14 



SUPPLEMENT. 



PRESSURE ERRORS OF MILLER-CASELLA THERMOMETERS. 



The following report is taken from "Nature" issue of April 21, 

 1881, page 595. It treats of a matter contained in the original volume 

 on Deep-Sea Sounding and Dredging, page 109: 



•'Royal Society, April 4. 

 "Professor Tait communicated the results of his experiments on the 

 pressure errors of the Challenger thermometers, the correction for which, 

 as originally furnished to the expedition, was 0°.5 F. per mile of depth. 

 The mode of experimenting was' to subject the thermometers to consider- 

 able pressure in a hydraulic press, which was essentially a strong steel 

 cylinder that was warranted to stand a pressure of 25 tons weight on the 

 square inch. It was supported in an upright position upon a strong 

 tripod stand. Water was filled in from above; and into the upper end of 

 the cylinder there was lowered a tight-fitting plug, which was fixed in 

 position by a transverse steel bolt. The lower end of the cylinder was 

 connected through a narrow copper tube to a hydraulic pump, which, by 

 pumping in water to the cylinder, raised the pressure to the required 

 amount. At three tons pressure an average effect of 1°.5 F. was produced 

 upon the inclosed thermometers. Before drawing any conclusions as to the 

 correction to be applied in deep-sea sounding, it was necessary to consider 

 how far this effect could be explained as resulting from the peculiar con- 

 ditions under which the experiments were made. From the known com- 

 pressibility of glass it was calculated that the volume of the bore of a 

 thermometer tube, closed at both ends, would be diminished by only one- 

 thousandth part for an increase of pressure of one ton weight on the square 

 inch; and from a direct experiment made with a metre-long tube this was 

 proved to represent very approximately the real effect. . Hence it was quite 

 out of the question that this could have any appreciable effect on such 

 comparatively short thermometers as those of the Challenger, which were 

 besides subject to much graver errors, such as those arising from the 

 shifting of the indices during the ascent from the depths, or even from 

 the effect of parallax when taking the reading. The direct action of 



