Recent Studies of Cosmic Rays*

A. H. Compton, Chicago, USA

In order properly to illuminate the results of the most recent studies of cosmic rays, let us very briefly recall their earlier history. It is well known that at the beginning of the present century Wilson[^1], and also Elster and Geitel[^2], established that ordinary air is weakly ionized, and that later McLennan and Burton[^3] in Toronto and Rutherford and Cook[^4] in Montreal, using absorbing screens, showed that a noticeable part of this ionization is caused by a penetrating radiation entering the ionization chamber from outside. It was assumed that this radiation consisted of γ-rays coming from radioactive substances which, as is known, are found in the soil and in the air. However, Wulf’s observations[^5], made around 1910 on high towers, as well as a series of balloon ascents by Gockel[^6], showed that the intensity of these ionizing rays decreases with height more slowly than could have been explained. These surprising results led to further balloon observations carried out by Victor Hess[^7][^8] and W. Kolhörster[^9]. These investigators established that the intensity of the penetrating radiation does indeed increase with increasing altitude, which of course would not have been observed if the source of the radiation were the soil. From these investigations Hess[^8] in 1912 drew the bold conclusion that the penetrating radiation enters our atmosphere from outside, from a source uniform in all directions. All new investigations confirm this interpretation of penetrating radiation.

During the war years, and also for several years afterward, little attention was paid to these rays, which had been discovered not long before. Some, including Hoffmann[^10] and Millikan,[^11] doubted their existence. Others, especially Kolhörster[^12], engaged in measuring their absorption and in studying the variation of their intensity with time. In 1925 Millikan, having convinced himself of the real existence of these rays by means of his greatly improved measurements of their absorption in water[^13], advanced the hypo—

* Report read at the joint session of the American Physical Society and Section B of the American Association for the Advancement of Science in St. Louis on January 1, 1936. Published in Review of Scientific Instruments, vol. 7/1, 1936; translated by L. V. Groshev.

thesis 14 on the origin of these rays. His idea was that the main part of the rays consists of photons arising in the recombination, in interstellar space, of protons and electrons into helium nuclei. However, this hypothesis had to be abandoned, since it became clear beyond doubt that the greater part of the rays consists of electrically charged particles possessing energies much larger in magnitude than had been supposed by Millikan. Nevertheless, Millikan’s alluring suggestion of a connection between the origin of cosmic rays and the origin of the universe proved very effective in stimulating an enormous amount of research and left a widely held hope that, if we learn how cosmic rays arise, we shall be able to read in them this ancient history of our universe.

Over the last 10 years, intensive study carried out by many researchers has clarified our knowledge of the properties of these unusual rays. Measurements of absorption have revealed two main components with absorption coefficients of about 0.6 and 0.08 per 1 m of water. The second component has approximately 100 times greater penetrating power in comparison with the hardest $\gamma$-rays. We have found that cosmic rays come to us from far beyond the limits of the earth’s atmosphere and that at great altitudes they have an intensity at the magnetic poles many times greater than near the equator. This gives confidence that they consist chiefly of electrically charged particles. The total amount of heat they bring to the earth is of the same order as the heat brought by the light from the stars. Individual cosmic-ray particles, however, possess unusually large energies of $10^9$ and, possibly, almost up to $10^{12}$ eV. This means 1 erg of energy for a single atomic projectile.

QUESTIONS OF THE PRESENT-DAY STUDY OF COSMIC RAYS

The present-day study of cosmic rays concerns, chiefly, two questions: the properties of these rays and their application in other investigations, as a means of action. Among the important properties of the rays, the most immediate interest is their composition, i.e. the nature of the particles of which they consist, and the energies these particles possess. What effects should the rays produce when they pass through matter? Where do they originate, and how do they arise? Among the applications of cosmic rays let us note their application to nuclear physics, where in Anderson’s hands they led to the discovery of positrons. From the study of their geographical distribution we broaden our knowledge of the earth’s magnetic field high above the atmosphere. Electrodynamics has been put to the test in an energy region hitherto inaccessible. In astronomy, cosmic rays seem as though they may give us a powerful means for studying the rotation of the galaxy and the early history of our universe. In biology it is not impossible that they

play an important role in the spontaneous variations on which evolutionary changes depend.

