DEEP-SEA EXPLORATION. 
337 
DENSITY OF SEA WATER. 
The density of sea water in dil'ferent latitudes and at different depths is an ele- 
ment of so great importance in tlie study of ocean physics as to have caused a great 
deal of attention to be paid lately to its determination. The instruments employed 
lor the iHirpose have been, almost without exception, areometers of various forms. 
The differences of density as arising fr()iu saltiiess are so small that it is necessaiy to 
have a very sensitive instrument. As the density of ocean water at the temperature 
of 150 (J. only varies between the limits 1.0-!3 and 1.028 it is necessary, in order to 
determine differences to the hundiedth part, that we should be able to observe 
accurately the half of a unit in the fourth decimal place. This gives a great exten- 
sion to the scale and involves the use of a series of lloats if the scale starts from fresh 
water, or else the instrument assumes dimensions which make it unlit for use on 
board shi]>. 
With a view to the convenient adaptation to ])ractical use, a salinometer (plate 
xxiii) was devised for the Coast Survey by Prof. J. E. Hilgard, which was subsequently 
adopted for use on board the vessels of the United States Fish Commission. Fahren- 
heit thermometers were used, and the densities given by the graduation of the hydrom- 
eter were referred to pure water at 00° F. The centigrade scale was adopted by the 
Coast and Geodetic Survey and by the United States Fish Commission in 1800, and 
all hydrometer observations since then are referred to that scale, the densities being 
reduced to the temperature of 15° C. referred to pure water at 4° C. 
Hilgard’s ocean salinometer is composed of a series of cylindrical glass floats num- 
bered 1, 2, and 3, respectively, their bodies inches in length by 1^- inches in diame- 
ter, the stems 44 inches in length and inch in diameter, the scales being marked on 
the interior of the stems. The range of No. 1 is from 1000, or fresh water, to 1011; 
No. 2 from 1010 to 1021, and No. 3 from 1020 to 1031, which gives sufficient range from 
fresh water to salt, including the etfect of temi)erature. Each unit in the third place, 
or thousands of the density of fresh water, is represented by a length on the stem of 
0.3 of an inch, which is subdivided into live parts, admitting of an accurate reading 
of a unit in the fourth place of decimals by estimation. 
The vessel for holding the specimen of water is of copper, cylindrical in form. Of 
inches in length and IJ inches in diameter, with a base of 3f inches. The water cup 
itself is 8^ inches long and I J inches in diameter. The attached thermometer is 
inclosed within a water-tight, glass-faced frame, secured to the cylindrical body of 
the cui), a section of the latter being removed to allow free circulation of water 
between its interior and the thermometer. The temperatures are read through the 
glazed front. The floats are packed separately in If by lOf inch tubes or cylinders of 
tin, in Avhich they are protected by cotton, and the stems have additional protection 
of hollow, cotton lined, wooden sleeves, which envelop them, slip inside of the tin 
tubes, and imj)inge upon balls of cotton, with which their covers are lined. 
A working set of three floats and a cup, the former in their tubes of tin, are 
packed in neatly fitting woolen-lined apertures in a handy Avooden box 11 inches long, 
10 inches wide, and 4 inches high, inside measurement. Sx>are floats are also ])rovided 
with tin tubes, in which they are always iiacked, either for transijortation or stowage. 
r. C. B. 1896— 2‘J 
