536 Prof. H. L. Callendar on a 



observations of Mr. Barnes on the Variation of the Specific 

 Heat of Water, in which the fundamental measm*ement of a 

 difference of temperature of 8° to 10° was obtained by a single 

 reading with a pair of differential platinum-thermometers 

 under conditions which precluded the employment of any 

 other method. 



Construction of Standard Thermometers. 



In considering the construction of standard thermometers, 

 there are many details which remain to be settled by the Com- 

 mittee, but which may be profitably discussed at the present 

 stage for the sake of obtaining suggestions. Although, theore- 

 tically, a single platinum-thermometer may be made to cover 

 the whole range, it is for many reasons desirable to construct 

 a series of such thermometers of different patterns for different 

 purposes. For use at low temperatures from 0° to 500° C, 

 the instruments are most conveniently made with containing- 

 tubes of glass, and adjusted to have a fundamental interval of 

 5 or 10 ohms. With a resistance of this order at ordinary 

 temperatures, it is possible to minimize the risk of error from 

 variable contacts, without, on the other hand, incurring 

 trouble from imperfect insulation. At higher temperatures 

 it is better to employ porcelain tubes, and to use a fundamental 

 interval of 1 or 2 ohms, because the same accuracy of measure- 

 ment, and the same perfection of insulation, are not attainable 

 at the higher points of the scale. A tenth of a degree at 

 1000° implies the same order of relative accuracy of measure- 

 ment as a hundredth of a degree at 100° (J.; but it is not so 

 easily attained, because the conditions are never so steady 

 and the insulation cannot be made so perfect. There is no 

 loss of accuracy, but rather a gain, in using the smaller funda- 

 mental interval for the higher points. At a temperature of 

 about 1100° C, the glaze on the porcelain tubes becomes 

 sticky, and the mica frames on which the wire is usually 

 wound soften and swell as they approach the fusing-point of 

 the material. The insulation deteriorates, and the wire is 

 strained as it cools. A pyrometer of this pattern, when it 

 has been in use for some hours at temperatures between 1000 Q 

 and 1100°, may often show a rise of zero of about a degree 

 due to strain. The effect is naturally more marked if the 

 instrument has been heated above 1100° and repeatedly cooled. 

 The best mercury thermometers show a temporary depression 

 of zero of the same order (1 in 1000) after each heating to 

 100° C; but the effect in this case is more troublesome because 

 the glass slowly recovers, and the reading at any time 

 depends on the previous treatment of the instrument in a 



