Is Technology Too Good for an Old-School Test of Einstein’s Relativity?

By Terena Bell | May 5, 2017 11:45 am
July 11, 2010 eclipse Image as viewed from Easter Island in the South Pacific. (Credits: Williams College Eclipse Expedition - Jay M. Pasachoff, Muzhou Lu, and Craig Malamut)

July 11, 2010 eclipse Image as viewed from Easter Island in the South Pacific. (Credits: Williams College Eclipse Expedition – Jay M. Pasachoff, Muzhou Lu, and Craig Malamut)

On Aug. 21, sky-gazers from around the world will converge in the United States as a total solar eclipse charts a path from Oregon to South Carolina. In between, on Casper Mountain in Wyoming, you’ll find Don Bruns with his telescope.

A retired physicist, Bruns is using the rare opportunity to test Albert Einstein’s general relativity like Sir Arthur Eddington, who was the first scientist to test the theory back in 1919. At that time, Newton’s law of universal gravity was still vogue, but Einstein shook the status quo by introducing his theory of general relativity, which fused concepts of time and three-dimensional space into a four-dimensional continuum called space-time. According to Einstein, gravity wasn’t a force; instead, it was a distortion in the fabric of space-time.

The First Test

His theory was just four years old when Eddington put it to the test during an eclipse in 1919. Both Einstein’s and Newton’s theories indicated that light from distant objects would warp, or bend, as it passed through gravitational fields of massive objects, such as the sun. Einstein’s theory indicated the sun would bend light from a star by a miniscule 1.75 arcseconds—just 0.000486 of a degree—while Newton’s gravity predicated starlight would bend by half that amount. To see who was correct, you’d just need to compare the apparent positions of stars at night to their position during the day, when their light passed through the sun’s gravitational field.

Einstein Newton angle

An eclipse was the perfect time to perform such a test. In January of 1919, Eddington photographed the real positions of stars in the Hyades cluster, because light from these stars passes through the sun’s gravitational field. But to perform the second half of the experiment, one needs to photograph the same stars during the day, but in darkness. In May, he photographed the same group during an eclipse, when stars in the Hyades remained visible. By laying an image of real star positions over images of their positions during the eclipse, Eddington could measure the apparent shift caused by the sun’s gravity.

When Eddington compared the two images, a shift was clearly visible. And based on his calculations, Eddington concluded that the theory of relativity more accurately predicted the amount light bent than Newton’s.


One of Eddington’s photo plates from his 1919 experiment. The shift in starlight is marked with faint lines in the image. (Credit: Wikimedia Commons)

Just like when scientists at the Laser Interferometer Gravitational-Wave Observatory confirmed the first detection of a gravitational wave in 2015, Eddington’s discovery was front-page news. He became a star. However, upon further inspection, the results left much to be desired for many scientists. Eddington used silver nitrate photographic plates to capture images of the sky, and the image resolution wasn’t precise. Further, one of the three photographic plates yielded a measurement that would have confirmed Newton’s theory of gravity, but it was removed from the experiment due to a “systemic error.”

Eddington was also a known pacifist and some scientists believed his experiment was wrought with confirmation bias. Following World War I, relations between scientists in Germany and the United Kingdom were frayed to say the least. Some speculated that Eddington, a Briton, disproved compatriot Newton in favor of Einstein, a German, in an act of international scientific diplomacy to cure old wounds.

Even Stephen Hawking had doubts about Eddington’s results, as he wrote in his Brief History of Time. Still, other physicists argue that if presented with Eddington’s data today, many physicists would come to the same conclusion.

Another Test

Thus, we return to citizen scientist Bruns who, in a sense, will try to vindicate Eddington and prove his findings stand up. But nearly 100 years later, Bruns’ attempt may be foiled by technology that’s actually too good.

Modern photography has long since moved beyond plates, but in this case, the march of technological progress doesn’t necessarily yield better results. To get clearer resolution, digital cameras use pixels that are too large, even on a micro level, to catch the apparent position shift that proves relativity.

Eddington’s original experiment showed that the sun’s starlight bends apparent position by a mere 1.75 arcseconds. When light falls on the line between two pixels, imaging software nudges it into a neighboring pixel. So if a star’s centroid falls on the line, the picture will reveal variations that are bigger or smaller than what really exists.


