HOGG. — FRICTION AND FORCE DUE TO TRANSPIRATION. 127 



larity the upper plate and the disk were lifted together, and the wire 

 carefully bent at a point close to the disk. After some trials it was 

 found possible to adjust, so that the maximum difference in height of 

 the two ends of a diameter of the disk was less than 0.07 mm.^^ The 

 width of the disk was 44 mm. The large plates were then fastened 

 together as described above, and the three clamps, one of which is 

 shown at N, were placed so as to fit the supporting wires, one of which 

 is seen at 0. 



Suspending cords were now placed temporarily around the glass 

 separating tubes and the plates raised by them, in an inverted position, 

 allowing the armature and its belongings to be suspended by the fibre. 

 The whole was then carefully lowered into the enclosing vessel, which 

 had been placed temporarily bottom upward. When the plates were 

 in position on the supporting wires, the various wires of the suspended 

 part were put in place by means of a tool made for the purpose, the 

 various nuts turned on, and the whole turned into its natural posi- 

 tion. The disk was raised by turning the armature, J, and the plates 

 then made parallel to it. This adjustment was made by turning the 

 supporting nuts on the wires 0, 0, 0. Finally, the check-nut, K, 

 being loose, the upper clamp was turned in the cylindrical nut, H, 

 until the mirror faced the window, D. The check-nut was then tight- 

 ened, and the suspended part lowered so as to rest on the tripod at C. 

 The instrument could now be handled with only moderate care. The 

 upper and lower ends were then sealed off, and it remained only to join 

 this to the other parts of the apparatus. 



The Transpiration Problem. 



The first experiments on Thermal Transpiration were made by Fed- 

 dersen,i7 but the full investigation of the phenomenon is due to Pro- 

 fessor 0. Reynolds. 1^ He showed that, if the two surfaces of a plate 

 of porous material which divides a mass of gas into two separate por- 

 tions are kept at different temperatures, the gas will force itself through 

 the channels in the material from the cold surface to the hot until a 

 certain difference of pressure is reached. In the case of air, he found 

 it possible to establish, in this way, as much as 6 mm. of mercury differ- 



" Stokes has shown that a small inaccuracy in this adjustment involves a 

 rather large error in the measurement of the resistance encountered by the 

 moving disk. If the layer of air between the latter and the tixed disk is wedge- 

 shaped, considerable energy is used up in crowding the gas between the fixed and 

 moving disks. The adjustment attained here is sufficiently accurate to avoid this 

 difficulty. 



" Pogg, 148, 302 (1873). " Phil. Trans., 170, 1880. 



