EXCITATORY AND INHIBITORY PROCESSES 297 



phase; i.e. the cell interior becomes transiently more negative in relation to 

 the outside in the post-spike period (Fig. 8) and the spike recovery phase 

 becomes progressively shorter. This is better shown in Fig. 9 where the 

 arrows indicate the turning point when the generator action takes over, 

 working in the opposite direction to the recovery phase, again depolarizing 

 the cell soma to the level determined by the extent of stretch. In addition, 

 during stretch the levels of repolarization, as referred to the relaxed resting 

 potential, become less complete. 



The after-positivity seen in these experiments is a hyperpolarization relative 

 to the reduced membrane potential maintained by a persisting generator 

 action during stretch. Actually, during the recovery phase of the action 

 potential the generator action is considerably reduced although it is not 

 completely wiped out since the recovery phase does not return to the full 

 resting potential level. Therefore, it seems legitimate to assume that this axon- 

 type of impulse does not propagate through the distal dendrite portions. Also, 

 progressive reduction in the extent of repolarization following an impulse as 

 stretch is increased suggests that the impulses penetrate less and less distally 

 as the dendrites become more depolarized. More direct evidence of a persist- 

 ing generator action during antidromic invasion of the dendrites by all-or- 

 none impulses will be shown later on. In fact, inhibitory impulses during the 

 peak of antidromic after-positivity may further repolarize the cell (see Fig. 18). 



Sometimes an axon impulse does not reach the soma but is blocked 

 somewhere along the axon, presumably near the axon hillock. This phe- 

 nomenon occurs almost exclusively in relaxed preparations. In such case the 

 axon impulse is detected by an intracellular lead recording from the soma as 

 a small potential of short duration wliich has spread electrotonically. In 

 other instances these blocked antidromic impulses set up a local response in 

 the soma which is distinguishable from a merely electrotonically conducted 

 impulse because of its amplitude and time course. It is, generally, considerably 

 slower and it may attain magnitudes of about 20 mV. If a second impulse is 

 sent through the axon it adds to the potential of the first and at a critical level 

 of about 20-25 mV a full soma impulse may arise. 



The converse picture is obtained in the same cell if its dendrites are de- 

 polarized by a small steady stretch reducing the resting potential by several 

 milHvolts. In this case, an antidromic impulse is capable of invading the 

 soma and it shows a small after-positivity. However if two antidromic 

 impulses are brought close together the second impulse disappears, leaving 

 only a local response. This observation shows that axon-soma transmission 

 can be greatly influenced by changing the membrane potential. Axon-soma 

 delays are longest when soma invasion is critical. Whenever such a delay 

 occurs it may be reduced or made to disappear by stretch. Some of these 

 delay times measured from the first inflection of the potential to the spike 

 peak are, occasionally, quite long (see Fig. 10). 



