90 Mr. W. Sutherland on two New 



Crookes's, and would therefore enable our formula to give 

 the coefficient of slipping £ with great accuracy, and con- 

 versely the pressure. 



The simplest form which the apparatus could take would 

 he that in which the two fixed plates form the top and bottom 

 of a shallow cylindrical drum strong enough to stand a 

 pressure of an atmo on its outer wall without appreciable 

 deformation ; within the drum the movable disk is suspended 

 equidistant from top and bottom by a quartz fibre passing- 

 down the axis of a fine vertical tube rising from the centre of 

 the top of the drum; a small tube in the centre of the bottom 

 would furnish connexion with the gas supply whose pressure 

 is to be measured, a side tube from the upper vertical tube 

 facilitating the operation of filling the drum with the pure 

 gas. A small mirror attached to the quartz fibre at a suitable 

 point near its entrance to the drum would give the means of 

 measuring the angular amplitude of successive vibrations of 

 the disk, which could be set in motion by means of a small 

 magnet attached to the back of the mirror. 



The whole apparatus ought to be made so that it could be 

 raised to as high a temperature as possible to secure thorough 

 drying and exhaustion. The special advantage of this form 

 of viscometer gauge is that its volume is small, only the 

 unoccupied space of the drum, say 2ir x 2'5 2 x *1 cub. centim., 

 together with the internal volume of the tube, which is, 

 say, 50 centim. long and 1 millim. radius, namely it X *1 2 x 50 

 cub. centim., that is a total volume of about 6 cub. centim. 



A more accurately constructable form of the viscometer 

 gauge would be one in which the two fixed disks and the 

 movable one with its suspension fibre passing through a 

 hole in the upper are all enclosed as in Maxwell's and in 

 Kundt and Warburg's apparatus in a suitable vessel for con- 

 nexion with the gas supply whose pressure is to be measured. 

 Now in these forms of viscometer gauge the expression 

 given by Maxwell for the relation between log. dec. and 

 viscosity simplifies at, say, 1/20 atmo to one which makes the 

 log. dec. minus the unknown log. dec. of an absolute vacuum 

 due to the viscosity of the fibre, proportional to the viscosity 

 of the gas. The part of the log. dec. due to the fibre could 

 be found by loading the fibre with a sphere of lead equal in 

 weight to the disk of the viscometer gauge and observing the 

 log. dec. in air at any ordinary pressure, the small effect of 

 the viscosity "of the air being allowed for by calculation, or 

 the sphere might be torsionally vibrated in a bulb exhausted 

 till the highest obtainable vacuum was reached, when the 

 log. dec. ought to become constant or tend to a constant limit 



