THE statement of Mr. S. C. Bradford in the second paragraph of his letter to NATURE of April 8, that I suppose the electrons to revolve in small circles without any constraining force, is erroneous. The fact that I reserved an opinion as to the nature of the constraining force does not imply, as he suggests, that I deny its existence. Thus (cf. Science Progress, April, 1920, and Phil. Trans. Roy. Soc, vol. ccxx., p. 247, 1920) an electron moving with speed v perpendicular to a magnetic field of intensity H (which may originate in the nucleus) describes a circular orbit of radius ρ = mI½/Hɛ, and the frequency of the electron is v = Hɛ/2πm, which (and this is an advantage in the case of a radiating orbit) is independent of the speed with which the electron describes the orbit. At present we know little about the actual value of v. If H is of the order 107 gauss, the value ascribed to the molecular, field from magnetic considerations, the frequency is that of infra-red radiation, and the correlation of the elastic, properties of the medium (which are determined by this molecular field) with the infra-red vibrations, as originally pointed out by Debye is apparent. Within an atom the controlling field may be of the order 108 gauss, which gives rise to vibrations of optical frequency. Closer to the nucleus a field of 109 gauss gives rise to frequencies comparable, with those of a K series.