Mr. G. J. Stoney on Crookes's Radiometer. 307 



this layer the air throughout the greater part of the chamber 

 will be very slightly denser and warmer than before the ad- 

 justment. There will also be another layer of varying tem- 

 perature and density in contact with the glass, having on 

 one side the temperature of the air in the little chamber, 

 and on the other the somewhat lower temperature of the 

 glass. This is the state which will come into existence during 

 the adjustment, and which will then continue as the permanent 

 state of affairs so long as the disk and glass are maintained at 

 their respective temperatures. While this is the condition of 

 the apparatus, there would be no transfer of heat from the disk 

 to the glass if convection-currents were absent : the contrivance 

 cannot act as a heat-engine ; and there is accordingly no pres- 

 sure of the kind which Mr. Crookes has detected, and to which 

 I would suggest that we should give the name of Crookes's 

 pressure. 



13. On the other hand, if the exhaustion be carried further, 

 the layers of varying temperature in contact with the disk and 

 with the glass will become thicker; and if the exhaustion pro- 

 ceed sufficiently far, these layers will be so dilated that there 

 will not be room between the disk and the glass for their full 

 development ; and then will arise the state of things in which 

 Crookes's pressure presents itself. The following description 

 will give an idea of what is now going on within the cham- 

 ber. As soon as the brief period of adjustment is over a pro- 

 cession of swift molecules at once commences, which is ever 

 starting from the front or warmed side of the disk (the heater 

 of the little engine), and jostling its way to the opposite hemi- 

 sphere of cool glass* (the cooler of the heat-engine). Another 

 procession is as constantly starting from the cool hemisphere 

 of glass and working its way both up to the disk and to the 

 regions behind the disk. These processions last so long as the 

 heater and cooler are maintained at their respective tempera- 

 tures. Any member of the slow procession that comes into con- 

 tact with the front of the disk is, in rebounding, converted into 

 a member of the swift procession ; and any member of the 

 swift procession that comes into contact with the glass is con* 

 verted into a member of the slow procession. 



14. To complete the picture we must suppose the members 

 of both processions to be mixed up with a crowd of molecules 

 moving about irregularly, and not belonging to either proces- 

 sion. The processionists are so buffeted that probably no 

 one follows his own procession far. But some of the mole- 

 cules against which he jostles generally carry a part of his 



* If the instrument has four arms, the back of one of the other disks at 

 right angles is to be regarded as part of the cooler of the little engine. 



Y2 



