Skip to main content

Full text of "Infection and resistance; an exposition of the biological phenomena underlying the occurrence of infection and the recovery of the animal body from infectious disease"

See other formats 


INFECTION  AND  RESISTANCE 


THE   MACMILLAN  COMPANY 

NEW  YORK   •    BOSTON   •    CHICAGO 
DALLAS   •    ATLANTA    •    SAN  FRANCISCO 

MACMILLAN  &  CO..  LIMITED 

LONDON    •    BOMBAY    •    CALCUTTA 
MELBOURNE 

THE  MACMILLAN  CO.  OF  CANADA,  LTD- 

TORONTO 


INFECTION  AND 
RESISTANCE 


AN    EXPOSITION   OF   THE   BIOLOGICAL    PHENOMENA 

UNDERLYING  THE  OCCURRENCE  OF  INFECTION 

AND  THE  RECOVERY  OF  THE  ANIMAL  BODY 

FROM  INFECTIOUS  DISEASE 


BY 

HANS  ZINSSER,  M.D. 

Professor  of  Bacteriology  at  the  College  of  Physicians  and  Surgeons,  Columbia  University, 

and  Bacteriologist  to  the  Presbyterian  Hospital,  New  York 

Major,  Medical  Officer's  Reserve  Corps,  U.  S.  A. 


WITH  A  CHAPTER  ON 

COLLOIDS  AND  COLLOIDAL  REACTIONS 

BY 

PROFESSOR  STEWART  W.  YOUNG 

Department  of  Chemistry,  Stanford  University 
SECOND  EDITION  REVISED 


gorfe 
THE  MACMILLAN  COMPANY 

1918 


COPTBIGHT.   1914  AND  1918 

BY  THE  MACMILLAN   COMPANY 


Set  up  and  electrotyped     Published  October,  1914 

Reprinted  January  1916 

Reprinted  January.  1917 

New  Edition,  Revised  and  Reset,  May.  1918 


TO 

a.  z. 

THIS  BOOK  IS  AFFECTIONATELY 

DEDICATED  B\  HIS 

SON 


387343 


PREFACE 

INFECTIOUS  disease,  biologically  considered,  is  the  reaction  which 
takes  place  between  invading  micro-organisms  and  their  products,  on 
the  one  hand,  and  the  cells  and  fluids  of  the  animal's  body  on  the 
other.  The  disease  is  the  product  of  two  variable  factors,  each  of 
them  to  a  certain  extent  amenable  to  analysis,  and  it  is  self-evident 
that  no  true  understanding  of  this  branch  of  medicine  is  possible 
without  a  knowledge  of  the  biological  principles  which  laboratory 
study  has  revealed. 

For  the  purpose  of  helping  to  render  such  knowledge  easily  ac- 
cessible this  book  was  written.  While  it  is  hoped  that  it  may  prove 
useful  to  the  practitioner  and  laboratory  worker,  it  is  intended  pri- 
marily for  the  undergraduate  medical  student.  To  many  it  will 
seem  that  the  subject  in  general  and  our  method  of  treatment  espe- 
cially are  too  technical  and  difficult  for  this  purpose.  Our  own  ex- 
perience contradicts  this.  During  the  past  three  years  the  writer  has 
had  the  opportunity  to  deliver  lectures  and  to  give  laboratory  courses 
on  this  subject  to  medical  students  of  2d,  3d,  and  4-th-year  classes  at 
the  Stanford  and  Columbia  Universities.  It  has  been  a  pleasant 
experience  to  find  the  medical  student  eager  for  the  opportunity  to 
obtain  this  knowledge  and,  under  the  present  increased  require- 
ments for  preliminary  training  at  our  best  schools,  fully  capable  of 
assimilating  it.  It  is  not  a  good  plan  to  attempt  too  extensively  to 
simplify  material  that,  in  its  close  analysis,  presents  complex  phe- 
nomena and  intricate  reasoning.  For  this  reason  no  attempt  has 
been  made  to  write  an  A  B  C  of  immunity  as  a  quick  road  to  com- 
prehension. No  true  insight  into  any  branch  of  medicine  or,  for  that 
matter,  into  any  other  science,  can  be  attained  without  a  certain 
amount  of  labor;  however  the  concepts  of  this  subject  are,  indeed, 
relatively  simple  after  the  first  principles  have  been  mastered,  and 
the  writer  has  attempted,  therefore,  at  the  risk  of  seeming  pedantic 
in  places,  to  treat  the  subject  critically,  separating  strictly  those 
data  which  may  be  accepted  as  fact  from  those  in  which  legitimate 
differences  of  opinion  prevail. 

As  far  as  was  feasible  every  chapter  has  been  written  as  a  sepa- 
rate unit.  This  has  necessitated  occasional  repetition,  but,  it  is 
hoped,  will  add  considerably  to  clearness  of  presentation  in  each  indi- 
vidual subject.  Theories  have  been  discussed  with  as  little  prejudice 
as  the  possession  of  a  personal  opinion  in  many  cases  has  permitted. 

vii 


viii  PREFACE 

The  chapter  on  Colloids  was  written  especially  for  the  book  by 
Prof.  Stewart  W.  Young,  of  Stanford  University.  Since  so  many 
analogies  between  serum  reactions  and  those  taking  place  between 
colloidal  substances  generally  have  been  observed,  it  has  seemed  best 
to  devote  this  chapter  entirely  to  the  elucidation  of  the  principles 
governing  colloidal  reactions,  so  that  its  contents  may  be  utilized 
as  explanatory  of  the  many  allusions  made  to  colloids  in  the  rest  of 
the  text. 

All  available  sources  of  information  have  been  freely  used.  In 
the  large  majority  of  cases  we  have  had  access  to  the  original  papers 
and  monographs.  However,  we  acknowledge  much  aid  from  care- 
ful reading  of  the  admirable  summaries,  written  by  acknowledged 
authorities,  in  the  works  edited  by  Kolle  and  Wassermann,  and 
by  Kraus  and  Levaditi.  Similar  acknowledgment  is  made  to 
equally  important  sources  in  Weichhardt's  Jahresbericht,  the  Bulle- 
tins of  the  Pasteur  Institute,  and  in  such  text-books  as  those  of  Paul 
Theo.  Miiller,  Emery,  Adami,  Gideon  Wells,  Marx,  Dieudonne,  and 
others.  It  is  needless  to  acknowledge  the  use  of  such  classics  as  that 
of  Metchnikoff  or  of  the  many  critical  writings  of  Bordet  and  of 
Ehrlich — masters  who  have  helped  to  shape  the  thoughts  of  all  men 
working  in  this  field. 

The  writer  takes  pleasure  in  acknowledging  many  helpful  sugges- 
tions from  his  associates,  Drs.  Hopkins  and  Ottenberg,  and  much 
aid,  in  the  verification  of  references,  from  Mr.  Walter  Bliss,  Fellow 
in  the  Department  of  Bacteriology. 


PREFACE  FOE  SECOND  EDITION 

THE  revision  for  the  second  edition  of  this  book  has  been  made 
under  difficult  conditions — since,  for  the  greater  part  of  the  time 
immediately  preceding  its  issue,  the  writer  was  removed  from  the 
facilities  of  libraries  and  journal  files.  There  are,  in  consequence, 
many  spots,  here  and  there,  which  have  not  been  as  thoroughly 
scrutinized  as  we  would  wish.  Nevertheless,  all  the  important 
changes  necessitated  by  the  lapse  of  time  since  its  first  printing 
have  been  made  and  much  new  material  added. 

The  chapters  on  anaphylaxis  have  been  almost  completely  re- 
written, as  demanded  by  changes  of  view  resulting  from  the  work 
of  others,  as  well  as  our  own.  The  Abderhalden  reaction  having 
been  proved  to  be  an  interesting  camouflage,  the  material  contained 
in  this  section  has  been  revised  and  the  more  recent  work  on 
enzymes  added.  The  development  of  conceptions  of  non-specific 
serum  and  cellular  reactions  has  been  discussed  especially  in  con- 
nection with  the  recent  important  work  of  Jobling  and  Peterson. 
A  section  of  Immunity  in  Syphilis  has  been  added,  based  on  our 
own  studies  with  Hopkins  and  McBurney,  and  the  chapter  dealing 
with  specific  therapy  in  various  infections  has  been  revised  and 
expanded.  In  addition  to  this,  many  minor  alterations  and  com- 
ments have  been  made,  paragraphs  omitted  and  inserted. 

Throughout,  there  has  been  no  attempt  to  make  the  book  what  is 
generally  known  as  a  "practical  manual."  There  are  many  books 
which  devote  themselves  particularly  to  the  instruction  of  workers 
in  measurements  and  manipulations,  and  we  believe  now,  as  we  did 
before,  that  the  present  volume  would  prove  most  useful  by  present- 
ing in  detail  the  fundamental  principles  underlying  the  biology  of 
infectious  diseases.  Such  treatment  of  the  subject  is,  in  our  opinion, 
indispensable  as  a  preliminary  training  for  those  who  deal  with  in- 
fection in  the  clinic  or  the  laboratory,  and  the  material  on  which 
the  book  is  based  is  included,  at  present,  in  lectures  required  of 
second  year  students  at  the  Columbia  School  of  Medicine  and  is  a 
prerequisite  to  further  practical  training  in  Serology. 

HANS  ZINSSER. 


CONTEXTS 

PAGB 

CHAPTER  I. — INFECTION  AND  THE  PROBLEM  or  VIRULENCE    ' .        .        *        .        1 

Scope  of  subject.  Conception  of  infection.  Attributes  of  pathogenic 
microorganisms.  Forms  of  infection.  Influences  of  biological  adapta- 
tion. Classification  of  parasites  on  the  basis  of  invasive  properties. 
Factors  which  determine  the  power  to  invade.  Fluctuations  in  viru- 
lence. How  microorganisms  defend  themselves  against  destruction. 
Serum  fastness,  arsenic  fastness,  capsule  formation,  inagglutinability, 
etc.  Resistance  to  Phagocytosis.  Development  of  offensive  properties 
on  the  part  of  bacteria.  Specificity  of  different  infections.  Chronic 
septicaemia.  "Sub-infection."  Selective  lodgment  in  tissues.  Locali- 
zation and  generalization.  Incubation  time. 

CHAPTER  II. — BACTERIAL  POISONS  .       ...       .        .        .        .        .        .        .      28 

Part  played  by  bacterial  poisons  in  clinical  manifestations.  Pto- 
maines. Importance  of  ptomaines  in  disease.  True  toxins  or  exo- 
toxins.  Endotoxiris.  Chemotactic  bacterial  extracts.  Basic  proper- 
ties of  true  toxins.  Other  substances  biologically  similar  to  them. 
Analogy  to  enzymes.  Snake  venoms.  Incubation  time  of  toxins. 
Conception  of  antitoxins.  Work  of  Vaughan.  Researches  of  Fried - 
berger.  Absorption  of  toxins.  Selective  action  of  toxins.  Distribu- 
tion of  tetanus  poison.  Causes  underlying  selective  action  in  general. 
Injury  done  during  the  excretion  of  toxins.  Union  of  toxins  with  sus- 
ceptible cells.  Importance  of  cell  lipoids. 

CHAPTER    III. — OUR    KNOWLEDGE    CONCERNING    NATURAL    IMMUNITY,    AC- 
QUIRED IMMUNITY  AND  ARTIFICIAL  IMMUNITY     -.•      -,  -      .        .      49 

The  struggle  between  the  infectious  agent  and  the  defensive  forces 
of  the  body.  External  defenses.  Skin  secretions.  Natural  vs.  arti- 
ficially acquired  immunity.  Species  immunity.  Racial  immunity.  Dif- 
ference between  individuals.  Inheritance  of  natural  immunity.  Im- 
munity resulting  from  an  attack  of  the  disease.  Jenner  and  smallpox. 
Pasteur's  work  with  chicken  cholera.  Active  immunization.  Passive 
immunization.  Pasteur's  studies  on  anthrax.  Different  methods  of  con- 
ferring active  immunity.  Methods  of  obtaining  bacterial  extracts.  De- 
velopment of  our  knowledge  of  passive  immunization.  Early  attempts. 
Behring  and  his  collaborators.  Ehrlich  's  work  on  ricin.  Snake  venom. 
Specificity. 

CHAPTER   IV. — THE   MECHANISM   OF   NATURAL   IMMUNITY   AND   THE   PHE- 
NOMENA FOLLOWING  UPON  ACTIVE  IMMUNIZATION.     ...      78 

Investigations  on  problems  of  inflammation.  Metchnikoff's  earlier 
studies.  Concentration  of  attention  upon  the  properties  of  the  blood. 
Grohman's  work.  Early  opposition  of  cellular  and  humoral  points 
of  view.  Buchner.  Nuttall.  Earlier  arguments  brought  forward  by 
the  two  schools.  Behring 's  summary  of  the  situation  at  this  time. 
Phenomena  following  upon  active  immunization.  Earlier  theories. 
Kxhaustion  theory.  Retention  theory.  Alkalinity  theory.  Osmotic 
theory.  Discovery  of  specific  antibodies  by  Behring  and  collaborators. 
Khrlich 's  study  on  ricin.  Antitoxins.  Pfeiffer's  discovery  of  lysins. 
Agglutinins.  Precipitiiis.  Opsonins.  Tropiiis.  Conception  of  anti- 


x  CONTENTS 

PAGE 

bodies  as  a  whole.     Generalization  of  the  facts  discovered  in  the  case 
of   bacteria.      Haemolysins.     Cytotoxins.      Haemagglutinins.     Precipitins 
to  unformed  proteins.     Conception  of  an  antigen.     Nature  of  antigens. 
Analogy  of  immunity  with  drug  tolerance. 
Origin  of  antibodies. 

CHAPTER  V. — TOXIN  AND  ANTITOXIN 104 

Nature  of  the  reaction.  Earlier  views.  Calmette's  work  on  snake 
poison.  Filtration  experiments.  Morgenroth  's  work  on  HCl-toxin  modi- 
fications. Ehrlich's  ricin  neutralization.  Development  of  the  neutrali- 
zation ideas  by  Ehrlich  and  Behring.  Conception  of  antitoxin  unit. 
Instability  of  toxin.  Ehrldch  's  experiments.  The  conceptions  of 
M.L.D.,  LO  and  L+  doses.  Discrepancy  between  L0  and  L+.  Toxoids 
and  toxons.  Method  of  partial  absorption.  Toxin  spectrum.  Opinions 
of  Arrhenius  &  Madsen.  Bordet's  opinion.  The  Danysz  effect.  THE 
SIDE  CHAIN  THEORY.  Early  work  of  Knorr.  Ehrlich's  analogy  of  cell 
with  chemical  substances.  Analogy  with  ferments.  Weigert's  law  of 
overcompensation.  Antibodies  and  cell  receptors.  Theoretical  con- 
clusions drawn  from  work  with  tetanus  poison.  Importance  of  lipoidal 
substances  in  brain  tissue. 

CHAPTER  VI. — BACTERICIDAL  PROPERTIES  OF  BLOOD  SERUM,  CYTOLYSIS  AND 

SENSITIZATION          ..........     134 

The  phenomenon  of  bacteriolysis  and  the  bactericidal  effect.  Haemo- 
lysis. The  mechanism  of  cytolysis.  Amboceptor  or  sensitizer?  The 
complement  or  alexin.  Iso-antibodies.  Discussion  of  views  of  Ehrlich 
and  Bordet.  Multiplicity  of  antibodies.  Multiplicity  of  complement 
or  alexin.  Anti-antibodies.  Neisser  and  Wechsberg  phenomenon  of 
complement  deviation.  Quantitative  relation  between  complement  and 
amboceptor.  Conglutinins. 

CHAPTER   VII. — DEVELOPMENT  OF   OUR   KNOWLEDGE   CONCERNING   COMPLE- 

.    MENT  OR  ALEXIN.     COMPLEMENT  FIXATION  .        .        ...     168 

Origin  of  alexin.  Microcytase  and  Macrocytase.  Anti-lysins.  Alexin  in 
osdema  fluids,  etc.  The  question  of  alexin  in  the  circulating  blood. 
Alexin  and  the  thyroid.  Alexin  and  the  liver.  Chemical  nature  of 
complement  and  alexin.  Cobra-lecithid.  Enzyme-like  nature  of  alexin. 
Filtration  of  alexin.  Complement  or  alexin  splitting.  Return  to  activ- 
ity of  inactivated  alexin  on  standing.  Inactivation  by  shaking. 
ALEXIN  FIXATION.  Bordet-Gengou  experiments.  Theoretical  explana- 
tion of  these  facts.  Albuminolysins  of  Gengou.  Alexin  fixation  by 
precipitates.  Views  of  earlier  writers.  Nicoll  's  view.  Writer 's  opinion. 
Conception  of  ' '  Bordet-antibody. ' '  Nonspecific  alexin  fixation.  Impor- 
tance of  lipoids.  Fixation  by  unsensitized  cells,  substances  in  sus- 
pension. Anticomplementary  properties  of  serum. 

CHAPTER      VIII. — PRACTICAL      APPLICATIONS      OF      COMPLEMENT-FIXATION 

METHOD.    THE  WASSERMANN  REACTION 198 

Historical.  Early  work  on  monkeys.  First  use  of  human  beings. 
Theories  of  the  Wassermann  reaction.  Methods  of  preparing  antigen. 
Titration  of  antigen.  Titration  of  haemolytic  sensitizer.  Alexin  titra- 
tion.  Performance  of  the  test.  Noguchi's  modification.  Modifica- 
tions of  Bauer,  Stern  and  others.  Results  of  test.  Reliability.  Spinal 
fluid,  etc.  COMPLEMENT  OR  ALEXIN  FIXATION  FOR  THE  DETERMINATION 
OF  UNKNOWN  PROTEIN.  Neisser-Sachs  method.  Principles  of  the 
method.  Performance  of  test.  COMPLEMENT-FIXATION  TESTS  IN  DIAG- 
NOSIS OF  MALIGNANT  NEOPLASMS.  Historical.  Von  DungenTs  method. 
Results  obtained.  Complement-fixation  in  glanders.  Complement-fixa- 
tion in  gonococcus  infections.  Complement-fixation  in  tuberculosis. 


CONTENTS  xi 

PAGE 

CHAPTER  IX. — THE  PHENOMENON  OF  AGGLUTINATION 218 

Discovery.  Applications  of  clinical  methods.  Clinical  usefulness.  Re- 
lation to  motility.  Passive  role  of  bacteria.  Bordet  's  discovery  of 
the  importance  of  electrolytes.  Nature  of  agglutinogen.  Alterations 
by  heat.  Alterations  in  agglutinability.  Reasons  for  agglutinability. 
Specificity.  Biological  relations  between  bacteria  parallel  to  agglutin- 
ins.  Castellani's  method  of  absorption.  Normal  agglutinins.  Agglu- 
tinoids.  Inhibition  zones.  Bordet 's  views.  * '  Two-phase  ' '  theory.  Phys- 
ical interpretation.  The  work  of  Neisser  and  Friedemann.  Acid  agglu- 
tination. Iso-agglutinihs. 

CHAPTER  X. — THE  PHENOMENA  OF  PRECIPITATION  .  *  ...  ,248 
Discovery.  Bacterial  filtrates.  Expansion  of  principle  to  proteins  in 
general.  Nature  of  precipitinogen.  Specificity.  Quantitative  rela- 
tions in  the  reaction.  Practical  uses.  Nuttall's  studies  on  precipitins 
and  historical  relationship.  Forensic  uses  of  the  test.  Performance 
of  the  test  as  advised  by  Uhlenhuth.  Influence  of  heat  upon  precipi- 
tinogen. Organ  specificity.  Ehrlich  's  view  of  the  nature  of  the  reac- 
tion. Physical  views  of  the  reaction.  Presence  of  precipitinogen  and 
precipitin  in  same  serum.  Analogy  with  colloids  of  known  constitution. 

CHAPTER  XI. — PHAGOCYTOSIS.     CHEMOTAXIS  .        .        ."       .        .        .        .    272 

Early  investigations.  Metchnikoff  's  first  studies.  Phagocytosis  in  lower 
animals.  Its  significance.  Importance  in  the  development  from  the 
larva  to  the  adult.  Its  importance  in  resorption  of  degenerated  cells. 
Varieties  of  phagocytosis.  Giant  cells.  Leucocytosis  in  response  to  the 
presence  of  bacteria.  In  the  peritoneum.  Phagocytosis  in  tuberculosis. 
CHEMOTAXIS.  Botanical  studies.  Early  studies  of  Leber.  Early  studies 
of  Buchner.  Methods.  Theories  of  chemotaxis.  Importance  of  surface 
tension. 

CHAPTER  XII. — PHAGOCYTOSIS,  Continued.     THE  RELATION  OF  THE  LEUKO- 
CYTES AND  OF  PHAGOCYTOSIS  TO  IMMUNITY   .        .        .        .        .    296 

Opsonins  and  tropins.  Metchnikoff 's  attempt  to  establish  parallelism 
between  phagocytosis  and  resistance.  Work  of  his  pupils.  Metchni- 
koff 's  interpretations.  Origin  of  bactericidal  substances  from  leuko- 
cytes. ' '  Macrocytase  "  and  microcytase.  Metchnikoff 's  interpretation 
of  the  Pfeiffer  phenomenon.  Origin  of  alexin.  Leukoeytic  bactericidal 
substances.  Their  nature.  Leukoeytic  ferments.  Leukoprotease.  Pet- 
terson  's  experiments.  Leukoeytic  extract  of  Hiss.  Bordet 's  views. 

CHAPTER  XIII. — PHAGOCYTOSIS,  Continued.    FACTORS  DETERMINING  PHAGO- 
CYTOSIS    H  .        *        .        .        .    311 

Opsonins.  Tropins.  Metchnikoff 's  conception  of  stimulins.  Work  of 
Denys  and  his  pupils.  Other  early  observations.  Work  of  Wright. 
Conception  of  opsonins  definitely  advanced.  Analysis  of  opsonic  action. 
Normal  and  immune  opsonins.  Neufeld's  opinions.  Bacteriotropins, 
Structure  of  opsonins.  Specific  absorption  of  opsonins.  Heat  stability 
of  immune  opsonins.  Relation  to  other  anti-bodies.  Relation  to  alexin. 
Variations  in  leukocytes  as  a  factor  in  opsonic  measurements.  Resistance 
to  opsonic  action  on  the  part  of  bacteria.  Relation  to  virulence. 

CHAPTER  XIV. — PHAGOCYTOSIS,  Continued.     OPSONIC  INDEX  AXD  VACCINE 

THERAPY  .        .        .        .        •        •        ...'.".        .    328 

Wright 's  work  on  typhoid  immunization.  Development  of  technique  for 
measuring  phagocytio  activity.  The  phagocytic  index.  Opsonic  index. 
Dilution  method.  Simon  and  Lamar's  method.  Accuracy  of  opsonic 
index.  Wright 's  work  on  the  staphylococcus  infections.  Relation  of 
opsonic  index  to  clinical  conditions.  Negative  phase.  Summation  of 
negative  phase.  Summation  of  positive  phase.  Clinical  value  of  op- 


xii  CONTENTS 

PAGE 

sonic  index  estimations.  Opsonins  and  tuberculosis.  Treatment  by  auto- 
inoculation.  The  value  of  opsonic  index  determinations.  The  value  of 
vaccine  therapy.  Prophylaxis.  Different  types  of  infection  and  the 
logic  of  vaccine  therapy  in  each  type.  The  production  and  standardiza- 
tion of  vaccines.  The  tuberculins. 

CHAPTER  XV. — ANAPHYLAXIS.    FUNDAMENTAL  FACTS    .        .        .        .    ..  .    358 

The  relation  of  immunity  to  hypersusceptibility.  Various  kinds  of 
hypersusceptibility.  Historical  development  of  our  knowledge  of  these 
phenomena.  The  work  of  Richet  and  others.  The  phenomenon  of 
Arthus.  The  phenomenon  of  Theobald  Smith.  Experimental  produc- 
tion of  the  anaphylactic  state.  Laws  governing  the  condition  as  at 
first  determined.  Symptoms  of  experimental  anaphylaxis  in  guinea 
pigs.  Autopsy  findings  and  causes  of  death.  Changes  in  blood  pres- 
sure. Changes  in  temperature.  Leucopenia.  Diminution  of  comple- 
ment. Symptoms  in  rabbits  and  dogs.  Variation  in  different  animals. 
Anaphylactic  antigen.  Specificity  of  anaphylactic  reaction.  Quantita- 
tive relations.  Variations  depending  upon  method  of  administration. 
Anti-anaphylactic  state.  Prevention  of  anaphylaxis  by  drugs.  Passive 
sensitization.  Conditions  governing  its  accomplishment.  Quantitative 
studies  of  Doerr  and  Russ. 

CHAPTER    XVI. — ANAPHYLAXIS    Continued.      FURTHER    DEVELOPMENT    AND 

THEORETICAL  CONSIDERATIONS 384 

Theory  of  Gay  and  Southard.  Besredka  's  theory.  Gradual  develop- 
ment of  the  antigen-antibody  conception.  Quantitative  work.  Identity 
of  sensitizing  and  toxic  substances.  Idea  of  sessile  receptors.  Ana- 
phylaxis and  precipitins.  The  work  of  Vaughan.  Diminution  of  alexin 
during  anaphylactic  shock.  Toxic  substances  obtained  by  action  of 
active  serum.  Friedberger 's  "anaphylatoxin. "  Obtained  from  pre- 
cipitates. Obtained  from  bacteria.  Is  the  mechanism  of  anaphylaxis 
intravascular  or  cellular?  Precipitins  and  albuminolysins.  Writer's 
opinion.  THE  MECHANISM  OF  ANTI-ANAPHYLAXIS.  Nature  of  ana- 
phylactic poison.  Peptone  shock.  PHENOMENA  CLOSELY  RELATED  TO 
ANAPHYLAXIS.  Toxicity  of  normal  serum.  Toxin  hypersusceptibility. 

CHAPTER  XVII. — ANAPHYLAXIS  Continued.     BACTERIAL  ANAPHYLAXIS  AND 

ITS  BEARING  ON  THE  PROBLEMS  OF  INFECTIOUS  DISEASE        .        .    412 

Early  work  on  sensitization  with  bacterial  protein.  Technique  for 
sensitizing  with  bacteria.  Revision  of  our  ideas  of  "  endo-toxin. " 
Vaughan 's  work  on  toxic  protein  split-products.  Friedberger 's  ana- 
phylatoxin.  Methods  of  production.  Quantitative  proportions  which 
must  be  observed.  Time  and  temperature  conditions.  Bearing  of  this 
work  upon  our  understanding  of  infectious  disease.  Friedberger 's 
interpretation.  Bacterial  toxemia.  Is  the  bacterial  antigen  the  matrix 
for  the  poison? 

CHAPTER  XVIII. — ANAPHYLAXIS  Continued.     THE  CLINICAL  SIGNIFICANCE 

OF  ANAPHYLAXIS *        •        •     428 

Serum  sickness.  Accelerated  reactions  and  immediate  reactions.  Meth- 
ods of  avoiding  anaphylaxis  in  antitoxin  injections.  Anaphylaxis  and 
bacterial  vaccines.  Asthma  and  hay  fever.  Sensitiveness  to  contact 
with  certain  animals.  Possible  anaphylactic  reason  for  eclampsia. 
Sympathetic  ophthalmia.  Diagnostic  reactions.  Tuberculin  reaction. 
Luetin  reaction.  Discussion  of  tuberculin  reaction.  Experimental  ana- 
phylaxis with  tuberculin.  Diagnostic  use  of  anaphylaxis. 

CHAPTER  XIX. — THERAPEUTIC  IMMUNIZATION  IN  MAN    .        ...        *    448 

THERAPEUTIC  USE  OF  DIPHTHERIA  ANTITOXIN.  Statistical  results. 
Amounts  to  be  injected.  Amount  of  antitoxin  normally  present  in  the 


CONTENTS  xiii 

PAGE 

human    blood    serum.      PRACTICAL    CONSIDERATIONS    CONNECTED    WITH 
DIPHTHERIA  ANTITOXIN  PRODUCTION  AND  STANDARDIZATION.     Toxin  pro- 
duction.    LO  and  L+   doses.    Methods  of  determination.      Production  of 
antitoxin.      Standardization    of    antitoxin,    II.   S.    Hygienic    Laboratory 
method.     Chemical  concentration  of  antitoxic  serum.    ACTIVE  IMMUNIZA- 
TION IN  DIPHTHERIA  WITH  MIXTURES  OF  TOXIN  AND  ANTITOXON.    Behr- 
ing's  work.     Use  of  the  method.     Results  obtained.     INTRACUTANEOUS 
METHOD  OF  DETERMINING  TOXIN  AND  ANTITOXIN  VALUES.     Principles  of 
the  method.  Uses.  Application  of  the  method  to  the  determination  of  anti- 
toxin  in   human  beings.     TETANUS   ANTITOXIN   AND   ITS   STANDARDIZA- 
TION.    Determination  of  the  unit.     ANTITOXIN  AGAINST  SNAKE  POISON. 
Calmette  's   work.      Differences   between   cobra   and    rattlesnake   poison. 
Production  of  antiserum.     PASSIVE  IMMUNIZATION  ix  DISEASES  CAUSED 
BY  BACTERIA  WHICH  Do  NOT  FORM  SOLUBLE  TOXINS.    General  considera- 
tion of  principles  involved.     Difficulties.     Serum  treatment  of  epidemic 
meningitis.     Work   of   Kolle   and   Wassermann.     Experiments  of  Joch- 
mann.     Flexner  and  Jobling  's  experiments.     Results.     Present  methods. 
Streptococcus  antiserum.     Differences  between  various  races  of  strepto- 
cocci.    Marmorek 's  serum.     Work  of  Aronson,  Tavel,  Van  de  Velde  and 
others.     Probable   manner  of   action.     Serum  treatment  in  pneumonia. 
Neuf  eld  's  work.    Recent  experiments  and  methods  of  Cole.    Serum  treat- 
ment of  typhoid  fever.     Earlier  experiments.    Attempts  to  produce  anti- 
endotoxin.      Principles   involved.      Immunization   with   trypsin    digested 
bacteria.     Immunization  with  sensitized  bacteria.     Prospects  of  success. 
Serum  treatment  of  plague.     Yersin's  attempts.     Kolle  and  Martini's 
serum.     Work  of  British  Plague  Commission.     Lustig's  serum.     General 
results  obtained.     FACTS  CONCERNING  ACTIVE  PROPHYLACTIC  IMMUNIZA- 
TION IN  MAN.     General  principles.     Typhoid  and  Paratyphoid  vaccina- 
tion.    Earlier  history.     Work  of  Wright,  Kolle,  and  others.     Russell's 
report  of  vaccination  in  the  United  States  army.     Statistics.     Work  of 
Metchnikoff  and   Besredka.     Therapeutic  vaccine  treatment  in  typhoid 
fever.     Prophylactic  immunization  against  cholera.     Methods.     Results. 
Plague  vaccination.     Difficulties.     Methods.     Results.    Smallpox  vaccina- 
tion.    Rabies.     Principles  and  methods  of  application.     INFECTION  AND 
IMMUNITY    IN    POLIOMYELITIS.      Epidemiology.      Relation    to    Animals. 
IMMUNITY    IN    SYPHILIS.      Superinfection.      Transmission    to    Animals. 
Infection  without  Disease.    INFLUENCE  OF  INJECTIONS  OF  NON-SPECIFIC 
SUBSTANCES  UPON  INFECTIOUS  DISEASES. 

CHAPTER  XX. — SERUM  ENZYMES.  LEUKOCYTIC  ENZYMES.  ABDERHALDEN 
REACTION.  PHYSICAL  PRINCIPLES  IN  SERUM  REACTION.  MEIO- 
STAGMIN  AND  EPIPHANIN  REACTIONS.  COLLOIDAL  GOLD  REACTION.  523 

CHAPTER  XXI. — COLLOIDS,  by  Professor  Stewart  W.  Young,  Stanford  Uni- 
versity, California  .  .  .  .  ...  .  .  .  543 

Introduction.  Definition.  Reversible  and  irreversible  colloids.  Sta- 
bility of  colloidal  systems.  Physical  properties  of  colloids.  Form  and 
size.  Osmotic  pressure.  Rate  of  settlement.  Brownian  movement. 
Electrical  properties  of  colloids.  Surface  tension.  Chemical  properties 
of  colloids.  Flocculation  of  colloids  by  electrolytes.  Salts  and  acid 
electrolytes.  Influence  of  concentration.  Difference  in  sensitiveness  to 
electrolytes.  Explanation  of  phenomenon.  The  "zone-phenomenon." 
Mutual  reactions  of  colloids.  Mutual  flocculation.  Protective  action. 
Theories  of  interaction.  The  preparation  of  colloid  solutions.  Applica- 
tions to  biology.  Living  tissues  as  colloids.  Agglutination  of  bacteria. 
Analogy  to  colloid  phenomenon.  Electrical  charge  Carried  by  bacteria. 
Sensitiveness  to  light.  Danysz  phenomena.  Conclusions. 


INFECTION  AND  RESISTANCE 

CHAPTER    I 

INFECTION  AND  THE  PROBLEM  OF  VIRULENCE 

THE  early  history  of  our  knowledge  of  infectious  disease  is  that 
of  fermentation.  It  was  a  philosopher,  Robert  Boyle,  writing  in  the 
17th  century,  who  prophesied  that  the  problem  of  infectious  dis- 
ease would  be  solved  by  him  who  elucidated  the  nature  of  fermenta- 
tion. His  prediction  was  fulfilled  200  years  later  by  the  train  of 
investigations  begun  by  Cagniard-Latour  and  by  Schwann,  and  car- 
ried to  a  brilliant  culmination  by  Pasteur.  It  was  the  discovery  of 
the  living  nature  of  ferments  and  the  specific  nature  of  the  various 
micro-organisms  which  caused  the  several  forms  of  fermentation, 
and  especially  of  putrefaction,  which  made  possible  rational  investi- 
gations in  the  field  of  infectious  disease  and  led  by  analogy,  first  to 
logical  speculation — then  to  actual  experimental  proof  of  the  etiolog- 
ical  relationship  between  the  minute  forms  of  life  and  the  com- 
municable diseases. 

It  is  not  much  more  than  50  years  since  Pollender  described  the 
anthrax  bacillus  in  the  blood  and  spleens  of  animals  dead  of  this 
disease.  In  this  short  period  the  large  number  of  maladies  of  ani- 
mals and  human  beings  caused  by  micro-organisms  belonging  both 
to  the  varieties  spoken  of  as  bacteria  and  to  those  classified  as 
protozoa  has  necessitated  the  segregation  of  this  branch  of  knowl- 
edge into  a  separate  chapter. 

The  period  of  etiological  investigation  is  now  approaching  its 
maturity.  The  causative  agents  of  most  of  the  more  common  infec- 
tious diseases  have  been  discovered,  and  the  biology  of  many  of  the 
pathogenic  micro-organisms  has  been  thoroughly  studied  both  in 
their  artificial  cultures  and  in  the  infected  animal  body.  In  spite 
of  a  considerable  accumulation  of  facts,  however,  the  science  of 
immunity,  that  is,  the  study  of  the  defensive  powers  of  the  living 
animal  body  against  infection,  is  still  in  its  infancy,  and  the  practi- 
cal therapeutic  successes  based  on  this  science  are  disappointingly  out 
of  proportion  to  the  really  large  amount  of  detailed  knowledge  of 
cellular  and  serum  reactions  at  our  disposal. 

The  study  of  putrefaction  and  of  fermentation — though  furnish- 
ing the  basic  analogy  from  which  the  first  impulse  was  obtained — 

1 


2      A  .  ... :t  :  .INFECTION    AND    RESISTANCE 

presented  after  all  a  problem  infinitely  more  simple  than  that  of  the 
infection  of  living  tissues  with  bacteria.  For,  given  any  organic 
material  containing  suitable  nutritive  constituents,  with  favorable 
environmental  conditions  of  moisture  and  temperature,  and  spon- 
taneously or  experimentally  inoculated  with  germs  of  a  proper 
species,  and  the  phenomena  which  ensued  were  merely  those  of  bac- 
terial growth,  in  which  an  active  part  was  played  by  the  bacteria 
only,  the  dead  organic  materials  serving  simply  as  a  passive  men- 
struum for  these  activities. 

During  the  earlier  days  of  the  development  of  bacteriology, 
therefore,  when  the  attention  of  investigators  was  concentrated  prij 
marily  upon  the  discovery  of  the  specific  causal  agents  of  various  in- 
fectious diseases,  it  seemed  that  the  simple  bringing  together  of 
pathogenic  germ  and  susceptible  subject  should  suffice  for  the  ac- 
complishment of  an  infection.  We  have  learned,  however,  that  the 
process  is  much  more  involved,  and  that,  fortunately  for  the  sur- 
vival of  the  higher  animals  and  man,  the  conditions  which  determine 
infection  are  intimately  dependent  upon  a  variety  of  secondary 
modifying  factors. 

Throughout  nature  bacteria  are  abundant,  and  the  environment 
of  man  and  animals,  the  outer  integuments  of  skin  and  hair,  and  the 
mucous  membranes  of  the  conjunctive,  the  intestinal  and  respira- 
tory tracts,  are  constantly  inhabited  by  a  thriving  bacterial  flora. 
The  distribution  of  certain  species  in  definite  localities  is  often  suffi- 
ciently constant  to  be  regarded  as  a  normal  condition.  Thus  the 
Bacillus  xerosis  is  a  characteristic  inhabitant  of  the  conjunctiva, 
certain  cocci  and  spirilla  are  always  present  in  the  mouth  and 
pharynx,  as  is  Doderlein's  bacillus  in  the  vagina.  The  fact  that 
bacilli  of  the  colon  group  are  invariably  present  in  the  bowels  of  ani- 
mals and  man  from  the  first  few  days  or  hours  after  birth  has  even 
been  interpreted  by  some  investigators  as  a  physiologically  beneficial 
condition.  In  the  course  of  ordinary  existence,  therefore,  and  much 
more  so  during  the  course  of  accidental  exposure  to  individuals  in 
whom  infection  is  present,  the  bodies  of  the  higher  animals  are  in 
intimate  contact,  not  only  with  ordinarily  harmless  bacteria  (sapro- 
phytes), but  also  with  many  varieties  of  the  micro-organisms  spoken 
of  as  "pathogenic"  or  disease-producing.  Perfectly  normal  indi- 
viduals have,  then,  on  occasion,  been  found  to  harbor  diphtheria 
bacilli  in  nose  and  pharynx,  meningococci  have  been  found  in  simi- 
lar localities,  and  tetanus  bacilli,  the  bacillus  of  malignant  edema, 
the  Welch  bacillus,  and  other  distinctly  pathogenic  germs  have  been 
isolated  from  the  intestinal  contents  of  individuals  who  showed  no 
evidence  of  disease.  In  fact,  the  problem  of  the  so-called  bacillus  car- 
riers— persons  who,  though  themselves  apparently  well  for  the  time 
being,  harbor  within  their  bodies  and  distribute  to  their  environ- 
ment bacteria  capable  of  causing  disease  in  others — is,  as  we  shall 


THE    PROBLEM    OF    VIRULENCE  8 

see,  now  recognized  as  one  of  the  most  important  difficulties  of  sani- 
tary prophylaxis.  In  the  case  of  typhoid  fever  this  is  particularly 
true,  for  it  is  now  well  known  that  a  perfectly  healthy  individual  may 
harbor  typhoid  bacilli  in  the  gall-bladder  for  years  and  constitute, 
through  all  this  time,  a  constant  focus  of  danger  to  the  public  health. 

The  accomplishment  of  an  infection,  then,  is  not  determined 
merely  by  the  fact  that  a  micro-organism  of  a  pathogenic  species 
finds  lodgment  in  or  upon  the  body  of  a  susceptible  individual,  but 
it  is  further  necessary  that  the  invading  germ  shall  be  capable  of 
maintaining  itself,  multiplying  and  functionating  within  the  new 
environment.  An  infection,  then,  or  an  infectious  disease,  is  the 
product  of  the  two  factors,  invading  germ  and  invaded  subject,  each 
factor  itself  influenced  by  a  number  of  secondary  modifying  circum- 
stances, and  both  influenced  materially  by  such  fortuitous  conditions 
as  the  number  or  dose  of  the  infecting  bacteria,  their  path  of  en- 
trance into  the  body,  and  the  environmental  conditions  under  which 
the  struggle  is  maintained. 

We  have  in  truth,  then,  a  battle  of  two  opposed  forces,  the  result 
of  which  is  infectious  disease.  And  it  is  the  systematic  analysis  of 
these  forces  in  their  variable  conditions,  and  the  laws  which  govern 
them,  which  constitutes  the  science  of  immunity.  It  is  the  initial 
skirmish  between  the  two  which  determines  whether  or  not  a  foot- 
hold shall  be  gained  upon  the  body  of  the  subject  and  an  infection 
thus  established,  and  it  is  the  balance  between  them  which  decides 
the  eventual  outcome  of  recovery  or  death.  And  though  it  is  un- 
fortunately true  that  much  of  the  knowledge  gained  by  such  studies 
has  yielded  no  direct  therapeutic  results,  the  facts  that  have  been 
revealed  are  fundamental  to  the  pathology  of  infectious  disease  and 
as  essential  to  the  clinical  understanding  of  these  maladies  as  is  the 
knowledge  of  the  mechanism  of  the  circulation,  the  chemistry  of 
metabolism,  or  the  structural  changes  of  the  tissues  to  the  compre- 
hension of  other  pathological  conditions. 

And  from  this  point  of  view  the  study  of  infectious  diseases  can 
be  made  an  eminently  logical  one,  in  that,  knowing  the  criteria 
which  govern  the  infection  of  a  human  being  with  a  given  germ, 
knowing  the  probable  path  of  entrance,  manner  of  distribution,  and 
biological  activities  of  the  micro-organism,  and  the  peculiarities  of 
the  mechanism  of  resistance  set  in  motion  in  the  body  by  this  par- 
ticular infection,  definite  clinical  deductions  can  often  be  made. 

One  of  the  most  fundamental  facts,  immediately  apparent  on 
considering  the  problems  of  infection,  is  the  phenomenon  that  among 
the  innumerable  varieties  of  bacteria  and  protozoa  present  in  nature 
there  is  a  very  limited  group  which  is  capable  of  becoming  parasitic 
upon  the  body  of  higher  animals,  and  among  these  a  still  smaller 
proportion  which  is  capable  of  being  "pathogenic"  or  causing  dis- 
ease. We  have  used  the  terms  pathogenic  and  non-pathogenic  as 


4  INFECTION    AND    RESISTANCE 

practically  synonymous  respectively  with  "parasitic"  and  "sapro- 
phytic." But,  as  we  shall  see,  although  as  a  rule  a  micro-organism 
must  be  parasitic  to  possess  pathogenic  powers,  some  of  the  true 
saprophytes  or  so-called  half-saprophytes  may,  be  pathogenic  under 
certain  conditions,  and  the  terms  do  not  cover  each  other  absolutely. 
It  is  reasonable  to  suppose  that  all  micro-organisms  were  origi- 
nally in  the  condition  which  we  designate  by  the  term  "saprophytic." 
By  this  term  we  imply  that  these  germs  maintain  themselves  only 
upon  dead  organic  matter  and  do  not  thrive  in  or  upon  the  living 
animal  tissues.  The  class  of  saprophytes  is  widely  distributed  and 
constitutes,  of  course,  the  most  important  group  of  bacteria  in  na- 
ture, since  upon  the  activities  of  these  germs  depends  the  unlocking 
of  nitrogen  and  carbon  from  the  organic  complexes  in  the  dead  bod- 
ies and  waste  products  of  animals  and  plants.  Such  bacteria  if 
strictly  saprophytic,  that  is,  entirely  unable  to  maintain  themselves 
upon  living  tissues,  have  little  importance  as  producers  of  disease, 
or,  expressed  in  technical  terms,  have  little  "pathogenicity."  Nev- 
ertheless, there  are  cases  in  which  strict  saprophytes  may  cause  dis- 
ease by  lodging  upon  and  growing  in  animal  tissues  which  have 
been  killed  by  other  causes,  so-called  necrotic  areas ;  and  these,  still 
being  in  relation  with  the  body  as  a  whole  through  the  blood  and 
lymph  channels,  furnish  an  area  of  saprophytic  growth  from 
which  products  of  putrefaction  or  even  bacterial  poisons  may  be 
absorbed.  While,  as  a  rule,  the  disease  following  the  invasion 
of  necrotic  tissue — such  as  gangrenous  amputation  stumps,  old 
unhealed  sinuses,  diabetically  gangrenous  areas,  etc.,  may  be  caused 
by  a  large  variety  of  saprophytic  bacteria,  there  are  a  few  very 
important  and  specifically  pathogenic  bacteria  which  are,  strictly 
speaking,  saprophytes.  Thus  tbe  form  of  meat  poisoning  caused 
by  the  Bacillus  botulinus  is  due  entirely  to  the  poison  formed  by 
this  bacillus  outside  of  the  body  within  the  substance  of  the  dead 
foodstuff,  and  disease  ensues  as  the  result  of  subsequent  ingestion  of 
this  poison  with  the  food.  In  the  same  way  the  tetanus  bacillus  and, 
less  strictly  speaking,  the  diphtheria  bacillus,  at  least  in  its  ordinary 
mode  of  attack,  are  rather  closer  to  the  class  of  saprophytes  than  to 
that  of  the  parasites,  since  neither  of  these  bacteria,  under  usual 
•circumstances,  invades  the  substance  of  the  tissues  beyond  the  point 
•of  initial  lodgment,  causing  disease  only  by  the  production  of 
specific  poisons,  a  condition  known  as  "toxemia"  or  intoxication. 
The  tetanus  bacillus,  moreover,  is  not  usually  capable  of  maintain- 
ing itself  and  multiplying  even  at  the  point  of  initial  lodgment 
unless  the  tissues  have  been  injured  by  trauma  or  irritated  by  the 
presence  of  foreign  bodies.  Bacteria  of  such  characteristics,  there- 
fore, though  pathogenic — that  is,  incitant  of  disease — remain  never- 
theless essentially  saprophytes  living  upon  the  dead  animal  tissues, 
not  invading  the  living  cells  or  body  fluids.  It  is  true  that  investi- 


THE    PROBLEM    OF    VIRULENCE  5 

gations  of  Frosch  1  have  shown  that  diphtheria  bacilli  may  often  be 
found  in  blood  and  organs  of  diphtheritic  patients,  and  tetanus 
bacilli  have  occasionally  been  found  in  the  spleen.  However,  such 
distribution  is  not  necessary  for  the  production  of  disease  by  these 
bacteria,  and  the  essential  point  remains  that  they  may  cause  violent, 
often  fatal,  disease  without  truly  departing  from  their  saprophytic 
mode  of  life  upon  dead  tissues.  Between  the  saprophytes  and  the 
true  parasites  or  invaders  of  living  tissue  many  transitions  occur, 
and  the  condition  of  parasitism  is  probably  a  form  of  specific  adap- 
tation. 

How  such  transition  may  be  biologically  developed  is  probably 
well  illustrated  by  the  investigations  of  Italian  bacteriologists  upon 
tetanus  bacilli.2  Tarozzi  3  inoculated  guinea  pigs  and  rabbits  with 
tetanus  spores  subcutaneously  and  found  that  these  spores  were 
rapidly  transported  to  the  liver,  spleen,  and  kidneys,  where  they 
could  maintain  a  latent  existence  for  as  long  as  51  days.  If  during 
this  period  trauma  or  any  injury  of  the  organs  was  practiced  which 
led  to  the  formation  of  necrotic  tissue  the  spores  would  develop  upon 
this  basis  and  cause  acute  or  chronic  tetanus.  Canfora,4  continuing 
these  studies,  likewise  found  that  tetanus  spores  inoculated  under  the 
skin  are  rapidly  distributed  throughout  the  circulation.  If  no 
trauma  has  taken  place  at  the  point  of  inoculation  the  locally  lodged 
spores  may  be  rapidly  destroyed,  probably  by  phagocytosis.  In  the 
circulation  they  appear  to  be  less  rapidly  eliminated  and  may  be 
present  for  from  ten  to  thirteen  days.  If,  during  this  period,  there 
is  produced  a  small  wound,  blood  clot,  or  necrotic  area  in  the  body — 
this  may  serve  as  a  focus  for  development  and  tetanus  may  ensue. 
After  ten  or  more  days  the  spores  disappear  from  the  blood,  but  may 
then  take  up  a  latent  existence  in  some  of  the  organs — as  stated  by 
Tarozzi.  Apart  from  their  importance  as  constituting  a  sort  of 
.  transitional  condition  between  pure  saprophytism  and  parasitism, 
these  investigations  would  seem  to  have  much  bearing  upon  the  so- 
called  cases  of  acryptogenic  tetanus." 

True  infection,  that  is,  the  invasion  of  one  species  by  individuals 
of  another,  and  the  ability  of  the  latter  to  multiply  and  functionate 
within  the  cell  complexes  of  the  former,  is  a  process  quite  out  of 
keeping  with  the  ordinary  plans  of  nature,  throughout  which  there 
seems  to  be  a  distinct  opposition  to  the  colonization  and  functiona- 
tion  of  one  living  being  within  the  living  substance  of  another. 
Thus,  as  Bail  5  has  pointed  out,  a  mass  of  frogs'  eggs  will  remain 

1  Froseh.     Zeitsdir.  f.  Hyg.,  Vol.  13,  1893. 

2Belfanti,  quoted  from  Canfora,  Centralblt.  f.  Bact.,  I.  Orig.  Vol.  45, 
1908. 

3  Tarozzi.     Centralblt.  f.  Bact.,  Orig.  Vol.  38,  1905. 

4  Canfora.     Centralblt.  f.  Bact.,  Orig.  Vol.  45,  1908. 

5  Bail.     "Das  Problem  der  Bakt.     Infection."    Klinkhardt,  Leipzig,  1911. 


6  INFECTION    AND    RESISTANCE 

entirely  iminvaded  while  alive,  though  the  water  surrounding  it  may 
swarm  with  bacteria  of  many  varieties,  but  when  by  some  accident 
such  a  mass  of  eggs  ceases  to  live,  it  immediately  falls  prey  to  bac- 
terial infection.  The  same  point  is  illustrated  by  the  rapidity  with 
which  intestinal  bacteria  will  spread  throughout  the  body  after 
death,  when  during  life  they  have  remained  confined  to  the  lumen 
of  the  intestine,  or,  at  most,  get  into  the  portal  circulation,  to  be  de- 
stroyed in  the  liver.  By  the  living  cell,  therefore,  an  opposition  is 
offered  to  invasion  by  bacteria,  a  vital  function  which  Bail  has  at- 
tempted to  make  clearer  by  formulating  it  as  a  law,  referring  to  it  as 
"Das  Gesetz  der  Lebensundurchdringlichkeit."  Upon  what  cell  func- 
tion this  vital  resistance  to  invasion  depends  is  to  a  large  extent  a 
mystery.  It  would  seem  to  rest  in  principle  upon  the  fact  that  the 
invading  cell  meets  the  invaded  one  under  conditions  peculiarly 
adapted  to  the  activities  of  the  latter,  and  is  overcome  before  condi- 
tions suitable  for  its  own  activities  have  been  established.  The  con- 
ditions here  are  not  unlike  those  observed  in  the  case  of  digestive 
enzymes,  a  comparison  which  becomes  more  than  an  illustrative  anal- 
ogy when  we  consider  that  apart  from  the  mere  mechanical  disturb- 
ance created  by  the  presence  of  bacteria  as  foreign  bodies  the  struggle 
between  invader  and  tissue  is  largely  one  of  enzyme  against  enzyme. 
Thus,  for  instance,  the  gastric  juice  does  not  act  upon  the  mucous 
membrane  of  the  stomach  during  life — but  after  death,  at  autopsy, 
partial  digestion  of  this  membrane  by  the  pepsin  is  often  seen. 

Whenever  this  vital  resistance  or  opposition  is  overcome,  and 
micro-organisms  enter  the  tissues  or  cells,  an  abnormal  process  is 
taking  place,  and  this  process  is,  strictly  defined,  infection.  Never- 
theless, it  is  by  no  means  necessary  that  such  infection  should  al- 
ways be  accompanied  by  manifestations  of  disease.  It  is  true  that, 
in  most  cases,  the  natural  resistance  is  such  that  a  struggle  ensues 
by  which  the  invader  is  destroyed  or  thrown  off,  or  in  which  the 
invaded  subject  is  functionally  injured  or  even  killed,  and  the  ac- 
companying evidences  of  such  a  struggle  constitute  what  we  know 
as  infectious  disease.  '  But  there  are  special  cases,  cases  of  adapta- 
tion, biologically  speaking,  in  which  neither  invader  nor  host  is  seri- 
ously harmed.6  In  the  field  of  protozoology,  especially,  there  are 
many  examples  of  true  parasites,  that  is,  invaders  truly  maintaining 
their  metabolism  at  the  expense  of  the  tissues  and  body  substances 
of  the  host,  which  do  not  ar,ouse  reactions  sufficiently  vigorous  to  be 
termed  "disease."  Thus  the  Trypanosoma  Lewisi  may  be  found  in 
the  blood  of  rats  7  without  noticeably  affecting  the  health  of  the  ani- 
mals, and  other  protozoa  have  similarly  been  found  in  organs  and 
blood  stream  of  a  number  of  other  apparently  healthy  animals.  Al- 
though such  conditions  have  been  frequently  spoken  of  as  "infection 

6  See  also  Bail,  loc.  cit. 

7  Doflein.    "Die  Protozoen  ais  Krankheitserreger." 


THE    PROBLEM    OF    VIRULENCE  7 

without  infectious  disease,"  the  distinction  is  probably  one  of  degree 
only — there  being  some  reaction  on  the  part  of  the  host  even  in  the 
mildest  cases,  if  only  in  the  weakening  by  withdrawal  of  body  sub- 
stance, which  distinguishes  the  infected  from  the  uninfected  animal. 
In  other  cases  there  may  even  be  advantage  to  the  host,  following  the 
infection,  to  the  detriment  of  the  invading  micro-organism,  a  phe- 
nomenon most  clearly  illustrated  by  the  invasion  of  the  root  hairs  of 
leguminous  plants  by  the  Nitrogen-fixing  " root-tubercle"  bacilli,  a 
condition  in  which,  as  Fischer  says,  the  plant  may  be  regarded  as 
parasitic  upon  the  bacteria. 

The  actual  harm  resulting  from  the  infection  must,  to  a  large 
extent,  depend  upon  the  degree  of  adaptation  to  the  new  conditions 
of  life  possible  on  the  part  both  of  the  invader  and  of  the  host.  If 
the  invader  can  acquire  resistance  to  the  defensive  properties  of  the 
host,  and  the  latter  can  be  similarly  adapted  to  the  harmful  effects  of 
the  invader,  a  prolonged  condition  of  infection  might  ensue,  a  sort 
of  truce  without  manifestations  of  the  disease.  Although  this  is  con- 
ceivable, such  mutual  adaptation  is  probably  very  rare  in  human 
disease. 

In  cases  of  so-called  chronic  septicemia  in  which  bacteria  may 
be  again  and  again  isolated  by  blood  culture  from  the  circulation  it 
is  more  than  likely  that  the  organisms  are  constantly  present,  not 
because  they  multiply  or  maintain  themselves  within  the  circulation, 
but  rather  because  they  are  being  continuously  discharged  into  the 
blood  from  an  established  focus  in  the  tissues — as,  for  instance,  on  a 
heart  valve.  We  have  examined  the  serum  of  patients  with  subacute 
and  chronic  septicemia  (endocarditis),  and  often  found  powerful 
opsonic  action  against  the  invading  germs  even  when  the  patient's 
own  serum  and  leukocytes  were  used  in  the  tests,  evidence  that  the 
bacteria  were  probably  being  successfully  disposed  of  after  they  had 
gained  entrance  into  the  blood  stream.  In  rabbits,  too,  in  our  ex- 
perience and  in  that  of  Miss  Gilbert  of  this  laboratory,  it  would 
seem  that  protracted  septicemia  is  present  only  when  secondary  foci 
have  been  established  from  which  the  bacteria  are  constantly  being 
discharged  into  the  blood.  This  we  believe  is  rather  the  rule  and 
the  establishment  of  a  balance  within  the  blood  stream  an  exception. 
When  bacteria  do  succeed  in  withstanding  successfully  the  opposing 
forces  active  within  the  circulating  blood  their  rapid  accumulation, 
the  collapse  of  the  defensive  mechanism,  and  death  of  the  patient  are 
probably  the  most  common  course. 

The  point  of  view  which  we  have  expressed  in  the  preceding 
paragraph  has  been  impressed  upon  us  with  particular  insistence  by 
the  observation  of  certain  cases  of  bacteriemia  following  infections  of 
the  middle  ear,  mastoid  processes,  and  thromboses  of  adjacent  veins. 
In  such  cases  it  appears  that  the  blood  may  be  flooded  with  bacteria 
which,  nevertheless,  disappear  after  the  focus  of  infection  has  been 


8  INFECTION    AND    RESISTANCE 

removed.  We  have  recently  had  the  opportunity  to  observe  this  in 
a  case  of  septicemia  caused  by  Streptococcus  mucosus,  in  which  blood 
culture  plates  showed  very  numerous  colonies,  and  in  which  recovery 
followed  promptly  upon  complete  excision  of  the  thrombosed  veins. 
It  would  seem  to  us,  therefore,  that  bacteriemia  offers  a  rather  better 
prognosis  than  was  formerly  supposed,  at  least  in  cases  in  which  the 
focus  is  surgically  accessible. 

The  same  principle  is  illustrated  in  the  ordinary  clinical  course  of 
typhoid  fever  in  the  human  being.  Here  the  disease  begins  as  a 
bacteriemia.  Very  rapidly,  usually  within  two  weeks,  the  bacteria 
disappear  from  the  blood  stream  and  a  high  serum  immunity  is 
established  in  the  patient.  Nevertheless,  the  bacteria  remain  actively 
growing  within  definite  foci  in  the  tissues,  where  they  are  to  a  cer- 
tain extent  protected  or  inaccessible  to  the  defensive  powers  so  suc- 
cessfully active  in  the  blood  stream.  At  any  rate  the  patient  remains 
diseased  and  the  bacteria  can  be  isolated  from  the  spleen,  gall-blad- 
der, and  intestines  at  a  stage  when  they  are  no  longer  present  in  the 
blood  stream,  and  during  which  a  measurement  of  the  bactericidal 
and  opsonic  powers  of  the  patient  will  reveal  a  serum  immunity 
much  higher  than  normal.  Just  why  the  organisms  are  protected 
from  these  influences  in  the  tissues  we  do  not  know. 

On  the  other  hand,  it  is  nevertheless  true  that  a  certain  amount 
of  actual  adaptation  between  the  bacteria  and  the  body  may  take 
place  and  contribute  to  the  chronicity  of  an  infection.  This  seems 
to  be  shown  especially  by  the  experiments  of  Walker  and  others, 
which  are  referred  to  in  other  places,  in  which  it  was  found  that  bac- 
teria grown  on  immune  sera  gain  a  certain  amount  of  resistance 
against  the  injurious  properties  of  these  substances,  and  evidence 
more  directly  bearing  upon  the  question  is  furnished  by  the  studies 
on  the  typhoid  carrier  state  in  rabbits  made  by  Chirolanza,8  Black- 
stein,9  Johnston,  10  and  recently  by  Gay  and  Claypole.11  The  last- 
named  writers  found  that  they  could  regularly  produce  the  typhoid 
carrier  state  in  these  animals  if  they  first  cultivated  the  typhoid 
bacilli  upon  a  medium  containing  defibrinated  rabbits'  blood.  Even 
in  these  cases,  however,  it  is  not  at  all  improbable  that  the  typhoid 
bacilli  establish  a  permanent  focus  from  which  they  are  discharged 
into  the  blood  stream. 

An  infectious  disease,  therefore,  may  be  interpreted  as  the  result 
of  parasitism  in  which  no  such  mutual  adaptation  has  taken  place, 
and  in  which  the  invasion  of  the  host  by  the  micro-organism  is 
marked  by  a  struggle,  the  local  and  systemic  manifestations  of  which 
constitute  the  disease.  The  disease  is  an  evidence  of  conflict  be- 

8  Chirolanza.     Ztschr.  f.  Hyg.,  Vol.  62,  1909. 

9  Blackstein.     Bull.  Johns  Hop.  Hosp.,  1891. 

10  Johnston.     Journ.  Med.  Res.,  27,  1912. 

11  Gay  and  Claypole.    Arch,  of  Int.  Med..,  12,  1913. 


THE    PROBLEM    OF    VIRULENCE  9 

tween  the  two  forces,  mild  and  locally  limited  if  the  protective  pow- 
ers far  outweigh  the  invasive  powers  of  the  micro-organisms,  violent 
if  the  balance  is  reversed.  This  conception  is  probably  a  correct  one 
in  the  case  of  the  large  majority  of  diseases — those  in  which  invasion 
is  accompanied  by  more  or  less  rapid  and  violent  inflammatory  and 
other  reactions.  In  diseases  like  leprosy,  tuberculosis,  and  a  few 
others  of  the  more  chronic  infections  it  is  also  possible  that  extensive 
invasion  of  the  body  depends,  not  so  much  upon  the  active  invasive 
powers  of  the  micro-organism,  powers  which  we  will  attempt  to 
analyze  presently,  but  rather  upon  the  fact  that  for  reasons  of  in- 
solubility and  lack  of  irritating  properties  on  the  part  of  the  invader 
no  reaction  is  set  up  at  first,  and  the  invasion,  though  progressive, 
elicits  no  violent  symptoms  and  no  energetic  opposition.  The  in- 
vader therefore  progresses  unopposed,  becoming  an  incitant  of  dis- 
turbed bodily  functions  to  the  degree  of  actual  disease  only  when  it 
has  gained  a  foothold  in  some  organ  and  begun  to  proliferate,  or 
has  multiplied  in  such  numbers  that  the  cumulative  effect  of  its 
toxic  powers  becomes  manifest. 

Such  a  conception  would  assign  the  slow  and  gradual  but  pro- 
gressively invasive  powers  of  such  diseases  as  tuberculosis,  leprosy, 
and  syphilis  in  which  systemic  symptoms  are  manifest  only  after 
the  disease  has  gained  an  extensive  foothold,  to  the  lack  of  acute 
physiological  reaction  resulting  from  the  presence  of  the  invading 
micro-organism.  In  the  case  of  such  infections  as  those  caused  by 
some  of  the  yeasts  or  blastomyces  we  have  seen  foci  of  blastomycotic 
lodgment  in  the  kidney  and  other  organs  surrounding  which  there 
was  neither  an  accumulation  of  mobile  cells — (leukocytes  or  lympho- 
cytes)— nor  any  evidence  of  cloudy  swelling  or  other  injury,  by 
poisons,  of  adjacent  parenchyma  cells.  Here,  as  in  tuberculosis  01 
leprosy,  the  reaction  induced  by  the  presence  of  the  micro-organisms 
is  slow  and  gradual — expressed  in  an  eventual  fixed  tissue-cell  reac- 
tion and  giant-cell  formation — similar  to  that  induced  by  insoluble 
foreign  bodies.  And  it  may  well  be  that  the  progressive  ability  to 
multiply  without  arousing  the  invaded  body  to  rapid  and  powerful 
reaction  may  account  for  the  .prolonged  period  of  apparent  well- 
being  in  the  early  stages  of  such  infections  and  permit  the  invaders 
to  pervade  the  body  so  extensively. 

This  point  of  view  has  been,  we  believe,  most  clearly  expressed 
by  Theobald  Smith.12  Bacteria  may  lack  invasive  or  pathogenic 
properties  and  be,  therefore,  immediately  destroyed  after  gaining 
entrance  to  the  host.  They  may  be  powerfully  invasive  and  because 
of  lack  of  adaptation  arouse  a  violent  defensive  reaction  on  the  part 
of  the  host.  "There  is  another  type  of  parasite,"  Smith  says,  "which 
may  dispense  largely  with  both  offensive  and  defensive  processes. 
We  can  conceive  of  this  type  as  exerting  a  metabolic  activity  approx- 
12  Theobald  Smith.  Journ.  of  A.  M.  A.,  May,  1913,  Vol.  60. 


10  INFECTION    AND    RESISTANCE 

imating  so  closely  to  that  of  the  host  that  the  latter  reacts  but  slightly 
and  then  only  after  a  long  period  of  stimulation."  Into  this  class 
he  places  the  syphilis  spirochseta  and,  in  a  somewhat  modified  sense, 
the  tubercle  bacillus. 

We  have  seen,  then,  that  a  micro-organism  may  be  pathogenic 
and  still  be  saprophytic  in  its  mode  of  life.  In  order  that  this  can 
occur,  however,  it  is  necessary  that  it  should  possess  the  power  of 
producing  at  the  place  of  lodgment  a  poison  or  toxin  which  can  be 
absorbed  and  cause  disease.  The  condition  wrhich  ensues  is  not,  prop- 
erly speaking,  an  infection,  but  rather  a  "toxemia"  differing  from 
the  toxemias  resulting  from  the  ingestion  of  drugs  or  other  poisons 
only  in  so  far  as  the  toxins  are  manufactured  at  some  point  of  bac- 
terial lodgment  within  the  body  of  the  victim.  Typical  tetanus  and 
diphtheria,  for  instance,  can  be  produced  as  readily  by  ingestion  of 
the  bacteria-free  culture  filtrates  as  by  inoculation  with  the  bacteria 
themselves.  And  although  these  bacteria  may,  on  occasion,  become 
invasive  and  thereby  satisfy  the  criteria  of  true  infection,  this  is  not 
necessary  for  their  pathogenicity. 

In  the  large  majority  of  bacterial  diseases,  however,  it  is  neces- 
sary that  the  germs  shall  be  capable  of  producing  a  true  infection 
before  they  can  become  pathogenic,  and  it  is  our  task  therefore  to 
attempt  to  analyze  those  bacterial  attributes  upon  which  the  invasive 
power  or  virulence  may  be  said  to  depend. 

In  the  realm  of  infectious  micro-organisms  a  wide  range  of  cul- 
tural variations  is  encountered  which  indicates  that  some  of  these 
germs  have  adapted  themselves  very  closely  to  the  specific  environ- 
mental conditions  found  in  the  living  animal  body,  while  others  can 
take  up  with  ease  and  under  the  simplest  cultural  conditions  a  purely 
saprophytic  existence. 

Many  pathogenic  micro-organisms  have  so  far  defied  all  attempts 
at  cultivation  in  artificial  media.  These  we  cannot  use  for  examples 
since  it  may  well  be  that  the  failure  of  attempts  in  many  of  them 
may  hinge  upon  such  simple  alterations  of  method  as  the  exclusion 
of  oxygen,  the  addition  of  fresh  tissue,  or  the  supplying  of  amino- 
acids,  which  have  made  possible  the  cultivation  of  the  spirochaeta 
pallida  and  the  leprosy  bacillus.  But  among  those  which  we  can 
cultivate  there  are  many  which  require  for  successful  cultivation 
the  production  of  artificial  conditions  simulating  closely  those  ob- 
taining in  the  living  body.  Thus  malarial  plasmodia  can  be  made  to 
multiply  only  if  furnished  with  uninjured  human  red  blood  cells, 
within  which  they  can  develop.  The  gonococcus  requires,  in  its 
first  cultures  outside  the  body,  a  medium  containing  human  protein ; 
and  the  hemophile  bacteria,  among  them  the  influenza  bacillus,  re- 
quire hemoglobin.  Other  organisms  like  pneumococci,  many  strep- 
tococci, diphtheria  bacilli,  and  many  others,  though  easily  grown  on 
artificial  media,  are  still  fastidious  in  their  requirements  and  develop 


THE    PROBLEM    OF    VIRULENCE  11 

sparsely  or  not  at  all  unless  definite  conditions  of  nutrient  materials, 
temperature,  reaction,  and  osmotic  pressure  are  observed.  On  the 
other  hand,  typhoid,  anthrax,  and  dysentery  bacilli,  staphylococci 
and  numerous  other  pathogenic  germs  grow  easily  and  luxuriantly 
on  the  simplest  laboratory  media  and  within  a  wide  range  of  envi- 
ronmental variations. 

Biologically  considered,  we  could  arrange  the  scale  of  adaptation 
to  parasitic  conditions  on  this  basis  and  it  would  seem,  a  priori,  that 
those  bacteria  which  had  thus  adapted  themselves  most  closely  to  the 
living  body  should  be  the  most  infectious.  There  is  not,  however, 
such  parallelism,  since  many  of  the  most  powerfully  invasive  or 
virulent  germs,  for  instance,  the  anthrax  bacillus,  have  retained 
their  capacity  for  saprophytic  life  to  the  fullest  extent.  It  is  more 
logical,  therefore,  to  classify  parasites,  not  according  to  their  ability 
to  revert  to  saprophytic  conditions,  but  rather,  as  Bail 13  has  done  it, 
on  the  basis  of  their  relative  powers  of  invading  the  living  body. 
His  classification,  of  course,  implies  that  the  position  of  each  micro- 
organism in  this  scale  must  be  determined  with  reference  to  a  given 
animal  species,  since  a  germ  which  is  highly  infectious  ("parasitic" 
in  Bail's  sense)  for  one  species  may  be  a  "half -parasite"  or  even  a 
pure  saprophyte  for  another. 

Briefly  reviewed,  his  classification  is  as  follows: 

I.  Pure  Saprophytes. — (Necroparasites,  superficial  parasites,  or 
external  parasites.) 

Micro-organisms  which  under  no  circumstances  can  be  made  to 
develop  within  the  living  tissues  of  a  given  animal.  This  does  not 
exclude  their  pathogenicity  for  this  animal,  since,  like  the  diphtheria 
or  tetanus  bacillus,  they  may  develop  and  produce  toxins  on  the 
basis  of  a  localized  area  of  dead  tissues. 

II.  Pure  Parasites. — Organisms  like  the  anthrax  bacillus  or 
the  bacilli  of  the  hemorrhagic  septicemia  group  which,  implanted  in 
small  quantity  in  an  animal,  will  rapidly  gain  a  foothold,  thrive, 
and  spread  throughout  the  body. 

III.  Half  parasites,  organisms  which  may  be  infectious  if  in- 
troduced into  the  animal  body,   but,  not  possessing  this   invasive 
power  to  the  same  degree  as  the  preceding  class,  require  the  inocula- 
tion of  considerable  quantities,  often  a  special  mode  or  path  of  in- 
oculation, or  even  possibly  a  preliminary  reduction  of  the  local  and 
general  resistance  of  the  infected  individual  in  order  that  they  may 
multiply  and   become  generalized.      This  class  includes  the  large 
majority  of  the  bacteria  pathogenic  for  man. 

This  property  of  invasive  power  is  spoken  of  as  virulence  in 
contradistinction  to  toxicity — the  latter  implying  merely  the  abil- 
ity to  produce  poisons,  and  not  necessarily  being  associated  with  the 
power  to  invade. 

13  Bail.      LOG.   cit. 


18  INFECTION    AND    RESISTANCE 

In  order  that  a  micro-organism  may  be  a  true  parasite  in  Bail's 
sense — or  invasive — for  any  given  species  of  animal  it  must  of 
course  possess  certain  basic  cultural  attributes  which  enable  it  to 
grow  in  the  environment  furnished  by  the  host.  For  instance,  a 
micro-organism  which  does  not  grow  at  temperatures  below  37.5°  C. 
cannot  very  well  become  parasitic  upon  cold-blooded  animals.  An 
excellent  illustration  of  this  influence  of  body  temperature  upon 
the  invasive  powers  of  bacteria  is  furnished  by  the  different  races 
of  acid-fast  bacilli  which  invade  the  bodies  of  man  and  of  birds. 
The  avian  tubercle  bacillus,  for  instance,  is  non-pathogenic  for  man 
and  in  cultures  will  not  develop  at  temperatures  below  40°  C., 
which  is  about  the  body  temperature  of  most  birds.  The  human 
tubercle  bacillus,  on  the  other  hand,  is  non-pathogenic  for  birds  and 
ceases  to  grow  in  artificial  cultures  when  the  temperature  is  raised 
above  40°  to  41°  C.  This  is  merely  one  of  a  number  of  examples 
which  might  be  cited  to  demonstrate  the  necessity  of  simple  cultural 
adaptation,  as  it  influences  the  property  of  virulence.  Again,  it  is 
probable  that  in  order  to  develop  in  the  animal  body  it  is  necessary 
that  a  micro-organism  shall  be  capable  of  developing  without  free 
oxygen.  While  this  point  is  not  definitely  certain,  it  is  not  probable 
that  any  of  the  virulent  bacteria  can  be  strict  aerobes.  As  a  matter 
of  experience  none  of  the  pathogenic  bacteria  at  present  known  are 
absolute  aerobes — though  many  of  them  grow  better  in  artificial  cul- 
ture when  oxygen  is  freely  present  than  when  it  is  absent. 

Furthermore,  the  conditions  encountered  by  bacteria  as  they 
enter  the  animal  body  will  vary  considerably  according  to  the  path 
by  which  they  gain  entrance.  Organisms  entering  by  the  intestinal 
canal  are  subjected  to  conditions  of  acidity  or  alkalinity,  the  action 
of  digestive  juices,  of  bile,  and  to  competition  with  other  intestinal 
bacteria,  forces  to  which  many  pathogenic  germs  will  succumb,  while 
others  may  survive  there  and  thrive.  Those  entering  into  the  tissues 
by  way  of  the  skin  and  mucous  membrane,  on  the  other  hand,  en- 
counter an  immediately  mobilized  protective  mechanism  which,  suc- 
cessfully resisted  by  some  of  them,  might  easily  and  quickly  dispose 
of  small  quantities  of  other  bacteria  more  resistant  to  conditions  in 
the  bowel.  It  is  but  natural  for  this  reason  that  the  accomplishment 
of  an  infection  by  any  given  germ  must  depend  to  a  great  extent 
upon  its  gaining  entrance  to  the  body  by  the  path  best  adapted  to  its 
peculiar  requirements. 

The  mechanical  protection  afforded  by  the  coverings  of  skin  and 
mucous  membranes  is  as  a  rule  sufficient  to  prevent  the  penetration 
of  any  bacteria  which  by  chance  may  have  found  lodgment  upon 
them.  In  the  case  of  the  most  usual  pyogenic  cocci  and  many  bacilli 
such  protection  is  probably  absolute,  and  a  distinct  break  of  con- 
tinuity, such  as  a  bruise  or  a  wound,  even  though  this  may  be  too 
small  to  attract  attention,  is  necessary  for  successful  infection.  In 


THE    PROBLEM    OF    VIRULENCE  13 

the  case  of  a  very  limited  number  of  diseases  infection  seems  to  take 
place  even  through  the  unbroken  skin,  and  the  method,  often  spoken 
of  as  the  vaccination  method  of  Kolle,  employed  in  many  instances 
when  it  is  desired  to  produce  experimental  plague  infection  in  rats 
or  guinea  pigs,  consists  in  merely  rubbing  a  small  amount  of  cul- 
tural material  into  a  shaven  area  of  the  skin.  However,  in  this 
case,  as  well  as  in  other  instances  where  mere  massage  of  bacteria 
into  unbroken  skin  has  led  to  successful*  inoculation,  it  is  more 
than  likely  that  success  has  depended  upon  either  microscopic 
lesions  or  possibly  the  violent  introduction  of  the  organisms  into  the 
sebaceous  glands,  the  sweat  glands,  or  hair  follicles.  The  defense 
of  intact  mucous  membranes,  however,  is  by  no  means  impervious. 
While  many  organisms  can  be  implanted  upon  mucous  membranes 
with  impunity,  there  are  a  number  of  others  that  can  cause  local 
inflammations  upon  these  and  can  further  pass  through  them  into  the 
deeper  tissues  and  thence  into  the  general  system.  Thus  gonorrhea 
is  ordinarily  a  disease  of  implantation  upon  a  mucous  membrane, 
and  diphtheria  bacilli  and  streptococci  give  rise  to  localized  disease 
on  the  pharyngeal  and  nasal  mucosse,  the  latter  not  infrequently 
penetrating  from  the  initial  point  of  lodgment  upon  the  mucosa  into 
the  deeper  tissues  and  the  circulation,  causing  a  condition  of  "septi- 
cemia"  or  "bacteriemia."  For  the  experimental  determination  of 
the  penetrative  power  of  organisms  through  mucous  membranes  the 
conjunctiva  has  been  a  favorite  test  object,  and  it  has  been  shown 
that  plague  14  and  glanders,15  as  well  as  hydrophobia,  may  be  trans- 
mitted by  simple  instillation  of  infectious  material  into  the  unin- 
jured conjunctival  sac.  In  the  case  of  hydrophobia  16  it  is  related 
that  in  Paris  a  young  man  contracted  hydrophobia  by  rubbing  his 
eyes  with  a  finger  contaminated  with  the  saliva  of  a  rabid  dog.  In 
the  case  of  syphilis,  though  often  claimed,  there  is  no  positive  proof 
to  show  that  infection  may  take  place  through  the  uninjured  sur- 
faces. It  has  been  definitely  shown,  however,  that  tubercle  bacilli 17 
may  pass  into  the  lymphatics  through  the  intestinal  mucosa  without 
there  being  any  traceable  injuries  on  this  membrane. 

It  may  well  be,  however,  that  even  without  the  existence  of 
demonstrable  morphological  lesions  penetrability  by  micro-organisms 
may  presuppose  local  physiological  or  functional  injury,  such  as  con- 
gestion or  catarrhal  inflammation. 

Thus  it  is  seen  that  the  mechanical  obstacle  to  the  entrance  of 
micro-organisms  offered  by  skin  and  mucous  membranes,  though 
important  and  not  to  be  underestimated,  is  by  no  means  a  perfect 
safeguard. 

14  Germ.  Plague  Com.    Arb.  a.  d.  kais.  Gesundheitsamte,  Vol.  16,  1899. 

15  Conte.     Rev.  veterin.,  Vol.  18,  1893. 

16  Galtier.     Compt.  rend,  de  la  soc.  biol,  1890. 
17Bartel.     Wien.  Klinikhandt,  1906-1907. 


14  INFECTION    AND    RESISTANCE 

However,  it  is  only  very  definite  species  of  micro-organisms 
which  can  cause  disease  at  all  when  introduced  into  the  body  by 
these  paths.  For,  although  the  rubbing  of  plague  bacilli  into  the 
skin,  or  the  inoculation  of  a  cut  surface  with  streptococcal  or  glan- 
ders bacilli,  will  rapidly  lead  to  progressive  infection,  similar  inocu- 
lation with  the  typhoid  bacillus  or  the  cholera  spirillum  would  lead 
to  no  such  result.  And,  though  the  swallowing  of  pus  cocci,  pneu- 
mococci,  and  a  number  of  other  micro-organisms  would  be  entirely 
without  effect,  similar  ingestion  of  the  typhoid  and  cholera  organism 
would  usually  result  in  typical  infection. 

The  path  of  introduction,  therefore,  is  an  important  considera- 
tion in  determining  whether  or  not  a  given  micro-organism  may  give 
rise  to  disease.  It  is  necessary  that  the  manner  of  gaining  entrance 
be  suited  to  the  cultural  and  other  peculiarities  of  the  germ  in  ques- 
tion. In  the  case  of  cholera,  for  instance,  the  spirillum  which  causes 
this  disease  is  peculiarly  susceptible  to  the  deeper  defences  residing 
in  the  body  fluids  and  cells,  and  cutaneous  infection  by  the  small 
numbers  of  bacteria  likely  to  be  introduced  in  this  way  would 
promptly  be  checked  by  these  agencies.  In  the  intestinal  mucosa, 
however,  the  cholera  spirillum  finds  conditions  most  favorable  for 
rapid  multiplication,  and  the  disease  is  caused  by  the  inflammation 
and  destruction  of  the  mucous  and  submucous  tissues  by  the  poison- 
ous substances  emanating  from  the  large  numbers  of  cholera  spirilla 
which  die  and  are  disintegrated,  as  well  as  by  the  absorption  of  these 
poisons  into  the  circulation.  The  bacteria  themselves,  however,  never 
gain  a  permanent  foothold  within  the  blood  or  other  organs.  In  the 
case  of  typhoid  fever  the  conditions  are  somewhat  similar,  although 
here,  during  the  earlier  weeks  of  the  disease,  we  have  an  actual 
penetration  of  the  bacilli  into  the  circulation.  This,  however,  prob- 
ably takes  place  only  after  intraintestinal  proliferation  has  taken 
place,  which  then,  on  the  injured  mucosa,  represents  a  dose  out  of 
all  proportion  great  when  compared  with  the  quantities  that  would 
spontaneously  come  into  contact  with  the  external  surface  of  the 
body. 

This  leads  us  to  another  important  factor  concerning  the  invad- 
ing forces,  in  the  determination  of  successful  infection,  namely,  that 
of  the  quantity  introduced  or  the  dosage. 

In  order  to  cause  infection,  even  when  the  bacteria  are  of  the 
variety  known  to  produce  disease  or  "pathogenic,"  and  are  brought 
into  contact  with  the  body  by  a  path  suitable  to  their  peculiar  re- 
quirements, the  initial  quantity  introduced  must  be  sufficiently  large 
to  preclude  complete  annihilation  by  the  first  onslaught  of  the  de- 
fensive powers  of  the  body.  It  is  plain,  therefore,  that  in  the  case 
of  bacteria  weak,  in  power  to  cause  disease,  given  the  subject  of  in- 
fection and  his  defences  as  a  constant,  the  quantities  to  be  introduced 
must  be  larger  than  in  the  case  of  micro-organisms  of  violent  disease- 


THE  PROBLEM  OF  VIRULENCE         15 

producing  properties.  The  dosage  necessary  to  cause  infection, 
therefore,  is  in  inverse  proportion  to  that  property  of  bacteria  spoken 
of  as  their  "virulence."  Thus  we  measure  the  degree  of  the  so-called 
virulence  of  bacteria  by  determining  the  smallest  quantity,  measured 
by  dilution  of  platinum  loops  or  by  fractions  of  agar  slant  cultures 
(both  very  inexact  methods),  which  will  still  cause  infection  and 
death  in  susceptible  animals  of  a  standard  weight.  In  the  case  of 
micro-organisms  of  extreme  virulence,  such  as  the  anthrax  bacillus 
or  bacilli  of  the  hemorrhagic  septicemia  group,  the  inoculation  of  a 
very  small  number  of  bacteria  may  suffice  to  initiate  infection.  In- 
deed, it  has  been  claimed  for  the  anthrax  bacillus  that  the  injection 
of  a '  single  bacterium  will  produce  fatal  disease  in  a  susceptible 
animal.  The  inverse  relation  existing  between  the  degree  of  viru- 
lence and  the  number  of  bacteria  inoculated  is  well  illustrated  by  the 
experiments  of  Webb,  Williams,  and  Barber,18  carried  out  upon 
white  mice  with  anthrax,  by  the  method  of  inoculation  devised  by 
Barber.19  This  technique  consists  in  picking  up  single  organisms  with 
a  capillary  pipette  under  microscopic  control,  from  a  very  thin 
emulsion  of  bacteria  and  injecting  directly  from  the  pipette  through 
a  needle  puncture  in  the  skin.  While  requiring  a  considerable  de- 
gree of  skill,  the  method,  when  successful,  permits  an  actual  accurate 
count  of  injected  bacteria  instead  of  the  merely  approximate  esti- 
mate which  can  be  made  by  consecutive  dilutions  of  thicker  emul- 
sions. In  their  experiments  with  anthrax  in  white  mice  Webb,  Wil- 
liams, and  Barber  found  that  the  inoculation  of  a  single  thread  of 
anthrax  bacilli  (3  to  6  individuals)  taken  directly  from  the  blood 
of  a  dead  animal  (that  is,  in  the  most  virulent  condition)  would 
regularly  cause  death,  and  it  was  impossible  for  this  reason  to 
immunize  with  such  bacilli.  On  the  other  hand,  if  taken  from 
12-hour  agar  cultures  of  the  same  strain  such  small  quantities 
would  often  fail  to  kill.  The  brief  period  of  growth  under 
artificial  conditions  had  sufficiently  lessened  the  virulence  of  the 
bacilli  so  that  2,  3,  and  more  threads  could  be  injected  with- 
out harm.  And  after  several  generations  of  such  cultivation  as 
many  as  27  and  more  threads  could  be  inoculated  with  im- 
punity. 

Another  example  of  the  measurement  of  relative  degrees  of 
virulence,  by  a  method  more  commonly  employed,  may  be  illustrated 
as  follows :  The  problem  in  which  this  particular  measurement  was 
used  consisted  in  the  comparison  of  the  virulence  of  two  strains  of 
pneumococcus,  one  (^2)  successively  passed  through  white  mice,  the 
other  (Ni)  kept  alive  for  several  weeks  on  serum-agar.  To  accom- 
plish this  graded  quantities  of  18-hour  broth  cultures  of  the  twa 

18  Webb,  Williams,  and  Barber.     Jour.  Med.  Res.,  1909,  Vol.  XV. 

19  Barber.     Kansas  Univ.  Science  Bulletin,  March,  1907. 


16  INFECTION    AND    RESISTANCE 

strains  were  injected  into  mice  of  approximately  the  same  weight  as 
follows:20 


Ni 

Result 

N2 

Result 

0.1  c.  c. 
0.05  c.  c. 
0.02  c.  c. 
0.01  c.  c. 

=  dead  24  hrs. 
=  lives 
=  lives 
=  lives 

0.1  c.  c. 
0.05c.  c. 
0.02  c.  c. 
0.01  c.  c. 

=  dead  24  hrs. 
=  dead  24  hrs. 
=  dead  24  hrs. 
=  lives 

This  example  further  illustrates  another  important  fact  in  con- 
nection with  the  problem  of  infection — namely,  that  within  the 
same  species  of  bacteria  different  races  of  strains  may  exhibit  widely 
varying  degrees  of  virulence.  This  has  been  known  since  the  days 
of  Pasteur,  and  it  is  indeed  of  great  importance  in  the  immunization 
of  animals  that  weakly  virulent  strains  of  a  given  micro-organism 
may  be  used  to  produce  a  gradual  immunity  against  the  same  species 
of  bacteria  in  their  fully  virulent  condition.  Though  observed  in 
almost  all  species  of  bacteria  such  variations  are  especially  notice- 
able in  the  cases  of  streptococci  and  pneumococci — organisms  in 
which  no  two  strains  may  be  alike  in  infectiousness,  and  in  which 
the  injection  of  some  strains  into  susceptible  animals  may  produce 
no  result  whatever,  while  other  strains  will  kill  if  administered  in 
the  smallest  measurable  quantities.  To  a  large  extent  these  fluctua- 
tions of  virulence  appear  to  represent  degrees  of  adaptation  on  the 
part  of  the  bacteria  to  the  conditions  met  with  in  the  living  body; 
and  the  ease  with  which  such  variations  can  often  be  artificially  pro- 
duced would  seem  to  furnish  another  proof  that  the  property  of  in- 
fectiousness is  a  biological  attribute  of  relatively  recent  acquisition. 
For,  although  no  general  statement  of  absolute  accuracy  can  be  made, 
it  is  a  fairly  uniform  rule  that  races  of  pathogenic  bacteria  gain  in 
virulence  as  they  are  passed  through  successive  animals  of  the  same 
species,  and  lose  in  virulence  as  they  are  preserved  upon  media 
under  conditions  of  artificial  cultivation. 

Further  showing  this  ability  to  rapidly  adapt  themselves  is  the 
observation  that  passage  through  animals  of  a  certain  species  will 
enhance  the  virulence  for  this  species,  but  often  reduce  it  for  animals 
of  another  kind.  Among  the  earliest  observations  on  this  point  are 
those  of  Pasteur  21  in  his  work  on  rabies.  He  found  that  the  virus 
of  hydrophobia  when  successively  passed  through  rabbits  gained  in 
virulence  until  a  degree  of  maximum  infectiousness  was  attained 

20  For  making  such  accurate  measurements  we  have  recently  found  very 
useful  the  Precision  syringe  described  by  Terry,  Jour,  of  Inf.  Dis.,  Vol.  13, 
1913. 

21  Pasteur  and  Thuillier,  Compt.  rend,  de  I'acad.  des.  sc.,  Vol.  XII,  1883. 


THE  PROBLEM  OF  VIRULENCE        17 

beyond  which  it  could  no  longer  be  enhanced.  After  only  three 
passages  through  monkeys,  however,  the  virulence  of  this  "virus 
fixe"  for  rabbits  was  reduced  almost  to  extinction.  His  experience 
with  swine  plague  was  similar.  Swine  plague  bacilli  successively 
passed  through  rabbits  and  pigeons  gained  enormously  in  virulence 
for  these  animals  respectively,  but  lost  in  virulence  for  hogs. 

There  are  numerous  methods  by  which  the  virulence  of  micro- 
organisms can  be  attenuated  by  laboratory  manipulations,  and  since 
many  of  them  are  of  great  importance  in  the  active  immunization  of 
animals  we  will  reserve  their  detailed  discussion  until  we  come  to 
consider  the  methods  of  immunization  themselves.  Suffice  it  to  say 
in  this  place  that  most  methods  of  attenuation  consist  in  subjecting 
the  bacteria,  in  artificial  culture,  to  deleterious  influences,  either  of 
unfavorably  high  temperature,  exposure  to  light  or  harmful  chemi- 
cal agents,  or  allowing  them  to  remain  in  prolonged  contact  with  the 
products  of  their  own  metabolism  by  infrequent  transplantation.  As 
a  rule  the  attenuation  which  inevitably  follows  any  form  of  arti- 
ficial cultivation  in  the  case  of  bacteria  like  streptococci  or  pneumo- 
cocci  can  be  delayed  by  preserving  them  in  media  containing  sera 
or  tissues.  In  the  case  of  the  pneumococcus,  for  instance,  one  of 
the  best  methods  of  conserving  virulence  in  storage  is  to  keep  them 
either  in  a  soft  rabbit-serum-agar  mixture,  as  practiced  by  Wads- 
worth,  or,  better  still,  to  store  them  within  the  spleen  of  a  mouse 
dead  of  pneumococcus  infection,  as  recommended  by  Neufeld.  The 
mouse  is  autopsied  and  the  spleen  kept  in  the  dark  and  cold  in  a  des- 
iccator, under  sterile  precautions.  This,  again,  as  well  as  the  en- 
hancement of  virulence  on  passage  through  the  same  species  of  ani- 
mal— or  the  reduction  of  virulence  for  one  species  by  passage  through 
another — shows  that  such  fluctuations  are  dependent  upon  a  very 
delicate  biological  adaptation. 

It  is  interesting,  moreover,  to  look  upon  this  process  of  adapta- 
tion as  a  sort  of  immunization  of  the  bacteria  against  the  defensive 
powers  of  the  host,  a  conception  early  suggested  by  Welch.  For  just 
as  the  animal  body  may  become  more  resistant  to  the  offensive 
weapons  of  the  invaders,  so  it  is  reasonable  to  suppose  that  the  bac- 
terial body  may  gradually  develop  increased  resistance  to  the  de- 
fensive mechanism  of  the  host.  And  this,  if  it  occurs,  would  of 
course  lead  to  an  increase  of  its  invasive  power  or  virulence.  The 
increase  of  virulence  by  passage  through  animals  would  alone  lead 
us  to  suspect  that  such  acquired  resistance  to  destructive  agents  on 
the  part  of  the  bacteria  might  be  responsible  for  the  enhancement, 
but  additional  evidence  pointing  in  this  direction  has  been  brought 
by  experiments  in  which  it  was  shown  that  bacteria  cultivated  in  the 
serum  of  immune  animals  not  only  gained  in  resistance  to  destruction 
by  the  serum  constituents,  but  at  the  same  time  were  rendered  more 
highly  pathogenic.  Experiments  of  this  kind  were  carried  out  by 


18  INFECTION    AND    RESISTANCE 

Sawtchenko,22  by  Danysz,23  and  by  Walker.24  The  results  of  Wal- 
ker are  especially  instructive.  He  worked  with  a  typhoid  bacillus 
which  he  cultivated  for  a  number  of  generations  upon  the  serum  of  a 
typhoid-immune  animal,  and  found  that  after  such  treatment  the 
organism  had  gained  in  virulence  and  lost  in  agglutinability  by  im- 
mune serum,  and  that  a  larger  amount  of  specific  immune  serum  was 
necessary  to  protect  animals  against  it  than  sufficed  for  protection 
against  normal  typhoid  strains  not  thus  cultivated.  We  will  refer 
to  these  results  more  in  detail  in  a  later  chapter,  since  the  conception 
will  be  easier  to  grasp  when  we  have  Considered  more  fully  the 
mechanism  of  defence  at  the  disposal  of  the  animal  body. 

That  this  power  of  gaining  resistance  against  deleterious  influ- 
ences on  the  part  of  bacteria  is  not  confined  to  their  resistance  to  the 
animal  defences  alone  is  well  shown  by  the  experiments  of  Danysz  25 
upon  the  immunization  of  anthrax  bacilli  against  arsenic.  In  in- 
oculating series  of  50  tubes  containing  arsenic  dilutions  (ranging 
from  1  to  10,000  to  1  to  200)  with  anthrax  bacilli  Danysz  found 
that  up  to  1  to  5,000  the  arsenic  increased  the  growth  of  the  bacilli ; 
in  concentrations  higher  than  this  growth  was  inhibited.  By  grad- 
ually progressive  cultivation  of  the  organisms  in  increasing  concen- 
trations of  arsenic  he  finally  succeeded  in  obtaining  growth  in  solu- 
tions five  times  more  concentrated  than  those  in  which  they  would 
develop  at  first. 

It  is  intensely  interesting  also  that  Danysz  found,  both  in  the 
case  of  his  serum-resistant  and  arsenic-resistant  strains,  that,  as 
they  became  less  sensitive  to  the  deleterious  effects  of  these  agencies, 
they  were  altered  morphologically  in  that  they  developed  capsules. 
Similar  in  significance  to  this  is  the  very  important  observation  that 
certain  strains  of  spirochseta  pallida  may  acquire  resistance  against 
salvarsan  or  "606."  26  These  so-called  arsenic-fast  strains  are  ap- 
parently unaffected  by  the  injection  of  this  preparation  into  the 
patient. 

.  *  The  experiments  of  Danysz  were  probably  the  first  to  call  atten- 
tion to  the  possible  relationship  of  bacterial  capsule  formation  to 
virulence,  and  this  particular  phase  of  the  subject  has  since  then 
been  extensively  studied.  It  is  a  matter  of  common  observation  that 
micro-organisms  like  the  pneumococcus,  the  anthrax  bacillus,  some 
streptococci,  and  a  number  of  other  germs  which  are  capable  of  pro- 
ducing capsules  under  suitable  conditions  are  most  virulent  in  the 
capsulated  stage.  As  the  strains  are  passed  through  animals  and 
their  virulence  increases  their  ability  to  form  capsules  becomes  more 

22  Sawtchenko.     Ann.  Past.,  Vol.  11,  1897. 

23  Danysz.    Ann.  Past.,  14,  1900. 

24  Walker.    Jour,  of  Path,  and  Pact.,  Vol.  8,  1903. 

25  Danysz.     Loc.  cit. 

26  Oppenheim.     Wien.  kl.  Woch.,  23,  1910,  No.  37. 


THE    PROBLEM    OF    VIRULENCE  19 

•and  more  apparent — whereas  the  diminution  of  virulence  which 
takes  place  on  artificial  media  is  accompanied  by  a  gradual  loss  -of 
capsule  formation.  Organisms  like  the  Friedlander  bacillus  which 
retain  their  ability  to  form  capsules  almost  indefinitely  in  artificial 
culture  moreover  do  not  lose  their  virulence  to  any  great  extent  as 
long  as  this  property  is  preserved.  It  is  also  well  known  that  cap- 
sulated bacteria  are  peculiarly  insusceptible  to  the  ordinary  aggluti- 
nating powers  of  specific  immune  sera.  This  has  been  noticed,  not 
only  in  the  case  of  heavily  capsulated  bacteria  like  those  of  the  Fried- 
lander  group  or  the  streptococcus  mucosus,  but,  in  the  case  of  plague 
bacilli — where  capsulation  is  usually  present  only  in  cultures  taken 
directly  from  the  animal  body  and  cultivated  at  37°  C.,  Shiba- 
yama  27  has  found  a  direct  relation  between  non-agglutinability  and 
a  slimy  condition  of  the  cultures.  Cultures  kept  at  5°  to  8°  C.  in 
the  ice-chest  were  easily  aggiutinable  and  lacked  the  slimy  property. 
Cultures  kept  at  37.5°  C.  were  slimy  and  thready  in  consistency  and 
were  not  as  easily  agglutinated  by  the  same  immune  serum. 
Forges  28  later  showed  that  inagglutinable,  capsulated  bacteria  can 
be  made  amenable  to  the  agglutinating  action  of  the  serum — which 
we  may  assume  to  indicate  vulnerability  by  the  serum  if  the  capsule 
is  previously  destroyed  by  heating  at  80°  C.  for  about  15  minutes  in 
!/4  normal  acid. 

Against  the  cellular  defences,  the  leukocytes,  capsulated  bacteria 
seem  also  to  be  more  resistant  than  are  the  non-capsulated.  This 
has  been  especially  studied  by  Gruber  and  Futaki,29  who  find  that  a 
capsulated  bacillus  is  rarely  taken  up  by  a  phagocyte  even  when 
these  cells  are  apparently  normal  and  able  to  take  up  the  uncap- 
sulated  organisms.  They  go  so  far  as  to  claim  that,  in  the  case  of 
anthrax  in  rabbits,  the  development  or  absence  of  a  capsule  deter- 
mines whether  or  not  infection  can  take  place.  The  same  conclusion 
is  reached  in  similar  studies  by  Preisz,30  who  does  not  believe  that 
anthrax  bacilli  can  ever  cause  infection  unless  they  possess  the 
power  of  forming  capsules.  All  this  experimental  evidence  points 
strongly  toward  a  probable  direct  relationship  between  capsule  for- 
mation and  virulence,  in  the  sense  that  a  thickening  of  the  ectoplasm 
may  in  some  way  protect  the  bacteria  from  the  destructive  forces 
aimed  at  them  by  the  cells  and  fluids  of  the  invaded  body. 

As  a  matter  of  fact,  even  when  no  distinct  capsule  is  visible,  it 
is  nevertheless  possible  that  ectoplasmic  changes  may  take  place. 
This  phase  of  the  subject  has  been  thoroughly  discussed  by  a  number 
of  writers,  more  especially  by  Eisenberg.31  It  appears  that  many 

27  Shibayama.     Centralbl.  f.  Bact.,  Orig.  Vols.  38,  1905,  and  42,  1906. 

28  Forge's.     Wien.  klin.  Woch.,  p.  691,  1905. 

29  Gruber  and  Futaki.     Munch,  med.  Woch.,  6,  1906. 

30  Preisz.     Centralbl  f.  Bakt.,  Vol.  49,  1909. 

81  Eisenberg.     Centralbl.  f.  Bakt.,  I,  45,  1908,  p.  638. 


20  INFECTION    AND    RESISTANCE 

bacteria,  in  which  true  capsule  formation  has  not  been  observed,  may 
show  swelling  or  enlargement  under  conditions  in  which  their  of- 
fensive activities  in  the  infected  animal  body  are  called  into  play.32 
Radziewsky33  has  noticed  such  swelling  of  B.  coli  in  fatal  guinea- 
pig  infections,  and  spoken  of  it  as  "one  of  the  characteristic  signs 
of  infectiousness."  Kisskalt  34  has  described  the  same  thing  in  the 
case  of  streptococci,  and  Eisenberg  interprets  this  as  signifying  an 
ectoplasmic  hypertrophy  comparable  in  principle  to  capsule  forma- 
tion. He  looks  upon  the  ectoplasmic  zone  as  a  protective  layer,  and 
calls  attention  to  the  observation  of  Liesenberg  and  Zopf)35  who 
showed  that  capsulated  strains  of  leukonostoc  mesenteroides  will 
withstand  85°  C.,  a  temperature  at  which  uncapsulated  forms  are 
rapidly  killed. 

There  is  a  considerable  amount  of  evidence,  then,  which  seems 
to  indicate  that  the  development  of  a  capsule  is  at  least  one  im- 
portant method  by  which  the  bacteria  can  protect  themselves  against 
the  onslaught  of  the  defences  of  the  invaded  animal  body  and  in  so 
doing  become  more  virulent. 

It  is  not  likely,  however,  that  this  merely  passive  increase  of  the 
resistance  to  injury  on  the  part  of  the  bacteria  accounts  for  the 
entire  train  of  phenomena  included  in  an  enhancement  of  virulence. 
It  has  been  suggested  by  a  number  of  observers  that  definite  active  .-- 
offensive  characteristics  distinguish  the  virulent  from  the  avirulent 
bacteria,  in  that  the  former  may  secrete,  within  the  living  body, 
substances  by  which  the  destructive  powers  of  serum  and  leukocytes 
are  neutralized  or  held  at  bay.  A  very  definite  suggestion  of  such  a 
possibility  we  find  expressed  in  the  now  classical  paper  of  Salmon 
and  Smith  36  on  hog  cholera  immunity,  published  in  1886.  They 
say:  ".  .  ..  the  germs  of  such  maladies  are  only  able  to  multiply  in 
the  body  of  the  individual  attacked,  because  of  a  poisonous  principle 
or  substance  which  is  produced  during  the  multiplication  of  these 
germs."  37  Bouchard  formulated  such  a  theory  in  1893  by  speaking 
of  the  "produits  secretes  par  les  microbes  pathogeniques,"  substances 
which  he  found  in  cultures  of  virulent  bacteria,  and  which  seemed 
to  reenf orce  the  invasive  powers  of  the  germs.  Kruse  38  also  within 
the  same  year  developed  a  similar  idea.  He  assumed  that  bacteria  A 
may  secrete  enzyme-like  substances  which  paralyze  the  destructive  ^Si- 
properties  of  animal  serum,  and  in  this  way  gain  the  power  to 

12  These  forms  Bail  has  spoken  of  as  "thierische  Bazillen." 
33  Radziewsky.     Zeitschr.  f.  Hyg.,  Vol.  34. 
54  Kisskalt.     Cited  after  Eisenberg,  loc.  cit. 

35  Liesenberg  and  Zopf.     Centralbl.  f.  Bakt.,  Vol.  XII,  1892. 

36  Salmon  and  Smith.     Proc.  Biol.  Soc.,  Washington,  D.   C.,  Ill,  1884. 
6,  p.  29. 

37  A  typewritten  copy  of  this  paper  was  kindly  put  at  my  disposal  by 
Prof.  Theobald  Smith. 

38  Kruse.     Ziegler>s  Beitrage,  Vol.  XII,  1893. 


THE    PROBLEM    OF    VIRULENCE  21 

invade.  As  a  matter  of  fact  we  have  learned,  since  that  time,  that 
staphylococci  may  secrete  soluble  substances,  "leukocidins,"  which 
injure  white  blood  cells,  and  that  many  bacteria  produce  similar 
poisons,  "hsemotoxins,"  which  specifically  injure  red  blood  cells — 
thereby  causing  anaemia  and  reducing  the  resistance  of  the  host. 
However,  the  correlation  and  further  elaboration  of  these  thoughts 
of  Salmon  and  Smith,  of  Bouchard  and  of  Krnse  was  left  to  Bail,39 
in  what  is  known  as  his  "aggressin  theory."  Bail  maintains  on  the 
basis  of  careful  experimentation  that  virulent  bacteria  can  produce 
within  the  animal  body  substances  which  he  calls  "aggressins,"  upon 
which  depend  their  invasive  powers  or  virulence.  These  substances 
are  secreted  only  under  stress  of  the  struggle  against  the  unusual 
defences,  are  not  demonstrable  in  test-tube  cultures,  and  are  in 
themselves,  according  to  Bail,  entirely  non-toxic. 

He  obtains  these  aggressins  by  injecting  virulent  bacteria  into  the 
peritoneal  cavity  of  a  guinea  pig  and  immediately  after  death  re- 
moving the  exudate.  This  he  centrifugalizes,  removes  the  bacteria 
and  cells,  and  sterilizes  the  supernatant  liquid  by  the  addition  of 
small  quantities  of  chloroform.  The  action  of  the  exudates  in  which 
aggressins  have  been  produced  by  the  bacteria  is  the  following: 
(We  take  this  tabulation  from  Bail's  own  paper  on  typhoid  and 
cholera  aggressins  in  the  Archiv  fur  Hygiene,  Vol.  52,  p.  342.) 

1.  Sublethal  doses  of  typhoid  bacilli  or  cholera  spirilla  become 
lethal  when  the  aggressin  is  injected  with  them. 

2.  Lethal  doses  of  bacilli  which  ordinarily  would  cause  a  slow 
infection  only  cause  a  rapid  and  severe  infection  when  aggressins 
are  added. 

3.  The  addition  of  aggressin  neutralizes  the  bacteria-destroy- 
ing power  of  immune  serum  in  the  peritoneal  cavity  of  a  guinea  pig. 

4.  The   injection  of  aggressin  alone  produces   subsequent  im- 
munity. 

It  is  impossible  to  discuss  with  completeness  the  arguments  ad- 
vanced for  and  against  the  correctness  of  Bail's  views  until  we  have 
described  in  detail  the  mechanism  of  protection  at  the  disposal  of 
animals.  But  the  main  objection  brought  against  this  theory  is  that 
of  Wassermann  and  Citron,40  who  claim  that  all  these  properties  of 
the  aggressive  exudates  can  be  explained  by  the  fact  that  they  con- 
tain extracts  of  the  bacteria  (endotoxins),  which,  injected  with 
a  sublethal  dose  of  bacteria,  merely  enhance  their  action  in  the  same 
way  that  this  would  have  been  accomplished  by  the  injection  of 
additional  dead  bacterial  bodies.  It  will  require  much  further  work 
before  this  point  is  settled,  and  the  problem  is  peculiarly  involved  and 

39  Ball.    Archiv  f.  Hijg.,  Vols.  52  and  53,  1905 ;  Folio  serologica,  Vol.  7, 
1911. 

40  Wassermann  and  Citron.     Deutsche  med.  Woch.,,  Vol.  31,  28,  1905. 


22  INFECTION    AND    RESISTANCE 

difficult.  However,  the  recent  work  of  Rosenow  41  on  pneumonococci 
seems  to  bring  some  reinforcement  to  the  ranks  of  those  who  main- 
tain the  existence  of  a  special  offensive  substance  at  the  command  of 
virulent  bacteria.  Rosenow  extracted  pneumococci  grown  on  serum 
broth  and  found  that  such  extracts  when  made  from  virulent  strains 
would  protect  avirulent  strains  from  engulfment  by  phagocytes.  The 
non-virulent  strains  left  in  these  extracts  for  24  hours  became  viru- 
lent. He  believes,  therefore,  that  the  virulence  of  pneumococci  de- 
pends largely  upon  the  possession  of  these  substances  which  he  calls 
"virulins,"  and  which  in  function  at  least  are  conceived  as  very 
similar  to  the  "aggressins." 

Recent  results  obtained  by  the  writer 42  with  Dwyer  seem  to 
indicate  that  anaphylatoxins  produced  from  the  typhoid  bacillus 
possess  some  of  the  properties  claimed  for  his  aggressin  by  Bail. 
It  is  not  impossible  that  the  "aggressins"  obtained  by  him  were  of 
this  nature. 

Virulence,  then,  may  be  analyzed  into  two  main  attributes :  one  a 
purely  passive  property  of  resistance  or  self-preservation  on  the 
part  of  the  bacteria,  perhaps  morphologically  expressed  in  ectoplas- 
mic  hypertrophy  and  capsule  formation;  the  other  an  actively  of- 
fensive weapon  in  the  form  of  substances  of  the  nature  of  the  "ag- 
gressins" of  Bail  or  the  "virulins"  of  Rosenow.  The  extent  of  our 
present  knowledge  of  details  does  not  warrant  a  statement  of  the 
case  in  more  definite  terms. 

From  the  facts  we  have  discussed  in  the  preceding  paragraphs 
it  now  becomes  manifest  that  the  elements  which  determine  the 
nature  of  an  infectious  disease  are  twofold.  On  the  one  hand 
each  variety  of  infectious  germs  possesses  certain  biological  and 
chemical  attributes  which  are  specific  and  peculiar  to  itself ;  by  these 
its  predilection  for  path  of  entrance  and  mode  of  attack  is  de- 
termined, and  upon  these  depends  the  nature  of  the  reaction  called 
forth  in  the  animal  body.  On  the  other  hand  the  degree  of  infec- 
tion in  each  case,  the  severity  of  the  reaction  and  the  ultimate  out- 
come are  determined  by  the  balance  which  is  struck  between  the 
virulence  of  the  entering  germ  and  the  protective  mechanism  opposed 
to  it. 

The  specific  properties  of  each  micro-organism  are  the  factors 
^vhich  account  for  the  clinical  uniformity  (within  definite  limits) 
-which  is  observed  in  the  maladies  produced  in  different  individuals 
iby  the  same  species  of  bacteria.  Thus  a  severe  typhoid  fever  is,  in 
•essential  characteristics,  entirely  similar  to  a  mild  case — since  in 
"both  instances  the  path  of  entrance,  through  the  intestine,  is  the 
same,  the  distribution  of  the  germs  after  entrance  differs  only  in 
degree,  and  the  reactions,  local  and  systemic,  which  are  called  forth 

41  Rosenow.     Jour,  of  Inf.  Dis.,  Vol.  4,  1907. 

4?  Zinsser  and  Dwyer.     Proc.  Soc.  Exp.  Biol.  and  Med.,  Feb.,  1914. 


THE  PROBLEM  OF  VIRULENCE         23 

are  alike.  And  cases  of  this  disease  in  general  differ  as  a  class  from 
the  maladies  caused  by,  let  us  say,  the  group  of  clinical  conditions 
resulting  from  anthrax  infection,  where  entrance  is  through  the 
skin,  and  generalized  infection  of  the  blood  ensues  without  definite 
or  regular  localization  in  any  given  organ.  Again,  a  localized 
staphylococcus  abscess  will  differ  materially  from  an  equally  local- 
ized focus  of  tuberculosis,  because  the  chemical  constituents  of  these 
bacteria  respectively  call  forth  each  a  characteristic  response  on  the 
part  of  the  defensive  mechanism. 

Such  specificity  of  the  various  micro-organisms  may  of  course 
be  due  partly  to  their  mode  of  attack  and  distribution,  and  partly, 
as  we  shall  see,  to  the  pharmacological  action  of  the  poisonous  prod- 
ucts given  out  by  them. 

That  both  factors  contribute  seems  beyond  doubt;  but  recent 
work,  especially  that  of  Friedberger,  which  is  fully  discussed  in 
another  place  (see  p.  415),  seems  to  show  that  clinical  differences 
Depend  much  less  than  was  formerly  supposed  upon  specificity  of  the 
intracellular  poisons,  and  much  more  upon  distribution  and  localized 
accumulation  of  the  germs,  conditions  which  are  determined  rather 
by  the  mode  and  extent  of  invasion  than  by  chemical  differences  of 
poison  production.  This  problem,  rather  difficult  to  discuss  on  the 
limited  basis  of  the  facts  so  far  outlined,  will  become  clearer  as  we 
proceed,  but  we  need  only  refer  at  present  to  the  essential  clinical 
uniformity  of  the  various  forms  of  septicemia,  where  organisms 
freely  circulate  in  the  blood — with  often  a  focus  of  distribution  on  a 
heart  valve — conditions  in  which  it  is  rarely  possible  to  determine 
the  species  of  the  responsible  germ  except  by  blood  culture.  Or, 
again,  as  Friedberger  43  points  out,  there  is  great  similarity  between 
the  ordinary  pneumococcus  pneumonia  and  that  caused  by  the  Fried- 
lander  bacillus.  In  both  cases  the  distribution  and  mode  of  attack  of 
the  bacteria  are  essentially  the  same,  though  the  micro-organisms' 
themselves  are  biologically  very  dissimilar. 

One  and  the  same  micro-organism,  on  the  other  hand,  may  cause 
entirely  different  clinical  conditions,  and  here  the  type  of  infection 
depends  purely  on  the  degree  of  invasion  possible  in  the  given  case — 
that  is,  the  balance  between  virulence  and  resistance.  A  germ  may 
enter  the  body  and  cause  an  inflammatory  reaction  at  the  point  of 
entrance,  the  process  remaining  purely  localized.  In  such  cases  the 
defensive  forces  have  been  so  efficient,  the  invasive  properties  of  the 
germ  so  relatively  weak,  that  progression  beyond  the  point  of  en- 
trance is  prevented  and  the  resultant  disease  takes  the  form  merely 
of  a  localized  abscess.  This  is  the  case  when  a  healthy  individual 
is  infected  with  an  attenuated  organism  or  by  one  whose  species' 
characteristics  do  not  include  a  powerful  invasive  property.  Thus 
streptococci,  if  entering  the  tissues  of  a  normal  subject  in  small 
43  Friedberger.  Deutsche  med.  Woch.,  No.  11,  1911. 


24  INFECTION    AND    RESISTANCE 

numbers  or  in  attenuated  form,  may  produce  a  purely  localized  in- 
fection, and  ordinarily  non-pathogenic  germs  like  proteus,  subtilis, 
or  colon  bacilli  may  produce  localized  abscesses  in  weak  and  debili- 
tated individuals,  though  implanted  upon  a  healthy  subject  they 
would  be  rapidly  disposed  of  without  gaining  even  a  preliminary 
foothold.  Such  tendency  to  localization  is  the  common  form  of  in- 
fection in  the  case  of  a  number  of  germs.  It  is  the  most  usual  type 
of  staphylococcus  infection,  for  instance,  in  which  the  degree  of 
virulence  of  the  strains  ordinarily  met  is  such  that  the  balance  struck 
by  them  with  the  average  defensive  powers  of  man  results  in  localiza- 
tion. However,  the  same  micro-organism,  enhanced  in  virulence,  or 
gaining  entrance  in  unusual  numbers  in  a  weakened  individual,  may 
rapidly  spread  from  the  point  of  inoculation,  at  first  by  contiguity, 
then  by  invasion  of  the  blood  and  lymph  channels,  and  become 
generalized. 

When  organisms  become  generalized  and  circulate  in  the  blood 
the  resulting  condition  is  spoken  of  as  septicemia  or  bacteriemia. 
This  is  the  form  of  infection  commonly  caused  by  streptococci, 
bacilli  of  the  hemorrhagic  septicemia  group,  anthrax  bacilli,  and 
many  others.  It  implies  a  powerful  invasive  property  and  always 
constitutes  a  condition  of  great-  gravity  when  persistent.  We  are 
learning  of  recent  years,  however,  that  in  many  infectious  diseases 
formerly  regarded  as  purely  localized  a  temporary  entrance  of  the 
bacteria  into  the  circulation  is  a  usual  occurrence.  Thus  Fraeiikel  44 
has  shown  that  lobar  pneumonia  is  almost  always  accompanied  dur- 
ing the  acute  stages  of  the  disease  by  pneumococcus  septicemia,  and 
in  typhoid  fever  we  now  know  that  the  organisms  circulate  freely  in 
the  blood  during  the  first  two  weeks  of  the  disease,  and  often  longer 
than  this. 

In  these  and  other  conditions  the  bacteria  may  be  gradually  de- 
stroyed and  disappear  from  the  blood  stream  as  the  immunity  of  the 
subject  increases.  In  other  cases  the  bacterial  activities  may  be 
partially  checked,  the  process  becoming  slower  and  more  chronic. 
This  is  especially  often  the  case  when  micro-organisms  after  entrance 
to  the  circulation  have  found  a  secondary  lodgment  upon  a  heart 
valve,  from  which  a  continuously  renewed  supply  of  bacteria  can 
be  given  off  to  the  blood.  A  special  form  of  such  "malignant  endo- 
carditis" caused  by  the  Streptococcus  viridans  is  particularly  apt  to- 
take  this  chronic  course. 

The  presence  of  bacteria  in  the  blood  is  not,  therefore,  as  for- 
merly supposed,  an  invariably  fatal  condition. 

Adami's  recent  work  would  indicate,  moreover,  that  bacteria  may 

normally  enter  the  portal  or  even  the  general  circulation  from  the 

intestine  during  health.      This  condition  of  "sub-infection,"   as  he 

calls  it,  is  more  fully  discussed  on  p.  234.     That  colon  and  other  in- 

44FraenkeL     V.  Ley  den  Festschr.,  1902. 


THE  PROBLEM  OF  VIRULENCE         25 

testinal  bacteria  may  often  penetrate  into  the  portal  circulation  is 
indicated  by  the  occasional  occurrence  of  colon  bacillus  abscesses 
after  trauma  of  the  liver.  In  most  septicemias,  however,  caused  by 
virulent  bacteria  the  invasion  of  the  blood  stream  persists,  rapid 
multiplication  occurs  and  leads  to  death. 

From  the  circulation  the  bacteria  may  gain  lodgment  in  various 
organs  and  cause  the  formation  of  Secondary  abscesses.  This  condi- 
tion is  known  as  ^pyemia,"  and  may  be  caused  by  almost  any  bac- 
teria which  are  capable^  of  producing  septicemia.  Thus  staphylo- 
cocci,  streptococci,  or  pneumococci  may  lodge  in  bones,  joints,  brain, 
or  kidneys,  in  fact  in  any  organ  in  which  they  can  gain  a  foothold. 
However,  there  are  evidences  of  distinct  tissue  predilections  on  the 
part  of  certain  germs.  Thus  the  virus  of  rabies  and  that  of  polio- 
myelitis, though  to  some  extent  universally  distributed,  seem  espe- 
cially to  concentrate  in  the  nervous  system;  cholera  spirilla  and 
dysentery  bacilli  appear  to  find  conditions  most  favorable  for  de- 
velopment in  the  intestinal  mucosa;  amebic  abscesses  are  most 
common  in  the  liver;  gonococcus  infections  when  generalized  find 
secondary  localization  with  particular  frequency  on  heart  valves  and 
joints;  leprosy  bacilli  have  a  predilection  for  the  nerve  sheaths;  and 
glanders  bacilli  injected  into  the  peritoneum  of  a  male  guinea  pig 
localize  with  such  regularity  in  the  testicles  that  the  experiment  has 
diagnostic  value  (Strauss  test).  Conversely  it  is  only  explicable  on 
the  assumption  of  such  selective  lodgment  that  tubercle  bacilli,  even 
though  otherwise  universally  distributed  through  the  body,  will  be 
absent  from  striped  muscle  tissue,  and  rare  in  the  walls  of  the  stom- 
ach. Such  selection — as  far  as  we  can  account  for  it  at  all,  seems  to 
depend  upon  the  varying  cultural  conditions  encountered  by  the 
germs  in  different  organs. 

On  the  other  hand,  localization  may  also  be  dependent  upon 
accidental  conditions  such  as  trauma.  Infections  in  which  the  en- 
trance of  bacteria  is  coincident  with  injury — as  in  the  case,  for  in- 
stance, of  compound  fractures — will  be  able  to  spread  throughout 
the  injured  region  much  more  easily  than  they  could  enter  the 
healthy  tissue.  In  fact,  it  is  well  known  that  local  tissue  injury  at 
the  point  of  inoculation  favors  infection  since  it  furnishes  a  rich 
substratum  for  growth  in  the  form  of  dead  cells  or  blood  clot  and 
interferes  with  the  accomplishment  of  a  normal  protective  reaction. 
In  cases  in  which  bacteria  are  circulating  in  the  blood  mechanical 
injury  may  create  a  focus  of  reduced  resistance  on  which  the  in- 
vaders can  gain  a  foothold.  It  is  in  this  way  perhaps  that,  among 
other  things,  we  can  explain  tuberculosis  of  joints  or  bones  which 
present  a  history  of  injury  preceding  the  development  of  the  infec- 
tion— or  the  pleurisy  and  lobular  pneumonias  which  have  been  known 
to  ensue  upon  the  fracture  of  a  rib. 

It  is  also  possible  that  bacteria  may  be  distributed  in  various 


26  INFECTION    AND    RESISTANCE 

organs  directly  from  the  initial  focus  by  embolism  or  by  the  massive 
invasion  of  a  blood  vessel.  It  is  by  such  breaking  into  a  vein  that 
Weigert  explains  the  generalization  of  miliary  tuberculosis. 

The  inflammatory  reaction  which  usually  ensues  at  the  point  of 
entrance  of  bacteria  is  merely  a  result  of  the  local  struggle  between 
invader  and  tissues,  and  the  violence -of  this  reaction  is  in  a  large 
measure  an  indication  of  the  resistance  of  the  infected  subject. 
When,  for  instance,  a  streptococcus  of  moderate  virulence  gains 
lodgment  in  the  skin  of  a  healthy  individual  the  rapid  mobilization 
of  leukocytic  and  other  defences  may  prevent  further  invasion  by 
the  bacteria  and  lead  to  a  struggle  which  is  clinically  evidenced  by 
severe  local  symptoms.  Did  the  virulence  of  the  streptococci  far 
overbalance  the  powers  of  resistance  the  local  struggle  might  be 
reduced  to  a  minimum,  the  infection  progressing  without  any,  or 
with  but  a  slight  local,  reaction.  The  fact  that  pneumococci  lodging 
in  the  human  lung  ordinarily  cause  lobar  pneumonia  is  merely  an  • 
evidence  of  a  considerable  degree  of  resistance  to  these  germs  on  the 
part  of  the  average  human  being.  Pneumococci  introduced  into  the 
pulmonary  alveoli  of  very  susceptible  animals  (rabbits)  may  pass 
directly  through  into  the  circulation,  causing  fatal  septicemia  with- 
out leading  to  a  more  than  mild  and  temporary  reaction  in  the  lungs 
themselves.  If,  as  in  Wadsworth's  45  experiments,  the  rabbits  are 
partially  immunized — that  is,  their  resistance  increased  before  the 
pulmonary  inoculation  is  carried  out — a  violent  local  reaction,  anal- 
ogous to  lobar  pneumonia,  may  follow,  the  severity  of  the  reaction 
at  the  portal  of  entry  being  manifestly  an  evidence  of  more  energetic 
opposition  to  further  penetration  of  the  bacteria. 

The  entrance  of  bacteria  into  the  deeper  tissues,  and  even  the 
circulation,  without  any,  or  with  but  slight,  local  evidences  of  infec- 
tion at  the  point  of  entrance  is  by  no  means  rare.  The  innocent 
appearance  of  the  site  of  the  entrance  of  the  bacteria  in  generalized 
streptococcus  infection  is  a  common  surgical  observation,  and  a  strep- 
coccus-infected  wound  of  the  hand  or  leg  in  a  patient  dying  of  septi- 
cemia may  appear  but  slightly  inflamed  and  edematous  and  incom- 
parably milder  in  appearance  than  a  staphylococcus  boil  with  which 
the  patient  is  walking  about  and  suffering  hardly  any  systemic  dis- 
turbance. 

Between  the  time  of  entrance  of  the  bacteria  into  the  body  and 
the  first  appearance  of  symptoms  of  disease  there  is  always  a  definite 
interval  which  is  spoken  of  as  "incubation  time."  This  period  is 
made  up  of  two  definite  divisions — one  the  time  necessary  for 
growth,  distribution,  and  accumulation  of  the  bacteria,  the  other  the 
time  necessary  for  the  action  of  the  toxin  or  poison  which  may  be 
secreted.  The  latter,  the  incubation  time  of  the  toxin,  is  a  subject 
which  is  still  unclear  in  many  of  its  phases,  and  will  be  discussed 
45  Wadsworth.  Am.  Jour,  of  the  Med.  Sc.,  Vol.  27,  1904. 


THE    PROBLEM    OF    VIRULENCE  27 

in  the  following  chapter  (see  p.  37).  The  former,  however,  is  easily 
comprehended,  in  fact,  is  to  be  expected.  For  the  small  number  of 
bacteria  which  gain  entrance  to  the  tissues  in  spontaneous  infection 
is  entirely  inadequate  in  itself  to  produce  symptoms.  It  is  neces- 
sary that  multiplication  shall  take  place  until  the  bacteria  have  ac- 
cumulated in  number  sufficient  to  cause  noticeable  physiological 
disturbance.  That  the  interval  necessary  for  this  must  vary  accord- 
ing to  the  number  of  bacteria  originally  introduced,  the  virulence  of 
these,  and  the  specific  resistance  of  the  patient  goes  without  saying. 
Von  Pirquet  and  Schick  have  suggested  also  that  the  incubation 
time  may  correspond  roughly  to  the  interval  during  which  the  sub- 
ject is  becoming  "allergic"  or  hypersusceptible  to  the  bacteria  or 
virus.  This  will  be  discussed  at  greater  length  in  the  chapter  on 
anaphylaxis.46 

But  within  the  limits  of  the  variations  introduced  by  these  fac- 
tors the  incubation  time  of  each  infectious  disease — if  spontaneously 
acquired — is  sufficiently  uniform  to  be  characteristic.  Thus  the  pri- 
mary lesion  in  syphilis  follows  the  inoculation  after  an  interval  of 
two  to  three  weeks,  rabies  follows  inoculation  with  street  virus  after 
about  four  to  six  weeks,  the  period  being  somewhat  dependent  on  the 
location  of  the  bite ;  typhoid  fever  takes  about  two  weeks  to  develop ; 
gonorrhea  about  five  to  seven  days;  small-pox  about  two  weeks; 
yellow  fever  three  to  five  days;  and  scarlet  fever  and  diphtheria 
about  two  to  six  days.  In  general,  it  may  be  stated  that  within  the 
limits  observed  for  each  particular  infection  the  shorter  the  incuba- 
tion time  the  more  severe  is  the  infection.  Thus  if  tetanus  follows 
inoculation  with  the  tetanus  bacillus  within  seven  days  the  prognosis 
is  far  more  grave  than  when  the  incubation  time  has  occupied  two 
or  three  weeks.  And  if  localized  and  general  symptoms  follow  rap- 
idly (within  twenty-four  to  forty-eight  hours)  after  a  streptococcus 
infection  it  is  likely  that  the  process  is  a  very  severe  and  virulent 
one. 

46  Von  Pirquet  u.  Schick.     Wien.  kl.  Woch.,  16,  1903,  pp.  758  and  1244. 


CHAPTER    II 

BACTEKIAL  POISONS 

WHEN  bacteria  have  gained  a  foothold  anywhere  within  the 
animal  body  the  local  and  general  disturbances  which  follow,  in  all 
but  the  mildest  and  most  trifling  cases,  are  such  that  we  cannot 
account  for  them  solely  on  the  basis  of  mechanical  injury. 

It  may  well  be  that  the  obstruction  of  capillaries  and  lymphatics 
and  the  pressure  upon  parenchyma  cells,  always  incident  to  inflam- 
matory reactions,  contribute  materially  to  local  destruction,  and 
thereby  indirectly  to  systemic  effects.  However,  even  in  diseases 
like  anthrax,  in  which  the  body  of  the  victim  after  death  is  found 
flooded  throughout  with  masses  of  bacteria,  these  factors  cannot  fully 
explain  the  clinical  manifestations.  And  such  cases,  indeed,  ara 
extreme  examples,  since,  in  the  large  majority  of  bacterial  diseases, 
the  illness  resulting  in  the  patient  is  severe  out  of  all  proportion  to 
the  extent  of  the  tissue  area  invaded. 

Moreover,  all  infections,  if  at  all  severe,  whatever  their  nature 
or  localization,  give  rise  to  fever,  and  this  symptom  alone,  if  care- 
fully observed  from  hour  to  hour,  may  be  sufficiently  characteristic 
to  indicate  the  specific  micro-organism  which  is  causing  the  illness. 
With  this  there  occur  alterations  of  the  blood  picture,  either  a 
numerical  increase  of  white  blood  cells  (leukocytosis)  or  a  change  in 
the  relative  proportions  of  the  different  kinds  of  leukocytes — or 
again  an  anemia  caused  by  the  destruction  of  red  cells.  There  may 
also  be  degenerative  changes  in  parenchyma  cells  of  organs  far  re- 
moved from  the  actual  site  of  bacterial  lodgment.  All  these  facts 
indicate  very  definitely  that,  apart  from  localized  tissue  destruction 
or  purely  mechanical  interference  with  function  by  capillary  ob- 
struction or  pressure,  there  is  at  the  same  time  an  absorption  of 
poisonous  substances  emanating  from  the  bacteria. 

From  the  earliest  days  of  logical  investigation  into  the  nature 
of  infectious  disease,  as  soon,  in  fact,  as  cultural  methods  had  been 
introduced,  bacteria  were  studied  with  the  purpose  of  throwing  light 
upon  this  phase  of  their  activity.  As  a  result  of  such  investigations 
Selmi,1  in  1885,  described  certain  basic  toxic  substances  which  he 
obtained  from  putrefying  human  cadavers  and  for  which  he  sug- 
gested the  designation  "ptomain"  (from  rrrw/xa  —  dead  body).  These 

1  Selmi.     Cited  from  Hammarsten,  "Textbook  of  Phyaol.  Chem.,"  p.  16. 


BACTERIAL    POISONS  29 

poisons  were  later  more  extensively  studied  by  Brieger,2  Gautier,3 
Griffiths,4  and  others,  and  it  was  at  first  surmised  that  the  formation 
of  such  substances  in  the  infected  animal  might  be  held  responsible 
for  the  toxemic  manifestations  which  accompany  bacterial  disease.5 
This,  as  we  shall  see,  is  not  the  case.  Ptomains  are  probably  not 
formed  in  traceable  quantity  in  the  living  tissues  and  are  not  in  any 
way  identical  with  the  specific  bacterial  poisons  which  are  respon- 
sible for  the  toxemia  of  infectious  diseases.  Nevertheless,  they  have 
some  pathogenic  significance,  since  they  are  invariably  products  of 
the  proteolysis  caused  by  bacteria  and  can  give  rise  to  illness  when 
ingested  with  putrefying  foodstuffs.  It  is  important,  therefore,  that 
we  discuss  them  briefly  and  consider  their  fundamental  distinction 
from  the  true  bacterial  poisons. 

Whenever  dead  organic  material,  meat,  fish,  vegetable  refuse, 
etc.,  is  left  to  itself  under  suitable  conditions  of  moisture  and  tem- 
perature, putrefaction  sets  in.  As  a  result  of  bacterial  growth  the 
protein  is  broken  up  and  among  the  intermediate  products  of  such 
proteolysis  ptomains  appear.  Chemically 6  T  8  these  substances  are 
basic  nitrogenous  compounds  which  may  or  may  not  contain  oxygen. 
Because  of  their  basic  and  often  highly  toxic  properties  they  have 
been  spoken  of  as  "animal  alkaloids."  Many  of  them  contain  only 
C,  H,  and  !N",  and  are  ammonia  substitution  products.  (See 
Yaughan  and  Novy,  loc.  cit.,  p.  248.)  Thus  some  of  the  simpler  ones 
are: 

Methylamin==(CH3)  KH2 

Dimethylamin—  (  CH3  )  2  NH 

Trimethylamin==(CH3)3  N 
Among  those  somewhat  more  complex  are: 

Putrescin=  KS2— CH2— CH2— CH2— CH2—  KH2 
and    Cadaverin=NH2— CH2— CH2— CH2  —  CH2— CH2—  KE2 
Samuely  classifies  the  ptomains  according  to  their  nitrogen  con- 
tents as  follows: 

1.  Those    with    one    nitrogen    atom     (CgH^N)     (C8H13!N") 
(C10H1BN) 

2.  Those  with  two  nitrogen  atoms  such  as  putrescin  (C4H12N2) 
and  cadaverin  (C5H14N2)  and 

2Brieger.  "Die  Ptomaine,"  Berlin,  1885;  Virchow's  Archiv.,  Vols.  112 
and  115;  Berl  klin.  Woch.,  1887,  1888. 

3  Gautier.     Cited  after  Pick,  Bull,  de  I'acad.  de  med.,  1886. 

4  Griffiths.     Compt.  Rend,  de  I'acad.  des  sc.,  Vol.  113. 

5  For  a  historical  outline  of  our  knowledge  of  these  poisons,  as  well  as 
for  a  thorough  treatment  of  their  nature,  see  Vaughan  and  Novy,  "Cellular 
Toxins." 

6  For  a  discussion  of  the  chemistry  of  the  ptomains   see  Vaughan  and 
Novy,  "Cellular  Toxins,"  Lea  Bros.,  Philadelphia,  1902. 

7  Also    Samuely  in   Oppenheimer's  "Handbuch   der  Biochemie,"   Vol.   I, 
pp.  794  et  seq. 

8  See  also  Wells,  "Chemical  Pathology,"  Saunders,  Phila.,  1907. 


30  INFECTION    AND    RESISTANCE 

3.  Those  with  three  nitrogen  atoms  such  as  methyl  guanidin 
(C2H7N,). 

4.  Finally  there  is  an  important  group  which  contains  oxygen, 
such  as  the  substance  sepsin  (C5H14N2O2)  obtained  by  Faust  from 
putrefying  yeast  cells. 

They  are  not  in  all  cases  protein  cleavage  products,  since  bodies 
of  the  cholin  group,  cholin,  neurin,  and  muscarin,  the  two  last  named 
highly  toxic,  are  lecithin  derivatives,  and  Samuely  points  out  that 
other  lipoid  cleavage  products,  always  present  in  decomposing  tis- 
sues, may  well  contribute  to  ptomain  production  in  the  presence  of 
a  source  of  nitrogen.  It  is  interesting  to  note  also  that  the  vegetable 
poison  muscarin,  isolated  by  Schmiedeberg  from  mushrooms,  is 
chemically  identical  with  a  toxic  base  found  by  Brieger  in  decom- 
posing fish. 

The  ptomains  are  not  poisonous  in  every  case.  The  chemically 
simpler  ones  like  methylamin,  di-  and  trimethylamin  possess  little 
or  no  toxicity.  Others  chemically  more  complex — like  cadaverin 
and  putrescin — may  be  capable  merely  of  causing  local  necrosis, 
while  sepsin,  closely  related  to  cadaverin  in  chemical  constitution, 
but  containing  oxygen,  is  a  powerful  poison  which  acts  violently 
upon  the  intestinal  blood  vessels,  causing  capillary  dilatation,  con- 
gestion, and  diapedesis.9  The  presence  of  oxygen  seems  indeed  to  be 
necessary  for  the  development  of  strong  toxicity  (Brieger,  Vaughan, 
and  ]N~ovy).  Again,  the  lecithin  derivative,  cholin,  is  but  weakly 
toxic,  while  neurin  is  exceedingly  poisonous.  In  putrefying  mix- 
tures these  toxic  bodies  appear  on  or  about  the  fifth  or  seventh  day 
after  putrefaction  sets  in,  and  disappear,  by  further  cleavage,  more 
or  less  rapidly,  yielding  less  complex  nitrogenous  substances  that  are 
non-toxic. 

With  the  limited  knowledge  regarding  bacteria  and  infectious 
diseases  at  the  disposal  of  the  earlier  investigators  it  was  but  natural 
that  the  discovery  of  ptomains  in  cultures  of  putrefactive  bacteria 
aroused  the  suspicion  that  these  bodies  were  responsible  for  the 
toxemia  of  infectious  disease. 

The  search  for  poisonous  substances  in  pure  cultures  of  patho- 
genic bacteria  was,  therefore,  assiduously  taken  up  by  Brieger  and 
his  pupils,  and,  in  truth,  ptomains  were  actually  found  as  products 
of  some  of  the  disease-producing  micro-organisms,  just  as  they  had 
been  found  in  the  mixed  cultures  involved  in  the  putrefaction  of 
meat.  Thus  cadaverin  was  found  in  cultures  of  the  cholera  spiril- 
lum, another  nitrogenous  poison,  typhotoxin,  in  those  of  typhoid 
bacilli,  and  still  another  in  tetanus  cultures,  all  of  them  producing 
more  or  less  severe  illness  when  injected  into  animals. 

In  spite  of  this  evidence,  however,  we  have  been  forced  to  con- 
clude that  the  ptomains  cannot  properly  be  held  responsible  for  bac- 

9  Meyer  and  Gottlieb.     "Experim.  Pharmakologie,"  2d  ed.,  p.  262. 


BACTERIAL    POISONS  31 

terial  toxemia  as  manifested  in  disease.  In  the  first  place  it  is 
doubtful  whether  ptomains,  in  noticeable  quantity,  are  ever  produced 
within  the  living  infected  body.  Then,  again,  potent  ptomains  are 
produced  in  culture  by  many  bacteria  having  absolutely  no  patho- 
genic power,  while  highly  pathogenic  bacteria  may  produce  little  or 
no  ptomains.  Ptomain  production,  moreover,  is  not  specific,  since 
the  same  ptomains  may  be  produced  by  many  different  bacteria  or 
mixtures  of  bacteria,  provided  the  conditions  of  nutrient  materials 
and  temperature  are  favorable  for  growth.  We  cannot  therefore 
account  for  bacterial  toxemia,  in  which  the  poison  produced  by  an 
individual  species  is  characteristic  and  invariably  the  same,  under 
varying  cultural  and  environmental  conditions,  by  the  production  of 
ptomains.  And  even  when  ptomains  are  produced  in  culture  fluids 
by  pathogenic  bacteria  their  physiological  action  is  usually  quite 
different  from  that  of  the  poisons  produced  by  the  same  micro-organ- 
isms in  the  infected  subject. 

Briefly  summarized,  therefore,  the  ptomains  are  poisons  elab- 
orated by  all  bacteria  that  are  capable  of  producing  protein  cleavage, 
if  planted  on  suitable  nutrient  materials  under  conditions  favoring 
growth.  The  matrix  of  these  poisons  is  the  protein  nutriment ;  they 
are  not  products  of  intracellular  metabolism  specifically  characteris- 
tic of  the  bacteria  which  produce  them. 

Their  importance  in  the  production  of  disease,  therefore,  is  really 
an  indirect  one.  They  may  cause  disease  if  putrid  meat  or  other 
material  is  ingested,  and  with  it  preformed  ptomains,  which  may  be 
taken  in  and  further  elaborated  by  continued  putrefaction  in  the 
intestines.  This  form  of  meat  poisoning,  without  bacteriological  in- 
vestigation, may  be  difficult  to  distinguish  from'  such  bacterial  forms 
of  meat  poisoning  as  those  caused  by  the  Gartner  bacillus  or  the 
bacillus  botulinus.  Novy  10  believes  that  true  ptomain  poisoning  of 
this  kind  is  rather  less  frequent  than  formerly  supposed.  However, 
in  such  cases  as  those  of  Vaughan,  who  isolated  a  poisonous  ptomain 
"tyrotoxicon"  from  cheese  and  milk,  their  importance  seems  rea- 
sonably certain.  It  is  also  probable  that  certain  forms  of  auto- 
intoxication may  be  caused  by  the  production  in  the  intestinal  ca- 
nal of  ptomains  resulting  from  bacterial  putrefaction  incident  to 
faulty  digestive  conditions.  It  is  the  antagonism  to  such  intes- 
tinal putrefaction  by  the  acid  production  of  the  bacillus  Bul- 
garicus  which  is  probably  the  basic  cause  of  any  favorable  thera- 
peutic effects  which  have  attended  the  soured  milk  therapy  of 
Metchnikoff.  Again  the  growth  of  saprophytes  in  necrotic  tissues 
such  as  gangrenous  extremities  in  diabetes  or  amputation  stumps, 
may  lead  to  the  formation  of  ptomains  which,  after  absorption,  can 
cause  disease.  In  all  such  cases  the  process  is  one  determined  by  the 
bacterial  putrefaction  of  dead  organic  materials,  and  the  absorbed 
10  Novy  in  Osier's  "Modern  Medicine,"  Vol.  1,  p.  223. 


32  INFECTION    AND    RESISTANCE 

poisons  are  not  true  bacterial  toxins,  since  they  do  not  emanate 
specifically  from  the  cell  substance  of  the  micro-organisms  but  rather 
represent  incidental  cleavage  products  of  the  nutrient  materials. 
Therefore,  also,  the  ptomains  are  unspecific — their  formation  a  com- 
mon attribute  of  a  large  variety  of  saprophytic  organisms,  their 
production,  as  to  quantity  and  kind,  primarily  dependent  upon 
the  nature  of  the  nutrient  materials  on  which  the  bacteria  are 
grown. 

In  contradistinction  to  the  ptomains,  the  specific  bacterial  poi- 
sons, in  the  technical  meaning  of  the  term,  are  substances  which  are 
characteristic  for  each  individual  species  of  bacteria  and  truly  the 
products  of  bacterial  metabolism  in  that  they  emanate  from  the  cell 
itself,  either  as  a  secretion  or  excretion  during  cell  life,  or  as  an 
inherent  element  of  the  cytoplasm  liberated  after  death  (or  possibly 
as  a  cleavage  product  of  the  disintegrating  bacterial  protein).11 
They  are  dependent  upon  the  nature  of  the  culture  medium  only  in 
so  far  as  this  favors  or  retards  the  normal  development  of  the  micro- 
organisms. While,  therefore,  a  diphtheria  bacillus  undoubtedly  pro- 
duces the  largest  quantities  of  its  specific  poison  on  bouillon  suitably 
prepared  for  this  particular  purpose,  it  will  also,  in  smaller  amount, 
produce  qualitatively  the  same  poison  on  all  media  on  which  its 
growth  is  free  and  uninhibited,  even  on  a  medium  such  as  that  of 
Uschinsky,  which  is  entirely  devoid  of  proteins.  The  toxins  are, 
therefore,  elements  of  intracellul&r  metabolism,  permanently  or 
transiently  constituent  parts  of  the  cell  body. 

A  specific  bacterial  toxin  was  first  obtained  from  the  diphtheria 
bacillus  by  Roux  and  Yersin12  in  1889.  They  discovered  that  if 
diphtheria  bacilli  were  grown  on  veal  broth  and  the  cultures  filtered 
through  porcelain  candles,  after  seven  days  at  37.5°  C.  the  filtrates 
were  highly  toxic,  producing  the  same  symptoms  and  autopsy  find- 
ings in  rabbits,  guinea  pigs  and  birds  which  followed  the  injection 
of  the  living  bacilli  themselves.  The  poison  was  therefore  a  soluble 
product  of  the  bacteria  during  the  period  of  their  vigorous  growth, 
apparently  given  up  by  them  to  the  culture'  fluid.  Very  soon  after 
this,  in  1891,  Kitasato  13  discovered  a  similar  specific  toxin  in  cul- 
ture filtrates  of  the  tetanus  bacillus,  and  it  was  the  hope  of  bacteri- 
ologists that  analogous  poisons  could  be  determined  for  all  patho- 
genic bacteria. 

This  hope,  however,  has  been  disappointed.  It  was  soon  found 
that  cultures  of  cholera  spirilla,  typhoid  bacilli,  and  many  other 
germs  did  not  yield  toxic  filtrates  of  this  kind  but  that  the  poisons 
in  these  cases  seemed  to  be  firmly  bound  to  the  bacterial  bodies  dur- 

11  In  connection  with  this  read  the  discussion  on  anaphylaxis  in  chapter 
XVII,  p.  413. 

12  Roux  and  Yersin.     Ann.  de  I'Inst.  Pasteur,  Vol.  2,  1889. 

13  Kitasato.     Zeitschr.  f.  Hyg.,  1891,  Vol.  10. 


BACTERIAL    POISONS  33 

ing  life,  and  given  tip  to  the  surrounding  media  only  after  death 
and  disintegration  of  the  cells. 

Pfeiffer  14  was  the  first  one  to  formulate  this  conception  in  his 
studies  upon  cholera  poisons.  He  found  that  when  cholera  spirilla 
were  grown  upon  broth  and  filtered  after  6  or  7  days,  the  filtrate  was 
hut  slightly  toxic,  but  that,  in  this  case,  unlike  the  conditions  pre- 
vailing in  diphtheria  and  tetanus  cultures,  the  residue  of  bacterial 
cell  bodies,  even  after  they  had  been  killed  by  chloroform,  thymol, 
or  drying,  were  powerfully  poisonous. 

We  have  then  two  main  classes  of  specific  bacterial  poisons.  One 
— typified  by  diphtheria  and  tetanus  poisons — is  produced  during 
the  period  of  energetic  growth  by  the  living  bacteria,  is  given  off  to 
the  surrounding  culture  fluid  as  a  secretion  or  excretion,  and  can  be 
obtained  in  bacteria-free  filtrates  at  a  time  when  few,  if  any,  of  the 
micro-organisms  have  died  or  disintegrated.  These  are  spoken  of  as 
"true  toxins"  or  "exotoxins." 

The  other  group — typified  by  the  cholera  poisons  as  described  by 
Pfeiffer — is  apparently  an  intracellular,  constituent  part  of  the  bac- 
terial body — not  given  off  during  life  and  not,  therefore,  obtained  in 
filtrates  of  young  living  cultures.  If  the  cultures  are  preserved  until 
cell  death  has  taken  place  and  the  dead  bodies  have  been  extracted 
by  the  culture  fluid,  the  filtrate  becomes  gradually  more  toxic.  The 
bodies  of  such  bacteria  are  in  themselves  powerfully  toxic  when 
injected,  dead  or  alive.  These  poisons  for  obvious  reasons  Pfeiffer 
has  named  the  "endotoxins"  since  he  regarded  them  as  specific  and 
definite  substances,  present  as  such  in  the  living  bacterial  cell. 

In  addition  to  the  endotoxins  the  bacterial  protein  contains  sub- 
stances which  attract  and  lead  to  the  accumulation  of  leukocytes.  In 
other  words,  they  exert  a  positive  chemotactic  influence.  This  was 
first  observed  in  1884  by  Leber,15  who  induced  the  formation  of  pus 
hy  injecting  dead  staphylococcus  cultures,  and,  later,  found  that  the 
same  effect  resulted  from  the  injection  of  alcoholic  extracts  of 
staphylococci.  These  chemotaxis-inducing  substances  were  later 
particularly  studied  by  Buchner.  Buchner  16  extracted  them  from 
many  varieties  of  bacteria,  independent  of  pathogenicity.  Although 
there  are  quantitative  differences,  all  bacteria  seem  to  contain  such 
substances,  and  Buchner  believed  the  chemotactic  property  to  be  a 
general  attribute  of  the  bacterial  protoplasm.  He  speaks  of  his  ex- 
tracts as  bacterial  proteins. 

The  true  toxins  or  exotoxins,  then,  appear  to  be  products  of  liv- 
ing bacteria  given  off  from  these  very  much  as  are  the  ferments  and 
enzymes  by  which  micro-organisms  cause  cleavage  of  carbohydrates 
or  proteins — and  indeed  the  French  school,  from  the  first,  compared 

14  Pfeiffer.     Zeitschr  /.  Hyg.,  Vol.  II,  1892. 

15  Leber.    "Tiber  die  Entziindung,"  Leipzig,  1884. 

16  Buchner.    Berl.  klin.  Woch.,  1890. 


34  INFECTION    AND    RESISTANCE 

these  toxins  to  enzymes,  with  which,  as  we  shall  see,  they  have  much 
in  common.  The  endotoxins — on  the  other  hand — at  least  as  con- 
ceived by  Pfeiffer,  are  structural  ingredients  of  the  bacterial  proto- 
plasm which  are  toxic  when  brought  into  solution  as  the  cells  break 
up. 

Concerning  the  accuracy  of  this  conception,  however,  much  doubt 
has  recently  arisen,  as  a  result  of  researches  which  will  be  discussed 
below. 

These  two  types  of  poison,  moreover,  differ  from  each  other  not 
only  in  mode  of  origin  but  in  biological  characteristics  far  more 
fundamental  than  this. 

The  discovery  of  diphtheria  toxin  by  Roux  and  Yersin  was  fol- 
lowed by  diligent  investigations  into  the  toxic  properties  of  all 
known  pathogenic  bacteria,  and  it  was  soon  found  that  a  few  only 
of  these  germs  could  produce  poisons  biologically  similar  to  that 
found  in  diphtheria  cultures.  It  was  in  the  course  of  investigations 
of  this  kind,  indeed,  that  Pfeiffer,  failing  to  discover  an  exotoxin  in 
cultures  of  cholera  and  other  germs,  formulated  his  endotoxin  theory. 

The  list  of  true  toxin  or  exotoxin  producers,  then,  is  short. 
Among  the  more  important  are,  in  addition  to  the  diphtheria  and 
tetanus  bacilli — which  have  been  mentioned  above — the  Bacillus 
botulinus,17  the  Bacillus  pyocyaneus,18  and  that  of  symptomatic  an- 
thrax.19 It  has  also  been  claimed  that  similar  toxins  are  formed  by 
the  cholera  spirillum  (Brau  and  Denier),20  by  the  dysentery  bacillus 
of  the  Shiga-Kruse  type  (Kraus  and  Doerr)  21  and  the  Bacillus  ty- 
phosus  (Arima).22  In  the  cases  of  the  three  last-named  organisms, 
however,  the  secretion  of  a  true  exotoxin  has  not  been  accepted  as  a 
fact  by  all  observers.  Indeed,  even  though  such  substances  may  pos- 
sibly be  produced  by  these  bacteria  in  small  amounts  it  is  not  likely, 
in  the  light  of  our  present  knowledge,  that  they  play  more  than  a  sec- 
ondary role  in  the  toxemic  manifestations  of  cholera,  dysentery,  and 
typhoid,  the  important  poisons  in  these  cases  being  those  derived 
from  the  bacterial  cell  bodies. 

Similar  in  essential  properties  to  the  true  exotoxins  also  are  the 
erythrocyte  poisons  (hemotoxins)  produced  by  many  bacteria  which 
cause  hemolysis  of  red  cells,  and  the  leukocyte-destroying  poison 
(leukocydin)  which  is  a  product  of  the  Staphylococcus  aureus. 

All  of  these  atrue  bacterial  toxins"  or  exotoxins,  apart  from  sim- 
ilarity of  origin,  as  soluble  secretions  of  the  living  bacteria,  possess 
certain  common  biological  characteristics  which  sharply  differentiate 

17  Kempner.     Zeitschr.  f.  Hyg.,  Vol.  26,  1897. 

18  Wassermann.    Zeitschr.  f.  Hyg.,  Vol.  22,  1896. 

19  Grassberger  and  Schattenfroh.    Wien  Deuticke,  1904. 

20  Brau  and  Denier.    Ann.  de  I'Inst.  Past.,  Vol.  20,  1906. 

21  Kraus  and  Doerr.     Wien  kl.  Woch.,  42,  1905. 
22 Arima.     Centralbl.  f.  Bakt.,  I,  Vol.  63,  1912. 


BACTERIAL    POISONS  35 

them  from  the  "endotoxins."  These  characteristics  they  share  with 
a  number  of  non-bacterial  substances  such  as  the  vegetable  poisons 
ricin,  crotin,  and  abrin,  with  animal  poisons  like  snake  venom  and 
spider  poison  (arachnolysin),  and,  in  certain  important  respects, 
with  the  substances  spoken  of  as  enzymes. 

Thus  the  bacterial  true  toxins  are  not  biologically  unique  sub- 
stances. Both  in  themselves  and  in  regard  to  the  reactions  they 
elicit  when  injected  into  the  animal  body,  they  share  certain  cardinal 
properties  with  analogous  substances  derived  from  the  higher  plants 
and  from  animals.  And  it  is  important  to  recognize  at  once  that  we 
are  dealing  here,  as  in  other  phases  of  the  study  of  bacterial  immu- 
nity, with  broad  biological  laws,  which  find  application  not  only  in 
bacteriology,  but  in  general  pathology  and  in  the  phenomena  of  pro- 
tein metabolism  in  general.  It  so  happens  that  these  phenomena 
have  been  studied  and  are  most  easily  elucidated  in  connection  with 
bacteria.  But  their  general  significance  must  not  be  lost  sight  of. 

The  cardinal  characteristic  which  unites  all  of  these  substances 
into  a  single  well-defined  biological  group  is  their  property  of  in- 
ducing the  formation  of  antitoxins  when  injected  into  animals.  This 
property  is  so  important  and  its  thorough  comprehension  so  essential 
that  we  may  be  permitted  to  digress  briefly  in  order  to  make  it  clear. 

As  we  shall  see,  in  subsequent  chapters,  all  substances  which  lead 
to  the  formation  of  specifically  reacting  antibodies  in  the  treated 
animal  are  spoken  of  as  "antigens"  or  "antibody-inducing  sub- 
stances." The  class  of  "antigens"  is  a  large  one,  including  all 
known  proteins,  and  possibly  some  of  the  higher  proteid  split  prod- 
ucts, and  protein-lipoid  combinations,  though  the  "antigenic"  prop- 
erties of  the  last  two  are  still  in  controversy.  But  among  this 
large  group  of  substances  it  is  only  the  bacterial  true  toxins  (exo- 
toxins),  obtained  in  broth  filtrates  of  living  cultures,  together 
with  the  vegetable  poisons  and  other  substances  we  have  classified 
with  them  above,  which  induce  in  the  blood  of  the  treated  animal  a 
neutralizing  antibody — (antitoxin) — which  inhibits  quantity  for 
quantity  the  activity  of  the  injected  toxin  or  vegetable  or  animal 
poison.  This  property  of  eliciting  the  production  of  antitoxin  in  the 
animal  body  alone  separates  these  substances  sharply  from  all  other 
antigens,  toxic  or  otherwise,  and,  in  this  respect,  they  differ  sharply 
from  the  so-called  "endotoxins"  against  which  no  antitoxins  can  be 
produced. 

As  an  important  secondary  characteristic  of  this  group  of  sub- 
stances we  may  regard  their  chemically  indefinable  nature.  In  the 
case  of  none  of  them  have  we  any  definite  knowledge  of  chemical 
constitution  except  in  so  far  as  it  has  been  hitherto  impossible  to 
separate  them  from  the  protein  molecule.  The  intensive  chemical 
study  of  the  toxins  has  universally  resulted  in  failure  to  obtain  a 
protein-free  product  which  has  the  characteristic  toxic  properties  of 


36  INFECTION    AND    RESISTANCE 

the  original  filtrate,  or  its  antitoxin-inducing  power.  Concerning 
the  methods  which  have  been  employed  in  the  study  of  the  chemistry 
of  these  substances  we  will  have  more  to  say  in  another  place.23  It 
is  safe  to  summarize  all  this  work  for  our  present  purposes,  by  stat- 
ing that,  whatever  the  method  employed,  until  now  all  of  the  prep- 
arations obtained  have  given  one  or  another  of  the  protein  type- 
reactions,  and  that  none  of  them  can  be  positively  accepted  as  pro- 
tein-free. The  results  here  obtained  have  been  entirely  analogous  to 
those  obtained  in  similar  investigations  upon  enzymes.  (See  also 
discussion  of  antigens,  chapter  4.) 

The  analogy  with  enzymes  is  indeed  a  striking  one  and  noted  by 
the  first  investigators  of  a  true  toxin,  Roux  and  Yersin.  Biolog- 
ically, of  course,  we  have  the  cardinal  similarity  in  that  the  injec- 
tion of  toxins  into  animals  induces  the  production  of  antitoxin,  and 
treatment  with  enzymes  induces  specific  and  neutralizing  anti-en- 
zymes. In  addition  to  this,  they  are  alike  in  their  susceptibility  to 
heat  (both  being  destroyed  when  in  solution  by  temperatures  over 
80°  C.),  in  their  gradual  deterioration  on  standing,  and  their  mys- 
terious activity  in  small  quantities  upon  disproportionately  larger 
masses  of  the  substances  they  attack.  There  is,  however,  one  impor- 
tant difference  between  the  two  in  their  mode  of  action.  For,  while 
the  toxins  are  apparently  bound  or  neutralized  by  the  tissues  they 
attack,  the  action  of  an  enzyme  seems  rather  to  be  a  process  in  which 
the  enzyme  unites  with  the  substance  it  acts  upon,  is  released  as  the 
result  is  attained,  and  freed  for  further  action,  without  noticeable 
loss  of  quantity.  Such  catalytic  properties  have  not  yet  been  satis- 
factorily demonstrated  for  the  bacterial  toxins.  However,  there  are 
other  modifying  factors  which  may  account  for  lack  of  similarity  in 
this  respect,  and  in  all  other  important  points  the  two  classes  of  sub- 
stances are  closely  analogous. 

The  property  of  heat  sensitiveness,  which  is  a  characteristic  of 
bacterial  exotoxins  and  enzymes,  is  shared  with  them  by  all  of  the 
substances  mentioned  above  except  snake  venoms.  Snake  venoms 
are  not  destroyed  completely  until  the  temperature  is  raised  to  75°- 
80°  C.  The  earlier  contention  of  Leclainche  and  Yallee,  that  the 
toxin  of  symptomatic  anthrax  possessed  similar  heat  stability  has 
been  satisfactorily  refuted  by  Grassberger  and  Schattenfroh,24  who 
.find  that  heating  it  to  50°  C.  for  an  hour  completely  destroys  it. 

There  is  another  important  attribute  of  the  true  toxin  which 
(deserves  discussion,  though  we  are  by  no  means  in  a  position  to  offer 
any  satisfactory  explanation  for  it.  We  refer  to  the  incubation  time 
which  elapses  between  the  administration  of  a  toxin  and  the  occur- 

23  An  extensive  and  authoritative  summary  of  this  phase  of  the  subject  is 
that  of  E.  Pick  in  "Kolle  u.  Wassermann  Handbuch,"  etc.,  2d  ed.,  Vol.   1. 

24  Grassberger    and    Schattenfroh.      "tiber    das    Rauschbrandgift,    etc.," 
Wien.  Deuticke,  1904. 


BACTERIAL    POISONS  37 

rence  of  symptoms.  Here  again  snake  poisons  form  an  exception — 
since  local  manifestations  may  appear  within  an  extremely  short 
period  after  the  injection  of  the  venom  or  as  the  result  of  a  snake 
bite.  However,  in  the  case  of  all  other  toxins  there  is  a  definite 
lapse  of  time  between  the  entrance  of  the  poison  and  the  first  symp- 
toms, local  or  general.  This  interval  is  longer  when  small  doses  are 
given — shorter  when  the  doses  are  large — but  is  never  entirely  elim- 
inated— even  when  many  times  the  fatal  dose  is  given. 

In  the  case  of  tetanus  poison,  for  instance,  injections  into  a  horse 
may  not  cause  symptoms  for  as  long  as  four  or  five  days.  In  mice, 
animals  that  are  extremely  susceptible,  the  incubation  time  may  be 
shortened  from  36  to  12  hours  if  we  inject  3,600  lethal  doses,  but, 
in  any  case,  whatever  the  dose,  this  interval  cannot  be  shortened 
below  8  or  9  hours.25  Many  attempts  have  been  made  to  explain 
this.  Ehrlich,  as  we  shall  see,  assumes  that  the  action  of  a  poison 
depends  upon  two  occurrences:  one,  the  union  of  the  poison  with 
the  vulnerable  cell,  the  other  the  gradual  injury  of  the  cell  by  the 
toxic  atom  groups  in  the  poison  molecule.  The  time  necessary  for 
the  institution  of  this  process,  he  believes,  explains  the  interval. 
Eichet  has  suggested  that  the  toxin  itself  may  not  be  potent  until 
acted  upon  by  the  body  of  the  recipient  and  transformed  into  a 
potent  form.  His  views  are  more  directly  related  to  the  phenom- 
enon of  anaphylaxis  and  are  discussed  in  another  section.  De  Waele 
has  recently  advanced  a  theory  which  implies  that  the  incubation 
time  represents  the  period  necessary  for  the  gradual  concentration 
of  the  poisons  in  the  vulnerable  tissues,  a  process  which  depends 
either  upon  chemical  affinities  or  solubility  of  the  toxins  in  the  cell 
lipoids.  A  little  at  a  time  would  then  be  absorbed  by  the  vulnerable 
cells  as  they  come  in  contact  with  the  poison,  through  the  circulation, 
and  the  symptoms  would  not  appear  until  a  definite  intracellular 
concentration  had  been  attained.  His  views  are  so  closely  bound  up 
with  the  theories  on  the  selective  action  of  the  toxins  upon  individual 
tissues  and  organs  that  they  will  be  rendered  clear  as  we  proceed 
with  a  discussion  of  the  latter. 

The  majority  of  pathogenic  bacteria  do  not,  as  we  have  seen,  pro- 
duce true  toxins  or  exotoxins.  Cultures  of  cholera  spirilla,  plague 
bacilli,  and  of  many  other  bacteria  do  not  yield  toxic  filtrates  until 
the  cultures  have  been  allowed  to  stand  for  prolonged  periods  during 
which  extraction  and  possibly  autolysis  have  occurred.  In  these 
cases,  moreover,  definite  toxic  properties  can  be  demonstrated  in  the 
dead  cell  bodies  or  in  extracts  prepared  by  various  methods.  In  no 
case,  however,  is  the  injection  of  these  "endotoxins"  followed  by  the 
production  of  antitoxins.  It  was  very  natural  to  suppose  that  in 
micro-organisms  of  this  class  the  toxic  principle  might  be  present  in 
the  form  of  a  preformed  intracellular  poison  which  could  be  ex- 
25  De  Waele.  Zeitschr.  /.  7mm.,  Vol.  4,  1910. 


38  INFECTION    AND    RESISTANCE 

tracted  or  which  became  free  as  cell-death  occurred  and  disintegra- 
tion ensued. 

It  was  assumed  that,  when  bacteria  entered  the  animal  body  and 
were  destroyed  by  the  action  of  the  serum  or  cells,  these  endotoxins 
were  liberated  and  poisoning  resulted.  The  very  protective  action 
of  the  serum,  which  prevented  the  extension  of  the  infectious  in- 
vasion, by  limiting  bacterial  growth,  was  thus  looked  upon  as  the 
agency  by  whicji  the  endotoxins,  toxalburnins,  were  set  free.  Ex- 
periments by  Radziewsky  and  others,  in  which  it  was  shown  that  large 
doses  of  bacteria  injected  into  immunized  animals  were  violently 
toxic  and  more  rapidly  fatal  than  corresponding  amounts  injected  into 
normal  animals,  were  taken  to  mean  that  in  the  immune  animals  a 
more  powerfully  cell-destroying  property  of  the  serum  led  to  a  more 
rapid  liberation  of  the  endotoxins. 

This  was  the  conception  of  Pfeiffer  and,  in  more  recent  theoret- 
ical discussions,  that  of  Wolff-Eisner.  Its  essential  features  con- 
sisted in  the  assumption  that  the  poisons  were  preformed  and  were 
contained  within  the  cell  body  as  such,  and  that  they  were  specific  for 
each  micro-organism,  determining  to  a  certain  extent  its  pathogenic 
properties.  Thus  typhoid  endotoxin,  cholera  endotoxin,  or  dysen- 
tery endotoxin  was  supposed  each  to  possess  its  own  particular 
pharmacological  properties  by  which  the  clinical  manifestations  of 
the  respective  diseases  were  partially  determined. 

It  is  chiefly  the  work  of  Vaughan  26  which  has  begun  to  throw 
doubt  upon  Pfeiffer's  original  views,  in  that  Vaughan  has  shown 
that  all  proteins,  bacterial  or  otherwise,  would  yield,  upon  cleavage 
with  alkalinized  alcohol,  toxic  split  products  which  possessed  many 
of  the  pharmacological  properties  of  the  so-called  endotoxins.  In 
fact,  Vaughan  succeeded  in  producing,  in  animals,  fever  and  other 
symptoms  which  are  generally  associated  with  infection,  merely  by 
injecting  into  them  graded  quantities  of  his  toxic  split  products. 

Following  Vaughan,  Friedberger  succeeded  in  showing  that 
toxic  substances  similar  to  Vaughan' s  split  products  are  formed 
when  bacteria  of  various  species  are  subjected  to  the  action  of  nor- 
mal or  immune  sera,  and  that  such  poisons  were  pharmacologically 
alike  and  produced  with  equal  ease  from  pathogenic  and  non-patho- 
genic micro-organisms.  These  phenomena  are  discussed  in  greater 
detail  in  our  section  on  bacterial  anaphylaxis.  It  is  necessary,  how- 
ever, to  point  out  in  this  place  the  uncertainty  in  which  these  re- 
searches have  left  the  conception  of  endotoxins.  They  suggest  that 
the  toxic  effects  following  upon  the  introduction  of  pathogenic  bac- 
teria into  the  animal  body  are  not  due  to  endotoxins,  but  are  rather 
the  result  of  the  action  of  toxic  cleavage  products  formed  in  the  re- 
action between  blood  plasma  and  bacterial  cell.  These  split  products 

26  For  a  complete  discussion  of  Vaughan's  work  see  Vaughan,  "Protein 
Split  Products,"  Lea  &  Febiger,  Phila.  and  N.  Y.,  1913. 


BACTERIAL  POISONS  39 

are  not  conceived  as  specific  for  individual  bacteria  but  may  be 
formed  from  all  bacterial  proteins,  both  the  pathogenic  and  the  non- 
pathogenic.  The  differences  in  pathogenicity  between  bacteria  of 
this  class  would  then  depend  entirely  upon  their  powers  to  invade — 
not  at  all  upon  their  possession  of  individually  peculiar  cell  poisons. 
The  differences  in  clinical  course  and  toxemic  manifestations  would 
be  taken  to  depend  entirely  upon  the  accumulation  and  the  distribu- 
tion of  the  invading  germs,  and  the  consequently  variable  energy  in 
the  production  of  the  toxic  split  products  from  them.  Considerable 
experimental  evidence  has  accumulated  in  favor  of  this  point  of 
view.  We  will  reserve  a  consideration  of  this  for  a  later  chapter. 
The  entire  question  of  "endotoxins,"  or  rather  the  problem  of 
the  mechanism  by  which  such  bacteria  as  typhoid  bacilli,  plague 
bacilli  (and  other  organisms  which  do  not  produce  exotoxins)  poison 
the  animal  body,  must  be  subjected  to  experimental  revision.  In 
addition  to  the  idea  of  toxic  split  products  in  the  sense  of  Vaughan 
and  Friedberger,  there  are  other  alternatives.  Jobling  and  Petersen 
have  shown  that  bacteria  injected  into  the  circulation  may  absorb 
lipoidal  substances  which  ordinarily  act  as  anti-enzymes.  In  con- 
sequence of  this,  serum  protease  may  be  liberated  to  act  upon  the 
plasma  itself,  and  produce  toxic  substances. 

Again,  it  is  well  known  that  the  bacterial  cells  are  relatively 
poor  in  coagulable  protein,  and  we  have  shown  with  one  of  our  stu- 
dents (Aronovitch)  that  primary  and  secondary  proteoses  may  be 
obtained  in  considerable  quantities  in  bacterial  extracts.  It  is  not 
impossible  that  these  in  themselves  may  have  toxic  functions  when 
liberated,  without  further  splitting.  This  particular  subject  finds  a 
more  extensive  discussion  in  the  chapter  on  bacterial  anaphylaxis. 

In  order  to  do  injury  to  the  infected  individual  the  bacterial 
poisons  must  be  produced  in  such  locations  that  they  can  easily  enter 
the  physiological  interior  of  the  body.  None  of  the  poisons  that 
have  been  so  far  investigated  can  produce  injury  when  introduced 
into  the  alimentary  canal.  In  this  location  they  are,  as  a  rule,  de- 
stroyed, or  they  pass  through  without  doing  harm.  Neither  diph- 
theria toxin  nor  tetanus  toxin  will  produce  symptoms  when  intro- 
duced intra-intestinally.27  28  29  30  Even  cholera  poison  does  not  pass 
through  the  uninjured  intestinal  wall.  Kruse  31  assumes,  and  Kolle 
and  Schiirmann  32  seem  to  agree  with  him,  that  the  absorption  of 
cholera  poison  does  not  occur  until  the  intestinal  wall  has  been  in- 

27  Meyer  and   Gottlieb.     "Exp.   Pharmakol.,"   Urban   &   Schwartzenberg, 
Berlin,  1911. 

28  Ransom.    Deutsche  med.  Woch.,  No.  8,  1898. 

29  Neneki.     Centralbl.  f.  Bakt.,  Vol.  23,  1898. 

30  Carriere.    Ann.  de  I'Inst.  Past.,  Vol.  13,  1899. 

31  Kruse.     "Allgemeine  Mikrobiologie,"  Vogel,  Leipzig,  1910,  p.  934. 

32  Kolle  and  Sclmrmann  in  "Kolle  u.  Wassermann  Haudbuch,"  2d  Ed.? 
Vol.  4. 


40  INFECTION  AND  RESISTANCE 

jured  by  the  actual  growth  of  the  living  bacteria.  Kruse  calls  atten- 
tion to  experiments  by  Burgers  in  which  enormous  quantities  of 
cholera  poison,  i.  e.,  200  cultures  of  dead  or  living  cholera  bacilli, 
could  be  administered  to  healthy  guinea  pigs  and  rabbits  by  mouth 
without  harm  in  spite  of  the  fact  that  these  animals  are  definitely 
susceptible  to  the  poisons  and  although  the  poisons  are  not  injured 
by  the  intestinal  ferments.  It  is  likely  therefore  that  the  absorption 
of  poison  begins  only  after  the  bacteria  have  extensively  invaded  the 
intestinal  mucosa  and,  by  injuring  tissue,  have  opened  paths  for 
absorption.  In  the  case  of  diphtheria  probably  a  similar  condition 
exists  in  that  the  localized  injury  to  the  mucous  membrane  at  the 
point  of  lodgment  of  the  primary  infection  prepares  a  portal  of 
entry.  The  poison  of  the  Bacillus  botulinus  alone  seems  to  form  an 
exception  to  this  rule,33  since  this  substance,  though  apparently  a 
true  bacterial  toxin,  is  absorbed  directly  from  the  intestinal  canal. 
With  most  bacteria  this  problem  does  not  arise,  since  the  poisons  are 
elaborated  within  the  tissues,  where  resorption  is  a  necessary  result. 
Like  alkaloids  and  other  organic  as  well  as  inorganic  drugs,  the 
action  of  many  bacterial  poisons  is  largely  selective.  Most  of  these 
poisons  may  excite  inflammatory  reactions  if  concentrated  in  any 
part  of  the  body,  but,  in  addition  to  this,  there  is  a  specific  distribu- 
tion after  introduction  which  indicates  that  the  poison  goes  into 
selective  relationship  with  certain  tissues  and  cells.  This  fact  is  most 
clearly  illustrated  by  the  bacterial  hemotoxins  which  specifically  in- 
jure the  red  blood  cells  of  the  infected  individual  and  by  such  sub- 
stances as  the  leukocidin  produced  by  the  Staphylococcus  aureus,,  a 
poison  which  directly  and  visibly  injures  the  white  blood  cells.  Here 
the  action  is  specifically  aimed  at  a  well-defined  variety  of  body  cell. 
In  considering  this  problem  in  connection  with  infectious  disease, 
it  is  of  great  importance  to  distinguish  between  selective  injury  by  the 
poisons  transported  through  the  body  by  the  lymph,  blood,  and  other 
channels,  on  the  one  hand,  and  the  selective  lodgment  of  the  micro- 
organisms themselves  on  the  other.  The  latter  may  occasionally  de- 
pend on  local  cultural  advantages  for  the  particular  bacteria  in  one 
organ  or  another,  but  may  just  as  often  be  determined  by  the  peculiar 
manner  of  entrance  to  the  body  which  is  most  suitable  for  lodgment 
of  the  germs  in  question,  and  the  degree  of  local  resistance  at  the 
point  of  entrance,  which  determines  whether  or  not  the  infection  shall 
be  locally  limited  or  permitted  to  invade  beyond  this  point.  In  the 
case  of  a  disease  like  acute  anterior  poliomyelitis,  where  our  knowl- 
edge of  the  microorganisms  which  cause  the  disease  is  yet  in  its  in- 
fancy, it  is  impossible  to  decide  whether  the  injuries  noted  in  the 
motor  areas  of  the  cord  and  medulla  are  due  to  toxins  or  the  lodg- 
ment of  the  germs  themselves.  In  the  case  of  rabies  it  seems  reason- 
ably sure  that  the  microorganisms  themselves  select  the  nervous  sys- 
33  Madsen  in  "Kraus  u.  Levaditi,  etc.,"  Vol.  1. 


BACTERIAL  POISONS  41 

tern.  In  such  instances  as  the  injury  of  the  motor  areas  by  tetanus 
poison,  that  of  certain  peripheral  nerves  by  diphtheria  toxin,  or  even 
the  characteristic  lesions  of  post-syphilitic  maladies  like  tabes,  we 
can  be  reasonably  sure  that  we  are  dealing  with  the  specific  action 
of  the  poisons,  independent  of  actual  localized  growth  of  the  infec- 
tious agents. 

Diphtheria  toxin,  after  distribution  through  the  body,  may  act 
upon  many  different  tissues,  as  is  evident  by  degenerations  in  the 
heart  muscle,  liver,  and  kidney,  and  the  petechial  hemorrhages  in 
serous  surfaces.  In  addition  to  this  general  action,  however,  there  is 
a  very  marked  selection  of  certain  nerve  centers.  By  Meyer  and 
Gottlieb  34  diphtheria  toxin  is  classed  as  a  specific  vascular  poison. 
Its  action  results  in  a  rapid  sinking  of  the  blood  pressure  with  final 
cardiac  death  in  spite  of  artificial  respiration.  These  manifestations 
seem  to  have  a  central  origin,  with  particular  action  upon  the  vagi 
and  the  phrenic  nerves.  Apparently  also  the  localization  of  the 
diphtheritic  lesion  may  influence  the  selection  of  individual  nerves, 
the  most  concentrated  action  taking  place  upon  the  nerves  whose 
endings  are  distributed  in  this  particular  region,  for,  as  Meyer  and 
Ransom  35  have  shown,  this  poison,  like  tetanus  toxin,  may  be  ab- 
sorbed into  the  nerves  directly  through  the  nerve  endings.  An  in- 
teresting selective  action  also  of  diphtheria  poison  is  the  apparently 
specific  alteration  of  the  suprarenal  glands  which  is  regularly  no- 
ticed, as  enlargement  and  congestion,  in  diphtheria-infected  guinea 
pigs,  and  which  has  been  associated  by  many  workers  with  the  char- 
acteristic drop  in  blood  pressure  which  accompanies  all  severe  cases 
of  the  disease.  Abramow  36  has  studied  this  lesion  particularly,  and 
believes  that  it  consists  in  a  degeneration  and  final  disappearance  of 
the  chromafnn  substance  and  of  the  medullary  cells.  He  believes 
that  this,  together  with  degeneration  of  the  heart  muscle  itself,  is  of 
great  importance  in  causing  the  characteristic  vascular  failure. 

In  botulinus  poisoning  there  is,  as  Marinesco  37  and  Kempner 
and  Pollack  38  have  shown,  a  direct  effect  upon  the  cells  of  the  an- 
terior horns  with  degenerative  changes  in  the  I^issl  granules. 

Tetanus  poison,  which  has  been  studied  extensively  by  pharma- 
cologists, shows  a  very  marked  affinity  for  the  nervous  system,  as,  in 
fact,  the  symptoms  of  tetanus  indicate.  Indeed,  while  many  of  the 
bacterial  poisons  are  distributed  by  the  blood  stream  to  the  point  of 
final  attack,  in  tetanus  the  absorption  of  the  toxin  from  the  lesion  or 

34  Meyer  and  Gottlieb.     "Pharmacology  Trans.  Halsey,"  Lippincott,  1914, 
p.  556. 

35  Meyer  and  Ransom.    Arch,  de  pharmacodyn.,  Vol.  15,  1905,  also  Meyer, 
Berl  klin.  Woch.,  25  and  26,  1909,  also  Arch.  f.  exp.  Path.  u.  Ther.,  Vol. 
60,  1909. 

36  Abramow.     Zeitschr.  f.  Imm.,  Vol.  15,  1912. 

37  Marinesco.     Compt.  rend,  de  la  soc.  de  biol.,  Vol.  3,  1896. 

38  Kempner  and  Pollack.    Deutsche  med.  Woch.,  32,  1897. 


42  INFECTION  AND  RESISTANCE 

the  point  of  injection  takes  place  entirely  by  the  path  of  the  nerves. 
That  this  method  of  poison  distribution  might  be,  among  others, 
an  important  one  was  suggested  as  early  as  1892  by  Bruschettini,39 
who  found  tetanus  toxin  in  the  nerves  but  not  in  the  adjacent  muscle 
and  other  tissues  surrounding  the  point  of  subcutaneous  injection. 
Similar  results  were  obtained  subsequently  by  Hans  Meyer,  whose 
experiments  were  confirmed  and  extended  by  Marie  and  Morax.40 
Finally  Meyer  and  Ransom41  furnished  complete  proof  that  the 
poison  was  absorbed  from  the  blood  and  tissues  by  the  peripheral 
nerve  endings  alone  and  was  transported  centripetally  only  by  the 
paths  of  the  neurons.  The  experimental  facts  elicited  may  be  sum- 
marized as  -follows : 

1.  When  tetanus  toxin  is  injected  into  the  thigh  muscles  of  a 
guinea  pig  the  poison  is  found  at  first  only  in  the  sciatic  nerve  of  the 
same  side  and  in  the  blood.     (The  determination  of  poison  was  made 
by  injecting  macerations  of  the  respective  tissues  into  mice.)     If  ex- 
amination was  delayed  until  the  symptoms  had  become  generalized, 
the  poison  was  found  in  the  opposite  sciatic,  but  the  muscle  bundles, 
fat,  etc.,  from  the  vicinity  of  the  injection  area  were  poison-free.42 

2.  When  a  nerve  is  cut  poison  absorption  ceases  as  soon  as  axis 
cylinder  degeneration  has  set  in. 

3.  If  the  nerve  is  cut  before  the  poison  is  injected  the  distal 
end  contains  poison,  the  proximal  end  does  not.     This  again  shows 
that  the  nerve  absorbs  the  toxin  not  from  its  capillaries  but  solely 
through  the  end  organs. 

4.  If  a  nerve  which  already  contains  poison  is  severed,  toxin 
will  disappear  rapidly  from  the  proximal  end,  since  it  no  longer 
obtains  a  renewed  supply  from  the  periphery. 

5.  If  antitoxin  is  injected  into  the  nerve,  above  the  point  of 
injection,  it  will  successfully  bar  the  way  for  the  ascending  toxin. 

6.  Severing  of  the  spinal  cord  prevents  the  passage  of  the  poison 
from  below  upward. 

These  facts  ascertained  in  the  case  of  tetanus  find  their  parallel  in 
the  phenomena  of  the  distribution  of  rabic  virus  43  as  well  as  in  that 
of  poliomyelitis,  in  both  of  which  there  seems  to  be  a  progressive  cen- 
tripetal transportation  through  the  nerves.  However,  in  these  condi- 
tions we  are  probably  dealing  not  with  a  poison  but  with  a  living  virus 
and,  though  analogous,  the  conditions  are  not  entirely  comparable. 

From  the  practical  point  of  view  these  facts  regarding  tetanus 

39  Bruschettini.     Riforma  medica,  1892. 

40  Marie  and  Morax.     Ann.  de  I'Inst.  Past.,  1902. 

41  Meyer  and  Ransom.    Archiv  f  exp.  Path.  u.  Pharm.,  49,  1903. 

42  In  view  of  our  discussion  of  the  importance  of  fats  in  the  absorption 
of  tetanus  toxin,  it  seems  inconsistent  that  the  toxin  does  not  concentrate 
in  fatty  as  well  as  in  nervous  tissues.     This  Meyer  explains  by  the  inactive 
and  poorly  vascularized  condition  of  the  fat  tissues. 

43  Di  Vestea  and  Zagari.    Fortschr.  d.  Med.,  Vol.  6,  1888. 


BACTERIAL  POISONS  43 

may  explain  the  frequent  failure  of  therapeutic  success  attending 
the  injection  of  tetanus  antitoxin  after  the  symptoms  of  the  disease 
have  set  in,  since  in  such  cases  the  poison  is  already  distributed  to 
the  nerves  and  is  largely  inaccessible  to  the  antitoxin.  They  also 
have  pointed  a  way  toward  a  more  hopeful  therapy,  namely,  the 
method  of  injecting  the  antiserum  directly  into  the  nerves  about  the 
point  of  injury.  It  is  not  surprising,  however,  in  view  of  the  stated 
facts,  that  even  this  is  unsuccessful  when  done  at  too  late  a  time, 
after  a  considerable  amount  of  poison  has  already  passed  above  the 
point  of  injection  to  the  spinal  centers. 

Such  selective  action  on  the  part  of  the  bacterial  poisons  is  en- 
tirely analogous  to  the  similar  specific  action  of  alkaloids,  narcotics, 
and  other  drugs.  In  order  that  the  poison  may  act  upon  a  cell  we 
must,  of  course,  assume  that  it  has  either  chemical  or  physical  affin- 
ity for  this  cell.  The  problem,  as  many  writers  have  pointed  out,  is 
.strongly  analogous  to  that  of  tissue  staining.  A  dye  must  be  able  to 
form  a  chemical  Tinion  with  the  cell  or  it  must  be  soluble  in  the  cell 
substance  in  order  to  stain  it.  The  chemical  difference  between  cells 
is  a  delicate  one  and  not  often  definable  by  our  present  methods.  We 
can  obtain  an  insight  into  the  principles  probably  underlying  selec- 
tive action  only  by  inference  from  the  relation  between  the  chemical 
constitution  of  drugs  or  their  physical  properties,  solubility,  etc.,  and 
their  respective  tissue  affinities.  These  problems  are  difficult  and,  to  a 
large  extent,  obscure.  They  cannot  be  directly  investigated  upon 
bacterial  poisons  since  these  are  themselves  of  chemically  unknown 
nature.  But  the  study  of  drugs  of  known  constitution  has  revealed 
certain  definite  relations  of  this  kind  which  have  furnished  analogies 
from  which  the  general  principles  of  selection  in  bacterial  poisons 
•can  be  surmised. 

It  is  a  well-known  fact  to  pharmacologists  that  there  is  a  definite 
relation  between  chemical  structure  and  toxicity.  Fraenkel 44  ex- 
presses it  as  follows :  "By  the  addition  of  identical  atom  groups  in 
an  identical  manner,  similarly  acting  substances  are  obtained."  He 
cites  the  well-known  example  of  curare ;  whichever  the  path  by  which 
this  poison  is  injected  it  leaves  intact  the  tissues  with  which  it  comes 
in  contact,  but  after  general  distribution  acts  specifically  upon  the 
nerve  endings.  It  had  been  discovered  by  Brown  and  Eraser  45  that 
by  introducing  methyl  radicles  (CH3)  into  molecules  of  various  alka- 
loids, strychnin,  morphin,  atropin,  and  others,  substances  were  ob- 
tained which  paralyzed  nerve  endings,  and  this  irrespective  of  their 
previous  physiological  action.  It  appears  that  the  combination  of 
four  methyl  radicles  attached  to  the  nitrogen  atom  (quaternary  bases) 

44  Sigmund  Fraenkel.     "Arzneimittel  Synthese,"  2d   Ed.,  Springer,  Ber- 
lin, 1906. 

45  Brown  and  Fraser.     Trans.  Royal  Soc.  of  Edinburgh,  25,  1868,  cited 
from  Fraenkel. 


44  INFECTION  AND  RESISTANCE 

universally  possesses  this  paralyzing  action.     Tertiary  bases  on  the 
other  hand  lack  this  property. 

CH3 
OH3 
CH3. 
CH3. 

"Ammonium  base"  ''Tertiary  base" 

Quaternary 

Subsequently  Bohm  46  47  discovered  that  curare  contains  two  bases 
— the  one,  ucurin,"  is  slightly  toxic  and  is  a  tertiary  base;  the  other, 
which  possesses  the  typical  curare  action,  "curarin,"  is  an  "ammo- 
nium base."  By  "methylizing"  curin,  curarin  could  be  obtained. 

From  these  and  other  examples  it  is  clear  that  in  a  certain  num- 
ber of  cases  actual  chemical  affinity  must  play  a  part  in  toxic  action ; 
on  the  other  hand,  there  are  many  cases  in  which  toxic  action  seems 
to  depend  merely  upon  physical  conditions  such  as  solubilities. 
Meyer  and  Overton's  well-known  theory  of  narcosis  maintains  that 
certain  narcotics  exert  their  action  by  passing  out  of  blood  and 
lymph  solution  into  solution  by  the  fat-like,  lipoidal  substances 
(lecithin,  cholestrin,  etc.)  contained  in  the  nerve  cells,  because  the 
latter  are  better  solvents  for  them  than  is  the  blood  plasma.  This 
theory  of  Meyer  and  Overton  has  stimulated  much  investigation  and 
speculation,  and  it  is  not  unlikely  that  it  is  valid  in  the  case  of  many 
narcotics,  although  it  does  not  explain  the  action  of  narcotics  in  gen- 
eral; for  Dickson  notes  that  chloral  hydrate,  for  instance,  is  more 
soluble  in  water  than  in  oils,  and  some  narcotic  drugs  like  alcohol 
exert  definite  action  on  proteins  and  are  oxidized  in  the  body.  These 
are  pharmacological  questions  of  which  we  cannot  speak  with  author- 
ity. We  wish  merely  to  point  out  that  the  action  of  poisons  upon  the 
body  may  depend  in  some  cases  upon  mere  physical  or  mechanical 
relationship  between  the  two.48  49 

As  regards  bacterial  poisons  the  union  between  poison  and  sus- 
ceptible cell  is  extremely  firm  and  difficult  to  dissociate  in  many  in- 
stances, and  this  points  to  the  possibility  that,  in  these  cases  at  least, 
true  chemical  union  takes  place  rather  than  merely  a  loose  combina- 
tion like  that  of  the  solution  of  one  substance  in  another.  Further- 
more, the  complete  inactivation  of  some  poisons  by  mixture  with  the 

46  Bohm.     Arch,  de  Pliarm.,  cited  from  Fraenkel. 

47  See  also  Dickson,  "A  Manual  of  Pharmacology,"  E.  Arnold,  London, 
1912. 

48  Ivar  Bang.     "Biochemie  der  Lipoide,"  Bergmann,  Wiesbaden,  1911. 

49  Meyer  and  Gottlieb.     "Experimented  Pharmakologie,"  2d  Ed.,  Urban. 
&  Schwartzenberg,  Berlin,  1911. 


BACTERIAL  POISONS  45 

cells  of  tissues  capable  of  binding  them  would  likewise  point  to  more 
than  mere  physical  union.  Nevertheless,  it  does  not  by  any  means 
exclude  the  thought  that  the  poisons  may,  in  fact,  go  into  selective 
relationship  with  special  cells  because  of  physical  properties,  such  as 
solubility  in  the  lipoidal  cell  membranes,50  51  and  may  subsequently 
be  bound  chemically  or  destroyed  by  oxidation  or  enzymotic  hydroly- 
sis after  such  entrance.  In  such  a  case  the  actual  specificity  would 
yet  depend  on  purely  physical  properties. 

In  addition  to  the  specific  physical  and  chemical  affinities  be- 
tween the  poisons  by  certain  cells  there  are  probably  also  certain 
fortuitous  factors  connected  with  the  distribution  and  local  accumu- 
lation of  the  poisons  which  have  some  weight  in  determining  the 
location  of  injury.  For  the  specific  selection  is  not  absolutely  strict 
and  there  are  probably  few  parenchyma  cells  in  the  body  that  are 
entirely  insusceptible  to  injury  if  the  poisons  are  sufficiently  con- 
centrated upon  them.  Thus,  to  cite  an  analog}7"  from  the  toxicology 
of  non-bacterial  poisons,  in  lead  poisoning,  as  Meyer  and  Gottlieb 
point  out,  the  paralysis  of  the  extensors  of  the  arm  occurs  chiefly  in 
adults  who  use  these  muscles  in  the  exercise  of  their  professions 
(painters,  type-setters),  while  in  children  and  in  animals,  in  which 
no  such  selective  use  of  particular  muscle  groups  is  habitual,  lead 
paralyses  are  atypical,  attacking  legs  as  well  as  arms.  It  is  not  un- 
likely that  the  frequent  injury  of  the  heart  muscle  by  bacterial  poi- 
sons or  the  irregular  parenchymatous  changes  in  various  organs  is 
determined  by  analogous  fortuitous  factors,  in  that  functional  activ- 
ity and  increased  metabolism  may  predispose  to  injury. 

Bacterial  poisons  also  may  produce  their  lesions  in  the  course  of 
excretion.  This  seems  likely  in  the  case  of  typhoid  poisons  in  which 
we  have  often  seen  bloody  diarrhea  in  rabbits  within  a  few  hours 
after  intravenous  injection  of  powerfully  toxic  culture  filtrates.  In 
connection  with  the  dysentery  bacillus  Flexner  and  Sweet  52  have 
studied  the  conditions  carefully.  They  succeeded  in  showing  first 
that  the  introduction  of  the  dysentery  poison  into  the  lumen  of  the 
intestine  does  no  harm  and  that  the  toxin  is  slowly  destroyed  by  peptic 
and  tryptic  digestion.  They  concluded  that  probably  no  absorption 
of  the  poison  through  the  uninjured  intestinal  mucosa  takes  place. 
They  then  showed  that  the  toxin  after  intravenous  administration 
is  excreted  by  the  intestine  and  that  the  inflammatory  reactions  and 
injury  of  the  mucosa  are  incident  to  this  act  of  elimination. 

Whether  or  not  the  kidneys  are  injured  in  the  same  way  it -is 
difficult  to  decide.  In  many  infectious  diseases,  of  course,  the  bac- 

50  For  Overtoil's  theory  of  osmosis  see  R.  Hober,  "Physikalische  Chemie 
der  Zelle  u.  Gewebe,"  Leipzig,  Engelmann,  1911. 

51  Compare  also,  regarding  this  entire  question,  the  discussion  in  P.  Th. 
Miiller,  "Vorlesungen  iiber  Immunitat,  etc.,"  Fischer,  Jena,  1910. 

52  Flexner  and  Sweet.    Jour,  of  Exp.  Med.,  Vol.  8,  1906. 


46  INFECTION  AND  RESISTANCE 

teria  themselves  pass  through  the  kidney  into  the  urine,  and  renal 
injury  may  result  from  the  actual  presence  of  the  bacteria  in  the 
kidney;  however,  renal  injury  may  also  occur  without  this,  and  it 
is  not  at  all  impossible  that  the  conditions  here  are  similar  to  those 
just  described  for  the  intestine. 

All  the  facts  which  we  have  considered  indicate  that,  although 
most  bacterial  poisons  can  injure  many  different  tissues,  yet  in  some 
cases  there  is  a  particular  susceptibility  on  the  part  of  an  individual 
tissue  which  is  independent  of  accidental  factors  and  seems  to  be 
due  to  specific  chemical  or  physical  affinity.  It  seems  even  that  in 
tetanus,  botulismus,  and  a  few  other  conditions  there  is  a  differential 
selection  of  particular  areas  within  a  tissue  like  the  nervous  system, 
just  as  this  occurs  in  the  case  of  certain  drugs.  As  stated  above  we 
have  no  satisfactory  scientific  explanation  for  this,  but  a  great  deal 
of  work  has  been  done  to  show  that  the  bacterial  poisons  actually 
unite  with  and  are  taken  up  by  the  susceptible  tissues. 

Indirectly,  proof  of  this  has  been  brought  by  the  demonstration 
of  the  rapid  disappearance  of  various  toxins  from  the  blood  streams 
of  susceptible  animals  and  their  persistence  in  the  circulation  of 
animals  insusceptible  to  them.  Thus  Donitz 53  has  shown  that 
tetanus  toxin  injected  into  the  blood  stream  of  a  susceptible  animal 
rapidly  diminishes  in  quantity,  and  Knorr,54  in  similar  experiments, 
showed  .that  the  demonstrable  disappearance  of  such  toxins  out  of 
the  blood  stream  is  synchronous  with  the  appearance  of  symptoms,  a 
fact  which  excludes  disappearance  by  excretion.  Conversely  Asa- 
kawa  55  showed  that  in  pigeons,  which  are  but  slightly  susceptible, 
tetanus  poison  could  be  demonstrated  in  blood,  liver,  spleen,  kidneys, 
and  muscles  six  days  after  injection,  but  not  in  the  brain,  showing 
that  in  this  organ,  at  least,  there  must  have  been  either  a  union  or  a 
destruction  of  the  poison.  Similar  to  these  results  are  those  of 
Metchnikoff,56  who  found  the  poison  unchanged  after  two  months  in 
the  circulation  of  insusceptible  animals  (lizards). 

Direct  evidence  of  union  between  susceptible  tissues  and  poison 
has  been  furnished  by  the  experiments  of  Wassermann  and  Takaki,57 
who  showed  that  the  brain  and  cord  tissues  of  rabbits  and  guinea 
pigs,  mixed  with  tetanus  toxin  before  injection,  served  to  neutralize 
its  harmful  effects.  And  it  appears  that  the  toxin-neutralizing  prop- 
erty of  the  brain  substances  of  various  animals  is  proportionate  to 
their  individual  susceptibility  to  the  poison.  Thus  Metchnikoff  58 
not  only  confirmed  the  results  of  Wassermann  and  Takaki  for  rab- 

63  Donitz.    Deutsche  med.  Woch.,  No.  27,  1897. 

54  Knorr.    Fortschr.  der  Medizin,  1897,  No.  17,  and  Munch,  med.  Woch., 
1898,  Nos.  11  and.  12. 

55  Asakawa.     Centralbl.  f.  Bakt.,  Vol.  24,  pp.  166  and  234. 

56  Metchnikoff.     "L'Imnmnite  dans  les  maladies  Infect.,"  Paris. 

57  Wassermann  and  Takaki.    Berl.  Tclin.  Woch.,  1898,  No.  1. 

58  Metchnikoff.    Ann.  de  I'Inst.  Past.,  1898,  p.  81. 


BACTERIAL  POISONS  47 

bits  and  guinea  pigs,  but  showed  further  that  the  brains  of  chickens, 
animals  that  are  but  moderately  susceptible,  possess  a  correspond- 
ingly slighter  neutralizing  power,  and,  further,  that  brain  tissues  of 
entirely  insusceptible  cold-blooded  animals,  turtles  and  frogs,  pos- 
sess absolutely  no  neutralizing  properties. 

The  original  interpretation  by  Wassermann  of  these  facts  was 
based  on  the  assumption  that  the  poison  was  bound  to  the  brain  tissue 
just  as  it  is  bound  to  antitoxin.  Experiments  by  Besredka  59  have 
cast  some  doubt  upon  this.  This  worker's  experiments  seem  to  indi- 
cate that  a  brain  emulsion  which  has  been  saturated  with  the  toxin 
can  be  rendered  capable  of  absorbing  more  toxin  if  tetanus  antitoxin 
is  mixed  with  it.  In  other  words,  the  affinity  of  the  antitoxin  for  the 
toxin  is  stronger  than  that  of  the  brain  substance  for  the  poison,  and 
that  the  union  toxin-brain  tissue  is  very  easily  dissociated ;  as  indeed 
it  should  if  tae  union  were  purely  a  physical  one  depending  on  solu- 
bility. 

After  it  had  been  shown  that  the  poisons  which  acted  specifically 
upon  certain  cells  were  actually  taken  up  by  these  cells,  a  number  of 
attempts  were  made  to  determine  chemically  the  tissue  element  which 
united  with  the  poisons.  Xoguchi  60  showed  that  cholesterin  and 
alcoholic  extracts  of  blood  serum  neutralized  tetanolysin.  The  same 
thing  was  later  shown  by  Miiller,61  and  Landsteiner  62  showed  that 
ether  extracts  of  red  blood  cells  likewise  neutralized  this  poison. 
In  a  later  study  by  Landsteiner  and  von  Eisler  63  the  relation  of  the 
tissue  lipoids  to  various  toxic  substances  was  still  more  definitely 
established.  They  studied  first  the  various  hemolysins  and  found 
that  extraction  of  blood  cells  with  ether  rendered  the  stromata  less 
capable  of  binding  the  hemolytic  substances.  The  same  thing  they 
showed  for  bacteriolysins,  in  the  latter  case  demonstrating  at  the 
same  time  that  the  ether  extracts  of  bacterial  bodies  possessed  slight 
binding  properties  for  the  bactericidal  substances  of  the  serum.  These 
experiments  have,  of  course,  a  merely  indirect  significance  in  the 
present  connection,  since  they  do  not  deal  with  the  type  of  poisons 
we  have  discussed.  However,  Landsteiner  and  von  Eisler  also 
worked  with  tetanus  toxin  and  found  that  the  treatment  of  the  brain 
substance  of  guinea  pigs  with  ether,  by  taking  out  lipoidal  sub- 
stances, considerably  reduces  the  power  of  this  tissue  to  bind  and 
neutralize  the  tetanus  poisons. 

Takaki,64  who  investigated  these  relations  in  great  detail,  iso- 
lated an  alcohol-soluble  element,  cerebron,  from  nerve  tissues,  a  sub- 

59  Besredka.    Ann.  Past.,  1903,  p.  138. 

60  Noguchi.     Univ.  Pa.  Med.  Bull,  Nov.,  1902. 

61  Miiller.     Centralbl.  f.  Bakt.,  Vol.  34,  1903. 

62  Landsteiner.     Wien.  kl.  Bundschau,  13,  1905. 

63  Landsteiner  and  von  Eisler.     Centralbl.  f.  Bakt.,  39,  p.  318,  1905. 

64  Takaki.     Beitr.  zur  chem.  Phys.  u.  Path.,  11,  No.  19,  1908. 


48  INFECTION  AND  RESISTANCE 

stance  to  which  he  ascribes  the  toxin-binding  properties.  Overton 
and  Bang  65  found,  furthermore,  that  cholesterin  and  lecithin  inhibit 
the  action  of  cobra  venom,  a  poison  which  is  in  so  many  ways  similar 
to  those  produced  by  bacteria.  Taking  into  consideration  all  avail- 
able evidence,  we  are  forced  to  admit  that  the  lipoids  seem  to  play  an 
important  role  in  determining  the  selective  action  of  the  nervous  sys- 
tem by  the  bacterial  poisons.  It  may  not,  of  course,  be  an  influence 
depending  merely  upon  the  solubility  of  the  harmful  substances  in 
the  lipoids  themselves.  For,  as  Bang  expresses  it,  "the  lipoids  pos- 
sess to  a  high  degree  the  property  of  altering  by  their  presence  the 
solubilities  of  other  bodies/'  and  it  is  quite  possible  that  in  the  tis- 
sues they  are  present  as  lipoid-protein  combinations.  Their  action 
in  determining  the  solubility  of  toxins  in  a  given  cell  may  therefore 
be  a  purely  indirect  one. 

It  is  of  some  interest  in  this  connection  to  recall  the  experiments 
of  De  Waele,66  which  bring  out  another  clear  analogy  between  alka- 
loids and  bacterial  poisons  in  their  relation  to  lecithin.  He  found 
that  the  addition  of  small  quantities  of  lecithin  increases  the  activity 
of  both  toxins  and  alkaloids  in  the  animal  body,  whereas  larger 
amounts  inhibit  both. 

65  See  Ivar  Bang,  "Biochemie  der  Lipoide,"  Bergmann,  Wiesbaden,  1911. 

66  De  Waele.    Zeitschr.  f.  Immunit.,  Vol.  3,  1909,  p.  504. 


CHAPTER    III 

OUK   KNOWLEDGE   CONCEKNING   NATURAL   IMMU- 
NITY, ACQUIEED  IMMUNITY,  AND  AKTIFICIAL 
IMMUNIZATION 

NATURAL  RESISTANCE  AGAINST  INFECTION 

IN  the  preceding  chapters  we  have  confined  ourselves  largely  to 
the  consideration  of  those  properties  of  the  bacteria  which  determine 
their  ability  to  infect.  In  this  discussion,  however,  we  have  repeat- 
edly emphasized  the  fact  that  every  infectious  disease  is  the  result  of 
a  struggle  between  two  variable  factors — the  pathogenic  powers  of 
the  bacteria  on  the  one  hand,  and  the  resistance  of  the  subject  on  the 
other,  each  of  these  again  modified  by  variations  in  the  conditions 
under  which  the  struggle  takes  place.  Thus  a  given  micro-organism 
may  be  capable  of  causing  fatal  infection  in  one  individual  but  may 
be  only  moderately  virulent  or  even  entirely  innocuous  for  another. 
Conversely  the  same  individual  may  be  highly  susceptible  to  one  va- 
riety of  bacteria,  but  highly  resistant  to  others.  Even  in  reactions 
with  one  and  the  same  micro-organism,  the  susceptibility  or  resist- 
ance of  the  individual  may  be  determined  by  variations  in  the  physi- 
ological state  or  by  the  environmental  conditions  under  which  the 
two  factors — invader  and  invaded — are  brought  together.  There- 
fore, the  conceptions  "resistance,"  "immunity,"  and  its  opposite 
"susceptibility,"  are  relative  terms  which  can  never  be  properly  dis- 
cussed without  careful  consideration  of  all  modifying  conditions 
which  influence  them. 

The  science  of  immunity  deals  with  a  detailed  analysis  of  these 
variables.  Its  ultimate  practical  aim  is  the  determination  of  meth- 
ods by  which  an  original  susceptibility  can  be  transformed  into  re- 
sistance or  even  immunity.  And  the  rational  method  of  approach- 
ing this  subject  consists  in  a  careful  study  of  the  conditions  of  sus- 
ceptibility and  immunity  as  they  exist  naturally  in  the  animal  king- 
dom. 

The  mere  fact  that  both  animals  and  man  are  in  constant  con- 
tact with  infectious  micro-organisms,  many  of  them  in  a  high  state 
of  virulence,  indicates  in  itself  that  the  animal  disposes  normally 
over  a  defensive  mechanism  of  considerable  efficiency. 

To  a  certain  extent,  of  course,  this  escape  from  harm  is  due  to 

49 


50  INFECTION    AND    RESISTANCE 

the  external  defences  of  skin  and  mucous  membrane  which,  in  the 
healthy  state,  mechanically  prevent  the  entrance  of  the  micro-organ- 
isms into  the  body.  For  we  have  seen,  in  another  place,  that  few  of 
the  bacteria  can  pass  through  the  uninjured  surfaces.  Moreover, 
added  to  this,  there  is  some  protection  in  the  bactericidal  properties 
of  the  secretions.  An  example  of  this  is  the  inhibitory  power  exer-\ 
cised  by  the  acidity  of  the  normal  gastric  juice  upon  the  choler^r  ' 
spirillum.  In  order  to  infect  the  intestinal  canal  of  guinea  pigs 
with  these  organisms  Koch  found  it  necessary  to  neutralize  the  gas- 
tric juice  with  sodium  carbonate  solutions,  and  other  observers  have 
found  it  necessary  to  inject  directly  into  the  duodenum.  But  even 
after  entrance  into  the  animal  tissues  a  second  line  of  defence  is 
normally  encountered  by  all  invading  germs  which  tend  to  inhibit 
their  further  progress  more  or  less  perfectly.  This  active  opposition 
to  the  bacteria  after  their  entrance  is  expressed  chiefly  in  the  anti-\ 
bacterial  (bactericidal)  activity  of  the  blood  serum,  and  the  pha-/ 
gocytic  powers  of  leukocytes  and  other  cells.  To  a  certain  extent 
these  forces  are  active  against  all  bacteria  in  all  animals,  but  they 
may  vary  in  different  species,  races,  or  even  individuals  in  potency 
against  any  given  infectious  agent,  and,  to  a  certain  extent,  varia- 
tions in  resistance  may  be  referable  to  this.  The  analysis  of  these 
forces,  both  in  the  normal  and  in  the  artificially  immunized  animal, 
forms  the  substance  of  the  systematic  discussions  which  are  to  fol- 
low, and,  for  the  present,  we  will  confine  ourselves  to  an  examination 
of  the  facts  that  have  been  gathered  regarding  the  actual  differences 
in  normal  resistance  or  "Natural  Immunity"  between  various  spe- 
cies of  animals. 

