5. TEMPERATURE RECEPTORS
129
having a range of relative touch-temperature sensitivities. An alternative
method could be to depend on stimulation arising from the fish‘s own
movements, in which case reafference (von Holst and Mittelstaedt, 1950)
could form the basis for the comparison.
The idea that one modality of stimulation can be detected by an
animal, using multimodal receptors only, is relatively recent. For instance,
the inadequacies of the old doctrine of specificity of receptors (and
especially of anatomical specificity) have been stressed for mammals in
various “pattern of stimulation” hypotheses ( e.g., Sinclair, 1955; Lele
and Weddell, 1959). However, modern views retain a considerable degree
of functional specificity, while admitting that especially within one modality the information from a single receptor may be ambiguous, and positive
identification of the stimulus may only be possible by comparison of the
responses in a number of channels (e.g., taste: Cohen et al., 1955; and
the review of cutaneous sensitivity in mammals by Melzack and Wall,
1962). In some cases, however, simpler explanations for multimodal units
may be adequate, either that one stimulus is mistaken for the other
(“seeing stars,” or the explanation of Weber’s Deception-cold weights
feel heavier-in terms of the stimulation of cold-touch units) or that the
two modalities can be separated by reference to coexisting specific units
for each of the modalities.
But neither of these can apply in teleosts, for temperature is evidently
not confounded with touch, and there are as yet no specific thermoreceptors known. Thus the computing problem for the CNS must be
complex, and the methods used are completely unknown. One particular
difficulty not found in the mammalian touch-temperature units is that the
fish mechanoreceptors are rapidly adapting, not tonic. But it is a reasonable hypothesis, all the same, since, as Spath points out, an aquatic animal
is warmed or cooled by convection or conduction, never under natural
conditions by radiation, and the former two necessarily involve mechanical stimulation. In fact, when Dijkgraaf (1940) tested the effect of
radiant heat on the exposed fish, he found that they responded to the
warming with a much higher threshold than in his other experiments.
The large number of touch-temperature units in mammals supports the
idea that the vertebrate CNS can make suitable computations, and
recently Bailey (1969) has shown that the most sensitive thermal
responses in lizard skin are given by touch-temperature units very similar
to those of the teleost. There is one extra source of information for the
lizard, namely, that the adaptation time for the phasic response is temperature dependent, as well as the initial impulse frequency.
The fact that the teleost touch-temperature units are “cold sensitive,”
i.e., have an increased response on cooling and decreased in warming,
129
having a range of relative touch-temperature sensitivities. An alternative
method could be to depend on stimulation arising from the fish‘s own
movements, in which case reafference (von Holst and Mittelstaedt, 1950)
could form the basis for the comparison.
The idea that one modality of stimulation can be detected by an
animal, using multimodal receptors only, is relatively recent. For instance,
the inadequacies of the old doctrine of specificity of receptors (and
especially of anatomical specificity) have been stressed for mammals in
various “pattern of stimulation” hypotheses ( e.g., Sinclair, 1955; Lele
and Weddell, 1959). However, modern views retain a considerable degree
of functional specificity, while admitting that especially within one modality the information from a single receptor may be ambiguous, and positive
identification of the stimulus may only be possible by comparison of the
responses in a number of channels (e.g., taste: Cohen et al., 1955; and
the review of cutaneous sensitivity in mammals by Melzack and Wall,
1962). In some cases, however, simpler explanations for multimodal units
may be adequate, either that one stimulus is mistaken for the other
(“seeing stars,” or the explanation of Weber’s Deception-cold weights
feel heavier-in terms of the stimulation of cold-touch units) or that the
two modalities can be separated by reference to coexisting specific units
for each of the modalities.
But neither of these can apply in teleosts, for temperature is evidently
not confounded with touch, and there are as yet no specific thermoreceptors known. Thus the computing problem for the CNS must be
complex, and the methods used are completely unknown. One particular
difficulty not found in the mammalian touch-temperature units is that the
fish mechanoreceptors are rapidly adapting, not tonic. But it is a reasonable hypothesis, all the same, since, as Spath points out, an aquatic animal
is warmed or cooled by convection or conduction, never under natural
conditions by radiation, and the former two necessarily involve mechanical stimulation. In fact, when Dijkgraaf (1940) tested the effect of
radiant heat on the exposed fish, he found that they responded to the
warming with a much higher threshold than in his other experiments.
The large number of touch-temperature units in mammals supports the
idea that the vertebrate CNS can make suitable computations, and
recently Bailey (1969) has shown that the most sensitive thermal
responses in lizard skin are given by touch-temperature units very similar
to those of the teleost. There is one extra source of information for the
lizard, namely, that the adaptation time for the phasic response is temperature dependent, as well as the initial impulse frequency.
The fact that the teleost touch-temperature units are “cold sensitive,”
i.e., have an increased response on cooling and decreased in warming,