ON LIVING MALAPTERURUS. 



401 



its neutral transverse section in presence of the leading-off cover, 

 but independently of the branch-current through the galvano- 

 meter. The resistance of the conducting wires leading from the 

 poles to the terminal B of the galvanometer was the same, and 

 thus the mean tension between the tensions at the two poles, i.e. 

 tension zero, prevailed at the point B ; consequently, when no 

 current was present in the galvanometer, the tension at the 

 terminal A was also nil, and, according to Bosscha's law, as there 

 was no current in the galvanometer-circuit, its removal would make 

 no difference (see Collected Papers, vol. ii. pp. 83, 84, note). 



I went a step further. I put a pair of saddles, having a constant 

 space between them (2 cm. between the edges of the bands opposite 

 each other) on the 'fish, at different points of its length. I got, as 

 the means of several experiments, the following deflections : 



From the anterior third 

 ,, middle 

 posterior 



26 scale divisions. 

 29 



13 >, >, 



Thus the curve of the tensions, the axis of the fish being taken as 

 the axis of abscissae, rises almost in a straight line from the front 

 backwards as far as the end of the middle third, and then becomes 

 convex towards the axis. It follows also from this, that the equator 

 is shifted towards the anterior end. Whether the somewhat smaller 

 amount of the deflection from the anterior third, in comparison with 

 the amount of the deflection from the middle third, was due to 

 the diminution of the transverse section of the organ towards 

 the anterior end (p. 399), or whether it was accidental, I cannot 

 say. 



I prepared a cover like the former one, except that all the gutta- 

 percha between the three linings was removed except three narrow 

 strips, one on the back and one on each side. In this case, the 

 ratio A : P ought to be greater than with the complete cover. 

 I obtained the following result : 



The experiment, as is seen, was a failure, inasmuch as the ratio 

 j i : 6 recurred with remarkable exactness. Evidently some chance 



Dd 



