Morphological Observations. ^99 



are, in all but A. rccitrvaia, Schonl., broader than the leaves, and 

 •completely cover them in. The buds and apices of the branches are 

 wrapped up by them thoroughly, and in dry weather, except again in 

 A. rcciirvata, they are closely pressed against the stem, so that 

 transpiration is stopped altogether, but when the atmosphere is 

 moister they open out slightly, so that the margin of the leaves is 

 just visible. This is especially striking in A. papyracea, but can 

 also be observed in A. ustulata and A. Ahtonii. They may even 

 then catch the dew, but again I have failed to satisfy myself that 

 they actually absorb it, and my observations seem even to [)oint to 

 ithe conclusion that such an absorption does not take place. 



The Flower. — The general structure of the flower is well 

 known. There are two sepals and five i:>etals, both of which are 

 <liflFerent in shape in the different species, though ovate and oblong 

 forms prevail. However, their minute description here would be 

 of no interest. The Flora Capensis ascribes to the genus i5-::o 

 stamens. In A. ustulata there are sometimes seven, but as a rule 

 ■eight stamens, five of which alternate with the petals, while three are 

 •epipetalous, thus indicating the presence of an epipetalous whorl. 

 In A. papyracea I counted about 16 stamens, in A. Ahtonii over 

 60, in A. filamentosa about 15, in A. arachnoides usually 27. It is 

 very desirable that the origin of these various numbers be carefully 

 traced, but I have not been able to do this yet. The pollen is 

 :globular, smooth. The ovary is globose in A. papyracea, A. Ahtonii. 

 A. ustulata, but more oblong in the species belonging to the section 

 Tele phi astr urn, especially in A. arachnoides. When we come to 

 -examine the placentation, we find that this has hithereto been either 

 loosely or in some instances even wrongly described. In the Flora 

 •Capensis it is stated that the seeds are affixed to a central placenta. 

 Unfortunately this term is applied both to a free central placenta 

 and to a central placenta which is connected with the roof of the 

 ■ovar) . Pax in Engler's Xatiirliche Pfianzenfamilien (III. i. 6. p. 53) 

 -expressly states that all Portulacaceae have a free central placenta 

 but I find that all species of Anacampseros (perhaps with the excep- 

 tion of A. ustulata. which requires re-examination) have a central 

 placenta which is connected with the roof of the ovarw Its shape 

 is different in the different species, but I need not here enter into 

 such details. The style is usually cylindrical, of various lengths. 

 but in A. papyracea it is practically absent, and the three stigmatic 

 lobes are therefore nearly sessile. 



The majority of the species of A. have very showv flowers. 

 These all agree in this : that they open for a few hours, in some 

 species in the afternoon only. Towards evening thev close 

 and never open again. In all these species self-fertilisation 

 is not only possible, but it takes place regularlv. I have never vet 

 seen an insect on any of their flowers. I have also kept them in 

 such a manner that insects could have no access to them, and \et 

 ripe seeds are in ever}- case produced. We have here, therefore, 

 the curious fact that the show-apparatus with which these flowers 

 iire so liberally pr(5vided is not necessary to them for the 



