PROPERTIES OE THE LEAP OP DIORLEA, 
27 
the readings, they were not repeated more frequently than appeared necessary. In 
each experiment the leaf was introduced into the warm chamber at 32° C. some time 
before the observation began. The observations were made at intervals of two 
minutes and a half, and the number of excitations necessary for each was from 20 to 
25. Consequently, each lasted nearly an hour. 
The observations given in. the table serve to show that the variation lasts during the 
whole of the first and the greater part of the second second, and that it culminates 
at from six to eight-tenths of a second after excitation. For these purposes it was 
thought best to employ long periods of closure (using the galvanometer in a very 
insensible state) and to make the end of one period coincide with the beginning of the 
next in such a way as to cover the whole period to be investigated. But such a 
method was quite inadequate to give information as to the question of greatest 
importance in relation to the excitatory variation—that of the time at which the 
effect commences. For this purpose it was necessary to shorten the periods of closure, 
and consequently to increase the sensibility of the galvanometer. We therefore 
substituted for the revolving rheotome the “ pendulum ” rheotome, of which the 
description is given in Part III. The instrument is so arranged for the experiment 
that on swinging from right to left it opens three keys in succession—K 1, K 2, 
K 3 (see fig. 8). Of these, the first occupies a fixed position, so that it is opened 
about O'l" after the pendulum is liberated. This position corresponds to zero of the 
pendulum scale. The other two keys, K 2 and K 3, can be slid along to any required 
distance from K 1. The binding screws of K 1 are connected with the primary 
circuit of the induction coil; those of K 2 with the shunt of the galvanometer, so 
that when this key is closed the galvanometer is “short circuited;” those of K 3 are 
interpolated in the galvanometer circuit between the leading-off electrodes and the 
shunt. Consequently, when the three keys (being previously closed) are opened in 
succession by the pendulum, the primary circuit is first opened by K 1, the galvano¬ 
meter circuit is unbridged—that is, closed—by K 2, and again broken by K 3. As 
the scale of the pendulum is divided into hundredths of a second and extends over a 
period of a quarter of a second, this arrangement affords the means of investigating 
with great exactitude the events of the beginning of the variation. 
The galvanometer is brought to such a degree of sensibility that with a resistance 
of 10,000 ohms, in addition to that of the galvanometer in circuit, O'OOOl D. gives a 
deflection of 97 scale. To facilitate accurate reading a slightly concave mirror is 
placed in front of the scale, by means of which an enlarged image of the part of 
the scale most nsed can be seen by the observer without fatigue. The leaf is led off 
and excited as in the fundamental experiment. For excitation, non-polarizable 
electrodes of the form described at D, p. 18, are used, the contacts being opposite, 
and corresponding in position to those of the leading off electrodes applied to the 
opposite lobe. The plant of which it forms part is placed in the chamber, the air of 
which is saturated, at 26° C. 
E 2 
