Some Solved and Unsolved Problems of Space Physics1

Lord Rayleigh.

After the first triumphs of spectral analysis, which showed that the lines in the spectra of the sun and the stars arise from elements that also exist on the earth, there still remained individual lines whose explanation encountered difficulties.

Among them belonged, above all, the line of helium—a riddle whose solution followed even during the lifetime of some of the pioneers of spectral analysis. Although in my youth I had the good fortune to be a close witness of this discovery, I shall not dwell on it here. When we read that millions of cubic meters of this gas are being extracted on earth for filling airships, we begin to understand that its discovery has already become the property of history.

Alongside the hypothesis of helium, so brilliantly confirmed by the discovery of this gas on earth, one must place the hypotheses of nebulium, geocoronium, and coronium. The first two problems may also already be considered solved, although the solution turned out to lie quite elsewhere than where the older generation of astrophysicists sought it.

The Spectrum of Nebulae

In the spectra of nebulae we observe lines that are never encountered in terrestrial spectra. What

Still more remarkable, they bear no resemblance whatever to those barely perceptible, as yet unrecognized lines that are observed in complex spectra, for example in the solar spectrum. On the contrary, they shine brightly and strikingly against a dark background and, thanks to their apparent simplicity, make the puzzle more intriguing. In spectral analysis it has been established as a precise and definite rule that all simple spectra belong to light elements. Since, moreover, the indicated lines in the spectra of nebulae are invariably accompanied by the lines of hydrogen and helium, the assumption arose that they too belong to some light elements of the class which, among the elements found on Earth, bears the name of the inert gases. But the matter was complicated by the fact that no one had yet succeeded in obtaining these lines in the laboratory, and the former expedient of assigning them to some unknown element became, with time, unacceptable. The point is that the table of elements had assumed definite limits, within which there was no room left for new light elements. This was one of those cases in which, in scientific investigation, the method of frontal attack proves not to attain the goal. To resolve the question, a more systematic study of spectra was necessary, especially the spectra of light elements.

The key to solving the problem was found thanks to the circumstance that the most important lines of nebular spectra are paired, as follows from their proximity to one another and from the constancy of the ratio of their intensities in different nebulae and in different parts of the same nebula. The study of such pairs or multiplets has more than once proved a convenient point of attack in spectroscopic investigations. It was precisely in this way that Hartley, studying the diffuse triplets of magnesium, first established the constancy of frequency intervals and thus arrived at the idea—then sounding like a paradox—that the addition and subtraction of certain frequencies is a good method for analyzing spectra. In the same way, the investigation of frequency intervals made it possible to decipher such complex spectra as the spectra of manganese and iron.

It turned out that the difference between the frequencies of the two green lines, originally discovered by Huggins and known as \(N_1\) and \(N_2\), is 193 waves per cm. J. S. Bowen, to whom we owe the final resolution of this riddle, sought the same interval in the spectrum of doubly ionized oxygen and found it between the lower levels, which are denoted as \(1^3P_2\) and \(1^3P_1\).

This, of course, is still not sufficient to prove the supposed origin of the mysterious lines. For proof it is necessary also to establish not only the interval between them, but also their position. These lines were attributed to the intercombination of an upper singlet level and two lower levels belonging to a triplet, in which a third line is excluded by the selection rule for inner quantum numbers. To establish the required differences of terms it was necessary to connect the singlet and triplet levels by an intercombination line observed in the spectrum of doubly ionized oxygen under laboratory conditions. This was done by A. Fowler, who, by combining Bowen’s laboratory data with his own, obtained a fairly accurate agreement with the observed position of the paired nebular line. In view of this there remains almost no doubt that other, less well-known spectral lines of nebulae can likewise be attributed to singly ionized nitrogen and singly ionized oxygen.

The identification of these lines was made with some disregard of the rules of quantum theory, which were derived on the basis of empirical material and received some theoretical justification in the works of Bohr and his followers. These rules exclude the possibility of certain lines that might be obtained on the basis of the combination principle. When the state of excitation of an atom is such that it cannot pass directly to a lower level without violating these rules, such a state is called metastable. This is precisely the case we have in the lines of nebulae. I shall return again to the question of metastable states and “forbidden” lines.

