472 



SCIENCE 



[N. S. Vol. XXXIII. No. 847 



Per cent. 



Temp. Obs'd E.M.F. Comp'd E.M.F. difference 



11.1 0.007300 0.007307 + 0.10 



15.4 7444 7439 —0.07 



19.8 7574 7575 + 0.01 



24.6 7720 7723 + 0.04 



29.4 7870 7871 +0.01 



32.8 7983 7976 —0.09 



36.6 8086 8094 + 0.10 



42.0 8262 8259 —0.04 



47.0 8417 8414 —0.04 



The greatest difference between the observed 

 values and those computed by the above equation 

 is 0.008 millivolt. 



The ZnSOi solution was then replaced by 

 ZnCl^ solution and the measurements were re- 

 peated. The results are best represented by the 

 same equation within the limits of temperature 

 10°.l and 49°.3 within which the observations 

 were made. Moreover, there is no break at the 

 transition point of ZnSO^ at 39°.0., 



This relation is strictly linear and is directly 

 and conclusively opposed to Nemst's assumption 

 that the constant a and o' are both necessarily 

 zero. 



Upon the Construction of the Wheatstone Bridge 

 for Besistanee Thermometers: Professor C. F. 

 Marvin, of the U. S. Weather Bureau. 

 The speaker mentioned the well-known fact that 

 the resistance of metals such as platinum, nickel, 

 etc., commonly employed in the construction of 

 resistance thermometers, does not change with 

 temperature according to a strictly linear law of 

 relation, therefore, the scales of temperature ob- 

 tamed directly from resistance thermometers are 

 not the same as the standard scale of temperature 

 by the gas thermometer. The object of the paper 

 was to call attention to certain interesting mathe- 

 matical relations between the bridge equations 

 and those for platinum and nickel resistances, 

 which, if availed of, enable the manufacturer to 

 give the arms of the bridge such resistances that 

 the indicated temperatures on the bridge scale 

 correspond to the true temperature of the ther- 

 mometer on the gas scale within a few hundredths 

 of a degree over ordinary meteorological ranges 

 of temperature, say from — 40° to + 60° Centi- 

 grade. This result is obtained, moreover, when 

 the subdivisions of the bridge scale are exactly 

 equal throughout, and when the intervals of re- 

 sistance on the bridge wire, or equivalent device, 

 are made as exactly equal to each other as pos- 

 sible. Numerical data shown for nickel indicated 

 that the logarithmic equation : B^a-\- bt fitted 

 the temperature resistance changes of some sam- 



ples of nickel. Other samples, however, require 

 an equation of three terms, viz: 



Log. Iiz=a -{- bt — ct'. 



Diagrams and equations of bridges with the re- 

 sistance placed in series with either the thermom- 

 eter in the one case, or the balancing coU in the 

 other, were explained. 



In all cases the bridge equations reduce to the 

 general form: M= {A -\- Mt/M ± t) • in which 

 A, M and N are constants fixing the numerical 

 value of the resistances in the arms of the bridge. 

 The plus ( + ) sign applies when the shunted 

 rheostat is in series with the balancing coU, and 

 the equation then represents a curve, mathemat- 

 ically similar to a parabola, or a curve convex 

 upward. The equation with the minus ( — ) sign 

 is required for the ordinary slide-wire connections, 

 and with the shunted rheostat in series with the 

 thermometer. In these cases the curve conforms 

 very closely to the logarithmic or other curve 

 concave upward. 



By computing the constants of the bridge equa- 

 tion from three points taken from the correspond- 

 ing temperature resistance curve for the platinum 

 or other thermometer that may be used, and then 

 adjusting the resistances in the bridge accord- 

 ingly, the indicated temperatures on the bridge 

 are either identical with the gas-scale tempera- 

 tures, or differ therefrom by only a few hun- 

 dredths of a degree at the extremities of the 

 range of eighty to one hundred degrees. 



Numerical examples were worked out and the 

 resistances computed for the arms of the bridge 

 to fit a platinum and a nickel thermometer. It 

 was mentioned in the discussion that all of the 

 equations apply equally to the case in which the 

 differential galvanometer is employed instead of 

 the bridge to determine the resistance of the 

 thermometer. 



It is expected that this paper in full will be 

 published in the Journal of the Franlclin Institute. 



(The abstracts of the two foregoing papers are 

 by their authors.) 



Professor Cleveland Abbe, of the TJ. S. Weather 

 Bureau, spoke informally concerning the altitude 

 of the aurora, describing briefly Professor Stoer- 

 mer's recent successful method of measuring the 

 altitude of an aurora, which consists essentially in 

 simultaneously photographing the aurora from 

 two points of known distance apart, some known 

 star being also simultaneously photographed on 

 the plates from which measurements can be made. 

 E. L. Faris, 

 Secretary 



