Interstellar Communication By Philip Morrison Cornell University [Read before the Society, October 7, 1960] IT is a great honor indeed to address the 1496th meeting of any organization, and particularly one founded by the man who was the second of the famous American physicists, after Benjamin Franklin, Joseph Henry. I should like to put very clearly the thesis to which I wish to speak and concerning which I hope to present a very plausible case. I propose to assert that near some star rather like our sun there now exists a civilization with scientific interests and with technical possibilities much greater than those now available to us. Moreover, to the beings of such a society, our sun must appear as a likely site for a similar civilization. It is probable that for a long time they have awaited specific development of science near the sun and I believe that they look forward patiently to signals from our solar system which would make known to them a new society ready to enter the community of intelligence. I would like to ask what sort of communication channel is open? What are the circumstances in those remote spaces, into which now, for the first time with some understanding, we begin to probe? Can we expect the extraordinary encounter with another civilization to which it seems to me we must inevitably come? The idea is not a new one. It has had a lengthy history, and a kind of efflorescence in fiction in the recent past. There is also some scientific literature—about a half dozen papers in the last year or two. But the first references—and I hope to quote one of the most eloquent at the close of my remarks—go back at least to the thirteenth century, and indeed this view prevailed in those great times, at the turn of the seventeenth century, when that science which we today serve was in fact born. Tonight, I would like first to speak to the question of whether or not we could recognize living things of a very different form from our own. I shall set up, if you like, a criterion of conservatism. Suppose we make a graph (Fig. 1), Graph: Complexity vs. Frequency of Occurrence Fig. 1 plotting on the horizontal axis some measure of complexity. On the vertical axis we present the population of some region as a function of the complexity of systems in the population. In general, if we look at a sterile sample, grabbed at random from an imaginary world without life, or from a part of our own world without life, taking for instance a sample of sea, beach, or sky, we would find a curve similar to Curve I (Fig. 1). In a random distribution of atomic structures of all kinds we must expect that complex forms will very rarely exist. Whether indeed there is ever enough time to permit the chance existence of even one genuinely complex form such as we find in a fern or a tree or a man, I very much doubt. But in principle, such a curve should contain an entry for any complex thing, but with a very small probability for extremely complex systems. This is the curve for a sterile world where no life has ever been seen. The coming of life, I think could be characterized quite objectively, without reference to its chemical nature or the place in which it is found, by modifying this curve in a very characteristic way. We will cut off the tail of the curve, replacing it by a rather rapid decline toward zero; but we will replace the missing tail by a pip (Curve II). At first the pip is small and, as time proceeds, gradually increases, representing the accumulation of living forms, who derive their sustenance by operating on the environment. They prevent the establishment chance processes of the intermediately complex products which, before the existence of life, would have filled in the part the curve between the pip and the main section. I suspect that such a two-piece curve would adequately characterize the existence of living forms. I would however, go further. If we imagine now that living forms evolve, as they have evolved on our own planet, to a certain capability of manipulating the free energy of the environment beyond that which with their own tentacles, they can touch, then we will find the growth of still another pip. This part of the curve represents still more complex systems, and if our experience is guide, it will grow still more, reflecting the cultural product of this second step. This is a third set of complexities. Here we include, for example, the books of the libraries of the world or, for that matter, the automobiles or the woven textiles or the architects' constructions. The presence of such breaks in the series of complexities, is, I think, a characteristic feature which we have to require for the existence of living forms, and then also for the existence of what we call cultural activities. As you see, I have made no reference whatever to whether the systems are liquid, solid or gaseous, whether they are magnetic or made of fluorine or carbon. We are dealing in matters in which we are at the moment simply too ignorant to be able to characterize them with even slight reliability. We are lucky if we can simply formulate a general program of such research. I should like now to make a much more modest yet less general proposal. I should like simply to say that if to the process which unites familiar atoms into complexes, there is available an adequate stream of free energy, then we must expect to obtain curves similar to Curve II of Fig. 1. There are more speculative thinkers who are not yet writing in serious journals but in, I assure you, quite serious works of fiction! Mr. Fred Hoyle for instance has described to us in his extremely suggestive and imaginative work, The Black Cloud, the emergence of a form of complexity built out of plasma, i.e., of magnetic fields, hot gas, and dust. I only mention this to show how far one can carry such speculations. No more imaginative discussion! Everything from now on is done with analogies of the most timid kind, avoiding any serious extrapolations. I shall look for precisely the material basis which we ourselves see on