GEOPHYSICAI, LABORATORY. 93 



irregularities were discovered, small in amount and as yet unexplained, but 

 apparently pointing to contamination of a different kind from that due to 

 iridium, which forms the chief source of error in the use of thermoelements 

 in platinum furnaces in an oxidizing atmosphere. 



(3) Thermokraftfreie Kompensationsapparate mit Kleinem Widerstand und Konstan- 



ter Galvanometerempfindlichkeit. Walter P. White. Zeitschr. f. Instr. 

 (Berlin), 27, pp. 210-219, 1907. 



Constant galvanometer sensibility can be attained for a wide range of 

 electromotive forces with considerable ease by a suitably constructed potentio- 

 meter, and when attained contributes greatly to the facility and rapidity with 

 which measurements of current and electromotive force are made. Such a 

 potentiometer is described in this article. The results are secured by the 

 use of compensating coils and, in instruments of especially great range, of 

 shunted resistances, a device which has already come into use in other kinds 

 of electrical measin-ements. The potentiometer described, besides giving con- 

 stant galvanometer sensibility, is less liable to certain other errors than any of 

 the other types of equal range. It is not more difficult to build or apprecia- 

 bly more complicated than the w-ell-know^i standard instruments. 



(4) Some new measurements with the gas-thermometer (abstract). J. K. Clement. 



Phys. Rev., 24, p. 531, 1907. 



A new gas-thermometer has been designed and built in the Geophysical 

 Laboratory, for the measurement of high temperatures, in which a con- 

 scientious effort has been made to reduce the magnitude of the errors com- 

 mon to the well-known European gas-thermometers. The constant-volume 

 principle and the thermoelement as the medium of comparison are retained, 

 but the sensibility of the instrument has been increased until the error of an 

 individual reading is no greater than the error in the corresponding barom- 

 eter reading. The bulb ( 10 per cent iridium, 90 per cent platinum, 200 c. c. 

 capacity) is suspended in an air-tight bomb, by means of which the pressure 

 outside of the bulb is maintained nearly equal to the inside pressure. This 

 makes it possible to use a high initial pressure — about 300 mm. Hg. — without 

 deforming the bulb at high temperatures, and thus to obtain a sensibility of 

 about I mm. of the manometer scale per degree centigrade without dimin- 

 ishing the range of the instrument. 



By taking advantage of the fact that the increased viscosity of the expand- 

 ing medium (nitrogen) due to rise in temperature affects only the heated 

 portion of the dead space connecting the bulb with the manometer, it has 

 been found possible to reduce the diameter of the cold portion of the capil- 

 lary connecting-tube to 0.5 mm-, so that the ratio of the volume of unheated 

 space to the volume of the bulb, v, /F = 0.0015. fHolborn & Day. v, /V 

 =: 0.0046; Jacquerod & Perrot, v, /F = 0.0180.) The entire correction for 

 the unheated space is therefore only 3.8° at 1,000° and 5° at 1,200°. 



Perhaps the most important source of error which has been overcome is 

 the variation of temperature along the bulb itself, due to the cooling at the 

 ends of the furnace tube. By the use of auxiliary heating-coils at either end 

 of the main coil, the gradient between the center of the bulb and either end 

 of it has been made less than 0.5°. 



We have made a number of comparisons between this instrument and 

 Pt-PtRh thermocouples from 300° to 1,200° at intervals of 25°, in which 

 the aggregate error from all sources appears to be well within 1°. There 

 has not yet been an opportunity to make a complete series of melting-point 



