IMMUNITY

AND

SPECIFIC THERAPY

BY

W. D'ESTE EMERY, M.D., B.Sc. LOND.

CLINICAL PATHOLOGIST TO KING'S COLLEGE HOSPITAL AND PATHOLOGIST TO THE CHILDREN'S
* HOSPITAL, PADDINGTON GREEN ; FORMERLY ASSISTANT BACTERIOLOGIST TO THE ROYAL
COLLEGES OF PHYSICIANS AND SURGEONS, AND SOMETIME LECTURER ON PATHOLOGY
AND BACTERIOLOGY IN THE UNIVERSITY OF BIRMINGHAM

WITH ILLUSTRATIONS

OF THE

UNIVERSITY

OF

PAUL B. HOEBER

69, EAST 59T.H STREET

NEW YORK

1909

PRINTED IN ENGLAND

PREFACE

IN writing this book I have attempted to give a connected and
symmetrical outline of the chief facts definitely known with regard
to the method in which the body protects itself against infections,
and of their applications in the diagnosis, prevention, and treat-
ment of disease. It is not written in support of the views of any
particular school of thought, and, when dealing with subjects still
under discussion, I have tried to give a fair and impartial, though
necessarily succinct, account of each of the rival theories. The
factors in many of the problems of immunity are so complex,
and our knowledge of the subject grows and alters so rapidly,
that it is quite impossible to deal with it dogmatically at the
present time. I have kept in view, as far as possible, the re-
quirements of the physician and surgeon who may require an
epitome of the theoretical basis of the modern methods of diagnosis
and treatment, now assuming so much importance, and of the
student who desires a general survey of the subject before com-
mencing more advanced studies.

My best thanks are due to Mr. H. K. Lewis for the ready and
courteous way in which he has acceded to all my suggestions and
requirements ; to Drs. Whitfield and Briscoe, from whom I have
received some valuable suggestions; and to Professor Herbert
Jackson, of King's College, for kindly reading the sections dealing
with the more purely chemical and physical questions and for
much useful information connected therewith. I have also to
thank Sir Almroth Wright and Drs. R. W. Allen, Eyre, and
Bolduan; Messrs. Macmillan and Co., Kegan Paul and Co., and
the proprietors of the Lavcet, British Medical Journal, and the St.
Bartholomew's Hospital Journal for permission to use illustrations
from their publications.

CONTENTS

CHAPTER PACE

GLOSSARY . . ' . . ix

I. INTRODUCTORY AND GENERAL . ... I

II. ON THE NATURE OF TOXINS . . . -37

III. THE PHENOMENA OF ANTITOXIN FORMATION . . 60

IV. INTERREACTIONS OF TOXIN AND ANTITOXIN . . 69

V. THE ORIGIN OF ANTITOXIN— THE SIDE-CHAIN THEORY. Q2

VI. IMMUNITY TO TOXINS . . . . • IT5

VII. BACTERIOLYSIS AND ALLIED PHENOMENA . . 139

VIII. THE AGGLUTININS . . . . 204

IX. THE PRECIPITINS .... 226

X. PHAGOCYTOSIS . . . . -. . 238

xi. "REACTIONS" AND SIMILAR PHENOMENA . . 300

XII. COLLOIDAL THEORY OF ANTIBODIES . . . 319

XIII. ON IMMUNITY TO BACTERIA .... 33!

XIV. PRACTICAL APPLICATIONS .... 358

BIBLIOGRAPHY . . . . . .421

LIST OF AUTHORITIES . . . . . 439

INDEX . . . . . . . 443

ERRATA

Page 9, line 5 from bottom, omit " to " after " -cytes."
14, line 13, for " rather of " read " than with."
36, line 22, for " of read "to."
,, 45, line 29, for " antitoxin " read " toxin."

48, line 14. for "became" read "become"; line 25, for " united

with" read "injured."

,, 52, line 10. for " supernatural " read " supernatant. "
,, 55, line 4 from bottom, for " properties " read " effects."

57, line i, for "is" read " are " ; bottom line, insert " upon " after

" toxins," and for " defends " read " depends."
,, 70. line 6. for " haemoglobin " read " haemolysin."
,, 78, lines 17 and 19, for " c.c. " read "parts"; and line 18, for

" 16-6 c.c. " read " i6'6 parts."
79, top line, for "antitoxin " read " toxin."
90, line 22, for " toxin " read " antitoxin."
,, 91, line 18, for " toxic " read " neutral" ; line 26, omit " as. "

93, line 22, for "injection" read "infection."
,, 106, line 26, for "it" read " the toxin "

,, 107, line 21, for " to tetanus antitoxin " read "of tetanus toxin."
,, no, line 9, for " antitoxin" read " toxin."
122, line 3, for " toxin " read " antitoxin."

128, line 20, for " leucocytes " read " bacteria ingested."

129, line 3 from bottom, for " toxin" read " antitoxin."
152, line 7, for "joined " read " formed."

171, line 10, for " rabbit" read " goat."

192, line 30. for " nephrotoxin " read " hepatotoxin."

220, line 7 from bottom, for " they " read "it."

252, line 12, for "leucocytes " read "bacteria."

256, line 15, for "complement" read "amboceptor."

284, line 13, for " opsonix " read " opsonic."

287, line 8, for "which in" read "in which."

290, line 19, for " bacteria can " read " leucocytes can '

296, line 8, for " bacteria " read " leucocytes."

307, line 12. for " local " read •' general."

310, line 6 from bottom, for " y " read "a."

323, line i, for "or" read "on"; line 28, for " complement " read

" amboceptor."
3-15 line 13, should reaJ, "those which hava no defensive layer, or

which have numerous receptors " etc.
348, line 14, for " bacteria" read " leucocytes."
351, line 14, for "installations " read "instillations."
355, last line, for "research " read "defence."
360, line 3, for " able " read •' unable."
365, line 29. for '" -stable ' ' read " labile. ' '
378, line 8 from bottom, read "are accompanied by but the slightest,"

etc.

386, line 21. for 'benefits " read '"benefit."
407, line 10 from bottom, for '• rise " read <;use."
419, line 25, for "heated " read "beaten."

GLOSSARY

Active immunity. Immunity due to an active struggle against some infective
material, vaccine, or toxin.

Addiment. See Alexin or Complement.

Agglutinin. A specific antibody which brings about agglutination — i.e.,
causes the bacteria, cells, etc., for which it is specific, to collect into
clumps. Non-specific substances (acids, etc.) have a similar action, but
are not properly termed agglutinins.

Agglutinogen. The antigen of agglutinin— i.e., the substance which, when
injected into a suitable animal, leads to the formation of agglutinin.

Agglutinoid, A modification of agglutinin which has retained the power of
uniting with the specific bacteria, etc., but has lost that of causing them
to clump.

Aggressin (aggredior, I attack). A substance secreted by bacteria and possess-
ing the power of inhibiting phagocytosis of the organism producing it.

Alexin (dX&jw, I ward off). A defensive substance having an injurious effect
on bacteria, and occurring in the serum of normal and immune animals.
It is analogous in many respects to the bacterial toxins, and, like them,
easily destroyed by heat, chemical agents, etc. It is probably identical
with complement, q.v. (For other synonyms, see p 143.)

Amboceptor (ambo, both, and capio, I take). A specific antibody produced
by the injection of bacteria, red corpuscles, cells, etc., and exerting, with
the help of alexin or complement, a solvent action on these substances.
The term is Ehrlich's, and its use should involve the acceptance of his
theory of its action. (For synonyms, see p. 142.)

Anaphylaxis (a or dm, privative, and 0uXa<r(rw, 1 guard). The opposite of
prophylaxis — i.e., a condition in which the susceptibility of the animal
(especially to toxins, serums, etc.) is abnormally increased. Practically
identical with hypersensitiveness.

Antibody. A substance formed by the injection into an animal of a substance
(its antigen) not normally found in the juices of that animal (and probably
in all cases of proteid constitution, or closely allied thereto), which unites
with its antigen and modifies it in some way.

Antiferment, or antienzyme. An antibody to a ferment or enzyme.

Antigen (dvTi, against, and yiyvu. I produce). A substance which, when
injected into a suitable animal, has the power of leading to the produc-
tion of an antibody. In most, if not in all, cases it is proteid in nature.

Antitoxin. An antibody to a toxin — i.e., a specific substance formed by an
animal in consequence of the presence in its tissues or juices of a given
toxin, which the antitoxin thus produced has the power of neutralizing.

Arthus' phenomenon. A form of hypersensitiveness to serum, where in a
sensitized animal local lesions (gargrene, abscesses, etc.) develop in a
region where serum is injected, and the animal may become cachectic
and die.

Atrepsy (d, privative, and rptyw, I nourish). A condition in which an infective
agent or collection of malignant cells dies in the animal body owing to
its being unable to obtain suitable nourishment : a conception intro-
duced by Ehrlich to explain certain forms of immunity — e.g., to malignant
growths.

Attenuation. The change which an organism undergoes whereby it becomes
less virulent.

X GLOSSARY

Autohsemolysin. A substance which has the power of dissolving the red cor-
puscles of the individual which produces it ; an autochthonous ambo-
ceptor or immune body to an animal's own red corpuscles.

Bacteriolysis. The phenomenon of solution of bacteria, more especially by
the action of specific antibodies, aided by alexin or complement.

Bacteriotropin, or bacteriotropic substance (Tptiru, I turn). A substance
(usually a specific antibody) which has the property of uniting with
bacteria, and in someway altering their properties, usually rendering them
more suitable for phagocytosis.

Bordet's phenomenon. — The absorption of alexin or complement which is
brought about by means of a cell or bacterium combined with amboceptor
or immune body, apart altogether from the alexin which is necessary for
the action ; the complete removal of all alexin from a fluid by means of
a cell-immune-body compound.

Chemotaxis (rdfa, an arrangement). The attraction or repulsion of leucocytes,
bacteria, etc., by substances in solution in the fluid containing the cells
in question.

Complement (compleo, I fill up). A synonym for alexin, q.v. The term was
introduced by Ehrlich, since on his theory this substance unites with the
amboceptor which has already united with the bacterium, etc., and thus
completes the conditions necessary for solution.

Complementoid. A modification of complement which possesses the combin-
ing powers of that substance, but which has lost its active solvent
properties.

Cytase. Metchnikoffs term for the digestive (proteolytic) enzyme secreted
by leucocytes ; a synonym for alexin or complement.

Cytolysin. An antibody (immune body or amboceptor) to a cell, having the
power of sensitizing that cell so that it is completely or partially dissolved
on the addition of alexin.

Cytolysis. The solution — usually partial — of cells by means of an antibody
(immune body or amboceptor) and alexin.

Cytotoxin. A substance acting as a cellular toxin, especially an antibody and
alexin ; practically identical with a cytolysin.

Danysz effect. The decrease in the neutralizing effect of antitoxin which is
manifested when the toxin is added in portions, with an interval between
each, rather than all at once.

Desmon (5e<r/<i6s, a bond). A synonym for immune body or amboceptor.

Deviation of complement. The phenomenon in which the solvent effect of
immune body on cells or bacteria (in presence of alexin) diminishes as an
excess of the antibody is added ; also known as the Neisser-Wechsberg
phenomenon.

Dominant complement. Where (on Ehrlich's theory) two or more different
complements unite with a complex molecule of amboceptor which has
united with a bacterium, corpuscle, or cell, that which has the more
potent action is known as the dominant. Its effect may be produced
without the action of the other complements, or the necessary amount
may be smaller.

Ehrlich's phenomenon. The fact that the difference between the amount of
toxin exactly neutralized by one unit of antitoxin and the amount which
(added to one unit of antitoxin) just leaves one lethal unit free is greater
than one lethal dose of simple toxin.

Endocomplement. A complement contained within a red corpuscle, probably
in all cases lecithin or an allied substance.

Endotoxin. A bacterial toxin contained within the substance of the bticterium,
and not liberated except when the cell is destroyed.

Ergophore group (tpyov, work, and 0^/>ar, I bear). The part of a molecule of
antigen or antibody on which the specific properties of the substance
depends (toxophore, zymophore, agglutinophore, etc.), in distinction from
the haptophore or combining part of the molecule.

GLOSSARY XI

Exotoxin. A soluble bacterial toxin which is excreted by the bacterium into
the surrounding fluid during the life of the organism.

Fixation of complement. A synonym for Bordet's phenomenon, q.v.
Fixator. A synonym for immune body or amboceptor.

Gastrotoxin. A cytotoxin or cytolysin acting on the cells of the mucous

membrane of the stomach.
Gengou's reaction. The removal of all alexin or complement from a fluid by

means of a compound of a precipitin and its antigen ; analogous with

Bordet's phenomenon, except that in this case the reacting antigen is a

soluble substance. The two are often grouped together as the Bordet-

Gengou reaction.
Group reaction. A reaction with an antibody (usually an agglutinin) which

is common to several species of bacteria, forming a well-defined group —

e.g. , the coli group, or the pasteurelloses.

Haemagglutinin. A substance which agglutinates red corpuscles.

Hsemolysin. A substance which dissolves red blood-corpuscles, or at least
releases the haemoglobin which they contain. The term is used mainly
for an antibody having, in conjunction with alexin, a solvent action of
this nature.

Haptin. A portion of a molecule of protoplasm having combining affinities
for food molecules, and forming an antibody when shed (v. Receptor).

Haptophore group, or Radicle (CCTTTW, I fasten). That portion of a substance
(whether antigen or antibody) which has the power of entering into com-
bination with its appropriate antibody or antigen, as the case may be.
Thus a molecule of toxin is supposed to contain a group of atoms which
can combine with a cell or molecule of antitoxin, and a second which can
then exert a toxic action. The former is known as the haptophore
group.

Immune body (immunis, exempt from public service). A specific antibody,
produced by the injection of bacteria or other cells, and having the
power of altering these substances in such a way as to render them com-
pletely or partially soluble on the addition of alexin. It is the same as
amboceptor, but the term implies no theory and is generally preferable.
Synonyms : substance sensibilatrice, desmon, preparator, copula, etc.

Incitor constituent of serum. A substance which aids phagocytosis, espe-
cially thermostable opsonin.

Isoagglutinins. An agglutinin which, ' occurring in the serum of a certain
animal, will agglutinate the red corpuscles of other animals of that
species, but not those of the individual which produces it.

Isohsemolysin. An immune body or amboceptor which, occurring in the
serum of a certain animal, dissolves (in conjunction with alexin) the red
corpuscles of other animals of that species, but not those of the individual
which produces it.

Koch's phenomenon. The tuberculin reaction, or rise of temperature and
sudden exacerbation of the local lesions occurring in a tuberculous animal
after injection of a culture of tubercle bacilli, living or dead, tuberculin,
or other specific tuberculous material.

Lactoserum. A serum containing a precipitin for milk proteids.

Leucotoxin. An antibody (immune body or amboceptor) which, in conjunc-
tion with alexin, exerts a toxic influence on leucocytes.

Lysis (Xims, a loosening). The solution of cells, bacteria, etc., mostly by
means of antibodies or other protective substances.

LO dose of toxin. The amount which is exactly neutralized by one unit of
antitoxin.

L+ dose of toxin. The amount which, added to one unit of antitoxin, behaves
just like one lethal dose of toxin, bringing about a fatal result in test
animals within the time-limit fixed. The fact that the L_|_ dose -the L0
dose is greater than one lethal dose constitutes the Ehrlich phenomenon.

Xll GLOSSARY

Macrocytase. In Metchnikoff's phraseology, the digestive enzyme secreted by
the large mononuclear leucocytes, and having a special action on cells
rather than on bacteria ; really a synonym for alexin, especially for one
acting on cells or red corpuscles.

Macrophage. Metchnikoff's term for a large phagocyte which, according to
him, is especially adapted to the ingestion of cells or corpuscles rather
than of bacteria They may be large lymphocytes, large hyaline cells,
endothelial or other tissue cells.

Microcytase. The digestive enzyme of Metchnikoff's microcytes or poly-
nuclear leucocytes ; supposed to have a special action on bacteria.
Practically identical with alexin.

Microphage. A small leucocyte supposed by Metchnikoff to be specially
active against bacteria, and to have little or no phagocytic action on cells
or corpuscles. They are polynuclear leucocytes.

Negative phase. The sudden diminution in the amount of an antibody (and
possibly of other defensive substances) in the blood which follows
immediately on the injection of an antigen.

Neisser-Wechsberg phenomenon. Deviation of the complement, q.v.

Nephrotoxin. A cytotoxin specific for renal cells.

-ogen. A suffix usually employed to denote an antigen in relation to its anti-
body— e.g., agglutinogen, the substance which on injection into an
animal leads to the production of agglutinin. Also used for a preliminary
non-active form of an active substance — e.g., opsoninogen, a substance
which under certain conditions becomes opsonin.

-oid (eldos, a figure or appearance). A suffix denoting a secondary modifica-
tion of an active substance in which it appears to retain its power of
entering into combination with its antibody or antigen, but has lost its
specific activity ; a molecule of antigen or antibody which has lost its
ergophore, but retained its toxophore, group— e.g., complementoid or
toxoid, q.v.

Opsonin (opsono — I cater for, I prepare for food. Derived from &\}/ov, cooked
meat, a sauce or relish). A substance or combination of substances of
whatever nature which has the power of combining with a bacterium,
cell, or other substance, and rendering it more easily ingested by a
leucocyte or other phagocyte.

Passive immunity. Immunity due to the injection of serum from an animal
which has acquired immunity to a toxin or infective agent.

Pfeiffer's phenomenon. The classical Pfeiffer's phenomenon consists in the
globular transformation, loss of staining reaction, and finally complete
disappearance of cholera vibrios, when introduced into the peritoneal
cavity of an immunized guinea-pig, or into that of a normal one if
immune serum be also injected. Also applied to the similar, but usually
less complete, destruction of other bacteria under similar conditions, or
to bacteriolysis in general.

Phytotoxin. A poisonous substance formed by one of the higher plants, but
otherwise closely resembling a bacterial toxin, more especially in its
power to give rise to the production of an antitoxin on injection — e.g.,
ricin, abrin.

Polyceptor. Amboceptor which possesses several haptophore groups capable
of anchoring several molecules of different sorts of complement, the
most important of which is termed the dominant (Ehrlich).

Polyvalent serum. A serum containing antibodies against several strains of
the same species of bacteria— e.g., streptococci.

Polyvalent vaccine. A vaccine composed of the dead bodies of several strains
of the same bacterial species. A vaccine composed of more than one
species of organism is termed a mixed vaccine.

Positive phase. The period during which the amount of antibody or other
protective body in the serum is increased owing to the injection of an
antigen. In general terms it corresponds to the period of exalted im-

GLOSSARY Xlll

munity due to vaccination, injection of toxin, etc., and is very variable
in duration.

Precipitin. An antibody to a soluble form of proteid, having the power of
precipitating or coagulating that proteid by a process of clumping its
molecules.

Precipitogen. The antigen to a given precipitin. Thus when a serum is
injected into an animal numberless substances are introduced, a certain
number of which only give rise to the formation of precipitin, and are
called precipitogens. Also called precipitable substance.

Precipitogenoid. Heated precipitable substance, which has retained its
power of combining with precipitin, but no longer forms a precipitate
after doing so.

Precipitoid. Precipitin which has lost its active or ergophore, but retained its
combining or haptophore, group ; the latter has also increased in
affinity for precipitable substance. The name is also applied to pre-
cipitogenoid.

Predisposition. The opposite of immunity ; the state of an animal, in virtue
of which it is readily infected with a given agent.

Preparator. Metchnikorf's term for immune body or amboceptor.

Prophylaxis. Any process by which the vulnerability of an animal by an
infective agent or toxin is diminished or removed ; a process for the
induction of immunity, more especially in its practical application to the
prevention of disease.

Prostatotoxin. Axcytolysin for the cells of the prostate.

Pro-zone. In constructing a curve indicating the action of an antibody at
different dilutions, it sometimes happens that stronger solutions have
less effect than more dilute ones. The region of the curve in which this
inhibition of the action is brought about by an excess of the active sub-
stance is termed the pro-zone. It occurs with substances other than
antibodies. Also called zone of inhibition.

Receptor. In Ehrlich's side-chain theory a part of a living molecule of pro-
toplasm which has the power of attracting and combining with a molecule
of food proteid (or of toxin, etc.) from the fluid with which it is bathed,
and of building it up into the whole molecule, and thus utilizing it as
nourishment, to aid which process it may also seize one or more
molecules of complement. When shed into the blood these receptors
constitute antibodies.

1. Simple (e.g., those constituting antitoxin). In the antibodies formed
by this group we can only distinguish one group o'f atoms — a haptophore
group having the power of combining with the specific antigen (e.g.,
toxin), and preventing its subsequent union with a living cell, thus render-
ing it inert.

2. Complex (e.g., agglutinin), in which we can recognize two separate
properties, presumably situate in different groups of atoms : (a) a hapto-
phore, combining group, as above ; and (b) an ergophore group, on
which the activity depends, and which may be destroyed whilst (a)
remains intact.

3. Compound (e.g., amboceptor, on Ehrlich's theory). In them there
are t-uv or more haptophore groups, one of which combines with the
antigen, the others with one or more molecules of complement.

Sensitization of bacteria, corpuscles, etc. The addition of immune body, so
that the objects are prepared or sensitized to the action of alexin.

Side-chain theory. The theory (Ehrlich's) which accounts for the develop-
ment of antibodies by supposing that the receptors (q.v.) which combine
with the specific antigen may, under certain circumstances, be produced
in excess and cast off into the surrounding fluid ; these receptors, retain-
ing their power of combining with antigen, constitute the antibodies in
question. A brilliant conception, which has been the cause of enormous
advance in our knowledge of problems connected with immunity.

XIV GLOSSARY

Smith's (Theobald) phenomenon. The acquisition of hypersensitiveness to
serum and other proteid substances (normally inert) which occurs in
some animals as a result of minute doses of these substances, and leads
to rapid death, with acute symptoms, when a second injection is given.

Specificity (species, an image). A direct relation of cause and effect between
two substances (such as diphtheria toxin and its antitoxin, the latter being
only produced by, and acting only on, the former), or between a substance
and a phenomenon (such as the tuberculin reaction, produced only by
tuberculous products in a tuberculous animal). The specific products of
a micro-organism are those produced only by that organism, so that their
recognition is proof of its presence. In the same way a specific disease
is one produced only by a certain bacterium (such as diphtheria or
anthrax), and not by several organisms (such as suppuration or actino-
mycosis) .

Spermotoxin. A cytolysin to spermatozoa.

Stimulin. A substance having the power of stimulating the action of the
leucocytes (more especially in regard to phagocytosis) by a direct action
on the leucocyte itself. The existence of these substances is doubtful,
most of the phenomena supposed to be caused by them being due (a) to
the action of opsonins, and (b) to substances which have a positive
chemotactic action, attracting leucocytes to the region.

Syncytiotoxin. A cytolysin acting on the cells of the placenta.

Thermolabile. Easily destroyed by heat. In general thermolabile substances
are destroyed, completely or partially, by an exposure to 55° C. for half
an hour or to 60° C. for 10 minutes.

Thermostable. The opposite to thermolabile, q.v.

Thyrotoxin. A cytolysin acting on the cells of the thyroid gland.

Toxin (ro&Kbv <j>dpiu.a.Koi> , the drug with which poisoned arrows were anointed.
TO^OV, a bow). The specific poison on which the pathogenic activity of a
micro-organism depends. The fact of its being specific excludes simple
chemical substances which may also exert a toxic action.

Toxoid. A secondary modification of a toxin which has lost its power of
producing toxic symptoms, but retained that of combining with antitoxin
or susceptible cells ; or, in Ehrlich's terminology, one that has lost its
toxophore, but retained its haptophore, group.

Toxone. A specific substance of feeble toxicity and slight affinity for anti-
toxin which is supposed to be produced by certain bacteria, notably that
of diphtheria, in which case it is believed to be the cause of paralysis.
Unlike toxoid, it is a primary product. Its existence is doubted, and the
effects attributed to minute amounts of toxin by some authors.

Toxophore group. The portion of a molecule of toxin on which the toxic
activity depends, the destruction of which converts the molecule into one
of toxoid.

Trichotoxin. A specific cytotoxin for ciliated epithelium.

Vaccination. The production of active immunity by some process less
severe than the induction of an ordinary attack of the disease in
question.

Vaccine. A substance (usually a dead culture or living culture of mitigated
virulence) the injection of which leads to the production of active im-
munity with less risk than that which accompanies an ordinary attack of
the disease.

Virulence (virus, a poison). The property or properties of an organism in
virtue of which it is able to give rise to disease in animals or to produce
a powerful toxin.

Zootoxin. A poisonous substance of animal origin which resembles in other
respects (and especially in that it can give rise to the production of an
antitoxin) the bacterial toxins — e.g., snake venom, eel serum.

Zymophore group (&M, leaven). The portion of an enzyme or enzyme-like
substance on which the specific properties depend, in contradistinction to
the combining or haptophore portion.

OF THE

UNIVERSITY

OF

>R^

IMMUNITY AND SPECIFIC
THERAPY

CHAPTER I
INTRODUCTORY AND GENERAL

IMMUNITY is the power which certain living organisms possess of
resisting influences which are deleterious to others. In its widest
form it includes the power of resisting poisons, adverse physical
influences, and diseases of all kinds. Thus, many men can and
do acquire some degree of immunity against nicotine, alcohol, and
other poisons ; some bacteria are immune to temperatures which
are quickly fatal to others ; and some individuals and races have
a very real immunity to gout and other metabolic diseases to
which their less fortunate brethren are more prone. In any
complete discussion of the subject these forms of immunity would
require some consideration, but in what follows we shall, in the
main, limit ourselves to the investigation of immunity against
the diseases of bacterial origin. In doing so we must not be
thought to consider the other diseases — metabolic and what not
— as being unimportant. The very reverse is the case, and the
subject which calls most urgently for research at the present day
is the nature and mechanism of immunity against malignant
tumours, and of this we have recently acquired a little know-
ledge. But the diseases other than those of bacterial origin will
not be dealt with, for the simple reason that our knowledge of
their intimate causes is still unknown, and until they are dis-
covered, and until the physiological disturbances of the economy
which occur in these diseases are more fully known, the nature
of the corresponding immunity is obviously extremely difficult
of study. The bacterial diseases are quite different, for here

2 INTRODUCTION

the causes are fully known ; the diseases themselves can be
reproduced (in most cases) at pleasure, and the physiological dis-
turbances which take place are fairly well investigated. There
are, of course, gaps, and those not inconsiderable, in our know-
ledge ; but, on the whole, the nature of these diseases is nearly as
well ascertained as the present state of normal physiology will
allow. Further, we can not only reproduce the diseases, but we
can reproduce in most cases any degree of immunity to them
which we may require for purposes of protection or research, and
we can investigate the differences between the cells and fluids of
the immunized person or animal and the corresponding parts of a
normal organism, and we can attempt to correlate them with the
production of the immune state. We have, therefore, a very
large amount of information on the subject, and although this
information is at present incomplete, we have already obtained
results of the highest practical and theoretical importance ; and
the value of these results leads us to believe with confidence that
our methods are right, that we are on the right track, and that a
solution of the problems that have at present baffled research will
come in the near future.

As denned above, immunity is a function of all living material,
and one of the highest importance. Biologists have compiled
lists of the essential properties of living protoplasm — nutrition,
reproduction, and the like — but have not realized that immunity
to bacterial action is the first necessity for continued life.
Consider for a moment a small water animal — say a hydra —
occurring in water which naturally contains saprophytic bacteria.
Whilst the animal lives these organisms do not affect its proto-
plasm in any way, the latter being immune to their action ; but
on the animal's death rapid putrefaction occurs, and in a few
hours its protoplasm is broken down by bacterial action: the
immunity has ceased. Immunity to putrefactive bacteria is
therefore a condition of life in the lower animals. But the same
is true in every respect for those of a higher grade, man included.
From the moment of birth we are surrounded with air containing
bacteria which are not pathogenic in the ordinary sense, but
which only fail to be so because of the inherent power of immunity
to saprophytic bacteria, which is a fundamental property of all
living material. Apart from this, the organisms present in the
air, alimentary canal, skin, etc., would flourish as rapidly as they
do in a corpse, and life would only be possible for a few hours,

INTRODUCTORY AND GENERAL 3

or perhaps minutes. Readers of one of Mr. Wells's ingenious
romances may perhaps remember how the strange beasts from
Mars which invaded this planet died rapidly, being evolved in a
region in which there were no bacteria, and in which this power
of resisting their action had not been developed. The example is
a striking one, and is strictly scientific, though we may wonder
how the rotation of nitrogen, in which bacteria play so essential a
part, is brought about in Mars ; for this process of the breaking
down of dead proteids by bacterial action, and the preparation of
its nitrogen for use in plants, is essential for continued life on the
planet. Without decomposition all the combined nitrogen of the
world would soon become locked up in the dead bodies of animals ;
plants would starve and die, and animals (which are all dependent,
directly or indirectly, on plant nitrogen) would likewise become ex-
tinct. It is a most marvellous natural phenomenon that these putre-
factive bacteria should be found wherever life occurs, and wherever
their aid may be required to deal with the protoplasm when dead,
and that this same protoplasm should have acquired such potency
in resisting their attacks whilst still alive. Absence of bacteria
or absence of immunity are alike incompatible with animal life.

Considerations of this nature lead us to a short discussion of
the difference between the pathogenic and non-pathogenic bacteria,
and we find that there is, theoretically, none. Any bacterium
will produce disease if it grows in the tissues of the living body,
and all bacteria1 will do so if the necessary degree and form of
immunity is not present. A pathogenic organism is one which
can grow in the living tissues, and it can do so only because those
mechanisms of immunity which are sufficient in the case of the
saprophytic bacteria are powerless to resist it ; but in most cases,
as we shall show, a higher degree of immunity can be produced
artificially, and the microbe in question then becomes non-patho-
genic to that particular animal. So, too, with the bacteria
ordinarily regarded as non-pathogenic. Under certain circum-
stances, some of which are known and some still unknown, the
resistance of the body or of a part of it may be broken down to
such an extent that these organisms may gain access, flourish,
and give rise to disease. Thus, B. proteus may give rise to
phlebitis, growing in the thrombosed vein, and giving off toxins
which have an injurious action on the tissues.

1 Bacteria growing only at very high or very low temperatures, or on media
very poor in nitrogen, perhaps excepted.

I — 2

4 INTRODUCTION — PATHOGENICITY

As a matter of high theory, therefore, there is no fundamental
distinction between pathogenic and non-pathogenic bacteria, and
we can imagine circumstances in which the tissues are vulnerable
to attack by almost any microbic species. Practically, however,
we shall consider an organism as pathogenic when the immunity
of the animal which it attacks is not so perfectly developed that
its presence in the tissues is but transient and unaccompanied by
any noticeable ill-effects, but in which there is a balanced contest
of longer or shorter duration between the injurious powers of the
microbe and the defensive mechanism of the host, accompanied
by more or less injury to the tissues and disturbances of the
physiological economy of the latter, and resulting either in the
death of the invader or of the patient. All grades occur. In
most staphylococcic infections the chances are enormously on
the side of the host, and the immunity is sufficiently high to
localize the process before it has gone far. In typhoid fever the
natural immunity and the pathogenic power of the organisms are
more nicely matched ; the contest between them is of long duration
and doubtful issue. And in some forms of human disease, but
more especially in artificial infections of animals with highly
virulent cultures, the power of immunity seems almost nothing,
the bacterium growing apparently unchecked and death occurring
within a few hours. We say that these organisms have different
degrees of pathogenicity, but it would be equally correct to say
that there are different degrees of resistance against them, since an
organism that is highly virulent towards one animal species may
be quite harmless to another, so that pathogenicity is not an
inherent property of certain bacteria.

Thus far we have considered the resistance of the host as if it
were fixed and definite, but this is not the case. It has been
known from time immemorial that certain diseases — especially
those due to infection — are followed by a greater or smaller degree
of immunity, so that a second attack is unlikely — at any rate, for
some time. Smallpox, scarlet fever, and measles are amongst the
most striking examples, and in them the protection given by the
disease is in most instances absolute and lifelong. This is known
as acquired immunity, and we shall enunciate it as a law that all
recovery from infective disease is due to, and followed by, some
degree of acquired immunity, though this may be slight, transient,
and perhaps local.

Take, for example, a case of pneumonia, a disease which may

INTRODUCTORY AND GENERAL 5

occur repeatedly and at short intervals in the same person.
Pneumococci are widely distributed, and are almost universally
present in the mouth ; the necessary exciting cause, therefore, is
always at hand. Under ordinary circumstances the power of
resistance is sufficient to ward off the infection, but when this
barrier of immunity is broken down by certain adverse circum-
stances— by excessive fatigue or starvation, by cold, or by an over-
dose of alcohol or other poison — the pneumococcus gains access
to the tissues, and infection1 occurs. The balanced contest
spoken of above then takes place. The pneumococcus grows in
' the lungs and blood and produces a toxin, which tends to reduce
the general health and the resistance of the body still further; and
looking at the problem only from this side, it would appear that
the process would go on until all the immunity was broken down,
and the pneumococcus could flourish unchecked. This, indeed,
might perhaps happen did not death supervene and bring with it
conditions unfavourable for the growth of this organism. But all
this time the tissues of the host have been reacting, and (in non-
fatal cases) sooner or later a condition is brought about in which
the noxious power of the coccus and the immunity of the patient
are exactly level, so that the disease neither advances nor retro-
cedes ; and the process goes still farther, and the patient develops
such a degree of resistance as will not only render him immune to
the spread of the infection, but will suffice to sterilize his tissues
of the pneumococci which have already gained access. In other
words, there has been an acquisition of immunity; the patient has
become immune to the pneumococcus, and it is this, and this
only, which has brought about the cure of the disease.

This process may be represented very diagrammatically, as
shown on p. 6.

The line ag represents the degree of immunity to the organism
in question, the pneumococcus. At b some event takes place
(e.g., exposure to cold) by which the resistance is lowered to such
a degree that infection can occur. This takes place at c, with the
result that the immunity falls still farther. At this time the
bacteria begin to flourish in the tissues in increasing numbers.
This is represented by the ascending line i. The immunity falls
and bacterial action increases until a certain point is reached,

1 I have elsewhere defined infection as the access of living, virulent,
pathogenic bacteria to a region whence their toxins may act on the tissues of
the body (Rose and Carless's " Surgery," sixth edition et seq., chap. i.).

6 RECOVERY FROM DISEASE

when the reserve forces of the patient have been brought into
action, with the result that the immunity rises (from d to e).
Somewhere during this rise (not necessarily or probably at its
commencement) the contest turns in favour of the host ; the bacteria
are rapidly destroyed, and the disease is cured. Usually, but not
necessarily, there is a rise to a level higher than the previous
normal one (e to/), of longer or shorter duration, and then a rever-
sion to the normal g. If exposure to cold again takes place, a
fresh infection may now occur.

Now it must be emphasized that natural recovery from disease
only takes place in virtue of an acquisition of immunity to the
infecting agent, and in no other way ; and, further, that, except in
a few instances, medical treatment simply aims in aiding this
phenomenon. If we exclude the various sera and vaccines, there

k

FIG. i.

are but two therapeutic agents which have a direct curative effect
— mercury in syphilis and quinine in malaria.1 In these diseases
the physician can apply a direct remedy, but in other cases the
aim and object of treatment is to support the patient's strength
until the natural development of acquired immunity takes place,
and in some cases to aid this development by certain empirical
means. It is found that all agents which tend to improve the
general vitality and facilitate the performance of the normal physio-
logical processes have this action ; hence the importance of suitable
food in amounts and at intervals suited to the patient's com-
plaint, of fresh air at a proper temperature, of the removal of
pain, and other symptoms which tend to impair the patient's

1 Arsenic and some other drugs in the treatment of various protozoal
infections (trypanosomiasis, etc.) may also be included. It is interesting to
notice that all the diseases directly combated by simple means are protozoal
in origin.

INTRODUCTORY AND GENERAL 7

strength. These agents are all-important in medical treatment 9
but in themselves they are useless, and they only act by hastening
the evolution of the immunity, without which the disease must
necessarily progress to a fatal issue. This is well seen in the few
diseases in which the development of immunity, in face of a natural
infection, is but slight, or perhaps altogether absent, such as
leprosy or hydrophobia. Here ordinary medical treatment is
powerless, and all our hopes for the future are concerned with the
discovery of a direct specific remedy.

It is this connection between immunity and recovery that
renders the subject so important to the physician, and the neglect
with which its study is treated by the general members of the pro-
fession a matter of such profound regret. In our medical educa-
tion at the present day we pay, and rightly, much attention to the
study of physiology, for without a knowledge of the processes of
the healthy body we can hardly hope to diagnose and treat its
derangements when diseased; and our physicians are in many
cases competent physiologists. But it is equally important to
understand the method in which the diseased body combats and
cures an infection ; and, although our knowledge of this is as yet
imperfect, it is increasing day by day, and results of the greatest
interest to the practising physician have already been obtained.
And I, for one, think that an intelligent appreciation of what is
actually taking place in the body, of the conservative and adverse
forces, and of the conditions necessary for cure, will always be of
value to the therapist, although it may not give any definite
information as to what drug is to be prescribed.

Let us revert to the subject of NATURAL IMMUNITY. We may
define it roughly as the immunity possessed by a certain individual
in virtue of its belonging to a given animal species ; it is inherent
to a greater or less extent in all members of that species, and is not
dependent on any event taking place during the life of the animal
in question. In most cases it is present at birth, though this is
not absolutely essential.

Examples are numerous. The lower animals are immune to
the gonococcus, and, with few exceptions (the higher apes), to
syphilis also. On the other hand, most of the diseases of the
lower animals do not affect man — fowl cholera, canine distemper,
and rinderpest are a few of many examples. In some cases all
animals, with a few exceptions, are immune : this is the case with
the venereal diseases, and in some of the protozoal infections of

8 NATURAL IMMUNITY

the lower animals. In others different types of the infecting
organism occur, and a given species is susceptible to one, immune
to others ; for example, there are three, and perhaps more, varieties
of tubercle bacillus, which resemble one another in many points,
and which attack respectively man, cattle, and birds, and each
animal species is more or less immune to bacilli from animals far
removed in the scale.

In general terms, the immunity or susceptibility of different
animals depends to some extent on their zoological affinities.
Thus man is pre-eminently susceptible to the Spirochata pallida,
the anthropoid apes less so, but still not immune, and the lower
animals entirely refractory. Rinderpest affects cattle, sheep,
goats, and other ruminants, and South African horse-sickness
horses, asses, and mules. But to this rule there are numerous
exceptions : thus, almost all warm-blooded animals are susceptible
to anthrax, but the Algerian sheep and white rat are relatively
immune, the wild rat being susceptible. And of the domestic
animals we find cattle to be highly susceptible to tubercle, whereas
goats, though closely allied zoologically, are almost immune.

Natural immunity does not exist to an equal degree in all
individuals of a species. This is well seen in man during an
epidemic, where, of a certain number of persons who are exposed
to an infection (and, as far as we know, receive the same dose of
the materies morbi), some escape the disease altogether, some have
a slight, and others a severe, attack, whilst yet others die rapidly.
Sex has some influence here, but it is usually difficult to trace,
since the males and females of a community are in most cases
exposed to an infection in varying degree.

Age is of more importance, and, in quite general terms, we may
say that the younger the infant the less its immunity. Certain
diseases, such as measles, scarlet fever, and whooping-cough, are
rarely seen except in infants, and this is not altogether due to
acquired immunity preventing a second attack in later life.
Epidemic diarrhoea due to bacilli of the dysentery group is
rarely seen — in this country, at least — except in the early years of
life, and the same is true of cerebro-spinal meningitis and some
other diseases. It is also interesting to notice that the variation
in immunity may take a qualitative rather than a quantitative
form. The best example is in the case of the pneumococcus.
This organism is the chief cause of suppurative processes of
whatever region in infants, whereas in adults it is (except in

INTRODUCTORY AND GENERAL 9

certain regions) a decidedly rare cause of abscesses and other
pyogenic processes. It is evident that the form of immunity
which prevents the pneumococcus from gaining access to the
tissues and giving rise to abscess formation is in abeyance in the
young and well developed in the adult ; yet the two are more
nearly equal in their resistance to this organism in its role of a
producer of pneumonia. There are also very marked differences
in regard to local immunity in the two ages, but of those we shall
speak subsequently.

Natural immunity must not be regarded as a fixed and definite
quantity, since all individuals vary enormously in their resisting
powers against various diseases at different times and under
different conditions. The factors which tend to break down the
immunity against any or all infections may be referred to as the
banal causes of the diseases in question. They are not in them-
selves sufficient to lead to these diseases, but when they come
into action and an infecting agent is present the disease will arise.
Hence they are often referred to as predisposing causes of disease,
and to the lay public they are the actual causes, since they are
usually open and obvious, and the real infecting agent is, of
course, unknown They are of the utmost importance in pre-
ventive medicine, and wherever the probability of an infection is
apprehended, a study of the patient's surroundings and habits
may often lead to the giving of advice by which these banal
causes of infection may be avoided and the disease warded off.
In general these predisposing causes are a study for the physician
rather than for the pathologist, and in some cases we are quite in
the dark as to the method in which they act. Their study cannot
be conveniently undertaken here before the mechanisms and pro-
cesses of immunity have been described, but it will be useful to
enumerate some of the more important.

Of these cold and wet, especially in combination, are unquestion-
ably the most important. The exact way in which they act is
not definitely known, but there are materials for a number of
suggestions. Thus, as we shall have abundant opportunity of
seeing, immunity is to a very large extent a function of the leuco-
cytes, to which are specialized cells to which the defence of the
body is entrusted. Now the functions (movement and phagocy-
tosis) which can be easily investigated are found to be dependent
in a very high degree on temperature, acting best at the tempera-
ture of the body, or slightly above ; and it is highly probable that

10 COLD, WET, AND FATIGUE

the more subtle functions of the leucocytes may be similarly
depressed by a low temperature. The exposure of the skin to
cold, especially if the animal heat be abstracted more quickly
by evaporation of moisture on the surface, will lead to a cooling
of the blood which circulates through it, and hence to a slight,
though appreciable, cooling of the whole blood. This, it is true,
is soon compensated for, and no great amount of cooling of the
whole body occurs; but even so, it is quite possible that the
periodical chilling of the leucocytes during their repeated passages
through the cold skin may be sufficient to diminish greatly their
functional activity, and to lower the resistance to a point at which
infection can occur, and when once pathogenic bacteria have
gained a foothold, the resistance will for a time tend to decrease.
There is also some evidence going to show that exposure to cold
may lessen the production of the defensive substances which occur
in the blood (alexin, antibodies, etc.), though this is not fully
proved. It is worthy of note that the loss of immunity due to the
action of cold and wet on one part of the body (such as the feet)
is a general one, and may result in a nasal catarrh, an attack of
pneumonia, acute rheumatism, etc., according to the nature of the
infection at hand. It is not necessarily a local infection of the
chilled region. This is very well shown experimentally. Fowls
are immune to anthrax, but are rendered susceptible if they are
kept for some time standing in cold water ; and this acquired
susceptibility is then a general one, and not merely of the feet.

Cold and wet, as is well known, have less action when accom-
panied by energetic muscular exercise, so long as this does not
reach the extent of undue fatigue. This is not because less heat
is lost during exercise. The reverse is the case. The suggested
explanation is that the muscular metabolism leads to an increased
production of heat, and at the same time the cutaneous capillaries
are dilated and the heart accelerated, or that the circulation of
blood through the skin occurs quickly ; further, the internal
temperature of the body may actually be raised several degrees.
The result is that the temperature of any given leucocyte never
falls much below normal, if at all, since it comes from the internal
regions where the temperature is raised, passes rapidly through
the skin, and returns again to the interior of the body.

The effect of fatigue, either alone or in conjunction with cold and
wet, is also well known, and is one reason for the excessive mor-
tality from disease of armies in the field. It is less explicable,

INTRODUCTORY AND GENERAL II

but may probably be connected in some way with the presence in
the blood of katabolic products of muscular activity, which have
an injurious action on the cells of the tissues in general and on the
leucocytes in particular. Further, the metabolic products formed
during the action of the muscles are acid in reaction, and it is
found that some at least of the protective substances which occur
in the blood (alexins and opsonins) act best in an alkaline medium.
This diminution of immunity after muscular fatigue is manifested
in animals as well as in man. White rats which have been made
to work in a revolving cage are more susceptible to anthrax than
normal white rats, the pre-existing immunity being broken down.

Insufficient or unsuitable food is a factor of importance, especially,
perhaps, in the aetiology of tuberculosis. It is, however, rarely
seen alone — in this country, at any rate — and in the poorer classes
its effects are usually complicated by insufficient clothing, un-
cleanly habits, and by insufficient ventilation of their houses. For
this reason we may perhaps be led to exaggerate its importance ;
and whilst it is, of course, true that semi-starvation, in common
with other weakening influences, does pave the way for infective
processes, we do not find that a supply of food restricted enough
to cause a marked reduction of the bodily strength and some
degree of anaemia is necessarily associated with any infective
disease, though the patient may live under conditions in which
infective material is present in abundance. This is well seen in
fasting men, in hysterical anorexia, and in patients with imperme-
able cesophageal strictures. The blood, it may be pointed out, is
not one of the tissues that suffers first in starvation, and its im-
portance to the body in many ways is so great that it is kept in
good functional activity whilst other regions waste quickly.

It is probable that insufficient food lowers the resistance of the
body in certain directions rather than in others. In the East
plague follows famine with some regularity, but there is little or
no connection between famine and cholera. But in these latitudes
at the present time the disease most commonly due to bad or in-
sufficient food is tuberculosis. Formerly it was relapsing fever, or,
as it was sometimes called, famine fever, a disease which is now
almost extinct as a result of the general cheapening of foodstuffs.

It is worthy of note that the number of leucocytes per cubic
centimetre diminishes in starvation, and is generally lower in the
badly-nourished than in the well-fed ; and these cells, as we shall
see, are pre-eminently concerned in immunity, and this in a great

12 EFFECTS OF A VITIATED ATMOSPHERE

many ways. It was recognized long ago that post-mortem wounds
are much more dangerous when received whilst fasting than during
the process of digestion, and it is possible that this may be due to
some extent to the increased number of leucocytes which occur in
the blood during the process.

Exposure to a vitiated atmosphere, if of long duration, is a most
potent cause of the breaking down of immunity, and when con-
sidered on a large scale, and in view of its effect on the general
death and disease rate, is probably of greater importance than all
other causes combined. It is especially important in connection
with tuberculosis, and nothing is more striking than to notice its
effect on the peasantry of some regions, in which, in spite of
exposure to abundant fresh air during the daytime, and a supply
of food which certainly does not fall below the physiological
minimum, and is usually more abundant, phthisis and other
tuberculous diseases are rife. These affections are in general
common in cold and windy climates, and less prevalent in warmer
countries, and there is little doubt that the main reason for this is
the habit which dwellers in cold countries frequently contract of
hermetically sealing all entrances to their rooms to keep out the
cold. But this is frequently seen in warmer regions, and even
throughout the South of England there is an almost universal
opinion amongst the lower classes that night air is injurious.
This is probably a survival from the time when malaria was
indigenous in this country.

Apart from tubercle, the effect of bad air is especially mani-
fested in the causation of diseases of the lungs, nose, throat, etc.,
and its effect is probably partly general and partly local. The
effect of irritating vapours is, of course, local. Thus exposure to
nitrous fumes is often followed by the rapid development of
pneumonia, and this is, or may be, due to the pneumococcus,
which is able to invade the injured lung.

We do not know the mechanism by which ordinary vitiated air
acts on the general immunity.

Prolonged anesthesia is probably a cause of considerable
importance, though one not easy to estimate. The prevalence of
ether-pneumonia is not yet ascertained, and has been hotly
debated. It falls, of course, into the same category as the
pneumonia due to irritating vapours, as described above. Apart
from this, however, there is reason to believe that prolonged
anaesthesia has some effect in lowering the general resisting

INTRODUCTORY AND GENERAL 13

power of the body to the common pyogenic bacteria, and that the
mere length of an operation should be an indication for the most
scrupulous care in antiseptic precautions. It is perhaps con-
ceivable that the anaesthetic drug present in the blood may be
sufficient to paralyze the leucocytes for a sufficient time to allow
bacteria to gain a foothold in the body.

Certain drugs, of which the most important is alcohol, have an
important action in this respect. The liability of alcoholic
subjects to pneumonia and some other infective diseases is well
known, and in them the prognosis is more than usually unfavour-
able. We have but little knowledge of the action of alcohol in
this respect. It may be that it acts as a direct inhibitant of the
activity of the leucocytes, and it is known to destroy certain
delicate defensive substances (alexins and opsonins) which play
some part in the defence of the body against microbic invasion,
but it is not known whether these effects are actually manifested
in the circulating blood. It is also possible that alcohol tends to
inhibit the formation of these defensive substances.

Alcohol tends to lower the temperature of the body by increas-
ing the amount of heat lost. It dilates the superficial vessels and
accelerates the heart's' action in a way somewhat similar to
muscular exercise, but does not, like it, raise the temperature of
the interior of the body. Hence the effect of alcohol in conjunc-
tion with cold and wet is to increase their ill-effects. More blood
is forced through the chilled skin and more heat is lost. The
injurious effect of alcohol during exposure to cold is well known.
The results, however, are different when alcohol is taken after
exposure, and when the sufferer has reached warmth and shelter.
There the increased flow in the cutaneous capillaries leads to a
warming of the skin and consequent cessation of the chilling of
the blood, although the loss of heat may go on.

Diseases — the most important of which are Bright's disease and
diabetes — lead to a general lowering of the level of immunity, and
a consequent predisposition to other diseases. We have no
knowledge of the way in which they act.

There are many causes which act locally, and cause a local
lowering of the resistance. Some of these have been hinted at
above, but their consideration will be deferred for the present.

In considering the nature, severity, and prognosis of any disease,
two factors have to be recognized : (i) the immunity of the patient,

14 VIRULENCE OF BACTERIA

and (2) the virulence of the infecting bacterium. A third — the
number of bacteria which gain access — is also of importance,
especially under experimental conditions, for it is found that,
within limits, lack of virulence can be compensated for by an
increase in the dose given. It is, however, one which we can
rarely estimate in natural disease ; besides which the growth of
bacteria is so rapid that, if not checked by the resisting power of
the body, a single organism would multiply in a very few hours
to an enormous extent, and render it a matter of but little impor-
tance whether one or a hundred bacteria had gained access at
first. The number of bacteria is probably of more importance in
connection with the occurrence or non- occurrence of infection,
rather of the severity of the disease when once infection has
occurred. Thus we find in epidemics of typhoid fever due to
water or milk that the disease is most prevalent in those who
take a large amount of the infective material, but it is not neces-
sarily more severe in them than in the patients who have appa-
rently become infected with a small dose. This is, however, not
the case with artificial infection of animals, for there the severity
of the disease (in animals as similar as possible in age, weight,
etc.) is fairly proportional to the dose given. But the conditions
are somewhat different in the two cases, and in the artificial injec-
tion of animals we eliminate altogether the steps by which, e.g.,
the typhoid bacillus passes the natural barriers, and gains access
to the tissues.

The question of virulence is of much greater importance, and
is one which must be more fully discussed subsequently, after we
have seen the methods in which the host immunizes itself against
the bacterium. Some general points must be mentioned here.

Cultures of the same organism, identical in all respects in
morphological, cultural, and chemical characters, may differ
enormously in this respect : thus a culture of streptococci may be
entirely devoid of virulence to rabbits, or may be so potent that a
minimal dose, containing probably but a single coccus or short
chain, may be inevitably fatal. Similar facts hold for pneumo-
cocci. According to Eyre, a virulent culture may kill when 20 to
200 cocci are injected, whereas an avirulent one may fail to do
so in massive doses. In most organisms there is, perhaps, not
such a marked difference, but all pathogenic bacteria vary greatly
in this respect, and cultures from different sources show marked
variations in pathogenicity.

INTRODUCTORY AND GENERAL 15

Further, the same culture can be made to undergo variation,
its virulence being either exalted or diminished, and this is a
subject of the utmost importance. An increase in virulence is the
more difficult to secure, and can practically only be procured by
passage through animals, or by other closely allied process.

Passage is carried out thus : the avirulent culture is made to
infect animals either by the administration of massive doses, or
by the simultaneous injection of some substance which lowers the
local or general resistance (lactic acid, alcohol, the toxins of
B. prodigiosus, etc.). In any case, the organism is made to cause
an infection which may or may not be allowed to progress to a
fatal issue. From the animal thus infected a second culture is
made, and the material used to inoculate a second animal, and
the organism will be found to have undergone a noticeable access
of virulence. The process is repeated as often as is necessary,
and ultimately the virulence of the culture will be brought to
its highest possible pitch. The simplest method, where available,
is to give the injections into the peritoneum, and to make the
cultures by withdrawing some of the peritoneal fluid in a sterile
pipette, and incubating it as it is, or after the addition of broth.

This method was introduced by Pasteur, and is of especial
value in preparing the vaccine used against rabies. The organism
of this disease is unknown, but the virus occurs in the brain, and
emulsions of this substance are used for inoculation. It is found
that the virus occurring naturally in rabid dogs (the " virus of
the streets ") is comparatively avirulent to rabbits. This is
shown by the long incubation period — fifteen to eighteen days
after intracerebral injection. After about fifty passages through
rabbits, the virus becomes so exalted that the incubation period
is shortened to six days, and the process cannot be carried further.
This virus is called the " fixed virus," and its potency is main-
tained unaltered, no matter how many more passages are made.

Passage does not necessarily raise the virulence of the culture
to all animals ; it may do so only for the species used for the pro-
cess, the action on other species remaining unaltered or even
falling. Nor is passage necessarily followed by an increased
degree of virulence — the virus of rabies diminishes in this respect
when passed through apes.

Phenomena suggesting a process akin to passage occur under
natural conditions. Pneumococci are frequently found in the
mouths of healthy persons, and are, as a rule, of feeble virulence,

l6 INCREASED VIRULENCE

whilst those which are isolated from the lungs in pneumonia, or
from pneumococcic lesions in general, are usually far more virulent.
Other explanations are possible, but it seems likely that the
sequence of events is as follows : The avirulent pneumococci gain
access to the body owing to a temporary loss of immunity, due to
one or other of the causes enumerated above, and then these are
transmitted to a process in all respects like passage, the result
being that they undergo a gradual increase in virulence. The
struggle of the conservative forces will then be increasingly difficult,
and the patient may succumb to an infection with an organism
which was at first but slightly virulent. This adaptation of an
organism to its environment during the course of a disease may
probably be found in the future to be of great importance, as indi-
cating a necessity for successive changes in the vaccine or serum
used in the treatment of a chronic infection.

An example worthy of notice has recently been given by
Ehrlich. It is not exactly on the same lines, since it deals with
an alteration in the body of the power possessed by the parasite of
resisting chemical agents of relatively simple composition, rather
than in the power of resisting the natural forces of the body, an
increase in which constitutes an increase in virulence. Ehrlich
investigated the preventive and curative action of atoxyl and of
various aniline dye-stuffs, such as fuchsin and trypanroth, on mice
infected with trypanosomiasis. He found in a certain number of
cases a cure might be obtained — e.g., by feeding infected mice with
fuchsin or by the injection of atoxyl — and that when this occurred
the trypanosomes were not entirely destroyed, but remained latent
in the body. This is a phenomenon of fairly frequent occurrence,
and is called by Ehrlich, " immunitas non sterilisans." After a
time a relapse occurred, and was cured by a fresh dose of the drug,
but after several of these recurrences this beneficial effect ceased.
It was then found that the trypanosomes had been immunized or
acclimatized to the agent in question — say, to fuchsin — and pos-
sessed the power of infecting mice previously treated with fuchsin
and immune to ordinary trypanosomes ; but the organism had
not altered in its susceptibility to other dye-stuffs or to atoxyl,
and mice infected with it could be cured by these agents, and not
by fuchsin. Further, it was found possible to create a race of
trypanosomes resistant to two or more of these agents, and these
acquired characters were made permanent after several passages.
If we substitute for the drugs used by Ehrlich the substances

INTRODUCTORY AND GENERAL 17

which are developed in the body as defensive agents during an
attack of disease, and imagine the same process to go on, we
shall have an exact reproduction of the rise in virulence occurring
during an attack of disease. If we defended ourselves against
trypanosomes by the development of fuchsin, Ehrlich's fuchsin-
resistant race would be extremely virulent for us.

The development of epidemics of diseases is probably due in
some cases to a spontaneous rise in virulence of the infecting agent,
but we have no knowledge of the causes by which this is produced.

The second method of increasing the virulence of a culture is
less general, and of greater theoretical than practical interest. It
consists in the cultivation of the organism for several generations
in the blood-serum of an animal which has been immuned to the
bacterium in question. It was discovered by Walker in the case
of B. typhosus, and is found in the case of some other organisms.
It is referred to subsequently, and we need only say here that it is
allied to passage ; the organism is immunized to the fluids of the
resistant animal in vitro instead of in vivo. And the virulence of a
culture is in general best sustained by a close approximation to
the conditions of the body. Thus it is more rapidly lost at the
temperature of the room than at that of the body, and the most
suitable culture medium is usually one containing body fluids
unaltered by heat. Thus Marmorek cultivates his virulent strepto-
cocci in broth to which one-third of its volume of ascitic fluid has
been added. In the case of diphtheria bacilli the virulence (as
estimated by its power of forming toxin) is best maintained by
daily transplantations into broth previously raised to the body
temperature, and when treated in this way shows little or no
change for years.

Diminution in virulence occurs, as a rule, when the organism is
submitted to conditions quite unlike those of the animal body, and
is usually the more rapid the greater the divergence. At the same
time, the growth under these conditions gradually becomes (in
most cases) more abundant. The organism gradually adapts itself
to a saprophytic habitat, losing in so doing its distinctive chemical
properties which made it virulent as a parasite. Old laboratory
cultures of bacteria which have been grown on artificial media for
many generations are usually almost devoid of virulence, though
here there are great variations, some species becoming inert far
quicker than others.

The subject is important, since cultures of diminished (" miti-

2

l8 DIMINUTION IN VIRULENCE

gated ") virulence are frequently employed as vaccines in the pro-
duction of artificial immunity. The following are some of the
chief methods employed :

1. By prolonged culture in artificial media, as described above.
This method was introduced by Pasteur in the case of fowl
cholera. The loss of virulence is a progressive one, and cultures
ten months old are devoid of virulence.

2. By cultivating the organism at a temperature above the
optimum for saprophytic growth. This was also introduced by
Pasteur, and is used in preparing the vaccines to anthrax. The
organism is cultivated at a temperature of 42-5° C., and all
virulence is destroyed in about six weeks, though the cultures
retain their power of growth unaltered. The first vaccine is prepared
by allowing growth to continue at this temperature for twenty-
four days. In appearance the bacilli are unaltered, but they have
lost the power of killing rabbits and guinea-pigs, though they
are still fatal to mice. The second vaccine is cultivated at a high
temperature for a fortnight only; it is virulent to mice and
guinea-pigs, but not to rabbits.

The process may also be carried out by a short exposure to a
higher temperature. Chauveau's vaccine consists of blood con-
taining anthrax bacilli, heated to a temperature of 50° to 55° C. for
ten or fifteen minutes. The bacilli remain alive, but are mitigated
in virulence.

3. In some cases the addition of various chemical antiseptics in
minute amounts to the culture medium has a similar effect. This
is the case with anthrax also. Addition of chemical substances is
also made with the idea of destroying toxins, but this is a different
phenomenon.

4. The virulence may be destroyed by drying. This method
was introduced by Pasteur for the preparation of a vaccine against
rabies. We have already described the method by which he
obtained the fixed virus and its action on rabbits. He found that
by suspending the spinal cords of rabbits dead of this fixed virus
over caustic potash at a temperature of 23° C. the virulence was
entirely removed in fifteen days. Drying for a shorter time
diminished the virulence, but did not remove it entirely.1

1 The more modern idea is that the process of drying kills off a large
number of the pathogenic organisms, and that the use of the dried cord is
merely another method of giving very minute doses of virus of normal
strength.

INTRODUCTORY AND GENERAL IQ

5. In some cases (as has been noted above) passage through
animals diminishes the virulence. In some cases this can be
exalted by passage through a series of animals of one species and
diminished by the use of another. Pasteur showed this to be the
case in swine erysipelas, the potency of which (as tested on pigs)
is increased by passage through pigeons and decreased by passage
through rabbits. Cultures thus attenuated are used as vaccines.

The term ACQUIRED IMMUNITY is one that is used to denote an
increased resistance to an organism dependent on some modifica-
tion in the animal's constitution due to some definite process to
which it is subjected, but not including the modifications due to
improvements in the general health due to betterment of the
environment. For example, a person living in insanitary sur-
roundings will undoubtedly acquire a higher degree of resistance
to the tubercle bacillus on being moved to more healthy ones, but
we do not speak of that as acquired immunity. The distinction
is this: The elevation of the natural resisting powers due to
improvement in the general vitality is a more or less general one,
and affects the immunity to most or all bacteria almost equally ;
whereas in acquired immunity in the narrower sense, to which
the use of the term is restricted by pathologists, the alteration is
in the powers of resistance to one bacterium only. For example,
a debilitated person removed to a more healthy environment,
given better food, tonics, etc., would become more resistant to
the attacks of smallpox, and to other diseases as well ; we should
speak of that as an augmentation of the natural immunity. But
after an attack of smallpox, or after vaccination, his immunity
to smallpox is enormously increased, whereas his resistance to
other organisms is unaltered ; this is acquired immunity.

This is expressed by the use of the word specific, embodying an
idea difficult to define, but implying a direct relationship of cause
and effect, and, moreover, that a certain effect is only produced
by a certain definite cause. Thus the toxin of diphtheria is
specific for the diphtheria bacillus in the sense that it is produced
by that organism, and by no other ; diphtheria antitoxin is specific
for diphtheria toxin, since it is produced only as a result of the
injection of that substance ; the reaction caused by the injection
of tuberculin into a tuberculous animal is specific, etc.

In most instances there is another difference between a rise in
natural immunity and the development of acquired immunity, in
that the latter is much stronger. Thus, the power of resistance

20 ACQUIRED IMMUNITY

to smallpox of a perfectly healthy person is probably not great,
whereas that produced by an attack of the disease or by vacci-
nation is for a time almost absolute. Yet all degrees of acquired
immunity exist, from the very slight amount which is developed
during an attack of pneumonia, and which is probably only just
sufficient to cut short the disease, to the enormous degree that
can be obtained in animals hyperimmunized to diphtheria or
tetanus toxin or hypervaccinated to B. typhosus. Perhaps our
conception of immunity in the past has been influenced too
strongly by a study of these latter conditions, which are readily
induced in the laboratory, but rarely if ever seen in the actual
practice of medicine. They represent in an extreme form the
changes which follow disease of natural origin, and possess the
theoretical interest which attaches to all extreme cases.

ACQUIRED IMMUNITY occurs in two distinct forms — active and
passive. A third form exists, which we may call mixed, since it is
brought about by a combination of the procedures necessary for
the development of the other two.

Active immunity may be defined as acquired immunity, due to
an attack of the disease in question in its normal form, or in a
modified and less severe form of artificial production. The
essential feature is that the cells and tissues of the person or
animal should be subjected to the action of the bacterium (or its
toxin), and by its own efforts, and as a result of an active struggle
with it, should become less susceptible to its toxin than before.
Active immunity is developed only as a result of an illness of the
host, due to the action of the microbe on its cells ; and this illness
may be of any degree of severity, ranging from an unmodified
attack of the disease which may threaten life down to the most
transitory and unimportant reaction due to an injection of a
minute dose of a mild vaccine. And one of the great aims of
modern preventive medicine is to reduce the severity of the disease
necessary to produce acquired immunity to a minimum. The
greatest step ever made in this direction was Jenner's substitution
of vaccination for inoculation. In each case the effect is the same
as regards the resulting immunity (though in different degree),
but the disease in the former case is mild and devoid of danger,
in the latter severe and dangerous. As a general rule, it may be
taken that the severer the disease the stronger and more lasting
the acquired immunity. A good example will be quoted when
dealing with mixed immunity. This is not necessarily the case,

INTRODUCTORY AND GENERAL 21

however, for the repeated injections of vaccines which are so mild
as not to cause any noticeable general and very little local reaction
may induce a high degree of immunity.

The main methods in which active immunity is acquired are
these :

i. A natural attack of the disease, or an attack which is natural
in course, but of artificial induction. The only example of the
latter in human medicine is the now disused practice of smallpox
inoculation, in which the person to be protected was inoculated
with the disease, which ran a perfectly usual course, and was not
infrequently fatal. As a rule, however, it was milder than the
naturally acquired smallpox, since the infective material was
taken from a favourable case, and the operation performed when
the patient was in good health and able to get proper attention
from the outset. Probably, too, the severity of the disease was
somewhat modified by the fact that the virus did not reach the
body by the usual route. But the infection was ordinary small-
pox, and might start an ordinary epidemic.

The process is used to a much greater extent in veterinary
practice, where an occasional death due to the induced disease is
of comparatively little importance if thereby the outbreak can be
controlled or the great majority of the flock saved. As a rule,
an attempt is made to render the attack as mild as possible,
either by (a) limiting the amount of the infective material used,
or (b) by introducing it in an abnormal way, or (c) inoculating
animals at a time when they are found to be least susceptible, or
by a combination of these methods. Thus Texas fever is a
disease of cattle due to a protozoon (Piroplasma bigemimim) which
is conveyed by the bites of ticks. One of the methods used for
the protection of cattle in infected districts is to expose calves
whilst still milk-fed to the bites of a few infected ticks ; another
is to inject blood from diseased animals (containing the parasite)
in small doses direct into the jugular vein. In favourable cases
the result is a severe attack of the disease, which, however, is
rarely fatal, and is followed after a time by complete immunity.
In some cases the disease is but slight, and in them a second or
even third dose, in each case larger than the preceding, is required.
The mortality from the injections is from 3 to 10 per cent., whilst
that of untreated animals in infected areas is about 90 per cent.

A similar method is in use for combating rinderpest, but here
bile from an animal dead of the disease is used as the infecting

22 METHODS OF VACCINATION

agent, since the blood frequently contains other infective materials
which would complicate the issue.

In pleuro-pneumonia of cattle the severity of the disease is
lowered by altering the route of infection. In the natural disease
the infection probably enters by the lung, and its course is severe
and dangerous. Protection is conferred by inoculating virus from
the lung of an animal dead of the disease into the subcutaneous
tissue near the tail ; much local swelling results, and general
immunity is established. Perhaps, strictly speaking, this method
of induction of active immunity should be put in a class of its
own ; it is one in which a local is substituted for a general disease,
with the obvious result of greatly lessening its severity.

The material used in the production of artificial immunity of
the type we are describing is sometimes called a vaccine. This
is undesirable, and it is advisable to use the word virus for material
containing the infective agent in its normal virulence, retaining
the word vaccine for that in which the bacterium has entirely lost
its power of producing the normal disease, whatever the dose and
whatever the channel of introduction. The term is a somewhat
unfortunate one etymologically, but it is in such general use that
it is hopeless to attempt to displace it.

2. By the use of living cultures of pathogenic bacteria of
diminished or altered virulence — i.e., of a living vaccine. There
are as many modifications of this method as there are ways of
mitigating the virulence of a culture, and different methods are
applicable to different diseases.

(a) By the use of vaccines diminished in virulence by passage
through animals. The most important example of this is, of
course, Jennerian vaccination. It would take us too far to
examine the evidence in favour of this view, but it may be taken
as fairly proved that ordinary lymph vaccine consists of a culture
of the smallpox organism modified by passage through " calves,
the modification being of such a nature that it has lost its power
of producing a general disease (smallpox), but retained that of
causing a local one (vaccinia) otherwise similar in nature.

We have already referred to the decrease in the virulence of the
bacillus of swine erysipelas on passage through rabbits, and the
use of these mitigated cultures as a vaccine for pigs. This is a
better example of the type of immunity we are considering, since
it has to do with a known organism.

(b) By the use of vaccines in which the virulence is diminished

INTRODUCTORY AND GENERAL 23

by drying. The only practical example is rabies. There the
process of immunization is carried out by means of the use of a
series of vaccines of gradually increasing degrees of virulence, the
degree dependent on the time for which drying has gone on. It
is, of course, necessary to proceed with extreme caution, since the
cords that have been dried for but a few days are still infective
and virulent, and the amount of natural immunity in man is
extremely small, so that an attempt to accelerate the process
might be fatal. The method varies somewhat at different labora-
tories, but the following may be taken as a type of the procedure
used. It is of interest as being the method used in the first case
treated — that of Joseph Meister.

Day i. — Inoculation with vaccine made by drying the cord for
fourteen days. A second injection with cord treated for ten
days.

Day 2. — Two injections ; cords dried for eleven and nine days.

Day 3. — One injection ; cord dried for eight days.

Day 4. — One injection ; cord dried for seven days.

Day 5. — One injection ; cord dried for six days.

Day 6. — One injection ; cord dried for five days.

Day 7. — One injection ; cord dried for four days.

Day 8. — One injection ; cord dried for three days.

Day g. — One injection ; cord dried for two days.

Day 10. — One injection ; cord from a rabbit which had died
the same day, and which was therefore unaltered in virulence.

The method in use in France at the present day is almost like
this, except that the latter stages are repeated twice, or, in severe
cases, three times — i.e., on the ninth and fourteenth days in mild
cases (and on the nineteenth also in severe ones) injections of
nine-day cords are started, and the strength increased rapidly, so
that three-day cords are used on the thirteenth, eighteenth, and
twenty-first. In Germany the treatment is begun with eight-day
cords, the older ones being considered inert.

(c) By the injection of living cultures modified by heat. The
classical example is vaccination against anthrax by means of
Pasteur's two vaccines, the method of preparing which is given
on p. 1 8. The first vaccine is injected, and is followed by the
second in about a fortnight, immunity being established in about
another fortnight.

(d) By the injection of cultures attenuated by prolonged cultiva-
tion in vitro. The use of this method in the case of chicken

24 VACCINES OF DEAD BACTERIA

cholera has been referred to already, and it is the one usually
employed in the laboratory, where old cultures are used in pre-
ference to more virulent ones in the early stages of immunization.

(e) By the use of very small doses of living cultures of full
virulence. This has been proved possible in anthrax, symptomatic
anthrax, and some other diseases. At present the process is more
interesting than practically useful, but it has been used clinically
in the case of tubercle, treatment being commenced by the injec-
tion of a single living bacillus, and promising results have been
obtained.

3; A third class of methods consists in the use of vaccines
composed of dead bacteria. The advantages are obvious : the
dose is under accurate control ; the disease which it induces is
self -limited, so that it is impossible for a general infective process
to be produced when used on a person of deficient natural
immunity ; and the vaccine is easy to keep in a condition ready
for immediate use. Hence this method is mostly used in human
medicine, whereas the use of mitigated or unmitigated viruses is
mainly confined to veterinary work. The methods used in the
preparation of the vaccines varies greatly in the different cases,
and here we can only glance at some of the general principles.
In preparing the cultures, the most careful precautions have to be
taken to insure the purity of the microbe used and absence of all
other pathogenic forms, especially perhaps the spores of the
tetanus bacillus. The age of the culture has to be determined by
the necessities of the case, but as a rule young cultures are
preferable. The method by which the bacteria is killed also
varies, but heat is generally employed, and as a rule the shorter
the exposure and the lower the temperature the better. In other
cases the bacteria are emulsified in saline solution and allowed to
undergo autolysis at the body temperature, sterility being ensured
subsequently by means of heat or chemical antiseptics ; or they
may be killed with a minimum of heat, and submitted to autolysis
at 37° C. subsequently.

There are numerous methods of determining the dose to be
used, (a) A definite fraction of an agar or other culture of
known age may be taken, or, what comes to much the same
thing, the growth from so many square centimetres or millimetres
of surface of the culture medium. (b) The amount may be
judged by the weight, and this is the method used in the case of
tubercle. When it is employed with other bacteria it is usually

INTRODUCTORY AND GENERAL 25

carried out by means of standard loops, each of which will pick
up a known amount of growth. (c) In Wright's ingenious
method of counting a vaccine a certain amount of the latter is
mixed with human blood in definite proportions, and films are
prepared and stained. The numbers of red corpuscles and of
bacteria in several fields of the microscope are then counted, and
(the numbers of red corpuscles in a definite volume of the blood
being known) the proportions of the two will permit of the calcu-
1 ation of the numbers of the bacteria, (d) Some determine the
strength of the vaccine by reference to a permanent standard,
usually consisting of a fine suspension of barium sulphate. A
strong emulsion of bacteria is prepared and diluted until it
matches the standard, (e) The volume of the bacteria in the
emulsion may be determined by centrifugalization in a graduated
tube, and a certain volume of sediment made up to a certain
volume of vaccine. (/) The number of bacteria present in the
emulsion may be counted directly by the use of the counting
chamber of the hsemocytometer, and this is the method I usually
employ. The emulsion is diluted (usually to twenty times its
volume) with a dilute solution of methylene blue or other stain,
boiled, and a drop placed in the counting chamber and prepared as
if it were a blood specimen in which the red corpuscles were to be
counted, A ^-inch lens and a high eyepiece are used, and, as a
rule, the process presents no difficulty.

In all cases an addition of a chemical antiseptic is advisable
to avoid subsequent contamination. Carbolic acid or lysol (0*25 to
0-5 per cent.) are most used; another good plan is to keep a few
drops of chloroform at the bottom of the bottle, so that the fluid
is always saturated.

This method is mostly used in plague, cholera, and enteric
fever in preventive medicine, and in the treatment of infective
processes by Sir Almroth Wright's method in curative medicine.
These will be discussed subsequently.

4. Inoculation with the chemical products or with the toxins
of the bacteria, the bodies of the bacteria themselves being
removed by filtration or in some other way. This is obviously
closely allied to the last method — the use of killed cultures.

It was introduced by Smith and Salmon in the case of hog
cholera, and is now chiefly used in the immunization of animals
for the production of antitoxic sera. It is considered fully in a
subsequent chapter.

26 PASSIVE IMMUNITY

PASSIVE IMMUNITY, the second form of acquired immunity, is
conferred by injecting into a susceptible animal the serum of one
which has acquired, an active immunity to the disease in question.
It is a kind of second-hand immunity, acquired in virtue of the
reception of substances actively formed by another animal which
has had to fight against the infecting agent in order to form them.
In its production there is no necessary illness, however slight.
Such may occur, it is true, but it is not more than would be pro-
duced by normal serum, and stands in no necessary relationship
to the development of the immunity. And when such illness does
occur, it does so after the production of the immunity, and may
be very severe when the protection given is but slight, and
vice versa.

For the production of passive immunity it is necessary to
inject the serum of an animal which has been artificially immun-
ized, that from one which is naturally immune being devoid
of action in this respect. To this general rule there are
one or two exceptions, which are perhaps more apparent than
real.

Passive immunity is sometimes called antitoxic. The term,
however, is not a good one, since there are several varieties of
passive immunity, only one of which is due to an antitoxin.

Passive immunity is specific — that is, the serum of an animal
which has acquired immunity against one organism will protect
a second against that, and against no other. In this, of course,
it resembles active immunity, but the two differ in several
important particulars.

i. As regards its production. Active immunity takes some
time— usually a week or so — to develop, dating from the infection
or injection of the vaccine, and in many cases at least its
appearance is preceded by a negative phase, in which the natural
immunity to the organism in question is lowered. But passive
immunity is established as soon as the serum has become mixed
with the blood of the person or animal injected, and there is no
negative phase.

Hence in severe infections our best hope in the way of specific
medication is in the production of passive immunity. It is but
recently that the injection of vaccines was thought of in face of
an infection already developed, and it is obvious that the method
will be useless or dangerous in very severe and rapid infective
processes. Passive immunity, on the other hand, can be induced

INTRODUCTORY AND GENERAL 27

at once and without a negative phase. Unfortunately, it is not
always or often possible.

2. As regards its duration. Active immunity lasts for a long
time, the length differing greatly in different diseases and after
various methods of induction. In many cases it lasts a year or
more. Passive immunity, on the other hand, is always of brief
duration, and lasts only about as long as the serum injected is

FIG. 2.— SHOWING THE SEQUENCE OF EVENTS IN THE PRODUCTION OF
ACTIVE IMMUNITY.

An injection of vaccine at b is followed by a decrease in the degree of
immunity (negative phase), a rise, and a gradual return to the normal
condition.

actually present in the blood. It depends to a certain extent on
the dose of serum given, and also on the species of animal from
which it was derived. An animal of a certain species is immu-
nized for a longer period by serum from another animal of the
same kind than from one of a different species. In general terms
the duration of passive immunity is three to six weeks. It is
renewable at pleasure, as far as we know indefinitely.

a

FIG. 3. — SHOWING THE SEQUENCE OF EVENTS IN PASSIVE IMMUNITY.
An injection of serum is given at b.

Hence passive immunity is chiefly of value to ward off* an
infection the danger to which is of short duration. Thus in
veterinary practice the passive immunity of horses conferred by
the injection of tetanus antitoxin is of the greatest possible value
before operations, or immediately after the infliction of a wound,
horses being so prone to tetanus that in some places any opera-

28 MIXED IMMUNITY

tion was a matter of great danger before the introduction of this
method.

Passive immunity is also useful as a basis for active immunity
This will be described under the heading of Mixed Immunity.

3. Passive immunity for a given bacterium or its products
cannot be made so potent as the active form for the same disease
in the same animal species. The reason is obvious : the passive
form only occurs in virtue of the presence in the blood of some of
the foreign serum, which can never form more than a fraction of
the whole fluid. The degree of the immunity may be sufficient for
all practical purposes, but can never reach the enormous height
met with in hypervaccinated animals.

4. Active immunity we believe to be developed to some extent
in all, or almost all, infections, but the production of passive
immunity is impossible in very many cases — e.g., tubercle (as far
as we know at present), infections with pyocyaneus, glanders,
malaria, and many other parasitic organisms. Perhaps in the
future we shall be able to procure active sera against all organisms,
but at present we have comparatively few of any value.

MIXED IMMUNITY is a combination of the two forms already
described, in which the dangers and delay incidental to the induc-
tion of active immunity are avoided by the use of a protective
serum. It is really a succession of the two forms, the passive
immunity being developed at once as a consequence of the injec-
tion of the serum, whilst the active form develops later in conse-
quence of the vaccination. The process is sometimes called
sero- vaccination.

It is not of great importance in human pathology, the chief
example being the form of typhoid inoculation suggested by
Besredka, and not yet used on a large scale. In it the killed
typhoid bacilli are submitted to the action of the immune serum,
from which they absorb certain protective substances and become
modified thereby. It is claimed that this treatment prevents the
development of the discomfort that follows the use of ordinary
typhoid vaccine, and that the immunity is developed very rapidly.
It may be followed by an injection of ordinary vaccine.

The method is used to a considerable extent by veterinary
surgeons, and there are several modifications in the process, the
serum being injected either mixed with the virus, or before, or
after, or simultaneously in different sides of the body. Thus in
the treatment of South African horse-sickness the virus (the blood

INTRODUCTORY AND GENERAL 29

of diseased animals) may be mixed with the serum from hyper-
immunized animals and injected subcutaneously. If the serum
and virus are injected separately the animal will in all cases
acquire passive immunity ; but unless there is some degree of ill-
ness (a " reaction") this will be but temporary, and no active
immunity will be superadded. Thus, if the serum be injected and
the virus given subcutaneously at the same time, no reaction
follows, and the immunity does not last more than a month ; but
if the injection is made into a vein a reaction occurs, and active
immunity, lasting for about a year, will follow (Stockman).

The method is also used in the early stages of antitoxin forma-
tion, the horse being treated with a mixture of toxin and anti-
toxin, the latter being in excess. But here it seems unquestionable
that active immunity is acquired, and the mechanism by which
this occurs is discussed subsequently.

FIG. 4. — MIXED IMMUNITY.
The presence of a negative phase, as shown in the diagram, is not essential.

LOCAL IMMUNITY.— We have hitherto spoken of the body as a
whole, assuming that all parts are equally resistant or susceptible.
This is not the case, and certain parts are found to have a marked
degree of immunity to certain bacteria. Here we have to be sure
that we are dealing with regions that are equally exposed to
infection. The stomach, for example, is comparatively rarely
attacked by infective processes, and this may be due to the fact
that the gastric juice is of a sufficient degree of acidity to kill or
inhibit most bacteria. Yet here it is probable that this does not
account for all the phenomena, and that some degree of true local
immunity does exist. Numerous other examples may be quoted.
Pneumococcic infections are common in the lungs and pleura, but
rarely spread further, and cause disease of the ribs and intercostal
muscles ; tubercle is common in the bones and extremely rare in
the muscles, whilst Trichina spimlis affects the muscles and never

30 LOCAL IMMUNITY

attacks the bones, and rarely any other tissues. Some of the best
examples may be taken from diseases that spread by continuity
from one tissue to another. Thus the gonococcus in either sex
spreads along the urethra with ease, but seldom involves the
mucous membrane of the bladder ; it practically never attacks the
vaginal mucosa (in adults), but spreads from the cervical to the
corporeal endometrium, and thence to the Fallopian tubes, but
comparatively rarely goes farther and produces a general peri-
tonitis. Diphtheria, too, though it may spread in any direction,
seldom creeps down the oesophagus. Many other examples might
be quoted.

There are marked differences in regard to local immunity
between the child and the adult. The most marked example,
perhaps, is in the almost perfect local immunity of the scalp to
ringworm in adults, which contrasts so markedly with the absolute
susceptibility of children, whereas the susceptibility of the skin of
the body to the same parasite is, if anything, greater in the former.
In most cases of differing immunity at different ages the child is
more susceptible, just as its resistance to general diseases is less,
and the few exceptions that may be quoted are perhaps rather
apparent than real.

Local immunity may be natural or acquired. Passive im-
munity, of course, cannot be local for long, as any serum which
is injected will rapidly diffuse away and be removed by the
lymphatics and blood-stream. The cases mentioned above are
all examples of natural local immunity. The difference between
the reactions of the tissues of children and adults do not neces-
sarily point to the acquisition of any active immunity in the sense
in which the word has been defined above, but rather to the
general rise in resisting power accompanying the general improve-
ment in strength and vitality, and in some cases, perhaps, to an
actual maturation of the tissues, as in the case of the adult vaginal
mucous membrane, which is immune to the gonococcus, whereas
the thin and immature infantile membrane is susceptible. The
immunity of the adult scalp to ringworm also is not acquired,
using the word in the narrow sense, for it occurs apart altogether
from an attack of the disease.

Our knowledge of acquired local immunity is very incomplete ;
it is a difficult subject for research, and more attention has been
paid to general immunity. A little consideration will demonstrate
the fact of its occurrence. For example, when a person develops

INTRODUCTORY AND GENERAL 3!

crops of boils it will often be found that one is undergoing involu-
tion whilst another is developing ; hence the cure of the first
cannot be due to any general immunity, but must depend on local
changes which do not affect the second. A similar line of argu-
ment will show the development of acquired immunity to the
streptococcus in erysipelas ; the healthy skin is susceptible, since
the disease spreads to it, but the process does not extend back-
ward into an area already affected, but now cured, or does so but
rarely.

The subject cannot be discussed further with advantage, and
will be deferred to a subsequent chapter, when the known factors
on which immunity depends have been elucidated.

There are important non-specific causes for alterations in local
immunity, as is the case with general. These practically resolve
themselves into the presence or absence of an adequate supply of
blood ; the more copious the supply of healthy circulating blood,
the greater the resistance to infections, and vice versa. Hence the
utility of fomentations and other hot applications in the initial
stages of an infective lesion ; hence, too, the application of Bier's
method of passive congestion, in which an excess of blood
(though partly stagnant) is made to flush the tissues. And there
is no doubt that the object of the dilatation of the vessels and
acceleration of the flow of blood through them which occurs in
the early stages of inflammation is a beneficial process which
has this improvement of the local resisting powers as one of its
objects, the influx of an increased number of leucocytes and the
dilution and removal of the soluble toxins being others. In acute
inflammation we may distinguish two stages. In the first, the
stage just mentioned, the conservative reaction of the vessels is
most obvious, and in the case of a mild infection, or if the
immunity is very strong, may suffice to destroy and remove the
infective material and its toxin. The stagnation and ultimate
cessation of the blood-flow are indications that the irritant is,
temporarily at least, getting the upper hand, and, by cutting off
the blood-supply, is neutralizing the most powerful defensive
factor. The acceleration of the flow may be regarded as physio-
logical, the retardation and cessation as pathological.

The causes of local reduction of immunity by obstruction of the
blood-stream are numerous, the most important being traumatism
(by injuring the vessels going to the region), endarteritis, throm-
bosis, tight bandaging, etc. They need not be discussed at

32 LOCAL IMMUNITY

length, but it is advisable to point out that severe traumatism, in
the form of violent laceration and contusion of a part, is an
extremely powerful predisposing agent, and that it acts in two
ways, or perhaps more. In the first place, there may be some
death of tissues, either in small or large amounts, and in these
dead tissues the natural resisting powers are of course in abeyance,
so that the bacteria will grow unchecked, as they would in dead
culture media ; and, secondly, that the blood does not reach this
dead material, and the leucocytes only do so with difficulty. The
importance of this is well seen in tetanus. The normal tissues
have a considerable degree of resistance to this organism, and
infection rarely takes place in a clean incised wound, even in
cases in which we can be almost certain that the spores of the
tetanus, bacillus have been introduced.

Another cause of reduced local immunity is the action of
irritants on the tissues. Here we must distinguish two cases.
If the irritant be but mild, it may be actually beneficial ; it
causes the earlier phenomena of inflammation which we have
previously referred to as being protective, and may tend to raise
the resistance of the part in consequence. Thus, according to
many observers (who do not agree precisely on the interpretation),
the injection of a small quantity of almost any bland (but never-
theless foreign) substance into the peritoneal cavity may protect
an animal against a lethal dose of a bacterial culture introduced
subsequently; normal saline solution, water, broth, serum, etc.,
all have this action. But if the irritant be more powerful, so that
the tissues are killed and the vessels occluded, or the leucocytes
killed, the susceptibility of the region is greatly increased.
Chemical antiseptics have this action, especially in certain
regions, such as the peritoneum. The same thing may be demon-
strated experimentally. Tetanus spores washed free of toxin
will not produce tetanus in rabbits, but will do so if an irritant
such as lactic or carbolic acid is injected simultaneously.

A few words may be said here on the phenomena of immunity
arid susceptibility in relation to the modifications they cause in
the infective processes. Where the immunity is great, or, as we
say, absolute, the result of an injection of the infective agent is
nil ; there is, of course, some degree of inflammation, but this
follows the injection of any fluid, even normal saline solution, and
the effect of the bacteria themselves is inappreciable. In this
case, therefore, the bacteria are immediately destroyed, and the

INTRODUCTORY AND GENERAL 33

substances which they produce are without deleterious effect on
the cells of the body. In another group of cases, referred to
above, the bacteria do not die, but their toxins remain harmless
to the host ; this is Ehrlich's immunitas non stevilisans, and it occurs
in the case of many of the lower animals which have in their blood
various protozoa (trypanosomes, etc.), without thereby suffering
the slightest appreciable injury. In man the condition is best
seen in its acquired form in the immunity possessed by negroes
to the action of malaria parasites, though the plasmodium may be
found in the blood. A closely allied phenomenon is in the latency
of bacteria. Thus a person may develop an attack of typhoid
osteitis years after an attack of typhoid fever, and we can only
assume that the bacteria have lain latent in his tissues for this
time ; in all probability they have been kept from infecting him as
a result of a sufficient degree of immunity, and when this breaks
down or wears off a renewed outburst occurs. The gonococcus
may be latent in a similar way for periods equally long. Another
similar phenomenon is the carriage of infection by persons who
remain themselves healthy. Diphtheria is a common example,
and it is no rarity to find a person in whose throat diphtheria
bacilli are present, but who remains unattacked. Here the
immunity suffices to prevent the bacillus from invading the body,
but not to destroy it.

At the opposite end of the scale occur those cases in which
immunity is practically absent. Here the result of the introduc-
tion of the bacteria is a rapid infection, both local and general,
with profound symptoms of intoxication ; the bacteria spread
through the tissues just as they would through a good culture
medium, and, in addition, invade the blood and multiply therein.
This is rarely seen in man, though some examples of septicaemic
plague and streptococcal septicaemia from post-mortem wounds
approach it closely. It can be produced experimentally in
animals, when large doses of virulent cultures are injected.
Death follows in a few hours, and the blood is found to be swarm-
ing with bacteria.

Between these two extremes come those cases in which the
introduction of the bacterium is followed by the production of a
local lesion. This always indicates some degree of local immunity,
and may be regarded as an attempt to localize the organism and
prevent its further spread. And the nature and severity of the
local lesion stand in close relation to the severity of the infection

3

34 THE LOCAL LESION

and the degree of the immunity. For example, in severe and
rapidly fatal infections from post-mortem wounds — i.e., where the
infection is virulent and the immunity but slight — there is very
little local reaction and very little glandular enlargement, the
process being septicaemic from the first. Where the infective and
protective forces are equally matched the local lesions are more
developed; inflammation, and usually suppuration, occur at the
site of the wound, and the glands enlarge and may suppurate ;
and when the infection is so feeble as to be quite unable to cope
with the immunity, the local lesion is the sole result of the infec-
tion. Eyre gives a similar example in the results of injecting
similar doses of pneumococci into rabbits of different ages. The
young animal is most susceptible, and in it death occurs within
forty-eight hours from septicaemia, and there is but little local
reaction. In half-grown animals the local lesion is more developed,
and is gelatinous or fibrinous, containing many leucocytes, and
the animal lives several days. In old rabbits quite definite pus is
formed, and the animal lives longer, and may recover completely.
Hence suppuration may be regarded as a proof that the defensive
and infecting forces are fairly balanced, and that either may be
victorious in the conflict.

The other local lesions need not be discussed at length, but the
case of tubercle and the allied diseases requires a brief notice.
Here the lesion indicates the presence of a very considerable
degree of immunity to the toxin, for the structure of a tubercle is
exactly similar to that of the cellular reaction to many feebly
irritating foreign bodies — e.g., unabsorbable ligatures, substances
from which it is clear no potent toxin can be given off; but it
also indicates that there is a defect in the mechanism by which
the bacilli should be removed, since the process is (for a time at
least) a progressive one. Here the walling-in of the infected area
which occurs as the result of the reaction of the tissues may be
taken to be a defensive process, but, as we shall have occasion to
see, it is one of doubtful utility.

EARLY THEORIES OF IMMUNITY. — Before turning to the dis-
cussion of the nature of immunity in the light of our present
knowledge, it will be convenient to insert a short account of some
of the early theories of the subject, which are in the main of
historic interest only. They have served their purpose as a point
of departure for subsequent research.

Of such nature was Pasteur's theory of exhaustion, the earliest

INTRODUCTORY AND GENERAL 35

attempt at a scientific explanation of the facts of recovery from,
and subsequent immunity to, the infectious diseases. Pasteur was
a chemist, and was only led to the study of bacteriology by the
pursuit of chemical investigations into examining reactions which
he proved to be due to micro-organisms. His theory was a
chemical one. A certain amount of food is necessary for each
bacterium, and when the total amount contained in a given solu-
tion is used up the growth of the bacteria must cease. For
example, if we take a dilute solution of sugar (containing the
necessary salts, etc.), and inoculate it with yeast, the cells will
begin to divide and multiply with great rapidity. After a time
the growth ceases, and it will not be resumed if we inoculate the
fluid with an additional amount of yeast. We may compare the
test-tube to the patient, the yeast to the pathogenic organism, and
the process of fermentation to the disease, and we may say that
the fluid has recovered from the disease and is now immune to it.
This immunity depends upon the absence of sugar, which was
used up by the yeast cells, and if more sugar be added the process
of fermentation may be restarted by a fresh inoculation, or by the
yeast still remaining.

The theory can easily be disproved, from the fact that bacteria
may grow well enough in the dead tissues and fluids of immune
animals ; and, secondly, because immunity, as we have seen, may
be produced (in some cases) by the injection of the chemical pro-
ducts of the bacteria, substances which can hardly use up food
materials. The theory has, however, been recently revived in a
modified form by Ehrlich, who considers that there is sufficient
evidence for the occurrence of this form of immunity in certain
cases. He calls it atreptic immunity.

The retention hypothesis of Chauveau is the exact opposite of
Pasteur's. Several observers showed that the growth of micro-
organisms in fluid media might cease spontaneously whilst
abundant food material remained unutilized. This was found to
be due to the presence of certain products of metabolism, which,
like carbon dioxide in the case of animals, act as poisons to the
organism which produces them. For instance, the fermentation
of sugar by yeast is found to cease when about 14 per cent, of
alcohol is present, and if a strong solution be taken the process
will stop at this point, but can be started again if the alcohol be
removed by distillation. Here the fermentation is stopped by
alcohol, a product of metabolism of the yeast cell, which acts as a

3—2

36 THEORY OF RETENTION

poison on the organism producing it. The theory of immunity
based on these facts is obvious. Bacteria growing in the body
will yield substances inimical to the continued growth of the
organism, so that they will die out and recovery ensue, and the
body will remain immune as long as these substances are retained
therein. This theory accounts well for the production of immunity
by injections of the toxins and other soluble products of bacteria.
It is negatived by the fact that bacteria may grow in the blood and
tissues of immune animals, and is improbable if we consider that
immunity may last for many years, and that it is extremely
improbable that substances (necessarily soluble) should be retained
in the body for so long a time.

We shall now proceed to a study of the more modern views,
and in doing so it will be convenient to deal with the subjects of
immunity to toxins and immunity to bacteria in separate sections.
Of course, in most cases they run parallel to one another: an
animal contracts a disease because its fluids and tissues cannot
kill the pathogenic bacteria offhand, and because its cells are sus-
ceptible to the action of the toxin, and vice versa. This is not
necessarily the case, however, and the two phenomena may be
entirely independent.

The subject of immunity of toxins is on the whole the more
important of the two, the simpler (though complex enough), and
the best understood. It will be best to deal with it first.

CHAPTER II
ON THE NATURE OF TOXINS

THE fact that the pathogenic action of any organism is dependent
entirely, or almost entirely, on that of the ^toxins which it pro-
duces renders it necessary to make a brief study of these
substances before considering the method in which the infected
animal reacts to the organism, and defends itself against infection.
In doing so we must distinguish clearly between the specific
toxins which are produced by any organism and the non-specific
and less important poisons which it may also elaborate. The
difference is a fundamental one. Numerous bacteria produce
by-products of metabolism, excreta, etc., which are comparatively
simple chemical substances of definite composition ; for example,
acids, alkalis, alcohol, ptomains, nitrites, etc. These may be
poisonous, and may, in some cases at least, play a part of some
importance in the production of the symptoms of the disease.
The cholera vibrio, for instance, produces nitrites in considerable
amount, and since the symptoms of cholera have some resem-
blance to those of nitrite poisoning, it is conceivable that those
substances may be, to some extent at least, the active causes of
the disease, and these nitrites might be regarded as the toxins of
the cholera vibrio. This, however, is not the case, and the true
toxins are quite different in nature, as is shown by many facts,
especially by the proof that cholera vibrios which have no longer
the power of producing nitrites may still cause infection in
susceptible animals.

The specific bacterial toxins differ from these poisonous sub-
stances in many important particulars. They are, as a rule,
formed only in very small amounts, and are extremely powerful.
For example, the toxin of tetanus may readily be obtained in so
poisonous a solution that ToV TF c-c- w*ll kill a guinea-pig in a day
or two, and of this solution only a very small fraction even of
the dried residue consists of toxin. They are not simple chemical

37

3 TOXINS — THEIR FRAGILITY

substances, and their exact nature is as yet unknown. This may
be due in part to the minute amounts which are formed, and in
part to the difficulties which prevent their being obtained in a
pure state ; but there are other reasons, to which we shall revert
later, for this complexity. Further, they are, with a few ex-
ceptions, very fragile substances, and are readily destroyed by
the action of many agents, and especially by heat. Nearly all
the bacterial toxins are rendered inert by boiling, and many of
them by a short exposure to a temperature of 60° or 70° C.
They are usually destroyed by gastric digestion, so that they are
without action when administered by the mouth.

A considerable amount of attention has been paid to this
question, since it would be desirable, if possible, to replace hypo-
dermic injections of vaccines, etc., by oral or rectal administration.
In general terms the statement made above holds good : toxins
administered by the mouth are not absorbed as such, and do not
produce the characteristic symptoms of the disease. In some
cases, however, there is reason to believe that a small amount
of absorption, probably of the toxin in an altered form, does
occur, and a certain degree of immunity may be produced by
the oral administration of killed cultures of typhoid bacilli, and
possibly of tubercle bacilli. But this method has only one advan-
tage— its painlessness — over the hypodermic method, whereas its
uncertainty renders it extremely undesirable. There can be no
doubt that the advantage of giving an exactly measured dose, with
the certainty that every particle will be absorbed and act in the
way desired, will, under ordinary circumstances, render the hypo-
dermic method infinitely preferable. To administer infinitesimal
doses of killed tubercle bacilli or of TR to an infant who may be
swallowing large doses of living and dead bacilli in milk, sputum,
etc., does not appear rational, and the clinical evidence in its favour
is entirely unconclusive. In the case of ricin, about a hundredth
part of the toxin given by the mouth is absorbed as such — i.e.t the
minimal lethal dose on oral administration is about 100 times
as large as the lethal dose of the same preparation given sub-
cutaneously (Stillmarck). Ricin is, however, far more resistant to
the action of digestive enzymes than are the exotoxins.

The most important feature of the bacterial toxins is their
relation to immunity. It is possible in all cases to render a
susceptible animal immune to their action by the injection of the
toxins in suitable doses at suitable intervals, though in some

ON THE NATURE OF TOXINS 39

cases the task is a difficult one. This is not the case with the
non-specific toxins. It is true that in a few isolated instances we
are able to increase slightly the resistance of an animal to the
simple chemical poisons (e.g., to alkaloids such as morphine), but
these apparent exceptions hardly interfere with the utility of the
general rule. Further, and more important, an animal immunized
to the action of a toxin is also protected against the pathogenic action
of the bacterium which produces it, and vice versa. Thus an animal
which has been rendered immune to the toxin of tetanus by re-
peated injections of that substance is also immune to infection
with the living cultures of the bacillus, and an animal which has
successfully survived an infection with the tetanus bacillus is
thereby rendered in some degree immune to the action of tetanus
toxin. This method of immunization with the bacterial toxins
(the so-called " chemical vaccination ") is of the utmost impor-
tance in practice. It was introduced by Smith and Salmon, who
showed that it was possible to immunize pigeons against living
cultures of the hog-cholera bacillus by means of the sterilized
products of that organism.

When this method is applicable it supplies us with a test as to
the specificity of a toxic substance which we have isolated from a
culture of a bacterium, or from the organs of an animal which
has been killed by an infection. The substance must be poisonous
for animals which are susceptible to the infection in question, and
it must be harmless to animals which have been immunized to
the organism ; on the other hand, it must immunize animals both
to its own action and to that of the bacterium when injected in
a living state. These conditions are never fulfilled by the non-
specific toxins.

There are a few apparent exceptions to this rule, but they fail
to stand investigation, being based on the fact that it is easier to
render an animal refractory to a living organism than to its toxin.
Thus an animal which has been injected with the filtered products
of certain organisms may be rendered immune to infection with
those organisms, but remain as susceptible as before to their
toxins. But this is due to the fact that the animal has been
immunized but partially ; if the process be carried further the
animal will be rendered refractory to both.

Again, an animal which has been immunized to the toxin of
one bacterium remains as susceptible as before to the action of
another toxin or bacterium. A horse which has been immunized

40 THE EXOTOXINS

to diphtheria toxin (e.g., in the production of diphtheria antitoxin)
will be just as susceptible to tetanus toxin as a normal animal.
In a very few cases the law does not hold. The only well-
authenticated example of this sort is the antagonism which
animals display to anthrax after injection with the products of
B. pyocyaneus.

These preliminary considerations will serve to show the more
important criteria by which the nature of a bacterial product may
be determined, and its nature as a true toxin established.

These toxins were soon found to fall into two main groups — the
extracellular or soluble toxins, or, as we shall call them, the exo-
toxins, and the intracellular insoluble toxins, or endotoxins. We
shall consider these substances in turn.

THE EXOTOXINS.

The exotoxim are substances which are given off in a free state
when the bacteria are grown in a suitable culture medium outside
the body, and can usually be separated by simple filtration (through
a Pasteur or Berkefeld filter) from the organisms which produce
them. We may consider them provisionally as the specific secre-
tions or excretions of the bacteria. They are not formed by all
pathogenic bacteria — that is, in the present state of bacteriological
science no suitable culture media have been found in which certain
organisms will produce a soluble toxin. The three most impor-
tant organisms which do so are the B. tetani, B. diphtheria,
and the B. botulismus. These toxins, the first two especially,
are substances of the greatest interest, since they have been
submitted to a most profound examination, and our knowledge
of the structure of bacterial toxins, of their action on the body,
and of the production of immunity thereto, is based almost entirely
on the results thus obtained. In addition to these, there are sub-
stances which are much less toxic — if, indeed, toxic at all — and
which fail to fulfil our definitions of a specific toxin, since an
animal which has been immunized thereto is not necessarily
immune to the organism, but which have many points in common
with the true toxins, and will be considered in this connection.
These are the bacterial cytolysins and hsemolysins, T substances

1 Haemolysis, or the liberation of haemoglobin from red blood-corpuscles,
may be brought about by a variety of agents, which fall under three main
headings : (i) Simple chemical substances, such as distilled water, ether,
acids, etc., which act osmotically, or by a direct solution of the strorna of

ON THE NATURE OF TOXINS 41

which have the power of dissolving living cells or red blood-
corpuscles respectively from susceptible animals.

In dealing with these substances we will consider firstly their
action, secondly their structure, and thirdly what has been estab-
lished concerning their chemical relationships with other sub-
stances. The last is comparatively unimportant.

i. Action of Toxins. — The results of the injection of a toxin
into a living and susceptible animal depend, in most instances,
on the dose injected. If, for instance, we inject a large amount of
the filtered broth in which the tetanus bacillus has been growing
for a month or so, and which in consequence contains tetanus
toxin, the animal (a guinea-pig, for example) will develop the
rigidities, spasmodic contractions of the muscles, etc., charac-
teristic of tetanus ; and these make their appearance after an
interval of some hours, during which period the animal shows
no symptoms whatever of the disease. Great stress was laid
at one time on the occurrence of this " latent period," since it was
thought to be peculiar to the bacterial toxins (and to the similar
substances of animal and vegetable origin), and to distinguish
them sharply from other poisons, alkaloids, etc. This is hardly
correct. It is true that in most cases of intoxication by bacterial
toxins there is a latent period, but in a few it is practically absent
The most interesting example is the " Nasik " vibrio, an organism
allied to that of cholera. This produces an exotoxin (though not
a very powerful one in the sense that it kills in small doses),
which proves fatal on intravenous injection into a rabbit after a
period of ten to thirty minutes, and symptoms are produced before
this. On the other hand, some of the alkaloids, and notably
colchicine, display a well-marked latent period. The phenomenon,
therefore, is not absolutely peculiar to, nor characteristic of, the
toxins ; but since it is so commonly displayed by them, it calls for
some investigation. Moreover, we must assume that part at
least of the incubation period of an infective disease is taken up
by the latent period of the bacterial toxin, a circumstance which
invests it with especial interest. Thus a horse which Madsen

the corpuscles or of parts thereof; (2) the simple organic haemolysins, which
include the bacterial haemolysins dealt with above, the haemolysins of vegetable
origin (such as ricin, etc.), and some of the haemolysins of animal origin ;
and (3) the compound haemolysins, all of animal origin, which will be dealt
with subsequently. These groups differ profoundly in their action, and must
be kept quite distinct.

42 ACTIONS OF TOXINS

was treating for the production of diphtheria antitoxin developed
tetanus, and tetanus toxin was found in a sample of blood collected
five days before the development of symptoms.

On diminishing the amount of the toxin which we inject, we
find that the latent period becomes gradually longer, and the
duration of the disease (i.e., the time between the first develop-
ment of symptoms of intoxication and the fatal issue) also lengthens.
By diminishing the dose gradually we can find an amount which
will just kill the animal in question in a given number of days,
and, provided the test animals used are approximately the same
in age and weight, we shall find that this amount, the " minimal
lethal dose," is fairly constant for animals of the same species.
Thus, in the standardization of diphtheria antitoxin the first step
is the estimation of the minimal lethal dose of the toxin, and for
this purpose it is customary to use guinea-pigs weighing from
250 to 280 grammes, and to fix a time-limit of four days. It is
found that the minimum lethal dose is the same, within close
limits, for all test animals, and that if a series similar in size and
weight be inoculated with the same dose, the majority will die
within a few hours of one another. This fact enables diphtheria
antitoxin to be titrated with some approach to chemical accuracy,
the test guinea-pig being used as the indicator.

On reducing the dose still further, we find that the incubation
period is still further prolonged, that the symptoms are less severe,
and that death may not take place, or only do so at a later period
than that which has been fixed for the minimal lethal dose. Thus,
in antitoxin-testing a dose of toxin which does not kill in five days
is regarded as a sublethal dose, although death may take place at
a later date — perhaps much later.

On giving still smaller doses the symptoms take still longer to
develop, are still slighter, and are followed by recovery, and the
animal may then present a certain degree of immunity to the toxin
and to the organism producing it. On the other hand, under
certain circumstances it may be more than usually sensitive to
the action of the toxin in question.

These phenomena present some points of comparison with
those which are presented in the action of the soluble enzymes,
such as pepsin. In each case an excessively minute amount of
the active substance will produce the given effect, and in each
the effect is more rapid if a larger amount be used. In either
case there is a latent period of longer or shorter duration before

ON THE NATURE OF TOXINS 43

the peculiar chemical action is manifested. There are several
other analogies between the soluble enzymes and the exotoxins.

(a) The soluble enzymes are, without exception, all produced
by living animal or vegetable cells, and are either secreted or
excreted by them, or remain locked in their protoplasm. The
bacterial toxins, in the same way, are all formed and eliminated
by living bacteria ; or, in the case of the endotoxins, retained
in the cell. In other words, both extracellular enzymes and
exotoxins are products of metabolism given off during the life of
a living organism. Further, both substances represent a method
in which the organism attempts to modify its environment and
render it more suitable : the animal secretes pepsin into its stomach
in order to modify the ingested proteids and render them suitable
for food, and the tetanus bacillus produces toxin in a living animal
because it is in itself but little adapted to grow in living tissues,
but can do so easily when these tissues have been injured by the
action of toxin. The spores of tetanus which have been washed
free from all traces of toxin have no power of producing tetanus
when injected into an animal, and are rapidly taken up by the
leucocytes, or otherwise dealt with by the tissues ; but if a minute
amount of toxin be injected at the same time the bacteria can
resist the leucocytes and tissues, which are injured thereby, and
continue to grow and produce fresh toxin, giving rise to fatal
tetanus.

(b) It is capable of proof that enzymes commence their action
on the substances which they attack by forming a combination
therewith. Thus the first effect of the addition of pepsin to fibrin
is the formation of a compound between the two substances, as
shown by the fact that, if the fibrin be thoroughly washed at
a temperature near the freezing-point until all traces of free
enzyme are washed away, it will still undergo digestion when
raised to the body temperature. Further, the enzyme is less
easily destroyed by heat when it has combined with the fibrin.

In a similar way it is capable of proof that the toxins unite
chemically with the cells of susceptible animals. The proof may
be deduced from the fact that if toxin be injected intravenously
into a susceptible animal it rapidly disappears from the blood,
although it does not escape, or only to a very small extent, in the
secretions. When the injection is made into insusceptible animals
it may disappear by a process to be discussed subsequently, or
may persist for long periods. Thus in one case Metchnikoff

44 COMBINATION OF TOXINS AND TISSUES

was able to demonstrate the presence of tetanus toxin in the
tortoise, which is insusceptible to the action of that substance,
at a period of four months after the injection. That the disap-
pearance which occurs in susceptible animals is actually due to
a combination of the toxin with the tissues of the body, and not
to its destruction or elimination, is shown by the fact that the
tissus of an animal which has been injected with tetanus toxin,
but which no longer contains that substance in the blood, may
produce tetanus when injected into a susceptible animal. In the
case of fowls it seems that this power of combining with tetanus
toxin is most marked in the leucocytes. Again, it is possible
to reproduce the absorption of tetanus toxin by fresh tissues in
vitro. This has been especially studied by Ignowtowsky, who
showed that emulsions of liver, kidney, spleen, etc., have the
power to absorb tetanus toxin, but that the subsequent injection
of these cells will produce the symptoms of the disease.

It ought to follow logically that the toxin will combine especially
with those cells and tissues which are acted upon by it in the
living body, and in all probability this is the case. The proof,
however, is somewhat difficult. Wassermann apparently proved
the point by his demonstration of the fact that tetanus toxin is
absorbed and neutralized by an emulsion of the central nervous
system, and not by that of any other organ, although, as has
been mentioned above, it is absorbed by other tissues. Now,
tetanus toxin acts entirely, or almost entirely, on the central
nervous system, and this well-known and oft-quoted experiment
appears to constitute a proof of the point at issue. The exact
interpretation of Wassermann's experiment appears, however, to
be doubtful, and it is hardly safe to rely on it as a proof of the
point.

With the bacterial haemolysins, which, although of feeble
toxicity, are in every other respect identical with the exotoxins,
we are on surer ground. A filtered broth culture of the tetanus
bacillus contains the specific toxin (tetanospasmin), and in addi-
tion a second substance, which has the power of dissolving red
blood-corpuscles when kept at a temperature near that of the
body. At a low temperature they do not act in this way ; but if
red corpuscles be added in suitable amount to a solution of tetano-
lysin at a temperature of o° C. and centrifugalized, the supernatant
fluid has no longer the power of producing haemolysis. On the
other hand, the red corpuscles, even after washing with normal

ON THE NATURE OF TOXINS 45

saline solution to remove all traces of free haemolysin, are dissolved
when raised to the body temperature. In other words, the
specific haemolysin of tetanus can form a combination with the
structures on which they act. Numerous similar examples will
be met with.

(c) In some of the specific exotoxins, notably that of tetanus, we
meet with a similar dependence on a suitable temperature for the
development of their toxic action, a property in which again
they resemble the soluble enzymes. The most striking example
is obtained by a study of the action of tetanus toxin on the frog,
which, in common with all cold-blooded animals, is but slightly
susceptible to its action. If, however, the frogs be kept in an
elevated temperature — 30° C. or higher — they develop the typical
symptoms of the disease after five days or thereabouts. Now
Morgenroth has shown that the toxin unites with the central
nervous system at a low temperature (8° C.), but without the de-
velopment of symptoms. For the production of these a high tem-
perature is necessary, exactly as in the case of the combination
of tetanolysin with red blood-corpuscles and the solution of the
latter.

(d) These and similar researches lead us to distinguish between
two faculties of a toxin — that of combining and that of injuring ;
and the fact that in some instances these processes can take place
at different temperatures leads us to the belief that they are quite
different properties of the toxin. In other words, the mere union
of a toxin with a cell is not sufficient to cause injury to the latter.
This is susceptible of proof. In Ehrlich's elaborate studies on the
standardization of diphtheria antitoxin he first obtained a speci-
men of diphtheria antitoxin, and determined its minimum lethal
dose for test guinea-pigs. For the sake of simplicity we will
suppose that for a given sample of toxin this was T^ c.c. — i.e.9
that amount of the filtered broth culture of the diphtheria bacillus
would just kill a guinea-pig weighing 250 grammes in four days.
Further, let us suppose that we have a standard sample of anti-
toxin of which i c.c. just neutralizes i c.c. of toxin (100 lethal
doses), so that the mixture of the two causes no symptoms when
injected into a test animal. Diphtheria antitoxin is a relatively
stable substance, and can be preserved in a dry state, at a low
temperature, for long periods if light and air are excluded. It is
thus possible to re-test the sample of toxin with a precisely similar
solution of antitoxin after some months. When this is done, it is

46 TOXINS AND TOXOIDS

found that it will have fallen off in potency ; for example, it may
take gL c.c. to kill a guinea-pig. It might be supposed that this
was due to a complete destruction of half the toxin, but this is not
the case. If it were so, we should find that to neutralize i c.c.
( = 50 lethal doses) we should require \ c.c. of antitoxin, since the
latter has not altered in potency. As a matter of fact, we find
that we still require i c.c. of antitoxin ; in other words, the
diminution of the toxic power of the solution has not been accom-
panied by a diminution in its combining capacity for antitoxin.
The explanation given by Ehrlich, and fully proved by analogy
with numerous other similar phenomena, is that part of the toxin
has altered into a substance which retains its power of uniting
with antitoxin (and, as we shall show later, with the tissue cells),
but which has been deprived of its toxicity. Toxin which has
undergone this change is called toxoid. Haemolysin also appears
to undergo a similar change into haemolysoid, and the rapid loss of

. #<

I J

FIG. 5.— A MOLECULE OF TOXIN WITH ITS HAPTOPHORE (a) AND
TOXOPHORE (6) GROUPS.

On the right a similar molecule, which has lost its toxophore group, and
become converted into toxoid.

activity which tetanolysin undergoes is very probably due to a
change into that substance.

The alteration of the toxin to toxoid can be best explained
by supposing that the power of entering into combination and the
power of intoxication reside in two different parts — which we may
regard as groups of atoms — of the molecule of toxin, and by
further supposing that the combining group is a relatively stable
one, and that the toxic group is easily destroyed. In the very
convenient nomenclature introduced by Ehrlich, and now uni-
versally adopted, the group of atoms which has the power of
entering into chemical combination with the living protoplas
or with antitoxin is called the haptophore grotip, whilst the portion
on which the toxic action depends is called the toxophore group.
The change of toxin into toxoid, or of haemolysin into haemolysoid,
consists in a destruction of the toxophore group, with retention of
the more stable haptophore group (Fig. 5.) From what has been
said as to the dependence of the phenomena of intoxication on a

ON THE NATURE OF TOXINS 47

temperature approaching that of the body, it follows that the hapto-
phore group can functionate at a low temperature (o° to 10° C.),
while the toxophore group can only do so at a fairly high one.
Looked at in this way, the process of intoxication with an endo-
toxin, or of haemolysis with a bacterial haemolysin, may be divided
into two stages : in the first place, the haptophore group of the
toxin or haemolysin combines with the protoplasm or with the
stroma of the red corpuscle, and in the second the toxophore
group exerts its action, and the cell is poisoned or the red corpuscle
dissolved. The phenomena of tetanus in frogs is thus readily
explicable.

We shall see several other examples of substances in which it
is possible to distinguish between a combining and an active
group, and the same terminology will be adopted throughout
(agglutinoids, complementoids, etc.).

The change of toxin into toxoid takes place in all exotoxins,
but at very different rapidities. Tetanolysin is transformed com-
pletely into haemolysoid in a day or so, whilst tetanospasmin, the
true toxin of the disease, is much more stable. The process is
accelerated by heat, light, and the access of oxygen, and by
certain chemical substances which are not sufficiently powerful
to destroy the toxin outright. Of these the most important are
a solution of iodine in iodide of potassium, and bisulphide of
carbon.

The exotoxins are destroyed outright by heating to the boiling-
point (to this rule there are a few exceptions, none of which has
been fully examined), by strong acids and alkalis, and by the
action of the digestive enzymes. They are, as a rule, precipitated
by the substances which precipitate proteids, and destroyed by
the substances that destroy those bodies. They have, further,
the power of attaching themselves to precipitates, of whatever
nature, which are thrown down in fluids containing them ; so that
formerly they were thought to be albumins, albumoses, nucleo-
albumins, etc., since they were carried down mechanically when
these substances were precipitated from a bacterial culture in
which they were present along with the exotoxin. In these points
again they closely resemble the enzymes.

They are substances the molecules of which must be small in
comparison with those of the coagulable proteids, since they
readily pass through filters (of unglazed porcelain permeated with
gelatin) which retain the latter. This fact was put to an ingenious

48 TOXINS — ANALOGIES WITH ENZYMES

use by Martin and Cherry in their demonstration that diphtheria
toxin and antitoxin combine chemically.

Enzymes are also substances of small molecule, and pass
through similar niters. When injected into suitable animals
enzymes give rise to the production of anti-enzymes, which are
exactly equivalent to antitoxins. Thus we see that in many
points the process of intoxication with the bacterial exotoxins
presents close analogies with the destruction of proteids, etc., by
enzymes; and to these we might add the suggestion that it is
very probable that these exotoxins act, partly at least, by a
process of hydrolysis. This suggestion is based partially on the
fact that the process of haemolysis is almost certainly one of
hydrolysis, and partially on the appearance of poisoned cells,
which look as if they had absorbed water and became partly
dissolved.

There is, however, one feature in which the exotoxins and
their allies, the bacterial haemolysins, are absolutely different from
the enzymes. In the case of the enzymes a molecule attaches
itself to the substance to be attacked, water is absorbed, and the
whole complex molecule breaks down ; and in this process the
molecule of enzyme is set free, and is again ready to attack
another molecule. Thus a very small amount of the active
substance can decompose a large amount of fermentable substance.
The toxins do not behave in this way, and, as far as we know,
a molecule of toxin which has united with one molecule of proto-
plasm is never set free to attack another.1 The proof of this is
not very direct, and rests mainly on the fact that the amount of
toxin necessary to kill two animals of the same species varies roughly
with their weight. Thus the minimal lethal dose of diphtheria
toxin for a guinea-pig of 250 grammes will not kill one of 400.
If the molecule of toxin could attack one molecule of cell
substance after another in the same way as an enzyme, we should
expect it to do so, though after a longer interval. It must be
confessed, however, that this proof is not very striking ; excep-
tions frequently occur, since, as a rule, older animals are less
susceptible than younger ones in proportion to the body-weight.
But it is certainly true with regard to the bacterial haemolysins,
since we can test them on the same sample of blood, and when

1 It may possibly undergo dissociation, and be set free to attack another
molecule, but this is a different process : the molecule first attacked is not
injured.

ON THE NATURE OF TOXINS 4Q

this is done we find they obey the law of multiple proportions
with great accuracy. Thus the exotoxins differ from the enzymes
mainly in the fact that each molecule of the former acts once, and
once only. We shall subsequently meet another group of sub-
stances, of very similar nature but of animal origin, which have
an enzyme-like action, but are destroyed in the process. They
are the complements (alexins, etc.), which resemble the exotoxins
in many respects, and might well be called the animal toxins.

On investigating more closely the action of the exotoxins, we
find that certain of them exert their pathogenic action mainly on
certain cells of the body. The most marked example of this is
in tetanus, which practically only affects the cells of the central
nervous system, causing in them definite histological changes,
and having a pharmacological action almost exactly like that of
strychnine. In the case of diphtheria also the action is most
marked on these cells ; this is best shown by the occurrence of
diphtheritic paralysis (associated with histological changes in the
ganglion cells similar to those of tetanus, and subsequent degene-
ration of the nerves), which occurs after the action of minute
doses of the toxin.1 We may fairly assume that when but an
excessively small amount of toxin is present, it will unite with
the cells with which it has most affinity — in this case with those
of the central nervous ganglia. But diphtheria toxin is not
limited in its action, as tetanus toxin is, and can act upon the
tissue cells almost without exception. Thus we find that the
injection of a large dose of toxin subcutaneously is followed by
the production of an acute inflammatory swelling, showing that
it can poison the connective tissues, and after death there may be
focal necrosis of the liver, degenerative changes in the renal
epithelium, fatty degeneration of the heart, etc., showing that
the toxin may act on all these organs and tissues. We may
regard it as a good example of a general protoplasmic toxin
having, as is so frequently the case, a special action on certain
cells. The toxins of most diseases come under this heading, the
specialized action of the tetanus toxin being unique.

In some cases we can study the action of the exotoxins and
allied substances on isolated cells in vitro, and these are of especial
interest from the ease with which they can be investigated, and
are of some importance in disease. They are the leucolysins, or
leucotoxins, and the haemolysins.

