A web exclusive story from Astronomy magazine


Lead illustration and drawings by Kellie Jaeger of Astronomy
The wild and woolly menagerie of ghost worlds that once haunted the halls of astronomical history.

by John Wenz

Our galaxy is lousy with planets. NASA’s exoplanet archive lists more than 3,400 confirmed planets outside our solar system, with more added every day. Nearly 4,500 planetary candidates from NASA’s Kepler spacecraft await confirmation, meaning thousands more could make the list in the near future.

Before 1995, there were only 11 known planets. There were the nine classical planets in our solar system, from Mercury through Pluto (then still a planet), and two bizarre objects that had been found around a distant pulsar.

But in 1995, we discovered 51 Pegasi b, the first confirmed planet around a Sun-like star. It was a weird one. It circled its star every four days and was eight times more massive than Jupiter. At the time, it received a lukewarm reception.

“In 1995, 51 Pegasi b was found, and about 50 percent of the astronomers or fewer believed it was an exoplanet,” says Debra Fischer, a professor of astronomy at Yale University.

Why did it take so long to find a world circling another star in the first place?

“You have to picture the size and the mass of the star and then a little planet like the Earth. And if you set them side by side, Earth has a diameter that’s 1/100th the size of the Sun and [a fraction of] the mass,” Fischer says. Thus, astronomers can’t directly see exoplanets, as even at “the edge of [their] solar system, [they] are drowned out by the light of the star.” To find planets, astronomers must employ other methods, many of which have only just come into their own.

With that in mind, it was a long road to get to 51 Pegasi b. The right technology had to reach maturity at the right time. But it wasn’t the first planetary claim. In fact, the history of claims of planets outside our solar system stretches back more than a century and involves outsized personalities, outright fraudsters, befuddled scientists, brown dwarfs, numerous retractions, bitter back-and-forths, and more.

Here are the tales of the planets that never were.

The grand fraud

Thomas Jefferson Jackson See (1866–1962) was a lot of things. He was a brilliant astronomer — at least, if you asked him. But his peers saw him as something else: arrogant and prone to plagiarism. Thomas J. Sherrill, a Lockheed engineer and astrophysicist, wrote in a 1999 paper published in the Journal for the History of Astronomy that few scientists of the early 20th century “inspire a degree of rancour comparable to that evoked” by See.

Thomas Jefferson Jackson See

Sherrill states that although See had a “solid background in celestial mechanics,” his latter work “[diverged] from his astronomical colleagues in striking ways.”

An 1895 paper published in The Astronomical Journal was especially audacious. “Since August 20, when I first announced to you the existence of peculiar anomalies in the motion of the companion of F.70 Ophiuchi, I have succeeded in showing conclusively that the system is perturbed by an unseen body,” See said, his arrogance readily apparent.

When an 1899 paper by Forest Moulton challenged the findings — claiming that the three-body problem proposed by See would fling a planet out of the binary star system — See wrote a letter to the journal so vitriolic that most of it was redacted, and he was nearly banned from further publication. (The editors instead said See would be heavily censored in future communications.)

See attempted a second career as a geologist before publishing books about the formation of the solar system, which Sherrill says had a few correct assertions, but “many more were speculations presented with little justification, and others were borrowed from his contemporaries.”

See’s eventual undoing was a supposed biography written by a journalist who turned out to be See himself, playing up his own brilliance. By the time he had taken to writing bitter letters (a few of which appeared in The New York Times) about Albert Einstein’s theories, few, if any, members of the science community heard him out.

“See was indeed a most colorful person, and probably quite brilliant, but he seemed to be extremely paranoid and double-dealing,” says David DeVorkin, senior curator of astronomy at the Smithsonian National Air and Space Museum.

See was also not the first person to propose a planet around 70 Ophiuchi. William Stephen Jacob of the Madras Observatory put forth the idea in 1855. After See, astronomers Dirk Reuyl and Erik Holberg brought it back in 1943. A. Vibert Douglas wrote in a 1955 article in the Journal of the Royal Astronomical Society of Canada that the planet is “more remote from its star than Jupiter from the sun. Jupiter is believed to be wholly incased in ice, so that the likelihood of life on 70 Ophiuchi C is negligible.”

But to date, no planet has been confirmed. So why has this binary star system so entranced astronomers?

DeVorkin says it’s because 70 Ophiuchi is a “close-by low-mass system that’s relatively easy to observe.

“Old ideas die hard,” he adds.

The man of wonder

See had no standing left in the astronomical community by 1905. He was eventually, as DeVorkin says, “banished to Mare Island,” a small observatory in San Francisco.

