rounded. Counts of aborted pollen in six plants ranged 
from 20.7 to 80.4 per cent. 
D. candida ‘Buyés’ is searcely distinguishable from 
many examples of D. candida in widely separated areas 
of tropical America. In the characters studied, its range 
of variation is less than that of the species as a whole. 
The rare Sibundoy cultivars of this species have never 
been found outside of the Valley of Sibundoy, and it is 
my assumption that they are all derived, at least in part, 
from D. candida ‘Buyés’.* 
Chromosome nu mbers 
Heiser (16) found the chromosome number of JD. can- 
dida as represented by the collection Heiser 6119 (IND), 
to be n=12, and Barclay (2,3) found 2n=24 in D. can- 
dida ‘Culebra’ (Barclay & Schultes 286) and in D. vul- 
canicola A.S. Barclay. 
Preparation of aceto-carmine stained microspore 
mother cells permitted observation of the chromosomes 
after 48-72 hours. The chromosome counts for each cul- 
tivar were made on either one or two plants as indicated 
by the voucher collection numbers in Table II. Seven of 
the nine D. candida cultivars were examined, and all 
were seen to havea chromosome complement of n=12. 
All counts were made after metaphase I, most at or 
following metaphase II. The data are summarized in 
Table IL. Figure la, illustrates the chromosomes of 
D. candida ‘Culebra’. 
Twelve is the haploid chromosome number of Datura 
sanguinea R.& P. (7, 16), of *‘D. suaveolens’’ (15), and 
of all the herbaceous species of Datura (1). 
‘TI recently encountered a tree referable to D. candida ‘Amar6n’ 
near Las Cruces in southernmost Narifo. Thus the question is raised 
as to whether this was brought from the Valley of Sibundoy, or 
whether it is part of an early population of plants from which one or 
more of the Sibundoy cultivars might have originated. 
[ 170 | 