1 If we accept Arrhenius's view of the interaction of toxin and antitoxin.

4

50 THE BACTERIAL LEUCOLYSINS

The leucolysins are substances which are formed by bacteria,
and which have the power of killing and dissolving, or partially
dissolving, the leucocytes of susceptible animals. Owing to the
comparative difficulty of obtaining emulsions of living leucocytes,
they have not been submitted to the same thorough examination
as have been the bacterial haemolysins ; but the important rela-
tions between the leucocytes and immunity lead us to*believe that
they are of very considerable pathological interest. The first
to be described was that formed by the Streptococcus pyogenes, the
action of which on the living leucocytes was shown by an ingenious
experiment to occur in vitro, and to be neutralized by means of
antileucolysin, this being one of the earliest proofs that toxin and
antitoxin form a chemical combination, and that the preventive
and curative effects of the latter are not due to some profound
influence on the tissues of the living body, by which they are
rendered immune before the toxin can attack them. The method,
invented by Neisser and Wechsberg, is based on the fact that
living leucocytes have the power of deoxidizing and bleaching a
solution of methylene blue. When a solution of the products of
growth of streptococci is added to an emulsion of living leuco-
cytes, together with a little of the dye, and a layer of liquid
paraffin added to prevent the further access of air and subsequent
oxidation of the methylene blue, the colour no longer disappears,
showing that the leucocytes have been killed. If, however, a
suitable amount of antileucolysin (obtained by injecting the
filtered products of the streptococci into an animal) be added to
the mixture the colour disappears, showing that the leucocytes
have been protected from the action of the leucolysin, which has
now been neutralized by the serum.

The action of the leucolysins can also be studied microscopi-
cally in vitro, when the cells are seen to become more transparent,
and their nuclei to become more indistinct, and ultimately to
disappear. The dissolving leucocytes look very much like those
found in pus.

Leucolysins are formed by the Streptococcus pyogenes, the staphy-
lococcus, and B. pyocyaneus, and probably by other organisms.

The bacterial hsemolysins are an interesting group of substances
which are closely allied to the exotoxins in their reactions, but
are little toxic, if at all. The most toxic appears to be that of
Streptococcus pyogenes, to which some observers, though not all,
attribute feeble poisonous powers when injected into animals.

ON THE NATURE OF TOXINS 51

At the same time, it is quite certain that these substances play
some part in the production of the symptoms of various diseases.
The anaemia which develops so rapidly in acute sepsis is well
known, and is one of the most constant symptoms of that affec-
tion ; it is to be ascribed, in part at least, to the destruction of
the red corpuscles by the haemolysins elaborated by the strepto-
cocci, staphylococci, or colon bacillus, if these happen to be the
infective organisms. The blood of an animal which has been
injected with virulent streptococci is found to contain haemolysin,
and that this is actually the haemolysin produced by the strepto-
coccus is shown by the fact that the action of this serum is
restrained by the addition of serum from an animal treated by
injections of streptococcic haemolysins. Thus it is proved that this
organism elaborates its haemolysin in vivo as well as in vitro ; and
several observers have found that it is those species of strepto-
coccus which are specially virulent to animals and man that
form haemolysins, the harmless ones doing so to a small extent,
if at all. The same is true for staphylolysins. Further, when
a culture of streptococci which is but slightly virulent and forms
but little haemolysin is rendered more virulent by " passage "
through rabbits, its power of forming streptocolysin is increased.

These facts render it certain that some at least of the bac-
terial haemolysins act, to some extent, as exotoxins, though the
organisms producing them certainly form other and more im-
portant specific poisons. We may consider them as accessory
toxins of comparatively little pathological importance.

The similarity in nature of the bacterial haemolysins and the
specific exotoxins is shown by the fact that (in the case of strep-
tocolysin, and probably in others) they can become converted into
htzmolysoids, analogous to toxoids. This is shown as follows :
Streptocolysin becomes inert in a week. If a small quantity of
blood-corpuscles be added to an excess of this inert solution, and
then thoroughly washed and added to a fresh and active solution
of streptocolysins, they will not be dissolved ; the corpuscles had
evidently become saturated with inert haemolysoid, and are now
unable to take up any haemolysin, their combining powers being
satisfied (see Fig. 6).

The chief bacterial haemolysins are those formed by the tetanus
bacillus, the staphylococcus, the Streptococcus pyogenes, the B. pyo-
cyaneus, B. coli, and the typhoid bacillus. Their more important
features will be recapitulated briefly.

4—2

TETANOLYSIN AND STAPHYLOLYSIN

Tetanolysin is formed along with the specific toxin, tetano-
spasmin, when the B. tetani is grown in broth, the two substances
being formed in variable amounts under different circumstances.
It is very unstable, disappearing entirely in a day or two at the
room temperature, and being destroyed by heating to 50° C. for
twenty minutes. It cannot be obtained free from tetanospasmin,
but a solution of tetanus toxin can be deprived of its lysin, and
only the specific toxin left, by adding some red corpuscles to the
solution, kept at a low temperature, and centrifugalizing them

FIG. 6. — A "SATURATION EXPERIMENT" SHOWING THAT H^MOLYSOID HAS
THE POWER OF COMBINING WITH RED BLOOD-CORPUSCLES, AND SHIELD-
ING THEM FROM THE ACTION OF H^MOLYSIN. (SCHEMATIC.)

In the first tube the corpuscles are shown in presence of an excess of old
or heated haemolysin ; in the second they are washed clear from this
excess, and are apparently unaltered ; in the third active haemolysin is
added, but the corpuscles are not dissolved, as they would be in a control-
tube with normal corpuscles.

off; the supernatural fluid will contain tetanospasmin, whilst the
tetanolysin will have combined with the corpuscles.

Staphylolysin appears in alkaline cultures on the fourth day,
and reaches its maximum between the tenth and twelfth. It is
an unstable substance, but more stable than tetanolysin, persist-
ing for a fortnight at the room temperature, and requiring a
temperature of 56° C. for twenty minutes for its complete destruc-
tion— and in this case the destruction appears to be really com-
plete, for the injection of the heated solution is said not to lead
to the production of an antistaphylolysin. Many normal sera,
especially those of man and the horse, contain antistaphylolysin ;

ON THE NATURE OF TOXINS 53

perhaps this is the reason why slight staphylococcic infections in
man are not associated with marked haemolysis.

Streptocolysin is formed in forty -eight hours when a virulent
streptococcus is incubated in broth containing blood-serum or
ascitic fluid, and it is a remarkable fact that the nature of the
serum used modifies the lysin produced. Thus if ox serum be
employed the lysin will act on the corpuscles of the guinea-pig,
rabbit, and man, but not those of the ox or sheep ; whilst all
these will be dissolved by that grown in broth to which human
serum has been added.

Streptocolysin is less thermolabile than tetanolysin and staphy-
lolysin, requiring an exposure of ten hours to 55° C. or of two
hours to 70° C. for complete destruction.

The other bacterial haemolysins — i.e., those produced by the
B. pyocyanens, B. typhosus, and B. coli — are quite different from
the foregoing in being thermostable. Thus pyocyanolysin re-
sists a temperature of 120° C. for thirty minutes. Typholysin
appears to be less resistant, but is definitely thermostable.
Colilysin is as stable as pyocyanolysin ; it is not destroyed by
a temperature of 120° C. for half an hour, and does not undergo
spontaneous weakening for months. It is obvious that these
substances are different in character from the other haemolysins
and exotoxins, and the fact that (in the case of pyocyanolysin and
colilysin, the most heat-resistant of the group) the haemolytic
property of the culture only appears when it becomes strongly
alkaline and is roughly parallel in degree to the amount of alka-
linity, has led some to think that the substances are not the true
haemolysins at all, but merely simple alkaline chemical products
of growth ; and this is corroborated by the fact that much of the
haemolytic power is taken away on neutralization with a weak
acid. It appears that this is not the case, since in a culture of
B. coli at a temperature of 23° C., the alkalinity reaches its maxi-
mum on the fifth day, whilst the haemolytic property does not
appear until later. The subject requires further investigation, and
at present it is advisable to disregard these substances which differ
so much from their allies.

The chemical nature of the exotoxins has been the subject of
much controversy, and is still very imperfectly understood. It
will not be discussed at great length, since from the point of
view of immunity it is not of very great importance.

The close analogy between the bacterial exotoxins and certain

54 CHEMICAL NATURE OF TOXINS

vegetable toxins, such as ricin and abrin (which were thought to
be definitely proteid in nature), led, very early in the history of
the subject, to the view that these toxins are proteid in nature,
and this view was strengthened by the fact that when the diph-
theria or tetanus bacillus is grown in an albuminous fluid, proteid
substances which are toxic and give the specific reactions of the
toxins in question can be precipitated therefrom. Thus Hankin
and Sidney Martin found toxic albumoses in bacterial cultures,
and apparently succeeded in proving that abrin is an albumose.
Brieger isolated a toxalbumin from diphtheria cultures, and
Sidney Martin showed that from cultures of the same organism
in alkali albumin it is possible to prepare an albumose which
he thought to be the specific toxin. Many similar researches
were published, and the exotoxins were regarded as being albu-
minoid in nature, and the term toxoprotein was applied to them.
Several writers — Duclaux in particular — argued that this was
not the case, and thought that these proteid substances merely
carried the true toxins with them mechanically on precipitation,
just as the precipitates of inert substances such as cholesterin
will carry enzymes down with them. This theory was sup-
ported by Brieger and Cohn, who purified tetanus toxin from all
ordinary proteids, and especially by the researches of Buchner
and Uschinsky, who cultivated tetanus and diphtheria bacilli in
solutions devoid of all albuminous material, the necessary nitro-
genous nutriment being provided by asparagin. Under these
circumstances the toxic solution contains neither albumoses,
peptones, nor known proteids of any description. The toxins
thus formed are present in infinitesimally small amount, and have
never been obtained in a pure form, nor submitted to ultimate
analysis. It is known, however, that they contain nitrogen, that
they are readily destroyed by heat, and that they are dialysable.
These considerations lead us to the supposition that they are
closely allied to the proteids, and especially to the albumoses or
peptones, but form a group differing from any of them, and
approximating more closely to the enzymes. That this is the
case appears certain from the facts brought out by researches on
the antibodies ; all the substances of known chemical composition
which lead to the production of antibodies on injection into suit-
able animals are either proteids or else substances of indefinite
composition similar to the toxins, and apparently all proteids will
lead to the production of antibodies on injection into suitable

ON THE NATURE OF TOXINS 55

animals. These facts lead us to the belief that the exotoxins are,
at any rate, allied to the proteids, and form with the enzymes a
group of the substances of peculiar composition.

We have referred above to ricin as a substance once thought
to be of definite proteid nature, and a few facts may be given
concerning this substance, which is closely allied in every way
to the bacterial toxins, and which may be taken as a type of the
vegetable toxins or phytotoxins. It occurs in the seeds of various
species of Ricinus, and was formerly regarded as being a proteid,
since, like the bacterial toxins, it is carried down mechanically
with proteid precipitates. Thus Stillmarck regarded it as a globulin,
since he prepared it from the seeds by a process which was adapted
to the separation of those substances (solution in 10 per cent. NaCl,
precipitation with sodium or magnesium sulphate, and dialysis).
But Jacoby thought he had succeeded in separating it entirely
from its proteid accompaniment, making use of the fact that when
a mixture of ricin and the other substances present in the seeds
are acted on by trypsin, the active principle is acted on but slightly,
if at all ; the ricin itself, in a pure state, is readily digested by
trypsin, like the other toxins. Jacoby digested an extract of
castor-oil seeds for five weeks, and then added enough ammonium
sulphate to render the fluid 60 per cent, saturated, and ricin was
thrown down in an almost pure state ; it was purified by repreci-
pitation, and then found not to give any of the proteid reactions,
though it retained the characteristic toxic properties of the sub-
stance. Quite recently, however, Osborne, Mendel, and Harris
obtained ricin in a very pure form, and found it to be either
proteid in nature or at least inseparably associated with coagulable
albumin ; its toxicity was removed by tryptic digestion or heat
coagulation. Its great potency (TTrV^ mgr- being a lethal dose
per i kilo of rabbit) suggests that the substance which they
prepared was really pure.

Ricin resembles the bacterial toxins in the following points :
It has a period of incubation ; it gives rise to an antitoxin when
suitably administered ; it is extraordinarily potent, the lethal dose
per kilo of weight (in rabbits) being a minute fraction of a
gramme ; it is destroyed by boiling ; and it is much less potent on
ingestion than on injection. Its main toxic properties are fever,
loss of weight, albuminuria, haematuria, and haemorrhage from
the intestine ; death occurs in about twenty-four hours with acute
nervous symptoms. It has a most interesting and characteristic

56 THE ENDOTOXINS

action on the blood, clumping the corpuscles in a peculiar way,
even at a dilution of i : 600,000, and also haemolyzing them.

Further research leads us to believe that the toxin molecule
may be, and under ordinary circumstances is, actually of more
complex constitution, being combined with a molecule of true
proteid. We have already pointed out the fact that streptocolysin
differs in its reactions according to the origin of the serum on
which it is grown. The best example, however, is derived from
diphtheria toxin when grown in broth containing blood-serum or
plasma, and subsequently heated. This solution is but feebly
toxic, probably from the toxins having undergone a change into
toxoids, yet it possesses the power of immunizing an animal
against diphtheria, and of stimulating the production of anti-
toxin to an unusual degree, but only on condition that it is
injected into an animal of the same species as that from which
the serum in which the bacillus grew was obtained. Thus horse-
serum toxin will stimulate the production of antitoxin in horses,
but not in goats or rabbits, and so forth. We are justified in sup-
posing that the essential toxin molecule formed by the diphtheria
bacillus exists in this fluid in a state of combination with a specific
proteid of horse serum, and that the resulting compound molecule
differs from that form when the bacillus is grown in goat serum,
in which the essential toxic molecule is united with a different
proteid. We may suppose that this essential toxin molecule is
produced in Buchner and Uschinsky's asparagin solution, but that
it is not produced under ordinary conditions, being in a state of
combination with proteid materials of more complex structure.
These facts render further research into the chemical nature of
the exotoxins of comparatively little importance.

THE ENDOTOXINS.

In the case of diphtheria and tetanus and a few other organisms
the mode of formation of toxins is a perfectly simple one, and
one exactly analogous to the formation of soluble enzymes. In
most other cases, however, the facts are less easy to understand,
and seem to point to the formation of a toxin which remains
under normal circumstances locked up in the substance of the
bacteria, just as invertase and diastase are contained within the
yeast cell, and not excreted by it into the surrounding fluid. A
satisfactory theory as to the nature of these toxins is not forth-
coming, and the experimental results obtained by various

ON THE NATURE OF TOXINS 57

observers is very contradictory and difficult to understand, the
difficulty being increased by the fact that in the earlier researches
the distinction between antitoxic and bacterial immunity was not
understood. As a result of this we have to be careful in in-
terpreting these early results, so as to make sure that when the
author speaks of a serum as containing antitoxin he does not
really mean that it contains a protective substance which may
not be an antitoxin at all. In many cases the data are not
sufficient for us to discover its actual nature.

The organisms on which the chief amount of experimental
work has been done are those of cholera, typhoid, tubercle,
anthrax, and the pneumococcus, and it is these which we shall
discuss in chief, excluding, however, the consideration of the
toxins of the tubercle bacillus for separate consideration in a
subsequent chapter.

The general facts brought out by experiments with organisms
such as those of typhoid and cholera are these : The germ-free
filtrate of a young and actively growing culture is very slightly
toxic, if at all. The nitrate of an older culture is usually feebly
toxic, but to a degree which can hardly be compared with that of
diphtheria or tetanus ; it may take several cubic centimetres to
kill a rabbit or guinea-pig. And even this feeble toxicity is
largely discounted by the fact that the nitrate may contain acids,
nitrites, etc., which are poisonous, but in no way related to true
toxins. Yet in some cases exotoxins do exist in the filtrate,
since it is possible to obtain an antitoxin for them. The re-
actions of these antitoxins, however, are peculiar, in that the law
of multiples does not seem to apply beyond a certain figure. This
is well seen in the case of B. pyocyaneus, which forms a sort of
connecting-link between cholera and diphtheria, in that it forms
a definite though feeble exotoxin, whilst the immunity to it is
bactericidal. Wassermann showed that it is possible to produce
a true antitoxin against this toxin, and to determine the amount
which will just neutralize one lethal dose. He found, however,
that a multiple of this amount of antitoxin beyond ten would not
protect an animal against a corresponding dose of toxin, With
larger doses of toxin even a great excess of antitoxin was power-
less to prevent a lethal issue. Similar results have been obtained
in the case of cholera. These and other results have led some
authorities to consider that these exotoxins are not the specific
toxins which the pathogenic action of the bacillus defends, but

58 MACFADYEN'S RESEARCHES

secondary products of but little importance. Thus, in the case of
B. pyocyaneus it is possible to immunize an animal by cautious
injections of living organisms, yet its serum has no antitoxic
powers against the so-called toxin.

These facts have turned attention to the bodies of the bacteria
themselves, with the result that they have been found to be
definitely toxic, although in many cases the toxicity is not great.
The theory has therefore been put forward — by Pfeiffer espe-
cially— that under normal circumstances these organisms do not
secrete a soluble toxin, but that their protoplasm itself is toxic,
and that it is only set free on the death and solution of the cell,
thus accounting for the slight toxicity of old cultures, in which
such a solution of the cells must take place. The symptom of
the disease caused by these organisms is attributed to the solu-
tion of the bacteria by the fluids of the body.

The study of these endotoxins has not left the matter clear.
They are present in the bodies of the bacteria, whether the latter
have been killed by heat, by antiseptics, or by drying. They are
apparently but slightly soluble in water, but can be obtained in
solution by autolysis of the bacteria in normal saline solution in
the incubator, by grinding the dead bacilli, or by the use of very
high pressure (the method introduced by Buchner for the extrac-
tion of endo-enzymes from yeast). But it did not appear possible
to produce an antitoxin against this poisonous material ; in
addition, animals which have been immunized against the living
organism might be as susceptible as before to the dead bacteria,
or to extracts of them.

The researches of Macfadyen and Rowland have apparently
disproved this, and tend to support the opinion that the endo-
toxin is a true toxin, for which an antitoxin can be obtained.
They obtained young cultures of various organisms, froze them
at the temperature of liquid air, and then ground them (whilst
solid) into an impalpable powder. This was made into a paste
with normal saline solution and centrifugalized, to remove any
solid particles. The juice thus obtained was sterile. It was
more powerful than endotoxins prepared in other ways, and it
acted very quickly, having a very short period of incubation, if
any. Thus, in the case of the typhoid toxin i c.c. killed in three
hours and J^ c.c. in less than two days, on intraperitoneal in-
jection. It was less active on subcutaneous injection — not more
so, in fact, than other toxins of the typhoid bacillus — requiring

ON THE NATURE OF TOXINS 59

£ to y1^ c.c. to kill in seven days. Macfadyen and Rowland
found that they could immunize animals against their toxin, and
that its serum was antitoxic. These researches are difficult to
harmonize with those of other observers. We must admit, how-
ever, that it is possible to prepare an antitoxin to the endotoxins.
The failure of other observers to do so may be owing to the fact
that their toxins were not prepared in so suitable a manner for
this purpose, and may have undergone some unknown secondary
alterations.

But these researches do not clear up the whole of the mystery,
for some observations of Metchnikoff and others show that the
V. cholera: can produce a soluble toxin whilst in the animal, and
apparently without being killed in the process. These observers
prepared collodion bags, which they filled with cultures of this
organism, hermetically sealed, and inserted into the peritoneal
cavities of guinea-pigs. The animals died in a few days with
the symptoms of cholera intoxication, although no bacteria had
escaped from the sacs ; the organisms in that situation were still
alive. Control experiments with dead organisms showed that
little toxin was present ; the animals remained alive, though they
might show some symptoms of toxic action. It appears, there-
fore, that the living bacteria do elaborate an exotoxin whilst
within the animal body, and that this exotoxin has the power of
diffusing through a collodion membrane. Welch has suggested
an explanation which cannot be discussed fully here, but which
may be mentioned briefly. He points out that when bacteria
are injected in living animals the tissues of the latter react and
produce substances — bacteriolysins, etc. — which are injurious to
the bacteria, and which determine in part the resistance of the
host, and suggests that the bacteria may also react in a similar
way to the cells with which they are brought in contact. Just as
the animal host only produces its toxins — the bactericidal sub-
stances— when the bacteria are brought into contact with it, so
the bacteria may only produce their protective substances — the
unidentified true toxins — when brought into contact with aggressive
animal cells. If this is the case it is obvious that we cannot expect
to produce these toxins in vitro, except perhaps by cultivation of
the bacteria in question in fresh serum from an immunized animal.

CHAPTER III

THE PHENOMENA OF ANTITOXIN
FORMATION

As a general rule, to which there are important exceptions, it is
necessary to make use of susceptible animals for the production
of antitoxin. When toxin is injected into animals in which it pro-
duces no injurious effects, it either disappears rapidly from the
blood or remains for a long time in that fluid or in the tissues
without leading to the formation of antitoxin. The most remark-
able exception to this rule is the way the cayman reacts to tetanus
toxin. The animal is immune, and if kept in the cold (20° C.) the
toxin soon disappears from the blood, no antitoxin being formed. If,
however, it is kept at an elevated temperature (32° to 37° C.), the
toxin disappears as before, but now antitoxin makes its appearance
(Metchnikoff). Such cases are exceptional, and when we wish
to procure antitoxin, we make use of an animal in which the
toxin in question produces symptoms of intoxication. The pro-
cess is usually much easier in large animals, such as horses or
goats, than in small ones, such as rabbits or guinea-pigs, the
immunization of which presents considerable difficulties. We
shall take as illustrations of the general phenomena of the pro-
cess the methods adopted for procuring diphtheria antitoxin and
tetanus antitoxin from horses, since these have become so familiar
from their extensive application.

On injecting a small dose of a potent diphtheria toxin sub-
cutaneously into a horse — say, \ c.c. — under the skin of the
neck we find there is a latent interval of a few hours or a
day before the development of symptoms ; then there is a l»cel
reaction, consisting in the formation of a hard brawny mass ©f
inflammatory oedema round the site of the inoculation, and a
general reaction, consisting in fever, anorexia, and symptoms of
general malaise. These symptoms last a day or two, according
to the dose of toxin injected, its potency, and the degree of sus-

60

THE PHENOMENA OF ANTITOXIN FORMATION 6l

ceptibility of the animal ; and when they have passed off a small
amount of antitoxin will be found in the blood, and the animal
will, as a rule, be found to be less susceptible to the action of the
toxin than before, so that the injection of the same dose will
produce less reaction, both local and general.

This, however, is not always the case, and careful research
leads us to the belief that the appearance of immunity is preceded
by a period of hypey sensitiveness, in which the animal betrays a
greatly increased susceptibility to the action of the toxin, and this
in spite of the fact that it may contain quite large quantities of
antitoxin in its blood. Thus it happens not infrequently that
after a horse has passed successfully through the early stages
of immunization to diphtheria toxin, and has developed far more
antitoxin than is necessary to neutralize the doses of toxin with
which it is being treated, it yet will die after the injection of an
amount which it would appear must be immediately rendered inert
as soon as it came into contact with the plasma. Such cases have
been reported from the Pasteur Institute, Behring and Kitashima,
and others, and by Brieger for tetanus. In the latter an immunized
horse died after an injection of tetanus toxin with the typical
symptoms of tetanus intoxication, and after death its blood con-
tained much free antitoxin. The phenomenon has probably been
witnessed by most observers who have been engaged in the manu-
facture of antitoxin, though it has become much less frequent since
the introduction of modern methods for the early treatment of
animals. It is an exceedingly puzzling one, and we shall leave
its further interpretation until later ; here it is sufficient to say
that Behring's theory of the occurrence of a stage in which the
tissues are hypersensitive to the toxin is well established.

The difficulty of immunizing the small animals of the laboratory
to these toxins appears to depend in large measure on the marked
development of hypersensitiveness. Thus Behring and Kitashima
found that they could kill a guinea-pig with ^J^ of the " minimal
lethal dose" of tetanus toxin, if this amount were divided into
several doses and given at suitable intervals, and similar facts
have been recorded by others.

The most striking proof of the occurrence of hypersensitiveness
in the process of immunization has been investigated by Behring,
who pointed out that normal horses show no local effects from the
injection of small quantities of tetanus toxin ; their connective
tissues are insusceptible to its action. As the animal becomes

62 CHOICE OF TOXINS

immunized to the action of the toxin this is not the case; the
tissues at the site of inoculation react to the poison with the pro-
duction of a mass of inflammatory redema similar to that seen in
a horse injected with diphtheria toxin. It is obvious that these
connective tissues have become more susceptible to the action of
the tetanus toxin, and this in spite of the antitoxin with which
they are bathed.

In order to avoid the difficulties arising from the occurrence of
hypersensitiveness in the early stages of immunization, the use
of unaltered toxin has now been practically abandoned, the follow-
ing methods, either alone or in combination, being employed
instead :

1. The use of mixtures of toxin and antitoxin, the latter being
present in amount sufficient to neutralize all the toxin, or in
excess. This is repeated several times, the amount of antitoxin
given being gradually reduced, until at last a small amount of
unaltered toxin is given.

It must not be thought that the immunity which is acquired in
this case is simply passive, and due to the free antitoxin which is
injected. The process is probably fundamentally different. We
shall revert to it subsequently.

2. The injection of toxoids. This method is of especial advan-
tage in the case of tetanus, to which toxin animals are extremely
sensitive, and the dangers of the early stages of the process of
immunization are very great. The toxin formed in the cultures
may be transformed into toxoids by the action of trichloride of
iodine, a solution of iodine in iodide of potassium, or by heat,
the filtered cultures being exposed to a temperature of 60° C. for
a time sufficient to destroy their toxicity. Toxin that has been
heated to a temperature much higher than this is completely
destroyed, and is useless for the process.

3. The use of serum toxin, which probably contains toxoids in
an unusual condition of activity. This method was introduced
by Cartwright Wood, and is now in general use in this country
for a part at least of the process of immunization, since it leads to
a more rapid production of antitoxin of high potency than can be
obtained by other methods in the same time. Ordinary alkaline
broth is inoculated with diphtheria bacilli, and incubated for a week
at 37° C. Then 15 to 30 per cent, of its volume of serum from an
animal of the same species as is to be used in the process of
immunization is added, and the incubation continued for a month

THE PHENOMENA OF ANTITOXIN FORMATION 63

or six weeks. It is then heated to 65° C. for half an hour and
filtered. It gives rise to marked febrile reaction and but little
local reaction. The initial dose is 200 to 300 c.c.

In giving these large doses the most convenient method is to
use a large wash-bottle, the side of which is graduated in
cubic centimetres. To the outflow arm there is attached 2 or
3 yards of pressure tubing, in the farther end of which a strong
exploring needle is inserted, and firmly wired in place. The
pressure is obtained by means of a bicycle pump attached to the
inflow tube of the wash-bottle by means of pressure tubing.
There should be a lateral branch communicating with a mano-
meter, by which the pressure can be regulated. Very high
pressure is sometimes necessary, especially in the later stages of
the process, when the subcutaneous tissues of the horse's neck
become sclerosed and dense from the repeated injections. The
apparatus is most easily sterilized by passing strong carbolic
lotion through it.

On testing the blood-serum from time to time, it is found that
the amount of antitoxin gradually rises, each injection being
followed by an increase in the antitoxic value of the serum. Thus
the process is a cumulative one, the antitoxic level being raised
step by step until a certain height is reached. This height differs in
different animals. Thus Atkinson, in summarizing his experience
of 100 horses, found that half of this number gave less than
300 units of antitoxin per cubic centimetre, a quarter between
300 and 500, whilst three gave more than 800. There appears
to be no method of investigation by which the value of a horse
as a source of antitoxin can be predicted early in the course
of treatment, and the great variability amongst different animals
is probably the reason that different observers have come to
such divergent opinions as to the best doses to give and the
most suitable intervals between each. Here are three chief
methods :

(a) By the use of large doses of toxin, 250 to 500 c.c. every day,
or almost every day, leaving an interval of a week or ten days
before the bleeding, so as to allow the last injection to produce its
maximum effect.

(b) The use of large injections (similar to the former) at longer
intervals — five to ten days.

(c) The use of relatively small doses of weak toxins repeated
every day.

64 THE NEGATIVE PHASE

All these methods have their advocates, and good results can
apparently be obtained by all.

On ceasing to inject toxin, it usually happens that the antitoxic
value of the serum commences to decline, and, in the absence of
further injections, would probably continue to do so until it had
entirely disappeared from the blood. In a few cases, however,
a period of antitoxic equilibrium is maintained for some time, the
amount of antitoxin lost by the excretions or destroyed in the
system being compensated for by a fresh production of the same
amount. When this is the case the phenomena resulting from
the injection of a single dose of toxin can be traced with ease, and
is of great importance, as will appear subsequently. The first
effect of the injection is the production of a negative phase, in which
the amount of antitoxin in the blood is suddenly and greatly
diminished. This production of a negative phase is apparently
a general phenomenon, and is found to occur in the development
of nearly all antibodies in which it has been investigated. If the
dose of the primary substance (toxin, etc.) is very small, the
negative phase may be short in duration and very slight in extent,
and may be overlooked, or may possibly be omitted altogether.
Its explanation is very uncertain and cannot be discussed here,
but it must be pointed out that it is not due to the neutralization
of the antitoxin in the blood by the toxin injected ; the proof of
this is that it is large out of all proportion to the latter. Thus in
one reported case the fall in antitoxic value of the serum which
occurred in the negative phase would have required an injection
of toxin 12,000 times as large as was actually given if it were due
to simple neutralization. The length of the negative phase varies
in different animals, and can only be learnt by experiment. It
appears to be roughly proportional (in the same animal) to the
amount of primary substance injected : the larger the doses of
toxin, the greater the fall and the longer its duration. It is,
of course, synchronous with the toxic symptoms, if any, of the
substance injected, since both are due to the action of this sub-
stance on the blood and tissues ; but the two do not appear to be
mutually dependent : a well-marked negative phase may appear
without any other symptoms of disease.

The negative phase is succeeded by a rise, the positive phase,
in which the antitoxic value of the blood reaches and usually
surpasses its previous level. It commonly reaches its maximum
in about a week, and then commences to decline ; hence it is

THE PHENOMENA OF ANTITOXIN FORMATION 65

advisable that the animal should be bled for antitoxin after a rest
of about a week from its last injection.

The bleedings are carried out at the laboratories of the Royal
Colleges of Physicians and Surgeons in the following manner :
The receptacles for the blood are 2-pound glass jam-jars, which
are sterilized by heat and covered with parchment paper which
has been soaked for some hours in i : 20 carbolic. Two layers
of this are used, and the lower one has two radial slits cut in it,
leaving a triangular wedge, which can be raised and access to the
bottle thus obtained. Twelve or fourteen of these are required
for each horse, and each is filled about two-thirds full.

The side of the horse's neck is shaved and washed with a
solution of lysol, or a lysol dressing is put on an hour or two
before the operation. The horse is placed in the stocks, and if
violent the head is restrained by a twitch. It is then necessary
to apply pressure at the lower part of the neck, in order to distend
the jugular vein ; this may be done by the thumb of an assistant,
or, better, by means of a firm leather plug, which is pressed into
the groove in front of the sterno-mastoid muscle by means of an
arrangement of straps devised by Dr. Cartwright Wood. In this
way the vein is temporarily occluded, and stands out clearly
above the region where the pressure is applied. The operator
(having sterilized his hands as for a surgical operation) then
makes an incision about 2 inches long and above or just in-
ternal to the vessel ; this should open the deep fascia, but
need not actually expose the vein. He then takes a trocar and
cannula having a diameter of about T\ inch, and pushes it firmly
downwards into the vein ; success in this is shown by the blood
oozing up by the side of the trocar. An assistant now stands
ready with a short metal tube which fits inside the cannula
and communicates with 2 or 3 yards of indiarubber tubing, with
a foot or so of glass tubing at its farther end. The whole has
been sterilized by being soaked in lysol or carbolic lotion. A
second assistant now reflects half of the outer parchment covers
of one of the jam-pots, reflects the triangular strip which has'been
already cut in the inner cover, and inserts the glass tube in the
opening. The operator then removes the trocar, and the first
assistant rapidly fits the metal tube attached to the rubber tubing
into the cannula; when this is done quickly hardly any blood
escapes. The blood now passes through the rubber tubing into
the jam-pot, which rapidly fills. When about two-thirds full the

5

66 PREPARATION OF ANTITOXIN ON A LARGE SCALE

assistant pinches the indiarubber tube and places the outflow tube
in a second pot. The outer cover is replaced on the first pot,
which is removed to a warm place to clot. The process is repeated
until twelve or fourteen pots have been filled.

Horse's blood coagulates slowly, and a well-marked buffy coat is
formed. In twenty-four hours this will have contracted, and
much of the serum will be squeezed out. In order to draw this
off use is made of a wash -bottle, the short tube of which is con-
nected with a water-pump, such as is used for filters, by which
a partial vacuum can be maintained. The long tube is connected
to a piece of indiarubber tubing terminating in a length of glass
tubing. A jam-pot is opened by half reflecting the outer cover
and lifting the triangular strip cut in the inner one, and the glass
tube is inserted. Air is now sucked out of the wash-bottle by
turning the tap which puts it into communication with the suction-
pump, and the serum siphons over. When all the serum has been
abstracted a second jar is treated in the same way, the parchment
cover of the first being replaced, and the process is continued with
all the jars. In twenty-four hours more serum will have appeared,
and the process is repeated, and a small amount may often be
obtained on the third day. In this way the total yield of anti-
toxin is usually nearly 50 per cent, of the total volume of blood
(4i to 5 litres).

The antitoxin thus obtained is usually sterile, the most careful
precautions being taken to prevent contamination. Carbolic acid
(0-3 per cent.), or trikresol (0-3 per cent.), or a mixture of the two,
must now be added to preserve it. It is then filtered through
a Berkefeld filter (not a Chamberland filter, through which it
passes with great difficulty, if at all), a low pressure only being
used, and finally tested for sterility by means of cultures, and for
the presence of toxins by the injection of large (10 c.c. or more)
amounts into normal guinea-pigs.

A specimen is taken at the time of the bleeding, and this is
tested for antitoxic value in the manner to be described subse-
quently. The results of this testing will give the amount necessary
to obtain the required dose, and this amount is placed in sterile
tubes or bottles ready for use. In most cases mixtures are made,
antitoxin of low potency being mixed with more powerful sera in
order to obtain the requisite dose in a given volume of serum.
An ingenious machine is used by which the tubes or bottles are
filled automatically with the antitoxin in the required amounts.

THE PHENOMENA OF ANTITOXIN FORMATION 67

In the earlier stages of immunization, as we have seen, each
injection is followed (after a negative phase) by a rise in the anti-
toxic value of the serum above its previous level. In the stage
which now follows this does not occur, or not definitely ; there is
a negative phase, but it is found impossible to force the antitoxic
value above a certain level, which varies in different horses.
This second stage, or period of maintained maximum, varies in
different horses, and may last a few months or a year. While it
lasts there are, of course, oscillations ; it falls, for instance, if the
animal contracts any disease or suffers in general health, but its
general average is about the same.

Sooner or later this state of affairs changes, and the antitoxic
value of the serum begins to fall, and cannot be raised or even

FIG. 7.

a, b, Normal resisting power ; b, c, period of hypersensitiveness ; c, d, period
of rise in immunity ; and d, e, maintained high level thereof. /, g, Normal
amount of antitoxin ; g, h, period in which it increases ; and h, i, gradual
fall and ultimate (theoretical) disappearance.

maintained at its former level in spite of very large doses of toxin.
It trends steadily downward, although the animal may continue
to give useful serum for a long time. Thus in one of Atkinson's
best horses (out of a series of 100) the serum contained 1,000 to
1,100 antitoxic units per cubic centimetre for ten months, and
then gradually sank, but remained above 300 units per cubic
centimetre for two years.

After a prolonged rest the power of manufacturing antitoxin
may return, but then only lasts a short time, and cannot be re-
newed again.

The third stage, therefore, consists in a gradual disappearance
of antitoxin from the blood, without any loss of the immunity to the
toxin. It would seem, indeed, as if the immunity reaches its
highest level at this point, in spite of the almost complete absence

5—2

68 IMMUNITY NOT DEPENDENT ON ANTITOXIN

of antitoxin. Thus the two phenomena do not run paripassu with
one another. This, is illustrated in the foregoing diagram, which
represents in a purely schematic way the period of immunization
and utility of an antitoxin horse, the height of the continuous line
from the base representing the degree of immunity, that of the
dotted line the amount of antitoxin in the blood.

CHAPTER IV

INTERREACTIONS OF TOXIN AND ANTITOXIN

STARTING from the facts that a suitable dose of antitoxin will
prevent the development of symptoms if toxin is injected shortly
before, at the same time, or shortly after, or that if antitoxin and
toxin be mixed in vitro and injected subsequently, no symptoms
develop, we have to inquire the mechanism by which this is brought
about. Two theories suggest themselves at once. The antitoxin
might act on the cells of the living body in such a way as to render
them insusceptible to the action of the poison, or, in other words,
render them immune, or the toxin and antitoxin might unite
chemically to form an inert and harmless compound. When the
fundamental facts of antitoxic action were first discovered, the
majority of pathologists probably inclined to the former alternative,
the latter seeming too simple and teleological. A certain amount
of experimental evidence was also forthcoming in favour of this
view, but as this has a merely historical value it will not be
considered. It is now fully proved that toxin and antitoxin form
chemical compounds, and that the prophylactic and curative value
of the latter is to be explained simply on the grounds that this
compound is inert, or devoid of toxic action on the animal cells.
The evidence in favour of the occurrence of this chemical com-
bination requires brief discussion.

The first group of experiments pointing in this direction are
those in which the toxin and antitoxin are mixed in vitro, and
the result tested by means of red blood-corpuscles as indicators,
the intervention of the cells of the living body being thus excluded.
(Many of these experiments can be repeated on corpuscles which
have been heated to a temperature sufficient to destroy the life of
isolated body cells, and the possible objection that the corpuscles
are " surviving " thus removed.)

The first of these researches was that of Ehrlich, who showed

69

70 FILTRATION EXPERIMENTS

that the agglutinative action of ricin on red blood-corpuscles
could be inhibited in vitro by means of the serum of an immunized
animal. Kanthack showed that the action of snake -venom in
inhibiting the coagulation of blood was similarly prevented in
vitro by its appropriate serum, whilst Kossel and others did the
same for the haemoglobin of eel's blood, and Ehrlich for
tetanolysin. The previous cases were not of true bacterial toxins,
and might possibly be open to objection on that account. The
experiment of Neisser and Wechsberg on the effect of leucocidin
on leucocytes in vitro, and its inhibition by means of an antiserum,
is another case in point. It is true that in this case the leucocytes
are living, but we can hardly imagine that they have become
immunized by the action of the serum, or that the phenomenon
can be explained on any hypothesis other than that the toxin and
its antiserum have combined.

The second and most important series of researches are those
of Martin and Cherry, who show that several toxins (e.g., that
of diphtheria and snake-venom) pass through a porcelain filter
which is impregnated with gelatin, whereas their appropriate
antitoxins, being composed of larger molecules, do not. (This had
previously been proved by Brodie.) They found, further, that
when a mixture of toxin and antitoxin was placed on such a filter
the first portion of the filtrate was toxic, but that the amount
diminished, and all toxicity disappeared a few minutes after the
mixture had been made. The inference is clear : the toxin had
united with the antitoxin to form a molecule as large as, or even
larger than, that of the latter, and therefore, like it, unable to pass
through the pores of the filter. These researches have been
confirmed by Brodie, and form, on the whole, the most striking
direct proof of the union of the two substances yet brought
forward.

Calmette found that snake-venom is more heat-resistant than
its antitoxin, withstanding a temperature of 80° or 90° C., whereas
the latter is rendered inert at 68° C. He was then able to show
that a neutral mixture of the two could be rendered toxic again
by exposure to a temperature of 70° C. ; and this fact was used first
as an argument against the chemical theory of combination, and
secondly as a proof that the toxin is not destroyed when it unites
with antitoxin. As a matter of fact, neither inference is
necessarily correct, and the experiment was shown by the further
researches of Martin and Cherry to constitute a proof of the

INTERREACTIONS OF TOXIN AND ANTITOXIN 71

chemical theory : for they found that if the mixture were allowed
to stand for some time at the temperature of the body before being
heated, its toxicity was not restored by a temperature of 70° C.
This seems to show that the toxin did not exist as such in the
mixture, otherwise it would not have been destroyed by the heat ;
it must, therefore, have become combined with the antitoxin, or
at any rate modified by it in some way. On the other hand, the
experiment does not prove that the toxin is completely destroyed
beyond all power of further activity ; it simply shows that, when
in a condition of combination with its antitoxin, it is less thermo-
stable than when free. Similar facts were adduced by Wasser-
mann with regard to the combination between pyocyaneus toxin
and its antitoxin, and are capable of a similar explanation.
Marenghi has also brought forward somewhat similar results
with diphtheria toxin.

Lastly, Ehrlich has shown that the conditions which favour the
occurrence of chemical combinations favour the union of toxin
and antitoxin — e.g., it is accelerated by heat, and takes place more
quickly in concentrated than in dilute solutions.

This brings us to the question as to whether the combination
takes place in accordance with the law of multiple proportions —
a question of great difficulty, but one which has lead in its
elucidation to the discovery of facts of much interest. As far as
concerns the action of the haemolysins and other toxins that can
be readily tested in vitro, there is no doubt that this question, in
its simplest form, must be answered in the affirmative. If it
requires x c.c. of a given solution of toxin to dissolve exactly
i c.c. of a 5 per cent, emulsion of red blood-corpuscles, then it
will require 2#, 3*, 4*, etc., c.c. to haemolyze 2, 3, 4, etc., c.c. of
the same emulsion. We assume in each case that the haemolysin
is added at once, and not in small consecutive amounts. To
study the effect of the partial neutralization of toxin by antitoxin
we will briefly outline Ehrlich's famous work on the standardiza-
tion of diphtheria toxin, and the conclusions he arrived at in
consequence of the results thus obtained.

We have seen that it is possible to determine with a close
approach to accuracy the minimal lethal dose of diphtheria toxin
for standard guinea-pigs— 4.e., those weighing about 250 grammes.
This amount is called the toxic unit (TU), and a toxin of which
T^ c.c. is just sufficient to kill a test guinea-pig in three or four
days is considered to be normal toxin of unit strength, and is

72 STANDARDIZATION OF DIPHTHERIA TOXIN

written DTN.1 A toxin of half this strength, of which -^ c.c.
is the lethal dose, is written DTN0.5. Toxins of other potencies
are numbered accordingly.

Ehrlich now proceeded to define a unit of antitoxin as the
amount that would just neutralize 100 lethal doses of toxin : this
is called IU ( = immunizing unit). This amount may be contained
in any quantity of the serum ; thus, in that used for clinical work
i c.c. contains anything between 300 and 1,000 units, or even
more. For the purpose of testing toxins it is convenient to use
an antitoxic serum which is much more dilute than this, and an
antitoxin of unit strength is defined as one which contains i unit
of antitoxin in i c.c. — i.e., i c.c. of the antitoxin will just neutra-
lize i c.c. (100 lethal doses) of standard toxin. The reaction
between these amounts is written thus :

i c.c. toxin (=100 lethal doses) + i c.c. antitoxin = L0,

where L0 (L = limes) indicates that the mixture is a truly neutral
one, and that it does not kill a susceptible animal within the time-
limit, or produce any pathogenic action whatever.

Now, if, as Ehrlich believes, the affinity of toxin for antitoxin
is a powerful one, similar to that of a strong acid for a strong
base, it should follow that if to the 100 lethal doses of toxin we
add only T9^ of i c.c. of standard antitoxin, then TJy of the original
amount — i.e., i lethal dose — should remain unneutralized, and the
animal should die in the same time as a similar animal which had
received i lethal dose and no antitoxin.

As a matter of fact, this is not what occurs. We find that when
we inject the mixture the animal does not die in a short time
with the ordinary symptoms of diphtheritic intoxication, but
develops local oedema, and possibly paralysis, which may bring
about death at a remote period. The same thing happens if we
add still less antitoxin to the 100 lethal doses of toxin. To take
a particular case, it is not until the mixture contains less than
T7^y of i c.c. of antitoxin that the animal dies acutely in the way
it does after an injection of i lethal dose of toxin. It seems,
therefore, that the whole of the toxicity of the toxin is removed
when only three-fourths of the amount of antitoxin necessary to
neutralize it has been added, or that a given amount of toxin can

1 DTN = diphtheria toxin normal. It is also written DTN1M250=DTN one
unit for a guinea-pig (Meerschweinchen) weighing 250 grammes.

INTERREACTIONS OF TOXIN AND ANTITOXIN

73

combine with one -fourth more antitoxin than is necessary to
neutralize it.

To account for this Ehrlich supposed that there are really two
substances present in the broth in which diphtheria bacilli have
been grown. There is the true toxin, which brings about local
inflammatory oedema, often going on to necrosis and causing local
alopecia, and causing acute death, and toxon, which produces only
soft and transient cedema locally and subsequent paralysis. Both
these substances combine with antitoxin, but the toxin has the
greater affinity for that substance, and when the total neutralizing
dose of antitoxin is added in successive small amounts, the whole
of the toxin is neutralized first, leaving the toxon free, and this
takes place when three-fourths of the whole amount of antitoxin
has been added. Ehrlich represents this result in the form of a
spectrum, thus :

0 25 50 75 100

FIG. 8.— SIMPLE SPECTRUM OF TOXIN.

The rectangle represents the L0 dose of toxin — i.e., in this simple
case i c.c. of the solution. The portion with the greatest affinity
for antitoxin is placed at the left hand of the "spectrum"; in
this case it is represented by the toxin. On the right are the
substances with the least affinity for antitoxin — in this case the
toxon.

Further investigation shows that the process is not usually so
simple as this. In certain samples of toxin we find that the
addition of small quantities of antitoxin causes no alteration in the
toxicity of the L0 dose. Thus, in a case of frequent occurrence
it happens that we may add J c.c. of normal antitoxin before
any loss of toxicity occurs ; i c.c. of the normal toxin will kill
100 guinea-pigs, and i c.c. of the same toxin + J- c.c. of normal
antitoxin will still kill 100 guinea-pigs. To explain this, Ehrlich
supposed that the solution contains a third substance, prototoxoid,
which is entirely devoid of lethal activity, but which has a power
of combining with antitoxin even greater than that which toxin
possesses. Thus, on the addition of small amounts (up to J c.c.) of
the antitoxin, this inert substance will seize on the antibody, unite

74

STANDARDIZATION OF DIPHTHERIA TOXIN

with it, and so render it incapable of neutralizing the true toxin.
The spectrum of this solution will be represented thus :

Pntto-
toxoid

Toxone

FIG. 9. — SPECTRUM OF TOXIN.

In this, as in the other diagrams, the lethal portion of the mixture
is shaded, the non-lethal portion left blank.

Ehrlich found on actual experiment that the constitution of the
solution was even more complex than this, and had to assume the
existence of yet other bodies. Thus, if the spectrum above were
a true representation of the constitution of i c.c. of the solution, it
follows that the first quarter and the last quarter of the antitoxin
added were without effect, so that the middle \ c.c. completely
neutralized the whole of the TOO lethal doses. Now let us
imagine this i c.c. of standard antitoxin divided into 200 equal
parts, and added part by part to the i c.c. of standard toxin, or
100 lethal doses. Then —

The first 50 parts added will combine with prototoxoid, and will

not affect the toxicity of the mixture ;
The next 100 parts added will neutralize 100 lethal amounts

of true toxin ; and
The last 50 parts will combine with toxon.

Now if the spectrum were as simple as is shown above, and if
the toxin were quite uniform in its combining capacity and its
toxicity, it would follow that the first -^^ part added after the
addition of ^^ part would just neutralize one lethal dose and leave
99 lethal doses over. Again, the addition of the amount necessary
to neutralize all the prototoxoid ( = -f^ c.c.) + -££$ c.c., which would
neutralize all the prototoxoid and all the toxin except yj^ part
(=i lethal dose), and all the toxon, should leave i lethal dose
of toxin free, and the animal should die in the limit of time for
i lethal dose. We might represent this as follows :

149 unnc'jfralised
'Toxin (

INTERREACTIONS OF TOXIN AND ANTITOXIN 75

in which the oblique shading represents the toxic portions, as
before, and the horizontal shading represents the amount neutral-
ized by the addition of ^-^ c.c. of antitoxin ; the portion with
oblique but no horizontal shading represents the toxic portion
which remains unneutralized : it constitutes TJ^ of the total
shaded portion, and is therefore i lethal dose.

Such a finding may occur, but is unusual. In most cases we
find that the amount of toxin left free on partial neutralization is
subject to laws which are far more complex. In a case given by
Madsen and described in the same way we find :

The addition of |££ parts of antitoxin left free no lethal substance
— a term which we shall use for the present, instead of " toxin,"
to denote the portion of the spectrum with the oblique shading.
In Ehrlich's language all had been neutralized except the toxon.

The addition of ££$ left 5 units of lethal substance free ; it
follows that i§S~ 29^j-=2fio°i7 had been necessary to neutralize
these 5 units.

The addition of -f^ left 55 lethal units free ; hence, if after
the addition of -f^ (as above, leaving 5 lethal doses free) we add
an additional -/^, the difference (•/<&) will neutralize 50 lethal
doses (55-5).

Hence the additon of •££§ will just neutralize the remaining
lethal doses — i.e., 45.

To account for facts like these, Ehrlich suggests that the
solution contains four or five substances. The first — i.e., that
which has the greatest power of combining with antitoxin, is called
prototoxin ; it is lethal, and it consists of two parts — an a part,
which is readily changed into inert prototoxoid, and a ft part,
which is more stable, but which may, after a time, change into
prototoxoid also. These two modifications have exactly the same
affinity for antitoxin, so that if they were present in equal
amounts, and if all the a modification were changed into proto-
toxoid, each addition of antitoxin would go to neutralize active
prototoxin and inert prototoxoid in equal amount ; hence half of
it would apparently be wasted.

Secondly, there is deuterotoxin, which also exists in an a and a /?
modification, of which the a part is readily transformed into
deuterotoxoid, whilst the (3 modification is very stable and is the
last lethal substance to disappear. The a and /3 modifications
have equal affinity for antitoxin, but this is less than that of the
prototoxin.

76

SPECTRA OF TOXINS

Thirdly, there is tvitotoxin, again in an a and a /3 modification,
with less affinity for antitoxin than deuterotoxin, and so are placed
on its right in the spectrum.

It is found, further, that the proportion of a modification to
P modification in the above forms of toxin is a simple one,
so that the ratio of toxoid to toxin present in any one part of the
spectrum is always simple (-J, ^, ^, etc.).

Fourthly, there is toxon (toxone) or epitoxoid, the characters of
which we have seen.

Lastly, some researches seem to prove that there is yet
another body, epitoxonoid, which has still less affinity for antitoxin
than has toxon, and which is entirely devoid of lethal or toxic
power. It will be left out of the further consideration of these
bodies.

The spectrum of a toxin on this theory is recorded thus :

Profofaxo/clA

Trifofoxoicf A

Toxon

Profotoxin B Deuterotoxin B Tritotoxin B

FIG. ii.

The spectrum of the example given by Madsen and quoted
above would be :

FIG. 12.
Another spectrum, given by Ehrlich, is appended :

DeuferotoxinB Tritoxoxm B

FIG. 13.

We must now turn to the experimental results which have led
to this idea of the change of the toxin into toxoid ; it has been

INTERREACTIONS OF TOXIN AND ANTITOXIN 77

referred to several times already, but not fully discussed in order
not to interrupt the main line of the argument.

L0 has been denned as the amount of toxic solution which is
exactly neutralized by i IU of antitoxin, and L+ is the amount
which, when added to i IU of antitoxin has i lethal dose left un-
neutralized. Now if the toxic solution contained a simple sub-
stance, we should expect the two quantities to have the following
relation in the simple standard toxin of which i c.c. contains 100
lethal doses.

i c.c. toxin(= 100 lethal doses) + i c.c. antitoxin (= i IU) = L0.
roi c.c. toxin(= 101 lethal doses) + i c.c. antitoxin ( = i IU) = L+.
.-. L+-L0 = o-oi c.c. = i lethal dose.

This, however, is not the case. If we take a neutral mixture of
toxin and antitoxin — e.g., of 100 units of the former and i of the
latter — add to it i lethal dose of toxin, and inject it into an animal,
it will not cause death ; there may be transient local cedema and
late paralysis, symptoms which are indicative of the presence of
free toxon. We must in general add very much more than
i lethal dose to the neutral mixture in order to bring about a fatal
result. For example, in our standard toxin it might happen that
the L+ dose was about 1-35 c.c. In other words—

i -oo c.c. toxin solution + i unit of antitoxin = L0.
I*35 c.c. toxin solution + i unit of antitoxin = L+.
L+-L0 = o-35 c.c.

This result can readily be explained on Ehrlich's assumption
of the existence of substances of differing combining powers for
antitoxin. For the sake of simplicity, we will take his earlier
nomenclature, and consider the substance as made up of proto-
toxoid (with a greater affinity for antitoxin than true toxin has),
toxin, and epitoxoid, with little affinity, and corresponding to
toxon. The spectrum of the toxin under discussion is :

150 200

FIG. 14.

In this diagram we represent the L0 dose — i.e., i c.c. divided
into its component parts. The oblique shading represents, as

78

DIFFERENCE BETWEEN L.

before,.Jthe acutely lethal portion, and the whole is shaded hori-
zontally to show that it is completely neutralized by the i unit of
antitoxin.

Now let us take 1*25 of the same solution and add to it i unit
of antitoxin. In this extra 0-25 c.c. of toxin (a quarter of
the original amount) there are 12-5 parts of prototoxoid, 25 of
toxin, and 12-5 of epitoxoid. There will now be 62-5 parts of
prototoxoid, 125 of toxin, and 62-5 parts of epitoxoid. The 200 parts
into which we imagine the unit of antitoxin is divided will now
neutralize the whole of the prototoxoid (62-5 parts), the whole of
the toxin (125 parts), and 12-5 parts of toxon. There will be
50 parts of epitoxoid left free, but no toxin. Hence, 1*25 c.c. of
the toxic solution is less than the L+ dose. The result may
be represented thus :

n-6 parts of epifoxotd

62-5 parrs

125 parrs.
FIG. 15.

200 62-5 parrs

Let us now imagine a third mixture of 1-33 c.c. of the toxic
solution and i unit of antitoxin. The 0-33 c.c. of toxin will
contain 16-6 c.c. of prototoxoid, 33-3 c.c. of toxin, and 16*6 c.c. of
toxon, and the total 1*33 c.c. will thus contain 66-6 c.c. of proto-
toxoid, i33'3 c.c. of toxin, and 66*6 c.c. of epitoxoid. The proto-
toxoid + toxin ( = 200 parts) will just absorb the whole of the unit
of antitoxin, leaving nothing but toxon free. Thus :

66 6 parrs

133-3 parrs

FIG. 1 6.

ZOO 66- 6 parrs

Then, if i extra lethal dose of toxin be added to the above
mixture, it will find all the antitoxin utilized by substances with
a combining affinity as great as, or greater than, its own, and
will be left free. Hence, the L+ dose is just greater than
1 33 c.c.

All this follows from what has previously been said concerning

INTERREACTIONS OF TOXIN AND ANTITOXIN

79

partial neutralization. If, however, we now keep this atrtitoxin
for some time, especially if it is exposed to warmth, light, air,
or certain chemical substances, we find a great change. The
L0 dose is unaltered : i c.c. is still exactly neutralized by i unit
of antitoxin, but we find that this amount is now much less lethal,

Prototoxoid

Epitoxoid

50 parrs

loo parrs
FIG. 17.

SO parrs

and the minimal lethal dose may have risen from o'oi c.c. to
O'O2 c.c., or higher.

If the cause for this increase in the lethal doses is investigated
by the partial neutralization method described above, it will be
found that the results obtained are such as will be readily ex-
plicable on the assumption that some of the molecules of toxin

Profotoxoid.

Toxoid (SO parts)

Epiroxoid

50 parts

100 parts
FIG. 18.

50 parrs

have ceased to be poisonous, but have retained their combining
power unaltered; whilst the non - poisonous portions of the
spectrum are unaltered. Thus, to take the simple case described
above, and shown in Fig. 17, in which protoxoid, toxin, and
epitoxoid are present in the proportion of 50, 100, and 50. If we
keep this, we may find the lethal dose doubled — i.e., ^ c.c. instead
of T<JTT c-c- But on working out the action of antitoxin, we may
find that the first -25 c.c. added removes none of the toxicity, and
the last -25 is equally without apparent effect. Thus the middle
•5 c.c. is used up in neutralizing 50 lethal doses, and of this
fraction further investigation shows that each one-fiftieth part
neutralizes one lethal dose; the antitoxin, therefore, appears to
have fallen off in potency ; but we know that this is not the case.
The explanation is that half the molecules have been changed into
the non-toxic form described above. The spectrum of this altered
form is shown in Fig. 18.

Thus, in a particular case Ehrlich found the minimal lethal

80 TOXIN AND TOXOID

dose of a toxin to be 0*003, and the L0 dose 0-305 (= 100 lethal
doses). Nine months later the L0 dose was still 0305, whilst
the minimal lethal dose was 0-009 c-c- (=33*3 lethal doses only).
Thus, the toxin had fallen to one-third of its former toxicity, but
had retained its power of combining with antitoxin unaltered ; and
it is found that the two numbers usually bear some simple ratio
to one another — e.g., i - ^ or i — J, as in the previous examples.

We are now in a position to understand all Ehrlich's results,
and especially the proof that a molecule of toxin consists of two
parts — a haptophore group, which has the power of combining
with antitoxin or with the protoplasm of the body cells, but which
has in itself no lethal action ; and a toxophore group, which has
the power, when linked to a body cell by means of the haptophore
group, of causing toxic symptoms, probably by a process akin to
enzyme action. This toxophore group is unstable, and when it is
decomposed the toxin molecule is converted into toxoid ; it then
retains its power of uniting with antitoxin and with the tissue
cells, but is devoid of toxicity. Further, there are many varieties
of toxin, which differ from one another — (i) in their avidity for
antitoxin, and presumably for the tissue cells ; (2) in the readiness
with which they are decomposed into inert toxoid; and (3) in
their toxic action, in that true toxins cause rapid death, local
inflammation, necrosis, etc., but no paralysis, whilst toxons
produce only transient soft oedema and subsequent paralysis.
Lastly, Ehrlich supposes that toxin and antitoxin have a great
affinity for one another, so that their combination resembles that
of a strong acid with a strong base, and the compound toxin-
antitoxin molecule when once formed does not dissociate, but
remains as a stable substance. It is on this point that his views
differ from the more recent ones of Arrhenius and Madsen, who
regard the union as being akin to that of a weak acid and a weak
base. We hope to be pardoned if, before describing their experi-
ments and conclusions, we give a brief outline of the difference
between the two reactions.

According to modern chemical theories, most substances in
watery solution decompose into ions, atoms or groups of atoms
carrying an electrical charge. Thus, a solution of hydrochloric
acid contains ions of H carrying a positive charge and of Cl
carrying a negative one. Different substances undergo ionization
to a different degree. Thus, HC1 and NaOH are ionized to a
great extent, boracic acid and ammonia hydroxide but slightly.

INTERREACTIONS OF TOXIN AND ANTITOXIN 8l

Now in general it is only free ions which enter into chemical
combination. For example, on adding HC1 to NaOH the posi-
tively charged H ion combines with the negatively charged OH
ion to form water, and the positively charged Na ion combines
with the negatively charged Cl ion to form NaCl. The two
substances HC1 and NaOH are strongly dissociated, and hence
the combination between the two is almost complete. This is an
example of the combination of a strong acid and a strong base ;
where it was expressed in older phraseology, the substances have
a strong affinity for one another. It is to this type that Ehrlich
conceives the union of toxin and antitoxin belongs.

When two substances dissociate but slightly — e.g., acetic acid
and alcohol or boracic acid and ammonia — the reaction takes
place in obedience to different laws (the law of "mass action"
of Guldberg and Waage). When we ^dd alcohol and acetic
acid we get ethyl acetate (ester) and water ; but if we take
equivalent combining quantities of the two primary substances,
the reaction is not complete, as it is in the case of HC1 and
NaOH. On the contrary, the solution will still contain free
alcohol, free acetic acid, ester, and water. This is due to the
fact that the process is reversible. Acetic acid and alcohol
combine on the one hand to form ester and water, and ester
and water combine on the other, and dissociate into free acid and
free alcohol.

In the first case, in the combination of a strong acid and a
strong base, the reaction is a simple one. If we take 100 com-
bining equivalents of NaOH, and add to it 10 combining equiva-
lents of NaOH, 10 parts of the alkali will be neutralized, and so
on, the whole of the alkali being neutralized when 100 combining
equivalents have been added. In the second case the reaction is
more complicated, and is expressed by the law that the products
of the concentrations of the substances on one side of the equa-
tion, divided by the product of those on the other, is a constant
(which varies in different reactions, and can be determined by
experiment). Thus in the case given :

Concentration of acid x concentration of alcohol

= constant.
Concentration of ester x concentration 01 water

It follows that when we add a relatively small amount of acid
to a given volume of alcohol, it is practically all used up to form
ester, much free alcohol remaining ; but each succeeding addition

6

82 REVERSIBLE REACTIONS

of acid is divided into two parts, of which one combines with
the alcohol and the other remains free. As we continue to add
successive small amounts, this latter uncombined portion gets
greater and greater, so that the addition of successive small
volumes use up less and less of the alcohol ; and it is only when
there is a great excess of acid that all the alcohol is used up.
Theoretically, some always remains.

The difference may be represented graphically thus :

FIG. 19.

The line a b represents the neutralization of a given amount of
NaOH by its equivalent of acid. It is a straight line, since each
successive addition of equal amounts of acid neutralizes the same
amount of alkali.

The line c d represents the reaction of alcohol and acetic acid,
or, to take the example used by Madsen, the neutralization of
ammonia by boracic acid. It is a hyperbolic curve. It is almost
a straight line to begin with, since each small addition of boracic
acid is almost all used up by the ammonia, there being enough
NH4 ions to seize on all the boracic ions which are added.
Farther along, however, it changes according to the rules already
described, and ultimately approaches infinitely near to the base
line, but never reaches it. There are always free boracic acid
and free ammonia in the solution, though the amount of the latter
is infinitely small when the former is in great excess.

The first suggestion that the reaction between toxin and anti-
toxin might be a reversible one was due to Myers in an investi-
gation on the haemolytic action of snake-venom, a substance

INTERREACTIONS OF TOXIN AND ANTITOXIN »3

which gives partial neutralization phenomena quite similar to
those we have described as occurring with diphtheria toxin, and
also declines in toxic strength, but not in combining capacity,
forming toxoids. It was, however, the researches of Arrhenius
and Madsen which led to the establishment of the theory on a
sound basis. Madsen had previously investigated the constitution
of the haemolysin of tetanus (tetanolysin), working on Ehrlich's
lines, and had found it to consist of proto-, deutero-, and trito-
toxin, and a large amount of toxone. The substance was much
more convenient to work with than diphtheria toxin, since a constant
emulsion of red blood-corpuscles could be used as test objects
instead of guinea-pigs, and thus the experiment could be multi-
plied at will ; and when Madsen added the antiserum in small
amounts he found that the apparent irregularities due to the
presence of these various forms of toxin disappeared, and the
curve of neutralization is represented by a curve similar to that
given above, as illustrating the reactions between boracic acid
and ammonia. Thus :

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

FIG. 20. CURVE OF NEUTRALIZATION OF TETANOLYSIN (MADSEN).

6—2

84

ARRHENIUS' AND MADSEN's VIEWS

The heavy line represents the curve of neutralization on
Ehrlich's principles, and apparently shows the presence of a
large number of varieties of toxin with different combining
capacities. When, however, the neutralization was carried out by
the addition of very small quantities of the antitoxin at a time,
the curve became a hyperbola of the character seen in the dia-
gram. Further, having found the value of k by determining by
experiment the amount of toxin left free after the addition of
a certain amount of antitoxin to a certain amount of the toxic
solution in two different instances, and hence, by use of the
formula :

Free toxin free antitoxin _ (toxin x antitoxin)
vol. vol. vol.

the amount of free toxin after any addition of antitoxin could be
calculated theoretically. It could then be determined by experi-
ment, and the results compared. In one case given by Arrhenius
and Madsen this was done, with the following result :

Amount of
Antitoxin added.

Amount of Toxin
(observed).

Amount of Toxin
(calculated).

0*05

3-67

3-67

o-i

3-13

2'95

0*15
0-2

2-32
1-62

2-29
172

0-3
0-4

07
I'D

0-97
0-63

o*45
0-27
0-18

1-03
0-62
0-46
0-28
0-18

1*3

1-6

0*12

0-13

O'll

2-0

0'08

0-09

The correspondence is certainly very close.

The case of diphtheria toxin was next investigated, and in the
figure which follows the " stair-step " curve, showing Ehrlich's
conception of its constitution, shows the presence of proto-, deutro-,
and trito-toxins, and of toxon. The curved line is that calcu-
lated after the constant of dissociation had been determined by
experiment.

INTERREACTIONS OF TOXIN AND ANTITOXIN

32
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2

\

\

v

\

\

(\

\

"^

^

^

)

\

\

\

\

\

f

\

^

v

2 4 6 8 10 12 14 16 -18'

FIG. 21. — CURVE OF NEUTRALIZATION OF DIPHTHERIA TOXIN (MADSEN).

This table shows the difference between the observed and calcu-
lated results :

Antitoxin.

Free Toxin
(observed).

Free Toxin
(calculated).

O'l

75'I

75*1

0-I5

62'6

627

0-2

47-6

50*6

0-25

45-8

38-6

0-3

25*9

27-3

o'35

i7'3

17*5

0-4

9-6

9'6

o'45

5'3

6-0

o'5

3'1

4'i

0-6

1-6

2-6

The conclusion to which Arrhenius and Madsen came was that
toxin and antitoxin react like a weak acid and a weak base ; that
the reaction is a reversible one ; and hence that the combination
of toxin and antitoxin is an unstable one, dissociating into free
toxin and free antitoxin.

86 ARRHENIUS* AND MADSEN'S VIEWS

On this theory many facts which were formerly very difficult
to explain become quite simple. Take, for instance, the exact point
of neutralization of toxin by antitoxin as seen in the determina-
tion of the L0 dose : this has always been a matter of great
difficulty — a difficulty formerly explained by assuming the last
substances to be neutralized are the toxons, which have a very
feeble and indefinite pathogenic action. On the physical chemistry
theory the difficulty disappears, because there is no point of exact
neutralization. In spite of the presence of an excess of antitoxin,
some dissociation of the toxin-antitoxin molecule will always
occur, and the mixture will always contain free toxin, though in
very small amount. Further, an exactly neutralized mixture con-
taining a few lethal doses of toxin is, of course, without action,
whereas a large bulk of the same mixture may be toxic ; this
also is readily explicable. Then there are some old experiments
of Buchner's, which showed that a mixture of tetanus toxin and
antitoxin which was neutral to mice would produce tetanus in
guinea-pigs, and several others (Roux's and Roux and Vaillard's)
of similar nature. The explanation of these is also easy.

The most interesting explanation of a previously known phe-
nomenon which Madsen offers in the light of his new theory is
that of the immunization of animals by means of a neutral mixture
of toxin and antitoxin. If we follow Ehrlich, and believe that the
compound is a stable one, this is very difficult to explain. Madsen's
solution is that the mixture contains free toxin, to which the
immunizing property is due.

Again, we can also explain the death of animals from specific
intoxication when it occurs in spite of the presence of free anti-
toxin in the blood in exactly the same way. It is true that the
free antitoxin would tend to inhibit dissociation ; but, on the other
hand, the hypersensitiveness of the tissues which occurs at the
early stages of the process of immunization — and it must be
remembered that it is only in these stages that death from in-
toxication takes place — would render the cells more susceptible
to minute amounts of toxin. The explanation may not be a
perfect one, but it appears to be the best forthcoming.

The importance of the whole question from the point of view
of immunity rests on this question of the dissociability of the
combination, for it is obvious that if this is the case, our views of
the action of antitoxin in the animal body will be very different
from those we shall hold if we regard the toxin-antitoxin molecule

INTERREACTIONS OF TOXIN AND ANTITOXIN 87

as an inert one of no further interest ; and this theory of dissocia-
tion is open to the objection urged by Nernst, and probably felt
by most bacteriologists on the first enunciation of the views of
Arrhenius and Madsen. Thus, if the combination of toxin and
antitoxin undergoes dissociation in the living animal and the toxin
is set free, it will immediately combine with the susceptible cells.
The equilibrium will now be disturbed and more toxin-antitoxin
molecules will be dissociated, more toxin set free, and more cells
poisoned ; and this process will go on until all the toxin has been
passed on to the cells and the antitoxin left free. Now this
dissociation takes place quickly, so that on this theory it would
seem that the antitoxin would only interpose a very temporary
barrier between the cells and the toxin, the lethal action of which
would be delayed, but in no way inhibited. We will return to
this question of dissociation shortly, and meanwhile state briefly
Ehrlich's objections to the physical-chemical theory. In the first
place, he points out that if we make a mixture of two alkaloids,
of which one is haemolytic and the other not, and neutralize them
by the addition of a strong acid, the result may be represented by
a hyperbolic curve ; this is put forward as a parallel experiment
to the neutralization by antitoxin of a substance containing active
toxin and inert toxoid. Secondly, some of the curves given by
Madsen and Arrhenius do not correspond very closely with the
observed results, and this is especially the case at their commence-
ment and termination. At the commencement of the curve it is
found in many cases that the addition of small amounts of anti-
toxin does not influence the toxicity ; this is readily explained on
the supposition of the existence of prototoxoid, but hardly on any
other hypothesis. The most interesting point, however, is the
behaviour of the curves at the termination — i.e., in what Erhlich
would call the region of the toxons ; here Arrhenius and Madsen
usually found figures which were lower than the calculated results.
Now Ehrlich holds that traces of toxin do not lead to paralysis,
and if the effect of a nearly neutral mixture of toxin and antitoxin
were due to dissociation we should expect no paralysis to occur,
the toxic action being due to toxin, and not, as Ehrlich thinks, to
toxon. Madsen and Arrhenius suggest that the action of this
trace of antitoxin may be modified by the presence of antitoxin in
excess. Further, Madsen and Dreyer claim to have found a
diphtheria poison, of which small quantities would cause paralysis
without the addition of antitoxin, so that the question of toxon

88 EVIDENCE IN FAVOUR OF DISSOCIATION

would not come in. And they also showed that certain mixtures
of toxin and antitoxin might act fatally on rabbits in a few days
and cause only paralysis in guinea-pigs, but that if a little more
antitoxin were added it would act as a toxon on rabbits and be
inert to guinea-pigs. To explain this, Ehrlich had to add yet
another body to his list of components of the diphtheria poison,
and to the proto-, deutero-, trito-toxin, and toxon he added
toxonoid, which is inert for guinea-pigs and produces paralysis
in the rabbit. The subject will not be followed farther, but
enough has been said to show its extreme difficulty.

To revert to the question of dissociation, Madsen and Walbum
have brought forward some definite evidence of its existence, of
which the more important are the following. They neutralized
ricin with antiricin, and to the neutral mixture added some red
blood-corpuscles. After allowing the mixture to stand for some
time, the latter were centrifugalized off, when it was found that
they had become charged with ricin and the fluid containe(i/free
antiricin. /

Secondly, they made use of the fact that diphtheria toxin is a
substance of comparatively small molecule, and can diffuse
through gelatin, whereas its antitoxin does so very slowly. They
prepared a neutral mixture of the two, and placed it on the
surface of a column of gelatin, and after some forty days' contact
examined slices at different depths, and found free toxin at a
certain distance from the surface. This they explained by assum-
ing that it had dissociated from the antitoxin, and diffused down-
wards into the gelatin. It was pointed out, however, that if the
mixture had been allowed to stand for some time the phenomenon
did not occur ; it would seem, therefore, that the combination
takes place slowly, but, once formed, does not dissociate.

There is, however, other and independent evidence in favour of
the theory that dissociation of a primary substance and its anti-
body does occur. Thus Muir and Morgenroth showed inde-
pendently that if red corpuscles are treated with as much
amboceptor as they will take up, and then mixed with normal
corpuscles, some of the amboceptor will pass to the latter, and
on the addition of a suitable complement-containing serum the
whole may be dissolved. (In this experiment the red corpuscles
correspond to the toxin, whilst the amboceptor is the antibody
and corresponds to the antitoxin.)

Bordet's explanation of the phenomena is quite different. Both

INTERREACTIONS OF TOXIN AND ANTITOXIN OQ

Ehrlich on the one hand and Arrhenius and Madsen on the other
agree that the combination of toxin and antitoxin is a chemical
union, and takes place in obedience to the laws of multiple
proportions : a single molecule of the one substance always com-
bines with the same number of the other on complete neutraliza-
tion. Bordet denies this, and compares the phenomenon with the
absorption of a stain by a colourable substance. His theory is
that a molecule of toxin requires many molecules of antitoxin for
its complete neutralization, and that it can be partially neutralized
or attenuated by a smaller number. Thus he holds that when a
small amount of antitoxin is added to a large amount of toxin,
there is not a mixture of free toxin and of molecules of toxin-
antitoxin (as would occur if either of the theories already discussed
was true), but a uniform solution of toxin of diminished potency.
Bordet shows that many of the experimental results of other
observers are explicable on his theory, and gives some new facts
in support thereof. For example, the amount of an emulsion of
red blood-corpuscles which can be haemolyzed by the addition of
a given quantity of haemolytic serum can be readily determined.
We will suppose that i c.c. of the serum just dissolves all the
corpuscles in 5 c.c. of the emulsion, and no more. We might
suppose that all the red corpuscles have all their combining
valencies exactly neutralized, so that we may regard them as an
exactly neutralized toxin-antitoxin mixture. This, however, is
not the case, for if we add the haemolytic serum in small amounts,
say 0-05 c.c. at a time, we shall find that only a small proportion,
perhaps 2 c.c., of the same emulsion can be completely haemolyzed.
Thus each corpuscle takes up more haemolysin in the second case
than in the first.

The objection may be raised that a red corpuscle is very
different from a molecule of toxin. It can undoubtedly combine
with a vast number of molecules of its antibody (haemolysin), but
it is quite easy to understand how complete haemolysis may take
place if only a certain number of its combining valencies are
occupied by antibody. But the molecule of toxin is, as we have
already seen reason to believe, much smaller than the molecule
of antitoxin ; and although this fact does not in itself render it
impossible that the toxin is of higher valency than the latter,
it certainly renders it unlikely that it is so much higher in this
respect that there can be an indefinite number of stages between
free toxin and a fully neutralized one. Bordet has, however,

90 ADSORPTION

brought forward evidence in favour of his view to which this
objection can hardly apply. It will not be discussed here, as it
involves certain questions concerning serum haemolysins which
have not yet been discussed.

Bordet's theory supplies an explanation, though hardly an
adequate one, of the negative phase. We have already seen that
when an animal is producing antitoxin an injection of toxin will
cause a fall in the antitoxic value of the serum far greater than
can be accounted for by the neutralization effected by the toxin
injected : it may be 2,000 times as great, or more. If we assume
that the combination does not take place in obedience to the laws
of multiple proportions, we can imagine that it may go on very
differently in the body, and that in that situation a small amount
of toxin may neutralize a large amount of antitoxin. But it is
very difficult to believe that this is the true explanation, for it
would involve an increase of the neutralizing power of the toxin
to 2,000 times that which it has outside the body — an increase
which certainly seems improbable. Further, we have been asked
to imagine a molecule of antitoxin spreading itself out and
weakening many molecules of toxin, and we are now asked to
imagine a molecule of toxin dividing itself amongst 2,000 molecules
of toxin, and completely neutralizing them.

The last conception which we have to consider is that of Biltz,
which approaches very closely to Bordet's. Starting with the
idea that toxins and antitoxins are both colloids, he attempts to
show that their union is analogous to adsorption rather than to
an ordinary chemical union.

Adsorption is a process akin to solution, and does not necessarily
involve a chemical union between the two substances which take
part in it. It may take place between colloids and colloids, or
between colloids and crystalloids, and in other ways. The two
substances taking part in the process are found to be electrically
positive and negative (as shown by their moving to the cathode
or anode in an electric field) ; thus a colloidal solution of silicic
acid (which is negative) is precipitated slowly by K2SO4, which
has a feeble positive charge ; more rapidly by CuSO4, which has a
stronger one ; and immediately by A12(SO4)3, which has a stronger
one still.

Biltz and his collaborators started with the idea that toxins and
antitoxins are both colloids. They attempted to find the electrical
nature of tetanus toxin by electrolysis, but found that it was

INTERREACTIONS OF TOXIN AND ANTITOXIN QI

destroyed ; the destruction occurred, however, sooner at the
cathode than at the anode, suggesting that it was a negative
colloid. They found it to be precipitated by positive colloids,
such as colloidal hydrated oxide of iron or chromium, though this
action occurred only in vitro, not in vivo. Tetanus antitoxin, how-
ever, did not lose its activity when exposed to mineral colloids.

Further, there is an approach to the phenomenon of specificity
in the adsorption of one colloid by another, since certain colloids
are only precipitated by certain other colloids ; the specificity,
however, is by no means so exact as in the case of the toxins, etc.,
and their antibodies.

A further analogy arises in the explanation of the difference
between the L0 and L+ dose, which is paralleled by the relation-
ship between colloidal ferric hydroxide and arsenious acid. When
these two substances are mixed adsorption and precipitation take
place ; hence the use of the former substance as an antidote for
the latter. Now it is found that if a solution of arsenious acid is
just rendered tox*e by the addition of ferric hydroxide, very much
more than one lethal dose of arsenic must be added to make the
mixture toxic again.

The further investigation into the evidence which they have
adduced would lead us too far into the study of the agglutinins to
be undertaken here. It will be dealt with subsequently (see
Chapter XII.).

It seems, on the whole, that no theory is absolutely sufficient to
explain all the phenomena, and that as soon after each new one is
adduced the supporters of the older ones bring forward evidence
which renders it untenable. The probability is, at the time of
writing, that Ehrlich's views are generally held, and are open to
the fewest objections. They are complicated, it is true, and have
had to undergo constant modifications as new facts have arisen ;
but the facts themselves are complicated. Yet it must be con-
fessed that there are some grave objections to its acceptance in its
present form, and it may become yet more involved before it can
be fully accepted as a complete explanation. Thus the analogy
with other bodies, and the phenomena of the death from intoxica-
tion of animals with antitoxin in their blood, seem to point strongly
to the theory that the toxin-antitoxin molecule dissociates strongly
both in vivo and in vitro. Yet this is not compatible with Ehrlich's
views.

CHAPTER V

THE ORIGIN OF ANTITOXIN— THE SIDE-CHAIN

THEORY

THE theory that antitoxins are derived from their appropriate
toxins deserves a short consideration, since it is so inherently
probable. When we consider the large numbers of toxins which
give rise to their antitoxins when injected into animals, and see
that each antitoxin has a direct action on its own toxin, but none
or practically none on others, the most likely explanation of the
phenomenon is that the animal tissues have the power of splitting
the toxin into two parts, or of otherwise modifying it, and that
this modified toxin can combine with unaltered toxin and form
antitoxin. Some experimental evidence is forthcoming to support
this view. Thus it is found that diphtheria toxin which has been
submitted to the action of an electric current loses its toxicity, but
retains that of producing immunity on injection. Some thought
that this was due to the direct transformation of the toxin into
" artificial antitoxin," and it was hoped that a similar change
might be brought about in the human body in disease by the use
of electricity. The explanation of .the phenomenon is very
simple. The electrolysis of the water in which the toxin is
dissolved gives rise to oxidizing substances which transform the
toxins into toxoids, the immunizing virtues of which we have
already seen. There is no method by which antitoxin can be
prepared other than by the injection of toxins into suitable
animals.

That antitoxin is not a direct transformation product of toxin
appears probable from the following considerations, none of which
perhaps is conclusive, but which together make up a body 'of
evidence of some importance :

i. The injection of a certain amount of toxin will give rise,
under suitable circumstances, to the formation of much more

92

THE ORIGIN OF ANTITOXIN — THE SIDE-CHAIN THEORY 93

antitoxin than it can neutralize. Thus Knorr showed that a
single unit of toxin might call forth 100,000 units of antitoxin.
This can scarcely be accounted for by supposing the transforma-
tion of the one substance into the other.

2. It frequently happens that antitoxin occurs, sometimes in
considerable amounts, in the blood of animals that have not
received injections of toxin, and have not suffered, as far as
known, from the disease in question — in other words, of " normal "
animals. Thus horses frequently have traces of diphtheria anti-
toxin in the blood, as much as 4 units per c.c. having been
observed. Metchnikoff suggests that this may be due to the
action of " pseudo-diphtheria " bacilli, which are so common.
But no one has been able as yet to produce diphtheria toxin or
symptoms of diphtheria by means of these bacilli, and it is not
usual to regard them as having any connection, other than a
superficial morphological resemblance, with the true diphtheria
bacillus.

If we may bring other antibodies into the argument the case
against Metchnikoff's supposition is enormously strengthened.
The blood of many animals contains agglutinins and hsemolysins —
often in large amounts, often under conditions in which we may
almost exclude the possibility of a preceding infection. Thus the
blood of a normal horse will almost invariably clump typhoid
bacilli, B. pyocyaneus, the cholera vibrio, and many other
organisms.

3. An animal which has received an injection of antitoxin
rapidly eliminates it from the blood, whereas the antitoxin which
is formed in the body remains a much longer time. This argu-
ment does not carry very much weight, since, as Metchnikoff
points out, the toxin might remain for long periods occluded
in the cells, and only undergo a gradual transformation into
antitoxin.

4. Roux and Vaillard showed that the whole of an animal's
blood might be removed by repeated venesections, and that the
newly regenerated blood might contain almost as much antitoxin
as was present before the haemorrhage, thus suggesting a new
formation. But the objection urged under the last heading is of
weight here also, and if we consider it probable that toxin may
remain latent in the tissues for long periods without causing
symptoms, we shall exclude this piece of evidence from the
argument, as well as the observations of Salomonsen and Madsen,

94 CONSTITUTION OF PROTOPLASM

that the production of antitoxin may be increased by an injection
of pilocarpin.

If we reject this explanation of the phenomenon, and regard
the production of antitoxin as being analogous to the process of
internal secretion, we must regard it as being due to the action of
the living cells when they are stimulated by the toxin. There is
at the present time but one theory which suggests how this may
be brought about : this is Ehrlich's celebrated side-chain theory,
which has played so important a part in the modern study of
problems concerning infection and immunity, and which, though
highly hypothetical, deserves a close and careful study.

Ehrlich points out that a cell has two functions. The first
is the physiological function, which it enacts in the body — secretory
in the case of a gland cell, conductive in the case of a nerve fibre,
etc. ; and the second is the function of nutrition, necessary to
supply the waste which is constantly taking place. We must
suppose a similar constitution for each of the complex molecules
of living protoplasm which build up the cell ; in each molecule
there is a portion which discharges the specific function of the
cell, and this portion has to be nourished. The functional portion
is the more important, and is called by Ehrlich the active centre
(Leistungskern). In our diagrams we shall represent it as forming
the central portion of the giant molecule, but it is hardly necessary
to caution the reader that this diagrammatic representation has
no necessary morphological basis. It is just as permissible to
regard this portion as diffused through the nutritive part of the
molecule as long as the difference in constitution and action of the
two are borne in mind.

The second portion, that concerned in nutrition, is more
important in relation to immunity. We may regard it as having
two functions — it " seizes " suitable molecules of food substances
from the blood or lymph in which it is bathed, and it alters them
in such a way as to build them up into living material which can
take its place in the molecule of protoplasm, so as to replace that
which has been used in the life of the cell.

The function of " seizing " molecules of food from the sur-
rounding tissues implies a selective faculty, for we cannot imagine
that all the molecules which circulate in the blood and lymph are
suitable for all cells at all times. This apparently intelligent
selection of suitable material is, of course, at bottom a chemical
process : the food molecule becomes attached to some portion of

THE ORIGIN OF ANTITOXIN— THE SIDE-CHAIN THEORY 95

the cell for which it has a chemical affinity. Now Ehrlich
supposes that this affinity for food molecules is situated in certain
portions of the molecule of protoplasm — in certain groups of atoms
which he calls side-chains, receptors, or haptines. [The name " side-
chain " was, perhaps, badly chosen. It denotes a possible analogy
with complex organic bodies — e.g., of the aromatic group — which
are composed of a central portion (such as the benzene ring) and
side-chains, on which many of their reactions depend. The com-
parison is not a very close one, and all that is necessary is to
regard the molecule of protoplasm as possessing numerous groups
of atoms, each of which has an affinity for one of the bodies
circulating in the body fluids, and necessary for the life of the
molecule in question.]

On this theory the nutrition of the molecule takes place as
follows : A molecule of suitable food substance in the fluid sur-
rounding the cell is brought into contact with one of the receptors
for which it has a chemical affinity, and the two unite. This
is the first step in the process. The food is "anchored " to the
cell by means of a receptor, for which it has a specific combining
affinity, or which, to use Ehrlich's analogy, fits it like a key fits a
lock. The second stage involves a process which we may compare
to digestion, by which the food molecule is altered in some pro-
found manner, and absorbed, in whole or in part, into the molecule
of protoplasm.

Let us apply this theory to the process by which a cell is
poisoned. We will imagine for a moment that the molecule of
toxin contains a group of atoms which will unite specifically with
the side-chain of one of the body cells. We have already shown
that this molecule of toxin contains a haptophore group of mole-
cules, in virtue of which it can combine with antitoxin, and a
toxophore group, on which its toxic action depends. We can now
go farther, and say that the first stage of the intoxication of a cell
by means of a true toxin consists in the union of the haptophore
group of atoms in the toxin to a receptor of a molecule of proto-
plasm, this receptor being one which " fits it like a key fits a
lock." Each molecule of protoplasm has innumerable receptors,
of which only a certain number are suitable for this toxin. This
is the first step. The toxin molecule is now " anchored " to the
living cell, and in the second stage the toxophore radicle of the
toxin comes into play. We may regard this toxophore group as
exerting an enzyme-like action on the protoplasm through the

96 INTOXICATION AND NUTRITION

haptophore group and the receptor, by which the two are united.
Thus:

CELL

FIG. 22. — CELL MOLECULE WITH RECEPTORS (E, E).

A, A, Molecules of toxin (B = toxophore, C = haptophore), D = a molecule

of toxoid.

The result of this is that the protoplasm is poisoned. If only a
few of its receptors are united to toxin molecules the result may
be but slight, whereas if more are occupied the functioning
centre of the molecule will be affected, and marked symptoms will
arise, and if still more molecules of toxin are linked up the mole-
cule of protoplasm may be killed.

The essential point in this process is that it is exactly analogous
to natural nutrition, except that in the latter case the receptor
unites with a molecule which is of use to the cell, whereas in the
former it unites with one which simply resembles the food molecule
in having a haptophore group with similar chemical affinities.
Thus, intoxication with bacterial toxins is essentially a process of
nutrition, but is perverted in its later stages by the nature of the
toxophore group of the toxin.

Now consider the case in which a molecule of protoplasm is
attacked by a number of molecules of toxin which is not large
enough to kill. A considerable number of receptors will be taken
up by toxin, and we must consider these receptors as being thereby
rendered useless. The protoplasm, however, has need of these
receptors, and the need may probably be more pressing since it is
poisoned, and metabolic changes may take place more rapidly, and
there is thus greater need for renewed nutriment. Fresh receptors
must therefore be formed, and Ehrlich compares their regeneration
to the budding forth of fresh tentacles in the hydra, to replace

THE ORIGIN OF ANTITOXIN — THE SIDE-CHAIN THEORY 97

those which have been lost. Now suppose a fresh dose of toxin
reaches the cell, and that these new receptors are in their turn
taken up by toxin, but yet not in numbers sufficient to kill the
cell. The same processes will occur : the receptors will be rendered
useless, and a fresh crop of the same nature will be produced.

Now in accordance with the well-known laws of habit, in virtue
of which a part of the body can gradually be trained to perform a
function which is difficult at first, but which becomes more easy
by usage, the molecule of protoplasm gradually acquires the
faculty of producing these receptors more and more easily, in
obedience to nicely graduated doses of toxin. If these are
repeated in proper amount and at suitable intervals the cell may
be trained, so to speak, to produce these receptors, and it may

FIG. 23.

ultimately come to do so in excess and far beyond its physiologi-
cal requirements. This also is analogous with known biological
facts, such as the production of callus in a fractured bone in
excess of the amount necessary for repair. We shall revert to
this point later. Now we have to imagine that these receptors
may be formed in excess so great that they cannot all remain
attached to the cell ; some of them are pushed off, so to speak, by
the younger ones, which are growing up to take their place. If
this occurs, these receptors will pass off freely into the blood. As
we have seen, they are so constituted that they will unite speci-
fically with the particular toxin that was injected. These cast-off
receptors constitute antitoxin, and the formation of that substance is
simply due to the pathological separation of the groups of atoms,
in virtue of which the toxin molecule is linked up to the living

7

98

WEIGERT'S HYPOTHESIS

protoplasm. The whole process is explained as one of perverted
nutrition, and (once admitting Ehrlich's theory of the constitution
of the molecule of protoplasm and of the way it is nourished)
without introducing any but well-recognized biological pheno-
mena. The theory is fascinating in its simplicity, and although
there are a few difficulties in the way of its full acceptance as a
complete explanation of the facts, it certainly accounts for them
far better, on the whole, than any other theory will do — if, indeed,
an alternative one has been put forth. And it has the great merit
in a theory that its application has led to the discovery of new
facts. *

A few remarks are necessary in connection with the question

FIG. 24.

of the mechanism in which the over-production of side-chains is
produced. Weigert's hypothesis with regard to the nature of
hyperplasia as the result of injury or irritation is now well known,
and a brief outline will suffice. He imagines that the maintenance
of the normal structure and physiological function of a tissue
depends upon a condition of equilibrium brought about by a
series of mutual restraints exerted by each cell on its neighbours.
In the same way, the structure of a single cell, or perhaps of a
single compound molecule of protoplasm, depends on a similar
equilibrium existing between minute living units. If one of these
components is killed or injured, the restraint which it exerts on
the surrounding units is removed, and unrestrained growth can

THE ORIGIN OF ANTITOXIN — THE SIDE-CHAIN THEORY QQ

take place ; and Weigert points out that this always goes on to
excess, more new material being formed than is necessary to
replace the amount lost. The application of this theory to the
mode of new formation of receptors is sufficiently obvious.

We will now consider certain phenomena in the light of the
side-chain theory.

The occurrence of antibodies in the blood of normal animals is
susceptible of a ready explanation. The receptors are, in general,
united to the cell, but it is easily conceivable that a few may be
desquamated accidentally and escape into the blood. It must be
remembered that the actual amount (weight or bulk) of these
antibodies in the normal blood is probably always infinitesimal,
although the effects of this small amount may be striking. Thus
the agglutination of typhoid bacilli at a dilution of i to 160
(which often occurs with the blood of normal horses) indicates
the presence of a very small amount of specific antibody, as is
apparent from the fact that the blood may be made to clump at
i to 1,000,000 without being altered chemically in any way that
can be detected. If an exceedingly minute proportion of all the
receptors are shed into the blood, it will probably account for all
the antibodies found in that situation under normal conditions.

The fact that toxoids produce antitoxin1 is also explicable. We
must regard the receptor which is occupied with toxoid as being
thereby rendered useless for the protoplasm, although the latter is
not poisoned as a result of the union.

Next arises a most important question, and one which is ap-
parently satisfactorily answered on the side-chain theory — the
fact that certain substances (bacterial toxins in particular) give
rise to the production of antibodies, whilst others, such as the
alkaloids, glucosides, mineral poisons, poisons of simple con-
stitution, such as alcohol, etc., do not. (This may be taken as
fairly proved, certain researches which go to prove the existence
of antibodies to morphine, alcohol, etc., being inconclusive.)

The explanation is founded on differences in the way the
various substances form combinations with the protoplasm. This
latter has, as an integral part of its constitution, certain receptors
which have a specific combined affinity for proteids, and we must
imagine the union of proteids, whatever be their nature, as taking
place by a direct union with these receptors. Other substances,

1 According to Ehrlich. The fact is not universally admitted, though all
agree that they will produce immunity.

7—2

IOO NATURE OF UNION OF TOXINS AND TISSUES

however, need not, and in all probability do not, unite with the
protoplasm in this way. Ehrlich, basing his theories on the fact that
certain alkaloids and other substances can be extracted from their
combinations with organic matter by simple means, has suggested
that the union of non-proteid poisons with protoplasm is less firm
than in the other case. This seems doubtful, for nothing could
be much firmer than the combination of nitrate of silver or similar
bodies with the tissues. And we know as a matter of fact that
the toxin-protoplasm compound is not necessarily a stable one.
Emulsions of tissues will abstract tetanus toxin, and retain it on
washing, but give it up again in a free state when allowed to stand
for some time in contact with normal saline solution. However
this may be — and the point is not of importance — we may readily
admit with Ehrlich that it is highly probable that proteids unite
with protoplasm in a manner fundamentally different from alka-
loids, etc. The former process follows on physiological lines,
whilst the latter is a pathological process entirely, and one which
has no counterpart in normal nutrition. The former may be
compared to the insertion of the key in the lock, the latter to the
violent smashing of the lock. If this is the case, it is easy to
see that "chemical" poisons cannot be expected to give rise to
the production of antibodies. It is true that if a cell is already
secreting antibodies, we might expect that a substance which
stimulates the general metabolism of the cell and increases the
rapidity of the vital processes might temporarily increase this
production, and we have seen that this is the case with pilo-
carpin, which stimulates the production of diphtheria antitoxin in
an immunized horse. This is an entirely different process, and
one that is easily explicable on the side-chain theory.

It is necessary, therefore, to determine whether all the sub-
stances which give rise to the production of antibodies on injec-
tion are proteids. We must admit at the outset that in many cases
this cannot be proved ; the chemical constitution of toxins, enzymes,
and many other antigens is as yet quite unknown. But in the
cases in which the chemical composition of the primary substance
is ascertained we find it to be invariably of proteid nature ; thus,
solutions of any of the coagulable proteids will give rise to the
production of precipitins; the proteid substances present in the
body of bacteria give rise to the production of agglutinins, which
are themselves of a proteid nature, and will, on injection into
suitable animals, give rise to anti-agglutinins ; the proteid con-

THE ORIGIN OF ANTITOXIN — THE SIDE-CHAIN THEORY IOI

stituents of the red blood-corpuscles, the substances forming their
stroma, call forth the production of haemolysins ; and a great many
other cases might be quoted. It may be taken as definitely
established that wherever the nature of the antigen is known and
falls into one of the known groups, that antigen is a proteid.

We may almost go a step farther, and say that all proteids, of
whatever nature, when injected into suitable animals, will give
rise to antibodies ; in fact, some writers actually enunciate this
as a law. There are, however, a few exceptions, which are
gradually diminishing in face of further researches, and it is
highly probable that time will show that the law is valid to the
fullest extent.

Reverting now to the question of the nature of the bodies of
unknown constitution which form antibodies — i.e., the toxins,
enzymes, etc. — we may ask whether these are not in reality
proteids, in spite of their failure to give the proteid reactions as
usually accepted by chemists. The question is largely one of
nomenclature. Accepting the definitions usually given in physio-
logical textbooks, and based on a study of ordinary proteids, such
as egg-albumin, etc., we may admit that they are not; but are
these definitions valid ? Might we not define them with greater
accuracy as substances from which animal organisms can obtain
food material containing combined (" organic ") nitrogen ? In
this case we should substitute for coarse chemical experiments
in vitro the more delicate reactions of the living animal body;
any nitrogenous substance which is recognized by living proto-
plasm as being a suitable pabulum for it would be defined as a
proteid. If this is the case, we are thrown back at once on
Ehrlich's theory, and can define the proteids as substances which
unite in a specific manner, haptophore to receptor, with the living
molecule of protoplasm. If we do this, we may admit that the
toxins and enzymes, whatever their chemical reactions, and
whether they are or are not food substances for the cells, combine
with protoplasm in the manner which is characteristic of proteids,
and are to be regarded as proteids when looked at from this point
of view. And we must not forget that many of the substances
which we regard as toxins are, as a matter of fact, of value in
nutrition to the animal which forms them, and only act as poisons
in strange species. Thus, the ichthyotoxin of eel serum must
nourish the cells of the eel, but is a powerful toxin to almost
all other species. Whether any animal can utilize tetanus toxin,

IO2 CONDITIONS OF ANTIBODY PRODUCTION

diphtheria toxin, etc., as cell pabulum is doubtful, but there are
some experiments which go to show that this is the case. Thus,
the rat is relatively immune to diphtheria toxin, and shows no
symptoms after the injection of an amount which is lethal for
many rabbits, and this immunity is not due to the presence of
antitoxin. This being so, it seems clear that the diphtheria toxin
disappears from the blood of the rat in virtue of forming a com-
bination with the cells of the latter without injuring them, and we
may fairly assume that it is of value in nutrition, although, of
course, this assumption is not open to direct proof. In a similar
way tetanus toxin disappears rapidly from the blood of scorpions,
no antibody being formed.

Thus it appears to be a logical sequel of the acceptance of
Ehrlich's theory that we may define proteids as substances which,
when injected into suitable animals, give rise to the production of
antibodies. We have now to glance for a moment at the ques-
tion of the suitability of different animals for the production of
antibodies.

Researches with precipitins show that the proteid substances
which are present in the serum of any species of animal are
different from those which occur in any other. We must regard
the molecule of proteid of any type (say of a globulin) as being of
great complexity, and capable of many slight modifications, which
are inappreciable to gross chemical tests, but which are perfectly
obvious to the living animal cells. Human globulin differs from
horse globulin, and this, again, from sheep globulin. Further,
some facts go to show that differences exist between the proteids
of animals of the same species, and that the proteids in the serum
of one man or horse are not exactly the same as those in another
man or horse. These differences come in as a result of the last
step in the process of digestion ; the food materials are broken
down into simpler bodies in the alimentary canal, and built up
again in their passage through the epithelial cells which line that
structure, and in this process receive certain distinctive features
peculiar to the species or to the animal in question. As a result,
the body fluids of any animal contain proteids which have been
deliberately adapted for the nutrition of the cells of that animal,
but which may be quite useless or even toxic for those of another
species. The prime requisite of their suitability for nutritive
purposes is that the proteid molecules should possess haptophores
which " fit " the receptors of the cell molecules ; the second,

THE ORIGIN OF ANTITOXIN — THE SIDE-CHAIN THEORY 103

equally important, is that the proteid molecules shall be " diges-
tible " by the cell. If the first requisite fails, the proteid molecule
fails to unite with the cells ; it may remain for long periods in the
blood (as is the case with tetanus toxin in the blood of Oryctes,
where it remains months after injection), or may probably be
eliminated with the excretions, or be otherwise dealt with. In
this case no antibody can be formed.

When the injected proteid molecule finds suitable receptors and
unites with them, but when the resulting compound is not suit-
able for the nutrition of the cell, and cannot be " digested " by it,
the case is different. The receptors which are occupied by the
molecules of proteid are useless for the cell, and the conditions
which we have discussed in dealing with the action of toxins are
reproduced, although there may be no toxic action ; the receptors
are useless and have to be regenerated, and under suitable circum-
stances may be produced in excess and form antibodies. The
criteria of the suitability of any animal for the production of an
antibody to a given proteid are therefore three in number :

1. This proteid must not exist naturally in the blood of the
animal.

2. It must possess haptophores which " fit " the receptors of
the animal cells.

3. The compound thus formed must be " indigestible " by the
cell and useless for its nutrition.

These are the theoretical criteria, which are derived from a
study of the side-chain theory, and, on the whole, experimental
research agrees with them, and thus corroborates the theory to an
extraordinary extent. For example, we need scarcely refer to
criterion i : we know that precipitins do not occur normally in
the blood of any animal to the proteids which circulate in that
fluid ; it is only the injection of a foreign proteid which causes
their production. Further, the more closely allied are two animal
species, the less the production of precipitin when serum from the
one is injected into the other. Thus, in order to prepare a potent
antihuman serum, the blood of a man must be injected into a
rabbit, fowl, etc., not into a monkey. If it is true, as appears to
be the case, that the serum of an animal causes the production of
precipitins when injected into another of the same species, this
does not affect the argument, for these precipitins only occur in
minute amount, and simply show that the fluids and tissues of
different animals of the same species are not in reality identical.

IO4 CONDITIONS OF ANTIBODY PRODUCTION

There is nothing surprising in that when we consider how different
persons vary in constitution and susceptibility. Again, it is not
uncommon to find auto-agglutinins in human blood — i.e., sub-
stances which agglutinate human red corpuscles. When this is
the case, it is found that these agglutinins have no action on the
red corpuscles of the persons from whom the serum is taken, but
only on those of other individuals ; the stroma of the corpuscles
cannot act as an antigen to the cells with which it comes normally
into contact. (The way in which these auto-agglutinins are
called into existence, and their meaning, if any, or use in the
economy, are still unexplained.)

The existence and mode of formation of the secondary anti-
bodies— i.e., the antibodies to antibodies — point in the same
direction. Thus, if we inject typhoid bacilli or their proteid
constituents into rabbits or other animals, the specific antibody —
the agglutinin — is produced, and accumulates in the blood in con-
siderable amount, but it does not produce anti -agglutinin. If,
however, we inject this serum into an animal of another species,
and preferably one far removed from the first zoologically, the
production of anti -agglutinin occurs. The agglutinin produced in
any one species of animal must have, in addition to its peculiar
clumping properties, the general constitution of a proteid charac-
teristic of that animal, and hence be devoid of the power of
forming antibodies whilst in the blood where it was produced.
Similar phenomena occur in the production of other haemolysins
(amboceptor), which will be referred to subsequently.

With regard to the second criterion — the possession of hapto-
phores which fit the receptors of the protoplasm of the animal
into which it is injected — it is only necessary to say that, as far
as can be traced, the substances which produce antibodies do, as a
matter of fact, disappear from the blood of the animal in a short
time— a few hours or days. As has been pointed out, the converse
of this is not necessarily true ; the toxin may disappear from the
blood, and yet no antitoxin be formed, as in the case of tetanus
toxin in scorpions. Here we have assumed that the scorpion
contains naturally a proteid with a haptophore closely allied to
that of tetanus toxin, that the toxophore of the latter is without
action on the cell with which it is linked, and that the latter can
use the former as pabulum in the same way as it uses the
molecules of normal occurrence in its blood.

The third criterion is that the proteid molecule shall not be

THE ORIGIN OF ANTITOXIN-— THE SIDE-CHAIN THEORY 105

capable of assimilation by the protoplasm to which it is anchored ;
it is not necessary that there should be a toxophore group which
injures the cell. In this case the production of the antibody will
not be accompanied by any symptoms of disease. The proto-
plasmic molecule will have some of its receptors occupied and
rendered useless by the foreign proteid, and will have to regenerate
others of the same nature, but it will not be injured in any way in
the process. It is in this way that we explain the formation of
antitoxins as a sequence to the injection of toxoids which have
retained their haptophore groups, but lost their toxophores, as
well as the formation of agglutinins (to non-pathogenic bacteria),
precipitins, etc.

We pass on to another question — that of specificity. We have
seen that the antitoxins are, in general, adapted only to neutralize
the toxins which lead to their production. To this rule there are
a few exceptions, real or apparent. Thus, antirobin serum has
also an action against ricin, tetanus antitoxin has apparently
some action on snake-venom, anti-snake venom serum neutralizes
scorpion venom, etc. The explanation is doubtless that the
venoms in question have haptophore groups which closely re-
semble one another in their chemical affinities, whilst differing
from those of other toxins. Thus, Ehrlich has suggested that
robin is the toxon of ricin, in which case the combining portions
of the molecules of the two would be identical.

In any case, the side-chain theory seems to fit quite well with
ascertained facts. It explains the specificity ; it is the receptors
which unite with the injected toxin which are produced in excess,
and which therefore form antitoxin. But in the complex changes
in metabolism which must take place in the poisoned cell it is
quite easy to imagine that, under certain circumstances, other
receptors may either be formed, in slight excess as a result of the
general stimulation of the cell, or may be cast off in slight excess
as a result of necrotic or autolytic processes taking place therein.
This may be the explanation of some of the apparent exceptions,
especially of those in which the serum has but slight antitoxic
power on the toxin which was not injected, whereas it is much
more potent on that which was.

The next question, and a much more difficult and important
one, deals with the site of production of the antibodies. Ehrlich's
original idea was that the antitoxins are produced from those cells
on which the toxins act — i.e., on the susceptible cells. For instance,

I06 TETANUS TOXIN AND BRAIN-TISSUE

in the case of tetanus, he supposed that antitoxin is formed by the
cells in the central nervous system, and explained the great difficulty
of immunizing animals to this toxin by pointing out the enormous
susceptibility of the cells to the action of this poison ; it is only
in the cells of the central nervous system that antitoxin can be
formed, and these cells are extremely easily killed by the toxin,
and are necessary for life. The chief evidence in favour of this
theory is derived from the experiments of Wassermann and of
Romer, which we shall now consider seriatim.

Wassermann's experiment may be regarded as a corollary to
that of Ransom, who found that, after injection of tetanus toxin
in pigeons (which are, in the ordinary sense of the word, in-
susceptible to that substance), he could extract it from all
substances except the brain. Wassermann and Takaki found
that the mixture of tetanus toxin with an emulsion of the central
nervous system was no longer toxic when injected into susceptible
animals ; in other words, that the emulsion of central nervous
system behaved like antitoxin. Emulsions of other organs are
devoid of this power. They will combine with tetanus toxin, but
will yield it again when injected into susceptible animals or
macerated in salt solution. It appears, therefore, that the central
rjervous system is the seat, and the only seat, of an antitoxin-like
substance, and it is thus rendered at least probable that when
antitoxin appears in the blood it is due to its release from the
cells which contain it normally — in other words, from the cells of
the central nervous system, the cells whichX attacks. The case
of tetanus is the only one in which a toxin appears to have a
definite selective influence on one group of cells, and is therefore
unique in providing a means whereby Ehrlich's supposition as
to the origin of antitoxin may be tested. It is no wonder that
Wassermann's experiment has been submitted to an extra-
ordinarily careful investigation, the results of which we must
outline briefly. Some of these experiments seem to point to the
truth of Ehrlich's assumption. Thus, Blumenthal showed that
after injection of tetanus toxin into living animals the spinal cord
lost its property of fixing the toxin in vitro, the receptors of the
cells being already occupied with that substance. Further, boiled
brain substance loses its power of neutralizing toxin, just as anti-
toxin does on being heated. He also showed that the central
nervous system of fowls, which are but slightly susceptible to
tetanus, has but little power of binding that toxin, so that emul-

THE ORIGIN OF ANTITOXIN —THE SIDE-CHAIN THEORY 107

sions of the brain must remain in contact with the toxin for some
time before the latter is neutralized.

It was urged by many writers that the brain cannot be the site
of production of antitoxin, since the injection of tetanus toxin
direct into the brain of immunized animals causes the develop-
ment of tetanic symptoms. This argument is obviously fallacious.
A cell that is secreting antitoxin still possesses side-chains suitable
for union with the toxin ; it possesses them, indeed, in increased
amount, and is therefore, if anything, more susceptible to the
action of the toxin. The antitoxin which is circulating in the
blood only prevents the toxin from reaching the cell, acting much
in the same way as a lightning conductor.

Further research, however, seems to prove definitely that the
neutralizing properties of brain substance are not due to the
presence of a substance having any real likeness to antitoxin.
Thus, Metchnikoff showed that when an emulsion of frog's brain
is mixed with tetanus toxin no neutralization takes place, although
the frog is, under certain circumstances, susceptible to the action
of the toxin ; these researches were corroborated by Courmont
and Doyon. Arguing from this and other facts, Metchnikoff
attributed the fixation fo tetanus -as*itoxin by the central nervous
system to the presence in the latter of fatty substances ; these are
absent from the brain of the frog. In support of this view, several
observers found that substances such as lecithin, cholesterin,
tyrosin, etc., have also the power of neutralizing various toxins.

It was objected to these researches of Metchnikoff and Cour-
mont and Doyon that these experimenters did not leave the
tetanus toxin in contact with the brain substance for a sufficiently
long time. But this seems unnecessary, since the emulsion of the
central nervous system of higher animals will neutralize tetanus
toxin if the mixture be injected immediately, or even if the
two substances are injected separately, though in this case the
necessary amount of brain substance is larger. And Danysz
showed that if a neutral mixture of brain and toxin be macerated
for a long time in salt solution the toxin is again set free, in which
again the phenomenon is unlike that presented by antitoxin in its
action on toxin.

The researches of Morax and Marie also point in the same
direction. They showed that the fixative power of the brain is
almost entirely destroyed by drying, whereas, as we know, that
process has practically no effect on antitoxin. The researches of

io8 ROMER'S EXPERIMENTS

Dmitrevsky are also of great interest, and almost constitute a
crucial experiment.

If Ehrlich's theory is true, and if antitoxin is produced by the
budding-off of the receptors in great numbers, it ought to follow
that the brain of an immunized animal, which contains these
receptors in abnormally great amount, ought to have a greater power
of neutralizing toxin than that of normal animals. But Dmitrev-
sky's experiments show that this is not the case, and constitute a
strong proof against the theory in its original form. It seems fair
to conclude that Wassermann's phenomenon is due to some
accidental property of the central nervous system, or possibly to
the presence of fats, and must not be quoted as evidence of the
origin of antitoxin from the cells on which it chiefly acts.

Romer's experiment was made with abrin, a substance which
has the power of causing violent conjunctivitis. He made use
of rabbits, instilling small but increasing amounts into the right
conjunctiva. After three weeks the animal was killed, and each
conjunctiva dissected off, ground up with one lethal dose of abrin,
and injected into animals. The right conjunctiva, that into which
the abrin had been instilled, was found to act as an antitoxin and
to neutralize the abrin, whereas the left was devoid of this power.
Of course, if the process were carried on for a long time the abrin
would be absorbed into the system, and there would be a general
production of antitoxin, but at first the process appears local.

These researches are interesting, but do not seem to have any very
direct bearing on the question at issue. They show — what does
not require proof — that tissues which are not reached by the toxin
do not produce antitoxin ; but the question as to whether the
latter substance is produced by the cells which are peculiarly
susceptible to the toxin is not elucidated. There are in the
conjunctiva numerous structures — connective-tissue cells, blood-
vessels, endothelium, epithelium, leucocytes, etc. — and these
experiments afford no means of gauging which of these are
affected by the toxin or which produce antitoxin. The behaviour
of the leucocytes is of especial interest ; they are present in the
inflamed conjunctiva in increased amounts, and the question may
be asked whether the apparent antitoxic action of the right con-
junctiva may not be due to these cells, which are present in but
small numbers in the left. This view is rendered more probable
from the further researches of Romer, who showed that in
immunized animals antiabrin is present in greater amounts in the

THE ORIGIN OF ANTITOXIN — THE SIDE-CHAIN THEORY

organs rich in leucocytes, such as the spleen, lymph glands, and
bone-marrow. This view harmonizes well with that of Metch-
nikoff, which we shall consider subsequently.

The chief arguments against the origin of antibodies from the
cells which are especially acted on are those of Metchnikoff on
the production of antispermotoxin, and from the study of other
cytotoxins. The first series of experiments are of fundamental
importance, and will be discussed here, although the substances in
question are not simple bodies like toxins, but have a much more
complex structure. The main facts are as follows : The injection
of spermatozoa into living animals is followed by the production
and appearance in the serum of certain substances which have
the power of immobilizing and clumping the spermatozoa of the
species from which the spermatozoa used for the injection was
taken ; we may call the substances spermotoxin. If the serum of
animals which has been prepared in this way is now injected into
animals of another species, the result will be the formation of an
antispermotoxin analogous with antitoxin, which inhibits the
action of the spermotoxic serum in vitro on the spermatozoa.

Now on Ehrlich's theory we may assume that this anti-
spermotoxin is derived from the cells from which the spermotoxin
is derived — i.e., from the spermatozoa. But Metchnikoff proved
that this is not the case, since the substance is developed as the
result of the injection of spermotoxic serum into castrated males
or into young animals of either sex. Antispermotoxins, therefore,
are not necessarily derived from sperm cells. This is certainly a
very striking result, and one that tells against the side-chain
theory, but it is not conclusive for this reason — the cytotoxins are
not usually sharply specific. Thus a cytotoxic serum made by the
injection of kidney cells has a profound action on the kidney, but
it usually has some haemolytic action also, and may affect other
cells as well. There do not seem to be any experiments bearing
on the point, but it is at least probable that spermotoxins have
some action on cells other than spermatozoa, and if this is the
case Metchnikoff 's experiments, considered as a proof against the
side-chain theory, fall to the ground.

The mode of formation of the antihaemolysins also calls for short
notice in this connection. The haemolysins are, in some cases at
least, more sharply specific than some of the cytotoxins, acting
apparently almost exclusively on the red blood-corpuscles. It
should follow on Ehrlich's theory that the antihaemolysins are

IIO SITE OF ANTITOXIN FORMATION

derived mainly from the red blood-corpuscles. This view,
however, seems very difficult to believe when we consider that
these substances are devoid of protoplasm and nuclei, and do not
present any of those appearances indicative of active metabolism
or secretion which we should expect in the case of a cell perform-
ing so profound a physiological function as the production of an
antitoxin.

Nor is there any evidence, except in the one doubtful case of
the action of brain substance on tetanus aa^koxin, of the presence
of substances like antibodies in the normal tissues and organs, in
spite of the fact that they must contain receptors which would
neutralize the toxins to which they are tested. Thus Calmette
did not find that emulsions of brain possessed any neutralizing
action on snake venom, although that substance certainly acts on
the central nervous system, and there are many similar researches
on the actions of other emulsions of organs on other toxins, all
with negative results. We may quote, for example, those of Blum,
who submitted various organs of the horse — liver, spleen, etc., all
of which are certainly acted on by diphtheria toxin, since they
show definite morphological lesions in acute cases of fatal intoxi-
cation— to prolonged autolysis, and found nothing in the nature of
an antitoxin for diphtheria or snake venom in the resulting
material. In one case — that of autolyzed lymphatic glands of the
calf — he found a substance which reacted like tetanus antitoxin.

On the whole, therefore, it would seem that there is no good
evidence for the theory that the cells which are acutely and
powerfully poisoned by means of toxin are those from which the
antitoxin is derived. It is certain, however, that this substance
must be formed from cells which are reached and affected in some
way by the toxin. The cells which especially require discussion
are those of the connective tissues and the leucocytes.

Attention is drawn especially to the connective tissues from the
fact that the most powerful production of antitoxin is usually
elicited by a subcutaneous injection of toxin ; intraperitoneal
injection is in general less potent, and intravenous injection less
still, and may not be followed by any production of antitoxin
at all. The same law holds, and even more strongly, in the
production of the other antibodies, although some exceptions
occur.

Now in some cases it is true that these toxins have a local
pathogenic action on the tissues into which they are injected, but

THE ORIGIN OF ANTITOXIN — THE SIDE-CHAIN THEORY III

in others this is extremely slight, or not more than that due to
the injection of any so-called " inert " fluid into that situation.
Consider in this connection the reactions of certain animals to
tetanus toxins. When this substance is injected into guinea-pigs,
the effect is practically the same whether the injection is sub-
cutaneous or intracerebral. In rabbits, on the other hand,
a very much smaller dose suffices to kill if injected into the brain
than is necessary if the injection is subcutaneous. Now rabbits
are less susceptible to tetanus toxin than are guinea-pigs, requiring
relatively larger doses to bring about a fatal issue. On the other
hand, it is easier to prepare tetanus antitoxin from rabbits than
from guinea-pigs by the injection of tetanus toxin, though difficult
in both cases. Compare these results with those obtained from
fowls. These animals are resistant against tetanus toxin (except
in absolutely enormous doses) when injected subcutaneously, but
are easily affected when the injection is made into the brain
or subarachnoid space. The probable explanation turns on the
varying affinity of the toxin for different parts of the body. We
may assume in all cases that the tetanus toxin acts as a poison
only on the brain and spinal cord, and that in the guinea-pigs it
has practically no affinity for the subcutaneous tissues, and
that when injected into this situation it passes ultimately to the
brain almost without loss. In rabbits some of the toxin is united
to the tissues, and only a part reaches the central nervous system,
whilst in fowls the affinity of the tissues for the toxin is so great
that the whole of the latter is absorbed and the brain escapes
injury. Now it is practically impossible to obtain antitoxin by
injecting toxin into the guinea-pig, difficult to obtain it from the
rabbit, but easy to obtain it from the fowl, so that it would appear
that it is the tissues which are not especially vulnerable to the
toxin which yield it. If this is true, it is no argument against the
validity of the side-chain theory. It simply indicates that the
cells which are profoundly affected with toxin are thereby
rendered unable to produce antitoxin in virtue of the toxic action
of the poison, the other conditions being suitable. We know this
to be the case from the fact that we can produce tetanus
antitoxin from guinea-pigs as a result of the injection of toxoid,
which we must assume to unite entirely with the cells of the
central nervous system, just as the toxin does, though without
injuring them. We may assume, therefore, that antitoxin is
produced from cells which unite with toxin or toxoid, but which

112 SIGNIFICANCE OF LEUCOCYTOSIS

are not too profoundly injured thereby to perform their functions
of self -nutrition in a normal manner.

The question whether we can narrow down the issue to one
particular set of cells which always occur in the connective and
other tissues — to wit, the leucocytes —is a more difficult one. We
have already adverted to the researches of Romer, who found that
antiabrin was present in immunized animals in greatest amount
in the organs that are rich in leucocytes, and Metchnikoff has
brought forward a whole series of researches which point strongly
in the same direction. The discussion of some of these will be
deferred to a subsequent chapter, since they hardly seem to
indicate that the leucocytes form antitoxin, though they do go
to prove that they discharge another and equally important
function in the production of immunity to toxins. Here it will
only be necessary to point out some well-known and important
facts which appear to have a bearing on the question of the
origin of antitoxin. We allude especially to the chemotactic
attraction for leucocytes which is showrn by almost all toxins
when not present in too great an amount. In virtue of this
attraction the most marked and constant feature of intoxication
is the presence of an excess of leucocytes, and this is, in general,
most marked when there is a balanced contest between the toxin
and the host — i.e., when the latter is neither destroyed at once
nor shows but little effect of the injection. It is practically a
general rule that wherever the host is victorious in the struggle
with the toxin there is an excess of leucocytes ; and even where
the fatal issue is not averted there is some leucocytosis, unless
the intoxication is a very grave one and very rapidly fatal. This
is well known in human pathology in the infective processes,
where a diminution of the number of leucocytes is a bad sign,
whereas hyperleucocytosis is a good sign in so far as it indicates
that the infection, though intense, is being combated by the
patient by the best means at his disposal. Many researches have
indicated that the same law holds in these intoxications. Thus
Chatenay, in an exhaustive study on the reactions of the more
important toxins, finds that a dose that is rapidly fatal brings
about a fall in the leucocytes, whereas smaller doses cause hyper-
leucocytosis; thus, in the guinea-pig about 100 lethal doses are
required to prevent the occurrence of the latter. Now we
can scarcely imagine that so widespread and constant a phe-
nomenon is without some meaning or without some use to the

THE ORIGIN OF ANTITOXIN — THE SIDE-CHAIN THEORY 113

organism, and since it always occurs and is most marked under
conditions in which antitoxin is formed, it is highly probable that
the leucocytes may be, entirely or in part, the origin of the
antitoxin. But we must not forget that in many cases the
leucocytes play other parts in the struggle against the infections ;
for example, it is plausible to argue that their presence in excess in
infected areas is an attempt to combat the bacteria themselves, and
not the toxins which they produce. In this case their access to
areas into which a toxin had been injected might be useless, and they
might simply be attracted there under the mistaken supposition that
there would be bacteria to combat as well as toxins. But this
view is rendered improbable in face of the numerous facts that
have been brought out by the French school with regard to the
absorption of toxins and other poisons by the leucocytes. Thus
Metchnikoff showed (and it is a fact of the utmost importance)
that if an aseptic exudate be produced in a fowl that has been
injected with tetanus toxin, that .substance can be demonstrated
in large amount in the leucocytes of the exudate. Again, Vaillard
and Vincent have shown that if tetanus bacilli or spores washed
free from toxin are injected into a guinea-pig, they become
surrounded by leucocytes, and that under such circumstances no
tetanic symptoms occur.

