A New Scale of Electromagnetic Waves

A. A. Glagoleva-Arkadieva.

Twenty-five years ago P. N. Lebedev published, under approximately the same title, an article containing a table of electromagnetic oscillations then available to researchers for their work. The scale of electromagnetic waves already then comprised fairly extensive regions, divided, according to the method by which they were obtained, into two groups. The first group encompassed the region of electrical oscillations obtained in dynamo machines and in capacitor discharges; to it belonged the oscillations in Weber’s inductor at one oscillation per second; the alternating current of a municipal electric lighting station at 50 oscillations per second; the 500 oscillations per second obtained by Lodge in the discharge of a large battery through self-induction and a spark, which produced a musical sound; oscillations of the order of \(10^4\) in Tesla’s dynamo machine; of the order of \(10^5\) per second, obtained by Feddersen in the spark discharge of a capacitor; then the waves of Hertz’s vibrator, of the order of \(10^8\) oscillations per second, corresponding to wavelengths down to 60 cm; the waves of Righi’s vibrator, from 15 to 3 cm in length; and, finally, Lebedev’s waves of 6 mm. Thus this region of the wave scale comprised 36 octaves, corresponding to wavelengths from 250,000 km down to 6 mm.

The second region of the wave scale, characterized by a different method of exciting electromagnetic oscillations—chiefly by the method of an incandescent body—was less extensive than the first: it included the visible spectrum, with adjoining, on one side, invisible infrared and heat rays down to 62 μ, and, on the other, likewise invisible ultraviolet rays down to 0.1 μ in wavelength. This region of the scale extended over only 9 octaves. Between it and the first region there remained a gap of more than three octaves, the filling of which was of very great importance in the sense of experimentally confirming the identity of the nature of light waves and electromagnetic waves.

Beyond the limits of the ultraviolet waves there was a completely unknown region. Although Röntgen rays at that time already

were discovered, they had no place on the scale of electromagnetic waves, since their nature still remained unexplained. Only as a hypothesis was the thought expressed of their electromagnetic nature; P. N. Lebedev writes in his article:

“In any case, in the study of shorter ultraviolet waves there opens up the attractive prospect of connecting the optics already known to us with the rays of Röntgen and Becquerel.”

These assumptions have at the present time already been justified. During the last 14 years a number of brilliant discoveries and most remarkable investigations have been made, which have widely enriched science. Investigators now have at their disposal a scale of waves which has not only greatly expanded its boundaries and been supplemented by a new enormous region—the rays of Röntgen and the γ-rays—but has also become continuous, owing to the filling of two gaps that for a long time had existed on it between adjacent regions. Therefore a brief survey of the whole scale of waves in the light of the new data is very timely.

The region of electrical oscillations arising in the discharges of capacitors and in dynamo machines at the present time extends over 40 octaves. For exciting oscillations by means of the discharge of capacitors, use is made of a circuit (Fig. 1) consisting of a capacitance \(C\) and a self-inductance \(L\). When the capacitor is discharged, damped electrical oscillations are established in the circuit, whose period is calculated by the formula

\[ T = 2\pi \sqrt{LC}. \]

Fig. 1. Diagram of an oscillatory circuit.

The wavelength corresponding to these oscillations is determined from the relation

\[ \lambda = cT, \]

where \(\lambda\) is the wavelength, and \(c\) is the velocity of light. With different values of \(L\) and \(C\), this method makes it possible to obtain electrical oscillations with a period varying within very wide limits. The lower limit, at very large values of \(L\) and \(C\), is constituted by the slowest oscillations known so far, corresponding to electromagnetic waves up to two million kilometers in length. These oscillations were observed by Martienssen in 1910 in the discharge of a capacitor having a capacitance of 1000 mF through a self-inductance of 1000 henrys. The period of oscillation, consequently, was equal to 6.3 seconds, i.e. one oscillation lasted a full 6.3 seconds; the alternating discharge current was observed with the aid of an ordinary direct-current ammeter.

A. A. GLAGOLEVA-ARKADYEVA

direct current, whose needle alternately deflected now to the right, now to the left, coming to rest after five complete periods, or after 31.5 seconds. This means that the entire oscillatory process of one discharge of the capacitor lasted more than half a minute.

To obtain rapid electrical oscillations, as is known, Hertz as early as 1887 gave the oscillatory circuit a somewhat different form; striving to reduce the capacitance and self-inductance, he composed the entire oscillatory circuit of two cylinders with spheres at the inner ends, i.e., he straightened the former circuit \(a\) (Fig. 2) into \(b\), replaced the self-inductance \(L\) of the spiral by the small self-inductance of the cylinders themselves, and achieved a reduction of capacitance by moving the plates of the capacitor \(C\) apart to the opposite ends of the straightened circuit and then completely eliminating them, leaving only the capacitance formed by the outer ends of the cylinders themselves. The Hertz vibrator \(c\), constructed in this way and excited by sparks from the inductor \(J\), emitted waves whose length proved to be related to the length \(l\) of the vibrator itself by the simple relation \(\lambda = 2l\).

Fig. 2. Hertz vibrator.

By the method of the Hertz vibrator, very short electromagnetic waves have now been obtained; they will be discussed below.

Electrical oscillations propagate in free space or in wires. In the first case, the wavelengths are calculated by the formula given above or are measured by the method of standing waves formed when they are reflected from a plane mirror; wave indicators are resonators with a spark gap or with a thermoelement connected to a galvanometer. In the second case, when the waves travel along wires, their lengths are measured by means of standing waves formed in the wires. A very convenient arrangement for measurements with waves propagating in wires was proposed by Lecher. His system (Fig. 3) consists of two parallel copper wires terminating in capacitors \(CC\); \(I\)—a spark gap—

Fig. 3. Lecher system.

the loop connected to the terminals of the inductor. A metallic bridge $a_1$ is thrown across the wires; when a spark jumps in the circuit bounded by the bridge, electrical oscillations of a definite period arise. The length of the wave corresponding to them is measured with the aid of a second bridge, $a_2$, which is moved along the wires until the circuit bounded by the bridges $a_1$ and $a_2$ is tuned to resonance with the primary circuit; the distance between the bridges is equal to $\dfrac{\lambda}{2}$; the waves are detected by means of a Geissler or wireless tube $g$, which glows only under the condition of resonance of the secondary circuit. In this way Lecher measured waves down to $4\ \mathrm{m}$.

The shortest waves traveling along wires were obtained by V. K. Arkadiev. The scheme that he used is shown in Fig. 4. $V$ is a Hertz vibrator $3\ \mathrm{mm}$ long, supported by glass tubes $aa$; $f$ is the cross-section of a rubber tube containing acetone, in which the exciting oscillations take place.

Fig. 4. Apparatus for obtaining currents of extremely high frequency.

Spark discharges of the inductor: $ee$ are the wires bringing the spark, $R$ is a resonator connected with two long parallel wires $DD_1$. Owing to the reflection of waves from the bridge with the thermoelement $T$, standing waves are formed between $R$ and $T$; by moving the thermoelement along the wires, one finds the positions at which the galvanometer $G$ connected with the thermoelement gives the greatest deflections; these places characterize a maximum of current in the bridge or a minimum of electric voltage, i.e. a node in the standing wave; having determined the positions of the nodes, the wavelength is measured directly. The shortest waves obtained in this way by V. K. Arkadiev, with a length of $11\ \mathrm{mm}$, are still the limiting ones. In this case an alternating current of extraordinarily high frequency flows along the wires—up to $2.7 \cdot 10^{10}$ periods per second.