And  if  we  glance  over  the  list  of  diseases  to  which  different  spe- 
cies and  races  of  animals  are  victim,  it  is  immediately  evident  that 
some  animals  are  never  spontaneously  infected  with  many  of  the) 
micro-organisms  that  cause  extensive  and  fatal  ravages  in  others. 
Also,  within  the  same  race  or  species,  an  epidemic  sweeping  through 
a  community  will  kill  many  individuals  and  leave  others  unscathed.) 
Such  differences  point  to  variations  in  the  defensive  mechanism, 
since  the  invader  in  these  cases  is  the  same.  We  speak,  therefore,  of 
Natural  Immunity  which  is  an  attribute  of  species,  that  which, 
within  the  same  species,  is  racial,  and  that  which,  within  the  same 
race,  is  individual.  And  the  attempts  to  discover  the  causes  under- 
lying such  differences  in  natural  resistance  have  elucidated  many  of 
the  fundamental  principles  of  immunity  in  general. 

Instances  of  natural  immunity  which  appear  to  depend  on  spe- 
cies are  common.  We  have  pointed  out,  above,  that  in  order  to  make 
infection  at  all  possible,  it  is  necessary  that  the  invading  germ  shall 
find  suitable  cultural  conditions  in  the  body  of  the  host.  It  is  this 
simple  principle  which  probably  explains  the  fact  that  bacteria  whick 
cause  disease  in  warm-blooded  animals  cannot,  as  a  rule,  cause  di/ 


NATURAL    IMMUNITY  51 

ease  in  those  that  are  cold-blooded,  and  vice  versa.  Thus  frequent 
attempts  to  produce  anthrax  in  turtles,  frogs,  and  other  cold-blooded 
species  have  failed.  Also  among  warm-blooded  animals  differences^ 
in  body  temperature  have  been  shown  to  influence  susceptibility. 
Thus  avian  tuberculosis  does  not  develop  in  mammals,  nor  do  the 
human  and  bovine  types  of  tubercle  bacilli  infect  birds.  And  this  is 
probably  due  to  the  fact  that  the  avian  bacillus  has  become  adapted 
to  growth  at  from  40°  to  45°  C.,  about  the  normal  temperature  of 
birds,  while  the  mammalian  bacilli  cease  to  grow  when  the  tempera- 
ture is  raised  above  40°  C.  Another  observation  which  clearly  illus- 
trates the  influence  of  body  temperature  upon  susceptibility  is  that 
made  by  Gibier  l  upon  anthrax.  Frogs  are  ordinarily  resistant  to 
this  disease.  When  they  are  kept  in  water  at  35°  C.  a  fatal  infec- 
tion can  be  produced.  Nuttall's  2  experiments  with  plague  infection 
in  lizards  illustrate  the  same  point.  Kept  at  16°  C.,  no  infection 
could  take  place.  Warmed  to  26°  C.,  they  could  be  readily  infected. 
It  is  ordinarily  assumed  that  these  results  are  explicable  upon  the 
basis  of  purely  cultural  and  temperature  considerations.  And  this, 
indeed,  is  most  likely.  It  is  possible,  however,  that  an  additional 
factor  involved  in  this  may  be  the  lowering  of  the  general  resistance 
of  cold-blooded  animals  when  warmed,  just  as  warm-blooded  animals 
can  be  rendered  susceptible  by  chilling. 

It  is  for  similar  simple  cultural  reasons,  possibly,  that  diseases 
which  occur  spontaneously  in  carnivora  do  not  occur  in  purely 
herbivorous  animals.  The  relative  resistance  of  dogs  to  anthrax 
and  to  tuberculosis  may  possibly  be  accounted  for  in  this  way. 
However,  there  are  many  micro-organisms  which  infect  easily 
both  carnivorous  and  herbivorous  animals,  and  it  may  well  be  that 
the  frequently  cited  cases  we  have  mentioned  above  depend  on  fac- 
tors more  complicated  than  mere  cultural  conditions  incident  to 
metabolic  differences.  In  most  cases  of  species  resistance,  indeed, 
simple  nutritional  conditions  alone  do  not  serve  as  valid  explana- 
tions. 

Species  resistance  may  be  so  perfect  that  it  amounts  to  an  ab- 
solute immunity.  This  is  apparently  so  in  the  cases  cited  above, 
namely  the  immunity  of  the  cold-blooded  species  to  certain  diseases 
of  warm-blooded  animals.  However,  such  examples  are  exceptional. 
When  we  are  dealing  with  diseases  of  warm-blooded  animals  only, 
natural  resistance,  in  all  but  a  limited  number  of  cases,  is  sufficient 
only  to  prevent  the  spontaneous  occurrence  of  the  particular  disease, 
or  to  prevent  infection  when  experimental  inoculation  with  moderate 
doses  is  practiced  upon  normal  animals.  In  most  of  these,  cases, 
however,  when  the  dose  experimentally  administered  is  excessive,  or 
the  resistance  is  lowered  artificially,  by  chilling  or  by  any  other 

1  Gibier.     Compt.  rend,  de  I'acad.  des  sc.,  Vol.  94,  1882. 

2  Nuttall.     Centralbl.  f.  Bdkt.,  Vol.  22,  1897. 


52  INFECTION    AND    RESISTANCE 

form  of  local  or  general  injury,  infection  can  be  accomplished.  In 
the  case  of  protozoan  diseases  species  adaptation  is  much  more  rigid 
and  parasites  that  infect  one  species  are  very  often  restricted  en- 
tirely to  that  class,  heing  unable  to  infect  any  other  animal,  even 
though  no  striking  difference  in  temperature  or  metabolism  exists. 

We  may  convey  the  clearest  conception  of  all  such  species  differ- 
ences by  a  tabulation  of  some  of  the  more  important  infectious  dis- 
eases of  man  with  a  statement  in  each  case  concerning  its  transmissi- 
bility  to  animals,  as  follows : 

Tuberculosis,  human  type,  spontaneously  infects  man.  It  is  very 
often  observed  in  monkeys  kept  in  captivity.  Cattle,  swine,  and 
sheep  are  probably  never  spontaneously  infected;  guinea  pigs  are 
highly  susceptible  to  experimental  inoculation.  Cattle,  swine,  sheep, 
and  rabbits  are  relatively  very  resistant  to  experimental  infection. 
Dogs  and  goats  are  still  more  so.  Birds  seem  to  be  entirely  refrac- 
tory. 

Tuberculosis,  Bovine  Type. — Spontaneous  infection  occurs  in  do- 
mestic animals,  chiefly  cattle ;  it  is  less  frequent  in  sheep,  hogs,  and 
horses;  it  has  been  reported  in  dogs  and  goats.  In  man  infection 
does  occur,  but  only  a  small  percentage  of  human  tuberculosis  is  of 
the  bovine  type,  and  these  cases  are  almost  exclusively  in  children. 
In  tabulating  1,042  cases  which  have  been  carefully  studied,  Park 
and  Krumwiede  3  report  the  following  figures : 

Cases  of  Tuberculosis  in  Man  (1042) 
Over  16  years 

Human  type  677,  bovine  type  9. 
5  years  to  16  years 

Human  type  99,  bovine  type  33. 
Under  5  years 

Human  type  161,  bovine  type  59. 

The  large  majority  of  bovine  infections  were  abdominal  or  in- 
volved cervical  lymph  nodes. 

Experimental  infection  is  successful  in  rabbits  and  guinea  pigs, 
both  of  these  animals  succumbing  more  rapidly  to  this  than  to  the 
human  bacillus.  In  fact,  the  relative  resistance  of  rabbits  to  the 
human  bacillus  is  such  that  rabbit  inoculation  is  one  of  the  most 
important  methods  of  differentiating  between  the  two  types.  Birds 
are  refractory. 

Tuberculosis  of  the  avian  type  occurs  spontaneously  in  birds.  It 
may  be  experimentally  produced  in  rabbits  (Strauss  and  Gamaleia). 
Injected  into  cattle  it  causes  a  local  reaction  only. 

Tuberculosis  of  cold-blooded  animals  is  not  transferable  to  warm- 
blooded animals. 

Syphilis  spontaneously  occurs  in  man  only.     It  can  be  inoculated 

3  Park  and  Krumwiede.    Jour,  of  Med.  Res.,  Vol.  23,  1910. 


NATURAL    IMMUNITY  53 

chimpanzees,  in  which  primary  and  secondary  lesions  develop, 
corresponding  mildly  to  human  syphilis.  Primary  lesions  can  be 
^produced  in  lower  monkeys.  It  can  be  transferred  by  intratesticular 
inoculations  to  rabbits. 

Gonococcus  infection  occurs  spontaneously  in  man  only.  No 
typical  lesions  can  be  produced  in  experimentally  inoculated  ani- 
mals, though  death  can  be  caused  by  large  doses,  probably  by  toxic 
action. 

Influenza  bacillus  spontaneously  infects  man  only.  Experi- 
mental infection  is  partly  successful  in  monkeys  only.  (Pfeiffer 
and  Beck,  Deut.  med.  Woch.,  1893.) 

Glanders. — Spontaneous  infection  occurs  in  horses  and  mules; 
less  frequently  in  sheep,  goats,  and  camels.  This  disease,  like  plague, 
may  be  regarded  as  primarily  a  disease  of  animals,  but  man  may  be 
infected  by  direct  or  indirect  contact  with  the  diseased  animal.  All 
domestic  animals  may  be  infected  experimentally  with  ease,  except 
cattle  and  rats,  in  which  cases  large  doses  are  necessary.  Birds  show 
local  reactions  only.  (Wladimiroff — in  "Kolle  und  Wassermann 
Handbuch,"  Vol.  5,  2d  Ed.) 

Plague  occurs  spontaneously  chiefly  in  man  and  in  rats.  It  has 
also  been  found  in  California  ground  squirrels  and  in  hogs  during 
plague  epidemics  in  Hong  Kong.  It  is  highly  infectious  for  guinea 
pigs  and  white  rats — slightly  less  so  for  mice ;  rabbits  are  much  less 
susceptible  than  guinea  pigs.  Dogs,  cats,  and  cattle  are  relatively 
resistant.  Birds  appear  to  be  immune.  Cold-blooded  animals  are 
immune  unless  artificially  warmed.  (See  above.) 

Malta  fever  occurs  spontaneously  in  man  and  in  goats.  It  is 
pathogenic  for  all  mammals,  but  it  is  not  fatal  for  lower  animals 
when  the  organisms  are  directly  cultivated  out  of  the  human  body. 

Diphtheria  occurs  spontaneously  in  man  only.  Experimental  in- 
oculation is  fatal  in  guinea  pigs,  rabbits,  dogs,  cats,  and  birds.  Rats 
and  mice  are  highly  resistant.  The  typical  pseudomembranous  in- 
flammation can  be  produced  in  susceptible  animals  only  after  pre- 
vious injury  of  the  mucous  membrane,  and  then  it  rarely  shows  any 
tendency  to  spread. 

Tetanus  is  spontaneous  in  man,  horses,  cattle,  and  sheep.  It  ia 
found  rarely  in  dogs  and  goats.  Birds  are  highly  resistant  to  ex- 
perimental inoculation. 

Anthrax  is  primarily  a  spontaneous  infection  of  cattle,  sheep, 
and  horses ;  it  occurs  in  man  largely  through  direct  or  indirect  contact 
with  these  animals.  Guinea  pigs,  rabbits,  and  white  mice  are  very 
susceptible  to  experimental  inoculation.  Rats  and  hogs  are  less  sus- 
ceptible, and  dogs  are  relatively  resistant,  though  they  can  be  regu- 
larly killed  by  moderate  doses  intravenously  injected.  Birds  and 
cold-blooded  animals  are  highly  resistant. 

Asiatic  cholera  develops  spontaneously  in  man  only.       Rabbits 


54  INFECTION    AND    RESISTANCE 

and  guinea  pigs  can  be  killed  by  injections  of  cultures,  but  die  prob- 
ably of  toxemia.  In  rabbits  a  cholera-like  condition  has  been  pro- 
duced by  injection  of  the  spirilla  into  the  duodenum  after  ligation 
of  the  common  bile  duct.  (Nikati  and  Rietsch,  Ref.  in  Deut.  med. 
Woch.;  Vol.  II,  1884,  p.  613.)  Ordinarily  no  multiplication  takes 
place  in  the  animal  body.  Pigeons  are  insusceptible,  a  fact  which 
helps  to  distinguish  this  organism  from  Spirillum  metchnikovi  and 
other  similar  bird-pathogenic  spirilla. 

Typhoid  fever  occurs  spontaneously  in  man  only.  It  has  recently 
been  produced  in  a  mild  form  in  chimpanzees.  Animals  are  suscep- 
tible to  the  endotoxins  and  can  therefore  be  killed  by  injections  of 
bacilli  and  extracts,  but  the  organism  is  not  invasive  as  in  the  case  of 
the  lower  animals.  Typhoid  septicemia  can  be  produced  in  rabbits 
by  inoculating  them  with  especially  virulent  cultures  of  the  bacilli, 
or  cultures  previously  grown  on  rabbit-blood  agar  (Gay).  The  ty- 
phoid-carrier state  may  ensue  for  considerable  periods  in  such  ani- 
mals. 

Pneumococcus  infection  in  various  forms  occurs  spontaneously 
in  man.  Rabbits,  mice,  and  guinea  pigs  are  highly  susceptible. 
Rats,  dogs,  cats,  cattle,  and  sheep  are  relatively  resistant. 

Staphylococcus  and  streptococcus  infections  may  occur  in  almost 
all  of  the  warm-blooded  animals,  chiefly  as  abscess  producers.  In 
horses  a  severe  form  of  pleuropneumonia  is  caused  by  them. 

Leprosy  occurs  spontaneously  in  man  only.  Lesions  simulating 
human  leprosy  have  been  produced  in  monkeys  by  inoculation,  and 
partially  successful  experiments  have  been  made  upon  the  Japanese 
dancing  mouse.  Other  animals  are  immune. 

Scarlet  fever  occurs  spontaneously  in  man  only.  Monkeys  may 
possibly  be  susceptible,  though  not  all  observers  have  been  successful 
in  such  experiments.  (Draper  and  Handford,  Journ.  of  Exp.  Med., 
Vol.  17,  1913.)  Landsteiner  and  Levaditi  (Ann.  Past.,  Vol.  25, 
1911)  have  succeeded  in  producing  the  disease  in  the  chimpanzee, 
though  they  failed  with  lower  monkeys. 

Small-pox  occurs  spontaneously  in  man  only.  It  is  probably  iden- 
tical with  cow-pox.  (See  reasons  for  this  assumption  given  by  Ha- 
cius  as  cited  by  Paul  in  "Kraus  and  Levaditi  Handbuch,"  etc.,  Vol. 
1.)  It  can  be  experimentally  produced  in  monkeys. 

•  Measles  develops  spontaneously  only  in  man.  Macacus  rhesus 
ias  been  successfully  inoculated  by  Anderson  and  Goldberger  (U.  S. 
Pub.  Health  Reports,  26,  1911).  Other  animals  are  immune. 

Typhus  fever  occurs  in  man  only.  Experimentally  it  has  been 
produced  in  chimpanzees,  Macacus,,  Cercopithecus,  Ateles,  and  My- 
cetes  monkeys.  Anderson  has  succeeded  in  producing  temperature 
reactions  in  guinea  pigs  by  injecting  blood  from  typhus  patients  or 
from  other  similarly  infected  guinea  pigs.  More  exact  information 
concerning  this  disease  will  probably  be  available  soon,  if  the  re- 


NATURAL    IMMUNITY  55 

ported  cultivation  of  the  organism  of  the  disease  by  Plotz  is  authen- 
ticated. 

Yellow  fever  up  to  the  present  has  been  observed  in  man  only. 

Poliomyelitis  is  spontaneous  in  man  only.  Can  be  transmitted  to 
monkeys  and — in  a  doubtful  form — to  rabbits.  No  other  animals 
are  known  to  be  susceptible. 

The  above  represents  an  incomplete  tabulation  of  the  variations 
in  susceptibility  in  the  animal  kingdom  for  infections  which  occur 
spontaneously  in  man.  They  will  illustrate  sufficiently,  however, 
the  facts  of  variable  species  susceptibility  as  we  have  stated  them. 
We  might,  with  equal  profit,  tabulate  the  infections  occurring  spon- 
taneously in  any  single  species  of  animal  and  show  how  variable 
would  be  their  pathogenic  powers  for  other  animals  and  for  man. 
Thus  man  is  immune  to  the  organism  which  causes  cattle  plague, 
and  to  that  of  chicken  cholera,  and  probably  to  many  other  diseases 
peculiar  to  animals,  though,  of  course,  in  the  case  of  infections  of 
the  human  being  we  are  entirely  dependent  for  such  information 
upon  observed  immunity  to  spontaneous  infection,  and  upon  a  few 
instances  of  accidental  inoculation. 

In  regard,  also,  to  differences  of  susceptibility  between  various 
races,  within  the  same  species,  many  interesting  facts  have  been  ob- 
served. Thus  gray  mice  are,  as  a  rule,  more  resistant  to  strepto- 
coccus and  pneumococcus  infection  than  are  wnite  mice.  Algerian 
sheep  are  said  to  be  more  resistant  to  anthrax  than  are  European 
sheep.  Of  black  rats  inoculated  by  Miiller  4  with  anthrax  over  79 
per  cent,  survived,  while  of  white  rats  similarly  inoculated  only  14 
per  cent,  survived. 

In  man,  too,  racial  differences  are  marked.  The  extraordinary 
susceptibility  of  the  negro  to  tuberculosis  is  familiar  to  all  American 
physicians,  and  it  is  well  known  that  Eskimos  transported  to  tem- 
perate climates  and  civilized  conditions  are  particularly  prone  to 
contract  this  disease.  Small-pox  is  considered  a  relatively  mild 
disease  in  Mexico.  Dr.  James  Carroll  5  stated  that  whites  are  more 
susceptible  to  yellow  fever  than  are  negroes,  and  that  among  the 
latter  those  living  nearest  the  equator  are  less  susceptible  than  are 
the  more  northern  races.  There  seems  to  be  no  doubt  about  the 
actual  occurrence  of  such  racial  differences,  although,  as  Hahn 6 
very  justly  points  out,  many  instances  formerly  regarded  as  racial 
differences  of  susceptibility  may  have  been  simulated  by  racial, 
or  often  religious,  differences  of  custom  that  influence  sanitary  con- 
ditions, and  consequently  the  incidence  of  epidemic  disease. 

Apart  from  the  explanations  furnished  in  a  few  instances  by 

4  Miiller.     Fortschr.  der  Med.,  1893.     Cited  from  Sobernheim,  in  "Kolle 
u.  Wassermann  Handbuch,"  2d  Ed.,  Vol.  3. 

5  Carroll  in  "Mense,  Tropenkrankheiten,"  Vol.  2,  p.  124. 

6  Hahn  in  "Kolle  und  Wassermann's  Handbuch,"  Vol.  1. 


56  INFECTION    AND    RESISTANCE 

gross  physiological  differences  such  as  body  temperature,  the  factors 
determining  species  resistance  are  largely  a  mystery,  and  in  the 
matter  of  racial  variations,  of  course,  we  have  no  instances  in  which 
such  very  obvious  physiological  factors  play  a  part.  In  attempting 
to  find  causes  for  differences  of  resistance  or  susceptibility  in  gen- 
eral, the  nature  of  the  problem  makes  it  necessary  for  us  to  examine 
it  from  a  number  of  different  points  of  view.  A  micro-organism 
may  be  infectious  for  a  given  species  of  animal  more  than  for 
another,  because  of  special  adaptation  to  the  conditions,  nutritive 
and  otherwise,  encountered  in  the  tissues  of  these  animals.  Such 
adaptation  is  illustrated  in  the  experience  of  Pasteur  with  "rouget" 
and  with  rabies,  where  passage  through  one  variety  of  animal  en- 
hanced the  virulence  for  this  species  but  reduced  it  for  others;  and 
the  same  thing  is  easily  demonstrated  in  the  laboratory  with  so  many 
bacteria  that  it  may  be  accepted  as  a  principle  underlying  enhance- 
ments of  virulence  in  general.  This  adaptation  implies  that,  to  a 
certain  extent,  the  part  played  by  the  animal  body  in  determining 
its  own  susceptibility  is  passive.  Gonococcus,  for  instance,  infec- 
tious for  man  only,  requires  human  protein  for  growth,  at  least  in 
its  first  generations  outside  the  body.  Its  ability  to  cause  disease 
in  man  may  be  largely  dependent  upon  its  cultural  need  of  human 
protein.  The  resistance  of  other  animals  to  this  disease,  then,  is,  in 
part,  due  to  their  failure  to  supply  proper  nutriment.  This,  as  Kolle 
points  out,  is  analogous  to  Atrepsie,  a  term  used  by  Ehrlich,  in 
speaking  of  the  insusceptibility  of  one  species  to  cancerous  growths- 
originating  in  another. 

Again,  "adaptation"  on  the  part  of  the  bacteria  may  imply,  not 
only  an  increased  ability  to  meet  altered  cultural  conditions,  but  an 
actual  acquisition  of  greater  offensive  or  invasive  powers  with  which 
to  meet  the  particular  defences  opposed  to  it  by  the  given  animal. 
Thus  the  increased  virulence  of  typhoid  bacilli  after  cultivation  in 
immune  sera  would  point  toward  an  increased  ability  to  survive 
under  the  adverse  conditions  encountered  in  the  animal  body.  An 
organism  may  possibly  acquire  particular  infectiousness  for  one 
species  to  the  exclusion  of  others,  by  a  succession  of  spontaneous 
inoculations — comparable  to  the  experimental  passage  of  the  micro- 
organism through  animals  of  the  same  species.  This  is  especially 
probable  in  diseases  such  as  gonorrhea,  syphilis,  and  some  others 
where  infection  is  usually  direct  from  one  person  to  another.  And 
it  is  these  diseases  particularly  in  which  infectiousness  is  rather 
strictly  limited  to  the  human  species. 

Regarding  the  matter  purely  from  the  point  of  view  of  the  ani- 
mal body  and  the  factors  which  determine  its  powers  to  ward  off  a 
given  infection,  we  may  justly  assume  that  natural  resistance  may 
be  largely  a  matter  of  inheritance.  Whether  this  is  to  be  interpreted 
as  purely  an  instance  of  survival  of  the  fittest  or  whether  immunity 


NATURAL    IMMUNITY  57 

acquired  by  an  individual  can  be  wholly  or  in  part  transmitted  to 
the  offspring  is  an  open  question — at  present  in  the  same  state  of 
unclearness  as  are  other  questions  relating  to  the  transmissibility  of 
acquired  characteristics.  However  this  may  be,  there  are  a  number 
of  facts  available  which  indicate  that  inheritance  plays  an  important 
part.  It  is  apparent  in  the  case  of  many  diseases  afflicting  human 
beings  that  infection  takes  a  milder  course  in  those  races  among 
which  it  has  long  been  endemic — whereas  the  same  disease,  suddenly 
introduced  among  a  new  people,  is  relatively  more  severe  and  spreads 
more  rapidly.  This  seems  to  be  the  case  with  yellow  fever  and  tuber- 
culosis, and  in  measles  and  small-pox,  too,  the  principle  seems 
to  hold  good.  Syphilis  when  first  described  authentically — as  epi- 
demically sweeping  through  Europe  toward  the  close  of  the  15th 
century — appears  to  have  been  a  far  more  acute  and  violent  disease 
than  it  is  among  us  to-day.  It  may  well  be  that  this  depends  upon 
a  gradual  elimination  (elimination  in  this  case,  especially  as  far  as 
reproduction  is  concerned)  of  those  individuals  that  are  fortuitously 
more  susceptible  and,  by  natural  selection,  a  higher  racial  resistance 
is  gradually  developed.  Whether  or  not  direct  inheritance  of  the 
individually  acquired  immunity  can  be  considered  at  all  as  a  con- 
tributing factor  is  difficult  to  decide.  That  immunity  can  be  trans- 
mitted from  mother  to  offspring  was  observed  by  Chauveau 7  as 
early  as  1888.  Lambs  thrown  by  anthrax-immune  ewes  possessed  a 
higher  resistance  against  this  infection  than  did  the  lambs  of  normal 
ewes.  The  extensive  experiments  of  Ehrlich,8  carried  out  chiefly 
upon  mice  with  the  vegetable  poisons  ricin  and  abrin,  showed  that  in 
these  cases  immunity  may  be  transmitted  from  mother  to  offspring, 
but  depends  upon  a  passive  transfer  of  the  specific  antitoxins  both 
by  the  blood  and  the  milk  of  the  mother.  The  sperm  of  the  father 
did  not  seem  to  have  anything  to  do  with  inherited  resistance,  since 
no  immunity  followed  in  the  offspring  when  immunized  males  were 
paired  with  normal  females.  From  the  complete  absence  of  im- 
munity in  the  second  generation  (grandchildren)  of  the  immunized 
female,  and  from  the  short  duration  (2  to  3  months)  of  its  per- 
sistence, he  concluded  that  the  ovum  itself  had  no  influence,  but  that 
the  entire  phenomenon  was  attributable  to  a  passive  transference  of 
antitoxins  from  mother  to  child  during  gestation  and  lactation.  He 
interpreted,  in  the  same  sense,  Chauveau's  anthrax  experiments,  and 
similar  experiments  of  Thomas9  and  Kitasato  10  with  symptomatic 
anthrax,  suggesting  that,  here  also,  a  transept  of  antibodies  -from 
mother  to  offspring  had  taken  place.  The  experiments  of  Ehrlich 
permit  of  no  doubt  as  to  the  validity  of  his  conclusions.  However, 

7  Chauveau.     Ann.  Pasteur,  1888. 

8  Ehrlich.     Zeitschr.  f.  Hyg.,  1892,  Vol.  12. 

9  Thomas.    Compt.  rend,  de  I'acad.  des  sc.,  Vol.  94,  cited  by  Ehrlich,  loc.  cit. 

10  Kitasato.    Cited  by  Ehrlich,  loc.  cit. 


58  INFECTION    AND    RESISTANCE 

we  must  remember  that  they  were  carried  out  with  antitoxic  im- 
munity only,  in  which  the  resistance  is  purely  dependent  upon  the 
circulating  antibody  and  is  never,  even  in  actively  immunized  indi- 
viduals, a  permanent  state.  In  immunity  such  as  that  acquired 
against  typhoid  fever,  plague,  cholera,  and  other  diseases  after  re- 
covery from  an  attack,  the  individual  remains  relatively  resistant 
long  after  the  demonstrable  antibodies  have  disappeared  from 
the  circulation,  and  we  must  assume  that  this  permanent  re- 
sistance depends  upon  a  physiological  alteration — inexplicable  for 
the  present,  but  surely  residing  in  the  body  cells.  In  such  cases 
it  is  by  no  means  certain  that  there  may  not  be  a  very  slight,  but 
through  generations  gradually  accumulating,  inheritance  of  im- 
munity. At  any  rate  the  experiments  of  Ehrlich  do  not  disprove 
such  a  possibility.  Moreover,  in  this  connection  it  must  not 
be  forgotten  that  natural  immunity,  unlike  acquired  immunity, 
cannot  be  passively  transferred  from  one'  animal  to  another,  and 
implies  therefore  a  fundamental  cellular  difference  rather  than 
a  condition  depending  merely  upon  antibodies  circulating  in  the 
l)lood. 

For  this  last  reason  also  it  has  been  unsatisfactory  to  attempt 
explanations  of  natural  immunity  purely  upon  grounds  of  bacteri- 
cidal and  other  properties  of  the  blood  serum.  These  points  we  will 
take  up  at  greater  length  when  we  discuss  the  mechanism  of  resist- 
ance in  general. 

An  important  observation  upon  the  inheritance  of  serum  prop- 
erties is  that  which  has  been  made  by  Ottenberg  and  Epstein  n  in 
connection  with  the  iso-agglutinins.  We  shall  see  in  another  section 
that  the  blood  serum  of  one  human  being  will  often  possess  the 
property  of  agglutinating  the  human  blood  cells  of  another  indi- 
vidual. These  iso-agglutinating  constituents  of  the  serum  are  ap- 
parently transmitted  from  parents  to  offspring.  Yon  Dungern  and 
Hirschfeld,12  in  studying  these  iso-agglutinins  in  72  families,  upon 
348  people,  not  only  confirmed  the  observations  of  the  preceding 
workers,  but  showed  that  such  inheritance  follows  Mendelian  laws. 
Not  only  is  this  of  great  biological  interest,  but  it  is  of  great  im- 
portance in  connection  with  our  present  discussion  in  showing  that 
such  properties  as  agglutinating  powers  of  serum  can  be  influenced 
by  inheritance  from  the  father  as  well  as  from  the  mother. 

The  individual  differences  in  resistance  which  unquestionably 
exist  among  members  of  the  same  species  and  races  are  very  difficult 
to  explain,  but,  as  far  as  we  can  tell  anything  about  them  at  all,  they 
seem  to  depend  upon  variation  in  what  is  popularly  spoken  of  as 
"general  condition."  The  laboratory  animals  with  which  most  ex- 
perimentation is  done,  rabbits  and  guinea  pigs,  if  healthy,  show  very 

11  Ottenberg  and  Epstein.     Proceedings  of  the  N.   Y.  Path.  Soc.,  1908. 

12  Von  Dungern  and  Hirschfeld.    Zeitschr.  f.  Immunitats.,  Vol.  4,  1910. 


NATURAL    IMMUNITY  59 

slight  individual  variations.  In  fact,  the  astonishing  uniformity  of 
reaction  on  the  part  of  guinea  pigs  of  similar  age  and  weight  against 
measured  quantities  of  bacterial  toxins  has  alone  made  it  possible 
to  utilize  these  animals  in  the  standardization  of  antitoxins.  Pneu- 
mococcus  and  streptococcus  cultures  can  be  measured  with  reason- 
able accuracy  upon  white  mice  of  approximately  uniform  weight, 
and  the  same  animals  are  relatively  uniform  in  their  reactions  to 
identical  amounts  of  tetanus  poison.  Many  other  examples  might 
be  cited  which  make  it  clear  that  healthy  animals  of  the  same  species, 
kept  under  the  same  conditions,  fed  upon  the  same  food,  and  of  ap- 
proximately the  same  age  and  weight,  differ  but  slightly  from  each 
other  in  reaction  to  the  same  infectious  agent.  This  would  indicate 
that  the  individual  differences  in  resistance  displayed  so  plainly  by 
human  beings  are  due,  not  to  any  fundamental  individual  variations, 
but  rather  to  such  fortuitous  factors  as  nutrition,  metabolic  fluctua- 
tions, temporary  physical  depression,  fatigue,  or  chilling.  A  person 
suffering  from  functional  impairment  of  any  kind  is  more  likely 
to  permit  the  invasion  of  a  pathogenic  micro-organism  than  is  a  per- 
fectly healthy  well-nourished  individual  of  the  same  species. 

Most  of  these  facts  we  know  from  the  accumulated  experience  of 
clinicians  who  also  have  given  us  much  valuable  information  con- 
cerning the  susceptibility  to  infection  on  the  part  of  chronically  dis- 
eased persons,  especially  diabetics  and  nephritics.  In  the  case  of  a 
few  of  these  influences,  chilling  and  fatigue,  experimental  data  on 
animals  are  available.  It  is,  however,  extremely  difficult  to  analyze 
the  causes  underlying  such  depression  of  resistance.  For  instance, 
with  fatigue  or  chilling  there  may  be  temporary  congestion  of  mucous 
surfaces,  due  to  vasomotor  influences,  which  alter  the  secretions  on 
mucous  surfaces,  or  interfere  with  the  normal  mobilization  of 
leukocytes,  permitting  penetration  of  bacteria  where  ordinarily 
they  would  have  been  held  back.  Our  ignorance  is  nowhere  more 
clearly  illustrated  than  in  the  fact  that  we  know  practically  nothing 
concerning  the  relation  between  a  thorough  chilling  and  the  acquisi- 
tion of  what  is  spoken  of  as  a  common  acold."  We  can  only  assume 
that  there  is  interference  in  some  way  with  the  normal  bactericidal 
and  phagocytic  mechanisms,,  making  possible  the  penetration  and 
lodgment  of  small  quantities  of  bacteria,  ordinarily  destroyed  imme- 
diately after  entrance  or  prevented  from  entering  at  all. 

Of  course  we  must  except  those  individual  differences  of  sus- 
ceptibility which  may  be  dependent  upon  inheritance.  We  know, 
for  instance,  that  in  such  diseases  as  diphtheria,  where  resistance 
depends  upon  antitoxins  circulating  in  the  blood,  there  may  be  a 
passive  immunity,  conferred  from  mother  to  offspring,  which  lasts 
for  several  weeks  or  months  after  birth.  It  is  important  to  remem- 
ber such  a  possibility  in  the  selection  of  guinea  pigs  for  diphtheria 
antitoxin  standardization,  as  Anderson  has  pointed  out.  Whether 


60  INFECTION    AND    RESISTANCE 

or  not  a  tendency  to  tuberculosis  can  be  inherited  is  still  an  open 
question.  In  most  cases  it  is  more  than  probable  that  the  supposedly 
inherited  tendency  to  tuberculosis  is  not  really  an  inherited  sus- 
ceptibility, but  rather  an  actual  infection  acquired  during  childhood 
from  the  parents.  Cornet  and  Kossel,13  who  have  recently  sum- 
marized the  statistics  dealing  with  this  problem,  have  come  to  the 
conclusion  that  this  factor,  namely,  infection  from  the  parents, 
probably  is  the  cause  of  the  greater  frequency  of  tuberculosis  among 
children  of  tuberculous  parents,  and  that  there  is  no  definite  proof 
of  inherited  susceptibility. 


ACQUIRED  IMMUNITY  AND  IMMUNIZATION 

We  have  outlined  in  the  preceding  pages  the  differences  in  sus- 
ceptibility to  various  diseases  apparent  among  different  species  of 
animals,  and  have  noted  that  the  degree  of  resistance  of  some  animals 
to  infection  with  germs  rapidly  fatal  to  others  is  often  sufficiently 
well-marked  to  be  termed  "immunity."  Such  immunity,  because  it 
is  a  natural  biological  attribute  of  the  species,  as  much  a  character- 
istic property  as  are  its  anatomical  or  physiological  properties,  has 
been  spoken  of  as  "Natural  Immunity." 

It  is  a  matter  of  common  knowledge,  however,  that  among 
species  of  animals  readily  susceptible  to  certain  infections  resistance, 
or  even  extreme  resistance,  i.  e.,  immunity,  may  be  acquired  by  an 
attack  of  the  disease.  Thus  human  beings  who  have  recovered  from 
plague,  small-pox,  typhoid  fever,  cholera,  the  exanthemata,  mumps, 
typhus,  yellow  fever,  and  a  number  of  other  conditions  do  not  ordi- 
narily contract  the  disease  a  second  time.  In  some  of  these  condi- 
tions, notably  cholera,  plague,  typhoid  fever,  and  small-pox,  the  rule 
is  almost  invariable.  In  others,  such  as  measles,  scarlet  fever,  and 
mumps,  a  second  attack  may  occur,  though  it  is  rare. 

The  following  table  briefly  indicates  infectious  diseases  in  which 
permanent  immunity  follows  an  attack : 

Infectious  Diseases  in  Which  One  Attack  Conveys  Lasting  Immunity 

Plague. 

Typhoid — second  attack  rare — about  2.4  per  cent.  (Curschmann). 

Cholera. 

Small-pox — second  attack  very  rare. 

Chicken-pox — second  attack  very  rare. 

Scarlet  fever — second  attack  about  0.7  per  cent. 

Measles — second  attack  uncommon,  but  less  rare  than  scarlatina. 

Yellow  fever. 

Typhus  fever. 

Syphilis — reinfection  rare,  though  more  common  than  formerly  supposed. 

Mumps-second  attack  rare  (Kraus). 

Poliomyelitis. 

13  Cornet  and  Kossel  in  "Kolle  u.  Wassermann,"  Vol.  5,  2d  Ed. 


ACQUIRED  IMMUNITY  61 

No  lasting  immunity  is  conferred  by  one  attack  in: 

Infection  with  the  Pyogenic  cocci 

Gonorrhea 

Pneumonia 

Influenza 

Glanders 

Dengue  fever 

Diphtheria  in  general  protection,  second  attack  In  0.9 
per  cent,  cases-0.01  antitoxin  unit  per  c.  c.  of  circu- 
lating blood  protects. 

Recurrent  fever 

Tetanus 

Erysipelas 

Beri  beri 

Tuberculosis 

There  is  another  group  of  diseases  in  which  the  immunological 
conditions  after  infection  are  as  yet  not  clear — namely,  protozoan 
infections  like  malaria  and  trypanosomiasis,  and  treponema  diseases 
like  syphilis.  In  these  conditions  reinfection  seems  to  be  impossible 
only  so  long  as  the  individual  still  harbors  the  microorganism,  but  no 
lasting  immunity  is  conferred.  We  have  discussed  these  conditions 
in  extenso  in  the  section  on  syphilis  immunity,  (see  p.  508). 

These  observations  actually  form  the  point  of  departure  of  that 
entire  branch  of  medical  science  which  devotes  itself  to  the  study  of 
resistance  to  infection,  serum  diagnosis,  and  specific  therapy,  and 
it  will  be  seen  that  all  the  facts  that  have  been  gathered  upon  these 
subjects  are  the  fruits  of  detailed  analysis  of  this  phenomenon  of 
acquired  immunity. 

Its  occurrence  in  many  instances  nas  been  so  striking  that  ancient 
observers,  long  before  the  birth  of  rational  medicine,  referred  to  it, 
and  often  drew  from  it  conclusions  of  great  hygienic  importance. 
Thucydides,  in  the  second  book  of  his  account  of  the  Peloponnesian 
Wars,  in  describing  the  plague  at  Athens,  notes  the  apparent  safety 
from  reinfection  of  those  who  had  recovered,  suggesting  the  possibil- 
ity of  their  being  therefrom  immune  against  disease  in  general.  The 
literature  of  the  Middle  Ages  and  of  earlier  modern  times  contains 
numerous  further  references  which  indicate  that  acquired  resistance 
was  clinically  recognized  as  a  result  of  recovery  from  many  diseases. 
The  phenomenon  was  not  only  observed,  but  put  to  practical  utiliza- 
tion by  the  ancients  of  China  and  India.  Thus  the  practice  of  inocu- 
lating children  with  small-pox  material  from  the  active  pustules  of 
patients,  or  making  them  sleep  in  beds  or  wear  the  shirts  of  sufferers 
was  a  dangerous  practice  but  logical,  on  the  reasoning  that  the  disease 
conveyed  to  a  person  in  full  health  and  good  condition  would  probably 
take  a  mild  course  and  confer  immunity,  while  the  naturally  acquired 
disease,  contracted  often  because  of  the  weak  and  debilitated  condi- 
tion of  the  individual,  would  be  more  apt  to  end  fatally. 


62  INFECTION  AND  RESISTANCE 

Such  methods,  though  barbaric  and  eventually  unjustified  by  the 
naturally  high  mortality  incident  upon  them,  were  actually  brought 
to  Europe  from  the  East,  and  for  a  time  practiced  in  European 
countries. 

The  first  great  advance  which  bridged  the  gap  between  the  obser- 
vations regarding  naturally  acquired  immunity  and  rational  experi- 
mental immunization  was  made  by  Edward  Jenner.  It  has  been  no- 
ticed before  Jenner  began  his  w.ork  that  milkmaids  and  others  who 
had  contracted  cow-pox  in  the  course  of  their  occupations  were  usually 
spared  when  a  small-pox  epidemic  occurred  in  their  community.  Spo- 
radic attempts  had  been  made  to  put  this  observation  to  practical  use, 
but  no  one  with  sufficient  intelligence,  persistence,  and  training  had 
taken  up  the  matter  seriously.  Jenner,  interested  by  the  reports  of  this 
nature  and  by  his  own  observations,  was  especially  impressed  by  the 
similarity  between  the  local  manifestations  of  small-pox,  cow-pox,  and 
a  disease  of  horses  spoken  of  as  "grease."  Though  at  first  disinclined 
to  identify  small-pox  with  cox-pox  (at  present  the  prevailing  opinion 
is  that  the  second  is  an  attenuated  form  of  the  former),  Jenner 
thoroughly  investigated  cases  of  alleged  protection  by  cow-pox,  a  claim 
which  before  this  had  been  hardly  more  than  a  rumor,  and  finally, 
with  the  encouragement  of  John  Hunter,  proceeded  to  the  vaccina- 
tion of  human  beings  with  cow-pox,  testing  the  result  by  subsequent 
inoculation  of  the  same  individual  with  small-pox.  His  report  to  the 
Royal  Society  in  1796  and  his  subsequent  publications  incorporate 
the  results  of  these  experiments  by  means  of  which  the  practice 
of  vaccination  against  small-pox  was  introduced  and  the  virtual 
eradication  of  the  disease  from  civilized  communities  was  attained. 

The  principles  underlying  small-pox  vaccination  are  extremely 
simple.  The  attenuated  virus  after  inoculation  incites  a  mild  and  lo- 
calized form  of  the  disease,  from  which  the  subject  recovers  rapidly 
and  completely.  The  recovery  implies  the  mobilization  of  certain  pro- 
tective forces  and  a  specific  physiological  alteration  of  the  body  in 
such  a  way  that  a  permanently,  or  at  least  prolongedly,  increased  re- 
sistance against  the  disease  remains.  In  consequence,  if  the  indi- 
vidual is  subsequently  exposed  to  spontaneous  infection  with  this  dis- 
ease, his  acquired  specific  resistance  suffices  to  prevent  invasion  by 
the  virus.  This  is  merely  an  artificial  imitation  of  the  conditions 
which  obtain  when  an  individual  recovers  from  an  attack  of  a  disease 
and  is  rendered  immune  thereby.  In  this  case,  however,  the  attenua- 
tion of  the  virus  has  eliminated  the  dangers  attendant  upon  an  actual 
attack.  The  immunity  thus  conferred  is  probably  never  as  perfect  nor 
as  lasting  as  that  following  a  seizure  of  the  disease  in  its  unattenuated 
form;  however,  it  suffices,  as  a  rule,  to  prevent  spontaneous  infec- 
tion which  is  never  as  severe  a  test  as  experimental  inoculation. 

In  contrast  to  the  "Natural  Immunity"  which  is  an  inherited  at- 
tribute of  race  or  species,  we  speak  of  such  increased  resistance  in  a 


ACQUIRED  IMMUNITY  63 

member  of  an  originally  susceptible  race  as  "Acquired  Immunity." 
When  the  immunity  has  been  attained  as  the  result  of  an  attack  of  the 
disease  itself  it  is  spoken  of  as  "Naturally  or  Spontaneously  Acquired 
Immunity"  When  produced  by  some  form  of  treatment  with  the 
virus  of  the  disease,  altered  in  such  a  way  that  an  actual  attack  is 
averted,  we  speak  of  it  as  "Artificially  Acquired  Immunity." 

The  premises  of  Jenner's  reasoning  were  valid  as  his  experiments 
were  convincing.  But  knowledge  regarding  infectious  disease  and 
its  causation  by  living  germs  was  not  developed  until  almost  one 
hundred  years  later,  by  the  work  chiefly  of  Pasteur.  For  this  reason 
no  direct  continuation  of  Jenner's  work  appeared  until  Pasteur 
made  his  communication  upon  Chicken  Cholera  to  the  Parisian 
Academy  of  Medicine  in  1880.  Though  his  investigations  differed 
entirely  from  those  of  Jenner  both  in  method  and  the  nature  of  the 
disease  with  which  they  dealt,  Pasteur  recognized  the  similarity  of 
the  fundamental  principles  underlying  both  discoveries. 

His  observations  took  origin  in  a  purely  accidental  occurrence. 
Cultures  of  chicken  cholera  which  had  been  allowed  to  stand  without 
transplantation  and  under  aerobic  conditions  for  periods  of  several 
months  were  found  to  have  diminished  in  virulence.  Inoculated  into 
chickens,  they  failed  to  kill,  giving  rise  in  many  cases  to  localized 
lesions  only.  It  occurred  to  Pasteur  that  inoculation  with  such  an 
attenuated  culture  might  protect  against  subsequent  infection  with 
fully  virulent  strains  and,  indeed,  experimental  investigation  of  this 
idea  proved  to  be  correct.  He  developed  a  method  of  "vaccination" 
against  chicken  cholera  which  consisted  in  injecting  -first  a  very 
much  attenuated  culture  of  the  organism  (premier  vaccin),  and, 
after  12  or  14  days,  another  less  perfectly  attenuated  (deuxieme 
vaccin),  since  he  observed  that  a  single  inoculation  was  often  in- 
sufficient to  confer  protection.  After  two  inoculations  a  degree  of 
immunity  could  be  attained  which  sufficed  to  protect  against  spon- 
taneous infection  as  well  as  against  experimental  inoculation  with 
doses  of  the  virulent  germs,  fatal  for  untreated  animals. 

These  experiments,  simple  as  they  are,  constitute  the  beginnings 
of  the  science  of  Immunity,  since,  for  the  first  time,  an  investigator 
working  with  a  pure  culture  of  a  pathogenic  microorganism  had 
succeeded,  in  planned  and  purposeful  experiments,  in  conferring 
artificial  immunity.  The  path  was  now  clearly  indicated  and  the 
years  immediately  following  were  fruitful  in  the  development  of 
many  methods  by  which  pathogenic  bacteria  may  be  attenuated  and 
changed  in  such  a  way  that  they  can  be  used  to  confer  immunity 
without  causing  more  than  a  transient  and  harmless  reaction  in  the 
subject.  Most  of  the  earlier  discoveries  of  this  kind  came  from 
Pasteur  himself  and  from  members  of  his  school. 

Since  in  all  these  methods  the  inoculated  animal  attains  its  in- 
creased resistance  by  reason  of  the  activities  of  its  own  tissues,  these 


64  INFECTION  AND  RESISTANCE 

processes  are  spoken  of  as  "Active  Immunization/'  ~No  protective 
factor  is  conferred  directly.  The  disease  itself  is  inoculated,  though 
in  an  altered  form,  and  the  subsequent  immunity  is  purely  the  result 
of  the  physiological  reaction  occurring  as  the  subject  struggles  against 
and  overcomes  the  injected  virus,  bacteria,  or  their  products.  Such 
"Active  Immunization,"  we  shall  see,  is  in  contrast  to  "Passive 
Immunization,"  a  procedure  in  which  the  serum  of  an  actively  im- 
munized animal  is  injected  into  another,  carrying  with  it  certain 
substances  by  which  protection  is  conferred.  The  recipient  here  is 
passively  protected  by  products  of  the  active  reaction  which  has 
taken  place  in  the  body  of  the  donor. 

After  his  success  in  active  immunization  against  chicken  cholera 
Pasteur  applied  the  principles  here  learned  to  experiments  upon  the 
protection  of  animals  against  anthrax.  This  problem  was  fraught 
with  considerable  difficulty  because  of  the  great  virulence  of  the 
anthrax  bacillus.  However,  successful  attenuation  was  attained  by 
a  method  which  depended  upon  the  cultivation  of  anthrax  cultures  at 
temperatures  above  the  optimum  for  its  growth.  Toussaint 14  had 
shown  that  the  resistance  of  sheep  could  be  increased  if  they  were 
inoculated  with  blood  from  animals  dead  of  anthrax  after  this  had 
been  heated  to  55°  C.  for  10  minutes.  Toussaint's  idea  had  been 
that  by  heating  the  blood  in  this  way  the  bacteria  themselves  were 
killed.  Pasteur  15  showed,  however,  that  this  was  not  the  case,  but 
that  what  actually  occurred  was  a  reduction  of  the  virulence  of  the 
strain  by  the  exposure  to  heat.  As  a  matter  of  fact,  moreover,  the 
method  of  Toussaint  did  not  furnish  a  reliable  means  of  attenuating 
anthrax,  and  Pasteur  succeeded  in  developing  a  far  more  satis- 
factory procedure  on  which  he  based  a  practical  method  for  the  pro- 
tective vaccination  of  sheep  and  cattle. 

His  method  was  as  follows :  16  Virulent  anthrax  bacilli  were  cul- 
tivated at  42°  to  43°  C.  on  neutral  chicken  bouillon  (Sobernheim 
states  that  horse  or  beef  broth — or  even  agar — answers  the  same  pur- 
pose). Cultivated  under  these  conditions  a  gradual  and  progressive 
reduction  of  virulence  occurs.  After  about  12  days  of  such  cultiva- 
tion the  culture  as  a  rule  no  longer  kills  rabbits,  but  is  still  virulent 
for  guinea  pigs  and  mice.  After  twenty-four  or  more  days  the 
virulence  for  rabbits  and  guinea  pigs  is  lost  and  mice  only  can  be 
killed  with  it.  The  latter — the  most  fully  attenuated  strain — was 
called  premier  vaccin  by  Pasteur,  and,  in  the  immunization  of  cattle 
or  sheep,  is  first  injected.  After  10  or  12  days  the  stronger  deuxieme 
vaccin  is  administered.  This  is  the  method  which  Pasteur  used  in 
his  now  classical  experiments  at  Pouilly-le-Fort,  in  which  he  con- 

14  Toussaint.     Compt.  rend,  de  I'acad.  des  sc.,  1880. 

15  Pasteur,  Chamberland  and  Roux.     Compt.  rend,  de  Vacad.  des   sc.,  Vol. 
91,  1881. 

16  Cited  from  Sobernheim.     "Kraus  und  Levaditi  Handbueh  der  Technik, 
etc,,"  Vol.  1,  1909. 


ACQUIRED  IMMUNITY  65 

vinced  a  hostile  audience  of  the  efficacy  of  his  immunization.  Sheep 
were  protected  in  the  manner  indicated,  and  14  days  after  the  last 
injection  a  fully  virulent  culture  was  inoculated  and  the  animals 
found  capable  of  successfully  resisting  it. 

In  the  train  of  this  work  many  other  methods  of  producing  active 
immunity  have  been  devised — all  of  them  of  considerable  theoretical 
interest  and  many  of  them  practically  adapted  to  some  special  case. 
We  may  conveniently  classify  these  methods  as  follows : 


I.     IMMUNIZATION  WITH  LIVING  BUT  ATTENUATED  CULTURES 

(1)  Methods  in  which  the  attenuation  is  obtained  by  heating. 
This  is  the  method  of  Toussaint  as  outlined  above,  in  which  anthrax 
blood  was  heated  to  55°  C.  for  10  minutes,  and  is  probably  the  least 
efficient  or  reliable  method  for  the  attenuation  of  the  anthrax  bacillus. 
It  has  been  applied  to  rabies  by  Babes  (cited  from  Kraus  in  "Kraus 
u.  Levaditi  Handbuch,  etc.,"  Vol.   1,  p.  708),  who  attenuated  the 
virus  by  heating  to  58°  C.  for  periods  varying  from  2  to  40  min- 
utes. 

(2)  Attenuation  by  prolonged  cultivation  of  the  bacteria  at  tem- 
peratures above  the  optimum  for  their  growth.    This  is  illustrated  by 
Pasteur's  anthrax  immunization  as  described  in  the  preceding  para- 
graphs. 

(3)  Attenuation  by  passage  through  animals.   Examples  of  this 
are  Pasteur's  experiments  with  the  arouget"  organism,  in  which  pas- 
sage through  rabbits  diminished  the  virulence  for  hogs.     The  attenu- 
ation of  rabic  virus  by  passage  through  monkeys  is  another  instance, 
and  Jennerian  vaccination  is  also  an  example  of  this,  although  here 
the  attenuation  by  passage  through  cattle  is  attained  naturally  and 
not  by  experimental  procedures.     Based  *on  the  same  principle  is 
Behring's  17  method  18  of  immunizing  cattle  against  tuberculosis  by 
inoculating  them  with  tubercle  bacilli  of  the  human  type. 

(4)  Attenuation  by  prolonged  growth  of  bacteria  on  artificial 
media  in  the  presence  of  their  own  metabolic  products.     This  is  the 
method  first  employed  by  Pasteur  in  chicken  cholera,  as  described 
above,  and  is  applicable  to  many  organisms,  such  as  pneumococci, 
streptococci,  and  others.     In  fact,  it  is  difficult  to  maintain  the  viru- 
lence of  many  of  these  bacteria  unless  special  methods  of  cultivation 
or  passage  through  animals  are  practiced.   Pasteur  believed  that  free 
access  of  oxygen  to  the  cultures  increases  the  rapidity  of  the  attenua- 
tion. 

(5)  Attenuation  by  drying.      The  classical   example  for  this 
method  is  the  Pasteur  method  of  prophylactic  immunization  against 

17Behring.     "Therapie  der  Gegenwart,"  April,  1907. 

18  See  also  Romer,  "Kraus  u.  Levaditi  Handbuch,"  1st  Suppl.,  p.  310. 


66  INFECTION  AND  RESISTANCE 

rabies.  Rabbits  are  inoculated  with  virus  fixe,  and  their  spinal 
cords  dried  for  varying  periods  in  bottles  containing  KOH  at  a  tem- 
perature of  about  25°  C.  The  virus  grows  progressively  weaker 
with  each  day  of  drying.  Greater  details  concerning  this  method  are 
given  in  another  place  (see  page  489). 

(6)  Attenuation  by  the  use  of  chemicals. — Chamberland  and 
Roux 19   succeeded   in   attenuating   anthrax  by  growing   it   in  the 
presence  of  various  antiseptics.     They  used  carbolic  acid  1  to  600, 
bichromate  of  potassium  1  to  1,500  and  sulphuric  acid  1  to  200, 
and  found  that,  after  a  short  time  of  cultivation  under  such  condi- 
tions, the  bacilli  lost  their  ability  to  form  spores  and  became  avirulent 
for  sheep.     Behring  20  and  others  have  applied  this  method  to  the 
attenuation  of  diphtheria  toxin;  Behring  adds  terchlorid  of  iodin, 
Roux  potassium  iodid — iodin  solutions.     The  principle,  of  course, 
is  not  exactly  the  same  in  the  last  cases,  since  here  the  attenuation 
is  not  of  the  bacteria  themselves,  but  rather  of  the  toxin. 

(7)  Attenuation  by  cultivation  under  pressure.     This  method 
is  difficult  to  apply,  and  has  no  striking  advantages  over  other  pro- 
cedures.    It  was  employed  by  Chauveau  21  for  the  attenuation  of 
anthrax.     He   succeeded  in   accomplishing  this  by  cultivation   of 
anthrax  bacilli  at  28-39°  C.  at  a  pressure  of  8  atmospheres. 


II.  ACTIVE  IMMUNIZATION  WITH  FULLY  VIEULENT  CULTURES  IN 

SUBLETHAL    AMOUNTS 

The  original  methods  of  Pasteur  carried  out  with  chicken  cholera 
and  anthrax  were  aimed  particularly  at  diminution  of  virulence, 
since  these  organisms,  as  isolated  from  the  diseased  animal,  are  so 
extremely  infectious  that  it  would  be  very  difficult — (in  the  case  of 
many  animals,  impossible) — to  inoculate  with  the  unattenuated 
germs  without  producing  fatal  disease.  However,  in  the  case  of 
many  other  infections  it  has  been  found  feasible  to  inoculate  normal 
animals  with  the  fully  virulent  germs  in  such  small  quantities  that 
the  body  can  successfully  overcome  them,  and,  in  doing  so,  acquire 
specific  resistance.  It  is  obvious  that  this  method  is  more  easily 
carried  out  with  the  organisms  which  Bail  terms  "half  parasites" 
than  with  organisms  as  highly  infectious  as  anthrax.  Ferran 22 
applied  this  method  both  to  animals  and  to  human  beings  with  broth 
cultures  of  cholera  spirilla.  Hogyes 23  has  introduced  a  similar 
procedure  for  immunization  against  rabies  by  injecting  dilutions  of 

19  Chamberland  and  Roux.     Compt.  rend  de  I'acad.  des  sc.,  96,  1882. 

20  Behring  and  Wernicke.     Zeitschr.  f.  Hyg.,  12,  1892. 

21  Chauveau.     Compt.  rend,  de  I'acad.  des  sc.,  Vol.  98,  1884. 

22  Ferran,     Compt.  rend,  de  I'acad.  des  sc.,  1885. 

23  Hogyes.    "Lyssa  Nothnagels  Handbuch,  etc.,"  Vienna,  1897. 


ACQUIRED    IMMUNITY  67 

fully  virulent  rabic  virus,  beginning  with  a  dilution  of  1  to  10,000 
and  rapidly  working  up  to  a  dilution  of  1  to  10.  In  tuberculosis 
immunization  with  fully  virulent  cultures  in  small  amounts  has  been 
attempted  by  Webb,  Williams,  and  Barber,24  using  the  Barber  method 
of  isolation,  and  giving  a  single  micro-organism  at  the  first  injec- 
tion. That  such  a  method  is  feasible,  if  carried  out  with  sufficient 
care,  even  with  the  most  virulent  germs,  was  demonstrated  by  the 
same  workers.  They  succeeded  in  immunizing  animals  against 
anthrax  (with  cultures  kept  12  hours  on  agar)  25  by  injecting  a 
single  thread  (3  to  6  bacilli)  as  the  first  dose,  and  then  gradually 
increasing  the  amount. 

In  the  general  laboratory  immunization  of  animals  treatment 
with  virulent  bacteria  in  sublethal  doses  is  of  considerable  value  and 
frequently  employed. 

It  would  seem  that  possibly  this  method  or  some  modification  of 
it  will  be  found  to  have  very  definite  advantages  over  methods  in 
which  either  attenuated  or  dead  bacteria  are  employed.  Bail's  work 
upon  the  aggressins  and  upon  anti-aggressin  immunity  (see  chapter 
I,  page  21)  has  opened  the  possibility  that  virulent  bacteria  pro- 
vide, within  the  living  body,  specific  aggressive  substances  which  are 
not  produced  in  the  test  tube.  If  this  proves  to  be  true,  and  the 
question  is  by  no  means  settled,  it  may  be  necessary  in  such  cases 
to  immunize  with  organisms  which  are  in  a  condition  capable  of 
producing  these  aggressins.  Sublethal  doses  of  fully  virulent  or- 
ganisms would  furnish  these  conditions  more  perfectly  than  at- 
tenuated avirulent  strains,  in  which  the  invasive  (aggressive)  power 
is  considerably  diminished. 

The  methods  of  active  immunization  so  far  described  differ  from 
those  which  are  to  follow  in  that  the  preceding  were  all  based  upon 
the  use  of  living  bacteria  or  virus,  whereas  the  methods  to  be  de- 
scribed below  depend  upon  the  treatment  of  animals  with  dead  bac- 
teria or  bacterial  products.  It  is  well  to  call  attention  in  this  place 
to  the  fact  that  a  number  of  recent  investigations  seem  to  point  to 
the  greater  efficiency  of  immunization  with  living  germs.  This 
method  has  recently  given  hopeful  results  in  the  case  of  plague  in 
the  hands  of  Strong;26  and  Metchnikoff  and  Besredka,27  in  their 
attempt  to  vaccinate  chimpanzees  against  typhoid  fever,  make  the 
statement  that  vaccination  with  dead  typhoid  bacilli  or  autolysates 
does  not  confer  adequate  protection,  but  that  this  can  be  attained  by 
treatment  with  small  doses  of  the  living  bacilli. 

24  Webb,  Williams,  and  Barber.    Jour,  of  Med.  Res.,  Vol.  15,  1909. 

25  This  was  not  possible  where  the  organisms  were  taken  directly  from 
the  blood  of  a  dead  mouse.    In  such  cases  even  a  single  thread  caused  fatal 
disease. 

26  Strong.     Jour,  of  Med.  Res.,  May,  1908. 

27  Metchnikoff  and  Besredka.     Ann.  Past.,  Vol.  25,  1911. 


68  INFECTION    AND    RESISTANCE 

In  speaking  of  this  subject  it  is  well  to  mention  recent  ob- 
servations upon  immunization  with  "sensitized"  bacteria,28  although 
this  necessitates  anticipatory  reference  to  subjects  not  so  far  dis- 
cussed. It  is  a  matter  of  common  experience  in  laboratories  that 
rabbits  and  other  animals  will  withstand  relatively  large  amounts  of 
pathogenic  bacteria  if  these  are  first  treated  with  heated  specific 
immune  serum  (sensitized).  This  is  probably  due  to  the  fact  that 
such  "sensitized"  micro-organisms  are  very  rapidly  taken  up  by 
phagocytes.  In  spite  of  the  phagocytosis,  immunity  is  developed. 
MetchnikofF  and  Besredka,  in  the  communication  alluded  to  above, 
state  that  typhoid  vaccination  with  unaltered  living  bacilli  is  efficient, 
but  is  attended  by  severe  local  and  general  reactions.  If  the  living 
bacilli  are  first  "sensitized"  no  such  severe  reaction  occurs  and  im- 
munization is  nevertheless  successful.  The  recent  work  of  Gay  points 
in  the  same  direction,  and  it  is  at  least  possible  that  by  the  practice  of 
sensitization  we  may  be  able  to  employ  living  unattenuated  organisms 
for  purposes  of  immunization  more  extensively  than  we  have  in  the 
past. 


III.    ACTIVE  IMMUNIZATION  WITH  DEAD  BACTERIA,  AND  BACTERIAL 

EXTRACTS 

This  method  is  the  one  most  extensively  practiced  in  the  labora- 
tory immunization  of  animals.  It  is  usual  in  most  experiments  of 
this  kind  to  inject  dead  organisms  once  or  twice  before  living  bac- 
teria are  administered.  High  degrees  of  resistance  can  in  some 
instances  be  attained  by  progressively  increasing  doses  of  dead  cul- 
tures only.  This  method  is  not  only  useful  in  experimental  work, 
but  is  clinically  employed  in  the  active  immunization  of  human 
beings  as  introduced  by  Wright  and  as  applied,  before  Wright,  to 
tuberculosis  (tuberculin  treatment).  But  it  is  very  probable  that 
"the  immunity  so  attained  is  not  entirely  comparable  to  the  immunity 
following  an  attack  of  a  disease,  nor  even  that  produced  by  the  in- 
jection of  living  bacteria. 

The  method  employed  for  killing  the  bacteria  is  of  considerable 
importance  since,  both  by  excessive  heating  as  well  as  by  too  vigorous 
•chemical  treatment,  the  immunizing  properties  of  the  bacterial 
protein  may  be  destroyed.  In  employing  heat  it  is  a  safe  rule  never 
to  expose  the  bacteria  for  prolonged  periods  to  temperatures  which 
considerably  exceed  the  thermal  death  point.  As  a  rule,  heating  non- 
spore-forming  bacteria  to  a  temperature  of  from  65°  to  70°  C.  for 

28  Refer  to  p.  159  and  the  discussion  of  the  conception  of  "sensitization" 
which  follows. 


ACQUIRED    IMMUNITY  69 

thirty  minutes  will  suffice  to  kill  them  without  too  radically  altering 
the  immunizing  properties  of  the  protein  constituents.29 

If  the  temperature  is  not  raised  above  60°  C.,  and  this  is  ad- 
vised by  many  workers,  the  suspensions  must  be  carefully  controlled 
by  cultural  tests  before  they  are  used,  at  least  for  the  treatment  of 
human  beings.  As  we  shall  see  in  a  later  section,  the  best  results  have 
been  obtained  when  heating  was  not  carried  beyond  53°  to  55°  C. 

When  bacterial  death  is  to  be  accomplished  by  chemicals  the 
antiseptics  most  commonly  used  are  carbolic  acid  (0.5  per  cent.), 
toluol  (removed  before  use  of  vaccine  by  filtration  or  evaporation), 
chloroform,  and  formaldehyd  (1  per  cent.). 

Pfeiffer,  who  was  one  of  the  first  to  practice  the  immunization 
of  animals  with  dead  bacteria  on  an  extensive  scale,  believed  that, 
in  the  case  of  bacteria  which  were  toxic  by  reason  of  their  intra- 
cellular  constituents  (endotoxins),  the  injection  of  the  cell  protein 
itself,  whether  dead  or  alive,  was  the  sole  essential  for  successful 
immunization.  The  method  developed  by  Kolle  30  and  by  Pfeiffer 
and  Marx  31  for  the  prophylactic  immunization  of  human  beings 
against  cholera  depends  upon  the  injection  of  cholera  cultures 
emulsified  in  salt  solution,  killed  by  exposure  to. 5 8°  C.  for  one  hour, 
and  further  insured  against  contamination  by  the  addition  of  0.5 
per  cent,  phenol.  The  application  of  this  method  to  other  diseases, 
both  prophylactically  and  therapeutically,  is  more  fully  discussed 
in  another  place.  (See  chapter  XIX.) 

Since  the  essential  point  in  such  immunization  is  the  introduc- 
tion of  the  bacterial  protein,  it  is  often  customary  to  inject  bacterial 
extracts  instead  of  the  whole  cells.  This  has  been  especially  desir- 
able in  the  case  of  such  insoluble  micro-organisms  as  the  tubercle 
bacillus,  where  the  injection  of  the  whole  dead  organism  produces 
localized  reactions  similar  to  those  caused  by  the  living  bacteria.32 
Thus  "Old  Tuberculin,"  as  commonly  used,  is  a  glycerin-broth  ex- 
tract of  tubercle  bacilli.  The  method  has  been  extensively  used  and 
a  variety  of  procedures  have  been  devised  for  bacterial  extraction. 
These  have  included  simple  autolysis  of  the  bacterial  bodies  in  alka- 
line broth,  shaking  in  salt  solution  in  mechanical  shakers,  trituration 
with  salt  or  sand,  trituration  after  freezing,  digestion  with  proteo- 
lytic  enzymes,  and  extraction  by  pressure  in  a  Buchner  press. 

We  may  mention  some  of  the  more  important  methods  for  pre- 

29  In  a  subsequent  chapter  (p.  258)  we  shall  see  that  the  physical  changes 
produced  in  an  antigen  by  heat  result  in  differences  in  the  antibodies  formed 
after  animal  inoculation.     This  point  has  practical  significance  in  the  present 
connection.     See  also  the  chapter  on  agglutinins,  the  work  of  Joos  there  dis- 
cussed, and  Friedberger  and  Moreschi,  Centralbl.  f.  Bakt.,  1905,  Vol.  39. 

30  Kolle.    Deutsche  med.  Woch.,  1897,  p.  4. 

31  Pfeiffer  and  Marx.    Deutsche  med.  Woch.,  1898. 

32  Prudden  and  Hodenpyl.    N.  Y.  Med.  Journal,  1891. 


70  INFECTION    AND    RESISTANCE 

paring  bacterial  extracts  for  purposes  of  immunization  and  antigen 
production  in  general  as  follows: 


A.     Extraction  of  Bacteria  by  Permitting  Them  to  Remain  for 
Prolonged  Periods  in  Fluid  Media 

The  bacteria  may  be  grown  upon  slightly  alkaline  bouillon  and 
kept  at  incubator  temperature  for  one  to  two  months.  They  are 
then  filtered  through  Berkefeldt  or  other  suitable  filters.  This  is  the 
common  method  of  producing  antigen  for  precipitin  reactions,  in  fact 
the  method  employed  by  Kraus  in  the  discovery  of  the  bacterial  pre- 
cipitins.  It  is  by  no  means  certain  whether  the  antigens  prepared  in 
this  way  represent  simple  extractions  or  autolytic  products  of  the 
bacteria;  probably  both  processes  take  place.  The  antigenic  value 
of  the  fluids  obtained  in  this  way  is  never  very  great.  From  such 
filtrates  Brieger  and  Mayer,  Pick,  and  others  have  attempted  to 
obtain  the  antigen  in  a  purified  form  by  chemical  precipitation. 
Pick  33  precipitates  the  bouillon  filtrate  by  saturation  with  ammonium 
sulphate;  the  precipitate  is  redissolved  in  water  and  again  precipi- 
tated with  ammonium  sulphate  and  the  resultant  precipitate  dried 
on  a  filter.  It  is  then  dissolved  in  water  and  precipitated  with 
alcohol.  The  sticky  substance  which  comes  down  represents  the 
antigen. 

Suitable  extracts  can  occasionally  be  obtained  also  by  emulsify- 
ing agar  cultures  in  physiological  salt  solution  and  allowing  them 
to  stand  for  twenty-four  hours  or  more  at  incubator  temperature. 
In  our  own  experience  we  have  found  this  method  rather  inefficient 
for  yielding  strong  extracts.  More  efficient  extraction  is  usually  ob- 
tained when  the  bacteria  are  suspended  in  alkaline  fluids  such  as 
N 
—  sodium  hydrate.  Lustig  and  Galleotti  digest  the  bacterial  mass 

for  24  hours  with  1  per  cent.  NaOH,  then  precipitate  with  am- 
monium sulphate,  dry  in  vacuo  and  pulverize.34 

Recently,  also,  Uhlenhuth  35  has  employed  the  proprietary  prep- 
aration aantiformin" 36  for  the  production  of  antigens.  This 

33  Pick.     "Hoffmeister's  Beitrage,  etc.,"  Vol.  1,  1902.     For  an  extensive 
discussion  of  the  various  methods  employed  for  the  production  of  bacterial 
antigens  by  chemical  methods  see  Pick  in  Kraus  und  Levaditi,  etc.,  Vol.  1,  and 
the  same  author  in  Kolle  u.  Wassermann,  etc.,  2nd  Ed.,  Vol.  1. 

34  See  Pick.     LOG.  cit. 

55  Uhlenhuth.     Centralbl.  f.  Bakt.,  I,  Ref.  Vol.  42,  Beilage,  p.  62. 

36  "Antiformin"  is  a  substance  largely  employed  for  the  cleansing  of  pipes 
and  vats  of  organic  matter  because  of  its  powerfully  solvent  action.  Its 
value  in  concentrating  tubercle  bacilli  out  of  sputum  and  other  mixtures 
depends  upon  its  power  to  dissolve  the  tissue  elements  and  all  bacteria  except 
those  that  are  acid-fast.  Rosenau  ("Preventive  Medicine  and  Hygiene," 


ACQUIRED    IMMUNITY  71 

substance  thoroughly  dissolves  all  but  the  acid-fast  bacteria  when 
used  in  concentrations  of  2.5  per  cent.  Since  it  is  alkaline  it  is 
necessary  to  neutralize  it  with  hydrochloric  or  sulphuric  acid  before 
use. 

For  the  preparation  of  antigen  from  pneumococci  Neufeld  37  has 
utilized  the  solvent  action  upon  these  organisms  of  bile.  He  adds 
the  bile  and  broth  cultures  just  as  this  is  done  in  the  diagnostic 
"bile  test"  (0.1-0.2  c.  c.  of  fresh  bile  to  a  broth  culture;  sodium 
taurocholate  solution  can  also  be  used).  Many  bacteria  can  also  be 
broken  up  by  emulsifying  them  in  17  per  cent,  salt  solution  and 
allowing  them  to  stand  for  some  time  in  this  medium  and  then 
diluting  this  to  0.85  per  cent. 


B.     Extraction  by  Mechanical  Methods 

One  of  the  most  useful  methods  for  obtaining  extracts  of  bacteria 
within  a  relatively  short  time  is  that  which  Besredka  38  has  applied 
mainly  for  the  preparation  of  typhoid  (endotoxin),  24-hour  agar 
cultures  washed  up  in  very  small  quantities  of  physiological  salt 
solution,  killed  by  heat  at  60-65°  C.  and  dried  in  vacuo.  The  dried 
mass  is  mixed  with  a  measured  quantity  of  dry  salt  and  the  mix- 
ture thoroughly  triturated  in  a  mortar  for  a  considerable  time. 
While  triturating  distilled  water  is  added  in  small  quantities  until 
the  fluid  represents  a  0.85  per  cent,  salt  solution.  This  is  allowed 
to  stand  for  anywhere  from  a  few  hours  to  a  week,  and  the  bacteria 
are  then  removed  by  centrifugalization.  This  method  has  been  modi- 
fied by  many  observers  and  gives  good  results  whenever  thorough 
trituration  is  practiced.  It  is  also  probable  that  the  exposure  to  the 
hypertonic  salt  solution  in  the  earlier  stages  of  the  trituration  may  aid 
considerably  in  breaking  up  the  bacteria. 

Trituration  after  freezing  is  a  method  which  has  yielded  excel- 
lent results  in  the  hands  of  Macfadyen  and  others.  This  requires  a 
rather  complicated  piece  of  machinery  originally  described  by  Mac- 
fadyen and  Rowland.  The  principle  of  this  is  one  of  mechanical 
trituration  in  a  steel  cylinder  which  is  surrounded  by  an  ice-brine 
mixture  so  that  the  bacteria  and  sand  may  be  kept  frozen  during  the 
process. 

Appleton,  1913,  p.  1020)  gives  its  composition  as  follows:  "Antiformin 
consists  of  equal  parts  of  liquor  sodae  chlorinate  of  the  British  Pharmacopoeia 
and  a  15  per  cent,  solution  of  caustic  soda.  The  formula  for  liquor  sodae 
chlorinate  is  as  follows: 

Sodium  carbonate    600 

Chlorinated    lime     400 

Distilled    water     4,000" 

37Neufeld.     Zeitschr.  f.  Hyg.,  Vol.  34,  1900. 

38  Besredka.    Ann.  de  I'Inst.  Past.,  19-20,  1905,  190C. 


72  INFECTION    AND    RESISTANCE 

Mechanical  trituration  is  also  the  principle  of  the  production  of 
the  new  tuberculins  as  advised  by  Koch. 

One  of  the  earliest  methods  of  obtaining  bacterial  substances 
by  mechanical  means  was  that  used  by  Buchner  and  Hahn  39  in  the 
production  of  their  "plasmines."  The  bacteria  were  grown  in  quan- 
tity on  large  agar  surfaces,  the  moist  bacterial  masses  triturated 
together  with  quartz  and  were  then  subjected  to  high  pressure  in  an 
especially  constructed  press  spoken  of  as  the  "Buchner  press." 

Mechanical  breaking  up  and  extraction  of  the  bacteria  also  under- 
lies in  principle  the  use  of  the  variously  constructed  shaking  ma- 
chines. There  are  many  models  of  such  machines  on  the  market,  all 
of  them  designed  to  accomplish  prolonged  agitation  of  bacterial  emul- 
sions. In  many  cases  the  apparatus  can  be  placed  inside  of  an  incu- 
bator and  shaking  carried  on  at  37.5°  C.  The  bacteria  are  suspended 
for  this  purpose  in  distilled  water  salt  solution,  weak  alkali,  or  in 
serum,  and  glass  beads  or  sand  may  be  added  to  aid  in  their  mechan- 
ical injury.  Shaking  must  be  continued  for  24  hours  or  more  in 
order  to  give  good  results. 

IV.     ACTIVE  IMMUNIZATION  WITH  BACTERIAL  PRODUCTS  ( TOXINS) 

As  soon  as  the  investigations  of  Roux  and  Yersin  had  shown  that 
in  some  diseases,  at  least,  the  injury  sustained  by  the  infected  animal 
was  largely  due  to  the  soluble  toxins  produced  by  the  bacteria,  it 
was  logical  to  attempt  to  immunize  animals  with  such  products. 
Probably  the  first  attempts  in  this  direction  were  those  made  by 
Salmon  and  Smith  in  hog  cholera.  The  experiments  of  these  writers 
have  attained  much  historical  importance  since  they  represent  the 
first  purposeful  attempt  to  immunize  animals  with  the  products  of 
bacterial  metabolism.  In  the  actual  experiment,  however,  the  im- 
munization practiced  by  Salmon  and  Smith  was  probably  a  combina- 
tion of  immunization  by  bacterial  products  and  by  dead  bacteria. 
Nevertheless,  the  thought  of  immunization  with  bacterial  products 
was  the  underlying  one  in  their  experiments.  Working  with  the 
hog  cholera  bacillus  which  they  had  recently  discovered  they  im- 
munized pigeons  in  the  following  way:  The  bacilli  were  grown  in 
broth  for  two  weeks,  and  the  cultures  were  killed  by  exposure  to  58° 
to  60°  C.  for  several  hours.  One  and  one-fifth  cubic  centimeter  of 
this  culture  liquid  was  then  injected  into  pigeons,  and  after  three 
such  injections  the  inoculated  pigeons  withstood,  without  harm,  doses 
of  the  bacilli  which  rapidly  killed  untreated  animals.  Salmon  and 
Smith  40  stated  distinctly  in  their  conclusions  that :  "Immunity  may 

39  Buchner  and  Hahn.     Munch,  med.  Woch.,  1897. 

40  Salmon  and  Smith  on  "A  New  Method  of  Producing  Immunity  from 
Contagious  Disease,"  Proc.  Biol  Soc.,  Wash.,   D.   C.,  Ill,  1884,   6,   p.   29, 
printed  Feb.  22,  1886. 


ACQUIRED    IMMUNITY  73 

be  produced  by  introducing  into  the  animal  body  such  chemical 
products  (results  of  bacterial  growth  in  culture  fluids)  that  have  been 
produced  in  the  laboratory.41 

Similar  attempts  to  immunize  rabbits  against  certain  forms  of 
septicemia  by  the  injection  of  culture  filtrates  were  made  by  Cham- 
berland  and  Roux  42  in  1888,43  and  the  same  investigators  applied 
this  method  to  anthrax  immunization  just  prior  to  the  discovery  of 
diphtheria  toxin  by  Roux.  However,  neither  in  hog 44  cholera  45 
nor  in  the  other  infections  upon  which  this  method  was  first  tried  do 
the  bacteria  produce  a  true  soluble  toxin,  and  the  immunization 
which  was  accomplished  depended  probably  upon  the  injection  of 
bacterial  extracts.  Nevertheless,  these  attempts  had  shown  the  way 
in  a  new  direction,  and  bore  immediate  fruit  in  the  investigations  of 
Brieger  and  Fraenkel,46  and  more  especially  in  those  of  Beh- 
ring  47  48  49  and  his  collaborators.  Fraenkel,  though  following  the 
method  of  injecting  filtered  diphtheria  culture  fluids,  came  to  the 
erroneous  conclusion  that  the  toxin  and  the  immunizing  substances  in 
the  cultures  were  not  identical  (loc.  cit.,  p.  1135).50  The  degree  of 
immunity  obtained  in  his  experiments,  moreover,  was  a  slight  one 
only. 

Behring,  in  his  first  work,  in  collaboration  with  Kitasato,  suc- 
ceeded in  immunizing  animals  with  culture  filtrates  and  with  pleural 
exudates  of  diphtheritic  animals.  Similar  results  were  accomplished 
with  tetanus.  Since  the  publication  of  these  results — especially  in 
consequence  of  the  epoch-making  discovery  of  passive  immunization, 
to  which  they  were  the  immediate  guides,  toxin  immunization  has 
been  investigated  and  accomplished  in  all  cases  in  which  a  true  solu- 
ble toxin  can  be  demonstrated.  It  has  accordingly  been  carried 
out  with  the  exotoxins  of  pyocyaneus 51  bacilli,  the  bacilli  of 
symptomatic  anthrax  52  and  botulinus,53  the  specific  leukocidins  54 

41  For  a  copy  of  the  original  paper  'iy  Salmon  and  Smith  I  am  indebted 
to  Professor  Theobald  Smith. 

42  Chamberland   and  Roux.     Ann.  Past.,  Vol.  1,  1888. 

43  Op.  Cit.,  Vol.  2,  1889. 

44Joest  in  "Kolle  u.  Wassermann  ^andbuch,  etc.,"  Vol.  3,  p.  632. 

45  Karlinski.     Zeitschr.  f.  Hyg.,  Vol.  28,  1898. 

46  Brieger  and  Fraenkel.    Berl.  kl.  Woch.,  1890,  Nos.  11,  12  and  49. 

47  Behring  and  Kitasato.     Deutsche  med.  Woch.,  No.  49,  1890. 

48  Behring.    Deutsche  med.  Woch.,  No.  50,  1890. 

49  Behring  and  Wernicke.     Zeitschr.  f.  Hyg.,  1892. 

50  De  Schweinitz,  indeed,  who  further  studied  hog  cholera  immunization 
(Medic.  News,  1892;   Centralbl.  f.   Bakt.,  Vol.   20,   1896),   claimed  that  in 
sterilized  milk  the  bacillus    produced   "enzymes"   with  which  immunization 
could  be  accomplished. 

51  Wassermann.     Zeitschr.  f.  Hyq.,   Vol.   22. 
52Kempner.     Zeitschr.  f.  Hyg.,  Vol.  23,  1897. 

53  Grassberger  and   Shattenfroh.     Deuticke,  Wien,  1904. 

54  Denys  and  Van  der  Velde.     "La  Cellule,"  Vol.  2,  1895. 


74  INFECTION    AND    RESISTANCE 

produced  by  staphylococci,  and  with  various  bacterial  hemolytic  poi- 
sons (tetanolysin  and  other  bacterial  hemotoxins).  The  result  of  all 
this  work  has  been  the  very  important  determination  that  susceptible 
animals  may  be  actively  immunized  both  against  the  effects  of  the 
toxin  alone,  as  well  as  against  the  virulent  bacteria  themselves,  by 
systematic  treatment  with  culture  filtrates  containing  the  toxins. 
Since  in  many  cases  the  effects  of  the  toxins  were  so  powerful  that 
their  attenuation  was  desirable,  Behring  and  others  have  advised  the 
addition  of  iodinterchlorid  and  other  chemicals  to  the  first  in- 
jections. 

PASSIVE    IMMUNIZATION 

In  the  logical  development  of  the  fundamental  facts  regarding 
immunization,  with  attention  focused  early  on  the  blood  and  body 
fluids  as  the  probable  carriers  of  immunity,  it  was  but  a  rational 
step  from  active  immunization  to  the  conception  that  such  acquired 
immunity  might  be  transferred  from  a  treated  to  a  normal  animal 
by  injecting  blood  from  the  former  into  the  latter.  This  was  prob- 
ably the  underlying  thought  of  Toussaint's  55  early  work  with  an- 
thrax, in  which  he  heated  anthrax  blood  to  55°  C.  and  injected  it 
into  other  animals,  wrongly  believing  that  the  bacteria  had  been 
killed  by  the  heating.  The  method  of  Toussaint,  however,  was  vague 
in  its  conception,  and  in  no  way  constitutes  an  example  of  true 
passive  immunization.  The  beginning  was  made  in  a  purposeful 
and  clearly  conceived  way  by  Richet  and  Hericourt.56 

These  investigators  actively  immunized  dogs  against  stapkyio- 
cocci,  and  then  attempted  to  transfer  the  immunity  to  normal  rabbits 
by  injecting  defibrinated  blood  from  the  immune  dogs.  Their  suc- 
cess was  a  partial  one  only,  for  reasons  that  we  will  discuss  directly. 
Reasoning  similar  to  that  of  Richet  and  Hericourt  was  applied  by 
Babes  and  Lepp  57  to  rabies  immunization.  When  the  blood  of 
rabies-immune  dogs  was  injected  into  normal  dogs  and  rabbits,  and 
these  inoculated  with  rabies  several  days  later,  the  treated  animals 
regularly  survived  the  controls,  but  in  one  dog  only  was  the  occur- 
rence of  rabies  absolutely  prevented.  Since  their  animals  were  nof 
experimentally  inoculated,  but  subjected  to  the  more  uncertain 
method  of  allowing  them  to  be  bitten  by  a  mad  dog,  and  since  the 
series  included  4  animals  only  (2  treated  and  2  controls),  Babes  and 
Lepp  were  unable  to  draw  definite  conclusions.  The  establishment  of 
passive  immunization  as  a  proved  scientific  fact  was  finally  accom- 

55  Toussaint.     Compt.  rend,  de  I'acad.  des  sc.,  1880. 

56  Richet  and  Hericourt.     Compt.  rend,  de  I'acad.  des  sc.,  1888,  Vol.  107, 
p.  750. 

67  Babes  and  Lepp.    Ann.  Past.,  Vol.  3,  1889. 


ACQUIRED    IMMUNITY  75 

plished  in  1890-1892  by  Behring  and  Kitasato,58  and  by  Bebring 
and  Wernicke.  Tbe  results  of  this  work — tbe  direct  outcome  of 
their  success  in  actively  immunizing  with  soluble  toxins,  is  sum- 
marized in  their  first  paper  as  follows:  "The  blood  of  tetanus- 
immune  rabbits  possesses  tetanus-poison-destroying  properties ;  these 
properties  are  demonstrable  in  the  extravascular  blood  and  in  the 
serum  obtained  from  this ;  these  properties  are  of  so  lasting  a  nature 
that  they  remain  active  in  the  bodies  of  other  animals,  so  that  one 
is  enabled  to  obtain  positive  therapeutic  results  by  transfusing  the 
blood  or  injecting  the  serum.  These  tetanus-poison-destroying  prop- 
erties are  absent  from  the  blood  of  non-immune  animals,  and  when 
the  tetanus  poison  is  inoculated  into  normal  animals  it  can  be  dem- 
onstrated as  such  in  the  blood  and  other  fluids  of  these  animals  after 
death." 

With  these  researches  begins  the  therapeutically  practicable 
method  of  passive  immunization  which  is  now  in  such  widespread 
and  successful  use  in  the  treatment  of  diphtheria,  in  the  prophylactic 
treatment  of  tetanus,  and,  to  a  less  common  and  less  successful  de- 
gree, in  the  treatment  of  dysentery,  typhoid  fever  (Besredka), 
plague,  and  a  number  of  other  bacterial  diseases.  The  same  method 
has  been  successful  in  the  treatment  of  various  diseases  of  domestic 
animals.  The  principle  was  also  applied  by  Ehrlich  59  to  ricin  and 
crotin  immunity,  in  the  formulation  of  which  he  succeeded  in  work- 
ing out  passive  immunization  on  a  quantitative  basis,  showing  that 
the  degree  of  immunity  in  such  cases  could  be  directly  referred  to 
the  amounts  of  the  specific  antitoxin  present  in  the  blood  of  the  im- 
munized animal.  Calmette,60  and  Physalix  and  Bertrand,61  then 
succeeded  in  producing  passive  immunization  against  snake  venoms. 

To  summarize  the  success  of  passive  immunization  in  general  we 
may  say  that  it  has  achieved  its  greatest  usefulness  in  the  case  of 
those  diseases  in  which  the  pathogenesis  depends  upon  a  true  exo- 
toxin — which,  as  we  have  mentioned  before,  leads  to  the  formation 
of  an  antitoxin  in  the  immunized  animal.  In  these  cases  the  passive 
immunization  is  accomplished  by  the  transfer  of  the  antitoxins  from 
the  treated  to  the  normal  animal. 

In  the  case  of  bacterial  infections  in  which  no  true  toxin  is 
formed — where  no  antitoxin  results  and  the  immunity  depends,  as 
we  shall  see,  upon  an  enhancement  of  the  bactericidal  and  phagocytic 
properties  of  the  blood  and  the  cells,  passive  immunization  has  not 
been  a  practical  therapeutic  success.  The  probable  reasons  for  this 
cannot  be  properly  discussed  until  we  have  examined  more  closely 
into  the  mechanism  by  which  the  immune  animal  is  protected  after 

58  Behring  and  Kitasato.    Deutsche  med.  Woch.,  No.  49,  1890. 

59  Ehrlich.    Deutsche  med.  Woch,  1891;  Fortschr.  d.  Med.,  p.  41,  1897. 

60  Calmette.    Compt.  rend,  de  la  soc.  de  biol.,  1894. 

61  Physalix  and  Bertrand.     Compt.  rend,  de  la  soc.  de  biol.,  1894. 


76  INFECTION    AND    RESISTANCE 

specific  treatment  with  bacteria  or  their  products.  The  moderately 
beneficial  effects  of  the  various  antiplague  sera  and  the  limited  suc- 
cess attending  the  use  of  antistaphylococcus,  antistreptococcus,  and 
antipneumococcus  sera  probably  depend,  as  recent  work  tends  to 
show,  not  upon  the  direct  action  of  antitoxic  bodies,  but  rather  upon 
the  indirect  opsonic  action  62  1(i3  64  65  which  renders  the  bacteria 
more  easily  amenable  to  phagocytic  action.  These  points  we  shall 
discuss  at  greater  length  in  a  succeeding  chapter. 


SPECIFICITY 

In  speaking  of  methods  of  immunization  in  the  preceding  sec- 
tions we  have  frequently  employed  the  terms  ''specific"  and  ''spe- 
cificity," without  sufficiently  defining  them.  It  will  be  necessary  to 
explain  them  since  the  principle  of  specificity  is  at  the  same  time  one 
of  the  most  important  and  one  of  the  most  mysterious  of  the  phe- 
nomena of  immunity.  When  an  individual  has  recovered  from  an 
attack  of  typhoid  he  is  thereafter  immune  to  typhoid — but  to  no 
other  disease — similarly  with  plague,  cholera,  small-pox,  etc.  The 
same  principle  governs  artificial  immunization.  Vaccination  against 
anthrax  protects  against  anthrax  only — and  active  or  passive  im- 
munization in  any  of  the  infectious  diseases  produces  a  resistance 
which  is  "specifically"  aimed  only  at  the  particular  infectious  agent 
writh  which  the  original  active  immunity  was  produced.  The  prin- 
ciple points  to  an  exquisite  chemical  difference  between  the  protein 
substances  which  constitute  the  bacterial  cell  bodies  or  their  meta- 
bolic products.  For  although  by  chemical  methods  we  can  detect  no 
differences  between  them — yet  the  reactions  of  immunity  are  sharply 
differentiating.  When  we  come  to  consider  the  antibodies  which 
specifically  precipitate  the  substances  by  which  they  are  incited  we 
shall  see  that  the  delicacy  and  consequent  differential  value  of  these 
reactions  far  outstrip  any  known  chemical  methods,  and  it  is  upon 
this  principle  indeed — inexplicable  as  it  is — that  the  great  diagnostic 
value  which  these  reactions  have  attained  depends.  The  conception 
of  the  specificity  of  the  causes  of  infectious  disease,  as  well  as  that 
of  the  specificity  of  toxins,  has  become  so  common  and  self-evident 
to  us  that  we  are  too  apt  to  forget  how  fundamental  to  progress  the 
establishment  of  this  fact  was  in  the  early  days  of  bacteriological 
research.  When,  in  1878,  Koch  published  his  treatise  on  the  "Eti- 
ology of  Wound  Infections"  specificity  was  not  generally  accepted, 
and  the  supposed  metamorphosis  of  bacterial  species,  as  asserted  by 

62Neufeld.     Deutsche  med.  Woch.,  No.  11,  1897. 

63  Neufeld  and  Rimpau.     Deutsche  med.  Woch.,  No.  40,  1904. 

64  Bail  and  Kleinhans.     Zeitschr.  f.  Imm.,  Vol.  12,  1912. 

65  Weil.     Zeitschr.  f.  Hyg.,  75,  1913. 


ACQUIRED    IMMUNITY  77 

Hallier  and  others,66  had  first  to  be  scientifically  refuted  by  Cohn, 
Koch,  and  their  pupils,  before  it  could  be  assumed  that  a  given  in- 
fectious disease  was  always  the  result  of  infection  with  a  definite 
and  constant  species  of  bacteria.  The  same  applied  to  the  specificity 
of  toxins — and  rational  investigations  into  the  reaction  of  the  animal 
body  against  bacterial  poisons  was  not  possible  until  the  works  of 
Roux  and  Yersin  on  diphtheria  and  that  of  Kitasato  on  tetanus  had 
differentiated  between  the  true,  specific  bacterial  poisons  and  the 
unspecific  ptomains  and  "sepsins"  of  Selmi,  Nencki,  and  Brieger. 

66  Hallier.      Cited    from    Behring,    "Bekampfung    der    Infektionskrank- 
heiten,"  Leipzig,  1894. 


CHAPTER    IV 

THE  MECHANISM  OF  NATURAL  IMMUNITY  AND  THE 

PHENOMENA   FOLLOWING   UPON    ACTIVE 

IMMUNIZATION 

ANTIBODIES  AND  ANTIGENS.       THE  ORIGIN  OF  ANTIBODIES 

THE   MECHANISM   OF    NATURAL   IMMUNITY 

PASTEUK'S  work  on  active  immunization  was  carried  out  in  the 
later  seventies  and  the  early  eighties.  During  and  immediately  after 
this  time  it  was  very  natural  that  the  attention  of  investigators 
should  have  concentrated  upon  the  elucidation  of  the  causes  under- 
lying both  the  natural  resistance  against  bacteria  observed  in  animals 
and  man,  and  the  changes  which  during  active  immunization  were 
fundamentally  responsible  for  the  acquisition  of  resistance. 

It  was  easily  determined  that  there  were  no  anatomically  and 
physiologically  determinable  differences  between  the  various  mam- 
malia which  could  account  for  the  observed  striking  variations  of 
susceptibility,  nor  could  gross  anatomical  or  histological  changes 
be  noted  in  an  animal  which  had  been  artificially  immunized.  Mor- 
phologically such  an  animal  was  indistinguishable  both  in  the  size 
and  appearance  of  its  organs,  and  in  the  arrangement  and  structure 
of  its  cells  from  any  other  individual  of  the  same  species  not  sub- 
jected to  treatment. 

It  was  a  natural  development  of  the  investigations  brought  to 
bear  upon  this  problem  that  attention  should,  for  a  time,  be  concen- 
trated upon  the  phenomena  of  inflammation,  processes  which  were 
regularly  associated  with  infections  of  all  kinds  and  seemed  indeed 
to  represent  a  sort  of  local  expression  of  tissue  resistance  to  the 
invading  micro-organisms. 

It  was  in  the  course  of  investigations  upon  the  nature  of  inflam- 
mation that  Metchnikoff  first  became  interested  in  problems  of  re- 
sistance. In  1883  he  presented  a  paper  at  the  Naturalists'  Congress 
at  Odessa,  in  which  he  referred  the  absorption  of  dead  or  foreign 
corpuscular  elements  in  the  bodies  of  invertebrates  to  a  process  of 
intracellular  digestion  carried  out  by  phagocytic  cells.  As  early  as 
1874  Panum  had  suggested  that  possibly  resistance  against  invading 
micro-organisms  might  be  due  to  a  similar  intracellular  destruction, 

78 


PHENOMENA    FOLLOWING    IMMUNIZATION         79 

and  Metchnikoff,  soon  after  his  first  communication,  extended  his 
phagocytic  studies  to  phenomena  of  infection.  His  first  investiga- 
tions concerned  themselves  with  an  infectious  disease  caused  by  a 
form  of  yeast  in  a  small  crustacean — the  daphnia  or  water  flea. '  He 
showed  that  recovery  or  death  from  the  disease  depended  upon  the 
completeness  with  which  the  invading  micro-organisms  were  taken 
up  by  the  white  blood  cells  found  in  the  body  cavity  of  the  daphnia. 
Immediately  subsequent  studies,  carried  out  with  the  aid  of  numer- 
ous pupils,  embraced  an  extensive  material  throughout  the  animal 
kingdom  in  which  he  attempted  to  show  parallelism  between  natural 
immunity  and  the  phagocytic  activities  mobilized  by  the  body  against 
the  invading  germs. 

Meanwhile  studies  along  another  path  were  in  progress.  It  had 
been  observed  many  years  before  this  by  the  physician,  Hunter,  that 
the  shed  blood  of  animals  was  not  as  easily  subject  to  putrefactive 
change  as  wTere  many  other  organic  substances.  Similar  observations 
by  Traube  and,  in  1881,  by  Lord  Lister  a  (the  latter  reported  at  a 
time  when  Pasteur's  experiments  were  reaping  their  first  practical  re- 
sults) further  stimulated  investigation  of  the  blood  as  the  possible 
seat  of  the  increased  antibacterial  property.  For,  indeed,  these 
observations  seemed  to  imply  that  by  resisting  decomposition,  even 
when  inoculated  with  putrefying  material,  the  blood  must  possess 
definite  means  of  inhibiting  or  even  destroying  the  putrefactive 
bacteria. 

In  1884,  in  a  dissertation  submitted  at  Dorpat,  Grohman  2  stated 
that  cell-free  blood  plasma  inhibited  the  growth  of  micro-organisms. 
But  Grohman  was  unable  to  determine  actual  bacterial  destruction. 
Similar,  but  inconclusive,  observations  were  published  by  Von  Fo- 
dor  3  in  1887.  In  1888,  however,  Nuttall,4  who  was  investigating 
the  validity  of  the  phagocytic  theory  of  Metchnikoff,  experimentally 
determined  that  normal  blood  possessed  the  property  of  killing  bac- 
teria— a  property  now  spoken  of  as  "bactericidal"  power.  The  atti- 
tude taken  by  Nuttall,  and  others  of  the  Fliigge  school,  toward 
Metchnikoff' s  opinions  was  one  of  doubt  as  to  the  fundamental  sig- 
nificance of  phagocytosis  in  determining  resistance.  They  argued 
that  Metchnikoff  had  not  yet  proved  that  living  bacteria  were  taken 
up  by  the  phagocytic  cell,  and  that  the  action  of  these  cells  might 
therefore  be  interpreted  as  merely  a  process  of  removal  of  the  dead 
bacteria,  after  these  had  been  killed  by  other  influences.  USTuttall, 
accordingly,  repeated  some  of  MetchnikofFs  experiments  on  anthrax 
in  frogs  and  rabbits,  essentially  confirmed  the  basic  observations, 
but  showed  also  that  the  cell-free  defibrinated  blood  of  these  and 

1  Lister.     Trans.  Intern.  Med.   Congress,  London,  1881. 

2  Grohman.     Cited  from  Lubarsch,  Centralbl.  /.  Bakt.,  6,  1889. 

3  Fodor.    Deutsche  med.  Woch.,  No.  34,  1887. 
*  Nuttall.     Zeitschr.  f.  Hyg.,  Vol.  4,  1888. 


80  INFECTION    AND    RESISTANCE 

other  animals  possessed  definite  bacteria-destroying  properties  (bac- 
tericidal power)  for  many  different  micro-organisms.  He  detected 
similar  properties  in  pleural  exudates,  pericardial  fluids,  and  aqueous 
humor,  and  determined  that  this  property  was  "inactivated"  or  de- 
stroyed when  the  fluids  were  heated  to  55°  C.  for  10  minutes  or 
longer.  Buchner 5  then  confirmed  NuttalPs  results  and  showed 
further  that  the  bactericidal  property  resided,  not  only  in  defibri- 
nated  blood,  peptone  blood,  and  plasma,  but  was  present  also  in  the 
serum  obtained  after  clotting.  He  applied  the  term  "Alexin"  to  this 
active  constituent  of  the  blood — likening  its  action  to  that  of  a 
ferment. 

The  immediate  theoretical  result  of  these  discoveries  was  an  at- 
tempt, begun  by  Fliigge's  school,  to  base  natural  as  well  as  acquired 
resistance  upon  the  bactericidal  properties  of  the  blood  and  body 
fluids  in  general.  For  the  observations  of  Nuttall  and  Buchner  were 
soon  extended  to  peritoneal  and  other  exudates  by  Stern,6  and  to 
ascitic  fluids  by  Prudden.7  By  these  two  groups,  that  of  Fliigge- 
Nuttall-Buchner  on  the  one  hand,  and  that  of  Metchnikoff  on  the 
other,  there  were  founded  the  two  schools  of  immunity — the  humoral 
and  the  cellular,  both  originating  in  attempts  to  explain  natural 
immunity,  and  later  extending  to  problems  of  acquired  resistance. 
And  it  is  to  the  diligent  and  ingenious  intellectual  and  experimental 
conflict  between  these  schools  that  we  owe  much  of  the  knowledge  we 
now  possess  concerning  the  phenomena  of  immunity.  A  bridge  be- 
tween them  was  early  established  when  Buchner  himself — (even 
before  Metchnikoff) — suggested  the  possible  leukocytic  origin  of  the 
bactericidal  serum  constituent  (alexin).  The  later  work  of  Denys, 
of  Gruber  and  Futaki,  of  Wright,  of  Neufeld,  of  Bail,  and  of  others 
has  demonstrated,  as  was  to  be  expected,  the  inadequacy  of  either 
point  of  view  by  itself,  and  the  intimate  interdependence  of  the 
humoral  and  the  cellular  processes. 

As  concerns  the  relation  of  bactericidal  serum  effects  and  natural 
immunity,  it  could  be  unquestionably  shown  by  Nuttall,  Buchner, 
Kissen,8  and  their  immediate  followers  that  the  blood  of  most  ani- 
mals possessed  bactericidal  properties  against  many  micro-organisms, 
their  experiments  being  so  planned  that  the  participation  of  leuko- 
cytes could  be  absolutely  excluded.  However,  a  parallelism  between 
bactericidal  power  and  the  degree  of  natural  resistance  could  not 
be  established.  Lubarsch,9  writing  during  the  early  periods  of  the 
controversy,  stated  that  "he  would  regard  the  (purely  humoral)  10 

5  Buchner.     Centralbl  f.'Bakt.,  1889. 

6  Stern.     Zeitschr.  f.  klin.  Med.,  Vol.  18   (cited  from  Hahn). 

7  Pnidden.     Med.  Eec.,  Jan.,  1890. 
8Nissen.     Zeitschr.  /.  Hyg.,  Vol.  6. 

9  Lubarsch.     Centralbl.  f.  Bakt.,  Vol.  6,  1889. 
*°  Bracketed  phrase  our  own. 


PHENOMENA    FOLLOWING    IMMUNIZATION         81 

experiments  of  Nuttall  as  decisively  contradicting  the  phagocytic 
theory  if  the  bactericidal  action  of  the  blood  (for  anthrax  bacilli) 
could  be  shown  to  be  more  potent  in  immune  than  in  suscep- 
tible animals."  Metchnikoff  1 1  himself,  taking  this  point  of  view, 
called  attention  to  the  fact  that  the  blood  serum  of  rabbits,  animals 
that  are  highly  susceptible  to  anthrax,  is  more  powerfully  bactericidal 
for  these  micro-organisms  than  is  the  blood  of  dogs  or  even  that  of 
immunized  calves,  both  of  which  are  much  more  resistant  than  are 
rabbits.  Nuttall  answered  this  by  reporting  that  the  blood  of  an- 
thrax-immunized calves  is  actually  more  powerfully  bactericidal  than 
is  that  of  normal  calves.  Although  this  argument  of  JSTuttall  was 
perfectly  valid  in  principle,  it  exerted  little  influence  on  opinions 
at  this  time  because  anthrax  happens  as  a  matter  of  fact  to 
belong  to  that  group  of  infections  in  which  bactericidal  protection 
is  actually  secondary  to  phagocytic,  and  Lubarsch  could  show  that 
the  differences  observed  by  Nuttall  were  often  less  than  those 
obtaining  between  specimens  of  blood  taken  from  individual  normal 
rabbits. 

Lubarsch  himself,  then,  in  carefully  planned  experiments, 
showed  that  rabbits  and  cats  could  be  killed  with  quantities  of  anthrax 
bacilli  far  less  than  the  number  which  the  extravascular  blood  of 
these  animals  can  destroy.  He  concluded  that  the  resistance,  in  these 
cases  at  least,  is  certainly  not  parallel  writh  the  bactericidal  properties 
of  the  blood,  and  suggested  the  possibility  that  the  intravascular  blood 
does  not  possess  bactericidal  power  to  the  same  degree  in  which  it  is 
possessed  by  the  extravascular  plasma  or  serum.  This  point,  first 
raised  by  Lubarsch — namely,  the  possibility  of  a  difference  between 
the  intravascular  blood  and  the  extravascular  blood  serum  or  plasma 
in  bactericidal  functions — soon  became  one  of  the  focal  points  of 
the  controversy,  since  Metchnikoff,  admitting  the  bactericidal  power 
of  the  shed  blood,  assumed  that  this  was  purely  the  result  of  sub- 
stances given  off  by  the  leukocytic  cell-bodies  after  extravascular 
injury. 

The  Metchnikoff  school  defended  its  premise  by  the  dual  method 
of  attempting  on  the  one  hand  to  establish  a  parallelism  between 
phagocytic  activity  and  natural  resistance,  and,  on  the  other  hand, 
by  showing  that  the  cell-free  blood  serum  of  naturally  resistant  ani- 
mals often  furnished  an  excellent  culture  medium  for  the  bacteria 
in  question.  Thus  Wagner  showed  that  anthrax  bacilli  grow  well 
in  the  blood  of  fowls  at  42°  C.,  and  Metchnikoff  himself  called  at- 
tention to  the  fact  that  pigeons'  blood  is  an  excellent  medium  for  the 
cultivation  of  the  Pfeiffer  bacillus,  whereas  the  living  pigeon  is  en- 
tirely insusceptible  to  influenzal  infection.  Arguments  based  on 
such  observations,  however,  have  lost  much  of  their  original  weight, 
for  we  have  since  then  learned  more  about  the  delicate  quantitative 

11  Metchnikoff.     Virch.  Archiv,  Vol.  97,  1884. 


82  INFECTION    AND    RESISTANCE 

conditions  and  the  difficulties  of  accurate  measurements  obtaining  in 
experiments  upon  in  vitro  bactericidal  phenomena.  For  although  a 
specimen  of  the  blood  of  a  naturally  immune  animal  may  be  capable 
of  destroying  a  considerable  number  of  bacteria  of  a  given  species,  the 
implantation  of  such  a  specimen  with  a  slight  excess  of  the  bacteria 
would  soon  exhaust  the  active  serum  constituents  and  profuse  growth 
could  then  take  place.  Furthermore,  the  conditions  of  temperature 
established  in  cultural  experiments  lead  rapidly  to  a  deterioration 
of  the  alexin  necessary  for  bactericidal  action,  and  any  bacteria 
remaining  alive  at  the  end  of  a  number  of  hours  would  then  have  un- 
opposed opportunity  to  multiply. 

The  attempts  to  establish  parallelism  between  phagocytic  activity 
and  natural  immunity,  though  somewhat  more  successful  than  the 
analogous  efforts  of  the  humoral  school,  nevertheless  also  failed  to 
furnish  complete  explanation  for  existing  conditions,  and,  as  we  shall 
see,  no  adequate  generalizations  could  be  made  until  later  years  re- 
vealed the  close  cooperation  between  cells  and  fluids.  We  must  post- 
pone any  attempts  to  do  justice  to  this  phase  of  the  problem,  there- 
fore, until  we  are  in  a  position  to  discuss  the  question  of  phagocytosis 
on  the  basis  of  a  fuller  knowledge  of  the  phenomena  which  influence 
it. 

The  clear  thinking  and  unprejudiced  logic  brought  to  bear  upon 
this  controversy  by  some  of  the  great  bacteriologists  of  this  time  are 
nowhere  more  instructively  illustrated  than  in  a  short  introduction 
published  by  v.  Behring  12  to  his  second  article  on  diphtheria.  He 
says:  "Neither  deduction  nor  theorizing  can  at  present  decide 
whether  a  compromise  will  be  found  in  the  future  between  the  two 
hypotheses  (humoral  and  cellular),  or  whether  the  one  or  the  other 
alone  will  be  found  correct.  As  yet  the  opinions  of  many  experiment- 
ing bacteriologists  are  in  direct  opposition  in  this  respect.  Mean- 
while, for  the  purposes  of  medical  advancement  and  therapeutic  suc- 
cess it  is  not  necessary  to  await  a  decision  of  this  question.  ...  It  is 
indeed  of  advantage  to  the  cause  if  the  struggle  against  infection  is 
undertaken  from  the  most  varied  points  of  view;  attempts  to  make 
proselytes  for  a  dogma  have  never  led  to  progress.  In  this  sense  I  will 
try  to  summarize  those  experimental  results  which  support  the  hu- 
moral point  of  view  without  attempting  particularly  to  detract  from 
the  importance  of  opinions  which  I  do  not  share." 

THE  PHENOMENA  FOLLOWING  UPON  ACTIVE  IMMUNIZATION 

The  cellular  and  humoral  points  of  view,  formulated  largely  upon 
the  facts  of  natural  immunity,  were  equally  applied,  almost  from 
the  beginning,  to  the  explanation  of  active  immunization.     The  light 
12  v.  Behring.     Zeitschr.  f.  Hyg.,  Vol.  12,  1892, 


PHENOMENA    FOLLOWING    IMMUNIZATION         83 

thrown  upon  these  phenomena  by  the  efforts  of  both  schools  rapidly 
led  to  a  complete  abandonment  of  those  earlier  theories  of  immunity 
which  had  conceived  the  acquired  resistance  of  animals  against  bac- 
teria as  a  purely  passive  development  in  the  body  of  conditions 
unfavorable  for  bacterial  gro  /vth. 

Among  these  earlier  theories,  now  of  historical  interest  only,  are 
the  "Exhaustion  Theory"  of  Pasteur  and  the  "Retention  Theory"  of 
Nencki,13  Chauveau,  and  others. 

Pasteur's  views,  defended  for  a  time  also  by  Garre,  held  that  the 
growth  of  any  given  variety  of  bacteria  in  the  animal  body  exhausted 
certain  specific  nutritive  substances  necessary  for  this  growth.  Sub- 
sequent lodgment  in  the  same  body  was  impossible  owing  to  the 
absence  of  proper  nutrient  material.  It  is  interesting  to  note,  as 
Kolle  14  points  out,  that  this  theory  is  in  principle  very  similar  to 
the  "Atrepsie"  idea  of  Ehrlich  advanced  in  explanation  of  species 
immunity  to  cancer. 

The  hypotheses  of  Chauveau,  of  Nencki,  and  others  were  the 
converse  of  those  of  Pasteur.  They  were  based  purely  on  inference, 
assuming  that  conditions  occurring  in  the  test  tube  could  be  applied 
also  to  those  existing  in  the  animal  body.  Baumann  15  had  shown 
that,  among  other  things,  phenol  was  produced  as  a  result  of  bacterial 
putrefaction.  Nencki  had  noticed  the  inhibition  of  bacteria  in 
culture  by  the  products  of  their  own  metabolism.  Wernicke,16  too, 
had  demonstrated  the  presence  of  phenol,  phenylacetate,  skatol,  and 
other  aromatic  compounds  harmful  to  bacteria  in  putrefying  mix- 
tures. The  reasoning  which  formulated  the  so-called  "Retention  The- 
ory," therefore,  was  the  following :  Bacteria  growing  in  the  animal 
body  produce  certain  substances  peculiar  to  their  own  metabolism, 
which  eventually  lead  to  inhibition  of  their  growth.  By  the  retention 
of  these  products  the  animal  is  rendered  immune.  Chauveau's  adher- 
ence to  this  theory  was  largely  based  on  the  fact  that  he  had  observed 
immunity  in  the  lambs  born  of  Algerian  ewes  which  had  recovered 
from  anthrax  shortly  before  or  during  parturition.  He  explained 
this  by  a  transference  of  the  retention  products  from  mother  to  off- 
spring. As  a  matter  of  fact  the  observation  could  just  as  well  have 
been  utilized  as  support  for  the  Exhaustion  Theory. 

Both  the  theory  of  "Exhaustion"  as  well  as  that  of  "Retention" 
could  not  long  withstand  experimental  criticism.  .Theories  which 
were  not  so  easily  disproved  and  which  have  given  rise  to  much  in- 
vestigation are  the  "Alkalinity  Theory,"  first  formulated  by  v.  Beh- 

13  Nencki.      Jour.    f.    prakt.    Chem.,    May,    1879,    cited   from    Sirotinin, 
Zeitschr.  f.  Ilyg.,  Vol.  4,  1888. 

14  Kolle  in  "Kolle  u.  Wassermann  Handbuch,"  2d  Ed.,  Vol.  1. 

15  Baumann.     Zeitschr.  f.  physiol.  Chem.,  Vol.  1. 
10  Wernicke.     Virch.  Archiv,  Vol.  78. 


84  INFECTION    AND    RESISTANCE 

ring,17  and  the  "Osmotic  Theory"  of  Baumgarten.18  In  the  iormer 
an  attempt  was  made  to  demonstrate  a  parallelism  between  blood  alka- 
linity and  bactericidal  action — the  latter  was  based  on  the  supposi- 
tion that  the  destruction  of  bacteria  in  the  body  was  largely  due  to 
harmful  osmotic  conditions.  Neither  of  these  theories  was  long 
seriously  maintained.  Behring  himself  took  an  active  part  in  the 
subsequent  development  of  our  present  views.  Baumgarten  10  still 
clings  to  his  own  opinion  in  a  modified  way,  in  that  he  maintains 
that  the  only  effect  produced  by  specific  antibodies  upon  cells — bac- 
terial or  otherwise — is  that  they  change  the  permeability  of  the  cell 
membranes  and  render  them  more  vulnerable  to  osmotic  injury. 

However  crude  or  vague  these  theories  may  seem  to  us  now,  it 
must  not  be  forgotten  that  they  were  conceived  at  a  time  when  no 
knowledge  had  been  gained  regarding  specific  "antibodies."  The 
phagocytic  powers  to  which  Metchnikoff  attributed  natural  immunity 
and  the  bactericidal  powers  of  the  blood,  regarded  in  the  same  light 
by  the  Fliigge  school,  were  general  properties  possessed  by  many 
animals  toward  many  different  micro-organisms.  That  immuniza- 
tion could  specifically  increase  these  functions  toward  the  particular 
micro-organisms  used  for  treatment  seemed  indicated  by  the  experi- 
ments of  Nuttall  in  which  higher  bactericidal  power  was  found  in 
the  blood  of  anthrax-immune  calves  than  in  that  of  normal  animals. 
However,  no  definite  and  conclusive  work  on  the  specific  increase  of 
measurable  serum  or  cell  properties  was  available. 

This  great  advance,  giving  new  energy  and  pointing  out  new 
paths  of  investigation,  carne  in  1890-1892  with  the  publication  of  the 
work  of  Behring  and  his  collaborators,  Kitasato  and  Wernicke,  on 
immunity  to  diphtheria  and  tetanus.  As  we  have  indicated  in  a  pre- 
ceding paragraph,  the  fundamentally  important  points  of  this  work 
were  as  follows : 

1.  The  establishment  of  the  fact  that  animals  may  be  actively 
immunized  with  products  of  bacterial  metabolism — true  toxins  or 
exotoxins. 

2.  The  discovery  that  such  active  immunity  was  dependent  upon 
specific   antibodies   formed   in   the   treated   animal   and  circulating 
freely  in  the  blood ;  and, 

3.  That,  by  the  transfer  of  the  blood  or  the  blood  serum  contain- 
ing these  specific  antibodies  other  normal  animals  could  be  passively 
protected — not  prophylactically  only,  but  even  after  active  disease 
had  set  in. 

These  observations  were  rapidly  confirmed  for  tetanus  by  Tiz- 
zoni  and  Cattani,  and  by  Yaillard,  and,  similar  but  less  successful 
attempts  at  passive  immunization  were  made  in  other  diseases  by 

17  v.  Behring.     Centralbl  f.  Jclin.  Med.,  1888,  No.  38. 

18  Baumgarten.     Berl.  klin.  Woch.,  1899,  1900. 

19  Baumgarten.     ILehrbuch,   etc.,  1912. 


PHENOMENA    FOLLOWING    IMMUNIZATION         85 

Foa,  Emmerich,  Bouchard,  and  many  others.  The  discovery  of 
passive  immunization  established  the  fact  of  specific  alteration  of 
the  blood  by  active  immunization,  and  represented,  for  the  time, 
a  distinct  triumph  for  the  humoral  hypothesis. 

Summarizing  the  knowledge  of  immunity  as  it  stood  at  the  close 
of  this  period,  Behring  says :  "In  the  case  of  natural  immunity  no 
generally  applicable  explanation  has  as  yet  been  found.  (By  this  he 
referred  to  the  lack  of  complete  parallelism  between  natural  im- 
munity and  either  the  bactericidal  or  the  phagocytic  activities.)  For 
artificial  immunization,  however,  it  has  now  been  shown,  in  a  number 
of  carefully  studied  infections,  that  we  can  surely  attribute  it  to 
properties  of  the  cell-free  blood." 

Within  a  very  short  time  after  Behring  and  Kitasato's  first  paper 
Ehrlich  20  demonstrated  that  the  principle  discovered  by  them  was 
not  limited  to  bacterial  poisons.  He  was  investigating  immuniza- 
tion against  ricin  in  mice,  and  showed  that  here,  too,  the  blood  of 
the  immune  animals  contained  a  body  which  would  antagonize  the 
toxic  action  of  ricin,  and  which,  injected  into  normal  mice,  would 
passively  protect  them.  He  spoke  of  this  blood  constituent  as  "anti- 
ricin." 

It  is  natural  that  extensive  generalization  followed  these  discov- 
eries. However,  while  it  was  found  that  the  blood  of  all  actively 
immunized  animals  possessed  a  certain  degree  of  protective  power 
for  normal  individuals,  it  was  soon  shown  that  this  was  not  due  in 
all  cases  to  antagonism  to  the  bacterial  poisons  on  the  part  of  the 
immune  blood  serum.  In  immunity  to  the  Vibrio  Metchnikovi — in 
pneumococcus  and  cholera  immunity — Sanarelli,21  Isaeff,22  Pfeiffer 
and  Wassermann,23  and  a  number  of  others  showed  that  here,  unlike 
diphtheria  and  tetanus,  the  protective  power  of  the  immune  serum 
did  not  rest  on  "antitoxic"  properties,  but  rather  on  antagonism  to 
the  bacteria  themselves.  It  soon  became  definitely  established  that 
antitoxic  immunity  resulted  only  in  the  cases  of  those  bacteria  in 
which  a  true  soluble  exotoxin  was  produced,  and  where  the  disease 
following  infection  was  primarily  due  to  the  absorption  of  these 
poisons.  The  antibodies  incited  in  the  blood  of  toxin-immune  ani- 
mals were  therefore  spoken  of  by  Behring  and  Ehrlich  as  "anti- 
toxins" and  their  action — after  a  number  of  false  hypotheses — was 
finally  recognized  as  a  direct  neutralization  of  the  bacterial  poisons. 

The  strict  specificity  of  these  antibodies  was,  from  the  first,  clear 
to  v.  Behring,  who  observed  that  diphtheria-immune  serum  and 
tetanus-immune  serum  acted  each  upon  its  respective  toxin  only.  It 
was  recognized  at  the  same  time  that  the  passive  immunity  produced 

20  Ehrlich.    Deutsche  med.  Woch.,  No.  32,  1891. 

21  Sanarelli.    Ann.  Past.,  Vol.  7,  1893. 

22  Isaeff.     Ibid,  and  Zeitschr.  f.  Hyg.,  Vol.  16,  1894. 

23  Pfeiffer  and  Wassermann.     Zeitschr.  f.  Hyg.,  Vol.  14.  1893. 


86  INFECTION    AND    RESISTANCE 

by  injecting  antitoxic  sera  is  almost  immediately  established ;  that,  by 
proportionately  increasing  the  amount  of  antitoxin,  immunity  can 
be  produced  against  any  amount  of  toxin;  and  that  this  passive  or 
transferred  immunity  is  of  relatively  short  duration. 

The  antitoxins,  then,  as  we  shall  see  in  the  more  detailed  analysis 
of  their  action  (in  chapter  V),  are  specific  poison-neutralizing  anti- 
bodies formed  in  the  blood  of  animals  immunized  with  a  true  bac- 
terial toxin  or  exotoxin — conferring  resistance  or  immunity,  not  by 
influencing  the  bacteria,  but  by  rendering  innocuous  the  specific  bac- 
terial poisons. 

The  therapeutic  successes  of  passive  immunization  achieved  with 
tetanus  and  diphtheria  very  naturally  led  to  a  careful  inquiry  into 
the  antitoxic  properties  of  the  blood  of  animals  immunized  with  all 
known  pathogenic  bacteria  and  bacterial  products,  and  with  many 
poisons  of  animal  and  vegetable  origin. 

Contrary  to  earlier  expectations,  however,  the  list  of  bacteria 
against  which  antitoxic  immunity  can  be  achieved  has  remained  rela- 
tively small,  limited  in  fact,  as  we  have  previously  stated,  to  those 
species  which  produce  a  soluble  exotoxin.  The  inciting  of  a  specific 
neutralizing  antibody  (antitoxin),  however,  is  also  a  property  of 
many  other  substances  of  proteid  nature  which  are  for  this  reason 
classified  biologically  with  the  true  toxins  or  exotoxins.  In  fact, 
the  one  absolutely  constant  attribute  which  defines  our  conception 
of  the  "true  toxins"  and  the  substances  classified  with  them  is  their 
antitoxin-inciting  power.  We  classify  a  bacterial  product  as  a 
"toxin"  or  "exotoxin"  only  if  it  incites  a  neutralizing  "antitoxin" 
in  the  serum  of  an  immunized  animal. 

The  first  discovery  of  a  non-bacterial  antitoxin-stimulating  sub- 
stance was,  as  we  have  stated,  that  of  ricin  by  Ehrlich,24  1891,  and 
this  was  soon  followed  by  similar  determinations  for  abrin  and  robin 
— other  vegetable  poisons.  In  1894  Calmette,25  and  Physalix  and 
IBertrand  26i  extended  the  principle  to  poisons  of  animal  origin  by 
demonstrating  antitoxin  formation  against  snake  poison.  And  that 
similar  specific  neutralizing  bodies  were  formed  in  response  to  im- 
munization with  ferments  was  shown  in  1900  by  Morgenroth.27 

The  more  important  individual  substances  which  may  be  bio- 
logically grouped  together  because  of  their  property  of  inciting  a 
specific  antitoxin  (or  toxin-neutralizing  body)  in  the  blood  of  im- 
munized animals  may  be  tabulated  as  follows : 

Diphtheria  toxin — (loc.  cit.  Behring  &  Wernicke). 
Tetanus  toxin — (loc.  cit.  Behring  &  Kitasato). 

24  Ehrlich.     Deutsche  med.   Woch.,  1891;  Fortschr.  d.   Med.,  1891,  1897. 

25  Calmette.     Ann.  Past.,  Vol.  8,  1894. 

28  Physalix  and  Bertrand.     Compt.  rend,  de  la  soc.  de  biol.,  1894. 
27  Morgenroth.    Centralbl.  f.  Bakt.,  26,  1899. 


PHENOMENA    FOLLOWING    IMMUNIZATION         87 

The  Toxin  of  the  Bacillus  of  Symptomatic  Anthrax — (Grassberger  & 
Shattenfroh,  Munch.  Med.  Woch.,  1900,  1901  and  10  c.  cit.). 

The  Toxin  of  the  Bacillus  Botulinus — (Kempner,  Zeitschr.  f.  Hyg.,VoL  26, 
1897). 

The  Toxin  of  the  Bacillus  Pyocyaneus — (Wassermann,  Zeitschr.  f.  Hyg., 
Vol.  22,  1896). 

The  Toxin  of  the  Dysentery  Bacillus  (?)  Shiga-Kruse  type — (Kraus  u. 
Doerr,  Wien.  klin.  Woch.,  1905). 

The  leukocyte  poison  of  the  Staphylococcus  pyogenes  aureus,  Leucocidin — 
(Denys  &  Van  de  Velde,  La  cellule,  1895). 

The  Hemolytic  Poisons  of  Various  Bacteria  (see  Pribram  in  "Kraus  und 
Levaditi  Handbuch,"  Vol.  II,  p.  223). 

Proteolytic  Ferments  of  the  Hog  Cholera  Bacillus  (De  Schweintz,  Medical 
News,  1892). 

The  Toxin  of  the  Cholera  Spirillum  (?)  Brau&  Denier,  Compt.  rend,  de  Vacad. 
des  sc.,  1906,  Kraus,  Centralbl.  f.  Bakt.,  1906,  and  Wien.  klin.  Woch.,  1906). 

Ricin — (Ehrlich,  loc.  cit.). 

Abrin — (Ehrlich,  loc.  cit.). 

Krotin — (Ehrlich,  loc.  cit.). 

Snake  venom — (Calmette,  loc.  cit.). 

Spider  poison— (Sachs,  " Hoffmeister's  Beitrage,"  1902,  and  Ehrlich,  "Ge- 
sammelte  Arbeiten,"  etc.). 

Lab.  enzyme — (Morgenroth,  loc.  cit.). 

Pepsin— (Sachs,  Fortschr.  d.  Med.,  1902). 

Trypsin— (Achalme,  Ann.  Past.,  1901). 

Leukocytic  ferments  Leukoprotease — (Jochmann  &  Miiller,  Munch,  med. 
Woch.,  1906). » 

The  period  of  investigation  which  was  initiated  by  the  discovery 
of  the  specific  antitoxins  was  replete  with  efforts  to  determine  true 
toxins  and,  consequently,  antitoxic  immunity  for  all  pathogenic 
bacteria.  We  have  already  mentioned  that  in  many  cases  these 
efforts  were  futile — the  bacteria  in  question  being  found  to  secrete 
no  exotoxin  and  the  immunity  established  against  them  developing 
without  the  formation  of  demonstrable  antitoxin.  Metchnikoff 29 
showed  this  to  be  the  case  with  hog  cholera  as  early  as  1892,  and  the 
investigations  of  Sanarelli,  IsaefT,  and  Pfeiffer  and  Wassermann 
pointed  in  the  same  direction. 

Perhaps  the  clearest  definition  of  the  conditions  prevailing  dur- 
ing immunization  of  animals  with  non-toxin-forming  bacteria  was 
that  formulated  at  this  time  by  Pfeiffer.  The  importance  of  the  bac- 
tericidal power  of  serum,  as  discussed  before  this  by  Fliigge,  Nut- 
tall,  and  others,  had  dealt  largely  with  variations  of  this  general 
property  in  relation  to  natural  immunity,  but  had  failed  to  recognize 
clearly  a  specific  increase  in  these  powers  during  active  immuniza- 

28  This  list  includes  all  the  important  antitoxin-inciting  substances.     For 
a  more  complete  tabulation  see  Wassermann  in  "Kolle  u.  Wassermann  Hand- 
buch, etc.,"  Vol.  IV,  1st  Ed.,  p.  498.     Our  own  list  is  adapted  from  the  one 
there  ^iven. 

29  Metchnikoff.     Ann.  Past.,  1892. 


88  INFECTION    AND    RESISTANCE 

tion.  Pfeiffer  with  Wassermann  30  had  studied  the  pathogenicity  of 
cholera  spirilla  for  guinea  pigs,  and  had  come  to  the  conclusion  that 
the  animals  died  of  toxemia  (and  not  of  bacteriemia,  as  claimed  by 
Gruber  and  Wiener),  and  that  this  toxemia  was  due  to  the  liberation 
of  poisons  from  the  dead  bodies  of  cholera  vibrios,  killed  by  the 
serum  of  the  infected  animals.  Pfeiffer  31  now  showed  that  the  in 
jection  of  cholera  spirilla  killed  with  chloroform  brought  about  a  tox- 
emia identical  with  that  following  inoculation  with  living  cultures. 
He  further  determined  that  the  resistance  of  animals  against  cholera 
was  due  to  the  bactericidal  effects  of  the  serum,  which  killed  the 
injected  cholera  spirilla,  and  not  to  any  poison-neutralizing  property. 

Isaeff,32  one  of  Pfeiffer's  pupils,  continuing  this  work,  expresses 
his  own  and  Pfeiffer's  conceptions  as  follows:  "Guinea  pigs  vac- 
cinated against  cholera,  in  spite  of  high  immunity  to  infection  with 
living  spirilla,  do  not  develop  any  immunity  to  cholera  [endo]33 
toxins.  The  blood  of  immunized  guinea  pigs  possesses  no  antitoxic 
properties.  The  maximal  dose  of  cholera  'toxin'  which  immunized 
guinea  pigs  can  withstand  is  not  higher  than  that  which  can  be  borne 
by  normal  animals,  and  but  slightly  higher  than  the  maximal  dose 
of  living  spirilla,  which  they  can  survive.  The  blood  of  cholera-vac- 
cinated guinea  pigs  possesses  strong  specific  protective  powers.  The 
same  specific  immunizing  properties  are  demonstrable  in  the  blood 
of  cholera  convalescents  toward  the  end  of  the  third  week  of  the 
disease." 

The  path  was  thus  cleared  for  a  definite  conception  of  cholera 
immunity,  and  this  was  formulated,  in  their  next  communication, 
by  Pfeiffer  and  Isaeff.34  35  36  In  this  paper  they  showed  that  the 
cholera  spirilla  injected  into  the  peritoneum  of  a  cholera-immune 
guinea  pig  were  subjected  to  a  rapid  dissolution,  a  process  which 
could  be  observed  by  taking  small  quantities  of  exudate  out  of  the 
peritoneum,  at  varying  intervals,  with  capillary  pipettes.  ]STo  such 
dissolution  occurred  in  normal  pigs  or  with  normal  serum.  But  the 
same  rapid  swelling,  granulation,  and,  finally,  dissolution  occurred 
when  the  spirilla  were  injected  into  the  peritoneal  cavity  of  a  nor- 
mal guinea  pig,  together  with  the  serum  of  an  immunized  animal. 
The  process  took  place  apparently  without  the  cooperation  of  the 
leukocytes  or  other  cells,  and  was  absolutely  specific.  For  instance, 
no  "lysis"  occurred  when  the  vibrios  "Nordhafen,"  "Massauah,"  and 
other  cholera-like  organisms  were  injected  into  cholera-immune  pigs, 

30  Pfeiffer    and    Wassermann.     Zeitschr.    f.    Hyg.,    Vol.    14,    1893;    also 
Pfeiffer,  Zeitschr.  f.  Hyg.,  Vol.  16,  1894. 

31  Gruber  and  Wiener.     Archiv  /.  Hyg.,  Vol.  15,  1893. 
2  Isaeff.     Zeitschr.  f.  Hyg.,  Vol.  16,  1894. 

83  Bracketed  word  our  own. 

34  Pfeiffer  and  Isaeff.     Zeitschr.  f.  Hyg.,  Vol.  17,  1894. 

35  Pfeiffer.     Ibid.,  Vol.  18,  1894. 

36  Pfeiffer  and  Isaeff.    Deutsche  med.  Woch.,  No.  13,  1894. 


PHENOMENA    FOLLOWING    IMMUNIZATION         89 

but  took  place  regularly  when  true  cholera  strains,  from  -  various 
sources,  were  used  in  the  experiment.  The  immunity  of  cholera- 
treated  animals,  therefore,  was  found  to  be  an  antibacterial  and  not 
an  antitoxic  one.  Cholera  spirilla  introduced  into  a  normal  animal 
were  permitted  to  multiply  and  accumulate  until  a  sufficient  number 
were  present  to  furnish,  upon  cell  death,  a  fatal  dose  of  poison.  In 
immunized  animals  the  small  quantities  of  bacteria  first  introduced 
succumbed  rapidly  to  the  lytic  properties  of  the  serum  and  accumu- 
lation was  prevented. 

By  these  experiments,  now  commonly  spoken  of  as  the  "Pfeiffer 
Phenomenon/'  it  was  definitely  proved  that  active  immunization 
with  bacteria  incites  in  the  serum  of  the  treated  animal  a  potent  in- 
crease of  bactericidal  properties — an  increase  which  is  entirely  spe- 
cific in  that  the  bactericidal  power  toward  bacteria  other  than  those 
employed  in  the  immunization  does  not  exceed  the  normal.  The 
immunity  in  these  cases,  then,  is  not  antitoxic,  but  rather  "antibac- 
terial" and  depends  on  the  development,  in  the  immune  sera,  of  anti- 
bodies quite  distinct  from  the  "antitoxins."  These  immune  serum 
constituents  were  spoken  of  by  Pfeiffer  as  "bacteriolysins"  or  "spe- 
cific bactericidal  substances." 

Not  long  after  the  discovery  of  the  specific  bacteriolysins  another 
property  of  immune  sera  was  described  by  Gruber  and  Durham.37 
They  had  been  studying  bacteriolytic  phenomena  with  colon  and 
cholera  organisms,  and  noticed  that  these  bacteria  were  rapidly  ag- 
glomerated and  gathered  in  small  clumps  when  emulsified  in  homolo- 
gous immune  serum.  Similar  clumping  had  indeed  been  described 
before.  Metchnikoff,  Isaeff,  Washburn,  and  Charrin  and  Eoger  had 
described  it  on  various  occasions,  but  had  not  recognized  it  as  a 
specific  property  of  immune  serum.38  Gruber  and  Durham  studied 
it  carefully,  determined  that  it  was  present  to  a  degree  roughly  pro- 
portionate to  the  degree  of  immunization  attained,  and  that  its 
specificity  was  such  that  it  could  be  utilized  for  bacterial  differen- 
tiation. They  believed  that  the  substances  in  the  immune  serum 
responsible  for  this  agglutination  were  independent  of  other  serum 
constituents  and  applied  to  them  the  term  "agglutinins." 

The  problems  of  immunization  had  now  considerably  expanded 
and  the  nature  of  the  new  serum  reactions  was  assiduously  studied. 
Primarily  the  phenomenon  of  agglutination  was  regarded  as  a  part 
of  the  struggle  of  the  body  against  the  living  bacteria  and  Gruber 
himself  believed  that  it  depended  upon  a  swelling  or  "klebrig  wer- 
den"  of  the  micro-organisms  which  tended  to  cause  their  sticking 
together,  and  rendered  them  more  readily  amenable  to  the  action  of 
the  bactericidal  powers  of  the  serum.  Bordet,39  however,  early  con- 

37  Gruber  and  Durham.     Munch,  med.  Woch.,  1896. 
ss  ;por  references  see  chapter  on  Agglutinins. 
39  Bordet.     Ann.  Past.,  1896. 


90  INFECTION    AND    RESISTANCE 

ceived  the  process  as  a  physical  phenomenon  in  which  the  bacteria 
themselves  were  entirely  passive,  and,  indeed,  Widal  40  soon  demon- 
strated that  bacteria  killed  by  heat  were  equally  as  agglutinable  as 
the  living  germs. 

This  naturally  suggested  that  the  reaction  between  specific  agglu- 
tinating serum  and  bacteria  was  based  on  individual  peculiarities  of 
the  bacterial  proteins,  and  it  occurred  to  Kraus,41  accordingly,  to 
investigate  whether  or  not  the  immune  sera  would  cause  any  sort  of 
reaction  when  mixed  with  the  dissolved  body  substances  of  homolo- 
gous bacteria.  Working  at  first  with  cholera  and  plague,  he  pre- 
pared solutions  of  bacterial  proteins,  both  by  allowing  broth  cultures 
to  stand  for  varying  periods  and  by  emulsifying  agar  cultures  in 
alkaline  broth.  The  extracts  were  then  filtered  through  Pukal  filters 
to  remove  the  bacterial  bodies.  When  the  sera  of  immunized  ani- 
mals were  added  to  these  clear  filtrates — cholera  serum  to  cholera 
filtrate,  and  plague  serum  to  plague  filtrate,  slight  turbidity  devel- 
oped and  was  followed  within  twenty-four  hours  by  the  formation  of 
small  flakes.  In  other  words,  it  was  found  that  the  mixture  of  a 
clear  filtrate  of  a  bacterial  culture  with  the  serum  of  an  animal 
immunized  against  these  bacteria  resulted  in  the  formation  of  a 
precipitate.  The  reaction  was  found  to  be  as  strictly  specific  as  that 
of  agglutination. 

Although,  from  the  beginning,  Paltauf  42  attempted  to  associate 
the  phenomena  of  agglutination  and  precipitation,  the  property  of 
precipitating  homologous  culture  filtrates  was  attributed  by  Kraus 
and  others  to  specific  antibodies  in  the  immune  sera,  distinct  and 
independent  of  those  previously  described,  and  spoke  of  them  as 
"preciptiins." 

The  discovery  of  the  various  "antibodies"  so  far  discussed  re- 
sulted from  the  study  of  the  direct  action  of  blood  serum  upon  bac- 
teria and  bacterial  products.  This  did  not,  however,  completely  de- 
flect the  attention  of  investigators  from  the  unquestionable  impor- 
tance of  phagocytosis  in  the  defence  of  animals  against  bacterial  in- 
vasion. Metchnikoff  and  his  school  continued  diligently  to  pursue 
this  other  phase  of  the  study  of  immunity  and,  although  the  increas- 
ing knowledge  of  serum  antibodies  continued  to  strengthen  the  prem- 
ises of  the  purely  humoral  point  of  view,  it  had  still  to  be  admitted 
that  in  some  diseases — particularly  anthrax  and  the  pyogenic  coccus 
infections,  phagocytosis  must  largely  be  held  responsible  for  recov- 
ery. It  was  found,  moreover,  by  the  later  investigations  of  Denys, 
Wright,  Neufeld,  and  others  that  phagocytosis  in  immunized  animals 
was  far  more  extensive  and  efficient  than  in  normal  ones,  and  that 

40  Widal.    La  semaine  medicale,  No.  5,  1897. 

41  Kraus.     Wien.  klin.  Woch.,  No.  32,  1897. 

42  Paltauf.    "Discussion  of  Kraus'  Paper,"  Wien.  kl  Woch.,  No.  18,  1897. 
p.  431. 


PHENOMENA    FOLLOWING    IMMUNIZATION         91 

this  depended  on  specific  constituents  of  the  immune  serum  which 
rendered  the  bacteria  more  amenable  to  the  phagocytic  action  of  the 
cells.  These  further  antibodies  we  will  discuss  in  a  subsequent  chap- 
ter, under  the  terms  "opsonins"  and  "bacteriotropins,"  designations 
applied  to  them  by  their  discoverers. 

We  have  thus  reviewed  briefly  the  various  specific  properties 
which  develop  in  the  serum  of  an  animal  when  it  is  systematically 
treated  (actively  immunized)  with  bacteria  or  bacterial  products. 
These  serum  activities  have  been  attributed  to  the  development  in 
the  serum  of  substances  which  we  speak  of  as  "antibodies." 

In  our  discussion  of  the  first  of  these  antibodies,  antitoxin,  we 
call  attention  to  the  fact  that  the  principle  discovered  in  the  case  of 
bacterial  toxins  was  rapidly  extended  to  vegetable  poisons,  snake 
venom,  spider  poison  and  enzymes.  It  was  found  that  the  power  of 
inciting  antitoxins  when  injected  into  animals  was  an  attribute  be- 
longing to  a  large  group  of  substances  in  nature,  and  not  limited  to 
bacteria  alone.  A  similar  generalization  of  conception  has  been  pos- 
sible with  other  antibodies.  Specific  lysins,  agglutinins,  and  pre- 
cipitins  may  be  produced  by  the  treatment  of  animals  with  many 
substances  not  of  bacterial  nature. 

The  first  observation  of  this  kind  was  made  almost  simultaneously 
by  Bordet  43  and  by  Belfanti  and  Carbone.44  They  observed  that 
the  serum  of  an  animal  that  had  been  treated  with  the  red  cells  of 
another  species  acquired  the  power  of  laking  these  cells.  That  the 
normal  serum  of  one  species  is  often  toxic  to,  and  causes  the  laking 
of,  the  erythrocytes  of  another  species  is  an  observation  that  dates 
back  to  the  earliest  experiments  on  transfusion,  and  had  been  studied 
in  considerable  detail  by  Landois  as  early  as  1875.  The  phenom- 
enon possesses  much  interest  in  its  bearing  on  the  problems  of  ana- 
phylaxis  and  will  be  discussed  more  particularly  in  that  connection. 
We  mention  it  in  this  place  to  show  that,  like  bactericidal  bodies, 
"hemolytic"  (erythrocyte  laking)  properties  may  be  present  in  nor- 
mal sera,  though  irregularly  and  by  no  means  occurring  in  every 
species  of  animal.  Incidentally  it  may  be  stated  that  this  is  true 
also  of  agglutinins  and  of  opsonins  which  may  be  found  in  consider- 
able amounts  in  normal  sera.  Of  precipitins,  however,  this  does  not 
seem  to  be  true. 

By  the  work  of  Bordet  it  was  found  that  "hemolysins"  could  be 
specifically  45  incited  in  an  animal  by  systematically  treating  it  with 

43  Bordet.    Ann.  Past.,  Vol.  12,  1898. 

44  Belfanti  and  Carbone.     Giorn.  della  R.  Acad.  di  Torino,  July,  1898. 

45  By  the  use  of  the  word  specific  in  this  case  we  imply  that  an  animal 
immunized  with  any  given  variety  of  red  blood  cells  will  form  hemolysins 
for  this  variety  only.     Thus  an  animal  treated  with  ox  blood  will  form  ox 
blood  hemolysins   only,   and   his   serum,   though  strongly  hemolytic   for   ox 
blood,  will  not  lake  sheep  cells,  dog  cells,  human  cells,  etc. 


92  INFECTION    AND    RESISTANCE 

the  red  blood  cells  of  another  species.  Apart  from  the  great  interest 
attaching  to  this  discovery  in  itself,  it  has  had  a  very  profound  influ- 
ence upon  investigations  on  immunity  generally,  since  it  has  fur- 
nished a  method  of  studying  lysis  far  more  simple  and  easily  con- 
trolled than  is  the  analogous  phenomenon  of  bacteriolysis.  And 
since,  in  fundamental  principles,  bacteriolysis  and  hemolysis  are 
essentially  alike,  much  of  our  knowledge  regarding  the  former  has 
been  arrived  at  by  experiments  upon  the  latter.  The  specific  hemo- 
lysins,  then,  are  antibodies  formed  in  response  to  "immunization" 
with  red  blood  cells,  analogous  to  the  similarly  produced  "bacterio- 
lysins."  Because  both  of  these  antibodies  exert  definite  injury  upon 
cells,  we  speak  of  them  by  the  group  names  of  "cytolysins"  or  "cyto- 
toxic"  substances. 

The  discovery  of  hemolysins  naturally  suggested  the  use  of  other 
cells,  and  the  following  years  brought  forth  many  reports  of  further 
specific  cytotoxins.  In  1899,  Metchnikoff,46  and  very  soon  after- 
ward Landsteiner,47  described  specific  "spermotoxins"  which  ap- 
peared in  the  blood  of  animals  treated  with  spermatozoa.  Von  Dun- 
gern  48  obtained  analogous  substances  by  injecting  ciliated  epithe- 
lium from  the  trachea.  Neisser  and  Wechsberg  49  produced  "leuko- 
toxiri'  by  injecting  leukocytes;  Delezenne  50  produced  "neurotoxin" 
and  "hepatotoxin"  and  Sunnont,51  pancreas  cytotoxin.  Subsequent 
years  have  added  to  these  " gastro-toxin'  (Bolton),52  thymotoxin 
(Slatineau),53  adrenal  cytotoxin  (Gilder sleeve),54  placentar  cyto- 
toxin (Frank),55  corpus  luteum  cytotoxin  (Miller),56  and  a  number 
of  others.  In  fact,  as  Roessle  57  puts  it,  in  a  review  of  the  literature, 
there  is  no  organ  in  the  body  for  which  it  has  not  been  claimed  that 
specific  cytotoxins  can  be  formed  by  the  injection  of  homologous 
macerated  tissues. 

Recent  critical  study  of  these  organ-cytotoxins  has  revealed,  how- 
ever, that  the  specificity  of  a  serum  produced  with  the  tissues  of  one 
organ  is  not  strictly  limited  to  this  organ  alone,  and  that  the  serum 
may  injure  other  organs  as  well.  It  is  true,  indeed,  that  there  are 
certain  cells  and  tissues  in  the  body  such  as  the  spermatozoa,  the 
tissues  of  the  testicles,  the  ovary,  the  lens  of  the  eye,  and,  possibly, 

46  Metchnikoff.    Ann.  Past.,  Vol.  13,  1899. 

47  Landsteiner.     CentralU.  f.  Bakt.,  Vol.  25,  p.  549,  1899. 

48  Von  Dungern.    Munch,  med.  Woch.,  p.  1228,  1899. 

49  Neisser  and  Wechsberg.     Zeitschr.  f.  Hyg.,  Vol.  36,  1901. 

50  Delezenne.    Ann.  Past.,  1900;  Compt.  rend,  de  Vacad.  des  sc.,  1900. 

51  Surmo  nt.     Compt.  rend,  de  la  soc.  de  biol,  1901. 

52  Bolton.     Lancet,  1908. 

53  Slatineau.     Cited  from  Roessle,  loc.  cit. 

54  Gildersleeve.     Cited  after  Roessle. 

55  Frank.     Jour.  Exp.  Med.,  1907. 

56  Miller.     CentralU.  f.  Bakt.,  47,  1908. 
"Roessle.     "Lubarsch  und  Ostertag,"  Vol.  13,  1909. 


PHENOMENA    FOLLOWING    IMMUNIZATION         93 

the  placenta  which  have  chemical  characteristics  so  well  defined  and 
individual  that  the  cytotoxic  sera  induced  by  them  have  definite 
organ  specificity.  The  same  to  a  more  limited  extent  seems  true 
of  kidney  substance  (Pearce).  In  most  cases,  however,  in  which 
originally  a  specific  cytotoxin  was  claimed,  it  has  been  possible  to 
show  subsequently  that  the  apparently  selective  injury  was  due  not  to 
organ  specificity  alone  but  to  the  fact  that  the  injection  of  tissue- 
macerates,  even  when  sufficiently  freed  from  blood,  induced  the 
formation  of  considerable  amounts  of  hemagglutinins  and  hemol- 
ysins. 

Pearce  58  expresses  it  as  follows :  "  ...  it  is  evident  that  the 
cells  or  the  various  organs  of  the  body,  while  differing  in  morphology 
and  function,  have  certain  (receptor)  characteristics  in  common, 
and  that  one  type  of  cell  may  therefore  produce  antibodies  affecting 
several  cells  of  differing  morphology,  but  with  like  (receptor) 
groups.  This  is  shown  by  the  sera  prepared  from  washed  liver, 
Ividney,  pancreas,  and  adrenal,  all  of  which  may  agglutinate  and 
hemolyze  red  blood  cells  and  may  cause  degenerative  changes  also 
in  the  liver  and  the  kidneys.  Some  of  these  cytotoxic  sera  have  no 
effect  upon  organs  for  which  they  are  supposed  to  have  a  morpho- 
logical affinity,  but  exert  a  powerful  lytic  influence  upon  other  cells. 
Aside  from  nephrotoxin,  which  has  a  distinct  injurious  action  upon 
renal  epithelium,  the  various  cytotoxins  studied  (kidney,  liver,  pan- 
creas, and  adrenal)  have  no  specific  action  in  the  morphological 


sense." 


This  opinion  seems  to  be  in  harmony  with  that  of  most  observers 
who  have  studied  the  problem  recently,  at  least  as  regards  most  of 
the  organ  cytotoxins.  Much  of  the  promised  light  upon  pathological 
processes — looked  for  when  cytotoxins  were  first  studied,  has  faded, 
moreover,  since  it  has  been  found  that  cytotoxins  cannot  be  produced 
by  injection  into  an  animal  of  cells,  tissues,  or  fluids  from  its  own 
body.  "Autocytotoxins"  in  general  cannot  be  produced,  a  question 
discussed  at  greater  length  in  the  chapter  on  lysis,  in  connection 
with  Ehrlich's  work  on  the  isolysins. 

The  work  outlined  in  the  preceding  paragraphs  had  thus  ex- 
tended the  principles  of  antitoxin  and  lysin  production  beyond  the 
scope  of  pure  bacteriology,  and  had  shown  them  to  possess  the  sig- 
nificance of  general  biological  laws.  Similar  generalization  was  soon 
attained  in  the  case  of  the  agglutinins  and  in  that  of  the  precipitins. 
In  the  former,  the  nature  of  the  reaction  limited  it  to  observations 
upon  cells  in  suspension,  and,  in  connection  with  the  earlier  experi- 
ments upon  hemolysis  it  was  soon  discovered  that  the  erythrocytes 
were  often  clumped  before  lysis  could  take  place,  when  brought  to- 
gether with  a  hemolytic  serum  of  moderate  or  feeble  potency,  or 
when  solution,  for  other  reasons,  was  delayed. 

58  Pearce.    Jour,  of  Med.  Res.,  N.  Sv  Vol.  7,  1914,  p.  13. 


94  INFECTION    AND    RESISTANCE 

The  first  observations  on  the  general  significance  of  the  precip- 
itin  reaction  we  owe  to  Tschistovitch  59'and  to  Bordet.60  Tschisto- 
vitch was  studying  the  toxic  action  of  eel  serum  upon  rabbits.  This 
serum,  as  Kossel  61  had  shown,  is  toxic  for  rabbits  and  possesses  the 
property  of  causing  hemolysis  of  rabbit  erythrocytes.  Its  similarity 
to  ricin,  in  this  respect,  stimulated  attempts  to  produce  an  antitoxic 
substance  against  eel  serum,  even  as  Ehrlich  had  produced  an  an- 
tiricin.  In  the  course  of  such  experiments  Tschistovitch  observed 
that,  when  eel  serum  was  mixed  with  the  serum  of  a  rabbit  which 
had  received  several  injections  of  this  substance,  the  mixture  became 
rapidly  opalescent  and  soon  a  flocculent  precipitate  was  formed. 
Coincident  with  this  discovery  Bordet  made  a  similar  observation. 
He  had  injected  chicken  blood  into  rabbits  in  the  course  of  experi- 
ments upon  hemagglutination.  He  found  that  the  serum  of  the  rab- 
bits so  treated  acquired  the  property  not  only  of  producing  hemolysis 
and  hemagglutination  of  chicken  cells,  but  also  of  giving  a  precipi- 
tate if  mixed  with  chicken  serum.62  Soon  after  this  precipitins  were 
produced  by  injecting  rabbits  with  milk  (Bordet),  egg  albumen 
(Ehrlich,  Uhlenhuth),  and  many  other  substances,  and  the  speci- 
ficity of  such  reactions  was  demonstrated  by  Fish,63  Wassermann  and 
Schiitze,64  Uhlenhuth,  and  many  others. 

It  is  apparent  from  the  preceding  paragraphs  that  the  discovery 
of  specific  antitoxins  merely  constituted  the  first  step  in  the  formu- 
lation of  a  fundamentally  important  biological  law.  There  is,  then, 
a  large  group  of  substances  of  animal  and  vegetable  origin  which 
call  forth  the  formation  of  specific  reacting  bodies  when  injected 
into  animals.  In  order  to  elicit  this  response  it  is  necessary  that 
these  substances  shall  penetrate  to  the  physiological  interior  of  the 
body  in  a  relatively  unchanged  condition.  For  this  reason  any  form 
of  injection,  subcutaneous,  intravenous,  or  into  a  serous  cavity,  is 
followed,  with  regularity,  by  antibody  formation,  whereas  feeding 
or  other  means  of  intraintestinal  administration  is  negative  in  result, 
unless  abnormal  conditions  prevail  which  permit  entrance  into  the 
blood  before  the  digestive  enzymes  have  decomposed  the  ingested 
materials. 

The  substances  with  which  antibody-formation  may  be  induced 
are  collectively  spoken  of  as  "antigens," 

59  Tschistovitch.    Cited  by  Bordet,  loc.  cit.,  and  also  Ann.  Past.,  13,  1899. 

60  Bordet.    Ann.  Past.,  Vol.  13,  1899. 

61  Kossel.    Berl  klin.  Wocn.,  No.  7,  1898. 

62  This,  we  know  now,  was  due  to  the  fact  that  the  blood  cells  injected 
were  not  washed  free  of  chicken  serum.     Thus  chicken  serum  precipitin  was 
formed  as  well  as  were  hemagglutinin  and  hemolysin. 

63  Fish.     St.  Louis  Med.  Cour.,  1900.     Cited  from  Uhlenhuth. 

64  Wassermann    and    Schutze.      Deutsche    med.    Woch.,    No.    30,    1900. 
Vereinsbeilage. 


PHENOMENA  FOLLOWING  IMMUNIZATION  95 

Antigens  are  all  substances  which,  injected  into  the  animal  body,  induce 
specific  antibody  formation.  They  form  a  large  group  in  nature  and  are 
chemically  proteins;  indeed,  we  may  say  that  all  known  proteins  may  act  as 
antigens.  Whether  or  not  this  term  may  also  include  lipoid-protein  com- 
binations, lipoids  or  the  higher  protein  derivatives  is  as  yet  uncertain  and 
need  not  in  the  present  connection  concern  us. 

We  may  divide  antigenic  substances  into  two  main  classes.1  One  of  these 
comprises  all  of  those  substances  of  bacterial,  animal  or  vegetable  origin 
which,  injected  into  the  animal  body,  give  rise  to  specific  neutralizing  or 
antitoxic  properties  in  the  blood  of  the  injected  animal.  These  are  the  bac- 
terial exotoxins,  the  snake  venoms,  some  powerful  vegetable  poisons  and  proteo- 
lytic  and  other  enzymes  of  animals  and  plants.  They  are  all  substances  which 
are  powerfully  active — some  of  them  strongly  toxic  to  the  living  animal, 
others  true  enzymes  or  ferments.  Indeed  all  of  them  possess  properties  which 
at  least  suggest  our  placing  them  into  the  class  of  enzymes  in  general.  The 
number  of  such  substances  known  is  limited.  The  reaction  they  call  forth  in 
the  animal  body  seems  aimed  directly  at  the  specific  neutralization  of  their 
respective  activities,  and  is  so  unique  and  different  from  that  induced  by  other 
antigens  that  it  would  be  convenient  had  we  another  term  like  "antitoxinogen" 
to  set  them  apart  by  themselves. 

The  other  class  of  antigens  comprises  all  proteins  which  are  inactive, 
showing  in  themselves  neither  toxic  nor  enzyme-like  properties.  Introduced 
into  the  animal  body  parenterally,  they  call  forth  a  response  of  a  nature 
entirely  unlike  that  of  the  antitoxins,  and  which  as  far  as  we  can  fathom 
its  purpose  seems  aimed  merely  at  the  assimilation  or  the  removal  of  the 
infected  substance.  For  the  cells  of  the  animal  cannot  utilize  the  foreign 
protein  as  such,  and  thus  it  is  only  foreign  proteins  injected  into  an  animal 
that  act  antigenically,  and  no  antibodies  are  formed  when  homologous  material 
is  injected. 

This  large  group,  composed  of  all  formed  and  unformed  substances  in 
nature  in  which  a  protein  structure  is  involved,  does  not  induce  the  formation 
of  anything  like  the  neutralizing  antitoxins  spoken  of  above.  The  antibodies 
appearing  in  animals  treated  with  such  substances  have  been  spoken  of  as 
cytolysins  or  cytotoxins — precipitins — and  in  the  case  of  formed  antigens 
like  bacteria  or  blood  cells — agglutinins  and  opsonins.  As  we  shall  see  in 
another  section,  it  is  our  opinion  that  all  these  various  antibodies  are  iden- 
tical in  structure  and  significance. 

We  must  not  forget,  however,  that  the  observation  of  antibodies  in  the 
circulating  blood  is  but  one  of  the  changes  that  have  taken  place  in  the 
treated  animal.  Much  has  been  made  of  this  phase  of  the  problem  because 
serum  antibodies  are  readily  studied  in  vitro;  but  their  origin  of  course  must 
be  sought  in  the  body  cell,  in  which  the  original  and  most  profound  changes 
must  necessarily  have  taken  place  during  such  treatment,  changes  the  nature 
of  which  are  to  a  large  extent  still  a  mystery,  but  on  which  ultimately  de- 
pend the  important  physiological  difference  between  treated  and  untreated 
animals.  For  such  changes — whether  we  refer  to  those  immediately  under 
discussion,  namely,  those  of  allergy  or  anaphylaxis,  or  whether  we  think  of 
the  so-called  immunity  remaining  after  attacks  of  many  diseases — remain 
present  long  after  the  circulating  antibodies  have  disappeared  and  must  there- 
fore be  regarded  as  associated  with  profound  alterations  in  the  ultimate  tissue 
unit,  the  body  cell. 

1  See  also  Zinsser,  "The  More  Recent  Developments  in  the  Study  of  Ana- 
phylactic  Phenomena."  Arch,  of  Int.  Med.,  Vol.  XVI,  1915,  pp.  223-256. 
Harvey  Lecture. 


96  INFECTION  AND  RESISTANCE 

Since  the  phenomenon  of  antibody  formation  is  not  at  all  limited 
to  bacteria  or  bacterial  derivatives,  it  cannot  be  looked  upon  merely  as 
a  mechanism  existing  for  the  primary  purpose  of  protecting  the  body 
against  infectious  disease.  This  latter  function  is  important,  indeed, 
but  is  probably  incidental  to  the  broader  significance  of  the  processes. 

In  the  course  of  normal  existence  substances  which  are  not  di- 
rectly assimilable  as  such — foreign  proteins,  for  instance — do  not 
penetrate  directly  into  the  blood  and  tissues.  Taken  into  the  ali- 
mentary canal,  they  are  first  hydrolized  into  peptons,  albumoses, 
polypeptids,  and  probably  amino-acids  before  absorption,  to  be  recon- 
structed from  these  cleavage  products  ("Bausteine"  is  Abderhalden's 
expression  for  the  amino-acids)  into  protein  biologically  identical 
with  that  of  the  tissues.  Digestive  and  other  accidents,  however,  on 
numerous  occasions  during  life  permit  the  direct  entrance  of  these 
materials  unchanged  or  insufficiently  changed  into  the  circulation. 
It  is  probably  by  the  action  of  digestive  powers  of  the  serum — or,  in 
the  case  of  'the  entrance  of  undissolved  foreign  particles,  by  the 
activity  of  the  phagocytic  cells — that  such  substances  are  then  dis- 
posed ot  and  assimilated.  For  each  particular  variety  of  substance 
(antigen)  a  specific  mechanism  is  called  into  play,  and  when  this 
mechanism  is  repeatedly  called  upon — as  in  successive  injections  of 
foreign  proteins — this  mechanism,  whatever  it  may  consist  of,  is  en- 
hanced in  efficiency — i.  e.,  increased  in  quantity.  How  this  increase 
of  specific  antibodies  is  theoretically  conceived  we  will  discuss  later 
in  connection  with  Ehrlich's  side-chain  theory. 

The  phenomena  of  antibody  formation  against  bacteria  on  this 
basis  may  be  taken  to  constitute,  then,  a  mechanism  for  the  digestion 
and  disposal  of  a  foreign  protein  which  has  penetrated  into  the  tis- 
sues and,  because  of  its  living  state,  increases  within  the  body  by 
multiplication,  furnishing  progressive  stimulation  to  the  antibody- 
producing  function.  Infectious  disease,  therefore,  from  this  point 
of  view  may  be  looked  upon  as  an  invasion  of  the  body  by  a  living 
foreign  protein  which  must  be  assimilated  and  disposed  of;  which, 
in  some  cases,  has  a  primary  toxicity  per  se ;  and  which  is  variously 
distributed  among  the  organs  and  tissues  according  to  the  biological 
peculiarities  of  the  particular  microorganism  in  question.  This 
general  conception  will  become  more  clear  as  we  analyze  the  phe- 
nomena associated  with  the  individual  antibodies.  It  is,  of  course, 
quite  plausible  as  far  as  it  refers  to  the  phagocytic  functions,  or  even 
bacteriolytic  and  cytolytic  phenomena.  It  has  been  less  clear  in 
connection  with  the  agglutinins  and  precipitins  in  which  a  direct  de- 
fensive or  bacteria-destroying  value  is  not  apparent.  However,  in 
our  discussions  of  these  phenomena  we  will  have  occasion  to  point,  out 
many  reasons  for  assuming  that,  even  in  these  phenomena,  there  are 
features  which  fall  into  direct  correlation  with  the  views  we  have 
just  expressed. 


PHENOMENA  FOLLOWING  IMMUNIZATION  97 

The  substances  which  possess  antigenic  properties — that  is,  which 
give  rise  to  antibody  production — with  the  exception  of  a  few  isolated 
and  contested  cases,  are  all  of  them  protein  in  nature.  Well-trained 
chemists  have  exerted  themselves  to  purify  antigenic  substances, 
in  attempts  to  determine  the  particular  fractions  of  the  complex 
protein  molecule  upon  which  the  antigenic  properties  depend.  In 
the  course  of  such  work  a  number  of  men  claim  to  have  obtained  a 
truly  antigenic  substance  which  no  longer  gave  protein  reactions. 
The  instance  most  frequently  cited  is  Jacoby's  65  announcement  of 
a  protein-free  ricin.  Jacoby  worked  with  an  apparently  very  impure 
"Ausgangsmaterial"  consisting  of  commercial  ricin,  which  he  di- 
gested for  five  weeks  in  trypsin  solution.  At  the  end  of  this  time  he 
obtained  a  ricin  which  still  possessed  the  properties  of  the  original 
castor-bean  extract,  but  no  longer  gave  protein  reactions.  His  "puri- 
fied ricin,"  however,  was  quickly  destroyed  by  further  trypsin  diges- 
tion, and  more  recent  work  by  Osborne,  Mendel,  and  Harris 66 
appears  to  have  fully  refuted  Jacoby's  results.  They  found  the 
purified  ricin  identical  with  the  coagulable  albumin  of  the  castor 
bean,  and  found  that  tryptic  digestion  destroys  the  characteristic 
ricin  properties. 

Less  easily  refuted  have  been  the  careful  experiments  of  Ford  6T 
upon  the  active  principle  of  a  mushroom  (Amanita  phalloides)  and 
upon  that  of  the  poison-ivy  plant — (Rhus  toxicodendron) .  These 
substances,  he  claims,  are  non-protein.  In  the  case  of  Amanita 
phalloides  Abel  and  Ford 68  have  shown  it  to  be  a  glucosid,  and 
similar  structure  has  been  claimed  for  Rhus  by  Syme.69  Yet  with 
both  of  these  substances  Ford  has  succeeded  in  producing  specific 
antitoxins.  Rabe  70  has  questioned  these  with  Amanita  phalloides. 
He  believes  that  the  poison  with  which  Ford  worked  is  not  a  glucosid, 
but  is  of  protein  nature.  In  the  case  of  Rhus,  however,  Ford's  con- 
clusions have  not,  to  our  knowledge,  been  challenged. 

With  these  and  a  few  other  less  important  exceptions,  however, 
observers  have  uniformly  concluded  that  antigenic  property  and  pro- 
tein structure  are  inseparably  associated.  All  procedures  by  which 
proteins  have  been  hydrolized  into  their  simpler  fractions,  chemical 
splitting,  tryptic  or  peptic  digestion  have  in  every  case  resulted  in  a 
simultaneous  loss  of  protein  reaction  and  antigenic  property.. 

Many  attempts  have  also  been  made  to  show  a  relation  between 
antigenic  properties  and  the  lipoid  constituents  of  cells.  These  en- 
deavors were  obviously  stimulated  by  the  observation  that  many  li- 

65  Jacoby.     Arch.  f.  exp.  Path.  u.  Pharm.,  Vol.  46,  1901. 

66  Osborne,  Mendel,  and  Harris.    Am.  Jour,  of  PhysioL,  1905,  Vol.  14. 

67  Ford.     Jour,  of  Inf.  Dis.,  Vol.  3,  1906 ;  Vol.  4,  1907. 

68  Abel  and  Ford.    Jour.  Biol.  Chem.,  1907, 

69  Syme.     Johns  Hopkins  Thesis,  1906. 

70  Rabe.     Zeitschr.  f.  exp.  Path.  u.  Therap.,  Vol.  9,  1911. 


98  INFECTION  AND  RESISTANCE 

poids  are  capable  of  binding  antibodies  in  miro,  and  that,  in  nervous 
tissues,  toxin  fixation  was  in  some  way  related  to  the  richness  in 
lipoids  of  these  structures.  Bang  and  Forsmann 71  accordingly 
treated  animals  with  ether  extracts  of  red  blood  cells — claiming  that 
this  resulted  in  the  production  of  hemolysins.  And  these  results 
have  been  confirmed  by  Landsteiner  and  Dautwitz.72  The  latter, 
however,  suggest  that  the  hemolysin  production  may  have  been  in- 
duced, not  by  the  lipoidal  substances  in  solution,  but  by  other  anti- 
genie  substances  which  had  gone  into  colloidal  suspension  in  the 
ether  extracts.  Much  similar  research  on  the  antigenic  nature  of 
lipoids  has  been  done,  but,  after  reviewing  this  very  thoroughly, 
Landsteiner  comes  to  the  conclusion  that  no  definite  proof  of  the  anti- 
genic nature  of  any  pure  lipoid  has  so  far  been  presented.  The  prob- 
lem is  experimentally  complicated  by  the  fact  that,  as  Landsteiner  7a 
suggests,  the  antigen  may  often  be  present  as  a  lipoid-protein  com- 
bination, and  as  such  go  into  solution  or  fine  emulsion  in  the  organic 
solvents ;  also  the  lipoids  possess  the  curious  property  of  altering  the 
solubilities  of  proteins  and  other  substances  by  their  presence. 

Summarizing  our  present  knowledge  of  the  chemical  nature  of 
antigens,  then,  we  must  conclude  that,  with  the  exception  of  Ford's 
glucosids,  no  protein-free  antigens  have  been  thus  far  demonstrated. 

In  the  light  of  this  fact  it  is  all  the  more  remarkable  that  antigen- 
antibody  reactions  are  specific.  For  we  possess  no  chemical  methods 
by  which  one  variety  of  protein  can  be  distinguished  from  another. 
And  yet  the  serum  antibodies  produced  with  each  species  of  bacteria 
react  with  this  species  only — and  the  hemolysins,  agglutinins,  or 
precipitins  produced  by  the  injection  of  bacterial,  cellular,  or  serum 
proteins  react  respectively  only  with  the  particular  variety  employed 
in  their  production.  This  indicates  that  each  of  these  antigens — of 
almost  unlimited  number — must  possess  a  chemical  structure  indi- 
vidually characteristic  and  different  from  all  the  others.  It  is  by 
means  of  the  biological  reactions,  indeed,  that  we  can  detect  protein 
in  dilutions  far  beyond  the  reaction-sensitiveness  of  chemical  tests 
and  can  distinguish  between  varieties  of  protein  when  the  chemical 
methods  will  indicate  only  protein  in  general.  Our  knowledge  of 
the  chemical  constitution  of  protein  has  not  yet  advanced  to  a  point 
at  which  specificity  can  be  based  upon  definite  variations  of  chemical 
structure,  and  the  complexity  of  the  problem  is  such  that  it  does  not 
seem  likely  that  we  can  hope  in  the  near  future  to  attain  such  knowl- 
edge. We  can  merely  accept  it  as  a  fact  that  the  antibody  produced 
with  one  protein  differs  materially  from  that  produced  with  another, 

71  Bang  and  Forsmann.    Hofm.  Beitr.,  1906;  Centralbl  f.  Bakt.,  40,  1906. 

72  Landsteiner  and  Dautwitz.     Hofm.  Beitr.,  9,  1907. 

73  Landsteiner.     "Wirken  Lipoide  als  Antigene?"     Weichardt's  Jahresbe- 
richt,  Vol.  6,  1910. 


PHENOMENA  FOLLOWING  IMMUNIZATION  99 

and  that  this  is  a  definite  indication  that  the  antigen  in  one  case  must 
be  chemically  different  from  that  in  another. 

The  range  of  such  variations  is  apparently  enormous.  For  each 
variety  of  bacteria  or  plant,  each  species  of  animal,  and  to  a  certain 
extent  each  individual  of  the  species,  possesses  certain  special  anti- 
genie  characteristics  peculiar  to  itself.  In  general  there  is  an  under- 
lying antigenic  similarity  which  is  peculiar  to  the  species.  This  is 
true  of  bacteria  and,  in  the  case  of  animal  and  vegetable  proteins,  an 
antibody  produced  with  material  from  an  individual  of  a  certain 
species  will  react  with  the  protein  derived  from  this  species  in  gen- 
eral. However,  that  there  are  also  antigenic  differences  between  in- 
dividuals within  the  same  species  is  indicated  by  Ehrlich's  experi- 
ments on  the  antibodies  produced  by  injecting  the  blood  cells  of  one 
goat  into  another.  And  we  have  further  indicated  that  within  the 
same  animal  different  organs  may  possess  individual  antigenic  char- 
acteristics. Added  to  this  we  know  that  certain  special  organs  like 
the  testicle,  the  lens,  and  some  others  contain  antigens  which  are 
peculiar  to  this  variety  of  organ,  irrespective  of  species — a  condition 
spoken  of  as  "organ  specificity."  Thus  an  antibody  produced  by 
injections  of  the  testicular  substance  of  one  animal  will  react  with 
testicular  protein  from  many  different  species — the  specificity  here 
depending  upon  the  organ  and  not  upon  the  zoological  relationship. 

It  is  clear,  therefore,  that  there  are  more  different  varieties  of 
protein,  biologically  distinguishable,  than  there  are  species  of  living 
beings  in  nature.  As  Abderhalden74  has  recently  pointed  out,  this 
is  a  conception  which  it  is  a  little  difficult  to  grasp  chemically,  since 
in  breaking  up  different  proteins  into  their  "building  stones"  (Bau- 
steine)  we  encounter  again  and  again  the  same  20  amino-acids.  By 
a  simple  arithmetical  consideration,  however,  he  shows  that  merely 
by  combining  these  twenty  amino-acids  in  different  groupings  an 
.enormous  number  of  isomeric  but  varying  compounds  can  be  formed 
— even  without  assuming  the  additional  possibility  of  quantitative 
variations.  He  reasons  that  3  "Bausteine" — A,  B,  and  C — could 
form  6  different  structures,  A  B  C,  A  C  B,  B  C  A,  B  A  C,  C  A  B, 
C  B  A.  Similarly  4  could  form  26,  and  finally  20  could  form  2,  432, 
902,  008,  176,  640,  000  different  compounds.75 

The  analogy  between  the  active  immunization  of  animals  with 
the  various  antigens  and  certain  chemically  well-defined  poisons, 
alkaloids,  etc.,  is  so  obvious  that  it  has  led  to  much  speculation  as  to 
a  possible  similarity  in  the  physiological  mechanisms  of  the  two  phe- 
nomena. As  a  matter  of  fact  the  acquired  tolerance  for  such  sub- 
stances as  morphin,  atropin,  and  other  alkaloids  is  not  really  anal- 
ogous to  the  physiological  reactions  which  follow  the  treatment  of  an- 

74  Abderhalden.    Munch,  med.  Woch.,  No.  43,  1913. 

75  "We  have  not  repeated  the  arithmetical  labor  and  take  Abderhalden's 
word  for  it. 


100  INFECTION  AND  RESISTANCE 

imals  with  bacterial  and  other  proteins,  for  whatever  toxic  properties 
there  are  in  the  latter  are,  as  we  shall  see  later,  rather  the  results  of 
the  interaction  of  these  injected  substances  and  the  reaction  products 
supplied  by  the  cells  and  fluids  of  the  body.  It  is  at  least  probable 
in  the  light  of  our  modern  conception  that  such  protein  antigens  are 
not  toxic  per  se,  in  the  native  state.  This,  however,  will  receive 
detailed  consideration  in  succeeding  sections.  The  analogy  of  drug 
tolerance,  however,  to  the  acquired  immunity  against  true  bacterial 
toxins  and  vegetable  poisons  like  ricin,  crotin,  and  others  is  a  strik- 
ing one,  since  in  both  classes  of  poisons  there  is  a  gradually  devel- 
oped tolerance  for  substances  toxic  in  the  native  state  and  often  very 
similar  in  physiological  effects  (strychnin  and  tetanus  toxin,  etc.). 
In  the  case  of  the  toxins,  however,  there  is  a  development  of  im- 
munity by  actual  neutralization  of  the  poisonous  principle  brought 
about  by  a  specific  antibody,  which  circulates  in  the  blood  of  im- 
munized animals  and  man — the  process  following,  within  certain 
limits,  the  law  of  multiple  proportions.  In  the  case  of  morphin 
and  other  alkaloids  no  such  neutralizing  antibodies  have  as  yet  been 
demonstrated.76  Whereas  toxin  immunity  is  passively  transferable 
from  one  animal  to  another  with  the  blood  serum,  and,  in  vitro,  the 
mixture  of  the  toxin  with  the  immune  serum  brings  about  a  neutrali- 
zation of  the  poison,  no  such  phenomena  have  been  observed,  as  a 
general  rule,  in  the  case  of  the  alkaloids.  We  say  "as  a  general 
rule"  since  an  exception  is  recorded  in  the  observations  of  Fleisch- 
mann,77  who  claims  to  have  found  antagonistic  action  to  atropin  in 
the  blood  of  normal  rabbits,  this  power  being  absent  from  the  blood 
of  rabbits  that  had  thyroid  hypertrophies  and  were,  in  consequence, 
atropin-susceptible.  Other  observations  of  a  similar  significance 
have  been  made  by  Physalix  and  Contejean  78  on  curare,  but  have 
not  been  confirmed,  and  the  investigations  of  all  other  workers  on 
this  subject  have  had  negative  results.  It  seems  from  available  evi- 
dence that  tolerance  (immunity)  against  drugs  is  due  to  cellular 
rather  than  to  serum  antagonism. 

THE    ORIGIN    OF  ANTIBODIES 

The  tissue  cell,  as  the  ultimate  functional  unit,  must,  of  course, 
})G  looked  upon  as  the  source  from  which  originate  the  various  pro- 
tective constituents  of  normal  and  immune  sera ;  and,  though  per- 
haps unrecognizable  by  the  coarse  tests  of  morphological  investiga- 
tions, it  is  in  the  cells  that  changes  must  take  place  primarily  when 

76  Hans  Meyer  and  Gottlieb.    "Exp.  Pharm.,"  2d  Ed.,  Neban  &  Schwart- 
zenberg,  Berlin,  1911,  p.  517. 

77  Fleischmann.     Archiv  f.   exp.  Path.  u.  Pharm.,  62,  1910,   cited  from 
Meyer  and  Gottlieb,  loc.  cit. 

78  Physalix  and  Contejean.     Cited  from  Meyer  and  Gottlieb. 


PHENOMENA  FOLLOWING 


the  animal  body  is  subjected  to  any  one  of  the  processes  spoken  of  as 
immunization.  The  exact  location  of  the  antibody-forming  cells  and 
tissues,  in  spite  of  much  investigation,  is  not  at  all  clear,  though 
many  data  seem  to  point  to  the  lymphatic  organs,  the  spleen,  and  the 
bone  marrow  as  particularly  concerned  with  this  process. 

Thus  Pfeiffer  and  Marx  '9  exsanguinated  animals  five  days  after 
injections  of  dead  cholera  spirilla  and  found  that  at  this  time  bac- 
teriolytic  antibodies  were  more  concentrated  in  the  spleen  than  in 
the  blood  serum  itself.  Wassermann's  80  analogous  experiments  with 
typhoid  bacilli  seemed  to  show  a  higher  antibody  content  in  spleen, 
bone  marrow,  thymus,  and  lymph  nodes  than  was  present  in  the 
blood  at  an  early  period  of  immunization.  Although  these  investiga- 
tions, as  well  as  many  others  of  Castellani,81  seem,  therefore,  to  indi- 
cate a  particular  association  of  the  special  lymphatic  organs  with 
antibody  formation,82  extirpation  of  the  spleen  83  before  immuniza- 
tion has  not  prevented  animals  from  responding  to  injections  of  bac- 
teria and  red  blood  cells  with  sharp  antibody  production.  The  ex- 
periments of  Deutsch,84  in  which  reduction  of  antibody  formation 
resulted  in  animals  in  which  splenectomy  was  practiced  three  or  four 
days  after  immunization  was  begun,  can  hardly  be  accepted  as  a  con- 
clusion, in  the  writer's  opinion  at  least,  since  any  severe  operation  or 
interference  with  the  normal  functions  of  an  animal  during  the 
severe  physiological  strain  of  active  immunization  would  naturally 
lead  to  a  less  perfect  response.  That  the  resistance  of  animals  and 
man  to  infection  with  bacteria  is  not  noticeably  diminished  by  sple- 
nectomy, moreover,  has  been  variously  shown.  In  unpublished  ex- 
periments by  the  writer  splenectomized  guinea  pigs  showed  no  differ- 
ence from  normal  animals  in  regard  to  their  susceptibility  to  tuber- 
culosis. And  though  these  and  similar  experiments  of  other  workers 
with  various  bacteria  are  not  entirely  devoid  of  interest,  their  nega- 
tive results  as  a  matter  of  fact  have  no  great  significance,  since  our 
knowledge  concerning  the  true  function  of  the  spleen  is  very  incom- 
plete, and  it  is  not  impossible  that  on  removal  of  this  organ  other 
elements  of  the  lymphatic  system  may  take  over  its  function  in  part 
or  as  a  whole. 

Removal  of  the  spleen  has  not  been  an  extremely  unusual  pro- 
cedure in  surgery,  and  there  -is  no  evidence  to  show  that  patients  so 
treated  have  been  abnormally  susceptible  to  infection  thereafter. 

Yet,  as  we  have  seen,  there  seems  to  be  an  early  concentration  of 
antibodies  in  the  lymphatic  organs  in  the  course  of  immunization, 
and  it  may  well  be  that  an  association  between  the  process  and  these 

79  Pfeiffer  and  Marx.     Zeitschr.  f.  Hyg.,  Vol.  27,  1898. 

80  Wassermann.     Berl.  klin.  Woch.,  p.  209,  1898. 

81  Castellani.     Zeitschr.  f.  Hyg.,  Vol.  37,  1901. 

82  Pfeiffer  and  Marx.    Loc.  cit. 

83  I.  Levin.     Jour.  Med.  Bes.,  Vol.  8,  1902. 

84  Deutsch.    Ann.  de  I'Inst.  Past.,  Vol.  13,  1899. 


KHJtM        mFECTION  AND  RESISTANCE 

tissues  exists  which  cannot  be  experimentally  demonstrated  with 
absolute  certainty. 

It  is  no  less  likely,  however,  that  similar  functions  are  exerted 
by  the  cells  of  other  organs.  In  fact,  it  is  more  than  probable  that 
antibodies  may  be  formed  anywhere  in  the  body — and  that  the  local- 
ity of  their  production  is  largely  dependent  upon  the  locality  in  which 
the  antigen  is  concentrated.  Wassermann  and  Citron  85  demonstrated 
this  by  injecting  typhoid  bacilli  into  rabbits  intraperitoneally,  in- 
travenously, and  intrapleurally,  and  nine  days  afterward  determining 
the  comparative  bactericidal  strength  of  blood  serum  and  of  aleuronat 
exudates  of  pleura  and  peritoneum  in  each  of  the  three  animals. 
Their  results  showed  that  the  bactericidal  titre  of  the  intravenously 
inoculated  animal  was  highest  in  the  blood  serum,  while  that  of  the 
intraperitoneally  and  intrapleurally  inoculated  animals  was  highest 
in  peritoneal  and  pleural  exudates  respectively.  Such  experiments 
point  to  the  possibility  of  a  "local"  immunity,  that  is,  a  production 
of  antibodies  directly  by  the  cells  with  which  the  antigen  comes  into 
contact  in  the  most  concentrated  and  direct  manner.  And,  indeed, 
another  isolated  experiment  of  the  same  authors,  alone  successful  of 
a  series  of  similar  attempts,  would  point  in  the  same  direction. 
Typhoid  bacilli  were  injected  subcutaneously  into  the  ear  of  a  rabbit 
and  the  ear  immediately  ligated  at  its  base  and  kept  so  for  several 
hours.  After  nine  days  the  bactericidal  titre  of  the  blood  serum  was 
determined  and  the  ear  amputated.  An  immediate  and  rapid  drop 
of  antibody  contents  occurred  after  the  amputation — indicating  that 
the  chief  source  of  antibody  function  had  been  removed.  More  strik- 
ing examples  of  the  same  thing  are  to  be  seen  in  the  experiments  of 
Romer,86  who  instilled  abrin  into  a  rabbit's  eye  and  found  that  the 
retina  of  the  eye  developed  an  antitoxic  power  against  abrin  which 
protected  mice  against  many  times  the  fatal  dose,  while  that  of  the 
other  eye  remained  practically  inactive. 

From  these  facts,  as  well  as  from  other  observations,  it  is  at  least 
reasonable  to  believe  that  antibody  formation  is  by  no  means  a  func- 
tion of  special  organs  and  that  many  cells  throughout  the  body  may 
take  part  in  the  process.  It  is  of  especial  importance  to  consider  this 
in  connection  with  the  possible  effects  of  the  treatment  of  infections 
by  means  of  bacterial  vaccines.  If  the  focus  of  the  infection  can 
possibly  become  also  a  local  source  of  antibody  production  then  such 
treatment  may  well  seem  rationally  founded,  even  in  generalized 
acute  infections  in  which  no  logical  basis  for  such  treatment  would 
exist,  were  the  production  of  antibodies  a  task  for  specialized  organs 
like  spleen  and  bone  marrow  only.  The  therapeutic  phases  of  this 
problem  are  more  extensively  considered  in  a  later  chapter. 

It  is  in  this  fact  also  that  we  must  seek  the  explanation  of  the 

85  Wassermann  and  Citron.    Zeitschr.  f.  Hyg.,  Vol.  50,  1905. 

86  Romer.    Arch.  f.  OphthaL,  52,  1901. 


PHENOMENA  FOLLOWING  IMMUNIZATION         103 

apparent  local  immunity  which  occurs  in  certain  infections  of  the 
skin.  Thus  it  frequently  happens  that  successive  crops  of  boils  may 
afflict  different  parts  of  a  patient's  skin — new  ones  arising  as  old 
ones  heal,  showing  that  the  process  of  the  limitation  and  healing  of 
the  infected  foci  is  not  due  to  any  increase  of  generalized  resistance, 
but  rather  to  local  causes.  In  the  same  way,  in  erysipelas,  the  process 
extends  along  the  edges  while  the  original  central  area  of  infection 
is  returning  to  the  normal  state,  and  it  rarely  occurs  in  adults  that 
the  erysipelatous  process  extends  back  into  the  originally  infected 
area.87  From  these  localized  laboratories  of  antibody  formation,  of 
course,  distribution  to  the  circulation  probably  takes  place  and  the 
complete  cure  of  the  patient  must  await  a  sufficient  concentration  of 
these  in  the  body  as  a  whole  before  further  local  foci  cease  to  arise. 

That  the  fixed  tissue  cells  of  any  part  of  the  body  can  and  do 
take  an  active  part  in  the  local  reaction  against  the  invasion  of  bac- 
teria and  other  foreign  materials  is  histologically  evident.  When 
a  more  or  less  insoluble  foreign  body — a  thread  of  lint,  paraffin, 
agar-agar,  or  other  material — is  deposited  in  the  subcutaneous  tissues 
anywhere  in  the  body,  and  is  accompanied  by  acute  infection  with 
bacteria,  there  is  a  characteristic  tissue  reaction  which  results  in  the 
surrounding  of  the  foreign  particle  by  multinucleated  cells  spoken 
of  as  giant  cells.  In  the  case  of  foreign  bodies  such  as  those  men- 
tioned the  process  is  purely  one  of  local  ingestion  of  the  particle 
which  later,  if  the  material  remains  absolutely  insoluble,  results  in 
encapsulation  by  connective  tissue.  If  soluble,  however,  there  may 
be  an  eventual  digestion  of  the  foreign  material  by  the  cell  with  a 
subsequent  degeneration  or  splitting  up  of  the  giant  cell  and  a  return 
to  normal.  This  also  occurs  in  the  case  of  such  infections  as  those 
due  to  yeasts  or  blastomyces,  in  which,  as  the  writer  has  seen,  the 
apparent  lack  of  liberation  of  toxic  products  gives  rise  to  a  purely 
local  giant-cell  reaction,  adjacent  tissue  cells  remaining  undegener- 
ated  and  apparently  unaffected.  In  the  case  of  infection  with  bac- 
teria like  the  bacillus  of  tuberculosis,  the  leprosy  bacillus,  that  of 
rhinoscleroma,  and  a  few  others  the  purely  local  picture  of  giant- 
cell  phagocytosis  is  complicated  by  secondary  reactions  arising  prob- 
ably from  the  liberation  of  toxic  products  from  the  living  or  dead 
invaders  which  both  stimulate  specific  cell  reactions  and  call  forth 
cell  degeneration  in  adjacent  tissues,  frequently  giving  the  individual 
infection  a  diagnostically  characteristic  appearance. 

87  In  children  erysipelas  not  infrequently  returns  within  a  few  days  over  a 
recently  healed  area. 


CHAPTER   V 

TOXIN   AND   ANTITOXIN 

THE  EEACTION  BETWEEN  TOXIN   AND  ANTITOXIN 
(EHBLICH'S    ANALYSIS) 

THE  TOXIN-ANTITOXIN  REACTION 

WHEN  Behring  and  his  collaborators,  Kitasato  and  Wernicke, 
had  definitely  shown  that  the  cell-free  blood  serum  of  animals  im- 
munized with  tetanus  and  diphtheria  toxins  respectively  possessed 
the  power  to  protect  other  animals  of  the  same  and  different  species 
against  the  poisons,  it  became  of  the  utmost  importance  to  deter- 
mine, if  possible,  the  mechanism  by  which  the  "antitoxic"  effect  was 
attained.  The  earlier  opinion,  expressed  by  Behring  himself,  held 
that  in  all  probability  the  toxin  was  directly  injured  or  destroyed  by 
the  action  of  the  antitoxic  serum.-  That  this  assumption  was  incor- 
rect was  soon  demonstrated  by  the  experiments  of  Eoux  and  Vail- 
lard  1  and  by  those  of  Buchner.2  The  worl:  of  the  former  investiga- 
tors showed  that  the  mixtures  of  tetanus  toxin  and  antitoxin,  meas- 
ured in  such  proportions  that  they  were  harmless  for  normal  guinea 
pigs,  could  still  be  found  toxic  for  animals  weakened  by  preliminary 
inoculation  with  other  bacteria.  Buchner  claimed  in  analogous  ex- 
periments that  similar  mixtures,  harmless  for  mice,  could  still  show 
toxicity  for  guinea  pigs.  He  inferred  from  this  that  the  nature  of 
the  cell  reactions  of  different  animal  species  influenced  the  antitoxic 
effect.  Both  investigations  led  the  workers  to  conclude  that  the  pro- 
tective action  of  antitoxin  was  not  due  to  a  direct  effect  upon  the 
poison  but  was  potent  by  acting  upon  the  tissue  cells  of  the  animal  by 
protecting  these  from  subsequent  harm  by  the  toxin.  Their  concep- 
tion implied  an  indirect  protective  function  on  the  part  of  the  anti- 
toxin, not  due  to  any  direct  reaction  between  it  and  the  poison. 

That  this  explanation,  too,  was  faulty  was  made  evident  by  a 
number  of  investigations  which  took  advantage  of  the  peculiar  dif- 
ferences in  resistance  to  temperature  between  certain  toxins  and 
their  specific  antitoxins. 

1  Roux  and  Vaillard.    Ann.  de  I'Inst.  Past.,  1894. 

2  Buchner.    Munch,  med.  Woch.,  p.  427,  1893. 

104 


TOXIN    AND    ANTITOXIN  105 

In  1894  Calmette  3  and  Physalix  and  Bertrand  4  had  indepen- 
dently succeeded  in  obtaining  an  antitoxin  against  snake  poison.  In 
the  course  of  further  study  of  these  bodies  Calmette5  determined 
that  the  venoms  of  certain  varieties  of  snakes,  the  naja  and  cobra, 
would  remain  potent  even  when  subjected  to  100°  C.  for  a  very 
short  time.  In  contrast  to  this  the  antitoxins  to  these  poisons  were 
destroyed  at  much  lower  temperatures.  Now  when  mixtures  of 
the  two  substances,  so  proportioned  that  their  injection  into  ani-. 
mals  was  innocuous,  were  heated  to  68°  C.  for  considerable 
periods,  toxia  properties  again  became  evident,  a  demonstration  that 
the  toxin  had  not  been  destroyed,  but  had  remained  neutral  only  in 
the  presence  of  the  intact  antitoxins.  These  experiments  were  con- 
firmed by  Wassermann,6  who  found  that  similar  conditions  pre- 
vailed in  the  combination  between  pyocyaneus  toxin  and  antitoxin. 

The  filtration  experiments  of  Martin  and  Cherry  7  are  not  con- 
vincing since  they  may  be  taken  as  indicating  either  neutralization 
or  toxin  destruction.  These  workers  subjected  mixtures  of  snake 
poison  and  its  specific  antitoxin  to  filtration  through  gelatin  filters, 
under  pressure.  Under  the  experimental  conditions  thus  estab- 
lished the  presumably  smaller  toxin  molecule  was  allowed  to  pass 
through  the  filter  while  the  larger  antitoxin  molecule  was  held  back. 
They  showed  that  if  filtered  soon  after  the  ingredients  have  been  put 
together  most  of  the  toxin  still  passes  through,  but  that,  as  this  inter- 
val is  prolonged,  less  and  less  comes  through,  presumably  because  of 
the  union  of  the  smaller  toxin  to  the  larger  antitoxin  molecule.  The 
chief  value  of  these  experiments  lies  in  their  proof  of  the  element  of 
time  as  an  important  factor  in  the  toxin-antitoxin  union. 

In  his  experiments  on  snake  venom  just  recorded,  Calmette  in- 
terpreted the  restitution  of  toxicity  after  the  heating  of  neutral  mix- 
tures of  cobra  neurotoxin  and  its  antitoxin  as  evidence  "qu'il  ne 
s'etait  pas  forme  aucune  combinaison  de  ces  deux  substances  ou  que 
la  combinaison  realisee  etait,  au  moins,  tres  instable."  Later  experi- 
ments of  Martin  and  Cherry  seemed  for  a  time  to  contradict  this  con- 
clusion. Observations  by  them,  analogous  to  those  of  Calmette,  but 
carried  out  with  the  poison  of  an  Australian  snake,  seemed  to  indi- 
cate that  when  the  toxin  and  antitoxin  were  allowed  to  remain  to- 
gether for  a  sufficiently  long  time  no  restitution  of  toxicity  could  be 
obtained  by  heating.  Apparently  the  application  of  heat  to  such  mix- 
tures merely  prevented  the  further  union  of  antitoxin  with  any  toxin 
that  was  not  yet  bound  at  the  time  that  the  heat  was  applied.  Accord- 

3  Calmette.    Compt.  rend,  de  la  soc.  de  biol.,  1894. 

4  Physalix  and  Bertrand.     Compt.  rend,  de  la  soc.  de  biol.,  1894. 

5  Calmette.     Ann.  Past.,  1895. 

6  Wassermann.     Zeitschr.  f.  Hyg.,  22,  1896. 

7  Martin  and  Cherry.     Proc.  of  the  Royal  Soc.,  Vol.  63,  1898. 


106  INFECTION    AND    RESISTANCE 

ingly  Morgenrotli  8  again  examined  these  relations  and  found  that  the 
addition  of  a  small  amount  of  hydrochloric  acid  to  mixtures  of  snake 
poison  and  the  antitoxin  resulted  in  the  dissociation  of  their  union. 
To  mixtures  of  the  venom  lysin  and  its  antitoxin,  neutralized  and 
even  overneutralized  so  that  they  were  perfectly  innocuous  to  suscep- 
tible animals  he  added  hydrochloric  acid  until  the  total  concentra- 
tion amounted  to  N/18.  By  this  method  a  toxin-HCl  modification 
was  produced  which  was  dissociated  from  its  union  with  the  anti- 
toxin and  was  extremely  resistant  to  heat.  In  such  a  mixture  of 
toxin  and  antitoxin  to  which  hydrochloric  acid  had  been  added,  heat- 
ing at  100°  C.  in  a  water  bath  for  30  minutes  destroyed  the  ther- 
molabile  antitoxin  and,  after  neutralization,  undiminished  toxic 
properties  could  again  be  demonstrated  by  animal  inoculation. 

These  researches  and  other  similar  ones  of  Morgenroth,  then, 
form  a  satisfactory  confirmation  of  the  original  experiments  of  Cal- 
mette  and  seem  to  show,  beyond  possibility  of  contradiction,  that  the 
inhibition  of  harmful  properties  of  any  true  toxin,  after  mixture 
with  its  antitoxin,  does  not  depend  upon  toxin  destruction.  But 
while  Calmette  interpreted  the  facts  as  pointing  toward  a  failure  of 
union  of  the  two  substances,  Morgenroth's  work  is  not  incompatible 
with  the  conception  of  a  neutralization  of  one  by  the  other  in  the 
chemical  sense.  These  experiments  of  Morgenroth  are  of  great  the- 
oretical importance  moreover  in  that  they  have  shown  that  dissocia- 
tion of  a  toxin-antitoxin  complex  can  occur. 

The  nature  of  such  neutralizations  in  regard  to  quantitative  rela- 
tions, speed  of  action,  and  relative  concentrations,  becomes  apparent 
partly  from  experiments  like  those  mentioned  above,  but  more  espe- 
cially from  those  carried  out  by  Ehrlich  with  ricin  and  antiricin,  ex- 
periments which  were  primarily  planned  to  demonstrate  that  the 
reaction  between  a  toxin  and  its  antibody  is  a  direct  one,  not  depend- 
ent upon  intervention  of  the  body  cells,  as  at  first  supposed. 

It  had  been  shown  by  Kobert  and  Stillmarck  that  ricin,  the 
powerfully  poisonous  principle  of  Ricinus  communis  (castor  oil 
bean)  would  agglutinate  the  red  blood  cells  of  a  number  of  animals. 
Ehrlich  recognized  from  the  beginning  how  closely  analogous  the 
neutralization  of  ricin  by  antiricin  was  to  that  of  diphtheria  toxin  by 
its  antitoxin.  The  former  reaction  furnished  him  with  a  simple 
method  of  test  tube  experimentation  in  that  the  agglutinating  effects 
of  ricin  upon  rabbits'  corpuscles  could  be  directly  inhibited  by  the 
preliminary  addition  of  antiricin.  A  visible  reaction  was  thus  avail- 
able, which,  of  course,  excluded  absolutely  the  participation  of  the 
tissue  cells  in  the  antigen-antibody  neutralization,  and  in  which 
careful  quantitative  measurements  were  possible. 

Ehrlich  9  determined  by  means  of  this  method  that  the  neutral 

8  Morgenroth.     Berl.  kl  Woch.,  No.  50,  1905,  p.  1550. 

9  Ehrlich.    Fortschr.  d.  Med.,  Vol.  15,  p.  41,  1897. 


TOXIN    AND    ANTITOXIN  107 

ization  was  accelerated  by  moderate  heat  and  by  concentration  of  the 
reagents  and,  most  important  of  all,  that  the  reaction  followed 
roughly  the  law  of  multiple  proportions,  characteristics,  all  of  them, 
which  were  entirely  analogous  to  chemical  reactions  in  general. 
When  he  added  0.3,  0.5,  0.75,  0.1,  etc.,  cubic  centimeters  of  serum 
from  a  ricin-immune  goat  to  constant  quantities  of  ricin,  and  then 
added  rabbit  cells,  the  hemagglutinating  properties  of  the  ricin  were 
inhibited  in  direct  proportion  to  the  amount  of  antiricin  mixed  with 
it.  And  his  test  tube  experiments  were  further  found  to  represent 
with  much  accuracy  the  occurrences  which  took  place  within  the 
animal  body.  For,  similar  mixtures  injected  into  mice  were  toxic 
in  direct  proportion  to  the  balance  of  ricin  and  antiricin  established 
in  the  injected  material. 

Although  the  views  of  Ehrlich  and  his'  followers  have  great  im- 
portance in  connection  with  the  union  of  antigens  and  their  anti- 
bodies in  general,  these  ideas  were  worked  out  by  him  most  elab- 
orately in  connection  with  his  efforts  to  arrive  at  a  practicable  and 
accurate  method  of  establishing  a  standard  of  strength  for  diphtheria 
antitoxin,  and  it  is  essential  that  we  consider  this  work  in  detail. 

The  earlier  attempts  to  standardize  diphtheria  antitoxin  by  the 
use  of  living  cultures  (Roux  and  Behring)  were  soon  abandoned, 
since  it  was  found  that  the  accurate  establishment  of  fixed  lethal 
doses  of  the  culture  was  not  possible.  When  the  facts,  just  recorded, 
concerning  the  interaction  and  quantitative  relations  of  the  soluble 
toxins  and  their  respective  antitoxins  came  to  light,  Behring  intro- 
duced the  standardization  of  the  curative  sera  by  the  use  of  toxins, 
both  in  the  case  of  tetanus  and  in  that  of  diphtheria.  In  order  to 
do  this  consistently  he  established  for  diphtheria  poison  an  arbi- 
trary toxin  unit  which  he  defined  as  the  amount  of  any  given  diph- 
theria filtrate  sufficient  to  cause  death  in  a  guinea  pig  of  250 
grams,  and,  borrowing  the  terms  from  chemical  nomenclature,  he 
designated  as  a  "normal"  diphtheria  poison  one  which  contained 
100  such  units  in  one  cubic  centimeter.  (D  T  !N",  M250  =  diph- 
theria toxin  normal,  Meerschweinchen  250  grams.) 

Together  with  Ehrlich,  Behring  then  established  an  antitoxin 
unit  (I-E,  Immunitats  Einheit).  They  designated  as  a  "normal" 
antitoxic  serum  one  "which  contained  in  one  cubic  centimeter  one 
antitoxic  unit"  (I-E),  and  state  further,  "of  this  serum  0.1  c.  c. 
neutralizes  1  c.  c.  of  the  Behring  normal  toxin."  (Conf.  Madsen  in 
"Kraus  u.  Levaditi  Handbuch,"  II,  p.  94.)  Alterations  were  subse- 
quently made  in  this  scale  of  standards  and  Ehrlich  later  desig- 
nated as  an  antitoxin  unit  a  quantity  of  an  antitoxin  which 
completely  neutralized  100  lethal  doses  (for  guinea  pigs  of 
250  grams)  of  a  poison  at  that  time  in  his  possession.  The  unit 
of  diphtheria  antitoxin  at  present  in  use  therefore  may  be  defined  as 
a  quantity  of  serum  sufficient  to  protect  a  guinea  pig  of  250  grams 


108 


INFECTION    AND    RESISTANCE 


against  100  times  the  fatal  dose  (M  L  D,  minima  dosis  lethalis) 
of  toxin.  Since  the  methods  of  antitoxin  standardization  employed 
at  present  in  the  United  States  were  worked  out  by  Rosenau  10  along 
the  lines  of  Ehrlich's  method,  and  the  standard  is  based  on  the  one 
introduced  by  Ehrlich,  the  antitoxin  unit  as  employed  in  this  country 
is  identical  with  the  one  spoken  of  in  the  following  paragraphs. 

In  measuring  the  neutralizing  value  of  antitoxin  for  toxin,  then, 
since  both  substances  are  chemically  unknown  and  no  purely  chem- 
ical indicator  of  neutralization  is  available,  it  was  necessary  to  select 

a  susceptible  animal  by  means  of  which  excess 
of  toxin,  in  mixtures  of  the  two,  could  be  de- 
tected. As  the  standard  test  animal  guinea 
pigs  of  250  grams  were  chosen,  and  improve- 
ments in  the  methods  of  measurement  were 
introduced,  in  that  the  toxin  and  antitoxin, 
instead  of  being  separately  injected  as  hereto- 
fore, were  mixed,  allowed  to  stand  for  15  to 
30  minutes,  and  then  injected  together  sub- 
cutaneously. 

By  means  of  this  technique  Ehrlich  set 
out  to  examine  a  large  number  of  toxins  and 
their  antibodies  and  obtained  results  which, 
aside  from  their  practical  value,  have  had  an 
important  influence  upon  the  development  of 
the  knowledge  of  antigen-antibody  reactions. 
These  investigations  were  considerably 

complicated  by  the  fact  that  neither  the  diphtheria  toxin  nor  the  anti- 
toxin is  very  stable  and  deterioration  occurs  unless  special  methods  of 
preservation  are  employed.  Since  the  antitoxin,  however,  is  much 
less  unstable  than  the  toxin,  the  former  is  employed  in  order  to  pre- 
serve the  standard,  and  is  preserved  in  sealed  U  tubes  (see  Fig- 
ure) with  anhydrous  phosphoric  acid.  Kept  in  this  way,  in  black, 
light-proof  boxes,  and  at  low  temperature,  it  may  be  preserved  for 
months  without  appreciable  loss  of  value  and  may  be  renewed  by  ac- 
curate comparative  measurements  from  time  to  time.  This  is  carried 
out  for  the  United  States,  at  the  present  time,  by  the  Government 
Hygienic  Laboratories  at  Washington. 

Preservation  of  the  toxin  is  much  more  difficult,  and  it  is  in 
connection  with  the  investigation  of  the  instability  of  the  toxin  that 
Ehrlich  gained  his  first  insight  into  the  nature  of  the  reaction.  He 
measured,  in  a  number  of  toxic  filtrates,  the  minimal  lethal  dose 
for  guinea  pigs  of  250  grams,  establishing  a  time  limit  for  death 
in  order  to  obtain  more  accurate  comparisons.  He  designated  as  the 

10  Rosenau.  U.  S.  P.  H.  &  Mar.  Hosp.  Service.  Hygienic  Laboratory  BulL> 
21,  April,  1905. 


TUBE  FOR  THE  PRESERVA- 
TION or  THE  STAND- 
ARD ANTITOXIN. 

Taken  from  Eosenau,  U. 
S.  Hygienic  Labora- 
tory Bulletin,  No.  21, 
1905,  p.  53. 


TOXIN    AND    ANTITOXIN  109 

new  M  L  D  or  "T"  (that  is:  toxic  unit)  the  quantity  of  toxin 
which  will  kill  a  guinea  pig  of  the  designated  weight  in  from  4  to  5 
days.  He  then  determined  for  a  number  of  poisons  the  exact  quan- 
tity just  neutralized  by  one  antitoxin  unit,  calling  this  quantity  L0. 
(L  meaning  Limes  or  threshold.) 

It  is  clear  that  in  judging  of  complete  neutralization  of  a  quan- 
tity of  toxin  by  antitoxin,  there  may  be  a  strong  subjective  element, 
since  any  very  slight  excess  of  toxin  may  cause  unimportant  local 
reactions  such  as  edema  or  small  hemorrhages,  which  could  escape 
the  attention  of  one  observer  while  being  noticed  and  recorded  by 
another.  In  order  therefore  to  exclude  definitely  all  subjective  fea- 
tures from  such  experimentation,  Ehrlich  now  established  another 
toxin  value — L+  dose  ("Limes  death" — now,  for  convenience,  writ- 
ten L  +  ) — which  eliminated  all  possible  variations  of  personal  per- 
ception. He  designated  by  this  symbol  that  quantity  of  toxin  which 
not  only  neutralized  one  antitoxin  unit  but  included  enough  toxin, 
in  excess  of  this,  to  give  the  result  of  one  free  toxin  unit,  that  is,  to 
cause  death  in  4  to  5  days  in  a  guinea  pig  of  250  grams.  Since 
the  three  values  just  defined  form  the  basis  of  Ehrlich's  experiments 
as  well  as  that  of  all  practical  diphtheria  serum  standardizations  we 
will  briefly  restate  them  for  the  sake  of  clearness. 

Thus: 

M  L  D  or  "T"  =  the  amount  of  toxin  which,  subcutaneously  injected, 
causes  death  in  a  250-gram  guinea  pig  in  from  4  to  5  days. 

L0  =  the  amount  of  toxin  which  is  just  neutralized  by  one  antitoxin  unit  so 
that  no  trace  of  reaction,  local  or  otherwise,  ensues 

and 

L+  =  that  amount  of  toxin  which  will  cause  death  in  4  to  5  days  in  a  guinea 
pig  of  250  grams  if  injected  together  with  one  antitoxin  unit. 


It  will  further  clarify  the  meaning  of  these  terms  to  examine 
experimental  protocols  which  show  how  these  values  are  determined. 
Thus  in  the  following: 

I.  Injections  of  toxin 

(1)  .005  c.  c. — guinea  pig  lives. 

(2)  .009  c.  c. — guinea  pig  dies  in  6  days. 

(3)  .01    c.  c. — guinea  pig  dies  in  4  days. 

(4)  .02    c.  c. — guinea  pig  dies  in  2  days. 
.01  =  M  L  D  or  T. 

II.  1  Antitoxin  unit  +  .19  toxin  =  late  paralysis. 

1  Antitoxin  unit  +  .20  toxin  =  sometimes  late  paralysis  and  sometimes 

acute  death. 

