As a result of recent experiments it can be said, then, that we have knowledge of an unbroken range of electromagnetic radia tions from wave-lengths of hundreds of metres down to wave lengths of a few hundred-thousand-millionths of a centimetre, and have evidence of cosmic radiations of still shorter wave-length.
Generation of the Radiations.—When we consider electro magnetic radiations in free space, any distinction into groups of different wave-length is purely arbitrary, for the velocity and mechanism of propagation is exactly the same throughout the spectrum. When, however, we turn to the relationship between radiation and matter, and in particular the emission and absorp tion of radiation, there are certain distinctions which, while giving no sharp boundaries of wave-length between different classes of radiation, do nevertheless divide the radiations into broad natural groups. These distinctions are based upon the nature of the vibra tor, or of the quantum mechanism, which gives rise to the undula tions.
Starting at the long wave-length end of the spectrum, the Hertzian waves are generated by oscillations of electricity in macroscopic systems of conductors. The atomic or molecular nature, or even crystal structure of the metals constituting these conductors does not enter directly into the calculations from which the frequency of a given system can be derived : the conductors can be treated as uniform and homogeneous, and the problem of the oscillations as strictly analogous to those of mechanical oscilla tions of systems of springs and masses. The size and geometry of the system determines the wave-length; the engineering stations of the wireless telegraph companies, with their mighty aerials, produce wave-lengths measured in hundreds of metres, while minute oscillators whose size is measured in fractions of milli metres produce wave-lengths of the same order as their own linear dimensions.
The infra-red radiations are produced by a variety of oscillators, if we merely consider an arbitrary group of wave-lengths from .8,u to 40012 say. The atomic mechanism of quantum jumps exe cuted by an outer electron, which is responsible for line spectra (see ATOM, QUANTUM THEORY, SPECTROSCOPY), leads to certain lines situated well in the infra-red: for instance Brackett (1922) has measured two hydrogen lines, of wave-length 4.05,u and 2.63/2, which can be represented by putting n'=4, n =5, 6 in the general Balmer formula v= nL — Such lines, how ever, are not representative of infra-red radiations, but should be classed with the optical spectra. More typical are the band
spectra consisting of nearly equidistant lines in the far infra red (in the neighbourhood of ioo,u), and the band-spectra con sisting of two branches of equidistant lines, with a gap between them, observed in the near infra-red (in the neighbourhood of 9.4). (See BAND SPECTRUM.) These spectra are usually observed as absorption spectra, but emission and absorption are effected by the same mechanism. It has been established beyond doubt that the bands in the far infra-red are due to molecules rotating, the possible rotations being fixed by quantum conditions, and inter changes between stationary states leading to emission or absorp tion, while the bands in the near infra-red are due to vibrations within the molecule, the distance between the nuclei of the con stituent atoms varying periodically. These vibrations are super posed on the rotations, and likewise governed by quantum con ditions. The details of their mechanism are described in the article BAND SPECTRUM; what it is desired to emphasise here is that the periodic changes in question, whose frequency is typical of the infra-red, are molecular in nature, the motion being of the molecule as a whole, often supplemented by motions of the atoms, treated as rigid wholes, within the molecule. As soon as changes within the atom itself are added to these motions the corresponding bands are no longer in the infra-red, but in the visible and ultra-violet.
Another case of vibrations of the infra-red class is offered by the rest rays, to which reference was made in the preceding section. The oscillations here in question are those of ionised atoms making up the structure of the crystal in question, and can be calculated, as shown by Born and his collaborators, by considering the possible vibrations of the crystal lattice. Here again we are dealing with vibrations of atoms considered as rigid wholes, governed by interatomic forces. It may be said, then, that frequencies of the orders usually associated with the infra red are connected with the vibrations of atoms or molecules as a whole, under the influence of their mutual attractions and repulsions, and not with changes within the atom.