STUDY OF THE COMPOSITION OF PRIMARY COSMIC RADIATION

S. N. Vernov, V. L. Ginzburg, L. V. Kurnosova,
L. A. Razorenov, M. I. Fradkin

INTRODUCTION

The study of the composition of the nuclear component and of the energy spectra of various groups of nuclei in primary cosmic radiation is essential both from the standpoint of the theory of the origin of cosmic rays and for the study of elementary acts of interaction of high-energy primary particles with atomic nuclei in the atmosphere.

According to the data presently available, the primary component of cosmic rays includes protons, α-particles and, in considerably smaller numbers, heavier nuclei. The distribution with respect to the charges of nuclei with \(Z > 2\) is still insufficiently studied, and with regard to the composition of the so-called nuclear component of cosmic rays there are a number of questions subject to experimental investigation. Most work on the study of the nuclear component of cosmic radiation has been carried out in the upper layers of the atmosphere, where the residual layer of matter above the apparatus is approximately 15 \(g/cm^2\). This leads to the fact that the true composition of the primary component beyond the boundary of the atmosphere can be obtained from direct measurements only by recalculation. For such a recalculation it is necessary to know reliably the relative number of light nuclei arising as a result of the destruction of heavier nuclei during their passage through the residual layer of the atmosphere above the apparatus. To eliminate the need for recalculation, measurements of the intensities of fluxes of nuclei with various \(Z\) should be carried out directly beyond the boundary of the atmosphere. The use of rockets for this purpose cannot give satisfactory results, since the flux intensity of nuclei with \(Z > 2\) is small and, during the very limited time that a rocket remains outside the atmosphere, too little data will be obtained.

Entirely new possibilities for the study of the nuclear component of cosmic rays are opened up by the launching of artificial Earth satellites. Placing apparatus on an artificial Earth satellite will make it possible to accumulate the necessary statistical material for determining nuclear fluxes of low intensity.

One of the most important questions concerning the composition of the nuclear component of cosmic rays is the question of the quantitative ratio between the fluxes of light nuclei Li, Be, B and the nuclei C, N, O, F. Knowledge of this ratio is important because, in all probability, the nuclei Li, Be, B, whose average abundance in the universe is very small, are emitted by the sources of cosmic rays only in negligible amounts, and the observed nuclei Li, Be, B are products of the spallation of heavier nuclei in their interaction with atomic nuclei of interstellar

media. Having made definite assumptions about the spatial distribution and power of the sources of cosmic rays and about the conditions of propagation of cosmic-ray particles in the interstellar medium, one can obtain values for the ratio of the fluxes of various groups of nuclei near the Earth.

Comparison of the ratio obtained in this way with that found as a result of measurements can serve as a criterion for the correctness of our theoretical conceptions. Thus, for example, the theory explaining the formation of cosmic radiation by acceleration of particles in expanding turbulent shells of supernova stars[^1] gives, for the ratio between the fluxes of the nuclei Li, Be, B and C, N, O, F, a value \(\gtrsim 0.1\). Owing to the inaccuracy of a number of parameters used in the estimate, this value may turn out to be several times larger; however, the theory is in sharp contradiction with the assumption that it is much less than 0.1.

The data of various authors on the relative number of Li, Be, B and C, N, O, F nuclei in the primary flux often contradict one another and, moreover, are insufficiently reliable, since they have been obtained by recalculating to the boundary of the Earth’s atmosphere the results of measurements in the stratosphere. Therefore, the final determination of the value of the flux ratio of these two groups of nuclei beyond the boundary of the atmosphere remains an essential problem.

With the aid of an artificial Earth satellite, clarity can also be brought to the question of the presence in the primary flux of nuclei with \(Z > 30\). Since the interaction cross section of such nuclei is very large, even in the case that such nuclei are present in the primary flux of cosmic radiation, at an altitude where the pressure is \(15\text{--}18\ \mathrm{g/cm^2}\), they will not be observed. There is no reliable information on the presence of nuclei with \(Z > 30\), but isolated cases of their registration give grounds for supposing that beyond the boundary of the atmosphere there is a noticeable flux of these nuclei. If such an assumption were experimentally confirmed, it would have very substantial significance for the theory of the origin of cosmic rays. Indeed, on average in the universe there are very few nuclei with \(Z > 30\)[^2], and, because of their large interaction cross section, they must have a comparatively short path in the interstellar medium. Therefore, the discovery of an appreciable number of such nuclei in the primary component would testify to an anomalous richness of the sources of cosmic rays in heavy elements.

