morning the end of the filament was bowed greatly downwards. It was
therefore removed and fixed lower down. The line traced during these two
days extended in the same general direction, and was in parts nearly
straight, and in others plainly zigzag, thus giving some evidence of
circumnutation.

As the arched epicotyl, in whatever position it may be placed, bends
quickly upwards through apogeotropism, and as the two legs tend at a very
early age to separate from one another, as soon as they are relieved from
the pressure of the surrounding earth, it was difficult to ascertain
positively whether the epicotyl, whilst remaining arched, circumnutated.
Therefore some rather deeply buried beans were uncovered, and the two legs
of the arches were tied together, as had been done with the epicotyl of
Tropaeolum and the hypocotyl of the Cabbage. The movements of the tied
arches were traced in the usual manner on
[page 32]
two occasions during three days. But the tracings made under such unnatural
conditions are not worth giving; and it need only be said that the lines
were decidedly zigzag, and that small loops were occasionally formed. We
may therefore conclude that the epicotyl circumnutates whilst still arched
and before it has grown tall enough to break through the surface of the
ground.

In order to observe the movements of the epicotyl at a somewhat more
advanced age, a filament was fixed near the base of one which was no longer
arched, for its upper half now formed a right angle with the lower half.
This bean had germinated on bare damp sand, and the epicotyl began to
straighten itself much sooner than would have occurred if it had been
properly planted. The course pursued during 50 h. (from 9 A.M. Dec. 26th,
to 11 A.M. 28th) is here shown (Fig. 22); and we see