206 PROCEEDINGS OF THE AMERICAN ACADEMY 



such observations we can readily calculate, by the method we have pre- 

 viously described in these Proceedings,* the volume of the small bulb 

 required to make the compensation perfect, and an experienced glass- 

 blower can readily make a bulb of the size indicated, which, before it 

 is sealed, may be adjusted as to weight by loading with mercury as 

 above described. 



The weights used in connection with the balance had been adjusted 

 for our previous work with the greatest care, and not only compared 

 with each other, but also compared with a copy of the Washington 

 standard. Although we shall deal in this paper solely with relative 

 weights, yet the absolute values that may be deduced from our results 

 must be correct far within the limits of the experimental errors of our 

 own work. The sum of the weights required to measure the differ- 

 ences of tare never exceeded twelve grams, and of this small mass 

 any possible difference between our copy and the Paris standard kilo- 

 gram must be wholly inappreciable. 



The Glass Balloon. 



We have used in this investigation the same glass balloon that was 

 fitted up for a somewhat different purpose in our work on the atomic 

 weight of oxygen, but which has proved equally well adapted to deter- 

 minations of gas density. Regnault in his work on the same subject 

 used a balloon of ten liters' capacity, having its neck closed with a 

 brass cap and stop-cock, and filled it by alternately exhausting the in- 

 terior with an air-pump and refilling from a gasometer until the com- 

 plete purity of the enclosed gas was assured. After long experience 

 with this mode of experimenting, we have found it more convenient 

 to fill the balloon by displacement directly from the generator, and we 

 have found that we can reach a satisfactory result with least labor in 

 this way. 



Our balloon is represented in Plate I., which shows all the details of 

 its construction. It was made by Emil Greiner of New York, and 

 all the joints are so finely ground that when properly lubricated they 

 will hold a vacuum indefinitely. The peculiar form of the stop-cocks 

 enables the experimenter to shut into the globe a definite volume of 

 gas and at the same time open a vent to the generator through the 

 side tube ; or to allow the gas to flow through the side tube until the 

 current is perfectly pure before it is admitted to the balloon ; and 

 again, when the stop-cocks are both closed and the balloon disconnected, 



* Vol. xviii. p. 55. 