EXPERIMENTAL DATA ON THE COMPOSITION OF PRIMARY RADIATION

After the discovery in 1948 of a noticeable number of heavy nuclei in the composition of primary cosmic rays, intensive study of this component began. The results of experiments carried out up to 1952–1953 are presented in reviews[^3],[^4], and here new data obtained in recent years will be briefly reported.

In the last few years our knowledge of the primary flux of cosmic rays has broadened very considerably. Measurements of the flux of \(\alpha\)-particles and nuclei with \(Z > 2\) have been carried out over a large interval of latitudes. The results of the measurements are presented in Tables I and II. From the tables it is clear that measurements carried out at one and the same geomagnetic latitude are not always in agreement with one another. Apparently, two circumstances play an essential role here: differences in the method used and differences in the geographic coordinates of the place of observation.

The significance of the latter circumstance is indicated by recent works[^5],[^6], according to which in a number of cases the difference in the magnitude of the fluxes

of heavy nuclei, measured at different points at one and the same geomagnetic latitude, can be eliminated if it is assumed that the geomagnetic coordinates determined on the basis of measurements of the Earth’s magnetic field at its surface do not correspond to the Earth’s magnetic field that acts on cosmic-ray particles. This assumption is consistent with the results of work ^7 on measuring the neutron flux in cosmic rays at various latitudes, according to which the geomagnetic equator does not coincide with the “magnetic equator for cosmic rays.” Unfortunately, there are no systematic data on the magnitude of the flux of primary nuclei at various points of the Earth, since in the measurements different methods were used (photographic emulsions, Wilson chambers, proportional, scintillation, and Cherenkov counters), the conditions under which the experiments were carried out (geographical and geomagnetic coordinates, altitude, time, etc.) were very diverse, and the reduction to the boundary of the atmosphere was performed under different assumptions about the magnitude of the parameters used.

In this connection it is necessary to carry out systematic measurements of the primary fluxes of the nuclear component of cosmic rays over the entire surface of the Earth, using apparatus of one and the same type for this purpose. This problem can apparently be solved only on an artificial Earth satellite. Carrying out such experiments will not only make it possible to determine more accurately the absolute values of the primary fluxes of various groups of nuclei, but at the same time will make it possible to form an idea of the character of the Earth’s magnetic field at large distances from its surface, since the trajectories of cosmic-ray particles are determined precisely by the region of the magnetic field remote from the surface.

Despite the appreciable scatter among the absolute values of the fluxes given in Tables I and II, an estimate of the relative values of the fluxes of nuclei at a given latitude (i.e., of nuclei with a given energy per nucleon) can be obtained with satisfactory accuracy if, in order to determine the relative abundance of various nuclei in cosmic rays, one takes the flux values obtained in one and the same experiment. In this way it is possible to exclude all uncertainties associated with the transition from one method to another. A comparison of the ratios found in this way is given in Figs. 1, 2, and 3, from which it is evident,

Fig. 1. Ratio of the flux of C, N, O, F nuclei to the flux of nuclei with \(Z > 10\). The dashed line is the mean value of the ratio.

Flux of α-particles in the primary

*) Value of the flux without allowance for fragmentations in the atmosphere above the apparatus.
**) Value of the flux with allowance for fragmentations.

cosmic radiation

Table I

Flux of C, N, O, F nuclei; \(Z>10\); \(Z>5\) in the primary.

*) Data are given for measurements at an altitude of \(\sim 30\) km.
*) The photographic emulsions were raised on a rocket. Ratio of the flux of C, N, O, F nuclei.
**) The value of the flux of nuclei with \(Z>2\) is given.

cosmic radiation

Table II

to the flux of nuclei with \(Z \geqslant 10\) in these measurements is \(2,66 \pm 0,30\).

that, within the limits of error (often very large), these ratios do not depend on the energy of the particles and are on the average equal to \(N_{\mathrm{He}}:N_{\mathrm{C,N,O,F}}:N_{Z>10}\simeq 53:3:1\). The fraction of \(\alpha\)-particles in the primary flux of nuclei with a given energy per nucleon amounts on the average (for the same cases as those used in Fig. 3) to \(11\pm 4.5\%\).

