Sometimes experiments lead to new discoveries, as with Galvani’s twitching frog’s legs (see here), and sometimes new ideas, or hypotheses, lead to experiments which prove the hypotheses correct and turn them into theories. Often, there is a synergy between the two processes as scientific knowledge advances. That is what happened with one of the two most important tools used by astronomers, now known as the Doppler effect.
Christian Doppler was working at the Prague Polytechnic, the forerunner of the Czech Technical University, in the 1840s, and knew about the relatively recent work of Thomas Young and Augustin Fresnel (see here) establishing the wave nature of light. He realized that if a source of light is moving towards an observer, the waves seen by the observer would be squashed to shorter wavelengths (towards the blue end of the spectrum); if it is moving away, the waves would be stretched to longer wavelengths (towards the red end of the spectrum). He thought that this would affect the observed colours of stars, and published a paper drawing attention to this in 1842 (the title translates as On the Coloured Light of the Binary Stars and some other Stars of the Heavens). He wrote: ‘nothing seems to be more intelligible than that, to an observer, the path length and the interim durations between two consecutive breakings of a wave must become shorter if the observer is hurrying toward the oncoming wave, and longer if he is fleeing from it.’
He was right, but the idea that this would affect the colour of stars turned out to be wrong. Stars do not move fast enough for their colour to be noticeably affected by their motion, although more than a century later a related effect turned out to be important in studies of much more distant objects known as quasars. But the Doppler effect, as he realized, also applies to sound waves moving through the air. If the source of the sound is moving towards you, the sound waves are squashed to a higher pitch; if it is moving away, they are stretched to a deeper note. This effect is very familiar today, as we all notice it when the note of the siren on an emergency vehicle shifts downward (a ‘down Doppler’) as the vehicle rushes past. But how could the idea be tested in the 1840s?
In 1845, the Dutchman Christoph Buys Ballot devised a brilliantly simple experiment to measure the effect. The fastest vehicle available at the time was a steam train, and he arranged for a group of horn players to play a particular note on the open carriage of a train as it steamed along the Utrecht-Amsterdam line past another group of musicians. These listening musicians had perfect pitch, and were able to describe precisely the way the note changed as the train steamed past them. So the experiment confirmed Doppler’s principal hypothesis; but Buys Ballot was also one of the first to point out that the idea that this would affect the perceived colour of stars was wrong.
This, however, was not the end of the story. In 1848, the Frenchman Hippolyte Fizeau realized that certain dark lines in the spectrum investigated by Joseph Fraunhofer would be shifted by the Doppler effect. The origin of these lines was not understood at the time (see here), but it was known that they occurred at precise wavelengths in the spectrum of light from the Sun. If the lines were shifted towards the red (redshift), the amount by which they were shifted could be used to measure how fast a star is receding from us, and if they were shifted towards the blue (blueshift), they could be used to measure how fast a star is approaching. This was put into practice by the British astronomer William Higgins in 1868, when he carried out the first measurements of the speed of several stars relative to the Earth. In binary systems, as one star orbits around another, the Doppler effect moves the lines to and fro, giving a measure of how fast the star is moving. Combined with other observations, this makes it possible to measure the masses of stars in binaries – a far cry from listening to trumpeters on a train.
It is worth pointing out, though, that the famous cosmological redshift is not a Doppler effect, even though many popular accounts (and even some textbooks) incorrectly refer to it that way. Although it is measured in the same way, in terms of features in the spectrum of light from distant objects being shifted towards the red end of the spectrum (to longer wavelengths), it is not caused by the motion of those objects (distant galaxies and quasars) through space, but by space itself stretching and stretching the light waves with it as the Universe expands.