by George Rhee
Herschel carried out his work using a reflecting telescope with a 47 cm mirror, although he earned fame and fortune by discovering Uranus with a telescope with a 15 cm diameter mirror. The method of star counts was also used by the Dutch astronomer Jacobus Kapteyn (1851–1922) to build a model of the galaxy. Kapteyn searched for evidence of interstellar absorption but found none. According to Kapteyn’s model, the galaxy was about 45,000 light years in diameter, with the Sun lying about 2,100 light years from the center. Kapteyn’s work involved collaboration between astronomers in many countries. It is one of the very positive aspects of the scientific enterprise that people from different countries and different cultural traditions can work together to achieve common goals.
The Realm of the Nebulae
Simon Marius (1573–1624) published the first observations of an island universe made with a telescope. He reports observations made on 15 December 1612 of what we now know is the Andromeda galaxy, a large spiral galaxy similar to our milky way galaxy. Marius was unaware of the nature of Andromeda but he gives an accurate description of its appearance seen through a small telescope. About 100 years later, an anonymous author believed to be Halley published “An Account of several Nebulae or lucid Spots like Clouds, lately discovered among the Fixt Stars by help of a Telescope”. The fact that these nebulae remained fixed relative to the stars suggested they were very far away from the solar system.
In 1845, William Parsons, the Earl of Rosse, built a reflecting telescope with a 72-inch mirror. With this instrument he established that some of the nebulae have spiral structure. The nature of these nebulae was by no means clear. Were they spiral-shaped clouds of gas inside our own galaxy, or were they huge galaxies in their own right, appearing small and blurry only because they were so far away?
Clues to resolving this problem were provided by William Huggins (1824–1910). Huggins was a pioneer of spectroscopy, a technique astronomers use to analyze the detailed color of the light emitted by stars and gas clouds. A spectrum consists of the brightness of the light detected at different colors. The elements of the periodic table leave their imprint on the light in the form of excess absorption or emission of light at certain very specific colors or wavelengths. Huggins and Miller in 1864 showed that there was a similarity between spectra of stars and the spectrum of the Sun which Fraunhofer had published in 1814. Huggins next took spectra of the nebulae that Herschel called planetary nebulae. The spectra he obtained were quite different to stellar spectra. The planetary nebulae spectra contained strong emission at certain specific wavelengths. It took until 1927 to realize that the emission was coming from twice ionized oxygen. It was clear to Huggins that the emission from planetary nebulae was due to gas and not distant star clusters. When Huggins took a spectrum of M31, the Andromeda nebula, he found that its spectrum resembled that of star clusters. The result implied that maybe this nebula was a distant collection of stars. In fact Huggins had obtained several spectra of nebulae that were different to planetary nebulae in that they did not show strong emission line of oxygen and hydrogen. This work implied that the spiral nebulae discovered by the Earl of Rosse were indeed distant aggregations of stars.
The issue of the distance to the nebulae was finally resolved by Edwin Hubble (1889–1953), who on the 6th of October 1923, discovered a variable star in the Andromeda nebula. Allan Sandage (1926–2010), Hubble’s student, who was born 3 years after that discovery, kept the photographic plate on which Hubble wrote “VAR!” in red ink next to the variable star he discovered. The plate is shown in Fig. 1.4. The luminosity of a variable star can be inferred from its period. It is as if the wattage of a light bulb in a light-house could be inferred from how fast the light appears to turn on and off. These stars can therefore be used to measure distances.