The fact that the ratio between the fluxes of different groups of nuclei, within the limits of error, does not depend on latitude indicates rather the identity of the energy spectra of different groups of nuclei than their difference, as is asserted in note \(^{41}\). However, the accuracy of the data at our disposal is too low for it to be possible to draw fully reliable conclusions about the identity of the energy spectra or, conversely, about their difference.

Thus, from the point of view of determining the energy spectrum of different groups of nuclei in primary cosmic radiation, the staging of experiments on an artificial Earth satellite also promises greater

Fig. 2. Ratio of the flux of \(\alpha\)-particles to the flux of nuclei with \(Z>5\).

possibilities, since fluxes of particles of different energy (at different latitudes) will be found with the aid of one and the same instrument, which will undoubtedly increase the reliability of our information about the energy spectrum of primary nuclei.

One of the most interesting questions concerning primary cosmic radiation is, as was already indicated above, the determination of the number of nuclei of the Li, Be, B group. Even at the present time the question of the relative abundance of these nuclei in the primary flux of cosmic rays cannot yet be considered definitively resolved. This is evident from the summary, given in Table III, of the results of various authors. With respect to the measurement of the flux of this group of nuclei, the importance of choosing the correct method was especially clearly manifested. As a result of work \(^{36}\) it was found that each of two different methods of identifying particle tracks in photoemulsions (determination of grain density using emulsions

of low sensitivity ^21 and the simultaneous measurement of the density of $\delta$-electrons and of the mean angle of multiple scattering ^38) leads to errors if it is used to determine the flux of light nuclei; moreover, in the first case an underestimate occurs, and in the second an overestimate in comparison with the true value of the flux. Apparently, the results obtained in work ^20 should be regarded as the cleanest. However, in all experiments carried out up to the present time, in order to obtain the flux beyond the “boundary” of the atmosphere it was necessary to make a recalculation, in which values known with little accuracy were used for the interaction cross sections and the probabilities of fragmentation of heavy nuclei, accompanied by the emission of fragments that are nuclei of light elements.

Fig. 3. Relative number of helium nuclei in the primary flux of particles with a given energy per nucleon.

In this connection it should be borne in mind that, at altitudes corresponding to a pressure of $15$–$18\ \mathrm{g/cm^2}$, where measurements are usually carried out, light nuclei of secondary origin constitute more than $60$–$70\%$ of all light nuclei.

Thus, at these altitudes what are actually measured are the fluxes of Li, Be, B nuclei that have appeared as a result of the fragmentation of heavier primary nuclei. For this reason there remain possibilities for criticizing the assertion that Li, Be, B nuclei are present in the primary flux, and experiments beyond the boundary of the atmosphere are necessary in order finally to resolve this question.

Table III

Relative number of Li, Be, B nuclei in the primary cosmic-ray flux

) The ratio of fluxes at an altitude of \(20\ \mathrm{g/cm^2}\) is given.
) Two sets of values correspond to different values of parameters used in recalculating to the boundary of the atmosphere.
) The ratio of the flux of Li, Be, B nuclei to the flux of nuclei with \(Z > 5\) is given.
) The photomultipliers were carried aloft on a rocket.
***) The ratio of the flux of Li, Be, B nuclei to the flux of nuclei with \(Z > 5\), measured at an altitude of \(15\ \mathrm{g/cm^2}\), is given.

EXPERIMENTAL METHOD FOR STUDYING THE SPECTRUM OF NUCLEI BY CHARGE IN PRIMARY COSMIC RADIATION

A large part of our information on the fluxes of nuclei with different \(Z\) has been obtained through the use of photographic emulsions lifted on balloon probes to altitudes of 20–30 km. Photographic emulsions make it possible, when determining the charge of each nucleus, to study in detail the behavior of the nucleus in the emulsion and to determine, simultaneously with the magnitude of the flux, the interaction cross section. When this method is used, the possibility of confusing a multiply charged particle with other phenomena—nuclear disintegration, a shower, etc.—is almost completely eliminated. However, in the presence of particles of moderate energy, the photographic-emulsion method can lead to errors in the measurement of charge. In some cases the efficiency of recording certain groups of nuclei may differ from 100%, and for this reason the charge distribution of nuclei may be very strongly distorted. From this point of view, methods in which there is no discrimination of particles with respect to their charge or mass are preferable. Among such methods is the use of particle counters in which the electrical pulse arising when a charged particle passes through them depends on the magnitude of the particle charge. The use of instruments of this kind on an artificial Earth satellite has the further advantage over photographic emulsions that it will make it possible to transmit information to Earth by radio. This circumstance facilitates the general processing of the data and, in particular, processing for the purpose of studying variations in intensity, determining the energy spectrum, etc. The necessity of ejecting from the satellite and subsequently searching on Earth for the instruments recording cosmic rays is also eliminated.

