CHAP, xciii EVERGREEN DICOTYLOUS FOREST 343 



tree-crowns. But on looking down upon forest from some eminence, one 

 sees dotted over it large yellow, white, violet, or red spots, which indicate 

 l)lossoming trees and lianes. In many cases, for instance among Laura- 

 ceae and the majority of Papilionaceae, the flowers are small but rendered 

 easily visible to insects by their great abundance. In many species it is 

 remarkable that the flowers spring from thick trunks or branches, or even 

 from the base of the bole, because year after year they develop from the 

 same dormant buds. The most widely known example of such cauliflorous 

 species is Theobroma Cacao, the chocolate tree ; other examples are 

 supplied by Myrtaceae, Sapotaceae, Leguminosae, Ficus Roxburghii, 

 Crescentia Cujete, and species of Swartzia.^ Wallace suggests that the 

 flowers of these cauliflorous species may be adapted for pollination by 

 Lepidoptera, which flutter within the calm forest. Whether or no this 

 is the case has not been determined. But judging by floral construction 

 it does not seem to be true of Theobroma, whose flowers are probably 

 pollinated by other kinds of insects or are self-pollinated. 



Periodicity. In tropical rain-forest there is generally neither summer 

 nor winter, neither spring nor autumn : the periodic habit of development 

 so distinct in other plant-communities is lacking or very feebly represented. 

 Some species acquire new foliage throughout the year : and even if certain 

 species do exhibit a distinct resting period, or are entirely leafless for 

 a short time, they are lost among the crowds of others that show either 

 no period of rest, or one at a different season of the year. Probably nearly 

 all the species have their own definite time of flowering ; but this is not 

 synchronous for the different species. The forest (like South American 

 savannah) is rich in flowers all through the year. Thus rain-forest shows 

 no periodicity in its life as a whole.- 



Leaves. Foliage-leaves in tropical rain-forest nearly always remain 

 attached to the tree for longer than one year, generally for thirteen or 

 fourteen months.- They probably often remain active for many months, 

 perhaps for more than a year a fact that is of profound oecological 

 importance to the plants, and explains the huge dimensions of these as 

 well as the large production of organic matter. The old leaves bend, 

 sometimes, according to Haberlandt,^ by active movements, in order to 

 provide space for the younger foliage. The play of colour of the foliage 

 has already been discussed on page ^^j. 



Leaf-shape. In tropical rain-forest leaves assume an extraordinary 

 number of shapes. On trees we find not only the forms familiar to us in 

 Europe namely, ovate, elliptical, and so forth, simple or simply com- 

 pound but also a number of new forms ; for example, the pinnate or 

 palmate fohage of palms ; the large, undivided, characteristically veined 

 leaves of Scitamineae; the pinnately compound leaves of Leguminosae, 

 and particularly the bipinnate or tripinnate mimosaceous leaves, whose 

 countless leaflets execute phototactic movements ; the digitate leaves of 

 Bombaceae and of the araliaceous Panax ; the palmatcly divided peltate 

 leaf of Cecropia ; the long-stalked, large, cordate or ovate-cordate leaves 

 of Araceae ; and the bamboo-leaves, which dispose themselves at the tip 

 of the branch in a digitate fashion ; and so forth. Yet the commonest 



' Wallace, 1891 ; Haberlandt, 1893 ; VVhitford, 1906. 



' See Warming, 1892; Holtermann, 1902; \"()lkens, ICXD3. ^ Haberlandt. 1893. 



