Composition of IVaters of Land- Drainage and of Rain. IGl 



that no considerable amount of condensation occurs in tlie pump 

 to impair its action. 



In this way we are enabled as it were to icash all the vapour, 

 causing it to leave behind it any ammonia ^vhich it may contain, 

 without greatly increasing the bulk of the liquid to be sub- 

 sequently tested. The operation proceeds regularly and without 

 requiring much care or attention, except in the occasional im- 

 provement of the vacuum as a slight leakage may occur. We 

 may in this way, if we please, distil over the greater part of the 

 water operated upon. 



The object of adding the salt to the liquid to be distilled is 

 to elevate the point of vaporization of the water, whilst at the 

 .same time tliat of the ammonia is diminished from the well- 

 known cii'cumstance that a liquid saturated with a salt has its 

 power of dissolving gases almost destroyed. We have found 

 that under these circumstances all the ammonia present is brought 

 over with one-fifth, and indeed in some instances with one-tenth, 

 of the w'ater ; but it is preferable to distil a larger quantity. 

 Bisulphale of potash is substituted for sulphuric acid as a means 

 of collecting the ammonia, as it may well be supposed to be less 

 volatile than the latter. The resulting liquid is tested witli a 

 standard solution of ammonia, of such a strength that each scptem 

 (7 grains) represents ■02714 grains of ammonia. A neutral solu- 

 tion of litmus is employed, as is usual in such experiments. It 

 is proper to state that when this apparatus was used in the way 

 described an apparent excess of ammonia was frequently ob- 

 tained, due, tliere can be no doubt, to a mechanical carrying over 

 of small quantities of the acid liquid in d by the bubbles of 

 vapour. I believe that Avith care such a result may be avoided.* 

 The results actually given were obtained by a modification of the 

 method of using the apparatus. The vessel d was kept cold by 

 means of an outer vessel of cold water, and the greater part or 

 the whole of the distilled liquid condensed there, and was tested 

 at one operation. Even in this form the use of a vacuous appa- 

 ratus is most desirable, as the process is so regular and so nmch 

 under control, whilst it is perfectly impossible to suffer loss by 

 escape of ammonia. Numerous direct experiments have been 

 made to prove the correctness of this process, but it may be 

 sufficient to give a few instances of duplicate analyses where the 

 quantities distilled over and other conditions have been unlike, 

 and the similarity of result cannot therefore be due to accident. 

 The following are sucli analyses of the rain-water of five different 

 months in the year 1855, as mentioned in the foregoing table. 



* A plug of asbestos was inserted in the tube at k, but failed to prevent the 

 acid lifiuid from being carried over. 



VOL. XVII. M 



