520 PROFESSOR W. E. AYKTON, MR T. MATHER AND MR. F. E. SMITH : 



tea 1-ohm coils, and ten 0'1-ohm coils in series with a mercury trough of resistance 

 0'12 ohm ; a sliding short-circuiting contact provides the final adjustment, a move- 

 ment of 3 millims. of the slider corresponding to a change in current of 1 in 

 100,000. The manganin coils are- wound on long brass tubes and are immersed 

 in paraffin oil, the capacity of the tank being 6 gallons ; a very constant 

 current was in this way ensured. Under favourable conditions, i.e., when manganin 

 formed by far the greater part of the resistance of the circuit, a current constant to 

 2 in 1,000,000 could be maintained for an hour or more ; when the coils of the 

 balance were in circuit a current steady to about the same limit could be held for 

 a few minutes only. This, however, is all that was desired. 



The potential circuit included the resistance coil R, the cadmium cell S, a contact 

 key, and a galvanometer G. The galvanometer was of the Broca type, having 

 a resistance of 1000 ohms. The controlling field was varied from time to time and 

 hence the sensitiveness was not the same in all of the determinations ; in general 

 a deflection of 5 millims. on the scale (T5 metres distant) corresponded to a change of 

 one-hundred-thousandth of an ampere in the main current. The galvanometer, 

 commutators and all of the auxiliary apparatus belonging to the balance were made 

 by Mr. MURFITT, the instrument maker attached to the National Physical Laboratory. 

 Much of the fitting was also very ably done by Mr. MURFITT. 



The Resistance Coil B, fig. 23, used as a secondary standard (numbered L. 87), is 

 made of thick manganin strip, wound non-inductively on six posts and insulated 

 therefrom by silk ribbon and shellac. The coil was built and annealed by Mr. MELSOM 

 in July, 1905, and its resistance changed very rapidly for many months afterwards ; it 

 is still rising in value. It is provided with potential points and can carry a current of 

 10 amperes without abnormal heating. In July, 1905, the coil was directly compared 

 with the mercury standards of the National Physical Laboratory, and again in March, 

 1906 ; the intermediate and subsequent evaluations were made by comparing it with 

 standard coils. The methods of comparison are described elsewhere.* The tempera- 

 ture coefficient was determined in 1905 and again in March, 1907 ; the mean 

 coefficient for the range 10C.-20C. is +0'0019 per cent, per 1C., but for the 

 reduction of values to a common temperature a resistance chart was used. Owing 

 to the rapid rise in resistance with time the coil was compared with practically 

 constant standard manganin coils on each day a determination of current was made ; 

 the secular change in resistance was thus eliminated as a source of error in the 

 comparison of results. 



The Main Commutator (M, fig. 23) is formed of four brass quadrants of square 

 section and an ebonite turning head carrying two springy copper contact pieces 

 insulated from each other. Connection with the concentric cable is made by 

 drilling two opposite quadrants, one aperture being inch in diameter and the 



* "Methods of High Precision for the Comparison of Resistances," F. E. SMITH, 'B.A. Report,' 

 Section A, 1906. 



