POLLINATION OF THE MANGO. 15 
A large seedling tree of the No. 11 race, growing on the property 
of Prof. P. H. Rolfs, at Buena Vista, near Miami, was selected for 
the principal experiments. On this tree two panicles were bagged, 
preventing the access of insects, and left to determine whether any 
fruits would develop without the agency of insects in effecting 
pollination. The result was that no fruit developed. While the 
experiment was on a very limited scale, the No. 11 race is exceed- 
ingly productive and usually carries a large number of fruits on 
each panicle, so that a failure to carry any fruits is much more sig- 
nificant than it would be in such a variety as the Mulgoba. This 
experiment was conducted on several other trees, with the same 
result in every case. 
Two panicles on the same tree were bagged before any of the buds 
had commenced to open (as in the others), and as rapidly as the 
flowers appeared the stamens were removed before they had com- 
menced to dehisce. Emasculation was continued until the last 
flower had opened, and the bags were then left on until ample time 
had elapsed for the fruits to develop ; but all dropped off, indicating 
that pollination is necessary, even though the egg cell may not 
develop into an embryo, as appears to be the case from Belling's 
investigations. This experiment was tried in about 10 other in- 
stances, with the same result in each. In a few cases fruits developed 
parthenocarpically to the length of about 1 inch and then dropped. 
Two more panicles on this tree were bagged for hand pollination, 
and as rapidly as the flowers opened the stigmas were pollinated 
with pollen from flowers on the same tree. On one panicle 36 stig- 
mas were thus treated, and on the other 30 stigmas. The first panicle 
carried seven fruits and the second two. This is somewhat more 
than the average of other panicles on the tree, considering the fact 
that only a small portion of the panicle was allowed to develop, 
ull the laterals having been removed. 
Two more panicles were bagged for cross-pollination and the 
flowers emasculated as soon as they opened, so that no pollen was 
liberated within the bag. Pollen of the Sandersha variety from a 
near-by tree was applied to all fresh stigmas once each day, a total of 
83 being pollinated in the first instance and 66 in the second. Seven 
fruits were carried by the first panicle and six by the second, a con- 
siderably larger number than is usually produced by an equal number 
of flowers on a panicle left to be pollinated naturally. Probably all 
of these fruits would not be carried to maturity. When last exam- 
ined, they were one-half to 1 inch long. 
Cross-pollination gave equally good results with other varieties. 
It was tried on the Madras variety, with the result that 8 fruits 
developed from 58 pollinations, the pollen being taken from a tree 
of the Cambodiana mango. The Alfonso, when pollinated by this 
