cause water to flow into this space at the anterior edges of the head. The
subcaudal flaps were perfectly motionless and tightly pressed between the
base of the tail and the surface of support, so that any movement of them
was impossible. The question arose, however, whether the tail and these
flaps acted as a sucker which aided in the adhesion. The flaps were
therefore cut off with scissors--an operation which caused practically no
pain or injury to the fish--and it adhered afterwards quite as well as
when the fin-flaps were intact. The subcaudal prolongations of the fins
are therefore not necessary to the adhesion, nor to the pumping action, of
the muscles and fins, which went on as before. It seemed probable,
therefore, that the pumping action was itself the cause of the adhesion.
But the difficulty in accepting this conclusion was that there was a
distinct though gentle respiratory movement of the jaws and opercula; and
if the pumping of the water from beneath the body caused a negative
pressure there, and a positive pressure on the outer side of the body, it
seemed equally certain that the respiratory movement must force water into
the space beneath the body and so cause a positive pressure there which
would tend to force the fish away from the surface with which it was in
contact. Examination of the currents of water around the edges of the
fish, by means of suspended carmine, showed that water passed in at the
mouth and out at the lower respiratory orifice, but also into the space
below the body at the upper and lower edges of the head, without passing
through the respiratory channel. It was thus proved that the rate at which
water was pumped out at the sides of the tail was greater than that at
which it passed in by the respiratory movements, and consequently there a
resultant negative pressure beneath the body. By means of a model made of
a thin flexible sheet of rubber, at each end of which on one side was
fastened a short piece of glass tube, I was able to imitate the physical
action observed in the fish. A long piece of rubber tube was attached to
one of the pieces of glass tube, and brought over the edge of the glass