The period of electrical oscillations in dynamos depends on the speed of rotation of the armature and on the number of magnetic poles, and can also be varied within very wide limits. It is well known to everyone that the alternating current of a municipal electric lighting station represents electrical oscillations of a rather large pe-

period—0.02 seconds. Since the wavelength corresponding to these oscillations reaches 6000 km, in the short conductors alternating current of an urban station represents a motion of electricity that periodically changes its direction simultaneously throughout the entire network of the city current.

Tesla’s dynamo machine (Nikolas Tesla) had 384 stationary poles and rotated at a speed of 3000 revolutions per minute; under these conditions the frequency of the electrical oscillations was equal to 9600, or the period of one oscillation lasted approximately one ten-thousandth of a second. With a further increase in the frequency of oscillations, great technical difficulties were encountered, connected with attaining high speeds of the machine armature. The greatest known frequency of electrical oscillations up to now, equal to 100000 oscillations per second, was attained by Alexanderson (Alexanderson) in 1910 in his high-frequency machine, used as a source of electromagnetic waves at radio stations. Alexanderson’s machine has an iron disk with 300 teeth, rotating at a speed of 20000 revolutions per minute; the stationary parts of the machine contain not only windings for excitation, but also armature windings. The wavelength corresponding, under such conditions, to the frequency of oscillations indicated above is 3 kilometers.

Fig. 5. Diagram of the operation of a vacuum tube with a grid.

Very great successes have recently been achieved in the field of obtaining electromagnetic waves with lengths from several kilometers to several centimeters—waves that have received enormous practical application in radio engineering. Hertz’s vibrator, as well as high-frequency dynamo machines, which played a large role in the initial and middle periods of the development of radiotelegraphy and telephony, in recent years, with the striving of technology to arrange the operation of radio stations on short waves, have been almost completely displaced by a new source of electromagnetic waves, the so-called tube generator of undamped oscillations. The basis of the new method is a vacuum tube having three electrodes. \(KK'\) is the filament (Fig. 5), representing the cathode, \(A\) the anode, \(G\) the grid. The circuit for connecting the tube has three circuits: the filament circuit, the anode or cylinder circuit, and the grid circuit. Point \(T\) is common to all three circuits. The tube may, as desired, serve as an amplifier, a detector, and a generator. The scheme of its operation is as follows: with a certain heating, the filament \(KK'\) begins to emit electrons; under the voltage of battery \(E\), a current begins to flow from the anode to the filament; if a potential is applied to the grid, then with a positive potential of the grid the current

in the anode circuit increases, since the positive potential of the grid promotes the motion of electrons toward the anode; if, however, a negative potential is imposed on the grid, the motion of electrons toward the anode is impeded or stops altogether.

Applying these considerations to the circuit of a generator of undamped oscillations, one can imagine its operation as follows. The primary winding I of a transformer is inserted in the circuit of the anode or cylinder (Fig. 6), and the secondary winding II of the transformer is inserted in the grid circuit. The secondary coil II must be connected in such a way that, when the current in the primary winding I increases, a negative potential is imparted to the grid. The generator begins to operate when the filament-heating circuit is closed. The current in the anode circuit, owing to the self-induction of coil I, increases gradually. As the current in the primary coil increases, a negative potential is imposed on the grid, and the current in the cylinder circuit decreases. In connection with this, the grid receives an induced positive potential, and the current in the cylinder circuit again increases. The mutual action of the cylinder circuit on the grid circuit and conversely creates in the cylinder circuit a current of one direction, but of varying strength—a pulsating current. If a capacitor \(C\) is connected in parallel with the self-induction of the primary coil, then in the closed circuit thus obtained there will occur undamped electric oscillations, whose period will depend only on the capacitance \(C\) and the self-induction \(L\), while the amplitude will depend on the operating regime of the tube, on the mutual arrangement of the primary and secondary coils, and on other causes1.

Fig. 6. Circuit of a tube generator of undamped oscillations.

Measurement of wavelengths is carried out with the aid of wavemeters, which are a second oscillatory circuit that can be tuned to resonance with the first circuit, or else by the method of waves in wires described above.

In connection with the requirements of modern radio engineering, great efforts are now being made to obtain undamped oscillations of as short a period as possible. The difficulty of obtaining such oscillations by means of a tube generator comes down to the fact that the decrease in wavelength is associated with a decrease in the dimensions of the tube,

that can be allowed only within known limits. Short waves of undamped oscillations at \(2.4\) m were obtained by Holborn. Barkhausen and Kurz (Barkhausen u. Kurz) obtained very short waves—down to \(42\) cm; by the Barkhausen and Kurz method Scheibe obtained the shortest waves—\(30\) cm—which so far appear to be the limiting ones.

Very short electromagnetic waves of this region of the scale, \(6\) mm in length, had already been obtained long ago, in 1895, by P. N. Lebedev; his source of radiation was a very small Hertz vibrator \(2.6\) mm long; the waves propagated in free space and were detected with the aid of a thermoelement. Further advance along the scale of electromagnetic waves toward the neighboring region was associated with great difficulties of a technical nature. The Hertz vibrator method, extremely convenient for obtaining long waves, became greatly complicated in attempts to obtain waves a few millimeters long and shorter. For almost three decades the region of Lebedev’s electromagnetic waves remained separated from the extreme waves of the neighboring region—the heat waves of Rubens—by an interval of several octaves. Since filling this gap was of very great fundamental importance, very many investigators devoted no little labor and effort to solving this problem. Concerning the difficulties which experimenters inevitably had to encounter in obtaining the shortest waves of this region, P. N. Lebedev writes in his article: “........to obtain oscillations lying between \(\lambda = 3\) mm and \(\lambda = 0.1\) mm we must find a new source.”

“At present we have no possibility of foreseeing how this difficulty will be resolved; in any case, considerable difficulties will be encountered here, and a method of obtaining still shorter waves will be a very large step forward in the field of experimental physics.”

These words, as will be seen from what follows, proved prophetic.

The difficulties associated with applying the Hertz vibrator method to obtain short electromagnetic waves are as follows. With a reduction of the dimensions of the vibrator to several millimeters, the simple relation between the wavelength and the length of the vibrator, \(\lambda = 2l\), is violated, and this violation occurs in the direction of increasing the technical difficulties in working with the Hertz vibrator, namely: as the length of the vibrator decreases, the ratio \(\frac{\lambda}{l}\) increases more and more; this was noted in 1911 by V. K. Arkadiev: a vibrator \(3\) mm long, with which he worked, emitted waves \(11\) to \(13\) mm in length; hence the value of \(\frac{\lambda}{l}\) was equal to \(4\), instead of \(2\).

Later, in 1923, Nichols and Tear, working by Hertz’s vibrator method with very short waves, had a vibrator $0.4\ \mathrm{mm}$ long, and the wavelength corresponding to it was $1.8\ \mathrm{mm}$; the ratio $\frac{\lambda}{l}$ had already increased to 4.5. This circumstance was one of the chief difficulties in working with Hertz’s vibrator; if, even with a further reduction of the dimensions of the vibrator, the ratio $\frac{\lambda}{l}$ remained approximately the same, i.e. equal to 4.5, then in order to obtain a wave with $\lambda$ equal, for example, to $0.4\ \mathrm{mm}$, it would be necessary to have a vibrator of length $\frac{0.4}{4.5}=0.089\ \mathrm{mm}$, i.e. each half of the vibrator, even under such conditions, would have to have a length of about $0.045\ \mathrm{mm}$; of course, work with so small a vibrator by Hertz’s method is impossible.

