1864.] 



and Air -pump. 



325 



absorbent fluid by an elastic tube, slipped on tbe a end of the cock ; which 

 will also convey the gas to any vessel. 



If it be required to fill a receiver for experiments in an atmosphere of 

 gas either at common pressure or a less one, it may either be exhausted 

 by an air-pump connected with K, and filled from A, or exhausted b^ A 

 and filled from N. The former can only be done with gases which ive 

 no action on brass. 



These operations seem complicated when described with so much 3tail, 

 but in practice they are very easy, and their result is good. Some pre- 

 cautions, however, are required to ensure it. The bottom of the bell-cock 

 and of its key must be ground, so as to leave no shoulder or hollow in 

 which air may be entangled when the bell is filled. Every part of the 

 metal work must be air-tight ; this can only be secured by covering, not 

 only its joints, but its whole surface with several coats of varnish-paint — ■ 

 best of white lead. When the first coat is applied, on exhausting the ap- 

 paratus, every hole or pore is revealed by an opening in the paint (often 

 almost microscopic), which must be filled up as it forms till all is tight. 

 It is almost needless to mention that the whole must be perfectly dry. If 

 the bell be filled a few times with undried air, enough of moisture will 

 adhere to its walls to prevent an exhaustion of more than 0*1 inch. In 

 such a case it must be dried by drawing air into it through sulphuric acid, 

 and this repeatedly. Moisture also occasionally finds its way into a part 

 still more troublesome, into the passage which connects the bell and cy- 

 linder ; it is probably condensed there when the mercury is colder than the 

 atmosphere. I remove this by connecting the tube of K with a desiccator ; 

 setting C to (r), opening K and E, and by working the air-pump drawing 

 a stream of dry air into D, which bubbles up through the mercury in the 

 passage, and at last sweeps away all trace of water and its vapour. In this 

 operation it is necessary to remove a portion of the mercury, as otherwise 

 it would be sacked into the pump ; indeed this mischief might occur in 

 ordinary work by some mistake in the manipulation — for instance, by leaving 

 E open with («). To prevent the possibility of this, D is connected with 

 the pump by a mercury trap, easily imagined, which intercepts any of that 

 metal that might come over. And lastly, the interior of the bell must be 

 perfectly clean if the highest degree of exhaustion is required. This state 

 is obtained by washing it with strong nitric acid, then with distilled water, 

 and when quite dry wiping it with linen, from which all traces of soap or 

 starch have been removed by boiling it in rain-water. Thus we reduce to 

 a minimum the film of air which adheres to the bell even when filled with 

 mercury, and lessens its vacuum. When all these precautions were taken, 

 I found that with a receiver containing 3*7 inches, the fifth operation 

 brought the gauge (which had been similarly cleaned and carefully boiled) 

 down to 0*01. The sixth brought it still lower, but my present means of 

 measurement* are not sufficient to determine the precise amount. In this 

 * A micrometer microscope put in the place of the telescope of my theodolite. 



2 B 2 



