at one time it crossed 5 divisions in 1 ½ h.  The pot had to be moved occasionally, as the end of the filament travelled beyond the field of vision; but as far as we could judge it followed during the daytime a semicircular course; and it certainly travelled in two different directions at right angles to one another.  It sometimes oscillated in the same manner as in the last species, some of the jerks forwards being as much as 1/1000 of an inch.  We may therefore conclude that the joints in this and the last species of grass long continue to circumnutate; so that this movement would be ready to be converted into an apogeotropic movement, whenever the stem was placed in an inclined or horizontal position.

Movements of the Flower-peduncles of Oxalis carnosa, due to apogeotropism and other forces.—­The movements of the main peduncle, and of the three or four sub-peduncles which each main peduncle of this plant bears, are extremely complex, and are determined by several distinct causes.  Whilst the flowers are expanded, both kinds of peduncles circumnutate about the same spot, as we have seen (Fig. 91) in the fourth chapter.  But soon after the flowers have begun to wither the sub-[page 504] peduncles bend downwards, and this is due to epinasty; for on two occasions when pots were laid horizontally, the sub-peduncles assumed the same position relatively to the main peduncle, as would have been the case if they had remained upright; that is, each of them formed with it an angle of about 40o.  If they had been acted on by geotropism or apheliotropism (for the plant was illuminated from above), they would have directed themselves to the centre of the earth.  A main peduncle was secured to a stick in an upright position, and one of the upright sub-peduncles which had been observed circumnutating whilst the flower was expanded, continued to do so for at least 24 h. after it had withered.  It then began to bend downwards, and after 36 h. pointed a little beneath the horizon.  A new figure was now begun (A, Fig. 188), and the sub-peduncle was traced descending in a zigzag line from 7.20 P.M. on the 19th to 9 A.M. on the 22nd.  It now pointed almost perpendicularly downwards, and the glass filament had to be removed and fastened transversely across the base of the young capsule.  We expected that the sub-peduncle would have been motionless in its new position; but it continued slowly to swing, like a pendulum, from side to side, that is, in a plane at right angles to that in which it had descended.  This circumnutating movement was observed from 9 A.M. on 22nd to 9 A.M. 24th, as shown at B in the diagram.  We were not able to observe this particular sub-peduncle any longer; but it would certainly have gone on circumnutating until the capsule was nearly ripe (which requires only a short time), and it would then have moved upwards.