270 GEOGRAPHY OF PLANTS, 



In general, the vegetation of the southern liemisphere is 

 very different from that of the northern, and there is a cer- 

 tain correspondence between tlie Floras of Southern Africa, 

 America, and New Holland. Most of the trees are woody, 

 with stiff leaves, blossoms sometimes magnificent, but fruit 

 of little flavour. In Southern Afiica, as well as in Nev/ 

 Holland, it is the form of the Proteae which prevails as if 

 appropriated to these regions. Instead of the South Ameri- 

 can Erica', we find the Epacrida of New Holland. LoheU^y 

 DiosmeeE, and a great number of rare forms of compound 

 blossoms, and of Uvihellata, are common to all these south- 



397. 



For understanding the growth and distribution of ])lants, 

 we must also attend to the soil. Similar plants are found in 

 similar soils, completely separated from each other, and with 

 respect to which no supposition of interchange can be enter- 

 tained, provided only the climate be not too different. Salt 

 soils produce almost every where particular Chenopodeae, 

 species of Chenopodium, A triplex, Salsola, Salicornia, and 

 Anabasis. Calcareous soils produce always the most nume- 

 rous and distinguished forms of plants. Volcanic mountains, 

 particularly basalt, produce few forms, but those of a distin- 

 guished kind and very variable. Alluvial mountains, parti- 

 cularly in the neighbourhood of streams, usually, like marshes, 

 produce forms that are always the same. The primitive 

 mountains, on the other hand, almost every where separate 

 the Floras of countries. Thus the Pyrenees, — the Alps 

 which divide Italy from France and Switzerland, — those 

 which separate Upper Italy from the Tyrol and Carindiia, — 

 and the Carpathian mountains, — divide the Floras of the 

 southern from those of the northern countries. 



It is hence of so much importance, along with Floras, to 

 describe the mountain rocks and the different soils, (258.) 

 The first example of a map of this kind was given by De 

 CandoUe, in the second volume of his Flore Fran^am, in 



