- 
CERTAIN SEEDS BURY THEMSELVES IN THE GROUND. 159 
the result. The cells are of an elongated form, and are of course closely attached to 
one another, so that unless they are isolated they cannot twist on their own axis, since 
this would require a sliding movement between each cell and its neighbours. "The 
following experiment demonstrates the way in which each cell satisfies its tendency to 
twist on its own axis. 
А number of S/ipa-awns (exclusive of feather and knees) are soaked in water, and 
when thoroughly untwisted are firmly tied into a cylindrical bundle or faggot. Here 
we have represented the state of things in an awn: each constituent awn in the bundle 
represents one of the cells of an awn, and, like them, tends to twist on its own axis on 
drying. It is found that, on drying, the bundle of awns is converted into a rope, its 
constituents passing helically round like the strands. And just in the same way 
the cells which make up ап awn (of which the bundle of awns is a schema) satisfy 
their tendeney to twist by forming a rope of cells twisted in the same direction as 
themselves. 
If, then, we have a number of cells so connected together that they are incapable of 
independent torsion (and this is the case in the awn of Stipa), and if, further, each cell 
has a tendency to twist on its own axis (as we know to be the case with Stipa), then it 
cannot be doubted that the torsion of the mass of cells as a whole will be the result. That 
this is actually the mechanism by which the awn twists is conclusively shown by the 
fact that the constituent cells of the feathery non-twisting portion have ло such power of 
independent torsion when isolated and dried in the manner already described. 
The Bending of the Awn at the Knees (k, & k.).—1f we compare a section of the feather- 
bearing portion of the awn with a section of the twisting part, we find that the difference 
between them lies in this: in the twisting part the cells are all thick-walled (excepting 
the central ones, с, and two small masses, m & m, fig. 11); but in the feather all the 
cells are hollow, and it looks as if the central mass of thin-walled cells had increased so 
as to fill up the whole interior of the section. And since this thin-walled tissue is not 
hygroscopic in the feathery portion, we may assume that it is not so in the twisting part. 
Comparing a section taken low down in the awn (fig. 11) with one taken at the upper 
knee (fig. 13)*, we find a marked difference in the distribution of the non-hygroscopic 
tissue, At the upper knee the cells are passing into the hollow condition found in the 
feather; but this change does not attack all parts simultaneously: the small masses of 
non-hygroscopic tissue, m & т, have enlarged and coalesced with the central mass (с); 
and it is the cells surrounding the coalesced masses which exhibit the commencement of 
the loss of thickening in their walls. We may divide the section by a line гу, on one 
side of which there are many hollow cells, on the other all thick-walled cells. The 
twisting-power, which is already weak in the portion of the awn between the two knees, 
must be disappearing at k», since in the feather it is quite gone. When drying com- 
mences, the mass of thick-walled cells contract longitudinally ; and the mass of hollow 
cells not being able to contract to the same extent, the awn will bend with the non- 
* In the drawings of the sections, figs. 11, 12, & 13, the stratification of the cell-walls is not represented; nor are 
the pit-channels given, which are especially numerous in the large internal cells, and which seem to communicato 
With the vascular bundle in the centre. 
SECOND SERIES.—BOTANY, VOL. I. Z 
