160 DR. FRANCIS DARWIN ON THE MECHANISM BY WHICH 
hygroscopic half on the convex surface, and the strongly contracting cells on the concave 
aspect. То prove that the internal cells do contract longitudinally, it is only necessary to 
split an awn longitudinally ; on drying, the split-off pieces bend with the internal surface 
concave. There can be no inherent tendency in the separate cells at 0; to bend in the 
direction assumed by the awn; for a small portion severed by a longitudinal section 
(such as is shown by the upper dotted line, fig. 13) bends with the internal surface con- 
cave, which is in the opposite direction to that in which the whole awn bends. 
The bending at the lower knee is more difficult to explain. The hollowing-out of the 
cells near m & m has begun even at this part of the awn: two hollow cells are shown in 
fig. 12; and in other sections I have found more of them occurring. Іп this section, 
again, we might draw a line which would divide the awn into a more and a less contractile 
half. In fact we have at the lower knee a bending-mechanism similar to that which 
exists at the upper knee, but with an important difference: at k, the torsion-power of 
the awn is exhausted, whereas at the lower knee this is not the case; and since the 
unequal longitudinal tension just described can satisfy itself by assisting the torsion, it 
does not seem evident how the bending of the awn is to be effected. To explain this a 
hitherto neglected point in the structure must be noticed. If we compare once more a 
section of the awn close to the seed with one taken at either of the two knees, we find 
that in the lower section (fig. 11) the ribs ^^, are separated from one another by approxi- 
mately equal portions of the circumference; but at the lower knee (fig. 12) the are on 
the side 2% m has diminished, while the other has increased. In a wet or untwisted awn, 
the fact that the ribs approach one another as they pass from below upwards, gives the 
appearance of a very slight and elongated left-handed screw. And I believe that the ap- 
proach of the ribs to one another is directly connected with the unequal arrangement of 
the more and less hygroscopic tissue at the knees. At the lower knee there is, moreover, 
a single sudden and short turn of a left-handed screw. This I call the ** reverse twist," 
since it is in the opposite direction to that in which the awn twists in drying. The 
reverse twist is perfectly distinct from the ordinary hygroscopic torsion of the awn ; for it 
is not obliterated by wetting, and is the result of the growth of the tissue into that 
shape. It is like a single turn of a fluted spiral column made of india-rubber, which 
could obviously be twisted in either direction independently of its shape. Now since the 
awn twists in the opposite direction to the reverse twist, the latter will be obliterated 
аз torsion proceeds ; and this is, in fact, the case: the awn is twisted both above and below 
the knee; but at the actual bending-place there is no torsion. When the reverse twist is 
obliterated, the unequal longitudinal tensions at the point б; are employed in producing 
the bend, instead of satisfying themselves in assisting the torsion. ‘To summarize this 
imperfeet explanation, the torsion-power satisfies itself in obliterating the reverse twist; 
and then the unequal tension, being brought into a longitudinal direction, satisfies itself 
by producing the bend. 
I suspect it was the reverse twist which deceived Max Wichura*, and led him to de- 
scribe the awn of Stipa as twisted to the right below the knee, and to the left above it. 
г * “On the winding of Leaves,” Taylors ‘Scientific Memoirs; May and Aug. 1853, p. 280; the original ™ 
* Flora,' 1852, Jan. and Feb. 
