1904] Cotton ly Water and ly Water Vapour. 233 



The telescope was generally so placed that one division of the 

 micrometer scale was equal to 0< 2, and readings were taken to 

 0'02. They may, in the case of Series V, be taken as free from 

 errors as large as 0'05, but the earlier observations are less reliable 

 for the reason stated. The actual temperatures of the immersion 

 bath, as given in the tables, were ascertained by a standard thermo- 

 meter that could be read to 0'01, and in all experiments after 

 Series II these temperatures were maintained constant to within 0> 03 

 by a thermostat. 



The experimental thermometer was provided with a jacket tube 

 of sufficient length to enclose the bulb and part of the stem, and 

 a smooth red-rubber cork was fixed at such a position on the stem 

 that the tube, when in use, was closed and air-tight. All that part 

 of the scale which was used in the observations lay above the level of 

 the cork, and there was sufficient length below it to admit of proper 

 immersion of the bulb and cotton covering without submerging the 

 cork itself. The thermometer was provided with platinum loops 

 so that it could, with its tube, be slung slantwise on the balance for 

 weighing. 



The cotton wool was tightly wound round the bulb so as to 

 efficiently cover it and a small part of the stem, leaving enough room 

 between it and the cork for subsequent operations. From J 1 gramme 

 was found to be a convenient quantity. When once properly wound 

 it remains in position without any special fastening, but for a long 

 series of experiments it is safer to secure it with a single turn of 

 cotton thread. After winding, it should be hung in distilled water 

 for a day to remove soluble impurities derived from the fingers or the 

 laboratory air, and should not afterwards be touched. 



In the latter experiments (Series V) a duplicate thermometer with 

 cork and tube, but without cotton wool, was used as a counterpoise in 

 weighing, and was also put through every treatment in the same way 

 and at the same time as the experimental apparatus. It was hoped 

 by this means to minimise any small errors that might result from 

 hygroscopic changes in the weight of the rubber cork. 



The oven used for drying the cotton covering before any experiment 

 was an ordinary air bath with temperature regulated to about 110 C. 

 and with a fairly free draught passing through it. The bulb of the 

 thermometer passed through an opening in the roof, and a slotted 

 wooden cover supported the rubber cork outside the bath and protected 

 it from the heat. It was proved that an hour's drying was more than 

 enough to give constant weight. 



It may be mentioned here, as a curiously extreme instance of the 

 ordinary wet and dry bulb effect, that the experimental thermometer 

 with its covering, when lifted out of cold water and placed in the oven, 

 rises steadily till it reaches a point some 50 or 60 below the oven 



VOL. LXXIV, T 



