386 MR. С. Е. М. SWYNNERTON ON 
and saw them treat the Cannas as already described, but always enter the 
aloes properly. I also examined all the flowers both of this aloe and of the 
Cannas on not less than eight separate oceasions during the month, each time 
just after the birds had visited them. On no occasion did I discover a 
damaged aloe blossom : on every occasion a greater or smaller proportion of 
Canna flowers had been pierced. The aloe flowers were not always con- 
never, I should say, very conveniently, for the bird had 
veniently situated 
at all times to lean out to reach them—but their very thin and very pliable 
pedicels made them so unstable that a sunbird would, I think, have found it 
very hard indeed to pierce them, however possible it might have been to bite 
a piece out of them. At the same time this pliability made it easy for the 
sunbird to pull them round to itself after inserting the point of its long, 
eurved bill in the mouth of the perianth. 
D. OBSERVATIONS ON (GARDENIA TIGRINA, Welw., and 
[)RACÆNA FRAGRANS, Gawl. 
I have on several occasions watched the bases of flowers of Dracæena 
fragrans and Gardenia tigrina being pierced in exactly the same way. The 
Dracwna is one of the commonest shrubs in the Chirinda Forest, and the culprit 
has been in each case Cinnyris olivacina (Peters) Gadow. The Gardenia, 
also common, is one of our larger forest-trees, but iny direct observations on 
it in this particular connection were carried out almost entirely on a young 
tree, 11 feet in height, that stands on my lawn. I have seen the bases of its 
corolla-tubes punctured by both Cinnyris chalybwus, Shelley, and Cinnyris 
venustus (Shaw) Shelley, var. niass, Reichw., few, if any, of the flowers on 
the tree escaping, and I do not now remember having ever seen them entered 
by the natural opening. Indirect evidence was afforded by similarly pierced 
fallen corollas that I have sometimes picked up in the forest. I have failed 
to find evidence of unsuccessful attempts to pierce. 
In the Gardenia, improperly entered by all individuals, the incentive to 
the robbery was probably to be found in the fact that the natural opening of 
the flower was high and inverted and, unless the bird should hover, was 
much harder to reach than the base was to pierce. The large, rather tiger- 
lily-like flowers are perched in a vertical position at intervals along the upper 
surface of the more or less horizontal twigs, and it must be quite difficult for 
a small sunbird to reach over and down to the honey-glands. ТЫ expla- 
nation will hardly by itself aecount adequately for some of the Canna 
instances. 
It seemed in any case unlikely that these attacks would tend to any great 
extent to hinder pollination in the Dracena and Gardenia: both are markedly 
entomophilous, and the punctures I have seen the birds make in their flowers 
probably constituted no great injury to them, seeing that they were usually 
