90 
PH YSICS: CLA RK AND D UA NE 
Proc. N. a. S. 
terminations in the second order we obtain for this ratio 1.97 as compared 
with 2.00 predicted by Bohr. 
The ratio of the intensities of j8 to 7 in the first order is 6.3 (one deter- 
mination) and in the second order is 5.46 (two determinations). 
^ Duane and Patterson, Proc. Nat. Acad. Sci., Washington, Sept. 1920; Phys. Rev., 
Dec. 1920, p. 526. 
2 Coster, Z. Phys., 5, 1921 (139); also 6, 1921 (185). 
3 De Broglie and Dauvillier, Paris, C. R., 173, p. 137. Dauvillier, Ibid., 172, also 
173, p. 647, also other papers in this volume. 
4 Smekal, Z. Phys., 5, 1921 (91, 121). Also 4, 1921. 
5 Wentzel, Ihid., 6, 1921 (2). 
6 Coster, Ibid., 5, 1921 (147). Also Physic i^ez;., Sept. 1921. 
7 Duane and Patterson, Physic Rev., 16, 1920 (534). 
8 Coster, Ibid., Jan. 1922. 
9 Hoyt, Proc. Nat. Acad. Sci., Washington, Nov. 1920. 
10 Overn, Physic. Rev., Nov. 1921, p. 350. 
A NEW METHOD OF USING X-RAYS IN CRYSTAL ANALYSIS 
By George L. Clark and Wiluam Duane 
Jefferson Physical Laboratory, Harvard University 
Communicated April 4, 1922 
The procedure described in this paper differs in some fundamental de- 
tails from the elegant methods of analyzing crystals by means of X-rays 
developed by Laue, the Braggs and others. 
The substance to be examined may be in the form of powder or of whole 
crystals. In the latter case the crystal need not be large. One having a 
volume of only .25 cubic mm. suffices, although slightly more accurate 
data may be obtained, if it has a somewhat larger volume. 
In our method we employ X-rays belonging to the continuous spectrum, 
and therefore containing rays of a great variety of wave-lengths, and not 
the line spectra of definite wave-lengths, as in most other methods. This 
procedure has certain advantages. An ordinary X-ray tube with a tung- 
sten target may be used, and high voltages may be applied to it. More 
intense X-radiation can be obtained from tungsten than from such chem- 
ical elements as rhodium and molybdenum. Further in some cases it is 
desirable to employ X-rays of shorter wave-lengths than those in the K 
series of rhodium and molybdenum, for these shorter rays are more pene- 
trating than the longer ones. As a rule we use the spectrum between .12 
and .80 Angstrom. 
In making an analysis of a small crystal we place it at the center of an 
X-ray spectrometer, provided with an ionization chamber and a quadrant 
electrometer. The slits of the instrument should be fairly narrow, and 