the Earth, forming what I have called a scientifically capable society. I shall look first for the existence of the abundant light elements: hydrogen, carbon, nitrogen, oxygen, magnesium, phosphorus, sodium, potassium, chlorine, calcium, iron. Hardly anything else is needed. We know that in the process of nucleogenesis their incidence is high. I would like to go so far as to assume that probably the chemical reactions which will lie at the base of the structures, unlike the plasma beings, are probably based on water in the liquid phase or, at the extreme of an extrapolation, to some other mode of utilization of the flexible and convenient hydrogen bond. This places severe limitations upon the physical environment in which such forms could evolve; namely, the limitations through which we ourselves have evolved. If the temperatures are very much below zero degrees centigrade the processes involving hydrogen-bond formation in solid water solutions are much too slow to give a very wide split of the original curve in the times available. So, with too low temperatures, while something may go on, it may be very slow indeed. With too high temperatures, one cannot make any such structures very easily: the structures involving the rather weak but flexible and highly manipulatable hydrogen bonds are next to impossible. Therefore we have a severe requirement on temperature. Besides thermodynamic temperature, we have to worry about the presence of currents of free energy of such concentrated kind that they can destroy complex structures, even though they may not represent a large contribution to the over-all thermal content of the environment. Here, of course, I refer to fluxes of high-energy particles, ionizing radiation, and the like: quanta of energies large compared to energies of chemical bonds. Such bombardment must be severely limited compared to chemical formation rates or those structures cannot evolve. We know we must have something like the atmosphere of the earth for this protection. Open exposure to the indiscriminate currents and fluxes of space will prevent elaboration of molecularly based complexities of the sort I describe. Andromeda Nebula Let us consider the great nebula in Andromeda in the north sky (left). It is a circular mass of unresolved stars some 120,000 light years across and appears to us as a tilted disk. We know that we live in a galaxy almost the twin of this one. Such a galaxy contains on the order of one hundred billions of stars, appearing to be merged together only because they are so far away that the photographic plate and the optical train through which the plate has been exposed cannot separate the images. If someone looks at our galaxy from the outside, he sees such a relatively bland, slightly spiral, lightly marked disk of whirling stars, rotating once every three hundred million years. Not far from the center of this light patch there is the sun, an inconspicuous member of a population typical of the entire galaxy. Around the sun there whirl, among others, two planets which may be suitable for life; namely, the Earth and Mars. On the Earth is a clear and present instance of the development of the society of modest scientific capability in the very recent past; and on Mars, a strong indication that conditions favor the independent evolution of life, not indeed to any stage of technical capability but (very plausibly) to a stage of considerable chemical elaboration. It is upon this sample of two that I think the argument we make must rest. I would like to stress the probable existence of some form of life on Mars—some form of growth of complexity at the expense of the free energy of the environment—because otherwise we are left to the uncomfortable sample size of one, the unique example of Earth. I learned in statistics that it is very hard to make a conclusion from a sample so small as one. Anything can happen once, but about things that happen twice—one can at least say each is not unique. This is the main reason I would like to see growing in the Smithsonian Institution a little sample of that Martian vegetation which Dr. W. Sinton has made so plausible from his magnificent work with infrared spectroscopy. Have we any certain knowledge of what goes on outside the solar system? I do not think so. But we have some conjectures and some plausible inferences which I should like to list briefly in order to establish the rest of my argument. We wish to look for those environments, resembling to a considerable degree the terrestrial environment and the Martian environment, which are hospitable enough to allow the elaboration of chemical complexity in the form of living beings. We are going to look therefore for terrestrial planets, with atmospheres, with free energy supplied by the sun, and with a temperature regime like our own. In spite of considerable effort we have not yet found a really sound basis for the origin of the planetary system. We could not, for example, calculate clearly the distribution and mass and position a priori from the knowledge of the type of star we have. But we do have, besides some rather plausible inferences of this kind, some observations which are not difficult to describe. Graph: Star Rotation vs. Color Fig. 2 Here again a graph (Fig. 2), and I will mark here the familiar letters which indicate to the astronomer, roughly speaking, a scale of temperature from about 50,000°K on the left to about 2500°K on the right, with our own sun being at the letter G. Next I would like to plot the measure of rotation of typical stars of these various classes. There is a great deal of rotation among the stars of classes of O, B, A, but we do not see much rotation in the F, G, K stars. If we look at the distribution of the angular momentum in our present solar system we find that 99.5 percent of the angular momentum lies, not in the sun, which has 99.9 percent of all the mass, but in the planets which go about the sun. Thus the angular momentum, which was presumably present in the whirling gases from which the sun condensed