E. W. Morley — Volumetric Composttio?i of Water. 229 



"been safe to use: but so many difficulties had to be sur- 

 mounted that the pleasure of evading some was not to be dis- 

 regarded, and I used no organic matter about the apparatus. 



The tube r yielding pure hydrogen is continued in the tube 

 a of fig. 2. And a leads directly to the stop- cock c just out- 

 side the calorimeter case containing e and to the stop-cock of 

 the globe e. The manipulation of c, d, e will be described in 

 a paper on the density of hydrogen and of oxygen ; the gas in 

 passing to e does not go near mercury. At g is seen a tube 

 connected to a McLeod gauge, and leading to a Geissler air- 

 pump having a capacity of 2*3 liters. This was kept in such 

 condition that it was not difficult to exhaust the pressure gauge 

 and the connecting tubes with a total volume of 700 cubic 

 centimeters to one part in five million. It is therefore obvious 

 that the amount of leakage through its ground joints which 

 could take place during the exhaustion was negligible, and 

 when the pump was not in use for exhaustion, it was shut off 

 by a mercurial valve which was perfectly tight. 



The vertical tubes seen immediately above the horizontal 

 part of a form a six-way mercurial valve, of which three ways 

 were used in the manipulation of hydrogen, and three in the 

 manipulation of oxygen. When mercury is lowered in the 

 tube h, a free passage exists between the Geissler pump with 

 its McLeod gauge and the tube a leading to the valve at o, fig. 

 1, as well as to the weighing globe e, fig. 2. When the mer- 

 cury is lowered in the tube k, there is a free passage from the 

 pump to the mano-barometer and the store globe, n. And 

 when the mercury is lowered in the tube j, there is a free pas- 

 sage between the tube a and the mano-barometer and the store 

 globe. Of course this valve will saturate the hydrogen with 

 mercurial vapor, but this part of the hydrogen is to be 

 analyzed over mercury. Any hydrogen stored in the globe n 

 is perfectly safe from contamination. The tubes were all most 

 scrupulously cleaned, all joints were made by fusion, when 

 possible, and the others are mercurial seals of sufficient depth ; 

 at the mercury six- way stop-cock, the hydrogen is shut off by 

 at least thirty centimeters of mercury, at the mano-barometer, 

 by thirty, and at the valve o by eighty centimeters. It was 

 accordingly found that that hydrogen stored at in n contained 

 no more nitrogen after keeping it for a month than on the 

 day of its preparation. 



Introducing hydrogen into the storage apparatus. — In filling 

 n with hydrogen, n was exhausted to a twenty-five thousandth 

 or a fifty thousandth part, and was then shut off from the 

 pump. Then the tube between d, fig. 2, and o, fig. 1, was 

 exhausted to a hundred thousandth. During these exhaustions, 

 hydrogen had been escaping at n, fig. 1. The tube at n was 

 then fused together, the stopcocks at m closed, and the valve o 



