This article discusses two classical papers in the history of astronomy, namely Christian Doppler’s (1803–53) announcement in May 1842 of the effect named after him, and a major article by Hermann Carl Vogel (1841–1907), which appeared in May 1892. Vogel’s work represented the first successful application of the Doppler effect to stars, to determine their space velocities in the line of sight.

Doppler’s theoretical paper and Vogel’s painstaking instrumental technique have together provided the basis for our knowledge not only of stellar motions, but have also indirectly contributed to our knowledge of stellar rotation, of thermal and turbulent Doppler line broadening in stellar (and other) spectra, of stellar masses in many binary stars, of galaxy masses from their rotation or velocity dispersion, of the missing mass problem in galaxies and clusters, of the expansion of the universe, of the nature of quasi-stellar objects and of the existence of the Big Bang. Indeed astronomical knowledge would almost certainly be vastly poorer if Doppler’s principle had never been applied in astronomy.

The circumstances in which these two great papers were born were totally different. Doppler’s was the work of a little known Austrian mathematics professor in Prague, who had no practical knowledge of stellar spectroscopy. His paper made several quite false assumptions, and its predictions for light waves were generally disputed until well after Doppler’s death in 1853, and its conclusions were mainly completely false.

Doppler’s theoretical prediction for the apparent frequency change of a light wave when either source or observer were in motion relative to the ether was a more or less preliminary result, which he then used as the basis for his paper entitled: ‘On the coloured light of double stars and of some other stars in the sky’. His attempts to explain star colours as a consequence of their motion were largely erroneous. Indeed he can be regarded as fortunate to be remembered for posterity, given the unproductive ideas to which most of his paper was devoted.

On the other hand, in 1892 Vogel was the director of the recently-founded Potsdam Astrophysical Observatory, which was then quickly becoming one of the most prestigious astronomical institutions of Europe. His observatory was equipped with the best instruments and employed outstanding scientists, whose reputation for careful technique and attention to detail was unequalled. Together with Huggins in London and Pickering at Harvard, Vogel was one of the pioneers in the application of photography to record stellar spectra, which proved to be an indispensable tool for measuring the minute Doppler shifts of spectral lines. The techniques which he established between 1887 and 1892 were copied and refined at many other observatories, as stellar radial-velocity research became one of the most popular branches of astrophysics over the ensuing half century.

Christan Doppler

In 1841, after attending the Polytechnic Institute in Vienna, Christian Doppler was appointed as a professor of mathematics and practical geometry at the State Technical Academy in Prague. Here he became an associate member of the Royal Bohemian Scientific Society in 1840. It was to this society that he delivered his famous lecture on 25 May 1842. The title of the paper showed that his principal aim was to account for the colours and magnitudes of double star components and of variable stars, since he believed both parameters to be influenced by stellar motion in the line of sight. Doppler treated light waves as transverse displacements in the ether and he made the analogy of light waves both with sound waves and waves in the sea. The cases of the motion of the source and of the observer were treated separately. But in either case a frequency change occurred, for which Doppler derived what is essentially his familiar equation of Δν/ν₀ = V/c.

Doppler made two incorrect assumptions in his analysis: first he supposed that the radiation from stars was largely confined to the visible region of the spectrum, and secondly he supposed that stars frequently moved through space at a significant fraction of the speed of light. Having made these incorrect premises, it was a fairly simple step for Doppler to predict that the apparent colour and brightness of any star would depend on its state of motion in the line of sight. An approaching object will appear brighter and bluer, a receding one fainter and redder than the same star at rest. He assumed most stars to be intrinsically white or slightly yellow and the existence of coloured stars to be direct evidence in support of his theory!

Title page of Doppler’s famous paper of 1842

Intrinsically white stars could be rendered completely invisible if they receded at more than 106000 km/s, because their radiation would be shifted entirely out of the visible region, whereas only 183 km/s sufficed to impart a perceptible orange colouration, while at 1040 km/s the colour change was described as striking. These arguments led Doppler to believe that stars in general have space motions typically of tens of thousands of kilometres a second and not infrequently of as much as 10⁵ km/s, values now recognized to be in excess of actual stellar velocities by a factor of some 10³.

Double stars provided, so he believed, a special confirmation of his theory, for he stated that they either consisted of a bright white primary and a fainter coloured companion, or alternatively of two stars of comparable magnitude but complementary in colour, such as one blue star and the other red. His source of double star photometry was the visual data of William Herschel and Wilhelm Struve, and indeed many visual doubles might well appear to conform to such a classification. The interpretation, based on orbital motion and Doppler’s principle, would then account very satisfactorily for such observations. The fact that Herschel and Struve sometimes gave differing colour descriptions for the same visual binary components was clearly due to orbital motion resulting in an apparent colour change from Herschel’s to Struve’s times of observation, so Doppler argued.