Spectrum of the Aurora

The next old cosmic problem on which I wish to dwell is the green line of the spectrum of the aurora. It was first noticed by A. I. Ångström in Uppsala in 1868, and he reported this observation in one of the supplements to his great work, in which the scale of wavelengths of the solar spectrum was first established. In the present case the mysterious line is even more isolated than in the spectrum of nebulae, since, with the exception only of cases of especially bright auroras, nothing else is visible in this spectrum. For a number of years I made use of every convenient opportunity to observe this spectrum, and its mysteriousness always threw me into the deepest perplexity. I understood that this mysterious line arises not from the depths of cosmic space, but in our own atmosphere, at a distance which even by our terrestrial measures is quite small. And nevertheless the most careful study of the spectra of terrestrial gases, by the combined efforts of very many investigators, could not solve the riddle of its origin.

As is known, the key to solving the riddle was found by McLennan, who succeeded in obtaining this line in strong electrical discharges in a mixture of oxygen with helium or, still better, oxygen with argon. Oxygen plays the chief role here, and there is no doubt that the auroral line is a line of oxygen; but the function performed in this case by the inert gas is not entirely clear, although various more or less probable conjectures on this subject are possible. In any case, the establishment of the fact that this line is connected with oxygen was a great step forward. But, of course, far from everything has yet been done, and we still do not know how to obtain this line by itself, or at least in the presence of the negative bands of nitrogen, as we observe it in the sky.

In the artificial spectrum we obtain both the ordinary lines of oxygen and the lines of the inert gas—helium or argon, depending on which of these gases is present.

The wavelength of the northern-light line cannot be calculated or predicted on the basis of the existing knowledge of the arc spectrum of oxygen. In this case we are dealing with only one line and are deprived of that valuable indication which is given to us by the magnitude of the splitting of doublet or triplet lines in the spectrum of the nebula. But we can, of course, without difficulty find a tentative place for it in the scheme of the arc spectrum of oxygen, on the basis of Hund’s theory. This theory, which may be regarded as a generalization of all our information concerning line spectra, gives a framework into which, with a considerable degree of confidence, we may fit the empirical data accumulated by us.

Proceeding from the fact that the nitrogen bands are not visible in the spectrum of the night sky, which, however, has the green line, MacLennan assumes that its excitation potential is less than 11.5 volts. This already excludes many possibilities. Strictly speaking, if one considers oneself bound by the selection rules, then this excludes absolutely all possibilities. Thus, for example, in the case of the nebula, MacLennan was compelled to depart from these rules and to ascribe the green line to a transition from one or another of the lower metastable states indicated by this theory.

But the very lowest state must be a triplet one, and since the green line has no satellites, this proposal can, with great probability, be excluded.

In that case, however, there remains only one alternative, with which there is in full agreement the successful determination of the Zeeman effect carried out in MacLennan’s laboratory. L. H. Sommer, soon after this, published a completely independent investigation of the same kind, which led him to the same conclusion. This is, of course, very noteworthy, but the position of the theory would be considerably strengthened if we could independently determine the energy levels and theoretically calculate the wavelengths for comparison with the observed facts. For this, more complete observations are necessary—

of Schumann’s region of the arc spectrum than has been done up to now. For the auroral line we already have an experimental reproduction of the phenomenon, but numerical spectroscopic relations are still lacking. For the nebular lines the matter stands exactly the opposite way.

Thus, the origin of the green auroral line is quite fully clarified, at least in the sense that it belongs to the arc spectrum of oxygen. But certain features of the auroral spectrum still remain enigmatic. I shall confine myself to only one of them, namely the red auroral line. Red aurora is observed comparatively rarely, but in this case the distribution of the colors presents very curious details. In some cases the ends of the rays are colored red, whereas the greater part of their length has a green coloration. The only case of red aurora which I was able to observe in my homeland, in the south of England (14 May 1921), had a distinctive character, since the color transitions were very sharp through various shades of purple. The light was distributed in irregular patches almost near the zenith, although the main part of the aurora was situated in the northern half of the sky. At the same time its position was extremely unstable, and the general impression produced by it resembled discharges of a high-voltage current in a very high vacuum. Vegard described a case when the whole sky assumed a bright-red coloration. He obtained a very good spectrogram of the red line on a small scale, according to which its position was equal to \(\lambda 6322\), with a probable error of not less than \(\pm 1\) Å. A determination made by V. M. Slipher at the Lowell Observatory gave \(\lambda 6320\).

