1893.] Waterhouse — Electrical action of Light upon Silver. 19 



sub-divided into millimetres, I have usually fixed the zero point at 30, so 

 that the readings above or below it may as far as possible show 

 different signs of E. M. F., and the direction of the currents has been 

 so arranged that a change in the position of the index to the apparent 

 left from 30 to shall indicate that the exposed plate is negative to 

 the unexposed, as copper to zinc, while a change to the apparent right, 

 30 to 50, shows that it is positive, or as zinc to copper. 



The coloured glasses used have been of the kinds ordinarily met 

 with in the bazar. A deep ruby, a brownish yellow, a medium green, 

 and a dark blue, and conditions being favourable it has generally been 

 possible to observe some trace of a current even with the deep ruby in 

 strong sunshine. 



"When observations were made with the spectroscope, whether with 

 a Rowland's diffraction grating or prisms, it was found that the 

 amount of light admitted through the slit for ordinary work, was quite 

 inadequate, even when the slit was open at its widest ; and it was 

 therefore necessary, in most cases, to use a much wider slit, or to 

 dispense with its use altogether ; also to use the directing magnet on 

 the galvanometer to increase the sensitiveness. 



In all cases sunshine has been reflected on to the sensitive plates 

 by means of a heliostat, as it was not convenient to use the direct rays 

 of the sun. With the flat cell there were thus two reflections, but 

 any loss of light was amply made up by the increased surface exposed. 



As is usual in such experiments, there were almost invariably 

 more or less strong local or polarisation currents generated between the 

 plates themselves, especially when they were freshly immersed in the 

 solutions, and it was generally found desirable to leave the cell from 12 

 to 24 hours before use, so as to give time for these currents to subside. 

 Sometimes, however, from half an hour to an hour, or even in some cases 

 a few minutes is sufficient. It was found, too, that even if there was 

 no polarisation current at the commencement of an experiment, the 

 action of light occasionally gave rise to fairly strong currents quite 

 independent of, and sometimes opposed to, the currents produced by 

 exposure to sunshine, while at others they were in the same direction. 

 Thus it was sometimes difficult to ascertain how far the currents 

 observed were due to light or to polarisation. The only test was the 

 retrograde movement of the needle after shutting off the light. 



Another difficulty in making these observations may be noted, and 

 that is, the apparent reversals of current which are due in many cases 

 to decrease in the strength of the light, though the decrease may be 

 almost imperceptible. For the same reason, if coloured glasses be 

 applied without first completely shutting off the light after the plates 