1  Antitoxin  unit  +  .21  toxin  =  death  fourth  day. 
1  Antitoxin  unit  +  -22  toxin  =  death  in  2  to  3  days. 
.21  =  L+  dose. 


110  INFECTION    AND    RESISTANCE 

III.     Antitoxin  unit  -f  .14  c.  c.  toxin  =  no  reaction. 

Antitoxin  unit  -j-  .15  c.  c.  toxin  =  no  reaction. 

Antitoxin  unit  +  .16  c.  c.  toxin  =  slight  congestion  about  point  of  injec- 
tion, scarcely  visible. 

Antitoxin  unit  +  .17  c.  c.  toxin  =  apparent  reaction  at  site. 

Antitoxin  unit  -j-  .18  c.  c.  toxin  =  edema  at  site. 
Lo  =  .16." 

In  determining  these  values  with  a  large  number  of  toxins  Ehr- 
lich  discovered  the  curious  fact  that,  although  there  was  a  rapid  and 
extensive  diminution  of  toxicity  in  every  toxic  filtrate  in  the  course 
of  time,  there  was  no  corresponding  alteration  in  the  L0  amount. 
In  other  words,  although  more  and  more  of  the  toxic  broth  was 
necessary  to  kill  a  guinea  pig  of  standard  weight  in  the  required 
time,  the  amount  of  the  same  broth  which  neutralized  one  antitoxin 
unit  remained  approximately  the  same.12 

In  seeking  an  explanation  for  this  apparent  paradox,  Ehrlich 
concluded  that  we  must  assume  that  the  toxin  is  complexly  con- 
structed, consisting  of  a  toxophore  and  a  haptophore  group.  Assum- 
ing that  chemical  union  between  the  toxin  and  the  antitoxin  (or,  in 
disease,  the  body  cell)  takes  place,  it  is  by  means  of  the  haptophore 
group  that  such  union  is  brought  about.  The  toxophore  group,  how- 
ever, is  the  element  by  which  toxic  action  is  exerted  after  union 
by  the  haptophore  group  has  been  accomplished.  It  would  be 
conceivable,  therefore,  that  in  deteriorating  in  toxicity  the  toxin 
might  undergo  alterations  in  the  toxophore  group  only,  its  hapto- 
phore group,  and,  therefore,  its  antitoxin  neutralizing  properties 
remaining  intact.  Such  modified  toxins,  modified  only  as  to  the 
toxophore  groups,  Ehrlich  now  refers  to  as  "toxoids." 

In  the  production  of  diphtheria  toxins  for  practical  purposes  it 
has  been  found  advisable  to  allow  them  to  "season,"  that  is  to  stand 
for  prolonged  periods  until  they  have  reached  a  state  of  "equili- 

"II  and  III  are  taken  from  the  article  by  Rosenau,  P.  H.  &  M.  H.  S., 
Hyg.  Lab.  Bulletin,  21,  1905. 

12  This  statement  plainly  contradicts  the  definition  of  a  toxin  unit;  i.  e., 
the  amount  which  neutralizes  100  toxin  units  and  often  leads  to  confusion 
among  students  or  others  who  are  unfamiliar  with  this  subject.  It  should 
be  borne  in  mind  that,  while  the  definition  of  an  antitoxin  unit  is  the  one 
accepted  when  Ehrlich  first  arbitrarily  established  it,  the  antitoxin  unit, 
as  at  present  in  use,  is  really  an  amount  of  antitoxin  standardized  against 
L+  quantities  of  toxin,  this  last  value  again  obtained  by  measurement 
against  the  original  unit.  It  represents  a  neutralization  value  equal  to  the 
original  one,  but  may  protect  the  guinea  pig  against  85,  110,  130,  etc. 
(variable)  toxin  units,  according  to  the  constitution  of  the  particular  toxic 
filtrate  employed  in  the  experiment.  Indeed,  if,  in  the  following  pages,  the 
reasoning  of  Ehrlich  is  consistently  adhered  to,  our  definition  of  an  anti- 
toxin unit  should  be:  The  amount  of  antitoxin  which  contains  200  binding 
affinities  for  toxin.  This  will  become  clearer  as  the  following  paragraphs 
are  read. 


TOXIN    AND    ANTITOXIN  111 

brium"  at  which  the  conversion  of  toxin  to  toxoids  has  been  reduced 
to  a  minimum  and  the  change  of  relationship  between  L0  and  aT" 
or  M  L  D  has  practically  ceased  to  go  on.  From  the  very  begin- 
ning of  the  growth  of  the  culture  in  the  incubator  the  process  of 
toxoid  formation  has  probably  occurred,  and  even  freshly  prepared 
toxic  filtrates  therefore  are  not  pure  "toxins,"  especially  since  the 
conversion  of  toxin  to  toxoid  seems  to  diminish  in  velocity  as  time 
goes  on. 

^sTow  in  spite  of  the  presence  of  such  alteration  products,  in  com- 
paring the  values  L0  and  L+  of  any  given  toxin  preparation,  one 
would  naturally  suppose  that  L+  minus  L0  should  be  equal  to  one 
M  L  D,  or  the  quantity  just  sufficient  to  kill  a  guinea  pig  of  250 
grams  in  4  to  5  days.  For  we  have  seen  that  L0  just  neutralizes 
one  antitoxin  unit  while  L+  is  the  quantity  which,  in  addition  to 
such  neutralizing  power,  has  an  excess  of  toxin  equal  in  action  to 
one  minimal  lethal  dose.  This,  however,  is  not  the  case.  Let  us 
illustrate  this  by  a  concrete  case.  One  of  Ehrlich's  toxins  on  meas- 
urement showed  a  minimal  lethal  dose  or  M  L  D  of  0.0025  c.  c. 

The  L+  dose  of  this  was  .  25 
while  The  L  o  dose  of  this  was  .  125 

The  difference  was  .125  or  50  M  L  D  instead  of  1  M  L  D  as  we 

would  suppose. 

Stated  in  words,  this  measurement  means  that  after  neutralizing 
completely  one  antitoxin  unit  with  the  toxic  filtrates,  in  order  to 
obtain  death  in  a  guinea  pig  in  4  days  with  such  a  mixture,  it  was 
necessary  to  add,  beyond  the  neutralizing  quantity,  50  M  L  D,  or 
again  as  much  as  was  necessary  for  neutralization. 

This  last  relation  is  merely  coincidence,  since  it  might  have  been 
30  or  40  or  60  M  L  D  just  as  well.  The  important  point  is  the  fact 
that  more  than  1  M  L  D  was  necessary,  and  by  this  fact  Ehrlich  was 
led  to  resort  to  an  assumption  which  forms  one  of  the  basic  princi- 
ples of  many  of  his  explanations  for  serum  phenomena,  namely,  the 
assumption  of  differences  in  combining  avidity  or  affinity. 

As  applied  to  the  present  problem  he  reasoned  as  follows: 

It  is  conceivable  that  the  toxoids  resulting  from  deterioration  of 
toxin  might  possess  three  different  degrees  of  affinity  for  the  anti- 
toxin. They  might  have  a  stronger,  an  equal,  or  a  lesser  affinity  than 
the  toxin  itself.  If  their  affinity  for  antitoxin  were  equal  to  that  of 
toxin  they  would,  of  course,  not  influence  the  L+  dose  itself;  if 
stronger  than  toxin  their  influence  would  be  so  exerted  that  toxin 
would  be  forced  out  of  combination  with  antitoxin,  giving  place  to 
the  toxoid,  and  the  effect  would  be  the  opposite  from  that  experi- 
mentally observed.  If,  however,  their  affinity  for  antitoxin  were 
weaker  than  that  of  toxin  each  additional  toxin  unit  added  to  the 
L0  dose  would  unite  with  antitoxin,  replacing  a  corresponding  quan- 


INFECTION    AND    RESISTANCE 

tity  of  the  toxoid  of  weaker  affinity.  In  consequence,  as  far  as  the 
poisonous  properties  of  the  mixture  are  concerned,  the  addition  of 
toxin  would  not  render  the  neutral  mixture  poisonous  for  guinea 
pigs  until  the  toxoids  had  been  completely  displaced  from  union  with 
antitoxin.  Finally,  after  all  the  antitoxin  had  been  bound  to  un- 
changed toxin,  further  addition  would  then  result  in  the  presence  of 
free  toxin  and  poisonous  properties  would  again  appear  in  the  mix- 
ture. Ehrlich  at  first  spoke  of  the  toxoids  possessing  weaker  affinity 
for  antitoxin  than  the  toxin  itself  as  "epitoxoids."  This  conception 
can  be  rendered  clear  by  the  following  example : 

In  the  case  cited  above  we  had 
TorMLD      =  0.0025  c.  c. 
L+  =  0.25      c.  c. 

Lo =  0.125    c.  c. 

The  difference  =  0.125  =  50  M  L  D. 

Supposing  that  the  toxoids  (epitoxoids)  present  in  the  mixture 
possessed  an  affinity  for  antitoxin  less  than  that  of  toxin,  the  follow- 
ing conditions  might  obtain : 

151  toxin-antitoxin  -|-  49  epitoxoid-antitoxin  =  L0. 
Add  1  M  L  D  or  T  and  we  have: 

152  toxin-antitoxin  +  48  epitoxoid-antitoxin  +  1  epitoxoid  free. 
Add  2  M  L  D  or  T  and  we  have: 

153  toxin-antitoxin  +  47  epitoxoid-antitoxin  -f  2  epitoxoid  free  until, 
finally,  adding  50  T,  we  get: 

200  toxin-antitoxin  +  49  epitoxoid  free  +  1  toxin  free  =  L+.13 

Later  experience  led  Ehrlich  to  abandon  the  opinion  that  the 
epitoxoids  were  deterioration  products  of  the  toxin.  He  found  that 
the  relation  between  L0  and  L+  which  we  have  just  outlined,  was 
demonstrable  in  the  same  way,  in  freshly  prepared  toxin  filtrates, 
in  which  there  had  been  little  time  for  toxoid  formation.  He  further 

13  An  example  identical  in  significance  with  the  one  just  given,  but  some- 
what simpler  in  its  arithmetical  conditions,  is  here  added  for  the  sake  of 
permitting  no  possibility  of  unclearness.  This  example  is  taken  from  Ehr- 
lich's  own  work. 

T  =  .01  c.  c.  of  the  toxin  bouillon. 

L+     (neutral,  of  antitoxin  unit  yet  killing  1  pig)  =  2.01  c.  c.  or  201  T. 
Lo      (complete  neutral,  of  1  antitoxin  unit)          =  1       c.  c.  or  100  T. 

Difference  =  1.01  c.  c.  or  101  T 


100  toxin-antitoxin  +  100  epitoxoid  antitoxin  =  L0; 
Add  1  T,  and  we  have: 

101  toxin-antitoxin  +  99  epitoxoid-antitoxin  +  1  epitoxoid  free; 
Add  101  T,  and  we  have: 

200  toxin-antitoxin  +100  epitoxoid  free  -f  1  T  free  =  L+. 


TOXIN    AND    ANTITOXIN  113 

noticed  that,  even  after  deterioration  had  occurred  to  a  considerable 
extent,  and  the  amount  necessary  to  kill  a  guinea  pig  had  been  much 
increased  (the  number  of  fatal  doses  in  L0  constantly  decreasing  as 
toxoids  replaced  toxin),  the  L+  nevertheless  remained  unchanged. 
This,  he  held,  could  mean  one  thing  only.  The  elements  present  in 
toxic  broth  which  possessed  a  weaker  affinity  for  antitoxin  than  the 
toxin  itself,  namely,  the  epitoxoids,  were  present  from  the  very  begin- 
ning and  were  probably  separate  and  primary  secretion  products  of 
the  diphtheria  bacilli,  remaining  relatively  stable  and  constant  as  the 
broth  was  preserved.  In  order  to  avoid  confusion,  therefore,  he  now 
referred  to  the  "epitoxoids"  as  "toxons" — to  preclude  their  confu- 
sion with  the  other  toxoids  or  true  toxin  deterioration  products. 
These  toxons  possess,  according  to  Ehrlich,  a  "haptophore"  group 
identical  with  that  of  the  toxin,  but  have  a  different  toxophore  group. 
For  there  is  reason  to  believe  that  the  toxon,  lacking  the  power  of 
causing  acute  death,  gives  rise  to  slow  emaciation  and  paralysis, 
finally  killing  after  a  subacute  or  chronic  course.  Thus,  in  the  tab- 
ulation just  preceding,  we  have  seen  that  the  toxic  broth  added  to 
neutral  mixtures  of  toxin  and  antitoxin  (containing  the  L0  dose), 
does  not  give  rise  to  the  acutely  toxic  effect  of  one  M  L  D  or  T  until 
an  amount  has  been  added  which  considerably  exceeds  one  toxin 
unit.  This,  we  explained,  by  Ehrlich's  reasoning,  on  the  supposi- 
tion that  "epitoxoids"  or  "toxons"  are  displaced  from  their  union 
with  antitoxin,  giving  place  to  toxin  and  becoming  free.  Such  toxin- 
antitoxin  mixtures — in  which  the  amount  of  toxin  broth  ranges  be- 
tween the  L0  and  the  L+  doses — therefore,  contain  no  free  toxin 
units  but  contain  varying  amounts  of  free  toxon.  An  injection  into 
guinea  pigs  is  not  followed  by  acute  death  in  these  cases,  but  leads 
with  considerable  regularity  to  emaciation,  paralysis,  and  death 
after  a  long  incubation  period. 

It  has  been  objected  to  this,  as  we  shall  see,  that  the  slow  poison- 
ing produced  by  such  mixtures  is  due,  not  to  a  qualitatively  differ- 
ent poison  but  to  the  presence  of  minute  quantities  of  free  toxin  too 
small  to  produce  acute  death,  yet  sufficient  to  produce  this  gradual 
injury.  This  Dreyer  and  Madsen  14  tried  to  disprove  by  experi- 
ments in  which  they  prepared  antitoxin-toxin  mixtures  so  bal- 
anced that  they  possessed  the  toxon  effect,  and  of  these  mixtures 
injected  increasing  multiples.  In  no  case  did  they  succeed  in  pro- 
ducing acute  death  even  when  the  amount  injected  had  been  multi- 
plied to  such  an  extent  that  free  toxin,  if  present,  must  have  asserted 
itself.  The  same  workers  were  able  to  show  that  the  injection  of 
these  mixtures,  in  which  free  toxons  were  assumed  to  be  present, 
produced  immunity  against  toxin,  thus  indicating  the  similarity  of 
the  haptophore  group  of  toxin  and  toxon — a  conception  which  will 
14  Dreyer  and  Madsen.  ZeitscJir.  /.  Hyg.,  Vol.  37,  1901. 


INFECTION    AND    RESISTANCE 

become  more  and  more  clear  as  we  consider  the  "Side-Chain  Theory" 
which  Ehrlich  evolved  as  a  result  of  his  toxin  analysis. 

Ehrlich  had  thus  elicited  facts  which  seemed  to  him  to  indicate 
the  presence  of  three  qualitatively  different  substances  in  toxic  fil- 
trates of  diphtheria  cultures.  Two  of  these,  the  toxin  and  the  toxon, 
were  present,  he  assumed,  in  freshly  prepared  filtrates,  as  indepen- 
dent primary  secretion  products  of  the  bacilli,  the  toxin  an  acute 
poison,  the  toxon  a  substance  with  slower  and  qualitatively  different 
poisonous  effects.  Both  of  them,  toxin  and  toxon,  possessing  similar 
haptophore  groups,  could  unite  with  antitoxin  and  neutralize  it,  but 
the  affinity  of  toxon  for  antitoxon  was  weaker  than  that  of  toxin.  For 
this  reason  toxin  could  displace  toxon  from  its  union  with  antitoxin, 
this  accounting  for  the  discrepancy  between  the  L+  and  the  L0 
doses.  The  third  class  of  substances,  the  toxoids,  were  deterioration 
products  of  toxin,  the  deterioration  implying  an  alteration  in  the 
toxophore  group  only,  the  haptophore  group  remaining  the  same. 

It  is  plain  from  this  reasoning  that  Ehrlich's  conception  implies 
complete  analogy  between  chemical  reactions  in  general  and  the 
neutralization  of  toxin  by  antitoxin.  Accordingly  it  is  but  another 
step  in  the  same  direction  to  speculate  concerning  the  actual  rela- 
tions of  valency  existing  between  the  two  substances.  It  seemed  to 
Ehrlich  that  there  were  many  reasons  for  assuming  that  the  union 
between  toxin  and  antitoxin  occurred  in  proportions  of  200  to  1 ; 
that  is,  just  as  the  formula  for  water  is  H2O,  that  of  toxin-antitoxin 
combinations  would  be  "Toxin20oAntitoxin." 

The  considerations  on  which  this  opinion  was  based  were  as  fol- 
lows: In  examining  a  large  series  of  toxic  filtrates,  Ehrlich,15  as 
well  as  Madsen,  had  found  that  the  number  of  toxin  units  ("T"  or 
M  L  D)  necessary  to  neutralize  one  antitoxin  unit  (that  is,  the 
number  of  toxin  units  contained  in  the  L0  dose)  corresponded, 
with  great  regularity,  to  multiples  of  100.  Values  of  25,  33,  50, 
100,  etc.,  recurred  again  and  again.  This  indicated  that  the  de- 
terioration of  the  toxin  into  toxoids  followed  a  certain  regularity  o*f 
progression  and  seemed  to  justify  the  assumption  that  the  absolute 
binding  power  possessed  by  antitoxin  was  represented  by  a  valency 
corresponding  to  a  multiple  of  100.  Since  the  number  of  toxin  units 
contained  in  an  Lo  dose  rarely  fell  below,  and  usually  above  100,  the 
valency  could  not  be  less  than  100.  On  the  other  hand,  repeated 
measurements  of  L0  and  L+  doses  never  showed  as  many  as  200 
toxin  units.  Ehrlich's  own  highest  value  was  133,  and  the  highest 
ever  obtained  by  any  one  was  a  measurement  by  Madsen  of  160. 
Now  considering  the  fact  that  no  toxin  is  "pure"  but  that,  in  every 
case,  it  contains  admixtures  of  toxoid  and  toxon,  the  values  133  or 
160  cannot  represent  all  the  valencies  of  an  antitoxin  unit.  They 
represent  only  that  part  of  the  antitoxin  unit  which  is  neutralized 
10  Ehrlich..  Deutsche  med.  Woch.,  No.  38,  1898,  Vol.  24, 


TOXIN    AND    ANTITOXIN  115 

by  the  "toxin,"  as  measurable  upon  guinea  pigs,  a  certain  fraction 
of  antitoxin  being  united  to  toxoid  or  toxon.  It  is  likely,  therefore, 
as  Ehrlich  reasoned,  that,  being  higher  than  100,  and  in  an  ob- 
viously impure  condition  approaching  but  never  reaching  200,  the 
valency  of  antitoxin  for  toxin  was  just  200.  The  correctness  of  this 
surmise  seemed  rendered  more  probable  by  Ehrlich's  further  studies, 
since  analysis  of  the  ingredients  of  various  toxic  filtrates,  that  is,  the 
determination  of  the  relative  contents  of  toxin,  toxoids,  and  toxon, 
appeared  to  show  constantly  definite  relations  to  the  valency  200. 

The  method  by  which  Ehrlich  carried  out  these  subsequent  stud- 
ies is  spoken  of  as  the  method  of  "Partial  Absorption."  In  prin- 
ciple it  represents  a  reversal  of  his  earlier  methods  of  measurement. 
In  these  he  had  titrated  various  amounts  of  toxin  broth  against  a 
constant  quantity  (one  unit)  of  antitoxin,  establishing  the  L+  and 
L0  values.  In  the  method  of  Partial  Absorption,  on  the  other  hand, 
he  measured  varying  fractions  of  an  antitoxin  unit  against  a  con- 
stant amount  of  toxin,  employing  for  this  a  previously  determined 
L+  and  L0  dose.  A  measurement  carried  out  in  this  way  is  shown 
in  the  following  tabulation  in  which,  at  the  same  time,  there  is  indi- 
cated how  many  valencies  each  antitoxin  fraction  represents,  on  the 
basis  of  an  assumed  total  of  200  for  each  unit.16 

0    antix.  unit  representing      0  valency    +  L+  =  85     free  T  units 

.1    antix.  unit  representing    20  valencies  -f  L+  =  85     free  T  units 

.25  antix.  unit  representing    50  valencies  +  L+  =  60     free  T  units 

.8    antix.  unit  representing  160  valencies  4"  L+  =  10    free  T  units 

.9    antix.  unit  representing  180  valencies  +  L+  =  3.5  free  T  units 

1.0    antix.  unit  representing  200  valencies  +  L+  =  1     free  T  unit 

It  is  immediately  evident  in  this  table,  as  it  would  be  evident  in 
any  other  citation  of  similar  measurements,  that  there  is  no  regu- 
larity in  the  progress  of  neutralization ;  or,  in  other  words,  that  addi- 
tion of  a  definite  fraction  of  antitoxin  does  not  remove  the  arithmet- 
ically corresponding  amount  of  toxic  properties  from  the  L+  dose. 
The  first  0.1  unit  of  antitoxin  in  this  table  has  removed  no  free 
toxin  whatever.  The  addition  of  the  next  0.15  of  an  antitoxin  unit, 
representing  30  valencies,  has  removed  f-f  or  T\  of  the  total  toxicity. 
Throughout  the  scale  there  is  not  the  regular  progression  of  neutral- 
ization, multiple  by  multiple,  which  would  be  expected  if  antitoxin 
could  be  titrated  against  a  pure  toxin.  This,  according  to  Ehrlich,  is 
due  to  the  presence  of  various  toxoids  which  possess  varying  affinities 
for  the  antitoxin  molecule.  The  first  0.1  of  a  unit  added,  in  this  case, 
does  not  diminish  the  toxicity  of  the  mixture  because  it  is  bound  by 
"protoxoids"  which  possess  a  higher  affinity  for  antitoxin  than  the 

16  This  measurement  is  taken  from  one  cited  by  Ehrlich  in  Deutsche  med. 
Woch.,  No.  38,  1898,  Vol.  24,  and  is  taken  literally  except  for  the  first  value 
of  1/10  antitoxin  unit,  which  is  inserted  to  illustrate  the  formation  of  pro- 
toxoids. 

9 


116 


INFECTION    AND    RESISTANCE 


toxin  itself.  Next  are  bound  the  toxins  themselves  together  with 
varying  amounts  of  "syntoxoicls"  which  possess  the  same  affinity  as 
toxin.  Finally  there  are  left  the  toxons  which  possess  a  lesser  affin- 
ity than  toxins  or  toxoids,  and  therefore  again  have  the  discrepancy 
between  the  L0  and  L+  dose.  Ehrlich  utilizes  this  method  in  order 
to  determine  the  composition  of  the  constituents  of  any  given  toxic 
filtrate  and  expresses  the  results  in  the  so-called  "toxin  spectra." 

The  construction  of  these  spectra  and  the  principles  underlying 
the  measurements  on  which  they  are  based  are  very  clearly  illus- 
trated by  Madsen,17  from  whose  article  the  following  type  spectra 
are  taken : 18 


9  »  20  K>  40  M  60  TO  00  so  too  no 

TOXIN   SPECTRUM  AFTER  MADSEN,  Ann.  de   I'lnst.  Past.,  Vol.  13,  1899,  p.  57. 


This  figure  represents  the  diphtheria  filtrate  composed  of  50 
valencies  of  protoxoid,  100  toxin  and  50  toxon  equivalents.  The 
measurements  in  this  case  may  be  represented  by  the  following  tab- 
ulation : 

LO  -h  1    antitox.  unit  =  200  valencies  =       0  lethal  dose 

Lo  -j-  .75  antitox.  unit  =  150  valencies  =       0  lethal  dose 

L0  +  .25  antitox.  unit  =    50  valencies  =  100  lethal  doses 

Lo  -j-  0    antitox.  unit  =      0  valency     =  100  lethal  doses 

The  following  diagram,  also  from  Madsen,  represents  the  same 
poison  after  it  had  deteriorated  to  ^  its  toxic  power.  Lo,  therefore, 
would  contain  only  50  toxic  doses. 


AFTER  MADSEN,  Ibid.,  p.  578. 

The  measurements  corresponding  to  this  table  are  as  follows: 

L0  +  I .       antitox.  unit  =  200  valencies  =  0  lethal  dose 

Lo  4"    .75    antitox.  unit  =  150  valencies  =  0  lethal  dose 

Lo  +    .745  antitox.  unit  =  149  valencies  =  0  lethal  dose 

Lo  4-    -74    antitox.  unit  =  148  valencies  =  1  lethal  dose 
etc.  until 

17  Madsen.     Ann  Past.,  Vol.  13,  1899,  p.  576. 

18  We  have  chosen  to  illustrate  the  principles  of  the  toxin  spectrum  from 
the  article  of  Madsen  rather  than  from  Ehrlich's  original  work,  because  the 
former  presents  this  difficult  phase  of  the  subject  in  a  hypothetical  toxin  of 
extremely   simple   structure.      Some   of   Ehrlich's   spectra   constructed   from 
actual  measurements  may  be  found  in  Deutsche  med.  Woch.,  No.  38,  1898. 


TOXIN    AND    ANTITOXIN 


117 


Lo  +    -25    antitox.  unit  =    50  valencies  =  50  lethal 
Lo  -f-    0      antitox.  unit  =      0  valency     =  50  lethal 


The  following  spectrum,  the  third  in  Madsen's  article,  represents 
the  same  toxin  described  in  the  preceding  spectrum  but  at  a  later 
period,  at  which  considerable  further  deterioration  had  taken  place. 
The  L0  dose  now  contained  but  30  M  L  D  or,  in  other  words,  the 
amount  of  toxin  contained  in  the  L0  dose  was  sufficient  to  kill  30 
guinea  pigs  only. 


PKOTOTOXOID 


ao  90100,10 

AFTER  MADSEN,  Ibid,.,  p.  579. 


Madsen's  description  of  the  method  in  which  this  spectrum  is 
constructed  is  the  following : 

LO  +  f  o&  of  an  antitoxin  unit  kills    0  guinea  pig 
LO  +  mi  of  an  antitoxin  unit  kills    0  guinea  pig 


L0  + 

H*of 

an 

antitoxin 

unit 

kills 

0  guinea 

Pig 

L0  + 

14D   nf 
3~OQ    OI 

an 

antitoxin 

unit 

kills 

1 

guinea 

Pig 

L0  + 

mof 

an 

antitoxin 

unit 

kills 

2 

guinea 

pigs 

L0  + 

H$  Of 

an 

antitoxin 

unit 

kills 

3 

guinea 

pigs 

L0  + 

M$  of 

an 

antitoxin 

unit 

kills 

5 

guinea 

pigs 

L0  + 

^r  of 

an 

antitoxin 

unit 

kills 

5 

guinea 

pigs 

L0  + 

/o8^  of 

an 

antitoxin 

unit 

kills 

6 

guinea 

pigs 

L0  + 

~/Wo  of 

an 

antitoxin 

unit 

kills 

6 

guinea 

pigs 

L0  + 

/A  of 

an 

antitoxin 

unit 

kills 

7 

guinea 

pigs 

L0  + 

A%  of 

an 

antitoxin 

unit 

kills 

10 

guinea 

pigs 

L0  + 

•2^V  Of 

an 

antitoxin 

unit 

kills 

30 

guinea 

pigs 

L0  + 

^^  of 

an 

antitoxin 

unit 

kills 

30 

guinea 

pigs 

The  amount  of  toxon  has  remained  the  same  in  spite  of  deteriora- 
tion. As  less  and  less  antitoxin  is  added,  between  the  values  of  -J£$ 
and  ^$J  of  an  antitoxin  unit,  there  are  now  liberated  only  5  fatal 
doses  of  the  toxin.  It  is  in  this  zone  that  deterioration  has  taken 
place,  since,  in  the  preceding  spectrum,  the  difference  between  the 
addition  of  %$-%  and  -J-J^  of  an  antitoxin  unit  represented  25  fatal 
doses  for  guinea  pigs.  When  in  this  last  spectrum  the  amount  of 
antitoxin  is  gradually  reduced  from  100  valencies  to  50  valencies  25 
fatal  doses  are  liberated,  a  quantity  corresponding  to  the  similar 
zone  in  the  preceding  spectrum.  Thus  in  this  particular  zone  of  the 
spectrum  no  change  has  taken  place.  The  same  is  true  of  the  pro- 
toxoid  zone. 

It  is  unnecessary  to  cite  a  larger  number  of  such  measurements 
in  this  place,  since  the  ones  given  sufficiently  illustrate  the  methods 


118  INFECTION    AND    RESISTANCE 

and  the  conclusions  drawn  from  them.     As  a  result  of  such  experi- 
ments Ehrlich  concludes: 

I.  That  the  diphtheria  bacillus  produces  primarily  two  kinds 
of  substances  (a)  toxin,  (b)  toxon,  both  of  which  bind  the  antibody. 

II.  The  toxins  (and  perhaps  also  the  toxons)  may  deteriorate 
and  be  modified  into  secondary  substances  (toxoids)  which  may  be 
distinguished  by  their  different  degrees  of  affinity  for  antitoxin. 

III.  This  classification  does  not  exhaust  all  possible  complica- 
tions, since  each  subdivision  of  toxin  consists  apparently  of  equal 
parts  of  two  different  modifications  which  are  similar  to  each  other 
in  their  relation  to  antitoxin  but  differ  in  varying  resistance  to  in- 
fluences of  deterioration.     A  more  complete  analysis  of  these  condi- 
tions may  be  found,  together  with  a  series  of  illustrative  spectra,  in 
Ehrlich's  article  in  the  Deutsche  med.   Wochenschr.,  Sept.,   1898, 
which  has  been  quoted  above. 

The  complex  deductions  arrived  at  by  Ehrlich  are  largely  de- 
pendent, as  we  have  seen,  upon  strict  adherence  to  the  analogy  be- 
tween the  toxin-antitoxin  reactions  and  those  occurring  between 
strong  acids  and  strong  bases.  In  such  cases  there  is  a  complete 
reaction,  in  which  chemical  change  ceases  only  when  there  has  been 
a  complete  neutralization  of  one  by  the  other.  If,  for  instance,  we 
mix  molecular  equivalent  amounts  of  H2SO4  and  NaOH,  an  ap- 
parently complete  change  into  Na2SO  and  H2O  occurs: 

H2S04  +  2  NaOH  =  Na2S04  +  2  H20 

The  reverse  process  does  not  seem  to  take  place,  and  if  traces 
of  uncombined  H2SO4  and  NaOH  are  present,  as  may  be  theoret- 
ically assumed,  they  are  so  slight  in  amount  that  they  are  not  dem- 
onstrable. There  are,  however,  many  chemical  reactions  in  which 
the  process  is  not  a  complete  one,  in  that  the  chemical  change  does 
not  proceed  until  the  reagents  are  completely  used  up.  Reaction  in 
these  cases  ceases  when  an  equilibrium  is  reached  at  which  there  are 
present  definite  amounts  of  the  reaction  products  and  of  the  original 
substances  at  the  same  time.19 

This  occurs  when  a  weak  acid  is  added  to  a  weak  base.  In  such 
cases  the  reaction  is  incomplete  and  reversible  and,  together  with  the 
neutralization  products,  both  free  acid  and  free  base  may  be  present. 
The  conditions  are  best  explained  by  citing  an  example  of  a  reversi- 
ble reaction  which  is  commonly  given  in  text-books  of  physical  chem- 
istry, namely,  the  reaction  between  ethyl-alcohol  and  acetic  acid. 
(Our  citation  is  taken  from  Philip's  "Physical  Chemistry,"  London, 
Arnold,  1910)  :  "When  one  gram  mol.  of  ethyl  alcohol  is  added  to 
one  gram  mol.  of  acetic  acid,  a  reaction  takes  place  which  results  in 

19  See  Cohn.  "Vortrage  f .  Artze  iiber  Physik.  Chem.,"  Engelman,  Leip- 
zig, 1901. 


TOXIN    AND    ANTITOXIN  119 

the  formation  of  ethyl  acetate  and  water;  the  reaction,  however,  is 
incomplete  and  stops  at  an  equilibrium  point  at  which  the  reaction 
mixture  contains  %  gram  mol.  alcohol,  %  gram  mol.  acid,  %  gram 
mol.  ethyl  acetate,  and  %  gram  mol.  water.  If,  on  the  other  hand, 
1  gram  mol.  of  ethyl  acetate  is  mixed  with  1  gram  mol.  of  water,  a 
reaction  sets  in  which  results  in  the  formation  of  ethyl  alcohol  and 
acetic  acid.  This  change  likewise  stops  in  equilibrium  at  a  point  at 
which  the  composition  of  the  reaction  mixture  is  the  same  as  that 
already  stated."  The  reaction  is  thus  reversible  and  may  be  written: 

C2H6OH  +  CHaCOOH  ^±1  CH3COOC2H6  +  H20 

Another  example  somewhat  simpler  and  more  easily  brought  into 
analogy  with  the  toxin-antitoxin  reaction  is  that  of  the  dissociation 
of  phosphorus  pentachlorid  into  phosphorus  bichlorid  and  chlorin 
(see  Alexander  Smith,  "General  Chemistry,"  Century  Company,  N. 
Y.,  1911,  p.  181). 

Here  the  reaction  takes  place: 

PC15  T-**  PCls  +  C12 

Chemical  equilibrium  is  reached  when  the  reaction  speed  is  the  same 
in  both  directions,  and  there  will  be  present,  at  equilibrium,  PC13, 
C12,  and  PC15.  Now  the  "Law  of  Mass  Action"  (Guldberg  & 
Waage)  states  that  reaction  goes  on  at  a  velocity  proportionate  to 
the  concentration  of  the  reacting  molecules.  It  is  plain,  therefore, 
that  at  the  point  at  which  the  reaction  takes  place  with  equal  veloci- 
ties in  both  directions,  that  is,  at  the  equilibrium  point,  a  very  defi- 
nite relation  of  molecular  concentrations  must  obtain,  and  this  rela- 
tion can  be  expressed  as  a  formula.  For  the  example  given  above 
this  may  be  written  as  follows : 

Cone.  PC13  X  Cone.  Cl       ^  ,  ,. 

7^ s™ =  K  (constant) 

Cone.  PCI 5 

This  formula  is  expressed  in  words  by  Alexander  Smith  as  follows: 
"If  we  change  the  amount  of  the  pentachlorid  placed  in  the  vessel, 
or  if  we  use  amounts  of  chlorin  and  trichlorid  which  are  not  equiv- 
alent, the  numerical  value  at  equilibrium  of  each  concentration  will, 
of  course,  be  different,  but  the  product  of  the  concentrations  of  tri- 
chlorid and  chlorin,  divided  by  the  concentration  of  the  pentachlorid, 
will  always  give  the  same  numerical  value  for  (K),  the  constant,  at 
the  same  temperature." 

Now  to  return  to  the  application  of  these  facts  to  the  neutraliza- 
tion of  toxin  by  antitoxin,  if  the  reaction  is  one  analogous  to  that  of 
a  strong  acid  and  alkali,  as  cited  above  in  the  case  of  H2SO4  and 


120  INFECTION    AND    RESISTANCE 

NaOH,  we  would  expect  a  complete  neutralization  of  one  by  the 
other,  multiple  for  multiple,  and  the  explanation  of  Ehrlich  based 
on  the  assumption  of  different  toxin  constituents,  of  varying  affin- 
ities, and  different  pharmacological  effects,  is  the  only  one  which  will 
account  for  the  experimental  results.  Assuming,  however,  that  the 
reaction  is  one  analogous  to  that  taking  place  between  a  weak  acid 
and  a  weak  base — such  as  boric  acid  and  ammonia — we  have  an  en- 
tirely different  state  of  affairs.  For  here  the  reaction  goes  on  to  a 
point  of  equilibrium,  and  in  mixtures  containing  equivalent  amounts 
of  the  weak  acid  and  the  base  there  will  be  present  the  reaction  prod- 
ucts and  also  small  amounts  of  unbound  free  acid  and  free  base. 
And  according  to  the  law  of  "Mass  Action,"  the  quantities  of  free 
acid  and  base  present  will  depend  entirely  on  the  masses  of  the 
reagents  put  together.  Thus,  for  each  particular  mixture  of  the  two, 
different  quantities  of  the  original  substances  will  be  found  uncom- 
bined,  and,  furthermore,  the  gradual  addition  of  one  to  the  other 
will  not  have  a  neutralizing  value  in  exact  proportion  to  the  amount 
added.  Were  the  toxin-antitoxin  reaction  analogous  to  such  chemical 
systems,  then  we  could  assume  that  every  mixture  of  the  two  sub- 
stances, whatever  the  relative  amounts,  would  contain  not  only  the 
united  toxin-antitoxin  molecule,  but  also  varying  amounts  of  disso- 
ciated free  toxin  and  free  antitoxin,  the  quantities  of  each  depending, 
according  to  the  law  of  mass  action,  upon  the  molecular  concentra- 
tions obtaining  in  the  individual  mixture.  This,  indeed,  is  the  con- 
ception of  toxin-antitoxin  union  formulated  by  Arrhenius  and  Mad- 
sen. 

Arrhenius  and  Madsen,20  21  bearing  in  mind  these  conditions,, 
made  comparative  studies  of  the  neutralization  of  tetanolysin  by  its 
antilysin  on  the  one  hand,  and  that  of  ammonia  by  boric  acid  on  the 
other.  Ammonia,  like  most  bases,  is  a  hemolytic  agent,  while  boric 
acid,  unlike  stronger  acids,  has  no  hemolyzing  properties.  For  this 
reason,  in  mixtures  of  the  two,  the  toxicity  is  proportional  to  the 
concentration  of  free  ammonia  (though,  as  Arrhenius  states,  "a  cor- 
rection must  be  made  for  the  lowering  action  of  the  ammonium  salt, 
as  indicated  by  experiments  on  this  action").  Because  the  reaction 
between  boric  acid  and  ammonia  is  reversible,  that  is,  the  salt  formed 
is  dissociated  by  the  hydrolytic  effect  of  the  water,  there  is  always 
present  a  slight  amount  of  free  ammonia  even  if  the  largest  possible 
quantities  (to  saturation)  of  boric  acid  are  added.  (See  Arrhenius, 
"Immunochem.,"  p.  174.)  The  curve  of  toxicity  indeed  descends 
as  more  boric  acid  is  added,  but  never  reaches  0. 

By  a  modification  of  the  formula  expressing  the  law  of  Mass 
Action,  Arrhenius  and  Madsen  could  calculate  the  amount  of  free 

20  Arrhenius   and  Madsen.      Zeitschr.   f.   physik.    Chem.,    44,    1903,    and 
Festschrift  Kopenhagen  Serum  Instit.,  1902. 

21  Arrhenius.     "Immunochemistry,"  Macmillan,  N.  Y.,  1907. 


TOXIN    AND    ANTITOXIN 


ammonia  present  in  a  series  of  mixtures  in  which  increasing  quanti- 
ties of  boric  acid  were  added  to  a  constant  quantity  of  ammonia,  and 


CURVE  EEPRESENTING  THE  NEUTRALIZATION  OF  TETANOLYSIN  BY  DIFFERENT  QUAN- 
TITIES OF  ANTITOXIN. 
Taken  from  Arrhenius,  "Immunochemistry,"  Macmillan,  1907,  p.  175. 

the  values  so  obtained  corresponded  with  much  accuracy  to  those  re- 
sulting from  measurements  of  toxicity  upon  red  blood  cells.  The  fol- 
lowing table  taken  from  Arrhenius  and  Madsen  illustrates  this : 


TOXICITY  (Q)  OF  0.1  N.  NH3  (1  EQUIVALENT)  WITH  N  EQUIVALENTS  OP  BORIC 
ACID.     (Taken  from  Arrhenius,  loc.  cit.  p.  176.) 


n    = 
Equivalents  of  boric 
acid    added 

Quantity  of  free 
ammonia  —  i.  e., 
toxicity  —  observed 

q  =  Ammonia 
toxicity  calculated 
from  formula 

Aq  oba. 

0 

100 

(100) 

0.17 

85 

79 

15 

0.33 

69 

64 

16 

0.67 

43 

42 

26:2  =  13 

1 

25 

27 

18:2  =    9 

1.33 

20 

18 

5:2] 

1.67 

13 

13 

7:2  }2.5 

2 

10 

10 

3:2  J 

Here,  in  the  last  column,  there  is  indicated  the  proportion  of 
toxicity  which  is  neutralized  by  the  successive  addition  of  %  of  an 
equivalent  of  boric  acid.  The  first  additions  lower  it  to  a  degree 
proportionate  to  the  amount  of  acid  added;  the  next  additions  neu- 
tralize it  to  a  much  slighter  degree,  and,  as  further  additions  are 


132  INFECTION    AND    RESISTANCE 

made,  each  successive  one  possesses  progressively  less  relative  neu- 
tralizing power  than  the  preceding. 

This,  it  is  plain,  is  closely  analogous  to  the  phenomena  observed 
by  Ehrlich  in  his  "Partial  Absorption"  method,  and  Arrhenius  con- 
cludes that  the  two  phenomena,  toxin-antitoxin  and  boric  acid-am- 
monia neutralization,  are  closely  analogous.  His  point  of  view  is 
further  strengthened  by  his  experiments  with  tetanolysin  and  its 
specific  antibody,  in  which  he  constructed  a  curve  similar  to  that 
given  for  boric  acid,  derived  a  formula  and  found  that  the  observed 
and  the  calculated  values  closely  coincided  for  various  mixtures  of  the 
two.  He  claims,  in  consequence,  that  the  phenomena  observed  by 
Ehrlich  should  not  be  interpreted  as  due  to  "partial  toxins" — toxoids 
or  toxons,  but  dependent  rather  upon  the  presence  of  varying  quan- 
tities of  free  toxin  dissociated  from  union  with  antitoxin  because  of 
the  reversibility  of  the  union. 

The  opinions  of  Arrhenius  and  Madsen  are  not  generally  ac- 
cepted. It  is  in  the  first  place  doubtful  whether  substances  like  toxin 
and  antitoxin,  which,  as  far  as  we  know  their  chemical  nature  at  all, 
belong  to  the  class  of  substances  spoken  of  as  colloids,  can  be  re- 
garded as  subject  to  the  laws  of  Mass  Action  in  their  reactions. 

Nernst 22  has  criticized  Arrhenius'  deductions  chiefly  on  the 
basis  of  their  assumption  of  the  reversibility  of  the  union  of  toxin 
and  antitoxin,  since  reversible  reactions  between  colloids,  though  not 
at  all  inconceivable,  have  so  far  not  been  definitely  shown.  Further- 
more, as  Nernst  states,  if  complete  reversibility  of  such  reactions 
were  possible  it  would  be  hard  to  understand  how  antitoxin  can  pro- 
tect the  animal  against  the  actions  of  toxin. 

Another  point  of  view  concerning  the  toxin-antitoxin  union  which 
has  been  gaining  ground  especially  through  the  work  of  Landsteiner 
and  his  pupils,  is  that  of  Bordet.23  Bordet  expresses  his  views  in 
the  following  way: 

I.  When  one  mixes  with  a  certain  quantity  of  toxin  an  amount 
of  antitoxin  which  is  insufficient  to  produce  a  complete  neutraliza- 
tion, the  molecules  of  antitoxin  are  not  taken  up  by  a  definite  frac- 
tion of  the  toxin  molecules,  satisfying  the  affinities  of  these  entirely 
while  other  units  remain  intact ;  on  the  contrary,  the  antitoxin  mole- 
cules distribute  themselves  equally  upon  all  the  toxin  molecules  pres- 
ent, and  these  are  therefore,  all  of  them,  partially  saturated,  and 
lose  proportionately  a  part  of  their  initial  toxicity.     One  could  say 
that  there  is  an  attenuation  of  the  toxin  since  there  is  a  formation  of 
a  less  poisonous  complex. 

II.  The  symptoms  of  poisoning  produced  by  such  a  complex  in- 
jected into  animals  or  placed  in  contact  with  sensitive  cells  cannot  be 

22  Cited  from  Landsteiner  in  "Kolle  u.  Wassermann  Handbuch,"  2d  Ed., 
Vol.  5. 

23  Bordet.     Ann.  de  I'Inst.  Past.,  Vol.  17,  1903. 


TOXIN    AND    ANTITOXIN  123 

identical  with  those  which  would  be  produced  by  a  fully  saturated 
mixture  of  toxin  and  antitoxin,  or  by  intact  toxin. 

III.  Between  these  two  extremes,  free  toxin  and  entirely  neu- 
tralized toxin,  one  can  imagine  many  transitions,  progressive  stages 
of  attenuation.  Every  time  that  one  mixes  toxin  and  antitoxin  in 
the  same  way  one  attains  the  same  degree  of  attenuation. 

Briefly  put,  this  means  that  Bordet  estimates  toxin-antitoxin 
combinations  of  different  degrees  of  toxicity  as  representing  differ- 
ent stages  in  the  completeness  of  the  saturation  of  the  individual 
toxin  units.  When  10  parts  of  toxin  are  added  to  1  part  of  anti- 
toxin, the  result,  according  to  him,  would  not  be  such  that  1  part  is 
neutralized  by  1  part  of  antitoxin,  leaving  9  parts  of  toxin  free.  He 
assumes  rather  that  each  unit  of  toxin  is  attenuated  by  the  absorp- 
tion of  T^th  of  a  part  of  antitoxin.  He  compares  this  process  to  the 
action  of  iodin  upon  starch.  Starch  can  absorb  variable  quantities 
of  iodin  and,  according  to  the  amount  taken  up,  is  colored  slightly 
or  deeply  blue.  This  mode  of  action  is  common  to  most  staining 
processes.  The  substance  that  is  stained  fixes  varying  quantities  of 
coloring  matter  and  the  coloring  matter  does  not  limit  itself  to  a 
definite  fraction  of  the  substance  stained  but  distributes  itself  equally 
to  the  material,  coloring  it  slightly  or  deeply,  in  its  entirety,  accord- 
ing to  the  relative  amount  of  color  added.  We  will  see  later  that 
there  are  many  reasons  for  regarding  other  antigen-antibody  com- 
binations as  following  similar  laws  of  proportion. 

Bordet  and  others  speak  of  this  point  of  view  as  the  "Absorption 
Theory,"  and  Biltz,  in  studying  this  point  of  view  by  physical 
methods,  comes  to  the  conclusion  that  the  observed  figures  of  the 
quantitative  relations  between  toxin  and  antitoxin  in  the  process  of 
neutralization  are  fairly  consistent  with  the  values  to  be  expected  if 
the  process  were  actually  an  absorption  phenomenon. 

A  curious  occurrence  which  seems  to  bring  the  toxin-antitoxin 
reactions  close  to  colloidal  reactions  in  general  is  that  which  is 
known  as  the  "Danysz  24  Effect"  or  as  the  "Bordet  25-Danysz  Phe- 
nomenon." Danysz  discovered  that  when  ricin  or  diphtheria  toxin 
were  brought  into  contact  with  their  Homologous  antibodies  the  de- 
gree of  neutralization  depended  upon  the  manner  in  which  the  two 
were  put  together.  When  the  toxin  was  added  to  the  antitoxin  in 
two  fractions,  a  considerable  time  being  allowed  to  elapse  between 
the  additions,  the  final  mixture  was  much  more  toxic  than  when  the 
total  amount  was  added  at  once.  In  other  words,  although  both 
mixtures  contained  exactly  the  same  quantities  of  the  two  reacting 
substances,  nevertheless  the  amount  of  toxin  left  free  varied  in  the 
two  cases,  according  to  the  speed  with  which  they  had  been  put  to- 

24  Danysz.     Ann.  de  I'Inst.  Past.,  Vol.  16,  1902. 

25  Bordet.     Ann.  de  I'Inst.  Past.,  Vol.  17,  1903. 


INFECTION    AND    RESISTANCE 

gether.  This  was  confirmed  in  1904  by  von  Dungern  26  for  diph- 
theria toxin,  and  Craw  27  was  able  to  observe  it  in  the  case  of  mega- 
theriolysin  and  its  antilysins. 

Von  Dungern  interpreted  this  in  the  sense  of  Ehrlich,  by  assum- 
ing it  to  be  due  to  what  he  calls  "epitoxonoids."  This  epitox- 
onoid  he  assumes  to  be  a  constituent  of  toxic  broth,  which  has  still 
less  affinity  for  antitoxin  than  the  toxon.  It  can  combine  with  diph- 
theria antitoxin,  but  not  until  all  the  true  toxin  is  bound.  However, 
when  it  is  once  united  with  diphtheria  antitoxin  it  is  not  very  easily 
displaced  from  the  union,  especially  when  a  considerable  time  has 
elapsed  since  the  union.  Therefore,  he  thinks,  when  the  toxin  is 
added  to  the  antitoxin  in  two  fractions,  this  epitoxonoid  is  bound  and 
keeps  the  toxin,  which  is  added  later,  out  of  combination.  Whereas 
if  the  toxic  broth  is  added  as  a  whole,  it  is  the  epitoxonoid  which  is 
left  unbound.  This  explanation  of  von  Dungern's  may  be  looked 
upon  as  an  ingenious  refinement  of  the  reasoning  introduced  by  Ehr- 
lich into  this  field.  As  a  matter  of  fact  reactions  similar  to  the 
Danyz  phenomena  have  been  very  commonly  observed  in  the  reac- 
tions between  various  colloids. 


THE  SIDE-CHAIN  THEORY 
Mechanism  of  Antibody  Formation 

The  discovery  of  antitoxins  in  the  blood  serum  of  toxin-immune 
animals  by  Behring  and  his  collaborators  furnished  a  point  of  new 
departure  for  the  investigation  of  the  phenomena  of  immunity,  and 
Ehrlich' s  work  upon  the  nature  of  the  reaction  between  toxin  and 
antitoxin,  both  in  vitro  and  in  the  animal  body,  firmly  established 
that  the  protective  effect  of  the  latter  was  one  of  direct  neutraliza- 
tion, and  not,  as  at  first  supposed,  one  of  toxin  destruction  or  of 
indirect  influence  through  the  mediation  of  the  body  cell.  As  we 
have  seen,  moreover,  it  was  quickly  noted  that  these  reactions  Were 
strictly  specific  in  that  an  antitoxin  produced  with  any  one  of  the 
known  toxins  reacted  solely  with  this  one  to  the  exclusion  of  all 
others.  All  these  facts  were  of  the  utmost  practical  importance  and 
gave  hope  of  ultimate  extensive  therapeutic  application,  a  hope 
which  has,  in  part,  been  realized. 

The  physiological  mechanism  by  which  these  phenomena  were 
brought  about,  however,  was,  and  is,  to  a  great  extent  still,  a  mys- 
tery, and  a  most  extensive  and  painstaking  series  of  researches  has 
occupied  itself  with  its  elucidation. 

26  Von  Dungern.    Deutsche  med.  Woch.,  30,  1904. 

27  Craw.     Jour.  Hyg.,  Vol.  7,  1907. 


TOXIN    AND    ANTITOXIN  125 

When  we  consider  the  invariable  production  of  a  specific  anti- 
toxin in  response  to  the  treatment  of  an  animal  with  a  toxin  it  is 
but  natural  that  Buchner  and  others  should  have  at  first  assumed 
that  the  antitoxin  is,  in  each  case,  a  product  obtained  by  the  action 
of  the  body  tissues  from  the  toxin  itself.  While  difficult  to  refute 
at  a  time  when  little  was  known  of  the  laws  of  antitoxin  production 
and  of  quantitative  relationships,  such  an  assumption  is  entirely  un- 
tenable in  the  light  of  more  recent  knowledge.  We  now  know  that 
such  a  simple  conversion  of  toxin  into  antitoxin  cannot  explain  the 
phenomenon  because  the  amount  of  antitoxin  incited  in  the  immu- 
nized animal  is  out  of  all  proportion  great  in  comparison  with  the 
amount  of  toxin  injected.  Thus  Knorr  28  has  found  that  100,000 
units  of  antitoxin  may  be  produced  by  the  injection  of  the  toxin 
equivalent  of  one  unit.  Moreover  the  discovery  by  Salomonsen  and 
Madsen 29  that  pilocarpin  injections  will  increase  the  amount  of 
antitoxin  produced  by  an  animal  distinctly  pointed  to  the  likelihood 
of  the  participation  of  the  general  physiological  activities  of  an 
immunized  subject  in  the  production  of  antibodies.  Unquestionable 
proof  of  this  was  also  brought  by  the  experiments  of  Roux  and  Yail- 
lard,30  in  which  antitoxin  production  in  immunized  animals  con- 
tinued even  after  the  entire  volume  of  blood  had  been  removed  by 
repeated  bleeding.  This  observation  points  distinctly  to  the  direct 
secretion  of  antibodies  by  the  tissue  cell,  in  the  nature  of  what  has 
been  termed  by  Roux 31  an  "internal  secretion.77  And  it  is  this 
activity  of  the  body  cell  in  the  production  of  antibodies  which  forms 
the  fundamental  premise  from  which  the  now  classical  "Side-Chain 
Theory7'  of  Ehrlich  takes  its  departure. 

In  order  to  approach  this  theory  logically  it  will  be  of  advantage 
to  consider  briefly  the  general  subject  of  the  assimilation  of  food- 
stuffs and  other  substances  distributed  by  the  circulation  to  the  cells 
of  the  animal  body.  For,  as  Ehrlich  has  expressed  it,  "The  Reac- 
tions of  Immunity,  after  all,  represent  only  a  repetition  of  the 
processes  of  normal  metabolism,  and  their  apparently  wonderful 
adjustment  to  new  conditions  is  only  another  phase  of  'Uralter 
Protoplasma  Weisheit.7  77  32  It  is  impossible  to  conceive  the  nutri- 
tion of  body  cells  without  assuming  that  the  assimilable  nutritive 
substances  come  into  physical  and,  eventually,  chemical  relationship 
with  the  protoplasm  of  the  nourished  cell.  Considering  the  large 
variety  of  substances  which  may  thus  be  brought  into  contact  with 
cells  in  the  course  of  normal  and  abnormal  metabolism,  the  body  cell, 

28  Knorr.     Munch,  med.  Woch.,  1898,  pp.  321,  362. 

29  Salomonsen  and  Madsen.    Ann.  de  I'Inst.  Past.,  Vol.  12,  1898. 

30  Roux  and  Vaillard.    Ann.  de  I'Inst.  Past.,  Vol.  7,  1893. 

31  Roux.     Ref .  in  Semaine  Medicale,  1899. 

32  Ehrlich.     Introduction    to    "Gesammelten    Arbeiten,"    Berlin,    Hirsch- 
wald,  1904. 


126  INFECTION    AND    RESISTANCE 

chemically  considered  as  a  complex  of  enormous  molecules,  must 
possess  a  correspondingly  great  variety  of  atom  groups,  by  means  of 
which  it  can  unite  with  these  substances  to  assimilate  them  and  make 
them  a  part  of  its  own  protoplasm.  In  order  to  enter  into  similar 
relationship  with  toxins  and  other  antigens,  then,  it  is  only  logical 
to  suppose  that  the  cell,  in  the  same  way,  unites  chemically  with  the 
antigenic  substance,  and  either  assimilates  it  without  sustaining 
harm,  as  in  the  case  of  non-poisonous  complexes,  or  is  injured  in  the 
process,  as  in  the  case  of  the  poisons. 

The  living  cell,  from  this  point  of  view,  is  conceived  as  consist- 
ing of  a  central  chemical  nucleus,  the  "Leistungskern,"  more  or  less 
stable,  in  that  the  specialized  tissue  function  is  dependent  upon  it, 
and  a  manifold  variety  of  "side  chains,"  or  atom  groups  by  means 
of  which  it  can  enter  into  relationship  with  the  nutritive  and  other 
materials  carried  to  it  by  the  body  fluids.  The  latter  term,  "side- 
chains,"  is  taken  from  the  nomenclature  of  chemistry,  and,  although 
the  analogy  is  a  loose  one,  it  serves  satisfactorily  to  elucidate  Ehr- 
lich's  meaning.  Thus  we  may  conceive  the  "Leistungskern"  as  the 
central  carbon  ring  of  any  compound  of  the  Benzol  series,  as,  for 
instance,  in  salicylic  acid  in  which  the  hydrogen  atoms,  the  hydroxyl, 

OH  OH 


° 


H— C        C— COOH  (  J  C02CH, 

H— C        C— H 

\    / 

Methyl  salicylate 
Salicylic  acid 


| 


and  the  acid  radicles  represent  "side  chains."  By  means  of  the  lat- 
ter the  compound  can  enter  into  relation  with  other  substances,  as, 
for  instance,  with  CH3  in  the  formation  of  methyl  salicylate. 
Graphically,  though  this  analogy  formulates  Ehrlich's  fundamental 
conception,  it  must  not  be  taken  as  too  literally  representing  the 
existing  conditions,  since,  in  actual  metabolic  interchange,  there  is 
an  infinite  variety  of  possible  "side-chain"  groups  ;  for  we  are  deal- 
ing with  an  enormous  number  of  assimilable  substances,  most  of 
them  of  chemically  unknown  constitution.  The  cell,  therefore,  is 
looked  upon  as  an  active  chemical  complex,  retaining  its  own  peculiar 
functional  characteristics  by  reason  of  the  "Leistungskern,"  but 
constantly  getting  rid  of  waste  products  and  entering  into  new  union 
with  extraneous  materials  by  virtue  of  its  "side  chains."  These  side 
chains,  because  of  their  "receiving"  function,  are  spoken  of  by  Ehr- 
lich  as  cell  "receptors." 


11 

J.1 

li 

f 

9! 


I 


128  INFECTION    AND    RESISTANCE 

That  the  chemical  structure  of  certain  bodies  determines  their 
ability  to  enter  into  relation  with  cell  derivatives  such  as  enzymes 
is,  of  course,  a  fact  well  established  by  experiment  and  explains  the 
specific  action  of  bacterial  and  other  ferments  upon  certain  sub- 
stances to  the  exclusion  of  others.  Thus  Pasteur  noted  the  fact  that 
bacterial  ferments  could  decompose  dextrorotatory  tartaric  acid 
while  they  did  not  affect  the  levorotatory  variety,  and  Emil 
Fischer  33  showed  that  only  those  carbohydrates  possessing  6  and  9 
carbon  atoms  were  subject  to  fermentation  by  yeasts,  and  of  these 
only  the  ones  belonging  to  the  "d"  series,  observations  which,  by 
demonstrating  the  relationship  between  these  active  agents  of  extra- 
cellular digestion,  and  the  stereochemical  configuration  of  the  mole- 
cules acted  upon,  lend  much  support  to  the  logic  of  Ehrlich's  con- 
tentions. 

Moreover,  the  recent  experiments  upon  the  growth  of  tissues  in 
plasma  outside  of  the  animal  body  in  which  cartilage  cells  produce 
cartilage,  kidney  cells,  etc.,  have  shown  that,  given  the  same  nutri- 
tive materials,  the  cells  themselves  must  command  a  certain  selective 
power  in  the  choice  of  these  materials,  which  can  only  depend  upon 
a  specific  element  in  the  structure  of  the  cell  receptors.  As  Fischer 
has  expressed  it  for  fermentation,  the  ferment  must  possess  an  atom 
group  which  fits  into  some  group  of  the  fermentable  substance  as  a 
"key  does  into  a  lock,"  an  analogy  which  is  equally  applicable  to 
Ehrlich's  conception  of  the  relation  of  the  "side  chain"  to  a  nutri- 
tive molecule. 

Now  the  toxins  and  other  antigens  are,  all  of  them,  so  far  as  we 
know,  complex  chemical  substances,  derivatives  of  animal  and  vege- 
table cells,  and,  for  this  reason,  should  have  much  in  common  with: 
the  materials  available  for  nutrition.     It  is  not  strange,  therefore,; 
that,  coming  into  contact  with  the  cells  of  the  body  during  the  acci- 
dents of  disease  or  other  abnormal  conditions,  they  should  find  re-% 
ceptors  by  means  of  which  they  can  combine  with  the  cell.     Under 
the  ordinary  conditions  of  nutrition  a  suitable  particle  taken  up  by 
the  cell  in  this  way  would  be  assimilated  and  the  receptor  either 
freed  for  further  use  or  regenerated  for  the  further  absorption  of  j 
similar  substances,  by  virtue  of  a  mechanism  delicately  coordinated 
to  the  needs  of  cell-nutrition.     In  the  case  of  the  absorption  of  sub- 
stances belonging  to  the  class  of  antigens,  however,  foreign  proteins 
difficult  of  assimilation,  or  of  toxins  even  directly  harmful,  the  re- 
ceptors occupied  by  these  substances  are  rendered  useless  to  the  cell, 
and,  if  the  cell  continues  to  live,  must  be  regenerated.     If  the  degree; 
of  poisoning  or  the  amount  of  other  antigen  introduced  has  been 
extremely  slight,  this  regeneration  may  possibly  take  place,  as  in 
the  course  of  nutritive  processes,  without  further  disturbance.     If, 
however,  the  amounts  of  antigen  are  greater  than  this,  or  are  repeat- 

33  See  Oppenheimer,  "Die  Fermente,"  Vol.  1. 


TOXIN    AND    ANTITOXIN  129 

edly  thrust  upon  the  cell,  the  process  of  regeneration  may  be  not  only 
sufficient  to  compensate  for  the  loss  of  the  eliminated  receptors,  but 
may  follow  the  general  law  of  overcompensation,  formulated  by  Wei- 
gert,  and  receptors  of  the  variety  occupied  by  the  antigen  are  pro- 
duced in  excessive  number. 

Here  again  Ehrlich  has  called  analogy  to  his  aid,  and  has  taken 
his  conception  of  "overcompensation"  from  the  well-known  phe- 
nomena of  pathological  anatomy  where,  for  instance,  in  the  restora- 
tion of  cellular  elements  after  injury,  there  is  often  an  overpro- 
duction of  granulation  tissue,  far  beyond  the  needs  of  simple  healing. 

Thus  the  restitution  of  cell  receptors,  if  sufficiently  stimulated 
by  large  quantities  or  repeated  administration  of  the  antigen,  far 
exceeds  the  quantity  normal  to  the  cell,  and  may  proceed  to  such  a 
degree  that  the  cell,  becoming  as  it  were  "top-heavy"  with  these 
elements,  sloughs  them  off  into  the  surrounding  lymph  and  blood, 
where  they  circulate  as  free  receptors.  These  free  receptors  then, 
having  specific  affinity  and  combining  power  for  the  antigen  which 
incited  their  production,  unite  with  subsequently  introduced  antigen 
in  the  blood  stream,  diverting  it  from  the  cells  themselves,  and,  in 
the  case  of  the  variety  of  antigens  spoken  of  as  toxins,  this  union 
with  the  free  receptors  in  the  blood  stream  would  serve  to  protect 
the  cells  from  harm,  exerting  thereby  an  antitoxic  action. 

The  antibodies  appearing  in  the  blood  of  immunized  animals, 
therefore,  represent  atom  complexes,  normally  parts  of  the  body  cells 
and  concerned  in  the  metabolic  processes,  but  now  produced  in  ex- 
cess and  extruded  into  the  body  fluids  under  the  influence  of  the 
stimulation  of  immunization.  The  very  substances,  as  Behring  has 
put  it,  which  make  possible  the  poisoning  of  the  cell  by  the  toxins  be- 
come protective  wThen,  detached  from  the  cell,  they  circulate  in  the 
blood.  Thus  the  theory,  beside  explaining  the  causes  leading  to  anti- 
body formation,  offers  a  plausible  reason  for  the  relatively  strict 
specificity  observed  in  antibody-antigen  reactions. 

Formulated  in  direct  connection  with  the  investigations  upon 
toxins  and  antitoxins,  the  side-chain  theory  has  been  extended  by 
Ehrlich  and  his  associates  to  all  known  phases  of  antibody-antigen 
reactions.  The  differences  in  the  nature  and  complexity  of  various 
antigens  would  naturally  necessitate  variation  in  the  receptors  capa- 
ble of  assimilating  them,  and  these  receptors,  appearing  subsequently 
in  the  blood  as  antibodies,  must,  of  necessity,  differ  from  each  other. 
On  this  basis  Ehrlich  has  conceived  of  three  main  varieties  or  "or- 
ders" of  receptors  or  "haptines,"  as  he  calls  them.  Of  these  the 
simplest  are  those  of  the  first  order  which  attach  to  the  toxins,  and 
by  over-regeneration  appear  in  the  blood  stream  as  antitoxins.  Those 
of  the  second  order,  adapted  to  the  assimilation  of  more  formidable 
protein  molecules,  are,  of  necessity,  of  greater  structural  complex- 
ity, appearing  in  immunized  animals  as  the  agglutinins  and  precip- 


130  INFECTION    AND    RESISTANCE 

it  ins,  while  those  of  the  third  order,  dependent  upon  the  coopera- 
tion of  alexin  or  complement,  for  proper  functionation,  appear  as 
the  cytotoxins  or  lysins.  The  detailed  structure  of  these  various 
haptines  will  be  discussed  in  connection  with  other  considerations 
dealing  with  their  special  reactions. 

Limiting  ourselves,  for  the  present,  to  a  broad  consideration  of 
the  theory  as  a  whole,  it  may  be  briefly  recapitulated  as  follows: 
Toxins  or  other  antigens,  in  order  to  exert  any  influence  upon  the 
animal  body,  must  enter  into  chemical  relationship  with  the  cells. 
This  they  do  by  virtue  of  union  with  chemical  units  or  atom  groups 
of  the  cells,  spoken  of  as  "side  chains."  These  side  chains  or  recep- 
tors, thrown  out  of  function  by  this  union,  and  necessary  for  the 
metabolic  processes  of  the  cell,  are  regenerated,  and  under  the  influ- 
ence of  repetition  of  this  process  are  produced  in  excess,  to  such  a 
degree  that  they  are  eventually  thrown  off  by  the  cells  and  enter  the 
circulation  as  antibodies.  Thus  far  the  theory,  comparing  the  union 
of  antigen  with  cells  to  the  processes  of  nutrition,  is  eminently  log- 
ical and  likely,  necessitating  the  assumption  of  over-regeneration  as 
the  only  criterion  not  directly  amenable  to  experimental  proof. 

That  the  antigen  can  be  bound  by  the  body  cells  has  been  vari- 
ously shown  in  a  large  number  of  investigations,  some  of  which  have 
been  reviewed  in  our  section  on  the  action  of  bacterial  poisons.  We 
have  there  seen  that  Donitz  demonstrated  the  rapid  disappearance 
of  tetanus  and  diphtheria  toxins  from  the  circulation  of  susceptible 
animals,  and  that  conversely  Metchnikoff  showed  that  the  poison  may 
persist  unabsorbed  and  unchanged  for  weeks  and  months  in  the 
blood  of  such  insusceptible  animals  as  the  turtle  and  the  lizard, 
facts  which  furnish  indirect  evidence  of  the  absorption  of  the  toxins 
by  the  body  cells.  More  direct  evidence  has,  of  course,  been  possible 
in  the  test  tube  experiments  with  hemolytic  and  other  cell  poisons 
where  a  directly  specific  combination  between  antigen  and  antibody 
has  been  easily  demonstrable.  Thus,  in  his  earlier  experiments  with 
spider  poison,  Sachs  was  able  to  show  that  rabbit  erythrocytes,  which 
are  sensitive  to  the  poison,  could  absorb  it  out  of  solution,  while  dog 
and  other  corpuscles,  which  were  insusceptible  to  the  poison,  did  not 
bind  or  absorb  it.  This  can  be  easily  demonstrated  for  many  anti- 
gens and  antibodies  and  may  be  accepted  as  a  fact. 

This  point  established,  and  repeatedly  confirmed,  and  the  origin 
of  antitoxins  from  the  cells  of  the  body  having  been  rendered  likely 
by  the  experiments  of  Salomonsen  and  Madsen,  and  by  those  of  Roux 
and  \7aillard  just  cited,  it  would  follow,  by  the  theory  of  Ehrlich, 
that  we  should  find  the  site  of  antibody  production  in  the  very  cells 
which  possessed  specific  affinity  (receptors)  for  the  antigen.  This 
question  has  been  variously  investigated,  chiefly  in  the  case  of  the 
toxins  and  antitoxins,  since  this  phase  of  the  subject  is  most  easily 
amenable  to  experiment.  It  will  be  remembered  also  that  Wasser- 


TOXIN    AND    ANTITOXIN  131 

mann  and  Takaki  discovered  that  emulsions  of  the  tissue  of  the  cen- 
tral nervous  system  of  rabbits  and  guinea  pigs,  shown  by  Meyer  and 
Ransom  and  others  to  be  the  special  points  of  attack  for  tetanus  toxin, 
possessed  the  power  of  neutralizing  this  poison  in  vitro,  while  emul- 
sions of  spleen  Jddney  and  other  organs  had  no  such  effect.  They 
assumed  from  this  that  the  poison  was  fixed  by  cell  receptors,  ante- 
cedents of  antitoxin  in  the  sense  of  Ehrlich.  Kempner  34  35  made 
similar  observations  with  botulinus  toxin  and  further  confirmation 
has  been  derived  from  experiments  like  those  of  Blumenthal,36  who 
found  that  the  toxin  was  neutralized  by  the  brain  tissue  of  susceptible 
animals  but  showed  conversely  that  the  brain  substance  of  the 
chicken,  an  animal  but  slightly  susceptible  to  tetanus,  possessed  little 
or  no  neutralizing  power.  Similar  results  were  obtained  by  Metchni- 
koff  in  the  cases  just  cited. 

The  great  importance  of  these  experiments  lies  not  only  in  show- 
ing that  body  cells  may  absorb  the  toxins,  but  that  there  is  direct 
relationship  between  the  susceptibility  of  tissue  and  the  toxin-bind- 
ing properties.  Furthermore  the  facts  demonstrated  by  Metchnikoff 
that  no  antitoxin  was  produced  by  those  animals  (turtle,  lizard)  in 
which  the  tissues  had  no  power  of  fixing  poison  and  which  are  con- 
sequently insusceptible,  furnish  powerful  evidence  in  favor  of  Ehr- 
lich's  view. 

It  becomes  of  great  importance,  therefore,  to  determine  whether 
in  the  case  of  the  fixation  of  tetanus  toxin  by  the  brain  cells  the 
union  between  cell  and  toxin  is  a  specific  and  chemical  one  compar- 
able in  every  way  to  the  union  of  toxin  with  antitoxin. 

Metchnikoff,  in  spite  of  his  results  in  the  experiments  just  cited, 
objected  to  this  interpretation  on  the  ground  that  although  the  brain 
emulsion  of  a  guinea  pig  neutralized  tetanus  toxin  in  vitro,  the  in- 
jection of  the  toxin  into  such  an  animal,  subdurally,  produced  the 
disease.  This  can  hardly  be  regarded  as  a  valid  argument  against 
Wassermann's  interpretation,  since  the  very  premises  of  the  Ehrlich 
theory  require  that  these  neutralizing  elements,  when  still  attached 
to  the  living  cell,  as  "sessile"  receptors,  are  the  cause  of  the  poison- 
ing, since  they  serve  to  "unlock"  the  cell  to  the  entrance  of  the  toxin. 
Similar  objections  on  the  part  of  Metchnikoff37  were  based  on  some 
of  his  own  experiments,  as  well  as  on  those  of  Courmont  38  39  and 
Doyen,  in  which  it  was  found  that  the  poison  disappears  but  slowly 
(in  2  to-  3  months)  from  the  circulation  of  frogs,  and  the  brain  cells 
show  hardly  any  toxin  neutralization  in  vitro,  whereas  these  animals 

34  Kempner  and  Pollak.     Deutsche  med.  Woch.,  1897,  No.  23,  p.  505. 

35  Kempner  and  Shepilewsky.    Zeitschr.  f.  Hyg.,  Vol.  36,  1901,  p.  1. 

36  Blumenthal.     Deutsche  med.  Woch.,  1898,  No.  12,  p.  185. 

37  Metchnikoff.    Ann.  de  I'lnst.  Past.,  Vol.  12,  1898. 

38  Courmont  and  Doyerf.     Arch,  de  Physiol.,  1893. 

39  Courmont  and  Doyen.     Compt.  rend,  de  la  soc.  de  biol,  1893. 


132  INFECTION    AND    RESISTANCE 

can  be  rendered  tetanic  if  they  are  warmed  to  25°  to  30°  C.  Further 
work,  however,  by  these  authors  as  well  as  by  Morgenroth 40  has 
satisfactorily  cleared  up  this  difficulty.  As  a  matter  of  fact,  tetanus 
poison  disappears  more  rapidly  (that  is,  is  bound  by  the  cells  more 
rapidly)  from  the  circulation  of  frogs,  if  the  frogs  are  warmed  to 
30°  C.  or  more.  Furthermore,  if  the  toxin  is  injected  into  these 
animals,  and  they  are  kept  at  low  temperatures,  no  disease  results, 
but  if  they  are  then  warmed  up  to  the  temperature  stated,  they  grad- 
ually succumb  to  the  disease.  Morgenroth  has  shown  that  the  ap- 
parently anomalous  behavior  of  frogs  in  this  respect  is  actually  a 
question  of  temperature.  At  low  temperatures  the  poison  is  bound, 
though  with  extreme  slowness,  but  the  toxophore  group  of  the  toxin 
does  not  functionate.  When  the  animals  are  warmed,  not  only  does 
the  binding  proceed  more  rapidly,  but  the  toxophore  group  becomes 
active.  He  thus  not  only  has  answered  Metchnikoff's  objections  to 
Ehrlich's  theory  on  this  ground,  but  has  furnished  an  additional  in- 
direct confirmation  of  the  dual  constitution  of  toxin,  that  is,  its 
constitution  of  a  haptophore  and  a  toxophore  atom  group,  suggested 
by  Ehrlich  in  his  diphtheria-toxin  analysis. 

There  is  apparently,  then,  a  strong  absorption  of  tetanus  toxin 
by  the  brain  and  nervous  tissue  of  all  animals  which  are  susceptible 
to  the  poison,  an  absorption  which  amounts,  as  we  have  seen,  to  neu- 
tralization, the  brain  emulsion  acting  like  antitoxin  when  mixed 
with  the  toxin  before  injection,  as  in  Wassermann's  and  Takaki's 
experiments. 

A  serious  objection  has  been  brought,  however,  to  the  assumption 
that  this  binding  can  be  identified  in  its  nature  with  the  similar  bind- 
ing of  toxin  by  antitoxin,  and  a  number  of  authors  have  claimed  that 
the  binding  by  the  brain  is  not  a  binding  by  specific  receptors,  but 
an  accidental  property  due  to  the  presence  of  some  fortuitous  fixing 
substance  in  the  central  nervous  system.  Besredka  41  showed,  for 
instance,  that  the  brain  of  susceptible  animals  could  bind  much  more 
toxin  than  it  could  actually  neutralize,  and  that,  if  antitoxin  was 
added  to  a  brain  emulsion  previously  saturated  with  the  toxin,  the 
toxin  is  removed  from  its  combination  with  the  brain  cells  and  these 
again  regain  their  original  absorbing  property.  These  experiments 
would  seem  to  point  to  a  difference,  especially  in  regard  to  affinity 
and  firmness  of  union  between  the  nature  of  the  combination  between 
toxin  and  brain  emulsion  on  the  one  hand,  and  toxin  and  antitoxin 
•on  the  other.  This,  of  course,  would  prove  a  serious  obstacle  to  the 
interpretation  of  the  binding  of  toxin  by  susceptible  cells  in  the  sense 
of  Ehrlich,  as  depending  as  it  were  upon  union  with  specific  recep- 
tors, or,  as  they  might  be  termed,  "sessile"  antitoxin.  Moreover,  to 
strengthen  such  objections  to  this  point  of  view,  the  work  of  Land- 

40  Morgenroth.     Arch,  internat.  de  Pharm.,  Vol.  7,  1900,  pp.  265-272. 

41  Besredka.    Ann.  de  VInst.  Past.,  Vol.  17,  1903,  p.  138. 


TOXIN    AND    ANTITOXIN  133 

steiner  and  v.  Eisler  42  has  brought  out  the  fact  that  extraction  of 
brain  tissue  with  ether  materially  reduces  its  toxin-binding  powers 
by  removing  fatty  or  lipoidal  substances,  such  as  cholesterin  and 
lecithin.  And  it  has  indeed  been  confirmed  that  lipoids  can  possess, 
in  many  instances,  binding  properties  not  only  for  toxins  but  for 
other  forms  of  antibodies.  On  the  basis  that  at  least  a  part  of  the 
toxin  absorption  by  brain  emulsions  depends  upon  such  lipoidal  fixa- 
tion, the  results  of  Besredka  are  readily  explained,  but  were  this  the 
sole  cause  of  toxin  fixation  by  these  tissues  it  would  indeed  be  diffi- 
cult to  interpret  the  phenomenon,  with  Wassermann  and  Takaki,  in 
support  of  Ehrlich's  theory.  For,  without  going  into  further  refine- 
ments, the  fact  of  the  probable  proteid,  certainly  not  lipoidal,  nature 
of  the  antitoxins,  discussed  in  a  previous  section,  would  alone  serve 
to  distinguish  the  two  modes  of  toxin  fixation. 

However,  a  number  of  facts  have  been  ascertained  which  show 
that,  although  the  lipoids  play  some  part  in  the  antitoxic  action  of 
brain  cells,  they  do  not  by  any  means  account  for  the  entire  process. 
In  the  first  place  it  is  found  that  the  heating  of  brain  emulsions  al- 
most completely  removes  their  power  to  bind  the  toxin,  while  no 
such  reduction  of  the  fixative  property  follows  the  heating  of  lipoids 
like  cholesterin  or  lecithin.  The  experiments  of  Marie  and  Tif- 
feneau  43  have  done  much  to  clear  up  the  confusion  regarding  this 
point.  They  determined  that  the  "lipoidal  binding"  constituted  only 
about  one-tenth  of  the  total  binding  power  of  the  brain  emulsions, 
by  showing  in  the  first  place  that  only  one-tenth  of  the  total  was  left 
after  heating,  and  that  all  but  one-tenth  could  be  destroyed  by  sub- 
jecting the  tissue  to  the  action  of  proteolytic  enzymes.  It  appears 
from  this  that  a  large  part,  at  any  rate,  of  the  toxin  fixation  of  the 
brain  tissues  is  dependent  upon  substances  of  an  albuminous  nature, 
a  smaller  but  definite  part  being  dependent  upon  fixation  by  lipoids, 
a  phenomenon  entirely  apart  from  the  former  in  underlying  princi- 
ples. This  would,  it  seems,  both  justify  the  original  interpretation 
of  Wassermann  and  still  explain  the  apparently  contradictory  results 
of  Besredka  and  others. 

42  Landsteiner  and  v.  Eisler.     Centralbl.  f.  Bakt.,  Vol.  39,  1905. 

43  Marie  and  Tiiteneau.    Ann.  de  I'Inst.  Past.,  Vol.  22,  pp.  289  and  644, 
1908. 

10 


CHAPTER   VI 

THE  BACTEEICIDAL  PKOPERTIES  OF  BLOOD  SEBUM, 
CYTOLYSIS,    AND    SENSITIZATION 

IN  spite  of  the  profound  physiological  alteration  of  the  animal 
body  which  is  implied  by  the  acquisition  of  immunity  against  any 
particular  infection,  we  have  seen  that  no  anatomical  or  histological 
changes  in  the  organs  and  tissues  accompany  such  alteration.  The 
same  is  true  of  the  difference  between  animals  of  different  species, 
in  which  the  most  marked  variation  in  resistance  against  any  given 
infection  is  inexplicable  on  the  basis  of  structural  or  microscopic 
characteristics  in  the  organs.  We  have  mentioned  briefly  the  at- 
tempts that  have  been  made  to  discover  chemical  and  physical  changes 
or  differences  to  account  for  such  conditions  and  have  seen  that  the 
attention  of  investigators  was  soon  attracted  to  the  blood. 

A  possible  relationship  between  the  blood  and  the  defence  of  the 
body  against  infection  had  been  foreshadowed  by  observations  made 
long  before  the  days  of  bacteriological  knowledge.  As  early  as  1792, 
John  Hunter,  in  his  " Treatise  on  the  Blood,  Inflammation  and  Gun- 
shot Wounds,"  had  noted  that  the  blood  did  not  decompose  as  readily 
as  other  putrescible  material,  and  a  century  later,  during  the  period 
of  great  interest  in  the  living  nature  of  fermentation  and  putrefac- 
tion, Traube  (1874)  expressed  the  opinion  that  blood  could  destroy 
bacteria.  Similar  observations  were  made  by  Lister  and  by  Groh- 
man  1  but  no  experimental  work  aimed  at  this  point  was  carried  on 
until  1886,  when  the  subject  was  taken  up  by  Nuttall,2  von  Fodor,3 
and  Fliigge,  and  a  little  later  by  Buchner.4  These  authors,  working 
with  defibrinated  blood,  peptone  blood,  and  blood  serum,  showed  that 
such  substances  all  exerted  a  definitely  measurable  destructive  influ- 
ence upon  bacteria,  and  Nuttall,  later  confirmed  by  Buchner,  further 
found  that  this  bactericidal  power  was  weakened  on  standing,  and 
could  be  rapidly  destroyed  by  heating  to  60°  C. 

Their  method  of  procedure  consisted  in  the  planting  of  controlled 
amounts  of  various  bacteria  in  measured  quantities  of  blood  and, 

1  Grohman.     Quoted  from  Adami,  "Principles  of  Pathology,"  Vol.  1,  p. 
497. 

2  Nuttall.     Zeitschr.  f.  Hyg.,  4,  1888. 

3  Von  Fodor.     Deutsche  med.  Woch.,  1887. 

4  Buchner.     Centralbl  f.  Bakt.,  Vol.  5,  1889. 

134 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      135 

after  several  hours  at  37°  C.,  pouring  plates  and  thus  determining 
the  numbers  of  surviving  organisms.  The  fact  of  bactericidal  power 
established,  there  was,  of  course,  much  early  difference  of  opinion 
as  to  the  mechanism  responsible  for  the  destruction  of  the  bacteria, 
and  a  number  of  simple  explanations  were  suggested  which,  though 
entirely  refuted  at  the  present  time,  still  possess  considerable  inter- 
est in  showing  the  stages  of  development  through  which  the  concep- 
tions of  the  mechanism  of  immunity  have  progressed. 

These  early  theories  were  formulated  chiefly  upon  the  under- 
lying thought  that  the  animal  body  was  primarily  passive  in  its  rela- 
tion to  the  invading  micro-organisms,  and  that  the  disappearance  of 
bacteria  in  the  body  fluids  was  due  to  the  existence  of  a  chemically 
or  physically  unfavorable  environment  which  prevented  their  multi- 
plication and  therefore  induced  gradual  mortality  among  them. 
Thus  Billroth  5  believed  that  bacteria  could  thrive  in  the  body  only 
after  a  preceding  putrefactive  change  had  prepared  a  favorable  pab- 
ulum. Others  attempted  to  discover  a  relation  between  the  degree 
of  alkalinity  of  the  blood  serum  and  the  destruction  of  bacteria.- 
This  argument  was  soon  refuted  by  the  experiments  of  Buchner,  who 
showed  conclusively  that  the  bactericidal  power  of  serum  was  not 
reduced  by  the  neutralization  of  its  natural  alkalinity  with  weak 
acetic  acid. 

Another  theory  which  has  been  kept  alive  until  the  present  day 
by  Baumgarten,6  and  in  favor  of  which  much  has  been  written  by 
Fischer,  is  the  so-called  "Osmotic"  explanation.  The  basis  of  this 
conception  is  the  observation  that  vegetable  and  other  cells,  which 
are  in  themselves  delicate  osmotic  systems,  undergo  changes  when 
they  are  placed  into  fluids  of  different  osmotic  tension.7  Thus,  of 
course,  cells  of  all  kinds  may  be  destroyed  by  being  placed  in  dis- 
tilled water  on  the  one  hand,  or  in  hypertonic  salt  solution  on  the 
other.  The  point  of  view  of  Baumgarten,  as  explained  in  a  recent 
edition  of  his  "Text-book  of  Bacteriology,"  is  the  following:  The  bac- 
terial (or  blood)  cell,  like  all  cells,  is  surrounded  by  a  semi-per- 
meable membrane.  Under  ordinary  conditions,  this  membrane  per- 
mits the  passage  of  certain  substances  which  must  enter  and  leave 
the  cell  in  the  course  of  normal  metabolism.  When  the  bacteria  are 
placed  in  a  specific  bacteriolytic  serum  there  is  a  chemical  union 
between  the  antibody  and  the  cell  membrane,  and  the  latter  is,  in 
consequence,  injured.  The  result  of  the  injury  is  that  now  the  cell 
becomes  permeable  for  salts  and  other  substances  to  which  it  was 
impermeable  before,  and  there  are  consequent  swelling  and  in- 

5  Billroth.    Quoted  from  Sauerbeck,  "Die  Krise  in  der  Immunitatsforsch.," 
Klinkhardt,  Leipzig-,  1909. 

6  Baumgarten.     "Lehrbuch   der  pathogenen   Mikroorg.,"   Hirzel,   Leipzig, 
1911. 

7  See  also  Pfeiffer's  "Pflanzen  Physiologic." 


136  INFECTION    AND    RESISTANCE 

creased  intracellular  pressure.  This,  in  turn,  brings  about  the  ex- 
trusion from  the  cell  of  proteins  and  other  ordinarily  non-diffusible 
substances,  and  destruction  of  the  cell  results.  This  explanation  is 
practically  an  adaptation  of  the. earlier  more  primitive  osmotic  the- 
ories to  the  facts  subsequently  discovered.  It  stands  in  direct  con- 
tradiction to  the  prevailing  opinion  that  the  process  of  bacteriolysis 
and  cytolysis  in  general  is  an  enzymotic  process,  brought  about  by 
the  injury  of  the  cell  by  specific  substances  comparable  to  digestive 
ferments..  Interesting'  though  the  suggestion  of  Baumgarten  is,  it 
can  hardly  receive  more  than  casual  attention  given  it  for  the  sake 
of  completeness,  since  careful  experimental  work  by  von  Lingel- 
sheim8  has  shown  definitely  that  altered  salt  contents  of  serum 
do  not  exercise  the  effect  upon  bacteriolysis  which  we  would  be  en- 
titled to  expect  from  Baumgarten's  reasoning. 

In  explanation  of  the  natural  immunity  possessed  by  many  ani- 
mals against  various  infections,  Baumgarten  has  offered  another 
explanation  which,  like  the  preceding,  we  may  classify,  in  agreement 
with  Sauerbeck,9  with  the  "passive"  theories.  This  theory,  which 
he  calls  his  "Assimilation  Theory,"  assumes  that  the  bacteria  do  not 
find  suitable  food  material  in  the  tissues  and  fluids  of  certain  ani- 
mals, and,  since  bacteria  do  not  have  to  be  killed  to  be  eliminated, 
but  may  be  checked  merely  by  their  inability  to  grow  and  multiply, 
they  must  soon  succumb  in  surroundings  in  which  they  find  no  suit- 
able foodstuffs.  This  point  of  view  approaches  somewhat  the  earlier 
exhaustion  theory  of  Pasteur,  which  has  been  mentioned  in  another 
place.10 

In  contrast  to  these  "Passive"  theories  of  immunity  are  the  now 
prevailing  and  well-founded  opinions  that  the  resistance  of  the  ani- 
mal body  against  bacterial  invasion  is  not  a  mere  fortuitous  result 
of  chemical  and  physical  conditions  encountered  by  the  infectious 
agents,  but  is  rather  the  result  of  the  struggle  against  the  invasion 
by  active  forces  of  the  body  cells  and  fluids.  The  part  played  by  the 
cells  had  already  been  emphasized  by  Metchnikoff  and  his  school 
when  the  discovery  of  the  bactericidal  power  of  the  normal  blood 
was  made.  The  study  of  the  antibacterial  powers  of  the  blood  now 
introduced  a  new  element  which  became  the  basis  of  the  so-called 
"humoral"  theories.  In  the  prolonged  controversies  waged,  with 
great  astuteness  and  experimental  skill,  between  the  adherents  of 
these  two  schools,  most  of  the  facts  which  we  possess  regarding  im- 
munity were  discovered,  and  it  is  only  within  recent  years  that  we 

8  Von  Lingelsheim.     Zeitsclir.  f.  Hyg.,  Vol.  37,  1901. 

9  Sauerbeck.      "Die    Krise    in    der    Immunitatsf  orschung,"    Klinkhardt, 
Leipzie,  1909. 

10  The  influence  of  foodstuffs,  temperature,  and  other  environmental  con- 
ditions upon  natural  immunity  has  been  discussed  in  an  earlier  section. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM   137 

have  obtained  information  which  has  made  possible  a  correlation 
between  these  two  main  paths  of  thought. 

The  humoral  theory  was  conceived  by  Buchner,  as  the  first 
important  theoretical  result  of  NuttalFa  discovery.  Buchner,  as 
we  have  seen,  confirmed  the  observations  of  Nuttall  both  as 
to  the  primary  fact  of  the  bactericidal  power  of  the  fresh  normal 
blood  and  as  to  the  unstable  nature  of  this  bactericidal  property. 
He  looked  upon  the  antibacterial  power  as  depending  upon  a  con- 
stituent of  the  fresh  blood  plasma,  which  he  named  <f  alexin'  (pro- 
tective substance),  and  which  he  believed  to  be  comparable  to  a 
proteolytic  enzyme.  The  action  of  this  alexin  was  conceived  as 
potent  against  all  bacteria  equally,  without  showing  specific  selection 
of  various  species  to  any  great  extent.  The  analogy  to  ferment 
action  was  formulated  by  Buchner  because  of  the  heat  sensitiveness 
and  the  instability  of  the  bactericidal  substance  on  standing ;  and  he 
suggested  that  this  alexin  might  possibly  be  a  product  of  the  tissue  or 
blood  cells,  possibly  leukocytic  in  origin. 

Buchner  found  that  the  action  of  the  ferment-like  alexin  upon 
bacteria  was  most  marked  at  the  temperature  of  the  body,  and  that 
it  was  capable  of  destroying  bacteria  in  the  subcutaneous  tissues  and 
the  serous  cavities  of  the  animal  body,  without  the  aid  or  coopera- 
tion of  cellular  elements.  He  inferred  that  there  was  a  direct  rela- 
tion between  the  potency  of  the  alexin  and  resistance  against  infec- 
tion. 

The  next  great  step  in  the  understanding  of  the  bactericidal 
processes  was  now  made  by  Pfeiffer  as  a  consequence  of  studies  upon 
the  nature  of  cholera  immunity.  Pfeiffer  n  12  found  that  the  injec- 
tion of  cholera  spirilla  into  the  peritoneal  cavity  of  a  guinea  pig 
which  had  recovered  from  a  previous  cholera  infection  was  fol- 
lowed by  a  rapid  destruction  of  the  bacteria.  If  small  quantities  of 
exudate  were  taken  out  of  the  peritoneum  at  varying  intervals  after 
the  injection,  a  granular  change  and  swelling  of  the  bacteria  were 
noticed,  followed,  soon  after,  by  complete  dissolution  and  disappear- 
ance. Such  animals  would  recover  from  doses  of  bacteria  which,  in 
control  animals  of  the  same  weight,  resulted  in  death.  He  further 
found  that  the  phenomenon  was  specific,  in  that  the  dissolution  of 
cholera  organisms  only  occurred  in  the  cholera-immune  animals, 
other  bacteria  being  unaffected.  In  other  words,  the  guinea  pig  had 
acquired  a  specific  antibacterial  power,  expressed  by  the  process  of 
"bacteriolysis,"  a  property  possessed  to  only  a  very  slight  extent  by 
the  peritoneal  exudate  of  a  normal  animal.  It  was  the  next  logical 
step  to  determine  whether  the  bacteriolytic  power  could  be  trans- 
ferred to  the  peritoneal  cavity  of  a  normal  animal  by  injecting,  to- 
gether with  the  bacteria,  a  small  amount  of  the  serum  of  such  an 

"Pfeiffer.     Zeitschr.  f.  Hyg.,  Vol.  18,  1894;  also  Vols.  19  and  20. 
12  Pfeiffer  &  Isaeff.    Deutsche  med.  Woch.,  No.  18,  1894. 


138  INFECTION    AND    RESISTANCE 

immune  animal.  This  was  indeed  found  to  be  the  case  and,  al- 
though such  immune  serum,  like  normal  serum,  is  deprived  of  its 
in  vitro  bactericidal  power  on  heating,  Pfeiffer  found,  in  his  intra- 
peritoneal  experiments,  that  heated  serum  is  quite  as  effectual  as 
fresh  immune  serum  in  transferring  passive  immunity  to  a  normal 
guinea  pig.  We  may  summarize  the  important  harvest  of  facts  ob- 
tained from  these  experiments  of  Pfeiffer  in  the  following  state- 
ments : 

1.  Rapid   dissolution    of   cholera    spirilla   takes   place   in   the 
peritoneal  cavity  of  a  cholera-immune  guinea   pig.     Similar  lysis 
takes  place  not  at  all,  or  only  to  a  slight  extent,  in  the  peritoneum 
of  a  normal  pig.     In  consequence  of  the  lysis  the  immune  pig  will 
survive  the  injection  of  quantities  of  bacteria  which  invariably  kill 
normal  animals  of  the  same  weight. 

2.  The  protection  obtained  in  this  way  is  specific. 

3.  The  protection  may  be  transferred  from  an  immune  to  a 
normal  guinea  pig,  by  injecting  a  little  immune  serum  together  with 
the  bacteria  into  the  peritoneum  of  the  normal  animal.    In  a  normal 
animal  so  treated  lysis  is  in  every  way  similar  to  that  observed  in 
the  immune  pig. 

4.  The  transfer  of  the  lytic  power  and  consequent  immunity 
can  be  brought  about  not  only  by  means  of  fresh  immune  serum  but 
by  heated  serum  as  well,  although  the  latter  has  lost  all  its  alexic 
power  because  of  the  heating. 

Of  the  phases  of  this  "Pfeiffer  phenomenon"  the  one  most  diffi- 
cult to  understand,  in  the  light  of  the  knowledge  of  that  time,  was 
the  transference  of  the  lytic  property  with  the  heated  serum.  Pfeiffer 
very  naturally  took  his  experiments  to  signify  that  the  actual  de- 
struction of  bacteria  in  the  animal  body  could  take  place  entirely 
without  the  phagocytic  participation  of  the  body  cells,  a  view  in 
sharp  contrast  to  that  of  the  Metchnikoff  school,  and  based  upon  his 
observation  of  the  complete  extracellular  disintegration  of  the  spir- 
illa in  the  peritoneal  exudate.  He  assumed,  however,  that  there  was 
an  indirect  participation  on  the  part  of  the  cells.  The  observation 
that  heated  serum,  inactive  outside  the  body,  was  efficient  when  in- 
troduced into  the  peritoneum,  persuaded  him  that  the  cooperation 
of  the  living  tissues  was  a  necessary  factor,  and  he  assumed  a  pos- 
sible activation  by  substances  derived  from  the  endothelial  cells  lin- 
ing the  peritoneal  cavity.  In  the  same  way  he  explained  his  failure 
to  observe  actual  bacterial  dissolution  in  hang-drop  preparations, 
even  when  fresh  serum  was  used  in  the  experiment. 

It  will  be  interesting  to  examine  a  protocol  of  an  experiment  such 
as  those  carried  out  in  the  performance  of  the  Pfeiffer  phenomenon 
in  order  to  make  the  actual  occurrences  entirely  clear.  In  such 
experiments  the  quantity  of  bacteria  used  must  be  chosen  with  some 
regard  to  the  virulence  and  toxicity  of  the  particular  culture  em- 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM    139 


ployed,  since,  as  we  shall  see,  protection  of  animals  by  bactericidal 
or  bacteriolytic  sera  does  not  follow  the  law  of  multiple  proportions 
as  in  the  case  of  the  protection  against  toxins  by  antitoxins.  While 
the  dose  of  bacteria  chosen  should  be  considerably  above  the  minimal 
lethal  dose  for  an  animal  of  the  weight  used,  it  should  nevertheless 
be  remembered  that  the  bactericidal  serum  does  not  possess  antitoxic 
properties  against  the  poisons  liberated  or  produced  as  the  bacteria 
undergo  dissolution,  and  at  best  the  protection  by  bacteriolysis  is 
limited  to  a  very  definite  maximum  of  bacteria,  beyond  which  no 
further  increase  of  serum  quantity  will  avail.  The  following  table 
will  illustrate  an  experiment  of  this  kind  in  which,  in  a  series  of 
guinea  pigs,  the  bacteriolytic  protective  power  (titre)  is  determined 
by  comparative  tests.13 

PFEIFFER  PHENOMENON 


Weight 
of 
guinea  pig 

Dose  of  bacteria* 
cholera  spirilla 

Amount  of 
inactivated 
immune  serum 

Result 

(1)  215  gm. 

2mg. 

0.1  c.  c.  in  1  c.  c. 

Complete  dissolution  in  less  than 

salt  solution. 

1  hour.    Lives. 

(2)  230  gm. 

2mg. 

0.05  c.  c. 

About  the  same  as  first. 

(3)  200  gm. 

2  ing. 

0.01  c.  c. 

Somewhat  slower  than  in  other 

, 

two;  a  few  unchanged  spirilla 

after  1  hr.    Final  dissolution. 

(4)  245  gm. 

2mg. 

0.005  c.  c. 

Pig  lives. 
Similar  to  (3)  but  complete  dis- 

solution in  2  hrs.    Pig  lives. 

<5)  220  gm. 

2mg. 

0.001  c.  c. 

After  30  min.  the  spirilla  seem 

to  have  begun  to  multiply. 

Dies  with  innumerable  active 

spirilla  in  peritoneum. 

Normal 

control 

(6)  210  gm.  ! 

2mg. 

0.1  c.  c.  normal 

Very  slight  lysis  at  the  beginning. 

inactive   rab- 

Soon    rapid     multiplication. 

bit  serum. 

Dies. 

"The  bacteria  may  be  measured  for  such  an  experiment  by  standard  loopfuls  (  1  loop  be- 
ing equal  to  2  milligrams),  or  by  volume  in  emulsion  with  salt  solution. 

Pfeiffer  has  established  a  system  of  standardization  for  the  meas- 
urement of  sera  by  this  technique.  He  speaks  of  one  immunity  unit 
as  the  smallest  amount  of  such  a  serum  which  is  capable  of  causing 

13  For  extensive  discussion  of  the  technique  of  such  tests  see  Boehme  in 
Kraus  u.  Levaditi  Handbuch,  etc.,  Vol.  2,  p.  366.  The  scheme  of  presenta- 
tion of  our  example  is  taken  from  that  used  by  him.  See  also  Pfeiffer, 
Zeitschr.  /.  Hyg.,  Vol.  19,  1895,  p.  77. 


140  INFECTION    AND    RESISTANCE 

complete  dissolution  of  2  milligrams  of  culture  material 14  (of  a 
standard  culture)  and  saving  the  life  of  the  animal.  The  unit  of 
the  serum  in  the  preceding  test  would  accordingly  be  0.005  c.  c.,  and 
the  ticre  of  the  serum,  expressed  in  Pfeiffer's  language,  would  be 
200  units  to  the  cubic  centimeter.  Owing  to  the  great  variation  in 
the  virulence  and  toxicity  of  different  strains  of  the  same  organism, 
and  also  because  of  the  difficulties  opposed  to  the  visible  dissolution 
of  many  bacteria,  which  may  be  killed  by  the  serum  without  show- 
ing much  evidence  of  solution,  the  practical  application  of  Pfeiffer's 
standardization  is  not  universally  possible.  In  doing  experiments 
by  this  technique,  whatever  their  purpose  may  be,  accurate  adjust- 
ment of  bacterial  amounts  and  preliminary  studies  of  virulence  must 
be  made  in  order  that  the  tests  may  be  of  real  value  and,  failing 
visible  lysis,  the  death  of  the  animals  must  be  taken  as  the  indicator 
of  the  titration.  Comparisons  of  results  obtained  with  two  different 
cultures  of  the  same  species  are  consequently  of  value  only  when  the 
minimal  lethal  dose  of  each  and  its  toxicity  have  been  studied  before 
the  final  tests  are  made. 

The  cardinal  points  of  Pfeiffer's  phenomenon  were  rapidly  con- 
firmed, but  his  assumption  that  the  process  could  take  place  only 
within  the  animal  body  was  soon  corrected  by  both  Metchnikoff  15 
and  Bordet.16  Both  of  these  investigators  succeeded  in  producing 
extracellular  lysis  of  cholera  spirilla  in  hang-drop  preparations.  The 
former  produced  the  phenomenon  by  adding  to  the  hang-drop  prep- 
arations small  quantities  of  extracts  of  leukocytes,  and  thus  at- 
tempted to  correlate  Pfeiffer's  observations  with  his  own  opinions 
regarding  the  importance  of  the  leukocytes  in  bacterial  destruction. 
The  latter,  however,  subjected  the  phenomenon  of  bacteriolysis,  both 
in  vivo  and  in  vitro,  to  a  careful  analysis  and  obtained  results  which 
definitely  disproved  the  necessity  of  cellular  intervention  in  this 
phenomenon,  and  furnished  facts  regarding  the  process  which  stand 
uncontradicted  to  the  present  day.  Upon  the  basis  of  these  our 
modern  views  of  the  mechanism  of  cytolysis  in  general  are  founded. 

Bordet  showed  that  the  bacteriolytic  properties  of  immune  serum 
are  indeed  destroyed  by  heating  to  from  50°  to  60°  C.  If,  however,, 
to  such  a  heated  immune  serum  there  is  added  a  small  quantity  of 
fresh  normal  serum,  the  bacteriolytic  power  is  restored  with  undi- 
minished  vigor.  He  recognized  in  consequence  that  there  were  two 
distinct  serum  elements  necessary  for  the  process.  Fresh  normal 
serum  by  itself  had  very  slight  or  no  bacteriolytic  power.  Fresh 
immune  serum  had  powerful  and  rapid  effects.  Heated  immune- 

14  The  standard  "loop"  used  in  many  laboratories  for  the  rough  meas- 
urement of  quantities  of  bacteria  from  agar  cultures  takes  up  approximately 
2  milligrams  of  the  material. 

15  Metchnikoff.     Ann.  de  I'Inst.  Past.,  Vol.  9,  1895. 

16  Bordet.    Ann.  de  I'Inst.  Past.,  Vol.  13,  1899. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM  141 

serum  had  lost  its  power  completely,  but  this  was  restored  to  it 
the  addition  of  the  fresh  normal  serum.  He  noted,  furthermore, 
that  the  specific  nature  of  the  bacteriolysis  by  the  immune  serum  was 
unchanged  after  it  had  been  inactivated  by  heat  and  reactivated  sub- 
sequently by  the  normal  serum.  The  inference  was  plain.  Immu- 
nization of  an  animal  incites  the  production,  in  the  blood  of  this 
animal,  of  a  "preventive"  substance,  which  is  moderately  resistant 
to  heat,  and  which  is  specific  for  the  bacteria  employed  in  the  im- 
munization. This  substance  cannot  act  upon  the  bacteria  alone,  how- 
ever, but  depends  for  its  effective  functionation  upon  the  coopera- 
tion of  another  substance  present  universally  in  normal  serum,  the 
"bactericidal"  substance,  which  is  non-specific,  corresponds  to  Buch- 
ner's  alexin,  and  is  apparently  not  increased  by  the  process  of  im- 
munization. These  are  the  fundamental  facts  revealed  by  the  early 
studies  of  Bordet,  and  they  are  stated  in  the  present  connection 
merely  as  experimental  facts,  without  further  elaboration  of  the 
later  theoretical  interpretation  placed  upon  them  by  Bordet  himself 
and  by  Ehrlich  and  his  followers. 

In  the  course  of  these  studies  Bordet 17  had  used  the  immune 
serum  produced  in  a  goat  by  injection  of  cholera  spirilla.  As  normal 
serum  he  had  used  guinea-pig  serum,  and  the  latter  frequently  con- 
tained a  few  blood  corpuscles.  He  noticed  that  these  corpuscle?  were 
frequently  clumped  in  the  goat  serum  and  correlated  this  with  the 
similar  clumping  (agglutination)  of  cholera  organisms  which  he 
had  noticed  in  this  and  other  sera.  In  his  incidental  observation  of 
the  phenomenon  of  agglutination  he  had  concluded  that  the  living 
nature  of  the  bacteria  had  no  importance  as  far  as  their  agglutina- 
tion was  concerned,  dead  organisms  being  as  readily  agglutinated  as 
living. 

Reasoning  from  this  similarity  between  blood  cells  and  bacteria 
in  their  behavior  in  serum,  it  occurred  to  him  that  the  phenomena 
both  of  agglutination  and  of  lysis  might  be  expressions  of  general 
biological  laws,  not  limited  to  bacteria.  Accordingly  he  injected 
rabbit  blood  into  guinea  pigs,  and  examined  the  serum  of  animals 
so  treated  for  its  action  upon  rabbit  corpuscles,  in  vitro.  He  found 
that  the  sera  of  "blood-immune"  animals  had  acquired  not  only  in- 
creased agglutinative  power  against  the  corpuscles  injected,  but  had 
also  acquired  specific  "hemolytic"  powers,  that  is,  the  property  of 
causing  a  solution  of  hemoglobin  out  of  the  red  cells.  (Eor  the 
process  of  serum  hemolysis  does  not  consist  of  a  complete  dissolution 
of  the  red  corpuscles,  but  rather  in  the  liberation  of  the  hemoglobin 
from  the  cell  stromata.)  The  latter  (shadow  forms)  can  be  recovered 
undisintegrated  by  the  centrifugation  of  hemolyzed  blood.  The 

17  See  Bordet's  own  account  in  a  "Resume  of  Immunity";  "Studies  in 
Immunity,"  Bordet,  collected  and  translated  by  Gay,  Wiley  &  Son,  N.  Y., 
1909. 


INFECTION    AND    RESISTANCE 

process,  like  that  of  bacteriolysis,  was  specific  in  that  the  hemolytic 
power  was  lost  if  the  serum  was  heated  to  from  50°  to  60°  C.,  but 
could  be  restored  undiminished  by  the  addition  of  a  little  fresh  nor- 
mal serum,  in  itself  possessing  no  hemolytic  properties  for  the  given 
species  of  cell.  The  specificity  of  the  phenomenon  again  was  seen 
to  reside  entirely  in  the  heat-stable  factor,  the  heat-sensitive  or 
"alexm"  factor  being  non-specific,  and  not  increased  during  the 
process  of  immunization. 

Observations  related  to  those  of  Bordet  concerning  hemolysis 
were  made  independently,  in  the  same  year,  by  Belfanti  and  Car- 
bone,  who  had  observed  that  the  serum  of  animals  treated  with  blood 
cells  of  another  species  became  toxic  for  this  species,  and  extensive 
confirmation  of  the  phenomenon  of  hemolysis  was  obtained,  in  the 
year  following,  by  the  work  of  von  Dungern,  and  by  that  of  Land- 
steiner. 

After  Bordet  had  thus  established  the  important  fact  that  hemol- 
ysis was  in  every  way  analogous  to  bacteriolysis  in  that,  like  bac- 
teriolytic  sera,  hemolytic  sera  could  be  inactivated  by  heat,  but  re- 
activated by  the  addition  of  small  quantities  of  fresh  normal  serum, 
Ehrlich  and  Morgenroth 18  undertook  an  elaborate  study  of  the 
mechanism  of  hemolytic  phenomena,  hoping  thereby  to  elucidate  the 
mechanism  of  lysis  in  general.  For  it  is  obvious  that  hemolysis 
lends  itself  far  more  easily  to  experimentation  than  does  bacteriol- 
ysis, and,  as  we  shall  see,  experiments  on  hemolysis  can  be  made 
with  a  considerable  degree  of  accuracy.  Ehrlich  and  Morgenroth 
approached  the  investigation  of  the  hemolysins  from  the  point  of 
view  of  the  side-chain  theory,  formulated  by  Ehrlich  in  connection 
with  his  work  on  the  toxins.  According  to  this  theory,  it  will  be 
remembered,  the  hemolytic  substances  in  the  sera  of  animals  treated 
with  blood  corpuscles  represent  the  receptors  or  side  chains  of  tissue 
cells.  These  receptors  were  originally  integral  chemical  elements  of 
the  body  cells,  by  means  of  which  the  cell  became  united  to  the 
injected  erythrocyte  (or  bacterial)  protein.  Since  union  with  the 
foreign  substance  blocked  these  receptors  or  side  chains,  thereby 
rendering  them  useless,  they  had  been  regenerated  and,  under  the 
influence  of  immunization,  regenerated  in  excess,  cast  off  by  the  cell, 
and  were  now  free  in  the  blood  stream  as  hemolysins  (or  bacteriol- 
ysins). 

If  this  conception  of  the  process  was  the  correct  one,  Ehrlich 
and  Morgenroth  argued,  the  hemolytic  substances  of  any  immune 
hemolytic  serum  should  possess  specific  chemical  affinity,  "hapto- 
phore  groups/'  as  they  expressed  it,  for  the  blood  cells  which  had 
been  used  in  the  immunization. 

In  order  to  show  this,  they  inactivated  at  5fi°  C.,  by  the  method 

of  Bordet,  a  goat  serurn  which  was  hemolytic  for  beef  blood,  left  it 

18  Ehrlich  and  Morgenroth.     Berl.  klin.  Woch.,  Nos.  1,  21,  and  22,  1900. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM     145 

in  contact  with  beef  blood  corpuscles  for  15  minutes  at  40°  C.,  and 
then  separated  the  cells  from  the  supernatant  fluid  by  centrifuga- 
tion.  To  the  blood  cells  they  then  added  a  little  normal  goat  serum 
(by  itself  not  hemolytic  for  beef  blood)  and  found  that  complete 
hemolysis  occurred.  The  addition  of  normal  goat  serum  and  beef 
blood  cells  to  the  supernatant  fluid,  however,  resulted  in  no  change. 
In  the  following  diagram  we  have  tried  to  represent  this  basic 
experiment,  giving  the  facts  only  of  the  experiment  without  using 
any  of  the  usual  symbols  which  imply  agreement  with  a  theory. 

EXPERIMENT   TO    SHOW    THAT   THE    ANTIGEN  (iN  THIS    CASE    RED    BLOOD    CELLS) 
ABSORBS    THE    SPECIFIC    HEAT   STABLE    ANTIBODY    OUT  OF  THE    IMMUNE    SERUM. 

f  4  c.  c.  of  5  per  cent,  emulsion  of  washed  beef  blood. 

In  a  test  tube    \  1  c.  c.  of  inactivated  blood  serum  of  a  goat  treated  with  beef 
i         blood. 

These  substances  are  left  together  at  37.5°  C.  for  one  hour  and 
then  centrifugalized  into: 

I  II 

Sediment  of  Corpuscles. — To  this  are  Supernatant  Fluid  Containing  the  Serum 

added  4  c.  c.  salt  solution  and  0.8  c.  c.  and  Salt  Solution. — To  this  are  added 

fresh  normal  goat  serum,  by  itself  washed  beef  corpuscles  and  0.8  c.  c. 

not  hemolytic  for  beef  corpuscles.  fresh  normal  goat  serum. 

Result  =  Complete  hemolysis.  Result  =  No  hemolysis. 

Summarized,  together  with  the  facts  we  have  already  outlined, 
this  basic  experiment  has  the  following  significance :  the  fresh  serum 
of  the  goat,  previously  injected  ("immunized")  with  the  beef  blood, 
possessed  the  property  of  dissolving  the  hemoglobin  out  of  beef  cor- 
puscles, viz.,  hemolyzing  them.  Heating  this  serum  to  56°  C.  for 
20  minutes,  as  Bordet  has  shown,  deprives  the  serum  of  all  hemolytic 
power,  i.  e.,  inactivates  it.  The  addition  of  a  little  fresh  goat  serum, 
in  itself  inactive,  completely  reactivates  the  hemolytic  properties  of 
the  heated  immune  serum.  So  far,  as  we  have  already  seen,  this 
shows  that  hemolysis  is  a  dual  process  in  which  a  heat-sensitive 
and  a  heat-stable  substance  co-operate,  neither  of  them  capable  of 
producing  lysis  by  itself.  The  heat-sensitive  ingredient,  correspond- 
ing to  Buchner's  "alexin,"  is  present  in  normal  serum,  and,  as  Bor- 
det 19  and  von  Dungern  20  had  shown,  is  not  increased  in  the  process 
of  immunization,  and  is  apparently  not  specific.  The  heat-stable 
substance,  therefore  specific  and  increased  in  immunization,  must 
represent  the  receptors,  overproduced  and  cast  off  into  the  circula- 
tion. And,  as  Ehrlich  and  Morgenroth  have  now  shown  in  the  ex- 
periment just  described,  this  heat-stable  element  is  actually  bound 
to  the  red  corpuscles,  and  renders  them  susceptible  to  the  action  of 

19  Bordet.     Ann.  de  I'Inst.  Past.,  Vol.  12,  1898. 

20  v.  Dungern.     Mimch.  med.  Woch.,  No.  20,  1900,  p.  677. 


144  INFECTION    AND    RESISTANCE 

the  heat-sensitive  substance  in  the  normal  goat  serum.  And  further- 
more, in  attaching  to  this  heat-stable  element,  the  blood  cells  have 
removed  it  from  the  solution.  For  we  have  seen,  in  the  experiment, 
that  addition  of  corpuscles  and  normal  serum  to  the  supernatant  fluid 
resulted  in  no  hemolysis,  showing  that  the  third  necessary  element, 
originally  in  the  mixture,  had  been  carried  down  with  the  red  cells. 
In  these  and  other  experiments  then,  it  was  shown  that  only  the 
heat  stable  substances  could  be  fixed  by  the  red  cells,  and  this  even 
at  temperatures  at  or  about  0°  C.  (a  fact  which  indicates  the  strong 
affinity  between  the  two  substances),  while  the  heat-sensitive 
"alexin,"  which  Ehrlich  now  called  "complement ,"  could  not  attach 
directly  to  the  red  cells.  For  if  such  complement,  in  the  form  of 
fresh  serum,  was  added  to  washed  red  blood  cells,  and  the  mixture 
after  standing  at  40°  C.  for  some  time  was  centrifugalized,  the  com- 
plement remained  in  the  supernatant  fluid,  as  could  be  easily  shown 
by  an  experiment  such  as  the  one  represented  in  the  following  proto- 
col. 

EXPERIMENT    TO    SHOW    THAT    COMPLEMENT    OR    ALEXIN    IS    NOT    ABSORBED    BY 

UNSENSITIZED    CELLS 

(4  c.  c.  of  5  per  cent,  emulsion  of  washed  beef  Llood. 
0.8  c.  c.  of  fresh  normal  goat  serum  (alexin  or  comple- 
ment), not,  by  itself,  hemolytic  for  beef  blood. 

These  substances  are  left  together  at  37.5°  C.  for  one  hour,  then 
centrifugalized  into: 

I  II 

Sediment  of  Cells. — To  this  is  added  Supernatant  Fluid  (salt  solution  and 

inactivated  serum  of  immune  goat  serum). — To  this  is  added  washed 

which   would    cause    hemolysis    if  beef  blood  and  inactivated  serum  of 

alexin  were  present.  immune  goat  containing  heat  stable 

element. 

Result  =  No  hemolysis.  Result  =  Complete  hemolysis. 

Although,  therefore,  the  red  cells  bind  the  thermostable  specific 
antibody  of  the  immune  serum  and  not  the  complement  or  alexin, 
it  was  shown  both  by  Bordet  and  by  Ehrlich  and  his  collaborators 
that  the  red  cells,  after  absorption  of  the  thermostable  substance, 
when  exposed  to  the  action  of  the  complement,  were  not  only  disin- 
tegrated by  hemolysis  but,  in  the  process,  fixed  or  attached  the  com- 
plement, so  that  this  was  no  longer  available  for  further  activation 
of  other  sensitized  cells. 

The  fact  that  the  alexin  or  complement  is  used  up  during  proc- 
esses of  lysis,  as  first  described  by  Bordet,  Ehrlich,  and  others,  has 
recently  been  made  the  subject  of  repeated  investigation,  srnce  this 
is  out  of  keeping  with  the  general  enzyme  or  fermentlike  nature  of 
complement  indicated  by  many  of  its  other  properties. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM     145 

Muir,21  who  studied  the  conditions  thoroughly ,  comes  to  the  con- 
clusion that  the  complement  is  in  truth  used  up  in  hemolysis,  but 
that  it  does  not  always  disappear  completely,  this  depending  upon 
the  relative  amount  of  sensitizer  or  amboceptor  present.  (He  con- 
firms the  quantitative  ratios  between  the  two  substances  found  by 
Morgenroth  and  Sachs  in  hemolytic  reactions,  a  subject  discussed 
by  us  in  another  place.) 

Liefmann  and  Cohn,22  in  a  more  recent  publication,  have  come 
to  different  conclusions.  They  believe  that  the  disappearance  of  free 
complement  from  hemolytic  complexes  is  not  due  to  its  chemical 
union  with  the  sensitized  cells  in  the  process  of  hemolysis,  but  is  due 
rather 

(1)  to  a  fixation  by  the  products  of  hemolysis  (stromata,  etc.) 
after  the  reaction  is  accomplished, 

(2)  to  dilution,  and 

(3)  to  weakening  because  of  prolonged  preservation  in  dilute 
solution  at  37°  C.23 

Theoretically  this  is  of  considerable  importance  if  confirmed, 
since  it  would  bear  out  strongly  the  conception  of  complement  as  a 
true  enzyme  or  ferment.  From  the  point  of  view  of  the  practical 
utilization  of  complement  fixation  for  various  purposes  it  makes 
little  difference,  since  here  the  disappearance  of  complement  is  the 
essential  thing,  irrespective  of  whether  this  occurs  in  the  course  of 
its  activity  or  because  of  fixation  by  the  products  of  its  own  action. 

We  now  have  the  basic  principle  of  hemolysis ;  facts  which  can 
easily  *be  shown  to  hold  good  for  bacteriolysis  and  for  the  bacteri- 
cidal processes  even  when  no  actual  solution  takes  place.  Briefly 
reviewed,  these  facts  are  as  follows:  The  antigen  (blood  cells,  bac- 
terial cells,  etc.)  undergoes  hemolysis  or  bacteriolysis  when  acted 
upon  by  two  factors,  one  a  thermostable  substance,  specific  and  in- 
creased during  immunization,  the  other  a  thermosensitive  substance 
present  in  fresh  serum,  not  increased  24  by  immunization  of  the  ani- 
mal with  the  antigen  and  not  specific.  The  specific  thermostable 
substance  becomes  united  with  or  fixed  to  the  antigen  regardless  of 
the  presence  or  absence  of  the  thermosensitive  alexin  or  comple- 
ment, and  with  such  avidity  that  the  union  takes  place  even  at  0°  C. 
The  alexin  or  complement,  however,  cannot  enter  into  relation  with 
the  antigen  unless  this  has  been  rendered  susceptible  to  it  by  attach- 
ment to  the  thermostable  specific  substance.  When  this  has  taken 

21  Muir.     Lancet,  Vol.  2,  1903,  p.  446. 

22  Liefmann  and  Cohn.     Zeitsch.  f.  Immunitatsforsch.  Or.,  Vol.  8,  p.  58, 
1911. 

23  In  tlie  ordinary  dilution  used  in  Wassermann  tests,  the  unit  of  comple- 
ment employed  may  deteriorate  entirely  within  several  hours  at  40°  C. 

24Bordet.  Ann.  de  I'Inst.  Past.,  Vol.  12,  1898.  Confirmed  by  v.  Dun- 
gern,  Munch,  med.  Woch.,  No.  20,  1900. 


146  INFECTION    AND    RESISTANCE 

place,  union  with  complement  occurs,  but  only  at  temperatures  above 
0°  C.  (the  speed  and  completeness  of  the  union  increasing  as  the 
temperature  approaches  40°  C.),  and  the  result  of  the  union  is  lysis 
or,  in  the  case  of  bacteria  not  easily  soluble,  the  bactericidal  effect. 

Early  in  their  researches,  Ehrlich  and  Morgenroth  were  led  to 
speculate  upon  the  possibility  of  the  formation  of  lytic  antibodies 
within  the  animal  against  its  own  tissue  cells.  It  would  be  of 
the  greatest  importance  to  pathology,  as  they  point  out,  if  it  could 
be  shown  that  an  animal  could  produce  hemolysins,  for  instance, 
against  its  own  blood  cells.  Thus,  if  an  extensive  internal  hemor- 
rhage occurred  from  trauma  or  other  cause,  in  the  course  of  which 
considerable  quantities  of  erythrocytes  are  subjected  to  disintegra- 
tion and  absorption,  it  is  at  least  conceivable  that  specific  "auto- 
hemolysins"  might  appear  which  would  lead  to  a  chronic  destruction 
of  the  red  cells,  with  consequent  anemia.  This  form  of  reasoning, 
as  we  shall  see,  has  been  extensively  applied  in  the  case  of  the  cyto- 
toxins  for  the  explanation  of  a  variety  of  pathological  conditions. 
Ehrlich  and  Morgenroth  approached  the  question  experimentally  in 
their  further  work  on  the  hemolysins  in  goat  blood.  They  found  that 
it  was  comparatively  easy  to  produce  hemolysins  in  one  goat  by 
treatment  with  the  erythrocytes  of  other  goats,  isohemolysins,  as 
they  called  them. 

Although,  however,  the  blood  serum  of  such  an  immunized  goat 
was  strongly  hemolytic,  not  only  for  the  blood  cells  of  the  goats 
whose  blood  had  been  injected,  but  also  for  the  erythrocytes  of  cer- 
tain other  goats  (though  not,  as  we  shall  see,  for  goats  in  general), 
it  was  never  in  any  case  active  against  this  goat's  own  cells.  More- 
over, while  the  other  sensitive  erythrocytes  could  absorb  the  hemo- 
lytic antibody  out  of  the  inactivated  serum,  the  insensitive  corpuscles 
of  the  goat  himself  seemed  to  possess  no  affinity  whatever  for  the 
lysin  of  his  own  serum ;  mixed  with  the  serum  they  failed  to  absorb 
out  the  hemolysin.  This  was  in  no  sense,  therefore,  an  autolysin. 

These  experiments  show  a  remarkable  individual  variation  be- 
tween the  similar  tissues  of  animals  of  the  same  species,  since  Ehr- 
lich and  Morgenroth  were  indeed  able  to  show  that  the  insensibility 
of  the  goat's  own  corpuscles  depended  upon  a  complete  absence  of 
receptors  for  the  isolysin.  For,  to  explain  the  lack  of  "autolytic" 
action  of  such  a  serum,  two  possibilities  could  be  assumed.  One,  as 
above,  that  the  corpuscles  of  the  goat  possessed  no  receptors  by  means 
of  which  the  isolysin  could  be  "anchored"  or,  second,  that,  although 
such  receptors  were  present,  they  were  already  satisfied,  or  saturated 
with  the  lysin  in  the  blood  stream.  In  the  latter  case  it  would  be 
hard  to  understand  why  hemolysis  had  not  taken  place. 

In  order  to  completely  disprove  the  latter  possibility,  Ehrlich 
and  Morgenroth  did  not  allow  the  matter  to  rest  upon  conjecture,  but 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM     147 

resorted  to  an  ingenious  method  of  experimentation  which  yielded  a 
further  important  result,  namely,  the  discovery  that  the  injection  of 
antibodies  into  animals  may  give  rise  to  "anti-antibodies."  They 
injected  inactivated  hemolytic  serum  into  goats  whose  corpuscles 
were  sensitive  to  its  action,  and  found  that  an  "anti-isolysin"  was 
formed,  which,  mixed  with  hemolysin  and  sensitive  corpuscles,  pre- 
vented hemolysis.  Injection  of  such  an  isolysin  into  the  goat  from 
which  it  had  been  obtained,  however,  did  not  yield  anti-isolysin,  and 
it  was  therefore  reasonable  to  suppose  that  its  tissue  cells  possessed 
no  suitable  receptors.  This  failure  of  the  production  of  antibodies 
by  an  animal  against  its  own  tissue  cell  has  been  spoken  of  by  Ehr- 
lich  as  "Horror  Autotoxicus." 

These  rather  involved  experimental  data  will  be  shown  to  have-  a 
more  than  purely  academic  value  when  we  come  to  speak  of  the 
problems  of  cytotoxin  formation,  and  although  they  seem  to  show 
that  auto-antibodies  do  not  form,  as  a  rule,  exceptions  to  this  gener- 
alization have  been  observed.  The  most  notable  of  these  is  the  ob- 
servation of  Landsteiner  and  Donath  25  made  in  connection  with  the 
condition  of  paroxysmal  hemoglobinuria.  It  was  found  that  in  such 
cases,  in  which  hemoglobinuria  follows  exposure  to  cold,  the  blood 
serum  of  the  patient  contains  an  "autohemolysin."  If  the  blood  of 
such  a  case  is  taken  into  oxalate  or  citrate  solution,  and  allowed  to 
stand  at  ordinary  or  incubator  temperature,  nothing  occurs.  If, 
however,  such  blood  is  cooled  to  0°  to  10°  C.  and  then  warmed  grad- 
ually to  the  temperature  of  the  body,  rapid  hemolysis  occurs.  In 
this  case  the  "amboceptor"  of  the  serum  is  apparently  fixed  or  an- 
chored by  the  blood  cells  only  at  a  low  temperature,  the  complement 
becoming  active  as  the  blood  is  warmed.  Although  Landsteiner' s 
observations  are  undoubtedly  accurate,  it  is  likely  that  this  mechan- 
ism does  not  explain  all  such  cases.  The  writer  has  had  occasion  to 
examine  carefully  a  number  of  clinically  diagnosed  cases  of  this 
sort  with  a  partially  successful  "Landsteiner"  phenomenon  in  one 
of  them  only.  Other  observers  have,  however,  confirmed  Land- 
steiner's  observation  in  well-established  cases  of  the  condition. 

Before  we  leave  the  subject  of  iso-antibodies  it  will  be  interest- 
ing to  discuss  for  a  moment  the  existence  of  isolysins  in  animals 
other  than  goats  and  more  especially  those  occurring  in  human 
beings,  phenomena  which  have  recently  assumed  considerable  impor- 
tance in  view  of  the  frequent  therapeutic  performance  of  blood 
transfusion. 

The  peculiar  facts  unearthed  by  Ehrlich  and  Morgenroth  26  indi- 
cated specific  differences  between  red  blood  cells  of  individuals  in 
the  same  species  (goats),  which  could  only  be  recognized  by  the 

25  Landsteiner  and  Donath.     Munch,  med.  Woch.,  1904,  p.  1590. 

26  Ehrlich  and  Morgenroth.     "tiber  Hamolysine,"  Berl.  kl.  Woch.,  1900, 
No.  21. 


148  INFECTION    AND    RESISTANCE 

development  of  immune-isolysins.  Work  on  other  species  of  animals 
has  indicated  that  this  fact  has  a  broad  significance  and  that  similar 
differences  between  individuals  of  the  same  species  occur  in  many, 
if  not  all,  species  of  animals.  Isolysins  similar  in  principle  to  those 
of  Ehrlich  and  Morgenroth  were  produced  by  Ascoli 27  in  rab- 
bits; by  Todd  and  White,28  in  oxen;  by  Ottenberg,  Kaliski,  and 
Friedmann  29  in  dogs ;  by  Ottenberg  and  Thalhimer  30  in  cats,  and  by 
Hada  and  Rosenthal  31  in  chickens.  In  all  these  instances  the  iso- 
lysins  developed  showed  the  same  peculiarities,  namely,  that  they 
attacked  the  cells  of  certain  individuals  and  left  the  cells  of  other 
individuals  of  the  same  species  unharmed.  Recent  work  on  the 
isolysins  occurring  naturally  in  the  human  blood  has  thrown  con- 
siderable light  on  the  nature  of  immune  isolysins. 

The  occurrence  of  isolysins  in  human  blood  was  first  noted  by 
Maragliano  32  in  1892,  and  a  large  amount  of  work  had  been  done 
before  it  was  clear  that  the  occurrence  of  isolysins  is  not  a  charac- 
teristic of  disease.  The  work  of  Moss  33  and  of  Grafe  and  Graham  34 
has  shown  that  the  occurrence  of  isolysins  is  parallel  with  that  of 
iso-agglutinins  (see  chapter  on  agglutination),  and  that  there  are  in 
human  bloods  two  isohemolysinogens,  A  and  B  (corresponding  to  the 
two  agglutinogens,  A  and  B),  and  two  isohemolysins,  a  and/3.  The 
hemolysinogens  occur  regularly  according  to  the  same  rule  as  the 
agglutinogens,  but  the  hemolysins,  while  they  always  follow  the 
same  rule  when  present,  may  be  present  or  latent.  Thus  a  person 
whose  red  cells  contain  A  may  or  may  not  have  fi  ,  and  never  has  a; 
a  person  whose  red  cells  contain  B  may  or  may  not  have  a,  but  can 
never  have  ft;  a  person  whose  red  cells  are  susceptible  to  both  a  and 
/?  never  has  any  hemolysin  in  the  serum.  It  seems  likely  that  the 
substances  A  and  B,  which  cause  the  susceptibility  of  red  cells  to 
the  corresponding  hemolysins,  are  definite  biochemical  structures 
which  possibly  may  be  inherited  in  a  similar  way  to  the  iso-agglu- 
tinogens  and  that  similar  substances  (probably  a  larger  number  of 
them)  are  present  in  the  blood  cells  of  various  species  of  lower  ani- 
mals. This  readily  explains  the  apparent  irregularity  attending  the 
development  of  isolysins  in  the  lower  animals.  The  reason  for  the 
natural  occurrence  of  such  isolysins  in  human  sera  and  occasionally 
in  the  sera  of  lower  animals,  however,  is  a  complete  mystery.  From 

27  Ascoli.     Munch,  med.  Woch.,  1901. 

28  Todd  and  White.     Nature,  June  23,  1910. 

29  Ottenberg,  Kaliski,  and  Friedmann.     Jour.  Med.  Bes.,  Vol.  28,  1913. 

80  Unpublished  personal  communication. 

81  Hada  and  Rosenthal.     Zts.  f.  Imm.,  1913,  16,  p.  524. 

32  Maragliano.     "IX  Kongr.  f.  Innere  Med.,"  1892. 

33  Moss.    Johns  Hop.  Hosp.  Med.  Bull,  March,  1910. 

34  Grafe  and  Graham.    Munch,  med.  Woch.,  1911,  pp.  2257,  2338. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      149 

the  work  of  Matsuo35  it  seems  likely  that  the  autolysins  of  par- 
oxysmal hemoglobinuria  are  not  identical  with  the  isolysins  a  and  ft  . 

Since  the  reintroduction  of  blood  transfusion  as  a  therapeutic 
measure  the  occurrence  of  hemolysis  between  the  blood  of  two  human 
beings  has  become  of  great  practical  importance.  A  number  of  seri- 
ous or  fatal  accidents  following  the  transfusion  of  hemolytic  blood 
have  been  reported.  Ottenberg  and  Kaliski36  have  shown  that  it 
is  possible  regularly  to  avoid  such  accidents  by  preliminary  blood 
tests. 

These  tests  are  easily  carried  out  by  obtaining  serum  and  washed 
blood  cells  from  both  prospective  recipient  and  donor,  and  testing 
them  one  against  the  other  for  hemagglutination  and  hemolysis,  as 
follows  : 

1.  Active  Serum  Donor,  0.5  c.  c.  +  Red  Cells  Recipient,  0.5  c.  c. 

(5%  emulsion  in  NaCl) 

2.  Active  Serum  Recipient  0.5  c.  c.  +  Red  Cells  Donor  0.5  c.  c. 


Controls  of  both  varieties  of  cells  in  salt  solution. 

Such  tests  should  be  observed  for  at  least  two  hours  before  final 
readings  are  taken. 

Although  we  have  by  no  means  covered  in  detail  the  entire  ex- 
perimental plan  followed  by  Ehrlich  and  his  collaborators  during 
their  early  work,  we  are  now  ready  to  consider  the  basic  views  on 
the  structure  of  the  lytic  antibodies  which  they  deduced. 

It  appears  from  the  preceding  that  the  thermostable  hemolytic 
antibody  must,  of  necessity,  unite  with  the  red  cell  before  the  com- 
plement or  alexin  can  exert  its  action  upon  it.  Ehrlich  conceives 
this  process  as  a  mediation  on  the  part  of  the  heat-stable  substance 
between  the  antigen  and  the  alexin  or  complement.  The  heat-stable 
body,  which  he  calls  "amboceptor,"  because  of  its  assumed  mode  of 
action,  possesses  two  combining  groups  —  one  the  "cytophile,"  by 
means  of  which  it  is  anchored  to  the  sensitive  cell,  the  other  the 
"complementophile,"  by  means  of  which  it  exerts  affinity  for  the 
complement.  The  original  cell  receptor,  from  which  such  an  "ambo- 
ceptor"  takes  its  origin,  is  one  which  not  only  can  combine  with  the 
antigenic  substance  offered  for  assimilation,  but  which  also  possesses. 
another  atom  group  by  means  of  which  it  can  enlist  the  aid  of  the 
digestive  ferment  of  the'  blood,  the  alexin  or  complement.  Cast  off 
into  the  blood  stream,  as  a  result  of  overregeneration,  it  now  appears 
as  a  "double"  receptor,  which  can  form  a  link  between  antigen  and 
complement. 

35  Matsuo.    D.  Arc.  f.  kl  Med.,  Bd.  107  H4,  p.  335. 
,  3G  Ottenberg  and  Kaliski.     J.  A.  M.  A.,  Vol.  61,  1913. 


150 


INFECTION    AND    RESISTANCE 


0NTIGEN 
BACTERIflLO 
OTHER  CELL 


flHBOCEPTOZ 
ANTIGEN 


& 

SCHEMATIC  EEPRESENTATIONS  OF  A  EECEPTOR  OF  THE  THIRD  ORDER. 
Ehrlich  ?s  conception  of  the  relationship  of  antigen,  amboceptor,  and  complement 
in  the  bactericidal  and  hemolytic    process.     In  A  the  receptor  is  still  a  part 
of  the  body  cell,  in  B  it  has  been  overproduced,  and  is  free  in  the  circulating 
blood. 

In  his  general  scheme  of  diagrammatic  representation  of  these 
processes  Ehrlich  refers  to  the  aamboceptors"  as  "haptines"  of  the 
third  order. 

Now  it  is  quite  plain,  from  the  extreme  specificity  which  results 
when  an  animal  is  immunized  with  any  given  variety  of  blood  cells 
or  bacteria,  that  there  must  be  as  great  a  variety  of  such  amboceptors 
as  there  are  different  antigens,  and  indeed  an  animal  immunized 
with  two  or  more  antigens  may  simultaneously  contain  in  its  blood 
serum  a  corresponding  number  of  different  amboceptors. 

This  assumption  of  the  multiplicity  of  "amboceptors"  in  the 
same  serum  is,  of  course,  forced  upon  us  by  the  fact  of  specificity, 
and  the  frequently  repeated  observation  that  the  same  serum  may 
contain  heat-stable  lytic  antibodies  against  a  variety  of  antigens,  each 
antigen  absorbing  out  of  such  a  serum  that  antibody  only  which 
specifically  reacts  with  it.  This  fact  has,  of  course,  never  been 
denied,  and  it  is  a  frequent  misunderstanding  of  the  views  of  Bor- 
det,  which  will  be  discussed  directly,  to  assume  that  he  has  combated 
the  "multiplicity  of  amboceptor"  in  the  sense  just  outlined.  Ehrlich 
.and  Morgenroth,  however,  have  expressed  themselves  in  favor  of  the 
conception  of  a  multiplicity  of  "amboceptor"  not  only  in  this  sense, 
but  as  occurring  in  response  to  immunization  with  one  and  the 
;same  antigen. 

Ehrlich  and  Morgenroth37  assume  that  any  cellular  antigen, 
blood  or  bacterial  cell,  substances  of  great  complexity  of  chemical 
structure,  must  necessarily  be  possessed  of  a  large  number  of  differ- 
ent side  chains  or  receptors.  When  immunization  is  practiced  with 
such  cells  a  correspondingly  varying  number  of  different  ambocep- 
tors must  result.  They  found,  for  instance,  that  when  rabbits  are 
37  Ehrlich  and  Morgenroth.  Berl.  klin.  Woch.,  Nos.  21  and  22,  1901. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      151 


immunized  with  ox  blood,  the  resulting  antiserum  is  capable  of  pro- 
ducing hemolysis  not  only  of  ox  blood  but  of  goat's  blood  as  well, 
though  to  a  lesser  degree.  They  conclude  from  this  that  the  hemo- 
lytic  action  of  the  serum  must  be  referred  to  the  presence  of  at  least 
two  kinds  of  amboceptor,  especially  since  repeated  experiments  with 
different  anti-ox-blood  sera  showed  that  there  was  no  regularity  in 
the  proportions  of  hemolysins  for  ox  and  goat  blood,  respectively. 
This  opinion  they  further  fortify  by  showing  that  exposure  of  the 
serum  to  ox  blood  deprives  it  of  all  its  hemolysins,  both  those  for 
ox  and  those  for  goat's  blood,  whereas  absorption  with  goat's  blood 
alone  removes  the  specific  goat's  blood  hemolysins  only.  They  trans- 
late their  understanding  of  the  conditions  to  graphic  form  by  the 
following  diagram:38 

If  ox  blood  is  in- 
jected, «  and  ft  receptors 
being  present,  <*  and  ft 
amboceptors  are  formed, 
and  ox  blood  can  conse- 
quently anchor  both  am- 
boceptors. The  presence 
of  ft  receptors  in  goat's 
blood  also  explains  the 
moderate  hemolysis  of 
this  blood  by  the  anti- 
serum,  but  lacking  the  <* 
receptors  which,  in  this 
case,  represent  the  larger 
proportion,  these  blood 
cells  cannot  remove  all 
the  amboceptor  for  ox 
blood  out  of  the  serum. 
The  example  given,  of 

course,  represents  the  simplest  assumed  case,  and  Ehrlich  and  Mor- 
genroth  believe  that  the  same  blood  or  bacterial  cells  may  possess  an 
entire  series  of  such  receptors,  some  of  them  being  dominant  for  the 
given  species,  others  being  merely  secondary  or  "partial,"  in  varying 
proportions. 

If  we  grant  the  fundamental  premises  of  Ehrlich  respecting  the 
"double  receptor"  or  "amboceptor"  nature  of  the  specific  antibody 
and  its  mediation  between  antigen  and  complement  by  means  of  a 
cytophile  and  a  complementophile  receptor,  certain  logical  conse- 
quences of  this  conception  suggest  themselves,  which,  in  their  many 
ramifications,  have  been  the  subject  of  much  investigation.  And 
although  many  phases  of  these  researches  are  no  longer  commonly 
accepted,  some,  indeed,  being  untenable  in  the  light  of  more  recent 
38  Ehrlich.  "Gesammelte  Arbeiten,"  p.  147. 


OX 


GO/77 


SCHEMATIC    KEPRESENTATION    OF    EHRLICH    AND 
MORGENROTH'S  CONCEPTION  OF  THE  COMPLEX 
STRUCTURE  OF  AN  ANTIGEN. 
(After  Ehrlich  and  Morgenroth.     Berl.  Iclin. 
Woch.,  Vol.  38,  1901.) 


152  INFECTION    AND    RESISTANCE 

discoveries,  the  influence  of  this  work  upon  the  development  of 
immunology  has  been  so  important  that  it  must  be  briefly  reviewed 
in  order  that  controversial  questions  may  be  justly  considered. 

The  comparison  of  the  action  of  hemolytic  sera  with  that  of  fer- 
ments, and  the  possibility  of  producing  antiferments  by  the  injection 
of  the  ferments  into  animals,  obviously  suggests  a  similar  induction 
of  antihemolysins  by  the  treatment  of  animals  with  lysins.  This, 
we  have  seen,  was  the  method  employed  by  Ehrlich  and  Morgenroth 
in  their  studies  of  the  causes  of  the  failure  of  autolysin  formation 
in  goats.  They  extended  this  work  with  the  purpose  of  ascertaining 
whether  or  not  there  were  differences  in  the  structure  of  the  cyto- 
phile  groups  of  the  various  amboceptors  formed  when  various  ani- 
mals were  injected  with  any  given  species  of  red  blood  cells.  After 
obtaining  a  strong  hemolytic  serum  by  injecting  ox  blood  into  a 
rabbit,  they  treated  a  goat  with  the  inactivated  serum  of  this  rabbit. 
The  result  was  that  the  serum  of  the  goat  so  treated,  when  mixed 
with  ox  blood  cells  and  the  hemolytic  serum,  prevented  the  sensiti- 
zation  of  the  cells  by  the  hemolysin.  They  then  measured  the  neu- 
tralizing power  of  such  an  "anti-amboceptor"  or  "anti-sensitizer" 
against  a  variety  of  hemolytic  sera  produced  with  ox  blood  in  differ- 
ent animals  and  found  that,  while  this  "anti-amboceptor"  neutralized 
the  hemolytic  action  of  an  antiserum  produced  in  rabbits,  it  had  but 
an  indifferent  or  entirely  ineffective  neutralizing  power  upon  sim- 
ilar ox  blood  hemolysins  derived  from  goats,  geese,  dogs,  rats,  or 
guinea  pigs.  They  concluded  from  this  that,  although  these  various 
lysins  had  been  produced  in  the  different  animals  by  the  injection  of 
the  same  antigen,  viz.,  ox  blood,  and  possessed  affinity  for  the  ox 
blood  in  consequence,  they  must  necessarily  differ  from  each  other  in 
some  way,  since  they  were  not  equally  neutralized  by  the  same  anti- 
lysin.  It  seemed  to  them  that  the  difference  in  such  cases  must 
depend  upon  variations  in  the  structure  of  the  cytophile  group  of 
the  amboceptor,  a  conclusion  which  they  based  upon  the  foregoing 
experiments  and  sought  to  support  by  the  following  reasoning: 
When  an  animal  is  treated  with  sensitizers  or  amboceptors,  they 
reasoned,  these  bodies  react  with  the  tissue  cells  by  means  of  the 
cell-receptors.  These  receptors  are  then  overproduced  and  extended 
into  the  circulation  as  free  atom-groups. 

They  now  act  as  "anti-amboceptor,"  free  in  the  serum,  but  are 
in  structure  merely  overproduced  cell  receptors,  identical  with  those 
which  originally  united  on  the  cell  with  the  injected  amboceptor. 

Ehrlich  and  Morgenroth,39  therefore,  believed  that  the  neutral- 
ization of  the  amboceptor  by  the  antilysin  depended  upon  a  union  of 
the  latter  with  the  "cytophile"  group  of  the  former,  preventing  its 
subsequent  union,  with  the  red  cells.  And  since  one  and  the  same 
antilysin  did  not  thus  invalidate  the  action  of  all  the  amboceptors 

39  Ehrlich  and  Morgenroth.    Berl.  kl  Woch.,  No.  22,  1901,  p.  600. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      153 


for  ox  blood  (derived  from  different  animals),  they  concluded  that 
these  "amboceptor"  must  possess  different  "cytophile  groups." 

That  this  conclusion  of  Ehrlich  and  Morgenroth  is  not  correct 
seems  to  follow  the  subsequent  work  of  Bordet.40  He  demonstrated 
that  it  is  not  necessary  to  inject  animals  with  specific  hemolytic  sera 
in  order  to  obtain  antilytic  sera,  but  that  the  same  object  may  be 
attained  by  injecting  animals  with  the  normal  serum  of  an  un- 
treated animal.  Moreover,  if  an  "antisensitizing"  serum  so  pro- 
duced was  added  to  corpuscles  which  had  al- 
ready absorbed  "amboceptor,"  it  prevented  the 
subsequent  union  of  these  sensitized  cells  with 
alexin  or  complement.  From  this  it  becomes 
clear  that,  in  the  first  place,  the  antisensitizer 
or  anti-amboceptor  cannot  be  identical  with  the 
cell  receptors  of  the  corpuscles,  and,  further, 
that  the  inhibition  of  the  hemolysis  which  such 
an  antisensitizer  exerts,  cannot  be  due  to  union 
with  the  "cytophile"  group.  This  both  contra- 
dicts the  Ehrlich  conception  of  the  mechanism 
of  "anti-amboceptors"  and  invalidates  his  ar- 
gument, in  this  instance,  in  favor  of  the  plural- 
ity of  the  amboceptors  produced  by  the  injec- 
tion. 

Bordet's  experiments  were  later  confirmed 
by  Ehrlich  and  Sachs,41  who  admit  the  error 
of  the  former  "anticytophile"  interpretation 
of  Ehrlich  and  Morgenroth's  experiments,  but 
they  still  maintain  that  Bordet' s  experiments 
do  not  disprove  the  conception  of  an  "ambo- 
ceptor" or  "Zwischenkorper"  of  Ehrlich.  They 
claim  that  Bordet's  results  merely  prove  that 
the  anti-amboceptor  or  anti-sensitizer  is  "anticomplementophile"  in- 
stead of  "anticytophile." 

The  principles  involved  we  will  discuss  in  another  place  in  con- 
nection with  Moreschi's  analysis  of  the  "anticomplements."  How- 
ever this  may  be,  we  may  conclude  that  Ehrlich  and  Morgenroth's 
differentiation  of  amboceptors  or  sensitizers  by  the  cytophile  group 
is  no  longer  valid. 

The  studies  of  Bordet  on  the  antisensitizers  (anti-amboceptor) 
had  important  results  apart  from  their  refutation  of  Ehrlich  and 
Morgenroth's  opinion.  In  addition  to  showing  that  such  antisensi- 
tizer did  not  represent  cell  receptors  identical  with  those  that  an- 
chored the  sensitizer  (amboceptor)  to  the  red  blood  cells,  his  experi- 
ments revealed  the  fact  that  such  an  antisensitizer  neutralizes  un- 

40  Bordet.    Ann.  tie  I'Inst.  Past.,  Vol.  18,  1904,  p.  593. 

41  Ehrlich  and  Sachs.    Berl.  klin.  Woch.,  No.  19,  1905. 


COMPLEMENT 


flVBOCEPTOR 


flNTMHBOCEPTOR 


SCHEMATIC  KEPRESEN- 
TATION  OF  EHRLICH 
AND  MORGENROTH  's 
CONCEPTION  OF  THE 
NEUTRALIZATION  OF 
A  HEMOLYTIC 
SERUM  BY  ANTILY- 
SIN  OR  ANTIAMBO- 
CEPTOR,  EEACTING 
WITH  THE  CYTO- 
PHILE GROUP.  (Ehr- 
lich and  Morgenroth, 
loc.  cit.) 

This  conception,  as  we 
shall  see,  has  be- 
come untenable. 


154  INFECTION    AND    RESISTANCE 

specifically  various  specific  sensitizers  as  well  as  normal  antibodies 
in  the  serum  of  the  same  animal;  and  this  showed  that  there  is  no 
necessity  of  assuming  a  variety  of  specific  antisensitizers,  as  had 
been  done  by  the  German  workers. 

As  regards  the  multiplicity  of  amboceptor  or  sensitizer,  however, 
though  the  proof  of  this,  by  means  of  anti-amboceptors,  has  had  to 
be  abandoned,  as  we  have  seen,  there  is  still  a  great  deal  of  evidence 
advanced  in  favor  of  such  an  assumption.  The  chief  support  for 
such  an  opinion  is  found  in  the  "group  reactions"  among  bacteria, 
similar  to  those  observed  for  blood  cells  by  Ehrlich  and  Morgenroth, 
and  described  above  (see  page  151).  For  it  is  frequently  observed 
that  the  antibodies  produced  by  immunization  with  one  species  of 
bacteria  may  have  a  certain  though  lesser  degree  of  action  upon 
other  related  forms,  these  in  turn  absorbing  only  a  part  of  the  ambo- 
ceptor out  of  the  serum,  while  the  species  originally  used  for  im- 
munization takes  out  all  the  amboceptor  present.  Considering  the 
great  chemical  complexity  of  the  bacterial  or  tissue  cells,  moreover, 
we  may  well  expect  such  multiplicity.  And  it  is,  indeed,  entirely 
reasonable  to  suppose  that  a  structure  as  complex  as  the  bacterial 
cell  may  contain  a  number  of  antigens  and  consequently  give  rise 
to  a  number  of  sensitizers  which  differ  in  that  each  is  specific  for  its 
particular  antigen  only.  This  is  merely  a  restatement  of  the  phe- 
nomenon of  specificity  and  has,  as  a  matter  of  fact,  no  modifying 
influence  on  the  general  principles  involved. 

From  the  point  of  view  of  a  general  understanding  of  the  proc- 
esses of  immunity,  however,  the  question  of  multiplicity  of  sensitizer 
is  not  so  fundamentally  important  as  is  the  similar  controversy  which 
has  been  waged  regarding  the  unity  or  multiplicity  of  alexin  or  com- 
plement. Here  again  there  has  been  some  misconception  as  to  the 
meaning  of  those  who  maintain  the  unity  of  alexin.  Neither  Bordet, 
nor  anyone  else  familiar  with  experimental  conditions,  has  ever  main- 
tained that  the  alexins  of  different  animals  were  functionally  iden- 
tical. It  is  a  well-known  fact  that  the  fresh  blood  sera  of  various 
animal  species  differ  from  each  other  considerably  in  their  power  to 
activate  bactericidal  or  hemolytic  systems.  In  regard  to  hemolysis, 
fresh  guinea-pig  serum  is  very  powerful  in  activating  many  sensi- 
tized blood-cell  complexes,  but  weak  in  activating  sensitized  guinea- 
pig  corpuscles.  Often  one  finds  that  the  alexin  of  an  animal  is  en- 
tirely impotent  or  but  weakly  capable  of  producing  hemolysis  of  the 
sensitized  cells  of  its  own  species,  though  this  is  not  a  general  rule. 

Again,  even  without  such  species  relationship,  a  given  alexin  may 
be  very  weak  for  certain  complexes  and  strong  for  others.  The 
alexin  of  horse  blood  can  even  be  fixed  to  sensitized  cells  42  without 

42  For  the  sake  of  clearness  it  may  be  repeated  here  that  by  sensitized 
cells  we  mean  cells  which  have  absorbed  specific  "amboceptor"  or  "sensitizer," 
and  have  thereby  become  amenable  to  the  action  of  alexin  or  complement. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      155 

producing  much,  if  any,  hemolysis.43  An  alexin  which  may  be  strong 
for  a  given  hemolytic  complex  may  be  weak  for  certain  bactericidal 
complexes,  or  vice  versa.  Thus  there  is  a  large  mass  of  evidence 
which  shows  that  no  two  alexins  are  exactly  alike,  though  the 
difference  between  them  can,  of  course,  be  defined  functionally 
only. 

The  difference  between  the  opinions  of  Ehrlich  and  his  school 
on  the  one  hand,  and  the  followers  of  Bordet,  on  the  other,  revolves 
not  about  this  point,  upon  which  all  agree,  but  about  the  question  of 
whether  one  and  the  same  serum  may  contain  more  than  one  alexin 
or  complement.  Ehrlich  and  Morgenroth44  and  Ehrlich  and 
Sachs  45  have  brought  forward  evidence  from  which  they  deduce  the 
existence  of  a  number  of  different  alexins  or  complements  for  hemo- 
lytic complexes  in  the  same  serum.  The  earlier  experiments  of 
Ehrlich  and  Morgenroth  on  this  question  were  carried  out  by  means 
of  the  filtration  of  normal  goat  serum  through  Pukall  filters ; 46  in 
these  it  appeared  that  the  serum  which  passed  through  the  filters  was 
complementary  for  sensitized  guinea-pig  cells,  while  that  part  which 
had,  in  the  original  serum,  activated  sensitized  rabbit  cells  was  left 
behind.  Similar  differentiation  of  complement  they  later  based 
upon  experiments  with  anticomplementary  sera  which,  they  showed, 
did  not  equally  neutralize  all  the  complementary  functions  of  a 
serum. 

In  support  of  their  contention  Neisser47  described  two  comple- 
mentary substances  in  rabbit  serum,  the  one  active  for  bactericidal 
complexes,  the  other  for  hemolytic,  and  similar  experimental  evi- 
dence has  been  brought  forward  by  Wassermann48  for  guinea-pig 
and  by  Wechsberg 49  for  goat  serum. 

The  evidence  advanced  by  these  writers  is  based  chiefly  on  ex- 
periments in  which  it  was  found  that  a  normal  serum  which  pos- 
sessed both  bactericidal  and  hemolytic  powers  could  be  deprived  of 
the  complement  for  one  or  the  other  of  these  activities  only,  by  ab- 
sorption with  the  respective  cells.  In  addition  to  this,  Ehrlich  and 
Morgenroth,  Ehrlich  and  Sachs,50  Wendelstadt,51  and  others,  claimed 
to  have  differentiated  various  complements  in  the  same  serum  by 
careful  heating,  by  the  action  of  weak  acids  or  alkalis,  or  such 
methods  as  the  digestion  of  sera  by  papain. 

43  Browning.     Wien.  klin.  Woch.,  No.  15,  1906. 

44  Ehrlich  and  Morgenroth.    Berl.  kl.  Woch.,  No.  31,  1900. 

45  Ehrlich  and  Sachs.    Berl.  kl  Woch.,  No.  21,  1902. 

46  Sachs.    Berl.  kl.  Woch.,  Nos.  9  and  10,  1902. 

47  Neisser.    Deutsche  med.  Woch.,  1900,  p.  790. 

48  Wassermann.     Zeitschr.  f.  Hyg.,  37,  1901. 

49  Wechsberg.    Zeitschr.  f.  Hyg.,  Vol.  39,  1902. 

60  Ehrlich  and  Sachs.    Berl.  kl.  Woch.,  Nos.  14  and  15,  1902. 
51  Wendelstadt.     Centralbl.  f.  Bakt.,  I,  Vol.  31,  1902. 


156  INFECTION    AND    RESISTANCE 

As  a  rule,  these  experiments  have  been  carried  out  with  normally 
hemolytic  serum  and  unsensitized  cells,  though  in  certain  cases 
Ehrlich  has  employed  sensitized  cells;  but  whenever  this  was  done 
exposure  to  complement  for  purposes  of  absorption  has  been  for 
much  briefer  periods  than  when  normal  serum  was  used.  This  point 
is  significant  when  we  come  to  consider  the  objections  to  the  inter- 
pretation of  the  preceding  experiment  in  favor  of  a  plurality  of  com- 
plement, objections  raised  chiefly  by  Wilde  52  and  by  Bordet. 

Wilde  refuted  particularly  the  experiments  of  Neisser,  who 
claimed  that  the  absorption  of  fresh  rabbit  serum  with  anthrax 
bacilli  deprived  this  serum  only  of  its  bactericidal  but  not  of  its 
hemolytic  complement.  Wilde  showed  that,  if  a  sufficient  excess  of 
anthrax  bacilli  (or  in  given  cases  of  typhoid  bacilli  or  cholera  spir- 
illa) were  added,  both  bactericidal  and  hemolytic  complement  could 
be  absorbed  from  normal  serum.  He  concludes  that  there  is  actually 
only  one  alexin  present,  but  that  the  red  cells  and  anthrax  bacilli 
differ  in  their  susceptibility  to  this  alexin  (or,  in  other  words,  that 
the  sensitization  of  these  cells  by  the  normal  serum  is  unequal,  a 
conclusion  which  seems  rational  in  view  of  the  fact,  now  well  known, 
that  one  and  the  same  complement  may  differ  greatly  in  the  degree 
of  its  activity  upon  different  sensitized  complexes. 

Bordet  has  analyzed  the  conditions  in  a  similar  way.  He  found 
that  absorption  of  normal  serum  with  unsensitized  cells  rarely  de- 
prived this  serum  of  all  of  its  alexin,  even  when  these  cells  were  used 
in  considerable  amounts.  This  he  attributed  to  the  feeble  sensitiza- 
tion of  the  cells.  If,  however,  strongly  sensitized  cells  were  added 
to  such  a  normal  serum,  all  the  alexin  would  be  taken  up.  He  refers 
the  phenomenon  of  specific  alexin  absorption,  observed  by  previous 
workers,  to  insufficiency  in  the  perfection  of  sensitization  on  the 
part  of  the  cells  used  in  the  preliminary  exposure ;  and  subsequent 
work  with  complement  fixation  seems  to  bear  him  out. 

Most  of  these  arguments,  though  they  seem  to  us  perfectly  valid 
in  the  light  of  the  experimental  facts,  have  been  answered  by  Ehrlich 
and  his  school  by  the  assumption  of  the  existence  of  so-called  "poly- 
ceptors."  Ehrlich  now  admits  that  the  amboceptors  cannot  be  shown 
to  differ  from  each  other.  However,  he  does  not  believe  that  differ- 
ences in  the  intensity  of  sensitization  explain  variation  in  the 
functional  efficiency  of  different  complements  upon  sensitized  cell 
complexes,  nor  does  he  accept,  for  proof  of  this,  the  fact  that  comple- 
ment may  be  entirely  absorbed  out  of  a  serum  by  a  complex,  even 
though  the  complement  may  be  comparatively  inefficient  as  an  acti- 
vator in  the  given  case.  He  assumes  that  the  sensitizer  or  "ambo- 
ceptor"  may  possess  a  number  of  complementophile  groups  (poly- 
ceptors),  by  means  of  which  a  number  of  different  complements  may 

52  Wilde.  Habilitations  Schrifft,  Munich,  1901.  Also  Berl  kl.  WocJi., 
No.  34,  1901. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      157 


become  active  in  the  given  case.  Thus,  although  such  a  polyceptor, 
of  course,  is  capable  of  uniting  with  the  complement  which  activates 
the  dominant  complement,  it  is  capable  also 
of  union  with  a  number  of  other  comple- 
ments  which  have  slight  or  no  functional 
action  whatever — the  non-dominant  com- 
plements. This  opinion  is  rendered  dia- 
grammatic by  Ehrlich  and  Marshall 53  in 
the  following  way : 

If  one  carefully  considers  the  reasons 
advanced  for  the  assumption  of  the  exis- 
tence of  such  polyceptors  it  does  not  seem 
that  they  are  sufficiently  forcible  to  lead 
one  to  desert  the  much  simpler  explanation 
of  Bordet. 

Related  to  the  problems  discussed  in 
connection  with 
the  production 
of  "anti-ambo- 
ceptors"  or  "an- 
t  i  sensitizers" 
are  those  which 
have  arisen  re- 
garding the  ex- 
istence of  "  anti- 
complement"  or 
"anti  -  alexins." 

Ehrlich  and  Morgenroth  claimed  that,  by 
the  injection  of  active  horse  serumvinto  a 
goat,  they  had  obtained  substances  in  the 
goat  serum  which  neutralized  horse  com- 
plement. They  believed  that  the  "anti- 
complements"  thus  produced  neutralized 
the  complement  by  uniting  with  its  hapto- 
phore  group,  thus  preventing  its  combina- 
tion with  the  acomplementophile  group" 
of  the  amboceptor.  This  was  their  con- 
clusion because  they  found  that  the  "anti- 
complementary"  serum  exerted  no  protec- 
tive influence  upon  sensitized  cells,  when 
these  were  exposed  to  the  serum  and  then 
removed,  but  that  it  protected  against 
hemolysis  when  added  to  the  cells  together 

with  the  complement.     There  was  apparently  no  union  of  the  pro- 
tective substance  with  the  "complementophile"  group  of  the  ambo- 
53  Ehrlich  and  Marshall.    Berl  kl.  Woch.,  No.  25,  1902. 


POLYCEPTOR    ACCORDING    TO 
EHRLICH  AND  MARSHALL. 

(a)  Eeceptor  of  the  Cell. 

(b)  Haptophore    Group    of 
the  Amboceptor. 

(c)  Dominant    Complement. 

(d)  Secondary  Complements. 
Complementophile  Groups  of 

the  Amboceptor: 

(1)  for  the  Dominant  Com- 
plement. 

(2)  for  the  Secondary  Com- 
plement. 

(After  Ehrlich  and  Marshall, 
Berl.  Uin.  Woch.,  No. 
25,  1902.) 


EHRLICH  AND  MORGEN- 
ROTH's  CONCEPTION  OF 
THE  ACTION  OF  ANTI- 
COMPLEMENT. 

A.  Scheme  of  Hemolysis. 

B.  Action    of    Anticomple- 
ment  upon  Hemolysin. 

b.  —  blood  cell,  c.  =  com- 
plement, i.  =  immune 
body,  a.  =  anticomple- 
ment. 

The  complementoids  are 
not  included  in  the 
scheme,  since  in  this 
case  they  are  without 
influence. 


158  INFECTION    AND    RESISTANCE 

ceptor,  but  the  protecting  substance  did  act  in  direct  antagonism  to 
the  complement  itself. 

From  the  fact  that  similar  anticomplements  could  be  produced 
when  inactivated  serum  was  injected  into  animals,  they  concluded 
that,  on  inactivation,  there  was  not  a  complete  destruction  of  the 
complement,  but  that  during  the  process  of  heating  the  zymophore 
group  of  the  complement  only  was-  injured,  the'  "haptophore  group," 
by  means  of  which  union  to  the  tissue  elements  would  take  place, 
and  through  which,  therefore,  specific  antibody  production  would  be 
incited,  remaining  intact.  Such  altered  complement  they  speak  of 
as  "complementoid." 

Bordet  has  made  similar  observations  upon  the  production  of  anti- 
alexins  by  the  injection  into  animals  both  of  active  and  of  inactive 
serum,  but  in  the  light  of  further  researches,  which  will  be  discussed 
in  connection  with  the  problems  of  alexin-fixation,  chiefly  those  of  Mo- 
reschi  and  of  Gay,  we  are  forced  to  the  conclusion  that  the  existence  of 
true  anticomplements-  is  by  no  means  certain,  and  that  the  older  evi- 
dence in  their  favor  is  found  to  be.  unconvincing  at  the  present  time. 

In  the  preceding  paragraphs  we  have  emphasized  the.  conceptions 
of  the  cytolytic  phenomena  formulated*  by  Ehrlich  and  his  followers, 
and  although  we  have  brought  out,  whenever  possible,  »the  objections 
of  other  investigators  to  many  of  these  opinions,  we  have  not  yet 
followed  out  in  a  systematic  manner  the  reasoning  of  any  of  Ehrlich's 
opponents.  In  opposition  to  the  views  of  his  school  the  leading 
position  has  been  taken  by  Bordet,  who,  after  all,  furnished  in  his 
investigations  the  fundamental  facts  which  have  led  to  a  comprehen- 
sion of  the  cytolytic  processes.  In  explaining  Bordet's  views  we  can 
do  no  better  than  to  follow  out  his  own  exposition  as  set  forth  in  his 
article,  "A  General  Resume  of  Immunity/'  54  published  with  a  collec- 
tion of  his  papers.  He  expresses  himself,  in  substance,  as  follows : 

That  the  antigen,  in  the  form  of  bacteria,  blood  cells,  or  cells  of 
any  other  nature,  meets  in  the  body  of  the  treated  animal  a  "recep- 
tor" complex  with  which  it  unites  is,  of  course,  plain  and  agreed  to 
by  everyone.  That  the  antibody  produced  by  the  tissues  in  response 
to  such  union  of  antigen  with  receptor  is  a  direct  product  of  the  cells 
containing  the  receptors  is  likely.  It  is  by  no  means  certain,  how- 
ever, or,  at  any  rate,  it  has  never  been  experimentally  demonstrated, 
that,  as  Ehrlich  maintains,  the  antibody  is  identical  with  the  original 
receptor  by  which  the  antigen  was  fixed  or  anchored  to  the  tissue 
cell.  It  might  be  assumed  with  equal  justice  that  the  cells  of  the 
immunized  animal  could  build  up  a  new  substance,  not  identical  with 
the  receptors,  in  consequence  of  stimulation  by  the  antigen.  It  is 
also  by  no  means  certain  whether  the  injected  antigen  reacts  with 
the  body  cells  themselves  or  with  the  normal  antibodies  which  we 

54  "Studies  in  Immunity"  oy  Bordet  and  collaborators.  Gay,  Wiley  & 
Sons,  N.  Y.,  1909. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      159 

know  to  exist  in  many  cases.  Thus  the  blood  serum  of  goats  may 
normally  often  contain  hemolysins  against  rabbit  corpuscles.  Is  it 
not  reasonable  to  suppose  that  possibly  these  may  furnish  the  point 
of  attachment  and  the  source  of  further  antibody  production  when 
rabbit  cells  are  injected  into  goats?  In  criticism  of  Ehrlich's  as- 
sumption of  the  mode  of  action  of  heat-stable  lytic  antibody,  Bordet 
very  justly  maintains  that  no  proof  whatever  exists  of  the  "ambo- 
ceptor"  nature  of  this  substance.  All  that  is  certain  is  that  the 
stable  substance  must  unite  with  the  antigen  before  the  alexin  or 
complement  can  exert  its  action  upon  it  or  be  fixed  by  it.  There  is 
no  entirely  valid  proof  of 
the  existence  in  this  anti- 
body of  a  "complemento-  co 
phile"  and  a  "cytophile" 

1  f«     f  SENSITIZE* 

group,  and  no  satistactory 
instance  has  been  observed 
in  which  alexin  has  united 
with  a  heat-stable  anti- 
body which  has  not  previ-  *********< 
ously  been  united  with  an 
antigen.55  All  that  has 
been  shown  is  that  the  an-  SCHEMATIC  EEPRESENTATION  OF  BORDET  's  VIEW 
,.  ,1  ..,  .  ,  CONCERNING  THE  INABILITY  OF  COMPLEMENT 
tigen,  together  with  its  T0  UNITE  WITH  ElTHER  ANTIGEN  OR  SEN- 

specific  antibody,  forms  a          SITIZER    ALONE   AND    ITS    ABILITY    TO    BE 

pornr>W    w  h  i  P  h    ha<*    an  FIXED    BY    THE    COMPLEX    FORMED    WHEN 

.ompiex   w  hi  c  h  THE  ANTIGEN  Is  SENSITIZED. 

avidity  for  alexin,  a  com-     Compare   this   figure   with   that   representing 
plex    which    IS    "endowed  Ehrlich's  conception  of  the  same  process, 

with  properties  of  absorp- 
tion for  complement  which  neither  of  its  constituents  alone  possesses." 
Bordet  speaks  of  the  "amboceptors,"  therefore,  as  "sensitizers," 
meaning  by  this  that  the  antigen,  by  union  with  its  antibody,  is  sensi- 
tized to  the  action  of  the  alexin.  The  term  "sensitizers"  in  no  way, 
therefore,  implies  a  preconceived  notion,  experimentally  unproved, 
of  the  mode  of  action  or  structure  of  the  sensitizer.  Since  we  have 
graphically  explained  Ehrlich's  opinions,  a  similar  diagrammatic 
representation  may  be  permitted  of  Borders  opinion  of  the  same 
process  of  union  of  antigen  and  heat-stable  antibody  with  the  conse- 
quent development  of  alexin-fixing  property. 

In  this  diagram  the  ability  to  absorb  or  unite  with  complement 
becomes  evident  only  after  a  complex  has  been  formed  by  the  union 
of  the  two  elements,  antigen  and  antibody.  The  diagram  must  not 
be  assumed  to  mean  that  the  notch  into  which  the  complement  fits 
symbolized  necessarily  an  "atom  group,"  but  merely  expresses  the 
idea  of  "ability  to  absorb  alexin,"  not  assuming  that  this  ability  is 

55  Refer  also  to  the  discussion  of  the  conglutinins  at  the  end  of  this 
chapter. 


160 


INFECTION    AND    RESISTANCE 


either  chemical  affinity  by  means  of  a  definite  atom  group  or  a  mere 
physical  change  of  molecular  equilibrium  permitting  a  specific  com- 
plement absorption.56 

It  will  be  seen  from  the  preceding  that  the  controversy  between 
Ehrlich's  "amboceptor"  conception  and  the  "sensitization"  idea  of 
Bordet  turns  largely  upon  the  existence  of  a  so-called  complemen- 
tophile  group  of  the  thermostable  antibody.  For  if  it  were  the  case 
that  this  antibody  possessed  an  atom  group  which  permitted  it  to 
unite  with  alexin,  independent  of  previous  union  with  antigen,  it 
would  go  far  to  support  Ehrlich's  view.  One  of  the  strongest  argu- 
ments brought  into  the  field  in  favor  of  such  an  occurrence  by  Ehr- 
lich's  followers  is  the  phenomenon  of  Neisser  and  Wechsberg,  which 
is  usually  spoken  of  as  "complement  deviation"  (Komplement  Ablen- 
kung). 

In  order  to  make  the  conditions  underlying  this  phenomenon 
clear,  it  will  be  of  advantage  to  consider  for  a  moment  the  methods 
of  determining  quantitatively  the  amount  of  bactericidal  antibody 
(sensitizer  amboceptor)  in  any  given  immune  serum,  since  it  was  in 
working  with  such  titrations  that  Neisser  and  Wechsberg  made  their 
observations. 

In  carrying  out  such  measurements,  -it  is  customary  to  add  in 
series,  to  constant  amounts  of  bacteria,  varying  amounts  of  inacti- 
vated antiserum  and  constant  amounts  of  complement  or  alexin. 
These  mixtures  are  set  away  in  the  thermostat  for  3  to  4  hours,  are 
then  mixed  with  agar  and  plates  are  poured.  The  colonies  which 
develop  will  give  an  indication  of  the  number  of  bacteria  killed  in 
each  mixture  when  compared  with  similar  plates  poured  from  tubes 
in  which  the  same  original  amounts  of  bacteria  had  been  mixed  with 
alexin  alone.  The  following  table  will  exemplify  such  a  test : 


Typhoid  bacilli 

Typhoid 
antiserum 
inactive 

Alexin 

Result  in  colonies 
after  3  hours, 
incubation 

Constant  quantity 

1           C    C. 

.07  c.  c. 

Many  thousand 

Constant  quantity  

.01        C.  C. 

.07  c.  c. 

Many  thousand 

Constant  quantity  

.005     C.  C. 

.07  c.  c. 

150  colonies 

Constant  quantity 

001    c.  c. 

.07  c.  c. 

200  colonies 

Constant  quantity  

.0005  c.  c. 

.07  c.  c. 

800  colonies 

Control  I,    constant  quantity.  . 

.07  c.  c. 

Many  thousand 

Control  II,  constant  quantity 

Many  thousand 

56  The  diagram  on  page  159,  though  possibly  not  expressing  with  absolute 
accuracy  the  idea  of  sensitization,  was  devised  because  it  will  remove  what 
seem  to  the  writer  frequent  misconceptions  of  Bordet's  views.  Statements  are 
found  in  the  literature  which  imply  (Ehrlich,  "Kraus  und  Levaditi  Hand- 
buch,"  Vol.  1,  p.  8)  that  Bordet  assumes  "dass  das  Komplement  direkt  an 
die  Zelle  angreift,"  and  deny  that  there  is  experimental  evidence  to  support 
this.  It  is  perfectly  true  that  there  is  no  evidence  to  show  such  "direktes 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      161 


MM 


Ufl 


In  this  table  it  is  noticeable  that,  although  there  has  been  con- 
siderable bactericidal  action  in  the  mixtures  in  which  0.005,  0.001, 
and  0.0005  c.  c.  of  antiserum  were  used,  the  mixtures  in  which  as 
much  as  0.1  and  0.01  c.  c.  were  present,  and  in  which  one  would  nat- 
urally expect  a  still  greater  antibacterial  action,  the  contrary  occurred. 
This  surprising  and  curious  phenomenon,  showing  that  an  excess  of 
antibody  could  ac- 
tually be  harmful 
to  the  functiona- 
tion  of  the  bacteri- 
cidal complex,  was 
explained  by  Neis- 
ser  and  Wechsberg 
by  t  h  e  following 
reasoning.  In  tests 
like  the  one  given 
above  a  limited 
amount  of  bacteria 
and  a  1  e  x  i  n  has 
been  mixed  with 
the  enormous 
amount  of  anti- 
body represented 
in  the  immune 
serum.  Although 
bacteria  can  absorb  more  of  this  antibody  than  is  necessary  for  their 
solution  or  destruction,  nevertheless  the  higher  concentration  given 
in  the  table  will  contain  quantities  of  "amboceptor"  so  far  in  excess 
of  the  amount  that  can  be  absorbed  that  much  of  it  must  remain 
free  in  the  fluid.  Now  this  amboceptor,  possessing  a  complemen- 
tophile  group,  is  able  to  anchor  complement  or  alexin  as  well  as  that 
which  has  become  united  with  the  bacteria.  In  consequence,  there 
being  only  a  limited  amount  of  complement,  some  of  this  is  deviated 
from  the  amboceptor-antigen  complexes  by  the  free  amboceptor,  and 
is,  in  consequence,  ineffective  so  far  as  bactericidal  action  is  con- 
cerned. In  the  higher  dilutions  of  the  antiserum,  in  which  no  such 
excess  is  present,  the  complement  will  be  concentrated  upon  the  "at- 
tached" or  "anchored"  amboceptor,  and  greater  efficiency  will  result. 
Graphically  Neisser  and  Wechsberg  express  their  idea  in  the  figure 
which  we  reproduce. 

As  to  the  accuracy  of  the  observations  of  Neisser  and  Wechsberg 
there  can  be  no  question,  and  everyone  who  has  occasion  to  carry  out 

Angreifen"  upon  the  unaltered  cell,  but  there  is  evidence  that  this  union 
takes  place  after  the  cell  has  absorbed  the  antibody,  and  no  satisfactory 
evidence  to  show  that  the  thermostable  body  is  an  intermediary,  that  is,  forms 
a  link  as  conceived  in  the  amboceptor  idea. 


COMPLEMENT  DEVIATION  AS  CONCEIVED  BY  NEISSER  AND 

WECHSBERG. 

The    complement   being   united   to    the   unbound    ambo- 
ceptor   is    thereby    deviated    from    the    amboceptor, 
which  has  gone  into  relation  with  the  antigen. 
(After  Neisser  and  Wechsberg,  Munch,  med.  Woch., 
1901,  p.  697.) 


162  INFECTION    AND    RESISTANCE 

bactericidal  tests  with  any  frequency  is  sure  to  meet  with  the  phe- 
nomenon again  and  again.  But  their  explanation,  which  involves 
the  assumption  of  union  between  free  sensitizer  or  amboceptor,  and 
alexin  or  complement,  without  the  participation  of  antigen,  cannot 
be  accepted  since,  search  as  we  may,  through  the  extensive  experi- 
mentation that  this  problem  has  inspired,  there  is  no  instance  on 
record  in  which  indisputable  evidence  of  such  an  occurrence  has 
been  advanced.  On  the  contrary,  there  is  a  mass  of  satisfactory 
evidence  available  which  indicates  clearly  that  amboceptor  or  sensi- 
tizer alone  cannot  absorb  alexin,  and  the  Neisser-Wechsberg  explan- 
ation seems  consequently  to  be  merely  an  interesting  and  cleverly 
conceived  but  improbable  possibility. 

What,  then,  is  the  explanation  of  the  diminution  of  bactericidal 
effect  in  the  presence  of  an  excess  of  sensitizer  ?  We  will  see  that,  in 
the  study  of  agglutinin  and  precipitin  reactions,  phenomena  exactly 
analogous  to  the  Neisser-Wechsberg  effect  have  been  noticed,  in  the 
case  of  the  agglutinins,  the  so-called  "pro-agglutinoid"  zone  being  a 
case  in  point.  For  these  phenomena,  as  well  as  for  that  of 
Neisser  and  Wechsberg,  explanations  have  been  advanced  by 
the  Ehrlich  school,  similar  in  principle  in  that  they  all  depend  upon 
more  or  less  arbitrary  assumptions  regarding  affinity  between  the 
reacting  bodies.  Such  explanations,  though  not  outside  the  realm  of 
possibility,  have,  however,  lost  much  force  since  it  has  been  recog- 
nized that  the  reactions  between  serum  antibodies  and  their  antigens, 
in  general,  take  place  according  to  laws  far  more  closely  analogous 
to  those  governing  reactions  between  colloids  than  to  those  governing 
chemical  reactions  in  which  the  laws  of  definite  proportions  can  be 
applied.  And,  indeed,  the  reacting  substances  in  antigen-antibody 
complexes  are,  beyond  doubt,  of  the  nature  of  colloids.  Now,  in 
many  precipitations  resulting  when  two  colloids  are  mixed,  an  ex- 
cess of  one  or  the  other  factor  will  completely  inhibit  the  occurrence 
of  the  precipitation;  the  reaction  taking  place  only  when  definite 
proportions  between  the  reacting  bodies  are  present.  The  occurrence 
of  such  inhibition  zones,  due  to  an  excessive  concentration  of  one 
reagent,  can  be  shown  for  agglutination  and  precipitation,  exactly 
as  it  can  in  ordinary  colloidal  reactions,  and  it  is  more  than  likely 
that  the  Neisser- Wechsberg  phenomenon  is  merely  an  example  of  a 
similar  phenomenon. 

Looked  at  from  this  point  of  view,  far  from  supporting  the  sup- 
position of  a  separate  complementophile  group  and  therefore  of  the 
"amboceptor"  nature  of  the  heat-stable  lytic  antibody,  the  Neisser- 
Wechsberg  phenomenon  indeed  becomes  rather  a  strong  argument 
in  favor  of  Bordet's  views,  and  against  those  of  Ehrlich.  For,  by 
introducing  the  analogy  between  the  lytic  and  bactericidal  processes 
with  colloidal  reactions,  it  takes  away  much  force  from  the  supposi- 
tion that  antigen-sensitizer  alexin  reactions  take  place  according  to 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      163 

laws  of  definite  proportion,  an  idea  which  still  underlies,  though 
somewhat  loosely,  many  of  the  more  important  views  of  antigen- 
antibody  reactions  as  conceived  on  the  basis  of  the  amboceptor  theory. 

Gay  has  suggested  also  that  the  ^sTeisser-Wechsberg  phenomenon 
may  well  be  explicable  on  the  basis  of  the  fixation  of  complement  by 
precipitates.  In  a  succeeding  section  we  will  discuss  the  fixation  of 
alexin,  which  occurs  when  a  dissolved  protein  is  brought  together 
with  its  specific  antiserum.  It  is  not  impossible  that  this  may  occur 
when  bacterial  emulsions,  from  which  a  small  amount  of  bacterial 
protein  may  well  go  into  solution,  are  brought  together  with  anti- 
serum  in  concentration.  Under  such  conditions  a  reaction  might 
readily  occur  which  would  lead  to  the  fixation  of  alexin  and  its  con- 
sequent deviation  from  the  sensitized  bacteria. 

Of  all  explanations  considered,  therefore,  that  of  Neisser  and 
Wechsberg  seems  to  be  the  least  likely.  It  would  seem  to  us  that 
Bordet's  interpretation  of  these  facts  is  borne  out  indirectly  by  cer- 
tain experiments  of  Morgenroth  and  Sachs57  themselves,  in  which 
the  mutual  quantitative  relations  between  complement  and  "ambo- 
ceptor" were  studied.  In  these  experiments  it  was  shown  th$,t  the 
more  highly  cells  were  sensitized,  the  smaller  was  the  quantity  of 
complement  which  was  needed  for  their  hemolysis,  and  vice  versa, 
the  less  the  sensitization  (the  smaller  the  quantity  of  amboceptor) 
the  more  complement  was  necessary  to  produce  the  same  result. 
The  following  extract  from  one  of  their  protocols  will  illustrate  this : 

BEEF  BLOOD  CELLS  5%,  1  C.  C.,    ANTIBEEF  GOAT  SERUM,  GUINEA-PIG  COMPLEMENT 


Amount  of 
amboceptor 

Relative  amount  of 
amboceptor 

Amount  of 
complement  for 
complete  hemolysis 

.05 

1 

.008 

.2 

4 

.0025 

.4 

8 

.0014 

A  similar  relation  may  be  observed  by  all  who  have  occasion  to 
work  with  hemolytic  reactions.  In  the  present  connection  this  seems 
to  bear  out  Bordet's  interpretation,  since,  knowing  the  differences  in 
functional  efficiency  of  various  complements  for  different  hemolytic 
and  bactericidal  complexes,  we  could  well  expect  that  insufficient 
sensitization  of  a  red  cell  or  bacterial  antigen,  not  particularly  amen- 
able to  the  complement  employed,  might  fail  to  absorb  it  completely 
out  of  the  serum,  thus  giving  a  negative  result  which  would  simulate 
complete  lack  of  affinity. 

This  research  of  Morgenroth  and  Sachs  seems  further  of  funda- 
57  Morgenroth  and  Sachs.    Berl  kl  Woch.,  No.  35,  1902. 


164 


INFECTION    AND    RESISTANCE 


mental  importance  in  its  contradiction  of  the  regularly  progressive 
quantitative  relations  which  strict  adherence  to  the  "amboceptor" 
idea  would  seem  to  impose. 

The  quantitative  relations  here  outlined  have  been  diagrammat- 
ically  represented  by  Noguchi  as  follows: 


-  Complement 
Purpto  -  Amboceptor 
Mtf  -  foemo/ysu 


I  unit  of  Amboceptor 
used  in  eac/i  tv/tn  various 
fractions  of- a  comp/emerrf 
unft 


* 


NOGUCHI  's  DIAGRAM  ILLUSTRATING  THE  QUANTITATIVE  EELATIONS  BETWEEN  ANTI- 
GEN, AMBOCEPTOR  AND  COMPLEMENT. 

< Taken  from  Noguchi,  "Serum  Diagnosis  of  Syphilis,"  Lippincott,  Philadelphia, 

1910.) 

The  essential  point  of  difference  between  the  opinions  of  Ehrlich 
and  Bordet  concerning  the  processes  of  hemolysis  and  bacteriolysis 
lies,  as  we  have  seen,  in  the  conception  of  the  union  of  alexin  or  com- 
plement with  amboceptor  or  sensitizer.  Although  Ehrlich  and  his 
followers  admit  that  the  union  of  complement  with  amboceptor  does 
not  usually  occur  unless  the  amboceptor  has  previously  united  with 
the  antigen,  they  still  maintain  that  this  may  occasionally  take  place 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      165 

in  the  case  of  special  complexes  in  which  the  complement  may  di- 
rectly unite  with  free  amhoceptor.  This,  we  have  seen,  is  the  hasis 
of  the  Neisser-Wechsberg  conception  of  complement — "Ablenkimg" 
or  deviation,  and  of  other  ramifications  of  this  theory.  Bordet,  on 
the  other  hand,  consistently  holds  that  alexin  or  complement  is  at- 
tached only  by  the  complex  antigen-sensitizer  (antigen-amboceptor). 
In  the  controversy  which  this  difference  aroused,  an  observation  was 
reported  by  Ehrlich  and  Sachs,58  which  seemed  to  represent,  as  they 
themselves  express  it,  an  "Experimentum  Crucis"  proving  Ehrlich's 
contention  of  the  intermediary  function  of  the  amboceptor  in  con- 
trast to  Bordet's  asensitization"  idea.  The  facts,  as  they  record 
them,  are  as  follows:  When  fresh  horse  serum  is  added  to  guinea 
pig  corpuscles,  slight  hemolysis  results.  When  inactivated  ox  serum 
alone  is  added  to  such  corpuscles,  of  course  no  hemolysis  results.  If 
the  corpuscles  are,  on  the  other  hand,  exposed  to  the  action  of  the 
inactive  ox  serum,  together  with  fresh  horse  serum,  very  active  hem- 
olysis is  brought  about.  Apparently  the  ox  serum  sensitizes  (or 
furnishes  amboceptor  to)  the  guinea  pig  corpuscles,  rendering  them 
amenable  to  the  action  of  the  complement  in  the  fresh  horse  serum. 
In  other  words,  inactivated  ox  serum  can  be  reactivated  by  the  addi- 
tion of  fresh  horse  serum.  From  this  one  would  expect  that  if  the 
guinea  pig  cells  were  exposed  to  inactive  ox  serum,  then  separated 
from  the  serum  by  centrifugalization  and  fresh  horse  serum  subse- 
quently added,  hemolysis  would  ensue.  However,  this  was  not  the 
case.  When  the  cells  were  so  treated  it  was  found  that  they  had 
not  been  sensitized,  and,  what  is  more,  it  could  be  shown  that  the 
ox  serum  so  employed  had  lost  none  of  its  ability  to  produce  strong 
hemolysis  when  added  to  another  complex  of  cells  and  fresh  horse 
serum.  Ehrlich  and  Sachs  concluded  that  this  experiment  defi- 
nitely showed  the  ability  of  the  amboceptor  in  the  ox  serum  to  unite 
with  alexin  independently.  The  relation  to  the  cell  occurred 
only  after  the  union  of  the  amboceptor  in  the  ox  serum  and  the 
complement  in  the  horse  serum  had  been  established,  and  if 
their  interpretation  is  correct,  of  course,  it  constitutes  strong- 
evidence  against  the  general  principle  of  "sensitization"  as  conceived 
by  Bordet. 

This  apparent  inability  of  the  corpuscles  to  absorb  amboceptor 
independently  out  of  the  inactivated  ox  serum,  and  the  fact  that 
hemolysis  results  only  if  the  corpuscles,  ox  serum,  and  fresh  horse 
serum  are  all  simultaneously  present,  are  extraordinary  and  not  at 
all  in  keeping  with  the  preceding  work  of  Ehrlich  and  Morgenroth, 
and  indeed  with  experience  of  these  phenomena  in  general.  It  is  log- 
ical therefore  to  examine  more  closely  the  peculiar  conditions  main- 
tained in  these  experiments  before  applying  the  reasoning  deduced 
from  obviously  different  phenomena  to  their  explanation. 

58  Ehrlich  and  Sachs.    Berl  klin.  Woch.,  No.  21,  1902. 


166  INFECTION    AND    RESISTANCE 

Bordet  and  Gay59  accordingly  studied  the  Ehrlich-Sachs  phe- 
nomenon carefully  and  obtained  results  which  confirmed  the  experi- 
mental data  of  these  writers  but  cast  much  doubt  upon  the  validity 
of  their  conclusions. 

In  going  over  the  experiments  of  Ehrlich  and  Sachs,  Bordet  and 
Gay  made  an  observation  which  had  apparently  escaped  the  atten- 
tion of  the  former  investigators.  Heated  bovine  serum  has  but  a 
slight  agglutinating  power  for  guinea  pig  corpuscles.  Fresh  horse 
serum  agglutinates  them  only  slightly  and  slowly.  On  the  other 
hand  a  mixture  of  the  two  sera  agglutinates  them  very  rapidly  and 
completely.  The  bovine  serum  apparently  possessed  an  accelerating 
or  fortifying  influence  both  upon  the  weakly  active  normal  hemoly- 
sins  and  agglutinins  in  the  horse  serum.  Bordet  and  Gay  conse- 
quently suspected  that  this  property  might  be  due  to  an  undescribed 
substance,  peculiar  to  the  bovine  serum.  To  eliminate  the  uncertain 
elements  obtaining  in  experiments  in  which  normal  sensitizer  is 
used  they  now  experimented  with  guinea  pig  corpuscles,  anti-guinea 
pig  sensitizer  (from  a  rabbit  immunized  with  guinea  pig  blood  cells) 
and  guinea  pig  alexin. 

They  found  that  sensitized  guinea  pig  cells  are  hemolyzed  by 
guinea  pig  alexin  very  slowly  and  imperfectly,  as  is  often  the  case 
when  the  alexin  comes  from  the  same  animal  species  as  the  cells. 
When  heated  bovine  serum  was  added  to  the  complex  of  sensitized 
cells  and  alexin,  rapid  agglutination  and  hemolysis  resulted.  Their 
experiments  may  be  tabulated  as  follows : 

1.  Cells  +  guinea  pig  alexin  +  heated  bovine  serum  =  no  agglutination; 

very  slight  hemolysis  on  next  day. 

2.  Cells  H-  sensitizer  +  heated  bovine  serum  =  slight  agglutination;  no 

hemolysis. 

3.  Cells  +  sensitizer  +  alexin  -f  bovine  serum  =  powerful  agglutination 

and  complete  hemolysis  in  10  minutes. 

4.  Cells  +  sensitizer  -f-  alexin  =  very  slight  agglutination  and  incomplete 

hemolysis  in  30  minutes. 

5.  Cells  +  sensitizer  =  slight  agglutination;  no  hemolysis. 

In  tube  (1)  the  slight  hemolysis  was  due  to  the  small  amount  of 
normal  sensitizer  present  in  the  bovine  serum,  and  the  slight  agglu- 
tination in  tube  (5)  is  referable  to  the  agglutinating  power  of  the 
sensitizer.  In  tube  (3)  we  see  the  powerfully  accelerating  effects 
exerted  both  upon  agglutination  and  hemolysis  when  bovine  serum 
acts  upon  sensitized  corpuscles  in  the  presence  of  alexin. 

Bordet  and  Gay's  interpretation  of  the  Ehrlich-Sachs  phenom- 
enon, in  the  light  of  these  new  experiments  then,  is,  in  their  own 
words,  as  follows:  "When  guinea  pig  corpuscles  are  added  to  a 
mixture  of  the  two  sera  they  are  affected  by  the  sensitizer  of  the 
horse  serum  and,  to  a  certain  extent,  by  the  sensitizer  in  the  heated 

69  Bordet  and  Gay.    Ann.  de  I'Inst.  Past.,  Vol.  20,  1906,  p.  467. 


BACTERICIDAL    PROPERTIES    OF    BLOOD    SERUM      167 

bovine  serum.  This  second  sensitizer  is,  however,  superfluous.  Its 
presence  is  by  no  means  necessary  for  the  experiment.  When  this 
sensitization  is  effected  the  corpuscles  are  then  in  condition  to  fix 
the  horse  alexin.  This  alexin,  however,  has  only  slight  hemolytic? 
power.  But  once  the  corpuscles  have  become  sensitized  and  laden 
with  alexin  they  are  modified  in  their  properties  of  molecular  ad- 
hesion to  such  an  extent  that  they  become  able  to  attract  a  colloidal 
substance  of  bovine  serum,  which  unites  with  them.  The  adhesion 
of  this  new  substance  produces  two  results :  it  causes  the  blood 
corpuscles  to  be  more  easily  destroyed  by  alexin  and  also  agglutinates 
them  energetically.  Consequently,  a  powerful  clumping,  followed 
by  hemolysis,  is  observed." 

Bordet  and  Gray,  therefore,  assume  that  the  action  of  the  bovine 
serum  is  due  to  a  new  substance  which  they  speak  of  as  "bovine 
colloid."  This  substance  resists  heating  to  56°  C.,  is  probably  al- 
buminous, and  has  the  property  of  uniting  with  cells  that  are  laden 
with  sensitizer  and  alexin,  but  remains  free  in  the  presence  of  nor- 
mal or  merely  sensitized  cells. 

They  fortify  this  opinion  by  showing  experimentally  that  the 
"colloid"  is  removed  from  bovine  serum  by  absorption  with  sensi- 
tized bovine  corpuscles  which  have  been  treated  with  horse  alexin.60 

Bordet  and  Streng61  later  studied  this  "colloid"  more  thor- 
oughly and  have  suggested  for  it  the  name  " congluiimn."  Streng  62 
later  showed  that  the  agglutinating  action  of  this  substance  could  be 
shown  not  only  for  sensitized  and  "alexinized"  red  blood  cells,  but 
also  for  similarly  treated  bacteria,  and  that  conglutinins  were  pres- 
ent not  only  in  bovine  serum,  but  in  that  of  goats,  sheep,  antelopes, 
and  a  number  of  other  herbivores,  but  apparently  absent  in  cats, 
dogs,  guinea  pigs,  and  birds. 

The  body  described  by  these  workers  as  conglutinin  is  probably 
identical  with  a  similar  heat-stable  serum  component  reported  by 
Manwaring63  and  called  by  him  "auxilysin." 

60  Browning  (Wien  kl.  Wochenschr.,  1906)  had  shown  that  horse  alexin 
may  be  absorbed  by  sensitized  beef  cells  without  causing  hemolysis. 

61  Bordet  and  Streng.     Centralbl  f.  Bakt.,  Orig.  Vol.  49,  1909. 

62  Streng.     Zeitschr.  f.  Immunitatsforsch.,  Orig.  Vol.  2,  1909,  p.  415. 

63  Manwaring.    Centralbl.  f.  Bakt.,  1906;  Orig.  Vol.  42. 


CHAPTER   VII 

FUKTHEK  DEVELOPMENT  OF  OUK  KNOWLEDGE  CON- 
CEKNING  COMPLEMENT  OE  ALEXIN.    COMPLE- 
MENT FIXATION 

IT  will  be  remembered  that  Buchner  in  his  first  studies  upon  the 
"alexin"  compared  its  action  to  that  of  an  enzyme  or  ferment,  and 
suggested  that  the  source  of  this  substance  might  possibly  be  found 
in  the  white  blood  cells.  This  thought  was  very  obviously  suggested 
by  the  observation  that  bacteria  were  destroyed  within  the  white 
blood  cells,  after  phagocytosis,  by  a  process  analogous  in  many  ways 
to  that  by  which  they  were  destroyed  by  the  serum  constituents. 
Hankin,1  in  an  elaborate  study  dealing  with  the  problem,  maintained 
the  leukocytic  origin  of  alexin  on  the  basis  of  the  observation  that 
increased  bactericidal  properties  closely  followed  upon  the  heels  of 
periods  of  leukocytosis.  He  assigned  the  particular  property  of 
alexin  production  to  the  eosinophile  cells,  proposing  for  them  the 
designation  "alexocytes."  Further  study,  however,  has  not  justified 
such  an  association  with  the  eosinophiles,  and  Hankin' s  opinion  has 
not  been  experimentally  upheld. 

After  Hankin  the  problem  occupied  the  attention  of  a  number 
of  other  investigators,  and  many  of  them  succeeded  in  showing  that 
there  was,  indeed,  an  increased  bactericidal  power  in  exudates  rich 
in  leukocytes,  and  further  that  bactericidal  substances  could  be 
directly  extracted  from  leukocytic  emulsions.  We  refer  particularly 
to  the  early  work  of  Denys  and  Havet,2  of  Hahn,3  of  Van  de  Velde,4 
and  others,  studies  which  will  be  described  in  our  chapter  on 
phagocytosis.  This  work  was  done  before  the  complex  nature  of  the 
bactericidal  constituents  of  serum  had  been  demonstrated  and  before 
the  work  of  Schattenfroh  and  others  had  shown  that  the  bactericidal 
substances  extracted  from  leukocytes  were  of  a  nature  quite  distinct 
from  the  active  elements  of  the  serum,  and  were  independent  of  the 
participation  of  alexin.  Although  these  earlier  investigations  cannot 
properly  be  regarded,  therefore,  as  proving  the  leukocytic  origin  of 

1  Hankin.    Centralbl.  f.  Bakt.,  Vol.  12,  1892. 

2  Denys  and  Havet.    La  Cellule,  Vol.  10,  1894. 

3  Hahn.    Archiv  f.  Hyg.,  Vol.  25,  1895. 

4  Van  de  Velde.    La  Cellule,  Vol.  10,  1894. 

168 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     169 

alexin,  Metchnikoff  and  his  school  have  nevertheless  adhered  to  this 
conception  for  various  additional  reasons. 

Metchnikoff  distinguishes  between  two  kinds  of  alexin — the 
microcytase,  which  is  the  bactericidal  complement  or  alexin,  and 
is  supposed  to  originate  from  the  microphages  or  polynuclear  leu- 
kocytes, and  the  macrocytase,  which  represents  the  hemolytic  and 
cytolytic  alexin  or  complement,  and  originates  from  the  mononuclear 
cells  or  macrophages.  As  in  the  case  of  the  bactericidal  alexin,  ex- 
traction methods  have  been  employed  to  demonstrate  that  the  hemo- 
lytic alexin  took  its  origin  in  the  macrophages,  and  at  Metchnikoff' s 
suggestion,  Tarassewitch  5  prepared  hemolytic  substances  by  extract- 
ing spleen  tissue  and  other  amacrophagic  organs"  in  various  ways. 
Here  again  the  identity  of  the  hemolytic  extracts  with  serum  hemoly- 
sins  has  been  placed  in  doubt.  Korschun  and  Morgenroth 6  have 
shown  that  the  hemolytic  organ  extracts  were  heat  stable  and  alcohol 
soluble ;  Donath  and  Landsteiner,7  and  others,  have  obtained  similar 
results.  It  would  be  quite  thankless  to  review  the  extensive  literature 
which  has  accumulated  upon  this  point.  It  would  seem,  in  summariz- 
ing it,  that  no  definite  proof  of  the  presence  of  true,  active  alexin, 
either  hemolytic  or  bactericidal,  within  the  leukocyte  or  mononuclear 
cells  has  been  brought  by  methods  of  extraction,  and  the  apparently 
positive  results  reported  by  earlier  observers  are  adequately  explained 
by  the  discovery  of  the  heat-stable  and  non-reactivable  bactericidal 
and  hemolytic  substances  in  extracts  of  such  cells  by  Schattenfroh, 
Korschun  and  Morgenroth,  and  many  others.  It  appears,  moreover, 
from  these  investigations  that  probably  the  intracellular  substances 
by  which  the  digestion  of  ingested  bacteria  or  blood  cells  is  brought 
about  are  of  a  nature  entirely  distinct  from  that  of  the  serum  anti- 
bodies and  alexins.  A  very  ingenious  demonstration  of  this  is  found 
in  an  experiment  first  made  by  RTeufeld.  Neufeld  8  allowed  leu- 
kocytes to  take  up  highly  sensitized  red  cells.  Instead  of  undergoing 
prompt  hemolysis,  as  they  would  if  small  amounts  of  alexin  had  been 
added,  they  were  slowly  broken  up  without  hemolysis,  fragments  of 
hemoglobin  remaining  after  complete  morphological  disintegration 
of  the  erythrocytes.  At  no  time  were  intraphagocytic  "shadow'* 
forms  observed. 

The  failure  to  extract  alexins  from  dead  leukocytes  does  not, 
however,  preclude  the  possibility  of  the  secretion  of  alexins  by  living 
leukocytes.  This  point  is  one  which  is,  of  course,  much  more  diffi- 
cult to  investigate  directly.  Indirectly  the  increased  bactericidal 
properties  of  exudates  rich  in  leukocytes,  as  found  by  Denys  and 
Havet,  would  point  in  this  direction.  However,  even  this  is  not 

5  Tarassewitch.    Cited  from  Metchnikoff. 

6  Korschun  and  Morgenroth.     Berl  kl  Woch.,  No.  37,  1902. 

7  Donath  and  Landsteiner.     Wien.  kl.  Rundschau,  Vol.  40,  1902. 

8  Neufeld.    Arb.  a.  d.  kais.  Gesundlieitsamt.,  Vol.  28,  1908,  p.  125. 


170  INFECTION    AND    RESISTANCE 

•conclusive,  since  at  the  time  when  these  investigations  were  carried 
out  no  discrimination  was  made  between  the  bactericidal  serum  sub- 
stances and  those  other  "endolysins"  which  might  well  have  been 
extracted  from  the  accumulated  white  blood  cells.  The  writer  some 
years  ago  attempted  to  approach  this  problem  directly  by  keeping 
leukocytes  alive  in  inactivated  serum  and  in  Ringer's  solution  at 
37.5°  C.  for  several  days  in  the  hope  that,  after  48  hours,  alexin, 
hemolytic  or  bactericidal,  might  appear  in  these  fluids.  The  experi- 
ments were  entirely  negative,  but  were  regarded  as  inconclusive,  since 
it  was  impossible  to  determine  accurately  how  long,  or  in  what  pro- 
portion, the  leukocytes  had  remained  alive. 

One  of  the  basic  premises  of  MetchnikofFs  theory  on  the  nature 
of  alexin  consists  in  the  conception  that  alexin  is  not  found  in  the 
circulating  blood  plasma,  but  appears  only  when  there  has  been  leu- 
kocytic  injury,  as  in  the  clotting  of  blood  or  in  the  "phagolysis" 
which,  as  we  have  seen  in  the  chapter  on  phagocytosis,  usually  occurs 
after  foreign  substances  have  been  injected  into  the  peritoneum,  pre- 
ceding a  local  accumulation  of  leukocytes.  This  point  of  view  seems 
to  be  rendered  improbable  because  of  the  rapid  hemolysis  which 
occurs  when  we  inject  sensitized  red  blood  cells  into  the  circulation 
of  an  animal,  but  we  might  here,  too,  assume  a  preliminary  injury 
to  white  blood  cells  resulting  from  the  intravenous  injection. 

Much  less  likely  to  be  accompanied  by  cell  injury  is  the  method 
of  obtaining  blood  serum  by  creating  an  area  of  artificial  edema  by 
ligating  a  limb — or,  as  in  MetchnikofFs 9  10  experiments,  the  ear  of 
a  rabbit.  And,  indeed,  in  edema  fluids  so  obtained  little  or  no  alexin 
is  ordinarily  found.  This  fact  has  been  interpreted  in  favor  of 
MetchnikoiPs  views,  as  has  also  the  curious  absence  of  alexin  in  the 
aqueous  humor  of  the  anterior  chamber  of  the  eye.11 12  In  this  fluid 
no  alexin  is  present  under  normal  conditions,  but  if  puncture  is  prac- 
ticed, and  the  fluid  again  taken  after  a  period  of  three  or  four  hours, 
alexin  is  now  found,  probably,  according  to  Metchnikoff  s  school,  be^ 
cause  of  the  coincident  entrance  of  leukocytes  into  this  space.  It  is 
conceivable,  however,  that  the  aqueous  humor  may  be  free  from 
alexin  for  other  reasons  than  the  absence  of  leukocytes ;  and  an  injury 
which  is  followed  by  the  invasion  of  leukocytes  is  pretty  sure  to  be 
followed  also  by  the  entrance  of  the  fluid  elements  of  the  blood ;  i.  e., 
alexin. 

Much  experimental  work  has  been  done  in  which  it  has  been 
attempted  to  demonstrate  directly  that  the  blood  plasma  contains  no 
complement  or  alexin.  The  most  important  investigation  of  this 

9  Metchnikoff.    Ann.  Past.,  Vol.  9,  1895. 

10  Bordet.     Ann.  Past.,  Vol.  9,  1895. 

11  Metchnikoff.     Loc.  cit. 

12  Mesnil.    Ann.  Past.,  Vol.  10. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     171 

kind  is  that  carried  out  by  Gengou13  in  1901.  It  was  Gengou's 
primary  purpose  to  obtain  the  plasma  of  mammals  in  such  a  way 
that  no  cell  injury  would  occur.  This  he  accomplished  by  special 
methods  in  which  coagulation  was  avoided  without  the  addition  of 
foreign  anticoagulants  like  hirudin,  etc.  His  technique  was,  in 
essence,  as  follows :  He  took  the  blood  directly  through  a  paraffined 
cannula  into  tubes  that  had  been  coated  with  paraffin,  and  centrifu- 
gal ized  it"  at  low  temperatures  until  cell  free.  This  plasma,  taken 
from  the  paraffin  tubes,  quickly  clotted,  and  the  material  with  which 
the  experiments  were  done  actually  consisted  of  blood  serum.  Upon 
examining  the  serum  so  obtained,  he  found  that  it  exerted  practically 
no  bactericidal  action.  As  a  result  of  this  investigation  he  claims 
to  have  demonstrated  the  truth  of  MetchnikofFs  contention  that  the 
circulating  blood  plasma  contains  no  alexin. 

If  borne  out,  it  is  true  that  Gengou's  results  would  very  power- 
fully support  this  theory,  and  for  this  reason  a  large  number  of  ex- 
periments have  been  made  since  then,  with  the  same  end  in  view. 

