42 THE WATER 



depths and brought to the surface under lower pressure 

 will undergo a sensible lowering of temperature. The 

 materials of which the water-bottle is made will 

 similarly be cooled, and the total error may reach a large 

 fraction of a degree. It is certainly possible to calculate 

 this error, but the rate at which the loss of heat of the 

 various parts affects the temperature of the central 

 chamber and the thermometer cannot be estimated 

 accurately. Ekman has drawn up a table in which 

 the calculated corrections are reduced by one-half 

 in order to allow for this latter uncertainty (p. 43). 

 It will be seen that the error becomes larger as the 

 temperature and the depth increase, and below 1,000 

 metres the insulating water-bottle should not be used.* 



Reversing Water-Bottles and Thermometers. 



The reversing thermometer is made so as actually to 

 register the temperature at any moment by the simple 

 process of turning it upside down, and, apart from faults 

 of construction, which unfortunately are difficult" to 

 avoid, it gives extremely good results. It differs 

 from an ordinary thermometer in having a constriction 

 and S-shaped dilatation immediately above the main 

 bulb, and in having a somewhat large secondary bulb 

 at the upper end of the stem. The graduations are 

 reversed so that the lowest temperature is marked 

 near the top of the capillary portion. There are two 

 well-known models, the original form, brought out by 



* For serial samples from great depths, for which the 

 insulating water-bottle is not adapted, the handy and inex- 

 pensive "Richard " or "Monaco " bottle, worked by a messenger, 

 should prove a useful instrument, if strengthened and provided 

 with a place for a second thermometer. It is a modification 

 of the large form used by Buchanan in H.M.S. "Challenger" 

 but has been somewhat over-lightened. — Ed. 



