Constitution of the Atmosphere. 251 



air has varied in experiments that have been carefully made 

 in Virginia. For a length of time these variations were im- 

 puted to the use of binoxide of nitrogen, and it was not until 

 we made the same observations in using hydrogen gas, that 

 the true cause was suspected. I believe that the binoxide of 

 nitrogen will always give accurate results, if added to atmo- 

 spheric air, in a stream of bubbles, by Hare's sliding rod eu- 

 diometer* : one fourth of the deficit is to be taken. 



In the course of these investigations, it has happened to 

 me to observe some instances of an action which you have so 

 fully described. In a paper inserted in the American Journal 

 of Medical Sciences for May 1836, it is remarked, "An im- 

 portant circumstance in gaseous analysis may here be no- 

 ticed. If a tissue, in the act of transmitting gas or ready to do 

 so, be placed in contact with another gas of a different nature, 

 disturbance immediately ensues. A cubic inch of nitrogen 

 made with phosphorus, but which was found to be contami- 

 nated with 4^ per cent, of oxygen, was agitated briskly in a 

 phial containing about an ounce of spring-water. In one 

 minute the nitrogen gained one per cent, by the agitation. 

 The same quantity of nitrogen agitated in a pint of water 

 gained no less than eleven per cent, of oxygen. Nor is agita- 

 tion or mechanical violence necessary to produce this im- 

 portant result. Into a bell filled with water and inverted 

 into another vessel, so as not touch it in any point, I placed 

 100 measures of a gas, 85 of which were oxygen. After four 

 weeks, an analysis was made, and the gas in the bell found 

 to contain only 72 per cent, of oxygen, the remainder being 

 nitrogen. In this way too, in the lapse of time, from an in- 

 verted vessel partially filled with atmospheric air, the oxygen 

 will escape into the water, and thence into the atmosphere ; and 

 I have twice known this event to take place, so that the resi- 

 due did not contain more than three or four per cent, of oxy- 

 gen. In many of the most delicate researches of chemistry 

 we have this disturbing cause in operation, which has for the 

 most part been overlooked. Water is uniformly employed 

 in our laboratories as a means of confining gases ; it enters 

 largely into our processes of pneumatic manipulation, and 

 though we have hitherto neglected its action, it silently dis- 

 turbs all our results. An air bell cannot pass to the top of 

 a jar without instant contamination; during its residence 

 there it is subject to a continued succession of changes ; at no 

 two moments is it the same in composition, a perfect freedom 

 of communication existing between it and the atmosphere." 



" As an instrument of rigid analysis, the pneumatic appara- 

 tus so arranged requires to be used with circumspection. It is 

 [* See Phil. Mag. and Annals, N. S., vol. vi. p. 114.] 



