HYDROGRAFI. 2 g 



addition of a solution of silver-nitrate in excess and determination of the excess by titration with a very 

 diluted solution of sulphocyanate of ammonium, and the employment of ammonia iron-alum as index 

 What I wished to obtain by this, was an easy and exact determination of the ratio between the quan- 

 tities of halogen in two samples of water. The analysis was executed in such a manner, that 10 cubic 

 centimetres of sea-water were placed in a small retort, supplied with a glass-cover ground for the 

 purpose, when the whole was weighed; previously the weight of the retort, as well as that of a 

 little pure nitric acid, which had been poured into it; was ascertained. Next the solution of silver- 

 nitrate, of such a strength that it corresponded to a similar volume of Atlantic sea-water, was added. This 

 process was effected by filling the solution into a Martottc's bottle, the drain-pipe of which was connected 

 with a glass-tube (terminating in a point) by means of a piece of india rubber tube supplied with a cock. 



If we open the cock, the solution will run out, and the air will bubble up from the lowermost 

 part of the glass-tube, which goes through the cork in the neck of the bottle. Every time that a 

 bubble of air is ascending, a certain volume of the liquid has made its way out, which is ascertained 

 by weighing. As in consequence of this, I knew how many bubbles of air there were to ascend, and 

 how many drops there still were to be added to neutralise a certain quantity of halogen, and knowing 

 approximatively the quantity of halogen in the water that was to be examined, it became possible 

 for me only to add a very small excess of the solution. The retort and its contents was then weighed 

 again. After this, the retort was heated a little, while at the same time it was shaken, and then 

 placed a few minutes under a cover of black pasteboard to protect its contents against the effects of 

 the light. After the lapse of some minutes, the halogen-silver had settled, and the supernatant 

 liquor was decanted off into a beaker. Next the halogen-silver was shaken during the heating 

 together with distilled water containing a little nitric acid, and deposited as before, on which the 

 liquor was decanted off into the beaker. In this manner the halogen-silver was washed out three 

 times, which proved to be sufficient, as further washing-out did not give any nitrate of silver. The 

 liquor decanted off from the halogen-silver, was now titrated by sulphocyanate of ammonium, applied 

 to it from a burette. The strength of this solution was found to be so, that i cubic centimetre of 

 same corresponded to 0.10157 grms of the solution of silver. 



Following this method, it proved that the added excess of the solution of silver-nitrate on an 

 average amounted to 0.25 grms, and never exceeded 0.7 grms. Making the excess so small, we run 

 of course the risk of getting a deficit instead of the desired excess of the solution. During the course 

 of 60 analyses, this has only been the case four times, and every time it occurred, the process was 

 repeated with a new sample of water. It appeared from the repeated titrations, executed at different 

 times, that the inaccuracy in the indication of the salinity in pro mille for inst of 35.24, did not exceed 

 0.006; consequently the system of titration employed had the desired exactness. By the adoption of 

 this mode of determination, errors, originating from differences of temperature, only manifest them- 

 selves in the result as a factor of the frequently mentioned excess of silver-nitrate, so that they may 

 be regarded as being of no consequence at all, and as the greatest weight was attached to have all 

 the analyses made in quite the same manner, it must be considered as an established fact that the 

 mutual deviations of the repetitions may be regarded as due only to differences in the judgement of the 

 colour-reaction, and as this latter, in this case, is much more distinct than the one produced by 



