formed by bacteria.
These toxins may become free in the culture fluid, and the living bacteria
may then be got rid of by filtering the fluid through a filter of unglazed
porcelain, whose pores are sufficiently small to retain them. The passage
of the fluid is readily effected by negative pressure produced by an
ordinary water exhaust-pump. The effects of the filtrate are then tested by
the methods used in pharmacology. In other instances the toxins are
retained to a large extent within the bacteria, and in this case the dead
bacteria are injected as a suspension in fluid. Methods have been
introduced for the purpose of breaking up the bodies of bacteria and
setting free the intracellular toxins. For this purpose Koch ground up
tubercle bacilli in an agate mortar and treated them with distilled water
until practically no deposit remained. Rowland and Macfadyen for the same
purpose introduced the method of grinding the bacilli in liquid air. At
this temperature the bacterial bodies are extremely brittle, and are thus
readily broken up. The study of the nature of toxins requires, of course,
the various methods of organic chemistry. Attempts to obtain them in an
absolutely pure condition have, however, failed in important cases. So that
when a "toxin" is spoken of, a mixture with other organic substances is
usually implied. Or the toxin may be precipitated with other organic
substances, purified to a certain extent by re-solution, re-precipitation,
&c., and desiccated. A "dry toxin" is thus obtained, though still in an
impure condition. Toxic substances have also been separated by
corresponding methods from the bodies of those who have died of certain
diseases, and the action of such substances on animals is in some cases an
important point in the pathology of the disease. Another auxiliary method
has been applied in this department, viz. the separation of organic
substances by filtration under high pressure through a colloid membrane,
gelatine supported in the pores of a porcelain filter being usually
employed. It has been found, for example, that a toxin may pass through
such a filter while an antitoxin may not. The methods of producing immunity
are dealt with below.

[Sidenote: Bacteria as agents of disease.]

The fact that in anthrax, one of the first diseases to be fully studied,
numerous bacilli are present in the blood of infected animals, gave origin
to the idea that the organisms might produce their effect by usin