COMPOSITION OF COSMIC RAYS

By analogy with γ-rays, the exceptionally great penetrating power of cosmic rays was at first considered an indication that they are photons. However, in 1929 the remarkable experiments of Bothe and Kolhörster[^15], carried out with counters operating in coincidence, gave strong evidence that the primary cosmic rays consist of electrically charged particles. These investigators drew attention to the fact that such particles, on approaching the earth, should be deflected by the earth’s magnetic field in such a way that they would reach the poles much more readily than the equator. This supposition led to a series of detailed investigations which gave us, as it were, direct evidence that the primary cosmic rays consist of electrically charged particles. I wish briefly to examine this evidence and then to indicate how subsequent experiments made it possible to carry out a tentative analysis of these electric rays into components that were identified with ordinary atomic particles.

Clay[^16] had by that time only just published his first measurements, which showed that the intensity of cosmic rays in Holland is greater than in Java. Bothe and Kolhörster attributed this difference in intensities to the expected action of the earth’s magnetic field. At first, the expeditions of Bothe and Kolhörster[^17], Millikan and Cameron[^18], Kerr Grant[^19], and others did not confirm the phenomenon repeatedly found by Clay[^20]; therefore it was generally accepted[^21] that no latitude effect existed. Later, however, extensive series of observations were made which confirmed Clay’s results and showed that the variation of intensity with geographical latitude is connected with the earth’s magnetic field just as Bothe and Kolhörster had predicted.

From 1931 to 1934 we sent out from Chicago 12 different expeditions, involving about 80 physicists; measurements were made at more than 1,000 sites widely distributed over the earth’s surface[^22]. The results obtained showed that, for magnetic latitudes greater than 50°, no dependence on latitude is observed at sea level. However, from the equator to a latitude of 50° there is, at sea level, an increase in intensity of approximately 16%. Similar contemporary measurements made by many different authors led essentially to the same results[^23]. Figure 1 gives a summary of the published observations carried out by various expeditions studying the geographical distribution of cosmic rays at sea level. The data are presented in the form of curves (they may be called isocosms) of equal intensity of cosmic rays; here all the results of the different investigators are reduced to one and the same scale.

Fig. 1. Curves of equal cosmic-ray intensity (isocosms), showing approximate parallelism with the parallels of geomagnetic latitude and with the curves of equal occurrence of auroras.

Solid dots represent the locations (at sea level) of observers of our Chicago expeditions; circles are plotted according to reports by other observers, among whom may be mentioned in particular Clay, Corlin, Millikan and Neher, Prins and their collaborators. Whereas Millikan and Neher found a mean latitude effect equal to 10%, in contrast to the zero effect of which they reported three years ago,^24 all other observers found a decrease, close to 14%, observed on moving from the pole to the equator. There are therefore some difficulties in quantitatively reconciling the results of Millikan and Neher with the data of other authors. However, the qualitative agreement among the various authors may be regarded as excellent.

Fig. 2. Increase of the latitude effect with altitude according to observations made in the mountains.

The isocosms plotted in Fig. 1 are close in their curvature to parallels of geomagnetic latitude, thereby emphasizing the dependence of the phenomenon on the terrestrial magnetic field. The parallelism between the isocosms and the lines of equal number of aurorae is manifested even more closely, as is shown by the dotted lines taken from Fritz’s map. This unusual similarity must mean that the earth’s magnetic field acts in the same way on aurorae and on cosmic rays. From the cited map there also follows a smaller equatorial intensity of cosmic rays for the eastern hemisphere than for the western, in accordance with the stronger field of the earth in the east. A similar longitude effect was found independently by Clay^23 and by Millikan and Neher^23. These details leave no doubt that the latitude effect is caused by the action of the earth’s magnetic field. The existence of the effect thus shows that, at the very least, a noticeable part of the primary cosmic rays consists of electrically charged particles.

A distinctive feature of the latitude effect is that it increases rapidly with altitude. This became clear from our high-mountain measurements carried out at various latitudes^22 (Fig. 2). The corresponding data were extended to great altitudes by the measurements of Bowen, Millikan and Neher^25 and Clay, made in an airplane, and especially by Regener’s observations^26,

Piccard and Cosyns[^27], Clay[^28], as well as Compton, Stephenson, and Millikan[^29] in the stratosphere. Some of these data are given in Fig. 3. The following fact is striking. Whereas at sea level the latitude effect amounts to \(15—20\%\), at the upper boundary of the atmosphere the intensity observed at geomagnetic latitude \(52^\circ\) is twice as great as the intensity at latitude \(42^\circ\), evidently

Fig. 3. Altitude measurements of cosmic rays as a function of depth, reckoned from the upper boundary of the atmosphere. \(52^\circ\)—Compton, Stephenson, Millikan; \(42^\circ\)—Millikan—Bowen; Peru—Millikan, Neher; Java—Clay.

by \(5—10\) times greater than that found in Peru and approximately 40 times exceeds the intensity detected by Clay near the equator in Java. This sharp change with latitude was observed directly by Cosyns[^30], when he was carried southward in Piccard’s balloon. At the upper boundary of the atmosphere the ratio of the intensities at the pole and at the equator is, of course, greater than all that is indicated by these experiments; it is quite possible that it is of the order of 100.