With regard to the action of other poisons the facts are still
*• more striking. Thus rabbits are far more susceptible to the
v intracerebral injection of atropin than to the same substance
injected into the circulation; and when the injection has been
made in the second manner, atropin can be demonstrated from
the leucocytes, whereas it is not present, or only in very small
amounts, in the plasma and red corpuscles. Similar facts have
been found for arsenic and other poisons. Still more striking are
the researches of Besredka with regard to arsenic. He studied
first the action of the trisulphide of arsenic, an almost insoluble
substance, yet very poisonous. When this was injected in small
amounts into the peritoneal cavity of a guinea-pig, there was a
marked increase in the mononuclear leucocytes in that situation,
and in a short time these cells had ingested the whole of the
arsenic, which could be recognized as fine granules in their
protoplasm. These gradually became smaller and ultimately
disappeared, being probably converted into a non-toxic compound
— perhaps by some process analogous with antitoxin formation.
In a second series of experiments he tested the action of the

8

114 ACTION OF LEUCOCYTES ON TOXINS

same substance when shielded from the leucocytes by being
enclosed in permeable bags before being placed in the peritoneum ;
in this case the animal died, the arsenic doubtless becoming
dissolved, and thus escaping the action of the leucocytes. He
found also that substances which diminished the number of the
leucocytes in the peritoneal fluid aided the action of the arsenic,
whilst those which caused a leucocytosis diminished it. He found
similar facts with relation to the action of a soluble salt of arsenic,
and showed that in animals which had been immunized to that
substance the arsenic is especially taken up by the leucocytes.
There can be no doubt, therefore, that the leucocytes play a role
of the utmost importance in the defence of the body against the
ordinary poisons.

All things being considered, we may deduce that antitoxin is
formed from any cell with which the toxin can combine, provided
that that cell is not too profoundly injured in the process, and
may attribute to the leucocytes an important, though not an
exclusive, role in this process. It is hardly necessary to say that
this does not disprove the side-chain theory ; we must regard the
leucocytes as being cells which are specially told off for the
defence of the organism against infections and intoxications.
They are, in consequence, specially immune to the action of
toxins, being able to resist amounts which are injurious to the
more highly organized tissues ; that this is the case is proved
by their invasion and persistence in a living state in areas in
which the tissues are profoundly injured by a toxin. It is not
difficult to believe that these resistant cells, which have a special
predilection for intoxicated areas in which antitoxin is especially
required, should be the source of that substance. We have, then,
only to regard them as possessing suitable haptophore groups and
as being relatively insusceptible to the toxophore groups — both of
which are inherently probable — to see that they fulfil in a special
way the criteria which the side-chain theory demands for cells
which are to produce antitoxin when exposed to the action of
toxins.

CHAPTER VI
IMMUNITY TO TOXINS

THE mechanisms by which the injurious effects of toxin in the
blood and tissues are combated are exceedingly complex, and
may be divided roughly into two series. In the first and less
interesting, though perhaps not less important, the mechanism
may be called (for want of a better term) non-specific. It acts
equally on all or many toxins, and recovery from the effects of the
poison is not necessarily or usually followed by any immunity
thereto. For example, the dilatation of the vessels and accelera-
tion of the blood-stream which takes place in an inflamed focus
may be regarded as one of the natural provisions for avoiding the
too intense action of a toxin in one small area. The poison pro-
duced is rapidly swept into, and diluted by, the whole blood-
stream, and if not produced in large amounts may not reach any
tissue in amount sufficient to cause much damage. Perhaps, also,
the proteolytic action of the enzymes frequently produced (mainly
from the polynuclear leucocytes) in the inflammatory focus are of
some importance. Toxins are very fragile bodies, and it is ex-
tremely probable that the potent enzymes which occur in all fluids
rich in polynuclear leucocytes may bring about their destruction in
considerable amounts.

The more purely chemical processes are hardly known in their
application to true toxins, but they have been largely studied in
regard to the methods by which the organism combats deleterious
substances produced in the course of metabolism, or absorbed
from the alimentary canal. The latter are for the most part
bacterial products, though not true toxins. It is, however, quite
possible that these latter substances are destroyed by the same
processes, which Herter classifies as (i) oxidation, (2) hydration
and dehydration, and (3) various syntheses.

As an example of the process of oxidation, we may consider the
formation of indol, a toxic substance, which is formed by B. coli

"5 8—2

Il6 NEUTRALIZATION OF POISONS

and other organisms in the alimentary canal, and which probably
plays some part in the production of the symptoms of auto-intoxi-
cation due to abnormal digestive processes occurring in the small
intestine. It undergoes oxidation to indoxyl, a much less toxic
substance, which then combines with sulphate of potash to form
indoxyl-potassium-sulphate, or indican. There is reason to believe
that this process takes place to a very large extent in the liver,
one of the chief functions of which is the destruction or neutraliza-
tion of poisonous substances conveyed to it by the portal circula-
tion. Thus the toxic dose of nicotin and of some other alkaloids
is much less when the injection is made into the general circulation
than if the substance be injected into the portal vein. This, it
must be noticed, is not due merely to the elimination of the poison
in the bile, though that may happen. The liver cells have the
power of actually destroying certain of the alkaloids. Thus
hyoscyamine mixed with surviving liver pulp is rapidly destroyed.
In the same way, indol, phenol, and skatol, all toxic substances of
bacterial origin, cannot be recovered by distillation after contact
with liver cells. It appears, therefore, that much of the immunity
to non-specific poisons is dependent on the functional activity of
this organ, which is, so to speak, interpolated in the blood-current
in order that it may rid the circulation of certain poisons, and
that these are eliminated in the bile or submitted to various
chemical modifications, in the course of which they become inert,
or at least less toxic.

Examples of the neutralization of poisons in the body are
numerous, the most important being, perhaps, the formation of
urea, a relatively non-toxic substance, by the synthesis of ammonia
and carbonic acid, followed by dehydration, this process taking
place mainly or entirely in the liver. Other syntheses, however,
take place in other organs — for example, the union of glycocoll
and benzoic acid is performed in the kidney. And we shall see
subsequently that the leucocytes and leucocytic organs have a
very special and important duty to perform in dealing with poisons
of all sorts, both organic and inorganic, though the exact chemical
processes that they bring about are unknown. It will be more
convenient to defer the consideration of these cells for the
present.

The exact application of these "non-specific" processes of
detoxication to the bacterial toxins has not yet been worked out,
but there can be little doubt of its great importance. Compare,

IMMUNITY TO TOXINS 117

for example, the severity of the symptoms due to an abscess due
to B. coli and the lack of symptoms due to the absorption of the
toxins of this organism which occurs from the alimentary canal,
and very probably from bacilli which actually make their way
through its walls into the lacteals. There can be but little doubt
that the symptoms of intoxication in the former case are due to
the direct access of the bacterial poisons into the blood-stream
without having first to traverse the liver. Again, in a case of
diphtheria of moderate severity virulent bacilli are often present
in abundance so great, that by comparison with the toxin-forming
powers of the organisms in vitro we might expect a rapidly fatal
issue from the toxin absorbed, and yet the symptoms of general
intoxication may not be severe. In all probability only a small
fraction of the toxin formed actually reaches the distant tissues.
Some may be dealt with by the zone of leucocytes which under-
lies the layer of diphtheria bacilli, but it is also possible that some
of the toxin is destroyed in the blood-stream, perhaps by a process
of oxidation. These non-specific methods of dealing with toxins
are matters more for the physiologist than for the pathologist,
depending as they do simply on the perfect discharge of the
normal functions of the body. They are of great importance —
greater, perhaps, than pathologists usually realize, the interest
attaching to them being so much less than that which is connected
with the study of the antitoxins and similar bodies. They are to
be regarded as the first line of defence against poisons of all sorts.
When a small dose of a specific bacterial toxin gains access it is
probable that the natural physiological methods in daily use for
dealing with the natural poisons (formed in metabolism or absorbed
from the alimentary canal) are in most cases applicable to it also,
and no specific process has to be brought into action. In confir-
mation of this view is the fact that diphtheria antitoxin is not
necessarily found in the blood after natural recovery from diph-
theria, suggesting that in these cases the processes at work are
non-specific.

But when the toxin is present in greater amount this process
may fail, and it may do so in consequence of the action of the
poison on the tissues and organs normally concerned in the defence
of the body. The neutralization of toxins, etc., in the liver and
other regions can be carried out best when these organs are in a
state of high functional activity, and when this is impaired some
of the deleterious substances will escape their action. Thus we

Il8 FORMS OF IMMUNITY TO TOXINS

get a vicious circle. The poison injures the liver or other organ
and impairs its defensive powers. More poison is allowed to
circulate in the blood, and the liver is injured still more. When
this occurs we are thrown back on the specific methods of dealing
with the toxins, which take longer to come into action, and which
may be regarded as the second line of defence. In one case,
however, we must regard them as of chief importance. Thus in
tetanus the toxin is produced locally, and ascends the nerve -
trunks without entering the blood-stream or being carried to the
liver or leucocytic organs. Here, then, in all probability, specific
methods have to be resorted to, and the chance of recovery depends
on their early development.

The question of specific immunity to toxins, and of recovery
from intoxication with true toxins, may be considered under three
heads :

1. Acquired immunity, due to disease or to vaccination with
toxins or toxoids (chemical vaccination).

2. Antitoxic or passive immunity, due to the injection of anti-
toxic serum, or possibly to its natural occurrence in the blood.

3. Natural immunity, not due to vaccination, and not accom-
panied by antitoxin in the blood.

It will be convenient to consider them in this order.

Natural recovery from toxic diseases such as diphtheria and
tetanus is not necessarily accompanied by the appearance of anti-
toxin in the blood — at least, not in demonstrable amounts. Further,
it appears from the researches of Abel that antitoxin, when it is
formed at all, does not make its appearance until about the eighth
day of the disease, at which period the severity of the toxic
symptoms may have already begun to decline. What may be
called the simple antitoxic theory cannot be maintained. Under
natural conditions the process of recovery is not due simply to the
production of antitoxin, though it is possible that this comes into
action in prolonged cases.

Further, acquired immunity to toxins is not due solely to the
presence of antitoxin in the blood. If it were so it should develop
proportionately to the development of antitoxin, and the two should
persist for the same length of time, and disappear together. This,
however, is not the case. We have already seen that animals in
the early stage of immunization to diphtheria and tetanus frequently
present a decreased amount of resistance, or hypersensitiveness,
to these substances, and may die with symptoms of acute intoxica-

IMMUNITY TO TOXINS IIQ

tion, in spite of the presence in the blood of amounts of antitoxin
sufficient to neutralize the toxin many times over. Similar appear-
ances may be seen in animals in which death from acute intoxica-
tion does not occur. Thus, in two horses which were treated at
the same time for the production of diphtheria antitoxin, one
showed but little sign of the action of the toxin, and improved in
general health during the treatment — it developed only 5 units of
antitoxin per cubic centimetre of serum ; the other suffered severely
in general health, and had ultimately to be killed — it developed
70 units per cubic centimetre. Similar phenomena are often met
with, and, as a general rule (to which there are numerous excep-
tions), the presence of a large amount of antitoxin in the blood
indicates susceptibility rather than immunity. The facts of the
later stages of antitoxin-formation may also be borne in mind.
Sooner or later in the history of any antitoxin horse there comes a
time when the amount of antitoxin begins to diminish, and would,
in all probability, disappear entirely if the injections were con-
tinued ; yet these horses are extremely resistant to the action of
toxin — more so, in fact, than animals with much antitoxin in the
blood. The degree of immunity, therefore, is not measured by
the amount of antitoxin in the blood, and we might argue that the
two are not related in any way. This, however, is absurd, in
view of the ascertained action of antitoxin in neutralizing the
effects of toxin in vitro, or of protecting against a dose surely fatal.
We may consider the role of antitoxin under two heads : (i) in
immunity, and (2) in recovery from disease.

Considering first the question why an animal dies in spite of the
presence of an excess of antitoxin in the blood, we must regard
the simplest and most probable explanation as one on which the
toxin-antitoxin molecule is looked upon as dissociable. This is
always the case on Arrhenius and Madsen's theory of the inter-
action of the two substances, whilst on the colloid theory the dis-
sociation only takes place for a short time after the compound has
formed, the union between the two substances gradually becoming
firmer and firmer. On this supposition the failure of antitoxin is
readily explicable : the two substances unite, and the inert mole-
cule is formed ; this dissociates, and the toxin-molecules, which
happen to be set free in the neighbourhood of susceptible cells,
unites with them. The removal of some of the molecules of toxin
allows more dissociation to take place, and ultimately the whole of
this substance is passed on to the tissues. We must assume that

I2O ROLE OF THE LEUCOCYTES

the compound formed between the toxin and the tissues does not
dissociate, as, indeed, appears probable.

But if this is the case, the effect of antitoxin in the blood would
be merely to delay the action of the toxin ; assuming this, how
are we to explain the preventive effect of this substance in passive
immunity and its curative effect in disease ? For it is only in com-
paratively rare cases in the early stages of antitoxin formation
that the phenomena under discussion occur, and in all other con-
ditions the presence of a sufficient amount of antitoxin in the
blood constitutes a perfect safeguard against the action of its
corresponding toxin. It is probable that the leucocyte is the all-
important factor necessary for the destruction of these specific
toxins, whether previously neutralized by antitoxin or not. We
must look upon the toxin-antitoxin molecule as one which can be
easily ingested and destroyed by the leucocytes, and the neutraliza-
tion of toxin by antitoxin as the first step in a double process, the
second being the destruction of the compound by the leucocytes.
If antitoxin is absent, the leucocytes may still deal successfully
with the toxin, if the latter be not too virulent, nor present in too
large an amount ; but if the leucocytes make default, the presence
of antitoxin may delay the lethal issue, though it is powerless to
avert it. We may, perhaps, compare antitoxin with opsonin,
which unites with bacteria and renders them suitable for ingestion
by the white corpuscles. Metchnikoff and his school have paid
great attention to the role of the leucocytes in intoxication, and
have brought forward very important and suggestive evidence,
pointing to the white corpuscles, and especially the large lympho-
cytes (macrophages) as the main source of antitoxin. This
question is dealt with elsewhere, and at present we have to discuss
only the role of the leucocytes in dealing with the toxins, either
alone or when neutralized by antitoxin.

The evidence proving the importance of the leucocytes in deal-
ing with unaltered toxins is abundant, and some of the more striking
facts brought forward by the French school have been briefly
summarized in the last chapter. The evidence for the removal
of toxin in combination with antitoxin by leucocytic action is less
direct, since it is obviously impossible to recognize this compound
by chemical or microscopical processes whilst in the leucocyte.
But it has been shown that where toxins are introduced in com-
bination with solid substances, such as carmine or brain tissue,
they become surrounded by large numbers of leucocytes, and the

IMMUNITY TO TOXINS 121

particles are taken up by these cells. Further, the injection of
toxin into an animal which already contains antitoxin in the blood
brings about a more or less marked leucocytosis. It is true that
the local leucocytosis brought about by the injection of neutral
mixtures of the two may be but slight, but we must remember
that substances in a state of solution are quickly absorbed and
carried to the lymph glands or bloodvessels, in either of which
situations leucocytes occur in plenty. Further, there are always
some leucocytes in the tissues, and the number may be sufficient
to deal with the amount of toxin-antitoxin injected, which, con-
sidered as mere weight, is always very small.

The effect of antitoxin on the toxin of B. pyocyaneus may also be
cited as a phenomenon capable of explanation on this supposition.
A few lethal doses of toxin are fully neutralized by antitoxin, the
law of multiple proportions holding up to a certain point. When,
however, more than ten lethal doses are injected the law does not
hold, and no amount of antitoxin will avert the fatal issue. Here
we must assume that the antitoxin has but a slight affinity for the
toxin, so that dissociation takes place rapidly, and the ultimate
cause of the destruction of the toxin to be leucocytic activity.
The single lethal dose is just more than the leucocytes can deal
with ; but when antitoxin is injected simultaneously the leucocytes
have time to come into action, and the toxin, gradually set free by
dissociation, is dealt with in detail instead of in a single dose.
There are, however, limits to this process, and when more than
ten lethal doses are present fully combined with toxin we must
imagine that dissociation goes on so rapidly that more toxin than
one lethal dose is set free before the toxin-antitoxin molecules can
be taken up. If this explanation is highly theoretical, it seems
the best that is available at present.

The difference between the effect of toxin in animals with anti-
toxin in the blood, but otherwise normal, and its effect in animals
in the early stage of antitoxin formation, is explicable if we regard
the hypersensitiveness as occurring in the leucocytes themselves
as well as in the tissues. When antitoxin is injected into a normal
animal, the leucocytes of which are functionally active, the con-
ditions for the immediate destruction of the inert molecule are
present, and no dissociated toxin gains access to the susceptible
tissues. But in hypersensitive animals the leucocytes will be
unable to deal with the molecule, or will perhaps be killed by toxin
liberated by dissociation taking place within their protoplasm, and

122 ROLE OF THE LEUCOCYTES

the toxin will ultimately reach the tissues. And it is probable
that the beneficial effect of early as compared with late injections
of diphtheria toxin may be dependent in part on the fact that in
this disease me leucocytes are very apt to suffer degenerative
changes which doubtless impair their defensive powers. Ewing
believes that the variations in the staining capacity of these cells
(loss of chromatin being an early and easily recognized sign of
their degeneration) might be utilized as a means of prognosis, and
points out that in cases which recover the staining quality of many
of the leucocytes undergoes a rapid improvement after the injec-
tion of antitoxin, whereas in fatal cases this change could not be
detected. As a rule the total number of leucocytes is diminished
immediately after the injection of antitoxin, but in some fatal
cases the previously existing leucocytosis may remain or even
increase. This fall in the total number of leucocytes is probably
to be accounted for by the neutralization of the toxin, so that
fresh white corpuscles are no longer attracted from the bone-
marrow by positive chemotaxis, whilst the old and injured cells
remain in the spleen or other internal organs, no longer circulating
in the blood. In cases which die the reduction in the number of
leucocytes may be followed by a subsequent rise.

Here also we must refer to numerous researches, emanating
more especially from the French school, which go to show that
substances which are certainly devoid of true specific antagonism
have nevertheless the power of preventing or diminishing the
action of toxins in virtue of causing an increased inflow of leuco-
cytes to the region into which the injections are made. This
question has been more especially investigated in regard to the
effect of this artificially produced local leucocytosis in causing
local bacterial immunity, but Calmette and Delecarde have shown
that the peritoneal injections of ordinary broth cause a certain
degree of immunity to abrin, whereas normal saline solution has
no such power. The former substance is a powerful agent in
attracting leucocytes, whilst the latter is much less potent in this
respect, though by no means devoid of activity. Substances which
have the power of increasing the action of the leucocytes in this
way are sometimes termed "stimulins," but the word should be
avoided. The action of leucocytes in any region may be in-
creased by many methods, the chief of which are : (i) attracting
larger numbers by means of substances having a positive chemo-
tactic action ; (2) stimulating the activity of the leucocytes ; and

IMMUNITY TO TOXINS 123

(3) altering the nature of the substances on which they are to act.
If the term "stimulin" is to be used at all, it should be rigidly
restricted to one that acts in the second way, and there is as yet
no strict proof that such a body occurs. Substances acting in the
third way are called opsonins, and, if the theories of antitoxic
action here outlined are correct, antitoxins have this action also.

Turning now to the process of recovery from bacterial intoxi-
cations in natural disease, we find but little evidence that the
formation of antitoxin plays any active part in the process. The
conditions are not as a rule suitable for the production of this
substance, since the constant flow of the toxin into the circulation
would lead rather to a summation of negative phases and the pro-
duction of hypersensitiveness. Further, the process of recovery
may have commenced before antitoxin appears in the blood — a
phenomenon which, when it occurs, we may regard rather as a
means of preventing reinfection than as a mechanism for the cure
of the original attack of the disease. (It may be pointed out that
in the case of diphtheria the studies of Ruth Tunnicliffe lead us to
believe that the process of recovery runs parallel to, and is due
to a rise in, the opsonic index, and that the main factor in the cure
of the disease is the removal of the bacilli by a process of phago-
cytosis, and consequent cessation of the absorption of toxin.) In
prolonged diseases it is possible that the production, and especially,
perhaps, the local production of antitoxin, may play some part in the
process. In other cases, perhaps, the toxin is dealt with by some of
the simpler chemical processes considered already and submitted to
various changes of oxidation, etc., and thus rendered inert ; but of
this we have but little knowledge, the main studies of the methods
in which the body deals with toxins having been devoted to the
phenomena occurring in highly immunized animals and brought
about by their sera.

Some experiments designed by Wassermann as evidence in
support of the side-chain theory may be alluded to in this con-
nection. He found that the injection of tetanus toxoids into
highly susceptible animals (guinea-pigs) has the power of render-
ing them partially immune to tetanus toxin for a short period.
Thus, when the toxoid was injected, followed in an hour's time
by unaltered toxin, the lethal dose of the latter was greater than
in a normal animal. On the other hand, after a day or two the
animal became more susceptible, being killed by less than the
minimal lethal dose for normal guinea-pigs. This he explained

124 RECOVERY FROM INTOXICATIONS

by supposing that the receptors of the sensitive cells were occupied
by the inert toxoid, so that part of the cells would be unable to
combine with the true toxin. The subsequent susceptibility he
explains by the over-production of new receptors. Now when
we consider that in acute disease of natural occurrence the pro-
duction of toxin (in cases that recover) is only temporary, and is
stopped sooner or later by the destruction of the bacteria, it seems
probable that some such process may occur and serve to shorten
the period during which the cells are exposed to the action of the
toxin. The true toxins are so fragile that it is highly probable
that a certain proportion of them are converted into toxoids during
or after the process of absorption into the blood-stream, and by
occupying their receptors diminish the susceptibility of these cells
to the true toxin, which, circulating in the blood, may be dealt
with by the leucocytes, liver, or other organs. Afterwards anti-
toxin would appear in the blood, but it would have no relation to
the early slight immunity of the tissues.

We may therefore recognize the following stages in the process
of recovery from a bacterial intoxication :

1. In the first the processes are mainly non-specific, the chief
being probably leucocytic action ; the dilution of the toxin in the
general blood-stream, and its elimination, either in a natural con-
dition or after it has undergone some chemical alteration ; the
partial destruction and conversion of toxin into toxoids, and sub-
sequent " blocking " of side-chains in highly important tissues.
In mild cases these may be sufficient for the restoration of health,
and no immunity, or the merest trace, may follow.

2. Where the process lasts longer antitoxin begins to be pro-
duced, and in all probability the time which must elapse before
this occurs varies greatly in different conditions, depending on :
(a) the constitution of the patient, (b) the dose of the toxin, and
(c) on the nature of the toxin. Thus, with regard to the last
point, it may be remarked that it appears to be extremely
difficult to prepare antitoxins to the endotoxins, and it is highly
probable that in recovery from such diseases as cholera and
typhoid fever the production of antitoxin is slight or absent.
Such diseases owe their symptoms mainly to the production of
endotoxin, and are combated mainly by the removal of the
bacteria which produce them. In the cases in which antitoxin is
produced, recovery at this stage depends on its presence, together
with the functional efficiency and sufficient numbers of the leuco-

IMMUNITY TO TOXINS 125

cytes, and the continuance of the factors by which the toxin is
combated in the first stage of the infection.

Passive antitoxic immunity consists in the artificial production
of this stage.

3. In the third stage, which is only reached in hyperimmunized
animals submitted to treatment for long periods, the conditions
are quite different, and do not appear to depend in any way on the
action of antitoxin, but approach more nearly to the state of
natural immunity to be treated subsequently. In this condition the
antitoxin in the blood falls greatly, and would probably disappear
entirely if the treatment were carried further, yet the degree of
immunity is very great. There are two or three theories which
have been invoked to explain this form.

In the first place, if we accept the side-chain theory, we may
ascribe it to the absence of receptors suitable for the toxin which
is being injected, and may suppose that the repeated and pro-
longed stimulation of the production of this particular receptor
has led firstly to a hypertrophy, and secondly to an atrophy of
these organs. This is readily conceivable, and might be compared
with the sequence of hypertrophy and failure of the heart so
frequently met with in Bright's disease, etc. In this case the
toxin would be unable to " anchor " itself to the cells, which would
thus escape its action, and the result would be one form of tissue
immunity. There is but little experimental evidence bearing
directly on this theory either for or against. The state is one
rarely seen even under experimental conditions, though of great
theoretical interest. It is found, however, that during the immuni-
zation of the rabbit with eel serum (a potent haemolytic agent) an
antitoxin or antihaemolysin is produced, whereas the corpuscles
themselves, if carefully washed from all traces of serum, are as
susceptible as before. If, however, the injections are repeated,
the antitoxin disappears, and, as Tchistovitch showed, about this
time the red corpuscles themselves become immune, not being
haemolyzed by eel serum, even although all traces of their own
serum is washed out. This we may fairly suppose to be due to a
loss of the receptors with which the molecules of eel serum
combine, and this may be taken as an experimental demonstration
of the occurrence of this form of immunity, which was predicted
by Ehrlich on theoretical grounds. Metchnikoff, it is true,
objects to this hypothesis, pointing out that if the receptors are so
necessary for the nutrition of the molecule it is impossible that

126 TISSUE IMMUNITY TO TOXINS

these latter should continue to live after the receptors had been
lost. But this seems based on a misconception of Ehrlich's
theories of cell-nutrition. We must imagine each molecule to be
provided with a very large number of varieties of side-chain, each
adapted to the seizure of a different form of food-molecule, so
that the complete loss of one variety does not imply that the cell
will suffer in nutrition. The only result is that more work will
be thrown on the receptors that remain. There is nothing im-
probable in this. It is in the highest degree likely that so
important a function as nutrition should not be dependent on the
functional integrity of every part of the molecule. The occurrence
of tissue immunity from loss of receptors is at least possible.

An alternative theory asserts that the cells of the body retain
their power of combining with toxin, but lose their susceptibility
to the action of the toxophore group of the latter, at the same time
losing their power of producing antitoxin. We may explain the
latter phenomenon by saying that the receptors in question
become sessile and incapable of being shed. There is experi-
mental evidence that some receptors are naturally of this nature,
and it is quite possible that ordinary deciduous receptors might
change in this way. As we shall see, tissue immunity in which
the toxin becomes anchored to the cell, but without injuring it or
stimulating it to produce antitoxin, is met with in some forms of
natural immunity, and it is quite possible that it may occur in the
process of hyperimmunization by vaccines, though direct proof is
lacking. Another process of somewhat similar nature may occur,
which has also its analogue in natural immunity. Thus when
animals are immunized to tetanus they still retain their suscepti-
bility to that toxin if the injection be made directly into the
brain. It is obvious, therefore, that the cells of the central
nervous system are still susceptible. Why, then, do the animals
not develop tetanus if the injections be made subcutaneously ?
We exclude, of course, the case in which antitoxin is present in
the blood and the leucocytes are functionally active. It will be
shown subsequently that some forms of natural immunity depend
upon the fact that some cells in the body unite with the toxin, and
have, indeed, a great affinity for it, but are not injured thereby,
whereas others have an affinity for it, but are killed by its action.
Obviously, if the toxin be injected into tissues of the former class
it will be absorbed, none will reach the susceptible second class of
cells, and no toxic symptoms will result. Now let us suppose

IMMUNITY TO TOXINS 127

that in immunizing a susceptible animal (rabbit or guinea-pig)
to tetanus the connective tissues acquire receptors suitable for
the toxin in question. They will absorb this toxin, be shed, and
the animal will develop antitoxin, whilst the tissues previously
susceptible — the brain and cord — will remain as susceptible as
before. This is probably what occurs, though the critical proof —
the demonstration of an increase in the toxin-absorbing power of
the tissues in acquired immunity, and thus of an increased number
of receptors — does not appear to be forthcoming. Dmitrevsky
investigated the amount of tetanus toxin neutralized by brain
tissue before and after immunization ; but since, as we have seen,
this process is probably different in nature from the neutralization
of toxin by antitoxin, his researches have not much weight in this
connection.

The third explanation is that of Metchnikoff, that the leucocytes
themselves have become immune to the toxin, and have thus
acquired the power of dealing with that substance in large
quantities, so that the tissues are sheltered from its action. This
theory simply shifts the immunity back to the leucocytes, and
thus does not really solve the problem, which, it must be
admitted, is unsolved whatever theory we may adopt, and perhaps
insoluble in the present state of our knowledge of the physical
chemistry of living matter. We may, however, combine it with
perfect propriety with the side-chain theory, and argue that the
leucocytes may have behaved in the same way as we have
supposed the tissue cells to have done in the last paragraph — that
is, to have retained, and perhaps increased, their receptors, whilst
losing their susceptibility to the toxophore group. Here, again,
we are confronted with the absence of any definite evidence either
way. There seems to be no experimental proof to show whether
leucocytes from a hypervaccinated animal have any increased
resistance to the poisonous action of toxins or any increased
power of combining with them. This is greatly to be deplored,
and might easily be remedied now that Sir Almroth Wright has
taught us a simple and convenient method of working with living
leucocytes. As far as our experimental knowledge goes, there is no
great difference in leucocytes from immunized and normal animals,
and the researches on opsonins have led to a tendency to disregard
the possible variations in the immunity of the leucocytes, since
the researches of Bulloch and others showed that leucocytes from
very varied sources would take up the same number of bacteria

128 IMMUNIZATION OF LEUCOCYTES

under identical conditions. Subsequently, as we have seen above,
some slight indications of a difference have been found. But on
Metchnikoff's theories there ought to be a very marked difference,
and we should expect that, the leucocyte being par excellence the
cell devoted to the protection of the animal body against infections,
it would be the first to acquire increased resistance in acquired
immunity, and we should expect, e.g., leucocytes from a convales-
cent case of pneumonia or from an animal vaccinated against the
pneumococcus to take up far more pneumococci than leucocytes
from a normal person under similar conditions, whereas the
difference, if one exists, is extremely slight. A careful considera-
tion of the conditions of opsonic experiments leads to the con-
clusion that the results obtained are not to be regarded in any
sense as an index of the immunity of the leucocytes. Ledingham
has shown that the number of bacteria taken up by the leucocytes
depends on the extent to which the former have been altered by
the action of serum, and not on the temperature at which the
phagocytosis takes place. Thus, if the bacteria are sensitized by
serum, mixed with leucocytes, and part kept at 18° C. and part at

-£*- c*^^«. ^^esoX<_tt.

37 C., the number of leucocytes is the same in the two cases.
Now at the lower temperature no active movements of the
leucocytes take place, so that we cannot regard the ingestion of
the bacteria as being entirely, as was formerly assumed, an active
process brought about by a seizure or surrounding of the
organism by pseudopodia, as can be seen so readily in the
amoeba. Such a process does occur to some extent, and can be
seen to occur under suitable conditions ; but it is also true that
bacteria can be seen to make their way into a cell which has
been watched continuously under a high power of the microscope,
and in which no movement of any sort has been witnessed. We
are led, therefore, to believe that the phagocytosis which occurs
in opsonin experiments in vitro is a process allied to agglutination
rather than to an actual physical seizing of an organism. It is
one which can take place without an actual conscious — if we may
use the term — movement of the leucocyte towards its prey.
An immunized leucocyte would be immune to bacteria which it
had ingested, to the endotoxins liberated in the process of
bacteriolysis, whether taking place within the leucocyte or outside
it, and to endotoxins ; and it would exert its physiological functions
of movement (pseudopodia- production and chemotaxis) equally
well whether these toxins were present or absent. Or the effect

IMMUNITY TO TOXINS I2Q

of a given toxin might be reversed in the immunized leucocytes,
so that in place of being repelled, as under normal conditions, it
might be attracted. As far as phagocytosis depends on the
chemotactic attraction of leucocytes and the active suggestion of
organisms, we might expect it to be greatly increased if these
cells became immune. But there is no reason to see how a
similar increase need occur if it depends on a process akin to
agglutination, and resulting, perhaps, from a change in the surface
tension between the leucocyte and the organism. And that this is
the way in which the vast majority of the bacteria are taken up
in ordinary opsonin preparations appears probable. Active move-
ments of leucocytes are rarely seen in emulsions prepared for the
opsonic technique and examined in normal saline solution on the
warm stage. In other words, the opsonic index, though useful as
an index of certain activities of the serum, would appear not to
throw much, if any, light on the immunity of the leucocytes.
That must be sought by other means, notably by experiments in
vivo, and these lead us to the belief that the leucocytes may
become immune and may play a part of importance in acquired
immunity.

But the facts which have been previously recorded concerning
the local reaction in animals which are being immunized to tetanus
toxin tend to support MetchnikofFs hypothesis, though the evi-
dence is somewhat indirect. We have already pointed out that
when normal horses are injected with small doses of this toxin
there is at first no local reaction, or but slight, and that as the
animal becomes more immune this local reaction becomes more
manifest — a phenomenon which we have attributed to the
hypersensitiveness of the tissues. Now the local reaction is
inflammatory in nature, and the tumefaction which occurs is due
partly to oedema and partly to an access of leucocytes, so that it
would appear that in the production of immunity these cells
acquired the power of invading a tissue which is permeated with
toxin.. If this is the case, we may assume that they have become
immunized to this toxin, and that they have changed in regard to
their chemotactic properties. This collection of leucocytes in
large numbers at the region of inoculation of toxins, where they
may form a sterile abscess of considerableextent, is a frequent
phenomenon in the production of diphtheria toxin ; and here, again,
it seems not improbable that the leucocytes have become, like the
tissue cells, more immune to the toxin. There are no inherent

9

130 BIOLOGICAL SIGNIFICANCE OF ANTITOXIN

difficulties in the supposition, and it certainly favours the
explanation of the phenomenon.

It will be remarked that the tendency of modern thought has
been rather to minimize the importance of antitoxin in the process
of natural recovery and in the subsequent immunity, and to regard
its appearance rather as an epiphenomenon, though doubtless one
that may under certain circumstances be of advantage to the
patient. Let us consider briefly the probable significance of anti-
toxin-formation in the animal economy. Is it a purposive pheno-
menon, developed and selected by natural selection with a view to
defend the species against natural dangers ? This was the natural
assumption when antitoxin was thought to play a role of the
utmost importance in natural recovery and acquired immunity.
There are great difficulties in the way of accepting such a hypo-
thesis. In the first place, immunity to bacterial infections is more
prevalent in the lower rather than in the higher animal types.
Doubtless very great susceptibility to a widespread infective
agent would operate unfavourably to the prospect of survival of
any animal species, but actual experience seems to show that
immunity to ordinary infections is a far less potent factor in
evolution than those studied by Darwin and his school. It would
appear in the case of tetanus at least that susceptibility is acquired
as the animal rises in the zoological scale as a secondary, though
perhaps necessary, result of the possession of a nervous system of
a certain degree of complexity. Secondly, the immunity due to a
successful struggle against a disease is an acquired factor, and as
such not transmitted, according, at least, to the majority of modern
authorities. It might be objected that the capability of producing
antitoxin might be a spontaneous feature, and one capable of being
transmitted by heredity. This is probably true, but in this case it
could only become a powerful agent in evolution if the disease
were prevalent and an attack usual in the life-history of many of
the individuals. This might occur in the case of tetanus, though
here, as we have seen, the tendency in evolution is for the production
of susceptibility rather than immunity ; but it is quite impossible
to see how the power of forming an antitoxin to eel serum after
the injection of that substance (to take one example of many
which might be quoted) can be of advantage to any animal unless
its habitat happens to be a pathological laboratory. Lastly, if the
acquirement of immunity be of great importance in natural
selection, we should expect those species to survive which de-

IMMUNITY TO TOXINS 131

veloped natural, rather than the power to develop acquired,
immunity, since the former would be always available, whilst the
latter only come into action after the animal has successfully sur-
mounted the hazard of an attack of the disease.

It appears more likely that the power to produce antibodies
under certain circumstances is to be regarded as one of the
essential properties of some forms of living protoplasm, and that
its occasional value in the cure or prevention of disease is a mere
accident. Thus the injection of eel serum leads to the production
of antitoxin in all mammals, as far as is known, and this could
only be of advantage to the animal if (i) eel serum happened to
gain access to the tissues ; (2) the animal recovered ; and (3) eel
serum again reached the tissues within a certain period. This is
extremely unlikely to occur. Nor is natural immunity from
poisons necessarily dependent on antitoxin ; in fact, the common
vegetable and other poisons do not lead to a production of anti-
bodies, though if they did the possession of these substances
would doubtless be of great value to animals in a wild state. If
there is any advantage attaching to the power of forming anti-
bodies, it is probably in regard to the solution of organized bodies,
such as bacteria, or their sensitization previously to phagocytosis.
It is quite conceivable that the advantage accruing from the
possession of the power of forming antibodies has led to the
selection of animals whose protoplasm has the property of develop-
ing an antibody to any foreign proteid, the useful side of this
property being the formation of bacteriolysins and opsonins,
the useless corollary being the production of antitoxins and
precipitins.

The theory of passive antitoxic immunity does not present any
difficulties of importance. It may be pointed out that it comes on
as soon as the antitoxin reaches the blood-stream — i.e., at once if
the injection be intravenous, and after a delay of some duration if
it be into the subcutaneous tissues or peritoneum.

Much attention has been paid to the question of the feasibility
of administering antitoxin by the mouth and rectum. In general
there is no doubt that this is useless, and under ordinary conditions
it is not absorbed as such, being probably digested, and thus
deprived of its peculiar characters. Under certain circumstances
this appears not to be the case ; thus it is held that absorption
from the stomach may take place in young animals. Hence,
perhaps, some of the undoubted benefit of the use of fresh raw

9—2

132 PASSIVE IMMUNITY

milk (in which the antibodies have not been destroyed by heat) in
the feeding of infants. It seems also that absorption of antitoxins
can take place in the intestine, and that the process may occur
when gastric digestion is impaired or is prevented by artificial
means ; thus McClintock and King obtained successful results (in
animals) in 93 per cent, of cases after the administration of a
mixture of diphtheria antitoxin, opium, and of a saturated solution
of salol in chloroform. It is possible that something of the same
sort may occur in disease where the digestive powers are notably
enfeebled, but this does not justify its exhibition in this way in
cases of disease where every hour is of the utmost value. Recent
researches have also shown that in some cases animals may be
immunized per rectum ; possibly success is only attained when the
antitoxin goes sufficiently high up.

As regards the duration of the immunity conferred by a single
dose of antitoxin, our knowledge is not very exact in the case of
human beings. According to von Behring, Goodman, and others, an
animal injected with diphtheria antitoxin retains its immunity for
rather more than three weeks, and the duration is not markedly
affected by the size of the dose given — for example, 20 units of
antitoxin per kilogramme immunized a rabbit twenty-three days,
500 units twenty-six days. According to Goodman, the degree of
immunity falls rapidly for the first two or three days, and then more
slowly. According to him also, the immunity is greater than the
amount of antitoxin in the blood would lead one to suppose, and it
continues when antitoxin is no longer demonstrable in that situation.
The duration of the immunity appears to depend upon the source of
the antitoxin injected, being longer if the serum is homologous —
i.e., derived from an animal of the same species : thus von Behring
showed that horses immunized by horse-serum antitoxin retain
their passive immunity almost as long as if they had acquired
active immunity from toxin injections, whereas the effect in
rabbits, etc., as we have seen, is more transient.

But little need be added to what has been already mentioned
incidentally with regard to the curative power of antitoxin. It
acts, in the main, by neutralizing all toxin present in a free state
in the blood or subsequently formed by the bacteria. In general,
as has been pointed out, it does not repair the damage that the
toxin has already done, nor remove the latter from its combination
with the susceptible cells. It is perhaps hardly safe to assume that
there is no trace of such an action, and it is by no means improb-

IMMUNITY TO TOXINS 133

able that a very large excess of antitoxin present in the blood may
have some slight power of attracting toxin from the tissues when
the union is but recent. This, however, is quite unproved, but it
is certain that cells which have been acted on by the toxin are not
benefited by the subsequent action of the antidote.

It is hardly necessary to point out that antitoxin is devoid of
bactericidal action, and that the removal of the infective agent is
brought about by other mechanisms — in most cases, perhaps, by
phagocytosis. Thus we shall see, in the case of diphtheria, that
there is a rapid rise in the opsonic index aj^out the time of the
improvement in the local lesions ; and we can hardly doubt, though
direct evidence is lacking, that the preservation of the leucocytes
from the deleterious action of the toxin when the latter is neutral-
ized by means of antitoxin also plays its part in facilitating
phagocytosis.

We now turn to the subject of natural immunity, an exceedingly
difficult one, and one that is very far from being understood. What
follows is for the most part extremely hypothetical.

It must be borne in mind that, though we speak of animals as
susceptible or as immune to a given toxin, no hard-and-fast line
can be drawn between the two conditions, and in some cases an un-
broken series can be constructed between the most susceptible and
the most resistant species. This is well seen in the case of tetanus.
According to Knorr, the horse is the most susceptible animal
to the toxin of this disease ; the guinea-pig requires twice as much
toxin per kilogramme of body- weight to constitute a lethal dose,
the goat 4 times as much, the mouse 13 times, the rabbit 2,000
times, and the hen 200,000 times. Von Behring confirms these
figures in general, but finds the mouse rather more susceptible
than the horse. With the exception of the fowl these animals are
all " susceptible." Of the " insusceptible " animals the hen is sus-
ceptible to large doses, but is also affected by much smaller ones
if the injection be intra-cerebral. Other animals are still more
resistant, such as the tortoise, lizard, cayman, larva of Oryctes,
etc. The effect of other toxins on different species of animals
has not yet been fully investigated.

It is necessary to realize that, before we can say that an animal
is insusceptible or immune to the action of a given toxin, it is
necessary for the animal to be kept, after injection, at a tempera-
ture at which this toxin can act. This subject has already been
mentioned in connection with the functions of the haptophore and

134 NATURAL IMMUNITY TO TOXINS

toxophore groups. We have already seen reason to believe that
the former can functionate (i.e., the molecule of toxin can unite
with the cell) at a low temperature, whereas the enzyme-like
action of the toxophore radical can only act at the temperature of
the human body, or one approximating thereto — this is in the case
of tetanus. It is apparent, therefore, at the outset that there are
two possible explanations for immunity to toxins : firstly, the
toxin may find no cell with which it can unite — or, to use Ehrlich's
terminology, no cell receptor having a combining affinity for the
haptophore molecule of the toxin molecule ; and, secondly, the
union takes place, but the cell is not injured as a consequence —
that is, the toxophore group is without action on the cell proto-
plasm.

1. Immunity due to an absence of suitable receptors.

This has been proved to exist in several cases. Most of these
are due to the researches of Metchnikoff, who, however, does not
express the fact in these words, but contents himself by saying
that the toxin does not combine with the tissues. Thus, when the
larva of Oryctes is injected with tetanus toxin, no symptoms
develop, and if the blood of the animal be collected several months
afterwards, it will be found to contain free toxin, as shown by the
fact that it will tetanize susceptible animals. These experiments
were carried out at a temperature of 30° to 36° C. Similar facts
were observed in lizards ; -^ c.c. of blood taken two months after
the injection of tetanus toxin produced fatal tetanus in a mouse ;
the blood of a turtle (Emys orUcularis) contained toxin no less
than four months after injection. In these cases the explanation of
the immunity is clear, however we choose to express it : the toxin
has no chemical affinity for the living protoplasm of any part of
the animal, and is, therefore, powerless to injure it.

2. Immunity due to the insusceptibility of the cells to the
action of the toxophore group.

Perhaps the best example of this is in the case, so often referred
to, of the action of tetanus toxin on the cold frog, but other
striking examples have been discovered by Metchnikoff. Thus,
when the alligator is kept at ordinary temperatures (about 20° C.)
and injected with tetanus toxin, that substance rapidly disappears
from the blood, yet without the production of any tetanic symptoms,
and without the appearance of antitoxin in the blood. If, how-
ever, the animal is kept at a temperature of 32° to 37° C., it pro-
duces antitoxin with great rapidity, though still without the

IMMUNITY To TOXINS 135

production of tetanus. Taking these two experiments together,
we may regard them as constituting a definite proof of the existence
of this form of immunity. The theory that the toxin disappears
because it is anchored to the cells appears to be clearly demon-
strated from the fact that antitoxin is produced, an occurrence
which we could scarcely explain on any other hypothesis.

But it is hardly necessary to look for experimental proof of the
occurrence of this form of immunity, since the fact that the blood
of many species of animals is toxic for other species appears to
constitute a ready-made demonstration of the fact. The most
striking example is given by eel serum, a substance which is toxic
for nearly all animals, except, of course, the eel itself. The serum
of the horse is perhaps the least toxic of all sera, but even this
has some poisonous properties. Whatever theory we may hold
with regard to the mechanism by which the cells of an animal
are nourished, it is hardly possible to avoid the conclusion that the
proteid substances of the plasma must combine with the living
protoplasm. And it is the proteids which make eel serum toxic
to other animals. It appears, therefore, that the immunity of the
eel to its own serum depends upon the fact that its protoplasm is
not injured by the toxophore group, although the molecules of the
toxin and protoplasm become united. (The toxicity of eel serum
is dependent on a more complex structure than that of ordinary
bacterial toxins, being, in fact, a cytolysin or cytotoxin, but this
does not affect the argument.)

A consideration of the use of these poisonous sera to the
animals which produce them may perhaps enable us to analyze
the process a step further, and to attribute their immunity (e.g.,
of eels to eel serum) to the power which their cells possess of
using these " toxic " substances as sources of nourishment ; and
if this be so, we might perhaps apply this suggestion to explain
the form of immunity to toxins under discussion. Thus, in the
case of the alligator injected with tetanus toxin and kept in the
cold, it is possible that the poison which combines with the cells
is " digested " by them and used as nourishment. If this is so, we
must assume that the molecule of toxin has, as far as its hapto-
phore group is concerned, a close chemical affinity with the
proteids normally present in the animal's blood and used by it in
cell nutrition ; and, further, that the toxophore group is not only
innocuous to the cell, but is also no bar to the assimilation of the
whole molecule by it. We have seen that in the case of the frog

136 NATURAL IMMUNITY TO TOXINS

the toxophore group only acts at a raised temperature. In the
alligator we must assume that the elevation of the temperature
has a similar but less marked effect ; it brings the toxophore group
into a state of activity in which it is inassimilable by the molecule
of cell protoplasm to which it is anchored, although the latter is
not injured in any way. The result is, of course, the production
of antitoxin, which approximates to the formation of the pre-
cipitins which occurs when non-poisonous foreign proteids are
injected.

Thus it would appear that at the last analysis this form of
immunity is in reality an expression of cell nutrition. A cell
which can nourish itself on a given toxin is naturally immune to
its action. It is hardly necessary, however, to say that there is
no direct proof that any cell can extract nourishment from a
bacterial toxin.

There is a third, and in some respects more interesting and
important, form of immunity to toxins which is readily conceivable
on theoretical grounds, and the occurrence of which is capable of
experimental proof. We may define it as immunity due to the
fact that some of the cells are insusceptible to the action of the
toxin, but have a great combining affinity therewith, and thus
shield from its action the susceptible cells, in which the combining
affinity is less. We have already referred to the most striking
and best known example, that of the immunity of the fowl to
tetanus toxin injected subcutaneously as compared with its sus-
ceptibility when the injection is made direct into the brain. Some
authorities have attempted to explain this on the assumption that
there^is a barrier en route which prevents the access of the toxin
to the central nervous system. But if by this we are to imagine a
physical barrier in the shape of a layer of endothelium or other
tissue between the blood and the cells of the brain, the explanation
is inadequate, since it would leave unexplained the nature of the
immunity of this living barrier, nor would it explain the rapid
disappearance of the toxin from the blood. The true explanation
is certainly that the toxin combines rapidly with the cells of the
body, or perhaps, as we shall see later, the leucocytes, and is
thereby prevented from gaining access to the central nervous
system. These relatively unimportant cells we must imagine to
have the same nutritive relations to the molecules of toxin as have
the cells of the heated alligator ; the two can unite, but the proto-
plasm is neither injured nor nourished. It seems probable that this

IMMUNITY TO TOXINS 137

affords a sufficient explanation of the differences in susceptibility
of the " susceptible " animals which figure in Knorr's list, and that
the cells of the central nervous system in the warm-blooded verte-
brates differ but little in their relation to tetanus toxin. Thus, in
the case of the horse and the guinea-pig little toxin is bound to the
body cells, and practically the whole amount makes its way to the
brain wherever the injection is made. In the rabbit the body cells
can absorb more, so that if only a small dose is given none may
reach the more important and susceptible organs. The fowl marks
the other end of the scale ; the tissue cells must have an enormous
avidity for toxin, and, unless absolutely gigantic doses are given,
absorb the whole of it.

The objection may be raised that, the receptors of the body cells
being of the same nature as antitoxin, the resulting combination
of body-cell receptor and toxin should undergo dissociation, the
poison being gradually passed on to the central nervous system,
in the cells of which dissociation does not occur, since poisoning
takes place ; this refers to the case, e.g., of the fowl. It is true
that such a process appears to occur in vitro. If tetanus toxin be
mixed with emulsions of many of the tissues, the two enter into a
loose combination, so that if the fragments of toxin-charged tissues,
after thorough washing in normal saline solution, are allowed to
soak in that fluid, toxin is gradually liberated. It is only in the
case of the central nervous system that stable non-dissociable
compounds are formed. Several interpretations of the apparent
anomaly might be suggested. It is, for instance, very probable
that here, as in so many other phenomena in immunity, the
real defensive cell is the leucocyte. Thus, Metchnikoff found that
if he injected tetanus toxin into the fowl, and then, after an appro-
priate interval, excited an aseptic exudate composed largely of
leucocytes, the fluid thus obtained would excite tetanus in a sus-
ceptible animal. Perhaps the chain of phenomena is as follows :
The toxin first unites with the connective and other tissue cells,
and so the amount that the leucocytes have to deal with at first is
greatly diminished ; dissociation takes place, and there is a steady
stream of toxin from the tissues to the blood. This is dealt with
by the leucocytes, which thus have time to increase in numbers
(and even in the insusceptible fowl the result of the injection is to
cause a marked leucocytosis), and perhaps to become immune. It
is useless to deny that these phenomena are difficult to harmonize
with the side-chain theory as first expounded, or that Metchnikoff

138 NATURAL IMMUNITY TO TOXINS

has brought forward very strong (though not conclusive) evidence
in support of his theory of the leucocytic origin of antitoxin ; and,
as we shall see, as far as our knowledge goes, we are led to believe
that all antibodies are derived from the spleen, bone-marrow, and
other organs rich in leucocytes.

The study of the affinity of bacterial toxins and of other poisons
to various tissues and organs of the body is destined to play a part
of great importance in our ideas of pathology and pharmacology.
It has been especially investigated by Ehrlich, and in connection
with immunity and susceptibility to toxins he recognizes four
conditions which may occur :

1. In which no specific receptors occur in any part of the
animal. This is the first form of natural immunity which we
have discussed ; the animal is absolutely immune, and cannot
produce antitoxin.

2. Receptors are present, but only in tissues on which the
poison does not act, or in tissues of little importance. Here the
animal is immune, but antibodies may be produced. This is the
case of the fowl vis-a-vis the tetanus toxin.

3. The receptors are distributed over various parts of the
organism, and are present in the organs which are sensitive to the
action of the poison. Here there is a relative immunity, and the
degree of intoxication depends largely on the region into which
the toxin is introduced. These animals may be immunized,
though the process is one of some difficulty, owing to the sensitive-
ness to the toxin of vital organs, and, of course, antibodies may
be produced.

4. The receptors are present only in vital organs which are
sensitive to the poison. Here the animal is very susceptible, and
immunization very difficult, involving the use of extremely minute
doses or of toxoids. The action of tetanus toxin on the guinea-
pig and horse provides examples.

CHAPTER VII
BACTERIOLYSIS AND ALLIED PHENOMENA

UP to the present we have dealt entirely with the mechanism by
which the injurious effects of the infective bacteria are nullified.
It is obvious that this is only one, though perhaps the most im-
portant, aspect of the question. We have now to study how the
bacteria themselves are removed.

The bactericidal properties of the blood attracted attention very
early in the history of bacteriology, and long before the beginnings
of the science Hunter showed that blood had the power of resisting
decomposition longer than other animal fluids. It was, however,
the controversy which took place between Metchnikoff and the
humoralist school which first focussed attention on this question,
and led to the discovery of the alexins by Buchner, Nuttall, and
others. The controversy is best discussed subsequently, and it is
sufficient now to point out that it was found that the circulating
blood had the power of killing certain bacteria, and that this
property was even more marked in the serum. Thus, according
to Lubarsch, 16,000 virulent bacilli will kill a rabbit if injected
intravenously — i.e., the blood has not the power of killing this
number ; yet i c.c. of fresh serum will destroy this number or
more. It is obvious that the bactericidal substance or substances
occur in the blood, and to a greater extent in the serum.

The properties of these bactericidal substances or alexins were
investigated, and it was found that they were fragile bodies, readily
destroyed by a moderate temperature (55° C.), and that they
disappeared spontaneously if the serum were kept for a few days.
They were destroyed by acids and alkalis, and what was most
important of all was that they were selective in their action — i.e.,
those from a certain animal might be potent antiseptics as regards
certain bacteria and inert towards others, whereas the serum of
another species might have quite different actions. They appeared
to be formed by the breaking down of leucocytes ; hence their

139

140

appearance in the blood after clotting (when fibrin ferment is also
liberated from these cells), and their absence from fluids containing
none, such as the aqueous humour.

It was obvious that this was a foundation for a theory of
immunity, but it soon became apparent that it did not explain the
whole of the phenomena. An animal might be immune to a
certain organism, yet its serum might not be bactericidal for it, or
it might be susceptible and yet possess suitable alexins. Thus,
rabbits are very susceptible to anthrax, yet have serum which
kills the bacilli in large numbers, whilst the dog is much less
susceptible, though its serum is very slightly bactericidal. Such
facts prevented the alexic theory of immunity from making head-
way until the discovery of Pfeiffer's phenomenon.

Pfeiffer found that when cholera vibrios were injected into the
peritoneal cavity of a highly immunized guinea-pig they were not
only killed, but dissolved, and this with great rapidity. The
vibrios are first rendered immotile, and then lose their proper
shape, becoming converted into spherical masses which stain
badly. In a little while they become smaller, and disappear alto-
gether, no trace being left.

Pfeiffer showed that the phenomenon could also be produced
by injecting into the peritoneum of a normal guinea-pig a mixture
of serum from an immunized animal and the culture of vibrios.
This happened when an old specimen of serum in which the
alexins had disappeared spontaneously was used, or a fresh
specimen that had been heated to 60° C. Lastly, he found that
if an old immune serum were injected into the peritoneal cavity
and allowed to remain for a while, it regained its bactericidal
powers, and could then dissolve the vibrios in vitro at the body
temperature. The phenomenon was specific — i.e., if serum from
an animal immunized against cholera were used as described
above, the vibrios of cholera were dissolved, but no others. If
typhoid serum were used, typhoid bacilli showed degenerative
changes, though they were not completely destroyed, but there
was no effect on cholera vibrios.

The explanation which suggested itself to Pfeiffer was this :
There is in the serum of highly immunized animals a substance
which can exist either in an active or inert state. In the blood-
serum or peritoneal fluid whilst in the body it occurs as an active
substance, but it passes into the inert condition when kept for a
few days, or very rapidly when heated to 60° C., and this inactive

BACTERIOLYSIS AND ALLIED PHENOMENA 141

substance can be rendered active again by contact with the living
endothelial cells of the peritoneum.

The true explanation was given by Bordet, who showed that
two substances are necessary for the phenomenon. The one is the
inert thermostable substance which occurs in the serum of highly
immunized animals, but not in normal animals, and which does
not disappear on keeping. The second is a thermolabile sub-
stance which occurs in fresh serum, whether from a normal or
from an immunized animal. This he showed by demonstrating
that, whereas neither heated nor stale immune serum alone, nor
fresh serum alone, had the power of leading to the production of
Pfeiffer's reaction in vitro, this was caused when the two were
used in combination. When a drop of fresh serum from a normal
animal, some heated immune serum, and the cholera vibrios are
mixed together and incubated, the whole train of phenomena was
repeated.

The thermolabile substance taking part in the process was
soon identified as the alexin of the earlier investigators, and the
new substance was called by Bordet substance sensibilatrice. His
theory was that alexin alone does not unite with the bacteria,
unless these have been previously sensitized by the action of the
substance in the immune serum. It was shown that the substance
sensibilatrice has the power of uniting with the bacteria as follows :
A culture exposed to the action of heated immune serum and
washed by repeated centrifugalizations with normal saline solution
is apparently unaltered, but the bacteria are dissolved by the
action of normal serum, which has no action on unsensitized
bacteria. Bordet supposed, therefore, that the immune serum
alters the constitution of the bacteria in some way, so that the
alexin can combine with them subsequently. He compared the
process to the opening of a lock, which can only be effected by
means of two keys, of which one (the sensibilatrice) must be
turned before the other (alexin) can be introduced.

Bordet found that the essential facts could be reproduced
exactly if red blood-corpuscles were substituted for bacteria. It
was previously known that the serum of some animals possesses
the power of liberating the haemoglobin (haemolysis) from the red
corpuscles of other species, just as the serum of some animals
can destroy certain bacteria. Bordet showed that the guinea-pig
serum has normally no action on the red corpuscles of the rabbit,
but that it becomes haemolytic to the latter after a few injections

142 HAEMOLYSIS — BORDET'S DISCOVERIES

of rabbit's blood. Just as an animal which was formerly destitute
of that power becomes able to dissolve the cholera vibrio after
a few injections of that organism, so a guinea-pig, the serum of
which had formerly no action on the rabbit's corpuscles, acquires
the power of dissolving them as the result of an injection or two
of rabbit's blood ; hence, following the analogy with the cholera
vibrio, the guinea-pig is said to be " immunized " to the rabbit's
corpuscles, though there is here no question of the avoidance of
any deleterious action. Further — and hence the importance of
these discoveries in the theory of immunity — Bordet showed that
the laws which govern bacteriolysis by means of immune sera
were apparently identical with those governing Pfeiffer's pheno-
mena. The fresh immune serum is haemolytic ; it loses its power on
heating to 55° C., and it regains it on the addition of fresh normal
serum. Further, the action is, to some extent at least, a specific
one. An animal (A) injected with corpuscles of another species
(B) will dissolve the corpuscles of that species, and may also have
some action on those of animals closely allied zoologically. It is
evident that bacteriolysis and haemolysis are closely akin, and as
researches on the latter phenomena are infinitely more easy than
on the former, much work intended to elucidate the mechanism
of immunity has been carried out on red corpuscles. This is to
some extent regrettable, since the processes, though similar, are
not identical, and it is unsafe to argue from one to the other
without experimental verification.

It was this analogy between bacteriolysis and haemolysis that
led Ehrlich to an investigation of the latter phenomenon, and
his researches led to a flood of new light being thrown upon the
subject. Ehrlich introduced fresh names for the substances which
Bordet had shown to be necessary for the phenomenon, and it
will now be convenient to give a list of the various terms which
the theories or caprices of various writers have applied to
each.

The thermostable substance has been called :

Substance sensibilatrice, or simply sensibilatrice.
Immune body. Philocytase.

Amboceptor. Immunisin.

Fixator. Desmon.

Intermediary body. Copula.

Interbody. Preparator.

BACTERIOLYSIS AND ALLIED PHENOMENA 143

Whilst the thermolabile substance is spoken of as :

Alexin. Complement.

Addiment. Cytase.

We shall speak of the substances as amboceptor or immune
body and as alexin or complement respectively. Objection may be

FIG. 24. — THE CONSTITUENTS OF FRESH IMMUNE SERUM, ON EHRLICH'S

THEORY.

a = complement, b = amboceptor, d being its complementophile, and c its cyto-
phile groups. Below, a red blood-corpuscle, showing its receptors.

This and the following figures are modified from Ehrlich, and are intended to
illustrate his theories of haemolysis. The black figures represent normal
substances (corpuscles, complement), the white ones antibodies. It need
hardly be said that they are absolutely diagrammatic.

FIG. 25. —THE IMMUNE SERUM SHOWN IN FIG. 24, AFTER HEATING.
The complement is destroyed, and amboceptor only remains.

FIG. 26. — NORMAL SERUM CONTAINING COMPLEMENT, BUT NO AMBOCEPTOR.

taken to Ehrlich's terms "amboceptor" and "complement," since,
as we shall show, they imply a theory, but they are too firmly
rooted to be displaced.

Ehrlich applied his side-chain theory to the study of the
phenomena, and argued that amboceptor must be an antibody
to the receptors of the red blood-corpuscle. If this is the case,
it ought to unite with the corpuscles, and Ehrlich showed in the

144

HAEMOLYSIS — EHRLICH'S RESEARCHES

following way that such actually occurred : He worked with the
serum of a goat which had been injected with and was haemolytic for
the red blood-corpuscles of a sheep. He heated this immune serum
to 56° C. to destroy complement, and added some sheep's cor-
puscles ; the mixture was then centrifugalized, the supernatant fluid
then pipetted off, and replaced by normal saline solution. The red
corpuscles were to all appearance unaltered, but it was now found

FIG. 27. — HEATED IMMUNE SERUM ADDED TO RED BLOOD - CORPUSCLES,

WHICH ARE APPARENTLY UNALTERED, BUT ARE IN REALITY " SENSITIZED."

FIG. 28. — EFFECT OF ADDITION OF FRESH NORMAL SERUM TO SENSITIZED

CORPUSCLES.

Complement is now linked up to the corpuscles, and haemolysis takes place
when they are incubated at 37° C.

that if a small amount of normal goat serum were added and the
mixture incubated, haemolysis occurred. It was evident, therefore,
that the corpuscles underwent some change in virtue of their stay
in the heated immune serum, though no alteration was obvious.

In a further experiment he showed that the change consisted in
the abstraction of the amboceptor from the fluid. This appeared
from the fact that if the supernatant fluid from the last experi-
ment were pipetted off, fresh normal goat's serum added (to supply
complement), and the mixture tested with sheep's corpuscles, no

BACTERIOLYSIS AND ALLIED PHENOMENA 145

haemolysis occurred. Evidently it had been removed by the first
addition of corpuscles.

Red corpuscles, therefore, have a combining affinity for ambo-
ceptor, and in further experiments Ehrlich showed that this
combination takes place at ordinary temperatures or at low ones,
down to o° C. Expressed in the language of the side-chain theory,
amboceptor has a haptophore group with a combining affinity
for the receptors of the corpuscle, bacterium, etc., with which it

'*'**

* t ~

FIG. 29. — NORMAL GOAT SERUM (COMPLEMENT), PLUS SHEEP'S CORPUSCLES.

No combination. This is shown as follows : The mixture is centrifugalized,
and to the corpuscles heated immune serum (amboceptor) was added
(Fig. 30), whilst to the supernatant fluid corpuscles and heated serum
were added (Fig. 31).

FIG. 30. — THE CORPUSCLES FROM THE PREVIOUS EXPERIMENT INCUBATED

WITH HEATED IMMUNE SERUM.

No solution, showing that no complement had been abstracted in combination

with them.

unites. We shall see reasons for believing that it may have a
second haptophore group, and shall distinguish this first by the
name of cytophile haptophore.

Ehrlich now investigated the behaviour of the second substance
— the complement — with the red corpuscles by a precisely similar
method, and found that the two had no power of entering into
combination. Thus, normal goat's blood (containing complement)
was added to sheep's corpuscles, and the mixture centrifugalized.
To the corpuscles heated immune serum (amboceptor) was added,
but there was no haemolysis. Again, to the supernatant fluid
sheep's corpuscles and heated serum were added, and haemolysis

10

146

HAEMOLYSIS — EHRLICH'S RESEARCHES

i

occurred ; complement had evidently not been withdrawn from the
fluid. Complement, therefore, will not unite with red corpuscles
direct. It has no haptophore group with an affinity for the
receptors of the latter.

This led Ehrlich to the theory that the complement united with
the red corpuscle only indirectly by means of the amboceptor.

FIG. 31. — THE SUPERNATANT FLUID FROM FIG. 29, TESTED WITH HEATED
IMMUNE SERUM AND SHEEP'S CORPUSCLES.

Solution takes place, showing that the complement had not been removed.

FIG. 32. — FRESH IMMUNE SERUM (OR A MIXTURE OF HEATED IMMUNE
SERUM AND FRESH NORMAL SERUM) ADDED TO CORPUSCLES AT o° C.

Amboceptor unites therewith, complement does not.

He pictured the latter as having two haptophore groups : a cyto-
phile, which we have already mentioned, and a complementophile,
with which the complement could unite after the former had
seized on a receptor of the red corpuscle. The process of haemo-
lysis was supposed to take place as follows : In fresh immune
serum or in a mixture of heated immune serum and fresh normal
serum (i.e., of amboceptor and complement) the two substances
occur independently of one another. When the mixture is kept
in the cold, and red corpuscles are added, the cytophile groups of
the amboceptor molecules attach themselves to the receptors of

BACTERIOLYSIS AND ALLIED PHENOMENA 147

the red corpuscles, but the molecules of the complement still
remain free, and do not attach themselves either directly or in-
directly to the corpuscles. Nevertheless, there is a change in the
combining affinities of the complementophile haptophore.1 This
is shown by the fact that if the mixture be raised to the body-
heat, the molecules of complement attach themselves to these
haptophores, exert a digestive or haemolytic action through the
amboceptor on the red corpuscles, and haemolysis occurs. Hence,
on Ehrlich's theory, the complement does not attack the cell or
corpuscle direct (as Bordet holds, and as is implied in the term
substance sensibilatrice for amboceptor), but it unites itself to
one part of the molecule of amboceptor after another part of the
latter has combined with the corpuscle. We will defer the
discussion of these theories for the present.

For the sake of simplicity, we have assumed that complement
will not unite with amboceptor until the latter has combined with
its antigen. Ehrlich, however, assumes (on no very clear evidence)
that the two substances may enter into a loose and easily dis-
sociated chemical combination. This is hastened by heat and
retarded by cold. The union between corpuscle and amboceptor
is a firm one, which takes place at low temperatures, and which
does not tend to dissociate, and after the combination amboceptor
has a stronger affinity for complement, though even then a firm
union only takes place at 30° C. or over.

Ehrlich explains the method of formation of amboceptor on the
side-chain theory in this wise : He points out that for the
absorption of substances of small molecule by the cell protoplasm
the simple receptors, which, when cast off, constitute antitoxin,
may suffice ; but he says that when a giant proteid molecule (e.g.,
of albumin) is anchored, it requires the action of a digestive
ferment before it can be brought into a condition to be of use
in the nourishment of the living cell. Hence it must be first
broken down by means of a proteolytic enzyme, and it is of this
nature that he imagines the complement to be. He holds, there-
fore, that the receptors or side-chains which are adapted to seize
molecules of coagulable proteid possess two haptophore groups,
the one to seize the nutrient material, and the other to seize a
molecule of digestive enzyme from the surrounding blood or
lymph. When the formation of such receptors is stimulated to

1 Ehrlich does not make a definite statement as to this increase of affinity,
but it seems a necessary deduction from the facts as he interprets them.

10 — 2

148 HAEMOLYSIS — EHRLICH'S RESEARCHES

excess, and they break loose, they retain both their haptophore
groups and constitute amboceptor. The stimulation and over-
production is, of course, produced by the molecules of the substance
injected (the specific antigens), be they red blood-corpuscles,
bacteria, or, as we shall see, a whole host of other cells.

As a rough illustration of the nature of these seizing arms
Ehrlich compares them with the tentacles of Drosera, which
seize the nutrient material and then secrete a digestive fluid.
The analogy is not exact, since the receptors do not form the
digestive enzyme, but simply select it from without.

Ehrlich compares the action of the haemolysins with that of
the toxins, and points out that in each case there is a haptophore
group adapted to seize on the red corpuscle to be dissolved or the
cell to be poisoned, and an actively functional group with functions
resembling those of an enzyme. He says that the haemolysins
are practically toxins in two parts, and that a combination of these
parts is necessary before any action can be effected.

These discoveries and theories of Ehrlich led to a series of
further researches, all of which were directly suggested by them,
and the results of which appear to corroborate them to the full.
Whatever the ultimate fate of the side-chain theory, there can be
no doubt that it has lead to an enormous increase of our knowledge
on a most intricate subject. These researches have in many
cases but little direct bearing at present on the subject of
immunity, but they are far too important to be passed over
without mention. We shall, therefore, summarize them as briefly
as is possible when dealing with so complicated a subject. We
will deal first with the explanation of the haemolytic action which
the serum of some normal animals has on the corpuscles of other
species ; for instance, normal goat serum will dissolve the red
corpuscles of the guinea-pig. This was shown conclusively to
depend upon the same mechanism of amboceptor and complement
as in the artificial immune sera. The proof was as follows :
Goat serum was mixed with guinea-pig's corpuscles, and the
mixture kept at o° C., and then centrifugal ized. The theory would
lead us to believe that the amboceptor had combined with the
corpuscles, but that the complement had remained free. If this
were so, the supernatant fluid would be found to have lost
amboceptor. To test this a further addition of corpuscles was
made, and the solvent power of the serum for these was found
to be lowered. Evidently, then, sera which are naturally haemo-

BACTERIOLYSIS AND ALLIED PHENOMENA 149

lytic to certain red corpuscles are so in virtue of containing
amboceptors with cytophile haptophore groups which fit the
receptors of these cells and complements which can dissolve
them. He was also able to show, by an ingenious experiment,
that when a normal serum can haemolyze the corpuscles from
different species of animals, it does so by the action of different
amboceptors, and in some cases of different complements also.
(,We must imagine, therefore, that normal serum contains numerous
antibodies of the amboceptor type, and adapted to dissolve
numerous foreign substances when they gain access to the blood ;
and it seems reasonable to believe that these normal antibodies
and those which are formed as a result of the presence of the
foreign substances play a part of the greatest importance in
immunity. \

The question then arose, Do animals possess or form haemolysins
adapted to the solution of their own corpuscles — in other words,
an autohtsmolysin ? For instance, if a person sustains a large
internal haemorrhage, and the blood is absorbed, is he thereby
stimulated to the production of a substance which dissolves his
own red corpuscles ? Ehrlich pointed out that such a phenomenon
is unknown, and set out to investigate the reason of its non-
occurrence. He selected goats as his experimental animals, and
injected into them large amounts of mixed blood-corpuscles from
other goats. In a week's time the serum was found to be power-
fully haemolytic, but only towards the corpuscles of other goats,
never towards its own; thus the serum of the first animal
experimented on was found to dissolve the corpuscles of goats i,
2, 4, 5, 6, and 9 easily ; those of 3 and 8 less powerfully ; and
those of 7 and of itself not at all. A second animal was treated
just like the first, and it also developed a haemolysin, but this
did not act on its own corpuscles, nor on those of other goats
in the same way as that of the first ; evidently a different series
of amboceptors had been formed. Ehrlich speaks of a haemolysin
formed by the injection of the corpuscles of a different species
as heterolysin, that which acts on the blood of the same species
as an isolysin ; one which would act on the blood-corpuscles of
the same animal would be called an autolysin, but this is never
formed.

In attempting to discover the reason for this non-formation
Ehrlich first proved that the amboceptor of the isolysin was
anchored by the corpuscles which it dissolved, but not by those

150 HAEMOLYSIS — ISOLYSIN

of the goat from which it was derived. An obvious explanation
would be that .the latter did not possess any receptors which it
would fit. Another suggestion might be that it had such receptors,
but that they were already fully occupied by amboceptors occur-
ring in the serum ; but in this case they would attract complement,
and solution would ensue. By a method1 which will not be
described here, since it would involve anticipation of facts not
yet described, Ehrlich was able to prove the former view the
correct one. The immunity, therefore, of the corpuscles of
an animal to its own isolysin is due to a complete absence
of suitable receptors in its corpuscles, and, we may add, in the
entire organism.

He deduced, therefore, that each blood-corpuscle possesses
numerous side-chains with haptophore groups, each of which,
when injected into a living animal, is able to combine with a
suitable receptor. If we take a particular variety of haptophore,
which we will call a, we can see that there are two possibilities
after the injection ; it may find no receptors a, in which case
there will be no antibody formed, or it may find such receptors.
In the latter case there are also two possibilities : there may be
receptors a only, or there may be receptors a and haptophore side-
chains a.

If there are only receptors a and no side-chains a, the injection
of corpuscles with side-chains a, will lead to an overproduction of
receptors a, and amboceptor will be produced. This will act as a
hsemolysin to the corpuscles injected, but not to those of the
animal itself, since they do not contain side-chains to which it can
attach itself ; it will be an isolysin, not an autolysin.

In the second possible case Ehrlich points out that the condi-
tions for the production of an autolysin do occur, and such a
substance might be produced and might do serious harm to the
animal. But it would be produced at first only in small amounts,
and the result might be that it would combine with receptor a,
which might then be stimulated and cast off and would form anti-
autolysin.

1 He first showed that the injection of an amboceptor from one animal into
another of a different species would cause the production of an anti-ambo-
ceptor, and then showed that the injection of a serum containing isolysin
into a goat whose corpuscles it dissolved produced anti-isolysin. These
corpuscles contained receptors, since they fixed the isolysin. In the goat
from which the isolysin came there was, of course, no production of anti-
isolysin. Hence, he argued, it had no suitable receptors in its entire body.

BACTERIOLYSIS AND ALLIED PHENOMENA 151

There are, therefore, three possibilities : no formation of haerno-
lysin, the formation of an isolysin, and the formation of an anti-
autolysin. In other words, if autolysins were ever developed they
would be followed immediately by the production of their specific
antibody.

But Ehrlich assumes — and our previous account of the iso-
agglutinins leads us to be ready to accept his assumption
unhesitatingly — that each corpuscle in every animal has numerous
side-chains, «, (3, 7, etc. If such a corpuscle is injected into
an animal of the same species, a may lead to no result, finding no
receptors ; ft may lead to the production of an isolysin, and 7 to
that of an anti-autolysin. The existence of the latter substance
could not, of course, be proved, since it was never possible to
prepare autolysin.

Hence what Ehrlich calls the " pluralistic conception of the
cellular immunity reaction." There are, he says, in each bac-
terial cell numerous side-chains, to each of which an antibody
is theoretically possible, and an ideal curative serum would
contain all these antibodies. But in some cases, perhaps in most,
only a few are found, and others cannot be developed ; for
instance, in some animals it is impossible to produce anti -enzymes
by the injection of enzymes. He explains this in two ways.
The specific receptors may be of peculiar constitution, so that they
cannot be cast off from the cell in the process of immunization :
these he calls sessile receptors. Secondly, it is conceivable that
the side-chains in question are normally produced by the animal
cells, and that no antibody is produced to them, or, at any rate, that
none accumulates.

But when either or both of the conditions arise, we may be able
to get the antibodies from a second animal which we are unable
to do from the first. Thus, if we imagine that the typhoid bacillus
has twenty different sorts of side-chains, we may be able to get
antibodies to some of them from the dog, to others from the rabbit,
etc. Hence he suggests as an important principle in the forma-
tion of curative sera to use many animal species, mixing the sera
of those which produce the antibodies required for use. (This, it
must be pointed out, is not what is meant by a polyvalent serum,
which is a serum formed by the injection of many strains of the
same organism.)

Leaving this subject for the present, we will pass on to Ehrlich's
researches into the nature of complements. Buchner, in his

152 HAEMOLYSIS — COMPLEMENT

studies on the alexins, assumed that the serum of each animal
contained one alexin, though the alexins of different animal
species were different. This view was also held by Bordet and
others, and constitutes the " Unitarian " view of alexin or com-
plement. Metchnikoff holds that there are two alexins, or, as he
calls them, cytases — macrocytase, which acts specially on red
blood-corpuscles, cells, etc., and i /jSaeJ by the large mononuclear
cells ; and microcytase, which acts on bacteria, and is formed by the
polynuclear leucocytes.

As against these views, Ehrlich advances the theory of the
multiplicity of the complements, the proof of which appears entirely
satisfactory. According to him there are numerous complements
in the serum of every animal, and these differ from one another in
their haptophore groups, so that one can reactivate one ambo-

* 3' #

FIG. 33.— SHOWING SOME OF THE CONSTITUENTS WHICH CAN BE DEMON-
STRATED IN THE SERUM OF A GOAT WHICH HAS BEEN INJECTED WITH
SHEEP'S CORPUSCLES.

a, b, and c are the different complements, a' , b' , and c' the amboceptors which
unite them to the corpuscles of the sheep (A), rabbit (B), and guinea pig (C).

ceptor, another another. They differ in some cases in other
respects. The first example which he was able to adduce in
which there were at least two complements in a sample of serum
was that of a goat which had been injected with sheep's corpuscles,
and which dissolved those of the sheep, guinea-pig, and rabbit.
But when this serum was heated to 56° C. for three-quarters of an
hour, it was found to have no action on the corpuscles of the latter
animals, whilst that of the former was unchanged. Evidently
then, the complement which activated the amboceptor combining
with the sheep's corpuscles was thermostable ; that acting on the
others, thermolabile. This was the first known example of a
thermostable complement. In a later experiment he was able to
prove that the complements taking part in the haemolysis of
rabbit's corpuscles and that taking part in the solution of guinea-
pigs were different : the latter passes through a Pukall's filter,
the former does not.

BACTERIOLYSIS AND ALLIED PHENOMENA 153

These methods of heating, filtration, etc., of course only separate
the complements in certain lucky cases, and the more general
method involves the removal of some of these bodies by means of
sensitized red corpuscles. Thus fresh goat serum will reactivate
heated normal goat serum in its action on rabbits and on guinea-
pig's red corpuscles, the amboceptors for which are contained in
untreated animals. It will also reactivate the heated serum
of goats injected with rabbit's corpuscles, ox corpuscles, or dog's
corpuscles. Ehrlich and Sachs found that if they added this fresh
serum to rabbit's corpuscles, the complements which took part in
the two latter reactions were absorbed, whilst the others were
unaltered. Evidently, therefore, two complements at least are
present in goat serum. When the serum was allowed to act
on guinea-pig's corpuscles, the complements which reactivated
the normal amboceptors for rabbit's and guinea-pig's corpuscles
were removed, also that which activated the artificial ox ambo-
ceptor. By an elaboration of these methods Ehrlich and Sachs
were able to demonstrate that these five hsemolytic actions depend
upon five different complements, each perfectly distinguishable
the one from the other. The subject will not be pursued farther,
though there is an abundance of evidence pointing to the same
end — that the serum of every animal contains a very large number
of complements, some of which take part in one reaction, some in
another. As a rule, complements which take part in haemolysis
have no action on bacteria, or but little, and this is all the truth in
MetchnikofT's theory of macro- and microcytase.1

Ehrlich lays it down as a law that amboceptors from a certain
species of animal are best complemented by complements from
this species, and this is true in general ; Muir has adduced some
exceptions.

The main experimental basis for the Unitarian theory was
furnished by Bordet and Gengou. They found that if they added
"sensitized" bacteria — i.e., bacteria which had been placed in
heated immune serum — in fresh serum, all the complements were
removed, and the fluid would no longer dissolve sensitized red
corpuscles, so that evidently the haemolytic complement had been
removed. Again, Bordet showed that if he added fresh serum to
sensitized red corpuscles and allowed solution to take place, all
complements, bacteriolytic as well as haemolytic, were removed.
These facts are not disputed, and we shall have occasion to refer

1 See p. 286.

154 HAEMOLYSIS — COMPLEMENT

to them again, as they are of great importance. Their interpreta-
tion is, however, still uncertain, and they cannot outweigh the
very precise demonstrations of numerous complements by Ehrlich
and his school. According to them the receptor, which, when
broken off, constitutes amboceptor, may have more than one com-
plementophile group, each adapted to seize and utilize different
complements. We must suppose that it can discharge its function
if it is supplied with one of these complements ; this is called the
dominant. It may also seize other complements, so that all its
affinities are supplied, and this is what takes place in Bordet's

FIG. 34. — PLURICEPTORS ATTACHED TO A CORPUSCLE, AND SHOWING THE
DOMINANT (a) AND NON-DOMINANT (b) COMPLEMENTS.

FIG. 35.— SOME OF THE CONSTITUENTS OF NORMAL GOAT SERUM.

a — Complement which dissolves sensitized sheep-corpuscles; b = that which
dissolves rabbit's corpuscles ; c = the normally occurring amboceptor for
rabbit's corpuscles. The next two diagrams shows how this is shown to
be a biceptor.

and Gengou's phenomenon ; these complements, which are not
essential to the lytic action, are called non-dominant or subordinate
complements.

Ehrlich and Sachs give an example of this. The amboceptor
of normal goats for rabbit's cells and that obtained by injecting
goats with ox corpuscles are, of course, both sensitized with
normal goat serum. Now if we take rabbit's corpuscles and add
them to fresh serum (amboceptor-complement), it is found that
both the complements which might theoretically be present are
removed, and the fluid will no longer reactivate heated immune
serum for ox corpuscles. But if the action is allowed to go on for
a short time only, and the corpuscles are then centrif ugalized down
and removed, it is found that only the non-dominant complement

BACTERIOLYSIS AND ALLIED PHENOMENA 155

is removed, and the supernatant fluid will still reactivate the action
of heated immune serum on rabbit's corpuscles. This is not the
invariable rule, and in other cases the non-dominant complements
are not seized until the dominant has been anchored. In either
case we can see an alternative explanation for the facts observed
by Bordet and Gengou. Ehrlich points out that the presence of
the power to bring these many ferments into relation with the
giant molecule of the protoplasm must be of advantage in cell
nutrition, as enabling the greatest possible effect to be produced,
and quotes Hoffmeister as showing that the liver cell contains ten
different ferments, and Delbruck as attributing five to the yeast
cell.

The discovery that the haemolytic action of snake venom
depends on the mechanism of amboceptor and complement made

FIG. 36.— SHOWING EFFECT OF ADDITION OF RABBIT'S CORPUSCLES TO THE
SERUM SHOWN IN FIG. 35.

The amboceptor unites with the corpuscle, and all the complements are with-
drawn. The supernatant fluid loses its power of dissolving ox corpuscles
previously sensitized. (See also Fig. 37.)

by Flexner and Noguchi, and the excellent subsequent work of
Kyes, made us realize that the process might be even more
complex. Flexner and Noguchi "^showed that snake venom
contains two substances, one of which is thermostable, not being
destroyed at 90° C., and which in itself has no power of bringing
about solution, and which is evidently equivalent to amboceptor.
It may be reactivated by the other thermolabile substance or by
the complements of normal serum. For example, horse corpuscles
can be dissolved by heated venom and fresh ox serum, ox blood
corpuscles by heated venom and guinea-pig serum, and so on.
Kyes found that some corpuscles, washed from all trace of
serum, could be dissolved by cobra venom alone ; others required
the concomitant presence of fresh serum of the same or other
species. Thus the corpuscles of man, the dog, guinea-pig, etc.,
are dissolved by the venom alone, whilst those of the ox, sheep,
and goat are not. He formed the theory (for reasons which will

HAEMOLYSIS — ENDO-COMPLEMENT

not be discussed) that the solution of corpuscles without the action
of serum was due to a complementing of the venom amboceptor
by means of complements contained in the corpuscle itself, which
he called endo- complements. This he proved experimentally. He
argued that if he dissolved the red blood-corpuscles which contain
endo-complements by means of distilled water, the solution should
act as though it contained free complements, and should reactivate
venom amboceptor, combined with corpuscles which it was unable
to dissolve without the aid of serum. This was found to be the
case. A laked solution of human corpuscles mixed with venom
would haemolyze the corpuscles of the ox, sheep, and goat ; but a
solution of ox corpuscles and venom would not dissolve the cor-
puscles of the goat or sheep. He showed that these endo-
complements are thermolabile, though more resistant than most

FIG. 37. — SHOWING EFFECT OF SHORT EXPOSURE OF RABBIT'S CORPUSCLES
TO THE SERUM SHOWN IN FIG. 35.

The non -dominant complement is absorbed, and the supernatant fluid no
longer dissolves sensitized ox corpuscles, whilst retaining its action on
sensitized rabbit corpuscles. (In most cases the dominant complement is
removed first.)

of those in sera. They are destroyed at 62° C. in half an hour.
Kyes gave an additional demonstration of his theory by showing
that if rabbit's corpuscles are soaked for twenty-four hours in
normal saline solution, the endo-complement dissolved out, and
could be demonstrated in the fluid, whilst the corpuscles were no
longer haemolyzed by snake venom without the addition of serum.
He holds, therefore, that the corpuscles contain endo-complements
able to dissolve them, but only when brought into organic relation
with their protoplasm by means of amboceptor.

He next went on to demonstrate that a definite chemical
substance — lecithin — can act as a complement to snake venom,
and that either it or the serum-complement could act as a
dominant so as to bring about solution, or that the two might act
together. Lastly, he was able to prepare a combination of cobra
venom and lecithin, which he calls cobra-lecithid, which is quite
different in its physical and chemical characters from either of its

BACTERIOLYSIS AND ALLIED PHENOMENA 157

components, and which acts haemolytically on all red corpuscles.
It also acts much more quickly than ordinary solutions of venom
even when activated with lecithin, and it is very thermostable,
resisting boiling for six hours. It is evidently a definite chemical
substance and of great theoretical interest, as being apparently
an amboceptor-complement compound preparable in a pure state.
The analogy between the complements and the exotoxins leads
to the inquiry whether the production of an anticomplement occurs
when an active serum is injected into an animal of another species.
Ehrlich and Bordet both proved this to be the case. He injected
fresh horse serum into goats, and obtained a serum which would
prevent the normal complementing action of horse serum.1 He
easily showed that this was not due to any action which it exerted
on the red corpuscles (rabbit's) themselves or on the amboceptor.

FIG. 38. — SHOWING THE EFFECT OF ADDING TO SENSITIZED RABBIT'S
CORPUSCLES (a, b) A MIXTURE OF FRESH HORSE SERUM (CONTAINING
COMPLEMENT, c), AND SERUM FROM A GOAT WHICH HAD BEEN INJECTED
WITH HORSE SERUM AND CONTAINED ANTICOMPLEMENT, d.

The complement and anticomplement combine, and the corpuscles are
unaffected ; this is shown by the fact that they will dissolve on the
addition of fresh complement.

It was clear, therefore, that it acted as an anticomplement. The
question now arose, Was this action exerted on the haptophore or
the zymophore group ? The following experiment shows that it
acts on the former : Sensitized red corpuscles were treated with a
rmxture of normal serum and of serum containing anticomplement:
no solution took place. The corpuscles were then centrifugalizecj.
ott andTresh serum added : haemolysis occurred. Evidently the
anticomplement had not acted on the zymophoric group, for if it
had the haptophore group would have combined with the ambo-
ceptors of the sensitized corpuscles, and would have shielded
them from further action when fresh serum was added. What

1 There is a possible fallacy here, to be mentioned under our description of
Bordet and Gengou's phenomenon. It may be simply a binding of the
complement in the precipitate. This is discussed later, and at present
Ehrlich's views will be set out as if they were definitely proved.

158 HAEMOLYSIS — ANTICOMPLEMENT AND COMPLEMENTED

occurred was that the haptophore radical of the complement had
been " blocked " by the anticomplement.

In a few cases anticomplements which will act on some only
of the complements of a sera can be produced, and from the study
of these additional evidence in favour of the multiplicity of com-
plements has been adduced.

As regards the nature of these anticomplements, this is readily
explicable on the side-chain theory. Ehrlich supposes that when
a foreign serum is injected into an animal it may find no receptors
for which it has an affinity ; in this case it forms no anticomple-
ment, and such cases are known to occur. Or it may find receptors
which it can activate completely, just as the normal complement
of the animal can do ; in this case also there will (under ordinary
circumstances) be no formation of anticomplement. Lastly, it
may find receptors with which it can combine, but the resulting
combination may be useless in the nutrition of the cell ; in this

FIG. 39.— SHOWING EFFECT OF ADDITION TO SENSITIZED CORPUSCLES OF A
MIXTURE OF COMPLEMENT AND COMPLEMENTOID.

The former unite with the complementophile haptophore groups of the ambo-
ceptor ; the latter, which have decreased in combining affinity, do not.

case the receptor will be cast off, and will constitute anti-
complement.

Lastly, Ehrlich was able to show that in some cases an
auto-anticomplement may be formed — i.e., an antibody which can
neutralize the complement normally present. The proof for this
is elaborate, and will not be given here.

The change of toxins into toxoids is paralleled by the change of
complements into complementoids. We have previously spoken of
the complements as being destroyed by heat, but this is not quite
correct ; they lose their zymophore groups, but retain their hapto-
phore portions, though these appear to lose some of their combining
affinity. The proof of the existence of complementoids is simple :
when injected into animals they call forth the production of
anticomplements, just as toxoids call forth the production of
antitoxin.

For some time it was found impossible to demonstrate the

BACTERIOLYSIS AND ALLIED PHENOMENA 159

presence of complementoids in test-tube experiments, and this
was owing to the fact that the affinity of the haptophore group of
the complements suffered loss during the process of heating, so
that when a mixture of complement and complementoid was
added to sensitized red corpuscles, the former combined just as
well as if the latter were not present. " Blocking " of the com-
plementophile group of the amboceptor appeared not to occur ;
but Ehrlich and Sachs were able subsequently to demonstrate
such a phenomenon in this wise : Dog serum contains an ambo-
ceptor for guinea-pig's corpuscles, and this can be activated either
by dog's or guinea-pig's serum. New when a mixture of heated
dog's serum, guinea-pig's corpuscles, and guinea-pig's serum was
made, solution took place ; but when the corpuscles and heated
serum were added together, allowed to stand, and then centrifuga-
lized and removed, no solution occurred on the addition of fresh

FIG. 40.— DEMONSTRATION OF " BLOCKING " OF COMPLEMENTOPHILE GROUPS
BY COMPLEMENTOIDS.

a = complementoid, and & = amboceptor in heated dog serum; r = rabbit's
corpuscles (see text).

guinea-pig's serum. Ehrlich proved1 that the combination of
amboceptor and corpuscles took place as usual, and that it was then
followed by a blocking of the free arms of these amboceptors by
the complementoids present in the heated dog's blood. In this
case, therefore, the complement had not suffered any appreciable
change in combining capacity as a result of its conversion into
complementoid.

It follows, therefore, that when we heat fresh serum, or allow it
to stand a few days at the room temperature, we do not get rid
of the complements altogether. This can be done, as shown by
Von Dungern and others, by shaking the serum with cells (liver,
kidney, etc.), yeast, fungi, certain bacteria, powdered charcoal,
etc. This abstraction of complements by inert bodies is a non-
specific process entirely unlike its fixation in haemolysis, etc.,

1 The proof is complete and conclusive, but too long to give here.

i6o

HAEMOLYSIS — ANTI AMBOCEPTOR

and Ehrlich compares it to a process of physical absorption. The
presence of the phenomenon must be borne in mind, since it must
make us careful how we interpret the mere disappearance of
complement from a fluid after it has acted on articulate sub-
stances. The true test of specificity is its action on these
substances, not its absorption by them. The yeast cells are in no
ways injured through having absorbed complement.

We have not yet finished our study of the complements, but it
will be convenient to continue it later, and to conclude here Ehrlich's
studies on the antilysins. Anticomplement has been already
discussed, and we will now turn to anti-amboceptor.

This is a substance of nature different to any that we have
previously studied, in that it is an antibody to an antibody. It is
prepared by injecting an immune serum into an animal other than
that from which it was derived. Thus, if an antityphoid (bacterio-
lytic) serum procured by immunizing a horse with typhoid bacilli
be injected into a rabbit, the serum of the latter acquires the
power of neutralizing the bacteriolytic action of the former.
Antibodies to antibodies cannot always be produced; thus, anti-
antitoxin is not yet known. Again, Metchnikoff was unable to
prepare antispermotoxin by injecting a cytolytic serum for sperma-
tozoa into the guinea-pig. In this case there can be no question of
the absence of suitable receptors, for the animal's spermatozoa are
attacked in vitro by the serum injected. Hence Ehrlich made
the assumption of the presence of sessile receptors, which are not
cast off when the suitable antibody is injected.

In studying the anti-amboceptor concerned in haemolysis,
Ehrlich first prepared a haemolytic serum by injecting a rabbit
with ox blood. When the serum of the latter was found to be
powerfully haemolytic, it was injected subcutaneously into a goat :
120 c.c. was given during an interval of two months. At the end
of this time this goat serum was powerfully antihaemolytic. This
was shown as follows : 0-00125 c.c. of (immune) rabbit serum was
found to haemolyze completely i c.c. of a 5 per cent, emulsion of
ox corpuscles, provided that sufficient normal guinea-pig serum
is added to provide complement. Now three times this amount
(0-00375), plus 0-5 of the goat serum, was added to i c.c. of the
emulsion of corpuscles, allowed to act at 40° C. for an hour, centri-
fugalized, and 0-15 c.c. of normal guinea-pig serum (complement)
added to the corpuscular sediment. No solution took place, so
that 0-5 c.c. of the anti-antiserum had completely neutralized three

BACTERIOLYSIS AND ALLIED PHENOMENA l6l

dissolving doses of the antiserum. A control test showed that
normal goat serum was without action.

Theoretically, two anti-amboceptors are possible : one which
will combine with the cytophile group of the amboceptor, and one
which will seize on the complementophile. In his earlier studies
Ehrlich thought that the former was that actually produced when
immune sera are injected into animals of another species. He
later altered his opinion, and now holds that the substance
actually formed is an antibody to the complementophile hapto-
phore of the amboceptor. He was led to this conclusion by a
discovery of Bordet's, which was adduced as evidence against the
amboceptor theory, but which was ingeniously adapted to its
defence by Ehrlich. The discovery was as follows : Normal
rabbit serum when injected into a guinea-pig leads to the production
of an anti-antibody, which neutralizes the haemolysin formed by

FIG. 41. — THE COMPLEMENTOPHILE HAPTOPHORE GROUPS OF AN AMBO-
CEPTOR "BLOCKED" BY ANTIAMBOCEPTOR (a, a}.

immunizing with ox blood. This, of course, is readily explicable
if normal rabbit serum contained an amboceptor for ox corpuscles,
but this is not the case. Ehrlich argued as follows : An anti-
amboceptor is obviously formed, but the normal rabbit serum
injected contains no amboceptor — i.e., no cytophile groups ; hence
the ant -amboceptors produced must be anticomplementophile.
For this to happen it is necessary that this normal rabbit serum
must contain complementophile groups, and this is obviously the
case, since other amboceptors, each with a complementophile
group, are present. Therefore, on the assumption that these
groups are similar, even when they occur in different amboceptors,
the case is clear. But are these complementophile groups similar ?
Does not this contradict Ehrlich's views on the multiplicity of
complements ? To avoid this difficulty, Ehrlich falls back on his
polyceptor theory. All or most amboceptors are really poly-
ceptors, differing only or mainly in their cytophile groups, and
possessing numerous complementophile groups, which are similar
or identical in different cases in the same species. But if this

ii

l62 RECAPITULATION OF EHRLICH'S THEORIES

is the case, anti-amboceptors should be non-specific ; they are

directed against the complementophile groups, which are the same

in all amboceptors from the same species. This subject has not

yet been fully investigated, but Pfeiffer and Friedberger have

shown that the antibody to the immune body against the cholera

vibrio also acts against the antityphoid serum. Ehrlich also

showed that if anti-amboceptor be added to sensitized cells^these

are protected against the subsequent addition of complement.

/This, of course, is attributable to the fact that the anti-amboceptor

I unites with the free complementophile groups, and " blocks " them

[^against the subsequent access of the complement molecule.

Ehrlich leaves undecided the question as to whether an anti-
cytophilic anti-amboceptor is ever produced, but holds that it
might be formed if by any process we could destroy the cytophile
haptophore of the amboceptor molecule. Under ordinary circum-
stances the complementophile group has a greater affinity for its
cell receptor than has the cytophile for the receptors with which
it could combine ; union takes place between the former, and the
latter is, so to speak, pulled out of the way of its affinity.

It is now advisable to recapitulate briefly some of the main
points in Ehrlich's theory of the structure of the bacteriolytic and
haemolytic sera. Firstly, as regards complements : Of these there is
very great number, and each is especially adapted for the solution of
one or more varieties of cells, which it can dissolve in the presence
of a suitable amboceptor ; it is known as the dominant comple-
ment. Other complements, however, may help in the process,
and these are termed non-dominant. In general terms the comple-
ments, which are especially active in haemolysis, have but little
action on bacteria, and vice versa. Secondly, as regards the ambo-
ceptor : This is really a polyceptor, and is so constituted that it
can combine with the cell to be dissolved, on the one hand, and
with a large number of molecules of complement, on the other.

Of the further complications of the theory which Ehrlich has
introduced to explain new phenomena as they arise, of the
amboceptoids, of loose and firm union, etc., we do not propose
to speak. That the theory needs these complications in order to
account for the phenomena is not necessarily in its disfavour, for
the phenomena themselves are complex in the extreme. And it
must be regarded as a strong argument in its favour that Ehrlich
has again and again deduced results from this theory which sub-
sequent research has shown to be correct ; and very few hypo-

BACTERIOLYSIS AND ALLIED PHENOMENA 163

theses can be adduced which have been so rich in leading to the
discovery of new facts.

It will now be necessary for us to glance briefly at some
alternative hypotheses, and it must be pointed out that the com-
paratively short notice they will receive must not be taken to
imply that there is less to be said in their favour. The trend of
modern research tends on the whole against Ehrlich's views ; yet
in the history of immunity his researches will always be regarded
as of the greatest interest and value.

It may be pointed out that the cast-off receptors of red
corpuscles, bacteria, etc., may combine with the complemento-
phile groups of an amboceptor, and thus appear to act as a
cytophilic anti-amboceptor.

Bordet holds that the immune body is not an amboceptor at
all — i.e., that it does not act as a link between the corpuscle or
bacterium and the complement, but that it sensitizes the former
and renders it susceptible to the action of the latter. On Ehrlich's
theory, amboceptor unites with cell and complement with ambo-
ceptor; on Bordet's, both substances unite with the cell direct.
On the latter theory many, if not all, of the results observed by
Ehrlich and others are explicable, and there appears to be no
experimentum crucis by which the truth can be determined. A diffi-
culty in the way of accepting Ehrlich's amboceptor theory is this :
there is no proof that complement and immune body ever unite
unless the latter has already combined with a cell, or with anti-
amboceptor, or with free receptors. The only example to the
contrary is supplied by Ryes' cobra-lecithid, which, though of
great interest, can hardly be quoted as a case in point, since
lecithin differs so markedly from the ordinary thermolabile
complements of serum. Another case (in the deviation of the
complements) in which this process appears to take place is
extremely complicated, and the evidence in favour of the direct
union of complement and antibody quite unsatisfactory. It
follows, then, that we must either assume that the mere union
of the cytophile group of the antibody with some other object
increases the affinity of the complementophile group for comple-
ment (as has been done here), or we must agree with Bordet that
cell and complement unite directly, but only after the latter has
been prepared by the action of immune body. A very important
observation of Muir's is of interest in this connection. Muir
showed that after red corpuscles had been saturated with immune

II — 2

164 HAEMOLYSIS — BORDET'S VIEWS

body and then with complement, some of the immune body, but
none of the complement, might be dissociated from the combina-
tion, and become free in the fluid. It is difficult, though perhaps
not quite impossible, to reconcile this experiment with Ehrlich's
theory, whilst it is really explicable on that of Bordet ; but Muir
(whose recent work on this difficult point should be consulted)
has not come to a definite conclusion as to the nature of the
combination.

Bordet also holds peculiar views on the subject of the union
of corpuscles and immune body. According to Ehrlich, the com-
bination is due to a strong chemical combining affinity between
the two substances, and the resulting compound is a stable one,
which is not readily dissociable. Cells thus activated possess a
strong affinity for complement, and the whole process is a chemi-
cal one, dependent on ordinary chemical laws and obeying the
laws of multiple proportions. This was, on the whole, confirmed
by Muir, who proved that corpuscles combined with multiple
doses of immune body took up multiple doses of complement ;
but he also showed that the corpuscle-antibody compound did
dissociate, so that when corpuscles which had been saturated
with immune body were placed in contact with normal corpuscles,
some of the amboceptor left the sensitized and attached them-
selves to the normal. Bordet adduced evidence to show that the
combination is of a remarkable nature, and dependent on obscure
chemico-physical reactions. He took a sample of haemolytic
serum, and determined the amount of blood which it could
haemolyze when the two were mixed together. We may assume,
on Ehrlich's theory, that all the suitable receptors in the cor-
puscles were saturated with amboceptor. But Bordet showed
that if the corpuscles were added a little at a time, a very much
smaller dose might take up all the immune body. Hence he
compared the process to the staining of filter-paper when im-
mersed in a dye. If the paper be added at once, it will be stained
a uniform colour, whereas if it be added a little at a time, the first
pieces will be stained deeply, the subsequent ones less and less,
until the dye is completely absorbed. It would appear to explain
the phenomenon equally well if we assumed that the corpuscles
could be haemolyzed by fewer amboceptors than they could take
up, and there is some experimental evidence of this. Thus Muir
showed that after red corpuscles have been dissolved by haemo-
lytic sera, the addition of more immune body will lead to the

BACTERIOLYSIS AND ALLIED PHENOMENA 165

absorption of more complement. It may be added, however, that
many of the recent researches on antibodies point to them as
being governed by the very complex laws, as yet not fully under-
stood, which regulate the actions of colloids on one another
and on crystalloids. This is discussed, though very briefly, in a
subsequent chapter.

As regards the evolutionary significance of the facts of cytolysis
(using the word to cover the solution of bacterial and all other
cells), there are two theories which, though apparently widely
different, have, on ultimate analysis, much in common. We have
seen how Ehrlich explains the process of cell nutrition, the role
which he attributes to complement, and the method in which he
conceives the complex antibodies to be formed. On his theories,
therefore, the fundamental process is one of cell nutrition, and
we may imagine it to have become adapted to serve as a means
of immunity during the course of natural selection : the animals
which could make use of their mechanism of cell nutrition as a
means of defence against invading micro-organisms would survive
and perpetuate their species ; the others could not.

We may add a few more considerations on this process of
cell nutrition. We must assume that the food molecules of
proteid which circulate in the blood are in an indifferent form,
and are equally well adapted for the nourishment of cells of the
brain, liver, etc., this method of transmission of nourishment
having been found the most convenient and economical. The
first requisite for the nutrition of the cell is that one of these
molecules shall be brought into close relations with its protoplasm.
Ehrlich speaks of this as being due to the selective action of the
cell receptors, but we must not be misled by terms or diagrams,
however convenient they may be to enable us to form a mental
picture of the process. Ehrlich simply means that certain
specialized molecules or groups of molecules have a chemical
combining affinity for certain food molecules. This union is the
first requisite, and in some cases it may be sufficient, and the
molecule may be forthwith incorporated with the cell protoplasm.
In most cases, however, this is not so, and the cell has to
transform an indifferent molecule of proteid material into a form
suitable for assimilation into itself. We now know that most, if
not all, of the activities of a living cell in metabolism are dis-
charged through the agency of enzyme action, and that each of
the metabolic actions of a cell appears to depend on a special

l66 CELL NUTRITION

enzyme. It is natural, therefore, that the cell should make use of
soluble ferments in preparing the molecule which it has seized.

We now (thanks to the researches of Fischer, Hierf elder, and
others) know that the action of these enzymes is strictly specific,
and dependent on a stereo-chemical relation between the ferment
and the body acted on, and that it is necessary in all cases for a
union to be brought about between the two substances before the
specific action is effected. Thus, if the two could unite in the blood,
the molecules would be broken down before they reach the cells, and
the food material would reach it in an indifferent form, which,
however well adapted to some cells, might be useless to others.
The substances, therefore, do not unite in the blood-stream at all,
but the cell first seizes a molecule of food-stuff, and then one or
more ferments which can change it into the exact form which the
cell requires. Thus, the process of cell nutrition, according to
Ehrlich, consists in the establishment of a link between the food
and a suitable ferment after the former has been selected by the
cell.

We may inquire why the cell does not form its own ferments.
In certain cases it may do so ; the existence of the endocom-
plements lends support to this view. But it is readily conceivable
that it may be more economical in every way for the animal
to limit this process of enzyme formation to certain specialized
cells or tissues, just as pepsin is formed by the cells of the gastric
glands. This is probably the case, and all experience goes to
show that the cells to which this duty is delegated are the
leucocytes.

Ehrlich, therefore, conceives of the cytotoxic mechanism of
amboceptors and complements as being adapted, in the first place,
to cell nutrition, and, secondarily, to the defence of the body by
bringing about solution of foreign cells. Metchnikoff 's views may
be summarized as follows: He holds that both the substances
taking part in cytolysis are ferments, and that both are adapted
for intracellular digestion. His views on phagocytosis will be
discussed subsequently, and here it will be sufficient to give the
merest outline. He has shown that in the lowest Metazoa — e.g.,
in the Coelenterata — digestion is an intracellular process entirely ;
the hypoblastic cells lining the alimentary canal seize food particles
from the lumen and digest them, but form no secretion like those
of the higher animals. In the cells which have seized and are
digesting food particles in this way he was able to demonstrate

BACTERIOLYSIS AND ALLIED PHENOMENA 167

ferments allied to pepsin or trypsin, which had evidently been
formed to digest the ingested food material. He then showed
the importance of this process in the absorption of cells, etc., in
the tissues of higher animals, and demonstrated that the two
processes are altogether similar. Hence his views on the nature
of the cytolysins arise easily and naturally. He holds that com-
plement, or, as he calls it, cytase, is the digestive secretion of
the leucocytes, and that, under ordinary circumstances, it is
retained within the leucocytes ; it is only set free when leucocytes
are dissolved (phagolysis), either as the result of an injection of a
foreign substance or in the process of clotting (this theory, as we
shall show, is held by many other authorities). Thus in Pfeiffer's
phenomenon the first result of the injection is phagolysis, and the
ferments set free in the process immediately attack the cholera
vibrios. Metchnikoff has proved clearly that the injection of
almost anything into the peritoneal cavity leads to a diminution in
the number of leucocytes, but the proof of the existence of
phagolysis under these circumstances is less convincing. Cytase,
therefore, is a digestive ferment adapted to deal with large masses
of food substance rather than with molecules, as Ehrlich supposes,
and is normally intracellular, being formed especially for intra-
cellular digestion.

Metchnikoff sees in amboceptor, or fixator, a substance alto-
gether analogous to enterokinase, and acting, like it, as an
accessory digestive ferment, which has for its object the linking
of the more potent ferment to the food to be digested. He
regards it also as being formed in the leucocytes, and for this
reason (amongst others, which will be enumerated subsequently) :
He states that the amount of fixator or amboceptor produced
is proportional to the amount of phagocytosis which occurs during
the absorption of the antigen. For example, he states that when
defibrinated goose blood is injected into the guinea-pig sub-
cutaneously there is but little phagocytosis, the blood being
dissolved extracellularly, and but little immune body is produced ;
but when the injection is made into the peritoneum there is much
phagocytosis and much development of antibody. There is cer-
tainly a remarkable difference between the various tissues as
origins of antibodies, and, as a rule, the subcutaneous and con-
nective tissues are most potent in this respect, the peritoneum
next, and the circulating blood worst ; but there is no sufficient
evidence to show that this depends on the amount of phagocytosis

168 METCHNIKOFF'S VIEWS

which occurs in these different situations. The site of the forma-
tion of an immune body will be referred to again.

Thus, whilst Ehrlich sees in the substances taking part in
cytolysis evidence of the nutrition of the living molecules of cell
protoplasm, Metchnikoff sees in them ferments which were also,
in their first appearance in the animal economy, designed for the
elaboration of the nourishment of the body or of certain cells
therein, the difference being that, according to him, their primitive
function was the digestion of large particles as well as of molecules.
But in the highly-organized animals with which we have chiefly
to deal this function has been changed, for in them phagocytes do
not ingest bacteria and red corpuscles for the sake of the nourish-
ment they contain, but to rid the body of invaders ; the body as a
whole is nourished through the agency of other extracellular
digestive ferments, which are secreted into the alimentary canal.
It follows, therefore, that cytase is unnecessary in the circulating
blood, and, on this supposition, does not usually exist in that situa-
tion. Metchnikoff admits that immune body does so exist — and has
indeed supplied a remarkable proof of the fact, since he showed
that the spermatozoa of immunized guinea-pigs are combined
with immune body, and only need the addition of fresh serum for
cytolysis to occur. Cytase and fixator, on MetchnikorPs explana-
tion of the phenomena, must be regarded as substances which in
the evolution of the animal kingdom were first developed as
digestive juices, but which now are entirely subservient to the
defence of the body against invaders. According to Ehrlich, they
are both in daily use in nourishment, and their defensive function
is of less importance than their value in cell nutrition. In either
case, the conception of immunity, as being fundamentally a process
of nutrition, is a most striking one.

We have now to turn to a phenomenon which is of the highest
theoretical importance, and which bids fair to be of great practical
value in diagnosis. We have already referred to the fact that
Bordet, the chief advocate of the Unitarian theory of complement,
showed that if red corpuscles be sensitized with amboceptor and
added to fresh serum, the latter is deprived of all complementary
activity. The same phenomena occur with sensitized bacteria.
When, for instance, typhoid bacilli which had been acted on by
heated typhoid serum were added to fresh blood, allowed to act, and
then centrifugalized, it was found that the supernatant fluid had no
longer the power of dissolving red corpuscles sensitized by suitable

BACTERIOLYSIS AND ALLIED PHENOMENA l6g

amboceptor. These facts are admitted by Ehrlich, but he shows
that they do not constitute any real evidence against the pluralist
conception, since in some cases it may be shown that, though all
complements are absorbed, this may be at different rates, and by
stopping the process at a proper time, some may have disappeared,
whilst others are left.

The practical importance of these observations arises from the
fact that they give us a method by which we can demonstrate the
presence or absence of an antibody to a given antigen, or of an
antigen to a given antibody. For example, the sensibilatrice or
amboceptor for the tubercle bacillus is very difficult to demonstrate,
since the organism is so resistant that it is never obviously dis-
solved, even partially, in the most potent serum we can obtain.
Nor are bactericidal experiments more promising, owing partly to
the resisting power of the organism and partly to the technical
difficulties. The only evidence (apart from the presence of
agglutinins) which we have in favour of the formation of specific
antibodies to tuberculosis is derived from an application of Bordet's
phenomenon. The experiments were carried out as follows : A
guinea-pig was injected with the bacillus of avian tuberculosis, to
which it is but slightly sensitive, and the blood was examined by
mixing it with an emulsion of the bacilli. If amboceptors were
present they would combine with the bacilli, and draw to them all
the complements of the fluid, which would thus lose its power of
activating suitably sensitized red corpuscles. This was found to
occur, and Bordet and Gengou deduced that the guinea-pig had
formed antibodies to the slightly virulent tubercle bacilli. When,
on the other hand, the guinea-pigs were injected with virulent
human tubercle bacilli, no such antibodies could be demonstrated.

As an example of the recognition of an antigen by means of
Bordet's phenomenon, we shall quote Bruck's demonstration of
the presence of tuberculin, or at least of some derivative of the
tubercle bacillus, in the blood of patients suffering from general
tuberculosis. Here the problem is changed. We have a specimen
of serum which we want to test, not for the antibody, but for the
antigen. The procedure is as follows : The serum is heated and
mixed with a serum known to contain antibodies to the tubercle
bacillus (antituberculin of Hochst). To this mixture is added
fresh guinea-pig's serum, and lastly sensitized red corpuscles. It
is found that no haemolysis takes place. In a control experiment,
in which normal human blood took the place of that from the

170 THE BORDET-GENGOU PHENOMENON

patient, haemolysis occurred ; it followed, therefore, that the
patient's blood contained derivatives of the tubercle bacillus, or
(as we may fairly say) its toxins.

The technique of these experiments is somewhat difficult, but
the results have been so important, and the method seems so
likely to play a part of importance in clinical diagnosis, that it will
be discussed in a further chapter.

Gengou's phenomenon is similar to Bordet's. It requires a
little anticipation of facts to be discussed subsequently concerning
the precipitins. These are antibodies obtained by the injection
of proteid solutions, and having the power of uniting with these
proteids to form insoluble precipitates. Gengou showed that in
this combination of antigen and antibody the same absorption of
complement occurred as in the case of sensitized red corpuscles
or bacteria. The process is an extraordinarily delicate one, and
it has been shown to be demonstrable with as little as o-oooooi c.c.
of the antigen (in this case normal human serum), and may occur
when the serum used is so dilute that no visible precipitation
occurs.

It appears, further, that this process of fixation of complement
is a general one, occurring whenever an antigen and its antibody
unite, whether the antigen occurs in solid or liquid form, and
whether the resulting compound forms a precipitate or remains in
solution. It has been shown by Nicolle and by Armand-Delille
to occur in the neutralization of tetanus and diphtheria toxins by
their appropriate antitoxins, and by Pozerski in the interaction of
papain and its antiferment. It has been shown by Guedini that
when hydatid fluid is mixed with the serum of an animal which
has been injected therewith a similar phenomenon occurs, and
this fact has been suggested and applied by Weinberg, Parvu, and
Lanbry to the diagnosis of hydatid cysts in man.

The theoretical importance of this phenomenon arises from the
light which it throws on the difficult subject of complementoids
and anticomplements. We have seen that heated serum has no
complementary activity, but that its injection into suitable animals
appears to call forth the presence of anticomplements. A little
consideration will show that this apparent anticomplementary
action can be explained equally well by the absorption of the
complements in a specific precipitate. We will take a particular
case.

Goat serum heated to 56° C., and therefore containing no active

BACTERIOLYSIS AND ALLIED PHENOMENA

complement, is injected into a rabbit. This is supposed to develop
an anticomplement in virtue of the complementoids it contains ;
we know that it also develops a precipitin which combines
with the proteids of' goat serum, and in doing so entangles any
complement which may be present. Now the complement of
goat serum can dissolve ox corpuscles when sensitized by a
suitable amboceptor (e.g., serum of a rabbit which has been
injected with ox corpuscles). When, however, the serum of the
normal goat is mixed with that of the rabbit which has been

<=V<7-txT"~

injected with rabbit serum and used in this way, no haemolysis
occurs. This Ehrlich explains on the supposition that the rabbit
serum contains anticomplement, but it is also explicable, as
Moreschi and Gay have shown, on the supposition that the com-
plements are all absorbed in the precipitate (perhaps an invisible
one) formed in the mixture.

It is obvious that either interpretation may be the correct one,
or that both processes may come into play. Moreschi has
investigated the process further, and his results tend to show that
the experiment is best explained on the absorption of complement
theory, and not by the presence of anticomplement. One of his
experiments, and that very ingenious, will be described. He
injected a rabbit with hen's egg albumin, and obtained a precipitin
to that substance ; this he found to act as an anticomplement
to fowl serum, but not to that of the rabbit, guinea-pig, or goat,
to which, of course, it was not a precipitin. He found, however,
that if he added to any of these latter sera a minute trace (^30 ws
of its volume) of egg-albumin, it appeared to become so, the
explanation being that a precipitate was formed, and that the
complements were entangled in it.

The phenomenon has also been invoked to explain some extremely
interesting researches of PfeifTer and Friedberger, who thought
they had demonstrated the presence of antibacteriolytic substances
in normal serum. They prepared a mixture of cholera vibrios
and normal serum, which they allowed to stand for some time, and
then removed the bacteria by centrifugalization. This would
remove any amboceptor which the serum might contain, and they
thought that they could demonstrate that an anti-amboceptor
still remained. They added to the serum thus prepared some
anticholera serum in suitable amount and a lethal dose of cholera
vibrios, and injected the whole into an animal, which invariably
died ; the amount of antiserum was sufficient to protect it if no

172 THE BORDET-GENGOU PHENOMENON

serum which had been exhausted by the action of cholera vibrio8
had been added. The anti-amboceptor which thus appeared to
be present in normal serum seemed to be strictly specific ; a
serum which had been weakened by the action of cholera vibrios
had no action on antityphoid serum, and vice versa.

Besredka attempted to explain these findings on the assumption
that " free receptors " passed from the bacteria into the normal
serum, and, when subsequently mixed with antiserum, combined
with it, and so prevented its union with the living bacteria. They
showed, in reply, that normal saline solution had no such action ;
but this is hardly conclusive, since receptors are dissolved from
the bacteria much less powerfully in normal saline than in serum.

According to Sachs, who demonstrated a similar occurrence in
haemolysis, the inhibiting substance which actually occurs in the
serum is an anticomplement. Gay, however, believes that the
whole series of phenomena can be explained by the absorption of
the complements in a specific precipitate. Thus, in Sach's
experiment normal rabbit serum heated to 55° C. was treated with
sheep's corpuscles. It was then removed, and added to a mixture
of heated serum immune to sheep's corpuscles (obtained from a
rabbit treated with sheep's corpuscles), fresh guinea-pig's serum,
and sheep's corpuscles. No haemolysis occurred, as it did in the
control experiment, in which no " exhausted " normal serum had
been added. His explanation was that the normal serum con-
tained amboceptors which combined with sheep's corpuscles,
leaving other amboceptors which have a greater affinity for
complement than those combined with the corpuscles used as a
test. Gay's explanation (which is almost certainly the correct
one) is quite similar to Moreschi's explanation of the anti-
complements. He believes that the immune serum contains also
a precipitin to sheep's serum, and that in the " exhaustion " of the
heated normal serum by the sheep's corpuscles some sheep's
serum was added to it, the corpuscles in Sach's experiment not
having been thoroughly washed. Thus a serum precipitate was
formed, and the alexins or complements combined therewith, so
that no haemolysis occurred on the subsequent addition of sensitized
corpuscles. He showed that if thoroughly washed corpuscles
were used for the exhaustion the phenomenon did not occur, and
gives other experimental proof.

Gay has also attempted to explain the process of deviation of
the complements in bacteriolysis (the Neisser-Wechsberg pheno-

BACTERIOLYSIS AND ALLIED PHENOMENA 173

menon) in a similar way. The process is described at length on
a subsequent page, but consists essentially in this : when bacteria
are acted on by an excess of amboceptor, no bacteriolysis may
occur, although sufficient complement is present. Gay's researches
on this point have not been fully published at the time of writing,
but he apparently thinks that soluble portions — receptors — may
pass off from the bacteria into the fluid in which the latter are
suspended, form a precipitate with the serum, and absorb the
complement. Much evidence will be required before this can
be established ; it might perhaps explain the phenomenon in
vitro, where the available amount of complement is limited, but
the effect is also demonstrable in the peritoneum, where we should
expect as much complement to be forthcoming as is required.

There is no doubt that the discovery of the absorption of the
complements in serum precipitates has rendered uncertain many
of the deductions in the process of haemolysis which have emanated
from the Ehrlich school, and the whole series of phenomena
requires re-investigation in the light of the observations of
Moreschi and Gay.

Experiments with bacteriolytic sera soon showed that, though
they protected against specific infections, they only did so in
certain doses. It is, of course, readily understandable that too
small an amount might be without action, but it was also found
that too large a dose might be equally inefficacious. Thus, Loffler
and Abel obtained a serum which protected against B. coli, and
found that when a lethal dose of this organism was given the
animal was only protected when it had received between 0-02 and
0*25 c.c. of serum, larger as well as smaller doses being equally
without effect. Similar results have been observed with sera
against cholera, dysentery, malignant oedema, and other organisms,
and seem to be general in dealing with bacteriolytic sera as
opposed to antitoxin.

