Wehner, R (2007). The desert ant's navigational toolkit: procedural rather than positional knowledge. In: Proceedings of the 63rd Annual Meeting of the Institute of Navigation, Cambridge, Massachusetts, 23 April 2007 - 25 April 2007, 1-4.
Full text not available from this repository.
Desert ants, Cataglyphis in North Africa and Melophorus in central Australia, exhibit amazing feats of navigation. This essay will focus on the architecture of the insect’s navigational toolkit. It will cover four main topics: First, vector navigation (path integration, PI) is the ant’s major means of navigation at least in so far as it is running all the time while the animal is on its foraging journey, irrespective of whether landmark-based systems of navigation are expressed as well, or are even dominating the insect’s behaviour. The path integrator relies on the simultaneous input from a compass and an odometer. Both tools of navigation have been studied extensively in Cataglyphis and are briefly summarized. In fact, among all animal navigators studied so far the insect’s skylight compass is the module of navigation that has been worked out best in neurophysiological terms. It has been computer simulated and implemented into an autonomous agent (Sahabot). The ant’s odometer – actually a stride integrator, or pedometer – computes distances covered along two- and three-dimensional paths, in the latter by projecting the 3-D path on to a 2-D plane. Second, while foraging within cluttered environments desert ants can learn a number of different routes, discriminate between inbound and outbound routes and follow these routes independently of the state of their path integrator (landmark-based route navigation, LN). If displaced to near-by locations, they rejoin a habitual route at any one point at which they come to hit it. Among the quite impressive amount of route memories which an ant is able to acquire, at any one time the proper memory is retrieved by a combination of internal (motivational) as well as external cues, and the proper route is selected. Third, PI and LN represent quite separate modules of navigation. PI is used to establish LN, but once established, LN can work independently of PI. In particular experimental paradigms this independence can lead to astonishingly aberrant behaviour. However, both systems might employ similar neural stratagems in so far as in either system information acquired during an animal’s foraging trip (a reference vector in PI and a stored sequence of visual and motor memories in LN) is later, on subsequent trips, continuously compared with currently incoming information (the current PI vector and the current sensorymotor data in LN). The ant moves so as to continually reduce the mismatch between reference “templates” on the one side and current sensory inputs and motor outputs on the other. Fourth, desert ants use the information provided by their PI and LN systems in a strictly cue-dependent procedural way rather than by acquiring and using a cartographical representation, or survey map, of their foraging terrain. Even if Cataglyphis moves along novel routes (a behavioural performance always regarded as \"the\" operational criterion indicative of using map information), it does so by memorizing different PI vectors and by merely matching current vector states to downloaded reference vectors. Following procedures rather than computing spatial positions is the insect’s way of navigation.
|Item Type:||Conference or Workshop Item (Paper), refereed, original work|
|Communities & Collections:||07 Faculty of Science > Institute of Zoology (former)|
|Dewey Decimal Classification:||570 Life sciences; biology
590 Animals (Zoology)
|Event End Date:||25 April 2007|
|Deposited On:||11 Feb 2008 12:14|
|Last Modified:||04 Apr 2012 12:10|
|Publisher:||Institute of Navigation|
Users (please log in): suggest update or correction for this item
Repository Staff Only: item control page