The Passing of Allan Sandage

By Julianne Dalcanton | November 15, 2010 10:37 pm

Earlier today I learned of the passing of Allan Sandage. Allan was a tremendously broad astronomer, who had a lasting impact on fields of astronomy stretching from stellar evolution to the largest cosmological issues. He is perhaps best known for his work on the distance scale, and measurements of the Hubble Constant, but he had equally significant contributions to our understanding of stars.

The prominence of his work on the Hubble Constant is in part due to the rather contentious history of this subject over much of the 90’s and early 2000’s. Allan was at heart a stellar astronomer, but one who found himself tied to Hubble’s legacy by virtue of being Hubble’s telescope assistant in the years leading up to Hubble’s unexpected death in 1953. As one of the earliest pioneers (with Martin Schwarzschild) of the technique of using main sequence turnoffs to assign ages to globular clusters, Allan was deeply (and understandably) bothered by experiments that returned large values of the Hubble Constant — these values implied ages for the universe that were younger than the oldest globular clusters, which was clearly an implausible contradiction. Instead, Allan and his collaborators published a long series of papers attempting to deal will every uncertainty and bias in the distance scale, and found a consistently smaller value of the Hubble Constant than the other competing team (the “Hubble Key Project”, led by Wendy Freedman with many collaborators). In time, Allan’s group’s on-going evaluations of the distance calibration gradually pushed their value of the Hubble Constant up somewhat, while the Key Project’s values were being nudged down a bit (although they never did actually meet, particularly as error bars shrank in more recent years). Simultaneously, the discovery of dark energy changed the age estimates for the universe, allowing old globular clusters to co-exist harmoniously with a moderate value of the Hubble Constant.

During this time, Allan developed a reputation for being, well, difficult. His scientific disagreements on this issue unfortunately veered occasionally into the personal. That said, I had the pleasure of being a postdoc at the Carnegie Observatories during this time, and had an office a few doors down the hall from him. Allan was invariably gracious and kind to the postdocs. He was scientifically engaged, and always willing to share his knowledge, which was both deep and wide. I enjoyed having him for a colleague for 4 years, during a very scientifically vibrant stage of my astronomical training, and I am very sorry to hear of his passing.

CATEGORIZED UNDER: News, Personal, Science
  • Pingback: Tweets that mention The Passing of Allan Sandage | Cosmic Variance | Discover Magazine --

  • Phillip Helbig

    I had the pleasure of hearing him lecture (together with Rich Kron and Malcolm Longair) at the 1993 Saas-Fee school. Interesting, especially from a historical perspective (“When I started out, computers were about 20 years old and usually female.”), but bitterness and contempt for the high-value camp was obvious. Surprisingly, his main motivation for “knowing” that H had to have a low value was that “grand unification” firmly predicts Omega=1 and thus the Einstein-de Sitter model (lambda was just ignored) which means that a low value of H is required to make the universe as old as globular clusters etc indicate that it is. I don’t know if he really thought observations indicate low H (he actually championed “42” in one paper!) and this tended to make him believe in Rocky-Kolb-style cardboard inflation or vice versa. Probably a bit of both, in a vicious circle. He sort of reminded me of Bill Haley playing “Rock Around the Clock” even after Sergeant Pepper had appeared. 😐

  • Pingback: RIP Allan Sandage « Letters to Nature()

  • Pingback: RIP Allan Sandage (1926-2010) « In the Dark()

  • Ben

    Sandage was, I believe, Edwin Hubble’s last student and in some sense the last link to that heroic age of early 20th C. astronomy and founding of extragalactic astronomy. The Hubble constant wars obscured that if Sandage had never touched the distance ladder, only produced his results on stellar populations and Galactic structure, he would still have been a giant figure (same could be said for Hubble).

    As I remember it, up to the mid-90s there was a strong theoretical prejudice for omega=1 and a cosmological constant was thought to be ruled out by the frequency of gravitational lenses, so he was hardly alone in insisting on that cosmology.

    Around that time, one of the Palomar telescope operators told me that before the autoguiders were ubiquitous, he could practically tell who was observing by the style of guiding with the hand paddle. (The astronomer would look through cross hairs at a guide star and punch buttons on the paddle to tune up the telescope tracking and keep the star centered throughout a long exposure.) Some astronomers had a very irregular rhythm, depending on the motion of the star (and their level of attention). Sandage, he said, kept a steady beat, punch, punch, punch, all throughout an exposure – which could be an hour or much more.

