When paleontologists find a structure, they ask what it was for. Horns for combat, feathers for thermoregulation (later flight), plates for display (or heat exchange, or defense — the stegosaur debate continues). The question assumes a function. The debate is over which one.
Huang and colleagues (Nature Ecology & Evolution, February 2026) described Haolong dongi, a 125-million-year-old juvenile iguanodontian from China covered in hollow cutaneous spikes — structures never before documented in any dinosaur. X-ray scans and histological sections revealed the architecture: a highly cornified outer layer over a multi-layered epidermis, with keratinocytes preserved down to their nuclei. The spikes are hollow and cylindrical. The researchers proposed three possible functions: defense (like porcupine quills), thermoregulation (like an elephant's ears), or sensory perception (like whiskers). Three hypotheses, one structure.
But hollowness is not a functional feature. It is a structural vacancy. A solid spike can do one thing well — resist compression, puncture a predator, support weight. A hollow spike can do many things adequately. Fill the interior with air and it insulates. Thread nerves through it and it senses. Leave it empty and it reduces mass. The hollow interior is an open parameter — a degree of freedom that different selection pressures can fill with different contents without modifying the external architecture.
The question “what were the spikes for?” contains an assumption that hollowness specifically undermines. Solid structures resist repurposing because their material is committed. Hollow structures invite it because their interior is uncommitted. A feather shaft is hollow. A bone is hollow. A hair follicle is hollow. In each case, the hollow interior was eventually co-opted for functions the original structure didn't serve — pneumaticity for weight reduction in dinosaur bones, vascular supply for antler growth, air trapping for insulation in mammalian fur. The hollowness didn't evolve for the later function. It preceded it. The vacancy was there first, and selection filled it with whatever the organism needed most.
The debate over Haolong dongi's spikes will likely continue as more specimens are found, especially adults (this fossil is a juvenile — the spikes may have changed or disappeared with maturity). But the debate's structure — defense or thermoregulation or sensory? — may be the wrong frame. The hollow architecture suggests all three, or none yet, or different ones at different life stages. The open parameter doesn't resolve to a single function. It holds space for functions that haven't been selected yet, or that shift with context, or that operate simultaneously without conflicting because the interior can serve multiple roles at once.
The general pattern: when a structure contains an uncommitted parameter, asking what it is for assumes an answer the architecture doesn't require. Hollow structures are pre-adapted for versatility because they carry an empty interior that selection can furnish differently in different contexts. The most interesting feature of the spike is not its exterior — which punctures, deflects, or signals — but its interior, which does nothing yet. The vacancy is the adaptation. The function will come later, and it will come more than once.