No lesser difficulties arise in connection with the burning away of the vibrator by the spark that excites the oscillations. The burning of the vibrator, when its dimensions are large, has no noticeable influence on the length of the wave emitted by it. But when the vibrator is small, together with its gradual shortening due to burning, the gradual decrease in the length of the wave emitted by it becomes noticeable; in other words, with a small vibrator it is difficult to obtain a stable wave of constant length.

Great difficulties are caused by the fact that, as the dimensions of the vibrator decrease, the energy of the emitted waves decreases; this circumstance makes it necessary to greatly increase the sensitivity of the measuring instruments.

Despite the difficulties just mentioned and many others not mentioned here in applying Hertz’s vibrator method to the production of the shortest electromagnetic waves, many investigators followed precisely this thorny path. By this method short waves were obtained by Lampa at $4\ \mathrm{mm}$, by Baeyer at $2\ \mathrm{mm}$; Möbius, in addition to stable waves of $7\ \mathrm{mm}$, obtained very unstable, non-reproducible, transient electrical oscillations corresponding to higher overtones of the vibrator and resonator, down to $0.1\ \mathrm{mm}$; the presence of overtones can be judged from the zigzags on the interference curves obtained for the fundamental oscillation of the vibrator.

With an extraordinary expenditure of technical effort and resources, Nichols and Tear in 1923 succeeded by this method in obtaining stable waves down to $4.2\ \mathrm{mm}$ and $1.8\ \mathrm{mm}$, and also in detecting, in the form of an overtone, a weak wave of $0.8\ \mathrm{mm}$. The work of Nichols and Tear was carried out with unusual elegance and delicacy. They obtained their shortest waves with a very small Hertz vibrator (Fig. 7), each half of which had a length and diameter of $0.2\ \mathrm{mm}$. These

microscopically small grains were soldered into the glass tubes \(a\) and \(b\). For such a small vibrator to operate properly, it was necessary to pass a strong jet of kerosene through the spark gap, and to blow the supplying spark with a jet of compressed air. The very small emission energy of the small vibrator led them to the necessity of constructing special, very sensitive measuring instruments—wave receivers.

Fig. 7. Vibrator for very short waves of Nichols and Tear.

Fig. 8. Receivers or indicators of short waves of Nichols and Tear.

Figure 8 shows various receivers constructed on the principle of resonance; on the vane \(a\) of the radiometer \(A\) there were small pieces of a thin layer of platinum or Wollaston wire; at resonance the incident waves were absorbed by them, as a result of which the radiometer vane was deflected. The sensitivity of the instrument was very high: it is enough to mention that the entire suspension system together with the mirror weighed from \(1\) to \(0.5\) mg.

Fig. 9. Staircase grating.

The measurement of wavelengths was carried out by the interference method with the aid of Boltzmann mirrors. The details of this method will be set forth below. Waves with \(\lambda\) equal to \(3.8\) and \(1.8\) mm were measured with a reflecting grating in the form of a staircase, or staircase grating, similar to that which Milentz first used in Moscow in 1914 when working with electromagnetic waves. The staircase grating of Nichols and Tear consisted of 8 well-prepared metal plates laid one upon another (Fig. 9), with the height of the steps equal to \(\frac{\lambda}{2}\) of the fundamental wave of the vibrator. The height.

the steps could be varied by changing the inclination of plate A. The waves reflected from the grating became more monochromatic.

The later work of Nichols and Tear was devoted to developing the method of overtones of the Hertz vibrator for obtaining the shortest electromagnetic waves. For this purpose they “tuned” the vibrator and the receivers’ resonators to one and the same wavelength, which corresponded to one of the higher harmonic oscillations relative to the fundamental oscillations of the vibrator and resonator. Thus, for example, a wave with \(\lambda\) equal to \(0.8\) mm was obtained as the 4th overtone of the vibrator and the 9th overtone of the resonator:

\[ \frac{\lambda_0}{5}=\frac{\lambda_r}{10}=0.8\ \text{mm}, \]

where \(\lambda_0\) and \(\lambda_r\) are the fundamental wavelengths of the vibrator and the resonator. In this way they obtained overtones with wavelengths down to \(0.22\) mm.

Using simultaneously several vibrators consisting of small metal balls, M. A. Levitskaya investigated the radiation of small spheres emitting waves corresponding both to the fundamental oscillations and to oscillations of higher orders. Calculations showed that the wavelengths for the vibrator dimensions she used range from \(\lambda_0=2.9\) to \(\lambda_3=0.62\) mm. The observations were made with a special thermoelement of tellurium and bismuth.

It proved considerably simpler to obtain waves lying in this intermediate region of the electromagnetic-wave scale by an entirely different method—the method of the “mass radiator.” The idea of the method, due to V. K. Arkadiev, was put forward by him as early as 1914. It became possible to realize this idea only in 1922. The method of the mass radiator consists in the fact that, in order to intensify the radiation energy, not one small Hertz vibrator is used, but a multitude of them. To avoid burning out from sparks, these vibrators are constantly replaced. To make it possible to use vibrators of extremely small size quite freely, the vibrators remain suspended in a liquid dielectric. The method of the mass radiator is something intermediate between the method of the Hertz vibrator and the method of the incandescent body; it constitutes a transition from the radiation of a single Hertz vibrator to the mass radiation of the molecules of a substance. By this intermediate method it was comparatively easy to obtain the intermediate—“ultra-Hertzian”—waves, the shortest waves of this region of the scale.

The mass radiator of A. A. Glagoleva-Arkadieva consists of a vessel \(A\) (Fig. 10), in which there is a mixture of brass or aluminum filings and oil; this mixture is kept in constant motion by the continuously operating stirrer \(M\) and has the appearance of a porridge-like homogeneous mass. A rotating carbolite wheel \(K\) picks up from the vessel the mass, which is located on

by the action of centrifugal force, with a ring like a liquid tire. The waves are excited by sparks from an inductor, brought up to the tire at \(f\). A great advantage of this source of waves is that it can operate for a very long time, for whole hours, without requiring special attention, like the burning of an electric bulb emitting heat and light waves. The radiation energy of the mass radiator is so considerable that the measurements were made with instruments of ordinary sensitivity.

Fig. 10. Mass radiator of Glagoleva-Arkad’eva.

Measurement of the wavelengths was carried out by the interference method with Boltzmann mirrors. In the focus of the parabolic mirror \(P_1\) (Fig. 11) was placed the radiating part of the mass radiator \(V\). The rays, after being reflected from \(P_1\), went as a parallel beam and fell upon two plane mirrors \(S_1\) and \(S_2\). After reflection from them the rays fell upon the second parabolic mirror \(P_2\), which collected them at the focus, where a thermoelement \(T\) was placed. By connecting the thermoelement with an armored galvanometer, the radiation energy can be detected. If the Boltzmann mirrors \(S_1\) and \(S_2\) are in one plane, then both parts of the beam of rays reflected from them arrive at the thermoelement in the same phase. Upon shifting one of the mirrors, for example \(S_1\), one half of the beam of rays will arrive at the thermoelement already in a different phase relative to the second half of the beam of rays; as a result, interference of the rays occurs. If

Fig. 11. Interference method with Boltzmann mirrors.

if one plots along the abscissa axis the displacements of the movable mirror \(S_1\), and along the ordinate axis the corresponding deflections of the galvanometer, then one obtains an interference curve of the radiation energy, the distances between the neighboring maxima of which will be equal to half the wavelength. Various types of thermoelements were used as indicators (Fig. 12). The measurements showed that the waves emitted by the mass radiator occupy a fairly broad region, about 7 octaves, from 50 to \(0.1\) mm. The radiation of the mass radiator proved to be “white”; one of the chief causes of this phenomenon must be considered to be the non-strict sifting of the grains of filings.

