Specific Gravity of Vapours and Gases, 5 



The exactness of the single weighings may be judged of from 

 the regularity of the differences in the second vertical column, 

 since all errors are there shown as irregularities in the oscillations 

 which were brought in by vibration of the ground, draughts of 

 air, and unequal heating of the balance. It is an essential 

 condition for the accuracy of the weighings to perform them ac- 

 cording to a system by which all errors arising from alterations in 

 the balance which do not occur within two closely consecutive 

 weighings shall be fully eliminated. Let r 2 , r 3 , r 4 , r h represent 

 the position of the pointer when the balance is in equili- 

 brium, calculated each time from the observed oscillations when 

 the left pan is weighed successively with the vessels II., III., 

 IV., V., whilst the vessel I. is placed on the right pan as 

 a tare. If we take one r (in the foregoing example r 5 ) as a point 

 from which the oscillations are to be reckoned, we have r 2 — r 5 , 

 ?* 3 — r 6 , r A —r b as the deviations, expressed in divisions of the 

 scale, which the several vessels produce when the beam is in 

 equilibrium. 



If we have estimated, once for all, the weight 5, which on 

 balancing with one of the vessels corresponds to a deviation of 

 one division of the scale, then we obtain the differences of weight 

 A which the separate vessels show with the vessel V. from the 

 equation 



A = S(rn— r 



Three weighings made at different times gave for 8 



0-00079 grm. 

 0-00071 „ 

 0-00073 „ 



Mean . . 0-00074 „ 

 And as differences of weight between V. and II., III., IV., 



+ 0-00005 -0-00031 -0-00060 



+ 0-00000 -0-00026 -0-00070 



-0-00005 -0-00032 -Q-Q0074 



0-00000 = A II. -000030 = A III. -0-00068 = A IV. 



In order to close the vessels air-tight, glass tubes are used 

 made somewhat narrow in the middle, and about 50 millims. 

 long and 5 millims. wide (fig. 2, a. a). Their inner walls are 

 lined with unvulcanized caoutchouc, and the end is closed with 

 a glass stopper prepared from glass rod at the blowpipe. The 

 tubes are lined with caoutchouc in the following manner: — A 

 thread of strong saddler's twine is bound round the end of a 

 seamless caoutchouc tube, which must be tolerably thick in the 



