REPORT ON THE COMPOSITION OF OCEAN-WATER. 
237 
Second Experiment. — Before entering upon a second trial, I thought I should endeavour 
to prepare an absolutely potash-free chloride of sodium, and succeeded in this (more nearly 
than I had done before) by decomposing commercially pure crystallised sulphate of soda 
with fuming hydrochloric acid, washing the precipitated chloride of sodium with hydro- 
chloric acid, drying, dissolving in water, and reprecipitating with hydrochloric acid. 
The dried salt contained only a trace of sulphuric acid (which was neglected). To test 
quantitatively for potash, a portion of the salt was made into normal sulphate, and 
23'25 grams of the latter (corresponding to 1 9 '1 grams of chloride) dissolved in water, 
mixed with 1 c.c. of pure chloride of platinum (converted into PtCl fi H 2 by addition 
of a measured volume of standard acid), evaporated to near dryness, and the residue 
mixed with 10 volumes of absolute alcohol and 5 volumes of ether, allowed to stand, and 
the deposit of salts washed with ether-alcohol. Saturated sal-ammoniac solution was now 
applied in successive instalments until about x 9 yths of the sulphate of soda had dis- 
solved. The residual salt was then ignited, finally in hydrogen, the platinum removed 
by dissolving in water and filtering, the salt in the filtrate reconverted into pure 
normal sulphate, and the Finkener (sal-ammoniac) process again applied. The chloro- 
platinate ultimately obtained was reduced in hydrogen, and gave 0‘8 mgrms. of 
Pt = 0 - 38 mgrms. of K 2 0, and the KC1 filtrate by Fresenius’s method 2‘4 mgrms. of 
PtCl 6 K 2 = 0 - 43 mgrms. of K 2 0. This latter chloroplatinate was identified by microscopic 
inspection. From the mean (0'40 mgrms.) of the two quantities of potash found, it 
follows that 100 grams of the chloride of sodium contained 0'0021 grams of K 2 0 or 
0'061 mgrms. for the 2‘9 grams which were employed in the test analysis to be reported 
on. 
The modus operandi was essentially the same as the one used in the first trial, 
except that a perfectly pure chloride of platinum (made from platinum purified by 
Schneider’s method) was employed, and that, to make assurance doubly sure, a fresh 
standard solution of chloride of potassium was used instead of the old one. 
Milligrams of Potash (K 2 0) present. 
Old 
New 
Synthesis. 
Atomic Weights. Atomic Weights. 
In the chloride of potassium, . 
50-04 
50-12 
In the chloride of sodium. 
0-06 
0-06 
Summa, 
50-10 
50-18 