They can also be obtained in vitro, and since this was first
shown by Neisser and Wechsberg, the phenomenon is often called
after them. Their method was as follows : They worked with
several organisms, amongst others with V. Afetchnikovi, and with
the serum of a rabbit which had been immunized against this
organism, and had acquired strong bactericidal powers. Equal
amounts of a broth culture of the vibrio were placed in a series
of test-tubes, and to each a dose of heated immune serum was
added. In all cases a uniform amount of normal rabbit serum

174

THE NEISSER-WECHSBERG PHENOMENON

(complement) was added, and the mixture made up to constant
volume with sterile normal saline solution. It was then incubated
for three hours at 37° C., and finally 5 drops from each tube
were plated out on agar and incubated. The result is shown on
the table.

Amount of Broth
Culture.

Heated Immune
Serum.

Fresh Normal
Serum.

Colonies Developing.

I C.C.

Infinity.

O'5 C.C.

Infinity.

0-25 c.c.

Many thousands.

O'l C.C.

Several hundreds.

o'O5 c.c.

About 100.

rTHHJ c-c- in aU

0-025 c-c-

0-3 c.c. in all

About 50.

cases.

o*oi c.c. cases.

None.

0-005 c.c.

None.

0-0025 c.c.

About 100.

O'OOI C.C.

Infinity.

o'ooo5 c.c.

Infinity.

This shows that when g-^Vzr c.c. of a one-day-old broth culture was
treated with 3 c.c. of fresh normal serum (complement) and in-
creasing doses of immune serum, there was no appreciable bacteri-
cidal action when more than 0-25 c.c. of the latter was used, or
with less than o-ooi c.c. When the amount was between o-i and
0-025 c.c., or about 0-0025 c.c., there was appreciable action, and
when it was between o-oi and 0-005 c.c. it was complete.

The same authors also demonstrated a very remarkable phe-
nomenon of exactly the same nature. A normal serum which has
a certain amount of bactericidal action (owing to the presence of
complement and of a small quantity of amboceptor) may have its
action increased, diminished, or nullified by the addition of a
powerful immune serum. This is shown by the following table,
which demonstrates the action of normal guinea-pig serum on
V. Nordhafen by itself and after the addition of variable doses
of heated immune serum.

This table shows that there is a definite relation between the
amounts of amboceptor and of complement which have to be
present to produce a maximal effect. In the case of the normal
serum, it is evident that there is more complement than can be
used by the amount of amboceptor present, and that the bacteri-
cidal action of the mixture increases as immune serum is added.
0-05 c.c. of normal serum was without obvious bactericidal effect
on c.c. of the culture, whereas the same amount plus 0*01 c.c. of

BACTERIOLYSIS AND ALLIED PHENOMENA

175

heated immune serum destroyed a very large number of the
bacteria. When, however, the immune serum is added in larger
proportion than this it does harm instead of good: 0-5 c.c. of

Number of Colonies after addition of Variable Amounts of

Amount of

Amount of

Heated Immune Serum.

Normal

Culture.

Serum.

No Serum
added.

i c.c. added.

o'i c.c. added.

0*01 c.c. added.

r-b- c.c.

I C.C.

None.

Many

A few.

None.

in all

thousands.

cases.

0-5 c.c.

None.

Almost in-

About 100.

None.

finity.

0-25 c.c.

A few.

Infinity.

Several

A few.

_

hundreds.

O'l C.C.

Several

Infinity.

Infinity.

About loo.

thousands.

0-05 c.c.

Infinity.

Infinity.

Infinity.

Many hun-

dreds.

o-o25 c.c. Infinity.

Infinity.

Infinity.

Infinity.

fresh serum completely sterilizes the amount of the culture used,
whereas after the addition of i c.c. of immune serum its action
is barely noticeable. An excess of amboceptor shields the bacteria
from the solvent action of the complements. These test-tube
experiments explain the results obtained in vivo, and enable us to
form some idea as to the mechanism of the process.

Neisser and Wechsberg offered the following explanation : They
assume that the molecules of complement and amboceptor unite
in the mixture before the latter has attached itself to the bacteria.
Owing to the excess of amboceptor, there will not be sufficient
molecules of complement to go round, and it will follow that a
variable proportion of molecules of amboceptor will be uncom-
plemented. Unless we suppose that the union of the complement
has changed the affinity of the cytophile group of the amboceptor
for the bacterium, it will follow that not all the amboceptors which
attach themselves to the bacteria will be charged with comple-
ment, and therefore able to exert a bacteriolytic or solvent action.
To take a concrete case, let us suppose that there are twice as
many molecules of amboceptor as of complement. Half the
molecules of the former will be complemented, and it will follow
that, though all the (appropriate) receptors of the bacterium are
occupied by amboceptor, only half of these will have their comple-
mentophile groups occupied, and this may not be enough to
injure it.

176 DEVIATIONS OF COMPLEMENT

They also imagine the possibility of the combination altering
the affinity of the cytophile group of the amboceptor for the cell.
It may diminish it, in which case fewer complemented molecules

FIG. 42. — FIRST POSSIBILITY.

The affinity of the amboceptor for complement is unaltered, as a result of
the union of the former with a bacterium.

of amboceptor would attack the bacterium than ever, and the
phenomenon of deviation of the complements would be even more
marked. The union might also (conceivably) increase this affinity ;
in this case the amboceptors which were complemented would

FIG. 43. — SECOND POSSIBILITY.

The affinity of the amboceptor for complement is diminished, as a result of the
union of the former with a bacterium.

seize on the bacterial receptors, to the exclusion of those which
were not, and the phenomenon of deviation of the complements
would not occur. This Neisser and Wechsberg think might occur
in the case of haemolysis, for in this process deviation has not

FIG. 44.— THIRD POSSIBILITY.
The affinity of the amboceptor for complement is increased.

been demonstrated, and solution occurs even when there is a
great excess of amboceptor. In the case of haemolysis by snake
venom, however, Myers and Stephens showed that the process
might only take place when medium doses of venom were added,

BACTERIOLYSIS AND ALLIED PHENOMENA 177

amounts too large or too small being without effect. This was
also corroborated by Kyes and Sachs, who showed that in
presence of an excess of venom (amboceptor) the addition of a
haemolyzing amount of complement might be without effect.
Now Kyes had already shown that venom does unite directly with
one of the substances (lecithin) which can complement it, and
this appeared strong evidence in favour of Neisser and Wechs-
berg's explanation. But Noguchi afterwards showed that the
protective action of large doses of venom protects the red cor-
puscles against the action of haemolytic agents other than comple-
ments. Corpuscles so treated are not dissolved by distilled water
or by tetanolysin. It appears, therefore, that strong solutions of
venom have a curious hardening effect on red corpuscles, render-
ing them resistant to haemolytic agents of all sorts. This effect
is apparently entirely different from the protection of a bacterium
by excessive doses of amboceptor, and should not be used as
evidence to support Neisser and Wechsberg's explanation.

Morgenroth has attempted to explain the absence of the devia-
tion of the complements in haemolysis by alleging that in this case
the amboceptors and complements do not unite until the former
have united with the red corpuscles ; this, of course, is the point
at issue. He finds that by adding anti-amboceptor (for the cyto-
phile group) to a mixture of amboceptor and complement the
phenomenon of deviation can be reproduced. But his experiment
can probably be explained by means of Gay's observations, to be
explained subsequently, and in any ease his anti-amboceptor would
only act as a free receptor of a red corpuscle, and the amboceptor
attached to it would be in the same condition as one anchored to
a corpuscle, which we know can absorb complement. In any case,
his experiment proves little or npthing.

Some experiments by Meakins afford the best evidence in
favour of the alteration in combining affinity undergone by ambo-
ceptor after union with complement, and call for short reference,
though their interpretation is somewhat uncertain. Meakins
experimented with corpuscles which had been very thoroughly
washed so as to remove all trace of serum. The importance of
this precaution will appear subsequently. He found that when
he added a large dose of heated immune serum and a small dose
of normal serum (complement) to corpuscles thus washed, and
allowed them to act, no haemolysis occurred. The corpuscles,
however, were sensitized ; for when they were centrifuged down,

12

178 NEISSER-WECHSBERG PHENOMENON IN HAEMOLYSIS

washed, and treated with more normal serum, haemolysis took
place. It thus appears that, in the first instance, the corpuscles
took up only those amboceptors which had no attached comple-
ment. It would appear, therefore, that amboceptor may unite
with complement whilst free, and that when it has done so its
affinity for the corpuscle is diminished. This is one of the
conditions which Neisser and Wechsberg imagined, and which
they showed would make the phenomenon of deviation even more
marked. And Meakins shows, though not in a very clear way,
that it does actually take place with corpuscles that are well
washed. The reason why it does not take place under ordinary
circumstances is explicable on the basis of Gengou's reaction.

But this theory cannot be regarded as entirely satisfactory. It
assumes, in the first place, a direct union between complement
and amboceptor before the latter has united with the cell or
bacterium, and of this there is no definite evidence apart from
this phenomenon. Ehrlich, it is true, supposes a feeble union
between the two which easily dissociates, even at low temperatures,
so that a cell which has been placed in contact with a mixture of
the two at o° will become saturated with amboceptor devoid of
complement. But this appears to be a fatal objection to Neisser
and Wechsberg's explanation ; for if the hypothetical amboceptor-
complement combination dissociates readily at a low temperature,
it should do so still more at a high one. Now the receptor-
amboceptor-complement will not dissociate, or only to a very
slight extent, since lytic action will at once commence. It follows
that the amboceptor-complement molecules will gradually all
dissociate, and the complements will sooner or later find their
way to the anchored amboceptors, and there remain. The process
may be delayed, but after a time all the attached amboceptors
will become complemented. And if, as seems implied in Ehrlich's
writings, the affinity of the complementophile group becomes
raised in virtue of the union of the cytophile group, with its
appropriate receptor, this process will take place all the more
quickly.

It appears much more likely that the Neisser- Wechsberg
phenomenon is an example of a reaction of a kind which has
already attracted much attention, and which may possibly modify
fundamentally our views on the antibodies, toxins, etc. They will
be referred to again when we deal with the agglutinins, and it is
sufficient to say here that in other cases, when there is no question

BACTERIOLYSIS AND ALLIED PHENOMENA 179

of two bodies, and therefore of the validity of Neisser and
Wechsberg's explanation, antibodies and other substances have
their action diminished or suspended when they are present in
excess ; thus with a great excess of agglutinin there may be no
agglutination. And Detre and Sellei have shown that phenomena
which are apparently somewhat similar occur in the haemolysis
induced by mercury perchloride.

In view of the important role in immunity which must be
ascribed to alexin — and, as we shall see, it acts not only as a
bacteriolytic agent, but also plays a part of great importance in
phagocytosis — an inquiry into its source becomes of paramount
importance. The theory was put forward very early in the
history of the subject, before the mode of action of alexin and its
dependence on immune body were understood, that it was derived
from the leucocytes, either by a process of secretion or disintegra-
tion. This was first suggested by Hankin, who prepared a
substance having bacteriolytic powers from extracts of the
lymphatic glands and spleen ; this, however, was probably not a
true alexin. Experimental evidence of a more convincing kind
soon followed ; thus Denys, Van de Velde, and others, caused the
production of aseptic exudates rich in leucocytes by injecting
various irritants into the pleural cavity of animals, and found that
the more leucocytes present the greater the bactericidal action of
the fluid. Buchner performed similar experiments, using aleuron
emulsions to produce the exudate, and found that the richly
cellular material had a greater bactericidal effect than blood or
serum ; in his experiments he avoided the possibility of phago-
cytosis (which probably explained some of the results of other
investigators) by freezing the fluids before determining their
potency. A large number of other researches were made about
this time, and all pointed in the same direction, and in the early
nineties it was fairly generally held that the alexins were given off
in some way or another by the leucocytes. This view was strongly
held by Metchnikoff, whose microcytase must be regarded as
identical with the alexin or complement which acts on the
bacteria; and this microcytase he holds to be produced by the
polynuclear leucocytes under certain circumstances, which will be
discussed subsequently.

With the discovery of the compound nature of the bactericidal
substances the question entered on a new phase, and new methods
of investigation were seen to be necessary. Since then very

12 — 2

l8o ORIGIN OF COMPLEMENT

many researches have appeared, the point at issue being whether
the alexin or complement is formed from the polynuclear leuco-
cytes (for there is no evidence pointing to the lymphocytes).
Some of the most important of these must be briefly summarized.

1. Various observers have investigated the relation between
the number of leucocytes in a given fluid and the amount of
complementary action. Thus Bulloch, working with hsemo-
complement, found its amount closely proportionate to the number
of polynuclear leucocytes. His figures are very convincing, and
the only alternative explanation of his results is that the injection
of any substance which stimulates leucocytosis stimulates at the
same time the hypothetical complement-producing organ or tissue.
Similar observations have been made under natural conditions
by Longcope, who found that in disease in man the occurrence
of hyperleucocytosis is accompanied by a rise in the amount of
complement. On the other hand, Guseff's investigations on
similar lines were entirely negative, and he was unable to trace
any parallelism between the two ; and Briscoe, dealing with
peritoneal fluids, arrived at a similar result. These results may
be taken as typical of the whole. A series of experiments by an
able worker is apparently conclusive in one direction, and appears
to establish the point beyond controversy, but is immediately met
with another on similar lines, pointing to a diametrically opposite
result.

Other investigations on similar lines are those of Bordet, who
found that the oedema fluid poured out in passive congestion in
an immunized animal — a fluid poor in cells — contains immune
body, but no complement, and that the aqueous humour is entirely
devoid of both substances. This latter fact has been corroborated
by Levaditi, who adds that when, after tapping, the aqueous
humour becomes rich in leucocytes, it acquires alexin at the same
time.

2. Many observers have attempted to prepare complement
from emulsions of polynuclear leucocytes, from bone-marrow,
or from lymphoid organs. The results have been diverse in the
extreme, and this diversity appears due in part to the fact that
substances other than true alexins or complement may cause,
or play some part in the causation of, both haemolysis and
bacteriolysis. In particular, we may refer to the important role
of lecithin and its allies in haemolysis, and possibly in the destruc-
tion of bacteria also. It is quite possible, for instance, that the

BACTERIOLYSIS AND ALLIED PHENOMENA l8l

haemolytic complement prepared by Levaditi from autolyzed
lymphoid tissue was one of these lipoid substances ; it was
soluble in alcohol and thermostable. The substance prepared by
Morgenroth and Korschun from extracts of various organs was
probably similar in nature, since they showed it to be thermostable,
and not to give rise to the production of antibodies on injection.
It may be at once admitted, as pointed out by Conradi, that
autolytic changes occurring in various organs give rise to the
production of bactericidal substances ; these may, perhaps, be
organic acids or other simple substances, but they do not help us
in the inquiry as to whether complement (having the power to
dissolve sensitized bacteria or red corpuscles) is formed from
the leucocytes.

The most important researches on this point are those of
Schattenfroh and Petrie, and they are mutually contradictory.
Schattenfroh washed the leucocytes thoroughly in saline solution,
added it to the heated serum, and obtained a bactericidal effect.
His results were corroborated by Lastchenko. Petrie's researches
were carried out with great care, and he was especially particular
to remove every trace of complement adhering to the leucocytes
by repeated washing in normal saline solution. (Ascher had
previously shown that some trace might remain when the process
had only been carried out three times.) The leucocytes were
obtained from aleuron exudates, and after washing they were
frozen to the temperature of liquid air, and ground to an im-
palpable powder, the method used being that of Macfadyen for
the preparation of endotoxin. His results were entirely negative,
and he failed entirely to reactivate heated bactericidal serum by
any of his extracts. Lambotte failed equally, working on lines
more like those of Schattenfroh.

Petrie's experiments seem conclusive on the point which he
investigated, and it seems quite certain that polynuclear leucocytes
do not contain alexin or complement as such ; they do not, how-
ever, negative the possibility that they may contain a precursor of
this substance, which may either be secreted or set free during
the natural solution of the leucocyte, in either case being converted
into an active form during the process. Both these views have
been widely held, and we shall discuss them later.

3. In the process of coagulation of the blood large numbers of
the leucocytes, and especially the polynuclears, are disintegrated
and partially dissolved, and these disintegrated leucocytes are

l82 ORIGIN OF COMPLEMENT

generally accepted as the source of the fibrin ferment. If the
complements are also set free during the solution of the leucocytes,
there should be far more present in the serum than in the circu-
lating plasma ; it might happen, indeed, that the plasma might be
devoid of any complementary action, though this does not neces-
sarily follow.

The difficulty of investigating this subject is very great, and
the evidence is mostly indirect. As far as it goes, it seems to
point to the fact that the plasma is alexin-free, or, at any rate,
poorer than the serum. The main direct researches into this
subject are those of Gengou and Falloise, and here, again, they
point in a diametrically opposite direction. Gengou (whose results
have been accepted in their entirety by Metchnikoff) worked with
plasma obtained by preventing coagulation by collecting the blood
in paraffined tubes and centrifugalizing forthwith — a method cer-
tainly less open to objections than any dependent on the addition
of anticoagulants, but difficult in execution. He found that plasma
thus prepared was devoid of bactericidal action, though the serum
from the same animal possessed it. In some cases the plasma
betrayed some action, perhaps because a partial coagulation had
taken place, but it was always less potent than the serum. These
results have been opposed by numerous investigators, using various
methods of inhibiting coagulation. Thus, Falloise used intra-
venous injections of peptone, the addition of sodium oxalate or
fluoride ; Petterson used oxalate and citrate, and Hahn histon,
and in all cases no appreciable difference in alexin content between
the plasma and serum was found. But this is hardly a fair test.
Peptone is generally believed to be a leucotoxic or leucolytic
agent ; and in the case of the citrated or oxalated blood it is
obvious that (if leucocytes are the origin of fibrin ferment) some
leucolysis has occurred, since the cell-free fluid coagulates on the
addition of calcium. Falloise also worked with paraffin tubes on
Gengou's method, and found that the plasma contained as much
haemolytic complement as the serum. Lastly, both he and
Lambotte worked with plasma obtained by centrifugalizing (or
by allowing sedimentation to occur in) the blood in a vein isolated
between two ligatures and kept at a low temperature, and with
the same results. Delezenne, however, criticizes these results by
pointing out that plasma obtained in this way coagulates instantly
in glass tubes at the ordinary temperature, a phenomenon in-
dicating that fibrin ferment is present, and that leucocytes have

BACTERIOLYSIS AND ALLIED PHENOMENA 183

become disintegrated, or have at least performed an act of secretion
which they do not perform in the living blood. The conditions,
therefore, are not altogether natural.

We are forced to rely to a large extent on indirect evidence, and
this, as far as it goes, is strongly in favour of Gengou's view.

4. Ainley Walker's experiments dealt with the amount of com-
plement present in the successive amounts of serum squeezed out
from a clot, and showed that this gradually rises, the first few drops
containing but little, and the quantity gradually increasing until it
attains its maximum in about twenty-four hours. Other explana-
tions are possible, but it is at least probable that the cause of this
increase is the destruction of the leucocytes which is known to
occur during the process of coagulation, and that the formation of
complement is quite analogous with that of fibrin ferment. The
serum first formed is weak in complement, the destruction of the
leucocytes having only just commenced ; and it seems a fair
deduction that, if we could examine the plasma when no leucolysis
has occurred, we should find that it contains much less, or none
at all.

This view is supported by Levaditi in a series of researches on
the fate of cholera vibrios injected into the circulatory system of
immunized guinea-pigs, care being taken to avoid as far as possible
the destruction of leucocytes. Under these circumstances he
finds that the Pfeiffer phenomenon does not take place, or does so
much more slowly ; the vibrios circulate in the blood for half an
hour or more without showing the least trace of granular trans-
formation, and are rapidly taken up by the leucocytes. In a
similar series of experiments on haemolysis he was able to prove
the existence of sensitized red corpuscles in the circulation,
although the animal did not present the least trace of haemo-
globinuria, and was of opinion that this was due to the absence of
free alexin in the plasrrr. The anaemia due to the injection of
haemolytic amboceptor he believes to be brought about by the
destruction of the red corpuscles within the macrophages of the
spleen.

As a third item of indirect evidence we may quote the experi-
ments of Lubarsch and others, to the effect that an animal may be
killed by the intravenous injection of a smaller number of bacteria
(e.g., of anthrax) than are destroyed by a small quantity (i c.c. or
less) of serum. The simplest explanation is that the blood contains
an abundance of immune body, and the failure of the defensive

184 ORIGIN OF COMPLEMENT AND IMMUNE BODY

mechanism is due to lack of complement. It would, however, be
unwise to attach too much importance to this proof.

The difficulties in coming to a conclusion on this subject are
great, but on the whole it would appear that the evidence is some-
what in favour of the views of Metchnikoff, and that in all proba-
bility there is no free alexin, or but little, in the plasma, and that
the main source of this substance is the polynuclear leucocyte,
but we cannot regard this as definitely proved. This being so,
the discussion as to whether the formation of alexin is a vital
secretory process or a phenomenon of cell death and solution
cannot be regarded as of great importance. Here again the
experimental work is most inconclusive, Lastchenko holding that
living leucocytes give off alexin when suspended in heated serum,
whilst Lazar found that it is only set free when some of the
leucocytes are destroyed. Kanthack's experiments may perhaps
be quoted at this point. He found that anthrax bacilli immersed
in frog's lymph become surrounded by eosinophile cells. After a
time these cells discharge their granules, and the bacilli soon
begin to show signs of injury, becoming less refractile and losing
their sharp-cut outline. After this these leucocytes move away
from the bacilli, from which we might argue that they are un-
injured, and that the solution of their granules is a vital action,
and exactly equivalent to the secretion of pepsin by the gastric
cells. This, of course, does not exclude the possibility that alexin
might also be liberated during the process of phagolysis, as
Metchnikoff maintains.

As regards the origin of the immune body the evidence is unani-
mous, showing that it originates from the lymphoid tissues, and
probably from the mononuclear leucocytes. It is found that if
animals be injected with cholera vibrios, killed at varying intervals
afterwards, and extracts made of the different organs, the first sites
in which the antibodies make their appearance are the lymphoid
organs, especially the bone-marrow, spleen, and lymphatic glands.
Similar facts have been observed for other bacteria and for red
blood-corpuscles (Pfeiffer and Marx, Deutsch, Wassermann). And
in Bulloch's experiments the amount of haemolytic amboceptor was
found to run roughly parallel with the number of mononuclear
leucocytes present in the blood. In all probability this increase of
circulating mononuclears must be regarded in such cases as an
indication of the activity of the lymphoid organs, which are to be
regarded as the main source of the protective antibodies. This is
a generalization of the highest importance in immunity, and it is

BACTERIOLYSIS AND ALLIED PHENOMENA 185

fortunate that the evidence in its favour is clear and direct ; and it
may be added that the proof is strengthened by the demonstration
of the fact that the agglutinins are formed in the same region. It
seems clear that the main, if not the only, function of the lympho-
cyte is the elaboration of the defensive antibodies.

There is no doubt that immune body, like the other antibodies,
circulates as such in the plasma. The most striking proof is that
of Metchnikoff in regard to spermotoxin, but other evidence is
available, and the point is not disputed.

Methods of Researches on Immune Bodies and
Complements.

The methods employed in the investigation of the haemolytic
sera are fairly simple in theory, though, as a rule, somewhat
tedious in practice, owing to the necessity, in most cases, for
quantitative work (so that slight degrees of haemolysis may not be
overlooked), and for numerous controls. The animal used in the
preparation of the immune serum will naturally depend on the
corpuscles to be employed, but in most cases the rabbit is most
convenient. The corpuscles used for immunization should be
collected in normal saline solution containing about i per cent,
sodium citrate, centrifugalized, and rewashed three or four times
in normal saline solution, so as to remove every trace of serum,
complements, etc. If this is not done the resulting serum may be
of great complexity, and results obtained by its action misleading.
The emulsion used for the injection may contain about half its
volume of corpuscles, and some 10 c.c. (or more in the case of
large rabbits) may be injected into the peritoneum. In most
cases three such injections at weekly intervals will cause the pro-
duction of a powerful haemolytic serum, but if there is any doubt
5 c.c. or so of blood may be withdrawn from the marginal vein of
the ear, and used to test the progress of the immunization. The
ear is well rubbed to make it hyperaemic, shaved, and washed with
alcohol. A small puncture is now made with a flat surgical
needle or fine knife, taking care that the vessel is merely incised
and not completely divided. As a rule the blood will now flow
quickly, drop by drop, and a sufficient amount may be collected in
a sterile test-tube. If the flow is sluggish the blood may be
milked out by passing the finger gently along the vein.

When the immunization is complete the rabbit is usually killed,
and as much blood as possible is collected. The best way to do this

l86 HAEMOLYSIS — METHODS OF RESEARCH

is to administer chloroform, and insert a cannula into the carotid
artery ; an easier method is to stick a small piece of capillary
glass tubing through the wall of the heart whilst it is still beating,
the animal being, of course, under chloroform. In either case the
blood is led into a sterile tube, allowed to clot — preferably in the
incubator, as the retraction of the clot is thereby increased — and
the clear serum 'withdrawn by means of a sterile pipette, and
stored in sterile tubes or small flasks, which may be plugged with
cotton- wool or hermetically sealed. Where the serum is to be heated
to destroy complement this may be accomplished in a thermostat,
or more simply by sealing it in a narrow tube which is tied to the
bulb of a thermometer, and placed in a beaker of warm water, and
the temperature regulated by hand by the application and removal
of a spirit-lamp. By this means any desired temperature can be
maintained within very close limits, and the necessity for main-
taining several thermostats or of altering the regulator is
removed.

The emulsion of corpuscles to be employed in the actual testing
is prepared as above, and must be rewashed at least four times,
since it is found that traces of complement may remain after three
washings. The emulsion is usually made up so that it contains
5 per cent, of corpuscles. This is readily accomplished by per-
forming the last centrifugalization in a graduated tube, going on
until the volume is constant. The bulk of the corpuscles is then
read off, and, the necessary amount of normal saline solution
being added, the corpuscles thoroughly mixed in. It is advisable
that aseptic precautions should be observed throughout, but as
this is troublesome it can often be omitted. It must be remem-
bered, however, that many common bacteria produce haemolysis,
and that if the mixtures of corpuscles, serum, etc., be incubated
for long periods fallacies may arise from this cause.

The actual experiment in its simplest form is carried out as
follows : The necessary amounts of 5 per cent, emulsion of
corpuscles, heated serum, and complement are placed in a narrow
test-tube, and in most cases normal saline solution is added to
bring the whole up to a definite volume. This is now incubated
for two hours, being stirred or shaken once or twice in the mean-
time. It is now removed and placed in a vertical position in the
ice-chest for twelve hours or so and examined. If there is com-
plete haemolysis the fluid will be deeply coloured, and there will be
no sediment, or only a minute deposit of stromata. With partial
haemolysis the fluid will be less deeply-coloured, and there will be

BACTERIOLYSIS AND ALLIED PHENOMENA

i87

a deposit of undissolved corpuscles, and when there is no haemo-
lysis the fluid will be untinged.

Modifications of the process are numerous, and almost every
investigator has his own. Thus it is very convenient to make the
mixtures by means of one of Wright's pipettes. The whole is
sucked into the pipette, which is sealed and incubated. At the
end of the process the mixture may be expelled on to white filter-
paper. Any unaltered corpuscles will form a solid dark deposit in
the centre of the drop, whilst the fluid which soaks through the
paper will be tinged or colourless, according to the amount of
haemolysis which has taken place. The amount of haemolysis may
also be determined more accurately by comparison of the super-
natant fluid with a series of colour-standards previously prepared.
The mere presence or absence of the phenomenon may be readily
shown by dropping some of the fluid on to filter-paper.

The amount of immune body present may be determined by
some such process as the following : In each of a series of narrow
test-tubes (having a mark to indicate 2 c.c.) is placed i c.c. of the
emulsion of corpuscles, and then varying amounts of the heated im-
mune serum — e.g., o-ooi c.c., 0-0025 c.c., 0-005 c-c-> etc-> or more if
the serum be a weaker one. As these small amounts are not easy to
measure accurately, the serum may be diluted ten or a hundred
times with normal saline solution and suitable multiples, these
amounts taken in the case of the smaller doses. The actual
measurements are done with graduated pipettes, which can be
procured from any instrument-maker. The complementing serum
is then added : the amount necessary to dissolve i c.c. of fully-
sensitized serum should have been previously determined by a
few rough tests (we will suppose it to be 0-2 c.c.). Lastly, sufficient
normal saline is added to bring the volume of each tube up to 2 c.c.,
and the whole series treated as above. Thus —

No.

Emulsion of
Corpuscles.

Heated Immune
Serum.

Fresh
Serum.

Haemolysis.

I.

ICC.

o-ooi c.c.

O'2 C.C.

None.

2.

0-0025 c.c.

M

3-

0-005 c.c.

5 i

4-

0-0075 c.c.

Trace.

5-

O'OI C.C.

Partial.

6.

0-025 c-c-

Complete.

7-

0-015 c c-

„

8.

0-0175 c.c.

»>

9-

O'O2 C.C.

f ,

10.

0-025 c'c-

••

l88 BACTERIOLYSIS — METHODS OF RESEARCH

Here 0-0125 c>c- °f ^he immune serum contained sufficient
immune body to sensitize fully i c.c. of a 5 per cent, emulsion of
corpuscles — i.e., a given volume of serum will sensitize 1-25 of its
own volume of corpuscles.

The determination of the amount of complement is made by an
inversion of this method. Thus Gay, who has made numerous
investigations as to the amount of complement present in human
serum, proceeds as follows : The sensitizing serum is derived from
a rabbit which has been injected with ox corpuscles. This is
heated, and the amount necessary for complete sensitization of a
definite amount of ox corpuscles is determined; thus in his
experiment 0-7 c.c. saturated 7 c.c. of a 5 per cent, emulsion. A
series of tubes, each containing i c.c. of a 5 per cent, emulsion of
fully-sensitized corpuscles, is prepared, and varying doses of the
serum to be tested are added ; the amount, which is small, is pre-
pared by dilution with normal saline to such an extent that the
actual bulk added is 0*1 c.c. The subsequent treatment is as
above. Gay and Ayer find that on the average about -£$ c.c. has
to be added to bring about complete haemolysis, the limits being
i1^ and g^ c.c.

Quantitative researches on the bacteriolytic action of the serum
are very much more difficult. The actual determination of the
amount of bactericidal action is by no means easy, and the results
obtained are of very little importance, since the serum may be
very deficient in complement, and deviation may occur. The
method which has been chiefly employed is that of plating out
after the bacteria and serum have been allowed to act together at
incubator temperature for a given period. The method is briefly
as follows : The emulsion of bacteria must be of constant strength.
As a rule, it is sufficient to take a twenty-four-hour broth culture,
and to dilute it to the same degree in all experiments ; or the
same loop may be employed throughout, or some one or other of
the counting methods which have been described may be used.
The emulsions should be dilute, so that all the bacteria may
be killed. - Klien recommends i : 8,000 of a twenty-four-hour
broth culture in the case of B. typhosus. Lastly, normal saline
solution is better than broth as a diluting agent, since it diminishes
the chance of error owing to the multiplication of bacteria during
the somewhat lengthy process of preparing the dilutions.

The actual process is as follows : Measured small amounts of
the serum to be tested are placed in a series of tubes, a uniform

BACTERIOLYSIS AND ALLIED PHENOMENA 189

amount of the emulsion added, each tube made up to a definite
volume, and all incubated for one to four hours. At the end of
this period a uniform quantity is withdrawn from each, and
plates prepared either by mixing with melted agar, or gelatin
where suitable, or by smearing over ready-poured agar plates.
The amount must, of course, be the same in each case, and may
be easily withdrawn by means of one of Wright's pipettes, which
is sterilized after use by being washed out several times with
boiling water or oil at 150° C. The plates are then incubated, and
the colonies which develop after twenty-four or forty-eight hours
are enumerated, and the amount of serum which kills all or the
greatest number of bacteria is noted.

Certain controls are necessary, the main being — (a) a tube
inoculated as above, but without the addition of serum ; and
(b) a tube also containing bacterial emulsion, and also a relatively
large amount of heated serum. The main error comes in from
the reduction of the number of colonies in consequence of aggluti-
nation, but this can be discounted in some measure by comparison
with the plate prepared from control (&).

Other methods are employed, notably that of Wright, for which
the original article should be consulted. The value of the pro-
cesses is not great, since it does not tell us even the actual
bactericidal value of the circulating blood (since we do not know
the amount of complement which is available) nor the amount of
immune body. In some cases a serum containing a large amount
of the latter substance will show little or no bactericidal power in
vitro, owing to the deficiency in complement, and may require the
addition of a hundred times its volume of normal serum to be
fully complemented.

To determine the relative amount of immune body present, the
principle of the method used for the measurement of the haemolytic
amboceptor is adopted, a series of mixtures of constant amounts
of bacterial emulsion and fresh normal serum is prepared, and
varying amounts of the heated immune serum to be tested are
added, the whole made up to uniform volume, and treated as
above. Here a further control is necessary, since the fresh
normal serum may contain some immune body or be otherwise
bactericidal. One of Neisser and Wechsberg's examples of this
process has been already quoted.

The determination of the amount of bactericidal complement is
simple enough theoretically, and follows the same lines as that

IQO THE CYTOLYSINS

for the determination of the haemolytic complement. In actual
practice these procedures are all so tedious that most of the
measurements of complement have been made on the latter
variety ; the two are believed to have the same origin, and there
is no reason to think that the one does not run parallel to the
other. Gay and Ayer employ a more direct method, adding
varying amounts of the serum to be tested to a definite volume
(0*5 c.c.) of a suspension of cholera vibrios, prepared by emulsify-
ing four twenty-four-hour agar cultures in 10 c.c. of normal saline,
and subsequently adding a sufficient sensitizing dose of serum
from an immunized rabbit. The action is allowed to go on for
one and a half hours at 37° C., films prepared, stained, and
examined as to the degree of the changes undergone by the
vibrios. They found that ^-^ c.c. of normal human serum was
sufficient to cause a complete Pfeiffer's reaction in 0-5 c.c. of
cholera emulsion tested as above, whilst when y^^ c.c. was used
there were distinct changes.

The Cytolysins.

Bordet's discovery of acquired haemolytic powers, arising from
the injection of foreign red corpuscles, proved the starting-point
of a most interesting series of researches, for it was soon shown
that the phenomenon was not an isolated one, but that it might be
produced when almost any animal cell took the place of the red
corpuscles. Thus, Metchnikoff in 1899 prepared a leucotoxic serum
by the injection of the cells from the spleen of a rat (mostly
lymphocytes) into a guinea-pig. The serum of the latter agglu-
tinated and partially dissolved the leucocytes, the lymphocytes
being most affected. Besredka studied the subject, and found
that, as in the hgemolysins, two substances — one thermostable
(sensibilatrice or amboceptor) and one thermolabile (alexin or
complement) — took part in the reaction. He studied the speci-
ficity of the substance, and found it was not sharply specialized
in its action to leucocytes of the animal used for the source of the
antigen ; it would attack those of most animals, but not man. It
was toxic, 3 c.c. of serum being a lethal dose. He also prepared
an antileucotoxin.

The next cytolysin to be prepared (by Landsteiner, and inde-
pendently by Metchnikoff) was spermotoxin. This was a very
suitable subject for study, since its action could be readily

BACTERIOLYSIS AND ALLIED PHENOMENA IQI

observed, the cells on which it acted being motile ; and it must be
pointed out that these cytolysins do not cause complete solution of
the cells. A red blood-corpuscle is a remarkable object, and
macroscopic evidence of its (partial) solution is easily obtained.
It is otherwise with the cytolysins, and here refined histological
methods are often necessary for the demonstration of a solvent
action. Agglutination of a suitable suspension of the cells is,
however, invariably present, and is easily observed. Further
evidence is also obtainable by observing the action of the serum
on live animals and the disturbances in function which it produces.
In the case of the spermotoxin, the spermatozoa are rendered
immotile, and are agglutinated, but are not dissolved.

Several interesting phenomena were brought to light by a study
of spermotoxin. Thus, Moxter showed that its action is. not
sharply specific, since a spermotoxic serum is also haemolytic.
MetchnikofT thought that this non-specificity is only apparent,
since haemolytic sera are not spermotoxic ; and he succeeded in
removing the haemolytic substance from the serum by the addition
of red corpuscles, leaving the spermotoxin intact.

It may be pointed out here that similar results have been
obtained with the other cytolytic sera ; they are not sharply
specific, all being haemolytic, and some attacking several cells, as
well as those which have been used as their antigens. This
subject has been thoroughly investigated by Pearce. Some of
his results may be briefly epitomized. Haemolytic sera act, of
course, most strongly on the red corpuscles, which they lake, and
give rise to haemoglobinuria. They also produce fatty degeneration
of the renal epithelium and necrosis of the cells of the liver.
With very small doses there may be no haemoglobinuria, bile-
pigment being present in the urine, but the lesions of the liver
and kidney are also present. A serum prepared by the injection
of kidney cells, thoroughly washed, so that no blood was injected
with them, was haemolytic in vitro, but did not produce haemo-
globinuria. It caused albuminuria, with presence of casts and
granular degeneration of the liver cells. A serum similarly pre-
pared from the suprarenal glands had no action on them, but
produced granular or fatty degeneration of the kidney and liver.
An animal injected therewith showed immediate pallor of the
mucous membranes and cardiac and respiratory failure. He
found that hepatotoxins and pancreatotoxins were without specific
action, behaving simply like haemolysins.

TRICHOTOXIN, HEPATOTOXIN, NEPHROTOXIN

It is obvious that these results are readily explicable if we
assume that the red corpuscles and tissue cells have receptors in
common, but that a particular sort of receptor is most abundant in
a particular species of cell. But, according to Beebe, sera which
are much more sharply specific can be prepared if, instead of
injecting the cells themselves, we employ the nucleo-proteid pre-
pared from them ; the method had also been employed by Bierry
and Pettit in the case of the nucleo-proteids of the liver and
kidney.

Another serum which was prepared early in the history of the
subject was trichotoxin, the cytotoxin for the ciliated epithelium.
This also, as Von Dungern showed, had a haemolytic action,
though he considered that there were no red corpuscles in the
substance used for the injections.

Hepatotoxin is produced by the injection of emulsions of liver
cells or of nucleo-proteid prepared from the liver. It causes con-
gestion of the liver, fatty or granular degeneration of the proto-
plasm, and dilatation of the bile canaliculi. If the serum has
been prepared by means of nucleo-proteid, no other organ is
affected. But the effects of hepatotoxin may also be produced by
nephrotoxic and lienotoxic serum, etc.

A considerable amount of interesting work has been done on
nephrotoxin, and the questions which have arisen are far from
having been settled. It is produced in the usual way, by injection
of animals with a fine emulsion of kidney cells (well washed
to remove blood-corpuscles, etc.) from a foreign species. It
produces albuminuria (but no glycosuria, according to Bierry),
and symptoms having at least some resemblance to uraemia (coma,
etc.) are occasionally produced. These symptoms are not specific,
and are frequently caused by injections of other cytolysins (SlpHpo-
toxin, etc.), or of emulsions of foreign cells. We have already
pointed out that Beebe and others have claimed to be able to
produce a truly specific nephrotoxin by means of injections of
nucleo-proteid from the kidney.

Of more interest is the question of the possible formation of
an autonephrotoxic body, which might conceivably be produced
when part of a kidney becomes disorganized whilst in the living
body. It has been thought, for instance, that when a toxin acts
on the kidneys it produces death and subsequent solution of the
renal epithelium, and that these soluble substances, being absorbed
into the system, call forth an autonephrotoxin, which reacts on

BACTERIOLYSIS AND ALLIED PHENOMENA

the kidney, dissolving more cells, which produce more of the anti-
body, a vicious circle being thus produced. Hence a pathology
for nephritis and uraemia on quite new lines was suggested by
Ascoli and Figari and Lindeman, etc. Thus the cardiac hyper-
trophy of renal disease is supposed to be due to a spasm of the
peripheral vessels and increase of blood-pressure due to the
nephrotoxic serum ; the nervous symptoms on the supposition
that there is a neurotoxin produced concurrently with the nephro-
toxin, and spontaneous recovery by the production of an anti-auto-
nephrotoxin, a substance for the existence of which there is a
little evidence.

There is a certain amount of experimental confirmation of this
theory. Thus Lindeman treated dogs with potassium bichromate,
causing nephritis, and found that the serum of these animals
(though free from bichromate) was toxic for other dogs. Again,
Le Play and Corpechot found that the injection of renal tissue
(of the guinea-pig) into the rabbit produced important organic
lesions : great increase in volume, fibrosis of the connective
tissues, cystic dilatations of the tubules, and desquamation of the
renal epithelium. That these changes may be due to the produc-
tion of a nephrolysin appears possible from the fact that when
these injections are made in gravid animals similar appearances
may be seen in the kidneys of the foatus, suggesting that the
nephrolysins traverse the placenta (Charrin and Delaware).
Albarran and Bernard also found that renal tissue is lethal on
injection, but Pearce denies this, and holds that their animals
were killed by bacterial infection. Further, Nefedieff ligatured
one ureter (in the rabbit), and found changes similar to those
seen in chronic nephritis. His results, might, of course, have
been due to the formation of a nephrotoxin in consequence of
the disintegration of the renal cells subsequent to ligature of
the ureter ; but Albarran pointed out that, according to Nefedieff
himself, the second kidney was unaffected at a time when the
serum was nephrotoxic, as tested on other animals. Sheldon
Amos failed to reproduce Nefedieff's results; according to her,
ligature of one ureter causes death after an average period of sixty-
nine and a half days in the guinea-pig, and fifty-two days in the
rabbit. There may be lesions on the control side, but if so these
are slight, and the liver is also affected. But that these results are
due to the action of a nephrotoxic serum appears most unlikely,
from the fact that when the whole pedicle of the kidney, or the

IQ4 GASTROTOXIN

artery and vein, are ligatured, no such results follow, though the
whole substance of the kidney is absorbed. These and other
researches make it very doubtful whether the facts observed in
nephritis are explicable on the nephrotoxic theory alone, but
further information on the subject is needed.

The degree of specificity of the nephrotoxic serum is not yet
settled. According to Pearce, the lesions which it produces may
be caused by other sera. This has been confirmed by other
observers, but Woltmann, though in accordance with Pearce on
the main question, thinks that nephrotoxin does exhibit some
degree of specificity : it produces marked congestion of the
medulla and swelling of the cortex, results not seen with other
sera. Beebe also finds nephrotoxic sera produced by the injection
of nucleo-proteid prepared from the kidney cause renal lesions,
whereas other cytotoxic sera produced by the injection of other
nucleo-proteids do not.

Gastvotoxic serum is especially interesting in view of its possible
action in the production of gastric ulcer. It has been very
thoroughly studied by Bolton, and was prepared by injecting
rabbits with emulsions or extracts of guinea-pig's gastric mucous
membrane into the rabbit. The serum thus obtained was injected
into guinea-pigs, and was found to be lethal, even in small doses
(i to 5 c.c.) ; a dose of 10 c.c. usually caused death in twenty-four
hours. The lesions were confined to the stomach, and were
striking and characteristic. They consisted of patches of necrosis
extending down to the muscularis mucosae, and often surrounded
by a haemorrhagic infiltration of the surrounding tissues. After
a time this necrotic tissue disappeared, leaving an ulcer presenting
some resemblance to the ordinary acute gastric ulcer. These
appearances (necrosis, etc.) were not seen if the acidity of the
gastric juice was neutralized by alkalis. No very definite action
could be demonstrated on gastric mucous membrane in vitro, but
isolated cells exposed to the action of the serum became hyaline
in appearance, resembling shadows. Further, the serum had a
powerfully agglutinating action on gastric cells, and produced a
precipitate in clear solutions obtained by filtration through a
Berkefeld filter.

Interesting facts were discovered as regards its specificity. It
is haemolytic, but this appears to be due to the fact that it
contains haemolysin as well as gastrotoxin. This is shown as
follows : If the serum is heated it loses its power to produce the

BACTERIOLYSIS AND ALLIED PHENOMENA

characteristic necrosis of the stomach, so that its immune body
cannot be reactivated by guinea-pig alexin; but the latter body
can reactivate haemolysin prepared by immunizing rabbits with
guinea-pig's corpuscles. If the serum is placed in contact with
an emulsion of guinea-pig's mucous membrane, it becomes
innocuous, both immune bodies being absorbed; but when
saturated with red corpuscles, it loses its haemolytic power, and
retains its necrotizing properties.

Rabbits injected with emulsions of rabbit's mucous membrane
develop a gastrotoxin which acts on guinea-pigs, but not on the
rabbit itself. Similarly for guinea-pigs treated with emulsions of
mucous membrane from the same species : their serum becomes
gastrotoxic for the rabbit, not for the guinea-pig.

To account for these remarkable facts it is suggested that the
gastrotoxin has two cytophile groups — one which combines with
the gastric cells of the animal which produces it, and one which
combines with those of the other species. Thus the gastrotoxin
of the rabbit has a cytophile group, a, which has an affinity for
rabbit's gastric cells, and a second, b, which unites with those of
the guinea-pig. During the process of immunization the animal
produces an anti-immune body, which combines with the cytophile
group a, but not with b. This is readily explicable on the side-
chain theory. It follows, therefore, that the gastrotoxin is never
efficacious against the species which produces it, being always
neutralized as regards these cells by a partial anti-antibody.

A nti -intestinal serum has been prepared. It is extremely toxic,
causing gangrene of the mucous membrane and death. Less
powerful sera cause non-fatal diarrhoea.

Syncytiolysin, or placentolysin, has been obtained by injections of
emulsions of placental tissue. According to Liepman the serum
thus obtained will give a precipitate with a solution of placental
tissue, with blood from the umbilical vein, or even with that of
a gravid woman, but not that of a non-gravid woman or a man ;
hence he proposed a serum test for pregnancy. But his results,
which seemed highly improbable, have been disproved by
Weichardt, who showed that the serum thus obtained acts equally
well on placental solutions and on all human blood. The question
of the action of the placenta when injected (in a fine emulsion)
into the tissues is of some importance in connection with a possible
pathology for eclampsia and the nephritis of pregnancy. Most
authorities (though not all) find that the animal thus treated

13—2

196 SYNCYTIOTOXIN, NEUROTOXIN

develops nephritis and lesions of the liver. Now it is known that
in some cases at least fragments of the placenta break loose and
circulate for a time in the blood during pregnancy, and it is not
difficult to suppose that dissolved products of these cells are
constantly being absorbed. Hence it seems possible that some
at least of the cases of nephritis during pregnancy and of eclampsia
may be produced in this way ; and Weichardt produced symptoms
resembling those of eclampsia by macerating placental tissue with
syncytiolysin, and injecting the result into normal rabbits. Hence
it was hoped that an antitoxin for puerperal eclampsia and
nephritis might be produced by immunizing animals with placental
tissue, so as to produce a serum which would dissolve the circulat-
ing placental cells, and prevent the destruction of the cells of the
liver and kidneys. This does not seem to have been put into
practice, and there are numerous theoretical objections which
might be raised.

Prostatotoxin has been prepared by Jungano by injecting an
emulsion of the prostates of young dogs into rabbits. The serum
clumps emulsions of prostatic cells, and when injected in vivo
produces fatty and granular degeneration of the epithelial cells
of the gland and a leucocytic infiltration of the stroma; it is
apparently fairly specific, there being no obvious lesion of other
organs.

Neurotoxin has been prepared by Delezenne, Centanni, Delille,
and others, by the treatment of one animal with the brain sub-
stance from another, which is often in itself somewhat toxic, so
that the process does not always succeed. It causes a remarkable
series of phenomena indicative of profound intoxication of the
nerve centres. These usually begin with somnolence and torpor,
which come on shortly after the injection, and may last some
hours, being succeeded by convulsive crises, in which there are
tonic and clonic spasms ; there may be one such attack, or a
series, with coma between each. The temperature is lowered, and
death usually occurs in one to twenty- four hours. The histological
changes are marked, and affect the ganglion and cortical cells;
they indicate a profound degree of destruction of these structures
(neurolysis). The substance is most active when injected into
the brain direct ; when introduced into the veins it is innocuous,
but forms an anticytolysin.

Schmidt has prepared a serum which he claims to be more or less
specific for the peripheral nerves. A guinea-pig which is injected

BACTERIOLYSIS AND ALLIED PHENOMENA IQ7

with an emulsion of the sciatic nerves of frogs develops in its
serum a substance which leads, when injected into frogs, to the
rapid production of symptoms of paralysis, which may become
complete, and resemble Landry's paralysis in man. Most of the
animals die in from twelve to forty-eight hours, and their nerves
show fragmentation of the axis cylinders, multiplication of the
nuclei in the sheath of Schwann, etc. The serum is also
haemolytic for frog's corpuscles, but neither normal serum nor a
simple haemolytic serum produce these paralytic symptoms.

The suggestion has been made that sympathetic ophthalmia
might be due to a specific cytotoxin formed by the disintegration
and absorption of the iris and ciliary body in the injured eye
(Bram Pusey). There is a certain amount of experimental proof
in favour of this interesting theory. Thus Le Play and Corpechot
prepared an ophthalmotoxic serum, and found that animals
injected therewith were less resistant than normal animals to
injections of B.pyocyaneus into the anterior chamber. The subject
has been more fully investigated by Golovine, who prepared his
serum by injecting into rabbits an emulsion of the ciliary bodies
of the dog (twelve to twenty in each animal). The ophthalmo-
toxic serum thus obtained was tested by injection into the anterior
chamber. It led to the production of a slight pericorneal injection,
a fibrinous exudate into the anterior chamber, and some appear-
ances of iritis. Microscopically it was found that the ciliary
processes presented evidence of inflammation and degeneration,
being infiltrated with leucocytes containing granules of pigment.
There was also marked evidence of degeneration of the epithelium
covering these processes. When the serum was injected into the
veins the macroscopic effects were not observed, but similar
microscopic changes were noted in the epithelium.

The pigment taken up by the leucocytes was derived from the
ciliary processes, which may become almost absolutely decolourized.
Hence Golovine holds that his serum contains not only a specific
cyclotoxin, but also a pigmentolysin.

Other cytolytic sera have been prepared, but are not of much
interest. A reference may be made to thyrotoxic serum, which
has been used in the treatment of exophthalmic goitre, though
without any considerable success. Indeed, the use of cytolytic
sera has proved most disappointing in practice. An anti-epithelial
serum which was very early suggested as a cure for cancer, but
proved inefficacious, and others have been tried. There are very

198 BACTERICIDAL SERA

many problems connected with cytolytic action that are unsolved,
and there can be but little doubt that future research in this
direction will yield results of great pathological importance, both
in theory and in practice, and whether the therapeutical advance
will take the form of a potent serum or of a juster knowledge of
the inner processes of the body in disease the future will show.

THERAPEUTIC APPLICATIONS OF BACTERICIDAL SERA.

The discovery of the great therapeutic value of diphtheria anti-
toxin naturally led to attempts at antitoxin treatment of other
diseases, but it was soon found that it was impossible to prepare
a potent toxin, and therefore antitoxin, in the great majority of
cases. The discovery of Pfeiffer's phenomenon, and the sub-
sequent researches on bacteriolysis and haemolysis, with the
demonstration of the nature of substances at work, indicated that
the problem was to be solved, if at all, on other lines, and anti-
sera were made by injecting the bacteria themselves into suitable
animals. The process need not be described at length, and of
course slight modifications are necessary in different cases. In
general the early part of the treatment consists in the injection
of small doses of dead or avirulent bacteria, or in some cases
(e.g., anthrax) of a more virulent vaccine and of a protective serum
from an already immunized animal. The animal (horses, donkeys,
or goats, are usually employed) is thus immunized, and now large
doses of virulent bacteria are given in order to stimulate the
production of antibodies to as great an extent as possible. This
part of the treatment is often prolonged, and may last for a year
or more. At the end of this time the animal is bled in the
manner described above, and the serum used for protective or
curative purposes. In some cases it is standardized, the usual
method being to determine the amount which will just protect
a small animal from a lethal dose of living bacteria, or from some
multiple thereof. Thus Sclavo's serum is tested by injecting
1*6 c.c. into six rabbits, each of which receives shortly afterwards
a known dose of virulent bacilli ; if three of the animals survive,
and the rest have their lives greatly prolonged (as compared with
controls), the serum is considered to be efficacious. Antistrepto-
coccic serum may be standardized in a similar way : according to
Hewlett, 0-05 c.c. should suffice to preserve a rabbit from ten
lethal doses of living streptococci injected intravenously. In
other cases a somewhat more refined method is adopted, and the

BACTERIOLYSIS AND ALLIED PHENOMENA IQQ

amount of antibody present is estimated. In the case of anti-
typhoid serum the simplest method is to measure the degree of
agglutination, which may rise as high as i : 1,000,000. This
cannot be taken as an absolute criterion of the amount of bactericidal
substance present, but in the great majority of cases the two
antibodies are developed roughly proportionately, and the agglu-
tination may be taken as a fair guide. Of course, the bacteriolytic
potency may be worked out by the method already described,
taking care that a sufficient amount of complement is added, and
that there is no deviation. This is probably the best method, and
is sometimes employed; thus Shiga found that O'oooi c.c. of his
antidysentery serum when reactivated by 0-3 c.c. of fresh serum
would kill all the bacilli in ^^ milligramme of a one-day-old agar
culture.

The results of tests of this nature have been to show that
extremely potent sera can be obtained against typhoid bacilli,
cholera vibrios, dysentery bacilli, and perhaps streptococci ; sera
of less but still of some power against plague bacilli, anthrax
bacilli, pneumococci, the gonococcus, and the meningococcus ;
whilst the results with staphylococci and tubercle bacilli have
been to all intents and purposes negative.

The method of action of these sera is not quite settled. In
some cases there is an abundance of bactericidal immune body,
and there is no reason to doubt that, when employed as a prophy-
lactic agent, this becomes complemented in the animal body, and
causes bacteriolysis of the infecting organism. This is certainly
the case with the sera directed against typhoid fever, cholera, and
dysentery. In other sera, which are, nevertheless, of definite
protective and even curative value, this effect cannot be demon-
strated. This is the case with anti-anthrax serum. Here we
have to assume either that the substance owes its value to the
presence of opsonins or of anti-endotoxin, or possibly (in some
cases) that it may contain free toxins, or at least specific antigens,
and act as a vaccine, producing active rather than passive
immunity, as was suggested by Wright in the case of Calmette's
typhoid serum, which is prepared in a manner somewhat different
from that just described. There is some reason for thinking that
Sclavo's serum acts opsonically, and with regard to the presence
of anti-endotoxin, it may be pointed out that the prolonged course of
immunization usually employed may lead to the production of
this substance in small amounts.

200 CLINICAL FAILURE OF BACTERICIDAL SERA

These sera, which for the purpose of convenience we shall
consider together as if they were all bactericidal, are in general
protective, but not curative. Thus the clinical use of antityphoid
and anticholera serum has shown them to be quite worthless or
even dangerous ; dysentery serum is of distinct value if used
early in the attack, and some of the other sera are of some value,
and their use is discussed in the final section of this book. In
general terms, however, and comparing them with diphtheria
antitoxin, we may say that they have proved most disappointing
in practice. The reason for this failure requires some discussion.

Antibacterial sera as ordinarily used are, of course, devoid of
complement, which has usually disappeared long before use ; it
is rendered inert on keeping, and is especially susceptible to the
antiseptics commonly added as a preservative. The first sugges-
tion is that the failure of the serum is due to lack of complement :
the union of the amboceptor and bacterium is supposed to take
place as usual, but the necessary alexin is not forthcoming. This
may be due to one of two causes : in the first place, there may be
(as is known to occur in certain diseases) a deficiency in the
amount of complement in the serum; in the second place, that
which is there may be unsuitable in nature.

As regards deficiency in complement, this has been found to
occur in certain diseases, and is very probably a common occur-
rence in pathological conditions and states of malnutrition in
general ; but when we consider the comparatively small amount
necessary to activate a large dose of sensitized bacilli, there is no
reason to think that it ever falls below that level. Again, the
facts known concerning the immunity of the dog and other
animals to anthrax are of such a nature as to render it improbable
(in this case, at least) that deficiency of complement can really
be of much importance ; for dog's serum contains abundance of
amboceptor, yet no suitable complement, and is devoid of bacteri-
cidal action. We shall see reasons for believing that amboceptor
may possibly act as opsonin, in some cases at least, without the
concurrence of complement, and this is probably the explanation
of the immunity of the dog to anthrax.

The other explanation, that of Ehrlich, is that the complements
present in human serum may be unsuitable to reactivate serum
derived from a horse, ass, or goat, or other animal used as the
source of the immune body. To obviate this, he has proposed
the use of sera from several species of animals, in the hope of

BACTERIOLYSIS AND ALLIED PHENOMENA 2OI

finding one that can be reactivated by human complement, and
has suggested the use of serum from the higher apes, the com-
plements of which closely resemble those of man. These
explanations are not very satisfactory. Thus Shiga's antidysentery
serum is certainly readily complemented by human blood; and
although it has certainly some beneficial action, it is useless in
the chronic stages of the disease, and this although the amount
injected must be very much greater than is necessary to dissolve
all the bacilli present in the body.

Another possible cause of failure is the deviation of complement.
If we admit the action of the bactericidal substances — by no
means undisputed — in the natural process of recovery from
disease, we can easily see how it is that this process does not
occur under normal conditions. Thus, when infection with the
typhoid bacillus occurs, there is at first little or no amboceptor
in the blood. The small amount present is quickly seized by
the bacteria and removed, and although a few bacilli may be
killed, the great majority flourish unchecked. But amboceptor
is soon put out in gradually increasing amounts, and at first is
used up as soon as it is formed. The two processes, proliferation
of bacilli and increase in the amount of amboceptor, now progress
side by side, and on their relative rapidity depends the outcome
of the disease. At first bacillary proliferation takes place more
rapidly than the production of antibody, and the symptoms
gradually become more and more severe. After a time the
antibodies are released in larger and larger amount, and (in a
favourable case) a time arrives when there is exactly enough
for all the bacteria present. We must assume that enough
complement is available, and in this case it is easy to see how
it can never become deviated ; for all the amboceptor is rapidly
linked up to the bacilli, and does not accumulate in excess in the
blood. It does seem possible, however, that an accumulation of
amboceptor might conceivably determine a relapse, bacteria
which escaped destruction owing to their having lain in the
tissues, gall-bladder, or other inaccessible region, being now free
to grow in the blood owing to the removal of complement by
deviation.

But when a dose of bactericidal serum, containing, it may be,
many times more immune body than is necessary for the solution
of the bacilli present, is suddenly thrown into the circulation,
the conditions are quite different. Here there is an excess of

2O2 FAILURE OF BACTERICIDAL SERA

immune body relatively both to the bacteria and to the com-
plement, and deviation of the latter may occur. Hence it is at
least conceivable that a dose of bactericidal serum may be
injurious in that it actually inhibits the normal bacteriolytic
processes that are at work in the blood-stream. We have already
quoted processes exactly parallel in describing the experimental
proof of the deviation of complement.

Another suggestion that has been made is to use perfectly fresh
immune serum, or to reactivate it by fresh serum from a normal
animal. But this seems not to be successful, and apparently
alien complements rapidly unite with the tissues of the animals
into which they are injected, and so become inert.

It would seem that no explanation based on deficiency in com-
plement will be found satisfactory : the facts concerning the
action of the dog's serum on anthrax bacilli appear to offer a
crucial experiment settling this point. Nor if the added ambo-
ceptor really acts as opsonin would the question of complement
come in. The most satisfactory explanation appears to be that
the sera do not actually come in contact with the bacteria in the
lesions, though they may, and very probably do, tend to sterilize
the blood, and so prevent further generalization of the infection.
This question — of the accessibility of the bacteria in the lesions
to the substances circulating in the blood — is probably one of prime
importance in immunity and recovery, and we shall meet with it
again in dealing with the opsonins. It seems to meet the facts
of the case very well with regard to the action of the serum in
dysentery. In acute cases it is of value ; and here the bacilli are
lying in regions which are fairly accessible to the blood. In chronic
dysentery it is almost useless, and in this form of the disease the
bacilli are shielded by a dense and impermeable layer of inflam-
matory tissues. And in cholera the bacilli are mostly lying in the
intestinal tract ; probably a few do gain access to the blood and
tissues, and are immediately destroyed. In typhoid fever the
bacilli are found in the blood early in the disease, and later,
roughly at the period at which antibodies make their appearance
in large amounts, they disappear. But there are always large
numbers in the lymph glands and spleen, regions in which it is
almost certain they are shielded from the action of the blood.
This explanation appears far more satisfactory than any depending
on deficiency in complement.

If the bacteria in the blood-stream are actually dissolved by the

BACTERIOLYSIS AND ALLIED PHENOMENA 203

added bactericidal substances, a new danger is involved — that of
the liberation of a large amount of endotoxin. This substance
we believe not to be liberated when bacteria are dissolved within
the leucocytes, but to be set free when extracellular solution takes
place. The essential fever of the early stages of typhoid fever
is very probably due to the endotoxin set free by the solution
of the bacilli in the circulating blood, and any sudden addition to
this amount occurring before the tissues have become immunized
or trained to produce anti-endotoxin may be fraught with danger.
There can be little doubt that if sero-therapy has any future
triumphs in store, they will be in the direction of the production
of anti-endotoxins.

The various antibacterial sera in common use are considered in
the last section.

CHAPTER VIII
THE AGGLUTININS

AN exceedingly interesting and important group of antibodies,
which were discovered by Gruber and Durham in 1896 (though
their effect had been observed by Charrin and Roger in 1889
in the case of B. pyocyaneus),1 are called the agglutinins, since
they have the power of agglutinating their antigens, or causing
them to adhere in masses. Their effect is best seen after the
addition of the serum of a patient convalescent from typhoid fever
(or of an animal which has been injected with typhoid bacilli)
to a living culture of the organism. The bacilli, which at first are
actively motile and are distributed uniformly throughout the fluid,
first lose their motility, and then individuals may be seen to move
nearer and nearer to one another, until they come into close contact.
It often happens, especially in weakly agglutinating sera, that this
approach of two bacilli may be seen to occur before their paralysis
has taken place. They then revolve rapidly round a common
axis, giving the observer the impression that they are united
together by a sort of invisible link, which they struggle to break.
This process continues, and fresh individuals are attracted to the
groups, until at last all the bacilli, instead of being scattered equally
throughout the fluid, are collected into masses, the intervening
fluid being free. The process may also be watched with the naked
eye, and the emulsion, which is at first uniformly turbid, will be
seen to lose its homogeneity, and take on a finely granular
appearance. This at first can only be realized by comparison
with a control specimen to which no serum has been added, but
in a little time it will be obvious that flocculi of bacilli are being
formed, and that between these flocculi the fluid is clearing. Soon

1 The effect had also been observed by Metchnikoff in the case of V. Metchni-
kovi in 1891 ; he was inclined to regard it as a general phenomenon, but failed
to find it in another case. Similar appearances had also been seen by Issaeff
in 1893.

204

THE AGGLUTININS 205

(if the emulsion is thick enough) all the organisms will be found to
have collected into a single mass or a few masses, the rest of the
fluid being quite clear. Finally, these masses will sink to the
bottom of the vessel, and it will be noted that if the bacilli in
the control specimen also sink (as happens with killed organisms),
the masses will be much more voluminous than the deposit of
unagglutinated bacteria. A microscopic examination of the de-
posit in the two cases will show why this is. In the deposit of
dead bacilli the separate rods have sunk down slowly, and have
packed themselves closely together, and will be found, to a very
large extent, to lie horizontally side by side. In the agglutinated
mass the bacilli point indifferently in all directions, and the explana-
tion suggests itself that they have been drawn forcibly together by
a centripetal force, and have not had time to adapt themselves so
as to take up as little room as possible. A result of this is that it
is easy to distinguish between a specimen that has agglutinated
and one in which the bacilli have simply settled, even although the
actual occurrence of the phenomenon has not been witnessed.

The reaction is given with the serum of immunized animals, and
is a general one. It is given with nearly all species of bacteria,
though to a very different extent in different cases, with red blood-
corpuscles, leucocytes, and with cells of all kinds. The occurrence
of motility is not necessary for it, and dead bacilli will clump
almost or quite as well as living ones. The reaction is, in general,
specific, and a serum which is strongly agglutinating as regards
one species of organism may be entirely devoid of action on others.
Hence it was proposed by Gruber and Durham as a test for the
identification of bacteria, and is of great value. Thus, when a
bacteriologist has isolated a culture of an organism resembling
B. typhosus from a patient suspected of having typhoid fever, or
from a sample of water supposed to be contaminated, the first step
in the identification is made by observing whether it is clumped by
a serum known to have agglutinating powers over typhoid bacilli
and not over others. Other tests are necessary for its complete
identification, but these are slower, and for some purposes un-
necessary. The clinical diagnosis of cholera by means of cultures
from the stools is carried out in the same way, and sera adapted
for either purpose can be obtained commercially.

The reaction, however, is not an absolutely specific one, and it
is found that a given immune serum may clump not only the
culture used in its production, but also those of closely allied

2C>6 SPECIFICITY OF AGGLUTININS

species. Thus typhoid serum clumps B. coli, the paratyphoid and
paracolon bacilli, the B. psittacosis, and others. This is called a
group reaction, and is of profound interest in classification. It is
not, as might be thought, a hindrance to the practical application
of the process as a method of identification of the nature of a
culture, since it is found that the action is exerted much more
strongly on the organism used for the immunization than on
others. This is determined by ascertaining the dilution necessary
to bring about agglutination in a certain time at a given tempera-
ture. For example, we may find that certain specimens of anti-
typhoid serum will agglutinate typhoid bacilli at a dilution of
i : 10,000 in one hour, whilst B. coli is not affected if the dilution
is greater than i : 50. In the practical use of this serum we
should not be certain that a given culture was one of B. typhosus
unless it reacted at i : 1,000 or more.

The explanation of these group reactions on Ehrlich's theory
offers no difficulties. Agglutinin is, as will be shown, a specific
antibody to the molecules of protoplasm contained in the bodies of
the injected cells. In each cell these will be of many varieties,
and to each a specific antibody will be produced. We must
imagine a typhoid bacillus as containing a large number of one
particular sort of molecule, a smaller one of another, whilst in the
colon bacillus these relations will be the reverse. A typhoid serum,
therefore, will contain much agglutinin which acts on the typhoid
molecules, and a little which acts on a few of those present in
B. coli ; it will agglutinate the former strongly, the latter feebly.
But the colon serum will contain antibodies to a few only of the
molecules present in the typhoid bacilli, and will clump it only in
strong dilutions.1

Agglutinins are formed, as we have seen, as the result of the

1 There are a few noteworthy exceptions which have been recorded to these
general rules. In a few cases of tuberculosis the power of agglutinating
B. typhosus has been seen to rise, and Park has quoted a case in which an
animal immunized against staphylococci increased ^its power against the same
bacillus from i : 10 to i : 160. In interpreting these results we must always
wonder whether they might not be explained by a rise in the sensitiveness of
the culture used. But this objection does not apply to the observations of
Posselt and Sagasser, who obtained an agglutinin which acted on bacteria
other than those used for the injection, and which was not removed from the
serum by these bacteria. And some cases have been recorded in which a
serum had less action on its own antigen than on others. All these exceptions
are rare and not full)' investigated, and do not affect the general law.

THE AGGLUTININS 2O7

injections of their specific antigens. They are also frequently
present apart from any interference. For example, normal human
serum clumps the second vaccine of anthrax powerfully, and in
most cases has a feeble action on both B. typhosus and B. coli.
Horse serum is very rich in agglutinins, clumping typhoid and
coli bacilli, the B. pyocyarmts, and the cholera vibrio, often in
dilutions as high as i : 100. In most cases agglutinins are present
in small amounts in the serum of infants and young children, and
become more abundant in later life. This suggests that they may
be formed — in part, at least — by a process of auto-inoculation with
bacteria, principally, perhaps, from the intestine. We have
already seen, however, that on Ehrlich's theory the presence of
antibodies in normal animals is readily explicable without such
assumption.

The injections of bacteria or cells of any sort leads to the pro-
duction both of agglutinins and of cytolysins, and in most cases of
haemolysis or bacteriolysis agglutination occurs as the first step in
the process. The question arises, therefore, whether they are the
same substance. It is easy to show that they are not, since sera
which contains agglutinin do not necessarily contain immune body,
or vice versa. In sera obtained by artificial immunization, of course,
the two are almost invariably formed side by side, and it is only
by special processes that we can obtain the one without the other.
Thus Frouin claims that if dried dog's corpuscles are washed with
acetone and injected into a rabbit, they cause the production of
agglutinin ; but no haemolysin. The residue from the evapora-
tion of the acetone, on the other hand, yields haemolysin, but no
agglutinin. But in sera from normal animals it is quite common
to find the one without the other. Thus the serum of healthy
human beings frequently clumps normal human corpuscles, but
haemolysis is extremely rare. The converse process — haemolysis
without agglutination — also occurs; and with regard to antibacterial
sera of artificial origin, Frankel and Otto found that when a dog
was fed on typhoid cultures it developed agglutinin, but no immune
body. Lastly, in many cases the action of agglutinin is destroyed
at a lower temperature than that of immune body, although both
substances are in a marked degree thermostable. We shall have
to discuss the effects of heat on agglutinin more fully subsequently.

There is, as a matter of fact, a kind of antagonism between
agglutination and cytolysis. Cells which are crowded firmly
together are naturally shielded, more or less, from the solvent

208 AGGLUTININS IN IMMUNITY

action of the fluid in which they are suspended ; and equally
naturally cells which are dissolved do not show ordinary agglu-
tination, though, as we shall see, they show a similar phenomenon.

The formation of agglutinins follows laws similar to those
governing the formation of other antibodies. After each injec-
tion there is a negative phase, followed by a rise, which, as a rule,
attains its maximum in about a week. In the case of typhoid
fever no agglutinin can be demonstrated, as a rule, during the
first week ; there is then a steady rise, which usually attains its
maximum at the commencement of convalescence. After this
the amount tends gradually downward, and disappears after a
time, which varies between a few months and several years.
On a single occasion the author has seen a marked drop in the
amount precede a relapse, during which a second rise occurred.
This was obviously a negative phase, and the occurrence of the
relapse might have been foretold therefrom.

Bacteria which have been acted on by agglutinin are not altered
thereby in appearance, viability, or virulence, and the process does
not appear to play a part of much importance in immunity. Two
suggestions have been made in this respect : Gruber thought it
caused the outer layer of the bacillus to swell up, so that it could
be attacked by alexin, and Walker suggested that the clumping
of the bacilli might render them more easily taken up in large
numbers by the leucocytes. Possibly, also, the paralysis is the
essential feature of the process, as a reaction of immunity, since
we should expect non- motile bacteria to be more easily ingested
by phagocytes. It is interesting in this connection to notice that
the bacteria for which strong agglutinating sera are obtainable
are all highly motile (B. typhosus, coli, and pyocyaneus, vibrios).
The recent researches on the thermostable opsonins have caused
a certain amount of attention to be directed to the agglutinins
from this point of view, but nothing is definitely proved.

That agglutinin, in common with the other antibodies, unites
directly with its antigen may be shown in several ways. In one
an agglutinating serum cooled to o° is added to a culture similarly
cooled, and the mixture kept on ice. The bacteria will gradually
settle down without agglutinating, and the supernatant fluid may
be pipetted off. This may be tested in the ordinary way, and
will be found to have lost much of its agglutinating power. The
bacteria, if suspended in warm saline solution, will immediately
clump. Evidently, therefore, the agglutinin has been removed in

THE AGGLUTININS 2Og

combination with the bacteria. Further, it is clear that we may
distinguish two properties of agglutinin (that of uniting with antigen
and that of clumping), and that these are discharged at different
temperatures : the agglutinin unites at o°, and only exerts its
specific action at higher temperatures. We may express this in
Ehrlich's terminology by saying that it possesses a haptophore
group which functionates at o°, and an ergophore group which
only acts in the warm.

Another proof is as follows : It was shown by Bordet that
agglutination only takes place when certain salts are present. Of
these sodium chloride appears to be the most generally efficient,
but Crendiropoulo and Amos have shown the calcium chloride
has a special adjuvant action in the agglutination of cholera
vibrios. To this subject we shall return. The proof of the union
between bacteria and their agglutinins is made as follows :
Bacteria are added to clumping serum, and the precipitate collected
and washed and shaken in a large quantity of distilled water.
No agglutination occurs until salt is added, when it takes place
rapidly, according to the thickness of the emulsion. In this case
also the two substances must have entered into the combination.

The substance with which agglutinin combines — i.e., that
which calls forth its production in the living animal — is evidently
not a toxin, since an agglutinating serum has, as such, no protec-
tive action. We know some of its characters. It is formed, of
course, in the bodies of the bacteria, and in young cultures is
entirely intracellular. In older cultures, however, it diffuses out,
being probably set free by a process of autolysis, and passes into
solution. This is especially the case in broth cultures, and this is
one of the reasons why, if liquid cultures are used in agglutination
tests, they must be young ; in agar cultures there is less diffusion
of the agglutinable substance, and the need is not so great. Its
presence may be proved in two ways : In the first place, this
filtrate, if injected into animals, will bring about the production
of agglutinin, as we should expect. In the second place, this
fluid, when added to a powerful clumping serum, will cause a
precipitate. This is Kraus's reaction, and it is a most interesting
phenomenon. It is best seen when the fluid portion of broth
culture of B. typhosns or V. cholera (at least a month old and
filtered through a Berkefeld filter to remove all solid particles) is
added in various proportions to a strong immune serum. Under
such circumstances the fluid will gradually become opalescent, or

210 AGGLUTINOID

even opaque, then granular, and finally flocculent. It presents a
most extraordinary resemblance to the clumping of an ordinary
culture, but a microscopic examination will show the flocculi
consist of amorphous granules instead of bacteria. It has been
suggested that it is due to a clumping of cilia which have passed
through the filter (Nicolle), but the phenomenon has since been
observed in the case of the pneumococcus (Panichi) and other
non-flagellated organisms. The agglutinable substance is thermo-
stable. It does not appear to be given off in all cases, and some-
times all attempts to get Kraus's reaction are unsuccessful.

This substance is the antigen of agglutinin, and our nomen-
clature would be more uniform if we were to call it agglutin and
its antibody anti-agglutin, but the terms are too firmly fixed to be
altered. We shall call it agglutinable substance, or agglutinogen.

The fact that heated serum still agglutinates shows that alexin
or complement plays no part in the process, but we have already
explained how we know that the molecule of agglutinin possesses
an ergophore or zymophore group. This group, as is the case
with the corresponding groups of the toxins and complements, is
less resistant than is the haptophore group, and is destroyed at
70° to 75° C. The substance left is called agglutinoid, and is
analogous to toxoid and complementoid. Its existence is demon-
strated thus : Heated serum (or serum which has been kept for a
long time) is added to a culture of bacteria. No agglutination
takes place. The bacteria are then centrifugalized off and placed in
a strongly agglutinating serum, but are found not to clump. It
is evident, therefore, that the bacteria have their receptors
occupied by some substance which prevents the union of the
agglutinin. The agglutinoid has combined with the agglutinogen,
and excludes the unaltered agglutinin.

In some cases at least agglutinoids, which have a stronger
affinity for bacteria than has normal agglutinin, may be present.
In this case, if bacteria be added to a mixture of the two sub-
stances, no agglutination occurs. The pro-agglutinoids (as they are
termed, the expression being taken from the prototoxoids) seize
on the agglutinable substance in the bacteria before the
agglutinin can do so. If to this mixture more bacteria be added,
more pro-agglutinoid will be taken up, until it is all exhausted,
and then any fresh bacteria that are added will be clumped.
This is one explanation of a phenomenon which is fairly frequently
observed (if looked for) in the clinical diagnosis of typhoid fever,

THE AGGLUTININS 211

and is probably a source of error often overlooked : the serum
clumps at a high dilution, and not at a low one. The author has
observed it three times in the last four years. Another explana-
tion, which is probably more often the true one, is that in the low
dilutions partial bacteriolysis takes place, and the partly dissolved
bacteria do not clump. The reason for this conclusion is that the
clumping may occur in low dilutions in the cold, when bacterio-
lysis does not take place. Yet other explanations have been
given.

Certain non-specific substances may bring about clumping
which has a close superficial resemblance to that caused by
agglutinin. This was first showed by Malvoz in the case of the
action of chrysoidin on V. cholera. He also showed that certain
stains, such as fuchsin, vesuvin, and safranin, and some anti-
septics, such as formalin (in fairly large amounts), corrosive
sublimate, and peroxide of hydrogen, have this action. Mineral
acids also possess this property, and also certain salts. In the
case of cholera vibrios, Ruffer and Crendiropoulo found calcium
chloride to have a powerful action, sodium phosphate to have a
very slight one. This must not be confused with the effect of
salts in favouring the action of agglutinating serum.

We are now in a position to discuss the mechanism of the
process. Numerous theories have been propounded. Thus
Gruber thought that the external membrane of the bacterium
became " sticky," so that organisms once brought into contact
remained adherent. But no visible alteration of the organisms
or red corpuscles can be seen. Further, it would not account for
the approach of two non-motile cells, which certainly appears to
take place in clumping, and would not explain why the cells or
bacteria were brought into contact in the first instance. Nicolle
propounded a similar theory. He, however, showed that when
inert and insoluble particles, such as of talc, were suspended in old
filtered cultures of typhoid bacilli (Kraus's fluid), and serum
added, they appeared to clump just as typhoid bacilli did, and it
is difficult to reconcile this with his theory. Dineur thought that
the flagella of the bacilli might have an adhesive material
deposited on them ; but many non -flagellated bacteria clump, to
say nothing of red corpuscles. Others have thought that Kraus's
reaction is the fundamental phenomenon, and that the bacteria,
etc., are entangled in it . like the particles of talc in Nicolle's
experiment. But no obvious precipitate can be seen in stained

14—2

212 MECHANISM OF AGGLUTINATION

films of clumped bacteria, whereas Kraus's precipitate is easily
demonstrated ; besides which, agglutination can be perfectly
easily demonstrated in young cultures (the fluid portion of which
will not precipitate with specific serum) or with carefully washed
bacteria. This explanation, though ingenious, may be disre-
garded, r

Bordet's view is undoubtedly the correct one.( It explains agglu-
tination as being due to a change in the molecular relations
between the objects and the fluids which bathe it — in other words,
it is practically an effect of surface tension^ It takes place in
many cases other than those in which it is produced by specific
sera acting on bacteria, red blood-corpuscles, etc. : thus, these
objects can be made to clump by the action of many aniline stains,
acids, antiseptics, etc. An emulsion of clay in distilled water will
remain turbid for a long time, but will rapidly clear, owing to the
formation of aggregates of particles, when salt is added. This
phenomenon (which explains the formation of mudbanks at the
mouths of rivers, where admixture of fresh and salt water occurs)
is of especial interest in view of the necessity for the presence of
salts in specific agglutination. Many bacteria, especially the
tubercle bacillus, clump spontaneously without the addition of
serum. In some cases this can be avoided by using a fluid poor in
or free from salt to make the dilution, as in Sir Almroth Wright's
method of estimating the agglutinating power of the serum on the
tubercle bacillus. A process fundamentally similar can be seen if
wooden matches smeared with grease are thrown on to the surface
of water, and may also be seen in the gathering together of
bubbles on the top of any fluid.

Two phenomena are involved : the approach of the particles the
one to the other, and their adhesion subsequently. The former
depends on certain physical laws investigated by Korn and others,
and not yet fully elaborated, in virtue of which two elastic
particles suspended in an inelastic fluid in which vibrations are
taking place tend to approach one another. It is probably fair to
assume that these conditions are always present in the case of
bacteria suspended in a fluid medium, and that, even in the
absence of any agglutinin, the individual organisms will tend to
approach one another and to form aggregates. But in the case
of most organisms the aggregates thus formed are quite instable,
breaking up when the slightest shaking of the fluid takes place.
Here the force of surface tension is all-important. It is a force

THE AGGLUTTNINS 213

which is generated wherever a fluid comes into contact with any
other substance, whether solid, liquid, or gas, and which acts
exactly as if the surface of the fluid in question were in a state of
tension, like a stretched film of indiarubber. If a relatively
small amount of any fluid be suspended in another fluid of the
same specific gravity with which it does not mix, it will assume
the form of a sphere : this is because the sphere has a smaller
surface for a given volume than any other solid body, and the
hypothetical film on the surface continually contracts until this
figure is assumed. Hence leucocytes, and most other free cells
consisting of fluid or semi-fluid protoplasm, tend to assume a
spherical form when in a resting condition ; hence also, of course,
the spherical form of soap-bubbles, oil-globules, etc. Now consider
the case of two spheres acted on by surface tension and just
touching one another ; for example, take two drops of oil
suspended in a fluid of about the same specific gravity. If we
regard the surface of the two spheres as continuous, it is obvious
that it is much larger than it would be if the two drops coalesced
to form a single sphere. (It is roughly larger in the proportion of
4:3.) The film, therefore, will contract until the two globules are
drawn into a single drop, with double the volume of each original
globule, but with a much smaller superficies than that of the two
separately. This process will take place whenever two bodies,
neither or both of which are wetted by the fluid, are brought in
contact or very close together : when one is wetted and the other
not, they tend to repel one another. The force of surface tension
only extends for an exceedingly minute distance into the fluid
from the surface, and therefore does not draw the substances
together if they are a finite distance apart. Its action comes into
play when the two bodies touch one another in one point, so that
the surfaces between the two bodies and the fluid join to become
one at this point. Thus, if two red blood-corpuscles touch one
another obliquely at one point, they become drawn together, and
slide the one on the other until they oppose as small a surface as
possible to the surrounding fluid. This, of course, is when the one
lies flat on the other, as in rouleaux formation. Two wooden
discs enclosed in a small indiarubber bag would act precisely
similarly.

The exact way in which the agglutinin affects the surface
tension between the bacteria and the fluid in which they lie is not
quite clear, and raises difficult questions in molecular physics, some

214 MECHANISM OF AGGLUTINATION

of which are glanced at in our section on Colloidal Chemistry. It
is intimately concerned with the subject of solubility. If a body
is soluble in a fluid — i.e., if the molecules of the latter have a
greater affinity for those of the former than these have for one
another — there will be no sharp line of demarcation between the two :
between the solid and the liquid there will be a zone in which mole-
cules of both substances are present, and this will shade gradually
off into the solid body on the one hand and the fluid on the other.
Here, then, there will be practically no surface between the two,
and surface tension will be small or absent ; ..and in a general way
substances present in a fluid which dissolves them have no
tendency to clump. Thus to prepare an emulsion of an oil, a
solution of a soap or of an alkali is used, and the emulsions thus
formed are comparatively stable ; but if the fluid be made acid,
the surface tension is increased, and the globules quickly run
together or clump. Now it is clear from the fact that the fluid
part of bacterial emulsions will give Kraus's reaction, and will
lead to the production of antibodies on injection, that a certain
amount of solution does take place. That agglutinin actually
renders the bacteria less soluble appears clear from the phenomena
of Kraus's reaction, though here the insoluble precipitate is formed
on and in the bacteria, rather than in the fluid. And the complete
absence of clumping which occurs when bacteriolysis takes place
(though there, is a large amount of agglutinin in the serum used)
is an indication of what takes place when the bacteria are rendered
more soluble, instead of less, by means of an antibody. Insolubility
does not account for the whole of the phenomena, but it is a feature
of great importance.

As regards the nature of agglutinin, all we know is that it is
precipitated with the globulins, and may be of that nature. It
does not dialyze, and is digested by trypsin, etc.

It appears to be formed in the lymphoid organs, red marrow,
and spleen, being found early in those organs after injections of
cholera vibrios (Pfeiffer and Marx). Metchnikoff found that the
peritoneal exudate might be richer in agglutinins than the blood,
and thought they came from the cells (leucocytes and endothelial)
in that fluid. The subject has also been investigated by Van
Emden, Deutsch, and Ruffer and Crendiropoulo, who all confirm
Pfeiffer and Marx as to the early presence of these substances in
the lymphoid tissues after inoculation.

So far the study of the agglutinins has not presented much

THE AGGLUTININS 215

difficulty, but further research has shown it to be full of com-
plexities. We will glance briefly at some more recent researches
on the subject, the exact explanation and significance of which
are not ascertained beyond dispute.

Several facts go to show that the agglutinin of B. typhosus is not
a simple substance, but that two or more bodies are concerned.
(It may be mentioned that this bacillus has been studied more
than any other in this connection.) Thus Joos, after a series of
ingenious researches, came to the conclusion that the bacillus
contains two agglutinogens, and that each has its corresponding
agglutinin. The agglutinogen which is present in largest amount
(and which he calls a) is thermolabile, being destroyed at 62° C.,
leaving only agglutinogen /3, which is thermostable. An animal
injected with living cultures will contain agglutinins (a and /3)
against both the substances. The first will combine with agglu-
tinogen a only, whilst the second will combine with both substances.
The two substances differ in their thermostability : a is thermo-
stable, but /3 loses its power of agglutination at 62° C. A couple
of examples of the facts which this complicated theory was intro-
duced to explain may be given. The serum of a horse treated
with living typhoid bacilli (and therefore containing agglu-
tinins a and /3) clumped a living culture at i : 20,000, and a
heated one at i : 1,000. When the supernatant fluid of this last
dilution was tested with heated typhoid bacilli, no agglutination
took place (agglutinin /3 had been removed), whereas it would
clump living bacilli readily enough. Agglutinin a was present in
larger amount than /3, and had not all been removed at this
dilution. Again, when heated serum is added to heated bacilli
there is no agglutination, since the thermolabile agglutinin /3 is
destroyed. The agglutinin a, it is true, is not destroyed, but its
agglutinogen (which is thermolabile) is. But when living bacilli
are now added clumping occurs, since the agglutinin a can find
unaltered agglutinogen a to affect.

Smith and Reagh (and their researches have been, in the main,
corroborated by others) found that typhoid bacilli and other
flagellated bacilli might form two agglutinins — the one acting
on the agglutinogen of the bodies of the bacilli, the other on that
of the flagella. The subject has also been investigated in a some-
what similar way by Buxton and Torrey, who find also two
agglutinins — the one to a substance which remains attached to the
body of the bacillus, whilst the other can be separated from it by

2l6 THE " PRO-ZONES " IN AGGLUTINATION

a temperature of 72°, followed by filtration. The action of the two
is specific. If the filtrate be injected into animals, the serum
which results clumps ordinary typhoid bacilli well, but has little
action on those from which the separable substance has been
removed. The serum obtained by injection of the bacilli deprived
of separable substance is weaker, and has an equal action on the
bacilli whether normal or heated and deprived of soluble substance.

It is evident that the subject is a complicated one, and this
is even more clear from the researches of Dreyer and Jex- Blake
on the agglutination of B. coli by its specific serum. Investigating
first the behaviour of the bacilli when heated, they found, as other
observers had done, no alteration at 60° C., but a sudden diminu-
tion in the power of undergoing agglutination when heated to
70° C. This, of course, i<s usually ascribed to the complete or
partial destruction of the agglutinogen, though this explanation is
incomplete, since (as Eisenberg and Volk had previously found)
the bacilli which do not clump will still combine with agglutinin.

But Dreyer and Jex-Blake found that the agglutinability is
partially or completely restored by prolonged heating to 100° C.
After exposure to this temperature for a period of from two to
thirteen hours, the susceptibility of the bacilli to the serum might
be as great, or almost as great, as at first. This is an extra-
ordinary fact, and one for which no adequate explanation is
forthcoming. The only parallel is the behaviour of megatheriolysin
(a bacterial haemolysin), also investigated by Dreyer. This is
destroyed, or at least rendered inert, at 60° C., and reactivated at
the boiling-point.

These authors also adduce evidence to show that the " zones of
inhibition," or "pro- zones," described by Eisenberg and Volk as
occurring with heated serum, cannot be explained by the assump-
tion of the presence of " agglutinoids " with a high affinity for
agglutinogen. The evidence against this view is briefly this : If
it were true, the more the serum were weakened by the heat
(i.e., the greater the production of agglutinoid), the larger should
be the zone of inhibition, and vice versa. This they found not
to be the case, for a serum which had not been appreciably injured
by the heat might have a large zone of inhibition. They also
found exactly similar zones in investigating agglutination caused
by means of acids, in which case, of course, nothing of the nature
of agglutinoids could occur. Thus, in a series of experiments it
was found that when orthophosphoric acid was added to a definite

THE AGGLUTININS 217

volume of emulsion of B. coli, agglutination occurred when between
118 centigrammes and 4 centigrammes of the acid was present,
and when between i-i milligrammes and 0*001 milligramme, but
not with intermediate amounts.

Similar phenomena (i.e., the presence of zones of inhibition) can
be seen in many of the actions of coagulants on colloid emulsions
or solutions, such, for instance, as the precipitation of gum mastic
from water by ferric chloride or trisulphide of arsenic. This
case also, as Dreyer points out, shows a marked analogy with
the clumping of coli bacilli by phosphoric acid, since in each case
the zone of inhibition becomes smaller when the agglutinating sub-
stance (bacilli or particles of gum) become more numerous. This
analogy between the agglutination of bacteria and the flocculation
of colloids has been investigated by Bechhold, Neisser and
Friedemann, and Biltz, and is a subject of the utmost interest,
and one which bids fair to revolutionize our views on the inter-
relations of the antigens and antibodies. At present our knowledge
of the subject is hardly sufficiently advanced to justify an account
of the experiments and theories on which these views are based,
and the original papers must be consulted for further information.

To revert again to the question of the specificity of the reaction :
The subject is a complex one, and we have already seen that the
phenomenon of the "group reaction " leads us to the supposition
that bacteria of different species must possess molecules of the
same nature. Further study shows that there are differences in
this respect between bacteria of the same species, which are in-
distinguishable the one from the other by ordinary morphological
and chemical tests, but which have had different origins. Thus it
is found that if the serum of an animal which is strongly immu-
nized to a given culture of V. cholera: be tested against cultures of
various stocks, that which was used for the injections will be
clumped most powerfully, the others to variable degrees. These
apparent anomalies, though inconvenient in practice, tend to make
us regard the reaction as more specific rather than as less —
sharply specific, that is, as regards a certain sort of chemical
substance which is formed in greater degree by certain races of a
given species than by others.

This modified specificity, sharp as regards chemical substances,
but not as regards bacterial species, is well illustrated in
Castellani's absorption reaction. If an agglutinating serum which ,
e.g., clumps the typhoid bacillus powerfully, and a colon less

218

CASTELLANl'S ABSORPTION REACTION

strongly (and which was obtained by immunizing an animal
against typhoid) be saturated with typhoid bacilli, centrifugalized,
and retested, it will be found that when the serum has lost its
power of agglutinating that organism, it will also be bereft of
power over the B. coli. On the other hand, if it be saturated
with colon bacilli until all its agglutinating power on that
organism is removed, its action on B. typkosus will be intact or
but slightly reduced. The mechanism by which this is brought
about is readily understandable by means of the following diagram
(Bolduan).

Fig. i represents a typhoid bacillus, shown as if it consisted of
three varieties of proteid material, A, B, and C, of which A is
present in largest amount. The agglutinin formed by its injection
will consist of three substances, each specific for its own proteid,
and the antibody to A (which we may call the main agglutinin)

B C

B

Typhoid Bacillus

Colon Bacillus
E H

Dysentery Bacillus

FIG. 45. (After Bolduan.)

will be most abundant. Fig. 2 represents a colon bacillus, and
is shown as consisting of four forms of proteid, of which B is
common both to it and to the typhoid bacillus. It follows that
the antityphoid serum will clump this bacillus as well as that of
typhoid, though not in so high a dilution ; the serum only acts in
virtue of its anti-B agglutinin, of which it possesses but a small
amount, and this can only act on proteid B, which forms but a
small part of the colon bacillus.

If the serum in question be saturated with typhoid bacilli the
whole of the agglutinins are removed in combination with the
bacilli, anti-B amongst them, and the serum will become devoid
of agglutinating action on both bacilli.

But if it be saturated with colon bacilli, the only effect will be
to withdraw the anti-B agglutinin ; the agglutinins to A and C
will remain, and consequently the serum will only lose its clump-
ing power on the typhoid bacillus to a slight extent. After

THE AGGLUTININS 2IQ

saturating the serum with all the allied bacteria on which it can
act, we might — theoretically, at least — remove all the partial
agglutinins, leaving only the main agglutinin, which acts on the
proteid formed by the typhoid bacillus, and by it only, and so
obtain a truly specific clumping serum.

Castellani's test seems to be generally correct, though excep-
tions have been recorded.

Further, bacilli of the same stock can be made to vary greatly
in their sensitiveness to the action of agglutinin by various
methods. These have been carefully studied by Bordet, Nicolle,
Kirschbruch, and others, and it has been found that the
sensitiveness of typhoid bacilli is diminished by washing, by
culture at high temperatures (40° C.) or low temperatures, by the
addition to the medium of minute traces of antiseptics, and is
less in old cultures and in those that have been recently isolated
from the living animal. It is found in general that bacilli just
isolated from a typhoid patient clump badly, and that they
gradually increase in sensitiveness for six months on cultivation in
artificial media : a very faintly acid medium is the most suitable.
Sometimes a culture alters very rapidly without obvious cause,
and this is a possible source of error in the clinical application of
Widal's reaction. The author in one case found a culture which
was clumped by a certain serum at i : 60 was clumped by the
same serum at i : 250 three days later.

Another method by which bacteria can be made to diminish in
their sensitiveness to clumping is by cultivation in specific
immune serum. This was first observed by Ainley Walker, and
since it is of some theoretical interest in connection with Welch's
theory of the nature of the unknown toxins, requires a short
description. He found that when typhoid bacilli were grown in
immune serum (of course, devoid of complement) diluted with
broth, they gradually lost their agglutinability, and became more
virulent. Thus both of the effects of" passage " were reproduced
in vitro, and Walker further found that bacilli thus treated re-
moved the agglutinin from the serum in which they were grown,
and believed his results could be best explained on the supposition
that the bacilli produced specific anti-agglutinins. This, of course,
tends to corroborate Welch's interesting suggestion that some of
the organisms for which no exotoxins have been discovered exert
their lethal action by producing antibodies against the blood of
their host.

22O IMMUNITY OF BACTERIA TO AGGLUTININS

The subject has been investigated by others, and their results
do not altogether corroborate those of Walker. Thus Miiller
found the same inagglutinability after culture in diluted typhoid
serum, but did not find any evidence .of the formation of an anti-
agglutinin ; on the contrary, he found that bacilli thus treated
had less power of weakening the action of typhoid serum than
had normal bacilli. We might explain his results by saying
the bacilli had lost their agglutinable receptors, or agglutinogen.
Bail, working with a different method, obtained comparable
results, though he explained them quite differently. He, as well
as Landsteiner and others, noticed that bacilli grown in the
immune serum grew into long branching threads, losing their
bacillary character entirely. This has been thought to be a sort
of end-to-end agglutination.

The change is a more or less permanent one. Thus Park and
Collins found that cultures (of B. coli and B. dysenteric) which
had lost their power to agglutinate might require to be grown for
months on ordinary agar before they retained their normal sensi-
tiveness. The change is evidently a very profound one, and one
that is handed on for many generations.

The subject has recently been carefully investigated by Marshall
and Knox, working with B. dysenteric. They found the same loss
of agglutinability when the bacillus was grown on immune sera,
but the same change also occurred when normal horse serum was
used ; and they further showed that the alteration was a rapid
rather than a gradual one, as Walker had found in the case of
B. typhosus. With regard to the mechanism of the process, they
proved conclusively that the modified bacilli had no power of
uniting with the agglutinin, as Miiller had also found. It is quite
clear, therefore, that the loss of agglutinability is due to the loss
of appropriate receptors, in this case at least. They point out,
apropos of Welch's hypothesis of toxins, that a bacterium which
attempted to protect itself against its host by the formation of
antibodies would have but a poor chance of surviving, whereas
by the simple process of losing some of its receptors tbey can
nullify some of the protective mechanisms possessed by the latter.

Their explanation of the process by which these resistant
bacilli are produced is interesting and suggestive. They do not
think, with Walker, that the bacilli acquire a new character —
i.e., the power of producing antibodies — and transmit it to their
descendants, but that there is simply a process of natural selection.

Of

«^£^

THE AGGLUTININS 221

In any culture of bacteria it will be found that some individuals
are more easily agglutinated than others. In the early cultures
in clumping serum the susceptible forms grow, but they sink to the
bottom in a dense network. The less sensitive forms also grow
from slight granules or a diffuse turbidity through the fluid. If
the subcultures are inoculated from this fluid, the change to the
non -clumping form occurs more quickly than if the masses were
used for the transfer. There is therefore a process of selection
of the non-clumping forms, and in time all the susceptible bacilli
become eliminated. It is this process, though in a more marked
form, that accounts for the increased virulence of the bacteria,
which is brought about by passage. For here all bacteria which
are not resistant to the bacteriolysins and to the phagocytic
action of the leucocytes will be killed off, and will not propagate
their species. We may therefore conclude on theoretical grounds
that virulent bacilli are those having (along with a potent toxin)
few receptors which can be attacked by amboceptor and opsonin.
We have seen that Pfeiffer and Friedberger's experiments do
not tend to corroborate this (in the case of amboceptor), but that
they cannot be regarded as conclusive.

The increase in the virulence of the organism by cultivation in
immune serum has been generally corroborated, though Roger
found the opposite to occur when streptococci were cultivated in
antistreptococcic serum.

The hamagglutinins occur in normal and in immunized
animals. Those in the latter call for no especial notice ; they are
developed after injection of red corpuscles side by side with
immune body, and their presence can be demonstrated if the
serum be heated or if the experiment be performed in the cold.
The normal agglutinins which one species may possess for the
red corpuscles of another species call for no special notice.

The iso-agglutinins are, however, worthy of a short description.
They are common, and have been most studied in human blood,
and will be found to occur to a greater or less extent in most
specimens of human sera. The phenomena they produce are
entirely similar to those produced by a specific serum in a typhoid
culture. Under the microscope the corpuscles will be seen to run
together into masses which are quite unlike the ordinary rouleaux
of shed blood in that the corpuscles are approximated together
without any trace of definite arrangement. When a strong serum
is used the cohesive force may be so great that the corpuscles

222 THE H^MAGGLUTININS

become absolutely fused together into a solid mass, in the centre
of which the outlines of the original corpuscles cannot be made
out. The macroscopic appearances are similar to those of a
Widal's reaction, the emulsion of corpuscles becoming granular
and flocculent, and clear rapidly, with the deposition of a red mass
at the bottom of the vessel. With a powerful serum a column of
emulsion 2 inches high may be completely cleared in five minutes,
and the clumping, as observed in a drop of the fluid on a slide,
may be complete in a few seconds.

It will be found that a certain serum does not clump all human
corpuscles equally well, and facts have been observed quite
comparable to those found by Ehrlich in the case of isohsemo-
lysin. According to Landsteiner (whose researches have been
corroborated by Hektoen), human bloods may be divided into
three groups, which Hektoen describes as follows :

Group i. — Here the corpuscles are not agglutinated by sera of
the other groups, whilst the sera agglutinate the corpuscles of
both groups.

Group 2. — In this group the corpuscles are agglutinated by the
sera of the other groups, whilst the sera agglutinate the
corpuscles of Group 3, but not of Group i.

Group 3. — The corpuscles are agglutinated by all other sera,
and the sera agglutinate the corpuscles of Group 2, but not of
Group i.

A few specimens of blood are found which do not fit exactly
into any of these categories.

It will be noted that in none of these groups is there an
agglutination of red corpuscles by its own serum — i.e., an example
of the presence of an auto-agglutinin. This substance, however,
may occur in the blood, but apparently only in disease. The
best example (in fact, the only one in which it has been positively
demonstrated) is in pernicious anaemia. In this disease it often
happens that the washed corpuscles are immediately and strongly
clumped by the serum of the same patient, and the same
phenomenon may occur when citrated plasma is used instead of
serum. Whether similar phenomena occur in the body, and if
not, the reason for its absence, is quite unknown : it is exceedingly
difficult to imagine that clumps similar to what we see in vitro
can be formed in the circulation, for they appear to be much too
large to pass through the capillaries. The only plausible explana-
tion is that the auto-agglutinin does not exist as such in the

THE AGGLUTININS 223

circulation, only coming into existence as the blood is shed. If
this is the case, it appears clear that its production is not the
result of clotting, since -the clumping may occur before coagula-
tion has taken place ; indeed, if the blood is carefully watched as
it flows from a skin puncture in a marked case of pernicious
anaemia, it maybe seen to become " streaky," pale and dark areas
being present, and a microscopical examination will show that
this is due to clumping of the corpuscles, which thus appears to
take place immediately the blood leaves the vessels.

The researches of Gay would appear to show that the clumping
of the red corpuscles by serum is not necessarily due to the
presence of an agglutinin at all, but may be caused by variations
in the tonicity of the corpuscles and serum. Thus he found that
in the bloods which have non-agglutinable corpuscles there is a
higher molecular concentration than in the other groups, indicating
a greater amount of salts both in the corpuscles and in the sera.
There are also differences in tonicity, though less marked, between
the members of the other groups. He also finds that if a serum
which is hypertonic to a certain sample of corpuscles, and which
therefore clumps them, is examined after contact with the
corpuscles in question, its tonicity is decreased until it reaches
that of the serum which normally accompanies those corpuscles,
which, of course, it now no longer agglutinates. This is a simple
saturation experiment, the old explanation of which would have
been that the agglutinin had all been removed by contact with the
corpuscles, but which Gay explains by absorption of salts by the
corpuscles. Lastly, a simple hypertonic solution of NaCl and
CaCl2, and according to Peskind a large number of acids and acid
salts, gave rise to appearances suggestive of clumping, though not
identical therewith. His researches are highly suggestive, and it
may be that we shall have to modify our ideas of the normal
iso-agglutinins ; but the subject requires further investigation.

Rouleaux formation ] presents some resemblances and also some
differences. Red blood-corpuscles, when washed free from serum
and suspended in normal saline solution, lie free side by side, and
show no tendency to run together ; but if placed in human serum
and some other fluids, the biconcave discs approach one another in
a peculiar orderly fashion, so as to form adherent rolls resembling

1 I am indebted to some unpublished researches of Dr. Wiltshire's for
much interesting information on this subject. Any new fact mentioned is
owing to him, unless the contrary is stated.

224 ROULEAUX FORMATION

piles of money. This phenomenon has attracted a good deal of
attention, though less than it deserves, and the exact mechanism
by which it is produced is still uncertain. It has obviously some
resemblance to agglutination. It is quite conceivable, for example,
that an agglutinating serum, when greatly diluted, might act so
weakly that the corpuscles might be drawn gradually together,
and so approach one another in such a fashion as to form the
peculiar rolls, very much as in the orderly deposition of molecules
to form a crystal, which occurs when a strong solution of a salt is
slowly evaporated, so as to allow the molecular attractions to
come into play in a regular fashion. But this is not the case,
since, as was shown by Descatello and Sturli, it is not possible to
dilute an agglutinating serum with normal saline solution so as to
convert it into a rouleaux-inducing one. And according to Lange,
an agglutinating serum which has had its agglutinin removed by
saturation with corpuscles will still give rise to the phenomenon
under discussion.

Both the corpuscles and the fluid in which they are suspended
require consideration in the discussion of the question why
rouleaux formation occurs in shed blood, though not in the tissues.
Human serum will always cause it, though its potency in this
respect varies greatly from time to time in the same person, and
also in different individuals (as tested on the same specimen of red
corpuscles). This power is quickly destroyed on dilution, being
almost always annulled on the addition of an equal quantity of
normal saline solution. The roulogenous principle also occurs in
human milk, but not — or not commonly — in that of cows, and is not
destroyed by boiling for five minutes. Viscosity appears to play
some part of secondary importance, and not clearly defined :
solutions of colloids, such as gum or gelatin, may induce it, as
was pointed out by Wharton Jones. The roulogenous substance
occurs to a very large extent in inflammatory exudates, but not in
transudates.

It is possible that the induction of rouleaux formation may be
due to the development of some substance in the serum at the
moment the blood is shed, and it is also possible that it may be
due to some change in the red corpuscles. It has been shown
that biconcave discs suspended in water and smeared with grease
will run together in this way, and Brunton has suggested that
when the blood is shed some fatty acid is liberated by the action
of carbon dioxide. It is also possible that the change which

THE AGGLUTININS 225

occurs when the blood is shed may be one of shape. Thus
Weidenreich and others hold that red corpuscles are normally
bell-shaped, and that the moment they leave the vessels they
become biconcave discs. If there is an increase in the tension
generated at their surface of contact with the serum, the con-
ditions for rouleaux formation would appear to be present.
Weidenreich's facts have been strongly disputed, and are not
generally accepted.

Rouleaux formation cannot be induced in corpuscles which
have been heated to 44° C.

CHAPTER IX
THE PRECIPITINS

AGGLUTININS are antibodies obtained by the injection of particu-
late substances, and have the property of causing these substances
to collect into clumps. In exactly the same way proteids in solu-
tion, when injected into suitable animals, bring about the formation
of another class of antibodies, which possess the power of clumping
the molecules of the proteid in a solution similar to that injected.
This manifests itself by the formation of a precipitate. After the
addition of the clear antiserum to the clear proteid solution, the
mixture becomes opalescent, and then opaque, and after a time a
precipitate is cast down, leaving a clear supernatant fluid. Hence
these substances are called precipitins, the substance with which
they form a precipitate, and which calls them into existence when
injected into a suitable animal (their antigen), being termed pre-
cipitable substance, or precipitogen, and the insoluble combination
of the two, precipitum. They are in all respects closely allied to
\ agglutinins, if not absolutely identical. The fact of their acting in
V a clear fluid does not prove that they exert their effect in a true
solution, since modern physico-chemical research has shown that
proteids do not form solutions, but merely emulsions or suspensions
of molecules or of complexes of molecules. The effect of the
addition of a precipitin is to cause an agglutination of these
molecules, which is entirely analogous with the agglutination of
typhoid bacilli. The laws which govern the reactions of the
precipitins and agglutinins are entirely similar, and, theoretically,
it would probably be more accurate to consider them under one
head. The practical applications of the two classes of antibodies
are, however, very different, and it is more convenient to treat
them as separate substances.

The first substances of this group to be discovered were the
bacterio - precipitins of Kraus, first investigated in 1897, anc^
referred to elsewhere. Kraus found that, if he took an old

226

THE PRECIPITINS 227

culture of typhoid bacilli and filtered it to remove the bodies of
the bacteria, and then added some typhoid serum to the clear
solution, a precipitate was formed. The same happened with
cultures of cholera and plague organisms, after addition of the
appropriate sera, so that the reaction is, within limits, specific.
The limits of the specificity are not yet thoroughly ascertained,
but there are distinct evidences of group reactions similar to those
seen in the agglutinins. Thus Norris found a common precipitin
for organisms of the cholera group, for some cocci, etc. There
were, however, some exceptions ; thus, a rabbit which had been
injected with cultures of B. prodigiosus developed a precipitin which
acted on filtered cultures of B. coli and V. Metchnikovi, as well as
on the filtrate from the organism injected. He thought, however,
that the reaction is a more intimate and constant test of group
relationships than is agglutination ; thus he prepared an antiserum
to a bacillus belonging to the hog-cholera group which did not
agglutinate typhoid or colon bacilli, but which gave the precipitin
reaction with their filtrates. The bacillus, it may be pointed out,
is a member of the same group as the other two, so that (in the
case of this particular serum) the agglutination reaction is mis-
leading.

Bacterio-precipitins may be prepared by the injection of cultures
of the bacteria, or the filtrates from the old cultures spoken of
above ; it is evident, therefore, that they are antibodies to substances
in solution. Some authorities, it is true, have thought that the
actual antigen and precipitable substance consist in reality of the
broken-off flagellae, which are sufficiently fine to pass through a
Berkefeld filter ; but this can hardly be the case, since a bacterio-
precipitin can be obtained for non -flagellate organisms. It is true
that the best-known and most powerful of these antibodies are for
organisms which possess flagellae, and it is quite probable that the
clumping of these filaments does occur in some cases, and intensi-
fies, or may be mistaken for, a true precipitation.

Bacterio-precipitins cannot be (or have not been) prepared to all
organisms. Thus, diphtheria antitoxin does not form a precipitate
with diphtheria toxin, a substance prepared on lines exactly similar
to those used in procuring the precipitating solution for typhoid
and cholera. Diphtheria toxin is an excreted substance, not a
substance dissolved out of the bacterial protoplasm.

The first to observe precipitins to other proteid solutions, and in
particular to serum, was apparently Tchistovitch, in 1898. He

15—2

228 PRECIPITOID

investigated the formation of an antitoxin for eel serum by inject-
ing that substance into rabbits, and found the serum thus obtained
not only neutralized the toxic effects of the eel serum, but formed
a precipitate with it. In an exactly similar way, he prepared a
precipitating serum to a non-toxic serum — that of the horse — and
investigated its properties. He found that the precipitate formed
by the interaction of the antiserum and its antibody was soluble in
dilute acids and alkalis, but insoluble in water, alkaline carbon-
ates, and neutral salts. These results were corroborated by Bordet,
whose classical researches on the haemolysins were made about
this time. He found that the injection of denbrinated fowl's blood
into rabbits caused the appearance of agglutinins, haemolysins, and
of precipitins in the serum of the latter, so that the fowl's red
corpuscles would be first clumped and then laked, after the
addition of the immune serum, and this substance, when added to
fowl serum, led to the formation of a precipitate. Bordet also
demonstrated the formation of a precipitating serum for milk
(lacto- serum). These researches were corroborated by Myers,
Uhlenhuth, Wassermann and Schiitze, Nuttall, and others, and
the reaction has now been very fully studied, and found to be of
considerable practical importance.

Precipitins are in all respects closely analogous to the agglutinins.
They are formed under the same conditions — i.e.t as a reaction of
the cells to a foreign material, probably in all cases of a proteid
nature — and appear themselves to be proteids. They are destroyed
by pepsin and other proteolytic enzymes, and by acids and alkalis.
They appear to be allied to or carried by the globulins. When
heated they appear to undergo a change into precipitoid, a sub-
stance which has the power of combining with the molecules of
precipitogen, but which does not precipitate them ; this is shown
by the fact that a further addition of precipitin does not cause
precipitation, indicating that the effect is not due to a mere
weakening of the substance or to its partial destruction. Hence
we deduce that the precipitin molecule consists of two portions —
a thermostable haptophore or combining group, and a thermolabile
functionating group, the action of which is necessary for agglutina-
tion of the molecules to occur. This change takes place at a
temperature of 50° to 60° C., but it also occurs slowly when
precipitin solutions are kept at ordinary temperatures, so that these
become gradually useless as tests for their antigens ; and the
change may be hastened by the action of light or of various

THE PRECIPITINS

22g

chemical substances. Precipitating sera should, therefore, be
kept in a dry state in a cool place, and preserved from light.

Precipitoids appear to have a stronger affinity for the precipitate
substance than has the unaltered precipitin — an alteration in
affinity similar to that which we have seen to occur sometimes in
the case of agglutinin, and which leads to the formation of what
we have termed pro-agglutinoid. Thus, if a serum which has
been heated until it has lost its precipitating power be mixed
with unheated precipitin, the mixture will not form any precipitate
after the addition of small amounts of the normal serum ; it is only
after enough of the latter has been added to combine with all the
precipitoid that a precipitate begins to appear. The same pheno-
mena may occur in working with old and degenerated sera, or
occasionally even with fresh ones. An example will make this
clear.

Normal Serum.

Antiserum.

Amount of Precipitate.

7

nil

6

nil

5

nil

4

nil

3

nil

2

0*5

I

I

i

I

2

1-25

I

3

2

I

4

3

I

5

3 '5

I

6

375

7

4

8

4

10

3-25

12

275

H

2-25

16

2

18

1-5

20

I

24

nil

3°

nil

Here the maximum precipitate was given when 8 parts of anti-
serum were mixed with i part of normal serum. When i part of
normal serum was mixed with 24 parts of antiserum, there was no

230 PRECIPITOID

precipitate — i.e., the precipitoids were at this point just saturated ;
but in a mixture of 20 parts of antiserum with i part of normal
serum, it appears that, after all the molecules of precipitoid were
saturated, there was still enough precipitable substance present to
combine with some of the precipitin, and thus to form a precipitate.
In the mixture of 8 and i the precipitoid was all saturated, and a
maximum amount of precipitable substance left over to combine
with precipitin.

This theory of the " specific inhibition " of the action of this
precipitin was the first explanation to be brought forward, and
appears to afford a fairly satisfactory explanation of the facts
observed. It is, however, quite probable that future physico-
chemical research may show it to be erroneous, and that these
zones of inhibition are in reality similar to those observed in the
non-specific precipitation of colloids, of which we have already
spoken. To this subject we shall return. Be the explanation
what it may, the phenomena are of considerable practical impor-
tance in the application of the precipitating sera to the diagnosis
of the nature of an unknown serum or other solution of proteid.
The mere fact of not obtaining a precipitate when a solution of
unknown strength (e.g., of a dried blood-stain) is added to an
anti-serum is not necessarily of any importance, and to obtain
accurate results it is necessary to perform a series of quantitative
experiments.

The experiments of Eisenberg, Michaelis, Fleischmann, and
others tend to show that the combination between the two obeys
the laws governing the combination of weak acids and bases,
which we have already discussed, and that when the two sera are
added together both free precipitin and precipitable substance may
be present at the same time. This, if true, would not explain
relations between the precipitin and precipitable substance similar
to those given above ; if the law of the mass reaction applied,
an excess of antiserum would tend to give the maximum
possible quantity of precipitum, though there would always be
some unaltered precipitable substance present in an unaltered
state. It is, however, doubtful whether the substances do actually
interact as weak acids and bases. Von Dungern, experimenting
with antisera obtained to the proteids of cold-blooded animals,
found that the reaction was rigorously quantitative, but that a
complication was introduced by the presence of two varieties of
precipitin, special and partial. The special precipitins are sup-

THE PRECIPITINS 23!

posed only to act on the proteid which has been used for the
immunization, whereas the partial ones act both on it and on
other foreign proteids. If these are formed at different rates of
speed, it is easy to see that at a certain stage in the process of
immunization an animal's serum may contain a precipitin (e.g.,
the special precipitin) and a precipitable substance (e.g., that which
acts as an antigen to the partial precipitin). In most of their
reactions the precipitins certainly act as if they had a powerful
combining affinity for their antigens, and von Dungern's observa-
tions may supply the explanation of the apparent discrepancies.

If precipitable substance (e.g., egg-albumin) be heated, it under-
goes a change comparable to that sustained by precipitin on
heating : it loses its power of becoming precipitated with anti-
serum, but retains its function of combining therewith. This
altered proteid is sometimes called precipitoid, since it is analogous
with the precipitoid derived from precipitin by the action of heat.
The term, however, is a bad one, since its use leads to confusion
between the two substances, and the only word which is at all
suitable is "precipitogenoid." Precipitogenoid, therefore, consists
of the haptophore portion of the molecule of precipitogen, another
portion of this molecule which must be present in order that the
reaction may take place being destroyed. From the fact that it
retains its haptophore group we should expect it to act as an
antigen, just as toxoid does ; and this is the case, for the injection
of precipitogenoid calls forth the production of ordinary precipitin.
In any complete quantitative study of the interactions of serum
and antiserum it is necessary to investigate the question of the
presence of precipitogenoid as well as of precipitoid, and the
problems thus may become complicated in the extreme.

Precipitins, like the agglutinins, act most quickly at body
temperatures, and show their essential similarity in the fact that
salts are necessary for both reactions ; short of absolute absence,
the amount of the precipitum form depends on the quantity of
salts present (Friedmann).

There is an interesting analogy between the agglutinins and the
precipitins, in that the latter as well as the former are occasionally
seen in the serum of unimmunized animals, though the precipitins
are much more uncommon in this situation. Thus, Hoke found
the presence of bacterio-precipitins to nitrates from B. typhosus
and V. cholera fairly frequently present in the serum of the ox,
more rarely in that of the goat, pig, and sheep. The presence of

232 SPECIFICITY

serum precipitins is rarer, but occurs, according to Noguchi, in
cold-blooded animals (crustacea, etc.), and Lamb found in normal
rabbit serum a precipitin for cobra venom, and Obermayer and
Pick one for dysglobulin of egg-white in the same fluid.
N^The relation of the precipitins as regards specificity forms a
difficult and most important question. It was at first thought
that the specificity was a sharp one, and that a serum prepared by
the injection of human serum would only precipitate with human
serum, and a lacto-serum prepared by the injection of cow's milk
would only precipitate with that fluid, and not with the milk of
other animals. Thus, Uhlenhuth prepared a precipitin by inject-
ing a rabbit with ox serum, and found it gave a precipitate with
that fluid, but not with the serum of the horse, donkey, pig, sheep,
dog, cat, deer, fallow-deer, hare, guinea-pig, rat, mouse, rabbit,
chicken, goose, turkey, or pigeon ; but he also showed that anti-
egg serum was not sharply specific, and would coagulate solutions
of albumin from eggs other than those of the species used for the
injection. Wassermann and Stern also showed that antihuman
serum would react, though but slightly, with the serum of the
baboon, and Stern confirmed this, and found it reacted with the
serum of other monkeys. Hence the idea gradually arose that
the precipitin obtained by the injection of a serum from one species
of animal is not specific for that species, but will give a precipi-
tate, though of less amount, with the sera of other species, provided
that they are sufficiently closely allied zoologically. This has been
especially studied by Nuttall, who expresses this relationship
between species close together in the animal scale as a " blood-
relationship." He showed that, provided the serum were powerful
enough, it would react with all the bloods of animals in the same
great division of the animal kingdom (mammalia, birds, reptiles,
amphibia, etc.). Thus, a strong antihuman serum will give a
precipitate with human serum, even when highly diluted, with
apes, monkeys, etc., but not in such high dilutions, and a slight
trace of precipitate after a long period when mixed with the sera
of more remote mammalia, but no precipitate with the blood of
birds, fishes, etc. Quite similar relationships hold with the
lacto-sera and with the precipitating sera for muscle proteids;
the anti-sera for egg proteids is apparently less specific.

Hence we deduce that the precipitins are not specific as regards
the animal species from which they are derived, but possess that
partial specificity seen in the cytotoxins and in the " group

THE PRECIPITINS 233

reactions " of the agglutinins ; that is to say, they are specific as
regards the antigens which bring them into existence, irrespective
of the source from which that antigen was derived. This appears
to be substantiated by experiments on purified and recrystallized
proteids, the precipitating sera for which show a high degree of
specificity. Too frequent recrystallization of proteids, however,
appears to injure their power of inducing the formation of
precipitins.

Further, proteids which have been altered in character by
chemical processes (iodized, nitrified, or denitrified) will cause
the formation of precipitins which are specific for the transformed
molecules of proteid, no matter from what source they were
derived. These observations are of profound interest, since they
appear to show that it is possible by artificial means to alter com-
pletely the haptophore group of a proteid molecule. According
to Uhlenhuth, there is little or no specificity in the antisera pre-
pared against the proteids of the crystalline lens. An antiserum
prepared by injections of serum will precipitate all albuminous
fluids except a solution of the lens, which in its turn will not pre-
cipitate with serum. But an anticrystalline serum prepared by
the injection of ox lenses into rabbits will give precipitates with
lens solutions from mammals, birds, amphibia, and even fish.

Hence, too, a practical point in connection with the employment
of precipitating sera in the detection of the source of blood. Here
it is necessary to use a high dilution of the blood to be tested, and
if a reaction is given, to try again with higher dilutions until the
limit is reached. In testing dried blood-stains for medico-legal
purposes it is, of course, usually impossible to determine the exact
amount of serum which has been taken up by the solvent (normal
saline solution). It is found, however, that a i : 1,000 dilution
of serum in normal saline solution will give a good froth when air
is allowed to bubble through them from a pipette, and this will
give a rough idea as to the amount of proteid material dissolved
out of the clot to be tested.

