REPORT ON THE COMPOSITION OF OCEAN-WATER. 
5 
dissolved in water to “ 1 litre,”* and the solution weighed exactly. Supposing it 
amounted to lOC^Ot grins., this weight was put down as the exact deci- equivalent of 
the solution, it being of course remembered at the same time that the approximate deci- 
equivalent volume was= 1000 c.c. 
The Solution of Nitrate of Silver was generally prepared on a large scale (40 to 50 
litres at a time), by dissolving a known weight of pure crystallised nitrate, containing 
an ascertained proportion of water, in very dilute nitric acid, the proportion being 
chosen so that every litre of solution contained as exactly as possible 17 grms. of 
nitrate and 20 c.c. of nitric acid of 1 '4 sp. gr. To ascertain the exact titre, equal 
volumes (50 c.c.t each at first ; 100 c.c.t for the final tests) of the silver and of the chloride 
of potassium solutions were measured out, mixed, and weighed at the same time ; the 
mixture shaken, allowed to settle, and the excess of silver titrated by means of centesi- 
mal solutions of silver and sulphocyanate of ammonium as above explained. When the 
chlorine happened to predominate, this was easily set right by neutralising the sulpho- 
cyanate added by its exact equivalent of silver, pouring the liquor back on the chloride 
in the precipitation phial, adding a sufficiency of measured silver for the chlorine and 
beginning de novo. We always took care at first to keep the silver solution a little 
above its intended standard, so that the correction could be effected by adding the 
calculated amount of water to the stock in the carboy. When the solution was as 
nearly as possible volumetrically correct, the exact titre was again determined three or 
four times, the ice igh £ - e qu i v al e n t calculated, and this number used subsequently for the 
calculation of the exact proportion of chlorine in the sea-water analysed. To give an 
example, I may state that for a supply of solution which served for about half of all the 
chlorine determinations made,| the equivalent was 1024'24 grms. ( = very nearly 1000 c.c.), 
so that the number (y) § of grms. of chlorine per kilo, of sea- water analysed followed from 
the computation of — 
Sx 3-546x1000 
X - 1024-24 x W 
where S means the weight of (corrected) silver-solution found equivalent to W grms. 
of sea-water. In no case was such a solution actually used before the correctness 
of its titre was confirmed by the analysis of exactly weighed samples of pure (dry) 
chloride of potassium or sodium. 
Tice Deci-normal Solution of Sulphocyanate was prepared from pure ammonium salt, 
and adjusted by means of the standard silver. After some experience we found it con- 
* Meaning the capacity of the (fairly correct) litre-flask used. 
t Measured in the same pipette. 
J A very large number of samples of water were analysed for chlorine only apart from those “ completely 
analysed. 
§ I may here state, once for all, that this symbol (x) is used consistently throughout my memoir to signify grnu. 
of halogen, calculated as chlorine, in 1000 grms. of sea-w T ater.” 
