498 ELECTRICAL SENSES 



(Figures 9C and 5B). The damping of the oscillations may result from 

 damping of the responses of individual receptor cells, decrease in the number 

 of cells oscillating, and desynchronization of the responses of individual 

 cells. Intracellular recording from single receptor cells is likely to be required 

 for a more complete analysis and for an estimation of the magnitude of the 

 responses. 



Oscillations of 20 /s in the clamped or short-circuited preparations are 

 considerably faster than the few-per-second oscillations seen in current 

 clamp, and their period is short compared to the duration of the late 

 outward current. In current clamp, the canal is electrically isolated and 

 outward current in the basal face can repolarize the lumenal face only 

 through the shunt resistance (Figure 6). Thus the lumenal action potential 

 apparently continues until there is activation of the long-lasting outward 

 current in this face. The long duration of this current accounts for the 

 relative slowness of repetitive discharges under current clamp. 



The question still remains as to how the high sensitivity arises. As pointed 

 out by Cole et al. (1970), the voltage gain of a membrane poised near 

 threshold becomes very large in that a small applied potential produces a 

 much larger change in potential, either generating an action potential or 

 returning the potential to the resting level. It seems likely also that the gain 

 of a tonically active cell can be quite large, and we ascribe the function of 

 increasing gain to the oscillations of the ampullae. Terzuolo and Bullock 

 (1956) found the crayfish stretch receptor to be sensitive to very small 

 currents applied extracellularly when it was tonically active but not when it 

 was inactive. (They were recording extracellularly, and transmembrane volt- 

 ages were not measured.) 



High sensitivity requires that the membrane stays poised near threshold. 

 From the observation that many receptors are spontaneously active, it 

 appears that adjustment to near threshold requires a tolerance of a certain 

 amount of spontaneous activity; that is, unless a cell occasionally exceeds 

 threshold it may not "know" that it is close to threshold. In addition, 

 activation and inactivation during subthreshold responses may have to be 

 reset by going through the action potential cycle. A further point about 

 tonic activity is that it allows the receptor to respond by a decrease in 

 activity and thus signal both increases and decreases in stimuli (or stimuli of 

 opposite sense). The hypothesized augmentation of sensitivity associated 

 with action potential generation requires confirmation by calculations from 

 voltage clamp data. Relevant comparative data are considered below. 



In considering the sensitivity of the receptor in other aspects than voltage- 

 to-voltage amplification, we might suspect a more sensitive relation between 

 voltage and transmitter release at the afferent synapses. (Since transmitter 

 release appears on morphological and comparative grounds to be quantized, 

 and since single quanta are detectable by many ordinary synapses, there is no 

 reason to suppose that electroreceptor synapses have larger quanta or more 

 sensitive transduction of transmitter concentration to postsynaptic impulse 

 frequency.) In tonic receptors of freshwater teleosts, there is indeed an 