The delicacy of the reaction is truly astonishing. Thus Ascoli
obtained an anti-egg albumin serum which gave a precipitate with
a i : 1,000,000 dilution of egg-albumin, and Stern an antihuman
serum which reacted with serum at a dilution of i : 50,000. These
are extreme figures, but sera active on solutions diluted i : 5,000 are
frequently obtained.

As regards the substances for which precipitins can be obtained,

234 ISOPRECIPITIN

we find them limited in all cases to the proteids. All the coagu-
lable proteids will give rise to the formation of precipitating sera.
As regards the formation of these substances by means of digested
proteids (peptone and albumoses), the facts are less certain.
Myers obtained a precipitin to Witte's peptone, but according to
Obermayer and Pick complete peptic digestion destroys both the
power of inducing the formation of antibodies and of being pre-
cipitated. The former is first to go, so that a partially digested
proteid solution which will no longer precipitate with an antiserum
will yet lead to the production of a precipitin when injected into a
suitable animal. The results with tryptic digestion appear to be
entirely similar. Complete destruction leads to loss of both
functions. It appears, therefore, that it is only the giant molecule
of proteid that can be agglutinated ; the smaller, diffusible
molecule of the products of proteolysis, like the molecules of
toxin, though they may, and probably do, unite with their anti-
bodies, do not manifest this combination by forming clumps.
Proteids which have been coagulated by heat have still the power
of forming a precipitin (which, of course, manifests its action on
solutions of the unaltered proteid) when injected.

As regards the species of animal from which the precipitins are
to be prepared, it is natural to choose an animal as remote in the
zoological scale as possible. In the case of human sera or in-
different substances, such as solutions of albumin, etc., the rabbit
is usually chosen, since it is easily obtained and handled, and will
yield a very considerable amount of serum. Where antisera to
animals closely allied to rabbits are to be obtained, fowls or other
large birds are most suitable. There are, however, some observa-
tions which go to show that there are differences between individual
rabbits which are comparable in every respect to those between
the different red corpuscles of goats and other animals, as seen in
Ehrlich's experiments on the isohaemolysins. Thus Schiitze
injected the serum of rabbits into rabbits. In two cases out of
ten he obtained a precipitin which reacted with the serum injected.
This substance is called isoprecipitin. It may often be noted
that a precipitate falls when different samples of diphtheria anti-
toxin (in animals in which the earlier stages of the process of
immunization have been carried out by means of serum toxin) are
mixed together, and this is probably an analogous phenomenon.
These precipitins are, however, very feeble as compared with
those obtained by the injection of the serum of a certain species

THE PRECIPITINS 235

into an animal far removed in the zoological scale, and their
explanation is obvious in the light of what has been already said
with regard to individual differences in receptors.

According to Ewing, an antihuman serum prepared from the
fowl shows a far higher degree of specificity than those obtained
by injections into rabbits.

A short account of the practical application of the precipitins
may not be out of place. The chief is, of course, the medico-legal
identification of blood-stains, the chief exponents of which are
Uhlenhuth and Wassermann. The antiserum is obtained from
rabbits, which are treated by intravenous or intraperitoneal in-
jections at intervals of three or four days. The material used
for the injection may be blood obtained by venesection or vein
puncture, or from the placenta or from the cadaver, or pleuritic or
ascitic fluid may be used ; in any case strict asepsis is necessary.
The amount given rises from i to 3 or 4 c.c. in the case of
intravenous injections, or twice as much or even more in the
peritoneum. The course of treatment lasts three or four months.
Another and simpler method is to give larger doses — up to 10 c.c.,
or even more — intraperitoneally at intervals of a week. The
animal is then chloroformed, and as much blood as possible
collected either from the heart or carotid artery.

The fluid to be tested is prepared by maceration of the clot,
piece of stained linen, etc., with normal saline solution, or with
i per cent. NaOH. In the case of a very old stain, Ziemka
recommends the use of a strong solution of potassium cyanide,
which is subsequently neutralized with tartaric acid. This is
examined with the microscope and tested spectroscopically to
determine the presence of blood-corpuscles and pigments. The
solution is then filtered. Air is then allowed to bubble through
the fluid to make sure that proteid material has actually passed
into solution ; this is indicated by the production of a stable foam.
Three tests are made. In the first tube one part of the fluid under
examination is mixed with two of the antiserum, the second
contains the fluid alone, and the third antiserum plus normal
saline solution. Further controls, in which the antiserum is
mixed with diluted serum from animals other than man, may also
be made if necessary. The tubes are usually incubated and
examined from time to time, and a positive result is obtained if
there is a precipitate in the first tube and not in the others.
Further tests are then made with greater dilutions, and with a

236 DEVIATION OF COMPLEMENT TEST

powerful antiserum a reaction can usually be obtained in dilutions
so high that proof of the presence of proteids is barely obtainable
by ordinary chemical means. The weak point in the method is
that it is never possible to say exactly how much of the proteid
matter of the clot has been dissolved, and thus to compare the
effect of the antiserum on the solution with its action on diluted
serum of man and of other animals. Given a sample of unaltered
and undried serum, the test can be carried out with almost
complete certainty ; but this is rarely, if ever, possible in actual
practice.

Another test, based on Gengou's reaction, has recently been intro-
duced by Neisser and Sachs. It is carried out in the following
way : A hsemolysin for ox corpuscles is prepared by injecting
these bodies into a rabbit. Another rabbit is injected with human
blood, so as to lead to the production of a precipitin. When the
test is to be made the fresh serum of the latter animal (or if only
stale serum is at hand, some fresh normal rabbit's serum must be
added to supply complement) is added to the fluid to be tested.
If human serum is present, even in an amount so small that no
precipitate is formed, the antigen and antibody combine and with-
draw the complement from solution. To test this ox corpuscles
are sensitized with the heated (decomplemented) serum of the first
rabbit and thoroughly washed, and some of the mixture added.
If the complement has been withdrawn, of course no haemolysis
will occur. Certain obvious controls are employed to demonstrate
that the corpuscles were actually sensitized, and that complement
was present in the rabbit's serum before the addition of the fluid
suspected of containing human blood.

This test is extraordinarily sensitive, Neisser and Sachs finding
that the millionth part of a cubic centimetre of human serum was
readily demonstrable. They claim also that it is more specific
than the ordinary precipitin test, it being necessary, for instance,
to use T^Viy c.c. of ape's serum to get the same result. But the
technique is complicated, and it appears, moreover, that comple-
ment may be abstracted in an altogether non-specific manner by
substances other than the combination of antigen and antibody.
Thus Uhlenhuth examined some spots of blood on a sack which a
cabdriver (who was found dead) used as a seat, and found it
brought about a fixation of complement, though it gave no reaction
with the precipitin test. He then tested the material of which
the bag was made, and found it also had the power of absorbing

THE PRECIPITINS

237

complement. According to Neisser and Sachs, however, this
power is destroyed if the serum solution is boiled, but is unaltered
in the case of non-specific substances. Another serious objection
is that a similar deviation may be brought about by means of
sweat, so that if the reaction were obtained in a stain on body-
linen it would be of little value.

The precipitin reaction has also been used for determining the
nature of meat (whether fresh, as in the case of beef suspected to
be horse-flesh, or prepared, as in sausages, etc.). The serum is
prepared by injecting meat-juice or an (unheated) watery extract
of the meat, and the test is carried out on lines similar to those
described above. It has also been employed to determine the
nature of bones, i.e., whether they are human or otherwise.

CHAPTER X
PHAGOCYTOSIS

METCHNIKOFF'S researches on phagocytosis in the lowly organized
animals formed a starting-point for an entirely new series of
researches on the subject of immunity, and his treatise on the
{< Comparative Pathology of Inflammation " must ever remain a
great medical classic, as well as a most fascinating work.
Metchnikoff was primarily a biologist, and his attention was
attracted to the spectacle of amoebae and other unicellular
organisms containing bacteria. In these lowly constituted
animals the process by which the cell takes in foreign particles
can be watched with ease, and the steps of the process traced.
The amoeba throws out arm-like processes, which surround the
bacterium, close on it, and in a few minutes an organism
which was previously lying free is deeply embedded in the
animal's protoplasm. Metchnikoff watched its fate, and found
it lost its sharp outline and clear appearance, became granular,
and in a little while disappeared altogether. He found also that,
in many cases at least, a minute vacuole was formed round the
ingested organism, and further research showed that this vacuole
contained a fluid which was acid in reaction and held in solution
a digestive ferment allied to pepsin. Considering this observation
more closely, it is obvious, firstly, that the amoeba must be regarded
as being immune to the organisms which it ingests and digests,
and, secondly, that in this case the processes concerned in im-
munity are those concerned in nutrition : the amoeba is immune
to the bacterium because it can make use of its protoplasm as
nourishment.

But this process, as Metchnikoff soon found, is not confined
to the unicellular organisms. His most beautiful and classical
series of researches deal with the properties and action of the
leucocytes in a small fresh-water crustacean (daphnia), which,
from its transparency and small size, is a very suitable object for

238

PHAGOCYTOSIS 239

observation. Leucocytes, it may be pointed out, are cells which,
from their isolation from one another, power of independent move-
ment, etc., are closely analogous with amoebae and other lowly
organized protozoa.

Metchnikoff found that daphnia is subject to a disease due to
the invasion of its body-cavity by the spores of a yeast— the
monospora — and that if these spores gained access in large
numbers they multiplied, formed into mature organisms, and
finally killed their host. When, however, few spores gained
access, he found that the daphnia's leucocytes approached them,

FIG. 46.— AN AMCEBA WHICH HAS INGESTED NUMEROUS SPECIMENS OF
MICROSPH^RA. (Metchnikoff.)

a, #, Vacuoles.

formed a wall round them, and finally digested and destroyed
them. It is obvious, therefore, that the immunity of these
animals is relegated, so to speak, to its leucocytes. If these are
efficient, the animal is preserved from its invaders; whereas,
if they make default, the latter multiply, and bring about the
lethal issue.

MetchnikofFs experiments were by no means confined to the
action of the leucocytes on bacteria. They included a careful
and exhaustive study of the method of absorption of all manner
of particles, organized and unorganized, in the tissues and body-
cavities of animals of all positions in the animal world, and they
proved to the full the importance of phagocytosis and intra-
cellular digestion, and one of the chief — if not, indeed, the only —

240

PHAGOCYTOSIS IN THE LOWER ANIMALS

method by which intruding particles are dealt with in the animal
economy. And, further, they correlate in the clearest possible
way this " scavenging " of the tissues with normal processes of

FIG. 47. — DAPHNIA CONTAINING LARGE NUMBERS OF MONOSPOR/E.
(Metchnikoff.)

FIG. 48. — HIND PART OF DAPHNIA.
At a the spores of monospora are shown surrounded by leucocytes.

digestion and nutrition. Take, for instance, the absorption of
the red blood-corpuscles of birds, which are very suitable for
research, being readily recognized (they are nucleated), non-
poisonous, and easily digested. Metchnikoff showed that these

PHAGOCYTOSIS 241

substances are absorbed from the alimentary canal of the lower
invertebrates by a process of intracellular digestion, and by that
alone. Thus, when the alimentary canal of a planarian (Dendro-
ccelmn lacteum, an animal resembling the liver-fluke in its general
anatomy) is filled with blood, the latter is found to undergo the
changes in colour familiar in bruises, and this is due to the fact
that the blood has been taken from the alimentary canal by the
cells lining it, and there has undergone the digestive changes.
These processes can readily be traced under the microscope, and
the blood-corpuscles can be seen, at first embedded in protoplasm
and of normal contour, and later enclosed in vacuoles and of altered
shape. Complete digestion takes several days, and every stage of
the process is easy to watch. Absorption of organized food particles
takes place from the alimentary canal in exactly the same way
in actinians, molluscs, and many other lower animals, and in
many of the cases which have been investigated the vacuoles are
found to contain a digestive ferment allied to trypsin or pepsin.
In higher animals this process of intracellular digestion does not
occur, or only to a slight extent (in the case of fats), the animal
finding it more advantageous to secrete the digestive juices into
the alimentary canal, and to absorb the products of their action
therefrom in a state of solution.

Absorption of particulate substances which have gained access
to the tissues takes place, to a very large extent, in a method
entirely similar. Thus Metchnikoff showed that when bird's
corpuscles are injected into the tissues of the larva of the cock-
chafer, or into the snail, earthworm, the peritoneal cavity of the
goldfish, etc., the process is also intracellular and entirely similar
to those which occur when these corpuscles are injected into the
alimentary canal of the planarian. In most cases the corpuscles
are absorbed by the leucocytes, in others by the cells of the part
(such as the endothelial cells lining the peritoneum) ; but in all
cases there is the same vacuolation, the same series of changes
in the ingested corpuscles, and the same final result. We shall
not be far wrong in associating the absorption of these corpuscles
in a very close way with processes of digestion and nutrition.

The French school have studied these processes of absorption
of particulate bodies from the tissues in a very complete manner,
and have shown beyond dispute the importance of phagocytosis
in this respect. Thus particles of carbon in the lung, the granules
of pigment left after interstitial haemorrhages, etc., are all ingested

16

2^2. METCHNIKOFF'S THEORY OF IMMUNITY

and removed by phagocytes of one sort or another, and this dis-
covery throws a flood of light on the meaning of many of the
phenomena of inflammation, more especially on the leucocytic
invasion of the injured tissues, which has for its object, in part
at least, the removal of particulate substances which are the cause
of the injury. Further, the dead or injured tissues, the result of
the action of the irritant, are eroded and removed by phagocytic
action. If a small volume of tissue, situated in a region to which
numerous leucocytes can gain access, be killed, it may be com-
pletely removed piecemeal in this way. We may quote as an
example the complete absorption of the central slough which
often takes place in acne pustules or small boils, especially after
treatment with staphylococcic vaccine. Very interesting in this
connection is the process of the absorption of the tail of the
tadpole, which is removed in an entirely similar way by the action
of phagocytes. And it is worthy of notice that the digestive
power of the phagocytes is a very powerful one, and substances
usually deemed entirely insoluble may be gradually removed by
their action.

We have already referred to the action of leucocytes in ab-
sorbing and neutralizing toxins, and have quoted the beautiful
experiments of Besredka on trisulphide of arsenic, which, when
placed 'in the peritoneal cavity, is absorbed by these cells and
prevented from exercising its poisonous action ; whilst, when
shielded from their attack by being enclosed in bags which are
permeable to fluids, but not to solids, it undergoes gradual
solution, and leads to fatal intoxication.

Metchnikoff applied these researches on phagocytosis to the
question of immunity, and formulated a complete and logical
theory on the subject, which he illustrated with many striking
researches and examples. For him — and modern advances tend
more and more to corroborate the truth of this view, though in a
much more complicated way than he thought — the defence of the
animal economy is entrusted entirely to the phagocytes, and
especially to the leucocytes. If a bacterium enters the tissues
these cells may at once make their way to the seat of infection,
and proceed to ingest the bacteria and kill them intracellularly.
In this case the animal recovers, with or without a transient illness,
and we say it is immune. If another species of bacterium enters,
the effects may be different : the leucocytes may. perhaps, be
repelled instead of being attracted, or, if attracted, may be killed

PHAGOCYTOSIS 243

by the action of the bacterial toxins, so that no phagocytosis
occurs ; or perhaps they may take up the bacteria and then be
killed by the toxins. In any case, the result is the same : the
bacteria continue to grow and to produce their toxins, and the
result is death. The animal is susceptible to the bacterium
because its leucocytes are unable to deal with it. Immunity,
therefore, is a function of the phagocytes.

So far we have dealt with natural immunity. The application
of Metchnikoff's theory to acquired immunity is equally simple,
though much less satisfactory. He argues that the leucocytes,
in their contest with a particular species of bacterium, become
educated to overcome this bacterium, and are able to deal with it
in future with great ease. In the first infection there may be a
balanced contest of some severity and duration, but as a result
the leucocytes, like war-trained veterans, are readily able to cope
with the invader a second time. This theory, though ingenious,
cannot be maintained at the present day. Its truth rests, of
course, on the truth of Metchnikoff's main thesis, which is only
partially true, and which is only one factor in the complicated
phenomena of immunity. We may just point out, however, that
the life of a leucocyte is, in all probability, a comparatively short
one, to be measured by days, or at most weeks, so that acquired
immunity due to the education of the leucocytes would be of
short duration. Nor is it of any assistance to argue that in the
struggle against the invading bacterium the fittest leucocytes
would survive, and so lead to the general improvement of the
leucocyte species : for leucocytes do not propagate themselves,
but are emitted from the bone-marrow, run their course in the
blood, degenerate, and die, their place being taken by others from
the same source. The education, therefore, must be one of the
bone-marrow, and we cannot conceive how this could take place
as the result of phagocytosis going on in a distant area. It is
possible that something of the sort may occur, but only by the
action of toxins and other bacterial products circulating in the
blood-stream, and being thus brought into the marrow. Other
considerations might be urged, but the theory has now but a
historic interest. It has served its purpose : it has been the
means of suggesting many researches which have helped greatly
in the elucidation of a most difficult subject.

Before discussing the role of phagocytosis in the infective pro-
cesses we must deal briefly with two subjects : firstly, the means

1 6 — 2

244 CHEMOTAXIS

by which the phagocytes are brought into contact with the
bacteria — in other words, of chemotaxis. This is a phenomenon
which is displayed by almost all motile and unicellular organisms,
whether animal or vegetable, and by the leucocytes of the higher
animals, and manifests itself in a movement of the organism in
response to a chemical stimulus. To take an example from the
bacteria : if a capillary tube containing a solution of meat-extract
be placed in a watery emulsion of B. coli or B. typhosus, the
bacteria will be seen to group themselves round the mouth of the
tube, which they ultimately enter. This is an example of positive
chemotaxis, the bacteria being attracted by the extractives, which,
indeed, they utilize as food. On the other hand, bacteria will
tend to remove themselves from an area from which alcohol and
/ similar substances are diffusing : this is negative chemotaxis. As
a general rule, we may say that motile organisms tend to be
attracted into a region rich in useful and nutritious substances,
and are repelled from injurious ones, but this is not invariably
true in the artificial conditions of experiment. An organism may
be lured by a useful substance into a region where there is a
sufficient quantity of an injurious one to kill it.

In cases where leucocytes make their way into a tissue infected
with bacteria it must be, therefore, because the latter give off a
substance which has a positive chemotactic action on them. This
action may readily be shown by experiment, and the easiest way
is to work with the lymph of a cold-blooded animal, since in this
way all difficulties connected with a warm stage are avoided. If,
for instance, we take a few anthrax bacilli and place them on a
slide, add a drop of frog's lymph (from the dorsal lymph sac), and
apply a cover-glass, the leucocytes can be easily seen to crawl
actively up to the bacilli. To this experiment (Kanthack's) we
shall have to recur.

In nearly all cases we find that leucocytes are attracted in large
numbers into the area in which the bacteria are situated — i.e..
nearly all bacteria give off substances which are positively
chemotactic for leucocytes. In a few cases, however, this appears
at first sight not to happen, for when tissues which are infected
with very virulent bacteria are examined microscopically, they
are often extremely poor in leucocytes. As an example, we may
take any example of acute spreading gangrene of bacterial origin.
But, as Kanthack pointed out, these are not necessarily examples
of negative chemotaxis, and it is quite probable that the paucity

PHAGOCYTOSIS 245

of the leucocytes is due to their paralysis and destruction by the
powerful toxins which are given off. Certainly since the use of
methods involving experiments on phagocytosis in vitro no valid
example of negative chemotaxis has been adduced, and it is highly
probable that leucocytes are attracted by all bacteria, whether
their toxins are mild or potent. In the former case the leucocytic
infiltration will be very obvious ; in the latter the cells will be first
paralyzed and then destroyed as soon as they reach an area in
which the toxin is present in a high degree of concentration.

We must not assume that this property of being attracted by
bacteria is of any advantage to the leucocyte itself, arguing from
the fact that the free-swimming unicellular organisms are attracted
by food and oxygen. The leucocytes are in many cases attracted
into the infected area to their own undoing, and it must not be
forgotten that even in an inflammatory process which is mild in
nature and favourable in result the number of leucocytes which
may be killed in the conflict is enormous. The leucocytes are not
independent protozoa inhabiting the blood and tissues, but an
integral part of the organism. It is to the advantage of the latter
that the former should be attracted at once to the seat of invasion,
and hence the processes of evolution have led to the development
of this function in the nomadic cells of the body. These are
extraordinarily susceptible to chemotactic influences ; they seem
to be attracted by any deviation from the normal constitution of
the tissues and fluid : a slight injury, a haemorrhage, the presence
of a poison, of a foreign body of any sort or of any dead or useless
tissue, and the leucocytes are immediately attracted into the area
affected. The more we regard the process, the more we must
regard it as one of the most exquisite examples of means to ends
met with in the animal economy.

Secondly, with regard to the nature of the cells which have the
power of acting as phagocytes. Of these the most important are
the leucocytes, and especially the polynuclear and large hyaline
cells of human blood. All the leucocytes, however, have phago-
cytic powers, as is well seen in opsonic estimations : the cells
which take up the fewest bacteria are the eosinophiles and the
small lymphocytes. The former are very deficient in this direc-
tion, and we may be certain that their main function is entirely
different. The lymphocytes, too, take up but a small number of
bacteria ; but when activated by a suitable haemopsonic serum, they
take up red blood-corpuscles in considerable numbers, and a

246 VARIETIES OF PHAGOCYTES

lymphocyte may then present a remarkable appearance, having
ingested two or even three red corpuscles, each as large as itself,
the narrow band of protoplasm being extended over the corpuscle
in a most extraordinary way.

It is noticeable, too, that even among the polynuclears there are
great differences in phagocytic powers. This is best seen in a film
of pus from a case of gonorrhoea, in which certain cells (indis-
tinguishable from the others) are packed full of cocci, whereas the
vast majority are entirely free. The same fact is brought out

FIG. 49. — RED CORPUSCLES INGESTED BY POLYNUCLEAR LEUCOCYTES
AND LYMPHOCYTES. (Original.)

clearly in opsonic experiments, where some leucocytes are often
found to take up very large numbers of bacteria, the general
average of the other cells being low.

Besides the leucocytes, some of the tissue cells, which are
either free or have the power of becoming so, are active phago-
cytes. Of these, the most important are the endothelial cells.
These are only flat plates of protoplasm when under normal
conditions. When submitted to the action of almost any irritant
they become cuboidal or columnar, and are then most active
phagocytes. A good example of this may sometimes be seen in
sections of a thrombosed vein at a certain stage : the endothelial
cells are columnar and contain much protoplasm, and this latter
is packed with pigment granules absorbed from the altered blood
in the lumen. But the process goes farther than this, and the
endothelial cell (whether of the serous membranes, vessels, or
lymph clefts) either breaks loose from its attachments or buds off

PHAGOCYTOSIS 247

fresh cells, in either case leading to the production of free and
motile endothelial cells, which have the closest resemblance to
the large hyaline cells of the blood, though they may be much
larger. These cells are most active and important phagocytes,
especially in the peritoneum. They have also the power of
undergoing organization, especially into fibrous tissue, and many,
if not all, of the fibroblasts of granulation tissue are endothelial in
origin. Further, some at least of the giant cells so familiar in
chronic inflammatory processes are derived from the endothelial
cells of the lymph clefts and lymph capillaries. This has been
proved to demonstration by Bergengriin in the case of the giant
cells in leprosy. Our knowledge of the origin of the giant cells of
tubercle is less exact, but analogy with those of leprosy would
lead us to infer that they are endothelial also. In both diseases
the giant cells are most important phagocytes.

Epithelial cells only exceptionally act as phagocytes. We
have already referred to the ingestion of fat by the columnar cells
of the intestine, and the other important example is supplied by
the epithelial cells lining the alveoli of the lungs. These are
flattened plaques under normal conditions, but in the presence of
an irritant they become cuboidal or columnar, detach themselves,
or bud off similar cells, and are powerful phagocytes. These are
the dust cells so frequently seen in sputum.

The nature of the cells which take part in phagocytosis is
determined to some extent by the nature of the irritant. Thus,
when the pyogenic bacteria are ingested it is usually by poly-
nuclear leucocytes, whereas it is extremely rare to find these
cells containing tubercle bacilli in the tissues, though they will
take them up readily enough under the artificial conditions of
opsonic experiments. MetchnikofF classifies phagocytes into two
groups — macrophages and microphages. His description of these
cells is not absolutely clear, but in general the microphages
correspond to the polynuclear leucocytes, and the macrophages to
the large hyaline cells of the blood and the endothelial cells of
the serous sacs and connective tissues. He claims that the
former are especially concerned with the phagocytosis of bacteria,
the latter with red blood-corpuscles and similar objects. This
distinction is not a valid one, since endothelial cells are ex-
tremely active phagocytes for bacteria, and polynuclear cells will
ingest red corpuscles with great readiness when provided with
a suitable opsonin (see Fig. 49). It would appear that under

248 PHAGOCYTES OF THE PERITONEUM AND LUNG

suitable conditions any phagocyte can ingest any abnormal body
of suitable size.

Under certain .conditions there may be traced a remarkable
sequence in the advent of various phagocytes to an infected area,
which almost suggests a symbiosis of the nomadic cells. Thus,
when a culture of a bacteria is injected into the peritoneum of the
lower animals, a very definite sequence of events takes place.
The peritoneal fluid normally contains some small mononuclear
cells, probably of endothelial origin, and a few polynuclears.
For an hour or so after the injection these cells are diminished in
numbers, and the eosinophiles disappear altogether. The exact
cause of this diminution is not quite clear. It may be due to the
dilution of the peritoneal fluid by the solution injected, or to the
destruction of the cells, or to their clumping together on the
omentum, but in any case is not due to negative chemotaxis. For
the next two hours or so there is a gradual increase -of the
polynuclears, at the end of that time an influx of small mono-
nuclears, until in about six hours these and the polynuclears are
present in approximately equal numbers. After this the mono-
nuclears (which are probably budded off from the peritoneal
cells, and are thus endothelial in origin) gradually become larger,
forming what Metchnikoff calls macrophages ; then the fluid
slowly becomes concentrated, and the polynuclears gradually
disappear, many being ingested by the large mononuclears.
Finally these too disappear, but it takes about a fortnight for the
animal to revert to its normal condition. Both varieties of cells —
endothelial and polynuclear — take part in ingesting the cocci.

Briscoe has shown that a remarkable series of processes also
occurs when organisms, etc., are injected into the alveoli of the
lungs. It varies according to the substance injected, and we
may take the result of the injection of the potato bacillus as an
example. In the first hour and a half phagocytosis is very active,
the bacilli being taken up exclusively by the pre-existing alveolar
epithelial cells. Up to this time practically no polynuclears have
made their appearance, but they now commence to be attracted
into the alveoli, where they occur in large numbers for about
twenty-four hours. They take, however, but a small share in the
phagocytosis of the bacteria, and are themselves taken up by the
alveolar cells. Some proliferation of these latter cells occurs,
and the eosinophiles increase for the first twenty-four hours, and
then gradually diminish. We cannot trace the reason for this

PHAGOCYTOSIS 249

sequence of phenomena, but it is evident that the division of
labour is carried to a high pitch amongst the phagocytes, and
that there must be some controlling influence which regulates the
appearance of the cell when it is required.

We must now turn to a discussion of the importance of these
facts in connection with immunity as it appeared to the patholo-
gists before the discovery of the antibodies. Much of this is

A '^"

l- & fi >

FIG. 50. FIG. 51.

FIG. 52.
FIGS. 50 TO 52. — FROM SCRAPINGS FROM THE LUNGS HALF AN HOUR, Two

HOURS, AND TWENTY-FOUR HOURS AFTER THE INJECTION OF POTATO

BACILLI INTO A BRONCHUS. (From films lent by Dr. Briscoe.)

The bacilli, which occurred in large numbers in the alveolar cells half an hour
after injection, are not shown.

mainly of historic importance, but it is of extreme interest, and it
is to the controversy which occurred between the cellular and
cellulo-humoral schools that we owe much of our knowledge of
the processes of inflammation and of the functions of the leuco-
cytes. This controversy was carried out with great skill on both
sides, and was the means of suggesting numerous experiments
of much beauty and ingenuity. To begin with, Metchnikoft's

25O OBJECTIONS TO THE THEORY

position was simple and logical. He pointed out that in mild
and non-fatal infections phagocytosis usually occurred, and the
bacteria could be readily seen inside the leucocytes, whereas in
fatal ones little phagocytosis took place, if any. He therefore
enunciated the paramount importance of the process in immunity,
and at one time considered it would cover the whole field of the
phenomena.

But his conclusions did not pass unchallenged, and the sup-
porters of the humoral school adduced numerous examples of
recovery from infection where little phagocytosis could be
observed, and went farther, and showed that recovery might
occur under conditions in which phagocytosis was impossible.
The best experiments of this sort were those of Baumgarten,
which were repeated by Sanarelli. These observers placed non-
virulent bacteria in the peritoneal cavities of animals enclosed in
bags of collodion or other substances which would permit the
free diffusion of the peritoneal fluids, but would prevent the access
of the leucocytes, and they found under such conditions that the
bacteria were completely destroyed. This was, of course, an
example of bacteriolysis of a type with which we are now
familiar. Other observers, including Metchnikoff himself, failed
to get these results ; but in an experiment of this sort a positive
result is of more value than a negative one. It is possible, for
example, that the walls of the bags which Metchnikoff prepared
may have been sufficiently impermeable to prevent the access of
the bacteriolytic substances. Then other observers found that
bacteria often underwent changes indicative of death and
destruction before they were taken up by the phagocytes. Thus
Nuttall found that when attenuated anthrax bacilli were placed
in a fine tube in the tissues of a rabbit's ear, the organisms showed
degeneration forms before they were taken up by the leucocytes,
and thought that they were injured by the serum before being
ingested. We have already alluded to this experiment as one of
the starting-points of the researches on the alexins. As a result
of experiments such as this, the humoralists relegated phagocytosis
to a part of quite secondary importance. They held that the
injury or death of the bacteria by the humours of the body was
the important factor, and admitted only that the phagocytes acted
as scavengers to remove the dead or disabled organisms. To
this Metchnikoff responded by allowing a leucocyte to take up a
living and virulent anthrax spore, and then isolating the leucocyte

PHAGOCYTOSIS

251

and planting it on a suitable culture medium, on which the cell
died ; but the spore survived, showing that it was taken up with-
out any previous injury. He also traced in a very clear and full
manner the steps by which a tubercle bacillus of absolutely normal
appearance, and apparently vigorous and healthy, undergoes

.d

F1G. 53.— PROCESS OF ABSORPTION OF TUBERCLE BACILLI IN GIANT CELLS.
a, Unaltered bacilli ; b, c, d, and 0, various stages in the process.

degeneration, death, and absorption in the giant cell. His case
was proved to the hilt in the case of certain bacteria, whereas his
opponents proved theirs in others. They were dealing with
immunity of different types, and the time was not ripe for the
solution of the problem.

The views of another school which sprang up at this point, and
which attempted to reconcile these two views, are of more impor-

252 CELLULO-HUMORAL THEORY

tance, in that they approach more closely to the modern theory of
opsonic immunity, and, indeed, are as close an approximation to
it as could have been formed in the then state of knowledge.
They were as follows : The importance of phagocytosis was
recognized, and it was also admitted that bacteria were frequently
prepared for ingestion by dissolved substances, but it was thought
that these substances emanated from the leucocytes. The phago-
cytes were thought to produce an alexin which injured the
bacteria, and then to devour them. Baumgarten's collodion-bag
experiments were explained by supposing that the leucocytes
which collected round the bags in the peritoneum gave off alexin,
which diffused through and was sufficient to kill the leucocytes,
though more slowly and with more difficulty than if the phago-
cytes had been able to give the coup de grace. In dealing with
organisms of very low virulence it was admitted that phagocytosis
might be all-sufficient.

Some of the experiments pointing in this direction may be
briefly referred to, though many have been alluded to before in
the chapter on the complements. Nuttall continued his experi-
ments on the destruction of anthrax bacilli by a comparison of
the action of blood and serum, and found that the latter was
enormously the more powerful ; and this he explained by the
assumption that the protective substances are given off in the
solution of the leucocytes which occurs in the process of clotting,
and many other experiments were forthcoming in support of this
view. But the most beautiful researches were those of Kanthack
and Hardy, alluded to previously, but now to be described at
greater length. When anthrax bacilli are placed in frog's lymph
and examined microscopically, the first phenomenon which occurs
is the approach of the eosinophile leucocytes to the bacilli. These
cells lose their granules, and at the same time the bacilli begin to
show signs of degeneration, the inference being that the granules
are dissolved, and that the solution acts injuriously on the
bacteria — i.e., is alexin. The next step is for the hyaline cells
to approach the area of conflict, and to fuse with the eosinophiles
to form a plasmodium around the bacilli. Then the oxyphile
cells separate themselves from the plasmodium and move away,
and then the hyaline cells can be seen to have taken up the
bacilli, fragments of which can still be seen within them. Lastly,
a number of cells with basophile granulations are attracted, but
their function is unknown. It is obvious that there is here a

PHAGOCYTOSIS 253

division of labour, the hyaline cells being the phagocytes and the
eosinophiles the mother cells of the defensive substances. The
granules may be regarded as a " pro-enzyme " stage of alexin.

It must not be thought from this experiment that it was held
that the eosinophile cells are always the cells which secrete the

V X X X \

\ \ \ \ i

"ft

S W ' ft

FIG. 54.— KANTHACK AND HARDY'S EXPERIMENT. (Original.)

1-6, An oxyphile leucocyte attacking a thread of anthrax bacilli ; the
figures were drawn at intervals of one and a half to two minutes, the
whole sequence occupying twelve minutes. 7-14, A thread repeatedly
attacked by three oxyphile leucocytes, one of which formed a plasmodium
with a hyaline cell when observations were commenced ; drawn at
intervals during a period of one hour. 15, A thread attacked by a plas-
modium, consisting of an oxyphile and a hyaline cell, the former having
lost its granules. The alteration in the bacilli, which was quite clear in
the specimens, is not shown.

The whole drawn from one preparation, the first series immediately after it
was put up, the second after half an hour, and number 15 after two hours.

alexin. This is certainly not the case in man, where these cells
play a very small part in inflammatory reactions of ordinary type.
Here we must assume that, if a similar process occurs at all, the

254 ACTION OF " CYTASE IN PHAGOCYTOSIS

injurious substance is provided by the polynuclears, which thus
play both parts.

Kanthack's experiment is the best and most direct evidence of
the extracellular injury of bacteria by substances derived from
the leucocytes occurring as a preparation for phagocytosis.

Metchnikoff resisted these views for a time, but soon had to
admit that phagocytosis is not the only factor in immunity ; and
he then altered his theory in an ingenious way, and regarded the
extracellular injury or solution of organisms as being essentially
the same process as that by which they are digested after being
taken in by the phagocytes. He considers that bacteria, red
corpuscles, etc., after being taken in, are digested by the action
of a proteolytic enzyme which he calls " cytase " — a term which
has been already alluded to as a synonym for complement or
alexin. Of this there are two sorts : macrocytase, which is
formed by the macrophages, and which digests corpuscles, cells,
etc. ; and microcytase, formed by the polynuclears, and powerful
against bacteria. Ordinarily these enzymes are restricted to the
cells which form them, and where ingested bodies are contained
in vacuoles, these latter contain a solution of the suitable cytase ;
but when solution of the phagocytes occurs the cytase is set at
liberty, and may then exert its action on cells or bacteria which
are lying free. Metchnikoff regards this as a process of much less
importance than phagocytosis, and points out that the solution
which it brings about is rarely complete : thus, when bird's
corpuscles are ingested, they are entirely absorbed, nuclei and
all ; whereas when they are acted on by a haemolysin (which
Metchnikoff regards as a macrocytase), the nuclei remain. This
is certainly true as regards the action of most sera on bacteria,
solution being rarely complete, and it is only in the highly potent
sera obtained by prolonged immunization to certain bacteria that
complete disappearance of the bacteria occurs as a result of the
action of serum ; yet when taken up by the leucocytes they are
digested altogether, sometimes with great rapidity.

The difference between these views and those of the cellulo-
humoralists is roughly this : Metchnikoff looks upon the protective
substance as a digestive enzyme which has for its object the
transformation of the foreign cells, etc., into proteids suitable for
the nourishment o£ the phagocyte ; whereas most bacteriologists
regard them as being allied to the toxins rather than to the
enzymes, and as being specially intended for the defence of the

PHAGOCYTOSIS

255

body against invaders. The point is one of theoretical interest
rather than of practical importance, and we have already pointed
out that the complement is apparently used up in its activity, and not
set free to attack other molecules, as is the case with the enzymes.
Another minor point is that Metchnikoff seems to regard the
setting free of cytase as only occurring when the mother cell is
dissolved, whereas most of the bacteriologists who admit the
origin of alexin from leucocytes regard it as a product of its
secretory activity. The point has been referred to before.
Metchnikoff explains the phenomena which occur in immunized

or.

C

FIG. 55. — PROCESS OF ABSORPTION OF ANTHRAX BACILLI IN THE LEUCO-
CYTES OF THE PIGEON. (Metchnikoff.)

(Showing various stages of alteration of the bacillus whilst in the protoplasm
of the leucocytes.)

as opposed to normal animals in this way : We will take the
absorption of bird's corpuscles from the peritoneum as an
example. When the injection takes place into normal animals,
there is no extracellular destruction of the corpuscles (haemolysis),
because there is no cytase free in the peritoneal fluid, no cor-
puscles having been broken down ; the corpuscles are taken up
by the phagocytes, but with some difficulty, since they have not
been prepared in any way for the process. When a second or
third injection is given, some haemolysis occurs, and this is
because the cells of the peritoneum are broken down by the
brusque introduction of the corpuscles ; this breaking down is
termed phagolysis, and is regarded as being a necessary pre-
liminary to the liberation of the cytase. The fresh leucocytes
which arrive now proceed to ingest the corpuscles with great

256 ANALOGY WITH DIGESTION

avidity, since they are already partially digested. We should
explain the phenomena very differently : the haemolysis is a result
of the action of amboceptor and complement, and the phagocytosis
of the action of an opsonin.

The explanation of bacteriolysis and haemolysis by means of
complement and amboceptor might appear to be difficult on the
theory of the reference of the whole process to the digestive
action of the phagocytes, but Metchnikoff has applied the
researches of Pawlow in a very ingenious way to show a
parallelism between cytolysis and digestion. It will be remem-
bered that pancreatic digestion depends upon the action of two
substances — an enzyme, protease, which occurs in the pancreatic
juice, and another substance, enterokinase, which occurs in the
succus entericus. Metchnikoff regards the protease as analoerous

0>wJrV--tX>ft&tr^

with cytase, and the enterokinase as analogous with compleitieftt
or substance sensibilatrice. Delezenne showed (though I believe
his results are not universally accepted by physiologists) that
protease has no power of attaching itself to proteids, whereas
enterokinase has such a power, and the substance thus sensitized
can then be attacked by protease. This, if true, is exactly
similar to the action of amboceptor and complement. We may
suppose, then, that amboceptor represents some substance used
by the phagocyte to assist the action of cytase or alexin on the
bacteria, etc., and normally retained in the protoplasm.

When an organism which is easy to deal with is injected it is
taken up by the phagocytes and dealt with in their protoplasm,
no preparatory action being necessary. Under other circum-
stances, when the infection is a more virulent one, some of the
phagocytes are killed and dissolved, and their digestive enzymes
escape, and partially digest the bacteria, which are then ready for
phagocytosis. When there is a balanced contest of long dura-
tion another substance is formed, which, under normal circum-
stances, is not necessary for intracellular digestion, but which
facilitates it in difficult cases ; this also may escape into the
juices, and still further facilitate the preparatory stages of diges-
tion. Lastly, as a rarity, enough of these soluble substances may
be set free to dissolve the bacteria altogether, and render phago-
cytosis unnecessary. To Metchnikoff cellular digestion and .
nutrition are the important factors in immunity ; extracellular
action is a less important and occasional phenomenon, and occurs
mainly or entirely as a preparation for phagocytosis. His theory

PHAGOCYTOSIS 257

is logical, complete, and well supported by evidence, but it does
not take into account the more recent work of Sir Almroth
Wright and his followers, and this now calls for discussion before
the role of phagocytosis in immunity can be profitably discussed
further.

It may be admitted that Wright did not discover the fact that
serum may aid phagocytosis by acting on the bacteria ; this had
been already shown by Denys and Leclef in 1895, by Mennes
and by Markl. And Neufeld and Rimpau had carefully investi-
gated the same property in the serum of animals immunized to
streptococci and pneumococci, and had described their bacterio-
tropic substances, which are apparently identical with what we
now know as thermostable opsonins. This does not detract in
the least from the credit due to Wright, who by devising a simple
quantitative method of examination, readily applicable in clinical
medicine, made a very great advance in our knowledge of the
theory of the subject, and has added a most important and useful
method of examination of the blood. The credit for the intro-
duction of the use of vaccines in the treatment of established
disease (as opposed to its prevention) is, of course, due to him
alone.

The name opsonin (opsono = I cater for, I prepare for food) is
given to substances which occur in the serum and have the power
of preparing bacteria and other cells for ingestion by the leuco-
cytes, and which are, or are held to be — for there is no absolute
proof — different from the substances which we have previously
considered. We shall discuss this question of identity or non-
identity subsequently, and shall be content at present with saying
that, whereas bacteria that have been exposed to the action of
alexin are, or may be, obviously injured, a bacterium may be
saturated with opsonin without being injured in the least, and
may still retain its viability and virulence uninjured.

The fundamental experiments of Wright and Douglas were of
this nature, and they are easy to repeat and unimpugnable in
accuracy. An emulsion of leucocytes, free from serum, is pre-
pared by receiving blood in normal saline solution containing
citrate of soda, centrifugalizing, removing the supernatant fluid
and replacing it with saline solution, mixing and recentrifugalizing.
This process must be repeated until all trace of serum is removed,
and the top layer of the deposit is then pipetted off, and will be
found to be rich in leucocytes.

17

258 FUNDAMENTAL EXPERIMENTS ON OPSONINS

The first experiment is to determine whether leucocytes thus
free from serum are able to ingest bacteria. To this end they
are mixed with an emulsion of staphylococci or tubercle bacilli,
enclosed in a capillary tube and incubated at 37° C. for a quarter
of an hour. At the end of that time the emulsion is expelled, and
films are prepared and stained in the ordinary way. It will be
found that the leucocytes have taken up very few bacteria, if any.
It is obvious, therefore, that phagocytosis goes on to a very small
extent in the absence of serum. Some species of non-pathogenic

£

"'\

' «»

"

FIG. 56. — ON THE LEFT, A PORTION OF AN OPSONIN FILM (OF PNEUMOCOCCI) ;
ON THE RIGHT, A PORTION OF A SIMILAR FILM, TAKEN FROM A PREPARA-
TION IN WHICH NO SERUM WAS USED. (Original.)

bacteria are taken up well in the absence of serum, and one micro-
coccus which I have met with was not ingested under any circum-
stances whatever.

Secondly, a mixture similar to the above is prepared, but with
the addition of one volume of serum, so that the mixture consists
of an equal volume each of leucocytic emulsion, bacterial emul-
sion, and serum. This is incubated and examined as above, and
it will be found that many bacteria are taken up ; the number
depends on the thickness of the emulsion and on the source of the
serum, but if the former be rich and the latter potent there may
be an average of twenty or even far more per polynuclear. The
bacteria which are not ingested show no signs of digestion, there
being no loss of sharpness of contour or of staining activity.

PHAGOCYTOSIS 259

It is clear, therefore, that serum has a great power in aiding
phagocytosis. Is this due to an action on the bacteria or to a
stimulation of the leucocytes ? Two experiments show that the
former occurs ; there is but little direct evidence for or against
the latter.

In one experiment Wright (having shown that the power of the
serum is destroyed by heating it to 55° C. for thirty to sixty
minutes, or to 60° to 65° C. for fifteen minutes) allowed serum to act
on bacteria, and then heated the mixture until the activity of the
serum was removed. He found bacteria thus treated were taken
up readily. This must have been due to an action which the
serum had exerted on them before it was heated, and any action on
the leucocytes is out of question, since they were only acted on by
heated and inactive serum.

Another and even better proof of the same fact may be obtained
by acting on bacteria with serum, centrifugalizing and removing
all trace of the latter by repeated washings with saline solution.
Bacteria thus treated are taken up with great readiness, and here
no free serum at all comes into contact with the leucocytes.1
Opsonin, therefore, combines with bacteria, and Bulloch showed
that this process goes on at ordinary temperatures and at o° C.2
Bacteria which have once been acted on by opsonin ("opsonized")
may be heated to 60° C. for five hours, and are still assimilable by
leucocytes ; this shows that they are profoundly affected, but they
may be absolutely unchanged in appearance.

There is no method by which an absolute measurement of the
amount of opsonin present in a specimen of blood can be made,
but comparative measurements can be made easily enough by the
process elaborated by Sir Almroth Wright. In order to do this
it is necessary to have as a standard either the serum of a normal
person, or preferably a mixture of sera from several normal
persons, so that slight individual variations or abnormalities may
be ruled out. The emulsion of leucocytes (" cream ") is prepared
as described above, and the emulsion of bacteria made by stirring
a little of a young culture of the organism in question in some
saline solution, taking care to remove clumps by sedimentation or
centrifugalization. When tubercle bacilli are being used it is
most convenient to employ dead and dried bacilli, which are

1 This experiment was performed by Markl in 1903, using plague bacilli.
- This is not altogether confirmed by Ledingham's more recent work,
which is discussed subsequently.

I7—2

260

OPSONIC TECHNIQUE

ground up in a mortar with saline solution before use. The
mixtures of leucocytes, bacteria, and serum are made in capillary
pipettes mounted with an indiarubber nipple, and furnished with
a unit mark about i inch from the free tip, which is drawn to
a fine point. The process is as follows : As much cream as will
reach to the unit mark is drawn into the pipette, then a little air
(to serve as an index), then a unit of the emulsion, another bubble
of air, and finally a unit of serum. These are then blown out on
to a glass surface, mixed intimately together, sucked into the
tube, the end of which is now sealed. The tube is now placed in
the incubator and the time accurately noted. Then the process

FIG. 57. — WRIGHT'S CAPILLARY PIPETTES, AS USED IN DETERMINATIONS
OF THE OPSONIC INDEX. (Emery's " Clinical Bacteriology and
Hsematology.")

The small figure shows the tip magnified. The middle figure shows the
pipette charged with leucocytic cream (in this case two volumes are
shown), emulsion of bacteria, and serum. In the lowermost figure these
are mixed together and the tip sealed.

is repeated in exactly the same way, but the control serum is used
instead of that which is being investigated. Each pipette is
incubated for exactly the same length of time, rerribved from the
incubator, the tip broken off, the contents expelled, and films
made. These are obtained in a suitable way, examined under the
microscope, and the number of bacteria which have been taken
up by 50 or 100 poly nuclear leucocytes is counted in each.
Thus we may find that in the control specimen (in which
healthy blood was used) there are 300 bacteria in 100 leucocytes ;
in this case we say the " phagocytic index " is 3. In the other
specimen (in which the patient's blood was used) we might find
150 bacteria in 100 leucocytes, giving a phagocytic index of 1*5.

PHAGOCYTOSIS 26l

We see in this case that the patient's blood has but half the
opsonic power of normal blood ; this we express by saying that
the opsonic index is 0-5. The opsonic index is obtained by dividing
the number of bacteria found in a certain number of leucocytes
in the films made with the patient's serum by the number of
bacteria in the same number of leucocytes in the films with the
control serum, and expresses the phagocytic power of the patient's
serum as compared with that of a healthy person. It is not
necessarily an exact measure of the amount of opsonin, since on
dilution of a serum the opsonic index falls at first slowly and then
more quickly, forming a flat-topped curve when plotted out in the
usual way (see Fig. 59, p. 265).

Other methods for the estimation of the opsonic index have
been suggested, and require some mention. In the earliest
method — that of Leishman — the patient's blood was mixed
directly with an emulsion of the bacteria in normal saline solution
in equal parts, and a drop of the admixture placed on a slide,
covered with a cover-glass, and incubated for a definite time. A
control specimen was prepared in a similar way, using normal
blood. After the incubation, films were prepared by sliding the
cover-glass off the slide, stained, and a count made as in the
method now in use. A similar but rather better method is also
employed, and is extremely convenient in some cases. The
bacterial emulsion is prepared as above, the organisms being
suspended in normal saline solution containing sodium citrate.
A mixture of this emulsion and of the patient's blood in definite
amounts (usually equal parts) is prepared, sucked into the pipette
(the tip of which is sealed), and incubated for a quarter of an
hour or twenty minutes. The process is the same as Leishman's
except that the mixture is incubated in a pipette, and not between
slide and cover-glass. A control is also prepared, using the same
emulsion and* normal blood, and is also incubated for a quarter
of an hour. At the end of this period films are prepared, and
the process finished in the ordinary way.

This method is theoretically more accurate as a test of the
phagocytic activity of the patient's blood as compared with normal
blood than is the opsonic index as determined in the ordinary
way, in which leucocytes from the same source are used in both
determinations — i.e., in that of the patient's serum and in that
of the control. Thus, if in any case the leucocytes were so
injured that they had very little phagocytic power, the opsonic

262 OPSONIC TECHNIQUE

index as determined by Wright's method might nevertheless be
normal ; yet this blood might have but little power of destroying
bacteria which gain access thereto. There is some experimental
evidence that alterations in the power of the leucocytes do
actually occur ; thus Shattock and Dudgeon, in some experiments
with granules of melanin (which, like bacteria, require to be
opsonized before they can be taken up by leucocytes), found that
either more or less might be taken up by the patient's leucocytes
as compared with normal ones, using the same serum in all cases.
The numbers varied between 0-46 and 2*9, taking the normal
number as unity. It must be pointed out that this method does
not give the opsonic index of the serum, and that in cases, e.g.,
in which a low result is obtained it affords no information as to
whether the leucocytes or the serum is at fault, or both. Further,
there is a possible error owing to the possible difference in the
number of the leucocytes in the unit volume of the two specimens
of blood. Where the patient has a leucocytosis — and this is very
common in the type of case in which opsonic estimations are
required — the difference may be very great. The result of this
has not been fully elucidated, but it is obvious that where the
bacterial emulsion is not very thick the number available per
leucocyte is very different in the two cases. This is a point
worthy of consideration in the determination of the opsonic index
by Wright's method.1 When the bacterial emulsion is very
dilute a large error is introduced, and even if very large numbers
of leucocytes are counted the results are untrustworthy. The
best results theoretically would be obtained where the emulsion
was so thick that every leucocyte would take up as many bacteria
as it was capable of doing in the given time. This is impracticable,
however, as the labour in counting leucocytes containing very
many bacteria is great, and the error in counting is also large.
Probably the best results are obtained where the phagocytic
index in the control is about 4, and it is a good plan to
perform an orientating experiment to determine the appropriate
strength of the emulsion before commencing a large series of
opsonic determinations.

1 It has been investigated by Ruth Tunnicliffe, who finds no very great
differences in a series of estimations in which the bacteria (diphtheria bacilli)
varied from 125,000 to 1,000,000 per cubic millimetre, all the other factors
being constant ; and by Walker, who finds that the index rises greatly if a
thicker emulsion is employed.

PHAGOCYTOSIS 263

Another modification of Wright's method, introduced by
Simon, concerns the method of counting only. A large number
of leucocytes are counted, and are classified simply into those
that contain bacteria and those that are free. Of course, the
emulsion must not be too thick, or practically all the leucocytes
will have taken some up. The process is repeated with the
control, and the results compared; thus, if in the control film
25 per cent, of leucocytes were empty, and in the patient's film
50 per cent., the index would be §£ = 0-5. A comparison of the
results obtained by this method and by careful counting show
that they are fairly comparable, and the process may be used
where it is only necessary to determine whether the index is
high or low.

Another and more important method is that of dilution or
extinction, as introduced by Dean and by Klien. It is especially
useful in the case of bacteria, such as B. typhosus and V. cholera,
which are dissolved by fresh serum when but slightly diluted.
Further, when an attempt is made to determine the opsonic index
to the former, and the pipette is incubated for but five minutes,
numerous shadows and partially digested bacilli are seen within
the leucocytes, thus introducing a new and very important error.
In order to avoid this, Klien determines the degree of dilution of
the serum necessary for the complete extinction of its opsonic
action. In preparations in which no serum is used the phagocytic
index is usually below 0*5, and the serum to be tested is diluted
until the degree of dilution is found, which gives a phagocytic
index no higher than this. Working by this method, Klien
obtained results very different from those obtained by Wright's
method. In the process of immunization of a rabbit the index
(by the latter method) remained low, varying only between 0-82
and 1-65, whereas by the process of dilution it was seen to be
actually greatly raised. Before the commencement of the im-
munization the opsonic power of the serum was extinguished
when the latter was diluted thirty times, whereas afterwards it
did not disappear until diluted 3,072 times. It appears clear that
in the case of bacteria like this the results obtained by Wright's
method are quite misleading. Klien states that the bacterial
emulsion should be a thick one, and should be of about the same
density in successive experiments, if these are to be comparable.
The main objection to this method is its tediousness: many
pipettes have to be prepared, and many films examined.

264

METHOD OF EXTINCTION

It has an advantage over Wright's method even in cases in
which the bacteria are not dissolved in the serum or leucocytes,
in that it provides a definite measure of the amount of opsonin
present, which the ordinary method does not do, as is shown by
the fact that the opsonic index of a mixture of equal parts of
serum and normal saline solution is more than half that of
undiluted serum.

An enormous number of opsonic determinations have been
carried out, and the results have been of extreme interest. It is

3000

2,800

2,600

2,400

2.200

2,000

1.800

1.600

1,400

1,200

1.000

800

600

400

200

IB 17 19 21 23 25 27 29 31 2 4 6 8 10 12 14

* = Leucocytes per cubic mm. (figures at right of chart).
— — — • — — - = Opsonic power of serum.

• = Bacteriolytic power of serum.

t-- — - = Agglutinative power of serum.

FIG. 58. — INFLUENCE OF INOCULATION OF TYPHOID VACCINE ON THE OPSONIC
POWER OF THE SERUM OF A RABBIT, AS SHOWN BY THE DILUTION
METHOD. (After Klien.)

found that the indices of healthy persons is approximately the
same, and does not vary much from day to day. In the case of
tubercle a very large number of determinations of the indices of
normal persons have been made, and it is found that, with one
or two exceptions, due perhaps to accidental errors in technique,
they all lie between o\S and 1-2, taking i as a standard. In reality
they agree very much more closely than this, for the great

PHAGOCYTOSIS

majority lie much nearer to i. When the estimations are carefully
carried out, very few will be found below 0*95 or above i'O5- We
may regard the opsonic index for a given organism as a definite
quantity in a healthy person. Some sera are lower or higher than
others, but the difference is but slight, and the index of the same
person is found to show but slight daily variations as long as he
remains in good health. A few observations go to show that the
index is slightly lowered in persons who, without being ill, are in a

Serum 4: 1

3:2 2:3

C4 N.Safine

FIG. 59. — SHOWING EFFECT OF DILUTION OF NORMAL SERUM ON THE
NUMBER OF BACTERIA TAKEN UP. (Original.)

state of lowered vitality, and that the onset of a mild disease,
such as a cold, may cause a fall in the index to tubercle or other
disease.

When the patient suffers from a disease due to a given organism,
and his index is tested against this organism, the results obtained
are of extreme interest. Taking the acute diseases first, we find
that as a rule the index is low at the commencement of the illness,
and that it rises, either gradually or suddenly, when recovery
takes place ; and in some cases there is a definite correlation
between the course of the index and that of the disease. For
example, Macdonald has shown that in an attack of pneumonia
the opsonic index of the patient's serum to the pneumococcus
remains at a constant low level until the crisis is reached, when it
shows a sudden rise, attaining a point above the normal level. It
remains elevated for a short time, and then relapses to normal or
below normal. It is difficult to believe that the rise in the opsonic
index and the consequent increase in phagocytosis which we

266

OPSONIC INDEX IN ACUTE INFECTIONS

should expect to be caused thereby is not the cause of the crisis
and the patient's recovery. The short duration of the high level
of the index is interesting, as we know that the immunity left
after an attack of pneumonia is but temporary.

A gradual rise of the index often takes place in staphylococcic
diseases — e.g., boils ; and when the index is traced from day to day,
it may be seen that it is low to begin with, during the onset and
increase of lesion, but that it rises more or less gradually until it

Days of disease

I 2 5 4 6 6 7 8 9 10 II 12 13 14 15

1-6 ^

1-6 /\

1-4 / \

B / V -

FIG. 60. — TYPES OF REACTION OF THE OPSONIC INDEX IN PNEUMOCOCCIC
INFECTION. (After Eyre.)

(a) Immediate, as seen in mild diseases ; (b) delayed ; and (c) progressive
decline, as seen in severe and fatal infections.

reaches a point well above normal. At the same time the disease
begins to improve, suggesting again that the phagocytosis de-
pendent on the amount of opsonin in the blood is the actual
cause of the recovery. Very many observations of this type have
now been made with many organisms, and as a general rule we
may say that in acute diseases (excluding tubercle) the index is, as
a rule, low during the onset and culmination of the disease, and
raised during involution and recovery. Exceptions may be met
with, but the sequence of events happens too often to be a mere
coincidence (see Figs. 60, 61, 63, and 64).

PHAGOCYTOSIS

267

Hence the opsonic school of immunity has formed a theory
which may be enunciated as follows : The immunity to certain
organisms (not to all) depends on phagocytosis, and this can only
take place in virtue of the preparation of the organism by the
action of opsonin. Where this substance is present in normal
amount the person is sufficiently immune to resist ordinary
infections; but if for any reason the amount is lowered or the

2

1-8
1-8
1-7
1-6
1-5
14
1-3
1-2
l-l
1-0
9
8
7
6
5
4
3
2

FIG. 61. — OPSONIC INDEX IN DIPHTHERIA. (Tunnicliffe.)

infection very virulent, phagocytosis cannot occur, and the disease
progresses. There is then a new formation of opsonin, just as
there is of other antibodies, and this goes on until there is
sufficient to sensitize all the bacteria and render them amenable to
phagocytosis, when recovery occurs. When this does not take
place the patient's phagocytes cannot ingest the bacteria, and the
disease progresses.

There is one assumption that will require critical consideration
subsequently, and that is that opsonin is an antibody.

The behaviour of the index in chronic infections is different,

268 OPSONIC INDEX IN CHRONIC INFECTIONS

and is difficult to explain on the opsonic theory of immunity. In
a chronic staphylococcic lesion, such as acne, the index may be
low, normal, or high, and this is also the case to a most marked
extent in tuberculosis. Wright classified the cases of this disease
into two groups: (i) strictly localized tubercle, such as lupus,
mild glandular cases, tuberculous abscesses, etc. ; (2) cases
associated with constitutional disturbances. In the former he
found the index uniformly low (from 0-13 to o!88), whereas in
the latter there was great variation, the index being below normal,
or as high as 2 or more. Further researches, however, have
not confirmed this, and the indices of patients with lupus will
often be found very high. As a rule, however, the patients with
localized tubercle, if kept at rest in bed, will be found to have
a constant index, whereas in those with a progressive disease it
will be found to vary from day to day, being often very high.

These variations are attributed to auto-inoculation — i.e., to the
discharge from the lesion of a few bacteria or of a small dose
of bacterial toxin, which makes its way into a region suitable
for the elaboration of a further amount of opsonin, acting just
as an injection of a vaccine, and causing a negative, followed by
a positive phase. When the patient is kept absolutely at rest in
bed this does not occur, or only to a comparatively slight extent,
and the index is more or less steady. If, however, the patient be
allowed to exert himself, even slightly, or if the lesions are
gently massaged, specific substances are set free, auto-inoculation
occurs, and the index exhibits its characteristic oscillations. It
is also dependent to some extent on the temperature, as has been
shown by Inman and others, tending (in phthisis) to fall with a
rise of temperature, and vice vevsz. In general, a fluctuating
temperature accompanies a subnormal index, a rise occurring
when the oscillations become less. The injection of a bacterial
vaccine may cause a rise of temperature, especially if the amount
is large, but does not always, and should not, do so.

In chronic infections a high opsonic index does not necessarily
imply that a patient is doing well. In general tuberculosis the
index is often normal or elevated, and a rise may occur just
before death. This is also the case in acute infections, such as
erysipelas, in which a sudden and great elevation may immediately
precede the fatal issue (Fig. 63).

These results are difficult to harmonize with the opsonic theory,
but Wright points out that it is not sufficient for there to be

PHAGOCYTOSIS

269

enough opsonin in the blood; it must reach the diseased
tissue.

Some observations go to show that it may be unable to do
this under certain conditions. Thus Bulloch found the liquor
puris from a staphylococcic abscess entirely devoid of opsonin to
staphylococci. This might have been due to absorption by the
bacteria in the pus, so he cleansed the abscess, and, taking the
first few drops which collected, found them also very deficient
in opsonin. It appears, therefore, that this substance, though
present in the blood, was unable to make its way through the

DATE

6

7

8

9

10

n

F*

104

103
102.
101
100
99
98
97
96

O.I.

2

2 i

2-

2-

2 :

A

2

n

2

/

\

I

\

1

\

j\

/

f

\

/

\

/

\

,

h

J

(

x ^

J

'

s

\

\

^

>

,

\.

/

\

/

'.

'

fi

^

\

1

\

\

/

w

--

i

*

'

'

\

•

•j

/

•

i

0

•-

!

•

j

j

\

N/

(/

3

FIG. 62. — SHOWING INVERSE RELATIONSHIP BETWEEN TEMPERATURE AND
OPSONIC INDEX IN PHTHISIS. (Inman.)

The continuous line shows the temperature.

wall of the abscess to the place where it was wanted. Again,
Wright has shown that the serous fluid in cases of tuberculous
pleurisy and peritonitis is very low in opsonin as compared with
the circulating blood, and has made use of this fact as a means
of diagnosis. It must be obvious that in the case of an extra-
vascular object like a tubercle, and especially a caseous mass,
that a slight alteration in the opsonic index of the blood can
have but a slight immediate effect ; any beneficial effect of a
high index must be slow in manifesting itself. To remedy this,
Wright attempts to flush the morbid tissues with blood or lymph

270

SPECIFICITY OF OPSONINS

by diminishing the viscosity of the blood by the exhibition of
citrates and other anticoagulants, by the use of hot applications,
and by Bier's method of congestion.

The first point which arises in a discussion of the opsonic
theory deals with the specificity of the opsonins themselves. Are
we to imagine that there is a specific opsonin to each organism,
and that during the process of immunization this increases,
whilst the others remain constant ? Unless this is the case, the
theory fails, for we know that immunity is specific.

101
100
99

2-5

1-5

Day
7 8

$ 10 II 12 13

FIG. 63. — BEHAVIOUR OF THE OPSONIC INDEX IN A MILD (a) AND SEVERE
(6) CASE OF ERYSIPELAS. (After Tunnicliffe. )

The latter shows the preagonal rise ; the broken line in the first chart
indicates the temperature.

The question may be investigated in two ways — by absorption
of the opsonins and by comparison of different sera.

The first method was employed by Bulloch and Western, who
added an emulsion of tubercle bacilli to normal serum, and found
but a slight reduction of the opsonic index to staphylococci,
suggesting the difference of the two opsonins. But these
results have not been confirmed by later writers, and it is quite
certain that a sufficient amount of tubercle bacilli will remove
practically the whole of the opsonin to staphylococci. These
experiments tend, therefore, to show that the opsonins are not

PHAGOCYTOSIS 27!

specific, and that any immunity due to them would be a general
one.

The second method is by a comparison of various sera in their
action on various organisms. For example, we may take two
sera, and compare them in their action on tubercle bacilli and on
staphylococci. If we find uniformly that a serum which is low
to one is also low to the other, it will tell strongly against the
theory of specificity. This, however, is what we do not find,
and it is quite usual to discover that a serum which is very low
to the tubercle bacillus as compared with a normal control has
a normal index to staphylococci as compared with the same
control. Of this there can be no doubt. Further, after the
injection of a vaccine composed of the dead bodies of certain
organisms, it is usual to find the opsonic index to that organism
rise, whereas to others it remains unaltered. This tends very
strongly to show that opsonins are specific bodies.

Quite similar results are seen when the behaviour of the
opsonic index to two or more bacteria is followed from day to
day in a patient suffering from an infection by one of them.
Thus, in a patient who was recovering from a severe furuncle
the index to staphylococci and tubercle bacilli was observed, with
the result shown in Fig. 64.

Here we may regard the tubercle opsonin as being normal
throughout, the slight variations met with being well within the
range of experimental error. The index to staphylococci, on the
other hand, ranged between 0*4 and 1*35, and showed a general
parallelism with the amelioration in the patient's condition. It is
obvious that the two indices are not due to the presence of a
single opsonin.

Reverting to the saturation experiments, we may perhaps
explain them as follows : Any opsonin can prepare any bacterium
for phagocytosis if it combines with it ; but there are different
opsonins, with very different degrees of affinity for different
bacteria.1 Thus we may suppose the tubercle opsonin to have a
powerful affinity for the tubercle bacillus, a slight one for the
staphylococcus, so that the addition of a few tubercle bacilli will

1 It now seems fairly clear that the explanation of these experiments is that
fixation of complement (which in this case acts as an opsonin) takes place.
Normal serum contains an amboceptor ( = thermostable opsonin) to staphylo-
cocci, though in small amount ; and this, when combined with staphylococci,
will attract all the opsonin to it, the staphylococcus opsonin most powerfully.

272

SPECIFICITY OF OPSONINS

remove it from a sample of serum, whereas a large number of
staphylococci are required. In this case opsonins will have a
sort of modified specificity comparable with that of the agglutinins
for the coli group, and this appears to harmonize Bulloch's results
with those of later observers.

An example of this selective absorption of opsonins may be
given, chiefly to illustrate the methods employed in this class

1-5
14
1-3
1-2

II

9
•8
•7
•6
•5
•4
•3
-2
•I

eo to

I

FIG. 64. — BEHAVIOUR OF OPSONIC INDEX TO STAPHYLOCOCCI AND TUBERCLE
BACILLI DURING NATURAL RECOVERY FROM AN ATTACK OF FURUNCU-
LOSIS. (Original.)

of experiment. A specimen of normal serum was mixed with an
equal amount of very thick emulsion of staphylococci, kept at
37° C. for one hour, and then centrifugalized until all the cocci were
removed, leaving the fluid (a). A second amount of serum was
treated similarly, but the staphylococci emulsion was diluted
100 times (b). It was hoped that the staphylococcic opsonin
would be completely removed from the first, and only partially
removed from the second specimen. This was tested as follows :

PHAGOCYTOSIS 273

Four experiments were carried out, in each of which the fluids
(£ unit of each) were mixed with i unit of leucocyte "cream,"
and of a fine emulsion of staphylococci, incubated, and films
prepared and counted. Thus :

Staphylococci in ,, l
50 Leucocytes.

1. Normal serum + normal saline ... ... 170 1*0

2. ,, ,, + supernatant fluid (b) ... 125 o'6g
3- ,, ,, -f- ,, ,, (*) ... 55 °'21
4. Normal saline -f normal saline ... ... 24

These fluids were then tested in exactly the same way with
regard to their action on tubercle bacilli. Thus :

Tubercle Bacilli , ,
in 50 Leucocytes,

1. Normal serum + normal saline ... ... 145 ro

2. ,, ,, -f supernatant fluid (b) ... 132 0*93
3- ,, + ,, ,, (a) ... go 0-6
4. Normal saline + normal saline ... ... 9

Here it is obvious that the staphylococci have removed the
staphylococcic opsonins more powerfully than the tubercle
opsonin. The strong emulsion removed 80 per cent, of the
former and only 40 per cent, of the latter.

A striking example of the fact that there is more than one
sort of opsonin is supplied by observations on the haemopsonins.
Most specimens of blood-serum are unable to act as opsonins
for the red corpuscles, which are not taken up by the leucocytes
under the ordinary conditions of opsonin investigation in vitro ;
but some specimens do possess the power of opsonizing red
corpuscles. It is obvious, therefore, that haemopsonin is not the
same as bacteriopsonin.

We must now discuss the nature of these opsonins. Are they
familiar substances (e.g., complements or amboceptors) mas-
querading under a new name, or are they essentially different ?
And if so, are they antibodies, or are they allied to other protective
substances, such as the alexins of the cellulo-humoralists or the
cytases of Metchnikoff ? This is an extremely difficult subject,
and one which has not yet been satisfactorily solved.

The main evidence in favour of the view that they are specific

1 The indices given are corrected by the deduction of the number of bacteria
taken up spontaneously (Expt. 4) from each of the totals. This may be
termed the corrected opsonic index, and ought to be given where great
accuracy is required.

18

274

RESULT OF INJECTIONS OF VACCINES

antibodies is derived from a study of their behaviour when a
patient is inoculated with their specific antigens. If, for instance,
a patient with a low index for tuberculosis is inoculated with a
small dose of new tuberculin (say y^^ milligramme), consisting
of the dead bodies of the tubercle bacilli, a very definite train of
phenomena, closely comparable to the results of an injection of diph-
theria toxin, is produced. In each case there is an immediate fall

30
•9
•8
•7
•6
•5
•4
•3
•2

• I
20

•9
•8
•7
•6
•5
••*
•3
•2

• I
1-0

•9
•8
•7
•6
•5

•3
•2
•I

da.1t.

-i :>-

-ri~

4 4 i-4--i i -4- 4-4- | -4-4 4-i-4-4-i-

| i : : • I :
. j ^ j,...^ j.-.^.-«^-..

1-fc'l

^ JH i M I 1 N Li i i-i l-i-l -i-4-i

rciij nijiciuiE xruinpi

"T— t~j~

..] I....1..

is |fcj n

lift! Mi :Uu ;I3|

FIG. 65. — EFFECT OF A SINGLE INJECTION OF TUBERCULIN, SHOWING THE
"FALSE RISE." (Wright.)

of variable duration, followed by a rise to a higher level than the
initial one — in other words, there is a negative followed by a
positive phase. (In some cases there is a sharp " false rise " of
short duration, which precedes the negative phase, a phenomenon
which, as far as I am aware, has not been found with the un-
doubted antibodies.) Now this rise, as has been already pointed
out, is to some extent at least a specific one ; an injection of
uterculin does not cause a rise in the opsonic index to staphy-

PHAGOCYTOSIS

275

lococci or pneumococci. There is, therefore, one important feature
possessed by the opsonins in common with the antibodies : in each
case an injection of the specific antigen causes first a diminution
and then an increase in the amount present.

There is, however, an important difference. In the other anti-
bodies— e.g., in diphtheria antitoxin — the amount present in the
blood can be raised to a point enormously above that of normal
blood by a series of inoculations of suitable doses of toxin at suit-
able intervals. Here the effect of repeated injections is a cumu-
lative one, the second raising the index above the level which it

FIG. 66. — RESULT OF A SINGLE DOSE OF STAPHYLOCOCCIC VACCINE,
SHOWING NEGATIVE PHASE. (Original.)

reaches after the first, and so on. But in the case of the opsonins
to most bacteria there is no such summation of results. An in-
jection of tuberculin may raise the opsonic index from 0-5 to 2 or
a little higher, but with a second injection it is not possible to
start with 2 as a base and raise the index to 3, and so on. The
maximum indices are not very much above the normal level. In
the case of tubercle it is very unusual to find an index as high as
2, whilst with the organisms of suppuration, etc., slightly higher
figures may occasionally be found.1 As we have already shown,
this does not prove that the amount of tubercle opsonin present

J The highest indices of all are met with in the case of the meningococcus.
I have seen them exceed 10 in patients treated with vaccine, and higher
figures have been recorded. The explanation of these figures will be given
subsequently.

1 8— 2

276 DIFFERENCE BETWEEN OPSONINS AND ANTIBODIES

in blood never exceeds twice the normal — and the actual amount
may be much more — but anything like the enormous amounts
which can be obtained when working with antitoxins or agglutinins
are never met with in the case of these bacteria at least.

FIG. 67.— SHOWING THE DIFFERENCE BETWEEN THE BEHAVIOUR OF THE
TRUE ANTIBODIES (DOTTED LINE) AND OPSONIN TO SUCH ORGANISMS AS
TUBERCLE (LOWER LINE) WHEN SUCCESSIVE INJECTIONS ARE GIVEN.
(Schematic.)

FIG. 68. — SUMMATION OF NEGATIVE PHASES IN OPSONIN FORMATION AS THE
RESULT OF INJECTIONS IN RAPID SUCCESSION. (Schematic )

When injections are repeated during the negative phase a
phenomenon of summation may be met with, as Wright first
pointed out. Here the first injection may lower the index, and
the second and third lower it still more, until a very low figure is
reached. A phenomenon similar to this may be seen after the
injections of toxins (Fig. 68).

PHAGOCYTOSIS 277

It is on these facts that Wright's vaccine therapy, or, as it is
sometimes called, opsonin therapy, is based. The object of the
treatment is to bring about an immunization of the patient by
means of an increase of the opsonin circulating in his blood, and
this is achieved by the injection of a suitable vaccine. This con-
sists in all cases of the dead bodies of the bacteria causing the
disease. In the case of tubercle Koch's new tuberculin (TR
or TE) is used in variable amount, but not usually more than
T^5- milligramme of dry material per dose. In the case of other
bacteria the vaccine is prepared by cultivating the organism on a
suitable solid culture medium, emulsifying with normal saline
solution, and heating to a temperature just sufficient to insure
sterility — usually 60° C. for one hour is requisite. The emulsion is
then inoculated on to a culture medium, incubated in order to test
its sterility, and the number of bacteria which it contains is counted,
in order to determine the amount to be used as a dose. Suitable
dilutions are then made. The dose varies with different bacteria.
Thus, with staphylococci 250,000,000 to 1,000,000,000 cocci may
be given, whereas with B. coli 25,000,000 is usually enough for
the first dose.

The treatment is controlled by a frequent estimation of the
opsonic index, and this is supposed to be advisable for three
reasons: (i) It avoids the possibility of a summation of the
negative phases, and so a worsening of the patient's condition by
lowering his immunity to the infective organism. As a rule, the
negative phase is but of short duration, but occasionally it is pro-
longed, and this is especially the case when large doses have been
given. I have seen it as long as three weeks in a case of tubercle.
(2) It enables a suitable dose to be selected. Thus, if we find a
certain number of bacteria cause a long negative phase, the next
injection should consist of a smaller one, when the negative phase
may be reduced and the rise may be greater. With a very small
dose the negative phase may be eliminated altogether, or may be
reduced so much that it is overlooked. (3) Whilst the index is
raised decidedly above normal it is assumed that the patient is
benefiting, and another injection is only required when it begins
to fall. As a rough general rule, the injections have to be repeated
at intervals varying from a week or fortnight, but individual
patients show decided differences in this respect.

Of the practical success of this treatment in certain diseases
there can be no doubt, and whatever we may think of its theoretical

278 " OPSONIN-THERAPY "

aspect, Sir Almroth Wright must receive the greatest credit for
its introduction. Before his researches the idea of injecting a
vaccine into a patient already suffering from a bacterial disease
was unthought of, although, of course, it was well known as a
method of producing immunity when disease was feared. The
question is often asked, Why inject more staphylococci into a
patient who has already too many ? The answer may, perhaps,
be as follows : The staphylococci which cause the lesion come
into contact with dead and diseased tissues only, and it is easily
conceivable that these may be very unsuitable to discharge so
vital a function as the formation of antibodies, whereas a few cocci
injected into the healthy tissues may cause a large amount. This,
however, does not explain the benefit which has been observed in
some cases of endocarditis and other haemal infections, for in them
the bacteria must be constantly gaining access to the healthy
endothelial cells, if to no others. But it is well known that not all
the tissues are equally adapted for the production of antibodies ;
thus, when diphtheria toxin is injected into the blood-stream little,
if any, production of antitoxin takes place. As a general rule,
when antibodies are required the blood-stream is the worst place
in which to inject the antigen, the serous membranes next, and
the connective tissues the best. Dr. Whitfield has suggested to
me that the reason may be that the stimulation of the opsonins
occurs best when dead bacteria are injected. Thus in the early
stage of the disease only living organisms are present, whilst later
we must suppose some are killed or die from some cause, and
then the stimulation of opsonin formation begins. The idea is
worth considering, but the subject is still obscure.

As regards the nature of these results : In tubercle, speaking
from my own experience, I can only report a moderate degree of
success, and this only in small lesions, such as tubercle of the iris
or cornea and of tuberculous ulcers. I have had but one or two
encouraging results and numerous failures with tuberculous glands,
bone disease, etc., though others have apparently been more
successful. In phthisis there appears to be some slight benefit
when combined with other treatment, and tuberculous sinuses
sometimes heal very quickly. I should only recommend the
treatment myself as an adjunct toother methods, or when surgical
interference is impossible or inadvisable.

With the diseases due to acute infections with staphylococci,
pneumococci, B. coli, and some other organisms, however, the

PHAGOCYTOSIS 279

results are most beneficial. We may often see boils apparently
on the point of bursting retrocede in a most striking manner after
a single injection of staphylococcic vaccine, and pustular acne is
often equally benefited. Localized lesions of pneumococcic origin
often clear up quickly under the action of pneumococcic vaccine :
thus a case of empyema of the frontal sinus, due to this origin and
of four years' duration, was cured in five injections, spread over a
period of about twqjnonths. Numerous cases of cure of chronic
infections of the urinary tract with B. coli have been recorded, and
some cases of gonorrhceal arthritis have been cured in a remark-
able manner. In a case of my own a patient, with five large and
numerous small joints affected, was completely cured in three
months, after having been crippled for over two years. One or two
undoubted cases of ulcerative endocarditis have been cured, and
others in which there was a haemic infection (with streptococci),
though the evidence in favour of a valvular infection is less con-
vincing. The results in cases of Malta fever are also very
encouraging. As a rule, however, we may say that the special
scope of the method is in the treatment of localized infections.

A point of great practical importance, and one that has some
theoretical interest as pointing to a high degree of specificity in
the opsonins, is the fact that good results are sometimes obtained
only when the vaccine used is from a culture of the organism in
question from the patient himself. This is sometimes seen in
staphylococcic infections ; acne is occasionally very resistant to
stock vaccines, and yields readily to treatment with an emulsion
prepared from a culture from the patient's own pus. This
phenomenon is specially marked in the case of streptococci and
B. coli.

In admitting the success of vaccine therapy, we do not neces-
sarily admit the truth of the theory on which it is based, nor the
necessity for the opsonic control of the doses. It is certainly true
in general that with acute lesions there is a low opsonic index,
and that when amelioration or cure takes place a rise to or
above normal occurs, but this is not invariably the case. Thus,
occasionally tuberculous patients improve whilst the index
remains low, and those with a meningococcal infection often go
steadily downhill whilst the index is very high, though in the
latter case the symptoms are in general more severe when the
index falls. Now it is quite true and perfectly conceivable that
the continued existence of a lesion in spite of a very high opsonic

280 OBJECTIONS TO THE OPSONIN THEORY

index may be due to a failure of the serum or leucocytes to gain
access to the lesions which are densely surrounded by inflam-
matory material. . We have adduced a similar reason to explain
the non-success of certain bactericidal sera. But it is otherwise
when we find that a patient improves when there is a low index,
for here we must admit that even this deficient amount of opsonin
is sufficient ; and, further, it is impossible to explain the formation
of- new lesions (e.g., staphylococcic) in patients in whom the
opsonic index is high — often very high — on any such grounds.
Again, a great rise in the opsonic index not infrequently occurs
just before death, as in the chart of the index in a fatal case of
erysipelas already figured. The more carefully the opsonic index
is considered, the more certain will it appear that a high index is
not an indication of immunity ; it neither proves that the lesion
is undergoing cure nor that a fresh infection will not occur. It
may, of course, occur concurrently with other properties in the
blood or tissues on which immunity does depend — indeed, since it
is commonly due to the presence of a natural or artificial vaccine,
it usually does so — but the parallelism between a rise in the
amount of opsonins and an increase in the grade of immunity is
not absolute. -Nor is a low index any proof of lack of immunity,
since patients may improve remarkably during a prolonged
negative phase. One of the most striking cases of amelioration
of a severe case of tubercle which I have ever seen occurred
during a negative phase lasting over three weeks. Allen has
noted a similar occurrence in gonorrhreal infections, from which,
however, he draws the assumption that the clinical signs are
a totally unreliable guide to the appropriate time for a fresh
injection — a deduction which is logical only if we regard the
raising of the opsonic index, and not the cure of the patient, as
the object of treatment.

It seems probable, from a consideration of the phenomena of
phagocytosis in vitro, that a very small amount of opsonin — even
less than that which is present in a serum in which the index is
very low — is quite sufficient to sensitize any bacteria that are
likely to gain access to the tissues or blood. In our laboratory
experiments the conditions are certainly much less favourable
than they are in the living body ; the leucocytes are certainly not
in the same state of functional activity as they are in the body,
and there is a limited supply of serum instead of a constant stream
thereof. In spite of this, an enormous number of bacteria are

PHAGOCYTOSIS 28l

taken up, and in some cases digested, within a few minutes. In
the body, of course, the action may go on for hours. The opsonin-
leucocyte mechanism would appear far stronger than is necessary
for the defence of the body. That it is not so indicates some
fallacy in the conclusions to be derived from these experiments
in vitro. We shall revert to this subject subsequently, and in the
meantime be content with pointing out that where a very small
amount of opsonin would appear sufficient for the resources of
the body, but little importance can be attached to small fluctua-
tions, or to a rise, e.g., from 0-8 to i.

The dread of a low opsonic index appears to have arisen on
purely theoretical grounds, and the only direct research on the
subject which seems to have been undertaken points rather in the
other direction. According to Pfeiffer and Friedberger, guinea-
pigs injected with bacterial vaccines (typhoid and cholera) do not
thereby become hypersensitive to doses of living cultures given
twelve or thirty-six hours afterwards ; on the contrary, they have
acquired an increased power of resistance, even after the shorter
period. And a very remarkable fact was noticed : this increased
resistance was not specific, since animals injected with heated
typhoid bacilli survived a lethal dose of cholera as well as of
typhoid. They conclude that the fear of a negative phase is
exaggerated ; and it must not be forgotten that the essence of
the " opsonin therapy " consists in administering a dose of vaccine,
in the first instance, while the index is low.

There is thus no direct proof that the period of the negative
phase is coincident with the period of hypersensitiveness to
infection. And when we compare it with the period of increased
sensitiveness to toxins, we find that, whereas the negative phase
comes on almost immediately, the hypersensitiveness to toxins
or tuberculin, or anaphylaxis to serum, takes some days to
develop.

Other theoretical interpretations of the undoubted good effects
of vaccine therapy are possible. Thus, a very probable explana-
tion is that it causes a local reaction in the form of an aseptic
inflammatory process in the neighbourhood of the lesion, which,
like the similar reaction caused by ultra-violet or X rays, has (in
some way not yet understood) a curative effect. The nature of
these " reactions " is considered subsequently ; in the meantime
it is sufficient to say that in the case of tubercle (and it is
probably a general effect) an injection of dead bacilli, or of the

282 ALTERNATIVE EXPLANATIONS

products thereof, causes a sharp rise in temperature and an
inflammatory process around the tuberculous focus. If the dose
of tuberculin be greatly reduced the local reaction takes place,
but there is no rise of temperature. This is best seen when
small doses of TR are used in the treatment of tuberculous
iritis, in which the iris can often be seen to become injected after
each dose ; and I have observed the same reaction in a very
marked form after the use of diluted old tuberculin in von
Pirquet's reaction. In this case the dose absorbed must have
been infinitesimal, since the temperature did not show the slightest
sign of a rise.

Other possibilities are that the vaccine may cause a general
tissue immunity, or that it may produce some degree of immunity
on the part of the leucocytes, or may at least alter them in some
way so that they are more able to perform their duties as phago-
cytes ; and, of course, other antibodies, such as antiendotoxins,
may be produced as a result of the injection, and of these the
opsonic index affords us no estimate.

In reverting to the question of the nature and properties of the
opsomns, the question of their thermo-stability first claims our
attention. The results obtained by various observers are not
quite in accord, and indicate very clearly that more than one
substance may have the same action. The opsonin present in
normal serum is in a high degree thermolabile. It is destroyed
by heating to 55° C. for half to one hour ; at 60° C. most disappears
in five minutes, the rest more slowly, little being left in fifteen
minutes. Wright and Reid, however, found that in cases of
tuberculosis some of the opsonin is more thermostable, and
whereas in heating a normal control to 60° C. for ten minutes
reduces the opsonic index to almost nothing, the same proceeding
may only lower the index of a tuberculous serum to 0-4 or so,
though the indices of both samples were formerly the same.
They suggested this as a means for the diagnosis of tubercle.
Other observers have failed to corroborate their results, and they
are certainly not true of all cases. Dean showed that in certain
sera obtained by the high immunization of animals to certain
bacteria (staphylococci, dysentery, and typhoid bacilli) there are
substances which act as opsonins, and which are thermostable.
His results have been corroborated for pneumococcic serum by
Macdonald and Rosenau, by Muir and Martin, and many others.
It is evident, therefore, that there is more than one substance

PHAGOCYTOSIS 283

which can prepare bacteria for phagocytosis. There is a thermo-
labile substance which occurs in normal serum, and a thermo-
stable one which is found in immune serum ; and this latter also
contains a thermolabile substance, since (as a rule) its index is
lowered by heat. Thermostable opsonin occurs in minute traces
in normal serum, since the index is never reduced quite to the
level seen in a control specimen made with normal saline by
heating to 60° C , and we need have no hesitation in recognizing it
as a specific antibody. It will be convenient to deal with it first,
and the question naturally arises, Is it amboceptor ? In other
words, Has amboceptor the power of preparing bacteria for
phagocytosis in addition to sensitizing them to the action of
complement ? The two substances arise under the same con-
ditions, and are identical in their power of resisting heat, faculty
of combining with bacteria, and in their specificity. The second
question arises, Assuming thermostable opsonin is amboceptor, is
the action of complement also useful in preparing bacteria for
phagocytosis, or does the process go on equally well without it ?
Now it is certain that complement is not necessary for the action
of thermostable opsonin; otherwise it would only exert its action
in a heated serum when subsequently activated by fresh serum,
and this is not the case. If thermostable opsonin is amboceptor,
therefore, it can exert its effects without the action of com-
plement. But some experiments go to show that thermostable
opsonin may be more potent when reactivated. Thus Crofton
found an antistreptococcic serum might stimulate phagocytosis
more when mixed with fresh human serum than with an equal
amount of normal saline.

Similar results have been obtained more recently by Dean,
who finds that the opsonic effect obtained by heated serum and
normal serum may be greater than the sum of the two effects
separately. The subject has been very carefully investigated by
Chapin and Cowie, who were able to avoid the possibility of
certain errors by performing their saturation experiments in a
cold room, kept at o° C. throughout the experiment. They found
that a normal human serum treated with staphylococci at this
temperature might have the whole of its opsonic power removed,
and yet would still reactivate a heated serum — i.e., the thermo-
stable opsonin combines with bacteria at o° C., and is probably
amboceptor. They found that staphylococci treated with normal
serum at o° C. and then washed are slightly more susceptible to

284 AMBOCEPTOR AND OPSONIN

phagocytosis than are normal ones, but the difference is not great.
They are, however, much more easily opsonized by normal serum,
or by serum that has had its amboceptor removed by treatment
with staphylococci in the cold.

In other cases the conditions are more complex, for when a
potent bacteriolytic serum is present, bacteriolysis may occur to
such an extent as to diminish . the number of organisms which
can be taken up by the leucocytes. We then get the " reversed
ratio " phenomenon described by Leishman and Dean. It is as
follows : Under ordinary conditions the index falls greatly on
heating, as has been shown. This is called the normal ratio.
But in some of the potent sera obtained from highly immunized
animals the opsoni^ index may apparently rise after heating to
two or three times that of the raw serum. This Dean explains
— and his explanation is an extremely rational one — by invoking
the bacteriolytic action of the unheated serum. The number of
bacteria in the emulsion is reduced, so that there are fewer for the
leucocyte to take up ; some that are not completely dissolved
may lose their power of retaining stains and become invisible ;
bacteria partially acted on may be readily digested within the
leucocyte, so that they are not counted ; and, lastly, the dissolved
bacteria may have a toxic effect on the leucocytes. The
phenomenon of the reversed ratio may be taken as an argument
in favour of the equivalence of thermostable opsonin and
amboceptor.

The strongest argument, however, is derived from the experi-
ments of Dean, who has shown that in different samples of
immune sera there is a distinct parallelism between the two
functions : when the serum is powerful as a bacteriolytic agent,
when activated with a suitable complement, it is also powerful as
an opsonin after heating. It must be admitted, of course, that a
serum may be opsonic, but not bacteriolytic ; but this is explicable
on the assumption that much less of the substance is required to
sensitize the bacterium to the attack of leucocytes than is necessary
to render it soluble by complement. This has been confirmed
by Neufeld and Bickel, who found that a very minute amount of
haemolytic serum, far less than would produce haemolysis, would
act as a haemopsonin.

The opsonic index does not rise pan passu with the bacteriolytic
power, but this is partly due to the fact that the criteria are

PHAGOCYTOSIS 285

different in the two cases. We have already shown that incre-
ments in the amounts of opsonin cause smaller and smaller
rises in the opsonic index as we proceed. There is, however,
a much closer parallelism between the bacteriolytic power and
the amount of thermostable opsonin present as shown by the
degree of dilution. This is well shown (in the case of typhoid
fever) by the chart given by Klien and inserted previously

(Fig. 58).

To sum up : Amboceptor appears to have the power of
sensitizing bacteria for phagocytosis, and this power appears to
be increased by the concurrent action of complement. Further,
there appears to be no sufficient evidence for the existence of
a thermostable opsonin apart from amboceptor, as has been
maintained by Neufeld and Hime.

(There is an additional possibility that the part of a thermo-
stable opsonin may be enacted by agglutinin.

I believe that in the case of the haemopsonins of normal
human serum the substance is a thermostable agglutinin with a
second thermolabile zymotoxic group. Natural haemopsonic sera
are, as far as I have seen, always powerful agglutinators of the
red corpuscles which they opsonize, and when they are heated to
60° C. the opsonic power is destroyed, but the agglutinative
faculty is unaltered.)

These facts may serve to explain the discordant results as to
the presence of thermostable opsonins in the sera of tuberculous
patients. It has been shown by Bruck that antibodies to the
tubercle bacillus are not always or usually present in the blood of
infected persons, and it is only when they are present that we
should find a thermostable opsonin.

If thermostable opsonins resemble amboceptor in their pro-
perties, there is an equally close resemblance between thermo-
labile opsonin and complement. Each occurs in normal serum,
and is destroyed by a short heating to 55° to 60° C. Are they the
same ?

The main fact against the theory of their identity is the
specificity, partial though it may be, of the opsonins ; for there is
no reason to think that different bacteria are attacked by different
complements, even if we accept the theory of the multiplicity of
these bodies to the fullest degree. But we have already seen
that the specificity of the opsonins is not complete, and that the

286 COMPLEMENT AND OPSONIN

whole of the staphylococcic opsonin may be removed by the
addition of sufficient amounts of tubercle bacilli.1

A second fact, closely allied and perhaps in reality identical
with the foregoing, is the rise in a particular opsonin after an
injection of a suitable vaccine, the others remaining constant.
This rise cannot be accounted for (in my opinion, at least) by
the appearance of small quantities of thermostable opsonin,
since it may occur when this substance cannot be found in the
serum.

On the other hand, there are very remarkable analogies between
the two substances. In each there is the same difference of
opinion as to whether it occurs in normal plasma, or is only
developed when clotting and destruction of leucocytes occur.
Wright and Douglas found the amount of opsonin present in
serum and in citrated plasma exactly the same, whereas Briscoe
found that very little phagocytosis took place when staphylococci
were injected into a surviving heart in which no clotting took
place. These divergencies are quite similar to those found by
different investigators in the case of complement.

Again, it has been already shown that when a blood-clot
contracts, the first serum which can be collected is poor in com-
plement compared with that which follows, and that after a time
the amount again diminishes. An exactly similar phenomenon
may sometimes, though apparently not always, be demonstrated
with opsonin (Henderson Smith). Hence an important practical
point : the patient's blood should always be collected at the same
time as the control in determinations of the opsonic index.

Thirdly, it has been shown by Levaditi that the aqueous
humour of the rabbit contains no complement and but a trace of
opsonin. But when the fluid which recollects after puncture was
examined, it was found to be rich in both substances. He found
a similar relation between the two substances in redema fluid.
As against these results we have to put the researches of Leding-

1 Since the above was written Muir and Browning have adduced very definite
evidence of a partial specificity in the case of the complements. They find
that the bactericidal action of normal serum may be due to the direct action
of complements, and that, on weakening normal serum by successive additions
of dead bacteria, the first effect is a falling off in the bactericidal action as
tested on that bacterium. Then the bactericidal action on the other bacteria is
diminished, and with a larger addition the haemolytic complement is absorbed.
This indicates features exactly like the partial specificity seen in opsonins,
and a similar absorption without the intervention of an immune body.

PHAGOCYTOSIS 287

ham and Bulloch, who found that when the number of leucocytes
in the blood was increased by injections of cinnamate of sodium,
there was an increase in the complement, but not in the
opsonin.

It may be pointed out that if opsonin and complement are the
same, we must suppose that the opsonin test is the most delicate
method of demonstrating this substance that we have, since
phagocytosis may be facilitated by substancesfwnicnjji} comple-
ments cannot be detected by ordinary tests. Further, we must
assume that it unites with bacteria direct, and sensitizes them for
phagocytosis without the intervention of amboceptor. There is
no serious difficulty in accepting both suppositions.

Lastly, Muir has shown that the substances which have the
power of absorbing complement (such as compounds of red blood-
corpuscles and their amboceptor) also remove the thermolabile
opsonin. We are forced, therefore, to the conclusion that com-
plements may play the part of opsonins. But to do this we must
necessarily broaden our ideas of the complements, and attribute
to them some degree of specificity ; otherwise the opsonic index of
any given sample of serum as measured against a given control
should be always the same, which, as we have already emphasized,
is not the case (see footnote, p. 286).

These results suggest another train of ideas as to the role of
bacteriolysis in immunity. We have already seen reason to
believe that this is not of the greatest importance, and have found
it difficult to think that so elaborate a mechanism should be of so
little apparent use. May it not be that the complements are
specially intended for use as opsonins, and that their action in
bacteriolysis is a secondary one, and comparatively of less
importance ? This, of course, is a close approximation to Metch-
nikoff's views, but there is this difference : his cytase is a
digestive ferment which, in the case of microcytase, is adapted to
attack all sorts of bacterial proteid. But with the opsonins or
complements we must assume that different molecules occur
which have different combining affinities for the protoplasm of
different bacteria, or, in other words, which differ slightly in their
haptophore groups. Yet this difference is one in degree and not
in kind, for they all have some power of uniting with all bacteria,
and a great power of uniting with the bacterium-amboceptor
combination.

On this theory the appearance of amboceptor will take on a new

288 SOURCE OF OPSONIN

significance, and we must regard this substance as a device for
attaching more complements and more varieties of complement to
an invading bacteria than can easily combine with it direct. In
other words, we must regard the cytophile group of the ambo-
ceptor as being specific, whilst the complementophile group has
the modified specificity which we attribute to the opsonins. The
presence of amboceptor will therefore enable the bacterium to be
prepared for phagocytosis by the concurrent action of many com-
plements which otherwise would only be able to attack it with
great difficulty.

And many facts, notably the liberation of endotoxin taking
place when bacteriolysis occurs, would lead us to believe that
this preparation for phagocytosis is the true function of ambo-
ceptor and complements, and that the appearance of the latter in
excess is a comparatively rare phenomenon in disease, and when
it occurs in enormous amounts (such as is seen in highly immunized
animals) is an artificial phenomenon comparable with the enormous
amounts of antitoxin seen in antitoxin -horses. Recovery from an
attack of disease caused by B. coli may occur without the appear-
ance of any amboceptor to B. coli demonstrable by ordinary tests ;
there may, nevertheless, be quite sufficient to act as a thermo-
stable opsonin. We are far from denying that bacteriolysis ever
occurs under natural conditions, but when there are plenty of
leucocytes of sufficient functional activity, it is difficult to avoid
the conclusion that they would ingest the bacteria when these
were sensitized by complement alone, or complement and a little
amboceptor, and before this latter substance had been developed
in amount sufficient to cause bacteriolysis. This latter process
may perhaps be the last line of defence, to be used only if the
leucocytes are injured by the toxins or by the high temperature,
or if they are present in insufficient numbers.

It has been pointed out already that there is some reason to
think that, whilst complement and amboceptor can each sensitize
for phagocytosis separately, they exert a more potent action when
both are present.

As regards the source of opsonin, little is definitely known.
If we regard the thermolable opsonin as identical with comple-
ment, we shall regard it as probably derived from the polynuclear
leucocytes, and this is corroborated by Levaditi's observations
on the aqueous humour. Eyre has also shown that the amount
of opsonin (to pneumococci) in the serum in pneumonia may be

PHAGOCYTOSIS

289

roughly parallel with the number of leucocytes per cubic milli-
metre (Fig. 69).

This, however, was not corroborated by Bulloch and Leding-
ham in the case of the hyperleucocytosis caused by cinnamate
of soda. But it is highly doubtful whether leucocytes hurried
prematurely from the bone-marrow, etc., are, as the result of the
injection of chemical substances, as active functionally as those
occurring normally in that situation ; and this is corroborated

FIG. 69. — RELATION BETWEEN LEUCOCYTES, OPSONIC INDEX, AND TEMPERA-
TURE IN A CASE OF PNEUMONIA. (Eyre.)

Dotted line = number of leucocytes per cubic millimetre ; thick line = opsonic
index; thin line = temperature.

by the fact that these observers found the leucocytes in question
deficient in phagocytic powers. The point is one of some im-
portance in connection with the lack of benefit which so often
follows an artificial leucocytosis brought about for therapeutic
purposes.

A few words on the subject of MetchnikofFs views on the op-
sonins may be added. He thinks that when bacteria gain access
to the blood or tissues, the presence of opsonins or other pre-
paratory substances is unnecessary, and the unaltered organisms
can be attacked by the fresh and vigorous leucocytes. The

19

2QO METCHNIKOFF S VIEWS ON OPSONINS

absence of phagocytosis in vitro in the absence of serum he
attributes to a weakening and injury of the cells, due to the
method by which they are prepared, and admits that these
weakened and altered leucocytes will ingest bacteria more easily
and more quickly if the latter have previously been prepared by
the action of serum. He admits, however, that the opsonic index
determines the defensive resources of the blood, and in doing so
would appear to range himself definitely amongst the adherents
of the opsonin theory. But there is no reason to think that
washed leucocytes are weakened in respect of their phagocytic
powers ; they can take up enormous numbers of (opsonized)
bacteria in a very short space of time, and it is difficult to believe
that they could take up more in the living body ; and if, as there
is reason to think, phagocytosis is a physical process akin to
agglutination, the functional activity of the leucocytes is a factor
of little importance in phagocytosis, though essential for the
other, equally necessary, phenomena of digestion and solution
which Jake place subsequently. Metchnikoff finds that washed
Bacteria can take up large numbers of bacteria slowly, even in
the absence of serum. This, however, proves nothing, since we
have seen there is some reason to believe that opsonin may be
formed from the leucocytes themselves. But, as a matter of fact,
the increase in phagocytosis in preparations incubated for one or
two hours as compared with those incubated for fifteen minutes
is slight as compared with that consequent on the addition of
serum.

The influence of the source of the leucocytes taking part in
phagocytosis is not yet fully investigated, and there are no facts
known at present which tend to show that those from an immunized
animal have any special powers in this direction. Bulloch showed
in a few cases that leucocytes from different sources would take
up the same number of bacteria if used with the same opsonic
sera. There are also observations tending to show that diseased
or abnormal leucocytes — e.g., those produced in excess as a result
of the injection of certain substances, such as nuclein — are deficient
in phagocytic activity.

In a very few cases some phenomena indicating an immunity,
and consequent increased phagocytic power of the leucocytes,
from an immune or infected person, as compared with the normal,
have been noticed. This, of course, is quite in accordance with
MetchnikofPs theoretical views. The examples are not numerous,

PHAGOCYTOSIS 2QI

and the best is, perhaps, that given by Bassett-Smith, who found
that in Malta fever the patient's leucocytes may be decidedly
more potent than normal ones, when used in conjunction with
the patient's serum, though, when normal serum is used, the

difference may disappear. Thus :

Cocci per Leucocytes.

Patient's serum -f patient's leucocytes + emulsion of cocci 23-0 46-0

,, ,, + normal leucocytes -f ,, ,, 8'6 25-0

Normal serum + patient's leucocytes + ,, ,, i6'o 30^0

+ normal leucocytes + ,, ,, 19*0 29-0

Rosenau has also brought forward evidence to show that
leucocytes from cases of pneumonia have greater phagocytic
powers than those from healthy persons, and are less easily
killed by heat.

Much attention was attracted of late by Bail's theory of the
aggyessins. Bail found that if washed tubercle bacilli were injected
in large amount into the peritoneum of guinea-pigs infected with
tubercle, the animals died rapidly — i.e., in eight hours or so.1
There was a fluid exudate (containing lymphocytes) in the peri-
toneal cavity, and this exudate (centrifugalized to get rid of cells
and bacteria) was found to have a remarkable action in increasing
the virulence of young tubercle bacilli to normal animals. Thus, if
a few cubic centimetres were injected together with the bacilli,
death occurred in about twenty hours, instead of in some weeks.
He found that this virulence was apparently due to an inhibitory
effect which the fluid exerted on phagocytosis. When bacilli
were injected into a normal animal without the exudate, many
polynuclears appeared in the peritoneal fluid and many large
mononuclear cells, and many of the bacilli were taken up ; but
when bacilli and exudate were injected, few cells other than
lymphocytes were seen, and there was no phagocytosis.

These observations were confirmed and extended by Bail and
others, and similar phenomena were found to occur in the case of
numerous other organisms, if not in all. A very striking example
was given by Weil in the case of the bacillus of chicken cholera,
which is extremely virulent to rabbits, so that a millionth of a
culture (containing perhaps but one bacillus) is certainly fatal.
A minute trace was injected into the pleura, and the animal died
in a few hours. Several cubic centimetres of turbid exudate, the

1 This, of course, is equivalent to the tuberculin reaction in an extreme
form.

19—2

2Q2 THE AGGRESSINS

cells of which had not taken up any bacteria, were collected, and
were found to have a most potent effect in increasing the lethal
action of the organism. This could not be tested on rabbits,
since they were loo susceptible, but in guinea-pigs it was found
to lower considerably the lethal dose. A most interesting obser-
vation was made : A guinea-pig which had received a small dose
of a culture of chicken cholera, and had apparently recovered
completely, was injected eight days after with some of the exudate,
and died of chicken cholera septicaemia, showing that the bacteria
were but latent, and had been allowed to become virulent and
active in virtue of the action of the exudate. Further, this fluid,
when injected into rabbits, was found to immunize them against
subsequent injections of the organism, even if mixed with the
exudate, and so rendered more virulent.

To these substances Bail gave the name of aggressins, and
considered them to be an entirely new type of specific substances
formed by the organism, and having the power of raising its
apparent virulence by checking phagocytosis and allowing the
invading microbe to flourish without hindrance : thus, by means
of the concurrent presence of its specific aggressin, an almost in-
nocuous organism, such as B. subtilis, becomes extremely virulent.
According to Bail and his followers, aggressins are only formed in
vivo ; but this is denied by others, who claim that a watery emulsion
of certain bacteria has many at least of their peculiar characters.

Immunity due to the injection of aggressin is supposed to depend
on the formation of a specific antibody, or anti-aggressin. It is
produced very rapidly after the injection of the aggressin, and lasts
several weeks or more, and is supposed to be due to the immediate
neutralization of any aggressin which the bacterium may form in
vivo by the anti-aggressin, so that phagocytosis in unchecked.

Agressins are sharply specific, except perhaps in the case of
those for B. typhosis and B. coli — i.e., the injection of one aggressin
will not prevent the phagocytosis of any species of bacterium
other than that by the action of which it was prepared ; hence
they are not mere leucocyte poisons : they are thermolabile.

A substance which prevents phagocytosis may act on the leuco-
cyte, the bacterium, or on the serum. The fact just described
(that phagocytosis of a bacterium A can go on in the presence of
an aggressin B) shows that the action of the aggressins is not on
the leucocytes. Further, as Weil and Nikayama have shown,
bacteria which have been acted on by their aggressins and the

PHAGOCYTOSIS 2Q3

latter removed by washing, are readily ingested. The action,
therefore, must be on the serum — i.e., aggressin must act as an
anti-opsonin.

This leads us to Wassermann and Citron's explanation of the
phenomena. They suppose aggressins to be simply solutions of the
bacterial protoplasm which have the power of combining with the
specific protective substances of the animal, and so disarming its
methods of defence. In other words, they are solutions of endo-
toxin of feeble toxicity. This view is strongly supported — indeed,
practically proved — by the researches of Doerr, who found that
aggressins caused a precipitate when mixed with their specific
immune sera, and that their presence might bring about an
absorption of the complements, just as if they were free bacterial
receptors. There are a few minor differences between aggressins
prepared in vivo and those obtained from cultures in vitro, but
not more than we might expect from the differences in their mode
of production.

If aggressins are merely free molecules of bacterial protoplasm,
we should expect them to combine with opsonins, just as do the
bacteria themselves, and hence to act as anti-opsonins. And this
supplies a striking proof of the specificity of the opsonins, for, as
already stated, an aggressin of one organism (e.g., B. coli) does
not prevent the phagocytosis of another organism (e.g., B. subtilis).
This must apply to the thermolabile opsonin, or opsonin proper,
since these experiments were made on normal animals.

The relationship between virulence and phagocytosis is an
interesting one. As a general rule, it will be found, as shown by
the extensive researches of Metchnikoff and his school, that there
is an inverse ratio between the two : when an organism is viru-
lent for an animal it will be ingested by the leucocytes to a very
slight extent, and vice versa. This refers, of course, mainly to
natural immunity, since in acquired immunity other factors, such
as the action of bacteriolysins or antitoxins, may come in. There
are, however, some exceptions. Thus, tubercle bacilli injected
into the peritoneum of normal guinea-pigs are readily taken up by
the phagocytes. We must assume in this case that an organism
may be taken up whilst it is alive and uninjured, that it may be
entirely indigestible by the leucocyte, and may continue to grow
and multiply in its interior. This is also sometimes seen in acute
infections : the common localization of the meningococcus in the
polynuclear leucocytes is well known, and Andrewes has described

2Q4 EFFECT OF VIRULENCE ON PHAGOCYTOSIS

a case of general haemic infection by this organism which ran a
rapid fatal course in spite of all the organisms (as far as could be
seen) being taken up by the leucocytes. In general, however, the
law holds good, and where there is abundant leucocytosis the
disease tends to recovery ; when there is little or none, to death.

As far as we know at present, the failure of phagocytosis which
occurs with virulent bacteria is due to their deficient opsonization ;
but whether this is because they require a large dose of opsonin
before they can be ingested, or whether the opsonin cannot com-
bine with them, has not yet been determined quite satisfactorily.
It is this resistance which very virulent bacteria exert to phago-
cytosis which causes the very high indices seen in meningococcic
infections. If the index is determined using the very virulent
organisms recently isolated from a case of cerebro-spinal fever,
very little, if any, opsonization and phagocytosis take place in
the specimen in which normal serum is used, whereas a fair
number are taken up when the serum from a patient is employed.
If, however, the index be determined using an old laboratory
culture, much more phagocytosis will be caused by normal serum,
and the index will be nearer unity. The relation between viru-
lence and lack of phagocytosis is discussed subsequently in the
section on immunity to bacteria.

Lastly, many bacteria form toxins, of one sort or another, which
prevent phagocytosis by a direct action on the leucocytes. It has
been shown that tetanus spores and bacilli, when washed per-
fectly free of toxin, are quite innocuous to all animals, and are
readily taken up by the phagocytes ; the presence of toxin, it may
be in small amounts, by killing or injuring the leucocytes, allows
the bacilli to grow in the tissues and elaborate more toxin. Similar
facts probably occur in the case of diphtheria. We have already
referred to the production of leucocidin by streptococci, and it is
obvious that when this is formed in the tissues in large amount
phagocytosis will be reduced or stopped altogether.

The nature of phagocytosis requires some discussion. We are,
perhaps, rather too apt to be influenced by the readily observed
phenomena of ingestion of bacteria, diatoms, etc., by amoebae,
and to assume that it is in all cases an active process on the part
of the leucocytes, which are usually considered to approach their
prey by active movements directed by positive chemotaxis, and
to seize them by means of their pseudopodia. Chemotaxis does,
of course, occur in the tissues, but it is clear that it does not take

PHAGOCYTOSIS 2Q5

place in the artificial conditions of opsonin estimations, where the
bacteria are uniformly distributed throughout the fluid, and there
is no reason why the leucocyte should be attracted in one direc-
tion rather than in another ; and movement of leucocytes either
does not occur at all or does so only to a very minute extent in
saline solution. It takes place much more actively in unheated
serum — a fact which gives some support to the theory of stimulins,
previously mentioned, but not discussed. It is quite possible that
all the facts related concerning opsonic action may be due to one
or more substances which occur in the serum, and which have
the power of stimulating the leucocyte, or of altering it in a
manner to be discussed subsequently. The phenomena of the
phagocytosis of sensitized bacteria in normal saline solution
would, of course, be due to a liberation of this stimulin from its
combination with the bacteria. This is known to occur, for
sensitized bacteria will yield some opsonin on prolonged soaking
in normal saline or heated serum ; the fluid acquires opsonic
properties, and the bacteria becomes insensitive to phagocytosis.
As Sellards points out, the fact that unorganized bodies, such as
carmine, particles of carbon, melanin, etc., are taken up more
readily in the presence of fresh serum is somewhat in favour of
this view. It is difficult to think that these substances are affected
in a way similar to bacteria or other antigens when combined
with their specific antibodies. There appears to be no crucial
test for determining the point.

And there is some reason for thinking that the actual process
of phagocytosis may be a physical one, akin to agglutination, and
entirely independent of any movements or other vital processes
on the part of the leucocytes. The chief evidence in favour of
this view arises from the fact that phagocytosis may occur under
conditions in which no movements of any sort take place. This
was first pointed out by Ledingham in a series of important
researches on the relation between temperature and opsonization.
He showed that when a series of opsonin mixtures were incu-
bated at temperatures varying between 18° and 37° C., the
latter temperature brought about much more phagocytosis than
the former ; and, further, that at the latter point there was very
little difference in the index between preparations incubated for
fifteen or thirty minutes, while in the former there was a long latent
period in which but little phagocytosis occurred. This he showed
to be due to the fact that opsonin combines with bacteria but very

2Q6 TEMPERATURE AND OPSONIZATION

slowly at 1 8° C. and rapidly at 37° C. Provided the bacteria
were sensitized at the latter, it mattered little or nothing whether
the mixture were incubated at either temperature, and a very
considerable amount of phagocytosis took place as low as 10° C.
Now at this point no movements of any sort occur, and it is quite
easy to satisfy oneself by actual observation under the microscope
that bacteria opsonized at 37° C. may be taken up at a low
temperature by Bacteria which remain absolutely motionless
during the process. This is even more easily observed by using
a modification of a method recently introduced by Ponder, and
of very great value in the direct observation of phagocytic and
other phenomena. If a drop of blood be placed in a glass cell
about o'2 millimetre deep (such as is used in mounting diatoms,
etc., in fluid), covered with a cover-glass, and incubated for fifteen
minutes or so, both slide and cover-glass will be found to be
dotted about with leucocytes which adhere so firmly that all the
red corpuscles can be washed off with warm normal saline solu-
tions, leaving the leucocytes adherent to the glass. If now the
cell be filled with serum mixed with bacteria, and incubated at
37° C., or with bacteria thus opsonized and thoroughly washed,
the process of phagocytosis can be readily watched, and is seen
to take place at 18° C. or lower. Under these circumstances, no
active movement or protrusion of pseudopodia takes place at all,
and it is easy to watch a sensitized coccus being gradually
attracted to and absorbed into the body of the leucocytes. The
process strongly recalls the agglutination of bacteria. A coccus
lying within a certain distance of the cell is seen, like the others,
to be in active Brownian movement, and the appearances would
suggest that it is slightly more easy for it to move towards the
cell than away from it; It oscillates in all directions, but
gradually approaches nearer and nearer the leucocyte, and is
finally taken in. Similar phenomena can be seen (using a hot
stage) when sensitized bacteria in an emulsion in normal saline
solution are added to leucocyte films at 37° C. ; and here also
no movement, or but little, takes place. If, however, serum be
added, there is usually some movement of the pseudopodia, but
little or no locomotion from place to place.

It is not easy to determine whether phagocytosis may take
place in dead leucocytes. I have not been able to detect it in
leucocytes killed either by heat or cold, but Rosenau states that
when leucocytes killed in the former way are mixed with opsonized

PHAGOCYTOSIS 2Q7

cocci, they collect round the cell, though they are not actually
ingested, and this is confirmed by Sellards. Killed leucocytes
probably undergo a sort of coagulation equivalent to rigor mortis,
which would prevent the ingress of bacteria.

Sellards has shown that salts are as necessary for phagocytosis
as for agglutination. The isotonic solution in which the
leucocytes were suspended was 5-5 per cent, of saccharose ; the
bacteria were opsonized by fresh serum, washed thoroughly, and
suspended in the same sugar solution. Little or no phagocytosis
occurred, but it took place if salts were added. This, again, does
not look like a vital process, but is quite analogous with agglutina-
tion, in which we have every reason to believe that the effect is
due to an alteration of surface tension. So also with the action of
serum in aiding the phagocytosis of substances such as carmine or
carbon. We have only to suppose that some substance is occluded
on the surface of the inert substance, the surface tension of which it
alters in the same way as opsonin alters that of the bacteria.

The degree of opsonization is determined to some extent by the
amount of salt present, and is found to be least (in the absence of
serum) in a 1*2 per cent, solution ; hence this strength of salt is
used by some observers in opsonic determinations in order to
reduce the amount of spontaneous phagocytosis as low as possible.
Hamburger and Hekma have also shown that a minute trace of
calcium chloride has a great influence in increasing the opsonic
power of the serum (we have already seen that it aids the ag-
glutination of cholera vibrios), and that the activity of the serum
is increased by alkalis and diminished by acids. Chloride of
potassium, unlike chloride of sodium, has also an unfavourable
effect on the leucocytes.

If we push our investigations a little farther, we may perhaps be
led to the belief that the amoeboid movements and protrusion of
pseudopodia which leucocytes display under suitable circumstances
may themselves be effects of surface tension rather than strictly
vital phenomena. Consider the case of the film of blood prepared
by Ponder's method, or by the use of a glass cell, as recommended
above. When this is incubated, large numbers of leucocytes
appear both on the lower and upper surfaces. Now in the latter
case the effect cannot be due to gravity, for the leucocytes are
heavier than the serum. It would be too great a strain on the
imagination to suppose the leucocytes capable of actual swimming
movements through the blood (and it may be remarked that many

2g8 PHYSICAL EXPLANATION OF PHAGOCYTOSIS

find their way to the top before coagulation occurs, though the
process appears to continue after that), and the only alternative is
a physical attraction between the glass and the leucocytes. This
is a perfectly feasible explanation, and if it is true the next stage
in the process would necessarily follow. This is the flattening
out of the leucocytes, so that they form thin plaques of very much
larger diameter than the same cells as seen in ordinary wet films.
This is very difficult to explain as any vital effect, but it is
exactly what we should expect to happen if the leucocyte (which,
like all, or almost all, forms of living protoplasm, is to be regarded
as a liquid) were pulled out under the influence of surface tension,
just as a drop of liquid paraffin is stretched out into an infinites! -
mally thin film when dropped on the surface of water.

The bizarre forms which the leucocytes assume in a preparation
made by Ponder's method, with long pseudopodia, are explicable
on the assumption that, owing to irregularities in the cover-glass,
the surface tension is not uniform in all directions, or 'that the
protoplasm of the leucocyte is not of the same degree of viscidity
throughout. Similar irregular protuberances can be produced in
globules of oil or water by purely physical means, and Pauli goes
so far as to say that " since the discovery of the amoeboid move-
ments of oil droplets, and the careful physical analysis of this
process by Quincke, the formation of pseudopodia has been
robbed of the characteristics of a specific life phenomenon, and
later investigations have shown that it is governed in all its
details by the laws of surface tension. The taking up of food and
the process of defalcation in rhizopods can also be explained in
the same way." The process of the ingestion of an opsonized
bacterium suspended in serum at the body temperature, in which
it is occasionally possible to see the protrusion and seizure of the
organism by a long, slender, and flexible pseudopodium, is
explicable as follows : Owing to the change of surface tension
induced by the action of the opsonin on the bacterium, there
is generated an attractive force which tends to draw the two
together. The leucocyte, being fixed to the cover-glass like a
sucker, does not move, but a small portion of its substance, being
liquid or semi-liquid in consistency, is drawn out until it meets
the bacterium, which is, of course, also attracted. The two meet,
and then it will be found that the organism is firmly held in
contact with the pseudopodium, so that it is not released even if
the latter be carried to and fro by currents in the fluid.

PHAGOCYTOSIS 2QQ

The effects of surface tension may also be traced in some of
the phenomena of inflammation, especially in the adhesion of the
leucocytes to the vessel wall. It has been abundantly shown
that this is due to an alteration in the latter, and it appears likely
that this is simply due to a change in the tension developed at
the surface between the endothelial lining and the serum, in
virtue of which the former behaves like the glass in Ponder's
method, attracting the leucocyte and causing it to adhere and
flatten itself out. This extension, so as to offer as large a surface
as possible, which is displayed by the leucocytes, and especially
of the polynuclears, when they come into contact with a resistant
surface, was noted long ago by Massart and Bordet, and in virtue
of it they are able to make their way through the finest pores,
even in compact bodies like bone and ivory. The remarkable
deformation in shape which leucocytes undergo in acutely inflamed
tissues is not usually appreciated. It was pointed out to me by
Whitfield, and may often be seen at the edge of the sections
where the fixation is perfect, provided the material has been
placed in the fixing fluid immediately after its excision. The
polynuclear leucocytes are often overlooked altogether, being
pulled out into long strands of protoplasm containing nuclear
filaments, giving the section a remarkable mossy appearance.
This change in the surface tension of the vessels, lymph clefts,
etc., probably plays a part of great importance in diapedesis. It
is somewhat doubtful, however, whether it can afford a complete
explanation of the phenomena of chemotaxis, in which a vital
and apparently quasi-intelligent action appears probable.

It must not be imagined that the vitality of the leucocyte is to
be regarded as unimportant in the consideration of phagocytosis
as a means of defence. Here the process has only begun when
the organism is ingested, and unless suitable digestive ferments
are secreted, the bacterium dissolved, and the endotoxin absorbed
or otherwise dealt with, the process is useless, or, by carrying
bacteria out of the lesion to other parts of the body, may even be
harmful.

CHAPTER XI
"REACTIONS" AND SIMILAR PHENOMENA

NOT long after the discovery of the tubercle bacillus Koch found
that the effects of an inoculation of living cultures of the organism
were quite different in normal and in tuberculous animals. If a
normal animal is inoculated by scarification of the skin the
wound soon heals, and in about a fortnight a hard nodule forms.
This ulcerates, and remains an open ulcer until the animal dies.
If a second inoculation be made after the first has run its course
to the stage of ulceration, the process is profoundly modified. No
nodule is formed at the site of the second inoculation, but the
tissue round the first becomes hard, dark-coloured, and finally
necrotic, and may be shed en masse and the lesion undergo com-
plete cure. Koch found, further, that this change might be
brought about by injections of dead cultures even after they had
been boiled. He found, too, that a large dose of these killed
cultures (which would cause nothing but local suppuration in
normal animals) would kill a tuberculous guinea-pig in a short
time — six to forty-eight hours — the symptoms being fever, acute
inflammation, running on to necrosis, in the region of the tubercu-
lous lesions, and in some cases generalization of the bacilli
throughout the body. When very minute doses were used he
found, on the contrary, that improvement might occur, and the
tuberculous ulcer become cicatrized over.