Methods based on the ionization of a medium by fast charged particles have the disadvantage that particles with small \(Z\) and comparatively low velocity can imitate a relativistic particle with large \(Z\). This is due to the fact that ionization increases both with increasing charge and with decreasing velocity. For this reason it is difficult, and sometimes impossible, to measure with sufficient accuracy the flux of nuclei with \(Z>2\), since protons and \(\alpha\)-particles of low energies, whose number may exceed the number of the nuclei under study, will produce ionization corresponding to the passage of a heavier relativistic nucleus. This remark applies also to proportional counters, pulse ionization chambers, and scintillation counters. An error of the same kind is produced by nuclear disintegrations in the material of the counter, since slow, strongly ionizing particles are then emitted, and the signal appearing as a result at the counter output will imitate the passage of a relativistic multiply charged particle.

An instrument free of the shortcomings indicated above is a counter based on the use of Vavilov—Cherenkov radiation (a Cherenkov counter)\(^{46}\). A Cherenkov counter consists of a detector (a transparent substance), a photomultiplier, and an amplifier. In the detector, when charged particles pass through with sufficiently high velocity (greater than the velocity of light in this substance), Vavilov—Cherenkov radiation arises. The latter, as is known, has the property that the angle between the direction of the particle and the radiation it emits is determined by the relation \(\cos \theta = \frac{1}{\beta n}\), where \(\beta = \frac{v}{c}\); \(v\) is the particle velocity, \(c\) the velocity of light, and \(n\) the refractive index of the detector material. Consequently, the minimum value of the velocity of a recorded particle is \(\beta_{\min} = \frac{1}{n}\), and particles having a velocity less than \(\beta_{\min}\) will not be

be registered by a Cherenkov counter. Therefore, in contrast to an ionization chamber and proportional and scintillation counters, a Cherenkov counter does not register nuclear disintegrations and nonrelativistic particles. The intensity of a flash of Cherenkov radiation is proportional to the square of the charge \(Z^2 e^2\) of the particle passing through the detector, depends on the velocity \(\beta\), on the refractive index \(n\) of the detector material, and also on the path length of the particle in the detector:

\[ \Delta N = \frac{4\pi^2 Z^2 e^2}{h c^2}\,\Delta \nu \left(1 - \frac{1}{n^2 \beta^2}\right) l, \]

where \(\Delta N\) is the number of photons in the frequency interval \(\Delta \nu\), emitted along the path \(l\) (cm), \(e\) is the elementary charge, \(c\) is the speed of light, and \(h\) is Planck’s constant. (It is assumed that the refractive index is constant for the given frequency interval.) Thus, at one and the same velocity of nuclei \(\beta\) and one and the same path \(l\), the magnitude of the flash is proportional to \(Z^2\), which makes it possible, by registering the amplitudes of flashes of Cherenkov radiation, to investigate the spectrum of nuclei by charge in primary cosmic radiation.

Since the Earth’s magnetic field does not admit to latitudes below \(40^\circ\) particles with velocities less than \(0.94c\), when measurements are carried out in the latitude interval \(\pm 40^\circ\) all particles will have approximately the same velocity. Constancy of paths in the detector can, within certain limits, be ensured by specifying the direction of the registered particles with a telescope of counters and registering at the output of the photomultiplier only those pulses from flashes in the detector that were accompanied by the triggering of the telescope counters.

To obtain a reliable result in measuring the spectrum of nuclei by the indicated method, it is necessary to ensure:

1. Constancy of the intensity of the light flash in the detector during the passage of particles possessing one and the same \(Z\).

2. The smallest scatter of pulse amplitudes at the photomultiplier output when registering light flashes of a given intensity.

