Crucial Tests?: The Michelson-Morley Experiment

by Douglas Allchin

Cantankerous philosopher Imre Lakatos used harsh words for crucial experiments. How could one experiment, even a well controlled one, arbitrate between two complex sets of theoretical concepts? No single decisive experiment was possible on a grand scale, Lakatos claimed, and he boldly pronounced the "end of instant rationality."

The notion that controlled experiments are not decisive may not sit easily with any teacher who struggles daily to convey the central notion of control to maturing students. Yet Lakatos was not abandoning science. Rather, he urged us to appreciate the complex and multifaceted nature of experimental evidence. Controls narrow the field of permissible conclusions. But because any experiment relies on many assumptions and concepts, one set of test results alone can never itself justify the final word.

Lakatos took aim at, among other targets, the mythic Michelson-Morley experiment. In standard textbook accounts, two physicists set out in 1887 to measure the speed of the earth through the ether, which in a Newtonian universe must serve as the medium for light waves. However, they found no effect—hence (apparently), no ether. And so, this version of history contends, Einstein was primed to develop (and other physicists to accept) relativity as an alternative theory that solved the anomaly of the elusive ether. The Michelson-Morley experiment was thus crucial to the development of modern physics.

In reality, the "crucial" 1887 experiment was part of series of studies—an initial clue that the conventional story is misleading. Moreover, neither Michelson, nor Morley, nor others at the time interpreted the results in the context of the existence or non-existence of the ether. More was at stake than a simple "ether-or" decision.

No one might have worried about the ether at all if it had not helped explain astronomical aberration—the observation that stars do not always seem to be where they belong. If, however, light waves traveled through a medium, and the medium itself moved, then starlight reaching Earth might be affected by the ether. Two possibilities arose: perhaps the Earth moved into and through the ether, creating an "ether wind" across the Earth's surface (as proposed by Augustin Fresnel); or the Earth held ether close to its surface, dragging it along as it orbited and leaving any movement effect higher up (proposed by G.G. Stokes).

Albert A. Michelson (1852-1931)—at first, working alone—was inspired by a technical challenge posed by this scenario. James Clerk Maxwell, major architect of the electromagnetic theory of light, suggested in 1878 how one might measure the movement of the Earth relative to the ether—outlining what would become the core of Michelson's experiment—by observing and comparing the velocity of light in different directions. Maxwell claimed, however, that no method would be able to detect the subtle difference involved. Of course, for a good experimenter like Michelson, that was more challenge than discouragement. Indeed, Michelson later wrote to Einstein that he could devote so much energy to precison for its own sake because he found it "fun." Michelson's initial aim, therefore, was to document the Earth's movement through space: a spectacular historical achievement, if he could succeed.

Michelson made every effort to detect the movement. He transfered his experiment out of town when horses on the street appeared to jiggle his apparatus. He dampened extraneous movement by floating his apparatus in a bath of mercury. But Michelson found no observable drift in light waves. A disappointing "negative" result. Michelson could point out, though, that--given the measurement accuracy of his equipment--Fresnel's explanation of aberration was not likely. That left Stokes's notion--with the ether still fully functional. Not much revolutionary in that--and most physicists took no note of the results.

Michelson turned to other investigations, including a replication of Fizeau's 1852 measurements of the velocity of light through running water. Michelson needed a glassblower with some finesse to construct the equipment and in 1886 he teamed up with a chemist, Edward W. Morley (1838-1923). Their result, incidentally, suggested that a moving medium had no effect on waves traveling through it— hence the whole speculation on the ether's role in astronomical aberration was suspect.

In the same year, the team elected to redo Michelson's earlier experiment, hoping to achieve more precision with better equipment. Published in 1887, this is the experiment most frequently cited. Yet the results were not wholly negative. Michelson and Morley could report only that the relative speed of the Earth and ether was less than one-sixth the velocity of the Earth—enough to discount Fresnel's explanation once again, but not enough to dispense with an ether wind altogether.

In any event, by 1895, Lorentz and others were suggesting ways to reconcile the apparent absence of movement with the presumed nature of the ether (for example, if bodies contracted while moving through the ether).

In 1897, Michelson was still interested in the basic experiment and, following an idea remaining from a decade earlier, he performed it at different altitudes (harkening back to the different notions of Fresnel and Stokes). Altitude made no difference. That might have given an advocate of Stokes's ideas a sense of how much ether was being carried by the Earth's atmosphere. It also allowed for other theoretical possibilities, such as Lorentz's noted above.

Michelson received a Nobel Prize in 1907, but his now acclaimed experiments were not a part of the citation. Instead, Michelson was acknowledged for his instruments—for example, the interferometer's contribution to establishing a new standard for defining the meter, through the velocity of light.

Finally, by 1925, another experimenter (D.C. Miller) had claimed to have detected an ether wind. Michelson was interested enough to recheck his experiment. The result, once again, was negative. But it is striking to note that the issue of ether was still alive--despite Einstein's theories. The changing horizon of Michelson's research across four decades shows how no individual result was "crucial" historically.

How did the complex history of Michelson become flattened into the sometimes mythic status of one experiment? As illustrated in Strick's analysis of Pasteur's swan-necked flasks, the experiment proved a powerful resource for showing progress in a synopsized fashion. Einstein's theories developed mostly independently of Michelson's specific results. Science did not progress from unambiguous, prophetic experimental results, but only captured those favorable results in its fold by looking backwards. And from the new vantage point, the results looked very different indeed. Thomas Kuhn has pointed out that textbooks often "rewrite" history from the perspective of current knowledge, erasing the context in which the results and even the aim of an experiment itself could have been conceived in wholly different terms. Good historical case studies can remind us, as always, that science understood prospectively is different than science understood retrospectively.

Further Reading