L E m N I N O BY MARINE INVERTEBRATES
13
p. 15). Later, the animals can certainly be trained to make simple
discriminations (Sanders and Young, 1940).
It would appear from the Sepia experiments that at least some
responses are in the f i s t instance innately determined; this animal
does not by any means attack indiscriminately, even when faced with
potential prey for the first time in its life. It seems, however, exceedingly unlikely that innate mechanisms determine the responses of the
animal towards its prey throughout life, if only because cuttlefish
increase in size by a factor of a thousand or more during the first two or
three years after hatching. The range of animals to be eaten or avoided
must keep changing during this time, and adults, like newly hatched
cuttlefish, are by no means indiscriminate. Innate prey recognition
would imply a very elaborate size-dependent programme of maturation
within the CNS. Knowing how readily cephalopods can be taught in
the laboratory, it seems only reasonable to suppose that their responses
in the wild are based largely upon individual experience.
7 . Detectable attributes of objects touched
Octopuses handle things that they touch with the suckers, and
animals blinded by section of the optic nerves can be taught to make a
number of tactile discriminations on the basis of information from
sense organs in the rims of the suckers. These latter are sensitive to
both texture and taste, and in chemotactile training experiments
octopuses have proved to be at least one hundred times as sensitive as
the human tongue to substances such as quinine sulphate, hydrochloric acid and sucrose (Wells, 1963a).
Rather more extensive experiments have been made on the capacity
of octopuses to distinguish the physical attributes of objects touched.
It seems that they can be taught to distinguish textures, but not
weights or shapes, except inasmuch as shape differences involve sharp
corners which are detected as differences in texture (Wells, 1964a).
Thus, Perspex cylinders made to differ in texture by cutting narrow
grooves into the surface were regularly distinguished provided that
the proportion of grooved to smooth surface was different ; the distribution of the grooves-whether they ran up and down or around the
cylinders-was irrelevant (Wells and Wells, 1956, 1967). In some more
recent experiments, octopuses were required to distinguish between
smooth cylinders of different diameter. The proportion of errors made
was found to depend on the difference in surface curvature, that is on
differences in the degree of distortion of the applied suckers (Fig. 8).
The overall diameter of the cylinders was shown to be irrelevant by
experiments with compound cylinders made up from rods of smaller
13
p. 15). Later, the animals can certainly be trained to make simple
discriminations (Sanders and Young, 1940).
It would appear from the Sepia experiments that at least some
responses are in the f i s t instance innately determined; this animal
does not by any means attack indiscriminately, even when faced with
potential prey for the first time in its life. It seems, however, exceedingly unlikely that innate mechanisms determine the responses of the
animal towards its prey throughout life, if only because cuttlefish
increase in size by a factor of a thousand or more during the first two or
three years after hatching. The range of animals to be eaten or avoided
must keep changing during this time, and adults, like newly hatched
cuttlefish, are by no means indiscriminate. Innate prey recognition
would imply a very elaborate size-dependent programme of maturation
within the CNS. Knowing how readily cephalopods can be taught in
the laboratory, it seems only reasonable to suppose that their responses
in the wild are based largely upon individual experience.
7 . Detectable attributes of objects touched
Octopuses handle things that they touch with the suckers, and
animals blinded by section of the optic nerves can be taught to make a
number of tactile discriminations on the basis of information from
sense organs in the rims of the suckers. These latter are sensitive to
both texture and taste, and in chemotactile training experiments
octopuses have proved to be at least one hundred times as sensitive as
the human tongue to substances such as quinine sulphate, hydrochloric acid and sucrose (Wells, 1963a).
Rather more extensive experiments have been made on the capacity
of octopuses to distinguish the physical attributes of objects touched.
It seems that they can be taught to distinguish textures, but not
weights or shapes, except inasmuch as shape differences involve sharp
corners which are detected as differences in texture (Wells, 1964a).
Thus, Perspex cylinders made to differ in texture by cutting narrow
grooves into the surface were regularly distinguished provided that
the proportion of grooved to smooth surface was different ; the distribution of the grooves-whether they ran up and down or around the
cylinders-was irrelevant (Wells and Wells, 1956, 1967). In some more
recent experiments, octopuses were required to distinguish between
smooth cylinders of different diameter. The proportion of errors made
was found to depend on the difference in surface curvature, that is on
differences in the degree of distortion of the applied suckers (Fig. 8).
The overall diameter of the cylinders was shown to be irrelevant by
experiments with compound cylinders made up from rods of smaller