LIVING TOBPEDOS IN BERLIN. 485 



on the belly 



0-0056 + b x = 0-003 J 5 

 which gives 



b = o-oon, x = 0-0036. 



When I wished to repeat the experiments last winter, on fish V, 

 killed on December 3ist, I came across a new kind of phenomenon 

 of the organ current, viz. in several organ preparations I found it 

 positive in the dorsal half agreeably to law, but on the other hand, 

 contrary to law, negative in the ventral half of the columns. As 

 the fish had been removed rapidly from the cold tank of the Aqua- 

 rium to warmer water, a disturbance owing to hydrothermal currents 

 at first occurred to me. But I was compelled to give up this idea, 

 because fish VI, which I killed on December ^oth, after I had kept 

 it for more than twenty-four hours in water at 17 -5, exhibited the 

 same phenomena, only with the difference, that this time the dorsal 

 half of the columns acted negatively. I now proceeded to make 

 measurements again, and obtained 



in the curve I (md") 0-0021 Eaoult. 



2 (mD) 0-0012 



3 (#">') + 00067 



4 (mT) + 0-0057 



5 (^i) + 0-0029 



6 (FJD) + 0-0029 



The first thing to be remarked is, that as in the first series and 

 in this, mV< mv', but in this series it happened that mD< md', 

 for now the skin current force is deducted from the organ current 

 force on the back also ; further, that the force in curves 5 and 

 6 by reason of the opposite action of the two halves of the columns, 

 was less than that of the stronger half, even if v l d ought to have 

 been somewhat < , and VD somewhat > 0-0045, instead of both 

 amounting to only 0-0029 without perceptible difference between 

 them. 



In order to determine x from our numbers, we must not proceed 

 as above, where we could assume that 8, which represents the 

 increase of organ current consequent on the lengthening of the 

 column tracts in the transition from md' to mD, and mv f to mV^ on 

 account of the sensibly equal activity of both halves of the columns, 

 is the same at the belly and the back. But if we assume that the 

 one entire half of the columns acted positively, the other entire half 

 negatively, then we can proceed in the following way. 



Let 8 d denote the increase at the back, b v that at the belly, then 



