May 4 1959 – The shape of the galaxy

The question of the shape of our star-system, or Galaxy, is obviously of particular importance in astronomy. The first man to give accurate figures for the size of the galactic system was Dr Harlow Shapley, whose results were published over forty years ago. Since that time, Dr Shapley has continued his astronomical work, and has been responsible for various fundamental advances. In the spring of 1959 he visited Britain, and I considered it a great honour when he joined me for the May Sky at Night programme. Dr Shapley is now Director Emeritus of the Harvard College Observatory

On a summer night when the Sun has long since set and the stars are shining brilliantly, one of the most wonderful sights in the heavens is the Milky Way. It stretches across the sky, making a band of radiance which catches the eye at once. As Ptolemy, the last great astronomer of Classical times wrote in his Almagest nearly 2,000 years ago: 'The Milky Way is not a circle, but a zone, which is almost everywhere as white as milk and this, has given it the name it bears. This zone is neither equal nor regular everywhere, but varies as much in width as in shade of colour, as well as in the number of stars in its parts, and by the diversity of its positions; and also because that in some places it is divided into two branches, as is easy to see if we examine it with a little attention.'

As a description of the Milky Way as seen with the unaided eye, Ptolemy's account could hardly be bettered; but it was not until the winter of 1609-10 that any great advance in astronomical knowledge of the familiar milky zone could be made. It was then that Galileo first applied the telescope to the skies, and discovered that the Milky Way is composed of a vast number of faint stars. These stars give the appearance of being crowded closely together, but in fact nothing could be further from the truth. Galileo and his contemporaries knew very little about the shape of our star-system; the first reasonably correct picture of how the stars are arranged was not drawn up until the late eighteenth century, due largely to the work of one man, William Herschel.

Herschel was Hanoverian by birth, but spent most of his life in England. Following his discovery of the planet Uranus, in 1781, he was able to give up his original profession, that of a music teacher and organist, in order to devote his life to astronomy. His main ambition was to solve the secrets of star distribution, and for many years he worked away at the problem, carrying out what he himself called 'reviews of the heavens'. Herschel's telescopes were home­made, and were probably the best of their time. The largest was a reflector with a focal length of 40 feet.

Herschel knew that he could not possibly count all the stars visible with his telescopes, so he decided to count the stars in certain selected regions of the sky. He believed that the apparent brightness of a star was a reasonably good measure of its relative distance from us, so that brilliant objects such as Sirius and Rigel were likely to be closer than fainter stars such as Polaris or the seven members of the familiar Plough; in consequence the regions of the sky containing the most stars would represent the greatest extensions of the stellar system. During the course of his 'star-gauging' he found that the faint stars became unexpectedly numerous near the Milky Way band; the increase in frequency was greater than for brighter stars. This led him on to a scheme according to which the stars were arranged rather in the manner of two plates clapped together by their rims, with the Sun lying near the centre of the system. At once the Milky Way appearance could be explained without the need for any actual crowding of the stars. An observer looking along the main plane of the system would naturally see many stars in almost the same line of sight - one behind the other; at right angles to this main plane, the sky would be relatively barren. Herschel's final conclusion was that the stellar system is shaped like a 'cloven grindstone'.

Many of Herschel's ideas have been confirmed by later research, and his picture of the stellar system is by no means wide of the mark. His main error was in supposing the Sun to occupy a more or less central position; we now know that the actual centre lies at a dis­tance of some 25,000 light-years from us, in the direction of the Sagittarius star-clouds. Yet Herschel's mistake was both natural and inevitable, since in his day there was no means of determining either the distances or the luminosities of the stars. All that could be said with certainty was that the stars were immensely more remote than our near neighbours such as the Sun, Moon and planets.

In 1838, sixteen years after Herschel's death, the German astronomer Bessel managed to measure the distance of a dim star in Cygnus. Other determinations followed, and for the first time the majestic scale of the stellar system became known. This in turn led to accurate estimates of the luminosities of the stars, and some of the results were decidedly unexpected. The stars differ so widely in absolute brightness that apparent magnitude is by no means a reliable guide for judging distance, as Herschel had believed. Of course, some apparently brilliant stars are genuinely near; Alpha Centauri, which shines as the third brightest star in the sky, is little over four light-years away (that is to say its light, travelling at 186,000 miles per second, takes more than four years to reach us), while Sirius lies at only eight and a half light-years, and must also be regarded as comparatively near. On the other hand, Rigel in Orion, which comes seventh in order of apparent brightness, may be over 900 light-years from us. It appears conspicuous not because it is close but because it is exceptionally luminous; accord­ing to one estimate it has 50,000 times the candle-power of the Sun.

