Do animals need to be able to answer the questions:
The answer, in many instances, is emphatically yes. Counting the number of eggs or offspring being tended can provide critical information to a parent. Assessing the number of food items, and being able to compare relative numbers of food items between patches of food, is elemental to efficient foraging. A reasonable hypothesis is that counting aids animals in surviving in many ways, and that an ability to count should be present in many, perhaps most, "higher" animals.
Testing this hypothesis, though, is difficult, and adequate tests have only been performed in a handful of animals. Test for counting are similar to tests for color vision, in that they usually involve a two-step process, first giving the test animal the opportunity to learn (count) objects, and then giving the test animal a choice between sets, one of which has the same number of objects as the learned set.
In honeybees, Apis mellifera, Chittka and Geiger (1995) trained workers to feed at a station along a 300 meter route from their hive. Small tents, which the bees could use as landmarks to guide their flight, were set up along the route. Bees which learned the route using a fixed number of tent landmarks changed their foraging distance--the location where they searched for the feeding station--if the investigators changed the number of landmarks. Inserting landmarks between the hive and the feeding station caused bees to fly a shorter distance before searching for food, while removing landmarks caused them to fly farther. This suggests that flying bees could count the number of landmarks, and that they extended their flight in search of additional landmarks when they did not find the expected number.
Coots face the challenge of preventing other birds from laying eggs in their nest. Female American coots, Filuca americana, often attempt to lay eggs in the nests of unsuspecting conspecifics. If they succeed, their young will be reared by another bird, a clear benefit to the cheating female. Coot eggs vary in the amount of speckling on their surface, a feature that sometimes allows coot females to tell their own eggs from eggs that have been imposed upon them by other females. Coot females who reject the eggs of cheaters must do so based on maintaining a count of their own eggs in the nest (Andersson 2003, Lyon 2003).
Tests of monkeys (Rhesus macaques, Macaca mulata) and chimpanzees (Pan troglodytes) suggest not only an ability to count, but also abilities to add and subtract. In Hauser's (1996, 2000) tests of counting and numerical representations in macaques, the basic experimental technique is to show the monkey a small number of objects and then hide objects with a screen. While the monkey cannot see the objects, the investigator may leave the same number (this is the control) or change the number by adding or subtracting objects (this is the experimental treatment). The screen is then removed and the amount of time the monkey spends looking at the objects is recorded. If the number hasn't changed (control) the monkey doesn't look very long, a second or so, but if the number has been changed, the monkey stares at the objects for three or four seconds. This and other tests suggest that macaques and chimpanzees are able to count at least small numbers of objects, and to perform simple addition.
Rats and pigeons, when tested using conditioned responses, show similar counting abilities. They can also generalize; rats trained to respond to a certain number of light flashes will respond to the same number of bell rings. Clever experiments are almost certain to reveal numerical abilities in a wide variety of animals.
Andersson, M, 2003, Behavioural ecology - Coots count NATURE. 422 (6931):
Return to beginning of Chapter 6
Return to Table of Contents
copyright ©2001 Michael D. Breed, all rights reserved