Vol. 8, 1922 
PHYSICS: CLARK AND DUANE 
95 
tween successive planes in the crystal that are alike. The curve in figure 
3 represents the ionization current as a function of the voltage in an 
experiment with potassium iodide powder. In addition to the point 
where the ionization current vanishes there are several well-marked 
places where the curvature changes abruptly. The voltages at these 
breaks in the curve are the voltages at which the rays reflected from other 
crystal planes cease to be produced. These voltages together with the 
angle 6, substituted in equation (3), give the corresponding distances be- 
tween the planes. The three most sharply determined voltages are 
21,740, 31,010 and 37,600 volts, respectively. The ratios of these numbers 
to each other differ from V2 and V3 by small fractions of one per cent 
only. They correspond, therefore, to the 100, 110 and 111 planes, re- 
spectively, of a cubic lattice. The angle d equals 4°. 51 in this experiment, 
and the average value of d, calculated from equation (3), is 3.534, in close 
agreement with that obtained in the above-described measurement with 
a small crystal of potassium iodide. 
The curve marked X in figure 2 corresponds to the peak marked X in 
figure 1. There appears to be a sharply indicated point of inflection in 
the former at a voltage of 30,000 volts. Substituting this in equation 
(2') we find that the corresponding wave-length is .412. This wave- 
length is a good mean value of the wave-lengths of a and /3 lines in the K 
series of iodine, namely, .437 and .388. 30,000 volts do not suffice to 
produce the K series of iodine. If, however, one sets a spectrometer on 
a characteristic line and reduces the voltage, the ionization current does 
not vanish when the voltage ceases to produce the characteristic X-rays. 
It vanishes only when the voltage becomes insufficient to produce an X- 
ray in the continuous spectrum of equal wave-length to that of the char- 
acteristic line. We have tested the existence of the peak X by turning 
the crystal as in the experiments represented in figure 1 with voltages 
in the neighborhood of 33,030 volts, which is the critical voltage for the 
production of the K series of iodine. We find no evidence of a peak at a 
voltage slightly below 33,000. The peak is just observable, however, 
if the voltage equals 33,500 volts. These points strongly favor the idea 
that the characteristic K radiation of iodine produces the peak X. The 
distance between the planes making an angle of 17°. 84 with the 100 planes, 
however, does not equal that calculated from equations (1) and (3). The 
former is 1.088 and the latter is 2.144, i.e., about twice the former. Per- 
haps this may be due to the fact that only half the atoms are iodine. 
If the characteristic radiation does produce the peak, and, if we adopt 
the interference of scattered rays as the explanation of X-ray reflection, 
we must assume that, on the average at least, definite time relations must 
exist between the passage of the primary X-ray across an atom and the 
emission of a secondary characteristic ray. If we adopt the quantum point 