Astronomer Peter van de Kamp (1901–1995) was nothing like See. While See was arrogant and unscrupulous, van de Kamp was gregarious and popular. While See failed to make lasting contributions to the field of astrophysics, van de Kamp wrote the book on 20th Century astrometry.

But both had one thing in common: planets that vanished upon further scrutiny.

Peter van de Kamp

Van de Kamp was a popular professor known for dynamic lectures, a love of classical music, and his amiable demeanor. As the director of Sproul Observatory at Swarthmore College, van de Kamp became a trusted adviser to several students, teaching them astrometry, a technique that measures the precise position of stars.

“Peter van de Kamp was one of the first to push this work down to much cooler and less massive stars,” says Eric Jensen, a professor of astronomy at Swarthmore. “So the work that he and others did at Sproul Observatory over many years, measuring orbits for systems with low-mass stars, was fundamental for our understanding of cool, red stars, which we now know to be by far the most common kind of stars.”

This work led to 61 Cygni. Van de Kamp and his graduate student Kaj Strand first proposed a planet in this binary system in 1942, based on their astrometry measurements. 61 Cygni A and B seemed to wobble slightly as they orbited each other, as if tugged by an unseen object. Van de Kamp added to his planetary claims in 1951, this time announcing a proposed planet around Lalande 21185 with graduate student Sarah Lee Lippincott.

Sarah Lee Lippincott

But 1963 saw van de Kamp’s most explosive announcement: There was a planet around Barnard’s Star, an M dwarf just 6 light-years away. It was a cold, inhospitable planet larger than Jupiter, in a 12-year orbit. Later, van de Kamp added a Saturn-sized world in a 20-year orbit, based on further assessment of the star’s motions.

“His work was focused on measuring the orbits of stars,” Jensen says. “But once you measure an orbit, you get masses for the orbiting object. And he thought he had found, from the orbit he measured for Barnard’s Star, that it implied an orbiting companion of planetary mass.”

A 1966 paper on interstellar travel focusing on Barnard’s Star kicked off what became Project Daedalus, one of the first modern engineering studies into interstellar travel within a human lifetime. A robotic probe would use nuclear explosions to propel a spacecraft toward Barnard’s Star at 12 percent the speed of light.

Project Daedalus

But by 1973, a new view on these planets was emerging: Sproul Observatory’s telescope was flawed, and none of the planets actually existed. Specifically, the photographic plates were underexposed, and when the telescope was calibrated a certain way, some stars appeared to move artificially.

“Astrometry, it turns out, is just tough to do from the ground because we’re looking through Earth’s atmosphere and the stars are twinkling,” Fischer says.

One of the flaw’s discoverers was the new Sproul Observatory head, Wulff-Dieter Heintz. Another observatory employee, John L. Hershey, published a paper in The Astronomical Journal in June 1973. Hershey had studied the tiny M-dwarf star Gliese 793, and as he pored through photographic plates, he noticed something: It had the exact same data discontinuity as Barnard’s Star.

But the United States Naval Observatory and Allegheny Observatory at the University of Pittsburgh found no shift in the motion of Barnard’s Star.

“When his work was checked over by other observatories using the more powerful plate constant method, his unseen planetary companions to Barnard’s Star vanished,” DeVorkin says.

That didn’t deter van de Kamp, but it did damage his friendships with Heintz and George Gatewood, the astronomer at Allegheny who confirmed the lack of perturbations in the motion of Barnard’s Star. As late as the 160th Meeting of the American Astronomical Society, held in Troy, New York, in 1982, van de Kamp was still pushing for his planets.

“Current analysis for one orbit clearly yields a perturbation with a period of 12 years,” his abstract read. “As before, a desirable improvement is made by an additional perturbation with a period of 20 years.”

Incidentally, at this same conference, van de Kamp’s protégé Lippincott presented the possibility of planetary-mass companions around three other stars. One of those stars, Luyten’s Star, has been recently confirmed to host planets. However, Lippincott suggested two long-period gas giants, whereas the planetary companions found in March 2017 were a super-Earth and an Earth-mass planet. These planets don’t account for the magnitude of Lippincott’s possible detection. Another 1982 paper by Lippincott suggested — but far from asserted — that a planet could cause flare activity on EV Lacertae.

“In general, people have been skeptical about planetary discoveries, often for good reason — [van de Kamp’s] is far from the only claim of a planetary discovery that turned out to be incorrect,” Jensen says. “And even the initial discoveries of exoplanets in the mid-1990s were met with some skepticism initially, though of course now we have overwhelming evidence of the rich diversity of planets around other stars.”

