Negro and Vaupés than in any other area. Here are rep- 
resented not only the greatest number of species of these 
two widespread groups, but endemism indicates that 
both Hevea and Micrandra may represent very ancient 
elements of the flora of this phytogeographically critical 
region. 
The genus Joannesia is not represented, so far as we 
know, in this area. Its distribution is rather local: one 
species occurs in the coastal region of Brazil near Rio de 
Janeiro: asecond is known from the Rio Tapajéz in the 
eastern Amazon. 
The flora of the caatingas and caatinga-forests in the 
upper Rio Negro-Vaupés basin shows very close affinities 
with that of the ancient Venezuela-Guiana land-mass. 
In fact, these caatingas (locally called ‘‘savanetas’’ in 
Colombia) are found in eastern Colombia either on the 
isolated quarzitic mountains of Cretaceous age (or on 
extensive sandy stretches associated with such moun- 
tains) or on granitic hills of the pre-Cambrian ‘‘ Brazilian 
shield.’” The very numerous rapids in the rivers of the 
Colombian Comisaria del Vaupés are likewise associated 
with these quartzitic rocky outcrops or, along the Brazil- 
ian boundary, with granitic outcrops. Vaupesia catarac- 
farum has been collected from caatinga-forests along both 
the quartzitic and the granitic cataracts: the type local- 
ity (the Falls of Jirijirimo on the Rio Apaporis) and the 
base of Cerro Isibukuri, the western and southernmost 
stations known for the genus are quartzitic; the Falls of 
Tapira on the Rio Vaupés and at Teresita on the Rio 
Papuri are granitic. Both of these geological associations 
support peculiar and, presumably, ancient floras and seem 
to be repositories for endemic and apparently primitive 
species. We might enumerate many, but it will perhaps 
suffice to name several allies of the new genus: Hevea 
rigidifolia (Spruce ex Benth.) Muell.-Arg., Micrandra 
[ B4 | 
