8 



DENSITIES OF OXYGEN AND HYDROGEN 



then reduced ti> .suit the engraved scales. Theimunietei' number 741 was of the 

 Jena normal ghiss ; G was of the glass of which Geisslei- made his thei-mometers. 



The boiling point was determined after half an lK)ur of ex- 

 posure to steam. The freezing point was detei'mined by noting 

 the ma.xiiiniiii ilepre.ssed zeiu within two or three minutes after 

 the last e.xposure to steam. 'i'he zero point w;i.s frequently 

 redetermined. 



Temjieratures determined with tiie two thermometers 

 agreed as closely as different readings of the same thermometer. 

 To the scale of these two thermometei-s were reduced the read- 

 ings of the other thermometers used for the temperatures of 

 Fig. 3.— coiiarandp.-in tjje water in the hydrostatic weighings. 



for submersion of (''<il>es. •' o o 



DETAIT-S OK A II VDHOSTATIC \V KH; II 1 .\(i. 



To give a dear idea of the details of the process, part of the determination of 

 the volume of globe ^'o. 1 is here given : Hr.st, the observations made on April 

 7th; secondly, the computation for this day ; and, thirdly, the results for eacli of 

 the other days on which tliis globe was weighed in watei'. 



HYDROSTATIC WKK; II I N<;, AI'KII. 7tll. 



On ])alance F,. Globe 1, cage and pan, weights ABCDEFCHM. l.-ft. 



HOUR. 



TEMPERATURE BY SIX THERMOMETERS. 



WEIGHTS, RIGHT. 



In the reduction, the indi<'ations of the tln-iinoinetei's at tlic top ;iii(l l)ottoni of 

 the cylinder of water were not included in the mean. One of tliese was somewhat 

 above the globe, and the other was somewhat below. They served to show the 

 presence or absence of tlicnnal inequalities producing currents; tlie other ther- 

 mometers fairly .showed the temperature of the water in whicli the globe was 

 lloatiiiiT. 



