Construction of Platinum Thermometers. 109 



these instruments depends largely on the adjustment of the 

 galvanometer and on the perfection of the optical means for 

 reading its deflexions. A convenient method is to use a 

 microscope * with a micrometer-eyepiece, and to observe the 

 reflected image in the galvanometer-mirror of a scale attached 

 to the objective. The power of the microscope in magnifying 

 the deflexions is proportional to the length of its body and to 

 the power of the eyepiece, and does not depend on the focal 

 length of the objective. A telescope does equally well, but 

 the arrangement is then less compact. Either method is 

 superior to the lamp and scale in point of accuracy and for 

 other reasons, particularly in not requiring a darkened room. 

 This is a great advantage, because a good light is preferable 

 for reading the scale of the bridge-wire and noting down the 

 observations. 



With regard to the formula to be used for reducing the 

 observed platinum-temperature pt to the temperature t by air- 

 thermometer, the very careful experiments made by Griffiths 

 on the boiling- and freezing-points of various substances t 

 seem to show that the simple parabolic formula 



*-jrt = 8{*/100| 2 -*/100} (d) 



holds over a wide range even more accurately than I had 

 previously imagined. The values of t for various boiling- and 

 freezing-points, deduced by the aid of this formula from 

 observations with thermometers of very different patterns and 

 with different coefficients, were rarely found to differ from 

 each other by more than a few hundredths of a degree over 

 the range 0°-450° C. The value of the constant 8 for any 

 thermometer may be readily deduced from a single observation 

 of the boiling-point of sulphur or mercury as described in the 

 paper above referred to J. 



The air-thermometer experiments § on which this formula 

 was founded did not extend beyond 650° C, but I have 

 recently succeeded in verifying it roughly at a higher tempe- 

 rature by an observation of the freezing-point of silver. This 

 point appears to be very clearly marked and well adapted for 

 a high-temperature standard. Several independent observa- 

 tions with a platinum thermometer gave the same result, 

 982°*1 C, within a small fraction of a degree. The specimen 

 of silver, however, was not perfectly pure, so that the result 



* I owe this suggestion to Mr. Horace Darwin of the Instrument Co., 

 Cambridge, to whom I am also much indebted for the care and skill 

 bestowed on the construction of my apparatus. 



t Phil. Trans. 1891, A. pp. 43 & 148. 



% Phil. Trans. 1891, A. p. 146. § Phil. Trans. 1887, A. p. 161. 