billions of years ago, now resides not in the massive sun but in the tiny planets. It is a plausible inference that the reason why the angular momentum of these young stars is still in the stars themselves, as we can see by the spectroscope, is that they have not made planets. Similarly the angular momentum which the older stars must certainly have had, if they were formed by the same processes that made the younger ones, has been whirled off perhaps in many forms; it is not unreasonable to say that some of these stars have bequeathed their spin to planets. Here the argument is of course uncertain. We cannot be sure that the conditions for condensation of planets were right. We can be fairly sure that the angular momentum went off with much gas. But since the angular momentum is still resident in the planetary system of our sun, it is plausible that most of the stars of masses like the sun, might well be surrounded by a suitable cortege of planets, which bear a small part of the mass of the original cloud but an appreciable fraction of its original endowment of angular momentum. Near stars of F class and fainter we then expect to find an appropriate distribution of planets. Let us imagine some considerable fraction do have planets. (We will recall the uncertainty in a final factor eventually.) There must be enough light; otherwise we would find too low a temperature for life. The planets which we impute to the stars must be in the right positions, receiving neither too much light, so that they are sun-baked like Mercury, or too little light, so that they are cold and sodden with mists of methane and hydrogen, like Jupiter. They must be somewhere analogous to the Mars-Earth region. In all this I follow the work of S. S. Huang. He draws this inference: Since the light from the sun is received here diluted by the spread of the light flux in free space—if you like, by the inverse square factor—we can scale all stars and their planets to have the heat and light of earthly conditions, provided we scale the distance to the planet according to the luminosity of its star. The planets' distance being called R, R2 will have to be proportional to the luminosity of the star. Let us say we have a spread of possible distance in our system something like the spread between Earth and Mars, or even between Mars and Venus. If we imagine that whirling planets have always formed disks, this allowable area is in turn proportional to R2. (The area of an annulus is proportional to the square of its radius.) In such an area a planet near any star would find conditions tolerably close to our own. Moreover, the calculations of William H. Guier and Robert W. Hart at the Johns Hopkins University Applied Physics Laboratory have demonstrated that the distribution of masses in any "solar" system ought to be rather flat in this region. It, is reasonable not to make any further correction, but simply to say the area available for Earthlike orbits is the only measure of how lucky a planet has to be in order to receive the right amount of light. Now, we see that a very faint star may have planets feeling the same light intensity as we enjoy on Earth, but only if they lie in a little disk hugging the star, to gain the benefit of the small warmth of their faint furnace. Our argument simply has shown that these habitable areas are proportional to the luminosity of the star. It might well be that there are such systems near faint stars, but they must be few, because the volume allowed in space for the statistical distribution of habitable planets is not large, since one must live so close to a faint star. Therefore, we can be pretty sure that the very numerous faint stars are not likely seats for planets endowed with the kind of warmth that we have here. One can calculate this nicely, using statistics on the distribution of stars in the galaxy, and we have done this, to find a curve rather like the one shown in Fig. 3. Graph: Star Luminosity vs. Incidence of Warm Planets Fig. 3 The size of the zone of warmth, multiplied by the fraction of stars having that luminosity, gives the relative probability of finding a planet around a star, dependent on the luminosity of the central star. In the middle lies the sun, to the right are fainter and fainter stars, very numerous ones. To the left brighter stars, but conspicuous ones. On this consideration alone, each of these stars would be very likely to have planets because there is ample warmth; each has a big useful volume. But since they are few in number the total contribution cannot be large, and the brightest ones of all still spin. They have no planets. The faint stars are not very important then; but the stars at the center of the range are not so important either because the brighter stars, even though they are not very numerous, are so much brighter that they make up for their scarceness by their favorable chance to have comfortable planets. So, if we have no other criterion, we would say the most likely thing is that such planets will be found in that very wide zone of tolerance near the very bright furnace of the high-temperature stars. But we have one more indispensable requirement—time. No doubt the spontaneous process of energy degradation and the transfers of free energy require a large amount of time—if you will, geologic amounts of time. Indeed, that is the story of paleontology. We have to allow billions of years for the elaboration of those many forms which necessarily precede the kind of complex beings we are looking for. This time is not available in the case of the bright stars, because they burn themselves out, and move off the sequence, perhaps to go through all sorts of catastrophic changes. Only the conservative, smoothly flowing sources of radiation, the weak stars, will work. Suppose we want the star to remain without an appreciable change in temperature, during a time of three or four billions of years, like the time since the first signs of life appeared in our solar system. Then we must multiply this curve by the fraction of that time