( li2 ) 



§ 5. The Thermostat. The thermostat which was used in these 

 control measurements and also in experiments upon the con)})ressi- 

 bility of h} drogen at ordinary temperature wliieli have already been 

 made but are not yet published was essentially the same as that 

 described in Comm. N". 70. For a description of the apparatus refe- 

 rence must be made in the first place to that paper. Certain modi- 

 fications have been introduced with a view to 



1. better constancy of the temperature during a great time interval 

 and consequently less necessity for constant supervision ; and 



2. easier adjustment to the desired temperature. 



The water vessels have remained the same. A small cylindrical 

 vessel has been introduced immediateh' after the large one (see 

 PI. III). The copper s[)iral witi» the xylol regulator of the vessel B 

 has been removed and in its place has been put (temporarily) a glass 

 spiral of three turns going from the bottom of the vessel B upwards. 

 This spiral is bent upwards tVom below- and ends in the gas regu- 

 lator. A side tube with ground joints connects the spiral with a large 

 thermometric vessel. A rubber connecting piece allows the system 

 to vibrate without damage. The thermometric vessel and the spiral 

 are filled with chloroform and mercury, the mercury being shown 

 black in the figure. The use of chloroform as thermometric liquid 

 offers the advantages of small specific heat, and small compressibility, 

 compared with a pretty large expansibility (cf. Comm. N" 70 III § 3). 

 The spiral contains about 100 cc. and the thermometric vessel about 

 600 cc. of chloroform. By means of a tap Sp^ the total mass of 

 mercury present in the apparatus may be altered. 



The action of the apparatus is l)riefly this: The glass spiral regulates 

 the temperature of the bath B just as the copper spiral did in the 

 Plate of Comm. N°. 70 III, while the large vessel C considerably 

 diminishes small temperature variations. The thermometric vessel, 

 which presents a relatively small surface compared with its large 

 volume, is not sensitive to these variations. The same applies equally 

 well to integral temperature variations. Should, for instance, a con- 

 tinuous rise in the temperature of the room cause the teu'perature 

 of the vessel C to rise a little, then mercury will fiow from the 

 thermometer vessel through S[>o, to the spiral, the regulating flame 

 will become smaller, and water will enter the large vessel at a 

 temperature that is lower but is still kept constant by the spiral, so 

 that the absorption of heat by the walls of the large vessel is thereby 

 to a great extent neutralised. 



The second purpose for which the modifications \vere introduced 

 is, as can easily be seen, also served. For, when once the burners 



