Here's one I find genuinely startling: the most abundant photosynthetic organism on Earth wasn't discovered until 1986.
It's called Prochlorococcus, a marine bacterium about 0.6 micrometers across — small enough that it slipped through the filters and nets oceanographers had been using for a century. There are estimated to be something like 10²⁷ of them, and they're thought to produce a meaningful share of the oxygen in the atmosphere. Sallie Chisholm and colleagues found them essentially by pointing a newly available tool (flow cytometry) at seawater and noticing a signal nobody could account for.
What I like about this isn't just the trivia — it's what it implies. We had detailed maps of the ocean floor, we'd been to the Moon, and we'd somehow missed the single most numerous photosynthesizer on the planet, hiding in plain sight in every bucket of seawater. It's a decent reminder that gaps in knowledge often aren't in exotic places; they're in the places our instruments happen to be blind to.
Happy to aim somewhere else if you have a preferred domain — math, history, language, something weirder.
Here's one I find genuinely delightful: you can read the length of an ancient day in fossil coral.
Corals lay down thin growth increments daily, bundled into annual bands — a bit like tree rings, but at two scales. In 1963 a paleontologist named John Wells realized this meant fossil corals were clocks. He counted the fine daily ridges between annual bands in Devonian corals, roughly 380 million years old, and got about 400 increments per year instead of 365.
That's exactly what physics predicted. Tidal friction — the Moon's pull dragging water and rock across a spinning Earth — bleeds rotational energy away, so days get longer while the Moon drifts outward to conserve angular momentum. A Devonian day was around 22 hours. The year's length barely changed, so you just fit more of them in.
Two things make this especially satisfying. First, the theory came first: Charles Darwin's son George Darwin worked out the tidal-braking math in the 1870s, decades before anyone could check it. Second, we now measure the other end of the same process directly — Apollo astronauts left laser retroreflectors on the Moon, and bouncing lasers off them shows it receding about 3.8 centimeters per year. A fossil in a drawer and a mirror on the Moon are telling you the same story from opposite ends of 400 million years.
There's a related trick with "tidal rhythmites" — layered sediments that record individual tides — which lets you extract not just day length but the ancient lunar month.
Want more in this vein (physical traces of deep time), or would you rather I aim somewhere else entirely — history where the popular story turns out to be wrong, weird math, linguistics?
Trees aren't a thing.
Not in the biological sense, anyway. "Tree" isn't a group of related organisms like "bird" or "beetle" — it's a growth strategy that plants have independently invented dozens of separate times. Wood (lignified secondary xylem) has evolved and then been abandoned again and again across the plant family tree.
The consequence is that a palm is more closely related to grass than to an oak. A tree fern and a pine share less recent ancestry with each other than you share with a goldfish.
The strangest part is how easily it happens. On islands, ordinary herbs repeatedly bulk up into trees — the Galápagos has "tree" daisies in the sunflower family, the Canary Islands have tree-sized relatives of forget-me-nots. Botanists call it insular woodiness. Give a weed enough time, a mild climate, and no grazers, and it will start reaching for the canopy.
So "tree" is more like "swimmer" than like "fish."