So far as can be concluded from the available observations, not one pair of lower levels in the scheme of the arc spectrum of oxygen which McLennan considered in connection with the spectrum of the aurora can be arranged so as to give this red line in combination. It is therefore natural to turn to the spectra of nitrogen, which, as is well known,

represented in the blue and violet regions of the spectrum of the aurora.

In 1922 I described a spectrum in which one of the first positive nitrogen bands, \(\lambda 6323\), was very clearly expressed in comparison with the neighboring red bands, which usually have almost the same brightness. This spectrum was obtained by adding a considerable excess of helium to the afterglow of nitrogen, and, owing to this, the glow visually acquired the red color determined by this band. I then expressed the supposition that this might also explain the origin of the red line of the aurora; approximately the same view was recently expressed by MacLennan in his recent Bakerian lecture. But here we encounter great difficulties. In photographic images the two yellow nitrogen bands stand out with the same brightness as the red ones, if not greater, and yet they are absent in the spectra of the aurora. Moreover, the data concerning the wavelength of the red line of the aurora are not so accurate that we could identify them on the basis of a single coincidence alone. One of the most important tasks in the study of the aurora is to determine the wavelength of this red line by means of large-scale spectrograms.

Coronium.

A problem that is usually included in the category of questions discussed here also concerns that line in the solar corona which is attributed to a hypothetical element—coronium. In the light of present-day knowledge the existence of such an element is highly improbable—one may even say, completely impossible. Many attempts have been made to identify these lines with elements known to us. The most recent of these attempts belongs to Freeman, working in the Ryerson laboratory in Chicago. He attributes these lines to argon and thinks, for example, that the bright green line which served as the occasion for creating the hypothesis of coronium may

may arise from two different transitions in the argon atom and is essentially double. One of the supposed transitions will give the fifth line in a possible series, and the other—the ninth line of the actual series. But not one of the preceding members of each of these series is observed in the corona, and this undermines, at the root, Friman’s theory. We cannot, of course, interpret the observed line \(\lambda 3771\) as \(H_1\), if \(H_\alpha\), \(H_\beta\), \(H_\gamma\), and the other earlier members of the same series are absent; and that would be tantamount to this.

I therefore believe that the origin of the bright lines in the corona remains an unsolved problem. One must reckon with the possibility that they represent the head lines of a molecular band spectrum.

Excitation of Various Spectra.

We have spoken about all these celestial spectra chiefly from the point of view of spectral analysis. It will be of interest also to discuss possible methods of excitation for some of them.

Let us first turn to the spectrum of the aurora borealis. The latter, as is well known, is closely connected with special conditions of magnetic disturbance, which, in turn, are due to the influence of the sun. As for the nature of this influence, the Birkeland theory, developed by Störmer, still remains in force. The sun is regarded as a source of radiation of localized streams of electrically charged particles from limited regions of its surface.

The undeniable advantages of this theory consist in the fact that it takes account of a precisely defined direction of the solar action and thereby explains the unexpected occurrence of magnetic storms over the whole globe, their tendency to recur every twenty-eight days, and the displacement of these phenomena along the night side of the earth under the influence of the earth’s magnetic field. But this theory, in its original simple form, required many emendations; the necessity of creating a number of ad hoc hypotheses can hardly afford much satisfaction.

The emergence of a stream of particles carrying charge of one sign is already, in itself, a weak point of this theory. As Shuster has already indicated, such a stream must scatter, owing to electrostatic repulsion, and lose the sharp outlines that are one of the essential features of the phenomenon. Lindemann proposed to get around this difficulty by taking the stream as a whole to be neutral and consisting of particles with charges of both signs. But in that case the possibility of magnetic deflection of the stream is lost to a considerable extent. Chapman preserves the well-known preponderance of particles of one sign, and thus thinks to create an acceptable compromise. It is clear, in any case, that before entering these theoretically obscure regions, some experimental support is necessary. The search for direct proofs at first sight promised nothing encouraging, but quite recently Størmer put forward a sensational supposition. Hals drew his attention to an echo accompanying short-wave radio signals (131 m) sent from Eindhoven in Holland. According to the investigations of Hals and Størmer, these echoes are heard after quite large intervals following the initial signal (up to 15 seconds).

