For the Perseverance rover team at NASA’s Jet Propulsion Laboratory (JPL), the early predawn hours of August 6 were like the night before Christmas. Hours earlier, the scientists had ordered the rover to drill into a rock within Mars’s Jezero Crater to extract and store the mission’s very first sample of Martian geology—one of up to 43 specimens that will one day be delivered to Earth and examined for signs of ancient life.

Around 2 A.M., as they began reviewing fresh data beamed back from Mars by the rover, everything looked to have proceeded perfectly: the sample appeared to have been acquired and placed in one of the rover’s 43 storage tubes. But a nasty surprise soon came courtesy of a probe designed to measure of the volume of that first precious sample: the tube, it seemed, was empty. A subsequent image of the tube confirmed the worst: it was Christmas Day, and Santa forgot to turn up with some geologic gifts. “It was definitely an ‘oh, shit’ moment,” says Ken Farley, Perseverance project scientist at JPL. Confusion peaked when the team commanded the rover to take snapshots of its surrounding—including inside the borehole itself—and still failed to find the missing rock sample. Where had it gone? And if Perseverance’s sample acquisition hardware was fatally flawed, what would happen to the elaborate, multi-billion-dollar plan to retrieve rocks from Mars?

After an anxious, urgent week of intense investigation, the team has concluded that the problem was not the rover, which performed flawlessly. Rather the rock itself had hoodwinked the scientists. “Basically, the rock was pulverized,” Farley says. Despite outwardly looking like a sturdy, solid stone that was ideal for drilling, inside the rock, it was a grainy, crumbly mess. When it was drilled, the sample simply fell apart. The “missing” rock core could be at least partly found as a fine powder in the pile of tailings around the borehole, although it is still not clear where most of the sample ended up.

The good news is that most of Jezero’s rocks will probably not be similarly deceptive, and no hardware mishap has doomed the grand quest to bring pieces of Mars down to Earth. “There is reasonable confidence that this is a one-off, weird rock that we were attracted to in some ways precisely because it was weird,” Farley says. Hopes are now high that the next sample attempt, which could take place within a month, will prove successful.

At first glance, the target rock for Perseverance’s, or “Percy’s,” ground-breaking ceremony looked almost too good to be true. Its face was fantastically flat: an easy exterior on which to inaugurate the drill. An abrasive tool brushed away the dust from a small surface patch, encountering enough resistance as it worked to suggest the rock was relatively rigid. And when team members compared data from instruments investigating the stone’s geochemical composition with photographs taken of the abraded patch’s gleaming crystals, they tentatively identified it as a hardy igneous rock known as gabbro, formed when magma slowly cools deep underground.

The team fully expected that the prize would provide a cohesive sample when drilled. Igneous rocks also contain radioactive compounds that can be used for precise geochronological dating—so this rock, one of the oldest in Jezero, could be used to constrain the timing of the region’s formation billions of years ago.

The team had transmitted the command to drill on the evening of August 5. From that point onward, right up until the sample tube was sealed and stored, the process was autonomous. The scientists could only wait; those whose nerves allowed it snatched a few hours of sleep. “We got up at 2 A.M. [the next morning] to see if the coring had worked, and it had,” says Jennifer Trosper, Perseverance project manager at JPL. Images revealed a picture-perfect drill hole in the rock, and telemetry indicated that a sample tube had been processed and stored inside a chamber on the rover’s underbelly. And a few hours later, images of a sealed sample tube came through. “We thought, ‘Wow, this is fantastic,’” she says—at least, until the probe beamed back more data showing empty space instead of stone within.

The emotional whiplash was severe. “We really thought we had it in the bag,” Farley says. “And we did not.”

In retrospect, it is clearer now that the gabbro was never going to get along with the coring drill. Covered in rusty stains and surrounded by fragments that had flaked away from its surface, the rock appears to have been altered and weakened by soaking in the waters that once flowed through Jezero, back when the crater held river-fed lakes. Remarkably, no one seems to have seen this plot twist coming. Engineers had tested the sampling apparatus more than 100 times back on Earth in a chamber replicating the ambient temperatures and pressures of the Martian surface. A Percy-style coring drill was pitted against a variety of rocks geologists expected to find in Jezero, from loosely packed sedimentary types to tough-as-nails volcanic basalt.

“We have never seen a failure mode where the core vanishes,” Farley says. But in rare instances, “we have seen a failure mode where the core pulverizes and falls back into the hole.” This happened when the drill bored into structurally flimsy sedimentary rock, though—not anything igneous.

It is not impossible that another mischievously mushy rock could defeat Percy’s coring drill in a future sampling attempt. “There are a variety of reasons for the core to collapse into a powder” even if the rock appears rigid, says Manish Patel, a planetary scientist at the Open University, based in England. Any rock can outwardly appear structurally sound only to prove precariously porous or fractured upon drilling.

Not getting a sample of this gabbro was frustrating, but scientists are already extracting valuable lessons from the momentary failure. “We now know more about the physical properties of the rocks on Mars, and this kind of knowledge will be vital for informing the collective future Mars drilling efforts on both Perseverance and beyond,” Patel says.

And at least Percy’s team received a consolation prize. Throughout the mission process, scientists have shown “tremendous interest in acquiring a sample of the atmosphere of Mars,” says María-Paz Zorzano, a researcher at Spain’s Center of Astrobiology and a European return sample adviser to the Perseverance team. Now, thanks to the tube containing little more than a tenuous wisp of Martian air, they have one.

The next sampling attempt will be approached with greater caution, Trosper says. The autonomous coring-and-storing process can be manually interrupted at points—for example, prior to sealing and storage, the team could ask Percy to look in the sample tube to ascertain it contains geologic gold before proceeding.

Percy may be directed back to this spot to try and grab some igneous rocks in the future, Farley says. But for now, the rover is headed to a new site in Jezero: the boundary of the elusive geologic unit where the first sample was attempted, known as the Cratered Floor Fractured Rough, and another layered unit in the crater dubbed Séítah. This second unit’s layers could be comprised of ancient sediments or veneers of volcanic ash. Either way, it is lithologically distinct from that once waterlogged gabbro—and the team members are optimistic that, this time, they will acquire their all-important first sample.

“Game on, Mars. Game on,” Trosper says. “We’ll getcha. We’ll get our core.”