REPORT ON THE COMPOSITION OF OCEAN-WATER. 
83 
heavy plunger, suspended from below within the liquid, kept the instrument steady, and 
ensured to it a vertical position. To determine the double weight of the hillock drawn 
up by this apparatus from water, I determined (l), the loss V in weight which it suffered 
when plunged into water up to the mark; (2), the loss l" which it suffered when 
suspended the other way; and (3), the loss L it suffered when totally immersed. 
Obviously the water displaced in (1) and (2) is l' + m and l" -f to, and 1' + 1" + 2m=L ; 
or, L-(l' + l") 
m — t • 
Now, m — au where u is the sum of the two peripheries, which in my instrument was 
= 140 mm., whence 
m 
a ~14b' 
Operating in this manner, I found the value of the constant a (the units being the gramme 
and millimetre) 
For pure water a 0 = 0-00456. 
For sea- water a x = 0-00402. 
Whence p , 0 = O'O46 and — 0*0406, and for the capillarity correction of Buchanan’s 
hydrometer (calling it S 0 - S) 
D a 0-0406-1-028 x 0-046 
On — 0= T7^r\ ’ 
or 
$0-S 
-0-0067 
160 
= - 0-000 042, 
or S 0 = S — 0'042 for water equal to 1000. The specific gravities, as found by the hydro- 
meter, should be too high by 0"042, water= 1000. When I actually used Mr. Buchanan’s 
hydrometer to determine the specific gravity of a sea-water which had already been 
ascertained exactly by means of the plunger, I obtained a number which differed from 
the plunger-result by considerably more than 0"04, and the difference lay in the opposite 
direction. To settle this matter I carried out the following two series of experiments. 
ls£ Series. A number of spirals of copper wire of known weights were prepared, and 
so adjusted that when one of them was attached to the top of the instrument, and the 
latter floated in water of 16° C. to 18° C., it sank down to somewhere between 40 
and 60 mm. of the scale. 
The hydrometer then was floated in a large mass of pure water, of about the 
temperature named (the temperature of the surrounding air being about the same), the 
several over-weights, iv j w 2 , &c., were attached successively, and in each case the hydrometer 
and the thermometer were read. This was repeated sixty-five times, the weights being- 
made to vary from reading to reading, so that each turned up about three to five times at 
