Ki;kk — Dischidia rafflesiana (tnrl Dischidia nummularia. 299 



corollas. It is bighly probable that pollination in this species is effected 

 by a bee, as is the case in B. nummuhria. 



D. nummularia has white flowers somewhat smaller than those of 

 D. rafflesiana; when the flowers open, the tips of the petals turn back, 

 affording a relatively large entrance to the interior of the corolla ; the 

 structure of tlie staminal column and style is similar to that of D. rafflesiana. 

 The flowers are freely visited by the Iridorayrmex living beneath the leaves 

 and by other ants, which, in this case, can readily enter the corolla. The circle 

 of hairs at the entrance seems to be no obstacle even to very minute ants; but I 

 have never seen an ant removing pollinia. The flowers are also visited by two 

 small bees ; one of these, AUodape, sp., is found at the flowers about mid-day : 

 of five bees caught on the flowers not one carried pollinia ; the other species, 

 Nomia, sp., appears from about 7 a.m. to 8 a.m. The only two of this bee I 

 was able to catch both had pollinia attaclied to their proboscides. The Nomia 

 is probably the chief pollinator of these flowers. Possibly an examination 

 of a larger series of the Allodape would have sliown some witli pollinia. 



Dispersal of Seeds. 



The seeds, like those of other Asclepiadacese, have a well-developed coma 

 of silky hairs, and appear well adapted for dispersal by wind ; but tliere is 

 evidence to show that tliey are commonly dispersed by another method. 



I had always found tlie young seedlings of botli J), rafflesiana and 

 D. nummularia coming up either in tlie runs or nests of Iridomyrmex, and 

 also had often seen a follicle, half or quite dehisced, still containing tlie silky 

 hairs but no seeds ; from these two observations I inferred that the 

 Iridomyrmex removed the seeds. I watched for some time before I was 

 able to prove this, as it is not easy to time the exact moment of dehiscence 

 of a follicle ; at last, however, I was fortunate enough on two occasions to see 

 the ants removing the seeds. On the first of these occasions I noticed a 

 number of Iridomyrmex round a follicle of D. nummularia whioli was just 

 commencing to dehisce ; when first seen it had a small slit at its base for 

 about one-fourth of its length. As this widened sufficiently the ants seized 

 bunches of protruding hairs in their mandibles, and began to pull on tliem, 

 doing so by walking round the follicle away from the slit on either side, by 

 this means accelerating dehiscence ; when the opening extended nearly the 

 whole length of the follicle, a seed was removed, with the hairs still attached, 

 and carried into the nest. The whole process, from the time the small slit 

 was seen till the fi.rst seed was carried into the nest, occupied about half an 

 hour, On the other occasion, also in D. nummularia, the follicle was already 