Bruns’ setup for eclipse day. (Courtesy: Don Bruns)

“It’s the most challenging experiment I’ve done in my career,” he says. In 1973, the last time a relativity experiment was conducted during a total solar eclipse, technology “was considered ancient. I mean glass plates, no computers, it’s really a different era. So I thought in 2017…it’d be a piece of cake, not a problem. Well, turns out it’s still pretty tough.”

Bruns claims that there’s “no scientific value” to what he’s doing; he says he’s just testing relativity for fun. In February 2016, you’ll remember, scientists announced the first gravitational wave detection, confirming Einstein was correct all along. Still, like the number-hunters who calculate pi to an ever-increasing number of digits, sometimes the act of science is a reward in itself.

Alex King, chair of physics and astronomy at Austin Peay State University, would disagree with Bruns. King believes there’s certainly a compelling scientific reason to conduct an old-school experiment. The resolution flaws in Eddington’s photos were reason enough for APSU to seek funding to conduct the same experiment this summer. But APSU’s efforts were blocked by what King calls the politicization of “funding for basic science”—they couldn’t get the money.

Fortunately, Bruns has the experience to pull it off on ecplise day. In 1992, Bruns founded a company called Stellar Products that made standard adaptive optics systems. Bruns left Stellar in 2014, but that experience gave him deep knowledge about equipment options that would work well. When he chose the Finger Lakes ML8051 camera, the manufacturer loaned him one for free. He’ll then mount the camera on a TeleVue NP101 telescope to optimize focal ratio. Unfortunately, so long as it relies on pixels to make a picture, there’s going to be some margin of error. After a few trial runs, though, Bruns thinks he’s trimmed the margin to roughly 1 percent. He says other citizen scientists who’ve attempted the experiment only managed to narrow their margins of error to 10 percent.

The camera Bruns opted for is monochrome—it shoots in black and white. Digital color photography works like pointillism; instead of painting a solid image, it captures the picture in tiny dots. According to Bruns, when a digital camera shoots in color every other pixel has a different response to light, which affects accuracy.

“They used to have these very long, 15-, 20-foot long telescopes operating at a very small focal ratio,” he says, referring to Eddington and other scientists in Einstein’s era. “But now we have a very short compact telescope. It’s only like 2 feet long, but optical distortion is now a problem.”

Shorter, modern telescopes correct coma and optical aberrations, but the error they cause in the distortion is bigger than the error from Einstein’s deflections, Bruns says. In other words, no matter what, Bruns is making trade-offs.

Just for Fun

In Bruns’ case, one of the largest trade-offs is his time. By the time he spoke with Discover, he’d already spent around 2,000 hours solving for distortion alone. That’s out of nearly 4,000 project hours total, the rest of which he devoted to sorting through star charts for base comparison data, writing software to better process his data and addressing every other part of the experiment that could affect margin of error. He even studied how to minimize the impact of breeze and atmospheric turbulence, which Burns says can spread starlight across three to four pixels.

During totality, Bruns plans to take calibration images, something that scientists of yore decided not to do. Ask why and he’ll tell you about Erwin Finlay-Freundlich, a physicist who “was a little frustrated [that] after 3 or 4 eclipses…no one did the experiment quite right.”

“You need to take pictures of the stars during the eclipse but you also need to take calibration images during the eclipse,” says Bruns citing Finlay-Freundlich’s method. “And a lot of people of course didn’t want to do that because the eclipses only last 3 to 5 minutes, and the exposures each took a minute or two, so they didn’t want to waste time during totality taking calibration images.”

At its point of greatest duration, the 2017 eclipse will last 2 minutes, 40.2 seconds. Fortunately, digital photography can get the required exposure in milliseconds. This means Bruns can test relativity Finlay-Freundlich’s way. And when he does, he says he’ll be the first person in history to do it from the ground.

“Just for fun” indeed.

CATEGORIZED UNDER: Space & Physics, Top Posts
MORE ABOUT: stargazing
  • Uncle Al

    Radio astronomy interferometry is not blinded by sunlight for solar edge-skimming sources. GR gravitational lensing includes higher order effects. Astrophys. J. Lett., 819(1) L8 (2016), arXiv:1512.04654, doi:10.3847/2041-8205/819/1/L8

    A 2159 mile diameter ball of rock will wholly block the sun. The Allais effect, specifically a Foucault pendulum going wonky during a solar eclipse, has been denied and obfuscated. I wrote to NASA suggesting the Allais effect as it is written be falsified or verified. “There is no reason to look because it cannot be true.” I mourn for science, Allais effect (arXiv:gr-qc/0408023) and Mpemba effect (arXiv 2000-2017, 13 arguments).

