Comparative Evolution of Biomes 323 



Festuca, Arabis, Draba, Cerastiiim, and Pedicidaris. It is virtually 

 certain that, since these genera evolved, tundra conditions have 

 not extended in a continuous belt across the equator. It is therefore 

 equally certain that in each region these plants have moved inde- 

 pendently from adjacent communities into tundra areas. The genera 

 mentioned above and the great bulk of other tundra genera contain 

 many species inhabiting stream bank or subclimax communities of 

 temperate or boreal biomes. The present tundra would therefore 

 appear to be populated by subclimax and stream bank species 

 arising in lower latitudes or elevations. It is reasonable to suppose 

 that these colonizing species were those with the wider ecological 

 tolerances and therefore were the most widely distributed both 

 ecologically and geographically. 



The tundra biome may therefore be regarded as a former sub- 

 climax aggregation which has become a series of climax commu- 

 nities by the independent colonization of distant areas. The remark- 

 able taxonomic similarity of distant regions of this biome is not 

 due to a joining of regions of the biome with a subsequent mixing 

 of biotas. It is due to the initial widespread nature of the species 

 which did the colonizing. 



The next theoretical stage in biome phylogeny would be the 

 evolution of climax species especially adapted to the ecological 

 situation of the biome and restricted to it. Certain desert biomes 

 seem to represent this stage. 



The cactus desert biome of southwestern North America and 

 certain areas such as Jamaica in the dry parts of the American 

 tropics (Asprey and Robbins, 1953) are physically and visually 

 similar to certain African deserts which are dominated by cactus- 

 like Euphorbiaceae. The vegetation of both associations produces 

 an overlay of similar ecological conditions and thus may be clas- 

 sified as the same biome. Taxonomically the situation is completely 

 the opposite. The Cactaceae evolved in the western hemisphere 

 and, with the exception of some species of Rhipsalis in Africa and 

 Ceylon which are probably introduced (Lawrence, 1951), are 

 rigidly confined to the Americas. The Euphorbiaceae is a large 

 cosmopolitan family in which cactus-like forms evolved in the east- 

 ern hemisphere and have apparently never dispersed from it. Each 

 of these cactus-like associations therefore arose independent!}' from 

 different ancestral types, and the resulting communities owe their 

 ecological similarity to the convergent evolution of dominants 

 rather than to phylogenetic affinities. In other words, this biome 

 is polyph)letic in origin. 