Since the earth’s magnetic field should not act on electrically neutral rays, this result at once shows

that the ionization produced by such rays at the boundary of the atmosphere near the poles probably amounts to no more than a few percent of the total ionization.

It should be noted that the rays reaching the earth’s surface at the equator are absorbed essentially in the same way as rays deflected by the earth’s magnetic field and therefore consisting of charged particles. In Fig. 4, \(A\) gives the curve of absorption in the atmosphere for cosmic rays observed at high magnetic latitudes; curve \(B\) gives the same for equatorial rays[^31]. Curve \(C\) gives the difference between \(A\) and \(B\) and, consequently, shows the absorption

Fig. 4. Curve \(C\), indicating the absorption of electrically charged particles, has the same form as curve \(B\), which refers to cosmic rays reaching the earth at the equator.

of electrically charged particles on which the earth’s magnetic field acts in such a way that they cannot reach the earth’s surface at the equator. The great similarity of curve \(B\), which refers to rays admitted by the magnetic field, and curve \(C\), solely for rays of charged particles, suggests that both kinds of rays are of one and the same nature.

Still more definitely, Johnson and Alvarez and others have established that near the equator cosmic rays show an east–west asymmetry[^32], caused by the action of the earth’s magnetic field. This means that, at the very least, an appreciable part even of those rays which reach the equator through the earth’s magnetic barrier consists of electrically charged particles, though of very high energies.

Further experiments with cosmic rays, which were an extension of the experiments of Bothe and Kolhörster (the coincidence method), made it possible

to clarify in greater detail the nature of cosmic rays. The investigations of Rossi^33 and Siuŋg^34 evidently lead to the inevitable conclusion that the most penetrating part of cosmic rays at sea level consists of very fast electrically charged particles. A vivid proof of this fact was obtained recently, independently, by Auger and Ehrenfest^35, Street, Woodward, and Stevenson^36. Fig. 5 shows schematically an apparatus with a Wilson chamber controlled by counters. In this apparatus, used in the work of the last-named investigators, it was possible to place 45 cm of lead between the upper counters. The Wilson chamber itself was placed between the lower counters. In 90% of the photographs taken with the chamber, tracks of single particles of very high energy are found.

Fig. 5. Arrangement of coincidence counters \(C_1, C_2, C_3\) and the Wilson chamber \(Ch\) for photographing tracks penetrating cosmic rays (Street, Woodward, Stevenson).

An example of this may be Fig. 6. From this it is perfectly clear that such coincidences are caused by primary ionizing particles and, consequently, particles carrying an electric charge and possessing very great penetrating power. If there were no

Fig. 6. Tracks of two particles of cosmic rays that have passed through 40 cm of lead.

latitude effect, it would be possible to suppose that these penetrating particles are secondary, formed in the upper parts of the atmosphere from easily absorbed photons of high energy. Since, however, the theory of the latitude effect shows that electrons, in order to pass through the terrestrial magnetic barrier, must have energies of the order observed in the experiments described above, it therefore becomes evident that the particles observed in these experiments do indeed belong to the primary cosmic rays. The presence of east–west and north–south asymmetry^37 in the distribution of cosmic rays, the approximate independence of the transition effect from latitude^31

and some other^38 facts are an additional confirmation that all components of cosmic rays observed at the present time consist of primary electrically charged particles. However, further work in this direction is required. If there are neutral particles in the primary rays, such as photons or neutrons, then they can constitute no more than a few percent of the number of particles entering our atmosphere; at sea level they should produce a very small ionization in comparison with the ionization created by the electrically charged particles present in much larger numbers.

ANALYSIS OF PRIMARY COSMIC RAYS

Our problem reduces to establishing the nature of the various components of the electrically charged particles that make up the cosmic rays falling upon our earth. For particles of low energies such an analysis is usually carried out with the aid of a mass spectrograph, in which electric and magnetic fields are employed. Recently, attempts have been made, with some success, to deflect cosmic rays in ordinary electric and magnetic fields. However, the energies of the primary particles are so great that they are difficult to deflect, and even if this does occur, it is very difficult to distinguish between primary cosmic particles and secondary ones created within the atmosphere. Fortunately, nature has provided us with a ready-made magnetic spectrograph suitable for the analysis of primary cosmic rays. The earth itself acts as a magnet, and instead of an electric field we have the retarding action of the earth’s atmosphere. A great advantage of this natural instrument is the fact that it analyzes the rays far above the atmosphere, where they are free of secondary rays. True, it leaves much to be desired with respect to the uniformity of the magnetic field. Moreover, we are not yet in a position to study accurately the calibration curve by means of which one could determine the energy of the particles, expressed in terms of their penetrating power in the atmosphere. Despite these limitations and the incompleteness of our knowledge, attempts to analyze cosmic rays by means of the earth’s magnetic field lead to valuable results and indicate the character of the data that should be obtained in a more rigorous analysis.^39

