CHAP. Lxxx TRUE SAVANNAH 297 



Lichens, mosses, and algae are entirely absent from the soil, and, at 

 , the most, are scantily represented on stones and trees. 

 I The leaves of the dicotylous plants show xerophily, particularly in 

 ^ their stiffness (for they are often so rigid and dry as to rattle in the wind), 

 ' in their lie, often in small size (though many are broad, oblong, or obovate, 

 I or compound ; cricoid and pinoid types being almost unrepresented), 

 ;aid in their very hairy nature or, if glabrous, in their waxy or lacquered 

 ' oatings. 



Ethereal oils are present in a whole series of the plants, and, in South 

 America, especially in Verbenaceae, Labiatae, and Myrtaceae. 



Many trees in tropical savannah are leafless during the dry season, 



or shed their foliage during this period. Yet this is the time at which 



many species are in flower. In this respect the Brazihan campos are not 



typical, but rather form a transition towards subtropical savannah, as 



; the leafless season is very short .^ This arises from the circumstance that 



I winter on the Brazihan upland is not so hot as in tropical lowlands, for 



j which reason there is less transpiration ; in addition the dry season in 



the campos is not without a scanty supply of rain. 



The physiognomy of savannah is subject to great variety, which 

 I depends partly upon the height of the vegetation and partly upon the 

 'share taken on the one hand by grasses and perennial herbs, and on the 

 other by shrubs and trees. 



In Brazil and Venezuela there are some savannahs campos in which 

 [trees, rising above ground that is clothed with vegetation varying from 

 [half a metre to a metre in height, are so close together as to produce a kind 

 jof forest, which is open, sunny, shadeless, and hot, and allows the pedes- 

 trian to walk freely and the horseman to ride in all directions : such are 

 the Brazilian campos serrados. There are other savannahs in which trees 

 are extremely scanty and low, or arc entirely wanting, and the carpet 

 of grass and perennial herbs is very short and almost sward-like. In 

 Matto Grosso, in the interior of Brazil, according to Pilger - the dry 

 jseason is not entirely rainless. Most of the trees are deciduous and 

 'blossom after defohation. Along the streamlets there extends forest 

 '.composed of Mauritia vinifera and evergreen trees and shrubs, with 

 iperennial herbs, sedges, and grasses carpeting the ground. 



Llanos. 



Belonging to a special type of savannah are the llanos, which form 

 Doundless plains to Venezuela and have been so admirably described by 

 Humboldt.^ In the llattos trees are very few in number ; indeed they 

 may be entirely wanting save in the moistest places, where palms, including 

 Mauritia flexuosa and a species of Corypha, together with other plants, 

 ^ive rise to forest, which does not itself really belong to savannah ; in 

 :>ther places isolated trees of the protcaceous Rhopala or other species 

 3ccur. Grasses form a vegetable covering, often as tall as man, and with 

 them are mingled Compositae, Leguminosae, Labiatae, and so forth, 

 ^arge portions of the llanos are under water during the rainy season, 



' Warming, 1892. ' Pilger, 1902. 



* In addition to Humboldt's works, that of C. Sachs should be consulted. 



i| 



