ions where they do no harm.
But if the pressure is relieved gradually they are not formed, because the
gas comes out of solution slowly and is got rid of by the heart and lungs.
Paul Bert exposed 24 dogs to pressure of 7-91/2 atmospheres and
"decompressed" them rapidly in 1-4 minutes. The result was that 21 died,
while only one showed no symptoms. In one of his cases, in which the
apparatus burst while at a pressure of 91/2 atmospheres, death was
instantaneous and the body was enormously distended, with the right heart
full of gas. [v.04 p.0959] But he also found that dogs exposed, for
moderate periods, to similar pressures suffered no ill effects provided
that the pressure was relieved gradually, in 1-11/2 hours; and his results
have been confirmed by subsequent investigators. To prevent caisson
disease, therefore, the decompression should be slow; Leonard Hill suggests
it should be at a rate of not less than 20 minutes for each atmosphere of
pressure. Good ventilation of the caisson is also of great importance
(though experiment does not entirely confirm the view that the presence of
carbonic acid to an amount exceeding 1 or 11/4 parts per thousand exercises a
specific influence on the production of compressed air illness), and long
shifts should be avoided, because by fatigue the circulatory and
respiratory organs are rendered less able to eliminate the absorbed gas.
Another reason against long shifts, especially at high pressures, is that a
high partial pressure of oxygen acts as a general protoplasmic poison. This
circumstance also sets a limit to the pressures that can possibly be used
in caissons and therefore to the depths at which they can be worked, though
there is reason to think that the maximum pressure (43/4 atmospheres) so far
used in caisson work might be considerably exceeded with safety, provided
that proper precautions were observed in regard to slow decompression, the
physique of the workmen, and the hours of labour. As to the remedy for the
symptoms after they have appeared, satisfactory results have been obtained
by replacing the sufferers in a compressed air chamber ("recompression"),
when the gas is again dissolved by the body fluids, and then slowly
"decompressing" them.

See Paul Bert, _La Pression barometrique_ (1878); and Leonard Hill, _Recent
Advances in Physiology and Biochemistry_ (1906), (both these works contain
bibliographies); also a lecture by Leonard Hill delivered at the Royal
Instit