266 BULL SYSTEM TECHNICAL JOURNAL 



One recent improvement in the apparatus is a change in the nature of 

 the scanning Hght, whereby, without sacrificing the general principle of 

 using visually inefficient hut photoelectrically efficient radiation, the 

 proper balance of tone values in the face is restored. This has been 

 accomplished by adding to the battery of blue sensitive potassium cells, 

 a group of red sensitive caesium oxide cells, and scanning by purple 

 instead of blue light, that is, both ends of the visible spectrum are 

 used in place of one end. 



In making this change, a number of others were involved, most of 

 which resulted in simplification or improvement. One important 

 alteration was the substitution for the arc lamps previously employed, 

 of incandescent lamps of a type available from motion picture projec- 

 tion practice, as shown in Fig. 1. The lamp employed has for its radia- 

 tor, four vertical helical coils of tungsten wire, and is furnished with a 

 reflector which images the coils back on the intervening spaces. An 

 efficient condenser system throws a brilliant rectangular image on the 

 back of the scanning disc, which is substantially uniform over the 

 whole field. With this unit, the scanning beam as it leaves the pro- 

 jection lens is somewhat larger in diameter than the beam as produced 

 from the arc. Consequently, for positions away from the focused 

 image of the disc holes, the scanning beam is larger than before, with 

 some resultant loss in the range of sharpest definition. Since, however, 

 the user of the two-way apparatus is seated in a fixed chair, he has little 

 opportunity to move far out of the plane in which the disc holes are 

 focused, so that this objection is not serious. The advantages of this 

 substitution were two-fold. First was a great gain in simplicity of 

 operation and maintenance. Second, the incandescent lamp, being a 

 lower temperature radiator, radiates relatively many times as much 

 red light as does the arc, for the same amount of blue. Consequently, 

 once an incandescent lamp unit was found which gave the amount of 

 blue light required for the potassium cells, the great e.xcess of red light 

 made possible the use of relatively few Ccesium oxide cells. Since 

 these are intrinsically somewhat more sensitive than the potassium cells, 

 the net result was that a red signal comparable with the blue signal 

 could be added by the installation of only two caesium cells, each of less 

 than half the electrode area of the potassium cells. 



It was found most convenient to mount the two Cccsium oxide cells 

 directly in front of the observer, to either side of the microphone, and 

 above the opening in the booth through which the scanning beam 

 enters, and through which the incoming image is seen. This arrange- 

 ment is shown in Fig. 2. The only objection to placing the cells in this 

 position is that they encroach somewhat into the region where reflec- 



