The Oscillator is diagrammatically shown in Fig. 15. M, M′ are the two electromagnetic coils, the free ends of the horseshoe being pointed. Facing them are the conical holes of the soft iron armature A. This armature carries two rods which slide through hollow tubes. The distal ends of the rods support the holder H, carrying the smoked glass plate. Under normal conditions, the plate-holder is held by suitable springs, somewhat to the right of, and free from contact with, the writer. A clockwork C carries a rotating arm, which makes periodic contact with a pool of mercury contained in the vessel V, once in a minute. On the completion of the electromagnetic circuit, the armature A is attracted, the recording glass plate being thereby moved to the left making contact with the writer. The successive dots in the record thus take place at intervals of a minute. Only a moderate amount of electric current is thus consumed in maintaining the oscillation of the plate. A 4-volt storage cell of 20 amperes capacity is quite sufficient to work the apparatus for several days. The responsive fall of the leaf of _Mimosa_ is completed in the course of about two seconds. The leaf remains in the fallen or ‘contracted’ position for nearly fifteen seconds; it then begins to recover slowly. As the successive dots of the Oscillating Recorder are at intervals of a minute, the maximum fall of leaf is accomplished between two successive dots. The dotted response record here obtained exhibits the recovery from maximum fall under stimulation (_cf._ Fig. 23). The recovery of the leaf in one minute is less than one-tenth the total amplitude of the fall, and is proportionately the same in all the response records. Hence the successive amplitudes of response curves that are recorded at different hours of the day afford us measures of the relative variations of excitability of the plant at different times. This enables us to demonstrate the reality of diurnal variation of excitability. In my experimental investigations on the subject I have not been content to take my data from any particular method of obtaining response, but have employed both types of recorders, the Resonant and Oscillating. It will be shown that the results given by the different instruments are in complete agreement with each other. EFFECTS OF EXTERNAL CONDITIONS ON EXCITABILITY.