It is very important that the short electromagnetic waves emitted by the mass radiator can be the result of independent electrical oscillations in the corresponding very small vibrators, and not overtones superposed on the fundamental tone of the vibrator or resonator, as was the case in the works of Möbius and of Nichols and Tear.

Fig. 12. Various types of thermoelements.

For isolating the shortest waves emitted by the mass radiator, the following principle was applied. If electromagnetic waves fall on a metallic plate, then standing waves are formed in front of it with a node at the surface of the plate. By placing the metallic plate behind the thermoelement at a very small distance from it, one can weaken the action of the long waves on it and strengthen the action of the short ones; the action on the thermoelement will be strengthened only for those waves in whose antinode it is located. This principle made it possible to isolate the most intense wave at \(150\,\mu\), then less intense ones at \(300\) and \(180\,\mu\), and a weak wave at \(82\,\mu\).

Thus the shortest waves of this region of the scale extended far into the neighboring region of the scale—the region of infrared waves, po-

emitting in an entirely different way, chiefly by the blackbody method.

How difficult it was to approach the boundary of infrared waves from the side of Lebedev’s electromagnetic waves follows from what has been said above; but it was still more difficult to extend the limits of infrared waves from the side of Rubens’s heat waves. Rubens’s heat waves are the longest infrared waves, with which Rubens and his students worked very extensively and with extraordinary success. Only thanks to the enormous talent and great labors of this outstanding scientist was the region of infrared waves advanced so far. In a whole series of brilliant, extremely finely executed works, with the aid of his most ingenious methods, Rubens steadily moved forward, step by step approaching the boundary of the neighboring region.

The sources of infrared waves are: the ordinary voltaic arc, giving a strong continuous spectrum from which the infrared part of the spectrum can be isolated; then the voltaic arc between electrodes containing alkaline and alkaline-earth metals; the Auer burner and the mercury-quartz lamp.

Measurements in the region of short infrared waves were made with the aid of a spectrometer with a wire grating. The indicators used were a thermoelement or a bolometer. The arrangement of the parts of the instrument is shown in Fig. 13. \(S\) is the source of waves, \(D_1\) and \(D_2\) are diaphragms in the form of slits, \(C_1, C_2, C_3\) and \(C_4\) are concave metallic mirrors, \(G\) is a wire grating, \(P\) is a prism of a substance transparent to infrared rays, for example sylvine. The rays from \(S\) pass through the slit \(D_1\), fall on the mirror \(C_1\), are reflected from it and pass as a parallel beam through the prism \(P\), which disperses rays with different wavelengths in different directions; the mirror \(C_2\) gives a real image of the spectrum on the diaphragm \(D_2\), through whose slit only a narrow beam of monochromatic rays is admitted. The monochromatic rays reflected from the mirror \(C_3\) fall as a parallel beam on the grating \(G\), which gives diffraction spectra. The mirror \(C_4\), set at a certain angle, concentrates the rays of a spectrum of a definite order on the thermoelement or bolometer \(T\). In this arrangement the lenses are replaced by mirrors because glass absorbs infrared rays. The wire grating is constructed as follows. Two very accurately cut screws \(P_1\) and \(P_2\), with the same pitch, are fastened by two metal plates \(L_1\) and \(L_2\) (Fig. 14). A thin wire is wound on the screw threads and then soldered to them. As a result of such

Fig. 13. Spectrometer for measurements in the infrared spectrum.

NEW SCALE OF ELECTROMAGNETIC WAVES

method of manufacture, two gratings are obtained on the two sides of the frame, of which one, after the wires have been soldered on, is cut away. The diffraction gratings come out very regular.

The region of infrared waves, on the side of the very shortest waves bordering on the visible red rays, begins at about \(0.760\,\mu\). With the aid of the apparatus described it was possible to measure wavelengths up to \(20\,\mu\). Further measurements with longer waves were impossible owing to the great absorptive power of the prism material for these rays. To overcome the difficulty, Rubens replaced the method of spectral decomposition of the rays by a prism with the method of selective reflection of the rays from certain substances. If rays of different wavelengths fall on a plate possessing the property of selective reflection, then in the composition of the reflected rays there increases, in percentage terms, the number of rays with a large coefficient of reflection. By repeated successive reflection from such crystalline plates, the admixture of rays with a small coefficient of reflection can be made so small that the reflected rays, or “residual rays,” become almost monochromatic.

Fig. 14. Wire diffraction grating.

Fig. 15. Rubens’ residual-rays method.

By the method of “residual rays” Rubens obtained very long heat waves up to \(151\,\mu\). The apparatus which he used is shown in Fig. 15. The source of radiation in his arrangement was an Auer burner \(A\) without glass. \(B\) is a concave mirror, which gives a real image of the burner mantle in the air layer \(C\) between two quartz plates of a quartz interferometer. The rays enter the chamber \(K\), where the reflecting crystalline plates \(F_1\), \(F_2\), \(F_3\), and \(F_4\) are located. Reflected successively from these

the plate, the rays fall upon the concave mirror \(G\), which casts the image of the burner tube onto the thermoelement \(I\). The rays, passing through the quartz interferometer (Fig. 16), consisting of two plane-parallel quartz plates—one, \(G'\), fixed and the other, \(G\), movable—will interfere, and the result of the interference will depend on the thickness of the air layer between the plates.

Fig. 16. Quartz interferometer.

With a gradual change in the thickness of the air layer there will change, as a consequence of interference, the energy of the rays received by the thermoelement. Plotting along the axis of abscissas the displacements of the movable plate of the interferometer, and along the axis of ordinates the deflections of the microradiometer connected with the thermoelement, Rubens obtained interference curves, from which he determined the wavelengths of the residual rays of sylvine, rock salt, potassium bromide and iodide. In this way he measured waves at \(52\); \(63.4\); \(82.6\) and \(94.1\,\mu\). The residual rays of silver bromide, potassium bromide and iodide made it possible to extend the heat spectrum to \(151.8\,\mu\). Further progress by means of the residual-ray method became impossible, owing to the very low intensity of the rays.

Fig. 17. Method of quartz lenses.

Making use of another property of bodies in the region of still greater wavelengths, Rubens found a new way of isolating them—the “method of quartz lenses.” The basis of this method was the strong increase in the refractive index of quartz precisely on passing into the region of long waves of the infrared spectrum. The source of radiation in the apparatus with quartz lenses (Fig. 17), with which Rubens worked jointly with Baeyer, was a mercury-quartz lamp \(A\). The quartz lens \(L_1\) made it possible to collect the long heat waves separately from the short ones, since the focus of the short waves lay considerably farther than the focus of the long waves. Therefore, through the aperture of the diaphragm situated in front of the focus of the long heat waves, only the converging bundle of rays with long waves passed freely, whereas the short waves were ...

retarded. To retard the short heat waves traveling along the main axis of the lens, its central part was covered by a small screen \(d_1\). The long heat waves that passed through the aperture of the diaphragm entered a quartz interferometer \(J\), then a second quartz lens \(L_2\), which collected them on the thermoelement of a microradiometer. The measurement of wavelengths was carried out with the aid of interference curves, as in the preceding method. The wavelengths turned out to be very large—equal to 343 and 218 \(\mu\).

