34 Transactions of the South African Philosophical Society. 



with muslin for about a week I found afterwards a fairly copious 

 supply of nectar in all flowers which had reached the stage above 

 mentioned, and at the same time I noticed that the epipetalous knobs 

 did not show any secretion whatever. Furthermore, by means of 

 anatomical investigation I could not demonstrate in them any tissue 

 that resembled the well-known excreting tissues of other vegetable 

 glands. The nectar inside the flowers appeared to have been secreted 

 by the small scale-like bodies, which in this plant, as in the majority 

 of crassulaceous plants, are found at the bases of the carpels on their 

 outsides. It seemed to me, therefore, certain that the above-mentioned 

 insects came to the flowers for the purpose of sucking nectar, and it 

 seemed, further, most probable that they were the means of effecting 

 the pollination of the plant. I may mention that no other insect was 

 observed visiting the flowers, and that Synagris emarginata (?) seemed 

 to prefer C. canescens although, I believe, it occasionally also visited 

 (after exhausting all flowers of the Crassula) flowers of mignonette, 5 " 

 which were growing in profusion close by. The five egg-shaped epipe- 

 talous knobs of each flower are, as stated above, seated behind the por- 

 tions of the petals which appear to be their true apices, and when they 

 have made their appearance it looks as if the flower had opened. 

 However, as a matter of fact, the flower never opens. To explain this 

 more clearly I must give a short description of a petal, which will, more- 

 over, help to explain some of the remarks which I have to make further 

 on. It can easily be seen that each petal is thickest along the median 

 line (Plate I, Fig. 2 A), and that it flattens out towards both its 

 sides — one might almost say that it has a kind of midrib with broad 

 wings. In making -a median longitudinal section (Plate I, Fig. 2 C) 

 through a petal it will be observed that the "epipetalous knob" 

 forms the natural continuation of this midrib, and thus forms the 

 true apex of the petal. The small, thin projection of the petal at the 

 base of the epipetalous knob, which has hitherto been considered the 

 apex of the petal, has, therefore, probably to be considered of the 

 nature of a ligula,t but as I have not studied carefully the develop- 

 ment of the petals I do not wish to insist too strongly on this 

 interpretation. What concerns us here are two points — namely (1) 

 each epipetalous knob is firmly connected with the thickest portion 

 of the petal, and thus a lateral pressure on it, if sufficiently strong, 

 will move the whole petal, and (2) the projecting thin portions of the 



* I have since observed the same insect sucking nectar from the flowers of a 

 Euphorbia at a time (May, '97) when no species of Crassula § Globulea was in 

 flower. 



f Similar ligular structures are found in Boraginacetc where they frequently 

 have the same function as in Crassula % Globulea. 



