( 155 ) 



black. This crest is broadest at one extremit}' of the seed ; it gradually 

 gets iianower towards the opposite end and ceases completely about 

 half way. The crest is composed of thin walled cells containing oil 

 and protein; it is clearly shown in the drawing of the seed of 

 D. Eafj1esi(inn published by Schimper (PI. 6 fig. 6). 



When walking on a quiet sunny day under the Mangif era's, 

 which are sometimes completely draped with the grey strands of 

 D. numinularia, one sees, when the seeds are ripe, the white pappus 

 floating in the currents of air. If the seeds come into contact with 

 a tree, tlie^^ attach themselves more or less firmly, but can also be 

 carried on again. At first sight one would therefore imagine, that 

 these plants are disseminated by the wind alone. In our own garden 

 and in that of the High School, in which a large number of trees 

 are grown, we have carried on germination experiments with the 

 seeds. They were stuck by means of a little water to the trunks 

 of various trees and were regularly watered during the first few 

 days. The seeds germinate extremely rapidly ; the seedling attaches 

 itself by the lower end of the hypocotyl and after a few days the 

 cotyledons already appear. (Later we hope to describe the germination 

 in detail). If the seedlings are now left on the trees without being 

 taken case of, i.e. if they are not watered regularly, they soon begin 

 to languish, although they are able to support much drought. The 

 cotyledons keep crumpling up and become again turgid after a 

 shower of rain. Although the past monsoon was very damp, not a 

 single one of all the seeds, which were sown, survived. They all 

 died off after a few weeks of drought. 



ScHiMPER writes that the seeds themselves may get down deep 

 into the fissures of the trunk, but we have never observed anything 

 of the kind ; on the contrary the hairs of the pappus completely 

 prevent the penetration of the seed itself into the fissures of the 

 bark. It is further known that these plants, and especially D. Raf- 

 fiesiana prefer to grow on trees with slight foliage or high up in 

 the crown, so that they are exposed to the intense rays of the sun. 

 It seemed to us worth while therefore to investigate in what manner 

 the distribution and germination of the seeds really takes place. 



After many abortive journeys through the habitats known to us, 

 we saw the first young seedlings appear on the trees at the begin- 

 ning of the wet season. They all, without exception, came up from 

 deep down out of the fissures in the bark. Moreover we now saw 

 them arise in all sorts of places out of the channels of the Jrido- 

 mi/rmex-SLnt, already well known to us. The seedlings were found 

 most numerous in those places where the ants build their very 



