LEARNING BY W I N E INVERTEBRATES
29
operation, the responses of the worms were, however, by no means at
random. Runs of trials at which individuals turned consistently in one
direction were significantly more frequent than would be expected by
chance (Evans, 1963b) ; the behaviour is perhaps comparable with that
of cephalopods after analagous operations, where, too, consistent but
perverse behaviour in discrimination experiments sometimes follows
removal of parts of the brain known to be concerned in learning
(Young, 1960b).
Flint (1965) has recently shown that Nereis’s failure to learn Tmazes after removal of the supraoesophageal ganglion is at least
partly due to sensory deprivation. The antennae, eyes, and palps are
disconnected by the operation, as are the tentacular cirri if the circumoesophageal connectives are cut close to the suboesophageal ganglion.
If all these sense organs are removed or disconnected, leaving the brain
intact, most trained worms revert to random choices. But naive worms,
similarly treated, can be taught and, indeed, learn rather faster than
unoperated controls; apparently the worms can learn to make the
discrimination on the basis of a wide variety of sensory cues. Removal
of the supraoesophageal brain, leaving the connexion between the
tentacular cirri and the ventral nerve cord intact, had no effect on the
learned behaviour of five out of ten Nerei8 diver8icolor.
IV. CRUSTACEANS-AND Limulus
Arthropods are satisfactory experimental animals for a variety of
reasons. They are tough and many of them apparently remain in good
condition for long periods in the laboratory. They respond to visual
stimuli, which we can ourselves measure. They have large neurones,
so that electrophysiological analyses are possible. And perhaps
most important, from the point of view of behavioural work, their vdry
construction means that movements can be described with a detail and
completeness that is simply not possible with soft-bodied animals.
One result of this has been the attraction of ethologists, and the
behaviour of some arthropods (fiddler-crabs and hymenopterans, for
example) is now known in sufficient detail for individual learning to be
recognized where it occurs in the normal behaviour of the animals.
This means that information about learning in arthropods can come
from two quite different sources, unlike evidence for learning in most
other invertebrate groups, which is almost always derived from formal
training experiments .
A. The results of formal training experiments
In aquaria, crustaceans soon become “ tame ” and cease to respond
to stimuli that at first lead to withdrawal reaponsea, A resent attempt
29
operation, the responses of the worms were, however, by no means at
random. Runs of trials at which individuals turned consistently in one
direction were significantly more frequent than would be expected by
chance (Evans, 1963b) ; the behaviour is perhaps comparable with that
of cephalopods after analagous operations, where, too, consistent but
perverse behaviour in discrimination experiments sometimes follows
removal of parts of the brain known to be concerned in learning
(Young, 1960b).
Flint (1965) has recently shown that Nereis’s failure to learn Tmazes after removal of the supraoesophageal ganglion is at least
partly due to sensory deprivation. The antennae, eyes, and palps are
disconnected by the operation, as are the tentacular cirri if the circumoesophageal connectives are cut close to the suboesophageal ganglion.
If all these sense organs are removed or disconnected, leaving the brain
intact, most trained worms revert to random choices. But naive worms,
similarly treated, can be taught and, indeed, learn rather faster than
unoperated controls; apparently the worms can learn to make the
discrimination on the basis of a wide variety of sensory cues. Removal
of the supraoesophageal brain, leaving the connexion between the
tentacular cirri and the ventral nerve cord intact, had no effect on the
learned behaviour of five out of ten Nerei8 diver8icolor.
IV. CRUSTACEANS-AND Limulus
Arthropods are satisfactory experimental animals for a variety of
reasons. They are tough and many of them apparently remain in good
condition for long periods in the laboratory. They respond to visual
stimuli, which we can ourselves measure. They have large neurones,
so that electrophysiological analyses are possible. And perhaps
most important, from the point of view of behavioural work, their vdry
construction means that movements can be described with a detail and
completeness that is simply not possible with soft-bodied animals.
One result of this has been the attraction of ethologists, and the
behaviour of some arthropods (fiddler-crabs and hymenopterans, for
example) is now known in sufficient detail for individual learning to be
recognized where it occurs in the normal behaviour of the animals.
This means that information about learning in arthropods can come
from two quite different sources, unlike evidence for learning in most
other invertebrate groups, which is almost always derived from formal
training experiments .
A. The results of formal training experiments
In aquaria, crustaceans soon become “ tame ” and cease to respond
to stimuli that at first lead to withdrawal reaponsea, A resent attempt