304 THE ATLANTIC. [chap. v. 



of the stem (which instead was widened into a funnel -shape 

 large enough to receive the ordinary decigramme weights), the 

 calibration of the stem was effected by loading the stem with 

 successive weights, and observing the consequent depressions 

 in distilled w\ater of known temperature. This done, the top 

 was sealed up and the instrument carefully weighed. The ex- 

 pansion of the body with temperature was determined in a 

 similar manner by reading the instrument in distilled water 

 of various temperatures. The co-efficient of expansion of the 

 glass was then found to be 0"000029 per degree centigrade. 



For using this instrument at sea about 900 cubic centimetres 

 of sea-water are taken, and the containing cylinder placed on 

 a swinging table in a position as near the centre of the ship as 

 possible. The observation with the hydrometer, loaded with 

 the necessary table and weight, is then effected in the ordinary 

 way, the accuracy of the readings being but little affected by 

 rolling. Pitching, however, is found to have a distinctly dis- 

 turbing effect ; and when it is in any way violent, it is advisa- 

 ble to store the specimen of water till the weather improves. 



The temperature of the water at the time of observation is 

 determined by one of Geissler's " normal " or standard ther- 

 mometers, graduated into tenths of a degree centigrade ; and 

 it is essential for the accuracy of the results that the water, 

 during the observation of the hydrometer, should be sensibly 

 at the same temperature as the atmosphere, otherwise the chang- 

 ing temperature of the water makes the readings of both the 

 hydrometer and the thermometer uncertain. At low tempera- 

 tures (below 10° or 12° C.) a tenth of a degree makes no sensi- 

 ble difference in the resulting specific gravity ; but at the high 

 temperatures always found at the surface of tropical seas, ris- 

 ing sometimes to 30° C, the same difference of temperature 

 may make a difference of 3 to 4 in the resulting specific gravity. 



Having obtained the specific gravity of the water in ques- 

 tion at a temperature which depends upon that of the air at 