3. Registration of as large a number of particles as possible during the experiment (statistics).

4. Reduction to a minimum of the role of secondary effects, such as showers, secondary relativistic nuclei with small \(Z\), which have appeared as a result of the destruction of primary heavier nuclei in interactions with the detector material and parts of the apparatus above it.

Fulfillment of these requirements should lead to good resolution of the individual peaks (belonging to different \(Z\)) on the curve of the distribution of particles by charge.

The main source of scatter of pulses at the photomultiplier output is fluctuations in the number of photoelectrons knocked out from the photocathode of the multiplier. Obviously, the relative fluctuations will be smaller the larger the number of photoelectrons. Therefore the questions of effective collection of light on the photocathode, transparency of the detector, and efficiency of the photocathode are very important. Since the number of photons in the flash is proportional to the path length traversed by the particle in the detector, it would be possible to increase the number of photons by increasing the dimensions of the detector. However, the dimensions of the detector are limited, since in the detector material there will occur formation of nuclei of light elements as a result of the splitting of heavier nuclei. The most suitable material for making detectors in our case is Plexiglas, which has refractive index \(n = 1.5\) and high transparency in the wavelength region longer than \(\approx 3500\) Å.

When choosing the type of photocathode for the multiplier, it is necessary to match in the best possible way the spectral characteristics of the photocathode with the spectral composition of the Vavilov–Cherenkov radiation. The most suitable from this point of view is an antimony–cesium photocathode. Figure 4 shows the spectral distribution of the Vavilov–Cherenkov radiation, the spectral characteristic of a typical antimony–cesium cathode, and the dependence on wavelength of the number of photoelectrons emitted by this photocathode under the action of Vavilov–Cherenkov radiation. When using a Plexiglas detector and an antimony–cesium photocathode, the effective wavelength range for registration is from 3800 to 5000 Å.

Fig. 4. 1—the spectral distribution of Vavilov–Cherenkov radiation; 2—the spectral characteristic of an antimony–cesium photocathode; 3—the number of photoelectrons emitted by an antimony–cesium photocathode under the action of Vavilov–Cherenkov radiation (the scale along the ordinate axis for the indicated curves is arbitrary).

Let us consider the question of fluctuations of pulses at the output of the photomultiplier. Even in the absence of any spread in the intensity of light flashes for particles with a given \(Z\), the pulses at the multiplier output will be obtained different because of statistical fluctuations in the number of photoelectrons knocked out from the photocathode, and fluctuations in the number of electrons of secondary emission. In the general case, the distribution in the number of photoelectrons is described by Poisson’s law \(^{47}\). However, for a large number of photoelectrons (practically for \(n \approx 10\)) the Poisson distribution is described with sufficient accuracy by a dependence of the form

\[ f(n)=Ae^{-\frac{(n-\overline{n})^{2}}{2\overline{n}}}, \]

where \(n\) is the number of electrons knocked out from the cathode of the multiplier, \(\overline{n}\) is the mean value of \(n\), and \(A\) is a normalization factor. The half-width \(p\) of the distribution curve, i.e., its width at half the maximum height, is equal to

\[ p=2\sqrt{2\overline{n}\ln 2}\simeq 2.35\sqrt{\overline{n}}. \]

The relative half-width of the distribution curve, equal to the ratio of \(p\) to the mean number of electrons \(\overline{n}\), is

\[ \Pi=\frac{2.35}{\sqrt{\overline{n}}}. \]

In order that the peaks belonging to nuclei with different \(Z\) be well resolved, it is necessary that the relative half-width of the distribution curve for a singly charged particle be approximately 50% (Fig. 5).

This requirement imposes a limitation on the thickness of the Plexiglas detector, which must be not less than 2 cm. At the same time, as indicated above, the detector thickness cannot be greatly increased because of

splitting of primary nuclei in the detector substance. In connection with this, the optimum detector size will be 3–4 cm.

Fig. 5. Distribution curves of pulse amplitudes at the output of a Cherenkov counter (the half-width of the distribution curve for singly charged particles is taken as equal to 50%).

