Hydrometers of Total Immersion. 501 



liberate the hydrometer the funnel was raised, but only so far 

 that its rim might catch the stem of the hydrometer should it 

 attempt, as it invariably did at first, to float to inconvenient 

 quarters of the beaker. 



Besides the standard thermometer [11806J which was 

 placed in the water, two other trustworthy thermometers were 

 placed in such positions that they indicated the temperatures 

 of the upper and lower strata of air in the chamber. A 

 mirror was placed at the back of the air-chamber to reflect 

 the light so as to facilitate the readings of the standard 

 thermometer. 



The position of the mercury column in the thermometer 

 was read by means of a microscope furnished with a 

 micrometer eyepiece. 



Fig. 3 (p. 502) gives a section of the apparatus. 

 With this apparatus it was possible to attain any tempera- 

 ture between the atmosphere and 40° C, and to regulate it so 

 that it could be made to rise as slowly as o, l per hour. 



By allowing the final adjustment of temperature to take 

 place with extreme slowness, and by frequent and effectual 

 stirring, the prejudicial influence of convection-currents can be 

 reduced to a negligible minimum. The best possible proof of 

 the absence of convection-currents is furnished by the 

 hydrometer itself, which, in the experiments described, often 

 for minutes, remained poised in the middle of the water. 



Generally at the time of reading the temperature was 

 apparently constant. Several successive readings were often 

 made which yielded identical results. In the neighbourhood 

 of the thermometer the temperature seemed to change only at 

 the time of stirring, remaining afterwards, for a time at least, 

 stationary. 



A glass envelope subjected to a change of temperature 

 does not for some time assume its final volume corresponding 

 to the new temperature. This interval of time is much greater 

 when the temperature is lowered than when it is raised. 

 Hence all the experiments were made at temperatures higher 

 than that at which the hydrometer had been for some time 

 previously, and no final reading was made until the hydrometer 

 had been exposed to within a few tenths of a degree of that 

 temperature for two or three hours. 



Method of carrying out an Experiment. 



Distilled water was poured into a beaker made of Jena glass 

 and having a capacity of 3000 c. c. The water was then 

 heated to within a degree of the required temperature, and 



Phil. Mag. S. 5. Vol. 48. No. 295. Dec. 1899. 2 N 



