412 Animal Memory and Cognition
under both laboratory and field conditions and includes a
variety of topics, ranging from the navigational abilities of
birds to maze learning in rats. Accordingly, spatial learning
occupies the attention of an equally diverse array of investi-
gators, ranging from evolutionary biologists to experimental
psychologists.
Consider an application of evolutionary thinking to spatial
learning. Gaulin and Fitzgerald (1989) reasoned that spatial
abilities would be genetically selected-for more often in males
than in females of polygymous species of meadow voles, be-
cause the male maintains a larger home range in which to seek
potential mates or resources to attract mates. They compared
the polygamous meadow voles to the monogamous pine voles.
They found, first of all, that sex difference in favor of a larger
home range occurred in male meadow voles but not male pine
voles. They compared the two species on a variety of mazes of
increasing difficulty, finding that males outperformed females
in meadow voles but not in pine voles. Moreover, the hip-
pocampus, which is considered to be of importance in spatial
learning, was found to be larger in meadow voles than in pine
voles (Jacobs, Gaulin, Sherry, & Hoffman, 1990).
Sex differences in spatial learning occur in a variety of
species, including humans. Human males outperform human
females in a variety of spatial learning tasks ranging from
mental rotation of objects to map reading. Human females
outperform human males in recalling the locations of objects
interspersed in a room, and the difference is larger for inci-
dental learning than for direct learning. Silverman and Eals
(1992) interpret these sex differences in humans in evolu-
tionary terms. It is believed that in the Pleistocene (the period
in which much human evolution is considered to have oc-
curred), males tended to hunt, and so traveled greater dis-
tances than females (favoring male spatial learning), whereas
females gathered items such as vegetables (which favored
learning the location of items by females).
Spatial learning thus provides a good illustration of the
evolutionary approach to animal and human cognition. The
evolutionary approachassumes that the cognitive ability
possessed by a species was designed by evolution to meet the
demands of its particular environment. Thus the evolutionary
approach is fully prepared to discover that a given species
may possess a unique adaptation. Unique adaptations are
hardly rare, and several have already been mentioned:
echolocation in bats, dancing in bees, language in humans.
That each of these has been considered to be instinctive by
some does not necessarily lessen their importance to cogni-
tion. For example, although our species’ ability to master lan-
guage may be instinctive to a great degree, there is reason to
suppose that language development was a major factor in
human problem solving, tool using, and related cognitive ac-
tivities (see, e.g., Bickerton, 1998).
The radial maze mentioned earlier is an important tool
used to investigate spatial learning and other problems in the
laboratory. A pellet of food is placed at the end of each arm of
the maze. The rat is placed on the central platform and is free
to enter the arms. Rats easily master radial mazes, as indi-
cated by their entering each arm only once. Efficient perfor-
mance of this sort may itself have an instinctive or unlearned
basis. It has long been known that in, say, a T-maze, the rats,
having obtained food in one arm, typically avoid that arm in
favor of responding to the other arm, a behavior sometimes
calledwin-shift. In foraging in the wild it may be of benefit to
animals such as rats to avoid going back to a location in
which food was obtained because the availability of food at
that source may be less likely than at some new source.
Not only are eight-arm radial mazes solved efficiently by
rats, but so, too, are mazes containing a greater number of
arms. The most impressive of these was a hierarchical maze
employed by Roberts (1998). That maze consisted of eight
primary arms. At the end of each primary arm were three
branching secondary arms. The rats performed very well in
this maze under a variety of conditions. For example, under
one condition the rats were allowed to enter the primary arms
with some of the secondary arms blocked off. In a subsequent
test phase the rats entered the previously blocked secondary
arms with a high degree of accuracy. This finding indeed sug-
gests that the rat’s memory for spatial location is well devel-
oped. Another indication of the rat’s impressive memory in
this situation is the difficulty of producing retroactive inhibi-
tion. In retroactive inhibition, as indicated earlier, memory of
an initial task is reduced by provision of a subsequent task
prior to testing of the initial task. Various means of producing
retroactive inhibition in the radial maze have been used, in-
cluding learning another maze in a different room. In an
impressive experiment by Beatty and Shavalia (1980), rats
exhibited little to no retroactive interference when they were
required to enter four arms of a different radial maze within a
4 hr retention interval.
Rats can learn radial mazes by employing either distal
cues,such as the shape of the room, landmark cues,such as
objects in the environment, or intramaze cues,such as light
differences in the appearance of the maze in different areas.
In a very interesting experiment Williams and Meck (1991)
reported that in the radial maze distal cues were used more
often by males than by females, with landmarks cues being
used more often by females than by males.
How animals represent spatial conditions is a topic of great
interest. According to one view, animals such as rats may pos-
sess acognitive mapof the environment that consists of both
a general framework, within which objects are located rela-
tive to each other, and general features of the environment,
such as its shape (O’Keefe & Nadel, 1978). Gallistel (1990)