﻿448 Notices respecting New Boohs. 



This differs from the value derived from Wien's formula by the 

 exponential terms. It is also shown that 



C 2 =4D65A Wi fl, approximately, 



where \ m is the wave-length for the maximum value of tlie radiation 

 function at the temperature 0. 



In a study of the current transformer with particular refer- 

 ence to iron bars by P. Gr. Agnew, the question of the meaning 

 of Steinmetz' index is discussed. The number obtained by 

 taking the slope of the logarithmic curve is not the index 

 when the index is variable. This slope is, however, the more 

 important datum in practice, and instead of existing names 

 (such as logarithmic derivative) it is proposed to call it. the ratio 

 of variation. We confess that we think that no name could be 

 chosen more successfully to hide the meaning of the quantity 

 than this. The term logarithmic ratio would be more explicit and 

 could be shortened in actual speaking in the well-known way. A 

 very plausible way of calculating the index is given based on the 

 employment of a point (as standard) at which the " ratio of 

 variation" passes through either a maximum or a minimum. 

 Values of this index are given for various flux densities for trans- 

 formers of ordinary and silicon steel for frequencies of 30 and 60. 

 They range from 1'36 to 1*71 for silicon steel, and from 1'49 to 

 4*71 for ordinary steel. But the method of calculation is in reality 

 based upon an erroneous mathematical deduction on p. 449. The 

 equation d 1 (log y)jd( log xf =0 cannot be integrated because it 

 applies only at one point of the curve. Nor is the ratio of variation 

 equal to the exponent when the former is a maximum or a minimum. 

 The truth is that any constant value of the factor K may be assumed 

 and will yield correct values of the exponent if inserted in the 

 equation y = Kx n . The present writer, taking the same data for 

 the transformer of ordinary steel, finds exponents which range 

 from 1*73 to 1*81 ; and between flux-densities 100 and 16000 the 

 greatest variation in the exponent is only 2-2 per cent. The value 

 taken for —log K was 6-888. 



Prof. H. S. Carhart discusses the thermodynamics of concen- 

 tration cells. In the course of this discussion he criticises Nernst's 

 third law of thermodynamics inasmuch as Nernst's assumption is 

 at variance with the behaviour of concentration cells in which the 

 free energy is proportional to the absolute temperature for very 

 dilute solutions, or is equal to a constant plus a term proportional 

 to the absolute temperature in other cases. Nernst's assumption 

 makes the coefficient of the first power of the absolute temperature 

 equal to zero. Numerous data are given showing that the e.m.f, 

 of such cells is a linear function of the temperature. 



The number also contains papers on the behaviour of high- 

 hoiling mineral oils on heating in the air, on the determination of 

 vanadium in vanadium and chrome-vanadium steels, and on a 

 comparison of American direct-current switchboard voltmeters and 

 ammeters. 



