24 THE VOYAGE OF H.M.S. CHALLENGER. 



The ends of the tube containing the rod must, of course, he made of glass, or some 

 other transparent material. In the first apparatus which I used, tubes of soda-water- 

 bottle glass were employed, their bore being about - 2 inch, and the thickness of the 

 walls about "3 inch. The image of the small enamel bead at the end of the glass rod 

 was very much distorted when seen through this tube, but the definition was greatly 

 improved by laying on it a concavo-plane cylindrical lens (which fitted the external 

 curvature), with a single drop of oil between them. I found, by trial, that, had it been 

 necessary to correct for the internal curvature also, the employment of winter-green (or 

 Gaultheria) oil as the compressing liquid would have effected the purpose completely : 

 — the refractive index being almost exactly the same as that of the green glass. 



As the construction and mode of support of this apparatus did not enable us 

 completely to get rid of air from its interior, there were occasional explosions of a 

 somewhat violent character when the glass tubes gave way ; and the operators who 

 were not otherwise protected (as by the microscopes, for instance) were obliged to hold 

 pieces of thick plate glass before their eyes during the getting up of pressure. The 

 explosions not only shattered the thick glass tube into small fragments, but smashed 

 the ends of the experimental glass rod, so that a great deal of time was lost after each. 

 Only on one occasion did we reach a pressure of 300 atm., and an explosion 

 occurred before the measurement was accurately made. On these accounts, after four 

 days experimenting (the first being merely preliminary), we gave up working with this 

 apparatus : — and the results obtained by means of it cannot be regarded as wholly 

 satisfactory, though they agreed very well with one another. 



As a sudden shock might have injured the Amagat gauge, all the pressures were 

 measured by the old external gauge, whose unit is now determined with accuracy. 

 Hence the readings are in tons-weight per square inch (152"3 atm.), which are 

 below called "tons" as in the vernacular of engineers. Three of us at least were 

 engaged in each experiment, one to apply and measure the pressure, and one at each 

 microscope. Pressure, in each group of experiments, was applied and let off six or 

 seven times in succession, readings of the two microscopes being taken before, during, 

 and after each application of pressure. To get rid of the possible effects of personal 

 equation, the observers at the microscopes changed places after each group of 

 experiments (sometimes after two groups), so that they read alternately displacements 

 to the right and to the left. 



The values of the screw-threads were carefully verified upon one of the subdivisions 

 of the scale which was employed to measure the length of the experimental rod ; these 

 subdivisions having been since tested among themselves by means of a small but 

 very accurate dividing-engine of Bianchi's make. 



These experiments were made in July 1887, when the day temperature of the 

 room was nearly 20° C. In the last two groups the compression tube was surrounded 



