Namibian desert ant: Ocymyrmex dekerus (photo: California Academy of Sciences)

Deserts are notoriously monotonous and contain little by way of visual cues to orient a traveller. So how do desert ants navigate from feeding sites to nests? It seems that ants use an internal calibration system to keeps track of the route and monitors where on the mapped route they are: “Many animals can navigate by means of path integration, in which an animal keeps a continuously updated record of its current direction and distance from some reference point as it moves away from that place. Equipped with such knowledge of its current location, the animal can reach other places of which it knows the path integration coordinates, with a precision that depends on the accuracy with which it estimates its position through path integration.” (Thomas S. Collett and Paul Graham, “Animal Navigation: Path Integration, Visual Landmarks and Cognitive Maps,” 14 Current Biology R475 (June 22, 2004) (full article).) This is similar to the method of dead reckoning used by mariners for centuries. Mariners plot direction and distances and maintain the information on a map or other record. Methods such as traverse board to allow even an illiterate sailor to record speed and direction during a watch. This can be shown to work in ants specifically by experiments which first train the ants and then displace them (see, e.g., Matthew Collett, Thomas S. Collett and Rüdiger Wehner, “Calibration of vector navigation in desert ants,” 9 Current Biology 1031 (1999) (summary with full article open access).

Path integration begins with an accumulator which is set to an initial state at a place, such as a nest. It is updated as the insect moves so that it always records the animal’s position relative to the initial place. Using this way method, vector navigation, is common in unfamiliar territory or when there is an absence of visual cues. In the short run, vector navigation produces acceptable results. Over a long course, however, minor errors in distance or vector will accumulate and cause large discrepancies. Therefore a combination of visual navigation and vector navigation is called for: the visual cues will help keep the vector navigation calibrated and the vector navigation will help in memorizing visual landmarks (Matthew Collett and Thomas S. Collett, “How do insects use path integration for their navigation?” 83 Biological Cybernetics 245 (2000) (abstract; article behind paywall)).

Cataglyphis fortis (Photo: Max Planck Institute for Chemical Ecology, Markus Knaden)

This method is used by rodents as well as insects. With respect to ants specifically, however, there is a global path integration system with path integration co-ordinates stored relative to the nest. Although ants use visual landmarks, landmarks seem to act independently of this path integration system and do not reset the ant’s global path integrator. Thus, in a experiment where the ants have been trained to make a run of a certain distance, when they are displaced they will overrun their nest, even bearing food and even with visual cues (David Andel and Rüdiger Wehner, “Path integration in desert ants, Cataglyphis: how to make a homing ant run away from home,” 271 Proc. R. Soc’y London B 1485 (2004) (full article open access)).

In nature, Cataglyphis desert ants seem also to be able to use olfactory cues. Various researchers from the Max Planck Institute for Chemical Ecology investigated Cataglyphis fortis in its native Tunisia. They used gas chromatography to isolate unique odour signatures in the desert. They then trained the ants in field experiments to recognise these odours pointing to a hidden nest entrance. Ants learned to associate their nest entrance with a single odour and discriminated the training odour against non-training odours (Kathrin Steck, Bill S. Hansson and Markus Knaden, “Smells like home: Desert ants, Cataglyphis fortis, use olfactory landmarks to pinpoint the nest,” 6 Frontiers in Zoology 5 (2009) (full article open access)).

Finally, Martin Müller and Rüdiger Wehner of the Brain Research Institute, University of Zürich, have conducted a study on how path integration relates to landmark learning in Namibian desert ants, Ocymyrmex. In “Path Integration Provides a Scaffold for Landmark Learning in Desert Ants,” 20 Current Biology 1368 (August 10, 2010) (full article) the authors show that these ants read out their path integrator when taking “snapshots” of landmark scenes even when they cannot see the goal. They do this by engaging in well-choreographed rotation movements integrated in spiral-like “learning walks.” During the rotations short stops occur, during which the ants orient themselves in the direction of the nest entrance. Since the entrances are invisible from those points, the ants must be reading out their current position by means of their path integrator. It has been speculated that this method is used by waps and bees, but this represents the first proof of such behavior among hymenopterans.