174 NATURAL SCIENCE. March, 



It seemed difficult to believe that the pumping action of the 

 posterior part of the fins could cause a negative pressure beneath the 

 body, because the respiratory movement was going on while the fish 

 adhered to the vertical surface. This respiratory movement forces a 

 current of water into the space below the body through the lower gill 

 cavity, and this would necessarily produce a positive pressure. How- 

 ever, the currents of water were ascertained by watching the course 

 taken by particles of suspended matter in the water, and afterwards 

 by dropping carmine with a pipette at points in the neighbourhood of 

 the head. The carmine was seen to pass in at the mouth and out at 

 both gill-openings, and no difference was observed in the respiratory 

 currents from those of an ordinary fish. The perforation of the gill- 

 septum was therefore not found to play any special part. But the 

 carmine was also drawn into the space below the body through con- 

 siderable passages above and below the jaws. This proves that more 

 water is pumped by the action of the posterior parts of the fins than 

 passes through the mouth and gill-openings. This, therefore, must 

 cause a pressure on the outside of the fish against the vertical surface 

 on which it rests. In order to test whether a pressure caused in this 

 way would be sufficient to produce the adhesion, I endeavoured to con- 

 struct a simple apparatus which would reproduce the conditions 

 observed in the fish. I took a rectangular piece of sheet india-rubber, 

 6 inches long by 4 inches broad, and at each end of its longer axis 1 

 fastened, with needle and cotton, a short piece of glass tube. The tubes 

 did not meet in the middle of the piece of india-rubber. A siphon was 

 placed on the edge of the glass front of a tank, and its end in the tank 

 connected by a long india-rubber tube to one of the tubes on the 

 india-rubber flap. When the flap was placed against the glass with 

 the glass tubes towards the latter, a current of water was drawn from 

 beneath the flap by means of the siphon. The longer edges of the 

 flap became pressed against the glass, while the entering current 

 passed through and at the sides of the glass tube which was not 

 attached to the siphon. In this apparatus the drawing of the siphon 

 represented the pumping action of the fins in the fish, while the piece 

 of glass tube in front represented the mouth and gill-opening of the 

 fish, the water being able to pass at the sides of the tube as it passes 

 at the edges of the jaws and head in the fish. The only differences 

 are that the respiratory movement in the fish is absent in the 

 apparatus, and that the tube representing the mouth communicates 

 only with the space beneath the rubber flap, not with the outer side of 

 the flap also, as in the fish. These differences are unimportant, because 

 it is certain that the respiratory movement in the fish does not cause 

 the adhesion, and there is in the fish a current caused by the pumping 

 action of the fins over and above the ordinary respiratory current. It 

 was found that the flap just described adhered to the vertical surface 

 of the glass front of the tank so long as the siphon was running, and 

 fell off as soon as the flow through the siphon was stopped. 



