E. W. Morley — Volumetric Composition of Water. 281 



stopcock. When the mercury rose to the next standard point, 

 level and temperature were measured, and another tube placed 

 to collect water, while the water already issued was weighed. 

 True volumes at zero were computed from the weights and 

 temperatures observed. The calibration was made four times ; 

 I give the values found for the points at which oxygen and 

 hydrogen were measured in the experiments, and the differ- 

 ences between the mean and the several determinations. It 

 will be seen that the mean error of a single determination of 

 volume is five cubic millimeters. A tube was previously cali- 







Errors (cubic millimeters). 



Scale Division. 



Volume found. 



| 











1. 2. 



3. 



4. 



600 



184-323 



• 



12 



- 7 



- 4 



625 



191-849 



-3 — 4 



4 



4 



650 



199-442 



1 - 2 



10 



-10 



675 



207082 



7 -15 



2 



7 



700 



214766 



-1 3 



5 



— 7 



725 



222-460 



1 4 



-10 



5 



750 



230-132 



1-1 — 3 — 1 



5 



brated w T ith a mean error only three fifths as much ; but it 

 broke during the first experiment. The sky was so cloudy 

 while this the present tube was calibrated that illumination 

 was defective, and work very trying. But in the determina- 

 tion of the volumetric composition of water, any slight errors, 

 either in the calibration of the standard points, or in the inter- 

 polation at intermediate points, was nearly eliminated by a 

 proper distribution of the points used in the measurements. 



Since, in my way of manipulating, the gas, scale, mercury, 

 and eudiometer are all at the same temperature, the effects of 

 the expansions of all were taken into account in one factor, 

 which moreover, was determined for the actual degrees of the 

 thermometer used. 



In measuring the volume of a quantity of gas in the 

 eudiometer, the mercury in the recurved capillary tube was 

 Drought to a certain mark. The level of the mercury in the 

 eudiometer was so adjusted that when the pressure tube was 

 opened, the mercury in it would stand near one of the Jolly 

 points. Then the level of the reading microscope was adjusted, 

 and the meniscus was brought into focus. During this time, 

 the water surrounding the eudiometer was stirred by a current 

 of air. When everything was ready, the current of air was 

 shut off, and the piston valve slowly moved till the mercury 

 exactly coincided with the Jolly point. Then, within five or 

 ten seconds, the thermometer was read and the fraction of a 



