84 



and secondary. Then from the various dimensions of the ring and the calibrat- 

 ing coil as regards number of turns of the primary and secondary, cross-section 

 and length of core, intensity of primary currents and throws of the galvanometer, 

 the permeability of the specimen can be calculated. 



The objections to the above method are the tediousness of observing the con- 

 stants (about a dozen in all) and making the calculation therefrom, and, furthei', 

 the inaccuracies involved in assuming the areas of windings and core to be the 

 same, in neglecting the difTerence in closeness of winding between the inside and 

 outside of the ring, etc. 



For the last two years the author has recommended the following method to 

 his students. An exact non-magnetic copy of the ring specimen is made in the 

 form of a plaster of Paris cast therefrom. This cast is wound precisely similarly 

 to the iron ring. The permeability is then simply the ratio of the throws given 

 by the iron ring and the plaster of Paris ring on making or breaking equal 

 currents in the primaries. The calculations are thus greatly simplified and the 

 inaccuracies involved in the above mentioned assumptions are greatly reduced 

 and can be completely eliminated by winding the primaries and secondaries in 

 alternate turns on the core. It is not claimed that by this method the galvano- 

 meter is more exactly calibrated, but it is calibrated under the exact conditions 

 under which the actual measurements on the specimen of iron are made. It is 

 calibrated, in fact, by the actual windings on the ring specimen, the iron core 

 being replaced by a non-magnetic core. 



With a view to testing the sum total of the errors inherent in the ordinary 

 ring method, simultaneous determinations of the permeability of the same speci- 

 men were made by the two methods. It was found that the total error involved 

 in the use of the ordinary calibrating coil was often large, amounting in some 

 cases to as much as thirty-eight per cent. 



Empirical Formula for the Temperature Variation of Viscosity. By 



A. WiLMER Duff. 



[Abstract.] 

 A careful determination was made of the viscosity of glycerine between zero 

 and thirty degrees. The method employed depended on Stokes' formula for 

 the rate of descent of a sphere through a viscous liquid. Several different forms 

 of formula have been proposed for the representation of this temperature varia- 

 tion. It was shown that nine of these would apply throughout a wide range of 



