Rutenbergii 
affords an extremely interesting case of geographical distribu- 
