558 REPORT— 1897. 



method used by Professor Dixou in some of his recent experiments) in which a 

 photograph is taken on a film fixed to the rim of a rapidly revolving wheel. 



On trial it was found that the molecular speed is sufficiently small to be 

 measured by this method. 



The experiments were conducted by M. Gustav Hemsalech, to whose care 

 end skill their success is largely due. Without entering into a detailed descrip- 

 tion of the apparatus it will be sufficient to say that the photographs exhibited 

 to the Section were taken on a film moving with a linear speed of about 80 ^^^ 

 in a direction at right angles to the slit of the spectroscope. While the air lines 

 appear perfectly straight, though slightly broadened, the metallic lines are 

 inclined and curved. The spark was taken from five Leyden jars, charged by 

 means of a Voss machine, the distance between the electrodes being about 

 1 cm. A single spark produces a good spectrum, reaching approximately from 

 X = 5000 to X = 4000. 



Thotograph I. is that of a spark taken between zinc poles on a stationary film. 

 It serves to show the sharpness of the lines in a spectrum of zinc and air. 



Fkotograph II. was taken immediately after /., but on the moving film ; the 

 curvature of the zinc lines shows that the velocity of the molecules is gradually 

 diminishing away from the poles. Close to them it must be very large. The 

 averao-e velocity up to a distance of one millimetre from the electrode is 

 2000 5?^ and at a distance of four millimetres. This is reduced to something 



second' " 



like 400. 



Photor/raph III. The upper pole is still zinc, while in the lower pole a piece 

 of metallic bismuth has been substituted. The three most refrangible ones, that 

 at 4259, being the strongest of them, are decidedly more inclined than the zinc 

 lines, while the line 4560 seems almost straight. 



Photoqraph IV . The poles are again zinc and bismuth, but both poles are 

 moistened with a solution of calcium chloride. The photograph reveals the 

 curious fact that Bi 4259, which was very much curved on HI., is now much 

 less so. The comparison with Ca 4226 clearly shows the greater inclination of 

 the calcium line. The latter is more inclined than H and K,and if the difference 

 in molecular velocity is due to dift'erences in molecular weight, this would show 

 that II and K belong to a simpler molecule than 4226. 



I do not desire to express any opinion respecting the bearing of these experi- 

 ments on the hypothesis of dissociation, as some of the photographs reveal rather 

 puzzling appearances which must first be cleared up before any certain conclusions 

 can be drawn. 



All the photographs show clearly that the luminosity of the metallic particles 

 is a plienomenon subsequent to the discharge proper which takes place through the 

 air. Even close to the pole the brightest parts of the metal lines are displaced, as 

 compared with the brightest part of the air line. If we could fix our attention 

 ■on a point halfway between two zinc poles we should see this point flash out 

 twice with a dark interval between the luminosities. At the moment the spark 

 passes, the air becomes luminous, and remains so for a period, which, in our 

 •experiments, did not exceed ttooooo ^^ ^ second. After an interval of about 

 __!__ part the zinc molecules arrive at the centre of the spark, and remain 

 luminescent for an appreciable time with diminishing intensity. The numbers 

 are of course approximate, as they must depend on the intensity of the spark. 



The photographs submitted had been enlarged about five times, but a few 

 prints taken from the original negatives were also shown. 



The experiments were made with comparatively rough appliances, and the 

 optical arrangement was defective in several respects. A more perfect apparatus 

 is in course of construction, and I hope to continue the research in conjunction 

 with M. Hemsalech. The preliminary results which have been described are suf- 

 ficient to show that the method is likely to furnish interesting information. 



