SMITH. — EXPANSION OF ETHER AND ALCOHOL. 433 



The resistance at 0° C. of the platinum thermometer which was in 

 the dilatometer was found to be 24.528 ohms. 



For the measurement of the change in the resistance of the plati- 

 num thermometer during the experiments on compression and expan- 

 sion a Carey-Foster bridge was used. The bridge wire, which was of 

 manganin, was calibrated in the usual way and had a resistance of 

 0.01317 ohm per cm. The platinum thermometer was connected to 

 the bridge by means of manganin lead wires with a resistance of 2.359 

 ohms and a temperature coefficient of 0.000015 per degree C. The 

 coils on the other side of the bridge were also made of manganin wire, 

 which had the same temperature coefficient as the lead wires. The 

 resistance of each of these coils was accurately known. The change 

 in the resistance of these coils due to change in their temperature 

 was very small, but as it introduced in the coefficient of expansion an 

 error of about 0.2 per cent for a change of 1° C, the temperature of 

 the room in the neighborhood of the coils was always noted and the 

 necessary correction applied. The temperature of the room never 

 changed more than 2° C. or 3° C. between observations on expansion. 



For a battery one Leclanch^ cell was used. A sensitive form of 

 Thomson galvanometer was found satisfactory for the resistance 

 measurements. 



The heating of the platinum thermometer by the current used in 

 measuring its resistance had, of course, to be considered. By putting 

 a resistance of 1 ohm in the battery circuit and by using coils with a 

 resistance of 0.1 ohm for balancing coils in the Carey-Foster bridge, it 

 was practicable to reduce the current through the platinum ther- 

 mometer to about 0.005 ampere. This current flowed for not more 

 than one second. The amount of heat thus generated was small, and 

 would, it seems, for the most part be communicated to the liquid which 

 surrounded the platinum wire. Moreover, as this heating effect was 

 always about the same, it could not at worst introduce more than a 

 small constant error in the determination of the temperature, and 

 in getting the change of temperature this small error would be elimi- 

 nated. It may therefore be neglected here. 



This method of measuring the change of temperature in the liquid 

 was found quite satisfactory. In the work on expansion the tempera- 

 ture of the liquid was changed only 2° C. or 3° C. The correspond- 

 ing change in the resistance of the platinum thermometer was never 

 greater than the resistance of the bridge wire, and care was taken to 

 measure the change of resistance in terms of the bridge wire alone. 

 The work was thus free from any error that might have been made in 

 determining the resistance of the standard coils. It was practicable to 



VOL. XLII. — 28 



