EXPLOSIONS BY A DISTANT ELECTRIC SPARK. 209 



a given solution. A trough, g, is to be provided, which may be formed by drilling a 

 hole two and a half inches in diameter through a thick piece of plate glass ; on each 

 side of this a thin piece of similar glass is laid, the trough having been filled with the 

 solution under investigation. Troughs made in this manner never leak ; they com- 

 pletely answer their purpose, and are easily washed and refilled. 



958. Let the substance under trial be a concentrated solution of bichromate of pot- 

 ash. Having adjusted the lamps and filled the trough, set it before one of them, sup- 

 porting it by a proper foot. On looking through the tube of the photometer, if the ab- 

 sorbent cell of the copper solution has been previously removed, the circle of light will 

 be seen very brightly illuminated ; one of its halves of a yellow tint, but not much less 

 luminous than the other. The copper cell being now restored to its place, on looking 

 again through the tube there is a striking contrast one half of the circle is of a bright 

 blue, but the other seems totally black ; with solutions which cut off the blue rays less 

 perfectly, this blackness is of course less intense ; in these cases the lamps are to be 

 moved into such positions that the two halves of the circle are equally illuminated, and 

 its dividing diameter invisible. As the eye is not disturbed with any difference of col- 

 our, the observation can be made within small limits of error. The calculation of the 

 relative intensities can then be made by the common photometrical law. 



959. As this photometer is affected by rays of the same refrangibility as those which 

 affect the tithonometer, it is clear that if it be the rays of light which are operative in 

 the union of chlorine and hydrogen, the results given by the two instruments should 

 correspond within certain small limits of error. 



960. With respect to the tithonometer, I have improved this instrument considera- 

 bly : by shading it with a glass case, which cuts off thermometric disturbance ; by ta- 

 king the observations through a small telescope, which avoids parallax ; by having the 

 scale movable, so as to slide along the tube; by making one charge of gas last for a 

 great number of experiments ; by completing the tithonization before commencing; by 

 altering the position of the adjusting wire, so as to bring it nearly down to the end of 

 the tube. The detail of these changes would, however, detain me now too long; they 

 will be given at some suitable time hereafter. 



961. I may, however, record as a striking fact, that so great is the sensitiveness of 

 chlorine and hydrogen, that a mixture will actually explode by the rays of an electric 

 spark. In Chapter XVI. I have already stated that silent combination would take 

 place under these circumstances, but more recently I have had instruments repeatedly 

 destroyed by explosions resulting in that way. 



962. It being understood that the indications of the photometer and the tithonome- 

 ter correspond, that they are affected by rays of the same refrangibility, we are enabled 

 to proceed to the direct solution of the question, and the determination of M. BECQUE- 

 REL'S hypothesis. 



963. A transparent medium, which absorbs, to a greater or lesser extent, the more re- 

 frangible rays, being selected, it is required to determine whether, when a given ray has 

 passed through it, the chemical effect diminishes as the intensity of the more refrangi- 

 ble rays of light diminishes. According to M. BECQUEREL, the effect should be in di- 



