imentation of these
elements, the rate depending partly upon differences in specific
gravity, and partly upon the tendency the corpuscles have to run into
clumps. Horse's blood offers one of the best instances of the clumping
of red corpuscles, and in this animal sedimentation of the red
corpuscles is most rapid.

If now such a sedimented blood is allowed to clot the process is found
to start in the middle two layers, i.e. in those containing the white
corpuscles and platelets. From these layers it spreads through the rest
of the liquid, being most retarded, however, in the red corpuscle layer,
and particularly so if the sedimentation has been very complete. Not
only does the clotting process start from the layers containing the
leucocytes and platelets, but in them it also proceeds more quickly.
These observations clearly indicate that the clotting process is
initiated by some change starting from these elements.

The object of the clotting of the blood is quite clear. It is to
prevent, as far as possible, any loss of blood when there is an injury
to an animal's vessels. The shed blood becomes converted into a solid,
and this, extending into the interior of the ruptured vessel, forms a
plug and thus arrests the bleeding. It is found that clotting is
especially accelerated whenever the blood touches a foreign tissue, for
instance, the outer layers of a torn blood-vessel wall, muscle tissue,
&c., i.e. in exactly those conditions in which rapid clotting becomes
of the greatest importance. Yet another very pregnant fact in connexion
with clotting is that if an animal be bled rapidly and the blood
collected in successive samples it is found that those collected last
clot most quickly. Hence the more excessive the haemorrhage in any case,
the greater becomes the onset of the natural cure for the bleeding, viz.
clotting.

When we begin to inquire into the nature of clotting we have to
determine in the first place whence the fibrin is derived. It has long
been known that two chemical substances at least are requisite for its
production. Thus certain fluids are known, e.g. some samples of
hydrocele or pericardial fluid, which will not clot spontaneously, but
will clot rapidly when a small quantity of serum or of an old blood-clot
is added to it. The constituent substance which is present in the
first-named fluids is known as fibrinogen, and that present in the serum
or the clot is known as fibrin-ferment or _thrombin_.

Fi