This was made the basis of a method for the treatment of
tubercle in man. But Koch found the use of killed cultures
inconvenient, since the bacilli wrere but slowly absorbed, and
might give rise to abscesses. He argued that the effect was
evidently due to some soluble substance which diffused out of the
bacilli, and after long research prepared the substance which is
now so familiar as the old tuberculin. It is a solution in 40 to 50 per
cent, glycerin of the soluble products of the tubercle bacillus, and
is prepared by cultivating that organism for several weeks in

300

AND SIMILAR PHENOMENA 3OI

glycerinated veal broth in a thin layer, so that there is an abundant
supply of oxygen. This culture is evaporated to one-tenth of its
volume and filtered through a Chamberland filter. There are
numerous slight modifications in the process of manufacture,
but they are unimportant.

Old tuberculin is a syrupy brownish-yellow fluid, with a faint
aromatic smell. It contains peptones and traces of other proteid
bodies, but the nature of the substance on which its extraordinary
power depends is quite unknown. It is in a sense to be regarded
as a toxin of the tubercle bacillus, but it is not a true toxin, like
those of diphtheria and tetanus, since it is practically non-toxic for
healthy animals or for man. Its injection in large quantity may
cause a slight febrile reaction, but not much more than a similar
injection of peptones, etc., from any other source. It differs, also,
in a marked degree from the exotoxins in that it is not destroyed
by a temperature of 100°, or even of 120° C. It is dialyzable.

When injected into tuberculous animals it causes the same
" reaction " as was produced by the living or dead culture, and this
in very minute amount. A dose of i milligramme will cause a sharp
reaction in a tuberculous patient, and, indeed, one-tenth of that
amount will sometimes suffice. When we consider that the
material consists mainly of the nutrient ingredients of the broth —
Koch thought that the active principle might form i per cent, of the
whole — its extraordinary potency is evident.

The phenomena of the " reaction " are as follows : There may
be, but usually is not, some inflammatory oedema at the seat of
injection. The temperature rises precipitously, often reaching
105° F. in a few hours, and falls almost as quickly. With this
there are the usual symptoms of fever, malaise, shivering, etc.
This is the general reaction. The local reaction occurs round the
pre-existing tuberculous lesion, and is best seen in lupus, tubercu-
lous ulcers, etc. Its severity depends upon the dose given. With
a small dose there is a little redness and swelling and some mild
inflammatory oedema, the whole lasting but a day or two. W^hen
it subsides the lesion often undergoes great improvement. After
larger doses the local reaction is more marked, acute inflammation
occurs, the tissues in and around the tuberculous foci undergo
coagulation necrosis, and are cast off. When this takes place in
the skin it may lead to complete cure, but in the internal organs
it is a source of grave danger, often leading to dissemination of the
bacilli and a consequent general infection. This occurred in the

302

THE TUBERCULIN REACTION

early days of the use of the fluid, when it was hailed as a specific
cure for the disease, and Koch's limitations of its use ignored. At
present it is used as a method of diagnosis, and found to be of
great value and devoid of danger if used with proper precautions.
And there can be no doubt that the bad results obtained when the

Normal
98°

FIG. 70. — SEVERE TUBERCULIN REACTION IN A CASE OF BAZIN'S DISEASE.
(Under Dr. Whitfield.)

potentialities of the substance were so little known have led to its
being unjustly abandoned as a method of cure. Properly applied
to suitable cases, it has proved of great value.

The reaction is a specific one, except that it is sometimes given
in patients with syphilis, leprosy, or actinomycosis. This is
unusual.

When patients are treated with gradually increasing doses of
tuberculin they become partially immunized, so that no febrile

" REACTIONS AND SIMILAR PHENOMENA 303

reaction is caused by large doses. Thus Wassermann records a
case in which 300 milligrammes caused no reaction. The patient
had been treated for a year, the dose being gradually increased.
It appears, too, that by careful treatment of animals an antituber-
culin can be produced which has the power of inhibiting the effects
of tuberculin in tuberculous animals. It has, however, little or
no effect in the treatment of the disease — another proof that tuber-
culin is not the specific toxin.

Quite recently important modifications have been introduced in
the diagnostic application of tuberculin. Von Pirquet's reaction,
or the cnti-veaction, is elicited by placing a drop of tuberculin
(undiluted or a 25 per cent, solution) on the skin, and performing
scarification just as for an ordinary Jennerian vaccination. It is
advisable to make a similar control scarification without using
tuberculin in order that the lesions may be compared. The
simple inoculation shows a little redness, which soon disappears.
That made with tuberculin does the same if the patient is not
tuberculous. If he is, a small red papule is formed, which increases
for three or four days and disappears in a week or so.

Calmette's method, the ophthalmo - reaction, is obtained by
instilling one or two drops of diluted tuberculin into the conjunc-
tival sac. He recommends the use of old tuberculin which has
been precipitated with absolute alcohol and redissolved in distilled
water, as being less irritating and less likely to cause a pseudo-
reaction in a non-tuberculous patient. The reaction in this case
consists of a mild attack of conjunctivitis, lasting twenty- four
hours, and accompanied by redness and swelling of the caruncle
and a small amount of mucoid secretion. The reaction should be
over in twenty-four hours, but in some cases undesirable results
have occurred from a secondary infection with organisms capable
of causing a more severe conjunctivitis or keratitis. For this
reason the test should be used with care, if at all.

These tests are most applicable in children, since in adults the
frequency of cured tubercle, also leading to hypersensitiveness,
may lead to reactions where there is no clinical tubercle, or the
patients may be more or less immune.

- Reactions are given in other diseases, the most important being
glanders. Mallein is a fluid obtained from cultures of B. mallei
exactly as tuberculin is obtained from tubercle bacilli. It is non-
toxic to normal animals, but it causes a febrile reaction in those
infected by glanders, even if there is but a small latent lesion.

304 OTHER REACTIONS

There is also a local reaction at the site of the lesion, though this
is less marked than in tubercle. There is, however, a very
marked production of inflammatory oedema at the site of the
inoculation, and this furnishes the most certain test for the
disease. A hard raised mass is formed, which increases in size
for twenty-four hours or more, becoming as large as the palm of
the hand, and persists some days. It often gradually travels
down the neck (in the horse), as if under the action of gravity.

It will be noticed that an essential feature in these reactions is
the rapid development of symptoms after an inoculation or
injection in a subject already infected with the disease. Von
Pirquet has brought forward other examples, which, if less
dramatic in their course than Koch's phenomenon, are at least
comparable with the cuti-reaction. Thus it has been noticed that
the second (Jennerian) vaccination, practised some years after the
first, runs a rapid course. Von Pirquet has shown that if a
revaccination be made a few months after the first, the reaction
takes place in twenty-four hours. It does not, of course,
develop, in the same way as a primary vaccination, but a small
papule, often surrounded by an areola, makes its appearance, and
lasts some two or three days.

Chantemesse has observed in typhoid fever phenomena similar
to those seen in Calmette's ophthalmo-reaction in tubercle. He
instils a single drop of typhoid "toxin" (obtained by cultivating
typhoid bacilli in extract of spleen digested with pepsin), and
finds that in normal persons there is but a little transient redness,
whilst in typhoid patients there is redness, lachrymation, and the
formation of a sero-fibrinous exudate. The process attains its
maximum in six to twelve hours. He makes use of this reaction
as a method of diagnosis.

The phenomena of the " negative phase," seen probably in all
antibodies, but specially studied in connection with the opsonins,
are probably similar in nature to these reactions, although the
doses of vaccines given are usually so small that the clinical
manifestations do not appear. Sometimes, however, this does
happen. Thus Irons found that a dose of 500,000,000 dead
gonococci caused no reaction in healthy persons ; but if given to
patients already suffering from a gonococcal infection, it produced
fever, pains in the joints, and general malaise. In most cases the
difference between the behaviour of a healthy and infected person
or • animal is traceable solely in the variations of the opsonic

" REACTIONS " AND SIMILAR PHENOMENA 305

index. When an ordinary dose of any vaccine is given to a
healthy person the opsonic index undergoes but slight changes,
and in particular there is no fall or negative phase. There may
be a slight subsequent rise. When the same dose is given to a
person infected with the same organism, the negative phase
(perhaps preceded by a " false rise ") is most marked, and is
followed by a positive rise, or sometimes by a series of rises and
falls, gradually dying away like a wave. Similar phenomena can
be produced in a healthy person, but here the dose must be much
larger. Evidently, therefore, the presence of an infecting agent
other than tubercle causes a condition of unstable equilibrium, in
which the tissues react in a different manner to healthy ones.
And the same condition of altered sensitiveness may persist for
long after the disease or injection of a vaccine, so that a dose of
dead bacilli that has but little action in health causes a great
output of antibodies. This reaction appears to be a general one,
occurring with bacteriolysins, agglutinins, etc.

Attempts have naturally been made to account for a phe-
nomenon so remarkable as the tuberculin reaction, and the large
number of explanations suggest that none is altogether satisfactory.
Many of them do not call for notice.

Koch's explanation, which was put forward more as a working
hypothesis than as an established fact, was this : The bacillus
formed a toxin, which diffused outwards from the colonies in the
tissues, and when in a sufficient state of concentration set up a
coagulation - necrosis going on to caseation. In the zone of
tissues just beyond this region the necrosis-producing substance is
present, but not in a sufficient degree of concentration to kill the
tissues. The injection of a little more of this substance— ?.£., of
tuberculin — is sufficient to turn the scale, and a rapid increase of
the necrosis takes place. He explains the beneficial effects of
the treatment in this wise : The necrotic tissue does not form a
suitable medium of growth for the tubercle bacillus (which is but
rarely seen in caseous material), and the extension of the process
may lead to the complete enclosure of the bacteria in dead and
altered tissues, in which they are incapable of further growth.

This theory assumes that the substance which produces necrosis
is identical with the active principle of tuberculin ; but tuberculin
in large doses will not produce necrosis in a healthy animal. It
seems also to fail to account for the remarkable rise in the
temperature, since it occurs in patients who are not febrile, as

20

306 EXPLANATIONS OF THE TUBERCULIN REACTION

we should expect them to be if tuberculin were diffusing from
their lesions.

Ehrlich's views are quite similar to Koch's, and he regards the
reaction as due to the effect of the tuberculin on tissues which are
injured by it at the time of the injection, and in which a slight
extra dose is sufficient to turn the scale.

Others have thought that the reaction is indicative of a hyper-
sensitiveness of the patient to tuberculin, using the term in the
sense in which we employed it in dealing with the toxins. This,
of course, is true, but it scarcely seems a sufficient explanation in
itself. We shall revert to the subject after giving an account of
some most remarkable discoveries that have recently been made
concerning this subject.

Marmorek holds — in all probability correctly — that tuberculin
is not to be regarded as in any sense the true toxin of the tubercle
bacillus. This is only formed when the organism is living para-
sitically in the tissues, or in artificial conditions bearing a very
close approximation thereto, not in such a simple medium as plain
broth. Tuberculin has this effect on a tuberculous animal : it
stimulates the tubercle bacilli to a sudden and energetic production
of toxin, which gives rise to the local reaction, and, passing into
the vessels, to fever and its concomitant general phenomena.
There is nothing inherently improbable in this suggestion, except
that no reason is forthcoming as to the way in which tuberculin
exerts this very remarkable action, but there is little direct evidence
in its favour. The toxin which Marmorek claims to have pro-
duced by the application of this principle is so weak as not to
be worth calling a toxin.

Wassermann and Briick point out that the extremely minute
amount which must be present in the blood at a given time leads
to the supposition that the tuberculin injected must leave the blood-
stream and become concentrated in the region of the tuberculous
focus. Thus, if a person with 5,000 c.c. of blood reacts to an injec-
tion of i milligramme of tuberculin, the dilution will be i : 5,000,000,
and they find that this dilution injected directly into a tuberculous
lesion gives absolutely no reaction. They then proceed to argue
that this attraction of the tuberculin from the blood must be due
to the presence in the tuberculous tissue of an antitoxin or anti-
tuberculin. They investigated the presence or absence of this
substance by means of the method of fixation of the complements.
Extracts of tuberculous tissues, when mixed with tuberculin,

AND SIMILAR PHENOMENA 307

acquired the power of absorbing haemolytic complements from
fresh serum, and so of inhibiting the haemolysis of sensitized red
corpuscles. Extracts of normal organs had no such power.

This is made the basis for their theory of the reaction. The
injected tuberculin circulates in the blood until it reaches the
antituberculin present in the lesions. The two combine, and in
doing so attract the complements which we must suppose to be
free in the plasma. This fixation is supposed to be followed by
cytolysis of the cells of the part. This accounts for the local
reaction. In this solution of the tissue cells products of disintegra-
tion are set free, pass into the blood, and give rise to fever, causing
the ioeal reaction. Thus neither the jbcal nor the general reaction
is due to the direct toxic action of the tuberculin itself. In this
the theory approaches somewhat to Marmorek's, and is in funda-
mental opposition to the older theories of "addition."

YVassermann and Briick bring forward an important piece of
evidence in favour of their theory by finding antituberculin present
in the serum of patients who had been treated with increasing
doses of tuberculin and had lost their power of reacting. In them
the tuberculin injected would be immediately neutralized in the
blood, and so never reach the lesion. The theory is ingenious,
and may possibly turn out to be the correct one, but there are
difficulties. Thus the authors find tuberculin as well as antituber-
culin in the diseased tissues, and it is difficult to see why the two
do not neutralize one another. And we might also ask why no
digestive phenomena should follow the union of the antituberculin
and tuberculin in the blood of injected patients, and the subsequent
absorption of the complements. Why should not the proteid
molecules be digested, liberate their products, and produce fever ?
It would seem that the antituberculin present in the lesion must
be in a state of fixation to the cells, or it must be carried away in
the blood-stream, and this, according to Wassermann and Briick)
rarely happens except after injections. But we do not know
definitely of any such antitoxin, the nearest approach to it being a
superabundance of suitable sessile receptors, which, if they occurred,
might very well make their way into the extracts used in the test,
and simulate an antitoxin. And if this were the case, there is no
explanation why these receptors are not shed in the normal tuber-
culous process, but are after the use of tuberculin. It is difficult,
too, to see why the presence of these abnormally numerous
receptors might not be made the basis for a " theory of addition "

20 — 2

308 EXPLANATIONS OF THE TUBERCULIN REACTION

without invoking the aid of the complements. But the whole
subject is theoretical to a degree, and needs, moreover, independent
experimental verification.

Bail's researches on the aggressins have been referred to already.
The application of his theory to Koch's phenomenon is obvious.
According to him the endotoxins are only set at liberty when
bacteriolysis occurs, not after phagocytosis. This is in all prob-
ability correct in most cases, though perhaps not in all. When,
therefore, tubercle bacilli are injected into the peritoneal cavity of
a normal guinea-pig and extensive phagocytosis occurs, there is
little or no febrile reaction ; but in the tuberculous animal the
bacilli produce aggressins, which paralyze the phagocytes, and,
wrhen a second injection is made, the bacteria undergo rapid
bacteriolysis, endotoxin is set free, and rapid death follows. Bail
compares the results of the solution of large quantities of cholera
bacilli in an immunized animal with what is seen after the injec-
tion of tubercle bacilli along with a small amount of the peritoneal
exudate from a tuberculous animal. In each case there is extra-
cellular bacteriolysis, and death in a few hours, obviously from the
toxin set free.

This might account in a satisfactory way for Koch's phenomena
when caused by the injection of cultures, but seems to fail utterly
when applied to the tuberculin reaction ; for tuberculin is neither
an " aggressin " in Bail's sense, nor an endotoxin of the tubercle
bacillus, and it cannot undergo bacteriolysis. The theory resembles
that of Marmorek.

Von Pirquet's explanation of the early reaction after Jennerian
vaccination calls for some notice, though it is not immediately
applicable to Koch's phenomenon. It introduces some new and
interesting conceptions. According to this author, the result of
an infection is to alter the way in which an animal reacts subse-
quently to a second infection with the same organism. This he
calls allergia. In some cases this may lead to hypersensitiveness,
but in the majority it leads to a temporary immunity, followed by
a condition in which the animal is no longer immune, but possesses
the power of forming antibodies in the region of inoculation more
quickly and easily than a normal one can do. When a second
inoculation is made, the bacteriolysins present in the blood may be
sufficient to destroy the bacteria introduced, setting free their
toxins, which act locally and cause the early reaction. Or this
may be delayed until local antibodies are formed. This occurs

"REACTIONS" AND SIMILAR PHENOMENA 309

more quickly than in the normal person. It leads to an early
development of the specific lesion of vaccinia.

The essential point of this theory is that an infection from which
recovery has taken place may lead to an alteration of the facilities
with which antibodies may be formed, which alteration persists
for a long time.

It seems desirable here to make a further reference to the
subject already mentioned briefly as " hypersensitiveness to
toxins," but now more generally termed anaphylaxis— i.e., the
opposite of prophylaxis. The term was introduced by Richet,
who studied especially the poison of the actiniae, which he found
to be extremely powerful, the lethal dose being about -009 gramme
per kilo of body-weight. He found that a non-lethal dose increased
the susceptibility of the animal to a second injection, and that this
hypersensitiveness might last as long as six months after the first
injection. This, of course, is quite similar to the phenomena we
have described in connection with diphtheria and tetanus, which
renders it so difficult to immunize small animals to these sub-
stances, and which is the cause of much danger in the early stages
of antitoxin formation in the higher animals. Richet has also
studied the poison formed by the common mussel, which he calls
" mytilo-congestine," and finds exactly similar facts; indeed, it is
probable that it is a general phenomenon of all the poisons which
can act as antigens. In the case of mytilo-congestine the measure
of the hypersensitiveness is simple, since one of the most constant
symptoms of its action is vomiting, which occur? almost as soon
as the injection is made. He finds that in an animal which has
previously been injected the emetizing dose is from a tenth to a
quarter of the amount originally necessary. Richet has elaborated
a theory to account for this phenomenon, and for anaphylaxis in
general. He holds that the condition is due to the presence in the
blood of a toxogenic substance, which gives rise to a poison after
reacting with the mytilo-congestine injected. This toxogenic sub-
stance is not formed immediately, for Richet does not find hyper-
sensitiveness to come on for five or six days, and it persists for
some fifty days, that being the average duration of the state. He
holds that the animal produces antitoxin also, but more slowly.
When the toxogenic substance has disappeared the antitoxin
remains, and the animal is immune. The main evidence in favour
of this theory is the fact that the serum of an anaphylactic animal
will produce a similar condition in a second animal. Currie has

3IO HYPERSENSITIVENESS OR ANAPHYLAXIS

enunciated a theory very like this in regard to serum anaphylaxis,
to be described shortly.

Other theories might be cited, but there is only one which gives
an explanation which is at all satisfactory without introducing many
unproved suggestions. It was introduced quite recently by Good-
man, and proceeds on lines somewhat similar to those we followed
when dealing with the question of immunity to toxins. The cells
of the body maybe classified into three groups: (i) The nerve
cells essential to life, and with a high degree of affinity for toxin ;
(2) cells not essential to life, but with a higher degree of affinity
for toxin than the nerve cells possess ; and (3) inert cells without
susceptibility to toxin. If a dose of toxin be injected, the second
class of cell will have its receptors satisfied first, and any toxin
which is left over will then attack the nerve cells, which we
assume to be the only region where it will do harm. A lethal dose
of toxin, therefore, is the amount which will satisfy the receptors
of the second group of cells and leave enough toxin to injure the
nerve cells sufficiently to cause death. Now if a first injection
just sufficient to combine with the receptors of Group 2 were
given, a very small additional amount would be sufficient to cause
death, since it would go straight to the nerve centres. So far the
theory is unsatisfactory, since it is simply a theory of summation,
and the total amount necessary to cause death, if given in divided
doses, should together form the amount necessary if given in one
dose, which is very far from being the case. We have seen that
TJ_ of the " lethal dose " of tetanus toxin may cause death if given
in divided doses. To account for this Goodman supposes that the
toxin which combines with the non-essential cells may cause a sort
of spreading necrosis of the receptors of the latter, or may interfere
with their nutrition ; in either case more of these receptors may
be destroyed than the toxin actually combines with. If we can
imagine one molecule of toxin destroying ten receptors, the animal
would become as susceptible as if ten times the dose were given
at once. Put in another way, if it takes x molecules of toxin to
satisfy the receptors of non-essential cells, and }( molecules to
combine with those of the central nervous system and kill, then
if in the sensitizing dose each molecule of toxin destroys ten
receptors, the lethal amount necessary for a second dose would be

but — + a. x we must suppose much larger than a.

He compares this process with the injury to the excretory

"REACTIONS" AND SIMILAR PHENOMENA 311

organs which often follows the action of poison. The kidneys,
etc., excrete the substance, but in doing so are injured, and a
smaller dose of the poison may now produce a great effect, since
it cannot readily be eliminated.

The main objection to this theory is that it is difficult to
imagine such a selective destruction of the receptors as seems
necessary to account for the fact that the hypersensitiveness is
specific. We should expect the creeping necrosis or interference
with nutrition to act more generally, so that an animal highly
sensitized to one toxin would show some degree of sensitiveness
to others. It seems also inadequate to explain the facts of serum
anaphylaxis, which will now be described, since here the animal
is sensitized with minute amounts of a substance which causes no
toxic symptoms in comparatively enormous doses in a normal
animal. Here the anaphylaxis appears to be the production of a
new sensitiveness rather than the exaltation of one previously
existing.

There are two of these phenomena of hypersensitiveness to
serum — Arthus' phenomenon and Theobald Smith's phenomenon,
both of which are referred to as " serum anaphylaxis." The latter
is the more important.

Arthus' phenomenon appears when a guinea-pig receives
several injections, at intervals of a few days, of normal horse
serum, a substance which in itself is scarcely more toxic than
normal solution. After a few such inoculations the animal becomes
hypersensitive, or anaphylactized, and after another injection an
cedematous mass, an aseptic abscess, or an area of necrosis,
appears at the site of a new inoculation, which need not be in a
region in which a previous injection has been made ; the altera-
tion is a general, and not a local, one. After several of these
injections the animal becomes cachectic, and dies after several
weeks. An animal thus sensitized will die rapidly after the
injection of 2 c.c. of serum into the veins.

It should be noticed that these results are not due to the accu-
mulation of the horse serum in the system, since they may be
brought about by the injection in divided doses of an amount
which an animal can stand with impunity if given in a single
dose.

Theobald Smith's phenomenon occurs when an animal has been
sensitized by a very small injection of horse serum (TJ^ c.c., or
even as little as T^nro^nnr c>Ct' an almost inconceivably small

312 THEOBALD SMITH'S PHENOMENON

amount to produce so great an effect), and kept for a fortnight or
more. If then a second injection of a larger amount of the same
serum be made (^ c.c. or more, the usual testing dose being 5 c.c.),
the animal develops a series of remarkable symptoms, the most
noteworthy being respiratory failure, paralysis, and clonic spasms.
Symptoms usually appear within ten minutes, and death occurs
within an hour. Death does not always follow. The less sensitive
the animal, the later the development of symptoms (which in
highly sensitive animals come on within ten minutes), and the
greater the chance of survival. The process evidently affects the
nervous system in a very special way, and the heart may continue
to beat for an hour after death. In some cases, but not in all,
there are definite haemorrhagic lesions present ; they usually occur
in the stomach, less frequently in the caecum, lungs, spleen,
adrenals, or other parts.

The phenomena had often been seen in the process of testing
diphtheria and other antitoxins for the presence of free toxin, in
which several cubic centimetres of the serum are injected intra-
peritoneally into guinea-pigs. Animals that have been previously
used for the standardization of the antitoxin are often employed,
and as these have received minute doses of the latter substance
they may be hypersensitive. The phenomenon is a familiar one,
but it is only recently that its true method of origin has been
apparent. It has no connection with the antitoxin as such, and
the same phenomena of hypersensitiveness may be produced by
means of egg-albumin.

The action is to a certain extent a specific one. An animal
sensitized with horse serum is less susceptible to the serum of the
cow, pig, sheep, etc., than to that of the horse. It may show
symptoms after the injection of one of these heterologous sera,
but usually recovers. And the same is true .for an animal sensi-
tized by small doses of another sera. Symptoms are not usually
produced by horse serum, and if they are, are not fatal. Animals
can be sensitized by feeding with horse serum or with horseflesh.
Rosenau and Anderson thought that children might be sensitized
in this way, and so develop toxic symptoms after the use of anti-
toxin, but abandoned the idea.

Otto and Rosenau and Anderson thought that small doses were
necessary for the production of this form of hypersensitiveness,
large ones appearing to bring about immunity ; but Gay and
Southard show that large doses simply delay the incubation

" REACTIONS " AND SIMILAR PHENOMENA 313

period. After an injection of TJ^ c.c. or Ti)ir c.c. the animal is
hypersensitive in a fortnight or less, whereas after a dose of 8 c.c.
the sensitiveness does not reach its maximum for some forty-five
days. The duration of this anaphylaxis is not exactly determined,
but it certainly lasts several months.

Gay and Southard further found that during the period of in-
sensitiveness which follows a large dose the animal actually
contains the substance which acts as a sensitizing agent. Thus a
guinea-pig which had received (in divided doses) 17 c.c. of normal
horse serum was bled fourteen days after the last dose : 1*5 c.c.
of its serum was found to sensitize a normal guinea-pig, so that it
died in ninety minutes after an injection of normal horse serum.
(Rosenau and Anderson had already found that the young of
sensitized animals are also sensitive.) Further, Gay and Southard
found the sensitizing substance present in the blood of sensitized
animals. Thus a guinea-pig received TJ<j- c.c. of horse serum,
and after twenty-nine days was bled, and 1-5 c.c. found to sensi-
tize another animal. The first pig was tested and found to be
sensitive.

These discoveries are sufficiently astonishing, and there appears
to be no satisfactory explanation for them. The period of incuba-
tion would suggest that we are dealing with antibodies of some
sort, but there is no evidence to show that the precipitins play any
part in the process. An animal may be highly sensitive when no
precipitin can be demonstrated in its serum.

Gay and Southard think that the hypersensitiveness depends on
the presence of a substance which occurs in horse serum, and
which they distinguish by the name anaphylactin.1 This they do
not consider to be the same as the toxic ingredient of the serum,
for this reason : a sensitized animal will not develop symptoms
after the injection of blood from an animal in the refractory stage,
though this, as we have seen, was sufficient to sensitize it. (The
injections were made into a vein, the sensitive animal reacting to
very minute doses administered by this channel.) They failed to
find any proof of the formation of antibodies in the animal during
the refractory stage. They regard the reaction as being one of
the cells of the sensitized animal, and as being due to a heightened
power of absorption of the " toxic substance " on the part of cells
which have been exposed for a certain period to the action of the
anaphylactic substance. The action of this toxic substance
1 Analogous to Richet's toxogenic substance.

314 THEOBALD SMITH'S PHENOMENON

appears to be to make the fatty constituents of the cells flow
rapidly together. This occurs especially in the endothelium of
the capillaries, and leads to haemorrhages.

This can hardly be regarded as a full explanation, and fails,
moreover, to explain a most remarkable fact discovered by
Besredka : that sensitized animals do not react to the injection of
horse serum if previously anaesthetized with ether. No theory
that has yet been suggested will explain all the extraordinary
phenomena connected with this subject, and no good purpose
would be served by discussing others that have been brought
forward. We seem to be on the eve of a series of discoveries
that may prove to have as great an effect on our ideas of immunity
and cell nutrition as the discovery of the antibodies themselves,
and until the facts are better known hypotheses seem out of
place.

We may, however, be permitted to point out that it is quite
possible that the tuberculin reaction may be a phenomenon of
exactly the same order as the serum anaphylaxis of the guinea-pig.
Tuberculin is in itself non-toxic for a normal animal, just as is
serum, but we may imagine that the tubercle bacilli in the lesions
produce it in small amounts until the animal becomes anaphy-
lactized, in which case it will react violently to a small injection.
The wasting and fever which accompany the evolution of the
disease may be phenomena akin to the phenomenon of Arthus,
and simply indicate the reaction of a hypersensitive animal to
repeated small doses of a non-toxic substance. If this is the case
we may regard tuberculin as being, after all, the true toxin of the
disease, though non-toxic to animals unless the processes of tissue
nutrition and metabolism have been profoundly modified by the
prolonged action of minute amounts of the same substance. That
a reaction does not occur in a normal person after two injections
of tuberculin does not surprise us, for the active principle is (as
shown by the fact that it dialyzes) of small molecule, and may be
eliminated more quickly than the large-moleculed toxic ingredient
of horse serum, so that for hypersensitiveness to occur it may be
necessary to have a continued stream of minute amounts from a
lesion in the tissues. It may be pointed out, however, that occa-
sionally a local reaction may be seen round the spots at which
tuberculin has been previously injected, suggesting that the tissues
in this region may be anaphylactic.

There appears to be no connection between this phenomenon of

" REACTIONS " AND SIMILAR PHENOMENA 315

serum anaphylaxis and the few cases that have occurred of sudden
death after the injection of antitoxin, since these have usually (not
invariably) occurred after the first dose. In some cases, if not
in all, they have happened in children the subjects of the lymphatic
diathesis, who are subject to sudden death on very slight provoca-
tion. And the remarkable lesions recorded by Gay and Southard
have never been seen after death from a single dose of antitoxin ;
nor do the extraordinary symptoms develop.

This seems the most suitable place in which to discuss the
" serum disease," or series of unpleasant though transient pheno-
mena which may occur after the use of antitoxin, though it must
not be considered that it is necessarily a phenomenon having any
connection with those that have been already described.

The symptoms have been carefully analyzed by von Pirquet,
and the following account is based very largely on his observations.
The most constant symptom is fever, which is usually remittent
in type. Next in frequency is a rash, which may be general or
confined to the vicinity of the region at which the injection was
made. The type of the rash varies in different cases, but all
forms are associated with the most unpleasant symptom of the
serum disease — namely, severe itching. The types of rashes are
those included under the term "erythema multiforme." The
severest, which is associated with the greatest degree of fever,
is the morbilliform ; the scarlatiniform is associated with less ;
and the urticarial rash, which is the commonest, is usually accom-
panied with but little elevation of temperature.

There is usually enlargement of the lymphatic glands corre-
sponding to the region where the injection was administered, and
this enlargement may be the first sign of the onset and of the cure
of the disease. The pains in the joints often form one of the most
unpleasant of the symptoms, and the articulations usually affected
are the metacarpo-phalangeal, wrist, and knee, in that order of
frequency. Albuminuria, not going on to nephritis, is occasionally
present, and slight oedema is common. The severity of the
disease is roughly proportioned to the amount of serum injected.

According to von Pirquet and Schick the leucocytes are
increased in number until the symptoms develop, when a rapid
fall (due chiefly to a decrease in the polynuclears) takes place.

The disease is painful, but not dangerous, complete recovery
invariably occurring. The few cases of sudden death that have
been recorded as taking place after an injection of serum appear

3*6 THE

to be quite different in nature, and in many cases were dependent
on the " status lymphaticus," or " thymicus," a condition in
which a slight disturbance of any kind may cause sudden
death.

In general terms, the severity and frequency of the serum
disease are proportionate to the amount of serum given — the larger
the dose, the greater the likelihood of its development and the more
severe the disease. There are, however, some exceptions :
(i) Certain samples of serum appear more potent in this respect
than others. The purified diphtheria antitoxin prepared by
Gibson (which consists of a solution of globulin) appears to
reduce the occurrence of the disease to a minimum. (2) The
serum disease is more likely to occur when a second injection is
given at an interval of three or four weeks or longer after the
first.

Under ordinary circumstances the disease manifests itself after
an interval of eight to twelve days, or sometimes longer, after the
injections. This period is insignificant, since it approximates
closely to the period of maximum development of most of the
antibodies after an injection of an antigen. Hence it was
suggested (by Hamburger and others) that the symptoms might
be due to the development of a precipitin, which, by combining
wTith the unaltered horse serum still present in the patient's blood,
might cause the production of small precipitates in the circula-
tion. These, being deposited in the minute capillaries of the
skin, joints, etc., might act as emboli, and cause the characteristic
symptoms. Now the blood of a patient who has been treated
with serum is frequently found to contain precipitin after a week
or ten days, so that the possibility of this explanation is obvious.
The phenomenon of the accelerated reaction also appears to lend
it support. A patient who has been injected with horse serum
may be presumed to have precipitin persisting in his system for
some unknown but possibly lengthy period afterward, and on the
injection of more horse serum (its antigen) might cause a precipi-
tate which would lead to an immediate or accelerated development
of the serum disease. It is found in practice that this immediate
reaction does occur, but is comparatively rare. In ninety cases
where a second injection was given the disease was developed
within six hours in nine ; in thirty-nine others it was produced
within a period varying from nineteen hours to five days (Goodall).
The " immediate effect" described by Goodall differs somewhat

"REACTIONS" AND SIMILAR PHENOMENA 317

from the usual " accelerated effect " occurring in patients injected
with serum for the second time. The symptoms are rigor,
pyrexia, vomiting, rash, and collapse, whereas the effects which
develop in a day or two consist of rash, joint pains, and pyrexia.

The precipitation theory is now generally abandoned, since
further investigation shows that there is no close parallelism
between the occurrence of the disease and the presence of
precipitins in the serum. Thus it is often found that the symptoms
are present when not the slightest trace of antihorse precipitin is
demonstrable. On the other hand, precipitins may occur when
the disease does not develop, though this is a fact of less impor-
tance, since it is necessary, on this theory, that the precipitin and
the unaltered horse serum should be present simultaneously, and
this latter fact is usually impossible of proof. But Widal and
Rostane have brought forward very strong negative evidence in
showing that the disease does not necessarily occur after the
intravenous injection of a powerful antihuman precipitating
serum in man.

These arguments, though strong, cannot be regarded as entirely
conclusive. As regards the absence of the precipitin in some
cases of serum disease, it may be pointed out (a) that it may have
all been removed in the form of precipitum when the sample was
collected, and (b) that the demonstration in vitro of a minute
amount of precipitin is by no means easy, and requires an appro-
priate correlation between the precipitating and normal sera, or
the phenomenon may be missed. Further — and this applies to the
negative experiments of Widal and Rostane — we do not know
exactly what conditions are necessary to the union of precipitin
and antigen to form an insoluble precipitate, and whether these
are always present in vivo. Some experiments, it is true, appear
to show that the combination never occurs in the animal body,
but they are inconclusive, and the supposition is in the highest
degree unlikely. It appears probable that precipitation, like
agglutination, depends on the presence of certain salts, as well as
on that of the antibody and antigen, and it may be that these are
sometimes absent or not present in an available form. In this
connection we may quote the researches of Netter, which are of
great practical importance. He showed that the administration
of calcium lactate at the time of the injection, and for a day or
two after, caused a great diminution in the number of cases of the
serum disease. His results have been generally substantiated,

318 THE

and, in addition, the drug has been found to be of decided
therapeutic value. Hence it appears that a diminished calcium
content of the patient's blood at the time of the injection may be
of some importance in bringing about the conditions necessary for
the development of the disease. If this is the case, it might
account for the symptoms in some cases and not in others,
although precipitin might be produced in both. And there are
probably other factors which are at present unsuspected, or the
use of the calcium salt would be efficacious in all cases.

The alternative theory is, of course, that the patient gradually
becomes hypersensitive to the serum, and that the grade of
hypersensitiveness necessary for a reaction is reached before the
serum is eliminated from the blood, or, in the case of an im-
mediate or accelerated reaction, has not passed off before the
second injection is given. It would make it a phenomenon
analogous to those of Koch or Theobald Smith.

A good argument in favour of this theory is the long interval
which may occur between the first injection and the second in the
case of an accelerated reaction. This may be a year or more — a
very long period for antibodies to persist in the blood after a single
injection of antigen.

CHAPTER XII
COLLOIDAL THEORY OF ANTIBODIES

THE fact that antigens and their antibodies are, as far as is known
at present, exclusively colloid in chemical character renders it
advisable to glance briefly at some of the main facts and theories
concerning these bodies and their reactions. They are substances
of the greatest possible interest to the biologist, since the living
tissues and the fluids of the living animal are, without exception,
colloids and colloidal solutions ; protoplasm, cell nuclei, etc., being
mixtures of colloids in a jelly-like form, whilst the fluid part of
blood-lymph, etc., is a colloidal solution. It is necessary to
remember this, since the reactions of colloids and of crystalloids
in presence of colloid appear to follow laws which are different
from those governing the reactions of crystalloids alone, and we
may doubt whether these latter ever take place in the living body.
Colloids are not necessarily organic bodies, since metals such as
gold and silver may be obtained in the colloidal state, as well as
many metallic salts, such as ferric oxide, silicic acid, etc. These
simple colloids appear to be governed by laws similar to those
concerned in the reactions of the organic colloids formed by the
action of vital processes.

The chief features which distinguish colloids from crystalloids
are — (i) that they do not undergo dissociation into their ions when
dissolved in water ; and (2) that this so-called solution is not a true
one, but merely a suspension of unaltered molecules or groups of
molecules. These two characters are probably fundamentally the
same. A colloid solution, therefore, is merely a very fine emulsion
or suspension of particles of the substance, and Siedentopf and
Zsigmondy have demonstrated a method by which these particles,
though infinitely small as compared with the most minute bacteria,
can be rendered visible and their numbers estimated. The method
is simple enough in theory, though the actual arrangements are
somewhat complicated. A powerful beam of light is passed trans-

319

32O NATURE OF COLLOIDAL SOLUTIONS

versely through a true solution placed in the field of a powerful
microscope, the ray passing at right angles to the optical axis.
No light enters the lens, and the whole field remains in darkness.
If, however, a colloid solution be similarly treated, these particles
will reflect the light in every direction, and some rays will enter
the lens. The result is that the particles are seen as minute
luminous points on a dark background. The process is analogous
with the examination of the stars with a telescope, wrhich, no
matter how powerful, never shows the star as a disc, but, merely
by collecting more light, renders it brighter, and shows stars so
small as to be invisible to the naked eye.

The particles in question are so small that they are prevented
from falling to the bottom of the fluid by friction ; if, however,
they clump together, the particles become heavier, whilst the
surface which they expose to the fluid does not increase at so great
a rate, until at last aggregates of molecules are formed, which fall
more or less rapidly. This is what takes place in coagulation of
a proteid, whether brought about by heat or by the addition of
electrolytes, etc. The factors which inhibit and which cause this
clumping of molecules thus come to be of vital importance. The
force which tends to cause this clumping is probably surface
tension (Hardy, Bredig, Perrin), which is developed at the junc-
tion of the molecule and the water, and which, as explained in the
section dealing with agglutination, tends to draw together any
particles within a certain distance of one another, so that the
surface may become as small as possible. This, of course, is not
a peculiarity of colloidal solutions, but occurs in all emulsions of
insoluble particles. Surface tension, therefore, is constantly tend-
ing to make the particles of colloid in a solution approach one
another, form aggregates, and so cause precipitation or coagula-
tion.

This action is counterbalanced in a stable solution by a force
of electrical repulsion. The colloid particles all carry a feeble
charge of positive or negative electricity, and therefore tend to
repel one another. The existence of this charge and its nature
can be shown by passing an electrical current through the fluid.
Under these circumstances the colloids do not undergo dissocia-
tion with passage of the + ions to the - pole, and vice versa, like
the electrolytes ; instead, the molecules pass bodily to one or other
pole, according to the sign of the electric charge they carry. Thus
Field and Teague find that antitoxin is carried towards the cathode,

COLLOIDAL THEORY OF ANTIBODIES 321

and must therefore carry a positive charge. Further researches
show that absolutely pure colloids, free from all traces of electro-
lytes, carry no charge at all, and are not conveyed by an electrical
current. The charge depends upon the nature of the ions present :
traces of acid and of acid salts give it a positive charge, whilst
alkalis and alkaline salts do the reverse (Pauli).

The process of coagulation of proteids, therefore, must depend
upon the neutralization of this electrical charge, and this can be
accomplished either by electrolytes or by colloids. Non-electro-
lytes (i.e., substances which do not split up into electrified ions in
solution) do not bring about coagulation in this way. Many of
them, such as sugar and urea, are inert ; others, such as alcohol,
act in an entirely different manner. The precipitation of an albu-
minous solution by means of a strong acid takes place thus : The
negatively-charged particles attract to themselves the positively-
charged hydrogen ions ; their charge is now neutralized, and the
force of attraction due to their surface tension is no longer counter-
balanced by an electrical repulsion ; the particles are drawn
together, form larger and larger aggregates, and finally cohere into
masses so large as to come under the influence of gravity, when
precipitation takes place. The coagulation of albumin by alcohol
is due to the fact that proteids are not soluble in that fluid, so that
when it is added to a watery solution of proteid, the water is with-
drawn and the particles brought so close together that surface
tension comes into play, and makes them coalesce into aggregates.
The truth of this explanation appears from Pauli's observation
that when no electrolytes are present alcohol acts very readily as
a precipitant — there is no electrical repulsion between the particles.
But if a little acid or alkali be added, and the molecules be thus
given a mutually repellent electrical charge, the precipitation is
inhibited or entirely prevented. Colloids can also be precipitated
by colloids as well as by electrolytes, but in this case they must
be of opposite sign. Thus, in testing for albumin in urine by
means of acetic acid and ferrocyanide of potash, the addition of
the acid insures that the particles shall acquire a positive charge
(if they have it not already), which is then neutralized by the
colloidal ferrocyanic acid of negative sign. Numerous other
examples might be quoted.

This taking up of substances by colloidal particles is an example
of the phenomenon known as adsorption. This process is difficult
of definition, but in general terms we may regard it as a combina-

21

322 ADSORPTION

tion between two substances dependent on physical attraction
rather than on chemical affinity, and taking place in very variable
ratios, rather than in simple ones dependent on combinations of
atoms or molecules, such as occurs in true chemical union. But
it must be admitted that there is no absolutely sharp line of
demarcation between the two classes of phenomena. In most
cases the two substances entering into the phenomenon of adsorp-
tion exist as such side by side in the compound, which is to be
regarded rather as an intimate admixture of the two than as an
entirely new substance. The question arises as to whether the
union between antigen and antibody is not really a process of
adsorption, and in the attempt to solve this problem several very
startling analogies with colloidal adsorption have been discovered.
It will be convenient to discuss some of these, and then to give
an account of the analogies met with in the chemistry of the
colloids.

Bordet has shown that the amount of haemolytic immune body
which can be taken up by a given volume of corpuscles varies
according to whether the corpuscles be added at the same time or
in successive small portions. Thus, in one example 0-4 c.c. of a
haemolytic serum dissolved 0-5 c.c. of corpuscles if added at once.
But if 0-2 c.c. of corpuscles were added first, and then successive
amounts of 0*1 c.c. put in, no solution took place after that of the
first portion added. This he explains, as we have already seen, by
invoking a physical process of the nature of adsorption, comparing
it with the adsorption of a dye by filter-paper. Other explanations
may be possible, but an exactly similar phenomenon may be seen
in the mutual adsorption of colloids of opposite sign. Thus, if to
a given amount of solution of an electro-positive colloid an amount
of solution of an electro-negative colloid exactly sufficient for
neutralization of the opposite charges be added, the result will be
the immediate commencement of the process of agglutination,
which will go on until all the colloids are precipitated. But
if a small amount of the second colloid be added to the same
volume of the other, a different state of affairs is brought about-
New aggregates of the two are formed, which are not necessarily
neutral in reaction, and experience shows that the conditions are
now less favourable to precip tation, which may require much
more of the second colloid for its complete accomplishment, the
intermediate bodies being apparently less easily precipitated than
the unaltered colloid. Similar phenomena may also be seen in the

COLLOIDAL THEORY OF ANTIBODIES 323

action of agglutinin on bacteria. Here also the solid particles may
take up much more antibody than is necessary for agglutination
to be induced if the serum be added all at once. It may be pointed
out that this has a practical value in the estimation of the degree
of agglutinating action as determined by the degree of dilution in
which the action will take place. It is not proper to make a strong
dilution of the serum in the bacterial emulsion, and to dilute this
with further amounts of the latter, unless the subsequent dilutions
are made quickly. Otherwise, all the agglutinin may be removed
by the bacilli in the first dilutions, and those subsequently added
be apparently unaltered. The most accurate method is to prepare
all the serum dilutions required of double strength, and to add to
each an equal volume of bacterial emulsion. On the other hand,
when a bacterial emulsion is added to a serum in successive doses,
more agglutinin is taken up than when the whole amount is added
at once.

Next as regards the phenomena in which an excess of antibody
has apparently a reversing action, the best-known examples of
which are : (i) the deviation of complement, or Neisser-Wechsberg
phenomenon ; (2) the presence of zones of inhibition in which no
precipitation occurs in mixtures of serum and precipitin ; and
(3) the similar phenomena observable in the agglutinins. All
these have been discussed previously and explanations suggested.
These explanations have the merit of not involving any pheno-
mena of nature different to those familiar in immunity reactions,
but there are objections to all. Thus, Neisser and Wechsberg's
explanation of the deviation of complement would have us believe

CLfiH^no C C fJ f-p-r-

that uncombined complement has a greater affinity for alexin than
that which has had its cytophile groups combined with cell re-
ceptors— a phenomenon exactly opposite to that which occurs in
haemolysis, where the process can be studied with greater accuracy.
We have also seen difficulties in the way of accepting Gay's
explanation, the full particulars of which are not yet available.
In the same way the inhibiting effects of large doses of precipi-
tating or agglutinating serum may apparently be explained on the
hypothesis of the existence of precipitoids and agglutinoids of
higher combining affinity than their unaltered antibodies, but it is
at least doubtful whether this is entirely satisfactory. It is difficult
to believe that the effect of moderate heat is to increase the com-
bining affinity of the agglutinin, whereas greater heat destroys it
entirely, and it is altogether different to what is observed in the

21 — 2

324 EXPLANATION OF SPECIFIC INHIBITION

case of toxin, which is exactly equal to toxoid in this respect, and
Dreyer and Jex-Blake bring forward very strong evidence against
this view. Some .of the more important of their facts may be
summarized thus : If it were true, the more the serum is weakened
by heat (and the greater the production of agglutinoid) the greater
should be the zone of inhibition, and vice versa. This they found
not to be the case, for a serum that had had its agglutinating power
very largely destroyed by heat might show a very small zone of
inhibition, whilst another which had been hardly injured might
have a very large one. Further, a serum might show a zone with
an emulsion of bacteria in saline solution, but not in broth.

The alternative explanation of all these phenomena is based on
facts observed in the mutual precipitation of colloids, for here
again exactly similar zones of inhibition are seen. For example,
as Neisser and Friedemann have shown, a suspension of particles
of mastic in water (made by dropping an alcoholic solution into
water) takes on a negative charge, and can be precipitated by
positive colloids or by positive ions. Thus, if ferric chloride be
added precipitation occurs, and if the dose be increased it gradually
becomes more and more rapid and abundant until a certain amount
of ferric chloride is present, after which it becomes less and less
until no reaction takes place at all. And similar facts have been
observed in numerous other cases of interaction of colloids. Their
explanation is somewhat as follows : When the two colloids are
present in such an amount that their electrical charges mutually
annul one another, and they are therefore formed into aggregates
which tend to run together, the addition of fresh colloid bearing
an electric charge disturbs the conditions, and may, by electrifying
the masses already formed, cause mutual repulsion, and, if suf-
ficient amount of the second colloid be added, a re-solution of all
.the masses. Hence solutions, say, of an electro-negative colloid
may be dependent on (a) the presence of mutually repellent mole-
cules or aggregates of molecules in an inert fluid, or (b) on the
presence of these molecules in a fluid also containing a large
number of molecules of opposite sign, whereas in the intermediate
mixtures the conditions for precipitation are present. Here the
precipitate is soluble in excess of both substances, just as pre-
cipitum is soluble in excess either of precipitin or of its antigen.

This theory of the action of agglutinins has been investigated
and strongly supported by Biltz, Neisser and Friedemann, Pauli,
and others, and Dreyer and Jex-Blake have in particular given

COLLOIDAL THEORY OF ANTIBODIES 325

strong evidence in its favour. For example, they find definite
zones of inhibition in the precipitation of B. coli by orthophosphoric
acid, a substance in which the existence of agglutinoids is out of
the question. Thus, in a series of tubes each containing 1-5 c.c.
of an emulsion of this organism in normal saline solution, agglu-
tination occurred in the tubes to which amounts from 118 centi-
grammes down to 4 centigrammes were added, and also in those
containing from i-i milligrammes to o-oi milligramme, but there
was none in which amounts from 4 centigrammes to i'i milli-
grammes were added. Further, there is a close analogy between
this phenomenon and the precipitation of gum mastic by ferric
chloride, for in each case the extent of the zone of inhibition
diminishes as the emulsions of the substance precipitated are made
stronger.

The facts at our disposal are not yet sufficient to enable the
phenomena of agglutination of bacteria to be fully explained, but
the process may take place somewhat as follows : Normal bacteria
pass toward the anode when an electric current is passed through
the fluid, and are therefore to be regarded as particles carrying a
negative charge. The addition of agglutinin removes this charge,
and the particles become electrically inert. This is shown by the
fact that when placed in an electric current they do not move in
either direction. If, however, very large amounts of agglutinin
are added, the point of neutrality may be reached and passed, and
the bacteria acquire a positive charge, again becoming mutually
repellent. Remarkable facts which may have some bearing on
this subject have been brought forward by Neisser and Friede-
mann. Emulsions of mastic are, as we have seen, precipitated by
ferric hydroxide, and it has been shown that the addition of a
small amount of organic colloid, whether positive or negative or
inert like gelatin, protects these particles, so that they are no longer
agglutinated by colloids or electrolytes of opposite sign. A pheno-
menon probably similar is seen when salts which form a colloidal
precipitate are mixed in a solution of gelatin, when the mixture
remains transparent, the colloidal particles being prevented from
undergoing flocculation. Silver chromate, formed by the inter-
action of potassium chromate and silver nitrate, may be obtained
in a transparent and stable form by this process. Particles thus
protected are not carried in either direction by an electric current.
Neisser and Friedemann have shown that mastic emulsions are
" protected " by the addition of a little serum, leech extract, or

326 COLLOIDS AND HAEMOLYSIS

extract of typhoid bacilli, and they think that normal bacteria
may be compared with these particles surrounded by a defensive
envelope which prevents the flocculating action of substances of
opposite sign. The action of agglutinin they believe to consist in
the removal of this layer, so that then the ions of opposite sign
can unite with the bacteria and bring about their agglutination.
This explanation would explain clearly the role of salts in agglu-
tination, and is supported by Kirstein's observation that typhoid
bacilli cultivated in medium free from albuminoid material is
clumped by salt without the addition of serum. The theory, how-
ever, appears at present not to account for the facts already cited
with regard to the transport of normal and charged bacteria in an
electric current.

Some remarkable analogies have been brought forward by
Landsteiner and Jagic between the process of haemolysis by
simple colloids and by specific antibodies. Thus a solution of
colloidal silicic acid acts in a manner closely recalling a specific
haemolysin. It clumps and dissolves the red corpuscles of rabbits,
and agglutinates their spermatozoa, but has no action on typhoid
bacilli ; it thus shows signs of selective action, though not of true
specificity. Its action is manifested in extremely small doses ; it
is rendered inert by heat, and gradually falls off even at the room
temperature. Further, phenomena are observable which strongly
recall the action of complements and of lecithin in reactivating
heated serum or in dissolving sensitized corpuscles, since red
corpuscles which have been agglutinated by colloidal silicic acid
are dissolved by traces of lecithin or of fresh serum, but not by
serum which has been heated to 60° C. ; but if the silicic acid
is present in excess, no haemolysis occurs.

A difficulty arises from the fact that emulsions of red corpuscles
are agglutinated by both positive and negative colloids (ferric
hydroxide, ferrocyanide of copper). Girard-Mangin and Henri
have given an explanation of this, which is briefly as follows :
When a red corpuscle is suspended diffusion of salts, especially
of sulphates of calcium and magnesium, takes place from their
surface, and these facilitate the precipitation of positive and
negative colloids respectively, so that the corpuscles come to be
surrounded by a layer of precipitated colloid material. It is this
zone of precipitated material which actually determines the agglu-
tination of the corpuscles. Thus these authors consider three
substances as taking part in the process : (a) the corpuscles ;

COLLOIDAL THEORY OF ANTIBODIES 327

(b) the salts which gradually diffuse therefrom ; and (c) the
colloid or agglutinin. They brought forward strong evidence
in favour of this theory by showing that agglutination is less
powerful when the corpuscles are suspended in normal saline
solution (the salts of which inhibit the exosmosis of the salts
in the corpuscles) than when they are in isotonic sugar solu-
tion (in which exosmosis is unchecked). Further deductions from
their theory (for which the original articles must be consulted)
were verified experimentally.

The analogies between colloid adsorption and the interactions
of toxin and antitoxin have been studied by Biltz, Pauli, and
others, and very remarkable facts adduced. For instance, there
appear to be phenomena indicating that the action of some anti-
toxins at least is reversible when a large amount is present. Thus,
according to Jacoby, the action of crotin (a phytotoxin of simple
nature and closely analogous with the bacterial exotoxins) is
increased by the addition of minute amounts of antitoxin, large
quantities of which neutralize its activity. Phenomena somewhat
similar are seen with the true toxins — for example, Danysz (cor-
roborated by von Dungern and Sachs) has proved that diphtheria
antitoxin will neutralize more toxin when added at once than when
added in successive portions. Thus if the amount of toxin which
is just neutralized by a certain amount of antitoxin be divided into
two parts, and added to the antitoxin at an interval of twenty-four
hours, the whole may be toxic, although the amounts of toxin and
of antitoxin present are exactly the same in the two cases. Both
these effects can be explained on Ehrlich's pluralistic conception.
Thus Jacobi's results may be explained on the assumption that
small amounts of anticrotin combine with the non-poisonous
prototoxoid, and that this renders the mixture more toxic, because
this prototoxoid would otherwise seize on the receptors of the
sensitive cells, to the exclusion of the active toxin. This, however,
is very difficult to believe. Taking the action of crotin in pro-
ducing haemolysis as a test, it would appear that the neutralization
of some of the toxin, even if inert, would render the substance
less haemolytic ; this appears to follow from Bordet's proof that
red corpuscles can take up much more than their haemolyzing dose
of an antibody. The Danysz effect is also explicable on the
theory of the presence of a non-toxic epitoxonoid of no toxicity
and of feeble combining affinity. This slowly forms a firm com-
bination with the antitoxin, and renders it useless for neutralizing

328 ANALOGIES WITH DANYSZ' EFFECT

more toxin. Explanations are also available on the Arrhenius-
Madsen explanation, regarding the combinations as examples of
mass reaction.

Regarding reactions such as these as interactions of colloids,
we find them paralleled in simple reactions. Thus in some cases
the addition to a colloidal solution of a small amount of a second
colloid of opposite sign may render the solution more stable, and
protect it from precipitation by an excess of the second substance.
The partially neutralized aggregates appear to possess more re-
pulsive power than those carrying their full electric charge.
Again, the effect of an addition of one colloid to another of
opposite sign is often brought about slowly, and the material only
gradually attains its permanent form. It is probably this fact
that renders solutions of antitoxin so unstable. Interreactions take
place between the colloids themselves and the electrolytes in the
serum, aggregates are formed, and the antitoxic potency falls off—
rapidly at first and subsequently more slowly. It is a matter of
great difficulty to prepare stable solutions of proteid materials.
As we have already pointed out, the amount of colloid necessary
to precipitate a constant amount of another of opposite sign is
reduced to a minimum if the addition be made at once, and
rendered much greater if it is made slowly in small amounts with
an interval between each. This is closely analogous with the
Danysz effect. And this must also be considered in its bearing
on Ehrlich's method of attempting to neutralize a certain dose of
toxin by successive addition of small amounts of antitoxin, the
method by which the whole of the elaborate pluralistic conception
of the structure of toxins has been built up. According to the
colloidal theory, this is explicable on the assumption that tran-
sitional compounds of toxin and antitoxin of very diverse nature,
not necessarily dependent on the proportions of the two substances,
are formed. This is practically the view put forward by Bordet
from a consideration of the reactions of the antibodies, and especially
alexin and anti-alexin, and supported by the workers who have
examined the question from the standpoint of colloidal reactions.

The difference between the L0 and L+ dose, which is so
important a feature in Ehrlich's work, also finds an analogy in
the reactions of simple colloidal substances. Thus ferric hydroxide
neutralizes arsenious acid (whence its use as an antidote in acute
arsenical poisoning), and it was found by Biltz that if one lethal
dose of the latter substance was added to a neutral mixture of the

COLLOIDAL THEORY OF ANTIBODIES 32Q

two, this mixture did not regain toxicity. Several lethal doses
must first be added, just as it is necessary to add several lethal
doses of diphtheria toxin to the L0 dose.

A difficulty arises from the fact that both toxin and antitoxin
pass in the same direction in an electric current. These experi-
ments are naturally difficult to carry out, since the effect of
electrolysis must be eliminated. When this is done, according to
Field and Teague, both substances travel towards the cathode, and
this whether the solution is acid or alkaline. But we must
remember that when toxin and antitoxin interact they always do
so in a very complex fluid, containing many other substances,
both colloids and electrolytes, and until we can determine the
electric charge of these substances in a pure form, these experi-
ments can hardly be considered to outweigh the remarkable
analogies described above. It is highly probable, judging from
analogy with the other antibodies, that free ions are essential to
the neutralization of toxin by antitoxin, and the researches of
Girard-Mangin and Henri show us how complex the conditions
may become.

We have seen that it is important to know whether the
compounds of antibody and antigen are dissociable, and that the
latter view explains some difficult phenomena seen in immunity.
On Ehrlich's theory the combination between the two is a strong
one, and once it has been allowed to take place, the two substances
are not dissociated into their components, whereas on Arrhenius's
and Madsen's theory the combination is an unstable one, and dis-
sociation constantly occurs. The experimental proof in favour of
the occurrence of this phenomenon appears on the whole satisfac-
tory. Red corpuscles saturated with immune body mixed with
normal corpuscles will part with some of their antibody, so that
the latter become sensitized. A mixture of tetanus toxin and
antitoxin, which causes no symptoms when injected into the leg of
a guinea-pig, causes tetanus when adrenalin has previously been
injected locally (explicable on the fact that toxin is absorbed by,
the nerves and antitoxin by the vessels, which in the second case
are constricted), and other examples might be quoted. It appears,
however, that this irreversibility only occurs if the two substances
are not kept in contact for a sufficiently long time, as was pointed
out by Martin and Cherry in their nitration experiments very
early in the history of the chemistry of the action of antitoxin.
Thus in the latter example, if the tetanus toxin and antitoxin be

I

330 COLLOIDAL CHEMISTRY

kept in contact for two hours, no tetanus is produced, in spite of
the local action of the adrenalin. This appears to harmonize best
with the colloidal theory. The interactions of colloids of opposite
sign and the precipitation of colloids by electrolytes may be
reversible if the conditions are changed early, but the compounds
formed gradually become more and more stable. But if the con-
tinuation were that of two substances of strong chemical affinity
we should expect it to be stable from the first, and if it followed
the laws of mass reaction it would be reversible, no matter how
long it had stood.

There appears to be no crucial experiment by which the truth
of these three theories can be tested. The arguments brought
forward by the upholders of each involve the study of the exceed-
ingly complicated laws connected with the true relations and
completeness of the reactions, and involve advanced mathematical
consideration. It is, however, doubtful whether the processes are
measurable with the accuracy required in the investigation of
these complicated formulae, and in some cases very slight altera-
tion of the observed results would render the facts explicable on
a theory other than that in support of which it was adduced. At
the present time the balance of the evidence appears to be decidedly
in favour of the colloid hypothesis.

CHAPTER XIII
ON IMMUNITY TO BACTERIA

WE are now in a position to discuss the mechanism by which a
bacterial infection is combated. It must be pointed out in the
first place, and kept in mind throughout, that different microbes
are dealt with in different ways. Some, such as many of the
protozoan blood parasites, may be tolerated, since the tissues of
their host are immune to the action of their toxins ; others are
removed by phagocytosis, after preparation by opsonins, or perhaps
by alexins, amboceptors, or other substances, or possibly without
previous treatment ; and yet others may be dissolved by the process
discussed under the heading of Bacteriolysis. Each of these
processes gives rise to a different form of immunity, which we
may call — (i) immunitas non sterilisans, (2) phagocytic immunity,
and (3) bacteriolytic immunity. And the latter two processes
(which are the only ones of practical importance) rarely, if ever,
exist in a pure state ; almost all invading bacteria are, or may be,
dealt with by a combination of the two.

We must also notice that the defensive processes will be greatly
modified by the nature of the region in which the combat between
the invaders and the defensive mechanism goes on. There are
three main cases : (i) the circulating blood, (2) the tissues, and
(3) regions, such as the peritoneum, combining^some of the
characters of the other two. We will take them Jh this order.

In dealing with the processes that take place when bacteria gain
access to the circulating blood, our main difficulty at first sight is
not to explain the existence of immunity, but the reason why
infection ever occurs ; for the defensive mechanism seems more
than adequate to combat any dose of bacteria which is likely to
gain access to the blood under natural conditions. The processes
by which the bacteria are removed are phagocytosis and bacterio-
lysis, or the two combined, and in either case the defence would
appear to be more than adequate. Thus, in an experiment for the

331

332 DIFFICULTIES OF THE PROBLEM

determination of the opsonic index in which thick emulsions of
bacteria are used the leucocytes are often found absolutely packed
with streptococci or tubercle bacilli after a short incubation ; if
on the average 100 cocci are taken up, calculation will show that
the whole volume of blood might phagocyte 2,000,000,000,000 in
a few minutes. Tubercle bacilli are perhaps hardly so easily
taken up, but it is difficult to see how the comparatively moderate
number of these bacteria which gain access to the blood after the
rupture of a small caseous gland into a vein can escape immediate
phagocytosis : yet we have no reason to believe that this accident
often occurs without causing general tuberculosis. Perhaps a
more convincing example is given by determinations of the opsonic
index in septicaemia, where the bacteria— e.g., streptococci — are
known to be circulating in the blood. Here the opsonic index is
usually low (0-5 or less), but even then it is sufficient to enable the
leucocytes to take up an enormously greater number of bacteria
than we can discover in the blood.

Similar phenomena occur in the case of bacteria which are com-
bated largely by means of bacteriolysins. Thus, normal blood
serum is markedly bactericidal to typhoid bacilli, yet infection
occurs, and many observers have noticed relapses when the
bacteriolytic power of the blood is extremely high. In these
cases, therefore, the bacteria can gain access to the blood (for
typhoid fever is always in part a septicaemia) in spite of barriers
which would appear unsurmountable. And we have already had
occasion to mention that anthrax may occur in animals the blood
serum of which is powerfully bactericidal ; the animals, indeed,
may be extremely susceptible.

Before attempting to explain the apparent discrepancy, it must
be pointed out that, as a matter of fact, septicaemia is a rare
disease in proportion to the opportunities for its occurrence, and
that the continued presence of bacteria in the blood without a local
lesion from which they are constantly discharged is extremely
uncommon. Thus in diphtheria, tetanus, the staphylomycoses,
etc., in which we should imagine that there is every chance for
bacteriaemia to occur, it is practically unknown. It is common in
diseases like ulcerative endocarditis, the early stages of typhoid
fever, and in pneumonia, in which we have reason to believe that
there is a constant shower of organisms discharged from the local
lesion direct into the blood. Here, however, we can feel fairly
certain that the bacteria which we find in the blood are only there

ON IMMUNITY TO BACTERIA 333

temporarily, and are rapidly destroyed : the reasons being, firstly,
that they are never present in large numbers, as they would be if
they lived and multiplied in the circulation, and, secondly, that
they do not as a rule form secondary lesions, as we might expect
them to do, when deposited in the tissues. We may fairly assume,
therefore, that in a disease like typhoid fever bacteria are constantly
passing from the lesions into the blood, but are rapidly destroyed,
so that the defensive mechanism of the fluid is sufficient to prevent
the occurrence of a true septicaemia, in which bacterial growth
continues in the blood. The same holds in pneumonia and in
ulcerative endocarditis when not accompanied by secondary
infective lesions, and probably most other diseases, and does not
militate against the view that the blood is a sufficient defence in
the majority of the cases in which the infective material gains
access to the interior of the vessels. Excluding these and similar
examples, septicaemia is a rare disease, and is commoner in the
case of the protozoal than in bacterial infections.

As regards the failure of the process of phagocytosis, which
must occur when an organism which is combated by phagocytosis
does gain a foothold in the blood, we may point out that we have
yet but little knowledge of the exact state of the bacteriotropic
substances in the living plasma. Wright, it is true, has shown
that citrated plasma contains the same amount of opsonin as
serum ; but the two fluids are not absolutely identical, and the fact
that opsonin increases in amount in the successive fractions of
serum squeezed out from a clot leads us to believe it probable
that it does not exist as such in the plasma ; and if, as we have
shown is likely, the naturally existent thermolabile opsonin is the
same as alexin, it is, on the whole, probable that there is none in
the living blood.

Briscoe's experiments reproduce the natural conditions more
closely. He injected emulsion of bacteria into the ventricle of a
heart after excision and clamping of the auriculo-ventricular
groove. After a time some of the blood was removed and
examined, and practically no phagocytosis was found to have
taken place. But many bacteria were taken up when the same
animal's leucocytes and serum and the emulsion of bacteria were
incubated outside the body in the ordinary way. Further, in
some cases a little clotting occurred in the heart, and in these it
was found that the leucocytes which were included within the
clot took up many more bacilli than those which remained free.

334 FUNCTION OF THE SPLEEN IN PHAGOCYTOSIS

These results argue strongly in favour of the fact that opsonin
does not occur in the circulating plasma as such, and is only set
free during the process of coagulation or of phagolysis. This
conclusion was corroborated by the fact that opsonized bacteria
were taken up freely, showing that the leucocytes were not at
fault,

We can get no information on this point from an examination
of the blood in these infections, since a small number of non-
virulent bacteria will be digested and completely disappear in a
few minutes if taken up by a leucocyte ; but as a matter of fact, it
is very uncommon to find ingested bacteria in the leucocytes of
the circulating blood in natural diseases.

It might appear that these explanations go too far, and would
lead us to the conclusion that phagocytosis is without value as a
defensive process when bacteria reach the blood-stream. This is
not the case, and it is probable that the great factor to consider is
the spleen. This acts as a sort of filter, and in its pulp the flow
of blood is slow, so that there is abundant opportunity for phago-
cytosis to occur. The same is true, though in a lesser degree, of
all regions of the body in which the circulation is tardy.1 Hence,
when bacteria gain access to the blood they will always be found
in considerable numbers in this organ, which Kanthack called
years ago " the graveyard of the bacteria." This is especially the
case in typhoid fever, and the method formerly in use in the
diagnosis of the disease, by making cultures from traces of spleen
pulp removed by puncture, almost invariably gave positive results.
Films or sections of the spleen pulp will often show bacteria
ingested by the leucocytes in many diseases, even although none
can be detected in the blood. The role of the spleen in immunity
is probably a complex one, and this may explain the divergent
views held as to its importance in this respect. Thus it was
found long ago by Tizzoni and Cattani that it was impossible to
immunize animals to tetanus after splenectomy ; but this was not
corroborated by other observers, who found the animals equally
immunizable after the results of the operation had completely
passed away. But in this connection we must remember that the
splenic tissue is very rapidly regenerated, and that in all prob-
ability the haemolymph glands and other structures take on its
functions vicariously after its ablation. ^

1 Leucocytes which have taken up bacteria whilst in the circulation
probably accumulate in the lungs also.

ON IMMUNITY TO BACTERIA 335

The role of the spleen in facilitating phagocytosis is well seen
in the spirilloses. Metchnikoff showed many years ago that after
death from relapsing fever the cells of this organ are packed with
spirilla, such cells being rarely, if ever, seen in the circulating
blood. When monkeys are inoculated with blood containing the
parasite, a febrile attack occurs, but recovery soon takes place.
During the disease free spirilla occur in the blood, but they soon
disappear ; but if the animals are killed after this, the organisms
may be found within the splenic leucocytes. Further, Souda-
kewitch showed that recovery often failed to occur in animals
inoculated after splenectomy, and in this case the spirillum
appeared in large and increasing numbers in the blood, and no
phagocytosis was observed. Very similar facts have been
observed by Levaditi and Manouelian in tick-fever. During the
height of the disease the spirillum occurs in the blood, and is
exclusively extracellular. After the crisis it is found only in the
phagocytes of the spleen and liver (Kupffer's cells).

Besides the spleen, the bone-marrow is also a region where
abundant phagocytosis occurs, and the functions of the two tissues,
like their structure, are similar. In each case there is a slow flow^
of blood in the pulp, close proximity of the blood to the tissue
cells of peculiar type, and an abundant supply of leucocytes.

The function of these organs in phagocytosis is probably two-
fold. It provides a region where the blood-stream is comparatively
tranquil, so that there is no mechanical obstacle to the process :
this is probably comparatively unimportant, since it can and does
take place under certain circumstances in the full torrent of the
circulation. And, secondly, if, as seems probable, thermolabile
opsonin is really complement, and if complement is formed by the
leucocytes, it is readily conceivable that it may be present in the
leucocytic organs when only present in small amount in the rest
of the blood. It may actually be called into existence by the action
of the toxin on the spleen cells or leucocytes, and be immediately
absorbed by the bacteria, so that it never occurs in appreciable
amount in the plasma.

This action will not come into play in the case of the bacterio-
tropic substances which are true antibodies, for we have every
reason to believe that these are present in the plasma as such, and
are not set free in the process of clotting ; yet even here we have
seen some evidence which suggests that they, in common with
other true antibodies, are formed inlymphoid tissue or from lymphoid

336 FUNCTIONS OF OTHER ORGANS

cells, or, as some recent research seems to prove, from the endo-
thelium.

The liver is also an important region with regard to the deposi-
tion of bacteria, and their subsequent destruction by phagocytosis
or other means, but in this case the main cells concerned are
Kupffer's cells. We must regard these two organs and the lungs
as being intercalated in the circulation with the object, amongst
others, of arresting bacteria and other foreign particles which gain
access to the circulation, and affording a region in which phagocytosis
can take place at leisure. It has been shown experimentally that
animals will resist a much larger dose of bacteria if the injection
be made into the portal vein than if it is thrown into one of the
peripheral veins. The importance of these organs is also well
shown from a study of the fate of inert particles, such as carbon
or carmine, when injected into the circulation or peritoneum. In
either case they make their way with great rapidity to the organs,
especially the liver and spleen, in which they occur in large
numbers within two hours of the injection. They are also found
within the bone-marrow, lymph glands, tonsils, and lungs, and,
whatever the region in which they occur, are mostly contained
within the leucocytes or tissue phagocytes. Here also intracellular
absorption may take place, though of course it is much slower
than is the case with bacteria, and according to Siebel the leuco-
cytes, with their load of unabsorbable particles, may leave the body
via the lung, tonsils, or lymphoid structures of the small intestine.

The lung plays a part of great importance, but one not fully
understood, in the process of phagocytosis of bacteria which gain
access to the circulation. According to some writers, it is the only
internal organ in which phagocytosis by polynuclears can be
demonstrated, but this is certainly not correct. It is true, how-
ever, that soon after the injection of streptococci (Tchistovitch) or
cholera vibrios (Levaditi) into a rabbit, polynuclears containing
these organisms may be found in large numbers in the wide
capillaries of the lung ; this is probably the reason for the sudden
fall in the number of the leucocytes in the circulating blood which
occurs soon after injection. The reason for this collection in the
lungs is very far from clear : the region, remote as it is from
lymphoid tissue, would appear an unfavourable battle-ground
against the infection ; nor are the later stages of the process fully
understood.

The question of the importance of the bacteriolysins in haemic

ON IMMUNITY TO BACTERIA 337

infections, already alluded to, now requires further mention. It
is obvious that the role which it plays varies greatly in different
diseases. There is, for instance, no evidence for believing that
the blood ever develops any substance which is bacteriolytic for
the tubercle bacillus. It is true that the researches of Wassermann
and his coadjutors show (by means of the method of fixation of
complement) that an antibody may occur in tuberculosis or in
tuberculous animals, but there is no evidence that this is a
bacteriolysin, or, if so, that it ever occurs in amounts sufficient to
bring about solution of tubercle bacilli ; in fact, it appears probable
that tubercle may run its whole course without leading to the pro-
duction of any recognizable antibody. So, too, with staphylococcic
infections. Here it is possible by special methods to produce a
serum which has a slight bactericidal effect, but normal human
serum or the serum of a patient who has been submitted to a
course of antistaphylococcic vaccination is quite powerless in this
respect. It is obvious, therefore, that neither the high grade of natural
immunity to staphylococci of normal persons nor the increased
amount present in active immunity is due to any bactericidal or
bacteriolytic substances occurring in the blood. It is quite true
that in old specimens of staphylococci pus-free cocci in all stages
of degeneration can be found, rendering it quite obvious that a
process of extracellular death and destruction of bacteria is at
work. This phenomenon is probably to be attributed mainly to
the action of the proteolytic enzymes formed by the leucocytes of
the pus. These are substances which are too often neglected in con-
sideration of immunity. They are, of course, non-specific, and will act
on any dead proteid, and on some living organisms. Their action may
be demonstrated by allowing some liquor puris from an old abscess
to act (in presence of thymol) on some staphylococci which have
been previously boiled to prevent autolysis. Another process in
which these degenerated forms of cocci may be produced is by
spontaneous autolysis in organisms killed by lack of suitable
nourishment, or atrepsy.

In sharp distinction from tubercle and the staphylomycoses, we
find such diseases as cholera and typhoid fever, in which bacterio-
lysins are formed in large amounts, and are readily demonstrable
by simple methods. It is in these and a few other diseases that
we may expect the role of these substances in recovery and
immunity to be most marked. We find, however, that their
action is most difficult to understand, and that its importance

22

ROLE OF THE BACTERIOLYSINS IN RECOVERY

appears to diminish the more it is investigated. In general terms,
there is no doubt as to the fact that the presence of abundant
bacteriolytic substances must be regarded as a cause of immunity :
thus, animals may be rendered passively immune, e.g., to intra-
peritoneal injections of typhoid or cholera by antecedent or
simultaneous injections of powerful bactericidal serum. But
further investigation shows that even in this phenomenon there
are certain features which must lead us to regard the process of
bacteriolysis as of subordinate importance, or even as harmful.
In the first place, the demonstration of the fact that the opsonic
action of a powerful antityphoid serum (as determined by the
process of dilution until extinction of the property occurs) is pro-
portionate to its bactericidal action, and other facts already noticed,
have led us to the belief that immune body or amboceptor can
play the part of an opsonin, and, further, that it does so when
present in amount so small that its bacteriolytic action is but slight
or not apparent. It is, therefore, in the highest degree probable
that the passive immunity conferred by the injection of a bacterio-
lytic serum is due to the opsonizing action of this serum, manifested
even when diluted with the juices of the animal into which it is
injected.

In the second place, there is this main difference between the
destruction of bacteria within the leucocyte after phagocytosis
has occurred and the extracellular solution by immune body and
alexin : in the former case it is unusual for the endotoxins to be
set at liberty, so that as soon as a bacterium has been taken up by
a leucocyte, we believe that, as a rule, its capacity for harm has
been removed. It undergoes solution within the protoplasm of the
leucocyte, which, as appears from the studies of the French school,
is peculiarly insusceptible to the action of toxins, is destroyed by
the digestive juices, or by other means, and is not, at least in the
majority of cases, set free to injure the tissue cells. It is other-
wise with the solution of bacteria which takes place from the
action of bacteriolytic antibodies, for here the endotoxins are set
free, so that the presence of these antibodies, though undoubtedly
developed as a part of the defensive mechanism, may be a source
of added danger to the animal. For example, if an intraperitoneal
injection of a large dose of cholera vibrios be made into the peri-
toneal cavities of two animals, and in one some anticholera serum
be added, it is often found that this animal dies very rapidly with
symptoms of acute intoxication, whilst the other survives longer

ON IMMUNITY TO BACTERIA 339

and dies of septicaemia. If dead vibrios are used the addition of
serum may bring about death, whilst an animal injected with the
organisms alone may recover with scarcely a symptom of intoxica-
tion. Similar phenomena, though less definite, often follow the
use of antistreptococcic serum for patients suffering from strepto-
coccic disease : the injection may cause a rapid rise of the
temperature and exacerbation of the symptoms. Other explana-
tions are possible, but it is highly probable that this is due to the
solution of some of the cocci and liberation of their endotoxin.
It is possible that in some cases this might be a source of danger,
and that this liberated toxin might be sufBcient to kill the patient ;
but there is no evidence that this actually occurs, and, as a rule,
the phenomena mentioned may be taken as proof that the serum
is a suitable one, and that its use should be continued.

We must, however, be careful in arguing from the phenomena
occurring in hypervaccinated animals to those suffering from a
natural infection. There can be no analogy between the train of
events when a massive dose of pathogenic bacteria is suddenly
injected into the circulation or peritoneal cavity of an animal in
which there is already an abundance of bacteriolytic substances,
and those which follow a natural infection. Considering the
latter case, and excluding all considerations of phagocytic action,
we may trace the sequence of events in the case of a disease such
as typhoid fever somewhat as follows : In the first place, since
normal human serum is bactericidal to typhoid bacilli, we must
regard the normal resistance against the disease as being due, in
part at least, to the presence of immune body in the serum,
though we have no means of knowing whether its action is mainly
opsonic or mainly bacteriolytic when occurring in the plasma.
When infection occurs, the amount of this substance must
diminish, since it will combine with the typhoid bacilli to which
it gains access, and may, indeed, be sufBcient to kill them all, and
thus prevent the development of the disease. Where, however,
the organisms gain access in sufficient numbers, so that the amount
of immune body or of alexin is insufficient to prevent the further
growth of some of them, those that escape will continue to grow,
and the disease will gradually increase in severity. At the same
time, the dissolved products of the organisms which succumb will
reach the seats of antibody production and stimulate a fresh pro-
duction of immune body. WTe have seen reason to believe that
this production takes place mainly in the lymphoid tissue and

22 — 2

340 SEQUENCE OF EVENTS IN GENERAL INFECTIONS

other regions rich in lymphocytes, and we may probably corrobo-
rate this fact with the hyperplasia of the abdominal lymph glands,
Peyer's patches, and spleen which occurs in the disease, regarding
them as conservative, defensive phenomena, rather than the
pathological effects of the bacillus.

After a latent period or negative phase of a few days (the dura-
tion being dependent on the severity of the infection), the antibodies
begin to make their appearance. At first put out only in small
quantities, they will be absorbed at once by the bacilli present, and
will not be demonstrable in the serum or plasma in a free state.
A race now ensues. On the one hand, the typhoid bacilli multiply
as rapidly as their environment will allow ; on the other, the
immune body and other defensive antibodies are being gradually
elaborated more and more rapidly. In a very severe infection
this latter process may perhaps make default altogether, and the
patient succumb during the negative phase. This has not been
demonstrated in the case of the bacteriolytic substances, but is
well known in the case of the agglutinins. In most cases, how-
ever, the increase of bacteria and of antibodies takes place at the
same time. After a time bacteriolysis occurs, but in small
amount, since but little immune body makes its appearance, and
that which is formed at first is probably utilized as opsonin.
Hence when solution of the bacteria and liberation of the toxins
does occur, it does so at first only to a small extent, and we may
regard it probable that the minute amount of endotoxin set free
does not add appreciably to the symptoms of intoxication already
present. Further, we have now the conditions under which
immunity to these toxins may be expected to develop : the setting
free of small but gradually increasing doses of endotoxin — condi-
tions, in fact, closely similar to those obtaining in Macfadyen's
experiments. Probably, therefore, by the time that much bacterio-
lysis occurs the patient's tissues are more or less immunized. It
is, however, not unlikely that the rapid oscillations of temperature
characteristic of the latter stages of typhoid fever may be due in
part to this liberation of endotoxin.

Some researches by Bail would appear to indicate that there is
a sort of automatic mechanism for the prevention of bacteriolysis
in the tissues. He found that no bacteriolysis took place when a
powerful serum was mixed with emulsions of cells from the liver,
spleen, etc, as well as with bacteria. The explanation is doubtless
that the complement is absorbed by these cells, as has been pointed

ON IMMUNITY TO BACTERIA 341

out by Hoke, Muir, and others. Now, if this process goes on in
the living body, it can only indicate that complement (and thermo-
labile opsonin) can never occur in a free state in the organs, and
hence that bacteriolysis does not occur in these regions. It must
be pointed out, however, that amboceptor (or thermostable opsonin)
is not absorbed in this way, so that there would be but little inter-
ference with phagocytosis. It is quite possible that this absorp-
tion of complement does take place in order to avoid the liberation
of endotoxin consequent on bacteriolysis.

The mobilization of the patient's defences is usually manifested
by an increase in the number of the leucocytes in the circulating
blood. This subserves two functions : there is a greater number
of leucocytes to act as phagocytes and there is an increase in the
possible number of cells which may produce alexin or complement.
Hence we find leucocytosis present almost invariably in cases in
which bacteria gain access to the blood; the exceptions being,
firstly, very severe infections in which the defensive reaction fails,
and, secondly, a group of diseases of which typhoid fever is the
best example. The meaning of the first exception is tolerably
clear, that of the second not at all apparent.

The mechanism by which this increased output of leucocytes is
produced is chemotaxis. The bacteria in the blood-stream produce
a toxin which, in comparatively mild infections or in severe
infections in an animal possessed of strong natural immunity,
attacks the leucocytes. These cells will, therefore, tend to leave
the region in which they are formed and in which they are lying
dormant and make their way into the blood, the amount of toxin
being greater in the latter than the former situation. At the same
time, as Muir has so conclusively shown, there is an increase in
the functional activity of the bone-marrow, which manifests itself
in an increase of the leucocyte (and especially polynuclear
leucocyte) producing cells. A double process goes on, leucocytes
being formed more rapidly and attracted from the bone-morrow
as soon as they are produced. That this is due to chemotaxis is
shown by the fact that it follows the injection of bacterial toxins
and other products into the blood-stream, if not too virulent or in
too large amount.

When either of these conditions occurs we see the opposite
result — a leucopaenia, or at least an absence of leucocytosis, and
this is always an extremely bad omen when it occurs in those
infectious processes where leucocytosis ordinarily takes place.

342 VIRULENCE OF BACTERIA

Thus in pneumonia the number of leucocytes per cubic millimetre
has a most important prognostic value, and in almost all cases it
will be found that a patient with a high leucocytosis will recover,
even if the attack is a severe one and the clinical signs unfavour-
able. This leucopaenia may be attributed either to negative
chemotaxis, or to the production of paralysis or death of the
leucocytes, or to inhibition of the functions of the bone-marrow,
or to a combination of these causes. In any case, it argues a high
grade of virulence on the part of the bacteria, which, by lessening
the action of the main natural protective forces, increases the
chance of a lethal issue to the disease. Let us, therefore, discuss
some of the main facts (many of which have been referred to
before) concerning the nature of the mechanism of an increase in
virulence.

For an organism to be virulent it must produce a toxin which
has a profound action on the cells (and, it may be added, on the
important cells) of the animal in question. This, of course, is
obvious, and without it the organism could only produce pathogenic
effects by such means as deprivation of oxygen or food-stuffs from
the host or the production of bacterial emboli — results which are
of no practical importance. Without an active toxin the bacteria
would either lie latent, as occurs in typhoid fever, gonorrhoea and
tubercle, or circulate in the blood as a harmless parasite, as is the
case with many protozoa in the lower animals.

As regards the special actions of toxins which render the bacteria
especially virulent, we may distinguish two — (i) a leucocytic or
leucolytic action, and (2) negative chemotaxis. The former is a
very common function of highly virulent bacteria; its action is
best seen in the production of pus and necrosis in local lesions,
but its presence may be inferred from the degenerative changes
frequently present in leucocytes in general infections. And these
leucocytic substances need not be true toxins, for there is
reason to believe that the non-specific enzymes and other
substances which the bacteria produce have this action, and
that the death and destruction of leucocytes which is so marked
a feature of abscess production may be due in part to their
action.

We have already referred to the question of negative chemotaxis.
It probably actually occurs, but the proof is hardly convincing,
and the subject needs investigation by modern methods in vitro.
There is no doubt, however, that in virulent infections the regions

ON IMMUNITY TO BACTERIA 343

containing the bacteria are often singularly free from leucocytes,
be the explanation what it may.

The first necessity for virulence, therefore, is the possession of
an active toxin, which, either by lowering the general vitality,
or by repelling, paralyzing, injuring, or killing the leucocytes,
diminishes the amount of phagocytosis which occurs.

A second defence against phagocytosis is the production of a
defensive envelope. This is well seen in the case of virulent
anthrax bacilli, which form thick gelatinous envelopes in the
animal body. Such capsulated forms are only taken up with
difficulty by the leucocytes, and the greater the power of capsule
formation the greater the virulence of the culture. Similar
phenomena have been seen in the case of the streptococci,
tubercle bacilli, and other organisms. The capsule of the
pneumococcus, which is only developed in the animal body or in
culture fluids approximating thereto, is probably a case in point.

Capsule formation is also probably a defence against bacterio-
lysins. Thus, according to Eisenberg, virulent typhoid or coli
bacilli stain blue by Giemsa's method, and show a pink edge.
Organisms which have undergone this modification are suscep-
tible to phagocytosis, but show great resistance to bacteriolysis.
And it is probable that the capsule of the anthrax bacillus is
intended largely as a defence against bacteriolytic substances. The
bacilli can be trained to produce it by cultivation in the serum of
the rat, which dissolves the unaltered bacilli in large numbers.

A third and more subtle method in which a bacterium can
increase its virulence is by loss of the receptors (which it possesses
in the avirulent state), which have the power of anchoring the
defensive substances of the blood. Thus the virulent pneumo-
cocci which (as shown by Rosenau) are not ingested by leucocytes
in presence of serum escape, as a result of the fact that they do
not absorb opsonin. Typhoid bacilli which have been cultivated
in immune serum lose their power of combining with amboceptor,
and also with agglutinin, and it may be taken as a general rule
that the more virulent a culture the less its reaction to its agglu-
tinin, and the less it absorbs that substance. Thus typhoid bacilli,
when isolated from the body, are usually refractory to agglutina-
tion, and gain the property only after several generations of culture
on artificial media.

Hence, therefore, we may picture the process which goes on
when a dose of pathogenic bacteria gains access to the blood of a

344 SEQUENCE OF EVENTS IN GENERAL INFECTIONS

susceptible animal somewhat as follows : The bacteria form a
toxin which attracts the leucocytes from the bone-marrow into
the blood, and which also stimulates their production in larger
numbers. It also attracts the leucocytes into the neighbourhood
of the bacteria, and this process is more marked in regions, such
as the spleen and bone-marrow, in which the flow is slow. As a
result some leucocytes are stimulated to produce complement, or,
on the other theory, are killed, and their complement set free. In
either case the bacteria are prepared for phagocytosis. This
process may now take place, the bacteria be destroyed, and the
infection come to an end. This is a type of a mild infection.

Or it may happen that the toxin is so powerful that the leuco-
cytes are repelled, or, if attracted, immediately killed. If this
occurs the sole resource of the patient is the bacteriolytic property
of the blood, which will then come into action — provided, of course,
that the bacterium is one to which amboceptor occurs in the normal
blood — because alexin is set free in the solution of the leucocytes.
Experiment, however, leads us to believe that this is a slower
and less important process than phagocytosis, and that if the latter
is in abeyance the outlook for the patient is bad in the extreme.
If both processes fail the infection must be rapidly fatal.

In an infection in which the defensive and offensive mechanisms
are nicely balanced, and an illness of some duration and severity
occurs, the processes will be much more complex. The early
stages will occur as in a mild infection, and the bacteria will be
surrounded by the leucocytes, and perhaps even ingested. In this
case it may happen that after ingestion solution will take place,
and an endotoxin be set free which will kill the phagocyte, and
perhaps also those in the immediate neighbourhood, and in this
way some of the bacteria may be shielded from further phagocy-
tosis for a time. This and similar processes give the bacteria
time in which to undergo their defensive modifications, which they
do in virtue of the " survival of the fittest " — i.e., those which are
best adapted to their environment in the host. Now it is clear
that in any culture of bacteria the different individuals have
marked differences with regard to their power of absorbing anti-
bodies. This is obvious when we consider that otherwise the
addition of gradually increasing amounts of an agglutinating
serum would cause no effect until a certain concentration was
reached, when all the bacteria would clump. As it is, those
organisms with the greatest affinity take up the agglutinin in

ON IMMUNITY TO BACTERIA 345

largest quantity, and when the number of molecules reaches a
certain amount they clump, whilst those with less affinity have
not yet absorbed sufficient. Probably exactly similar phenomena
occur in the opsonization of bacteria. Some organisms are avid
for opsonin ; others take it up with difficulty, and a large concen-
tration of the substance is necessary before they are prepared for
phagocytosis. This probably explains the fact that the number of
bacteria taken up by the leucocytes does not increase pari passu
with the amount of opsonin present.

The bacteria which remain immune to the defensive mechanisms
will be enabled to live, and in their descendants the conditions for.
natural selection will occur — variations and an adverse environ-
ment. The weaker forms — i.e., those which, by the absence of a
defensive layer or .the presence of numerous receptors on which
the opsonins or bacteriolysins of the host can seize, or those
which do not elaborate a powerful toxin — will be destroyed, and
the more virulent forms will survive, so that a gradual selection
of the latter will ensue, and the invaders rapidly increase in
virulence. And it must not be forgotten that bacteria multiply
with great rapidity, division sometimes occurring in a quarter of
an hour, so that nearly a hundred generations are passed through
in a day, and abundant opportunity for evolutionary selection
occurs. From what we know of the virulence of typhoid bacilli
and of pneumococci when causing disease and when living
parasitically without the body, there is reason to believe that this
increase of virulence always takes place in infections.

Here, then, the standing forces of the body are insufficient to
deal with the invader, and the latter increases in virulence and
numbers during the early stages of the struggle. This process
will continue until the latent reserve forces are mobilized and
fresh defensive substances brought into action. These are

(1) antitoxins, whether to exotoxins or endotoxins, or the toxins
may be dealt with in one or other of the methods discussed under
Antitoxic Immunity ; but in any case it is probable that the develop-
ment of some degree of toxic immunity, especially, perhaps, of
the leucocytes, must precede the destruction of the bacteria.

(2) Increased amounts of alexin-opsonin having a specific action
on the organism in question. We have already noted some of
the gaps in our knowledge of this subject, and shall, perhaps,
again have occasion to refer to the difficulties in explaining
its action, but there can be little doubt of its importance in

346 LOCAL LESIONS

generalized infections. We notice, for example, that in pneumonia
it remains at a low level during the attack, and increases rather
suddenly at the time of the crisis ; the pneumococci disappear
from the blood at this period, though they remain living in the
lung until much later. In cases which recover by lysis, on the
other hand, there is a gradual increase in the opsonic index, and in
acutely fatal pneumococcal septicaemia the index falls gradually
and continuously. When dealing with generalized infections the
opsonic index (as far as our researches have gone at present) may
be taken as a very rough guide to the degree of immunity and
chance of recovery. This does not apply in local infections, but
here we are probably justified in believing that the higher the
index, the less the chance of generalization. It is worthy of
notice that the disease — cerebro-spinal meningitis — in which the
highest indices (sometimes as high as ten) are found is one in
which bacteriaemia very rarely occurs, though by analogy with
other similar diseases we should rather expect it to do so. As
regards the source of this increased amount of alexin-opsonin,
there is nothing to be added to what has been already stated.
(3) In some cases recovery may be brought about by the produc-
tion of thermostable opsonins, whether amboceptor (as appears
most probable) or agglutinin. This is rarely, if ever, the case in
the pneumococcic diseases, but probably occurs constantly in the
maladies, like typhoid fever and cholera, in which agglutinins and
bacteriolysins are formed in large amount. We may, indeed,
classify diseases roughly into two main groups in this respect :
(i) those in which the acquired immunity is due (inter alia) to the
presence of thermolabile opsonins : in these we may expect the
degree of the immunity to be but slight and its duration short;
and (2) those in which it is due to true antibodies : here we may
expect the opposite conditions to hold. Pneumonia and typhoid
fever may be taken as examples. We do not know accurately
the duration of the immunity to the latter disease, but the anti-
bodies on which we believe it to depend may be traced for long
periods, whereas the blood returns to its normal state very shortly
after recovery from a pneumococcic infection. And experience
with typhoid vaccine leads us to suppose that the immunity to that
disease lasts a year or more. The occurrence of relapses may
seem to argue that the protection is in reality very evanescent,
but other interpretations are possible.

The processes which occur in local lesions are similar in

ON IMMUNITY TO BACTERIA 347

nature, but modified by their occurrence in the tissues and not in
a fluid medium, and they tend to deviate more from test-tube
conditions than do the blood infections.

At the commencement of the infection the conditions are quite
comparable to those of a haemic infection. Leucocytes are soon
attracted into the area, and, if the bacteria are not too virulent,
may remove them entirely, the infection being thus cut short at
its very commencement. Probably myriads of slight wounds are
thus dealt with. But when the bacteria resist phagocytosis, and
form virulent toxins, a new set of factors are brought into
existence. These have been glanced at previously, and will now
require further discussion.

The dilatation of the vessels of the infected region and the
consequent acceleration of the blood-flow is, as has been pointed
out, altogether a conservative influence, having for its object the
removal of toxins and their dilution in the general mass of the
blood. But in all cases except the most trivial the local reaction
extends further, and the more extensive changes tend, partly at
least, to favour the spread of the infection by shielding the
bacteria from the full action of the blood. In other words, the
middle stages in the evolution of an inflammatory lesion indicates
a victory for the bacterium in so far that it has succeeded in
altering the tissues to such an extent as to shield itself from the
action of the protective forces. At a still later stage (in lesions
going on to recovery) the tissues and juices gain the victory, and
there is an alteration in the nature of an increase in the local, and
usually of the general, immunity.

Let us trace what we know of this process in the case of a
small staphylococcic lesion. The early stages consist of the usual
inflammatory reaction. The increased access of blood, increased
number of leucocytes, and probably increased amount of opsonin,
present in the tissues may suffice to cut short the process, the
whole of the bacteria being removed by phagocytosis. In this
case the only clinical signs will be those of acute inflammation,
resulting in rapid recovery.

If, however, the bacteria are present in too great numbers, or
are too violent, or if the immunity (local or general) is deficient,
the toxins formed will kill the tissues in the neighbourhood, and
the bacteria will then lie in a smaller or larger mass of necrotic
tissue, which is surrounded by an inflammatory zone. The tide
has now turned very definitely in favour of the bacteria, for four

348 SEQUENCE OF EVENTS IN LOCAL LESIONS

reasons : (i) There is a mechanical obstacle to the access of
leucocytes to the bacteria, which make their way through the
slough with some difficulty. In ordinary healthy tissues, permeated
by capillaries, the leucocytes have never very far to go to reach a
given area; but when death of these tissues has occurred, the
leucocytes have to crawl in from the still patent capillaries of the
periphery, a distance, maybe, of several millimetres. (2) The
bacteria in the centre of the lesion have time to elaborate a
powerful toxin, which is not diluted by the blood or lymph, since
the vessels are all thrombosed, and can only escape by diffusion.
As a result, there is a central zone where the toxin is present in so
high a concentration that it may either repel the leucocytes by
negative chemotaxis or kill them outright. The latter process
certainly takes place, as the large number of dead bacteria present
in pus sufficiently proves. (3) There is an insufficient supply of
opsonin, and perhaps of other defensive substances also, in the
fluids at the centre of the lesion. This is brought about partly by
removal of these substances by the bacteria of the region. In the
case under consideration the staphylococcus opsonin is absorbed
by the cocci which, in the absence of the leucocytes, it is power-
less to injure. Another possible factor in the removal of these
substances is the action of proteolytic enzymes, the defensive
materials being digested and rendered inert before they reach the
bacteria. Lastly, there is reason to think that opsonins make
their way with difficulty through the inflamed tissues. Thus
Bulloch found that the serum or liquor puris which was collected
from an abscess immediately after it had been cleansed was almost
devoid of opsonic power, and this is the case with the fluid portion
of pus in general. In some cases I have been able to demonstrate
the existence of an antiopsonin, perhaps consisting of cast-off
receptors of the bacteria. (4) There is an increase in the virulence
of the bacteria, due to conditions already discussed.

The absence of opsonin from the fluid at the centre of the
lesion suggests several considerations of some importance. We
see, for instance, that it is not sufficient in the cure of a local
lesion for there to be an abundant supply of opsonin in the blood :
two other factors are required— a permeable lesion and a region
suited for the activity of leucocytes. This is very well seen in the
case of tubercle, where the opsonic index may be greatly raised (to
i '8 or more), and the lesion show no indication of recovery. There
is, as a rule, little or no tendency of the disease to spread or

ON IMMUNITY TO BACTERIA 349

become generalized with a high opsonic index, yet even this may
occur. We may trace a certain degree of parallelism between the
ease of access of opsonin to all parts of the local lesion and the
likelihood of a cure following an elevation of the opsonic index.
For example, there is, according to TunniclifFe, a definite relation
between the rise of the opsonic index and the disappearance of the
membrane in diphtheria, the latter clearing rapidly when the index
rises above normal. Here the conditions are these : there is a
membrane, a few millimetres thick, composed of dead material, in
which the specific bacilli are elaborating their toxins ; below this
there is a zone of acute inflammation, hyperaemic and infiltrated with
phagocytes. Now it is easier for the toxins to diffuse outward and
be washed away by the secretions of the mouth than to pass
inwards, and probably but a small fraction actually formed reaches
the blood-stream. On the other hand, it is comparatively easy for
the protective substances in the serum to pass outwards, the con-
ditions being quite different from those in a closed abscess, where
opsonins, etc., can only reach the centre of the lesion by diffusion,
and not by actual transudation. Here there is a constant stream of
lymph from the pervious vessels to the free surface. Thus the
low concentration of the toxins renders it easy for the leucocytes
to gain access to the bacilli, and the latter have abundant oppor-
tunities of becoming opsonized ; hence the conditions for successful
phagocytosis are produced.

A similar train of phenomena follows the free drainage of an
acute abscess. The toxins escape outwards, and are no longer
forced into the tissues, and at the same time there is a flow of
protective lymph into the abscess cavity, and consequent sensitiza-
tion of the bacteria, and the removal of the toxins allows the
phagocytes to act. If, however, the wall of the abscess is very
thick and impermeable, it may be, as in Bulloch's experiment,
that the lymph which exudes is deprived of its opsonin and other
protective substances during filtration, and the conditions are then
less favourable.

Now consider a large staphylococcic lesion completely embedded
in the tissues. There is a large core of dead material, in which the
cocci constantly produce toxins, and in which they are completely
shielded both from fresh lymph or plasma and from leucocytes.
Here we should expect the lesion to be progressive, no matter how
high the opsonic index, and this is usually the case. Hence, other
things being equal, it is in the smaller lesions that we may expect

350 SEQUENCE OF EVENTS IN LOCAL LESIONS

benefit from vaccine injections which raise the index, and more
especially in cases in which there is a free drain for the toxins —
ulcers, open abscesses, sinuses, etc., or in those in which there is
no dead material (slough or caseous matter), and the blood is
brought into close contact with the bacteria. Small isolated
tubercles in the iris often yield rapidly to tuberculin injections*
whereas large caseous glands are most refractory.1 Chronicity is
also an important factor : a long-standing thick-walled abscess is
less amenable to treatment than a small freshly formed boil.

In order to aid this flow of plasma or lymph laden with pro-
tective substances through the lesion, Wright has suggested the
exhibition of substances such as citric acid or leech extract, which
have the power of diminishing the coagulability of the blood. In
certain cases the beneficial effects of these remedies may be most
striking. And the local use of hot fomentations, etc., which dilate
the vessels, has a similar effect ; in addition to which they probably
aid phagocytosis by raising the temperature of the part.

Next as regards the other pre-existing defensive substances of
the plasma — the bacteriolysins. In the case of the staphylococci
there is no reason to think that they are of any importance, since
most authorities hold that there is no proof that serum has any
bactericidal action on these organisms under any circumstances.2
And when the blood does normally contain the amboceptor-
complement apparatus, the establishment of the lesion sufficiently
demonstrates the insufficiency of this apparatus for defence.

At a later stage, supposing the pre-existing defences — opsonin
and leucocytes, etc. — fail to bring about cure, a second series of
factors come into action. These are the antibodies, the chief
being antitoxin, amboceptor, and thermostable opsonin, the latter
being possibly either agglutinin or amboceptor, as we have seen
already. These usually take about a week to be produced, and
may be looked upon as the second line of defence.

As regards their place of production, this may be remote or
local. The sites of general production of the antibodies have

1 In saying this I do not wish to imply that the cure in these cases, and after
the use of vaccines in general, is brought about solely by an increase in the
opsonins of the blood,

2 I must point out, however, that in old staphylococcic abscesses it is
common to find cocci which have lost their power of retaining Gram's stain,
an invariable proof (where applicable) of the early stages of bacteriolysis.
This is probably due to the action of the peptic enzyme formed by the
leucocytes.

ON IMMUNITY TO BACTERIA 351

been discussed already, and we have seen reason to believe that
the chief regions in which it takes place are the lymphoid organs,
the lymph glands, bone-marrow, and spleen especially. These
organs appear to have as one of their functions the elaboration of
defensive antibodies as an internal secretion in response to toxins
and other bacterial products (free receptors, etc.) circulating in the
blood. We have also seen that the phagocytes, and especially the
polynuclear leucocytes, may act as sources of these bodies, but
that the evidence is less convincing than in the case of the
lymphoid organs.1

The local production of antibodies is probably more important,
since it takes place at the spot at which these substances are
required. As an example we may take Romer's demonstration of
antitoxin in the conjunctiva as a result of installations of abrin,
when no antitoxin had yet appeared in the general circulation.
The source of these antibodies, as first suggested by Whitfield, is
probably the lymphocytes, which form so important a factor in
chronic inflammatory lesions, and which occur after a few days in
the periphery of acute lesions. These cells do not act as phago-
cytes in vivo, though they may do so under experimental condi-
tions, and on any other hypothesis their presence in inflammatory
lesions is entirely inexplicable. We may feel certain, however,
that they are attracted there for some good purpose, and it is in
the highest degree probable that this purpose is the elaboration of
protective antibodies. It seems fairly certain that the source of
the haemic antibodies is the lymphocyte cells of the adenoid
tissues, and the cells of a chronic inflammatory lesion are exactly
similar. And in many cases if we examine the region in which
inflammation has taken place some time previously, we shall find
that the cells which were attracted thither have become organized
into definite adenoid tissue.

It seems probable, too, that these small round cells are the basis
of the local acquired immunity, concerning which so little is
definitely known. After the bacteria have been destroyed the
polynuclear leucocytes remove the debris of dead tissues, etc., and
then retire, but the lymphocytes persist in the region for long
periods. During this time they are perhaps continuously elaborat-
ing small amounts of antibodies, and are certainly on the spot and
ready for action should a fresh infection occur. It is probable,

1 More recent researches tend to point to the endothelial cells as the more
probable source of origin of the antibodies.

352 SEQUENCE OF EVENTS IN LOCAL LESIONS

too, that they may have become so altered in virtue of having
once been stimulated to produce antibodies that they can do so
more quickly and easily than normal cells. Thus Cole has shown
that an animal which has once been inoculated — e.g., with typhoid
bacilli — will produce on a second injection a much greater output
of antibodies than will a normal animal, and that this increased
defensive reaction persists for months, long after the animal has
apparently become normal. This observation is probably of the
highest importance, both as regards general and local acquired
immunity.

Thus in the case of a local lesion we may distinguish two
stages : ( i ) That in which the natural resources of the body alone
come into action, and in which the main mechanism for com-
bating the bacteria is phagocytosis, the main cell the polynuclear
leucocytes, and the main defensive substance the thermolabile
opsonin. In this stage there is in general a decline of the local
immunity due to the action of toxins on the tissues, and the only
rise of the general immunity, if present at all, is due to an increase
of this opsonin. (2) The second stage is that of the antibodies.
The factors taking part in the first stage persist, but there is a
new defensive cell, the lymphocyte, and new defensive substances,
the antibodies. Here the immunity, both local and general, tends
to be raised. We may also distinguish a third stage, in which the
polynuclears have retired and the lymphocytes remain— a stage in
which the natural immunity of the part is reinforced by the actual
or potential opsonins due to the lymphoid cells, and in which
these are better equipped (in virtue of their previous training) to
produce a large amount of antibody in response to a slight
stimulus.

Lastly, we must study briefly the defensive reactions of the
peritoneum, a region which has been subjected to a full investiga-
tion and which resembles the tissues in some points and the blood
in others. In all probability the process of absorption from the
other serous sacs is, in general, similar. We have already glanced
at the subject, and in what follows much use has been made of
the admirable researches of Buxton and Torrey. The case of an
intraperitoneal injection of a pathogenic organism of moderate
virulence into an animal possessed of a certain amount of natural
immunity will be considered — e.g., of a laboratory culture of
B. typhosus into the peritoneal cavity of the rabbit.

Under these circumstances, one or other of two trains of events

ON IMMUNITY TO BACTERIA 353

may take place. In the first the immunity appears to depend
mainly on bacteriolysis, in the second on phagocytosis.

In the former case an enormous number of bacilli are killed in
the space of a few minutes, very few being recoverable from the
peritoneum by washings with normal saline solution, and when
this takes place it is found that few or no bacilli make their
way into the blood or organs. In such cases the typhoid bacilli
which are recovered show marked signs of degeneration, similar,
though less marked, to those seen in cholera vibrios in the classical
PfeifFer's reaction; the staining material becomes collected into
the centre of the sheath of the bacillus, which subsequently breaks
down into minute granules. This process is usually entirely
extracellular, but after a time some of the altered bacilli may be
taken up by the phagocytes. It would appear at first sight, there-
fore, that in this case the complement -amboceptor mechanism
is present in the normal peritoneal fluid, ready to bring about
immediate bacteriolysis. This, however, is by no means certain,
and it is highly probable that the complement does not exist in
this fluid, and that its appearance follows the injection, and is due
in some way to the leucopaenia which has been already noted.
The nature of this leucopaenia has been much discussed, and is of
some importance in regard to this question of the nature of the
complement. Metchnikoff and the French school generally
maintain that the diminution is due to the actual destruction of
the leucocytes— phagolysis — and that during this destruction or
solution of these cells alexin or complement, or, as Metchnikoff
calls it, microcytase, is set free. As a matter of fact, there is no
evidence that this phagolysis does occur, and more recent researches
appear to point to another solution of the leucopaenia. It is true
that the fluid withdrawn from the peritoneum by means of a
pipette is markedly deficient in cells, but those that are present
do not show the marked degenerative changes we should expect
if they were undergoing rapid solution. And in any case, it is
difficult to believe that an injection of substances such as normal
saline, which we know preserves the leucocytes in a lively condition
for many hours, should have such a profound destructive effect on
them in the peritoneum. Normal saline, it may be pointed out, is
capable of producing a most marked leucopaenia when injected
into the peritoneal cavity.

The most probable explanation of the phenomenon is that first
suggested by Pierallini, and corroborated by Buxton and Torrey.

23

354 INFECTIVE PROCESSES IN THE PERITONEUM

According to them, there is a deposition of masses of fibrin on
the omentum, and in and upon these masses there are numerous
polynuclear leucocytes.

The formation of fibrin, it need hardly be pointed out, is
universally regarded as being due to the liberation of fibrin
ferment by the leucocytes. It is not yet settled beyond controversy
whether this is to be regarded as a process of secretion, a vital
phenomenon, or a process occurring only during the solution —
phagolysis — of the leucocytes. The point is of no importance :
the remarkable feature is that, according to the opinions we have
considered as most probably correct, complement is set free at the
same time as fibrin ferment. The deposition of fibrin on the
surface of the omentum may be taken as sufficient proof of the
liberation of this substance, so that the rapid destruction of the
bacteria which occurs in the peritoneum in some cases cannot be
regarded as constituting definite proof that it occurs preformed in
that situation.

In some cases, therefore, bacteria injected into the peritoneum
are killed rapidly, almost instantaneously, by a process resembling
bacteriolysis, and doubtless due to amboceptor existing in the
peritoneal fluid as such at the time of the injection, together with
complement which is probably set free from the leucocytes as a
result of the presence of the foreign body, but which may possibly
also be present in the normal state. As an example of this process,
Buxton and Torrey quote the case of a rabbit which received an
intraperitoneal injection of about 4,000,000,000 typhoid bacilli, and
was then immediately killed, the peritoneum washed out with
normal saline solution and plated out. The fluid contained only
about 1,000 bacilli, and there were none in the blood or internal
organs. In a case which was allowed to survive for two hours no
bacilli were found in any part of the body. It is most remarkable
that, with all this sudden, almost explosive, solution of bacilli there
were no symptoms indicative of the liberation of endotoxins, the
temperature remaining constant. The culture was one of very
moderate virulence.

In other cases the whole process differs, and the defence of the
body appears to be entrusted to the phagocytes rather than to the
bacteriolytic substances, and in these a most interesting train of
phenomena occurs. Some destruction of the bacilli may occur in
the peritoneum, showing that the bacteriolytic properties do not
fail entirely, but comparatively few organisms are destroyed by

ON IMMUNITY TO BACTERIA 355

this process ; instead, large numbers of bacilli make their way
into the circulation by a route not fully known. These rapidly
decrease in number, the diminution being quite noticeable within
half an hour, and in six hours or less the blood may be entirely
sterile. There is, further, a remarkable amount of deposition of
the bacilli in the organs, especially the liver, spleen, lung, and
bone-marrow ; and in these situations the numbers of organisms
decrease for about two or three hours, and then show a very
definite rise, which lasts for two or three hours more, and the
organisms may attain numbers approximating to those present in
these organs immediately after the injection. The numbers then
gradually decrease for the next two days or so. Here we may see
in a very clear manner an example of the mobilization of the
defensive forces discussed previously. The diminution which
takes place immediately after the access of the bacilli to the
organs is doubtless due to the bacteriolytic substances, either
present in the blood as such, or immediately available after
phagolysis or secretion of alexin. After a time these substances
are exhausted, and it is only after some hours that a fresh supply
is elaborated, and destruction of bacilli continues. Further, the
process that now takes place is mainly phagocytosis, whereas the
earlier defensive process was bacteriolysis. This sequence of
events is probably to be interpreted as follows : As soon as infec-
tion takes place the small amount of immune body naturally
present in the blood combines with the bacilli, and, since some
complement is present, bacteriolysis occurs. In this way all the
immune body is removed in combination with the bacilli, and all
the complement is also absorbed, since the conditions for the
Bordet-Gengou phenomenon are present, and complement not
actually required for solution of bacilli will be taken up, as well
as that which is actually used in the process. Hence all the
defensive substances are removed, and the bacilli can flourish
unchecked for a time. Soon, however, more complement is
produced. This has little or no power of producing bacteriolysis in
the absence of amboceptor, but it can, and does, act as opsonin,
and abundant phagocytosis occurs. In most diseases this stage
would be marked by leucocytosis, but in typhoid fever this does
not occur. But we shall shortly see that a polynuclear reaction
does take place in the peritoneum, if not in the other regions.

In the series of researches we are discussing the development
of the last line of research — the true antibodies — was not studied.

23-2

35^ INFECTIVE PROCESSES IN THE PERITONEUM

These only come into action later. In the case of a mild infection
the whole process of cure may occur without their appearance, and
they are only of importance in so far as they are the cause of the
subsequent immunity.

Let us now turn to the sequence of events in the peritoneum in
these animals in which explosive bacteriolysis does not occur, or
only to a small extent, in that region. Here some phagocytosis is
found to take place in the cells lying free in the peritoneal fluid,
(the changes in which have been noted previously), but the greater
part of the process— and this is significant in view of the theories
which we have adopted with regard to the liberation of comple-
ment-opsonin during the process of coagulation — takes place in
the masses of fibrin found on the omentum. At first the bacilli lie
free in this deposit, often lying more or less parallel ; but in a
short time vast numbers are taken up by the macrophages, and
very few free organisms may be seen in an hour or two after the
injection. Then one of two series of phenomena may occur. The
animal may recover ; and in this case the bacilli which are con-
tained in the macrophages become pale and granular, and soon
disappear altogether, and the fibrin becomes infiltrated and eroded
by polynuclear leucocytes. Or the animal may die ; and in this
case the bacilli, both those which have been ingested and those
which have not, commence to grow rapidly, and the organisms,
after diminishing in numbers for a few hours, proliferate so quickly
that the animal dies in some twenty- four hours. In these cases
the bacilli which have been ingested are not destroyed, and retain
their normal appearance and staining reactions throughout.
Here it is obvious that the first line of defence, the rapid initial
phagocytosis, has been unsuccessful, and that the secondary
reserve forces have not had time to come into action. It is
specially pointed out by Buxton and Torrey that in these cases the
invasion of the fibrin by the polynuclear leucocytes does not occur,
a fact strongly confirmatory of the view that it is these cells that
give rise to the opsonin or complement, the deficiency in which
brings about the lethal issue.

Here we must conclude a short and altogether inadequate con-
sideration of a subject of the highest importance. Much more
might be written on the subject, but the very fact that the
phenomena do not lend themselves to a brief discussion only
shows how very imperfect is our present knowledge of the events
which actually take place in the juices and tissues during the

ON IMMUNITY TO BACTERIA 357

process of recovery from an infective process. The mechanisms
themselves are now fairly well known, the nature of the substances
concerned therein thoroughly investigated, and their sources ascer-
tained with some degree of probability; but when we come to
apply our knowledge to the actual events of the living body, our
difficulties begin in earnest. It is in this field of research that
future advances are most to be expected.

In conclusion, let us emphasize the enormous importance which
we have been led to attach to the leucocytes in the struggle against
the infective diseases. In all probability the substance we call
alexin, thermolabile opsonin, or complement, and which, in one
or other of its actions, constitutes the first line of defence, is
formed by the polynuclear leucocytes. In itself its action is but
slight, and it requires to be supplemented either by the leucocyte
itself, in case the reaction is mainly phagocytic, or by the action
of amboceptor, which we believe to be derived from the masses of
leucocytes known as lymphoid tissue. Again, the bacterium may
give off toxin, especially if it has escaped being taken up by the
leucocyte, and in this case these cells exhibit another phase of
their protean activities, for that they can actually absorb toxin in
solution appears quite certain. If this action fails the last resource
is the neutralization of the toxin by antitoxin ; and although this
cannot be regarded as definitely proved, there is at least some
reason to believe that this may be formed in the lymphoid tissues.
Lastly, certain facts would seem to show that it is quite likely
that even the compound of toxin and antitoxin is by no means
inert unless it has been taken up by the leucocytes, and that the
action of the latter substance is simply to prepare the former so
that these cells may attack it. In every phase of the struggle
against the bacteria the leucocyte appears, on certain or presump-
tive evidence, as the particular cell to which the defence of the
body is entrusted.

CHAPTER XIV

PRACTICAL APPLICATIONS
Staphylococcic Infections.

COMMON as is disease due to the staphylococcus (boils, carbuncles,
pustular acne, osteomyelitis, etc.), our knowledge of the inner
mechanism of the pathogenicity of this organism is still somewhat
scanty. The only toxins definitely known to exist are a staphylo-
lysin and a leucocidin. The former is produced in three or four
days in faintly acid broth, and reaches its maximum in about
fourteen days. It is destroyed by a temperature of 56° C., and in
other respects resembles the true toxins. Many animals, and
notably the horse, contain a natural antistaphylolysin, which
neutralizes the action of this haemolysin, and is apparently a true
antibody. It is present in human serum, and may perhaps
account for the fact that anaemia is not a striking feature of
Staphylococcic diseases. Animals from which it is absent, such
as the rabbit or goat, can be made to produce it by immuniza-
tion with solutions of the lysin. Leucocidin is produced under
the same conditions as the haemolysin, and the two are usually
produced side by side ; but they are distinct substances, the former
being the more easily destroyed by heat. Its action is not
entirely specific : it kills the leucocytes and dissolves them, but
has also some action on the ganglionic and other cells. As in the
case of the hsemolysin, a natural antibody exists in the serum of
man, the horse, etc., and can be prepared from other animals.

These substances, especially perhaps the latter, may play some
role in the production of disease, in that they may inhibit phago-
cytosis by a poisonous action on the leucocytes. But there
is little doubt that there is some other toxic body, probably an
endotoxin, since young cultures killed by heat, as used for vaccines,
are decidedly toxic, causing local irritation, and, if used in large
doses, a rise of temperature. These vaccines contain no haemo-
lysin or leucocidin.

358

PRACTICAL APPLICATIONS 359

Immunity. — The resistance against staphylococci appears to be
almost entirely dependent on phagocytic action. Ordinary
methods fail to demonstrate any bactericidal action, either in
normal or immune serum. Andrewes and Gordon, however, by
the use of special methods, have succeeded in showing that such
an action, though very slight, occurs even in the serum of normal
animals, and to a greater extent, though still but slight, when the
animal has been immunized. Perhaps the antibodies to the soluble
toxins mentioned above may have some protective action.

The staphylococcus is very susceptible to phagocytic action,
and there is a general correlation between the opsonic index and
the stage of evolution of the lesion, as shown in Fig. 64. The
opsonic index is greatly raised in the serum of immunized animals,
and it would appear that injections of dead staphylococci have the
power of increasing both the thermolabile and thermostable
opsonin. The duration of the immunity is not definitely known,
but is certainly short.

Diagnosis. — This is made entirely by the demonstration of the
infective organism — an easy task. In a few cases the opsonic
index may afford some help, being usually low during the acute
stage.

Agglutinating sera can be prepared, but their action is not
powerful, and this reaction is useless in diagnosis.

Treatment. — Antistaphylococcic sera have been prepared, but
the results of its use are not encouraging, and the only valid
method is the use of vaccines. These should be prepared either
from virulent cultures derived from severe boils or carbuncles
or from the lesion it is desired to treat. In practice it is a good
plan to commence with a stock vaccine prepared from a virulent
culture, and to prepare an autochthonous vaccine for use if the
first does not succeed. In general opsonic control is not necessary,
and the doses may vary between 100 and 1,000 millions, repeated
every ten or fourteen days.

Streptococcic Infections.

The toxin of the Streptococcus pyogenes is still not definitely
known. We have already referred to the haemolysin which it
produces. There is some reason for believing that this substance
has some clinical action in acute infections, and that it is produced
to a greater extent by virulent than by non-virulent cultures.

360 STREPTOCOCCIC INFECTIONS

Apart from this, the existence of a soluble toxin is doubtful.
Filtered broth cultures of the organism are poisonous, but
most observers have been^able to produce immunity therewith,
and the toxic substances are probably merely simple metabolic
products. Parascandalo, it is true, claimed to have been more
successful, but his results have not been corroborated, and there
is some reason to believe that the nitrate which he used was not
free from bacteria, living or dead, and that his animals were
really vaccinated with a vaccine. The toxin is probably an
endotoxin, though of this but little is known. The bodies of the
bacteria are highly irritating, and very small doses of vaccines
have to be given at the commencement of the treatment.

Diagnosis. — This is made by the demonstration of the micro-
organism in all cases. Agglutinating sera may be prepared,
but the appearance of the property in the serum in human
disease is not sufficiently marked or constant to be of value.