    • OWilson

      A penny can block the sun (from the eye) if held at the right distance. In this case “size doesn’t matter”. :)

      But there are lots of abiguities and paradoxes when contemplating the Foucalt phenomena, absolute and relative motion, then there’s Newton’s Bucket with centrifugal force, rotating space ships with artificial gravity, accelerating elevators creating gravity, conflicting with the theory that gravity is just a “distortion” in spacetime.

      Then there’s the universal tendency to spirals and vortexes, in everything from a DNA molecule to seeds and plants, to the kitchen drain, to the galaxies.

      And we still haven’t touched the uncertain world of the atom!

      • Uncle Al

        lots of abiguities and paradoxes when contemplating the Foucalt phenomena” Foucault . Videotape it from above before, during, and after. Frame rate is the time stamp. You want a high latitude re sine(latitude) for angular rotation rate. Clouds don’t matter.

  • Xinhang Shen

    I would like to inform you that Einstein’s relativity theory has already been disproved both logically and experimentally (see “Challenge to the special theory of relativity”, March 1, 2016). It is a complete waste of money and time to verify general relativity again.

    The most obvious and indisputable experimental evidence, which everybody with basic knowledge of special relativity should immediately understand: is the existence of the absolute time shown by the universally synchronized clocks on the GPS satellites which move at high velocities relative to each other while special relativity claims that time is relative (i.e. the time on each reference frame is different) and can never be synchronized on clocks moving with relative velocities.

    The corrections of the atomic clocks on the GPS satellites are nothing to do with relativistic effects because the corrections are absolute changes of the clocks, not relative as claimed by special relativity. After all corrections, the clocks are synchronized not only relative to the ground clocks but also relative to each other.

    Some people may argue that the clocks are only synchronized in the earth centered inertial reference frame, and are not synchronized in the reference frames of satellites. If it were true, then the time difference between a clock on a GPS satellite and a clock on the ground observed in the satellite reference frame would grow while the times of the same clocks observed on the earth centered frame keep synchronized. If you corrected the clock on the satellite when the difference became significant, the correction would break the synchronization of the clocks observed in the earth centered frame. That is, there is no way to make a correction without breaking the synchronization of the clocks observed in the earth centered frame. Therefore, it is wrong to think that the clocks are not synchronized in the satellite frame. Actually, on the paper mentioned above, I have proved that if clocks synchronized in one inertial reference frame, then they are synchronized in all inertial reference frames because clock time is absolute and universal.

    The simplest thought experiment to disprove special relativity is the symmetric twin paradox: two twins made separate space travels in the same velocity and acceleration relative to the earth all the time during their entire trips but in opposite directions. According to special relativity, each twin should find the other twin’s clock ticking more slowly than his own clock during the entire trip because of the relative velocity between them as we know that acceleration did not have any effect on kinematic time dilation in special relativity. But when they came back to the earth, they found their clocks had exact the same time because of symmetry. This is a contradiction that has disproved special relativity. This thought experiment demonstrates that relativistic time is not our physical time and can never be materialized on physical clocks.

    • Uncle Al

      ” Einstein’s relativity theory has already been disproved” Grammar aside, a confluence of overwhelming ignorance with overweening arrogance.

      .,.with pictures, for the textually-challenged.

      …1) As for GPS, you are not even wrong.

      …short form

      …complete DOI:10.1017/S174392130999010X

      …2) 1.74 solar- mass 465.1 Hz PSR J1903+0327 and a 1.05 solar- (AP4 model) vs. -0.000369% gravitational binding energy (GBE); 1.8×10^11 vs. 30 surface gees, 2×10^8 gauss vs. 5 gauss magnetic fields; compressed superconductive protons and superfluid neutrons vs. 11.8:1 atom-abundance hydrogen:helium plasma; extreme isospin and lepton number divergences; and pulsar 11% of lightspeed equatorial spin velocity are GR violation-inert for orbit, periastron precession, and gravitational radiation orbital decay.

      …3) 2.01 solar mass 25.58 Hz PSR J0348+0432 (-18.7% GBE) plus 0.172 solar-mass white dwarf (-0.0012% GBE) form a 2.46-hour binary system. Gravitational radiation orbital decay is GR-predicted -8.6 μsec/yr. DOI:10.1126/science.1233232, arXiv:1304.6875

      …4) Galaxy cluster MACS J1149.5+2223 delayed image gravitational lensing of SN Refsdal is accurately GR-predicted.
      DOI:10.3847/2041-8205/819/1/L8, arXiv:1512.04654

      …5) General Relativity perfectly models black hole mergers: LIGO Event GW150914 35 + 30 solar mass BHs inspiral and merge LIGO Event GW151226[3], 14.2 and 7.5 solar masses, ditto.


