Makch 16, 1900.] 



SCIENCE. 



407 



a number of mercury copies will also be 

 taken up. The construction of primary 

 mercury standards, which is of funda- 

 mental importance, not only from a scien- 

 tific standpoint but also on account of the 

 legal questions which will surely arise, will 

 be undertaken as soon as time permits. 



The OfiSce has acquired in addition to the 

 unit coils, standards of the following de- 

 nominations : 



1 2 



1 5. 



4 10, 



4 100. 



3 1,000. 



2 10,000. 



2 100,000 



3 0. 



3 0.01 



3 0.001.!. 



3 0.0001.. 



The temperature coefficients of these coils 

 were first carefully determined, then the 

 coils of the same denomination were inter- 

 compared and the observations reduced to 

 difterences at 20° C. 



The next step consisted in determining 

 the resistance of the multiples and sub- 

 multiples in terms of the unit. Thus, two 

 unit coils placed in series by means of a 

 connecting link of known resistance were 

 compared with the two ohm coil. The five 

 and ten ohm coils were similarly evaluated 

 in terms of the Unit. From the known ratio 

 of a 10 ohm coil to one of the units, the step 

 can be made to the 100 ohm coils by means 

 of a second ten ohm coil also known in 

 terms of the unit. Similarly the values of 

 the coils of still higher value were deter- 

 mined and those of the sub-multiples of the 

 unit, by a slightly modified method. 



THE METHOD OF COMPARISON ADOPTED. 



The practice in this country and in 

 England has been overwhelmingly in favor 

 of the Carey Foster method, but the con- 

 struction of the Reichsanstalt standards, 



with their terminals 16 cm. apart, makes the 

 use of the Carey Foster Bridge almost out 

 of the question. The design of a suitable 

 bridge capable of comparing coils of this 

 tj'pe with those having a different distance 

 between the terminals, introduces still fur- 

 ther complications. Moreover, in the Carey 

 Foster method, additional resistances are 

 introduced in the mercury cups of the com- 

 mutator which are only eliminated in a 

 perfect mechanical construction. Besides 

 the resistances connecting the coils to the 

 commutator, unless equal, are not elimi- 

 nated. While the Carey Foster method is 

 at first glance superior in being a zero 

 method, the Wheatstone Kelvin Bridge, 

 more simple and far less expensive in 

 construction, excels the former especially 

 where the intercomparison of low resistance 

 standards is concerned, provided the coils 

 to be compared are first made nearly equal 

 by means of a shunt of known value applied 

 to the greater. The value of the shunt 

 need not even be known to a high degree 

 of accuracy in case of fairly well adjusted 

 coils. 



The coils are connected to one another 

 by copper forgings 1.5 cm. thick and 2 cm. 

 wide, having therefore a resistance of about 

 0.6 micro-ohm per cm. of length, their ter- 

 minals resting in mercury cups. To permit 

 the comparison of coils differing widely from 

 the standards, provisions have been made 

 to enable one of the coils compared, to be 

 placed in parallel with an accurately known 

 coil by means of a second pair of mercury 

 cups. 



The sources of error characteristic of the 

 direct deflection method are due to the fol- 

 lowing causes : 



(1) Variation of E. M. F. of test bat- 

 tery. 



(2) Variation of galvanometer sensibil- 

 ity, due either to the variation of propor- 

 tionality between deflection and current, or 

 to a change in the actual sensibility. 



