REPOET ON THE COMPOSITION OP OCEAN-WATER. 
55 
S, = S 0 — at -rbt I 2 , which was found to do sufficient justice to the experiments (the “ errors ” 
ranged from — 0'05 to + 0'06). # 
II. Series. — This series was quite similar to the first, except that a more dilute water 
(y= 18-023) was operated upon — in 20 experiments — the temperatures for which 
ranged from + 0 o, 8 to 29 0, 8. Here also a formula of the second degree was found to sum 
up the results satisfactorily. 
As the excess of the specific gravity over 1000 must be presumed to be at 
least approximately proportional to y, I tried to sum up the two series in one 
formula — 
4&t — 1000 = x(a 0 — fit±yt 2 ), 
but did not succeed in establishing a sufficient general agreement between calculation and 
experiment. 
III. Series. — The object of this series was to determine the exact relation between 
4 S t and x f° r a wider range of concentration at a constant “ordinary” temperature. 
Starting with a sea-water of y=20‘345, I mixed three separate portions of it with known 
proportions by weight of pure water, so as to produce in all four waters of convenient 
strengths. Each of the four waters was analysed, the results were calculated as so many 
determinations of the y of the original water, and the mean, 20"345, adopted as the value 
of x for it- From this value (and the data of the syntheses) the y’s of the other three 
waters were calculated and found— 
Original Water. Mixture A. Mixture B. Mixture C. 
(20-345) 19-290 18‘322 17‘322 
The specific gravity of the original water was ascertained by means of the plunger 
and I found that — 
4 S 19 . 3 = 1026-396. 
I then proceeded to determine the differences between the specific gravity of this 
water and those of the mixtures A, B, and C, by means of a method of my own inven- 
tion.! A cylindrical specific gravity bottle, of about 30 c.c. capacity, whose glass stopper 
was perforated by a capillary aperture, was filled with “ original ” water, and from the 
one pan of a balance, by means of a thin platinum wire, suspended within a mass of 
water of the same kind, kept at a constant temperature (about 19° C.), its tare taken and 
noted down as W 0 . The bottle was then emptied out, rinsed and charged with, say, 
water A, up to near the edge, and the “ original ” water in the cylinder used as a bath 
* Iu the final calculations ( see Table II.) I adopted for each of certain groups of experiments such an intermediate 
value between the initial and the final % as seemed to me most probable. The arbitrary element thus introduced is of 
no significance. 
t See Chem. News, vol. xliv. p. 51. 
