YET another feature of human cognition that supposedly marks us out from the rest of nature has turned up in a wild animal. The New Caledonian crow, famous for its tool-making prowess, can upgrade previous designs to make more specialised devices. Even chimpanzees do not make these step-by-step improvements to their tools.
Our ability to improve technology, such as the way our ancestors made sequential improvements to stone tools, is central to our success. Now Gavin Hunt and Russell Gray at the University of Auckland in New Zealand have shown how crows on Grande Terre and Mar茅 Islands make step-by-step improvements to their tools. 鈥淥ur findings remove an important technological difference between humans and other animals,鈥� says Hunt. They will report their work in Proceedings of the Royal Society of London B.
Last year a team led by Alex Kacelnik at the University of Oxford found that New Caledonian crows could customise objects to act as tools for retrieving food (New Scientist, 17 August 2002, p 44).
Advertisement
The Auckland team studied four designs of probing tools made from Pandanus leaves: wide, narrow, single-stepped or multi-stepped. Wide tools are the simplest design, but can鈥檛 be slotted through cracks to gouge out prey. The narrow tools are an improvement, but are liable to bend. Stepped tools are probably best because they are more robust but have a narrow tip (see Graphic).
To deduce which came first, the researchers studied the geographical distribution of the different designs. That was relatively easy, as the birds cut the tools straight from the leaf, leaving behind an outline of the design. Some trees have over 100 of these impressions in their foliage.
Each of the four designs is found in its own continuous distribution, but all four overlap in one particular place, strongly suggesting that each design originated just once, as crows learned to improve an existing design and then copied each other. Kacelnik says the evidence is persuasive, but not yet conclusive. 鈥淭he final test must come from observing the behaviour.鈥�
All the same, different tools did not correlate with different ecological features, implying they weren鈥檛 simply designed separately for different tasks. And the overall distribution is also consistent with wide tools having been improved along two different routes.