which the life of the star represents. The very brightest stars do not live anything close to that time. Therefore, they contribute nothing. By the time we multiply this curve by an appropriate time factor we have produced the effect shown in Fig. 4. Graph: Star Temperature vs. Probability of a Planet with Evolved Life Fig. 4 This is the probability, taking into account both light and time, among only those stars with planets. Recall that the brightest stars do not have planets because they too still rotate. The sun lies tolerably close to the maximum of the final curve. About 90 percent of all the stars that are plausible homes of life in our hypotheses are contained in the small range of surface temperature which astronomers would call late F to K classes. These stars vary from our sun's temperature by perhaps 10 percent one way or the other. We should probably exclude multiple stars (an argument also due to Huang), although multiple stars may well have planets. They illuminate their planets so differentially that instead of mere seasons, the planets undergo extremely complex rhythms of heating, unlike our rather smooth evolutionary history. A skeptic might well say this would be a kind of stimulation and challenge to early life. But consistent with our determination to extrapolate hardly at all, we shall exclude all multiple stars as being possible sites of life like our own. Maybe they are sites for other forms, more suitable to a climate which may change enormously in a few million years and then change back a few million years later, but we shall not discuss them further. We can say that around the simple dwarfs of the main sequence, from what are called dG0 to dK2, we subsume perhaps 90 percent of all the possibilities for having planets with atmospheres and temperatures like the terrestrial planets of the solar system. Now, the number even of these special stars is not small. The number of such eligible stars is a few hundred million in our galaxy alone. If we exclude the multiple stars, we cut this by a factor of three or four, to above a hundred million instead of several hundred million. If we allow that the galaxy is several times older than our sun and allow for the fact that some stars have played out within that time, we still come to many times 107, or say roughly one hundred million. Where are these stars located? They are stars of the disk population, found not far from the central galactic plane, anywhere from near the center quite far out to the rim. One can compute that eligible stars are sitting fifty to eighty light years apart throughout the whole bulk of the galaxy. A remote astronomer observing our galaxy sees a bright mass like the Andromeda nebula, with fifty to a hundred million star-spots at which he might plausibly argue that living forms occur. We now know he would be right about exactly one of those spots; namely, our own planet. It seems to me an irresistible inference to say he may be right about many of those spots. There is no central feature, there is no great arrow in the heavens to mark where we live. We are but one mote in this enormous Keplerian ring that runs around the galaxy, democratically indistinguishable from our fifty million dG0 to dK2 counterparts. Now we must ask: What is the history of life as we see it here? Can we not expect this to have some kind of counterpart in these other possible, still unknown seats? I plot in Fig. 5 time as abscissa and, as ordinate, the number, population, the amount, or some other measure of quantity for a number of different interesting phenomena which we know have gone on in time on the surface of a planet near our sun. Graph: Development of Various Characteristics over Time Fig. 5 Here we must be rather flexible. I will mention the significant features of the plots (which are not to he interpreted literally!). At the right-hand end of Fig. 5 is 1960. At the left is five billion B.C. I ask, for example, what would the expert observer, who knows everything, say about the plot of the planetary mass, the total mass congealed into good working planets around our sun? We know it has not changed very much for the last 4.5 to 6 billion years and that earlier there was a time when there were no planets at all, but only a kind of gas. So the planets' mass had to rise from zero up to its final value along a curve something like the top one in the figure. Next, I plot on the same axes the total mass, not of planets, but of living forms in the solar system; i.e., the mass of life upon Earth (Fig. 5, middle curve). Life itself begins much later than the planets. Things rapidly grew, and built up smoothly until the time when the first land forms began after life had filled the seas for a thousand million years. Here occurs a little bump in the curve. Then came the flowering of land life, which represents an increase in total life, but not a great one. Even to this day most life probably exists not on the land but still in the sea, the first home of life. Again experts may disagree whether the lands hold twenty percent or eighty percent, of life but broadly speaking, that does not change the look of the curve. As we come close to the present I would like to add one more small increase, not easy to calculate. In the very last moments of geologic time, a little spike protrudes from the curve which represents the replacement of the forest by crop lands and the irrigation of new lands. It is the first effect of culture, the third rise in my original curve. The middle curve of Fig. 5 then reports the total mass of living beings as a function of time. I have also tried to estimate such a curve for flame. It is easy in principle for an observer to measure flame; he can distinguish the flame of a fire from most of the gases and glowing liquids of a volcano. When I thought this through, I was surprised to realize that long before there were men there was fire, burning in the grass lands and in the forests where lightning ignited it, over much of geologic time. I have made a rough, but reasonable