Allan Rex Sandage, who died of pancreatic cancer on 13 November 2010 in San Gabriel, California, was one of the greatest astronomers of the 20th century. He gave the first reliable size of the universe—seven times larger than the estimate by his mentor, Edwin Hubble—and the first reliable age of the universe. Among his many honors were the 1991 Crafoord Prize and the 2000 Gruber Cosmology Prize.
Born on 18 June 1926 in Iowa City, Iowa, Sandage enrolled briefly at the University of Miami and then joined the US Navy for the last year of World War II. He told astronomer Martin Johnson years later that his subsequent path was decided by reading Johnson’s book Time, Knowledge and the Nebulae (Faber and Faber, 1945). Sandage graduated in physics and philosophy in 1948 from the University of Illinois. In 1949 he embarked on a PhD at Caltech under the supervision of Walter Baade, and in 1950 he started work at the Carnegie Observatories in Pasadena as Hubble’s observing assistant. After becoming a staff member in 1952, he remained associated with the observatories all his life.
Besides his association with Hubble and Baade, the two giants of extragalactic observational research of the time, he also worked with theoretician Martin Schwarzschild on models for stars at the end of their hydrogen-burning phase. Their seminal 1952 paper explained the abrupt turnoff of the bright end of the observed “main sequence” in globular clusters and the fact that red giants were dying stars rather than young stars as had been previously supposed. Sandage used that insight the following year to estimate the age of the globular cluster Messier 3 as 5 billion years.
In 1953, the year that Sandage submitted his PhD, Hubble died and Sandage, at age 27, became responsible for developing the Carnegie Observatories cosmology program, which used the 60- and 100-inch telescopes on Mount Wilson and the newly commissioned 200-inch one on Mount Palomar. Being on the Carnegie staff gave him ample observing time, and he used it, with Hubble’s associates Milton Humason and Nicholas Mayall, to extend Hubble’s velocity–distance diagram for galaxies. Their 1956 paper contained redshifts for 920 galaxies, compared with the 24 originally used by Hubble. Sandage also extended Hubble’s galaxy classification scheme with his 1961 The Hubble Atlas of Galaxies (Carnegie Institution of Washington).
By 1958, however, Sandage realized that there were problems with the distance scale, even after its revision by Baade. Sandage showed that what Hubble had thought were the brightest stars in galaxies were extended clouds of ionized gas, or HII regions. And he showed that for Cepheid variable stars, the period–luminosity relation, which had provided crucial distance estimates for Hubble, should be thought of as a period-luminosity-color relation. He concluded that the Hubble velocity/distance constant H0 was 75 km s−1 Mpc−1, one-seventh of Hubble’s estimate and one-third of Baade’s. The universe was correspondingly larger.
Sandage knew that his estimate was still uncertain, and in a classic 1961 paper “The Ability of the 200-inch Telescope to Discriminate Between Selected World Models,” he laid out the goal of using observations to accurately determine cosmological parameters. The three key parameters were the age of the universe, the Hubble parameter H0, and the deceleration parameter, which measures how much the expansion of the universe is slowing due to self-gravity. In 1970 he published studies of the main-sequence turnoff point in four globular clusters, which allowed him to estimate a mean age of 11.5 billion years for them. With Olin Eggen and Donald Lynden-Bell, he had shown in 1962 that the globular clusters date back to the era when our galaxy was undergoing its first collapse under gravity, so the age of the globular clusters is the age of our galaxy and not far short of the age of the universe.
Between 1972 and 1973 Sandage published a series of papers extending the Hubble diagram to greater distances; he used the brightest galaxies in clusters to try to estimate the deceleration parameter. Unfortunately, as the late Beatrice Tinsley pointed out, there were two competing sources of uncertainty in the luminosity of the brightest cluster galaxies: Their luminosity declines as their stars age, and mergers between galaxies increase the mass and luminosity of individual galaxies.
In 1974–75 Sandage published a series of papers with Swiss astronomer Gustav Tammann on the distance scale, in which they concluded that H0 is about 55 km s−1 Mpc−1. For the next 25 years they were involved in an H0 controversy with other astronomers, especially Gérard de Vaucouleurs, who favored values as high as 100. Sandage did not relish the controversy, and it was almost always left to Tammann to defend their work at conferences. In the 1990s the Hubble Space Telescope devoted much time to estimating the distances to galaxies using Cepheids, and two teams used those to estimate H0. The team led by Wendy Freedman concluded that H0 = 72, while the team led by Sandage found 62. The Wilkinson Microwave Anisotropy Probe analysis of cosmic microwave background fluctuations appears to confirm the higher value, but some recent analyses have shown some drift downward.
Allan Sandage was always charming to younger scientists who visited him in his Pasadena office. When I first visited him around 1970 he took the trouble to show me his measuring machines in the basement and demonstrate how one searches for Cepheids in a galaxy. What he really loved were the nights at the telescope. I shall always think of him as one of the small handful of truly great scientists I have been lucky enough to meet.
