Observations on South African Plants. 35 



petals close the entrance to the flower .completely, and no access to 

 it can be gained unless the petals are pushed apart and they them- 

 selves bent. Now it is by no means easy to push the petals apart, 

 for we find that they are closely interlocked as they stand very close 

 together, and are, moreover, imbricate (Plate I, Fig. 2 H pj, but in 

 addition to this they are closely surrounded by the comparatively thick 

 sepals which have to be pushed aside at the same time. We can under- 

 stand now that the entrance to the flowers is completely barred to 

 small insects and all insects wmich have only soft mouth-parts. But 

 a middle-sized hymenopterous insect, such as Synagris emarginata, 

 with its wedge-shaped, hard mouth-parts, has no difficulty in gaining 

 access to the flower. To these wedge-shaped, hard mouth-parts the 

 epipetalous knobs form excellent points of resistance when the insect 

 wants to prise a flower open and suck the nectar contained in it by 

 means of its soft mouth-parts enclosed in the former. This is, as far 

 as I can make out, their whole purpose in the economy of the 

 flower. 



The sepals and petals together form an elastic structure, so that 

 the flower closes again immediately the insect has left it, and the 

 same process has to be gone through again if another insect wants 

 to gain access to the nectar. 



Turning now to the sexual organs of the flower, we find that both 

 anthers and stigmata are in the line which the mouth organs of the 

 insect have to traverse'to reach the nectar-producing glands (Plate I, 

 Fig. 2 G gl). When examining the stigmata of a flower which 

 shows the epipetalous knobs boldly, but in which the anthers have 

 not opened yet, we frequently find germinating pollen-grains on them. 

 Thus the stigmata are ready to receive pollen before the anthers of 

 the same flower have discharged theirs. The flower is pro tog y nous. 

 Pollen germinating on the stigmata at this stage must have been 

 brought from some other flower by the visiting insects, and these 

 insects, therefore/ effect cross-fertilisation. When the anthers open 

 the pollen remains attached to them, but from their position it can 

 be inferred that visiting insects cannot fail to brush some of it on 

 their heads and carry it to the stigmata of the next flower which they 

 visit. The stigmata remain receptive for a considerable period, and 

 if the pollen of a flower is not removed part of it is deposited on the 

 stigmata of the same flower. In flowers which were protected from 

 insect visitors by a covering of muslin I found repeatedly large 

 numbers of germinating pollen grains on the stigmata, and although 

 I did not carry the experiment so far as to see whether these flowers 

 would produce ripe seed, I think it is not far-fetched to conclude from 

 the fact just mentioned that self-fertilisation is also possible. 



