Unpolarized electrotonic connections between the two lateral giant axons of crayfish nerve cord have been demonstrated. Presumably they are also the sites of the tightly coupled unpolarized ephaptic junctions by which activity from one axon may be transmitted to the other. Commissures between the segmental branches of the two s o n s are probably the sites of the junctions. If septa1 transmission is blocked in one axon, the inactive segment can be excited by the activity of the other axon. The segmentally arrayed ladder of ephaptic junctions forms a series of delay lines and circus activity can occur in the loops between the two axons. Two segments form a loop with a delay line of 4 msec. The frequency of repetition is about 250/second. Each additional segment introduces another delay of about 4 msec. For a three-segment loop the repetition rate is about 125/second. The delays at ephaptic junctions between neurons in vertebrate nervous systems are probably too brief to cause repetitive activity.
Functional interconnections between the two lateral giant axons of the crayfish nerve cord were demonstrated by a number of observers (Furshpan and Potter, '59; Takeda, '59; Watanabe and Grundfest, '61). Watanabe and Grundfest ('61) suggested that these interconnections were due to ephaptic junctions between the commissural branches which are given off at the caudad end of each axon segment and which come together near the midline (Johnson, '24). The existence of this ladder-like system of interconnections, it was suggested, could likewise account for the trains of repetitive activity which were observed or which could be induced experimentally in the axons (Furshpan and Potter, '59; Watanabe and Grundfest, '61 ). The repetitive activity might result from circus reexcitation in a loop that was established when one of the components had failed to respond in the normal tempo of activity. Activity that has been ascribed to circus reexcitation was observed in earthworm giant axons (Wilson, '61) and in leech neurons (Hagiwara and Morita, '62). However, in none of the work cited has this effect been demonstrated directly with intracellular re\ cording.
Reverberatory activity is probably responsible for some of the arrhythmias that