﻿Investigating Transpiration of Gases through Tubes. 253 



rate of flow, and the influence o£ end-effects, can be examined 

 by using successively smaller and smaller portions of the 

 same capillary, the tap T 3 being fixed in position throughout. 

 To compare the transpiration rates for different gases 

 under the same circumstances, a single capillary is used for 

 all the gases. The time of transpiration of the gas is first 

 found for a moderate pressure, and then, using air, an 

 adjustment of T 3 is made to give the same pressure on the 

 manometer. A trial adjustment is made with air to suit 

 the pressure obtained with each gas ; and thus the various 

 gases, being compared with air, can be compared with one 

 another. The particular pressure taken is immaterial, as 

 the ratio of the transpiration times of different gases under 

 the same circumstances depends only upon the gases*. This 

 ratio is the ratio of their viscosities, and hence if the viscosity 

 of air be known, the viscosities of other gases can be obtained 

 by comparison. 



The theory of these experiments, which are obviously 

 similar to those carried out by Poiseuille and Hagen on the 

 behaviour of water and other liquids in passing through a 

 capillary tube, may be easily developed by the analogy of 

 the theory of these hydraulic experiments. This theory f 

 leads in the case of gases to a law of speed which corresponds 

 fully with that found by Poiseuille for liquids, and like the 

 latter can be used to calculate the numerical value of the 

 coefficient of viscosity from the measured amount of the gas 

 transpired. 



I£ _/?i, p 2 be the pressures at the ends of the tube, Yj the 

 volume entering, V 2 the volume leaving per unit time, and jjl 

 the coefficient of viscosity, then 



2 ™ 2\ „ „,4 



/>iVi-ftV 2 -- 1M 



For another gas under the same circumstances 



whence 



- v '^ v '- ( -4^ 



or 



v\ 



-H 



v, 



H> 



t 



_/* 



t 1 



^ 



where /, t' are the transpiration times for a given volume. 



* Graham (he. ciL), 



t Meyer, Poggr. Ann, 1866. 



