214
GEORGE SZÉKELY
of optic fibers with receptive fields of different size and function. By
similar analysis of the histology of bipolar cells, they assumed that an
outer sheet of the inner plexiform layer is concerned with average illumination, since those bipolars that connect only to cones have a broad
dendritic spread in this sheet. This, they think, is consistent with the
concept that so-called dimming detector fibers are sensitive only to
changes of light intensity, originating presumably from ganglion cells
with broad, single-level, dendritic arborization in the outer sheet of the
inner plexiform layer. An inner sheet of the inner plexiform layer is
concerned with boundaries; here are found the terminals of bipolars
making contact with both cones and rods. Similarly, they have identified
fibers mediating only boundary detection, and fibers with other functions
combining information from both sheets. The size of receptive fields
relative to the dendritic arborization patterns was such that the anatomical and functional groups could be matched.
This remarkable example of a highly probable form-function relationship, in which the form, as it were, lays down the function, may have
an interesting developmental relevance. The connections between the
receptor and ganglion cells are carried out by the parallel-oriented bipolar cells. One cannot see fibers crossing or bending to look for their
corresponding mates with which to make synaptic contact, and it is
also improbable that the parallel organization would be the result of a
secondary rearrangement of the components. One is under the impression
that in the retina the random connections of morphologically welldefined elements result in a very complex function. Instead of determining the specificity of a few million connections, the morphological determination of a few types of neurons and their statistical distribution
within the appropriate cell layers ensures the specific activity of the
retina.
The functional significance of the dendritic arbor is also emphasized
by the strong suspicion that electrotonic transmission may occur between
adjacent dendrites of separate neurons and lead to propagating depolarization of the dendritic surface. The work on which this assumption is
based was done on motoneurons of fishes (Bennett, 1960; Bennett et al.,
1959, 1963) and of frogs (Washizu, 1960). The assumption may be
ventured that the cellular interaction, which is effected partly by electrotonic transmission and partly by internuncial synaptic connections,
is more or less determined by the form and extent of the dendritic arbor.
Thus, the density and shape of cells within a given region may also have
a functional relevance in the determination of the output pattern of a
pool of neurons. An example supporting this point of view can be found
GEORGE SZÉKELY
of optic fibers with receptive fields of different size and function. By
similar analysis of the histology of bipolar cells, they assumed that an
outer sheet of the inner plexiform layer is concerned with average illumination, since those bipolars that connect only to cones have a broad
dendritic spread in this sheet. This, they think, is consistent with the
concept that so-called dimming detector fibers are sensitive only to
changes of light intensity, originating presumably from ganglion cells
with broad, single-level, dendritic arborization in the outer sheet of the
inner plexiform layer. An inner sheet of the inner plexiform layer is
concerned with boundaries; here are found the terminals of bipolars
making contact with both cones and rods. Similarly, they have identified
fibers mediating only boundary detection, and fibers with other functions
combining information from both sheets. The size of receptive fields
relative to the dendritic arborization patterns was such that the anatomical and functional groups could be matched.
This remarkable example of a highly probable form-function relationship, in which the form, as it were, lays down the function, may have
an interesting developmental relevance. The connections between the
receptor and ganglion cells are carried out by the parallel-oriented bipolar cells. One cannot see fibers crossing or bending to look for their
corresponding mates with which to make synaptic contact, and it is
also improbable that the parallel organization would be the result of a
secondary rearrangement of the components. One is under the impression
that in the retina the random connections of morphologically welldefined elements result in a very complex function. Instead of determining the specificity of a few million connections, the morphological determination of a few types of neurons and their statistical distribution
within the appropriate cell layers ensures the specific activity of the
retina.
The functional significance of the dendritic arbor is also emphasized
by the strong suspicion that electrotonic transmission may occur between
adjacent dendrites of separate neurons and lead to propagating depolarization of the dendritic surface. The work on which this assumption is
based was done on motoneurons of fishes (Bennett, 1960; Bennett et al.,
1959, 1963) and of frogs (Washizu, 1960). The assumption may be
ventured that the cellular interaction, which is effected partly by electrotonic transmission and partly by internuncial synaptic connections,
is more or less determined by the form and extent of the dendritic arbor.
Thus, the density and shape of cells within a given region may also have
a functional relevance in the determination of the output pattern of a
pool of neurons. An example supporting this point of view can be found