In  all  such  investigations  the  technical  procedures  are  extremely 
difficult  and,  as  Addis  14  has  recently  said,  in  our  opinion  quite  cor- 
rectly, it  would  be  impossible  to  carry  out  bacteriolytic  or  hemolytic 
experiments  with  mammalian  paraffin  plasma  without  obtaining 
coagulation,  and  for  this  reason  most  of  the  writers  who  have  re- 
peated Gengou's  experiments  have  worked,  as  did  he,  not  with 
plasma,  but  with  serum.  Falloise,15  following  Gengou's  method 
exactly,  obtained  results  diametrically  opposed  to  those  of  Gengou; 
Schneider/6  also  with  the  same  technique,  failed  to  confirm  Gengou's 
results;  Herman,17  on  the  other  hand,  confirms  Gengou. 

In  order  to  overcome  the  technical  difficulties  encountered  in 
working  with  mammalian  plasma  a  number  of  writers  have  more 
recently  experimented  with  bird  blood,  which,  as  is  well  known, 
coagulates  much  more  easily  than  does  mammalian  blood.  Hewlett,18 
who  worked  with  goose  plasma  and  peptone  plasma,  could  not  con- 
firm Gengou' s  results.  Lambotte,19  examining  the  plasma  of  chick- 
ens, found  no  difference  between  the  serum  and  plasma  in  their  con- 
tents of  bactericidal  alexin,  as  measured  against  cholera  spirilla. 
Yon  Dungern,  working  with  fish  plasma,  obtained  similarly  negative 
results,  and  recently  Addis,  in  a  careful  comparative  study  of 
chicken  plasma,  found  no  evidence  of  differences  between  plasma 
and  serum  in  either  the  bactericidal  or  the  hemolytic  alexin.  As  far 

13  Geng-ou.    Ann.  de  I'Inst.  Past.,  Vol.  15,  1901. 

14  Addis.    Journ.  of  Inf.  Dis.,  Vol.  10,  1912. 

15  Falloise.     Bull,  de  I'Acad.  Roy.  de  Med,,  1905,  p.  230. 

16  Schneider.    Archiv  /.  Hyg.,  1908,  Vol.  65,  p.  305. 

17  Herman.     Bull,  de  I'Acad.  Eoy.  de  Med.,  1904,  p.  157. 

18  Hewlett.    Archiv  f.  exp.  Path.  u.  Pharmk.,  1903,  Vol.  49,  p.  307. 

19  Lambotte.     Centralbl.  f.  Bakt.,  I,  Orig.,  1903,  Vol.  34,  p.  453. 


INFECTION    AND    RESISTANCE 

as  we  can  tell  at  present,  therefore,  we  cannot  accept,  as  conclusively 
proven,  the  contention  that  the  circulating  plasma  contains  no  alexin. 
Nevertheless  the  Metchnikoff  school  have  not  been  discouraged  by  the 
various  contradictions  of  Gengou's  work,  found  in  the  experiments  we 
have  enumerated,  because  they  are  not  satisfied  that  the  technique  of 
other  workers  has  conclusively  excluded  cell  injury.  Owing  to  the 
great  difficulties  of  investigations  of  this  kind,  when  carried  out  with 
mammalian  blood,  it  is  not  impossible  that  they  are  justified  in  this, 
but  nevertheless  the  assumption  of  the  absence  of  alexin  in  the 
plasma  finds  so  many  objections  in  other  observations  that  the  bur- 
den of  proof  would  certainly  rest  with  Gengou  and  his  supporters. 
Not  the  least  important  of  these  objections,  it  seems  to  us,  is  based 
on  the  very  simple  experiment  of  injecting  bacteria  into  the  veins  of 
a  living  animal  and  finding  a  very  rapid  and  active  phagocytosis. 
And  considering  the  very  probable  participation  of  alexin  in  the 
opsonic  functions  this  would  seem  to  point  strongly  toward  the  pres- 
ence of  these  substances  in  the  circulating  blood.  The  evidence  also 
furnished  by  the  recent  developments  of  our  understanding  of  ana- 
phylaxis  would  further  tend  to  strengthen  our  belief  in  the  presence 
of  alexin  or  complement  in  the  normal  circulation.  For,  in  the 
process,  as  we  shall  see  in  a  later  chapter,  complement  plays 
an  important  role.  When  3  per  cent,  salt  solution  is  administered  (as 
in  Friedberger's  experiments),  and  the  action  of  complement  is 
thereby  inhibited,  anaphylactic  shock  may  be  greatly  diminished. 

It  has  also  been  claimed,  chiefly  by  Walker  20  and  by  Henderson 
Smith,21  that,  as  serum  stands  upon  the  clot  it  at  first  gains  in  alexin 
or  complement  contents,  an  occurrence  which  they  attribute  to  the 
liberation  of  alexin  from  the  leukocytes.  This  observation  has  not 
been  universally  borne  out  and,  even  were  it  unquestionable,  it  might 
be  dependent  upon  any  one  of  the  numerous  factors  involved  in  the 
complicated  process  of  coagulation  rather  than  upon  leukocytic 
changes  only. 

The  failure  to  obtain  definite  proof  of  the  origin  of  alexin  from 
the  white  blood  cells  has  led  to  search  for  the  source  of  these  sub- 
stances in  various  organs.  An  interesting  series  of  investigations 
on  this  subject  are  those  of  Mile.  Louise  Fassin,22  who  believes  that 
she  has  found  reasons  for  definitely  associating  the  thyroid  gland 
with  alexin  production.  She  found  that  the  subcutaneous  injection 
of  thyroid  extract  into  dogs  and  rabbits  was  followed  by  a  rapid 
increase  of  alexin,  both  hemolytic  and  bactericidal,  and  that  the 
same  thing  was  true  when  thyroid  substance  was  administered  by 
mouth.  When  the  thyroid  gland  was  removed  from  rabbits  a  reduc- 
tion of  alexin  resulted.  Although  important,  these  researches  do  not 

20  Walker.    Journ.  of  Hyg.,  Vol.  3,  1903. 

21  Smith.     Proc.  Roy.  Soc.,  Series  B,  Vol.  79,  1906. 

22  Louise  Fassin.     C.  E.  de  Soc.  Biol.,  Vol.  62,  1907. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     173 

necessarily  prove  that  the  thyroid  can  be  looked  upon  as  a  source 
of  alexin,  and,  indeed,  Fassin  gives  experimental  results  without 
drawing  any  very  sweeping  conclusions.  It  might  well  be  that  the 
thyroid  secretion  is  simply  concerned  in  stimulating  the  production 
of  alexin  from  another  source.  Marbe  2B  has  similarly  associated 
the  thyroid  gland  with  the  production  of  opsonins,  which,  when  we 
consider  the  probable  identity  of  alexin  and  normal  opsonin,  may  be 
taken  as  a  confirmation  of  Fassin's  work. 

Of  great  interest  also  are  the  series  of  investigations  which  asso- 
ciate the  liver  with  the  production  of  alexin.  The  basis  of  such 
investigations  is  found  in  the  observations  made  by  Morgenroth  and 
Ehrlich  24  that  there  is  a  diminished  production  of  complement  or 
alexin  in  dogs  subjected  to  phosphorus  poisoning,  with  consequent 
degeneration  of  the  liver.  The  first  investigator  to  study  this  ques- 
tion experimentally  was  Nolf.25  Nolf  tried  to  approach  it  by  extir- 
pating the  liver  in  dogs,  and  found  that  his  results  were  unreliable 
by  this  method.  He  then  experimented  with  rabbits  and  found  that 
when  the  liver  was  extirpated  in  these  animals  and  the  vena  cava 
anastomosed  with  the  portal  vein  (Eck  fistula)  the  animals  would 
survive  for  three  or  four  hours.  This  period,  though  short,  was  suffi- 
cient to  show  definite  changes  in  the  blood.  Taken  just  before  death 
it  differed  from  that  taken  just  before  the  operation  in  a  number  of 
important  respects.  There  was  relative  incoagulability,  there  was 
autohemolysis,  and  with  these  there  occurred  an  extreme  fall  of 
alexin  or  complement.  Serious  objections  may  be  brought  against 
Golf's  experiments.  In  the  first  place  the  operation  as  performed  by 
him  results  in  shock  and  injury  so  profound  that  rapid  death  ensues, 
conditions  under  which  not  only  the  complement-producing  functions 
but  all  functions,  secretory  and  otherwise,  are  reduced.  Miiller  26  ob- 
jects to  Nolf 's  experiments  chiefly  for  the  reason  that  he  did  not  pre- 
vent the  absorption  of  toxic  substances  from  the  intestine,  materials 
which  could  now  enter  the  general  circulation  without  any  longer  be- 
ing neutralized  by  the  liver  functions.  M tiller,  for  this  reason,  re- 
peated ~Nolf 7s  work  but,  by  a  complicated  technique,  temporarily  shut 
off  the  intestinal  circulation  in  addition  to  extirpation  of  the  liver. 
He  found,  in  agreement  with  Nolf,  that  exclusion  of  the  liver  from 
the  circulation  resulted  in  the  prompt  diminution  of  complement  or 
alexin.  In  all  such  experiments,  however,  the  very  profound  shock 
which  necessarily  occurs  in  the  animals  would  seem  to  us  to  vitiate 
the  results.  Moreover,  Liefmann 27  has  repeated  Miiller's  experi- 
ments without  being  able  to  obtain  the  same  results.  Not  satisfied 

23  Marbe.    C.  E.  de  la  Soc.  Biol,  Vols.  64,  et  seq.,  1908-1909. 

24  Morgenroth  and  Ehrlich.     In  Ehrlich's  "Gesammelte  Arb.,"  etc. 

25  Nolf .    Bull,  de  I'Acad.  de  Science  de  Belg.,  1908. 

26  Miiller.     Centralbl.  f.  Eakt.,  Vol.  57,  1911. 

27  Liefmann.     Weichhart' s  Jahresbericht,  Vol.  8,  1912,  p.  155. 


174  INFECTION    AND    RESISTANCE 

with  these  experiments,  however,  Liefmann  experimented  on  frogs ; 
in  whom,  as  Friedberger  has  shown,  extirpation  of  the  liver  is  not 
so  rapidly  fatal  as  in  warm-blooded  animals.  He  removed  the  livers 
of  frogs  in  a  number  of  cases  and,  although  his  animals  lived  about 
a  week,  there  was  no  definite  diminution  of  the  hemolytic  properties 
of  the  serum.  It  seems,  therefore,  that  the  origin  of  alexin  in  the 
body  is  by  no  means  settled  and  requires  further  investigation. 

Equally  unsatisfactory  have  been  the  attempts  to  define  the  chem- 
ical nature  of  the  complement  or  alexin.  In  the  investigations  deal- 
ing with  the  hemolytic  action  of  cobra  venom  it  seemed  at  first  as 
though  a  clue  to  this  problem  had  been  found.  Flexner  and  JSTogu- 
chi  28  made  the  interesting  observation  that  cobra  poison  alone  does 
not  hemolyze  the  blood  cells  of  certain  animals,  namely  those  of 
cattle,  goats,  or  sheep,  if  these  cells  are  washed  entirely  free  of 
serum.  This  seemed  to  suggest  that  the  serum  of  these  animals  con- 
tained some  activating  substance.  It  also  seemed  to  indicate  that  the 
cells  of  other  animals,  which  were  easily  hemolyzed,  even  when  en- 
tirely freed  of  serum,  might  contain  such  an  activating  substance 
within  themselves.  The  behavior  of  this  activating  substance  toward 
snake- venom  hemolysis  was  therefore  very  similar  to  the  action  of 
complement,  except  in  one  important  respect,  namely,  as  Calmette  29 
showed,  almost  all  sera  were  rendered  more  efficient  for  the  activa- 
tion of  snake  venom  when  heated  to  65°  C.,  whereas  complementary 
properties  of  sera  for  other  hemolyzing  complexes  are,  of  course, 
destroyed  at  56°  C.  Kyes,30  31  on  further  studying  these  phenomena, 
extracted  the  red  blood  cells  of  rabbits  and  other  animals  whose  cells 
were  hemolyzed  by  snake  venom  alone,  by  shaking  them  up  with 
distilled  water,  and  showed  that,  with  these  extracts,  he  could  activate 
the  venom  against  ox,  goat,  and  sheep  corpuscles,  cells  which  were 
not  ordinarily  hemolyzed  by  the  venom  without  the  addition  of 
serum.  Similar  activation  of  the  venom  with  extracts  of  the  ox, 
goat,  or  sheep  corpuscles  was  not  possible.  He  concluded  from  this 
that  the  blood  cells  of  the  rabbit,  dog,  guinea  pig,  and  man  pos- 
sessed an  "endocomplement"  for  the  snake  venom;  that  is,  a  com- 
plementary substance  contained  within  the  cells,  while  in  the  other 
species  it  was  found  in  the  activating  serum  only. 

The  thermostability  of  such  venom  "complements"  encouraged 
him  to  attempt  their  isolation,  and  he  found  that  they  were  ether- 
soluble,  indicating  their  lipoidal  nature;  and,  finally,  after  several 
negative  attempts  with  activation  by  other  lipoids,  he  determined 
that  lecithin,  added  to  the  corpuscles  and  the  snake  venom,  brought 

28  Flexner  and  Noguchi.    Journ.  Exp.  Med.,  Vol.  6,  1902 ;  Univ.  Pa.  Med. 
Bull.,  1902  and  1903. 

29  Calmette.     C.  R.  de  I'Acad.  des  Sciences,  p.  134,  1902. 

30  Kyes.    Berl  klin.  Woch.,  Nos.  38  and  39,  1902. 

31  Kyes  and  Sachs.    Berl.  klin.  Woch.,  Nos.  2-4,  1903. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     175 

about  a  rapid  hemolysis.  This  seemed  to  explain  both  why  heated 
serum  could  activate  the  venom  in  some  cases,  and  why  some  varie- 
ties of  blood  cells  could  be  hemolyzed  without  serum,  since  lecithin 
is  a  substance  widely  distributed  both  in  the  fluids  and  cells  of  the 
animal  body.  His  further  studies  seemed  to  show  that,  by  proper 
chemical  manipulation  (bringing  together  cobra  poison  with  lecithin 
in  chloroform  solution),  he  could  produce  a  combination  of  the  two 
which  he  called  "cobra  lecithid,"  a  substance  which  apparently  "acti- 
vated" cobra  venom.  He  conceived  it  as  the  "amboceptor-comple- 
ment"  complex  of  the  cobra  hemolysin,  which  acted  hemolytically 
upon  all  varieties  of  blood  cells. 

These  researches  of  Kyes  aroused  much  interest,  chiefly  because 
they  seemed  to  furnish  an  example  of  a  chemically  definable  com- 
plement, lipoidal  in  its  constitution.  Recent  researches  by  Von 
Dungern  and  Coca,32  however,  seem  to  prove  that,  while  Kyes'  ex- 
perimental facts  were  perfectly  accurate,  his  conclusions  do  not  seem 
to  have  been  warranted.  Von  Dungern  and  Coca  showed  that  the 
cobra  venom  contains  a  lipoid-splitting  ferment  which  acts  upon  the 
lecithin,  liberating  substances  from  it  which  hemolyze  in  the  same 
way  as  do  many  other  non-specific  substances.  The  cobra-lecithid, 
according  to  this,  would  represent  merely  a  lecithin  derivative  which 
happens  to  have  hemolytic  action  without  any  specific  -relationship  to 
the  hemolytic  properties  of  the  venom  itself.  Thus,  even  in  this  case, 
unfortunately,  we  are  not  in  possession  of  facts  which  bring  'us 
nearer  to  a  chemical  understanding  of  the  complementary  substances 
or  alexins. 

In  the  further  development  of  attempts  to  define  alexin  or  com- 
plement chemically,  two  further  researches  are  of  importance, 
namely,  those  of  Von  Liebermann  33  and  of  Noguchi.  In  both  in- 
vestigations it  is  suggested  that  the  alexin  may  consist  of  a  combina- 
tion of  soaps  and  proteins.  Noguchi  34  showed  that  the  hemolytic 
organ-extracts  described  by  various  observers  were  soaps,  a  possi- 
bility which  had  been  previously  considered  by  Sachs  and  Kyes.35 
Noguchi  further  established  analogies  between  his  soaps  and  com- 
plement as  follows:  Sensitized  blood  cells  are  hemolyzed  by  mix- 
tures of  soaps  and  inactivated  guinea  pig  serum,  while  normal  ery- 
throcytes  are  not  hemolyzed  by  similar  mixtures;  furthermore,  like 
normal  complement,  such  serum-soap  mixtures  are  inactivated  by 
prolonged  preservation  and  by  heating  at  56°  C.  Objections  were 
soon  made  to  the  findings  of  both  Noguchi  and  Von  Liebermann  by 
Hecker,36  whose  experiments  seemed  to  show  that  when  sensitize^ 

32  Von  Dungern  and  Coca.    Munch,  med.  Woch.,  1907,  p.  2317. 

33  Von  Liebermann.     Biochem.  Zeitschr.,  Vol.  4,  1907. 
34Noguchi.     Biochem.  Zeitschr.,  Vol.  6,  1907. 

35  Sachs  and  Kyes.    Berl.  kl  Woch.,  2-4,  1903. 

36  Hecker.    Arb.  auf  dem  konig.  Inst.  f.  exp.  Ther.,  Heft  3,  1907. 


176  INFECTION    AND    RESISTANCE 

blood  cells  were  thoroughly  washed  free  of  serum  soaps  did  not  have 
this  hemolyzing  action,  and  Friedemann  and  Sachs37  claimed  that 
they  were  unable  in  any  case  to  inactivate  the  hemolytic  serum  soap 
mixtures  by  heating  to  56°  C.  These  writers,  as  well  as  others, 
attribute  Noguchi's  results  to  the  fact  that  the  sera  which  he  used  to 
produce  his  "artificial  complement,"  i.  e.,  his  serum  soap  mixtures, 
were  heated  to  50-51°  C.  only,  a  fact  which  would  justify  doubt  of 
complete  inactivation.  Knaffl-Lenz38  has  more  recently  carried  out 
experiments  on  the  same  question.  His  results  seem  to  show  that  the 
hemolytic  action  exerted  by  fatty  acids  or  soaps  is  a  phenomenon 
quite  incomparable  to  true  complement  action,  and  that  these  hemoly- 
sins  are  heat  stable,  remaining  unchanged  by  heating  at  56°  C.  We 
have  referred  in  a  number  of  places  to  the  analogy  between  alexins 
and  ferments  or  enzymes.  The  chief  objection  to  this  conception 
formerly  brought  forward  was  based  upon  the  fact  that  the  comple- 
ment or  alexin,  unlike  an  enzyme,  was  used  up  during  its  reactions, 
and  that  a  definite  quantitative  relationship  existed  between  the 
alexin  and  the  amount  of  cells  or  bacteria  upon  which  it  could  act. 
Recent  experiments  by  Kiss  39  seem  to  show  that  this  quantitative  re- 
lationship is  not  as  strict  and  regular  as  was  formerly  supposed.  He 
showed  that  the  action  of  complement  depends  very  largely  upon  its 
concentration.  For  instance,  to  cite  his  work  directly:  "0.05  com- 
plement is  sufficient  to  hemolyze  completely  a  definite  quantity  of 
sensitized  blood  cells  if  the  experiment  is  done  in  a  total  volume  of 
5  c.  c.  0.02  c.  c.  of  complement  gives  absolutely  no  hemolysis  in  a 
similar  volume.  When,  however,  the  total  volume  is  reduced  to  2.5 
c.  c.,  then  0.02  c.  c.  of  the  complement  begins  to  act,  and  it  produces 
complete  hemolysis  if  the  total  volume  is  reduced  to  1.25."  In  fur- 
ther developing  this  observation  he  showed  that,  if  sufficiently  con- 
centrated, a  very  small  amount  of  complement  can  act  upon  an  ex- 
tremely large  amount  of  red  blood  cells,  an  amount  incomparably 
larger  than  those  acted  upon  in  more  dilute  solutions.  These  observa- 
tions would  tend  to  strengthen  considerably  the  conception  of  the  fer- 
ment nature  of  alexins  in  general. 

Kiss'  observations  are  furthermore  in  agreement  with  the  inves- 
tigations of  Liefmann  and  Cohn,40  whose  work  we  have  mentioned 
in  the  preceding  chapter  on  Cytolysis.  These  writers  assert  that  the 
fixation  of  complement  during  hemolysis  is  not  due  to  its  chemical 
union  with  the  sensitized  cells,  but  is  due  to  fixation  by  the  end 
products  of  the  reaction ;  in  other  words,  by  the  stromata  of  the  red 
cells  and  possibly  by  other  substances  given  up  by  these  cells.  A 
further  factor  contributing  to  the  disappearance  of  complement  in 

37  Friedemann  and  Sachs.    Biochem.  Zeitschr.,  Vol.  12,  1908. 

38  Knaffl-Lenz.    Biochem.  Zeitschr.,  Vol.  20,  1909. 

39  Kiss.     Zeitschr.  f.  Imm.,  Vol.  3,  1909. 

40  Liefmann  and  Cohn.    Zeitschr.  f.  Imm.,  Vol.  8, 1911. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     177 

such  reactions  is,  they  claim,  its  rapid  deterioration  at  37°  to  40°  C., 
when  diluted.  If  they  are  right,  these  considerations  also  remove 
important  objections  to  the  conception  of  complement  as  a  ferment. 

It  is  clear,  therefore,  that  although  we  have  gained  much  detailed 
information  regarding  the  'functional  activity  of  the  complement  or 
alexin,  and  may  assume,  in  a  general  way,  that  its  action  is  similar 
to,  if  not  identical  with,  that  of  an  enzyme,  we  are  nevertheless  still 
very  much  in  the  dark  concerning  its  chemical  nature.  The  same 
thing  may  be  said  in  regard  to  its  physical  characteristics.  One 
method  of  investigating  the  physical  properties  of  complement  has 
been  that  of  filtration.  It  may  be  remembered  that  one  of  Ehrlich 
and  Morgenroth's  41  arguments  in  favor  of  the  multiplicity  of  com- 
plement was  the  fact  that,  when  goat  serum  was  filtered  through  a 
Pukal  candle,  the  complement  which  was  active  upon  rabbit  corpus- 
cles was  retained,  while  that  which  acted  upon  guinea-pig  cells 
passed  through.  Immune  bodies  or  amboceptor  always  passed 
through. 

Vedder,42  in  similar  experiments  upon  bactericidal  complements, 
claims  to  have  been  able,  in  the  same  way,  to  separate  the  comple- 
ments acting  upon  different  bacteria.  The  problem  has  been  more 
recently  investigated  by  Muir  and  Browning.43  Their  conclusions 
are  briefly  as  follows:  In  the  early  stages  of  filtration  through  a 
Berkefeldt  filter  complement  is  often  completely  held  back.  After 
continued  filtration  it  begins  to  pass  through.  If  the  complement  is 
inactivated  by  the  addition  of  hypertonic  salt  solution  (5  per  cent.), 
it  passes  through,  and  the  filtrate  can  be  reactivated  by  dilution  to 
isotonicity.  Sensitizer  or  amboceptor  always  passes  through.  Just 
how  these  experiments  are  to  be-  interpreted  is  a  little  obscure.  The 
fact  that  the  addition  of  salt  renders  the  complement  capable  of 
passing  through  the  filter  would  seem  to  indicate  that  its  original 
inability  to  permeate  did  not  depend  upon  the  size  of  the  molecule. 
On  the  other  hand, "it  is  also  possible  that  the  addition  of  salt  to  the 
complement  may  increase  its  dispersion  in  such  a  way  that  the  indi- 
vidual particles  are  rendered  smaller.  This,  however,  is  purely 
speculative,  and  we  are  at  a  loss  for  a  fully  satisfactory  explanation  of 
the  results  of  Muir  and  Browning.  We  have  repeated  some  of  the 
experiments  of  Muir  and  Browning  and,  in  substance,  'confirmed 
their  results.  It  is  our  opinion  that  new  filters  remove  complement 
by  adsorption,  just  as  this  is  accomplished  when  complement  is 
shaken  up  with  kaolin  or  other  finely  suspended  material. 

That  the  addition  of  salts  of  various  kinds  in  quantities  greater 
than  isotonicity  (or  more  than  the  equivalent  of  0.85=0.9  per  cent. 

41  Morgenroth  and  Ehrlich.     Ehrlich's  "Gesammelte  Arbeiten,"  etc. 

42  Vedder.     Journ.  Med.  Res.,  Vol.  9,  1903. 

43  Muir  and  Browning.    Journ.  of  Path,  and  Bact.,  Vol.  13,  1909. 


178  INFECTION    AND    RESISTANCE 

!N"aCl)44  exerts  a  profound  action  upon  the  activity  of  complement 
is  well  known.  ISTolf  45  noted  this  in  1900,  and  the  problem  has  heen 
studied  since  that  time  by  many  investigators.  Von  Lingelsheim,46 
who  studied  it  in  connection  with  his  work  on  the  refutation  of  the 
"osmotic"  theories  of  immunity,  showed  that  increasing  the  salt  con- 
tents of  serum  (KNO3,  NaCl,  K2HPO3,  etc.)  progressively  dimin- 
ished its  bactericidal  power.  Hektoen  and  Ruediger  47  also,  after  a 
very  thorough  study  of  this  phenomenon,  conclude  that  the  action  of 
the  salts  in  such  cases  is  exerted  upon  the  alexin  or  complement  and 
not  upon  the  heat-stable  sensitizers,  and  that  it  probably  depends 
upon  "physicochemical"  causes.  However,  the  manner  in  which 
such  salt-inactivation  is  brought  about  is,  to  a  great  extent,  obscure. 
There  is  no  visible  precipitation  from  serum  after  the  addition  of 
salts  sufficient  in  quantity  to  weaken  its  action.  Nothing  is,  as  far 
as  we  can  tell,  removed  from  solution,  and  yet  there  is  temporary 
inactivation  which,  at  the  same  time,  renders  the  complement  fil- 
trable,  facts  from  which  we  can  only  surmise  some  physical  altera- 
tion. 

Inactivation  of  the  complement  also  follows  the  removal  of  salts, 
but  here  the  process  is  accompanied  by  a  definite  chemical  change  in 
that  the  serum  globulins  are  precipitated. 

Studies  of  this  process  have  led  to  important  modifications  in  our 
conception  of  the  nature  of  alexin,  since  they  have  shown  that  this 
body,  formerly  assumed  to  be  single  and  homogeneous,  may  be  sub- 
divided into  at  least  two  component  parts  by  a  number  of  experi- 
mental procedures.  Ferrata  was  the  first  one  to  point  this  out  as  a 
consequence  of  investigations  undertaken  by  him  primarily  with  the 
purpose  of  determining  the  nature  of  the  influence  of  salts  upon 
hemolytic  processes.  Older  studies  of  Buchner  and  Orthenberger  48 
had  shown  that  bactericidal  action  was  inhibited  when  salts  were 
removed  from  the  medium,  but  the  causes  underlying  such  inhibi- 
tion had  not  been  made  clear.  Ferrata  49  found,  in  the  first  place, 
that  the  absence  of  salts  exerted  no  effect  upon  the  mechanism  of 
sensitization,  but  that  amboceptor  or  sensitizer  became  attached  to 
the  cellular  elements  as  readily  when  salts  were  absent  as  when  the 
reagents  were  suspended  in  normal  salt  solution.  It  was  a  natural 
inference,  therefore,  that  the  failure  of  hemolysis,  which  he  observed 
in 'salt-free  media  (analogous  to  the  similar  experiences  of  Buchner 

44  Alexin  can  be  preserved  in  the  refrigerator  for  long  periods  if  hyper- 
tonic  salt  solution  (15  to  25%)  is  added.    It  will  again  become  active  if  iso- 
tonicity  is  restored  with  distilled  water. 

45  Nolf .    Ann.  Past.,  Vol.  14,  1900. 

46  V.  Lingelsheim.    Zeitschr.  f.  Hyg.,  Vol.  37,  1901. 

47  Hektoen  and  Ruediger.    Journ.  of  Inf.  Dis.,  Vol.  1,  1904. 

48  Buchner  and  Orthenberger.    Archiv  f.  Hyg.,  Vol.  10,  1C90. 

49  Ferrata.     Berl.  kl  Woch.,  1907,  No.  13. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     179 

in  the  case  of  bacteriolysis),  must  be  attributed  to  failure  of  func- 
tionation  on  the  part  of  the  complement.  On  further  investigation 
he  obtained  a  very  simple  explanation.  Ferrata  removed  the  salts 
from  his  sera  by  dialyzing  for  twenty-four  hours  against  distilled 
water.  In  this  process,  of  course,  there  is  a  precipitation  of  the 
globulins  while  the  water-soluble  albumins  remain  in  solution.  The 
former  may  be  redissolved  in  normal  salt  solution  and  the  latter 
rendered  is'otonic  by  the  addition  of  calculated  amounts  of  concen- 
trated salt.  In  this  way  the  original  serum  components  are  divided 
into  two  parts,  neither  of  which,  as  Ferrata  found,  is  alone  capable 
of  producing  hemolysis  of  sensitized  cells.  In  order  to  obtain  the 
complementary  action  possessed  by  the  original 
serum  it  is  necessary  to  combine  the  two.  This 
principle  discovered  by  Ferrata  is  probably  re- 
sponsible also  for  the  results  obtained  by  Sachs 
and  Teruuchi,50  who  likewise  noted  the  destruc- 
tion of  the  complementary  function  in  sera 
diluted  with  distilled  water,  but  attributed  this, 
in  their  publication,  to  the  action  of  a  comple- 
ment-destroying ferment,  which  is  assumed  to 
be  active  in  salt-free  media  only. 

In  his  first  experiments  Ferrata   reported 
that    the    precipitated    globulin    fraction    was 
thermostable,  the  thermolability  of  complement      CONCEPTION  OF  COM- 
being  due  entirely  to  the  unprecipitated  albu-  TING^AS 


t^Mjiii/ji-j 

•  END' PIKE 

•  (ALBUMEN  FIMCT/OK) 

SB    H, DP, tec 

^•^B   (GLOBULIN  FftACTIOH) 
^Z 


min  fraction.     The  work  of  Ferrata  was  soon  SUGGESTED  BY 

continued,  however,  by  a  number  of  other  work-  BRAND. 
ers,  who  confirmed  the  essential  fact  of  the  par- 
tition of  the  complement  but  modified  and  considerably  extended  the 
original  observations.  Brand  51  found  that  both  fractions  were  equally 
thermolabile,  and  that  the  globulin  sediment,  after  being  redissolved 
in  salt  solution,  could  not  be  preserved  in  an  active  condition  for  more 
than  a  few  hours.  Preserved  in  distilled  water  or  as  sediment,  it 
may  retain  its  activity  for  several  days,  but  dissolved  in  salt  solution 
it  becomes  inactive  within  3  to  4  hours,  at  room  temperature. 
Michaelis  and  Skwirsky  52  have  since  shown  that  the  globulin  frac- 
tion, thermolabile  when  free,  is  unaffected  by  a  temperature  of  56° 
C.  after  it  has  become  attached  to  sensitized  cells.  Brand  further 
studied  the  relationship  of  the  two  fractions  to  the  sensitized  cells 
and  found  that  the  globulin  fraction  may  attach  directly  to  such 
antigen-antibody  complexes,  but  that  the  albumin  fraction  cannot  be 
bound  in  this  way  unless  the  globulin  fraction  has  been  previously 
attached.  For  this  reason  he  has  referred  to  the  former  as  the  "end- 

50  Sachs  and  Teruuchi.    Berl  kl  Woch.,  1907,  Nos.  16,  17,  and  19. 

51  Brand.    Berl.  kl.  Woch.,  1907,  No.  34. 

52  Michaelis  and  Skwirsky.     Zeitschr.  f.  Imm.,  Vol.  4,  1910. 


180  INFECTION    AND    RESISTANCE 

piece"  and  the  latter  globulin  sediment  as  the  "mid-piece,"  assum- 
ing, on  the  basis  of  the  conception  of  Ehrlich,  that  the  globulin  frac- 
tion serves  to  establish  a  link  between  the  sensitized  cell  and  the 
end-piece  analogous  to  that  formed  by  the  "amboceptor"  between  the 
cell  and  the  whole  complement.  It  is  possible,  therefore,  to  treat 
sensitized  cells  with  mid-piece  in  such  a  way  that  they  are  thereafter 
susceptible  to  hemolysis  by  the  end-piece  alone.  Such  cell-sensitizer- 
mid-piece  combinations  have  been  spoken  of  by  Michaelis  as  "per- 
sensitized"  cells. 

Tsurusaki53  confirmed  the  findings  of  Brand  as  to  the  thermo- 
lability  of  both  "mid-piece"  and  "end-piece,"  but  was  unable  to  sep- 
arate the  complement  of  normal  hemolysins  into  the  two  components 
in  the  same  way,  since  he  found  that  hemolytic  power  was,  in  such 
cases,  completely  destroyed  after  twenty-four  hours  of  dialysis.  It 
seems  to  us  not  impossible  that  the  natural  deterioration  of  alexic 
power  which  takes  place  during  such  periods  of  time,  at  temperatures 
of  from  16°  to  20°  C.,  may  easily  be  held  accountable  for  this,  since 
the  very  feeble  sensitization  of  cells  in  normal  hemolysin  complexes 
requires  a  correspondingly  larger  amount  of  alexin  for  activation. 

We  have  mentioned  that  the  so-called  "mid-piece"  undergoes  a 
rapid  change  when  dissolved  in  salt  solution  and,  after  3  or  4  hours, 
may  lose  its  ability  to  induce  hemolysis  when  added  to  sensitized 
cells  together  with  end-piece.  Although  Hecker54  was  able  to  con- 
firm this,  he  nevertheless  showed  that  this  fact  does  not  imply  a 
destruction  of  the  mid-piece.  For  when  such  apparently  inactive 
"mid-piece"  was  added  separately  to  sensitized  cells,  and  end-piece 
wTas  subsequently  allowed  to  act  upon  the  complex,  hemolysis  re- 
sulted. This  seems  to  show  that  the  "mid-piece"  undergoes  a  change 
on  standing  in  salt  solution  which  does  not  alter  its  ability  to  com- 
bine with  the  sensitized  cells,  but  which  subjects  it  to  inhibition  of 
such  union  when  end-piece  is  present.  It  is  also  a  peculiar  fact,  evi- 
dent in  many  of  our  own  experiments,  that  when  "mid-piece"  and 
"end-piece"  are  first  mixed  and  then  added  to  sensitized  cells  the  ef- 
fect in  hemolysis  is  less  powerful  than  when  the  "mid-piece"  is  added 
to  the  cells  first,  and  the  "end-piece"  later.  This  effect  is  so  instan- 
taneous that,  if,  in  a  series  of  experiments  in  which  combinations  of 
mid-  and  end-piece  are  used,  the  end-piece  is  run  into  the  tubes  con- 
taining the  cells  just  before  the  mid-piece  is  added  instead  of  the 
other  way  round,  hemolysis  is  inhibited. 

In  working  with  dialysis,  also,  we  have  regularly  had  an  experi- 
ence which  may  explain  the  difficulties  which  many  other  investiga- 
tors have  had  in  such  experiments.  The  globulin  precipitate,  which 

53  Tsurusaki.     Biockem.  Zeitschr.,  Vol.  10,  1908. 

54  Hecker.     Arb.  a.  d.  konig.  Inst.  f.  exp.  Ther.,  Frankfurt  a/M.,  Heft 
3,  1907.     See  also  Guggenheimer.     Zeitschr.  f.  Imm.,  Vol.  8,  1911. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     181 

fell  out  after  dialysis  of  24  hours  or  more,  almost  without  exception, 
retained  moderate  or  slight  hemolytic  properties,  which  could  not  be 
removed  until  the  precipitate  had  been  dissolved  in  salt  solution  and 
reprecipitated  with  distilled  water  two  or  three  times.  This  would 
imply  that  a  minute  amount  of  the  end-piece,  carried  down  in  pre- 
cipitation, must  suffice  to  activate  the  mid-piece  and  would  seem 
to  point  to  the  fact  that  in  whole  serum  the  two  fractions  are  present 
as  a  complex  and  not  separately.  This  question  has  been  much  dis- 
cussed and  many  facts  have  been  brought  out  on  both  sides.  Hecker 
showed  that  the  combination  of  mid-piece  with  the  sensitized  cells 
can  take  place  at  a  temperature  of  0°  C.,  while  that  of  end-piece 
with  the  "persensitized"  cells  requires  a  considerably  higher  tem- 
perature. The  bearing  this  fact  may  have  upon  similar  earlier  ex- 
periments of  Ehrlich  and  Morgenroth  upon  the  thermal  conditions 
governing  the  union  of  amboceptor  and  complement  with  antigen  is 
self-evident.  In  the  present  connection,  however,  the  fact  that  the 
two  fractions  may  be  separately  absorbed  out  of  the  serum  by  sensi- 
tized cells  at  0°  C.  would  suggest  the  probability  of  their  being  sep- 
arate in  the  whole  blood.  ~No  crucial  experiment  has  so  far  been 
possible,  and  there  is  not  enough  evidence  on  either  side  as  yet  to 
justify  a  definite  opinion.  However,  the  experiments  of  Michaelis 
and  Skwirsky  and  later  ones  of  Skwirsky  alone  have  much  indi- 
rect bearing  on  this  question,  though  final  interpretation  is  as  yet 
impossible.  Michaelis  and  Skwirsky,55  after  determining  that  an 
acid  reaction  inhibits  the  hemolysis  of  sensitized  blood  cells,  found 
that  under  such  conditions  "mid-piece"  alone  is  bound,  but  that 
"end-piece"  or  the  albumin  fraction  is  left  unbound.  They  recom- 
mend the  use  of  strongly  sensitized  cells  in  an  acid  medium  as  a 
method  of  obtaining  free  "end-piece"  from  serum. 

Skwirsky  56  subsequently  found  that  during  the  ordinary  Wasser- 
mann  reaction  the  complex  of  syphilitic  serum  and  antigen  binds 
the  mid-piece  only.  If  the  Wassermann  reaction  has  been  strongly 
positive;  that  is  if  there  has  been  absolutely  no  hemolysis,  and  we 
remove  the  supernatant  fluid  by  centrifugation,  active  end-piece  can 
be  demonstrated  in  it  by  the  addition  of  persensitized  cells.  Bron- 
fenbrenner  and  Noguchi  have  also  studied  this  phenomenon,  but  da 
not  believe  that  Skwirsky' s  experiments  prove  that  end-piece  is  free 
in  such  "fixation"  supernatant  fluids.  These  supernatant  fluids,  ac- 
cording to  them,  differ  from  all  other  "end-pieces"  in  that  they  are 
active  upon  persensitized  sheep  corpuscles  only,  but  not  upon  other 
cells.  An  explanation  for  this  is  lacking. 

There  is  much  that  is  confusing  in  the  facts  so  far  revealed  about 
the  two  component  parts  of  the  alexin.  The  most  difficult  fact  to 
explain  is  the  peculiar  inactivation  of  the  mid-piece  in  salt  solution, 

55  Michaelis  and  Skwirsky.     Zeitschr.  f.  Imm.,  Vol.  4,  1910. 

56  Skwirsky.     Zeitschr.  f.  Imm.,  Vol.  5,  1910. 


182  INFECTION    AND    RESISTANCE 

which  prevents  its  functionation  in  the  simultaneous  presence  of 
end-piece,  but  does  not  seem  to  interfere  with  its  ability  to  combine 
with  the  sensitized  cells.  As  was  to  be  expected,  explanation  for 
this  has  been  sought  by  the  Ehrlich  school  in  changes  of  affinity. 
Sachs  suggests  that  the  mid-piece,  by  its  preservation  in  salt  solution, 
has  lost  its  avidity  for  the  sensitized  cells  and  has  gained  in  avidity 
for  the  end-piece,  an  alteration  which  therefore  prevents  its  union 
with  the  cells.  The  same  idea  was  suggested  by  Hecker  himself.  It  is 
a  little  difficult  to  reconcile  this  explanation,  however,  with  the  fact 
that  whole  serum  can  be  preserved  and  remain  active  in  its  comple- 
mentary function  for  a  number  of  days,  mid-piece  and  end-piece 
being  present  together,  in  a  medium  which,  as  far  as  salt  contents  are 
concerned,  is  isotonic  with  the  salt  solution  in  which  mid-piece  de- 
teriorates so  rapidly  when  alone. 

That  there  is,  after  all,  much  similarity  between  the  alexins  of 
different  animals  is  evident  from  the  fact  that,  as  Marks  and  others 
have  shown,  the  end-piece  of  one  animal  may  activate  the  mid-piece 
of  another  species.  It  appears  also  from  experiments  like  those  of 
Ritz  and  Sachs  5T  that  an  animal  may  possess  a  mid-piece  for  certain 
sensitized  cell  complexes  without  possessing  a  corresponding  end- 
piece.  Thus  they  found  that  the  serum  of  mice  contained  a  mid- 
piece  but  not  an  end-piece  for  sensitized  guinea  pig  corpuscles. 

Much  that  has  been  found  out  about  the  so-called  globulin  portion, 
moreover,  tends  to  engender  doubt  as  to  the  wisdom  of  applying  to 
these  complement  fractions  the  terms  "mid-piece"  and  "end-piece," 
an  objection  which  is  based  upon  reasons  similar  to  those  which  pre- 
vent iBordet  from  accepting  the  term  amboceptor.  For  so  little  is 
actually  known  concerning  the  mechanism  of  complement  functiona- 
tion, that  it  seems  unwise  to  establish  on  a  firm  basis  a  preconceived 
idea  of  the  mechanism  by  adapting  the  terminology  to  a  theory.  The 
most  confusing  feature  of  the  problem  lies  in  the  surprising  quantita- 
tive relations  which  seem  to  exist  in  the  reactions  of  the  two  frac- 
tions. Thus  Liefmann  and  Cohn58  claim  that  in  the  presence  of 
moderately  sensitized  cells  no  measurable  amount  of  the  so-called 
mid-piece  or  globulin  fraction  is  bound,  that  is,  removed  from  solu- 
tion ;  and  yet,  when  both  fractions  are  added  to  such  cells,  rapid  and 
complete  hemolysis  results.  In  the  presence  of  heavily  sensitized 
•cells  (20  to  50  units)  a  small  quantity  only  is  removed.  Nevertheless 
this  fraction  has  had  a  demonstrable  effect  on  the  cells,  since  it  has 
Tendered  them  amenable  to  the  action  of  the  albumin  fraction.  In  all 
such  experiments,  therefore,  as  Liefmann  justly  points  out,  the  de- 
gree of  sensitization  must  be  taken  into  consideration  before  conclu- 
sions are  formulated.  It  is  curious  also  that  a  slight  excess  of  the 
globulin  fraction  may  prevent  complement  action  completely.  In 

57  Ritz  and  Sachs.     Zeitschr.  f.  Imm.,  Vol.  14,  1912. 

58  Liefmann  and  Cohn.     Zeitschr.  f.  Imm.,  Vol.  7,  1910. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE    183 

experiments  cited  by  Marks  59  it  appears'  that  the  most  ineffective 
complement  is  obtained  when  "mid-piece"  and  "end-piece"  are  added 
to  the  sensitized  cells  in  proportions  of  1  to  1.  If  the  proportion  of 
"mid-piece"  is  increased  two  or  threefold  over  that  of  "end-piece," 
hemolysis  is  inhibited.  This,  however,  is  true  only  when  the  two 
fractions  are  simultaneously  added  to  the  sensitized  cells.  When  the 
sensitized  cells  are  exposed  to  the  excessive  quantity  of  the  "mid- 
piece"  separately,  and  "end-piece"  added  later,  the  effect  is  one  of 
stronger  hemolysis  than  when  smaller  amounts  are  used.  It  is  thus 
seen  that  the  relations  between  the  complement  fractions  in  hemolysis 
are  very  involved.  All  that  we  can  be  sure  of  is  that  there  are  at  least 
two  separable  parts,  that  one  of  these  acts  directly  upon  the  sensitized 
cells,  forming  a  so-called  persensitized  complex  and  rendering  them 
amenable  to  the  subsequent  action  of  the  unprecipitated  albumin 
fraction. 

The  many  difficulties  encountered  in  the  interpretation  of  the 
confusing  phenomena  observed  in  connection  with  this  problem  have, 
very  naturally,  led  to  a  corresponding  multiplicity  of  opinion.  Most 
observers  at  present  incline  to  the  opinion  that  the  globulin  and 
albumin  portions  of  fresh  serum,  separated  by  Ferrata's  or  any  other 
of  several  common  methods,  represent  actually  two  complement  frac- 
tions. This  is  not,  however,  accepted  by  all  workers.  Bronfenbren- 
ner  and  Noguchi  60  believe  that  the  entire  active  complement  is  con- 
tained in  the  albumin  fraction  or  so-called  "end-piece."  They  hold 
that  "complement-splitting"  by  dialysis  or  other  methods  is  an  inacti- 
vation  of  end-piece  by  change  of  reaction.  In  their  experiments  they 
were  able  to  restore  the  functional  activity  of  end-piece  by  the  adjust- 
ment of  reaction,  either  with  acid  or  alkali,  respectively,  or  by  the 
addition  of  amphoteric  substances.  The  mid-piece  activates,  they  be- 
lieve, by  reason  of  its  amphoteric  nature  and  consequently  adjusts 
any  excessive  acidity  or  alkalinity  of  the  medium.  They  were  able 
to  substitute  for  mid-piece  indifferent  amphoteric  substances  such  as 
alanin.  Liefmann  61  has  been  unable  to  confirm  the  experiments  of 
Bronfenbrenner  and  Noguchi,  and  believes  that  their  results  were 
caused  by  incomplete  splitting  of  the  complement.  Incidental  to  a 
study  of  normal  opsonins  the  writer  has  also  repeated  the  experi- 
ments of  ;Bronfenbrenner  without  being  able  to  confirm  them.62 

The  method  of  Ferrata  for  the  separation  of  the  two  parts  of  the 
complement  is  successful  only  if  dialysis  is  very  thorough  and  suffi- 
ciently prolonged  to  lead  to  complete  precipitation  of  the  globulins. 
iNeufeld  and  Haendel  63  have  had  difficulty  in  thus  separating  the 

89  Marks.     Zeitschr.  f.  Imm.,  Vols.  8  and  11,  1911. 

60  Bronfenbrenner  and  Noguchi.     Jour,  of  Exp.  Med.,  Vol.  15,  1912. 

61  Liefmann.     Weichhardt's  Jahresbericht,  Vol.  8,  1912. 

62  Zinsser  and  Gary.     Journ.  of  Exp.  Med.,  Vol.  19,  1914. 

63  Neufeld  and  Haendel.     Arb.  a.  d.  kais.  Gesund.,  1908. 


184 


INFECTION    AND    RESISTANCE 


fractions,  and  the  writer  has  noticed  similar  failures  but  has  always 
been  able  to  obtain  eventual  separation  by  sufficient  prolongation  of 
the  dialysis.  Because  of  the  occasional  difficulties  and  because  of 
the  time-consuming  and  inconvenient  nature  of  the  method  other 
means  of  separation  have  been  devised.  The  one  used  with  success 
by  many  workers  has  been  that  introduced  by  Sachs  and  Altmann,64 
namely,  precipitation  of  the  sera  with  weak  hydrochloric  acid  -g^ 
to  -j^.  Liefmann  has  separated  the  components  by  precipitation  of 
the  globulins  by  the  introduction  of  CO2.  In  carrying  out  this 
method,  Fraenkel  65  has  found  it  advantageous  to  dilute  the  serum 
ten  times  with  distilled  water,  then  allowing  the  CO2  to  flow  in  at 
low  temperatures.  It  is  likely  that  any  of  the  usual  methods  of 
globulin  separation  will  serve  for  complement  partition.  The  salt- 
ing out  methods  are,  however,  extremely  inconvenient  because  of  the 
prolonged  dialysis  subsequently  necessary  to  remove  the  salts. 

The  inactivation  of  complement  or  alexin  by  the  addition  of 
salts  or  by  splitting  is,  very  apparently,  a  temporary  inactivation  in 
which  prompt  restitution  can  be  practiced  by  bringing  back  original 
conditions  either  by  dilution  to  isotonicity  or  by  reconstruction  of 
the  divided  substance,  respectively.  Heating  to  56°  C.,  the  simplest 
and  most  commonly  employed  method  of  inactivations  was,  until  of 
late,  regarded  as  an  irreversible  process,  the  complement  being  irre- 
trievably destroyed  in  the  procedure.  Gramenitski  66  has  recently 
carried  out  experiments  which  seem  to  show  that  this  opinion  is 
erroneous.  His  experiments  were  suggested  by  the  fact,  observed  by 
Bach  and  Chodat,67  that  certain  oxydases  and  diastases  may  spon- 
taneously regain  some  of  their  activity  after  inactivation  by  heat. 
His  work  with  complement  indicated  a  similar  gradual  return  to  an 
active  condition  after  moderate  heating.  The  great  theoretical  im- 
portance of  this  observation  will  justify  our  insertion  of  one  of 
Gramenitski's  protocols. 

Experiment  1.  Complement  10  times  diluted  was  heated  to  56° 
C.  for  7  minutes.  It  was  then  tested  against  sensitized  beef  blood 
at  varying  intervals  as  follows: 


Time  after  heating  at  which  test 
was  made 

Quantity  of  hemoglobin  gone  into  solution  after 

%  10  min. 

%  20  min. 

%  30  min. 

%  40  min. 

Immediately  after  heating  
1]/2  hour  

0 
0 
20 
10 

20 
30 
70 
40 

40 
60 

80 
70 

70 

80 
100 

24  hours 

48  hours  

64  Sachs  and  Altmann.    Cited  from  Sachs  in  "Kolle  u.  Wassermann  Hand- 
buch,"  Vol.  2,  p.  877. 

65  Fraenkel.     Zeitschr.  /.  7mm.,  I,  Vol.  8,  1911. 

66  Gramenitski.    Biochem.  Zeits.,  Vol.  38,  1912. 

67  Bach  and  Chodat.     Cited  from  Gramenitski,  loc.  cit.,  p.  511. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     185 

In  other  experiments  in  which  heating  was  more  prolonged  a 
similar  regeneration  was  observed,  though  not  as  pronounced  as  in 
the  one  cited  above.  The  largest  amount  of  restored  complement 
seemed  to  be  present  after  about  24  hours.  After  this  gradual  de- 
terioration again  ensued.  It  is  quite  impossible  to  offer  an  adequate 
explanation  for  this  at  the  present  time.  Gramenitski  68  acknowl- 
edges this,  but  permits  himself  certain  speculations  which  we  repeat 
in  nearly  his  own  language,  since  there  is  much  in  them  which  seems 
to  us  reasonable.  The  complement,  as  indeed  all  other  active  serum 
constituents,  must  be  looked  upon  as  colloidal  in  nature.  When  heat 
is  applied  to  such  substances  alterations  occur  which  gradually  lead 
to  coagulation.  As  this  occurs  there  is  an  aggregation  of  particles 
and  a  consequent  diminution  of  surface  tension.  This  last  point  has 
been  experimentally  demonstrated  by  Traube,69  who  has  regularly 
found  a  fall  of  surface  tension  as  serum  was  heated  to  56°  C.  And 
of  greatest  interest  in  this  connection  is  the  further  determination  by 
Traube  that  a  gradual  restoration  of  the  surface  tension  takes  place 
as  the  serum  is  allowed  to  stand.  It  is  not  inconceivable,  therefore, 
that  the  inactivation  of  complement  by  heat  may  depend  upon  an 
alteration  of  its  colloidal  state,  i.  e.,  an  aggregation  of  the  particles, 
which,  if  not  carried  too  far,  may  be  reversible  and  followed  by  a 
gradual  dispersion  as  the  serum  is  kept  24  hours.  On  the  same 
grounds  the  gradual  deterioration  of  complement  on  standing  may  be 
compared  to  the  slow  settling  out  of  colloidal  suspensions  which 
eventually  results  in  spontaneous  precipitation,  a  process  which 
occurs  not  only  in  chemically  well-defined  colloids,  but  is  often  ob- 
served in  sera.  Bechold  has  referred  to  this  as  "das  Altern  Kol- 
loidaler  Losungen." 

Of  great  interest,  furthermore,  in  connection  with  the  physical 
properties  of  complement  is  the  discovery  made  by  Jacoby  and 
Schiitze  70  that  complement  can  be  inactivated  by  shaking.  This 
astonishing  observation  has  been  confirmed  by  Zeissler,71  Noguchi 
and  Bronfenbrenner,72  Ritz,73  and  others.  It  appears,  according  to 
these  observers,  that  guinea  pig  serum,  when  subjected  to  active 
shaking,  can  eventually  be  robbed  thereby  of  its  activating  prop- 
erties. The  success  of  such  experiments  depends  somewhat  upon  the 
concentration  of  the  serum,  and  is  best  observed  in  a  dilution  of  1 
part  to  10  parts  of  salt  solution.  Under  such  conditions  complete 
inactivation  may  be  observed  within  20  to  25  minutes.  Between  the 
inactivation  of  complement  by  heat  and  that  which  results  from 

68  Gramenitski.    Loc.  cit.,  p.  504. 

69  Traube.    Zeitschr.  f.  1mm.,  Vol.  9,  1911,  and  Biochem.  Zeitschr.,  1908. 

70  Jacoby  and  Schiitze.    Zeitschr.  /.  1mm.,  Vol.  4,  1910. 

71  Zeissler.    Berl  kl.  Woch.,  No.  52,  1909. 

72  Noguchi  and  Bronfenbrenner.     Journ.  of  Exp.  Med.,  Vol.  13,  1911. 

73  Ritz.    Zeitschr.  f.  1mm.,  Vol.  15,  1912. 


186  INFECTION    AND    RESISTANCE 

shaking,  there  are  certain  similarities  which  seem  to  strengthen  the 
opinion  regarding  the  nature  of  heat  inactivation  which  we  have 
'cited  above.  For  it  has  been  variously  shown  that  prolonged  shaking 
of  protein  solutions,  like  heating,  gradually  leads  to  coagulation.  It 
would  be  important  to  determine  whether  or  not  the  inactivation  by 
shaking,  like  that  produced  by  heat,  is  accompanied  by  a  fall  of  sur- 
face tension. 


ALEXIN  OR  COMPLEMENT  FIXATION 

The  controversy  regarding  the  multiplicity  of  alexin  and  the 
existence  of  a  "complementophile  group"  cannot,  of  course,  be  re- 
garded as  closed,  however  much  we  may  lean  toward  the  acceptation 
of  Bordet's  point  of  view,  since  German  experimenters  of  eminence 
still  adhere  to  the  Ehrlich  interpretations.  Moreover,  it  is,  of  course, 
extremely  difficult  to  disprove  such  an  assumption  as  that  of  the 
"polyceptor"  conception  of  the  complementophile  group.  However, 
we  may  safely  assert  that  the  functional  unity  of  complement  (and, 
after  all,  that  is  all  that  Bordet  has  maintained)  is  being  upheld  by 
the  constantly  increasing  evidence  in  its  favor  which  is  being  fur- 
nished by  the  practical  and  experimental  application  of  the  phe- 
nomenon of  "alexin  fixation"  described,  in  1901,  by  Bordet  and 
Gengou.74  It  will  be  well  to  bear  in  mind  that  this  phenomenon 
should  be  strictly  distinguished  from  the  so-called  "complement  de- 
viation" ("Ablenkung"),  described  by  Neisser  and  Wechsberg.  The 
latter  was  advanced  as  an  explanation  of  the  inactivity  of  bacteri- 
cidal sera  when  used  in  too  great  concentration,  as  described  in  an- 
other place  {p.  160)  (Neisser  and  Wechsberg  phenomenon),  and  has 
been  variously  utilized  as  support  for  the  assertion  that  alexin  can 
unite  with  unattached  sensitizer.  It  is  regarded  by  most  observers, 
moreover,  as  untenable  in  the  light  of  later  investigation.  In  spite  of 
this,  the  term  "Komplement-Ablenkung"  has  been  employed  by  a 
number  of  German  writers  (see  Citron,  Vol.  2,  "Kraus  und  Levaditi 
Handbuch")  as  synonymous  with  "fixation"  in  the  sense  of  Bordet 
and  Gengou. 

The  phenomenon  of  Bordet  and  Gengou,  briefly  described,  is 
nothing  more  than  an  experimental  utilization  of  the  fact  which  we 
have  discussed  at  length,  that  alexin  is  fixed  by  antigen  and  antibody 
after  union,  but  by  neither  alone. 

The  condition,  as  observed  by  them,  may  be  best  described  by 
submitting  the  protocol  of  the  first  experiment  detailed  in  their 
communication : 

,  An  .emulsion  of  a  24-hour  slant  of  plague  bacilli  was  used  as 

74  Bordet  and  Gengou.    Ann.  de  VInst.  Past,,  1901,  Vol.  15,  p.  289. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     187 


antigen,  heated  antiplague  horse  serum  represented  the  antibody, 
and  fresh  guinea  pig  serum  was  used  as  alexin.  A  series  of  tubes 
was  then  prepared  as  follows : 

1. 
2. 
3. 
4. 
5. 
6. 


v 

Alexin  +  plague  bacilli  +  inactivated  antiplague  serum. 
Alexin  +  plague  bacilli  -j-  inactivated  normal  horse  serum.    ' 


Alexin  --  inactivated  antiplague  serum. 
Alexin  -j-  inactivated  normal  horse  serum. 
Plague  bacilli  +  inactivated  antiplague  serum. 
Plague  bacilli  and  normal  horse  serum. 


These  mixtures  were  left  together  for  5  hours  and,  at  the  end  of 
this  time,  sensitized  rabbit  corpuscles  were  added  to  each  tube.  The 
result  showed  hemolysis  in  all  the  tubes  except  "I,77  in  which  there 
were  plague  bacilli,  antiplague  serum,  and  alexin,  and  in  tubes  5  and 
6,  which  had  contained  no  alexin  from  the  beginning.75 

It  was  plain,  therefore,  that  the  bacilli  when  specifically  sensi- 
tized had  become  capable  of  absorbing  alexin  and  preventing  its  sub- 
sequent action  upon  the  sensitized  erythrocytes.  That  the  occurrence 
was  not  exceptional  was  stiown  by  the  fact  that,  in  the  same  series, 
similar  results  were  obtained  with  anthrax,  typhoid,  and  proteus 
bacilli,  and  their  respective  antisera. 

Schematized  in  accordance  with  the  conceptions  of  Ehrlich  our 
diagram  would  be  as  follows : 


*-  COMPLEMENT  otALEXfN 


SPECIFIC 
ANTIBODY  — » 
(PRESENT  Of?  NOT?} 


HAEMOLYTIC 
ANTIBODY 


REDBLOOD. 


0NTIGEM   _^ 

(BACTERIA ETC) 

COMPLEMENT  FIXATION  SCHEMATIZED  ACCORDING  TO  EHRLICH'S  VIEWS. 
If  the  antibody  in  I  is  present  then  complement  is  fixed  by  the  antigen-antibody 
complex,  and  is  no  longer  free,  to  act  upon  the  hemolytic  complex  II.  In  the 
same  way  antigen  I  could  be  determined  if  a  known  antibody  I  were  used. 
For,  in  the  absence  of  either  of  these  parts  of  the  complex  I  complement 
would  remain  unfixed  and  free  to  act  on  complex  II. 

We  represent  the  phenomenon  graphically  in  the  symbols  of  Ehr- 
lich  merely  because  they  facilitate  clearness  of  exposition. 

In  the  presence  of  both  parts  of  Complex  I  the  alexin  is  held  and 
75  We  will  see  later  that  imsensitized  bacteria  in  emulsion  will  non-specifi- 
cally  fix  small  amounts  of  complement. 


188  INFECTION    AND    RESISTANCE 

is  no  longer  available  for  Complex  II.  If  either  of  the  reacting 
parts,  antigen  or  sensitizer,  of  Complex  I  are  lacking  the  alexin  is 
left  unfixed  and  free  to  react  with  Complex  II. 

With  this  technique  Bordet  and  Gengou  were  able  to  demonstrate, 
by  indirect  experiment,  the  presence  of  specific  sensitizers  in  the 
sera  of  animals  immunized  with  various  bacteria,  a  fact  which  was, 
of  course,  surmised  but  had  been  amenable  to  proof  heretofore  only 
in  the  case  of  bacteria  like  the  spirillum  of  cholera  in  which  lysis 
under  the  influence  of  immune  serum  and  alexin  could  be  directly 
observed  under  the  microscope.  The  practical  possibilities  of  their 
method  were,  of  course,  immediately  apparent.  By  the  use  of  a 
known  antigen  specific  sensitizers  can  be  demonstrated  in  this  way, 
and,  vice  versa,  in  the  presence  of  a  known  antibody,  the  method 
will  serve  to  identify  the  nature  of  a  doubtful  antigen.  Thus  bac- 
terial differentiation  can  be  carried  out  by  adding  to  the  suspected 
bacteria,  in  emulsion,  a  small  quantity  of  a  known  antiserum  and 
alexin,  and  determining  whether  or  not  the  alexin  has  become  fixed. 
And,  conversely,  Bordet  and  Gengou  76  have  more  recently  utilized 
the  method  in  support  of  their  claim  of  the  specific  etiological  im- 
portance of  the  bacillus  isolated  by  them  from  whooping  cough,  by 
showing  that  the  serum  of  children  suffering  from  this  disease 
formed  a  specific  alexin-fixing  complex  when  treated  with  the  ba- 
cillus. 

The  phenomenon  of  Bordet  and  Gengou  thus  found  rapid  prac- 
tical application  in  the  diagnosis  of  a  number  of  infectious  diseases, 
and  has,  of  course,  attained  great  clinical  importance  in  the  diag- 
nosis of  syphilis  in  the  form  of  the  "Wassermann"  reaction  and  its 
many  modifications.  Before  discussing  these  practical  features  in 
greater  detail,  however,  it  will  be  useful  to  discuss  more  particularly 
the  many  important  theoretical  considerations  which  have  followed 
in  the  train  of  the  complement-fixation  phenomena. 

A  year  after  the  publication  of  Bordet  and  Gengou's  paper 
Gengou 77  made  another  fundamentally  important  observation  by 
showing  that  complement  or  alexin  fixation  was  not  limited  to  the 
complexes  of  cellular  antigens  and  their  antibodies,  but  that  the  sera 
of  animals  immunized  with  dissolved  proteins  (animal  sera,  etc.), 
when  brought  together  with  their  specific  antigens,  likewise  formed 
combinations  which  fixed  alexin.  Thus  egg-white  or  dog  serum, 
brought  together  with  "anti-egg-white"  or  "anti-dog"  rabbit  serum, 
respectively,  strongly  fixed  alexin,  whereas  neither  the  antigenic  sub- 
stances nor  the  antisera  exerted  such  fixation  alone.  The  interpre- 
tation put  upon  this  by  Gengou  was  the  following :  "In  sera  obtained 
by  injecting  rabbits  with  large  doses  of  cow's  milk,  etc.,  there  are, 
in  addition  to  the  precipitins  of  Bordet  and  Tschistovitch,  substances 

76  Bordet  and  Gengou.     Ann.  de  I'Inst.  Past.,  1906. 

77  Gengou.    Ann  de  I'Inst.  Past.,  16,  1902. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     189 

•analogous  to  the  sensitizers  described  by  Bordet  in  bacteriolytic  and 
hemolytic  sera,  and  later  found  in  the  majority  of  antimicrobial 
sera."  The  important  point  in  this  interpretation  is  that  Gengou 
conceived  the  existence  of  antiprotein  sensitizers,  in  addition  to  the 
precipitins,  formed  as  a  response  to  immunization  with  amorphous 
protein.  Moreschi  78  soon  confirmed  Gengou's  experimental  deter- 
minations, and  Neisser  and  Sachs79  took  the  further  logical  step  of 
applying  this  knowledge  to  the  determination  of  proteins  for  foren- 
sic purposes.  This,  too,  we  will  further  discuss  when  we  speak  of 
the  practical  features  of  these  phenomena.  It  thus  appears  that  the 
fixation  of  alexin  is  a  generalized  property  of  all  mixtures  in  which 
an  antigen  is  brought  into  contact  with  its  specific  antibody,  whether 
the  antigen  is  in  the  form  of  the  whole  bacterial  or  other  cell,  or  in 
that  of  a  dissolved  protein,  animal  serum,  or  egg-white,  etc. 

The  observation  of  Gengou,  though  for  a  time  insufficiently  val- 
ued, has  had  a  profound  influence  upon  the  subsequent  under- 
standing of  serum  reactions.  The  fundamental  importance  of 
this  work  was  not  fully  recognized  until  his  studies  had  found 
logical  continuation  in  the  investigations  of  Gay80  and  in  those  of 
Moreschi. 

Moreschi 81  studied  the  antihemolytic  properties  possessed  by  the 
serum  of  a  rabbit  which  had  been  treated  with  normal  goat  serum. 
He  found  that  such  a  serum  had  distinct  anticomplementary  powers 
when  it  was  added  to  a  hemolytic  system  of  ox  blood  sensitizer  (ob- 
tained against  ox  blood  from  rabbits),  and  goat  complement.  With 
such  a  hemolytic  system,  however,  there  was  anticomplementary  ac- 
tion only  against  goat  complement  and  not  against  rabbit  or  guinea 
pig  complement.  If,  however,  he  used  a  hemolytic  system  in  which 
the  amboceptor  or  hemolytic  sensitizer  employed  was  one  obtained 
from  a  goat,  the  serum  was  anticomplementary  for  all  complements 
which  were  used.  Moreschi  concluded  from  this  that  the  apparent 
anticomplementary  action  of  the  serum  could  not  be  interpreted  as 
the  action  of  a  specific  anticomplement  in  the  sense  of  Ehrlich,  but 
that  it  resulted  from  the  reaction  which  took  place  as  the  consequence 
of  union  of  the  antibody  in  the  anti-goat  rabbit  serum  and  goat  pro- 
tein, which  was  introduced  into  the  tubes,  in  the  first  case  with  the 
complement,  and  in  the  second  with  the  amboceptor.  He  proved  his 
contention  by  obtaining  similar  universal  anticomplementary  action 
when  he  added  a  little  normal  goat  serum  to  the  tubes  set  up  as  above 
described.  It  is  plain,  therefore,  that  anticomplementary  action  can 
be  explained  in  observed  cases  by  the  simple  consideration  of  the  phe- 
nomenon of  Gengou.  Similar  findings  were  later  recorded  by  Muir 

78  Moreschi.    Berl.  kl  Woch.,  No.  37,  1905. 

79  Neisser  and  Sachs.     Berl  kl.  Woch.,  No.  44,  1905. 

80  Gay.    Centralbl.  f.  Bakt.  I  Orior.  Vol.  93,  1905,  p.  603. 

P1  Moreschi.    Berl.  kl.  Woch.,  1905,  No.  37,  ibid.,  No.  4,  1906. 


190  INFECTION    AND    RESISTANCE 

and  Martin,82  and  it  may  well  be  doubted,  as  a  result  of  these  and 
other  researches,  whether  we  are  at  all  justified  in  assuming  the  ex- 
istence of  anticomplements. 

The  work  of  Gay,  published  independently  in  the  same  year  as 
that  of  Moreschi,  has,  in  a  general  way,  the  same  significance,  out 
Gay  recognized  the  relation  of  the  conditions  observed  by  him  to  the 
precipitin  reaction,  a  feature  absent  from  both  the  original  study  of 
Gengou  and  the  work  of  Moreschi.  Gay  noticed  that  an  inactivated 
hemolytic  immune  serum,  left  for  some  time  in  contact  with  its  spe- 
cific cells,  and  then  separated  from  them  by  centrifugation,  would 
often  possess  anticomplementary  or  anti-alexic  properties.  He  fur- 
ther noted  that  after  such  a  serum  had  been  freed  from  the  cells  by  a 
short  centrifugation,  if  it  was  again  vigorously  centrifugalized,  a 
slight,  cloudy  sediment  would  appear  at  the  bottom  of  the  tubes.  If 
this  sediment  was  removed  the  serum  lost  its  alexin-fixing  properties. 
He  recognized  that  the  precipitate  formed  in  these  tubes  was  a  spe- 
cific precipitate  resulting  from  the  union  of  a  precipitinogen  and  its 
antibody.  The  reaction  was  due  entirely  to  the  fact  that  insufficient 
washing  of  the  cells  used  in  producing  the  hemolysin,  gave  rise  to 
the  formation  of  precipitin  against  the  serum  of  the  animal  from 
which  the  cells  had  been  taken,  and  subsequently  insufficient  washing 
of  the  cells  of  this  same  species  employed  in  the  tests  furnished 
enough  antigen  to  give  a  precipitin  reaction  in  the  tubes  in  which 
the  inactivated  hemolytic  (and  precipitating)  serum  was 'mixed  with 
the  'cells.  Subsequently,  numerous  investigations 83  have  shown 
Gay's  interpretation  to  be  correct,  and  we  may  now  accept  it  as  a 
fact  that  precipitates  formed  by  the  union  of  specific  antigen  with 
its  antibody  possess  the  power  of  fixing  alexin  and  that,  in  a  general 
way,  this  fixation  is  proportionate  in  energy  to  the  amount  of  pre- 
cipitate which  is  formed. 

Gay  utilized  his  results  primarily  to  contradict  certain  assertions 
of  Pfeiffer  and  Friedberger  concerning  antibacteriolytic  substances 
supposed  to  occur  in  normal  sera.  These  authors  had  found  that,  if 
normal  sera  possessing  no  "antagonistic"  properties  in  the  first  place 
were  left  in  contact  with  certain  bacteria,  they  acquired  antibac- 
teriolytic properties  for  these  particular  bacteria.  Thus  normal 
inactive  rabbit  serum,  left  in  contact  with  typhoid  bacilli,  and  again 
separated  from  the  bacteria,  now  prevented  the  lysis  of  sensitized 
typhoid  bacilli  if  tested  by  the  intraperitoneal  method  spoken  of  as 
the  Pfeiffer  reaction.  Sachs  84  applied  these  observations  to  analo- 
gous hemolytic  reactions  and  obtained  similar  results.  He  found 
that,  if  normal,  inactive  rabbit  serum  was  left  in  contact  with  sheep 

82Muir  and  Martin.  Journ.  of  Hyg.,  Vol.  6,  1906.  See  also  Muir's 
"Studies  on  Immunity,"  Froude,  London,  1909. 

83  Dean.  Zeitschr.  f.  Imm.,  I,  Vol.  13,  1912. 

84  Sachs.    Deut.  med.  Woch.,  1905,  No.  18. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     191 


or  guinea  pig  corpuscles,  it  acquired  the  property  of  preventing  the 
hemolysis  of  these  corpuscles  if,  later,  it  was  brought  together  with 
them  in  the  presence  of  specific  hemolysin  and  alexin.  Gray  now 
showed  hy  experiment  that  Sachs'  method  was  referable  to  insuffi- 
cient washing  of  the  corpuscles.  When,  in  the  first  contact,  the  rab- 
bit serum  was  exposed  to  the  sheep  corpuscles,  a  certain  amount  of 
sheep  serum  adherent  to  the  cells  was  carried  over  into  the  rabbit 
serum.  This  sheep  antigen  later  reacted  with  the  antisheep  pre- 
cipitin  present  in  the  hemolytic  immune  serum  and,  in  this  way, 
fixed  alexin  and  prevented  hemolysis. 

It  seems  that  the  analysis  of  Gay  is  correct,  and  that  Sachs' 
conclusion  as  well  as  thpse  of  Pfeiffer  and  Friedberger,  by  analogy, 
cannot  be  taken  as  demonstrating  the  existence  of  specific  anticom- 
plements  or  anti-amboceptors.  Gay  has  further  offered  the  same 
mechanism  as  an  explanation  of  the  Neisser-Wechsberg  phenomenon, 
which  has  been  discussed  in  another  place. 

To  summarize,  then,  we  have  learned  that  there  are  a  number  of 
varieties  of  specific  alexin  absorption  or  fixation  processes,  one  that 
is  exerted  by  cells  treated  with  specific  sensitizer,  be  they  blood  or 
bacterial,  the  other  that  which  occurs  when  unformed  protein  is 
brought  into  contact  with  its  specific  antiserum.  The  latter  has  been 
correlated  with  the  precipitin  reaction,  in  that  it  has  been  found  that, 
whenever  a  specific  precipitate  is  formed  in  such  reactions,  it  is  this 
precipitate  on  which  the  fixation  depends.  On  the  other  hand,  it  is 
necessary  to  note  that  the  formation  of  a  precipitate  is  by  no  means 
necessary  for  the  fixation,  for,  as  is  well  known,  if  a  series  of  pre- 
cipitin tubes  are  set  up,  in  each  successive  one  of  which  the  amount 
of  antigen  is  diminished,  a  degree  of  dilution  will  soon  be  reached 
at  which  no  visible  precipitate  will  occur,  but  which  nevertheless 
will  show  alexin  fixation.  The  following  is  an  illustration  of  such 
an  experiment: 


Sheep  serum  -f  antisheep  serum 

Precipitate 

Fixation  of  0.5  c.  c. 
guinea  pig  complement 

05  c  c.  (1-20)  4-  0.5  c.  c        .     .. 

4- 

Complete 

0.5  c.  c.  (1:50)         0.5  c.  c  

+4- 

Complete 

05  c  c  (1-100)       05  c  c 

+++ 

Complete 

0.5  c.  c.  (1  :200)       0.5  c.  c  

4-4-4- 

Complete 

0  5  c  c  (1  -500)       0  5  c  c 

+  +  + 

Complete 

0.5  c.  c  (1-1000)    0.5  c.  c       

4- 

Complete 

0.5  c.  c.  (1:2,000)    0.5  c.  c  

4- 

Complete 

05  c  c.  (1-5000)    05  c.  c                  

Partial 

0.5  c.  c.  (1:10,000)  0.5  c.  c  

Partial 

0  5  c.  c.  (1  '20  000)  0  5  c.  c              

None 

From  such  experiments  it  follows  moreover  that  the  fixation  of 
alexin,  carefully  titrated,  is  a  more  delicate  method  of  determining 


192  INFECTION    AND    RESISTANCE 

the  presence  of  an  antigen  or,  vice  versa,  of  an  antibody  than  is  the 
observation  of  a  visible  precipitate,  a  fact  which  has  been  made  use 
of,  as  we  have  mentioned,  by  Neisser  and  Sachs  and  others  for  for- 
ensic antigen  determinations. 

It  should  also  be  remembered  that,  if  to  such  a  precipitate  there 
is  added  an  excess  of  the  antigen,  the  precipitate  may  be  partially 
dissolved,  and  this  dissolved  precipitate,  as  Gay  85  has  shown,  may 
possess  fixation  properties.  This,  too,  accounts  for  the  fact,  observed 
by  a  number  of  workers,  that  if,  in  a  series  of  precipitin  tests  the 
supernatant  fluids  and  the  washed  precipitates  are  separately  exam- 
ined for  alexin  fixation,  the  fixation  properties  reside  entirely  in  the 
precipitates  except  in  those  tubes  in  which  a  considerable  excess  of 
antigen  was  used  and  in  which,  as  in  tubes  1  and  2  of  the  preceding 
protocol,  the  precipitates  were  relatively  slight.  The  subject,  though 
involved,  is  worthy  of  detailed  consideration  in  this  place  since  it 
seems  to  us  to  have  an  important  bearing  on  certain  theoretical  con- 
ceptions which  will  be  taken  up  below. 

The  important  question  now  arises:  what  is  the  nature  of  the 
alexin  fixation  by  the  complexes  formed  by  unformed  proteins  with 
their  antibodies  and,  more  especially,  what  is  the  nature  of  the 
alexin  fixation  exerted  by  specific  precipitates?  There  have  been 
much  experimentation  and  speculation  concerning  this,  and  a  number 
of  different  views  are  held.  Gengou  assumed,  as  we  have  seen,  that 
this  fixation,  as  studied  by  him,  was  entirely  analogous  to  the  fixation 
by  sensitized  bacterial  or  blood  cells.  He  expressed  the  belief  that 
treatment  of  an  animal  with  an  unformed  protein  produced  not  only 
specific  precipitins  but  also  specific  sensitizers,  analogous  to  those 
produced  in  response  to  treatment  with  bacterial  or  other  cells.  He 
noticed  the  parallelism  between  the  quantity  of  the  precipitate 
formed  and  the  alexin  fixation,  but  did  not  associate  the  two  proc- 
esses. 

His  conception  of  specific  antiprotein  sensitizers  was  accepted  by 
a  number  of  workers,  and  Wassermann  and  Bruck,86  Friedberger  87 
and  several  others  brought  out  the  facts  that  actual  precipitate  forma- 
tion is  not  a  necessary  criterion  of  fixation.  Thus  the  last-named 
writer  showed  that  the  precipitating  power  of  a  serum  may  be  de- 
stroyed by  moderate  heat  without  a  corresponding  destruction  of  its 
fixing  property.  A  similar  independence  of  the  precipitation  from 
the  complement-fixing  property,  in  the  presence  of  an  antigen,  has 
been  observed  by  Muir  and  Martin.88 

85  Gay.    Univ.  of  Cal  Public,  in  Pathol,  Vol.  2,  No.  1,  1911. 

86  Wassermann  and  Bruck.    Mediz.  KL,  1905,  Vol.  1,  No.  55. 

87  Friedberger.    Deut.  med.  Woch.,  1906,  No.  15. 

88  Muir  and  Martin.    Jour,  of  Hyg.,  Vol.  6,  1906. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     193 

Gay,89  90  also,  though  he  was  the  first  definitely  to  associate 
precipitin  formation  with  the  alexin-fixing  property  and,  indeed, 
determined  a  rough  parallelism  between  the  amount  of  precipitate 
and  the  degree  of  alexin  fixation,  has  nevertheless  recently  declared 
himself  in  favor  of  the  assumption  of  the  presence  in  protein  anti- 
sera  of  two  antibodies,  the  alexin-fixing  lysins  and  the  precipi- 
tins.  This  he  does  on  the  basis  of  certain  experiments  from  which 
he  concludes  that  the  antigen-antibody  complex  which  fixes  alexin 
is  distinct  from  the  precipitin-precipitinogen  complex,  but  is  usu- 
ally "brought  down  in  its  formation  in  such  a  way  as  to  simulate 
fixation  by  the  precipitate."  Nicolle91  goes  even  further  than 
this  in  declaring  that  the  "coagulins"  or  precipitins  are  "anti- 
corps  bons,"  which  prevent  the  action  of  the  albuminolysin  upon 
the  antigen,  thereby  inhibiting  the  liberation  of  poisonous  cleavage 
products. 

It  seems  to  the  writer9293  that  the  assumption  of  a  separation 
between  the  precipitin  and  the  albuminolysin  is  a  needlessly  compli- 
cated interpretation  of  the  phenomena.  In  order  to  elucidate  this 
point  a  comparison  was  made  between  the  fixing  properties  of  a 
mixture  of  a  protein  (sheep  serum)  and  its  antibody  and  a  mixture 
of  typhoid  filtrate  and  antityphoid  serum  in  which  it  is  known  that 
both  precipitins  and  antibacterial  sensitizers  are  present.  It  was 
shown  that,  as  stated  before,  in  the  former  mixture  the  alexin-fixing 
property  resided  entirely  in  the  precipitate,  whereas  in  the  latter 
case  both  the  precipitate  and  the  supernatant  fluid  fixed  alexin. 
From  this  it  seems  to  follow  that  immunization  with  the  more  com- 
plex cellular  elements  has  given  rise  to  the  precipitating  antibody 
present  also  in  the  antisheep  serum,  and,  in  addition  to  this,  to  sensi- 
tizers which  are  not  precipitable  (remaining  in  the  supernatant 
liquid)  and  not  present  in  the  antisheep  serum.  The  precipitates, 
moreover,  were  found  to  fix  "end-piece"  and  "mid-piece,"  frac- 
tions of  alexin,  in  the  same  way  as  these  are  fixed  by  sensi- 
tized cells. 

Without  going  into  further  complicated  detail,  it  would  seem  to 
us  94  to  be  justified  that  we  look  upon  the  so-called  precipitins  not  as 
separate  antibodies  but  as  identical  with  so-called  albuminolysins. 
They  unite  with  the  antigen,  producing  an  alexin-fixing  complex. 
Since  both  reacting  bodies  are  colloidal  in  nature,  they  precipitate 
each  other  in  the  test  tube,  but,  following  the  laws  governing  other 
mutually  precipitating  colloids,  they  do  so  only  when  brought  to- 

89  Gay.    Loc.  cit. 

90  Also  Univ.  of  Cal.  Publ.  in  PathoL,  Vol.  2,  No.  1,  1911. 

91  Nicolle.    Ref.  in  Bull,  tie  VInst.  Past.,  Vol.  5,  1907. 

92  Zinsser.    Journ.  Exp.  Med.,  Vol.  15,  1912. 

93  Zinsser.    Proc.  of  Soc.  of  Exp.  Biol.  and  Med.,  April,  1913. 

94  Zinsser.    Journ.  of  Exp.  Med.,  Sept.,  1913. 


194  INFECTION    AND    RESISTANCE 

gether  in  concentrations  which  lie  within  definite  zones  of  relative 
proportions.  The  visible  precipitation  would  seem,  therefore,  to  be 
merely  a  secondary  phenomenon,  the  essential  one  being  the  union  of 
an  antigen  with  a  sensitizer  by  which  it  is  rendered  amenable  to  the 
action  of  the  alexin.  This  would  enable  us  to  comprehend  also  the 
experiments  of  Friedberger,  discussed  in  the  section  on  anaphylaxis, 
in  which  it  was  shown  that  the  action  of  alexin  upon  precipitates 
gives  rise  to  the  formation  of  toxic  bodies  just  as  this  occurs  when 
alexin  acts  upon  sensitized  cells.  It  leads,  moreover,  to  a  compre- 
hension of  the  processes  of  the  digestion  of  intravascularly  intro- 
duced foreign  proteins,  which  are  rendered  amenable  to  the  digestive 
action  of  the  alexin  by  the  antibodies  spoken  of  as  precipitins,  which 
functionally  and  in  structure  are  conceived  as  identical  with  other 
sensitizers. 

Dean,95  who  has  lately  analyzed  the  relation  between  precipita- 
tion and  alexin  fixation  on  the  basis  of  extensive  experimentation, 
comes  to  the  conclusion  that  the  proportions  of  antigen  and  antibody 
which  are  favorable  for  rapid  and  complete  precipitation  do  not 
favor  the  most  complete  alexin  fixation.  He  states  that  the  two  reac- 
tions do  not  run  a  parallel  course  but  believes  that  this  does  not  mean 
that  they  are  necessarily  distinct  phenomena.  He  says :  "They  rep- 
resent two  phases  of  the  same  reaction  ...  a  flocculent  precipitate 
represents  the  final  stage  of  a  change  which  can  be  recognized  in  its 
earliest  and  incomplete  stage  by  means  of  a  complement  fixation." 

Our  view  differs  from  this  only  in  that  we  believe  that  the  pre- 
cipitation is  merely  a  secondary,  colloidal  phenomenon,  which  may, 
or  may  not,  coincide  with  the  phase  of  greatest  alexin  fixation,  ac- 
cording to  other  fortuitous  conditions  which  may  favor  or  retard 
flocculation.  Indeed,  if  our  view  be  accepted,  rapid  compact  pre- 
cipitation may  possibly  be  assumed  to  interfere  with  alexin  fixation 
in  that  it  would  inhibit  perfect  contact  of  the  alexin  with  the  antigen- 
antibody  complexes. 

Another  view  of  the  mechanism  of  alexin  fixation  is  that  which 
has  been  advanced  by  Neufeld  and  Haendel.96  These  workers  have 
found  that  sensitized  cholera  spirilla  will  fix  hemolytic  complement 
at  0°  C.,  whereas  the  same  bacteria  at  37°  C.  will  fix  both  the  hemo- 
lytic and  the  bactericidal  complement.  They  conclude  from  this  that 
the  fixation  at  37°  C.  was  brought  about  by  virtue  of  the  bactericidal 
amboceptor,  whereas  at  0°  C.  fixation  was  brought  about  by  an  anti- 
body which  is  distinct  from  amboceptor  or  sensitizer.  They  believe 
from  this  and  other  observations,  which  we  cannot  consider  in  detail, 
that  alexin  fixation  may  be  brought  about  by  a  special  fixing  anti- 

95  Dean.    Zeitschr.  f.  Imm.,  Vol.  13,  1912. 

96  Neufeld  and  Haendel.    Arb.  a.  d.  kais.  Gcsund.,  Vol.  28,  1908. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     195 

body,  the  "Bordetscher  Antikorper,"  which  is  not  identical  with  any 
of  the  other  known  antibodies. 

In  all  experiments  which  deal  with  alexin  fixation  by  specific 
antigen-antibody  complexes  it  is  of  the  greatest  importance  that  we 
should  guard  against  the  errors  easily  introduced  by  fortuitous  non- 
specific antihemolytic  agencies.  Thus  there  are  a  number  of  factors 
which  will  interfere  with  the  functionation  of  alexin  upon  a  sensi- 
tized antigen,  either  by  direct  non-specific  absorption  of  the  alexin 
itself  or  by  producing  physical  conditions  in  the  presence  of  which 
alexin  cannot  act. 

Thus  many  animal  tissue  cells,  in  emulsion,  will  absorb  alexin, 
and  the  same  property  may  be  possessed  by  tissue  extracts.  Von 
Dungern  97  was  the  first  to  call  attention  to  this,  and  his  observations 
have  been  variously  confirmed.  Muir  98  showed  that  the  stromata  of 
hemolyzed  red  blood  cells  exert  strong  anticomplementary  action, 
and  that  this  is  due  to  a  firm  union  with  the  complement.  It  is  not  un- 
likely that  the  action  of  cells  in  this  respect  is  referable  to  their 
lipoidal  contents.  This  suggestion  was  first  made  by  Landsteiner 
and  von  Eisler,"  who  found  that  the  petroleum-ether  extracts  of  red 
blood  cells  possessed  strong  anticomplementary  action  which,  to  a 
limited  extent,  was  specific  toward  the  particular  corpuscles  from 
which  the  extracts  had  been  made.  Similar  observations  have  been 
made  by  Noguchi,100  who  speaks  of  the  substance  he  extracts  as  "pro- 
tectin."  In  general,  the  protective  action  of  the  lipoidal  extracts 
seems  to  depend  largely  upon  cholesterin,  and,  since  this  substance  is 
present  to  some  extent  in  many  tissues,  their  antihemolytic  action  is 
easily  understood.  In  another  section  we  have  discussed  the  similar 
neutralizing  action  of  lipoidal  substances  upon  poisons  of  various 
kinds  (saponin,  tetanolysin,  and  snake  poison),  but,  as  we  have  noted 
there,  the  neutralizing  properties  of  the  extracts  do  not,  as  a  rule, 
equal  those  of  the  whole  tissues.101  It  is  not  unlikely  that  in  such 
cases  as  Landsteiner  suggests  the  potent  agent  is  not  the  lipoid  itself 
but  rather  a  lipoid-protein  combination,  a  class  of  substances  of  which 
we  know  very  little,  but  the  importance  of  which,  in  many  phases  of 
serum  reactions,  seems  assured. 

We  have  already  mentioned  that  yeast  cells  may  absorb  alexin. 
And  it  has  been  found  by  Wilde  102  and  others  that  almost  all  bac- 
teria in  emulsion  may  possess  varying  degrees  of  alexin-fixing  prop- 
erties even  though  unsensitized.  There  seems  to  be  no  regularity 
either  qualitatively  or  quantitatively  in  regard  to  this,  but  the  fixa- 

97  Von  Dungern.     Munch,  med.  Woch.,  Nos.  20  and  28,  1900. 

98  Muir.     "Studies  in  Immunity,"  London,  Vol.  19. 

99  Landsteiner  and  von  Eisler.     Wien.  kl  Woch.,  No.  24,  1904. 

100  Noguchi.    Journ.  Exp.  Med.,  Vol.  8,  1906,  p.  726. 

101  See  also  Ivar  Bang,  "Bioehemie  der  Lipoide." 

102  Wilde.    Berl.  kl.  Woch.,  1901,  Vol.  38,  and  Archiv  f.  Hyg.,  39,  1902. 


196  INFECTION    AND    RESISTANCE 

tion  is  usually  sufficiently  marked  to  render  the  use  of  whole  bacteria 
unreliable  for  specific  fixation  experiments.  For  this  reason,  as  we 
will  see,  bacterial  extracts  must  be  used  in  such  work  unless  careful 
quantitative  controls  are  made.  Upon  what  this  fixation  depends  it 
is  difficult  to  determine.  It  may  be  that  it  is  purely  non-specific  and 
due  to  absorption  of  the  fine  emulsion  of  the  bacteria  comparable  to 
that  observed  on  the  part  of  kaolin  or  quartz  sand  emulsions,  or,  pos- 
sibly fixation  by  such  bacterial  emulsions  may  occur  because  of  the 
small  amounts  of  normal  sensitizer  almost  always  present  in  the 
serum  employed  as  alexin. 

Apart  from  the  lipoids,  a  number  of  other  substances  have  been 
found  to  fix  alexin  and  exert  consequent  antihemolytic  action.  Thus 
Landsteiner  and  Stankovic,103  and  Landsteiner  and  von  Eisler  104 
describe  the  anti-alexic  action  of  various  proteins  coagulated  or  pre- 
cipitated. They  refer  this  action  not  to  particular  chemical  struc- 
ture but  to  the  colloidal  state,  since  they  obtained  similar  antilytic 
action  with  such  inorganic  emulsions  as  quartz  sand  and  kaolin 
(aluminium-orthosilicate).  Since  anticomplementary  action  has, 
moreover,  been  noted  in  the  case  of  a  large  number  of  extracts  of 
such  materials  as  wool,  leather,  etc.,  it  is  clear  that  the  methods  of 
alexin  fixation,  as  applied  to  the  forensic  differentiation  of  blood, 
must  be  carefully  controlled  with  this  point  in  view.105 

Among  the  most  practically  important  non-specific  agencies 
which  fix  alexin  there  are  some  which  appear  under  certain  condi- 
tions in  normal  serum.  Noguchi  106  has  found  that  serum  will  often 
develop  anticomplementary  properties  as  a  consequence  of  heating 
during  the  process  of  inactivation.  On  more  detailed  investigation 
he  determined  that  the  anticomplementary  action  increased  as  the 
serum  was  heated  to  about  90°  C.  Above  this  temperature  it  is  de- 
stroyed. He  refers  this  property  to  the  serum  lipoids,  since  he  was 
able  to  remove  it  by  extraction  with  ether,  the  ether  extract  possessing 
the  same  anticomplementary  power  as  the  original  serum. 

Neisser  and  Doring 1  °7  have  noticed  anti-alexic  or  anticomple- 
mentary properties  of  human  sera  which  were  destroyed  on  heating, 
and  which  they  associate  with  disease  of  the  kidneys,  since  they 
noted  it  in  sera  of  uremic  patients.  Browning  and  McKenzie  1( 
have  observed  a  similar  heat-sensitive  anti-alexic  action  on  the  part 
of  normal  serum,  and  the  subject  has  been  studied  by  Zinsser  and 
Johnston.109  It  was  found  that  all  normal  sera  will  develop  anti- 

103  Landsteiner  and  Stankovic.     Centralbl  f.  Bakt.,  1906,  Vols.  41  and  42. 

104  Landsteiner  and  von  Eisler.     Wien.  kl.  Woch.,  1904,  No.  24. 

105  Uhlenhuth.     Deut.  med.  Woch.,  1906,  Nos.  31  and  51,  and  Centralbl.  f. 
Bakt.,  1906,  I,  Ref.,  Vol.  38. 

106  Noguchi.    Journ.  of  Exp.  Med.,  Vol.  8,  1906,  p.  726. 

107  Neisser  and  Doring     Berl.  kl.  Woch.,  1901,  No.  22. 

108  Browning  and  McKenzie.    Journ.  of  Path,  and  Bact.,  Vol.  13,  1909. 

109  Zinsser  and  Johnston.     Journ.  of  Exp.  Med.,  Vol.  13,  1911. 


FURTHER    DEVELOPMENT    OF    KNOWLEDGE     197 

alexic  properties  on  preservation  at  room  temperature  within  a  few 
days,  and  more  slowly  but  no  less  regularly  in  the  ice  chest.  This 
anti-alexin  is  destroyed  on  heating  to  56°  C.,  and  may  be  precipitated 
out  with  the  globulins  of  the  serum.  There  appeared  in  these  studies 
no  particular  association  between  the  anti-alexic  property  and  ne- 
phritis. 

The  action  of  alexin  upon  sensitized  cells  may  be  prevented,  also, 
by  physical  or  chemical  conditions  without  actual  fixation  or  binding 
of  the  alexin.  We  refer  to  the  effects  of  the  addition  of  salts,  prob- 
lems which  have  been  considered  above. 


CHAPTER   VIII 

PEACTICAL    APPLICATIONS    OF    THE    COMPLEMENT- 
FIXATION  METHOD 

THE  WASSERMANN  REACTION 

THE  principle  of  specific  alexin  fixation  has  been  practically 
utilized  in  the  diagnosis  of  disease  and  in  the  forensic  determination 
of  the  nature  of  spots  of  blood  or  other  protein  material. 

Soon  after  Bordet  and  Gengou's  experiments  Wassermann  and 
Bruck  1  showed  that  bacterial  extracts  could  be  successfully  substi- 
tuted for  whole  bacteria  in  these  reactions.  Citron,2  too,  made  sim- 
ilar observations,  and,  indeed,  we  now  know  that  the  use  of  bacterial 
extracts  is  more  suitable  for  these  experiments  than  are  emulsions 
of  whole  bacteria,  since,  as  we  have  mentioned  above,  bacterial  emul- 
sions may  often  fix  small  amounts  of  complement  of  themselves 
(without  specific  sensitization),  thereby  confusing  the  results  of  the 
reaction. 

On  the  basis  of  their  experience  with  bacterial  extracts  Wasser- 
mann and  Bruck  3  then  determined  that  complement  fixation  could 
be  carried  out  in  tuberculosis  when  the  various  tuberculin  prepara- 
tions were  used  as  antigen.4  These  investigations  fell  into  the  period 
during  which  active  research  upon  the  Spirochoeta  pallida  in  syphilis 
was  going  on,  and  it  occurred  to  Wassermann  that  the  technique  of 
complement  or  alexin  fixation  might  be  utilized  in  the  diagnosis  of 
syphilis.  Together  with  Neisser  and  Bruck  5  he  subjected  this  idea 
to  experimental  test.  The  publication  of  their  first  results  appeared 
in  1906.  They  used  in  their  experiments  the  syphilitic  monkeys 
which  were  being  observed  in  Neisser's  clinic.  Their  method  con- 
sisted in  mixing  inactivated  serum  from  syphilis-inoculated  monkeys 
with  organ  extracts,  serum,  etc.,  of  syphilitic  human  beings,  and 

1  Wassermann  and  Bruck.     M ed.  Klinik,  Vol.  55,  1905. 

2  Citron.    Centralbl.  f.  Bakt.,  Vol.  41,  1906. 

3  Wassermann  and  Bruck.    Deut.  med.  Woch.,  No.  12,  1906. 

4  Complement  fixation  in  tuberculosis  is  not  yet  on  a  practical  or  reliable 
basis.     Recent  claims  of  Besredka   (Ann.  Past.,  1913)   for  his  new  antigen 
promise  a  successful  technique,  but  no  extensive  confirmation  has  followed  up 
to  the  present  time. 

6  Wassermann,  A.  Neisser,  and  Bruck.    Deut.  med.  Woch.,  No.  19,  1906. 

198 


PRACTICAL    APPLICATIONS    OF    METHOD         199 

adding  a  small  amount  of  fresh  guinea  pig  complement.  After  these 
materials  had  been  together  for  a  certain  time,  sensitized  red  blood 
cells  were  added.  If  the  complement  was  bound  during  the  first 
exposure  no  hemolysis  resulted  and  the  reaction  was  regarded  as 
positive.  From  their  results  they  drew  the  following  conclusions : 

1.  Immune  serum  from  monkeys,  produced  by  treatment  with 
syphilitic  material,   will  sensitize  syphilitic  material  from  human 
beings  or  monkeys,  so  that  an  alexin-fixing  complex  is  formed. 

2.  Complement  fixation  results  only  when  the  syphilitic  immune 
serum  of  monkeys  is  added  to  similar  material  from  men  or  mon- 
keys, but  not  when  added  to  organ  extracts  of  normal  men  or  mon- 
keys. 

3.  Normal  monkey  serum  has  no  such  action. 

They  concluded  that  their  results  justified  them  in  assuming  a 
specific  fixation  due  to  specific  antisyphilitic  immune  bodies  in  the 
blood  of  the  treated  monkeys.  They  excluded  experimentally  the 
possibility  of  fixation  by  a  precipitin  reaction  resulting  from  the 
treatment  of  the  monkeys  with  human  material.  It  might  well  have 
happened  that  precipitins  against  human  protein  appearing  in  the 
serum  of  the  treated  monkeys  might  subsequently  react  with  the 
human  protein  material  used  as  antigen,  a  complement-fixing  com- 
plex resulting.  This,  however,  was  excluded  by  the  fact  that  they 
obtained  positive  reactions  only  when  the  human  material  was  ob- 
tained from  luetic  lesions. 

The  same  authors,  with  Schucht,6  very  soon  after  this,  extended 
their  method  to  the  diagnosis  of  syphilis  in  human  beings.  The 
same  thing  had  been  done  shortly  before  their  publication  appeared 
by  Detre7  on  a  smaller  material.  By  these  and  many  other  investi- 
gations it  was  very  soon  shown  that  syphilis  may  be  reliably  diag- 
nosed by  complement  fixation  when  extracts  of  the  syphilitic  organs, 
employed  as  antigen,  are  mixed  with  the  inactivated  serum  of  syphi- 
litic individuals.  It  was  incidentally  shown  by  Wassermann  and 
Plaut  8  that  the  reaction  could  be  obtained  not  only  with  blood  serum 
but  also  with  spinal  fluid  in  paralytic  cases. 

It  was  generally  assumed,  at  this  time,  that  the  reaction  in  syph- 
ilis depended,  as  in  the  case  of  other  infections,  upon  the  presence  in 
the  syphilitic  serum  of  specific  antibodies.  For  it  seemed  reasonable 
to  suppose  that  the  specific  antigen  obtained  in  the  extracts  was  de- 
rived from  the  extraction  of  large  numbers  of  spirochetes  demonstra- 
ble in  the  extracted  organs. 

This,  of  course,  is  the  most  logical  and  simple  theoretical  concep- 
tion of  the  reaction,  and  is  justified  on  the  basis  of  analogy.  Un- 

6  Wassermann,  Neisser,  Bruck,  and  Schucht.     Zeitschr.  f.  Hyg.,  Vol.  55, 
1906. 

7  Detre.     Wien.  kl.  ~Woch.,  Vol.  19,  No.  21,  1906. 

8  Wassermann  and  Plaut.    Deut.  med.  Woch.,  No.  44,  1906. 


200  INFECTION    AND    RESISTANCE 

fortunately,  however,  it  was  soon  found  by  a  number  of  workers, 
Marie  and  Levaditi,9  Weygant,  Kraus  and  Volk,  Landsteiner, 
Miiller,  and  Potzl,10  and  others  that  antigens  perfectly  capable  of  fix- 
ing complement  in  the  presence  of  syphilitic  serum  could  be  pro- 
duced from  normal  organs.11 

Theoretically  it  must  be  admitted  that  we  are  very  much  in  the 
dark  at  present.  The  fact,  now  entirely  unquestionable,  that  the 
sera  of  syphilitic  patients  will  give  fixation  with  antigens  derived 
from  extracts  of  normal  organs,  as  well  as  from  those  of  syphilitic 
organs,  seems  to  throw  doubt  upon  the  simple  specific  antigen-anti- 
body conception  at  first  held. 

In  order  to  understand  the  questions  involved  in  the  theories  of 
the  Wassermann  reaction  as  at  present  conceived  it  will  be  necessary 
to  consider  the  types  of  antigen  which  are  now  employed. 

Wassermann' s  original  method  of  antigen  preparation  consisted 
in  using  the  liver  or  spleen  of  a  congenitally  syphilitic  fetus.  The 
organs  were  finely  divided  and  emulsified  in  4  to  6  parts  of  normal 
salt  solution.  This  mixture  was  shaken  for  24  hours,  eentrifugal- 
ized,  and  the  clear  supernatant  fluid  used  as  antigen.  Later  the 
specific  organ  substances  were  extracted  by  Forges  and  Meier  12  in 
five  times  the  volume  of  absolute  alcohol  for  24  hours.  This  alco- 
holic extract  was  evaporated  in  vacuo  and  the  residue  taken  up  in 
salt  solution  and  shaken  until  an  even  suspension  resulted. 

After  it  had  been  discovered  that  normal  organ  extracts  could 
serve  as  antigen  as  well  as  the  extracts  of  syphilitic  organs,  Land- 
steiner,  Forges  and  Meier,  and  others,  introduced  antigens  produced 
by  alcoholic  extraction  of  normal  organs  of  animals  and  of  man. 
Landsteiner  introduced  the  alcoholic  extract  of  normal  guinea  pig 
organs,  especially  extracts  of  the  heart  and  liver,  and  Weil  and 
Braun  13  made  use  of  extracts  of  normal  human  organs.  There  are 
various  methods  of  preparing  extracts  for  this  purpose.  We  may 
mention,  to  illustrate  these  methods,  the  one  suggested,  first,  we  be- 
lieve, by  Noguchi,  a  procedure  which  is  applicable  to  the  extraction 
of  normal  human  organs  (spleen),  beef  hearts,  and  guinea  pig  hearts. 
The  finely  divided  or  triturated  organ  substance  is  shaken  up  with 
five  times  its  weight  of  absolute  alcohol  and  allowed  to  stand  in  the 

9  Marie    and    Levaditi.      Cited   from   Mclntosh    and   Fildes'    "Syphilis." 
Longmans  &  Co.,  1911,  p.  94. 

10  Landsteiner,  Miiller,  and  Potzl.     Wien.  kl  Woch.,  Vol.  20,  1907. 

11  An  extensive  historical  review  of  the  development  of  the  Wassermann 
reaction   is  found  in  the  book  of  Boas,   "Die  Wassermannsche  Reaktion," 
Karger,  Berlin,  1911.     Since  these  earlier  publications  have  appeared  the 
literature  of   the  Wassermann  reaction  has   become  very   extensive.     It  is 
enumerated  more  fully  than  we  can  afford  space  for  here  in  the  book  of 
Noguchi  ("Serum  Diagnosis  of  Syphilis")  and  that  of  Boas,  mentioned  above. 

12  Forges  and  Meier.    Berl  kl  Woch.,  No.  15,  1908. 

13  Weil  and  Braun.     Berl.  kl.  Woch.,  No.  49,  1907. 


PRACTICAL    APPLICATIONS    OF    METHOD         201 

incubator  at  37.5°  C.,  for  from  5  to  7  days.  At  the  end  of  this  time 
it  is  filtered  through  cheesecloth  and  then  through  coarse  paper,  and 
the  filtrate  placed  in  a  large  crystallizing  dish  in  which  it  is  evapo- 
rated at  room  temperature  with  the  aid  of  an  electric  fan.  A  gummy 
yellow  residue  is  left,  which  is  then  taken  up  in  as  small  a  quantity 
of  ether  as  possible.  This  ether  solution  is  then  precipitated  with 
4  times  its  volume  of  acetone,  in  consequence  of  which  there  is  a  pro- 
fuse precipitation  of  coarse  white  flakes.  This  acetone-insoluble 
substance,  which  is  at  first  white,  later  yellowish,  in  color,  is  the 
stock  antigen.  A  little  of  this  is  taken  up  in  a  very  small  quantity 
of  ether,  and  this  ethereal  solution  is  shaken  up  in  salt  solution  until 
the  ether  has  evaporated  or  the  material  has  gone  into  very  fine  col- 
loidal suspension  in  the  salt  solution.  This  is  the  antigen  ready  to 
be  used. 

It  is  immediately  evident  that  these  antigenic  substances  must 
consist  very  largely  of  lipoidal  extractives  of  the  organ  substances, 
and  it  has  been  found  that  such  antigen  contains  sodium  oleate,  leci- 
thin and  cholesterin.  Indeed,  Forges  and  Meier  have  claimed  that  a 
1  per  cent,  solution  of  commercial  lecithin  may  be  used  with  success. 
Browning  and  Cruikshank  14  have  found  further  that  the  addition 
of  small  amounts  of  cholesterin  to  syphilitic  antigen  very  largely 
increases  its  specifically  diagnostic  value,  and  this  idea  has  since 
been  utilized  more  especially  by  Sachs,15  Walker  and  Swift,16  and 
others.  Sachs,  especially,  has  obtained  excellent  antigens  in  the 
following  way:  1  gram  of  moist  guinea  pig  heart  substance  was 
extracted  with  5  c.  c.  of  alcohol  and  left  at  room  temperature  for 
twelve  hours  or  in  the  ice  box  for  two  days ;  it  was  then  filtered  and 
0.5  to  1  per  cent,  of  cholesterin  was  added;  frequently  the  alcohol 
extract  had  to  be  diluted  two  or  three  times  before  use.  Sachs  and 
Rondoni  17  have  also  recommended  artificial  mixtures  of  lipoids  con- 
taining sodium  oleate,  lecithin,  and  oleic  acid. 

The  fact  that  cholesterin  added  to  alcoholic  organ  extracts  in- 
creases the  antigenic  value  of  these  for  the  Wassermann  reaction  is 
all  the  more  curious  inasmuch  as  cholesterin  alone  has  practically  no 
antigenic  action.  Walker  and  Swift  have  recommended  an  antigen 
in  which  alcoholic  extracts  of  human  or  guinea  pig  hearts  were  made 
up  to  0.4  per  cent,  of  cholesterin,  0.4  per  cent/having  been  found  by 
comparative  test  to  be  the  most  favorable  concentration.  Cholesterin- 
liver  extracts  or  even  alcoholic  extracts  of  syphilitic  livers  without 
cholesterin  were  found  to  be  inferior  in  specific  antigenic  value  to 
0.4  per  cent,  cholesterin-heart  antigens.  From  the  experience  of 
many  investigators  it  now  seems  unquestionable  that  additions  of 

14  Browning-  and  Cruikshank.     Journ.  of  Path,  and  Bact.,  Vol.  16,  1911- 

15  Sachs.    Berl  kl  Woch.,  No.  46,  1911. 

16  Walker  and  .Swift.     Journ.  of  Exp.  Med.,  Vol.  18,  1913. 

17  Sachs  and  Rondoni.    Zeitschr.  f.  Imm.,  Vol.  1,  1909. 


INFECTION    AND    RESISTANCE 


cholesterin  increase  the  delicacy  of  the  reaction  in  that  more  cases 
react  positively  with  such  an  antigen  than  with  the  imcholesterinized 
preparations.  The  experience  of  Hopkins  and  Zimmermann,  how- 
ever, would  indicate  that  great  caution  must  be  exercised  when  the 
reaction  is  done  in  this  way,  since  occasional  positive  results  are 
obtained  with  cases  clinically  not  syphilitic.  These  workers  believe 
that  cholesterinized  antigen  is  extremely  useful,  but  advise  its  use 
only  parallel  with  the  ordinary  lipoidal  antigens  and  together  with 
careful  study  of  the  clinical  aspects  of  the  case. 

The  fact  that  these  antigens  are  non-specific  in  origin  naturally 
necessitates  careful  determination  of  their  usefulness  before  they 
are  used.  Before  any  antigen  can  be  regarded  as  reliable,  therefore, 
a  titration  must  be  carried  out  in  the  following  way :  Two  series  of 
tubes  are  prepared,  in  the  first  of  which  antigen  and  complement  are 
added  to  normal  serum,  and  in  the  second  the  same  substances  are 
added  to  known  syphilitic  serum.  The  antigen  must,  of  course,  be 
such  that  in  no  test  tube  does  it  cause  alexin  fixation  in  the  presence 
of  normal  serum,  but,  in  the  quantities  used,  it  must  give  fixation 
regularly  with  syphilitic  serum.  An  example  of  such  a  titration 
may  be  tabulated  as  follows : 


EXAMPLE  OF  ANTIGEN  TITEATION 

Antigen  by  Landsteiner's  method :  normal  guinea  pig  heart 
freed  from  fat  and  ground  up  in  a  mortar.  To  each  gram  is  added 
5  c.  c.  of  absolute  ethyl  alcohol  and  the  mixture  allowed  to  extract 
at  60°  C.  for  12  hours  (or  several  days  at  37.5°  C.).  It  is  then  fil- 
tered through  paper.  The  following  titration  is  then  carried  out : 


A 

Tube  1 

Tube  2 

Tube  3 

Tube  4 

Tube  5 

Normal  serum  

0.2 

0.2 

0.2 

0.2 

Antigen 

0  05 

0  1 

0  2 

0  3 

0  6 

Alexin      

0.1 

0  1 

0  1 

0.1 

0    1 

B 

Tube  1 

Tube  2 

Tube  3 

Tube  4 

Syphilitic  serum 

0  2 

0  2 

0  2 

0  2 

Antigen  

0.05 

0.1 

0.2 

0.8 

Alexin  

0  1 

0  1 

0  1 

0  1 

The  volume  in  all  of  these  tubes  is  brought  to  3  c.  c.  with  isotonic 
salt  solution.     After  one  hour  at  37.5°  C.,  sensitized  red  cells  are 


PRACTICAL    APPLICATIONS    OF    METHOD         203 

added  to  each  tube.18  If  the  antigen  is  suitable  in  that  it  does  not 
fix  alexin  by  itself  or  in  the  presence  of  normal  serum,  hemolysis 
will  result  in  all  of  the  tubes  of  series  A.  If  it  is  suitable  in  that  it 
fixes  in  the  presence  of  syphilitic  serum,  the  tubes  in  series  B  will 
show  no  hemolysis ;  if  there  is  slight  hemolysis  in  B  1,  it  is  inferred 
that  0.05  c.  c.  of  the  antigen  is  insufficient,  and  the  smallest  amount 
(0.1  c.  c.),  which  completely  fixes  0.1  G.  c.  of  alexin  in  the  presence 
of  the  positive  serum,  is  the  quantity  used.  Again  the  antigen  may 
be  able  to  cause  hemolysis  by  itself  if  used  in  too  large  amounts.  If 
this  is  the  case  in  tube  B  4,  then  this  antigen  is  suitable  only  in 
amounts  varying  between  0.1  c.  c.  and  0.2  c.  c. 

The  titration  is  done  with  varying  quantities  because  too  little 
antigen  might  fail  in  fixing  the  alexin,  even  if  the  serum  were  posi- 
tively syphilitic,  whereas  too  much  antigen  might  possess  alexin-fix- 
ing  properties  in  itself,  even  in  the  presence  of  normal  serum,  or 
possibly  without  any  serum  at  all,  an  attribute  which  is  not  uncom- 
monly possessed  by  lipoidal  extracts. 

It  is  thus  seen  that  Wassermann  reactions  can  be  carried  out 
with  antigens  which  do  not  contain  extracts  of  syphilitic  lesions  or 
of  the  micro-organisms  which  give  rise  to  syphilis.  This  fact  alone 
would  exclude  the  possibility  of  considering  the  fixation  of  comple- 
ment as  at  present  carried  out  in  the  Wassermann  reaction  as  being 
due  to  a  specific  antigen-antibody  union. 

This  conclusion  is  strengthened  by  the  recent  discovery  that  a 
specific  antigen  prepared  from  cultures  of  8pirocJiceta  pallida  cannot 
be  successfully  used  in  diagnostic  Wassermann  tests.  The  first  in- 
vestigations of  this  kind  were  made  by  Schereschewsky,19  who  used 
as  antigen  extracts  of  mixed  cultures  in  which  the  spirochete  was 
present ;  his  results  were  inconclusive.  Noguchi  20  later  investigated 
this  phase  of  the  problem,  preparing  his  antigens  by  the  extraction  of 
pure  cultures  and  of  syphilitic  rabbit  testicles  in  which  the  spirochetes 
were  very  profuse.  He  found  that  positive  tests  with  such  an  antigen 
were  obtained  only  in  isolated  cases  of  prolonged  syphilis  which  had 
been  thoroughly  treated,  and  that  the  ordinary  Wassermann  reaction, 
as  obtained  in  active  cases,  is  not  due  to  antibodies  which  combine 
specifically  with  the  pallida  antigen.  Craig  and  ISTichols  21  also  have 
found  that  cases  of  untreated  syphilis  which  gave  positive  reactions 
with  syphilitic  liver  extracts  gave  absolutely  negative  results  when 
culture  antigens  were  used. 

18  Tube  "5"  is  the  antigen  control  which  shows  that  the  antigen  in  large 
amounts  is  neither  anticomplementary  nor  hemolytic  by  itself.     It  is  well,  in 
addition,  also  to  test  out  various  amounts  of  the  antigen  and  alexin,  without 
either  normal  or  syphilitic  serum,  to  determine  the  largest  amount  of  antigen 
which,  by  itself,  is  devoid  of  the  actions  mentioned  above. 

19  Schereschewsky.     Deut.  med.  Woch.,  1909,  p.  1653. 

20  Noguchi.     Journ.  A.  M.  A.,  Vol.  58,  1912. 

21  Craig  and  Nichols.    Journ.  of  Exp.  Med.,  Vol.  16,  1912. 


204  INFECTION    AND    RESISTANCE 

From  these  results  also  we  may  infer  that  the  Wassermann  reac- 
tion does  not  represent  a  fixation  of  alexin  by  the  union  of  a  specific 
syphilitic  antigen  with  antibodies  found  against  the  Spirochceta  pal- 
lida.  Noguchi  concludes  that  it  is  caused  by  "lipotropic"  substances 
in  the  sera  of  syphilitic  human  beings ;  a  conclusion  which  is  justi- 
fied by  the  fact  that  the  antigens  used,  all  of  them,  contain  large 
quantities  of  lipoids.  It  must  be  acknowledged,  however,  that  we 
have  no  definite  information  concerning  the  nature  of  the  reaction 
beyond  this.  Schmidt  22  believes  that  it  is  a  colloidal  reaction,  and 
depends  upon  the  union  of  the  serum  globulins  with  the  extract 
colloids  in  the  antigen.  In  normal  serum  such  a  union  is  prevented 
by  the  albumins  which  act  as  a  sort  of  protective  colloid.  In  syph- 
ilitic serum  the  globulins  are  increased  quantitatively  or  are  changed 
qualitatively  in  the  degree  of  their  dispersion,  or  possibly  in  both 
characteristics.  He  regards  the  serum  globulins  in  the  Wassermann 
reaction  as  directly  uniting  with  the  extract  colloid. 

Levaditi  and  Yamanouchi  23  also  conclude  that  the  Wassermann 
reaction  depends  upon  the  union  of  two  colloidal  substances — one  a 
non-proteid  constituent  of  syphilitic  serum  (cholesterin  derivatives 
or  fatty  acids),  the  other  the  lipoidal  constituents  of  the  antigen. 
Like  others  they  found  that  the  active  substances  in  the  antigenic 
extracts  are  non-protein  and  alcohol  soluble. 

It  is  interesting  to  note,  moreover,  that  Forges  and  Meier  24  ob- 
served actual  precipitation  when  syphilitic  serum  was  added  to 
lecithin  emulsions.  In  consequence,  attempts  have  been  made  to 
make  the  diagnosis  of  syphilis  by  direct  precipitation  of  syphilitic 
serum  by  such  emulsions  of  lecithin  and  of  sodium  glycocholate 
(Merck).  The  results  of  these  investigations  as  well  as  those  of 
Klausner,25  who  claims  that  syphilitic  sera  are  more  easily  precipi- 
tated by  distilled  water  than  are  normal  sera,  have  led  to  no  diag- 
nostically  reliable  results,  but  they  have  seemed  to  show  that  the 
serum  globulins  are  probably  more  plentiful  and  more  easily  pre- 
cipitated out  of  syphilitic  than  out  of  normal  sera. 

The  inference  of  many  workers,  therefore,  has  been  that  the 
Wassermann  reaction  is  primarily  due  to  the  precipitation  of  (prob- 
ably) globulin  by  the  lipoidal  colloids  of  the  antigen,  the  resulting 
precipitate  being  capable  of  absorbing  alexin.  Jacobsthal  26  has  ex- 
amined mixtures  of  syphilitic  serum  and  antigen  by  the  ultramicro- 
scopic  method,  and  claims  that  precipitates  are  always  present  even 
when  they  are  not  macroscopically  visible.  Bergel,26  who  has  re- 

22  Schmidt.     Zeitschr.  f.  Hyg.,  Vol.  69,  1911. 

23  Levaditi  and  Yamanouchi.     C.  R.  de  la  Soc.  de  BioL,  1907,  Vol.  63,  p. 
740. 

24  Forges  and  Meier.    Berl  kl.  Woch.,  No.  15,  1908. 

25  Klausner.     Wien.  kl  Woch.,  No.  7,  1908. 

26  Jacobsthal.     Munch,  med.  Woch.,  1910. 

27  Bergel.    Zeitschr.  /.  Imm.,  Vol.  17,  1913. 


PRACTICAL    APPLICATIONS    OF    METHOD         205 

cently  suggested  the  importance  of  specific  lipase  production  as  a 
cause  of  hemolysis,  suggests  that  the  Wassermann  reaction  is  due  to 
fixation  exerted  by  the  products  of  the  action  of  a  specific  lipase 
formed  in  the  syphilitic  body  against  "lues-lipoids."  This  theory  is 
open  to  objections  similar  to  those  mentioned  above,  namely,  that 
the  antigen  need  not  necessarily  be  a  lues-lipoid,  but  may  be  derived 
from  normal  organs.  Other  theories  have  been  brought  forward  by 
Bruck,  Weil,  Braun,  Manwaring,  and  more  recently  by  Rabino- 
witch.28  The  data  supporting  most  of  these  theories  are,  as  yet,  too 
speculative  to  justify  our  discussion  of  them  at  any  length.  The 
only  fact  which  seems  established  with  any  reasonable  certainty  is 
the  independence  of  the  Wassermann  test  from  a  specific  antigen- 
antibody  reaction  in  the  usual  sense. 

Although  the  Wassermann  reaction  is  thus  apparently  not  based 
on  those  principles  in  the  investigation  of  which  it  was  discovered, 
its  practical  diagnostic  value  is  not  therefore  diminished.  .  For  its 
proper  performance  any  of  the  methods  of  antigen  preparation  con- 
sidered above  may  be  employed,  provided  that  the  usefulness  of  the 
preparation  utilized  is  carefully  controlled  in  each  case  as  indicated. 
Since,  of  course,  a  hemolytic  system  is  used  in  such  tests  as  an  in- 
dicator, it  is  necessary  also  to  titrate  sensitizer  and  alexin. 

From  what  has  been  said  in  another  place  concerning  the  quanti- 
tative relations  of  alexin  and  amboceptor  or  sensitizer  (see  reference 
to  work  of  Morgenroth  and  Sachs,  p.  163),  it  is  evident  that  the  use 
of  too  strongly  sensitized  cells  might  result  in  hemolysis,  if  a  slight 
fraction  of  alexin  were  left  unbound  by  a  weak  syphilis  reaction. 
Conversely  the  use  of  too  large  a  quantity  of  alexin  would  result  in 
hemolysis,  since,  even  if  the  amount  of  syphilitic  fixation  were  con- 
siderable, a  sufficient  excess  of  alexin  might  remain.  The  use  of 
uniform  amounts  of  fresh  guinea  pig  serum  in  each  case  does  not 
control  this  adequately,  for  different  specimens  of  guinea  pig  serum 
may  vary  considerably  in  alexin  content.  In  consequence,  titra- 
tions  of  both  sensitizer  and  alexin  should  be  made.  For  practical 
purposes  it  is  quite  enough  to  titrate  the  hemolytic  sensitizer  every 
few  weeks  and  use  a  stated  amount  in  successive  reactions.  The 
alexin  or  complement  can  then  be  titrated  individually  for  each  set 
of  reactions.  Examples  of  such  preliminary  titrations  follow: 

Titration  of  Hemolytic  Amboceptor  or  Sensitizer 

Rabbit  injected  3  times  at  5-day  intervals  with  washed  sheep 
corpuscles  .  .  .  . ,  3,  4,  and  5  c.  c.,  and  bled  10  days  after  the  last 
injection.29 

28  Rabinowitch.     Centralbl.  f.  Bakt.,  Orig.,  1914. 

29  In  immunizing1  animals  with  blood  cells  for  this  or  any  other  purpose 
it  is  necessary  to  wash  the  cells  very  carefully  in  salt  solution.    Unless  this  is 


206  INFECTION    AND    RESISTANCE 

This  serum  is  inactivated  at  56°  C.  for  20  minutes. 


Washed  sheep  corpuscles 
5%  emulsion 
in  salt  solution 

Sensitizer 

Fresh 
g-  P- 
serum 

Hemolysis 

1 

1  c.  c  

0  01 

0  1 

-f-  -4-4- 

2 

1  c.  c  

0.005 

0.1 

4-4-4- 

3 

Ice 

0  003 

0  1 

4-4-4- 

4 

1  c.  c  

0.001 

0.1 

4-4-4- 

5 

Ice 

0  0005 

0  1 

4-4- 

6 

1  c.  c. 

0.0002 

0  1 

=b 

7 

Ice 

0  1 

8 

1  c.  c  

salt  sol. 

In  this  case  0.001  c.  c.  still  causes  complete  hemolysis  of  1  c.  c. 
of  a  5  per  cent,  emulsion  of  sheep  cells  (volumetric  measurement  of 
cells  sedimented  in  the  centrifuge),  and  this  amount  (-rsW.  c.  c.) 
is  called  the  "hemolytic  unit"  of  sensitizer ;  two  units  are  then  used 
in  the  reactions. 

Against  these  cells  alexin  can,  in  each  case,  be  titrated  as  follows : 

Alexin  Titration: 

Fresh  Guinea  Pig  Serum  Pipetted  from  Clot 


Red  cells 
5%  emulsion 

Sensitizer 
as  above 
determined 

Guinea  pig 
serum 

Hemolysis 

1 

2 
3 

4 

1  C.  C. 
1  C.  C. 

1  c.  c. 
1  c.  c. 

2  units  (.002) 
2  units  (.002) 
2  units  (.002) 
2  units  (.002) 

0.1      c.  c. 
0.05    c.  c. 
0.025c.  c. 
0.01     c.  c. 

rf 

The  smallest  amount  of  alexin  which  completely  hemolyzes  the 
red  cells  (0.05  in  this  case)  is  the  amount  used.  Since  it  is  easier 
to  measure  larger  volumes  with  accuracy,  the  alexin  is  diluted  1  to 
10  in  salt  solution  before  use.  A  typical  Wassermann  reaction  can 
then  be  carried  out  as  follows : 

done  blood  serum  or  plasma  will  be  injected  with  them  and  the  treated  animal 
will  respond  by  the  formation  not  only  of  hemolysin  but  of  precipitins  for 
the  serum  proteins  as  well.  When  a  subsequent  hemolytic  test  is  carried  out, 
a  precipitin  reaction  between  the  precipitin  in  the  antiserum  and  serum  ad- 
hering to  the  corpuscles  will  follow,  and  this,  as  we  have  seen,  will  fix  alexin, 
obscuring  other  reactions  which  may  be  under  observation. 


PRACTICAL    APPLICATIONS    OF    METHOD 


207 


SCHEME  FOR  WASSERMANN  TEST 
ADAPTED  TO  ORIGINAL  WASSERMANN  SYSTEM  AFTER  SCHEME  OF  NOGUCHI 


Test  with  known 

Test  with  known 

Test  without  serum 

Test  with 

positive  syphilitic 

negative  normal 

to  control  efficiency 

unknown  serum 

serum 

serum 

of  hemolytic  system 

Serum  2  c.  c. 

«    Serum  .2  c.  c. 

Serum  .2  c.  c. 

d 

i 

o                      _j_ 

_i_ 

*l 

O  Complement 

„-  O  Complement 

O  Complement 

O  Complement 

2  c3 

.1  c.  c. 

.1  c.  c. 

.1  c.  c. 

.1  c.  c. 

11 

4- 

-}_ 

-|_ 

&| 

Salt  sol. 

1        Salt  sol. 

Salt  sol. 

Salt  sol. 

•I 

3  c.  c. 

3  c.  c. 

3  c.  c. 

3  c.  c. 

2. 

4. 

6. 

8. 

Serum  .2  c.  c. 

Serum  .2  c.  c. 

Serum  .2  c.  c. 

O  Complement 

O  Complement 

O  Complement 

O  Complement 

.1  c.  c. 

d             .1  C.  C. 

.1  c.  c. 

.1  c.  c. 

j_ 

d                     -U 

-j- 

-|- 

d 
*  & 

Antigen 

™        Antigen 

Antigen 

Antigen 

ef 

(required  amount 

| 

£* 

in  1  c.  c.  salt  sol.) 

| 

2£ 
fc-g 

Salt  sol. 

3             + 

Salt  sol. 

Salt  sol. 

Salt  sol. 

2  c.  c. 

2c.  c. 

2c.  c. 

2c.  c. 

1. 

3. 

5. 

7. 

O  =  Test  tube. 

Place  in  water  bath  at  40°  C.  for  one  hour,  then  add  to  all  tubes  red  blood 
cells  and  amboceptor.  These  are  previously  mixed  so  that  2  c.  c.  contains  the 
equivalents  of  1  c.  c.  of  a  5  per  cent,  emulsion  of  sheep  corpuscles  and  2  units  of 
amboceptor.  Again  expose  to  40°  C.  If  the  serum  tested  is  positive,  tubes  1 
and  3  should  show  no  hemolysis,  all  the  other  tubes  showing  complete  hemolysis 
in  one  hour. 

Since  many  human  sera  normally  contain  small  amounts  of  antisheep 
sensitizer,  it  is  the  habit  of  many  workers  to  add  the  sheep  corpuscles,  without 
the  sensitizer  or  amboceptor,  and  incubate  for  a  half -hour.  If,  at  the  end  of  this 
time,  no  hemolysis  has  occurred  either  in  the  front  or  the  back  ^  row,  then 
amboceptor  may  be  added.  This  technique  avoids  the  possible  error  introduced 
by  an  excess  of  amboceptor,  a  condition  which  easily  occurs  when  any  large 
amount  is  normally  present  in  the  serum  and  in  addition  to  this  2  units  are 
added  as  in  the  test  described  above. 

The  above  represents  the  typical  "Wassermann7'  as  at  present 
carried  out  in  most  laboratories.  It  may  be  carried  out  just  as  well 
and  with  greater  economy  of  material  by  using  one-half  the  amounts 
throughout.  It  is  evident  that  the  performance  of  the  reaction  calls 
for  experience  of  serum  technique,  and  knowledge  of  such  reactions, 
so  that  fortuitous  irregularities  may  be  intelligently  controlled.  It  is 
our  opinion  that  the  performance  of  routine  Wassermann  tests  by 
workers  without  a  thorough  knowledge  of  the  fundamental  facts  of 


208  INFECTION    AND    RESISTANCE 

serum  phenomena  is  worse  than  useless  in  that  insufficient  attention 
to  special  conditions  and  to  details  may  easily  result  in  a  positive  re- 
action when  syphilis  is  not  present,  and  vice  versa. 

Recently  Archibald  McNeil  and  others  have  exposed  the  mix- 
tures of  complement,  antigen,  and  patients'  serum  at  refrigerator 
temperature  for  a  number  of  hours  instead  of  in  the  water  bath  or 
thermostat  at  37.5°  C.,  before  adding  the  sensitized  cells.  It  is  a 
curious  fact,  which  has  not  yet  been  satisfactorily  explained,  that 
such  a  procedure  increases  the  delicacy  of  the  reaction.  It  may  be 
that,  when  the  tubes  containing  the  antigen,  patient's  serum,  and 
alexin  are  left  at  incubator  temperature,  partial  alexin  fixation  only 
can  take  place  during  the  brief  period  of  30  minutes  to  one  hour, 
which  is  usually  employed.  More  prolonged  exposure  at  this  tempera- 
ture would  not  be  advisable  on  account  of  deterioration  of  the  alexin. 
On  the  other  hand,  at  ordinary  ice-box  temperatures  of  about  8°  to 
10°  C.,  the  exposure  can  be  continued  for  as  long  as  10  hours  without 
extensive  complement  deterioration,  and  meanwhile  more  complete 
fixation  can  occur.  This,  however,  is  a  surmise.  The  actual  condi- 
tions are  not  clear.  As  a  matter  of  fact  in  our  laboratory  Dr.  Otten- 
berg,  in  120  cases  so  far  done  in  parallel  series,  one  being  exposed 
for  fixation  for  30  minutes  at  37.5°  C.,  the  other  at  8°  to  10°  C.  for 
three  hours,  found  discrepancies  between  the  two  methods  in  15 
cases.  In  all  of  these,  positive  reactions  were  obtained  by  the  ice-box 
method,  whereas  by  the  water  bath  method  the  results  were  negative. 
Of  these  cases  7  were  clearly  unquestionable  syphilitics,  two  were 
treated  syphilis,  and  four  were  probably  syphilitic. 

Many  modifications  of  the  Wassermann  test  have  been  suggested. 
Probably  the  most  important  is  that  of  Noguchi.  The  chief  justifi- 
sation  for  this  modification  is  the  fact  that  many  normal  human  sera 
contain  hemolysins  for  sheep  corpuscles.  For  this  reason  many 
workers  carry  out  the  ordinary  Wassermann  technique  without  add- 
ing antisheep  sensitizer  or  amboceptor  until  they  have  first  observed 
whether  or  not  the  tested  serum  (in  the  "back  row,"  without  antigen) 
will  not  hemolyze  the  corpuscles  without  such  an  addition,  adding 
the  sensitizer  only  when  this  does  not  take  place.  This  is  advisable 
since  the  presence  of  any  considerable  amount  of  normal  antisheep 
sensitizer  in  the  human  serum  which  is  being  examined  (if  added  to 
the  amount  used  in  the  ordinary  reaction,  2  units),  may  so  increase 
the  total  quantity  that  hemolysis  wrill  result  even  after  most  of  the 
alexin  has  been  fixed.  Noguchi  excludes  this  uncertainty  by  avoid- 
ing the  use  of  the  "sheep  cell-antisheep  sensitizer"  system  entirely, 
substituting  a  hemolytic  complex  consisting  of  human  cells  and  anti- 
human  sensitizer,  produced  by  injecting  washed  human  corpuscles 
into  rabbits. 

His  technique  may  be  best  illustrated  in  the  following  tabula- 
tion: 


PRACTICAL    APPLICATIONS    OF    METHOD         209 

Reagents 

1.  Sensitizer  prepared  by  injecting  washed  human  blood  corpuscles  into 
rabbits. 

2.  1  per  cent,  emulsion  of  washed  human  blood  cells. 

3.  Alexin — fresh  guinea  pig  serum  diluted  with  one  and  one-half  volumes 
of  salt  solution,  40  per  cent. 

The  reaction  is  performed  in  the  following  way : 
Noguchi's  Method  of  Complement  Fixation  for  the  Serum  Diagnosis  of  Syphilis 


Rear  row  | 

Set  for  diagnosis 
Test  with  the  serum  in 
question 

Positive  control  set 
Test  with  a  positive  syphi- 
litic serum 

Negative  control  set 
Test  with  a  normal  serum 

a.  Unknown  ser- 
um, 1  drop* 
b.  Complement, 
2  units 
O  c.  Corpuscle 
susp.,  1  c.  c. 

a.  'Positive  syph. 
serum,  1  drop* 
b.  Complement, 
2  units 
O  c.  Corpuscle  susp., 
1  c.  c. 

a.  "Normal  serum, 
1  drop* 
b.  Complement, 
2  units 
Oc.  Corpuscle 
susp.,  1  c.  c. 

| 
j| 

k 

1 

. 
ll 

°1 
1? 

I" 

1  Incubation  at  37°  C.  for  2  hours  longer, 
then  at  room  temperature. 

|  Front  row 

a.  Unknown  ser- 
um, 1  drop* 
b.  Complement, 
2  units 
O  c.  Corpuscle 
susp.,  1  c.  c. 
+  Antigen 

a.  'Positive  syph. 
serum,  1  drop* 
b.  Complement, 
2  units 
O  c.  Corpuscle  susp., 
1  c.  c. 
+  Antigen 

a.  "Normal  serum, 
1  drop* 
b.  Complement, 
2  units 
Oc.  Corpuscle 
susp.,  1  c.  c. 
-h  Antigen 

*  When  working  with  inactivated  serum  4  drops  (0.08  c.  c.)  should  be  em- 
ployed. With  cerebrospinal  fluid,  0.2  c.  c.  (not  inactivated)  is  used. 

(Taken  from  Noguchi's  " Serum  Diagnosis  of  Syphilis,"  Lippincott,  1910, 
p.  57.) 

Bauer  30  has  introduced  a  modification  in  which  he  utilizes  the 
presence  of  normal  sheep  sensitizer  in  many  human  sera.  He  per- 
forms his  tests  without  the  addition  of  antisheep  sensitizer  at 
first,  adding  this  only  to  those  tubes  in  which  controls  have  shown 
that  no  normal  sensitizer  is  present.  Stern,31  on  the  other  hand, 
utilizes  the  alexin  normally  present  in  human  serum.  The  syphi- 
litic serum  to  be  tested  is,  therefore,  not  inactivated,  and  the 
sheep  cells  are  more  heavily  sensitized  (9  to  12  units).  It  seems 
to  us  that  this  method  is  objectionable  chiefly  because  of  the 
anticomplementary  action  which  develops  in  most  normal  human 
sera  if  kept  for  a  short  time,  and  which  can  be  removed  only  by 
inactivation. 

Other  modifications  of  the  Wassermann  reaction  are  those  of 


30  Bauer.     Semaine  Medicate,  28,  1908. 

31  Stern.    Zeitschr.  f.  Imm.,  Vol.  1,  1909. 


810  INFECTION    AND    RESISTANCE 

Jacobaeus  32  and  of  Wechselman.33  It  seems,  however,  that,  as  the 
reaction  is  gaining  in  importance  in  clinical  diagnosis,  most  labora- 
tories are  adhering  to  the  original  system  used  by  Wassermann  and 
his  associates,  except  for  the  substitution  of  the  non-specific  lipoidal 
antigens  for  the  originally  employed  organ  extracts. 

The  value  of  the  Wassermann  test  in  the  diagnosis  of  the  various 
stages  of  syphilis  is  a  problem  which  can  be  approached  only  by 
careful  statistical  analysis  of  the  results  obtained.  This  has  been 
done  by  various  investigators,  and  some  of  the  results  have  been 
tabulated  in  the  books  of  Noguchi,  of  Boas,  and  of  Mclntosh  and 
Fildes.  The  figures  we  cite  are  those  largely  taken  from  Boas,  as 
summarized  in  F.  C.  Wood's  "Chemical  and  Microscopical  Diag- 
nosis" (D.  Appleton  &  Co.,  1911),  pp.  706  et  seq. 

Primary  syphilis,  974  cases,  56.5  per  cent,  positive. 

The  reaction  may  appear  before  the  primary  sore,  but  this  is 
very  rare.  Usually  it  is  positive  in  from  5  to  6  weeks  after  infection. 
Secondary  syphilis,  2,762  cases,  88  per  cent,  positive.  In  untreated 

cases  they  are  stated  to  be  100  per  cent,  positive. 
Tertiary  syphilis,  830  cases,  80  per  cent,  positive. 
Tabes,  360  cases,  70  per  cent,  positive. 
Dementia  paralytica,  95  to  100  per  cent,  positive. 

The  tabulation  on  the  following  page,  taken  directly  from  Boas, 
will  give  a  comprehensive  summary  of  this  phase  of  the  problem. 

Since  the  reaction  is  not  a  specific  antigen-antibody  union  but 
depends  on  some  substance  liberated  or  produced  by  reason  of  the 
syphilitic  injection,  it  is  not  out  of  question  that  other  infections 
may  give  rise  to  a  "positive  Wassermann."  And  this,  indeed,  is 
the  case.  It  was  claimed  for  a  time  that  a  positive  reaction  may 
be  obtained  in  tuberculosis,  but  this  has  been  refuted  by  subsequent 
experience,  and  the  earlier  positive  results  probably  depended  upon 
faulty  technique.  There  can  be  little  doubt,  however,  that  occasional 
positive  reactions  are  obtained  in  cases  of  leprosy,  scarlet  fever, 
malaria,  and  trypanosoma  infections. 

The  spinal  fluid  may  be  used  instead  of  the  blood  serum  in  cases 
of  syphilis  of  the  central  nervous  system,  but  even  here,  as  Citron  34 
has  shown,  the  results  with  blood  serum  are  more  frequently  positive 
than  those  done  with  the  spinal  fluid  itself.  In  isolated  cases  posi- 
tive reactions  have  been  obtained  with  ascitic  fluids,  pleural  and 
pericardial  exudates.  Bab  35  reports  a  case  of  positive  reaction  in 

32  Jaeobaeus.     Zeitschr.  f.  Imm.,  Vol.  8,  1911. 

33  Wechselmann.     Zeitschr.  f.  Imm.,  Vol.  3,  1909. 

84  Citron.    Deut.  med.  Woch.,  1907,  No.  29,  p.  1165. 

85  Bab.  Munch,  med.  Woch.,  Vol.  46,  1907. 


PRACTICAL    APPLICATIONS    OF    METHOD 

Table  Compiled  by  Boas,  loc.  tit.,  p.  138. 


Stage  of  disease 

Number  of 
cases 

Positive 
reaction 

Negative 
reaction 

Control  cases  (not  syphilitic)  

1,064 

1 

1  063 

Induration  

76 

(scarlatina) 
56 

20 

Secondary 
Early  untreated  

269 

269 

o 

Recurrent  after  treatment  

199 

187 

12 

Tertiary 
No  treatment  of  early  tertiary  mani- 
festations   

63 

63 

o 

Treatment  

20 

16 

4 

Latent  syphilis 
Within  3  yrs.  after  infection  
After  3  yrs  

243 
111 

89 
44 

154 

87 

Tabes 
Untreated  

17 

17 

o 

Treated   . 

26 

11 

15 

Dementia  paralytica 
Serum 

139 

139 

o 

Spinal  fluid  

67 

61 

6 

Congenital 
With  symptoms 

54 

54 

o 

Without  symptoms   

10 

7 

3 

the  milk  of  a  syphilitic  mother.  Serum  obtained  at  autopsy  is  not 
suitable  for  the  reaction,  since  this,  for  unknown  reasons,  may  often 
give  a  positive  reaction  in  non-syphilitic  cases. 


COMPLEMENT   OK   ALEXIN   FIXATION   AS   A   METHOD    OF 

DETERMINING  THE   NATURE   OF  UNKNOWN 

PROTEIN 

FORENSIC  ALEXIN  FIXATION  TESTS 

Our  preliminary  discussions  of  the  principles  underlying  alexin 
or  complement  fixation  have  revealed  that  alexin  is  bound  not  only 
by  sensitized  cells  but  also  by  the  specific  precipitates  formed  when 
an  unformed  protein  antigen  is  mixed  with  its  specific  antiserum. 
This  discovery,  made  by  Gengou,  was  attributed  by  him,  it  will  be 
remembered,  to  the  presence  of  "albuminolysins,"  or  protein  sensi- 
tizers,  antibodies  which  have  been  by  many  observers  regarded  as 
separate  from  the  precipitins,  but  which  we  believe,  for  stated  rea- 
sons (see  p.  193),  to  be  very  probably  identical  with  the  precipitating 
antibodies  or  precipitins.  However  this  may  be,  when  a  dissolved 
antigen  is  mixed  with  its  antiserum  alexin  fixation  is  exerted  by  the 


INFECTION    AND    RESISTANCE 

complex,  and  this,  even  when  the  reacting  quantities,  antigen  and 
antibody,  are  so  small  that  visible  precipitation  will  not  take  place. 
For  this  reason,  it  is  plain,  it  should  be  possible  by  means  of  com- 
plement fixation  to  detect  amounts  of  a  foreign  protein  too  small  to 
be  demonstrable  by  direct  precipitation  with  an  antiserum. 

The  method  has,  therefore,  been  suggested  chiefly  by  Neisser  and 
Sachs  36  for  the  forensic  determination  of  unknown  proteins,  as  an 
adjuvant  to,  and  improvement  upon,  the  forensic  precipitin  test.  Our 
discussion  of  the  principles  involved  in  the  introductory  paragraphs 
of  this  chapter  will  render  unnecessary  an  extensive  discussion  of 
the  reasoning  upon  which  this  reaction  is  based.  It  is  well  to  remind 
the  reader,  however,  of  the  facts  which  we  have  discussed  regarding 
the  quantitative  proportions  which  govern  the  occurrence  of  precipi- 
tation when  an  antigen,  say  human  serum,  is  mixed  with  its  antibody, 
in  this  case  antihuman  rabbit  serum.  The  actual  precipitation  may 
be  absent  either  when  an  excess  of  the  antigen  is  used  or  when  the 
antigen  is  present  in  too  small  a  quantity.  Thus  a  given  quantity 
of  the  antiserum  may  precipitate  strongly  dilutions  of  the  antigen 
ranging  from  1-50  to  1-10,000.  No  precipitation  or,  at  least,  a  very 
slight  one  only  may  occur  when  concentrations  stronger  than  1-50 
are  used  and  when  the  dilution  is  greater  than  1-10,000.  Neverthe- 
less, in  both  cases,  alexin  fixation  may  be  exerted  by  the  complex 
although  no  precipitation  takes  place.  As  Gay  37  has  shown,  com- 
plement fixation  may  be  exerted  even  when  a  formed  precipitate  has 
been  redispersed  by  the  subsequent  addition  of  more  antigen.  The 
importance  of  the  forensic  reaction  of  Neisser  and  Sachs,  however, 
lies  chiefly  in  its  application  to  the  detection  of  quantities  of  un- 
known protein  too  small  to  be  detected  by  precipitin  reactions. 

The  tests  are  carried  out  by  mixing  a  dilution  of  unknown  pro- 
tein with  given  quantities  of  antiserum,  adding  small  quantities  of 
alexin  (quantities  determined  best  by  previous  alexin  titration  as 
indicated  in  our  section  on  the  Wassermann  reaction)  ;  these  reagents 
are  left  together  for  a  given  time  at  37.5°  C.,  and  then  sensitized  cells 
are  added  to  determine  whether  or  not  the  alexin  has  been  bound. 

The  table  on  the  following  page,  taken  directly  from  the  article 
of  Neisser  and  Sachs,  loc.  cit.,  will  not  only  illustrate  the  method  of 
carrying  out  the  reactions  but  will  also  give  an  indication  of  their 
extreme  delicacy. 

It  will  be  seen  that  0.00001  c.  c.  of  the  normal  human  serum  still 
gave  almost  complete  complement  fixation  of  0.05  c.  c.  of  comple- 
ment in  the  presence  of  0.1  c.  c.  of  the  antihuman  serum.  The  table 
also  shows  that  this  reaction  follows  a  general  law  of  relative  specific- 
ity so  often  noted  in  other  reactions,  namely  that,  of  all  the  ani- 
mals tested,  the  serum  of  monkeys  alone  gave  reactions  with  the 

86  Neisser  and  Sachs.     Berl  kl  Woch.,  Vol.  42,  No.  44,  1905,  p.  1388. 

87  Gay.    Univ.  of  Cal.,  "Publications  in  Pathology,"  1912. 


PRACTICAL    APPLICATIONS    OF    METHOD 


Table  Taken  from  Neisser  and  Sachs,  loc.  cit.,  p.  1388 

0.1  human  antiserum  +  0.05  complement  and  variable  amounts  of  different 
normal  sera  (brought  to  1  c.  c.  volume  with  salt  solution) ;  the  mixtures  kept 
1  hour  at  room  temperature.  Then  added  1  c.  c.  5  per  cent,  washed  beef 
blood  +  0.0015  c.  c.  amboceptor  and  left  1-2  hours  at  37°  C. 

The  results  are  as  follows : 


Amounts 


Hemolysis  on  addition  of  serum  of: 


normal 
serum 

Man 

Monkey 

Rat 

Pig 

Goat 

Rabbit 

Ox 

Horse 

0.01 

0 

0 

* 

- 

• 

- 

- 

0.001 

0 

0 

0.0001 

0 

moderate 

com- 

com- 

com- 

com- 

com- 

com- 

0.00001 

slight 

complete 

"  plete 

"  plete 

plete 

"  plete 

*"  plete 

^  plete 

0.000001 

complete 

complete 

0 

complete 

complete 

j 

J 

„ 

„ 

human  antiserum;  and  this  in  quantities  as  small  as  0.001  cubic 
centimeter. 

The  forensic  complement  fixation  reaction  of  Neisser  and  Sachs 
is  both  theoretically  and  practically  valid.  Its  extensive  use  in  many 
investigations  for  theoretical  purposes  has  well  established  its  reli- 
ability. However,  it  is  more  complicated  and  requires  much  more 
experimental  training  and  care  than  does  the  simpler  precipitin  test, 
and  it  will  rarely  occur  that  an  unknown  protein  is  available  in 
quantities  too  small  to  permit  of  successful  precipitation. 


THE   USE   OF   COMPLEMENT   FIXATION   TESTS   IN   THE    DIAG- 
NOSIS   OF    MALIGNANT    NEOPLASMS 

A  great  many  attempts  have  been  made  to  establish  a  method  of 
complement  fixation  by  which  a  diagnosis  of  malignant  tumors  could 
be  made.  It  had  been  hoped  that  the  substance  of  malignant  tumors 
might  contain  a  form  of  protein  or  protein  lipoid  combination  which 
might  represent  substances  specific  for  such  tumors,  and  might  there- 
fore functionate  as  a  specific  antigen.  On  this  basis  it  might  be 
possible  that  the  serum  of  tumor  patients  would  contain  a  specific 
antibody  which  could  react  with  a  specific  antigen  in  tumor  extracts,, 
with  the  resulting  formation  of  an  alexin-fixing  complex. 

No  experimental  facts  have  so  far  justified  our  assumption  of 
the  presence  of  either  specific  antigen  in  tumor  extracts,  or  that  of 
a  specific  antibody  in  the  serum  of  such  patients.  However,  we  have 
seen  that  the  Wassermann  reaction  is  a  perfectly  useful  clinically 
diagnostic  method,  in  spite  of  the  fact  that  the  antigen  need  not  be 
specific,  and  the  purely  empirical  basis  on  which  the  syphilis  reac- 


INFECTION    AND    RESISTANCE 

tion  is  at  present  based  has  justified  extensive  attempts  to  establish 
an  analogous  empirical  method  for  tumor  diagnosis. 

The  literature  on  this  question  is  confusing.  A  number  of  ob- 
servers using  antigens  variously  prepared  from  tumor  substances 
have  reported  favorable  results.  Simon  and  Thomas  38  report  many 
positive  reactions,  as  do  Sanpietro  and  Tesa,39  and  a  number  of 
others.  Clowes  40  has  carried  out  a  reaction  on  sarcoma  rats  and 
obtained  positive  reactions  in  animals  in  which  the  tumors  were  small, 
negative  ones  when  the  tumor  had  grown  to  a  large  size.  Ranzi, 
on  the  other  hand,  obtained  negative  results  throughout.  Ranzi  41 
found  that  normal  serum  would  often  give  complement  fixation  with 
carcinoma  extracts,  also  that  many  tumor  extracts  and  sera  of  tumor 
patients  inhibited  complement  by  themselves.  The  reactions  were  so 
irregular  that  he  assumed  them  to  be  without  value.  Recently  the 
subject  has  been  very  thoroughly  investigated  by  v.  Dungern.42 

Von  Dungern  claims  to  have  finally  evolved  a  method  by  which 
the  diagnosis  of  malignant  disease  can  be  made  with  reasonable 
accuracy.  Like  the  Wassermann  reaction  his  method  is  purely  em- 
pirical. He  admits  that  probably  it  is  not  a  specific  antibody  deter- 
mination and  depends  rather  upon  the  presence  of  pathological  prod- 
ucts of  metabolism  in  the  sera  of  tumor  patients.  The  reliability  of 
his  method  depends  upon  the  observation  of  a  number  of  details 
which  he  has  determined  empirically. 

He  obtains  his  antigen  in  a  purely  non-specific  manner,  using, 
as  just  stated,  for  this  reaction  acetone  extracts  of  human  blood  cells. 
We  take  the  description  of  the  reaction  entirely  from  his  own  article 
in  "Weichhardt's  Jahresbericht."  The  antigen  is  prepared  in  the  fol- 
lowing way:  Blood  is  taken  from  a  vein,  preferably  from  a  para- 
lytic patient,  since  v.  Dungern  claims  that  individual  specimens  of 
blood  vary,  and  he  hafe  had  the  best  results  with  that  of  paralytic 
cases.  Clotting  is  prevented  by  sodium  oxalate  and  the  blood  cells 
are  thoroughly  washed  in  the  centrifuge.  To  the  sediment  are  added 
19  volumes  of  pure  acetone  (Merck).  This  is  allowed  to  stand  three 
days  at  room  temperature  and  is  occasionally  shaken  during  this 
time.  It  is  then  filtered,  the  acetone  evaporated  in  the  incubator  at 
37°  C.,  and  the  residue  taken  up  in  96  per  cent,  alcohol.  This  alco- 
holic extract  is  diluted  before  use  with  four  parts  of  salt  solution.  Of 
this  final  preparation  0.8  c.  c.  is  used  in  the  individual  test. 

Particular  precautions  must  also  be  taken  in  the  handling  of  the 
serum  of  the  patient.  In  his  earliest  tests  v.  Dungern  determined 

38  Simon  and  Thomas.    Journ.  Exp.  Med.,  Vol.  10,  1908. 

39  Sanpietro  and  Tesa.     Cited  from  v.  Dungern  in  "Weichhardt's  Jahres- 
bericht," etc.,  Vol.  8,  1912,  p.  163.  ' 

40  Clowes.    Journ.  A.  M.  A.,  1909,  Vol.  52. 

41  Ranzi.     Wien.  kl.  Woch.,  1906,  p.  1552. 

42  V.  Dungern.    Munch,  med.  Woch.,  Nos.  2,  20,  52,  1912;  Berl.  kl.  Woch., 
1913,  "Weichhardt's  Jahresbericht,"  Vol.  8,  1912,  p.  163. 


PRACTICAL    APPLICATIONS    OF    METHOD         215 


that  the  inactivation  of  the  tumor  sera  greatly  diminishes  their  spe- 
cific fixation  properties,  and  for  this  reason  he  at  first  advised  that 
the  serum  be  used  unheated.  He  has  found  recently  that  the  best 
results  are  obtained  when  the  serum  is  heated  to  54°  C.,  together  with 
a  little  sodium  hydrate  solution.  He  handles  'the  blood  in  the  fol- 
lowing way:  After  being  taken  from  the  patient  it  is  allowed  to 
stand  1  to  2  days  in  the  refrigerator;  just  before  use  he  adds  two  parts 
of  an  -^5-  NaOH  solution  with  one  part  of  serum  and  heats  it  for  half 
an  hour  at  54°  C.  As  it  is  important  that  the  sodium  hydrate  should 
contain  no  sodium  carbonate,  he  advises  the  use  of  the  Kahlbaum 
preparation.  In  setting  up  the  test  he  uses  graded  quantities  of  the 
mixture  corresponding  to  0.2,  0.1,  0.05,  and  0.025  c.  c.  of  the  original 
serum.  To  each  of  these  quantities  he  adds  the  stated  quantity,  0.8 
c.  c.  antigen  preparation  described  above,  and  the  0.05  guinea  pig 
complement.  Controls  must  be  set  up  with  the  antigen  alone  and 
with  the  patient's  serum  alone  to  prevent  error  from  independent 
fixation  by  these  substances.  These  reactions  are  allowed  to  stand 
three  hours  at  room  temperature,  and  then  one  cubic  centimeter  of 
a  5  per  cent,  solution  of  beef  blood  sensitized  with  two  units  of  hemo- 
lytic  serum  is  added  (as  in  the  Wassermann  reaction).  It  is  im- 
portant to  use  a  strongly  sensitizing  serum,  so  that  not  too  much  of 
the  hemolytic  rabbit  serum  must  be  added  to  the  tubes.  Experi- 
ments done  in  this  way  with  normal  sera  usually  result  in  complete 
hemolysis  within  one  hour,  although  in  certain  other  diseases,  i.  e., 
tuberculosis  and  syphilis,  slight  inhibition  may  result.  However, 
fixation  with  the  patient's  serum  in  quantities  of  0.1  c.  c.  or  less  is,  ac- 
cording to  von  Dungern,  fairly  specific  for  malignant  tumors,  since 
normal  sera  treated  in  the  way  described  usually  do  not  cause  fixa- 
tion in  quantities  of  less  than  0.2  c.  c.  and,  in  syphilis  and  tubercu- 
losis, if  fixation  is  at  all  present,  it  is  usually  not  evident  in  quanti- 
ties less  than  0.1  c.  c. 

With  a  reaction  so  carried  out  von  Dungern  has  examined  244 
cases.     The  following  tabulation  states  his  results: 


Malignant  tumor  of 

No.  of  cases 

Reaction  positive* 

Pharynx  

3 

3 

Esophagus       

6 

6 

Stomach  

15 

11 

Rectum  

14 

10 

Larynx  

2 

2 

Tongue  

5 

5 

Bladder  

1 

1 

Breast  

22 

22 

Uterus  

10 

10 

Skin  

8 

7 

Ethmoid  bone  

1 

1 

Upper  maxilla  .  . 

1 

1 

*  Taken  from  von  Dungern,  "Weichhardt's  Jahreabericht,"  Vol.  8,  1912,  p.  174. 


216  INFECTION    AND    RESISTANCE 

We  report  von  Dungern's  results  exactly  as  lie  states  them  in 
his  last  summary,  since  his  well-known  experimental  ability  necessi- 
tates serious  consideration  of  all  of  his  work.  We  may  say,  however, 
that  a  survey  of  the  entire  literature  of  complement  fixation  in  the 
diagnosis  of  malignant  tumors  does  not  yet  justify  our  acceptation 
of  this  method  as  of  anything  like  the  established  value  which  the 
similar  method  has  attained  in  syphilis. 

COMPLEMENT  FIXATION  IN  GLANDEES 

The  diagnosis  of  glanders  by  the  mallein  test  and  by  agglutina- 
tion has  been  recently  reenforced  by  the  method  of  complement  fixa- 
tion. In  carrying  out  these  tests  the  method  of  preparation  of  the 
antigen  is  of  the  greatest  importance.  The  directions  which  we  give 
are  those  employed  in  the  Diagnostic  Laboratory  of  the  New  York 
Department  of  Health,  under  the  immediate  supervision  of  Dr. 
McNeil  and  Miss  Olmstead,  from  whom  we  have  our  information. 

The  particular  strain  of  glanders  bacilli  employed  seems  to  be  of 
little  importance.  The  organisms  are  grown  on  1.6  per  cent,  acid 
glycerin-potato-agar.  This  stock  culture  is  transplanted  every  other 
day.  From  it  cultures  are  planted  upon  salt-free  veal  peptone  agar. 
It  is  of  the  greatest  importance  that  this  medium  shall  be  neutral  to 
phenolphthalein.  After  twenty-four  hours  in  the  incubator  the 
growth  is  washed  off  with  distilled  water,  which  also  should  be  neu- 
tral, and  the  emulsion  heated  for  from  four  to  six  hours  at  80°  C.  in 
a  water  bath.  It  is  then  filtered  through  a  Buchner  filter  simply  to 
facilitate  subsequent  filtration  through  a  Berkefeld  "N"  or  "V" 
filter.  After  filtration  this  antigen  is  again  sterilized  for  one  hour 
at  80°  C.  and  then  on  two  successive  days  at  56°  C.  for  one-half  hour. 

The  fixation  tests  carried  out  with  these  antigens  have  yielded 
excellent  results  as  reported  by  Dr.  McNeil 43  at  the  New  York 
Serological  Society. 

It  is  unnecessary  to  give  further  directions  as  to  the  technique  of 
this  reaction,  since  it  is  simply  that  of  complement  fixations  in  gen- 
eral, the  chief  difficulty  being  that  of  antigen  preparation. 

COMPLEMENT  FIXATION  IN  GONOEEHEAL  INFECTIONS 

There  are  certain  conditions  following  gonococcus  infection  of 
the  genito-urinary  tract  which  are  not  easily  distinguished  from  a 
number  of  other  tests  unless  the  organisms  can  be  cultivated  or  a 
specific  serum  reaction  can  be  applied.  Most  important  of  these  are 
gonorrheal  rheumatism,  salpingitis,  and  endocarditis.  Complement 
fixation  with  the  sera  of  such  patients,  and  an  antigen  produced  from 

43  McNeil,  Archibald.    N.  Y.  Serological  Soc.  Meeting,  April  4,  1914. 


PRACTICAL  APPLICATIONS  OF  METHOD 

gonococci,  has  been  employed  by  many  observers  during  recent  years, 
and  promises  to  be  of  great  value. 

Here,  too,  the  production  of  the  antigen  is  the  only  feature  of 
the  reaction  which  has  offered  difficulties.  Since  the  researches  of 
Torrey  have  shown  that  not  all  races  of  gonococcus  are  antigenically 
alike,  it  seems  necessary  to  produce  a  polyvalent  antigen.  At  the 
Xew  York  Department  of  Health  at  present  the  antigen  is  prepared 
by  using  the  ten  Torrey  strains.  Stock  cultures  are  carried  on  neu- 
tral veal  agar  and  cultures  are  planted  upon  salt-free  veal  agar. 
Twenty-four-hour  growths,  washed  off  in  neutral  distilled  water,  are 
kept  in  a  water  bath  at  56°  C.  for  two  hours,  and  then  filtered 
through,  first,  a  Buchner  and  then  a  Berkefeld  filter.  They  are  then 
sterilized  for  one  hour. 


COMPLEMENT  FIXATION  IN  TUBERCULOSIS 

Attempts  to  apply  complement  fixation  to  the  diagnosis  of  tuber- 
culosis have  been  made  by  as  many  as  thirty  or  more  investigators 
with  varying  results.  Wassermann  and  his  collaborators  attempted 
it  before  they  succeeded  in  developing  the  Wassermann  reaction  in 
syphilis.  Recently,  intensive  work  has  been  done  on  the  subject  by 
Besredka,44  Petroff,45  Craig,46  Bronfenbrenner,47  and  Miller  and 
Zinsser.48  Results  have  warranted  the  application  of  the  reaction  to 
clinical  tuberculosis,  although  the  actual  degree  of  usefulness  of  the 
reaction  must  still  await  the  multiplication  of  cases  tested.  The  diffi- 
culty has  of  course  consisted  in  the  development  of  a  suitable  antigen. 
The  antigen  of  Petroff  has  consisted  of  a  filtrate  of  potato  broth. 
Besredka  has  used  a  filtrate  of  cultures  made  upon  egg  broth.  Craig 
has  used  Besredka's  antigen  and  suspensions  of  ground  bacilli.  Mil- 
ler and  Zinsser  48  have  employed  an  antigen  made  by  triturating  liv- 
ing and  dead  bacilli  with  crystals  of  table  salt,  then  adding  distilled 
water  to  isotonicity.  With  all  of  the  antigens  favorable  results  have 
been  obtained.  Our  own  results  seem  to  check  up  with  clinical 
diagnoses  in  over  80  per  cent,  of  the  cases  and  so  far  have  appeared 
to  give  practically  no  positive  results  in  negative  cases  and  indicated 
only  active  tuberculosis  and  not  healed  lesions. 

The  reactions  are  carried  out  by  methods  entirely  analogous  to 
those  employed  in  the  Wassermann  reaction,'  but  careful  antigen 
titrations  must  be  done. 

44  Besredka.     Ztschr.  f.  Immunit.,  1914. 

45  Petroff. 

46  Craig-.     Am.  Jour.  Med.  Sc.,  Dec..  1915,  p.  781. 

17  Bronfenbrenner:  Arch.  Int.  Med.,  1914,  No.  6,  786.  Ztschr.  f.  Immunit., 
1914,  XXIII,  2,  21. 

48  Miller  &  Zinsser:  Proc.  Soc.  E.rper.  Biol  and  Med.,  1916,  134.  Jour. 
Lab.  and  Clinical  Med.,  Vol.  I.  p.  817. 


CHAPTER    IX 

THE   PHENOMENON   OF   AGGLUTINATION 

WHEN  bacteria  are  added  to  homologous  immune  serum  there 
occurs  a  peculiar  agglomeration  of  the  individual  micro-organisms 
into  small  clumps.  The  phenomenon  is  so  general  and  so  easily 
observed  that  it  is  not  surprising  that  it  was  noticed  and  reported  by 
a  number  of  workers  during  the  period  of  active  investigation  upon 
serum  reactions  which  preceded  and  followed  the  discovery  of  the 
Pfeiffer  phenomenon.  Thus,  in  the  years  from  1891  to  1895, 
Metchnikoff,1  Charrin  and  Roger,2  Isaeff  and  Ivanoff,3  Washburn,4 
and  several  other  workers  made  this  observation  with  a  variety  of 
bacteria  and  immune  sera.  But  all  of  these  observers  failed  to  follow 
up  or  analyze  the  process  they  incidentally  noticed  in  the  course  of 
other  investigations.  A  thorough  study  of  the  phenomenon  was  not 
made  until  1896,  when  Gruber  and  Durham,5  in  Vienna,  in  the 
course  of  their  studies  upon  bacteriolytic  reactions  with  colon  bacilli 
and  cholera  spirilla,  again  noticed  the  agglutination  of  these  bacteria 
in  homologous  immune  sera,  recognized  the  specificity  of  the  reaction, 
and  called  attention  to  its  apparent  independence  of  other  previously 
studied  serum  phenomena. 

The  process  known  as  agglutination  consists  in  the  following 
train  of  occurrences.  If  we  add  to  an  even  emulsion  of  bacteria  a 
small  amount  of  homologous  immune  serum  the  micro-organisms 
may  be  seen  to  collect  rapidly  in  groups  or  masses,  with  a  resultant 
clearing  of  the  fluid  in  which  they  have  been  suspended.  The  clumps 
of  bacteria  gather  in  flakes  which,  not  unlike  flakes  of  snow, 
sink  to  the  bottom  of  the  test  tube.  The  speed  and  completeness 
with  which  this  phenomenon  occurs  depend,  as  we  shall  see, 
upon  the  agglutinating  strength  and  other  qualities  of  the  serum 
which  is  employed,  but  the  essential  process  of  clumping  is  alike 
in  all  cases. 

There  are  a'  large  number  of  different  methods  by  which  this 

•    1Metchnikoff.    Ann.  de  I'lnst.  Past.,  1892. 

2  Charrin  and  Roger.     C.  E.  de  la  Soc.  de  Biol,  1889. 

3  Isaeff  and  Ivanoff.    Zeitschr.  f.  Hyg.,  Vol.  17,  1894. 

4  Washburn.    Journ.  of  Path,  and  Bact.,  1896,  p.  228. 
6  Gruber  and  Durham.    Munch,  med.  Woch.,  1896. 


THE    PHENOMENON    OF    AGGLUTINATION 


occurrence  can  be  observed,  each  one  particularly  adapted  to  some 
special  purpose  for  which  the  reaction  is  carried  out.  Gruber  and 
Durham,  who  were  investigating  the  properties  of  bacteriolysins 
when  they  observed  agglutination,  naturally  recognized  the  specific 
nature  of  the  reaction  and  proposed  to  make  use  of  it  for  the  purpose 
of  bacterial  differentiation  and  species  determination.  For  this  pur- 
pose, which  has  become  one  of  the  most  important  of  the  practical 
applications  of  the  agglutination  reaction,  the  phenomenon  is  best 
observed  by  the  so-called  "macroscopic  method,"  in  which  a  series 
of  serum  dilutions  are  mixed,  in  small  test  tubes,  with  equal  volumes 
of  emulsions  of  the  bacteria.  Thus,  if  we  wish  to  determine  the 
nature  of  an  unknown  bacillus,  suspected  of  belonging  to  the  typhoid 
bacillus  group,  by  this  meth- 
od, we  may  determine  its  ag- 
glutination in  the  serum  of 
an  animal  immunized  with 
a  known  strain  of  typhoid. 
The  tubes  are  incubated 
after  the  mixtures  have  been 
made,  and  the  agglutination 
which  has  taken  place  in  the 
various  tubes  is  recognized 
by  a  clearing  up  of  the  fluid 
and  the  flaking  of  the  bac- 
teria after  from  one  to  three 
hours.  The  test  tube  method 
has  the  advantage  of  permit- 
ting the  use  of  larger  quanti- 
ties of  reagents  than  can  be 
used  in  the  other  methods 

described  below,  and  therefore  more  exact  quantitative  measurements 
can  be  made. 

Although  this  method  for  the  determination  of  bacteria  has  found 
universal  application,  it  is  probably  most  frequently  employed  at  the 
present  time  for  the  rapid  identification  of  colonies  of  doubtful 
typhoid  or  dysentery,  incident  to  the  isolation  of  these  organisms  for 
stools  by  such  methods  of  plating  as  those  of  Conradi-Drigalski,  of 
Endo,  or  of  Hiss.  The  suspicious  colonies  can  thus  be  fished  directly 
to  an  agar  slant,  and  the  cultures,  when  developed,  emulsified  and 
identified  by  agglutination.  The  advantages  of  such  a  method  for 
the  determination  of  the  biological  interrelationship  of  the  organ- 
isms of  a  given  group,  like,  for  instance,  that  of  the  dysentery  bacilli, 
are  obvious. 

An  ingenious  use  of  this  reaction  was  also  made  by  Shiga  when 
he  determined,  among  various  bacteria  isolated  from  the  stools  of 
dysentery  cases,  the  particular  one  which  was  specifically  aggluti- 


MICEOSCOPIC  AGGLUTINATION. 


220  INFECTION    AND    RESISTANCE 

nated  by  the  patient's  serum,  thus  discovering  the  dysentery  bacillus 
which  bears  his  name. 

Within  a  few  months  after  the  publication  of  Gruber  and  Dur- 
ham's work,  Widal  and,  apparently  independently  of  him,  Griiii- 
baum,6  by  a  process  of  reasoning  the  converse  of  that  detailed  above, 
applied  the  reaction  to  the  diagnosis  of  infectious  disease. 

It  is  obvious  that  a  human  being  or  an  animal  infected  with  a 
given  variety  of  bacteria  may  develop  agglutinating  properties 
against  these  bacteria.  It  is  of  great  value,  therefore,  to  determine 
the  agglutinating  power  of  the  serum  of  a  patient  for  the  bacteria 
which  are  known  to  cause  the  disease  suspected  in  the  particular  case 
in  which  a  diagnosis  is  desired.  This  method  has  become  a  routine 
measure  in  the  early  diagnosis  of  typhoid  fever  under  the  name  of 
"Widal"  or  "Gruber-Widal"  reaction  and,  since  the  quantities  of 
serum  which  can  easily  be  obtained  from  a  patient  are  usually  small, 
it  is  convenient  to  carry  out  the  reaction  by  the  microscopic  method. 
This  consists  in  mixing  serum  and  bacterial  emulsion  in  hang-drop 
preparations  and  observing  them  with  the  microscope.  An  excellent 
method,  also,  is  the  so-called  Proescher  7  method  in  which  serum  and 
bacterial  emulsion  are  mixed  in  small  watch-glasses  or  salt  cellars. 
Proescher  used  this  method  extensively  in  the  study  of  staphylococcus 
agglutinations.  The  mixtures  in  the  salt  cellars  were  set  away  at 
37°  C.  for  two  hours,  and  then  observed  with  a  magnification  of  GO 
to  70  diameters. 

Close  observation  of  the  occurrence  under  the  higher  power  of  a 
microscope  shows  that  the  bacteria,  if  motile,  lose  their  motility,  if 
non-motile  the  Brownian  motion  is  arrested.  They  are  then  rapidly 
gathered  in  small  clumps,  isolated  individuals  between  these  clumps 
being  gradually  drawn  into  them,  until  finally  the  fluid  between  the 
masses  is  entirely  clear.  This  complete  clearing,  of  course,  happens 
only  when  there  is  not  an  excess  of  bacteria,  for,  like  other  serum  re- 
actions, this  phenomenon  is  a  quantitative  one  in  which  definite  pro- 
portions must  be  maintained. 

Clinically  the  most  frequent  use  of  the  agglutination  reaction  is 
in  the  diagnosis  of  typhoid  fever.  The  technique  used  for  this  test 
is,  in  the  large  majority  of  cases,  the  microscopic  hang-drop  method. 
In  Germany  the  Proescher  method  is  sometimes  used,  and  the  micro- 
scopic method  with  dead  organisms,  as  first  introduced  by  Ficker,  is 
also  not  uncommon  at  the  present  day. 

Since  the  serum  of  normal  human  beings  very  often  contains 
moderate  agglutinating  powers  for  the  typhoid  bacillus,  the  diag- 
nostic value  of  the  reaction  in  this  disease  depends  upon  the  elimina- 
tion of  this  error  by  sufficient  dilution.  If  dilutions  of  the  serum  of 
from  1-40  to  1-60  are  used  diagnostic  errors  on  this  point  are 

6  Griinbaum.    Lancet,  1896,  Vol.  2. 

7  Proescher.    Centralbl  f.  Bakt.,  Vol.  34,  1903. 


THE    PHENOMENON    OF    AGGLUTINATION 

avoided,  since  the  normal  agglutinating  power  of  human  beings  is 
never  such  that  typhoid  bacilli  will  be  clumped  by  it  in  these  dilu- 
tions within  one  hour.  Prompt  clumping  in  serum  dilutions  of  1-20 
is  fairly  reliable,  but  does  not  absolutely  exclude  an  unusually  high 
normal  agglutinating  power.  In  carrying  out  tests  clinically  dilu- 
tions of  1-20,  1-40,  and  1-80  are  usually  made  and  observed  for  one 
hour.  From  such  tests  diagnosis  can  be  made  without  danger  of 
error.  In  rare  cases  of  icterus  the  agglutinating  power  for  typhoid 
bacilli  may  be  increased.  Just  what  is  the  cause  of  this  is  not  cer- 
tain ;  Wood  8  reports  cases  in  which  agglutination  of  1-40  was  pres- 
ent with  slight  jaundice  (hepatic  cirrhosis).  On  the  other  hand  he 
has  frequently  failed  to  notice  agglutination  in  other  cases  of  intense 
jaundice.  It  is  not  impossible,  as  Wood  suggests,  that  the  occasional 
presence  of  unusual  agglutinating  power  in  individuals  with  jaun- 
dice has  some  relation  to  the  frequent  persistence  of  typhoid  bacilli 
in  the  gall  bladder. 

Occasionally  it  will  be  noticed  that  dilutions  of  the  patient's 
serum  of  1-5  to  1-20  fail  to  agglutinate,  while  higher  dilutions  will 
give  positive  tests.  This  is  referable  to  the  so-called  "pro-agglu- 
tinoid  zone,"  the  principles  underlying  which  we  shall  discuss  in  an- 
other place. 

The  Widal  test  in  typhoid  cases  rarely  appears  before  the  end 
of  the  first  week,  and,  in  the  majority  of  cases,  is  present  before  the 
end  of  the  second  week.  It  may  proceed  for  months,  although  Wood 
states  that  he  has  seen  it  disappear  at  the  end  of  three  to  six  weeks. 

In  paratyphoid  fever  the  diagnosis  can  often  be  made  by  agglu- 
tination, and  in  dysentery,  as  we  have  seen,  the  fact  that  the  pa- 
tient's blood  agglutinated  the  bacteria  was  one  of  the  important  facts 
utilized  by  Shiga  in  his  discovery  of  the  organism  which  bears  his 
name. 

In  pneumonia  agglutination  of  the  pneumococcus,  isolated  from 
the  patient's  sputum  by  sera  prepared  by  immunization  with  various 
types  of  pneumococci,  has  become  of  considerable  importance  clin- 
ically, since  Neufeld  and  Haendel  and,  in  this  country,  Cole,  Dochez, 
and  Gillespie  have  determined  that  there  are  a  number  of  different 
types  of  this  micro-organism.  The  use  of  pneumococcus  serum  in  the 
disease  will  be  of  value  only  if  a  serum  is  used  which  has  been  pro- 
duced with  an  organism  of  the  same  type  as  the  one  infecting  the 
patient.  Therefore,  determinations  of  the  type  by  highly  potent 
agglutinating  sera  give  a  finger-point  to  the  variety  of  serum  to  be 
used.  Whatever  may  be  the  eventual  outcome  of  the  serum  treat- 
ment in  pneumonia,  no  results  whatever  can  be  expected,  according 
to  our  present  knowledge,  unless  such  determinations  are  made.  The 
technique  of  agglutinations  in  pneumococcus  work  is  facilitated  by 

8  Wood.    "Chemical  and  Microscopical  Diagnosis,"  Appleton  &  Co.,  p.  242. 


222  INFECTION    AND    RESISTANCE 

growing  mass  cultures  of  organisms,  as  advised  by  Hiss,  in  flasks 
of  glucose  broth  containing  1  per  cent,  calcium  carbonate. 

The  same  method  of  growing  micro-organisms  is  useful  in  the 
case  of  streptococcus  agglutinations,  since  the  insoluble  calcium  car- 
bonate, if  thoroughly  shaken,  breaks  the  chains  of  streptococci  and 
thereby  facilitates  judgment  as  to  the  reaction. 

Agglutination  reactions  have  been  of  considerable  usefulness  also 
in  the  diagnosis  of  glanders  in  horses.  The  early  work  on  this  sub- 
ject was  done  chiefly  by  MacFadyean,9  and  the  reaction  has  been 
particularly  studied  by  Wladimiroff.10  Since  the  serum  of  normal 
horses  will  often  agglutinate  glanders  bacilli  in  dilutions  of  as  much 
as  1-500,  Wladimiroff  advises  making  the  positive  diagnosis  on  dilu- 
tions only  higher  than  1-1,000,  since  he  states  that  normal  horses  may 
occasionally  reach  an  agglutination  titre  of  1-1,000.  The  same  writer 
states,  moreover,  that  glanders  bacilli  are  subject  to  great  variations 
in  agglutinability,  and  that  for  this  reason  the  choice  of  a  suitable 
strain  is  of  great  importance. 

The  motility  of  bacteria  has  absolutely  no  relation  to  the  reac- 
tion, and  their  agglutination  is  entirely  passive. 

Some  of  the  earlier  investigators  of  agglutination  associated  the 
reaction  with  alteration  in  the  flagellar  mechanism  of  the  micro- 
organisms. It  is  now  well  known,  however,  that  non-motile,  as  well 
as  motile,  bacteria  are  subject  to  the  phenomenon,  and  that  no  visible 
change  in  the  appearance  or  arrangement  of  flagella  accompanies  the 
clumping.  Although  this  is  the  case,  observation  of  the  motility  of 
such  organisms  as  the  bacillus  of  typhoid  fever,  while  subjected  to 
the  action  of  agglutinating  serum,  may  be  of  great  value  in  aiding 
in  the  determination  of  the  degree  of  completeness  with  which  the 
reaction  is  taking  place. 

Agglutination,  furthermore,  does  not  lead  to  the  death  of  the 
bacteria.  Of  course,  whenever  the  reaction  is  carried  out  in  un- 
heated  serum  the  concomitant  effects  of  the  bactericidal  substances 
bring  about  bacterial  death.  Agglutination  does  not,  however, 
depend  upon  the  cooperation  of  alexin,  and  serum  may  be  inactivated 
-  without  interference  with  its  power  of  agglutination.  In  such  heated 
serum  clumping  takes  place  without  bactericidal  effects,  and,  more 
than  this,  the  bacteria  may  grow,  if  exposed  to  proper  temperature 
conditions,  when  suspended  in  the  serum.  In  fact,  it  is  of  consider- 
able interest  to  carry  out  the  reaction  in  this  way,  for  the  bacteria 
growing  in  agglutinating  serum  form  long  convoluted  threads  and 
skeins  even  when  in  ordinary  culture  they  habitually  occur  as  sep- 
arate individuals.  Thus  colon  bacilli,  typhoid  bacilli,  pneumococci, 
cholera  spirilla,  and  other  organisms,  which  ordinarily  grow  as  free 
single  cells,  or,  at  most,  in  chains  of  two  or  three,  if  kept  in  the 

9  Macfadyean.    Journ.  of  Comparative  Path,  and  Ther.,  Vol.  9,  1896. 
10  Wladimiroff.     "Kolle  u.  Wassermann  Handbuch,"  2d  Ed.,  Vol.  5. 


THE    PHENOMENON    OF    AGGLUTINATION 

incubator  for  ten  to  twelve  hours  together  with  homologous  serum, 
will  grow  in  long,  delicate  chains,  like  those  of  streptococci.  This 
form  of  reaction  has  been  especially  studied  by  Pfaundler,11  who 
attributed  particularly  delicate  specificity  to  it.  However,  the 
"Thread  Reaction"  of  Pfaundler,  as  it  is  sometimes  called,  is  merely 
another  manifestation  of  the  phenomenon  of  agglutination  and  sub- 
ject to  the  same  laws  and  limitations  of  specificity  which  apply  to 
other  methods. 

The  purely  passive  role  played  by  the  bacteria  in  agglutination 
is  best  shown  by  the  fact  that  dead  bacteria,  killed  in  various  ways,"*" 
are  specifically  clumped  just  as  are  the  living  cultures.12  On  this 
fact  depends  the  method  spoken  of  as  "Ficker's  Reaction,"  in  which 
emulsions  of  typhoid  bacilli,  killed  by  formaldehyd  or  carbolic  acid 
(distributed  commercially),  are  agglutinated  in  small  test  tubes  by 
the  serum  of  typhoid  patients.  The  original  method  of  Ficker  is 
said  to  be  a  proprietary  secret ;  however,  a  number  of  other  methods 
which  attain  the  same  purpose  are  in  use  in  various  places.  Volk 13 
describes  the  method  used  in  Vienna,  and  states  that  there  carbolic 
acid  is  used  to  kill  the  cultures.  Similar  to  this  is  the  method  de- 
scribed by  J.  H.  Borden,14  who  proceeds  as  follows : 

The  bacilli  are  grown  on  agar  slants  in  large  tubes  for  24  hours. 
They  are  then  washed  from  the  medium  with  a  sterile  mixture  of 
salt  solution  450  parts,  glycerin  50  parts,  and  95  per  cent,  carbolic 
acid  2.5  parts.  After  this  solution  has  been  kept  a  week  it  becomes 
translucent  and  by  this  time  the  bacilli  are  dead.  The  preparation  is 
then  ready  for  use  and  can  be  kept  a  long  time  in  dark  bottles  in  a 
cool  place.  Borden  very  carefully  controlled  this  bacterial  emulsion 
with  positive  and  negative  typhoid  sera  and  found  it  reliable.  The 
great  advantage  of  all  these  methods,  of  course,  consists  in  the  possi- 
bility of  furnishing  the  general  practitioner  with  materials  for  clini- 
cal agglutination  tests  in  which  the  necessity  of  preserving  and  sus- 
pending living  cultures  is  eliminated. 

The  facts  which  we  have  just  considered  tend  to  show  that  agglu- 
tination is  not  a  vital  phenomenon  15  dependent  in  any  way  upon 
the  living  nature  of  the  bacterial  cell,  but,  like  other  phenomena  of 
antigen-antibody  reactions,  a  purely  chemical  or  physical  process  in 
which  the  substance  of  the  bacterial  cell  enters  specifically  into  rela-  ^ 
tion  with  the  agglutinating  factor  of  the  serum.    In  uniformity  with 
other  analogous  reactions  the  antigenic  substance  is  here  spoken  of  ^ 
as  "agglutinogen,"  the  antibody  as  "agglutinin."  N 

11  Pfaundler.     Wien.  kl.  Woch.,  1898,  and  Centralbl.  f.  Bakt.,  I,  Vol.  23, 
1898. 

12  Bordet.     Ann.  Past.,  Vol.  10,  1896. 

13  Volk.     "Kraus  und  Levaditi  Handbuch,"  Vol.  2. 

14  Borden.    Medical  News,  N.  Y.,  Mar.,  1903. 

15  Bordet.    Ann.  Past.,  Vol.  10,  1896. 


INFECTION    AND    RESISTANCE 

The  agglutinogen,  or  agglutinin-inducing  substance  in  the  bac- 
teria is  apparently  an  inherent  part  of  the  bacterial  protein,  and 
agglutinins  may  be  produced  in  animals  by  injection  not  only  of 
living  and  dead  whole  bacteria,  but  by  bacterial  extracts,  prepared 
in  various  ways.  And,  furthermore,  just  as  the  agglutinins  of  serum 
are  absorbed  out  of  a  serum  by  the  whole  bacteria,  they  may  be  neu- 
tralized by  the  dissolved  bacterial  extracts. 

Just  what  the  nature  of  the  agglutinogen  is  has  been  a  matter 
of  prolonged  controversy,  Pick  16  and  others  claiming  that  it  is  pos- 
sible to  obtain  an  agglutinogen  by  alcohol  precipitation  from  old 
bacterial  cultures  which,  upon  further  purification,  can  be  found  to 
give  none  of  the  usual  protein  reactions  (Biuret,  Millon).  It  is  by 
no  means  certain,  however,  that  Pick's  results  are  correct.  In  fact, 
many  objections  have  been  advanced  against  them,  and  the  accept- 
ance of  an  antigen  of  non-protein  nature  is  so  opposed  to  all  our 
knowledge  regarding  antigens  in  other  cases  that  Pick's  results 
should  not  be  taken  as  final  until  very  careful  revision  of  the  experi- 
mental facts  has  been  carried  out.  That  the  agglutinogen  is,  to  a 
certain  extent,  subject  to  dialysis  has  been  claimed  because  of  ex^ 
periments  in  which  specific  agglutinins  have  appeared  in  the  sera  of 
animals  into  whose  peritoneal  cavities  celloidin  sacs,  filled  with  bac- 
teria, have  been  placed.17 

There  has  been  a  great  deal  of  discussion  regarding  the  possible 
localization  of  the  agglutinogen  of  bacteria  in  the  ectoplasmic  layers 
of  the  cells,  and  especially  in  the  flagellar  substance.  We  have  seen 
that,  as  a  matter  of  fact,  nonmotile  bacteria  are  subject  .to  the  phe- 
nomena of  agglutination  just  as  are  the  motile  forms,  but  numerous 
attempts  were  made  during  the  earlier  stages  of  our  knowledge  of 
this  reaction  to  demonstrate  that  changes  in  ectoplasm  and  flagella 
accompanied  the  actual  agglutination.  Gruber18  himself  held  the 
opinion  for  a  time  that  the  clumping  was  due  to  an  eotoplasmic 
swelling  which  rendered  the  bacteria  sticky,  causing  them  to  hold 
together  after  chance  approximation.  He  soon  gave  up  this  idea 
himself,  but  a  similar  theory  was  for  some  time  upheld  by  ISTicolle  19 
and  others. 

Malvoz  20  in  1897  devised  an  ingenious  method  by  which  he  be- 
lieved that  he  could  show  that  the  agglutination  of  bacteria  depended! 
upon  their  ectoplasmic  substances.  He  passed  the  typhoid  emulsion 
through  Chamberland  filters  and,  when  the  bacilli  had  been  caught 

16  Pick.     "Hof meister's  Beitrage,"  1901,  1902. 

17  This  would  be  in  keeping  with  Pick's  work  just  referred  to,  and  should 
be  subjected  to  the  same  criticism  before  final  acceptance.     For  a  more  de- 
tailed discussion  of  these  conditions  the  reader  is  referred  to  the  article  by 
Paltauf,  "Kolle  u.  Wassermann  Handbuch,"  Vol.  4,  part  1. 

18  Gruber.    Munch,  med.  Woch.,  1896. 

19  Nicolle.    Ann.  de  I'Inst.  Past.,  1898. 

20  Malvoz.    Ann.  de  I'Inst.  Past.,  Vol.  11,  1897. 


THE    PHENOMENON    OF    AGGLUTINATION       225 

upon  the  filters,  he  subjected  them  to  prolonged  washing.  The  ba- 
cilli, now  removed  from  the  filter  by  passing  fluid  through  in  the 
opposite  direction,  were  no  longer  motile  or  agglutinable  either 
by  formalin,  safranin,  or  other  chemical  agents,  nor  by  agglutinating 
sera.  Dineur,21  repeating  the  experiments  of  Malvoz,  came  to  the 
same  conclusions.  He  decided  that  in  agglutination  the  bacteria 
became  entangled  with  each  other  by  means  of  the  flagella.  Harri- 
son,22 in  later  studies  working  under  Tavel,  attempted  to  dissolve 
out  the  ectoplasmic  layers  of  bacteria  with  pyocyanase,  and  from  his 
experiments  also  came  to  the  conclusion  that  the  agglutinogen  was 
contained  in  the  external  layers.  Similar  results  were  obtained  by  De 
Eossi.23 

Further  studies  on  the  same  problem  are  those  of  Smith  and 
Reagh.24  These  investigators  worked  with  two  strains  of  bacilli, 
both  of  which  they  regarded  as  belonging  to  the  hog-cholera  group, 
though  the  one  was  motile  and  the  other  nonmotile.  When  rabbits 
were  immunized  with  the  nonmotile  bacillus  an  agglutinin  was  ob- 
tained which  acted  upon  this  bacillus  differently  and  less  powerfully 
than  did  the  agglutinin  produced  with  the  motile  one.  Contact  witK 
the  nonmotile  bacillus  did  not  deprive  the  serum  produced  with  the 
flagellated  organism  of  the  agglutinins  for  the  latter.  They  con- 
clude that  two  agglutinins  were  involved — one  incited  by  the  ecto- 
plasm and  flagellar  substance,  the  other  by  the  bacterial  cell  body 
proper.  Rehns  as  well  as  Paltauf  have  criticized  these  results  by 
questioning  the  species  identity,  of  the  two  bacilli  employed  in  the 
experiments,  referring  the  phenomenon  to  the  occurrence  of  group 
agglutination.  •» 

As  a  matter  of  fact  our  present  knowledge  of  agglutination  no 
longer  justifies  the  association  of  agglutination  with  flagella,/  !N"on- 
motile  as  well  as  motile  bacteria  are  readily  agglutinated,  and  we 
have  much  evidence  which  will  be  discussed  presently  which  shows 
that  the  agglutination  reaction  is  governed  by  many  of  the  laws 
which  obtain  in  colloidal  flocculations.^  This,  however,  does  not  ex- 
clude the  possibility  that  it  is  the  ectoplasmic  zone  chiefly  which 
takes  part  in  the  reaction..  Furthermore,  loss  of  motility,  which 
always  accompanies  agglutination  when  a  motile  organism  is  under 
observation,  is  an  extremely  valuable  aid  in  guiding  us  in  our  judg- 
ment of  incomplete  reactions.. 

That  changes  may  be  brought  about  in  bacterial  agglutinogen  by 
various  methods  of  treatment  has  been  shown  by  a  number  of  work- 

21  Dineur.    Bull,  tie  I'Acad.  de  Med.  de  Beige,  1898,  cited  from  Smith  and 
Beagh. 

22  Harrison.     Centralbl.  f.  Bakt.,  Vol.  30,  I,  Orig.  1901. 

23  De  Rossi.     Centralbl.  f.  Bakt.,  I,Vols.  36  and  40. 

24  Smith  and  Reagh.    Journ.  of  Med.  Ees.,  Vol.  10,  1903. 


226  INFECTION    AND    RESISTANCE 

ers,  although  the  fundamental  principles  underlying  such  changes 
are  not  at  all  clear. 

Joos  25  was  the  first  to  study  agglutination  with  particular  refer- 
ence to  the  effects  upon  the  reaction  of  heating  both  the  antigen  and 
the  antibody.  On  the  basis  of  extensive  and  complicated  experiments 
upon  the  agglutinin  produced  in  horses  by  immunization  with  heated 
and  unheated  typhoid  bacilli,  he  drew  the  conclusion  that  agglu- 
tinogen  (agglutinin-inducing  substance)  in  bacteria  was  not  a  single 
element  but  consisted  of  at  least  two  definite  parts  of  which  he  speaks 
as  a  and  /3-agglutinogen.  a-agglutinogen  is  a  constituent  of  the 
living  bacteria,  and  is  easily  destroyed  at  60°  to  62°  C.  The  /8-agglu- 
tinogen  is  also  present  in  normal  bacilli,  but  is  more  heat-resistant 
and  will  withstand  60°  to  62°  C.  for  several  hours.  The  injection 
of  living  unheated  bacilli  then  induces  the  formation  of  both  a 
and  /?-agglutinin,  which  have  respectively  a  particular  affinity  for 
a  and  /8-agglutinogens.  The  injection  of  heated  bacilli,  on  the 
other  hand,  induces  the  formation  only  of  /?-agglutinjn  .  and  not  a 
trace  of  a-agglutinin..  The  a-agglutinin  is  considerably  heat- 
resistant,  resisting  60°  to  62°  C.,  whereas  the  /?-agglutinin  loses  its 
agglutinating  property  when  heated  to  60°  C.  The  a-agglutinin 
is  entirely  incapable  of  uniting  with  /?-agglutinogen.y  However, 
/?-agglutinogen  can  combine  or  react  with  both  the  a  and  p  con- 
stituents of  the  bacilli.  For  this  reason  Paltauf  has  spoken  of  agglu- 
tinin produced  with  the  heated  bacteria  as  "umfanglicher."  This 
is  a  point  of  great  interest,  and  if  Joos  is  right  is,  of  course,  of  con- 
siderable practical  importance. 

However  one  may  look  upon  these  experiments,  as  well  as  the 
similar  ones  of  other  workers,  it  seems  established  that  heating  bac- 
teria leaves  them  still  capable  of  inciting  agglutinins  powerfully 
and  rapidly,  perhaps  of  an  "umfanglicher"  nature  than  those  pro- 
duced with  the  native  cells. 

Heating  bacteria  may  also  alter  their  agglutinability.  Thus,  ac- 
cording to  Eisenberg  and  Volk,26  heated  above  65°  C.  the  bacteria 
no  longer  agglutinate  in  the  presence  of  specific  immune  serum,  but 
still  absorb  agglutinin.  Eisenberg  and  Yolk,  therefore,  distinguish 
between  a  heat-sensitive  constituent  of  the  antigen,  which  is 
particularly  associated  with  the  clumping,  whereas  the  thermo- 
stable substance  represents  the  haptophore  or  combining  portion.. 
It  seems  simpler,  in  this  case  also,  to  assume  a  change  in  the 
colloidal  stability  of  the  bacteria  by  heating  than  to  seek  it  in  a 
differentiation  into  combining  and  agglutinable  parts  of  the  same 
antigen. 

The  points  raised  by  Joos'  work  have  been  followed  up  particu- 

25  Joos.     Centralbl.  f.  Bakt.,  Vol.  33,  1903. 

26  Eisenberg  and  Volk.     Zeitschr.  /.  Hyg.,  Vol.  40,  1902. 


THE    PHENOMENON    OF    AGGLUTINATION 

larly  by  Kraus  and  Joachim  27  and  by  Scheller.28  Scheller  sum- 
marizes the  results  of  his  work  as  follows :  First,  in  agreement  with 
Joos  he  found  that  immune  sera  obtained  by  injection  of  bacteria 
modified  by  heat  vary  considerably  from  each  other.  Secondly,  im- 
munization with  living  typhoid  bacilli  produces  sera  which  agglu- 
tinate living  bacilli  very  highly  and  less  highly  bacilli  heated  to  60° 
C.  The  titre  of  agglutinating  serum  is  altered  very  little  toward 
living  bacilli  after  heating  to  60°  to  62°  C.,  but  is  sometimes  dimin- 
ished toward  bacteria  that  have  been  heated.  Bacilli  jthat  have  been 
heated  to  100°  C.  but  slightly  agglutinate  unheated  serum.  Sera 
produced  by  the  injection  of  typhoid  bacilli  heated  to  60°  to  62°  C. 
agglutinate  with  both  living  and  heated  bacilli. .  Very  important 
furthermore  in  Scheller's  work  are  the  determinations  that  typhoid 
bacilli  which  have  been  heated  absorb  agglutinins  out  of  the  sera 
more  actively  than  do  the  unheated  bacteria,  and  that  the  highest 
agglutinin  titres  can  be  obtained  by  agglutination  with  bacilli  that 
have  been  heated  to  60°  C.  The  analogy  of  Scheller's  results  with 
similar  work  done  in  connection  with  the  precipitin  reaction  is  strik- 
ing and  will  be  referred  to  in  another  place. 

Alterations  in  the  agglutinability  of  bacteria  may  also  occur  spon- 
taneously, without  previous  heating,  as  in  the  preceding  experiments. 
It  has  been  frequently  noticed  that  typhoid  bacilli,  recently  culti- 
vated out  of  the  human  body,  will  not  agglutinate  in  sera  which  have 
high  agglutinating  power  for  strains  kept  for  some  time  on  labora- 
tory media.  Much  investigation  has  been  focused  upon  the  deter- 
mination of  the  cause  for  this,  and  although  many  explanations  have 
been  suggested  no  satisfactory  solution  has  been  found.  •  Most  work- 
ers who  have  attempted  to  attack  this  problem  have  based  their  rea- 
soning upon  the  receptor  conception  of  Ehrlich  and  have  assumed 
that  such  inagglutinable  bacteria  are  characterized  by  a  diminished 
equipment  in  "receptors."  Such  strains  have  been  especially  well 
studied  in  the  case  of  typhoid  bacilli  and  cholera  spirilla.  Inagglu- 
tinable typhoid  bacilli  have  been  cultivated  by  many  investigators 
from  the  spleen,  gall  bladder,  and  urine  of  typhoid  patients,  and 
many  of  these,  when  studied  for  prolonged  periods,  have  been  found 
to  regain  normal  agglutinability  after  several  generations  of  culti- 
vation upon  artificial  media..  Apparently  some  alteration  of  the 
bacilli  had  taken  place  in  the  presence  of  the  body  fluids  (immune 
serum)  which  affected  their  sensitiveness  to  the  agglutinins,  i.  e., 
their  ability  to  unite  with  or  absorb  this  antibody.  The  phenomenon 
involves  an  important  principle,  emphasized  some  years  ago  by  Pro- 
fessor Welch,  namely,  that  the  bacteria  may  acquire  a  quasi- 
immunity  against  the  attacking  forces  of  the  body,  a  property  which' 
may  be  responsible  for  the  increase  of  virulence  noted  when  some 

27  Kraus  and  Joachim.     Centralbl  f.  Bakt.,  I,  Vol.  36,  1904. 

28  Scheller.     Centralbl.  f.  Bakt.t  Vol.  36,  1904,  and  Vol.  38,  1905. 


228  INFECTION    AND    RESISTANCE 

bacteria  are  repeatedly  passed  through  the  bodies  of  animals,  and, 
indeed,  alterations  of  virulence  signify  biologically  a  process  of 
adaptation  on  the  part  of  the  bacteria  just  as  increased  immunity 
indicates  a  similar  process  on  the  part  of  the  invaded  subject. 

This  lessened  susceptibility  to  antibodies  is  noticeable  not  only  in 
strains  cultivated  from  the  body  in  disease,  but  can  be  produced  arti- 
ficially by  cultivating  the  bacteria  in  inactivated  homologous  immune 
serum.  This  has  been  accomplished  by  Walker  29  especially,  and  by 
Miiller,30  with  both  typhoid  bacilli  and  cholera  spirilla  cultivated 
apon  broth  mixed  with  serum.  Such  strains  not  only  increase 
in  virulence  but  lose  in  both  agglutinability  and  susceptibility  to 
bactericidal  effects.  Sacqueppee  31  obtained  similar  results  by  keep- 
ing the  organisms  in  collodium  sacs  in  the  peritoneal  cavity,  and 
Bail  32  found  similar  inagglutinability  of  typhoid  bacilli  taken  from 
the  peritoneal  exudates  of  guinea  pigs  dead  of  infection. 

Zinsser  and  Dwyer 33  have  noticed  similar  inagglutinability  in 
typhoid  bacilli  recovered  from  the  peritoneal  cavities  of  guinea  pigs 
injected  with  anaphylatoxin  and  bacteria.  The  anaphylatoxin  in 
these  cases  possessed  distinct  aggressive  action,  and  the  conditions 
here  were  probably  very  similar  to  those  observed  by  Bail. 

There  are  various  possible  explanations,  the  most  prevalent  ones 
all  representing  variations  of  the  opinion  that  such  inagglutinable 
strains  possess  an  inadequate  receptor  apparatus.  Cole34  advances 
this  because  he  found  these  cultures  possessed  less  power  to  absorb 
agglutinin  than  others,  and,  injected  into  animals,  produced  sera 
which  were  not  highly  agglutinating  for  the  injected  strain.  Some 
of  Cole's  experiments  show  clearly  the  variable  agglutinability  dis- 
played by  different  strains  of  the  same  species.  Thus  the  agglutina- 
tion in  the  same  serum 

Of  strain  E  =  1:8,000 

Of  strain  H  =  1:7,000 

Of  strain  I    =  1:4,500 

Of  strain  W=  1:4,500 

Of  strain  C  =  1:4,000 

The  difference  here  between  E  and  C  actually  amounted  to  a 
relation  of  1  to  2.  A  rabbit  immunized  with  strain  I  furnished  a 
.serum  which  agglutinated  strain  E  more  powerfully  than  I  itself. 

Miiller's  experiments  have  the  same  general  significance.  It  has 
;also  been  suggested  that  the  inagglutinable  bacteria,  especially  those 
from  the  peritoneal  exudate,  which  Bail  found  were  neither  agglu- 

29  Walker.    Journ.  of  Path,  and  Bact.,  Vol.  8,  1902. 

30  Miiller.    Munch,  med.  Woch.,  1903. 

31  Sacqueppee,   Ann.  Past.,  Vol.  4,  1901. 

32  Bail.    Archiv  f.  Hyg.,  Vol.  42. 

33  Zinsser  and  Dwyer.    Proc.  Soc.  for  Exp.  Biol.  and  Med.,,  Feb.,  1914. 

34  Cole.     Zeitschr.  f.  Hyg.,  Vol.  46,  1904. 


THE    PHENOMENON    OF    AGGLUTINATION       229 

tinable  nor  absorbed  agglutinin,  may  have  taken  up  altered  agglu- 
tinin  or  agglutinoid.  We  will  have  occasion  to  recur  to  this  problem 
in  connection  with  our  discussions  of  the  capsulated  bacteria  and  of 
virulence.  The  explanations  given  above  do  not  seem  on  the  whole 
satisfactory,  and  the  problem  is  an  exceedingly  complex  one.  It  has 
been  found  indeed  that  the  acquired  resistance  of  bacteria  against 
agglutinins  is  not  at  all  unique,  and  that  acquired  resistance  against 
serum  lysins  may  be  observed.35  The  extensive  investigations  of 
Bail,  Walker,36  and  others,  on  the  nature  of  changes  in  virulence  in 
many  invasive  bacteria,  and  the  knowledge  more  recently  gained  on 
the  resistance  to  phagocytosis  of  virulent  strains  of  streptococci  and 
pneumococci  are  facts  closely  related  in  underlying  principle  to  the 
inagglutinability  of  typhoid  strains  cultivated  in  immune  sera. 

That  no  two  strains  of  bacteria  of  the  same  species  are  exactly 
similar  in  their  agglutinability  in  the  same  serum,  moreover,  is  an 
observation  wrhich  is  made  by  every  one  who  is  in  a  position  to  carry 
out  routine  Widal  tests  in  hospital  practice.  The  spontaneous  ag- 
glutination which  occasionally  occurs  in  the  broth  cultures  of  typhoid 
bacilli  used  for  this  test  in  many  laboratories  37  may  often  be  re- 
ferred, at  least  in  the  cases  which  have  come  to  the  writer's  notice,  to 
an  excessive  acidity  of  the  broth,  a  phenomenon  which  will  be  dis- 
cussed in  a  subsequent  paragraph.  As  far  as  the  phenomenon  of 
variable  agglutinability  inherent  in  different  strains  is  concerned, 
however,  it  is  of  great  practical  importance  in  carrying  out  routine 
Widal  tests  to  bear  this  in  mind  and  to  control  the  strain  of  typhoid 
bacilli  employed  in  the  reactions  from  this  point  of  view.  A  strain 
also  which  has  been  in  use  for  a  long  time  should  be  titrated  with 
agglutinating  animal  sera  from  time  to  time  to  determine  whether  or 
not  alterations  in  agglutinability  have  occurred. 

That  the  reaction  of  bacterial  agglutination  was  specific  was 
noted,  we  have  seen,  by  Gruber  and  Durham  from  the  very  begin- 
ning. The  closer  study  of  the  reaction  in  its  application  to  bacterial 
identification  has  led  to  interesting  data  which  have  revealed  certain 
definite  limitations  of  this  specificity.  It  has  been  found,  for  in- 
stance, that,  while  immunization  with  any  given  species  of  bacteria 
gives  rise  to  a  very  marked  increase  of  agglutinins  for  this  species, 
there  are  formed  at  the  same  time,  though  to  a  lesser  degree,  agglu- 
tinins for  bacteria  of  other  species.  This  has  been  referred  to  as 
"group  reaction/'  and  the  agglutinins  appearing  in  such  sera  are 
spoken  of  by  German  observers  as  "Haupt  Aggluiinine"  and  "Neben" 
or  "Mit  Agglutinine"  In  English  texts  they  are  usually  referred  to 
as  "chief"  or  "major"  agglutinins  and  "para"  or  "minor"  agglu- 
tinins. Although,  as  a  general  rule,  such  group-agglutinin  formation 

35  Eisenberg.     Centralbl.  f.  Bakt.,  Vol.  34,  p.  739,  1903. 

36  Walker.     Centralbl.  f.  Bakt.,  Vol.  33,  1903. 

37  See  section  on  Aggressins. 


230  INFECTION    AND    RESISTANCE 

is  evident  most  markedly  in  the  cases  of  biologically  related  micro- 
organisms like  the  typhoid,  paratyphoid,  and  colon  bacilli,  this  is  not 
necessarily  the  case,  and  in  some  instances  the  immunization  of  an 
animal  with  a  given  bacterium  may  produce  minor  agglutinins  for 
other  bacteria  that  have  no  morphologically  or  culturally  demonstra- 
ble biological  relation  to  that  which  reacts  with  the  major  agglutinin. 
We  may  obtain  the  most  graphic  survey  of  these  conditions  by  exam- 
ining one  of  a  number  of  experimental  protocols  in  which  such  major 
and  minor  agglutinin  formation  is  illustrated.  Thus  in  the  work  of 
Hiss 38  on  the  dysentery  bacilli  the  following  relations  were  ob- 
served : 

A  serum  produced  in  rabbits  by  immunization  with  the  Shiga 
bacillus  agglutinated  the  Shiga  bacillus  in  dilutions  of  1  to  20,000, 
the  "Baltimore,"  "Harris,"  "Gray,"  and  "Wollstein"  bacillus  1  to 
1,200,  the  Y  bacillus  and  others  1  to  200. 

An  Anti-Y  bacillus  serum,  which  agglutinated  this  bacillus  1  to 
6,400,  agglutinated  the  Baltimore  bacillus  1  to  1,600,  and  the  Shiga 
bacillus  1  to  100. 

Anti-"Baltimore"  bacillus  serum  agglutinated  this  bacillus  1  to 
3,200,  and  the  Y  bacillus  1  to  400,  and  the  Shiga  bacillus  1  to  100. 

