tion, Dr. Lyman B. Smith (in Contrib. Gray Herb. no. 
161 (1946) 81) has said: ‘‘....movement on the are 
was from the Andean region eastward. Finally most of 
the species [i.e. of certain bromeliads and cyperaceous 
plants he was discussing ] are narrow endemics, indicating 
that the movement was relatively ancient. ’”’ 
Rhytidanthera mellifera is an endemic of the cretaceous, 
quartzite hills of the upper Apaporis basin. We might 
well expect Rhytidanthera to occur in collections from 
other sandstone mountains in eastern Colombia and 
southern Venezuela, but even now we may venture to 
say that it is not acommon element of the flora of these 
repositories of endemics. On the Cerro Castillo, Rhyti- 
danthera mellifera is extremely rare. Special search for 
more abundant flowering material failed to furnish addi- 
tional individuals, so the species is at present known only 
from one bush. This bush, growing on the edge of a rock- 
ledge covered with a scrubby vegetation, was visible to 
us far below as we were climbing the face of the moun- 
tain. Like the other species of the genus, it makes an 
unusual display with the large, white caesalpinaceous-like 
flowers. Rhytidanthera was not seen on the nearby moun- 
tains, Chiribiquete and Campana. 
Rhytidanthera mellifera, so named because of the sweet 
and mucilaginous exudation of the flower which attracts 
swarms of wild bees to the plant, is immediately set apart 
from the other known species by having 64 stamens, a 
much larger number than has hitherto been found in the 
genus. 
The closest ally of Rhytidanthera mellifera would ap- 
pear to be BR. magnifica (Gleason) Dwyer, from which 
it differs in having twelve or thirteen alternate (instead 
of nine or ten opposite) leaflets ; a very broadly cuneate- 
lanceolate, obtuse (instead of a narrowly lanceolate- 
elliptic, acute) terminal leaflet; much smaller lateral 
[ 35 ] 
