488 



SCIENCE. 



[N. S. Vol. VIII. No. 198. 



suggested, I prepared a quantity of pulver- 

 ized soda glass free from lead and placed 

 about 120 grams of it in a large combus- 

 tion tube connected with the apparatus 

 already described. The glass was made 

 from an old stock of tubing, and was of 

 such a degree of fineness that it all passed 

 through a sieve of 90 meshes to the linear 

 inch, but would not pass through a 140- 

 mesh sieve. The combustion tube was 

 adapted to be heated by a gas furnace with 

 automatically regulated gas supply, so as 

 to be maintained at any desired tempera- 

 ture. 



While prepared to make analysis, in the 

 usual way, of the gas evolved by the pul- 

 verized glass in case it appeared in sufficient 

 quantity, I relied chiefly on its heat con- 

 ductivity for the detection of any large pro- 

 portion of hydrogen in the last few mill- 

 ionths, not being prepared at that time for 

 spectroscopic examination ; and thinking 

 that, perhaps, some hydrogen might be 

 evolved in the last stages of exhaustion be- 

 fore heating, I tested the conductivity of 

 the residual air from time to time as the 

 preliminary exhaustion progressed. 



When the exhaustion approached a good 

 vacuum an astounding phenomenon de- 

 veloped. At 36 millionths pressure the 

 residual gas conducted heat twice as well 

 as air, and nearly as well as hydrogen ; at 

 3.8 millionths it conducted seven times as 

 rapidly as hydrogen; at 1.6 millionths, 

 fourteen times, and at .96 of one millionth, 

 twenty times as rapidly. These results are 

 plotted in the curve e /. I did not carry 

 the exhaustion lower than .96 M. At this 

 pressure the time of cooling of the ther- 

 mometer from 15° to 10° was only 177 

 seconds, while pure hydrogen would have 

 required 288 seconds. 



Evidently a new gas of enormous heat 

 conducting capacity was present, mixed 

 with the last small portion of air. It must 

 have come from the pulverized glass, and 



probably formed only a small fraction of 

 the mixture. The last cc. of gas pumped 

 out was collected, and upon subsequent 

 analysis of it nothing but air was found. 



The discovery of the new gas, as above 

 outlined, was made nearly a year and a 

 half ago, on March 10, 1897. On the fol- 

 lowing day the pressure had increased to 

 nearly five millionths ; while the time of 

 cooling of the thermometer, instead of di- 

 minishing as usual with increase of pres- 

 sure, had raised from 177 seconds to 245 

 seconds, showing that the new gas had 

 largely disappeared. This was subsequently 

 found to be due to absorption by the phos- 

 phorus pentoxide used in desiccation. 



Long continued moderate heating of the 

 pulverized glass caused the evolution of 

 some air, much carbon dioxide and hydro- 

 gen, some carbon monoxide, and more of 

 the new gas. The selective absorption 

 which occurred on cooling was confined 

 almost wholly to the hydrogen and new gas. 

 After further moderate heating and thor- 

 ough exhaustion, I raised the temperature, 

 continued the exhaustion, and got the re- 

 sults embodied in the curve g h, showing 

 that much of the new gas was present ; the 

 conductivity at .38 millionth, the last sta- 

 tion in the curve, being tiventy-seven times 

 that of hydrogen. Several days of high 

 heating, with frequent exhaustion, failed to 

 wholly deprive the pulverized glass of its 

 new gas, though the output was greatly di- 

 minished. For this and other reasons, I 

 believe that the new gas resides in and not 

 simply on the surface of glass. 



To make sure that the apparatus was not 

 deceiving me, I sealed off the combustion 

 tube, heated the vacuous parts several 

 hours by means of air and water baths as 

 before indicated, admitted air dried over 

 phosphorus pentoxide, exhausted step by 

 step, and got the data for the ' small-bulb ' 

 air curve shown here and in last year's pa- 

 per. Evidently I had not been deceived 



