Researches on the Discharge of the Electric Organ. 19 



Referring to tlie tiiblo we see tliat the value of tlie modal latent 

 period is toleral)l\' constant in spite of the wide changes of the 

 magnitude of the stimulus, while the latent period decreases with 

 the height of the discharge curve. The small increase of the modal 

 latent period may be due to the progressive decrease of tempera- 

 ture, as is given in the data of the experiment. In this case it is 

 a remarkable fact, that o is always positive except in Curve No. 5, 

 and the constancy of the modal latent periods corrected becomes 

 rather worse than that of the observed modal latent periods. The 

 values of .^•o and of <^" vary with the increase of the height of the 

 stimulus and the ^'alues of b" are equal to 5-0 in order of magnitude. 

 Here we ma^^ remark ; that it is very difhcult to determine accurate- 

 ly the value of .iVj, but b" may compensate for the error of Xq in 

 the final result. So we cannot put too nmch weight on the 

 changes of the value of Xq and of ^^ At anyrate we calculated the 

 values of 2/ for every one-thousandth of a secondi from, the values 

 of constants in the table, and plotted these calculated values on the 

 reproduced curves. These are shown in Plate VI. The agreement 

 between the calculated and the observed curves is veiy satisfactoiy 

 except for a small part on both feet of the curves, wehere the calcu- 

 lated points are in general lower than the observed. The discrep- 

 ancy on the descending foot ma}^ be due to the superposed 

 secondary discharge. 



Next Oscillogram No. 37 (Plate XIII.) was examined. This 

 was a series of experiments for in\'estigating the relation between 

 the duration of the stimulating current and the magnitude of the 

 discharge. The durations of the stimuli were very small compared 

 with those of Oscillogram No. 54. The cur\'es on the oscillogram 

 were measured and reproduced in the same way as in No. 54 

 (Plate VIL). 



In this series of experiments, the previous method for the 

 determination of a^o failed to give consistent values to x^ on account 

 of the smallness of ^-^—^v or, in other words, the values of Xo were 

 greater than those of No. 54. After many laborious trials, we at 

 last found that, when we took the origin of the discharge at the 

 beginning of the stimulus, the square root of the product of Xi and 