Theoretical investigations by Størmer,^40 Lemaitre and Vallarta^41 and others show that, for charged particles approaching the earth from remote regions of space, for a given magnetic latitude there exists, roughly speaking, a critical energy for rays of a given kind. This means that particles of these rays with energies below the critical value are bent and pass away from the earth, whereas particles with higher energies freely reach the earth. Corresponding to this critical energy there must exist

minimum path in air that the particles can have at the given latitude.

Our method of magnetic analysis consists in comparing the minimum paths calculated for particles of various kinds with the experimental minimum paths determined from measurements of the intensity of cosmic rays at different altitudes. It has been found that the calculated minimum paths for $\alpha$-particles, electrons, and protons correspond to the minimum paths of three different groups of cosmic rays. Although this comparison at present cannot be carried out with sufficient accuracy, nevertheless the results show that, under known circumstances, this method could give us a complete and reliable analysis of primary cosmic rays. In Fig. 7 data from various balloon flights are presented; they reveal the presence of a minimum path for rays passing through the atmosphere.

Fig. 7. Comparison of the intensity of cosmic rays for the vertical direction. The humps on the curves indicate the presence of a minimum path.

1. Compton—Stephenson—Millikan. 2. Bowen—Millikan—Neher. 3. Regener—Gross. 4. Piccard—Cosyns. 5. Kolhörster. 6. Bowen—Millikan. 7. Clay. 8. Compton—Stephenson.

The quantity $\psi$ plotted along the ordinates is not the directly measured intensity of cosmic rays arriving from all directions, but rather the intensity of the component that passes through the atmosphere in the vertical direction. This component is calculated from the measured total intensity with the aid of

formula proposed by Gross. Recently Regener\(^{42}\) confirmed Gross’s analysis by making a direct measurement of the vertical component with the aid of counters operating in coincidence. It should be noted that, for one and the same effective magnetic latitude \(\mu\), the curves obtained by different authors are in good agreement with one another.

The horizontal portions of the curves shown indicate a minimum range; if there existed no particles with ranges smaller than this limit, the ionization would not increase with a further increase in altitude. The figure also indicates the calculated minimum ranges for electrons of \(\alpha\)-particles capable of penetrating through the earth’s magnetic field at these latitudes. The calculated minimum range for protons is too large to fit on the diagram shown. All these calculations can be carried out only approximately, chiefly because we do not know exactly the loss of energy by the various particles in passing through matter. It will be seen later, however, that a comparison of the minimum range found with that predicted indicates that the least penetrating group of cosmic rays consists of \(\alpha\)-particles, while the more penetrating group consists of electrons. Observations of the latitude effect at sea level give analogous evidence for the existence of an still more penetrating group, which has been identified with protons.

A detailed discussion of this powerful method of analysis has been given by me elsewhere\(^{38}\). Time, however, does not permit us to go further into its development. On the basis of data obtained in our own and also in other experiments, we can arrive at the following preliminary analysis. The principal part of the cosmic rays observed at sea level consists approximately of equal numbers of positive and negative electrons. At sea level and below there is a very penetrating component which, apparently, must be taken to consist of protons; however, there are certain difficulties in this latter conclusion. At very great altitudes there is found a relatively strongly absorbed component which, apparently, consists of \(\alpha\)-particles.

The existence of three separate components of cosmic rays is proved by measurements of the intensity of the rays at various depths measured from the boundary of the atmosphere. Fig. 7 shows the presence of two components, which we shall call \(A\) and \(B\). Component \(A\) plays an important role only at very great altitudes. Component \(B\) is the component that we have identified as consisting of electrons. The data for depths lying below sea level are best summarized by Fig. 8, where are shown the results of Eckart’s detailed analysis of the absorption coefficients of cosmic rays for an altitude of 7000 ft., on the basis of the measurements of Regener, Millikan, and Benade for great depths in water. Our component \(A\) is not manifested at such small altitudes, but component \(B\) may be identified with the soft component obtained by Eckart,

and possessing a mean absorption coefficient of 0.6 per 1 m of water. Its more penetrating component, with an absorption coefficient of approximately 0.08, according to our preliminary analysis, consists of protons.

