INSULATING WATER-BOTTLE 41 



The interval which should be allowed for the water- 

 bottle to take up the temperature of the water depends 

 on the variations to which it is exposed. Six or seven 

 minutes should be sufficient for a very wide difference, 

 and when the water is of the same temperature at all 

 depths, and the bottle has been kept in a place of 

 nearly similar temperature, half a minute will suffice. 

 In the form described the releasing gear is actuated 

 by a messenger. Bottles can also be bought fitted 

 with a propeller release, but the latter is uncertain in 

 action, and the time saved by its use is negligible in 

 the moderate depths to which the insulating water- 

 bottle should be confined. 



The advantages of the Nansen-Pettersson instru- 

 ment are its accuracy and sensitiveness within its 

 proper limits of depth, but it is somewhat heavy, and 

 its construction renders it impossible to use a number 

 on the same wire. 



The thermometers used with it are of the ordinary 

 form, graduated to T V° C. on the stem, and protected 

 against pressure by an outer glass tube. The space 

 between the bulb and the outer tube should be partly 

 filled with mercury to insure its taking the temperature 

 of the water quickly. 



The theory of the insulating water-bottle is too 

 complicated for treatment here, but a very full dis- 

 cussion is to be found in a paper by Dr. Walfrid Ekman 

 (Pub. de Circ, No. 23). There is, however, one serious 

 defect, of which mention must be made here, which 

 renders all such instruments useless for the determina- 

 tion of the temperature at more than about 1,000 metres, 

 and it is unfortunately in the deeper layers that the 

 greatest accuracy is desirable. When water is strongly 

 compressed and then allowed to expand, it becomes 

 colder at the same time, and water collected at great 



