Letter lost for 75 years leads to the accidental discovery of one of Earth's rarest minerals

A long-lost letter, tucked away for 75 years, has led scientists in Germany to uncover one of the rarest minerals on Earth—humboldtine.

While digitizing its extensive collection of rocks and minerals, the Bavarian Environment Agency (LfU) stumbled upon a forgotten envelope in a basement drawer. Inside was a handwritten letter from 1949, sent by a coal mine operator who had discovered an unusual yellow mineral in the Mathias coal mine near Schwandorf, Bavaria.

Curious, researchers at the LfU revisited samples collected from that mine decades ago. Among them, they found a hazelnut-sized piece of a yellowish mineral labeled “humboldtine.” Using modern techniques like X-ray diffraction and chemical analysis, they confirmed it was indeed the elusive substance mentioned in the letter.

Humboldtine is an iron oxalate mineral (FeC₂O₄·2H₂O) that typically forms under rare conditions involving organic carbon, iron, and water—usually in coal beds. First discovered in 1821 and named after famed explorer Alexander von Humboldt, humboldtine has only been documented in a few dozen locations around the world. It’s normally found in tiny, barely visible crystal formations—making the hazelnut-sized piece discovered in Bavaria exceptionally rare.

According to the researchers, this one sample may have doubled the total amount of humboldtine known to exist on Earth.

Unfortunately, the Mathias mine where it was originally found has been closed and flooded since the 1960s, so scientists may never fully understand how the conditions came together to create this mineral.

Still, the discovery is a major win for science—and a testament to the value of archival research. The rediscovered humboldtine specimen will go on public display at Munich’s Mineralientage show this October, offering a rare glimpse at a mineral few people have ever seen.

If extremely rare minerals like humboldtine can go unnoticed for decades, what else might we be overlooking in existing collections or data?