152 THE POLLINATION OF SATYR1UM BICALLOSUM, THUNB. 
being warm, but dull and cloudy. Later, on the same day, at 2.30 p.m., I saw 
another fly running up the spike of a Satyrium, amongst the bracts and the 
bases of the ovaries. Presently, it ascended a flower, and entered the left hand 
opening. The fly was evidently exploring the spur, and, as it stood in the 
opening, it placed itself at an angle of about 45° to the horizontal. This tilted 
position is exactly that required to bring the base of the wing into contact 
with the viscid disc. After a few seconds, the fly withdrew, bearing a pollinium 
attached to the thorax. I have removed numerous pollinia artificially, but 
have never observed any subsequent curvature or movement of the caudicle. 
Indeed, owing to the position of the stigmatic surface immediately above 
the horizontally placed anther lobe, scarcely any subsequent movement of the 
caudicle is needed. When a fly bearing a pollinium, say on its right shoulder, 
enters the right half of a flower, the pollen mass will almost inevitably come 
into contact with the stigmatic surface. It seems probable that the curved 
lobes which over-arch, and are in intimate contact with the stigmatic surfaces, 
will help in this matter, as, when a fly bearing a pollinium on the correct side 
withdraws from the passage, the hooked process will retard the pollen mass, 
and cause it to strike the stigmatic surface. Self-pollination is avoided, as the 
pollinium is not withdrawn from the pollen sac until the fly is leaving the 
flower. 
As regards the food substance which attracts the fly, I have made the 
following observations. No honey is secreted at any time within the spur, 
and an examination of the internal surface always showed this to be quite dry. 
The base and sides of the spur are studded with numerous papillate, or shortly 
clavate hairs (Fig. 6), and these unicellular hairs contain abundant protoplasm 
and cell-sap (Figs. 8 and 9). In some cases, the hairs are elevated upon small 
humps of tissue (Fig. 8). Small drops of a highly refractive, oily substance are 
to be observed on the outside of each hair. The flies of the genus Sciara have 
a comparatively short proboscis, consisting of two muscular pads, adapted for 
sucking up liquids; they have no hard mouth parts with which to bite or 
perforate. It does not seem probable that the insects bite or puncture the 
clavate hairs in order to suck out the cell sap ; they are more probably attracted 
by, and lick up, the oily extra-cellular secretion on the surfaces of the hairs. 
This is a matter somewhat difficult to investigate, as it is impossible to observe 
the fly when actually at work inside the flower. I did not determine the exact 
origin of the oily secretion, and the question of the attractive food substance, 
and the mode of obtaining it is still somewhat problematical. 
For purposes of comparison, I studied the structure of the spur of Satyrium 
coriifolium, Sw., a species in which, as is well known, honey is secreted abun- 
dantly. I found the interior surface of the spur in newly opened flowers, to 
be quite dry, but covered with elongated, juicy hairs (Fig. 5). The hairs are 
longer towards the opening of the spur. In a slightly older flower, the epidermis 
