great variability in a practically exclusive association 
with man, it seems entirely likely that much of the ob- 
servable variation in the group has resulted from man’s 
preferential maintenance of conspicuously different 
forms. The size of the flowers, among the largest of all 
flowering plants, facilitates, both for the native and for 
the botanist, the detection of differences that often go 
unnoticed in flowers of smaller size. Because of the 
paucity of seedlings, implying ineffective sexual repro- 
duction, we cannot yet confidently infer whether the 
numerous cultivars have arisen primarily through hy- 
bridization and gene recombination or through the ap- 
pearance of mutations as bud sports or chimeras. In any 
case, most of our attempts to delimit species amidst 
these circumstances have been unwarranted. 
The characters used by Safford (11) to distinguish 
species are not always unique to the entities that he at- 
tempts to separate. For example, in studying a large 
population of D. candida in the Valley of Sibundoy in 
Andean southern Colombia, I noted the presence of both 
spathe-like and several-pointed calyces (as in D. sawveo- 
lens), of rounded, as well as emarginate, corolla margins 
(asin D. arborea and D. cornigera) and of corollas vary- 
ing from under 20 to over 82 cms. in length. Seeds of 
four of the five species illustrated by Lagerheim (9), 
though apparently distinctive, can be seen in a single 
fruit of D. candida. 
It is hoped that controlled hybridization experiments 
and more intensive observation of natural populations 
will eventually contribute to a better understanding of 
the great variability in the tree Daturas. 