The transverse dimensions of the detector, in the case when it is connected to the multiplier cathode as shown in Fig. 6, are determined by the dimensions of the photocathode. It is desirable to make the transverse dimensions of the detector as large as possible, in order to obtain the maximum number of registered nuclei. In addition to a large photocathode area, to collect large statistics it is necessary to have the largest solid angle within which the directions of the registered particles lie. However, the magnitude of this solid angle is limited by the tolerance on the spread of particle paths in the detector. In addition to the spread in the path length traversed by particles in the detector, the inequality of the intensities of the light flashes of Vavilov—Cherenkov radiation caused by the passage of particles with a given \(Z\) is also due to other causes. First of all, the intensity of the flash depends on the velocity of the particle and, although \(\beta\) is confined within comparatively narrow limits, this will nevertheless lead to a considerable smearing of the peaks corresponding to each value of \(Z\). An estimate shows that if the particle distribution in energy has the form \(N(>E)=\dfrac{A}{E^{1.2}}\), then for singly charged particles with \(\beta \geq 0.66\) (which corresponds to the minimum velocity at which the particle produces Vavilov—Cherenkov radiation in a detector with \(n=1.5\)) one obtains the smearing of the distribution curve shown in Fig. 7. However, for low latitudes \(\beta_{\min}\) will be determined by the Earth’s magnetic field, which screens out particles with low velocity, and the smearing of the distribution curve will not be so large.

The effective area of the detector and the tolerance on the spread of paths determine the choice of the telescope arrangement, its geometrical factor*), and consequently also the total number of particles registered,

(Continued in issue 16)

*) The geometrical factor \(\Gamma\) of a telescope consisting of two rows of counters of areas \(S_1\) and \(S_2\) is defined as the product of the area of one row of counters by the solid angle under which the second row is seen. The geometrical factor has the dimension sterad · cm\(^2\) and is equal to \(\Gamma \simeq \dfrac{S_1S_2}{R_{12}^2}\), where \(R_{12}\) is the distance between the rows, with \(R_{12}^2 \gg S_1\) and \(S_2\).

by the apparatus during the given interval of time. The telescope must be such that not a single particle passing through the telescope passes through the detector, touching it only partially.

For obtaining the number of particles registered by the apparatus, it is sufficient to multiply their flux, expressed in the number of particles per steradian·cm², by the value of the geometrical factor of the apparatus \(\Gamma\). If the geometrical factor of the apparatus is chosen equal to \(5\ \text{cm}^2\cdot\text{steradian}\), then, on the basis of the results given in Tables I and II, the following statistics of the registered nuclei can be obtained: for \(Z \geq 6\), seven, and for \(Z=2\), 50 nuclei per hour.

Fig. 6. Schematic representation of an apparatus for registering nuclei of primary cosmic radiation.
1, 2—counters of the telescope that selects the particle direction; 3—side counters that give a mark in the event of passage of side showers; 4—photomultiplier; 5—Vavilov–Cherenkov radiation detector. The shaded counters 1 and 2 are included in a circuit of twofold coincidences; in addition, the triggering of more than one counter in each of groups 1 and 2 is recorded, in order to exclude cases of showers passing through the apparatus.

Fig. 7. Distribution curves for \(Z=1\) without taking account of the effect of the velocity distribution (1) and with this effect taken into account (2).

Thus, for example, in a week of observations it is possible to register approximately 1000 nuclei with \(Z \geq 6\), 7000 \(\alpha\)-particles, and the corresponding number of Li, Be, B nuclei. The proposed experiments on the study of the spectrum of nuclei envisage registration of the differential spectrum of nuclei in \(Z\) in the interval from \(\alpha\)-particles to oxygen. Such a method is possible only in the event that the apparatus can resolve each peak corresponding to different values of \(Z\).

\[ * \quad * \quad * \]

The possibility of installing instruments that register cosmic rays on an artificial Earth satellite opens broad prospects for posing new problems in the study of the primary flux. Among such problems should be included measurement of the flux of primary protons, clarification of the role of the “albedo” of the Earth’s atmosphere, determination of the lower limit for the flux of the electron–positron component, study of the interactions of primary particles with matter, study of temporal variations of the intensity of cosmic rays, of the energy spectrum, etc. In connection with the problems mentioned above, we shall make several remarks.

Experiments carried out on an artificial Earth satellite for measuring the primary flux of nuclei will already make it possible to obtain a whole series of additional data on the properties of primary cosmic rays, in particular on variations in the flux intensity of different groups of nuclei.

