60 PROCEEDINGS OP THE AMERICAN ACADEMY 



prove, the accuracy is sufficient for the purpose in view. When the 

 volume of the load is large, it becomes necessary to measure the tem- 

 perature and pressure with great precision, and to protect the balance 

 from radiation, and from all causes of rapid change of temperature. 

 It was a great satisfaction to the author to find that by so simple 

 means the relative weight of glass vessels of considerable size may 

 be determined with accuracy to the tenth of a milligramme, — an 

 accuracy which is fully equal to that of the most refined processes of 

 chemical analysis. 



From the data given, and from the known normal density of the 

 air, it can easily be calculated that in the first series of weighings the 

 volume of the tube and mountings exceeded that of the weights by 

 about 75 cubic centimeters. With this difference of volume we have 

 a variation of ^^j of a milligramme of weight for a difference of ^^ of 

 an inch of mercury in tension, or one degree in temperature. Hence, 

 with a difference of volume of 100 cubic centimeters, we should have 

 a variation of weight amounting to about four milligrammes for every 

 difference of one inch in the barometer, or of ten degrees of the Cen- 

 tigrade thermometer ; and these data will furnish the basis for a rude 

 estimate of the effect in any given case. If the difference of volume 

 amounts to 2,500 cubic centimeters, then a difference of xoVo ^^ ^^ 

 inch in the barometer, or of y^„ of a degree in the thermometer, would 

 cause a variation of ^g of a milligramme in the weight ; so, also, a 

 variation in the intensity of gravity amounting to only ^qoito ^^ ^^^ 

 whole amount would produce a similar effect, and a sensible variation 

 would follow any marked change in the purity of the air. Hence, the 

 balance might be used to detect exceedingly minute changes in any 

 one of these variables, provided the others could be exactly controlled ; 

 and although with our better methods these applications of the balance 

 may be of no practical value, yet the considerations here adduced will 

 serve to show how sensitive the instrument is to the slightest changes 

 in the density of the air when loaded with vessels of large volume. 

 The best method of controlling the weight in such cases is that adopted 

 by Regnault in his classical work on the density of the more perma- 

 nent gases. This consists in balancing the vessel whose contents are 

 to be weighed with a second vessel of equal volume, the two hanging 

 side by side in a case lined with thick felt. The balance is set over 

 the case, and the vessels are suspended from the pans by means of 

 platinum wires, which swing freely through holes made in the base 

 of the instrument. 