Variable stars also appeared to lend striking support for Doppler’s hypothesis. If these were members of a binary system, then the periodic changes in brightness of stars such as Algol and Mira must result from their periodic orbital motions. Even the novae of 1572 and 1604 (in reality supernovae) could be explained on the binary hypothesis, their rapid rise and slow decline being a consequence of eccentric orbits. Their increasingly reddish colours on fading were naturally explained by an increasing recessional orbital velocity.

On reading Doppler’s paper, written as it is in a ponderous and repetitive Teutonic style, one is nevertheless impressed with the inevitable logic of his conclusions, even though they were based on faulty premises. And in searching for observational evidence to support his theory, one must admit that the scanty available data on double and variable stars – only 53 were known of all types of variable in the mid-19th century – must have seemed as compelling evidence in support of his ideas.

Doppler’s paper was immediately controversial. Devising empirical laboratory tests of the optical Doppler effect was impossible in the mid-19th century, so astronomical evidence to support Doppler’s work had to be found. In the event this evidence was over a quarter of a century in coming, which gave Doppler’s critics ample time to raise often spurious arguments to attack him. C. H. Buys-Ballot in Holland was one of those who contested Doppler’s prediction of colour changes of stars, on the grounds that the ultraviolet or infrared regions should be shifted into the visible spectrum and hence no colour change would then result. However Buys-Ballot was soon (in 1845) able to confirm the correctness of Doppler’s principle for sound, by conducting an experiment on the Dutch railway to observe the apparent change in pitch of wind instruments played on passing trains.

Hermann Carl Vogel

In Germany, Vogel repeated this experiment many years later, using the relatively pure note of the steam whistle from a locomotive as it passed the observer. In the United States, B. Sestini, an Italian Jesuit at Georgetown College, studied the colours of 400 double stars in 1850, and believed he found some weak evidence in support of Doppler. Doppler in 1852 enthusiastically wrote an article showing how the unreliable visual data of Sestini confirmed the application of his theory to starlight. Altogether Doppler wrote at least five further papers on the coloured light of double stars, mainly to defend his original ideas of 1842 from the criticism of others.

It should be recalled that Doppler’s paper made no reference to stellar spectroscopy. The only significant observations of stellar spectra up to the 1840s had been those of Fraunhofer in 1814–15 and again in 1823. It is therefore quite remarkable that A. H. Fizeau, in a lecture to the Société Philomatique in Paris in December 1848, should have grasped the true nature of Doppler’s principle as applied to starlight, and predicted a shift in stellar spectral lines: ‘… each ray … will take the place of the ray which possessed this same wavelength when the luminous body was at rest; all the rays will thus replace each other in such a way that the lines will no longer be in the same places, but are all displaced towards the red or towards the violet, according to the direction of motion of the luminous body. The colours on the other hand … will suffer no displacement … It is noted that this result only depends on the velocity of the luminous body and not at all on its distance. Such observations could thus lead to data on the intrinsic velocities of the most distant stars … .’ Unfortunately this far-sighted prediction remained unpublished and hence largely unknown until 1870, by which time Ernst Mach in Vienna had independently reached the same conclusion (in 1860). The theoretical basis for measuring the line-of-sight velocities of stars had thus been laid down before any substantial observations of stellar spectra after Fraunhofer were undertaken.

The rebirth of stellar spectroscopy in part came in 1862–63 with the independent work of Rutherfurd, Huggins, Secchi and Airy. The controversy over the reality of the optical Doppler effect continued until astronomers had developed spectroscopic instrumentation able to detect and measure the small line shifts produced by objects with known line-of-sight velocities. In 1870 Secchi proposed observing the east and west limbs of the solar photosphere to detect solar rotation and indeed he reported a difference in the line positions in spectra from the two limbs. This result was confirmed quantitatively by Vogel in 1871 by using a spectroscope to display simultaneously the two spectra from opposite limbs. His result was close to that predicted from sunspots to determine the solar rotation rate. This can therefore be regarded as the year when the validity of the Doppler effect for light first received direct observational confirmation.

A number of observers during the 1870s and 1880s tackled the solar rotation rate using the Doppler principle, including C. A. Young and H. Crew in the United States, L. Thollon and A. Cornu in France and N. C. Dunér in Sweden. Cornu’s work in 1884 on the D line shifts was notably precise. The measured shift was within 3 per cent of the predicted value. In 1891 Vogel was able to confirm the validity of the effect for bodies shining in reflected light – in this case by observing the shift of the lines in a photographic spectrum of Venus and comparing the results with the known orbit of this planet.