Immunity. — Human blood appears to contain an antilysin which
neutralizes the action of streptocolysin, and it is quite possible
that this substance may play some part in the defence of the body
against infections. There is also a little evidence for the forma-
tion of a true antitoxin against the actual and efficient toxin of the
streptococcus, whatever this may be. This is the fact that in
some cases, though unfortunately not in all, the injection of anti-
streptococcic serum is followed by a marked immediate benefit in
cases of septicaemia, etc. The temperature may fall, the delirium
and headache pass off, and the pulse improve in a striking manner
within an hour or two of the injection. Now Wright has pointed
out that but little is known of the mode of action of the serum,
and that it may contain a toxic ingredient and really be a vaccine
in disguise. If this is the case we should not expect it to produce
good effects so early as actually happens in favourable cases, in
which its action resembles that of diphtheria antitoxin, and
certainly suggests that some poison is neutralized. Probably,
therefore, acquired immunity to streptococci is partly antitoxic or
anti-endotoxic.

Bacterial immunity is probably mixed, partly bactericidal or
bacteriolytic and partly phagocytic. The serum of normal
animals has usually some bactericidal action, and this can be
raised by immunization to a high figure ; further, the serum of a
immunized animal is a powerful agent in conferring passive
immunity, a property not belonging (apparently) to sera of high

PRACTICAL APPLICATIONS 361

opsonic value. But probably phagocytosis is of the main value
in the struggle against the streptococcus in the body. The
organisms are very readily taken up by the leucocytes in presence
of serum, and are frequently seen inside the pus cells in suppura-
tive diseases- In general the opsonic index may be taken as a
fair index of the degree of immunity, since it is usually low during
the course of the infection and rises as cure is brought about.
This has been already shown in the case of erysipelas : the index
returns to normal in one to three days after the rise, indicating
that the immunity in this case is but transient (see Fig. 63). This
is of interest in view of the frequency with which this disease is
recurrent. I have seen a case in which attacks occurred every
three weeks for a year. But the solid immunity conferred by
vaccination and the use of massive cultures is more lasting, due
probably to the formation of immune bodies and thermostable
opsonins.

Local immunity probably also plays a part of great importance
in streptococcic infections, as shown by the fact that the ery-
sipelatous lesion usually spreads at one margin and heals at
another simultaneously. Yet even here the spread may be due to
failure of access of the blood to the bacteria. This is suggested
by the frequent success of the local use of iodine and other
rubefacients in arresting its spread. A similar effect may some-
times be seen in chronic cases by the use of citrates or one or
other of the means suggested by Wright for lowering the coagu-
lability of the blood. A case in which the disease had been
present for five weeks recovered in two days when treated in
this way.

Treatment. — Protective treatment is not employed. It is, how-
ever, interesting to. note a fact pointed out by Sir James Paget
many years ago. Pathologists are as a rule more or less immune
to septic infections, and these comparatively rarely occur in those
who are in the daily habit of performing post-mortem examina-
tions. The attack which proved so nearly fatal to him occurred
after a long absence from the post-mortem room. Probably most
pathologists are vaccinated against the disease.

The curative treatment (apart, of course, from that of the
local lesion, if any, and the use of general toxic and sustaining
measures) resolves itself into the question whether a serum or
vaccine should be used. This question cannot be definitely
settled at the present time, and the recommendations which are

362 STREPTOCOCCIC INFECTIONS

given here may require great alteration in the future. In several
and rapidly progressing septic cases, such as those due to post-
mortem wounds, etc., the use of serum is indicated, and offers the
best chance of success. In view of the very great benefit which
is frequently derived from this measure, it appears highly im-
proper to wait until a vaccine is prepared. This should be taken
in hand at once, so that a homologous vaccine may be ready if
subsequent events suggest its use, but in the meantime serum
should be employed. The use of minute doses of a stock vaccine
(5 to 10 millions per dose) may be considered, and there
appears to be no reason why they should not be given along with
the serum. If practicable, opsonic estimations should be taken,
but the streptococcus is not as a rule a very satisfactory organism
to work with. In the absence of such observations, doses of the
size mentioned above may be given every three or four days.

In chronic septicaemia or ulcerative endocarditis of streptococcic
origin the use of vaccines offers more prospect of success, though
even here cures have been brought about by means of the serum,
which should be used whilst cultures are being taken from the
blood and a vaccine prepared. At present it seems desirable to
use careful opsonic control in cases of this class. An excellent
illustrative case treated by Douglas should be referred to for
details. Here the doses were 5 tc 12 millions, and the
injections were given each time the index fell. The case was
certainly one of septicaemia, but the evidence for the existence of
ulcerative endocarditis is not conclusive.

In chronic local inflammatory disease of streptococcic origin
vaccine treatment is probably the best, but here ordinary methods
(especially Bier's) may be of more advantage. In the absence of
opsonic control, the doses may begin at 10 millions and rise to
100 millions, and be given once a week.

Erysipelas does not usually require specific treatment, and in
severe cases serum often answers well. The use of vaccines has
not been tried on a sufficiently large scale to allow us to judge of
its value. A case under my care of recurrent erysipelas, in which
attacks had occurred about every three weeks with some degree
of regularity for a year, was apparently cured by a few doses of
vaccine, commencing at 25 millions and rising to 50 millions, and
administered once every ten days. In some chronic cases of
erysipelas it is probable that there is a considerable amount of
haemic immunity, but the protective substances are unable to

PRACTICAL APPLICATIONS 363

reach the bacteria. In these the use of citrates or citric acid (as
recommended by Wright) is occasionally of great value.

The mode of action of antistreptococcic serum is still quite
uncertain, and its use is purely empirical. Probably in most
cases it acts as a bacteriolytic substance. Wright has suggested
that it may contain free toxin, and so act as a vaccine ; but the
fact that it will passively immunize animals treated with it is
adverse to this supposition. It is prepared by gradually immuniz-
ing horses and other animals until they can withstand large doses
of the living organisms. The differences between the various
brands mainly arise from the nature of the culture used for the
purpose. The earlier sera to be prepared were procured by the
injection of a single strain of streptococcus isolated from a case
of erysipelas, puerperal fever, etc. It was found, however, that
this serum was not successful in every case, and attempts were
made to improve it. Starting from the supposition that the
reason for the non-success was the fact that the streptococcus
used was not exactly the same variety as that causing the infec-
tion in the patient, polyvalent sera were prepared by treating horses
with cultures from many various sources, and it is sera of this
nature that are now in general use. Another explanation was
that the cultures used were not of sufficient virulence, and as
antibodies to an attenuated organism may not act against the
same culture when the virulence is exalted, strains of great
potency were obtained by means of passage through rabbits. Yet
a third method has been adopted, based on the fact that an
organism which is highly virulent for one animal may be harmless
to another species. Here passage is avoided, and the cultures
used in the immunization of the horse are taken direct from
human disease, and are used soon after isolation— «'.£., before
virulence has been lost by prolonged cultivation in vitro. There
is no very clear evidence that any special advantage attaches to
any of the sera thus prepared ; in any given case it is a matter of
chance which sample will prove successful. The criterion as to
the suitability of a given sample of serum is purely an empirical
one. The patient's pulse, temperature, nervous condition, etc.,
should be watched with the closest attention, and any definite
improvement occurring within the first twenty-four (and usually
within the first four) hours of the first doses should be an indica-
tion to continue with the use of the same brand of serum. In
some cases each injection causes a slight but definite rise of

364 PNEUMOCOCCIC INFECTIONS

temperature, and the symptoms undergo a brief exacerbation,
followed by an improvement in the general condition. The
meaning of this phenomenon is somewhat doubtful, Wright hold-
ing that the rise is due to the presence of toxin in the serum. It
seems, however, more probable that it is due to a solution of some
of the streptococci and consequent liberation of endotoxin. It is
not a contra-indication'bto the use of the serum, though it may
suggest care in its use. If no obvious result follows the use of
the serum, the brand should be changed, another specimen being
administered forthwith.

Antistreptococcic serum is, as a rule, not standardized, and the
initial dose should not be less than 10 c.c., whilst 20 to 25 c.c. is
not too much. Subsequently 10 c.c. may be given each day as
long as it appears to be of benefit. In a generalized infection it
seems reasonable to give the injections intravenously, but there is
no proof that the method is of more value than the ordinary
process of injecting it into the cellular tissue of the flank.

Good results have been obtained in severe cases of scarlet fever
by the use of serum obtained by the use of streptococci obtained
from scarlatinal throats (S. conglomerate} . It appears tolerably
clear in this disease that a homologous serum is of advantage, the
results being better than those got by the use of ordinary anti-
streptococcic serum.

Pneumococcic Infections.

The marked remote toxic symptoms frequently met with in
pneumonia would suggest that the pneumococcus forms an
exotoxin, and there is a certain amount of experimental support
for this supposition. The Klemperers made use of nitrates in
their experiments in the production of an antitoxic serum, and
found them fairly toxic, and possessed of undoubted immunizing
powers ; and similar results have been obtained by Washbourn,
Isaeff, and others. But the potency of these filtrates is slight
compared with that of a true exotoxin, and the results obtained
are quite consistent with the supposition that they are due to
slight autolysis and liberation of an endotoxin. The fact that
lesions occur remote from the lung cannot be taken to suggest
that an exotoxin is produced, since we know that in pneumonia
considerable numbers of cocci make their way into, and are
destroyed in, the blood-stream. Vaccines of the dead cocci are
but slightly toxic.

PRACTICAL APPLICATIONS 365

According to Casagrandi, non-pathogenic varieties of the
pneumococcus form a haemolysin, whilst virulent ones do not.
The same observer states that certain cultures produce a leuco-
cidin.

The nature of the immunity to the pneumococcus, and in
especial the causation of the crisis (when present), has been much
discussed, and now seems to be fairly well elucidated. It is
doubtful whether any. bactericidal substances are formed in man
after an attack of pneumonia ; they may be produced in rabbits as
a result of artificial immunization, but are not invariably present,
and it is useless to attempt to explain the characteristic crisis by
their agency. Anti-endotoxin may perhaps be formed artificially,
but its presence in human disease is very doubtful. It is, how-
ever, now tolerably clear that recovery in pneumococcic infections
is brought about mainly by the action of opsonins, and more
especially thermolabile opsonins. This was well shown by
Mac Donald, whose investigations have been referred to already.
The opsonic index is low during the course of the attack, and rises
suddenly to normal at or near the crisis. Eyre distinguishes three
types of opsonic index in different forms of pneumococcic infec-
tions. The first is the pneumonic type, his results precisely
corroborating Mac Donald ; the second is the form in which
recovery takes place by lysis, in which the rise is gradual ; and
the third, in which the disease progresses to a fatal termination, is
characterized by a steady fall in the index. Eyre holds that the
resistance of the individual may be measured by his opsonic
response, as represented by these three curves.

The opsonin present in natural disease is, as mentioned above,
mostly thermootablc ; but in the immunized sera of hypervaccinated
animals thermostable opsonin, the bacteriotropin of Neufeld and
Rimpau, is present. In all probability it is to these substances
that antipneumococcic serum owes its effects. These are not
great — according to Eyre, i c.c. of the most powerful serum yet
obtained will only protect against some 300 lethal doses — as we
should expect to be the case if the immunity conveyed were due
to a rise in the opsonic value of the blood.

Agglutinins are usually found in cases of pneumonia, but they
are not present in large amount (the serum rarely clumps at a
greater dilution than i : 60), though a much more powerful action
may be obtained by artificial immune serum. The reaction is of
little value in diagnosis, since it may be absent even in pneumonia,

366 PNEUMOCOCCIC INFECTIONS

and is often absent in the more localized infections. The disease
is recognized by the demonstration of the diplococcus.

In acute pneumococcic infections (pneumonia, septicaemia,
peritonitis, etc.) attended with severe constitutional disturbance
the use of serum may be tried. Several have been prepared, the
methods used differing somewhat in the different cases ; but, with
the exception of that prepared by the Klemperers, the animals
used are immunized by the use of dead, and subsequently of living,
cultures. The best known are Pane's and Romer's. The latter is
markedly polyvalent, an important point, since various strains of
pneumococci probably differ largely inter se. This was very well
shown by Washbourn and Eyre, who found Pane's serum
protective against four cultures from different sources, but power-
less against a fifth.

The results of the use of antipneumococcic serum have not
been such as to lead to its general use, although several observers
have reported very good effects. According to Tauber, the tem-
perature falls to normal after one to three injections of 20 c.c.,
and there is marked mental and general improvement. Stress
is laid on this by other physicians, and it may perhaps be due to
the cutting off of the supply of endotoxins brought about by the
ingestion by the leucocytes of the pneumococci after opsonization
by the added serum.

The results are sufficiently encouraging to lead us to hope that
the serum may become of more value in the future, possibly as a
result of the discovery of a method by which the dosage and
spacing of the injections and the time for bleeding may be more
scientifically determined.

In general, the use of serum for localized lesions does not
appear to be of much value ; but great benefit has been claimed to
follow the use of Romer's serum in ulcus serpens of the cornea,
used either alone or, as Axenfeld recommends, in conjunction
with a vaccine. The serum is to be dropped into the conjunctiva
or injected beneath it.

Vaccines have also been used in pneumonia and in pneumo-
coccal septicaemia with good results. Thus Eyre has reported
a case of the latter disease in which there were metastases
(purulent) in the subcutaneous tissues, the hip, the gluteal
muscle, etc., and which recovered after five injections. In another
case Eyre notes, what is of great importance, that no benefit
whatever was derived from a stock vaccine, and it was only when

PRACTICAL APPLICATIONS 367

the patient's own culture was employed that improvement began.
It is in general advisable to treat the case with a stock vaccine,
and to commence to prepare a special one if this should fail. In
prolonged cases it is sometimes of benefit to prepare a fresh
vaccine after a time, in order to counteract any possible change of
type of the organism in the body.

As a rule, the first dose should not exceed 50 millions for an
adult and 10 millions for an infant, and, if the disease is febrile
and the symptoms acute, may be decidedly less. Subsequent
doses may be larger, but as a rule it is not necessary to exceed
200 or 250 millions. Probably the opsonic control is of more
value here than in most diseases, but in the absence of this doses
may be given every week or ten days. The opsonic control, it
may be noted, is of more value in acute cases than in chronic
ones, since in the latter the index may be persistently normal or
high, and yet the disease yields readily to treatment. Thus a case
of pneumococcic empyema of the frontal sinus of four years'
duration, which had been twice submitted to operation, had an
index well above normal, and was completely and permanently
cured by four injections of a homologous vaccine at intervals of a
fortnight. As a general rule, for the treatment of a small lesion
in an adult, such as an unhealed sinus from an empyema, the
injections may be 25, 50, 100, and 200 millions, with an interval
of a week between the first two and ten days between the
remainder, of course subject to any indications which may be
derived from clinical observation. It need scarcely be pointed
out that but little benefit can be expected in cases in which there
is a mechanical obstacle to the escape of pus or the closure of an
abscess. A case of very chronic pulmonary abscess, due to
pneumococci, in which the X rays showed a considerable amount
of thickening, derived no benefit from a long and careful course
of vaccine treatment, with and without opsonic control.

Vaccine treatment has been used in acute pneumonia (according
to Allen, routine injections of 25 millions may be given), and is of
especial value in unresolved consolidation, which often clears up
after its use in a most satisfactory manner, the moist sounds
clearing up in a very short time, and the general condition
improving rapidly.

Space forbids mention of all the conditions in which the
pneumococcus, perhaps the most protean of all bacteria in its
pathogenic effects, has been combated by vaccine - therapy.

368 GONOCOCCIC INFECTIONS

Special mention should, however, be made of such serious con-
ditions as ulcus serpens and other diseases of the eye, in which
good results have -been obtained by Allen and others. In general,
all localized pneumococcic diseases should be subjected to vaccine
treatment, if not easily amenable to surgical operation.

Gonococcic Infections.

Here again we have to deal with an organism which exerts its
pathogenic effects mainly or entirely by means of an endotoxin.
The lower animals are entirely refractory to infection with living
cultures, whereas the dead bodies of the cocci produce local
inflammatory changes (peritonitis, etc., according to the region
into which the inoculation is made), or death if the dose is suffi-
ciently large. The toxicity, however, is but slight, and that of
the filtrate from the cultures is extremely small. Several
observers claim to have produced a soluble exotoxin, but there is
considerable doubt as to the interpretation of their results, and in
all probability the substances present in these fluids are simply
traces of endotoxin, or free receptors let loose by the autolysis of
a few of the cocci. Analogy with other diseases would rather
suggest the absence of a powerful exotoxin. In the great majority
of cases the disease is a purely local one, and the symptoms of a
general intoxication very slight. The endotoxin is said to be very
stable, resisting the temperature of boiling water for some hours.

The hgemic indications of immunity are but slight. Bacterio-
lysis has not been demonstrated,1 but there is some evidence to
think that immunization may be due — at least, in some cases — to a
rise in the opsonic index, and since some part of the newly-formed
opsonin is, or may be, thermostable, it is possible that some small
amount of immune body may be produced. According to Allen,
the opsonic index in acute gonorrhceal infections is somewhat
below normal — 0-6 or 07; and when spontaneous cure takes place
it rises gradually, going as high as 1-6. If, however, it remains
subnormal, the case passes on into one of intractable gleet. But
in acute gonorrhceal conjunctivitis in adults the index may be as
high as 2-5, showing, as has been already seen in other diseases,
that a high index is no proof of immunity. Agglutinins also occur
in the serum of immunized animals (Torrey treated a rabbit the

1 Torrey has recently shown that an antigonococcic serum which he has
prepared possesses powerful bactericidal properties.

PRACTICAL APPLICATIONS 369

serum of which clumped at a dilution of i : 700,000), but this does
not occur in man, and the reaction is useless in diagnosis. An
interesting point brought out by Torrey's studies is that gonococcic
cultures exhibit well-marked differences inter se. The serum of the
rabbit alluded to above, which clumped its homologous culture at
i : 700,000, clumped another culture only when diluted fifty times
or less. This is of importance in connection with the prepara-
tions of vaccines and sera. Torrey also showed that antigono-
coccic serum contained a bacterio-precipitin.

Clinical facts would suggest that the immunity to the gonococcus
is of very peculiar nature. Thus, as Ricketts well points out,
recovery is not due to a loss of virulence of the cocci, for they
remain potent to produce infection during all stages of the disease.
Nor is it due to the production of acquired local immunity, unless,
indeed, this is of such a nature that it can be very easily broken
down ; for the patient can be reinfected immediately after an
attack, or whilst the disease is in a chronic stage in some part of
the urethra or its diverticula. It is conceivable that the gono-
coccus is very easily modified by passage through other human
beings, and so altered that it is able to reinfect a person who has
just recovered from an attack caused by it in its unaltered form ;
but this would hardly explain the recrudescence of an apparently
cured discharge after excessive indulgence in alcohol. Even the
relationship between phagocytosis and recovery is not easy to
make out. Unlike other organisms, the gonococci do not appear
to undergo the usual morphological changes indicative of intra-
cellular digestion (loss of staining power and of sharp outline)
usually seen in bacteria after phagocytosis ; nor do the leucocytes
which have taken them up show degenerative changes (Ricketts).
And the fact that the gonococci are practically all intracellular
from a very early period in the disease — indeed, whilst it is in
active progress — would seem to indicate that it is a most inefficient
protective mechanism. Further, there is no obvious difference
in the phagocytosis according to the height of the opsonic
index, which would lead us to believe that, even when the index
is low, there is sufficient opsonin to enable all the available cocci
to be taken up. As far as it goes, the evidence leads us .to think
that the process of cure is due to some local change — possibly to
some exhaustion of a necessary nutrient material — and not to a
general haemic reaction.

Natural local immunity is very marked in the case of this

24

370 GONOCOCCIC INFECTIONS

organism. A spread of the disease beyond the mucous membrane
of the urethra is extremely rare. Less rare, though still un-
common, is extension to the bladder. Haemic infections are
also rare, but do occur ; in them, as in other general infections, it
is difficult to see how the cocci can run the gauntlet of the leuco-
cytes.

Diagnosis. — In most cases, of course, this is made by the
demonstration of the specific coccus— usually an easy task. If no
material is forthcoming (as in gonorrhceal arthritis, etc.), two
methods are available — the opsonic index and the absorption of
complement. According to Allen, the normal index ranges
between 0-8 and 1-2, and in chronic infections it may be low
(down to o!2) or high (up to 2 or more). Further, a dose of
75 millions of dead cocci causes but little disturbance in a non-
gonorrhoeal patient, but a marked rise if the patient is infected.
The method of absorption of complement has been employed
with marked success by Meakins, and appears to be of great
value. The difficulty of the technique, however, renders it much
less easily available than the opsonic method.

Treatment. — The use of serum need hardly be discussed. Some
cases, it is true, have apparently been benefited, but it is always
possible that the serum may have contained specific toxic bodies
which acted as a vaccine.

The vaccine treatment, on the other hand, is of the greatest
possible value. The dose varies between 5 and 500 millions.
In general, large amounts are not well tolerated, especially early
in the treatment. The doses may, of course, be regulated by the
opsonic index, but this is probably not necessary. In acute cases
two or three doses of 50 millions may be given ; but it is in the
chronic infections, especially, perhaps, in gonorrhceal arthritis,
that the method is of especial value. Here the dose may begin
with 50 millions, rising to 500 millions, or even twice this amount,
and the intervals may be seven to ten days. If no benefit is
obtained a further course of treatment under opsonic control may
be administered. The results are usually beneficial in the
extreme.

Gonorrhceal conjunctivitis is to be treated on the same lines as
acute arthritis, but here, of course, ordinary local antiseptic
treatment is all-important. Gonorrhceal iritis is treated like
gonorrhceal arthritis, and the results are usually excellent.

PRACTICAL APPLICATIONS 371

Meningococcic Infections.

Very little is known concerning the toxin of the meningococcus.
Cultures are of very feeble toxicity for animals, and large doses
of the vaccine are usually (though not invariably) well tolerated
by human patients. It is probably an endotoxin which is only
produced under certain conditions. Its effect in the human
subject is mainly a local one, manifested chiefly on the tissues in
or near which the cocci are localized. The disease is in most
cases a local one, the organism being rarely found in the blood
or organs other than the brain and cord.

The frequency with which the organism is found within the
polynuclear leucocytes would lead us to believe that the organism
is combated in the main by phagocytosis ; and this is confirmed
on the whole by the results obtained by a study of the disease by
Wright's method. When opsonic determinations are made using
a virulent culture, such as one derived from the patient himself,
it is found as a rule to be taken up very badly in preparations
made with normal serum and leucocytes, and very well when
some of the patient's serum is present, so that very high indices
are obtained. This is obviously because the cocci have become
animalized, like the virulent pneumococci studied by Rosenow
and others. They require a large dose of opsonin, such as is
present in serum from a patient who is attempting to combat the
disease, and not in that from a normal person. The result is
that the opsonic index of these patients is often extremely high,
figures of 10 or more being common, and 40 has been recorded.
Houston has pointed out that old laboratory cultures which have
lost their virulence are phagocyted more easily in presence of
normal serum, so that the opsonic index of meningitis patients,
as determined by the use of these non-virulent cultures, is com-
paratively low. Houston has proposed this test for distinguishing
between the " true " meningococcus and allied cocci of similar
morphological characters. He determines the opsonic index of a
patient suffering from true cerebro-spinal meningitis against a
normal control, using both a known meningococcus culture and
that under consideration. In what he terms the positive reaction
there is, in the case of the normal blood, very little phagocytosis,
and no agglutination of the cocci which are not ingested, whilst
with the blood from the cerebro-spinal case there is much phago-
cytosis and marked clumping of the free cocci. It would seem,

24—2

372 MENINGOCOCCIC INFECTIONS

however, that the test is simply one of the virulence of the
culture, rather than that of its specific nature. As regards
agglutination, Houston and Rankin point out that this property
also is lost on prolonged cultivation on artificial media ; and as
regards the behaviour of the coccus in opsonic estimations, exactly
the same phenomena are seen when cultures isolated from what
are apparently ordinary cases of basic meningitis are tested against
the blood of the patient and a normal control. It seems more
reasonable to suppose that the coccus from the basic meningitis
cases are either of lower virulence, or that their virulence has
developed along different lines from a common non-virulent
stock, rather than a different species.

The opsonin present in the serum in meningitis is to a large
extent thermolabile. Some thermostable opsonin is present, the
amount being roughly proportionate to the height of the index.

Agglutination is somewhat difficult of study in the case of this
organism, owing to the variations presented by different cultures
in this respect and the difficulty in procuring a homogeneous
emulsion. Tested by ordinary methods, the blood of meningitis
cases does not clump in high dilutions : according to Davis, i : 50
is about the average. Kutscher states that the phenomenon is
much more marked at 55° C., and that a culture which was not
agglutinated at all by a specific serum at 37° C. was clumped in
twenty-four hours at 55° C. in dilutions of i : 500, or even i : 1,000.
If this is correct it may prove important in the clinical diagnosis
of the disease, which at present is based mainly on the characters
of the cerebro-spinal fluid, and especially on the presence of the
coccus. The cytological and chemical examination of the fluid
affords definite evidence of the presence of a meningitis, but the
cocci are not always discoverable, especially in cases of internal
meningitis, or those in which the foramina at the base of the brain
are closed. The opsonic index may be of great value, especially
if a virulent culture is at hand. According to Houston and
Rankin, the positive " reaction " described above is not usually
present before the fifth or sixth day in the epidemic form.

The serum of immunized animals contains a substance which
gives the phenomenon of fixation of complement when com-
bined with meningococci, and this fact is used by Kolle in the
standardization of his therapeutic serum ; but whether this is a
bacteriolysin or bactericidal substance is not definitely known.
A cording to Davis, normal blood-serum is bactericidal to menin-

PRACTICAL APPLICATIONS 373

gococci, and this power is increased in meningitis. He states,
however, that the opsonic power of the blood was not altered
during the disease.

The treatment differs in the acute and chronic stages. In the
very acute cases the only hope appears to be in the use of
a serum, coupled of course with the more ordinary remedial
measures, such as repeated lumbar puncture. This probably acts by
removing the inert cerebro-spinal fluid (which is almost deficient
both in opsonin and in complement), and causing the exudation
of fluid containing a larger amount of protective substances. In
more chronic cases, and especially in posterior basic meningitis,
the use of vaccines, either alone or in conjunction with serum, is
more promising.

Several sera have been prepared — Kolle and Wassermann's,
Jochmann's, Ruppel's, Flexner's,and Burroughs and Wellcome's —
and very diverse reports have been published concerning their
value. Small differences exist in the methods of preparation,
but in all cases large doses of organisms, dead or living, are
injected, either subcutaneously or into the veins. The potency of
the serum is- tested either by agglutination or by the absorption of
complement. According to Houston and Rankin, some of the
therapeutic sera in common use have very slight opsonic
power and are devoid of agglutinating properties. It seems quite
clear that the serum is useless if given hypodermically, and the
observers who have obtained beneficial results (Flexner, Levy,
and others) have injected the serum into the spinal canal after
removing an equal amount of cerebro-spinal fluid by lumbar
puncture. The dose is 20 to 30 c.c., repeated several times at
intervals of twenty-four hours. Some patients (adults) have
received as much as 340 c.c. The injections may cause vomiting
and unconsciousness, but no permanent inconvenience ; they are
painful, and this is attributed to the use of carbolic acid as a
preservative agent. The treatment must be commenced early,
and in view of the results of Levy, in whose practice the deaths fell
from 78-57 per cent, to 6-25 per cent., and Flexner, who reduced
the mortality to 20 per cent., seems to be of decided value. Other
observers have, however, been less successful.

The use of serum from convalescent cases of meningitis (which
they find to possess marked bactericidal properties) has been
suggested by MacKenzie and Martin. The blood is collected by
venesection, whipped, and centrifugalized, and the serum used

374 MALTA FEVER

for intraspinous injection. In some cases the blood used was
taken from the patient himself, and they obtained encouraging
results after the use of both methods. The method is perfectly
rational, and we might expect on a priori grounds that fresh serum,
containing its full amount of complement and thermolabile opsonin,
would be of more benefit than stale immune serum. With the
idea of causing an increased outflow of preventive substances
from the patient's blood by producing a mild aseptic inflammatory
process, Briscoe has injected dilute carbolic acid solutions and
other fluids into the cerebro-spinal canal, but without success.

The treatment by vaccines has not attracted much attention,
but in the subacute and chronic cases it is well worthy of a trial.
I have used it in four cases, with three recoveries and one death.
These figures are not sufficiently great to argue about, but what
was most obvious was the very decided clinical improvement
which followed almost every dose. This occurred too frequently
to be mere coincidences, and as these patients were all young
children the question of their being due to a mental effect need
not be considered. The dose has been 250 millions, increasing to
500 millions, and 1,000 millions, and no bad effects have been
noticed. They have been usually regulated by opsonic control,
and the improvement in the general condition as the index rises
appears to me to be more definite in this disease than in most
others. It need hardly be pointed out that no form of specific
treatment will cure the obliteration of the foramina in the roof of
the fourth ventricle and consequent hydrocephalus, which is so
frequently present in these chronic cases.

In some cases the index shows a very marked rise (to 10 or
more) as a result of a single injection of a homologous vaccine,
whilst in others the reaction is much less, and the level does not
rise much above unity. It is probable that these two types of
reaction correspond to infections with the cerebro-spinal and basic
meningitis types of organism, but the cases show no obvious
clinical difference.

Malta Fever.

The toxin of Malta fever appears to be an endotoxin. Killed
cultures are decidedly toxic for animals, and their prophylactic
use in moderate doses has been followed in some cases by the
development of chronic aseptic abscesses (Eyre, Bousfield).
The type of immunity is not known ; no bactericidal substance is

PRACTICAL APPLICATIONS 375

known to be developed in the blood, and Eyre's attempts to
produce a powerful curative or preventive serum were unsuccessful.
There is more evidence pointing to opsonic action in the cure of
the disease and the subsequent immunity : the index early in the
disease is as a rule low, and it rises during the progress of the
case. In some points the immunity reactions of the disease
resemble tubercle (slow development of the immunity, frequency
of relapses, absence of known bacteriolytic properties in the
serum, local toxicity of the cultures), but there is this marked
difference, that Malta fever is essentially a septicaemia, the cocci
being present in the blood in small or large numbers in practically
all cases.

The agglutinative reaction, first studied by Wright and Smith
is of enormous importance in the diagnosis of the disease, and
it was Zammit's discovery of clumping powers in the blood of
the goat that led to the discovery of the fact that the milk of
these animals was frequently infective even when the animal had
no signs of disease. Remedial measures based on this phenomenon
have been of enormous advantage in checking the disease.

Various criteria are adopted in the application of the reaction
to diagnosis of disease in man. Critien uses a dilution of i : 10,
and a time-limit of half an hour, and finds results thus obtained
as conclusive as if higher dilutions were used. Bassett-Smith
recommends i : 30, with a time-limit of four hours, or twenty-
four if the macroscopic method be used. Eyre, however, points
out that the blood not infrequently clumps in a high dilution, but
not when more concentrated — e.g., giving no reaction between
i : 10 and i : 100, but clumping strongly at i : 200. It
appears, therefore, that a series of dilutions should be made if
negative results are obtained with the test adopted as a standard.
Agglutination may be manifested in very high dilutions, even as
high as i : 500,000. It usually appears about the fifth day.
The only other certain method of diagnosis is the isolation of the
organism from the blood, usually a fairly easy matter.

The treatment of the disease by means of vaccines has been
studied by Reid, and more recently by Bassett-Smith. The
latter recommends a ten-day-old agar culture, emulsified in dis-
tilled water, sterilized at 60° C. for an hour, and sufficient carbolic
acid added to bring the strength to 0-5 per cent. It is standardized
by drying 20 c.c. (of course, before the addition of the carbolic
acid), and weighing the solid residue. The results thus obtained

376 TUBERCULOSIS

are used to determine the degree of dilution necessary, so that
20 c.c. of the vaccine should contain 4 milligrammes of solid
substance. The doses given were between 0*25 and 0*5 c.c. They
were checked by opsonic control twice weekly, and were in
general given once a week or fortnight. Bassett-Smith does not
approve of the method in acute cases, but his chronic ones
seemed decidedly benefited.

The prophylactic use of the vaccine has not been tried on a
scale sufficiently large to enable any very definite inferences to
be drawn therefrom. The doses may be 200 to 400 million
cocci, and in general two are given, the results being judged by
their effect on the agglutination reaction. As noted above, some
bad local results have been noticed, and there is in all cases a
good deal of fever, malaise, and inflammation at the site of injection
and corresponding lymph glands.

Tubercle.

The actual toxin and mode of action of the tubercle bacillus
remain unknown. The most interesting substance which has
been derived from cultures of the organism is tuberculin, which
has been discussed previously. That it is not the true toxin
appears from the fact that animals immunized to it are still
susceptible to the tubercle bacillus, and a serum can be prepared
which appears to be an antitoxin for tuberculin, but it has no
curative or preventive effects in tubercle. The fatty substances
which confer on the bacillus its peculiar staining properties are
not devoid of toxicity. They cause chronic inflammatory and
caseous changes in the tissues, and may perhaps play a larger
role in the evolution of the lesion than is generally thought. And
it must be pointed out that tuberculosis may be an afebrile
disease throughout. The temperature of phthisis is mainly due to
other organisms, notably to the pyogenic organisms which form
such frequent secondary contaminations of the vomicae. The
chief pure tuberculous affection in which fever is a marked and
constant symptom is tubercle of the meninges, and here the local
processes are in close proximity to the cerebral cortex. In general
tuberculosis there is usually fever, but there is also usually a
focus exposed to secondary infections, or a broncho-pneumonia in
which other organisms play a part. The main toxic effect which
we recognize as being due to the tubercle bacillus is exerted on
the surrounding tissues, and there is no disease which is more

PRACTICAL APPLICATIONS 377

truly a local one ; meaning thereby that in pure tuberculosis the
general symptoms indicative of a spread of the toxins into the
general circulation are but slight, and the local lesion constitutes
almost the whole of the disease. Note, for instance, the good
general health which often occurs in patients with tuberculous
glands in the neck, even if of some size. And where the general
health is impaired, it will often be found to be a predisposing
cause rather than a result of the tuberculous process.

Immunity. — These facts would tend to show, what I believe to
be the case, that the immunity to tubercle is local rather than
general. That general immunity exists there can be no doubt,
but we must distinguish between general immunity due to a
condition of the blood and that due to a resisting power inherent
in all the tissues of the body. Thus to take two patients, one
strong and robust and the other of enfeebled vitality : the latter
may fall a ready victim to tubercle, whilst the former resists
exposure to the most virulent infection : yet there may be no
demonstrable difference in the serum of the two persons. There
may, it is true, be a slight difference in the opsonic index, but this
is not invariably the case.

The importance of local immunity in tubercle appears clearly
from the fact that areas of advance and of cure may occur in the
same patient, or even in different parts of the same lesion. This
is often very well seen in cases of lupus, but is even more
marked in sections from chronic tuberculosis — e.g., of a synovial
membrane, in which areas of cure (as indicated by fibrosis and
organization of the giant-cell systems) and of extension (as in-
dicated by caseation and the formation of new tubercles) may
almost always be found in the same lesion. In general, tubercle
tends to spontaneous cure, and when a lesion continues to spread
it will often be found that some second influence (such as a
secondary infection) is at work ; even in debilitated subjects there
is usually an effort at spontaneous cure in some parts of the
lesion.

As regards the nature of this process of cure, we can say but
little. There can be but little doubt that the main curative
agency is phagocytosis, but there is no reason to think that this
is dependent on the same mechanism of opsonic action that we
reproduce so readily in vitro. Polynuclear leucocytes are con-
spicuous by their absence from tubercles, and the only cells in
which tubercle bacilli have actually been demonstrated in the

37^ TUBERCULOSIS

living body are the giant and endothelial cells. In these, as
Metchnikoff pointed out many years ago, tubercle bacilli can be
seen in all stages, of degeneration, from bacilli possessed of normal
staining properties to mere " ghosts." Either the endothelial
and giant cells actually engulf the bacilli, or, what is perhaps more
likely, they grow round them, being stimulated to growth and
proliferation by the action of the same toxin which, when more
concentrated, leads to caseation and death. Whether any
opsonin is necessary for this process to occur, or whether this
opsonin is produced locally or makes its way from the blood, we
do not know. What we do know, however, is that a high opsonic
index does not necessarily imply that the local lesions are under-
going cure ; chronic tubercle, and especially lupus of some stand-
ing, is often associated with a very high index. But Bulloch
found that the cases of lupus which did well on X-ray treatment
(which, amongst other effects, causes a permeation of the lesions
with plasma) were those that had a high index, and that in the
cases in which it was low originally better results could be
obtained if it were raised by means of tuberculin injections. It
is, therefore, not improbable that opsonin may actually soak
through the lymphoid zone and sensitize the bacilli, thus aiding
phagocytosis by the giant cells. But this is by no means certain.
The results of an injection of tuberculin are probably very
complex, and it is at least possible that the curative effect is
mainly or entirely due to the production of a local reaction in the
neighbourhood of the lesion, as a result of which it is flushed
with blood, and perhaps simply increased as regards nutrition.
Very small doses of tuberculin are sufficient to cause a slight but
definite reaction. This may be quite inappreciable in most
regions, but perfectly obvious in the case of tubercles of the iris,
which may be seen to become surrounded by a hyperaemic zone
after injections of T^-^ milligramme of new .tuberculin, or even
less. Reactions of this nature are, of course, entirely harmless,
and are unaccompanied byu!he slightest rise of temperature, if any.
The only role which we can assign with any degree of probability
to the polynuclear leucocyte in the struggle against tubercle has
regard to its action on bacilli which have made their way into the
blood. Here the conditions are much more like those which
occur in our opsonin experiments, and it is quite likely that
opsonization and phagocytosis occur. Even this is not certain,
however, for we do not yet know definitely whether opsonin exists

PRACTICAL APPLICATIONS 379

in the circulating blood-plasma, and that the mechanism is of
comparatively little value appears from the fact that rupture of a
caseous gland into a vein is so often (as far as we know, always)
followed by general tuberculosis. It is possible that living and
virulent bacilli may be taken up by the polynuclear leucocytes,
carried to distant organs, such as the lymph glands or spleen, and
there continue to grow, producing anatomical tubercles.

As regards the nature of tuberculo-opsonin, the normal opsonin
is completely thermolabile, and the rise in the index due to
injections or the natural disease is mainly of that nature also.
Occasionally the presence of thermostable opsonin is demonstrable,
but it never becomes abundant.

We know but little concerning any other antibody or defensive
mechanism. No bacteriolytic or bactericidal action is demon-
strable. This may possibly be due to the technical difficulties
incidental to investigations of this nature on a bacillus which
owes its staining properties to the presence of fats (which we
cannot expect a serum to dissolve), and which grows so slowly.
Wassermann and others have shown that antibodies occur in
patients immunized with tuberculin, but we do not know their
nature. Nothing really definite is known concerning an antitoxin.
An agglutinin is often formed, but it may be absent throughout
the whole course of the disease, so that we cannot regard it as of
importance.

Diagnosis. — Where possible this should be made by the recogni-
tion of the bacillus, an achievement which modern methods
have made possible in many cases in which it would formerly
have been regarded as out of the question. Into this and into
questions of cytology, etc., we need not enter. The main methods
to be considered are — (i) the tuberculin reaction, and (2) the
opsonic index.

i. The Tuberculin Reaction. — The old tuberculin is used, and is best
bought ready prepared, as issued with the German Government
stamp. It may be standardized, but this process is uncertain,
animals differing markedly in susceptibility. Several methods
have been introduced, but are not in general use. Behring's
method is to determine the lethal dose for guinea-pigs on sub-
cutaneous injection, whilst Von Lingelsheim injects directly into
the brain, the susceptibility of which is much greater. As a rule,
a normal guinea-pig will stand a dose of i c.c. with impunity. It
should be diluted before use with sterile normal saline solution

380 TUBERCULOSIS

containing 0-5 per cent, carbolic acid or lysol, and kept in sealed
" ampoules," each containing i c.c. A convenient method to
adopt is to take 99 c.c. of diluent and add the whole of a i c.c.
bottle of tuberculin. Each cubic centimetre of this contains 10 milli-
grammes of fluid, or yj^ c.c. ; several ampoules are prepared, and the
remainder of the fluid diluted with an equal amount of the diluent.
Each cubic centimetre now contains 5 milligrammes, or -^^ c.c., and
several ampoules are filled with this dilution. Lastly, what remains,
or part of it, is diluted with four times its volume of diluent, so that
each cubic centimetre contains i milligramme, or y^^- c.c., of tuber-
culin. It will be noted that the fraction of a milligramme refers to the
tuberculin as sold, and not to any active constituent it is supposed to
contain. It is recommended to err on the side of caution, and to com-
mence with y^1^ c.c., and then to go on to ^J^ and finally y^-; but
with proper selection of cases it is probable that the risk with ^ J^ is
infinitesimal. The rise in temperature which constitutes a reaction
should be at least i° F., and is usually more. It usually com-
mences in eight to twelve hours, rises for another two hours,
remains up for a few hours longer, and then falls rapidly. There
is often a considerable amount of general malaise.

It should be used only in cases in which the diagnosis cannot
be made by other methods with ease and certainty, and in which
it is of great importance that it should be made definitely and
rapidly. Thus, a patient living under favourable conditions who
developed ambiguous indications of phthisis might fairly be
watched for a time, his weight recorded, an attempt made to
obtain sputum, etc. But the same signs occurring in a person
who was about to get married, or a medical man thinking of
taking a resident appointment in a hospital, would be a strong
indication for the diagnostic use of tuberculin. Perhaps its main
value is that it enables a negative diagnosis to be made with some
degree of confidence, and is the only agent which will do so. No
reaction after two injections of yj^ c.c. may be taken as definite
proof that the disease is absent.

It should not be used (i) when the temperature is irregular —
obviously, since the rise might not be due to the tuberculin — and
(2) when there are secondary infections.

Some cases of syphilis, leprosy, and actinomycosis are said to
react, but this is unusual, and the concomitant presence of tubercle
has not always been ruled out.

In cattle the doses are larger. The tuberculin is usually diluted

PRACTICAL APPLICATIONS 381

ten times, and the dose varies from 4 c.c. for a large bull to i c.c.
for a calf under one year. The temperature is taken for a day or
two before the injection, and at the ninth, twelfth, fifteenth, and
eighteenth hours afterwards. The rise is gradual, reaching its
maximum in twelve to fifteen hours, and then falling gradually to
normal. According to Nocard, anything over 2*5° F. is diagnostic,
from i -4° to 2-5° F. suspicious, and under i'4° F. unimportant. The
temperature of the animal should not exceed 103° F. when the
injection is made.

Von Pirquefs reaction or the cutireaction : This was introduced
as a method of avoiding the dangers supposed to be incidental to
the use of tuberculin sub cute. A very gentle scarification of the
skin is made, just as for Jennerian vaccination, avoiding drawing
blood, if possible, and the abraded surface covered with diluted
tuberculin (i part with 3 of normal saline containing 0-25 per cent,
carbolic acid). A control scarification is made, preferably on the
other arm, and covered with the carbolized normal saline solution.
The reaction takes the form of a red papule of varying size,
sometimes extending for ^ inch or more in all directions from the
site of the original scarification. I have seen the skin so sensitive
that a vivid reaction was obtained where the diluted tuberculin
had accidentally run down the skin. The redness increases for a
day or two, and may remain for four or five days, but in general
disappears earlier. It is to be compared with the control side,
which has also been treated with dilute carbolic acid, but no
tuberculin. Von Pirquet now uses undiluted tuberculin, the
control scarification not being treated at all. This is probably
the better method, and seems devoid of danger. In Moro's
method an ointment containing tuberculin is rubbed into the skin.

The process is probably quite devoid of danger. Some tuber-
culin is certainly absorbed, since I have seen a quite definite local
reaction round a tubercle in the iris; the temperature, however,
does not usually rise, but may do so ; this is a sign that the
scarification has been too deep. No reaction is given in advanced
cases of the disease, probably because the resisting powers are
too low for the production of the necessary antibodies. The
same, it may be noted, is true for the ordinary tuberculin test,
if it were ever justifiable to use it in such cases.

Of the value of the test there is no doubt, and a well-marked
reaction is conclusive. A good deal of difficulty arises from
slight and doubtful reactions, and it sometimes appears to fail in

382 TUBERCULOSIS

cases definitely tuberculous. A positive reaction is here of more
value than a negative one.

Calmette's reaction was introduced soon after Von Pirquet's, and
consists in a slight conjunctivitis and oedema of the caruncle, with
sero-fibrinous discharge, which occurs in tuberculous subjects after
the instillation of a drop of well-diluted tuberculin into the eye.
Since glycerin has an irritant action, a special preparation devoid
of this substance is sold for the test. It is put up in a dry form
(being precipitated by alcohol), and when dissolved in water
according to the instructions gives a dilution corresponding to
i : 100 or i : 200 of the original tuberculin. The strength has
been gradually decreased owing to calamitous results having been
obtained when the stronger dilutions were used. It should
certainly never be used when the eyes are affected, and even
when they are healthy I personally do not consider the risk worth
running unless the diagnosis is of great importance, and all other
methods have been tried and have failed.

2. The Opsonic Index. — Numerous observations by different
observers have shown that in health the opsonic index to tubercle
lies between 0-8 and 1*2, the exceptions being very rare and due
perhaps to slips in the technique. As a matter of fact, these are
outside figures, and in the great majority of cases the index will
be found to lie between 0*95 and 1*05. If, therefore, the diagnosis
lies between health and tubercle, an index below 0-8 or above 1-2
is very strong evidence in favour of the latter, becoming more
convincing the remoter it is from these figures, either above or
below. Indices between 0*8 and 1*2 are not of much value, yet a
figure of, say, 0-85 on several occasions is very suggestive.

There is not sufficient evidence as yet to show how other
diseases influence the tuberculo-opsonic index. In the majority
of cases the figures lie within the normal limits, but there are
occasional exceptions. In general, therefore, the results given
above will apply, but the conclusions are less certain.

A rapid variation of the index, either (a) spontaneous, or
apparently so, or (b) due to auto-inoculation from exercise or
massage of the lesion, or (c] caused by an injection of a small
dose of tuberculin, is much more suggestive — indeed, practically
diagnostic, supposing no errors are made in the determination.
And it must be emphasized that when the opsonic index is
required for diagnostic purposes, the most careful technique and
attention to every detail is absolutely necessary.

PRACTICAL APPLICATIONS 383

The agglutination reaction has been recommended as a means of
diagnosis, but it is not always present at any stage of the disease,
and there are practical difficulties in the way of its determination.
It does not appear to have come into general use even in France,
where much attention has been paid to it.

Treatment. — The essays at a serum treatment have been many,
and no method has attained any appreciable degree of success.
Good results have been obtained, it is true, by several, especially
in the hands of their inventors, but recent work by Hort and
others has shown that normal horse serum has some value as a
curative agent. Maragliano's serum appears to be prepared by
injecting animals with various extracts of tubercle bacilli, and
is supposed to be both bactericidal and antitoxic. It is given
either alone or in conjunction with a vaccine. McFarland
prepared an antituberculin by immunizing donkeys for long
periods with tuberculin, and found that it annulled the effects of
tuberculin on tuberculous animals, but had no protective or
curative powers as tested on guinea-pigs. Clinical evidence
appeared to show that it was of some value.

Marmorek's toxin is quite different from any of the others,
which are various extracts of tubercle bacilli, or of their soluble
products. It is prepared by cultivating young bacilli (prepared
in such a way that they form homogeneous emulsions) in a
medium containing leucotoxic serum, prepared by injecting calves
with guinea-pig leucocytes. It is supposed to be the actual
toxin which is developed in vivo. This substance (which is of
feeble toxicity) is used to immunize horses. Some good results
have been obtained, but perhaps we may attribute them to minute
amounts of tuberculin which may be formed in the cultures and
remain unabsorbed in the blood of the horse when it is bled. The
experience of most observers has not been favourable to its use.

The other sera which have been introduced do not call for
further notice.

The most hopeful method of combating tubercle — other than
the all-important use of fresh air, good food, and careful regimen —
consists in the use of a vaccine. There are many to choose from
— old tuberculin, TR, BE, bovine tuberculin, oxytuberculin,
antiphthisin, etc.; but of these, Koch's preparations (old tuber-
culin, TR, and BE) are the only ones in general use, and appear
at least as good as any.

TR (tuber culinum residmtm) is prepared by triturating living, dry,

384 TUBERCULOSIS

virulent tubercle bacilli by a mechanical process in an agate
mortar. The powdered bacilli are suspended in distilled water
and centrifugalized. The supernatant fluid has the properties of
the old tuberculin, and is called TO (O = obey, = upper). The
residue is dried, and again powdered, emulsified, and centri-
fugalized, and the bacilli are now found to have been reduced to
a uniform mass. This constitutes TR, which thus corresponds
somewhat to an endotoxin. It is diluted with 20 per cent,
glycerin, and when diluted it is advised that no carbolic acid
should be added. There has been a good deal of confusion with
regard to the dosage. It arose from the fact that i gramme of the
dried tubercle bacilli is used in the preparation of 100 c.c. of the
remedy, so that each i c.c. of the latter contains the material
derived from 10 milligrammes of bacilli, and is supposed to be
equal in immunizing power thereto. But the amount of dry
residue which it contains is only one-fifth of this amount, the fluid
being standardized before issue, so that i c.c. contains 2 milli-
grammes of solid substance. The dosage adopted in this book
concerns this dry residue only.

Tuberculins are also prepared from bovine tubercle (perlsucht)
bacilli, and its use mixed with the human form (as recommended
by Allen) appears rational and quite worthy of a trial. The
majority of cases of human tubercle are due to bacilli of the
human type, but at the worst the material derived from the bovine
bacilli will be inert and do no harm.

Another substance used as a vaccine is Koch's Bazillen-
emulsion, or Neutuberculin, usually known as BE. It consists
of a suspension of dried and ground up bacilli in equal parts of
glycerin and water. When used as a vaccine it is diluted with
normal saline solution and heated to 60° C. to insure sterility.
This may also be done with TR. It contains 5 milligrammes
solid substance per c.c.

As regards the use of these substances : We may recognize
two, or perhaps three, methods — (i) the intensive ; (2) the opsonic,
and (3) the uncontrolled use of small doses.

i. In the intensive method the idea is to bring about as
much immunity as possible as quickly as possible, by the use of
rapidly increasing doses, taking care always to avoid a reaction.
The commencing dose is very small, about T\j- milligramme
( = TTrijoo c-c-) °f °ld tuberculin, or T-^^ milligramme (of solid
substance) of TR. The dose is repeated two or three times a

PRACTICAL APPLICATIONS 385

week, and is gradually increased until 500 milligrammes ( = J c.c.)
of old or 2 milligrammes of TR is given. If a reaction occurs
an interval of a week is given, and the treatment recommenced
with a smaller dose. The dose of BE may be the same as that
of TR.

There are numerous slight modifications in detail, but the fore-
going outline will serve as a general description of the process as
it is used in many of the Continental clinics. As a rule no
attempt is made to estimate the degree of immunity, but Koch
suggests the agglutination reaction for this purpose. It usually
rises to i : 100, or thereabouts, and may go much higher.

As a result of the treatment, the patient is certainly immunized
to tuberculin, since he fails to react to fairly large doses. The
process is then stopped for a time and recommenced. The
beneficial effects as recorded (as I have seen on a small scale) are
undoubted, and the risks in careful hands and in properly selected
cases appear to be slight.

2. The opsonic method consists in the use of such doses at such
intervals as will cause the maximum increase in the opsonic
index. The doses here are very much smaller than in the last
method. TR is usually employed, and the amounts range
between yo^Vc^ anc^ IWQV milligramme. The idea, of course, is
to avoid the possibility of a summation of negative phases, and
to call forth the greatest possible formation of protective sub-
stances. It is extremely difficult to carry out, since the index
alters so quickly in most cases of progressive tubercle that very
frequent determinations of the index are necessary. Perhaps it
is the ideal method, and all who undertake tuberculin treatment
on a large scale ought to acquire some experience of it, but con-
siderations of time will prevent its ever coming into general use
as a routine treatment. In practice a sort of modified opsonic
method is sometimes used. The first few doses are controlled by
the index, and the optimum dose and interval determined. These
are then preserved throughout the treatment, with perhaps
occasional determinations of the index at times.

It must be pointed out that the theoretical necessity for the
opsonic control depends on the acceptance of the fact that
tuberculin exerts its beneficial influence by causing an increase
in the amount of opsonins present in the blood. But this is not
certain. Admitting that the immunity is due to antibodies, these
may be bactericidal substances or anti-endotoxins, and we have

386 TUBERCULOSIS

no reason whatever for thinking that the opsonic index affords
any clue to the amounts of these substances formed. And, of
course, as is held by many, if the beneficial effects are due to
repeated slight local reactions, acting in a way as yet not under-
stood, the opsonic index is of little value. It should not be
forgotten that a polynuclear leucocyte will engulf an enormous
number of tubercle bacilli in a short time even when the opsonic
index is low, and it seems improbable that it ever falls so low
(even when the negative phase is most marked) as to prevent
these leucocytes from doing their work, presuming the bacilli are
accessible. A high opsonic index may have some action in
preventing generalization of the bacilli— e.g., at a surgical opera-
tion— but even this is not certain.

It must be noted that the beneficial effects of vaccine treatment
by the opsonic control are not denied, but that the opinion is
expressed that the same benefit^ may be obtained by simpler
methods.

Space forbids more than a reference to the use of auto- inocula-
tion by graduated exercise, massage, etc. This is supposed to
expel some bacilli or products thereof into the lymph spaces, etc.,
where they act as a vaccine. The benefits of these measures is
undeniable, but it may be doubted whether it is due entirely or
in part to auto-inoculation, and it would appear preferable to
employ vaccines in known doses.

3. The use of small doses without opsonic or other control. This
combines the advantages of practicability and safety, and often
gives good results. How they compare with those obtained
by the other methods cannot be accurately known. Probably
no harm has ever resulted from TT^Ty milligramme of TR, and if
this amount is given once in every ten days, or even once a week,
many cases improve in a most striking manner. My experience
has been mainly confined to its use in surgical cases, and though
I have met with numerous disappointments (which I believe are
not unknown with the most strict opsonic control), I have seen
others in which rapid cure — e.g., of a chronic sinus or tuberculous
ulcer — has been obtained. I should not recommend it or any
other method of vaccine treatment in place of surgical or
sanatorium treatment, where applicable ; in conjunction with
other methods, it has its place.

Preventive Treatment. — This is at present of but little importance
in human pathology. A person who is in danger of becoming

PRACTICAL APPLICATIONS 387

tuberculous should be removed from the infection and placed
under good hygienic conditions. A few injections of TR may be
given, however, and analogy with a similar procedure in cattle
would lead us to believe that it may be of some value.

Attempts to immunize lower animals to tubercle have been
made from a very early period ; it was, indeed, the discovery of
the fact that dead bacilli are absorbed with great difficulty that
led Koch to devise his TR. The vaccines that have been employed
are numerous — cultures of low virulence, dead cultures, avian
and reptilian bacilli, etc. — but the subject entered oh a new and
most important phase in 1901, when von Behring published his
method, which is of great theoretical interest; and from which
important practical results in the stamping out of bovine tubercle
(so great a pest to infants, from the prevalence of bacilli in milk)
have been esteemed possible. Here the vaccine consists of living
tubercle bacilli of human origin. This is of feeble virulence for
cattle. It is prepared by being dried in vacuo, emulsified with
glycerin in a mortar, and diluted with normal saline solution
containing 0*15 per cent, of sodium carbonate. It is standardized
in such a way that i c.c. = 2 milligrammes of dried bacilli. The
injections are made intravenously; two are given — the first of
i c.c., and the second, about twelve weeks later, of five times this
amount.

There is no doubt that this process confers a certain amount
of immunity, or rather of increased resistance. The process is
harmless to calves, but sometimes kills older animals from
pulmonary oedema. The immunity lasts for a year or so.

The method has been carefully examined at Melun by Rossignol
and Vallee, and their results were quite favourable. The
vaccinated animals were tested (along with controls) by sub-
cutaneous injection of bovine tubercle, by injections of these
bacilli into the veins, and by keeping them in a shed with animals
known to be tuberculous, and in all cases a greatly increased
degree of resistance was found. But Lignieres pointed out a
remarkable fact, that animals which have been treated in this
way and which do not react to tuberculin, and are apparently
normal at the autopsy, nevertheless may contain living tubercle
bacilli in their lymphatic glands. This is of extreme interest in
connection with the question of the latency of bacilli in the tissues.

Von Behring's latest vaccine is known as tuberculase, or tulase.
It appears to be prepared by the action of chloral on tubercle

25—2

388 TYPHOID FEVER

bacilli, but the exact details are not available, and the preparation
is not yet obtainable. It seems not to contain living bacilli.

Attempts have been made to vaccinate against tubercle (and to
treat it in human subjects) by ingestion. Some degree of success
appears to have been obtained by feeding calves with living
human tubercle bacilli, and some physicians have claimed good
results by administering a minute dose of TR by the mouth ; but
in what way these methods, so uncertain as to the dose which is
actually absorbed, are preferable to the use of the svringe is
not very clear.

This is a very brief epitome of a field of research which would
require many volumes for its adequate treatment. I have
attempted to give the main essentials only.

Typhoid Fever.

The pathogenic action of the typhoid bacillus appears to be due
entirely to the action of an endotoxin which is set free when the
bacillus undergoes solution in the body. This toxin has a local
and a general action. The former is marked in the regions which
harbour the bacillus, and in which it undergoes solution, either
by autolysis or by the action of bacteriolysis— the Peyer's patches,
abdominal lymph glands, spleen, etc. In these regions it produces
hyperaemia, followed by inflammatory hyperplasia of the lymphoid
and endothelial cells, which diminish the blood-supply and may
lead to necrobiosis of the inflamed tissues. In the case of the
Peyer's patches this commonly occurs, the lymphoid tissue being
cast off as a slough, and the wound, becoming infected by intestinal
organisms, may extend deep into the bowel wall and cause haemor-
rhage or perforation. The process is less likely to occur in the
solid organs, which are nourished by blood from all sides and in
which secondary infections are less frequent. The general effects
of the toxin — fever, degenerations of the kidney, liver, etc.— are
not characteristic. The failure of a leucocytic reaction of the
bone-marrow, and consequent leucopaenia, is worthy of notice.

The fact that typhoid fever is a septicaemia and that the living
organisms circulate in the blood must not be forgotten.

Immunity. — A short time ago typhoid fever was regarded as a
disease in which the immunity was purely bacteriolytic. During
an attack of the disease or the process of artificial immunization the
amount of immune body (which occurs in small amounts in
health) shows a steady rise, and the blood soon becomes extremely

PRACTICAL APPLICATIONS 389

potent in this respect. The bacteriolytic power of the blood as
tested by Wright's method does not necessarily show a great
increase ; this is due probably to the lack of complement, and it is
possible that relapses may be due to the same cause, deviation of
complement occurring owing to the excess of immune body.

There is, however, no doubt that phagocytosis plays a part of the
utmost importance in the natural cure of the disease, and that the
opsonic content of the blood (as tested by one of the dilution
methods) undergoes a steady rise during the process of immuniza-
tion. We may regard the mechanism by which the infection is
combated as a double one, bacteriolysis and phagocytosis being
jointly concerned. That the latter is of chief importance appears
probable, from the fact that typhoid bacilli set free their endotoxin
when dissolved by means of bacteriolytic sera.

The antitoxin or anti-endotoxin of typhoid bacilli can be readily
prepared from animals by a process (somewhat prolonged) of
immunization with the endotoxin, prepared either by grinding the
bacilli, by dissolving them in bacteriolytic sera, or by allowing
them to undergo aseptic autolysis. There is, however, no proof
for the belief that this substance is developed during the natural
process of cure, or that it is a cause of the subsequent immunity.
It is quite as possible that the cessation of the febrile process
is due to the destruction of the bacilli, and consequent cessation
of the supply of toxin.

The effect of the true typhoid toxin is seen during the earlier
stages of the disease, when the temperature is continuous. The
intermittent temperature of the later stages is probably due in
part to the absorption of other toxins from the ulcerated surfaces
of the Peyer's patches, and in part to the liberation of endotoxins
which occurs when bacilli are dissolved before being ingested.

The duration of the immunity after a natural attack of the
disease is not accurately determined, but it is certainly long —
perhaps several years. That due to preventive inoculation is
thought to be six months at least, but here again exact figures
cannot be obtained. The duration of the antibodies produced in
typhoid fever varies. The agglutinins usually disappear within
one or two years, but they may persist for seven or more. The
bacteriolysin is thought to go sooner, but the observations on
which this idea is based were mostly on the bactericidal powers
of the blood, and open to fallacy.

Diagnosis. — Usually the W7idal reaction is all that is required.

39° TYPHOID FEVER

The standard that is adopted varies in different laboratories, but
in the great majority of cases it will be found that the serum at
the end of the first week will clump in a dilution of i : 30 in one
hour at the body temperature, and that this is very seldom the
case in health or in other diseases. The amount of agglutinin
soon increases, and may reach a degree at which it will agglu-
tinate at i : 1,000 or more. The naked-eye reaction as carried out
by a modification of Wright's method renders it very easy to
perform exact quantitative work, and it is an advantage to do so in
all cases in which the agglutination does not reach i : 50 on the
first examination, since a rise in the power of the serum is
practically diagnostic. My routine method is as follows : A unit
mark is made about i inch from the end of a rather wide
Wright's pipette, and with this 9, 2, 4 and 9 units of a watery
emulsion of a young agar culture of typhoid bacilli are placed in
four depressions on a porcelain slab ; dead cultures may also be
used in the same way. A unit of serum is now sucked into the
pipette and mixed with the first pool of 9 units of emulsion.
One unit of this is then mixed with each of the remaining three
pools, the process being carried out quickly, so as to avoid the
absorption of agglutinin from the powerful serum. This gives a
series of i : 10, i : 30, i : 50, and i : 100. Each pool is now sucked
into a Wright's pipette sealed at the end, and incubated at 37° C.
A control of emulsion without serum is also prepared and in-
cubated. The tubes are examined at the end of one hour, and
agglutination, if present, is most obvious.

The diagnosis might also be made from a determination of the
opsonic index by the dilution method, or of the amount of immune
body, but these processes are much more tedious.

Where a very early diagnosis is required (i.e., before the
appearance of the agglutination reaction) the bacilli may be
sought for in the stools or blood. This was formerly difficult, but
it has been greatly facilitated by the introduction of new methods
involving the use of special culture media, in which the bacilli
grow rapidly and form characteristic colonies. These are beyond
the scope of this work, and a description of them will be found in
Hewlett's " Bacteriology," second edition.

Treatment.— The curative treatment is not satisfactory. The
use of a simple bacteriolytic serum is useless or even dangerous,
for the reasons given earlier. The best hope for the future is
in the use of an anti-endotoxic serum, such as was prepared by

PRACTICAL APPLICATIONS 3QI

Allan Macfadyen and others, and is now put on the market by
Messrs. Burroughs and Wellcome, under the supervision of
Professor Hewlett. Some promising results have already been
obtained by its use, and it seems to be quite harmless. It seems
only to be of advantage when used in the early stages of the
disease, as is naturally the case with any serum which acts by
neutralizing a soluble toxin.

Chantemesse has apparently attained a very high degree of
success by the use of a serum obtained by injecting horses with a
soluble toxin, prepared by cultivating the bacillus in an extract of
spleen digested with pepsin and subsequently neutralized. It is
very doubtful that this is the genuine toxin, since it is feeble in
action and not destroyed at 100° C. The serum thus obtained is
used in very small doses, and appears to act (as pointed out by
Wright) rather as a vaccine than as a means of conferring passive
immunity. The results, however, appear to have been excellent,
but the serum is not yet obtainable commercially.

Typhoid fever has also been treated by the use of small doses
of vaccine, and some promising results obtained ; but the method
is still on its trial.

Preventive Treatment. — The earliest method was that of Wright,
and it has probably not been surpassed. It consists in the use of
vaccines of typhoid bacilli cultivated in broth for twenty-four
hours, killed at 60° C., counted by admixture with red corpuscles,
and preserved by means of lysol or carbolic acid. Two doses are
given, the first being 750 to 1,000 millions, the second twice this
amount, the injection being usually made deep in the subcutaneous
tissue of the flank. The injections are given at intervals of one
or two weeks. Local and general symptoms of some severity
follow : redness and swelling at the site of inoculation, lymphan-
gitis, etc., with fever, headache, and general malaise. These last
but a short time, and no evil consequences have been recorded.
It is necessary, however, for the patient to be prepared to stay
in bed during the day of inoculation and the next day : the
unpleasant effects may be reduced to a minimum by the simul-
taneous administration of chloride of calcium by the month
(15 to 45 grains).

As a result of these injections, protective substances, and espe-
cially agglutinins, make their appearance in the blood, and may
persist for several years. These may be taken as affording some
test as to the degree of immunity conferred, and as to its duration.

392 TYPHOID FEVER

Space does not permit us to discuss at length the evidence in
favour of this method of inoculation. It has been strongly, and
perhaps unfairly, opposed in certain quarters, but a careful and
unprejudiced study of the statistics of the British Army in South
Africa and elsewhere seems to render it quite clear that the process
develops a real though not absolute protection against an attack
of the disease, and is still more valuable in diminishing the
mortality rate amongst those attacked. The analysis of these
figures is by no means easy, but certain isolated cases are in
themselves very convincing. As an example (one out of many)
we may quote the Manchester Regiment, in which, out of
200 men inoculated there were 3 cases, without a death, whilst of
the 517 uninoculated 23 were attacked and 3 died. In general
terms we may say that the mortality from an attack fell from
over 12 per cent, in the uninoculated to below 6 per cent, in the
inoculated, taking the results from the whole of the Transvaal
and Natal. One caution is necessary : the first result of the in-
jection is the production of a negative phase okjncreased suscep-
tibility, and injections should not be practised, if avoidable, when
the patient is already exposed to the infection.

The vaccine is prepared by the Lister Institute, and can be
bought ready for use.

Numerous modifications of the process have been adopted.
Pfeiffer and Kolle prepare their vaccine from agar cultures.
Bassenge and Rimpau do the same, but use very small doses —
TJ to f milligramme. Friedberger and Moreschi give intra-
venous injections of infinitesimally small doses of bacilli killed at
120° C. Wassermann's vaccine is prepared by killing a culture at
60° C., allowing it to undergo autolysis at 37° C. for five days,
filtering, and desiccating in vacua at 35° C. It forms a yellowish -
white powder, of which the dose is 0*0017 gramme, equal to
12 milligrammes of the culture. Numerous other methods might
be enumerated, but they have mostly been investigated on a small
scale only, and Wright's vaccine is of proved efficacy and easy to
prepare.

Antityphoid serum has often been employed, and in all proba-
bility is of value in that it affords a rapid (or practically instan-
taneous) means by which immunity can be produced, and eliminates
the negative phase altogether. The serum is prepared from the
horse, which is injected with gradually increasing doses of dead
bacilli, and subsequently of living ones. It can be obtained of

PRACTICAL APPLICATIONS 393

great potency, as determined by its power of agglutination, which
may be manifested when it is diluted a million times. But the
immunity it confers (judging from laboratory experiments and
analogy with other diseases) is but temporary, and the method
can only be of occasional value.

The plan of injecting the serum and vaccine in combination
suggested by Besredka is more promising. It is carried out as
follows : A twenty-four-hour agar culture of bacilli is mixed with
typhoid serum and incubated at 37° C. for twenty-four hours.
The bacilli are, of course, agglutinated, and the mass is washed
free from serum by repeated centrifugalizations with sterile
normal saline solution. The bacilli are then suspended in this
solution and heated to 60° C. for one hour. This vaccine is said
to produce very rapid immunity, with practically no negative
phase, and the local and general reactions it produces are but
slight. It has not yet been tested on a large scale.

Simultaneous injections of serum and vaccine have also been
employed by Calrpette and others. The results do not appear to
be as satisfactory as those obtained from the use of the vaccine
alone.

Bacillus Coli.

The Bacillus coli, in addition to its haemolysin (colilysin) already
mentioned, produces an endotoxin, on which its pathogenic action
doubtless depends. Its vaccine is moderately toxic, causing
severe local irritation and general febrile reaction if the initial
dose be a large one. Filtrates from broth cultures are practically
devoid of toxicity.

In regard to its immunity reactions B. coli closely resembles
B. typhosus. Bacteriolytic and opsonic substances are developed
during the course of the disease, and in both cases the opsonic
index must be estimated by the dilution method, Wright's original
process giving inaccurate results. There is some difference with
regard to the production of the agglutination reaction, which is
invariably present in typhoid infections, except in those acute
cases of the disease in which the patient dies before it has time to
develop. In infections with B. coli this is not the case, and in
many of the local infections, such as cystitis, pyelitis, etc,, no
appreciable agglutinative properties are developed. When animals
are immunized with massive doses extremely powerful clumping
sera are obtained, but this is not always observed in the treatment
of the human patient with vaccines.

394 BACILLUS COLI

The main difference, pathologically, between the two organisms
is that, whereas the disease due to the typhoid bacillus is usually
a septicaemia, localized infections (typhoid osteitis, etc.) only
occur as sequelae, the diseases due to B. coli are almost in-
variably local inflammations, blood infections, except, perhaps, as
mere terminal phenomena, being rare in the extreme. These
inflammatory lesions are extremely common and important, and
of most diverse nature. The majority are in connection with
the urinary (cystitis, pyelitis, etc.) and alimentary systems
(cholangitis, cholecystitis, appendicitis, etc.). It also causes
acute peritonitis, but usually in association with other bacteria.
It may affect almost any part of the body, causing broncho-
pneumonia, otitis media, endometritis, metritis, and a host of
other diseases.

The only specific treatment of any avail is the use of a vaccine.
A serum has been prepared, but appears to be quite useless.
The vaccine should always be prepared, if possible, from the
patient's own culture, since various strains classified as B. coli
differ slightly the one from the other. In addition to this, there
are very marked differences in virulence, cultures isolated from
the stools being in general less virulent than those derived from
the diseased tissues. In cases in which treatment has to be pro-
longed it is sometimes an advantage to prepare fresh vaccines
from time to time, since the organism may change its type to
accommodate itself to the immune substances produced.

The initial doses should be small (10 to 40 millions), and it will
be rarely found advisable to exceed 250 millions. Too large an
initial dose may be badly tolerated, causing rise of temperature
and a good deal of local reaction, but no permanent bad results
seem to have been observed. On the other hand, it may some-
times be noted that in severe febrile cases the vaccine acts more
like an antitoxin, causing a sudden drop in the temperature and
a rapid amelioration of the symptoms, which may or may not be
associated with a great improvement in the local lesion. This is
shown in the accompanying chart, from a severe case of cystitis
under the care of Mr. Burghard at King's College Hospital. As
far as his general condition was concerned, he was practically well
within ten days of the commencement of the treatment. The
amount of pus in the urine fell to less than an eighth of its
original volume, but had not disappeared entirely when he was
discharged. This has been my usual experience of cystitis due

PRACTICAL APPLICATIONS

395

to B. coli — vaccine treatment cures all the symptoms and reduces
the pus and bacilli present in the urine to a small fraction of its
original amount, but fails to remove them entirely. Other
observers have been more fortunate (Wright, Western, Norton,
and others), but I believe my experience is the general one, and
that complete cures are unusual. There is no doubt, however,
that the treatment is the best available, the disease being
notoriously resistant to simple methods of treatment. If the
injections are made with opsonic control, the dilution method of

FIG. 71. — CHART FROM A SEVERE CASE OF CYSTITIS DUE TO B. COLI.

The continuous line shows the temperature, the dotted line the amount of

pus in the urine.

estimating the index should be used. In default of this, the
injections may be given every eight to twelve days, doubling the
dose until the upper limit is reached.

The treatment of inflammatory lesions of other regions (chole-
cystitis, persistent sinuses, etc.) appears to be more satisfactory,
remarkable cures having been obtained. Butler Harris has
obtained good results in endometritis, cervical catarrh, and
mucous colitis. In the former disease he gives a small dose a
week before and after each menstrual period.

It is in infections due to B. coli that some of the most striking

3Q6 DYSENTERY

successes of vaccine therapy have been achieved. It has been
used in some extremely severe conditions with profound toxaemia,
and in no case did it render the patient's condition worse. Small
doses should, of course, be used in such cases, but the severity of
the disease need be no bar to the use of the remedy.

Dysentery.

Bacillary dysentery, under which heading we may include some
at least of the cases known as ulcerative colitis, asylum dysentery
and infantile diarrhoea, is in its general pathology closely akin to
typhoid fever. Its toxin is an endotoxin which is only set free
when the bacilli undergo solution, and its cure is associated with,
and perhaps due to, the development of a large amount of bacterio-
lysin. The main difference between the two diseases is that,
whereas typhoid fever is a self-limited disease, in which the patient
either dies or immunizes himself in a fairly constant period after
infection, in dysentery the disease has a tendency to become
chronic, the degree of immunity produced being sufficient to slow
the course of the disease, but insufficient to arrest it altogether.
The bacilli are in the main limited to the intestinal lesions, but
the fact that they have been found in the heart-blood of a new-
born foetus whose mother was suffering from the disease leads us
to believe that they may circulate in the blood ; in most cases
they are probably quickly destroyed, for normal serum has some
bacteriolytic action.

The toxin is contained in the bodies of the organisms, which
are much more poisonous than those of typhoid bacilli. It may
be obtained by Macfadyen's method of grinding at the temperature
of liquid air, by aseptic autolysis at 37° C., or simply by filtering
old broth cultures in which some of the bacilli have undergone
solution. Conradi's toxin, prepared by submitting emulsions to
autolysis for forty-eight hours, was fatal to rabbits in doses of
i c.c., causing diarrhoea, subnormal temperature, etc. According
to Todd, the toxin is developed best in alkaline broth, and attains
its maximum in about six weeks, after which time it begins to
diminish in potency. It is more thermostable than the exotoxins,
resisting heating to 70° C. for one hour, and only being destroyed
slowly at the boiling-point. It is possible to obtain an anti-endo-
toxin, but the treatment of the animals has to be conducted with
great care. It is probable also that the serum of animals
immunized to the bacilli themselves contains some antitoxic

PRACTICAL APPLICATIONS 397

properties, especially if the bacilli are injected into the veins
(Besredka). In preparing this serum it is a great advantage to
follow Besredka's process, and make use of sensitized bacilli. If
the unaltered bacilli are used the process presents some difficulties,
especially in the case of small animals.

The serum thus prepared possesses powerful agglutinating and
bacteriolytic properties, and is extremely powerful as a prophylactic
agent. Kruse's serum protected a guinea-pig against a lethal
dose of the bacilli in doses of ginn^ gramme> and ToiiRr c-c- °f
Shiga's serum (activated with a suitable amount of complement)
would sterilize i c.c. of a twenty-four-hour broth culture of the
bacilli.

Antidysentery serum has now fully proved its value in the
treatment of acute dysentery. According to Shiga, it reduces the
mortality of the disease by nearly 50 per cent. Kruse claims that
his serum causes a rapid diminution in the number of the stools,
such as is effected by no other agent, a general improvement in
the patient's condition, a shortening of the convalescence, and a
diminution of the mortality. Large doses, frequently repeated,
are required.

It has probably a complex action, being at the same time
antitoxic and bacteriolytic, and contains opsonins and bacterio-
precipitins.

In chronic cases it is of much less value, and in these forms
reliance must be placed on vaccine therapy. This has been care-
fully studied by Captain Forster and others. In Forster's patients
the mortality fell from 6-3 per cent, to 0-9 per cent., several cases
of the extremely chronic type which defies all ordinary treatment
for years being completely cured. He uses no opsonic control,
and standardizes his vaccines by determining the minimal lethal
dose ; this is necessary, since the various strains differ greatly in
toxicity. If the minimal lethal dose for a rabbit is about 0*4 c.c.,
doses of o'i, 0-2, 0-3, and 0-4 c.c. are given at intervals of about ten
days. If symptoms of an overdose are produced, the amount given
at the next injection is reduced.

Prophylactic treatment by injections of killed cultures, either
as they are or after autolysis or sensitization with an immune
serum, or injected in conjunction with immune serum, have been
used on a large scale by Shiga and others, with apparent good
results as far as the case mortality is concerned, though with less
obvious effect on the prevalence of the disease. This phenomenon

3Q8 CHOLERA

is readily intelligible if we regard the subsequent immunity as
being a condition in which the body, having been once trained to
do so, readily manufactures antibodies and other protective sub-
stances when infection occurs. Since no protective substances
are present at the time, infection occurs as in a normal person,
but the defensive substances are very quickly produced. Shiga's
vaccine was prepared by emulsifying a twenty-four-hour agar
culture of the bacillus in 5 c.c. of normal saline, heating to 60° C.
for one hour, and submitting the dead bacilli to autolysis at 37° C.
for two days. It is then filtered and used in doses of 0^05 to 0-5 c.c.
The serum becomes strongly agglutinating and bactericidal.
Other methods have been proposed.

Immunity reactions, especially that of agglutination, are of
great value in the diagnosis of dysentery, and especially of the
type of the bacillus present. In acute cases this is hardly
necessary, since modern methods have rendered the task of
isolating the bacilli from the stools an easy one. In the chronic
forms this is extremely difficult, and recourse must be had to the
agglutination test. The reaction is not as strong as in typhoid
fever, a positive result at a dilution of i : 50 being diagnostic.
The blood should be tested against any strains of dysentery
bacilli which may be available, especially if vaccine treatment is
to be used.

The method of absorption of complement has also been used.

Cholera.

In cholera the living organisms are strictly limited to the
intestinal contents, and the disease appears to be a pure intoxica-
tion, without access of living bacteria to the tissues. It is, how-
ever, probable that this is not the case, and that the vibrios enter
the blood and there suffer rapid and complete bacteriolysis, their
endotoxins being liberated in the process. But there is nothing
that can be called a local lesion, and the disease is not a septi-
caemia in the ordinary sense of the word.

The toxin of cholera is a typical endotoxin. The nitrates from
broth cultures are of very feeble toxicity, though they possess
immunizing properties, due doubtless to some degree of autolysis
which has taken place, and to the presence of free receptors.
Bacilli killed with agents such as chloroform or thymol are highly
toxic, especially if injected along with an immune serum, so that
they can be rapidly dissolved. The endotoxin can be prepared

PRACTICAL APPLICATIONS 399

by aseptic autolysis, or by the freezing and grinding method of
Macfadyen. In either case it is thermolabile, being largely
destroyed at 60° C., so that cultures from which the toxin is to be
prepared must not be killed by heat. Metchnikoff and others
claim to have produced a soluble exotoxin by the use of very
virulent cultures in broth : it is thermostable and not very
potent. An antitoxin to it was prepared, but only very low grades
of potency were obtainable. Macfadyen's toxin was much more
toxic, and an anti-endotoxin of high potency was obtainable.
There is no demonstrable antitoxin in the ordinary bacteriolytic
serum obtained by the immunization of animals to the bodies of
the bacilli, or in the serum of cholera convalescents.

Cholera presents the best example of an apparently pure
bacteriolytic immunity, and presents a good example of the
difficulties inherent in the explanation of this subject. The
serum of an immunized animal or that of a person who has
recently recovered from cholera is powerfully bacteriolytic, giving
Pfeiffer's phenomenon in its earliest discovered and most marked
form : it is almost the only organism which is completely dissolved
in vitro under suitable conditions. Such a serum is also strongly
protective, shielding animals against several times the lethal dose
of living vibrios, and it seems difficult to avoid the conclusion
that its preventive properties are due to its bacteriolytic action.
But this is very difficult to maintain in view of the fact that the
serum increases the toxic effect of dead vibrios (and under some
circumstances of living ones), owing to the liberation of endotoxin.
It seems rather as if the presence of bacteriolytic substances is
actually harmful to the animal, allowing the organisms to set free
their toxin, instead of being taken up by the phagocytes and
remaining harmless. I am not aware that any opsonic experi-
ments by the dilution method (which alone would be of value)
have been carried out.

Diagnosis. — In dealing with cases of supposed sporadic cholera
the main problem is the recognition of the vibrio isolated from
the stools, usually an easy matter. The morphological and
cultural characters will of course afford great help, but they take
some time to work out, and more reliance is to be placed on the
immunity tests, which are quicker and more conclusive. The
agglutination reaction is most convenient, and can be carried out
on the dejecta themselves, if the suspected organisms are present
in large numbers. Some of the mucus is broken up in a little

400 CHOLERA

peptone solution, and two hanging-drop preparations are made,
one with the addition of normal serum in i : 50 dilution, the
other with a i : 500 dilution of a powerful anticholera serum,
such as can be obtained commercially. Cholera vibrios become
paralyzed and agglutinated in the second specimen, not in the
first. When smaller numbers are present a culture (probably
impure, but with the vibrios in sufficient abundance to serve for
the test) may be made by incubating peptone-water inoculated
with a flake of mucus for eight to twelve hours. This is to be
tested with the serum in the ordinary way, and should agglutinate
at nearly the same dilution as a known cholera culture. The
serum should be a powerful one, clumping at i : 10,000 or
more.

The Pfeiffer's reaction is perhaps more conclusive, and is
carried out as follows : The test immune-serum is diluted with
broth or normal saline, so that i c.c. contains o-ooi c.c. of serum;
i c.c. of this fluid is used to emulsify a loopful of a young agar
culture of the suspected organism, and the emulsion injected
intraperitoneally into a young guinea-pig. After a few minutes a
little peritoneal fluid is withdrawn by means of a capillary tube,
and the vibrios will be seen to have become non-motile, and to be
undergoing the characteristic change into slightly refractile
rounded masses. After a short time more they will be found
to have disappeared altogether. A control experiment with
normal serum may be made. This test is of great value, many
closely allied organisms failing to react. But no test is absolutely
conclusive, since a few cultures (notably the El Tor vibrio) have
been found to give all or most of them, and yet have been isolated
in a region in which cholera is not known to occur. The subject
is not yet settled, but in the meantime the probability that any
organism which reacts positively to the agglutination and
Pfeiffer's tests is true cholera is enormous.

The serum of persons convalescent from cholera agglutinates
the vibrios at dilutions of i : 100 or more for some months after
the attack, a fact which may be of some value in determining
the nature of a previous disease and a possible immunity to
cholera.

As regards treatment, the ordinary bacteriolytic serum is quite
useless, and, as far as I am aware, no potent anti-endotoxic serum
has been tried. The prophylactic treatment is on a sounder
footing. It was introduced by Ferran, of Barcelona, as early as

PRACTICAL APPLICATIONS 40!

1884, very soon after the discovery of the V. cholera by Koch.
His results were of doubtful value, his vaccines being made of
cultures of feeble virulence, and perhaps impure. The method
was placed on a scientific basis by Haffkine, who showed the
necessity for the use of cultures of great virulence. These are
prepared by passage through guinea-pigs. A more than lethal
dose of a laboratory culture is injected into the peritoneum of a
guinea-pig, and the peritoneal fluid (rich in vibrios) is collected
after death. This fluid is incubated in a thin layer, so as to allow
of thorough aeration, for fifteen hours, and is then administered
intraperitoneally into a second animal. After about twenty or
thirty passages the culture will have attained its maximum
virulence, the lethal dose being some ^ of the original. Its
potency falls off in some ten days, and a few further passages are
required to restore it.

The treatment is commenced by a dose of attenuated virus.
This is prepared by cultivating an ordinary laboratory stock in
broth at 39° C. in conditions of complete aeration. An inocula-
tion on agar is made every day, until (after a few days) the fluid
is found to be sterile. The process is now recommenced, using
the last agar culture that grew, and after several generations a
culture of very feeble virulence is obtained. It causes oedema,
but no necrosis, when injected under the skin. The vaccines are
prepared by cultivating the organisms on agar slants of definite
size (10 centimetres long) for twenty-four hours, and emulsify-
ing with 8 c.c. of broth, or 6 c.c. of 0-5 per cent, solution of
carbolic acid. The dose is i c.c. One or two injections of the
attenuated virus, followed by one of the exalted, all at intervals
of three to five days, may be given, or the exalted virus only may
be used. The injections cause moderate fever, headache, and
general malaise, and local tenderness, swelling and enlargement of
the corresponding lymph glands, all of which pass off in a few days.

This method (with various slight modifications with regard to
dosage) has now been used on a very large scale in India, with
strikingly good results. The immunity lasts for at least a year,
and probably decidedly longer if large doses of strong vaccines
are used, and, what is somewhat unusual, it manifests itself more
in a reduction of the incidence of the disease than in the
case- mortality. The value of the method is best seen from
statistics from isolated regions in which some persons were
vaccinated and others not, all living under the same conditions.

26

4O2 PLAGUE

Thus, in the tea-plantations at Catchar, in 6,549 persons who were
not vaccinated, there were 198 cases, with 124 deaths, whilst in
5,778 vaccinated, there were 27 cases, with 14 deaths — i.e., the
incidence fell from 3 to under 0-5 per cent, the case-mortality
being 62 and 51 per cent, respectively. Numerous other
examples might be quoted, and the value of the method is now
proved to the full.

Plague.

The plague bacillus produces a powerful endotoxin, cultures
killed by heat being markedly irritating. There is some evidence
that a true exotoxin may be produced, though in small amounts.
Filtrates from young cultures are devoid of toxicity, whereas
those from older ones may be fairly potent. The fluid portion
of culture (in broth grown at 20° C. and kept well aerated)
two months old was found by Markl to kill rats in doses of
c'i c.c. This might, of course, be due to an autolysis of the bacilli,
but this seems improbable from the fact that the toxicity of the
filtrate is very easily destroyed by heat, whereas the endotoxin is
thermostable. These filtrates have slight immunizing properties,
but the plague anti-endotoxin has not been closely studied.

Immunity appears to be due to the production of bacteriolytic
substances : antiplague serum, prepared by immunizing horses
first with dead and then with living bacilli, is powerfully bacteri-
cidal. According to Wright, the plague bacillus is quite insensible
to the bactericidal action of human blood, and recovery is due to
opsonization followed by phagocytosis.

The agglutination reaction is well marked in artificially
prepared immune serum, which may clump at i : 1,000 or more
and may be of use in the identification of a doubtful bacillus. It
s not usually marked, and may be absent in human cases of the
disease, and the diagnosis is most frequently made by the
identification of the bacillus in fluid from a lesion or from the
blood or sputum. According to Cairns, the blood does not usually
clump until the disease has been in progress for about a week.
The strength of the reaction is not great, rarely rising above i : 50,
and is sometimes as low as 1:3 or 1:5. The macroscopic
method is advisable.

The curative treatment of the disease by specific methods
resolves itself into the use of a serum, vaccines not having been
tried, as far as I am aware. Several sera are prepared, but not all

PRACTICAL APPLICATIONS 403

have had an extensive trial. Yersin's serum is prepared at the
Pasteur Institute, the process being to immunize horses for long
periods — up to a year and a half — by weekly intravenous injections
(which do not cause abscesses, as is the case if the injections are
given subcutaneously). For the first three months or so dead
bacilli are used, afterwards living ones, and a very high degree of
immunity is attained. The potency of the serum is estimated by
finding the smallest amount which, given twenty -four hours
previously, will save a mouse from a lethal dose of living bacilli :
this may be as low as 0-02 c.c. Lustig's serum is supposed to be
antitoxic as well as bactericidal. It is prepared by the immuniza-
tion of horses with a "toxin" prepared by dissolving plague bacilli
in i per cent, caustic soda solution, filtering and precipitating with
dilute hydrochloric acid. (This has also been suggested as a
vaccine.) The precipitate is dissolved in 0*5 per cent, sodium
carbonate before use.

All observers are not agreed as to the efficacy of these sera, but
there is a decided preponderance of opinion in their favour.
Yersin's serum is most used, and is probably of the greater value.
A most important point in connection with its use is that large
doses are necessary, and those observers who have not obtained
good results have in some cases used quantities which were far
too small. Cairns used Yersin's serum in the Glasgow epidemic,
and in severe cases gave 150 to 200 c.c., part in the region draining
into the affected glands and part intravenously. Choksy, as the
result of large experience, urges the importance of a very early
use of the remedy, and gives 60 to 100 c.c. for adults and 10 c.c.
for infants, giving fresh injections of gradually diminishing amounts
every twenty-four hours, until six or eight have been given in all —
150 to 300 c.c. He used Lustig's serum. In any case the effect
of serum is not a great one, a lowering of the case-mortality by
about i o to 20 per cent, being apparently the utmost to be hoped
for at present. It appears, however, that no other treatment
available is so successful.

The question of the preventive treatment is much more im-
portant. In some cases the serum may be used, and is probably
most efficacious ; but its effects are but transitory, and its only
legitimate use is to tide the person over the time until vaccination
can be performed and active immunity acquired.

Haftkine's plague prophylactic consists of a virulent broth
culture of the bacillus, killed by heat and preserved by the

26 — 2

4°4

PLAGUE

addition of 0*5 per cent, carbolic acid. Cultures are made in
peptonized broth to which a small amount of oil is added. This
floats on the surface, and serves as a point of attachment for the
characteristic " stalactites." The flasks are kept at the ordinary
temperature (of Bombay — about 27° C.) and shaken occasionally,
to break up the stalactites. Incubation lasts five to six weeks.
The vaccine is sterilized at 65° C. for one hour. The dose is
2-5 c.c. Constitutional and local symptoms of moderate severity,
and lasting for a few days, are produced, but the patient is as a
rule able to follow his ordinary occupation. The immunity
seems to be developed quite quickly, so that there is no reason to
fear any ill-effects from the injections when the patient is actually
exposed to plague, and perhaps even infected. According to
Bannerman, the protection is developed in twenty-four hours, and
asts about eighteen months.

Of the value of the method there can be no doubt, and statistics,
both those on a large scale and those dealing with communities,
some of whom are vaccinated and some not, prove clearly that
the treatment lowers the likelihood of infection, and also the case-
mortality. Thus, in twelve districts in the Punjab in which
plague was raging in the winter of 1902-03 the following results
were obtained :

Total.

Cases.

Per
Cent.

Deaths.

Per Case-
Cent. Mortality.

Uninoculated (average

l

population of district)

639,630

49,433

77

29,733

47 60- 1

Inoculated (average

population of district)

186,797

3-399

1-8

814

0-4 23-9

With regard to the second group of statistics, the experience in
Umarkadi Gaol may be quoted, as one out of many. Half the
prisoners, selected purely by chance, were inoculated, and all
lived together under exactly the same conditions. Some of each
group were liberated, and of the remainder there were 127 non-
inoculated, with 10 cases and 6 deaths, and 147 vaccinated, with
3 cases and no death.

The German Commission recommended the use of vaccines
prepared from two-day-old agar cultures, sterilized by heat. This
is more easily and quickly prepared than Haftkine's fluid.

The combined method (use of vaccine and serum) has been

PRACTICAL APPLICATIONS 405

recommended by Calmette, by Besredka, and by Shiga ; the last-
named obtained very good results by its use in an epidemic in
Kobe.

Anthrax.

The nature of the toxin of anthrax is quite unknown, although
it has been the subject of much experimental investigation. No
exotoxin is formed in ordinary media. If coagulable or coagulated
proteids are present in the medium, they will be broken down into
peptones, etc., which have some toxic action, but no true toxin is
produced. Some observers have found that the filtrate from broth
cultures of anthrax, though devoid of toxicity, may have some
immunizing powers, a result which we should now attribute to
the presence of free receptors. The only importance attaching to
these facts is that they may explain the results obtained by some
investigators, who obtained albumoses and other bodies of very
feeble toxicity from various culture media, and considered them
to be the true toxin because they served to immunize animals.
And, according to Conradi, there is no evidence in favour of the
existence of an endotoxin. Bacilli killed by various methods and
disintegrated by Buchner's process yielded a non-toxic fluid.
The clinical nature of the disease in some of its manifestations
(especially pulmonary anthrax) would rather lead us to believe
that a powerful toxin is produced, but of this there is not the
slightest shred of experimental verification.

The process of recovery and the subsequent immunity are also
very difficult to understand. Local immunity is very marked,
the skin being highly resistant in comparison with the lungs, an
infection of which region forms one of the most rapid and intract-
able diseases known in man. There are very marked differences
with regard to the immunity of different animals. The fowl is
highly immune, as are cold-blooded animals. The rat and dog
are partially immune, whereas sheep, cattle, and the small animals
of the laboratory are very susceptible.

It is especially noteworthy in the case of anthrax that the
presence of bactericidal substances in the blood is no indication
whatever as to the degree of immunity* The serum of the rabbit,
a highly susceptible animal, has an extremely powerful bactericidal
effect, whereas that of the dog and rat have but little. The
classical Pfeiffer's phenomenon is not seen in the case of this
bacillus, but the altered bacteria may be readily recognized from

406 ANTHRAX

the fact that they fail to stain by Gram's method. This change
is brought about very quickly by a suitable serum, the change
being often complete in ten minutes at 37° C.

There have been numerous attempts to explain the apparent
anomalies of the reaction in question. Bail found that dog serum
(normally a good culture medium for the anthrax bacillus)
becomes highly bactericidal after the addition of a small amount
of rabbit serum, even when this is only present in amount so
small that it is devoid of bactericidal action per se. This appears
to be due to the presence of immune body in the dog's blood, but
no complement. If the action of the rabbit's serum is due to the
presence of complement, this must be thermostable, for the effect
is not annulled by heating to 50° C. Bail and Petterson found
that many other sera could be reactivated with rabbit serum
(man, ox, calf, pig, etc.), and that extracts of leucocytes or of
organs (liver, bone-marrow) might be equally effective. Malvoz
also investigated the presence of immune body by means of the
Bordet-Gengou reaction (absorption of complement), and found
that the amount in the serum was some index as to the degree of
immunity. Thus the blood of the ox and guinea-pig contain
none, as is the case with the newly-born puppy, an animal
susceptible to anthrax, whereas the adult dog contains a large
amount. Remy has also studied the question of the reactivation
of sera of various species by complements from others, and
notably that of the fowl. Thus the serum of the white rat (an
immune animal) contains an immune body, for after heating to
55° C. it can be reactivated with fowl serum. On the other hand,
the serum of the goat after heating cannot be rendered bacteri-
cidal in this way. He holds that there is an absolute concordance
between the bactericidal power of the blood, the presence of
immune body, and the resistance of the animal to infection with
this organism.

Sobernheim and others have explained the susceptibility of the
rabbit by supposing that the immune body has a greater affinity
for the cells of the animal than for the anthrax bacillus, and is
thus absorbed and rendered useless.

On the other hand, Metchnikoff holds that the immunity is
entirely due to phagocytosis, and finds that the extent to which
the bacteria are taken up by the leucocytes is proportional to the
degree of resisting power. Anthrax bacilli (and especially the
second vaccine, which forms a very good emulsion) are very

PRACTICAL APPLICATIONS 407

suitable objects for the study of phagocytosis. They are taken up
with great rapidity, and quickly undergo solution within the
leucocyte, first losing their sharp outline and power of retaining
Gram's stain, and disappearing altogether in ten minutes or less.
This makes the study of the opsonic index a matter of some
difficulty, which can be overcome by using isolated spores in test-
tube experiments. When no serum is used very few bacilli or
spores are taken up, and before the discovery of the opsonins
Metchnikoff noted that when rats are injected on the one side
with anthrax bacilli and on the other with the same organisms
mixed with blood-serum, oedema occurs only at the former place,
and it is from this that generalization occurs. Sawtchenko also
found that when the injection of the needle causes haemorrhage
the rat survives. The very careful and full researches of Metch-
nikofF on the degree of phagocytosis in susceptible and non-
susceptible animals are probably sufficient to lead us to believe
that the ingestion of the bacilli by the leucocytes is the all-im-
portant process in the cure of the disease, and the discovery of
the opsonins supplies the missing link necessary for us to account
for all the facts in a fairly satisfactory manner. We can only
conclude that the bactericidal effect of the serum plays a part of
comparatively small importance in combating the disease — the
elaborate researches of Bail, Petterson, Sobernheim, etc., to the
contrary — possibly, but by no means certainly, owing to the
absence of complement.

The facts of passive immunity are not so fully explained.
There are, however, some reasons for thinking that the active
substance is an opsonin, perhaps a thermostable one. Thus
Sclavo's serum (according to Cler) will render bacilli fit for
ingestion after five hours' contact, and it does not lose its efficiency
on keeping. On the other hand, the remarkably rapid improve-
ment sometimes seen after the -fise of Bandi's serum rather

suggests the presence of an antitoxin.

Diagnosis. — This is made in all cases by the demonstration of
the bacillus.

Treatment. — The preventive treatment is used for animals only.
Pasteur's method has already been noticed : it has been largely
used, and the results have, on the whole, been good. The
mortality from the inoculation is about \ per cent, of all cases,
but in some herds the number of deaths is much higher, and
serious loss is caused. The immunity is supposed to last for less

408 ANTHRAX

than a year, when a reinoculation is necessary. The method is
not free from objections, but its use in regions of France where
anthrax was very prevalent proved of enormous value, and areas
in which raising cattle and sheep was rapidly becoming im-
possible were practically cleared of the disease. The weak point
of the process is that the immunity to infection through the
alimentary canal, if it exists, is extremely feeble.

To remedy the defects of Pasteur's system, Sobernheim has
introduced a method of conferring mixed immunity. An anti-
anthrax serum and a culture resembling Pasteur's second vaccine
are injected simultaneously into different parts of the body, and
no second inoculation is given. The doses are 5 to 15 c.c. of the
serum and 0^5 to i c.c. of culture. This method of treatment is
said to be free from danger, to protect against infection via the
intestinal tract ; it has also the advantage of requiring only a
single visit. The serum is also curative.

Curative Treatment. — Here the use of serum is indicated. Sclavo's
serum is most used in this country. It is obtained by immuniz-
ing the animals with Pasteur's vaccines, and then by giving large
doses of virulent bacilli mixed with gelatin, which seems to
prevent the formation of abscesses. The dose is 20 to 40 c.c.,
repeated in twenty-four hours if necessary, or four or five doses
of 20 c.c. each : the first injection may advantageously be
intravenous. It is usually followed by improvement within
twenty-four hours, and often causes sweating and a rise of
temperature. Sobernheim's serum is obtained by a somewhat
different method, and appears to be equally efficacious. The dose
recommended is 20 c.c.

The results of the use of serum in malignant pustule (which is
not so dangerous a disease as was once thought, even if untreated
by serum, the knife, cautery, etc.) have been very satisfactory:
there do not seem to be any observations on its use in the
far more serious woolsorter's disease or pulmonary anthrax.
Malignant pustule is also treated by the use of very hot fomenta-
tions, the idea being to bring about the attenuation of the bacillus.
There is little doubt that vaccines might be used if thought
desirable in the absence of serum.

Diphtheria.

Diphtheria presents a close approach to our idea of a disease the
immunity to which is antitoxic, but it is erroneous to imagine that

PRACTICAL APPLICATIONS 409

the neutralization of the toxin or its destruction or elimination
constitutes the whole process of cure. There is a little evidence
in favour of the formation of bacteriolytic substances, though
experimental evidence on this point is not unanimous. Bandi, it
is true, claimed to have been able to immunize animals to the
bacilli themselves, and prepared a serum which was supposed to
have bactericidal properties ; it has been prepared by others, and
can be obtained commercially. It is supposed to be used locally,
either in the form of a powder or of lozenges, and is intended to
supplement the action of antitoxin. Rist, however, failed to
immunize animals to the bodies of the bacilli, and though Lipstein
was more successful, his serum was apparently inert as a protective
agent. It contained, however, an agglutinin, and the interesting
fact was noticed that it only clumps bacilli of the culture used
for the injection. This is of some interest, since the Klebs-Loffler
bacillus has always been looked upon as a very definite bacterial
species; the toxins it produces are always neutralized by the
same antitoxin, and though they may be produced in larger or
smaller amounts and may contain varying proportions of proto-
toxoids, etc. (on Ehrlich's theory), appear to be the same substance
in all cases. These experiments would tend to show that, though
the bacilli of various types agree in their metabolic products, they
may differ in the constitution of their protoplasm.

The observations referred to previously, show clearly that the
process of cure of the local lesion is assisted by the produc-
tion of an opsonin. And there is every reason to believe that
it is by phagocytosis that the bacilli are combated, bacteriolysis
being very doubtful and of comparatively small importance.
The cure of the disease therefore is accomplished partly by one
or more of the methods discussed in Chapter VI., and partly
by phagocytosis.

Diagnosis. — This is made by the demonstration of the bacillus.
If necessary, the opsonic test might be used, and Bordet and
Gengou have shown by their method of fixation of complement
that " sensibilatrices " circulate in the blood. These methods are
quite unnecessary. The absolute recognition of diphtheria
bacillus in cultures can best be made by an application of an
immunity reaction. A pure culture in broth is divided into two
parts, and each injected into a guinea-pig. One of the animals
receives a large dose of antitoxin, and should this remain unaffected
whilst the other dies, the culture is certainly diphtheria. The

410 TETANUS

method is usually only required in cases where a healthy person
contains diphtheroid bacilli in his mouth, nose, skin, etc., and
considerations of public health render a determination of their
exact nature necessary.

Treatment. — This consists in the early use of antitoxin and the
treatment of the local lesion with antiseptics, and the only question
of importance concerns the dosage of the former remedy. As a
rule, 4,000 to 8,000 units should be given at once, and a second
injection at the end of twelve or twenty-four hours ; subsequent
doses are given if required. Unless a case is seen very early, a
part at least of the first dose may be given intravenously, and
this is always advisable in severe cases not seen until the disease
has been present for two or three days. Larger doses may be
given, but are of doubtful advantage ; a smaller amount should
not be given, except perhaps in mild cases.

The sole preventive treatment in actual use consists in the use
of comparatively small doses of antitoxin. The protection which
is conferred is usually a strong one, but exceptions have been
recorded. It lasts about a month. Essays in vaccination have
been made, but not on a large scale.

Tetanus.

The pathology of tetanus is akin to that of diphtheria in that it
is a local disease with remote symptoms due to the action of a
soluble exotoxin on distant structures. It differs from diphtheria
mainly in two points : the bacilli are strictly localized to the region
inoculated and the immediate neighbourhood, and the toxin, which
acts entirely on the central nervous system, reaches it entirely, or
almost so, by ascending the nerves from the region in which
infection occurs, and not by circulating in the blood-stream. This,
at least, is the usual course of events, and when, as occasionally
happens, the toxin actually gains access to the blood, it seems
likely that even then it does not act on the brain direct, but enters
the peripheral nerves at their distal endings and then ascends
them to their origin.

The diagnosis is made entirely by the recognition of the
organism in the wound, no agglutination or other tests being used.
If (as usually happens) the culture obtained from the w?ound is
impure, it is divided into two parts, the one of which is injected
alone, the other in conjunction with tetanus antitoxin. If no
other pathogenic bacteria are present the animal that has received

PRACTICAL APPLICATIONS 4! I

the mixture will survive, whilst the other will develop tetanic
symptoms and die. Even if other pathogenic bacteria are present
the indications are usually clear, since spasms will commonly
develop (in the animal which has received no antitoxin) before the
lethal issue. It is best to use a broth culture for this test, so that
there may be a good development of toxins.

The nature of the toxins of tetanus have been already mentioned.
There are two, both exotoxins — the real poison, tetanospasmin, and
tetanolysin. Tetanospasmin is readily prepared by cultivation of
the organism in pure culture in almost any medium under anaerobic
conditions. It is even more fragile than diphtheria toxin, being
rapidly rendered inert in a few days if exposed to air at ordinary
temperatures. It is destroyed in eight to eighteen hours by
sunlight, by a temperature of 55° C. in one and a half hours, and by
exposure to agents such as alcohol, potassium permanganate, and
trichloride of iodine. It can be preserved by means of dilute
carbolic acid (0-6 per cent.) or chloroform without much loss.
Inert solutions have in general powerful immunizing properties,
the toxin being converted into toxoids, and not absolutely
destroyed.

It can be prepared so as not to give the reactions for proteid,
and is formed when the bacillus is grown on Uschinsky's proteid-
free medium. Its potency is enormous. Thus Vaillard prepared
a toxin of which the lethal dose for a guinea-pig was 0*001 c.c.,
containing about 0*000025 gramme of solid matter, only a small
portion of which was pure toxin. Brieger and Cohn calculated
that the lethal dose of an (impure) toxin for a man was 0-00023
gramme.

The effect of tetanus toxin is manifested almost solely on
the central nervous system, and the post-mortem lesions are
practically confined to the ganglionic cells, especially of the
anterior cornua. It appears probable that there is no direct
action on the nerves themselves, but the toxin, like the virus of
rabies, reaches the central nervous system mainly, if not entirely,
by ascending the nerves leading from the area of inoculation.
According to Meyer and Ransom, toxin which gains access to the
blood only affects the brain by entering the peripheral nerves via
the nerve endings, especially the end-plates, but this is not
universally accepted. As in the case of rabies, the richer the area
of inoculation in nerves, the more powerful the action of the toxin
and the shorter the period of incubation. The brain and cord are

412 TETANUS

the most susceptible regions, the peripheral nerves next, then
regions with an abundant nerve supply, such as the face ; and
lastly, regions poorly supplied, such as the subcutaneous and
peritoneal tissues. The incubation period of tetanus is thus seen
to be composed of : (i) the time necessary for the production of the
toxin in the tissues ; (2) for its ascent of the nerves to the brain
being longer, other things being equal, if infection takes place at
a long distance therefrom ; and (3) the latent period which elapses
after the toxin has united with the ganglion cells of the central
nervous system, and before the development of symptoms — i.e., that
in which the enzyme-like action of the zymophore group is being
gradually exerted on the protoplasm. The fixation of tetanus
toxin in the system is extremely rapid: in rabbits it may dis-
appear entirely from the blood in one minute, whilst in other
susceptible animals it circulates for slightly longer periods. The
importance of this arises from the fact that toxin which has once
entered the nerves is thereby shielded from the action of antitoxin.
The dose of antitoxin necessary to save the life of an animal which
has received a few lethal doses of toxin rises enormously if the
injection of the former is delayed more than a few minutes.

Tetanolysin is even more fragile than tetanospasmin, being
converted into toxoids in a few hours at the room temperature.
It can be preserved in a dry state. The role which it plays in
natural infections, if any, is unknown.

As regards immunity, there is but little to add to what has been
discussed previously. The bacilli are not powerful parasites,
being readily ingested by the leucocytes, and destroyed if the
conditions are favourable for phagocytosis. In most of the cases
which develop tetanus there is a contused or lacerated wound,
with much killed and bruised tissues and an abundant con-
comitant infection with other bacteria, which still further paralyze
the natural resistance of the part. These organisms may have
an additional influence in securing a condition of anaerobiosis :
tetanus bacilli grown in symbiosis with certain other bacteria
which have powerful oxygen-absorbing properties will develop
vigorously, and develop toxin in spite of the free access of air.
No observations with regard to the opsonic index in tetanus
appear to have been recorded. The question of immunity to
tetanus toxin has been dealt with already, but we may add that
in all probability much of the toxin is destroyed in loco by the
unspecific action of the peptic enzyme formed by the leucocytes

PRACTICAL APPLICATIONS 413

Antitoxin is rarely, if ever, found in human patients who have
survived an attack of the disease.

Treatment. — The main question, of course, concerns the use of
antitoxin, and two general rules may be laid down : (i) It is of
great value as a prophylactic agent, and (2) it is of some value in
chronic tetanus— i.e., the form with mild symptoms developing after
a long period of incubation.

Its preventive application is indicated in the treatment of all
wounds which experience has shown to be followed by tetanus
— i.e., lacerated and contused wounds, especially if contaminated
with garden soil, road debris, etc. Gunshot wounds are especially
dangerous, and tetanus is usually extremely prevalent in warfare.
It is, of course, somewhat difficult to estimate precisely the value
of the treatment, inasmuch as tetanus is not a common disease ;
but experience derived from horses, which animals are extremely
prone to it, is more conclusive. In some veterinary practices it
was so common as to counterindicate any operative measure, and
has now been completely eradicated. The duration of the
immunity conferred by a single dose is about three weeks, and
in the prophylactic treatment of wounds, whether accidental or
due to operation, two doses should be given, at intervals of ten
to fourteen days. The prophylactic treatment of dirty wounds by
means of antitoxin is now a routine method in several Continental
clinics, and, as far as I am aware, there has been no case
recorded in which it has been followed by the development of the
disease, excepting those in which the injection has been given
some days after the injury, when the toxin has already gained
access to the nerves. One case (under Mr. Lenthal Cheatle)
from which I isolated a bacillus identical in cultural and morpho-
logical characters with that of tetanus, and in which the organisms
occurred in great abundance, was treated with antitoxin at the
outset, and healed without a symptom of the disease : the culture
was unfortunately not tested by inoculation.

Calmette prepares a powder of dry antitetanus serum to be used
as a dressing for wounds, but its use is very doubtful. Anti-
toxin is a good culture medium for bacteria, and unless the wound
is fairly clean may decompose and become offensive. The most
scrupulous antiseptic technique should be adopted, and it seems
probable that the dry dressing presents no advantage over the
subcutaneous administration of the serum when this is done.

The doses should be 5 to 10 c.c. for a man, and 10 to 20 c.c. for

4*4 TETANUS

a horse. The best method of standardization is that of Roux, who
determines the amount of serum necessary to protect a guinea-pig
weighing 500 grammes against ten lethal doses of toxin. The result
is expressed in terms of the weight of guinea-pig protected against
one lethal dose of antitoxin by i c.c. of serum — e.g., if ^ c.c. pro-
tected a guinea-pig weighing 500 grammes against ten lethal doses,
the potency would be 50,000. A potency of 1,000,000 is the least
that should be employed.

The use of tetanus antitoxin in the developed disease is less
satisfactory, a fact readily explicable nowr that the pathology of
the disease is more fully understood. In acute tetanus it is
practically worthless, though a few cures have been reported. In
many cases of chronic tetanus it is without action ; in a few,
however, it is decidedly beneficial, each injection greatly alle-
viating the patient's suffering. It is always worthy of trial, but
it is hardly necessary to say that the non-specific treatment
should not be neglected. If the patient has a sufficient degree of
immunity to resist the toxin which has already gained access to
his nervous system, the antitoxin will be of value in preventing any
more from doing so, inasmuch as it will neutralize it as soon as it
is formed.

The doses should be large — 20 c.c. or more at first, and 10 c.c.
every day, or every alternate day, subsequently. The site of
inoculation is of some importance. The injections may be given
subcutaneously in a distant region, as in the use of diphtheria
antitoxin ; but, in view of the fact that it takes an appreciable
time for it to be absorbed — and time is of the utmost value if the
remedy is to be of any use — it seems advisable to give the first
dose either in the region of the wound or intravenously.

Various methods have been proposed by which the antitoxin
can be brought into closer relation with the nerve elements.
The intracerebral injection has most to recommend it on theoretical
grounds, and several very decided successes have been recorded
in severe cases of the disease. The method is as follows : A
small flap of the scalp (with its base downwards) is reflected so
as to expose the skull a little to one side of the middle line, and
just in front of the fronto-parietal suture. A small trephine hole
is made through the skull, and an exploring needle is inserted
until the lateral ventricle is reached, and cerebro-spinal fluid
escapes through the needle. Ten c.c. or more of the serum are in-
jected. This- passes down the ventricular system, and bathes the

PRACTICAL APPLICATIONS 415

respiratory and cardiac centres at the floor of the fourth ventricle.
Another method is to inject small quantities of the fluid directly
into the spinal cord by means of a needle introduced between the
sixth and seventh cervical vertebrae. This procedure would
appear to be dangerous, but this is said not to be the case.
Lastly, the simplest method of all is to perform lumbar puncture,
draw off some of the cerebro-spinal fluid, and replace it with
serum, just as in the process of spinal anaesthesia.

Ransom and Meyer have advocated the direct application of
antitoxin to the nerves supplying the region in which the wound
is situated, the idea being, of course, to intercept any further
access of toxin to the brain and cord. The nerves are exposed
by operation as near to their origin as possible, and infiltrated
with serum by means of a hypodermic syringe.

Analogy with other diseases would fully justify the use of
vaccine in chronic tetanus. Its preparation would present some
difficulties, owing to the heat-resisting power of the spores.

Syphilis.

Little is known definitely concerning the mode of cure or of the
type of immunity of syphilis. It used to be regarded as one of
the diseases which are followed by practically complete immunity
of long duration, but .Neisser has brought forward some evidence
for thinking that this is not the case, and that it only lasts as
long as the disease itself — i.e., as soon as it is completely eradicated
the patient is again susceptible. Nothing is known as to the
toxins of syphilis, and, as regards the method of cure, the only
point worth mentioning is the fact that spirochaetes which have
been ingested by the leucocytes can be ma^e out occasionally.
They stain badly, and are doubtless on the way to complete
absorption. The fact that the organism cannot be obtained in
pure culture renders researches with regard to the opsonic and
bacteriolytic action of the serum very difficult. Indirect researches
by means of the deviation of complement — constituting the
Wassermann reaction, a special method of application of the
Bordet-Gengou reaction — have led to results of great interest
which have recently attracted much attention.

The first necessity was, of course, the preparation of an antigen,
and for this purpose Wassermann made use of the internal organs
of a syphilitic foetus, which were swarming with spirochaetes. In

416 SYPHILIS

its main outlines the technique is exactly the same as that already
described. The serum to be tested is heated, to remove comple-
ment, and diluted with sterile normal saline solution. A dilution
of i : 20 or i : 40 is generally correct, but the point may be deter-
mined by preliminary tests with a known syphilitic serum ; and in
any case it is an advantage to perform a series of tests with
different dilutions, so that a rough idea of the amount of antibody
present in the serum may be obtained. This is mixed with an
extract of the syphilitic organ (antigen), and some fresh guinea-
pig serum (complement) added. The proportions may be i c.c.
of diluted serum, 0*1 or 0-2 c.c. of organ extract, and 0*2 c.c. of fresh
serum. The whole is incubated for one hour, at the end of which
time all the complement will be removed from the fluid if
syphilitic antibody is present. Next, corpuscles (e.g., of a sheep or
pigeon) are added, together with heated serum from a rabbit
which has been injected with the corpuscles in question ; or the
corpuscles may previously be sensitized with the inactivated
serum, washed, and then added. The whole mixture is then
incubated for two hours, with occasional stirring or shaking,
and kept some hours in the ice-chest. A positive reaction is
shown by the absence of haemolysis. Control tests are also
advisable — e.g., the corpuscles must be completely dissolved by
the heated immune serum and the guinea-pig's serum if the other
two ingredients are not added, and there should be no haemolysis
if all the substances except the guinea-pig's serum are used.

Ledingham and Hartoch have shown independently that
opsonin is absorbed as well as complement, and this fact may
be used as a test of the presence of the reaction. In this case the
first part of the test is performed as beforehand the fluid used as
the serum in an opsonin estimation, using staphylococci or any
other organism, and using as a control guinea-pig serum diluted
with normal saline to the same extent as it was in the mixture of
organ extract, human serum, and guinea-pig serum. In a positive
reaction the phagocytic index in the first preparation will be
much below that in the second ; in a negative one they will be
equal.

The exact value of the test is not yet quite definitely settled.
It is very rarely present in health, and not common in diseases
other than syphilis ; but it does occur, especially in diseases which
(like syphilis) are due to animal parasites, such as malaria or
trypanosomiasis, and is not uncommon in leprosy and scarlet

PRACTICAL APPLICATIONS 417

fever. It is rarer in other diseases, but isolated examples have
been met with in systematic investigations in a great many
maladies ; but here it is the exception, whereas in syphilis it is
the rule. In primary and secondary cases it occurs in 90 per
cent, or more, and is present in the majority of patients suffering
from tertiary syphilis and " metasyphilitic " affections. It is very
frequently found in the cerebro-spinal fluid of general paralytics
(80 per cent, 90 per cent., or more), even when it is absent from
the blood. It is not so common in tabes, and is extremely rare
(if it ever occurs) in the cerebro-spinal fluid in non-syphilitic
diseases, with the curious exception of scarlet fever, in which it
is almost constant.

So far there is no theoretical difficulty in the interpretation of
the phenomenon, but a new fact discovered by Landsteiner,
Miiller and Potzl seems to show that the reaction is of a nature
entirely different from the ordinary Bordet-Gengou phenomenon.
They found that an alcoholic extract of a normal organ (e.g., of a
guinea-pig's heart muscle) might be used instead of a tissue rich
in spirochsetes ; and further researches have shown that the lipoid
substances isolated therefrom, or even comparatively simple
substances, as lecithin and taurocholate and glycocholate of soda
(Levaditi and Yamanouchi) give the reaction, although apparently
not so frequently, as when an extract from a syphilitic organ is used.
The "antigen" is soluble in hot alcohol, and this fact alone removes
it from the group of true antigens, which, as we have seen, are
apparently all proteid in nature. According to Levaditi, the sub-
stance occurring in the blood or cerebro-spinal fluid is not an anti-
body at all, but either lipoid substances or salts, or the two in
combination, and they are set free when tissues are broken down in
a certain way, which occurs most frequently in syphilis, but may
take place in other diseases. Under ordinary circumstances they
are present in a colloid state/but form a precipitate with the lecithin
and allied substances extracted from normal organs by hot alcohol,
and to this precipitate the complement attaches itself. According
to Forges, the serum of syphilitics has the power of precipitating an
emulsion of lecithin (0*5 gramme, shaken up with 0*5 per cent,
solution of carbolic acid in normal saline) when mixed therewith
in equal parts. This he proposed as a test for syphilis, and Nobl
and Arzt found it successful in 80 per cent, of cases. Subsequently,
Forges replaced the lecithin (which as usually bought is not con-
stant in composition) by a recently prepared I per cent, solution

2?

418 RABIES

of glycocholate of soda. A mixture of this with an equal amount
of serum is incubated for five hours, and observed after it has
stood sixteen to twenty hours at the room temperature. The
precipitate is specially obvious near the surface.

Fornet, Schereschewsky, Eisenzimmer, and Rosenfeld find
that the sera of syphilitics in the early stages of the disease
contain a precipitogen which forms an insoluble compound with
a substance or precipitin present in the serum of tabetics or
general paralytics. When the one is floated on the other, a
characteristic ring appears at the area of contact. They say that
normal serum rarely contains the precipitin, but not the
precipitogen. What relation this has to any immunity reaction
is unknown.

Rabies.

The actual causal agent of rabies is not yet definitely ascertained.
The peculiar structures known as the corpuscles of Negri which
occur in the brain, and especially in the hippocampus major, of
rabid animals appear to be quite characteristic of the condition,
and may possibly be the actual parasite, although this is not yet
universally accepted. It seems, however, fairly certain that their
recognition constitutes a sufficient proof of the presence of the
disease ; and this is of great importance in view of the necessity
for the early commencement of the treatment, which is entirely
preventive, and not curative. If the dog by which the patient has
been bitten is forthcoming, the corpuscles of Negri can be demon-
strated in a short time by simple methods, and the need for
Pasteur's treatment ascertained ; apart from this the only method
is by animal inoculation, an emulsion of brain substance being
injected into the brains of rabbits after trephining.

Rabies presents one of the most striking examples of local
immunity ; the action of the virus is manifested almost entirely
on the central nervous system, and in whatever part of the body
the inoculation is made the effects are only caused when it has
reached the brain and spinal cord ; and in doing so it does not
gain access to the blood, but ascends the peripheral nerves.
Hence the central nervous system is extremely susceptible to
injection, and the other tissues in proportion to their richness in
nerves. Subcutaneous (unless into a region like the paw), intra-
venous, or intraperitoneal injections only convey the disease if a
large amount of extremely potent virus is used. Hence it seems

PRACTICAL APPLICATIONS 419

reasonable to suppose that there is a fair amount of immunity
inherent in all the tissues except in the nervous structures, and
that the living virus deposited elsewhere may be entirely destroyed
by bacteriolysis or phagocytosis.

