LEARNINQ BY MARINE INVERTEBRATES
5
and the orientation of things that they see. The degree of transfer to
similar shLpes of different size depends on the size difference, and is
made most readily to shapes somewhat larger than those used in
training (Sutherland and Carr, 1962). Visual acuity is not a limiting
factor in any of these experiments (Sutherland, 1963a), as might be
S hope,
Horizmtol
Vertical
Proportion
projection
projection of correct
responses
n
1 1
1
2
T J-4 I-4- 1 5 9 %
t 560’0
--I
FIQ. 2. Pairs of shapes that octopuses can and cannot learn to distinguish. Percentage
figures show the proportion of correct responses made by groups of octopuses in the
first 60 trials. Projections, see text. (From facts given in Sutherland, 1967.)
expected from the extremely close packing of the retinal elements and
the very large numbers of optic nerves (Young, 1962, 1963b).
Differences in orientation are most readily distinguished when they
are in the horizontal and vertical planes. Oblique rectangles are not
distinguished at all under the same conditions. The results summarized
in Fig. 2 and most other shape discrimination experiments can be
summarized as showing that octopuses tend to classify the things that
they see on a basis of their horizontal and vertical extents ; by comparing the projections of any two shapes in these planes it is possible to
.
5
and the orientation of things that they see. The degree of transfer to
similar shLpes of different size depends on the size difference, and is
made most readily to shapes somewhat larger than those used in
training (Sutherland and Carr, 1962). Visual acuity is not a limiting
factor in any of these experiments (Sutherland, 1963a), as might be
S hope,
Horizmtol
Vertical
Proportion
projection
projection of correct
responses
n
1 1
1
2
T J-4 I-4- 1 5 9 %
t 560’0
--I
FIQ. 2. Pairs of shapes that octopuses can and cannot learn to distinguish. Percentage
figures show the proportion of correct responses made by groups of octopuses in the
first 60 trials. Projections, see text. (From facts given in Sutherland, 1967.)
expected from the extremely close packing of the retinal elements and
the very large numbers of optic nerves (Young, 1962, 1963b).
Differences in orientation are most readily distinguished when they
are in the horizontal and vertical planes. Oblique rectangles are not
distinguished at all under the same conditions. The results summarized
in Fig. 2 and most other shape discrimination experiments can be
summarized as showing that octopuses tend to classify the things that
they see on a basis of their horizontal and vertical extents ; by comparing the projections of any two shapes in these planes it is possible to
.