Fig. 1.4The image on the left is taken from the ground showing the Andromeda galaxy. The image at the bottom right is Edwin Hubble’s discovery image of a cepheid in the Andromeda Nebula. These variable stars were crucial to determining the distance to the Andromeda Nebula and showing that it was not part of the Milky Way galaxy. This same cepheid star was observed many years later by the Hubble Space Telescope (see image at top right). The boxes in the left hand image show the location of the two images on the right hand within the galaxy (Credit: E. Hubble, NASA, ESA, R. Gendler, Z. Levay and the Hubble Heritage Team)
The most useful variable stars for measuring the distances to galaxies are called Cepheid variable stars. One can determine the luminosity of a Cepheid star to an accuracy of about 10 % by measuring the time period over which its brightness varies. The Cepheid star that Hubble discovered was clearly in the Andromeda nebula and was at such a distance that it could not be in our galaxy thus proving that the Andromeda nebula lay outside our own galaxy. By the end of 1924, Hubble had discovered 36 variable stars in Andromeda, 12 of which were Cepheid variables. Using those stars he concluded that Andromeda was 900,000 light years away. Today we believe Andromeda to be about 2 million light years away.
In the roughly 70 years between the discovery of spiral nebulae and Hubble’s solution of the problem, there was much discussion about their nature, which culminated in a famous so-called debate conducted in April 1920 at the annual meeting of the National Academy of Sciences held at the Smithsonian Institution. On one side of the debate, was Harlow Shapley (1885–1972) who argued that the Milky Way was so large that the spiral nebulae had to be within it. Shapley believed the Milky Way to have diameter of 300,000 light years. Heber D. Curtis (1872–1942) disagreed, holding that the nebulae were “inconceivably distant galaxies of stars, or separate stellar universes, so remote that an entire galaxy becomes but an unresolved haze of light”. Curtis thought that the Milky Way’s diameter is 30,000 light years. The debate attempted to settle three key questions: 1.What are the distances to spiral nebulae?
2.Are spiral nebulae composed of stars or gas?
3.Why do spiral nebulae avoid the plane of the Milky Way?
The key argument concerning the distances to spirals centered on some measurements made by Adriaan van Maanen (1884–1946) a Dutch astronomer. Van Maanen attempted to measure proper motions of spiral galaxies. By proper motions we simply mean the motion of objects in the sky. Proper motions are caused by the motion of the object itself, as opposed to parallax, which, as we have seen, is an effect caused by the Earth’s motion around the Sun. He compared photographs of a given spiral taken some time apart and detected motion of the spiral nebulae. If Shapley’s model of the Milky Way was correct and other spiral nebulae were of the same size as the Milky Way, van Mannen’s measurement implied that they were rotating close to or faster than the speed of light. To avoid this conclusion, Shapley decided that the spiral nebulae were smaller objects within our own galaxy.
Curtis decided to ignore van Maanen’s data, believing them to be spurious. Curtis was correct: The proper motion measurements were in error. When one looks at the data published by van Maanen in the Astrophysical Journal, they seem highly convincing. It is hard to imagine what observational error could produce the effect that he measured. On reading one of van Maanen’s papers, I noticed, interestingly, that one of the photographic plates he examined was taken by Curtis. It is probable that, as an observer, Curtis had a good feel for the limitations of the data. To measure the positions of blobs of emission in spiral galaxies to an accuracy of 1/30,000 of a degree using photographic plates taken at the turn of the twentieth century sounds like an impossible task. Yet van Maanen claimed to have done precisely this. One moral to be drawn from the so-called Great Debate is that the two protagonists drew more reliable conclusions when discussing data that they were fully familiar with.
As to whether spirals were composed of gas or stars, Shapley argued in favor of the former. Spiral nebulae appeared bluer in their outer portions than in their centers, suggesting they did not consist of stars. Those favoring the view of spirals as galaxies argued that we could not resolve the individual stars because the galaxies were so fa
r away.
Who, If Anybody Won the Great Debate?
