In the first article of the first issue for 1993 of The Astrophysical Journal, Allan Sandage of the Carnegie Observatories in California presents a new technique for measuring the Hubble constant, Ho. One need not read the paper to know his result, for the title proudly begins ‘Ho = 43 +- 11. . .’
But even before the paper was published, other astronomers attacked it – the latest skirmish in the long and bitter war over the size of the Hubble constant. The war is hard-fought because astronomers’ estimates for the age of the Universe depend on it.
Because the Universe is expanding, a distant galaxy recedes from us faster than a nearby one; the Hubble constant expresses how much faster. If its value were 43, as Sandage found, a galaxy 1 megaparsec farther away than another would recede from us 43 kilometres per second faster. (A megaparsec is 3.26 million light years.)
Advertisement
The Hubble constant is the clue to the distance of far-off galaxies. If the constant is 43, a galaxy with a 1 per cent red shift in its spectrum – the apparent ‘stretching’ of the light waves coming from it, caused by the expansion of the Universe – must be 230 million light years away. But other astronomers believe the Hubble constant is twice as large as Sandage’s figure. The unresolved controversy means astronomers must allow for an uncertainty factor of 2 in the distance to any except the nearest galaxies. A value for the constant of 86 would put the galaxy with a 1 per cent red shift only 115 million light years away.
The Hubble constant has a direct bearing on the age of the Universe because the faster it is expanding, the less time it must have taken to reach its present size since the big bang, and so the younger it must be. A high value – above about 70 – suggests the Universe is younger than its oldest stars, a logical contradiction that would destroy the big bang theory.
The Hubble constant was first measured in 1929, by Edwin Hubble himself, though huge errors in his data gave him a value of over 500. During the 1960s and 1970s Sandage and the Swiss astronomer Gustav Tammann found a value of about 50.
Born in Iowa in 1926, Sandage was seen as Hubble’s heir: he worked with him while in graduate school at the California Institute of Technology and continued Hubble’s work after he died in 1953. Sandage became a senior member of, among others, the American Astronomical Society and the National Academy of Sciences, and in 1967 received the Royal Astronomical Society’s gold medal.
But in 1976 Gerard de Vaucouleurs, a professor of astronomy at the University of Texas in Austin, dared to say that the value was 100 – twice the prevailing value accepted by nearly every astronomer. The war proper began.
De Vaucouleurs was the outsider. Born in France in 1918, he gained a doctorate in physics from the University of Paris and in the early 1950s briefly looked after the science programme broadcast by the BBC’s French-language unit. He first crossed Sandage’s path in the late 1950s, when in an apparent misunderstanding he published some photographs of galaxies borrowed from Sandage, who had thought they would be used just for measurements.
Shifting positions
Today, the two are still at odds. Since de Vaucouleurs’ challenge, the fronts in the war have shifted, but a factor of two still separates the opponents. Proponents of a low value for the constant now find values in the upper 30s and 40s as often as in the 50s, whereas the ‘highs’ favour a value in the 80s. In the past five years, however, many younger and formerly uncommitted astronomers, using new techniques, have swung towards the ‘high’ side, supporting de Vaucouleurs.
In principle, the Hubble constant can be determined simply by comparing a galaxy’s red shift with its distance: the farther away the galaxy is, the smaller the Hubble constant. Unfortunately, while measuring red shifts is easy, determining distances to galaxies is not. Furthermore, the galaxies must be distant enough to be part of the so-called ‘Hubble flow’ caused by the expansion of the Universe; otherwise, the gravity of our Galaxy and others nearby perturbs the observed motions. Most of the galaxies whose distances are agreed upon lie within 10 million light years of Earth, whereas galaxies in the Hubble flow are much farther away. Consequently, astronomers construct different methods to measure distances of galaxies in the Hubble flow. Sandage’s 1 January paper gives nine methods that he considers reliable. All of them, including his new one, give large distances for galaxies in the Hubble flow and therefore low values for the Hubble constant – between 37 and 55.
But in possibly the longest paper ever written on the subject (excluding a 1985 book by Michael Rowan-Robinson), in the August 1992 issue of Publications of the Astronomical Society of the Pacific, George Jacoby of Kitt Peak National Observatory in Arizona and colleagues combine weighted results from seven different methods of observation and reach the opposite conclusion: the galaxies in the Hubble flow are closer than Sandage says, and the value is 80 +- 11.
