NA TURE 



\J\ larch 13, li 



junction if tlie metal is iron. In passing from compressed to 

 stretclied portions the converse effects will occur. 



It follows from the above considerations that if the metal 

 plate (which is acted upon by a force from south to north and is 

 traversed by a current from west to east) be of copper, heat m ill 

 be de\'eloped in the western half of the northern division and 

 absorbed in the eastern half ; while heat will be absorbed in the 

 western half of the southern division and developed in the 

 eastern half. But the resistance of a metal increases vi'ith its 

 temperature. The resistance of the north-western and south- 

 eastern districts of the plate will therefore be greater, and that 

 of the north-eastern and south-western districts smaller than 

 before it was strained ; and an equipotential line through the 

 centre of the plate, which would originally have been parallel 

 to the west and east sides, will now be inclined to them, being 

 apparently rotated in a counter-clockwise direction. 



if the plate were of iron instead of copper, the Peltier effects 

 w ould clearly be reversed, and the equipotential line would be 

 rotated in the opposite direction. 



The peculiar thermo-electric effects of copper, and iron dis- 

 covered by Thomson are thus seen to be sufficient to account for 

 Hall's phenomenon in the case of those metals. It became ex- 

 ceedingly interesting to ascertain whether the above explanation 

 admitted of general application, and the author therefore p-iro- 

 ceeded to repeat '1 homson's experiments upon all the metals 

 mentioned by Hall. The results are given in the following table, 

 where those metals which in Hall's experiments behave like gold 

 are distinguished as negative, and those v\hich behave like iron 

 as positive : — 



It will be seen that in every case, excepting that of aluminium 

 and one out of five specimens of gold there is perfect correspond- 

 ence between the direction of the thermo-electric current and the 

 sign of Hall's effect. With regard to the aluminium, a piece 

 of the foil was mounted on gla^s, and Hall's experiment per- 

 formed with it. As was anticipated, the sign of the "rotational 

 coefficient" was found to be positive like that of iron, zinc, and 

 cobalt. Either, therefore, Mr. Hall fell into some error, or the 

 aluminium \\ith which he worked differed in some respect from 

 that used by the author. The anomalous specimen of gold, 

 being in the form of wire, could not be submitted to the same 

 test. It probably contained some di^lurbing impurity. 



It is submitted that the considerations and experiments above 

 detailed render it aljundantly evident that the phenomenon de- 

 scribed by Mr. Hall involves no new law of nature, but is merely 

 a consequence of certain thermo-electric effects which had been 

 observed nearly thirty yeais ago. 



"Some Relations of Chemical Corrosion to Voltaic Current." 

 By G. Gore, F.K.S., LL.D. 



The author states that the chief object of this research was 

 to ascertain the amounts of voltaic current produced by the 

 chemical corrosion of known weights of various metals in 

 different liquids, and to throw some light upon the conditions 

 which determine the entire conversion of potential molecular 

 energy into external (< e. available) electric current. The metals 

 used were magnesium, zinc, cadmium, tin, lead, aluminium, 

 ' S meins stretched i U means unitretched. 



iron, nickel, copper, and silver ; some of them being also 

 used in an amalgamated stale. The liquids employed were 

 .solutions of nitric, hydrofluoric, hydrochloric, sulphuric, fluo- 

 silicic, and acetic acids ; and potassic hydrate and cyanide, also 

 of different degrees of strength. 



The chief numerical results are given in a series of ten tables, 

 a table for each metal. Each table contains the electromotive of 

 the current, the lo.ss and rate of loss of the corroded metal, and 

 of a comparison sheet not producing a current ; and the percentage 

 of current obtained in ninety-seven different cases. 



The results show that the proportion of loss of the positive 

 plate by "local action" to that by corrosion producing external 

 current varied greatly in different cases, viz. from I'jto 95'25 

 per cent. In no case was the whole of the metal dissolved by 

 "local action," nor did the whole of the corrosion produce 

 external current. In about 6 per cent, of the cases the com- 

 parison plate was more corroded thnn the one which w as used to 

 produce a current. Whilst also the contact of a negative metal 

 w ith the corroding plate usually increased the total corrosion, it 

 commonly decreased the corrosion due to " local action." 



The proportion of corrosion attended by external current to 

 that due to " local action," varied with the kind of metal and of 

 liquid ; with cadmium it averaged 75'63, and with copper 3033 

 percent, of the total corrosi^m ; with solution of potassic cyanide it 

 averaged 63*27, and viith dilute nitric acid 31 '14 percent. It 

 varied also with other conditions ; and the kind of metal had 

 more influence than that of the liquid. Amalgamation of the 

 metal also had distinct effects upon the proportion, but opposite 

 in different cases. The rate of total corrosion of the positive 

 plate appe.ared to be related to the degree of electromotive force 

 of the current. The chief cause of the grtat variation in the 

 proportion of corrosion by "local action'* to that producing 

 external cuiTent was probably a variation of electric conduction 

 resistance. 



March 6.- — "Magnetic Polarity and Neutrality." By Prof. 

 D. E. Hughes, F.R.S. 



The author, citing the researches of Page, Marianini, Wertheim, 

 Joule, Wiedemann, De la Rive, Weber, Beetz, and Maxwell, to- 

 gether with his o\\"n published researches, demonstrating that the 

 molecules of magnetic bodies, such as iron, have inherent polarity, 

 and that all the known effects of magnetism can be explained by 

 the demonstrable rotation of the molecules whenever a change of 

 polarity occurs, now gives a new series of experiments verified 

 by several independent methods, in which he shows that the 

 penetration of the apparent polarity diminishes rapidly from the 

 exterior to the interior of a bar, due to the frictional resistance 

 of its molecules. In rotation, as when the rod or bar is vibrated 

 whilst under the exciting induence, the penetration is four times 

 greater than previously. In all cases, however, there is no re- 

 ver>al of polarity in the interior whilst under the influence of its 

 exciting cause. The instant this is withdrawn neutrality takes 

 ]5lace in soft iron, or a partial return to the same state even in the 

 hardest of steel. 



The author has discovered that this neutrality is not caused by 

 a mixing of the fluids as assumed by Coulomb, or a heterogeneous 

 arrangement of the molecules as assumed by AmpC-re and all 

 other theories up to the present time, but that a reaction takes 

 place between the outside or strongest polarity with that of the 

 weaker inside, completely reversing it to a remarkable extent. 



A bar of iron under the influence of its exciting cause may be 

 represented by three series of letters, the centre representing the 



; N N N ; 



inside of a bar, thus — ■ ■ , but when this influence is 



: s s s : 



withdrawn v\ e should have- 



N S N : 

 S N S 



And if the inner 



reversed polarity exactly balanced the exterior the sum of both 

 would be zero, and consequent neutrality. 



The paper describes the methods emjiloyed, and gives diagrams 

 of these curves. In certain cases the exterior becomes reversed, 

 as shown by magnetising a soft strip of steel half a millimetre 

 thick, and then reducing it to a nearly j erfect neutral state, cither 

 by mechanical vibrations, or by heating the strip to red heat. 

 That the outside is reversed is shown by dissolving the exterior 

 n dilute nitric acid, when its previous polarity reappears. 