Why do spiral nebulae avoid the plane of the Milky Way? Shapley argued that since they avoid the Milky Way plane they must be influenced by it and therefore be close by. Curtis countered by saying that many spiral nebulae exhibit a central belt of obscuring material. If we imagine that the Milky Way also has such a belt and that spirals are external to the Milky Way system, the zone of avoidance could be explained. The spiral galaxies are blocked from our view by the dust, but, in fact, are really actually there. If we imagine the Milky Way as a disk containing dust, then, when we look in the plane of the Milky Way, the dust blocks our view of the galaxies, whereas when we look perpendicular to the plane of our galaxy, we see many galaxies because we do not have to look through as much dust. The Sun looks red at sunset for similar reasons. When the Sun is on the horizon at sunset, the sunlight passes through more of the Earth’s atmosphere than at noon. The atmosphere contains molecules that scatter the blue light from the Sun, making the Sun look redder. Again, Curtis’ explanation was correct.
When looking at the protagonists and the arguments used in the debate we see how confusing science can be. Neither Shapley nor Curtis was 100 % correct. Some key measurements that had been recently published by a reputable scientist (a former student of Kapteyn) proved to be incorrect. Both Shapley and Curtis made reasonable arguments to support their cases. We currently believe the Milky Way to be three times larger than Curtis thought and three times smaller than Shapley believed it to be. To summarize, one might say that Shapley made some good arguments, and came up with the wrong answer; Curtis did not necessarily have the best arguments but got the answer right. For want of a better word, intuition can play a big part in arriving at one’s conclusions when the evidence is incomplete or misleading.
The opinions of both Shapley and Curtis were most reliable when they were discussing observations they had obtained themselves or were familiar with. The moral for astronomy students is to “know the data”. When one is familiar with observations, one has a feel for the experimental accuracy and possible errors that is hard to get by simply reading the literature.
A key issue in the Great Debate was that of the amount of absorption of starlight by gas and dust. Robert Trumpler (1886–1956) was the first to demonstrate the existence of an absorbing medium in-between the stars. Born in Switzerland, Trumpler came to work in the United States, spending most of his research career at Lick Observatory, where he studied star clusters. He could measure their angular size on the sky (by ‘sky’ in this case we mean ‘on a photographic plate’) and used this to estimate their distances, assuming these clusters had similar physical sizes. He found that the more distant clusters appeared to be intrinsically less luminous than the closer ones. He chose not to take this fact at face value but postulated instead the existence of an absorbing medium, which would make distant objects look fainter than they really were. This could immediately explain why the external spiral nebulae avoided the plane of the Milky Way. Spiral galaxies behind the plane of our galaxy are obscured by it. More direct evidence for an interstellar medium came from spectra of binary stars. These stars are orbiting each other and produce spectral lines that shift back and forth in wavelength, reflecting the motion of the stars radially toward and away from the observer. These spectra contained some lines that did not shift at all, suggesting that they were associated not with the stars, but with an interstellar medium. Taking into account interstellar absorption helped explain why astronomers were calculating different sizes for the Milky Way using different methods.
During the late 1920s people were analyzing the motions of stars in our galaxy and concluding that our galaxy must consist of a disk that rotates and a halo that does not. Most of the key insights were provided by Jan Oort (1900–1992). Oort began his studies at Leiden and went on to study in Groningen under Kapteyn. Like Kapteyn, Oort promoted international collaboration and was instrumental in bringing about the creation of the European Southern Observatory. He left his substantial mark on many fields of astronomy, from the study of comets up to cosmology. Oort showed that our galaxy rotates differentially, that is to say not as a solid body. From the study of stellar motions perpendicular to the galactic plane and close to the Sun, Oort showed that, in the solar neighborhood of the galactic disk, twice as much mass must exist gravitationally as can be accounted for by luminous matter. This is known as the missing mass problem–something of a misnomer, since we believe the mass is actually there. The problem is not that the mass is missing but that it does not emit light. The mass is present but invisible because it does not emit light. The missing mass problem is one that pervades all of cosmology, and we shall encounter it again and again. To infer the presence of dark matter one must make use of a theory of gravitation, namely Newton’s theory or its more accurate extension developed by Albert Einstein, General Relativity. This theory has not been tested on the physical scales (thousands to millions of light years) on which we are applying it. There is the possibility that gravity does not follow an inverse square law on these large scales. Nevertheless, most astronomers prefer to admit the existence of dark matter.
