The definitive study of the herd instincts of astronomers has yet to be written, but there are times when we resemble nothing so much as a herd of antelope, heads down in tight parallel formation, thundering with firm determination in a particular direction across the plain.
—J.D.Fernie
Once he had gotten used to the idea, the enthusiastic Dr. Shapley embraced the new universe with more gusto than the reserved Dr. Hubble would let himself show. Sometimes it almost seemed that Hubble was being willfully perverse. Hewing to tradition, he steadfastly reserved the term “galaxy” for the Milky Way, continuing to refer to Andromeda and the other island universes as nebulae—“extragalactic nebulae,” to be exact. Etymologically this might be correct: “galaxy” comes from the Greek word for milk. But Shapley, like everybody else, quickly generalized the term, calling all the island universes galaxies. The bandwagon had almost taken off without him. Now he was back on board.
Though Shapley’s old view of the Milky Way as the sole galaxy seemed quainter by the year, he appeared to be right about its enormous size. After all, the same calibration of the Cepheids—Shapley’s curve—that had revealed a Milky Way a whopping three hundred thousand light-years from end to end had also been used by Hubble to plumb the distance to Andromeda. From there he had extrapolated outward, using redshift to measure a hundred million light-years into space. If Hubble was right about the size of the universe, it seemed, Shapley must be right about the size of the Milky Way.
And that led to a dilemma. With their new measuring techniques, astronomers could now judge how large other galaxies were. Just take the apparent size and adjust for distance to get the true diameter. Simple enough. But the results of the calculations were disconcerting. None of the galaxies came out to be anywhere near the size of our own. Andromeda measured only a tenth as wide, while the others ranged from a mere thousand light-years to perhaps seventy-five hundred light-years across. The term “island universe” took on a new meaning with the emphasis shifted to the first word. If these distant spirals were islands, Shapley contended, then our own galaxy was a continent.
A few years earlier it would have been acceptable for people to find themselves living in the biggest galaxy around. But perceptions had changed. Shapley had moved the sun from the center of the galaxy, and Hubble had moved the galaxy from the center of the universe. This reversal of perspective was becoming so ingrained that it was a gold standard by which astronomical ideas were judged. If a theory or observation seemed to suggest that we, the observers, happen to occupy an exalted place in the heavens, then it was probably wrong.
Of course it was possible that chance had conspired to put earthlings in a special location. But other discrepancies were harder to dismiss. If the big bang theory was correct, then the size of the universe was an indicator of its age. The larger it was, the longer it had been expanding since the primordial explosion. If the galaxies at the circumference were two billion light-years away, as recent measurements had suggested, it must have taken them two billion years to get there.
Two billion years seemed like a reasonable number. The Earth, however, measured using the technique of radioactive dating, came out to be four billion years old—twice as ancient as the universe that contained it. Something somewhere would have to give.
2
When kinks like these develop in the fabric of knowledge, the fault might lie anywhere in the weave—the result of bad data or a false assumption, the malfunction of a dumb machine or of a human brain. People may see things that turn out to be chimeras. Or miss what is right in front of their telescopes.
At Lick Observatory in California, a Swiss-born astronomer named Robert Trumpler had been studying structures called open clusters in the Milky Way. These stellar aggregations— Pleiades is an example—are smaller and more loosely packed than the globular clusters Shapley used to map the galaxy. Comparing each cluster’s true brightness with its apparent luminosity, Trumpler calculated its distance. With this information in hand, he could convert the cluster’s apparent diameter into its true diameter—how big it really was.
After measuring a number of them this way, he was forced to a bizarre conclusion: the farther a cluster was from Earth, the larger it appeared to be. What were we doing sitting at the center of so symmetrical an arrangement, surrounded in all directions by increasingly larger star clusters?
More likely, Trumpler reasoned, he was being fooled by an optical illusion. All his calculations, like those of most every astronomer, took for granted that space was generally transparent, an empty medium through which light could travel unimpeded. If, however, the Milky Way was permeated with a fine cosmic dust, the measurements would be skewed—especially those where the dust was thickest, along the galactic plane. The dimness of a star or galaxy might be due not only to distance but also to this cosmic pollution. The farther the galaxy, the more pronounced the effect. Once Trumpler corrected for the distortion, the clusters turned out to be approximately the same size.
Astronomers had known there was dust in the Milky Way. The surprise was that it could be so pervasive. From almost the beginning, stargazers had contended with light and air pollution here on Earth. As civilization developed and the clarity of the sky degenerated, they placed their observatories higher and higher in the mountains. It hadn’t occurred to them that space itself could be so dirty.
And so came the first step toward resolving the Big Galaxy problem, the anomalous size of the Milky Way. When Shapley had taken his measurements, he had been looking through a fog. That made some of his beacons seem much farther than they really were. Once dust was factored into the equations, the home galaxy began to contract in size. That still left it larger than the others, but the adjustment felt like a step in the right direction. More revisions were about to come.
3
While Shapley’s map of the galaxy shrunk, the one for the universe grew larger. Again the reason was cosmic dust. Because of the galactic haze, Shapley had underestimated the true brightness of the Milky Way Cepheids, the ones at the foundation of the period-luminosity scale. When Hubble relied on this same standard to measure the distance to Andromeda he was, in effect, mistaking 75-watt bulbs for 60-watt bulbs. If these Cepheids were in fact burning brighter, then we were seeing them across a greater expanse. Using redshift, other galactic distances had been gauged in terms of Andromeda’s, so the error rippled outward. Everything was farther than it had appeared.
