parably higher than any that can be produced by a spark from an electric machine. Mr. R. W. Wood[1] obtained reversal with a single spark, when the photographic lens was wide open, but there was no reversal with four sparks, the lens aperture being reduced to one-fourth. The quantity of light was the same in the two cases, but the time-rate of illumination was different. This curious result would no longer appear anomalous, if we bear in mind the experiment in which the influence of time-rate was shown.
(2.) In trying to obtain photographs by heat radiation on sensitised papers coated with a mixture of silver and mercury iodides, the following curious effect was observed. The sensitised paper was exposed to heat radiation and became uniformly reddish in colour. A mask with cut-out letters was now put on it, and the sensitised paper was allowed to cool. The rate of cooling was very rapid at the places exposed by the cut-out letters, whereas at the covered portions the rate of cooling was very much less. After a long time when the sensitive paper had cooled down to a uniform temperature, prints were still visible, the effect being evidently due to the different rates of emission in the screened and unscreened parts.
(3.) Major-General Waterhouse in his paper[2] mentions an anomalous case which seems to be explicable from considerations given above. He took a polished silvered glass plate, and put it into a printing frame with a cut-out paper mask and mica screen in which were cut-out initials, just as if it were going to be exposed to the sun; but instead of exposure to light the plate was gently warmed for about 5 minutes over a spirit lamp, and then developed with mercury. The cut-out initials came out distinctly in dark lines. It seems to me that in this experiment, as the plate was uniformly warmed, the difference between the screened and unscreened portions could only be in the different rates of emission.
9. Phenomenon of Recurrent Reversals.
The fourth stage in the curve for the action of light (see fig. 16) will be found specially interesting with reference to photographic reversals. These reversals are found to be recurrent. Thus, starting with a neutral condition, we obtain the first negative with a moderate exposure; longer exposure will tend to reduce the intensity of the negative and give rise to a neutral condition. Further exposure gives rise to a positive, then a second neutral, and again a succeeding negative stage, and this often goes on in recurrent series.
Such recurrent reversals are also exhibited (see fig. 18) by a substance under continuous mechanical vibration. In my paper on "Elec-
