CliAi'. IV] TROPICAL DISTRICTS CONSTANTLY MOIST 389 



stance that frequently natives can correctly name individual species of 

 trees b\- mcrtly observing the characters that can be seen from the ground, 

 such as the nature of the bark, the thickness of the trunk, the occasional 

 plank-buttresses, which will be subsequently dealt with, and, at times, the 

 mode of branching. Analytical tables founded upon such characters, 

 which however are often very difficult to describe, would be of immense 

 value. 



Most lianes — the woody climbing plants which usually occur in such 

 abundance in every virgin tropical forest — also withdraw their crown of 

 foliage from the view of the naturalist, who is confined to the ground, and 

 an attempt to pull them down is as a rule as useless as to fell them. But 

 in their case the anatomical structure of the stem exhibits so much that is 

 characteristic, that the determination, at least of the genus, is possible, 

 thanks to the excellent works of Radlkofer, Bureau, and Schenck '■. 



In order to obtain an idea of the systematic composition of the forest- 

 canopy, the botanist, even if he has frequentl)' tra\-ersed and carefully 

 searched the virgin forest, must rely on ' Floras,' which, prepared chiefly 

 from collections made b}' natives, usually convey very incomplete informa- 

 tion. Much more useful are the works of the foresters, but these un- 

 fortunatel}- arc still entirely wanting in respect to tropical America 

 and troi^ical Africa. The works of Brandis, Kurz, and Koorders and 

 Valeton are sources of information regarding the systematic composition 

 of the forests of India and further India. Apart from the lianes and 

 herbaceous flora, it is comparatively easy to acquire the necessary know- 

 ledge of the systematic composition of the underwood, without which 

 a work on the physiology of the virgin forest is valueless. Not only 

 is the variety of foliage much greater than in Europe, but a number 

 of species bear flowers and fruit, though often in small quantities, for 

 months if not throughout the year. At the first glance, in a damp, cool 

 forest, the tree-ferns strike the e}-e, and these, like the ferns in general, can 

 be easily determined from the available herbaria. Small palms are seldom 

 absent ; for instance, in Brazil species of Geonoma occur, in Java species 

 dF Pinanga. Here and there appear thickets of bamboos, or of climbing 

 erns like Lygodium, or of species of Selaginella. The main mass of the 

 underwood and bushes, however, is formed by Dicotyledones. The Urti- 

 :aceae are seldom absent, for instance species of Boehmeria, and in Asia 

 also species of Laportea, which are easily recognizable, even when not 

 ilowering, by the form of their leaves and their hairs. In company with 

 these, shrubby Piperaceae will be found, in particular species of Artanthe, 

 md species of Ottonia with knotty stems and vertical, white, taper-like 

 spadices ; also the varied species of Croton, with inconspicuous flowers and 

 eaves with scales beneath. Less abundant, but highly characteristic, are 



' See p. 196. 



SCHIMPER Tj 



