338 
BULLETIN OF THE UNITED STATES FISH COMMISSION. 
To observe tbe specific gravity of a specimen of sea water, fill the cup, clean and 
dry the float, lower it carefully into the water, causing it to overflow from the top of 
the cup, and, when the float has come to a rest, read off the scale at the surface of the 
water, not from the water immediately surrounding the stem, where it is slightly 
elevated by the effects of capillary attraction ; note temperature of water by attached 
thermometer, and correct obserwed density by means of subjoined table. 
As before stated, the variation in specific gravity of sea water is so small that the 
greatest accuracy is required to give value to observations, and experience alone will 
teach the observer how difticnlt it is to obtain satisfactory results on shipboard when 
the vessel is under way, i)articularly if there is much motion. In order to avoid this 
fruitful source of error, a supply of the best quality of glass bottles, with ground-glass 
stoppers, were procured from the manufacturers, 0. Dorfiiuger »& Sons, who describe 
the material as follows : 
This grade of glass is what is called “lead glass” and has been made especially for the work of 
the United States Fish Commission and the National Mnsenm. It is a glass that is A’ery suitable for 
making specimen jars and work of this character, and is composed of sand, red lead, pearl ash, nitre, 
arsenic, and manganese. The quantity of arsenic is so small that it burns out in the melting and 
there is no trace of it in the glass, neither is there anything in it that sea water will affect. 
The water specimens are carefully sealed in these bottles until calm, smooth 
weather, or till the vessel reaches iiort, when the temperature of the specimens is 
brought as nearly to the required standard of 15° O. as convenient and the densities 
carefully observed. It has been said that the stoppers might not be tight, foreign 
substances might be accidentally introduced into the bottles, or that the sea water 
might attack the material of the glass itself, thereby cliangiiig the si)ecitic gravity, 
all of which might have happened had not proper xirecautious been taken. 
In the first jdace, the quality of glass is xu'oof against the action of sea water, 
at least for the short sjiace of time the specimens are exi>osed ; the bottles are carefully 
cleansed and dried before using, and they are sealed with as much care as though they 
contained volatile matter. The water specimens are retained no longer than necessary, 
usually from a day to a week, never more than two weeks, and the densities are 
observed with great care, under the most favorable conditions. 
Hear- Admiral Makaroti’, Imperial Eussian navy, author of Le Vitiaz et I’Oceau 
Pacitiqne, St. Petersburg, 1894, and one of the highest authorities on the specific 
gravity of sea water, thought the saliuometer cu}) was too small, but his opinion was 
based nx)on a casual iusx>ection only, as he had never seen it in ox)eration. Its adoiition 
by the United States Coast Survey and Fish Commission was niiou the approval of 
the most eminent physicists in the United States, and it has been in constant service 
for many years without eliciting unfavorable comment in a single instance; on the 
contrary, it has been commended in the highest terms as the most simx)le and thoroughly 
X>ractical appliance ever introduced for the piuq)ose. 
The table for the reduction of observed densities to 15° C. is taken from Dittmar, 
Physics aud Chemistry, Challenger Exi^edition, vol. 1, after af)i)lying to his densities 
one-half of the correction given by him for reducing them to Thorpe and Pucker’s 
results. The teiux^erature atAvhich the density of standard water is 1.02600 has been 
shifted from 15.56° to 15° C. for the sake of getting an integer number. The table has 
been rearranged so as to give the densities of the standard water multiplied by 1000, for 
whole degrees and tenths from 0° to 30.9° C. on one page. The values given should 
have the figures 10 i)refixed; these are omitted for brevity’s sake. Thus in the table 
the density for 2.5° is jirinted 27.92, but this must be understood to mean 1027.92. 