In  this  series  there  is  fair  correspondence  between  cultural  bio- 
logical relations  and  agglutination,  and  from  many  such  investi- 
gations it  would  seem  that  "group"  agglutination  might  be  taken 
to  represent  a  method  of  determining  biological  classifications  simi- 
lar to  the  zoological  relations  revealed  by  the  precipitin  reaction. 
While,  in  a  general  way,  this  is  undoubtedly  true,  nevertheless  great 
caution  must  be  exercised  in  relying  upon  such  evidence  for  classifi- 
cation, since  notable  exceptions  have  been  observed.  Park,39  for 
instance,  cites  a  case  in  which  a  horse,  immunized  with  a  paradysen- 
tery  bacillus,  agglutinated  a  colon  strain  in  dilutions  up  to  1  to 
10,000.  Similarly  Durham  40  found  that  two  members  of  the  colon 
group — one  saccharose  fermenting — reacted  almost  identically  with 
the  same  agglutinating  serum,  while  the  agglutinations  of  two  cul- 
turally identical  bacilli  of  the  hog  cholera  group  were  entirely  at 
variance. 

The  cause  for  the  phenomenon  of  group  agglutination  must  un- 
questionably be  sought  in  the  nature  of  the  bacterial  agglutinogens, 
and  it  is  but  reasonable  to  assume  that  living  cells  so  little  differen- 
tiated biologically  and  morphologically  should  have  much  in  common 
chemically  as  well.  The  bacterial  cell,  moreover,  may  contain  sev- 
eral antigenic  complexes  and,  beside  its  specifically  peculiar  constitu- 
ents, therefore,  we  may  suppose  that  every  bacterium  contains  addi- 
tional antigenic  substances  which  it  has  in  common  with  other 

38  Hiss.     Jourri.  of  Med.  Res.,  13,  N.  S.,  Vol.  8,  1904. 

39  Park.     "Pathogenic  Micro-organisms,"  1910,  p.  166. 

40  Durham.     Journ.  of  Med.  Res.,  Vol.  5. 


THE    PHENOMENON   OF    AGGLUTINATION        231 


to.ooo 


ZONE  OF 

ABSOLUTS 

SPECIFICITY 


species.     It  is  the  specific  antigen  in  response  to  which  the  "chief" 
agglutinin  is  formed,  while  the  others,  present  in  smaller  quantity,     / 
lead  to  the  formation  of  the  minor  or  paraagglutinins  with  an  in-^ 
tensity  proportionate  to  the  amounts  present  in  the  bacterial  cell. 
Thus,  as  Durham  expresses  it,  if  we  assume  one  micro-organism  to 
contain  antigenic  substances  a,  b,  c,  and  d,  and  another  d,  e,  f,  and  g, 
the  antibodies  produced  by  injections  of  the  former  would  react  with 
the  common  element  d  in  the  latter. 

The  diagnostic  value  of  the  specificity,  however,  is  plainly  not 
affected  by  the  phenomenon  of  group 
agglutination,  since  the  action  of  minor 
agglutinins  can  be  always  easily  elim- 
inated by  sufficient  dilution.  Thus  if 
we  possess  a  typhoid-immune  serum 
which  agglutinates  the  typhoid  bacillus 
in  dilutions  of  1  to  10,000,  the  para- 
typhoid bacillus  1  to  1,000  and  the 
colon  bacillus  1  to  100  (as  in  the  fig- 
ure), we  may  still  utilize  this  serum  for 
the  identification  of  suspected  typhoid 
cultures,  as,  let  us  say,  in  the  isolation 
of  unknown  bacteria  from  stools  or 
urine,  by  using  potent  sera  in  dilutions 
as  high  or  higher  than  1  to  1,000,  be- 
yond which  point  the  action  of  minor 
agglutinins  is  eliminated.  The  dia- 
gram illustrates  our  meaning  in  the 
hypothetical  case  of  a  typhoid-immune 
serum  which  agglutinates  typhoid  in 
dilutions  of  1  to  10,000,  paratyphoid 
bacilli  1  to  800,  and  colon  bacilli  1  to 

100.  The  relation  of  agglutination  to  biologic  relationship  is  not~a 
simple  problem  in  that  individual  strains  even  of  the  same  species 
may  vary  considerably  in  agglutination  by  the  same  serum.  Smith 
and  Reagh41  have  studied  particularly  these  conditions  as  they  pre- 
vail in  the  colon,  hog  cholera  and  allied. groups.  They  found  that 
biologic  relationship  usually  may  be  concluded  from  close  agglutina- 
tion affinities,  and  that  minor  biologic  differences  such  as  colony 
appearance,  etc.,  do  not  exclude  such  affinities.  On  the  other  hand, 
closely  related  bacteria  vegetating  on  mucous  surfaces  (different 
strains  of  diphtheria,  dysentery,  and  colon  bacilli)  may  vary  con- 
siderably in  their  agglutinative  characteristics,  while  invasive  species 
show  a  greater  homogeneity  among  their  varieties  or  races.  This 
brings  in  another  important  feature — that  is,  the  modification  in 

41  Smith  and  Reagh.     "Studies  from  the  Rockefeller  Institute,"  Vol.  1, 
1904,  p.  270. 


BOO 


SOO 

T 


TYPHOJD  PARATYPHOID  COLON 

DIAGRAMMATIC  EEPBESENTATION 
OF  GROUP  AGGLUTINATION. 


232 


INFECTION    AND    RESISTANCE 


agglutinative  characteristics  induced  in  bacteria  when  they  become 
parasitic  upon  different  hosts,  and  Smith  and  Reagh  conclude  that 
such  changes  of  parasitic  habitat  may  modify  the  agglutinative  prop- 
erties (probably  by  adaptation  to  the  peculiar  reactions  of  each  host), 
some  of  them  being  weakened  and  others  strengthened. 

The  animal  species  used  for  immunization  indeed  influences  the 
quantity  and  nature  of  the  produced  agglutinin  considerably.  For 
instance,  in  Pfeiffer's  42  experiments,  a  dog,  a  chicken,  and  a  rabbit 
were  immunized  with  the  same  strain  of  cholera  spirilla.  The 
sera  obtained  from  these  animals  agglutinated  this  and  other  strains 
of  cholera  spirilla  in  an  entirely  irregular  manner — showing  that 
the  constitution  of  the  agglutinins  in  each  case  was  an  absolutely 
different  one  in  regard  to  the  relative  concentrations  of  "major"  and 
"minor"  constituents. 

Castellani 43  found  that  the  immunization  of  an  animal  with 
two  or  more  different  species  of  bacteria  results  in  the  formation  of 
agglutinins  against  all  of  these,  (^upposing,  for  instance,  that  species 
A  and  B  are  used  for  treatment,  agglutinins  against  both  A  and  B 
are  formed  in  quantity,  depending  upon  the  intensity  of  the  treat- 
ment in  each  case.  Now,  if  to  the  serum  so  produced  an  emulsion  of 
A  is  added,  agglutinin  A  only  will  be  removed,  while  agglutinin  B 
will  remain  in  the  serum  almost  undiminished.  An  example  of  this 
is  seen  in  the  following  protocol  taken  from  Castellani's  paper : 


Titre  of  the  serum 

Titre  after 
absorption  with 
B.  typhi 

Titre  after 
absorption  with 
B.  coli  "31" 

Titre  after 
absorption  with 
B.  coli  and  B.  typhi 

B.  typhi  4,000 
B.  coli'1  31"  1,000 

B.  typhi  0 
B.co&"31" 

1,000  >  300 

B.  typhi  4,000 
B.  coli"  31"  Q 

B.  typhi  0 
B.  coli"  31"  0 

In  the  preceding  paragraphs,  however,  we  have  seen  that  im- 
munization with  a  single  organism,  say  B.  typliosus,  will  induce  the 
formation  of  agglutinins,  not  only  for  this  species,  but  also  of  para 
or  minor  agglutinins  for  biologically  similar  strains  as  well.  In 
such  cases,  as  Castellani  showed,  absorption  of  the  serum  with  the 
organism  used  for  immunization  takes  out,  not  only  the  major  ag- 
glutinins, but  rather  all  of  the  agglutinins,  major  and  minor.  Con- 
versely, however,  absorption  of  such  a  serum  with  the  species  ag- 
glutinated by  the  minor  agglutinins  takes  out  these  antibodies  only^ 
leaving  the  major  substances  intact.  These  relations  are  well  illus- 

42Pfeiffer.     Quoted  from  Paltauf  in  "Kolle  u.  Wassermann  Handbuch," 
Vol.  4. 

43  Castellani.     Zeitschr.  f.  Ilyg.,  Vol.  40,  1902. 


THE    PHENOMENON  OF    AGGLUTINATION 


trated  by  the  two  following  protocols,  also  taken  from  Castellani's 
paper : 


Serum  of  rabbit  No.  1  immunized  with  B.  typhi  only 

Agglutination  titre  of  serum 

Titre  after  absorption  with 

B.  typhi 

B  typhi  5,000 
B.  coli  .     600 

B.  typhi  0* 
B.  coli     0 

Serum  of  rabbit  No.  7  immunized  with  B.  typhi  only 

Agglutination  titre 
of  serum 

After  absorption  with 
B.  typhi  • 

After  absorption  with 
B.  coli 

B. 
B. 

typhi  10,000 
(heavy  clumps) 
coli  800 

B.  typhi  0 
B.  coli  0 

B.  typhi  10,000  , 
(small  clumps) 
B.  coli  0  - 

Note:  All  of  these  protocols  are  taken  from  Castellani's  communication,  loc.  cit. 

These  facts,  variously  confirmed,  tend  to  corroborate  the  concep- 
tion of  the  production  of  major  and  minor  agglutinins  outlined 
above. 

It  is  of  practical  and  theoretical  importance  to  mention  that 
complete  absorption  of  specific  agglutinin  by  a  single  exposure  to 
homologous  bacteria,  however  thickly  emulsified,  is  not  possible.  It 
is  always  necessary  to  absorb  repeatedly,  and  even  then  a  minute 
trace  of  agglutinin  may  eventually  remain.  Eisenberg  and 
Yolk,44,  45  who  have  studied  these  conditions  particularly,  attribute 
this  to  the  nature  of  the  union  of  agglutinogen  with  agglutinin, 
which  they  conceive  as  following  the  laws  of  mass  action — this  ac- 
counting for  the  persistence  of  a  small  "rest"  of  free  agglutinin, 
even  after  repeated  absorption  by  partial  dissociation.  The  prin- 
ciple involved  here  is  identical  with  that  discussed  in  connection  with 
antigen-antibody  union  in  general.  / 

It  is  not  only  in  the  sera  of  immunized  animals,  however,  that 
agglutinins  are  found.  Just  as  the  other  antibodies,  antitoxins,  and 
bactericidal  sensitizers  may  be  found  in  the  blood  of  normal  animals, 
so  agglutinins  for  various  bacteria  may  be  normally  present.  These 
normal  agglutinins  do  not  in  any  respect,  further  than  that  of  quan- 
tity, differ  from  the  immune  agglutinins  and  follow  the  same  laws 
of  specificity  which  have  been  described  for  the  latter.  It  has  been 
shown  a  number  of  times  that  such  normal  agglutinins  are  not  pres- 

44  Eisenberg  and  Volk.    Zeitschr.  f.  Hyg.,  Vol.  40,  1902. 

45  Eisenberg.    Centralbl.  f.  Bakt.,  Vol.  34,  1903. 


INFECTION    AND    RESISTANCE 

ent  in  the  new-born  animal,  but  are  acquired  later  in  life,  possibly 
because  of  the  absorption  of  bacterial  products  from  the  intestinal 
canal.  It  has  been  variously  shown,46  too,  that  living  bacteria  them- 
selves may  enter  the  lymphatics  and  the  portal  circulation  from  the 
intestine  during  apparently  perfect  health  of  the  individual. 

This  subject  is  of  interest,  not  only  in  connection  with  the  ag- 
glutinins,  but  has  bearing  upon  the  existence  of  normal  antibodies 
in  general.  Ruffer,47  who  has  studied  particularly  the  penetration 
of  leukocytes  and  bacteria  through  the  intestinal  mucosa,  demon- 
strated micro-organisms  in  the  sub-mucous  lymph  nodes  of  normal 
rabbits,  and  Ribbert  48  and  Bizzozero  49  have  shown  the  presence  of 
bacteria  in  apparently  normal  mesenteric  lymph  nodes.  Adami 
and  Nichols  even  claim  that  during  health  a  certain  number  of  liv- 
ing bacteria  enter  the  portal  circulation  from  the  intestine,  and  from 
here  may  get  into  the  systemic  circulation,  and  are  ordinarily  de- 
stroyed by  either  leukocytes,  liver  lymphatic  organs,  or  the  kidneys. 

It  is  thus  not  surprising  that  normal  agglutinins  should  occur, 
and  that  they  should  be  qualitatively  identical  with  the  so-called 
"immune"  agglutinins,  since  they  probably  arise  by  a  sort  of  spon- 
taneous immunization  through  the  intestinal  canal.  From  the  in- 
vestigations of  Ford  especially  we  may  conclude  that  the  immune 
agglutinin  may  be  regarded  as  merely  a  quantitative  increase  of  the 
normal  antibody,  if  this  has  been  present  before  immunization. 
Ford  50  found  that  when  an  animal  is  treated  with  an  agglutinating 
serum  an  anti-agglutinin  may  be  obtained  which  neutralizes  the 
action,  not  only  of  immune,  but  also  of  homologous,  normal  ag- 
glutinin. 

An  interpretation  of  the  process  of  agglutination,  according  to 
the  theory  of  Ehrlich,  conceives  it  as  a  chemical  union  of  agglutinin 
and  bacteria   (agglutinogen).     The  agglutinin  is  regarded  as  con- 
sisting of  two  atom  complexes,  one  the  "haptophore,"  having  af- 
finity for  the  bacterial  protein,  and  concerned  with  the  union,  the 
other  the  "ergophore"  or  "zymophore,"  by  means  of  which  the  actual 
agglutination  is  brought  about  after  the  union  has  taken  place.    Un-  j 
like  the  antibody  concerned  in  the  processes  of  hemolysis  or  bac-  ! 
teriolysis,  the  agglutinins  are  not  dependent  in  their  action  upon  the   , 
cooperation  of  alexin,  and  the  agglutination  power  of  a  serum  is 
therefore  not  destroyed  by  inactivation  or  heating  to  56°  C.,  as  is 
the  case  with  the  former.     Although  the  accurate  point  of  thermal 
destruction  varies  with  different  agglutinins  (the  agglutinins  for  the 
Bacillus  peslis  and  a  few  other  bacilli  are  said  to  be  destroyed  at 

46  Adami.    Jour.  Am.  Med.  Assoc.,  Dec.,  1899. 

47  Ruffer.    Brit.  Med.  Journal,  2,  1890. 

48  Ribbert.     Deutsche  med.  Woch.,  1885. 

49  Bizzozero.     Centralbl.  f.  d.  Med.  Wiss.,  Vol.  23,  1885,  p.  49.     Quoted 
from  Adami. 

50  Ford.    Zeitschr.  f.  Hyg.,  Vol.  40,  1902. 


THE   PHENOMENON  OF    AGGLUTINATION        235 

56°  C.),  as  a  rule  agglutinins  will  not  disappear  from  serum  until 
the  temperature  is  raised  to  between  70°  and  80°  C.  Once  de- 
stroyed, however,  no  reactivation  takes  place  upon  the  addition  of 
fresh  normal  serum.  Ehrlich,  for  this  reason,  has  conceived  the 
structure  of  agglutinins  as  '  Ilaptines  of  the  Second  Order,"  in 
which  he  supposes  that  the  zymophore  and  the  haptophore  groups  are 
inseparably  connected,  and  in  which  we  could  assume  an  alteration 
of  the  less  stable  zymophore  group  without  interference  with  the 
functional  integrity  of  the  haptophore  group.  Such  an  altered  ag- 
glutinin  could  be  spoken  of  as  "agglutinoid,"  and  this  could  become 
united  with  a  bacterial  cell  without 
inducing  agglutination,  but,  by  its 
union,  prevent  subsequent  combina- 
tion of  the  cell  with  unaltered  agglu- 
tinin.  This  process  of  Aagglutinoid 
Yerstopfung"  has  been  held  respon- 
sible for  the  failure  of  agglutination 
when  bacteria  that  have  been  in  con- 
tact  with  heated  serum  are  subse- 
quently  exposed  to  the  action  of  ac- 

tively    agglutinating    serum.      It    is 

•^     i  ;  i    . .      .  ,        ,  .  ,       DIAGRAMMATIC  REPRESENTATION  OF 

assumed  that  the  agglutmoids  which  EHRLICH  's  CONCEPTION  OF  THE 

were    present    in   the    heated    serum  STRUCTURE  OF  AGGLUTININ. 

have  occupied  the  bacterial  receptors 

and  have  thereby  prevented  the  union  of  these  with  the  agglutinins 
later  added. 

The  work  of  Eisenberg  and  Yolk  51  has  gone  very  thoroughly 
into  these  conditions  and  forms  the  strongest  bulwark  of  this  point 
of  view.  These  workers  showed  that  bacteria  thus  exposed  have 
not  only  become  less  sensitive  to  agglutinins,  but  have,  at  the  same 
time,  lost  much  of  their  power  to  absorb  agglutinins  when  compared 
with  normal  bacteria.  The  same  loss  of  agglutinating  power  which 
is  observable  in  heated  agglutinating  serum  is  evident  to  a  lesser 
extent  in  serum  which  has  been  preserved  at  room  temperature. 
Eisenberg  and  Yolk  have  shown  that  such  serum,  in  addition  to  a 
loss  of  its  total  agglutinin  content,  loses  the  power  to  agglutinate 
in  high  concentrations.  Thus  a  serum  which  has  been  preserved  in 
this  way  will  no  longer  agglutinate  bacteria  in  concentrations  of  1 
to  20,  1  to  40,  or  even  1  to  100,- but  will  agglutinate  as  before  in 
higher  dilutions.  This  is  taken  to  mean  that  the  agglutmoids  formed 
during  the  period  of  standing  possess  a  higher  affinity  for  the  bac- 
terial antigen  than  do  the  true  unaltered  agglutinins.  Since  these 
so-called  "proagglutinoids"  are  relatively  small  in  amount,  their 
action  is  masked  when  considerable  dilution  has  sufficiently  di- 
minished their  quantity,  in  proportion  to  the  more  plentiful  un- 

51  Eisenberg  and  Volk.     Zeitschr.  f.  Hyg.,  Vol.  40,  1902. 


236  INFECTION    AND    RESISTANCE 

altered  agglutinins.  In  support  of  this  assumption  it  has  been 
further  shown  that  sera  which  'have  been  rendered  inhibitory  by 
either  of  the  methods  named  can  be  deprived  of  their  inhibiting 
characteristics  by  absorption  with  homologous  bacteria.  Together 
these  observations  constitute  a  strong  argument  in  favor  of  the  ag- 
glutinoid  theory. 

In  practical  experience  the  existence  of  such  an  inhibition  zone 
is  of  great  importance,  since  freshly  taken  sera  will  occasionally 
show  this  failure  of  agglutination  in  concentration,  while  strong 
agglutination  follows  when  the  dilution  is  increased.  In  clinical 
tests,  as  in  the  Widal  reaction  for  the  diagnosis  of  typhoid  fever, 
we  not  infrequently  encounter  sera  which  will  give  no  agglutination 
in  dilutions  of  1  to  20  and  even  1  to  40,  and  the  reaction  would 
therefore  be  regarded  as  negative  unless  the  possibility  of  the  pro- 
agglutinoid  zone  were  recognized  and  further  dilutions  carried  out. 

While  there  is  no  question  of  the  accuracy  of  the  experimental 
data  cited  in  the  preceding  paragraphs,  the  interpretation  of  the  phe- 
nomena on  the  basis  of  Ehrlich's  haptine  conception  has  not  been 
universally  accepted. 

The  fact  that  the  agglutinins  lose  their  agglutinating  power  after 
heating,  while  retaining  their  power  to  prevent  the  subsequent  agglu- 
tination of  the  bacteria,  may  be  more  simply  explained  in  analogy 
with  the  observation  of  Forges  on  the  influence  of  heated  serum  upon 
the  agglutination  of  mastic  suspensions.  He  found  that  unheated 
serum  will  flocculate  such  suspensions,  while  heated  serum  of  the 
same  concentration  will  prevent  the  flocculation,  acting  probably  as 
a  protective  colloid  (see  chapter  on  Colloids).  In  the  same  way  the 
heated  agglutinating  serum  may  prevent  subsequent  flocculation  by  a 
similar  protective  action.  We  suggest  this  as  a  possible  explanation 
of  the  proagglutinoid  phenomenon,  although  of  course  it  is  a  mere 
conjecture  as  opposed  to  the  painstaking  experiments  of  Eisenberg 
and  Volk.  It  has  the  advantages  of  simplicity,  but  does  not,  it  is 
true,  explain  the  apparent  specific  absorption  of  the  agglutinin- 
inhibiting  properties  out  of  heated  sera  with  homologous  bacteria,  as 
claimed  by  these  authors  as  well  as  Kraus  and  Joachim.  The  simi- 
larity of  the  proagglutinoid  phenomenon  to  the  inhibition  zones  oc- 
curring in  the  flocculation  of  colloids  will  be  referred  to  in  a  subse- 
quent paragraph. 

In  describing  the  investigations  which  led  to  the  discovery  of 
the  mechanism  of  "the  lytic  phenomenon,  in  the  chapter  on  Cytolysis, 
we  mentioned  that  Bordet  and  others  had  noticed  the  frequent  ag- 
glutination of  red  blood  cells,  in  the  sera  of  animals  treated  with 
such  cells  after  the  hemolytic  property  had  been  destroyed  by  heat- 
ing to  56°  C.  Such  hemagglutination  is  a  phenomenon  entirely 
analogous  to  the  agglutination  of  bacteria  by  serum,  and  hemag- 
glutinins  regularly  result  when  an  animal  is  systematically  treated 


THE    PHENOMENON    OF    AGGLUTINATION  237 

with  the  red  blood  cells  of  another  species.  Like  the  bacterial  ag- 
glutinins,  the  hemagglutinins  are  relatively  thermostable  and  are 
best  observed,  therefore,  after  the  sera  are  inactivated.  Otherwise 
rapid  hemolysis  will  often  obscure  the  agglutination.  Like  other 
agglutinins  the  hemagglutinins  thus  produced  are  specific,  acting 
only  upon  that  variety  of  cells  which  are  used  in  their  production. 
Moreover,  certain  sera  may  normally  contain  hemagglutinins  for  the 
blood  cells  of  animals  of  another  species.  An  illustration  of  this  is 
the  hemolytic  and  hemagglutinating  property  of  normal  goat  serum 
for  rabbit  cells — but  there  are  many  other  examples  which  might  be 
cited.  JSuch  normal  hemolytic  and  hemagglutinating  properties  for 
the  cells  of  other  animals  usually  render  thejsera  toxic  for  these  ani-> 
malsj  and  some  observers  have  attributed  the  toxicity  T;o  this  agglu-1 
tinating  action,  believing  that  the  clumped  erythrocytes  form  emboli/ 
around  which  clotting  is  initiated.  The  writer's  own  investigations, 
however,  seem  to  show  that  this  is  not  the  case,  since  the  toxicity  of 
such  sera  is  completely  removed  after  they  have  been  heated,  in  spite 
of  the  fact  that  the  hemagglutinative  property  remains  unchanged. 

In  discussing  hemolysins,  also,  we  called  attention  to  the  observa- 
tion that  the  sera  of  animals  may  develop  the  property  of  hemolyzing 
blood  cells  of  other  individuals  of  the  same  species  when  immunized 
with  such  cells,  and  that  on  occasion  such  isolysins  may  be  normally 
present. 

Analogous  to  iso-lysins,  iso-agglutinins  also  jiaye  been  observed. 
They  were  described  first  in  human  blood  in  lyoOTTndependently,  by 
Landsteiner,52  and  by  Shattock,.  As  the  result  of  the  work  of  a 
number  of  men,  in  particular  that  of  Landsteiner,  of  Ascoli,53  and 
of  Descatello  and  Sturlii,54  it  became  evident  that  all  human  blood 
fell  into  four  sharply  separated  and,  for  the  individual,  permanent 
groups,  according  to  the  way  in  which  they  interagglutinate.  The 
members  of  the  first  group  have  blood  cells  which  are  not  agglutinated 
by  the  serum  of  any  human  blood.  The  serum  of  the  members  of 
this  group  agglutinates  the  blood  cells  of  persons  belonging  to  any 
of  the  other  three  groups.  This  group  constitutes  between  40  to  50 
per  cent,  of  all  human  beings.  Members  of  the  second  group  have 
blood  serum  which  agglutinates  the  cells  of  persons  belonging  to  the 
third  and  fourth  groups ;  while  the  cells  of  the  second  group  are  ag- 
glutinable  by  the  serum  of  the  first  and  third  groups.  The  third  is 
the  reciprocal  of  the  second,  and  the  serum  of  the  third  group  ag- 
glutinates the  cells  of  the  second  and  fourth  groups ;  while  the  cells 
of  the  third  group  are  agglutinated  only  by  the  serum  of  the  first  and 
second  groups.  The  fourth  group  (which  is  very  rare)  is  the  recipro- 

52  Landsteiner  and  Richter.     Zeitschr.  f.  Med.,  3,  1902. 

53  Ascoli.     Munch,  med.  Woch.,  1901. 

54  Descatello  and  Sturlii.    Munch,  med.  Woch.,  1902,  49,  p.  1090. 


238 


INFECTION    AND    RESISTANCE 


TABLE   FOR   ISO-AGGLUTININS  * 

Sera 


I 

II 

III 

IV 

1 

2 

3 

4 

5 

6 

7 

8 

9 

10 

r 

1 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

2 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

I' 

3 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

4 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

[ 

5 

+ 

+ 

+ 

+ 

0 

0 

0 

+ 

+ 

0 

fm 

6 

+ 

+ 

+ 

+ 

Q 

0 

0 

+ 

+ 

0 

i 

7 

+ 

+ 

+ 

+ 

0 

0 

0 

+ 

+ 

0 

\ 

8 

+ 

+ 

+ 

•  + 

+ 

+ 

0 

0 

0 

in\ 

i 

9 

+ 

+ 

+ 

+ 

+ 

+ 

+ 

0 

0 

0 

ivj 

10 

+ 

* 

i 

t 

* 

+ 

+ 

+ 

+ 

0 

*  For  this  table  as  well  as  for  much  direct  information  concerning  the  iso-agglutinins  and  iso- 
lysin  we  are  indebted  to  Dr.  Ottenberg  of  this  laboratory. 

cal  of  the  first,  the  serum  having  no  agglutinin,  the  cells  being  ag- 
glutinated by  the  serum  of  any  other  group.  (See  table.)  It  is 
evident  from  examining  this  grouping  that  the  phenomena  can  be 
explained  (as  Landsteiner  has  suggested)  if  it  is  assumed  that  there 
are  two  agglutinins  (a  and  ft)  and  two  corresponding  agglutinogens 
present  in  the  red  cells  (A  and  B).  The  blood  of  the  first  group  pos- 
sesses both  agglutinins,  but  no  agglutinogens,  the  blood  of  the  second 
group  possesses  agglutinin  a?  agglutinogen  B,  the  blood  of  the  third 
group  possesses  agglutinin  /?,  agglutinogen  A,  the  blood  of  the  fourth 
group  possesses  no  agglutinin  but  both  agglutinogens. 

These  agglutinins  are  present  in  weak  dilution  only,  being  gen- 
erally active  in  dilutions  only  of  1-15  to  1-30.     They  are  separately 
absorbed  from  the  serum  by  the  suitable  red  cells  (Hektoen).55     Ot- 
55  Hektoen.     Jour,  of  Inf.  Dis.}  1907,  p.  297. 


THE    PHENOMENON    OF    AGGLUTINATION  239 

tenberg  noticed  that  they  were  inherited,  and  this  was  also  shown 
in  1908  and  in  1910  by  von  Dungern  and  Hirschfeld,56  who  further 
found  that  this  inheritance  followed  the  Mendelian  law  strictly. 
The  agglutinogens  are  the  unit  characters.  The  agglutinogens  ap- 
parently are  present  at  an  earlier  embryonic  stage  than  the  ag- 
glutinins.  On  account  of  their  hereditary  nature  and  permanence 
for  the  individual  the  iso-agglutinins  may  possibly  be  of  medicolegal 
value.  They  may  also  be  of  some  practical  importance  in  selecting 
donors  for  blood  transfusion,  as  phagocytosis  of  red  blood  cells  in  the 
circulation  after  transfusion,  first  described  by  Hopkins,  was  proved 
by  Ottenberg  to  occur  only  when  the  patient's  serum  was  agglutina- 
tive toward  the  donor's  red  cells,  and  several  such  transfusions  have 
had  fatal  results.  Similar  iso-agglutinins  have  been  observed  in  the 
blood  of  lower  animals,  in  horses  (Klein,57  1902)  ;  rabbits  (Boycott 
and  Douglas,58  1910)  ;  cats  (Ingebrigtsen)  ;  dogs,  rats,  and  steers 
(Ottenberg).59  The  iso-agglutinins  have  been  developed  in  dogs 
(von  Dungern  and  Hirschfeld).60  In  most  of  the  lower  animals 
they  have  occurred  with  peculiar  irregularity,  indicating  probably 
the  presence  of,  not  two,  but  of  a  larger  number  of  agglutinins  and 
agglutinogens.  In  steers,  however,  they  fall  into  simple  groups,, 
indicating  the  presence  of  only  one  agglutinin  and  agglutiuogen.  In 
many  animals  the  agglutinins  are  entirely  latent,  and  the  biochemical 
differences  represented  by  the  agglutinogens  are  present  in  the  red 
cells,  and  the  correct  agglutinin  is  developed  by  the  animal  only 
when  it  is  immunized  with  blood  whose  cells  contain  agglutinogen 
not  present  in  the  animal's  own  blood  cells. 

The  fundamental  principle  underlying  all  the  Ehrlich  hypotheses, 
is  the  conception  that  these  reactions  take  place  as  do  purely  chemical 
reactions^  following  the  law  of  multiple  proportions.     Such  reasoning" 
has  often  necessitated  the  assumption  of  differences  of  affinity  which, 
critically  examined,  are  really  ex  post  facto  explanations,  forcing          / 
the  phenomena  to  conform  with  the  theory.     As  a  matter  of  fact,    ^/ 
the  bodies  which  participate  in  the  antibody-antigen  reactions  are 
probably  of  the  nature  of  the  substances  which  are  spoken  of  as  col-^ 
Ipids,  and  it  is  therefore  more  than  likely  that  the  quantitative  and 
other  relations  governing  the  union  of  these  reagents  should  be  anal- 
ogous to  those  governing  colloidal  reactions  in  general.     The  reaction 
of  agglutination,  like  that  of  precipitation,  has  lent  itself  particularly 
to  the  study  of  the  principles  of  the  union  from  this  point  of  view, 
and  the  first  and  fundamental  progress  made  in  this  direction  is  j 
found  in  the,  work  of  Bordet. 

56  Von  Dungern  and  Hirschfeld.     Zeitschr.  f.  Imm.,  4,  1909-1910,  p.  53; 
also  ibid.,  1910,  p.  284. 

"Klein.     Wien.  kl.  Woch.,  1902,  p.  413. 

58  Boycott  and  Douglas.     Jour,  of  Path,  and  Pact.,  Jan.,  1910. 

59  Epstein  and  Ottenberg.     Tr.  N.  Y.  Path.  Soc.,  1908. 

60  Von  Dungern  and  Hirschfeld.     Zeitschr.  f.  Imm.,  1909,  1910,  p.  531. 


240  INFECTION    AND    RESISTANCE 

Bordet 61  compared  the  formation  of  precipitates  in  bacterial 
emulsions  to  the  precipitation  of  such  inorganic  emulsions  as  clay  in 
distilled  water,  and  noted  that  the  precipitation  of  homogeneous 
emulsions  of  such  substances  is  "often  controlled  by  such  insignifi- 
cant causes  as  the  presence  of  salts."  Applying  this  analogy  to  the 
agglutination  of  bacteria,  he  performed  the  following  experiment: 
Cholera  spirilla,  emulsified  in  salt  solution,  were  treated  with  homol- 
ogous immune  serum  and,  after  agglutination  had  taken  place,  the 
bacteria  were  thrown  to  the  bottom  by  centrifugation  and  divided  into 
two  parts.  One  part  was  again  suspended  in  salt  solution,  and  the 
other  was  washed,  and  then  suspended  in  distilled  water.  The  bac- 
teria in  the  tube  of  salt  solution  rapidly  agglutinated,  while  those 
in  the  distilled  water,  after  thorough  shaking,  remained  indefinitely 
suspended  in  an  even  emulsion.  If,  however,  to  these  unagglutinated 
bacteria  a  small  amount  of  pure  sodium  chlorid  was  added  agglutina- 
tion occurred. 

The  conclusions  that  can  justly  be  drawn  from  this  experiment 
are,  first,  that  the  bacteria  could  not  agglutinate,  even  though  they 
\/  had  been  bound  to  agglutinin,  when  salt  was  removed  from  the  en- 
vironment, and,  second,  that  the  addition  of  salt  to  such  emulsions 
brought  about  immediate  agglutination.  The  same  principle  can  be 
demonstrated  in  other  ways.  If,  for  instance,  a  bacterial  emulsion 
is  rendered  free  of  salt  by  dialysis,  and  this  is  added  to  an  aggluti- 
nating serum  similarly  dialyzed,  no  agglutination  occurs.  The  sus- 
pension may  remain  evenly  clouded  indefinitely  unless  salt  is  added. 
As  soon  as  a  little  salt  is  added,  however,  perfect  agglutination 
occurs.  To  this  technique  the  very  obvious  criticism  may  be  applied 
that  perhaps  the  absence  of  salt  has  precipitated  the  agglutinins, 
which,  as  we  know,  are  precipitated  with  globulin,  which  is  insoluble 
in  the  absence  of  salt.  However,  this  source  of  error  is  excluded  by 
the  first  experiment  cited,  and,  moreover,  it  can  be  shown  by  the  last 
experiment  that,  even  though  the  bacteria  are  not  agglutinated  in 
the  salt-free  serum,  they  have  nevertheless  absorbed  agglutinin.  For, 
if  such  a  salt-free  mixture  is  centrifugalized,  the  bacteria  washed 
and  suspended  in  distilled  water,  and  salt  is  then  added,  agglutination 
occurs.  The  supernatant  fluid  of  the  original  suspension,  further- 
more, can  be  shown  to  have  been  deprived  of  agglutinins  by  suitable 
•  experiment. 

These  facts,  first  observed  by  Bordet,  and  further  elaborated  by 
the  studies  of  Joos,62  Friedberger,63  and  others,  have  been  inter- 
preted in  different  ways.  Joos  claims  that  there  is  a  ^chemical  union 
between  the  bacteria  and  the  salt,  and  bases  this  upon  me"  observation 
that  the  salt  added  to  a  salt-free  mixture  cannot  be  demonstrated  in 

61  Bordet.    Ann.  de  I'Inst.  Pasteur,  1896,  1899. 

62  Joos.     Centralbl  f.  Bakt.,  1,  Vol.  33,  1903. 

63  Friedberg-er.    Berl  kl  Woch.,  1902;  Centralbl.  f.  Bakt.,  1,  Vol.  30. 


THE    PHENOMENON    OF    AGGLUTINATION 

the  supernatant  fluid  after  agglutination  has  taken  place.     His  ob- 
servations in  this  respect  have  not  found  confirmation  at  the  hands 
of  Friedberger  and  other  workers,  and  it  is  generally  agreed  to-day 
that  the  role  of  the  salts  is,  as  Eordet  first' assumed  it  to  be,  a  purely 
physical  ona    Bordet's  opinion  is  often  spoken  of  as  the  "two  phase"   » 
theory,  in  that  he  conceives  the  process  of  agglutination  to  consist  of  [ 
two  distinct  occurrences,  first,  an  absorption  of  the  agglutinin  by  • 
the  bacteria,  and,  second,  an  agglutination  of  the  new  complex  by  the  « 
salt.     It  is  not  the  agglutinin  which  causes  agglutination,  but  by 
union  with  the  agglutinogen  forms  a  complex  which  is  altered  in 
"colloidal  stability/7  and  therefore  is  flocculable  by  the  electrolyte. 

The  opinion  of  Bordet  becomes  clearer  as  we  consider  the  con- 
ditions governing  the  flocculation  of  colloids  in  general.  Without 
wishing  to  enter  in  this  place  into  detail  regarding  the  nature  of 
colloidal  suspensions,  it  nevertheless  seems  necessary  in  order  to  do 
justice  to  this  phase  of  the  question  to  recall  briefly  the  conditions 
governing  such  flocculation.  The  so-called  colloidal  solutions  are 
not  true  solutions  as  the  term  is  applied  to  dissociable  substances,  but 
are  looked  upon  as  consisting  of  small  particles  in  suspension.  The 
particles  are  similarly  charged,  as  can  be  demonstrated  by  their 
wandering  when  subjected  to  an  electric  current,  and  it  is  supposed 
that  it  is  this  fact  of  similarity  of  charge  which,  in  the  "sol"  state, 
permits  them  to  remain  in  suspension.  For  the  similarity  of  the 
charges  of  the  individual  particles  prevents  their  mutual  approxima- 
tion. 

The  state  of  suspension  of  these  substances,  then,  represents  a 
delicately  balanced  equilibrium  between  the  two  forces  of  electrical 
repulsion  and  of  surface  tension,  an  equilibrium  which  may  be 
disturbed  by  the  action  of  a  number  of  factors.  Thus,  studies  on 
inorganic  colloids  have  shown,  long  before  these  considerations  were  / 
applied  to  the  explanation  of  serum  reactions,  that  the  stability  of  \/ 
these  suspensions  could  be  disturbed  both  by  electrolytes  and  by  the 
addition  of  other  colloids.  Thus  they  may  be  precipitated  by  vari- 
ous salts,  acids,  and  bases  and,  as  Schultze  64  has  shown,  they  react 
with  that  ion  of  the  electrolyte  which  carries  an  opposite  charge  to 
that  of  the  colloidal  particles.  For,  although  the  colloidal  units  are 
similarly  charged,  this  may  be  either  negative  or  positive,  according 
to  the  nature  of  the  particular  'substance.  In  the  case  of  the  so-called 
amphoteric  colloids  reaction  may  take  place,  according  to  Pauli,65 
with  both  ions  of  the  electrolyte. 

The  probable  mechanism  of  the  process  is  postulated  by  Pauli 
in  describing  the  precipitation  of  a  colloidal  metal  by  salts,  acids,  or 
bases  in  the  following  way: 

64  Schultze.    Journ.  f.  prakt.  Chem.,  25,  1882,  and  27,  1883. 
5  Pauli.      "Hofmeister's    Beitra^e,"    1905,    and    "Physical    Chemistry   in 
icine,"  Wiley  &  Son,  N.  Y.,  1907. 


65 

Medicine 


INFECTION    AND    RESISTANCE 

"The  introduction  of  such  electrolytes  into  a  colloidal  suspension 
is  of  course  accompanied  by  a  certain  amount  of  dissociation.  In 
consequence  the  weakly  charged  particles  of  the  colloid  collect  ahout 
the  ions  of  opposite  charge  until  a  sufficient  accumulation  of  the 
particles  leads  to  an  electrical  neutralization  of  the  ion,  and  the  ag- 
gregation, if  of  sufficient  size,  will  sink  to  the  bottom,  forming  a 
precipitate." 

In  regard  to  the  mutual  influences  exerted  upon  each  other  when 
two  colloids  are  mixed,  it  has  been  shown  by  Biltz,  Hardy,  and 
many  other  observers  that  oppositely  charged  'colloids  precipitate 
each  other,  though  this  is  not  an  absolute  rule,  as  experiments  by 
Professor  Stewart  Young,  of  Stanford,  have  shown.  Thus  colloidal 
gold  and  platinum  will  be  precipitated  by  such  colloids  as  ferric 
oxid  or  aluminium  oxid.  When  such  a  precipitating  colloid  is 
added  to  another  oppositely  charged  suspension  in  quantities  too 
small  to  bring  about  flocculation,  moreover,  the  addition  of  a  quan- 
tity of  salt,  likewise  too  small  to  precipitate  alone,  will  in  many  cases 
bring  about  the  flocculation. 

These  and  other  phenomena  of  colloidal  reaction  have  found 
close  analogy  in  antibody-antigen  studies,  and  have  given  support  to 
the  interpretation  of  the  latter  in  the  sense  of  Bordet. 

To  return  to  the  consideration  of  bacterial  agglutination,  we  have 
spoken  of  the  dependence  of  the  reaction  upon  the  presence  of  salts, 
and  have  seen  that  the  researches  of  Friedberger  and  others  have 
/  refuted  the  assumption  that  the  action  of  the  salt  in  bringing  about 
agglutination  depends  upon  chemical  union  of  the  salt  with  the 
bacteria.  It  is  probable,  therefore,  that  here,  as  in  other  colloidal 
precipitations,  the  function  of  the  salt  is  to  be  regarded  purely  as 
an  electrophysical  phenomenon. 

The  analogy  becomes  still  closer  when  we  consider  the  researches 
of  Bechold,66  Neisser  and  Friedemann,67  Sears  and  Jameson,68  and 
others,  which  have  shown  that  bacteria  in  suspension  are  to  be  com- 
pared very  closely  with  true  colloidal  suspensions  in  that  the  bac- 
terial cells  carry  a  definite  and  uniform  electrical  charge. 

Bacteria  in  salt  solution  emulsion,  for  instance,  wander  to  the 
anode,  thus  giving  evidence  of  their  carrying  a  negative  charge. 
This  charge  may  be  altered  by  adding  to  the  emulsions  definite  con- 
centrations of  acids  or  bases,  a  reversal  of  the  charge  taking  place 
under  the  influence  of  NaOH  or  other  hydroxids.  Just  how  this  is 
brought  about  is  by  no  means  clear,  but  it  is  not  impossible  that 
there  is  a  selective  absorption  of  OH  ions  by  the  bacteria,  which 
therefore  take  on  the  charge  of  the  ion. 

66  Bechold.     Zeitschr.  f.  physik.  Chem.,  48,  1904. 

67  Neisser  and  Friedemann.     Munch,  med.  Woch.,  Vol.  51,  pp.  465,  827 
1904. 

68  Sears  and  Jameson.    "Thesis  for  M.  A.  Stanford  University,"  1912. 


THE    PHENOMENON    OF    AGGLUTINATION 

However  this  may  be,  and  we  must  admit  that  explanations  of 
these  phenomena  are  as  yet  largely  speculative,  a  fact  which  interests 
us  particularly  in  connection  with  the  phenomena  under  discussion 
at  present  is  the  influence  exerted  upon  the  charge  of  bacteria  by 
exposure  to  the  influence  of  serum.  Bechold,69  as  well  as  Neisser 
and  Friedemann,70  assert  that  bacteria  which  have  absorbed  ag- 
glutinin  no  longer  wander  to  the  anode,  but  act  as  though  they  had 
been  deprived  of  electrical  charge,  and  precipitate  between  the  elec- 
trodes. 

Bechojo!  has  suggested,  for  this  reason,  that  it  may  be  possible 
that  bacteria  in  the  normal  condition  are  protected  from  the  action 
of  the  electrolyte  by  a  membrane  or  coating  of  protoplasm  which 
acts  as  a  protective  colloid.  The  absorption  of  agglutinin  may  alter 
this  in  such  a  way  that  they  become  amenable  to  the  flocculating 
effects  of  the  salt  ions.  In  keeping  with  such  an  opinion  is  the  well- 
known  observation  of  the  inagglutinability  of  capsulated  organisms, 
which,  as  Forges  71  has  shown,  become  agglutinable  as  soon  as  the 
capsules  have  been  destroyed  by  heating  in  a  weak  acid. 

That  the  absorption  of  agglutinin  indeed  alters  the  electric  sta- 
bility of  the  emulsified  bacteria  further  appears  from  the  fact  that 
"agglutinin"  bacteria  72  are  agglutinated  by  concentrations  of  salts 
which  are  too  slight  to  affect  the  normal  micro-organisms.  In  this 
respect  there  is  close  similarity  between  the  flocculation  of  agglutinin- 
bacteria  and  such  emulsions  as  kaolin  and  mastic,  whereas  bacteria 
without  agglutinin  require  much  higher  concentrations  of  the  salts 
to  produce  like  effects.  The  absorption  of  agglutinin  may  have  re- 
moved a  factor  which  protected  the  bacteria  against  the  influence  of 
the  salt.  On~fhe" other  hand,  it  is  equally  just  to  assume — and  this 
is  more  likely  and  corresponds  with  Eordet's  views — that  the  ab-\ 
sorption  of  agglutinin  has  "sensitized"  the  bacteria  to  the  action  of] 
the  electrolyte.  The  experimental  facts  upon  which  the  above  state- 
ments are  formulated  are  largely  found  in  the  important  papers  of 
K"eisser  and  Friedemann — whose  work  brought  out,  likewise,  inter- 
esting analogies  of  the  colloidal  precipitations  with  the  phenomenon 
which  we  have  described  above  as  the  proagglutinoid  zone. 

In  regard  to  the  greater  amenability  of  agglutinin  bacteria  to 
flocculation  by  electrolytes,  the  following  protocol,  adapted  from  the 
work  of  these  authors,  will  explain  itself.  They  were  tabulated  from 
experiments  in  which  different  quantities  of  normal  T  solution 
of  various  salts  were  added,  on  the  one  hand  to  emulsions  of  unal- 

69  Bechold.     "Die  Kolloide  in  der  Biologie  u.  Medizin,"  Steinkopf,  Dres- 
den, 1912. 

70  Neisser  and  Friedemann.     Munch,  med.  Wocli.,  Vol.  51,  1904,  pp.  465 
and  827. 

71  Forges.     Ztschr.  f.  exp.  Path.  u.  Therapie,  1905. 

72  "Agglutinin"  bacteria — bacteria  which  have  absorbed  specific  agglutinin. 


244 


INFECTION    AND    RESISTANCE 


tered  bacteria,  and,  on  the  other,  to  bacteria  which  had  absorbed 
agglutinin.  It  is  seen  that,  with  some  salts,  agglutination  of  the 
unaltered  bacteria  did  not  occur  at  all,  whereas  agglutination  was 
brought  about  in  the  treated  bacteria  with  comparatively  small 
amounts ;  in  other  cases  the  difference  is  a  quantitative  one  only : 

Protocol  constructed  from  the  tables  of  Neisser  and  Friedemann,  loc.  cit. 


y  solution  of  salt 

Quantity  of  salt  sol. 
which  brought  about 
agglutination  of 
1  c.  c.  of  normal  bacteria 
in  emulsion. 
0  =  no  agglutination 
by  the  salt 
solution 

Quantity  of  salt  sol. 
which  agglutinated 
1  c.  c.  of 
agglutinin  bacteria 
in 
emulsion 

NaCl.  . 

o 

025 

NaNOg  

o 

.025 

Na2S04  

0 

025 

Rbl  

o 

.025 

MgS04  

0 

.0025 

ZnS04  

01 

001 

CaCl2  

0 

.005 

BaCl2  .    .    . 

o 

005 

Cd(N03)2  

.01 

.001 

CuS04  

.0025 

.0001 

CuCl2  

0025 

.0005 

Pb(N03)2  

.0025 

.0001 

HgCL.. 

.0025 

.0005 

The  analogy  between  the  experiment  tabulated  in  the  preceding 
protocol  and  the  following  from  the  work  of  the  same  writers  is  self- 
evident.  Just  as  the  absorption  of  agglutinin  by  bacteria  rendered 
these  more  amenable  to  precipitation  by  salts,  so  the  addition  of 
minute  quantities  of  gelatin  to  mastic  emulsions  had  a  similar  sensi- 
tizing effect  upon  these. 


1  c.  c.  mastic 

1  o.  c.  mastic  -f  .0001  o. 

0. 

NaCl 
10%  solution 

(1-10  original  emulsion) 
diluted  to  3  c.  c. 

of  a  2%  gel.  solution, 
the  whole  diluted  to  3  c. 

c. 

1.0 

+  +  + 

+  +  + 

0.5 

0 

+  +  + 

0.25 

0 

+  +  + 

0.125 

0 

+  +  + 

0.05 

0 

0 

0.025 

0 

0 

Finally,  one  of  the  most  important  analogies  yielded  by  the  work 
of  the  above  investigators  is  illustrated  in  the  following  protocol : 


THE    PHENOMENON    OF    AGGLUTINATION 


Colloidal  iron 
hydroxid 

Precipitation 
emulsion 

of  mastic 
1  c.  c. 

1. 

0 

0.5 

0 

0.25 

0 

0.1 

+- 

h 

0.05 

+ 

0.025 

++ 

+ 

0.01 

++ 

+ 

0.005 

++ 

+ 

0.0025 

+- 

h 

0.001 

0 

Here  we  have  an  inhibition  zone  in  the  tubes  containing  the 
highest  concentrations,  accurately  analogous  to  the  previously  dis- 
cussed proagglutinoid  zone.  It  is  a  phenomenon  similar  also  to  the 
inhibition  zones  noticed  in  precipitin  reactions  and  observed,  though 
by  a  different  technique,  in  bacteriolytic  phenomena  discussed  in  an- 
other place  in  connection  with  the  Neisser-Wechsberg  notion  of  com- 
plement-deviation or  "Komplement  Ablenkung."  It  seems  to  be  a 
universal  fact  governing  the  union  of  colloidal  substances,  that  defi- 
nite quantitative  proportions  must  be  maintained  in  order  to  lead 
to  reaction,  this  being,  possibly,  explicable  on  the  basis  that  actual 
union  can  take  place  only  after  disturbance  of  the  electrical  balance 
which  keeps  the  particles  apart.  These  reactions  will  be  found  more 
accurately  discussed  in  another  place.  Whatever  the  mechanisms 
may  be,  however,  these  and  similar  experiments  have  seemed  to 
render  unnecessary  and  unlikely  the  assumption  of  proagglutinoids, 
proprecipitoids,  etc.,  to  explain  the  inhibition  zones  so  frequently 
observed  in  all  reactions  of  this  kind. 

A  peculiar  observation,  which  does  not  coincide  with  the  above 
interpretation  of  these  phenomena,  and  the  significance  of  which  is 
indeed  doubtful,  is  one  which  Friedberger  73  made  in  researches  in 
which  he  confirmed  the  work  of  Bordet  on  the  absence  of  agglutina- 
tion in  a  salt-free  environment.  He  found  that  not  only  the  addition 
of  various  salts  would  bring  about  agglutination  under  such  condi- 
tions, but  that  organic  substances  such  as  dextrose  and  asparagin 
could  be  substituted  for  salts  and  had  similar  agglutinating  effect — 
although  higher  concentrations  of  these  than  of  the  salts  were  re- 
quired. Were  these  substances  at  all  dissociable  it  might  be  pos- 
sible that  they  acted  by  a  mechanism  identical  with  that  of  the  salts — 
but  since  such  substances  as  dextrose  either  do  not  dissociate  at  all  or 
do  so  to  an  infinitesimal  degree  only  there  does  not  seem  any  pos- 
sibility of  reconciling  these  results  with  Bordet's  theory. 

73  Friedberger.     Centralbl.  f.  Bakt.,  30,  1901. 


246  INFECTION    AND    RESISTANCE 

It  is  difficult  to  explain  Friedberger's  results.  Possible  impurity 
of  his  preparations  and  the  presence  of  traces  of  electrolyte  seem 
to  be  excluded  by  the  fact  that  he  was  quite  conscious  of  this  possi- 
bility of  error  and  used  only  substances  which  yielded  no  ash  on 
combustion. 

It  may  be  that  the  results  of  Friedberger  in  which  glucose  and 
asparagin  were  used  may  have  brought  about  agglutination  by  an 
entirely  different  mechanism  from  that  which  we  are  discussing -and 
form  no  analogy  to  this. 

In  one  of  the  preceding  paragraphs  we  have  mentioned  the  phe- 
nomenon spoken  of  as  "acid  agglutination."  By  this  is  meant  the 
spontaneous  clumping,  not  only  of  bacteria,  but  of  small  particles  of 
any  kind,  in  suspension,  in  the  presence  of  certain  concentrations  of 
acid.  Michaelis,74  Beniasch,75  and  others  who  have  studied  this 
phenomenon  in  detail  have  come  to  the  conclusion  that  it  is  the  con- 
centration of  the  hydrogen  ions  which  is  responsible  for  the  ag- 
glutination. This  explanation  is  also  applicable  to  the  agglutination 
often  observed  about  the  anode  when  bacteria  are  subjected  in  sus- 
pension to  the  action  of  a  direct  current.  In  such  experiments  the 
organisms  after  concentrating  at  this  electrode  often  flocculate,  and 
it  is  here,  of  course,  that  hydrogen  ions  are  present  in  the  greatest 
concentration.  How  this  takes  place  is  problematical,  but  the  reason- 
ing of  Pauli,  if  applied  to  this,  would  favor  the  assumption  that  the 
weakly  charged  bacteria  group  themselves  about  the  ions  and,  when 
a  sufficiently  large  aggregation  has  formed,  fall  to  the  bottom  as 
precipitate.  This  phenomenon  of  acid  agglutination  is  of  course 
entirely  different  in  nature  from  the  specific  serum  agglutination 
which  we  are  discussing.  Nevertheless,  Schidorsky  and  Reim,76 
Jaffe,77  and  others  have  attempted  to  apply  acid  agglutination  to  the 
isolation  and  differentiation  of  bacteria,  on  the  conception  that  dif- 
ferent species  are  agglutinated  by  varying  concentrations  of  hy- 
drogen ions.  The  former  investigators,  even,  claim  to  have  been 
successful  in  isolating  typhoid  bacilli  from  the  stools  by  this  method 
in  that  the  typhoid  bacillus  was  agglutinated  by  concentrations  of 
acid  which  had  no  effect  upon  the  Bacillus  coli.  Sears  78  has  gone 
over  this  work  carefully,  and,  while  he  has  obtained  results  which 
bear  out  the  contention  that  the  agglutination  is  probably  due  to  the 
'concentration  of  the  H  ions,  his  experiments  have  revealed  an  irregu- 
larity in  the  behavior  of  bacteria  of  the  same  species  in  acid  solutions 
:and  an  overlapping  of  those  of  one  species  with  those  of  another. 
Therefore  the  use  of  acid  agglutination  for  differential  purposes 

74  Michaelis.    Folia  Serol,  7,  p.  1010,  and  Deutsche  med.  Woch.,  37,  969. 

75  Beniasch.     Zeitschr.  f.  Imm.,  Vol.  12,  1912. 

76  Schidorsky  and  Reim.     Deutsche  med.  Woch.,  Vol.  38,  p.  1125. 
*7  Jaffe.    Arch.  f.  Hyg.,  Vol.  76. 

78  Sears.    Proc.  Soc.  of  Exp.  Biol  and  Med.,  1913. 


THE    PHENOMENON    OF    AGGLUTINATION  247 

seems  to  us  entirely  hopeless.  And  indeed  it  would  be  surprising  if 
any  such  distinctive  and  regular  reaction  differences  between  simple 
bacterial  cells,  after  all  chemically  and  physically  so  essentially  alike, 
could  be  found. 


CHAPTER   X 

THE    PHENOMENON    OF   PRECIPITATION 

(Precipitins) 

THE  establishment  of  the  agglutinin  reaction  as  a  constant  and 
specific  serum-phenomenon  by  the  work  of  Gruber  and  Durham  led 
immediately  to  assiduous  investigation  of  the  many  problems  sug- 
gested by  it,  and  among  them,  as  we  have  seen,  the  question  of  the 
nature  of  the  agglutinogen.  It  was  found  that  agglutinins  could  be 
produced,  not  only  by  the  injection  of  whole  bacteria,  but  equally 
as  well  by  treatment  with  dissolved  bacterial  extracts  or  with  filtrates 
from  old  broth  cultures.  This  naturally  led  to  the  thought  that  there 
might  be  a  definite  reaction  if  such  extracts  (instead  of  the  bacteria 
themselves)  were  added  to  agglutinating  sera  in  vitro.  Rudolf 
Kraus  -1  was  the  first  to  perform  this  very  logical  experiment.  He 
was  working  with  broth  filtrates  of  Bacillus  pestis  and  of  the  cholera 
spirillum,  and  found  that  when  he  mixed  the  perfectly  clear  filtrates 
of  such  cultures  with  their  respective  antisera  the  mixtures  would  at 
first  become  turbid  and  finally  show  a  light  flocculent  precipitate.  He 
named  the  reaction  the  "precipitin  reaction"  and,  in  analogy  to 
agglutinins,  spoke  of  the  bodies  in  the  serum  which  caused  the  pre- 
cipitation as  "precipitins/"  The  reaction  was  found,  like  that  of 
agglutination,  to  be  specific;  the  cholera  serum  gave  no  precipitate 
with  the  plague  extract  and  vice  versa,  and  Kraus,  after  extending 
his  observations  to  other  bacteria,  pointed  out  the  practical  diagnostic 
possibilities  of  his  discovery. 

Though  Kraus7  first  observations  were  made  entirely  with  bac- 
terial culture  filtrates  and  antibacterial  sera,  it  was  soon  discovered 
that  his  results  were  merely  isolated  instances  of  a  broad  biological 
law,  and  that  specific  precipitins  were  produced  whenever  animals 
were  treated  with  injections  of  any  kind  of  foreign  protein.  Thus 
Tschistovitch,2  in  1899,  found  that  the  blood  serum  of  rabbits  im- 
munized with  eel-serum  gave  specific  precipitates  when  mixed  with 
eel-serum,  and  Bordet  3  obtained  analogous  results  by  treating  rab- 
bits with  defibrinated  chicken  blood  and  with  milk.  Thus  rapidly 

1  R.  Kraus.     Wien.  Idin.  Woch.,  No.  32,  1897. 

2  Tschistovitch.    Ann.  de  I'Inst.  Past.,  13,  1899. 

3  Bordet,    Ann.  de  I'Inst.  Past.,  Vol.  13,  1899,  pp.  225-273. 


THE   PHENOMENON    OF    PRECIPITATION          249 

the  discovery  of  Kraus  was  developed  into  the  generalization  that 
the  sera  of  animals  that  have  been  treated  with  foreign  proteins  of 
any  kind — bacterial,  animal,  or  vegetable — will  develop  the  property 
of  causing  precipitates  when  mixed  with  clear  solutions  of  the  re- 
spective antigens. 

The  substances  which,  after  injection  into  the  animal  body,  lead 
to  the  formation  of  precipitating  antibodies  are  spoken  of  in  the 
language  of  immunology  as  "precipitinogen."  In  the  case  of  bac- 
teria it  has  been  shown  that,  while  the  injection  of  the  whole  bac- 
terial cell — dead  or  alive — will  lead  to  precipitin  formation,  bacterial 
extracts  produced  in  a  variety  of  ways  will  lead  to  the  same  result. 
Such  precipitinogen  extracts  can  be  obtained  by  allowing  the  bacteria 
to  grow  in  flasks  of  slightly  alkaline  bouillon,  keeping  them  in  the 
incubator  for  from  three  weeks  to  three  months,  and  then  filtering 
them  through  Berkefeldt  candles.  Again,  useful  extracts  can  be 
more  rapidly  produced  by  growing  large  quantities  of  bacilli  on 
agar,  emulsifying  in  salt  solution,  and  shaking  in  any  one  of  the 
ordinary  types  of  shaking  machine  for  48  hours  or  longer.  On  filter- 
ing an  extract  is  obtained  which  will  form  precipitates  with  homol- 
ogous immune  serum,  or  will  incite  precipitins  when  injected  into 
animals.  In  fact,  any  one  of  the  customary  vigorous  methods  of 
extracting  bacterial  or  other  cells  will  yield  precipitinogen.  A  rela- 
tively purified  precipitinogen  in  the  form  of  a  dry,  water-soluble 
powder  has  been  obtained  by  Pick  by  the  precipitation  of  culture 
filtrates  with  alcohol. 

Regarding  the  chemical  nature  of  the  precipitin-inducing  sub- 
stances, or  precipitinogens,  the  same  problems  have  arisen  which 
have  been  discussed  in  connection  with  antigens  in  general.  We  may 
say  that  all  soluble  native  proteins  possess  precipitin-inducing  prop- 
erties. Yet  this  does  not  sufficiently  define  the  term,  since  many 
observations  have  been  published  which  show  that  physically  and 
chemically  altered  proteins  may  still  induce  specific  precipitins;  a 
few  investigators,  furthermore,  have  claimed  that  they  have  produced 
non-protein  precipitinogen  by  various  methods  of  breaking  up  the 
molecule  of  the  original  antigen.  In  the  section  on  agglutination 
we  have  seen  that  moderate  heating  (56-65°  C.)  rather  increases 
than  decreases  the  agglutinogen  characteristics  of  bacteria,  and  it 
is  equally  true  that  such  heated  bacteria  or  bacterial  extracts  may 
induce  precipitins.  However,  regarding  the  action  of  higher  de- 
grees of  heat  (boiling)  upon  precipitinogens  in  general  we  will  have 
more  to  say  in  another  place. 

Of  more  immediate,  indeed  of  fundamental,  importance  is  the 
problem  of  a  non-protein  antigen.  The  most  important  claims  in 
this  regard  have  been  made  by  Pick,4  Obermeyer  and  Pick,5  and  by 

4  Pick.    "Hof  meister's  Beitrage,"  Vol.  1,  1901. 

5  Obermeyer  and  Pick.    Wien.  klin.  Woch.,  1904,  p.  265. 


250  INFECTION    AND    RESISTANCE 

Jacoby.  6  7     Jacoby,  working  with  a  vegetable  antigen,  ricin,  fount 
that  by  trypsin  digestion  he  could  obtain  a  substance  which  still 
retained  antigenic  properties,  but  no  longer  gave  any  of  the  pro- 
tein reactions.     Obermeyer  and  Pick,  by  the  same  method,  claim, 
that  they  have  produced  a  non-protein  precipitinogen  from  egg  al- 
bumen.    On  the  other  hand,  others  have  had  negative  results,  and 
Kraus8  himself,  after  reviewing  the  evidence  on  both  sides,  comes 
to  the  conclusion  that  available  data  do  not  justify  us  in  separating 
^  the  antigenic  properties  from  the  protein  molecule.     In  unpublished 
^-experiments  which  the  writer  carried  on  in  the  laboratory  of  Profes-J 
sor  Friedemann  in  Berlin  also  attempts  to  produce  a  non-protein\ 
precipitinogen  from  horse  serum  by  bacterial  putrefaction  were  en-£ 
tirely  negative.     The  putrefaction  of  the  serum,  though  carried  out  <L 
in  dialyzing  bags  for  the  removal  of  diffusible  products,  was  ex-    .i 
tremely  slow,  and  when  finally  the  Biuret  reaction  disappeared  the      J 
serum  was  no  longer  precipitable  by  potent  antisera.     Hxpwever,  the       ) 
flaw  in  these  experiments  is  that  the  true  tesj_of  the  j)resence_pf 
precipitinogen  is  not  the  precipitable  character  of  the  solution  SPV 
question,  since  actual  precipitation  is  dependent,  as  we  shall  see, 
upon  many  modifying  secondary  factors,  but  rather  the  ability  o& 
the  substance  to  induce  precipitins  in  treated  animals. 

The  fact  that  ISTicolle,9  and  later  Pick,10  were  unable  to  obtain 
alcohol-soluble  substances  from  bacteria  and  bacterial  extracts  which 
were  still  precipitable  might  also  be  taken  to  point  toward  the  non- 
protein  character  of  the  precipitinogens,  suggesting  that  these  sub- 
stances may  be  of  a  lipoidal  nature.  However,  as  Landsteiner  11 
points  out,  mere  solubility  in  organic  solvents  can  no  longer  be  taken 
as  a  proof  of  lipoidal  character,  since  it  is  more  than  probable  that 
non-lipoidal  substances  may  go  into  alcoholic  and  other  organic  solu- 
tion when  lipoids,  such  as  lecithin,  are  present.  Thus  Miiller  12 
found  that  the  antigen  of  typhoid  bacilli  was  soluble  in  chloroform 
in  the  presence  of  old  preparations  of  lecithin.  Pick  and  Schwartz,13 
who  had  previously  studied  similar  antigen  solubilities  in  the  pres- 
ence both  of  lecithin  and  of  other  organ  lipoids,  suggest  that  possibly 
such  solutions  represent  lipoid-protein  combinations — colloidal  "so- 
lutions"— which  permit  the  presence  of  protein  mechanically  or 
chemically  united  to  the  lipoid  in  the  organic  solvents-^— alcohol, 
chloroform,  etc.  Here,  too,  then  there  is  no  evidence  for  the  ex- 
istence of  non-protein  precipitinogen. 

6  Jacoby.    "Hofmeister's  Beitrage,"  Vol.  1,  1901. 

7  Oppenheimer.     "Hofmeister's  Beitrage,"  Vol.  4,  1904,  p.  259. 

8  Kraus  in  "Kolle  u.  Wassermann  Handbuch,"  Vol.  4,  p.  605. 

9  Nicolle.     Ann.  de  I'Inst.  Past.,  12,  1898. 

10  Pick.    "Hofmeister's  Beitrage,"  Vol.  1,  1901. 

11  Landsteiner.     "Weichhardt's  Jahresbericht,"  Vol.  6,  1910,  p.  214. 

12  Miiller.     Zeitschr.  f.  Imm.,  Vol.  5,  1910. 

13  Pick  and  Schwartz.    Biochem.  Zeitschr.,  Vol.  15,  1909. 


THE   PHENOMENON    OF    PRECIPITATION          251 

Of  importance  in  connection  with  the  problem  of  the  nature  of 
precipitinogen,  also,  is  the  claim  of  Myers,14  that  specific  precipitins 
may  be  produced  in  rabbits  by  treatment  with  Witte  peptone,  a  sub- 
stance complex  in  constitution,  but  consisting' largely  of  albumoses. 
This  observation  has  failed  of  confirmation  in  the  hands  of  Ober- 
meyer  and  Pick,  Michaelis,15  ISTorris,16  and  others,  and  cannot,  there- 
fore, be  accepted  as  an  established  fact. 

Whichever  method  of  precipitinogen  production  is  used  bacterial 
precipitins  appear  in  the  serum  of  the  immunized  animal  only  after 
careful  and  continued  immunization,  usually  later  than  the  demon- 
strable appearance  of  the  bactericidal  or  agglutinating  properties  of 
the  serum.  The  most  convenient  material  for  such  immunization 
consists  of  salt  solution  emulsions  of  agar  cultures,  killed  at  60°  to 
TO0  C.  These  may  be  injected  subcutaneously,  intraperitoneally,  or 
intravenously,  the  last  method  leading  to  the  most  satisfactory  and 
rapid  results  and,  therefore,  best  employed  unless  great  inherent 
toxicity  of  the  particular  bacteria  contraindicates.  When  rabbits 
are  used  it  is  generally  necessary  to  inject  3,  4,  or  5  times  at  5  or  6- 
day  intervals,  and  to  bleed  the  animals  on  the  8th  or  9th  day  after 
the  last  injection. 

The  bacterial  precipitins  so  produced  are,  as  we  have  said  above, 
specific — but,  again,  specificity,  as  in  the  case  of  agglutinins,  is 
limited  by  the  so-called  "group  reactions."  In  the  chapter  dealing 
with  agglutination  we  have  seen  that  the  serum  of  a  typhoid-immune 
animal  which  agglutinates  typhoid  bacilli  strongly  will  also  aggluti- 
nate, though  far  less  powerfully,  paratyphoid  bacilli  and,  in  some 
cases,  even  colon  bacilli,  this  appearance  of  "minor"  agglutinins 
being  probably  due  to  a  close  group  relationship  of  these  bacteria  to 
the  typhoid  bacillus.  In  the  case  of  bacterial  precipitins  the  same 
thing  is  true,  and  has  been  made  the  subject  of  special  studies  by 
Zupnik,17  Kraus,18  Nor r is,19  and  others.  As  in  the  case  of  ag- 
glutination, however,  this  fact  does  not  in  any  way  interfere  with 
the  practical  value  of  the  specificity  of  the  reaction  because  elimina- 
tion of  the  secondary  group  reactions,  which  in  agglutination  is 
obtained  by  dilution  of  the  antiserum,  can  liere  be  obtained,  as  Kraus 
points  out,  by  diminishing  the  quantity  of  the  undiluted  precipitat- 
ing serum  added  to  the  bacterial  filtrates.  Thus,  while  one  volume 
of  serum  added  to  one,  two,  or  three  volumes  of  culture  filtrate  may 
still  give  error  due  to  non-specific  group  reactions,  a  proportion  of 

14  Myers.     Centralbl  f.  Bakt.,  Vol.  28,  1900. 

15  Michaelis.     Deutsche  med.  Woch.,  1902. 

16  Norris.     Jour,  of  Inf.  Dis.,  Vol.  1,  1904. 

17  Zupnik.     Zeitschr.  f.  Hyg.,  49,  1905. 

18  Kraus.     Wien.  klin.  Woch.,  1901,  No.  29. 

19  Norris.    Jour,  of  Inf.  Dis.,  Vol.  1,  1904. 


INFECTION    AND    RESISTANCE 

one  part  of  serum  to  8  or  10  parts  of  the  filtrate  will  usually  elimi- 
nate all  secondary  reactions  and  prove  strictly  specific. 

An  illustration  of  such  an  elimination  of  "partial"  or  "minor" 
precipitins  by  diminution  of  the  amount  of  the  homologous  anti- 
serum  is  given  in  the  following  table  taken  from  the  work  of 


ANTICOLI   RABBIT  SERUM 
TABLE   III 

The  precipitating  action  of  the  anticoli  rabbit  serum  upon  its  corresponding 
filtrates  and  upon  the  nitrates  of  B.  N°  1  (hog  cholera)  and  B.  typhosus. 

Coli  filtrate        Anticoli  serum 

0 . 5  c.  c.  0 . 05  Cloudiness  in  all  tubes  in  1  hour  at  37.5°  C.  which 

0.5  c.  c.  0.10  increases  rapidly.   Six  hours  well-marked  precipita- 

0.5  c.  c.  0. 15  tion — most  copious  in  tube  containing  0.25  serum. 

0.5  c.  c.  0.25  Fluid  in  all  tubes  becomes  clear. 

B.  N°l  filtrate    Anticoli  serum 

O.Sc.c.          0.10        At  6  hours  a  slight  precipitate  in  the  form  of  fine 
0 . 5  c.  c.          0 . 25  granules  appears  on  the  sides  of  the  tubes.    After 

24  hours  the  precipitate  in  the  tube  containing 
0.25  c.  c.  serum  compares  in  amount  to  that 
formed  in  the  homologous  filtrate  with  0.05  c.  c. 
of  serum. 

B.  typh.  (Coll) 

filtrate  Anticoli  serum 

O.Sc.c.  0 . 10        Similar  reaction  obtained  to  that  with  B.  N°  1  filtrate. 

0.5  c.  c.          0.25 

B.  typh.  (Pfeif- 
fer)  filtrate 

6.5  c.  c.  0.10        Similar  delay  in  reaction  as  obtained  with  B.  typh. 

O.Sc.c.          0.25  Coll. 

And,  indeed,  though  the  great  practical  value  of  the  precipitin 
reaction  has  not  been  in  the  special  field  of  bacteriology,  it  has  been 
successfully  utilized  in  the  diagnosis  of  glanders  by  Wladimiroff,21 
and  constitutes  a  valuable  adjuvant  to  our  methods  of  determining 
the  biological  relationship  between  micro-organisms. 

The  production  of  precipitins  against  unformed  proteins,  egg 
albumen,  animal  sera,  etc.,  is  much  more  easily  accomplished  than 
the  production  of  bacterial  precipitins,  and  three  intravenous  injec- 
tions of  from  2  to  5  c.  c.  of  the  protein  at  5  or  6-day  intervals  usually 
give  rise  to  a  formation  of  potent  precipitins.  When  a  small  quan- 
tity of  the  serum  of  such  an  animal,  taken  9  or  10  days  after  the 
third  injection,  is  mixed  in  a  test  tube  with  an  equal  quantity  of  a 

20  Norris.    Joyr.  of  Inf.  Dis.,  Vol.  1,  1904,  p.  472. 

21  Wladimiroff.   "Kolle  u.  Wassermann  Handbueh  "  article  on  "Glanders." 
Vol.  5,  2d  Ed. 


THE    PHENOMENON    OF    PRECIPITATION 


253 


dilution  of  the  protein  which  was  injected,  turbidity  and  rapid  floccu- 
lation  will  result.  In  tests  of  this  kind,  unlike  the  bacterial  precipitin 
tests  in  which  the  delicacy  of  the  reaction  is  ordinarily  determined 
by  diminution  of  the  amounts  of  antiserum,  the  same  object  may  be 
more  conveniently  attained  by  dilution  of  the  antigen.  Thus,  in  test- 
ing the  precipitating  potency  of,  let  us  say,  the  serum  of  a  rabbit 
immunized  with  sheep  serum,  we  would  proceed  by  setting  up  a 
series  of  small  tubes,  each  of  which  contains  a  constant  amount  of 
antiserum  (precipitin),  but  a  progressively  diminishing  amount  of 
antigen  in  the  same  volume — i.  e.,  in  dilution  with  iso tonic  salt  solu- 
tion. The  following  example  will  make  this  clear : 


Anti  sheep  serum 
from  rabbit 

Sheep  serum  0.5  c.  c. 
of  following  dilutions: 

Precipitation 

0.5  c.  c. 

+ 

1:10 

=t 

0.5  c.  c. 

+ 

1:100 

+  +  + 

0.5  c.  c. 

-j- 

1:500 

+  +  + 

0.5  c.  c. 
0.5  c.  c. 
0.5  c.  c. 

+ 
+ 
+ 

1:1,000 
1:5,000 
1:10,000 

— 

In  this  test  it  will  be  noticed  that  the  strongest  concentration  of 
the  antigen  (1:10)  gave  a  relatively  slight  precipitation  only.  This 
phenomenon  is  analogous  to  the  inhibition  zones  noticed  in  the  case 
of  agglutination  and  other  antibody  reactions  and  will  be  more  espe- 
cially discussed  in  a  succeeding  paragraph. 

The  delicacy  of  the  above  example,  moreover,  is  by  no  means 
unusual,  and  sera  have  been  obtained  by  careful  immunization  with 
which  the  specific  antigen  could  be  detected  in  dilutions  as  high  as  1 
to  100,000  (Uhlenhuth).  A  serum  which  will  react  with  antigen 
dilutions  of  1  to  10,000  and  1  to  20,000  is  not  at  all  uncommon  nor 
difficult  to  obtain.  Apart  from  the  additional  advantage  of  the 
specificity  of  the  reaction,  therefore,  this  biological  method  of  de- 
tecting proteins  is  more  delicate  than  that  of  any  of  the  known 
chemical  methods;  neither  the  Biuret  nor  Millon's  reaction  will  far 
exceed  a  delicacy  of  1  to  1,000.  By  a  modification  known  as  the 
method  of  Complement  or  Alexin-fixation,  furthermore,  the  delicacy 
of  the  biological  reactions  can  be  still  further  enhanced.  This  is 
discussed  in  detail  in  another  place  (see  page  212). 

The  practical  value  of  the  precipitin  reaction,  however,  lies,  not 
in  the  mere  detection  of  protein,  but,  by  virtue  of  its  specificity,22  in 
the  determination  of  the  variety  of  protein  under  examination.  And 

22  Wassermann  and  Schiitze.  Deutsche  med.  Woch.,  1900,  Vereinsbeilage, 
p.  178;  Berl.  kl.  Woch.,  1901;  Deutsche  med.  Woch.,  1902;  Bordet,  Ann.  Past., 
Vol.  13,  1899;  Nolf,  ibid.,  Vol.  14,  1900;  Fish,  Medical  Courier,  St.  Louis, 
1900,  cited  from  Wassermann. 


254  INFECTION    AND    RESISTANCE 

here  again  the  specificity,  like  that  of  bacterial  precipitation,  ag- 
glutination, and  other  serum  tests,  is  relative  rather  than  absolute. 
Thus  a  serum  which  has  been  obtained  by  the  immunization  of  an 
animal  with  human  serum  may  react,  not  only  with  human  serum, 
but  also  with  relatively  higher  concentrations  of  the  sera  of  some  of 
the  higher  apes.  However,  such  non-specific  partial  reactions  can  be 
eliminated  entirely  by  employing  higher  dilutions  of  antigen.  Thus 
Uhlenhuth,23,  24?  25  on  the  basis  of  a  large  experience,  has  established 
a  -standard  of  antigen  dilution  at  1  to  1,000,  beyond  which  no  "para" 
or  "minor"  precipitation  will  occur.  Since  potency  far  exceeding 
this  is  easily  procured,  absolute  specificity  can  be  ensured  by  the 
very  simple  precaution  of  a  sufficient  dilution. 

The  most  important  practical  use  for  the  reaction  has  been  found 
in  forensic  medicine,  where  it  is  possible  in  this  way  to  determine 
the  species  of  animal  from  which  have  emanated  the  blood,  sperm, 
etc.,  found  in  spots  on  wearing  apparel,  weapons,  or  other  articles. 
The  extensive  investigations  of  Nuttall  26  upon  this  subject  have  inci- 
dentally been  of  much  value  in  furnishing  a  further  method  for  the 
determination  of  zoological  species  relationships.  Nuttall  carried 
out  16,000  precipitin  tests,  with  precipitating  sera,  upon  900  speci- 
mens of  blood  which  he  obtained  from  various  sources.  He  not  only 
confirmed  many  of  the  accepted  zoological  classifications,  but  shed 
much  light  upon  a  number  of  disputed  points.  In  working  out  the 
tests  upon  monkeys  he  found  that  the  reactions  carried  out  with  anti- 
human  serum  become  weaker  as  the  species  examined  is  farther  re- 
moved from  man  zoologically.  Thus  as  we  read  down  the  column 
from  man  to  the  hapalidsB  the  precipitate  becomes  less  and  less  in 
amount. 

Nuttall's  Tests  with  Antihuman  Serum.     (Nuttall,  loc.  cit.,  p.  165.} 

ANTIHUMAN   PRECIPITATING   SERUM 
Tested  against  Precipitate 

34  Specimens  human  blood 100%27 

8  Simiidse,  3  species 100% 

36  Cercopithecidae 92% 

13  Cebidse 78% 

4  Hapalidse 50% 

2  Lemuridse 0 

23  Uhlenhuth.     Deutsche  med.  Woch.,  1900,  1901;  Rob.  Koch  Festschrift, 
1903. 

24  Uhlenhuth  and  Weidanz.     "Kraus  u.  Levaditi  Handbuch,"  etc.,  Vol.  2, 
1909. 

25  Uhlenhuth  and  Weidanz.     Loc.  cit.,  where  other  publications  are  sum- 
marized. 

26  Nuttall.     "Blood  Immunity  and  Blood  Relationship,"  Cambridge  Uni- 
versity Press,  1904. 

27  The  percentages  refer  to  the  volume  of  precipitate  formed  on  standing 
for  a  given  time,  the  amount  formed  by  the  antiserum  with  its  specific  antigen 
being  taken  as  100  per  cent.     Antigen  dilutions  correspond  throughout. 


THE    PHENOMENON    OF    PRECIPITATION  255 

In  another  series  he  finds: 

ANTIHUMAN   PRECIPITATING   SERUM 

Tested  against  Precipitate 

Man 100% 

Chimpanzee (loose  precip.)  130% 

Gorilla 64% 

Ourang 42% 

Cynocephalus  mormon 42% 

Cynocephalus  sphinx 29% 

Ateles 29% 


Among  the  primates  the  highest  figures  with  antihuman  serum  are 
given  by  the  chimpanzee.  Other  bloods  than  those  of  the  primates 
gave  slight  reactions  or  none  whatever  with  the  antihuman  serum. 

In  addition  to  these  results  the  relationships  within  the  dog 
family,  the  horse  family,  and  many  other  kinships  similar  to  these 
were  confirmed.  In  every  case  the  precipitin  reaction  was  con- 
sistent with  the  results  of  other  methods  of  classification,  and  Nut- 
tali's  work  is  an  extremely  valuable  aid  to  zoologists  in  disputed 
questions  of  animal  relationships. 

These  facts  are  the  more  surprising  in  that  they  demonstrate 
species  differences  between  the  proteins  of  various  animals  which 
are  not  determinable  by  known  chemical  methods.  How  funda- 
mental these  differences  are  and  how  delicate  the  reaction,  is  further 
shown  by  experiments  of  Uhlenhuth,  in  which  he  obtained  a  specific 
antihare  serum  by  treating  rabbits'  with  hares'  blood,  an  astonishing 
result  in  view  of  the  close  zoological  relations  between  these  animals. 

Isoprecipitins,  that  is,  precipitins  resulting  from  the  treatment 
of  animals  with  blood  of  another  individual  of  the  same  species,  have 
also  been  described  by  Schiitze  and  others.  They  are  not,  however, 
regular  in  their  appearance,  nor  are  they  very  potent  when  obtained. 

Since  the  reaction  is  equally  applicable  to  vegetable  proteins, 
similar  investigations  on  the  interrelationship  of  different  varieties , 
of  wheat  have  been  carried  out  by  Magnus.28 

The  methods  of  performing  precipitin  tests  for  forensic  or  other 
purposes  is  extremely  simple.  Nevertheless,  there  are  a  number  of 
theoretical  considerations  which  we  must  take  up  in  order  to  make 
clear  the  limitations  of  accuracy  and  conditions  of  control  which  are 
involved  in  these  reactions.  From  our  discussion  of  the  nature  of 
precipitinogen  it  follows  that  blood  stains,  etc.,  on  linen  or  articles 
of  any  kind  will  be  suitable  for  precipitin  tests  even  after  they  have 
been  exposed  for  considerable  periods  to  unfavorable  conditions,  that 
is,  an  environment  in  which  they  are  subjected  to  exposure  to  light, 
moderate  heat,  or  drying.  Thus  blood  spots,  etc.,  if  kept  dry  and  in 

28  Magnus.     Cited  from  Uhlenhuth,  loc.  cit. 


£56  INFECTION    AND    RESISTANCE 

the  dark,  may  give  positive  reactions  even  after  years,  as  experi- 
ments by  Uhlenhuth  have  shown.  Meyer  29  claims  even  to  have 
obtained  a  precipitation  with  extracts  of  the  material  of  mummies. 
One  of  his  specimens  was  a  mummy  dating  back  to  the  first  Egyptian 
Empire  (5,000  years),  the  other  about  2,000  years  old.  Pieces  of 
the  leg  and  neck  muscles  of  these  specimens  were  chopped  up  finely, 
extracted  for  24  hours  with  salt  solution,  then  filtered  until  clear. 
With  antihuman  serum  they  gave  turbidity  after  one  hour  at 
37.5°  C. 

Under  conditions  of  putrefaction,  of  course,  the  precipitinogen 
is  more  rapidly  destroyed,  though  blood  putrefies  with  surprising 
slowness,  even  if,  as  in  our  own  experiments,  the  conditions  of  mois- 
ture, temperature,  and  reinoculation  with  putrefactive  bacteria  are 
constantly  observed.  Under  such  conditions  a  weak  reaction  may  be 
obtained  after  as  long  as  a  month  or  six  weeks. 

In  carrying  out  the  tests  with  any  material  it  is  first  necessary 
to  get  it  into  clear  solution,  a  result  which  is  best  accomplished  by 
soaking  it  in  a  small  quantity  of  isotonic  salt  solution.  Preliminary 
to  this  it  is  always  necessary  to  scrape  off  a  bit  of  the  specimen  and 
examine  it  microscopically  to  discover,  if  possible,  whether  blood  cells, 
sperm,  or  other  cellular  constituents  can  be  detected.  The  infusion  in 
salt  solution  should  be  continued  for  several  hours — if  necessary  for 
12  to  24  hours.  After  the  first  few  hours  in  the  incubator  the  material 
should  be  placed  at  room  or  refrigerator  temperature  so  that  the 
yield  in  unchanged  protein  may  not  be  diminished  by  the  action  of 
bacterial  growth.  After  extraction  the  solution  may  be  filtered  in 
order  to  clear  it,  but  often  mere  centrifugation  suffices  for  this  pur- 
pose. The  concentration  of  antigen  in  such  an  extract  is  always  an 
uncertainty,  but  may  be  determined  with  sufficient  accuracy  for 
practical  purposes  by  shaking  and  observing  the  formation  of  a 
lasting  foam.  Protein  solutions  will  show  foam  on  shaking  in  dilu- 
tions as  high  as  1  to  1,000,  and  if  the  original  amount  of  salt  solution 
used,  in  washing  out  the  material  is  properly  gauged  to  the  amount 
of  blood  available  in  the  stain,  and  the  solution  shaken  and  observed 
for  the  formation  of  foam,  it  is  usually  a  simple  matter  to  obtain  a 
final  concentration  approximating  one  to  one  thousand.30 

The  antiserum  which  is  used  should  be  of  such  a  potency  that 
preliminary  titration  with  the  specific  antigen,  diluted  1  to  1,000, 
should  give  an  almost  immediate  cloudiness  at  room  temperature. 

By  testing  this  serum  against  a  number  of  other  varieties  of 

29  Meyer.     Munch,  med.  Woch.,  Vol.  51,  No.  15,  1904. 

30  If  there  is  enough  material,  .the  amount  of  dissolved  protein  can  be 
also  approximately  gauged  by  adding  to  a  little  of  it  a  drop  of  acid,  boiling 
and  observing  the.  heaviness  of  the  cloud  which  forms.     A  control  test  of  a 
known  dilution  of  the  suspected  variety  of  blood  can  be  made  at  the  same 
time  and  the  heaviness  of  this  cloud  compared  with  that  in  the  test  solution, 


THE    PHENOMENON  OF    PRECIPITATION  257 

protein — dog  serum,  beef  serum,  etc. — it  must  be  determined  that 
the  precipitin  in  this  case  is  strictly  specific. 

The  reaction  can  be  observed  with  greater  delicacy  if  it  is  first 
set  up  by  the  method  recommended  by  Fornet  and  Miiller,31  which 
we  may  speak  of  as  the  "ring  test."  The  antiserum  is  put  into  the 
tubes  and  the  solution  to  be  tested  is  allowed  to  flow  slowly  over  this 
—as  in  Heller's  nitric  acid  albumin  test.  -  At  the  line  of  contact  be- 
tween the  two  a  fine  white  ring  will  rapidly  appear,  thickening  and 
growing  heavier  as  the  preparation  is  allowed  to  stand.  After  taking 
the  final  readings  from  such  a  test,  let  us  say  after  an  hour  or  so,  it 
is  well  to  shake  up  the  tubes,  set  them  away  in  the  ice-chest,  and  again 
read  the  amount  of  precipitates  formed  in  the  various  tubes  the  next 
morning.  Since  every  test  of  this  kind  necessitates  a  number  of 
controls,  the  following  example  will  serve  as  a  basis  for  discussion : 

Forensic  Blood  Examination 

Material:  Blood  spot  on  trouser  pocket,  washed  up  in  salt  solution.  Clear 
after  paper  filtration. 

Antiserum:  Rabbit  treated  with  three  intravenous  injections,  2,  5,  and  5  c.  c. 
of  human  serum  at  six-day  intervals;  bled  on  tenth  day  after  last  injection. 
This  serum  has  been  titrated  against  human  serum  and  gives  precipitation 
in  dilutions  up  to  one  to  ten  thousand.  With  one  to  one  thousand  there  is 
clouding  which  begins  in  three  minutes  and  is  very  distinct  in  eight  minutes, 
at  room  temperature.32 

Test  Jtw**** 

Tube  1.  Known  human  serum  1  to  1,000.  .   1.0  c.  c.  +  Antiserum 0.2  c.  c.    f  u 

Tube  2.  Unknown  solution  to  be  tested. ...   1.0  c.  c.  +  Antfserum 0.2  c.  c. 

Tube  3.  Unknown  solution  to  be  tested 1.0  c.  c.  +  Normal   rabbit 

serum 0.2  c.  c. 

Tube  4.  Salt  solution 1.0  c.  c.  +  Antiserum ....  0.2  c.  c.      — 

Tube  5.  Unknown  solution 1.0  c.  c.  -f-  Salt  solution. .  .0.2  c.  c.     — 

In  this  test,  if  the  original  material  was  human  blood,  tubes  1 
and  2  should  show  ring  formation  within  5  minutes — while  the 
other  tubes  remain  clear.  In  addition  to  these  controls  it  is  well  to 
be  sure  that  the  test  extract  is  neither  strongly  acid  nor  alkaline,  and 
that,  as  Uhlenhuth  suggests,  the  material  from  which  it  is  extracted 
does  not  contain  other  substances  which  can  give  precipitates  by 
themselves  when  added  to  serum.  This  is  especially  necessary  in  the 
case  of  cloth  fabrics,  and  a  recent  instance  in  our  own  experience 
has  suggested  to  us  the  possibility  that  such  materials  may  also  con- 
tain colloidal  dye  stuffs  or  other  extractable  substances  which  can 
cause  inhibition  of  the  precipitation.  In  an  apparently  positive 
case  the  reactions  with  a  blood  extract  from  trouser  cloth  were  suffi- 
ciently heavy,  but  regularly  delayed,  as  in  the  flocculation  of  such 

31  Fornet  and  Miiller.     Zeitschr.  f.  Hyg.,  Vol.  66,  1910. 

32  A  mixture  of  too  specific  antisera  should  never  be  used,   since  such 
sera  may  often  precipitate  each  other  for  reasons  that  are  discussed  below. 


258  INFECTION    AND    RESISTANCE 

colloidal  suspensions  as  arsenic  trisulphide  in  the  presence  of  a 
protective  colloid. 

In  the  ordinary  criminal  or  civil  case  which  would  come  under 
consideration  for  precipitin  tests  the  spots  or  stains  are  made  by 
blood  as  it  flows  from  the  wound  and  unchanged  by  chemical  or 
physical  agencies  except  as  these  are  encountered  afterward,  by 
exposure.  In  the  case  of  meat  inspection,  in  which  the  precipitin 
test  is  useful  in  detecting  admixtures  of  horse  flesh,  dog  flesh,  or 
other  less  desirable  varieties  of  meat,  in  sausages,  chopped  meat, 
etc.,  it  often  happens  that  such  procedures  as  heating  or  smoking 
may  vitiate  the  results  of  precipitin  reactions.  It  is  of  practical  im- 
portance, therefore,  that  we  should  know  exactly  what  the  effects  of 
heating  (boiling)  may  be  upon  precipitinogen.  Moreover,  this  ques- 
tion possesses  considerable  theoretical  interest  since  the  coagulation 
of  proteins  by  heat  seems  to  involve  chiefly  a  physical  rather  than  a 
chemical  change. 

Cohnheim  33  says  in  discussing  this  question :  "It  is  still  unclear 
what  the  changes  are  that  take  place  in  coagulation.  It  may  be  that 
there  is  merely  an  intramolecular  'Umlagerung' — or  there  may  be 
cleavage ;  or  the  process  may  be  comparable  to  the  flocculation  of  col- 
loidal clay  emulsions  by  salts.  ..  .  .  With  coagulation  all  proteins 
have  lost  the  differences  which  they  possess  in  the  native  state  in 
respect  to  solubility  or  precipitability  by  salts.  Physically  all  coagu- 
lated proteins  are  alike ;  they  are  no  longer  native  proteins,  and  with- 
out further  decomposition  are  insoluble.  Chemical  differences,  how- 
ever, variations  of  composition,  and  the  cleavage  products  which 
they  yield  still  distinguish  them." 

The  question  has  been  experimentally  approached  by  Obermeyer 
and  Pick  34  in  connection  with  their  general  investigations  upon  the 
influence  of  chemical  and  physical  alterations  upon  precipitinogen. 
They  found  that  precipitin  produced  with  unchanged  (native)  beef 
serum  does  not  react  with  heated  beef  serum,  even  if  immunization 
was  prolonged  and  a  very  potent  serum  was  produced.  On  the  other 
hand,  when  animals  were  immunized  with  beef  serum  which  had 
been  boiled  for  a  short  time  ("Kurz  aufgekocht"35)  the  precipitin 
so  produced  reacted,  not  only  with  native  beef  serum,  but  also  pre- 
cipitated the  boiled  serum  and  a  whole  row  of  split  products  which 
give  no  reaction  to  normal  precipitin.  The  "coctoprecipitin"  so 
produced,  furthermore,  was  found  by  them  to  be  specific,  acting 
only  upon  beef  protein  or  its  derivatives. 

33  Otto  Cohnheim.     "Chemie  der  Eiweiss  Korper  Vieweg  Braunschweig," 
1900,  p.  8. 

34  Obermeyer  and  Pick.     Wien.  kl  Wocli.,  12,  1906. 

35  Sera  or  other  proteins  used  in  such  tests  are  boiled  in  dilutions  of  1 
to  10  or  more,  in  order  to  avoid  the  formation  of  heavy  flakes  which  cannot 
be  injected.    Boiled  in  sufficient  dilution,  an  opalescent  suspension  is  formed 
which  easily  passes  through  the  syringe. 


THE    PHENOMENON  OF    PRECIPITATION  259 

It  is  immediately  evident  that  these  investigations  are  closely 
analogous  to  those  of  Joos  and  others  on  the  agghitinins.  The  anti- 
serum  produced  with  the  heated  antigen  here  again  reacts  both  with 
the  native  and  with  the  heated  antigen,  whereas  the  antiserum  pro- 
duced with  the  native  unheated  antigen  reacts  only  with  the  un- 
heated.  The  "heat-precipitins"  therefore  may  be  also  called  "um- 
fanglicher" — the  term  applied  by  Paltauf  to  the  agglutinins  pro- 
duced with  heated  bacteria. 

Schmidt,36  who  has  studied  the  problem  extensively,  finds  that 
heating  serum  protein  to  70°  C.  for  as  long  as  30  to  60  minutes 
alters  its  precipitability  by  "native  precipitin"  (precipitin  produced 
by  immunization  with  native  unheated  serum)  only  in  so  far  as  it 
diminishes  the  delicacy  of  the  reaction  by  10  to  30  per  cent.,  and' 
that  heating  to  90°  C.  for  as  long  as  an  hour  does  not  render  it  en- 
tirely non-precipitable,  so  that  protein  so  treated  may  yet  be  detect- 
able by  ordinary  specific  precipitins  produced  by  injections  of  un- 
heated serum,  though  the  delicacy  of  the  reaction  is  lessened.  Boil- 
ing, according  to  Schmidt,  renders  the  antigen  no  longer  precipitable 
by  such  "native  precipitin,"  but,  on  the  other  hand,  it  does  not  seem 
to  destroy  its  antigenic  property  of  inciting  precipitins  on  injection 
into  animals.  Fornet  and  Miiller,  on  the  other  hand,  claim  that  even 
boiled  protein  can  be  detected  by  "native  precipitins,"  though  the 
reaction  is  only  about  one-tenth  as  delicate  as  it  is  with  unheated 
protein. 

Schmidt  studied  these  relations  especially  as  they  affect  -the  per- 
formance of  specific  precipitin  reactions  in  the  identification  of 
boiled  meat.  He  found  that  when  he  immunized  rabbits  with  serum 
protein  that  had  been  heated  at  TO0  C.  for  30  minutes  the  antiserum 
so  obtained  gave  strong  and  practically  useful  reactions  with  its 
specific  antigen  even  if  this  had  been  boiled.  Since  "native  pre- 
cipitin" gives  weak  reactions  only  with  such  a  boiled  protein, 
Schmidt  recommends  the  use  of  the  "70°  precipitin"  (produced  by 
injections  of  heated  serum)  for  tests  in  which  a  heated  antigen  is  to 
be  identified. 

He  states,  however,  that  very  prolonged  heating  may  so  com- 
pletely coagulate  the  antigen  that  none  of  it  can  be  gotten  into  "solu- 
tion" (suspension),  and  in  such  cases  results  can  be  obtained  neither 
with  the  "native"  nor  with  the  "70°  precipitin."  He  has  at- 
tempted, therefore,  to  find  a  method  whereby  even  such  entirely 
insoluble  proteins  may  be  identified,  and  claims  to  have  succeeded 
by  preparing  what  he  calls  his  "heat-alkali-precipitin."  37  He  di- 

36  Schmidt.     Biochem.  Zeitschr.,   14,   1908;   also   Zeitschr.  f.  Imm.,  Vol. 
13,  1912. 

37  "Native  precipitin"  =  preeipitin  produced  by  injections  of  normal  un- 
heated serum. 

"70°  precipitin"  —  precipitin  produced  by  injections  of  serum  heated 
to  70°  C.  for  30  minutes. 


260 


INFECTION    AND    RESISTANCE 


lutes  serum  with  equal  parts  of  isotonic  salt  solution  and  heats  it  to 
70°  C.  for  30  minutes  in  a  water  bath.  To  60  c.  c.  of  such  a  solution 
he  now  adds  10  c.  c.  of  f  NaOH,  and  continues  heating  for  15  to 
20  minutes.  At  the  end  of  this  time  he  neutralizes  with  HC1,  cools, 
and  injects  20  c.  c.  intraperitoneally  into  rabbits.  (The  neutraliza- 
tion is  not  absolutely  necessary.)  Five  or  more  injections. yield  a 
serum  sufficiently  potent  for  use. 

A  precipitin  so  produced  will,  according  to  Schmidt,  react  spe- 
cifically with  heated  proteins,  and  also  with  protein  which  has  been 
solidly  coagulated  and  brought  into  solution  by  means  of  NaOH  and 
heat.  It  will  not,  however,  react  with  normal  unheated  antigen. 

He  tested  this  by  coagulating  horse  serum  by  boiling  for  3  hours. 
The  coagulum  was  washed  with  salt  solution,  dried,  and  powdered. 
Tests  were  then  made  to  prove  that  this  powder  was  entirely  in- 
soluble in  NaCl  solution.  A  little  of  it  was  then  treated  with  10 
c.  c.  of  salt  solution  containing  enough  NaOH  to  correspond  to  an 
^  solution.  The  exposure  was  continued  for  20  minutes  in  a  water 
bath  at  60°  to  70°  C.  Before  the  entire  mass  was  dissolved  the  solu- 
tion was  filtered  and  neutralized  with  -^  HC1. 

The  rather  complicated  relations  described  by  Schmidt  are  easily 
surveyed  in  the  following  protocol  taken  from  his  work : 


TABLE   I 

(W.  A.  Schmidt,  Zeitschr.  /.  7mm.,  Vol.  13,  1912,  p. 


173) 


Solution 
of 

Native 
precipitin 

Heat  (70°) 
precipitin 

Heat-alkali- 
precipitin 

Native  serum  

Strong  reaction 

Good  reaction 

0   (very  slight 

70°  serum  (heated  30  min.) 
100°    serum    (heated    30 
min  )                       

Good  reaction 
0 

Strong  reaction 
Good  reaction 

turbidity) 
Strong  reaction 

Strong  reaction 

70°    serum   treated   with 
NaOH  (used  to  produce 
heat-alkali-precipitin)  . 
Boiled    insoluble    serum, 
brought    into    solution 
with  NaOH   

0 
0 

0 
0 

Strong  reaction 
Good  reaction 

.Native  serum  treated  with 
«    NaOH  in  the  cold  

0 

o 

Good  reaction 

Schmidt  speaks  of  the  "heat-alkali-precipitin"  also  as  "alkali- 
albuminate-precipitin."  It  can  be  produced  only  if  the  NaOH  treat- 
ment of  the  serum  is  cautiously  performed.  If  the  sodium  hydroxid 
is  allowed  to  act  too  vigorously  in  strong  concentrations  or  for  too 
long  a  time  the  antigen  is  completely  destroyed,  is  no  longer  pre- 


THE    PHENOMENON    OF    PRECIPITATION         261 

cipitable,  and  no  longer  produces  precipitin  when  injected  into 
animals. 

The  striking  feature  of  these  experiments  is  that  they  show  a 
gradual  alteration  of  the  protein  first  by  heat,  then  by  alkali  and 
heat,  in  such  a  way  that  the  antigenic  properties  are  changed  but 
not  destroyed.  Each  precipitin,  moreover,  seems  to  react  most 
strongly  with  the  particular  antigen-alteration  which  produced  it, 
and,  according  to  Schmidt,  retains  its  species  specificity.  This  is 
not  the  case  with  the  iodized  proteins  and  nitroproteins  and  diazo- 
proteins  produced  by  Obermeyer  and  Pick.38  Here  iodized  beef 
protein  injected  into  animals  produced  a  precipitin  which  reacted 
with  the  iodized  protein,  not  only  of  the  beef,  but  also  similarly 
altered  proteins  of  other  animals — and  the  same  was  true  of  the 
nitro  and  diazo  modifications. 

Although  the  experiments  of  Schmidt  have  great  theoretical 
value,  their  practical  utilization  must  depend  upon  the  degree  of 
specificity  possessed  by  the  heat-precipitins  or  the  heat-alkali-pre- 
cipitins.  In  Obermeyer  and  Pick's  original  investigations  we  have 
seen  that  they  found  the  precipitin  produced  with  heated  serum  as 
strictly  specific  as  that  induced  by  native  serum.  This  has  also  been 
the  experience  of  Schmidt.  Fornet  and  Miiller,39  on  the  other  hand, 
report  that  the  precipitins  produced  by  them  with  heated  muscle- 
protein  were  not  as  strictly  specific  as  those  produced  with  the  un- 
heated — in  that  the  former  gave  precipitates,  not  only  with  homol- 
ogous protein  solutions,  but  with  foreign  proteins  in  moderate  con- 
centration as  well.  In  experiments  carried  out  by  the  writer  with 
Ostenberg  40  it  was  attempted  to  determine  whether  or  not  precipi- 
tins could  be  produced  by  injecting  animals  with  protein  that  had 
been  boiled,  and  if  so  what  the  action  of  these  substances  would  be 
upon  boiled  proteins.  Contrary  to  the  results  of  Fornet  and  Miiller, 
it  was  actually  found  that  sera  boiled  for  3  to  5  minutes  injected  into 
rabbits  induced  precipitins  which  acted  upon  boiled  proteins,  but  at 
the  same  time  it  was  determined  that  the  antibodies  so  produced 
were  no  longer  strictly  specific.  The  protocol  given  at  the  top  of  the 
next  page  will  illustiate  these  experiments. 

Summarizing  these  results  together  with  those  of  Fornet  and 
Miiller  and  of  Schmidt  it  would  seem  that  the  injection  of  boiled 
proteins  induces  precipitins  which  .no  longer  act  on  native  antigen, 
which  act  powerfully  on  boiled  antigen,  but  are  no  longer  strictly 
specific.  This  seems  to  us  of  great  theoretical  interest  as  showing 
an  alteration  by  heating  in  the  species  adherence  of  the  antigen. 
Practically,  therefore,  precipitins  produced  with  boiled  protein  are 
of  little  value,  and  forensic  determinations  of  boiled  proteins  should 

38  Obermeyer  and  Pick.     Wien.  klin.  Woch.,  No.  12,  1906. 

39  Fornet  and  Miiller.     Zeitschr.  f.  Hyg.,  Vol.  66,  1910. 

40  Zinsser  and  Ostenberg.    Proc.  N.  Y.  Pathol.  Soc.,  1914. 


INFECTION    AND    RESISTANCE 

Experiments  on  Cocto-precipitin.     Table  II  (March  23,  1913). 

Cross  titrations — dilutions  of  sera  in  salt  solution  boiled  5  minutes, 
precipitated  with  antisera  produced  by  injections  with  similarly  boiled  material. 

The  readings  here  indicated  were  taken  by  "ring"  test  at  the  end  of  30 
minutes. 


Beef 

Beef 

Beef 

Dog 

Dog 

Dog 

Sheep 

Sheep 

Sheep 

serum 

serum 

serum 

serum 

serum 

serum 

serum 

serum 

serum 

vs. 

vs. 

vs. 

vs. 

vs. 

vs. 

vs. 

vs. 

vs. 

Dilution 

anti- 

anti- 

anti- 

anti- 

anti- 

anti- 

anti- 

anti- 

anti- 

beef 

dog 

sheep 

dog 

beef 

sheep 

sheep 

dog 

beef 

precipi- 

precipi- 

precipi- 

precipi- 

precipi- 

precipi- 

precipi- 

precipi- 

precipi- 

tin 

tin 

tin 

tin 

tin 

tin 

tin 

tin 

tin 

:20 

+ 

+ 

+ 

+  + 



+ 

+  + 

+  + 

+ 

:50 

+  +  + 

+ 

+  +  + 

+  + 



+  + 

+  +  + 

+ 

+  +  + 

:100 
.cnn 

+  +  + 

_|  1 

+ 

+  + 

+ 



+ 

+  + 

+ 

+  + 

.OUU 

:1,000 

T  T" 

± 

— 

•§• 

•§• 

± 



± 

± 





Controls  of 

boiled  serum 

alone* 

1:20 

— 





1:50 

_  — 

1 

. 

Serum 

control 

*  These  controls  were  necessitated  by  the  fact  that  the  boiled  serum  suspensions  were  them- 
selves turbid  and  occasionally  showed  slight  settling  on  standing. 


be  done,  as  advised  by  Schmidt,  by  the  "70°  precipitins,"  or  with 
native  precipitin  as  practiced  by  Fornet  and  Miiller. 

The  specificity  which  is  the  basis  of  the  practical  value  of  the  reac- 
tions that  we  have  discussed  is  spoken  of  as  "species"  specificity 
since  it  has  been  found  that  the  blood  serum  of  rabbits  or  other  ani- 
mals into  which  the  serum  of  another  animal  has  been  injected 
reacts,  not  only  with  the  homologous  blood  serum,  but  also  with 
extracts  of  the  various  organs  of  the  particular  species  of  animal 
which  furnished  the  serum.  Thus  if  we  immunize  rabbit,  let  us  say, 
with  sheep  serum  the  resulting  precipitin  will  react,  not  only  with 
sheep  serum,  but  also  with  extracts  of  sheep  spleen,  sheep  liver,  etc. 
It  seems  that  every  species  of  animal  possesses  throughout  its  tis- 
sues a  particular  variety  of  protein,  fundamental  to  its  general 
metabolism  and  peculiar  to  its  species.  On  the  other  hand,  we  have 
seen  in  the  preceding  discussions  how  chemically  slight  the  changes 
in  a  protein  may  be  which  can  alter  materially  its  antigenic  nature, 
and  it  is  a  logical  deduction  that  different  organs  of  the  same  animal 
might  contain  antigenic  constituents  qualitatively  different  from  the 
general  serum  protein.  There  are  undoubtedly  in  many  organs 
protein  complexes  which  are  peculiar  to  them  and  not  present  in 
other  organs,  and  it  would  be  reasonable  to  expect  therefore  that 
immunization  with  separate  organ  substances  would  lead  to  the  pro- 
duction of  sera  of  specific  precipitating  power  for  the  protein  of  that 
particular  kind  of  organ.  This  is  not  ordinarily  obtainable,  how- 


THE    PHENOMENON    OF    PRECIPITATION         263 

ever,  because  it  has  been  impossible  to  isolate  from  organs  their  pe- 
culiar, characteristic  proteins,  and  immunization  of  animals  with 
organ  extracts  or  solutions  has  necessarily  implied  the  injection  of 
much  blood  protein  and  other  albuminous  material  of  a  character 
general  to  many  organs  of  the  animal,  i.  e.,  to  the  species.  These 
quantitatively  overshadow  the  organ-specific  substances  which  may  be 
present,  and  give  rise,  therefore,  to  a  "species"  precipitin.  That 
"organ  specificity/7  however,  is  a  fact  has  been  shown  by  the  experi- 
ments of  Uhlenhuth  with  the  protein  of  the  crystalline  lens  of  the  eye. 
Immunization  with  this  substance  induces  a  precipitin  which  does 
not  react  with  the  serum  of  the  animal  from  which  the  lens  was  taken, 
but  does  react,  not  only  with  the  crystalline  lens  proteins  of  this  spe- 
cies of  animal,  but  also  with  crystalline  lens  proteins  in  general, 
though  taken  from  another  animal  species.  Analogous  to  this  are  the 
experiments  of  von  Dungern  and  others  upon  the  protein  derived 
from  the  testicle. 

In  both  of  these  cases,  as  well  as  in  other  less  sharply  defined 
examples,  the  specificity  is  attached,  not  to  the  species  of  animal, 
but  rather  to  the  nature  of  the  organ  from  which  the  particular 
protein  is  derived.  These  facts — first  ascertained  by  means  of  the 
precipitin  reaction — have  been  recently  confirmed  by  means  of  the 
reaction  of  anaphylaxis  by  Uhlenhuth  and  Haendel,  and  by  Kraus, 
Doerr,  and  Sohma.  (See  chapter  on  Anaphylaxis.)  They  have 
been  discussed,  moreover,  in  connection  with  the  problem  of  spe- 
cificity in  general. 

Biologically  they  probably  signify  that,  although  there  are  fun- 
damental species  differences  between  the  general  body  proteins  of 
various  animals,  there  are  still,  in  certain  highly  specialized  organs, 
varieties  of  protein  which,  possibly  because  of  functional  exigencies, 
have  developed  similar  chemical  characteristics.  These  have  been 
determinable  by  our  present  methods,  however,  only  for  organs  like 
the  lens,  the  testicle,  and  the  placenta  from  which  the  organ-specific 
protein  can  be  gotten  in  a  relatively  pure  state.  The  pathological 
importance  of  these  phenomena  lies  in  the  fact  that,  although  guinea 
pig  serum  injected  into  a  guinea  pig  will  not  give  rise  to  antibodies, 
lens  protein  apparently  will  do  so — -an  observation  which  opens  the 
possibility  of  autocytotoxins.  The  significance  of  this  is  indicated  in 
such  investigations  as  those  of  Homer,41  who,  using  the  complement- 
fixation  technique  to  determine  antibody,  found  that  the  serum  of 
adult  human  beings  possessed  antibodies  for  their  own  lens  protein, 
but  that  such  antibodies  were  absent  in  the  sera  of  children. 

The  study  of  agglutination  and  that  of  precipitation  reveal, 
throughout,  a  close  similarity  between  the  two  reactions,  and  indeed 
in  physical  principles  they  are  probably  the  same,  although  the  one 

41  Rb'mer.     Klin.  Monatsbl.  f.  Augenlieilkunde,  Sept.,  1906.     Ref.  from 
"Weichhardt's  Jahresber.,"  Vol.  2,  1906,  p.  348. 


264  INFECTION    AND    RESISTANCE 

(agglutination)  consists  in  the  flocculation  of  large  particles  in  sus- 
pension— the  bacteria — while  in  the  other  the  precipitation  is  one 
of  smaller  units — the  precipitable  colloidal  particles  of  the  protein 
solutions.  This  phase  of  the  subject  will  be  more  thoroughly  dis- 
cussed directly. 

Meanwhile,  it  is  noticeable  also  that,  even  without  drawing  the 
physical  parallel  between  the  two  reactions,  there  is  much  in  the 
behavior  of  the  antibodies — the  agglutinins  and  the  precipitins  as 
conceived  by  Ehrlich,  which  led  him  and  his  school  to  attribute  to 
them  a  similar  receptor  structure.  Like  the  agglutinins,  the  pre- 
cipitins are  not  inactivated  by  56°  C.,  but  when  once  rendered  in- 
effectual by  higher  temperatures  (70°  C.  or  over)  they  can  no  longer 
be  reactivated  by  the  addition  of  fresh  normal  serum.  For  this 
reason  chiefly  Ehrlich  has  conceived  that  both  agglutinins  and  pre- 
cipitins are  "haptines"  of  the  second  order. 


ZYMOPHOfSE 


CELL   OR  ('BOW  CELL 

OTHER  ANTIGEN 

SCHEMATIC  KEPRESENTATION  OF  EHRLICH 's  VIEWS  ON  THE  STRUCTURE  OP  PRE- 
CIPITINS. 

Ehrlich  assumes  that  when  dissolved  protein  substances — ordi- 
narily suitable  for  body  nutrition — are  injected  into  animals,  they 
become  anchored  to  the  cells  by  such  receptors  of  the  second  order. 
When  overproduction  occurs  in  response  to  repeated  stimulation  of 
the  cells  by  consecutive  injections  (see  Side-Chain  Theory),  these 
haptines  of  the  second  order  circulate  as  agglutinins  or  pre- 
cipitins. Since  they  act  without  the  apparent  cooperation  of  alexin, 
he  supposes  that  they  carry  within  themselves  the  "zymophore,"  or 
ferment  groups,  by  means  of  which  the  agglutination  or  coagulation 
is  accomplished.  It  is  this  zymophore  group  which,  it  is  assumed, 
accomplishes  the  digestion  of  the  foreign  protein  before  its  assimila- 
tion, when  these  receptors  are  still  parts  of  the  living  cell. 

Thus  the  conception  of  precipitins  is  identical  with  that  formu- 
lated by  the  same  school  concerning  the  agglutinins,  and  the  deduc- 
tions from  these  premises  have  been  essentially  similar.  Thus,  anal- 
ogous to  the  conditions  prevailing  in  agglutination,  Pick,42  and 
Kraus  and  v.  Pirquet  43  have  shown  that  when  precipitating  serum 
is  inactivated  by  heat,  and  then  is  added  to  bacterial  filtrates,  it  will 

42  Pick.     "Hofmeister's  Beitrage,"  Vol.  1,  1902. 

43Kraus  and  von  Pirquet.     CentralU.  /.  Bakt.,  Vol.  32,  1902. 


THE    PHENOMENON    OF    PRECIPITATION         265 

prevent  their  subsequent  precipitation  by  active  precipitin.  An 
illustration  of  this  is  found  in  the  following  protocol  taken  from  the 
paper  by  Kraus  and  v.  Pirquet  (loc.  cit.,  p.  69). 

(a)  5  c.  c.  cholera  filtrate  +  0.5  c.  c.  inactiv.  (60°)  cholera  serum  =  no  precipitate 

after  10  hours  at  37°  C. 
After  10  hours  add  0.5  c.  c.  active  cholera  serum  =  no  precipitate. 

(b)  Omitted. 

(c)  Omitted. 

(d)  5  c.  c.  cholera  filtrate  +  &-5  c.  c.  active  cholera  serum  =  after   10   hrs. 

typical  precipitate. 

From  this  it  was  concluded  that  heat  may  destroy  the  zymophore 
or  coagulating  group  of  precipitins,  leading  to  the  formation  of 
"precipitinoids"  which,  like  agglutinoids,  may  have  a  higher  affinity 
for  the  antigen  than  is  possessed  by  the  uninjured  antibody. 

Subsequently  there  were  opposed  to  these  views  the  physical  in- 
terpretations which  have  been  outlined  sufficiently  in  the  section  on 
Agglutination  (see  p.  240).  In  the  case  of  precipitation  the  anal- 
ogy between  colloidal  reactions  and  the  serum  phenomena  is  fully  as 
striking  as  in  the  former,  an  analogy  in  the  delineation  of  which  the 
first  credit  belongs  to  Landsteiner,44  45  and  important  further  contri- 
butions have  been  made  by  Neisser  and  Friedemann,  Porges,  Gen- 
gou,  and  a  number  of  others.  As  in  agglutination  and  colloidal  floc- 
culation,  the  presence  of  salts  (electrolytes)  fundamentally  influ- 
ences the  occurrence  of  precipitin  reactions;  and  in  both  colloidal 
and  precipitin  reactions  the  relative  concentration  of  the  reacting 
bodies  is  paramount  in  determining  whether  or  not  precipitation 
takes  place.  In  this  connection  the  most  frequently  observed  inhibi- 
tion occurring  in  serum  precipitations  is  that  which  is  caused  by  an 
excess  of  antigen.  An  example  of  this  is  as  follows : 


Sheep  serum  0.5  c.  c. 

Antisheep  rabbit  serum 

Precipitate 

1:10                     +                0.5  c.c.                                    — 
1:50                     +                0.5  c.c. 
1:100                   +                0.5  c.c.                                  ++• 
1:500                   +                0.5  c.c.                                +  +  + 
1:1,000                +                0.5  c.c.                                  ++ 
1:5,000                +                0.5  c.c.                                     -f 

This  is  entirely  analogous  to  the  inhibition  which  may  occur 
when,  let  us  say,  a  weak  gelatin  solution  is  added  to  a  colloidal  sus- 
pension of  arsenic  trisulphid ;  or  blood  serum  is  added  to  mastic  or 
arsenic  suspensions.  In  both  cases  inhibition  zones  appear  which 

44  Landsteiner  and  Jagic.     Munch,  med.   Woch.,  No.  18,  1903;  No.  27, 
1904;  Wien.  kl.  Woch.,  No.  3,  1904. 

45  Landsteiner  and  Stankovic.     Centralbl  f.  Bakt.,  Vols.  41  and  42,  1906. 


266  INFECTION    AND    RESISTANCE 

show  that  the  relative  quantities  of  the  two  reacting  bodies  are  quite 
as  significant  as  their  chemical  or  physical  constitution  in  determin- 
ing the  occurrence  of  flocculation.  This,  according  to  Bechold,  Bil- 
litzer,46  and  others  depends  upon  the  fact  that  the  reason  for  floccu- 
lation is  one  of  electrical  charge.  One  hydrosol — say  arsenic  tri- 
sulphid — can  be  flocculated  by  the  oppositely  charged  colloidal  alu- 
minium hydroxid,  but  this  will  occur  only  when  the  quantitative 
relations  are  properly  adjusted.  If  one  or  the  other  is  in  excess,  no 
flocculation  may  occur,  and,  if  subjected  to  a  direct  current,  both 
colloids,  though  ordinarily  wandering  in  opposite  directions,  will 
now  wander  in  that  of  the  one  which  is  now  present  in  the  largest 
amount.  We  will  not  elaborate  here  upon  the  causes  for  this,  since 
they  have  been  indicated  in  the  section  on  Agglutinins,  and  are  set 
forth  more  accurately  by  Prof.  Young  in  the  special  chapter  on  Col- 
loids. 

This  effect  of  quantitative  proportions  would  explain  not  only 
the  absence  of  precipitation  in  the  presence  of  too  much  antigen,  but 
also  the  converse  phenomenon,  already  mentioned,  that  precipitation 
may  be  inhibited  when  the  precipitin  is  in  excess. 

The  fact  that  heated  precipitating  serum  when  added  to  its  an- 
tigen not  only  does  not  cause  flocculation,  but  may  even  prevent  sub- 
sequent precipitation  by  active  precipitin,  also  finds  its  analogy  in 
colloidal  reactions  in  the  so-called  protective  colloids.  Thus  arsenic 
trisulphid  may  be  protected  from  precipitation  by  gelatin,  if  a  small 
amount  of  gum  arabic  is  added,  and  the  analogy  has  been  brought 
even  closer  by  Porges,47  who  showed  that  heated  serum  will  protect 
mastic  suspension  from  precipitation  by  normal  serum.  This  obser- 
vation of  Porges  is  so  closely  similar  to  the  results  obtained  by  Kraus 
and  v.  Pirquet  and  others  on  the  inhibition  of  precipitation  by  heated 
precipitating  serum  that  it  would  seem,  on  first  consideration,  effec- 
tually to  refute  the  conception  of  "precipitoids." 

However,  it  does  not  explain  the  specificity  of  such  inhibition  on 
the  part  of  heated  precipitating  serum,  as  reported  by  Kraus  and  v. 
Pirquet,  an  observation  which  is  one  of  the  strongest  arguments  in 
favor  of  the  derivation  of  the  inhibiting  factor  from  the  specific 
precipitin  (a  precipitoid).48 

In  spite  of  the  strong  evidence  in  favor  of  the  colloidal  inter- 
pretations, such  contrary  evidence,  brought  forward  by  careful  and 

46  Billitzer.     Cited  from  Bechold,  "Die  Kolloide,  etc.,"  p.  79. 

47  Porges.     Chapter  on  "Colloids  and  Lipoids"  in   "Kraus   u.   Levaditi 
Handbuch,"  Vol.  1. 

48  Although   normal   sera   may   gradually   precipitate    on    standing,    this 
takes  place  much  more  rapidly  in  precipitin-sera,     The  spontaneous  precipi- 
tation of  normal  sera  as  well  as  of  those  under  consideration  is  analogous 
to  what  Bechold  and  others  call  the  "ageing"    (altern)   of  colloidal  suspen- 
sions, which,  though  originally  stable,  will  eventually  settle  out,  even  in  the 
presence  of  protective  colloids. 


THE    PHENOMENON    OF    PRECIPITATION         267 

experienced  workers,  must  be  borne  in  mind  and  positive  acceptance 
of  the  colloidal  explanations,  however  attractive,  must  be  withheld 
until  much  further  investigation  has  been  done. 

Another  important  and  interesting  phase  of  the  study  of  precipi- 
tins  is  that  associated  with  the  occasional  presence  in  the  same  serum 
of  remnants  of  antigen  and  of  precipitins  which,  though  present 
side  by  side,  do  not  unite  to  form  precipitates.  This  condition 
is  frequently  seen  in  such  sera  as  those  produced  by  Fornet  and 
Mliller 49  for  rapid  precipitin  production  for  forensic  work,  a 
method  in  which  the  foreign  serum  is  injected  into  rabbits  in  large 
amounts  (2  to  10  c.  c.),  on  consecutive  days,  and  the  animals  are 
bled  6  to  8  days  after  the  last  injection.  That  such  sera  contain 
both  antigen  and  antibody  is  shown  by  the  fact  that,  though  clear 
when  taken,  they  will  show  precipitation  not  only  when  mixed  with 
dilutions  of  the  antigen,  but  also  when  added  to  homologous  precipi- 
tating sera.50 

This  phenomenon  has  been  noticed  by  Linossier  and  Lemoine,51 
Eisenberg,52  Ascoli,53  and  others,  and  has  been  extensively  studied 
by  von  Dungern.54  Gay  and  Rusk  55  have  recently  observed  it  in 
connection  with  the  rapid  method  of  precipitin  production  of  Fornet 
and  Miiller,  and  have  noted  that  such  sera,  although  containing  both 
antigen  and  precipitin,  do  not  possess  complement-fixing  properties. 
According  to  Uhlenhuth  and  Weidanz,56  the  antigen  may  persist  in 
the  sera  of  protein-immunized  animals,  in  demonstrable  amounts, 
as  long  as  fifteen  days  after  the  last  injection,  and  it  is  constantly 
present  during  this  period,  but  in  progressively  diminishing  amounts. 

We  are  thus  confronted  by  the  apparently  paradoxical  phenom- 
enon of  the  presence  in  these  sera,  side  by  side,  of  an  antigen  and  its 
homologous  precipitin,  incapable  of  reacting  with  each  other,  al- 
though each  of  them  readily  reacts  with  precipitin  or  antigen,  re- 
spectively, when  these  are  added  from  another  source. 

Many  attempts  have  been  made  to  account  for  this.  A  number 
of  observers,  notably  Eisenberg,  have  concluded  from  extensive  an- 

49  Fornet  and  Miiller.  Zeitschr.  f.  biol.  Technik  u.  Methodik,  Vol.  1, 
1908. 

so  j>or  instance,  a  rabbit  was  injected  on  three  consecutive  days  with 
sheep  serum.  It  was  bled  on  the  fifth  day  after  the  last  injection.  The 
serum  was  clear  when  taken,  but  a  precipitate  was  formed  when  it  was 
added  to  sheep  serum  and  also  when  it  was  added  to  serum  from  another 
rabbit  similarly  treated  and  containing  sheep  serum  precipitin. 

51  Linossier  and  Lemoine.     C.  R.  de  la  Soc.  de  Biol.,  54,  1902. 

52  Eisenberg.     Centralbl.  f.  Bakt.,  34,  1903. 

53  Ascoli.    Munch,  med.  Woch.,  Vol.  49,  No.  34,  1902. 

54  Von  Dungern.     Centralbl  f.  Bakt.,  34,  1903. 

55  Gay  and  Rusk.     "Univ.  of  Cal.  Public,  in  Pathology,"  Vol.  2,  1912. 

56  Uhlenhuth    and    Weidanz.      "Praktische    Anleitung    zur    Ausfiihrung, 
•etc.,"  Jena,  1909. 


268  INFECTION    AND    RESISTANCE 

alyses  of  quantitative  relationships,  both  of  agglutinin  and  precipitin 
reactions,  that  these  take  place  according  to  the  laws  of  mass  action. 
In  consequence,  in  addition  to  the  combined  precipitin-antigen  com- 
plex present  in  all  mixtures  of  the  two,  there  should  also  be  present 
free  dissociated  fractions  of  each,  in  amounts  dependent  upon  rela- 
tive concentrations.  This  might  explain  conditions  such  as  those 
described  above. 

Von  Dungern,  whose  paper  forms  one  of  the  most  extensive  studies 
of  the  phenomenon  with  which  we  are  concerned,  does  not  believe 
that  precipitin  reactions  can  follow  the  laws  of  mass  action,  and 
explains  the  simultaneous  presence  of  precipitin  and  antigen  in  the 
same  serum  by  assuming  a  multiplicity  of  precipitins.  He  believes 
that  every  proteid  antigen  contains  a  number  of  related  partial  an- 
tigens which  give  rise  in  the  immunized  animal  each  to  a  partial 
precipitin.  In  sera  in  which  both  antigen  and  precipitin  are  found 
side  by  side  and  free,  he  believes  that  the  antigen  is  of  a  nature  that 
has  no  affinity  for  the  particular  partial  precipitin  present  with  it. 
He  says :  "Auch  hier  handelt  es  sich  nicht  um  zwei  reaktionsf  ahige 
Korper,  deren  Verbindung  aus  irgend  welchen  Griinden  unterbleibt, 
sondern  um  Substanzen,  welche  keine  Affinitat  zu  einander  besitzen. 
Die  betreffenden  Kaninchen  haben  zu  dieser  Zeit  noch  nicht  alle 
moglichen  Teilprazipitine  gebildet,  sondern  nur  einzelne  derselben. 
Diese  zunachst  produzierten,  nur  auf  bestimmte  Gruppen  der  prazi- 
pitablen  Eiweisskorper  passenden  Partialprazipitine  sind  es,  welche 
nach  der  Absattigung  aller  zur  Verfiigung  stehenden  zugehorigen 
Gruppen  der  prazipitablen  Substanz  in  Serum  nachweisbar  werden. 
Daneben  bleibt  aber  ein  anderer  Teil  der  prazipitablen  Substanz, 
der  keine  Affinitat  zu  dem  gebildeten  Prazipitin  bestizt,  bestehen, 
solange  bis  ein  anderes  Partialprazipitin  von  den  Kaninchenzellen 
geliefert  wird,  welches  sich  mit  Gruppen  der  in  Losung  geliebenen 
Eiweisskorper  vereinigen  kann." 

Zinsser  and  Young57  have  also  studied  these  phenomena  and 
have  attempted  to  explain  them  on  the  basis  of  protective  colloidal 
action.  In  considering  the  theories  that  have  been  advanced  to  ex- 
plain these  occurrences,  the  conception  of  mass  action  as  accounting 
for  the  simultaneous  presence  of  the  two  reacting  bodies  in  the  same 
serum  seemed  entirely  incompatible  with  our  own  observations  and 
with  those  of  Gay  and  Rusk,  that  these  sera  do  not  of  themselves 
fix  alexin.  Were  the  conception  of  the  manner  of  union  of  these 
two  reagents,  according  to  the  laws  of  mass  action,  representative 
of  the  true  state  of  affairs,  it  would  be  necessary  to  assume  the  pres- 
ence, in  such  sera,  not  only  of  the  two  reacting  bodies  free  and  disso- 
ciated, but  also  of  a  definite  quantity  of  the  united  complex  of  the 
two,  a  state  of  equilibrium  being  established.  If  this  were  the  case 
the  sera  should,  in  agreement  with  all  experience  on  the  phenomenon 

57  Zinsser  and  Young.    Jour,  of  Exp.  Med.,  1913,  Vol.  17. 


THE    PHENOMENON    OF    PRECIPITATION         269 

of  complement  fixation,  exert  definite  complement-binding  power. 
Moreover,  it  has  not  been  experimentally  shown  that  colloidal  sub- 
stances react  in  accordance  with  the  laws  of  mass  action  as  observed 
for  simpler  chemical  substances. 

As  regards  the  opinion  of  von  Dungern,  this  seemed  incom- 
patible with  another  occurrence,  observed  by  many  writers,  namely, 
that  such  sera,  although  clear  at  first,  eventually,  after  prolonged 
standing,  do  actually  precipitate  spontaneously;  that  is,  the  union 
of  the  precipitin  and  the  precipitinogen  does  actually  take  place,  but 
goes  on  with  extreme  slowness. 

Now  a  notable  and  strange  feature  of  this  phenomenon  is  the 
fact  that  two  such  sera,  both  containing  antigen  and  precipitin,  but 
neither  of  them  precipitating  by  itself,  will  precipitate  each  other 
when  mixed.  For  this  reason  Uhlenhuth  has  advised  against  the 
use  of  mixtures  of  precipitin  sera  for  forensic  tests.  For  it  is  not 
unusual  that  precipitin  sera,  even  when  produced  by  the  slow  method, 
may  contain  traces  of  antigen,  and  this  may  lead  to  precipitate 
formation  if  such  a  serum  is  mixed  with  another  homologous  pre- 
cipitin and  thereby  simulate  a  positive  forensic  test. 

In  seeking  analogy  for  this  serum  phenomenon  with  the  various 
colloidal  suspensions,  the  problem  consisted  in  protecting  two 
mutually  precipitating  colloids  by  a  third,  and  this  in  such  propor- 
tions that  the  mixing  of  two  such  protected  suspensions,  each  con- 
taining all  three  of  the  elements,  would  be  followed  by  precipitation. 
This  was  obtained  by  the  use  of  gum  arabic,  gelatin,  and  arsenic  tri- 
sulphid.  Thin  emulsions  of  gelatin  will  precipitate  arsenic  tri- 
sulphid  suspensions.  Small  amounts  of  gum  arabic  will  act  as  a 
protective  agent,  preventing  the  precipitations. 

The  amount  of  the  protecting  substance  necessary  to  prevent 
precipitation  in  any  one  mixture  varies  apparently  with  every 
change  in  the  relative  proportions  of  the  two.  Thus  a  considerable 
number  of  mixtures  of  the  three  can  be  made  which  will  remain 
stable  for  days,  the  actual  and  relative  quantities  of  the  three 
ingredients  differing  in  each  of  the  mixtures.  When  two  such  mix* 
tures  are  poured  together,  in  many  cases  precipitation  will  result, 
varying  in  speed  and  completeness,  according  to  the  particular  quan- 
titative relationship  arrived  at  in  the  mixture. 

An  example  of  such  an  experiment  follows : 

Two  solutions  of  colloidal  arsenic  sulphid  were  prepared,  one  containing 
1  gm.  per  liter,  the  other  containing  5  gm.  per  liter.  With  Kahlbaum's  "Gold- 
ruck"  gelatin  a  solution  containing  1  gm.  per  liter  was  prepared.  A  solution  of 
gum  arabic  was  prepared  which  contained  10  gm.  per  liter,  this  being  made 
stronger  than  the  gelatin  solution  to  avoid  too  great  dilution  in  the  final  mixtures. 
The  gelatin  solution  was  prepared  twenty-four  hours  before  being  used,  as 
freshly  prepared  gelatin  has  but  slight  precipitating  power  for  arsenic  sulphid, 
this  power  appearing  to  increase  greatly  with  the  ageing  of  the  solution. 


270  INFECTION    AND    RESISTANCE 

For  the  purpose  of  demonstrating  this  analogy  two  protected 
solutions  were  prepared  as  follows : 

Solution  1. — This  consisted  of  2  drops  of  gum  arabic,  2  c.  c.  of  gelatin,  and 
5  c.  c.  of  the  weaker  arsenic  solution. 

Solution  2. — This  consisted  of  10  drops  of  gum  arabic,  1  c.  c.  of  gelatiD, 
and  about  4  c.  c.  of  the  stronger  arsenic  solution. 


In  each  case  the  arsenic  sulphid  was  added  until  there  were  signs 
of  increasing  opalescence  or  turbidity,  this  being  done  in  order  that 
the  two  solutions  should  each  be  as  little  overprotected  as  possible. 

Portions  of  the  two  solutions  were  then  mixed  in  equal  propor- 
tions. In  the  course  of  a  few  minutes  the  mixture  was  noticeably 
more  turbid  than  either  of  the  original  solutions.  This  turbidity 
continued  to  increase  quite  rapidly,  and  on  the  following  morning 
after  about  sixteen  hours  of  standing,  the  mixture  was  found  to  be 
completely  flocculated  out,  while  the  original  protected  mixtures  re- 
mained unprecipitated  and  showed  about  the  same  degree  of  opales- 
cence as  on  the  preceding  night.  The  same  condition  of  affairs  was 
found  to  have  persisted  after  five  days.  On  the  fifth  day  the  less 
concentrated  of  the  clear  protected  suspension  began  to  settle  out, 
and  was  completely  precipitated  within  twenty-four  hours.  The 
other  remained  clear  for  four  days  more,  but  on  the  ninth  day  it 
began  to  precipitate  slightly,  the  precipitation  remaining  incom- 
plete. 

In  these  cases  it  appears,  therefore,  that  a  complete  analogy  to 
the  observed  conditions  of  the  serum  reactions  has  been  found,  and 
that  all  data  observed  in  connection  with  sera  in  which  antigen  and 
precipitin  are  found  side  by  side  without  reacting  can  be  most  simply 
explained  on  the  conception  of  protective  colloid  action.  Moreover, 
the  chemical  nature  of  the  substances  involved  seems  to  add  weight 
to  this  point  of  view. 

These  relations  have  been  gone  into  here  at  some  length,  since 
they  seem  to  us  to  possess  considerable  theoretical  and  practical  sig- 
nificance. For  it  may  be  that  the  presence  of  a  protective  colloid 
may,  by  inhibiting  the  union  of  antigen  and  precipitin  within  the 
body,  protect  the  animal  from  intoxication  during  the  early  stages 
of  immunization  when  antigen  and  antibody  are  present  simulta- 
neously for  longer  or  shorter  periods.  Were  union  between  the  two 
possible  at  such  times  in  the  circulation,  an  assumption  necessitated 
both  by  the  hypotheses  of  mass  action  and  of  multiplicity  of  precip- 
itins,  there  would  probably  be  an  absorption  of  complement  by  these 
complexes,  with,  as  shown  by  Friedberger,  a  consequent  formation  of 
powerful  toxic  products.  (See  chapter  on  Anaphylaxis. )  It  is 
not  impossible  by  any  means,  therefore,  that  the  injection  of  anti- 
gen in  an  animal  in  which  such  a  balance  has  been  established  may 


THE  PHENOMENON  OF  PRECIPITATION  271 

lead  to  a  sudden  elimination  of  the  colloidal  protective  action,  union 
of  the  antigen  and  antibody,  and,  by  the  mechanism  just  outlined, 
a  naphy lactic  shock. 

The  fact,  moreover,  that  mere  heating  will  change  the  precipi- 
tating action,  which  certain  sera  have  on  inorganic  colloids,  to  a 
protective  one  seems  to  show  that  this  latter  function  may  justly 
be  associated  with  delicate  physical  or  chemical  alterations  of  animal 
sera. 

Furthermore,  this  point  of  view  is  strengthened  by  the  fact  that 
the  mutual  precipitation  of  sera,  such  as  those  described,  takes  place 
slowly,  as  does  the  mutual  precipitation  of  two  protected  colloidal 
mixtures,  in  contradistinction  to  the  more  rapid  precipitation  which 
takes  place  when  any  of  these  sera  is  added  to  an  antigen  dilution, 
where  the  element  of  protection  may  be  assumed  to  be  practically 
eliminated  by  more  extensively  changed  quantitative  relations. 

This  point  of  view  has  lately  been  disputed  by  Weil,  who  has  gone 
back  to  the  older  view  of  von  Dungern,  largely  on  the  basis  of  pre- 
cipitation experiments  he  carried  out  with  crystallized  egg  albumen. 
Weil  claims  that  if  a  pure  protein,  like  crystallized  egg  albumen,  is 
used  for  immunization,  antigen  and  antibody  are  never  found  simul- 
taneously in  the  blood  stream.  If  this  were  true  it  would  indeed  con- 
stitute a  very  important  contradiction  of  our  point  of  view.  In  con- 
sequence Bayne-Jones  carried  out  similar  experiments  in  our  labora- 
tory, and  found  that  even  with  the  purest  obtainable  recrystallized 
egg  albumen  both  antigen  and  antibody  are  at  times  demonstrable 
in  the  blood.  He  showed  this  both  by  precipitation  and  by  comple- 
ment fixation.  We  do  not  consider  the  question  closed,  however,  be- 
cause it  is  indeed  true  that  when  working  with  a  purified  protein  it 
is  more  difficult  to  demonstrate  the  two  substances  in  any  quantity, 
than  it  is  when  the  crude  serum  antigen-  is  used.  This  may  be,  of 
course,  due  to  the  fact  that  pure  egg  albumen  may  be  more  rapidly 
assimilated  and  remains  in  the  circulation  for  a  less  extensive  period 
than  does  the  crude  antigen.  However,  further  experiments  in  this 
direction  will  unquestionably  clear  the  matter  up  because  it  is  a 
simple  question  of  fact  amenable  to  experiment. 

As  far  as  the  separation  of  a  pure  albumen  from  an  albumen 
globulin  mixture  is  concerned,  we  feel  rather  doubtful  whether  such 
a  sharp  separation  can  be  made  and  are  inclined  to  lean  towards  the 
view  expressed  by  Wells  some  years  ago,  that  the  various  proteins 
(that  is,  albumens  and  globulins)  of  the  same  animal  material  are  not 
sharply  separable  but  shade  one  into  the  next  as  in  a  spectrum.  This 
view  is  the  more  likely  when  we  consider  that  our  methods  of  separa- 
ting these  various  proteins  are  purely  physical  methods  of  heating 
and  of  precipitation  by  salts. 


CHAPTER   XI 

PHAGOCYTOSIS 

EARLY  investigations  into  the  fate  of  bacteria  within  the  infected 
animal  body  were  largely  carried  out  by  pathological  anatomists,  and 
the  observation  of  the  presence  of  micro-organisms  within  the  cells 
of  the  animal  and  human  tissues  was  definitely  made  as  early  as 
1870.  Hay  em,1  Klebs,2  Waldeyer,3  and  others,  saw  leukocytes  con- 
taining bacteria  but  failed  to  interpret  this  in  the  sense  of  possible 
protection.  The  process  was  regarded  rather  as  a  means  of  trans- 
portation of  the  bacteria  through  the  infected  body,  or  it  was  as- 
sumed that  possibly  the  micro-organisms  had  entered  these  cells  be- 
cause erf  the  favorable  nutritive  environment  thus  furnished. 

The  first  to  suggest  that  such  cell  ingestion  might  represent  a 
method  of  defence  was  Panum,4  who  referred  to  it  as  a  vague  possi- 
bility. A  similar  but  more  convinced  expression  of  this  opinion 
was  made  in  1881,  according  to  Metchnikoff,5  by  Roser  in  explaining 
the  resistance  of  certain  lower  animals  and  plants  against  bacteria. 
But  Roser  brought  no  experimental  support  for  his  contention,  and 
little  attention  was  paid  to  his  assertion. 

The  significance  of  cell  ingestion  as  a  mode  of  protection  against 
bacterial  invasion,  therefore,  was  hardly  more  than  a  vague  sugges- 
tion when  Metchnikoff,  who,  though  a  zoologist,  had  become  intensely 
interested  in  the  problem  of  inflammation,  began  to  experiment  upon 
the  cell  reaction  which  followed  the  introduction  of  foreign  material, 
living  or  dead,  into  the  larvae  of  certain  starfishes  (Bipinnaria). 

Pathologists,  at  this  time,  held  complicated  views  of  inflamma- 
tion which  involved  complex  coordinated  reactions  of  vascular  and 
nervous  systems,  and  Metchnikoff's  primary  purpose  was  to  observe 
reactions  to  irritation  in  simple  forms  devoid  of  specialized  vascular 
or  nervous  apparatus.  He  noted  in  these  transparent,  simple  forms 
of  life  that  the  foreign  particles  were  rapidly  surrounded  by  masses 
of  ameboid  cells  and  reached  a  conclusion  which,  in  his  own  words, 
is  expressed  as  follows: 

1  Hayem.     C.  E.  de  la  Soc.  Biol,  1870. 

2  Klebs.     Pathol.  Anat.  der  Schusswinden,  1872. 

3  Waldeyer.     Arch.  f.  GynekoL,  Vol.  3,  1872. 

4  Panum.     Virch.  Arch.,  Vol.  60,  1874. 

5  Metchnikoff.     "L'lmmunite  dans  les  Maladies  Infectieuses." 


PHAGOCYTOSIS 

"L'exsudat  inflammatoire  doit  etre  considere  comme  une  reac- 
tion centre  toutes  sortes  de  lesions  et  Fexsudation  est  un  phenomena 
plus  primitif  et  plus  ancien  que  le  role  du  systeme  nerveux  et  des 
vaisseaux  dans  rinflamrnation."  6 

He  compared  the  process  of  cell  ingestion  or  phagocytosis  of  for- 
eign particles,  as  here  observed,  to  that  taking  place  in  the  most 
simple  intracellular  digestion  which  occurs  in  unicellular  forms,  a 
hereditary  cell  function  now  specialized  in  certain  mesodermal  cells, 
and  passed  on  in  the  evolution  of  higher  forms  to  other  specialized 
cells.  And  indeed  in  animals  of  the  most  complex  structure  the 
leukocytes  which  carry  on  this  phagocytic  process  may  be  considered 
as,  in  a  way,  representing  a  primitive  form  of  cell,  since  they  are 
only  nucleated  elements  of  the  body  which  wander  from  place  to 
place,  and  are  anatomically  independent  of  nervous  control.  In 
1883,  at  the  Naturalists'  Congress  in  Odessa,  Metchnikoff 7  first 
expressed  his  views  and  communicated  the  first  of  the  splendid  re- 
searches upon  which  our  modern  conception  of  phagocytosis  is  based. 

His  earlier  studies  were  carried  out  with  a  small  crustacean,  the 
daphnia,  in  which  he  studied  the  reaction  which  followed  the  intro- 
duction of  yeast  cells.  He  observed  the  struggle  which  ensued  be- 
tween the  ameboid  leukocytes  of  the  crustacean  and  the  infecting 
agents  and  determined  that  complete  enclosure  of  the  yeast  within 
the  leukocytes  assured  protection  to  the  daphnia,  while  a  failure  of 
this  process,  either  from  fortuitous  causes  or  because  of  too  large  a 
quantity  of  the  infecting  agents,  resulted  in  disease  and  rapid  death. 

This  early  work  of  Metchnikoff  forms  the  beginning  of  a  long 
train  of  investigations  to  which  we  owe  most  of  the  basic  facts  we 
possess  concerning  the  role  of  the  phagocytic  cells  in  the  protection 
of  the  body  against  infection.  Just  as  the  various  serum  phenomena, 
of  which  we  have  spoken,  have  a  general  biological  significance  apart 
from  their  importance  in  relation  to  bacterial  invasion,  so  the  process 
of  phagocytosis  must  be  looked  upon  as  an  attribute  of  the  animal 
and  vegetable  cell  which  has  important  physiological  bearing  entirely 
apart  from  infection. 

In  fact,  the  ingestion  of  bacteria  and  other  foreign  particles  by 
the  leukocytes  and  other  phagocytic  cells  of  higher  plants  and  ani- 
mals is  entirely  analogous  to  the  intracellular  digestive  processes 
which  take  place,  as  the  ordinary  manner  of  nutrition,  among  the 
unicellular  forms.  Among  the  rhyzopods,  in  general,  food  is  taken 
in  by  means  of  the  ingestion  of  other  smaller  forms  of  life,  bacteria, 

6  Inflammatory  exudation  should  be  considered  as  a  reaction  against  all 
sorts  of  injuries,  and  exudation  is  a  phenomenon  more  primitive  and  ancient 
than  are  the  parts  played  by  nervous  system  and  blood  vessels  in  the  process 
of  inflammation. 

7  Metchnikoff.     Arb.  a,  d.  zool.  Inst.,  Wien,  Vol.  5,  1883. 


274  INFECTION    AND    RESISTANCE 

algse,  etc.  (or  particles  of  dead  organic  matter),  into  the  cell  bodj- 
of  the  protozoon. 

These  materials  are  gradually  engulfed  by  the  body  of  the  ameba, 
which  flows  about  them  with  its  pseudopods,  and  within  the  cyto- 
plasm undergo  gradual  digestion.  The  process  has  been  carefully 
studied  by  Mouton.8  In  symbiotic  cultures  of  amebse  with  colon 
bacilli  on  agar  plates,  the  bacteria  are  taken  up  in  large  numbers 
and  about  them  are  formed  small  vacuoles.  That  the  digestion  takes 
place  in  a  slightly  acid  medium  with  the  vacuoles  can  be  proved  by 
adding  a  drop  of  neutral  red  to  the  hang-drop  preparation  of  ameba3 
and  observing  the  brownish-red  color  taken  by  the  materials  in  the 
vacuoles.  Mouton  was  able  to  obtain  a  digestive  ferment  from  the 
amebse,  by  glycerin  extraction,  which  exerted  strong  proteolytic  action 
upon  various  albuminous  substances,  liquefied  gelatin,  and  digested 
dead  colon  bacilli  in  vitro,  acting  best  in  slightly  alkaline,  but  also 
in  slightly  acid,  reactions.  It  is  plain,  therefore,  that  the  most  prim- 
itive form  of  digestion  is  an  intracellular  one  carried  on  by  ferments 
comparable  in  every  way  to  the  secreted  digestive  enzymes  which 
accomplish  the  same  purpose  outside  of  the  cells  in  higher  animals. 
In  essence,  however,  there  is  no  fundamental  difference  physiolog- 
ically between  intra-  and  extracellular  digestions,  and  the  intracellu- 
lar manner  of  assimilating  solid  nutritive  particles  may  be  retained 
in  forms  much  higher  in  the  scale  of  evolution  than  the  rhizopods. 
It  has  been  studied  by  Metchnikoff  and  others  in  certain  of  the  flat 
worms  (Dendrocelum  lacteum)  in  which  typical  phagocytosis  is  car- 
ried on  by  the  cells  of  the  intestinal  mucosa.  Many  of  these  plan- 
aria  obtain  their  nourishment  by  sucking  the  blood  of  higher  ani- 
mals. Placed  under  a  microscope  after  feeding,  it  may  be  seen  that 
the  foreign  blood  cells  are  rapidly  taken  up  by  the  intestinal  epithe- 
lial cells,  which  engulf  them  by  means  of  pseudopodia  not  unlike 
those  of  the  ameba.  After  ingestion,  here,  too,  the  blood  cells  are 
surrounded  by  vacuoles  within  which  their  gradual  disintegration  or 
digestion  is  accomplished.  Similar  intracellular  digestion  seems  to 
be  general  among  the  coelenterates,  and  has  been  thoroughly  studied 
by  Metchnikoff  in  the  actinia.  Here  the  food  particles  are  carriecf 
by  the  tentacles  into  the  esophagus,  and  are  taken  up  by  the  endo- 
dermal  cells  of  the  so-called  "mesenteric  filaments/7  where  they  are 
digested  by  a  trypsin-like  enzyme.  In  these  animals  digestion  is 
entirely  intracellular,  though  the  ingesting  cells  are  the  parts  of  a 
specialized  tissue.  In  other  forms,  still  higher  in  the  scale,  although 
there  is  persistence  of  intracellular  digestion,  the  extracellular 
process  begins  to  be  developed.  Thus  in  certain  mollusca  the  solid 
food  is  taken  into  the  intestinal  canal,  where  it  first  undergoes  a 
preliminary  digestion  by  secreted  intestinal  juices.  After  it  has 

8  Mouton.     C.  E.  de  I'Acad.  des  Sciences,  Vol.  133,  1901. 


PHAGOCYTOSIS  275 

been  reduced  to  small  amorphous  particles  in  this  way,  these  are 
seized  by  the  ameboid  cells,  and  intracellular  digestion  completes  the 
process  which  has  been  begun  extracellularly. 

As  we  study  the  process  among  higher  animals,  it  appears  that, 
among  vertebrates,  the  intracellular  methods  of  digestion  have  been, 
at  least  for  normal  metabolism,  entirely  displaced  by  the  extracellu- 
lar as  it  occurs  in  the  intestine,  where  solid  particles  are  rendered 
completely  amorphous,  dissolved,  and  reduced  to  a  diffusible  condi- 
tion by  the  digestive  juices  before  they  are  offered  to  the  cells  for 
utilization.  However,  the  capacity  for  intracellular  digestion  is  not 
entirely  lost,  and  is  retained  of  necessity  in  certain  body  cells.  For 
were  there  not  such  an  emergency  arrangement  the  body  would  lack 
an  available  mechanism  with  which  to  meet  such  accidents  as  ex- 
travasations of  blood,  or  the  entrance  of  bacteria  and  other  foreign 
solid  particles  into  the  tissues.  It  seems  reasonable  to  classify  both 
the  phagocytic  action  of  body  cells  and  the  formation  of  antibodies 
in  the  blood  plasma,  primarily  as  emergency  devices  for  the  diges- 
tion of  foreign  materials  both  formed  and  unformed  which,  under 
abnormal  conditions,  penetrate  into  the  physiological  interior  of  the 
body  (blood  stream  or  tissue  spaces),  and  must  be  disposed  of. 

In  the  lowest  animals  the  single  cell  is  called  upon  to  perform 
all  necessary  functions.  In  the  course  of  evolution,  however,  as  the 
body  becomes  more  and  more  a  community  of  many  cells,  a  division 
of  labor  takes  place  which  is  expressed  morphologically  in  the  differ- 
entiation of  tissues  and  organs,  and  physiologically  in  the  adaptation 
of  individual  tissue  cells  to  the  performance  of  specialized  functions. 
Nevertheless,  it  is  necessary,  both  for  certain  normal  processes,  as 
well  as  for  provision  against  such  complex  emergencies  as  those 
mentioned,  that  certain  cells  of  the  complex  community  should  retain 
the  primitive  abilities  of  the  more  independent  cells  of  the  lower 
forms.  Thus,  among  many  animals,  the  phagocytic  action  of  cells 
performs  definite  services  in  the  course  of  normal  development. 
This  is  seen  most  markedly  in  some  insects  (diptera)  in  which  the 
destruction  of  larval  organs,  useless  to  the  adult  animal,  may  be  en- 
tirely accomplished  by  the  action  of  phagocytic  cells,  and  a  similar 
process  may  accompany  the  transformation  of  the  tadpole  to  the  adult 
in  many  amphibia.9  In  higher  animals  the  removal  of  extravasations 
of  blood  is  accompanied  by  a  train  of  occurrences  which  is  readily 
subjected  to  study.10  In  such  cases  the  leukocytes  rapidly  enter  the 
area  of  extravasation  and  an  engulfment  of  the  blood  cells  occurs, 
followed  by  a  process  of  digestion  entirely  analogous  to  the  digestion 
of  similar  blood  elements  by  the  various  forms  of  intestinal  hem- 
amebse.  In  the  latter  case  it  is  a  process  of  normal  digestion,  in  the 

9  See  Henneguy.     "Les  Insectes,"  Paris,  1904,  p.  677. 
10  Langhans.     Virchow's  Archiv,  Vol.  49,  1870. 


276  INFECTION    AND    RESISTANCE 

former  an  emergency  procedure  carried  out  by  virtue  of  the  retained 
ancestral  characteristics  of  the  special  phagocytic  cells. 

The  leukocytes,  whose  chief  functions  seem  to  be  associated  with 
such  processes  of  intracellular  digestion,  may,  therefore,  be  looked 
upon  as  cells  retaining  primitive  characteristics  for  definite  physio- 
logical purposes.  We  shall  see,  however,  that,  to  meet  exceptional 
conditions,  the  process  of  phagocytosis  may  be  carried  out  also  by 
many  other  cells  which  are  associated  ordinarily  with  functions  en- 
tirely apart  from  this  phenomenon. 

During  normal  life  in  higher  animals,  too,  constant  destruction 
of  red  blood  cells  by  phagocytosis  takes  place  in  the  spleen  and  liver, 
and  is  described  by  Dickson  1  1  as  occurring  in  the  bone  marrow  as 
well ;  and  similar  phagocytosis  of  red  cells  is  seen  in  the  hemolymph 
nodes.  It  is  claimed  by  Metchnikoff,  furthermore,  that  many  of  the 
degenerative  and  retrogressive  processes  which  take  place  in  the 
human  body  are  carried  on  by  the  mechanism  of  phagocytosis.  The 
rapid  return  of  the  puerperal  uterus  to  the  normal  state  is  explained 
in  this  way,  and  work  by  Helme  12  seems  to  show  that  there  is  an 
actual  phagocytosis  of  the  hyperplastic  uterine  musculature  during 
this  period.  The  atrophic  changes  of  senility,  too,  are  attributed  by 
Metchnikoff 13 14  to  the  same  processes.  The  involution  of  the 
ovaries  is  accompanied  by  active  phagocytosis  of  portions,  of  this 
organ,  and  Metchnikoff  claims  further  to  have  shown  that  the  de- 
generation of  the  nervous  system  during  old  age  is  accomplished  by 
the  phagocytosis  of  nerve  cells  by  phagocytic  elements  derived  either 
from  the  leukocytes  or  the  neuroglia,  or  from  both.15  The  whitening 
of  the  hair,  both  in  human  beings  and  in  old  animals  (dogs),  is  simi- 
larly due,  he  claims,  to  phagocytosis  of  the  pigment  by  cells  which 
wander  in  from  the  root  sheaths.  It  is,  up  to  the  present  time,  im- 
possible to  determine  the  stimulus  to  which  this  phagocytosis  is  due. 

Since  the  subject  is  a  very  important  one,  many  studies  have  been 
made  to  determine  which  cells  of  the  body  of  higher  animals  can 
take  in  and  digest  foreign  particles  and  to  classify  them  according 
to  this  power.  Metchnikoff  has  distinguished  between  the  "motile" 
and  "fixed"  phagocytes,  the  former  the  leukocytes  of  the  circulating 
blood,  the  latter  certain  connective  tissue  cells,  endothelial  cells, 
splenic  pulp  cells,  and  certain  cellular  elements  of  the  lymph  nodes, 

11  Dickson.     "The  Bone  Marrow,"  Longmans,  Green,  London,  1908. 

12  Helme.     Transact.  Roy.  Soc.  of  Edinburgh,  Vol.  35,  1889.     Cited  from 
Metchnikoff. 

13  Matschinsky.     Ann.  de  I'Inst.  Past.,  Vol.  14,  1900. 
u  Metchnikoff.     Ann.  de  I'Inst.  Past.,  Vol.   15,  1901. 

15  That  the  leukocytes  are  concerned  in  the  destruction  and  resorption 
of  dead  tissues  has  been  shown  by  Leber  especially  (Leber,  "Die  Entstehung 
der  Entziindung,"  Leipzig,  Engelmann,  1891).  An  accumulation  of  leukocytes 
about  a  bacterial  focus  or  from  any  other  stimulus  is  followed  by  tissue 
lysis  due  to  leukocytic  enzymes. 


PHAGOCYTOSIS 


277 


POLYNUCLEAR    LEUKOCYTES     TAKING    UP     STA- 
PHYLOCOCCI. 


the  neuroglia  tissue,  and,  in  fact,  all  phagocytic  cells  which  are 
ordinarily  confined  to  some  definite  localization  in  the  body.  Among 
phagocytic  cells  Metchnikoff  further  distinguishes  between  "micro- 
phages,"  by  which  he  designates  the  polymorphonuclear  leukocytes 
of  the  circulating  blood  and 
"macrophages."  The  ma- 
crophages  include  the  fixed 
cells  mentioned  above,  to- 
gether with  the  large  mono- 
nuclear  elements  of  the 
blood,  in  short,  all  phago- 
cytic cells  except  the  micro- 
phages. 

Although  no  absolute 
functional  differentiation  is 
possible  between  the  two,  it 
is  true,  in  a  general  way, 
that  the  microphages  are 
concerned  primarily  with 
the  phagocytosis  of  bacteria 
and  especially  of  those 
which  invade  acutely,  while 
the  macrophages  are  con- 
cerned especially  with  the  resorption  of  cellular  detritus,  foreign 
bodies,  and  such  bacteria  as  are  more  chronic  in  their  activities,  or 

are  peculiarly  insoluble. 
On  the  other  hand,  micro- 
phages may  take  up  foreign 
particles  and  bacteria  of  all 
kinds  under  suitable  condi- 
tions, and  no  sharp  line  can 
be  drawn  between  the  two 
varieties  in  this  respect. 
Metchnikoff  further  be- 
lieves that  the  two  classes 
of  phagocytic  cells  differ  in 


the  nature  of  the  protective 
substances  they  secrete  and 
furnish  in  the  blood 
plasma.  This,  however,  is 
a  problem  concerning  which 
there  is  much  difference  of 
opinion  and  which  calls  for 
KUPFER  CELLS  CONTAINING  MALARIAL  PIG-  ~prmrfltp  rliomiccinTi  in  an 

MENT.       UNGRAMMATICALLY    DRAWN     FROM 

A   SECTION   OF    MALARIAL   LIVER   KINDLY      other  place. 

FURNISHED  BY  DR.  E.  LAMBERT.  The  property  of  phago- 


278 


INFECTION    AND    RESISTANCE 


EAT  LEPROSY  BACILLI  GROUPED  IN  THE  EEMAINS   OF 
DEAD  SPLEEN  CELLS  GROWING  IN  PLASMA. 


cytosis  is  therefore 
an  attribute  of  a 
considerable  num- 
ber of  different  va- 
rieties of  cells.  In 
the  circulating 
blood  the  polynu- 
clear  leukocytes  are 
the  most  actively 
motile  and  phago- 
cytic  elements.  The 
eosinophile  cells 
may  also  take  up 
foreign  particles 
and  bacteria,  as 
may  also  the  large 
lymphocytes.  The 


Drawn  after  illustration  in  Zinsser  and  Carey,  Journal  small  lymphocytes 
of  the  A.  M.  A.,  Vol.  58,  1912.  and  mast  cells  are 

either  entirely  inac- 
tive in  this  respect,  or,  at  least,  possess  phagocytic  powers  under  ex- 
ceptional circumstances  only.  This  does  not  mean,  however,  that 

these  last-named  cells  may 

not  accumulate  at  the  point 

of   invasion   nor  that  they 

may  not  play  an  important 

part  in  the  defence  of  the 

body.     It  is  well-known,  of 

course,  that,  in  tuberculosis 

and  a  number  of  other  con- 
ditions, the  lymphocytes 

may  form  the  majority  of 

the  cellular  elements  which 

accumulate    at   the   site    of 

the     lesion.       Among     the 

fixed  cells  of  the  body  it  is 

probable   that   phagocytosis 

may  be  carried  on  by  cells 

<of   many  different   origins, 

though  the  identification  of 


PHAGOCYTOSIS  or  SENSITIZED  PIGEON  COR- 
PUSCLES BY  ALVEOLAR  CELLS  OF  LUNG. 

Drawing  made  after  photomicrograph  pub- 
lished by  Briscoe,  Journal  of  Path,  and 
Bact.,  Vol.  12,  1908. 


•cells  in  tissues  is  often  a 
purely  morphological  prob- 
lem, and  therefore  fraught 
with  many  possibilities  of 
error.  Probably  the  most  active  fixed  tissue  cells  are  the  endothelial 
cells  of  the  blood  vessels  and  those  which  line  the  serous  cavities,  the 


PHAGOCYTOSIS 


279 


sinuses  of  the  lymphnodes,  and  of  the  spleen.  However,  there  are 
many  other  cells  in  addition  to  these  which  may  be  phagocytic.  The 
writer,  with  Carey,16  has  observed  the  active  phagocytosis  of  leprosy 
bacilli  by  cells,  probably  of  connective  tissue  origin,  growing  from 
plants  of  rat  spleen  in  plasma.  Phagocytosis  by  the  cells  lining  the 
alveoli  of  the  lungs  has  been  observed  by  Briscoe.17  This  author  made 
the  interesting  observation  that  in  cases  of  mild  infection  such  cells 
can  free  the  lungs  of  micro-organisms  entirely  without  aid  from  the 
leukocytes  of  the  circulating  blood.  It  is  these  cells,  too,  which,  in 
the  ordinary  conditions  of  life,  take  up  the  inhaled  particles  of  dust 
and  are,  therefore,  often  spoken  of  as  dust  cells.  The  origin  of  the 
dust  cells  has  often  been  the  subject  of  controversy.  In  the  embryo 
the  alveoli  of  the  lung,  like  the  bronchi,  are  lined  with  columnar  cells 
which  are  transformed  into  flattened  epithelium  as  the  alveoli  ex- 
pand at  the  first  inspirations  after  birth.  These  flattened  cells, 
which  constitute  the  alveolar  or  dust  cells,  are  probably  of  epithelial 
origin,  and  as  such  are  probably  the  only  epithelial  cells  which  act 
as  phagocytes  under  ordinary  conditions.  Although  no  positive 
general  statement  is  justified,  we  can  yet  say  with  reasonable  accuracy 
that  among  the  phagocytic  fixed  tissue  cells  the  most  important  are 
the  connective  tissue  and  endothelial  cells. 

The  type  of  phagocy- 
tosis and  the  variety  of  cell 
whic,h  participates  in  it 
seem  to  depend  to  a  great 
extent  upon  the  nature  of 
the  substance  which  incites 
the  process,  or  rather  at 
which  the  process  is  aimed. 
Thus  the  large  cells  which, 
in  tissues,  take  up  the  lep- 
rosy bacillus,  those  which 
are  characteristic  of  tuber- 
culous foci,  or  those  caused 
by  blastomycetes,  or  by  for- 
eign bodies,  all  have  special 
appearances  which  are  suf- 
ficiently characteristic  to 
have  diagnostic  value. 

However,  it  is  difficult  to  determine  with  certainty  the  origin 
of  the  cells  which  participate.  The  chemical  nature  of  the  substances 
taken  up,  moreover,  often  complicates  the  phagocytic  process  in  such 
a  way  that  different  cellular  elements  are  enlisted  in  succession  in 
order  that  the  ingested  substances  may  be  disposed  of.  Thus  tubercle 

16  Zinsser  and  Carey.    Jour.  A.  M.  A.,  March,  1912,  Vol.  58. 

17  Briscoe.     Jour,  of  Path,  and  Bacter.,  Vol.  12,  1907. 


GIANT  CELL  IN  TUBERCULOSIS. 


280  INFECTION  AND  RESISTANCE 

or  leprosy  bacilli  which  are  injected  into  an  animal  may  be  at  first 
taken  up  by  polynuclear  leucocytes  or  microphages,  by  which  they  may 
even  be  carried  into  the  lymph  channels  and  distributed,  perhaps  to 
the  detriment  of  the  host.  But  these  cells,  probably  because  they  lack 
a  lipolytic  ferment  by  means  of  which  the  waxes  of  the  acid-fast  or- 
ganisms can  be  digested,  cannot  destroy  the  bacteria,  which  are  then 
attacked  by  other  cellular  elements  at  the  site  of  their  final  deposit. 

That  fixed  tissue  cells  as  a  matter  of  fact  play  a  very  important 
role  in  the  disposal  of  invading  bacteria  is  becoming  more  and  more 
clear.  Kyes  ls  showed  a  few  years  ago  that  the  immunity  of  the 
pigeon  to  pneumococcus  infection  is  largely  due  to  active  removal  of 
the  bacteria  by  the  Kupfer  cells  in  the  liver.  Bull's  19  observations 
on  the  intravascular  agglutination  of  typhoid  bacilli  which  are  then 
phagocyted  by  cells  in  the  liver  and  spleen  point  in  the  same  direc- 
tion, and  recently  Hopkins  and  Parker  19a  in  our  laboratory  have 
shown  that  streptococci  injected  into  rabbits  and  cats  are  rapidly  re- 
moved from  the  circulation,  the  removal  being  to  a  great  extent  due 
to  phagocytosis  carried  on  by  the  endothelial  cells  in  the  lungs  and 
by  similar  cells  in  the  liver  and  spleen. 

In  many  such  cases  the  further  resolution  of  the  foreign  sub- 
stance is  accomplished  by  an  important  type  of  phagocytosis  which  is 
characterized  by  the  formation  of  the  so-called  giant  cells.  These 
cells  are  of  varying  appearance  in  different  conditions  and  locations. 
Thus  the  giant  cells  which  form  about  foreign  bodies,  such  as  the 
small  cotton  fibers  occasionally  left  in  wounds,  or  injected  particles  of 
paraffin  or  iron  splinters,  etc.,  are  quite  characteristic  and  distinct 
from  the  giant  cells  of  tuberculous  foci,  or  of  rhinoscleroma,  glanders, 
or  leprosy.  They  are  all  large  cells,  containing  often  numerous 
nuclei  which  form  either  by  the  fusion  of  several  cells,  as  claimed  by 
Borrell,19b  Hektoen,19c  and  others,  or  by  the  cleavage  of  the  nuclei 
alone,  without  coincident  divisions  of  the  cytoplasm. 

Although  it  is,  of  course,  impossible  to  decide  definitely  upon 
purely  morphological  grounds,  the  researches  of  Hektoen  especially 
would  lead  one  strongly  to  favor  the  former  view.  It  is  equally 
difficult  to  decide  the  origin  of  giant  cells,  and  endothelial,  connec- 
tive tissue,  and  even  leucocytic  origin  has  been  claimed  for  them. 
Yet  in  no  case  has  it  thus  far  been  possible  to  actually  observe  their 
formation  by  a  method  which  could  positively  decide  this  point. 

In  order  to  gain  a  clear  conception  of  the  participation  of  phago- 
cytes in  the  response  of  the  body  to  injury  or  invasion,  it  will  be 
useful  to  follow  out  the  process  of  inflammation  as  it  occurs  in  the 

18  Kyes.    Journal  of  Infectious  Diseases,  Vol.  18,  1916,  p.  272. 

19  Bull.    Journal  of  Exp.  Med.,  Vol.  XX,  p.  237. 
19a  Hopkins  and  Parker.    Paper  in  Manuscript. 
19b  Borrell.    Ann.  de  I'Inst.  Past.,  7,  1893. 

19c  Hektoen.    Jour.  Exp.  Med.}  3,  1898,  p.  21. 


PHAGOCYTOSIS 


281 


higher  animals.  Inflammation  may  be  incited  by  a  large  number  of 
agencies — chemical  irritants,  mechanical  injury,  or  even  by  the  in- 
troduction of  inactive  and  isotonic  substances  such  as  broth  or  salt 
solution.20  21  Yet  in  these  cases  the  response,  though  essentially 
'similar  in  principle  to  that  following  invasion  by  bacteria,  lacks 
certain  features  especially  interesting  in  the  present  connection,  and 
it  will  be  most  profitable  for  our  purpose  to  consider  in  detail  the 
result  of  infection  with  pathogenic  micro-organisms. 

If  an  emulsion  of  pyogenic  staphylococci  is  injected  into  an  ani- 
mal subcutaneously  the  site  of  injection  will  soon  become  reddened 
and  swollen  and  microscopic  examination  will  show,  within  a  few 
hours,  a  swelling  and  engorgement  of  the  blood  vessels. 

The  injected  cocci  will  be  found  to  lie  partly  scattered  in  the 
tissue  spaces,  in  part  within  polynuclear  leukocytes  and  connective 
tissue  cells  which  have  begun  to  ingest  them.  The  tissue  spaces 
will  be  swollen  and 
stretched  by  the 
exudation  of  blood 
serum  from  the  ves- 
sels. This  condi- 
tion will  begin  in 
from  4  to  6  hours 
after  injection  and 
increase  during  the 
next  24  hours  in  ex- 
tent and  severity, 
according  to  the 
quantity  and  viru- 
lence of  the  cocci 
injected.  The  con- 
ditions which  pre- 
cede the  wandering 
of  the  p  o  1  y  m  o  r- 
phonuclear  leuko- 
cytes out  of  the  ves- 
sels have  been  care- 
fully studied  in  such  thin  tissues  as  the  mesentery  of  a  frog  after 
injury  by  trauma  or  acid.  Within  the  vessels  of  the  affected  area 
there  is  at  first  an  acceleration  of  the  blood  stream,  then  a  dilatation 
of  the  capillaries  and  a  slowing  of  the  current.  Leukocytes  may  now 
be  observed  moving  more  slowly  than  the  main  stream,  and  keeping 
close  to  the  periphery  along  the  walls  of  the  vessels.  Here  and  there 
they  seem  interrupted  in  their  movements  and  adhere  to  the  vascular 
wall.  A  little  later  these  cells  appear  to  pass  through  the  wall  of  the 

20  See  Adami.     "Inflammation,"  Macmillan,  London,  1909. 

21  See  Adami,  loc.  cit. 


DIAGRAMMATIC  EEPRESENTATION  OF  LEUKOCYTES  WAN- 
DERING THROUGH  CAPILLARY  WALLS. 
Adapted    from   Eibbert,    "Lehrbuch    der    Allgemeinen 
Pathologie,"  p.  337. 


282  INFECTION    AND    RESISTANCE 

vessel  by  sending  out  pseudopodia  which  slowly  penetrate  it.  Adami 
states  that  if,  at  this  stage,  the  tissues  be  excised,  fixed  in  osmic 
acid,  and  stained,  leukocytes  may  be  seen  crowding  the  inner  sur- 
face of  the  vessel  in  all  stages  of  transition  from  its  anterior  to 
the  lymph  spaces  on  the  outside. 

In  the  staphylococcus  infection,  after  from  12  to  4$  hours,  there 
will  be  seen  the  results  of  an  active  and  destructive  struggle  between 
the  invading  bacteria  and  the  defending  cells.  In  the  center  of  the 
area  of  invasion  tissue  has  been  destroyed  and  disintegrated.  Amid 
the  necrotic  detritus,  closely  packed,  lie  leukocytes  and  cocci  and 
active  phagocytosis  has  taken  place.  In  some  cases  the  intracellular 
bacteria  appear  swollen  and  disintegrating,  in  others  the  leukocyte 
itself,  overcome  by  the  larger  number  of  bacteria  it  has  taken  in, 
becomes  vacuolated,  indefinite  in  outline,  and  apparently  is  being 
itself  destroyed.  The  presence  of  blood  serum,  which  is  aiding  in 
the  destruction  of  bacteria  both  by  its  bactericidal  powers  and  its 
reenforcement  of  the  phagocytic  process,  renders  this  mass  fluid  or 
semi-fluid,  and  the  whole  mixture  constitutes  what  is  known  as  pus. 
Around  the  periphery  cocci  and  leukocytes  become  more  scattered 
and  sparse,  and  bacteria,  together  with  leukocytes,  loaded  with  cocci, 
may  be  seen  lying  within  large  mononuclear  cells  (macrophages). 
Whether  the  process  goes  on  to  further  extension  or  is  eventually 
walled  off  into  a  distinct  abscess  by  the  formation  of  granulation 
tissue  and  new  connective  tissue  depends  upon  the  balance  of  forces 
between  attacking  agent  and  defensive  factors. 

If  we  inject  a  similar  emulsion  of  cocci  into  the  pleural  or  peri- 
toneal cavity  of  an  animal  a  process  similar  in  principle  may  be 
observed. 

Normally  the  peritoneum  contains  a  small  amount  of  this  serous 
fluid  and  a  moderate  number  of  white  blood  cells,  chiefly  lympho- 
cytes. When  any  substance,  broth  or  salt  solution,  an  aleuronat  or 
a  bacterial  emulsion,  is  injected  into  the  peritoneal  cavity,  there 
follows  a  brief  period  during  which  there  is  a  diminution  of  the 
free  cellular  elements  in  the  peritoneal  fluid.  At  this  time  there  is 
a  clumping  of  cells  in  the  folds  of  the  omentum  and  mesentery,  a 
transient  stage  of  flight  away  from  the  point  of  injury.  This,  how- 
ever, is  soon  over.  Within  one  to  two  hours  an  active  immigration 
of  leukocytes  into  the  serous  cavity  occurs  and  if,  during  the  next 
12  to  24  hours  small  quantities  of  fluid  are,  from  time  to  time,  with- 
drawn with  a  capillary  pipette,  a  rapid  and  constant  increase  of 
leukocytic  elements,  chiefly  of  the  microphage  or  polynuclear  type, 
is  observed.  If  the  injected  substance  has  been  a  sterile,  harmless 
fluid,  a  gradual  return  to  normal  within  48  hours  then  ensues.  If, 
however,  we  have  injected  bacteria,  a  struggle  similar  to  the  one 
described  above  takes  place  within  the  peritoneum,  and  active 
phagocytosis  of  the  micro-organisms  takes  place. 


PHAGOCYTOSIS  283 

Let  us  suppose  that  the  injected  bacteria  have  been  small  in 
quantity  and  moderate  in  virulence.  In  such  a  case  a  rapid  phago- 
cytosis gradually  rids  the  fluid  of  micro-organisms  and  within  24 
hours  after  injection  few,  if  any,  free  bacteria  are  visible. 

A  little  exudate  taken  at  this  time  shows  large  numbers  of  micro- 
phages  varyingly  crowded  with  well-preserved  and  disintegrating 
bacteria.  Some  of  the  phagocytes,  having  literally  taken  up  more 
than  they  can  digest,  are  vacuolated  and  disintegrating,  but,  in  gen- 
eral, the  victory  lies  with  the  cells.  A  little  later  large  mononuclear 
elements  appear,  and  here  and  there  will  be  seen  to  take  up  dead 
leukocytes  together  with  ingested  cocci.  In  this  way  gradually  a 
cleaning  out  of  the  peritoneum  takes  place,  the  animal  recovers,  and 
the  peritoneum  returns  to  normal. 

Let  us  suppose,  on  the  other  hand,  that  the  bacteria  injected  are 
in  larger  doses  and  of  greater  virulence.  In  such  a  case,  after  a 
period  of  active  phagocytosis,  there  may  be  a  gradual  increase  of 
bacteria  over  leukocytes.  The  phagocytic  cells  are  found  to  be  under- 
going degeneration  in  larger  numbers,  the  free  bacteria  increase, 
and  the  impending  death  of  the  animal  can  often  be  foretold  by  the 
appearance  of  the  exudate.  Finally,  the  peritoneal  fluid  may  con- 
sist chiefly  of  free  and  rapidly  multiplying  bacteria  with  a  practical 
absence  of  phagocytic  cells. 

In  all  of  the  processes  so  far  as  described  the  burden  of  the 
defence  has  fallen  upon  the  microphages  or  polynuclear  leukocytes, 
while  the  macrophages — endothelial  and  connective  tissue  cells — 
have  taken  a  purely  secondary  part  in  the  reaction,  forming,  to  some 
extent,  a  second  line  of  defence,  or,  more  probably,  taking  part  only 
in  the  final  removal  of  degenerated  and  disintegrating  combatants 
and  tissue  detritus.  In  order  to  obtain  a  complete  conception  of 
phagocytosis  in  its  entire  significance  it  will  be  necessary  to  consider 
a  further  example,  namely,  the  process  which  takes  place  within  tis- 
sues in  the  course  of  the  efforts  of  macrophages  to  remove  bacteria 
and  other  substances  which,  either  because  of  their  insolubility  or  for 
other  unknown  reasons,  are  refractory  to  the  attacks  of  the  mi- 
crophages. Since  we  are  interested  in  this  subject  chiefly  from  the 
point  of  view  of  the  defence  against  bacteria,  we  may  illustrate  this 
process  best  by  the  description  of  the  reaction  which  takes  place  when 
tubercle  bacilli  become  localized  anywhere  within  the  animal  body. 

When  tubercle  bacilli  are  injected  into  the  peritoneum  they  are 
actively  taken  up  by  the  polynuclear  leukocytes  just  as  are  other 
bacteria  and  many  entirely  inactive  solid  particles.  A  similar  inges- 
tion  by  microphages  may  take  place  in  the  folds  of  the  intestinal 
mucosa  if  tubercle  bacilli  are  fed  to  guinea  pigs.  However,  this 
preliminary  phagocytosis  is  probably  of  but  secondary  significance 
in  the  combat  of  the  body  against  tuberculosis,  since  it  has  still  to 
be  shown  that  polynuclear  leukocytes  are  capable  of  digesting  and 


284  INFECTION    AND    RESISTANCE 

destroying  acid-fast  bacilli.  Indeed,  much  evidence  tends  to  show 
that  the  ingestion  of  tubercle  bacilli  by  microphages  may  be  a  detri- 
ment to  the  host,  since  the  bacilli  by  this  means  are  carried  through 
the  lymphatics  and  variously  distributed  throughout  the  body.  Poly- 
nuclear  leukocyte  extracts,  though  containing,  as  we  shall  see,  pro- 
teolytic  enzymes,  do  not,  according  to  Tschernorutzky,  contain  any 
lipase,  and  it  may  well  be  that  for  this  reason  they  are  unable  to 
attack  the  waxy  substances  which  form  an  integral  part  of  these  or- 
ganisms. This  is  in  keeping  with  the  observations  made  by  Terry 
in  our  laboratory,  that  rat  leprosy  bacilli  may  be  kept  within  leu- 
kocytes for  weeks  without  losing  their  acid-fast  properties,  whereas 
the  same  bacilli,  as  the  writer  and  Gary  found,  were  rapidly  disin- 
tegrated in  spleen  cells  growing  in  plasma.  Moreover,  it  is  well 
known  that  the  estimation  of  tuberculo-opsonin  contents  of  the  sera 
of  tuberculous  patients  has  been  peculiarly  unsatisfactory  in  throw- 
ing light  on  the  progress  of  the  disease.  It  would  seem,  therefore, 
that  in  this  disease,  as  well  as  in  others  caused  by  acid-fast  organ- 
isms, the  microphages  play  only  an  unimportant  part  in  the  defence 
of  the  body. 

On  the  other  hand,  when  tubercle  bacilli  are  deposited  either  in 
a  lymphnode  (through  the  vehicle  of  leukocytes)  or  in  a  capillary 
anywhere  by  the  blood  stream,  a  train  of  cellular  changes  is  initiated 
in  which  the  predominant  part  is  played  by  the  macrophages.  The 
tubercle  bacilli  so  deposited  are  rapidly  surrounded  by  large  mono- 
nuclear  cells,  probably  endothelial  in  origin.  Some  of  the  micro- 
organisms may  even  be  phagocyted  and  taken  into  these  cells.  These 
cells,  spoken  of  as  "epithelioid  cells,"  surround  the  clump  of  bacteria 
in  more  or  less  concentric  rings,  and  around  these  there  is  an  accumu- 
lation of  leukocytes,  largely  of  the  lymphocyte  variety,  with  an  ad- 
mixture of  a  very  few  microphages.  Then  by  the  fusion  of  endothe- 
lial cells,  or  possibly  by  division  of  the  nuclei  of  some  of  these  cells 
within  the  individual  cell  bodies,  giant  cells  are  formed  which  take 
up  the  bacilli.  The  further  progress  of  the  tubercle  now  greatly 
depends  upon  the  balance  of  power.  Often  such  a  tubercle  may 
heal,  possibly  because  of  complete  intracellular  digestion  of  the  ba- 
cilli. On  the  other  hand  growth  and  multiplication  may  lead  to  a 
slow  and  dry  necrosis  of  the  center  of  such  a  mass  of  cells,  leading 
to  the  condition  spoken  of  as  caseation.  Epithelioid  cells  lose  their 
outlines  and  staining  properties,  and  go  to  pieces.  The  center  of  the 
lesion  is  a  grumous  mass,  the  periphery  shows  a  few  giant  cells  and 
connective  tissue  proliferation. 

It  is  always  surprising  to  those  who  study  these  lesions  for  the 
first  time  how  rarely  they  succeed  in  finding  tubercle  bacilli  in 
microscopic  sections  prepared  from  such  tubercles  by  the  ordinary 
Ziehl-E"eelsen  method  of  staining.  Repeated  and  careful  examina- 
tion of  such  material  may  fail  to  reveal  any  acid-fast  organisms, 


PHAGOCYTOSIS  285 

though  inoculation  into  guinea  pigs  is  nevertheless  successful,  pro- 
ducing typical  tuberculosis.  Much  22  has  studied  this  peculiar  state 
of  affairs  particularly  and  has  shown  that,  although  such  lesions  may 
show  no  tubercle  bacilli  by  the  Ziehl-Neelsen  carbol-fuchsin  method, 
staining  by  a  modified  Gram  technique  will  reveal  numerous  Gram- 
positive  rods  and  granules  which  have  lost  their  acid-fast  properties. 
This,  too,  if  true,  and  the  evidence  is  very  much  in  its  favor,  would 
point  to  an  ability  of  the  macrophages  to  digest  the  waxy  substance 
•of  the  tubercle  and  other  acid-fast  bacilli,  a  property  not  possessed 
by  the  microphages.  It  may,  of  course,  mean  on  the  other  hand  that 
the  tubercle  bacilli  in  the  lesion  have  not  developed  the  waxy  condi- 
tion. 

ClIEMOTAXIS    AND    LEUKOCYTOSIS 

The  part  played  by  the  phagocytic  cell  in  the  defence  of  the  body 
against  the  entrance  of  bacteria  and  other  foreign  substances  consists, 
then,  of  two  functionally  different  phases.  The  first  is  an  active 
motion  of  the  cells  toward  the  point  attacked,  and  their  accumulation 
about  the  noxious  agent,  the  second  consists  in  the  act  of  ingestion 
itself. 

The  motion  of  the  leukocytes  toward  the  invading  substances 
indicates  a  sensibility  on  the  part  of  the  cell  to  changes  in  its  environ- 
ment incited  by  the  foreign  agent,  and  since  the  stimuli  most  likely 
to  reach  the  leukocytes  and  bring  about  this  alteration  in  the  direc- 
tion of  their  movements  are  chemical  in  nature,  the  phenomenon  is 
spoken  of  as  "chemotaxis."  This  term  was  borrowed  from  Pfeffer,23 
who  studied  similar  phenomena  in  connection  with  many  freely 
motile  plant  cells,  spermatozoa,  and  bacteria.  Since  the  change  of 
direction  brought  about  in  a  moving  cell  by  such  influences  may  be 
such  as  either  to  attract  or  to  repel,  the  term  "positive  chemotaxis" 
is  used  to  designate  the  former  and  that  of  "negative  chemotaxis" 
the  latter. 

The  property  of  chemotaxis  is  of  vital  interest  in  the  present 
connection,  since,  whatever  may  be  our  opinion  regarding  the  relative 
values  of  phagocytosis  and  serum  protection  in  immunity,  the  great 
importance  of  the  phagocytic  process  cannot  be  questioned,  and  any 
agency  which  repels  the  approach  of  the  phagocytes  must  be  a  detri- 
ment, while  any  factor  which  attracts  them  is,  of  necessity,  a  power- 
ful means  of  defence.  In  the  investigations  upon  the  nature  of  infec- 
tious diseases  attention  'has  been  concentrated  upon  the  phenomenon 
of  phagocytosis,  and  the  relations  governing  the  act  of  ingestion 
have  been  very  thoroughly  studred.  The  details  of  the  chemotactic 
phenomenon,  however,  though  of  equal  importance,  are  much  more 

22  Much.     "Beitrage  zur  Klinik  der  Tuberk.,"  Vol.  8,  1907,  Hft.  1  and  4. 

23  Pfeffer.    "Untersuch.  a.  d.  Botan.  Inst.  Tubingen,"  Vols.  1  and  2,  1884 
sincl  1S88. 


286  INFECTION    AND    RESISTANCE 

obscure.     A  large  part  of  our  sparse  knowledge  in  this  connection, 
moreover,  has  been  gained  by  studies  not  related  to  infection. 

The  stimuli  which  determine  the  motion  of  cells  are,  of  course, 
not  necessarily  chemical,  and  extensive  studies  have  been  made  upon 
the  effect  of  light  waves  in  this  connection.  Although  these  inves- 
tigations are  of  great  biological  importance,  they  have  little  direct 
bearing  upon  the  problems  of  tropism  as  related  to  bacteria  and  leu- 
kocytes and  cannot  therefore  be  considered  here. 

Some  of  the  earlier  researches  upon  chemotaxis  were  those  made 
by  Stahl  24  upon  the  slime-molds  or  myxomycetes.  These  organisms 
possess  the  power  of  ameboid  motion,  and  were  observed  by  Stahl  to 
move  toward  or  away  from  any  given  region,  according  to  the  nature 
of  the  substances  with  which  they  came  in  contact.  Pfeffer  sub- 
jected this  phenomenon  to  closer  analysis.  Working  with  the  sperma- 
tozoa of  ferns,  swarm  spores,  bacteria  and  infusoria,  he  elaborated 
an  ingenious  technique  by  means  of  which  he  was  enabled  to  de- 
termine directly  the  negative  or  positive  chemotactic  properties  of 
various  substances  in  solution  upon  these  motile  forms.  His  tech- 
nique was  exceedingly  simple.  Capillary  glass  tubes,  about  8  to  10 
mm.  long  and  0.1  mm.  in  diameter,  were  sealed  at  one  end  in  the 
flame,  and  then  dropped  into  a  watch-glass.  The  solution  which  was 
to  be  tested  was  poured  over  the  tubes  and  the  watch-glass  then 
placed  under  the  bell  of  an  air-pump.  When  the  air  was  evacuated 
and  pressure  reduced  the  tubes  became  partly  filled  up  with  the 
liquid.  They  were  then  removed,  washed  in  water,  and  placed  under 
a  cover  slip  under  which  a  preparation  of  the  motile  cells  was  swim- 
ming. Positive  chemotaxis  was  indicated  by  entrance  of  the  cells 
into  the  tubes,  negative,  by  their  refusal  to  enter.  Failure  of  the 
solution  to  exert  any  chemotactic  influence  resulted  in  their  moving 
into  and  out  of  the  tubes  indiscriminately.25 

By  this  technique  a  large  number  of  interesting  observations  were 
made  which  threw  much  light  upon  the  causes  underlying  the  move- 
ments of  plant  cells.  For  instance,  in  investigating  the  spermatozoa 
of  the  ferns  it  was  found  that  they  were  attracted  strongly  by  malic 
acid  and  its  salts,  while  no  other  substance  investigated  approached 
these  compounds  in  the  intensity  of  positively  chemotactic  stimula- 
tion. From  this  Pfeffer  concludes  that  the  bursting  of  the  fern 
archegonia  is  accompanied  by  the  liberation  of  malic  acid,  this  at- 
tracting the  male  to  the  female  cell. 

Similar  experiments  have  been  carried  out  since  then  by  numer- 
ous naturalists,  among  them  Buller,26  Lidforss,27  and  Jennings,28 

24  Stahl.     Botanische  Zeitung,  1884. 

25  Buller.     Annals  of  Botany,  Vol.  16,  No.  56,  1900. 

26  Buller.    Loc,  tit. 

27  Lidforss.    "Jahrbiicher  f.  wissensch.  Botanik,"  41,  1904. 

28  Jennings.     "Behavior    of    Lower    Organisms/'    Columbia   Univ.    Press, 
Macmillan,  1906. 


PHAGOCYTOSIS  287 

and  it  has  been  found  that  in  addition  to  malic  acid  compounds  many 
other  substances,  organic  and  inorganic,  occurring  in  plant  cells  and 
cell-sap  exert  positive  chemotactic  power.  Lidforss  has  shown,  for 
instance,  that  calcium  chlorid  in  0.1  per  cent,  solution  may  strongly 
attract  plant  spermatozoids  (equisetum — horsetail).  When  the  solu- 
tion is  concentrated  to  1  per  cent.,  attraction  is  still  exerted,  but  the 
spermatozoids  immediately  lose  their  motility  upon  entrance  into  the 
fluid. 

The  same  worker  has  shown  that  a  substance  which  is  positively 
chemotactic  for  one  variety  of  plant  cell  may  be  negatively  chemo- 
tactic for  another,  showing  a  certain  selective  variation  which  should 
be  of  great  biological  importance.  Thus  capillaries  with  a  1  per  cent, 
solution  of  potassium  malate  actively  attracted  the  spermatozoids  of 
marchantia  (a  liverwort),  while  not  a  single  spermatozoid  of  equi- 
setum would  enter  these  tubes.  Low  29  has  applied  these  methods  of 
study  to  the  investigation  of  the  chemotaxis  of  mammalian  sperma- 
tozoa and  found  that  these  cells  were  actively  attracted  by  weakly 
alkaline  solutions. 

Studies  upon  the  factors  determining  the  movement  of  bacteria 
and  ameba3  toward  some  substances  and  away  from  others  have  been 
numerous,  and  are  valuable  for  the  understanding  of  leukocytic 
chemotaxis,  because  they  have  led  to  the  formulation  of  a  number  of 
important  general  theories.  The  fact  that  the  motions  of  bacteria 
in  suspensions  are,  to  a  certain  extent,  determined  by  the  negative 
electrical  charge  which  they  all  carry  in  neutral  media,  has  been 
touched  upon  in  the  section  on  agglutination.  Attempts  on  the  part 
of  Young  and  the  writer  to  determine  whether  the  attraction  of 
leukocytes  toward  bacteria  might  be  due  to  the  carrying  of  an  elec- 
tropositive charge  by  the  white  cells  have  met  with  no  result,  owing 
so  far  to  the  failure  to  elaborate  a  reliable  technique.  However,  this 
thought  is  not  an  impossible  one  and  should  be  borne  in  mind. 

That  certain  bacteria  will  wander  actively  toward  a  source  of 
oxygen  was  shown  by  Engelmann' s  30  classical  experiment  in  which  a 
diatom,  half  in  the  shade  and  half  in  the  light,  was  surrounded  by  an 
emulsion  of  bacteria,  and  these  were  seen  to  collect  about  the  lighted 
half  only,  where  oxygen  was  being  liberated  by  virtue  of  the  chloro- 
phyll. The  extreme  delicacy  of  chemotactic  reactions  is  illustrated 
in  these  experiments  in  that  Engelmann  calculated  that  the  bacteria 
reacted  to  one  one-hundred  billionth  of  a  milligram  of  oxygen.  The 
selective  reaction  of  bacteria  to  various  chemical  substances,  further- 
more, has  been  shown  by  allowing  different  solutions  to  diffuse  into 
bacterial  emulsions  from  capillary  tubes,  and  by  observing  attraction 
or  repulsion  from  the  point  of  contact. 

The  chemotaxis  of  leukocytes  has  opposed  more  difficulties  to 

29  Low.     Sitzungs  Berichte  kais.  Akad.  d.  Wiss.,  Wien,  Vol.  3,  Abf.  3. 
80  Engelmann.     Arch.  f.  d.  ges.  PhysioL,  Vol.  57,  p.  375. 


288  INFECTION    AND    RESISTANCE 

direct  study,  since  the  conditions  within  the  living  body  are  subject 
to  a  large  number  of  modifying  factors,  and  experiments  upon  the 
isolated  cells,  in  vitro,  even  under  conditions  of  the  most  careful 
technique,  are  fraught  with  much  unavoidable  injury  to  the  cells. 
However,  enough  has  been  learned  to  indicate  that  these  cells  are 
subject  to  the  phenomena  of  chemotaxis  or  tropism  just  as  are  inde- 
pendent unicellular  forms,  and  that  they  may  be  attracted  or  re- 
pelled by  a  variety  of  organic  and  inorganic  substances.  Leber  31 
was  one  of  the  first  to  study  this  in  his  work  upon  inflammation. 
He  found  that  leukocytes  were  actively  attracted  by  powdered  cop- 
per and  mercury  compounds,  but  not  by  powdered  gold  or  iron.  He 
also  observed  that  dead  bacteria  exerted  a  similar  positive  chemo- 
tactic  influence,  and  Buchner  32  later  succeeded  in  extracting  sub- 
stances from  various  bacteria  which  possessed  similar  properties.  It 
appears,  from  these  and  other  investigations,  that  the  power  of  stim- 
ulating positive  chemotaxis  is  a  general  property  of  bacterial  pro- 
teins, equally  evident  in  bacterial  extracts,  dead  bacteria,  or  the  liv- 
ing organisms.  It  is  likely,  therefore,  that  the  attraction  of  leu- 
kocytes toward  the  point  of  bacterial  invasion  is,  in  part  at  least, 
due  to  the  properties  of  the  bacterial  proteins  themselves.  That  this, 
however,  is  not  the  whole  story  is  evident  from  the  work  of  Massart 
and  Bordet,33  who  showed  that  the  products  of  cell  destruction  and 
disintegration  possess  similar  positively  chemotactic  properties.  This 
is  true  not  only  of  the  products  of  disintegrated  tissue  cells,  but  of 
those  of  the  destroyed  leukocytes  themselves.  Thus  it  appears  that 
when  any  injury  of  tissue  takes  place,  a  stimulus  which  attracts 
leukocytes  results,  even  when  the  injury  is  not  accompanied  by  bac- 
terial invasion.  This  would  explain  the  participation  of  leukocytes 
in  reactions  to  injury,  and  in  inflammations  not  of  bacterial  origin, 
and  their  local  accumulation  following  the  injection  of  insoluble 
inorganic  substances. 

When  bacteria  are  actually  present,  however,  the  added  stimulus 
due  to  the  diffusion  of  bacterial  proteins  probably  increases  the 
process  to  a  degree  often  sufficient  to  meet  the  added  requirements 
for  protection.  Following  this,  both  the  destruction  of  tissues,  of 
bacteria,  and  of  leukocytes  may  together  exert  a  cumulative  chemo- 
tactic power  which  continues  the  process  proportionately  with  the 
extent  of  the  lesion. 

It  is  of  the  utmost  importance,  therefore,  to  ascertain  whether 
or  not  any  substances  derived  from  bacteria  may,  under  any  circum- 
stances, exert  a  repellent  or  negatively  chemotactic  power.  If  we 
infect  an  animal  intraperitoneally  with  virulent  bacteria,  in  doses 

31  Leber.     FortscJir.  der  Med.,    1888;    also    "Die    Entstehung    der    Ent- 
ziindung,"  Engelmann,  Leipzig,  1891. 

32  Buchner.    Berl.  klin.  Woch.,  Vol.  27,  No.  30,  1890. 

33  Massart  and  Bordet.     Ann.  de  I'Inst.  Past.,  Vol.  5,  1891. 


PHAGOCYTOSIS  289 

sufficient  to  lead  to  death,  and  examine  the  peritoneal  exudate  just 
before  the  lethal  outcome,  we  may  observe  that  leukocytes  are  gradu- 
ally disappearing,  and  that  finally  but  a  few  will  be  present  and  the 
fluid  will  be  swimming  with  free  micro-organisms.  In  the  same  way 
it  is  well  known  that  the  diminution  of  leukocytes  in  the  circulating 
blood — or  even  the  failure  of  these  cells  to  increase  in  the  circulation 
in  the  course  of  such  diseases  as  pneumonia,  or  general  infections 
with  staphylococci  or  streptococci — is  seriously  prognostic  of  fatal 
outcome.  The  conditions  here  observed  point  strongly  to  the  ex- 
istence of  substances  of  negative  chemotactic  influence  which  protect 
the  bacteria,  not  from  phagocytosis  itself,  but  from  that  necessary 
forerunner  of  phagocytosis,  the  approach  of  the  leukocyte.  It  is 
necessary  to  draw  this  distinction  since  these  phenomena  are  not 
merely,  as  often  believed,  "antiopsonic,"  but  in  truth  largely  "anti- 
chemotactic."  It  is  true  that  Kanthack,34  and  more  especially 
Werigo,35  have  denied  the  existence  of  negatively  chemotactic  bac- 
terial products,  the  latter  basing  his  assertion  upon  the  observation 
that  active  phagocytosis  occurs  in  the  lungs,  liver,  and  spleen  of 
animals  dying  of  infection  with  virulent  germs.  However,  the*  argu- 
ments of  these  authors  are  not  conclusive  and  the  mass  of  experi- 
mental and  clinical  evidence  which  points  to  a  direct  failure  of 
leukocyte  accumulation  in  the  presence  of  virulent  bacteria  in  the 
animal  body  would  alone  suffice  to  render  such  conclusions  unlikely. 
Moreover,  strong  evidence  in  favor  of  the  existence  of  negatively 
chemotactic  influences,  is  brought  by  the  extensive  experiments  of 
Bail  upon  the  so-called  aggressins,  discussed  in  another  place,  and 
such  observations  as  those  of  Vaillard  and  Vincent  36  and  Vaillard 
and  Rouget,37  which  showed  that  the  injection  of  a  little  tetanus 
toxin  together  with  tetanus  spores  would  prevent  the  ingestion  of  the 
spores  by  leukocytes,  and  thereby  furnish  an  opportunity  for  germi- 
nation and  consequent  fatal  toxemia. 

Similar  observations  have  been  made  by  Besson  38  in  the  case  of 
the  bacillus  of  malignant  edema  by  the  use  of  the  original  technique 
of  Pfeiffer.  Capillary  tubes  containing  the  toxin  remained  free  of 
leukocytes  after  subcutaneous  introduction  into  guinea  pigs,  while 
similar  tubes  containing  the  culture  medium  alone,  or  the  bacilli  and 
their  spores,  attracted  leukocytes  in  considerable  numbers. 

It  is  possible,  of  course,  to  interpret  such  phenomena  as  due  to  a 
failure  of  positive  chemotaxis  rather  than  to  an  active  negative 
chemotaxis. 

Although  the  phenomena  of  chemotaxis  are  most  easily  studied 

34  Kanthack.     Quoted  from  Adami,  loc.  cit. 

35  Werigo.     Ann.  de  I'Inst.  Past.,  Vol.  8,  1894. 

36  Vaillard  and  Vincent.    Ann.  de  I'Inst.  Past.,  Vol.  5,  1891. 

37  Vaillard  and  Rouget.     Ann.  de  I'Inst.  Past.,  Vol.  6,  1892. 

38  Besson.    Ann.  de  I'Inst.  Past.,  Vol.  9,  1895. 


290  INFECTION    AND    RESISTANCE 

in  extravascular  inflammatory  changes,  there  is  none  the  less  a 
regular  and  apparently  purposeful  attraction  or  repulsion  of  leuko- 
cytes evident  in  the  circulating  blood  during  infectious  diseases. 
That  infection  of  the  body  with  many  micro-organisms  results  in  the 
increase  of  leukocytes,  and  that  in  others  there  is  either  no  increase 
or  even  a  decrease,  is  too  well  known  and  too  generally  applied  in 
diagnosis  and  prognosis  to  warrant  our  giving  up  much  space  to  a 
review  of  the  facts.  Nevertheless,  the  causes  which  lead  to  a  leuko- 
cytosis  in  the  one  case,  a  leukopenia  in  the  other,  are  still  very 
obscure  and  deserve  discussion. 

I  In  the  first  place  it  is  by  no  means  certain  whether  a  leuko- 
cytosis signifies  an  active  discharge  of  new  leukocytes  from  the  bone 
marrow  or  whether  it  means  simply  an  altered  distribution  in  that 
the  phagocytes  accumulated  in  the  lymphatic  and  other  organs  are 
attracted  by  chemotaxis  into  the  peripheral  circulation.  Studies  of 
the  bone  marrow  during  infection  as  well  as  the  occasional  appear- 
ance of  myelocytes  and  other  cells  ordinarily  found  only  in  the  bone 
marrow  during  health  would  point  toward  a  participation  of  active 
bone-marrow  hyperplasia  in  the  increase  of  peripheral  leukocytes. 
There  is  no  good  reason  to  doubt,  moreover,  that  a  chemotactic  stimu- 
lus exercised  in  the  circulation  should  withdraw  leukocytes  from 
any  place  of  accumulation  to  the  circulation.  Probably  both  proc- 
esses take  part.  When  bacteria  are  injected  into  the  circulation  of 
an  animal  there  is,  at  first,  a  moderate  diminution  of  the  leukocytes 
just  as  there  is  after  injection  of  bacteria  or  other  substances  into 
the  peritoneum.  This  is  soon  followed  in  most  cases  by  a  rapid  and 
progressive  increase,  in  which,  whenever  the  leukocytosis  is  one  of 
considerable  degree,  the  polynuclear  leukocytes  preponderate.  The 
extensive  clinical  study  of  the  white  cells  in  infectious  disease  of  the 
human  being  give  us  more  material  for  reasoning  in  this  respect 
than  we  have  available  from  animal  experiment.  Infection  with 
invasive  bacteria  such  as  the  pneumococcus  (and  Neufeld  and  others 
have  shown  that  most  lobar  pneumonias  are  accompanied  by  pneu- 
mococcus bacteriemia),  streptococci,  staphylococci,  and  others  is 
always  accompanied  by  an  increase  of  the  leukocytes,  while,  in 
typhoid  fever,  influenzal  infection,  tuberculosis,  and  a  number  of 
other  infections,  the  leukocytes  do  not  increase  and  may  even  de- 
crease. How  are  we  going  to  account  for  this  ?  That  all  these  bac- 
teria contain  a  substance  which  is  positive  in  its  chemotactic  effects 
is  easily  demonstrated  by  injecting  them  into  the  peritoneum  and 
observing  an  accumulation  of  leukocytes  and  a  consequent  phago- 
cytosis, even  in  the  cases  of  those  organisms  which  do  not  call  forth 
a  leukocytosis  in  the  blood  of  the  diseased  human  being.  Thus  it 
has  been  our  experience  as  well  as  that  of  others  invariably  to  ob- 
serve the  rapid  and  complete  polynuclear  phagocytosis  of  both 
leprosy  bacilli  and  tubercle  bacilli  after  injection  of  these  micro- 


PHAGOCYTOSIS  291 

organisms  into  the  peritoneal  cavities  of  guinea  pigs.  Yet  a  chronic 
tuberculous  peritonitis  or  pleurisy  is  characterized  usually  by  an 
exudate  which  contains  but  few  polynuclears  and  relatively  many 
lymphocytes.  A  final  explanation  of  these  conditions  is  not  pos- 
sible at  present.  No  adequate  explanation  for  the  selective  accu- 
mulation of  lymphocytes  and  the  absence  of  polynuclear  cells  about 
tuberculous  foci  has  yet  been  advanced.  The  absence  of  polynuclear 
leukocytosis  may  possibly  be  due  to  the  great  insolubility  of  these 
bacilli,  in  consequence  of  which  little  or  no  leukocytosis-stimulating 
substances  are  liberated. 

Pearce  39  has  suggested  a  similar  reason  for  the  absence  of  poly- 
nuclear  accumulations  about  chronic  localized  lesions  of  any  kind 
in  which  tissue  encapsulation  may  prevent  the  contact  of  the  inciting 
agents  with  the  body  fluids  and  there  is  a  consequently  slow  or  slight 
production  of  such  chemotactic  stimulating  materials. 

In  typhoid  f ever,  where  the  slight  primary  leukocytosis  is  rap- 
idly succeeded  by  a  leukopenia  with  a  relative  lymphocytosis,  the 
conditions  are  somewhat  different.  Here,  as  in  some  other  infec- 
tions, as  Friedberger  and  others  have  shown,  we  are  dealing  with  a 
generalized  infection  by  an  organism  which  is  easily  subject  to  the 
action  of  alexin  with  consequent  production  of  anaphylatoxin.  (See 
chapter  on  Anaphylaxis.)  This  poison,  it  seems,  exerts  a  nega- 
tive chemotaxis,  and  probably  during  the  height  of  the  disease, 
therefore,  leads  to  the  low  leukocyte  count  observed.  That  this  is 
at  least  likely  seems  to  follow  from  the  studies  which  have  been 
made  upon  the  nature  of  the  typhoid  poisons,  and  also  from  the 
observation  of  Gay  and  Claypole,  that  typhoid-immune  rabbits  react 
to  the  infection  of  typhoid  bacilli  with  a  rapid  and  powerful  increase 
in  the  polynuclear  leukocytes,  whereas  similar  injections  into  the 
normal  animal  lead  to  leukopenia. 