extrapolation, based on present experience in remote countries, which would lead me to believe that the plot of fire would look something like the bottom curve of Fig. 5. It begins when the land forms begin. When there was no life on land there was not much fire. Fire grew nicely as the forests and grass became well developed; then it grew very much indeed when first men came on the scene. Thereafter it did not grow much until rather recently, when agriculture and then cities were invented. In the last couple of hundred years when industry was developed, it went up again though yet not very far. Of course this last time interval cannot be shown in the figure; there is actually a spike at the end. Finally, I will show one more curve which is really the key to what I am driving at. It is a very easy curve to draw, the steepest possible. That is the curve for the population of telescopes in the solar system. Up to 300 years ago, there were absolutely none; and then whatever number there are, now appear effectively all at once. You can not fairly represent the time since Galileo by the thickness of a fine line; the curve is an absolute step function. I take the population of telescopes to be a very good representation of the beginnings of a technically competent cultural inventory. Therefore, no matter what we think is the distribution of the histories of cultures in these other parts of the universe, since the rise time of science is so small compared to the spread in their starting times, to spread in the rates of evolution, and to every other cause of spread that we can imagine, the starting points of culture are distributed more or less uniformly over a time very large compared to the difference between Galileo or even the Chaldeans and 1960. That means if one simply assumes the cultures of these stars we talk about did not have exactly the same starting time and exactly the same evolutionary rate as our own—that is, unless the synchronization is exact to a wholly unreasonable degree—because of this very short rise, we can be sure, that if civilizations exist, then about half of them are far older culturally than our own. But what is the probability that these older cultures also have an appreciable longevity? They might, of course give up science or even die out. Here we come to points still harder to calculate than the very difficult problems of evolution and planetary formation with which I began. We are trembling on the edges of speculation which our science is inadequate to handle. Our experience, our history, is not yet rich enough to allow sound generalization. I beg those who are historically minded and socially trained to consider whether any general remarks may he made to form some guide for us in this problem. How likely is it that populations of men, or manlike things, would evolve along that curious path which leads to the swift succession of those steeply rising functions which I think are characteristic of the artifacts of man? I do not know. We will say, then, that a certain number, say ε×107 stars in the galaxy contain living forms superior in culture to us. Of those some may continue to exist, some may still be scientifically interested, some may have remarkable scientific ability. The factor ε conceals the following probabilities which we know nothing about: (1) The probability that under the same conditions something like "men" will rise from other living forms; (2) the probability that those societies will remain interested; (3) that they maintain technical capabilities of an increasing sort; and (4) that those societies have a longevity great compared to the span of human history, if not comparable with the span of geologic time. Let each person put in his own guess for ε. Those who are very pessimistic will say ε=0. I think if we approach the problem with the usual hopeful hypotheses of scientific investigation, we will say: "no reason to put it zero." I do not know what ε is, but we ought to try some schemes of measurement to find out. Therefore, we argue that near a number which may be somewhere up to 200 million stars, at most, certainly not much more, and perhaps as small as one, certainly not less, somewhere near this number of stars in our galaxy there are astronomers, telescopes, and the rest, and most of them understand much better than we do stellar evolution, planetary formation, radio propagation, etc. They do so not for any reason intrinsic to the mental forms with which they may describe these things, but for reasons intrinsic for the survival of these organisms in a bath of sunlight, protected by an atmosphere. These reasons force them, step by step, if they are to investigate their environment, to carry through the same sort of measurements and to obtain the same sort of information about the spaces between the stars and about the stars themselves as we have, only very likely much more. That, then, is the situation in which we place ourselves when we look at the problem: Do these beings communicate, and how will they choose to send their communications? First, what kind of communications would such advanced societies be likely to undertake? Would they go traveling? I submit that the motives for travel, even in a less advanced culture like our own, are becoming fewer and fewer from the point of view of the explorers of old time. Explorers seeking sources of raw material, like migrations of people seeking new crop lands, have relatively less importance each decade. The major explorations of today, even the major travel of today, is for gathering information, even here on the surface of the Earth. On a much larger scale, if one must dispatch a rocket ship to the Pleiades to bring back a carload of plutonium iodide, it simply is not worth it. If you are in a position to do thing like that, you are in a much better position to make plutonium, or to do without it. To dream of bringing that cargo is to put the thinking of the merchant adventures of the sixteenth century into the framework of a technical ability enormously greater than that