Hubble versus Slipher

Since I’m here at a conference celebrating the scientific achievements of Vesto M. Slipher, I thought I’d take the opportunity to make a few remarks about Slipher’s work and legacy.

I often use this picture in popular talks to illustrate the correlation between distance (x-axis) and apparent recession velocity (y-axis) that has become universally known as Hubble’s Law. This is an early version of such a plot published by Edwin Hubble in 1929.

In public talks I rarely have time to go into the details of this, but it is worth saying that only the results on the x-axis were Hubble’s own measurements. Hubble only contributed half of the above plot, i.e. the distance measurements, and these turned out to be wrong by a factor of about 10 owing to an incorrect identification of the stars used as standard candles. All the recession velocities on the y-axis – obtained by looking at the displacement of lines in the target galaxy’s spectrum – were in fact obtained by Vesto Slipher at the Lowell Observatory here in Flagstaff, Arizona. Hubble used these data from Slipher with permission, but gave no credit to Slipher in the references to his 1929 work. A later, and more convincing, version of this plot published in 1931 by Hubble and Humason, was accompanied by a generous acknowledgement to Slipher’s contribution. However, by then, Hubble’s name was firmly associated with the plot and Slipher’s contribution was largely forgotten for many years subsequently.

This episode isn’t at all atypical of Hubble’s behaviour. He was an extremely ambitious man who was an expert in the art of promoting himself and the Mount Wilson Observatory where he worked. Slipher was a very different type of man: quiet, self-effacing, and very much a team player, dedicated to scientific accuracy rather than his own reputation.

It’s worth saying further that the key observation that led to the understanding that the Universe is expanding is the fact that most of the spectra obtained by Slipher, over the years subsequent to his first measurement of the spectrum of the Andromeda Nebula (M31) celebrated by this conference, showed a redshift indicating velocity away from the observer. Even without distance measurements this leads directly an interpretation in terms of cosmic expansion. Ironically, the first spectrum he obtained, M31 shows a blue shift, as do a few others plotted with negative velocities in the above diagram, but the more distant sources exclusively show a redshift.

As a scientist should be, Slipher was very careful about the interpretation of this result. The more distant objects are fainter and thus more difficult to observe. Could it arise from some systematic artifact? Or could there be an unknown physical effect that produces a redshift dependent on the size of the source? These questions could only be answered when accurate distances to the nebulae were established, so Hubble’s contribution was by no means negligible. It’s completely untrue, however, to say that Hubble discovered the expansion of the Universe, so there’s yet another example of Stigler’s Law of Eponymy whenever anyone talks about the Hubble expansion.

One of the great things about coming to this meeting was the chance to meet Alan Slipher, grandson of Vesto Slipher. He and other members of his family refer to Vesto as “VM”, by the way, which I hadn’t realised before. VM lived a long life, dying in 1969 just short of his 94th birthday, so Alan knew him well until age 17 or so. He spoke most warmly and movingly after yesterday’s conference dinner about his memories of his grandfather, who he clearly looked up to. His words confirmed the impression I’d already formed, that Slipher was an extremely cautious and serious scientist as well as a kindly and humble man.

The contrasting personalities of Slipher and Hubble are further illustrated by correspondence between the two that is archived at the Lowell Observatory. Slipher comes across as kindly and cooperative, Hubble as pompous and self-regarding. I know which of the two I admire the best, both and scientist and human being.

About these ads

22 Responses to “Hubble versus Slipher”

  1. My view is that the demonstration of the expansion of the Universe required redshift measurements, distance estimates and a clear identification of a linear relationship between them. That is what Hubble achieved, although Slipher’s redshift results alone were suggestive of an expansion, and Slipher was the giant on whose shoulders Hubble stood.

    I find it puzzling that Slipher dominated the field of measuring redshifts of the `spiral nebulae’ for some years without so much competition. The Lowell Observatory had some excellent equipment, but many other observatories had equipment to match, including ones outside the United States. Work with new facilities at Mount Wilson soon enabled new things to be done, and uniquely, particularly by Hubble. But there were several years when Slipher stood almost alone as a pioneer of galaxy redshift measurement. Did nobody else bother that much?

    It must be a very interesting conference. I hope everybody there enjoys themselves greatly.

  2. Thanks for this interesting story. Where did Slipher publish his extragalactic redshifts that Hubble used? Is there a citeable reference?

  3. Hubble, as you know, did his PhD work at Yerkes. So, during my 6 years there I learned a reasonable amount about him. He apparently had visited England during his time at U Chicago, and came back with a posh accent and also a pipe to complete the posh gentleman look.

    • Rhodri,

      Yes, that’s right, Hubble was a Rhodes Scholar at Oxford University, where he studied humanities including law and, oddly, picked up an upper-class English accent while he was there, with affected manners as well. (He had already obtained a BSc degree from the University of Chicago before going to Oxford.)

      Curious.

      Bryn.

      • Yes, he sojourn to Oxford was between his undergraduate and postgraduate studies at Chicago. He did his PhD work on gaseous nebula using the 24-inch reflecting telescope which Ritchey had designed and built. This was located in the south tower of Yerkes, above where my office was. It was replaced by a 40-inch reflecting telescope in 1967 or 1968.

    • “He apparently had visited England during his time at U Chicago, and came back with a posh accent and also a pipe to complete the posh gentleman look.”

      As for the accent, there are people who do it intentionally and people who don’t. (This is not correlated with whether or not they say they do it intentionally.) Since in Hubble’s case we also have the whole “posh bloke” persona, it was presumably intentional.

      Are there any recordings of him? Was it convincing? I read somewhere that his pronunciation was not bad but his diction and vocabulary were a bit off.

