503 



Returning to onr equation (1) we have for a monalotnic gas 

 2E\ 

 dp 



where for helium (', = 0.1190 



We l)ave now to express T as a function of p and ).. 



. / 7' \0.G47 



As tiie relation t\ = ?/„ 1 holds down to the boiling point 



V273.1, 



of hydrogen (comp. Homm. N°. 1346 March 1913) and as the thernio- 

 nieter-corrections are almost entirely due to that part of the ca- 

 pillary which is at a higher temperature than 20° K , we may 

 apply this formula to the whole temperature-range in the form 



,273.1 

 According to the expression for A given above, we have: 



dlijp = (1 + n) dlgT — dlj-k. 

 or also 



dk,p = {\-\n)dhjT^dUnj 

 if 



^ = T- 



With y as independent variable we may therefore write: 



it, dy 



J V J I 



The correction consists in our case in the sum of three corrections 

 for the different parts of the capillary, each with a different R. 

 For each of the three parts the integral might be easily found by 

 mechanical quadrature, taking into account the changing valu^ of 

 h„ as soon as the limits of the integration are known. We may 

 also for the sake of simplicity divide each part into smaller parts 

 such, that in the integration a mean value may be assumed for k\. 

 The limits are each time determined by the value of the viscosity 



dp 

 1) It follows from this expression, that there is a maximum value of — ^, 



(S. Weber Coinm. N'l 137c Sept. 1913). In arranging the measurements in 

 question care must be taken that at the place where this maximum occurs the 

 distribution of temperature is known as accurately as possible. 



Tlie determination of this maximum may possibly be of importance in the 



2ii 



investigation of the relation between fcj and ---. 