When the Potsdam Astrophysical Observatory building was completed, Carl Vogel became the first director in 1882. His institution attracted some of Germany’s most outstanding astrophysicists. Vogel’s personal interest was the application of photography to stellar spectroscopy, for the purposes of spectral classification and the determination of Doppler shifts. J. Wilsing and J. Scheiner were his close collaborators in these spectroscopic researches and in particular Scheiner undertook much of the work to determine stellar radial velocities.

Vogel’s two-prism spectrograph on the 30-cm refractor at Potsdam

Vogel began his first trials in the spectrographic programme with Scheiner at Potsdam in 1887. He declared that: ‘The observation of the line shifts in stellar spectra belongs to one of the most difficult astronomical measurements…’. His apparatus consisted of a two-prism slit spectrograph on the 30-cm Potsdam refractor. A thin hydrogen Geissler discharge tube, mounted 42 cm ahead of the slit inside the telescope, was the comparison source. The Hγ emission line from the Geissler tube gave the zero-point in velocity against which the stellar spectral plates were measured. It was exposed superimposed on the stellar spectrum.

The main Doppler-shift programme at Potsdam was undertaken from 1888–92 by Vogel and Scheiner, using an improvement on the spectrograph initially employed. In his classical paper of 1892 Vogel reported the Doppler shifts of 51 bright stars based on 252 stellar spectrograms exposed at the 30-cm refractor. The spectra were widened by tracking the telescope in right ascension. The exposure times were limited to about 1 hour, because of the problem of thermal changes in the instrument affecting the quality of focus. Presumably spectrograph flexure would also have been a factor limiting the exposure time, even though the spectrograph weighed only 12 kg. The line positions in stellar and comparison spectra were measured in a travelling microscope and an interpolation formula devised by Hartmann at Potsdam used to convert these to wavelengths and hence Doppler shifts. Finally a correction was applied for the orbital motion of the Earth.

Title page of Vogel’s famous paper of 1892, in which the Doppler shifts of stars were first recorded spectrograhically

One of the greatest early triumphs of photographic stellar spectroscopy was the discovery of spectroscopic binary stars. In 1889 Vogel had found that the famous eclipsing double star Algol showed radial-velocity variations. The system comprises a luminous B dwarf which orbits in about 69 hours a fainter K subgiant companion. As the B star dominates the light, only its lines are seen. This was the first single-lined spectroscopic binary to be discovered. Vogel’s paper submitted in December 1889 came just one month after Pickering at Harvard had announced periodic doubling in the lines of ζ UMa observed on objective prism spectrograms. This and β Aurigae discovered shortly afterwards at Harvard were the first double-lined spectroscopic binaries, in which the two stars in the system are of comparable brightness but still unresolved in the telescope. Vogel also announced the discovery of a second single-lined spectroscopic binary, Spica, in 1890. Spectroscopic binary stars were thereafter discovered in considerable numbers, especially by W. W. Campbell and his colleagues at Lick.

The overall success of Vogel’s photographic techniques allowed reliable space motions of stars in the solar neighbourhood to be determined for the first time. As it happens all the stars observed at Potsdam had what we now recognize as low radial velocities, ranging from +12.0 km/s for Aldeberan to –9.6 km/s for γ Leonis. No high velocity objects were included. But in less than a decade, Campbell at Lick had published a list of seven stars with radial velocities in excess of 76 km/s.

Photographic spectra of bright stars at H gamma recorded by Vogel at Potsdam, showing also the emission line from a discharge tube used for the velocity zero-point.

Apart from the discovery of the single-lined spectroscopic binary stars, the most important legacy of the Potsdam radial-velocity programme was the establishing of a reliable technique for the determination of stellar Doppler shifts. This method was quickly copied by A. A. Belopolsky at the Pulkova Observatory, who entered this field as early as 1891, and also by H. A. Deslandres in Paris from about 1892, by H. F. Newall in Cambridge from 1896, by Henry Lord in Columbus, Ohio, also from 1896, and above all by Campbell at the Lick Observatory from 1894. Thus at least five other observatories undertook photographic Doppler shift measurements in the 1890s. Although the techniques of photographic radial-velocity determination were greatly refined, both in instrumentation (diffraction gratings, coudé spectrographs, Schmidt cameras, faster photographic emulsions) and reduction methods (standard velocity stars, selection of unblended lines at different spectral types, use of new laboratory rest wavelengths), the fundamental principles employed for about eighty years or more were those originally devised by Vogel. Because of its profound influence on other observers of stellar radial velocity, Vogel’s paper of 1892 was a landmark in stellar astronomy. It was also a final vindication of at least some of Doppler’s original ideas of a half century earlier, and in hindsight a remarkable confirmation of the less heralded though more perspicacious theories of Fizeau and of Mach.

(written 24 May 2009; based on an article by the author in Vistas in Astronomy 35, 157 (1992))

This entry was posted on Thursday, July 16th, 2009 at 12:00 am and is filed under Doppler effect, History of astronomy. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.