II. ELECTRICAL RHEOMETRY. 33 



thermo-electric pile belonged to a complete and very excellent apparatus of Melloni. 

 The difficulty, however, of keeping a perfectly constant temperature in these piles, 

 when made only for i-adiant heat, renders this way of finding resistances more 

 difficult than the other, although they are very convenient on account of their small 

 interior resistance. When sufficient care is used, the results cannot differ more 

 than a tenth of a turn, or half a turn when resistances are very long. 



The diameter of wires in the rheostat has been already given ; I will only add 

 that each turn of the rheostat used for the experiments made with the globe was 

 4.73 inches long (12™""'"), and that used for the circle 5.885 inches (14.9884'^=°"'"). 



IV. Circle and Ohhe. The circle on which the current is disposed was made of 

 a wooden ring, half an inch thick, one broad, and 17.7 inches (45™"'™) in diameter. 

 A greater thickness to make it stronger was not allowed by reason of the small 

 distance from the currents at which the needle was to be settled. I observed, 

 however, that small changes of figure did not alter sensibly the results. The 

 sphere or globe also, which was enveloped by the wires, was likewise made with 

 thin wooden circles, crossing one another at the poles under an angle of 30°, and 

 connected at the equator with a full circle like the circles in armillary spheres. 

 Several other small arcs of circles representing the parallel of a sphere, connected 

 the two extreme meridians, so as to make a frame of two lunary surfaces^ opposite 

 to one another 60° wide. On this armature the copper wire was coiled, covered 

 with silk. It was about 330 feet long (nearly 100 metres), 0.036 of an inch in 

 diameter (0.92"'"'), making 68 turns, all crossing each other at the poles, and cover- 

 ing with the currents the whole of the above-mentioned surface. Both the radius 

 of the circle and of the globe were equal, and divided into 10 equal parts. Some 

 difficulty was found in the determination of the length of the polar diameter of the 

 globe, on account of having to ascertain what point should be taken, as the mathe- 

 matical pole or centre of action in the wires crossing at the poles and forming a 

 bundle 6°™- thick (0.236 in.). But it was evident that it could not fall far from 

 the surface of the sphere. They were spread on that surface as soon as they had 

 crossed one another, and it was therefore supposed to be at M of the said thickness, 

 which quantity of 2™°- added to the radius of the globe constituted the total radius 

 used in the calculations 22.5™°'™- = 8.8585 inches. A true appreciation of the 

 place of the real pole of all the currents surrounding the globe is a point of the 

 utmost importance, as will be seen hereafter. 



In the circle I found no such difficulty, having only used a wire making 38 

 turns, disposed in two layers occupying 0.38 inches in breadth (9.65™°'), the 

 diameter of the wire when uncovered being only Voth of an inch. Half the 

 thickness of both the layers was also considered in the measure of the radius of 

 the circle. 



The circle and the globe were both fixed upon a steady triangular stand, as is 

 seen in Fig. IV., on which they could take all the positions required for the experi- 

 ments. The stand could be made level by means of three brass screws S, S, S. The 



* Fuseaux, in French. 



