Electro-optical Investigation of Polarized Light, 219 



silicates, and cements composed of phosphoric acid and me- 

 tallic oxides produce a decrease in sensitiveness to light after 

 a time, probably in consequence of the increase of gas-pressure 

 due to the evolution of hydrogen by the action of metallic 

 vapours on the water contained in the cement. A better 

 result was obtained with molten glacial phosphoric acid if 

 the precaution was taken of covering the joints, immediately 

 after the plate had been cemented on, with zinc oxide which 

 had been washed and ignited, and then covering the joint 

 with a layer of a mixture of wax and resin. It is true that 

 with the apparatus so constructed, the sensitiveness also de- 

 creased for the first few days after sealing-off from the 

 air-pump ; but this falling off soon ceased. We were thus 

 able to construct vessels with windows of parallel glass, which 

 might serve at least for control-experiments. 



These difficulties induced us to use generally simple glass 

 bulbs (cells) blown before the blowpipe, in which the alkali 

 metal and its vapours were in contact only with the glass 

 walls and the platinum electrodes. For most of the experi- 

 ments here described we employed such bulbs of 50 millim. 

 diameter which were half filled with the alloy of sodium and 

 potassium. Each ray, therefore, that strikes the centre of 

 this metallic mirror must cut the glass wall at right angles, 

 and its intensity will not vary with the azimuth. The cross 

 section of the beam was therefore made as small as the sensi- 

 tiveness of the cell permitted. 



In order to obtain a beam of light of small section and great 

 intensity, we obtained a projection lantern {sciopticon) with a 

 little disk of zirconia heated in the oxygen-coalgas-flame as 

 source of light. After this had been so adjusted that the 

 image of the piece of zirconia was obtained sharply at a con- 

 siderable distance, a screen with a slit, about 3 milKm. long 

 and 1 millim. broad, was placed between the condenser and 

 the projecting lens so that its outline was defined upon the 

 wall of the cell. The sciopticon, which with the screen 

 formed a rigid whole, was capable of rotation in a vertical 

 plane, and in any position could be inclined to the horizontal 

 through about 50° and held fast in that position. If it had 

 such a position that the spot of light which marked the 

 entrance of the ray of light was distinctly seen at A upon the 

 glass wall, then the point of emergence at A' could also be 

 distinctly recognized. The distances AB and A'B 7 of corre- 

 sponding edges of these spots of light from the horizontal 

 surfaces of the fluid metal were measured with a pair of 

 compasses, and the sciopticon and cell were so placed that AB 

 was equal to A'B'. We were thus sure that the ray of light 

 struck the centre of the kathode, and cut the glass wall at a 



Q2 