of our own day. There is only one real motive for travel (aside from ceremonial or symbolic travel, involved say in Mr. Khrushchev's visit to the United States) and that is to gain information. But to information, it is not necessary to travel; it is necessary only to signal. And the signal has one great advantage over every possible means of travel; information can be transmitted, as no travel can be carried out, at the speed of light itself. The maximum rate of information gain will be obtained from system which transmits, not things or people, but signals at speed of light. Therefore, I think that, perhaps after a few temporary explorations in the near neighborhood, these ε ×107 stars around whom these superior fellows are now living, have long been in intercommunication, over splendid channels of high complexity, using light-velocity signals, carried by fields of some sort, probably electromagnetic (although for all I know they may use neutrinos). The question is: Are they interested in doing anything besides that? There is not much more science left to do at their level, if one dies, say, stellar evolution. What is still interesting is clearly experience of our fellows, because we know that the most complex and the most unpredictable of these forms of complexity are the things that we plotted at the far right of Fig. 1, in the cultural area. What are the novels? What are the art histories? What are the anthropological problems of those distant stars? That is the kind of material that these remote philosophers have been chewing over for a long time. Do they want to know about the Earth? I would say if there are many of these stars, if ε is a largish number, comparable to 10-4 or 10-3, then they do not specially want to know, because they have already seen many new societies emerge. But there may be, however, a little corner of interest still retained, and there are many societies who might be seeking. If I may risk a somewhat frivolous statement, I will say that our earth is not the concern of the great enterprises of knowledge among those far societies, or even of their great enterprises of art; rather, it is the activity of a Department of Anthropology. They may well maintain a certain small subsidiary interest in looking around for new entrants into their great community. Of course, if there are very few of them, if ε is a number comparable to 10-6 or 10-7 then they will be strongly interested in finding us; but they will likely live very far away; the means of contact will be difficult, and even very advanced civilizations will have a hard time making many round trips across the galaxy in search of this curious planet. They would therefore try electromagnetic signalling as the simplest means to call the attention of even the most primitive follows to what is going on. We need not look for sophisticated means of signaling. If we wish to land on the Queensland coast near Port Darwin and communicate with the Australians (I do not mean the Australian astronomers from Sydney, but the aboriginals), we would hardly set up a TV station and broadcast a program. We would rather use some simple audible means, like a steam whistle and a drum. Then these people who are sure to have that kind of communication will come to see what it is we have to sell, give, or trade, or what news we have to spread. So it is with the civilizations of the universe. Their Department of Anthropology will maintain primitive signaling devices meant to catch those people who cannot do very much better. The anthropologists will feel that it would be nice to see how the primitives could enter their interesting society. Here the points of view diverge. I will mention the opinions of three different authors on this subject. First, Professor R. Bracewell, who has what appears to me a rather tendentious scheme, not so good as one that will come later. He asks: How would I go about this? I would dispatch automatic probe ships to every plausible solar system in the neighborhood, to idle about each solar system like satellites. They would listen; when they heard radio signals or TV debates from their near neighbors, these satellites would mount up a big antenna and report home. He says, moreover, that such satellites would try to encourage communication directed at themselves by echoing what they heard; e.g., if they heard dot-dash they would echo back dot-dash; if they heard a commercial, they would echo back a commercial. This, it seems to me, is a frightening degree of pessimism! I do not think that any drone in orbit would simply echo. Mr. Bracewell does point to the fact that echoes of mysterious origin are well known. They were unmistakably heard 25 or 30 years ago. Nobody knows their origin for sure. He thinks perhaps these were drone-orbiters echoing back to show they heard us. I would not, myself, build such unintelligent orbiters. If I heard a signal I would send back, not an echo, but something unmistakably meant to attract attention. I agree his is a possible scheme. It does not depend on the abilities of the local people to do anything very good. If they can reach a neighboring drone by accident, that would be enough to tip off the news of their existence. However, it is very expensive to maintain the drones; to maintain them in space is perhaps relatively easy, but to maintain them in time, against the erosion of space, is very difficult, since they must sit in orbit for millions of years before they have the expectation of hitting the evolution of science. I think it is better to mount signal beams at home, beams of a simple and unmistakable kind, directed preferentially toward those points where we think listeners may sooner or later arise. The beams should be of a kind best suited to attract attention, and to carry the information over the distance of galactic space. Figure 6 is a demonstration of the kind of transmissions which might succeed in space. Graph: Transmission of Signals vs. Frequency Fig. 6 Here you see the transmission in percent plotted all the way from very slow