These data belong to the year 1910. For many years the waves of 343 \(\mu\) remained the limiting ones. Apparently, Rubens had reached the limit of what was possible in obtaining still longer heat waves. Attempts by other investigators in this direction did not yield major results. In 1924 Nichols and Tear repeated the experiment of Rubens and Baeyer with a mercury-quartz lamp, using as an indicator not the thermoelement employed by Rubens and Baeyer, but their own receivers for detecting short Hertzian waves, described above. These experiments of Nichols and Tear confirmed the existence of the long waves of Rubens and Baeyer and detected another wave at 420 \(\mu\), close to Rubens’s extreme wave of 343 \(\mu\). In the experiments of Gerda Lasky (G. Lasky), who was engaged in the study of questions connected with the radiation of the mercury-quartz lamp, a wave of 400 \(\mu\) was obtained.

In Fig. 18 are compared the interference curves obtained by various investigators in the transitional region of the scale from Hertzian waves to infrared rays. \(A\)—one of the curves of Möbius, obtained by him with Hertz vibrators; the short overtone waves, superposed on the fundamental wave in the form of serrations, were for him an inconstant, irreproducible phenomenon. \(B\)—the curve of Nichols and Tear, also obtained with a Hertz vibrator, reveals the shortest fundamental waves obtained by them, of 1.8 mm. Curve \(C\)—also of Nichols and Tear, constructed with a Hertz vibrator by the overtone method of the vibrator and resonator; it shows the shortest wave attained by this method, 0.22 mm. Curve \(D\)—of Glagoleva-Arkad’eva, obtained with a mass radiator; the shortest waves attained with it, of 150, 300, and 180 \(\mu\), are the shortest waves on the side of the Hertzian waves. Curve \(E\)—of Rubens and Baeyer, obtained with a mercury-quartz lamp; it reveals the longest heat waves, of 218 and 343 \(\mu\).

In this intermediate region of the wave scale, quite recently, in 1926, M. A. Levitskaya carried out new experiments on obtaining short electromagnetic waves, continuing her work with the radiation of a system of small vibrators. The experiments were performed with two systems. The first system consisted of small

pieces of molybdenum wire with a diameter of \(0.2\) mm and a length of \(0.1\)—\(0.4\) mm; the second system differed from the first in that the lengths of the small vibrators were taken to be identical and equal to \(0.1\) mm. The vibrators, in chains or rows parallel to one another, were glued with Canada balsam to a glass plate at the smallest possible distance from one another. Each of the two systems occupied an area of \(1\ \mathrm{cm}^2\), having 10 rows each \(1\) cm long. Each row of the system was fed from a separate coil of a transformer, which for this purpose had the corresponding number of secondary coils. The primary coil of the transformer was fed from the secondary coil of an ordinary induction coil, through a spark and a capacitor connected in series. The system of vibrators was placed at the focus of a parabolic mirror with a focal length of \(2.5\) cm.

Fig. 18. Interference curves obtained with various emitters in the intermediate region of the wave scale.

Measurement of the wavelengths was carried out at first with the aid of two plane diffraction gratings and a rock-salt lens. Since the latter has absorption in the wavelength region from \(100\) to \(20\ \mu\), for measurements in the region of the shortest waves two concave reflecting gratings were used. In the latter case both systems each consisted of a single row of small vibrators, with a total length of \(3\)—\(5\) cm. Thermoelements of various kinds served as indicators in all cases. Measurements with plane gratings revealed waves with lengths of \(138\)—\(450\ \mu\) and \(138\)—\(475\ \mu\). With the aid of concave gratings, waves with lengths from \(30\) to \(508\ \mu\) and from

34 to 470 μ, and in one case up to 915 μ. M. A. Levitskaya believes that the short waves obtained by this method belong partly to the thermal radiation of spark gaps and partly to the radiation of small vibrators; which of these waves belong to the radiation of the vibrators has not yet been resolved. The energy of the electrical oscillations of the vibrators described, as M. A. Levitskaya says, is extraordinarily small, and therefore she considers it necessary, for the investigation of the laws and properties of the radiation, to seek an especially intense source of these short waves.

Immediately adjacent to the very shortest infrared rays is the visible spectrum, occupying a very small region, less than one octave, from 0.760 to 0.400 μ. Beyond the extreme violet rays of the visible spectrum lie the invisible ultraviolet rays.

The sources of radiation of ultraviolet waves are: the voltaic arc between iron electrodes, the mercury and amalgam-quartz lamp, the spark between electrodes of different metals—zinc, aluminum, platinum, etc., whose spectrum contains the shortest waves, and other special sources.

Measurements of wavelengths were made with the aid of a spectrograph or a vacuum spectrograph. The ultraviolet spectrum begins, as was mentioned above, at about 4000 Å. Down to 3000 Å the ultraviolet rays were photographed with the aid of a spectrograph with glass lenses and prism. With an optical system of quartz the boundary of the ultraviolet spectrum was advanced farther, to 1850 Å. In order to measure still shorter waves, for which both glass and quartz were opaque, Schumann used a prism and lenses of fluorite; moreover, to reduce the absorption of short rays by air along their path through the instrument, he placed his spectrograph in a vacuum. He photographed the spectrum on special sensitive plates prepared by himself. By these means he succeeded in advancing the ultraviolet spectrum to 1230 Å. Further extension of the spectrum toward still shorter waves was limited by the strong absorption of them by the material required for the optical system. A great step forward along this path was made by Lyman, who devised a way to dispense with the ray-absorbing optical system by replacing it with Rowland’s concave reflecting grating.

Diffraction gratings made of wires, used for measurements in the infrared spectrum, are unsuitable for measurements with ultraviolet rays, since for short waves gratings with a small period are necessary. Meanwhile, in the manufacture of a wire grating, the limit of technical possibility is reached very soon: the production of correctly cut screws with a very small pitch and

properly stretched wires of very small diameter presents great difficulties. The ordinary diffraction gratings used for measurements in the visible spectrum, consisting of glass plates with fine, equidistant grooves parallel to one another ruled on them with a diamond, also proved inapplicable, since glass is opaque for the greater part of the ultraviolet spectrum. Therefore, for the investigation of short ultraviolet rays, reflecting diffraction gratings proved very convenient; in these the grooves are ruled with a diamond on a metallic mirror. Diffraction is observed in the reflected rays. If the grooves are cut on a concave mirror, then such a grating requires no additional collecting mirrors.

The Lyman vacuum spectrograph, constructed by him for the investigation of the spectra of gases, consisted of two parts: the spectrograph and the vacuum chamber in which it was enclosed1. The spectrograph itself consisted of a brass tube about 10 cm in diameter and about 1 m long, at one end of which was placed a concave reflecting grating, and at the other—a slit and a “Schumann” sensitive plate. The grating had 15,028 lines per 1 inch. The source of radiation in it was a discharge tube of quartz or with a fluorite window for the exit of rays into the vacuum spectrograph; the rarefaction inside the tubes was from 1 to 3 mm of mercury. In this way Lyman gradually advanced the boundary of the extreme ultraviolet rays first to 900, then to 600 and 510 Å.

