X-RAYS, NATURE OF: Röntgen Rays), in the region noted, and Hol weck's experiments have shown in addition a continuous, or non characteristic, radiation from 25 volts to 28o volts (X =5oo A.U.
to X=44 A.U.). Typical characteristic lines, for which concordant results have been obtained by several different workers, occur, for the L series of magnesium, aluminium, chlorine, argon, at 46, 68, 200 and 245 volts respectively, corresponding to wave-lengths 269. 182, 62 and so A.U. For the K series of carbon, nitrogen, oxygen, fluorine, neon, Holweck has obtained, by his ionisation method, lines at 280, 397, 528, 684 and 864 volts, corresponding to wave lengths 44.1, 31.1, 23.4, 18.1, 14.3 A.U. respectively. It should be noted that the first three of these lines are in good agreement with the measurements of Mohler and Foote, while Holweck by his has further obtained an aluminium line at 1,555 volts (X = 7.95 A.U.) which agrees precisely with the K absorption edge determined by Fricke by the method of crystal diffraction. This agreement is important, as it gives additional assurance that the results of the ionisation potential methods have been correctly interpreted.
By these measurements of critical potential wave-lengths have, then, been measured from 269 down to 14.3 A.U. which comfort ably bridges the gap between wave-lengths measured by the help of ruled gratings, and those measured by crystal diffraction. The methods of crystal diffraction have been used with X-rays to measure wave-lengths down to .107 A.U., the K absorption edge of uranium, while the shortest 7-ray measured by Rutherford and Andrade, with the crystal method, was .072 A.U. C. D. Ellis and L. Meitner have independently measured much shorter 7-ray wave-lengths by the help of the natural and secondary /3-ray spectra, a method which is essentially photoelectric. (See PHOTO ELECTRICITY.) The measurements of Ellis stretch from wave length .26 A.U. to .005 A.U., and thus have an ample overlap with the lengths measured by crystal diffraction. This .005 A.U. repre sents the shortest wave-length of which the measurement has much claim to accuracy.
The estimates of the wave-lengths of the very penetrating cosmic radiations investigated by Kohlhorster and later by Milli kan depend upon the extrapolation of the absorption law far beyond the limits of our experimental knowledge.
Millikan's latest measurements (Physical Review, xxxii., 533, 1928) show that the cosmic radiation is not homogeneous, but con sists of two main bands, one having an absorption coefficient /2 = .35 per metre of water Cup =.0035 per cm.), and the other being resolvable into two radiations for which ,u = .08 and ,u = .04 per metre of water respectively. This rough resolution into a dis continuous spectrum invalidates the method of generation of the rays formerly suggested, namely that the energy of the radiation might be supplied by the impact of high speed electrons against the nuclei of atoms, for this would be expected to produce a gen eral radiation. The fundamental assumptions underlying any explanation so far imagined are: (I) that a frequency v can be derived from the absorption coefficient of the cosmic radiation by the use of some formula to extrapolate the frequency-absorption relation into regions of low absorptionfor which no laboratory measurements are available; (2) that the quantum of energy hv must be released by some single atomic process. The problem is to
find an atomic process which will liberate so large an amount of energy as the very high frequencies derived by (I) require.
The Einstein relation between mass and energy (see RELATIVITY and also NUCLEUS, section Stability of Nucleus: Atomic Energy) associates any loss of mass with an emission of energy. There is evidence for a change of mass when protons and electrons are packed together to form a nucleus (see NUCLEUS) and Aston has, from his experiments, obtained a curve showing the "packing effect" for all elements. (See ISOTOPES.) From this curve it is clear that for the lighter elements, of atomic weight less than 8o or so, energy can only be given out when the nuclei are built up from protons, and not when they break down. The heavier ele ments would, it is true, give out energy on disintegrating, for they lie beyond the minimum in Aston's curve, and the radioactive ele ments provide actual examples of atoms disintegrating with emis sion of energy in the form of a- and 3-rays (with 7-rays as a secondary phenomenon). However, the energy given out per i atom by radioactive disintegration not to account for the cosmic radiation. Using a formula of Dirac to obtain a frequency from the absorption coefficient Millikan has argued that the cosmic radiation for which p. =•35 per metre oi water must originate in the formation of helium from hydrogen in outer space. It is shown in NUCLEUS that there is a loss of energy 4.6X ergs (4.3X ergs with the latest figures, used by Millikan) per atom of helium formed, which from the quantum relation E= by corresponds to a wave-length of .00046 A.U. Dirac's formula gives an absorption coefficient .3o per metre of water for such a fre quency, which agrees well enough with the necessarily very rough figure derived from experiment. Reasoning on similar lines, the formation of oxygen from hydrogen in a single act gives rise to a radiation of wave-length .000i A.U., which has an absorption coefficient of .o8 per metre of water : this agrees exactly with that of one of the other cosmic radiations observed by Millikan. The formation of silicon from hydrogen leads to a radiation for which =.041 per metre of water, agreeing with the second of the more penetrating cosmic radiations of Millikan. Helium, oxygen and silicon are elements which appear in great abundance everywhere in the heavens, which lends plausibility to this theory of the gen eration of cosmic rays, although the creation of an element like silicon in one act out of 28 atoms of hydrogen is a process which offers grave difficulties to the physical imagination.