  • Paddy

    During the 1970s and early 1980s, Sandage was widely considered to be the leading observational cosmologist and, as you mention, the inheritor of Hubble’s mantle. It was also at this time (not the 1990s and later) that the contention about the value of the Hubble’s constant reached its height. Sandage’s epic series of papers co-authored with Gustav Temmann, “Steps towards the Hubble constant”, began in 1974 and over the following twenty years repeatedly reached a value of 50 to 55 km/s/Mpc, with ever-decreasing formal uncertainty. Meanwhile a rival series by Gerard de Vaucouleurs and collaborators homed in on a value of 90-100, also with a claimed 10% error. Both sides lost credibility by repeatedly getting the same answers, with the same huge discrepancy, from a wide range of different methods. Not too surprisingly, the Hubble Key Project found a value half-way between; and even the Hubble telescope couldn’t give errors as low as Sandage had claimed in 1982, mostly based on photographic plates from Palomar. (To be fair, Sandage would admit that his quoted errors did not include systematic uncertainties).

    There is a sensitive portrait of Sandage in Dennis Overbye’s book “Lonely Hearts of the Cosmos”.

  • Sean

    I remember going to a talk by a theorist, way back in 1988, who used a value of 75 km/sec/Mpc in her simulations. “That way we won’t offend anyone, because nobody believes that’s the right number.”

  • Richard

    A touching and accurate tribute. Very nicely done. For anyone who is interested, I profiled Sandage for the NYTimes Magazine in 1999:

  • ben

    See lengthy obituary in today’s New York Times at

  • Phillip Helbig

    Lest my comment above sound too negative, let me note that Sandage’s Saas-Fee lectures were quite enjoyable (especially considering that he is not actually a lecturer), comparable to those of Longair (which is saying a lot). Buy the book!

    You can see parts of it here:

    A similar review is available online:

  • Phillip Helbig

    “As I remember it, up to the mid-90s there was a strong theoretical prejudice for omega=1 and a cosmological constant was thought to be ruled out by the frequency of gravitational lenses, so he was hardly alone in insisting on that cosmology.”

    The first is certainly true. Again, Overbye paints an interesting picture in Lonely Hearts in the Cosmos. This is quoted in Coles’s and Ellis’s excellent book:'re+thinking+like+an+astronomer+and+not+like+a+physicist+simon+white+never+understood+inflation&source=bl&ots=n2oSTY3Lld&sig=l5dSOZMwCH5jGZcd8x9rIs75Ktk&hl=de&ei=bgnkTMC8BMWBOr7RiZIB&sa=X&oi=book_result&ct=result&resnum=3&ved=0CCQQ6AEwAg#v=onepage&q=you're%20thinking%20like%20an%20astronomer%20and%20not%20like%20a%20physicist%20simon%20white%20never%20understood%20inflation&f=false

    (It’s at the top of the page; scroll up to get the whole quote. Ironically, Schramm co-authored (with Gott, Gunn and Tinsley) one of the most famous papers advocating a low Omega, not too many years before. Yes, the Chicago physicist is definitely Schramm (RIP).)

    As to the second, this was just as badly established (i.e. not really established at all) and not as prevalent as the first.

    Yes, Sandage was not alone, but he was one of the few observers who strongly supported such a cosmological model. Some observers assumed it casually, but AFAIK none defended it as strongly as Sandage.

  • Anonymous

    Allan Sandage always made me sad. Brilliant guy who was so attached to his own cognitive prejudices that it interfered with his ability to think straight. I’ve sometimes wondered if the same mindset that never let him acknowledge that the Hubble constant was 75 was behind his unusual religiosity (see

  • Fermi-Walker Public Transport

    I remember Alan Sandage visiting our department and giving a talk when I was a grad student. Afterward he chatted with us grad students for some time and then he asked us what were our interests. Someone said quasars for which he responded by shaking his head, I said radio galaxies which got a similar response and this went on till someone said astrometry which got him to smile, pat the guy on his back and say “astrometry is the most important topic in astronomy and that everything else depends on the astrometry being right” or words close to that. We all laughed at that, but he was making an important point.

  • Nicholas Suntzeff

    I knew Allan very well. He was a truly brilliant and complicated man. I experienced the complicated part many times: for 10 years he did not talk to me because of the work on Type Ia supernovae which we showed to be standardizable candles. He believed them as standard candles. He later wrote me a letter out of the blue apologizing for his actions, and we started talking again, especially about RR Lyraes which was one of his favorite subjects. I am so sorry to see that people only remember him for the H0 arguments. Remember though, that with Omega=1, with no cosmological constant, and H0 75, the Universe is younger than the oldest stars. Allan berated me that we did not include a reference to a negative q0 as being a prediction of the SS theory. The High-Z group would not include such a reference and I felt bad (and still do) that we did not give a nod to at least one consistent measurement with SS.

    So, those of you who (rightfully) criticize Allan for his staunch support for H0<60, be careful because he was motived by stellar ages, and H0=75 with a mass filled universe is too young. He was precisely right on both counts, and lamented to me that it was not seeing that q0 could be negative rather than H0<60 which was his error. His final values for H0 were around 63 by the way, not too far from the best value we have right now.

