Who first published a Hertzsprung-Russell diagram? Hertzsprung or Russell? Answer: neither!
Most astronomers, when asked the question of who first published a Hertzsprung-Russell diagram would probably guess that it was Ejnar Hertzsprung, the famous Danish astronomer who went to Göttingen in 1908 to visit Karl Schwarzschild. In 1911, Hertzsprung published colour-magnitude diagrams for both the Hyades and Pleiades clusters. A few might guess that it was Henry Norris Russell at Princeton, who independently published a diagram of absolute visual magnitude against Harvard spectral type for nearby field stars using trigonometric parallaxes. However Russell’s first publication of his diagram was in 1914, several years after Hertzsprung.
Whether you answer Hertzsprung or Russell, your answer is wrong!
The first diagram of magnitude versus spectral type was published by Hans Rosenberg (1879-1940) at Göttingen. Rosenberg was a colleague of Schwarzschild, and his diagram of the Pleiades star cluster showing photographic apparent magnitude against an index based on the strength of the Balmer lines Hδ and Hζ relative to the K line of calcium, was submitted in June 1910 to the Astronomische Nachrichten. It was the first published ‘Hertzsprung-Russell’ diagram. This interesting fact has been brought to light by Leos Ondra in the Czech Republic. A full English translation of Rosenberg’s paper can be found at www.leosondra.cz/en/ .
Rosenberg’s colour-magnitude diagram for the Pleiades, published in 1910. This was the first CM diagram for a cluster to be published.
It is well known that Hertzsprung, while an amateur astronomer in Copenhagen, produced early versions of his diagram for clusters as early as 1908. It is these diagrams that he showed to Schwarzschild that same year. It is interesting that Rosenberg published his paper just a year after Schwarzschild had met Hertzsprung in Göttingen in 1908. Rosenberg explicitly mentions that Schwarzschild had encouraged him to undertake this work, based on his own objective prism spectra of the Pleiades cluster. Did Schwarzschild leak Hertzsprung’s original ideas to Rosenberg? We do not know, but the coincidence in time and place seems extraordinary.
Meanwhile Russell appears at first to have been unaware of this new work of Hertzsprung, as well as the work of Rosenberg. He presented his own so-called Russell diagram, based on trigonometric parallaxes and Harvard spectral types for field stars, at a Royal Astronomical Society (RAS) meeting in London in June 1913, while on his way to Bonn. At the London meeting he explicitly referred to the terms ‘giant’ and ‘dwarf’ stars for the two distinct categories for later spectral types, although he erroneously ascribed this terminology to Hertzsprung. He expanded on these ideas at a meeting of the American Astronomical Society in Atlanta in December that year, and the published account of his paper contains the first Russell diagram to appear in print. The Russell diagram employed Harvard spectral types for the temperature-related parameter. On the other hand, Hertzsprung had used the effective wavelength, a kind of colour index measured from very low dispersion photographic objective grating spectra.
There are other curious events in the history of astronomy that are quite surprising. It is often assumed that Hertzsprung in 1905 was the first to draw attention to the dichotomy in luminosity displayed by late-type stars. Some are very luminous and distant stars of low proper motion, while others are relatively nearby high proper motion stars of low luminosity. They are respectively supergiants and dwarfs, and Hertzsprung correctly recognized that they can be distinguished spectroscopically, even though he didn’t use that terminology. He showed that stars with the narrowest lines, labelled as c-type stars by Antonia Maury in 1897, were the low proper motion supergiants.
Less well known is the fact that William Monck (1839-1915) in Ireland, an amateur astronomer in Dublin who was also a philosophy professor at Trinity College, had also deduced that later type stars can have either high or low luminosities, like for example Canopus (of high luminosity and low proper motion) and stars like alpha Centauri, Procyon or Capella, which have high proper motion, and are in the latter low luminosity group. No spectral charateristics were used, so Monck did not go as far as Hertzsprung. But this work predated that of Hertzsprung by a decade and was published in 1895.
One final comment. Who was the astronomer who first found the correlation between apparent magnitude and a colour index for stars along the main sequence of a cluster? Was it Rosenberg? Or perhaps Hertzsprung? In fact it was the Swedish astronomer Carl Charlier (1862-1934) at Lund Observatory in 1889! This was well before the concept of colour index was used (that had to await Karl Schwarzschild’s introduction of an index m(pg) – m(vis) in 1900).
What Charlier did in 1889 was study the difference m(pg) – m(vis) for Pleiades stars as a function of apparent photographic magnitude, m(pg). This is the difference in apparent magnitude determined photographically by measuring star images on a plate, and determined visually by the eye through the telescope, in this case using the visual magnitudes of Max Wolf in Heidelberg. To his surprise, Charlier found that the fainter Pleiades stars had progressively larger values of the magnitude difference. He ascribed this to a progressively increasing systematic error in Wolf’s visual magnitudes for the fainter Pleiades stars. He tabulated but did not plot his results graphically. Had he plotted them, he would have plotted the first colour-magnitude diagram showing a main sequence! Nevertheless, he was completely unaware of how close he came to discovering the colour-magnitude diagram. He narrowly missed a great opportunity. Instead his 1889 paper is now almost forgotten. It was published in the Publicationen der Astronomischen Gesellschaft 19, 1 (1889).
(Written 30 September 2009)
