32 RESEAKCHES ON II. 



holes or cups, and so one or more of the coils B, R' was easily excluded or included. 

 To avoid the errors of a longer or shorter part of wire immersed in the quicksilverj 

 the ends of the wires were made a good deal thicker than the rest. 



The rheostat used in the experiments on the globe had a brass wire 0.0078 of 

 an inch in diameter (o™r). That used in the experiments on the circle had a cop- 

 per wire 0.012 of an inch diameter {o^T)- It was chosen not very thin in order to 

 avoid a part of the errors arising from the elevation of the temperature, and from 

 small differences in the measure of length. On the contrary, had a thicker wire 

 been used, it would have been necessary to introduce too great length of coil, and 

 thus render the operations troublesome. 



A rheostat joined with a galvanometer is a more convenient apparatus for 

 measuring the forces of the currents than any other means used by the German or 

 French philosophers for the same purpose. It is very regular in its action, and it 

 is also very easy to discover by it the slightest variations of strength in the current, 

 which would pass almost without observation, or certainly without exact appre- 

 ciation with the galvanometer alone. If, for instance, the variation of a degree 

 require ten turns of the rheostat, it is certain that the motion made by only a 

 quarter of a turn will be easily perceived, and so ^^th of a degree will be easily 

 appreciated, and even smaller fractions by means of the microscope, and this by 

 using only common galvanometers, with a divided circle three or four inches in 

 diameter; but by using an apparatus on a larger scale, more accuracy can be 

 obtained. Thus the errors of measure are all rejected on the rheostat and on the 

 resistances. When the total resistances of the circuit are small and short, the 

 variation of the needle produced by the tenth of a turn is perceptible ; but when 

 resistances are very great, as for instance 200 or 300 turns, the variation of the 

 needle produced by a quantity less than a turn is scarcely perceptible without the 

 use of the microscope, but with it the variations of half a turn are sensible also in 

 this case. 



The method of determining the absolute resistance of coils and other wires 

 entering the circuit, is the same as used by Wheatstone : the wire, whose resistance 

 is required, is introduced into the circuit, the rheostat being near zero. The 

 position taken by the needle is then observed carefully, and by the rheostat is 

 reduced to a marked division. Then a bridge of thick copper ribbon is thrown 

 across the two cups in which the ends of the wire are immersed; the needle of the 

 galvanometer rises immediately, the resistance being now an unappreciable one, but 

 coilino- more wire on the wooden cylinder of the rheostat, it is reduced to the 

 same degree as it was before : the number of turns to be introduced to obtain this 

 constitutes the reduced length of that coil. It is necessary to have at least two of 

 these coils whose resistance is less than that of the rheostat by a few turns, and 

 two or three others equivalent to twice or thrice that resistance. By the same 

 means the resistance of a galvanometer is also determined, using another galvano- 

 meter to measure the deviations ; but we shall see hereafter a case in which only a 

 single galvanometer is required to obtain the same result. 



Thermo-electric piles were also used to appreciate better the resistance of wires, 

 and the results were identical with those obtained with voltaic combinations. The 



