472 BELL SYSTEM TECHNICAL JOURNAL 



No means were provided for measuring the current of electrons 

 incident upon the first crystal. We had found, however, from a 

 preliminary investigation of the characteristics of the gun, that currents 

 of the order 2 X 10~^ amperes could be obtained from it. It was known 

 also from these tests that the electrons ejected from the gun are very 

 nearly homogeneous in speed. Given this value for the current in the 

 primary beam, it was possible from our previous observations on the 

 regular reflection of electrons at 45 degrees incidence to estimate the 

 order of magnitude of the current of full speed electrons which might 

 reach the collector after two such reflections. The estimated magni- 

 tudes were from lO^^^ to 10"" amp, and the currents of selectively 

 reflected electrons actually observed have had values within this range. 



In measuring these small currents we have had the use of a direct 

 current vacuum tube amplifier designed and built by Dr. J. M. Eglin. 

 It is the type of amplifier described recently by Wynn-Williams,^ 

 but embodies certain improvements described by Dr. Eglin at a 

 recent meeting of the American Physical Society.^ Conditions for 

 observing were best when the amplification factor was about 

 2,000, so that the currents actually measured were of the order of 

 10~^ amp. 



A few preliminary observations were made before heating the crystals 

 to free their surfaces from adsorbed gas. The relation between the 

 current entering the collector and the bombarding potential for a 

 fixed angle 6 was quite difi'erent in these first tests from that observed 

 after the crystals had been heated. The principal feature of this 

 initial current-voltage relation is a strong maximum at 20 volts. Tests 

 were made for polarization with the crystals in this condition but no 

 evidence of such a phenomenon was obtained. 



The current-voltage curve characteristic of reflection from the 

 crystals in a thoroughly cleaned condition is shown as Curve A in Fig. 4. 

 The data from which this curve has been plotted were obtained with 

 the faces of the reflectors parallel to one another as illustrated in Figs. 

 1 and 2. It will be convenient to designate this position of the movable 

 system as the position 6 = 0°. The "parallel" positions are then the 

 positions 0=0° and 6 = 180° and the "transverse" positions are those 

 for which 6 = 90° and 6 = 270°. A curve similar to Curve A is ob- 

 tained whatever value is chosen for 6. In this and in all other tests the 

 inner box of the collector was maintained at a potential 2 volts above 

 that of the midpoint of the filament. 



Curve B of Fig. 4 exhibits, on a different scale of ordinates, the rela- 

 tion between current and voltage observed for angle of incidence 45 



* C. E. Wynn-Williams, Proc. Camb. Phil. Soc, 23, 811 (1927). 

 ^ J. M. Eglin, Phys. Rev., Z3, 113 (1929). 



