IJA'I'K) OF SPEOIFK! IIRATS. 3 



tciiii.cnitiiiv 7; loiiuvi' tliMii an cxmssively l)ricl' interval, secinij that lip;it is con- 

 tinually lUMiied into it IVoni without. 



II. TllM MlOTirOD IN PRAdTICIC. 



(1. PLAN OK TlIK EXPERIMENTS. 



To liold the i-har^v of gas we made u.se of a lai-gc receiver of s|Mni copper 

 nearly spherical in form ami about 90 litres' cajtacity. 



Figure i. 



This receiver 7? (fig. 1) contains three tnliidated openings, one of which, 

 shown laterally at O-^, 'I'i^y be closed with a stopcock h. This coinniunicates in the 

 manner seen in the figure with the manometers JSl and /;/, the drying apparatus 2\ 

 and the compression pum[) 6', by which the gas can be condensed in any necessary 

 amount. If commercial oxygen, hydrogen, or carbon-dio.xide is used, the gas may 

 be tapped directly out of the high-pressure cylinders, through the drying train 1 

 into the receiver. The cock // is closed whenever the charge of gas is allowed 

 to expaud into the atiiKisphere. 



The second lateral tubiilure, O.,, is adapted to receive the lioloineter strip x, 

 which represents one branch ( 7) of theWheatstoue bridge II'. The other branches 

 (5,5, 4) consist of suitablv chosen resistance standards. For reasons relating to the 

 heat-conduction discrepancy, the form given to the strip of the bolometer is as shown 

 in fig. 2. The strip is cut from a piece of platinum-silver foil, in Avliich the 



^ ° ~ >=-, r==== ^-=^ 



FiGtIRE 2. 



lavei- of silver is al>ont ten times as tliicdv as the cop.xtensive ]>latinuni layer. In 

 our definitive experiments the thickness of platiniini foil used did not exceed 