If  the  supposition  regarding  tuberculosis,  made  above,  is  correct, 
it  would  follow  that  a  sudden  and  considerable  increase  in  the  poly- 
nuclear leukocytes  in  a  case  of  tuberculosis  would  indicate  a  dis- 
charge of  organisms  into  the  circulation  and  a  tendency  toward  gen- 
eralization of  the  infection  in  this  manner.  (See  Weigert's  view  of 
the  manner  in  which  tuberculosis  may  spread  by  the  destruction  of 
the  wall  of  a  vein  by  a  localized  lesion.)  However,  although  specu- 
lation in  the  absence  of  experimental  proof  is  justified,  it  must  not 
be  forgotten  that  the  problems  of  selective  chemotaxis  are  too  ob- 
scure to  permit  of  our  laying  much  weight  on  any  of  these  views. 

Gabritchewsky,40  who  investigated  this  subject  extensively,  has 
classified  various  substances  according  to  their  positively,  negatively, 
or  neutral  chemotactic  activities.  It  is  not  necessary  to  recapitulate 
these,  but  it  is  interesting  to  note  that  he  found  that  some  substances 

39  Pearce.    Jour.  A.  M.  A.,  Vol.  61,  1913. 

40  Gabritchewsky.     Ann.  Past.,  Vol.  4,  1890. 


INFECTION    AND    RESISTANCE 

which  were  positively  chemotactic  in  certain  concentrations  became 
neutral  or  even  negative  when  the  concentration  was  altered. 

We  have  seen  that  the  action  of  the  leukocytes  in  moving  toward 
some  substances  and  away  from  others  is  entirely  analogous  to  simi- 
lar phenomena  occurring  among  lower,  unicellular  forms  of  life,  and 
the  explanations  applied  to  the  apparently  conscious  acts  of  the 
ameba,  such  as  the  motion  toward  and  the  engulfment  of  food,  have 
been  applied  to  the  activities  of  the  leukocytes  as  well.  Many  of 
the  theories  developed  concerning  the  free  living  forms,  however, 
have  been  easily  excluded  in  the  case  of  the  leukocytes,  because  of  the 
environment  in  which  their  activities  are  developed.  Thus  the  many 
interesting  reactions  of  paramecia  and  other  organisms  to  light 
(heliotropism)  have  little  bearing  upon  this  subject,  and  the  views 
based  on  the  theories  of  orientation  may  be  excluded  on  the  ground 
of  the  symmetry  of  the  normal  leukocyte.  The  observations  of 
Garrey,41  that  indicate  that  it  is  the  dissociated  ions  of  various  acids 
and  bases  which  are  responsible  for  the  directive  stimuli  exerted 
upon  certain  flagellates,  may  yet  result  in  throwing  some  light  upon 
leukocytic  movements,  especially  if  we  can  come  to  accept  the  con- 
ceptions of  ion-proteins  upheld  by  Loeb  42  and  his  pupils.  However, 
the  facts  concerning  these  phenomena,  as  well  as  the  possibility, 
previously  mentioned,  of  the  opposite  electrical  charges  carried  by 
the  leukocytes  and  the  substances  attracting  them  cannot  be  regarded 
at  present  as  more  than  interesting  thoughts.  Of  more  than  merely 
speculative  interest,  however,  are  the  views  of  chemotaxis  which  are 
based  upon  the  study  of  conditions  of  surface  tension.  In  order  to 
consider  these  properly  it  will  be  useful  to  review  briefly  the  funda- 
mental principles  governing  these  conditions. 

The  molecules  of  any  fluid  are  held  together  by  mutual  attraction 
due  to  the  force  generally  spoken  of  as  cohesion.  This  force  is  ex- 
erted by  like  molecules  upon  each  other  in  solids  more  strongly  than 
in  liquids,  and  in  gases  less  strongly.  Since  we  are  dealing  in  this 
•connection  with  occurrences  taking  place  in  liquids,  we  will  restrict 
our  consideration  to  these.  The  force  of  cohesion  is  influenced  in 
a  number  of  ways.  Thus,  for  instance,  heat  reduces  it,  and  this 
is  the  cause  that  solids  are  converted  into  liquids  and  liquids  into 
gases,  provided  of  course  that  the  heat  brings  about  no  chemical 
change.  In  large  masses  of  fluids  the  force  of  gravitation  over- 
•comes  that  of  cohesion  and  larger  masses  of  liquid  assume  the  shape 
•of  the  containing  vessel.  In  smaller  masses  the  force  of  cohesion 
lends  to  bring  about  the  spherical  shape.  This  comes  about  in 
the  following  way:  Within  the  interior  of  a  drop  of  liquid  all  the 
molecules  attract  each  other,  and  since  the  force  of  attraction  is 
equal  in  all  directions  it  neutralizes  itself,  and  the  molecules  are 

41  Garrey.  Am.  Jour,  of  Phys.,  3,  1900. 

42  Loeb.  Am.  Jour,  of  Phys.,  3,  1900. 


PHAGOCYTOSIS  293 

uninfluenced  by  it,  mobile  and  free.  The  molecules  on  the  surface 
are  in  a  different  condition,  however.  They  are  subjected  to  the 
force  of  cohesion  from  within,  but  not  from  without,  and  are  there- 
fore drawn  strongly  toward  the  center.  The  result  is  the  same  as 
though  the  drop  were  subjected  to  pressure  from  without  and  the 
surface  layers  were  in  a  state  of  compression.  There  is  in  conse- 
quence a  constant  tendency  of  all  the  surface  molecules  to  be  drawn 
toward  the  center  and  a  resulting  tendency  to  a  diminution  of  the 
surface  area.  It  is  as  though  the  surface  of  such  a  drop  were  a  thin, 
elastic  membrane  which  tended  to  contract  and  diminish  in  size  and 
surface.  The  force  with  which  this  takes  place  is  spoken  of  as  sur- 
face tension,43  and  the  energy  underlying  it  is  called,  by  Ostwald, 
surface  energy.  Since  a  drop  of  one  fluid  suspended  in  another  with 
which  it  cannot  mix  is  relieved  of  the  disturbing  factor  of  gravitation, 
its  surface  tension  has  the  effect  of  contracting  the  small  mass  into  a 
form  which,  for  the  given  volume,  will  expose  the  smallest  possible 
surface,  and  this  is,  of  course,  the  sphere.  It  is  for  this  reason  that,  if 
we  shake  up  such  systems  as  water  and  chloroform,  or  oil  and  water, 
the  chloroform  or  the  oil  will  be  distributed  through  the  water  as 
small  droplets.  The  degree  of  surface  tension  of  any  fluid  is  meas- 
urable by  a  number  of  reasonably  accurate  methods  which  may  be 
found  in  any  text-book  of  physics  and  which  we  need  not  consider 
here.  It  is  of  course  dependent  in  each  case  upon  the  nature  of  the 
surrounding  medium.  We  have  taken  into  consideration  above  only 
the  force  which  is  exerted  within  the  drop  by  the  cohesion,  that  is, 
the  attraction  toward  the  center.  This  would  be  uninfluenced  from 
without  only  in  a  vacuum.  In  nature  the  surface  molecules,  though 
forcibly  drawn  toward  the  center,  are  also  affected  from  without  by 
the  attraction  exerted  by  the  molecules  of  the  substances  surrounding 
the  drop.  There  is  a  constant  balance,  therefore,  at  any  part  of  the 
surface  of  a  drop  of  fluid  between  the  cohesion  tension  from  within 
and  attractions  from  without.  The  resultant  of  the  two  forces  de- 
termines the  surface  tension,  which  will  be  greater  or  less  in  inverse 
ratio  to  the  attraction  from  without  for  any  given  drop,  and  a  varia- 
tion of  the  external  attraction  at  different  points  on  the  periphery  of 
the  drop  will  naturally  influence  the  shape  of  the  drop.  For  a  relief 
of  attraction  at  one  point  would  tend  to  permit  that  part  of  the  sur- 
face to  retract,  and  an  increase  in  this  attraction  would  tend  to 
allow  it  to  bulge,  with  the  formation  of  a  sort  of  pseudopod. 

In  studying  the  importance  of  surface  tension  44  in  determining 
the  motions  of  unicellular  organisms  a  number  of  important  attempts 
have  been  made  to  imitate  cell  motion  by  means  of  the  suspension  of 
various  substances  of  strong  cohesive  properties  in  liquid  media.  The 

43  Michaelis.     "Dynamik  der  Oberflachen,"  Steinkopf,  Dresden,  1909. 

44  For  a  thorough  discussion  of  this  phenomenon  see  also  Gideon  Wells, 
"Chemical  Pathology/'  Saunders,  1911. 


294  .INFECTION    AND    RESISTANCE 

idea  was  suggested  by  Quincke,45  and  later  by  Biitschli,46  but  has 
been  most  expensively  studied  by  Rhumbler.47  The  result  has  been 
the  production  of  a  number  of  "artificial  amebse"  which  in  almost 
all  respects  behave  like  the  living  organisms.  Thus  if  a  small  mass 
of  mercury  is  placed  into  a  dish  filled  with  water  acidified  with 
nitric  acid,  and  a  small  crystal  of  dichromate  of  potassium  is  dropped 
near  the  mercury,  the  dichromate*  will  -dissolve  and  a  yellow  cloud 
will  gradually  diffuse  from  it  toward  t*he  mercury.  As  soon  as  the 
yellow  cloud  touches  this  it  will  begin  to  show  change  of  form  and 
to  elongate  in  the  direction  of  the  dichromate,  often  moving  to- 
ward it.  The  motion  of  the  quicksilver  will  resemble  with  con- 
siderable accuracy  that  of  an  ameba  moving  toward  a  particle  of 
food  or  sending  out  pseudopodia.  A  more  striking  and  complete 
imitation  is  that  obtained  by  Rhumbler  when  he  placed  a  drop  of 
clove  oil  into  a  mixture  of  alcohol  and  glycerin.  The  changes  of  sur- 
face tension  produced  upon  the  surface  of  the  clove  oil  by  the  alcohol 
give  rise  to  movements  in  the  oil  entirely  analogous  takjthose  of  mo- 
tile cells  in  favorable  media.  The  similarity  has  been  ^extended 
even  to  the  processes  of  engulfment  of  the  food  as  observed  among 
amebse.  Thus  a  drop  of  chloroform  in  water  will  flow  about  a 
particle  of  shellac  and  dissolve  it.  If  a  piece  of  glass  coated  with 
shellac  is  placed  in  contact  with  the  drop  it  will  engulf  it, 
but  will  cast  out  the  glass  after  the  shellac  coating  has  been  dissolved 
away. 

The  similarity  between  phenomena  purely  referable  to  surface 
tension  and  those  taking  place  in  the  living  cells  is  therefore  very 
striking  and  has  been  clearly  analyzed  in  regard  to  its  bearing 
upon  leukocytic  chemotaxis  by  Gideon  Wells  in  his  "Chemical  Path- 
ology.77 The  chemotactic  substances,  diffusing  to  the  leukocyte,  will 
lower  its  surface  tension  on  the  side  at  which  they  come  in  contact. 
Pseudopodia  will  be  thrown  out  on  this  side  in  consequence,  and  the 
leukocyte  will  move  in  this  direction.  The  motion  will  be  continued 
in  this  direction  as  long  as  the  concentration  of  the  chemotactic  sub- 
stance, and  therefore  the  diminution  of  surface  tension  is  greater  on 
this  side  than  on  other  parts  of  the  periphery,  until  a  point  is  reached 
at  which  the  chemotactic  substance  is  equally  diffused  on  all  sides,  and 
motion  will  cease.  The  actual  engulfment  may  then  occur  or  the 
nature  and  concentration  of  the  chemotactic  substance  may  be  so 
great  that  injury  is  done  to  the  leukocyte.  Whether  or  not  the  purely 
physical  explanation  of  chemotaxis  tells  the  whole  story  it  is  of  course 
not  possible  to  decide.  At  any  rate,  it  furnishes  a  rational  basis  for 

45  Quincke.     Quoted  from  H.  G.  Wells,  "Chemical  Patholo£rv,"  Saunders. 
1907. 

46  Biitschli.     "Untersuch.  uber  mikroskopische   Schaume  und  das  Proto- 
plasma."     Leipzig,  1892.     See  also  H.  G.  Wells,  loc.  cit.,  pp.  220  et  seq. 

47  Rhumbler.     Arch.  f.  Entwickelungs  Mechanik,  1898. 


PHAGOCYTOSIS  295 

the  study  of  the  phenomenon  more  promising  than  any  of  the  others 
so  far  offered. 

It  is  true,  on  the  other  hand,  that  such  a  theory  in  no  way  ac- 
counts for  the  apparently  selective  positive  chemotaxis  which  is 
exerted  by  different  substances.  Thus  the  preponderance  of  poly- 
nuclear  leukocytes  in  foci  and  serous  cavities  containing  organisms 
like  staphylococci,  meningococci,  streptococci,  and  others  is  in  con- 
trast to  the  lymphocytic  accumulation  in  the  pleural,  subarachnoid', 
and  peritoneal  spaces  infected  with  tubercle  bacilli.  Some  writers 
have  spoken,  therefore,  of  active  and  passive  leukocytosis  according 
to  whether  or  not  the  cells  attracted  seemed  to  possess  ameboid 
motility.  That  surface  tension  phenomena  alone  do  not  account  for 
this  is  clear.  But  it  must  be  remembered  that  even  tubercle  bacilli, 
though  eventually  attracting  few  polynuclears  and  many  lympho- 
cytes, will  cause  an  active  polynuclear  accumulation  in  the  perito- 
neum and  pleura  when  first  injected,  and  are  actively  phagocyted. 
Later  when  the  lesion  is  established  and  the  bacilli  are  lodged  in  the 
tissues  the  polynuclears  give  way  to  the  lymphocytes,  which  even 
then  never  accumulate  in  such  proportion  as  do  the  microphages  in 
acute  suppurative  lesions.  It  may  well  be  that  the  chemotaxis  origi- 
nally attracting  the  polymorphonuclear  leukocytes  is  the  same  in 
every  case,  but  that  a  continued  irritant,  especially  one  well  sur- 
rounded by  tissue  elements  as  are  the  organisms  within  the  tubercles, 
may  cease  to  exact  any  chemotactic  influence,  the  accumulation  of  in- 
active lymphocytes  possibly  being  due  to  a  progressive  death  of  these 
cells  carried  into  the  neighborhood  of  the  lesion  by  the  normal  circu- 
lation of  the  lymph. 


CHAPTER    XII 

THE  EELATION  OF  THE  LEUKOCYTES  AKD  OF 
PHAGOCYTOSIS  TO  IMMUNITY 

IN  MetchnikofFs  earliest  work  upon  the  daphnia  or  water  flea 
he  observed  clearly  that  there  was  a  direct  relation  between  the  de- 
gree of  phagocytosis  and  the  outcome  of  the  infection.  When 
phagocytosis  of  the  invading  yeasts  was  energetic  and  complete  the 
daphnia  recovered.  When  the  yeast  cells  penetrated  the  intestinal 
wall  of  the  daphnia  in  large  numbers,  and  were  enabled  to  multiply 
before  the  phagocytic  cells  could  accumulate  in  sufficient  numbers  to 
engulf  them,  then  the  body  of  the  daphnia  was  soon  swamped  with 
the  parasites  and  death  ensued. 

This  simple  observation  fostered  the  thought  that  the  basic  prin- 
ciple underlying  all  processes  of  immunity  was  represented  in  this 
struggle  between  the  invading  bacteria  and  the  phagocytic  cells. 
To  the  activity  of  the  latter,  entirely,  he  attributed  the  power  of 
resistance. 

In  support  of  this  contention  Metchnikoff  and  his  pupils  have 
marshaled  many  facts,  most  of  which  are  set  forth  in  his  classical 
work  "L'Immunite  dans  les  Maladies  Infectieuses."  It  will  be 
manifestly  impossible  here  to  do  more  than  outline  the  plan  of 
study  which  these  investigations  have  followed  and  the  conclusions 
to  which  they  gave  just  foundation. 

The  original  study  upon  the  infectious  disease  of  daphnia  led  to 
analogous  experiments  upon  higher  animals  and,  by  the  prolonged 
and  patient  investigations  of  Metchnikoff  and  his  pupils,  it  was 
shown  that,  throughout  the  field  of  infectious  disease,  there  was  a 
striking  parallelism  between  the  resistance  of  the  infected  subject 
and  the  degree  of  phagocytosis  which  occurred. 

Earlier  studies  concern  themselves  chiefly  with  the  natural  im- 
munity possessed  by  many  animals  against  certain  infection.  The 
infectious  disease  which  at  this  time  had  been  most  thoroughly 
studied  was  anthrax,  and  Koch  had  shown  that  frogs  and  other  cold- 
blooded animals  were  markedly  resistant  against  this  micro-organism. 
Taking  advantage  of  this  observation,  Metchnikoff  studied  the  phago- 
cytosis of  anthrax  bacilli  in  frogs  and  found  that  it  took  place  rap- 
idly and  effectively,  all  of  the  injected  bacilli  being  soon  engulfed 
by  the  accumulating  cells.  Similarly,  active  phagocytosis  of  anthrax 

296 


RELATION    OF    LEUKOCYTES    TO    IMMUNITY    297 

bacilli  was  demonstrated  in  such  naturally  resistant  animals  as  dogs 
and  chickens,  while  almost  no  cell  ingestion  occurred  in  delicately 
susceptible  animals  like  guinea  pigs  and  rabbits.  Rats,  on  the  other 
hand,  more  resistant  to  anthrax  than  guinea  pigs,  less  so  than  dogs, 
showed  a  degree  of  phagocytosis  intermediate  between  that  observed 
in  the  cases  of  the  other  animals  mentioned  above.  And  yet,  in  these 
more  susceptible  animals,  the  normal  bactericidal  action  of  the  blood 
upon  anthrax  bacilli,  though  never  extreme,  was  often  more  marked 
than  that  of  the  naturally  immune  animals  mentioned  above. 

It  is  well  known,  for  instance,  that  the  serum  of  dogs  possesses 
almost  no  bactericidal  properties  for  anthrax  bacilli,1  although  the 
animals  are  highly  resistant  to  this  infection,  while  the  serum  of 
rabbits  is  probably  more  strongly  bactericidal  for  these  bacilli  than 
the  serum  of  most  other  animals,  and  yet  rabbits  are  extremely 
susceptible.  That  the  lack  of  bactericidal  powers  of  the  serum  is 
not  always  a  sign  of  susceptibility  on  the  part  of  the  animal  was 
shown  as  early  as  1889  by  Lubarsch.  (We  must  remember,  how- 
ever, that  lack  of  bactericidal  power  does  not  necessarily  mean  lack 
of  sensitizer.  For  bacteria  may  be  sensitized  without  being  killed 
extracellularly  as  can  be  shown  by  the  alexin-fixation  reaction.) 

The  study  of  anthrax  infections  was  a  peculiarly  fortunate 
choice  of  subject,  since  in  this  bacillus  resistance  to  serum  lysis  is 
especially  well  marked  and  phagocytosis  seems  indeed  to  be  the  chief 
mode  of  bacterial  destruction.  Studies  analogous  to  those  originally 
made  with  anthrax,  however,  were  subsequently  carried  on  with 
streptococci,  pneumococci,  and  staphylococci  chiefly  by  Bordet,2 
Marchand,3  and  others,  and  results  coinciding  with  those  of  Metchni- 
koff  were  obtained.  In  every  case  naturally  resistant  animals 
showed  marked  phagocytosis,  and  susceptible  ones  failed  to  show  it 
to  the  same  degree.  It  is  a  strong  support  of  the  same  opinions,  too, 
that  Marchand's  studies,  later  extensively  confirmed,  showed  that 
the  more  virulent  and  invading  strains  of  streptococci,  the  less  active 
is  the  phagocytosis — a  converse,  but  equally  conclusive,  observation. 

Further  support  for  this  point  of  view  is  manifold  and  cannot  be 
considered  with  anything  like  completeness.  We  may  refer  briefly, 
however,  to  the  experiments  of  Vaillard,  Vincent,  and  Rouget 4  with 
tetanus,  and  those  of  Leclainche  and  Vallee 5  with  symptomatic 
anthrax,  because  they  are  especially  valuable  in  illustrating  the 
importance  of  phagocytosis  in  another  class  of  infection.  The  pois- 
ons of  these  micro-organisms  are  extremely  toxic  for  rabbits,  and  if 

1  Petterson.     CentraM.  f.  Bakt.,  1,  39,  1905. 

2  Bordet.    Ann.  de  I'Inst.  Past.,  Vol.  11,  1897. 

3  Marchand.     Archiv.  de  med.  Exp.,  Vol.  10,  1898. 

4  Vaillard,  Vincent,  and  Rouget,     Ann.  de  I'Inst.  Past.,  Vols.  5,  6,  1891- 
1892. 

5  Leclainche  and  Vallee.    Ann.  de  I'Inst.  Past.,  Vol.  14,  1900. 


298  INFECTION    AND    RESISTANCE 

a  small  amount  of  culture  material,  together  with  agar,  broth,  or 
any  foreign  substance  which  may  inhibit  or  divert  phagocytosis  from 
the  spores,  is  injected  into  these  animals  rapid  proliferation  and 
death  with  toxemia  result.  If,  on  the  other  hand,  the  spores  are 
carefully  washed  of  foreign  material  and  toxin  rapid  phagocytosis 
results  and  the  animals  recover. 

The  parallelism  which  was  followed  out  so  extensively  between 
natural  immunity  and  phagocytosis  was  even  more  closely  marked 
in  the  case  of  artificially  acquired  immunity.  The  first  observations 
of  this  kind  made  by  Metchnikoff,  again  on  the  subject  of  anthrax 
infection,  were  carried  out  by  the  active  immunization  of  rabbits. 
The  subcutaneous  injection  of  virulent  anthrax  bacilli  into  normal 
rabbits  is  usually  followed  by  a  rapid  growth  of  the  bacteria,  with 
much  serous  exudation  but  hardly  any  leukocytic  accumulation.  In 
immunized  animals,  on  the  other  hand,  the  bacilli  are  taken  up  by 
hosts  of  phagocytes,  just  as  this  occurs  in  naturally  resistant  dogs  or 
other  animals.  Similarly  Bordet  6  has  shown  that  cholera  spirilla 
injected  into  the  blood  stream  of  cholera-immune  animals  are  taken 
up  by  leukocytes  even  before  they  can  be  subjected  to  lysis  by  the 
circulating  lytic  antibodies. 

It  would  add  little  to  clearness  were  we  to  multiply  the  examples 
in  which  it  has  been  demonstrated  that  the  acquisition  of  increased 
resistance  is  accompanied  by  enhancement  of  the  phagocytic  process. 
This  statement  may  be  regarded  as  an  axiom,  and  indeed  our  later 
discussions  of  the  opsonins  and  bacteriotropins  will  show  clearly  why 
such  a  state  of  affairs  is  to  be  expected.  Taken  by  itself,  however, 
it  does  not  necessarily  prove  that  the  destruction  of  the  invading 
germs  is  entirely  due  to  the  leukocytes.  It  might  still  be  possible 
that  the  bacteria  are  injured  or  even  killed  by  the  antibacterial 
serum  constituents  before  they  can  be  taken  up  and  carried  away 
by  the  cellular  elements;  the  phagocytes  then  would  act  only  as 
scavengers  for  the  removal  of  the  dead  bodies.  Indeed,  this  opinion 
was  long  held  by  a  number  of  the  adherents  of  the  purely  humoral 
school.  However,  such  a  point  of  view  is  no  Tonger  tenable — espe- 
cially in  the  light  of  the  later  opsonin  studies  just  referred  to. 
Moreover,  long  before  these  more  recent .  studies  it  was  clear  that 
bacteria  may  often  grow  within  the  leukocytes — finally  destroying 
these — and  that  they  may  even  remain  fully  virulent  after  ingestion. 
For,  as  Metchnikoff  showed,  if  guinea  pigs  were  injected  with  a 
little  of  the  exudate  formed  after  the  injection  of  anthrax  bacilli 
into  immunized  rabbits  (an  exudate  in  which  there  were  no  longer 
any  extracellular  bacteria  because  of  energetic  phagocytosis)  death 
often  resulted.  It  was  clear,  therefore,  not  only  that  the  ingested 
bacteria  were  still  alive,  but  that  they  were,  at  least  in  part,  still 
fully  virulent. 

6  Bordet.    Ann.  de  I'Inst.  Past.,  Vol.  9,  1895. 


RELATION    OF    LEUKOCYTES    TO    IMMUNITY  299 

A  further  method  of  investigation  employed  by  Metchnikoff  in 
Ms  endeavors  to  prove  his  point  consisted  in  the  attempt  to  demon- 
strate that  virulent  bacteria  could  be  protected  from  destruction  in 
the  bodies  of  resistant  animals  if  the  leukocytes  could  be  held  at  bay. 
This  resulted  in  a  number  of  ingenious  experiments,  the  most  con- 
vincing of  which  is  the  one  carried  out  with  anthrax  bacilli  and 
frogs  by  Trapetznikoff.7  Anthrax  spores  were  inclosed  in  little  sacks 
of  filter  paper  and  these  were  introduced  subcutaneously  into  frogs. 
In  consequence  the  spores,  bathed  in  the  tissue  fluids,  but  protected 
from  phagocytosis,  developed  into  the  vegetative  forms,  multiplied, 
and  remained  virulent  for  several  days.  Taken  up  by  the  frog's 
phagocytes  under  ordinary  conditions,  they  would  rapidly  have  been 
taken  up,  digested,  and  destroyed.  Here  again  it  was  shown  that  the 
body  of  fluids  alone  were  unable  to  dispose  of  the  bacteria  and  that 
the  natural  resistance  of  the  frog  was  due  entirely  to  phagocytic 
processes. 

Other  experiments  have  been  aimed  at  a  general  reduction  of 
phagocytic  activity  by  the  use  of  narcotics.  Thus,  Cantacuzene  8 
showed  that  animals  treated  with  opium  are  very  much  more  sus- 
ceptible to  infection  than  are  normal  controls.  And  since  opium 
markedly  inhibits  the  activity  of  the  white  cells  it  may  possibly  be 
that  these  experiments  furnish  a  further  support  for  Metchnikoff  s 
opinion.  At  any  rate,  it  is  worth  noting  that,  even  though  these 
experiments  are  not  convincing  in  their  assertion  that  the  increased 
susceptibility  was  due  entirely  to  the  interference  with  the  leuko- 
cytes, they  indicate  very  definitely  the  inadvisability  of  using  mor- 
phin  and  similar  narcotics  in  infectious  diseases. 

It  is  quite  clear  at  any  rate,  then,  that  the  process  of  phagocy- 
tosis increases  in  energy  as  immunity  is  acquired  and,  so  far,  Metch- 
nikoff' s  assertions  are  entirely  upheld  by  later  knowledge.  In  his 
contention  that  all  properties  upon  which  the  resistance  of  the  ani- 
mal against  infection  depends  center  directly  or  indirectly  in  the 
phagocyte,  however,  many  subsequent  amendments  have  been  neces- 
sary, which  will  become  self-evident  in  the  following  discussions  of 
individual  phases  of  the  destruction  of  invading  bacteria. 

We  cannot  at  the  present  time  attempt  to  correlate  these  extreme 
views  of  Metchnikoff  with  the  equally  extreme  opinions  of  those  in- 
vestigators who  formerly  attributed  immunity  entirely  tojjthe  prop- 
erties of  the  body  fluids,  assigning  to  the  cellular  activities  a  merely 
secondary  role.  Many  of  the  apparently  opposed  contentions  have 
become  reconciled,  and  we  now  realize  that  neither  process  alone  tells 
the  whole  story,  both  being  parts  of  the  complicated  correlated  proc- 
esses which  together  constitute  the  mechanism  of  resistance.  It  was 

7  Trapetznikoff.     Ann.  de  I'Inst.  Past.,  1891. 

8  Cantacuzene.    Ann.  de  I'Inst.  Past.,  Vol.  12,  1898. 


300  INFECTION    AND    RESISTANCE 

indeed  to  the  eager  controversy  between  the  two  schools  that  we  owe 
much  of  the  clearness  of  conception  which  recent  years  have  given. 

After  the  bacteria  are  taken  up  by  phagocytes  they  undergo  a 
gradual  disintegration  or  dissolution  comparable  to  that  by  which 
a  particle  of  food  is  digested  within  the  cell  body  of  a  rhizopod. 
With  the  exception  of  such  particularly  insoluble  micro-organisms  as 
the  tubercle  bacillus,  the  leprosy  bacillus,  blastomyces,  and  a  few 
others,  there  is  in  all  cases  an  eventual  complete  resolution  of  the 
bacterial  body.  As  in  amebse  the  digestion  takes  place  often  after 
the  formation  of  digestive  vacuoles,  and  by  staining  at  this  time  with 
neutral  red  it  may  be  demonstrated  that  the  process  goes  on  in  a 
weakly  acid  environment. 

Metchnikoff  naturally  assumed,  therefore,  that  the  intracellular 
digestion  of  bacteria  by  microphages  (polynuclear  leukocytes),  or  of 
cellular  elements,  etc.,  by  macrophages,  was  a  process  carried  on 
most  probably  by  enzymes,  and  that  these  enzymes  were  identical 
with  the  bactericidal  bodies  described  as  "alexin"  and  "sensitizer" 
or  "amboceptor"  in  the  blood  stream.  To  follow  without  confusion 
the  development  of  his  ideas,  however,  it  is  necessary  to  bear  in  mind 
that  much  of  his  earlier  work  was  done  at  a  time  when  the  discovery 
of  the  cooperation  of  two  substances  in  bacteriolysis  and  hemolysis 
(the  amboceptor  and  the  complement)  had  not  yet  been  made  by 
Bordet,  and  when  the  bactericidal  effect  of  normal  serum  was  at- 
tributed entirely  to  a  single  substance — the  alexin  of  Buchner. 

Buchner  9  himself  had  suggested  that  alexin  was  an  enzyme-like 
body  probably  derived  from  the  leukocytes. 

In  his  experiments  Buchner  had  noticed  that  exudates,  produced 
by  intrapleural  injections  of  aleuronat  in  rabbits  and  dogs,  possessed 
a  bactericidal  value  for  Bacillus  coli  which  exceeded  the  bactericidal 
power  of  the  blood  serum  itself.  The  influence  of  active  phagocytosis 
could  be  excluded  by  the  fact  that  the  leukocytes  of  the  exudate  had 
been  killed  by  repeated  freezing  and  thawing.  Similar  results  were 
obtained  by  Hahn  10  with  B.  typhosus. 

Denys  and  Kaisin,11  working  along  similar  lines,  found  that  the 
pleural  exudates  of  rabbits,  obtained  by  the  injection  of  dead  staphy- 
lococci  and  freed  of  cells  by  centrifugalization,  were  more  highly 
bactericidal  for  staphylococci  than  the  blood  serum  of  the  same  ani- 
mals, but  found  also  that  the  inactivated  exudate  could  not  be  reacti- 
vated by  the  addition  of  leukocytes.  Denys  offered  as  an  explanation 
for  these  phenomena  that  the  living  leukocytes  in  the  original  exudate 
secreted  alexin  or  complement  which  enhanced  the  bactericidal 
activity  of  the  exudate,  that  the  leukocytes,  subsequently  added  -to 

9  Buchner.    Munch,  med.  Woch.,  No.  24,  1894. 
10  Hahn.    Archiv  f.  Hyg.,  Vol.  25,  1895. 

"Denys  and  Kaisin,  Denys  and  Havet.  La  Cellule,  Vol.  9,  1893;  VoL 
10,  1894. 


RELATION    OF    LEUKOCYTES    TO    IMMUNITY    301 

inactivated  exudate,  however,  had  lost  vitality  during  the  process  of 
isolation  and  washing,  and  no  longer  possessed  secretory  power. 

Hankin,12  Kanthack  and  Hardy  13  had  gone  even  farther  than 
this,  and  had  attributed  the  production  of  alexin  to  the  eosinophile 
leukocytes  particularly. 

Metchnikoff,14  basing  his  opinion  on  his  own  studies,  those  of 
his  pupils,  and  many  other  investigations  similar  to  those  mentioned 
above,  came  to  the  conclusion  that,  under  ordinary  conditions,  the 
normal  blood  contains  no  free  bactericidal  substances.  He  assumes 
that  these  substances  are  entirely  intracellular,  being  constituents  of 
the  various  phagocytic  elements,  by  means  of  which  the  cells  digest 
the  foreign  elements  they  take  up.  He  believes  that  there  are  essen- 
tially two  varieties  of  such  digestive  enzymes  or  "cytases" — just  as 
there  are  two  varieties  of  phagocytes.  The  microphages,  chiefly  con- 
cerned in  the  digestion  of  bacteria,  secrete  the  bactericidal  alexin,  or, 
as  Metchnikoff  calls  it,  "microcytase."  The  macrophages,  the  large 
mononuclear  lymph  and  endothelial  cells,  primarily  concerned  in  the 
phagocytosis  of  cellular  elements  (red  cells,  etc.),  contain  another 
variety  of  digestive  enzyme,  the  "macrocytase,"  or  cytolytic  (hemo- 
lytic)  alexin.  The  supposition  that  the  hemolytic  "cytase"  is  largely 
derived  from  the  macrophages  was  based  particularly  upon  the  in- 
vestigations of  MetchnikofFs  pupil,  Tarassewitch,15  who  found  that 
the  extracts  obtained  from  lymph  nodes,  and  other  organs  rich  in 
macrophages,  possessed  hemolytic  properties.  Both  this  work  and 
the  preceding  studies  regarding  the  extraction  of  alexin  from  poly- 
nuclear  leukocytes  will  be  more  fully  discussed  below. 

Maintaining  that  these  cytases  are  purely  intracellular  under 
ordinary  conditions,  Metchnikoif  believes  that,  in  normal  animals, 
the  destruction  of  invading  bacteria  or  of  injected  cellular  substances 
(blood  cells,  etc.)  is  accomplished  entirely  by  the  phagocytic  process, 
with  subsequent  intracellular  digestion.  In  immunized  animals, 
however,  there  is  present  in  the  circulating  blood  another  substance, 
not  identical  with  the  cytases,  but  also  derived  from  the  leukocytes  or 
from  the  blood-forming  organs — the  afixateur"  (Ehrlich's  "ambo- 
ceptor" — Borders  "sensitizer").  This  specific  "fixateur"  sensitizes 
the  bacteria  or  other  antigens  to  the  action  of  the  cytases.  For  his 
assumption  regarding  the  origin  of  this  sensitizer  he  finds  support 
largely  in  the  researches  of  Pfeiffer  and  Marx,  and  others  mentioned 
in  our  section  on  the  origin  of  antibodies,  as  well  as  in  the  simi- 
lar investigations  of  Deutsch,16  carried  on  under  MetchnikofFs  per- 
sonal supervision. 

12Hankin.     Centralbl  f.  Bakt.,  Vol.  12,  1892. 

13  Kanthack  and  Hardy.     Proc.  Eoy.  Soc.,  Vol.  52,  1892. 

14  Metchnikoff.     Ann.  de  VInst.  Pasteur,  Vol.  7,  1893;  Vol.  8,  1894;  Vol. 
9,  1895. 

15  Tarassewitch.    Ann.  de  VInst.  Past.,  Vol.  16,  1902. 

16  Deutsch.     Ann.  de  VInst.  Pasteur,  Vol.  13,  1899. 


302  INFECTION    AND    RESISTANCE 

Final  digestion  of  the  sensitized  antigens  (bacteria  or  blood 
cells),  however,  can  take  place  only  under  the  influence  of  the 
cytases  intracellularly,  unless  by  previous  leukocytic  injury  these 
enzymes  have  been  liberated  into  the  blood  stream. 

While  it  is  admitted,  then,  that  bacteria  may  be  killed  and  di- 
gested both  intra-  and  extracellularly  in  the  animal  body,  the  cytases, 
which  accomplish  this,  are  regarded  as  the  same  in  both  cases,  being 
derived  from  the  phagocytic  cells.  In  immunized  animals  "fixateur" 
may  be  produced  under  the  stress  of  active  immunization  and  fur- 
nished to  the  blood  stream  by  the  blood-forming  organs.  By  this 
substance  bacteria  and  cells  may  be  sensitized.  However,  the  enzyme 
by  which  digestion  is  actually  accomplished,  "cytase"  or  alexin,  is 
not  present  free  in  the  blood  even  in  immune  animals  unless  it  has 
become  free  and  extracellular  by  injury  to  the  leukocytes. 

How,  then,  on  this  basis  does  Metchnikoff  account  for  the  Pfeiffer 
phenomenon,  in  which  the  extracellular  destruction  of  bacteria  takes 
place  rapidly  in  the  peritoneal  exudate  ?  His  explanation  is  the  fol- 
lowing: When  bacteria  or  other  substances  are  injected  into  the 
peritoneum  there  is  a  preliminary  injury  of  leukocytes  (phagolysis), 
and  by  this  alexin  or  cytase  is  liberated.  When  cholera  spirilla,  for 
instance,  are  injected  into  the  peritoneal  cavity  of  an  immunized 
guinea  pig  there  follows  a  short  period  during  which  few  if  any 
leukocytes  are  present  in  the  exudate,  but  many  may  be  found  gath- 
ered in  motionless  clumps  in  the  folds  of  the  peritoneum  and  mesen- 
tery, incapable  of  phagocytosis  and  apparently  injured.  If  such 
leukocytic  injury  can  be  avoided  Metchnikoff  claims  that  the  extra- 
cellular lysis  of  cholera  spirilla  will  fail  to  take  place.  Thus  if 
sterile  broth  or  salt  solution  is  injected  into  the  peritoneum  of  a 
guinea  pig  a  preliminary  phagolysis  will  be  followed  by  an  accumu- 
lation of  leukocytes.  If  cholera  spirilla  are  now  introduced  no 
extracellular  digestion  is  seen,  but,  instead  of  this,  rapid  phagocytosis 
takes  place.  This  he  says  is  due  to  the  fact  that  the  freshly  accumu- 
lated, healthy  phagocytes,  collected  in  response  to  the  preliminary 
broth  injection,  are  not  easily  injured  and  do  not  undergo  phago- 
lysis; no  cytase  is  liberated  and,  in  consequence,  no  serum  bacterio- 
lysis can  take  place.  In  the  same  way  he  claims  that  the  hemolysis 
of  red  blood  cells  (goose  blood)  in  the  peritoneum  of  specifically 
immunized  guinea  pigs  may  be  prevented  if,  by  a  previous  injection 
of  broth,  healthy  leukocytes  have  been  caused  to  accumulate.  In 
such  a  case  the  goose  blood  corpuscles  are  rapidly  ingested  by  the 
phagocytes  and  no  hemolysis  occurs. 

It  is  self-evident  that  the  validity  of  this  interpretation  of  the 
occurrences  is  strictly  dependent  upon  the  demonstration  that  the 
circulating  blood  normally  contains  no  alexin  or  complement.  This 
is  rigidly  maintained  by  MetchnikofF,  and  is  indeed  one  of  the  most 
important  uncertainties  of  serology.  He  asserts  that  alexin  appears 


RELATION    OF    LEUKOCYTES    TO    IMMUNITY    303 

in  the  blood  serum  only  because  changes  in  the  leukocytes  take  place 
during  coagulation.  It  is  not,  by  any  means,  settled  that  Metchni- 
koff  is  right  in  this — in  fact,  more  recent  investigations  seem  to  show 
that  he  is  wrong,  and  that  we  may  assume  definitely  that  alexin  is 
present  in  considerable  amounts  in  the  circulating  blood  plasma  of 
normal  animals. 

Metchnikoff  s  denial  of  this  is  based  chiefly  on  the  experiments 
of  Gengou.  Gengou  17  took  the  blood  from  various  animals  into 
paraffined  tubes  and  centrifugalized  it  at  low  temperature  before  it 
could  clot.  This  freed  the  plasma  from  the  cells  before  clotting, 
though  coagulation  of  course  took  place  as  soon  as  this  plasma  was 
removed  to  tubes  and  kept  at  room  temperature.  The  serum  ex- 
pressed from  this  clotted  plasma  he  compared  for  alexin  contents 
(bactericidal  properties)  with  that  obtained  from  clotted  whole 
blood. 

He  found  that,  in  all  cases  examined  (dogs,  rabbits,  and  rats), 
the  plasma  contained  practically  no  bactericidal  substances,  or  at  any 
rate  an  incomparably  smaller  amount  than  was  present  in  the  serum 
obtained  from  the  clotted  blood. 

These  experiments  of  Gengou  would  be  conclusive  in  establishing 
MetchnikofFs  theory  if  they  were  confirmed  by  other  observers. 
This,  however,  has  not  been  the  case.  Petterson  18  found  no  differ- 
ence between  the  bactericidal  properties  of  serum  and  oxalate  plasma, 
and  Lambotte  19  arrived  at  similar  results  when  he  compared  serum 
with  the  coagulable  plasma  obtained  by  tying  off  a  section  of  a  vein 
and  centrifugalizing  the  blood  without  opening  the  vessel.  Hew- 
lett,20 Falloise,21  Schneider,22  and  more  recently  Dick  23  and  Addis,24 
whose  work  has  been  done  with  particular  attention  to  technical 
accuracy,  fail  to  confirm  Gengou,  finding  no  appreciable  difference 
between  plasma  and  serum. 

In  favor  of  Gengou's  results  are  the  investigations  of  Herman  25 
and  the  more  recent  ones  of  Gurd.26  Further  supporting  Gengou's 
conclusion  is  the  observation  recorded  by  a  number  of  workers  (Wal- 
ker,27 Longcope,28  and  others)  that  the  complement  or  alexin  con- 
tents of  serum  will  increase  somewhat  as  the  serum  is  allowed  to 

17  Gengou.     Ann.  de  I'Inst.  Past.,  Vol.  15,  1901. 

18  Petterson.     Arch.  f.  Hyg.,  Vol.  43,  1902. 

19  Lambotte.     Centralbl  f.  Bdkt.,  Vol.  34,  1903. 

20  Hewlett.     Zeitschr.  f.  Heilkunde,  24,  1903. 

21  Falloise.    Bull,  de  I'Acad.  Eoy.  de  Med,,  1905. 

22  Schneider.     Arch.  f.  Hyg.,  65,  1908. 

23  Dick.     Jour.  Inf.  Dis.,  Vol.  12,  1913. 

24  Addis.    Jour.  Inf.  Dis.,  Vol.  10,  1912. 

25  Herman.    Bull,  de  I'Acad.  Roy.  de  Med.,  1904. 

26  Gurd.     Jour.  Inf.  Dis.,  Vol.  11,  1912. 

27  Walker.     Jour,  of  Hyg.,  3,  1903. 

28  Longcope.    Med.  Bull.  Univ.  of  Pa.,  1902,  Vol.  15,  p.  331, 


INFECTION    AND    RESISTANCE 

stand  on  the  clot.  This  observation,  too,  has  been  rendered  incon- 
clusive by  contrary  reports  from  other  investigators.  Longcope,29 
further,  has  found  that  alexin  was  more  plentiful  in  the  blood  of 
individuals  suffering  from  leukemia — in  which  of  course  a  larger 
percentage  of  leukocytes  is  present  in  the  circulation.  This,  too, 
has  been  contradicted  by  other  workers,  but  even  if  upheld  would  not 
influence  the  possibility  of  there  being  alexin  in  the  normal  circula- 
tion. On  the  whole  Gengou's  contentions  with  their  consequent 
bearing  upon  Metchnikoff's  theory  cannot  be  accepted  as  final.  In 
fact,  the  greater  part  of  available  experimental  evidence  seems  to 
point  to  the  actual  presence  of  alexin  in  the  normal  circulating  blood. 
This  seems  also  indicated  by  the  unquestionable  fact  that  active 
phagocytosis  may  take  place  in  the  circulation  of  an  animal  and,  as 
we  shall  see  below,  free  alexin  is  probably  necessary  (as  opsonin)  in 
this  process.  Further  evidence  in  this  direction  also  is  furnished  by 
the  immediate  anaphylactic  shock  which  follows  the  injection  of 
antigen  into  the  blood  stream  of  a  sensitized  animal,  a  process  in 
which  we  have  much  reason  to  believe  that  alexin  takes  an  active 
part.  However,  the  problem  is  a  difficult  one,  and,  while  we 
favor  the  opinion  that  free  alexin  is  present  in  the  intravascular 
blood,  we  must  admit  that  a  crucial  experiment  has  not  yet  been 
formulated. 

Now,  as  regards  the  apparent  extraction  of  alexin  from  leuko- 
cytes and  lymphatic  organs,  we  have  already  called  attention  to  the 
fact  that  most  of  the  researches  associating  these  cells  with  the  bac- 
tericidal substances  were  carried  out  before  the  dual  mechanism  of 
sensitizer  and  alexin  in  bacteriolysis  had  been  fully  worked  out.  In 
consequence  conclusions  were  formulated  from  the  mere  facts  of  the 
presence  of  bactericidal  or  hemolytic  properties  in  cell-extracts  with- 
out the  further  determination  of  heat  stability  or  the  possibility  of 
reactivation.  Most  of  the  earlier  work  also  was  done  without  suffi- 
cient attention  to  the  separation  of  the  cells  and  the  serum  of  leuko- 
cytic  exudates.  The  first  one  to  do  this  carefully  was  Hahn,30  who, 
like  his  predecessors,  concluded  that  the  bactericidal  leukocytic  sub- 
stances, undoubtedly  encountered  by  him,  were  identical  with  alexin. 
Doubt  was  first  cast  upon  this  by  Schattenfroh,31  who  worked  with 
leukocytes  suspended  and  extracted  in  physiological  salt  solution. 
He  found  that  bactericidal  substances  were,  indeed,  obtained  in  this 
way,  but  that,  unlike  alexin,  these  substances  were  thermostable, 
withstanding  exposure  to  a  temperature  of  56°  C.  and  destroyed  only 
by  exposure  to  temperatures  as  high  as  75°  C.  to  80°  C.  for  thirty 
minutes. 

29  Longcope.    Jour,  of  Hyg.,  Vol.  3. 

30  Hahn.     Arch.  f.  Hyg.,  Vol.  25. 

31  Schattenfroh.     Arch.  f.  Hyg.,  Vols.  31  and  35;  1897. 


RELATION    OF    LEUKOCYTES    TO    IMMUNITY    305 

Moxter,32  a  little  later,  working  with  cholera  spirilla,  also  came 
to  the  conclusion  that  the  leukocytic  bactericidal  substances  were  not 
identical  with  those  found  in  the  blood  serum.  Petterson,33  too,  made 
thorough  investigations  into  the  nature  of  the  bactericidal  substances 
extracted  from  the  leukocytes,  and,  working  chiefly  with  B.  proteus 
and  B.  anthracis,  found  such  substances  in  the  leukocytes  of  dogs, 
rabbits,  and  guinea  pigs  active  against  the  bacteria  mentioned  above, 
but  failed  to  find  them  active,  at  least  in  guinea  pig  and  cat  leuko- 
cytes, against  B.  typhosus  or  the  cholera  spirillum.  He  expresses  the 
opinion  that  bactericidal  leukocytic  substances  are  normally  given 
up  to  the  blood  in  minute  quantity  only  or  not  at  all,  and  that  these 
substances  hold  no  definite  relationship  to  the  bactericidal  sub- 
stances found  in  blood  serum.  In  a  later  investigation  he  showed 
that  the  "endolysins,"  as  he  now  calls  the  leukocytic  bactericidal 
substances,  may,  like  many  enzymes  and  serum  bacteriolysins,  be 
precipitated  out  of  solution  with  alcohol  and  ether ;  but  he  separates 
them  absolutely  from  serum  lysins  and  complement.  The  latter, 
while  they  may  be,  in  part  at  least,  secreted  by  the  leukocytes,  are, 
according  to  Petterson,  induced  easily  to  come  out  of  the  cells  during 
life  by  slight  injury  or  other  stimulation,  while  the  endolysins  them- 
selves are  abstracted  from  the  cells  only  after  extensive  extraction  or 
maceration. 

Schneider  34  has  come  to  similar  conclusions  and  speaks  of  the 
endocellular  bactericidal  substances  as  "Leukine."  In  a  recent  in- 
vestigation of  the  same  subject  the  writer  35  has  in  all  essentials  con- 
firmed Schattenfroh's  original  conclusions  regarding  the  heat  sta- 
bility of  the  extracted  leukocytic  bactericidal  substances,  and  has 
shown  that  after  inactivation  by  heat  these  substances  are  not  reacti- 
vable  by  the  addition  of  fresh  leukocytic  extracts,  and  that  the  yield 
obtained  from  the  leukocytes  of  immunized  animals  is  not  greater 
than  that  obtained  from  normal  leukocytes. 

It  appears,  therefore,  that,  contrary  to  MetchnikofFs  first  sup- 
position, the  enzymes  which  bring  about  the  digestion  of  phagocyted 
bacteria  within  the  cell  are  not  identical  with  those  which  bring 
about  a  similar  extracellular  digestion.  In  addition  to  the  demon- 
stration of  a  definitely  different  structure  possessed  by  the  bacteri- 
cidal leukocytic  extracts,  as  evidenced  by  their  heat  stability,  we  have 
the  negative  evidence  that  neither  true  alexin  nor  sensitizers  have 
ever  been  successfully  extracted  from  such  cells. 

It  is  still  possible  that  this  may  eventually  be  done — but,  al- 
though indirect  evidence  like  that  of  Denys,  Longcope's  observations 
in  leukemia,  and  the  occasional  increase  of  the  alexic,  powers  of  serum 

32Moxter.     Deutsche  med.  WocJi.,  1899,  p.  687. 

33  Petterson.     Centralbl  f.  Bakt.,  i,  39,  1905;  46,  1908. 

34  Schneider.    Archiv  f.  Hyg.,  Vol.  70,  1909. 

35  Zinsser.     Jour.  Med.  Res.,  Vol.  22,  1910. 


306  INFECTION    AND    RESISTANCE 

after  standing  on  the  clot  points  to  a  probability  of  this,  no  direct 
evidence  has  so  far  been  satisfactorily  produced.  In  the  hope  that 
the  leukocytes  would  give  up  alexin — possibly  as  a  secretion  as  sug- 
gested by  Denys — the  writer,  with  Cary,  some  years  ago  kept 
washed  leukocytes  at  37.5°  C.  in  Locke's  solution,  but  was  unable 
to  find  any  evidence  of  alexin  production  within  48  hours. 

The  apparent  extraction  of  hemolysin  from  macrophages  by 
Tarassewith,  moreover,  has  met  with  a  similar  refutation.  Korschun 
and  Morgenroth  36  have  shown  that  these  hemolytic  extracts  are  ex- 
tremely heat  resistant,  are  alcohol  and  ether  soluble,  and  do  not  act 
as  antigens.  They  are  quite  different  from  the  serum  hemolysins, 
therefore,  and  probably  closely  related  to  the  hemolytic  lipoidal 
substances  described  by  Noguchi  and  others. 

Further  strong  arguments  against  the  assumption  of  the  pres- 
ence of  hemolytic  alexin  in  the  body  of  the  macrophages  have  been 
advanced  by  Gruber  37  and  by  Neufeld.38  Gruber  showed  that  no 
extracellular  hemolysis  takes  place  when  leukocytes  are  brought 
together  with  sensitized  red  blood  cells,  and  Neufeld  showed  that 
even  after  the  phagocytosis  of  such  sensitized  cells  the  hemolysis  is 
very  much  slower,  and  of  a  different  character  from  extracellular 
serum  hemolysis.  In  the  intracellular  digestion  there  are  no  mere 
solution  of  the  hemoglobin  and  formation  of  a  shadow  form 
(stroma),  but  there  occur  a  gradual  degeneration  with  the  forma- 
tion of  a  granular  detritus  of  hemoglobin. 

It  is  probable,  then,  that  the  digestion  of  bacteria  and  cells 
within  the  phagocytes  is  carried  out  by  substances  not  identical  with 
those  taking  part  in  serum  lysis.  It  is  not  unlikely  that  the  intra- 
cellular process  is  a  quite  complicated  one,  not  depending  on  a  single 
enzyme. 

In  addition  to  the  bactericidal  substances  extracted  from  leuko- 
cytes a  number  of  true  enzymes  have  indeed  been  obtained  by  various 
investigators.  We  have  mentioned  in  another  place  that  one  of  the 
earliest  observations  in  this  respect  was  that  of  Leber,39  who  noticed 
that  sterile  pus  could  liquefy  gelatin.  It  may  be  commonly  observed 
in  paraffin  or  celloidin  sections  of  staphylococcus  abscesses  that  a 
ring  of  apparently  digested  or  degenerating  tissue  is  formed  about 
an  accumulation  of  leukocytes — in  foci  in  which  the  bacteria  may 
be  too  sparse  to  be  held  accountable  for  the  changes.  These  leuko- 
proteases  have  subsequently  been  carefully  studied  by  Opie,40  Joch- 
mann  and  Miiller,41  and  a  number  of  others. 

36  Korschun  and  Morgenroth.    Berl  klin.  Woch.,  1902. 

37  Gruber.    Quoted  from  Sachs,  in  "Kraus  u.  Levadiii  Handbuch,"  Vol.  2, 
p.  991. 

38  Neufeld. .  Arb.  a.  d.  kais.  Gesundh.  Ami.,  Vol.  28,  1908. 

39  Leber.     "Die  Entstehung  der  Entziindung,"  Leipzig,  1891. 

40  Opie.     Jour.  Exp.  Med.,  Vol.  7,  1905;  Vol.  8,  1906;  Vol.  9,  1907. 

41  Midler  and  Jochmann.    Munch,  med.  Woch.,  Nos.  29  and  3~L,  1906. 


RELATION    OF    LEUKOCYTES    TO    IMMUNITY    307 

Opie  found  that  two  distinct  proteolytic  enzymes  could  be  ex- 
tracted from  the  cells  of  exudates  obtained  by  turpentine  injections. 
One — peculiar  to  the  polynuclear  leukocyte,  and  similar  to  one  pre- 
viously described  by  Miiller  42 — acts  in  a  weakly  alkaline  medium. 
The  other,  present  particularly  in  exudates  containing  a  predominat- 
ing number  of  mononuclear  cells,  acts  in  a  weakly  acid  reaction. 
Tschernorutski  also  found  proteolytic  ferments  in  both  micro-  and 
macrophages,  but  found  no  lipase  in  the  polynuclear  extracts.  This 
seems  particularly  interesting  in  view  of  the  great  resistance  to 
intracellular  digestion  noticed  in  acid-fast  bacteria,  a  point  of  some 
importance  in  connection  with  the  destruction  in  the  body  of  such 
micro-organisms  as  the  bacilli  of  tuberculosis  and  leprosy.43  Joch- 
mann  44  states  that  the  action  of  the  leukoprotease,  which  acts  in  an 
alkaline  medium  upon  casein,  results  in  the  formation  of  tryptophan 
and  ammonia,  and  believes  it  to  be  functionally  very  similar  to 
trypsin.  It  is  interesting  to  note  that  the  most  active  protease  is 
obtained  from  pus  as  it  forms  about  acute  infections  or  other  stimuli 
which  lead  to  the  accumulation  of  polynuclear  leukocytes,  whereas 
it  is  apparently  completely  absent  from  tuberculous  pus. 

The  question  immediately  arises,  are  these  leukoproteases  identi- 
cal with  the  bactericidal  substances  extracted  from  leukocytes  as  de- 
scribed above  ?  For  it  might  well  be  that  bacterial  death  resulted 
merely  from  the  digestive  action  of  the  enzyme  upon  the  bacterial 
protein.  Jochmann,45  who  has  approached  this  problem  experi- 
mentally, has  answered  it  in  the  negative.  By  repeated  alcohol 
precipitation  of  glycerin  extracts  of  leukocytes  he  obtained  an  en- 
zyme preparation  which  possessed  absolutely  no  bactericidal  prop- 
erties, though  it  was  still  actively  proteolytic. 

Not  only  did  this  relatively  pure  ferment  possess  no  bactericidal 
action,  but  bacteria  actively  proliferated  when  suspended  in  it.  Joch- 
mann  believes  that  living  bacteria  are  not  amenable  to  the  enzyme 
possibly  because  of  their  possession  of  an  "antiferment,"  at  least 
this  would  follow  in  some  cases  from  the  experiments  of  Kantoro- 
wicz.46 

The  leukoproteases,  therefore,  appear  to  possess  no  direct  sig- 
nificance in  bacterial  immunity.  Their  function  seems  rather  to 
lie  in  the  resorption  of  dead  tissues,  fibrin,  blood  clots,  etc.  Friedrich 
Miiller  47  has  pointed  out  their  possible  importance  in  the  rapid  de- 
struction and  liquefaction  of  the  massive  fibrinous  exudates  remain- 
ing after  the  crisis  in  lobar  pneumonia. 

42  Miiller.     Congr.  f.  inn.  Med.,  Wiesbaden,  1902. 

43  Zinsser  and  Gary.     Jour.  A.  M.  A.,  1911. 

44  Jochmann.     Leucozyten  Fermente,  etc.,  "Kolle  u.  Wassermann  Hand- 
buch,"  2d  Ed.,  Vol.  49,  2. 

45  Jochmann.     Zeitschr.  f.  Hyg.,  61,  1908. 

46  Kantorowicz.     Munch,  med,  Woch.,  No.  28,  1909. 

47  Friedrich  Miiller.     "Verhand.  d.  Kongr.  f.  inn.  Med.,"  1902. 


308  INFECTION  AND  RESISTANCE 

From  the  discovery  of  antibacterial  properties  in  the  extracts 
of  leukocytes  it  is  but  a  logical  step  to  the  attempt  to  utilize  these 
substances  therapeutically.  Petterson 48  was  probably  the  first  to 
study  this  phase  of  the  problem  systematically  in  connection  with 
anthrax  infection  in  dogs  and  rabbits.  In  preliminary  studies  he 
claimed  to  have  determined  that  when  leukocytes  are  left  in  con- 
tact with  serum  for  four  hours  or  longer  there  develops  in  the  mix- 
ture a  bactericidal  power  far  superior  to  that  which  is  possessed 
by  these  elements  when  separately  kept  in  salt  solution  and  mixed 
only  just  before  the  bactericidal  tests.  He  attributes  this  to  the 
fact  that  in  dogs,  at  least,  the  leukocytes  furnish  bactericidal  sub- 
stances to  the  serum — an  assumption  which  is  entirely  in  accord 
with  the  earlier  opinion  of  Denys  and  Kaisin,49  which  we  have  men- 
tioned in  another  place.  In  direct  continuance  of  these  experiments 
he  injected  leukocytes  into  dogs  at  the  same  time  at  which  he  in- 
fected them  with  anthrax  and  observed  a  moderately  protective  in- 
fluence, which,  however,  he  admits  was  not  very  great.  He  followed 
this  work  in  1906  with  similar  observations  on  the  protective  influ- 
ence of  leukocytes  in  intraperitoneal  infections  of  guinea  pigs  with 
typhoid  bacilli.  In  these  experiments  50  he  made  the  curious  ob- 
servation that,  although  such  protective  influence  was  unquestionable, 
the  guinea  pig  leukocytes  contained  no  bactericidal  substances  active 
against  typhoid  bacilli.  In  consequence  he  concluded  that  the  de- 
struction of  these  bacteria  in  the  guinea  pig  was  due  entirely  to  the 
serum-antibodies  absorbed  by  the  microorganisms  before  phago- 
cytosis, even  when  the  actual  destructive  process  took  place  intra- 
cellularly.  The  protective  effect  following  on  the  injection  of  the 
leukocytes  he  attributed  to  an  indirect  influence  of  the  leukocytic 
substances  in  stimulating  the  more  rapid  accumulation  of  alexin  or 
complement  in  the  peritoneum,  with  consequently  more  powerful 
phagocytosis.  Following  this,  in  1908,  Opie  51  carried  out  experi- 
ments in  which  he  observed  that  leukocytes  injected  intrapleurally 
into  dogs,  together  with  tubercle  bacilli,  exerted  a  distinct  protec- 
tion. 

In  the  same  year  extensive  observations  on  the  protective  prop- 
erties of  leukocyte  extracts  were  published  by  Hiss. 

Hiss52  worked  at  first  with  extracts  of  dog,  rabbit,  and  guinea 
pig  leukocytes;  later  he  confined  himself  entirely  to  rabbit  leuko- 
cytes. He  extracted  the  leukocytes  at  first  by  repeated  freezing  and 
thawing  in  physiological  salt  solution,  but  the  technique  of  hi»  sub- 
sequent work  was  uniformly  as  follows:  Intrapleural  injections  of 

48  Petterson.     Centralbl  f.  Bakt.,  Vol.  36,  1904. 

49  Denys  and  Kaisin.     "La  Cellule,"  Vol.  9,  1893. 

50  Petterson.     Centralbl  f.  Bakt.,  Vols.  40  and  42,  1906. 

51  Opie.    Jour.  Exp.  Med.,  1908. 

52  Hiss.    Jour,  of  Med.  Res.,  new  series,  Vol.  14,  1908. 


RELATION  OF  LEUKOCYTES  TO  IMMUNITY       309 

aleuronat  emulsions  were  made  in  rabbits  and,  after  about  24  hours, 
the  resulting  exudates  were  taken  away  with  sterile  pipettes  and 
centrifugalized  before  clotting  could  take  place;  the  serum  was  de- 
canted and  the  leucocytes  then  emulsified  in  distilled  water,  in  quan- 
tity about  equal  to  the  amount  of  serum  poured  off.  In  this  the  leu- 
cocytes were  allowed  to  stand  for  a  few  hours  at  incubator  tempera- 
ture, and  then  in  the  ice-box  until  used.  For  his  experimental  work 
in  both  animals  and  man,  in  most  instances,  not  only  the  clear  super- 
natant fluid  was  injected,  but  the  cell  residue  as  well. 

With  leucocytic  extracts  so  prepared  Hiss  treated  staphylococ- 
cus,  typhoid  bacillus,  pneumococcus,  streptococcus,  and  meningococ- 
cus  infections  in-  rabbits  and  obtained  results  which  justified  him  in 
concluding  that  the  leucocyte  extract  exerted  strong  protective  action 
in  all  of  these  cases.  Many  of  his  animals  survived  infections  fatal 
to  controls  even  when  the  treatment  was  delayed  as  long  as  24  hours 
after  infection.  Subsequently  Hiss  and  Zinsser  53  treated  series  of 
patients,  ill  with  pneumonia,  meningitis,  and  staphylococcus  infec- 
tions, with  leucocyte  extracts  prepared  by  the  method  of  Hiss,  and 
felt  that  they  were  justified  in  concluding  that  in  many  cases,  at 
least,  the  course  of  the  disease  was  favorably  influenced  by  the  leuco- 
cyte extract.  Favorable  results  have  since  then  been  obtained  also  by 
Lambert  in  erysipelas,  and  by  Hiss  and  Dwyer  in  a  variety  of  con- 
ditions. 

While  there  seems  to  be  little  question  about  the  actually  favor- 
able influence  of  the  leucocyte  extract,  both  in  experiments  with 
animals  and  in  the  treatment  of  human  cases,  there  has  been  consid-  • 
erable  difficulty  in  determining  the  reasons  for  this  influence.     In  . 
subsequent  studies  Hiss  and  Zinsser  (loc.  cit.)  were  able  to  show  that 
/-"the  extracts  did  not  favor  phagocytosis  and  that  the  moderate  bacteri- 
cidal properties  possessed  by  the  leucocytic  substances  could  not  ac-  » 
count  for  their  effectiveness.     There  did  seem  to  be  a  more  rapid 
v  accumulation  of  phagocytes  in  the  peritoneal  cavities  of  guinea  pigs 
infected  with  cholera  spirilla  when  leucocyte  extract  was  injected   . 
with  the  bacteria,  and  it  is  not  impossible,  in  fact,  it  seems  probable 
to  the  writer,  from  subsequent  experience,  that  the  protective  prop-    « 
erties  of  the  leucocyte  extracts  are  attributable,  in  part  at  least,  to     « 
their  positively  chemotactic  effect. 

We  are  inclined  to  believe  at  present  that  the  beneficial  effects  of  leucocyte 
extracts  are  based  on  the  same  principles  as  those  which  determine  the  reactions 
following  on  the  injection  of  bacterial  and  any  other  protein. 

In  this  connection  a  very  interesting  problem  has  arisen — namely,  that  spoken 

of  as  the  phenomenon  of  Specific  Hyperleucocytosis.     Bordet,  as  early  as  1896, 

made  the  following  statement,  "Active  immunity  has  also  other  characteristics,  in 

that  there  is  an  increase  of  the  number  of  leucocytes,  that  is,  an  'exaltation'  of 

the  chemotactic  sensibilities  of  the  leucocytes."     He  suggests  herein  that  an  im- 

^  munized  animal  may  respond  with  a  more  powerful  leucocytic  reaction  to  the  in- 

"  jection  of  the  infectious  agent  than  would  a  normal  animal  similarly  treated.    This 

53  Hiss  and  Zinsser.    Jour,  of  Med.  Res.,  new  series,  Vol.  14,  1908. 


310  INFECTION  AND  RESISTANCE 

idea  has  recently  found  experimental  elaboration  in  the  work  of  Gay  and  Claypole, 
who  found  that  the  reinjection  of  immune  animals  with  the  homologous  bacteria 
produced  a  specific  hyperleucocytosis,  that  is,  typhoid  immune  animals  receiving 
typhoid  bacilli  would  respond  with  counts  ranging  up  to  150,000  leucocytes  per 
cu.  mm.,  whereas  the  normal  animals  rarely  showed  more  than  40,000  to  50,000. 

This  observation  would  tend  to  indicate  a  great  advantage  of  the  specific  over 
the  non-specific  methods  of  treatment. 

Unfortunately  the  results  of  Gay  and  Claypole  M  have  not  found  confirmation. 
McWilliams55  in  similar  experiments  found  no  differences  in  the  degree  of  leuco- 
cytosis  between  normal  and  immune  animals  in  response  to  the  injection  of  bacteria 
and  reported  that  the  same  degree  of  response  followed  in  typhoid  immune  animals 
when  injected  with  Bacillus  coli  as  when  typhoid  bacilli  were  administered.  In 
part,  this  is  also  stated  to  be  the  experience  of  Jobling  and  Petersen. 

The  question  is  such  an  important  one  that  the  writer,  with  Dr.  Tsen,56  re- 
investigated  it  in  connection  with  work  on  the  therapeutic  effect  of  leucocytic  ex- 
tract. Our  conclusions  showed  little  agreement  with  those  of  Gay  and  Claypole. 

We  found  that  when  homologous  Gram-negative  bacilli  are  injected  into  im- 
munized animals  there  seems  to  be  a  definitely  higher  leucocytosis  in  the  immu- 
nized animals  than  in  normal  controls  similarly  treated.  The  contrasts  in  our 
experiments,  however,  were  nothing  like  as  striking  as  those  reported  by  Gay  and 
Claypole.  Indeed  the  contrast  in  general  is  so  slight  and  so  irregular  that  in  the 
case  of  the  Gram-negative  bacilli  we  were  at  first  inclined  to  agree  with  McWil- 
liams. There  was,  however,  a  sufficiently  definite  difference  in  an  average  of  many 
counts  to  convince  us  that  this  was  more  than  coincidence. 

In  the  case  of  the  Gram-positive  cocci  there  was  a  more  marked  difference,  in 
that  the  immunized  animals  reacted  more  promptly  and  very  much  more  ener- 
getically than  did  the  normal  animals. 

It  seems  reasonably  clear,  then,  that  an  animal  reacts  more  energetically  as 
far  as  its  mobilization  of  leucocytes  is  concerned  when  reinjected  than  does  a 
normal  animal  treated  with  the  same  variety  and  quantity  of  bacteria. 

The  reaction  is  dependent  upon  a  number  of  factors,  chief  among  which  are: 
(1)  the  condition  of  the  animal  (loss  of  weight,  etc.) ;  (2)  the  amount  of  bacteria 
injected,  and  (3)  the  interval  between  injections.  These  factors  all  very  naturally 
signify  the  necessity  of  avoiding  too  profound  an  intoxication  of  the  animal. 

When  immune  animals  are  treated  with  heterologous  bacteria — that  is,  when 
prodigiosus  bacilli  or  colon  bacilli  are  injected  into  typhoid  immune  animals  and 
vice  versa — there  seems  to  be  no  specific  difference  in  response.  That  is,  the  in- 
jection of  colon  bacilli  into  typhoid  animals  has  shown  no  marked  difference  in 
leucocytic  response  from  that  observed  when  typhoid  bacilli  were  injected  into  a 
typhoid  animal.  In  this  our  figures  correspond  with  those  of  McWilliams.  They 
are  also  in  keeping  with  the  clinical  experience  of  Kraus,  Ichikawa,  and  others 
which  have  been  mentioned  above. 

The  injection  of  leucocytic  extract  does  not  arouse  as  vigorous  a  leucocytie 
Tesponse  as  does  the  injection  of  bacillary  protein. 

In  reading  these  facts  superficially  they  at  first  seem  to  be  contradictory  in 
significance,  inasmuch  as  specificity  seems  to  exist  in  the  fact  that  typhoid 
immune  rabbits  or  streptococcus  immune  rabbits  respond  somewhat  more  vigor- 
ously than  do  normal  controls  injected  with  the  same  substance. 

However,  we  think  that  these  relations  are  explained  by  the  fact  that  animals 
that  have  reacted  to  such  organisms  as  the  typhoid  bacillus,  etc.,  develop  a  certain 
amount  of  non-specific  tolerance  against  the  proteose-like  substances  which  are 
probably  responsible  for  a  not  unimportant  part  of  the  symptoms  caused  by  the 
bacteria.  Such  tolerance  has  been  shown  in  the  experiments  made  by  the  writer 
with  Dwyer. 

To  summarize,  therefore,  we  do  not  think  that  at  present  a  specific  leucocytosia 
In  the  sense  of  Gay  and  Claypole  has  been  demonstrated,  but  believe  that  an  ani- 
mal, immune  to  one  microorganism,  will  have  a  slight  non-specific  increase  of 
resistance  to  other  organisms. 

This  does  not  mean  that  the  immunity  is  non-specific.  The  destruction  of 
living  bacteria  is  still  a  purely  specific  process,  and  this,  of  course,  would 
determine  the  occurrence  of  outcome  of  an  infection. 

5*  Gay  and  Claypole.    Arch,  of  Int.  Med.,  V,  XIV,  1914,  p.  662. 

55  McWilliams.    Journal  of  Immunology,  Vol.  I,  No.  2,  1916,  p.  159. 

56  Zinsser  and  Tsen.    Journal  of  Immunology }  Vol.  II,  No.  3,  1917,  p.  247. 


CHAPTER   XIII 

FACTOID    DETEKMINING   PHAGOCYTOSIS 

OPSONINS,   TBOPINS 

FROM  the  very  beginnings  of  his  researches  upon  phagocytosis 
Metchnikoff  recognized  that  the  process  was  profoundly  influenced 
by  the  properties  of  the  fluid  constituents  of  the  blood  plasma  in 
which  the  phenomenon  occurred.  Both  he  and  his  pupil  Bordet,1 
at  this  time  working  in  MetchnikofFs  laboratory,  noticed  that  the 
phagocytic  activity  of  leukocytes  was  greater  in  immune  than  in 
normal  sera  and  associated  this  with  the  specific  properties  of  the 
immune  substances  or  antibodies  in  these  sera ;  Metchnikoff  himself 
interpreted  the  phagocytosis-enhancing  power  of  the  serum  as  a 
stimulation  of  the  leukocytes  and  referred  to  the  serum  constituents 
by  which  this  effect  was  produced  as  "stimulins."  A  closer  analysis 
of  the  factors  involved  in  this  interrelationship,  however,  was  not 
attempted  at  this  time  by  him  or  his  pupils,  although  indirect  ref- 
erence was  made  to  it  in  a  number  of  articles  emanating  from  this 
school  in  the  course  of  investigations  on  kindred  problems  of  phago- 
cytosis. Thus  Gabritschewsky,2  in  1894,  published  a  paper  on 
"Leukocytose  dans  la  Diphtheric/7  in  which  he  concluded  that  the 
poison  of  diphtheria  bacilli,  among  other  harmful  effects,  diminished 
the  phagocytic  power  of  the  leukocytes,  and  that  one  of  the  beneficial 
influences  of  the  curative  serum  was  to  render  these  and  other  cells 
""less  sensitive  to  the  bacterial  poisons."  This  may  be  interpreted 
as  indicating  an  assumption  that  the  action  of  an  immune  serum  in 
increasing  phagocytic  activity  rested  rather  upon  its  influence  upon 
the  bacterial  products  than  upon  any  stimulation  of  the  phagocytes 
themselves.  However,  in  diphtheria  the  action  of  the  leukocytes 
was,  even  at  this  time,  recognized  as  a  merely  secondary  one,  and 
Gabritschewsky's  results  did  not  materially  influence  the  "stimulin" 
conception. 

The  first  extensive  investigation  which  occupied  itself  directly 
with  these  problems  was  that  of  the  Belgian  bacteriologists  Denys 
and  Leclef.3  The  publication  of  these  workers  deals  primarily  with 

1  Bordet.     Ann.  de  I'Inst.  Past.,  1895. 

2  Gabritschewsky.     Ann.  de  I'Inst.  Past.,  1894. 

3  Denys  and  Leclef.    La  Cellule,  11,  1895. 


312  INFECTION    AND    RESISTANCE 

the  nature  of  streptococcus  immunity  in  rabbits.  It  established,  first 
of  all,  the  paramount  importance  of  phagocytosis  in  the  resistance 
of  animals  against  these  bacteria,  and  made  clear  that  the  destruction 
of  bacteria  was  carried  out  equally  as  well  by  the  leukocytes  of  nor- 
mal as  by  those  of  immune  animals,  but  was  powerfully  enhanced 
when  either  normal  or  "immune"  leukocytes  were  combined  with 
immune  serum.  Their  work,  therefore,  indicated  again  that  the  in- 
creased phagocytosis  of  virulent  bacteria,  taking  place  in  immune 
animals,  depended  clearly  upon  alterations  in  the  functions  of  the 
serum  rather  than  in  those  of  the  cells,  and  they  suggested  that 
the  influence  of  this  serum  was  not  necessarily  one  of  leukocyte 
stimulation,  but  might  rather  consist  in  action  upon  the  bacteria, 
rendering  them  less  resistant  to  phagocytosis.  They  say  in  sub- 
stance: "A  notre  avis,  on  pourrait  tout  aussi  bien  admettre  que  la 
substance  vaccinante  ou  antitoxique  agit,  non  pas  sur  le  leukocyte, 
mais  sur  un  poison  renferme  dans  le  corps  du  microbe  ou  dissous 
dans  le  milieu,  et  qui  preserve  le  micro-organisme  centre  les  at- 
teintes  du  leukocyte."  4 

In  this  statement  we  have,  in  brief,  the  distinct  formulation  of 
our  present  view  of  the  "dpsonins."  5 

Observations  with  pneumococci  and  streptococci  carried  out 
after  this  by  Marchand  6  and  by  Mennes,7  whose  investigations  we 
cannot  discuss  in  detail,  beside  confirming  most  of  the  observations 
of  Denys  and  Leclef,  brought  out  especially  the  relation  of  the 
virulence  of  micro-organisms  to  phagocytosis,  showing  that  very 
virulent  strains  were  taken  up  to  a  slight  degree  only  in  the  presence 
of  normal  serum,  but  were  subject  to  active  phagocytosis  when  im- 
mune serum  was  employed.  This,  too,  seemed  to  point  primarily  to 
the  fact  that  the  serum  influenced  rather  the  bacteria  than  the 
phagocytes,  although  no  convincing  proof  is  brought  for  this  in  their 
publications.  Though  much  that  had  bearing  indirectly  on  this 
problem  was  written  during  the  following  years,  no  definite  progress 
was  made  beyond  the  results  of  Denys  and  his  pupils  until  1902, 

*  In  our  opinion  one  can  just  as  well  believe  that  the  vaccinating  or  anti- 
toxic substance  acts  not  upon  the  leukocyte  but  upon  a  poison  inclosed  within 
the  body  of  the  bacteria  or  dissolved  in  the  medium,  which  preserves  the 
micro-organism  against  the  attacks  of  the  leukocyte. 

5  Denys  formulated  this  view  with  still  greater  clearness  and  positiveness 
at  the  Congress  of  Hygiene  held  at  Brussels  in  1903.     We  take  our  citation 
from  the  discussion  on  opsonins  by  Gruber  (3d  meeting  Freie  Vereinigung  f. 
Mikrobiol.,  Vienna,  1909,  Centralbl.  /.  Bakt.,  I  Ref.,  Vol.  44,  Suppl.  p.  3). 
Following  is  Denys'  statement :     1.  The  phagocytosis  in  immune  sera  is  de- 
pendent   upon    substances    which    are    precipitated    with    the    euglobulins. 
2.  These  substances  cause  phagocytosis  by  inciting  a  physical  alteration  of 
the  micro-organisms.     3.  These  substances  are  specific. 

6  Marchand.    Arch,  de  Med.  Exp.,  1898. 

7  Mennes.     Zeitschr.  f.  Hyg.,  Vol.  25. 


FACTORS    DETERMINING    PHAGOCYTOSIS 

when  Leischman  8  introduced  a  technique  by  means  of  which  it  be- 
came possible  to  observe  the  process  of  phagocytosis  with  fresh  serum 
and  leukocytes  in  vitro. 

By  utilizing  this  technique  and  improving  upon  it  Wright  and 
Douglas  in  the  following  year  (1903)  evolved  a  method  by  means  of 
which  phagocytic  activity  could  be  quantitatively  measured  with 
reasonable  accuracy.  They  worked  at  first  with  staphylococcus 
phagocytosis  by  human  leukocytes  in  the  presence  of  human  citrate 
plasma,  a  research  undertaken  primarily  because  Wright,9  in  collabo- 
ration with  Windsor,  had  previously  determined  that  human  blood 
serum  possessed  practically  no  bactericidal  power  for  this  organism, 
and  that  phagocytosis  was  probably  the  chief  mechanism  of  protection 
which  the  human  body  possessed  against  these  bacteria.  The  re- 
searches of  Wright  and  Douglas  10  were  carried  out  chiefly  by  mixing 
equal  volumes  of  bacteria,  serum,  and  leukocytes  (in  citrate  sus- 
pension),11 allowing  these  elements  to  remain  together  at  37.5°  C. 
for  varying  periods,  then  staining  on  slides  and  determining  the 
degree  of  phagocytosis  by  counting  the  numbers  of  bacteria  taken  up 
by  each  polynuclear  leukocyte.  Though  many  technical  difficulties 
had  to  be  overcome,  and  although  the  method  at  its  best  still  permits 
of  much  personal  error,  careful  work  and  untiring  repetition  made 
possible  a  considerable  degree  of  accuracy,  and  definite  facts  regard- 
ing the  mechanism  of  phagocytosis,  heretofore  merely  suspected,, 
could  be  recorded.  The  most  important  result  of  these  investiga- 
tions was  the  unquestionable  establishment  of  the  function  of  serum 
in  the  process  of  phagocytosis,  namely,  that  it  in  no  way  "stimu- 
lated" the  leukocytes  in  the  sense  of  Metchnikoff,  but  rather  acted 
entirely  upon  the  bacteria,  preparing  them  for  ingestion.  For  this 
reason  Wright  coined  the  word  "opsonins"  (o^oi/«o=  I  prepare  food) 
for  the  serum  constituents  which  brought  about  this  effect,  believing 
them  to  be  new  antibodies,  entirely  distinct  from  the  other  serum 
antibodies  heretofore  recognized. 

Wright  and  his  followers  now  concluded  that  the  role  of  the 
leukocyte  in  taking  up  bacteria  was  entirely  dependent  upon  the 
opsonin  contents  of  the  serum.  In  a  menstruum  containing  no 
serum,  or  in  a  serum  in  which  the  opsonins  had  been  destroyed  by 
heat,  they  found  practically  no  phagocytic  action  on  the  part  of 
washed  serum-free  leukocytes,  and  they,  therefore,  doubted  the  oc- 
currence of  spontaneous  phagocytosis  on  the  part  of  leukocytes 
themselves. 

8  Leischman.    Brit.  Med.  Jour.,  Vol.  2,  1901,  and  Vol.  1,  1902. 

9  Wright  and  Windsor.     Jour,  of  Hyg.,  Vol.  2,  1902. 

10  Wright   and   Douglas.     Proc.  Roy.   Soc.,   72,   1903,   73   and   74,   1904. 
See  also  Wright,  "Studien  liber  Immunisierung,"  Fischer,  Jena,  1909. 

11  At  first  bacteria  were  merely  mixed  in   equal  volumes  with   citrated 
whole  blood. 


INFECTION    AND    RESISTANCE 

In  this  point  it  is  not  unlikely  that  Wright  is  mistaken,  since 
•other  observers,  notably  Lohlein,12  have  observed  the  phagocytosis  of 
various  bacteria  by  washed  leukocytes  in  indifferent,  opsonin-free 
media.  Although  we  may  take  it  as  assured  that  such  spontaneous 
phagocytosis  may  take  place  (Metchnikoff  and  a  number  of  others 
having  obtained  results  similar  to  those  of  Lohlein),  this  is  prob- 
ably never  very  intense. 

In  fact,  Wright,  in  some  of  his  later  work,  does  not  insist  rigidly 
upon  the  non-occurrence  of  spontaneous  phagocytosis,  but  attempts 
to  associate  such  phenomena  with  the  salt  contents  of  the  medium 
in  which  it  occurs.  Together  with  Reid,13  he  determined  that  spon- 
taneous phagocytosis  of  tubercle  bacilli  unquestionably  takes  place, 
is  most  intense  at  a  concentration  of  about  0.6  per  cent.  JSTaCl,  and 
diminishes  as  the  concentration  is  increased.  This,  as  we  shall  see, 
has  bearing  on  the  possible  physical  explanations  advanced  to  ac- 
count for  opsonic  action,  and  has  its  parallels  in  experiments  on  the 
influence  of  electrolytes  on  agglutination  and  precipitation. 

The  fact  remains  that  Wright  demonstrated  by  his  work  that 
Metchnikoff's  original  view,  which  interpreted  the  difference  be- 
tween susceptibility  and  immunity  as  a  difference  between  the  in- 
herent phagocytic  powers  of  the  leukocytes,  is  incorrect,  and  that 
the  essential  regulating  influence  affecting  phagocytosis  rests  upon 
the  action  of  the  serum  upon  the  bacteria. 

The  following  experiment  from  the  work  of  Hektoen  and  Rue- 
diger  -14  illustrates  this  point  with  exceptional  clearness.  It  shows 
that  human  leukocytes  in  the  presence  of  normal  defibrinated  blood 
will  take  up  bacteria  energetically.  When  the  leukocytes,  however, 
are  washed  free  of  blood  and  added  to  untreated  bacteria  phago- 
cytosis is  practically  nil.  If,  however,  such  washed  leukocytes  are 
mixed  with  bacteria  that  have  been  previously  in  contact  with  serum 
active  phagocytosis  will  take  place.  In  other  words,  the  bacteria 
have  been  altered  by  the  serum  in  such  a  way  that  they  are  now 
amenable  to  phagocytosis  by  washed  leukocytes.  The  serum  then 
acts  upon  the  bacteria  and  not  upon  the  leukocytes. 

TABLE    II 

Phagocytosis  by  Human  Leukocytes  of  Sensitized  Bacteria 

Average 
Phagocytosis 

Human  leukocytes  (defibrinated  blood)  +  Staphylococcus  aureus 22 . 

Human  leukocytes  (washed  in  NaCl  solution)  +  Staphylococcus  aureus.     1 . 2 
Human  leukocytes  (washed  in  NaCl  solution)  -f-  Staphylococcus  aureus 

(treated  with  human  serum) 10 . 

Human  leukocytes  (defibrinated  blood)  +  Streptococcus  300 22 . 

12  Lohlein.  Centralbl.  f.  Bakt.,  38,  1906,  Beihef t,  p.  32 ;  also  Munch,  med. 
Woch.,  1907,  p.  1473. 

is  Wright  and  Reid.    Proc.  of  Eoyal  Soc.  B.,  Vol.  77,  1906. 

14  Hektoen  and  Ruediger.     Jour.  Inf.  Dis.,  Vol.  2,  1905,  p.  132. 


FACTORS    DETERMINING    PHAGOCYTOSIS        315 

Average 
Phagocytosis 

Human  leukocytes  (washed  in  NaCl  solution)  -f-  Streptococcus  300 ....  1 . 
Human  leukocytes  (washed  hi  NaCl  solution)  +  Streptococcus  (treated 

with  human  serum) 14 . 

Human  leukocytes  (washed  in  NaCl  solution)  +  Streptococcus  (treated 

with  guinea  pig  serum) 12 . 

Human  leukocytes  (washed  in  NaCl  solution)  -f-  Streptococcus  (treated 

with  rabbit  serum) 14 . 

Wright  and  Douglas' 15  work  was  done  at  first  with  normal 
serum  or  normal  citrate  plasma,  and  in  this  case  they  found  that  the 
opsonins  were  essentially  unstable,  being  easily  weakened  by  ex- 
posure to  light,  or  heat,  and  even  when  preserved  in  sealed  tubes  in 
the  dark  they  diminished  noticeably  on  standing  for  5  or  6  days. 
Other  writers  who  have  worked  with  the  opsonic  substances  in  nor- 
mal serum  have  confirmed  this  instability  of  the  normal  opsonin, 
although  even  Wright  himself  admits  that  heating  to  60°  C.  does 
not  entirely  destroy  the  opsonic  power,  though  it  reduces  it  to  a 
minimum.  A  protocol  from  Wright  and  Douglas'  first  paper  will 
best  illustrate  the  degree  of  reduction  of  opsonic  power  resulting 
from  the  exposure  of  normal  serum  to  60-65°  C.  for  10  to  15  minutes. 

A.  Unheated  serum  Wright — Staphylococcus  suspension  1  vol. — Blood  cells 

Wright  3  vols. 

(1)  Phagocytic  average  20  cells 17 .4 

(2)  Phagocytic  average  20  cells 19.8 

B.  Heated  serum  as  above. 

(1)  Phagocytic  average  52  cells 0.6 

(2)  Phagocytic  average  46  cells 3.4 

The  experiments  just  cited  refer  only  to  the  opsonic  powers  of 
normal  serum.  When  an  animal  is  immunized  with  any  particular 
micro-organism  or  other  cellular  antigen,  such  as  red  blood  cells, 
etc.,  a  marked  specific  increase  of  opsonins  occurs,  but  unlike  the 
opsonins  of  normal  serum  these  newly  formed  elements  in  the  im- 
mune serum  seem  to  possess  a  much  greater  resistance  to  heat. 

Neufeld  and  Rimpau,16  who  have  studied  these  constituents  of 
immune  serum  with  especial  thoroughness,  have  shown  that  heating 
to  62°  to  63°  C.  for  as  long  as  three-quarters  of  an  hour  does  not 
destroy  them,  and  that  such  sera  may  be  preserved  for  as  long  as 
several  years  without  their  complete  disappearance.17 

We  may  accept  as  definitely  determined,  therefore,  that  there  is 
a  qualitative  difference  between  the  serum  components  which  initiate 
phagocytosis  in  normal  serum  (normal  opsonins)  and  those  which 
carry  out  the  same  function  to  a  much  more  powerful  degree  in 

is  "Wright  and  Douglas.  Cited  in  Wright,  "Studien  iiber  Immun.,  etc.," 
p.  9. 

16Neufeld  and  Rimpau.  Deutsche  med.  Woch.,  No.  40,  1904;  Zeitschr. 
f.  Hyg.,  Vol.  51,  1905. 

17  Leishman.     Trans.  London  Path.  Soc.,  Vol.  56,  1905. 


316  INFECTION    AND    RESISTANCE 

immune  serum.  This  is  the  more  surprising  since,  in  the  case  of  all 
other  antibodies  (lysins,  agglutinins,  etc.),  it  has  been  shown  that  in 
structure  and  mode  of  action  the  antibodies  of  immune  serum  are 
in  every  way  qualitatively  similar  to  the  corresponding  ones  of  nor- 
mal serum,18,  19  representing  merely  a  specific  quantitative  increase 
of  substances  originally  present  in  small  amount. 

This  difference  between  the  normal  and  immune  opsonic  sub- 
stances has  added  much  difficulty  to  the  investigation  of  the  nature  of 
these  bodies,  and  we  may  approach  the  problem  with  greater  clearness 
by  considering  them  separately,  at  first,  attempting  to  define  the 
relations  between  them  after  we  have  set  down  the  facts  ascertained 
in  connection  with  each. 

In  their  earliest  investigations  upon  the  normal  opsonins  Wright 
and  Douglas  20  regarded  them  as  new  antibodies,  separate  and  dis- 
tinct from  those  already  known.  There  is  no  convincing  proof  of 
this,  and  a  number  of  other  interpretations  of  the  observed  phe- 
nomena are  possible.  Indeed,  the  burden  of  proof  is  rather  upon 
those  who  would  establish  the  existence  of  a  new  antibody,  for  before 
this  can  be  done  it  must  be  shown  that  the  new  function  is  not  merely 
another  property  of  the  serum  constituents  already  known.  For,  as 
Gruber  has  justly  said,  "One  of  the  most  important  attributes  of  the 
natural  scientist  is  economy  of  hypotheses."  And  in  the  case  of  the 
normal  opsonins  there  are  many  good  reasons  for  regarding  them  as 
possibly  identical  with  known  serum  constituents.  The  two  possi- 
bilities suggested  have  been  (1)  Are  the  opsonic  substances  identical 
with  the  alexin  or  complement?  or  (2)  Do  they  represent  the  com- 
bined action  of  the  normal  sensitizer  of  the  serum  activated  by  the 
alexin  ? 

The  similarity  of  normal  opsonin  with  alexin  or  complement  has 
been  brought  out  especially  by  Muir  and  Martin,21  by  Baecher,22 
and  by  Levaditi  and  Inmann.23  The  fact  that  both  are  thermolabile 
has  been  mentioned  above. 

In  addition  to  this,  as  Muir  and  Martin24  have  shown,  all  antigen- 
antibody  complexes  which  absorb  alexin  out  of  serum  at  the  same 
time  remove  the  normal  opsonin.  Thus  sensitized  red  corpuscles, 
sensitized  bacteria,  and  specific  precipitates  added  to  normal  serum 
take  out  its  opsonic  substances.  From  this  fact  they  also  concluded 
that  the  normal  opsonins  like  alexin  were  non-specific.  For  just  as 

18  Dean.     Proc.  Royal  Soc.,  76,  1905. 

19  Neufeld  and  Hiine.    Arb.  a.  d.  kais.  Gesundh.  Amt.}  Vol.  25,  1907. 

20  wright  and  Douglas.     LOG.   cit. 

21  Muir  and  Martin.     Br.  Med.  Jour.,  Vol.  2,  1906;  Proc.  Royal  Soc.  B., 
Vol.  79,  1907. 

22  Baecher.     Zeitschr.  f.  Hyg.,  Vol.  56,  1907. 

23  Levaditi  and  Inmann.     C.   R.   de  la  Soc.  Biol.,  1907,   pp.   683,  725, 
817,  869. 

24  Muir  and  Martin.     Loc.  cit. 


FACTORS    DETERMINING    PHAGOCYTOSIS        317 

the  alexin  of  a  serum  may  serve  to, activate  a  considerable  variety  of 
sensitized  antigens,  so  the  opsonic  action  of  a  normal  serum  may 
functionate  upon  a  large  variety  of  bacteria.  Muir  and  Martin  were 
probably  wrong  in  this  and,  as  we  shall  see  below,  normal  opsonins, 
like  normal  sensitizers,  may  be  regarded  as  specific. 

Similar  to  the  observations  of  Muir  and  Martin  are  those  of 
^vTeufeld  and  Hiine,25  which  showed  that  yeast  cells  will  absorb  both 
alexin  and  opsonin  out  of  serum. 

A  further  similarity  between  the  two  serum  constituents  is  the 
fact  that  both  are  absent  from  the  normal  fluid  of  the  anterior  cham- 
ber of  the  eye,  but  they  together  appear  in  it  after  injury  (puncture 
for  the  first  removal  of  fluid).  A  like  parallelism  between  the  ab- 
sence and  presence  of  both  has  been  shown  for  edema  fluids. 
Furthermore,  phosphorus  poisoning  which  reduces  alexin  likewise 
reduces  opsonin. 

Although  this  parallelism  is  very  striking,  it  does  not  on  this 
account  mean  that  necessarily  the  two  are  identical.  It  may  signify 
merely  that  the  alexin  is  a  necessary  participant  in  normal  opsonic 
action,  essential  in  that  it  activates  a  thermostable  opsonic  constitu- 
ent just  as  it  activates  hemolytic  or  bactericidal  sensitizer. 

This  opinion  has  been  expressed  by  Levaditi,  Neufeld,26  Dean,27 
and  others,  and  indeed  it  is  a  conception  which  seems  most  logical. 
For  the  procedures  which  remove  both  alexin  and  opsonin,  as  stated 
above,  do  not,  as  a  matter  of  fact,  remove  all  the  opsonic  action. 
(Although  Neufeld  maintains  this.28)  Studies  of  Hektoen  and 
others  have  definitely  proved  that,  though  reduced  to  almost  nil, 
nevertheless  heated  serum  shows  definite  though  slight  opsonic  action 
as  compared  with  indifferent  menstrua  such  as  salt  solution.  A 
similar  slight  remnant  of  opsonic  action  after  absorption  of  normal 
serum  with  sensitized  cells,  bacteria,  and  precipitates  is  evident  in  the 
protocols  of  Muir  and  Martin.  The  significance  of  this  point  be- 
comes immediately  clear  when  we  consider  the  properties  of  the  bac- 
teriotropins  or  immune  opsonins,  which  are  heat  stable  and  capable 
of  initiating  opsonic  action  in  the  entire  absence  of  alexin  or  comple- 
ment. It  is  possible,  therefore,  that  there  may  be  present  in  normal 
serum  a  slight  amount  of  specific  thermostable  opsonin,  which, 
though  capable  of  acting  feebly  by  itself,  is  nevertheless  powerfully 
activated  by  alexin — just  as  bactericidal  or  hemolytic  antibody  is 
similarly  activated. 

One  of  the  most  thorough  studies  upon  this  question  is  that  of 

25  Neufeld  and  Hiine.    Arb.  a.  d.  kais.  Gesundh.  Amt.,  Vol.  25,  1907. 

26  Neufeld.      "Kolle    u.    Wassermann's    Handbuch,"    Erganzungsband    2, 
p.  313. 

27  Dean.     Brit.  Med.  Jour.,  2,  1907,  p.  1409. 

28  In  fact  he  states  that  heated  normal  serum  may  be  used  as  a  control 
in  opsonic  experiments  instead  of  salt  solution,, 


318  INFECTION    AND    RESISTANCE 

Cowie  and  Chapin.29  Dean  30  had  previously  shown  that,  although 
heated  immune  serum  was  capable  of  exerting  opsonic  action  by 
itself,  this  action  could  nevertheless  be  enhanced  by  the  addition  of 
a  little  diluted  fresh  normal  serum.  The  particular  significance  of 
Dean's  work  will  be  discussed  later.  Cowie  and  Chapin,  however, 
carried  on  similar  experiments  with  normal  serum  in  which  they  at- 
tempted to  reactivate  heated  normal  serum  by  the  addition  of  small 
amounts  of  diluted  fresh  serum,  by  itself  but  slightly  opsonic.  One 
of  their  experiments  may  serve  to  illustrate  this  point,  as  follows : 

Experiment  10.    June  13,  1907 

Phagocytic 
count 31 

1.  Unheated  serum 15 . 44 

2.  Salt  solution 0. 18 

3.  Heated  serum,  57°  C 1 .08 

4.  Diluted  serum  (1 :15) 1 .56 

5.  Heated  serum  57°  C.  +  diluted  serum  (1 :15) 12 . 40 

6.  Unheated  serum  +  unheated  serum 16 . 08 

This  experiment  and  others  like  it  seem  to  demonstrate  clearly 
that  the  opsonic  action  of  normal  serum,  though  dependent  largely 
upon  alexin,  is  nevertheless  also  dependent  upon  a  heat-stable  body, 
comparable  to  the  sensitizer  or  amboceptor,  in  that  it  is  reactivable  to 
almost  the  full  power  of  the  original  condition  (before  heating)  by 
slight  amounts  of  alexin — in  themselves  almost  inactive.32 

These  findings  were  later  confirmed  by  Eggers,34  and  it  is  plain 
from  this  work  that  the  apparent  opsonic  inactivation  of  normal 
serum  by  heat  depends  upon  the  destruction  of  the  heat-sensitive 
constituent  only — the  heat-stable  substance — surely  involved  in  the 
process,  remaining  intact,  and  reactivable. 

Closely  associated  with  this  phase  of  the  problem  is  that  of  the 
specificity  of  the  normal  opsonins.  For  if,  as  at  first  supposed,  the 
normal  opsonins  are,  like  complement  or  alexin,  non-specific,  the 
above  amboceptor-complement  structure  of  this  mechanism  would 
be  rendered  unlikely.  Earlier  work  upon  this  question  was  con- 
tradictory. Bulloch  and  Western,35  working  with  staphylococci  and 

29  Cowie  and   Chapin.     Jour.  Med.  Ees.,  Vol.  17,  1907,  pp.  57,  95  and 
213. 

30  Dean.     Loc.  cit. 

31  Phagocytic  count  =  average  number  of  bacteria  in  each  leukocyte. 

32  In   earlier   experiments   Hektoen   and   Ruediger 3S   did  not   succeed   in 
reactivating  heated  sera  and  concluded  that  normal  opsonins  had  the  hypo- 
thetical structure  of  toxins  in   that   they   possessed   a   haptophore   and   an 
opsonophore  group.     From   this   point   of   view   Hektoen   has   subsequently 
receded  largely  because  of  work  done  under  his  own  direction. 

33  Hektoen  and  Ruediger.    Jour.  Inf.  Dis.,  1905. 

34  Eggers.     Jour,  of  Inf.  Dis.,  Vol.  5,  1908. 

35  Bulloch  and  Western.     Proc.  Roy.  Soc.  B.,  77,  1906. 


FACTORS    DETERMINING    PHAGOCYTOSIS         319 

tubercle  bacilli,  found  that  each  of  these  organisms  absorbed  out 
separately  specific  opsonins  from  normal  serum,  leaving  those  for 
other  bacteria  but  slightly  reduced.  Slight  reduction  of  the  opsonic 
action  for  other  micro-organisms  might  easily  be  explained  by  a 
partial  removal  of  complement  which  is  bound  to  take  place  in  such 
experiments.  Simon,  Lamar  and  Bispham,36  and  some  others  failed 
to  find  any  such  specificity.  Russell,37  Axamit  and  Tsuda,38  and  a 
number  of  others  obtained  similar  negative  results — in  that  a  num- 
ber of  different  bacteria  seemed  to  absorb  opsonins  out  of  normal 
serum  indiscriminately  and  without  specificity.  On  the  other  hand, 
more  recent  careful  work  by  Rosenow,39  by  Macdonald,40  and  by 
Hektoen  41  has  upheld  the  original  contention  of  Bulloch  and  West- 
ern. The  work  of  Rosenow,  in  which  pneumococci  were  shown  to 
absorb  out  their  specific  opsonins  from  normal  human  serum,  taking 
out  in  part  only  those  for  streptococci,  staphylococci,  and  tubercle 
bacilli,  is  especially  convincing,  and  the  experiment  of  Hektoen  with 
normal  hemopsonins  (opsonins  which  cause  the  phagocytosis  of  red 
blood  cells)  bear  him  out. 

It  seems  fair  to  conclude,  therefore,  that  normal  opsonins  de- 
pend upon  the  cooperation  of  a  heat-stable  and  a  heat-sensitive  body. 
The  heat-stable  body,  analogous  to  normal  sensitizer  or  amboceptor,  is 
specific  and  reactivable  by  the  heat-sensitive  body  which  appears  to 
be  identical  with  alexin.  This  statement  merely  asserts  the  facts  of 
the  dual  mechanism  of  the  process  without  assuming  necessarily  the 
identity  of  the  heat-stable  body  with  sensitizer  or  that  of  the  heat- 
sensitive  one  with  alexin,  though  this  seems  extremely  probable. 

This  question  we  will  discuss  again  more  particularly  in  connec- 
tion with  the  bacteriotropins  or  immune  opsonins. 

Further  proof  for  such  a  complex  constitution  of  the  normal 
opsonins  has  been  adduced  by  means  of  absorption  experiments  at 
0°  C. — by  Cowie  and  Chapin.  In  our  discussion  of  the  lytic  anti- 
bodies we  have  seen  that  sensitizer  or  amboceptor  may  be  absorbed 
from  serum  by  its  specific  antigen  at  0°  C. — but  that  the  attachment 
of  alexin  takes  place  only  when  the  temperature  is  raised  above  this. 
Practically  no  alexin  is  bound  at  the  low  temperature.  Cowie  and 
Chapin,  applying  this  method  of  investigation,  showed : 

1.  That  normal  human  serum  may  have  its  opsonic  power  for 
staphylococci  removed  by  absorption  with  staphylococci  at  0°  C. 

2.  Serum  so  treated  retains  the  power  of  reactivating  the  op- 
sonin  of  heated  normal  serum. 

36  Simon,  Lamar,  and  Bispham.    Jour.  Exp.  Med.,  Vol.  8,  1906. 

37  Russell.     Johns  Hopk.  Bull.,  Vol.  18,  1907. 

38  Axamit  and  Tsuda.     Wien.  klin.  Woch.,  Vol.  20,  No.  35,  1907. 

39  Rosenow.     Jour.  Inf.  Dis.,  Vol.  4,   1907. 

40  Macdonald.     Quoted  from  Hektoen,  loc.  cit.;  Aberdeen  Univ.  Studies, 
Vol.  21,  1906,  p.  323. 

41  Hektoen.    Journ.  Inf.  Dis.,  Vol.  5,  1908. 


320  INFECTION    AND    RESISTANCE 

3.  Staphylococci  so  treated  are  more  easily  subject  to  phago- 
cytosis in  the  presence  of  dilute  normal  serum,  or  normal  serum 
which  has  been  inactivated  by  contact  with  staphylococci  in  the  cold, 
than  are  the  same  bacteria  untreated. 

Kurt  Meyer  42  has  carried  out  similar  experiments  with  paraty- 
phoid bacilli  and  normal  serum,  and,  though  his  work  is  less  exten- 
sive, he  reaches  the  same  conclusion  as  Cowie  and  Chapin. 

We  may  accept,  therefore,  as  fairly  well  established  that  the 
opsonic  power  of  normal  serum  depends  upon  a  complex  mechan- 
ism consisting  of  (a)  a  thermostable  substance  comparable  to 
amboceptor  or  sensitizer,  probably  specific,  but  present  in  very 
small  amount,  and  (b)  a  thermolabile  substance  probably  identical 
with  alexin  or  complement  which  powerfully,  but  non-speci- 
fically,  enhances  the  slight  opsonic  power  of  the  thermostable 
substance. 

In  considering  this  conception,  together  with  the  subsequent  dis- 
cussion of  bacteriotropins  or  immune  opsonins,  it  will  be  well  to 
remember  that  in  normal  inactivated  sera  the  thermostable  opsonic 
constituent  differs  in  its  action  from  the  bodies  we  speak  of  as  ambo- 
ceptors  or  sensitizers  in  that  it  may  functionate  for  phagocytosis  by 
itself — entirely  without  alexin — while  neither  bactericidal  nor  he- 
molytic  effects  can  be  brought  about  by  sensitizer  alone.  Does  this 
definitely  exclude  the  identity  of  this  thermostable  opsonic  substance 
and  sensitizer  ?  It  is  indeed  an  argument  against  identification,  but 
in  opsonic  action,  we  must  remember,  there  is  merely  a  sensitization 
to  the  action  of  the  phagocyte.  This  phagocyte  may  in  itself  be 
capable  of  furnishing  a  small  amount  of  substance  comparable  in 
action  to  alexin — in  fact,  we  have  seen  that  the  origin  of  alexin  from 
leukocytes  is  still  suspected  by  a  number  of  workers.  At  any  rate 
the  phagocyte  is  a  living  cell  which  may  well  be  capable  of  supplying 
in  itself  to  some  degree  the  necessary  activation,  and  therefore  the 
difference  cited  above  is  not  necessarily  a  proof  that  the  normal 
thermostable  opsonic  constituent  is  different  from  normal  sensitizer 
or  amboceptor. 

The  difference  between  the  opsonic  action  of  normal  serum  and 
that  of  immune  serum,  then,  is  the  fact  that  heating  to  from  56°  to 
60°  C.  almost  completely  destroys  the  former,  whereas  it  has  but 
slight  if  any  diminishing  effect  upon  the  latter.  The  immune  op- 
sonins, or,  as  Neufeld  and  Kimpau  have  called  them,  bacteriotropins, 
therefore  are  thermostable.  This  was  determined  as  early  as  1902 
by  Sawtschenko,43  and  was  subsequently  studied  with  great  accuracy 

42  Kurt  Meyer.     Berl  klin.   Woch.,  1908,  p.  951. 

43  Sawtschenko.    Ann.  de  I'Inst.  Past.,  Vol.  16,  1902,  quoted  from  Levaditi. 


FACTORS    DETERMINING    PHAGOCYTOSIS 

by  Neufeld  and  Rimpau,44  Neufeld  and  Topfer,45  Dean,46  Hektoen,47 
and  others.  It  was  shown  that  when  an  animal  is  immunized  with 
any  given  bacterium  or  other  cellular  antigen  (blood  corpuscles, 
etc.)  opsonic  substances  specific  for  the  particular  antigen  appear  in 
considerable  quantities,  and  these  are  but  slightly,  if  at  all,  dimin- 
ished when  the  serum  is  heated ;  Neufeld  and  Hiine  48  found  that 
heating  for  as  long  as  three-quarters  of  an  hour  to  63°  C.  did  not 
noticeably  reduce  the  activity  of  the  bacteriotropins  of  immune 
serum,  and  that,  again,  unlike  the  normal  opsonins,  prolonged  pres- 
ervation, under  sterile  conditions,  changes  them  but  slowly. 

These  facts  alone  indicate  a  close  similarity  between  the  bac- 
teriotropins and  the  other  well-known  thermostable  constituents  of 
immune  sera,  and  the  question  here  again  immediately  arises  whether 
we  are  to  regard  them  as  identical  with  any  of  the  other  specific 
antibodies  or  as  distinct  substances  independent  of  these. 

It  was  suggested  early  in  these  investigations  by  Muir  and  Mar- 
tin that  bacteriotropins  might  be  identified  with  agglutinins,  inas- 
much as  they  possessed  resistance  to  heat,  were  active  without  ap- 
parent dependence  upon  alexin,  and  could  not,  at  least  in  the  earlier 
studies,  be  reactivated  by  the  addition  of  fresh  normal  serum  when 
once  inactivated.  The  supposition  was  that  for  this  reason  the  bac- 
teriotropin  might  have  a  structure  like  the  hypothetical  "haptines 
of  the  second  order77  which  Ehrlich  attributes  to  the  agglutinins. 
This  supposition  has  found  no  experimental  support  in  that  ag- 
glutination and  bacteriotropic  effects  did  not  run  parallel.  We  our- 
selves are  not  ready  to  admit  that  such  lack  of  parallelism  is  proof 
against  their  identity.  However,  since  it  is  very  probable  that  both 
agglutination  and  precipitation  are  merely  phenomena  of  colloidal 
flocculation  effects  which  follow  certain  quantitatively  adjusted  com- 
binations of  antigen  and  specific  antibody,  and  that  it  is  not  at  all 
necessary  to  assume  separate  agglutinating  or  precipitating  serum 
constituents,  this  problem  becomes  merely  another  version  of  the 
question  of  the  identity  of  bacteriotropins  and  sensitizer  or  ambo- 
ceptor. 

Apart  from  thermostability,  further  similarity  lies  in  the  fact 
that  bacteriotropins  are  strictly  specific  and  may  be  specifically  ab- 
sorbed out  of  immune  sera  by  their  respective  bacteria. 

Like  amboceptor  or  sensitizer  they  are  specifically  increased  to  a 
powerful  degree  by  the  treatment  of  animals  with  any  given  micro- 
organism and  may  be  incited  not  only  by  the  injection  of  bacteria 

44  Neufeld  and  Rimpau.     Deutsche  med.  Woch.,  No.  40,  1904;  Zeitschr, 
f.  Hyg.,  51,   1905. 

45  Neufeld  and  Topfer.     Centralbl.  f.  Bakt.,  1,  38,  1905. 

46  Dean.     Proc.   Roy.   Soc.   B.}   76,   1905. 

47  Hektoen.     Jour.  Inf.  Dis.,  3,  1906,  and  loc.  cit. 

48  Neufeld  arid  Hiine.    Arb.  a.  d.  kais.  Gesundh.  Amt.,  Vol.  25,  1907. 


INFECTION    AND    RESISTANCE 

but  by  that  of  blood  cells  as  well.  In  spite  of  these  points  of  likeness, 
however,  Neufeld  49  and  his  associates  maintain  rigidly  that  the  two 
substances  are  not  the  same  and  that  the  bacteriotropins  are  distinct 
and  independent  antibodies. 

Among  the  reasons  advanced  in  support  of  this  opinion  are  the 
facts  that  certain  immune  sera,  both  antibacterial  and  hemolytic, 
may  contain  bacteriotropins  without  containing  lysins  and  vice  versa. 
That  this  is  undoubtedly  true  has  been  shown  not  only  by  ISTeufeld 
and  his  associates  but  by  Hektoen  50  and  others,  and  it  is  likewise  a 
fact  that  in  sera  in  which  both  functions  are  demonstrable  they  fre- 
quently do  not  run  quantitatively  parallel.  These  are  unquestionably 
strong  arguments,  but  their  force  is  somewhat  weakened,  as  Levaditi 
has  pointed  out,  by  the  fact  that  there  are  many  varieties  of  bacterial 
immune  sera  which  undoubtedly  sensitize  the  specific  bacteria  (as 
can  be  shown  by  alexin  fixation),  but  which  do  not  lead  to  bacterio- 
lysis. Wassermann  51  also  attaches  little  value  to  the  lack  of  parallel- 
ism between  the  lytic  and  opsonic  functions,  expressing  the  belief 
that  the  solubility  of  the  particular  antigen  may  determine  whether 
sensitization  leads  to  phagocytosis  or  to  lysis.  With  bacteria  like 
the  cholera  spirillum  rapid  lysis  takes  place,  but  when,  as  in  pneumo- 
cocci  or  streptococci,  there  is  great  resistance  to  lysis,  sensitization 
may  lead  to  delayed  lysis  anticipated  by  leukocytic  accumulation, 
phagocytosis,  and  intracellular  digestion. 

It  by  no  means  follows  from  mere  lack  of  parallelism,  therefore, 
that  the  two  serum  functions  are  dependent  upon  separate  antibodies, 
although  the  argument  is  sufficiently  strong  to  impose  conservatism 
in  identifying  them. 

Another  important  argument  advanced  against  the  identification 
of  bacteriotropins  with  the  bactericidal  sensitizers  or  amboceptors  is 
the  fact  that  the  former  lead  to  phagocytosis  without  the  participa- 
tion of  alexin,  whereas  the  latter  become  active  for  lysis  only  when 
alexin  is  present. 

This  point  also  has  constituted  Neufeld's  strongest  support  for 
maintaining  that  the  bacteriotropins  or  immune  opsonins  are  entirely 
distinct  from  the  normal  opsonins.  It  is  true,  indeed,  that  immune 
serum,  unlike  normal  serum,  may  opsonize  powerfully  even  after 
heating  to  temperatures  which  destroy  alexin. 

If  we  regard  the  heat-stable  lytic  antibody  as  an  amboceptor  in 
the  strict  sense  of  Ehrlich,  as  a  specific  "Zwischenkorper"  with  a 
complementophile  group,  this  argument  would  have  considerable 
weight.  Even  in  this  case,  however,  strong  sensitization  of  the  bac- 
teria may  make  them  amenable  to  the  living  cells — the  phagocytes — 

4&Neufeld  and  Topfer.     Centralbl.  f.  Bakt.,  1,  38,  1905. 

50  Hektoen.     Jour,  of  Inf.  Dis.,  6,  1909. 

51  Wassermann.    Deutsche  med.  Woch.,  Vol.  33,  No.  47, 1907. 


FACTORS    DETERMINING    PHAGOCYTOSIS        323 

which  in  itself  may  furnish  a  slight  amount  of  alexin  or  alexin-like 
substances. 

We  may  regard  the  action  of  the  immune  serum  upon  the  antigen 
as  rather  a  sensitization  in  the  sense  of  Bordet,  and  it  does  not  seem 
logical  to  assume  that  the  heat-stable  bodies,  similar  in  other  respects, 
are  different  merely  because  they  can  sensitize  bacteria  both  to  the 
action  of  an  alexin  and  to  that  of  a  living  cell,  which  in  itself  surely 
contains  a  number  of  different  enzymes,  comparable  functionally  to 
alexin,  though  possibly  not  identical  with  it. 

Indeed,  the  experiments  of  Dean  have  given  much  positive  evi- 
dence in  favor  of  regarding  the  immune  opsonins  or  bacteriotropins 
as  true  amboceptors  or  sensitizem  Dean 52  found  that,  although 
heated  immune  serum  may  unquestionably  opsonize  by  itself,  its 
action  may  be  still  further  enhanced  by  the  addition  of  a  little  diluted 
normal  serum  (compare  these  results  with  those  of  Cowie  and  Chapin 
on  normal  opsonins).  Hektoen's 53  experiments  with  hemopsonic 
immune  sera  are  analogous.  We  cite  one  of  these  as  illustrating  the 
point  in  question : 

TABLE   I 

Phagocytosis  of  Goat  Corpuscles  under  the  Influence  of  Goat-blood-immune  Rabbit 
Serum,  and  Normal  Guinea  Pig  Complement  (Table  from  Hektoen,  loc.  cit.) 


Immune  serum 

Normal  guinea  pig  serum 

Phagocytosis 

.001  4. 
.001  +  -01  20. 
.01  0 


Here,  therefore,  the  diluted  immune  serum,  but  slightly  cyto- 
tropic  in  itself,  was  powerfully  activated  by  a  diluted  unheated  nor- 
mal serum,  which  in  itself  was  entirely  inactive. 

Indeed,  an  experiment  by  Neufeld  himself,  with  Bickel,54  points 
in  the  same  direction.  They  found  that,  when  a  heated  specific  hemo- 
lytic  serum  was  added  to  the  homologous  cells  in  such  small  quanti- 
ties that  it  no  longer  exerted  cytotropic  (opsonic)  action,  the  addi- 
tion of  a  small  amount  of  alexin,  too  small  to  lead  to  hemolysis  of 
the  cells  (and  not  by  itself  cytotropic  or  hemopsonic),  caused  active 
phagocytosis.  Analogous  experiments  upon  bacterial  antisera  were 
carried  out  by  Levaditi  and  Inmann.  It  thus  appears  that,  even  in 
the  case  of  the  immune  opsonins  or  bacteriotropins,  we  may  think  of 
a  participation  of  two  substances — a  sensitizer-like  one  and  one  com- 
parable to  alexin  or  complement.  We  may,  at  least,  infer  that  the 
full  opsonic  action  both  of  normal  and  immune  sera  is  dependent 

52  Dean.     Proc.  Royal  Soc.  B.,  79,  1907. 

53  Hektoen.     Jour.  Inf.  Dis.,  Vol.  6,  1909,  p.  67. 

5-  Neufeld  and  Bickel.    Arb.  a.  d.  kais.  Gesundh.  Ami.,  Vol.  27,  1907. 


INFECTION    AND    RESISTANCE 

upon  the  cooperation  of  two  such  bodies.  It  is  likely,  therefore,  that 
the  mechanism  of  normal  and  of  immune  opsonic  action  may,  after 
all,  differ  only  in  quantitative  relations  between  the  two. 

For  assuming  this  to  be  an  antibody-alexin  mechanism  like  hemol- 
ysis,  we  may  recall  the  work  of  Morgenroth  and  Sachs  on  the  rela- 
tions between  amboceptor  and  complement  in  hemolysis.  There  we 
saw  that  a  large  amount  of  amboceptor  would  cause  hemolysis  in  the 
presence  of  a  small  amount  of  complement  and  vice  versa.  There- 
fore, here,  too,  in  normal  serum  the  small  quantity  of  amboceptor  or 
specific  thermostable  opsonin  (bacteriotropin)  may  act  very  power- 
fully in  the  presence  of  the  alexin.  When  the  latter  is  destroyed, 
however,  the  minute  quantity  of  specific  thermostable  opsonin  is 
hardly  enough  to  do  more  than  initiate  slight  phagocytosis  of  com- 
paratively non-resistant  bacteria,  whereas  the  large  amount  of  spe- 
cific sensitizer  left  in  immune  sera  after  inactivation  may  still  lead 
to  strong  bacteriotropic  action.  In  outlining  this  explanation  we 
have  consistently  drawn  upon  the  analogy  between  thermostable  op- 
sonin and  amboceptor  or  sensitizer.  Whether  or  not  these  two  sub- 
stances are  identical  is  by  no  means  positively  determined  and  must 
be  considered  an  open  question  for  the  present.  However,  from  the 
above,  it  seems  to  us  that  much  testifies  in  favor  of  such  an  identi- 
fication.55 

The  preceding  discussions  have  ignored  the  possibility  that  apart 
from  opsonic  or  bacteriotropic  action  on  the  bacteria  there  may  be 
a  difference  in  phagocytic  energy  which  depends  upon  inherent  prop- 
erties of  the  leukocyte  itself. 

Indeed,  the  technique  by  which  the  researches  of  Wright  and 
his  followers  were  carried  out  does  not  in  any  way  take  into  account 
the  source  of  the  leukocytes  as  a  possible  variable  factor.  There  is, 
however,  a  considerable  amount  of  evidence  which  points  to  differ- 
ences in  phagocytic  powers  residing  in  the  leukocytes  themselves 
independent  of  the  serum.  Park  and  Biggs 56  have  demonstrated 
such  differences  for  the  leukocytes  of  normal  persons  in  the  phagocy- 
tosis of  staphylococci,  and  more  extensive  researches  have  been  made 
with  similar  results,  in  the  case  both  of  staphylococci  and  tubercle 
bacilli  by  Glynn  and  Cox.57 

The  last-named  authors,  moreover,  recognized  the  necessity,  in 
making  such  investigations,  of  experimenting  with  leukocyte  emul- 
sions containing  approximately  the  same  number  of  cells,  for,  as 
JFleming  5^Piad  shown,  if  unequal  leukocytic  emulsions  are  used, 

55  Pfeiffer   (quoted  from  P.  Th.  Miiller)   regards  opsonic  action  as  due 
to  a  combined  action  of  amboceptor  and  complement  and  speaks  of  it  as  an 
"Andauung"  of  the  bacteria  for  the  leukocyte — which  we  may  translate  best 
as  a  partial  predigestion. 

56  Park  and  Biggs.     Jour.  Med.  Ees.,  Vol.  17,  1907. 

57  Glynn  and  Cox.    Jour.  Path,  and  Bact.,  14,  1910. 

58  Fleming.     Practitioner,  London,  Vol.  80,  1908. 


FACTORS    DETERMINING    PHAGOCYTOSIS         325 

less  phagocytosis  per  cell  occurs  in  the  emulsion  containing  the 
greater  number  of  leukocytes.  This  phase  of  the  subject  has  been 
taken  up  most  thoroughly  by  Hektoen  59  and  his  associates,  and  Rose- 
now  60  has  made  careful  comparative  studies  on  pneumococcus 
phagocytosis,  in  which  he  standardized  the  leukocytic  suspensions  by 
actual  cell  counts.  His  work  as  well  as  that  of  Tunnicliff,61  of  the 
same  school,  has  shown  definitely  that  the  inherent  phagocytic  power 
of  leukocytes  may  vary  not  only  in  health  and  disease,  but  differences 
may  exist  between  the  cells  of  apparently  normal  people.  Tunni- 
cliff showed,  for  instance,  that  at  birth  the  leukocytes  are  less  active 
than  in  adult  life. 

For  accurate  experimental  work,  therefore,  as  well  as  in  theoret- 
ical reasoning  upon  problems  of  phagocytosis,  it  is  necessary  to  bear 
in  mind  the  possible  inherent  variations  in  the  leukocytes  themselves. 

Of  the  three  factors  concerned  in  the  process  of  phagocytosis, 
then,  we  have  considered  two,  the  serum  and  the  leukocytes.  The 
former  we  have  seen  exerts  a  powerful  determinative  influence  on 
the  process,  the  latter  a  less  marked  influence,  though  still  definite 
and  measurable.  We  have  still  to  discuss  the  bacteria  themselves 
as  variable  factors  in  determining  the  degree  to  which  phagocytosis 
may  take  place. 

This  problem  was  first  investigated  by  Denys  and  Marchand  in 
connection  with  their  work  upon  streptococcus  immunity,  and  was 
further  studied  in  detail  by  Marchand.  Marchand  62  showed  that 
leukocytes  would  readily  take  up  non-virulent  streptococci  in  the 
presence  of  normal  serum,  but  that  under  similar  conditions  virulent 
streptococci  were  not  phagocyted  at  all  or  to  a  very  slight  degree 
only.  He  determined  further  that  this  resistance  to  phagocytosis 
remained  unchanged  after  the  virulent  organisms  had  been  killed 
by  heat,  and  washed  clean  of  culture  fluid.  It  seemed,  therefore, 
that  the  resistance  depended  upon  a  condition  of  the  bacterial  body 
and  not  upon  substances  secreted  and  given  off  to  the  environment. 
These  experiments,  as  well  as  similar  work  by  Mennes,63  Gruber  and 
Futaki,64  and  others  makes  it  clear  that  differences  in  virulence 
between  different  species  of  bacteria,  as  well  as  between  different 
strains  of  the  same  micro-organism,  depend,  at  least  in  part,  upon 
the  resistance  which  the  bacterial  bodies  oppose  to  ingestion  by  the 
leukocytes.  We  must  distinguish  clearly  here  between  these  appar- 
ently purely  "antiopsonic"  bacterial  properties  and  those  supposedly 
"antichemotactic"  substances  which  are  conceived  as  a  cause  for 

59  Hektoen.    Jour,  of  A.  M.  A.,  Vol.  57,  No.  20,  1911. 
60Rosenow.     Jour,   of  Inf.  Dis.,  7,  1910. 

61  Tunnicliff.     Jour.   Inf.  Dis.,  8,   1911. 

62  Marchand.     Arch,  de  Med.  Exp.,  No.  2,  1898. 

63  Mennes.     Zeitschr.  f.  Hyg.,  Vol.  25,  1897. 

64  Gruber  and  Futaki.     Munch,  med.  Woch.,  1906. 


INFECTION    AND    RESISTANCE 

virulence  by  Deutsch  and  Feistmantel  65  and  by  Bail  66  in  his  so- 
called  "aggressins."  The  latter  are  supposed  to  be  secreted  bacterial 
substances  by  means  of  which  the  leukocytes  are  held  at  bay.  The 
properties  we  are,  at  present,  considering  are  probably  in  no  way 
antichemotactic,  but  oppose  purely  the  actual  ingestion  by  the 
leukocyte,  nor  do  they  seem  to  depend  upon  the  secretion  of  sub- 
stances which  injure  the  leukocytes.  For,  in  the  first  place,  a 
profuse  accumulation  of  leukocytes  may  follow  the  injection  of 
virulent  micro-organisms,  and  Denys  (quoting  from  Gruber)  has  seen 
active  phagocytosis  of  virulent  pneumococci,  but  none  of  virulent 
streptococci  when  antipneumococcus  serum  was  injected  with  the 
mixture. 

Eosenow  67  has  carried  out  a  thorough  investigation  dealing  with 
these  relations  in  pneumococcus  infection.  Seventy-five  strains  of 
this  organism  were  all  found  non-phagocytable  when  first  isolated 
and  the  resistant  condition  was  associated  with  virulence  for  rabbits 
and  guinea  pigs.  It  was  found,  moreover,  that  the  resistance  to 
phagocytosis  was  dependent  upon  the  inability  to  absorb  opsonin. 
For,  while  phagocytable  non-virulent  pneumococci  absorbed  specific 
opsonin  from  serum,  the  virulent  ones  failed  to  do  this  in  proportion 
to  the  degree  of  their  virulence.  Furthermore,  extraction  of  the 
bodies  of  the  virulent  organisms  in  KaCl  solution  yielded  a  substance 
which  inhibited  the  action  of  pneumococcus  opsonin — a  true  anti- 
opsonin — which  he  speaks  of  as  "virulin."  This  discovery,  if  con- 
firmed, would  supply  us  with  a  very  simple  explanation  for  some 
phases  of  the  problem  of  virulence.  It  is,  indeed,  likely  that  the 
antiopsonic  property  is  closely  bound  up  with  chemical  and  struc- 
tural changes  which  take  place  in  the  bacterial  cell  as  it  adapts  itself 
to  the  parasitic  conditions.  This  is  plain  from  the  fact  that  pneumo- 
cocci and  some  other  bacteria  will  rapidly  lose  their  virulence  when 
cultivated  on  artificial  media  devoid  of  animal  serum,  will  retain  it 
longer  if  grown  on  some  serum  media,  and  will  rapidly  regain  it  if 
passed  through  animals.  The  formation  of  a  capsule  is  unquestion- 
ably a  morphological  evidence  of  such  a  change.  Habitually  capsu- 
lated  bacteria,  like  the  Friedlander  bacillus,  and  Streptococcus  muco- 
sus,  are  of  fairly  constant  virulence,  while  in  other  micro-organisms 
like  the  pneumococci,  anthrax  bacillus,  plague  bacillus,  and  certain 
other  streptococci,  the  formation  of  a  capsule  goes  hand  in  hand  with 
an  increase  of  virulence.  By  the  aid  of  this  morphological  earmark 
of  virulence,  moreover,  Gruber  and  Futaki  have  obtained  further 

65  Deutsch   and   Feistmantel.      Quoted   from    Sauerbeck.     Lubarsch   und 
Ostertag,  Vol.  2,  1906. 

66  Bail.     Arch.  f.  Hyg.,  Vol.  52,  1905. 

67  Rosenow.    Jour.  Inf.  Dis.,  Vol.  4,  1907. 


FACTORS    DETERMINING    PHAGOCYTOSIS         327 

proof  that  the  resistance  to  phagocytosis  in  these  cases  is  due  to  the 
nature  of  the  bacterial  cell  body  rather  than  to  any  secreted  anti- 
opsonic  substances.  For,  after  the  injection  of  anthrax  bacilli  into 
guinea  pigs,  they  saw  that  leukocytes  would  take  up  uncapsulated 
bacilli,  apparently  picking  them  out  of  the  midst  of  surrounding 
capsulated  organisms  which  they  were  unable  to  ingest. 


CHAPTER   XIV 

THE  OPSONIC  INDEX  AND  VACCINE  THERAPY 

WEIGHT'S  1  investigations  upon  phagocytosis  were,  indirectly, 
the  outcome  of  his  earlier  work  upon  antityphoid  vaccination.  His 
purpose  .in  these  studies  had  been  a  purely  practical  one,  and  he 
had  attempted  to  obtain  a  guide  for  the  dosage  and  the  interval  be- 
tween injections  by  measuring  the  bactericidal  and  agglutinating 
powers  of  the  blood  serum.  In  the  case  of  typhoid  immunization 
this  was  indeed  a  practicable  method  of  control,  since  the  bacteri- 
cidal power  of  the  blood  serum  rose  directly  as  the  immunization  of 
the  patient  was  attained.  In  the  cases  of  many  other  bacteria,  how- 
ever, this  method  of  study  was  not  practicable,  and  Wright,  as  others 
before  him,  did  not  find  a  regularly  increased  specific  bactericidal 
power  in  the  blood  sera  of  immunized  animals  or  of  patients  con- 
valescing from  infections  with  such  bacteria  as  the  staphylococcus, 
streptococcus,  Micrococcus  melitensis,  the  Bacillus  pestis,  and  a  num- 
ber of  others.  In  fact,  together  with  Windsor,2  he  showed  that  nor- 
mal human  blood  has  practically  no  bactericidal  power  for  pyogenic 
staphylococci  and  that  antistaphylococcus  inoculations  or  recovery 
from  an  infection  do  not  result  in  the  production  of  such  proper- 
ties in  the  serum.  These  determinations  are  practically  identical 
with  Nuttall's  3  earlier  studies  on  the  same  bacteria  and,  indeed,  cor- 
respond with  the  data  obtained  by  Metchnikoff  and  his  followers  in 
their  work  on  anthrax  infection.  For,  in  discussing  these  investi- 
gations, we  saw  that  very  often  the  serum  of  a  comparatively  resist- 
ant animal  is  less  potently  bactericidal  than  that  of  a  more  suscep- 
tible one.  We  need  only  recall  the  difference  between  rabbits  and 
dogs  in  this  respect.  The  serum  of  the  former  is  more  strongly  bac- 
tericidal than  that  of  the  latter,  and  yet  rabbits  are  the  far  more 
susceptible  animals.  These  relations  have  been  studied  with  great 
care,  also,  by  Petterson.4  It  was  logical  in  such  cases  to  look  for 
the  cause  of  resistance  in  the  activity  of  the  phagocytes,  and  this, 
we  have  seen,  Metchnikoff  did  successfully  in  a  large  series  of  cases, 
both  as  regards  natural  and  acquired  immunity. 

1  Wright.     Lancet,  1902;  Practitioner,  Vol.  72,  1904. 

2  Wright    and    Windsor.     Jour,    of   Hyg.,    Vol.    2,    1902 ;    and    Wright, 
Lancet,  1900  and  1901. 

3  Nuttall.    Zeitschr.  f.  Hyg.,  Vol.  4,  1888. 

4  Petterson.     Centralbl.  f.  Bakt.,  Vol.  39. 

328 


OPSONIC    INDEX    AND    VACCINE    THERAPY      329 

Yet  the  controversy  between  the  strictly  humoral  and  the  cellular 
schools  was  by  no  means  regarded  as  closed,  especially  since,  in  such 
cases  as  typhoid  infection,  the  parallelism  between  increased  resist- 
ance and  extracellular  bactericidal  power  was  so  plainly  evident, 
while  in  this  disease  particularly  (for  technical  reasons  which  will 
become  clear  as  we  proceed)  no  such  parallelism  with  phagocytosis 
could  at  first  be  shown.  It  was  because  of  such  apparent  confusion 
that  Leishmann  5  undertook  to  study  again  the  relation  of  phago- 
cytosis to  active  immunity,  chiefly  upon  staphylococcus  cases  that 
were  being  "vaccinated"  therapeutically  by  Wright  himself. 

In  order  to  obtain  a  numerical  measure  of  the  degree  of  phago- 
cytosis, he  developed  a  simple  technique  which,  though  crude,  served 
to  give  him  the  information  he  sought.  It  consisted  in  taking  small 
quantities  of  the  blood  of  patients  and  mixing  these  in  capillary 
pipettes  with  equal  volumes  of  bacteria  suspended  in  salt  solution. 

The  mixtures  were  then  placed  on  slides,  covered  with  a  cover- 
slip,  and  incubated  at  37°  C.  for  varying  periods.  At  the  end  of 
incubation  the  preparations  were  smeared  upon  slides  and  stained 
by  Leishmann's  modification  of  the  Romanowski  method,  the  num- 
ber of  bacteria  in  a  large  series  of  leukocytes  counted  and  an  average 
taken. 

This  method  had  many  serious  flaws,  chief  among  them  being 
the  liability  to  coagulation  of  the  preparations  and  the  fact  that,  in 
each  test,  the  fluid  constituents  as  well  as  phagocytes,  both  of  them 
variable  factors,  came  from  -the  same  individual.  While,  therefore, 
it  was  possible  to  estimate  an  increase  or  decrease  of  general  phago- 
cytic  power,  it  was  impossible  to  analyze  this  in  reference  to  its  de- 
pendence either  upon  the  condition  of  the  cells,  on  the  one  hand,  or 
that  of  the  plasma  or  serum,  on  the  other.  Moreover  the  relation  of 
the  number  of  leukocytes  to  that  of  bacteria  in  individual  tests  neces- 
sarily differed,  and  this,  we  have  seen,  adds  a  variable  factor  which 
renders  it  impossible  to  compare  any  two  experiments  with  accuracy. 

In  spite  of  these  difficulties,  however,  Leishmann  succeeded  in 
establishing,  in  a  number  of  cases  of  staphylococcus  infection,  that 
an  increased  resistance  was  accompanied  by  an  increased  energy  of 
phagocytosis. 

Leishmann,  however,  went  no  further  than  this,  and  interpreted 
his  results  on  the  basis  of  the  "stimulin"  theory  of  Metchnikoff. 

The  subsequent  studies  of  Wright,  which  began  at  the  point  at 
which  Leishmann  stopped,  have  been  described  in  the  preceding  chap- 
ter and  had,  as  their  main  result,  we  have  seen,  the  discovery  of  the 
opsonins  and  the  final  confirmation  of  Denys'  conception  of  the  true 
mechanism  of  cooperation  between  serum  and  leukocytes  in  phago- 
cytosis. In  order  to  carry  out  these  studies  the  technique  of  Leish- 

5  Leishmann.  Br.  Med.  Jour.,  1,  1902 ;  Transact.  Lond.  Path.  Soc.,  Vol. 
56,  1905. 


330 


INFECTION    AND    RESISTANCE 


mann  was  quite  inadequate,  and  Wright's  first  task  was  to  modify  it 
in  such  a  way  that  reasonably  accurate  comparative  estimates  of 
phagocytosis  could  be  made. 

It  is  necessary  to  outline  Wright's  method  briefly  in  this  place 
in  order  that  we  may  consider  possible  sources  of  error  and  obtain 

a  clear  understanding  of  the  conclu- 
sions he  based  on  his  observations. 
Wright  recognized  that  the  deter- 
mination   of    the    degree   of   phago- 
cytosis,  induced  by  the  opsonin  of 
WR.GHT     CAPSULE    ro*     TAKING    any  given  serum  in  a  single  test,  is 

BLOOD  TO  OBTAIN  SERUM  FOR    by  itseli  oi  no  value,  since  the  actual 

OPSONIC  TESTS.  number  of  bacteria  taken  up  by  each 

leukocyte,    apart    from   the    opsoiiic 

contents  of  the  serum,  depends  also  upon  such  purely  technical  fac- 
tors as  the  concentration  of  the  bacterial  emulsion,  the  relative  num- 
ber of  leukocytes,  and  the  length  of  time  of  incubation.  Two 
individual  tests,  therefore,  carried  out  with  the  serum  of  the  same 
patient  at  the  same  or  at  different  times,  with  different  bacterial 
emulsions  or  leukocytes  in  each, 
would  give  variable  results,  even 
though  the  opsonin  contents  them- 
selves were  entirely  alike. 

In  order,  therefore,  to  obtain 
a,  relative  estimate  of  the  opsonic 
contents  of  any  serum  it  is  neces- 
sary to  compare  the  phagocytic  ac- 
tivity induced  by  this  serum  with 
the  similar  power  of  another  sup- 
posedly normal  serum,  both  tests 
being  carried  out,  under  exactly 
similar  conditions,  with  the  same 
bacterial  emulsion  and  the  same 
leukocytes.  The  average  number 
of  bacteria  found  in  each  leuko- 
cyte in  each  one  of  the  prepara- 
tions is  then  the  "phagocytic  in- 
dex." The  relation  of  the  phago- 
cytic index  of  the  unknown  serum  METHOD 
to  that  of  the  supposedly  normal 
serum  constitutes  what  Wright  has 
called  the  "opsonic  index." 

Instead  of  using  the  whole  blood  of  the  patient  Wright  takes  a 
small  amount  of  blood  in  glass  capsules,  allows  it  to  clot,  and  uses 
the  expressed  serum  in  his  test.  For  comparison  with  this  he  em- 
ploys a  "pool"  of  a  number  of  specimens  of  serum  from  supposedly 


OF  PRODUCING  AN  EVEN 
EMULSION  OF  BACTERIA  FOR  OP- 
SONIN DETERMINATION. 


OPSONIC    INDEX    AND    VACCINE    THERAPY      331 

normal  individuals.  By  the  use  of  such  a  serum  mixture  any  slight 
possible  variations  from  the  normal  in  any  one  of  the  sera  are  likely 
to  be  equalized,  and  a  closer  approach  to  a  normal  standard  is  at- 
tained. 

The  leukocytes  used  in  both  tests  are  the  same  and  taken,  as  a 
rule,  from  the  blood  of  the  worker  or  from  some  other  supposedly 
healthy  person.  They  are  obtained  by  taking  15  or  20  drops  of 
blood  from  the  finger  or  ear  into  5  to  10  c.  c.  of  sodium  citrate  solu- 
tion, in  which  the  blood  does  not  clot.  Brief  centrifugalization  throws 
down  the  blood  cells,  with  a  thin,  buffy  coat  of  leukocytes  on  top,  and 
these  are  gently  taken  off  with  a  pipette.  This  constitutes  the  leu- 
kocytic  cream  of  Wright's  experiments,  and  furnishes  a  uniform  leu- 


METHOD  OF  TAKING  UP  EQUAL  VOLUMES  OF  LEUKOCYTES,  BLOOD  SERUM  AND 
BACTERIAL  EMULSION  IN  WRIGHT'S  TECHNIQUE  FOR  OPSONIC-INDEX  DETER- 
MINATION. 


kocyte  factor  for  the  two  tests  which  are  to  be  compared.  The  bac- 
teria are  obtained  by  emulsifying  carefully  in  salt  solution.  It  is 
very  important  to  obtain  an  emulsion  free  from  clumps  and  neither 
too  thick  nor  too  thin,  a  result  which  can  be  secured  only  by  experi- 
ence. 

Equal  quantities  of  serum  (unknown  and  normal  "pool"  respec- 
tively) are  mixed  with  equal  quantities  of  the  bacterial  emulsion  and 
the  leukocytes  in  capillary  pipettes,  and  the  mixtures  are  incubated 
for  fifteen  to  thirty  minutes  under  exactly  similar  conditions.  At 
the  end  of  this  time  smears  are  made  upon  slides,  the  preparations 
stained,  and  the  numbers  of  bacteria  in  a  hundred  or  .more  leuko- 
cytes counted  in  each  of  the  two  experiments.  The  average  is  taken, 
and  from  the  phagocytic  indices  thus  obtained  the  opsonic  index  is 
calculated.  For  instance,  if 

Phagocytic  index  (normal  pool)          =  8 
Phagocytic  index   (patient's  serum)  =  6 

then  the  opsonic  index  (patient's  serum)  =  0.75.  Or,  if  the  phago- 
cytic index  of  the  normal  pool  had  been  10.  and  that  of  the  patient's 
serum  15.,  then  the  opsonic  index  of  the  patient's  serum,  higher  than 
normal,  would  be  1.5. 

For  the  insurance  of  accuracy  in  carrying  out  this  method  Wright 
calls  especial  attention  to  the  caliber  of  the  capillary  pipettes  that 
are  used,  the  concentration  of  the  sodium  citrate  solution,  which 
should  be  1.5  per  cent.,  and  the  freshness  of  the  leukocytes.  But 
it  is  still  necessary  to  remember  that  with  the  greatest  care  in  tech- 


INFECTION    AND    RESISTANCE 


nique  uncontrollable  sources  of  error  influence  this  method.  Most 
important  among  them  are  the  differences  necessarily  existing  be- 
tween different  normal  sera  used  for  comparison  and  differences 
in  the  agglutinative  powers  of  the  sera  used  in  the  two  specimens. 
For  it  is  plain  that  different  degrees  of  agglutination  may  bring 
about  great  variations  in  the  number  of  bacteria  with  which  the 
individual  leukocyte  comes  into  contact. 

Wright's  method  has   also  been  particularly  unsatisfactory  in 
taking  the  opsonic  index  against  such  bacteria  as  the  typhoid  bacillus 

and  the  cholera  spirillum,  or- 
ganisms which  are  very  rap- 
idly digested  after  being 
taken  up  by  the  leukocytes. 
In  consequence,  even  after  as 
short  an  incubation  time  as 
five  or  ten  minutes,  the  in- 
gested bacteria  are  partly  dis- 
integrated, are  stained  in- 
distinctly, and  cannot  be 
counted  with  accuracy.  In 
order  to  avoid  this  source  of 
error  Klien 6  has  devised  a 
modification  which  depends 
upon  gradual  dilution  of  the 
serum  in  a  series  of  pha- 
gocytic  tests  with  the 
LEUKOCYTES  CONTAINING  BACTERIA.  DRAW-  same  leukocytic  and  bacterial 
ING  OF  FIELD  AS  SEEN  IN  WRIGHT'S  mrml«irm«  Tn  tliic  wav  IIP 
METHOD  OF  OPSONIC-INDEX  ESTIMATION,  emulsions.  In  tllia  way  ne 

determines  the  degree  of  di- 
lution   of   the    serum   to   be 

tested  at  which  phagocytosis  no  longer  exceeds  that  taking  place  in 
salt  solution  alone.  The  degree  of  dilution  at  which  this  result  was 
obtained  has  been  called  by  Simon  the  "coefficient  of  extinction."  A 
comparison  of  sera  with  regard  to  this  value,  it  is  clear,  furnished  an 
estimate  of  their  quantitative  opsonic  properties  quite  as  instructive 
as  the  direct  estimations  by  the  Wright  method,  and  in  our  opinion, 
at  least,  more  reliable.  Though  also  subject  to  some  of  the  objections 
advanced  against  the  Wright  method,  it  has  the  definite  advantages 
mentioned  above,  and  is  not  so  closely  dependent  upon  irregularities 
in  counting,  agglutinin  influences,  and  differences  in  relative  pro- 
portions of  bacteria  and  leukocytes  employed.  Jobling 7  has  used 
this  method  with  success  for  the  standardization  of  antimeningitis 
serum. 

A  further  modification  suggested  by  Simon,  Lamar,   and  Bis* 

6  Klien.     Johns  Hop.  Hosp.  Bull.,  Vol.  18,  1907. 

7  Jobling.    "Studies  from  the  Rockefeller  Inst.,"  Vol.  10,  1910,  p.  614. 


OPSONIC    INDEX    AND    VACCINE    THERAPY      333 

pham  8  depends  upon  a  combination  of  the  dilution  method  and  a 
modification  in  the  method  of  counting.  They  make  comparative 
tests  of  the  same  serum,  diluted  from  1  to  ten  to  1  to  one  hundred  in 
salt  solution,  and  estimate  the  opsonic  power,  not  by  determining 
the  average  number  of  bacteria  to  the  leukocyte,  but  by  taking  a  per- 
centage of  the  total  number  of  leukocytes  which  take  part  in  the 
phagocytosis,  that  is,  contain  any  leukocytes  at  all.  The  bacterial 
emulsion  for  this  method  should  be  so  thin  that,  in  normal  serum, 
only  about  50  per  cent,  of  the  leukocytes  will  contain  bacteria. 

That  Wright's  method,  or  any  of  the  others,  gives  absolutely 
accurate  results  will  hardly  be  claimed  by  any  one  who  has  worked 
upon  opsonic-index  estimations.  There  are  certain  uncontrollable 
variable  factors,  some  of  which  have  been  pointed  out  above;  and, 
apart  from  these,  the  delicacy  of  the  technique  is  such  that  reliable 
results  can  ordinarily  be  'obtained  only  by  trained  workers  after  con- 
siderable practice  and  experience.  Even  in  such  hands  the  per- 
centage of  personal  error  is  more  likely  to  be  above  than  below  10 
per  cent.  For  ordinary  clinical  purposes,  therefore,  in  the  control 
of  cases  the  estimation  of  the  opsonic  index  is  not  often  practicable. 

On  the  other  hand,  there  can  be  little  doubt  about  the  fact  that 
careful  comparative  estimation,  by  Wright's  method  and  by  some 
of  the  modifications,  carried  out  by  workers  with  experimental  train- 
ing and  consequent  attention  to  extensive  controls,  have  yielded 
results  of  sufficient  accuracy  to  permit  the  recognition  of  definite 
facts  concerning  opsonins.  It  is  beyond  question,  therefore,  that  the 
conclusion  regarding  the  relation  of  opsonic  fluctuations  to  clinical 
conditions  and  the  general  significance  of  opsonins  emanating  from 
laboratories  like  those  of  Wright,  E~eufeld,  Hektoen,  and  some  others 
may  be  accepted  as  fact — especially  since  in  most  essentials  such 
workers  have  agreed.  In  consequence  we  are  now  in  possession  of 
knowledge  regarding  the  opsonic  constituents  of  the  blood  in  health 
and  disease,  and  in  the  course  of  active  immunization  with  bacterial 
vaccines,  which  is  of  the  greatest  practical  importance.  We  may 
summarize  the  results  of  such  investigations  by  saying  that  in  many 
of  the  infections  of  man  the  resistance  of  the  patient  is  roughly  pro- 
portionate to  the  opsonic  index — and  that  properly  spaced  inocula- 
tion with  suitable  quantities  of  dead  bacteria  (vaccines)  will  raise 
the  opsonic  index  and  lead  to  recovery  in  many  of  the  localized 
subacute  and  chronic  conditions. 

As  to  the  usefulness  of  the  treatment  in  various  infections  and 
the  limitations  within  which  we  may  hope  for  results  opinions  differ, 
and  these  will  be  discussed  more  fully  below.  Before  we  proceed 
to  this,  however,  it  will  be  useful  to  consider  the  studies  upon  which 

8  Simon  and  Lamar.  Johns  Hop.  Hosp.  Bull,  Vol.  17,  1906;  Simon, 
Lamar,  and  Bispham,  Jour.  Exp.  Med.,  Vol.  8,  1906 ;  Simon,  Jour.  A.  M.  A., 
Vol.  48,  1907,  p.  139. 


334 


INFECTION    AND    RESISTANCE 


the  parallelism  between  opsonic  index  and  clinical  condition  was 
founded. 

Wright's  own  earlier  studies  were  made  chiefly  upon  staphylococ- 
cus  infections  and  tuberculosis.  Since  then  the  method  has  been 
applied  to  almost  all  known  infections  with  varyingly  successful 
results. 

One  of  the  first  steps  in  determining  such  a  parallelism  between 
the  resistance  of  a  patient  and  the  opsonic  index  consisted,  of  course, 
in  comparing  the  index  of  the  sera  of  normal  individuals  with  that 
of  patients  suffering  from  infection.  Wright  and  Douglas  did  this 
in  a  large  series  of  studies.  In  the  case  of  staphylococcus  infections 
the  following  experiment  will  illustrate  their  results: 


TABLE   I 


(Wright  and  Douglas,    Proc.  Royal  Soc.,  Vol.  74,  1904.) 
Showing    the    ratio    in  which  the  phagocytic  or  opsonic  power  of  the 
patient's  blood  stood  in  each  case  to  the  phagocytic   or  opsonic  power  of  the 
normal  individual  who  furnished  the  control  blood.     (The  phagocytic  power 
of  the  control  blood  is  taken  in  each  case  as  unity.) 


Initials  of  Patient 

Form  of  Staphylococcus  Invasion 

Opsonic  Index 

E.  E. 

Furunculosis  

0   48 

F.  F. 

Sycosis 

0  49 

J.  E. 

Acne  

0  64 

J.  H. 

Furunculosis  

0  87 

W.  B. 

Acne    

0  55 

E.  H. 

Acne  

0  82 

W.  H. 

Furunculosis                .    .    . 

0  79 

R.  G. 

Furunculosis  

0  7 

G.  L 

Acne  and  sycosis 

0  74 

S.  C. 

Furunculosis  . 

0  87 

W.  L. 

Furunculosis  

0  88 

W.  P. 

Furunculosis 

0  39 

S.  F. 

Very  aggravated  sycosis  

0  1 

E.  F.  D. 

Acne  

0.73 

D.  C. 

Sycosis                             

0  8 

J.  M. 

Acne      

0.48 

W.  M. 

Sycosis  

0.37 

E.  P. 

Acne                                  

0  6 

M.  S 

Pustular  affection  of  lips 

0  6 

F.  V. 

Repeated  sta/ph  infection         .         .    . 

0  47 

In  this  series,  as  in  others  investigated  by  Wright  and  his  col- 
laborators, staphylococcus  infection  was  uniformly  associated  with 
a  low  index.  He  concludes  that  there  is  probably  a  causative  rela- 
tion between  the  two  facts,  in  that  under  conditions  of  depressed 
phagocytic  powers  staphylococci  may  gain  a  foothold,  while  under 


OPSONIC    INDEX    AND    VACCINE    THERAPY      33*5 

ordinary  normal  conditions  they  would  fall  prey  to  phagocytic  de- 
struction soon  after  entering  the  body. 

The  study  of  the  opsonic  index  during  the  treatment  of  such 
cases  with  dead  staphylococcus  cultures  (usually  with  the  organisms 
cultivated  from  the  patient's  own  lesions — "autogenous  vaccines") 
revealed  a  striking  coincidence  between  the  rise  of  the  opsonic  index 
and  improvement  in  the  clinical  conditions.  A  number  of  further 
interesting  and  practically  important  points  were  brought  out  by 
the  systematic  study  of  these  relations  which  may  be  illustrated  by 

5% 6  769/0//IZ/3/4/S/6/7/8/9202/&23&&Z& 


2.0 
/.8 

/S 

tz 

10 

$ 

8 
7 

.6  \ 

J 

r 

S 

$ 

s 

/ 

1 

c| 

0 

1 

1 

? 

^ 

«Q 

/ 

$ 

£ 
* 

^ 

J 

/ 

| 

v) 

1 

/ 

^ 

$ 

f\ 

i 

/ 

£ 

j 

\ 

^ 

/ 

0 
(V 

I 

/ 

A 

x 

^ 

/ 

f 

X* 

* 

y 

> 

y 

1 

^ 

^ 

V 

/ 

| 

o 

^ 

5 

CURVE  I. — RESULT  UPON  OPSONIC  INDEX  OF  VACCINE  TREATMENT  IN  Two  CASES  OF 

CHRONIC  STAPHYLOCOCCUS  FURUNCULOSIS. 

(After  Wright  and  Douglas,  Proc.  Eoyal  Soc.,  Vol.  74,  1904,  p.  156;  also  from 
" Studies  on  Immunity,"  p.  41.) 


reproducing  a  plan  of  the  opsonic  index  curves  constructed  from 
cases. 

The  curve  shown  above,  and  taken  from  a  paper  by  Wright  and 
Douglas,  illustrates  the  course  of  the  opsonic  fluctuations  in  the 
case  of  a  medical  student  who  had  suffered  for  four  years  from  boils. 

When  first  seen  the  opsonic  index  (1.  being  normal)  was  0.6, 
and  there  were  2  boils  on  the  neck.  For  3  days  after  this  there  was 
a  spontaneous  rise  in  the  index  accompanied  by  an  improvement  of 
the  lesions. 

On  the  third  day  2  billion  staphylococci  were  injected.  This 
was  followed  by  an  immediate  drop  of  the  phagocytic  power — (the 
negative  phase)  ;  together  with  this  a  new  boil  began  to  form.  Soon, 
however,  the  opsonic  power  began  again  to  rise,  this  time  consid- 
erably above  normal,  reaching  its  highest  point  on  the  8th  day,  when 


<Q 


5 

K 


'OO 


07A 


AO 


07, 


(h 


7|7/ 


It 


OCS 


o/ttw 


02 


Q 

I 

i 


D 


•a 


336 


OPSONIC    INDEX    AND    VACCINE    THERAPY      337 

it  again  began  to  diminish.  A  second  inoculation  on  the  12th  day 
was  followed  by  a  similar  preliminary  negative  phase,  then  a  steady 
and  rapid  positive  phase,  which  was  accompanied  by  cure. 

Another  curve — Curve  2  of  the  same  publication  (Wright  and 
Douglas,  Proc.  Royal  Soc.,  Vol.  74,  1904,  p\  156) — is  similar.  This 
case  suffered  from  severe  sycosis  (barber's  itch),  had  been  ill  for 
17  months,  and  had  been  unsuccessfully  treated  during  this  time  with 
antiseptics.  Staphylococci  were  isolated  from  a  hair  follicle,  and 
from  this  the  vaccine  was  made  which  was  used  in  the  treatment. 
Here  the  originally  low  opsonic  index  (0.8)  rose  after  the  first  in- 
jection without  a  preliminary  negative  phase — but  after  the  second 
treatment  a  sharp  fall  preceded  the  subsequent  rise.  Finally  a  sus- 
tained high  index  accompanied  complete  cure. 

The  rise  and  fall  of  the  opsonins  after  the  injection  of  bacteria 
is  entirely  analogous  to  the  similar  fluctuations  of  other  antibodies 
after  antigen  injections.  Measurements  of  this  kind  are  numerous 
in  the  literature.  Thus  Saloinonsen  and  Madsen,  measuring  the 
antitoxin  contents  of  the  blood  and  milk  of  a  mare  which  were  being 
immunized  by  injections  of  diphtheria  toxin,  obtained  the  following 
curve,  which  is  entirely  similar  in  essential  features  to  those  con- 
structed for  the  opsonic  index  by  Wright  and  Douglas : 


/90 
180 

Jj 

I 

a 

\ 

§ 

75 

1 

1 

i 

^ 

^ 

> 

-i 

M 

g 

i 

_^ 

160 
150 
140 
/SO 

- 

, 

' 

\ 

3 

VF 

50 

k 

| 

f 

y 

,\ 

3 

fit 

1 

^ 

J 

1 

j 

\\ 

y 

\ 

^ 

A, 

3 

LI 

UO 
110 
100 
90 
80 

r 

/" 

v 

2 

- 

III 

\ 

-•' 

-- 

i 

\ 

1 

/ 

j 

.., 

- 

V 

A 

/ 

1 

"-- 

i 

A 

'     J 

70 
60 
50 
4n 

«i 

i 

I—  •- 

/ 

X1A7 

/ro 

K/C 

Pot 

v£^ 

o/^ 

THE 

/w.L 

. 

1 

.... 

... 

4/V7 

/rox 

1C 

POM 

£*• 

Of 

THC. 

/*4>o 

% 


2  4  6  8  JO  /2/4I6  /6 

CURVE  DESCRIBING  QUANTITATIVE  MEASUREMENTS  OF  ANTITOXIN  IN  A  MARE  IN 

EESPONSE  TO  TOXIN  INJECTIONS. 

(Taken  from  article  by  Salomonsen  and  Madsen,  Ann.  de  I'Inst.  Pasteur,  Vol.  11, 

1897,  p.  319.) 

Results  having  the  same  general  significance  are  apparent  in  the 
measurements  made  upon  a  tetanus  toxin  goat  by  Ehrlich  and 
Brieger,9  and  in  the  observations  upon  the  fluctuations  of  bacteri- 

9  Ehrlich  and  Brieger.     Zeitschr.  f.  Hyg.,  Vol.  13. 


338 


INFECTION    AND    RESISTANCE 


'0  II 


£  13/4/5 


PROLONGATION  OF  THE  NEGATIVE  PHASE  DUE  TO 
Too  VIGOROUS  TREATMENT  WITH  TYPHOID 
VACCINE. 

(After  A.  E.  Wright,  Brit.  Med.  Journ.,  May 
9,  1903.  Also  from  "Studies  on  Im- 
munity," p.  179.) 


cidal  power  of  the  sera  of  patients  treated  with  typhoid  vaccines 
made  by  Wright10  himself.  Similar,  again,  are  the  various  ag- 
glutinin  curves  constructed  by  Jorgensen  and  Madsen  n  and  others. 
Apart  from  the  purely  theoretical  value  of  such  measurements, 
they  demonstrate  features  which  are  therapeutically  of  the  greatest 
importance.  They  show  that  in  all  processes  of  active  immunization 
the  injection  of  antigen  is  followed  almost  immediately  by  a  rapid 
decline  of  specific  antibodies  in  the  blood  serum.  This  "negative" 
phase,  as  it  is  called,  is  probably  due  to  a  neutralization  of  existing 

antibodies  and  lasts  for 
varying  periods,  which 
must,  of  course,  depend 
upon  complex  relations  be- 
tween the  degree  of  resist- 
ance (or  amount  of  anti- 
body constituents  of  the 
serum),  the  quantity  of 
antigen  injected,  and  the 
general  recuperative  pow- 
ers of  the  subject.  There- 
fore, without  some  control 
like  that  furnished  by  the 
measurement  of  opsonins 

or  other  antibodies  it  is  impossible  to  determine  whether  the  negative 
phase  has  ended  or  is  still  in  progress  unless  the  clinical  condition  is 
of  such  a  nature  or  location  that  degrees  of  improvement  or  exacerba- 
tion are  well  marked  and  easily  observed.  Even  then  clinical 
observation  alone  is  at  best  not  an  absolutely  reliable  guide. 

The  practical  importance  of  the  question  lies  in  the  harm  which 
may  accrue  to  the  patient  if  a  second  injection  is  practiced  before 
the  cessation  of  the  negative  phase.  Wright  himself  accentuates 
this  danger  by  expressing  the  opinion  that,  in  typhoid  inoculations, 
an  excessive  dose  administered  to  a  patient  in  the  physiological  con- 
dition of  the  negative  phase  may  be  followed  by  a  prolongation  of 
this  phase  into  a  period  of  several  months. 

In  the  case  of  successive  inoculations,  as  in  vaccine  treatment,  a 
too  rapid  repetition — i.  e.,  a  repetition  of  injection  during  such  a 
period  of  depression — leads  to  what  Wright  speaks  of  as  a  "summa- 
tion of  the  negative  phase,"  which  obviously  may  seriously  aggra- 
vate the  condition  of  the  case.  ' 

It  is  to  such  a  cumulation  of  the  negative  phase  that  Wright 
attributes  the  failures  attendant  upon  the  use  of  tuberculin  during 
the  early  days  after  its  introduction,  since  injections  at  this  time 

10  Wright.     Practitioner,  Vol.  72,  1904,  p.  118. 

11  Jorgensen    and    Madsen.      Festschrift.    Serum    Institut.    Kopenhagen, 
1902. 


OPSONIC    INDEX    AND    VACCINE    THERAPY      339 

were  carried  out  without  any  control  of  serum  reactions  in  the 
patient  and  with  comparatively  large  doses. 

The  danger  to  be  carefully  avoided,  therefore,  is  a  too  rapid 
succession  of  inoculations  and  too  large  a  dosage,  since  both  of  these 
procedures  may  be  followed  by  cumulation  of  the  "ebb  tide  of  im- 
munity," and  great  harm  may  result.  On  the  other  hand,  if  the 
treatment  is  so  spaced  and  measured  that  the  successive  inoculations 
are  given  just  before  the  positive  phase  has  ended — in  other  words, 
just  before  the  apex  of  the  curve  is  reached — a  moderate  negative 
phase  may  be  then  followed  by  a  second  positive  phase  still  higher 
than  the  first,  and  corresponding  improvement  will  result.  It  is  even 
possible  to  occasionally  obtain  a  summation  or  cumulation  of  the 
positive  phase — in  which  the  negative  phase  will  be  entirely  sup- 
pressed. This  is  illustrated  in  the  following  curve,  in  the  case  of 
the  first  and  second  inoculation  indicated  on  the  chart.  This  case, 
too,  was  a  staphylococcus  infection  occurring  in  a  laboratory  at- 
tendant : 


STAPHYLO- 
OPSONIC     20, 


'6&Z8&X3/  I 


IS  l6l7t8'9&ZI 


STAPHYLOCOCCUS  INDEX  AS  DETERMINED  BY  WRIGHT  IN  A  CASE  OF  ACNE  TREATED 

WITH  STAPHYLOCOCCUS  VACCINES. 

Note  summation  of  positive  phase  after  third  injection.     (After  A.  E.  Wright, 
"Studies  on  Immunity,"  p.  348.) 


Such  a  summation  of  positive  phase,  though  of  course  the  ideal 
to  be  aimed  at,  cannot  be  produced  with  regularity,  however  carefully 
we  may  attempt  to  control  the  treatment.  It  is  worth  mentioning, 
moreover,  a  fact  which  should  become  evident  from  the  preceding 
and  is  too  often  overlooked,  that  a  summation  of  the  negative  phase 
can  certainly  be  attained  by  the  frequent  repetition  of  larger  doses. 
This  is  practiced  not  infrequently  in  the  false  hope  of  hastening  the 
acquisition  of  immunity,  and  does  harm  more  often  than  good. 

Ordinarily  the  opsonic  index  when  raised  to  a  level  considerably 
above  normal  will  gradually  recede  to  the  normal  or  even  to  a  sub- 


340 


INFECTION    AND    RESISTANCE 


normal  condition.  In  isolated  cases,  however,  especially  in  tubercu- 
losis, the  index  may  remain  high  for  periods  as  long  as  a  month. 
This  Wright  speaks  of  as  a  sustained  "high  tide"  of  immunity. 
These  laws  of  fluctuation  are  all  of  them  entirely  analogous  to  those 
long  well  known  in  the-  cases  of  other  antibodies,  for  even  in  diseases 
in  which  the  immunity  following  an  attack — (typhoid  fever,  cholera, 
plague,  and  others) — is  continued  through  life  the  antibodies  disap- 
pear from  the  blood  after  varying  periods  and  we  are  forced  to  seek 
the  cause  of  the  permanently  high  resistance,  not  in  the  circulating 
blood,  but  in  the  ultimate  physiological  units — the  cells  and  tissues. 

According  to  Wright  also,  the  treatment  with  vaccines  may 
be  either  reenforced  or  entirely  replaced  by  a  process  of  autoin- 
oculation  from  the  patient's  own  lesion  by  increasing  the  local  cir- 
culation, thereby  throwing  more  of  the  specific  antigen  into  the 
blood  stream. 

This  reasoning  has  been  applied,  not  only  to  the  treatment  of 
tuberculosis  and  other  conditions,  but  has  been  utilized  to  explain 
fluctuations  in  the  opsonic  indices  of  untreated  patients  under  the 
influence  of  unusual  motion  of  the  diseased  parts — as  in  walking  or 
other  exercise.  Wright's  meaning  is  well  illustrated  by  the  follow- 
ing curve  of  opsonins  in  a  case  of  gonorrheal  polyarthritis  in  which 
massage  of  the  joints  resulted  in  reactions  similar  to  those  ordinarily 
elicited  by  vaccine  injections: 


J/V/C 

rx 

40 

|Bi 

ji 

3.5 
iO 

25 
?0 

= 

I 

lj 

t 

4- 

/- 

/.f 

/<9 

^~ 

5! 

JL 

S 

I 

i; 

~v 

(95 

^N 

V 

Ez 

i  ~ 

^6 

27 

2fl 

29 

30 

3/ 

1 

2 

3 

4 

5 

6    ?• 

a  5 

/t7 

7 

12 

13 

/4 

/? 

/6 

/7 

/<9 

/P 

2f 

^/ 

22 

23 

f\2S\Z6Z 

7Zfl 

29 

30/ 

OPSONIC  CURVE  IN  A  CASE  OF  GONORRHEAL  ARTHRITIS  IN  WHICH  AUTO-INOCULA- 
TION BY  MASSAGE  WAS  PRACTICED. 

(After  Wright,  Douglas,  Freeman,  Wells  and  Fleming,  "Studies  on  Immunity,7' 

p.  373.) 


A  further  modification  of  the  vaccine  treatment  of  Wright  origi- 
nated in  the  observation  that  the  exudate  present  in  many  infected 
foci  is  often  very  much  less  rich  in  opsonins  than  is  the  blood  serum 
of  the  same  patient.  This  is  not  unlikely  to  be  due  to  an  absorption 
of  the  antibodies  by  the  bacteria — as  well  as  by  the  tissue  detritus  in 
the  lesion.  But  Wright  has  interpreted  it  as  a  purely  specific  ab- 


OPSONIC    INDEX    AND    VACCINE    THERAPY      341 

sorption  by  the  bacteria,  and  has  utilized  it  for  diagnostic  purposes. 
Thus,  with  Reid,12  he  has  examined  in  this  way  the  comparative 
amounts  of  tubercle  bacillus-opsonins  in  the  blood,  and  in  the  local 
exudates  (peritoneal  fluid)  in  cases  suspected  of  tuberculosis,  and 
has  determined  the  tuberculous  nature  of  the  condition  by  showing 
a  discrepancy  between  the  two.  These  results  have  not  been  uni- 
versally confirmed.13  But  therapeutically,  because  of  this  supposed 
lack  of  opsonin  in  the  fluid  of  lesions,  Wright  has  advised  the  in- 
crease of  the  local  flow  of  lymph  by  poulticing,  heat,  drainage,  Bier's 
cups,  X-rays,  Finsen  light,  and  other  means  of  accomplishing  this 
purpose. 

All  that  has  gone  before  (most  of  it  taken  directly  from  the 
staphylococcus  studies  of  Wright  and  his  immediate  followers)  has 
tended  to  show  a  very  close  correspondence  of  clinical  improvement 
with  the  increased  opsonin  contents  of  the  blood. 

As  applied  to  other  infections,  such  as  gonococcus  arthritis,  colon 
bacillus  cystitis,  localized  pneumococcus  lesions,  and  many  other 
conditions  of  a  localized  character,  observations  of  a  similar  general 
significance  have  been  made.  Such  reports  have  been  made,  apart 
from  the  Wright  school,  by  Emery,14  Potter,  Ditman,  and  Bradley,15 
Potter,16  Tunnicliff,17  Whitfield,18  Cole  and  Meakins,19  and  many 
others,  and  we  may  say  with  reasonable  accuracy  that,  in  localized 
infections  particularly,  there  is  much  evidence  to  show  that  clinical 
improvement  and  rise  of  the  opsonic  index  go  hand  in  hand. 

There  have  been  many  exceptions  to  this — which,  in  view  of  the 
complicated  factors  involved  in  immunization,  as  well  as  the  diffi- 
culty of  the  technique,  is  not  surprising. 

In  tuberculosis — in  which  many  of  Wright's  earlier  studies  were 
made — the  parallelism  has  not  been  so  consistent.  Thus  even  the 
early  work  of  Bullock  20  showed  that,  in  contrast  to  similar  staphy- 
lococcus investigations,  the  tuberculo-opsonic  indices  of  patients  may 
occasionally  be  higher  than  normal,  and  similar  observations  were 
made  by  Lawson  and  Stewart 21  in  cases  of  acute  pulmonary  tubercu- 
losis. 

Various  investigations,  too  numerous  to  be  reviewed  in  detail  in 

12  Wright  and  Reid.    Lancet,  1906 ;  Proc.  Royal  Society,  Vol.  77,  1906. 

13  Opie.     Assoc.  of  Am.  Phys.,  Washington,  1907. 

14  Emery.     "Immunity,   etc.,"  Lewis,,  London,   1909. 

15  Potter,  Ditman,  and  Bradley.    Journ.  A.  M.  A.,  Vol.  47,  1906,  p.  1722. 

16  Potter.    Jour,  of  A.  M.  A.,  Vol.  49,  1907,  p.  1815. 

17  Tunnicliff.     Jour,   of  Int.   Dis.,   Vols.   4   and   5,   1907   and  1908.  , 

18  Whitfield.     Practitioner,  May,   1908. 

19  Cole  and  Meakins.     Johns  Hop.  Hosp.  Bull,  Vol.  18,  1907. 

20  Bullock.     Transact,  of  Lond.  Path.  Soc.,  Vol.  56,  1905,  and  Lancet, 
1905,  Vol.  II,  p.  1603. 

21  Lawson  and  Stewart.    Lancet,  1905,  Vol.  II,  p.  1406. 


342  INFECTION    AND    RESISTANCE 

this  place,  indicate  in  a  general  way  that  localized  tuberculosis  of 
the  skin,  joints,  intestines,  and  glands,  with  the  patient  quiet  and  at 
rest,  is  apt  to  show  a  low  index,  while  a  high  index  may,  under 
such  conditions,  often  point  to  an  active  pulmonary  lesion.  Accord- 
ing to  Wright,  this  depends  upon  the  following  factors:  In  a 
localized  lesion,  with  the  body  at  rest — and  when  systemic  symptoms 
such  as  fever  are  absent — the  focus  is,  very  probably,  quiescent  and  in 
but  slight  communication  with  the  circulation,  even  though  it  may 
be  slowly  progressive.  In  such  cases  little  or  no  antigen  is  being 
discharged  and,  in  consequence,  no  antibody  formation  is  stimulated. 
Indeed,  even  the  small  amount  of  antibody  which  is  present  comes 
into  but  indifferent  contact  with  the  lesion  because  of  its  compara- 
tive insulation  from  the  body  fluids.  Such  a  lesion  may  be  benefited 
by  rest,  in  that  spreading  is  inhibited,  and  autointoxication,  with 
the  production  of  a  negative  phase,  prevented ;  but  it  cannot  be  com- 
pletely cured  unless  the  antibodies  are  increased.  This  can  be  ac- 
complished by  carefully  controlled  vaccinations  with  tuberculin.  At 
the  same  time  more  effective  contact  of  these  antibodies  with  the 
lesion  may  be  attained  by  local  applications,  X-ray,  etc.  Or,  again, 
the  same  purpose  may  be  accomplished  by  carefully  controlled  and 
graded  motion  or  massage  of  the  diseased  part — which  may  be  used 
both  to  increase  the  opsonin  contents  by  auto-inoculation  and  to  en- 
hance the  local  circulation.  If  this  is  done  with  care  it  may  serve  to 
substitute  entirely  for  the  treatment  with  vaccines. 

On  the  other  hand,  such  treatment  with  auto-inoculation,  it  must 
be  remembered,  is  entirely  uncontrollable  as  to  dosage,  and,  there- 
fore, not  to  be  generally  recommended.22  In  active  pulmonary  tu- 
berculosis, when  there  are  systemic  symptoms  such  as  rise  of  tem- 
perature, the  body  is  very  probably  already  receiving  excessive 
amounts  of  antigen  and  vaccine  treatment  of  any  kind  may  be 
dangerous. 

However  we  analyze  the  work  done  on  tuberculo-opsonins — and 
the  investigations  on  this  subject  are  far  too  numerous  to  be  here 
reviewed — -we  are  forced  to  the  conclusion  that  in  this  disease  the 
opsonic  fluctuations  are  far  more  irregular  than  in  most  other  condi- 
tions. Much,23  for  instance,  found  no  regular  differences  between 
the  tubercle  bacillus  opsonins  of  healthy  and  of  diseased  individuals, 
.and  Koehlisch  24  obtained  similar  results,  adding  the  important  ob- 
servation that  animals  that  show  a  high  natural  resistance  to  the 
liuman  type  of  the  tubercle  bacillus  invariably  show  an  opsonic 
index  much  lower  than  that  of  man. 

We  may  question  with  much  justice,  therefore,  whether  in  the 

22  Meakin  and  Wheeler.     Br.  Med.  Jour.,  2,  1905. 

23  Much.     Munch,  med.  Woch.,  p.  496,  1908. 

24  Koehlisch.    Zeitschr.  f.  Hyg.,  Vol.  68,  1911. 


OPSONIC    INDEX    AND    VACCINE    THERAPY      343 

case  of  this  bacillus  opsonic  investigations  can  be  looked  upon  as 
indicators  of  immunity  with  as  much  confidence  as  in  cases  of  other 
bacterial  invasions.  It  is  true,  indeed,  that  tubercle  bacilli — as  well 
as  leprosy,  rat  leprosy,  and  other  acid-fast  bacteria — are  eagerly 
taken  up  by  polynuclear  leukocytes  when  they  are  injected  into  the 
peritoneal  cavity  of  a  guinea  pig  or  rat  or  other  experimental  ani- 
mal. On  the  other  hand,  we  have  much  evidence  which  seems  to 
show  that  such  phagocytosis  is  not  in  these  cases  a  direct  method 
of  bacterial  destruction.  In  another  place  we  have  cited  the  experi- 
ments of  Tschernorutski,25  which  showed  that  polynuclear  leuko- 
cytes, though  containing  other  ferments,  were  devoid  of  lipase.  And 
Carey  and  the  writer — experimenting  with  rat  leprosy  bacilli — found 
that  these  acid-fast  bacteria  were  not  disintegrated  within  leukocytes 
in  the  course  of  weeks,  while  they  were  often  subject  to  rapid  de- 
struction in  the  presence  of  living  spleen  cells  in  plasma.  Further- 
more, in  the  discussion  of  the  tuberculin  tests  we  have  reviewed 
evidence  which  points  to  the  fact  that  in  the  reactions  to 
tubercle  bacilli  we  have  probably  to  deal  more  particularly  with 
sessile  receptors  on  fixed  tissue  cells  than  with  specific  circulating 
antibodies.  Bartel  and  Neumann  26  have  concluded  that  the  phago- 
cyte which  takes  up  tubercle  bacilli  represents  only  a  preliminary 
vehicle  by  which  the  micro-organisms  are  conveyed  to  the  spleen  and 
lymphatic  tissues,  in  which  actual  destruction  then  takes  place. 
While  no  final  conclusions  can  be  drawn  from  the  available  evidence, 
all  these  data  render  it  uncertain  whether  the  opsonic  index  as  de- 
termined for  polynuclear  phagocytosis  may  be  at  all  regarded  as  a 
reliable  indication  of  increased  or  diminished  resistance,  and  on 
this  basis  the  control  of  therapy  in  tuberculosis  by  opsonin  estima- 
tions is  of  course  placed  upon  an  uncertain  basis. 

We  have  then  very  briefly  traced  the  work  done  upon  opsonin 
determinations  from  the  purely  practical  point  of  view.  There  is 
of  course  no  question  about  the  scientific  accuracy  of  the  observa- 
tions upon  which  rests  our  knowledge  of  the  opsonic  properties  of 
blood  serum.  There  is  also  no  doubt  concerning  our  ability  to  in- 
crease the  immunity  of  an  individual  by  systematic  treatment  with 
vaccines  made  of  pure  cultures  of  bacteria.  However,  the  work  of 
Wright  has  concerned  itself  with  two  distinct  questions  which  must 
be  separately  answered.  Briefly  stated  these  are:  1.  What  is  the 
value  of  opsonic  estimations  in  controlling  the  therapeutic  vaccina- 
tions of  patients?  2.  To  what  degree  and  in  which  particular 
conditions  may  the  process  of  vaccination  (active  immunization)  be 
regarded  as  a  hopeful  method  of  therapy  ? 

25  Tschernorutski.     Hoppe-Seyler's   Zeitschr.    f.    Phys.    Chem.,   Vol.    75, 
1911. 

26  Bartel  and  Neumann.     Wien.  kl  Woch.}  Nos.  43  and  44,  1907;  Cen- 
tralbl  f.  Bakt.,  Vol.  48,  1909. 


344  INFECTION    AND    RESISTANCE 

The  first  question  has,  in  part,  been  answered  in  the  preceding 
paragraphs.  Reasonably  accurate  comparative  estimations  of  the 
opsonic  properties  of  serum  can  unquestionably  be  made  by  Wright's 
method,  or  some  of  its  accepted  modifications,  in  the  hands  of  trained 
workers  who  look  upon  each  estimation  as  an  experimental  problem 
and  have  time  for  control  and  repetition.  That  even  in  such  cases 
the  matter  is  difficult  is  amply  testified  to  by  such  reports  as  that  of 
E.  C.  Hort,27  who  states  that  two  of  the  most  skilled  experts  28  in 
London,  working  with  samples  of  the  same  serum  taken  before  and 
after  vaccination,  reported — "the  one  that  the  index  was  raised,  the 
other  that  it  was  lowered  by  the  treatment."  This,  and  similar  ex- 
periments of  other  observers,  do  not,  of  course,  invalidate  the  results 
obtained  in  special  researches  like  those  of  Wright,  Neufeld,  and 
others,  but  they  do  indicate  that  the  control  of  clinical  cases  by 
opsonic  estimations  is  not  a  matter  that  can  profitably  be  made  a 
routine  procedure  by  which  the  treatment  of  the  cases  can  be  regu- 
lated. As  a  problem  of  clinical  research  in  a  given  series  of  patients 
opsonin  studies  are  unquestionably  valuable  and  the  comparative  data 
so  obtained  have  proved,  and  will  continue  to  prove,  of  great  value. 
But  we  cannot  hope  as  yet,  it  seems  to  us,  to  utilize  this  method,  ex- 
cept in  cases  in  which  much  time  and  care  can  be  centered  upon  a 
few  patients  under  the  best  conditions.  Opinions  essentially  similar 
to  this  have  been  expressed  by  experienced  clinicians  (Potter,29  for 
instance),  who  have  followed  out  series  of  cases  on  which  systematic 
opsonin  determinations  were  made. 

As  to  the  opsonic  index  in  tuberculosis,  we  believe  that  the  ex- 
perimental evidence  at  present  available  does  not  show  that  such 
measurements  are  reliable  measures  of  resistance,  and,  in  this  dis- 
ease, even  when  the  index  is  taken  with  a  degree  of  care  which 
precludes  gross  error,  it  is  doubtful  whether  its  estimation  is  of  as- 
much  value  in  controlling  treatment  as  are  the  data  obtained  by 
skilled  clinical  observation. 

This  leaves  us,  therefore,  for  the  control  of  vaccine  treatment  in 
the  routine  work  of  the  clinic  only  the  information  gleaned  from 
such  indications  as  alterations  in  any  visible  or  palpable  lesions, 
general  systemic  symptoms,  temperature,  leukocytosis,  etc.  Since 
these  will  present  such  manifold  and  variable  pictures  in  different 
conditions,  generalization  is  useless. 

The  second  question  concerning  the  value  of  vaccine  treatment 
in  infectious  disease  of  human  beings  cannot  be  so  briefly  answered, 
and  is  one  of  the  greatest  importance  in  medicine.  It  is  well  known 

27  Hort.    Br.  Med.  Jour.,  Feb.,  1909,  p.  400. 

28  Quoted  from  Adami,  Trans.  Amer.  Phys.  &  Surg.,  Vol.  8,  1910.     See- 
also  Pearson,  Biometrica,  1911. 

29  Potter.     Loc.   cit. 


OPSONIC    INDEX    AND    VACCINE    THERAPY      345 

that  tuberculin  therapy  has  come  into  carefully  controlled  use  in 
recent  years  only,  although  it  was  introduced  early  in  the  history 
of  specific  therapy  by  Koch.  The  misuse  and  failure  of  this  treat- 
ment during  the  years  following  its  introduction  are  easily  explained 
by  the  defective  knowledge  of  antibody  reactions  and  the  general 
principles  of  immunity — a  condition  which  was  removed  only  by 
the  subsequent  assiduous  work  of  numerous  investigators.  At  the 
present  time  the  value  of  this  method  of  treatment  is  being  acknowl- 
edged, though  its  limitations  and  possible  dangers  are  properly 
recognized.  The  Wright  method  of  vaccine  treatment  is  also  an 
unquestionably  powerful  therapeutic  weapon,  and  yet,  owing  to  com- 
mercialization, unskilful  application,  and,  more  especially,  because 
of  extensive  attempts  to  apply  it  in  unsuitable  cases,  it  may  easily, 
like  tuberculin  therapy,  enter  into  a  period  of  neglect  and  disrepute. 
It  is  very  necessary  to  accentuate  at  the  present  time  that  the  active 
immunization  of  human  beings  with  any  form  of  bacterial  product 
is  a  serious  procedure  which  requires  painstaking  and  skilled  control, 
and  should  not  be  undertaken  without  the  same  degree  of  preliminary 
experience  and  study  which  is  considered  prerequisite  in  any  other 
branch  of  specialized  medicine. 

Any  opinion  expressed  regarding  the  ultimate  value  of  a  method 
of  treatment  which  is  still  undergoing  active  clinical  investigation 
must  of  course  be  purely  tentative.  Moreover,  there  are  so  many 
differences  of  judgment  that  we  wish  to  emphasize  the  purely  per- 
sonal character  of  the  views  expressed. 

In  passing  judgment  upon  the  value  of  active  immunization  in 
man  we  must  distinguish  sharply  between  active  immunization  which 
is  prophylactic  and  that  which  is  carried  out  after  the  disease  has 
gained  a  definite  foothold  in  the  body.  In  the  former  case  we  are 
dealing  with  a  new  method  and  with  one  upon  which  the  very  foun- 
dations of  our  knowledge  of  immunity  have  been  built.  It  is  the 
method  of  Jenner  in  small-pox.  It  is  that  of  Pasteur  in  chicken 
cholera,  in  anthrax,  and  in  many  other  infections.  It  has  been  used 
as  a  routine  in  animal  experimentation  in  laboratories  since  the  first 
days  of  the  systematic  study  of  infections.  There  is  no  question 
about  its  being  a  rational  and  logical  procedure.  The  immunity 
which  can  be  easily  conferred  upon  a  healthy  individual  in  this 
way  need  not  be  extensively  above  the  normal  in  order  to  protect 
from  invasion  by  the  small  numbers  of  pathogenic  germs  which 
may  gain  entrance  under  conditions  of  accidental,  spontaneous  in- 
fection. 

The  possibilities  of  the  method  were  recognized  by  Ferran,  a 
pupil  of  Pasteur,  who  applied  it  to  cholera,  and,  since  his  time,  it  has 
been  extensively  attempted  in  many  of  the  infectious  diseases  which 
occur  epidemically,  and  therefore  justify  attempts  in  this  direction. 


346  INFECTION    AND    RESISTANCE 

In  essence  also  Pasteur's  method  of  active  immunization  in  rabies 
represents  such  prophylactic  vaccination,  since,  in  this  case,  although 
treatment  is  begun  after  infection  has  taken  place,  nevertheless  the 
process  of  immunization  is  carried  out  during  the  incubation  period 
before  active  manifestations  of  the  disease  have  set  in.  Prophylactic 
vaccination,  therefore,  is  a  valuable  procedure  which  has  reaped 
remarkable  results  of  recent  years,  especially  in  protection  against 
typhoid  fever.  In  a  subsequent  section  this  phase  of  vaccination 
is  more  extensively  discussed,  and  we  may  therefore  leave  it  for  the 
present. 

In  this  place  we  are  more  particularly  concerned  with  the  prob- 
lem of  the  treatment  of  existing  disease  with  vaccines  prepared 
from  the  bacteria  by  which  the  disease  is  caused.  In  how  far  this  is 
justifiable  or  even  logical  is  a  question  which  depends  upon  the  con- 
ditions of  each  individual  case.  We  can  approach  the  problem  best 
by  roughly  classifying  the  various  forms  in  which  infection  occurs 
in  the  human  being. 

When  bacteria  gain  entrance  into  the  tissues  of  the  human  body, 
granted  that  the  organisms  are  pathogenic,  an  immediate  struggle 
ensues  between  the  offensive  properties  of  the  micro-organisms  and 
the  defensive  powers  of  the  tissues.  The  factors  which  determine 
the  outcome  of  such  a  combat  have  been  more  fully  considered  in 
Chapter  I.  Briefly,  if  the  defensive  powers  of  the  body  greatly 
preponderate  the  result  is  localization  and  rapid  destruction  of  the 
micro-organisms — with  cure.  In  such  a  case  any  form  of  treatment 
is  unnecessary.  On  the  other  hand,  the  balance  of  power  may  be 
turned  in  the  opposite  direction,  in  which  case  the  infectious  process 
becomes  rapidly  generalized,  the  bacteria  enter  the  blood  stream 
and  lymphatics,  and  the  defensive  powers  are  overwhelmed.  In, 
such  a  case  also  active  immunization  with  vaccines  is  entirely  use- 
less. 

There  are  cases,  however,  in  which  the  struggle  is  a  more  equal 
one,  and  in  which  the  infectious  process  is  held  in  check  by  the 
defenses,  so  that  it  takes  a  slow,  chronic,  localized  form,  and  spreads, 
if  at  all,  very  slowly.  What  is  it  in  such  a  case  that  prevents  com- 
plete healing  of  the  process  ?  The  answer  to  this  may  be  found  both 
in  local  and  in  systemic  causes.  Locally  the  lesion,  after  the  pre- 
liminary skirmishes,  may  become  encapsulated  either  by  fibrin 
formation,  clot,  or  other  tissue  changes  so  that,  as  Wright  suggests, 
the  fluid  constituents  of  the  blood-plasma  cannot  easily  approach  the 
organisms  in  the  lesion.  The  same  effect  may  result  from  internal 
pressure  by  fluid  and  possibly  by  the  presence  of  considerable  quan- 
tities of  tissue  detritus,  by  which  protective  serum  constituents  are 
fixed  and  thus  diverted  from  the  bacteria.  Against  these  factors,  of 
course,  no  form  of  immunization  can  be  of  value.  Wright  recog- 
nizes this,  and  suggests  the  use  of  surgical  evacuation,  Bier's  method, 


OPSONIC    INDEX    AND    VACCINE    THERAPY     347 

X-rays,  Finsen  light,  heat,  and  a  number  of  other  localized  methods 
of  increasing  the  blood  supply.  This,  too,  may  be  the  reason  for 
the  benefits  derived  from  wet  dressings,  in  that  they  keep  the  tissues 
macerated,  soft,  and  moist.  At  any  rate,  it  is  a  matter  of  local 
surgical  treatment.  At  the  same  time,  however,  there  may  be  sys- 
temic causes  which  prevent  the  complete  healing  of  such  lesions, 
namely,  an  insufficient  supply  of  circulating  antibodies,  opsonic  or 
bactericidal  substances.  These  may  be  sufficient  to  hold  the  lesion 
in  check,  but  since  small  quantities  of  bacteria  only  are  in  contact 
with  the  blood  stream,  relatively  small  amounts  of  antigen  are  ab- 
sorbed and  antibody  formation  is  consequently  deficient.  Here  we 
have  an  ideal  condition  for  vaccine  therapy.  By  isolation  of  the 
organisms  from  the  patient's  lesion,  for  which,  in  this  case,  there 
is  time,  and  the  careful  immunization  of  the  patient  with  these 
organisms,  the  immunity  may  be  considerably  increased  and  cure 
effected. 

Closely  related  to  this  type  of  lesion  are  those  conditions  in 
which  there  are  localized  infections  which  heal  rapidly  but  recur  in 
quick  succession  again  and  again.  Such  are  the  common  cases  of  con- 
secutive crops  of  boils ;  and  not  dissimilar  are  the  manifestations  of 
erysipelas  where  the  lesion  extends  along  the  edges  while  it  heals  in 
the  center.  There  is  in  this  type,  probably,  a  very  close  balance 
between  protection  and  offense ;  the  defensive  reaction  is  sufficient  to 
overcome  the  localized  lesion,  but  insufficient  to  set  up  a  permanent 
systemic  protection.  A  certain  amount  of  local  immunity  acquired 
by  the  tissues  of  the  affected  areas  may  suffice  to  throw  slight  weight 
into  the  balance  on  the  side  of  protection,  enough  at  least  to  decide 
the  struggle ;  and  this  element  of  locally  acquired  tissue  resistance  is 
in  all  probability  also  the  cause  for  the  failure  of  these  lesions  to 
recur  immediately  in  the  same  area.  Here,  too,  treatment  with  vac- 
cines is  not  illogical  and  may  yield  good  results  if  properly  carried 
out. 

In  generalized  systemic  infections  we  must  sharply  distinguish 
between  cases  of  acute  sepsis  in  which  the  bacteria  are  actively  grow- 
ing and  multiplying  in  the  circulation  and  cases  in  which  blood  cul- 
tures are  positive  only  because  the  bacteria  are  being  constantly 
discharged  into  the  circulation  from  a  focus  in  the  tissues.  In  the 
former  the  defenses  of  the  body  are  overwhelmed  by  an  extensive 
flooding  with  the  bacteria,  and  vaccines,  if  not  harmful,  are,  at  any 
rate,  utterly  useless  since  the  antigen  is  already  so  extensively  dis- 
tributed throughout  the  tissues  that  if  the  body  were  capable  of 
responding  with  sufficient  antibody  formation  this  would  unques- 
tionably occur  without  the  small  additional  amount  furnished  in  the 
bacterial  emulsion.  Vaccination  in  such  cases  is  entirely  analogous  to 
an  attempt  to  stimulate  a  degenerated  heart  muscle  with  strychnin — 
the  whipping  of  a  tired  mare. 


348  INFECTION    AND    RESISTANCE 

Such  cases  of  septicemia,  however,  are  not  in  our  opinion  the 
most  common  ones  in  the  human  being.  It  is  probable  that  all 
localized  infections  of  more  than  a  very  trifling  nature  discharge 
living  bacteria  into  the  circulation  from  the  very  beginning.  How- 
ever, in  most  cases  the  bacteria,  though  able  to  hold  their  own  in 
their  entrenched  position  at  the  focus  where  accumulated  offensive 
factors  and  local  injury  reenforce  them,  are  yet  rapidly  destroyed 
when,  in  small  detachments,  they  get  into  the  open  circulation  where 
the  plasma  antibodies  and  phagocytes  are  freely  active.  There  are 
cases  which  take  a  middle  course  between  such  purely  localized 
lesions  and  the  acute  septicemia,  conditions  in  which  a  well-estab- 
lished focus  continues  to  furnish  bacteria  to  the  blood  stream  as  fast 
as  they  are  destroyed.  An  example  which  illustrates  our  meaning 
well  is  that  of  the  so-called  subacute  endocarditis  caused  by  the 
Streptococcus  viridans  and  its  close  biological  kin,  where  blood  cul- 
tures are  often  consistently  positive  for  a  long  period  or  may  show 
occasional  intervals  in  which  the  blood  is  bacteria-free.  The  focus 
on  the  heart  valves  apparently  can  continue  uncured  in  spite  of  a 
relatively  high  or  at  least  normal  systemic  resistance  to  the  micro- 
organisms. If,  as  we  ourselves  have  done,  we  isolate  the  organisms 
by  blood  culture  from  such  cases,  and  then  measure  the  opsonic  prop- 
erties of  the  patient's  own  serum  against  them,  using  the  patient's 
own  leukocytes,  we  may  often  find  that  active  phagocytosis  takes 
place,  in  a  degree  equal  or  even  superior  to  that  taking  place  in  the 
serum  of  normal  individuals.  Neither  does  there  seem  to  be  a  dimin- 
ished phagocytic  power  of  the  patient's  own  leukocytes.  For  a  long 
time  these  conditions  may  continue,  with  a  constant  destruction  of 
bacteria  in  the  blood  and  a  corresponding  renewal  of  the  supply  from 
the  lesion.  The  same  condition  can  be  observed  in  rabbits  in  which 
chronic  endocarditis  with  persistently  positive  blood  culture  has  been 
produced  by  injections  of  these  bacteria.  In  such  animals  measure- 
ments similar  to  those  described  above  have  been  made  by  Miss  Gil- 
bert in  our  laboratory,  and  it  has  seemed  as  though  persistently 
positive  blood  cultures  could  be  obtained  only  when  a  localized  focus 
was  set  up  in  the  animals.  Unless  this  is  the  case  the  blood  cultures 
rapidly  become  negative. 

Conditions  essentially  similar  may  exist  in  any '  other  form  of 
severe  localized  infection.  Positive  blood  cultures  do  not  necessarily 
mean  a  multiplication  of  the  bacteria  in  the  blood  stream  and  a  rapid 
overwhelming  of  the  body.  We  have  had  occasion  to  see  a  number  of 
cases  of  bacteriemia  in  which  the  focus  of  infection  was  surgically 
accessible ;  and  in  some  of  these  cases  early  removal  of  the  focus  an( 
purely  surgical  treatment  resulted  in  a  clearing  up  of  the  infection 
Similar  experiences  have  been  reported  by  Libman  and  a  number  of 
others,  and  for  this  reason  general  septicemia,  if  not  fulminating, 
may  still  be  less  desperate  than  ordinarily  supposed. 


OPSONIC    INDEX    AND    VACCINE    THERAPY      349 

Now,  having  outlined  the  conditions  obtaining  in  such  cases,  let 
us  briefly  consider  whether,  under  the  circumstances,  vaccine  therapy 
may  logically  be  regarded  as  a  hopeful  form  of  treatment.  We  may 
assume,  on  the  one  hand,  that  the  bacteria,  being  consistently  present 
and  destroyed  in  the  blood,  should  furnish  antigen  sufficient  to 
stimulate  the  body  tissues  to  their  utmost  reactive  ability.  This 
would  seem  a  strong  argument  against  vaccine  therapy.  On  the 
other  hand,  we  must  take  into  consideration  another  phase  of  the 
subject,  one  which  has  some  experimental  justification.  In  discus- 
sing the  origin  of  antibodies  in  another  section  it  will  be  remembered 
that  we  called  attention  to  the  fact  that  many  different  tissue  cells 
probably  participate  in  the  production  of  these  protective  reaction- 
bodies.  We  cited  an  experiment  of  Wassermann  and  his  pupils  in 
which  they  proved  that  antibodies  were  produced  most  energetically 
in  the  tissues  about  the  point  of  injection  of  the  antigen,  namely,  in 
the  place  at  which  it  came  into  most  concentrated  contact  with  the 
cells.  They  injected  bacteria  into  the  subcutaneous  tissues  of  the  ear 
of  a  rabbit,  measured  the  progressively  increasing  appearance  of 
antibodies  in  the  blood  stream,  and  then  amputated  the  ear.  A 
sudden  drop  of  antibody  contents  followed,  showing  that  the  supply 
of  antibodies  had  largely  emanated  from  the  tissues  surrounding 
the  injection  point.  Park 30  has  pointed  out  another  reason  why 
vaccine  treatment  may  be  expected  to  exert  beneficial  action  in  such 
cases.  He  calls  attention  to  the  fact  that  when  very  large  amounts 
of  antitoxin  are  added  to  toxin  before  injection  no  antibody  produc- 
tion results,  and  assumes  that  in  chronic  or  subacute  general  infec- 
tions the  circulating  bacteria  are  in  contact  with  specific  antibodies, 
partially  "sensitized,"  and  therefore  not  efficient  as  antigen.  In 
consequence  the  injection  of  homologous  unsensitized  bacteria  may 
hasten  antibody  formation.  This  assumption  of  Park  is  theoretically 
valid,  but  it  is  not  in  accord  with  the  more  recent  experiments  of 
Metchnikoff  and  Besredka,  who  claim  to  have  obtained  the  best 
results  in  prophylactic  typhoid  vaccination  by  the  injection  of  sensi- 
tized bacteria. 

Thus  the  use  of  vaccines  in  the  subacute  or  chronic  cases  of  in- 
fection with  bacteria  in  the  blood  stream  may  be  theoretically  justi- 
fied, and  no  one  can  say  at  the  present  time  whether  or  not  it  has 
therapeutic  promise.  At  any  rate,  it  cannot  be  absolutely  condemned 
on  theoretical  grounds. 

Like  so  many  other  phases  of  this  question,  it  must  be  answered 
ultimately  by  clinical  experience,  for  in  experimentation  upon  ani- 
mals, while  it  is  easy  to  produce  a  purely  localized  lesion  followed 
"by  rapid  healing,  or  a  generalized  lesion  leading  to  rapid  death,  it 
is  not  easy  to  produce  prolonged  infections  with  anything  like  regu- 
larity, and  there  are  so  many  modifying  accidental  factors  which 

30  Park.     Trans,  of  Amer.  Phys.,  Vol.  8,  1910. 


350  INFECTION  AND  RESISTANCE 

influence  the  course  of  such  infections  in  animals  that  the  results 
of  vaccine  treatment  in  them  are  difficult  to  judge. 

In  acute  diseases  which  run  a  definite  course,  typhoid  fever, 
pneumonia,  dysentery,  cholera,  plague,  and  a  number  of  other  con- 
ditions, vaccine  treatment  during  the  course  of  the  disease  has  not 
much  theoretical  justification.  In  typhoid  fever,  especially,  specific 
antibodies  appear  in  the  blood  in  amounts  enormously  increased 
above  the  normal  at  periods  when  the  patient  is  still  actively  ill  in 
spite  of  the  fact  that  the  blood  stream  has  been  freed  of  the  micro- 
organisms. Whatever  may  be  our  opinion  as  to  the  continuance  of 
the  disease  after  bacteria  have  been  driven  out  of  the  blood  stream, 
the  use  of  vaccines  can  only  tend  to  further  increase  of  antibodies 
which  are  already  present  in  amounts  far  exceeding  normal.  In 
pneumonia  the  microorganisms  seem  curiously  resistant  against  the 
attack  of  the  serum  antibodies,  and  in  spite  of  the  presence  of  large 
amounts  of  antigen  both  in  the  lungs  and,  for  a  time,  in  the  circula- 
tion the  development  of  immunity  is  delayed  until  just  before  or  near 
the  crisis.  Since  this,  however,  is  usually  only  a  matter  of  7  or  8 
days,  it  is  hardly  likely  that  the  injection  of  vaccines  during  this 
period  could  markedly  alter  the  ultimate  outcome.  In  a  later  sec- 
tion we  shall  see  that  vaccine  treatment  in  typhoid  fever  is  neverthe- 
less being  extensively  tried  and  gives  reactions  of  a  nature  which 
cannot  entirely  be  explained  on  the  basis  of  the  above  considerations. 

As  we  have  said  before,  the  opinions  expressed  above  are  given, 
with  the  purpose  of  stating  as  clearly  as  we  can  the  logic  of  vaccine  32 
therapy  as  we  see  it  at  present.  The  next  ten  years  of  clinical  ex- 
perience may  largely  modify  these  views.  One  thing  is  certain,  how- 
ever, and  that  is  that  the  problem  can  only  be  settled  if  treatment  by 
this  method  is  undertaken  with  the  guidance  of  an  accurate  bacteri- 
ological diagnosis,  and  with  bacteriological  control  of  the  individual 
case,  so  that,  when  occasion  arises,  estimations  of  antibodies  can  be 
made. 

To  protest  against  the  random  use  of  commercial  stock  vaccines 
without  laboratory  diagnosis  and  without  control  is  almost  a  plati- 
tude. 

In  the  case  of  tuberculosis  the  problem  had  been  actively  investi- 
gated before  Wright,  and  there  seems  little  question  that  tuberculin 
therapy  properly  and  cautiously  applied  has  an  established  value  in 
the  treatment  of  initial  and  localized  tuberculous  disease.  Whether 

81  See  also  Theobald  Smith,  Jour.  A.  M.  A.,  Vol.  60,  1913,  and  R.  M. 
Pearce,  Jour.  A.  M.  A.}  Vol.  61,  1913. 

32  For  discussion  of  various  clinical  applications  of  vaccine  treatment 
see  symposium  on  vaccine  treatment,  Trans,  of  Ass'n  of  Amer.  Phys.  and 
Surg.,  Vol.  8,  1910. 


OPSONIC    INDEX    AND    VACCINE    THERAPY      351 

or  not  its  use  in  actively  progressive  tuberculosis  may  or  may  not  be 
hopeful,  in  which  particular  cases,  and  by  what  methods,  it  is  to  be 
applied,  these  are  problems  that  we  have  neither  the  space  to  deal 
with  nor  the  experience  to  summarize  properly.  They  constitute  a 
special  field  of  clinical  research,  a  survey  of  which  may  be  obtained 
in  such  works  as  that  of  Bandelier  and  Roepke,33  or  the  more  espe- 
cial experimental  studies  of  Denys.34 

THE  PRODUCTION  AND   STANDARDIZATION   OF  VACCINES 

Vaccines  in  the  sense  of  Wright  consist  merely  of  killed  cultures 
of  the  bacteria  with  which  the  patient  is  infected.  In  all  cases  it  is 
extremely  desirable  to  make  such  vaccines  "autogenous,"  by  whicK 
we  mean  that  the  organism  used  is  one  which  has  been  isolated  from 
the  case.  The  difference  between  various  strains  of  the  same  species 
of  bacteria  seems  to  make  this  imperative  whenever  it  is  at  all  pos- 
sible. The  recent  investigations  of  Neufeld  and  Haendel  in  de- 
termining that  there  are  a  number  of  types  of  pneumococcus  which 
are  antigenically  distinct  illustrates  this  point.  The  same  principle 
is  made  clear  by  the  recent  work  of  Rosenau  on  the  streptococcus- 
pneumococcus  group.  Especially  important  is  Rosenau's  observa- 
tion that  a  pneumococcus  which  he  had  been  able  to  transform  cul- 
turally by  special  methods  was  found  to  be  altered  also  in  its  reaction 
to  agglutinins. 

In  the  development  of  prophylactic  methods  of  vaccination 
against  epidemic  disease  like  typhoid,  cholera,  plague,  etc.,  many 
different  methods  of  antigen  preparation  have  been  developed.  In 
typhoid  prophylaxis  the  bacteria  have  been  used  dead,  living,  and 
sensitized,  and  even  extracts  have  been  employed.  In  cholera  the 
early  use  of  living  cultures  by  Ferran  has  given  way,  in  the  hands 
of  Kolle  and  others,  to  that  of  dead  bacterial  emulsions.  In  plague 
and  a  number  of  other  conditions  the  impression  seems  to  be  general 
that  the  bacteria  should  be  used  in  the  living,  but  attenuated,  state. 
Special  methods  which  have  been  developed  in  these  cases  are  dis- 
cussed in  another  section. 

In  treatment  of  developed  diseases  with  vaccines  the  method  most 
commonly  used  is  that  which  has  been  introduced  by  Wright,  namely, 
the  use  of  dead  cultures.  In  his  earlier  experiments  Wright  culti- 
vated the  bacteria  on  agar  slants  for  about  24  hours,  then  washed  off 
the  growth  with  10  c.  c.  of  sterile  salt  solution.  It  will  be  well  to 
describe  in  detail  the  preparation  of  such  a  vaccine. 

The  bacteria  must  be  isolated  from  the  patient  by  the  usual 
method  of  plate  cultivation  and  colony  fishing  on  suitable  media. 

33  Bandelier  and  Roepke.     "Lehrbuch  der  spez.    Diagnostik  und  Therapie 
der  Tuberkulose,"  6th  Ed.,  Kabitzsch,  Wiirzburg,  1911 

34  Denys.    "Le  Bouillon  Filtre,"  Louvain,  1903. 


INFECTION    AND    RESISTANCE 


We  do  not  think  that  any  satisfactory  substitute  for  careful  isolation 
by  plating  has  been  devised.  After  a  pure  culture  of  the  organism 
has  been  obtained  this  is  grown  on  relatively  large  surfaces  of  agar, 
glucose-agar,  or  ascitic  agar,  as  the  case  may  require.  These  culti- 
vations may  be  made  in  Kolle  flasks  or,  as  Wright  and 
others  have  suggested,  on  large  agar  surfaces  obtained 
when  the  culture  medium  is  allowed  to  harden  in  a 
square  3-oz.  medicine  bottle  laid  on  its  side.  Any  de- 
vice of  this  kind  in  which  a  large  surface  of  agar  is 
exposed  may  be  used. 

After  suitable  growth  of  the  micro-organisms  has 
taken  place,  24-48  hours,  the  growth  is  gently  washed 
off  with  10  c.  c.  or  more  of  sterilized  salt  solution. 
Care  must  be  taken  to  do  this  in  such  a  way  that  no 
agar  is  drawn  away  with  the  emulsion.  The  thick 
emulsion  so  obtained  is  removed  from  the  culture  bottle 
with  sterile  nipple  pipettes  or  Pasteur  pipettes  and 
transferred  to  a  sterile  thick-walled  test  tube  into  which 
glass  beads  have  been  placed.  By  drawing  out  the 
neck  of  this  test  tube  in  the  flame  a  glass  capsule  is 
formed,  in  which  we  now  have  our  so-called  stock 
emulsion.  (See  figure.) 

The  next  thing  to  be  done  is  to  standardize  this 
stock  emulsion,  or,  in  other  words,  determine  approxi- 
TO  HOLD  STOCK   mately  the  number  of  bacteria  to  the  cubic  centimeter. 

VACCINE    There  are  a  number  of  methods  by  which  this  can  be 
EMULSION     , 
FROM     WHICH    clone. 