  • Phillip Helbig

    “Remember though, that with Omega=1, with no cosmological constant, and H0 75, the Universe is younger than the oldest stars.”

    OK, so what? Maybe Omega is less than 1. There was certainly NEVER ANY observational evidence that Omega was anywhere near 1. As an observer, Sandage should have appreciated that. Also, why rule out a cosmological constant with no good reason to do so? Let’s face it, despite all the great stuff he did, Sandage here is no better than Aristotle writing that women have fewer teeth than men (he was married twice; maybe, as the “ultimate observer”, he limited his sample to his personal experience). Also, this is a “two men say they’re Jesus; one of them must be wrong” situation. One knew that the scenario described above was wrong but not, for sure, which part was wrong. With all the uncertainties, maybe the stellar ages were wrong. (Whenever it looked that H might be high after all, some bright theorist would always come up with a way “to get those globular-cluster ages down”.)

    “Allan berated me that we did not include a reference to a negative q0 as being a prediction of the SS theory. The High-Z group would not include such a reference and I felt bad (and still do) that we did not give a nod to at least one consistent measurement with SS.”

    SS predicts q of exactly -1. It has the same metric as the de Sitter model. “Negative q” is too vague. q=1 is ruled out by the SNIa data, AFAIK.

    “lamented to me that it was not seeing that q0 could be negative rather than H0<60 which was his error"

    Well, the first implies the second. I’m glad he admitted it. At the Saas-Fee lectures, he justified his prejudice several times with hand-waving arguments involving “grand unification”.

  • réalta fuar

    I’m amazed at how historically tone-deaf the original post here was. Thanks to Paddy above for setting the record straight. As Paddy said, the real “war” (and it was a bitter one, on BOTH sides) was between Sandage’s group and de Vaucouleurs’. Dark energy never entered into that battle at all. Sandage won that battle, as only befits a master stellar astronomer. It is a shame (and sad) that Sandage could never reconcile himself to the Hubble Key Project results, but as N.S. correctly pointed out, his final value for the Hubble Constant was really quite good (and completely consistent with both globular cluster ages and the mass density of the universe). Given the intense ridicule Sandage was subjected to for daring to suggest that the cosmologists might be grossly wrong (100 km/sec/mpc) , it’s really not surprising that he became bitter and intransigent.

  • Phillip Helbig

    I don’t follow you. Yes, dark energy had nothing to do with the debate over the value of the Hubble constant. Both groups produced various values over the years, not always compatible with their own other results, much less those of the other group. True, the initial values of the high group came down, but then Sandage’s value of 55 was higher than the 42 he advocated in another paper. Yes, the HST key project found a value somewhere in-between, but within the errors not really compatible with Sandage’s low value. Note also that Sandage used HST as well, and got a value lower than the HST key project.

    Note also that the determination of the Hubble constant via the methods Sandage used is not sensitive to the other cosmological parameters. It’s not like he assumed the Einstein-de Sitter model and this somehow made his value come out too low. Rather, it was his theoretical prejudice in favour of the Einstein-de Sitter model which, somehow, made his value come out low enough to be consistent with the age of the universe.

  • réalta fuar

    Sandage, wrongly, made assumptions to JUSTIFY why he HAD to get low values of H, but those assumptions, except perhaps (or perhaps not) influencing his analysis, were not necessary to make his results consistent with what was known at the time. Actually, to my knowledge (I could be wrong here) he didn’t even do that until after the main battle was over. The point I suppose I was trying to make is that Sandage was wrong (obviously, no shame in that) but not STUPIDLY wrong, as his primary nemesis was. The discovery of dark energy was not motivated by the, by then, well determined value of H.

  • Phillip Helbig

    I agree with your first sentence, except that “known at the time” needs to be quantified. :-) (As Bill Press once quipped, someone knows the Hubble constant to 4 significant figures; we just don’t yet know who that person is!) He mentioned “grand unification” at least a few times per day in his Saas-Fee lectures in 1993. The main battle was by then not yet over; the HST Key Project did not yet have the “right” value of H0; other methods had wide scatter and big error bars; the cosmological constant was far from being accepted as even a theoretical possibility, much less as the interpretation of observations. (We have come a long way since then.)

    I don’t see a qualitative difference between Sandage getting a too low value and de Vaucouleurs getting a too high value. Since the value of the cosmological constant was essentially unknown back then, even H0 of 100 would have been compatible with observations.

    The discovery of dark energy has little to do with the value of H0. The most direct evidence is the SNIa data, and these are independent of H0 (one is looking for a departure from linearity in the m-log(z) relation; H0 shifts it up and down (it is a log-log plot)). Yes, other methods can (and do) give a similar value of the cosmological constant if other things, including the Hubble constant, are well known, but this is more “supporting evidence” than “discovery”.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

See More

Collapse bottom bar