frequencies, like turning on and off light switches, up to gamma rays. We notice the two famous windows through the Earth's atmosphere explored by terrestrial astronomers. Here we have, at the bottom of Fig. 6, the absorption of interstellar space.There are two very wide windows, but the ultraviolet and soft x-rays are cut off by the atmospheres of planets which will also cut off radiation in the millimeter and the decameter ranges. If we look at these plots, then, it seems likely that we will want to use one of these windows which we ourselves find, either in the far gamma-ray region or in the radio region, or possibly in the visible. Certainly no one will use the uv and soft x-ray region. The designer of such equipment will choose some optimum. On which of these frequencies is the random noise of space most serious? Since he knows the conditions in space, he will choose rationally. If we too understand his rationale, we can predict his design. There are two important kinds of noise in space. We know the visual Milky Way and the radio Milky Way. We are looking at a sunlike star, because the planets of life are in orbit around such stars. It turns out that the noise from the star itself both in the visible and the gamma-ray regions is high, and the much more plausible channel appears to be in the radio region. Now in the radio region, at very low frequencies, the sky is very bright; at very high frequencies, the sky becomes dark, but the stars become bright. Therefore, there is an optimum for simple receivers, which are not capable of resolving star from sky, namely the broad intermediate radio-frequency region, somewhere near a few thousand megacycles (Fig. 7). Graph: Radio Noise Power vs. Frequency Fig. 7 If we had to look at random for such signals, we would be searching indefinitely. But there is right here a unique frequency, as everyone knows, the one major spectral line in the radio band, the 21-centimeter emissions from neutral hydrogen atoms. This line, at 1420 megacycles, or 21 centimeters, is a frequency for which there must be sensitive receivers in use on the part of anyone who would understand the nature of galactic space and matter in it. I suggest that for a signaling distance between ten and a few hundred light years, this channel remains indispensable. If you want to communicate to someone who does not know you are sending, you usually choose a frequency near the frequency he is already prepared to listen to. If I want people to listen to my illegal radio station, I always choose a frequency close to the frequency of the official broadcasting station, because I know listeners at that frequency can become listeners of a frequency close to it. They will hear me call from the Sierra Maestra. That is exactly what these remote people would do (though I would welcome detailed historical or social study of how one makes signals known to persons who do not expect to find them). Therefore, set at 1420 megacycles. Look as Dr. Frank Drake very courageously did last year, with his rather small mirror but good receivers, at a few nearby stars, to see if he was lucky enough to pick up the ethnological beams from Tau Ceti or Epsilon Eridani. He was not successful, but one can not expect to be so lucky the first time. We cannot expect to have neighbors as close as twenty light years. Maybe we will have to go to a thousand years if we are to find any at all. This is the Ozma project. I would very much like to lend my support to this investigation. I do not think it in the least foolish. I think it is worthwhile. I feel that there is no more philosophical or practical conclusion to be derived from astronomy than the conclusion that such signals would immediately bring. I should like now to devote only a few paragraphs to suggest how the code would be sent. Writers often say it is indispensable for communication that the partners have something in common. Communication is not possible between completely isolated systems. This is indeed true, but it is only a tautology. Communication is possible only when there is something in common, but there is always one thing in common whenever there is communication; namely, the signal. A signal is by definition some common physical properties of the transmitter and the receiver. Therefore, by denoting in the signal itself we can make communication; we can invent a language, so to speak, by pointing. What we point at is not some other object. We point at and with the signal itself. How would I point? I am now going to present a little experiment. I am going to pretend that I am communicating to others without the use of language. I want to formulate it very simply, in a few moments, but realize that in fact this project could employ cryptographic computing machines, and many clever people; then, even a much more difficult problem could be solved in no time at all. The decoding probably would be relatively trivial. I can go a long way in three minutes. I will show that I can send signals which would elicit guaranteed response. But I cannot in this restricted space do anything very abstract. I must be allowed a few ground rules. Since the receiver would in reality be getting pulses at the rate of ten thousand or more a second, he would have much more information than I can illustrate. Therefore, I will use symbolic boxes, which I will call "A," "B," "C," and so on (Fig. 8). Graph: Hypothetical Pulse Messages Fig. 8 These lettered boxes stand for many pulses in a certain repeated pattern. Any pattern will do; if the receiver hears, say 200 pulses, then the same series comes again, I will call the whole series of 200 pulses pattern "A" (Fig. 8a). 