While the spectrum was first extended by investigation of the prolongation of the optical spectrum on either side, and thus showed a continuous range of wave-lengths, the Hertzian waves as first investigated had wave-lengths very much greater than those of any measured infra-red radiation, while the X-rays first measured, although forming a continuous range with the 7-rays on the short wave-length side, proved, on the long wave-length side, to be very much shorter than the shortest ultra-violet. The gaps between the Hertzian waves and the infra-red, and between the X-rays and the ultra-violet were until recently unexplored regions, in spite of the many efforts that had been made to pro duce and detect waves there. Since 1922, however, the gaps have been closed, and it is now possible to carry out experiments with waves at any part of the sixty odd octaves of the electromagnetic spectrum.
The gap between the infra-red and the Hertzian waves has been bridged by advances from both sides. The production of shorter and shorter Hertzian waves has been effected by the use of smaller and smaller oscillators. Lebedew (1895) and Lampa (1896) respectively obtained waves which they estimated as of wave-length 6 mm. and 4 mm., but Mobius, who published his results in 192o, considers that these waves must have been of length I cm. and 7 mm. respectively. He himself obtained waves 7 mm. long, with which appeared overtones of much shorter wave-length, but such measurements could not be reproduced exactly. The quest for short Hertzian waves was carried much further by Nichols and Tear, who published their final results in 1923. Lebedew and Lampa had constructed oscillators which consisted of minute platinum cylinders, 1.3 mm. long and .5 mm.
thick, arranged axially end to end, with a gap between them. The current from an induction coil had to leap across small spark gaps to reach the oscillator. Nichols and Tear worked on similar lines, but with great experimental skill succeeded in reducing the dimensions of the oscillator still further. The two . cylinders which formed the oscillator proper were of tungsten, each being .2 mm. long and .2 mm. thick, with an oil-immersed gap of .or or .02 mm. between them. (It appears from a note published by K. K. Darrow in the Bell System Technical Journal, Vol. 3, Number 3, 1924, that the dimensions of the cylinders were subsequently reduced to mm. long and •1 mm. diameter.) The waves proceeding from the gap were focussed into a parallel beam by a lens of paraffin wax, and then fell on the interferometer used for measuring the wave-length, which was of the simple form due to Boltzmann. This consists of two plane mirrors mounted one above the other, one stationary and the other movable in such a way that the two always remain parallel. Half the beam falls on each mirror. The reflected beam was brought to a focus by a second paraffin wax lens. The intensity
distribution of the resulting radiation was measured, as heating effect 'caused by the rays falling on a conductor, by a modified form of radiometer (q.v.). With such apparatus Nichols and Tear obtained waves the length of whose fundamental was as low as 1.8 mm., and in certain cases very prominent harmonics of wave-length .22 mm. These harmonics were apparently the fundamental wave-length of the part of the minute oscillator cylinders which projected through the glass into the oil, and pre dominated, so that it may be claimed that Nichols and Tear extended the Hertzian spectrum down to .22 mm.
More recently still Glagolewa-Arkadiewa has pushed things further by the use of an entirely new type of generator. This consists of a suspension of fine metal filings in machine oil, kept in constant motion by a rotating disc. The spark from an induc tion coil passes through a mass of closely packed filings separated by oil. This disposition has the advantages that the vibrators, consisting each of a pair of filings, are exceedingly small, which makes the wave-length small: that there is a very large number of them, which gives the requisite intensity: and that they are constantly renewed, which prevents the spark burning them away at the end. The wave-lengths are investigated with a Boltzmann interferometer, while as a detector special thermocouples are used. With this apparatus Glagolewa-Arkadiewa has measured wave-lengths of from 5o mm. down to .1 mm. This represents the limit reached at the present time (1928) in the way of short wave-lengths produced by what, in spite of all modifications, is essentially the method of Hertz.
The approach from the infra-red side was mainly carried out by Rubens and his collaborators. Important advances were made by the method of Reststrahlen ("rest-rays" or "residual rays"). The method depends upon the fact that certain crystals have a strong absorption, accompanied by metallic reflection, for a small range of wave-lengths in the infra-red, at a frequency correspond ing to the period of vibration of the crystal itself. (See LIGHT, Refraction in Absorbing Media; SOLID STATE.) The reflection is strongly selective, that is, very much stronger for the particular narrow range of wave-lengths than for wave-lengths on either side. In consequence repeated reflection from a few (say four) surfaces leaves only the selectively reflected rays present in any intensity. With rock salt and sylvite (potassium chloride) rest-rays of wave length 52/Land 63/.4 respectively are obtained, while with thallium iodide rest-rays as long as 152/.4 result. An advantage of the method of rest-rays is that the experimenter has at his disposal a broad beam, with the consequent relatively high intensity.