D  i  L  u  T  i  ONS  The  method  most  extensively  used  by  Wright  and 
^Q  followers  was  that  in  which  the  bacteria  are  counted 
against  red  blood  cells.  The  bacteria  in  the  capsule 
are  snaken  thoroughly  with  glass  beads  so  that  clumps 
until  may  be  broken  up  and  even  distribution  obtained.  A 
cooled  little  of  the  emulsion  is  then  put  into  a  clean  watch 
it  may  crack  glass,  a  step  which  can  be  accomplished  most  easily  by 
when  quickly  breaking  off  the  tip  of  the  drawn-out  part  of  the  cap- 
sule, tilting  it  very  gently  and  heating  the  closed  end 
over  a  small  flame,  so  that  some  of  the  emulsion  will  be 
driven  out  by  the  expanding  air.  With  a  nipple  pipette  marked 
about  an  inch  from  the  tip,  as  in  the  taking  of  an  opsonic  index,  a 
little  of  the  emulsion  is  drawn  up.  This  is  placed  into  another  clean 
watch  glass  and  is  mixed  with  about  2  volumes  of  salt  solution  and 
one  volume  of  blood  from  the  finger,  these  quantities  being  measured 
with  the  same  nipple  pipette.  We  then  have  a  mixture  in  which, 
in  a  total  of  4  volumes,  there  are  equal  parts  of  blood  and  of  bac- 
terial emulsion.  After  this  emulsion  has  been  thoroughly  mixed  by 
drawing  in  and  out  through  the  nipple  pipette  smears  are  made  on 


CAPSULE    MADE 
OF  TEST  TUBE 


ARE  MADE. 
(It    is    well    to 


open  to 
lary  tip 
it  has 


OPSONIC    INDEX    AND    VACCINE    THERAPY      353 


slides  and  stained  with  Jenner  or  any  other  suitable  blood  and  bac- 
terial stain.  Under  the  field  of  the  microscope  the  ratio  between 
the  bacteria  and  blood  cells  is  then  determined,  and  from  our 
knowledge  of  the  number  of  the  red  blood  cells  in  this  blood  to  each 
c.  mm.  we  can  easily  calculate  the  number  of  bacteria  to  the  c.  mm. 
or  c.  c.35 

A  more  accurate  method  of  enumerating  the  bacteria  in  a  sus- 
pension to  be  used  for  vaccine  is  by  direct  count  of  an  accurately 
made  dilution  in  a  hemocytometer  chamber,  as  was  first  suggested 
by  Malory  and  Wright  in  1908.36    The  bacterial  suspension  is  diluted 
in   blood-counting   pipettes, 
1-20   to   1-100   dilutions  of 
thick  bacterial   suspensions 
being  as  a  rule  satisfactory. 
As  a  diluent  one  may  use 
either  salt  solution  or  some 
dilute  anilin  dye,  such  as  one 
made  by  mixing  one  part 
alcoholic     methylene     blue 
with  40  parts  of  1  per  cent, 
carbolic   acid.      The   dilute 
suspension  is  then  placed  in 
an     ordinary     Thoma-Zeiss 
chamber,     which    was     de- 
signed  for    counting   blood 
platelets  and  has  a  depth  of 
0.02  mm.     This  enables  one 
to  use  an  oil  immersion  lens 
or  high  power  dry  system 
with  a  short  working  dis- 
tance.     From  such   a  count  one  may  readily   estimate  the  num- 
ber  of  bacteria   in   the   original   suspension ;    for   example,    if   20 
squares  in  the  Helber-Zeiss  chamber  are  counted  the  result  gives  the 
number  of  bacteria  in  0.001  c.  mm.37 

Another  method  of  standardization  of  vaccines  wThich  is  suffi- 
ciently accurate  for  clinical  purposes  is  that  of  Hopkins,  which  con- 
sists in  measuring  the  volume  of  the  sediment  38  after  centrifugaliz- 
ing  the  preparation  under  standard  conditions  in  a  graduated  tube. 
The  tubes  may  be  made  with  a  capacity  of  10  to  15  c.  c.  with  a  capil- 

35  For  such  counts  it  is  convenient  to  contract  the  field  of  the  microscope 
by  using-  a  diaphragm  or  simply  marking  a  circle  on  the  eyepiece  with  a 
grease  pencil. 

36  Malory  and  Wright.     "Pathological  Technique,"  4th   Ed.,  New  York, 
1908. 

37  Glynn,  Powell,  Rees,  and  Cox.    Jour,  of  Path,  and  Bact.,  Vol.  18,  1914, 
p.  379. 

38  Hopkins.     Jour.  A.  M.  A.,  1913,  Vol.  60,  p.  1615. 


MICROSCOPIC  FIELD  AS  SEEN  IN  STANDARDIZA- 
TION OF  VACCINES  BY  WRIGHT'S  METHOD. 


354  INFECTION    AND    RESISTANCE 

lary  tip  about  one  inch  in  length,  having  a  capacity  of  about  0.05 
c.  c.  graduated  in  0.01  c.  c.  The  bacterial  suspension,  after  being  fil- 
tered through  sterile  cotton  to  remove  fragments  of  the  agar  or  other 
foreign  bodies,  is  centrifugalized  in  such  a  tube  for  half  an  hour  at 
about  2,800  revolutions  a  minute.  The  supernatant  fluid  and  bac- 
teria are  removed  down  to  the  0.5  c.  c.  mark  and  the  sediment  resus- 
pended  in  5  c.  c.  sterile  salt  solution  by  means  of  a  capillary  pipette 
which  gives  a  1  per  cent,  suspension.  0.05  c.  c.  of  streptococci  sedi- 
mented  in  this  way  represent  quite  constantly  16  mm.  of  dried  bac- 
terial substance.  The  number  of  organisms  per  cubic  centimeter 
contained  in  1  per  cent,  suspension  in  this  way  are  as  follows : 

Streptococcus  aureus  and  albus 10  billion 

Streptococcus 8      " 

Gonococcus 8      " 

Pneumococcus  (capsulated) 2.5   " 

Bacillus  typhosus 8      " 

Bacillus  coli. .  4      " 


After  the  vaccines  have  been  standardized  suitable  dilutions  can 
be  made  in  salt  solution  to  which  0.5  per  cent,  carbolic  acid  or  some 
other  antiseptic  has  been  added.  The  dilutions  are  usually  so  made 
that  from  100  to  500  million  bacteria  are  contained  in  the  cubic  cen- 
timeter, this  being  a  suitable  initial  dose  of  most  organisms.  The 
dilutions  are  placed  in  sterile  bottles  containing  beads  and  fitted 
with  rubber  caps.  These  bottles  can  be  shaken 
before  use,  the  emulsion  thoroughly  distributed, 
and  the  desired  quantity  can  be  taken  out  with  a 
sterile  hypodermic  syringe  thrust  through  the 
rubber  cap  after  this  has  been  covered  with  a 
small  amount  of  lysol  or  strong  carbolic  (see  fig- 
ure). After  the  dilutions  have  been  made  both 
these  and  the  stock  vaccines  should  be  sterilized. 
Some  workers  sterilize  always  the  stock  vaccines 
VACCINE  STOCK  and  make  the  dilutions  with  aseptic  proportions 
EMULSION  IN  j  ^  ^  t  further  sterilization  is 

EUBBEB     TOP  J.        .  £         ,,        , 

BOTTLE.  necessary.      This   is   preferable   because   the   less 

heat  that  is  applied  the  better  it  is  for  the 
preservation  of  their  antigenic  properties — sterilization  is  usually 
accomplished  by  heat  in  the  water  bath.  Wassermann's  earlier 
technique  called  for  heating  to  60°  C.  for  one  hour  for  a 
number  of  consecutive  days.  It  is  generally  considered  at  the 
present  time  that  it  is  better  not  to  heat  above  55°  C.  After  the 
vaccine  has  been  heated  its  sterility  must  be  controlled  to  aerobic 
and  anaerobic  cultivation,  and  possibly  by  animal  inoculation,  al- 
though, except  in  special  cases,  this  is  unnecessary.  Some  workers, 


OPSONIC    INDEX    AND    VACCINE    THERAPY      355 

especially  when  the  vaccine  is  to  be  extensively  used,  as  in  typhoid 
immunization,  inject  some  of  the  vaccine  into  white  mice  to  exclude 
the  possibility  of  contamination  with  tetanus.  In  such  cases  also  it 
is  not  inadvisable  to  test  out  the  antigenic  value  of  the  vaccine  upon 
animals,  measuring  the  agglutinins,  etc.,  which  result  from  a  number 
of  inoculations.  In  the  preparation  of  a  therapeutic  vaccine  where 
speed  is  required  this  of  course  is  not  feasible.  Moreover,  it  is  un- 
necessary in  view  of  the  fact  that  we  wish  to  inject  that  particular 
organism  into  the  patient  from  whom  it  has  been  cultivated.  What- 
ever its  antigenic  value  may  be  from  animal  experiments,  it  is  pre- 
ferable for  the  given  purpose  to  any  other  strain. 

Sensitized  vaccines  are  easily  made  by  exposing  emulsions  of  the 
bacteria  to  moderate  amounts  of  a  strong  immune  serum  which  has 
been  heated  to  56°  C.  to  destroy  the  complement.  Bacteria  will 
usually  agglutinate  under  these  circumstances  and  can  easily  be 
centrifugalized  to  the  bottom.  The  excess  serum  is  then  washed  off 
and  the  bacteria  emulsified  as  in  the  case  of  the  preparation  of  vac- 
cines with  dead  organisms. 


THE   TUBERCULINS 

Since  we  shall  not  attempt  to  discuss  critically  tuberculin  treat- 
ment, as  this  is  a  subject  upon  which  many  special  studies  have  been 
made  both  by  clinicians  and  by  laboratory  workers,  and  is  entirely 
too  extensive  to  be  reviewed  in  a  book  like  this,  on  the  other  hand, 
we  deem  it  a,  part  of  our  task  to  discuss  at  least  the  methods  by  which 
the  antigen  or  tuberculin  preparations  are  obtained.  There  has  been 
much  discussion  concerning  the  nature  of  the  antigenic  substances 
obtained  from  the  tubercle  bacillus.  It  has  been  claimed  by  Denys 
and  others,  for  instance,  that  the  tubercle  bacillus  may  give  rise  to 
small  quantities  of  a  true  exotoxin  with  consequent  endotoxin- 
inducing  properties.  Again,  most  observers  have  believed  that  the 
poison  of  the  tubercle  bacillus  consists  of  substances  comparable  to 
the  endotoxin  of  other  micro-organisms.  The  matter  is  by  no  means 
settled,  and  without  going  into  the  theoretical  aspects  of  the  problem 
we  will  confine  ourselves  in  this  place  to  a  description  of  the  pro- 
duction of  the  various  forms  of  so-called  "tuberculin." 

OLD  TUBERCULIN  (Keen) 

The  first  tuberculin  prepared  by  Koch  is  made  in  the  following 
way :  Tubercle  bacilli  of  the  human  type  are  grown  for  from  4  to  6 
weeks  upon  a  5  per  cent,  glycerin  broth.  The  cultures  are  then 
sterilized  in  an  Arnold  sterilizer  and  are  evaporated  at  about  80°  C. 


356  INFECTION    AND    RESISTANCE 

to  one-tenth  the  original  volume.  This  60  per  cent,  glycerin  extract 
of  the  tubercle  bacilli  is  then  filtered  clear  and  constitutes  the  tuber- 
culin. 

The  old  tuberculin  is  a  preparation  which  is  extensively  used 
in  the  subcutaneous  and  intracutaneous  tests  upon  human  beings  and 
cattle,  and  forms  the  basis  of  the  various  preparations  by  von  Pir- 
quet,  Moro,  and  others  in  the  cutaneous  tuberculin  reactions.  In 
his  earliest  work  von  Pirquet  used  a  25  per  cent,  solution  of  the  old 
tuberculin.  At  present  an  undiluted  old  tuberculin  is  used  for  these 
purposes. 

The  old  tuberculin  also  is  the  material  from  which  the  prepara- 
tion for  the  ophthalmotuberculin  test  is  made.  For  this  purpose 
Calmette  advises  precipitating  old  tuberculin  with  double  the  volume 
of  95  per  cent,  alcohol,  allowing  the  precipitate  to  settle  and  repeat- 
edly washing  the  sediment  with  70  per  cent,  alcohol.  The  powder 
which  results  is  thoroughly  dried,  pulverized,  and  made  up  for  use 
in  0.5  per  cent,  solutions.  Bandelier  and  Roepke  recommend  the 
use  of  the  diluted  old  tuberculin  directly  for  these  tests,  employing 
a  1  per  cent,  solution. 


TUBERCULIN  (T  R  AND  T  O) 

The  description  of  the  preparation  of  these  tuberculins  we  take 
from  Ruppell  in  the  Lancet,  March  28,  1908.  Virulent  cultures  of 
tubercle  bacilli  are  dried  in  the  vacuum  and  are  then  thoroughly 
pulverized  by  specially  constructed  machinery,  and  the  grinding  is 
continued  until  no  intact  bacilli  are  found  in  the  preparation.  One 
gram  dry  weight  is  then  shaken  up  in  100  c.  c.  of  sterile  distilled 
water.  The  mixture  is  then  centrifugalized  at  high  speed — the 
supernatant  fluid  is  T  O  (tuberkulin  oberschicht).  This  contains 
the  water-soluble  substances  of  the  bacillus  and  gives  no  precipitate 
with  glycerin.  The  residue — T  R  (tuberkulin  ruckstand) — is  again 
dried  and  ground  up,  shaken  up  in  water,  and  centrifugalized.  This 
is  repeated  3  or  4  times,  the  total  volume  of  water  used  for  all  the 
repetitions  not  exceeding  100  c.  c.  At  the  end  of  several  repetitions 
.all  the  T  R  goes  into  emulsion,  and  the  various  supernatant  fluids 
obtained  during  these  repeated  grindings  and  shaking  are  mixed 
together  and  constitute  the  final  T  R  preparation.  This  preparation, 
according  to  Koch,  contains  important  antigenic  substances,  it  gives 
a  precipitate  with  glycerin,  and  it  is  standardized  by  the  determina- 
tion of  the  solid  substances  contained  in  a  cubic  centimeter.  This, 
for  a  standard  preparation,  should  be  0.002  gram  to  a  cubic  centi- 
meter. 


OPSONIC    INDEX    AND    VACCINE    THERAPY      357 


TUBERCULIN  BACILLARY  EMULSION 

This  preparation  consists  of  a  combination  of  T  O  and  T  R.  It 
represents  an  emulsion  of  pulverized  tubercle  bacilli  in  100  parts  of 
50  per  cent,  glycerin.  The  preparation  as  marketed  contains  0.005 
gram  solid  substance  to  the  cubic  centimeter.  It  is  prepared  simply 
by  mechanically  grinding  the  bacteria  as  in  the  new  tuberculin,  but, 
instead  of  centrifugalizing  for  the  separation  of  T  O  and  T  R,  the 
bacteria  are  allowed  to  sediment  after  the  addition  of  glycerin.  This 
is  the  preparation  which  is  extensively  used  in  many  places  at  pres- 
ent for  the  treatment  of  tuberculosis.  It  was  adopted  by  Koch  par- 
ticularly because  of  experiments  in  which  he  showed  that  the  treat- 
ment of  animals  with  such  preparations  greatly  increased  the  ag- 
glutinins  for  tubercle  bacilli. 

BOUILLON  FILTRE   (DENYS) 

Denys  cultivates  the  tubercle  bacilli  upon  5  per  cent,  glycerin 
bouillon  as  in  the  preparation  of  old  tuberculin,  but  does  not  heat, 
sterilizing  his  cultures  by  filtration  through  porcelain.  Denys  be- 
lieves that  the  application  of  heat  in  sterilization  destroys  exotoxins 
which  have  valuable  antigenic  properties. 

SENSITIZED  TUBERCULIN 

Following  the  introduction  of  sensitized  vaccines  in  other  dis- 
eases by  Besredka,  Meyer  39  has  introduced  the  sensitized  tuberculin. 
This  tuberculin  is  prepared  in  the  following  way :  Tubercle  bacilli 
of  the  human  type  are  washed  and  dried  and  are  mixed  with  a  con- 
siderable quantity  of  the  serum  of  animals  immunized  with  tubercle 
emulsions  and  containing  considerable  quantities  of  tubercle-agglu- 
tinins.  These  serum  mixtures  are  kept  at  37°  C.  for  several  days 
and  are  then  shaken  in  a  shaking  machine  until  intact  tubercle  bacilli 
are  no  longer  to  be  found.  The  tubercle  bacillus  fragments  are  then 
thrown  down  in  the  centrifuge,  washed  in  salt  solution,  and  emulsi- 
fied in  40  per  cent,  glycerin,  0.5  per  cent,  carbolic  acid  being  used. 
The  emulsion  contains  0.005  gram  dry  weight  to  a  cubic  centimeter. 
We  take  the  description  of  the  preparation  from  that  cited  by 
Bandelier  and  Roepke. 

The  above  tabulation  contains  the  most  important  tuberculin 
preparations  as  they  are  at  the  present  time  in  use.  For  detailed 
studies  of  their  clinical  application  we  refer  the  reader  to  the  very 
valuable  book  of  Bandelier  and  Roepke,  "Lehrbuch  der  spezifischen 
Diagnostik  und  Therapie  der  Tuberkulose,"  Curt  Kabitzsch,  Wiirz- 
burg. 

39  Meyer.  Cited  from  Bandelier  and  Roepke,  "Lehrbuch  d.  spez.  Diagn. 
u.  Ther.  d.  Tuberkulose,"  Kabitzsch,  Wiirzburg,  6th  ed.,  1911,  p.  186. 


CHAPTER  XV 

ANAPHYLAXIS 

FUNDAMENTAL    FACTS 

THE  fundamental  principle  of  active  immunization  is  the  fact 
that  the  treatment  of  animals  with  bacteria  or  bacterial  products, 
carried  out  according  to  certain  empirically  determined  methods, 
leads  to  increased  tolerance  or  resistance.  The  limitations  within 
which  this  statement  is  true,  and  the  variable  factors  to  which  it  is 
subject,  we  have  considered  in  the  foregoing  discussions  dealing  with 
the  antibody-antigen  reactions. 

Although  these  reactions  were  studied  at  first  purely  from  the 
point  of  view  of  increased  resistance  to  infection,  the  most  extensive 
studies  of  antibody  formation  have  been  made  with  such  antigens  as 
blood  cells,  serum,  and  other  substances  which  are  in  themselves  en- 
tirely harmless.  For,  in  such  reactions,  great  simplicity  and  ease  of 
experimentation  could  be  attained.  For  a  time,  therefore,  the  pri- 
mary problem  of  increased  tolerance  or  resistance  was  relegated  to  a 
secondary  position,  or,  at  least,  dealt  with  chiefly  by  analogy,  and  the 
phenomena  of  increased  antibody  formation  and  increased  resistance 
to  the  antigen  were  assumed  to  maintain  a  more  or  less  strict  paral- 
lelism. 

That  the  problem  is  not  as  simple  as  this  has  gradually  become 
obvious.  We  have  come  to  recognize  that  the  treatment  of  animals 
with  any  antigen,  bacterial  or  otherwise,  though  leading  to  increased 
tolerance  under  certain  conditions  and  within  definite  limits,  may, 
under  other  conditions,  give  rise  to  the  very  opposite,  that  is,  to  an 
intolerance  or  increased  susceptibility. 

The  development  of  this  knowledge,  like  much  else  that  serum 
study  has  revealed  in  the  last  fifteen  years,  takes  root  in  isolated 
observations  scattered  throughout  the  early  literature,  but  often 
regarded  as  merely  noteworthy  accidents  or  technical  errors.  This 
particular  problem,  moreover,  was  confused  by  the  fact  that  some  of 
the  earliest  observations  regarding  hypersusceptibility  were  made  in 
the  course  of  experimentation  with  diphtheria  and  tetanus  toxins, 
antigenic  substances  toxic  in  themselves  and,  therefore,  as  we  shall 
see,  clouding  some  of  the  basic  principles  apparently  involved  in  the 
phenomenon  of  which  we  now  speak  as  anaphylaxis.  We  will  for  the 
present,  therefore,  limit  our  discussion  to  the  development  of  the 

658 


ANAPHYLAXIS  359 

knowledge  of  anaphylaxis  merely  as  it  concerns  the  hypersuscepti- 
bility  incited  in  animals  and  man  by  treatment  with  various  antigens, 
such  as  animal  sera  and  other  proteins,  which  possess  but  slight 
native  toxicity  or  no  toxicity  whatever  in  themselves. 

The  special  problem  of  toxin  hypersusceptibility  ( "Gif tuber emp- 
findlichkeit"  of  von  Behring)  we  will  deal  with  later  in  a  separate 
section,  since  it  is  as  yet  very  doubtful  whether  these  phenomena 
may  justly  be  incorporated  with  true  anaphylaxis  as  we  now  define 
it,  despite  the  admitted  fact  that  attention  was  called  to  the  prob- 
lems of  acquired  susceptibility  largely  because  of  these  toxin  in- 
vestigations. 

The  earliest  observation  having  direct  bearing  upon  protein 
anaphylaxis  is  one  which  Morgenroth  discovered  in  the  writings  of 
Magendie.  Morgenroth  l  mentions  that,  in  his  "Vorlesungen  iiber 
das  Blut,"  published  in  1839,  Magendie  describes  the  sudden  death 
of  dogs  which  had  been  repeatedly  injected  with  egg  albumen.  Al- 
though Morgenroth,  whose  paper  was  written  before  the  present 
facts  regarding  hypersusceptibility  were  fully  developed,  attributes 
these  results  to  the  action  of  precipitins,  there  can  be  little  doubt  as 
to  the  anaphylactic  nature  of  Magendie's  results. 

A  clear  statement  of  the  fundamental  phenomena  was  given,  also, 
by  Flexner,2  in  1894.  In  describing  certain  experiments  he  says: 
" Animals  that  had  withstood  one  dose  of  dog  serum  would  succumb 
to  a  second  dose  given  after  the  lapse  of  some  days  or  weeks,  even 
when  this  dose  was  sublethal  for  a  control  animal." 

One  of  the  experiments  cited  to  justify  this  statement  is  as 
follows : 

"Two  rabbits  received  %  of  1  per  cent,  and  1  per  cent,  of  their 
body  weight  respectively  of  dog's  serum,  twenty-four  hours  old,  on 
January  19,  1894.  With  the  exception  of  hemoglobinuria,  indisposi- 
tion to  move,  and  increased  respiration,  no  ill  effects  were  noted. 
The  animals  still  showed  hemoglobinuria  on  the  following  day. 
These  symptoms  disappeared  and  apparently  the  rabbits  entirely 
recovered.  On  February  12,  1894,  each  received  1  per  cent,  of  their 
body  weight  of  dog's  serum  intravenously.  A  control  animal  also 
received  1  per  cent,  of  its  body  weight  of  the  same  serum.  The  two 
animals  that  had  been  previously  inoculated  died  in  two  and  twelve 
hours  respectively;  the  control  animal  showed  only  hemoglobinuria 
which  disappeared  after  a  day  or  two." 

The  experiment  here  quoted  is,  as  a  matter  of  fact,  a  perfect  ex- 
ample of  what  we  now  know  as  "active  sensitization." 

However,  the  isolated  observations  recorded  above  were  neither 
correlated  nor  followed  out  to  their  logical  developments,  and  a 

1  Morgenroth.     "Ehrlich    Gesamme'lte    Arbeiten,"    Transl.,  Wiley  &  Son, 
N.  Y.,  1906;  p.  332  footnote. 

2  Flexner.     Medical  News,  Vol.  65,  p.  116,  1894. 


360  INFECTION    AND    RESISTANCE 

systematic  and  purposeful  study  of  the  problem  was  deferred  until 
Richet  and  Portier  3  attacked  it  in  1902. 

Richet  and  Hericourt4  had  observed  in  1898  that  dogs  treated 
with  eel  serum,  which  is  toxic  per  se,  could  be  killed  by  a  second 
injection  of  an  amount  too  small  to  injure  normal  untreated  animals. 
Some  years  later  Richet,  in  collaboration  with  Portier,5  determined 
a  similar  fact  in  the  case  of  a  poisonous  substance,  "actinocongestin," 
which  they  isolated  by  extraction  of  the  tentacles  of  actinia. 

Some  of  the  facts  of  Richet  and  Hericourt's  observations  are  as 
follows :  Actinocongestin  injected  intravenously  into  dogs  in  quan- 
tities of  0.05  to  0.075  gram  per  kilo  weight  may  cause  illness,  with 
vomiting,  diarrhea,  and  respiratory  distress,  but  does  not  kill.  A 
dose  of  0.002  gram  per  kilo  causes  no  symptoms  in  a  normal  dog. 
If,  however,  0.002  gram  of  the  poison  is  injected  into  a  dog  which 
has  previously  received  a  sublethal  dose  and  recovered,  the  result 
is  violent  illness  and  often  death.  It  was  obvious,  and  this  was 
clearly  stated  by  Richet,  that  the  first  dose  had  induced  a  condition 
of  markedly  greater  susceptibility  to  the  poison. 

He,  therefore,  spoke  of  the  phenomenon  as  "anaphylaxis"  ("ac- 
tion anaphylactique  de  certains  venins")  to  express  its 'antithesis  to 
prophylaxis  or  protective  effects. 

Although  it  has  been  disputed  by  a  number  of  writers  that 
Richet's  investigations  constitute  the  beginnings  of  our  modern 
understanding  of  the  anaphylactic  phenomena,  yet  his  recognition 
of  the  distinct  dependence  of  the  hyper  susceptible  condition  upon  a 
preceding  inoculation  with  the  same  substance,  and  his  conclusion 
that  a  definite  incubation  time  must  elapse  after  the  first  injection 
before  susceptibility  is  developed,  defined  two  of  the  most  important 
criteria  of  the  condition  and  initiated  purposeful  investigations  in 
this  field.  It  is  true,  on  the  other  hand,  that,  like  v.  Behring  and 
most  of  his  other  predecessors,  he  was  working  with  primarily  toxic 
substances,  and  the  final  recognition  of  the  general  biological  sig- 
nificance of  the  anaphylactic  phenomenon  was  necessarily  deferred 
until  a  similar  development  of  hypersusceptibility  was  noted  in 
animals  injected  with  various  antigens  which  of  themselves  were 
entirely  harmless.  In  this  the  history  of  anaphylactic  investigations 
is  similar  to  that  of  other  reactions  to  antigen  injections,  lysin,  ag- 
glutinin,  and  precipitin  formation,  in  which  the  first  observations 
were  made  upon  pathogenic  bacteria  or  their  products,  and  in  which 
subsequent  extension  of  the  investigations  revealed  that  the  response 
to  inoculation  with  bacterial  proteins  represented  merely  a  single 
phase  of  a  general  biological  reaction  on  the  part  of  animals  to  treat- 
ment with  the  large  class  of  substances  known  as  antigens. 

8  Richet  and  Portier.    C.  E.  de  la  Soc.  Biol,  p.  170,  1902. 

4  Richet  and  Hericourt.     C.  E.  de  la  Soc.  Biol,  1898. 

5  Portier  and  Richet.     C.  E.  de  la  Soc.  Biol,  p.  170,  1902. 


ANAPHYLAXIS  361 

This  generalization  of  Richet's  observations  had  really  been 
foreshadowed  by  the  observations  of  Magendie  and  by  the  experi- 
ments 'of  Flexner  quoted  above,  but  this  work  had  been  lost  sight  . 
of  and  the  attention  of  investigators  was  again  focused  upon  the 
problem  mainly  by  the  publication  of  Arthus  6  in  1903  on  the  re- 
peated injection  of  horse  serum  into  rabbits,  and  some  observations 
made  upon  guinea  pigs  by  Theobald  Smith  and  communicated  by 
him  in  1904  to  Ehrlich. 

Arthus  7  found  that  horse  serum  injected  into  rabbits  by  any  of 
the  usual  paths  of  entrance  is  entirely  innocuous.  It  is  possible  to 
inject  10,  20,  or  even  40  c.  c.  without  harm.  If,  however,  one  re- 
peatedly injects  small  amounts,  5  c.  c.  or  less,  subcutaneously,  at 
intervals  of  several  days,  eventually  the  later  injections  will  give  7 
rise  to  infiltrations,  edema,  sterile  abscesses,  and  even  gangrene  at/ 
the  points  of  injection.  He  recognized  that  this  was  not  due  to 
cumulative  action,  and  that  it  was  not  necessary  to  inject  several 
times  in  the  same  place  to  produce  the  characteristic  response.  For 
instance,  the  early  injections  might  be  made  into  the  peritoneum, 
the  subsequent  ones  into  the  skin,  and  the  local  reactions  to  the  later 
injections  might  nevertheless  ensue.  In  other  words,  he  recognized 
the  systemic  nature  of  the  phenomenon  and  regarded  it  as  analogous 
to  the  observations  of  Richet  in  that  he  spoke  of  the  hypersensitive 
rabbits  as  "anaphylactises"  by  a  series  of  preparatory  injections. 

The  "phenomenon  of  Theobald  Smith"  is  closely  related  to  that 
of  Arthus,  and  was  made  in  the  course  of  the  standardization  of 
diphtheria  antitoxin  in  guinea  pigs.  It  was  noticed  that  guinea 
pigs  which  had  been  used  for  this  purpose  and  had  survived  had 
acquired  great  susceptibility  to  subsequent  injections  of  normal  horse 
serum  made  several  days  or  weeks  later. 

With  these  observations  as  points  of  departure,  together  with  the 
studies  of  v.  Pirquet  and  Schick  8  upon  the  clinical  manifestations 
of  antitoxin  injections  into  human  beings,  a  number  of  investigators 
took  up  the  problem,  chief  among  them  Rosenau  and  Anderson,  of 
the  United  States  Hygienic  Laboratory,  and  R.  Otto,  of  the  Frank- 
furt Institute  of  Experimental  Therapy. 

Although  the  paper  of  Otto  9  appeared  in  print  a  little  earlier 
than  did  the  first  one  of  the  American  workers,  the  investigations 
were  independent  and  almost  synchronous.  Their  results,  moreover, 
confirm  each  other  in  all  essentials.  Otto  showed  that  the  Theobald 
Smith  phenomenon  was  entirely  independent  of  the  toxin  or  anti- 

6  Arthus.    C.  R.  de  la  Soc.  Biol.,  Vol.  55,  p.  817,  Reunion  biol.,  Marseille, 
June,  1903. 

7  Arthus  et  Breton.     C.  E.  de  la  Soc.  Biol,  55,  p.  1478. 

8  Von  Pirquet  u.  Schick.     "Die  Serumkrankheit,"  Deuticke,  Wien,  1906. 

9  Otto.     "Das  Theobald  Smithsche  Phaenomen,  etc.,  v.  Leuthold  Gedenk- 
schrift,"  Vol.  1.  1905 ;  also  Otto  in  Erganzungsband  2,  "Kolle  u.  Wassermann 
Handbuch,"  etc. 


362  INFECTION    AND    RESISTANCE 

toxin  contents  of  the  injected  serum,  but  could  be  produced  (though 
somewhat  less  markedly)  with  horse  serum  alone.  He  also  showed 
that,  while  a  preliminary  injection  of  horse  serum  "sensitized"  a 
guinea  pig  to  a  subsequent  dose  given  after  an  interval  of  10  to  12 
days,  the  repeated  injection  of  considerable  quantities  at  short  inter- 
vals produced  a  condition  of  "antianaphylaxis"  or  immunity  to  the 
later  injections.  Otto,  too,  excluded  from  his  results  the  direct  rela- 
tion of  the  anaphylactic  state  with  the  possible  presence  of  serum 
precipitins,  a  thought  suggested  by  Morgenroth  in  his  interpretation 
of  the  observations  of  Magendie  mentioned  above. 

Eosenau  and  Anderson  10  had  attacked  the  problem  with  the  pri- 
mary purpose  of  throwing  light  upon  the  occasional  accident  of  sud- 
den death  following  the  injection  of  diphtheria  antitoxin  into  human 
beings.  Since  the  detailed  description  of  their  extensive  investiga- 
tions would  tend  to  render  more  difficult  the  exposition  of  an  already 
sufficiently  complicated  subject,  it  will  be  best  to  tabulate  the  chief 
results  of  this  classical  series  of  their  earlier  papers.  Briefly,  these 
are  as  follows : 

1.  A  single  injection  of  horse  serum  into  guinea  pigs,  harmless 
in  itself,  renders  these  animals  hypersusceptible  to  a   subsequent 
injection  given  after  a  definite  interval  or  incubation  time. 

2.  This  interval,  with  the  ordinary  dosages  employed  (about  1 
to  2  c.  c.),  was  about  10  days.    Properly  carried  out  injections  after 
this  period  were  usually  fatal. 

3.  The  known  antibodies,  antitoxins,  hemolysins,  and  precipi- 
tins, are  not  responsible  for  the  reaction. 

4.  The  reaction  is  "quantitatively"  specific,  injections  of  horse 
serum  sensitizing  to  horse  serum  only.     (The  question  of  specificity 
will  be  further  discussed  below.) 

5.  The  sensitive  condition  is  transmissible  from  mother  to  off- 
spring,11 the  young  of  sensitized  mothers  being  hypersusceptible  to 
a  first  injection  of  horse  serum. 

6.  The  reaction  is  extremely  delicate.     Rosenau  and  Anderson 
succeeded  in  sensitizing  in  one  case  with  0.000001  c.  c.   (one  one- 
millionth)  of  horse  serum. 

7.  The  hypersusceptible  state  is  not  a  transient  condition,  but 
may  last  a  long  time. 

8.  Sensitization,  or  the  production  of  the  hypersusceptible  con- 
dition, can  be  carried  out,  not  only  with  the  various  animal  and  vege- 
table proteins  employed  in  the  first  experiment,  but  can  be  brought 

10Rosenau  and  Anderson.  U.  S.  Pub.  Health  and  M.  H.  S.  Hyg.  Lab. 
Bull  29,  1906;  30,  1906;  36,  1907;  Journ.  Med.  Ees.,  Vol.  15,  1906,  Vol.  16, 
1907;  also  Jour.  Inf.  Dis.,  Vol.  4,  1907,  Vol.  5,  1908. 

11  It  is  important  practically,  as  Anderson  points  out,  that  a  female 
guinea  pig  may  transmit  to  its  young  sensitiveness  to  horse  serum  and  im- 
munity to  diphtheria  toxin. 


ANAPHYLAXIS 

about  by  the  use  of  extracts  of  various  bacteria.  In  such  cases  also 
the  reaction  is  specific.  The  first  determinations  with  bacterial  ex- 
tracts carried  out  by  Eosenau  and  Anderson  were  made  with  colon, 
anthrax,  typhoid,  and  tubercle  bacilli. 

By  these  observations,  then,  the  possibility  of  a  direct  relation 
between  the  phenomena  of  anaphylaxis  and  infectious  diseases  in 
animals  was  indicated. 

This,  in  essence,  is  the  harvest  of  the  two  earliest  purposeful  re- 
searches into  this  problem.  A  large  number  of  investigators  now 
took  up  the  question,  and  its  further  elucidation,  as  we  shall  see,  has 
proved,  not  only  the  most  directly  fruitful  of  the  phases  of  recent 
immunological  studies,  but  has  thrown  much  indirect  light  upon 
antigen-antibody  reactions  apart  from  the  anaphylactic  phenomena 
themselves. 

Before  entering  into  the  further  discussion  of  the  experimental 
data,  however,  it  will  be  necessary  to  describe  briefly  the  clinical 
manifestations  which  follow  upon  the  second  injection  of  an  anaphy- 
lactic antigen  into  a  sensitized  animal,  manifestations  which  we  have 
heretofore  summarized  in  the  phrase  "anaphylactic  shock."  For 
there  has  been  much  controversy  regarding  the  physiological  mech- 
anism which  lies  at  the  bottom  of  these  symptoms,  and  the  matter 
has  been  complicated  by  the  unquestionably  different  reactions  oc- 
curring in  various  species  of  animals  in  response  to  the  anaphylactic 
experiment. 

Since  anaphylactic  studies  were  begun  largely  as  the  result  of 
Theobald  Smith's  observations  upon  guinea  pigs,  and  subsequent 
study  has  revealed  these  animals  as  peculiarly  susceptible  to  the 
anaphylactic  poison,  the  large  bulk  of  the  experimental  data  at  our 
disposal  was  worked  out  upon  these  animals.  In  consequence  our 
understanding  of  the  mechanism  of  the  reaction  is  based  largely 
upon  guinea  pig  studies. 

If  a  properly  sensitized  guinea  pig  receives  a  second  injection  of 
an  antigen  after  a  suitable  incubation  time  a  very  characteristic 
train  of  symptoms  ensues.  There  is  usually  a  short  preliminary 
period — lasting  either  a  fraction  of  a  minute  or  several  minutes  ac- 
cording to  the  violence  of  the  reaction  and  the  mode  of  administra- 
tion— during  which  the  pig  appears  normal.  At  the  end  of  this  time 
the  animal  will  grow  restless  and  uneasy,  and  will  usually  rub  its 
nose  with  its  forepaws.  It  may  sneeze  and  occasionally  emit  short 
coughing  sounds.  At  the  same  time  an  increased  rapidity  of  res- 
piration is  noticeable  and  the  fur  will  appear  ruffled.  In  light  cases 
the  animals  may  remain  in  this  condition,  with  further  irregularity 
and  difficulty  of  respiration,  possible  discharges  of  urine  and  f eces ; 
then  gradual  slow  recovery  may  set  in,  with  complete  return  to 
normal  in  from  30  minutes  to  several  hours.  In  more  severe  cases 
these  preliminary  stages  are  rapidly  followed  by  great  apparent 


364  INFECTION    AND    RESISTANCE 

weakness.  The  animals  fall  to  the  side,  the  legs  and  trunk  muscles 
twitch  irregularly,  and  the  respiration  becomes  slow  and  shallow; 
the  thorax  never  entirely  contracts,  but  remains  in  a  more  or  less 
expanded  condition.  The  very  evident  dyspnea  is  of  an  inspiratory 
character.  The  excursions  of  the  lung  itself  seem  to  grow  shallower 
and  shallower  in  spite  of  apparent  strong  inspiratory  efforts — the 
volume  of  the  thorax  and  lung  remaining  in  the  expanded  condition. 
At  this  stage  evidences  of  motor  irritation  may  appear,  in  that  the 
animal  may  arise  and  attempt  to  run.  More  often,  however,  in  this 
phase  general  convulsions  set  in,  often  several  times  repeated,  and 
in  these  the  animals  usually  die. 

On  the  other  hand,  after  cessation  of  convulsions  they  may  lie 
perfectly  still  on  the  side  as  though  paralyzed,  the  breathing  becom- 
ing gradually  slower  and  more  shallow,  finally  ceasing  entirely.  The 
heart  may  continue  to  beat  for  a  considerable  time  after  the  breath- 
ing has  stopped. 

If  such  an  animal  is  immediately  autopsied  a  very  characteristic 
condition  is  found — to  which,  in  the  essentials,  attention  was  first 
called  by  Gay  and  Southard.12  They  speak  of  finding  "pulmonary 
emphysema  as  a  constant  feature  at  autopsy/'  and  attribute  the 
anaphylactic  death  in  guinea  pigs  to  cessation  of  respiration  in  the 
inspiratory  phase  under  the  influence  of  respiratory  central  intoxi- 
cation. 

The  lungs  of  such  guinea  pigs  after  death  are  found  distended 
and  completely  filling  the  thorax.  They  are  usually  pale  and  blood- 
less and  do  not  collapse  as  the  pleurae  are  opened.  On  microscopic 
examination  the  alveoli  are  seen  to  be  distended  and  small  hemor- 
rhages may  appear  upon  the  serous  surfaces.  According  to  Gay 
and  Southard,  furthermore,  histological  study  of  the  other  organs 
shows  also  hemorrhages  in  the  brain,  stomach,  heart,  cecum,  and 
spleen — more  rarely  in  other  organs,  and  there  are  local  fatty 
changes  in  the  capillary  endothelium  which  they  regard  as  causa- 
tively  related  to  the  hemorrhages. 

That  the  respiratory  symptoms  are  the  most  striking  feature  of 
the  clinical  picture  of  guinea  pig  anaphylaxis  had,  as  a  matter  of 
fact,  been  noticed  by  Rosenau  and  Anderson.  A  detailed  physiologi- 
cal study  of  the  mechanism  of  the  respiratory  death  in  these  cases 
was  first  made,  however,  by  Auer  and  Lewis  13  in  1909. 

These  investigators  showed  that,  during  the  later  respiratory 
symptoms,  little  or  no  air  enters  the  lungs,  although  the  animal 
makes  violent  respiratory  efforts.  This  is  due,  as  they  found,  to  a 
tetanic  contraction  of  the  small  bronchioles,  which  practically  oc- 

12  Gay  and  Southard.     Jour.  Med.  Ees.,  Vol.  16.  1907:  Vols.  18  and  19, 
1908. 

13  Auer  and  Lewis.     Jour,  of  the  A.  M.  A.,  Vol.  53,  p.  458,  1909 ;  Jour. 
Exp.  Med.,  Vol.  12,  1910. 


ANAPHYLAXIS  365 

eludes  the  air  passages.  That  the  origin  of  this  contraction  is  not,  as 
previously  supposed,  of  central  origin,  but  is  referable  to  peripheral 
cause,  they  proved  by  showing  that  the  same  phenomena  occur  in 
the  guinea  pigs  even  after  the  cord  and  medulla  have  been  destroyed 
and  the  vagi  divided.  In  such  cases,  of  course,  with  the  cord  and 
medulla  destroyed,  artificial  respiration  had  to  be  done,  and  when 
the  symptoms  set  in  it  was  found  that  the  lungs  could  no  longer  be 
expanded  by  the  same  force  of  artificial  respiration  which  before  this 
had  been  sufficient. 

They  showed  also  that  the  non-collapsible  expansion  of  the  lungs 
after  death  was  due  to  imprisonment  of  the  air  in  the  alveoli  by  the 
contracted  musculature  of  the  small  bronchioles,  and  further  con- 
firmed their  opinion  of  the  peripheral  origin  of  this  contraction  by 
the  important  discovery  that  atropin  will  markedly  protect,  often 
preventing  death  or  hastening  recovery.  It  is  noteworthy,  too,  that 
Auer  and  Lewis  speak  of  occasionally  finding  slight  pulmonary 
edema,  a  feature  which  Biedl  and  Kraus  consider  incompatible  with 
true  anaphylaxis.  * 

Anderson  and  Schultz,14  who  have  (bn^rmed  much  of  the  work 
of  Auer  and  Lewis,  find  that  not  only  atrojrin  will  prevent  asphyx- 
iation in  these  cases,  but  methane,  chloral  hydrate,  adrenalin,  and 
pure  oxygen  will  exert  a  similar  effect.  The  animals  may  be  saved 
from  suffocation  in  this  way,  but  may  nevertheless  die,  probably  as 
the  result  of  lowered  blood  pressure. 

The  observations  of  Auer  and  Lewis  have  been  further  confirmed 
especially  by  Biedl  and  Kraus,15  who  regard  it  as  well  established 
that  anaphylactic  death  in  guinea  pigs  is  caused  primarily  by  suffo- 
cation, due  to  tetanic  spasms  of  the  musculature  of  the  small  bronchi. 
These  spasms  are  not  of  central  origin,  but  are  peripherally  initiated, 
possibly  by  direct  action  upon  the  smooth-  muscle  itself.  The  fact 
that  atropin  is  not  effective  in  preventing  death  in  all  severe  cases  is 
no  argument  against  this,  since  such  an  eff ect^jKOuld  naturally  de- 
pend upon  the  relation  between  the  amount  of  atropin  given  and  the 
severity  of  the  attack.  In  this  connection  the  studies  that  have  been 
made  upon  the  irritability  of  smooth  muscle  fibers  in  normal  and  in 
sensitized  animals  are  of  great  interest.  Schultz,16  following  out  an 
observation  made  by  Rosenau  and  Anderson,  studied  the  intestinal 
muscle  of  normal  sensitized  guinea  pigs  excised  and  suspended  in 
HowelPs  solution.  In  this  way  he  showed  that  during  the  period  of 
hypersusceptibility  the  smooth  muscle  is  abnormally  sensitive  to 
treatment  with  the  antigen.  The  contraction  which  normally  occurs 

14  Anderson  and  Schultz.    Proc.  Soc.  Exp.  Biol.  and  Med.,  7,  1909,  p.  32. 

15  Biedl   und   Kraus.      Zeitschr.   f.   Immunitatsforschung,    Vol.    7,    1910; 
Centralbl  f.  PhysioL,  1910;   Wien.  klin.   Woch.,  No.  11,  1910. 

16  Schultz.     Jour.  Pharm.  and  Exp.  Therap.,  1,  1910 ;  2,  1910. 


366  INFECTION    AND    RESISTANCE 

in  smooth  muscles  under  the  influence  of  serum  is  markedly  aug- 
mented if  the  preparations  are  taken  from  sensitized  animals. 

In  addition  to  these  predominant  features  of  the  anaphylactic 
symptomatology  in  guinea  pigs,  there  are  a  number  of  secondary  re- 
actions which,  though  less  prominent,  are  nevertheless  of  considerable 
interest  and  theoretical  importance.  The  conditions  in  the  circula- 
tion are  probably,  to  a  great  extent,  dependent  upon  the  respiratory 
condition,  and  the  fall  of  blood  pressure  in  guinea  pigs  is  regarded  by 
some  investigators  as  merely  a  secondary  manifestation  just  preceding 
death.  The  fall  of  temperature  first  described  by  H.  Pfeiffer,17 
however,  seems  to  be  an  occurrence  which,  though  standing  in  no 
causative  relation  to  the  symptoms  as  a  whole,  is  so  constant  and  well 
marked  that  it  has  been  taken  by  a  number  of  workers  as  one  of  the 
necessary  criteria  for  the  characterization  of  the  anaphylactic  con- 
dition. 

There  is,  indeed,  an  almost  regular  drop  of  several  degrees  in  the 
rectal  temperature,  and  a  close  observation  of  this  may  be  of  much 
aid  in  determining  the  occurrence  of  mild  reactions,  when  other 
symptoms  of  shock  are  not  strongly  marked.  Pfeiffer  18  himself 
goes  so  far  as  to  claim  that  by  this  symptom  alone  delicate  anaphy- 
lactic reactions  may  be  determined  when  all  other  symptoms  are 
lacking. 

Friedberger,19  20  too,  has  found  the  sudden  drop  of  temperature 
a  very  regular  occurrence,  and  has  employed  this  method  of  study 
for  the  analysis  of  the  intensity  of  anaphylactic  shock.  He  calls 
attention  to  the  apparent  difference  between  infection  and  anaphy- 
laxis  in  this  respect  in  that  in  the  former  there  is  fever,  in  the  latter 
there  is  depression  of  body  heat ;  but,  at  the  same  time,  he  points  out 
that  this  discrepancy  is  an  apparent  one  only,  and  determined  by 
quantitative  differences,  for  when  he  treated  sensitized  animals  with 
varying  doses  of  antigen  he  found  that  quantities  which  produced 
other  anaphylactic  symptoms  of  noticeable  degree  would  regularly 
depress  the  temperature  as  Pfeiffer  had  shown.  It  was  possible, 
however,  to  determine  a  minimal  dose  necessary  for  temperature 
reduction.  Quantities  just  below  this  left  the  temperature  un- 
changed, and  still  smaller  quantities  produced  fever  or  even  in- 
creased the  temperature.  This  fact  is  extremely  significant  in  that, 
as  we  shall  see,  it  has  an  important  bearing  upon  views  which  inter- 
pret bacterial  infection  as  a  series  of  anaphylactic  poisonings,  the 
multiplying  bacteria  furnishing  the  constant  supply  of  minute 
amounts  of  antigen.  This  thought,  indeed,  based  also  on  the  study  of 

17  H.  Pfeiffer.     Wien.  klin.  Woch.,  No.  1,  1909. 

18  Pfeiffer  u.  Mita.     Zeitschr.  f.  Immunitatsforschung,  Vol.  4,  1910. 
9  Friedberger.     Deutsche  med.  Woch.,  No.  11,  1911. 

20  Friedberger  und  Mita.  Zeitschr.  f.  Immunitatsforschung,  Vol.  10,  1911. 


ANAPHYLAXIS  367 

temperature  curves  in  animals,  was  expressed  by  Vaughan  21  as  early 
as  1909,  and  was  developed  by  him  with  Gumming  and  Wright22  in 
an  extensive  study  upon  what  he  called  "protein  fever."  It  was 
shown  in  these  experiments  that  continued  fever,  not  unlike  that  of 
infectious  diseases,  could  be  produced  in  rabbits  by  repeated  subcu- 
taneous injections  of  primarily  harmless  substances,  such  as  egg 
white  and  vegetable  proteins.  The  conditions  observed  and  the  con- 
clusions drawn  from  them  in  this  work,  as  well  as  in  the  similar  in- 
vestigations of  other  workers,  were  clearly  foreseen  by  Vaughan  in 
his  early  investigations  on  proteid  split-products  studies,  which  we 
will  find  occasion  to  discuss  in  a  later  section. 

The  rigidity  of  the  diagnostic  value  of  the  temperature  relations 
for  anaphylactic  shock  in  particular,  as  advanced  by  Pfeiffer,  was 
somewhat  weakened  by  Ranzi's  23  observations  that  foreign  serum 
may  produce  temperature  depression  when  injected  into  -perfectly 
normal  animals  and  that,  injected  into  sensitized  animals,  the  same 
reaction  may  follow  if  other  proteins  than  the  original  antigen  were 
administered. 

Although  these  objections  of  Ranzi  are  perfectly  just,  yet  there 
is  such  a  marked  quantitative  difference  between  the  reaction  in  nor- 
mal and  in  sensitized  animals  that,  in  principle,  Pfeiffer's  claim  is 
not  invalidated.  Friedberger24  very  logically  remarks  that,  after 
all,  the  phenomena  of  sensitization  as  well  as  those  of  immunity  are 
merely  an  exaggeration  of  normal  physiological  conditions,  and  in 
experiment  he  has  shown  that,  whereas  noticeable  depressions  of  tem- 
perature will  follow  in  the  normal  animal  only  upon  quantities  of 
antigen  exceeding  0.5  c.  c.,  the  temperature  of  the  sensitized  animal 
may  be  depressed  by  amounts  as  small  as  0.0005  c.  c. 

Apart  from  the  symptoms  so  far  discussed,  there  are  other  less 
apparent  characteristics  of  anaphylaxis  in  guinea  pigs,  all  of  which, 
however,  possess  considerable  importance  theoretically.  The  most  sig- 
nificant of  these  is  the  reduction  in  the  amount  of  alexin  or  comple- 
ment, first  noticed  by  Sleeswijk,25  which  occurs  after  the  injection 
of  the  second  or  toxogenic  dose — during  the  development  of  shock. 
This  phenomenon  is  so  closely  interwoven  with  the  later  theoretical 
aspects  of  anaphylaxis  that  we  will  defer  its  discussion  until  we 
have  completed  a  more  general  survey  of  the  field. 

In  guinea  pigs,  as  in  dogs,  Friedberger  and  others  have  also  seen 
a  lowered  coagulability  of  the  blood  and  a  temporary  diminution  of 
the  polynuclear  leukocytes  (leukopenia)  during  shock. 

21  Vaughan.     Zeitschr.  f.  Immunitatsforschung,  Vol.  1,  1909. 

22  Vaughan,   Gumming,  and  Wright.     Zeitschr.  f.  Immunitatsforschung, 
Vol.  9,  1911. 

23  Ranzi.     Zeitschr.  f.  Immunitatsforschung,  Vol.   2,   1909 ;   Wien.  klin* 
Woch.,  No.  40,  1909.     ' 

24  Friedberger  u.  Mita.    Loc.  cit. 

25  Sleeswijk.     Zeitschr.  f:  Immunitatsforschung,  Vol.  2,  1909. 


368  INFECTION    AND    RESISTANCE 

During  the  earlier  periods  of  experimentation  there  was  a 
marked  discrepancy  in  the  ease  with  which  guinea  pigs  could  be 
sensitized  by  American  and  German  investigators,  on  the  one  hand, 
and  by  Besredka  and  Steinhart  in  France,  on  the  other.  The  mor- 
tality, upon  second  injection,  was  much  higher,  with  like  quantities 
of  horse  serum  in  the  hands  of  the  first-named.  In  attempting  to 
explain  this,  Rosenau  and  Anderson  carried  out  typical  experiments 
with  horse  serum  sent  to  them  by  Besredka  and  obtained  high  per- 
centages of  fatal  results.  They  believe,  for  this  reason,  that  the  dif- 
ferences cannot  be  accounted  for  by  variations  in  the  toxicity  of  the 
horse  sera,  but  conclude  that  probably  there  are  varying  grades  of 
susceptibility  to  the  reaction  in  guinea  pigs  of  different  breeds. 

Next  to  guinea  pigs  the  animals  most  commonly  employed  for 
anaphylactic  experiment  are  rabbits  and  dogs.  In  both  of  these  the 
symptoms  and  autopsy  findings  differ  markedly  from  each  other  and 
from  those  observed  in  guinea  pigs. 

In  sensitized  rabbits  the  injection  of  a  second  dose  of  the  antigen 
is  usually  followed,  after  a  short  but  definite  incubation  time,  by 
great  weakness  with,  often,  discharge  of  urine  and  feces.  The  ani- 
mals sink  down  until  the  abdomen  touches  the  ground,  the  legs  are 
stretched  out  weakly  but  not  paralyzed,  and  the  head  may  drop 
forward  or  to  one  side.  After  this,  the  animal  may  gradually  fall 
upon  its  side  and  lie  motionless  except  for  labored  and  irregular 
breathing  and  occasional  twitching  of  the  legs  and  head.  Sometimes 
this  gradual  relaxation  may  be  interrupted  by  a  sudden  motor  irri- 
tation, the  rabbit  suddenly  getting  up  and  running  a  short  distance 
but  soon  falling  down  again  apparently  from  a  sudden  return  of  the 
muscular  weakness.  During  these  running  spells  it  seems  as  though 
there  was  no  sense  of  direction  or  purpose — the  animals  running  into 
obstructions  or  off  tables  as  the  case  may  be.  During  this  period  gen- 
eral convulsions  and  a  drawing  back  of  the  head  by  a  tetanic  spasm 
of  the  muscles  of  the  neck  are  not  uncommon.  Death  may  occur 
within  a  few  minutes,  or  it  may  follow  a  gradually  increasing  weak- 
ness in  the  course  of  several  hours.  The  fall  of  blood  pressure  here 
seems  to  be  purely  secondary  to  the  general  failure  of  all  the  func- 
tions.26 

Anaphylaxis  in  dogs  has  been  very  extensively  studied,  especially 
by  Bie'dl  and  Kraus,27  and  by  Pearce  and  Eisenbrey.28  The  symp- 
toms in  dogs  are  characterized  by  a  rapid  progressive  fall  in  the 
blood  pressure,  followed  by  the  symptoms  of  cerebral  anemia.  Ana- 
phylactic dogs,  after  injection,  will  at  first  grow  restless,  vomit,  and 

26  Arthus.     Arch.  Internat.  de  Physiol.,  7,  1909. 

27  Biedl  and  Kraus.     LOG.  cit.;  also  in  "Kraus  u.  Levaditi  Handbuch," 
Erganzungsband  1. 

28  Pearce  and  Eisenbrey.     Proc.  Soc.  Exp.  Biol  and  Med.,  7,  1909,  p.  30; 
Transact.   Congr.  Am.  Ph.  and  S.,  Vol.  8,   1910. 


ANAPHYLAXIS 

pass  urine  and  feces.  They  then  grow  rapidly  weak,  fall  to  the 
ground,  and  continue  to  twitch  and  vomit  and  the  respiration  be- 
comes labored  and  irregular.  There  is  general  weakness  of  the  mus- 
cles, but  no  paralysis.  The  marked,  constant,  and  characteristic 
feature  of  the  condition  in  these  animals  is  the  fall  of  blood  pressure. 
There  is  also  a  lessened  coagulability  of  the  blood,  much  more 
strongly  developed  than  in  guinea  pigs  and  rabbits. 

According  to  Biedl  and  Kraus  this  may  amount  to  almost  a  pre- 
vention of  the  coagulation  in  anaphylactic  dogs. 

As  in  other  animals  the  blood  picture  is  changed  in  that  there  is 
a  falling  off  of  the  total  number  of  leukocytes  with  a  relative  diminu- 
tion of  polynuclear  cells. 

Quantitative  measurements  by  Calvary,29  moreover,  have  shown 
that  anaphylaxis  in  dogs  is  accompanied  by  a  marked  increase  of  the 
lymph  flow  ( 7  times  the  amount  observed  in  normal  dogs  in  the  same 
time)  and,  by  controlling  the  blood  pressure  with  barium  chlorid, 
that  this  lymphagogue  action  is  not  directly  dependent  upon  the  low 
pressure.  This  observation  is  of  especial  interest  in  connection  with 
the  similarity  of  anaphylaxis  to  peptone  poisoning  in  which  Heiden- 
hcim  30  noticed  a  similar  increase  of  the  lymph. 

Pearce  and  Eisenbrey  found,  at  autopsy  of  dogs  dead  of  anaphy- 
lactic shock,  subserous  petechial  hemorrhages  in  the  rectum  and  gall 
bladder,  hemorrhagic  spots  on  the  gastric  and  duodenal  mucosa,  and 
in  the  colon.  According  to  these  workers,  in  agreement  with  Biedl 
and  Kraus,  the  fall  of  blood  pressure  is  not  due  to  central  causes  but 
depends  upon  influences  exerted  upon  the  peripheral  vasomotor  sys- 
tem. Biedl  and  Kraus  believe  that  this  action  is  exerted  upon  the 
muscle  cells  themselves  rather  than  on  the  nerve  endings.  They 
admit  the  inconclusiveness  of  their  experimental  data,  but  take  the 
above  standpoint  because  of  the  fact  that  adrenalin,  which  acts  by 
stimulation  of  the  vasomotor  nerve  endings  particularly,  does  not 
raise  the  low  pressure  in  dogs  during  anaphylaxis  while  barium 
chlorid,  which  acts  upon  the  smooth  muscle  fibers  themselves,  strongly 
raises  the  blood  pressure  in  such  animals.  Pearce  and  Eisenbrey  are 
inclined  to  believe  that  the  action  is  chiefly  upon  the  nerve  endings, 
though  both  factors,  nerve  and  muscle,  may  be  involved.  They 
worked  with  apocodein,  a  substance  which,  in  large  doses,  paralyzes 
the  vasomotor  nerve  terminals.31 

When  a  sensitized  dog  was  treated  with  apocodein  and  the  anti- 
gen then  injected,  no  further  drop  of  pressure  was  obtained.  Appar- 
ently a  paralysis  of  the  vasomotor  nerve  endings  had  removed  the 
point  of  attack  upon  which  the  anaphylactic  poison  could  act. 

In  addition  to  the  symptoms  already  enumerated  Weichhardt  and 

20  Calvary.     Munch,   med.   Woch.,  No.   13,  1911. 
30  Heidenheim.    P finger's  Archiv,  49,  1891. 
31Brodie  and  Dixon.     Jour,  of  Phys.}  30,  1904. 


370  INFECTION  AND  RESISTANCE 

Schittenhelm  32  claims  that  anaphylaxis  in  dogs  is  invariably  accom- 
panied by  a  severe  local  reaction  in  the  gut.  The  intestinal  rnucosa 
is  swollen  and  contains  miliary  hemorrhages  and  the  lumen  is  often 
filled  with  a  mucus  mixed  with  blood.  In  the  further  analysis  of  the 
anaphylactic  reaction  in  dogs,  Manwaring  33  has  recently  reported 
observations  of  great  interest.  He  investigated  the  participation  in 
anaphylactic  shock  of  the  various  organs  and  determined  that  shock 
did  not  occur  when  the  abdominal  vessels  were  ligated  just  above 
the  diaphragm.  In  further  localizing  the  source  of  shock  he  found 
that  exclusion  of  the  spleen,  stomach,  kidneys,  suprarenals,  and 
ovaries  from  the  circulation  had  no  effect  upon  the  occurrence  of 
anaphylactic  shock.  However,  when  he  operated  in  such  a  way  that 
the  liver  was  thrown  out  of  circulation,  none  of  the  seven  dogs  that 
he  used  reacted  with  anaphylactic  shock  to  the  injection  of  serum. 
He  concludes  from  this  that  the  liver  is  directly  responsible  in  some 
way  for  the  production  of  anaphylaxis.  The  intestines,  too,  were 
found,  by  a  similar  procedure,  to  take  part,  though  to  a  less  important 
extent  than  the  liver. 

I  Other  animals  than  those  mentioned  have  been  little  used  for 
^anaphylactic  experiment.  Observations  incidental  to  other  work, 
however,  have  shown  that  horses  and  goats  are  particularly  sensitive. 
In  goats  the  writer  has  observed  both  serum  and  bacterial  anaphy- 
laxis, and  the  symptoms  here  were  those  of  general  trembling,  weak- 
ness, labored  respiration,  and  involuntary  evacuation  of  urine. 

The  occurrence  of  anaphylaxis  in  man  will  be  discussed  in  a 
subsequent  section. 

Strange  as  it  may  seem,  it  is  nevertheless  a  fact  which  we  must 
emphasize  that  the  occurrence  of  anaphylactic  shock  is  not  physi- 
ologically identical  in  different  species  of  animals.  The  phenomenon 
in  guinea  pigs  is  unlike  that  in  rabbits.  The  isolated  rabbit  uterus 
will  not  react  when  the  guinea  pig  uterus  does.  The  rabbit  heart 
seems  to  react  when  the  guinea  pig  heart  does  not,  and  the  character- 
istic pulmonary  picture  of  the  guinea  pig  is  entirely  lacking  in  the 
rabbit.  In  dogs,  again,  the  liver  seems  to  be  the  organ  most  definitely 
involved,  and  intestinal  symptoms  are  particularly  severe.  It  is  en- 
tirely unsafe,  therefore,  to  draw  conclusions  by  analogy  from  one  ani- 
mal to  another,  and  naturally,  also,  it  is  unwarranted  to  apply  directly 
to  man  conditions  revealed  by  animal  experimentation. 

The  manifestations  of  "active  anaphylaxis  "  therefore,  consist  in 
the  profound  physiological  changes  occurring  in  animals  when  rein- 
jected  after  a  definite  interval  with  certain  substances  which,  on  first 
injection,  were  practically  harmless.  The  factors  which  are  of  funda- 
mental importance  in  determining  the  development  of  this  hypersus- 
ceptible  or  anaphylactic  state  consist  in  the  nature  of  the  injected  sub- 

32  Weiehhardt  and  Schittenhelm.    Deutsche  med.  Woch.,  19,  1911. 

33  Manwaring.    Zeitschrift,  f.  Immun.,  Vol.  18,  1911. 


ANAPHYLAXIS  371 

stance,  the  quantity  injected,  and  the  interval  between  administra- 
tions.    To  a  great  extent,  too,  the  violence  of  the  reaction  is  depend- 
,ent  upon  the  path  by  which  the  particular  substance  enters  the  body. 

Each  of  these  factors,  therefore,  requires  detailed  consideration 
before  we  can  intelligently  proceed  with  a  further  analysis  of  the 
condition. 

The  substances  with  which  animals  may  be  sensitized  are,  in  all 
particulars,  identical  with  the  class  of  substances  which  we  have  char- 
acterized as  "antigens."  In  fact,  up  to  the  present  time,  there  has 
not  been  a  single  authenticated  exception  to  this,  and  from  our  pres- 
ent understanding  of  the  mechanism  of  anaphylaxis  we  may  safely 
predict  that  no  such  exceptions  will  be  found.  It  is  the  large  class  of 
proteins,  therefore,  whatever  their  source,  which  may  act  as  the  "ana- 
phy lactic  antigens."  However,  in  this  connection  as  well  as  in  the 
larger  problem  of  the  nature  of  antigens  in  general,  it  has  been  diffi- 
cult to  decide  whether  or  not  the  antigenic  property  is  entirely  con- 
fined to  proteins  or  whether  other  substances,  such  as  the  lipoids, 
must  be  included  in  the  definition.  The  problem  has  been  the  same 
here  as  in  other  serum  phenomena,  but  much  special  experimenta- 
tion has  been  done  upon  the  question  with  particular  reference  to 
anaphylaxis  and  the  possibility  of  sensitizing  animals  with  lipoids. 

As  in  the  case  of  similar  investigations  in  regard  to  antibody  for- 
mation, the  results  obtained  in  this  work  have  been  somewhat  confus- 
ing. Pick  and  Yamanouchi  34  extracted  beef  and  horse  sera  with  alco- 
hol, and  evaporated  and  redissolved  the  solutions  until  neither  con- 
tained coagulable  protein  nor  gave  the  Biuret  reaction.  With  this  ma- 
terial they  obtained  a  few  positive  anaphylactic  experiments.  Simi- 
larly curious  are  the  results  of  Bogomolez,35  who  succeeded  in  sensi- 
tizing and  producing  shock  with  the  lipoids  extracted  from  egg  yolks. 
Although  such  experiments  would  tend  to  persuade  us  that  lipoidal 
substances  may  actually  have  sensitizing,  and  therefore  antigenic, 
functions,  this  does  not  follow  necessarily.  As  Pick  and  Yamanouchi 
themselves  point  out,  it  is  practically  impossible  to  demonstrate  with 
certainty  the  presence  of  slight  traces  of  proteins  as  impurities  in  li- 
poid  preparations,  and  we  know  especially  from  Rosenau  and  Ander- 
son's work  how  minute  are  the  quantities  of  antigen  which  still  serve 
to  sensitize.  It  is  possible,  moreover  (a  thought  developed  particu- 
larly by  Pick  and  Schwartz36  and  by  Landsteiner  37),  that  we  are 
dealing  in  many  cases  with  combinations  of  protein  and  lipoid — a 
form  of  chemical  substance  of  which  very  little  is  known  analytically, 
but  the  existence  of  which  many  biological  facts  lead  us  to  assume. 

That^h&cmaphAilacMr,  reaction  is  specific  we  have  mentioned  in 

34  Pick  and  Yamanouchi.    Zeitschr.  f.  Immunitatsforschung,  Vol.  1,  1909. 

35  Bogomolez.     Zeitschr.  f.  Immunitatsforschung,  Vols.  5  and  6,  1910. 

36  Pick  and  Schwartz.     Biochem.  Zeitsch.,  15,  1909. 

37  Landsteiner.     Referat.   "Weichhardt's  Jahresbericht,"   6,  1910. 


372  INFECTION  AND  RESISTANCE 

the  brief  summary  we  have  given  of  Rosenau  and  Anderson's  work. 
These  authors  use  the  adjective  "quantitative,"  by  which  they  simply 
mean  to  convey  that  the  specificity  here  is  not  absolute,  any  more  than 
it  is  absolute  in  the  case  of  any  of  the  known  serum  reactions.  An 
animal  sensitized  with  a  certain  variety  of  protein,  animal  serum,  etc., 
reacts  with  disproportionately  greater  delicacy  to  a  second  injection 
of  the  same  variety  than  of  any  other  substance.  In  fact,  apiart  from 
a  few  cases  mentioned  by  Gay  and  Southard,  there  are  not  many  in- 
stances of  marked  non-specific  anaphylactic  reactions.  Still  we  would 
expect  here,  as  in  other  serum  reactions,  a  certain  limitation  in  the 
degree  of  specificity,  and  Otto  recommends  the  less  delicate  subcu- 
taneous method  of  testing  for  all  experiments  in  which  questions  of 
specificity  are  involved.  This  point  we  will  touch  upon  a  little  later. 

An  interesting  addition  to  our  knowledge  of  such  specificity  was 
made  by  experiments  of  Rosenau  and  Anderson,38  which  showed  that 
a  guinea  pig  could  be  rendered  separately  sensitive  at  one  and  the 
same  time  to  blood  serum,  eggwhite,  and  milk,  reacting  specifically 
to  each  on  second  injection. 

In  anaphylaxis,  again  analogous  to  antibody  reactions  in  general, 
the  specificity,  as  a  rule,  is  one  of  species.  In  other  words,  the  pro- 
tein of  any  animal  is  specific  for  the  proteins  of  its  particular  spe- 
cies generally,  there  being  definitely  similar  characteristics  in  the 
body  proteins  of  animals  of  like  species  which,  though  chemically 
indefinable,  are  nevertheless  delicately  determinable  by  biologic  reac- 
tions. In  considering  specificity  of  precipitins,  however,  we  have 
seen  that  there  are  exceptions  to  the  specificity  of  species  expressed 
in  the  phenomenon  of  so-called  organ  specificity.  The  same  thing- 
has  been  shown  for  anaphylaxis.  Kraus,  Doerr,  and  Sohma  39  were 
able  to  show  that  animals  sensitized  with  protein  from  the  crystalline 
lens  were  hypersusceptible  to  lens  protein  generally,  whether  this 
came  from  the  species  from  which  the  original  lens  was  taken,  or 
whether  some  other  variety  of  animal  had  furnished  it.  On  the  other 
hand,  animals  so  sensitized,  while  hypersusceptible  to  lens  protein 
generally,  did  not  react  to  injections  of  homologous  blood.40  In 
other  words,  this  organ  contains  a  characteristic  variety  of  antigen 
(protein)  peculiar  to  this  kind  of  organ  throughout  the  different 
animal  species,  but  not  common  to  other  tissues  and  organs  of  the 
same  animal.  Results  similar  to  these  were  obtained  by  von  Dun- 
gern  and  Hirschfeld 41  in  the  case  of  testicular  protein,  although 
here  the  phenomenon  seemed  to  be  less  rigidly  organ-specific  than 
in  the  preceding  case.  These  writers  worked  not  with  the  systemic 

38  Rosenau  and  Anderson.    Jour.  Inf.  Dis.,  Vol.  4,  1907. 

39  Kraus,  Doerr,  and  Sohma.     Wien.  klin.  Wocli.,  No.  30,  1908. 
40Andrejew.    Arb.  a.  d,  kais.  Gesundh.  Ami.,  Vol.  30,. 1909. 

41  von   Dungern   and   Hirschfeld.     Zcitschr.   f.   Immunitatsforschung,  4, 
1910. 


ANAPHYLAXIS  373 

anaphylactic  reaction,  but  with  the  localized  (allergic)  reaction 
described  above  as  the  phenomenon  of  Arthus.  They  injected  ex- 
tracts of  the  testicular  materials  into  the  ears  of  rabbits  and  inci- 
dentally made  the  very  curious  observation  that  pregnant  females 
would  not  infrequently  react  to  a  first  injection  without  previous 
sensitization. 

Of  great  importance  also  in  connection  with  the  subject  of  organ 
specificity  is  the  further  discovery  by  Uhlenhuth  and  Haendel42 
that  animals  can  be  sensitized  with  their  own  lens  protein,  a  fact 
which  opens  the  possibility  of  other  forms  of  "autosensitization"  and 
consequently  of  much  opportunity  for  clinical  speculation.  Rosenau 
and  Anderson,43  indeed,  have  found  that  guinea  pigs  can  be  sensi- 
tized by  means  of  extracts  of  guinea  pig  placenta.  They  have  applied 
this  to  the  possible  explanation  of  eclampsia,  and  similar  reasoning, 
as  we  shall  see,  has  been  utilized  in  many  other  conditions.  At- 
tempts have  also  been  made  to  show,  by  the  anaphylactic  reaction, 
that  the  tissue  of  malignant  tumors  possess  such  ''tissue-specific"  or 
"organ-specific"  qualities.  Yamanouchi,44  indeed,  claims  to  have 
shown  this,  but  his  results  were  not  confirmed  by  Apolant,45  and  the 
writer  has  carried  out  a  series  of  entirely  negative  experiments  upon 
the  same  subject.  However,  in  view  of  the  great  difficulty  of  obtain- 
ing any  kind  of  anaphylactic  reaction  in  mice,  the  animals  in  which 
the  tumor  experiments  were  carried  out,  there  is  little  information  to 
be  obtained  from  negative  results  of  this  kind. 

The  delicate  quantitative  method  of  studying  problems  of  speci- 
ficity, which  the  reaction  of  anaphylaxis  supplies,  has  further  served 
to  revive  the  unsettled  question  of  the  "organ-specific"  properties  of 
the  tissues  of  such  organs  as  the  liver,  spleen,  kidney,  blood,  etc. 
Indeed,  Pfeiffer  46  has  published  results  which  would  seem  to  en- 
courage the  belief  of  the  existence  of  such  specificity.  However, 
Ranzi 47  had  previously  obtained  entirely  negative  results,  and 
Pearce,  Karsner,  and  Eisenbrey 48  have  recently  made  a  careful 
inquiry  into  the  same  problem  with  similar  failure  to  determine  such 
organ-specificity. 

In  this,  then,  as  well  as  in  other  respects,  the  substances  by  wkich\ 
animals  may  ~be  sensitized  are  entirely  similar  to  antigens  in  general.  J 

The  substances  which  sensitize,  therefore,  are  those  which  have' 
the  property  of  antibody  formation,  a  statement  self-evident  from 

42  Uhlenhuth  and  Haendel.     Zeitschr.  f.  Immunitdtsforschung,  4,  1910. 
43Rosenau  and  Anderson.     U.  S.  Pub.  Health  and  M.  H.  S.  Hyg.  Lab. 
Bull.,  45. 

44  Yamanouchi.    C.  E.  de  la  Soc.  Biol.,  Vol.  66.  1009,  p.  754. 

45  Apolant.     Zeitschr.  f.  Immunitatsforschung,  Vol.  3,  1909. 

46  Pfeiffer.     Zeitschr.  f.  Immunitatsforschung,  Vol.  8,  1910. 

47  Ranzi.     Zeitschr.  f.  Immunitatsforschung,  Vol.  2,  1909. 
48Pearce,  Karsner,  and  Eisenbrey.    Jour.  Exp.  Med.,  Vol.  14,  1911. 


374  INFECTION  AND  RESISTANCE 

what  has  been  said  before,  but  which  is  again  emphasized  because  of 
its  very  important  bearing  upon  later  theoretical  considerations. 

Variations  in  experimental  anaphylaxis  are,  to  some  extent,  de- 
pendent upon  the  manner  in  which  the  antigen  is  introduced  into  the 
body.  It  is  now  well  known  that  sensitization  may  be  accomplished 
by  a  first  injection  given  subcutaneously,  intravenously,  intraperi- 
toneally,  or  intrapleurally.  At  the  second  or  toxogenic  administra- 
tion shock  may  be  probably  best  induced  and  with  the  smallest 
quantities  by  the  intravenous  method.  Besredka  and  Stein- 
hardt,49  50  51  52  who  began  their  studies  soon  after  the  first  publica- 
tions of  Eosenau  and  Anderson,  came  to  the  conclusion  that  the  most 
effectual  and  rapid  method  of  producing  the  anaphylactic  shock 
consisted  in  direct  injection  into  the  brain.  Curiously  enough,  while 
Besredka  and  Steinhardt  obtained  the  most  violent  reactions  by 
injection  of  the  second  or  toxogenic  dose  into  the  brain,  they  were 
unable  to  sensitize  by  this  path,  at  least  with  doses  of  1/4000  c.  c., 
which  sufficed  to  sensitize  by  the  intravenous  method.  Rosenau  and 
Anderson,  in  repeating  this  work,  obtained  similar  results  with  very 
minute  amounts,  but  found  that  intracerebral  sensitization  could  be 
accomplished  by  doses  of  .0001  c.  c.,  or  more.  According  to  them, 
animals  intracerebrally  sensitized  become  anaphylactic  more  rapidly 
than  those  in  which  the  injections  were  subcutaneous.  In  the 
former  the  incubation  time  was  about  7  days,  while  in  the  latter  it 
was  never  less  than  9.  Lewis,53  in  his  thorough  study  on  the  same 
subject,  made  extensive  use  of  the  direct  intracardial  method  of 
injection.  In  other  words,  any  method  of  introducing  the  foreign 
protein  into  the  blood  or  tissues  seems  to  lead  both  to  sensitization 
and  to  toxic  effect,  and  those  methods  which  introduce  the  substance, 
on  reinjection,  directly  into  the  blood  stream  or  the  brain  induce 
the  most  violent  symptoms  with  the  relatively  smallest  dosage.  Ac- 
cording to  Otto  and  others,  the  subcutaneous  method,  while  followed 
by  less  violent  symptoms,  is  the  method  to  be  preferred  when  ques- 
tions of  specificity  are  involved,  for,  while  the  reaction  is  specific  in 
the  ordinary  sense,  yet  it  is  extremely  delicate  and  therefore,  as 
Eosenau  and  Anderson  put  it,  "quantitatively  specific.'7  The  less 
violent  subcutaneous  method,  therefore,  might  be  said  to  have  the 
same  purpose  here  that  dilution  of  the  antigen  or  immune  serum  has 
in  safeguarding  against  error  when  carrying  out  specific  precipitin  or 
iigglutinin  reactions. 

49  Besredka  and  Steinhardt.     Ann.  de  VInst.  Past.,  p.  117,  1907;  ibid., 
p.  384. 

50  Besredka.    Ibid.,  p.  777,  p.  950,  1907;  ibid.,  p.  496,  1908;  p.  166,  1909; 
p.  801,  1909. 

51  Also:   Bull,  de  VInst.  Past.,  Nos.  19,  20,  21,  1908;  No.  17,  1909. 

52  Also:    C.  E.  de  la  Soc.  Biol,  p.  478,  1908,  Vol.  65;  p.  266,  1909,  Vol. 
67. 

53  Lewis.    Jour.  Exp.  Med.,  Vol.  10,  1908. 


ANAPHYLAXIS  375 

Whether  or  not  sensitization  can  be  accomplished  by  introduction 
of  the  antigen  into  the  intestinal  canal,  feeding,  in  other  words,  is 
still  to  some  extent  an  open  question  and  of  great  importance  in 
view  of  the  many  clinical  manifestations  (urticaria,  albuminuria, 
etc.)  which  are  attributed  to  possible  individual  hypersusceptibility 
to  certain  proteins  taken  in  the  diet  (idiosyncrasies).  Rosenau  and 
Anderson,  in  their  earliest  paper,  report  success  in  sensitizing  guinea 
pigs  by  the  feeding  of  horse  meat  and  horse  serum.  McClintock  and 
King  54  failed  to  confirm  this,  and  the  observations  of  other  writers 
seem  to  bear  them  out.  However,  when  we  consider  that  Ascoli, 
Oppenheimer,  and  others  have  shown  that  proteins  fed  to  animals  in 
large  quantities  may  be  subsequently  demonstrated  not  only  in  the 
circulating  blood  but  occasionally  even  in  the  urine  by  means  of  the 
precipitin  reaction,  there  seems  to  be  little  room  for  doubting  that 
antigen  may  enter  the  circulation  unchanged,  though  possible  only 
under  abnormal  local  conditions  of  the  intestine.  This,  together  with 
Eosenau  and  Anderson's  demonstration  of  the  extremely  small  amount 
of  antigen  necessary  to  sensitize,  furnishes  all  the  conditions  n( 
sary  for  anaphylaxis  by  way  of  the  intestinal  canal. 

A  study  made  by  Lesne  and  Dreyfus  55  seems  to  us  to  have  ex- 
plained the  contradictory  results  of  other  workers  on  this  phase  of 
the  problem.  Without  being  able  to  associate  the  destruction  of  the 
sensitizing  function  with  either  the  gastric  or  pancreatic  secretions, 
they  were  nevertheless  successful  in  showing  that  sensitization  could 
be  carried  out  regularly  if  the  antigen  were  injected  after  laparotomy 
into  the  large  intestine,  whereas  similar  injections  into  the  stomach  or 
small  intestine  wrere  negative.  In  these  experiments  we  must  take 
into  consideration  that  the  conditions  following  laparotomy,  such  as 
temporary  intestinal  atony  and  congestion,  may  have  exerted  con- 
siderable influence  upon  the  positive  outcome  of  their  large  intestine 
injections.  Whereas  they  do  not,  therefore,  permit  us  to  assume  the 
possibility  of  sensitization  through  the  normal  alimentary  canal,  they 
nevertheless  confirm  the  assumption  of  the  possibility  of  sensitiza- 
tion by  this  path  under  the  influence  of  slightly  abnormal  local  con- 
ditions. 

In  this  connection  Besredka's  56  experiments  on  the  production 
of  anti-anaphylaxis  by  the  intestinal  administration  of  protein  are 
of  interest.  He  found  that  if  sensitized  animals  were  given  5  c.  c. 
of  the  antigen  (milk)  by  rectum,  they  were  thereby  protected  from 
the  reaction  following  in  controls  upon  a  second  injection.  In  his 
later  experiments  with  egg  white  it  appeared  that  the  protection 
could  also  be  conferred  by  mouth,  but  that  in  this  case  it  developed 

54  McClintock  and  King.    Jour.  Inf.  Dis.,  3,  1906.    See  section  on  normal 
antibodies. 

65  Lesne  and  Dreyfus.     C.  E.  de  la  Soc.  BioL,  Vol.  70,  p.  136,  1911. 
66Besredka.    C.  E.  de  la  Soc.  Biol.,  Vol.  65,  1908;  Vol.  70,  1911. 


376  INFECTION  AND  RESISTANCE 

more  slowly,  it  being  necessary  to  wait  two  days  after  ingestion  be- 
fore the  anti-anaphylaxis  had  developed  sufficiently  to  protect.  Since 
attempts  by  mouth  were  not  as  rapidly  successful  as  those  per  rec- 
tum, it  is  clear  that  these  facts  are  in  keeping  with  Lesne  and  Drey- 
fus' results  in  showing  that  the  antigen  is  probably  absorbed  chiefly 
or  solely  from  the  large  intestine.  In  Lesne  and  Dreyfus'  experi- 
ments the  sensitizing  dose  was  given  into  the  intestine,  the  toxogenic 
or  second  dose  being  administered  intravenously,  and  since,  as  we 
shall  see,  minute  doses  may  suffice  to  sensitize,  whereas  100  or  more 
times  the  sensitizing  amount  is  necessary  to  produce  intoxication,  it 
is  easy  to  understand  why  sensitization  followed  in  Lesne  and  Drey- 
fus' work,  but  no  toxic  effects  followed  in  the  experiments  of  Bes- 
redka.  Furthermore,  the  slow  absorption  from  the  intestine  in  these 
experiments  explains  the  development  of  anti-anaphylaxis  in  Bes- 
redka's  work,  in  that  they  are,  in  this  respect,  analogous  to  later 
experiments  of  Friedberger,  cited  below,  in  which  it  was  shown  that- 
sensitized  guinea  pigs,  which  could  (in  controls)  be  killed  by  rapid 
intravenous  injection  of  0.1  c.  c.  of  antigen  and  less,  would  withstand 
without  symptoms  many  times  this  amount  when  it  was  gradually 
administered  by  slow  injection  covering  an  hour  or  longer. 

In  referring  to  the  quantities  of  antigen  by  which  sensitization 
may  be  accomplished,  we  have  already  called  attention  to  the  very 
small  amounts  which  have  been  found  sufficient  for  this  purpose. 
There  seems,  indeed,  to  be  a  wide  latitude  in  this  regard,  the  re- 
quired quantities  ranging  from  as  little  as  a  millionth  of  a  cubic 
centimeter  (Rosenau  and  Anderson)  to  as  much  as  10  c.  c.  or  more. 
rw  oonnTirj  ipj^tinri;  however,  trmc  effects^are  never  produced  by 

q-nflyit.it.ipjp  as  -minute  as  thoSft  TyVnVh   anffW  for 


tual  amounts  exists.  An  important 
problem,  moreover,  is  the  relation  which  has  been  said  to  exist  be- 
tween the  sensitizing  dose  and  the  interval  necessary  for  the  devel- 
opment of  the  hypersusceptible  state  (anaphylactic  incubation  time). 
In  their  first  publications,  Rosenau  and  Anderson,  Otto,  and  others 
expressed  the  opinion  that  the  length  of  incubation  time  was  in- 
versely proportionate  to  the  size  of  the  sensitizing  dose;  in  other 
words,  animals  sensitized  with  small  quantities  (0.01  c.  c.  or  less) 
would  become  hypersusceptible  and  react  to  a  second  injection  in 
from  8  to  12  days,  whereas  animals  receiving  two,  three,  or  more 
cubic  centimeters  of  the  antigen  would  take  weeks  or  months  to  be- 
come anaphylactic.  The  same  opinion  was  expressed  by  Otto,57  and 
is  now  generally  found  in  the  literature.  Later  experiments  of 
Rosenau  and  Anderson,58  however,  have  seemed  to  show  that  this 

57  Otto.    Loc.  cit.     See  also  in  "Kolle  u.  Wassermann  Handbuch,"  Ergan- 
zungsband  II,  p.  241. 

58  Rosenau  and  Anderson.     U.  S.  Pub.  Health  and  M.  II.  S.  Hyn.  Lab. 
Bull.  45,  1908. 


ANAPHYLAXIS  377 

relation  is  not  as  definite  as  at  first  assumed.  In  the  tables  given 
by  them  guinea  pigs  receiving  0.01  c.  c.  reacted  severely  after  14,  17, 
and  155  days;  others,  receiving  1  c.  c,,  after  14,  17,  and  155  days; 
and,  again,  another  series  sensitized  with  8  c.  c.  reacted  severely 
after  similar  intervals.  All  of  these  series  reacted  but  mildly  after 
245  days,  showing  apparently  that  the  anaphylaxis,  contrary  to  gen- 
eral belief,  does  not  last  so  much  longer  after  the  larger  than  after 
the  smaller  sensitizing  doses. 

These  experiments,  however,  as  well  as  similar  ones  by  other 
workers,  have  shown  that,  once  sensitized,  animals  may  remain  so 
for  very  extensive  periods.  In  the  work  of  Rosenau  and  Anderson  59 
the  limit  of  horse  serum  sensitization  was  245  days.  A  few  guinea 
pigs,  sensitized  with  toxin-antitoxin  mixtures,  gave  positive  reac- 
tions after  732  days;  more  recently  they  have  obtained  a  reaction 
after  1,096  days.60 

This  seems  to  indicate  that  perhaps  individuals  once  sensitized 
may  remain  so  for  life.  And,  as  we  shall  see,  a  hyper  susceptible 
subject,  temporarily  desensitized  by  antigen,  will  again  gradually^ 
become  sensitive. 

In  properly  sensitized  animals  the  result  of  a  sufficient  dose  of 
antigen,  given  at  the  proper  time,  is  very  often  death.  When  the 
time  and  quantity  are  so  chosen  that  instead  of  death  there  is  merely 
a  more  or  less  severe  anaphylactic  shock,  the  animals  are  immediately 
thereafter  in  a  refractory  condition.  That  is,  they  are  no  longer 
sensitive  to  further  injections  of  the  antigen.  This  observation  was 
made  by  Otto  and  by  Rosenau  and  Anderson  in  their  pioneer  inves- 
tigations, was  confirmed  by  Gay  and  Southard,  and  was  subsequently 
very  thoroughly  studied  by  Besredka  and  Steinhardt.61  The  last- 
named  workers  named  this  refractory  or  immune  condition  "anii- 
anapliylaxis."  There  is  obviously  a  great  deal  of  both  practical  and 
theoretical  significance  in  this  fact,  and  methods  were  sought  by 
which  such  an  anti-anaphylactic  state  might  be  induced  in  sensitized 
animals  without  subjecting  them  to  the  dangers  of  actual  shock.  It 
was  found  that  this  could  be  accomplished  in  a  number  of  ways. 
According  to  Besredka  and  Steinhardt  the  injection  of  moderate 
quantities  of  the  antigen  at  a  time  just  preceding  the  development  of 
hypersusceptibility,  in  the  preanaphylactic  period,  will  render  them 
refractory  to  later  injections.  This  preventive  administration,  how- 
ever, must  be  given  during  the  later  days  of  the  anaphylactic  incu- 
bation time.  If  given  too  soon  after  the  first  injection  it  does  not 
prevent  eventual  sensitization,  though  it  may  occasionally  delay  its 

80Kosenau  and  Anderson.     U.  S.  Pub.  Health  and  M.  H.  S.  Hyg.  Lab. 
Bull,  50,  1909. 

60  They  express  the  belief  from  this  that  a  guinea  pig  may  remain  sensi- 
tive throughout  life. 

61  Besredka  and  Steinhardt.    Loc.  cit. 


378  INFECTION  AND  RESISTANCE 

development,  acting  then  simply  as  though  a  larger  dose  had  been 
given  in  the  first  place.  Thus  if  antigen  is  given  by  a  method  of 
introduction  and  in  a  quantity  which  would  justify  us  in  expecting 
hypersusceptibility  to  be  developed  at  the  end  of  12  days,  we  can 
render  the  animal  "antianaphylactic"  by  a  second  administration 
given,  say,  on  the  8th,  9th,  or  10th  day.  If  we  give  it  on  the  2d,  3d, 
or  4th  day  after  the  first  injection,  it  is  very  likely  that  sensitization 
will  proceed  nevertheless.  Eosenau  and  Anderson  have  also  investi- 
gated the  repeated  injection  of  antigen  during  the  incubation  time, 
and  their  results  would  also  seem  to  emphasize  the  necessity  of  mak- 
ing the  preventive  injection  close  to  the  time  at  which  hypersuscepti- 
bility may  be  expected.  If  quantities  of  2  c.  c.  were  injected  10  times 
in  the  course  of  17  days,  and  15  to  17  days  thereafter  6  c.  c.  of  horse 
serum  were  given,  the  animals  showed  symptoms  proving  that  anti- 
anaphylaxis  was  but  partial.  If  amounts  of  0.001  c.  c.  were  given 
5  times  in  a  period  of  8  days,  and  the  animals  were  tested  23  days 
later,  death  often  ensued.  It  is  also  possible,  as  a  number  of  investi- 
gators have  shown,  to  produce  the  antianaphy lactic  state  by  the  injec- 
tion of  sublethal  doses,  even  after  the  time  has  set  in  at  which  the  ani- 
mals are  hypersusceptible.  '  This  method  can  be  carried  out  success- 
fully according  to  Besredka  by  injecting  very  small  amounts  into  the 
brain  (1/50  to  1/400  of  a  cubic  centimeter).  Within  a  few  hours 
after  such  an  injection  the  animals  may  withstand  an  otherwise 
fatal  dose  with  slight  or  no  symptoms;  although  it  is  generally 
stated  that  intraperitoneal  injections,  carried  out  after  hypersus- 
ceptibility has  set  in,  must  be  of  considerable  quantity  (large  enough 
to  cause  symptoms)  in  order  to  induce  antianaphylaxis.  Besredka  62 
states,  in  a  recent  resume,  that  1/50  to  1/100  cubic  centimeter  in- 
jected intraperitoneally  and  giving  "practically  no  symptoms"  in  a 
sensitized  guinea  pig,  after  the  anaphylactic  state  has  set  in,  may  ren- 
der the  animal  entirely  refractory  after  5  hours. 

On  the  other  hand,  Rosenau  and  Anderson,63  working  with  sub- 
cutaneous injection,  obtained  results  which  differ  considerably  from 
those  of  Besredka.  They  sensitized  a  series  of  guinea  pigs  with 
mixtures  of  toxin  and  antitoxin,  and  48  days  later,  at  a  time  when 
the  animals  were  hypersusceptible,  gave  20  subcutaneous  injections 
of  0.001  c.  c.  daily.  Two  days  after  the  last  injection,  0.2  c.  c.  of 
horse  serum  was  given  intracerebrally,  and  all  of  the  animals  showed 
symptoms,  and  many  of  them  died.  They  conclude,  therefore,  that 
the  repeated  injection  of  small  amounts  of  antigen  into  sensitized  ani- 
mals has  no  appreciable  effect.  The  same  worker  has  shown  by  ex- 
periment that  the  introduction  of  large  amounts  of  antigen  into  the 
previously  cleansed  rectum  of  sensitive  animals  is  entirely  without 

62  Besredka  in  "Kraus  u.  Levaditi  Handbuch,"  Erganzungsband  I. 

63  Rosenau  and  Anderson.     Loc.  cit.,  U.  S.  Pub.  Health  and  M.  H.  S. 
Hyg.  Lab.  Bull.  45,  1908. 


ANAPHYLAXIS  379 

danger  and  will  produce  an  antianaphylaxis,  which  becomes  evident 
after  12  hours.  This  is  probably  dependent  upon  the  very  slow  pene- 
tration of  small  amounts  of  antigen  into  the  circulation  from  the  gut, 
and  has,  therefore,  an  effect  similar  to  the  repeated  injection  of  small 
amounts  directly,  or  the  very  slow  and  gradual  method  of  intrave- 
nous injection  advocated  by  Friedberger  for  the  prevention  of  serum 
sickness  in  man.  This  phase  of  the  subject  is  considered  in  greater 
detail  in  a  subsequent  discussion  of  serum  sickness. 

This  observation  has  recently  been  confirmed  by  Coca. 

As  far  as  we  can  understand  it  at  the  present  time,  the  desensiX 
tization  depends  upon  a  saturation  of  the  antibodies  in  the  animal! 
with  antigen,  which  temporarily  removes  them  from  possibility  of  I. 
further  reaction.     A  gradual  administration  of  antigen  to  the  sensi-V 
tized  animal,  either  by  fractional  dosage  or  slow  administration,  may 
bring  this  about  without  the  fatal  shock  resulting  from  a  too  sudden    I 
and  violent  reaction.    This  is  the  rationale  of  Besredka's  and  Fried- 
berger's  methods.  * 

Antianaphylaxis  produced  in  this  way  is  specific,64  although,  as 
we  shall  see,  there  are  other  methods  by  which  it  is  claimed  that  a 
nonspecific  antianaphylaxis  can  be  produced.  One  of  these  consists 
in  the  injection  of  anaphy lactic  animals  with  peptone.  The  problem 
of  peptone  poisoning  and  its  relation  to  anaphylaxis  will  receive  sep- 
arate consideration. 

Banzhaf  and  Steinhardt  65  have  reported  that  0.5  gram  of  lecithin 
given  to  sensitized  guinea  pigs  protects  them  against  second  injec- 
tion. Rosenau  and  Anderson  66  have  failed  to  confirm  this. 

The  above  methods  of  rendering  animals  antianaphylactic  apart 
from  the  bearing  they  may  have  on  purely  therapeutic  possibilities, 
serve  to  throw  much  light  upon  the  possible  mechanism  of  the  reac- 
tion within  the  animal  body.  It  is  of  great  interest  for  the  under- 
standing of  the  physiological  conditions  underlying  anaphylaxis  also 
to  consider  briefly  the  influence  upon  anaphylactic  shock  which  may 
be  exerted  by  certain  drugs.  The  preventive  influence  of  atropin 
we  have  already  mentioned  in  connection  with  the  work  of  Auer  and 
Lewis.  Besredka,  who,  as  we  shall  see,  attributes  the  major  part  of 
anaphylactic  manifestations  to  reactions  proceeding  from  the  central 
nervous  system,  claims  to  have  succeeded  in  injecting  ordinarily 
fatal  doses  of  antigen  without  harm  into  guinea  pigs  previously 
anesthetized  with  ether.  Banzhaf  and  Famulener  67  have  similarly 

64  Pfeiffer  has  recorded  an  exception  to  this  in  that  he  claims  to  have 
rendered  a  horse-serum  sensitive  animal  refractory  by  an  injection  of  swine 
serum. 

65  Banzhaf  and   Steinhardt.     Proc.  Soc.  Exp.  Biol   and   Med.,  Vol.   7, 
1910. 

66  Rosenau  and  Anderson.    Hyg.  Lab.  Bull.,  64,  1910. 

6T  Banzhaf  and  Famulener.     Studies  N.  Y.  Dep.  Health  Ees.  Lab.,  1908, 
p.  107. 


380  INFECTION  AND  RESISTANCE 

prevented  shock  by  large  doses  of  chloral  hydrate.  Rosenau  and 
Anderson,08  carrying  on  similar  investigations  with  urethane,  paral- 
dehyd,  chloral  hydrate,  and  magnesium  sulphate,  came  to  the  con- 
clusion that  none  of  these  drugs  has  any  noticeable  effect  upon  ana- 
phylactic  shock  in  guinea  pigs. 

Up  to  the  present  time  AVC  have  confined  ourselves  to  the  descrip- 
tion of  the  basic  anaphylactic  experiment,  which  is  spoken  of  as 
''active  sen&itization"  in  analogy  to  the  expression  "active  immuniza- 
tion,'7 since,  like  the  latter,  it  conveys  the  conception  that  the  state 
of  hypersusceptibility  (like  the  immunity  in  active  immunization)  is 
here  acquired  by  reason  of  physiological  changes  directly  induced  in 
the  treated  animal  in  reaction  to  the  first  injection  of  the  foreign 
antigen.  There  is  another  method  of  inducing  hypersusceptibility 
which,  in  continviance.  of  the  analogy  to  immunization,  is  spoken  of 
as  "passive  anaphylaxis/'  since  it  consists  in  transferring  the  hyper- 
susceptible  condition  to  a  perfectly  normal  animal  by  injecting  into 
it  serum  from  an  actively  sensitized  one.  The  normal  animal  is  thus 
merely  the  passive  recipient  of  the  reaction  bodies  produced  in  the 
sensitive  animal  by  preliminary  treatment. 

That  such  a  passive  transference  of  anaphylaxis  is  possible  was 
shown  by  a  number  of  investigators  almost  simultaneously  and  M. 
Xicolle,69  in  February,  1907,  published  a  study  on  the  phenomenon 
of  Arthus  in  which  he  showed  that,  if  the  serum  of  a  hypersusceptible 
rabbit  (sensitized  with  horse  serum)  was  injected  into  a  normal 
rabbit,  the  recipient  was  rendered  sensitive,  so  that  the  subcutaneous 
injection  of  horse  serum,  made  24  hours  later,  produced  typical 
infiltrations.  Richet 70  soon  after  this  succeeded  in  transferring 
hypersusceptibility  toward  mytilocongestin  (a  mussel  poison)  from 
a  sensitized  to  a  normal  dog  by  injecting  considerable  amounts  of  the 
blood  from  the  former  into  the  latter.  In  this  case,  too,  the  hyper- 
susceptibility of  the  second  dog  did  not  appear  until  one  or  two  days 
after  the  injection  of  the  blood.  At  almost  the  same  time  Otto  71 
and  Friedemann  72  independently  succeeded  in  transferring  serum 
anaphylaxis  from  hypersusceptible  to  normal  guinea  pigs  in  a  similar 
way.  Experiments  of  Gay  and  Southard,73  published  during  the 
same  year,  may  possibly  be  also  interpreted  as  instances  of  passive 
anaphylaxis,  although  their  experimental  procedure  renders  this 
doubtful,  even  in  their  own  opinions.  They  injected  0.1  c.  c.  of 
serum  from  both  sensitive  and  refractory  guinea  pigs  into  normal 
animals  and  followed  this,  after  10  days,  with  injections  of  antigen. 
The  fact  that  such  animals  reacted  may  be  interpreted  in  a  number 

68  Rosenau  and  Anderson.    Jour.  Med.  Ees.,  Vol.  21,  N.  S.,  16,  1909. 

69  M.  Nicolle.    Ann.  de  I'Inst.  Past.,  Vol.  21,  1907. 

70  Richet.    Ann.  de  I'Inst.  Past.,  Vol.  21,  1907. 

71  Otto.    Munch,  med.  Woch.,  No.  34,  1907. 

72  Friedemann.    Munch,  med.  Woch.,  No.  49,  1907. 

73  Gay  and  Southard.    Jour.  Med.  Ees.,  Vol.  16,  1907. 


ANAPHYLAXIS  381 

of  ways.  They  themselves  regarded  the  hypersusceptibility  which 
the  injected  animals  developed  as  a  "purely  active  one/'  and  it  is 
more  than  likely  that  this  was  the  case,  the  recipient  animals  being 
actively  sensitized  by  traces  of  antigen  remaining  unassimilated  in 
the  blood  of  the  actively  sensitized  donors.  In  the  following  year 
(1908)  the  facts  of  passive  sensitization  were  rapidly  confirmed  and 
extended  by  Besredka,74  Lewis,75  and  others,76  and  information  of 
the  greatest  value  for  the  comprehension  of  the  anaphylactic  reaction 
was  obtained. 

Otto  showed  that  passive  sensitization  could  be  carried  out  with 
the  serum  of  an  actively  sensitized  animal  8  days  after  the  antigen 
injection,  at  a  period  when  this  animal  itself  had  not  yet  become 
hypersusceptible.  He  also  showed  that  the  passive  transfer  of  ana- 
phylaxis  need  not  be  confined  to  animals  of  the  same  species,  but  that 
guinea  pigs  could  be  Tendered  passively  anaphylactic  with  the  blood 
serum  of  sensitized  rabbits.  From  the  work  of  Gay  and  Southard,77 
moreover,  it  appeared  that  not  only  by  the  blood  of  sensitive  animals 
can  anaphylaxis  be  transferred,  but  that  this  can  also  be  done  by 
injecting  the  blood  of  animals  that  have  once  been  sensitive  but  have_ 
subsequently  been  rendered  antianaphylactic  or  refractory.  Analo- 
gous to  this  observation  is  the  fact  observed  by  these  authors  as  well 
as  by  Friedemann  that  the  young  of  antianaphylactic  mothers  are  not  _ 
refractory  but  hypersusceptible.  This  observation,  unquestionably 
correct,  since  it  has  been  confirmed  by  several  other  workers,  is 
astonishing  and  contrary  to  expectation.  It  has  had  no  inconsider- 
able bearing  upon  our  theoretical  understanding  of  anaphylaxis. 

It  was  soon  found  out,  too,  that  hypersusceptibility  was  conveyed 
not  only  by  the  sera  of  sensitive  and  of  refractory  animals,  but  that 
it  could  likewise  be  transferred  by  the  precipitating  sera  of  animals 
systematically  immunized  with  a  foreign  proteid. 

This  method  was  later  employed  by  Doerr  and  Russ  78  in  their 
quantitative  studies  on  the  relations  between  anaphylactic  antigen 
and  antibody.  We  are  confronted,  then,  with  the  curious  facts  that 
animals  may  be  passively  sensitized : 

(a)  by  the  serum  of  a  sensitized  animal. 

(b)  by  the  serum  of  an  animal  not  yet  sensitive — in  the  pre- 
anaphylactic  period  (8th  day,  Otto). 

(c)  by  the  serum  of  an  antianaphylactic  animal. 

(d)  by  the  precipitating  serum  of  an  "immunized"  79  animal. 

74  Besredka.     Ann.  de  I'Inst.  Past.,  Vol.  22,  1908. 

75  Lewis.    Jour.  Exp.  Med.,  Vol.  10,  1908. 

76  Kraus  and  Doerr.     Wien.  klin.  Woch.,  No.  28,  1908. 

77  Gay  and  Southard.    Jour.  Med.  Res.,  Vol.  18,  1908. 

78  Doerr  and  Russ.     Zeitschr.  f.  Immunitcitsforschung,  Vol.  3,  1909. 

79  We  must  never  forget  that  the  term  "immunized"  as  applied  to  animals 
treated   with   harmless    protein   is    an    analogy   and   not    absolutely   correct. 
Such  animals,  though  probably  capable  of  assimilating  larger  quantities  of 


S82  INFECTION  AND  RESISTANCE 

Lewis  further  showed  that  normal  guinea  pigs  could  be  rendered 
hypersusceptible  with  the  blood  of  congenitally  sensitive  animals. 

Passive  sensitization  is  carried  by  the  blood  serum  purely,  since, 
in  ordinary  cases,  as  Rosenau  and  Anderson  have  shown,  the  blood 
corpuscles  and  tissues  of  a  sensitive  animal  do  not  convey  the  hyper- 
susceptibility.  An  exception  to  this  will  be  noted  later  when  we 
come  to  discuss  Bail's  experiments  on  the  passive  transfer  of  tuber- 
culin sensitiveness. 

Passive  sensitization,  once  established,  may  persist  for  as  long 
as  3  or  4  weeks,  though  Rosenau  and  Anderson  found  that  animals 
tested  26  days  after  treatment  reacted  but  weakly.  In  the  young 
of  anaphylactic  mothers  Otto  has  observed  positive  reactions  as  long 
as  44  days  after  birth,  though  fatal  results  were  obtained  in  pigs  only 
a  few  days  old. 

o     To  summarize  the  matter  briefly,  we  may  state  that  passive  sen- 
sitization may  be  accomplished  with  any  serum  that  contains  anti- 
/ bodies,  and  that  the  qualitative  power  of  such  a  serum  to  convey 
I  passive  sensitization  is  in  direct  proportion  to  its  antibody  concen- 
I  tration;  in  short  the  process  of  sensitization  consists  in  the  introduc- 
tion of  antibodies.     This  fact,  which  is,  of  course,  of  the  greatest 
I  theoretical  importance,  will  be  further  discussed  in  the  succeeding 
\chapter. 

^  Throughout  the  earlier  investigations  upon  passive  sensitization 
the  curious  fact  recurs  in  the  experiments  of  successive  workers  that 
a  definite  period  must  elapse  between  the  injection  of  the  sensitive 
blood  and  that  of  the  antigen. 

Both  Friedemann  and  Otto  found  that  when  the  sensitive  serum 
was  injected  subcutaneously  the  best  results  were  obtained  by  ad- 
ministration of  the  antigen  24  to  48  hours  after  this.  On  intra- 
peritoneal  injection  of  the  sensitizing  serum  Doerr  and  Russ  80  ob- 
tained the  best  results-  by  permitting  an  interval  of  24  hours  to 
elapse,  and  the  same  investigators  still  further  shortened  this  period 
to  4  hours  by  injecting  the  sensitive  serum  intravenously.  Beyond 
this,  the  interval  could  not  be  shortened  with  success.  Indeed,  some 
writers,  notably  Gay  and  Southard,  have  claimed  that  the  maximum 
hypersusceptibility  in  guinea  pigs  treated  with  sensitive  serum  is 
reached  only  after  10  or  more  days,  and  Rosenau  and  Anderson, 
Lewis,  and  others  have  obtained  results  which  seemed  to  point  in 
the  same  direction.  However,  as  we  have  already  indicated,  the 
testing  of  animals  so  long  after  the  injection  of  sensitive  serum 
leaves  us  in  doubt  whether  we  are  dealing  with  true  "passive"  trans- 
foreign  injected  protein  than  normal  ones,  and  this  more  rapidly,  may  never- 
theless be  not  a  whit  more  tolerant  of  the  antigen — sometimes  even  extremely 
sensitive  and  vulnerable. 

80  Doerr  and  Russ.     Zeitschr.  f.  ImmunitdtsforscJiung,   Vol.   3,   p.   181f 
1909. 


ANAPHYLAXIS  383 

ference  of  anaphylaxis  or  with  active  sensitization  due  to  traces  of 
antigen  carried  over  with  the  serum  of  the  sensitive  animal.  For 
the  purposes  of  theoretical  deduction,  therefore,  it  is  better  to  ignore 
these  cases  and  consider  chiefly  passive  transference  in  which  reac- 
tions are  obtained  within  24  hours  or  less  after  the  injection  of  the 
aiiaphylactic  serum — an  interval  so  short  that  active  sensitization 
can  hardly  be  considered  as  a  reasonable  possibility. 

The  important  point,  in  this  connection,  is  the  fact  that  it  was 
found  that  between  the  administration  of  sensitive  serum  and  of 
antigen  a  definite  interval,  however  short,  was  invariably  necessary.81 

From  these  observations  the  natural  deduction  was  made  that  the 
anaphylactic  symptoms  were  the  result  of  cellular  occurrences,  and 
that  the  antigen  could  act  only  after  the  sensitizing  substance  (how- 
ever conceived)  had  become  attached  to  certain  cells,  probably  to 
those  of  the  central  nervous  system.  It  was  thought  that  a  meeting 
of  antigen  and  the  sensitized  body  in  the  circulation  would  result  in 
no  reaction;  that,  in  other  words,  the  effective  reaction  was  not  a 
direct,  but  an  indirect,  one  after  the  anaphylactic  "antibody"  of  the 
sensitive  serum  had  become  bound  to  the  cells.  It  will  be  neces- 
sary to  recur  to  this  problem  when  we  discuss  the  various  theories 
of  anaphylaxis,  where  we  will  see  that  this  point  has  been  one  of  the 
crucial  ones  in  the  controversy  between  the  two  main  directions  of 
thought  on  anaphylaxis. 

81  An  exception  to  this,  contradicting  the  then  prevailing  opinion,  were 
the  researches  of  Weill-Halle  and  Lamaire  (C.  R.  de  la  Soc.  de  Biol.,  Vol. 
65,  July,  1908,  p.  141),  who  showed  that,  under  certain  conditions,  guinea 
pigs  would  react  with  typical,  often  fatal,  anaphylaxis  if  injected  simul- 
taneously with  the  serum  of  sensitized  rabbits  and  the  antigen  horse  serum. 
According  to  them,  the  success  of  such  experiments  depended  entirely  upon 
the  condition  of  the  sensitive  serum — that  is,  the  time  at  which  the  rabbits 
treated  with  horse  serum  were  bled.  These  experiments,  we  shall  see,  were 
later  confirmed.  We  record  them,  though  important,  in  a  footnote,  since  we 
wish  at  present  to  emphasize  the  reasoning  which  led  to  the  assumption  of  a 
cellular  participation  in  the  reaction. 


CHAPTER   XVI 

ANAPHYLAXIS  (Cont.) 

FURTHER  DEVELOPMENT   AND   THEOEETICAL 
CONSIDERATIONS 

WE  have  now  briefly  considered  some  of  the  fundamental  facts 
which  the  earlier  investigations  upon  anaphylaxis  have  revealed  and, 
although  there  are  still  many  important  observations  to  record,  the 
material  so  far  outlined  will  serve  as  a  basis  for  a  brief  consideration 
of  the  views  that  have  been  formulated  concerning  the  mechanism 
of  aiiaphylactic  phenomena. 

It  is  clear  that  the  chapter  of  anaphylaxis  is  hardly  more  than 
well  begun.  In  the  earlier  stages  of  the  investigations  into  this  prob- 
lem many  opinions  were  advanced  which  served  the  valuable  func- 
tions of  working  hypotheses,  but  were  quickly  altered,  trimmed,  or 
expanded  as  new  and  incompatible  facts  were  revealed  in  astonish- 
ingly rapid  succession.  The  final  solution  is  probably  still  far  be- 
yond our  present  horizon,  but  the  recent  knowledge  of  the  toxic 
derivatives  of  proteins,  "the  anaphylatoxins,"  foreshadowed  by  the 
work  of  Vaughan  and  his  associates,  more  definitely  determined  by 
Eriedemann  and  especially  by  Friedberger,  has  furnished  hope  that 
we  are  not  only  on  the  right  path  toward  understanding  anaphylaxis, 
but  has  given  us  a  new  clue  to  the  correlation  of  this  condition  with 
immunity. 

It  will  greatly  facilitate  exposition  of  the  various  theories  which 
have  been  advanced  if  we  bear  in  mind  that,  although  there  have 
been  many  discrepancies  on  minor  phases,  the  differences  of  opinion 
have  centered  upon  the  cardinal  points. 

These  are:  1.  Is  the  anaphy lactic  phenomenon  a  true  antigen- 
antibody  reaction  in  which  the  sensitizing  injection  gives  rise  to  the 
formation  of  a  specific  antibody  with  which  it  reacts  on  second  injec- 
tion ?  2<  Is  sensitization  the  result  of  effects  exerted  upon  the  tissue 
cells,  which  participate  directly  in  the  reaction,  or  may  the  reaction 
take  place  entirely  in  the  circulation,  the  tissue  cells  being  affected 
secondarily  only  ? 

Upon  these  two  questions  we  can  logically  classify  theories  of 
anaphylaxis. 

384 


ANAPHYLAXIS  385 

Among  the  earliest  definitely  stated  theories  is  that  of  Gay  and  Southard.1  "~ 
These  workers  were  emphatic  in  denying  that  anaphylaxis  has  the  nature  of  an 
antigen-antibody  reaction.     Their  views   are  summarized  in  the  following  para- 
graphs given,  as  nearly   as  space  permits,   in  their  own  words: 

Increased  susceptibility  in  the  sensitized  animal  is  due  to  the  continued 
presence  in  the  circulation  of  an  unneutralized  element  of  the  antigen  (in  their 
case  horse  serum),  which  they  call  ' ( anaphylactin, "  which  acts  as  an  irritant  or 
stimulant  to  the  body  cells,  and,  in  some  way,  causes  them  to  assimilate  over 
rapidly  certain  other  elements  of  horse  serum.  These  assimilated  or  toxic  ele- 
ments are  the  same  as  those  eliminated  without  producing  intoxication  during  the 
incubation  period  following  the  first  dose.  This  overassimilation  after  anaphy- 
laxis is  the  cause  of  the  intoxication. 

Gay  and  Southard  find  much  support  for  their  contentions  in  the  results  of 
experiments  done  with  the  so-called  ' '  passive ' '  transfer  of  hypersusceptibility. 
As  mentioned  above,  hypersusceptibility  may  be  transferred  to  a  normal  animal 
with  the  blood  serum  not  only  of  a  sensitive  animal,  but  even  more  surely  and 
effectually  with  that  of  a  refractory,  or  '  *  antianaphylactic, ' '  animal.  They  believe 
that  such  transfer  is  not  "passive"  but  "active"  sensitization,  being  accom- 
plished by  the  transfer  of  "anaphylactin"  to  the  normal  animal.  The  re- 
fractory animal  has  received  more  horse  serum  than  the  merely  sensitive  one, 
since  antianaphylaxis  is  produced  by  massive  injections.  Therefore  its  blood 
contains  more  anaphylactin  and  is  consequently  more  active  in  transferring  sen- 
sitiveness. The  fact  that  a  considerable  incubation  .time  is  necessary  in  active 
sensitization  they  attribute  to  the  gradual  action  of  the  anaphylactin. 

In  passive  sensitization,  therefore,  they  assumed  a  similar  gradual  irrita- 
tion of  the  vulnerable  cells  by  the  anaphylactin  and,  as  we  have  seen,  obtained 
their  reactions  in  animals  so  treated,  usually  10  to  14  days  after  the  sensi- 
tive serum  had  been  given.  This  conception  of  the  mechanism  of  passive 
anaphylaxis  was,  of  course,  rendered  unlikely  by  the  demonstrations  by  Friede- 
mann,  Otto,  and  others  that  shock  could  be  elicited  in  passively  sensitized  animals 
within  24  hours  or  less  after  transfer  of  the  anaphylactic  serum. 

To  this,  however,  Gay  and  Southard2  answer  by  implying  that  this  greater 
speed  of  development  of  sensitiveness  in  the  experiments  of  Otto  is  due  to  the 
larger  doses  used  by  him.  They  say  "if  the  doses  are  sufficient  it  (transmitted 
sensitiveness)  may  be  shown  in  a  single  day  (Otto)."  However,  it  is  very  likely 
that  the  sensitiveness  noted  by  them  in  animals  two  weeks  after  the  transference 
of  anaphylactic  serum  was  actually  positive  sensitization  with  antigen  rests,  en- 
tirely comparable  to  the  usual  ' '  Theobald  Smith ' '  phenomenon. 

Gay  &  Southard's  definite  objections  to  the  possibility  of  an  antigen-antibody 
reaction  are  found  in  the  following  arguments  based  on  experimental  observations: 

1.  Sensibility  persists  for  a  long  time,  antibodies  disappear  rapidly. 

2.  In   the   serum   of   animals   sensitive   to   horse    serum   antibodies   to   this 
serum  are  not  demonstrable  by  complement  fixation. 

3.  Although  sensitiveness  can  be  transferred  to  a  normal  animal,  neverthe- 
less a   definite  period  of   incubation  must   elapse   before   the   recipient   becomes 
sensitive. 

To  the  first  of  these  arguments  Besredka3  objects  by  saying  that,  while  it 
is  true  that  sensitiveness  persists  for  a  long  time,  the  power  to  transmit  anaphy- 
laxis passively  disappears  rapidly,  as  Otto,  Eichet,  and  others  have  shown. 

The  second  contention  is  contradicted  by  the  work  of  Nicolle  and  Abt.4 
But  since  these  workers  made  their  observations  upon  rabbits  their  experiments 
do  not  necessarily  contradict  those  of  Gay  and  Southard.  This  point  at  best  is 
a  difficult  one  to  determine,  especially  as  recent  investigations  haye  shown  us 
that  under  certain  circumstances  antigen  and  antibody  may  be  found  side  by 
side  in  the  same  serum  without  uniting  and  without  therefore  fixing  alexin  or 
complement. 

*Gay  and  Southard.  Jour.  Med.  Ees.,  Vol.  16,  1907;  Vol.  18,  1908; 
Vol.  19,  1908. 

2  Gay  and  Southard.     Jour.  Med.  Ees.,  p.  427,  1908. 

3  Besredka.     Bull,  de  I'Inst.  Past.,  6,  1908,  p.  826. 

4  Nicolle  and  Abt.    Ann.  de  I'Inst.  Past.,  Vol.  22,  p.  132,  1908. 


386  INFECTION  AND  RESISTANCE 

The  point  of  their  third  argument  has  been   discussed  above. 
It  is  clear  that  Gay  and  Southard  separate  distinctly  the  substance  in  the 
antigen  which  sensitizes  from  that  which  exerts  the  toxic  action  on  second  injection. 

Another  theory  which  is  based  on  such  a  separation  of  a  sensi- 
tizing  and  shock-producing  element  in  the  original  antigen  is  that 
of  Besredka. 

Besredka5  assumes  that  in  the  injected  antigen  (serum)  there  are  two 
separate  substances.  One  of  these,  the  sensibilisinogen,  induces,  during  the  time 
of  incubation,  a  specific  antibody  (sensibilisin) .  This  antibody  remains  in  part 
attached  to  tissue  cells  and  in  part  circulates  freely  in  the  blood.  The  other 
substance  in  the  antigen  he  calls  "  antisensibilisin."  This,  at  the  second  injection, 
reacts  with  the  sensibilisin  and  anaphylactic  shock  results.  The  nature  of  the 
symptoms  is  explained  by  the  fact  that  the  antibody  or  sensibilisin  is  attached  to 
cells  of  the  central  nervous  system,  and  shock  can  result  only  when  such  attach- 
ment is  present.  Thus,  in  passive  transference  of  sensitization,  the  property  of 
hypersusceptibility  is  bestowed  upon  the  normal  animal  by  the  sensibilisin  or 
antibody  present  in  the  circulating  blood,  but  the  significance  of  this  body  for 
anaphylaxis  is  not  in  evidence  until  a  connection  with  the  central  nervous  system 
has  been  established. 

There  is  much  in  Besredka's  theory  which  is  at  variance  with 
prevailing  conceptions  of  biological  phenomena  of  this  category.  The 
fact  that  an  antigen  should  give  rise  to  an  antibody  which  reacts 
not  with  the  substance  that  induced  it,  but  with  a  third  body,  is 
quite  out  of  keeping  with  experience. 

However,  it  is  clear  that  in  both  theories,  that  of  Gay  and  South- 
ard, as  well  as  that  of  Besredka,  the  cardinal  point  is  this  separa- 
tion in  the  antigen  of  two  substances,  a  sensitizing  and  a  toxic  or 
shock-producing,  and,  since  this  forms  the  chief  argument  against  an 
antigen-antibody  conception  of  anaphylaxis,  it  will  be  necessary  to 
examine  the  experimental  evidence  on  which  it  is  based. 

Gay  and  Adler  6  attempted  to  show  such  a  dual  function  of  the 
original  antigen  by  chemical  methods.  They  report  that,  by  frac- 
tional precipitation  of  horse  serum  with  ammonium  sulphate,  the 
successive  protein  fractions  obtained,  as  saturation  is  increased,  are 
found  to  be  less  sensitizing  and  more  toxic  as  more  and  more  am- 
monium sulphate  is  added.  The  first  fraction  (euglobulins)  obtained 
by  %  saturation  is  as  sensitizing  as  whole  serum  and  corresponds  to 
anaphylactin,  but  is  nontoxic  when  injected  into  sensitive  animals. 
The  last  fraction,  while  distinctly  less  sensitizing  than  either  the 
whole  serum  or  the  first  fraction,  is  at  least  as  toxic  as  the  whole 
serum. 

In  these  experiments,  therefore,  we  have  a  strong  argument  in 
favor  of  the  separate  presence  in  an  anaphylactic  antigen  of  two 
bodies,  the  one  sensitizing  and  the  other  toxogenic.  However,  this- 
assertion  has  not  been  borne  out  by  later  work. 

5  Besredka.     Loc.  cit. 

6  Gay  and  Adler.    Jour.  Med.  Res.,  Vol.  13,  1908. 


ANAPHYLAXIS  387 

Pick  and  Yamanouchi,7  whose  extensive  investigation  cannot  be 
fully  reviewed  here,  were  unable  to  obtain  such  a  separation;  in 
fact,  they  conclude  that  the  same  substances  which  sensitize  are  also 
toxic,  and,  working^with  a  large  variety  of  methods,  find  that  both 
the  sensitizing ^a~nd  toxogenic  properties  of  proteins  show  no  differ- 
ences either  in  chemical  condition  or  in  resistance  to  chemical  agents 
or  heat. 

The  work  of  Pick  and  Yamanouchi,  however,  was  done  with 
rabbits  and,  therefore,  as  bearing  on  the  theory  of  Gay  and  Southard, 
the  work  of  Doerr  and  Russ  8  is  more  directly  to  the  point.  These 
workers  using  guinea  pigs,  and  both  horse  and  beef  sera,  obtained 
results  which  are  practically  diametrically  opposed  to  those  of  Gay 
and  Adler.  They  found  that  the  euglobulins,  obtained  by  %  satura- 
tion with  ammonium  sulphate,  are  the  most  strongly  sensitizing  and, 
at  the  same  time,  the  most  toxic  of  the  fractions  of  the  sera.  As 
saturation  with  the  salt  is  increased,  the  proteins  which  come  down 
decrease  progressively  and  in  parallelism,  both  as  regards  the  power 
to  sensitize  and  the  faculty  of  exerting  toxic  action  on  second  injec- 
tion. The  albumin,  which  finally  comes  out  on  total  saturation,  is 
devoid  both  of  sensitizing  and  of  toxic  properties.  Similar  results 
were  obtained  by  Doerr  and  Russ  with  the  precipitation  of  serum 
proteins  with  CO2. 

The  weight  of  evidence,  therefore,  seems  to  point  against  a  chem- 
ical separation  of  the  two  functions  in  the  antigen. 

Besredka's  contentions  in  favor  of  such  a  separation  were  based 
chiefly  upon  a  difference  in  resistance  to  heat. 

His  experiments  showed  that  the  sensitizing  properties  of  serum 
are  not  lost  even  if  it  is  heated  to  120°  C.,  while  the  toxogenic 
powers  are  destroyed  by  much  lower  temperatures.  The  results  of 
Besredka  as  to  the  differences  in  thermostability  between  the  two 
properties  have  found  confirmation  by  Kraus  and  Yolk  9  and  others, 
and  there  can  be  little  doubt  that  the  sensitizing  function  is  extremely 
heat-resistant,  since  this  has  also  been  shown  by  Wells,10  Rosenau 
and  Anderson,  and  many  others.  However,  researches  by  Doerr  and 
Russ,1  *  and  notably  by  Wells,  have  shown  that,  though  not  destroyed 
by  high  temperatures,  even  moderate  heating  markedly  diminishes 
the  sensitizing  function,  and  that  larger  doses  have  to  be  given  as 
the  temperature  is  increased;  and  since  the  smallest  quantities  of 
antigen  necessary  for  inducing  shock  at  the  second  injection  must  be 
anywhere  from  100  to  1,000  times  as  large  as  the  smallest  sensi- 
tizing doses,  it  is  quite  likely  that  a  combination  of  such  conditions 

7  Pick  and  Yamanouclii.     Zeitschr.  f.  Immunitatsforschung,  1,  1909. 

8  Doerr  and  Russ.     Zeitschr.  f.  Immunitatsforschung,  Vol.  2,  1909. 

9  Kraus  and  Volk.     Zeitschr.  f.  Immunitatsforschung,  Vol.  3,  1909. 

10  Wells.     Jour.  Inf.  Dis.,  Vol.  5,  1908. 

11  Doerr  and  Russ.     Loc.  cit. 


388  INFECTION  AND  RESISTANCE 

might  stimulate  an  actual  difference  in  heat  resistance.  In  fact,  this 
is  the  view  expressed  by  Wells  12  and  borne  out  by  experiments  car- 
ried out  by  Doerr  and  Russ. 

Wells,  too,  confirms  the  identity  of  sensitizing  and  toxic  sub- 
stance by  his  experiments  on  the  influence  of  tryptic  digestion  upon 
these  properties  of  the  antigen.  He  concludes  that  both  sensitizing 
and  intoxicating  properties  are  attacked  and  slowly  decrease  as  the 
coagulable  protein  disappears. 

As  to  that  aspect  of  Besredka's  theory  which  deals  with  the 
indirect  participation  of  the  central  nervous  system,  his  arguments 
are  based  mainly  on  the  fact  that  ether  narcosis  seemed,  in  his 
experiments,  to  prevent  anaphylactic  shock  when  animals  were 
deeply  anesthetized  during  the  second  injection,  and  also  upon  the 
regularity,  severity,  and  speed  with  which  anaphylactic  symptoms 
follow  injections  directly  into  the  brain.  The  former  contention 
regarding  narcotics  cannot,  by  any  means,  be  accepted  as  yet,  since 
Rosenau  and  Anderson  failed  to  confirm  it  and  claim  that  ether 
narcosis  merely  masks  the  symptoms  but  does  not  prevent  death. 
If  we  admit  the  beneficial  effects  of  ether,  moreover,  it  may  well  be 
that  this  is  accomplished  by  relaxation  of  the  bronchial  spasms, 
known,  since  Auer  and  Lewis,  to  be  the  cause  of  death  in  guinea 
pigs,  and  the  action  of  ether  could  hardly  be  utilized,  therefore,  to 
argue  in  favor  of  a  central  localization  of  the  anaphylactic  process. 

That  phase  of  the  two  theories  so  far  mentioned,  therefore,  which 
depends  upon  the  assumption  of  two  separate  substances  in  the  orig- 
inal antigen  does  not  seem  established  nor  even  sufficiently  likely 
to  warrant  the  formulation  of  a  theory  upon  it. 

The  second  premise  is  the  necessary  participation  of  the  body 
cell,  in  that  the  reaction  cannot  take  place  unless  the  cells  are  ren- 
dered vulnerable  by  preliminary  alteration.  In  Gay  and  Southard's 
theory  this  is  assumed  to  occur  by  irritation  exerted  by  the  "anaphy- 
lactin,"  in  Besredka's  scheme  it  is  attributed  to  the  antisensibilisin 
which  is  attached  to  the  nerve  cells. 

It  is  plain,  therefore,  that  both  Gay  and  Southard  and  Besredka 
admit  a  preliminary  preparation  of  the  cells  of  the  body,  and  this  is, 
as  we  shall  see,  an  important  factor  in  anaphylaxis,  though  not 
exactly  in  the  sense  of  any  of  the  observers  named. 

All  other  views  of  the  mechanism  of  anaphylaxis  have  held  from 
the  beginning  that,  in  substance,  this  reaction  is  a  true  antigenr 
(antibody  reaction.  The  injected  antigen  gives  rise  to  a  specific 
antibody.  This,  on  second  injection,  unites  with  the  first  antigen 
and  the  result  is  anaphylactic  shock.  Such  a  point  of  view  was  held 
by  v.  Pirquet,  Rosenau  and  Anderson,  and  others,  who  reached  this 
conclusion  from  the  nature  of  the  anaphylactic  antigens,  the  speci- 

12  Wells.    Jour    Inf.  Dis.,  Vol.  6,  p.  521,  1909. 


ANAPHYLAXIS  389 

ficity  of  the  reaction,  the  incubation  time,  and  the  phenomena  of 
passive  sensitization. 

This  assumption  gained  much  further  support  when  Doerr  and 
Kuss  13  succeeded  in  applying  quantitative  methods  to  the  study  of 
the  anaphylactic  antibody.  Their  methods  consisted  in  precipitating 
sera  in  rabbits.  With  these  they  then  passively  sensitized  guinea 
pigs,  subsequently  testing  them  with  antigen  24  hours  later.  To 
arrive  at  quantitative  results  they  developed  two  reliable  methods. 
These  consisted  in:  1.  Intraperitoneal  sensitization  of  guinea  pigs 
with  constant  quantities  of  titrated  precipitating  serum.  Twenty- 
four  hours  later  intravenous  test  with  diminishing  amounts  of  specific 
antigen.  2.  Intraperitoneal  sensitization  W7ith  diminishing  quanti- 
ties of  the  titrated  precipitating  serum,  and  24  hours  later  intrave- 
nous tests  with  constant  amounts  of  antigen. 

In  this  way  they  showed  that  there  was  a  direct  relationship 
tween  the  power  of  a  serum  to  convey  anaphylaxis  passively  and  its 
contents  of  precipitins.  We  may  elucidate  this  by  an  example  from 
their  work.  They  possessed  a  rabbit  serum  which  gave  precipitation 
with  sheep,  goat,  beef,  pig,  human,  and  horse  sera,  but  not  with 
chicken  serum.  The  precipitation  titre  of  this  serum  for  the  sera 
mentioned  varied  from  1  in  20,000  in  the  case  of  sheep  and  goat 
sera,  to  1  in  100  in  the  cases  of  the  human  and  horse  sera.  When 
guinea  pigs  were  injected  intraperitoneally  with  1  c.  c.  of  this  serum, 
and  after  24  hours  were  intravenously  injected  with  the  various  sera 
mentioned  above,  in  decreasing  quantities,  the  sera  which  were  pre- 
cipitated in  the  highest  dilutions  gave  anaphylactic  shock  in  the 
smallest  quantities.  Those  sera  in  which  no  precipitin  or  little  had 
been  present  gave  little  or  no  reaction  by  this  method  even  where  con- 
siderable quantities  were  used.  Thus  in  animals  prepared  by  1  c.  c. 
of  the  antiserum  (precipitated  in  dilutions  of  1  in  20,000)  sheep 
serum  caused  death  when  injected  in  doses  of  0.006  c.  c.,  whereas 
horse  serum  (which  was  precipitated  only  in  concentration  of  1  to 
100)  gave  slight  symptoms  only  when  2  c.  c.  were  employed  for  re- 
injection  and  chicken  serum  (non-precipitable  by  the  antiserum) 
gave  no  reaction  in  similar  doses. 

In  this,  then,  we  have  a  definite  quantitative  analysis  which 
proves  that  the  power  to  sensitize  passively  is  in  direct  relation  to  the 
antibodies  against  the  protein  present  in  the  sensitizing  serum. 
Whether  or  not  this  means  the  precipitins  we  consider  relatively  un- 
important inasmuch  as  we  have  already  made  clear  that  we  believe 
precipitins,  complement-fixing  antibodies  and  agglutinins  to  be  one 
and  the  same  thing. 

We  may  therefore  accept,  as  a  definitely  determined  fact,  that 
anaphylaxis  is  directly  or  indirectly  the  result  of  the  reaction  within 

13  Doerr  and  Russ.     Zeitschr.  f.   Immunitatsforscliung,  Vol.  3,   pp.  181 
and  706,  1909. 


390  INFECTION  AND  RESISTANCE 

the  animal  body  of  an  antigen  with  its  specific  antibody.  In  "active" 
anaphylaxis  the  antibodies  are  present  as  the  reaction  to  a  preceding 
antigen  injection.  In  the  "passive"  condition  they  were  conveyed 
with  the  injected  antiserum. 

We  are  now  prepared  to  follow  individually  the  development  of 
those  theories  in  which  the  anaphylactic  mechanism  was  looked  upon 
purely  as  the  result  o^  the  union  of  an  antigen  with  its  antibody. 

The  conception  which  gradually  grew  out  of  the  antigen-antibody 
mechanism  of  anaphylaxis  was  the  following:  When  a  specific  an- 
tigen meets  its  antibody  the  reaction  between  them  gives  rise  to  a 
toxic  product,  and  this  causes  the  characteristic  symptoms.  A  simi- 
lar idea,  it  will  be  remembered,  is  found  in  the  original  endotoxiii 
theory  of  Pfeiffer.  According  to  this,  the  action  of  the  specific 
lysin  liberated  from  bacteria  a  preformed  poison,  the  endotoxin.  In 
1902  Weichhardt,  14  bearing  this  conception  in  mind,  subjected 
syncytial  protein  of  rabbit  placenta  to  the  action  of  specific  antisera 
and  obtained  substances  toxic  for  normal  rabbits. 

This  work  was  done  long  before  the  days  of  anaphylaxis  studies, 
and  the  results  were  interpreted  in  keeping  with  Pfeiffer's  theory. 
However,  as  Weichhardt  himself  now  claims,  it  is  not  unlikely  that 
he  was  dealing  with  a  phenomenon  analogous  to  the  ones  we  are  now 
discussing.  A  similar  opinion  of  the  production  of  toxic  substances 
by  specific  cytolysis  was  expressed  by  Wolff-Eisner  15  in  1904. 

Probably  the  most  important  of  the  earlier  investigations  along 
these  lines,  at  least  in  its  direct  bearing  on  anaphylaxis,  was  the 
work  of  Vaughan  and  Wheeler/6  published  in  1907. 

In  its  general  significance  this  work  ranks  among  the  most  im- 
portant contributions  to  our  understanding  of  hyper  susceptibility,17 
though  the  theoretical  deductions  made  from  it  have  had  to  be  sub- 
jected to  considerable  alteration.  Their  conception  of  anaphylaxis 
takes  root  in  the  earlier  investigations  of  Vaughan  18  and  his  pupils 
upon  the  extraction  of  a  poisonous  group  from  the  protein  molecule. 

Vaughan  and  Wheeler  19,  20  believe  that  the  sensitizing  and  the 
toxogenic  properties  of  the  anaphylactic  antigens  are  in  truth  con- 
tained within  the  self-same  proteid  molecule;  but  can  be  chemically 
separated  from  each  other.  They  have  been  able  to  split  egg  al- 
bumen and  other  proteids  by  treatment  with  absolute  alcohol  (con- 

14  Weichhardt.     Deutsche  med.  Woch.,  1902,  p.  624. 

15  Wolff- Eisner.     Centralbl  f.  Bakt.,  Vol.  37,  1904. 

16  Vaughan  and  Wheeler.     Jour.  Inf.  Dis.,  Vol.  4,  1907. 

17  This  work  also  contains  the  germ  of  the  more  recent  ideas  upon  the 
nature    of   toxemia    in    infectious    disease,    advanced   more    particularly    by 
Friedberger.     This  will  be  considered  in  detail  in  the  next  chapter. 

18  Vaughan.     Transact.  Ass'n  Am.  Phys.,  Vol.  16,  1901;  Jour.  A.M..  A., 
Vol.  36,  1901;  Am.  Med.,  1901;  Jour.  A.  M.  A.,  Vol.  43,  1904. 

19  V  C.  Vaughan,  Jr.     Jour.  A.  M.  A.,  Vol.  44,  1905,  p.  1340. 

20  V.  C.  Vaughan,  Jr.     Boston  Med.  and  Surg.  Jour.,  Vol.  155,  1906. 


ANAPHYLAXIS  391 

taining  2  per  cent.  NaOH)  into  2  fractions — a  toxic  alcohol-soluble 
and  a  non-toxic  alcohol-insoluble  one.  The  former  fraction  gave 
protein  reactions,  and  they  regard  it  as  a  true  protein — while  Wells,21 
considering  the  hydrolytic  nature  of  the  cleavage  resorted  to,  con- 
siders this  fraction  as  possibly  a  soluble  peptone  or  polypeptid  (the 
positive  protein  reactions  being  possibly  due  to  ammo  acids).  The 
non-alcohol-soluble,  non-toxic  fraction  also  gives  proteid  reactions. 
Injections  into  guinea  pigs  of  the  toxic  fraction  produce  symptoms 
riot  unlike  anaphylaxis — but  do  not  sensitize  against  protein.  The 
alcohol-soluble  portion  is  non-toxic  and  sensitizes  against  protein  in 
doses  of  0.001  to  0.005  gm.  Based  on  these  results,  their  views  of 
mechanism  of  anaphylaxis  are  as  follows : 

At  the  first  injection  a  slow  lysis  (cleavage)  of  the  injected  pro- 
tein gradually  liberates  a  fraction,  corresponding  to  the  alcohol-in- 
soluble substance — and  this  by  its  antigenic  action  gives  rise  to  the 
formation,  in  excess,  of  an  enzyme  (lysin),  which  on  reinjection 
brings  about  the  rapid  cleavage  of  the  injected  protein — with  an  ex- 
plosive liberation  of  the  toxic  fraction  and  consequent  symptoms.22 

Nicolle  believes  that  the  injection  of  a  protein  into  an  animal  induces 
the  production  in  the  subject  of  antibodies.  These  are  preeminently  two — 
albuminolysins,  which  cause  its  cleavage,  and  albuminocoagulins  or  precipi- 
tins,  which  coagulate  and  prevent  the  action  of  the  lysin.  At  the  time  at 
which  an  animal  is  hypersusceptible  or  anaphylactic  there  has  been  a  pro- 
duction of  albuminolysins  which  cause  cleavage  of  the  protein,  with  the 
rapid  liberation  of  toxic  substances;  but  the  albuminocoagulins  or  precipi- 
tins  have  not  yet  adequately  developed.  In  a  refractory  animal  the  neu- 
tralizing action  of  the  albuminoprecipitins  prevents  the  harm  which  the  lytic 
action  might  otherwise  accomplish.  The  relative  amounts  of  these  two  anti- 
bodies present  in  the  circulation  of  the  animal  at  any  particular  time  deter- 
mine whether  the  animal  is  anaphylactic  or  refractory  or  immune.  This 
theory  assumes  arbitrarily  the  protective  nature  of  precipitation,  an  idea 
which  has  no  foundation  in  experiment  and,  in  fact,  is  rendered  extremely 
unlikely  by  more  recent  developments  of  our  knowledge  of  the  precipitating 
antibodies. 

Given,  then,  a  reasonable  hypothesis  in  which  anaphylaxis  is 
associated  with  the  cleavage  of  protein  by  lysis,  given,  in  other  words, 
an  antigen-antibody  conception,  it  is  but  natural  that  experimenters 
should  ask  themselves:  What  is  the  relation  of  the  alexin  to  this 
cleavage  ?  For  in  all  known  lytic  reactions,  of  course,  the  union  of 
antigen  and  antibody  leads  to  the  absorption  of  alexin,  by  means  of 
which,  then,  the  lysis  has  been  assumed  to  take  place.  This  problem 
suggested  itself  to  a  number  of  the  earlier  investigators  who  attempted 
to  approach  it  by  determining  whether  or  not  the  sera  of  sensitive 

21  Wells.     Jour.  Inf.  Dis.,  Vol.  5,  1908. 

22  For  the  sake  of  completeness  it  is  well  also  to  mention   Nicolle's 23 
iheory,  which,  though  attractive,  is  not  borne  out  by  recent  knowledge  con- 
cerning the  nature  of  precipitins. 

2:!  Nicolle.     Am.  de  I'Inst.  Past.,  Vol.  22,  1998. 


392  INFECTION  AND  RESISTANCE 

animals,  added  to  antigen,  would  fix  alexin.  Gay  and  Southard, 
Sleeswijk,24  and  others  obtained  negative  results,  while  Nicolle  and 
Abt,25  and  Doerr  and  Russ  26  obtained  positive  results.  As  far  as 
this  particular  method  is  concerned,  therefore,  no  conclusions  can 
be  drawn.  Sleeswijk,  however,  has  approached  the  question  in  an- 
other way  and  examined  whether  or  not  there  is  a  diminution  of 
alexin  in  the  blood  of  an  animal  immediately  after  anaphylactic 
shock.  He  found  that  this  was  indeed  a  regular  occurrence,  and  his 
results  have  been  confirmed  by  Friedberger  and  Hartoch  27  and  a 
number  of  others. 

Jt  was  shown  by  these  workers  that,  both  in  active  and  passive 
anaphylaxis  in  rabbits  and  dogs,  as  well  as  in  guinea  pigs,  there  is  a 
definite  and  considerable  diminution  of  complement  immediately 
after  anaphylactic  shock. 

The  question  now  arises :  What  is  the  significance  of  this  dimi- 
nution of  alexin?  Do  the  animals  die  because  of  a  sudden  loss  of 
circulating,  physiologically  necessary  alexin,  or  does  the  alexin  take 
an  active  part  in  producing  the  conditions  which  cause  death  ? 

Either  of  these  possibilities  might  follow  from  the  mere  fact  of 
alexin  diminution,  but  the  former — the  possibility  that  complement 
depletion  is  the  cause  of  death — was  ruled  out  by  Friedberger  and 
Hartoch.28  They  showed  that,  by  supplying  fresh  complement  to 
sensitive  animals  at  the  time  of  rein ject ion,  shock  cannot  be  pre- 
vented. They  now  proceeded  to  demonstrate  the  active  participation 
of  complement  in  the  production  of  anaphylaxis.  They  did  this  in  an 
ingenious  way  which  depended  on  utilization  of  the  fact  observed  by 
Nolf,29  Hektoen  and  Ruediger,30  and  others  that  hypertonic  salt 
solution  (1.5-2  per  cent.)  will  prevent  the  combination  of  comple- 
ment with  its  sensitized  cells.  By  slowly  injecting  into  sensitized 
guinea  pigs  0.3  cubic  centimeter  of  concentrated  NaCl  solution 
just  before  the  injection  of  antigen  they  were  able  to  markedly 
diminish  anaphylactic  shock — saving  animals  from  injections  which 
invariably  killed  the  controls.  The  force  of  this  experiment  we  think 
has  been  largely  eliminated  by  work  of  the  writer  with  Dwyer  and 
Lieb31  in  which  it  was  shown  that  the  effect  of  the  salt  is  upon  the 
rmooth  muscle  which  is  rendered  less  irritable  by  the  salt  and  there- 
fore less  susceptible  to  the  antigen. 

24  Sleeswijk.     Zeitschr.  f.  Immunitatsforschung,  Vol.  2,  1909. 

25  Nicolle  and  Abt.     Ann.  de  rinst.  Past.,  Vol.  22,  1908. 

26  Doerr  and  Russ.     Zeitschr.  f.  Jmmunitcitsforschung,  Vol.  3,  1909. 

27  Friedberger  and  Hartoch.     Zeitschr.  f.  Immunitatsforschung.  Vol.   3, 
1909. 

28  Friedberger  and  Hartoch.    Loc.  cit. 
9  Nolf.     Ann.  de  I'Inst.  Past.,  1900. 

30  Hektoen  and  Ruediger.     Jour.  Inf.  Dis.,  Vol.  1,  1904. 

31  Zinsser,  Lieb  and  Dwyer,  Proc.  Soc.  for  Exn.  Biol.  &  MecL,  Vol.  12, 
May,  1915,  p.  123. 


ANAPHYLAXIS  393 

An  ingenious  attempt  to  demonstrate  the  important  role  played 
by  complement  in  anaphylaxis  has  recently  been  furnished  by  Loef- 
fler.  Loeffler,32  using  guinea  pigs  sensitized  with  horse  serum,  com- 
pletely depleted  their  complement  by  injecting  intraperitoneally  con- 
siderable quantities  of  sensitized  sheep  corpuscles.  Tested  by 
injection  of  horse  serum  one  hour  later  no  anaphylaxis  occurred, 
while  controls  regularly  succumbed.33 

It  seemed  thus  established  that  the  complement  or  alexin  played 
an  important  active  part  in  the  production  of  anaphylaxis,  and  the 
next  logical  step  was  to  attempt  to  produce  the  anaphylactic  poison 
by  the  action  of  alexin  upon  an  antigen-antibody  complex  in  vitro. 
This  was  first  done,  with  direct  reference  to  anaphylaxis,  by  Ulrich 
Friedemann.34  Friedemann  chose  as  his  antigen-antibody  complex 
the  sensitized  red  blood  cell  after  he  had  demonstrated  by  preliminary 
experiment  that  the  basic  anaphylactic  experiment  could  be  carried 
out  in  rabbits  with  washed  beef  corpuscles.  He  found  that  if  3  c.  c. 
of  such  corpuscles  were  injected  into  rabbits  and  the  injection  re- 
peated after  3  weeks  anaphylactic  symptoms  were  regularly  elicited. 
He  then  allowed  alexin  to  act  upon  sensitized  beef  blood  in  vitro, 
interrupted  the  action  by  cooling  at  a  time  just  preceding  the  occur- 
rence of  hemolysis  (to  exclude  the  supposed  toxic  action  of  hemoglo- 
bin), and  injected  the  supernatant  fluid  of  such  mixtures  into  normal 
rabbits.  The  result  was  marked  illness  resembling  anaphylaxis,  and 
Friedemann  thus  had  succeeded  in  producing  the  anaphylactic  poison 
in  vitro  under  conditions  as  nearly  as  possible  similar  to  those  occur- 
ring in  the  circulation  of  the  anaphylactic  rabbit.  In  the  conclusions 
drawn  from  his  experiments  he  expresses  the  opinion  that  the  poisons 
were  not  preformed  in  the  red  blood  cells,  but  were  formed  by  the 
proteolysis  exerted  by  "amboceptor"  and  complement.  In  this  state- 
ment he  sets  down  the  basic  conception  of  the  production  of  anaphy- 
lactic poisons  now  generally  held. 

Friedemann,  then,  in  attempts  to  apply  the  same  methods  to  the 
study  of  serum  anaphylaxis,  attempted  to  produce  similar  poisons 
by  the  action  of  rabbit  alexin  upon  the  washed  precipitates  formed 
by  mixtures  of  antigen  and  precipitating  sera.  In  this  he  failed — 
probably  because  of  his  choice  of  rabbits  as  subjects  for  experiment. 
Where  he  had  failed,  however,  Friedberger 35  succeeded  by  using 
guinea  pigs.  Doerr  and  Russ  36  had  previously  shown  that  feeble 
symptoms  of  shock  could  be  produced  by  the  injection  of  serum  pre- 

32  Loeffler.     Zeitschr.  f.  Immunitatsforsch.,  8,  1910. 

33  For  additional  evidence  pointing  in  the  same  direction  see  also  TJhlen- 
huth  and  Haendel,  Zeitschr.  f.  Immunitatsforsch.,  Vol.  3,  1909. 

34  Ulrich  Friedemann.     Zeitschr.  f.  Immunitatsforsch.,  Vol.  2,  1909. 

35  Friedberger.     Berl.   klin.    Woch.,  32   and  42,   1910 ;   also  Zeitschr.  f. 
Immunitatsforsch.,  Vol.  4,  1910. 

36  Doerr  and  Russ.     Zeitschr.  f.  Immunitatsforsch. y  Vol.  3,  p.  181,  1909. 


394  INFECTION  AND  RESISTANCE 

cipitates  into  normal  guinea  pigs.  With  this  additional  evidence 
in  favor  of  his  reasoning,  Friedberger  proceeded  as  follows: 

One  c.  c.  of  a  rabbit  serum  which  precipitated  sheep  serum  in  a 
dilution  of  1  to  10,000  was  mixed  with  30  c.  c.  of  a  1  to  50  sheep 
serum  dilution.  This  was  kept  one  hour  at  37.5°  C.  and  over  night 
in  the  ice-chest,  when  a  heavy  flocculent  precipitate  had  formed. 
This  precipitate  was  washed  to  remove  all  traces  of  serum,  and  to  it 
were  added  2  c.  c.  of  fresh  normal  guinea  pig  serum — as  comple- 
ment. This  was  again  allowed  to  stand  for  12  hours  and  then  the 
supernatant  fluid  was  injected  into  a  guinea  pig  intravenously.  In 
most  cases  the  pigs  so  treated  showed  marked  symptoms  soon  after 
the  injection  and  died  within  a  few  hours. 

Friedberger  concludes,  therefore,  that  anaphylactic  shock  is  a 
true  intoxication  due  to  a  poison  produced  from  the  products  of  a 
precipitin-precipitinogen  reaction  by  the  action  of  a  complement; 
he  speaks  of  the  formed  poison  as  anaphylatoxin.  The  experiment 
just  outlined,  moreover,  seems  to  show,  contrary  to  Friedberger's 
first  ideas,  that  the  entire  reaction  may  go  on  under  certain  circum- 
stances in  the  blood  stream  without  intervention  of  sessile  precipitins 
upon  the  cells. 

We  have,  thus,  in  the  cited  work  of  Friedberger  the  culmination 
of  a  long  series  of  investigations — the  end  result  being  the  conclusion 
that  in  all  probability — at  least  as  far  as  experimental  ingenuity 
has  permitted  us  to  penetrate  into  this  very  difficult  problem  up  to 
the  present  time — the  phenomenon  of  anaphylaxis  must  be  regarded 
as  an  acute  intoxication,  the  poison  which  calls  it  forth  being  the 
result  of  the  union  of  an  antigen  and  its  antibody,  the  complex  being 
subsequently  subjected  to  proteolysis  by  the  action  of  alexin  or  com- 
plement. The  experimental  extension  of  this  conception  to  the  phe- 
nomena of  bacterial  anaphylaxis  has  promised  to  exert  such  an  im- 
portant influence  upon  our  conceptions  of  infectious  disease  that  we 
will  take  up  these  investigations  in  a  separate  section. 

37  Here,  then,  we  have  a  simple  and  apparently  logical  explanation 
of  anaphylaxis,  entirely  in  accord  with  Vaughan's  views  of  parenteral 
digestion.  An  antigen  is  injected  into  an  animal,  specific  antibodies 
and  enzymes  against  it  develop  in  the  animal ;  reinjectiori  of  this 
antigen  results  in  relatively  rapid  proteolysis  in  the  course  of  which 
poisonous  substances,  the  anaphylatoxins,  are  produced  and  anaphy- 
laxis is  the  result.  This  hypothesis  although  very  attractive  does 
not  entirely  meet  with  the  facts  as  they  have  been  developed  since 
Friedberger's  first  work.  The  premises  on  which  it  is  based  assume 
in  the  first  place  that  the  poison  or  "anaphylatoxin' '  is  formed  out  of 
the  matrix  of  the  antigen ;  further,  it  is  definitely  assumed  that  in  the 

37  Much  of  this  discussion  is  adapted  from  Zinsser,  "More  Recent  De- 
velopments in  the  Study  of  Anaphylactic  Phenomena."  Harvey  Lecture 
Arch,  of  Int.  Med.,  Aug.,  1915,  Vol.  16,  p.  223. 


ANAPHYLAXIS  395 

production  of  the  poison  after  the  antigen  and  antibody  have  met,  the 
complement  or  alexin  plays  an  active  part.  Friedberger's  hypothesis 
as  stated  by  him,  moreover,  seems  to  assume  that  the  entire  process 
takes  place  intravascularly,  a  matter  which  we  will  discuss  at  consid- 
erable length  in  a  short  time.  It  is  important  to  note  also  that  Fried- 
berger,  with  Nathan,  was  able  to  show  that  this  anaphylatoxin  produc- 
tion could  take  place  within  the  animal  body ;  that  is,  within  the  peri- 
toneum of  a  guinea-pig  into  which  bacteria  had  been  injected. 

The  simplicity  of  Friedberger's  explanation  and  the  correctness 
of  his  experimental  data  soon  persuaded  many  investigators  that,  in 
essence,  his  hypothesis  probably  contained  the  nucleus  of  the  solution 
of  this  difficult  problem.  However,  even  his  own  early  experiments 
aroused  some  misgivings  concerning  the  matrix  of  the  poisons  pro- 
duced, for  he  found  that  the  poisons  could  be  obtained  as  well  when 
boiled  antigen  was  used  as  when  the  fresh,  unheated  substances  were 
employed,  and  the  poisons  were  easily  obtained  from  such  organisms 
as  the  tubercle  bacillus,  which  is  extremely  insoluble  and  unamenable 
to  serum  influence.  It  was  also  doubted  whether  one  could  truly 
assume  the  participation  of  this  specific  antibody  or  sensitizer  in  the 
production  of  Friedberger's  poisons,  since  it  soon  developed  that  from 
bacteria,  at  least,  the  poison  could  be  produced  when  the  organisms 
were  directly  exposed  to  the  action  of  fresh  guinea-pig  serum  without 
the  presence  of  any  immune  serum. 

Experiments  which  soon  cast  a  definite  doubt  on  the  assumption 
that  the  poisons  were  produced  by  a  decomposition  of  the  antigen 
were  reported  by  Keysser  and  Wassermann,38  These  workers  sub- 
stituted insoluble  substances  like  barium  sulphate  and  kaolin  for  the 
antigen;  that  is,  the  precipitates  or  bacteria  used  in  Friedberger's 
experiments.  They  found  that  if  kaolin  were  treated  with  horse 
serum  and,  then  exposed  to  the  action  of  guinea-pig  serum  or  comple- 
ment, poisons  were  produced  identical  in  every  respect  to  those  pro- 
duced by  Friedberger's  method.  The  conclusions  they  drew  were 
that  the  poisons  were  produced,  not  by  action  of  the  complement  on 
the  antigen,  but  by  its  action  on  the  horse  serum  absorbed  by  the 
kaolin.  In  other  words,  they  transferred  the  matrix  of  the  poison 
from  the  antigen  to  constituents  in  the  serum  itself,  possibly  the 
•sensitizer  or  amboceptor.  Bordet  39  also  was  able  to  show  that  poisons 
similar  to  those  of  Friedberger  could  be  produced  by  the  action  of 
fresh  guinea-pig  serum  on  agar,  and  recently  Bordet  has  further 
shown  that  this  is  the  case  even  when  the  agar  has  been  by  special 
methods  deprived  entirely  of  its  nitrogenous  components  and  repre- 
.sents  simply  a  complex  of  carbohydrates.  Agar-guinea-pig  serum 
mixtures  of  this  kind  show  an  increase  in  total  nonprotein  nitrogen 

38  Keysser  and  Wassermann.  Folia  serol,  1911,  vii;  Ztschr.  f.  Hyg.  u. 
Infectionskrankh.,  1911,  Ixviii. 

39  Bordet.     Compt.  rend.  Soc.  de  biol,  1913,  Ixxiv,  877. 


396  INFECTION  AND  RESISTANCE 

which  would  prove  that  the  proteolytic  action  of  the  guinea-pig  serum 
must  have  been  active  against  its  own  proteins. 

An  interesting  further  development  of  this  work  has  recently 
appeared  in  the  experiments  of  Jobling  and  Petersen.40  They 
showed  that  when  bacteria  are  mixed  with  fresh  active  serum  they 
adsorb  the  antienzymes  normally  present  in  blood.  They  have  shown 
this  experimentally  and  have  proved  that  similar  antienzyme  re- 
moval can  be  accomplished  by  the  addition  of  kaolin  or  agar,  and  by 
treatment  with  chloroform.  Serums  so  treated  become  toxic,  the 
actions  of  the  poisons  formed  showing  great  similarity  to  that  pro- 
duced by  Friedberger's  anaphylatoxins.  According  to  them,  the 
poisons  are  formed  because  of  the  fact  that  antienzymes  are  adsorbed 
by  the  antigen,  thus  setting  the  normal  ferments  in  the  fresh  serum 
free  to  act  on  their  own  serum  protein. 

It  should  be  recalled  that  Friedemann,  who  was  really  the  first 
one  to  show  that  the  toxic  substances  could  result  from  the  interaction 
of  fresh  serum  and  sensitized  antigens,  although  he  succeeded  only 
in  doing  this  with  red  blood  cells,  suggested  rather  early  that  the 
success  of  such  an  experiment  does  not  necessarily  mean  that  the 
antigen  furnishes  the  matrix  entirely.  He  had  studied  the  metabo- 
lism in  anaphylactic  poisoning  and  with  Isaac  has  shown  that  the 
nitrogen  output  following  reinjection  in  a  sensitized  animal  is  far  in 
excess  of  that  which  could  be  derived  solely  from  the  injected  antigen, 
and  in  this  he  has  been  confirmed  by  many  other  workers,  notably  by 
Vaughan.  Moreover,  recent  investigations  of  Jobling  and  others  have 
seemed  to  show  that  proteolysis  is  not  necessarily  a  function  of 
antibodies  but  is  accomplished  by  non-specific  proteoses  in  the  blood. 

It  would  seem  to  us  that  our  present  knowledge  of  this  phase  of 
anaphylactic  investigation  permits  us  only  to  conclude  that  wherever 
proteolytic  changes  take  place  these  "proteotoxins"  may  be  formed. 
That  they  can  be  produced  from  a  protein  antigen  has  been  shown 
beyond  doubt  by  Vaughan  and  his  collaborators  for  both  formed  and 
unformed  antigens.  Also  this  is  evident  from  the  experiments  of 
many  workers  and  has  been  confirmed  in  our  own  experience  with 
poisons  appearing  during  the  autolysis  of  bacterial  emulsions.  On 
the  other  hand,  it  is  also  clear  that  the  antigen  need  not  represent 
the  matrix  which  furnishes  the  poison,  and  that  in  the  reactions  as 
they  are  generally  performed  both  in  the  test  tube  and  in  the  animal 
body,  it  is  more  than  likely  that  if  an  antigen  participates  at  all  in 
furnishing  the  substratum  for  the  poison,  this  is  probably  less  impor- 
tant than  that  furnished  by  the  animal's  own  proteins.  However, 
this  does  not  weaken  the  importance  of  the  knowledge  that  the 
antigen-antibody  reaction  in  the  presence  of  normal  serum  and  cer- 
tain antigens  in  the  presence  of  normal  serum  alone,  induce  a  reac- 
tion in  the  course  of  which  such  poisons  are  formed.  And  the  fact 

40  Jobling  and  Petersen.     Jour.  Exper.  Med.,  1914,  xix,  No.  5. 


ANAPHYLAXIS'  397 

that  they  can  be  produced  experimentally  in  the  peritoneal  cavity 
of  a  living  guinea  pig  renders  their  participation  in  such  reactions  in 
the  animal  body  a  likely  assumption. 

Our  own  work  41  on  these  substances  induces  us  to  believe  that 
proteotoxins  so  formed  are  identical  with  Bail's  aggressins,  a  point 
to  which  we  will  later  refer. 

It  is  plain  that  the  foregoing  work  has  done  away  almost  entirely 
with  the  idea  that  anaphylactic  phenomena  were  chiefly  intravascular. 
The  brilliant  experiments  of  Vaughan,  of  Friedemann,  and  of  Fried- 
berger  seemed  entirely  convincing.  They  had  succeeded  in  producing 
a  toxic  substance  apparently  by  cleavage  of  typical  antigens  in  the 
test  tube  by  chemical  and  serological  methods,  and,  on  the  face  of  it 
there  seemed  no  reason  to  doubt  that  such  reactions  might  occur  in 
the  circulating  blood,  thus  explaining  anaphylaxis.  It  has  been  on 
the  basis  of  the  work  of  these  pioneers,  and  experiments  of  others 
based  on  the  principles  of  this  early  work,  that  the  intravascular 
theories  of  anaphylaxis  mechanism  have  been  founded.  For  a  time 
these  views  held  the  center  of  the  stage,  and  in  their  elaboration  many 
conditions  were  revealed  which  we  believe  have  important  bearing 
on  anaphylaxis  in  rabbits,  perhaps  in  dogs,  and  in  such  special  forms 
of  anaphylaxis  as  that  produced  by  cellular  antigens  such  as  bac- 
teria. However,  from  the  beginning  there  have  been  irrefutable 
experimental  data  which  have  prevented  the  complete  acceptation 
of  the  intravascular  theory,  indeed  have  caused  it  to  seem  likely  that 
in  guinea  pigs  anaphylaxis  was  purely  a  cellular  phenomenon  and 
that  in  anaphylaxis  in  general  the  intravascular  occurrences  were, 
if  at  all  important,  secondary  to  the  cellular  mechanism.  The  most 
important  point  which  has  always  inclined  workers  to  come  back  again 
and  again  to  the  assumption  of  a  participation  of  the  body-cells,  is 
the  necessary  interval  that  must  elapse  (certainly  in  guinea  pigs,  and 
with  almost  equal  certainty  in  other  animals)  between  the  time  of 
injection  of  an  antiserum,  and  the  development  of  the  anaphylactic 
state  in  the  animal. 

This  observation,  made  by  many  of  the  earlier  writers,  has  been 
the  crucial  point  of  controversy  not  yet  completely  settled.  The  work 
of  Weill-Halle  and  Lemaire,  which  we  have  cited  in  a  preceding- 
section,  seemed  to  show  that  the  interval  was  not  always  necessary. 
Eichet,  too,  obtained  experiments  with  crepitin  which  seemed  to  ex- 
clude the  necessity  of  the  interval.  It  is  interesting  to  note  that  he 
spoke  of  his  experiments  as  "reaction  anaphylactique  in  vitro/'  He 
sensitized  a  dog  to  crepitin,  then  bled  him  during  the  hypersusceptible 
period,  mixed  the  serum  with  a  harmless  dose  of  crepitin,  and  in- 
jected the  mixture  into  a  normal  dog.  Violent  anaphylaxis  resulted 
almost  immediately. 

At  about  the  same  time  Friedemann  published  his  very  impor- 

41  Zinsser  and  Dwyer.  .  Jour.  Exper.  Med.,  1914,  xx';  No.  6. 


398  INFECTION  AND  RESISTANCE 

tant  studies  on  the  mechanism  of  anaphylaxis  in  rabbits.  He  found 
that  passive  sensitization  in  these  animals,  in  contrast  to  the  work  of 
others  upon  guinea  pigs,  was  best  obtained  by  the  simultaneous  in- 
travenous injection  of  antigen  and  anaphy lactic  serum.  If  the  injec- 
tion of  the  sensitive  serum  preceded  that  of  the  antigen  by  as  much 
as  24  hours,  the  reaction  became  indistinct  (undeutlich),  and  Friede- 
mann  concluded  that  here,  at  least,  there  could  not  be  assumed  the 
necessity  of  preliminary  sensitization  of  the  body  cells  by  the  anaphy- 
lactic  serum,  as  is  the  case  of  guinea  pig  anaphylaxis.  The  anaphy- 
lactic  poison,  whatever  it  may  be,  Friedemann  concludes  is,  in  rabbits 
at  least,  formed  in  the  circulating  blood.  In  1910  Biedl  and  Kraus  42 
obtained  immediate  and  severe  symptoms  in  guinea  pigs  when  they 
injected  intravenously  mixtures  of  horse  serum  together  with  the 
serum  of  sensitized  guinea  pigs.  Briot  43  in  the  same  year  obtained 
reactions  in  young  rabbits  into  which  he  had  injected  mixtures  of 
horse  serum  and  anti-horse  serum.  Gurd  44  in  a  recent  publication 
obtained  reactions  in  guinea  pigs  when  he  injected  intravenously 
immune  rabbit  serum  (anti-sheep  serum)  and  immediately  thereafter 
sheep  serum.  We  ourselves  have  been  able  to  obtain  occasional  and 
distinct  results  in  rabbits  and  guinea  pigs  both  by  simultaneous  and 
immediately  consecutive  intravenous  injections  of  antigen  and  anti- 
body, though  we  did  not  succeed  in  attempts  to  duplicate  exactly  the 
experiments  of  Friedemann  and  of  Biedl  and  Kraus. 

Recently  too  Manwaring  and  his  co-workers  studying  the  isolated 
hearts  of  rabbits,  and  the  isolated  lungs  of  guinea  pigs,  by  physio- 
logical methods,  have  claimed  that  they  can  observe  reactions  due  to 
the  meeting  of  antigen  and  antibody  in  the  blood  vessels. 

There  is  thus  a  considerable  amount  of  evidence  that  in  rabbits, 
and  some  other  animals,  the  meeting  of  antigen  and  antibody  in  the 
blood  stream  may  lead  to  anaphylactic  reaction.  However,  the  evi- 
dence in  guinea  pigs  that  this  may  occur  is  very  slim,  and  to  any  one 
who  has  worked  in  anaphylaxis  consistently  it  is  quite  plain  that  in 
these  animals  the  simultaneous  injection  of  the  two  substances  leads  to 
either  no  symptoms,  or  such  very  slight  ones  that  they  play  no  im- 
portant part  in  the  fatal  reaction.  Even  in  other  animals,  further- 
more, we  ourselves  feel  that  mere  contact  within  the  blood  stream  be- 
tween antigen  and  antibody  cannot  account  for  the  entire  train  of 
phenomena,  and  one  is  forced  to  assume  that  any  considerable  degree 
of  hypersusceptibility  must  be  essentially  cellular  and  due  to  the 
reaction  of  injected  antigen  with  antibody  that  has  previously  be- 
come attached  in  some  way  to  tissue  cells. 

The  idea  in  itself  is  not  new.  Wassermann  had  first  suggested  it  in 
an  attempt  to  explain  the  peculiar  hypersusceptibility  to  toxin  pos- 

42  Biedl  and  Kraus.     Ztschr.  f.  Immunitatsforsch.,  1910,  iv. 

43  Briot.     Compt.  rend.  Soc.  de  biol,  1910.  Ixviii,  402. 

44  Gurd.     Jour.  Med.  Besearch,  1914,  xxxi,  205. 


ANAPHYLAXIS  399 

sessed  by  some  of  Behring's  toxin-immunized  animals.  He  assumed 
that  in  such  animals  the  formation  of  antitoxins  may  indeed  have 
been  stimulated,  but  that  much  of  it  might  still  be  attached  to  the 
generating  cells  themselves,  thereby  rendering  these  proportionately 
more  vulnerable  to  the  injected  toxin. 

Such  a  conception  of  "sessile  receptors"  was  applied  by  Fried- 
berger  to  anaphylaxis  in  his  first  attempts  to  formulate  an  hy- 
pothesis. He  assumed  that  at  the  first  or  sensitizing  injection  the 
production  of  antibodies  (precipitins)  was  stimulated.  These,  how- 
ever, were  not  produced  in  great  quantity  and  were  not  discharged 
into  the  circulation,  possibly  owing  to  the  small  single  dose  given 
for  sensitization.  They  were  present  at  the  end  of  the  anaphylactic 
incubation  time  as  sessile  receptors  or  sessile  antibodies  (precipitins). 
On  the  second  injection  a  reaction  occurred  between  the  injection 
antigen  and  these  sessile  precipitins  and  the  cell  was  injured  because 
the  reaction  occurred  on  its  substance,  a  reaction  which,  it  is  sug- 
gested, might  have  been  harmless  had  it  taken  place  in  the  blood 
stream.  In  passive  sensitization,  conversely,  no  injury  could  result 
until  considerable  quantities  of  the  antibody  had  become  united  to 
body  cells  in  the  course  of  several  hours.  That  the  antibody  injected 
into  passively  sensitized  animals  indeed  disappears  from  the  circula- 
tion with  relative  speed,  has  been  shown  by  Doerr  and  again  recently 
by  Weil. 

Direct  study  of  the  cellular  conception  was  made  possible  by 
methods  such  as  the  transfusion  method  employed  in  anaphylactic 
dogs  by  Pearce  and  Eisenbrey45  and  the  technique  of  observing 
isolated  tissues  from  anaphylactic  animals  as  used  by  Schultz  4( 
work  which  appeared  as  early  as  1910.  Pearce  and  Eisenbrey  work- 
ing with  two  normal  and  one  sensitized  dog,  transfused  the  blood 
of  one  of  the  normal  animals  into  the  sensitized  one,  transferring  the 
blood  of  the  latter  to  the  normal  dog.  "At  the  proper  moment 
the  normal  dog  containing  the  blood  of  the  sensitized  animal  and 
the  latter  containing  the  blood  of  the  normal  dog,  each  received 
intravenously  the  toxic  dose  of  horse  serum."  The  normal  dog  having 
the  sensitized  blood  did  not  react,  the  sensitized  dog  having  the  nor- 
mal blood  showed  typical  fall  of  blood  pressure.  Pearce  and  Eisen- 
brey drew  the  conclusion  "that  the  phenomenon  of  anaphylaxis  is  due 
to  a  reaction  in  the  fixed  cells  and  not  either  primarily  or  secondarily 
in  the  blood."  Similar  experiments  have  since  been  done  by  Weil. 

In  the  same  year  Schultz  began  to  work  with  what  is  now  spoken 
of  as  the  physiological  method.  He  determined  that  smooth  muscle 
— freshly  excised  from  various  animals — will  react  with  contraction 
when  brought  into  contact  with  serum.  When  such  muscle  was 
taken  from  anaphylactic  animals  after  being  thoroughly  washed  free 

45  Pearce  and  Eisenbrey.     Cong.  Am.  Phys.  and  Surg.,  1910,  viii. 

46  Schultz.     Jour.  Pharmacol.  and  Exper.  Therap.,  1910,  i. 


400  INFECTION  AND  RESISTANCE 

of  blood,  it  would  react  more  energetically  and  to  small  amounts  of 
the  homologous  serum.  There  are  many  interesting  by-products  of 
Schultz's  work,  such  as  the  differences  between  fresh  arterial  blood 
and  blood  serum  in  their  abilities  to  stimulate  contraction,  but  this 
and  other  points  will  not  be  discussed  at  present.  The  important  and 
incontrovertible  fact  established  by  Schultz  is  the  changed  reaction- 
energy  or,  in  truth,  "allergie"  of  the  smooth  muscle  of  anaphylactic 
animals  to  the  stimulus  of  the  sensitizing  antigen.  Dale  47  has  con- 
firmed and  extended  these  observations  of  Schultz.  He  removed  the 
uteri  from  guinea  pigs  after  thoroughly  perfusing  them  with  Ringer's 
solution  to  remove  all  blood.  He  then  suspended  them  in  baths  of 
Rimer's  solution  and  by  the  customary  physiological  methods  meas- 
ured the  contractions  following  the  addition  of  various  amounts  of 
foreign  protein  in  the  form  of — among  other  things — horse  serum 
and  beef  serum.  He  found  that  the  uterus  of  an  animal  sensitized 
to  horse  serum  would  react  to  this  substance  in  dilutions  of  1 :2,000 
or  1 : 10,000,  while  the  organ  taken  from  a  normal  guinea  pig  reached 
its  limit  of  reactionability  at  dilutions  often  less  than  1 :200.  A 
uterus  that  had  reacted  strongly  was  found  to  be  subsequently  desen- 
sitized. A  normal  uterus  could  not — strangely — be  passively  sen- 
sitized by  immersion  into  a  solution  containing  serum  antibodies. 
This  method  of  investigation  has  recently,  also,  been  taken  up  by 
Richard  Weil  48  who  has  fully  confirmed  the  principles  laid  down 
by  Schultz  and  Dale.  He  has  incidentally  also  answered  an  objec- 
tion to  the  conclusions  of  Dale  and  Schultz  (never  indeed  a  very 
valid  objection),  namely,  that  the  reaction  of  the  muscle  tissue  of  a 
sensitized  animal  might  be  in  part  due  to  the  fact  that  the  blood, 
i.e.,  the  antibodies,  had  not  been  entirely  washed  out  of  the  tissue 
spaces  by  perfusion.  Weil  performed  the  very  simple  and  ingenious 
experiment  of  injecting  a  normal  guinea  pig  with  large  amounts  of 
immune  serum  (anti-horse  serum)  and,  after  a  few  minutes,  killing 
the  animal.  He  then  suspended  the  uterus  in  Ringer's  solution  in 
the  usual  manner  without  washing  it  completely  free  of  blood.  Con- 
tact with  the  homologous  antigen  produced  no  response.  We  may 
accept  as  definitely  established  by  these  researches  of  Schultz,  Dale, 
and  Weil  that  the  fixed  cells  of  anaphylactic  animals  possess  an  in- 
creased reaction-ability  toward  the  antigen  which  is  in  no  sense  sec- 
ondary to  processes  involving  the  circulating  antibodies.  Moreover, 
the  work  of  We1"!  seems  to  indicate  that  desensitization  of  a  pas- 
sively prepared  guinea  pig  deprives  the  uterus  of  its  power  to  re- 
spond and  that  the  gradual  spontaneous  diminution  of  hypersuscep- 
tibility  on  the  part  of  the  guinea  pig  is  accompanied  by  an  entirely 
parallel  loss  of  reaction-capacity  on  the  part  of  the  isolated  uterus. 

47  Dale.    Jour.  Pharmacol.  and  Exper.  Therap.,  1913,  iv. 

48  Weil,  R.     Jour.  Med.  Research,  27,  1913;  30,  1914;  Proc.  Soc.  Exper. 
Biol.  and  Med.,  1914,  xi,  86. 


ANAPHYLAXIS  401 

The  recent  work  of  Coca,49  too,  has  further  fortified  the  cellular 
point  of  view  by  a  method  which  in  principle  is  similar  to  that 
employed  by  Pearce  and  Eisenbrey.  Coca  succeeded  in  perfusing 
actively  and  passively  sensitized  guinea  pigs  with  the  defibrinated 
blood  of  normal  guinea  pigs  in  such  a  way  that  the  original  blood  of 
the  sensitized  animals  was  reduced  to  a  necessarily  slight  residue. 
Animals  so  treated  could  be  kept  alive  for  as  long  as  six  hours  after 
the  transfusions  and  remained  delicately  hypersusceptible  in  spite 
of  the  blood  substitution. 

Limiting  ourselves  for  the  present  to  the  phenomena  of  anaphy- 
laxis  in  which  noncellular  antigens  are  employed,  we  may  safely 
say  that  the  evidence  furnished  by  the  incubation  time  necessary  in 
passive  anaphylaxis,  by  the  transfusion  experiments  of  Pearce  and 
Eisenbrey  and  of  Coca,  and  most  conclusively  by  the  work  on  iso- 
lated tissue  by  Schultz,  by  Dale  and  by  Weil,  shows  conclusively 
that  the  hypersusceptible  state  is  largely  determined  by  a  changed 
reaction-capacity  to  the  specific  antigen  on  the  part  of  the  fixed  tissue 
cells — an  "alergie"  which  is  probably  due  to  the  presence  of  specific 
antibodies  in  the  substance  of  the  cell  protoplasm,  and  incidentally 
accounts  for  such  effects  as  the  skin  reactions.  It  is  probable  that 
the  acute  symptoms  and  death  of  anaphylactic  guinea  pigs  (and 
perhaps  of  other  animals)  is  in  most  cases  of  experimental  ana- 
phylaxis due  to  the  reaction  which  takes  place  between  the  injected 
antigen  and  these  sessile  receptors. 

Now,  as  to  the  identification  of  the  anaphylactic  antibody  with 
some  one  of  the  well-known  antibodies,  the  assumption  is  that  ia 
cellular  anaphylaxis  (as  in  the  corpuscle  experiments  of  Friede- 
mann  and  in  the  bacterial  experiments  of  Friedberger  and  others) 
the  so-called  sensitizer  or  amboceptor  is  to  be  held  responsible.  This 
seems  reasonable,  and  there  is  much  evidence  in  its  favor,  no  reliable 
evidence  against  it. 

In  the  case  of  serum  anaphylaxis  extensive  work  has  been  done 
to  show  a  parallelism  between  the  anaphylactic  antibody  and  the 
precipitins.  This  we  have  seen  principally  in  the  experiments  of 
Doerr  and  Russ,  and  those  of  Friedberger. 

The  problem  becomes  a  complicated  one  when  we  attempt  then 
to  define  the  nature  of  the  precipitins  and  their  relation  to  the  anti- 
bodies hypothetically  advanced  as  "albuminolysins"  by  Gengou> 
Without  going  into  this  point  extensively  at  present,  it  may  be  per- 
mitted to  refer  the  reader  to  the  chapters  on  alexin  fixation  and 
precipitins,  and  to  reiterate  the  writer's  50  own  opinion,  which  is 
that  much  reasonable  evidence  points  to  the  fact  that  the  so-called 
precipitins  are  in  truth  protein-sensitizers,  in  structure  and  function 
identical  with  the  sensitizers  or  amboceptors  of  cytolytic  processes. 

49  Coca.     Ztschr.  f.  Immunitatsforsch.,  1914,  xx. 

50  Zinsser.     Jour.  Exp.  Med.,  Vol.  15,  1912,  and  Vol.  18,  1913. 


T 


402  INFECTION  AND  RESISTANCE 

The  fact  that  precipitation  occurs  when  these  antibodies  are  added 
to  the  homologous  dissolved  antigen  is  merely  a  secondary  colloidal 
phenomenon.  Antigen  and  antibody  react,  forming  a  complex  which 
is  then  amenable  to  the  action  of  alexin.  Being  colloidal  in  nature, 
and  mixed  under  suitable  quantitative  and  other  conditions  which 
favor  flocculation,  they  precipitate.  This  point  of  view,  then,  identi- 
fies the  so-called  precipitins  with  the  protein-sensitizers  or  albumino- 
lysins  first  hypothetically  suggested  by  Gengou.  It  leads  necessarily 
to  the  conception  that  in  cytolysis  as  well  as  proteolysis,  in  fact,  in 
all  reactions  in  which  antigen  is  sensitized  to  the  action  of  alexin, 
there  is  functionally  but  one  variety  of  antibody — the  sensitizer 
, — precipitation  and  agglutination  being  incidental  physical  phe- 
nomena not  dependent  upon  special  antibodies  as  heretofore  sup- 
posed. 

In  this  sense,  then,  the  "precipitins"  or  albuminolysins  may  be 
regarded  as  identical  with  the  anaphylactic  antibody. 

That  animals  in  whose  circulation  antigen  and  antibody  are 
simultaneously  present  do  not  suffer  from  symptoms  of  anaphylaxis 
has  been  referred  by  Zinsser  and  Young  51  as  possibly  due  to  the 
action  of  a  protective  colloid  which  prevents  the  union  of  the  two. 


THE  MECHANISM   OP   ANTIANAPHYLAXIS 

The  conditions  under  which  these  animals,  previously  anaphy- 
lactic, may  be  rendered  refractory  or  "antianaphylactic"  have  been 
discussed  in  another  place.  This  condition  is  not  entirely  comparable 
to  immunity  since  it  is  a  purely  temporary  state,  lasting  possibly  a 
few  weeks,  but  after  this  the  animals  do  not  return  to  the  normal 
condition,  but  gradually  become  again  moderately  hypersusceptible. 
(Rosenau  and  Anderson — Otto  and  others.)  Thus  a  guinea  pig 
which  has  been  sensitized,  then  rendered  antianaphylactic  by  a  mas- 
sive injection  of  antigen,  may  react  with  mild  symptoms  to  an  in- 
jection made  20  to  30  days  later.  Such  returning  sensitiveness, 
according  to  Rosenau  and  Anderson  52  is  usually  mild,  fatal  reac- 
tions rarely  occurring. 

A  satisfactory  theory  of  antianaphylaxis  has  not  yet  been  ad- 
vanced. 

Besredka,53  as  we  have  seen,  believes  that  the  anaphylactic  reac- 
tion takes  place  by  the  union  of  the  toxic  factor  in  the  serum  (anti- 
sensibilisin)  with  a  specific  antibody  sessile  upon  the  cells  of  the 

61  Zinsser  and  Young.    Jour.  Exp.  Med.,  Vol.  17,  1913. 

52  Rosenau  and  Anderson.     Pub.  Health  and  M.  H.  S.  Hyg.,  Bull.}  36, 
1907. 

53  See  Besredka,  "Kraus  u.  Levaditi  Handbnch,  etc.,"  Erganzungsband  ], 
p.  246. 


ANAPHYLAXIS  403 

central  nervous  system.  If  the  antigen  is  injected  slowly  or  in  small 
amount  these  sessile  receptors  are  gradually  united  to  antigen  with- 
out fatal  shock,  and  the  animal  is  thereby  rendered  insensitive. 

In  his  own  words,  this  "desensitization"  amounts  to  a  return  of 
the  cells  to  their  normal  preanaphylactic  or  naturally  unsensitive 
condition.  With  the  refutation  of  his  theory  of  anaphylaxis,  his 
theory  of  anti anaphylaxis  also  falls  to  the  ground,  and  neither  of  the 
two  can  be  accepted  as  valid  at  present. 

If  we  look  upon  anaphylaxis  as  a  reaction  taking  place  entirely  in 
the  circulation  we  may  accept,  with  Rosenau  and  Anderson  54  Fried- 
berger,  and  others  the  explanation  that  antianaphylaxis  is  due  to  a 
saturation  of  the  anaphylactic  antibody  with  antigen.  Hypersus- 
ceptibility  is  then  subsequently  reestablished  because  a  gradual 
formation  of  circulating  antibody  continues,  and  eventually  free 
antibody  will  again  be  present  in  the  blood.  This  view  is  only  in 
part  satisfactory,  as  Friedemann  55  points  out.  For  it  does  not 
explain  the  antianaphylaxis  which  Biedl  and  Kraus  56  have  noticed 
after  the  injection  of  mixtures  of  antigen  and  antibody,  nor  the  non- 
specific antianaphylaxis  which  the  same  workers  have  observed  after 
peptone  injections.  It  is  clear  that  the  nature  of  antianaphylaxis 
remains  for  the  present  obscure,  and,  in  view  of  the  recent  attempts 
to  account  for  certain  phases  of  infectious  disease  by  the  anaphy- 
lactic phenomena,  is  one  of  the  most  important  problems  of  im- 
munity. 

Bearing  upon  this  condition  of  antianaphylaxis  is  the  tolerance 
to  the  anaphylactic  poison  which  has  been  observed  to  develop  in 
animals  once  or  twice  injected.  Vaughan  57  has  noticed  this  in  ani- 
mals injected  with  his  toxic  split  products,  produced  by  alkaline- 
alcohol  splitting  of  colon  bacilli.  By  repeated  injection  of  the  guinea 
pigs  he  showed  that  a  tolerance  was  developed  which  protects  the 
animal  from  about  double  the  fatal  dose,  but  the  animal  is  not  pro- 
tected against  larger  multiples,  and  the  condition  is  not  an  immunity 
in  the  sense  in  which  we  have  used  the  term.  Similar  observations 
have  been  made  by  Bessau.58  Bessau  passive  sensitized  guinea  pigs 
with  1  c.  c.  of  anti-horse  serum  intraperitoneally,  and  on  the  follow- 
ing day  injected  them  intravenously  with  1  c.  c.  of  horse  serum.  He 
gauged  his  dose  so  that  the  animals  should  have  severe  shock  but 
survive.  One  or  two  days  later  he  injected  the  amount  of  typhoid 
anaphylatoxin  which  was  fatal  for  normal  pigs,  and  found  that  those 

54  Rosenau  and  Anderson.     U.  S.  Pub.  Health  and  M.  H.  S.  Hyg.  Lab. 
Bull.,  64,  1910. 

55  Friedemann.     "Frei   Vereinigung   f.    Mikrobiol.,"   Berlin,    1910.     Ref. 
Centralbl.  Bakt.  I,  Vol.  47;  Beiheft,  p.  1,  1910. 

56  Biedl  and  Kraus.     Zeitschr.  f.  Imm.,  Vol.  4,  1910. 

57  Vaughan.     "Protein  Split  Products,  etc.,"  Lea  &  Febiger,  Philadelphia, 
1913,  p.  139. 

58  Bessau.     Centralbl.  f.  Bakt.,  Vol.  60,  1911. 


404  INFECTION  AND  RESISTANCE 

which  had  been  treated  as  described  were  now  able  to  withstand  the 
anaphylatoxin  These  experiments  of  Bessau  would  indicate  that 
antianaphylaxis  was  to  a  certain  extent  due  to  tolerance  to  the  poison, 
and  that  it  was  non-specific.  Friedberger,  together  with  Szymanow- 
ski,  Kumagai,  Odaira  and  Lura,  later  studied  this  problem  and  came 
to  the  conclusion  that  antianaphylaxis  is  strictly  specific,  depending, 
as  Friedberger  had  suggested,  upon  the  diminution  of  specific  anti- 
bodies rather  than  upon  tolerance  to  the  poison.  They  claimed  that 
animals  that  had  been  sensitized  and  then  had  survived  the  "shock" 
dose  of  homologous  protein  showed  no  tolerance  for  anaphylatoxin, 
and  that  animals  that  had  been  treated  with  the  sublethal  dose  of 
anaphylatoxin  are,  for  24  hours,  as  sensitive  to  anaphylatoxin,  how- 
ever prepared,  as  are  normal  animals.  Recent  studies  along  the 
same  lines  by  Zinsser  and  Dwyer  59  have  yielded  results  differing 
from  these  conclusions.  Working  with  typhoid  anaphylatoxin  they 
found  that  guinea  pigs  treated  with  a  sublethal  dose  of  anaphylatoxin 
develop  a  tolerance  which  enabled  them  to  resist  1%  to  2  units  of 
poison,  the  tolerance  developing  within  three  days  and  lasting,  to  a 
slight  degree,  for  as  long  as  two  months.  It  seemed  to  them  that 
animals  treated  with  a  second  dose  of  anaphylatoxin  within  24  hours 
after  the  first,  if  the  results  of  this  first  injection  have  been  severe,  as 
they  usually  are,  are  still  weak  and  generally  depressed  in  vitality 
so  that  a  developed  tolerance  may  be  clouded  by  this  condition.  The 
tolerance  did  not  seem  to  be  strictly  specific  in  that  typhoid  anaphy- 
latoxin seemed  to  produce  a  moderate  tolerance  to  prodigiosus 
anaphylatoxin. 

It  would  seem,  therefore,  that  in  antianaphylaxis  we  might  have 
two  very  important  elements.  The  one  strictly  specific  depends  upon 
the  depletion  of  antigen  from  the  body,  a  true  "desensitization."  The 
other  non-specific,  and  probably  of  secondary  importance  since  so 
far  it  has  not  been  shown  to  any  very  powerful  degree,  consists  of 
the  development  of  tolerance  by  the  body  cells  for  the  anaphy lactic 
poison. 

NATURE    OF   ANAPHYLACTIC    POISON 

As  to  the  nature  of  the  anaphylactic  poison  we  are  also  to  a  large 
extent  in  the  dark.  From  the  experimentation  upon  the  production 
of  these  poisons  in  vitro  it  appears  that  they  are  protein  cleavage 
products.  This  is  indirectly  indicated  also  by  metabolism  experi- 
ments— such  as  those  of  Friedemann  and  Isaak,60  and  of  Weichhardt 
and  Schittenhelm.61  It  appeared  from  this  work  that,  as  measured 

59  Zinsser  and  Dwyer.    Reported  at  the  meeting  of  Am.  Ass.  of  Path,  and 
Bact.,  Toronto,  April,  1914. 

60  Friedemann  and  Isaak.    Zeitschr.  f.  exp.  Path.  u.  Ther.,  Vol.  1,  1905. 

61  Schittenhelm  u.  Weichhardt.     Munch,  med.  Woch.,  1910,  No.  34,  and 
1911. 


ANAPHYLAXIS  405 

by  nitrogen  output,  the  cleavage  of  foreign  protein  injected  into 
specifically  sensitized  or  immunized  dogs  occurred  with  much  greater 
energy  and  speed  than  occurred  in  normal  animals  after  first  in- 
jection. 

Attempts  to  obtain  the  poison  by  non-specific  methods — that  is, 
by  purely  chemical  processes  without  the  agencies  of  alexin  and  sen- 
sitizer  or  antibody,  have  been  made  with  apparent  success  by 
Vaughan  and  Wheeler,62  whose  toxic,  alcoholic-soluble  fraction  (ob- 
tained by  boiling  egg  white  in  absolute  alcohol  containing  2  per  cent. 
NaOH),  seems  to  produce  typical  anaphylaxis  in  guinea  pigs.  This 
substance  Vaughan  and  Wheeler  regard  as  a  protein,  whereas  Wells  63 
states  that  it  may  be  this,  or  a  "soluble  peptone  or  polypeptid,  con- 
taining enough  of  the  different  aminoacids  to  give  all  the  usual  reac- 
tions." Weichhardt,64  too,  has  obtained  similar  poisons  by  a  method 
similar  in  principle  to  that  of  Vaughan  and  Wheeler. 

This  substance  is,  according  to  him,  pharmacologically  identical 
with  his  "keno  toxin/'  or  fatigue  toxin,  obtained  in  the  washings 
from  the  muscles  of  excessively  fatigued  animals. 

Accurate  chemical  definition  of  the  anaphylactic  poison  has  not 
so  far  been  accomplished,  and  it  is  obvious  that  the  problem  is  an 
extremely  difficult  one.  Biedl  and  Kraus,65'  66  however,  have  drawn 
a  very  close  parallelism  between  anaphylactic  intoxication  and  pep- 
tone poisoning  in  dogs.  They  have  shown  that  peptone  (0.3  gr.  to  the 
kilo.)  injected  into  these  animals  gives  rise  to  the  same  clinical 
symptoms  that  characterize  anaphylaxis.  It  is  accompanied  also  by 
typical  fall  of  blood  pressure,  delayed  coagulability  of  the  blood,  and 
leukopenia.  Furthermore,  they  claim  that  the  injection  of  sublethal 
doses  of  Witte  peptone  into  serum-sensitized  dogs  leads  to  a  non- 
specific anti-anaphylaxis.  They  claim  a  physiological  identity  of  the 
Witte  peptone  with  the  anaphylactic  poison. 

This  last  observation  could  not  be  confirmed  by  Manwaring,67 
who  found  that  dogs  that  had  been  rendered  anti-anaphylactic  to- 
horse  serum  still  reacted  strongly  to  peptone — an  observation  which 
does  not  indeed  weaken  the  contention  of  Biedl  and  Kraus  as  to  the 
similarity  of  peptone  shock  to  anaphylaxis,  but  has  significance  in 
contradicting  the  doubts  their  experiments  have  thrown  on  the 
specificity  of  antianaphylaxis. 

Observations  similar  to  those  of  Biedl  and  Kraus  on  the  toxic 
action  of  peptone  have  been  made  by  Arthus.68 

62  Vaughan  and  Wheeler.    Loc.  cit. 

63  Wells.     Jour.  Inf.  Dis.,  Vol.  5,  1908. 

64  Weichhardt.     Zentralbl.  /.   die  ges.  Pliys.  u.  Path,  des  Staff wechsels, 
No.  15,  1909.     Ref.  "Weichhardt's  Jahresbericht,"  1910,  p.  554. 

65  Biedl  and  Kraus.     Wien.  klin.  Woch.,  No.  11,  1909. 

66  Also  "Kraus  n.  Levaditi  Handhuch,  etc.,"  Erg-anzung-sband  1,  p.  26 l. 

67  Manwaring.     Zeitschr.   f.   Imrmtnitntsforschung,  Vol.  8,  p.  589,  1911. 

68  Arthus.     C.  E.  tie  I'Acad.  des  Scienc.,  Vol.  148. 


406  INFECTION  AND  RESISTANCE 

Biedl  and  Kraus  have  found  a  similar  parallelism  in  guinea  pigs 
in  which  they  determined  the  typical  bronchial  spasms  after  peptone 
administration.  This  is  in  contrast  to  Werbitzky,69  who  found  even 
large  doses  of  peptone  non-toxic  for  guinea  pigs.  Nevertheless,  there 
is  no  question  that  the  similarity  between  peptone  shock  and  anaphy- 
laxis  is  very  striking  and  of  great  theoretical  importance.  It  does 
not,  however,  bring  us  much  nearer  to  a  chemical  understanding  of 
the  nature  of  the  poisons  since  the  "Witte"  peptone  used  in  these  ex- 
periments is  a  mixture  of  many  different  substances.  Brieger,70 
for  instance,  found  toxic  and  non-toxic  lots  of  Witte  peptone.  The 
toxic  ones  yielded  on  extraction  a  body  which  he  calls  peptotoxin. 
This  variation  in  the  constitution  of  different  samples  of  so-called 
"pepton"  may  account  for  some  of  the  conflicting  results  obtained  in 
guinea  pigs.71 

Recently  Dale  has  suggested  that  B-imidazolethylamin  or  his- 
tamin  may  be  the  active  principle  concerned  in  anaphylactic  shock. 
Intravenous  injection  of  0.5  mg.  of  this  substance  into  large  guinea 
pigs  results  in  typical  respiratory  difficulties,  convulsions  with  death 
and  distention  of  the  lungs  typical  of  anaphylactic  shock.  Treat- 
ment with  atropin  diminishes  this  action,  just  as  Auer  and  Lewis 
found  this  to  be  the  case  in  true  anaphylaxis,  and  fall  in  blood 
pressure  also  occurs.  It  would  seem  then  that  substances  represent- 
ing cleavage  products  of  native  proteins  of  highly  complex  nature, 
the  result  of  proteolytic  cleavage  not  very  far  advanced,  are  prob- 
ably concerned  in  the  production  of  anaphylactic  shock.  The  ana- 
phylatoxins  of  Friedberger  cannot  of  course  be  studied  chemically 
by  the  methods  to  which  Vaughan's  poisons  are  amenable. 

PHENOMENA  CLOSELY  RELATED  TO  ANAPHYLAXIS 

There  are  a  number  of  well-defined  phenomena  of  acquired 
hypersusceptibility  or  sensitiveness  which,  in  nature,  seem  closely 
analogous  to  true  anaphylaxis  as  we  understand  it  to-day,  but  re- 
garding the  mechanism  of  which  the  opinions  of  experimenters  are 
still  to  some  extent  at  variance. 

Among  the  most  important  of  these  is  the  toxic  action  of  normal 
«era  when  injected  into  animals  of  another  species — a  phenomenon 
which  is  now  generally  accepted  as  belonging  in  principle  to  the  true 
anaphylactic  phenomena,  though  this  opinion  is  of  comparatively 
recent  formulation.  The  subject  is  of  sufficient  theoretical  and 
practical  importance  to  be  considered  in  some  detail. 

The  older  studies  of  phenomena  belonging  to  this  category  fol- 

69  Werbitzky.     C.  E.  de  la  Soc.  de  BioL,  Vol.  66,  1909. 

70  Brieger.     "Die  Ptomaine,"  1,  p.  14. 

71  For  analysis  of  Witte  peptone,  see  Hammarsten,  "Physiological  Chem- 
istry," English  translation. 


ANAPHYLAXIS  407 

lowed  closely  in  the  footsteps  of  experiments  on  transfusion,  and  as 
early  as  1666  a  commission  of  the  London  Koyal  Philosophical  So- 
ciety reported  deaths  following  transfusion,  alleging  intravascular 
coagulation  as  the  probable  cause  of  deatk 

The  cause  of  death  following  injections  of  foreign  whole  blood, 
blood  cells,  and  serum  has,  since  that  time,  occupied  the  attention  of 
many  workers  whose  studies  need  not  be  reviewed  for  our  present 
purposes.  Chief  among  them  were  Magnani,  Brown-Sequard,  Ma- 
gendie,  and,  more  recently,  Naunyn,  Landois,  and  Ponfick.72 

The  work  of  Landois  is  of  special  interest  in  that  he  worked 
with  blood  serum  free  from  cells,  and  attempted  to  correlate  the 
occurrences  after  the  injection  of  animals  with  the  action  of  the 
serum  upon  the  cellular  blood  elements  in  vitro.  Landois  observed 
both  the  solution  of  hemoglobin  and  hemagglutination,  and  was 
led  to  regard  the  action  of  serum  upon  erythrocytes  as  the  primary 
cause  of  death  after  transfusion.  His  conception  of  the  mechanism 
is  apparently  twofold.  On  the  one  hand,  he  believed  that  when 
small  quantities  of  blood  were  transfused,  a  formation  of  fibrin 
(stroma-fibrin)  was  initiated  in  the  stroma  of  the  injured  erythro- 
cytes which  led  to  coagulation  and  thrombosis  in  the  capillaries  of 
the  central  nervous  system  and  lungs.  In  the  case  of  the  transfusion 
of  rabbit's  blood  into  dogs  he  attributed  death  to  embolism  in  the 
pulmonary  vessels  due  to  "Massenhafte  Verklebung  der  Kaninchen- 
zellen  im  Hundeblut" — or,  in  other  words,  to  hemagglutination. 

Ponfick  and  others  have  disputed  the  validity  of  Landois'  con- 
clusions, but  the  basic  principles  of  his  explanations  have  been  up- 
held within  recent  years  by  workers  who  have  gone  over  the  same 
ground  with  the  aid  of  more  modern  methods.  Two  careful  re- 
searches have  appeared  during  the  last  two  years  in  which  the  prob- 
lem has  been  approached  by  different  routes,  but  in  which  the  gen- 
eral conclusions  show  much  agreement.  Coca,73  investigating  the 
cause  of  death  following  the  injection  of  washed  blood  cells  into  ani- 
mals of  different  species,  concludes  that  in  these  cases  death  is  due  to 
mechanical  obstruction  of  the  pulmonary  circulation  owing  to  ag- 
glutination of  the  injected  cells.  It  is  important  to  note,  however, 
that  he  adds  in  his  conclusions  the  following  paragraph:  "The 
mere  presence  of  specific  agglutinins  does  not  suffice,  in  the  injection 
of  'toxic7  erythrocytes,  to  occlude  the  pulmonary  circulation.  The 
cooperation  of  another  factor  must  be  assumed — a  factor  probably 
found  in  the  capillary  walls." 

Loeb,  Strickler,  and  Tuttle,74  investigated  the  cause  of  death  fol- 
lowing the  injection  of  normal  dog  and  beef  sera  into  rabbits.  They 

72  A  brief  historical  review  of  this  work  can  be  found  in  the  paper  of 
€oca   (1),  Virchow's  Arch.  f.  path.  Anat.,  1909,  Vol.  196,  p.  92. 

73  Coca.     Virchow's  Archiv,  Vol.  196,  1909. 

74  Loeb,  Strickler,  and  Tuttle.     Virchow's  Archiv,  Vol.  201,  1910. 


408  INFECTION  AND  RESISTANCE 

correlated  their  animal  experiments  carefully  with  the  action  of  the 
sera  in  vitro  upon  the  blood  elements  of  rabbits,  and  utilized  the 
property  of  hirudin  to  inhibit  the  coagulation  of  blood,  finding,  in 
the  case  of  dog  serum,  that  injections  of  hirudin,  while  not  always 
preventing  death,  at  any  rate  prolonged  life  or  necessitated  an  in- 
crease in  the  lethal  dose.  The  conclusions  of  these  authors  are  as 
follows:  "Death  following  the  injection  of  foreign  serum  is  brought 
about  by  obstruction  of  the  pulmonary  circulation  either  by  heaps  of 
agglutinaled  erythrocytes  or  by  fibrinous  plugs.  Dog  serum  and  beef 
serum  represent  two  different  types.  In  the  case  of  dog  serum  hem- 
olysis  of  the  blood  cells  of  the  recipient  liberates  substances  at- 
tached to  the  stromata,  which  hasten  coagulation.  In  consequence 
fibrin  is  formed  which  is  carried  into  the  pulmonary  vessels  and 
occludes  them.  In  the  case  of  beef  serum  death  is  due  to  hemag- 
glutination." 

The  more  recent  understanding  of  the  liberation  of  toxic  bodies 
from  blood  cells  by  immune  hemolytic  sera,  especially  by  the  experi- 
ments of  Friedemann  cited  above,  have  rendered  it  likely  that  a 
similar  anaphylatoxin  formation  from  the  cells  of  the  recipient  may 
lie  at  the  bottom  of  the  toxic  action  of  normal  sera.  And  it  is  a 
fact,  indeed,  that  such  toxic  sera  are  always  hemolytic  for  the  cor- 
puscles of  the  susceptible  animal. 

An  analysis  of  the  toxic  action  of  certain  normal  sera  from  this 
point  of  view  has  been  made  by  Uhlenhuth  and  Haendel,75  who,  in 
studying  the  necrotizing  action  of  beef  serum  injected  into  guinea 
pigs,  attribute  this  action  of  the  serum  to  a  "complex  process  de- 
pending upon  the  cooperation  of  complement,"  but  not  identical 
with  the  hemolytic  mechanism.  The  toxic  action  of  such  serum,  how- 
ever, they  separate  from  the  necrotizing  action,  concluding  that  this 
is  independent  of  complement,  and  more  thermostable  than  either 
the  mechanism  causing  necrosis  or  that  responsible  for  hemolysis. 

Recent  studies  of  the  writer76  on  the  toxic  action  of  goat  serum 
for  rabbits  have  shown  that,  contrary  to  Loeb,  Strickler,  and  Tut  tie, 
hemagglutination  and  blood  coagulation  can  be  excluded  as  causes 
of  death,  and  that,  in  agreement  with  Uhlenhuth  and  Haendel,  the 
toxic  action  is  due  to  a  proteolyHc  action  on  the  part  of  the  serum 
not  necessarily  identical  with  the  hemolysins,  but  producing  from 
the  protein  of  the  recipient  a  poison  similar  to  the  anaphylatoxins. 
Unlike  Uhlenhuth  and  Haendel,,  however,  it  seemed  clear  that  the 
participation  of  alexin  was  definitely  necessary — the  process  being 
probably  entirely  analogous  to  Friedemann's  results  with  immune 
hemolytic  (cytolytic)  sera.  The  poisonous  action  of  dissolved  hemo- 
globin could  be  excluded.  In  principle,  therefore,  the  toxic  action 

75  Uhlenhuth  and  Haendel.     Zeitschr.  f.  Immunitatsforscli.,  Vol.  7,  1910. 
76Zin,sser.     Jour.  Exp.  Med.,  Vol.  14,  1911. 


ANAPHYLAXIS  409 

of  normal  sera  would  seem  to  depend  upon  a  mechanism,  similar  to 
that  of  other  anaphylactic  phenomena. 

Toxin  hyper  susceptibility,  which  is  often  acquired  by  animals 
in  the  course  of  immunization  with  diphtheria  and  tetanus  toxin,  is 
usually  classified  with  anaphylaxis,  indeed  is  often  cited  as  the 
earliest  observation  of  this  phenomenon.  However,  it  is  by  no 
means  clear  that  the  two  conditions  are  actually  analogous,  since  in 
the  case  of  the  toxins  we  are  dealing  with  antigens  which  are  not  only 
toxic  in  themselves,  but  against  which  neutralizing  antibodies  are 
formed  in  the  reacting  animal.  This  last  fact  alone  would  separate 
toxin  hypersusceptibility  sharply  from  true  protein-anaphylaxis  in 
that  entirely  different  reacting-mechanisms  seem  to  be  called  into 
play  by  the  two  varieties  of  antigen.  It  will  be  necessary,  therefore, 
to  discuss  toxin  hypersusceptibility  at  some  length. 

