Measurement of Electromagnetic Radiation. 57 



of air passes from one side of the tube to the other, it has to 

 pass through the narrow space between these cylinders. 



A hole was bored through the top of B, through which a 

 capillary glass tube could be passed. This tube was filled 

 with phosphorus, and the glass broken away for about one 

 millimetre, so as to expose a clean piece of phosphorus when- 

 ever the stream of oxide had to be observed. The tube was 

 then passed through the hole (K), so that the exposed piece of 

 phosphorus came just beneath the partition and between the 

 glass cylinders. It is shown in place at H in fig. 5. 



By this means the stream of phosphorous oxide was formed 

 at the narrowest part of the air channel, and so was affected 

 to the maximum extent by any passage of air from one side 

 of the partition to the other. The resulting deflexion of the 

 phosphorous-oxide stream could be read with an accuracy of 

 about one-tenth millimetre. In this form, when the phosphorus 

 index is deflected it is carried to a part where the channel is 

 much wider, owing to the curvature of the glass half-cylinders. 

 On this account the deflexion of the index is not any simple 

 function of the rate at which the air is passing. With 

 a view to do away with this want of proportionality, and 

 also, if possible, to render the instrument more sensitive, the 

 glass half-cylinders were replaced by a pair of narrow flat 

 glass plates, near together, with a wide channel on each side 

 communicating with the tube above. The instrument was, 

 however, found to be less sensitive in this form than when the 

 half-cylinders were used. 



The sensitiveness of the instrument could, to a certain 

 extent, be altered by moving M and N (fig. 6) further apart, 

 so that the phosphorus stream fell in a wider channel. 



The divided tube was surrounded by another glass tube 

 (Gr, fig. 5), which was about 10 centimetres longer. This 

 formed the guard-tube, and was suspended from one of the 

 axles of a clock (K, fig. 11), which made one revolution in 

 about eight seconds. 



By means of three levelling-screws the guard-tube w r as 

 made to hang free of the inside tube, so that it did not touch 

 it at any part of its revolution. Thus no heat due to friction 

 was developed near the divided tube. 



The phosphorus could not be placed at the top of the tube, 

 as the falling column of oxide was found to go on one side of 

 the partition, and its weight was sufficient to keep up a down- 

 current of air on this side and an up -current on the other side, 

 and so make the instrument unstable. 



Inside the divided tube, and on one side of the partition, 

 two fine platinum wires were stretched 3 reaching from the top 