In the region of the shortest waves of the ultraviolet spectrum, Millikan (R. A. Millikan) worked very extensively and successfully with his collaborators. Using the principle of Lyman’s vacuum spectrograph, Millikan and Sawyer constructed a vacuum spectrograph for the investigation of the spark spectra of metals. To weaken still further the absorption of rays by the medium, he removed the fluorite window from the apparatus, placing the wave source—the spark—inside the vacuum spectrograph. The vacuum in the apparatus reached less than \(10^{-4}\) mm of mercury. The source of radiation was a spark in the same vacuum between aluminum electrodes, produced by the discharge of a large capacitance charged to a high potential—several hundred thousand volts. Under these conditions Millikan succeeded, with great difficulty, in extending the ultraviolet spectrum to 202 and even to 136 Å. These Millikan waves are the shortest measured up to the present time.

In the very most recent period, measurements in the region of the extreme ultraviolet waves were made by Weinberg (M. Weinberg), who

he investigated the spark spectra of gallium and indium. In the spectrum of gallium, among the multitude of other spectral lines, at its extreme end, he found on one plate a weak line at \(126.3\ \text{\AA}\). His source of radiation was a spark between electrodes, one of which was made of aluminum, while the other consisted of a small quartz vessel containing the indicated metals. The measurements were made by means of a vacuum spectrograph with a concave grating. The region of the ultraviolet spectrum, represented by Millikan in the form of a graph (Fig. 19), at present occupies more than five octaves. The limiting ultraviolet rays overlap the longest waves of the neighboring region of the scale—the region of X-rays.

X-rays are excited by the method of bombarding bodies with rapidly moving electrons. The sources of radiation of this kind are X-ray tubes of very varied construction, according to the different purposes for which they are used. The necessary parts of every X-ray tube are the cathode, from which, under the influence of high voltage, a stream of electrons rushes forth, and the anticathode, on which X-rays arise.

Fig. 19. Ultraviolet spectrum.

The tubes are divided into two principal types: first, ion tubes, containing rarefied gas and having a cold cathode; second, electron tubes, having a high vacuum and a heated cathode. To obtain short X-ray waves, or “hard” X-rays, shorter than \(0.2\ \text{\AA}\), electron tubes are used, for example, Coolidge and Lilienfeld tubes; for work with waves of medium length, up to \(1.2\ \text{\AA}\), ion tubes are usually taken; to obtain “soft” X-rays, possessing wavelengths greater than \(1.2\ \text{\AA}\), special tubes are constructed with windows of a material more transparent than the glass of the tube itself, which strongly absorbs soft X-rays. To excite X-ray waves, a high voltage from an induction coil or transformer is applied to the tube, usually measured in tens and hundreds of kilovolts.

After the discovery of his rays by Röntgen in 1895, investigators soon succeeded in establishing their nature. Despite the extraordinary efforts of the most experienced experimenters to confirm the hypothesis of the wave character of these rays by means of ordinary optical methods—slits and wedge—all attempts proved unsuccessful and, at best, led only to an approximate estimate of the order of magnitude of the wavelengths of X-rays. This value came out to be approximately \(10^{-8}\ \text{cm}\). Sommerfeld’s theoretical interpretation of the experimental data gave results consistent with the same

assessment. Only seventeen years after Roentgen’s great discovery, in 1912, was it possible to produce interference of X-rays, thanks to Laue’s brilliant idea of using as a diffraction grating a natural, very fine grating formed by the regularly arranged atoms of matter in crystals. The distance between neighboring atoms in crystal lattices is of approximately the same order as the wavelengths of X-rays. Of course, diffraction phenomena in a crystalline spatial lattice must be much more complex than in an ordinary diffraction grating; these considerations were taken into account by Laue, who also gave a theoretical treatment of his method.

The experiment, proposed by Laue, was carried out by Friedrich and Knipping (Friedrich und Knipping). A narrow beam of X-rays fell upon a crystal of zinc blende, behind which there was a photographic plate. After development, dark, regularly arranged spots were found on the plate, testifying to the interference of X-rays. From the positions of the spots on the plate it was possible to determine the wavelengths accurately. From the first experiments Laue found wavelengths equal to 0.127; 0.190; 0.224; 0.355; 0.483 Å.

Laue’s discovery created an entire epoch in the study of questions concerning the structure of matter and the atom. By his experiments Laue opened a broad path to a whole series of most important works, as a result of which a large field of physics—X-ray spectroscopy—was created.

X-radiation is divided into two kinds: “braking” rays and “characteristic” rays. Braking rays are nothing other than a series of individual electromagnetic impulses arising as a result of the loss of velocity by rapidly moving electrons upon encountering the anticathode. This kind of radiation corresponds to white radiation in optics. It follows from this that braking rays can be resolved into a continuous X-ray spectrum. The curves of intensity distribution in the continuous X-ray spectrum have a maximum sharply bounded on the side of the short waves. This maximum, with increasing voltage (Fig. 20), shifts toward the short waves. Duane and Hunt (Duane a. Hunt) found a relation between the shortest wavelength in the spectrum and the corresponding voltage \(V\):

\[ \lambda_{\min} V = \mathrm{const}, \]

obtained from the formula \(eV = h\nu\), where \(e\) is the charge of the electron, \(h\) is Planck’s constant, \(\nu\) is the frequency of oscillation. If \(\lambda\) is expressed in Å, and \(V\) in kilovolts, then

\[ \lambda V = 12.3. \]

Consequently, in this kind of radiation \(\lambda\) does not depend on the substance of the anti-

cathode, i.e. on the kind of atoms, but depends only on the voltage \(V\), or on the velocity of the cathode rays.

The formula relating the minimum wavelength and the voltage proved valid over a very wide range—from several volts to 150 kilovolts. Using this formula, one can determine the smallest wavelength corresponding to a given voltage. The limit on the side of the shortest waves is set by the technical possibilities in manufacturing high-voltage apparatus and X-ray tubes. In X-ray installations of the newest types, voltages of about 250 kilovolts have now been attained, which corresponds to a wavelength of \(0.049\ \text{Å}\).

The second kind of X-ray radiation is characteristic rays, which depend on the substance of the anticathode. They correspond to monochromatic rays in optics. The method for studying X-ray spectra, thanks to the work of very many investigators—Laue, the Braggs (W. H. Bragg and W. L. Bragg), Moseley, Barkla, de Broglie, Siegbahn with his pupils, and many other researchers—has at the present time been developed with such precision and completeness that it constitutes a very extensive chapter of physics, the principal propositions of which have served as the subject of special works published recently.

Fig. 20. Curves of intensity distribution at various voltages—from 20 to 50 kilovolts.

X-ray spectra of the elements from sodium to uranium fill the region of the wave scale from \(17.3\) to \(0.1\ \text{Å}\). Beyond the indicated limit of the longest waves lie the characteristic frequencies of the light atoms; investigations in this region are made difficult by the fact that soft X-rays are strongly absorbed by media. The longest wavelengths of X-rays, \(493\ \text{Å}\), were obtained by Holweck. The source of radiation for him was an electron tube with an anticathode of special shape; the rays were detected by means of an ionization chamber, which had a pressure of \(1\ \text{mm}\) of mercury and was separated from the tube by a thin sheet of celluloid. The wavelengths were calculated from the formula given above, \(eV = h\nu\). At present this intermediate region—of X-rays and ultraviolet rays—is the subject of extensive

works aimed at establishing a connection between the spectral series of rays of the general type.

