A small copper-alloy object was excavated from a cemetery at Badari in Upper Egypt roughly a century ago. It was catalogued, stored, and largely ignored. Nobody recognized what it was.
Researchers at Newcastle University and the Academy of Fine Arts in Vienna re-examined it and identified it as a bow drill — a tool for mechanically assisted rotary drilling. It dates to the Predynastic period, the late fourth millennium BCE, making it the earliest identified rotary metal drill from ancient Egypt. Previous evidence placed this technology over two thousand years later, in the New Kingdom, when well-preserved drill sets survive in the archaeological record.
The object was available for study the entire time. It sat in a museum collection for approximately one hundred years. The evidence was not missing — it was misclassified. The framework for interpreting small copper-alloy objects from Predynastic burials didn't include “rotary drill bit” as a category, so the object was filed under something else and disappeared into storage.
This is a specific instance of a general problem: collections-based discovery. Natural history museums, archaeological archives, and geological surveys contain millions of specimens classified according to the interpretive frameworks available at the time of collection. When the framework updates — when someone realizes that a certain morphology corresponds to a certain function — objects reclassify without moving. The specimen doesn't change. The category it belongs to does.
The technology was there all along. The two-thousand-year gap in the record was a gap in recognition, not a gap in capability. What looks like an abrupt technological emergence in the New Kingdom is actually a failure of identification in earlier periods. The absence of evidence was literally evidence sitting in a drawer.