262 Scott- Elliot. — On the Fertilisation of Musa , 
The breadth across the broadest part of the arrowhead is 
about ten lines. The odd petal is small, three quarters of an 
inch long and dome-shaped, completely covering the entrance 
to the honey. 
The anthers occupy the whole length of the broad part of 
the arrowhead (some 2| inches), while the style projects 
almost an inch in front of the petals. If the flanges are 
pressed down, the tubular cavity opens and exposes the six 
anthers. 
The superior edges of the united petals have a peculiar structure. In the 
broadened portion they are only about a line broad (though overlapping), but 
where the narrow part of the arrow-head begins they suddenly expand to about 
half an inch, at the same time overlapping in the peculiarly perfect way shown in 
Fig. 30. Insects must be completely prevented therefore from entering the flower 
from above. 
Professor Maccowan informs me that the bird walks along 
the flanges (probably developed for this purpose) and in 
bending down beneath the dome-shaped odd petal probably 
causes the petaline sheath to open so as to dust its breast 
with pollen. The stigma being in front of the petals will of 
course be touched first. 
I have often seen the honey-bee and Diptera (. Lucilia 
argyrocephala and others) sucking the gummy juice which 
exudes from the spathe, but it is not I think possible for 
insects to produce fertilisation as Hildebrand 1 supposes may 
happen, unless very exceptionally indeed. Unfortunately I 
could not visit its proper habitat, and, in spite of several days 
watching, I never saw any birds near it in the Cape Town 
Gardens. 
1 Hildebrand, loc. cit. 
