6 2 
and color of the flowers perfectly preserved, on a table for several 
days ; but the flies completely ignored it. 
Now, while color and perfume may attract the flies from afar, 
surely something more substantial is required to retain them on 
the shrub, and induce them to visit flower after flower, day after 
day. Moreover, 1 have definite evidence that they do obtain 
some nourishment from the flowers. I have repeatedly caught 
and imprisoned a pollinator in a bottle. If a Marianthus blossom 
was enclosed with the fly, the latter lived for several days, indeed 
until killed ; but if shut up without anything it always died 
within twenty-four hours. Of course, plenty of air was supplied 
in all cases. 
Each petal is slightly saccate at the base, and just above its 
sides are directed inwards, making the cross-section U-shaped. 
The trough thus formed is pressed close against the stamens, so 
that a small tube is completed by the stamens closing up the 
front. We have, therefore, five apparently ideal nectaries at 
the base of the flower ; but 1 have often sought nectar therein 
and always jailed to find it. Yet it is into these empty nectaries 
that the flies thrust their probosces. Thinking that possibly the 
nectar dried or crystallised on the walls of the nectary as fast as 
secreted (this happens in some plants I have examined), I carefully 
washed out the nectaries of a number of flowers with distilled 
water and tested with Fehling Solution. I tried this experiment 
several times, but the result was always negative. The micro- 
scopic appearance of a section of the nectary is not that of secre- 
tory tissue, and I have never observed crystals or dried secretion 
when examining sections. Each epidermal cell of the inside 
of the base of the petal is produced into a small papilla, more of 
less bulbous at the top. The cuticle is thickened in ridges, and 
these ridges continue over the papillae in a slightly spiral manner. 
I have frequently seen within the papillae tiny drops of oil, which 
is doubtless the flower’s scent. The petal sap is rich in glucose 
as shown by its reduction of Fehling Solution. There are two 
ways in which flies might obtain this sweet sap — by breaking the 
papillae or by squeezing it through them. I have rejected the 
idea again and again and sought some other solution, but now I 
feel convinced that the latter is the method adopted. I think the 
papillae are too strong to be broken by the flies. If, however, 
they were broken, surely broken ones would be seen in micro- 
scopic examination ! I have examined a great number of 
sections, but I have never found a broken papilla. If the broken 
papillae retained their virgin shapes sap would surely ooze out 
through the split cell-walls and form nectar in the nectary. A 
flower is visited not only once, but many times, and, if breaking 
were the rule fresh cells would probably be broken each visit, 
and broken ones would not mend. This would surely mean a 
copious flow of nectar, and a quick withering of the flower, but 
