THURAS: DENSITY MEASUREMENTS 609 



which is wound 75 ohms of "advance" resistance ribbon, whereby 

 the bath can be rapidly heated electrically. The base contains 

 a stirrer and inlet and outlet tubes for cooling water. The upper 

 part of the bath is made of ordinary glass tubing and contains a 

 copper cover through which a thermometer and the test tube 

 containing the liquid can be inserted. 



A standard stem type thermometer graduated in 0?1 is se- 

 cured in the cover of the bath with its bulb immersed in the 

 liquid of the bath. The temperatures can easily be read to 0.1 

 of a division. 



The method of making a determination depends on noting the 

 precise temperature at which the liquid to be tested is exactly 

 of the same density as the bobbin. It is apparent that this 

 point, where the bobbin neither sinks nor rises, must occur in a 

 particular liquid at a perfectly definite temperature. In an 

 actual test this equilibrium temperature is not obtained, but 

 pairs of readings approaching it from a higher and from a 

 lower temperature are successively made until a pair is ob- 

 tained covering such a small range of temperature that their 

 mean may be regarded as sufficiently close to the equilibrium 

 temperature to give the required accuracy in density. The 

 temperature coefficient of sea water is such that a sufficient 

 degree of accuracy can be obtained by determining this equi- 

 librium temperature to 0?05 C. The rapidity in reaching these 

 points depends on the viscosity of the water and, consequently, 

 increases with the temperature. This in some measure com- 

 pensates for the decreased sensitivity caused by the increased 

 temperature coefficient at higher temperatures. Therefore a 

 high degree of accuracy in a limited time is not necessarily ob- 

 tained by cooling the sea water to such a temperature that the 

 coefficient of expansion is small. Consequently that part of the 

 temperature-density curve of sea water which is most conven- 

 ient to work with has been selected in the present work, and 

 the bobbins have been made of such a density that all of the 

 equilibrium temperatures come between room temperature and 

 about 40°C. 



In building this density apparatus the requirements for use 

 at sea have been constantly kept in mind. The essential 