The question of variations in the intensity of the nuclear component is among those questions for which no answer has yet been obtained, since the observational results cannot be reconciled with one another, as is clearly shown in Table IV. It is quite possible that regular diurnal variations in the intensity of heavy nuclei do not exist at all, but, apparently, there are certain fluctuations of intensity, the character and causes of which are still unclear. The carrying out of continuous and prolonged measurements

Table IV

Data on diurnal variations of the intensity of the nuclear component of the primary flux of cosmic rays

) Ratio of the flux magnitude in the afternoon to the flux magnitude in the morning hours.
*) Ratio of the flux magnitude at 13 hours to the mean flux magnitude during the interval from 10 h 30 min to 15 h 30 min.

of the intensity of the fluxes of different groups of nuclei with the aid of apparatus installed on the satellite will make it possible to solve the question of variations of the nuclear component in the best way; it will make it possible to determine which groups of nuclei are subject to variations, how they are connected with variations in the intensity of protons, etc. Registration of the intensity by a “differential” method, i.e., of each group of nuclei separately, can provide clearer information on variations in the intensity of the nuclear component than the “integral” method of recording changes in the total ionization produced by all primary cosmic radiation.

A very important question is that of the energy spectrum of cosmic-

cosmic rays both as protons and as heavier nuclei. In the energy region up to \(30\cdot 10^9\) eV for protons and up to \(15\cdot 10^9\) eV per nucleon for particles with \(Z \geq 2\), the energy distribution can be determined on the basis of measurements of the latitude effect. However, even in this case a number of circumstances (for example, the lack of information on the nature of the magnetic field at large distances from the Earth’s surface, the presence of secondary particles emerging from the Earth’s atmosphere) make it difficult to obtain reliable data on the energy spectra of primary particles. In the region of high and ultrahigh energies the situation is still more difficult: the only source of information about particles with energy greater than \(10^{12}\) eV is extensive air showers. Determination of the spectrum of primary particles on the basis of studying the distribution of extensive air showers by size is connected with a whole series of assumptions and for this reason is insufficiently reliable and unambiguous. In this connection, it appears very promising to place aboard an artificial Earth satellite instruments for direct measurement of the energy distribution of high-energy particles. As a possible method for solving this problem one may use measurement of the ionization produced by the particles of an electron-photon shower, formed by a high-energy particle in a layer of absorber above an ionization chamber \(^{48}\). The number of particles (and, consequently, the ionization produced by them) within certain limits is proportional to the energy of the primary particle. Thus, using a pulse ionization chamber and recording the magnitude of the pulse produced in it by the particles of an electron-photon shower, one can find the energy of the primary particle and, after collecting sufficient statistical material, determine the energy spectrum of primary particles in the high-energy region. An estimate shows that an installation with an area of \(300\ \text{cm}^2\) will register in a week about 300 particles with energy \(10^{13}\) eV.

Another question connected with the investigation of particles of ultrahigh energies is whether, at energies \(10^{12}\div 10^{14}\) eV, the ratio between the fluxes of protons and of heavier nuclei observed in the region of moderate energies is preserved. At present the composition of cosmic rays in the region of ultrahigh energies is completely unknown. Therefore it is unclear whether extensive showers with very large energy release are produced by protons or by nuclei. It will apparently be possible to answer this important question (see, in particular, \(^{1}\)) only by carrying out the corresponding experiment on an artificial Earth satellite. Even in the case where the main fraction of extensive showers is produced by protons, and the flux of nuclei with energy \(E > 10^{12}\) eV is small, one may hope that on a satellite it will be possible to collect sufficiently rich statistical material for determining the magnitude of this flux. In this case, for registering nuclei with \(E > 10^{12}\) eV, one can use a combination of a Cherenkov counter, by means of which the charge of the particle will be determined, and the device described above, which makes it possible to estimate the particle energy.

Cited Literature

1. V. L. Ginzburg, UFN 62, issue 2, 37 (1956).
2. H. E. Suess, H. C. Urey, Rev. Mod. Phys. 28, 53 (1956).
3. B. Peters, in the collection Physics of Cosmic Rays, ed. by J. Wilson, Foreign Literature Publishing House, Moscow, 1954, p. 155.
4. M. I. Fradkin, UFN 53, 305 (1954).
5. G. J. Waddington, Nuovo Cim. 3, 930 (1956).
6. R. E. Danielson, P. S. Freier, J. E. Naugle, E. P. Ney, Phys. Rev. 103, 1075 (1956).
7. J. A. Simpson, K. B. Fenton, I. Katzman, D. C. Rose, Phys. Rev. 102, 1648 (1956).
8. S. F. Singer, Phys. Rev. 80, 47 (1950).

