REPORT ON THE COMPOSITION OF OCEAN-WATER. 
235 
with nitrate of silver, give only an opalescence of chloride of silver. The precipitate of 
chloroplatinate is dissolved on the filter in hot water, and the solution allowed to run 
into a tared porcelain crucible, recovered by evaporation and weighed, first after drying 
at 100° to 105° C., then after further drying at 130° C. From the weight of the pre- 
cipitate, the potash is calculated as K,0, in order to see what the exact but unreasoning 
application of Fresenius’s method would lead to. 
But the precipitate can neither be presumed to be pure nor to contain the whole of 
the potassium, hence, — Firstly, the crude precipitate is reduced in hydrogen, the residue 
treated with water, the platinum metal weighed, and from the solution the potassium 
present reprecipitated by a renewed application of Fresenius’s method. To check the 
result the pure chloroplatinate, after having been weighed, is reduced in hydrogen, its 
platinum thus separated out and weighed. The mother-liquor obtained in the prepara- 
tion of the pure chloroplatinate of potash is evaporated to dryness, the residue heated in 
hydrogen, the salt extracted with water, made into normal sulphate, and subjected to 
Finkener’s method to extract and determine the potassium. Secondly, the same process 
is applied to the filtrate from the crude chloroplatinate, to determine the potassium, 
which from the first escaped precipitation. This method was faithfully adhered to, 
except that we soon discarded the drying of the chloroplatinate at two successive 
temperatures. Haying found that the additional loss of weight involved in the passing 
from 105° to 130° amounts to less than the thousandth part of the weight of the jwe- 
c ip it ate, we subsequently dried at once at 130°, and calculated from the weight thus 
found. 
First Experiment . — The preparations used for the synthesis were a perfectly pure 
chloride of potassium made from recrystallised chlorate, and a chloride of sodium made 
from recrystallised bicarbonate and pure hydrochloric acid, and purified by crystallisation 
in the heat, the mother-liquor being rejected. This mode of preparation I thought 
would ensure complete absence of potash ; and yet, when the results came to be calculated 
it turned out that the sum of the several instalments of potash found came to more than 
the potash of the chloride of potassium taken. Hence the chloride of sodium used was 
suspected to contain potassium ; 5 '042 grms. of it were made into normal sulphate, and the 
latter subjected to that form of Finkener’s method in which the mixture of sulphate of 
soda and chloroplatinate of potassium is lixiviated with concentrated sal-ammoniac 
solution, to eliminate the sodium salt. The residual chloroplatinate was ignited in 
hydrogen and the potash calculated from the weight of the metallic platinum. We 
obtained 6'5 mgrms. of platinum, equal to 3'09 mgrms. of K 2 0,' :< or 179 mgrrns. of K 2 0 
for the 2'91 grms. of chloride of sodium which had been used in the test experiment. 
The chloride of potassium obtained from the chloroplatinate was not weighed, but 
identified by re-conversion into chloroplatinate. In the following statement of the results 
* Or 4-89 mgrms. of chloride of potassium = 0 '097 per cent, of the whole. 
