290 THE DEPTHS OF THE SEA. [CHAP. vn. 



stant in its action, as the amount to which the 

 bulb is compressed depends upon its form and upon 

 the thickness and quality of the glass ; thus the 

 error of good thermometers of the Hydrographic 

 Office pattern varies from 7 C. to 10 0> 5 C. at a pres- 

 sure of 6'817 Ibs. on the square inch, representing 

 a depth of 2,500 fathoms. In thoroughly well- 

 constructed thermometers, however, such as those 

 made by Casella and Pastorelli for the English 

 Admiralty, the pressure error is tolerably constant ; 

 and Captain Davis, R.N., who has lately conducted 

 important experiments on this point, expresses his 

 opinion that by an extended series of observations 

 a scale might be obtained to correct the ther- 

 mometers hitherto in use to a close approximation 

 to the truth, and thus utilize to some extent obser- 

 vations which have been already made with our 

 ordinary instruments. 



In the ( Lightning ' expedition in 1868 we used 

 the ordinary Hydrographic Office pattern, and a 

 large number by different makers were sent with 

 us for testing and comparison. The depths not 

 being very great, the general temperature results 

 came out well, and were among the most singular 

 phenomena which we had to record. Many of the in- 

 struments were very wild at a few hundred fathoms, 

 and several gave way under the pressure. On our 

 return in April 1869, Dr. W. A. Miller, V.P.R.S., 

 attended a meeting of the Deep-Sea Committee of 

 the lloyal Society at the Hydrographic Office, - 

 and proposed encasing the full bulb in an outer 

 covering of glass containing air, in order to permit 

 the air to be compressed by the pressure of the 



