Before 2026, nearly all spinosaurid fossils — bones, teeth, fragments — came from coastal deposits. The interpretation followed the evidence: spinosaurids were coastal predators, semi-aquatic hunters adapted to marine or near-marine environments. The anatomy supported it — elongated snouts, conical teeth, sometimes interpreted as fish-eating adaptations. The ecology was reconstructed from the depositional environment, and the depositional environment was coastal. The inference seemed secure.
Sereno and colleagues (Science, February 2026) describe Spinosaurus mirabilis, a new species from the Jenguebi site in Niger's central Sahara, found in river sediments 500 to 1,000 kilometers from the nearest marine shoreline. The animal has a scimitar-shaped cranial crest — blade-like, likely keratinized and brightly colored — so large and unexpected that the researchers initially failed to recognize it. Its skeletal proportions suggest a “hell heron”: sturdy legs for wading in shallow water, comfortable in two meters of depth but spending most of its time in shallower traps along forested river systems. Alongside S. mirabilis, the site yielded intact partial skeletons of long-necked dinosaurs, placing the predator firmly in an inland ecosystem.
The finding doesn't overthrow the aquatic interpretation. Spinosaurids clearly interacted with water. What it overturns is the assumption that the fossil record's geographic distribution reflected the animal's actual range. Coastal deposits preserve bones well — water, sediment, rapid burial, mineral replacement. Inland environments preserve bones poorly — exposed surfaces, biological recycling, erosion. An animal that lived across both environments would leave disproportionately more fossils in the one that preserves better. The fossil record selected for the depositional environment, not the behavior.
This is taphonomic bias operating at the level of ecological inference. The standard correction for it is well known in paleontology: you can't treat absence of evidence as evidence of absence. But the spinosaurid case is sharper than that general warning. Here, the presence of evidence in coastal deposits was treated as positive evidence for coastal specialization. The bias wasn't in what was missing — it was in what was found. The existing record wasn't incomplete; it was systematically misleading. Every coastal fossil made the coastal interpretation look stronger, and each one was equally compatible with an animal that happened to die near the coast.
The through-claim is about how the evidence selects itself. The deposit doesn't just preserve the specimen; it shapes the story. An organism buried in a biased archive doesn't just leave gaps — it leaves a coherent, internally consistent, and wrong narrative. The correction comes not from better theory but from a different archive: an inland site with different preservation conditions, producing a different slice of the same animal's life.