the Photoelectric Effect. 511 



glass joint A. The neck of this bulb is closed by a quartz 

 window Q. It is waxed firmly in a brass collar which 

 screws into the telescope tube of the monochromatic illu- 

 minator at T, in such a position that the slit L is just 

 focussed on the strip S so that the image of the slit for any 

 particular wave-length of light may be made to coincide 

 with the position of the strip. The final adjustment is made 

 by using visible lio-ht as the source and looking through M. 

 The strip S is shifted a little until it just shuts out the image. 

 By turning the screw-head \ any desired wave-length may 

 be thrown on $ and its magnitude read off on the graduated 

 scale. In the figure the illuminator is shown rotated through 

 a right angle from its true position, about the axis of the 

 telescope. The collimator should project out perpendicular 

 to the plane of the paper. A quartz mercury lamp at L was 

 used as the source of light. The apparatus was evacuated 

 by a Gaede pump connected at N, and usually the tube B of 

 coconut charcoal was used to obtain a still better vacuum. 



The velocities of the liberated electrons were determined 

 by varying the field between S and D by adjusting tlie 

 sliding contact R, and the applied field was read directly 

 from the voltmeter V. The distribution of velocity curves 

 were obtained in the ordinary way. 



The peculiarities of this apparatus are that the field between 

 S and D is approximately radial, and that the area of the 

 emitting electrode is very small compared with the area of 

 the receiving electrode. In the first place, this arrangement 

 makes it possible to measure the distributions of velocity 

 of all the electrons irrespective of their angles of emission 

 from S, and thus it gives the distribution of total energy of 

 the electrons. In the second place, it makes the effect of 

 electron reflexion practically negligible. A certain per- 

 centage of the electrons reaching D are reflected, this reflexion 

 being diffuse. But none of these reflected electrons will 

 return to S, except a very small proportion which start 

 almost straight towards S. All the others will traverse 

 comet- like paths past S and again strike D. 



The metal strip to be tested was carefully scraped with a 

 clean knife-blade. It was then placed quickly in position 

 and the vacuum-pump started. The readings were taken as 

 soon as the requisite vacuum was obtained. 



The Distribution of Velocities. 

 The general character of the distribution of velocity curves 

 is the same for all the metals, so that it will be sufficient to 

 consider in detail the case of platinum only. With this 