Thus the region of X-rays, extending over more than 13 octaves, on the one hand penetrates deeply into the adjacent region of ultraviolet rays, and on the other hand overlaps the next region of the wave scale—the region of \(\gamma\)-rays.

The source of \(\gamma\)-rays, as is well known, is radioactive substances, which emit, along with \(\gamma\)-rays, two other kinds of rays, \(\alpha\)- and \(\beta\)-rays. It is also well known that \(\alpha\)- and \(\beta\)-rays carry charges, \(\alpha\) a positive charge and \(\beta\) a negative one, whereas \(\gamma\)-rays carry no charge and in their properties resemble the hardest Röntgen rays, differing from them only by a considerably greater penetrating power. The latter circumstance indicates that \(\gamma\)-rays have shorter wavelengths than Röntgen rays.

Fig. 21. Rutherford and Andrade’s method.

The determination of the wavelengths of \(\gamma\)-rays was made by Rutherford and Andrade (Rutherford and Andrade). For this purpose they applied one of the spectrometric methods of measuring the wavelengths of Röntgen rays, a somewhat modified method of the crystal grating. The principle of Rutherford and Andrade’s method is as follows (Fig. 21).

At \(S\) there is a point or linear source of \(\gamma\)-rays, consisting of a thin wire activated by RaC and RaB (Fig. 23). The \(\gamma\)-rays fall on the crystal \(K\); the rays \(A\) and \(A'\), which have passed through the crystal, are stopped by the diaphragm \(B\); the rays \(R\) and \(R'\), reflected from the internal planes of the crystal, fall on the plate \(P\) and produce on it dark interference lines. By this method Rutherford and Andrade measured waves from 0.428 to 0.072 Å. For measuring shorter wavelengths, for which the crystal grating had already proved too coarse, it was necessary to find another method, the basis of which was the relation existing between \(\beta\)- and \(\gamma\)-rays. If \(\gamma\)-rays fall on some substance, electrons begin to be ejected from the latter, i.e. the body begins to emit \(\beta\)-rays. The energy of the \(\gamma\)-ray and of the \(\beta\)-ray are then connected by a relation which can be used to determine the wavelength of the incident \(\gamma\)-ray:

\[ \text{Energy of the excited } \beta\text{-ray} = \text{energy of the exciting } \gamma\text{-ray minus} \]

energy expended in tearing the electron away from one of the orbits.

The latter quantity depends on from which orbit and from an atom of what substance the electron has been torn away, i.e., it is characteristic for each element. It can be found from X-ray spectrometric measurements, from the boundary of the absorption bands. The energy of the excited \(\beta\)-rays can be measured from their behavior in a magnetic field, under the action of which the \(\beta\)-rays curve from their rectilinear path: first the velocity of the \(\beta\)-ray is calculated from the radius of curvature, and then its energy. From Einstein’s formula, relating the magnitude of the energy of a \(\gamma\)-ray to its wavelength, it is not difficult to calculate the wavelength of the incident \(\gamma\)-ray:

\[ E_{\gamma}=h\nu_{\gamma}=\frac{hc}{\lambda_{\gamma}}, \]

where \(h\) is Planck’s constant, \(c\) is the speed of light, and \(\nu_{\gamma}\) and \(\lambda_{\gamma}\) are the frequency and wavelength of the \(\gamma\)-ray.

Since \(\gamma\)-rays and Röntgen rays have very small wavelengths, the angstrom proved to be a very large unit for them. Therefore a new unit was introduced—\(X\); \(1\,X=10^{-11}\) cm, or \(1\,X=10^{-3}\) Å.

De Broglie and Cabrera (de Broglie et Cabrea) measured \(\gamma\)-ray wavelengths of \(171;\ 59.7;\ 53.0;\ 37.0;\) and \(29.7\,X\). Mesothorium was taken by them as the source of the \(\gamma\)-rays; the \(\beta\)-rays were excited when the \(\gamma\)-rays passed through plates of silver, tin, barium, gold, and uranium. With radium bromide as the source of the \(\gamma\)-rays and plates of silver, tantalum, platinum, gold, lead, and uranium as the source of the excited \(\beta\)-rays, they obtained shorter \(\gamma\)-ray wavelengths: \(51.9;\ 42.6;\ 35.6;\ 30.2;\ 20.6\,X\). The shortest \(\gamma\)-ray wavelengths were obtained by Thibaud (Jean Thiebaud) in 1924. His source of \(\gamma\)-rays was RaC. The \(\beta\)-rays were excited in plates of uranium, platinum, lead, tungsten, antimony, silver, selenium, and others. He obtained the RaC spectrum consisting of \(\lambda=20.5;\ 11.0;\ 10.0;\ 7.04\,X\). The last wavelength is so far the shortest in the region of \(\gamma\)-rays.

Very recently there appeared Millikan’s report on the “penetrating rays” of cosmic origin found by him, lying beyond the region of \(\gamma\)-rays. The idea of the existence of very penetrating rays was first expressed by Rutherford as early as 1903. He observed that the charge of an electroscope is retained longer if the electroscope is enclosed in a thick-walled metal box. Experiments carried out at first by Gockel, then by Hess and Kohlhörster (Hess u. Kohlhörster) and by other investigators confirmed the fact of the existence of strongly penetrating rays. In order finally to clarify the question of these unknown rays, Millikan in 1923 undertook a whole series of grandiose experiments, conducted by him not within the walls of his scientific laboratory, but on one of the highest mountains

of the United States of America. On the basis of the results he obtained, he came to the conclusion that rays of cosmic origin do indeed exist; they resemble γ-rays, differing from them by a greater penetrating power; they are not homogeneous, their spectrum lying between 0.67 and 0.4 X.

Thus the modern scale of electromagnetic waves occupies a very wide interval, extending over 76 octaves. Investigators have at their disposal waves from the very longest, of two million kilometers (50 terrestrial meridians!), with a number of oscillations of less than one per second, to the very shortest—in 0.4 X, or \(4 \cdot 10^{-12}\) cm, with a number of oscillations of ten thousand trillion, or \(10^{-22}\), in one second.

LITERATURE

1. Lebedev, P. N. The scale of electromagnetic waves in the ether. Phys. Rev. 2, 1901. Collected Works, p. 303. Moscow, 1913.

2. Die Kultur der Gegenwart, B. 1, Leipzig u. Berlin. 1925.

3. Ergebnisse d. exakten Naturwissenschaften, B. III, Berlin. 1924.

4. Martienssen. Langsame oscillatorische Entladung eines Kondensators von 1000 Mikrofarard. Ber. d. D. Phys. Ges. 12, 2, 1910.

5. Hertz, H. Über sehr schnelle elektrische Schwingungen, Wied. Ann. 31, 421, 1887.

6. Arkadiev, V. K. The absorption of electrical waves in wires. Zhurn. R. F.-Kh. O. 44, 165, 1912; Ann. d. Phys. 58, 105, 1919.

7. Tesla, Nikolas. Elektrotechnische ZS, Heft 25, 327, 1891.

8. Alexanderson, E. F. W. Wechselstrommaschine für die Frequenz. 100.000 E. T. Z. 30, 1003, 1909.

9. Holborn, F. Über Versuche mit Kurzen ungedämpften elektrischen Wellen ZS. f. Phys. 6, 328, 1921.

10. Barkhausen u. Kurz. Die kürzesten mit Vakuumröhren herstellbaren Wellen. Phys. ZS. 21, 1, 1920.

11. Scheibe, A. Untersuchungen über die Erzeugung sehr kleiner Wellen mit Glühkathodenröhren nach Barkhausen und Kurz. Ann. d. Phys. 73, 54, 1924.

