Yapp. — Fruit- dispersal in A denostemma viscosum. 313 
stains a deep red with alcannin, and with osmic acid takes on a faint brown 
colour, though it is not blackened as an oil would be. In fact, it behaves 
generally like a body akin to resins, though, as seen above, it does not 
quickly harden, as an ordinary resin should. Possibly the secretion may 
be more or less allied to caoutchouc, or to the supposed caoutchouc-like 
substance which has received the name of ‘ Viscin V 
The pappus-setae are formed at a very early stage in the development 
of the flower. Each contains a single vascular bundle 2 ; and, as the fruit 
matures, the walls of the cells surrounding this bundle become indurated 
and thickened, except in the region of the glandular hairs, and in the 
swollen base of the seta. 
The structure of this basal swelling is interesting. It is composed, for 
the most part, of large, thin-walled cells (with no inter-cellular spaces), 
elongated in a direction at right-angles to the axis of the seta itself (Fig. 6). 
After the gland-stalks have assumed their final horizontal position, sections 
taken through the base of a seta show that these large cells have collapsed, 
their walls being much crumpled. This, together with the internal structure 
of the swelling, and the fact, observed in the field, that the swelling itself 
disappears during the movement of the gland-stalks, furnish, I think, 
sufficient proof that the dilated bases of the setae constitute the mechanism 
of movement. The base of each pappus-seta is in fact a pulvinus, composed 
largely of thin-walled motor cells, which probably act by losing water 
during the drying and ripening of the fruit. 
The horizontal position assumed by the pappus-setae is probably 
advantageous in that it provides the fruit with a more extended (and there- 
fore more secure) base of attachment. 
In view of the fact that the feathery pappus of many Compositae 
executes similar movements, I examined some of our British species, with 
the result that, in a number of instances, a mechanism resembling that 
described above was found. In the cases examined, instead of the down- 
ward movement of the hairs being due, as stated by Haberlandt 3 (apparently 
for Compositae generally), to an unequal thickening of the walls at the 
bases of the hairs themselves, it was effected by means of a continuous 
annular pulvinus, on the edge of which the pappus hairs were borne (Figs. 
7 - 9 )- 
I have since found that two other authors (Taliefif 4 and Hirsch 5 ) have 
1 Cf. Parkin, Observations on Latex and its functions, Ann. of Bot., 1900, vol. xiv, p. 203. 
2 The Composite pappus rarely contains vascular tissue. Hoffmann, loc. cit., p. 97, states that 
only in some four genera (amongst which he does not mention Adenostemmd) have even traces of 
vascular bundles been found in the pappus. 
3 Haberlandt, loc. cit., p. 465. 
4 Talieff, Ueber das hygroskopische Gewebe des Compositen-Pappus — abstract by Rothert : Bot. 
Centralblatt, 1895, vol. Ixiii, p. 320. 
5 Hirsch, Ueber den Bevvegungsmechanismus des Compositen-Pappus. Dissertation, Wurzburg 
Berlin, 1901. 
Y % 
