72 THE ROYAL SOCIETY OF CANADA 



colloidal potassium enclosed in a glass bulb containing helium at low 

 pressure. The potassium surface was kept at a negative potential 

 of about 13 volts and the central electrode was connected to the electro- 

 meter. The cell was enclosed in a light-tight box and the metallic 

 surface illuminated directly, through a narrow slit, by the light from 

 the glowing wire. The slit was largely protected from illumination 

 by outside light and in all cases, the net effect, due to light from the 

 glow-lamp alone, was used in obtaining the results given below. 



III. Preliminary Measurements. 



1. The cell was first illuminated by light from a Nernst lamp 

 sent through a monochromatic illuminator, and the photo-electric 

 current measured for successive portions of the spectrum. The re- 

 sults obtained are shown by the curve in Fig. 2. It will be seen that 

 the light effective photo-electrically with the glass cell used lay be- 

 tween the wave-lengths 2500° A.U. and 7500° A.U., with a maxi- 

 mum due to light intensity and cell sensitiveness, acting in con- 

 junction, at about 4600° A.U. 



2. In order to gain familiarity with the apparatus, an ordinary 

 32 c.p. incandescent lamp was attached by a glass tube, sealed into 

 the top, to a mercury pump. The bulb was next exhausted, and 

 hydrogen from a Kipp apparatus admitted through the pump; the 

 process was then repeated and the readings finally taken. The lamp 

 was attached to the 110 volt mains and the times of deflection over 

 a fixed portion of the reading scale were taken as the pressure was 

 varied in the bulb. The result of the readings is given by the curve 

 shown in Fig. 3. Detailed readings for a typical set of observations 

 are given later. 



Approximate results were also obtained for nitrogen in a similar 

 bulb. It was, however, impossible to work satisfactorily with the 

 ordinary incandescent filament; hence, for the readings with nitrogen 

 and air, two glow-lamps were constructed by fusing similar pieces 

 of platinum wire axially in bulbs of different dimensions. These bulbs 

 were of the shape indicated in Fig. 4, and their greatest diameters 

 were 2 and 7 cm. respectively. The length of the wire was al)out 

 3-5 cm. 



This type of lamp was used throughout the remaining experi- 

 ments. A potential difference of 3 • 5 volts was constantly maintained 

 between the ends of the filament, and the lamp, attached to the mer- 

 cury pump, was so placed in front of the narrow slit that the light 

 from the glowing filament fell upon the sensitive surface of the photo- 

 electric cell. Measurements were then taken of the photo-electric 