We have already glanced briefly at Pasteur's earlier work on
antirabic inoculations, and the method by which immunity is
produced. Numerous modifications of the process have been
introduced since Pasteur's time. Thus Hogyes of Budapest
makes use of fully virulent cords, but given in extremely small
doses ; and there is some reason to think that his process does
not really differ from that of Pasteur, and that in drying the cords
the virus is gradually destroyed and not really attenuated, so that
a dose of a fourteen-day cord really contains a small trace of
virus of full virulence. A true vaccine — i.e., a virus of mitigated
virulence— can be obtained by passage through monkeys or birds.
Further, though the fixed virus is so potent for rabbits, it is quite
possible that its virulence for man is slight or nil. Nitsch was so
sure of this that he injected 4 to 5 milligrammes of the fresh cord
subcutaneously unto himself (in the abdominal region, a part
comparatively poor in nerves) without evil results.

Another method, introduced by Marie, consists in the use of injec-
tions of a mixture of virus and serum from an immunized animal.
This serum is prepared in a variety of ways, the simplest being to
give the virus intravenously. The animal usually employed is the
sheep, and the injection consists of rabid brains, hcatcclf up into a
fine emulsion with normal saline solution, and filtered through
linen. The serum prepared from animals treated in this way
possesses powerful ancirabic properties : when mixed with a
potent virus it removes entirely all harmful properties, so that it
is quite innocuous even on intracerebral injection. It can be
titrated against an emulsion of fixed virus of definite strength,
and by appropriate treatment a very potent serum can be obtained.
It is apparently quite useless in the treatment of the developed
disease or of an infected animal, even before the development of
symptoms. If it is mixed in excess with fixed virus and injected
into animals these do not develop rabies ; on the other hand, but
little immunity is produced, and this is supposed to be due to the
fact that the virus is so quickly absorbed that it does not act as
an antigen. But if the mixture be allowed to stand for some
time, and the virus then recovered by centrifugalization and
washing with normal saline solution, the clot thus obtained has

27—2

42O RABIES

powerful immunizing properties. In the preventive treatment of
rabies on Marie's system the fresh fixed virus, made into a fine
emulsion with normal saline solution, is partially neutralized with
immune serum, and a dose of 6 c.c. (2 c.c. of i : 10 emulsion of
virus and 4 c.c of serum) is given in two places under the skin of
the abdomen. This is done for four days, and then injections of
dried cord, beginning at that of the sixth day, are commenced.

Other methods involve the use of heat, of chemical methods
(e.g., partial digestion with gastric juice, as practised by Centanni),
in order to bring about attenuation or partial destruction of the
virus.

Whatever the method, it appears necessary that the patient
should undergo a course of active immunization, various causes
(e.g., the long incubation period and the localization of the virus
in the nerves) rendering passive immunity an unsafe method of
protection.

Of the value of the process there cannot be the slightest doubt.

The incidence of hydrophobia after the bite of a rabid animal
is variously estimated, the figures usually given being about
15 per cent, in the case of dog-bites, and 40 to 80 per cent,
in bites from wolves. The probability of the patient's developing
the disease depends on the severity of the bite, its position
(i.e., whether in regions rich in nerves or the reverse), and
on whether the bite is through the clothing, so that some of
the virus is wiped from the teeth. In the twenty-two years (down to
1907 inclusive, the last year of which the figures are available),
^0,359 patients have been treated at the Pasteur Institute in
Paris, with 126 deaths — a death-rate of 0-31 per cent. (The
patients dying within fifteen days of the commencement of the
treatment — a small number — are excluded from the figures, since
in them the disease was too far advanced for a preventive treat-
ment to be of value.)

BIBLIOGRAPHY

[/4 complete bibliography of the subject being obviously impossible in any reasonable
space, an attempt has been made to include important articles, and especially
those referred to in the text, and articles dealing with the subjects in a complete
manner, especially those with a good account of the literature.'}

CHAPTER I

GOOD accounts of the general phenomena of immunity may be found in
Metchnikoff's " L'Immunite dans les Maladies Infectieuses " (English trans-
lation by F. G. Binnie, University Press, Cambridge) ; Ricketts' " Infection,
Immunity, and Serum Therapy " (American Medical Associated Press,
Chicago) ; in Clifford Allbutt's " System of Medicine," vol. ii., part i. ; in
Muir and Ritchie's " Bacteriology." Also Levaditi's " La Nutrition dans
ses Rapports avec I'lmmunity " (Masson et Cie.) ; discussion on Immunity,
Brit. Med. Assoc., 1904 (" B. M. J.," September 10, 1904). The admirable
abstracts and collected articles in the " Central blatt f. Bakteriologie "
(Referate), in the " Bulletin de 1'Institut Pasteur," and in " Folia Haemato-
logica," will be found invaluable.

Cold and Wet. — See Trommsdorf, Arch. f. Hyg., vol. lix., p. i, and
Vincent, Bull. Acad. Med., 1908. Ciuca, Comptes Rendus Soc. Biol.,
vol. Ixii., pp. 858, 883. Alcohol. — See Friedberger, Congres internat.
d'Hyg. and Demog. (Brux.), 1903. Rubin, Journ. Inf. Dis., 1904, p. 424.
Trommsdorf (vide supra}. Laitinen, Zeit. f. Hyg., vol. Iviii., 1907, p. 139.

Anesthesia. — Snell, Berlin. Klin. Woch., 1903, p. 212. Rubin, Journ.
Inf. Dis., vol. i., p. 424.

Ehrlich (Trypanosomiasis, Atreptic Immunity, etc.), Harben Lectures
(H. K. Lewis).

Walker, Ainley, Journ. Hyg., vol. iii., p. 52 ; Cent. f. Bak., vol. xxxiii.,
p. 297 ; Journ. Path. Bact., 1903, p. 34.

Papers on The Early Work on Immunity against Anthrax, etc., will be
found in Microparasites in Disease, New Sydenham Soc., 1886. See also
Pasteur, Comptes Rendus de 1'Acad. des Sci., 1880. Pasteur, Roux,
and Chamberland, ibid., 1883, xcvii. Rabies. — See Sims Woodhead's
article in Clifford Allbutt's System of Medicine, with a full bibliography
up to 1906. See also Chapter XIV.

A useful account of the Use of Vaccines, etc., in Veterinary Practice is
given in Jowett's Blood-Serum Therapy (Bailliere, Tindall and Cox, 1907).

CHAPTER II

A full account of the subject will be found in Oppenheimer's " Toxine und
Antitoxine " (English translation by Ainsworth Mitchell : Charles Griffin
and Co.). This contains a most useful bibliography extending to 1904.

Antagonism of B. pyocyaneus and B. anthracis. — Woodhead and Wood,
Edin. Med. Journ., 1890. Nasik vibrio. — Kraus, R., Centr. f. Bakt. I. O.,

421

422 BIBLIOGRAPHY

vol. xxxiv., 1903, p. 488, and Rothberger, ibid., vol. xxxviii., 1905, p. 165.
Absorption of Toxins by Tissue. — Ignowtowsky, Cent. f. Bakt. T. O.,
vol. xxxv., p. 4. Vaillard, quoted by Metchnikoff (L'Immunite).

Wassermanris Experiment. — See Chapter IV.

Combining Reactions of Tetanolysin. — Ehrlich, Berlin. Klin. Woch.,
1898, p. 273. Madsen, Zeit. f. Hyg., 1899, xxxii., 214.

Action of Tetanus Toxin on Frogs at Different Temperatures. — Courmont
and Doyon, Comptes Rendus de la Soc. Biol., 1893, p. 618. Morgenroth,
Archives Int. de Pharmacodyn, 1900, p. 265. Constitution of Toxin
Molecule. — Ehrlich, Croonian Lecture, Proc. Roy. Soc., 1900. Ibid.,
Congres internat. de Med., Paris, 1900, Klin. Jahrbuch, vol. vi. (Die Werth-
bemessung des Diphthericheilserums).

Leucolysins or Leucotoxins. — Neisser and Wechsberg, Zeit. f. Hyg.,
vol. xxxvi., 1901, p. 300. Kerner, Julius, Cent. f. Bakt. I. O., vol. xxxviii.,
p. 223. Christian, H. A., Deut. Arch. f. Klin. Med., vol. Ixxx., p. 333.
Denys and van de Velde, La Cellule, vol. xi., p. 359.

Bacterial Hcemolysins in Relation to Toxicity. — Besredka, Ann. de 1'Inst.
Past., vol. xv. Ruedinger, Journ. Amer. Med. .Assoc., 1903. Breton,
Comptes Rendus de la Soc. Biol., vol. lv., p. 886. Schlesinger, Zeit. f. Hyg.,
vol. xliv., p. 428. Bacterial H&molysins. — A full bibliography will be
found in Oppenheimer, and a good general account of the subject by
Besredka, Bull, de 1'Inst. Pasteur, vol. i., 1903, pp. 547, 579.

Ricin. — Ehrlich, Deut. Med. Woch., 1891. Fortschr. d. Med., 1897.
Stillmarck, Arb. pharm. Inst. Dorpat (quoted by Oppenheimer). Jacoby,
Arch. exp. Path., xlvi., p. 28. Osborne and Mandel, Amer. Journ. Phys.,
vol. x., p. 36.

Serum Toxin. — Cartwright Wood. See Chapter II.

Endotoxins (Pyocyaneus}. — Wassermann, Zeit. f. Hyg., xxii., p. 263.
Endotoxins in general: Macfadyen, Proc. Roy. Soc., 1903, p. 76; 1903,
p. 351. Macfadyen and Rowland, Cent. f. Bakt. I. O., vol. xxxiv., p. 618.
Macfadyen, Lancet, 1904, p. 494. Macfadyen and Rowland, Journ. Phys.,
vol. xxiii. Macfadyen, B. M. J., 1906, p. 776. Also Vaughan and Wheeler,
Journ. Amer. Med. Assos., 1905. Ransom, Deut. Med. Woch., 1895, P- 457-
Petterson, Cent. f. Bakt., vol. xlvi., p. 405. Pfeiffer and Friedberger, Cent,
f. Bakt. I. O., vol. xlvi., p. 98. Metchnikoff, Roux and Taurelli-Salembeni,
Ann. de 1'Inst. Past., vol. x., p. 257. Besredka, Ann. Inst. Past., 1906,
pp. 81, 304.

See also the discussion on Endotoxins (Kraus especially), Cent. f. Bakt.
(Ref.), vol. xlii.

CHAPTER III

Methods of Preparing Antitoxin, etc. — Levaditi, in Kraus and Levaditi's
Handbuch, vol. ii., p. 62. Woodhead, Report of M. A. B., 1901. Hewlett's
Serumtherapy (Churchill, 1903). Dean, Trans. Path. Soc., vol. 1L, p. 15.
Madsen, Zeit. f. Hyg., vol. xxiv., p. 425. Hibbert, Journ. Exp. Med\,
vol. vii., p. 176. Martin, Ann. Inst. Past., vol. xii,, p. 26. Park and
Williams, Journ. Exp. Med., 1896, No. i. Atkinson, Journ. Med. Res.,
vol. ix., p. 173. Salomonsen and Madsen, Ann. Inst. Past., vol. xi.,
p. 315, and vol. xii., p. 763. Hibbert, Journ. Exp. Med., vol. vii., p. 176.

Serum-toxin. — Cartwright Wood, Proc. Roy. Soc., vol. lix., p. 290 ;
Cent. f. Bakt., vol. xxxi., p. 241.

CHAPTER IV

Action of Ricin on Red Blood-Corpuscles. — Ehrlich, Fortsch. der Med.,
1897, P- 41- Of Snake Venom. — Stephens and Myers, B. M. J., 1898,
vol. Ixiii., p. 20. Of Eel Serum. — Kossel, Berlin. Klin. Woch., 1898, p. 152.
Camus and Gley, Ann. Inst. Past., 1899, P- 779- Tchistovitch, ibid., 1899.
Action of Leucocidin. — Neisser and Wechsberg, Zeit. f. Hyg., 1901, p. 299.

BIBLIOGRAPHY 423

Filtration Experiments. — Martin and Cherry, B. M. J., 1898, p. 1120.
Brodie, Journ. Path, and Bact., 1897, p. 460. Action of Heat on Snake
Venom, etc. — Calmette, Ann. Inst. Past., 1895, p. 225. Martin and
Cherry, Proc. Roy. Soc., 1898. Wassermann, Zeit. f. Hyg., vol. xxii.,
p. 263. Marenghi. Cent. f. Bakt. I. O., vol. xxii., p. 521.

Constitution of Diphtheria Toxin. — Ehrlich, Die Wertbemessung des
Diphtherieheilserums, and numerous other papers, some of the more
important of which are in his Collected Papers. Madsen, B. M. J., 1904,
p. 567. Oppenheim, Toxin and Antitoxin. Gruber and Pirquet, Munch,
med. Woch., vol. 1., pp. 1193, 1259.

Arrhenius and Madsen' s Theories are discussed at great length in the
former's Immuno-Chemistry (Macmillan), where full references are given.
See also Madsen, B. M. J., 1904, September 10. Myers, Lancet, 1898, vol. ii.,
p. 23, and Journ. of Path, and Bact., 1900. Mouton, Bull, de 1'Inst. Past.,
vol. v., 1907, p. 449 (with bibliography). Arrhenius and Madsen, Zeit.
f. Phys. Chem., vol. xliv., p. 6. Gruber and v. Pirquet, Miinch. Med.
Woch., vol. 1., p. 1193. Arrhenius, Bull. Inst. Past., vol. ii., 1904, p. 553
(good general account). Madsen and Walbum, Cent. f. Bakt. I. O.( vol.
xxxvi., p. 242. Mainwaring, W. H., Journ. Inf. Dis., vol. iii., p. 638.
Morgenroth and Pane, Biochem. Zeit., vol. i., p. 354. Nernst, Zeit. f.
Electrochemie, x., p. 177. Ehrlich's Reply to Arrhenius Theory, Berlin.
Klin. Woch., 1903, Nos. 35 and 37 (XXXVII. in Collected Studies).

Bordet's Views. — Bordet, Ann. Inst. Past., vol. xvii., p. 161. Eisenberg,
Cent. f. Bakt., xxxiv., p. 259. Biltz, Zeit. f. Phys. Chem., 1904, p. 615.
See also Chapter XII.

CHAPTER V

Action of Electricity on Toxins. — Kruger, Deut. Med. Woch., 1895, p. 331.
D'Arsonval and Charrin, Comptes Rendus de la Soc. de Biol., 1896, pp. 122,
280. Marmier, Ann. de 1'Inst. Past., vol. x., p. 468. Knorr, Miinch.
Med. Woch., 1898, p. 321.

Antibodies in Normal Blood. — Metchnikoff, L'lmmunite, p. 598. Neisser,
Deut. Med. Woch., 1900, p. 791. Cobbett, Lancet, 1899, p. 332. Ibid.,
Cent. f. Bakt., vol. xxvi., p. 548. Meade, Bolton, Journ. Exp. Med.,
vol. i., p. 543.

Regeneration of Antitoxin after Bleedings. — Roux and Vaillard, Ann. Inst.
Past., vol. vii., p. 64. Salom onsen and Madsen, ibid., vol. xii., p. 763.
Action of Pilocarpin. — Salomonsen and Madsen, Comptes Rendus de
1'Acad. des Sciences, vol. cxxv., p. 122.

Side-Chain Theory. — Ehrlich, Croonian Lecture, Proc. Roy. Soc., vol.
Ixvi., p. 424.; Ver. f. Innere Med. Berlin, 1901. Numerous articles in
Collected Studies. See also Aschoffs Ehrlich's Seitenkettentheorie
(Fischer, 1902), with a very full bibliography. Levaditi, La Nutrition dans
ses rapports avec I'lmmunite (Masson and Cie.), which also gives numerous
references. Wassermann, Berlin. Klin. Woch., 1898. Plimmer, Journ.
Path, and Bact., 1898, p. 489. Figs. 22, 23, 24 are from Emery, The
Specific Antibodies, St. Bart.'s Hosp. Journ., 1902. Weigert, Verhandlung
der ges. Deutscher Naturforscher und Aerzte, 1896. Bruck, Zeit. f. Hyg.,
vol. xlvi., p. 176.

Union of Tetanus Toxin with Brain Substance. — Wassermann and
Takaki, Berlin. Klin. Woch., 1898. Metchnikoff, Ann. Inst. Past., vol. xii.,
pp. 81, 263. Marie, ibid., p. 91. Courmont and Doyon, Comptes Rendus
de la Soc. Biol., 1898, p. 602. Joukowsky, Ann. Inst. Past., vol. xiii.,
p. 464. Morax and Marie, ibid., vol. xvii., p. 335, and Comptes Rendus
Soc. Biol., vol. liv., p. 1535. Dmitrevsky, Ann, Inst. Past., vol. xvii.,
p. 148. Besredka, ibid, p. 138. Muller, Cent. f. Bakt. I. O., vol. xxxiv.,
p. 567. Landsteiner and Boteri, ibid., vol. xlii., p. 562. Wolff-Eisner and
Rosenbaum, Berlin. Klin. Woch., 1906, p. 945. Takaki, Beitr. z. Chem.

424 BIBLIOGRAPHY

Phys. und Path., vol. xi., p. 238. Morax and Tiffaneau, Comptes Rendus
de la Soc. Biol., vol. Ixii., p. 15. Noon, Journ. Hyg., vol. vii., p. 101.

Romer's Experiments. — Arch. f. Opthal., vol. lii., p. 72.

Antispermotoxin. — Metchnikoff, L'Immunite, p. 130. Blum, Beit. z.
Chem. Phys., 1904. Vaillard and Vincent, Ann. Inst. Past., vol. v., p. i.
Besredka, Ann. Inst. Past., vol. xiii., pp. 49, 209.

CHAPTER VI

Non-Specific Processes. — Herter, Lectures on Chemical Pathology
(Smith, Elder, 1902).

Function of Liver. — Brunton, Sir L., and Bokenham, Journ. Path. Bact.,
1905, p. 50.

Antitoxin in Blood after Diphtheria, etc. — Wassermann, Zeit. f. Hyg.,
vol. xix., p. 408. Abel, Deut. Med. Woch., 1894, pp. 899, 936. Vincenzi,
ibid., 1898, p. 247. Knorr, Munch. Med. Woch., 1898, p. 363.

Absence of Correlation between Immunity and Antitoxin. — Roux and
Vaillard, Ann. Inst. Past., vol. vii., p. 64. Behring and Kitashima, Berlin.
Klin. Woch, 1901, p. 157. Metchnikoff, L'Immunite, pp. 386 et seq.
Behring, Allgemeine Therapie der Infectionskrankheiten, in Eulenberg
and Samuel's Lehrbuch der Allg. Therapie.

Leucocytes in Intoxications. — Metchnikoff, loc. cit., p. 413, where nu-
merous references are given. Besredka, Ann. Inst. Past., vol. xiii., pp. 49,
205 and 465. Dean, Journ. Path. Bact., 1908, p. 154. Ewing, Clinical
Pathology of Blood (Kimpton, 1904), p. 292. Vincent, Ann. Inst. Past,
vol. xviii., p. 450.

Stimulins. — Metchnikoff, loc. cit., p. 284.

Wassermann, N. Y. Med. Journ., 1904.

Immunization to Eel Serum. — Tchistovitch, Ann. Inst. Past., vol. xiii.,
p. 406.

Specific Processes. — See Ehrlich's Croonian Lecture, Harben Lectures,
and various papers in his Collected Studies. Wassermann and Bruck,
Deut. Med. Woch., 1904, p. 764. Jacoby, Beit. z. Chem. Phys., vol. vi.,
p. 113. Bruck, Zeit. f. Hyg., vol. xlix, p. 282. Ricketts, Trans. Chicago
Path. Soc., vol. vi. Metchnikoff, L'Immunite, Chapters XI, XII. Leva-
diti's L'Immunite dans ses Rapports avec la Nutrition.

Passive Immunity. — McClintock and King, Journ. Inf. Dis., vol. iii.
p. 701. Goodman, ibid., vol. v., p. 184. Bulloch, Journ. Path. Bact.,
1898, p. 274. Schiitze, Koch's Festschrift, p. 657. Pfeiffer and Fried -
berger, Cent. f. Bakt. I. O., vol. xxxvii., p. 131. Wassermann and Bruck,
Zeit. f. Hyg., vol. 1., p. 309. Weil-Halle and Lemaire, Comptes Rendus
Soc. Biol., 1906, p. 114. Henderson Smith, Journ. Hyg., vol. vii., p. 205
Goodman, Journ. Inf. Dis., vol. v., p. 184.

Susceptibility to Tetanus Toxins. — Knorr, Munch. Med. Woch., 1898,
pp. 321, 362. Behring, Fortschr. der Med., vol. xvii., p. 501. Behring's
Beitrage, August, 1903. See also Chapters V., XIV.

CHAPTER VII

Alexins.— Nuttall, Zeit. f. Hyg., vol. iv., 1888, p. 353. Behring, Cent,
f. Klin. Med., 1888, No. 32. Behring and Nissen, Zeit. f. Hyg.,
vol. viii., p. 412. Buchner, Cent. f. Bakt. I. O., vol. v., p. 817. Ibid.,
Arch. f. Hyg., vol. x., p. 84. Ibid., Arch. f. Hyg., vol. xvii., p. 112. Ibid.,
Munch. Med. Woch., vol. xlvii., p. 277. Lubarsch, Cent. f. Bakt., vol. vi.,
p. 481, 529. Pfeiffer, Zeit. f. Hyg., vol. xviii. ; Deut. Med. Woch., 1896,
pp. 97, 119. Bordet, Ann. Inst. Past., vol. ix., p. 462, and ibid., vol. xii.,
p. 688. Landsteiner, Cent. f. Bakt. I. O., vol. xxv., p. 546.

Ehrlich's Researches are given in his Collected Studies, the main chapters

BIBLIOGRAPHY 425

being I. -VIII., XVII., XIX., XXI., XXII., XXXII., XXXIII., and XL.

Marino, Ann. Inst. Past., vol. xvii., p. 321.

A full account of the subject of Hcemolysins, with an excellent biblio-
graphy, is given by Sachs in Kraus and Levaditi, vol. i.( and the work of
Muir and his school has just been published in collected form (The Oxford
Press, 1909). See also Flexner and Noguchi, Journ. Exp. Med., vol. vi.
Kyes, Berlin. Klin. Woch., 1902 (reprinted in Ehrlich's Studies). Kyes
and Sachs, Berlin. Klin. Woch., 1903, p. 21, 57, 82. Kyes, Berlin. Klin.
Woch, 1903, p. 956, 982. Bordet's Views. — Bordet, Ann. Inst. Past.,
vol. xiii., 1899, PP- 225> 273 > v°l- xiv., p. 257 ; 1906, p. 467. Muir and
Browning, Proc. Roy. Soc., vol. Ixxiv., p. 298 ; Journ. Path, and Bact.^.
vol. xiii., p. 76. Muir, Lancet, vol. clxv., p. 446, and B. M. J., Septem-
ber 10, 1904. Muir and Ferguson, Journ. Path, and Bact., 1906, p. 84.
Metchnikoff, L'Immunite, Chapters VII., VIII.

Bordet and Gengou' s Phenomenon (Fixation of Complement). — Bordet,
Ann. Inst. Past., vol. xv., p. 289. Bordet and Gengou, C. R. Acad. Sci.,
vol. cxxxvii., p. 351. Gengou, Berlin. Klin. Woch., 1906, p. 1532. Muir
and Martin, Journ. of Hyg., vol. vi., p. 265. Heller and Tomarkin, Deut.
Med. Woch., 1907, p. 795. Cruveilhier, Comptes Rendus Soc. Bio.,
vol. Ixii., p. 1027. Schutze, Berlin. Klin. Woch., 1907, p. 800. Seligmann,
ibid., 1907, p. 1013. Widal and le Sourd, Comptes Rendus de la Soc.
Biol., 1901, pp. 673, 841. Wassermann and Bruck, Deut. Med. Woch.,
1906, p. 449. Bruck, Deut. Med. Woch., 1906, June, p. 945. Also Camus
and Pagnier, Comptes Rendus Soc. Biol., 1901, July.

For References re Gengou' s Phenomena see also Chapter IX.

Deviation by Toxins, etc. — Armand-Delille, Comptes Rendus Soc. Biol.,
1908. Poyerski, ibid., 1908, p. 896. Weinberg and Parvu, ibid., Novem-
ber, 1908. Laubry and Parvu, Soc. Med. des Hop., December, 1908.

In Explanation of Complementoid. — Moreschi, Berlin. Klin. Woch.,
vol. xiii., September, 1905, p. 1181. Gay, Cent. f. Bakt., vol. xxxix., 1905,
pp. 172, 603 ; also vol. xl., p. 695. Pfeiffer and Friedberger, Deut.
Med. Woch., 1905, p. 6. Besredka, Ann. Inst. Past., 1905. Sachs, Deut.
Med. Woch., May, 1908. Pfeiffer and Friedberger, Deut. Med. Woch.,
1905, p. 1145. Bordet, Ann. Inst. Past., vol. xv., p. 289. Sachs, Cent. f.
Bakt. I. O., vol. xl., p. 125. Bordet, Berlin. Klin. Woch., 1906, p. 17.
Pfeiffer and Moreschi, Berlin. Klin. Woch., 1906, p. 33.

Deviation of Complement. — Loffler and Abel, Cent. f. Bakt. I. O.( vol. xix.,
p. 51. Pfeiffer, Zeit. f. Hyg., vol. xx., p. 198. Neisser and Wechsberg,
Munch. Med. Woch., 1901. Lipstein, Cent. f. Bakt. I. O., vol. xxxi., p. 460.
Morgenroth, ibid., vol. xxxv., p. 501. Myers and Stephens, Journ. Path.
Bact., vol. v. Kyes, Berlin. Klin. Woch., 1902, and Kyes and Sachs, ibid.,

1903 (both these are in Ehrlich's Collected Studies). Meakins, Johns
Hopkins Bull., 1907, p. 259.

Origin of Complement, Alexin, etc. — Hankin, Cent. f. Bakt. I. O., xii.
and xiv., p. 853. Denys and Havet, La Cellule, 1894, vol. x., p. 7.
Havet, ibid., 1894, vol. x. Denys, Cent. f. Bakt. I. O., vol. xvi., p. 781.
Buchner, Munch. Med. Woch., 1894, p. 589; Metchnikoff, L'lmmunite.
Bulloch, Trans. Path. Soc., 1901, p. 208; ibid., B. M. J., September 10,

1904 (with full bibliography). Longcope, Journ. of Hyg., vol. iii.,
p. 28. Guseff, quoted by " Petrie, loc. cit. Briscoe, Orth Festschrift,
1903. Levaditi, Ann. Inst. Past., xvii., p. 187. Korschun and Morgenroth,
Berlin. Klin. Woch., 1902. Marino, Comptes Rendus Soc. Biol., vol. lv.,
p. 689. Schattenfroh, Arch. f. Hyg., vol. xxxi., p. i ; ibid., xxvii.,
p. 234; ibid., Munch. Med. Woch., 1897, P- 4T4 '> ibid., 1898, p. 1109 ;
ibid., 1897, P- 4' ibid., Arch. f. Hyg., vol. xxxv., p. 135, Petrie,
Journ. Path. Bact., vol. ix., p. 130. Lastschenko, Munch. Med.
Woch., 1899, p. 475 ; ibid., Arch. f. Hyg., xxxvii., p. 290. Lambotte, Cent,
f. Bakt. I. O., vol. xxxiv., p. 453. Lambotte and Stienon, Cent. f. Bakt.
I. O., vol. xl., p. 224. Donath and Landsteiner, Zeit. f. Hyg., vol. xliii.,
p. 552. Lowit and Schwarz, Zeit. f. Heilk., vol. xxiv., pp. 205, 301.

426 BIBLIOGRAPHY

Gengou, Ann. Inst. Past., vol. xv., p. 68 ; ibid., vol. xv., p. 232. Falloise,
Comptes Rendus Soc. Biol., vol. Ivi., p. 324. Falloise, Bull, de 1'Acad.
Royale de Belgique, 1903. Levaditi, Ann. Inst. Past., 1901, vol. xv.,
p. 894, and vol. xvi., p. 233, 1902. Ainley Walker, Journ. Hyg.,
vol. iii., p. 52, and Cent. f. Bakt., vol. xxxiii., p. 297. See also Hahn,
Arch. f. Hyg., vol. xxv., p. 105. Wauters, Arch, de Med. Exp., vol. x.,
p. 751. Moxter, Deut. Med. Woch., 1899, p. 687. Tromsdorff,
Arch. f. Hyg., vol. xl., p. 382. Van de Velde, Cent. f. Bakt. I. O.,
vol. xxiii., p. 692. Bail, Hyg. Rundschau, vol. viii., p. 1066. Sweet,
Cent. f. Bakt. I. O., vol. xxxiii., p. 208. Malvoz, Ann. Inst. Past., vol. xvi.,
p. 623. Lazar, Wien. Klin. Woch., 1904, p. 439. Kanthack, vide
Chapter X. Steinhardt, Journ. Med. Research, vol. xix., p. 161.
Gousseff, abstract in Bull. Inst. Past., vol. i., p. 175. Longcope, Journ.
Hyg., vol. iii., p. 28.

Origin of Immune Bodies, etc. — Pfeiffer and Marx, Deut. Med. Woch.,

1898, p. 47 ; Deutsch. Ann. Inst. Past., vol. xiii., p. 689 ; and Cent. f.
Bakt. I. O., vol. xxviii., p. 45. Kraus and Schiffmann, Ann. Inst. Past.,
vol. xx., p. 226. Bulloch, vide ante. Wassermann, Deut. Med. Woch.,

1899, p. 141. Wassermann and Citron, Zeit. f. Hyg., vol. 1., p. 331.
Pfeiffer and Marx, Zeit. f. Hyg., vol. xxvii., p. 272. Donath and Land
steiner, Zeit. f. Hyg., vol. xliii., p. 552. Kraus and Schiffmann, Ann. Inst.
Past., vol. xx., p. 226. Kraus and Levaditi, Comptes Rendus Acad. Sci.,
vol. cxxxviii. Emden, Zeit. f. Hyg., vol. xxx., p. 19.

Methods (Haemolysis}. — Papers in Ehrlich's Collected Studies, especially
Chapter XXIX. (Morgenroth) . Sachs in Kraus and Levaditi's Handbuch
der Technik and Methodik der Immunitatsforschung (Fischer, Jena,
1907). Moro, Munch. Med. Woch., vol. liv., p. 1026. Gay and Ayer,
Journ. Med. Research, vol. xvii., p. 341. Longcope, Univ. of Penn. Med.
Bull., xv., p. 331.

Methods (Bacteriolysis}. — Ehrlich's Studies, Chapters IX., XXX. (Neisser
and Wechsberg). Klien, Johns Hopkins Bull., 1907, p. 245. Stern and
Korte, Berlin. Klin. Woch., 1904, p. 213. Wright, Lancet, December,
1900 ; ibid., 1901, March 2 and September 14 ; ibid., Proc. Roy. Soc.,
Ixxi., p. 54. Gay and Ayer, loc. cit. Andrewes and Gordon, Report of
L.G.B. (supplement), 1906-7, p. 141. Goodwin, Proc. N. Y. Path. Soc..
vol. v.

Cytolysins, etc. ; Leucolysins. — Metchnikoff, L'lmmunite. Funk, Cent,
f. Bakt. I. O., vol. xxvii., p. 670. Flexner, Univ. of Penns. Med. Bull.,
vol. xv. Bunting, ibid., vol. xvi., p. 200. Goodman, Journ. Inf. Dis.,
vol. v., p. 173. Christian, Deut. Arch. f. Klin. Med., vol. Ixxx., p. 333.

Spermotoxin. — Metchnikoff, Ann. Inst. Past., vol. xiv., p. i, 369.
Metalnikoff, ibid., p. 577. Moxter, Deut. Med. Woch, 1900, p. 61.
Landsteiner, Cent. f. Bakt. I. O., vol. xxv., p. 546. London, Arch, de
Sci. Biol. St. Petersburg, vol. ix. Weichardt, Ann. Inst. Past., vol. xv.,
p. 8^3.

3 ^'specificity and General. — Sachs, Biochem, Cent., 1903. Pearce, Journ.
/.^ Exp. Med., vol. viii. ; ibid., Journ. Med. Res., vol. xii., pp. i, 329. Beebe,
Journ. Exp. Med., vol. vii., p. 730. Armand-Delille and Leenhardt, C. R.
Soc. Biol., vol. Ixii., p. 31. Woltmann, Journ. Exp. Med., vol. vii., p. 119.
Forsner, Munch. Med.' Woch., vol. Hi., p. 892. Flexner and Noguchi,
Journ. Med. Res., vol. ix., p. 257. Bierry and Pettit, C. R. Soc. Biol.
vol. Ivi., p. 238. Dudgeon, Panton, and Ross, Proc. Roy. Soc. Med.,
vol. ii., No. 2.

Trichotoxin. — Von Dungern, Munch. Med. Woch., 1899. Hoyton,
B. M. J., 1902.

Nephrotoxin. — Nefedieff, Ann. Inst. Past., vol. xv., p. 17. Ascoli and
Figari, Berlin. Klin. Woch., 1902. Lindemann, Cent. f. Allg. Path.,
vol. vi., p. 184. Pearce, Univ. Penns. Med. Bull., vol. xvi., p. 217.
Bierry, C. R. Acad. Sci., vol. cxxxii. Bierry, C. R. Soc. Biol., vol.
lv., p. 496. Le Play and Corpechot, ibid., p. 206. Sheldon, Amos,

BIBLIOGRAPHY 427

Reports of Med. Staff, Egyptian San. Council, 1906. Albarran and
Bernard, Arch, de Med. Exp., vol. xv., p. 13. Woltmann, Journ. Exp.
Med., vol. vii., p. 119.

Gastrotoxin. — Bolton, Proc. Roy. Soc., vol. Ixxvii., p. 426, and Ixxix.,
p. 533 ; ibid., Proc. Roy. Soc. Med., vol. ii., No. 2. Theobary and Bates,
Comptes Rendus Soc. Biol., 1903, p. 459.

Anti-intestinal Serum. — Belonowski, Comptes Rendus Soc. Biol., 1907,
P- 9-

Syncytiolysin. — Liepmann, Deut. Med. Woch., 1902, p. 911. Weichardt,
ibid., 1902, p. 624. Ascoli, Cent. f. Gynekol., 1902. Wormser, Munch.
Med. Woch., 1904, p. 7.

Neurotoxin. — Delezenne, Ann. Inst. Past., vol. xiv., p. 686 ; ibid., Comptes
Rendus Soc. Biol., 1901, p. 1161. Armand-Delille, Ann. Inst. Past.,
vol. xx., p. 838 ; ibid., Enriquer and Sicard, Comptes Rendus Soc. Biol.,

1900. Pirone, Arch. Sci. Biol., vol. x., p. 75.

For Peripheral Nerves. — Schmidt, Ann. Inst. Past., vol. xx., p. 601.

Ophthalmotoxin. — Bram Pusey, quoted by Ricketts. Le Play and
Corpechot, Comptes Rendus Soc. Biol., 1904, p. 1021. Golovine, Russie
Vratch, 1904, abstracted in Bull. Inst. Pasteur, vol. ii., p. 1009.

Hepatotoxin. — Delezenne, Comptes Rendus Acad. Sciences, vol. cxxxi.,
p. 427. Pease and Pearce, Journ. Inf. Dis., vol. iii., p. 619. Bolton,
Proc. Roy. Soc., vol. Ixxiv., p. 135. Bierry and Mayer, Comptes Rendus
Soc. Biol., vol. Ivi., p. 1016.

Adrenotoxic Serum,. — Bigart and Bernard, Comptes Rendus Soc. Biol.,

1901, p. 161. Yates, Univ. Penns. Med. Bull., vol. xvi., p. 195.
Thyrotoxic Serum. — Gontscharnkow, Cent. f. Allg. Path., vol. lix., p. 76.

Portis, Journ. Inf. Dis., vol. i., p. 127.

CHAPTER VIII

Gruber and Durham, Munch. Med. Woch., 1896, p. 285 ; ibid., 1899,
p. 1829. Charrin and Roger, Comptes Rendus Soc. Biol., 1889, p. 667.
Metchnikoff, Ann. Inst. Past., vol. v., p. 473. Durham, Journ. Path.
Bact., vol. iv., p. 13, and vol. vii., p. 240. Grunbaum, Lancet, Septem-
ber 19, 1896; ibid., Munch. Med. Woch., 1897, No- T3-

Group Reactions. — Pfaundler, Munch. Med. Woch., 1899, November 15,
p. 472. Posselt and Sagasser, Wien. Klin. Woch., 1903, p. 691. Park,
Journ. Inf. Dis., 1906, February, p. i. Frouin, Comptes Rendus Soc.
Biol., vol. Ixii., p. 154. Crendiropoulo and Amos, Reports of Egyptian
Sanitary Council, 1906. Bordet, Ann. Inst. Past., vol. xiii., p. 225.

Bacterio-precipitins. — Kraus, Wien. Klin. Woch., 1897, August 12.
Norris, Journ. Inf. Dis., vol. i., p. 463. See Chapter IX.

Agglutination of Flagella. — Smith and Reagh, Journ. Med. Res., vol. x.,
p. 89. Buxton and Torrey, Journ. Med. Res., vol. xiv.

Theories as to the Mechanism of the Process. — Nicolle, Ann. Inst. Past.,
vol. xii., p. 161. Paltauf, Wien. Klin. Woch., 1897. Dineur, Bull. Acad.
Med. Belg., 1898, p. 652. Bordet, Ann. Inst. Past., vol. x., p. 195, and
vol. xiii., p. 225 (the latter especially). Lowit, Cent. f. Bakt. I. O., vol.
xxxiv., pp. 156, 251. Kraus and Joachim, ibid., vol. xxxvi., p. 662, and
xxxvii., p. 71.

Site of Origin of Agglutinin. — Pfeiffer and Marx, Deut. Med. Woch.,
1898, p. 47. Emden, Zeit. f. Hyg., vol. xxx. Wassermann, Deut. Med.
Woch., 1899, p. 141. Deutsch, Cent. f. Bakt., vol. xxviii., p, 45. Ruffer
and Crendiropoulo, vide ante.

Colloid Chemistry. — Biltz, Zeit. f. Phys. Chem., vol. xlviii., p. 615.
Neisser and Friedemann, Munch. Med. Woch., 1904, p. 827. Bechhokl,
Zeit. f. Phys. Chem., vol. xlviii., p. 385. See also Chapter XII.

Absorption Test. — Castellani, Zeit. f. Hyg., vol. xl., p. i.

Park, Journ. Med. Res., vol. vii. Hirschbruch, Arch. f. Hyg., vol. Ivi.,

428 BIBLIOGRAPHY

p. 280. Ballner, Arch. f. Hyg., vol. li., p. 245. Lowit, Cent. f. Bakt. I. O.
vol. xxxiv., pp. 156, 251.

Constitution of Agglutinins, A gglutinoids , etc. — Wassermann, Zeit. f.
Hyg., vol. xlii., p. 267. Buxton and Vaughan, Journ. Med. Res., vol. xii.,
p. 115. Eisenberg and Volk, Zeit. f. Hyg., vol. xl. Shibayama, Cent. f.
Bakt. I. O., vol. xlii., pp. 68, 144. Joos, Cent. f. Bakt. I. O., vol. xxxiii.,
p. 762 ; ibid., Zeit. f. Hyg., vol. xxxvi., p. 422. Scheller, Cent. f. Bakt. I. O.,
vol. xxxvi., p. 694. Smith and Reagh, Journ. Med. Res., vol. x., p. 89.
Buxton and Torrey, Journ. Med. Res., vol. xiv., April. Dreyer and
Jex-Blake, vide Dreyer, B. M. J., September 10, 1904, p. 564 ; Journ.
Path. Bact., vol. xi., p. i.

Modifications of Bacteria grown in Agglutinating Serum. — Ainley Walker,
Journ. Path. Bact., vol. viii., p. 34. Welch, Johns Hopkins Bull., vol.
xiii., p. 291. Muller, Munch. Med. Woch., 1903, p. 56. Bail, Arch. f.
Hyg., vol. xlii., p. 307. Landsteiner, Wien. Klin. Woch., 1897, P- 439-
Marshall and Knox, Journ. Med. Res., vol. xv., p. 325. See also
Chapter XIII.

Htzmagglutinins. — Landsteiner, Cent. f. Bakt. I. O., vol. xxvii., p. 357.
Landsteiner and Leiner, ibid., vol. xxxviii., p. 548. Hektoen, Journ.' Inf.
Dis., vol. iv., p. 297. Gay, Journ. Med. Res., vol. xvii., p. 321. Peskind,
Amer. Journ. Phys., 1903. Biffi, Ann. d'Ig. Sperim., vol. xiii., abstracted
in Bull. Inst. Past., vol. i., p. 526. Shattock, Journ. Path. Bact., vol. vi.,
p. 303. Ford and Halsey, Journ. Med. Res., vol. xi., p. 403. Eisenberg,
Wien. Klin. Woch., 1901, p. 1020. Griinbaum, B. M. J., 1900, p. 1089.

CHAPTER IX

Precipitins in Normal Sera. — Hoke, Wien. Klin. Woch., vol. xx., p. 347 ;
Rodet, Comptes Rendus de la Soc. Biol., vol. Iv., p. 1626. Noguchi, Bull.
Univ. Penns., vol. xv., p. 301. Ascoli, abstracted in Bull. Inst. Past.,
vol. i., p. 343.

Specificity of Serum Precipitins. — Vide Nuttall, loc. cit., in which the
main references are given. Uhlenhuth, Deut. Med. Woch., 1901, pp.- 82,
499. Wassermann and Schiitze, Berlin. Klin. Woch., 1901, p. 187 ; ibid.,
I9°3> P- J92. Ewing and Strauss, Proc. N. Y. Path. Soc., vol. ii., p. 152.

Ewing, ibid., vol. iii., p. 14. Deutsch, Cent. f. Bakt. I. O., vol. xxix.,

E. 661. Stern, Deut. Med. Woch., 1901, p. 135. Wassermann, Congr. f.
in. Med., 1900. Strube, Deut. Med. Woch., 1902, p. 425. Lenossier and

Lemoine, Sem. Med., 1901, No. 4. Stern, Deut. Med. Woch., 1901,

P- 135-

Precipitoids, etc. — Michaelis, Beit. z. Chem. Phys., vol. iv., p. 59. Ober-
mayer and Pick, Wien. Klin. Woch., 1903, No. 22, and 1904, p. 265.
Von Dungern, Cent. f. Bakt. I. O., vol. xxxiv., p. 355.

Kraus's Reaction. — Wien. Klin. Woch., 1897, P- 73^ ; ibid., 1901, p. 693.
Panichi, Cent. f. Bakt. I. O., vol. xliii., p. 188. Norris, Journ. Inf. Dis.,
vol. i., p. 463 (with bibliography). Hoke, Wien. Klin. Woch., vol. xx.,
p. 347. Eisler, Wien. Klin. Woch, vol. xx., p. 377. Dopter, Comptes
Rendus de la Soc. Biol., vol. lix., p. 69. Smith and Reagh, Journ. Med.
Res., vol. x., p. 89.

Serum Precipitins. — Tchistovitch, Ann. Inst. Past., vol. xiii., p. 406.
Bordet, ibid., p. 225. Myers, Cent. f. Bakt. I. O., vol. xxviii., p. 237.
Wassermann and Schutze, Berlin. Klin. Woch., 1901, p. 187. Nuttall,
Blood Immunity and Blood Relationship (Cambridge, 1904), in which
there is a full bibliography to the date of issue. Uhlenhuth, Deut. Med.
Woch., 1900, p. 734. Michaelis and Fleischmann, Zeit. f. Exp. Path,
and Ther., vol. i., p. 537. Von Dungern, Cent. f. Bakt. I. O., vol. xxxiv.,
p. 355. Obermayer and Pick, Wien. Klin. Woch., 1903, No. 22 ; ibid.,
I9°3. P- 265. Oppenheimer, Beit. z. Chem. Phys., vol. iv., p. 259.

BIBLIOGRAPHY 429

Precipitins for Crystalline Lens. — Uhlenhuth, Deut. Med. Woch., 1906,
p. 1244 ; also Koch's Festschrift.

Practical Application. — An excellent account of the technique is given
by Welsh and Chapman, Australian Medical Gazette, January 21, 1907-
See also Ewing, Clinical Pathology of the Blood, seond edition (Kimpton,
London). Graham-Smith and Sanger, Journ. Hyg., vol. iii., pp. 258, 354.
Buckmaster, Morphology of Blood (Murray, 1906). Bruck, Berlin. Klin.
Woch., 1907, pp. 793, 1510. Zebrowski, C. R. Soc. Biol., vol. Ixii., p. 603.
Uhlenhuth Deut. Med. Woch., 1906, p. 1244. Ziemke, Deut. Med. Woch.,
1 90 1, pp. 424, 731.

Deviation of Complement. — Neisser and Sachs, Berlin. Klin. Woch., 1905.
Uhlenhuth, Deut. Med. Woch., 1906, p. 1244. Muir and Martin, Journ. of
Hyg., 1906, July, p. 265. Friedberger, Deut. Med. Woch., 1906, p. 578.

Recognition of Foods. — Pniiger, Arch. f. Phys., 1906, pp. 465, 540.
Schmidt, Bioch. Zeit., vol. v., p. 422. Uhlenhuth, Deut. Med. Woch.,
1901, p. 780. Schutze, Zeit. f. Hyg., vol. xlvii., p. 144.

Recognition of Bones. — Schutze, Deut. Med. Woch., 1903, p. 62.

CHAPTER X

Metchnikoff's views and experiments are fully set forth in his " L'lm-
munite dans les Maladies Infecteuses " (English translation by Binnie,
Cambridge University Press, 1905), with numerous references, and his
" Comparative Pathology of Inflammation " (translated by F. A. and E. H.
Starling, Kegan Paul, Trench and Co., 1893). Buchner, vol. xvii., p. 138;
Marchand, Arch. Med. Exp., vol. x., p. 253 ; Massart, Ann. Inst. Past.,
vol. vi., p. 321 ; Petersson, Cent. f. Bakt. I. O., vol. xxxix., p. 423 ;
Savtschenko, Ann. Inst. Past., vol. xvi., p. 106 ; and numerous articles
from the French School published in the Annales de 1'Institute Pasteur,
Comptes Rendus de la Soc. Biol., etc. An excellent account of the main
phenomena is given in Adami's article on Inflammation in Clifford Allbutt's
" System of Medicine."

A bsorption of Tail of Tadpole. — Mercier, Arch. Zool. Exper. , vol. v. , p. 151.

Cells in Peritoneal Fluid. — Metchnikoff, loc. cit. Buxton and Torrey,
Journ. Med. Res., vol. xv., p. i. Kanthack and Hardy, Journ. Phys.,
vol. xvii., p. 81. Durham, Journ. Path, and Bact., vol. iv., p. 338.

Phagocytosis in the Lungs. — Briscoe, Journ. Path, and Bakt., 1907.
Baumgarten, Cent. f. Inn. Med., 1888, Zeigler's Beit., 1889, and Berlin.
Klin. Woch., 1884. Sanarelli, Cent. f. Bakt. I. O., vol. x., p. 514. Kant-
hack and Hardy, Phil. Trans., 1894, Journ. of Phys., 1894.

Enterokinase, etc. — Delezenne, vide Levaditi, L'Immunite.

Opsonins. — Sir Almroth Wright's researches have recently been pub-
lished in book-form (Studies in Immunization, Constable, 1909), to which
the reader is referred for a full account of the main researches on the
subject. See also the Practitioner, special number, May, 1908, and the

discussion on Phagocytosis, B. M. J., November 16, 1907. See also

Rimpau, Deut. Med. Woch., 1904, p.
B. M. J., 1902. Dean, Proc. Roy. Soc., 1905, vol. Ixxvi., p. 506, and

Neufeld and Rimpau, Deut. Med. Woch., 1904, p. 1458. Leishman,

May 30, 1907. Muir and Martin, B. M. J., 1907, p. 1783. Noguchi,
Journ. Exp. Med., vol. ix., p. 455. Rosenow, Journ. Inf. Dis., vol. iv.,
p. 285. Gruber and Futaki, Munch. Med. Woch., vol. liii., p. 249. Hektoen
and Ruediger, Journ. Inf. Dis., vol. ii., p. 128. Bulloch and Atkin, Proc.
Roy. Soc., vol. Ixxiv., p. 379. Neufeld, Arb. der Kais. Gesundh., vol. xxv.,
p. 164, and Berlin. Klin. Woch., 1908, p. 993. Lohlein, Ann. Inst. Past.,
vol. xix., p. 647, and vol. xxx., p. 939. Weil, Cent. f. Bakt. (Ref.), 1908,

P- 337-

Technique of Opsonin Estimations, etc. — Leishman, B. M. J., January n,
1902. Wright and Douglas, Proc. Roy. Soc., vols. Ixxii., Ixxxiii. Fleming,

430 BIBLIOGRAPHY

Practitioner, May, 1908. Walker, R. E., Journ. Med. Res., vol. xix.,
p. 237. Klien, Bull. Johns Hopkins Hosp., 1907, p. 245. Simon, Journ.
Amer. Med. Assoc., 1907, p. 139. Hektoen, Journ. Inf. Dis., vol. iii.,
p. 434. Veitch, Journ. Path, and Bact., January, 1908. Brown, Journ.
Amer. Med. Assoc., 1908. Morland, Inaugural Dissertation (Bern, 1908).
Emery, Clinical Pathology and Bacteriology, third edition (H. K. Lewis,
1908).

Opsonic Index in Health. — Bulloch, Trans. Path. Soc., vol. Ivi. Fleming,
Practitioner, May, 1908. Hollister, quoted by Bergey, Monthly Cyclop,
of Prac. Med., August, 1907. Urwick, B. M. J., 1905, July 22. Frazer,
Glas. Med. Journ., March, April, etc.

Opsonic Indices in Diseases. — See under the appropriate headings below.
Accuracy of Opsonic Determinations. — Greenwood, Proc. Roy. Soc. Med.,
vol. ii., No. 5, where a full bibliography is given.

Nature of Opsonins. — Crofton, Journ. Hyg., vol. v., p. 949. Chapin and
Cowie, Journ. Med. Res., vol. xvii., p. 213. Dean, Proc. Roy. Soc., 1907,
p. 399. Levaditi and Inman, Arb. Kais. Gesund., vol. xxv., p. 164.
Ledingham, Proc. Roy. Soc., 1907. McFarlane, Journ. Amer. Med. Assoc.,
vol. xlix., p. 1178. Noguchi, Journ. Exp. Med., vol. ix., p. 455. Simon,
Journ. Exp. Med., vol. ix., p. 487. Eggers, Journ. Inf. Dis., vol. v., p. 268.
Graham, ibid., p. 273. Bohme, Munch. Med. Woch., 1908, p. 1475.
Neufeld and Bickel, Ar.b. Kais. Gesund., vol. xxvii., p. 310. Levaditi and
Inman, C. R. Soc. Biol., vol. Ixii., p. 683. Eggers, Journ. Inf. Dis., vol. v.,
p. 263. Hektoen and Ruediger, Journ. Inf. Dis., vol. ii., p. 128. Hektoen,
Journ. Inf. Dis., vol. iii., p. 434. Browning, Journ. Med. Res., vol. xix., p. 201.
Specificity of Opsonins. — Bulloch and Western, Proc. Roy. Soc., vol.
Ixxvi. Simon, Journ. Exp. Med., 1906, p. 651. Muir and Martin (W. B. M.),
B. M. J., 1906, vol. ii., p. 1783. Potter, Ditman, and Bradley, Journ.
Amer. Med. Assoc., vol. xlvii., p. 1793. Russell, Bull. Johns Hopkins
Hosp., 1907, p. 252. Hektoen, Journ. Inf. Dis., vol. v., p. 249. McFarland
and L'Engle, Journ. Amer. Med. Assoc., vol. xlix., p. 1178.

Thermolability of Opsonins. — Wright and Douglas, Proc. Roy. Soc.,
vol. Ixxii. Wright and Reid, ibid., vol. Ixxvii. Macdonald, Studies in
Path. Aberd. Uni., 1906. Rosenow, Journ. Inf. Dis., vol. iii., p. 683.
Muir and Martin, B. M. J., 1906, vol. ii., p. 1783 ; and Proc. Roy. Soc.,
vol. Ixxix., p. 187. Neufeld and Hime, Arb. Kais. Gesund., vol. xxv.,
p. 164. Dean, B. M. J., Nov. 16, 1907 (with an excellent general account
of the subject to date). See also under Nature of Opsonins.

Influence of Temperature. — Bulloch and Atkins, Proc. Roy. Soc., vols.
Ixxii. and Ixxiii. Ledingham, ibid., 1908.

Influence of Source of Leucocytes. — Wright and Douglas, Proc. Roy. Soc.,
vol. Ixxiv. Bulloch and Ledingham, Studies in Path. Univ. Aberdeen,
1906. Fleming, Practitioner, May, 1908. Rosenow, Journ. Inf. Dis.,
vol. iii., p. 683. Lowenstein, Zeit. f. Hyg., vol. lv., p. 429. Bassett-
Smith, Journ. Hyg., 1907, p. 115. Shattock and Dudgeon, Proc. Roy.
Soc. Med., vol. i., No. 6.

Virulence.— See Chapter XIII.

Influence of Salts, etc. — Wright and Reid, Proc. Roy. Soc., vol. Ixxvii.
Hamburger and Hekma, Biochem. Zeit., vol. ix., pp. 275, 512. Sellards,
Journ. Inf. Dis., 1908, June. Noguchi, Journ. Exp. Med., vol. ix., p. 455.
Influence of Dilution of Serum. — Wright and Douglas, Proc. Roy. Soc.,
vol. Ixxii. Emery, Trans. Med. Chi. Soc., vol. Ixxxix. Marshall, Journ.
Path. Bact., 1908, p. 378.

Influence of Thickness of Bacterial Emulsion. — Tunnicliffe, Journ. Inf.
Dis., 1908, January. Walker, Journ. Med. Res., vol. xvi., p. 521.

Hcemopsonins. — Neufeld and Bickel, Arb. Kais. Gesund., vol. xxvii.,
p. 310. Neufeld and Topfer, Cent. f. Bakt. I. O., vol. xxxviii., p. 456.
Barratt, Wakelin, .Proc. Roy. Soc., 1905, p. 524. Keith, Proc. Roy.
Soc., 1906.

Aggressins. — Bail, O., Wien. Klin. Woch., vol. xvii., p. 846; ibid.,

BIBLIOGRAPHY 43!

vol.
Woch

xviii., p. 428. Munch. Med. Woch., 1905, pp. 1212, 1865 ; Deut. Med.
h., 1905, p. 1788. Bail and Weil, Cent. f. Bakt. I. O., vol. xl., p. 371.
sermann and Citron, Cent. f. Bakt. I. O., vol. xliii., p. 373 ; and Deut.

Wassermann and Citron, Uent. l. .tsakt. l. u., vol. xlni., p. 373 ;

Hyg., vol. liii., p.
Cent. f. Bakt, vol. xli., p. 230. Weil, Deut. Med. Woch., 1906, p. 382 ;

Klin. Woch., 1905, p. 1102. Citron, Zeit. f. Hyg., vol. liii., p. 515 ; ibid.,

ibid., Wien. Klin. Woch., 1905, p. 406; ibid., Arch. f. Hyg., vol. liv., p. 297 ;
and Berlin. Klin. Woch., 1905, p. 430. Salus, Arch. f. Hyg., vol. lv.,
P- 335 >' ibid., Wien. Klin. Woch., vol. xviii., p. 660. Especially Lancet,
August 17 and 24, 1907 (Collected Studies, p. 317).

Vaccine Treatment. — Wright's Collected Studies. Especially Lancet,
August 17 and 24, 1907 (Collected Studies, p. 327). Practitioner, May,
1908. Allen's Vaccine-Therapy (H. K. Lewis, 1908). Pfeiffer and Fried-
berger, Cent. f. Bakt. I. O., vol. xlvii., p. 503. See also under the separate
headings.

CHAPTER XI

Tuberculin Reaction. — Koch, Deut. Med. Woch., 1890 and 1891. Wasser-
mann and Bruck, Deut. Med. Woch., 1906, p. 449 (in which there is a good
account of the earlier theories). (See also Chapter XIV.)

Modifications of Tuberculin Reaction. — See under Tubercle.

Mallein Reaction. — Vide Jowett's Blood-Serum Therapy, p. 156. Kraus
and Levaditi, vol. i.

Reactions in Gonococcal Infections. — Irons, Arch. Int. Med., vol. i., p. 433.
Difference in Reactions between Healthy and Infected Persons. — Lawson and
Stewart, Proc. Med. Chi. Soc., 1905. See also Allen's Vaccine Therapy.

Anaphylaxis to Toxins. — Richet, Comptes Rendus Soc. Biol., vol. Iviii.,
p. 109 ; Ann. Inst. Past., vol. xxi., p. 497 ; and Comptes Rendus Soc. Biol.,
vol. Ixii., pp. 358, 643. Goodman, Journ. Inf. Dis., vol. iv., p. 509.

Hyper sensitiveness to Serum. — Arthus, Comptes Rendus Soc. Biol.,
vol. lv., p. 817. Nicolle, Ann. Inst. Past., vol. xxi., p. 128. Remlinger,
Comptes Rendus Soc. Biol., vol. Ixii., p. 23.

Theobald Smith's Phenomenon. — Rosenau and Anderson, Journ. Med.
Res., vol. xv., p. 179 ; ibid., vol. xvi., p. 381 ; and Journ. Amer. Med.
Assoc., 1906, p. 1007. Besredka and Steinhardt, Ann. Inst. Past., vol. xxi.,
p. 117. Besredka, Comptes Rendus Soc. Biol., vol. Ixii., p. 477; ibid.,
vol. Ixiii., p. 294 ; ibid., Ann. Inst. Past., vol. xxi., p. 950 ; and Bull.
Inst. Past., vol. vi., p. 841. Gay and Southard, Journ. Med. Res., vol. xv.,
p. 143. Vaughan and Wheeler, Journ. Inf. Dis., 1907, p. 476. Otto,
Munch. Med. Woch., 1907. Doerr, Wien. Klin. Woch., 1908. Gay and
Southard, Journ. Med. Res., vol. xviii., p. 407. Weil-Halle and Lemaire,
Comptes Rendus Soc. Biol., vol. Ixiii., p. 748. Lewis, Journ. Exp. Med.,
vol. x.

Serum Disease. — Von Pirquet and Schick, Die Serum-Krankheit (Leipzic
and Wien, 1905). Currie, Journ. Hyg., vol. vii., p. 35. Goodall, Journ.

Hyg., vol. vii. Hamburger and Moro, Wien. Klin. Woch., vol. xvi., p. 445.

vii., p. 807, and xx., p. 817. Wic
and Rostane, Bull. Soc. Med. des Hop. de Paris, 1905, p. 424. Marfan

Hamburger and Dehne, ibid., vol. xvii., p. 807, and xx., p. 817. Widal

and Le Play, ibid., p. 274. Netter, Comptes Rendus Soc. Biol., vol. lx.,
p. 279. Park and Throne, Trans. Assoc. Amer. Phys., vol. xxi., p. 259.
Saunders, Interstate Med. Journ., 1908, p. 576.

CHAPTER XII

A good general outline of the subject may be found in Pauli's " Physical
Chemistry in the Service of Medicine," 1907, translated by Fischer (Chap-
man and Hall). See also Findlay's " Physical Chemistry in Medical and
Biological Science " (Longmans, Green and Co., 1905).

Biltz, Zeit. f. Phys. Chem., vol. xlviii., p. 615. Biltz and Siebert, Beitr.
z. Exp. Therap., 1905, p. 30. Field and Teague, Journ. Exp. Med., vol.

432 BIBLIOGRAPHY

viii., p. 222 ; and vol. ix., p. 86. Teague and Buxton, Journ. Exp. Med.,
vol. ix., p. 254. Craw, Proc. Roy. Soc., vol. Ixxvi., p. 179 ; and vol. Ixxvii.,
p. 311, and other articles. Bordet, Ann. Inst. Past., vol. xvii., p. 161.
Nernst, Zeit. f. Electrochemie, vol. x., p. 377. Girard-Mangin and Henri,
Comptes Rendus Soc. Biol., vol. Ivi., p. 541, and numerous other articles
in the same periodical and in Comptes Rendus Acad. Sci. Landsteiner
and Stancovic, Cent. f. Bakt. I. O., vol. xli., p. 108. Landsteiner and
Urlirz, Cent. f. Bakt. I. O., vol. xl., p. 265. Flexner and Noguchi, Journ.
Exp. Med., 1906, p. 547. Bechhold, Zeit. f. Phys. Chem., vol. xlviii.,
p. 385. Neisser, Cent. f. Bakt. I. O., vol. xxxvi., p. 671. Neisser and
Friedemann, Munch. Med. Woch., 1904, p. 465. Michaelis and Fleisch-
mann, Zeit. f. Exp. Path. u. Ther., vol. i., p. 547. Gengou, Ann. Inst.
Past., vol. xviii., p. 678. Dreyer, B. M. J., September 10, 1904.

Danysz Effect. — Danysz, Ann. Inst. Past., vol. xvi., p. 331. Jacoby,
Hoffm. Beit., vol. iv., p. 212. Sachs, Cent. f. Bakt. I. O., vol. xxxvii.,
p. 251. Craw, Proc. Roy. Soc., 1905.

Precipitation of Colloids. — Spiro, Beit. z. Chem. Phys., vol. iv., p. 300.
Perrin, Comptes Rendus Acad. Sci., vol. cxxxvi., p. 564.

Hcemolysis by Silicic Acid. — Landsteiner and Jagic, Wien. Klin. Woch.,
vol. xvii., p. 63.

CHAPTER XIII

Phagocytosis in Peritoneum. — Buxton and Torrey, Journ. Med. Res.,
vol. xv., p. 5. Petterson, Cent. f. Bakt. I. O., vol. xl., p. 537. Weil, ibid.,
vol. xliii., p. 190, and vol. xliv., p. 164 ; and Arch. f. Hyg., vol. Ixi., p. 293 ;
Journ. Inf. Dis., vol. iv., p. 582. Metchnikoff, L'lmrrhinite. Pierallini, Ann.
Inst. Past., vol. xi., p. 308. Wolff, Berlin. Klin. Woch., 1903, Nos. 17-20.

Bacterial Immunity in General. — Metchnikoff, L'Immunite, especially
chapters vi. to x. Sauerbeck, Die Krise in der Immunitatsforschung,
Folia Serologica, vol. ii., p. i, with full bibliography. Hahn, Kolle, and
Wassermann's Handbuch, Fasc. xviii. and xix. Cole, Rufus, Zeit. f.
Hyg., vol. xlvi., p. 371. Kisskalt, Zeit. f. Hyg., vol. xlv., p. i. Hoke, Zeit.
f. Hyg., vol. xxv., p. 197. Bail, Arch. f. Hyg., vol. Hi., p. 272. Neufeld,
Arb. a. d. Kais. Gesundh., vol. xxviii., p. 125. W'erigo, Ann. Inst. Past.,
vol. viii. Bail, Arch. f. Hyg., vol. liii., p. 272. Hoke, Cent. f. Bakt. I. O.,
vol. xxxiv., p. 693. Sir Watson Cheyne, Lancet, June 27, 1908.

In Tick Fever. — Levaditi and Manouelian, Comptes Rendus Soc. Biol.,
vol. Ixi., p. 566, and vol. Ixii., pp. 619, 815.

Virulence. — Walker, Ainley, Cent. f. Bakt. I. O., vol. xxxiii., p. 297.
Shaw, B. M. J., 1903, May 9, p. 1074. Cohn, Zeit. f. Hyg., vol. xlv., p. 61.
Pfeiffer, Koch's Festschrift, 1903. Stiirtz, Zeit. f. Klin. Med., vol. Hi.,
p. 422. Bail, Wien. Klin. Woch., vol. xvii., p. 846. Petterson, Cent. f.
Bakt. I. O., vol. xxxviii., p. 73. Steinhardt, Proc. N.Y. Path. Soc., vol. iv.
Day, Journ. Inf. Dis., 1905, p. 569. Marshall and Knox, Journ. Med.
Res., vol. xv., p. 325. Friedberger, Cent. f. Bakt. I. O., vol. xliv., p. 32.
Rosenow, Journ. Inf. Dis., vol. iv., p. 285.

Formation of Envelope, etc. — Metchnikoff, L'Immunite, chapter i.
Danysz, Ann. Inst. Past., vol. xiv., p. 641. Bordet, ibid., vol. xi., p. 177.
Gruber and Futaki, Munch. Med. Woch., 1906, p. 249. Preis, Cent. f.
Bakt. I. O., vol. xliv., p. 209. Bail, Wien. Klin. Woch., vol. xix., p. 1278.
Bail and Rubritius, Cent, f. Bakt. I. O., vol. xliii., p. 641. Stienon,
Comptes Rendus Soc* Biol., vol. xii., pp. 604, 841.

CHAPTER XIV

Staphylococci ; Staphylolysin. — Van de Velde, Ann. Inst. Past., vol. xv.,
p. 580. Kraus and Clairmont, Wien. Klin. Woch., 1900. Neisser and
Wechsberg, Zeit. f. Hyg., vol. xxxvi., p. 299.

Leucocidine. — Van de Velde, loc. cit. Bail, Arch. f. Hyg., vol. xxxii.,
p. 133. Neisser and Wechsberg, Munch. Med. Woch., 1902, p. 1261.

BIBLIOGRAPHY 433

Immunity. — Nuttall, Zeit. f. Hyg., vol. iv., p. 353. Wright and Windsor,
Journ. of Hyg., vol. ii., p. 397. Andrewes and Gordon, Suppl. Report
Med. Officer L.G.B., 1906, p. 141. Wright and Douglas, Proc. Roy. Soc.,
vol. Ixxxii., and other articles in Wright's Collected Studies.

Vaccine Treatment, Opsonins, etc. — Wright, Lancet, March 29, 1902 ;
B. M. J., May 7, 1904, etc. Allen's Vaccine Therapy. Chapman and
Cowie, Journ. Med. Res., vol. xvii., p. i.

Streptococcic Infections; Streptocolysin. — Besredka, Ann. Inst. Past.,
vol. x., p. 880. Casagrandi, quoted by Oppenheimer.

Toxins. — Parascandalo, Wien. Klin., Woch. 1897, P- 86 1. Marmorek,
Ann. Inst. Past., vol. ix., p. 593. Roger, Comptes Rendus Soc. Biol.,
vol. xliii., p. 538. Schenk, Wien. Klin. Woch., 1897, P- 937- Breton,
Comptes Rendus Soc. Biol., vol. Iv., p. 886. Simon, Cent. f. Bakt. I. O.,
1903, pp. 308, 440. Schlesinger, Zeit. f. Hyg., vol. xliv., p. 428.

Serum Treatment. — Marmorek, Ann. Inst. Past., vol. ix., p. 593 ; and
Berlin. Klin. Woch., 1902, No. 14. Besredka, Ann. Inst. Past., vol. xviii.,

p. 363. Aronson, Deut. Med. Woch., 1903, p. 439. Tavel, Cent. f. Bakt.
I. O., vol. xxxiii., p. 212, and vol. xxxv., p. 513. Neufeld, Zeit. f. Hyg.,
vol. xliv., p. 161. Simon, Cent. f. Bakt. I. O., vol. xxxv., pp. 308, 440.

Bordet, Ann. Inst. Past., 1897, p. 177. Wright, Clin. Journ., 1906, p. 78.
Neufeld, Zeit. f. Hyg., vol. xliv., p. 161. Sommerfeld, Cent. f. Bakt. I. O.,
vol. xxxiii., p. 722.

Vaccine Treatment. — Wright, Practitioner, May, 1908 ; and Lancet,
August 24, 1907. Douglas, Lancet, February 23, 1907. Crowe and
Wynn, B. M. J., August 8, 1908, p. 303. Sutcliffe and Bayley, Lancet,
August 10, 1907. Tunnicliffe, Journ. Inf. Dis., vol. v., p. 268. Banks,
Journ. Path. Bact., 1908, p. 113.

Pneumococcic Infections: Toxin. — Klemperer, Berlin. Klin. Woch, 1891,
and Zeit. f. Klin. Med., vol. xx., p. 165. Washbourn, Journ. Path. Bact.,
vol. iii., p. 214. Isaeff, Ann. Inst. Past., vol. vii., p. 259. Casagrandi,
quoted by Oppenheimer. Mennes, Zeit. f. Hyg., vol. xxv., p. 413. Carnot
and Fournier, Arch. Med. Exp., 1900, p. 357.

Serum Treatment. — Washbourn, B. M. J., February 27, 1897, p. 510 ;
and with Eyre, ibid., 1899, p. 1247 ; and Journ. Path, and Bact., vol. v.,
p. 13. Eyre, vide infra. Pane, Cent. f. Bakt. I. O., vol. xxi., p. 664.
Knauth, Deut. Med. Woch., 1905, p. 452. Castresana, Rev. de Ther.,
1905, No. 1 8. Tyler, Journ. Amer. Med. Assoc., 1901, p. 1540. Mennes,
vide supra.

Vaccine Therapy, Opsonins, etc. — MacDonald, Path. Studies, Univer.
Aberdeen. Eyre, Lancet, February 22, 1908. Neufeld and Rimpau, Zeit.
f. Hyg., vol. li., p. 283. Graham, Journ. Inf. Dis., vol. v., p. 273. Butler
Harris, Practitioner, May, 1908. Briscoe and Williams, ibid.

Gonococcic Infections; Toxin. — Wassermann, Berlin. Klin. Woch.,
1897, P- 685 ; and Zeit. f. Hyg., vol. xxvii., p. 298. Christmas, Ann.
Inst. Past., vol. xi., p. 609. Nicolaysen, Cent. f. Bakt. I. O., vol. xxii.,

P- 305-

Serum Diagnosis, Immunity, etc. — Torrey, Journ. Med. Res., vol.- xvii.,
p. 347, and vol. xix., p. 471. Teague and Torrey, ibid., vol. xvii., p. 223.
Meakins, Johns Hopkins Hosp. Bull., 1907, p. 255. Ricketts, Infection
and Immunity. Bruckner and Christeanu, Comptes Rendus Soc. Biol.,
vol. Ix., May, June. Miiller and Oppenheim, Wien. Klin. Woch., vol. xix.,
p. 894. Bruck, Deut. Med. Woch., 1906, p. 1368. Vannod, ibid., 1906,
p. 1984. Rogers, Cent. f. Bakt. I. O., vol. xxxix., p. 279.

Vaccine Treatment, Opsonins, etc. — Wright, Lancet, August 17 and 24,
1907. Allen, Vaccine Therapy. Rons, Arch. Int. Med., vol. i., p. 433.
Cole and Meakins, Bull. Johns Hopkins Hosp., 1907, p. 223. Butler and
Long, Journ. Amer. Med. Assoc., 1908, p. 744.

Meningococcic Infections : Toxins, Immunity. — Lepierre, Journ. Phys.
et Path. Gen., vol. v., p. 547. Houston and Rankin, B. M. J., Novem-
ber 16, 1907. Davis, Journ. Inf. Dis., vol. ii.

28

434 BIBLIOGRAPHY

Agglutination. — Kutscher, Deut. Med. Woch., 1906, p. 1849. Alice
Taylor, Lancet, July 6, 1907.

Serum Treatment. — Kolle and Wassermann, Deut. Med. Woch., 1906,
p. 609. Ruppel, ib'id., 1906, p. 1366. Markl, Cent. f. Bakt., vol. xliii.,
p. 95. Levy, Deut. Med. Woch., 1908, p. 139. Emmett Holt, B. M. J.,
October 31, 1908. Flexner and Jobling, Journ. Exp. Med., 1908, pp. 141,
690. Jochman, Deut. Med. Woch, 1906, p. 788. Meyer and Ruppel,
Mediz. Klin., 1907, No. 4, and Cent. f. Bakt. I. O., 1907. Wassermann,
Deut. Med. Woch., 1907, p. 1585.

Vaccine Therapy, Opsonins, etc. — McKenzie and Martin, ibid., October 31,
1908, and Journ. Bact., 1908, vol. xii., p. 539. Davis, Journ. Inf. Dis.,
vol. ii., and vol. iv., p. 538. Houston, B. M. J., November 16, 1907.
Mackenzie, ibid., June 15, 1907.

Malta Fever. — Wright and Smith, Lancet, March 6, 1897. Birt and
Lamb, Lancet, September 9, 1899. Eyre, J. W. H., and Shaw, H. E. A.,
Report of Royal Society's Comm. on Med. Fever, part v. Bassett-Smith,
Journ. Trop. Med. and Hyg., 1907, and Journ. Hyg., vol. vii., p. 115.
Eyre, Lancet, 1908, June 13, 20, and 27.

Tubercle ; Tuberculin Reaction. — Koch, Deut. Med. Woch., 1890, p. 1028,
and 1891, pp. 101, 1188. (See also 1890, p. 1053 et seq.) Babes, Zeit. f,
Hyg., vol. xxxii. Marmorek, Comptes Rendus Soc. Biol., 1903, p. 1650.
Ehrlich, Inter. Kong. f. Hyg., 1900. Trudeau, Baldwin, and Kinghorn,
Journ. Med. Res., vol. xii., p. 169. Weil and Nakajama, Munch. Med.
Woch., 1906, p. 1001. Cohn, Berlin. Klin. Woch., 1908, p. 1309. Richet,
Comptes Rendus Soc. Biol., 1905, p. 109. Citron, Berlin. Klin. Woch.,
I9°7> P- TI35- Marmorek, Lancet, December 12, 1903 (in diagnosis
especially). Arloing, Journ. de Phys. et Path. General, 1903, p. 677.
V. Bergmann, Deut. Med. Woch., 1890, p. 1073, and Munch. Med. Woch.,
p. 824. Beck, Arch. f. Kinderheilkunde, 1903, p. i. Lowenstein, Kraus,
and Levaditi, vol. i., p. 1019 (with full bibliography). Lingelsheim,
Deut. Med. Woch., 1898, p. 583. Armand-Delille, These de Paris, 1903,
abs. in Bull. Inst. Past., vol. ii., p. 73. For an account of the main forms
of the tuberculins, see Allen's Vaccine Therapy, and Gamble, Pharma-
ceutical Journal, February 16, 1909.

Tuberculinin Treatment. — Koch, Deut. Med. Woch., 1891, No. 3. Denys,
Comptes Rendus Congr. Tuberc., 1898, p. 497. Petruschky, Berlin. Klin.
Woch., 1899, pp. 1 1 20, 1141. Moller and Kayserling, Zeit. f. Tuberkulose,

1902, p. 4. Bandelier, Deut. Med. Woch., 1898, p. 798 ; ibid., Zeit. f. Hyg.,
vol. xliii., p. 335. Pardoe, Lancet, December 16, 1905. Spengler, Deut.
Med. Woch., 1897, No. 36. Lowenstein and Rappoport, Zeit. f. Tuber-
kulose, vol. v., p. 6. Stone and Miller, Medical Record, March 28, 1908.
Hamburger, Munch. Med. Woch., vol. Iv., p. 1741.

Serum Treatment. — Maragliano, Berlin. Klin. Woch., 1903, pp. 563, 593.
Marmorek, Berlin. Klin. Woch, 1903, p. 1108.

Tuberculin in Immunization of Animals. — Macfadyen, Journ. Comp. Path,
and Therap., 1901, p. 136; 1902, p. 60. Behring, Berlin. Klin. Woch.,

1903, p. 233, and Deut. Med. Woch., 1903, p. 689. Behring, Romer, and
Ruppel, Beitr. zur Exp. Therap., vol. v. Pearson and Gilliland, Univ.
Penns. Med. Bull., vol. xviii., No. 2. Neufeld, Deut. Med. Woch., 1904.
Baumgarten, Berlin. Klin. Woch., 1905, No. 3. Vallee and Rossignol,
Bull. Soc. Med. Vet. Pratique, 1906, p. 39.

Cuti-Re action. — V. Pirquet, Klin. Studien iiber Vaccination and Vac-
cinale Allergic (Deut. Wien., 1907) ; Berlin. Klin. Woch., 1907. Vallee,
Comptes Rendus Acad. des Science, 1907, No. 22. Ferrand and Lemaire,
La Presse Medicale, 1907, p. 617. Dufour, Bull. Soc. Med. Hop. de Paris,
1907. Engel and Bauer, Berlin. Klin. Woch., 1907, p. 1169. Lignieres,
Bull. Soc. Cent. Med. Vet., 1907, p. 517. Wolff-Eisner and Teichman,
Berlin. Klin. Woch., 1908, p. 65.

Ophthalmo-Reaction. — Wolff-Eisner, Berlin. Klin. Woch., 1907. Cal-
mette, Comptes Rendus Acad. des Sciences, 1907. Vallee, ibid. Moro

BIBLIOGRAPHY 435

and Dagonoff, Wien. Klin. Woch., 1907, August. Calmette, La Clinique,
August, 1907. Chantemesse, Comptes Rendus Acad. de Med., July 20,
1907. Deut. Med. Woch., September 26, 1907.

Vaccine Treatment. — Vide numerous articles by Wright and his fellow-
workers (in his Collected Studies), especially Clinical Journal, Novem-
ber 9, 1904. Trans. Med. Chi. Soc., vol. Ixxxix., and the succeeding
articles in the discussion, Lancet, August 17 and 24, 1907. Reyn and
Peterson, Lancet, April 4, 1908. Latham, Spitta, and Inman, Proc. Roy.
Soc. Med., April, 1908. Torton, International Clinics (eighteenth series),
vol. ii., p. 23. Riviere, B. M. J., October 26, 1907. Whitfield, Practitioner,
May, 1908. Briscoe and Williams, ibid. Allen, Vaccine Therapy. Car-
malt Jones, Science Progress, April, 1909. Patterson, Lancet, Janu-
ary 25, 1908. Inman, ibid.

Typhoid Fevsr ; Toxin. — Chantemesse, Prog. Med., 1898, p. 245 ; Deut.
Med. Woch., 1907, p. 1572. Presse Med., 1904, p. 681. Macfadyen and
Rowland, Proc. Roy. Soc., vol. Ixxi., p. 77. Conradi, Deut. Med. Woch.,
1903, p. 26. Pfeiffer and Kolle, Zeit. f. Hyg., vol. xxi., p. 203. Besredka,
Ann. Inst. Past., vol. xx., pp. 149, 304. Neisser and Shiga, Deut. Med.
Woch., 1903, p. 6r.

Immunity, Bactericidal Power of Blood, Opsonins, etc. — Leishman, Jour.
R. A. M. C., 1907. Evans, Laming, Journ. Path. Bact., 1904, p. 42.
Shiga, Berlin. Klin. Woch., 1904, p. 79. Richardson, Journ. Med. Res.,
vol. xiii. Wright, Lancet, September 14, 1901. Harrison, Journ. R. A.
M. C., 1907, p. 472. Stern and Korte, Berlin. Klin. Woch., 1904. Klien,
Johns Hopkins Hosp. Bull., 1907, p. 245. Neufeld and Kuhn, Arb. a. d.
K. Gesundh., vol. xxv., p. 164.

Vaccine Treatment (Prophylactic). — Wright, Short Treatise on Anti-
Typhoid Inoculation (Constable, 1904) ; ibid., Lancet, September 6, 1902 ;
ibid., B. M. J., October 10, 1903. Luxmore, Journ. R. A. M. C., 1907. A
good account of the subject is by Netter, Bull. Inst. Past., vol. iv., pp. 873,
921, 969, and 1024. Shiga, Berlin. Klin. Woch., 1904, p. 78. Friedberger
and Moreschi, Deut. Med. Woch., 1906, p. 1986.

Curative. — Richardson, Boston Med. and Surg. Journ., vol. Ivii., p. 449.
Cholera : Toxin. — Wassermann, Zeit. f. Hyg., vol. xiv., p. 35. Westbrook
Ann. Inst. Past., vol. viii., p. 318. Pfeiffer, Zeit. f. Hyg., vol. xi., p. 373,
and vol. xvi., p. 268, and vol. xx., p. 198. Metchnikoff, Roux, and Tau-
relli Salimbeni, Ann. Inst. Past., vol. x., p. 257. Kraus, Wien. Klin.
Woch., vol. xix., p. 655. Brau and Demei, Ann. Inst. Past., vol. xx.,
p. 578. Macfadyen, Cent. f. Bakt., vol. xlii., p. 365.

Serum Treatment. — Kraus, Wien. Klin. Woch., 1909, No. 2. Macfadyen,
Lancet, August 25, 1906.

Vaccine Prophylaxis. — Haffkine, Bull. Inst. Past., vol. iv., pp. 697, 737.
Fischera, Cent. f. Bakt. I. O., vol. xli., pp. 576, 671, and 771 (with full
bibliography).

Plague : Prophylaxis. — Haffkine, B. M. J., June 12, 1897 '< ibid., B. M. J.,
September 24, 1898 ; ibid., Proc. Roy. Soc., 1899., vol. Ixv. ; ibid., Gov.
Central Press, 1900, 1903, 1904. Burch, N. Y. Med. Journ., September,
1902. Forsyth, Lancet, December 12, 1903. Lustig and Galeotti,
B. M. J., October 9 and November 27, 1897. Bannerman, Cent. f. Bakt.
I. O., vol. xxix., p. 857. Kolle and Otto, Deut. Med. Woch., 1904, p. 493.
Lustig and Galeotti, Deut. Med. Woch., 1897, P- 227-

Sero-Therapy. — Yersin, Ann. Inst. Past., vol. xi., p. 8-1. Metchnikoff,
ibid., vol. xi., p. 737. Zabolotny, ibid., vol. xiii., p. 833. Calmette and
Salimbeni, ibid., vol. xiii., p. 865. Dugardin-Beaumetz, Bull. Inst. Past.,
1906, p. 453. Choksy, Report on Treatment of Plague, Bombay, 1906,
and Lancet, 1900, p. 291. Clemow, Lancet, May 6, 1899, p. 1212. Cairns,
Lancet, 1903, May 9. Symmers, Cent. f. Bakt. I. O., vol. xxv., p. 460.
Markl, Zeit. f. Hyg., vol. xlii., p. 244.

Glanders : Immunity. — Nicolle, Ann. Inst. Past., vol. xx., pp. 625, 698,
and 801 (especially p. 828). Kleine, Zeit. f. Hyg., vol. xliv., p. 183.

28—2

436 BIBLIOGRAPHY

Mallein. — The directions given at the Royal Veterinary College, London,
are given in Hewlett's Serum-Therapy. See also Jowett's Blood -Serum
Therapy.

The only full account of the subject is in Kraus and Levaditi, vol. i.,
p. 1090 (Wladimoroff).

Agglutination. — Bonome, Cent. f. Bakt. I. O., vol. xxxviii., p. 60 1.
Heanly, Lancet, February^, 1904, p. 364. Feodorowsky, Bull. Inst.
Past., vol. ii., p. 127.

Dysentery : Toxin. — Rosenthal, Deut. Med. Woch., 1904, p. 235. Todd,
,/ Journ. of Hyg., vol. iv., p. 480. Conradi, Deut. Med. Woch., 1903, p. 26.
Ludke, Berlin. Klin. Woch., 1906, pp. 3, 54. Besredka, Ann. Inst. Past.,
vol. xx., p. 304. Neisser and Shiga, Deut. Med. Woch., 1903, p. 61.

Serum. — Kruse, Deut. Med. Woch., 1903, pp. 6, 49. Shiga, Cent. f.
Bakt. I. O., 1903, No. 7 ; ibid., Deut. Med. Woch., 1901, pp. 744,
765, and 783 ; ibid., Zeit. f. Hyg., vol. xli., p. 355 (in Ehrlich's Collected
Studies), and vol. lx., p. 75. Vallard and Dopter, Ann. Inst. Past., vol. xx.,
p. 321. Flexner, Bull. Johns Hopkins Hosp., vol. xi., p. 231. Besredka,
vide supra. Doerr in Kraus and Levaditi's Hardbuch (with bibliography).
Coyne and Auche, Comptes Rendus Soc. Biol., vol. Ixiv., p. 829. Ruffer
and Willmore, B. M. J., October 17, 1908, vol. ii., p. 1176. Heller, Cent,
f. Bakt. I. O., vol. xlii., p. 30.

Vaccine Treatment. — Shiga, Cent. f. Bakt. I. O., vol. xxxiv., p. 392.
Forster, Indian Med. Gaz., 1907, p. 201 (quoted by Allen). Newman,
Lancet, May 16, 1908, p. 1410. Kolle and Strong, Deut. Med. Woch.,
1906, p. 413.

Anthrax: Toxin. — Conradi, Zeit. f. Hyg., vol. xxxi., p. 287 (with full
bibliography to date).

Immunity, Serum Reactions, etc. — Sobernheim, Berlin. Klin. Woch.,
1897, p. 910 ; ibid., Zeit. f. Hyg., 1899, p. 891. Bail, Cent. f. Bakt. I. O.,
vol. xxvii., p. 10 ; ibid., vol. xxxiii., pp. 343, 610 ; vol. xxxvi., pp. 266,
287 ; vol. xxxvii., p. 270. Bail and Petterson, ibid., vol. xxxiii., p. 756,
and vol. xxxiv., pp. 450, 540. Gengou, Ann. Inst. Past., vol. xiii., p. 642.
Hektoen, Journ. Inf. Dis., vol. iii., p 103. Horton, ibid., vol. iii., p. no.
Ascoli, Zeit. f. Hyg., vol. Iv., p. 44. Bandi, Cent. f. Bakt., vol. xxxvii.,
p. 464. Gruber and Futaki, Deut. Med. Woch., 1906, p. 1589. Cler,
Arch. Sc. Med., vol. xxix., 1905.

Serum Treatment. — Legge, Lancet, March 25, 1905, in which a good
account of the subject and the more important references are given.

Diphtheria: Immunization to the Bacilli. — Bandi, Cent. f. Bakt. I. O.,
vol. xxxiii., p. 535. Rist, Comptes Rendus Soc. Biol., 1903, p. 978.
Lipstein, Cent. f. Bakt. I. O., vol. xxxiii., p. 305.

Opsonic Action. — Tunnicliffe, Journ. Inf. Dis., vol. v. Reque, ibid.,
vol. iii., p. 441.

No literature concerning the use of diphtheria antitoxin need be given.

Tetanus. — A full account of the toxin is given in Oppenheimer, with full
bibliography.

Action on the Nervous System. — Gumprecht, Deut. Med. Woch., 1894,
p. 546. Meyer and Ransom, Arch. Exp. Path., vol. xlix., p. 369. Marie
and Morax, Ann. Inst. Past., vol. xvi., p. 818, and vol. xvii., p. 335. Roux
and Borrel, ibid., vol. xii., p. 225. Vaillard and Vincent, ibid., vol. v.,
p. i. Marie, Bull. Inst. Past., vol. i., p. 633. Fletcher, Brain, 1903,

P- 383-

Immunity. — Vide Metchnikoff, L'Immunite, especially p. 179 (English
edition, p. 169) and p. 412 (p. 392). In the same work much information
will be found regarding the action of tetanus toxin on different animals.

Antitoxin. — Vide Hewlett's Serum-Therapy, where the process of manu-
facture is given.

Local Application of Antitoxin. — Calmette, Comptes Rendus Acad. Sci.,
vol. cxxxvi., p. 1150. £

Syphilis. — The literature of the serum diagnosis of syphilis has already

BIBLIOGRAPHY 437

assumed formidable proportions. Wassermann, Neisser, Bruck, and
Schucht, Zeit. f. Hyg., vol. lv., p. 451. Wassermann, Berl. Klin. Woch.,
I9O7. P- 1599- Wassermann and Plaut, Deut. Med. Woch., 1906, p. 1769.
Wassermann and Meier, ibid., 1907, p. 1287. Neisser, Bruck, and Schucht,
Deut. Med. Woch., 1906, p. 1937. Bruck and Stern, ibid., 1908, p. 401.
Schutze, Berlin Klin. Woch., 1907, p. 126. Levaditi and Marie, Comptes
Rendus Soc. Biol., vol. Ixii., p. 872. Levaditi and Yamanouchi, ibid.,
vol. Ixiii., p. 740, and vol. Ixiv., pp. 275, 349, and 720. Marie, Levaditi,
and Yamanouchi, ibid., p. 169. Citron, Berlin. Klin. Woch., 1907, p. 1370.
Michaelis, ibid. Meier, ibid., p. 1636. Weil and Braun, Berlin. Klin.
Woch., 1907, p. 1570, and Wien. Klin. Woch., 1908, p. 151. Klausner,
Wien. Klin. Woch., 1908, p. 214. Landsteiner, Miller and, Potzl, Wien.
Klin. Woch., 1907. Forges and Meier, Berlin. Klin. Woch., 1908, p. 731.
Elias, Neubauer, Forges, and Salmon, Wien. Klin. Woch., 1908, p. 748.
Simplified forms of technique are given by Noguchi, Journ. Exp. Med.,
1909, p. 392, and Caulfeild, Journ. Med. Res., 1908, p. 507.

Rabies. — -An excellent account of modern views on the immunity to
rabies is given by Marie, Bull. Inst. Past., vol. vi., 1908, pp. 705 and 753.

See also Schneder, Zeit. f. Hyg., vol. xlii., p. 362. Remlinger, Bull.
Inst. Past., vol. ii., pp. 753 and 792.

INDEX OF AUTHORITIES CITED

ABEL, 1 18, 173
Albarran, 193
Allen, 367, 370, 384
Amos, 193, 209
Anderson, 312
Andrewes, 293, 359
Armand-Delille, 170, 196
Arrhenius, 80, 83 et seq., 119
Arthus, 311
Arzt, 417
Ascoli, 193, 233
Atkinson, 63, 67
Axenfeld, 366
Ayer, 188, 190

Bail, 220, 291, 292, 308, 340, 406

Bannerman, 404

Bassenge, 392

Bassett-Smith, 291, 375

Baumgarten, 250

Bechhold, 217

Beebe, 192

Bearing, von, 61, 132, 379

Bergengriin, 247

Bernard, 193

Besredka, 28, 1 13, 172, 190, 242, 314,

393

Bickel, 284
Bier, 31
Bierry, 192

Biltz, 90, 217, 324, 327, 328
Blum, no
Blumenthal, 106
Bolton, -.94
Bordet, 88, 153, 168, 180, 209, 212,

219, 228, 299, 322
Bousfield, 374
Bram Pusey, 197
Bredig, 320
Brieger, 54, 61, 411
Briscoe, 180, 248, 286, 333, 374
Brodie, 70
Browning, 286
Briick, 169, 285, 306
Brunton, Sir L., 224
Buchner, 54, 58, 86, 179

Bulloch, 127, 139, 1 80, 184, 259, 269

270, 286, 289, 290, 348, 378
Buxton, 215, 352

Cairns, 402

Calmette, 70, no, 122, 199, 3°3,

393- 403. 4T3
Casagrandi, 365
Castellani, 217
Cattani, 334
Centanni, 196, 418
Chantemesse, 304. 391
Chapin, 283
Charrin, 193, 204
Chatenay, 112
Chauveau, 18, 35
Cherry, 48, 70, 329
Choksy, 403
Citron, 293
Cler, 407
Cohn, 54, 411
Cole, 352
Collins, 220
Conradi, 181, 396, 405
Corpechot, 193, 197
Courmont, 107
Cowie, 283

Crendiropoulo, 209, 211, 214
Crofton, 283
Currie, 309

D

Danysz, 327
Davis, 372
Dean, 263, 283
Delaware, 193
Delbriick, 155
Delecarde, 122
Delezenne, 182, 196, 256
Denys, 179, 257
Descatello, 224
Dineur, 211
Dmitrevsky, 108, 127
Doerr, 293
Douglas, 257, 286
Doyon, 107
Dreyer, 87, 216, 324

439

440

INDEX OF AUTHORITIES CITED

Duclaux, 54

Dudgeon, 262

Dungern, von, 159, 192, 230, 327,

39i
Durham, 204

Ehrlich, 16, 33, 45, 69 et seq., 87, 94,

105, 125, 142 et seq., 200
Eisenberg, 216, 230, 343
Eisenzimmer, 418
Emden, von, 214
Ewing, 235
Eyre, 14, 34, 266, 288, 365, 366, 374

j.'

Falloise, 182, 183

Ferran, 400

Field, 320, 329

Figari, 193

Flexner, 155, 373

Forner, 418

Forster, 397

Frankel, 207

Friedberger, 162, 171, 281, 392

Friedemann, 217

Frouin, 207

G

Gay, 171, 172, 188, 190, 223, 312

Girard-Mangin, 326, 329

Gengou, 153, 169, 182

Golovine, 197

Goodall, 316

Goodman, 132, 310

Gordon, 359

Gruber, 204, 208, 211

Guedini, 170

Guldberg, 81

Guseff, 180

H

Haffkine, 401, 403
Hahn, 182

Hamburger, 297, 316
Hankin, 54, 179
Hardy, 252, 320
Harris, 55
Harris, Butler, 395
Hartoch, 416
Hekma, 297
Hektoen, 222
Henri, 326, 329
Herter, 115
Hewlett, 198
Hime, 285
Hoffmeister, 155
Hogyes, 419
Hoke, 231, 341
Houston, 371, 372, 373

Ignowtowsky, 44
Inman, 268
Irons, 304
Isaeff, 204, 364

J

acobi, 55

agic, 326

enner, 20

ex-Blake, 216, 324

ochmann, 373

ones, Wharton, 224
Joos, 215
Jungano, 196

K

Kanthack, 70, 184, 244, 252, 334

King, 132

Kirschbruch, 219

Kitashima, 61

Klein, 188

Klemperer, 364

Klien, 263

Knorr, 73, 133

Koch, 300, 305

Kolle, 373, 392

Korschun, 181

Kossel, 70

Kraus, 209, 226

Kruse, 397

Kutscher, 372

Kyes, 155

Lamb, 232

Lambotte, 181

Landsteiner, 190, 220, 222, 326, 417

Lastschenko, 181, 184

Laubry, 170

Lazar, 184

Leclef, 257

Ledingham, 128, 259, 286, 295, 416

Leishman, 261, 284

Le Play, 193, 197

Levaditi, 180, 183, 286, 335, 336, 417

Levy, 373

Liepmann, 195

Lignieres, 387

Lindemann, 193

Lingelsheim, von, 379

Lipstein, 409

Loffler, 173

Longcope, 180

Lubarsch, 139, 183

Lustig, 403

M

Macdonald, 265, 282, 365
Macfadyen, 58, 181, 390, 396, 399
Mackenzie, 373

INDEX OF AUTHORITIES CITED

44I

Madsen, 41, 76, 80, 82, 83 et stq., 86,

93. H9

Malvoz, 211, 406

Manouelian, 335

Maragliano, 383

Marenghi, 71

Marie, 419

Markl, 257, 259, 402

Marmorek, 17, 306, 583

Martin, 282, 373

Martin, S., 54

Marx, 184, 214

Massart, 299

McClintock. 132

McFarland, 383

Meakins, 177, 370

Mendel, 55

Mennes. 257

Metchnikoff, 43, 59, 60, 93, 107, 109,
113, 125, 134, 137, 152, 166, 179,
184, 190, 204, 238 et seq., 289, 293,

335, 4°6
Meyer, 411, 415
Morax, 107

Moreschi, 171, 172, 392
Morgenroth, 45, 88, 177, 181
Moro, 381

Muir, 88, 163, 164, 282, 286, 341
Miiller, 220, 417
Myers, 82, 228, 234

N

Nefedieff, 193
Neisser, 50, 70, 173, 217, 236, 323,

324. 4J5
Nernst, 87
Netter, 317

Neufeld, 257, 284, 285, 365
Nicolle, 170, 209, 211, 219
Nikayama, 292
Nocard, 381
Noguchi, 155, 232
Norris, 227
Norton, 395
Nuttall, 139, 228, 232, 250

Obermayer, 232, 234
Osborne, 55
Otto, 207, 312

P

Paget, Sir James, 361
Pane, 366
Panichi, 209
Parascandalo, 360
Park, 206, 220
Parvu, 170

Pasteur, 15, -8, 19, 34
Pauli, 298, 321, 324, 327

Pearce, 191 193

Perrin, 320

Peskind, 223

Petrie, 181

Petterson, 182

Pfeiffer, 58, 140, 162, 171, 184, 214,

281, 392
Pick, 232, 234
Pierallini, 353

Pirquet, von, 303, 308, 315, 381
Ponder, 296
Porges, 417
Posselt, 206
Potzl, 417
Pozerski, 170

R

Rankin, 372
Ransom, 106, 411, 415
Reagh, 215
Reid, 282, 375
Remy, 406
Richet, 309
Ricketts, 369
Rimpau, 257, 365, 392
Rist, 409
Roger, 204, 221
Rdmer, 106, 108, 351, 366
Rosenau, 297, 312,
Rosenfeld, 418
Rosenow, 282, 291, 343
Rossignol, 387
Rostane, 317
Roux, 86, 93, 414
Rowland, 58
Ruffer, 211, 214
Ruppel, 373

S

Sachs, 154, 172, 236, 327
Sagasser, 206
Salmon, 25, 39
Salmonsen, 93
Sanarelli, 250
Satchenko, 407
Schattenfroh, 181
Schereschewsky, 418
Schick, 315
Schmidt, 196
Schutze, 228, 234
Sclavo, 198, 407, 408
Sellards, 295, 297
Shattock, 262
Shiga, 199, 397, 405
Siedentopf, 319
Simon, 263
Smith, 25, 39, 215
Smith, Henderson, 286
Smith, Theobald, 311
Sobernheim, 406, 408
Soudakewitch, 336
Southard, 312

442

INDEX OF AUTHORITIES CITED

Stern, 232, 233
Stillmarck, 38, 55
Stockman, 29
Sturli, 224

T

Takaki, 106
Tauber, 366

Tchistovitch, 125, 227, 336
Teague, 320, 329
Tizzoni, 334
Todd, 396

Torrey, 215, 352, 368
Tunnicliffe, 123, 262, 267, 270

U

Uhlenhuth, 228, 232, 235, 236
Uschinsky, 54

Vaillard, 86, 93, 113, 411

Vallee, 387

Van de Velde, 179

Vincent, 113

Volk, 216

W

Waage, 81
Walker, Ainley, 17, 183, 208, 219

Walker, R., 262
Washbourn, 364

Wassermann, 44, 57, 71, 106, 184, 228,
232, 235, 293, 303, 306, 337, 373,

392, 4*5

Wechsberg, 50, 70, 173, 323
Weidenreich, 225
Weichardt, 195
Weigert, 98
Weil, 291, 292
Weinberg, 170
Welch, 59, 219
Western, 270, 395
Whitfield, 278, 299
Widal, 317
Wiltshire, 223
Woltmann, 194
Wood, Cartwright, 62, 65
Wright, Sir A., 25, 127, 189, 199 211,

257 * seq., 333, 350, 360, 375, 4°2

Yersin, 403

Zammit, 375
Ziemka, 235
Zsigmondy, 319

INDEX

ABRIN, 54; action on conjunctiva, 108

Abscess, cure of, 349

Absorption of complement. See Fixa-
tion of complement

Acne, 358

Acquired immunity, 19

Active immunity, 20. See Glossary

Addiment, 143. See Glossary

Adsorption, 90, 392

Age in relation to immunity, 8

Agglutination : by chemical substances,
211 ; mechanism of, 212; salts in,
209, 326

Agglutinins, 204 (see Glossary) ; action
of heat on, 205 ; chemical nature of,
214 ; to B. diphtheria, 409 ; to B.
dysenteria, 398 ; effects of tempera-
ture on, 208, 216 ; formation of, 208 ;
in normal blood, 99 ; to gonococci,
368 ; mechanism of action, 212 ; to
meningococci, 372 ; to pneumococci,
365 ; relation with cytolysins, 207 ;
role in immunity, 208 ; sensitiveness
of bacteria to, 219; specificity of, 205,
217; tostaphylococci, 359; to strepto-
cocci, 360; to tubercle bacilli, 383; to
typhoid bacilli, 389 ; to V. cholera, 400

Agglutinogen, 210. See Glossary

Agglutinoids, 47, 210, 216, 323. See
Glossary

Aggressins, 291 (see Glossary) ; speci-
ficity of, 292

Air, vitiated, 12

Albumoses in bacterial cultures, 54

Alcohol, 13

Alexins, 139 (see Glossary) ; source of,
179

Allergia, 308

Amboceptor. 141 (see Glossary) ; action
as opsonin, 285 ; formation of, 147 ;
methods of investigating, 185; source
of, 184

Amoeba, phagocytosis in, 238

Anaesthesia. 12

Anaphylaxis, 309. See Glossary

Anaphylactin, 313

Anthrax bacilli : immunity to, 405 ;
phagocytosis of, 244, 250, 252 ;
prophylaxis, 23, 407 ; toxins of, 405 ;
treatment of, 408 ; vaccination
against, 18, 23, 407

Anti-abrin, 108

Anti-agglutinin, 219

Anti-aggressin, 292

Anti-amboceptor, 160

Anti-antibodies, 104
j Anti-autolysin, 150

Antibodies, site of production of, 105,351

Anticomplement, 157

Anticrotin, 327

And enzymes, 48

Anti-epithelial serum, 197

Antigen, 101. See Glossary

Antihsemolysin, 109

Anti-intestinal serum, 195

Antileucolysin, 50

Antileucotoxin, 190

Antilysin, 160

Antispermotoxin, 109, 160

Antistaphylolysin, 52, 358

Antistreptocolysin, 360

Antitoxin : administration by mouth,
131 ; formation of, 60, 97, 114; in
normal blood, 93, 99 ; production by
toxoids, 99 ; reactions with toxin, 69 ;
role in immunity, 119; role in re-
covery, 119 ; unit of, 72

Antituberculin, 161, 307

Arsenic, absorption by leucocytes, 113

Arthus' phenomenon, 311. See Glossary

Atoxyl in trypanosomiasis, 6, 16

Atreptic immunity, 35. See Glossary

Atropin, absorption by leucocytes, 113

Aqueous humour, opsonin in, 286

Auto-agglutinin, 104, 222

Auto-anticomplement, 158

Autohsemolysin, 149. See Glossary

Auto-inoculation, 268, 382

Autonephrotoxin, 192

443

444

INDEX

B

B. anthracis. See Anthrax
Bacillus of hotulismus, toxin of, 40
B. coli : diseases due to, 393 ; im-
munity to, 393 ; toxins of, 393 ;

vaccine treatment, 394
B. diphtheria. See Diphtheria
Bacillus of dysentery, toxins of, 396
B. pyocyaneus : antagonism to anthrax,

40 ; antitoxin, 121 ; hsemolysin of,

53 ; leucolysin, 50
B. tetani. See Tetanus
B. typhosus. See Typhoid
Bacteria, immunity to, 331
Bacterial hsemolysins, 40, 44
Bactericidal serum, therapeutic use of,

198
Bactericidal power of blood, 139 ; of

serum, measurement of, 188
Bacteriolysis, 139. See Glossary
Bacterio-precipitin, 226, 231
Bacteriotropin. See Glossary
Bazillen emulsion, 384
Bleeding large animals, 65 ; small,

185

Blood, human, test for, 233, 235
Blood-relationship, 234
Boils, 358
Bone-marrow, reaction of, in infections,

34i
Bordet-Gengou phenomenon, 153, 168.

See Glossary

Bovine tuberculosis, diagnosis of, 380
Brain substance and tetanus toxin, 44,

106
Bright's disease, 13, 193

Calcium chloride, agglutination by,

211

Calcium lactate, use of, 317, 391

Calmette's test, 303, 382

Capsules (bacterial), function of, 343

Carbuncles, 358

Castellani's absorption reaction, 217

Cayman, reaction to tetanus toxin, 60

Cellulo-humoral theory, 249

Cerebro-spinal fever, 371

Cerebro-spinal fluid, 373, 374

Cervical catarrh, 395

Chemotaxis, 112, 244, 294, 341. See
Glossary

Chicken cholera, aggressin to, 291

Cholecystitis, 395

Cholera, 398; bacteriolysis in, 140, 337 ;
diagnosis of, 399 ; endotoxin of, 57,
59 ; Pfeiffer's test in, 140, 400 ; pro-
phylaxis, 400 ; toxins, 398

Cholesterin, action on toxins, 107

Coagulation of blood, liberation
complement in, 182

Coagulation of proteids, 321

Cobra-lecithid, 156

Cold in causation of disease, 9

Colchicine, latent period of, 41

Colitis, mucous, 395

Colloidal chemistry, 319

Colloids, 90 ; agglutination of, 217

Complement, 14.2, 145 (see Glossary) ;
as opsonin, 285 ; deviation of, I73>
323; (endo-), 156; fixation of, 170;
methods of research on, 185 ; ori-
gin of, 179, 252 ; specificity of,
286

Complementoid, 158. See Glossary

Complementophile haptophore group,
146

Complementoids, 47. See Glossary

Conjunctivitis, 368, 370

Copula, 141

Crisis (in pneumonia), 365

Cuti-reaction, 303, 381

Cystitis (B. coli}, 394

Cytase, 142, 167, 254. See Glossary

Cytolysins, 190 et seq. (see Glossary) ;
bacterial, 40

Cytophile haptophore group, 145

Cytotoxin, 197

1)

Danysz effect, 327. See Glossary

Daphnia, phagocytosis in, 238

Dead bacteria as vaccines, 24

Dendroccelum, digestion in, 241

Desmon, 141. See Glossary

Deuterotoxin, 75

Deviation of complement, 173, 323.
See Glossary

Diabetes, 13

Digestion, intracellular, 241

Diphtheria antitoxin : dosage of, 410 ;
in normal blood, 93 ; standardiza-
tion of, 45

Diphtheria bacillus, antiserum against,
409

Diphtheria : diagnosis of, 409 ; latency
of, 33 ; local immunity to, 30 ; pro-
phylaxis, 410 ; toxin of, 40 ; action
of, 49 ; neutralization of, 72, 85 ;
standardization of, 45

Diphtheritic paralysis, 72, 80, 87

Dissociation, 28, 85

Dominant complement, 154. See
Glossary

Dosage of vaccines, 24

Dysentery, 396 ; bacillus, agglutination
of, 220 ; prophylaxis of, 397 ; treat-
ment of, 397

INDEX

445

Eclampsia, cytolytic theory of, 195

Eel serum: immunity to, 125, 130, 135 ;
precipitin for, 228

Ehrlich's phenomenon, 328. See Glos-
sary

Electrolysis of toxins, 90, 92

Endocomplement, 156. See Glossary

Endothelial cells as phagocytes, 246

Endotoxin, 56, 339. See Glossary

Enterokinase, 256

Enzymes : analogies with toxins, 42 ;
proteolytic, in pus, 337

Epitoxoid, 76

Epitoxonoid, 327

Ergophore group. See Glossary

Erysipelas, treatment of, 362

Evolution, 130, 165 ; of bacteria, 221,

344

Exhaustion, Pasteur's theory of, 34
Exotoxins, 48 (see Glossary) ; chemical

nature of, 53

False rise, 274

Fatigue, 10

Fixation of complement, 153, 168, 236.

See Glossary

Fixator, 141, 167. See Glossary
Flagella, agglutination of, 227
Food, insufficient, II
Fowl cholera, 18, 22
Frog, action of tetanus toxin on, 45
Frontal sinus suppuration, 367

Gastrotoxin, 194. See Glossary

Gengou's reaction. See Glossary

Giant cells, 378

Gleet, opsonic index in, 368

Gonococci : immunity to, 369 ; local
immunity to, 30, 369 ; opsonic index
to, 368 ; vaccines in disease due to,
370

Group reactions, 217. See Glossary

H

Hsemagglutinin, 221. See Glossary
Hsemolysins : bacterial, 40, 44, 50 ;

serum, 141 et seq.
Haemolysis, 40, 141 (see Glossary) ;

by silicic acid, 326 ; methods of

research, 185
Hsemolysoids, 46, 51
Hsemopsonin, 245, 273, 285
Haptines, 95. See Glossary
Haptophore group, 46. See Glossary
Hepatotoxin, 192
Heterolysins, 149
Hog cholera, vaccination against, 39

Horse-flesh, test for, 237
Horse-sickness, vaccination against, 28
Hydatids, diagnosis of, 170
Hypersensitiveness to toxins, 61, 121.
See Anaphylaxis

Ichthyotoxin, 101, 125

Immune body, 141. See Glossary

Immunisin, 141

Immunitas non sterilisans, 16, 33, 331

Immunity : acquired, 19 ; active, 20 ;
atreptic, 35 ; bacterial, 331 ; defini-
tion of, I ; due to loss of receptors,
125 ; local, 29 ; mixed, 28 ; natural,
7 ; of leucocytes, 128 ; passive, 26 ;
to toxins, 115; to toxins, natural,

134

Incitor element. See Glossary
Indol, 115
Infection, definition of, 5 ; predisposing

causes of, 9
Interbody, 141
Intermediary body, 141
Ions, 80

Iritis, gonococcal, 370
Isoagglutinin, 221. See Glossary
Isolysin, 149
Isoprecipitin, 234

K

Koch's phenomenon, see tuberculin.

See Glossary
Kraus's reaction, 209

Latency of bacteria) 33 ; of tubercle
bacilli, 387

Latent period of toxins, 41

Lecithin : action on toxins, 107 ; role
in haemolysis, 156, 180

Lens, crystalline, precipitin to, 233

Lethal dose, minimal, 42

Leucocytes : absorption of toxins by,
44, 113 ; as source of complement,
179 ; chemotactic attraction of, 112 ;
degeneration of, 122 ; during starva-
tion, etc., 12 ; immunity of, 128 ;
in combating toxins, 120, 137 ; in
Metchnikoff s theory, 242 ; prepara-
tion of emulsions of, 257

Leucocytosis in prognosis, 112, 341

Leucolysins, 49, 358

Leucopsenia, 341

Leucotoxic serum, 190

Leucotoxins, 49, 358

Liver, phagocytosis in, 336

Local lesion, 33 ; cure of, 346

Local immunity, 29, 125

Lungs, phagocytosis in, 248, 336

446

INDEX

M

Macrocytase, 152, 254. See Glossary

Macrophage, 247. See Glossary

Malaria, immunity to, 33

Mallein, 303

Malta fever, 374 ; treatment of, 375

Meats, recognition of, 237

Meningococcus, toxins of, 371 ; phago-
cytosis of, 371

Meningitis : serum treatment, 373 ;
vaccine treatment, 374

Micrococcus melitensis, agglutination of,

375

Microcytase, 152, 254. See Glossary
Microphage, 247. See Glossary
Minimal lethal dose, 42
Monospora, phagocytosis of, 239, 240
Mytilo-congestine, 309

N

Nasik vibrio, toxin of, 41
Natural immunity, 7
Negative phase, 62 (see Glossary) ; in

opsonic index, 274 ; summation of,

276
Neisser-Wechsberg phenomenon, 173,

323. See Glossary.
Nephrotoxin, 192. See Glossary
Nerves, peripheral, cytolytic serum for,

196

Neutralization of poison?, 116
Neurotoxin, 196
New tuberculin, 384
Nicotin, absorption by liver, 116
Nitrites, production of, in cholera, 37
Nucleo-proteids as antigens, 192

Ophthalmo-reaction, 303, 304, 382

Ophthalmotoxic serum, 197

Opsonic index, 261 ; in acute diseases,
265; in chronic diseases, 268; in
diphtheria, 267 ; effect of dilution of
serum, 264 ; effect of vaccines, 274 ;
in erysipelas, 270 ; false rise in, 274 ;
to gonococci, 270 ; to meningococci,
371 ; pre-agonal rise in, 280 ; to
pneumococci, 265, 361, 365 ; in
staphylococcic diseases, 266 ; tubercle
bacilli, 268, 382

" Opsonins-therapy," 277 ; in tubercle,

385

Opsonins: etTect of temperature on their
action, 295 ; fundamental experi-
ments on, 257 ; Metchnikoffs views
on, 289 ; nature of, 273 ; origin of,
288 ; presence in plasma, 286, 333 :
specificity of, 270 ; thermostability,
259, 282 ; technique of researches on,
259-264; thermolabile, 285 ; thermo-
stable, 259, 282

Orthophosphoric acid, agglutination by,

216
Osteomyelitis, 358

Passage, 15, 219

Passive immunity, 26. See Glossary

Peritoneum, phagocytosis in, 248, 250,

352

Perlsucht tuberculin, 384

Pfeiffer's phenomenon, 141. See Glos-
sary

Phagocytic index, 260

Phagocytosis, 238 et seq.; action of
salts in, 297; in circulating blood,
333 ; in peritoneum, 248, 250, 352 ;
influence of source of leucocytes, 290 ;
influence of temperature, 295 ; influ-
ence of virulence, 343 ; nature of, 295

Phagolysis, 181, 353

Philocytase, 141

Phthisis, 12

Pigmentolysin, 197

Piroplasma bigeminum, 21

Pirquet's (von) reaction, 303, 381

Placentolysin, 195

Plague, 402 ; prophylaxis of, 403 ;
serum treatment of, 403

Plasma, complement in, 182; opsonin
in, 286, 333

Pleuralistic conception, 151

Pleuropneumonia of cattle, 22

Pneumonia. See Pneumococci

Pneumococci : agglutinins to, 365 ;
immunity to, 365 ; in childhood, 8 ;
opsonic index to, 266 ; serum treat-
ment, 366 ; vaccine treatment, 366 ;
virulence, 366

Poisons : difference from toxins, 37 ;
neutralization of, 116

Polyceptor. See Glossary

Polyvalent serum, 363 (see Glossary) ;

vaccine (see Glossary)
! Positive phase, 62. See Glossary
: Potato bacillus, phagocytosis of, 248
j Precipitins, 226 (see Glossary) ; speci-
ficity of, 232, 235

Precipitoid, 228, 323. See Glossary

Precipitogenoid, 231. See Glossary

Predisposing causes, 9

Preparator, 141. See Glossary

Pro-agglutinin, 210

Prostatotoxin, 196

Proteids, coagulation of, 321

Prototoxin, 75

Prototoxoid, 73

Pro-zones, 216, 323. See Glossary

Pus, enzymes of, 337

Pyocyaneus (B. ) : antagonism to anthrax,
40 ; toxin of, 57

Pyocyanolysin, 53

INDEX

447

R

Rabies, vaccination against, 23, 419 ;

virus of, 15
Reactions : curative effects of, 281 ;

tubercle, etc., 300
Receptors, 95 (see Glossary) ; loss of,

I25. 343 ; sessile, 126, 160 /

Relapsing fever, recovery from, 335 r
Retention theory. Chauveau's, 35
Reversible reactions, 81
Ricin, 38, 54, 55

Rinderpest, vaccination against, 21
Ringworm, local immunity to, 30

Salts, role of, in agglutination, 209,

211

Septicsemia, 332

Serum : anti-anthrax, 408 ; anticholera,
400 ; antidiphtheritic, 409 ; anti-
meningococcic, 373 ; antipneumo-
coccic, 366 ; antiplague, 403 ; anti-
streptococcic, 363 ; antitetanic, 413 ;
antityphoid, 390-392 ; bacteriolytic,
use of, 198; disease, 315; toxin,
62

Side-chains, 95

Side-chain theory, 94. See Glossary

Smith's (Theobald) phenomenon, 311

Specific inhibition, 230

Specificity, 19, 105 (see Glossary) ;
of agglutinins, 205, 217; of cyto-
lysins, 191; of precipitins, 232

Spectrum of toxin, 73

Spermotoxin, 109, 190

Spleen, phagocytosis in, 334

Staphylococcus pyogenes : bacteriolysis,
337; leucolysin of, 50, 358; im-
munity against, 359 ; recovery, 347 ;
toxins of, 388 ; vaccine treatment of,

359

Staphylolysin, 51, 52, 358
Starvation, II
Sdmulins, 122, 295
Stomach, immunity of, 29
Streptococci : immunity to, 360 ; serum

treatment of disease due to, 362 ;

toxins of, 359 ; vaccine treatment of

disease due to, 362
Streptococcus pyogenes : hcemolysin of,

5°) 359 I leucolysin of, 50
Streptocolysin, 50
Substance sensibilatrice^ 141
Surface-tension, 213, 298
Swine erysipelas, 19, 22
Symbiosis of leucocytes, 248
Sympathetic ophthalmia, 197
Syncytiolysin, 195
Syphilis, 415 ; Wassermann's reaction

in, 415

; Teianolysin, 52, 83, 411

Tetanospasmin, 52, 411

Tetanus: diagnosis of, 410; immunity
to, 412 ; passive immunity to, 27 ;
prophylaxis of, 413; treatment of,
414 ; toxin, 40, 47, 411 ; absorption
of, by brain, 44, 106 ; leucocytes, 44;
tissues, 44; action of, 49, 116; on
various animals, 133 ; antitoxin to,
413 ; effect of temperature on, 45
i Texas fever, vaccination against, 21
! Thyrotoxin, 197. See Glossary

Tick fever, 335

Tissue immunity. See Local immunity

Toxalbumin, 54

Toxins : action of, 41 ; composition of,
47 ; electrolysis of, 90, 92 ; hyper-
sensitiveness to, 61, 309; immunity
to, 115 ; spectrum, 73 ; standardi-
zation of, 45, 71 ; union with tissues,
100

Toxoids, 46, 47, 80 (see Glossary) ; pro-
duction of antitoxin by, 99 ; use in
immunization, 62

Toxone, 73, 80. See Glossary

Toxonoid, 88

Toxophore group, 46. See Glossary

Trichina spiralis, 29

Trichotoxin, 192. See Glossary

Tritotoxin, 76

Trypanosomiasis, 6, 1 6

Tubercle bacillus : antibodies for, 337 ;
immunity to, 377 ; opsonic index
to, 377 ; phagocytosis of, 251, 377 ;
toxins of, 314, 376 ; toxins of, Mar-
morek's, 383

Tuberculin : dilution of, 380 ; immuni-
zation to, 303 ; old, 300, 379 ; re-
action, 301, 378 ; theories of reaction,
305, 306, 308 ; reaction in cattle,
380 ; R, 383 ; therapeutic use of,

384

Tuberculosis : diagnosis of, 379 ; op-
sonin therapy of, 385 ; prophylaxis
of, 386 ; prophylaxis in cattle, 387

Tulase, 387

Typhoid bacillus : agglutinins to, 206,
210, 215 ; agglutinins in normal
blood, 99 ; endotoxin of, 53 ; haemo-
lysin of, 53 ; immunity, 388 ; in
peritoneum, 353 ; latency of, 33 ;
phagocytosis of, 263, 389 j virulence
of, 17, 343

Typhoid fever : bacterisemia in, 332 ;
bacteriolysis in, 337 ; opthalmo-re-
action in, 304 ; prophylaxis of, 28,
390

Typholysin, 53

Tyrosin : action on toxins, 107

448

INDEX

U

Ulcerative endocarditis, 332
Ulcus serpens, treatment. of, 366
Unitarian theory of complement, 152
Unit of toxin, 72 ; of antitoxin, 72
Uraemia, cytolytic theory of, 192

Vaccine treatment : for gonococcic
diseases, 370 ; for Malta fever, 375 ;
for meningitis, 374 ; for pneumo-
coccic diseases, 366 ; for staphylo-
coccic diseases, 359 ; for strepto-
coccic diseases. 362 ; theory of, 276 ;
tubercle (see Tuberculin)

Vaccines : chemical, 25, 39 ; of dead
bacteria, 24 ; dosage of, 24 ; method
of preparation, 1 8

Vaccinia, 22

Venom (snake), '155

Vibrio, Nasik, 41

Vibrio Metchnikovi, bacteriolysis of,

173

Vibrio Nordhafen, bacteriolysis of, 174
Virulence, 14 ; capsule formation,
effect of, on, 343 ; changes in, 15, 17 ;
diminution in, methods of produc-
tion, 18 ; increase in, methods of
production, 15, 17, 219; influence
on phagocytosis, 295 ; mechanism
of, 343

Virus (see Glossary), 22 ; fixed, 15 ; of
the streets, 15

W

Wassermann's reaction, 416
Wet, in causation of disease. 9
Widal reaction, 389

Zones of inhibition, 216

Zoological affinities : immunity in re-
lation to, 8 ; relations to precipitins,
234

K. LEWIS, 136, GOWER STREET, LONDON, W.C.

Yf 8^/35

fet

BIOLOGY

LIBRARY

G

UNIVERSITY OF CALIFORNIA LIBRARY