1 will use other distinct patterns "B," "C," and so on. From these we will infer what the statements mean and what the meanings of the patterns are. Figure 8 represents the voltage on the output of the receiver. In Fig. 8.1 is the first continuous sequence. Imagine the sequence occurs a few thousand times in a few seconds. The next sequence (Fig. 8.2) too goes on a few hundred times or a few thousand times. Do I have any takers for what I would mean by A and B? It is, of course, clear that "A" is plus, "B" is identical with equals. I can skip the rest. I could have gone through all the algebraic symbols in the same way. When I have zero, minus, equals, I have no trouble signaling multiplication and division. Now we receive another pattern block X and the pattern block for equals, followed by a series of pattern blocks and numerics. We evaluate this series and we find the series says 3.14159265358979323846. If X equals that series, what then is "X"? We have a name for it. X equals pi. Other series are then transmitted to us and each of them defines the number pi. These people have shouted at us for many seconds, "pi, pi, pi, pi," using infinite expansions. Then we get the following signal. A narrow pulse, a long time with nothing, and another high narrow pulse. The next signal, the high narrow pulse, but the same time elapsing before another high narrow pulse. Again a high narrow pulse, and then a little pulse in the middle. Then a high narrow pulse, and then two spaced pulses in the middle, like Fig. 9, and then more and more of such spaced pulses. Pulsed Signal with Pictorial Message Fig. 9 Now, this is a curious thing. I will give a slight hint. In the first pictures, we saw pulses whose numbers changed, but their spacings showed no interesting features. Now, these pulses are distinct. There is a constant numerical pattern. Here is one pulse, then two, two, two, two, hundreds of these pairs of pulses, but spaced more and more. Of course I get the cryptographers to work. They begin to do all kinds of operations on these spacings. Maybe if the group is clever, after a while, some physicist plots up the way these spaces vary with the number of these high pulses. Soon they start coming together again as in Fig. 9. Then the signal shouts "pi" at us again, and it is all algebraic forms, pi equals, pi equals, and the whole rigmarole starts again. What then is going on? As you have guessed, it is a circle. We now know the TV code. Of course, they may not scan linearly. Maybe they scan in logarithmic spiral. It makes no difference to the method. As long as they supply us with a simple geometric pattern and some algebraic clue to it, we cannot take very long to make out the nature of their scanning raster. Once the television pictures enter, I retire from the field in favor of linguists, language teachers, and elementary-school teachers. Can pictures alone convey adequate information? Teaching a language seems not at all hopeless. Guessing what will be in these pulses is not very profitable. We had better look for them, and not merely guess. (H. Freudenthal of Leiden has done the whole job even without pictures, by symbolic logic alone—in my view, too hard.) There is, at least, no intrinsic difficulty in communicating, to the level of being able to send decodable two-dimensional scans (even three-dimensional if we like) which a little bit of algebraic ingenuity and geometric intuition can lead anybody to decode. We will have, say, two weeks of pulses to work on; it is not a hard problem. Then we would see displayed before us animated films, so to speak, of whatever it was they wanted to teach us. I do not think it would be very long before we would have a rudimentary language. This is the first communication, which would last a few years. Of course, we cannot expect to answer in a few years. We would be doing at least pretty good secondary-school work before we could hope to have acknowledged the first signal. Someone said to me, "This would tend to divide scientific questions in two kinds: Those that can be answered on earth within twice the transit time, and those that should be put on this channel—very much in demand—and sent off to Them to get the answers." I have spoken too long, but I think that one can demonstrate the probable existence of these beings, in what number I do not know. Their communications are most plausible; and it is even likely they will try to communicate with us. We should listen only, and not yet try to do anything more ambitious. It is worthwhile at least mentioning the problem of communicating to the Andromeda galaxy, that neighbor collection of a hundred billion stars. It is very hard for me to believe that nowhere in its great disk, containing a hundred billion stars, 250 million of them likely to have planets like terrestrial ones, is there any scientifically competent civilization. I can think of only one or two ways to signal them, which sound far beyond the capacity of men. Maybe one of the stars can be modulated by interposing an opaque screen. It would have to weigh about 1020 grams (the mass of a comet), distributed in micron-size particles over a five-degree zone of a sphere surrounding the star, and moving in an orbit like the orbit of a planet. If this could be modulated every six months or so, taken away and put back again, or changed to affect the interstellar intensity, we could make it beam a series of algebraic equations at us. Perhaps in that remote galaxy, some patient signalers have for fifty million years tried to modulate a star. These ideas are real ones, and not meant wholly lightly. I should like to conclude with that impression. I think I am not producing science fiction, but legitimate speculation of demonstrable plausibility. I should like to close with an early reference to these ways of thought. Teng Mu, a scholar of the Sung Dynasty in China, wrote this. I cannot close more fittingly than by reproducing the words now seven hundred years old, of a man who thought as we think, but who lacked the technical capability to verify in real life what his imagination was capable of foreseeing: Empty space is like a kingdom, and earth and sky are no more than a single individual person in that kingdom. Upon one tree are many fruits, and in one kingdom there are many people. How unreasonable it would be to suppose that, besides the earth and the sky which we can see, there are no other skies and no other earths.