12. Lebedev, P. N. On the double refraction of rays of electric force. Zhurnal R. F.-Kh. O. 27, 213, 1895; Wied. Ann. 56, 1, 1895.

13. Möbius, W. Über die Dispersion von Wasser und Äthylalkohol zwischen 7 und 35 mm Wellenlänge und Vorversuche zum Verwendung noch kürzerer elektrischen Wellen. Ann. d. Phys. 62, 293, 1920.

14. Nichols, E. F. and Tear, J. D. Short electric Waves. Phys. Rev. (2), 21, 587, 1923; Astroph. Journ. 61, 17, 1925.

15. Tear, J. D. The optical constants of certain liquids for short electric Waves Phys. Rev. (2) 21, 611, 1923.

16. Arkadiev, V. K. Über die Herstellung von schwachgedämpften kurzen Hertzschen Wellen. Versuche von B. Milentz. Phys. Zs. 23, 35, 1922.

17. Levitskaya, M. A. Ein Versuch von kurzen elektrischen zu den langen Wärmewellen überzugehen. Vorläufige Mitteilung. Phys. Zs. 25, 107, 1924, Elektrische Wellen im Gebiete des äusseren Ultrarot. Phys. Zs. 27, 177, 1926.

1. Glagoleva-Arkad’eva, A. A. A new source of short electromagnetic waves of ultrahigh frequency. Proceedings of the G. E. E. I. Issue 2, 1924; Nature 113, 640, 1924; Zs. f. Phys. 24, 153, 1924.

2. Rubens, H. and Hollnagel, H. Messungen im langwelligen Spektrum. Sitzungsber. d. Königl. Preuss. Ak. d. Wiss. p. 26, 1910; Rubens u. Baeyer. Über die Energieverteilung der von der Quarzquecksilberlampe ausgesandten langwelligen Strahlung. Sitzungsber. d. Königl. Preuss. Ak. d. Wiss. p. 666, 1911.

3. Lasky, G. Die langwellige Strahlung der Quarzquecksilberlampe bei verschiedener Belastung. Zs. f. Phys. 10, 353, 1922.

4. Schumann, V. Nature 69, 262, January 14, 1904.

5. Lyman, Th. The Spectrum of Hydrogen in the region of extremely short wave-length. Astroph. Journ. 23, 181, 1906; The extension of the Spectrum beyond the Schumann region. Astroph. Journ. 43, 89, 1916; The Spectrum of Helium in the extreme Ultra-Violet. Phys. Rev. 17, 34, 1921; The Vacuum Grating spectrograph. Journ. of the Opt. Soc. Amer. 7, 495, 1923.

6. Richardson und Bazzoni. The limiting Frequency in the Spectra of Helium, Hydrogen and Mercury in the extreme Ultra-Violet. Phil. Mag. 34, 285, 1917.

7. Millican, R. A. The extension of the Ultra-Violet Spectrum. Astroph. Journ. 52, 47, 1920; and Sawyer R. A. Extreme Ultra-Violet Spectra of Hot Sparcs in High Vacua. Phys. Rev. (2), 12, 167, 1918; and Bowen, J. S. Extreme Ultra-Violet Spectra, Phys. Rev. (2), 23, 1, 1924.

8. Weinberg, M. Sparc Spectra of Indium and Gallium in the Extreme Ultra-Violet Region. Proc. of Roy. Soc. A, 107, 138, 1925.

9. Röntgen, W. C. Über eine neue Art von Strahlen. Sitzungsber. d. Würzburg. Phys.-Med. Gesellsch. 1895.

10. Laue, M. Eine quantitative Prüfung der Theorie für die Interferenz-Erscheinungen bei Röntgenstrahlen. Friedrich W., Knipping P. und Laue M. Interferenz-Erscheinungen bei Röntgenstrahlen. Münch. Ber. pp. 303 and 363, 1912; Ann. d. Phys. 41, 971 and 989, 1913.

11. Bragg, W. H. and Bragg, W. L. X-rays and the structure of crystals. Trans. Prof. G. V. Wolf. Moscow, 1916.

12. Siegbahn, M. Spektroskopie der Röntgenstrahlen. Berlin, 1925, and Thoracus. Eine Erweiterung der röntgenspektroskopischen Gebietes. Phys. Ber. Heft. 10, 730, 1925.

13. Holweck, F. Recherches expérimentales sur les rayons x de grande longueur d’onde. Ann. de Phys. 17, 5, 1922.

14. Rutherford and Andrade. Spectrum of Penetrating γ-Rays from Radium B and Radium C. Phil. Mag. 28, 263, 1914.

15. De-Broglie et Cabrera. Etude des rayons γ au moyen de leur effect photoelectrique. Journ. de Phys. et le Radium. (4), 6, 224 S., No. 5.

16. Thibaud, J. Les rayon γ de très grand quantum et l’origine photoélectrique du spectre γ naturel du radium. C. R. 179, 165, 1924.

17. Millican, R. A. High Frequency Rays of Cosmic Origin. Nature 116, 823, 1925. December, 5.

Diagram labels (top to bottom):

• Martinsen 1910
• Weber inductor
• Alternating current of a central lighting station
• Lodge 1893
• Tesla dynamo machine 1893
• Alexanderson dynamo machine 1910
• Feddersen 1862
• Hertz 1887, 1889
• Barkhausen and Kurz 1900
• Lodge 1890
• Righi 1894
• Arkadiev 1912
• Möbius 1920
• Lebedev 1895
• Nichols and Tear 1923
• Glagoleva-Arkadieva 1923
• Rubens and Baeyer 1911
• Nichols and Tear 1925
• Levitskaya 1926
• Rubens and Hollnagel 1910
• Rubens and Aschkinass 1894
• Rubens and Nichols 1896
• Langley 1886
• Visible spectrum
• Knopp 1887

Vertical labels:

• Undamped oscillations
• Radio-frequency waves
• Hertzian waves
• Vortices
• Ultra-Hertzian
• Long-wave infrared
• Rays
• Electric oscillations in capacitor discharges
• Vacuum-tube generator
• Hertz vibrator method
• Method of the mass radiator
• Method of incandescent substance
• Method of the glowing substance

Scale markings visible on the left:

• 100 km
• 10 km
• 1 km
• 100 m
• 10 m
• 1 m
• 1 cm
• 1 mm
• 1 μm
• 250
• 500
• 1, 2, 4, 8, 16, 32, 64, 125, 250, 500 repeated along the scale.

• Visible spectrum
• Substances
• Ultraviolet rays
• Cornu, 1887
• Schumann, 1892
• Lyman, 1906
• Holweck, 1922
• Millikan, 1920
• Metallic electronic bombardment
• X-rays
• Siegbahn and Thoreus, 1925
• W. H. Bragg, W. L. Bragg
• Moseley
• Barkla
• de Broglie
• Siegbahn
• Coster et al.
• 1912–1925
• Rutherford and Andrade, 1914
• Laue, 1912
• de Broglie and Cabrera, 1923
• Rutherford and Andrade, 1914
• de Broglie and Kibber, 1923
• Thibaud, 1924
• X-radiation of elements
• Radioactive decay of matter
• γ-rays
• Cosmic rays
• Millikan, 1925

1. For details of the experimental method of investigating the vacuum-ultraviolet region, see Shponer’s article. Ed. ↩↩