Additional data confirming this analysis can be obtained from the observations of Anderson and others43, which showed that the high-energy particles observed in Wilson chambers at sea level and at Pike’s Peak consist of approximately equal numbers of positively and negatively charged particles, with a probable excess of positive ones (at sea level). These measurements dealt mainly with the component which we have denoted by \(B\), and therefore they confirm the fact that it consists of positrons and electrons. The experiments of Johnson, Alvarez, Rossi* and others, in which the distribution of cosmic rays by direction was investigated, showed that near the equator the particles producing coincidences come chiefly from the west, revealing a certain predominance in favor of positive particles, which in all probability are protons; at the same time, for the primary particles producing showers there is a symmetric distribution, although a latitude effect is observed for them as well; they can probably be identified with electrons and positrons occurring in equal quantities. The fact established by Rossi38 and Johnson37, that the relative number of showers increases with altitude, supports the identification of the most penetrating component with protons. However, the impossibility of identifying protons of very high energies in investigations with the Wilson chamber leaves some doubt as to the composition of this most penetrating part of cosmic rays.

Fig. 8. Spectrum of the absorption coefficient of cosmic rays at an altitude of \(\sim 2000\) m (after Eckart).

* See in particular footnote38.

There can hardly be any doubt that continuing the investigation in the same direction will make it possible to carry out a complete and accurate analysis of the composition of cosmic rays.

DISTRIBUTION OF COSMIC RAYS BY ENERGY

There are several methods for measuring the energy of cosmic-ray particles; the most direct of them consists in bending the paths of the particles in the magnetic field of an electromagnet or of the earth. The most fruitful experiments for determining the energies of particles from their deflection in the magnetic field of an electromagnet were carried out by Anderson. In his apparatus a Wilson chamber was placed between the poles of a magnet, put into operation by counters according to the Blekett and Occhialini scheme. Fig. 9 shows the distribution curve obtained by Anderson, by energies, of cosmic-ray particles observed at sea level[^44]. Many particles having energies less than \(10^9\) eV are secondary; for higher energies most of them are primary. For energies greater than \(5 \cdot 10^9\) eV, however, this method is unreliable because of the small curvature of the tracks. Of course, the curve shown gives the energies of the particles after they have passed through the atmosphere.

Fig. 9. Distribution by energy of the total number of cosmic-ray particles observed at sea level (Anderson).

The study of the latitude effect gives us a magnetic analysis of the distribution of particles by energy, free from the question of whether the particles are primary or secondary. A detailed distribution of particles by energy obtained by this method has not yet been published. However, the existence of the latitude effect at high altitudes up to \(55^\circ\) means that there are primary particles with energies of about \(2 \cdot 10^9\) eV (for electrons). At higher latitudes there probably exist primary particles with still lower energy. On the other hand, an electron reaching the equator in the vertical direction must have an energy of about \(20 \cdot 10^9\) V. Investigations of the distribution of particles by direction show that at the equator many of them possess still greater energies. From such observations one can state with confidence that primary cosmic-ray particles are encountered with energies from \(2 \cdot 10^9\) eV up to more than \(60 \cdot 10^9\) eV.

An estimate of the energy of particles from their penetration to great depths and from the energy released in the so-called coll-

as Hoffmann’s, is less reliable; nevertheless, in order of magnitude it agrees with the energies found from the investigation of the latitude effect. These methods show that cosmic particles with energies greater than \(600 \cdot 10^9\) eV sometimes occur.^45

EFFECTS CAUSED BY COSMIC RAYS

Two different kinds of effects are known that are caused by the direct action of the primary particles of cosmic rays. These are: 1) direct ionization of atoms in collisions with electrons, as occurs in the case of \(\beta\)-rays of ordinary energies, 2) nuclear collisions accompanied by the emission of photons, i.e. the formation of \(\gamma\)-rays. In Fig. 6 we saw an example of ionization of the first kind. The proof of the formation of photons from primary rays is not so direct. Fig. 10 shows Anderson’s photograph of a shower of high-energy electrons and positrons produced in lead by a non-ionizing particle, probably a photon. From such observations one may conclude that showers are produced directly upon the absorption of photons. However, in similar experiments it became clear that the photons themselves arise in the immediate vicinity of the apparatus and therefore are themselves secondary rays. Johnson’s experiments,^37 performed by the coincidence method, showed that a latitude effect is also observed for showers. This must mean that they are produced by primary electrically charged particles of cosmic rays.