The story of the discovery of the structure of our galaxy and the existence of external galaxies is not as elegant as that of planetary motion. This is because a number of key discoveries whose interpretations influenced each other were made in a short span of time. It may also be that we have a more accurate picture of relatively recent events that are better documented than debates taking place several 100 years ago. It is certainly fascinating to look at the actual data that were published in the 1920s and 1930s. As happens today, some of the observational evidence was rather thin at times.
Percival Lowell (1855–1916) founded the observatory that is named after him in 1894. Lowell was fascinated by the idea that there might be life on Mars. In 1901, he hired a man named Vesto Slipher to assist him in his work. Lowell also was interested in the spiral nebulae, because he believed they were possibly solar systems in formation. This hypothesis implied they should rotate, and Slipher used a spectrograph to measure the shift in the wavelength of spectral lines emitted by atoms in gaseous form and hence infer the speed at which these atoms are moving. The wavelengths of spectral lines can be measured in the laboratory and compared with the wavelengths of lines measured with a spectrograph. In 1912 Slipher had obtained a spectrum of the Andromeda nebula and concluded that it is approaching the Earth with a velocity of 300 km per second. This was a surprisingly large speed. The Andromeda nebula is unusual in that it is the only large galaxy that is approaching the Milky Way galaxy. We estimate that Andromeda will collide with the Milky Way in about 4 billion years.
By 1914 Slipher had obtained spectral line shifts for 14 spiral nebulae. These shifts were all toward the red and are commonly known as redshifts. The fact that shifts were towards the red and quite large implied that they had large positive radial velocities. In other words, the spiral nebulae appeared to be moving away from the Milky Way at high speeds. Interestingly, the high recession speeds of the spirals were used as an argument against the island universe hypothesis. Slipher also measured rotational velocities for the spiral nebulae. His results contradicted the proper motion results of van Maanen! Both thought the spirals were rotating, but they disagreed about the rotation direction. We shall discuss these rotational velocities in more detail when we discuss the nature of dark matter in the universe. Slipher achieved worldwide fame for his spectrographic observations of the nebulae and eventually became director of Lowell Observatory. It was under his directorship that the planet Pluto was discovered in 1930 by Clyde Tombaugh.
The Expansion of the Universe
The situation, even after the Great Debate, was one of great confusion. It was by no means clear if the spiral nebulae lay inside or outside the Milky Way. Hubble finally solved the problem by measuring distances to variable stars in spiral nebulae. By 1929 Hubble had obtained distances for 18 of the 46 objects for whi
ch velocities were then available. The most distant of Hubble’s nebulae was moving away from us at 1,000 km per second. Hubble estimated its distance as 700,000 light years, thereby placing this nebula well beyond the edge of the Milky Way, even according to Shapley’s large diameter for our galaxy of 300,000 light years.
The result that was to make Hubble famous was the correlation he found between distance and velocity. He found that the more distant nebulae were moving away from us faster than the nearby ones. The proportionality constant between these two quantities is now known as Hubble’s constant. Hubble was unaware that there are two kinds of variable stars. Because of this, he underestimated the distances to the nebulae by a factor of ten or so.
Hubble does not deserve all the credit for the discovery of an expanding universe. In 1927 George Lemaître (1894–1966) published his model of an expanding universe (Fig. 1.5). He predicted a relationship between velocity and distance in the form that we now credit to Hubble. He went further and used data to calculate a value of the Hubble constant (roughly 8 times larger than the accepted value today). Hubble found the same relation 2 years later using data that had mostly been gathered by Slipher. It is possible that Lemaître’s paper did not have a great impact because it was published in French. Even more intriguing is the fact that the version of the paper which was published in English with the help of Arthur Eddington in 1931 does not contain the estimates of the Hubble constant from the 1927 paper.