Astronomers now routinely adjust for the amount of cosmic pollution. Like dust in the atmosphere intensifying the redness of the sunset, dust in interstellar space can be measured by how much it reddens starlight. The lesson, though, took a decade to sink in. For years, one paper after another ignored the dust factor and, errors canceling out other errors, continued to find confirmation for Shapley’s original calibration. Looking back years later, the astronomer J. D. Fernie attributed the blindness to a herd instinct: “most of the astronomers of the day simply could not bring themselves to believe that interstellar absorption played any important role.”
Dust turned out to be just part of the problem. As far back as the Great Debate, Heber Curtis had suggested that Shapley was overreaching when he assumed that the variables in the Milky Way’s globular clusters shared the same relationship between period and brightness as those Henrietta Leavitt had found in the Magellanic Clouds. Both kinds had been lumped together to draw Shapley’s curve, the yardstick Hubble had used to measure to Andromeda.
The principle of uniformity encouraged these kinds of generalizations. But were the two variables really the same? If not, the whole distance scale might be askew.
Certain celestial anomalies hinted that this might be true. Even when one allowed for distance, the brightest of the globular clusters in Andromeda—the ones most easily detected— appeared to be inherently dimmer than their counterparts in the Milky Way. A German astronomer named Walter Baade later remembered discussing the discrepancy with Hubble, on cloudy winter nights at Mount Wilson as they waited for the sky to clear. Hubble argued that this might be a case where the principle of uniformity should not be so slavishly followed. After all, he noted, the clusters in the more distant galaxy M33, or Triangulum, were even fainter. Maybe this kind of variation was normal.
Baade had a different idea. Maybe the distance scale was wrong. The clusters in Andromeda and Triangulum were not really emitting less light. They were simply farther than had been reckoned. If so, uniformity would be restored. He soon had a chance to put the hypothesis to a test.
As the mid-1940s approached, many astronomers were off serving in the war. Hubble himself was soon directing ballistic missile tests at the Aberdeen Proving Ground in Maryland. Once Baade, technically an enemy alien, persuaded government authorities that he was not a security threat, he found it easy to book telescope time at Mount Wilson. The periodic blackouts, staged to discourage aerial attacks of Los Angeles, restored the night sky to a primitive blackness. Aiming the 100inch telescope at Andromeda, he could actually resolve individual stars, not just in the spiral arms but inside the galaxy’s dense core.
He discovered what appeared to be two different kinds of starlight. The stars in the galaxy’s center and in its globular clusters were colored differently from the “ordinary” stars in the galaxy’s outer reaches. That meant the two types must have different chemical makeups. While Leavitt’s “classical” Cepheids belonged to what is now called Population I, Shapley’s cluster variables belonged to Population II. It seemed a greater stretch than ever to assume that they obeyed the same law relating period and brightness.
When the new 200-inch Hale telescope came on line at Mount Palomar, ninety miles southeast of Mount Wilson, Baade zeroed in on Andromeda for a closer look. The classical Cepheids, he observed, were on average 1.5 magnitudes brighter than the cluster variables. “Instead of one period-luminosity relation,” he concluded, “there are actually two.”
When the new, brighter value of the classical Cepheids was plugged into the inverse square law, Andromeda turned out to be twice as distant as Hubble had reckoned. And so was the distance to everything else. As the newspapers put it, the universe doubled in size overnight. And, from the perspective of the big bang theory, it doubled in age. It was no longer younger than the Earth.
Baade’s discovery completed the explanation of why the other galaxies had seemed so much smaller than our own. That too had been an illusion. If they were farther away, then they were also larger.
Finally, with the new adjustments, the Milky Way was taken down to about 100,000 light-years in diameter—right in between where Curtis and Shapley had put it. It was, in the end, as unremarkable as its stars.
HUBBLE DIED IN 1953. Over the next few years, Allan Sandage, the young astronomer who had served as his last assistant, continued to clock redshifts and make adjustments to the distance scale. For some of his calibrations, Hubble had relied on the brightest star method. Sandage showed that what his old boss had taken as individual stars were actually entire stellar regions. Their intrinsic brightness was therefore much greater, putting the galaxies still farther away. Hubble, as Shapley might have put it, had mistaken trees for asters. The universe was expanding, not just because of the big bang but because of the explosion in astronomical knowledge.
Another way to say it is that the Hubble constant—the number by which you divide a galaxy’s velocity to get its distance—was growing smaller and smaller. And so, because of the reciprocal nature of the relationship, the size of the universe continued to grow. Hubble had initially set the constant at 150 kilometers per second per million light-years. More commonly the ratio is expressed using “parsecs” instead of light-years. As the Earth orbits the sun, a star that shows a parallax of 1 arc-second (1/3,600 of a degree) is, by definition, a single parsec (about 3.26 light-years) away. On that scale the Hubble constant had been around 500, Baade knocked it down to 250, and now Sandage to 75. Later he would reduce it again—to 50, ten times smaller than the original value. “The incredible shrinking constant,” one astronomer has called it. Every time it gets smaller, the map of the universe grows. So much is packed into that one little number. At its core lie Miss Leavitt’s stars.