1899.] 



on Measuring Extreme Temperatures. 



109 



give so uniform a scale, or so simple a formula. It has the great 

 advantage, however, that the theoretical conditions of flow may bo 

 realised with unlimited accuracy, and that the transpiration resistance 

 can be measured with a degree of precision very little, if at all, 

 inferior to the corresponding electrical measurement. 



The complication of the transpiration problem arises from the 

 fact that the flow depends on the increase of the viscosity of the gas, 

 as well as on its expansion. The viscosity of liquids in general 

 decreases very considerably with rise of temperature. That of water, 

 for instance, is six times less at the boiling point than at the freezing 

 point. If the viscosity of gases diminished in a similar manner, it 

 might happen that the transpiration resistance would decrease with 

 rise of temperature. Maxwell was the first to give a theoretical 

 explanation of the behaviour of gases in this respect. On certain 

 simple kinetic assumptions, he showed that the viscosity should 

 increase in direct proportion to the absolute temperature. Since the 

 expansion follows the same law, the transpiration resistance on 

 Maxwell's hypothesis should increase in proportion to the square of 

 the temperature. This would give a fairly simple formula, and would 

 make the transpiration thermometer a very sensitive instrument, but 

 the scale would be very far from uniform. Maxwell made some 

 experiments on the temperature variation of the viscosity between 0° 

 and 100° C, which appeared to give support to his mathematical 

 assumptions ; but his apparatus did not happen to be of a very suitable 

 type for temperature measurement, and it is clear that he did not 

 regard this part of his experimental work with great confidence. 



The question of the viscosity of gases was next attacked with 

 great vigour in Germany by a number of different physicists. They 

 ultimately succeeded in proving that the law was not quite so simple 

 as Maxwell had supposed, and that the rate of increase of viscosity 

 was less than that of volume. A summary of some of the principal 

 results obtained, over the range 0° to 100° C, is given in the following 



Table III. — Variation of Viscosity v with Temperature 

 Formula, v/v = (T/T ) n . 



T. 



Observers. 



Maxwell 



Meyer 



Puluj 



Obermeyer 

 Wiedemann 



Warburg 



„ and Kundt 

 Holman 



Dates 



Values of Index n (0° to 100° C.) 



Air. 



C0 2 



■94 

 •93 



