CONDUCTION OF THE NERVE IMPULSE 



95 



applied between No and Ni sets up through the mem- 

 brane of Ni a current, the time course of which is 

 distorted by current flow through the myelin sheath. 



The records in figure i6 show the behavior of the 

 membrane potential at threshold as observed with 

 this arrangement. The duration of the stimulating 

 pulse was varied in the range between 0.05 and 6.4 

 msec. At every stimulus duration the stimulus inten- 

 sity was adjusted to threshold, and without changing 

 the intensity, five to seven sweeps of the oscillograph 

 beam were superposed on each record. Because of 

 spontaneous variation in the property of the nerve 

 fiber (14, 99), the node sometimes responded with a 

 full-sized action potential and sometimes failed to 

 produce an action potential. 



We may define the 'threshold membrane potential' 

 as the highest potential level of the membrane which, 

 after the end of the applied stimulating pulse, decays 

 without producing an action potential (63, 126). The 

 level of the threshold membrane potential measured 

 from the resting potential level is often called the 

 'threshold (or critical) depolarization.' It is seen in 

 the records that the threshold depolarization is 

 practically independent of the stimulus duration. 

 When the duration is short (e.g. 0.05 msec), a very 

 large voltage (200 mv) is needed to excite the node; 

 the observed fact is that this high a voltage is required 

 to raise the membrane potential within a short period 

 of time to the threshold level, which is about 15 mv 

 above the resting potential. This is exactly what has 

 been assumed in most of the classical theories of nerve 

 excitation. 



As we have discus.sed in a previous section, the 

 surface membranes of the nerve fiber, both the myelin 

 sheath and the nodal membrane, have relatively large 

 capacities. Consequently, in order to raise the mem- 

 brane potential by a constant amount, higher stimu- 

 lus intensities are required at shorter stimulus dura- 

 tions. 



However, there is in this type of experiment one 

 complication that has not been fully understood by 

 previous investigators who worked only on nerve 

 trunks. It is the gradual rise in the membrane po- 

 tential that precedes the rapid rising phase of the 

 action potential in stimulation by a long pulse (see 

 fig. 16, record for 6.4 msec). In response to a long 

 stimulating pulse, an action potential either appears 

 within a few msec, (within 10 msec, at the most) 

 after the start of the pulse or fails to appear at all. 

 When the action potential fails to appear, the be- 

 havior of the membrane potential does not diverge 

 from what is expected from the physical constants of 



the resting nerve fiber. When the membrane potential 

 starts to diverge distinctly from the simple time 

 course, provided that the applied pulse has not been 

 withdrawn within 5 msec, or so, there is alw-ays (at 

 least in a normal node) an action potential. 



Action potentials evoked by long stimulating 

 pulses have a more-or-less gradual rising phase 

 followed by a phase of rapid potential rise. If the 

 applied stimulating pulse is withdrawn before the 

 start of the rapid potential rise, the production of 

 a full-sized action potential is prevented. Such a 

 gradual potential rise followed liy a sudden potential 

 fall caused by a withdrawal of the applied pulse is 

 seen in the record labelled 46 mv (1.6 msec.) in 

 figure 16. 



The nonlinear phenomenon just described is con- 

 sidered at present to indicate the following. The pro- 

 duction of an action potential is a kind of 'regenerative' 

 or 'autocatalytic' process similar to the explosion 

 induced by heating of a mass of gunpowder (105). 

 The heat applied from outside causes combustion in 

 only some of the gunpowder particles; the heal arising 

 from these particles in turn induces combustion in 

 other neighboring particles. Similarly, when the 

 stimulus duration is sufficiently long, the start of a 

 ■ response' (the start of comljustion in the analogy 

 above) tends to raise the membrane potential (tem- 

 perature) together with the applied stimulus (applied 

 heat). If the external source of current (heat) is 

 maintained, this process eventually raises the mem- 

 brane potential (temperature) to a critical explosive 

 point. If, however, the applied pulse is withdrawn 

 before the critical level of the membrane potential is 

 reached, the potential returns to its resting level 

 along a variable time course. With very short current 

 pulses, the membrane potential has to be raised by 

 the external source up to the critical level. ^ 



In the excitation of the invertebrate axon with 

 rectangular current pulses, results similar to those in 

 figui-e 16 have been obtained by several investigators 

 [e.g. Hodgkin & Rushton (63)]. To stress the similar- 

 ity between the vertebrate myelinated ner\e fiber 

 and the squid a.xon, unpublished records obtained 

 by Hagiwara and others are presented in figure 1 7. 

 The arrangement of the stimulating and recording 

 electrodes used is similar to that in figure 14; two 

 metal wires about 30 mm in length were inserted 

 along the axis of an axon. Pulses of constant current 



' It should be pointed out that some physiologists have 

 slightly different viewpoints in regard to the statement made 

 in this sentence (104, 107). 



