Berlin, 23th April 1993
.
        Coming back to the question of Ed Southwick on bees' orientation,
the theme is a rather exciting one. It also generated a polemic that
actually continues. To summarize it, the study of how bees represent large-
scale spatial relationships between landmarks in memory has intensified
since JL Gould (Science 232: 861-863, 1986) claimed that bees can acquire
topographically accurate cognitive maps of a landscape. "Cognitive maps"
are the mental analogue of a topographic map, i.e., an internal represen-
tation of the geometric relations among some points of the animal's environ-
ment. This kind of map- demonstrated for some vertebrates- allows a navi-
gator to compute and therefore to choice the shortest course from one
familiar site to any other familiar site, even if he has never traveled
this particular route and cannot see any landmark previously used to reach
this goal.
.
        Gould's view strongly contrasts with the classical hypothesis that
insects learn several stereotyped routes  through a landscape but cannot
derive a global knowledge of the lay of the land (Baerends, Tidjschr voor
Entomol 84: 68-275, 1941). Moreover, this hypothesis was thereafter extended
by Cartwright and Collett (J Comp Physiol 151: 521-543, 1983;Biol Cybern 57:
85-93, 1987) and Wehner (in: Neuroethology and Behavioural Physiology, Huber
& Markl eds, pp 366-381, Springer) who stated that a homing insect
(experiments were performed with honeybees and ants) uses a succession of
snapshots activated sequentially as the insect approaches its nesting site.
Thus, snapshots would act as a sequence of "memory pictures" leading to the
nest. In this case, distant marks guide to the broad area of the goal whereas
close marks allow for finally finding the exact position of the goal.
        The last point was demonstrated by Cheng et al (J. Comp Physiol A 161
469-475, 1987): They showed that in their final search for the goal, bees
weigh near landmarks more heavily than distant ones.
.
        In any case, several researchers tried to replicate Gould's experi-
ments (see for example Menzel et al, Z Naturforsch 45c 723-726, 1990;
Wehner & Menzel, Ann Rev Neurosci 13: 403-414; Dyer, Anim Behav 41:239-246,
Dyer in Proceedings og the Third International Congress of Neuroethology,
Montreal, 1992, p.16; Wehner, Menzel & Lehrer, in Dynamics & Plasticity in
Neuronal Systems, Elsner & Singer eds, 1988, Thieme).
        The basic experiment consisted in capturing previously trained bees
at a feeding place and displacing them in the dark to a novel site. There,
they were released and their flight direction was recorded to determine
whether bees actually were able to use a cognitive map, in which case they
should point to the hive also in this "new situation".  In all experiments,
performed by different researchers and in different places of the world, bees
flew, upon release, in the direction predicted by the snapshot route hypothe-
sis, i.e. in the compass direction in which they used to fly to the hive
from the feeding place, and not in the direction predicted by Gould's map
hypothesis (here I should enclose a draw!!!!!).
.
        Anyway, the controversy continues but it seems that honeybees do
not use our familiar cognitive maps.
.
 
.
                     Martin Giurfa
.
                     Institut fur Neurobiologie
                     Freie Universitat Berlin
                     Konigin Luise Strasse 28/30
                     1000 BERLIN 33
                     GERMANY
.
                     [log in to unmask]
.