coated with shellac varnish to render it highly insulating. The index is projected to a short distance on the opposite side, for attachment of a counterpoise; this takes the form of a vertical vane of mica which acts as a damper. The galvanometer itself is of an aperiodic type, and the addition of the damper makes it perfectly dead-beat. The sensitiveness of the galvanometer is such that a micro-ampere of current produces a deflection of 10 mm. of the index. The recording index has attached to it a short vertical piece of thin platinum wire pointed at its two ends; this end of the index moves between a sheet of metal M, on which is spread the recording paper, and a semi-circular piece of narrow metal sheet C. The metal sheet M is mounted on wheels and moves at an uniform rate by clockwork. Record is made by sparks. One electrode of the sparking coil is in connection with C, and the other with M. The sparking thus takes place simultaneously, above and below the vertical and double-pointed platinum wire carried at the end of the index. There is thus no resultant kick, and the index remains undisturbed. The sparking, as previously stated, takes place three seconds after exposure of the selenium cell to light, by which time the deflection reaches its maximum. The record thus consists of successive dots at intervals of 15 minutes, the dots representing the maximum deflections of the galvanometer corresponding to the intensities of light.
The record given in figure 107 was taken about the end of January; the sun rose at about 6-45 a.m. and set at 5-30 p.m. The twilight is very short in the tropics; the sky is feebly lighted about 6 a.m.; it becomes dark about 6 p.m. The record shows the intensity of light
