152 DR. FRANCIS DARWIN ON THE MECHANISM BY WHICH 
Turn. | Completed in | Turn. Completed in | 
M. 8. | м. 8. | 
No. 1 Sw Ио 6 130 -—| 
2 9:20 | 7 1 45 
3 1 45 | 8 2 10 
4 id 9 3 90 
5 19571 10 | 
- 
Thus the rate increases up to the 5th revolution, and then diminishes quickly. 
The lower bend or knee is obliterated in from 20’ to 30’; the awn becomes com- 
pletely untwisted in from 3 to 1 hour; in doing so it increases in length by about 7 per 
cent.; a well-grown awn measures from the point of the seed to this lower bend 70 mm. 
dry, 75 wet. The upper knee disappears in periods of time which are rather variable, 
and lie between two and three hours. The movement which takes place as the awn 
untwists or twists is a combination of two rotations: the first is that due to the torsion 
of the vertical portion of the awn (50-55 mm.) from the lower knee down to the seed; 
the other rotation is due to the torsion of the short portion (15-18 mm.) between the 
knees. This latter axis is carried round quite independently of its own torsion, by the 
torsion of the long vertical axis; thus the portion Ё, k, (fig. 1) describes a cone about an 
axis of rotation which is continuous with v; and, again, owing to the torsion of the part 
kı b, the feather is made to rotate, describing a figure of rotation having a line con- 
tinuous with А, 0; for its axis. We have seen that the feather is usually nearly hori- 
zontal when dry, and vertical when wet (the seed being held vertically in both cases); 
it follows that in passing from the wet to the dry state the tip of the feather does not 
describe а regular helix, as would be the case if there were only one bend, but a complex 
figure varying with the varying relations between the rotations due to the two axes 2 
and А, Ё. The rotations due to tlie untwisting of Ё, Ё, are slow (7-18 minutes) ; and since 
but two revolutions are possible, owing to the shortness of this part, their effect is not 
always great on the revolutions due to the untwisting of the vertical axis, which are from 
13 to 16 in number. Тһе rotation of the алуп is best studied by fixing the seed while 
the upper end is allowed to revolve; and by the experience gained in this way it is 
possible to understand what occurs in the experiments on artificial burying where the 
awn is fixed while the seed is free to rotate. When an awn has been placed for seyeral 
hours in the damp chamber, fixed in the kind of support described, it will not always 
be found to have buried itself. I have often found that the point of the seed has been 
dragged across the soil into which it should have thrust itself; and in these cases it has 
. happened that the seed has been pushed down over the edge of the vessel, so that it 
would have tended to bury itself if the soil had been of a larger area. It is this 
dragging motion which I believe to be produced by the torsion of the short axis between 
kı and №, (fig. 1); for just as, when the awn is free, irregular movements are impressed 
on е rotation, so ће tendency to irregular rotation is transferred to the seed when ће 
awn is fixed. In other cases the experiment is successful, and the seed is found buried 
to а certain depth; the sand is often found disturbed close around the point, owing, DO | 