Fig. 10. Shower of secondary particles of cosmic rays, produced by a non-ionizing particle, probably a photon (Anderson).

On this basis, Geiger and Fünfer^46 established five gradations in cosmic rays: \(A\)—a primary electron or positron arriving from outer space; \(B\)—photons excited

primary electron; \(C\)—electrons and positrons produced by the absorption of these photons; \(D\)—photons produced in turn by these electrons; and, finally, \(E\)—electron pairs arising from such photons. This complicated scheme of secondary rays is apparently confirmed by Clay’s experiments[^47] with counters, as well as by photographs obtained with a Wilson chamber. Thus an electron with an energy of the order of \(10^{11}\,\mathrm{eV}\) gives rise to many electrons with energies of the order of \(10^7\,\mathrm{eV}\). For energies less than \(10^7\,\mathrm{eV}\), the loss of energy directly in collisions of electrons with atoms is more probable than the production of photons by an electron; for this reason secondary rays with small energies are encountered comparatively rarely. It now seems probable that the Hoffmann bursts observed in cosmic rays are nothing other than this fragmentation of the particle’s energy, occurring chiefly in the gas of the ionization chamber.

From what is known about the behavior of low-energy rays, the fact that for these large energies of secondary cosmic rays a photon is absorbed much more readily than an electron of the very same energy must seem surprising.

According to the new theory of Bethe and Heitler[^48] and Oppenheimer[^49], we should expect that the greater part of the energy of very fast electrons is spent on the excitation of bremsstrahlung radiation. For protons, on the other hand, the part lost in their collisions with electrons should predominate. In agreement with this theory, the experiments of Rossi[^38] and Johnson[^37], carried out at various altitudes by the coincidence method, indicate the existence of two kinds of primary particles, some of which are more effective in creating showers than others. Johnson’s experiments[^37], carried out to determine the spatial distribution of cosmic rays (with the aid of counters), also confirm the view that the primary particles producing showers consist in equal numbers of positrons and electrons, whereas the particles that do not create showers carry a positive charge and, in all probability, are protons. However, a test of this theory of the production of bremsstrahlung radiation by fast particles shows the presence of a substantial discrepancy between the experimental data and the theoretical data given by quantum electrodynamics. Up to energies of \(70\,\mathrm{MeV}\), Anderson’s investigations of the energy loss of electrons passing through matter reveal good agreement of the data obtained with the theory of Bethe and Heitler (Fig. 11).

However, for greater energies, where the wavelength corresponding to the electron becomes smaller than the electron radius of classical theory, the energy loss is much less than this theory predicts.[^50] In this region of very large energies a new modification of electrodynamics is required, similar to Lorentz–Einstein’s modification, which extended electrodynamics into the region of large velocities. If someday such an improved-

... theory is created, cosmic rays will give us an instrument—and, apparently, the only one—suitable for testing this theory.

WHERE DO COSMIC RAYS ORIGINATE?

The existence of the latitude effect shows that cosmic rays originate far beyond the limits of the earth’s atmosphere. The earth’s magnetic field is not large enough to bend appreciably any radiation arising within the atmosphere before it is slowed down by collisions with molecules. Moreover, as Blekett[^51] noted, if cosmic rays arose in the earth’s atmosphere, then, owing to the bending action of the earth’s magnetic field on particles traveling outward, the greater intensity should be observed at the equator. This is the opposite of what is observed in the latitude effect.

It was found that, if one excludes the deflecting action of the earth’s field, cosmic rays fall upon the earth almost uniformly from all directions. Outside the earth’s atmosphere it is not possible to find within our galaxy an isotropic distribution of matter where cosmic rays could originate. On the other hand, extragalactic nebulae, or space itself, could satisfy the condition of spherical symmetry. Calculations by both Eddington and Lemaitre showed that the probable absorption of cosmic rays in their passage through interstellar space at approximately the speed of light for \(10^{10}\) years could be quite neglected. However, if these rays underwent the same redward shift as is observed for the light of distant nebulae, then rays formed at distances of the order of \(10^{10}\) light-years would reach the earth with only a small fraction of their initial energy. Consequently, if the rays arise continuously, their isotropic distribution indicates that the greater part of them appears in distant galaxies, or in distant space, located at distances between \(10^9\)—\(10^{10}\) light-years. It could be assumed, as Lemaitre did, that they were formed at the moment when the expansion of the universe began, and since then have been traveling through space.

Fig. 11. Loss of energy (referred to \(1\ \mathrm{cm}\) of fast electrons) in passing through a layer of lead (experimental data according to Anderson, theoretical data according to Bethe and Heitler).

