330 ACTION AT IHE BOl'TOM. 
Thirdly, It remains to consider the eft'ect of 
the air (supposing it to be retained continually 
within the chambers,) at the bottoni of the sea. 
Here, if the position of the moving animal be 
beneath the mouth of the shell, like that of a 
snail as it crawls along the ground, the air within 
the chambers would maintain the shell, buoyant, 
and floating at ease above the body ; and the 
The air within each chamber remains nnder compression, as 
long as the siph uncle contiiruts distended by tlie pericardial 
fluid ; and returning, by its elasticity, to its former state, as soon 
as the pressure of the body is withdrawn from the pericardium, 
cooperates with the muscular coat of the siphuncle, to force the 
fluid back again into the pericardium ; and the shell, thus dimi- 
nished as to its specific gravity, has a tendency to rise. 
The place of the pericardial fluid, therefore, wmU be always in 
the pericardium, excepting when it is forced into and retained in 
the siphuncle by pressure of the body on the pericardial sac, 
during the contraction of the animal within its shell. When the 
arms and body are expanded, either on the surface, or at the bot- 
tom of the sea, the water will have access to the branchial cham- 
bers, and the movements of the heart proceed freely in the dis- 
tended pericardium; which will have great part of its fluid with- 
drawn at those times only, when the body is contracted into the 
shell, and the access of water to the branchiae consequently 
impeded. 
The following experiments shew that the weight of fluid requi- 
site to be added to the shell of a Nautilus, in order to make it 
sink, is about half an ounce. 
I took two perfect shells of a Nautilus Pompilius, each weigh- 
ing about six ounces and a half in air, and measuring about seven 
inches across their largest diameter ; and having stopped the 
Siphuncle with wax, I found that each shell, when placed in fresh 
water, required the weight of a few grains more than an ounce to 
make it sink. As the shell, when attached to the living animal, 
was probably a quarter of an ounce heavier than these dry dead 
