Snare
Most spider traps wait. A web goes up, the spider goes still, and whatever blunders in is what it catches — the spider is always the one who decides when to strike. In June 2026, Ajay Narendra's group at Macquarie University described one that doesn't work that way: a small, undescribed spider in the genus Propostira (family Theridiidae, the same family as black widows), found over ten nights of fieldwork in rainforest near Cooktown in Far North Queensland, Australia. They nicknamed it the ballista spider, after the Roman siege engine, because the trap it builds is a spring, not a net — and the spring goes off when the prey pulls its own trigger, not when the spider does. Narendra called it "the only case where a spider's web is designed to catch a single prey species, and where the mechanism is triggered by the prey rather than by the predator."
The build happens every evening, on a schedule. About thirty minutes before sunset the spider leaves its daytime hideout on a leaf's underside, descends fifty centimeters or more, and spends up to four hours weaving fifteen to sixty silk tension lines into a fan, bundled at the center into a small wrapped cone. Then it retreats back upward — out of its own blast radius — and waits. Its one customer is the green tree ant, Oecophylla smaragdina, a notoriously aggressive weaver ant most predators leave alone entirely. When one investigates the cone and bites its base, that bite releases the whole tensioned fan at once: the ant is launched more than thirty centimeters straight up, at an acceleration exceeding 1,300 meters per second squared — on the order of 140 times Earth's own gravity — into the spider's own main web waiting above — where it lands entangled, and the spider comes down to wrap it.
Send an ant in below. Nothing here fires until it bites.
The slider draws six to twenty-four strands rather than the real fifteen to sixty — a browser tab can't usefully show sixty overlapping lines at this size — but the label keeps the real range in view rather than quietly rounding it away. Nothing about the count changes what happens when the ant bites: the whole fan releases together either way, the same all-or-nothing discharge a real tensioned line does.
Honest gap: the real launch takes well under a tenth of a second — this room stretches it to under half a second so it's watchable at all, the same liberty /cone and /touch already take with their own timings, and the approach walk beforehand is invented outright; the sources describe the bite and the launch, not how an ant spends the seconds before it. The sources also don't say whether a fired snare is rebuilt from scratch regardless of whether it caught anything, or only after it's been sprung — this room assumes the more conservative reading, that firing spends it, and lets you rebuild instantly rather than making you wait out a real evening's four hours. And every ant sent in here gets caught; the field observations behind this room describe the mechanism working, not a success rate against a failure one.
Source: Narendra, Joshi, Liprandi, Anderson & Wolff, Ballistic high-powered spider webs overcome dangerous prey defenses, Current Biology 36(12), R691 (2026), DOI: 10.1016/j.cub.2026.04.066.