Research Reveals Spinosaurus as Predominantly Aquatic Predator

New analysis of extinct vertebrate anatomy suggests the notorious dinosaur Spinosaurus spent most of its time submerged rather than hunting from shorelines, settling a long-standing paleontological debate. Researchers compared skeletal measurements and CT scans of extinct species with modern animals to distinguish between aquatic and semi-aquatic lifestyles. The study employed advanced techniques including machine-learning models to reconstruct how this unusual predator from 113 to 94 million years ago adapted to its watery African environment.
The research team examined over 11,000 anatomical measurements alongside CT imaging and fossil photographs to develop their conclusions. By training machine-learning algorithms on modern vertebrates with well-documented behaviors, scientists could then apply those patterns to extinct species, revealing lifestyle adaptations that single skeletal features alone might not clarify. This methodology proved particularly valuable for Spinosaurus, whose crocodile-like snout and sail-backed spine had sparked competing theories about whether it hunted as an aquatic diver or a shoreline forager during the Early Cretaceous period in Africa.
The study also demonstrates how similar reversals from terrestrial to aquatic existence played out differently across species. Mesosaurs, despite bearing live young—a trait sometimes interpreted as evidence of full water adaptation—showed limb proportions suggesting they remained partially land-dependent, resembling modern alligators in their behavioral ecology rather than fully aquatic seals.
This research may refine how paleontologists interpret fragmentary fossil evidence and resolve longstanding anatomical debates. The machine-learning methodology could enhance future species classification efforts, potentially shifting educational content about prehistoric ecosystems and predator behavior. Museums and textbooks may need updating, while the approach itself could influence how researchers approach other ambiguous cases in the fossil record. Such methodological advances in paleontology could also strengthen interdisciplinary collaboration between computer science and natural history fields.