REPORT ON THE PRESSURE ERRORS OF THE THERMOMETERS. 17 



Accurate Measurement of great Pressures. 



It will be obvious from what has been said, especially as regards the old apparatus which 

 was carried about in the Challenger, that one of the most essential requisites of the whole 

 investigation was the accurate measurement of pressure. All the ordinary forms of 

 pressure-gauge were found to be untrustworthy. It was necessary that in all cases the 

 pressure should be measured with certainty to about 1 per cent. No attempt was made 

 to secure any greater degree of accuracy, as the indications of the thermometers them- 

 selves could not in any case be trusted to less than 0°'l Fahr. It is a vain but too 

 common custom to try to make some parts of an experimental measurement exact to a 

 greater degree than can possibly be attained in the rest. But it is mere waste of time. 



The basis on which, after a great many trials, I finally founded my determination of 

 pressures, was Amagat's T remarkable measurements of the volume of air and other gases 

 at high pressures. Amagat's data were obtained in the most direct and satisfactory 

 manner, inasmuch as he measured his pressures by means of an actual column of mercury 

 extending sometimes to 300 metres, and more. All other means of measuring pressure 

 are as it were valueless in comparison with this. We know by these experiments the 

 compressibility of nitrogen, and of air, up to pressures of at least 1\ tons weight per 

 square inch, with almost all desirable accuracy. 



All that was necessary therefore in order to determine the pressures in the operating 

 cylinder, and thus to calibrate the gauges employed, was to compress once for all a quantity 

 of air, measure the volume to which it was compressed and the corresponding indications 

 of the gauges, and then by the help of Amagat's tables compute the pressure actually 

 attained. The apparatus I employed for this purpose is figured in section in the diagram 

 below. 



This apparatus, filled with dry air, was allowed to come exactly to the temperature of 

 the water inside the compression apparatus ; then, the lower end of it being dipped into 

 a large vessel of mercury, it was let down full of air into the compression cylinder and 

 pressure was applied. The effect was of course to compress the air, force up the mercury 

 until it gradually filled the vessel and forced the air entirely into the smaller bulb. After 

 a few trials we found roughly what amount of pressure was necessary in order just to 



1 " Memoire sur la compressibility ties gaz a des pressions elevees, par M. E.-H. Amagat " (Ann. de Clmnie et de 

 Physique, 1880). 