Bessel and his contemporaries measured star distances by means of trigonometrically parallax. Their results were good, but the method breaks down for all but relatively near stars, since at great distances the parallax shifts become too small to be measured. Gradually other methods were developed, but it was not until the present century that a completely unexpected discovery provided the means of spectacular advances in knowledge. This was the period-luminosity law of variable stars of short period.

Most stars shine so steadily that their luminosity output does not central position; we now know that the actual centre lies at a dis­tance of some 25,000 light-years from us, in the direction of the Sagittarius star-clouds. Yet Herschel's mistake was both natural and inevitable, since in his day there was no means of determining either the distances or the luminosities of the stars. All that could be said with certainty was that the stars were immensely more remote than our near neighbours such as the Sun, Moon and planets.

In 1838, sixteen years after Herschel's death, the German astronomer Bessel managed to measure the distance of a dim star in Cygnus. Other determinations followed, and for the first time the majestic scale of the stellar system became known. This in turn led to accurate estimates of the luminosities of the stars, and some of the results were decidedly unexpected. The stars differ so widely in absolute brightness that apparent magnitude is by no means a reliable guide for judging distance, as Herschel had believed. Of course, some apparently brilliant stars are genuinely near; Alpha Centauri, which shines as the third brightest star in the sky, is little over four light-years away (that is to say its light, travelling at 186,000 miles per second, takes more than four years to reach us), while Sirius lies at only eight and a half light-years, and must also be regarded as comparatively near. On the other hand, Rigel in Orion, which comes seventh in order of apparent brightness, may be over 900 light-years from us. It appears conspicuous not because it is close but because it is exceptionally luminous; accord­ing to one estimate it has 50,000 times the candle-power of the Sun.

Bessel and his contemporaries measured star distances by means of trigonometrically parallax. Their results were good, but the method breaks down for all but relatively near stars, since at great distances the parallax shifts become too small to be measured. Gradually other methods were developed, but it was not until the present century that a completely unexpected discovery provided the means of spectacular advances in knowledge. This was the period-luminosity law of variable stars of short period.

Most stars shine so steadily that their luminosity output does no alter perceptibly over many thousands or even millions of years. Fortunately for us, the Sun belongs to this category. There are, however, some stars which fluctuate in brightness, rising to maxi­mum and then falling to minimum before starting to increase once more. Particularly interesting is Delta Cephei, which lies in the northern hemisphere of the sky and never sets in the latitude of Britain. The magnitude changes between 3-7 and 4-3 in a period of five days nine hours, and the period is absolutely regular, so that it may be determined to within the fraction of a second. The light- curve is not entirely smooth; as the increase to maximum is steeper than the subsequent drop, but we can always tell how bright Delta Cephei will be at any particular moment. It is by no means unique; other stars behave in the same way, and are hence termed Cepheid Variables.

What makes the Cepheid’s so important is that their period is a measure of their luminosity: the longer the period, the more lumi­nous the star. Hence a Cepheid of period five days nine hours will have the same absolute brightness as Delta Cephei itself, while a Cepheid of period ten days will be considerably more powerful. By studying the period of a Cepheid, we can therefore find out its real luminosity; its apparent magnitude is easy to measure, and hence its distance may be determined, so that these convenient variables act as our standard candles in space. The exact cause of the period-luminosity law remains uncertain, but there can be no doubt of its validity.

Variables known as RR Lyrae stars, after the prototype object, have shorter periods than classical Cepheid’s - less than an hour and a half in the case of one star, CY Aquarii - and appear to be of more or less uniform brilliancy: about ninety times that of the Sun. As soon as we observe an RR Lyrae star, we can therefore make a good estimate of its distance. Many such variables are found in the globular clusters, vast spherical groups of stars found here and there in the sky, and they were originally known as cluster- Cepheid's. The name has now become obsolete, partly because the stars are not true Cepheid's and partly because some of them - including RR Lyrae; itself - are not members of globular clusters.

The globular clusters themselves are of vital importance in our studies of the shape of the stellar system or Galaxy. The average globular contains perhaps 100,000 stars, many of which are far more luminous than the Sun; yet the clusters appear as faint objects in our skies. Only three are bright enough to be seen without a telescope. Two of these, Omega Centauri and 47 Tucanae, lie too far south to rise in Britain; the third, Messier 13 Herculis, may be seen on a clear night as a very dim, misty patch, while a telescope of moderate power resolves it into a glorious sphere of stars. Alto­gether, about 100 globulars are known.

Since the globular clusters contain highly luminous stars, and yet appear faint, they must be extremely distant. This had been obvious for many years, but until the revelations concerning short- period variables it had been impossible to find out how far away they really were. The 'standard candles' came to the rescue in no uncertain fashion. As we have seen, globulars contain RR Lyrae variables; the distances of the RR Lyra; variables may be deter­mined because of their uniform luminosities; and hence it was at once possible to judge the distances of the globular clusters in which the variables lie. Omega Centauri has proved to be 22,000 light- years from us, while Messier 13 Herculis is 34,000 light-years away. Most of the others are more remote still.

