0. Barns — Viscosity of Gases at high temperatures. 409 



son is made in the fourth column and the residual errors 

 inserted in the final column. The data given are typical 

 values. £ symbolizes Helmholtz's " Gleitungs-coefficient." 



From the table it follows that for the range of temperatures 

 within which I have observed the mean increase of gaseous 

 viscosity takes place proportionally to the two-thirds power of 

 absolute temperature. Interpreted by aid of the well-known 

 Clausius-Maxwell relations, the results of the table may be 

 stated succinctly thus : The mean free path of the molecule of 

 a perfect gas varies directly as the sixth root of its absolute 

 temperature. I had hoped to find that at temperatures suffi- 

 ciently high the mean free path, would be independent of tem- 

 perature, a law to be regarded as a criterion of a perfect gas 

 and for which the experiments of E. Wiedemann when used 

 to interpret the low-temperature results of O. E. Meyer, 

 Puluj, Warburg, Obermayer, and particularly the admirable 

 researches of Holman* seemed to contain suggestive evidence. 

 But after applying many devices for the removal of errors, I 

 found that my original results were not essentially changed. 

 Accepting the law of sixth roots as indicating perfect gaseity 

 (i. e. the non-occurrence of ephemeral mechanically cohering 

 molecular aggregates) it appears that the linear magnitude, 

 mean free path, is proportional to the cube root, of the 

 velocity of the mean square, — a singularly suggestive result. 



The chief discrepancy of my work is this, that the tempera- 

 ture measured externally is not identical with the temperature 

 at which transpiration actually occurs. Taking the transpira- 

 tion data alone they show a surprising degree of accordance 

 even above 1300°. If [#"] be the temperature computed from 

 transpiration data under assumption of the above law, I found 

 in successive measurements, for instance : 



6" 



in 



0" 



m 



0." 



m 



0" 



[H 



436° 



446 



455 



450° 



459 



469 



568° 

 570 



575 



575° 



577 



581 



975° 



981 



990 



971° 

 972 



982 



1210° 



1210 



1209 



1245° 



1247 



1245 



If the law governing the thermal variations of the viscosity 

 of a gas were rigorously known, Poiseuille-Meyer's equation 

 applied to transpiration data would enable us to measure tem- 



* After a careful consideration of his own results and those of all earlier ob- 

 servers, Mr. Holman has discarded exponential relations altogether. For my own 

 part, I believe that at the present stage of research a conservative policy is the 

 wiser. In chemistry the hypothesis of residual affinity is fast gaining ground. 

 Hence before final decision can be made, it will be necessary to have exhausted 

 data for an interval of temperature (say 500° to 1000°) within which the ephem- 

 eral molecular aggregates in question may reasonably be assumed to be absent. 

 Discussion must be reserved for the Bulletin. 



