NARRATIVE OF THE CRUISE. 99 



Notwithstanding previous experience, when H.M.S. "Lightning" was employed in 

 sounding and dredging in the Fseroe Channel in 1868, she was supplied with unprotected 

 thermometers. On her return a number of interesting experiments were made by 

 Professor \V. A. Miller and Mr. Casella, to find the " pressure correction " for the instru- 

 ments used. 1 The corrections so found, though good for the thermometers actually 

 experimented on, are of no use for correcting other instruments, even though they may 

 be of the same pattern. This is due to the fact that the bulbs of even the most 

 carefully made thermometers are never uniform in thickness of glass, and consequently 

 yield differently to pressure. It has been mentioned above that the stem of the 

 thermometer suffers compression though the effect so produced on the reading 

 of the thermometer is insignificant. If the stem be uniform, the effect will be pro- 

 portional to the length utilised. In the case of the thermometers supplied to the 

 Challenger, the bore of the stem was not uniform. Close to the bulb there was a swelling, 

 and at the bend there were other swellings. As it was the minimum limb that was almost 

 exclusively used, the effect of pressure on the reading was limited to that produced on 

 about 2 inches of tube with a slight swelling near the neck of the bulb. Along with 

 each instrument was supplied, as " pressure correction," the amount to be deducted from 

 the reading according to the depth to which the instrument had been sent. During the 

 first part of the cruise this correction was applied without question, and the results 

 embodied in reports with sections sent home to the Admiralty. As, however, observations 

 multiplied, and side by side with the thermometric observations experiments were made on 

 the effects of pressure on various substances contained in piezometers, the readings of which 

 required to be " cleared for temperature " the question of the validity of the " pressure 

 correction " came to be seriously considered, and the conclusion was come to that it had been 

 improperly applied. It was obvious that the correction referred to could have been obtained 

 only in one way, namely, by submitting the thermometer to pressure in a hydraulic machine, 

 and noting the rise of the maximum index. This rise would be caused by the compression 

 of the stem forcing the liquid up the tube, and by the actual rise of temperature produced 

 by the compression of the water of the hydraulic receivers. It was at once evident that 

 the part due to actual rise of temperature caused by compression must be rejected 

 altogether, because in use the thermometer takes the temperature of the water in which 

 it is immersed. Of the residual amount due to actual compression of the stem, only so 

 much ought to be taken as is applicable to the portion of stem between the bulb and 

 the mercury on the minimum leg. This would as a rule be about one-sixth of the 

 length of stem from the bulb to the mercury meniscus in the maximum leg, without 

 adding anything for the swellings at the bend. The errors for a pressure of 3 tons per 

 square inch varied from 1° to 1^° F., and even if nothing be rejected for heat effect, the sixth 

 part would be considerably less than the probable error of observation. The temperature 



1 Depths of the Sea, p. 295, 1873. 



