* TRANSACTIONS OF THE SECTIONS. 39 



dark band which appears in the luminous stratified dischai-ge. Whatever may be 

 the cause ttf the difference in the action of the electrical discharge between the posi- 

 tive and the negative, the disruption of the metal in the latter is merely mechanical; 

 the minute particles are disrupted by the force of the discharge, which at the nega- 

 tive meets with resistance, and which resistance, imder certain conditions,is attended 

 with considerable heating effects ; for if the vdres are thin, the negative invariably 

 fuses, whether the discharges are made in air or in vacuo. 



On a Probable Cause of the Diurnal Variation of Magnetic Dip and Declina- 

 tion. Bif Professor HenisT)sst, F.R.S. 



The author called attention to the researches of Mr. Faraday relative to atmo- 

 spheric magnetism, whereby it appears that vai'iations in the density and tempe- 

 ratm-e of oxygen ai-e always accompanied bj' corresponding variations in its mag- 

 netic properties. Variations in temperature of oiu: atmosphere may occur not 

 only horizontally, but vertically, and they may occur not only between colimins 

 of great extent, but even among extremely small portions of air. This ques- 

 tion had been already submitted to the consideration of the Section in 1858 by 

 Professor Hennessy, and an account of his experiments is contained in the volume 

 published for that year. He had shown that certain abnormal serrations in the 

 thermometrical curve which occur in May and June, and generally_ dm-ing the 

 months of greatest simshine, as exhibited 'by the photographical register kept at 

 the Eadcliffe Observatory, ai-e explicable by the convection of minute ciurents of 

 air. Thus the summer months of greatest simshine con-espond with the period of 

 greatest inequality of temperature between small atmospheric masses. But the 

 same months are also those dm-ing which dim-nal magnetic variation is greatest. 

 The hours of maximum magnetic deviation correspond with the hours of greatest 

 thermometiical serration. This result appeai-s not only from the facts disclosed by 

 Sabine and Hansteen, but also from the facts disclosed in the paper just communi- 

 cated by the Astronomer Royal. 



On Permanent Thermo-Electric Currents in Circuits of one Metal, 

 Bij Fleeming Jexkin. 

 '' In the com-se of some thermo-electrical experunents, I was led to examine the 

 effect of various distributions of heat in circuits formed by one metal. I verified 

 the conclusion arrived at by Professor Magnus, that no distribution or movement 

 of heat in a continuous and homogeneous piece of metal will produce a current of 

 electricity. I also repeated, with some variations, the experiments of Seebeck and 

 Mao-nus, which show that if one end of a wire be heated, the other remaining cold, 

 a momentary or transient ciUTent of electi-icity will be developed when contact is 

 suddenly made between the hot and cold ends ; the du-ection of the cun-ent de- 

 pending on the metal employed. I foimd that I could obtain permanent currents 

 in the same direction from each metal, if I simply looped the two ends of the wire 

 together and heated one of the two loops ; and, moreover, that the cun-ent was 

 usually much greater when there was a loose contact between the two wires, than 

 when the two loops were tightly di-awn together. It is to these currents, due to 

 loose contact between a hot and cold wii-e of the same metal, that I wish to direct 

 the attent'on of the Section. 



I will first shortly describe the apparatus used, and the experiments which 

 showed the existence and importance of these cmTents, and I will then endeavour 

 to repeat some of the experiments before you. I used a retlecting galvanometer of 

 the fomi constructed by Professor W. Thomson of Glasgow. A. very light mirror, 

 attached to a very small mao;net hung inside the galvanometer coil, reflects the 

 light of a lamp upon a scale about two feet ofi". Very small deflections of the mag- 

 net are distinctly shown by the movement of the reflected spot of light, while the 

 slio-ht inertia of" the moving parts has great advantages when rapidly varying cur- 

 rents are to be observ^ed. A common spirit-lamp was used to heat the wires, which 

 were fi-om 0-02 in. to 0-05 in. in diameter. 



When two pieces of similar cojjper wire are connected with a galvanometer, and 

 the end of one wii-e is heated, a momentary current flows from the hot wire across 

 the joint to the cold one whenever th-ey are suddenly brought in contact. While 



