THE FERTILISATION OF SALVIA AND OTHER FLOWERS. 26 ^ 
opened carefully a drop of nectar will usually be seen jus 
above the constriction. The use of the constriction seems, then, 
to be, to prevent the nectar from escaping below, and getting 
out of reach of the insect’s proboscis. The glandular hairs are 
most abundant just above the constriction, and get scantier and 
smaller higher up. The purpose which is here served by the 
constriction is in many other species attained by the glandular 
hairs themselves, which are set so thickly as to form a dense 
fringe, which only just leaves room for the passage of the style, 
and with this completely blocks the tube. (See fig. 2.) 
I come now to a matter which seems to me of considerable 
interest, but concerning which I would speak with diffidence. 
On examining a number of blossoms of Salvia patens, I found 
that there were two kinds. The great majority, in the arrange- 
ment of stamens and pistil, accorded with the description I have 
just given. In a certain number, however, the arrangemeni 
was different. In these the style was much shorter than in the 
others, and only passed once between the anthers, namely, at the 
lower part of the hood. (Fig. 14.) It thus projected from the hood 
below the fertile anther cells, and not above, as in the majority 
of blossoms. The consequence of course would be, that whert 
the stigma and the anther cells are brought down on to an insect’s 
back, the stigma would strike at a point nearer to the insect’s 
head than would the anthers. (Fig. 15.) It is plain, that an insect 
visiting first one form of blossom and then the other would have 
the same points on its back in contact first with the anther cells of 
one blossom, and then with the stigma of another. This dimor- 
phism would therefore be a second way of insuring crossing 
between different flowers. The blossoms with the long style 1 
found very many times more numerous than the blossoms with 
the short style ; and it may therefore be that these latter were 
only accidental, though tolerably frequent, deformities. It is plain, 
however, that if such dimorphism be of use to the plant by in- 
suring intercrossing, the plant, when growing wild and subjected 
to a struggle for existence, might avail itself of this “ accidental ’* 
occurrence, and that in time the short-styled flowers might come 
to be equally numerous with the long-styled ones. 
The flower is not, as far as I can make out, fertilised in thii- 
country by any insect. Growing only in a cultivated condition, 
it is not subjected to any other struggle for existence than that 
entailed by the changing fashion and caprice of horticulturists. 
This, however, has been so severe, that I was unable to obtaiii 
last summer from Covent Garden shops any large number of 
specimens, and thus I am unable to say in what proportion th< 
two forms of blossoms exist. 
I will not weary the reader with descriptions of other species. 
I have examined some thirty, and in all have found some con- 
