REPORT ON THE COMPOSITION OF OCEAN-WATER. 
13 
analyses were made before any analysis of sea-water was attempted. Let me at once add 
another experiment, made at a far later date, with specially purified chloride of platinum. 
A known weight of chloride of potassium equivalent to 5 2 -02 mgrms. of metallic platinum 
w T as dissolved in 50 cubic centimetres of potash -free artificial sea- water, prepared from 
absolutely potash-free materials, and evaporated down with a quantity of pure chloride of 
platinum, containing 3 *2 grms. of metal, to the consistence of a paste, from which the 
foreign chloroplatinates were washed away successively by a little water (forming a strong 
PtCl 4 solution with the excess of reagent), 5 per cent, solution of platinum chloride, and 
lastly alcohol, — according to Tatlock’s directions. The cliloroplatinate of potassium 
obtained (dried at 105° C.), amounted to 107'9 mgrms. = 43'62 mgrms. of metal. This 
chloroplatinate, not presenting a perfectly normal appearance, was purified by redissolv- 
ing it in hot water, adding 1 cubic centimetre of chloride of platinum solution, containing 
10 per cent, of metal, re-evaporating and again applying Tatlock’s washing process. 
The purified product weighed 101 '7 mgrms. = 41 T 8 mgrms. of metal. The filtrate from 
the original chloroplatinate was evaporated to dryness, reduced in hydrogen, the salts 
extracted with water, made into sulphates and worked up by Finkener’s process. ( See 
below.) Platinum obtained from the PtCl 6 K 2 =8'9 mgrms. Calculating the potash in 
terms of metallic platinum, we have : — 
(1) In purified precipitate, ..... 4148 mgrms. 
(2) Lost in purification, an unknown weight x, sure to be less than 244 ,, 
(3) From PtCl 6 K 2 filtrate from the crude chloroplatinate, . 8 - 90 ,, 
Accounted for . 52 - 52 „ ( + £-244) 
Due, . . 52-02 „ 
Even the unpurified precipitate (1) represents only 83'85 per cent, of the potash to be 
determined. 
Seeing that the customary process even in its most exact form, would not work, I 
thought I should give a trial to that ingenious modification of the platinum process which 
was invented some years ago by Finkener,* the more so as this process, far more readily 
than the old one, lends itself to the laying down of a hard and fast sequence of operations, 
which, although it may fail to ensure perfect precision, was certain (I thought) to give at 
least constant results, susceptible in the worst case of subsequent correction. For this 
purpose I brought the process into the following form : — Measure off 50 c.c. of sea-water, 
and after having determined its weight, evaporate with sulphuric acid ( vide infra under 
“Soda,”) and ignite the residue to convert all the bases into normal sulphates. Dissolve 
these in 10 to 20 c.c. of water, filter, add excess of chloride of platinum, i.e., more than It? 
[PtCl*.] per 1 [K 2 S0 4 ], evaporate to a very small volume, allow to cool, and add, first 10 
volumes of absolute alcohol, and then 5 volumes of ether. After some hours’ standing wash 
* Poggendorff, Annalen, vol. cxxi. p. 637, et seq. ; also Handbueh der analytischen Chemie von H. Rose, 6th ed. by R. 
Finkener, vol. ii. p. 929. 