      …6) The twin “paradox” is trivially resolved by the triplet Gedankenexperiment in which no clock is ever accelerated. Never.

      • Xinhang Shen

        Please refute my points!

        • Uncle Al

          arXive:1611.06025 “Solar-system tests of the relativistic gravity,” 166 citations.

          arXiv:0911.4527, 1702.08412 Single ion aluminum atomic clocks measure general relativity speeding of time flow in Earth’s gravitational potential from a one foot increase in altitude

          You are overwhelmingly refuted by common observation.

          • Xinhang Shen

            This is not a refutation at all. If y a theory fails only one experiment , no matter how many so called other evidences you have, the theory is wrong.

          • Victor Cole

            I quit earning a living at shopritte and afterwards now I am earning Seventy five $ – Ninety Seven $ per/hr. How? Now I am working via internet! My work did not actually make me satisfied therefore I decided to take an opportunity on something new…after 4 years it wasn’t easy to discontinue my day employment however right now I couldn’t be more satisfied.>>> OKNO.UK/r/279jn

          • Uncle Al

            General Relativity is accurately predictive to the limits of measurement in all venues at all scales. Propose a better theory, git:
            arXiv:1705.01597 EIGHT FAILS
            arXiv:1611:07878 SCORE! Oh, wait – that’s also Einstein.

          • Xinhang Shen

            It doesn’t help if you claim that new experiment supports general relativity because it has failed in logic and basic experiments!

          • Doug Nusbaum

            I love when a person totally ignorant of science makes the above statement. Lets look at the germ theory of disease. According to Mr. Shen it is wrong since.

            1. Cancer is not caused by germs
            2. COPD is not caused by germs
            3. There are few “germs” that cause disease in all people.
            4. Some germs make people healthy, as in infections causing some people with some cancers to have complete remission.
            5. Some experimental results, even when reproduced, are simply matters of error.

            Other than that. great show Mr. Shen

          • Xinhang Shen

            What’s wrong with you? If you don’t agree with me, use your logical refutation please! Otherwise, please shut up!

    • OWilson

      Marry me!

      I’m about to divorce Uncle Al!

    • mbk

      Regarding your thought experiment: When the twins turn around to return to Earth, they change reference frames, so each twin’s view of the other twin’s clock makes a sudden change that compensates for the apparent dilation. You cannot use time dilation by itself.

      • Xinhang Shen

        No, you are wrong. The kinematic time dilation is determined only by the square of the relative velocity between the twin as shown in the Lorentz factor. It doesn’t matter which direction the relative velocity has. According to special relativity, each twin would always see the clock of the other twin ticking more slowly than his own clock as long as there is a relative velocity between them.

        • mbk

          Yes, each twin sees the other’s clock as ticking more slowly than her own clock. You have, however, forgotten that the time seen in another frame also depends on position–“trailing clocks lead”, which does depend on the direction of the relative velocity. When the twins turn around, each sees the other’s clock as suddenly changing to a later time, compensating for the slower ticking. You need to properly use the Lorentz transformations.

          • Xinhang Shen

            No, you are wrong. The sudden turn won’t change the clock time abruptly. Each clock time is continuous. The differential times of the two clocks are related by: dt = dt’/γ and the times then can be calculated through integration.

          • mbk

            Again, you are neglecting part of the Lorentz transformations. The velocity changes abruptly.
            It’s not all about time dilation and length contraction. Each clock, to itself, is continuous, but not to the other.

          • Xinhang Shen

            Of course, both clocks are continuous. If you don’t agree, please show me how you get your conclusion in detail math.

          • mbk

            Okay, the clocks change continuously if you do not want a sudden change in velocity. That does not change the fact that at turnaround each sees the other’s clock change very quickly due to the observer’s acceleration. You could also think of acceleration as being like in a gravity field–each sees the other as being much “higher” in the field and so, during turn around, the “usual” time dilation is not applicable.

          • Xinhang Shen

            Acceleration does not have any effects on clocks (called Clock Hypothesis).

  • StanChaz

    Einstein does not play dice.

  • Todd Chambers

    Will he also measure the frequency of the light of the star before and after?


The Crux

A collection of bright and big ideas about timely and important science from a community of experts.

See More


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar