52 Mr Shaw, On measurement of Temperature, etc. [Nov. 26, 



can be determined to one or two minutes, but in an instrument 

 intended for practical use such a degree of delicacy is not available. 

 One difficulty arises from the depolarizing properties of most speci- 

 mens of heavy glass. Arrangements are in progress for a redeter- 

 mination of the rotation in bisulphide of carbon. 



(2) On the measurement of Temperature by Water-vapour 

 pressure. By W. N. Shaw, M.A. 



[Absteact.] 



The method practically adopted for accurately measuring 

 temperature is to use a mercury thermometer as a thermoscope 

 and express its indications in terms of some accepted standard of 

 temperature by direct or indirect comparison. 



The methods of comparison usually employed are by the air- 

 thermometer, by calibration and repeated determinations of the 

 fixed points and the method of direct comparison adopted at the 

 Kew Observatory. The last is the one which is most generally 

 applicable, but it necessitates the sending of the instrument to 

 Kew, and examples of its application shew that it cannot be relied 

 upon to give greater accuracy than a tenth of a degree. 



A water-steam thermometer, as suggested by Sir W. Thomson, 

 (Ency. Brit. Edit, ix., Art. 'Heat') may be used to give a standard 

 scale of temperature and mercury thermometers compared with it; 

 and instead of measuring directly the saturation-pressure e of 

 water- vapour, it can be calculated by the formula 



e _ l + it f 

 T&)~~^' d' 



from an observation of the weight of water contained in a known 

 volume of air artificially saturated, where f is the quantity of 

 moisture per unit volume of the air, d the specific gravity of water- 

 vapour referred to air at the same temperature and pressure, A the 

 density of air at 0°C. and 760 mm. pressure, and t the temperature 

 of the air (which maybe measured by an uncorrected thermometer) 

 and a the coefficient of expansion. The temperature is then accu- 

 rately given by Regnault's table of tensions of water-vapour at 

 different temperatures and thus the indication of the thermometer 

 to be tested, placed in the saturating space, can be corrected. 



The quantities A and a of the formula are known constants, 

 the value of d was shewn by Kegnault's experiments (Ann. d. 

 Chim. [3] XV.) to be very nearly constant and equal to "622. 



The experimental part of the comparison of any thermometer 

 with a water-steam thermometer by this method consists in the 

 determination of /. This requires only the same experimental 



