Sequential organization and optimization of the nut-cracking behavior of semi-free tufted capuchin monkeys (Sapajus sp.).

Stone-aided nut-cracking requires the coordination of three elements: the agent must assemble nuts, a "hammer" stone and an "anvil." Under naturalistic settings, nut-cracking sites, constituted of anvil-like surfaces and already containing a hammer stone, can be fairly stable, lasting as long as the "hammer" stays in place. In an experiment with a semi-free-ranging group of tufted capuchin monkeys (Sapajus sp.) we observed the behavioral sequences leading to nut-cracking. We positioned nuts, hammer, and anvil at the vertices of a 10-m-sided equilateral triangle. Thus, to crack a nut the individuals had to visit the vertices and gather the movable elements (nut and hammer) at the anvil. Under such conditions, the monkeys systematically employed a nut-hammer-anvil vertex visit sequence, one of the shortest and more cost-effective possible routes. In the following experiment, we examined whether the gathering of the hammer after the nuts resulted solely from a "nut first" strategy or if the monkeys were also minimizing hammer transport costs. We positioned two hammers, of the same weight, at different distances from the nuts and anvil, so the cost of hammer transportation (energy and risk of injury) would be higher or lower depending on the choice of hammer (the hammer closer to the nuts being farther from the anvil). We found that, instead of collecting the closest hammer, after collecting the nut, the monkeys systematically chose the hammer closer to (and beyond) the anvil, thus minimizing transport costs.