512
M. V. L. BENNETT
tor, particularly when the overlying water is removed (Figs. 15C and F).
(This procedure increases re in Fig. 9 and thus increases the fraction of
the potential produced by the nerve impulse across ri that is recorded
across T , . ) This small spikelike potential is shown to be the nerve impulse
by the fact that antidromic stimulation of the afferent fiber produces a
spike nearly identical to the orthodromic one (Fig. 15C). Thus, one can
directly measure orthodromic conduction time from receptor to the site
of recording in the afferent fiber. When large stimuli obscure the impulse
recorded from the receptor itself, one can still determine the time of impulse initiation from the orthodromic conduction time and recorded
afferent impulse.
If the strength of an anodal stimulus is increased, the time of initiation of the first impulse is soon reduced to a plateau value of about 1.5
msec which is the synaptic delay plus time for the postsynaptic potential
(PSP) to excite the nerve fiber ( Fig. 15D). In view of the high frequency
of h i n g the rise time of the PSP would be expected to be very small. At
a stimulus strength some 104 times greater than threshold, the latency
suddenly jumps to a new shorter value that is close to zero (Fig. 15E).
Apparently, the shorter latency results from direct stimulation of the
afferent fiber. A delay of 0 5 1 msec or more is characteristic of chemically mediated transmission [at room temperature (cf. Katz and Miledi,
1965; Bennett et al., 1967a)], and the delays at receptor synapses are
reasonable for this mode of transmission. If the receptor synapses were
electrically transmitting, one would expect increasing stimulus strength
to cause a gradual decrease in latency down to a very small value.
The responses of mormyrid tonic receptors are virtually identical to
those of the gymnotid receptors. A minor difference is that a single nerve
fiber innervates only a single receptor with several receptor cells rather
than a cluster of receptors. The same three lines of evidence indicate that
transmission is chemically mediated: (1) the synapse has a synaptic gap
and presynaptic vesicles; ( 2 ) a persistent excitation follows a brief excitatory stimulus and this excitation is little affected by an inhibitory
stimulus; and ( 3 ) there is an apparent synaptic delay of about 1 msec
( Bennett, 1965, 1967).
Morphologically similar receptors are found in Gymnurchus but have
not been studied physiologically ( Mullinger, 1969).
B. Electrical Equivalent and Input-Output Relationships
If the water over a receptor opening is replaced by air, a local electrode need pass much less current to produce a given voltage change
outside the receptor; the shunting by the water is reduced, that is, re in
M. V. L. BENNETT
tor, particularly when the overlying water is removed (Figs. 15C and F).
(This procedure increases re in Fig. 9 and thus increases the fraction of
the potential produced by the nerve impulse across ri that is recorded
across T , . ) This small spikelike potential is shown to be the nerve impulse
by the fact that antidromic stimulation of the afferent fiber produces a
spike nearly identical to the orthodromic one (Fig. 15C). Thus, one can
directly measure orthodromic conduction time from receptor to the site
of recording in the afferent fiber. When large stimuli obscure the impulse
recorded from the receptor itself, one can still determine the time of impulse initiation from the orthodromic conduction time and recorded
afferent impulse.
If the strength of an anodal stimulus is increased, the time of initiation of the first impulse is soon reduced to a plateau value of about 1.5
msec which is the synaptic delay plus time for the postsynaptic potential
(PSP) to excite the nerve fiber ( Fig. 15D). In view of the high frequency
of h i n g the rise time of the PSP would be expected to be very small. At
a stimulus strength some 104 times greater than threshold, the latency
suddenly jumps to a new shorter value that is close to zero (Fig. 15E).
Apparently, the shorter latency results from direct stimulation of the
afferent fiber. A delay of 0 5 1 msec or more is characteristic of chemically mediated transmission [at room temperature (cf. Katz and Miledi,
1965; Bennett et al., 1967a)], and the delays at receptor synapses are
reasonable for this mode of transmission. If the receptor synapses were
electrically transmitting, one would expect increasing stimulus strength
to cause a gradual decrease in latency down to a very small value.
The responses of mormyrid tonic receptors are virtually identical to
those of the gymnotid receptors. A minor difference is that a single nerve
fiber innervates only a single receptor with several receptor cells rather
than a cluster of receptors. The same three lines of evidence indicate that
transmission is chemically mediated: (1) the synapse has a synaptic gap
and presynaptic vesicles; ( 2 ) a persistent excitation follows a brief excitatory stimulus and this excitation is little affected by an inhibitory
stimulus; and ( 3 ) there is an apparent synaptic delay of about 1 msec
( Bennett, 1965, 1967).
Morphologically similar receptors are found in Gymnurchus but have
not been studied physiologically ( Mullinger, 1969).
B. Electrical Equivalent and Input-Output Relationships
If the water over a receptor opening is replaced by air, a local electrode need pass much less current to produce a given voltage change
outside the receptor; the shunting by the water is reduced, that is, re in