ELECTRICAL FISHES 459 



interval between closure of galvanometer circuit and moment of 

 stimulation still finer, by means of the same apparatus. He 

 found, in accordance with his previous observations, that when 

 the closure of the galvanometer was shortened so much that 

 the induction -current alone could affect the galvanometer 

 (#,-#3 = 0" -0-0025") there was no difference in the deflec- 

 tions produced by the homo- and heterodromous currents ; at the 

 next moment the electromotive action of the organ-preparation 

 discharged by the shock makes its appearance (S 2 S 3 = 

 0'0025 // O'OOS") as the homodromous after - current, and 

 increases rapidly with further augmentation of closure. The 

 apparent irreciprocal conduction thus arises only when the action 

 upon the galvanometer of the exciting induction-shock is com- 

 bined with that discharged from the organ-preparation (du Bois- 

 Eeymond's positive polarisation). 



In further confirmation of this view we have the results of 

 experiments on the influence of varying temperature upon the con- 

 sequences of direct excitation of organ-preparations. Gotch was able 

 to show easily, by means of his spring rheotome, that the intensity, 

 and more particularly the time -distribution of the excitation 

 effects, on applying single homo- and heterodromous induction- 

 currents, are essentially influenced by temperature, and that in 

 the direction we should a priori expect in the matter of dis- 

 charging an excitation. As appears from comparison with the 

 curves in Fig. 281, which give a graphic representation of the 

 experimental results (the ordinates corresponding with the gal- 

 vanometer deflections, the abscissae with the interval after the 

 moment of excitation, at 0), the intrinsically less important 

 effects of excitation in the cooled preparation are considerably 

 retarded, and only become appreciable on the galvanometer long 

 after the induction-current has been made, so that there is a 

 long " latent period " during which the closure of the galvano- 

 meter circuit has no effect. It must thus be relatively easy to 

 separate the effect of the induction-current upon the galvano- 

 meter from that of the excitation (supposed positive polarisation), 

 and hence to determine whether irreciprocity of conduction 

 really exists or no. Gotch succeeded in demonstrating that 

 apparent irreciprocity does appear at higher temperature (22 C.) 

 with the same closure of the galvanometer, while the same pre- 

 paration cooled, under otherwise uniform conditions, conducts 



