495 . 



It will be seen tliat these requirements cannot all be fultilled 

 at the same time. A small value of (1) goes together with a large 

 value of (2) and similarly a minimum of (3) corresponds to a 

 maximum of (4). 



All we can do therefore is choosing the construction of the 

 thermometer such that in the intended measurements an optimum 

 is attained as regards satisfying the mutually conflicting requirements. 



A calculation of the order of magnitude of each of the afore- 

 mentioned disturbances, uncertainties or corrections will in general 

 sufficiently enable us to reach our object. 



It is clear, that we have to devote our attention particularly to 

 the capillary which connects the reservoir (at low temperature) with 

 the manometer (at (he ordinary temperature). 



As regards (1), the uncertainty regarding the distribution of tem- 

 perature along the capillary makes itself principally felt in the lowest, 

 coldest part of the capillary, where the density of the gas is highest ; 

 the narrower this part in proportion to the rest, the smaller the 

 uncertaiuty will be. 



It would of course be advisable, if possible, to avoid the calcu- 

 lation of the correction for that part of the dead-space which is 

 dependent on the capillary by placing an auxiliary c-ipillary according 

 to Chai'I'Uis beside (he capillary of the thermometer. ') In our case 

 we were imable to utilize this device owing to want of space in 

 the cryostat. It was all the more important, therefore, to take the 

 lower part of (he capillary as narrow as possible, from which it 

 follows in view of (2), that the capillary must be taken wider 

 higher up. 



At the lowest temperatures the question becomes of importance, 



whether helium still follows the gaseous laws. On the one hand the 



B B 

 term — in the equation of slate pv ^ A -\ has (o be considered, 



V V 



where B for a reduced temperature say of 0.2 acquires a fairly 

 high value, so that the cori'ection to be made on account of /J may 

 obtain an important intluence. As long as the equation of state for 

 helium is no belter known than is at present the case and the 

 calculation has to be made with the "luean" equation of state accord- 

 ing to the law of corresponding states, great uncertainty exists with 

 regard to this correction. On the other hand it might be a question, 

 whether yl may still be taken directly proportional to 7' or whether 



1) We are dealing here exclusively with the constant-volume Ihermomfter. A 

 subsequent Gommunicalion will deal with the use of thermometers at constant 

 pressure. 



