Nine Bengal slow lorises, rescued from the illegal pet trade, were fitted with radio collars and released into a national park in northeastern Bangladesh. Within ten days, three were dead. Within six months, four more had died. Of the seven bodies recovered, four showed bite wounds on the head, face, and digits — the signature of territorial attack by other slow lorises. The animals released to save them were killed by their own species.
Slow lorises are the only venomous primates. They deliver toxic bites through specialized teeth, and they are intensely territorial. The national park was not empty habitat waiting to be filled. It was occupied territory, defended by resident lorises that responded to the newcomers as intruders. The two survivors traveled farther than those that died, crossing through established territories fast enough to avoid lethal confrontation.
The study (Global Ecology and Conservation, February 2026) tracked all nine animals and documented a 78% mortality rate, with conspecific aggression as the primary cause. The conservation logic — rescue from captivity, rehabilitate, release into protected habitat — assumed the bottleneck was captivity. The actual bottleneck was territory. The habitat was protected. It was also full.
The structural observation: the intervention created the threat it was designed to prevent. The lorises were rescued because they were endangered. They were released into habitat selected because it already supported a healthy population. But a healthy population means occupied territories, and occupied territories mean that any newcomer faces the resident animals' primary defense mechanism. The same ecological success that made the release site attractive made it lethal for the released animals. The refuge was the danger.
This is not a failure of execution but a failure of model. The conservation framework treats habitat as the limiting resource — find suitable habitat, place the animal in it, let nature work. But for territorial species, habitat is not a resource that can be saturated by adding individuals. It is a partitioned space where each unit is claimed and defended. Adding an animal to occupied territory is not supplementation. It is invasion, and the resident population responds accordingly. The model that says “more protected habitat means more capacity for animals” is correct for species limited by food or shelter. For species limited by territory, more habitat does not help unless it is unoccupied habitat — and unoccupied habitat in a national park hosting a healthy population is a contradiction in terms.