Some confirmation of this point of view regarding the origin of the rays at great distances from the earth is furnished by the fact that, apparently, the intensity of cosmic rays depends on the rotation of the galaxy[^52].

According to Oort and other astronomers, this rotation carries us in the direction of approximately \(47^\circ\) north latitude and 20 h. 55 min. right ascension, with a speed of about 300 km per second—\(1/1000\) the speed of light. If the source of cosmic rays lies beyond our galaxy and is at rest relative to its center of gravity, then, as calculation shows, at our latitude this motion should produce a diurnal variation (in sidereal time) in the intensity, whose magnitude is of the order of \(0.1\%\).

Fig. 12. Change in the intensity of cosmic rays with sidereal time (1932). Data of Tesse and Steinmaurer; theory, assuming the rotation of the galaxy,—Compton, Getting.

The most suitable record of cosmic-ray intensity, obtained by Tesse and Steinmaurer for 1932 (Fig. 12), shows the presence of a variation of intensity (with change in sidereal time) of the expected magnitude and with a maximum located very close to the predicted time. Although further results are needed before other possible interpretations of this variation of intensity with sidereal time are excluded, nevertheless the complete agreement with the prediction may justify the assumption that this effect is indeed caused by the rotation of the galaxy. This should lead to the conclusion that a significant part of the rays originates beyond the Milky Way, thereby justifying the name “cosmic,” which formerly had a purely heuristic meaning.

HOW DO COSMIC RAYS ARISE?

Of the large number of hypotheses concerning the origin of cosmic rays, not one has received sufficient experimental confirmation for its general acceptance. Hypotheses asserting that cosmic rays consist of photons are in clear contradiction with the observed latitude effect. Hypotheses attributing the origin of cosmic rays to nuclear transformations proceeding with a decrease in their mass are unable to explain the presence of the enormous energies, from \(10^{10}\) to \(10^{12}\) eV, which, according to the most recent investigations, cosmic rays possess. Swann imagines that electrons are accelerated by means of electromagnetic induction caused by changing magnetic fields of the “sunspots” of giant stars. Millikan supposes that the particles acquire their energy at the expense of the gravitational attraction of the universe. Lemaitre’s hypothesis consists in the idea that “super-radioactive particles” were created in the initial explosion of his expanding universe. At present we cannot subject these assumptions to experimental verification.

Time does not permit us to discuss questions concerning the use of cosmic rays as an instrument for the study of other phenomena. It is enough to say that, by virtue of their enormous energy, cosmic rays occupy in the atomic artillery of physicists a place that has no equal.

Our analysis of the composition of cosmic rays is moving forward. Their cosmic place of origin, although possibly not established, nevertheless now seems more definite than before. Their origin is still obscure; however, the increasing knowledge of their characteristics makes it possible to restrict the types of admissible hypotheses. Used as an active agent, cosmic rays have made it possible to discover the positron; they may serve as a supplement to the telescope for collecting astronomical data; they provide means for expanding our knowledge of the laws of electricity and the properties of matter for energies a thousand times greater than those available to us from any other known sources.

LITERATURE

1. C. T. R. Wilson, Proc. Camb. Phil. Soc., 11, 520, 1900; Proc. Roy. Soc., A 68, 151; A 69, 277, 1901.
2. H. Geitel, Phys. Z., 2, 116, 1900—1901; J. Elster a. H. Geitel, ibid., 2, 560, 1900—1901.
3. I. C. McClennan u. E. Burton, Physik. Z., 4, 533, 1902—1903; Phys. Rev., 16, 184, 1903.
4. E. Rutherford a. H. L. Cooke, Phys. Rev., 16, 183, 1903.
5. T. Wulf, Physik. Z., 10, 152, 1909, 11, 811, 1910.
6. A. Gockel, Physik. Z., 10, 152, 1909; 12, 595, 1911.