Luyten’s Star is the only star in van de Kamp’s cadre with a confirmed planetary system. To date, no planets have been found around Barnard’s Star, Lalande 21185, or 61 Cygni.

Gatewood did present a paper in 1996 claiming that he had found several planetary-mass companions around Lalande 21185. This result, too, was based on astrometry. Gatewood’s planets also were cast in doubt and remain unconfirmed. A February 2017 paper suggested evidence of a 3.8-Earth-mass planet, though it also remains ambiguous.

Lalande 21185 just can’t catch a break.

Still, van de Kamp’s work left him “highly respected” at the end of his career, according to DeVorkin. And Sproul Observatory has since been replaced at Swarthmore with a new one: Peter van de Kamp Observatory. As part of the Kilodegree Extremely Little Telescope program, the observatory has turned up planets — this time, real ones.

“Our collaboration has published discoveries of about 15 exoplanets so far, and many of the discovery papers include data from our telescope here,” Jensen says. “Although van de Kamp was wrong about Barnard’s Star, he made important contributions to astronomy, and I’m pleased that we can honor and remember his work by having his name on our observatory.”

Van Biesbroeck 8B

When astronomers discovered an object around van Biesbroeck 8 (vB8) in 1984, they weren’t quite sure what they were seeing. The New York Times ran the headline “Possible Planet Found Outside the Solar System,” with author John Noble Wilford stating, “If this is indeed a planet, the discovery would be a clear breakthrough in the long search for extrasolar planetary systems and the first direct evidence to support a premise underlying theories of possible extraterrestrial life, which is that planetary systems are not unique to the Sun and may even be common in the universe.”

But there was a problem with the puff of gas that had been spotted in the infrared: Although it was the size of Jupiter, further studies showed an outsized influence on its parent star. In fact, this is a common problem in the hunt for planets. There are objects that are neither stars nor planets, called brown dwarfs, which can be easily mistaken for planetary companions. Mass estimates on vB8’s companion were hard to nail down — somewhere between 30 and 80 Jupiter masses. To several astronomers, this indicated a brown dwarf.

Jill Tarter of the SETI Institute first gave brown dwarfs their name. Brown dwarfs accumulate matter in the same way as stars, but fail to attain enough mass to ignite hydrogen fusion. Stars fuse hydrogen into helium, which can only occur above a certain temperature and pressure.

Objects above that threshold are stars. Brown dwarfs initially produce heat by fusing an isotope of hydrogen called deuterium into helium-3, which can occur at lower temperatures and thus, lower masses.

TRAPPIST-1, an ultra-cool dwarf star, is 84 Jupiter masses. The largest known brown dwarf is 90 Jupiter masses. In other words, the realm of large brown dwarfs and small stars is a bit murky.

Astronomers discovered the first brown dwarf in 1995, the same year they confirmed the first planet around a Sun-like star. So, what happened to vB8’s companion 11 years earlier?

The vB8 discovery sent shock waves through the astronomical community; a conference was convened in 1986 on the topic of brown dwarfs. According to New Light on Dark Stars by I. Neill Reid and Suzanne L. Hawley, that very conference torpedoed the case for vB8’s companion, as other infrared observations failed to find it. “The only possible conclusion is that the original detection was an observational artefact, probably due to the chromatic effects of atmospheric refraction,” they write.

The first planet — or possibly the first brown dwarf — evaporated almost as quickly as it had emerged.

Brown dwarfs are just one of many red herrings astronomers face when searching for planets. The mass dividing line between the smallest brown dwarfs and the largest planets remains murky. NASA/JPL-Caltec

Pulsar planets

In 1992, astronomers officially found the first planetary system when they discovered two (later found to be three) objects around PSR B1257+12, a pulsar 2,300 light-years away. Pulsars — a type of rapidly rotating neutron star — are often too small to be seen in optical light, so the planets’ presence was inferred from subtle changes in the normally precise radio signals coming from the pulsar.

Pulsar planets are weird — neutron stars are formed in supernovae or stellar mergers, violent events that tend to consume or sweep away materials from any planet in the vicinity, destroying them. Thus, the scant few pulsar planets discovered to date are likely to have formed after the supernova in a second planetary genesis.

A year earlier, in 1991, there were claims of a planet around PSR B1829−10, a different pulsar 30,000 light-years away.

The progression of papers in the journal Nature tells it all. In August 1991, a trio of astronomers published their results: “A planet orbiting the neutron star PSR 1829–10.” But a January 1992 paper by two of the three original authors was titled “No planet orbiting PSR 1829–10.” The planet detection had been a simple miscalculation created by failing to model Earth’s orbit accurately when assessing measurements of the star.

