On August 12, 2026, a total solar eclipse crossed northern Spain, the first to darken the peninsula in 121 years 1. Millions gathered along a path stretching from A Coruña to the Balearic Islands, turning a celestial mechanics lesson into a nationwide spectacle 2. But the event was more than a photo opportunity. It was a reminder that astronomy, at its core, is the discipline of measuring shadows—and that the most interesting findings are often the ones that leave us staring at the dark.
Consider the Hubble Space Telescope data published this week in Nature Astronomy. Led by Davide Massari of the Observatory of Astrophysics, the team identified one of the Milky Way’s earliest known mergers: a collision with a dwarf galaxy roughly 11.8 billion years ago, just 2 billion years after the Big Bang 35. The evidence rests on stellar kinematics—the precise motions of ancient stars that still carry the gravitational fingerprints of their birth galaxy. It is a statistical inference, not a direct observation; no telescope can watch a merger happen. But the numbers are consistent, and the interpretation is robust: our galaxy grew not only by forming stars but by consuming smaller ones 5.
The institutional context matters here. Hubble, now in its fourth decade, remains the workhorse for this kind of archaeology. It is not the newest instrument, nor the sharpest, but its long baseline of observations allows astronomers to measure proper motions with exquisite precision. That longevity is a feature, not a bug. The same cannot be said for the International Space Station, whose aging hardware prompted a spacewalk on August 18. Sophie Adenot became the first French woman to walk in space, spending six and a half hours outside the station with NASA astronaut Anil Menon to replace the Space-to-Ground antenna, a critical link for high-speed data transmission 4. The repair was necessary, unglamorous, and entirely routine—except for the history it carried.
Skepticism is built into these measurements, but it takes different forms. For the eclipse, the uncertainty was about weather and crowds, not physics. For the galactic merger, it is about whether the dwarf galaxy’s identity can be confirmed with future data from the James Webb Space Telescope or the Vera Rubin Observatory. For the Mallorca plague of 1820, published in PNAS, the uncertainty is more unsettling. Researchers used daily mortality records from Son Servera and Capdepera to model the outbreak, concluding that the disease was not solely bubonic; pneumonic and septicemic forms likely played a substantial role 6. The overall case fatality rate was 78 percent 6. That figure is a statistical inference drawn from parish registers—not a clinical diagnosis, which is impossible two centuries later. The study does not settle the debate; it reframes it.
What remains unknowable is the human texture of these events. We can model transmission dynamics but not the fear in a Mallorcan village. We can date a galactic collision but not the fate of every star it scattered. We can predict an eclipse to the second but not the expression on a child’s face when the sky goes dark.
The tradeoff that matters most: every measurement we make narrows one uncertainty while opening another. The eclipse will not return to Spain for over a century. The Milky Way’s merger happened once, and we are still reading its aftermath. The plague’s true toll will never be fully known. That is not a failure of science. It is the condition of doing it.