CHAP, vii.] DEEP-SEA TEMPERATURES. 289 



liquid and slightly compressed air. A small steel 

 index, with a hair tied round it to act as a spring 

 and maintain the index in any position which it 

 may assume, lies free in the tube among the creo- 

 sote at either end of the column of mercury. This 

 thermometer gives its indications solely by the con- 

 traction and expansion of the liquid in the large full 

 bulb, and is consequently liable to some slight error 

 from the effect of variations of temperature upon 

 the liquids in other parts of the tube. When the 

 liquid in the large bulb expands, the column of mer- 

 cury is driven upwards towards the half-empty bulb, 

 and the limb of the tube in which it rises is graduated 

 from below upwards for increasing heat. When the 

 liquid contracts in the bulb, the column of mercury 

 falls in this limb, but rises in the limb terminating in 

 the full bulb, which is graduated from above down- 

 wards. W T hen the thermometer is going to be used the 

 steel indices are drawn down in each limb of the tube 

 by a strong magnet, till they rest on each side on 

 the surface of the mercury. When the thermometer 

 is brought up, the height at which the lower end of 

 the index stands in each tube indicates the limit to 

 which the index has been driven by the mercury, 

 the extreme of heat or cold to which the instrument 

 has been exposed. 



Unfortunately, the accuracy of the ordinary Six's 

 thermometer cannot be depended upon beyond a 

 very limited depth, for the glass of the bulb which 

 contains the expanding fluid yields to the pressure 

 of the water, and, compressing the contained fluid, 

 gives an indication higher than is due to tem- 

 perature alone. This cause of error is not con- 



u 