1. M. A. Pomerantz, J. Franklin Inst. 258, 443 (1954).
2. G. W. McClure, Phys. Rev. 96, 1691 (1954).
3. M. A. Pomerantz, Phys. Rev. 95, 1691 (1954).
4. L. Goldfarb, H. L. Bradt, B. Peters, Phys. Rev. 77, 751 (1950).
5. B. Peters, Proc. Indian Acad. Sci. A40, 230 (1954).
6. G. J. Perlow, L. R. Davis, C. W. Kissinger, J. D. Shipman, Phys. Rev. 88, 321 (1952).
7. A. De Marco, A. Milone, M. Reinharz, Nuovo Cim. 3, 1150 (1956).
8. N. Horwitz, Phys. Rev. 98, 165 (1955).
9. J. Linsley, Phys. Rev. 101, 826 (1956).
10. W. R. Webber, F. B. McDonald, Phys. Rev. 100, 1460 (1955).
11. F. B. McDonald, Phys. Rev. 104, 1723 (1956).
12. W. R. Webber, Nuovo Cim. 4, 1285 (1956).
13. H. L. Bradt, B. Peters, Phys. Rev. 77, 54 (1950).
14. J. H. Noon, A. J. Herz, B. J. O’Brien, Nature 179, 91 (1957).
15. E. P. Ney, D. M. Thon, Phys. Rev. 81, 1068 (1951).
16. L. R. Davis, H. M. Caulk, C. Y. Johnson, Phys. Rev. 101, 800 (1956).
17. J. Linsley, Phys. Rev. 97, 1292 (1955).
18. C. J. Waddington, Phil. Mag. 45, 1312 (1954).
19. K. Gottstein, Phil. Mag. 45, 347 (1954).
20. D. Lal, Y. Pal, M. F. Kaplon, B. Peters, Phys. Rev. 86, 569 (1952).
21. H. J. Taylor, M. Sitaramaswami, P. N. Krishnamoorthy, Proc. Indian Acad. Sci. A36, 41 (1952).
22. R. E. Danielson, P. S. Freier, J. S. Naugle, E. P. Ney, Phys. Rev. 96, 829 (1954).
23. R. F. Hourd, J. R. Fleming, J. J. Lord, Phys. Rev. 95, 647 (1954).
24. P. S. Freier, G. W. Anderson, J. E. Naugle, E. P. Ney, Phys. Rev. 84, 322 (1951).
25. H. Fay, Zs. f. Naturf. 10a, 572 (1955).
26. T. H. Stix, Phys. Rev. 95, 782 (1954).
27. H. Yagoda, Phys. Rev. 99, 1644 (1955).
28. M. F. Kaplon, J. H. Noon, G. W. Racette, Phys. Rev. 96, 1408 (1954).
29. M. F. Kaplon, B. Peters, H. L. Reynolds, D. M. Ritson, Phys. Rev. 85, 295 (1952).
30. A. D. Dainton, P. H. Fowler, D. W. Kent, Phil. Mag. 43, 729 (1952).
31. G. W. Anderson, P. S. Freier, J. E. Naugle, Phys. Rev. 94, 1317 (1954).
32. M. F. Kaplon, G. W. Racette, D. M. Ritson, Phys. Rev. 93, 914 (1954).
33. S. F. Singer, Bull. Amer. Phys. Soc. 2, No. 1, 53 (1957).
34. J. J. Lord, M. Schein, Phys. Rev. 78, 484 (1950); Phys. Rev. 80, 304 (1950).
35. P. S. Freier, E. P. Ney, J. E. Naugle, G. W. Anderson, Phys. Rev. 79, 206 (1950).
36. G. W. McClure, M. A. Pomerantz, Phys. Rev. 84, 1252 (1951).
37. V. H. Yngve, Phys. Rev. 92, 428 (1953).
38. Problems of Modern Physics, ser. 5, issue 7, IL, Moscow, 1953.
39. V. L. Granovskii, Electrical Fluctuations, Moscow—Leningrad, ONTI, 1936.
40. O. N. Vavilov, Comptes Rendus (Doklady) d’Acad. Sci. USSR 33, 3 (1941).