( 412 ) 
water or glucose on the outside of the florets in the capitula of 
those Compositae, might also be found in other places. 
This expectation has been realized. In Cosmos hybridus and also 
in Dahlia n j the leaves are already covered in very early youth by 
two kinds of trichomes, firstly with such as correspond structurally 
to the water- and glucose-secreting glands of the florets and secondly 
with longer {Dahlia) or shorter hairs {Cosmos), which from a broad 
base are continued to a conical point, and have a thick wall with a 
striped cuticle. The former consist of a row of 15—20 thin-walled 
cells, succeeding one another somewhat like a chain of pearls, 
which in the middle of the trichome are as long as broad but which 
gradually become longer towards the apex. These cells, especially 
those of the upper half of the trichome, are filled with dense granular 
protoplasm. They are readily stained by aniJine-dyes (methylene blue, 
rosaniline violet). In Dahlia they are chiefly found above the median 
and lateral veins, pressed against the surface on both sides of the 
leaf. In Cosmos hybridus , where the leaf-segments are very narrow, 
they are also found in largest numbers on the upper surface of the 
leaves above the vein, but here and there also on the lower surface. 
The other hairs have a different shape in Cosmos and in Dahlia, but 
it is not always constant in different parts of the same plant. 
In Cosmos those occurring on the lower surface and on the margin 
of the leaf are composed of 6 cells; they arise from a broad and 
occasionally multicellular base, are conically pointed and curved like 
a sickle; the content is more or less highly refractive and the cuticle, 
especially that of the apical cell, is striped. 
In Dahlia sickle-shaped trichomes occur on the margin of the leaf- 
segments, but those on the upper and lower surface of the leaf are 
much longer, are not curved and are not restricted in their distri¬ 
bution to the veins; the cuticle of all cells is striped. These hairs are 
much more numerous on the lower surface of the leaf than on the 
upper. 
It is easy to verify that in the very youngest leaves the glandular 
hairs secrete mucilage (resin or balsam). In a somewhat more advanced 
state of development the glands above the median and lateral veins 
are partly, as it were, surrounded by mucilage; the cells of the upper 
half of the trichome have not infrequently perished during the muci¬ 
lage secretion. 
With cut plants placed in a glass of water under a bell-jar we 
observe exactly the same phenomenon as in Calystegia and the 
Solanaceae. 
q In Coreopsis I did not investigate this point. 
