Non-luminous Radiation emitted by a Gas Flame. 



67 



portion of the non-luminous radiation of an Argand gas burner, but 

 that when the thickness of the water is increased to 15 cm. the 

 transmitted radiation consists exclusively, or almost exclusively, of 

 those kinds of radiation which affect the eye as light. 



(2.) That with the form of apparatus employed (a thermopile and 

 g-alvano meter) there is no measurable difference between the dia- 

 thermancy of pure water and of a solution of alum. 



(3.) That the radiation from an Argand gas burner consists of 

 about 1*75 per cent, luminous and 98*25 per cent, non-luminous 

 radiation. 



[Note. — After this paper had been presented to the Royal Society, 

 I was made aware for the first time, by means of a reprint in the 

 December number of ' Wiedemann's Annalen ' (vol. 38, 1889, p. 640), 

 of a paper on "the mechanical equivalent of light," which 0. Tumlirz 

 had communicated to the Vienna Academy in the summer of this 

 year. He states that the total radiation from the amyl acetate lamp 

 he employed contained 2'4 per cent, of luminous radiation. He 

 obtained this result by allowing either the whole radiation from the 

 lamp, or that portion of it which had traversed a glass cell containing 

 water, to fall upon the face of a thermopile, and noting in the two 

 cases the deflections of the needle of a galvanometer in the thermo- 

 electric circuit. — December 27, 1839.] 



*• Observations on the Spark Discharge." By J. Joly, M.A., B.E. 

 Communicated by Prof. G. F. Fitzgerald, F.R.S. Re- 

 ceived June 15, — Read June 20, 1889. 



[Plates 1—5.] 



Path of Discharge over the Surface of a Dielectric. 



The subject of dust-figures produced by electrical discharge has 

 received mach attention at various times. In the following notes I 

 have refrained, to the best of my knowledge, from going over old 

 ground. The subject has been reviewed in Lehmann's recently 

 published ' Molekularphysik.' It is a sufficient substitute for the 

 customary resume of past observations to refer to that work. 



(1.) When a spark discharge occurs in a homogeneous dielectric 

 medium, the path of discharge, as is known, in general lies in a fairly 

 straight line between the points of discharge. If the dielectric 

 medium be heterogeneous in character, the path chosen by the spark 

 will vary with the circumstances. Thus, if the straight line between 

 the conductors be interrupted by a layer of a substance offering a 

 higher resistance to discharge than the surrounding dielectri •, the 



VOL. XLVII. G 