In the years following the end of the first World War, the American astronomer Harlow Shapley made a determined attack on the whole problem. It had long been known that the globulars are not spread uniformly around the sky; most of them are in the south, in the region of Scorpio and Sagittarius, so that the distribu­tion appears to be lop-sided. Shapley assumed that the globulars form a kind of outer surround to the main star-system, and by distance determinations he was able to show that this is in fact the case. The mystery of the lop-sided distribution was solved; the Sun, with its family of planets, lies well away from the centre of the Galaxy, so that we have an unsymmetrical view.

Shapley's work led to the first really reliable picture of the scale and shape of the Galaxy. The central nucleus lies, as expected, in the direction of the Sagittarius star-clouds; the whole system measures approximately 100,000 light-years from end to end, and the Sun is 25,000 light-years from the centre, while the thickness of the system is some 20,000 light-years. (Of course, these figures are bound to be uncertain, and different authorities give different values, but they are certainly of the right order.) Surrounding the main system is the 'galactic corona', a sort of outer skeleton of globular clusters and individual stars; in this corona the ratio is roughly 100 stars to each globular. It is believed that the total number of stars in the Galaxy is about 100,000,000,000.

Our galaxy seen edge on. The cross marks the position of the Sun.

If we could go far out into space and look at the Galaxy 'edge- on', we would see a flattened system with the bulge of the central nucleus showing up very noticeably. If however we could look from right angles, it would become clear that the Galaxy is spiral, not unlike a tremendous Catherine-wheel. A spiral shape has been suspected for many years, but an entirely new branch of astrono­mical research was needed to prove it.

Telescopes show us other galaxies, lying at distances of many millions of light-years; the most conspicuous, Messier 31 Andro­meda: has been found to lie at about 2,200,000 light-years. (Here again the short-period variables, this time the classical Cepheids, provided the means of making a sound estimate.) Many of these galaxies are spiral, and there would be nothing surprising in finding our own system to be of the same shape. But while suspicion is one thing, proof is quite another; we cannot see our Galaxy from 'outside'.

Indications of spirality were obtained by what are generally regarded as conventional methods. W. Baade, in America, drew attention to the fact that there are two distinct types of stellar 'populations' - the first (Population I) in regions where there is considerable interstellar gas and dust and where the brightest stars are very white and luminous; the second (Population II) in regions relatively clear of interstellar material, and where the leading members are red supergiants. It appeared that globular clusters and the centres of external galaxies were mainly Population II, while the spiral arms of galaxies were mainly Population I. By plotting the distribution of the highly luminous white stars in our own system, inconclusive signs of spiral structure appeared.

All this was uncertain, but a solution was to hand. We know that of all interstellar gas, hydrogen is much the most plentiful. It tends to collect into huge clouds, and is very cold, with a tem­perature of perhaps — I50°C. Naturally it is very rarefied, and there are on an average less than ten atoms per cubic centimetre, which judged by our everyday standards is equivalent to what scientists term a vacuum; needless to say, normal telescopes will not show these hydrogen clouds at all. In 1944, however, the Dutch scientist van de Hulst worked out that the hydrogen should be emitting radio energy on a wavelength of 21 • 1 centimetres. He believed that this energy would be detectable by means of radio telescopes, and six years later his prediction was brilliantly con­firmed.

Radio astronomy is still young, but already it has made remark­able strides, and the detection of the 21-centimetre 'noise' is one of its major achievements. It has become possible to measure the distances and velocities of the interstellar hydrogen, and there can no longer be any serious doubt that the Galaxy is indeed a spiral, with the Sun lying near the inner edge of one of the arms. To an outside observer, then, it seems that the Galaxy would appear as a rather loose Catherine-wheel, very much like many of the remote systems that we ourselves can see in the sky.

Like a Catherine-wheel, too, the Galaxy is rotating. The Sun shares in the general motion, and seems to take roughly 225,000,000 years to complete one journey round the centre. This period has been aptly termed the 'cosmic year', and it is interesting to note that the Sun has made only one revolution since the time when the lords of the Earth were dragonflies and amphibians, with mammals lying far in the future. Man himself proves to be very much of a newcomer to the scene.

Our views of the universe have changed out of all recognition during the last two centuries. We know now that the Sun is an unimportant star, one of 100,000,000,000 such bodies in our stellar system; it occupies a position of no significance whatsoever, and the Earth itself is puny beyond all understanding. Yet we are justified in being proud of our achievements, and it has been no mean feat to prove that the Galaxy in which we live takes the form of a whirling spiral.