If one bears in mind that, at the speed of light, the maximum terrestrial distances can give intervals of \(1/7\) second, then one must think of some extraterrestrial reflector. Størmer seeks it in the stream of particles bending around the earth under the influence of terrestrial magnetism. For all the boldness of this conception, it is difficult to devise anything in its place. Størmer at the same time points out that “the variability of this phenomenon, established by experiments, is in very good agreement with the corresponding changes of the aurora and of magnetic deviations.”

T. L. Eckersley reports natural electrical disturbances he has observed which, in his opinion, are analogous to the phenomena described by Størmer. If a telephone is connected to a large antenna, clicks can be heard in it, after which, some time later,

for three seconds a “whistle” or musical note of short duration is heard. Further whistles follow at intervals of 3.8 seconds, with each subsequent sound having a longer duration than the preceding one. He regards these sounds as a consequence of the dispersive action of the medium on the electrical impulse. These phenomena are often observed only during magnetic storms.

Further observations in this field are awaited with the keenest interest.

But let us return to our nebular spectrum. Although the main problem, as already indicated, has been solved, it is very important to reproduce this spectrum in the laboratory—not so much to confirm the theory of the origin of the lines as to clarify the conditions under which the spectrum arises. Attempts of this kind have not yet yielded results, but it is already clear in what direction one must seek a solution. It is necessary to create the conditions for exciting the lines of doubly ionized oxygen and to find a way of working in a large volume at a high degree of rarefaction.

A large volume and a high degree of rarefaction (rarity of particle collisions) are considered characteristic of nebulae and, as Bowen explained, are essential requirements. It must be said, however, that the results obtained so far in attempts to obtain transitions from metastable states do not provide direct confirmation of this.

One should not, however, neglect what Darwin called “foolish experiments,” and what oil prospectors call “hunting wildcats.” Many highly fruitful discoveries have been made in precisely this way. Logic is already an afterthought. This is exactly what happened with the three-electrode thermionic amplifier.

Thanks to the work of Wright, Hubble, and others, the source of excitation of the bright lines of nebulae is no longer inexplicable. We already know the important fact that in almost all cases stars of an early type, capable of high-frequency radiation, are associated with these nebulae. Two or three exceptions, though also

deserve attention, can no longer shake the generalization which rests upon a large number of observations. True, we cannot observe these short waves, since the maximally intense part of the spectrum is hidden from our sight by the layer of ozone, of which I shall speak in more detail below. But we can judge with confidence of the existence of these waves on the basis of extrapolation from what is visible to us, making corrections for the atmospheric absorption known to us. The existence of central nuclei in some planetary nebulae is especially convincing in the sense of establishing a definite relation between the star and the nebula and the character of the star itself. In 1918, i.e. even before these views had taken shape, W. T. Wright, not yet thinking of any theory on this score, wrote as follows: “This remarkable abundance of ultraviolet rays, which gives a peculiar appearance to the spectra of the nuclei of nebulae, is, in my deep conviction, despite the differences in their bright bands, the predominant feature and distinctive mark of this group of luminaries.”

The view has been expressed that the penetrating cosmic rays, about which so much has been said lately, excite also the spectra of nebulae; but in view of the facts already known, such a hypothesis is hardly necessary or desirable.

Dark Spots in Nebulae.

There is one feature in diffuse galactic nebulae which remains dark in the literal and figurative sense. It is especially conspicuous in such objects as the “Trifid” Nebula in Sagittarius. Here the dark places are so intermingled with the bright ones that the thought naturally suggests itself that here we have a phenomenon of the same order as the connection of emission with absorption in a fluorescing body. However, it is difficult here to say anything beyond this general idea. Opacity has no relation to emission, and the peculiarity of this phenomenon consists in the absence of any

peculiarity. Apparently, all parts of the spectrum of the stars lying behind are darkened to the same degree. Anyone engaged in experimentation in the field of optics knows well how difficult it is to obtain results of this kind in the laboratory, especially when it is desirable to include in the region of investigation the ultraviolet part of the spectrum as well.

