159 DR. FRANCIS DARWIN ON THE MECHANISM BY WHICH 
The hygroscopic action of the awn may be shown by holding the seed so that the twisted 
axis is vertical; on wetting the awn, the feathery portion will revolve, describing a circle 
in a horizontal plane; the vertical part untwisting, and its strand-like ribs becoming 
more and more oblique ($. e. making a smaller angle with the axis of the awn), until at 
‘last, when the awn is thoroughly wet, it shows no signs of torsion, and the two ribs 
run parallel to its axis down two of its opposite surfaces. Аз the awn untwists, the 
two knees, А, and k, are obliterated (k, disappearing first); by this means the feather 
is brought into line with the twisting part, and the awn forms a long straight pliable 
rod, pointing vertically upwards. It follows from this, that though the end of the 
feather begins (at the commencement of the wetting) by describing a circle, it ultimately 
describes a helix. 
As the awn dries again, the action is exactly reversed : the axis becomes twisted; the 
two knees reappear ; and the feather becomes again subhorizontal. 
It was this remarkable property exhibited by Stipa of untwisting and twisting in 
response to alternations of moisture and drought, which convinced my father and myself 
that the whole structure must be adapted to some very definite end. We concluded that 
this object must be that of forcing the seed into the ground (in the manner to be imme- 
diately described). The sequel gives the result of my investigation of this point. 
If in the above-described experiment the feather is prevented from revolving while 
the seed is left free, the untwisting of the vertical part of the awn will make the seed 
rotate on its axis. Іп consequence of the untwisting, the vertical part of the awn from 
the knee (2,) to the tip of the seed becomes longer by about 7 per cent. Therefore, if we 
fix a dry Stipa-awn іп a vertical position, the seed resting freely on the ground, and the 
feather being prevented from revolving, on wetting the awn the seed will rotate and 
will be pressed against the ground. But this requires that the expansion of the awn 
shall not be permitted to take place in an upward direction; i.e. the upper extremity 
must be fixed to givea point d'appui, from which the expansion may take effect in 
pressing the lower extremity against the soil. 
In describing the untwisting of the axis it was stated that the angles k, and 4; increase 
as the амп absorbs water, until the whole awn is converted into а straight rod. Now in 
the same way that, by interfering with its movements, the rotation of the feather is 
transferred to the seed, so by preventing the feather from rising to its full extent the 
tendency of theawn to become straight results in increased pressure of the point against 
the ground. Тһе seed now resembles a brad-awl in exhibiting vertical pressure оп 4 
revolving point. In order that this action may take place in its entirety, three poinls 
d'appui are theoretically needed, although one is practically capable of performing the 
duties of all. The first is required, to transfer the rotation from the feather to the seed ; 
the second, to force the expansion resulting from the untwisting of the awn to take effect 
downwards; the third, to convert the tendency of the awn to straighten itself into 
pressure of the seed against the ground. It is evident that if an awn be placed vertically, 
the seed resting on the ground, so that the feather can neither rotate nor describe any 
great angle in a vertical plane, all the required conditions will be fulfilled. 
How can this be effected in a state of nature? Prof. Hildebrand has pointed 
