Mr. 11. Phillips on the Magnetism of Pewter Coils. 235 



170. In the experiment with the pewter coil (159.), there is a 

 difference of temperature between the exterior and interior sur- 

 faces of the tube, this difference of temperature being greater at 

 those parts of the tube forming the exterior of the coil than at 

 those portions internally situated. Now the difference of tem- 

 perature will, from a thermo-electric action, throw the surfaces 

 into opposite electrical states, the intensity of which will vary 

 with the difference of temperature ; consequently those parts of 

 the tube which have the greatest difference of temperature will 

 produce electricity, which will circulate by overcoming the resist- 

 ance opposed by those pai*ts where the difference of temperature 

 is less. Suppose a small portion of pipe cut out from the coil 

 in the shape of a right cylinder. Now looking at one end, which 

 gives an annulus, and bisecting this annular space, not the metal 

 itself, by a straight line, and supposing that in one of the halves 

 so obtained the difference of temperature is greater than in the 

 other half; then a chain of the metallic particles producing elec- 

 tricity will cause the current to flow along the curved boundary of 

 the metal until it comes to a part where the difference of tempera- 

 ture is less, and at this point it will more or less force its way across ; 

 thus forming a circuit, the contour of which is some segment of 

 the annulus, and the tube may be regarded as built up of many 

 such annuli. Such a system of currents, it can easily be shown 

 by direct experiments, produces a magnetism parallel in direction 

 with the axis of this bit of tube ; consequently, the system of 

 currents existing in the pewter coils in fact resembles that of the 

 copper helix. 



171. But to account for the regular magnetism of the pewter 

 coils, it is necessary to suppose that the direction of the currents 

 in each part of the pipe is, on the whole, the same with regard 

 to the coil ; and this requires the admission of some force or pro- 

 perty in the coil capable of giving uniform direction to the cur- 

 rents. This view involves the least assumption, nevertheless its 

 application to many of the experiments is exceedingly difficult. 

 If an experiment could be devised in which this magnetism could 

 be developed under circumstances where thermo-electric cmTents 

 could not exist, it would perhaps be necessary to look upon the 

 effect as the direct conversion of heat into magnetism. 



172. The magnetism of these pewter coils affords, as far as it 

 goes, a sufficient explanation of the cause of terrestrial mag- 

 netism ; for the internal heat of the earth continually passing 

 outwards produces electric currents ; then, assuming the exist- 

 ence of a structural force, and which can hardly be unconnected 

 with stratification, capable, as with the pewter coils, of giving 

 regularity to the arrangement of these currents, and the product 

 will be a regular magnetic force, such as is terrestrial magnetism. 

 That Humboldt^s isothermal bands should coincide with Sabine's 



