911 
channel tor some distance beyond the point where the duct 
stops. Like the duct itself, this conducting channel is sur- 
rounded by sacchariferous tissue. A similar functional provision, 
which also occurs in some other Agaves, has been indicated 
by Bronenrart for Agave yuccaefolia as a closed septal stylar 
channel debouching above the floor of the flower. In the swollen 
base of the style of Furcraea the three openly triangular septal 
sulci are separated by broad and nearly flat carpellary ridges 
against which the outer filaments are closely applied. The 
walls of the sulci are longitudinally keeled, so that each channel 
has a rather wide median groove and two narrower lateral 
grooves above. Opposed to the sulcus is one of the inner 
whorl of filaments, its ventral face keeled above. The apposition 
of the greatly thickened and closely crowded filaments to the 
stylar column thus closes access to the top of the ovary except 
by way of three channels, each leading directly to the open 
mouth of a nectar duct, and each incompletely forked above 
through the ridging of style and filament. The ecological 
result of this structure is to permit access to each nectar 
outlet rather more easily by way of two free lateral grooves 
than directly on the septal radius, thus ensuring two or eyen 
three trials, with correspondingly multiplied contact with 
stigma or anther, during the exploration of each nectary. 
Though less brilliantly colored than most flowers of their 
size, the flowers of Furcraea are marked as entomophilous, 
like those of Agave, in such deviation from green coloring as 
they show and in the production, even though limited, of 
nectar and its advertisement by means of odor. The long open 
duct, about half a millimeter wide and some six mm. long, 
With its length about doubled in effect by the stylar-staminal 
channel, points to pollination by insects with suctorial mouth 
parts long and slender enough to permit the duct itself to 
be probed in the search for nectar. Field observations on the 
floral visitors of Furcraea have not been recorded, but it 1s 
not improbable that these will be found to include moths 
and large long-tongued bees, — both known to be attracted . 
