Dec. 26, 1878] 



NATURE 



185 



the equation of a hyperbola whose semi-transverse axis is '1025 

 centimetres, and semi-conjugate axis 6*8623 C.G.S. units. 

 From the above equation we infer that — 



R = 66-94 V I I + •205 - I 



where /* denotes the electrostatic force; from which it is evi- 

 dent that as s becomes smaller R becomes greater. A similar 

 curve was obtained when hydrogen was substituted for air. 



When the disks were heated before taking the sparks, the 

 curve obtained satisfies the equation — 



V - 87-04? - I9'S6j', 



a parabola, from which we deduce — 



J? — 87-04 - 19-56^. 



It was found that when the capacity of the charged conductor 

 was changed, the difiference of potential required to produce a 

 spark remained constant. 



When the discharge was continued so as to keep the spot of 

 light at a fixed deflection, the reading was always less than for 

 the corresponding single discharge, but the curves were similar. 



Readings were taken of the difference of potential required to 

 produce a "5 centimetre spark through air at different pressures 

 from the atmospheric to 20 mm. They give — 



v= -0458 v{/' + 203/} 



where / denotes the pressure in millimetres of mercury. 



The electric strengths of several gases were determined by 

 comparing the differences of potential required to pass a "5 

 centimetre spark through the gas at the atmospheric pressure. 



" Electric strength" is the term used by Prof, Clerk-Maxwell 

 to denote the physical constant in question. I have added, for 

 the sake of comparison, values deduced from the results of 

 De la Rue and Miiller and of Faraday, but the ratios given do 

 not strictly give the relative electric strength, but the ratio of the 

 lengths of spark when the difference of potential is kept constant. 



The difference of potential required to produce a spark 

 between two spherical balls is approximately proportional to 

 the square root of the length of the spark. This we have veri- 

 fied up to 15 cm. 



On proceeding to investigate the discharge through insulating 

 liquids, we first took up oil of turpentine. The liquid was placed 

 in a glass jar of 7 inches diameter and 5 inches height. A screw 

 passing through the bottom of the jar served to fix the lower 

 electrode, and also to afford conducting connection with the 

 earth. W^e observed four modes of discharge : by means of 

 threads of solid particles, by motion of the liquid, by a dis- 

 ruptive discharge, and by motion of gas bubbles. When a 

 chain was formed the index of the electrometer behaved as if a 

 current were passing. The discharge, when sufficiently great, 

 broke the thread and turned into a spark. The liquid was more 

 easily set in motion when its surface was not much higher than 

 the upper plate. The bubbles of gas appeared to be formed by 

 the passing of the spark. They were always attracted to the 

 negative electrode. When the electrification was neutralised 

 they of course adhered to the under surface of the upper disk ; 

 wheri the disk was electrified negatively they still adhered ; when 

 positively they were repelled so as as to remain suspended in the 

 liquid or to adhere to the lower electrode, according to the 

 greater or less distance between the electrodes. At a dimin- 

 ished pressure the bubbles produced at the upper surface were 

 observed to effect the discharge by carrying the electricity with 

 tiiem to the negative electrode. The fact that it is possible 

 to cause a shower of electrified bubbles to descend and pro- 

 duce a flash and sound on impinging on the lower surface 

 appears to throw some light upon the nature of lightning balls. 



Similar phenomena were obser\'ed in paraffin oil, excepting 

 that the gas bubbles produced were generally attracted to the 

 positive surface. 



We observed the differences of potential required to pass a 

 spark through paraffin oil and oil of turpentine between plates 

 for distances up to *5 cm. It was impossible to observe for 

 greater distances, as our insulated wire allowed the charge to 

 escape. For paraffin oil, 



F=75oj - IS; 

 therefore, ^ = 750. 



The above has not been reduced to absolute measure. Thus 

 R is constant in the case of the liquids, but variable in the case 

 of the gases. 



Electric Strength of Liquid Dielectrics 



Air I 



Paraffin Oil (kind used for burning) 4 



Oil of Turpentine ... ... 3'7 



Sparks were taken between two platinum wures placed at 

 right angles to one another. When one of the wires was heated 

 by a voltaic current the electrometer deflection was diminished 

 by afeout one-fourth of its amount. 



W^e have also investigated the effect upon the electric spark of 

 heating the air round the disks, the pressure being kept con- 

 stant. The deflections of the electrometer for a constant spark 

 for temperatures from 20° C. to 280° C. indicate a curve which 

 slopes down gradually as the temperature is increased, while 

 the deflections during cooling give a ciu:\-e which is somewhat 

 lower at the lower temperatures. 



These experiments were made in Prof. Tait's laboratory, to 

 whom we are indebted, not only for the use of apparatus, but 

 also for ever ready advice. 



SCIENTIFIC SERIALS 



Annalen der Physik und Chetnie, No. lo. — The loss of electricity 

 by an insulated charged body in rarefied gas in an envelope that 

 has conductive connection with the earth is here stated by Herr 

 Narr to be due to two processes distinct in time and intensity, 

 the first, one of outflow, rapid and intense, the other, one of 

 dispersion, slow and weak. The intensity of the former increases 

 with decreasing density of each of the gases used (COo air and 

 H), and also on substituting one gas for the other in the order 

 just given, the density remaining constant. These differences 

 between the gases decrease with the density, and in vacuum fal 

 within the limits of errors of observation. In discussing these 

 results, Herr Narr is led to regard the condensed layer of gas on 

 the conducting system as an insulator, not as a conductor. — Dr. 

 Holz finds that the specific magnetism of magnetic ironstone is 

 the greatest of all magnetic substances hitherto examined. Its 

 maximum permanent magnetism is nearly as great, and partly 

 greater than that of steel as hard as glass. Its permanent magnetism 

 is sooner removed in demagnetisation with the same external 

 forces than that of steel, &c. — Dr. Strouhal enunciates the laws 

 of a mode of sound-production not much studied hitherto, that, 

 viz. , of rapid swinging of a rod, a blade, or the like, in air, or 

 the passage of air-currents over strong wires or sharp edges, &c. 

 — Herr Braun contributes a long and [interesting paper on the 

 development of electricity as equivalent of chemical processes. — 

 Herr Koch demonstrates the applicability of the method of deter- 

 mining coefficients of elasticity from the bending of short bars 

 supported at the two ends, the sinking in the middle being mea- 

 sured by means of Newton's interference-bands, and he suggests 

 a more thorough investigation of the elasticity of crystals, by the 

 improved means he describes. — Some remarks on the atomic 

 weight of antimony, with reference to Cooke's recent research, 

 are communicated by Herr Schneider. 



American Journal of Science and Arts, November. — In the 

 opening paper Prof. Dana considei^ the value of some distinc- 

 tive characters generally accepted in defining certain kinds of 

 rocks, as, "older and younger," foliated or not, and porphy- 

 ritic structure ; showing them to be often trivial and inapplicable. 

 — With regard to the relative agency of glaciers and sub-glacial 

 streams in the erosion of valleys. Prof. Miles considers that the 

 streams are of primary importance in working in advance of 

 the ice in deepening and enlarging these valleys, and that the 

 glaciers abrade, modify, and reduce the prominent portions left 

 by the streams, and give them the well-known glaciated sur- 



