410 
PROCEEDINGS OF SECTION D. 
the stigma are narrower, and there are less glands on the style. The 
stigmatic lobes, when straightened out, measure from one-third to one- 
half of the rest of the style (Fig. 6), while in the other variety the 
lobes are as long as the remainder (Fig. 7). 
The tube of the corolla in the long-styled flowers has five rows 
of long cottony hairs (Fig. 10) reaching over so as to touch the style; 
but this character, though it generally occurs, is not constant. These 
hairs often catch and retain much pollen, and thus form a means for 
self-fertilisation if cross-fertilisation fails, as the lobes of the stigma 
bend back into the hairs. The percentage of long-styled forms, as 
determined from gatherings of some hundreds of flowers at various 
times and in different localities, is 58*8, and of short-styled 41*2. 
The plant is remarkably variable, but, undoubtedly, these may be 
taken as the two main varieties. The following arc some of the 
directions in which it varies : — 
1st. In size of plant and flower. A specimen in my herbarium 
is about 1 inch in height, and the flower is barely ^Y-inch across. 
2nd. In the number of parts. Sepals and j)etals may be 4, 5, 6, 
or 7, and one well filled-in double flower was observed. Anthers mav 
be 3, 4, or 5. The stigma may have 2, 3, or 4 divisions. 
3rd. The style may be long or short. 
4th. The inside of the corolla may be hairy or glabrous. 
5th. The calyx may be downy or glabrous. 
6th. The leaves vary in being broader or narrower, and hairy or 
glabrous. J 
7th. The free limbs of the corolla vary in breadth. 
8th. The colour of the corolla is white, pale blue, deep blue, and 
rarely pink, and the outside is white, faintly blue, or rusty buff. 
8o far as my observations go, I believe the small-flowered variety 
reproduces itself constantly, as I have known clumps of it for years 
in one place, close to colonies of the large-flowered variety. But the 
small variety is more common on poor, dry soils ; and I have not 
observed it on the uplands near the sea, where the large variety 
flourishes. 
The flowers close in the evening and on cloudy days, and then, by 
a bending of the stalk immediately below the ovulary, hang with the 
corolla mouth downward. 
CANDOLLE ACEiE. 
LeewenlioeJcio dnbia , Sond. — I have never succeeded in making an 
observation of the actual process of fertilisation in this plant; but, 
with the aid of a knowledge of the ‘method in Candollea (4), I have 
formed a theory which I give here, in the hope that some member of 
the association who has the opportunity will observe the plant and 
publish his results. The flower consists of five petals, four of which 
open in the usual position, while the fifth, which is folded into a 
slipper-shape, hangs down the side of the tube, and is irritable. The 
stamens and style are united in a column which stands up in the 
middle of the flower. The anthers are strongly proterandrous, and 
hide the stigma till all the pollen is shed. Then the stigma enlarges 
and becomes mature. I imagine the process to be as follows : — An 
insect alights on the flower, and in probing the tube for nectar touches 
the irritable petal, which then flies up and imprisons it against the 
