Dec. 4, 1873J 



NA TURE 



87 



ON THE SCIENCE OF WEIGHING AND 

 MEASURING, AND THE STANDARDS OF 

 WEIGHT AND MEASURE* 

 IX. 



T N the comparison of standard weights, the difficulty 

 ^ and risk of error in determining the weight of air dis- 

 placed by them is to be avoided by weighing them not in 

 air, but in a vacuum. Two methods are employed for 

 this kind of weighing. 



In the first and simplest method, when an ordinary 

 balance of precision is used, each standard weight is 

 placed in an exhausted receiver just large enough to hold 



it, and is weighed separately against a counterpoise by 

 Borda's method. Sensible discordances have, however, 

 been found in the results of this method of weighing in 

 exhausted receivers, which render its use inexpedient 

 i when scientific accuracy is required. These discord- 

 I ances are perhaps attributable to a small quantity of air 

 1 being present in the receiver during the weighings, the 

 amount of which cannot be accurately determined. An- 

 other probable cause is a chmge in the temperature and 

 I atmospheric pressure affecting the balance itself and the 

 weights in the pans during the long time necessarily 

 occupied in the whole process of this method of weighing. 

 Indeed it may be generally stated as a rule that the risk 

 of discordances in the results of weighings is in proportion 



19.— New V 



to the time occupied in the operation. Such discord- 

 ances are not found when the weighings are made by the 

 second method, when a vacuum balance is used, that is to 

 say, when the balance case itself is made an exhausted 

 receiver. 



A vacuum balance has been constructed at Paris by 

 1»\. Deleuil, and is now used at the Conservatoire des 

 Arts et Mdtiers, consisting of a balance of the best con- 

 struction placed in a very strong cast-iron case that can 

 be made perfectly air-tight. This case has four circular 

 openings for giving admittance and light to the inside, 

 which are closed with strong glass covers. It has a 

 * Continued from p. 40. 



Dcp.litn 



stuffing box for the handle of the lever by which the 

 balance is put in action and arrested. This balance has 

 been found to give very accurate results of weighing in a 

 vacuum. But the comparison of standard weights in this 

 vacuum balance takes a considerable time from the ne- 

 cessity of opening the case and re-establishing a va- 

 caam at least a second time in order to change the 

 weights in the pans even when Borda's method is used ; 

 and occasionally this must be done again if a small 

 additional balance weight is required to be placed in 

 either pan, in order to obtain a sufficiently approximate 

 equilibrium, so that the pointer may not exceed the limits 

 of the index scale. 