      • I’ve always assumed that Hubble’s accent and manner were deliberate, but that was no more than an assumption.

        I can’t remember ever hearing a recording of Hubble, so can’t comment on how convincing he was.

    • Alan Heavens Says:

      There is a recording of Hubble explaining his ideas, in fact. I was quite surprised when I heard it – not what I expected at all.

      • Well, that’s certainly not an Oxford accent!

        It’s an entertaining video. Admittedly, I’m not sure I’m going to cite that clip as primary source material in the history of science.

  4. John Peacock Says:

    Bryn: the problem is that Hubble didn’t achieve an objective proof of the linear relation that you desire. His sample barely made it to 10 Mpc, on which scale peculiar velocities dominate. But Hubble knew that there was a theoretically predicted linear relationship; his paper was one of several trying to find it in the data. What he did was take a model velocity field of the assumed expansion, plus a dipole from the random motion of the Sun. The latter is the largest effect, and he adjusted the fitted dipole to maximise the signal for the linear term – and then plotted the “corrected” redshifts against distance. If we took a proper approach and marginalized over an unknown dipole, I don’t believe there would be a significant signal: looking at modern HST Cepheid & SNe distance data, you would need to go to 2 to 3 times the depth to get a solid signal where the expansion dominates over peculiar velocities. This shows the value of having a theoretical prior. The first person who predicted this linear distance-redshift prior was Weyl, in 1923.

    And anyone who wants to see the Slipher papers can find them at http://www.roe.ac.uk/japwww/slipher – he had the bad luck to publish some of his key work in journals that died not long after. They are beautiful pieces of writing, never mind being great science.

    • Alan Heavens Says:

      John,
      It seems that Hubble’s result actually stands up. With Hubble’s original 24 galaxies, and gathering RAs and DECs from the well-known astronomical database ‘wikipedia’, a quick MCMC analysis gives mean and marginal errors, in his units, of

      H0 = 463 +/- 54 (MCMC) vs 465 +/- 50 (Hubble)
      X = -72 +/- 78 vs -65 +/- 50
      Y = 239 +/- 150 vs 226 +/- 95
      Z = -197 +/- 79 vs -195 +/- 40

      (X,Y,Z) = Milky Way dipole (km/s)

      Comparing with Hubble’s on the right it seems he basically got it right (apart from the well-known problem with the distances). There is some degree of correlation of the H0 estimate with that of the dipole, but it’s not enormous. Note that the MCMC error bars depend on the assumed dispersion in the speeds, which I have set at 200 km/s here. A smaller value of around 100 brings them more in line with Hubble’s.

    • John,

      Yes, absolutely. I was thinking more of Hubble’s later work for a clearer demonstration of the linear relation. I think he achieved this, for example, in his 1931 paper with Milton Humason which used new Mount Wilson data.

      I agree very much that peculiar velocities complicate the interpretation of Hubble’s 1929 paper. I had underappreciated the importance of his fitting of the solar motion in the 1929 paper. Fortunately for Hubble, there were very few Virgo Cluster galaxies in his 1929 sample (and none from Fornax) – the velocities of cluster galaxies would have greatly confused the results had he included several of them.

      Your webpage about Slipher’s publications is very useful: I wasn’t aware of this and am grateful for the link. It’s now in my bookmarks.

      Bryn.

    • John,
      I like your page with Slipher’s papers. I suggest you add links to the ADS records for the papers so that people easily can get the bibtex record, check the citations, etc. Hopefully these papers will be cited more in the future

  5. But are we sure that universe is expanding and not merging into the outer antimatter universe on opposite entropy path. Read balloon inside balloon theory of matter and antimatter universe on opposite entropy path .

  6. John Peacock Says:

    Alan: do you plan to publish this first correct analysis? At least we could keep the symbol H_0 for the Heavens constant?

    Two points: even if the free dipole makes (to me) surprisingly little difference, Hubble didn’t know that.

    More seriously, his distances are completely wrong in any case. I knew this was true empirically: even if you calibrate them with the modern distance to M31, the slope is still more than twice what it should be. At this meeting, I learned an interesting anecdote from Owen Gingerich, which he got from Sandage: for the more distant galaxies (which give any signnificance to the detection) the brightest stars he was using as distance indicators are actually HII regions.

  7. John Peacock Says:

    Bryn: Hubble & Humason 1931 assumed that galaxies were standard candles, for which there was little/no justification at the time. And in any case, that was the same assumption made by Lundmark, when he found a distance-redshift correlation in 1924. I think the 1931 paper wasn’t challenged as it might have deserved to be because people thought (incorrectly) that the issue was already settled in 1929.

    • My interpretation of the 1931 Hubble and Humason paper is that they sampled brightest galaxies in a number of clusters, assumed (crudely) that the mean of these brightest cluster galaxies was the same and used them to probe the redshift-distance relation out to distances where peculiar velocities were not significant. The method was clumsy and there would have been systematic errors, but it did establish the linear relationship.

      The samples of galaxies used to estimate distances were very poorly defined, but I think that some valid results were obtained. They would have sampled less far down the luminosity function at greater distances, estimating mean magnitudes that were too bright as a consequence, and estimating distances that would have been too small at higher redshift (given the fact that the distance scale was already out by a significant factor even at low redshift). This would have propagated into the value of the Hubble constant. They also included some field galaxies, which would have confused matters.

      Hubble’s understanding of the luminosity function of galaxies was, to be charitable, poor, or being less charitable was completely wrong.

      However, I still think their 1931 study obtained an important result.

  8. Phil Uttley Says:

    Leaving the so-called Hubble Law aside, didn’t he also play a major role in settling the ‘great debate’ about the distance to galaxies? And there is also galaxy classification.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 3,799 other followers

%d bloggers like this: