M. E. Pringsheim on the Radiometer. 103 



furnished with glass cocks, proceeded, by which the apparatus 

 could be connected with the air-pump and a gasometer. To 

 produce a vacuum in the apparatus a Geissler mercury air- 

 pump was employed ; and the exhaustion was always carried 

 on till the volume of air expelled by the rise of the mercury- 

 level ceased to diminish, in spite of continued pumping. This 

 pump being employed, the finally remaining pressure could 

 not be determined, because the pump possessed only an ordinary 

 mercury manometer ; in this, however, on exhausting, the 

 mercury stood somewhat higher in the open than in the 

 closed leg ; so that the rarefaction was at all events consider- 

 able, and, as may be inferred from the sensitiveness of the 

 apparatus, not far from the pressure at which the mobility 

 reaches its maximum. Moreover, an exact knowledge of the 

 amount of pressure is immaterial in our experiments. 



For the experiments the instrument was fixed, as perpendi- 

 cular as possible, on a tripod, so that the vane could rotate 

 with perfect freedom. The reading-off of its position was 

 secured by the following arrangement: — The image of a fine 

 slit illuminated by a petroleum lamp was concentrated upon 

 the mirror by a lens, and reflected thence to a scale divided 

 into millimetres at about 1100 millim. distance. In order, 

 however, to be quite safe from any disturbing action of this 

 light, the latter should first be caused to pass through a layer 

 of water and solution of alum, and thus deprived of the rays 

 which are effective upon the apparatus. 



To protect the whole from air currents it might be put into 

 a tin case stuffed with wadding, having only two apertures to 

 permit the luminous index and the rays of the source of heat 

 to fall upon the mirror and the disk respectively. 



This apparatus permitted a relatively convenient and quick 

 changing of the disks, although this is a sufficiently delicate 

 manipulation; and in lowering in and taking out the apparatus 

 more than one glass thread broke, since the disk with the mirror 

 could only slide through the narrow glass tube lengthways, so 

 that the thread had to bend sharp round directly on the 

 mirror. This inconvenience could have been remedied only 

 by inserting the apparatus in a proportionally wide cylindrical 

 tube ; but then the space to be exhausted would have been 

 too large, or the thread must have been short, and hence the 

 sensitiveness too little ; moreover the desired similarity of the 

 glass case in the vicinity of the vane to the globe of the 

 radiometer would have been lost. 



In most of the experiments the cock of one of the side tubes 

 was closed, while the other tube was connected with the air- 

 purap by very short india-rubber connexions and a long 



