Reports of an exomoon’s death may have been greatly exaggerated. In late April, a trio of astronomers suggested that the best-yet candidate for an exomoon—a moon orbiting a world in another planetary system—does not actually exist, being instead a statistical mirage. The candidate’s original discoverers disagree, saying the case is anything but closed. For now, alas, the debate is at an impasse: One way or the other, definitive proof for or against this exomoon may be years away.

In October 2018, astronomers Alex Teachey and David Kipping from Columbia University in New York said they had potentially found the first known exomoon, a Neptune-sized world orbiting a gas giant exoplanet about 8,000 light-years away in the Kepler-1625 system, dubbed Kepler-1625b I. The discovery, using observations from the Hubble Space Telescope gathered in October 2017, would open the door to finding many more such worlds, some of which might even be habitable. “At the time, the best interpretation of the evidence was that a moon is there,” says Teachey.

In February 2019, however, René Heller from the Max Planck Institute for Solar System Research in Germany and his colleagues released their independent analysis of the data. They suggested the potential exomoon signal—an additional dip in the star’s light trailing like a dog on a leash behind the larger dip produced by the gas-giant planet “transiting” across the face of the star—was likely not a moon at all. Instead, they suggested the exomoon-like dip was instead the transit of another shadowy planet in the system. “I’d say this interpretation sounds more likely than an exomoon interpretation,” says Heller.

Now a new twist has emerged, via another independent analysis by Laura Kreidberg from the Harvard-Smithsonian Center for Astrophysics in Massachusetts and colleagues. In a paper posted on the pre-print server arXiv, they re-examine the Hubble data used by Teachey and Kipping, finding no evidence for a second dip in the star’s light at all. “I’m not convinced that this particular exomoon is real,” says Kreidberg. She thinks Teachey and Kipping were mistaken in their original analysis, and that efforts would be better placed looking for exomoon candidates around other stars. The discrepancy arises from the fact that Hubble observations struggle to reach the necessary levels of precision required to distinguish between random fluctuations (in starlight or in instrumental performance) and the faint-but-genuine signal that would be associated with even the largest conceivable exomoons. Discerning the difference is a matter of guesswork using sophisticated, computationally intensive numerical models.

Even so, all three research papers confirmed a transit timing variation (TTV) for the planet—a finding that arguably favors the exomoon hypothesis. In other words, the planet’s transit began slightly earlier than would be expected based on its known orbit. This suggests it was experiencing a gravitational tug from another body, perhaps indeed an exomoon or another planet. Furthermore, Heller’s conclusions overlap with Teachey and Kipping’s in that Heller also sees evidence for a genuine dip in the star’s light; of the three papers examining the data, only Kreidberg’s deems the dip to be illusory.

Unfortunately, attempts to use Hubble again to observe this system have been in vain; Teachey and Kipping’s request for further time on the telescope to study the next predicted transit of Kepler-1625b and its potential exomoon in May 2019 was recently denied. “There’s always stiff competition for observing time on Hubble,” says Neill Reid, the associate director for science at the Baltimore-based Space Telescope Science Institute, the organization responsible for allocating time on Hubble. “There’s always twice as many proposals that [we] would like to see get observations than we have orbits to support.” The failure to secure additional Hubble time is problematic, Kipping notes, because subsequent transits of the potential exomoon become much more difficult to predict the further off they occur in the future.

All hope is not lost, though, because it remains possible to dismiss the exomoon hypothesis using a method called radial velocity. This involves precisely measuring the position of the star, and looking for noticeable wobbles in its location owing to the gravitational tugging from other bodies. If, as Heller maintains, an as-yet-unconfirmed planet is responsible for the exomoon-like dip in the starlight from Kepler 1625, radial velocity measurements should eventually detect that planet’s wobbly influence upon the star. “I think we will definitely know if it’s there or not,” says Teachey, although he notes it could be 10 to 15 years before we have an answer.

The whole saga highlights the intense difficulty in finding these worlds, as their comparatively small size compared to their parent star and planet makes them difficult to spot. Uri Malamud from the Technion—Israel Institute of Technology in Haifa, Israel, notes that our current technology simply is not good enough to find most of them yet. But the future remains bright for exomoon hunting, even if the prospects for Kepler-1625b I have dimmed. Soon, a new generation of enormous observatories such as the Giant Magellan Telescope and the James Webb Space Telescope will begin their stargazing, bringing unprecedented levels of sensitivity and precision to the ongoing search. “Ten years from now we might have data that makes all of this stuff obsolete,” says Teachey. “But you don’t wait for those giant telescopes. We work with what we’ve got now and see what we can do.”