OF ARTS AND SCIENCES. 2 1 1 



When the globe is filled by displacement it conduces to greater ac- 

 curacy to determine independently the tare in question ; and as soon 

 as the details of the new method had been worked out, we began our 

 final determinations by taking according to the old method the tare 

 of the balloon abov r e described when empty, so as to obtain a definite 

 standard of comparison for further results. We however varied the 

 method of Regnault in so far that we always filled the globe with hy- 

 drogen before exhausting the interior. As with this exceedingly light, 

 material the weight of the residual gas seldom exceeded one half a 

 milligram, an error as great as one fifth in the capacity of the bal- 

 loon or in the tension of the residual gas would cause no appreciable 

 difference in the value of the small weight we sought to estimate. 

 The details of a single example will be sufficient to illustrate the 

 procedure. 



Determination No. 1. 



A current of hydrogen gas from the electrolytic generator described 

 in our previous paper * was run through the balloon from Saturday, 

 May 25th, at 4 o'clock p. m., to Monday, May 29th, at 10 o'clock A. M., 

 1889, the balloon standing in calorimeter case connected with larger 

 balloon to receive overflow. At moment of closing, the height of 

 barometer and temperature of the case were observed, but these data 

 are not required for the present determination, and will be given 

 hereafter when used for calculating the specific gravity of hydrogen. 

 The balloon was exhausted with an excellent mechanical air-pump, f 

 and the tension of the residual gas measured by means of a siphon 

 manometer interposed between the balloon and the pump. The form 

 of manometer used is shown by Figure 2. It was made in the labo- 

 ratory, and in filling it with mercury the liquid metal, purified by 

 repeated distillation, was boiled in the tube with the greatest care. 

 For convenience, the difference of level of the mercury in the two 

 arms was measured with a cathetometer constructed by the Societe 

 Genevoise, but it might have been estimated with sufficient accuracy 

 by the millimeter scale of the instrument. Tension in balloon after 

 exhaustion 456.95 — 455.55 = 1.40 mm. 



Assuming 0.4164 as the approximate value of the weight of hydro- 

 gen gas filling the balloon at 763.10 mm. and 21°.08 C. (p. 227), the 

 ht of the residual gas at 1.4 mm. and same temperature would be 



weig 



* These Proceedings, vol. xxiii. p. 1G8. 

 t Author's Chemical Physics, p- 881. 