The 1991 claim wasn’t even the first possible pulsar planet that never was. In a November 1979 Nature letter, Mieczysław Prószyński and Marek Demiański of Warsaw University found timing variations in pulsar PSR 0329+54 suggesting something was affecting the dense star husk. They proposed a change in shape, a change in magnetic fields, or a planet half the mass of Earth (or less) orbiting it. A 1995 article by Tatiana V. Shabanova in The Astrophysical Journal tried to bolster the case, but subsequent investigations found the signal variations are likely a consequence of the star’s variable rotation rate.

There is also an October 22, 1969, article in The New York Times asserting a planet orbiting the Crab pulsar, based on a “wobble” observed in the pulsar that could be like the wobble that occurs in the Sun as the planets orbit. The results were published in Nature in 1970, but subsequent papers on the wobble observed in the Crab and Vela pulsars suggested that it was instead caused by stresses and friction in the star’s crust, leading to starquakes.

Gamma Cephei

Some planet stories have a happy ending.

Gamma (γ) Cephei A was initially thought to have a planetary-mass companion in 1988, though the evidence was only tentative. At the time, the emerging technology of radial velocity was just barely refined enough to detect planets.

The 1988 data did confirm one thing: Gamma Cephei was a binary star system with a low-mass red dwarf, Gamma Cephei B, circling Gamma Cephei A. Even after accounting for the companion’s effects, Stephenson Yang and his associates still saw evidence in their signal for a low-mass, possibly planetary object in the system.

Stephenson Yang

But a 1992 follow-up study cast doubt on the tentative planet. Gamma Cephei B dominated the radial velocity measurements, making the supposed planet’s signal scientifically unreliable. “We had a very weak signal, one and a half sigma,” says Yang, a professor at the University of Victoria and a principal investigator.

Also, Gamma Cephei A was believed to be much younger than it actually is. “We thought we were looking at a much more variable star than one that was main sequence,” Yang says. They chalked up their weak signal to variations in the star and retracted their planetary claim.

Fast-forward to 2003. A planet around Gamma Cephei A was announced with roughly the mass and orbit suggested by the 1988 results. Yang and his compatriots had, in fact, found the first exoplanet. They just hadn’t been able to confirm it.

In 1992, Gamma Cephei B wasn’t well constrained. By gathering more data, it was possible to extract the signature of the smaller star from the wobbles of Gamma Cephei A. And from that information, a Jupiter-sized world with a 2.5-Earth-year orbit emerged.

“When you look at the radial velocity of the star, you do see a large radial velocity change because it has a companion,” Yang says. “If you took out the changes, you see ripples, which [are] caused by the planet.”

And thus, a planet called Gamma Cephei Ab was found around a future North Star, based on the small changes it makes to its star’s orbit, which had been drowned out by the tug of a much larger star.

“I still remember the first time we looked at the ripple and said, ‘Oh wow, it fits so nicely to the orbit,’ ” Yang says.

So 51 Pegasi b got the glory. But Gamma Cephei Ab was there first. (Fischer also points out HD 114762b, which was discovered in 1989 and confirmed in 2012. Technically it could be considered the second, bumping 51 Pegasi b to third.)

A nearly 150-year hunt had drawn to a close. Now armed with a few confirmed planets, astronomers could begin to build a real catalog of stars with planetary systems. All it took was a handful of phantom planets to get there.

Modern Ghosts

This artist’s impression shows three planets in the Gliese 581 system: b, c, and d. The status of other proposed planets in the system remains unconfirmed. ESO

The early years of planetary detection sparked debate as discoveries accumulated. Of the first 19 planets discovered, half were actually brown dwarfs. The problem persists today, as more accurate mass measurements of hefty planets reveal instead brown dwarfs.

In 2009, the Kepler telescope inundated astronomers with new data. Previous false detections were relabeled candidates, while some candidates became false detections. But only one confirmed Kepler planet has ever been retracted: Kepler-32e, the result of a clerical error.

The Gliese 581 system has seen quite the volley back and forth. In 2010, astronomers found two or three potentially habitable planets: Gliese 581b, c, and e. But the proposed Gliese 581f was soon ruled out, while the status of Gliese 581g, the most potentially habitable planet, has long been cast in doubt.

Alpha Centauri Bb, discovered in 2012, was such a weak detection that it became a high-profile retraction. Many of the planet’s discoverers joined the Pale Red Dot team, which successfully found the Earth-mass planet Proxima Centauri b in the same system in 2016. — J. W.

John Wenz is a former associate editor of Astronomy magazine.