

LIVING TORPEDOS IN BERLIN. 527 



first momnt, falls with longer duration of the current, and 

 shows itself again in full force when the current is again closed 

 for a short time. 



It seems rather, that we must suppose that the heterodromous 

 current, as soon as its density surpasses a certain value, the 

 threshold value in the preparation, meets with a special resistance, 

 and that a certain time is required to overcome it. This would 

 explain why irreciprocity of conduction is more strongly developed 

 with current impulses of short duration, while comparatively little 

 is left if the current is of longer duration. 



The small amount of irreciprocity when the flow is continuous, 

 may perhaps not be dependent on the same cause as the extreme 

 irreciprocity with currents of minimal duration. For a remark 

 made in the first communication (p. 456) ought not to be forgotten 

 here. Inasmuch as there is no doubt of the existence of homo- 

 dromous positive polarisation, it must be regarded as a concomitant 

 cause of apparently irreciprocal conduction, even though this is 

 actually traced to an unequal resistance in the two directions. 

 But when the time of closing is long, the irreciprocity becomes so 

 small, that we have no longer the same ground for holding it 

 improbable, as at first sight seemed necessary, that it is solely due 

 to polarisation, that ground being the enormous electromotive force, 

 amounting to many Groves, which that explanation obliged us to 

 ascribe to polarisation. When the time of closing is long, the 

 difference of current strengths, so far as the disturbances due to 

 heating etc. allow of its being estimated, amounts to a few per- 

 centages, so that with twenty Groves in the circuit, a polarisation 

 force of less than one Grove would suffice to explain the irreciprocity. 

 If this view of the matter should be confirmed, then irreciprocity 

 .of conduction might be entirely connected with short duration 

 of current. The curve r 2 , 3 in Fig. 22 would then merge into 

 the straight line w 2 , 3 , instead of a line above it and parallel 

 to it. 



It is not necessary to say that the mechanism of irreciprocal 

 conduction in the long direction of the columns remains in com- 

 plete obscurity. We must not enter here, upon what is already 

 known as to irreciprocal conduction in galvanic and induction 

 circuits, all of which is to be found collected in Prof. Christiani's 

 book 1 . We may state as probable that it is not nerves, vessels, or 



1 Beitrage zur Electricitatslehre. Uber irreciproke Leitung electrischer Strome, 

 &c., Berlin, 1^6. 



