M. V. L. BENNETT
566
trically excitable responses of receptor cells of electroreceptors are insensitive to tetrodotoxin (Zipser, 1971). In the same species, and in most
other forms, the ordinary Na channels mediating propagated spikes are
blocked by tetrodotoxin. However, tetrodotoxin resistant spikes can be
obtained in prcsynaptic terminals of the squid giant synapse and apparently of the neuromuscular junction as well (Katz and Miledi,
1969a,b). These neuronal responses are dependent on divalent ions and
presumably mediate the normal Ca influx required for transmitter
secretion. Although the ionic dependence of the receptor cell responses
has yet to be studied, it is tempting to suppose that their spike generating mechanism represents evolution of an increased number of electrically excitable channels that carry divalent ions. Smaller numbers
of these channels would presumably be present in receptor cells lacking
an obvious electrical response where they would mediate transmitter
release. Divalent ions certainly affect transmitter release in electroreceptors, although there is not the clear-cut antagonism between Ca and
Mg that has been observed at a number of interneuronal synapses
( Steinbach and Bennett, 1971).
An interesting question about receptor synapses is whether transmitter is released in small packets or quanta as has been so elegantly
documented for the neuromuscular junction ( see Katz, 1969). The presence of presynaptic vesicles at receptor synapses suggests that release is
quantized, and a certain amount of indirect confirmatory evidence involving analysis of intervals between afferent impulses has been obtained
for lateral line receptors ( Harris and Flock, 1967). The irregularity of intervals between impulses from tonic receptors could be a result of quanta1
variations in the PSP, and intracellular recording from nerve terminals
should provide more direct evidence ( Fig. 18).
It is likely that an input-output relationship similar to that of tonic
receptors will be found in other receptor cells, where more or less linearly
graded information about stimulus amplitude is transmitted centripetally.
The absolute sensitivity is also likely to be similar to that in electroreceptors, at least in the most sensitive receptors for a given modality. In all
the sensitive receptors there should be strong selection pressure for evolution of a presynaptic face that can detect very small signals generated by
the outer face. The cochlear microphonic is estimated to have a value of
0.01 pV at threshold for hearing (Machin and Lissmann, 1960). The
potcntial across the inner face of the hair cells is undoubtedly much larger
since the receptor should have evolved so that most of the voltage drop is
developed across the inner face. A hundred- to a thousandfold larger
potential across the secretory membrane is not unreasonable and brings
the actual potential across the secretory face into the range where elec-
566
trically excitable responses of receptor cells of electroreceptors are insensitive to tetrodotoxin (Zipser, 1971). In the same species, and in most
other forms, the ordinary Na channels mediating propagated spikes are
blocked by tetrodotoxin. However, tetrodotoxin resistant spikes can be
obtained in prcsynaptic terminals of the squid giant synapse and apparently of the neuromuscular junction as well (Katz and Miledi,
1969a,b). These neuronal responses are dependent on divalent ions and
presumably mediate the normal Ca influx required for transmitter
secretion. Although the ionic dependence of the receptor cell responses
has yet to be studied, it is tempting to suppose that their spike generating mechanism represents evolution of an increased number of electrically excitable channels that carry divalent ions. Smaller numbers
of these channels would presumably be present in receptor cells lacking
an obvious electrical response where they would mediate transmitter
release. Divalent ions certainly affect transmitter release in electroreceptors, although there is not the clear-cut antagonism between Ca and
Mg that has been observed at a number of interneuronal synapses
( Steinbach and Bennett, 1971).
An interesting question about receptor synapses is whether transmitter is released in small packets or quanta as has been so elegantly
documented for the neuromuscular junction ( see Katz, 1969). The presence of presynaptic vesicles at receptor synapses suggests that release is
quantized, and a certain amount of indirect confirmatory evidence involving analysis of intervals between afferent impulses has been obtained
for lateral line receptors ( Harris and Flock, 1967). The irregularity of intervals between impulses from tonic receptors could be a result of quanta1
variations in the PSP, and intracellular recording from nerve terminals
should provide more direct evidence ( Fig. 18).
It is likely that an input-output relationship similar to that of tonic
receptors will be found in other receptor cells, where more or less linearly
graded information about stimulus amplitude is transmitted centripetally.
The absolute sensitivity is also likely to be similar to that in electroreceptors, at least in the most sensitive receptors for a given modality. In all
the sensitive receptors there should be strong selection pressure for evolution of a presynaptic face that can detect very small signals generated by
the outer face. The cochlear microphonic is estimated to have a value of
0.01 pV at threshold for hearing (Machin and Lissmann, 1960). The
potcntial across the inner face of the hair cells is undoubtedly much larger
since the receptor should have evolved so that most of the voltage drop is
developed across the inner face. A hundred- to a thousandfold larger
potential across the secretory membrane is not unreasonable and brings
the actual potential across the secretory face into the range where elec-