REPORT ON THE PRESSURE ERRORS OF THE THERMOMETERS. 39 



The fifth column gives the correction supplied to the Admiralty by Captain Davis for those 

 thermometers which he tested. The correction is, in all cases, for 2500 fathoms. With reference to 

 the numbers in this column, the following extract from a letter addressed to me by Captain Tizard (of 

 date 11th January 1SS1) must be kept well in view : — 

 m 



" The method employed by Captain Davis, in experimenting on the thermometers, was to place in 



the press, with the instruments on trial, one of Phillips' thermometers, enclosed in a tube on Sir William 

 Thomson's principle. He took it for granted that this perfectly enclosed thermometer would not be 

 affected by pressure ; but that any alteration of its index would be due to the generation of heat in 

 forcing the water into the press. The alteration of its index, which was always of very small amount, 

 was deducted from the alterations in the indices of the instruments on trial, and the differences 

 assumed to be the errors of the thermometers at a given pressure. But, as T mentioned to you before, 

 this alteration was always on the maximum side of the tube, and not the miminum. Consequently it 

 appeared to us to require considerable modification." 



On this it is necessary to remark that the indication of the Phillips' thermometer (as is obvious 

 from the text above) is not due to heating of the water in the press alone, but also to the heating 

 effect of pressure upon the strong protecting tube. Thus I have no direct means of comparing my 

 results quantitatively with those given by Captain Davis. 



Under the circumstances, I have done what appeared to me best for obtaining a rough comparison. 

 I have given in column 7 the observed effects of a pressure of three tons (nearly 2500 fms.) on each 

 thermometer. In column 8 the temperatures are given, and in column 9 the corresponding 

 temperature-change (by Thomson's formula. — Appendix C, ante), Column G gives the differences of 

 the numbers in columns 7 and 9 ; and these differences may be roughly compared with those of 

 Captain Davis in column 5. There is a general agreement, but my reduced numbers are, on the 

 whole, rather greater than those of Captain Davis. 



This may be ascribed, in part, to the fact that in Captain Davis' apparatus (as I understand) the 

 water was pumped in from above, and thus the heat developed by friction did not affect his results. 

 And it may be due in part to inadequate measurement of pressure, — a point which was impressed on 

 me from the very commencement of my work. I have learned from Mr Casella that the pressure 

 gauge employed by Captain Davis has been broken ; so that it is impossible now to verify his scale 

 of pressures. To show how possible is a serious mistake in this matter, I append a comparison of 

 the indications of the very elaborate gauge attached to the old Challenger apparatus with those 

 of my steel external gauge already described. The scale of the Challenger gauge is divided to 

 cwts. on the square inch. My gauge gives very nearly 20 mm. per ton ; so that, for a rough com- 

 parison, we may take 1 mm. as equivalent to 1 cwt. The two instruments were simultaneously 

 attached to the pump, and the pressure was therefore the same in both at each reading. There can 

 be no doubt whatever, from repeated comparisons with glass gauges of all sizes and shapes, that my 

 gauge follows Hooke's law with great accuracy. The only possibility of serious error is in the 

 actual value of the unit. This important determination has, however, been very carefully repeated 

 by the aid of Amagat's numbers and the indications of the silvered gauge already described ; and the 

 result is as above stated. 



