Because their experimental apparatus seemed to be at rest in the aether, Michelson and Morley initially thought that a bubble of aether might be stuck to the earth, moving with us through the sea of aether that fills outer space. This wasn’t a new idea. Decades earlier, other physicists had speculated that moving objects might drag some aether along, in the same way that water is carried by the sodden planks of an old ship. So, when the Michelson-Morley experiment failed to detect any flow of aether, this aether drag hypothesis seemed to provide a ready explanation. The experiment had been conducted in a basement laboratory, underground; but if aether doesn’t flow freely through matter, a basement might be shielded from the “wind” of aether blowing past. In the concluding section of their experimental report (published 1887), Michelson and Morley suggested that aether wind might be detected if the experiment were performed at a higher altitude, “at the top of an isolated mountain peak, for instance.”Albert A. Michelson and Edward W. Morley, “On the Relative Motion of the Earth and the Luminiferous Ether,” American Journal of Science 34 (1887): 341. Available here.
Unfortunately, the mystery was not so easily solved. Such experiments were eventually performed atop a mountain, to no avail.In the early 1920s, Dayton Miller conducted a series of experiments at the Mount Wilson Observatory, and claimed to observe a very small amount of aether wind, much less than expected. (You can read his report here.) His results were inconsistent, however, and later physicists found that he had misinterpreted his data. See this article for more information. Moreover, serious problems with the aether drag hypothesis were well-known, long before the Michelson-Morley experiment. The strongest evidence against aether drag, in fact, involved a phenomenon that had been discovered nearly a century before the aether drag hypothesis was even proposed. In 1728, astronomer James Bradley discovered that when stars are viewed through a telescope, their positions seem to shift slightly in the direction of Earth’s motion around the sun. For example, a star located directly overhead in the midnight sky will appear through a telescope as though it is located at an angle more than 20 arcseconds (0.006 degrees) from its actual position.In most cases, the apparent displacement is less than 20.5 arcseconds. The maximum aberration (known as the constant of aberration) only occurs with starlight that is exactly perpendicular to the earth’s direction of motion. This phenomenon, called stellar aberration, provides compelling evidence against the aether drag hypothesis, as I will explain.
Stellar aberration—the shift or displacement in the apparent position of a star viewed through a telescope—is most significant when starlight enters a telescope at an angle perpendicular to the earth’s motion (relative to that star’s reference frame). To see why this happens, imagine a ray of starlight traveling directly toward the sun. Suppose Earth moves into the path of this ray as we orbit the sun, and the ray enters the top of a telescope, which is pointed toward the night sky. As the ray travels at 300,000,000 m/s (the speed of light) from the top lens of the telescope toward the eyepiece at the bottom, the telescope itself moves in a perpendicular direction at 30,000 m/s (the speed of Earth’s orbit around the sun).
Because the telescope is moving while the starlight passes through it, the telescope must be tilted slightly in order for the star to be visible at the center of the eyepiece. For example, if an astronomer wants to look at a star located directly overhead, she has to angle her telescope slightly forward, toward the direction in which Earth is moving. That way, the top lens is centered beneath the starlight when the ray first enters the telescope, and the eyepiece moves to the center of that ray by the time it reaches the bottom.
If the aether drag hypothesis were correct, however, stellar aberration would not occur. Any telescope that is at rest relative to Earth would also be at rest relative to the bubble of aether that supposedly engulfs our planet. The reference frame of the telescope and the reference frame of the aether bubble would be one and the same. In that case, rays of starlight should stay parallel to the telescope as they travel from the top lens to the eyepiece, since the telescope does not move sideways through the aether bubble. The telescope would not have to be tilted at an angle to the incoming starlight in order for that light to hit the center of the eyepiece. Stellar aberration would not occur. But stellar aberration does occur! Therefore, the aether drag hypothesis is incorrect.
Why didn’t the Michelson-Morley experiment detect any aether wind, then? Dutch physicist Hendrik Lorentz suspected that aether had played a cunning trick on the experimenters. We’ll consider his hypothesis on the next page.