But how can both sides use ‘reliable methods’ to reach such widely different conclusions? And what constitutes a ‘reliable method’? The Canadian astronomer Sidney van den Bergh frequently writes reviews that try to sort out the mess, and is aligned with neither side. One that he wrote last year began with a quote from Mark Twain: ‘The researches of many commentators have already thrown much darkness on this subject, and it is probable that, if they continue, we shall soon know nothing at all about it.’
However his latest review, in October’s Publications of the Astronomical Society of the Pacific, surely sent joy through the high-Hubble side. It evaluates over a dozen methods for determining the Hubble constant, and finds a value of 76 +- 9 – which is consistent with the result of Jacoby and his collaborators. Does this mean that the value is now known? Not necessarily: after all, if enough people make enough determinations, two of them are bound to agree sooner or later.
Interestingly, it was van den Bergh who attacked Sandage’s new method. In his 1 January paper, Sandage compares the diameter of the nearby spiral galaxy M101 with the diameters of more distant spiral galaxies to deduce those galaxies’ distances and thereby the Hubble constant. This method would work if all spiral galaxies had the same diameter. But van den Bergh’s review has a devastating photograph of two apparently similar spiral galaxies that differ in true size by a factor of three – rendering Sandage’s technique useless.
Every war thrusts into prominence places that few people have heard of before. Last year, the Hubble war gave us IC 4182. A small galaxy from an obscure constellation, IC 4182 is a fierce battleground for astronomers. In 1937 a Type Ia supernova exploded there. Type Ia supernovae are special because they arise only in a particular kind of star, and so might all have the same intrinsic brightness.
Unfortunately, there have been no Type Ia supernovae recently in any galaxies with a known distance. But IC 4182 is fairly close. If its distance could be determined we would know the intrinsic brightness of the Type Ia supernova that occurred there – and thus of all Type Ia supernovae. From this it would be possible to work backwards and measure the distances to more remote galaxies that have spawned Type Ia supernovae. These distances would then yield the Hubble constant.
How far to the stars?
In 1982, Sandage and Tammann found that IC 4182 was 14 million light years away, giving a Hubble constant of 50 +- 7. Last year, however, Michael Pierce at Kitt Peak and his colleagues found the distance to be only 8 million light years. As they reported in Astrophysical Journal Letters of 10 May 1992, this gives a Hubble constant of 86 +- 12 (91av, Science, 4 April 1992).
Sandage’s team quickly fought back, firing off a press release last summer that reported a new distance of 16 million light years for IC 4182 and a Hubble constant of 45 +- 9 (91av, Science, 18 July 1992). Their paper appeared in Astrophysical Journal Letters of 10 December 1992. Cleverly, it contained no references at all to Pierce’s work, so preventing it from racking up gains in the Science Citation Index – one yardstick of a scientific paper’s importance. The battle over IC 4182 continues: the high-Hubble side now claims that the galaxy contains so much dust that it looks fainter and farther away than it really is.
Like all wars, this one abounds with propaganda. The high-Hubble side will tell you that the question is no longer whether the value is 50 or 100, but whether it is 80 or 90. Those on the low-Hubble side talk about ‘Malmquist bias’, a statistical effect that can make distant galaxies seem nearer than they are, causing overestimates of the Hubble constant.
And as in most wars, the media frequently favours one side or the other. A review of 1992 in Science News mentioned Sandage’s work on IC 4182 and assured readers that ‘researchers moved toward pinning down the much-debated numerical value of the Hubble constant.’ A four-page article in November 1992’s Physics Today also adopted the Sandage-Tammann line and gave only a single dismissive paragraph to the other side.
Curiously, while many astronomers have been finding high values, the French astronomer who once championed just that position has remained strangely silent. But later this year, de Vaucouleurs will report in The Astrophysical Journal a new way to measure the Hubble constant. By counting the number of globular star clusters around a galaxy, he says, it is possible to derive the galaxy’s intrinsic brightness, and hence its distance. He finds a Hubble constant of 85 +- 4.
But before de Vaucouleurs’ paper appears, Sandage will have launched another missile, in which he uses the diameter of M31, the Andromeda galaxy, to determine the Hubble constant. At least, that is what can be inferred from the paper’s title. Although I requested a preprint, all that came back was a curt letter from the publications editor of the Carnegie Observatories. ‘Thank you for your request,’ it read. ‘Preprints were not available for this paper; however, it will be published in the February 20, 1993 issue of The Astrophysical Journal.’
This time, Sandage has not announced the result in the title. But if the past is any guide, next week’s paper will conclude that the Hubble constant is around 45 – with an error of +- 10.
