720 



Mr. A. Holt on the 



or the volume o£ gas to be collected was very small, 

 the percentage o£ oxygen absorbed was large, but 

 if a considerable volume o£ gas was rapidly collected 

 the percentage absorption diminished. 

 (2) When the taps B and C were opened in order to pump 

 off the gas and water vapour at the end o£ each 

 experiment, the hydrogen, on account of its great 

 power o£ diffusion compared with oxygen, passed 

 rapidly through the condenser to the pump, leaving 

 some o£ the oxygen behind condensed on the glass 

 surfaces o£ the apparatus and dissolved in the water 

 in the condenser. On carefully warming the glass 

 this oxygen could be pumped out. The larger 

 the amount of electrolytic gas produced the smaller 

 the percentage of oxygen remaining on the glass 

 surfaces. No perceptible quantity of oxygen 

 was absorbed by the platinum wire, nor was 

 hydrogen evolved from it when it was heated to 

 incandescence. 



This excess of hydrogen was always estimated and calcu- 

 lated as electrolytic gas. 



The two following tables give the author's results. 



The first consists of experiments carried out at almost 

 constant temperature, and under as similar conditions as 

 possible; the second gives the values obtained at a series of 

 temperatures ranging from 1233° to 1637°. 



Table I. 



Temp. abs. 



Percentage decomposition 

 at 760 mm. pressure. 



K. 



1620° 



005107 



3-786 



1621 



005421 



3-808 



1620 



0-04965 



3-774 



1620 



0-04686 



3-749 



1620 



005102 



3-786 



1620 



005116 



3-787 



1620 



0-04808 



3-760 



1620 



004605 



3-741 



1620 



005014 



3-778 



1620 



0-05037 



3-780 



1620 



0-05111 



3-786 

 Mean =3'776 



