REPORT ON THE PRESSURE ERRORS OF THE THERMOMETERS. 13 



effects produced by pressure upon the material of a tube, according as the pressure is 

 applied from without or from within. 



First, with regard to the thermometers themselves, which are exposed to external 

 pressure, but have comparatively very slight pressure applied in the interior of their 

 bore ; and second, the corresponding effect when pressure is applied, as in the press 

 itself, from the inside and tends to stretch the walls. [This second case has occurred 

 with one or two of the Challenger thermometers also. Its source is usually defective 

 strength of the terminal bulb of the maximum end of the tube. This bulb implodes, 

 then the pressure is applied to the interior of the protected bulb, which, in its turn, 

 explodes.] 



In the diagrams below, the first three figures refer to part of the walls of the glass 

 tube, which is exposed to pressure from the outside, but has no corresponding 

 pressure applied within. The effects of pressure indicated are those in a transverse 

 section of the tube. The circles represent (on a large scale) transverse sections of very 

 small spherical elements of the glass wall of the tube, the first close to the outside, the 



second in the middle of the wall of the tube, and the third close to the inner surface. 

 The ellipses which are drawn along with the circles represent (of course, with much 

 exaggeration) the corresponding transverse sections of the ellipsoids into which the 

 spheres are distorted by the external pressure. The sphere near the outside is com- 

 pressed in all directions, but much less in a radial direction than it is in a direction 

 perpendicular to the former. The greatest amount of compression is tangential as it were, 

 and the circular section of the sphere has been compressed into an ellipse which has a 

 major axis in the radial direction very nearly equal to its original length, while the 

 minor axis is very considerably reduced. The second figure refers to a small spherical 

 portion inside the glass wall originally situated at a distance from the axis equal to 1 '6 

 times the internal radius of the tube. (It is curious that the number 1/6, though 

 obtained from a totally different source, should be so nearly the same as that already 

 quoted as the refractive index of the glass.) The little spherical element at that place 

 suffers no radial compression, but there is considerable tangential compression. Close to 

 the interior surface of the glass tube we find large compression in a tangential direc- 

 tion and actual extension in the radial direction. These diagrams have been purposely 

 exaggerated to make the effects visible. They represent what would be the effect of a 



