258 Origin of Biomes and Succession 



draw any sharp line at points where one community ended and 

 another began. Each major type would merge gradually into the 

 next. These types would be definable as associations comparable 

 to those existing side by side today. 



These transects through time might not correspond to actual 

 borings "through time" made in pollen profiles or well drillings. 

 The latter may cut through sections of many different biomes, such 

 as marine versus terrestrial types or tundra versus deciduous forest, 

 because the ranges of the biomes moved in unison with their 

 necessary ecological conditions. 



Although having a short life in relation to that of the biome, 

 individual sets of communities or associations may exist for many 

 millions of years. In his review of the fossil flora of the West Indies, 

 Hollick ( 1924 ) showed that the generic elements of the early and 

 middle Cenozoic floras are identical with those of the present. 

 Furthermore, in genus after genus the Cenozoic species appear to 

 be identical with or remarkably similar to existing Neotropical 

 species occurring in the West Indies and adjacent Central and 

 South America. Hollick concluded that the present northern Neo- 

 tropical evergeen rain forest has continued with only minor changes 

 in species composition since early in the Cenozoic. 



These records show that, since land floras became well estab- 

 lished, large evergeen trees of some sort have dominated uninter- 

 ruptedly tropical land areas of low elevation having abundant avail- 

 able moisture throughout the year. In an ecological sense conditions 

 beneath this tree canopy have therefore changed little if at all 

 since mid-Paleozoic. To subdominant organisms living in the under- 

 story portion of the community it would not matter whether tree 

 ferns, cycads, conifers, or angiosperms were the agents producing 

 the shade. In the present-day Brazilian evergreen rain forest this 

 principle is well illustrated by the bromeliad epiphytes and by the 

 mosquitoes living in the water which collects in bromeliad leaf 

 axils. Veloso et al. (1956) showed that each species of bromeliads 

 is dependent not on specific tree species but on physical conditions 

 of light and humidity, which in turn are determined by distance 

 from the ground. Production and distribution of bromeliad-inhabit- 

 ing mosquito species are not related to the species of bromeliads 

 but to the water-holding capacity of the bromeliads, the level of 

 bromeliad occurrence, and density of suitable bromeliads in a unit 

 area of forest. Thus two levels of subdominants, the bromeliads 

 and the mosquitoes, are dependent, not on the taxonomic composi- 

 tion of the forest, but on the physical conditions superimposed on 