1. V. F. Hess, Wien. Ber., 120, 1575; Physik. Z., 12, 998, 1911; Wien. Ber., 122, 1053, 1481, 1913.

2. V. F. Hess, Wien. Ber., 121, 2001, 1912.

3. W. Kolhörster, Verh. d. Phys. Ges., 15, 1111, 1913; 16, 719, 1914.

4. G. Hoffmann, Physik. Z., 26, 40, 669, 1925.

5. R. A. Millikan a. R. M. Otis, Phys. Rev., 23, 778, 1924.

6. W. Kolhörster, Berlin. Ber., No 34, 1932; No 7, 1925.

7. R. A. Millikan, Proc. Nat. Acad. Sci, 18, 38, 1926; R. A. Millikan a. G. H. Cameron, Phys. Rev., 28, 851, 1926.

8. See reference¹³ and Phys. Rev., 31, 921; 32, 533, 1928.

9. W. Bothe u. W. Kolhörster, Z. Physik, 56, 751, 1929.

10. J. Clay, Proc. Acad. Sci. Amsterdam, 30, 1115, 1927; 32, 1091, 1928.

11. W. Bothe u. W. Kolhörster, Berlin Ber., No 26, 450, 1930.

12. R. A. Millikan a. G. H. Cameron, Phys. Rev., 31, 163, 1928; R. A. Millikan, ibid., 36, 1595, 1930.

13. Kerr Grant, Nature, 127, 924, 1931.

14. J. Clay, Proc. Acad. Sic. Amsterdam, 33, 711, 1930.

15. G. Hoffmann, Physik. Z., 32, 633, 1932.

16. A. H. Compton, Phys. Rev., 41, 111, 681, 1932; 43, 387, 1933; Trans. Amer. Geophys. Union p. 154, 1933.

17. J. Clay a. H. P. Berlage, Naturwiss., 20, 687, 1932; J. Clay. Physica, 1, 363, 829, 1934; H. Hoerlin, Nature, 132, 61, 1933; R. A. Millikan, Phys. Rev., 43, 661, 1933; R. A. Millikan a. H. V. Neher, Phys. Rev., 47, 205, 1935; J. Prins, Nature, 132, 781, 1932.

18. See R. A. Millikan, New-York Times, Dec. 31, 1932, p. 6.

19. I. S. Bowen, R. A. Millikan a. H. V. Neher, Phys. Rev., 44, 248, 1933.

20. E. Regener, Physik. Z., 34, 306, 1933.

21. A. Piccard a. M. Cosyns, C. R., 195, 605, 1932.

22. J. Clay, Physica, 1, 363, 1934.

23. A. H. Compton a. R. J. Stephenson, Phys. Rev., 45, 441, 1934; I. S. Bowen, R. A. Millikan a. V. Neher, Phys. Rev., 46, 641, 1934.

24. M. Cosyns, Nature, 135, 313, 1935.

25. A. H. Compton a. R. J. Stephenson, Phys. Rev., 45, 448, 1934.

26. T. H. Johnson a. J. C. Street, Phys. Rev., 43, 381; T. H. Johnson, ibid., 43, 834; 44, 856, 1933; L. Alvarez a. A. H. Compton, ibid., 43, 835, 1930; B. Rossi a. S. de Benedetti, ibid., 45, 214, 1934.

27. B. Rossi, Z. physik. 68, 64, 1931; 82, 151, 1933.

28. D. S. Hsiung, Phys. Rev., 46, 653, 1934.

29. P. Auger a. Ehrenfest, C. R., 199, 1609, 1934.

30. U. C. Street, R. H. Woodward a. E. C. Stevenson, Phys. Rev., 47, 891, 1935.

31. See also T. H. Johnson, Phys Rev., 47, 318, 1935.

32. B. Rossi, International Conference on Physics, I Nuclear Physics, (London 1935), p. 233.

33. A. H. Compton a. R. I. Stephens, Phys. Rev., 45, 441, 1934, A. H. Compton, Proc. Phys. Soc. London, 47, 717, 1935.

34. C. Störmer, Z. Astrophys., 1, 237, 1930; Oslo. Abos. Rub. N 10, 1934.

35. G. Lemaitre a. M. S. Vallarta, Phys. Rev., 43, 87, 1933.

36. E. Regener, Nature, 136, 718, 1935.

37. C. D. Anderson, Proc. Amer. Phys. Soc. Jan. 2, 1936.

38. C. D. Anderson, Intern. Conf. on Physics, I, Nuclear Physics (London 1935) p. 171.

39. See also W. Kolhörster, Berlin., Ver., 686, 1933; F. Steinke, Ergeb. d. exakt. Naturwiss., 13, 89, 1934, Compton, Nature, 134, 1006, 1934.

1. H. Geiger and Fünfer, Z. Physik, 93, 543, 1935.
2. J. Clay, Physica, 2, 551, 1935.
3. H. Bethe and W. Heitler, Proc. Roy. Soc., A 146, 83, 1934.
4. I. R. Oppenheimer, Phys. Rev., 47, 44, 1935.
5. See references 49, 50 and A. H. Compton, Proc. Phys. Soc. London, 47, 747, 1935, note 46.
6. P. M. S. Blackett. La Radiation Cosmique. Paris, 1935, p. 10.
7. A. H. Compton and J. A. Getting, Phys. Rev., 47, 817, 1935.