Why Did It Take Fourteen Months To Discover The Expansion of the Universe?

In a Fog

In his book "Blind Watchers of the Sky", Rocky Kolb describes the work that culminated in Hubble's 1929 paper (PNAS, 15, 168) showing a linear correlation between velocity and distance. Kolb characterizes the impact of Hubble's work as follows:

"... the fog shrouding the expansion of the universe had finally lifted ... Once Hubble's paper appeared, most physicists realized that expansion and Einstein's theory are naturally compatible."

Is that really true? Did Hubble's paper immediately open people's eyes to the realization that we are in an expanding universe?

Not at all. Hubble's paper appeared in March 1929. Plaskett (1932 PASP, 44, 215) conveyed the situation at the time:

"I distinctly remember being present at the meeting of this section [the Pacific Division] of the American Association for the Advancement of Science at Berkeley in [June] 1929, when NO ONE appeared to have the temerity to suggest that Hubble's recently announced velocity-distance relation represented an actual recession of the nebulae."

I have found seven papers published over the next fourteen months that cited Hubble. Not a single one mentioned the expansion of the universe. Hubble himself thought his data would shed light on the de Sitter effect , but at the time de Sitter's metric was interpreted as representing a static (or stationary) universe.

Tolman (1929, ApJ, 69, 245) followed up with an immediate analysis of Hubble's discovery, finding that a de Sitter model DOESN'T work.
Tolman (1929, PNAS, 16, 320) wrote another followup paper, this time developing a non-static line element but assuming it was driven by matter being converted to radiation, a model that convinced no one.
Menzel mentioned it in a review talk presented at the AAASPD (1929, PASP, 41, 224), and thought Hubble's data had actually been demonstrated to verify the de Sitter model.
Zwicky (1929, PNAS, 15, 773) presented various possible physical explanations for the effect, including a theory now called "tired light", but not including an expanding universe.
Shapley (1929, PNAS, 15, 565) wrote a paper in which he constructed his own version of the Hubble diagram but was less certain about a linear v-D relation ("... the evidence from the integrated magnitudes does not appear to support strongly the deduction that the velocities of external galaxies are directly proportional to their distances ..."). He offered no physical explanation for the effect.
Lundmark had lost interest in the velocity-distance business, but in an early 1930 paper (PASP, 42, 31) he mentioned Hubble's discovery tangentially (not actually citing the paper itself), mainly commenting that Hubble has assumed a K term linear in distance.
De Sitter (1930, BAN, 5, 157), like Shapley, also felt compelled to construct his own version of a velocity-distance diagram. The main conclusion of the paper was similar to that reached by Tolman - a de Sitter model does not work.

At the very end of his paper, de Sitter finally mentioned the work of Lemaitre but indicated he would defer discussion to a future paper. Thus, de Sitter's paper is properly the last in line of papers that failed to find that Hubble's relation implies an expanding universe.

Lifting the Fog

The fog was finally lifted in May/June 1930. Lemaitre sent 2 reprints to Eddington, who passed along one to de Sitter, of an overlooked paper that Lemaitre had published in 1927 (Annales de la Societe Scientifique de Bruxelles, A47, 49), which was the first paper to connect a model of an expanding universe that included ordinary matter to the radial velocities of galaxies. Both Eddington (MNRAS, 90, 668) and de Sitter (BAN, 5, 211) immediately responded and published papers that essentially expanded on Lemaitre's paper. De Sitter appears to have been the first to use the term "Expanding Universe" in the title of a paper.

Plaskett also related a story about attending a RAS meeting: "... I was ... present in London at the May 1930 meeting of the Royal Astronomical Society when Eddington announced - the FIRST TIME to my knowledge that it had been DIRECTLY and UNAMBIGUOUSLY stated at an astronomical meeting - that the RED-SHIFT was due to the EXPANSION of the UNIVERSE." Eddington was giving a preview of his MNRAS paper. De Sitter was also present (1930, Observatory, 53, 161).

Assigning Credit

Kolb continues:

"In my opinion, Hubble justifiably deserves credit for the discovery of the expansion of the universe."

As Sandage has pointed out (1989, in "Hubble Deep Field"), Hubble did not believe in an expanding universe. The term appeared only once in the title of a paper of his (1942, American Scientist, 30, #2) and he considered an expanding universe model to be unsatisfactory, preferring a model that is static and infinite, but with the redshifts due to some unknown cause. E.G., "If red shifts are velocity shifts which measure the rate of expansion, the expanding models are definitely inconsistent with the observations unless a large positive curvature (small, closed universe) is postulated." (1936, ApJ, 84, 517).

Lemaitre would seem to be the most deserving of the accolade since, in 1927, two years before Hubble's paper, he wrote explicitly, "The receding velocities of extragalactic nebulae are a cosmical effect of the expansion of the universe." However, as Ian Steer has pointed out (2011, JRASC, 105, 18), "Success goes not to the first to make discoveries to their own satisfaction but rather to the first to prove them to the satisfaction of others." Lemaitre's paper was published in a little-read journal, unnoticed by others, the copy he sent to Eddington was forgotten, and the one person who knew about it, Einstein, did not believe it. Eventually he did convince others, but it was done through the voices of Eddington and de Sitter. Further, one could say that Weyl (1923 Raum Zeit, Materie; 1923 Phys. Zeitschr, 24, 230) was actually the first to derive Hubble's law in an expanding universe, although his derivation was rather opaque, and his expanding universe was embedded in pure de Sitter space, which is the end-state of Lemaitre's.

Robertson (1928, Phil. Mag., 7, 835) unknowingly repeated much of Weyl's and Lemaitre's work, including deriving the velocity-distance relationship, but he did not explictly describe his cosmological model as representing an expanding universe. Curiously, Eddington is listed as the one who communicated it to the journal, but he seems to have missed its significance. So Robertson cannot be awarded credit either.

In summary the question of who discovered the expansion of the universe has no easy answer. It is much easier to answer when it was discovered, and the the most fitting time would seem to be Plaskett's account of listening to Eddington at the May, 1930 RAS meeting. Fourteen months late.

Addenda

Although Eddington and de Sitter wrote the two key papers that put Lemaitre's model of an expanding universe on the map, neither cited Hubble's 1929 paper. De Sitter referenced his own velocity-distance paper from a month early. Eddington quoted Hubble's "round number" value for the Hubble constant (500 km/s/Mpc) but did not give a reference; his only reference to Hubble was in regard to the 1926 paper, where Hubble estimated the mean density of the universe. (It is interesting to note that Eddington also [mis]quoted de Sitter's form for the Hubble constant ["about 1/2000 of the velocity of light for a million light-years distance,"] although de Sitter's value properly refers to a distance of 10^24 cm, slightly more than one light year, and is slightly lower than Hubble's value.) Neither paper cited Lemaitre's original value of 625 km/s/Mpc. In 1934, Lemaitre (PNAS, 20, 12) would refer to "Hubble's ratio of distances to spectroscopic velocities".

Weyl also wrote a paper (1930, Phil. Mag., 9, 936) in response to Hubble's, and even though he he had derived Hubble's law in 1923, he expressed caution about saying that his model was the proper explanation for Hubble's data (calling it "speculative"), as opposed to, say, Zwicky's tired light theory or some other effect. Thus, he straddled the fence. The text was written in 1929; the actual data of publication is a little unclear (possibly June of 1930). The bulk of the paper was actually a clarification of his 1923 paper, stating that it was equivalent to Robertson's 1928 work. Weyl referred to his model as one of relativistic cosmology offering a "geometric" explanation of Hubble's "strange phenomenon".