Almost all the more or less opaque gases at our disposal, such as, for example, iodine vapor, also possess a sharply expressed selective capacity for absorption, and under terrestrial conditions we usually have to resort to partial darkening by means of a solid body, such as, for example, a rotating sector or a wire mesh invisible in the focus. In astronomical investigations the same role is played by a swarm of meteorites, and however necessary their aid may be, that rarefied gaseous atmosphere which is capable of giving the line spectrum of hydrogen, helium, nitrogen, and oxygen is hardly conveniently combined with a swarm of meteorites, or can have any special connection with it. From this point of view it is decidedly impossible to understand in what way the bright lines of the nebula are often observed against an entirely dark background, devoid or almost devoid of the signs of a continuous spectrum.

Comets.

Analogous to the preceding problem is the luminosity of comets. Zanstra1, in his recently published article, discusses this question. He holds the view that the Swan bands of carbon are resonance bands excited in the visible spectrum by the light of the Sun, while the gases have a temperature no higher than that which occurs under terrestrial conditions. If this view is correct, then imitating a comet is an ideally easy matter from the laboratory point of view. The Swan spectrum should appear as the absorption spectrum of carbonaceous gases enclosed in a vessel at ordinary temperature, and should be observed in lateral emission. I think, however, that if

For such an experiment nothing else would be required; it would have been done long ago. In the case of the sodium \(D\) line, which Zanstra considers to be entirely analogous, P. V. Bydom has long since described the phenomenon of resonance radiation.

Metastable States

In considering the question of the spectra of nebulae and of the aurora borealis, we have encountered the concept of a “metastable state.” At present this concept has not yet received a sufficiently clear formulation. Initially it was understood to mean a state in which an atom, while radiating energy, cannot pass directly into the normal stable state. Let us liken the energy levels of an atom to the floors of a building, and the optical electron to a person located inside this building. The normal state of the atom corresponds to the situation in which the person is on the lower floor, while the metastable state corresponds to his moving to the second floor. But the internal architecture of our atom is peculiar. The second floor is connected by a staircase with the third; another staircase connects the third floor with the lower floor; but between the first and second floors there is no direct connection, and in order to get from one to the other one must first go upstairs and then come down.

Such, I repeat, was the original conception, but the facts that were subsequently discovered required a certain modification of it.

In nebulae the electron somehow contrives to slip out of its confinement and descends to the lower floor not by the usual route through the second floor, but by breaking through the floor, contrary to all the rules established in the building.

Now leaving this metaphor, I shall say that the selection rule which does not allow transitions unconnected with a change in the azimuthal quantum number is not observed in all these cases. The selection rule for the internal quantum number, which requires that the internal quantum number not change from 2 to 0 or from 0 to 0,

is also violated in a number of cases, and is expressed very inaccurately in another. According to this rule, in doubly ionized oxygen only one pair of green lines is possible; in fact, two separate lines are observed instead of the three that might have been expected, taking into account that the ground state is a triplet.

Nevertheless, we find the blue singlet line \(\lambda 4363\) of this ion, which violates the indicated rule; the same applies to the spectrum of the aurora borealis, if we adopt MacLennan’s view of the position of the lines in the scheme of the arc spectrum.

In the case of the mercury spectrum, which is a convenient object of experiment and has been well studied, we have a laboratory example of the violation of the same rule, as follows from the experiments of Takamine, Fukuda, and other Japanese physicists. These lines were originally obtained under the action of a strong electric field, and in this the cause of the non-observance of the rule was sought. In addition, the lines were of very low intensity, which also served as a certain justification.

Whatever one’s attitude toward these explanations, their untenability, I believe, is clearly proved by my own experiments, in which I succeeded in obtaining some of the “forbidden” mercury lines, whose brightness occupied second place in the entire emission spectrum. This was observed when vapors were passed through a discharge, but away from the actual site of the discharge and, consequently, in the absence of an external electric field.

In another experiment I succeeded in obtaining another “forbidden” line as an absorption line in unexcited mercury vapor and therefore in the absence of any perturbing conditions whatsoever. In this experiment the amount of vapor was very large—approximately ten million times greater than that required to obtain the resonance line of mercury in the absorption spectrum. The probability of the transition was thus very small, and for the other forbidden lines still smaller.

Nevertheless, as we have seen, this forbidden line can be obtained with considerable brightness in the emission spectrum. A necessary condition in these experiments with mercury vapor is a large accumulation of mercury atoms in a metastable state, so that even with a very small probability of transition for each individual excited atom, a certain number of such transitions nevertheless takes place.

An attempt has even been made to define the metastable state as a state characterized by a very small probability of transitions. But this takes us very far from the original conception and makes “metastability” merely a matter of degree. Some of the most recent results seem to indicate that even the normal state of excitation may last considerably longer than has hitherto been supposed. If this view is accepted, then our present views must undergo a radical revision. The general softening of outlines in our picture of atomic processes, as a result of replacing particles by groups of waves, apparently opens up this possibility and in some cases allows atoms to pass to a lower level from a metastable state.

Ozone

The spectrum of nebulae may serve as an illustration of how the theory of spectra, on the basis of laboratory data for the far ultraviolet region, enables us in part to circumvent the difficulties caused by the impossibility of studying this region in the spectra of celestial bodies. The ozone covering surrounding us conceals from us the spectra of the Sun and the stars and conceals a considerable part of the ultraviolet rays, which is one of the chief obstacles for astrophysics. It must not, of course, be forgotten that this same covering also protects us from the harmful action of ultraviolet rays, and that without it there would be no one at all to engage in astrophysics. Since atmospheric ozone is formed owing to the absorption of short waves in the solar spectrum, Cario, R. W. Wood, and others have expressed the supposition that it

may be absent in the Arctic regions during the polar night. This supposition was tested in practice by Rosseland, who obtained negative results. True, his observing station was not sufficiently far north and isolated from sunlight for these results to be regarded as final. But other facts are likewise not very reassuring. Thus, Chalonge found that the amount of ozone at night (with the Moon as the source of the spectrum) is considerably greater than by day. Dobson, Harrison, and Lawrence found that under such meteorological conditions, when air masses are carried to us from the Arctic, the amount of ozone increases, and that this is observed especially in spring, i.e., after the prolonged nocturnal period in the Arctic. All these facts are, of course, of enormous importance in themselves, and they further obscure the question of how ozone is formed. In any case, as Dobson says, they substantially shake the view that ozone is the result of the Sun’s ultraviolet radiation.

As for the atmospheres of the planets, it appears that no searches for ozone have yet been undertaken there. In the case of Mars, Jupiter, and Saturn, the task at first glance seems not to present any special difficulties, if, of course, the layer of ozone there reaches at least the same thickness as on Earth.

The possibility of the existence of unknown elements with high atomic weight.

Although at the present time we can no longer count on the discovery of any new light elements, there are no logical grounds for excluding the possibility of the existence of some elements with a heavier atomic weight than those known to us on Earth. Jeans, as is well known, makes use of this hypothesis to explain the origin of stellar energy. Together with several other authors, he sees its source in the destruction of matter with the emission of an equivalent quantity of energy \((mc^2)\), as required by the theory

relativity. In this form the theory encounters almost no objections. But a difficulty arises as soon as we approach the question of stability, and here opinions begin to diverge. Jeans holds to the view that the source can release energy at a rate independent of temperature. I do not consider myself competent and shall not discuss this question. The purpose of postulating the unknown heavy elements is to acknowledge their ability to cease their existence at a rate independent of external conditions, with the sole exception that ionization, consisting in the removal of some of the electrons from the neighborhood of the proton, tends to delay this process.

The radioactive elements known to us are, of course, examples of similar unstable forms of matter, and Jeans considers them transitional. It must be recognized, however, that in substances subject to spontaneous decay one can hardly see suitable intermediate stages between perfectly stable bodies and those which spontaneously cease to exist. It is necessary, further, to find an explanation of the fact that these heavy atoms are not found on the earth, which once, as is unanimously acknowledged, was part of the solar mass. Jeans himself mentions this difficulty and suggests that the heavy elements must have sunk into the inner regions of the solar mass, so that the earth, formed from the outer parts, may not contain them. However plausible this explanation may be, it is nevertheless rather difficult to admit that this process could have produced so exact a result. The list of elements known to us ends with uranium, and, as we know, all the elements occupying the ninety-two places up to and including uranium fully correspond to their atomic numbers. There are only two exceptions: 85 and 87, but one cannot attach serious significance to these gaps, which may be filled any day. Thus all the elements up to and including uranium exist in fact, while all the heavier ones attributed to the stars are absent—

exist. To ascribe so perfect a separation to the mechanism of gravitation would mean assigning it too difficult a task. The inventors of mechanical ore-dressing could not dream of such a superior mechanism.

Nature works on a very broad scale and is not pressed for time, and therefore one may say that we should not measure her achievements by our own yardstick. But we have the right to seek more direct indications that she has produced such a separation of the elements. If the line of division passed between 92 and 93, then one would have had to expect that the greater part of class 92 would likewise have crossed the boundary, if nothing at all remained of class 93. But class 92 (uranium) turns out to be by no means a rare element on earth, since in its content in volcanic rocks it, according to the determination of Clarke and Washington, occupies the 25th place. On the other hand, we observe, at least on earth, that uranium has not only not sunk inward, but is concentrated chiefly on the surface. To this conclusion we are led by investigations of the earth’s heat emission, which is difficult to reconcile with the observed quantity of radioactive substances near the surface and is wholly inconsistent with the existence of similar quantities within the earth.

Assuming that uranium exists on the sun as it does on the earth, we have every reason to believe, as Lindemann has correctly indicated, that it is there in the process of formation. The lifetime of uranium is too short to allow anything else, given the probable age of the sun. Those who remember the early stages of the study of radioactivity know that this consideration gave Rutherford the right to speak of the constant origination of radium on earth, before this had been directly confirmed. Radium, as was subsequently proved, arises from the parent element—uranium, which has a greater atomic weight. Jeans admits an analogous origin for uranium as well, but in that case it is necessary to allow for the existence of an element of still greater atomic weight, capable of undergoing

radioactive decay, but is incapable of being converted entirely into radiation.

There is no doubt that here we are stepping onto very shaky ground. Here only such a theory of atomic structure can help us as would be in full agreement with the experimental data concerning the elements known to us, and could give us an idea of the properties of elements of greater atomic weight than class 92. On the general question of whether the evolution of the elements proceeded from simple to complex or from complex to simple, it seems to me there is no need to appeal to evolutionary doctrine or to seek analogies in organic evolution in the interests of the first view. Is it not a more important consideration that all the facts observed by us relate precisely to the second alternative (radioactive changes and changes caused by radioactive bombardment)? For the time being this is still a matter of scientific taste. But perhaps it will not be superfluous to note that even in the processes of organic evolution there sometimes occurs a degeneration of organisms; and can our biologist colleagues confidently assert that someday, with a change of conditions, the whole course of organic evolution will not acquire a reverse current? In any case, for us it is also important that we can now pose a question that is, though more limited, more precisely formulated: what is the character of the process of the formation of uranium on the Sun—analytical or synthetic? It seems to me that the question is posed quite clearly, and that there can be only one answer to it.

Conclusion.

The modern successes of theoretical research naturally impel enterprising minds to resort to it not only for explaining what we already know or can verify by means of observation, but also for excursions into those regions where no verifying experiments are possible. I consider this far from harmful at the present time, when there is no danger that the theories of even the most authoritative scholars will be upheld even contrary to the testimony of

facts. But, of course, we should not expect too much from the work of the intellect in a region inaccessible to observation. Theories that do not withstand the test of time for the most part sink into the tangle of oblivion, and we tend to forget how many collapses occur in this field. The next generation remembers only those theories that have proved viable, and forgets how many conclusions, which had seemed unquestionable, were refuted by reality. If our reasoning does not contain confirmation within itself, thanks to some precise and, in advance, unpredictable numerical coincidence, then it is often difficult even to say beforehand whether we are on the right path or not.

Although some of the problems considered above have been solved only in part, and others have not yet received any solution at all, many paths toward a solution are opening before us.

The war with the mysteries of nature now requires a very broad front. As soon as the attack weakens at any point, new crowds of fighters, unhampered by the bonds of traditional scientific thinking, are ready to fill the ranks and rush into battle. A whole army of trained workers in the field of pure knowledge, existing in modern society, is reinforced by workers in applied science, who return to the physical laboratory the apparatus created by the latter, considerably improved, strengthened, and perfected. Reinforced by this new weapon, pure science attacks new regions, and this process goes on continuously. Nevertheless, there is no reason to fear that our descendants will complain of us for having left them with nothing to do.

1. Zanstra. Monthly Notices. December, 1928. ↩↩