strikingly similar to several of the species of Micrandra 
with which it grows that even the Indians of the region 
use the same name for Vaupesia eataractarum and tor 
Micrandra Rossiana R. i. Schultes and MW. Sprucecana 
(Baill) R. ke. Schultes. Close examination, however, dis- 
closes the extraordinary distinctness of the new concept. 
All of the characters of Vaupesia place it in the Mi- 
crandra- Hevca-Joannesia complex. The leaves are very 
definitely micrandroid, departing sharply from the three- 
foliolate leaves of FZevea and the usually tive-foliolate 
ones of Joannesia. The capsule, valves and seeds of 
Vaupesia likewise resemble those of Micrandra very 
strongly. Indeed, if I had collected nothing but fruiting 
material, | would, without hesitation, have described it 
as anew species of Wicrandra. 
The type of inflorescence and the petaloid flowers re- 
semble those of Joannesia to a remarkable degree, even 
though the calyx is quite micrandroid and entirely unlike 
the undivided calyx of Joannesia. The presence of a 
watery red resinous liquid in the bark of many trees of 
Vaupesia recalls a similar condition in Joannesia heve- 
vides Ducke. 
The whorled disposition of the anthers on a fleshy 
staminal column definitely indicates a relationship with 
Hevea. Nevertheless, the morphology of the column 
might be interpreted as rather intermediate between 
Joannesia and Hevea. We are able to point out 2 com- 
plete series, tending towards fusion of the filaments, from 
Micrandra (with tree filaments) and Joannesia (with the 
filaments basally connate) through Maupesia (with the 
filaments totally fused into a column which, however, 
shows by fluting the position of the individual filaments) 
to Hevea (with a perfectly terete column). 
A comparison of the disks of both the pistillate and 
staminate flowers suggests, perhaps, a closer alliance with 