Probably  the  earliest  authentically  recorded  observation  is  that 
of  von  Behring,77  who  determined,  both  for  diphtheria  and  tetanus 
toxins,  that  animals  once  inoculated  with  these  poisons  were  oc- 
casionally more  sensitive  to  them  subsequently  than  were  normal 
animals.  He  spoke  of  "Gift  Ueber  empfindlichkeit"  as  a  property 
acquired  by  reason  of  a  preceding  injection,  and  the  observation  was 
further  developed  by  Knorr78'in  1895,  and  by  v.  Behring  himself, 
in  collaboration  with  Kitashima  79 — a  few  years  later.  These  writers 
showed  that  guinea  pigs  which  are  treated  repeatedly  with  small 
doses  of  diphtheria  toxin  may,  under  certain  circumstances,  not  only 
fail  to  show  immunity,  but  may  even  develop  a  susceptibility  in- 
creased to  such  an  extent  that  doses  far  too  small  to  injure  a  normal 
animal  will  cause  their  death.  Again,  in  the  case  of  diphtheria  toxin 
similar  observations  were  made  upon  horses  by  both  Salomonsen  and 
Madsen,80  and  by  Krete.81  The  last-named  worker  observed  that 
horses  that  had  been  immunized  with  diphtheria  toxin  would  often 
react  to  neutral  mixtures  of  toxin  and  antitoxin  by  which  normal 
horses  were  unaffected.  This  so-called  "paradox  phenomenon"  was 
much  discussed,  and  many  theories  advanced  to  explain  it,  a  most 
ingenious  adaptation  of  the  side-chain  theory  being  applied  to  it  by 
Kretz  82  and  by  Wassermann.83  They  assumed  that  the  partial  im- 
munization in  such  treated  animals  had  in  truth  induced  the  forma- 
tion of  excessive  receptors ;  that,  in  the  stages  of  hypersusceptibility. 
however,  these  receptors  had  not  yet  been  cast  off  from  the  cells.  In 

77  Von  Behring.     Deutsche  med.  Woch.,  1893. 

78  Knorr.     Quoted  from  Otto,  "Dissertation,"  Marburg,  1895. 

79  Von  Behring  u.  Kitashima.     Berlin.  Jclin.  Woch.,  1901. 

80  Salomonsen  et  Madsen.     Ann.  de  I'Imt.  Past.,  1897. 

81  Kretz.     Quoted  from  Otto  in  "Kolle  u.  Wassermann  Handbuch,"  Er- 
ganzungsband  2,  p.  232. 

82  Kretz.     Zeitschr.  f.  Heilknnde,  1902. 

83  Wassermann.     "Kolle  u.  Wassermann  Handbuch,'"  Vol.  4,  479.      .;... 


410  INFECTION  AND  RESISTANCE 

consequence  there  was  an  excess  of  "sessile  receptors" — by  means  of 
which  the  cell  was  rendered  more  exposed  to  toxin  action  than  it  was 
normally — it  being  still  unprotected  by  the  presence  of  freely  cir- 
culating "antitoxin"  receptors.  The  difficulties  arising  from  the 
observation  of  similar  hypersusceptibility  in  animals  whose  blood 
contained  free  antitoxin  were  disposed  of  by  Wassermann  by  the 
convenient  assumption  of  variations  of  affinity. 

He  assumed  that  the  treatment  with  toxin,  i.  e.,  the  intoxication, 
may  induce  a  condition  of  higher  affinity  for  the  poison  on  the  part 
of  the  sessile  cell  receptors,  leading  to  a  selective  toxin-absorption  by 
the  cells  and  consequent  greater  susceptibility  to  injury.  With 
Behring,  he  speaks  of  this  as  a  "histogenic  hypersusceptibility," 
implying  an  increased  vulnerability  of  the  tissue  cells. 

The  analogy  between  these  early  observations  and  the  phenomena 
which  we  now  classify  as  anaphylaxis  is  unquestionably  a  striking 
one.  However,  it  is  doubtful,  as  Friedemann  suggests,  whether  the 
two  processes  depend  upon  similar  mechanisms.  For,  as  we  have 
seen  in  the  case  of  the  sensitiveness  to  toxin,  we  are  dealing  with 
primarily  poisonous  substances  against  which  in  the  reacting  animal 
neutralizing  antibodies  are  found — a  combination  of  conditions  quite 
different  from  those  with  which  we  are  confronted  in  hypersuscepti- 
bility against  primarily  harmless  proteins.  It  is,  of  course,  possible 
that  the  toxin  hypersusceptibility  is  a  true  anaphylaxis  against  the 
toxin-protein — independent  of  the  specifically  poisonous  nature  of 
this  substance.  However,  this  is  unlikely,  since  Lowi  and  Meyer  84 
have  shown  that  with  tetanus  toxin,  the  symptoms  of  such  hypersus- 
ceptibility are  not  those  of  anaphylaxis,  but  of  increased  but  charac- 
teristic tetanus  poisoning.  The  fact  that  toxin  hypersusceptibility 
cannot  be  passively  transferred  with  the  serum  of  a  susceptible  ani- 
mal does  not  seem  to  us  a  good  argument  against  its  anaphylactic  na- 
ture, since  this,  as  we  shall  see,  is  equally  impossible  in  the  case  of 
tuberculin  susceptibility,  which  is  in  all  probability  a  modified  exam- 
ple of  true  anaphylaxis.  Lowi  and  Meyer  regard  tetanus  toxin  hyper- 
susceptibility as  a  "'summation" — meaning  thereby  that  it  depends 
upon  an  alteration  of  the  cells  of  the  spinal  cord  because  of  traces  of 
the  poison  retained  in  them.  When  the  toxin  was  given  intraneurally 
no  antitoxin  formation  occurred,  but  the  animals  developed  a  marked 
hypersusceptibility  in  the  course  of  several  weeks,  showing  that 
here,  unlike  true  anaphylaxis,  specific  antibodies  play  no  part. 

Not  unlike  toxin  hypersusceptibility  is  that  which  is  noticed  in 
the  case  of  certain  medicinal  substances.  Such  are  the  so-called 
idiosyncrasies  against  cocain,  pilocarpin,  morphin,  quinin,  and  other 
drugs.  These  conditions  have  no  direct  relation  to  anaphylaxis,  and, 

84  Lowi   and   Meyer.     Festschrift,   Schmiedeberg   Suppl.,    Arch.   f.    exp. 
Path.  u.  Therap.,  1908,  p.  355. 


ANAPHYLAXIS  411 

according  to  Hans  Meyer,85  depend  probably  upon  the  chemical 
peculiarities  of  the  tissues  of  the  individual,  such  as  calcium  con- 
tents, etc.  Hunt  86  has  also  shown  that  poison  susceptibility,  in  cer- 
tain cases,  may  be  influenced  by  the  diet. 

85  Meyer  u.   Gottlieb.     "Experimentelle  Pharmakologie,"  2d  Ed.,  Urban 
&  Schwartzenberg,  pp.  520  et  seq.,  1911. 

86  Reid  Hunt.  U.  8.  Pub.  Health  and  M.  H.  S.  Hyg.  Lab.  Bull,  69,  1910. 


CHAPTER  XVII 

BACTERIAL  ANAPHYLAXIS  AND  ITS  BEARING  ON 
THE    PROBLEMS    OF    INFECTIOUS    DISEASE 

IN  the  case  of  most  serum  reactions  the  original  observations 
were  made  upon  the  sera  of  bacteria-immune  animals,  and  later  ex- 
panded into  generalizations  applicable  to  antigens  as  a  class.  This 
was  the  case  with  the  phenomena  of  lysis,  agglutination,  and  precipi- 
tation. In  the  case  of  anaphylaxis  the  reverse  was  true.  The  fun- 
damental observations  were  made  with  non-bacterial  antigens,  but 
the  thought  that  analogous  observations  could  be  made  with  bac- 
terial proteins  was  an  obvious  one,  and  since  the  problem  was  one  of 
altered  susceptibility  there  was  great  promise  that  investigation  of 
this  subject  might  prove  of  profound  significance  for  our  knowledge 
of  the  pathology  of  infectious  diseases. 

Accordingly  Rosenau  and  Anderson,1  in  one  of  their  earliest 
researches,  carried  out  experiments  upon  the  sensitizing  properties 
of  bacterial  proteins.  They  were  successful  in  sensitizing  guinea 
pigs  with  extracts  of  colon,  tubercle,  anthrax,  and  typhoid  bacilli, 
with  Bacillus  subtilis  extracts,  and  with  those  of  yeast.  In  most 
cases  they  used  considerable  quantities  of  bacterial  extracts  and 
obtained  but  slight  or  moderate  symptoms.  However,  their  results 
were  conclusive  in  showing  that  the  anaphylactic  experiment  could 
be  carried  out  with  bacterial  proteins  and  was,  in  every  detail, 
analogous  to  the  similar  phenomena  of  serum  anaphylaxis. 

Not  only  could  the  basic  experiment  of  active  sensitization  be 
carried  out  with  these  substances,  but  it  was  found  that  the  -reaction 
here,  as  in  other  cases,  was  specific,  and  that  shock  was  followed  by 
a  period  of  "antianaphylaxis"  or  "immunity."  Rosenau  and  An- 
derson suggested  that  the  incubation  time  of  many  infectious  dis- 
eases may  be  represented  by  the  period  necessary  for  the  development 
of  susceptibility  after  a  first  injection,  and  that  tlie  crisis  of  pneu- 
monia might  possibly  find  an  explanation  in  the  analogy  with  anaphy- 
laxis. 

The  criteria  governing  the  successful  production  of  bacterial 
anaphylaxis  were  then  studied  especially  by  Kraus  and  Doerr,2  Holo- 

1  Rosenau  and  Anderson.     U.  S.  Pub.  Health  and  M.  II.  S.  Hyg.  Lab. 
Bull.  36,  1907. 

2  Kraus  and  Doerr.     Wien.  klin.  Woch.,  No.  28,  1908. 


BACTERIAL    ANAPHYLAXIS  413 

but,3  Delanoe,4  and  others,  and  the  essential  points  of  Rosenau  and 
Anderson's  experiments  were  confirmed.  Although  Kraus  and 
Doerr  succeeded  in  frequently  sensitizing  guinea  pigs  with  a  single 
injection  of  bacteria,  this  was  not  found  to  be  the  most  favorable 
method  for  sensitization.  Braun  5  obtained  entirely  negative  results 
by  such  a  procedure,  but  this  may  well  have  been  because  in  the  first 
place  single  sensitization  with  bacteria  is  evidently  irregular  in 
result,  and  because  Braun  carried  out  his  intravenous  test-injection 
slowly,  a  technique  by  which  Friedberger  found  later  that  shock 
could  be  avoided.  Delanoe,  in  the  main,  confirmed  the  fact  that 
bacterial  sensitization  was  possible,  but  denied  the  specificity  of  the 
resulting  anaphylaxis,  in  that  he  succeeded  in  producing  shock  in 
tubercle-sensitized  guinea  pigs  with  comparatively  large  amounts 
of  typhoid,  paratyphoid,  and  other  bacilli,  and  conversely  found 
typhoid-sensitized  guinea  pigs  hypersusceptible  to  tubercle-injec- 
tions. Other  workers,  however,  notably  Kraus  and  Doerr,  Holobut, 
and  Kraus  and  Admiradzibi,6  agree  that  the  reaction  is  specific,  at 
least  in  the  same  limits  within  which  other  serum  reactions  may  be 
called  specific. 

Holobut  then  developed  a  technique  of  sensitization  with  bac- 
teria more  reliable  than  any  which  had  been  previously  employed  by 
other  workers.  He  found  that  the  most  regularly  successful  results 
were  obtained  when  he  injected  small  quantities  of  bacteria  (1/100 
loopful)  daily  for  ten  days,  subcutaneously,  and  tested  with  fairly 
large  amounts  (1-2  c.  c.  of  an  emulsion  of  the  bacteria)  intrave- 
nously about  3  weeks  after  the  last  sensitizing  injection.  This  is  in 
keeping  with  later  experience,  and  in  our  own  work  with  typhoid 
immunization  in  young  goats  we  have  found  that  anaphylactic 
reactions  were  not  observed  unless  the  goats  had  previously 
received  several  injections.  A  second  injection  never  elicited 
symptoms. 

It  is  not  at  all  unlikely  that  this  difference  between  serum  sensi- 
tization and  bacterial  sensitization  is  due  to  the  comparatively  larger 
amounts  of  protein  injected  with  very  small  volumes  of  serum  than 
is  the  case  with  even  the  thickest  bacterial  emulsions.  When  larger 
sensitizing  quantities  of  bacteria  are  used — (which  is  often  difficult 
because  of  the  primarily  toxic  nature  of  some  of  the  bacteria) — a 
single  sensitization  gives  positive  results  in  guinea  pigs  more  fre- 
quently than  when  the  smaller  amounts  are  used. 

Since  it  was  objected  to  many  of  the  results  at  first  obtained  with 
bacterial  sensitization  that  they  might  have  been  due  to  the  primarily 

3  Holobut.     Zeitschr.  f.  Immunitatsforsch.,  Vol.  3,  1909. 

*  Delanoe.    C.  E.  de  la  Soc.  de  Biol,  Vol.  66,  1909,  pp.  207,  252,  348,  389. 

5  Braun.     Quoted  by  Bail  and  Weil,  Zeitschr.  f.  Immunitatsforsch.,  Vol. 
4,  1910. 

6  Kraus  u.  Admiradzibi.    Zeitschr.  f.  Immunitatsforsch.,  Vol.  4,  1910. 


414  INFECTION    AND    RESISTANCE 

toxic  nature  of  the  bacteria  or  their  extracts,  it  is  important  to  note 
that  Kraus  and  Doerr  and  later  Kraus  and  Admiradzibi  succeeded 
in  well-controlled  experiments  in  transferring  bacterial  anaphylaxis 
"passively"  with  the  serum  of  previously  sensitized  animals — not 
only  of  the  same,  but  of  other  species — (rabbit  serum  to  guinea  pigs). 
These  experiments  add  the  final  link  to  the  chain  of  complete  analogy 
between  bacterial  and  serum  anaphylaxis. 

This  analogy  was  partly  established  and,  in  its  completeness, 
clearly  foreseen,  when  Friedemann's  work  upon  the  poisons  produced 
from  cells  by  hemolytic  sera,  and  Friedberger's  similar  work  upon 
serum  precipitates,  turned  the  trend  of  anaphylactic  experimentation 
into  new  channels. 

It  will  be  remembered  that,  before  this  time,  the  toxic  action  of 
most  bacteria  (exclusive  of  "true  toxin"  producers  like  diphtheria 
and  tetanus  bacilli)  had,  since  Pfeiffer,  been  attributed  to  the  libera- 
tion of  preformed  "endotoxins"  from  the  bacterial  body  during  the 
process  of  lysis. 

This  idea  is  fundamental  to  the  opinion  of  hypersusceptibility 
expressed  by  Wolff-Eisner7  as  early  as  1904. 

The  underlying  concept  of  these  ideas  is  really  a  morphological 
one  in  which  the  "endotoxin"  is  regarded  as  something  present  in  the 
antigen  which  is  set  free  by  disintegration  of  the  cell.  In  applying 
this  to  serum  anaphylaxis  Wolff-Eisner  8  preserves  this  morphological 
simile  in  that  he  speaks  of  the  dissolved  protein  antigen  (serum, 
etc.)  as  "nur  scheinbar  gelost"  and  "dass  es  erst  durch  die  Lysine 
wirklich  resorbierbar  wird." 

Indeed  the  sudden  liberation  of  endotoxins  by  immune  sera  had 
been  regarded  by  Pfeiffer  and  others  as  the  cause  of  the  rapid  death 
often  ensuing  in  immunized  guinea  pigs  when  more  than  a  definite 
maximum  of  cholera  spirilla  or  other  organisms  was  injected.  In 
all  these  opinions  the  basic  conception  was  that  certain  bacteria  con- 
tained a  characteristic  preformed  poison  (endotoxin)  upon  the 
pharmacological  properties  of  which  the  peculiar  symptoms  caused 
by  each  organism  depended. 

The  earliest  unambiguous  statements  of  a  conception  differing 
from  this  original  view  of  the  nature  of  bacterial  endotoxins,  and 
approaching  the  later  conceptions  of  Friedberger,  are  found,  we 
believe,  in  the  work  of  Vaughan.9  In  an  article  by  him,  published  in 
1908,  Vaughan,  after  describing  the  incubation  time  occurring  in 
man  and  animals  after  inoculation  with  typhoid  bacilli,  says :  "The 
sickness  begins  when  the  animal  body  becomes  sensitized  and  begins 
to  split  up  the  bacilli."  By  "splitting  up"  he  means  here,  as  in  his 

7  Wolff-Eisner.     Centralbl.  f.  Bakt.,  Vol.  37,  1904. 

8  Wolff- Eisner.     "Handbuch    der    Serum    Therapie,"    p.    24,    Lehmanns, 
Miinchen,  1910. 

9  Vaughan.    A  m.  Jour,  of  Med.  Sci.,  Sept.,  1908. 


BACTERIAL  ANAPHYLAXIS  415 

other  work  10  on  protein  split  products,  not  a  mere  liberation  of  pre- 
formed poisons,  but  a  chemical  (enzymotic)  proteolysis  by  which  a 
poisonous  group  of  the  bacterial  protein-molecule  is  set  free. 

The  essential  difference  of  this  point  of  view  from  the  endotoxin 
theory  at  first  sight  seems  a  trivial  one — in  the  one  case  liberation  of 
a  preformed  poison  molecule — in  the  other  liberation  of  a  poison  by 
the  breaking  up  of  a  molecule.  The  difference,  however,  is  a  funda- 
mental one.  For,  in  the  earlier  theory,  the  specific  element  of  the 
toxemia  was  in  the  nature  of  the  different  poisons — whereas  in  the 
view  of  Vaughan  the  lysin  which  breaks  up  the  protein  molecule  is 
alone  the  specific  element,  the  formed  poisons  being  concerned  as 
non-specific  and  alike,  whether  produced  from  colon  bacilli,  tubercle 
bacilli,  or  egg  white. 

Friedberger,1 1  finally,  in  1910,  repeating  with  bacteria  his  ex- 
periments upon  "anaphylatoxin"  liberation  from  specific  precipitates, 
succeeded  in  obtaining  such  poisons  in  the  test  tube  by  allowing 
fresh  guinea  pig  complement  to  act  upon  sensitized  bacteria. 

These  results  were  confirmed  by  extensive  experiments  carried 
out  soon  after  this  by  Friedberger 12  himself  with  a  number  of 
collaborators. 

The  results  of  these  investigations  may  be  summarized  as  follows : 

1.  The  action  of  alexin  upon  sensitized  or  unsensitized  bacteria 
yields  toxic   substances  which,   injected  into   normal  guinea   pigs, 
produce  the  characteristic  symptoms  of  anaphylaxis,  with  frequent 
death  and  typical  autopsy  findings. 

2.  These  poisons   ("anaphylatoxins")   may  be  produced  from 
any  variety  of  bacteria,  pathogenic  and  non-pathogenic.13     (The  or- 
ganisms used  in  the  earlier  experiments  were  Vibrio  metchnikovi, 
the  bacillus  of  tuberculosis,  the  typhoid,  prodigiosus,  and  subtilis 
bacillus,  and  Aspergillus  fumigatus.) 

3.  The  successful  production  of  the  poisons  depends  intimately 
upon  the  relative  amounts  of  antigen  (bacteria)  and  alexin  used,  and 
upon  the  time   and   temperature   conditions  under   which  the   ex- 
posures are  made. 

10  Vaughan.     Zeitschr.  f.  Immunitatsforsch.,  Vol.  1,  1909. 

11  Friedberger.     Berl.  klin.  Woch.,  Nos.  32  and  42,  1910. 

12  Friedberger;  Friedberger  and  Goldschmid;  Friedberger  and  Szymano.w- 
ski;   Friedberger  and   Schiitze:   Friedberger  and  Nathan.     Zeitschr.  f.  Im- 
munitatsforsch., Vol.  9,  1911. 

13  Neufeld  and  Dold,  comparing  virulent  and  avirulent  strains  of  pneu- 
mococcus  in  this  regard,  have  found  that  the  virulence  of  the  race  has  no 
relation  to  its  yield  of  anaphylatoxin.     Indeed  the  anaphylatoxins  from  va- 
rious bacteria  seem  to  be  qualitatively  entirely  alike. 

NOTE. — We  wish  to  note  here  that  we  are  giving  Friedberger's  theories  and 
experiments  at  some  length  because  they  have  had  considerable  influence  on 
our  conceptions  of  infection.  We  do  not  wish  to  create  the  impression  that  we 
accept  them,  however,  and  will  discuss  their  fallacies  later  in  the  same  chapter. 


416 


INFECTION  AND  RESISTANCE 


4.  The  poisons  can  be  produced  from  boiled  as  well  as  from  na- 
tive bacteria. 

Although  unsuccessful  with  none  of  the  bacteria  with  which  ex- 
periments were  carried  out,  different  species  yielded  the  poison  with 
varying  degrees  of  intensity,  though  qualitatively  the  poisons  were 
similar.  Bacillus  prodigiosus,  though  non-pathogenic,  seems,  in 
general,  to  be  one  of  the  most  favorable  microorganisms  for  such 
experiments. 

Since  a  clear  understanding  of  Friedberger's  basic  experiments  is 
essential  to  the  further  development  of  the  theoretical  conceptions 
which  have  been  based  upon  them,  it  will  be  useful  to  insert  here  a 
protocol  taken  from  his  paper  with  Goldschmid. 

Experiment  VI.  30,  VI,  1910.  Ten  3-day  agar  cultures  of 
typhoid  bacilli  washed  up  in  salt  solution — 5  c.  c.  to  1/2  culture. 
Varying  amounts  of  inactivated  typhoid  immune  serum  are  added, 
the  tubes  brought  to  11  c.  c.,  24  hours  in  refrigerator.  1,  VII — 
Centrifugalized  and  to  sediment  added  4  c.  c.  guinea  pig  complement 
(active  or  inactivated),  24  hours  in  refrigerator.  2,  VII — Centrif- 
ugalized and  supernatant  fluids  injected  into  guinea  pigs  of  200 
grams  intravenously. 


Exp. 
N6. 

Amount  of 
culture 

Specific 
immune 
serum 
sensitiz. 

Amount  of 
complement 

No.  of 
animal 

Symptoms 

Result 

1 

^  slant  agar 

0 

4  c.  c. 

G61 

Severe 

Dead  4  min. 

2 

%  slant  agar 

0 

4c.  c. 

G64 

Slight 

Dead  18  hre. 

3 

l/2  slant  agar 

0 

4  (heated  56°  C.) 

G62 

No  symptoms 

Lives 

4 

J^  slant  agar 

0 

4  (heated) 

G63 

No  symptoms 

Lives 

5 

H  slant  agar 

1.0 

4 

G66 

Severe  anaph. 

Lives 

6 

H  slant  agar 

1.0 

4 

G69 

Very  severe 

Dead  4  min. 

7 

J^  slant  agar 

1.0 

4  (heated) 

G65 

0 

Lives 

8 

H  slant  agar 

1.0 

4  (heated) 

G68 

0 

Lives 

9 

^  slant  agar 

0.1 

4 

G67 

Very  severe 

Dead  8  min. 

10 

H  slant  agar 

0.1 

4 

G70 

Very  severe 

Dead  11  min. 

11 

J^  slant  agar 

0.1 

4  (heated) 

G73 

0 

Lives 

12 

H  slant  agar 

0.01 

4 

G71 

Very  severe 

Dead  2  min. 

13 

^  slant  agar 

0.01 

4  (heated) 

G75 

0 

Lives 

14 

Y<i  slant  agar 

0.001 

4 

G72 

Very  severe 

Dead  5  min. 

15 

3/2  slant  agar 

0.001 

4  (heated) 

G74 

0 

Lives 

16 

1.0 

o 

G76 

0 

Lives 

17 

0  1 

o 

G77 

o 

Lives 

18 

H  culture 

0 

0 

G79 

0 

Lives 

From  Friedberger  and  Goldschmid,  Joe.  «*.,  p.  402. 
sion  of  control  19.) 


(Changes  made  only  in  wording  and  onus- 


This  series  alone  shows  that,  under  the  given  conditions,  4  c.  c. 
of  alexin  will  produce  the  poison  with  1/2  slant  of  typhoid  bacilli, 
without  sensitization  (tubes  1  and  2),  with  sensitization  ranging 
in  degree  from  1  c.  c.  to  0.001  c.  c.  of  the  given  immune  serum 
(tubes  5,  6,  9,  10,  12,  and  14),  and  that  inactivation  of  the  alexin 
'serum  in  all  cases  prevented  the  poison  formation.  Normal  guinea 


BACTERIAL  ANAPHYLAXIS 


417 


pig  serum  alone,  active  or  inactivated,  the  bacteria,  or  the  immune 
serum  alone  were  without  toxicity  in  all  of  numerous  controls.14 

The  experiments  of  Friedberger  and  his  associates  were  rapidly 
confirmed  by  Neufeld  and  Bold,15  Kraus,16  Ritz  and  Sachs,17  and 
many  others,18  and,  though  the  conditions  under  which  the  anaphy- 
latoxin  formation  took  place  were  defined  with  slight  variation  by 
different  workers,  the  essential  features  of  Friedberger' s  claims 
were  upheld. 

As  was  to  be  expected,  it  was  soon  found  that  instead  of  the  pro- 
longed exposures  at  refrigerator  temperature  the  poisons  could  be 
obtained  more  rapidly  by  digestion  for  shorter  periods  in  water  19 
baths  at  37°  C.  And  with  this  method  accurate  studies  on  the  rela- 
tions between  time  of  exposure  and  proportions  of  reagents  (antigen, 
sensitizer,  alexin)  were  made,  relations  the  importance  of  which 
was  apparent  from  Friedberger's  first  studies.  The  outcome  of 
this  work  was  as  follows:  1.  There  are  a  definite  minimum  and  a 
definite  maximum  quantity  of  bacteria  with  which  anaphylatoxin 
can  be  produced  by  a  given  fixed  quantity  of  guinea  pig  serum. 
Thus,  in  one  of  the  experiments  of  Friedberger  and  Goldschmid,  4 
loopsful  of  typhoid  bacilli  with  4  c.  c.  of  complement  produced  a 
fatal  poison,  24  loopsful  with  the  same  amount  produced  none.  (In 
some  of  the  writer's  20  experiments  with  typhoid  bacilli  a  similar 
principle  of  proportions  was  evident,  though  much  larger  quantities 

TITBATION   EXPERIMENT   WITH   TYPHOID-IMMUNE   SERUM 

Rabbit  79 


Dilution  of  serum 

Agglutination 

Bactericidal  titre  with  modified 
Stern-Korte  method 

1:100 

+  +  + 

480  colonies 

1:200 

+  +  + 

556  colonies 

1:500 

+  +  + 

750  colonies 

1:1,000 

+  + 

Over  10,000  colonies 

1:2,000 

=fc 

+  +  +  +  + 

1:5,000 

— 

+  +  +  +  + 

1:10,000 

— 

+  +  +  +  + 

14  Injury  of  the  animals  by  mere  volume  of  injection  can  be  definitely 
excluded.  The  writer  has  frequently  injected  5  to  6  c.  c.  of  salt  solution 
into  guinea  pigs  of  200  to  300  grams  without  symptoms  in  any  way  resem- 
bling anaphylaxis. 

15Neufeld  and  Bold.  Berl.  Wn.  Woch.,  No.  2,  24,  1911;  Arb.  a.  d.  kais. 
Gesundheits  ami.,  Vol.  38,  1911. 

16  Kraus.    Zeitschr.  f.  Immunitatsforsch.,  Vol.  8,  1911. 

17  Ritz  u.  Sachs.     Berl.  klin.  Woch.,  No.  22,  1911. 

18  Izar.     Zeitschr.  f.  Immunitatsforsch.,  Vol.  11,  1911. 

19  Friedberger  u.   Mita.     Zeitschr.   f.   Immunitatsforsch.,   Vol.   10,   1911. 
See  also  Bold,  "Das  Bakterien  Anaphylatoxin,"  Fischer,  Jena,  1912. 

20  Zinsser.     Jour.  Exp.  Med.,  Vol.  17,  1913. 


418 


INFECTION  AND  RESISTANCE 


of  typhoid  bacilli  could  be  successfully  used  if  the  time  of  exposure  at 
37°  C.  was  prolonged.)  2.  If  sensitized  bacteria  are  used  an  excess 
of  sensitization,  beyond  a  definite  limit,  weakens  the  formation  of 
anaphylatoxin.  It  may  be  permitted  to  illustrate  this  with  a  protocol 
of  one  of  the  writer's  experiments  with  typhoid  bacilli,  since,  though 
merely  confirming  the  principle  laid  down  by  Friedberger,  it  in- 
cluded a  careful  titration  of  the  bactericidal  contents  of  the  anti- 
typhoid serum.  (See  table  on  p.  417.) 

Two-tenths  c.  c.  of  this  serum  added  to  1  c.  c.  of  typhoid  filtrate 
gave  a  very  slight  clouding  in  about  15  minutes. 

ANAPHYLATOXIN   EXPERIMENTS 


Amount 

Number 
in 

Typhoid 
bacilli 

of 
inactive 
serum 

Amount 
of 
complement 

Weight 
of 
animal 

Result 

antityphoid 

1 

2 

|  slant 
|  slant 

5.0c.  c. 
3.5  c.  c. 

4  C.  C. 

4  c.  c. 

215  gm. 
200  gm. 

Very  sick,  recovers 
Typical  death  2  min. 

3 

£  slant 

3.0c.  c. 

4  c.  c. 

198  gm. 

Typical  death  2  min. 

4 

£  slant 

2.0  c.  c. 

4  c.  c. 

225  gm. 

Typical  death  2  min. 

5 

|  slant 

l.Oc.  c. 

4  c.  c. 

200  gm. 

Sick,  recovers. 

The  weak  character  of  the  antiserum  used,  and  the  fact  that, 
in  this  experiment,  the  digestion  was  at  0°  to  5°  C.,  explain  the 
failure  to  obtain  a  strong  anaphylatoxin  with  1  c.  c.  of  sensitizing 
serum. 

The  negative  experiment  resulting  from  a  too  vigorous  sensi- 
tization is  practically  a  corollary  of  the  next  point  ascertained  by 
Friedberger,  namely,  that : 

3.  With  constant  amounts  of  reagents  a  too  prolonged  exposure 
at  37°  C.  will  result  in  failure  to  obtain  the  poison. 

How  are  we  to  explain  these  experimental  results  ?  The  first  of 
the  three — namely,  the  fact  that  an  excess  of  bacteria  inhibits  the 
formation  of  anaphylatoxins — seems  to  the  writer  most  easily  ex- 
plained by  accepting  the  views  of  Bordet  on  the  manner  of  the 
union  of  an  antigen  with  its  antibody.  For,  unlike  the  opinion  of 
Ehrlich,  who  assumes  a  union  of  the  two  according  to  the  laws  of 
multiple  proportions,  Bordet 21  believes  that  the  distribution  of 
serum  substances  upon  an  antigen  is  such  that  the  entire  amount  of 
antibody  is  distributed  equally  among  the  antigenic  elements.  In 
the  case  of  an  excess  of  bacteria,  as  in  these  experiments,'  therefore, 
the  quantity  falling  to  each  unit  is  insufficient,  at  least  in  the  time  of 
exposure  here  practised,  to  accomplish  the  cleavage  necessary  for 
poison  production. 

21  Bordet.     Ann.  de  I'Inst.  Past.,  17,  p.  161,  1903. 


BACTERIAL  ANAPHYLAXIS  419 

As  regards  the  second  and  third  point — the  failure  of  producing 
anaphylatoidns  if,  on  the  one  hand,  too  intense  sensitization  was 
employed — or,  on  the  other,  the  time  of  exposure  was  too  prolonged — 
these  seem  to  indicate  that  anaphylatoxin  is  not  the  end  product  of 
the  complement  action,  but  rather  an  unstable  intermediate  sub- 
stance which,  once  formed,  is.  rapidly  further  decomposed  ("abge- 
baut")  into  non-toxic  derivatives. 

Indeed,  Neufeld  and  Dold,22  in  experiments  with  the  cholera 
spirillum,  found  that  whenever  lysis  was  permitted  to  proceed  as  far 
as  the  actual  disintegration  and  granulation  of  the  bacteria  no  pois- 
onous substances  were  obtained.  They  conclude  from  this  that  rapid 
lysis  actually  prevents  the  production  of  the  poison,  and  that  the 
anaphylactic  antibody  has  no  relation  to  the  bacteriolytic  sensitizer. 
They  fortify  this  opinion  by  experiments  in  which  they  easily  ob- 
tained powerful  poisons  with  pneumococci,  organisms  which  are  but 
slightly,  if  at  all,  subject  to  actual  lysis.  They  suggest  identity  of 
the  anaphylactic  antibody  with  the  opsonins,  or  possibly  with  the 
"Bordetsche  Antikorper"  of  Neufeld.  This  latter  conclusion  does 
not  seem  valid,  since  the  mere  fact  that  one  microorganism  under- 
goes lysis  and  another  does  not  is  not  necessarily  an  argument  for  a 
difference  in  the  sensitizers  produced  in  animals  by  immunization 
with  these  bacteria.  It  may,  and  probably  does,  depend  upon  varia- 
tions in  the  ease  of  disintegration  of  the  different  cell-bodies,  and, 
as  a  matter  of  fact,  not  many  bacteria  undergo  actual  complete  lysis 
as  easily  as  does  the  cholera  spirillum.  Moreover,  there  is  much 
evidence  in  favor  of  the  so-called  "unitarian"  point  of  view,  which 
holds  that  no  fundamental  structural  and  functional  differences 
between  the  various  heat-stable  antibodies — sensitizers  (ambocep- 
tors),  precipitins,  immune  opsonins  (bacteriotropins),  and  the  so- 
called  "Bordet"  alexin-fixing  antibodies — have  as  yet  been  proved. 

However  this  may  be,  it  seems  conclusively  established  that  a  too 
vigorous  and  prolonged  action  does  not  yield  poisons — and  that,  since 
less  vigorous  sensitization  or  early  interruption  of  the  exposure  will 
lead  to  positive  results,  the  mechanism  is  one  of  rapid  poison  forma- 
tion with  equally  rapid  further  decomposition  into  a  non-toxic  sub- 
stance. In  some  cases  chis  is  jiore  rapid  than  in  others.  In  Neufeld 
and  Dold's  experiments  with  cholera  spirilla  the  exposure  of  2  loops- 
ful  of  the  organisms  sensitized  with  0.02  antiserum  and  treated  with 
2  c.  c.  of  alexin  resulted  in  complete  lysis  and  failure  of  demonstrable 
anaphylatoxin  in  2  hours  at  37°  C.  In  some  of  the  writer's  experi- 
ments with  typhoid  bacilli  the  most  regular  positive  results  were  ob- 
tained when  the  exposures  at  37°  C.  were  prolonged  to  several  hours 
and  powerful  poisons  were  determined  even  after  as  long  as  15  hours 
at  37°  C. 

That  the  production  of  the  poison  can  under  no  circumstances  be 

22Neufeld  and  Dold.    Loc.  cit. 


420  INFECTION  AND  RESISTANCE 

regarded  merely  as  a  giving  up  from  the  bacterial  cell  of  preformed 
endotoxins  under  the  influence  of  lytic  substances  which  produce 
greater  permeability  of  the  cell  membrane  was  shown  by  Neufeld 
and  Dold,  who  extracted  bacteria  with  lecithin  salt  solution  and  pure 
salt  solution,  and  from  these  extracts  (but  moderately  toxic  in  them- 
selves) produced  typical  anaphylatoxins  by  the  action  of  complement. 
The  matrix  of  the  poison  thus  is  shown  by  direct  experiment  to  be  a 
soluble  ingredient  of  the  bacterial  cell. 

It  was  further  shown  by  Friedberger  and  Nathan  that  the  con- 
ditions prevailing  in  the  test  tube  experiment  in  truth  represent  the 
processes  taking  place  within  the  animal  body.  This  they  accom- 
plished by  injected  bacterial  emulsions  into  the  peritoneal  cavities  of 
guinea  pigs,  killing  the  animals  after  several  hours  and  examining 
the  peritoneal  exudates  for  their  toxic  properties.  Centrifugalized, 
cleared  of  bacteria,  and  injected  intravenously  into  other  guinea 
pigs,  these  exudates  produced  the  typical  acute  symptoms  character- 
istic of  the  poisons  obtained  in  test-tube  experiments. 

It  was  on  these  premises,  then,  that  Friedberger  23  was  led  to 
formulate  his  views  of  the  nature  of  bacterial  infections,  which  give 
promise  of  introducing  a  new  understanding  of  these  diseases.  It 
has  been  shown  in  the  researches  upon  serum  anaphylaxis  that  the 
injection  of  small  quantities  of  a  foreign  protein  may  produce  reac- 
tions of  temperature  which  simulate  very  closely  those  prevailing 
in  infectious  diseases,  and  variations  in  the  quantities  injected,  the 
path  of  administration,  and  the  interval  between  injections  may  lead 
to  conditions,  local  and  systemic,  which  may  affect,  more  or  less 
profoundly,  many  different  organs  and  tissues  of  the  body.  These 
matters  we  have  considered  in  the  general  discussion  of  anaphylactic 
phenomena.  Friedberger  now  suggests  that  we  may  regard  bacterial 
infection,  after  all,  as  the  presence  in  the  body  of  a  living  foreign 
protein — in  this  case  varying  in  distribution  and  quantity  by  reason 
of  the  particular  invasive  properties  of  the  given  germ  and  the 
balance  between  these  and  the  resistance  of  the  host.  It  is  not  neces- 
sary, therefore,  to  assume  that  the  character  of  the  disease  is  deter- 
mined by  the  existence  of  different  preformed  "endotoxins."  He 
believes  that  we  may  justly  assume  that  the  toxic  substances  appear 
only  after  proteid  cleavage  of  the  bacterial  bodies  has  been  initiated 
Iby  the  action  upon  them  of  the  serum  components,  and  that  the  ap- 
parent specificity  of  the  poisons,  or  differences  between  the  toxemic 
.manifestations  of  various  diseases,  may  depend,  not  on  differences 
in  the  pharmacological  actions  of  these  poisons,  but  rather  upon 
variations  in  the  invasive  properties  of  the  bacteria,  both  as  concerns 
their  quantitative  distribution  and  their  accumulation  and  localiza- 
tion in  the  infected  body. 

23  Friedberger.     Loc.  cit.;  also  Deutsche  med.  Woch.,  No.  11,  1911 ;  Berl. 
klin.  Woch.,  No.  42,  1911. 


BACTERIAL  ANAPHYLAXIS 

If  we  leave  out  of  consideration  bacteria  which,  like  the  diph- 
theria bacillus,  produce  true  secretory  poisons,  it  would  be  the  ability 
to  gain  a  foothold  in  the  body,  the  degree  of  invasive  power,  the  pre- 
dilection in  the  choice  of  a  path  of  entrance,  and  the  specific  local 
accumulation  upon  which  the  speed  and  quantity  of  anaphylatoxin 
production  and  absorption  would  depend,  and  which  consequently 
would  give  character  to  variations  in  the  clinical  pictures  of  different 
diseases.  Besides  simplifying  considerably  our  comprehension  of 
bacterial  toxemia  this  point  of  view  again  brings  out  the  great  im- 
portance of  the  work  of  Vaughan,  and  of  Vaughan  and  Wheeler, 
on  the  non-specific  poisonous  fraction  obtained  by  hydrolysis  of  bac- 
terial and  other  proteins. 

To  support  this  assumption  Friedherger  points  out  the  similarity 
in  the  clinical  manifestations  of  several  diseases  in  which  the  inciting 
bacteria  are  biologically  very  different,  but  in  which  the  distribution 
and  invasive  properties  are  alike.  For  instance,  lobar  pneumonia 
caused  by  the  pneumococcus  is  clinically  very  similar  to  that  caused 
by  the  Friedlander  bacillus,  though  the  microorganisms  inciting 
them  are  extremely  unlike  each  other.  He  draws  a  similar  parallel 
between  true  cholera  and  cholera  nostras,  and  we  may  add  another 
striking  example  in  the  great  similarity  existing  clinically  between 
the  various  forms  of  acute  and  subacute  septicemia  in  which  a  defi- 
nite bacteriological  diagnosis  can  rarely  be  made  except  by  blood 
culture. 

Conversely,  the  same  microorganism  may  call  forth  diseases 
which  clinically,  apart  from  the  purely  local  manifestations,  are  very 
dissimilar,  according  to  the  localization  and  distribution  of  the  bacteria. 

Granted  that  we  accept  this  view,  then  the  subsidence  of  the 
disease  might  depend  merely  upon  limitation  of  the  supply  of  an- 
tigen, as  the  increasing  bactericidal  action  of  the  blood  constituents 
comes  into  play,  and  upon  the  consequent  diminution  of  the  anaphy- 
latoxin. For,  as  the  bacteria  diminish  and  the  sensitizer  increases,  a 
changed  proportion  between  them  is  established  which,  finally,  as 
experiment  has  shown,  results  in  a  failure  of  anaphylatoxin  produc- 
tion. For,  although  experiments  have  shown  that,  within  a  wide 
latitude  of  relative  proportions  of  bacteria  and  antibody,  anaphyla- 
toxin can  be  formed,  beyond  this  range  an  excess  of  one  or  the  other 
element  eventually  will  prevent  their  formation. 

Infectious  disease,  then,  according  to  this  point  of  view,  repre- 
sents merely  the  reaction  of  the  body  against  a  foreign  protein,  the 
bacteria.  These  gain  a  foothold  in  the  body,  and  at  first,  during  the 
so-called  incubation  time,  cause  no  symptoms,  since  the  slight  amount 
of  bacterial  destruction  with  correspondingly  slight  cleavage  of  the 
bacterial  protoplasm  liberates  too  small  an  amount  of  anaphylatoxin 
to  incite  noticeable  deviations  from  the  normal  condition.  As  these 
slight  quantities  of  bacterial  cleavage  products  are  absorbed,  however, 


INFECTION  AND  RESISTANCE 

a  reactionary  formation  of  specific  antibody  occurs.  Meanwhile, 
also,  the  foreign  protein  increases  and  is  distributed  by  bacterial 
growth.  In  consequence  of  these  parallel  processes  changes  of  pro- 
portion between  the  reacting  substances  are  created  and  a  constantly 
greater  amount  of  anaphylatoxin  is  liberated  and  the  disease  pro- 
gresses. This  may  kill  the  patient  if  the  proportions  become  such 
that  the  amount  of  poison  formed  exceeds  the  lethal  dose.  At  any 
rate,  the  symptoms  may  vary  and  fluctuate  according  to  the  relations 
maintained  between  the  reacting  bodies,  modified  somewhat  by  the 
supply  of  alexin  or  complement.  If  recovery  is  to  take  place  the 
amount  of  antibody  (sensitizer,  amboceptor)  may  become  so  great 
that  the  bacteria  are  subjected  to  rapid  destruction,  the  chemical 
cleavage  of  their  bodies  taking  place  so  vigorously  that  practically  no 
anaphylatoxin  is  distributed  and  vigorous  phagocytosis  is  initiated. 
Finally  the  antigen  is  completely  removed.  On  the  other  hand,  an 
excessive  increase  of  the  bacteria  or  a  defective  supply  of  alexin 
might  also  lead  to  a  final  cessation  of  the  formation  of  anaphylatoxin ; 
in  this  case,  however,  we  would  expect  death  by  the  metabolic  dis- 
turbance occasioned  by  the  life  processes  of  the  great  masses  of  bac- 
teria. It  is  not  unthinkable,  moreover,  that  the  bacterial  enzymes  in 
such  a  case  might  produce  substances  comparable  to  the  anaphyla- 
toxins  from  the  destroyed  tissues  of  the  host. 

It  is  perfectly  true,  as  Friedberger  says,  that  on  the  basis  of  this 
theory,  rendered  so  likely  by  experimental  fact,  the  assumption  of 
the  existence  of  endotoxins  to  explain  the  various  manifestations  of 
infectious  disease  is  not  necessary.  The  poisons,  according  to  the 
view  just  outlined,  are  alike  and  non-specific.  It  is  the  reaction 
bodies,  the  sensitizers,  induced  by  the  bacterial  protein  which  in  each 
case  are  these  specific  elements. 

While  it  is  not  necessary  to  assume  specific  endotoxins,  however, 
it  is  not  possible  on  present  evidence  to  entirely  exclude  the  partici- 
pation of  such  substances  in  the  genesis  of  infectious  disease.  The 
rapid  toxic  action  of  bacterial  extracts  obtained  in  various  ways 
has  been  taken  to  argue  in  favor  of  this. 

It  is  a  difficult  question  to  settle,  and  must  undoubtedly  remain 
an  open  one  until  a  method  is  found  by  which  crucial  experiments 
can  be  formulated.  Since  Neufeld  and  Dold  have  succeeded  in  pro- 
ducing anaphylatoxin  from  bacterial  extracts,  the  primary  toxic 
action  of  every  bacterial  extract,  however  rapidly  produced  from  the 
bacteria,  can  be  regarded  as  possibly  furnishing  merely  an  antigen 
for  anaphylatoxin  production,  and  indeed  such  a  supposition  is  ren- 
dered more  likely  by  the  almost  invariable  incubation  time  following 
upon  the  administration  of  endotoxic  extracts,  even  when  they  are 
introduced  directly  into  the  circulation.  Pfeiifer 24  himself  still 

24  R.  Pfeiffer.    "tiber  Bakterien  Endotoxine,  etc.,"  Weiehhardt's  Jdhres- 
bericht,  Vol.  6,  p.  29,  1910. 


BACTERIAL  ANAPHYLAXIS 

believes  in  specific  endotoxins,  basing  his  opinion  on  the  individually 
characteristic  nature  of  the  infections  caused  by  supposedly  endo- 
toxic  bacteria.  The  differences  in  the  degrees  of  toxicity,  moreover, 
of  extracts  obtained  by  the  same  technique  from  different  micro- 
organisms would  certainly  tend  to  add  some  weight  to  his  argument. 
We  need  only  to  recall  to  memory  the  greater  toxicity  of  bouillon 
culture  extracts  of  B.  dysenterice  Shiga-Kruse  as  compared  with 
similar  extracts  of  B.  dysenterice  Flexner  or  Hiss-Russell,  or  the 
similar  difference  between  typhoid  and  colon  extracts.  Altogether 
the  problem  is  an  involved  one,  for  the  recent  claims  of  Kraus,25 
Doerr,26  27  and  others  of  having  discovered  true  (antitoxin-forming) 
soluble  toxins 28  in  such  cultures  as  those  of  cholera,  dysentery 
Shiga,  and  typhoid  bacilli  add  another  complication. 

As  a  matter  of  fact,  the  entire  problem  of  endotoxins  is  one  which 
calls  for  reexamination  in  that  the  knowledge  gained  of  recent  years 
has  opened  a  number  of  alternative  explanations  for  the  primary 
toxicity  of  such  bacteria  as  the  typhoid  bacillus.  Briefly  summarized 
they  are:  (1)  The  actual  intracellular  existence  of  specific  endo- 
toxins in  the  sense  of  Pfeiffer  29  (toxalbumin).  (2)  The  production 
of  toxic  split  products  in  the  animal  body  from  the  bacterial  protein 
by  proteolytic  cleavage  brought  about  by  non-specific  serum  protease 
(Jobling  and  Petersen)  29  or  by  the  cooperation  of  antibody  and 
alexin  (Friedberger).  (3)  The  absorption  of  antienzymes  by  the 
bacteria  with  consequent  activation  of  the  serum  protease  which  then 
splits  off  toxic  substances  from  the  plasma  protein  (Jobling  and 
Petersen).  (4)  The  presence  of  non-specific  toxic  substances  in  the 
bacterial  cell  body,  of  the  nature  of  peptones,  primary  and  second- 
ary albumoses,  etc.,  which  are  liberated  by  lysis  from  the  bacterial 
cell  after  cell  death.  This  conception  would  differ  from  that  of 
Pfeiffer  in  that  the  intercellular  substances  are  conceived  as  in  no 
sense  specific  toxic  proteins,  but  rather  entirely  non-specific  constit- 
uents representing  the  type  of  poisons  conceived  as  proteolytically 
produced  from  the  antigen  by  Vaughan  and  others.  This  last  view, 
though  hitherto  not  particularly  considered,  should  nevertheless  in 
our  opinion  be  regarded  as  at  least  a  possible  explanation  for  a  part 
of  the  toxic  manifestations  resulting  from  the  injection  of  bacteria 
of  this  class.  Moreover,  such  a  possibility  is  suggested  by  the  fact 
that  bacterial  protein  is  relatively  poor  in  antigenic  properties. 
Doerr,  also,  has  considered  this  in  stating  that  he  believed  the  diffi- 
culty of  producing  anaphylaxis  with  bacteria  was  in  part  due  to 
the  fact  that  their  body  substances  were  relatively  poor  in  coagulable 

26  Kraus.     MonatscJir.  f.  Gesundheitspflege,  No.  11,  1904. 

26  Kraus  u.  Doerr.     Wien.  klin.  Woch.,  No.  42,  1905. 

27  Kraus.     "Kraus  u.  Levaditi  Handbuch,"  Vol.  1,  p.  180. 

28  Exotoxins. 

29  Zinsser  and  Parker.     Journal  of  Exp.  Med.,  September,  1917. 


424  INFECTION  AND  RESISTANCE 

(antigenic)  proteins.  We  have  not  been  able  as  yet  to  study  the 
matter  extensively,  but  we  have  carried  out  a  few  experiments. 

Typhoid  bacilli  from  twenty-four  agar  cultures  were  weighed  as 
a  moist  mass,  ground  up  with  salt,  and  then  taken  up  in  distilled 
water  to  isotonicity.  After  the  addition  of  0.2  c.c.  of  N  sodium 
hydroxid  to  100  c.c.  the  suspension  was  heated  to  60°  C.  for  30 
minutes  to  prevent  autolysis,  and  was  then  shaken  for  4  to  5  hours 
with  a  motor,  at  room  temperature.  After  filtration  through  a 
Berkefeld  candle  the  clear  solution  gave  a  definite  cloud  after  boiling 
and  adding  acetic  acid.  The  filtrate  was  treated  with  heat  and  acid 
to  remove  coagulable  protein.  On  the  advice  of  Professor  Gies  the 
suspension  was  first  brought  to  a  boil  and  then  small  amounts  of 
acid  were  added  to  prevent  possible  hydrolysis  which  might  have  oc- 
curred had  the  acid  been  added  first.  The  filtrate  from  this  was 
then  half  saturated  with  ammonium  sulphate.  Again  a  definite  cloud 
was  obtained,  and  when  this  was  filtered  clear,  a  second  turbidity 
could  be  produced  by  complete  saturation  with  the  sulphate. 

Although  we  have  not  yet  obtained  toxic  reactions  with  these 
substances  after  isolation,  perhaps  because  of  the  difficulty  of  obtain- 
ing them  in  sufficient  amount,  the  presence  of  albumoses,  substances 
which  have  often  been  found  to  possess  primary  toxicity  for  animals, 
suggests  the  possibility  that  their  existence  in  the  bacterial  body 
might  indeed  contribute  to  the  injury  done  by  injections  of  bacteria. 
We  found  albumoses  in  extracts  of  typhoid  bacilli  not  only  when 
the  bacteria  were  grown  on  the  ordinary  peptone  media  but  also  on 
agar  made  without  peptone  to  which  nutrose  or  sodium  casemate  has 
been  added  as  an  enriching  substance. 

Moreover,  we  have  also  had  the  experience  with  the  typhoid 
bacillus  which  is  referred  to  repeatedly  by  Vaughan  when  dealing 
with  the  colon  bacillus;  namely,  that  bacterial  suspensions  sub- 
jected to  boiling  are  quite  as  toxic  and  often  more  so  than  are  the 
unheated  and  living  suspensions. 

Thus,  if  three  guinea  pigs  are  injected  with  equal  amounts  of 
typhoid  suspensions,  one  with  the  living  bacteria,  the  second  with 
bacteria  heated  to  60°  C.  for  15  minutes,  and  the  third  with  bacteria 
boiled  for  5  minutes,  the  guinea  pig  receiving  the  boiled  bacilli  will 
often  be  the  first  one  to  grow  sick  and  die  several  hours  before  the 
others. 

This  may  mean,  of  course,  as  Vaughan  suggests,  that  the  heated 
protein  is  more  promptly  split  by  the  ferments  of  the  body. .  It 
also  suggests,  however,  that  in  addition  to  this  the  heating  has  left 
unchanged  non-coagulable  toxic  constituents  of  the  cell. 

It  still  remains  for  us  to  consider  certain  experimental  facts 
which  have  had  some  influence  upon  extending  and  altering  the  con- 
ceptions of  anaphylatoxin  formation  which  we  have  just  outlined. 


BACTERIAL  ANAPHYLAXIS  425 

In  the  earlier  work  of  Friedberger,  Neufeld  and  Dold,  and  others  the 
poisons  were  formed  from  the  bacteria  by  the  action  of  alexin  at 
low  temperatures.  This  suggested  the  possibility  that  the  alexin 
fractions — "Endstiick"  and  "Mittelstiick" — might  not  both  be  in- 
volved in  the  reaction,  since,  from  previous  studies,  it  was  known  that 
at  low  temperatures  the  midpiece  (the  globulin  fraction)  was  bound, 
but  that  the  end  piece  did  not  become  active  until  the  temperature 
was  increased.  This  point  was,  therefore,  made  the  object  of  a 
special  investigation  by  Friedberger  and  Ito,30  who  found  that 
neither  fraction  alone  would  suffice,  but  that  bacterial  anaphylatoxins 
were  formed  only  under  the  influence  of  the  intact  whole  alexin,  or 
by  that  of  the  two  fractions,  reunited  after  separation. 

Because  of  the  reasoning  along  which  the  investigations  of  ana- 
phylatoxin  formation  were  developed,  it  is  not  surprising  that  it 
seemed  self-evident  that  the  matrix  of  the  poison  was  represented  by 
the  bacterial  protein — the  antigen  of  the  lytic  complex.  The  only 
fact  which,  in  the  earlier  experiments,  might  have  cast  some  doubt 
upon  this  was  the  ease  with  which  anaphylatoxins  were  produced 
from  boiled  bacteria  and  precipitates  and  from  such  very  insoluble 
organisms  as  the  tubercle  bacillus. 

Such  vague  suspicion  becomes  a  very  definite  doubt,  however, 
in  the  light  of  the  experiments  of  Keysser  and  Wassermann.31 
Keysser  and  Wassermann  utilized  the  fact  that  certain  serum  ele- 
ments may  be  absorbed  out  of  serum  if  this  is  shaken  up  with  such 
indifferent  suspensions  as  barium  sulphate  or  kaolin  (aluminium 
orthosilicate).32  They  therefore  substituted  these  insoluble  sub- 
stances for  antigen,  allowed  them  to  absorb  serum  constituents,  as- 
sumed by  them  to  be  amboceptor,  out  of  normal  and  inactivated  im- 
mune sera,  and  then  allowed  complement  or  alexin  to  act  upon  the 
"sensitized"  kaolin. 

In  this  way  they  obtained  active  and  powerful  anaphylatoxin, 
and  claim,  in  consequence,  that  the  matrix  of  the  poison  is  not  in 
the  bacterial  antigen,  but  in  the  sensitizer  or  amboceptor,  which  is 
mechanically  absorbed  by  the  bacteria  (as  by  the  kaolin),  and  thus 
made  amenable  to  the  alexin  action. 

The  experiments  of  Keysser  and  Wassermann  have  found  con- 
firmation in  the  hands  of  other  investigators,  although  the  results  of 
Neufeld  and  Dold,  as  well  as  our  own,  with  this  method  were  far 
more  irregular  than  were  those  of  Keysser  and  Wassermann.  !N~eu- 

30  Friedberger  and  Ito.     Zeitschr.  f.  Immunitatsforsch.,  Vol.  11,  1911. 

31  Keysser  and  Wassermann.     Folia  Serologica,  Vol.  7,  1911;  Zeitschr. 
f.  Hyg.,  Vol.  68,  1911. 

32  Kaolin  emulsions  will  absorb  amboceptor  only  out  of   diluted  serum. 
Out  of  concentrated  serum  complement  is  completely  absorbed.     Friedberger 
u.    Salecker,   Zeitschr.   f.   Immunitatsforsch.,   Vol.   11,   1911;   Zinsser,   from 
Journ.  Exp.  Mecl.,  Vol.  18,  1913. 


426  INFECTION  AND  RESISTANCE 

feld  and  Dold 83  and  Friedberger,34  suggest  that  the  horse  serum 
absorbed  by  the  kaolin  may  act  as  an  antigen  itself,  and  is  acted  upon 
by  normal  sensitizer  present  in  the  guinea  pig  serum.  This  is  in 
keeping  with  the  well-known  fact  that  small  amounts  of  sensitizers  to 
many  varieties  of  foreign  proteins  are  present  in  normal  serum,  and 
is  further  borne  out  by  the  fact  that  Neufeld  and  Dold,  unlike  Keys- 
ser  and  Wassermann,  were  never  able  to  produce  anaphylatoxin  by 
allowing  the  alexin  alone  to  act  upon  kaolin — without  previous 
absorption  of  horse  serum. 

We  say  "never,"  though  the  protocols  of  Neufeld  and  Dold35 
show  a  single  successful  experiment.  This  they  explain,  however, 
by  assuming  the  accidental  presence  of  some  antigen  in  the  alexic 
serum.  That  is,  the  entire  complex,  antigen,  sensitize^  and  alexin, 
is  assumed  to  have  been  present  in  this  particular  guinea  pig  serum. 
The  same  explanation  may  be  applied  to  the  occasional  inherent 
toxicity  which  develops  in  normal  guinea  pig  sera  on  standing. 
Whether  the  above  complicated  explanation  is  necessary  or  whether 
we  may  assume  an  autolytic  process  in  the  guinea  pig  serum  by  which 
anaphylatoxin-like  substances  are  formed  is  an  open  question. 

At  any  rate,  it  has  been  shown  that,  even  with  bacteria,  the 
action  of  alexin  is  not  the  only  way  in  which  acute  poisons  may  be 
obtained  from  them.  And,  indeed,  if  we  look  upon  the  action  of 
alexin  as  analogous  to  that  of  an  enzyme — an  assumption  for  which 
we  have  much  supporting  evidence,  we  may  well  expect  that  other 
methods  of  proteolysis  will  give  similar  toxic  cleavage  products. 
And  various  methods  of  bacterial  autolysis  have  actually  yielded 
such  results.  Thus  Neufeld  and  Dold  obtained  poisons  by  digesting 
typhoid  bacilli,  cholera  spirilla,  and  other  microorganisms  for  sev- 
eral hours  in  salt  solution,  lecithin  salt  solution,  and  inactivated 
guinea  pig  sera.  Their  extracts  killed  guinea  pigs  within  several 
hours.  Rosenow  36  has  even  succeeded  in  obtaining  acutely  toxic  sub- 
stances which  caused  typical  anaphylactic  death  in  guinea  pigs  by 
suspending  pneumococci,  typhoid  bacilli,  and  other  bacteria  in  salt 
solution  at  37°  C.  for  varying  periods,  and  the  writer,37  though  never 
producing  acute  death,  was  able  to  cause  typical  anaphylactic  shock 
in  isolated  cases  with  similar  salt  solution  extracts  of  typhoid  bacilli. 
It  is  not  impossible  that  poisons  obtained  in  this  way  are  formed 
by  autolysis  due  to  proteolytic  enzymes  of  the  bacterial  cell. 

In  cases  in  which  bacteria,  suspended  in  salt  solution  and  other 
indifferent  fluids,  represent  the  only  source  of  protein  present  it 
must,  of  course,  be  assumed  that  they  are  the  substratum  or  matrix  of 

83  Neufeld  and  Dold.     Loc.  cit. 

84  Friedberger  and  Salecker.    Zeitschr.  f.  Imm.,  Vol.  11,  1911. 

85  Dold.    Loc.  cit. 

88  Rosenow.    Jour.  Inf.  Dis.,  Vol.  9,  1911;  Vol.  10,  1912. 
87  Zinsser.    Loc.  cit. 


BACTERIAL  ANAPHYLAXIS  427 

the  anaphylactic  poison.  They  are  also,  of  course,  to  be  regarded  as 
the  source  of  the  poison  in  such  experiments  as  those  of  Vaughan,  in 
which  the  poison  was  produced  by  chemical  hydrolysis  of  the  bac- 
terial bodies.  In  the  case  of  anaphylatoxin  production  by  fresh 
serum  in  the  presence  of  bacteria,  kaolin,  precipitates,  etc.,  the  ques- 
tion is  much  more  complex. 

As  we  have  stated  before,  it  is  only  natural,  considering  our  pre- 
vious knowledge  of  bacteriolysis  in  serum,  that  the  first  conclusion 
arrived  at  should  look  for  the  source  of  the  poisons  in  the  bacterial 
cells.  The  doubt  which  has  been  cast  upon  this  assumption  by  the 
work  of  Keysser  and  Wassermann  and  others,  however,  rests  upon  a 
sufficiently  sound  experimental  basis  to  prevent  our  absolute  accept- 
ance of  this  view.  Jobling  and  Petersen  38  have  recently  carried  out 
experiments  which  may  serve  to  throw  much  light  upon  anaphyla- 
toxin. They  believe  that,  by  the  ordinary  technique  of  anaphylatoxin 
production  with  bacteria  and  serum,  most  of  the  toxic  substances  orig- 
inate from  the  serum  proteins.  The  bacteria  act  merely  by  remov- 
ing the  antiferments  from  the  serum,  thereby  setting  free  the  fer- 
ments normally  present  in  the  serum,  and  permitting  them  to  act 
upon  the  serum  proteins.  The  result  is  cleavage  and  the  production 
of  toxic  split  products.  This  would  explain  such  results  as  those  of 
Keysser  and  Wassermann.  Jobling  and  Petersen  have  supported 
their  contention  by  experiments  in  which  they  have  obtained  typical 
anaphylatoxins  by  removing  serum  antiferments  with  chloroform, 
kaolin,  and  agar.  They  have  further  shown  that  emulsions  of  bac- 
teria actually  do  remove  antiferments  from  fresh  serum,  and  that 
the  bacteria  used  in  the  process  become  more  resistant  to  tryptic 
digestion  in  consequence. 

This  does  not  necessarily  weaken  the  force  of  Friedberger's  view 
of  infectious  disease.  For,  whatever  the  source  of  the  toxic  sub- 
stances, the  result  is  still  the  same.  Wherever  proteolysis  takes 
place,  and  certain  quantitative  relations  between  cleavage,  energy, 
and  substratum  exist,  it  seems  toxic  bodies  may  be  liberated. 

And  the  result  of  such  proteolysis,  at  some  stage  of  the  process, 
yields  apparently  the  same  non-specific  toxic  substance,  whatever  the 
particular  nature  of  the  proteolysis  and  whatever  the  variety  of  the 
original  protein  matrix. 

38  Jobling  and  Petersen.    Jour,  of  Exp.  Med.,  June,  1914. 


CHAPTER   XVIII 

THE  CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS 

SERUM  SICKNESS 

WE  have  mentioned  that  Rosenau  and  Anderson  attacked  the 
problem  of  hypersusceptibility  primarily  in  the  hope  of  casting  light 
upon  the  nature  and  cause  of  the  distressing  symptoms  which  in 
human  beings  often  ensue  upon  the  injection  of  diphtheria  antitoxin. 
It  has  been  one  of  the  staple  objections  of  lay  opponents  to  the  use 
of  antitoxins  that  the  injections  are  apt  to  cause  severe  illness  and 
occasionally  death,  and  indeed  a  few  cases  are  on  record  in  which 
sudden  death  has  followed  the  first  injection  of  diphtheria  antitoxin. 
Since  it  was  known  by  accumulated  clinical  experience  as  well  as 
by  experiments  like  those  of  Bertin,1  of  Johannesen,2  and  others 
that  the  harmful  effects  were  not  dependent  upon  the  antitoxin  con- 
tents, but  could  be  produced  by  injections  of  normal  horse  serum,  it 
was  but  natural  to  bring  these  ill  effects  into  analogy  with  the  phe- 
nomena of  hypersusceptibility.  A  large  number  of  references  to 
such  antitoxin  illness  or  SERUM  SICKNESS  have  appeared  in  the  lit- 
erature since  the  first  beginnings  of  the  therapeutic  use  of  sera,  yet 
no  careful  analysis  of  the  condition  was  made  until  von  Pirquet  and 
Schick,3  in  1905,  published  their  studies. 

As  a  rule  the  results  of  serum  injection  have  been  mild  and  with- 
out danger,  though  sufficiently  frequent  and  troublesome  to  call  for 
thorough  study  and  attempts  to  discover  the  prophylactic  measures. 
As  stated  above,  a  few  cases  are  on  record  in  which  sudden  death 
has  followed  a  single  first  injection.  There  are  no  reports  in  the 
literature  known  to  us,  however,  of  fatalities  after  second  injections, 
although  not  infrequently  such  cases  have  taken  on  alarmingly  seri- 
ous aspects. 

The  percentage  of  incidence  and  the  variety  of  symptoms  have 
been  the  subjects  of  many  reports.  The  most  frequent  and  striking 
single  occurrence  has  been  an  urticarial  rash.  Rolleston,4  in  a  large 

1  Bertin.     Gaz.  Med.  de  Nantes,  1895.    Quoted  from  Levaditi. 

2  Johannesen.     Deutsche  med.Woch.,  No.  51,  1895. 

3  Von   Pirquet  u.   Schick.     "Die   Serum  Krankheit,"  Deuticke,  Leipzig, 
1905.    Also  Munch,  med.  Woch.,  53,  p.  67,  1906. 

4  Rolleston.     The  Practitioner,  Vol.  74,  1905. 

428 


CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS       429 

series  of  cases,  found  urticaria  in  all  but  17  of  289  cases  of  serum 
eruptions  occurring  between  the  first  and  tenth  days  after  injection, 
and  in  all  but  ten  of  ninety-four  later  eruptions. 

Rashes  occurred  in  from  69.4  to  81.9  per  cent,  of  the  600  anti- 
toxin cases  which  Rolleston  reports. 

Joint  pains  commonly  accompany  the  appearance  of  the  rash,  and 
frequently  there  is  adenitis,  involving  the  glands  adjacent  to  the  point 
of  injection,  and  even  remotely  in  the  submaxillary,  axillary,  or  in- 
guinal glands.  Albuminuria  is  quite  common,  and  with  it  oliguria 
and  relative  concentration.  Fever  is  rarely  absent,  though  usually 
slight,  together  with  general  malaise.  Rolleston  in  his  purely  clini- 
cal study  does  not  classify  his  cases  into  those  reacting  after  a  first 
injection  and  those  showing  symptoms  after  repeated  treatment.  He 
states,  however,  that  the  serum-reaction  may  be  extremely  severe  in 
cases  of  "relapse  or  second  attack  of  diphtheria"  in  which  urticaria 
with  "pronounced  edema  surrounding  the  wheals,  vomiting,  rigors, 
and  collapse  may  ensue  within  a  few  hours  of  injection,"  and  further 
asserts  that  these  severe  symptoms  are  more  apt  to  follow  upon  large 
than  upon  small  doses. 

Von  Pirquet  and  Schick  have  studied  the  condition  with  careful 
reference  to  a  comparison  between  the  symptoms  occurring  in  sub- 
jects after  a  first  injection  of  serum  and  those  following  upon  re- 
peated treatments.  Their  studies  revealed  the  very  important  fact 
that  the  ill  effects  following  a  second  injection  were  not  only  more 
severe  than  those  occurring  after  the  first  injection,  but  developed 
after  much  shorter  periods  of  incubation.  In  the  ordinary  "first 
injection"  case  the  symptoms  appear  usually  in  from  one  to  twelve 
days.  After  a  second  injection  this  incubation  period  may  be  con- 
siderably shortened  and  symptoms  may  appear  in  from  five  to  seven 
days,  the  local  and  general  reactions  being  much  more  marked  than 
those  subsequent  to  a  first  injection.  Indeed,  in  some  of  the  cases  re- 
ported they  may  attain  very  alarming  degrees  of  severity.  This  is 
the  so-called  accelerated  ("beschleunigte")  reaction  of  von  Pirquet 
and  Schick,  and  is  different  from  the  "first  injection"  symptoms  only 
in  its  greater  severity  and  speedier  onset.  In  addition  to  this,  how- 
ever, the  "second  injection"  cases  may  show  a  train  of  immediate 
symptoms5  (sofortige  Reaktion),  which  occur  within  twenty-four 
hours  after  injection,  and  are  characterized  by  marked  local  erythema 
and  edema  with  often  urticaria  and  constitutional  disturbance.  Both 
reactions  may  occur  in  the  same  individual,  the  "accelerated"  reac- 
tion setting  in  as  the  "immediate"  reaction  subsides. 

Again,  one  reaction  or  the  other  may  occur  alone.  The  analogy 
between  the  immediate  reaction  and  the  anaphylaxis  of  animal  ex- 
periment is  obvious.  The  cases  may  be  classified  on  the  basis  of 

5  Rankin  in  the  Lancet,  Dec.,  1911,  reports  a  case  of  "immediate"  reac- 
tion 15  minutes  after  injection. 


430  INFECTION    AND    RESISTANCE 

these  reactions,  according  to  von  Pirquet  and  Schick,  the  nature  of 
the  reaction  being,  within  certain  limits,  determined  by  the  interval 
ensuing  between  the  first  and  the  second  injection.  Thus,  when  the 
interval  was  twenty-one  days  or  less  the  immediate  reaction  alone 
was  noticed.  When  the  interval  was  between  two  and  six  months 
both  the  immediate  and  accelerated  reactions  were  present,  and  when 
the  interval  was  still  longer  (seven  months  or  more)  the  accelerated 
reaction  alone  was  present.  Isolated  exceptions  to  this  are  noted  in 
the  series  of  sixty-one  cases  so  reported. 

Currie,6  7  who  has  made  similar  studies,  confirms  the  results  of 
von  Pirquet  and  Schick  in  all  essentials,  and  agrees  with  their  state- 
ment that  the  nature  of  the  reaction  is  chiefly  dependent  upon  the 
interval  between  injections. 

That  the  entire  train  of  symptoms,  as  well  as  the  mere  fact  of 
their  dependence  upon  an  injection  of  a  foreign  protein,  rather  than 
upon  the  antitoxin  itself,  force  upon  us  the  analogy  with  anaphy- 
laxis  is  clear.  Moreover,  this  analogy  becomes  almost  an  identity 
when  we  can  show,  as  von  Pirquet  and  Schick  have  done,  that  the 
first  injection  has  apparently  sensitized  the  subject,  in  that  the 
second  administrations  are  fraught  with  more  violent  and  serious 
reactions,  dependent  to  a  great  extent,  as  in  experimental  anaphy- 
laxis,  upon  the  time  intervening.  If  serum-sickness  is  truly  an 
anaphylactic  phenomenon,  however,  it  is  still  by  no  means  clear  why 
symptoms  should  at  all  ensue  after  the  first  injection.  Many  ex- 
planations have  been  offered  for  this;  none  of  them,  however,  from 
the  very  nature  of  the  problem  itself,  can  be  finally  accepted  as 
proved.  Two  possible  explanations  appear  from  the  experimental 
work  of  Rosenau  and  Anderson  quoted  above.  These  workers,  we 
have  seen,  showed  among  other  things  that  the  state  of  hypersus- 
ceptibility  could  be  transmitted  from  mother  to  offspring,  and  that 
sensitization  by  way  of  the  intestinal  canal  was  at  least  possible. 
Both  of  these  factors  may  have  determinative  significance  in  the 
present  case.8  There  may  be,  because  of  such  conditions,  a  pre- 
existent  sensitization  which,  especially  in  cases  of  accidental  injec- 
tion of  the  antitoxin  directly  into  a  small  vein  (an  accident  prob- 
ably not  infrequent  in  deep  muscular  injections),  may  possibly  ex- 
plain the  few  instances  of  sudden  death  following  the  first  antitoxin 
injection  and  the  isolated  instances  of  "immediate"  reaction  follow- 
ing "first"  injections.  Rosenau  has  also  suggested  recently  that  sensi- 
tization may  be  unconsciously  acquired  against  various  forms  of 
protein  by  absorption  through  the  lungs  of  the  organic  matter  car- 
ried in  the  expired  breath  of  animals.  In  this  way  possibly  hyper- 

6  Currie.     Jour,  of  Hyg.,  Vol.  7,  1907. 

7  See  also  Goodall,  Jour,  of  Hyg.,  7,  1907. 

8  Regarding  intestinal  sensitization  see  also  Richet,  C.  R.  de  la  Soc.  cU 
Biol.,  Vol.  70,  1911 ;  Lesne  et  Dreyfus,  C.  E.  de  la  Soc.  de  Biol,  Vol.  70,  1911 


CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS       431 

susceptibility  against  horse  protein  may  be  acquired  and  subsequently 
be  expressed  by  a  reaction  to  the  first  injection  of  antitoxin.9  While 
this  must  be  considered  a  possibility,  however,  not  all  investigators 
are  ready  to  accept  it  and  its  significance  is  at  present  very  uncertain. 

In  the  ordinary  case  of  serum-sickness  after  first  injection,  how- 
ever, the  long  incubation  time  elapsing  between  the  injection  of  the 
serum  and  the  onset  of  symptoms,  often  more  than  10  days,  would, 
it  seems  to  us,  tend  to  argue  against  a  previous  hypersusceptible 
state  of  the  patient.  On  the  other  hand,  we  have  learned  since  10 
the  earlier  studies  of  Eisenberg,  von  Dungern,  and  others  that  for- 
eign proteins  injected  into  rabjbits  may  be  excreted  very  slowly,  and 
that  even  after  the  formation  of  antibodies  (precipitins)  the  antigen 
may  still  be  demonstrable  in  the  blood  serum  of  the  rabbit.  Thus 
at  periods  eight  to  twelve  days  after  the  injection  of  comparatively 
large  amounts  there  is  often  present  in  the  same  individual  both  an- 
tigen and  its  specific  antibody  side  by  side,  and  the  essential  condi- 
tions for  the  production  of  anaphylatoxin  are  thus  established.  That 
the  two  bodies  do  not,  as  a  rule,  unite  in  such  serum  in  quantities 
sufficient  to  be  demonstrated  by  alexin  fixation  has  been  discussed 
in  another  place,  but  this  by  no  means  excludes  a  gradual,  slow  union 
of  small  amounts  of  antigen  with  antibody,  consequent  fixation  of 
alexin,  and  the  liberation  of  anaphylatoxic  products.  In  fact,  al- 
though there  does  not  seem  at  present  to  be  any  way  to  bring  experi- 
mental proof  to  support  it,  it  seems  very  likely  that  a  slow  splitting 
of  the  antigen  begins  by  virtue  of  the  normal  antibody,  and  as,  in 
the  course  of  eight  to  ten  days,  the  antibody  appears  in  relatively 
larger  amounts,  the  toxic  products  of  the  reaction  are  sufficient  to 
give  rise  to  symptoms.  Such  a  point  of  view  is  supported  only  by  the 
experimental  knowledge  that  antibody  may  appear  in  considerable 
concentration  before  the  antigen  has  disappeared  from  the  circula- 
tion, and  upon  the  facts  we  know  concerning  the  toxic  substances 
which  arise  from  the  union  of  two  such  reagents  subjected  to  the  in- 
fluence of  alexin. 

In  fact,  it  seems  likely  that  this  process  of  antibody  formation 
may  represent  merely  an  emergency  mechanism  for  the  purpose  of 
ridding  the  body  of  foreign  dissolved  proteins  which  have  penetrated 
into  the  circulation,  cannot  diffuse  unchanged  through  the  healthy 
excretory  channels,  and  must  remain  in  the  blood  stream  until  sub- 
jected to  proteolysis  by  the  enzymes  of  the  blood.  In  the  course  of 
ordinary  life  the  quantities  of  such  substances  gaining  entrance  into 
the  circulation  are  necessarily  small,  and  would  call  forth  but  slight 
reactions.  The  sudden  injection  of  large  amounts  of  serum,  not 

9  Weichhardt,  Arch.  f.  Hyg.,Vo\.  74,  1911,  has  made  similar  studies  and 
claims  to  have  found  toxic  protein  cleavage  products  similar  to  his  kenotoxin 
in  exposed  air. 

10  See  Zinsser  and  Young,  Jour.  Exp.  Med.,  Vol.  17,  1913. 


432  INFECTION    AND    RESISTANCE 

easily  disposed  of,  would,  on  the  basis  of  the  preceding  assumptions, 
result  in  a  very  gradual  antigen  destruction  with  consequent  antibody 
formation,  so  that,  at  the  end  of  eight  to  ten  days,  there  would  be 
present  side  by  side  remnants  of  unchanged  antigen  and  newly 
formed  specific  antibody.  The  destroyed  antigen  fraction,  in  other 
words,  gradually  sensitizes  the  body  to  the  fraction  which  persists  and 
has  not  yet  been  assimilated  or  excreted  at  the  end  of  this  time.  Such 
a  point  of  view  would  explain,  not  only  the  reaction  after  a  first  in- 
jection, but  would  account  for  the  incubation  time  in  such  cases,  and 
for  the  differences  between  these  reactions  and  both  the  "immediate" 
and  the  "accelerated"  reactions  of  cases  twice  injected.  The  bearing 
which  this  point  of  view  would  have  on  the  problems  of  incubation 
time  in  general  is  obvious. 

The  recognition  of  the  anaphylactic  nature  of  serum  sickness 
has  led  to  many  attempts  to  develop  methods  of  antitoxin  adminis- 
tration by  which  these  reactions  could  be  avoided.  Since  it  was  de- 
termined that  the  degree  of  reaction  was  directly  dependent  upon 
the  amount  of  the  foreign  serum  injected,  it  was  an  obviously  logical 
procedure  to  attempt  in  antitoxin  production  to  concentrate  as  high 
a  potency  of  antitoxin  into  as  small  as  possible  an  amount  of  serum. 
Attempts  have  also  been  made  to  alter  the  serum  itself  in  such  a  way 
that  it  would  lose  its  properties  of  acting  as  an  anaphylactic  antigen 
without  suffering  materially  in  antitoxin  value.  Bujwid11  found 
that  serum  sickness  was  less  frequent  after  the  use  of  sera  which 
had  been  allowed  to  stand  for  prolonged  periods,  and  we  have  seen 
that  Besredka  and  others  have  claimed  a  reduction  of  toxic  property 
in  sera  heated  repeatedly  to  60°  C.  It  was  hoped,  moreover,  that 
the  so-called  concentration  methods — such  as  those  of  Gibson,  Banz- 
haf,  and  others — would  yield  an  antitoxin  that  would  be  devoid  of 
anaphylactic  properties.  None  of  these  methods  of  altering  the 
serum  can,  however,  be  said  to  have  been  satisfactory  in  that  the 
antitoxic  property  seems  to  be  closely  associated  with  the  globulins,12 
which  we  have  seen  are  at  the  same  time  closely  associated  with  the 
production  of  anaphylaxis. 

The  conclusions  of  Rosenau  and  Anderson  13  regarding  this  are 
based  on  direct  experimentation  with  concentrated  antitoxin  made 
at  the  New  York  Department  of  Health  by  the  Gibson  method. 
They  found  the  refined  antitoxin,  volume  for  volume,  quite  as  toxic  as 
the  unrefined,  but  since  the  same  amount  of  antitoxin  is  by  this  and 
other  methods  concentrated  in  a  considerably  smaller  amount  of 

11  Bujwid.       Quoted  from  Friedberger  and  Mita,  Deutsche  med.  Woch., 
No.  5,  1912. 

12  Among  others  previously  mentioned  see  also  Turro  and  Gonzales,  C.  R. 
de  la  Soc.  de  Biol.,  Vol.  69,  1910. 

13  Rosenau  and  Anderson.     U.  S.  Pub.  Health  and  M.  H.  S.  Hyg.  Lab. 
Bull.  36,  April,  1907. 


CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS       433 

protein  solution  there  is  a  distinct  gain  for  safety  in  the  use  of 
such  preparations.  Endeavors  to  produce  potent  antitoxic  sera  by 
chemical  or  physical  methods  without  any  sensitizing  properties  have 
thus  been  unsuccessful. 

On  the  other  hand,  the  knowledge  gained  by  animal  experimen- 
tation regarding  the  influence  upon  the  anaphylactic  manifestations 
exerted  by  various  methods  of  administering  the  antigen  has  led  to 
results  which  have  proven  of  much  value,  both  in  the  immuniza- 
tion of  experimental  animals  and  in  human  serum  therapy.  Prob- 
ably the  most  carefully  studied  of  these  methods  is  the  one  which 
Besredka  14  has  recommended  on  the  basis  of  his  work  on  antianaphy- 
laxis  in  animals.  He  found  that  sensitized  guinea  pigs  could  be 
injected  with  quantities  of  serum  amounting  to  about  one  half  or 
less  of  the  fatal  dose  without  showing  symptoms,  and  subsequently, 
at  intervals  of  2  to  5  minutes,  further  injections  of  the  serum  could 
be  given,  the  total  amount  five  to  twenty  times  exceeding  the  lethal 
dose  without  causing  symptoms  of  any  kind.  From  these  experi- 
ments he  has  developed  a  method  of  serum  injections  the  principle 
of  which  is  very  simply  a  division  of  dose.  In  lieu  of  injecting  into 
an  animal  or  man  the  entire  quantity  of  serum  at  once,  small,  gradu- 
ally increasing  amounts  are  administered  in  two,  three,  or  more 
doses,  the  intervals  varying  from  five  minutes  to  several  hours,  ac- 
cording to  the  necessities  of  speed  indicated  by  clinical  considera- 
tions. The  process  as  applied  to  man  consists,  then,  in  preceding 
the  injection  of  the  larger  quantity  of  the  serum  by  one  or  two  sub- 
cutaneous injections  of  smaller  amounts.  With  this  principle  well 
defined  it  would  be  quite  unwise  to  lay  down  definite  rules  of  quan- 
tity or  interval  at  present,  since  in  no  instance  will  it  be  possible 
to  estimate  the  exact  condition  of  susceptibility  of  the  particular 
case.  It  goes  without  saying  that  the  precautions  should  be  par- 
ticularly respected  in  children  in  whom  the  relation  of  5  or  10  c.  c. 
of  serum  volumes  to  the  body  weight  approaches  the  dangerous  pro- 
portions dealt  with  in  animal  experiments. 

Besredka  has  also  shown  that  if  the  rectum  of  a  sensitive  animal 
is  cleaned  out  by  enema,  and  a  relatively  large  amount  of  the  an- 
tigen then  introduced,  an  injection  may  be  given  in  within  twelve  to 
twenty-four  hours  later  without  danger,  however  delicate  the  hyper- 
susceptibility  of  the  animal  has  been.  This  method  apparently 
must  depend  upon  a  slow,  gradual  absorption  of  antigen,  and  would 
seem  to  furnish  a  most  convenient  and  advisable  method  to  apply  in 
man. 

14  Besredka.  Ann.  tie  I'Inst.  Past.,  Vol.  24,  1910;  C.  E.  de  la  Soc.  de 
Biol,  65,  1908,  p.  478;  C.  E.  de  la  Soc.  de  Biol.,  Vol.  66,  1909,  p.  125;  ibid., 
67,  1909,  p.  266;  C.  E.  de  I'Acad.  des  Sc.,  Vol.  150,  1910,  p.  1456;  ref.  Bull, 
de  I'Inst.  Past.}  Vol.  8,  1910,  p.  735. 


434  INFECTION    AND    RESISTANCE 

Friedberger  and  Mita  15  have  suggested  another  method  which 
also  depends  upon  very  slow  administration  rather  than  division  of 
dose.  In  experiments  upon  guinea  pigs  they  had  found  that  sensi- 
tized animals  which,  as  tested  by  controls,  would  succumb  to  intrave- 
nous injections  of  0.01  c.  c.  of  sheep  serum  per  100  grams  weight 
when  the  entire  quantity  was  injected  within  one  minute,  would  sur- 
vive a  similar  administration  of. as  much  as  0.1  c.  c.  if,  by  means  of  a 
specially  constructed  apparatus,  the  injection  was  made  gradually, 
extending  over  a  period  of  100  minutes.  While  this  method  offers 
many  practical  difficulties  to  ordinary  bedside  application,  it  does 
show  that  the  intervals  of  injections  by  the  Besredka  method  do 
not  need  to  exceed  fractions  of  an  hour — or,  at  most,  a  few  hours — 
in  order  to  add  materially  to  the  safety  of  injection. 

There  is  another  phase  of  specific  therapy  in  which  the  question 
of  possible  anaphylaxis  must  be  taken  into  consideration,  and  that 
is  the  treatment  of  patients  with  bacterial  vaccines.  As  a  matter  of 
fact  the  danger  of  anaphylaxis  in  such  cases  is  probably  very  remote 
— both  because  of  the  shortness  of  the  intervals  at  which  these  injec- 
tions are  usually  made  and  because  of  the  extremely  small  amounts 
of  protein  represented  by  the  usual  dose  of  100  or  200  millions  of 
bacteria.  However,  the  possibility  cannot  be  disregarded,  especially 
in  children,  and  two  cases  were  verbally  described  to  the  writer  by 
Dr.  Philip  Van  Ingen,  in  which  gonococcus  vaccines  caused  immedi- 
ate symptoms  of  such  a  character  that  anaphylaxis  could  not  be  ex- 
cluded. 

Ohlmacher  16  also  has  described  localized  reactions  at  the  place 
of  inoculation  as  well  as  swelling  and  tenderness  at  points  of  former 
inoculations  following  bacterial  vaccine  injections.  He  has  oc- 
casionally seen  slight  systemic  symptoms  (dizziness,  nausea,  etc.) 
which  he  explains  on  the  basis  of  anaphylaxis. 

Moreover,  it  must  be  remembered  that  active  sensitization  with 
bacterial  antigens  has  been  most  regularly  successful  in  the  hands 
of  Kraus  and  Doerr,17  Holobut,18  and  Kraus  and  Admiradzibi,19  as 
well  as  in  confirmatory  experiments  carried  out  in  the  Stanford 
laboratory,  when  repeated  injections  at  short  intervals  were  made, 
rather  than  when,  as  in  serum  anaphylaxis,  a  single  injection  only 
was  given.  This  would  lend  an  even  closer  analogy  to  the  procedures 
carried  out  during  vaccine  treatment.  For  instance,  in  the  successful 
experiments  of  the  last-named  writers  ten  daily  injections  of  1/100 
of  a  slant  culture  of  dead  colon  bacilli  were  made  for  the  purpose  of 

15  Friedberger  and  Mita.     Deutsche   med.   Woch.,  No.   5,   1912;   figures 
taken  from  Versuch.,  3. 

16  Ohlmacher.    Jour.  Med.  Res.,  Vol.  19,  1908,  p.  113. 

17  Kraus  u.  Doerr.     Wien.  klin.  Woch.,  No.  28,  1908. 

18  Holobut.     Zeitschr.  f.  Immunitatsforsch.,  Vol.  3,  1909. 

19  Kraus  u.  Admiradzibi.    Zeitschr.  f.  Immunitatsforsch.,  Vol.  4,  1910. 


CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS       435 

sensitization — the  toxic  dose  of  1/2  slant  being  given  fifteen  days 
after  the  last  of  these. 

In  order  to  obtain  some  opinion  regarding  the  possible  dangers 
of  vaccine  therapy  in  this  regard  the  writer  a  few  years  ago  ob- 
served carefully  a  pair  of  young  goats,  animals  extremely  favorable 
for  anaphylactic  experiment  (and  of  approximately  the  weight  of  a 
child  of  three)  in  the  course  of  frequent  and  irregularly  spaced 
intravenous  injections  of  typhoid  bacilli.  In  both  cases  marked 
anaphylactic  symptoms  were  observed  after  the  animals  had  attained 
a  considerable  agglutinative  and  bactericidal  power  (1  to  5,000  to 
1  to  20,000),  but  in  each  case  only  after  intravenous  injections  of 
large  quantities  of  bacteria,  1/2  to  2  slant  cultures.  While,  of 
course,  such  experiments  are  not  conclusive  in  any  way,  from  these, 
as  well  as  from  a  number  of  laboratory  accidents  in  the  course  of 
animal  immunization,  it  is  the  writer's  impression  that  the  intrave- 
nous injection  of  bacteria  or  bacterial  products  in  human  beings 
would  be  a  procedure  involving  some  risk,  unless  more  thorough  ex- 
perimental data  than  we  at  present  possess  were  available  to  guide 
us  as  to  dosage  and  intervals.  The  ordinary  subcutaneous  treatment 
of  patients,  however,  with  bacteria  in  the  amounts  customarily  em- 
ployed in  vaccines  would  seem  to  be  practically  without  risk  as  far 
as  acute  anaphylaxis  is  concerned. 

In  the  treatment  of  animals  with  vaccines  of  various  kinds  Le- 
clainche2021  has  repeatedly  called  attention  to  the  fact  that  inocu- 
lation with  a  vaccine  may  lead  to  a  condition  of  hypersusceptibility, 
serving  to  light  up  a  latent  lesion  which  'might  have  been  held  in 
check  if  the  normal  resistance  had  not  been  interfered  with.  This 
objection,  we  have  seen,  has  been  made  on  numerous  occasions  against 
tuberculin  therapy,  and  is  one  of  the  factors  which  have  led  to  the 
great  caution  in  dosage  and  control  of  all  therapy  based  on  active 
immunization.  These  considerations,  even  more  than  the  rather 
remote  dangers  of  serious  active  anaphylaxis,  require  that  all  forms 
of  specific  therapy  should  be  carried  out  only  under  the  safeguards 
of  thorough  familiarity  with  the  experimental  phases  of  such  work. 

Our  own  recent  studies  on  anaphylatoxins,  moreover,  have  in- 
clined us  to  believe  that  hypersusceptibility  to  bacterial  protein  may 
well  be  a  strong  predisposing  factor  in  infection. 

Serum  sickness,  occurring  as  a  direct  consequence  of  the  injection 
of  a  foreign  protein  into  a  human  being,  forces  itself  upon  us  as 
manifestly  related  to  anaphylaxis.  There  are  a  number  of  other 
clinical  conditions  which  are  less  obviously  anaphylactic  in  nature, 
but  in  which  we  have  many  good  reasons  for  attributing  an  important 
part  of  the  etiology  to  a  state  of  hypersusceptibility.  Thus  the  pe- 

20  Leclainche  and  Vallee.     Ann.  de  I'Inst.  Past.,  1902. 

21  Leclainehe.     Eevue  Gen.  Med.  Vet.,  Sept.,  1911;  Bull,  de  I'Inst.  Past., 
9,  1911,  p.  1089. 


436  INFECTION    AND    RESISTANCE 

culiar  so-called  "idiosyncrasies"  observed  in  many  people  who  suffer 
from  urticarial  skin  rashes,  gastro-intestinal  difficulties,  and  even 
severe  systemic  illnesses  after  certain  varieties  of  food  seem  to  de- 
pend upon  an  acquired  or  possibly  inherited  hypersusceptibility  to 
the  particular  proteins  involved,  which,  at  certain  times  of  abnormal 
gastro-enteric  conditions,  can  get  into  the  circulation  in  small  quan- 
tity. It  is  not  impossible,  furthermore,  that  such  unfortunate  cases 
as  the  severe  forms  of  angioneurotic  edema,  which  seem,  at  least  in 
part,  to  be  associated  with  gastro-intestinal  disturbance,  and  which 
may  be  transmitted  from  mother  to  child,  have  their  root  in  anaphy- 
laxis. For  this,  however,  we  have  only  inference  based  on  clinical 
observation. 

ASTHMA  AND  HAY  FEVER 

Conditions  in  which  there  seems  to  be  more  definite  ground  for 
association  with  anaphylaxis  are  ASTHMA  and  HAY  FEVER.  In 
asthma  the  analogy  has  been  clearly  set  forth  by  Meltzer.22  He 
points  out  that  in  both  asthma  and  anaphylaxis  the  symptoms  consist 
in  a  tonic  stenosis  of  the  small  bronchioli  of  peripheral  origin,  and 
that  both  conditions  are  favorably  affected  by  the  administration  of 
atropin.  It  is,  of  course,  not  certain,  but  it  seems  extremely  likely 
that  so-called  "nervous  asthma"  is  nothing  else  than  an  anaphylactic 
attack  in  a  hypersusceptible  individual  when  the  particular  protein 
to  which  he  is  sensitive  gains  access  either  by  the  alimentary  or 
respiratory  path. 

Very  closely  related  to  asthma  is  the  condition  known  as  "hay 
fever."  This  disease  has  been  of  recent  years  most  thoroughly  stud- 
ied by  Dunbar.23  Dunbar  has  ascertained  that  the  hay  fever  preva- 
lent in  Europe  is  dependent  chiefly  upon  a  protein  substance  found 
in  the  pollen  of  most  grasses,  while  that  of  America,  which  occurs 
chiefly  in  the  autumn,  is  caused  by  the  proteins  of  the  pollen  cells  of 
the  ambrosiacese  and  solidaginese — plants  which  are  generally  dis- 
tributed on  the  North  American  continent  and  bloom  in  August  and 
September.  The  disease  occurring  in  China  is  caused  by  another 
plant,  the  Ligustrum  vulgare.  The  suggestion  that  the  disease  was 
due  to  anaphylactic  action  of  these  pollen  proteins  upon  hypersus- 
ceptible individuals  was  first  made  by  Wolff-Eisner.24  Dunbar  has 
gone  into  the  question  with  great  thoroughness,  and  has  come  to  the 
conclusion  that  the  disease  has  much  in  common  with  anaphylaxis — 
though  he  believes  that,  in  addition  to  a  hypersusceptibility  to  the 
pollen  "toxin,"  there  must  be  present  in  the  patients,  at  the  same 

22  Meltzer.    Jour,  of  the  A.  M.  A.,  Vol.  55,  1910,  p.  1021. 

23  Dunbar.    Berl.  klin.  Woch.,  Nos.  26,  28,  30,  1905;  Zeitschr.  f.  Immuni- 
tatsforsch.,  Vol.  7,  1907;  Deutsche  r.ied.  Woch.,  Vol.  37,  1911,  p.  578. 

24  Wolff- Eisner.     "Das  Heufieber  sein  .wesen  u.  seine  Behandlung,"  1906. 


CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS       437 

time,  an  abnormal  "Durchlassigkeit"  or  penetrability  of  the  cutis 
and  mucosa  for  the  pollen  substances.  He  claims  to  have  shown  that 
a  solution  of  pollen  protein  instilled  into  the  eye — or  even  dropped 
upon  the  skin  of  a  hay-fever  patient — gives  rise  to  a  prompt  .and 
severe  reaction,  while  it  produces  no  effect  upon  normal  persons. 
Unlike  experimental  serum  anaphylaxis,  the  repeated  instillation  of 
the  pollen  substances  rather  increases  than  diminishes  the  suscepti- 
bility even  when  these  are  carried  out  daily.  Furthermore,  unlike 
serum  anaphylaxis,  against  the  manifestations  of  which  no  direct 
passive  immunization  has  so  far  been  possible,  Dunbar  claims  to 
have  produced  a  curative  immune  serum  by  the  treatment  of  horses 
with  the  pollen  extracts  ("Pollantin").  Dunbar,  therefore,  while 
admitting  an  anaphylaxis-like  hypersusceptibility  of  the  patients, 
still  believes  that  the  antigen  in  this  case  is  a  true  "toxin"  against 
which  an  antitoxin  can  be  produced — the  condition  being  more 
directly  comparable  to  the  sensitization  against  diphtheria  and 
tetanus  toxins  observed  during  the  earlier  phases  of  these  investiga- 
tions by  v.  Behring  and  his  associates  rather  than  to  the  phenomena 
of  serum  anaphylaxis  themselves. 

Schittenhelm  and  Weichhardt,25  on  the  other  hand,  regard  hay 
fever  as  truly  anaphylactic  in  every  sense.  They  speak  of  it  as 
epithelial  anaphylaxis  (hay  fever  being  specifically  designated  as 
"conjunctivitis  and  rhinitis  anaphylactica,"  in  distinction  from 
other  forms  of  cellular  anaphylaxis,  i.  e.,  enteritis  anaphylactica). 
They  believe  that  the  manifestations  of  the  disease  result  from  a 
local  hypersusceptibility  in  which  a  toxic  substance  (Abbau  Produkt 
— similar  to  anaphylatoxin)  is  produced.  The  so-called  "antitoxin" 
of  Dunbar  acts  favorably  only  when  locally  applied,  and  not  on  sub- 
cutaneous administration.  For  this  reason  they  do  not  regard  it  as 
a  true  antitoxin,  but  think  it  acts  as  a  local  antiferment  which  pre- 
vents or  delays  the  cleavage  of  the  pollen-substance  into  its  toxic 
split-product — thereby  preventing  or  ameliorating  the  attacks. 

Similar  to  hay  fever  are  the  sudden  attacks  of  catarrhal  naso- 
pharyngitis  and  conjunctivitis — often  of  asthma-like  respiratory  dif- 
ficulty, with  itching  of  the  nose  and  eyes  and  sneezing  which  many 
individuals  experience  when  coming  close  to  horses,  cats,  or  other 
animals.  In  the  Stanford  University  laboratory  the  writer  had  an. 
assistant  who  invariably  had  such  attacks,  sudden,  violent,  and  of 
several  hours'  duration,  when  handling  guinea  pigs  for  experiment. 
The  character  of  such  attacks  has  long  aroused  the  suspicion  that 
the  reaction  was  anaphylactic  in  nature,  especially  since  it  was  known 
that  extremely  slight  amounts  of  antigen  could  give  rise  to  symptoms 
in  susceptible  subjects.  The  difficulty  in  these  cases  was  the  ques- 
tion of  the  nature  of  the  antigen  which  emanated  from  the  animal 
to  excite  an  attack.  Recently,  however,  observations  having  impor- 
25  Schittenhelm  and  Weichhardt.  Deutsche  med.  Woch.,  37,  No.  19,  1911. 


438  INFECTION    AND    RESISTANCE 

tant  bearing  upon  this  problem  have  been  made  by  Weichhardt  26 
and  by  Rosenau,27  who  have  demonstrated  the  presence  of  organic 
matter  in  expired  breath.  Rosenau  condensed  the  moisture  of  the 
expired  breath  of  man  and  injected  the  liquid  so  obtained  into 
guinea  pigs.  After  two  weeks  these  animals  were  injected  with  nor- 
mal human  serum,  and  out  of  99  test  animals  26  responded  with 
symptoms  of  anaphylaxis.  This  demonstrated  not  only  the  presence 
of  organic  matter  in  the  breath,  but  showed  at  the  same  time  that 
such  organic  matter  was  probably  protein  in  nature  or  at  least  surely 
capable  of  acting  as  anaphylactic  antigen.  Rosenau  surmises,  there- 
fore, that  such  protein  may  be  slightly  volatile  under  the  given  con- 
ditions. He  suggests  that  sensitization  in  this  manner  may  explain 
the  harmful  effects  resulting  from  a  first  injection  of  horse  serum 
into  patients,  previous  sensitization  having  occurred  by  close  associa- 
tion with  horses.  Surely  it  would  explain  logically  the  "cellular" 
or  epithelial  anaphylaxis  experienced  by  certain  people  in  the  pres- 
ence of  animals.  In  our  opinion  this  is  rendered  more  likely,  since, 
in  the  case  mentioned  as  occurring  at  Stanford  University,  the  in- 
halation of  washings  (both  aqueous  and  alcohol  soluble)  obtained 
from  the  hair  and  skin  of  guinea  pigs,  and  dried  in  Petri  dishes, 
produced  absolutely  no  effects  in  the  susceptible  individual,  whereas 
continued  handling  of  a  living  pig  almost  invariably  caused  such 
marked  effects  that  the  person  in  question  often  became  useless  as 
an  assistant  because  of  violent  attacks  of  sneezing.  It  must  not  be 
omitted,  however,  that  not  all  observers  have  confirmed  Rosenau's 
work,  and  his  explanation  must  therefore  be  regarded  as  merely  tenta- 
tive. 

An  interesting  train  of  suggestions  connecting  human  pathology 
with  anaphylaxis  has  followed  the  discovery  of  "organ-specificity" 
in  the  case  of  hypersusceptibility  similar  to  that  noted  by  Uhlenhuth 
in  connection  with  the  precipitin  formation  and  described  in  another 
.chapter. 

It  was  shown  by  Kraus,  Doerr,  and  Sohma,28  we  have  seen,  that 
animals  sensitized  with  the  crystalline  lens  protein  of  one  animal 
species  would  react  to  lens  protein  in  general,  and  not  necessarily 
to  the  tissue  protein  of  the  animal  species  from  which  it  was  taken. 
In  other  words,  the  ordinary  "species"  specificity  did  not  hold  good. 
Specificity  was  determined  in  this  case  by  the  character  of  the  organ 
rather  than  by  that  of  the  species.  The  same  thing  was  shown  for 
testicular  protein  by  v.  Dungern  and  Hirschfeld.29  The  proteins  of 
these  organs  from  various  animals  have  therefore  a  certain  common 
antigenic  property  which  is  independent  of  the  antigenic  element 

26  Weichhardt.     Arch.  f.  Hyg.,  Vol.  74,  1911. 

27  Rosenau  and  Amoss.     Jour.  Med.  Res.,  Vol.  25,  Sept.,  1911. 

28  Kraus,  Doerr,  and  Sohma.     Wien.  klin.  Woch.,  Vol.  21,  1908,  p.  1084. 
2*  Von  Dungern  u.  Hirschfeld.     Zeitschr.  f.  Immunitatsforsch.,  4,  1910. 


CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS       439 

common  to  the  particular  species.  Further  than  this,  Andre  jew  3d 
claims  to  have  shown  that  it  is  possible  to  sensitize  an  animal  with 
its  own  lens  protein.  A  few  guinea  pigs  injected  by  him  with  their 
own  lens  proteins,  and  reinjected  with  the  same  substances  after  a 
suitable  interval,  reacted  with  definite  anaphylactic  symptoms.  The 
possibility  is  thus  given  than  an  animal  or  human  being  could  become 
sensitized  by  its  own  organ  proteins  if  these  were  traumatically  or 
otherwise  destroyed  and  absorbed.  The  train  of  reasoning  is  similar 
to  that  which  has  given  much  hope  of  enlightenment  to  pathologists 
when  the  earlier  work  upon  the  cytotoxins  was  done.  Rosenau  and 
Anderson,31  for  instance,  injected  guinea  pigs  with  guinea  pig  pla- 
centa, and  found  that,  after  the  usual  period  of  incubation,  the  ani- 
mals reacted  to  a  second  injection  with  marked  symptoms  of  anaphy- 
laxis.  On  the  basis  of  these  experiments  Rosenau  and  Anderson 
suggest  that  certain  of  the  toxemias  of  pregnancy  are  of  anaphylactic 
origin.  They  believe  that  it  is  possible  that  a  mother  may  become 
sensitized  by  the  "autolytic  products  of  her  own  placenta,"  the  result 
being  eclampsia. 

By  a  similar  process  of  reasoning  Elschnig32  has  attempted  to 
explain  sympathetic  ophthalmia.  He  claims  to  have  shown  that  the 
laws  of  organ  specificity  apply  to  the  proteins  (especially  the  pig- 
ment) of  the  uveal  tract.  The  destruction  and  absorption  of  injured 
uveal  tissue,  according  to  him,  induce  the  formation  of  organ- 
specific  antibodies  by  which  the  remaining  uveal  structures  of  the 
same,  as  well  as  of  the  opposite,  eye  are  sensitized.  The  consequence 
is  a  "sympathetic"  inflammation  which  "is  to  be  regarded  purely  as 
an  anaphylactic  reaction." 

These  and  other  similar  suggestions  less  well  founded  experi- 
mentally illustrate  the  possibilities  for  clinical  reasoning  furnished 
by  a  knowledge  of  the  anaphylactic  phenomena.  In  no  cases  of  this 
sort,  however,  can  the  association  with  anaphylaxis  be  as  yet  re- 
garded as  more  than  an  extremely  interesting  suggestion. 

From  all  that  has  gone  before  it  is  quite  evident  that  most  of 
the  positive  facts  which  may  be  regarded  as  determined  concerning 
the  phenomena  of  anaphylaxis  have  been  obtained  in  experiments 
with  small  and  very  sensitive  animals,  comparatively  large  and 
measured  quantities  of  antigen,  and  often  by  the  violent  method  of 
intravenous  injection  in  which  the  entire  mass  of  antigen  comes 
rapidly  into  contact  with  the  available  antibodies  and  the  vulnerable 
tissues.  We  cannot,  therefore,  draw  rigid  parallels  between  these 
experiments  and  clinical  manifestations  in  human  beings  in  whom 

30  Andrejew.    Arb.  a.  d.  kais.  Gesundh.,  Vol.  30,  1909. 

31  Rosenau  and  Anderson.     U.  S.  Pub.  Health  and  M.  H.  S.  Hyg.  Lab. 
Bull  45,  1908. 

32  Elschnig.     Von  Graefe's  Archiv   f.  Ophthal,  Vol.  75,  p.  459;  Vol.  76, 
p.  509;  Vol.  78,  p.  549. 


440  INFECTION    AND    RESISTANCE 

the  localization,  quantitative  discharge  of  antigen,  and  consequent 
production  of  antibodies  is  of  necessity  irregular  and  different  in 
each  individual  case.  We  have  learned,  as  a  general  conception, 
however,  that  the  introduction  into  the  animal  body  of  antigenic  sub- 
stances of  all  varieties  leads,  under  certain  conditions,  to  increased 
tolerance  or  resistance — under  other  circumstances  to  a  state  of 
greater  susceptibility — these  diametrically  opposed  physiological 
consequences  being,  in  all  probability,  determined  by  relative  concen- 
trations of  antigen  and  antibody,  their  speed  of  contact,  and  their 
quantitative  relationship  to  available  alexin.  The  problems  of  clini- 
cal medicine  in  which  the  possibility  of  anaphylaxis  can  be  at  all 
considered,  therefore,  are  extremely  complicated,  and  few  of  them 
can  be  approached  by  direct  experiment. 

In  the  case  of  serum  sickness  the  analogy  has  been  so  clear  and 
the  experience  with  human  beings  so  extensive  that  practically  no 
doubt  can  exist  as  to  the  common  mechanism  of  this  condition  with 
that  of  experimental  anaphylaxis.  In  the  other  conditions  men- 
tioned the  connection  is  one  of  great  likelihood,  but  after  all  is  in- 
ferential, and  calls  for  much  further  investigation.  For  this  reason 
it  is  best  to  abstain  from  a  further  enumeration  of  many  other 
maladies  in  which  the  condition  of  hypersusceptibility  has  been  sug- 
gested as  a  vaguely  possible  etiological  factor. 


ANAPHYLAXIS  AND  THE  TUBERCULIN  REACTION 

There  is  one  class  of  phenomena,  however,  which  calls  for 
further  discussion  in  this  connection,  since  its  dependence  upon 
anaphylaxis,  while  generally  assumed,  is  still  opposed  by  many 
authorities.  This  consists  of  the  various  DIAGNOSTIC  REACTIONS  in 
which  extracts  of  micro-organisms  are  injected,  or  brought  into  con- 
tact with  the  skin  or  conjunctiva  of  infected  subjects.  Such  are  the 
various  forms  of  the  tuberculin  reaction,  the  typhoid  reaction  of 
Chantemesse,  the  one  of  Gay,  and  the  luetin  reaction  of  Noguchi.  In 
the  tuberculin  reaction  the  conditions  have  been  thoroughly  studied, 
and  we  may  make  a  detailed  consideration  of  this  example  serve  to 
bring  out  the  general  principles  involved. 

In  all  forms  of  the  TUBERCULIN  REACTION  there  is  a  very  evident 
hypersusceptibility  to  various  forms  of  antigen  derived  from  the 
bacillus.  When  the  tuberculin  is  injected  subcutaneously  the  reac- 
tion is  systemic  and  also  localized  to  a  certain  extent  in  any  tubercu- 
lous foci  which  may  be  present.  When  the  v.  Pirquet  or  Moro  skin 
reactions  are  carried  out,  or  the  Calmette  ophthalmic  test  is  made,  the 
reactions  are  almost  purely  local.  In  all  cases  reactions  are  induced 
by  quantities  of  antigen  which  cause  no  effect  whatever  in  normal 
individuals. 


CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS       441 

The  -basic  observation  leading  to  the  diagnostic  use  of  tuberculin 
was  made  by  Koch  33  upon  guinea  pigs.  He  describes  his  observa- 
tion as  follows: 

" Tuberculin  may  be  injected  into  normal  guinea  pigs  in  consid- 
erable quantities  without  causing  noticeable  symptoms.  Tuberculous 
guinea  pigs,  on  the  other  hand,  react  to  comparatively  small  doses 
in  a  very  characteristic  manner." 

Since,  in  Koch's  experiments  upon  tuberculin,  it  was  desirable 
for  his  particular  purposes  at  the  time,  to  obtain  very  sharp  reac- 
tions, he  did  not  content  himself  with  the  production  of  moderate 
symptoms  by  the  injection  of  slight  amounts  of  tuberculin  into  in- 
fected animals,  but  increased  his  dosage  until  the  guinea  pigs  were 
killed.  He  showed  that  guinea  pigs  having  a  moderately  advanced 
infection — 4  to  5  weeks  after  inoculation — could  be  killed  by  doses 
of  0.2  to  0.3  gram,  while  animals  in  very  advanced  stages  would  suc- 
cumb within  6  to  30  hours  to  quantities  as  small  as  0.1  gram  sub- 
cutaneously.  In  the  animals  so  studied  he  determined  not  only  a 
systemic  effect,  but  a  very  marked  local  reaction  as  well  in  the  skin, 
areolar  tissues,  and  adjacent  lymph  nodes. 

Koch's  observations  upon  guinea  pigs  were  applied  by  him,  Gutt- 
stadt,84  Beck,35  and  others  to  man,  and  the  result  was  the  develop- 
ment of  the  present  important  diagnostic  test.  The  fundamental 
fact  in  this  as  well  as  in  other  tests  of  this  kind,  then,  is  the 
appearance  of  local  and  systemic  reactions  in  infected  subjects  to  con- 
tact with  specific  antigenic  material  which,  at  least  in  the  same 
quantities,  produces  no  effects  in  normal  individuals.  The  analogy 
with  the  phenomena  of  anaphyJaxis  is  thus  indicated. 

Koch's  original  interpretation  of  the  phenomenon  was  of  course 
unaided  by  any  of  the  later  observations  on  anaphylaxis.  According 
to  him  the  tuberculin  contained  substances  which  caused  tissue 
necrosis.  The  necrotizing  action  was  particularly  powerful  upon 
tissues  which  were  tuberculous,  and  therefore  already  saturated  with 
the  toxic  material.  The  destruction  of  such  tissues  resulted  in  sys- 
temic symptoms. 

Very  similar  to  this  view  is  the  one  later  expressed  by  Babes  and 
Broca,36  who  attribute  the  systemic  symptoms  to  a  sudden  lighting 
up  of  the  existing  lesions  by  the  small  amount  of  extra  tuberculin 
added  to  that  already  present  in  these  foci. 

The  first  suggestion  of  the  possible  association  of  the  tuberculin 
reaction  with  the  union  of  an  antigen  and  its  antibody  was  made 
by  Wassermann  and  Bruck.37  They  accepted  Ehrlich's  assumption 

33  Koch.     Deutsche  med.  Woch.,  No.  43,  1891. 

84  Guttstadt.     "Klin.  Jahrbuch"  Erganzung-sband,  1891. 

35  Beck.     Deutsche  med.  Woch.,  No.  9,  1899. 

36  Babes  u.  Broca,     Zeitschr.  f.  Hyg.,  Vol.  23,  1896. 

37  Wassermann  and  Bruck.    Deutsche  med.  Woch.,  No.  12,  1906. 


442  INFECTION    AND    RESISTANCE 

that  certain  cells  of  the  tuberculous  foci  (those  situated  just  below 
the  periphery  and  already  affected  by  the  tubercle  toxin,  though  still 
resistant)  were  possessed  of  an  increased  receptor  apparatus  for  the 
tubercle  antigen.  For  this  reason  the  injected  tuberculin  was  con- 
centrated in  these  foci,  attracted  out  of  the  circulation  by  the  in- 
creased avidity  of  these  cells,  the  consequence  being  increased  ac- 
tivity of  the  lesions  and  systemic  symptoms.  The  tuberculin  reac- 
tion, according  to  these  writers,  therefore,  would  be  caused  by  the 
union  of  the  tuberculin  with  the  "sessile  receptors"  upon  the  diseased 
tissues — a  point  of  view  which  would  specify  the  diseased  tissues  and 
their  products  as  the  sources  from  which  emanated  the  toxic  factors 
inciting  the  systemic  symptoms. 

The  theories  of  Koch  and  of  Babes  do  not,  as  Meyer  points  out, 
explain  the  frequent  absence  of  the  tuberculin  reaction  in  very  ad- 
vanced cases  of  human  tuberculosis,  as  contrasted  with  its  frequency 
and  regularity  in  the  earlier  cases.  For,  according  to  both  of  these 
views,  the  more  severe  the  existing  lesions  the  more  actively  would 
the  injected  tuberculin  initiate  tissue  necrosis  and  consequent  symp- 
toms. The  theory  of  Wassermann  and  Bruck  avoids  this  objection 
since  it  presupposes  the  acceptance  of  Ehrlich's  view  that  the  in- 
creased receptor  apparatus  is  present  and  free  only  in  those  cells  in 
which  necrotic  destruction  has  not  yet  set  in.  In  the  necrotic  areas 
the  receptor  apparatus  is  already  saturated  or  satisfied  as  to  its 
affinities,  and  extensive  areas  of  necrosis,  therefore,  are  unaffected 
by  contact  with  further  quantities  of  tuberculin. 

All  of  these  theories,  however,  inasmuch  as  they  refer  the  tubercu- 
lin reaction  to  alterations  taking  place  in  more  or  less  active  lesions, 
are  unable  to  account  for  the  occurrence  of  the  reaction  in  persons 
in  whom  healed  foci  only  are  present,  and  are  entirely  inconsistent 
with  the  facts  we  now  possess  regarding  the  cutaneous  and  ophthal- 
mic tests  in  which  the  reactions  occur  in  previously  healthy  tissues. 

These  facts  practically  exclude  the  acceptation  of  any  theories 
which  regard  the  tuberculous  focus  as  the  sole  source  of  the  reaction. 
We  may  still  accept  the  Koch  or  Wassermann  views  to  explain  local 
swellings  and  other  changes  in  infected  lymph-nodes  or  other  lesions, 
but  we  must  assume  in  addition  to  this  a  generalized  hypersuscepti- 
bility  at  least  analogous  to  the  phenomena  of  anaphylaxis. 

This  ability  of  previously  healthy  tissues,  remote  from  any  center 
of  tuberculous  infection,  to  react  to  the  application  of  tuberculin 
was  discovered  by  von  Pirquet38  in  the  development  of  his  skin 
reaction,  and  by  Calmette  39  and  Wolff-Eisner  40  in  their  work  upon 

88  v.  Pirquet.  Berl.  Urn.  Woch.,  No.  20,  p.  644,  and  No.  22,  p.  699,  1907; 
also  "Klinische  Studien  iiber  Vaccination,"  Deuticke,  Wien,  1907. 

39  Calmette.     C.  E.  de  I'Acad.  des  Sciences,  June,  1907. 

10  Wolff-Eisner.  Berl.  klin.  Woch.,  1907,  p.  1052.  Discussion  of  paper 
by  Citron. 


CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS       443 

the  ophthalmoreaction.  The  principle  involved  in  these  reactions 
was  then  further  emphasized  by  the  introduction  of  the  Moro  tuber- 
culin-ointment method  and  the  intracutaneous  test  of  Romec.  The 
mere  observation  that  the  infection  with  tuberculosis  results  in  a 
general  tissue  hypersusceptibility  immediately  suggests  the  interpre- 
tation of  the  tuberculin  reaction  as  a  manifestation  of  anaphylaxis. 
Von  Pirquet,  accordingly,  on  the  basis  of  his  previous  studies  upon 
serum  sickness  includes  the  reaction  in  the  category  of  what  he  calls, 
"allergic." 

He  assumes  that  the  reaction  depends  upon  the  presence  in  the 
system  of  antibodies,  which  form  a  union  with  the  applied  tubercu- 
lin, the  result  being  the  formation  of  poisons  and  a  reaction.  This 
assumption,  according  to  the  anaphylatoxin  theory  of  Friedberger, 
would  imply  the  participation  of  alexin  in  the  reaction — acting  upon 
the  united  tuberculin-antituberculin  complex,  though  v.  Pirquet 
does  not  express  himself  positive  as  to  this. 

This,  moreover,  is  the  clearly  expressed  opinion  of  Friedberger41 
himself.  Consistently  with  his  general  theory  of  anaphylaxis  he 
assumes  that  the  injected  tuberculin  comes  into  relation  with  specific 
antibodies  with  which  it  unites,  the  alexin  then  splitting  off  anaphy- 
latoxin from  the  complex.  He  bases  this  view  upon  his  experimental 
demonstration,  mentioned  above,  of  the  production  of  anaphylatoxin 
from  tubercle  bacilli  by  in  vitro  digestion  with  guinea  pig  comple- 
ment. 

In  principle  the  view  of  v.  Pirquet  is  similar  to  that  previously 
expressed  by  Wolff-Eisner42  that  the  union  of  tuberculin  with  its 
lytic  antibody,  present  in  the  tuberculous  animal,  gave  rise  to  pois- 
ons as  the  result  of  lysis.  Both  of  these  theories  simply  apply  to  the 
special  case  of  the  tuberculin  reaction  theories  of  mechanism  applied 
to  anaphylactic  reactions  in  general. 

We  must  admit  that  the  facts  of  the  "allergic"  reactions  as  a 
class  seem  to  force  upon  us  the  acceptation  of  von  Pirquet's  views. 
Apart  from  the  purely  clinical  observations  made  in  carrying  out  the 
routine  tests  we  have  the  additional  evidence  that  the  instillation  of 
tuberculin  into  the  eye  of  normal  individuals  gives  rise  to  no  reac- 
tion, but  a  repetition  of  the  instillation  into  the  same  eye  after  ten 
or  more  days  results  in  a  marked  and  typically  positive  test.  Further- 
more, von  Pirquet  43  states  that  individuals  showing  no  clinical  tu- 
berculosis and  negative  to  a  first  test  will  often  react  ("sekundare 
Reaktion")  to  a  second  test  carried  out  a  few  days  after  the  first. 

These  facts  all  seem  to  indicate  acquired  hypersusceptibility  more 
analogous  to  true  serum-anaphylaxis  than  to  the  toxin  hyper suscepti- 

41  Friedberger.     Munch,  med.  Woch.,  Nos.  50  and  51,  1910. 

42  Wolff-Eisner.    Berl.  klin.  Woch.,  Nos.  42  and  44,  1904. 

43  Cited  from  Lowenstein  in  "Kraus  u.  Levaditi  Handbuch,"  Vol.  1,  p. 
1039. 


444  INFECTION    AND    RESISTANCE 

bility  of  von  Behring,  in  that  the  tuberculin  is  but  slightly  toxic  in 
itself. 

If  the  analogy  is  such  a  close  one,  therefore,  it  should  be  easy  to 
formulate  experiments  by  which  the  phenomena  now  ascertained 
regarding  serum-anaphylaxis  could  be  demonstrated  for  tuberculin 
hypersusceptibility.  The  obvious  procedure,  therefore,  would  be  to 
attempt  to  passively  transfer  tuberculin  sensitiveness  to  a  normal 
animal  with  the  serum  of  a  tuberculous  one.  This  has  indeed  been 
attempted  by  Friedemann,44  later  by  Bauer 45  and  a  number  of 
others — usually  with  negative  result.  The  writer,  hoping  to  develop 
a  diagnostic  method  for  tuberculosis,  has  also  attempted  this  by  the 
transference  of  human  tuberculous  blood  to  guinea  pigs,  but  invari- 
ably obtained  negative  results.  Yamanouchi46  alone  has  reported 
positive  experiments  by  a  similar  procedure  with  rabbits,  but  so  far 
his  results,  according  to  Friedemann,  have  completely  failed  of  con- 
firmation. Austrian  succeeded  by  sensitizing  guinea  pigs  with  5 
c.  c.  of  titrated  whole  blood,  using  for  the  second  injection  a  tuber- 
culo-protein  prepared  by  the  method  of  Baldwin.47  In  this  particu- 
lar, therefore,  the  analogy  between  anaphylaxis  and  the  tuberculin 
reaction,  though  not  easily  worked  out,  has  nevertheless  been  estab- 
lished. Another  objection  which  has  been  made  by  a  number  of  ob- 
servers is  the  fact  that  anaphylaxis  is  accompanied  by  temperature 
depression  while  tuberculin  reactions  are  followed  by  a  rise.  This 
objection  may  be  regarded  as  invalid,  however,  in  the  light  of  Fried- 
berger's  48  experiments  which  showed  that  temperature  depression 
follows  only  when  large  doses  of  the  antigen  are  injected  into  the 
sensitized  animals,  smaller  doses  often  giving  rise  to  increased  tem- 
perature. 

We  gain  a  certain  amount  of  insight  into  the  conditions  here 
prevailing  by  considering  the  information  which  has  been  obtained 
from  the  study  of  the  antibodies  formed  in  animals  in  tuberculosis. 
It  appears  from  the  work  of  Christian  and  Rosenblatt  49  that  anti- 
bodies to  the  tubercle  bacillus  are  formed  by  tuberculous  animals 
only.  Normal  animals  form  these  to  a  very  slight  degree  only,  if  at 
all,  when  immunization  with  tuberculin  is  attempted.  In  other 
words,  as  Friedemann  (aWeichhardt's  Jahresbericht,"  6,  1910) 
points  out,  the  specific  reaction  of  antibody  formation  in  tuberculosis 
seems  to  be  closely  associated  with  the  tuberculous  tissues  themselves. 

44  Friedemann.  "tiber  anaphylaxie,"  "Weichhardt's  Jahresbericht,"  Vol.  6, 
1910. 

45  Bauer.     Cited  ibid.;  also  Munch,  med.  Woch.,  1909,  p.  1218. 
[Q  Yamanouchi.     Wien.  klin.  Woch.,  1908,  p.  1263. 

47  Austrian  Bull,  of  the  Johns  Hop.  Hosp.,  Vol.  24,  1913 ;  Baldwin,  Journ. 
Med.  Res.,  Vol.  17,  1910. 

48  Friedberger.    Deutsche  med.  Woch.,  No.  11,  1911. 

49  Christian  and  Rosenblatt.    Munch,  med.  Woch.,  1908. 


CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS       445 

The  same  inference  can  be  made  from  Bail's  50  experiments  on 
passive  sensitization.  For,  although  passive  sensitization  of  guinea 
pigs  with  the  serum  of  tuberculous  animals  has  been  unsuccessful, 
Bail  succeeded  in  obtaining  lethal  anaphylactic  reactions  by  injected 
macerated  tuberculous  tissues,  following  these  on  the  next  day  by 
injections  of  tuberculin.  It  is  plain  from  this,  as  Friedemann  cor- 
rectly argues,  that  we  must  assume  that  the  antibodies  (receptors) 
formed  against  tubercle  bacilli  are  closely  bound  up  with  the  tissue 
cells,  the  reaction  of  tuberculin  being  largely  with  "sessile  receptors." 
Indeed,  it  seems  as  though  the  antibodies  formed  against  tubercle 
bacilli  undoubtedly  remain  in  close  relation  to  the  cells  of  the  tis- 
sues except  in  cases  of  active  tuberculosis  in  which  localized  areas 
of  cells  are  under  the  influence  of  a  very  intense  action  of  the  poi- 
sons, and  a  consequent  overproduction  and  discharge  of  receptors 
(using  the  Ehrlich  nomenclature)  may  occur.  This  would  corre- 
spond with  considerable  accuracy,  moreover,  to  the  histological  facts, 
for  in  this  infection,  similar  to  leprosy  and  a  number  of  other  con- 
ditions, but  unlike  most  acute  infections,  the  battle  against  the  micro- 
organisms is  carried  out  chiefly  by  the  adjacent  tissue  cells. 

We  might  assume,  therefore,  that,  in  tuberculous  individuals, 
there  is  indeed  a  reaction,  at  first  local,  then  generalized  to  a  slight 
degree,  in  which  antibodies  are  actually  formed.  These  antibodies, 
however,  remain  to  a  preponderant  extent  sessile,  or  incorporated  in 
the  reacting  cells.  Upon  the  injection  or  application  of  tuberculin 
the  reaction  takes  place  in  or  upon  the  cells.  Whether  or  not  the 
further  cooperation  of  complement  or  alexin  is  then  necessary  for  the 
lysis  and  poison  production  from  the  antigen,  as  in  the  similar  reac- 
tions taking  place  in  the  circulation,  or  whether  the  intracellular 
ferments  themselves  suffice  for  this,  cannot  be  decided  at  present. 
It  is  certainly  not  unlikely  that  the  circulation  of  tubercle-antigen — 
even  in  small  quantities — throughout  the  body  may  produce  such 
hypersusceptibility  of  cells  (represented  graphically  by  the  concep- 
tion of  sessile  receptors  in  many  parts  of  the  body  remote  from  the 
lesion — a  quality  remaining  constant  for  prolonged  periods  and  ex- 
plaining the  subsequent  skin  and  ophthalmic  reactions  obtained  upon 
test).  Certain  clinical  observations  cited  by  v.  Pirquet  51  would  seem 
to  support  such  a  view.  For  instance,  he  states  that,  having  em- 
ployed his  left  forearm  repeatedly  for  tests,  he  was  able  to  obtain 
positive  reactions  in  this  area  with  tuberculin  diluted  1  to  1,000, 
whereas  his  right  forearm  was  negative  to  tuberculin  ten  times  as 
concentrated.  Furthermore,  as  Kohn  52  has  shown  that,  while  the 
first  injection  of  tuberculin  into  the  eye  of  a  normal  person  produces 
no  reaction,  this  eye  will  not  only  react  to  a  second  instillation, 

50  Bail.     Zeitschr.  f.  Immunitatsforsch.,  Vol.  4. 

51  V.  Pirquet,  "Kraus  u.  Levaditi  Handbuch,"  Vol.  1,  p.  1050. 

52  Kohn.     Quoted  from  Lowenstein,  Kraus  and  Levaditi,  Vol.  1,  p.  1033. 


446  INFECTION    AND    RESISTANCE 

but  will  show  a  reaction  when  the  second  application  is  made  sub- 
cutaneously.  Negative  evidence  pointing  in  the  same  direction  is 
the  observation  that  absolutely  no  influence  is  exerted  upon  the  out- 
come of  tuberculin  reaction  if  the  tuberculin  is  previously  mixed 
with  blood  serum  from  either  positively  or  negatively  reacting 
cases.53  This  would  tend  to  show  that,  whether  reacting  or  not,  the 
factors  which  determined  this  are  certainly  not  present  in  the  cir- 
culating plasma. 

That  the  circulation  of  tubercle  antibodies  in  the  blood  may  even 
interfere  with  the  localized  tuberculin  reaction  is  rendered  likely  by 
the  fact  that  skin  reactions  are  often  negative  in  cases  of  advanced 
tuberculosis,  and  that,  as  we  are  told  by  Dr.  Blair,  of  the  N".  Y. 
Zoological  Park,  such  reactions  are  usually  negative  in  tuberculous 
monkeys  in  which  the  disease  is  invariably  very  rapid  and  acute. 

PRACTICAL  DIAGNOSTIC  USES  OF  ANAPHYLAXIS 

The  specificity  of  the  anaphylactic  reaction  has  led  to  extensive 
attempt  to  utilize  it  for  forensic  protein  determinations  in  the  same 
way  in  which  the  precipitin  test  is  used.  Uhlenhuth,54  Thomsen,55 
Pfeiffer,  and  others  have  carried  on  extensive  experimentation  in 
this  problem,  the  technique,  in  general,  consisting  in  sensitizing 
guinea  pigs  with  solutions  of  the  unknown  protein  (dissolved  blood 
spots,  etc.)  and  testing  them  by  a  second  injection  of  the  suspected 
protein  after  the  usual  anaphylactic  incubation  time.  The  results  of 
such  work  have  shown  that  indeed  positive  reaction  may  be  obtained 
and  diagnosis  made  in  this  way.  However,  the  reactions  are  not  ordi- 
narily very  striking,  and  this  method  is  not  as  reliable  as  the  precipi- 
tin method.  Uhlenhuth 56  believes  that  the  anaphylactic  reaction 
has  value  only  in  cases  in  which  the  amount  of  unknown  protein  is  so 
small  or  so  changed  by  preservation  or  decomposition  that  its  pre- 
cipitable  qualities  have  been  lost. 

Yamanouchi's  57  attempt  to  utilize  anaphylaxis  for  the  diagnosis 
of  tuberculosis,  by  passively  sensitizing  guinea  pigs  with  the  serum 
of  tuberculous  patients  and  testing  subsequently  with  tuberculin,  has 
been  mentioned  before.  Although  he  claims  positive  experiments, 
our  own  experience  with  a  similar  technique  has  given  us  results 
which  were  so  irregular  that  we  feel  that  this  technique  has  very 
slight  practical  value,  if  any. 

Pfeiffer  58  has  attempted  to  apply  anaphylaxis  to  the  diagnosis 

53  V.  Pirquet.     Loc.  cit. 

54  Uhlenhuth.     Deutsche   milit.  Zeitschr.  No.  2,  1909.     Cited  from  same 
author,  Zeitschr.  f.  Immunitatsforsch.,  Vol.  1,  1909. 

55  Thomsen.    Zeitschr.  f.  Immunitatsforsch.,  Vol.  1,  1909. 

56  Uhlenhuth.     Zeitschr.  f.  Immunitatsforsch.,  Ref.  Vol.  1,  1909,  p.  525. 

57  Yamanouchi.     Wien.  klin.  Woch.,  No.  47,  1908. 

58  Pfeiffer.     Zeitschr.  f.  Imm.,  Vol.  4.  1910. 


CLINICAL  SIGNIFICANCE  OF  ANAPHYLAXIS       447 

of  malignant  tumors.  '  Together  with  Finsterer 59  he  sensitized 
guinea  pigs  with  the  serum  of  carcinomatous  patients  following  the 
injection  48  hours  later  with  press-juices  of  tumors.  His  conclu- 
sions were  drawn  from  the  anaphylactic  temperature  reaction,  and 
he  claims  that  animals  so  sensitized  are  hypersusceptible  to  the 
juices  obtained  from  carcinomata,  whereas  animals  sensitized  with 
normal  serum  or  the  serum  of  sarcoma  patients  show  no  hypersus- 
ceptibility.  Ranzi  60  has  not  been  able  to  confirm  this.  The  signifi- 
cance of  such  experiments,  if  correct,  apart  from  their  practical 
value  for  the  diagnosis  of  carcinoma,  would  be  considerable  in  that 
they  tend  to  show  that  cancer  tissues  contain  a  specific  protein 
which  is  antigenically  distinct  from  the  other  tissue-proteins  of 
the  afflicted  individual.  However,  we  cannot  yet  accept  these  facts 
as  absolutely  established. 

59  Pfeiffer  and  Finsterer.     Wien.  klin.  Woch.,  No.  28,  1909. 

60  Ranzi.    Zeitschr.  f.  Imm.}  Vol.  2, 1909. 


CHAPTER    XIX 

THEEAPEUTIC    IMMUNIZATION  IN  MAN 

FACTS  CONCERNING  ANTITOXIN  TREATMENT  IN  MAN 

THERAPEUTIC  USE  OF  DIPHTHERIA  ANTITOXIC 

IT  is  not  consistent  with  the  purpose  of  this  brief  treatise  to  dis- 
cuss extensively  the  therapeutic  benefits  obtained  by  serum  therapy  in 
diphtheria.  We  can  convey  briefly  an  adequate  idea  of  this  by  citing 
some  of  the  tables  given  by  Northrup  in  Nothnagel's  "Encyclopedia 
of  Practical  Medicine,"  American  Edition,  Volume  on  Diphtheria, 
etc.,  p.  143.  These  figures  are  taken  from  the  statistics  of  the  New 
York  Board  of  Health,  which  began  treatment  of  diphtheria  with 
antitoxin  in  January,  1895.  Dr.  Northrup  states,  however,  that 
serum  treatment  cannot  be  considered  to  have  been  in  general  use 
until  some  time  later. 

Without  Antitoxin 


Year 

Cases  reported 

Deaths 

Mortality,  per  cent. 

1891  

5,364 

1,970 

36.7 

1892     

5,184 

2,106 

40  0 

1893  

7,021 

2,558 

36.4 

1894            

9,641 

2,870 

29  7 

Total 

27210 

9,504 

Avg.  34  9 

With  Antitoxin 


1895 

10353 

1  976 

19  0 

1896  

11,399 

1,763 

15  5 

1897 

10896 

1  590 

14  5 

1898  

7  173 

919 

12  8 

1899  

8,240 

1,085 

13.1 

1900  

8364 

1  176 

14  0 

Total 

56425 

8509 

Avg.  15  0 

Table  taken  directly  from  Northrup,  loc.  cit. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         449 

From  this  table  there  appears  a  reduction  of  58  per  cent,  in 
mortality  and  a  similar  drop  is  evident  from  the  German  statistics 
of  Dieudonne,1  from  those  of  Welch,  and  many  others. 

It  should  be  considered,  moreover,  in  reading  such  statistics  that 
they  are  made  on  gross  mortality  reports  without  elimination  of  the 
many  cases  that  have  not  come  under  observation  until  too  severely 
diseased  to  react  to  any  form  of  treatment.  The  reason  for  the  fail- 
ure to  obtain  results  with  antitoxin  when  the  cases  have  proceeded 
beyond  a  certain  stage  of  intoxication  will  become  evident  when  we 
consider  the  manner  of  absorption  of  the  poison  in  a  succeeding  para- 
graph. The  mortality  sinks  to  between  8  and  9  per  cent.,  when  such 
cases  are  omitted,  as  is  shown  by  the  collective  investigations  of  the 
American  Pediatric  Society  in  1896 — figures  which  we  take  also 
from  Northrup's  comprehensive  study.  This  purely  statistical  evi- 
dence, however  good,  is  further  reenforced  by  the  unquestionable 
and  considerable  diminution  of  emergency  operations,2  such  as  intu- 
bation and  tracheotomy,  since  introduction  of  the  antitoxin.  More- 
over, there  is  the  manifold  clinical  evidence  of  benefit,  after  the 
serum  treatment,  familiar  to  every  practicing  physician. 

Although  the  injection  of  antitoxin  is  of  benefit  by  whatever 
route  and  in  whatever  quantity  it  may  be  given,  nevertheless  recent 
experimental  investigations  have  taught  us  much  regarding  the 
proper  use  of  this  therapeutic  agent.  Especially  interesting  are  the 
investigations  of  Meyer,3  who  showed  the  extreme  importance  of  an 
early  use  of  the  antitoxin.  Apparently,  as  we  have  mentioned  in 
another  place,  like  tetanus  antitoxin,  the  diphtheria  poison  may  be 
in  part  absorbed  directly  by  the  nerves.4 

There  is  apparently  a  great  difference  in  therapeutic  efficiency, 
according  to  the  method  in  which  the  serum  is  administered,  a  differ- 
ence probably  depending  upon  speed  of  absorption.  Berghaus5 
showed  that  intravenous  injection  is  500  times  more  potent  therapeu- 
tically  than  the  subcutaneous,  and  80  to  90  times. more  so  than  the 
intraperitoneal  injection.  Schick,  in  discussing  this  problem  from 
the  clinical  point  of  view,  for  this  reason  lays  special  stress  upon  the 
speed  of  administration.  He  says:  "Not  only  days  but  hours  are 
of  great  importance."  He  bases  this  opinion  largely  upon  the  fact 
that  the  toxin  which  has  already  united  with  the  nerve  substance  can 
probably  no  longer  be  influenced  by  the  injection  of  the  serum. 

According  to  the  experiments  of  Meyer  and  Ranson  diphtheritic 

1  Dieudonne.    Arb.  aus  dem  kais.  Gesund.,  XIII,  1897. 

2  Siegert.     "Jahrbuch  f.  Kinderheilkunde,"  Vol.  52,  cited  after  Wernicke. 

3  Meyer.    Berl.  kl  Woch.,  25,  26,  1909 ;  Arch.  f.  exp.  Path.  u.  Ther.,  Vol. 
60,  1909,  and  Berl.  kl.  Woch.,  No.  45,  1911. 

4  For  a  thorough  discussion  of  these  conditions  see  Schick,  Centralbl.  f. 
Bakt.,  Rev.  Vol.  57,  1933,  "Report  of  7th  Meeting  of  the  Mikrobiol.  Gesell.," 
Berlin,  1913. 

5  Berghaus.     Cited  from  Schick,  loc.  cit. 


450 


INFECTION    AND    RESISTANCE 


paralysis  may  follow  even  when  vigorous  serum  treatment  has  been 
employed.  For,  according  to  them,  only  the  toxin  which  has  reached 
the  central  nervous  system  through  the  circulation  can  be  influenced 
by  the  serum,  but  no  effect  is  possible  upon  the  fraction  which  has 
been  absorbed  from  the  nerve  endings  directly. 

Schick,6  on  the  basis  of  extensive  experiments,  comes  to  the  con- 
clusion that  the  subcutaneous  injection  of  1,000  to  2,000  units  in 
diphtheritic  cases  has  an  immunizing  value  only,  which  protects  the 
tissues  from  further  injury  and  leads  to  cure  if,  at  the  time  of  injec- 
tion, the  lethal  dose  has  not  yet  united  with  the  sensitive  cells.  "If," 
he  states,  awe  wish  to  obtain  antitoxic  action  upon  toxin  which  has 
already  gone  into  action  before  the  injection  of  the  serum,  then  re- 
sults can  be  obtained  both  in  man  and  in  animals  only  if  a  great  deal 
of  antitoxin  is  injected  intramuscularly  or  intravenously."  7 

Interesting  also  from  a  clinical  point  of  view  are  the  studies  of 
Schick,8  Hahn,9  and  others10  upon  the  presence  of  antitoxin  in  the 
blood  of  normal,  untreated  individuals  at  different  ages.  These 
investigations  were  carried  out  by  the  intracutaneous  method  of 
toxin  and  antitoxin  determination  described  in  greater  detail  in  a 
later  section.  The  following  table,  taken  from  the  article  of  Hahn, 
illustrates  the  experience,  in  such  investigations,  both  of  Schick  and 
of  Hahn  himself.  The  determinations  were  carried  out  upon  indi- 
viduals who  had  never  had  diphtheria,  as  far  as  could  be  learned. 


Age 

Cases  with 
antitoxin  serum 

Cases  without 
antitoxin  serum 

Highest  antitoxin 
value  in  1  c.  c.11 

Schick  - 

r  Newborn  .... 
L0-l  year  

11 
1 

0 
3 

under  1.5    units 
0.11  unit 

:-  2-10  years.. 

7 

5 

1.0    unit 

11-20  years.  . 

8 

9 

0.75  unit 

Hahn  

21-30  years 

9 

5 

2  5    units 

3  1-40  years.. 

5 

1 

0.25  unit 

L  4  1-65  years.  . 

2 

8 

2  .  5    units 

The  table  shows  that  in  newborn  children  there  is  almost  regu- 
larly a  definite  and  sufficient  protective  value  in  the  serum  which 
diminishes  up  to  the  first  year,  so  that  at  the  end  of  the  first  year 
three  out  of  four  individuals  had  no  antitoxin  in  their  serum.  In 
subsequent  years  up  to  the  age  of  40  an  increasing  percentage  of 

6  Schick.    Loc.  cit. 

J  Schick.    Loc.  cit.,  p.  32. 

8  Schick.  '  "tiber  Diphtherimmunitat,"  Wiesbaden,  1910. 

9  Hahn.     Deutsche  med.  Woch.,  Vol.  38,  No.  29,  p.  1366,  1912. 

10  Karasawa  and  Schick.    Zeitschr.  f.  Kinderkrankheiten,  1910,  and  " Jahr- 
buch  f.  Kinderheilkunde,"  1910. 

11  Table  taken  directly  from  Hahn,  loc.  cit. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         451 

people  have  sufficient  amounts  of  diphtheria  antitoxin  in  their  blood. 
After  the  age  of  40  an  increasing  percentage  is  without  such  protec- 
tion. The  first  observation,  that  newborn  children  usually  possess 
considerable  amounts  of  antitoxin,  is  very  probably  due  to  passive 
immunization  by  the  blood  of  the  mother,  a  fact  which  we  have  men- 
tioned in  another  place.  The  exact  method  by  which  such  measure- 
ments are  made  is  described  in  a  subsequent  section  on  the  intra- 
cutaneous  method  of  determining  toxin  and  antitoxin. 

The  work  of  Schiek,  that  of  J.  Henderson  Smith,  and  recent 
studies  by  Park  and  Biggs  promise  to  alter  considerably  the  methods 
of  antitoxin  therapy  as  at  present  in  use  in  diphtheria.  Smith  meas- 
ured the  speed  of  absorption  of  antitoxin  injected  subcutaneously  into 
the  abdominal  wall  of  a  healthy  man.  His  results  are  shown  in  the 
following  table,  which  we  take  from  his  communication  (page  213)  : 

TABLE  V 

One  c.  c.  of  the  patient's  serum  contained: 

Before  injection No  demonstrable  antitoxin 

5  hours  after  injection 0.1      unit  antitoxin 

14  hours  after  injection 0 . 225  unit  antitoxin 

32  hours  after  injection 0.68    unit  antitoxin 

44  hours  after  injection 1.0      unit  antitoxin 

3  days  after  injection 1.3  units  antitoxin 

4  days  after  injection 1.3  units  antitoxin 

6  days  after  injection 0 . 68  unit  antitoxin 

13  days  after  injection 0 . 17  unit  antitoxin 

15  days  after  injection 0 . 14    unit  antitoxin 

20  days  after  injection 0 .08    unit  antitoxin 

27  days  after  injection No  demonstrable  antitoxin12 

Park  and  Biggs  13  have  made  similar  studies  and  have  contrasted 
the  speed  of  absorption  after  subcutaneous  administration  with  that 
after  intravenous  injection,  basing  their  curves  upon  careful  measure- 
ments of  the  sera  of  the  treated  patients.  We  reproduce  their 
charts  as  given  in  their  recent  publication. 

It  is  apparent  from  these  charts,  as  well  as  from  the  work  of  Hen- 
derson Smith,  that  antitoxin,  subcutaneously  given,  is  slowly  ab- 
sorbed, and  does  not  reach  its  maximum  concentration  in  the  blood 
stream  until  forty-eight  hours  or  more  after  the  injection.  It  fol- 
lows that,  as  Park  and  Biggs  point  out,  it  is  more  rational  to  inject 
a  single  adequate  dose  than  to  divide  the  dosage  and  inject  at  inter- 
vals. They  have  obtained  results  in  animal  experiment  which 
graphically  illustrate  this  principle.  A  rabbit  which  had  received 
ten  fatal  doses  of  toxin  intravenously  was  given  a  total  of  500  anti- 
toxin units  in  divided  doses  as  follows :  100  units  after  twenty  min- 

12  J.  Henderson  Smith.     Journal  of  Hygiene,  Vol.  7,  1907,  p.  205. 

13  Park  and  Biggs.     Collected   Studies  from  the  N.   Y.   Department   of 
Health,  Bureau  of  Laboratories,  Vol.  7,  1912-1913,  p.  27. 


452 


INFECTION    AND    RESISTANCE 


6.0 


CHART  I. — Showing  the  extent  and  rapidity  of  absorption  of  10,000  units  of 
antitoxin  given  subcutaneously.  Each  line  represents  the  antitoxin  content 
of  1  c.  c.  of  blood  at  different  intervals  of  time.  (From  Park  and  Biggs, 
loc.  cit.) 


THERAPEUTIC    IMMUNIZATION    IN    MAN 


453 


CHART  II. — The  antitoxic  power  of  human  blood  after  an  intravenous  injection 
of  10,000  antitoxic  units.     (From  Park  and  Biggs,  loc.  cit.) 

utes,  100  after  40  minutes,  and  150  units  each  after  60  and  80  min- 
utes. This  rabbit  died.  Another  animal  given  the  same  dose  of 
toxin  received  200  units  of  antitoxin  twenty  minutes  later  and  lived. 
The  amount  necessary  to  save  life  in  rabbits  receiving  ten  fatal  doses 
intravenously  was  as  follows : 

Given  after  10  minutes 5  units  antitoxin 

Given  after  20  minutes 200  units  antitoxin 

Given  after  30  minutes 2,000  units  antitoxin 

Given  after  45  minutes 4,000  units  antitoxin 

Given  after  60  minutes 5,000  units  antitoxin 

Given  after  90  minutes . .  No  amount 


These  extremely  important  experiments  of  Park  and  Biggs  bear 
out  the  opinion  of  Schick  and  show  beyond  question  that  the  proper 
way  to  give  antitoxin  is  to  give  a  single  adequate  dose  as  early  as 
possible.  They  emphasize  the  fact  that  probably  the  most  important 
single  point  in  the  specific  therapy  of  diphtheria  is  the  speed  with 
which  the  diagnosis  can  be  made  and  the  antitoxin  given.  At  the  De- 
partment of  Health  the  dosage  now  employed,  as  given  by  Park  and 
Biggs,  is  the  following : 


UNITS  ii 

i  CASES 

Mild 

Moderate 

Severe 

Very  Severe 

Infants  under  1  year 

2,000 

3,000 

10,000 

10000 

Children  1  to  5  years       

3,000 

5,000 

10,000 

10,000 

Children  5  to  9  years  
Persons  over  10  years     

4,000 
5,000 

5,000 
10,000 

10,000 
10,000 

15,000 
20,000 

454  INFECTION    AND    RESISTANCE 

PRACTICAL  CONSIDERATIONS  CONNECTED  WITH  DIPHTHERIA  ANTI- 
TOXIN PRODUCTION  AND  STANDARDIZATION 

The  conditions  which  govern  the  active  production  of  toxins  by 
bacteria  in  culture  media  are  not  only  of  great  theoretical  interest 
but  possess  unusual  practical  value  in  that  the  most  important  factor 
for  the  successful  production  of  a  strong  antitoxin  consists  in  the 
preliminary  preparation  of  a  potent  toxin.  The  bacterial  true  toxins 
are  all  "exotoxins"  in  that  they  are  soluble,  moderately  diffusible 
substances  which  pass  readily  from  the  bacterial  bodies  to  the  en- 
vironment, and  for  this  reason  can  be  obtained  most  readily  by  the 
cultivation  of  the  bacteria  upon  fluid  media  and  subsequent  filtration 
of  the  cultures  through  earth  or  porcelain  filters. 

The  choice  of  culture  or  strain  is  an  important  element  in  this 
procedure,  since  within  the  same  species  of  toxin-producing  micro- 
organisms there  is  much  variation  in  the  speed  and  energy  of  toxin 
production.  Thus  for  unknown  reasons  some  strains  of  diphtheria 
bacilli  will  far  outstrip  others  in  this  respect.  An  excellent  illustra- 
tion of  this  is  the  experience  of  Park  and  Williams  14  with  two  diph- 
theria cultures — a  very  virulent  and  a  very  weak  one.  Of  the 
former,  0.002  c.  c.  of  a  forty-hour  bouillon  culture  killed  a  guinea 
pig,  while  of  the  latter  0.1  c.  c.  of  a  similar  culture  was  necessary 
for  the  same  result. 

In  the  case  of  tetanus,  cultural  differences  do  not  seem  to  be  as 
common.  Individual  strains  also  may  gain  or  lose  in  toxin-pro- 
ducing powers,  according  to  the  method  of  handling  them  which  is 
practiced.  It  is  stated,15  for  instance,  that  a  diphtheria  culture  will 
lose  in  energy  of  toxin  production  if  permitted  to  grow  without  suffi- 
ciently frequent  transplantation.  However,  transplanted  on  solid 
media  with  reasonable  frequency,  these  bacteria  show  a  remarkably 
constant  toxin  production.  A  well-known  strain,  the  Park-Williams 
No.  8,  now  in  use  in  many  antitoxin  laboratories  throughout  the 
world,  has  persisted  for  over  15  years  in  producing  a  strong  toxin. 
There  are  occasional  strains  among  toxin-forming  species  which  are 
entirely  devoid  of  this  property.  Diphtheria  bacilli  which  were 
virulent  while  possessing  all  the  other  cultural  characteristics  of  the 
group  have  been  described,  but  appear,  from  the  experience  of  the 
writer,  to  be  rather  uncommon.16  Of  tetanus  bacilli  little  is  known 
in  this  respect. 

Given  a  powerfully  toxic  strain  of.  the  proper  bacteria  the 
method  of  cultivation  is  also  of  great  importance  in  influencing  the 
eventual  yield  of  poison.  These  relations  have  naturally  been  stud- 
ied with  the  greatest  care  in  the  case  of  diphtheria  and  tetanus 

14  Park  and  Williams.     "Pathogen.  Micro-organ.,"  N.  Y.,  1910. 

15  Park  and  Williams.     Loc.  cit. 

16  Zinsser.    Jour.  Med.  Ees.,  N.  S.,  Vol.  12,  1907. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         455 

bacilli,  since  in  these  cases  there  has  been  the  greatest  practical  appli- 
cation for  such  knowledge. 

In  the  case  of  diphtheria,  though  toxin  will  be  produced  on  all 
media  on  which  the  bacillus  grows  easily,  the  most  favorable  medium 
for  this  purpose  is  a  slightly  alkaline  broth  made  of  lean  beef  or 
veal  infusion  and  containing  peptone.  Since  acid  formation  hinders 
the  production  of  toxin,  Martin  17  has  suggested  fermentation  of  the 


APPARATUS   ARRANGED  FOR  THE  STERILE  FILTRATION  OF  DIPHTHERIA  CULTURES 

IN  TOXIN  PRODUCTION. 
(After  Eosenau,  U.  S.  Hyg.  Lab.  Bull.  21,  1905,  p.  38.) 


muscle  sugar  with  yeast,  while  Theobald  Smith  18  recommends  pre- 
liminary fermentation  with  Bacillus  coli. 

Park  and  Williams  19  regard  this  as  unnecessary.  They  recom- 
mend a  2  per  cent,  peptone  broth  made  of  veal.  This  is  neutralized 
to  litmus  and  7  to  9  c.  c.  of  normal  NaOH  solution  to  the  liter  are 
added.  In  such  a  medium  at  37.5°  C.  the  production  of  toxin  begins 
within  24  hours  and  reaches  its  highest  point  in  from  five  to  ten 
days.  When  at  its  height  the  process  must  be  stopped  and  the  cul- 
tures exposed  to  a  lower  temperature,  otherwise  rapid  deterioration 
takes  place  because  of  the  instability  of  the  toxin.  Even  when  kept 
cold  and  in  the  dark  this  deterioration  proceeds  steadily  though 
slowly.  At  first,  however,  even  under  these  conditions  a  compara- 
tively extensive  loss  of  toxin  goes  on — a  process  sometimes  spoken 
of  as  "maturing  of  the  toxin" — after  which  the  poison  strikes  a 

17  Martin.    Ann.  Past.,  1896. 

18  Th.  Smith.    Journ.  Exp.  Med.,  IV,  1899,  p.  373. 

19  Park  and  Williams.     Journ.  Exp.  Med.,  Vol.  1,  1896. 


456  INFECTION    AND    RESISTANCE 

fairly  constant  and  very  gradual  rate  of  weakening,  and  is,  com- 
paratively speaking,  stable. 

In  the  United  States  Hygienic  Laboratory  in  Washington,  ac- 
cording to  Rosenau,  the  recommendations  of  Theobald  Smith  are 
largely  followed  in  the  production  of  toxin.  The  procedure  is  as 
follows : 

The  culture  medium,  "Smith's  Bouillon,"  is  prepared  from 
chopped  beef  from  which  fat  and  tendon  have  been  cut  out.  This  is 
adjusted  by  phenolphthalein  titration  to  0.5  per  cent,  acidity.  It  is 
then  placed  into  Fernbach  flasks  and  inoculated  on  the  surface  with 
a  Park-Williams  bacillus  ~No.  8.  The  flasks  are  incubated  for  7  days 
at  37.5°  C.  The  reaction  of  the  medium  after  such  incubation  is 
determined,  and  flasks  showing  an  acidity  of  1.5  or  over  are  dis- 
carded. The  usual  reaction  at  the  end  of  incubation  is  0.6  to  0.8 
per  cent,  acidity.  This  broth  is  filtered  through  Berkefeld  filters  or 
porcelain  candles. 

Toxin  so  prepared  is  now  tested  and  its  L0  and  L+  doses  deter- 
mined by  the  methods  described  above.  Rosenau20  states  that  poi- 
sons are  discarded  as  containing  too  large  a  proportion  of  toxon  if 
the  difference  between  L0  and  L+  is  greater  than  15  M  L  D.  The 
toxin  is  now  set  aside  in  flasks  for  the  process  which  Rosenau 
calls  "seasoning."  At  intervals  of  about  a  month  it  is  retested 
and  finally  it  is  found  that  the  rate  of  toxoid  formation  decreases 
and  the  poison  reaches  a  period  of  equilibrium.  It  can  now  be  used 
for  accurate  determination  of  the  L+  dose,  and  this  is  done  from 
careful  measurements  on  a  large  number  of  guinea  pigs. 

Examples  21  of  such  measurements,  abbreviated  for  the  sake  of 
simplicity,  are  given  in  the  following  tables : 

Toxin  Determinations  of  M  L  D  or  "T" 

Dose  in  c.  c.  Result 

0.03  =  death  in  1^  days 

0.02  =  death  in  \y%  days 

0.01  =  death  in  2      days 

0.008  =  death  in  3      days 

0.006  =  death  in  3}^  days 

0.005  =  death  in  4      days  M  L  D 

0.004  =  death  in  6      days 

0.003  =  death  in  8      days 

0 . 002  =  late  paralysis 

0.001  =  well  in  16  days. 

Toxin  Determination  of  L+  Dose 

1  Antitoxin  unit  -f-  0 . 2    c.  c.  =  0 

1  Antitoxin  unit  -j-  0.21  c.  c.  =  0  =  L0 

1  Antitoxin  unit  +  0.22  c.  c.  =  local  infiltration 

20  Rosenau.    Hyg.  Lab.  Bull  No.  21,  April,  1905. 

21  Examples  are  taken  from  measurements  reported  by  Rosenau,  loc.  cit. 


THERAPEUTIC    IMMUNIZATION    IN   MAN         457 

1  Antitoxin  unit  +  0.23  c.  c.  =  fatal  in  17  days 

1  Antitoxin  unit  -j-  0.24  c.  c.  =  fatal  in  14  days 

1  Antitoxin  unit  -f  0 . 26  c.  c.  =  fatal  in    9  days 

1  Antitoxin  unit  -j-  0 . 28  c.  c.  =  fatal  in    6  days 

i  Antitoxin  unit  -f  0.29  c.  c.  =  fatal  in    4  days  =  L+ 

1  Antitoxin  unit  +  0.3    c.  c.  =  fatal  in    3  days 

The  production  of  antitoxin  is  carried  out  by  the  graded  injec- 
tion of  antitoxin  into  horses.  Young,  healthy  horses  are  chosen, 
tested  for  freedom  from,  glanders,  and  the  first  injections  are  made 
either  with  toxin  attenuated  by  the  addition  of  Lugol's  solution  or 
terchlorid  of  iodin,  or,  as  in  the  New  York  Health  Department,  the 
first  injections  consist  of  mixtures  of  toxin  and  antitoxin.  We  take 
our  description  largely  from  the  account  given  by  Park.22  The  first 
injection  consists  of  12  c.  c.  of  toxin  (M  L  D  1/400  c.  c.),  together 
with  100  units  of  antitoxin.  After  the  reaction  from  such  an  injec- 
tion has  completely  subsided — after  3  to  5  days- — a  second  injection 
is  given  of  toxin  without  antitoxin;  then  15  c.  c.,  45  c.  c.,  55  c.  c., 
65  c.  c.,  80  c.  c.,  95  c.  c.,  115  c.  c.,  140  c.  c.,  etc.,  the  intervals  be- 
tween injections  being  about  three  days  and  depending  upon  the 
reaction  of  the  horse  and  the  speed  with  which  it  entirely  recovers 
from  the  preceding  injection.  In  a  particular  case  cited  by  Park 
675  c.  c.  of  toxin  could  be  given  by  the  60th  day;  in  this  case  by  the 
28th  day  the  horse  was  yielding  225  units  to  the  c.  c. ;  on  the  40th 
day,  850  units;  on  the  60th  day,  1,000  units. 

The  determination  of  the  antitoxin  unit,  carried  out  from  time 
to  time  on  the  serum  of  such  a  horse  against  the  L+  dose  described 
in  our  preceding  table,  would  be  carried  out  as  follows : 

In  all  such  standardization  great  care  must  be  taken  in  employ- 
ing accurately  standardized  glassware.  Rosenau  recommends  em- 
ploying "capacity  instruments"  rather  than  "outflow  instruments." 
Dilutions  of  unknown  antitoxin  are  made  in  0.85  per  cent,  sterile 
salt  solution.  As  a  basic  dilution  one  part  of  the  antitoxic  serum  to 
nine  of  the  salt  solution  gives  1/10  c.  c.  to  each  cubic  centimeter, 
and  from  this  initial  dilution  further  dilutions  may  be  easily  made 
as  follows :  1  c.  c.  of  dilution  I.  +  9  c.  c.  salt  solution  =  1-100,  etc. 
A  series  of  mixtures  is  then  made  in  each  of  which  the  quantity  of 
toxin  equals  the  L+  dose,  and  in  which  the  quantity  of  antitoxin 
varies  within  a  wide  margin  of  the  limits  of  strength  to  be  expected. 
This  is  illustrated  in  the  following  table : 

L+  (0.29  c.  c.)  +  1/500     c.  c.  of  antitoxic  serum  =  lives 

L+  (0.29  c.  c.)  -j-  1/600     c.  c.  of  antitoxic  serum  =  lives 

L+  (0.29  c.  c.)  -j-  1/700     c.  c.  of  antitoxic  serum  =  lives 

L+  (0.29  c.  c.)  -j"  1/800     c.  c.  of  antitoxic  serum  =  dies  in  8  days 

L+  (0.29  c.  c.)  4"  1/900     c.  c.  of  antitoxic  serum  =  dies  in  4  days 

L+  (0.29  c.  c.)  +  1/1,000  c.  c.  of  antitoxic  serum  =  dies  in  2  days 

22  Park  and  Williams.     "Pathogenic  Bacteria,"  p.  213. 


458 


INFECTION    AND    RESISTANCE 


In  the  above  tables,  according  to  our  previous  definition  of  the 
antitoxin  unit,  the  serum  would  contain  900  units  to  the  cubic  centi- 
meter, since  1/900  c.  c.,  injected  together  with  the  L+  dose  of  the 
standard  toxin,  resulted  in  the  death  of  the  guinea  pig  in  four  days. 
In  order  to  allow  a  margin  of  safety  Rosenau  and  others  have  sug- 
gested that  the  unit  should  be  determined,  not  by  the  quantity  of 
antitoxin,  which  delays  death  by  the  L+  dose  for  four  days,  but 
rather  by  the  quantity  which,  with  the  L+  dose,  results  in  saving 
the  life  of  the  guinea  pig.  According  to  this  latter  standard  the 
serum  employed  in  the  table  would  be  spoken  of  as  containing  700 
units  to  the  cubic  centimeter.  Of  course  the  tabulated  measurements 


fnMMMMT 


BATTERY  OF  EOSENAU  SYRINGES  PREPARED  FOR  ANTITOXIN  STANDARDIZATION. 
(Taken  from  Eosenau,  U.  S.  Hygienic  Bulletin  21,  1905.) 


are  rough,  leaving  an  undetermined  zone  of  100  units.  The  exact 
number  of  units  to  the  cubic  centimeter  could,  of  course,  be  deter- 
mined with  greater  accuracy  by  now  carrying  out  another  series  of 
tests  in  which  the  amount  of  serum  varied  between  1/700  and  1/900 
of  a  cubic  centimeter. 

In  carrying  out  such  a  standardization  the  toxin  is  diluted  so 
that  the  L-j.  dose  is  contained  in  2  c.  c.  This  can  easily  be  done. 
For  instance,  in  the  above  the  L+  dose  being  0.29,  it  merely  necessi- 
tates adding  to  each  0.29  c.  c.  of  toxin  1.71  c.  c.  of  salt  solution, 
to  each  2.9  c.  c.,  17.1  c.  c.  The  antitoxin  also  is  made  up  in  such  a 
way  that  the  required  dilution  is  contained  in  two  cubic  centimeters, 
since  a  total  volume  of  4  c.  c.  has  been  agreed  upon  as  standard  for 
these  tests,  the  injected  volume  having  much  influence  upon  the  speed 
of  absorption.  In  using  the  so-called  Rosenau  syringe,  shown  in  the 
figure  for  the  standardization  of  antitoxin,  the  antitoxin  is  made  up 
to  1  c.  c.  in  each  case,  so  that  1  c.  c.  of  salt  solution  may  be  added  to 
wash  out  the  syringe  after  injection  of  the  mixture.  The  mixtures 


THERAPEUTIC    IMMUNIZATION    IN    MAN         459 

can  be  made  directly  in  these  syringes  or  in  test  tubes,  and  are 
allowed  to  stand  one  hour  at  room  temperature,  so  that  there 
may  be  time  for  complete  union.  If  the  mixtures  are  made  di- 
rectly in  the  syringes  the  needles  are  dipped  into  sterile  vaselin, 
which  closes  them  and  prevents  leakage  while  standing.  The 
mixture  is  then  forced  out  of  the  syringe  with  a  rubber  bulb,  thus 
ensuring  complete  injection  of  all  the  fluid.  As  Rosenau  states, 
much  depends  on  the  guinea  pigs.  They  must  be  of  standard 
weight,  about  250  grammes,  well  fed  and  cared  for,  and  must  not 
be  descendants  of  pigs  that  have  shown  marked  or  unusual  resist- 
ance to  diphtheria  toxin.  This,  as  Theobald  Smith  has  shown, 
occasionally  happens. 

The  antitoxic  serum  as  obtained  from  the  horse  directly  may 
be  concentrated  in  a  number  of  ways,  representative  of  which  is  the 
method  developed  at  the  New  York  Department  of  Health  by  Gib- 
son,23 Banzhaff,  and  others.24  The  original  method  consisted  in 
heating  horse  serum  to  56°  C.  for  12  hours,  by  which  some  of  the 
pseudoglobulin  was  converted  into  euglobulin,  the  antitoxin  remain- 
ing in  the  pseudoglobulin  fraction.  After  this  an  equal  volume  of 
saturated  ammonium  sulphate  solution  is  added  and  the  globulin 
precipitated.  After  several  hours  the  precipitate  is  filtered  off  and 
again  taken  up  in  water  corresponding  in  amount  to  the  original 
volume  of  serum.  After  filtration  this  solution  is  precipitated  with 
ammonium  sulphate  and  this  precipitate  is  treated  with  saturated 
solution  of  E"aCl  in  quantity  twice  that  of  the  original  serum.  After 
standing  for  12  hours  the  supernatant  fluid  containing  the  antitoxin 
is  decanted,  and  this  is  precipitated  with  0.25  per  cent,  acetic  acid. 
The  resulting  precipitate  is  dried  by  pressing  it  between  filter  papers 
and  is  placed  in  a  parchment  dialyzing  bag,  after  neutralization  with 
sodium  carbonate.  At  the  end  of  seven  or  more  days  of  dialyzation 
against  running  water,  the  globulin  solution  remaining  in  the  dial- 
yzer  is  filtered  and  made  isotonic. 

More  recently  the  method  as  modified  by  Banzhaff  is  as  follows : 
The  serum,  as  obtained  from  the  horse,  is  diluted  by  one-half  the 
volume  of  water,  and  to  this  a  saturated  solution  of  ammonium 
sulphate  is  added  up  to  30  per  cent,  saturation.  This  is  heated  to 
61°  C.  for  two  hours.  It  is  then  filtered  and  the  residue  on  the  filter 
paper,  which  contains  the  antitoxin,  is  thoroughly  dried  by  pressing 
between  filter  papers  and  is  directly  dialyzed. 

Observations  by  Park  and  Throne  25  have  shown  that  this  con- 
centrated antitoxin  which,  according  to  Gibson,  represents  a  yield  of 
about  TO  per  cent,  original  antitoxic  power  of  the  serum,  is  equally 
efficient  for  therapeutic  purposes  as  an  unconcentrated  preparation 

23  Gibson.     Journ.  of  Biol  Chem.,  Vol.  1,  1906. 

24  Gibson  and  Collins.    Journ.  of  Biol.  Chem.,  Vol.  3,  1907. 

25  Park  and  Throne.     Amer.  Journ.  of  Medical  Science,  Vol.  132,  1906. 


460  INFECTION    AND    RESISTANCE 

and  has  the  advantage  of  introducing  less  foreign  protein  into  the 
human  body.  It  retains  its  potency,  according  to  Park  and  Throne, 
as  long  as  does  the  whole  serum. 

ACTIVE  IMMUNIZATION  IN  DIPHTHERIA  WITH  MIXTURE  OF  TOXIN 

AND  ANTITOXIN 

Recently  Behring26  has  advocated  the  immunization  of  human 
beings  with  mixtures  of  diphtheria  toxin  and  antitoxin.  This 
method  represents  essentially  active  immunization  with  toxin  ren- 
dered harmless  by  neutralization  with  antitoxin.  The  use  of  such 
mixtures  had  previously  been  studied  with  considerable  care,  in  the 
case  of  the  toxin  of  symptomatic  anthrax,  by  Schattenfroh  and 
Grassberger,27  and  the  procedure  had  been  used  in  the  New  York 
Department  of  Health  for  some  years  in  the  initial  treatment  of 
antitoxin  horses.  Theoretically  considered  on  the  basis  of  Ehrlich's 
opinions,  one  would  be  inclined  to  wonder  at  the  fact  that  relatively 
neutral  mixtures  of  toxin  and  antitoxin  should  possess  any  antitoxin- 
inciting  properties.  Behring  explains  the  immunizing  value  of  such 
mixtures  by  the  reversible  nature  of  toxin-antitoxin  union  in  the 
animal  body.  He  calls  attention  to  the  fact  that  our  analyses  of 
diphtheria  toxin-antitoxin  mixtures  have  been  made  entirely  with 
guinea  pigs  as  indicators.  In  studying  such  mixtures  in  other  ani- 
mals Behring  has  come  to  the  conclusion  that  complete  detoxication 
of  the  poison  in  vitro  does  not  occur.  He  found,  for  instance,  that  a 
toxin-antitoxin  mixture  that  wyas  entirely  innocuous  for  guinea  pigs 
produced  an  active  febrile  reaction  in  an  ass.  In  monkeys  (Maca- 
cus  rhesus)  he  finally  found  an  animal  in  which  he  obtained  evidence 
satisfactory  to  him  that  toxin  may  be  powerfully  active  in  the  animal 
body,  even  if  it  has  been  previously  mixed  with  antitoxin.  If,  for 
instance,  he  gave  a  monkey  a  mixture  in  which  as  much  as  20  to  40 
antitoxin  units  were  mixed  with  one  toxin  unit,  and  repeated  the 
injection  two  or  three  times,  the  animal  died  of  subacute  diphtheria 
toxin  poisoning.  The  mixture  ceased  to  be  poisonous  for  monkeys 
only  when  the  relation  of  antitoxin  to  toxin  became  one  of  80  to  100 
antitoxin  units  to  one  toxin  unit.  This  final  detoxication  when  suffi- 
cient amounts  of  antitoxin  were  used,  it  seems  to  us,  may  be  taken 
as  sufficient  evidence  that  Behring' s  monkeys  did  not  die  of  ana- 
phylaxis. 

We  gather  from  Behring' s  writings  that  he  attributes  these  dif- 
ferences in  susceptibility  to  toxin-antitoxin  mixtures  in  various  ani- 
mals to  differences  in  the  reversibility  of  the  toxin-antitoxin  com- 
plex in  the  bodies  of  the  individual  species. 

26  Behring.    Deutsche  med.  Woch.,  Vol.  39,  No.  19,  1913. 

27  Schattenfroh  and  Grassberger.     Deuticke,  Wien,  1904 ;  see  also  Schat- 
tenfroh, Wien.  kl  Woch.,  No.  39,  Sept.,  1913. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         461 

Human  beings  are  less  susceptible  to  such  mixtures  than  are 
monkeys,  but  nevertheless  more  so  than  guinea  pigs.  It  also  appears 
that  diphtheria  bacillus  carriers  or  such  persons  who,  because  of  a 
previous  infection,  have  antitoxin  in  their  blood  are  much  more  sus- 
ceptible to  these  mixtures  than  are  others.  Newborn  children  are 
less  susceptible  than  are  children  from  4  to  15  years.  Mixtures 
which  are  entirely  neutral  for  the  newborn  may  incite  febrile  reac- 
tion in  older  children.  In  all  cases  the  injection  of  such  mixtures  is 
followed  by  a  more  or  less  active  production  of  antitoxin. 

The  mixtures  which  von  Behring  advocates  at  present  are  so 
prepared  that  the  toxin  action  upon  guinea  pigs  is  practically  nil; 
in  other  words,  the  mixture  is  completely  neutralized. 

The  method  represents  in  purpose,  and  apparently  in  achieve- 
ment, a  safe  process  of  actively  immunizing  against  diphtheria. 
Heretofore  the  method  of  protecting  human  beings  prophylactically 
against  diphtheria  has  consisted  in  the  injection  of  antitoxic  serum. 
This,  unquestionably  a  wise  procedure,  has  nevertheless  the  disad- 
vantage of  bringing  about  an  immunity  of  short  duration  only. 
Within  20  to  30  days  the  antitoxin  injected  may  have  completely  or 
almost  completely  disappeared  from  the  blood  stream.  Prophylactic 
immunization  with  the  toxin-antitoxin  mixtures,  however,  repre- 
senting as  it  does  an  active  immunization,  is  likely  to  be  more  pro- 
longed in  its  effects.  According  to  Behring  a  human  being  possess- 
ing 0.01  antitoxin  unit  in  1  c.  c.  of  blood  may  be  regarded  as  still 
moderately  protected  against  diphtheria.  According  to  his  estima- 
tion a  decline  to  this  amount,  in  a  person  actively  immunized  by  the 
mixtures  (an  estimation  based  upon  curve  measurements  of  treated 
cases),  would  take  about  two  years.  He  has  observed  that  horses 
that  had  been  actively  immunized  by  him,  and  subsequently  used  in 
agricultural  work,  retained  measurable  antitoxin  values  in  their 
blood  after  five  years  without  treatment. 

Schreiber28  and  others  state,  also,  that  this  method  of  active 
immunization  with  mixtures  of  toxin  and  antitoxin  has  the  advan- 
tage of  avoiding  the  anaphylactic  dangers  incident  to  the  injection 
of  antitoxin  alone.  Their  opinion  is  probably  erroneous,  since  it  is 
most  likely  that  whatever  anaphylactic  dangers  there  are  result  from 
the  injection  of  horse  serum  rather  than  from  the  antitoxin  con- 
tained in  the  injected  substance.  Moreover,  the  recent  studies  of 
Park  have  shown  satisfactorily  that  the  danger  of  anaphylaxis  in  the 
injection  of  antidiphtheritic  sera  is  practically  nil.  Among  330,000 
cases  on  record  there  were  but  five  deaths. 

The  chief  value  of  this  new  method  of  immunization  is  that  it 
represents  a  safe  technique  for  the  prophylactic  treatment  of  indi- 
viduals exposed  to  the  disease  and  possibly  for  the  general  prophy- 
lactic immunization  of  school  children,  nurses,  physicians,  etc.  In 
28  Schreiber.  Deutsche  med.  Woch.,  Vol.  39,  No.  20,  1913. 


462  INFECTION    AND    RESISTANCE 

the  case  of  children  during  the  ages  at  which  they  are  most  suscep- 
tible to  the  disease,  the  prolonged  immunity  resulting  from  the  treat- 
ment should  strongly  recommend  it  as  a  method  of  promise  for  the 
gradual  eradication  of  epidemics.  Behring  also  suggests  it  as  a 
hopeful  method  of  treatment  in  the  case  of  bacillus  carriers. 

Schreiber  and  others  have  reported  upon  the  effects  of  treatment 
when  carried  out  with  Behring's  mixtures.  In  the  earlier  experi- 
ments of  Hahn,  mixtures  were  used  in  which  there  was  a  slight  excess 
of  toxin.  The  later  experiments  were  made  with  mixtures  which 
were  completely  neutralized  for  guinea  pigs.  In  Schreiber's  cases 
from  two  to  six  injections  were  made  at  intervals  of  three  to  five 
days,  most  of  them  subcutaneously,  and  some  of  them  intramuscu- 
larly. In  no  case  were  there  serious  reactions,  although  occasionally 
there  were  slight  swelling  of  regional  lymph  nodes  and  a  little  fever. 
The  effects  of  immunization  were  noticeable  about  23  to  25  days 
later.  When  two  injections  only  had  been  made,  at  least  0.075  of  an 
antitoxin  unit  to  the  cubic  centimeter  was  present.  The  highest  value 
obtained  after  two  injections  was  one  unit  to  one  cubic  centimeter. 
In  nine  patients  who  had  been  treated  by  four  to  seven  injections 
with  gradually  increasing  doses,  as  much  as  10  to  75  antitoxin  units 
to  the  cubic  centimeter  resulted.  It  appears,  therefore,  that  in  med- 
ical practice  this  method  is  safe,  and  that  with  as  little  as  two  injec- 
tions antitoxin  values  may  be  obtained  which  entirely  suffice  for  the 
protection  of  human  beings  against  the  ordinary  dangers  of  diph- 
theria infection,  an  immunity  which,  as  far  as  we  can  judge  at 
present,  may  last  about  two  years. 

Another  advantage  which  Behring  claims  for  his  method  is  the 
production  of  homologous  antitoxin  in  human  beings  for  the  passive 
immunization  of  other  human  beings.  Mathes  has  tried  this  in 
children  with  the  idea  of  thereby  avoiding  the  dangers  of  anaphy- 
laxis.  Incidentally  it  was  claimed  in  this  case  that  the  passive  im- 
munization, when  carried  out  with  homologous  serum,  lasted  longer 
than  did  that  conferred  by  horse  serum.  However,  one  case  is 
hardly  enough  to  establish  such  a  fact. 


THE  INTRACUTANEOUS  METHOD  OF  DETERMINING  TOXIN  AND 
ANTITOXIN  VALUES 

Marks  29  was  the  first  to  utilize  the  prevention  of  local  edema  or 
injury  for  the  determination  of  antitoxin  values.  He  mixed  diph- 
theria antitoxin  and  toxin  and  injected  them  subcutaneously  into 
guinea  pigs,  claiming  that  this  method  was  considerably  more  deli- 
cate than  the  Ehrlich  method,  since  the  amount  of  toxin  capable  of 
causing  localized  edema  amounted  to  as  little  as  one-twentieth  of  a 

29  Marks.     Centralbl.  f.  Bakt.,  Orig.  Vol.  36. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         465 

minimal  lethal  dose.  This  method  has  many  points  in  its  favor,  and 
has  been  recently  utilized  and  improved  upon  by  Romer. 

Romer  30  31  32  has  developed  a  method  of  diphtheria  antitoxin 
standardization  which  depends  upon  intracutaneous  injections  into 
guinea  pigs.  The  principle  of  this  test  consists  in  the  observation 
that,  when  very  slight  amounts  of  diphtheria  toxin  are  injected  intra- 
cutaneously  into  the  abdominal  skin  of  guinea  pigs,  small  areas  of 
local  necrosis  result  within  about  48  hours.  When  such  injections 
are  made  with  mixtures  of  toxin  and  antitoxin  the  presence  of  free 
toxin  is  indicated  by  the  appearance  of  such  necrosis. 

Before  proceeding  to  the  standardization  by  this  method  it  is 
necessary  to  determine  the  "limes-necrosis"  (just  as  Ehrlich  deter- 
mines his  L+  dose),  that  is,  the  amount  of  toxin  which,  together 
with  a  given  amount  of  toxin  (1/50,  1/200,  or  1/2,000),  will  still 
produce  a  minimal  amount  of  necrosis  after  intracutaneous  injection 
into  guinea  pigs.  It  is  necessary,  therefore,  arbitrarily  to  choose  a 
certain  definite  fraction  of  an  antitoxin  unit  and  mix  this  with  vary- 
ing amounts  of  toxin  and  inject  the  mixtures  into  guinea  pigs  intra- 
cutaneously.  Those  mixtures  in  which  the  toxin  is  fully  neutralized 
will  give  rise  to  absolutely  no  lesion  further  than,  possibly,  a  slight 
local  edema.  Those  in  which  there  is  a  large  excess  of  toxin  will 
cause  extensive  necrosis.  Between  the  two,  in  the  series,  there  will 
be  a  mixture  in  which  slight  local  necrosis  results  from  the  injection. 
In  this  mixture  the  amount  of  toxin,  just  sufficient  to  cause  notice- 
able necrosis  in  spite  of  admixture  with  the  antitoxin,  contains  the 
L-n  (limes  necrosis)  dose. 

When  this  has  been  determined,  then  unknown  antitoxin  can  be 
similarly  measured  against  this  L-n  dose  of  the  standard  toxin.  The 
method  has  the  advantage  of  permitting  one  to  work  with  very  small 
quantities,  since  only  a  small  fraction  of  a  cubic  centimeter  need  be 
used  for  intracutaneous  injections;  also  it  permits  great  economy  of 
animal  material,  since  four  or  five  tests  can  be  simultaneously  car- 
ried out  upon  the  abdominal  wall  of  the  same  guinea  pig. 

The  technique  is  not  easy.  We  have  found  in  studying  this 
method  in  connection  with  some  work  carried  on  in  our  laboratory  by 
Dr.  M.  C.  Terry,  that  a  considerable  amount  of  practice  and  experi- 
ence is  necessary,  both  in  carrying  out  the  procedure  accurately  and 
in  judging  the  lesions.  However,  when  carefully  and  consistently 
done  by  an  experienced  worker,  this  method  gives  results  which  cor- 
respond with  fair  accuracy  to  measurements  made  of  the  same  anti- 
toxin by  the  Ehrlich  method.  This  has  been  the  experience  of 
Lewin,33  and  also  of  Terry  in  the  few  experiments  carried  out  by  him. 

30  Romer.     Zeitschr.  f.  7mm.,  Vol.  3,  p.  208,  1909. 

31  Romer  and  Sarnes.     Ibid.,  p.  344. 

32  Romer  and  Somogyi.     Ibid.,  p.  433. 

33  Lewin.    Centralbl.  f.  Bakt.,  Orig.  Vol.  67,  1913. 


464  INFECTION    AND    RESISTANCE 

The  Homer  method  has  been  recently  used  by  clinicians  for  the 
determination  of  the  presence  of  free  toxin  or  antitoxin  in  the  circu- 
lating blood  of  patients  suffering  or  convalescent  from  diphtheria. 
Homer  himself  suggested  this,  since  his  method  is  adapted  to  the 
determination  of  extremely  slight  amounts  of  either  substance.  A 
recent  study  by  Harriehausen  and  Wirth  34  illustrates  the  results 
obtained  in  such  tests,  formal  human  serum  injected  intracuta- 
neously  into  guinea  pigs  never  caused  necrosis.  Neither  did  the 
similar  injection  of  the  sera  of  children  suffering  from  varicella  and 
other  diseases.  Of  twelve  children  suffering  from  diphtheria,  how- 
ever, serum  taken  before  the  administration  of  antitoxin  caused 
necrosis  upon  intracutaneous  injection  into  guinea  pigs,  in  every 
case.  In  spite  of  the  administration  of  antitoxin,  toxin  was  demon- 
strable in  the  blood  in  five  cases  as  long  as  the  35th  day.  Of  ten 
cases  of  post-diphtheritic  paralysis,  toxin  was  demonstrated  in  the 
blood  of  five. 

Since  this  method  of  determining  antitoxin  values  in  the  blood 
of  human  beings  is  of  considerable  importance  and  may  have  much 
practical  value,  it  may  be  useful  to  insert  an  example  of  such  an 
application  of  this  method  as  used  by  Hahn  35  in  a  series  of  investi- 
gations mentioned  elsewhere. 

The  standard  toxin  was  obtained  from  Marburg.  In  a  series 
of  guinea  pigs  a  determination  was  made  of  the  smallest  quantity  of 
this  standard  poison  which  would  produce  just  noticeable  necrosis 
of  the  skin  if  injected  into  the  pig  intracutaneously,  together  with 
1/2, 000th  of  a  unit  of  a  standard  antitoxin.  The  toxin  and  antitoxin 
were  left  together  for  24  hours  before  injection,  3  hours  in  the  incu- 
bator, and  21  hours  in  the  refrigerator. 

When  this  quantity  of  the  antitoxin  had  been  determined,  it 
could  be  used  in  similar  experiments  and  similarly  mixed  with  vary- 
ing amounts  of  the  patient's  serum.  The  amount  of  antitoxin  pres- 
ent in  such  serum  could  then  be  easily  computed.  For,  let  us  sup- 
pose that  this  amount  of  toxin,  together  with  1/5 00th  of  a  c.  c.  of 
the  serum  injected  intracutaneously  into  the  guinea  pig,  gave  the 
same  amount  of  necrosis  in  the  same  time  as  the  identical  quantity 
of  the  toxin,  together  with  1/2, 000th  of  a  standard  unit.  Then 
l/500th  of  a  patient's  serum  was  equivalent  to  1/2, 000th  of  a  stand- 
ard unit,  and  the  patient's  serum  would  contain  0.25  of  a  unit 
per  cubic  centimeter. 

Michiels  and  Schick  36  have  carried  out  intracutaneous  reactions 
with  diphtheria  toxin  directly  upon  the  human  body  to  determine 
whether  or  not  diphtheria  immunity  was  present.  They  injected 
0.1  c.  c.  of  a  1  to  1,000  dilution  of  toxin  and  claim  that  a  positive 

34  Harriehausen  and  Wirth.     Zeitschr.  f.  Kinderlieilkunde,  Vol.  7,  1913. 

35  Hahn.    Deutsche  med.  Woch.,  Vol.  38,  No.  29,  1912. 

36  Michiels  and  Schick.     Zeitschr.  f.  Kinderheilkunde,  Vol.  5,  1912. 


THERAPEUTIC    IMMUNIZATION    IN    MAN          465 

intracutaneous  reaction  with  this  amount  indicates  an  absence  of 
antitoxin  from  the  blood,  or  at  least  an  insufficient  protection.  The 
Schick  reaction  is  at  present  carried  out  at  the  New  York  Depart- 
ment of  Health,  under  the  direction  of  Park,  with  1/5 Oth  M  L  D 
intracutaneously  injected.  The  dilutions  are  so  made  that  this 
quantity  is  contained  in  a  total  volume  of  0.1  c.  c. 


TETANUS  ANTITOXIN  AND  ITS  STANDARDIZATION 

The  methods  employed  in  the  production  and  standardization  of 
tetanus  toxin  are  in  every  way  analogous  to  those  used  in  the  case 
of  diphtheria  antitoxin.  A  strong  toxin  is  obtained  by  growing  the 
organisms  under  anaerobic  conditions  on  suitable  media.  Accord- 
ing to  Vaillard  and  Vincent37  it  is  essential  that  the  media  upon 
which  the  tetanus  bacilli  are  grown  should  be  freshly  made  and 
sterilized.  Apparently  this  precaution,  which  has  been  similarly 
recommended  by  Wladimiroff,  ISTovy,  and  others,  is  made  necessary 
by  the  gradual  absorption  of  oxygen  which  takes  place  if  the  media 
are  allowed  to  stand  for  a  long  time  without  heating.  It  is  further 
necessary  in  preparing  tetanus  toxin  that  the  culture  medium  should 
not  be  acid,  and  a  weakly  alkaline  initial  titre  is  advised.  For  the 
same  reason,  also,  most  workers  have  advised  against  the  use  of 
glucose  or  other  carbohydrates  in  the  media,  since  the  acid  formed 
by  the  fermentation  of  these  substances  inhibits  growth  and  toxin 
production.  Recently  Hall  38  has  advised  the  use  of  a  simple  meat 
extract  broth  to  which  have  been  added  1  per  cent,  of  dextrose  and 
0.5  per  cent,  of  finely  powdered  magnesium  carbonate.  The  last- 
named  substance,  by  neutralizing  any  acid  that  is  formed  from  the 
glucose,  prevents  the  harmful  acidity.  Anaerobic  conditions  are  ob- 
tained by  growing  the  organisms  under  a  layer  of  oil  in  tightly  stop- 
pered flasks. 

Although  mice  were  formerly  used  in  the  standardization  of 
tetanus  toxin  and  antitoxin,  the  more  recent  usage  has  been  to  sub- 
stitute guinea  pigs  as  in  diphtheria  standardization.  According  to 
the  recent  directions  of  Rosenau  and  Anderson  39  the  purposes  of  the 
standardization  are  carried  out  as  follows: 

The  unit  of  antitoxin  is  arbitrarily  designated  as  10  times  the 
smallest  amount  of  serum  necessary  to  preserve  the  life  of  a  guinea 
pig  weighing  350  grams  for  96  hours,  when  given  together  with  an 
official  test  dose  of  toxin.  The  test  dose  of  toxin  contains  100  min- 
imal lethal  doses.  And  the  minimal  lethal  dose  is  measured  against 
a  350-gram  guinea  pig. 

37  Yaillard  and  Vincent.     Ann.  Past.,  1891. 

38  Hall.    "Univ.  of  Cal.,  Publ.  in  Path.,"  Vol.  2,  No.  11,  1913. 

39  Rosenau  and  Anderson.     U.  S.  P.  H.  Service  Hyg.  Lab.  Bull.  43,  1908. 


466 


INFECTION    AND    RESISTANCE 


In  carrying  out  the  standardization  the  L+  dose  of 'the  toxin  is 
used,  but,  unlike  diphtheria  standardization,  in  this  case  the  L+ 
dose  means  an  amount  of  toxin  which  will  kill  a  guinea  pig  of  350 
grams  in  four  days,  although  united  with  0.1  unit  of  antitoxin  (it 
must  be  noted  that  the  L+  dose  in  this  case  is  measured  against  one- 
tenth  unit  of  antitoxin  rather  than  against  1  unit,  as  in  the  case  of 
diphtheria. 

In  determining  the  value  of  an  unknown  antitoxin,  mixtures  are 
made,  each  containing  the  L_j_  dose  of  the  toxin  and  varying  quan- 
tities of  antitoxin.  As  in  diphtheria  measurements,  the  various  in- 
jection volumes  are  brought  to  4  c.  c.  with  salt  solution,  and  are  then 
injected  subcutaneously  into  guinea  pigs  of  about  350  grams.  The 
table  given  below  is  taken  from  the  Bulletin  of  Rosenau  and  Ander- 
son. 


Subcutaneous  injection  of 

a  mixture  of 

No.  of 

Weight  of 

guinea 

guinea  pig 

Time  of  death 

pig 

(grams) 

Toxin 
(Test  dose) 

Antitoxin 

(gram) 

(C.  C.) 

1 

360 

0.0006 

0.001 

2  days  4  hours 

2 

350 

0.0006 

0.0015 

4  days  1  hour 

3 

350 

0.0006 

0.002 

Symptoms 

4 

360 

0.0006 

0.0025 

Slight  symptoms 

5 

350 

0.0006 

0.003 

No  symptoms 

In  this  experiment  0.0015  equals  0.10  antitoxin  unit. 


ANTITOXINS  AGAINST  SNAKE  POISONS 
(Antivenin) 

Antitoxins  against  snake  poisons  have  been  produced  by  a  num- 
ber of  different  workers,  but  the  subject  has  been  most  extensively 
studied  by  Calmette.  As  early  as  1887  Sewall  40  succeeded  in  in- 
creasing the  resistance  of  pigeons  to  snake  poison.  Later  Calmette 
and  Physalix  and  Bertrand  independently  succeeded  in  producing 
immunity  in  rabbits  and  guinea  pigs  with  the  poison  of  the  cobra. 
The  serum  of  animals  treated  with  snake  poisons  gradually  acquires 
antitoxin  properties,  but  the  process  of  immunization  is  not  a  simple 
one,  and  considerable  time  is  needed  for  the  immunizations. 

Snake  poisons,  as  we  have  seen,  have  attracted  considerable  atten- 
tion because  of  their  peculiarities  in  being  antigenic  and  yet  differ- 
ing in  heat  resistance  and  a  number  of  other  properties  from  the 

40  Sewall.    Cited  from  Calmette. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         467 

bacterial  toxins.  It  was  with  snake  poisons  that  Calmette  definitely 
showed  that  the  union  of  toxin  and  antitoxin  is  a  true  neutralization 
and  is  not  accompanied  by  the  destruction  of  the  toxin.  These  ex- 
periments, as  we  have  seen,  were  elaborated  later  by  Morgenroth, 
who  succeeded  in  producing  the  snake  poison  HC1  combination.  It 
is  these  poisons  also  that  have  been  the  subject  of  extensive  study  by 
Flexner  and  Noguchi,  by  Kyes,  and  later  by  von  Dungern  and  Coca. 
This  work  has  been  sufficiently  discussed  in  other  places  and  need 
not  occupy  us  here.  The  important  poisonous  snakes  may  be  divided 
into  the  colubridse,  to  which  class  the  cobra  belongs,  and  the  viper- 
.idse,  which  includes  the  ordinary  European  vipers,  the  rattlesnake, 
and  most  of  the  poisonous  snakes  of  North  and  South  America.  Ac- 
cording to  Calmette  the  poison  of  the  cobra  is  much  more  heat-stable 
than  that  of  the  rattlesnake.  Pharmacologically  the  poisons  of  these 
two  main  classes  of  snakes  show  considerable  difference.  In  the  case 
of  the  cobra  there  is  very  little  local  disturbance  and  the  systemic 
symptoms  dominate  the  clinical  picture.  Calmette  describes  the 
cobra  bite  as  being  followed  only  by  a  feeling  of  stiffness  at  the  site 
of  the  bite,  followed  very  soon  by  great  general  weakness,  difficulty 
in  respiration,  slow  heart  action,  and  finally  death  with  unconscious- 
ness. In  the  case  of  the  vipers  the  local  symptoms  are  very  much 
more  marked,  there  being  great  pain  and  swelling  and  apparent  clot- 
ting of  the  blood  about  the  point  of  the  bite,  with  a  rather  slower 
onset  of  systemic  symptoms.  In  a  description  by  Sparr  41  of  a  case 
of  bite  by  Russell's  viper  there  was  almost  immediate  swelling  of  the 
limb  with  a  faint  bluish  tint  around  the  pin-point  puncture,  and 
within  15  minutes  great  weakness,  restlessness,  and  retching.  In 
spite  of  very  active  local  treatment,  within  a  short  time  after  the 
bite,  the  patient  died  within  24  hours  of  asphyxia  and  heart  failure. 

According  to  Calmette  0.0002  gm.  of  cobra  poison  will  kill  a 
guinea  pig;  Noguchi  states  that  0.0005  gm.  of  rattlesnake  venom  will 
kill  a  guinea  pig  of  250  gr.  within  24-30  hours  when  injected  intra- 
peritoneally.  The  snake  poisons  apparently  contain  substances  which 
are  especially  active  upon  nerve  cells  (neurotoxins),  and  hemolysins 
which  act  particularly  upon  the  red  blood  cells.  Flexner  and 
Noguchi  42  also  speak  of  another  poison  which  acts  particularly  upon 
the  endothelium  of  the  blood  vessels  producing  hemorrhages. 

According  to  Calmette  the  antisera  which  are  produced  by  im- 
munization with  cobra  poison  are  most  strongly  potent  against  neuro- 
toxic  poisons  of  the  colubridse  and,  to  a  certain  extent,  against  some 
of  the  poisons  of  the  vipers.  However,  the  action  of  the  cobra  anti- 
toxin against  viper  poison  seems  at  best  to  be  weak.  On  the  other 
hand,  antitoxins  produced  with  rattlesnake  poison  are  not  potent 
against  the  cobra  venom  since,  as  Calmette  states,  the  rattlesnake 

41  Sparr.    Biochem.  Bull.,  Dec.,  1911,  No.  2. 

42  Flexner  and  Noguchi.     Univ.  of  Pa.  Bull.,  Vol.  15,  1902. 


468  INFECTION    AND    RESISTANCE 

poison  contains  hardly  any  neurotoxin.  Antitoxins  may  be  produced 
by  the  gradual  immunization  of  horses,  and  have  been  produced  in 
this  way  by  Calmette  in  the  Pasteur  Institute  of  Lille  for  some 
years.  Calmette  standardizes  his  antitoxin  by  determining  the 
amount  of  serum  which  completely  neutralizes  in  vitro  0,0001  gm. 
of  the  poison  as  tested  upon  white  light.  He  also  determines  the 
protective  power  by  injecting  a  rabbit  with  2  c.  c.  of  the  serum  and 
two  hours  later  gives  1  gm.  of  the  poison. 

JSToguchi  has  studied  rattlesnake  poison  particularly  and  suc- 
ceeded in  preparing  a  strong  antitoxin  by  the  gradual  immunization 
of  a  goat.  Great  difficulty  has  always  been  experienced  in  attempts 
at  immunization  with  rattlesnake  poison  because  of  the  very  violent 
local  injury  produced  by  injections  of  the  venom.  The  potency  of 
the  serum  produced  by  him  was  such  that  2%  c.  c.  of  goat  serum 
protected  guinea  pigs  against  12  times  the  fatal  dose  of  rattlesnake 
poison  if  given  at  the  same  time.  If  the  antivenin  was  given  one 
hour  later,  5  times  the  amount  of  serum  had  to  be  given. 


PASSIVE  IMMUNIZATION  IN  DISEASES  CAUSED  BY  BAC- 
TERIA WHICH  DO  NOT  FORM  SOLUBLE  TOXINS 

As  we  have  stated  in  another  place  the  greatest  therapeutic  suc- 
cesses with  passive  immunization  have  been  achieved  in  bacterial 
diseases  in  which  the  malady  is  essentially  a  toxemia  due  to  a  soluble 
toxin.  In  such  cases  the  serum  of  actively  immunized  animals  con- 
tains specific  antitoxins  by  virtue  of  which  the  toxins  circulating  in 
the  blood  of  the  patient  are  directly  neutralized,  quantity  for  quan- 
tity, with  consequent  therapeutic  benefit.  In  the  case  of  bacteria  in 
which  no  toxins  are  formed,  the  immunization  of  an  animal  is  not 
followed  by  the  formation  of  any  poison-neutralizing  principle. 
Here  the  injection  of  bacteria,  dead  or  alive,  or  the  invasion  of  the 
bacteria  in  the  course  of  spontaneous  disease,  is  followed  by  the 
formation  of  specific  antibacterial  substances,  lytic,  opsonic,  agglu- 
tinating, or  precipitating  bodies,  the  nature  of  which  we  have  dis- 
cussed in  other  chapters.  The  toxemia  which  occurs  in  such  cases 
is  due  as  we  have  seen  to  derivatives  of  the  bacterial  protein  which 
by  some  observers  are  regarded  as  preformed  endocellular  poisons 
liberated  by  the  lytic  action  of  the  serum,  and  by  others  as  split 
products  of  the  bacterial  protein,  non-existent  until  the  bacterial  cell 
has  been  acted  upon  by  the  serum  components  and  destroyed.  How- 
ever this  may  be,  the  recovery  from  diseases  of  this  nature  is  accom- 
plished by  bacterial  destruction ;  this  may  be  direct,  by  the  bacteri- 
cidal action  of  the  serum,  or  indirect  by  opsonic  properties  which 
induce  phagocytosis.  The  poisons  which  are  liberated  from  the  bac- 
terial bodies,  if  free,  can  do  their  injury,  and  no  neutralizing  sub- 


THERAPEUTIC    IMMUNIZATION    IN    MAN         469 

stance  is  formed  in  the  body  fluids  to  prevent  their  action  as  far  as 
we  know.  Immunity  in  such  cases,  then,  is  not  an  antitoxic  immu- 
nity in  any  sense  of  the  word ;  it  is  rather  an  antibacterial  immunity 
in  which  the  disease  is  prevented  or  cured  only  when  complete  de- 
struction of  the  bacteria  has  taken  place.  If  an  animal  or  a  human 
being  is  prophylactically  immunized  against  diseases  of  this  kind 
(typhoid,  cholera,  etc.),  it  is  easy  to  see  that  an  increased  presence 
of  antibacterial  substances,  bactericidal  or  opsonic,  in  the  circulation 
would  serve  efficiently  and  rapidly  in  disposing  of  the  small  numbers 
of  invading  micro-organisms  which  ordinarily  enter  the  body  in  spon- 
taneous infections.  And,  indeed,  experience  has  shown  that  prophy- 
lactic immunization  can  be  successfully  carried  out  in  the  case  of 
cholera,  typhoid  fever,  plague,  and  other  diseases  which  are  suffi- 
ciently prevalent  endemically  or  epidemically  to  justify  prophylaxis 
on  an  extensive  scale. 

However,  when  in  diseases  of  this  kind  the  body  is  already  exten- 
sively infected  and  has  begun,  as  is  usually  the  case,  to  respond  spon- 
taneously with  the  formation  of  specific  antibodies,  it  has  been  a 
matter  of  doubt  whether  or  not  passive  immunization,  that  is,  the 
introduction  of  specific  antibodies  in  the  form  of  the  serum  of  a 
highly  immunized  animal,  is  therapeutically  of  value.  Indeed,  it 
has  been  feared  that  the  use  of  such  sera  may  even  be  harmful  in 
that  the  sudden  introduction  of  large  amounts  of  bactericidal  sub- 
stances might  lead  to  a  sudden  liberation  of  large  quantities  of  poi- 
sonous products  and  consequent  rapid  toxemia. 

The  conditions  in  such  cases  are  exceedingly  complex  and  many 
gaps  exist  in  our  knowledge  concerning  them.  The  bacteria  when 
invading  the  body,  immediately  enter  into  conflict  with  the  protective 
forces,  as  we  have  stated  in  the  chapter  on  Infection.  If  a  consider- 
able degree  of  resistance  exists,  let  us  say  as  the  result  of  preceding 
immunization  or  a  recent  attack  of  the  disease,  there  is  a  rapid 
destruction  of  the  bacteria,  probably  by  active  phagocytosis.  It  has 
been  shown  by  Bordet  in  the  case  of  cholera  and  more  recently  by 
Gay  with  typhoid,  that  injection  of  the  organisms  into  immunized 
animals  is  followed  by  prompt  and  high  leukocytosis,  whereas  sim- 
ilar injections  into  normal  animals  usually  induce  a  temporary  leuko- 
penia.  When  the  invaded  animal  is  not  particularly  resistant  the 
bacteria  may  accumulate  and,  as  in  the  case  of  pneuraococci  and 
streptococci,  develop  phagocytosis-resisting  properties  (capsule  forma- 
tion, etc.)  ;  or,  as  in  the  case  of  typhoid  bacilli,  there  may  be  an 
immediate  liberation  of  toxic  substances  (anaphylatoxins)  by  reac- 
tion between  bacterial  cell  and  blood  plasma,  which  can  induce  leuko- 
penia,  and  by  this  means  the  organisms  may  be  protected  from 
phagocytic  destruction.  Experience  with  curative  sera  in  all  of  the 
conditions  of  this  class  has  yielded  promising  results  only  when  the 
cases  have  been  treated  with  the  sera  at  early  stages  of  the  disease, 


470  INFECTION    AND    RESISTANCE 

either  when  the  invading  germ  was  still  localized  or,  at  least,  wnen 
the  septicemic  condition  was  not  yet  thoroughly  established.  It  may 
be  that  the  doses  heretofore  given  have  been  insufficient,  and  indeed 
recent  experiences  with  pneumonia  seem  to  indicate  that  this  may 
have  been,  in  part,  the  cause  of  earlier  failures.  Yet  in  pneumonia 
the  septicemia  probably  does  not  represent  the  firm  establishment  of  a 
foothold  by  the  pneumococcus  in  the  circulation  but  rather  a  con- 
tinuous discharge  of  new  organisms  into  the  blood  from  the  localized 
lesion  in  the  lung. 

It  is  our  own  opinion  moreover  that  septicemia  as  usually  ob- 
served clinically  represents  in  most  cases  exactly  this  condition,  that 
is",  a  more  or  less  continuous  discharge  of  the  bacteria  into  the  blood 
from  some  active  focus  with  a  continuous  destruction  of  the  organ- 
isms after  they  have  entered  the  blood  stream.  It  is  only  when  the 
resistance  of  the  body  is  overwhelmed,  in  the  later  stages  of  the 
disease,  that  the  bacteria  can  continue  to  grow  and  develop  in  the 
circulation,  and  this  stage  probably  does  not  occur  until  death  is 
imminent.  In  such  septicemic  diseases  as  streptococcus  infection, 
typhoid  fever,  plague,  anthrax,  and  many  others  the  presence  of  the 
bacteria  in  the  blood  at  the  time  when  the  patient  is  still  in  a  condi- 
tion of  powerful  resistance  probably  means  that  the  bacteria  are 
being  supplied  to  the  blood  from  the  local  lesions.  There  is  prob- 
ably just  such  a  continuous  discharge  of  bacteria  from  the  focus  into 
the  blood  with  active  destruction  after  the  bacteria  have  entered  the 
circulation.  This  seems  especially  probable  from  the  fact  that  in 
many  of  these  diseases  the  protective  antibodies,  bactericidal  and 
opsonic,  can  often  be  demonstrated  in  the  blood  serum  in  quantities 
higher  than  normal  at  the  very  time  when  blood  culture  yields  posi- 
tive results.  In  typhoid  fever,  of  course,  it  is  well  known  that  bac- 
tericidal titres  of  over  1-50,000  are  often  present  while  the  patient 
may  still  be  very  sick,  and  in  the  more  chronic  streptococcus  condi- 
tions with  malignant  endocarditis  we  have  often  seen  that  opsonic 
properties  on  the  part  of  the  patient's  serum  against  the  very  organ- 
ism invading  him  are  considerably  higher  than  normal.  We  take 
this  to  mean  that  the  injection  of  immune  sera  would  simply  aid  in 
more  rapidly  freeing  the  blood  stream  of  the  bacteria,  the  cure  of' 
the  disease,  however,  involving  a  destruction  of  the  focus.  This,  of 
course,  is  not  possible  merely  by  the  injection  of  the  serum.  When, 
as  in  some  cases  of  streptococcus  infection,  the  focus  can  be  surgically 
reached,  the  septicemia  will  often  disappear  and  cure  result,  as  we 
have  ourselves  had  the  opportunity  to  observe.  When  the  focus 
•cannot  be  reached  surgically,  it  may  nevertheless  be  a  wise  procedure 
to  inject  considerable  amounts  of  immune  serum,  for,  by  keeping  the 
Wood  stream  free  of  bacteria,  the  case  may  be  influenced  favorably. 
Pneumonia  is  an  example  of  this.  Former  failures  have  recently 
been  turned  into  partial  success  by  the  work  of  Neufeld  and  of  Cole 


THERAPEUTIC    IMMUNIZATION    IN    MAN         471 

merely  by  the  use  of  larger  quantities  of  immune  sera  essentially 
similar  to  sera  used  at  previous  times,  and  Cole  attributes  the  appar- 
ently favorable  results  to  the  fact  that  the  blood  stream  can  be  cleared 
of  bacteria  although  the  focus  cannot  itself  be  affected. 

Cure  of  such  diseases,  therefore,  by  serum  treatment  can  hardly 
be  expected.  Favorable  influence  of  the  disease  by  energetic  serum 
treatment  may,  however,  be  hoped  for. 

In  discussing  this  subject  it  must  not  be  forgotten,  however,  that 
in  most  of  the  diseases  which  we  have  classified,  on  the  basis  of  pre- 
vailing opinions,  as  caused  by  bacteria  that  do  not  form  true  toxins, 
the  formation  of  such  poisons  has  been  claimed  by  a  number  of  care- 
ful and  eminent  observers.  In  the  case  of  the  typhoid  bacillus,  espe- 
cially, Chantemesse,  Kraus  and  Stenitzer,  and  others  have  claimed 
the  existence  of  a  true  toxin  and  a  consequent  antitoxin  in  immune 
sera.  Similar  claims  have  been  made  for  the  cholera  spirillum  by 
Kraus  and  Doerr,  for  the  streptococcus  by  Marmorek,  and  for  the 
plague  bacillus  by  Markl  and  Rowland.  Since  these  claims  have 
been  made  on  the  basis  of  extensive  experimentation  by  competent 
men  the  question  must  be  left  open,  and  the  possibility  of  antitoxic 
properties  on  the  part  of  the  sera  cannot  be  completely  ignored. 
Since  in  most  cases,  however,  the  poison-neutralizing  properties  of 
the  immune  sera  in  this  disease  have  not  exceeded  more  than  1  to  2 
multiples  of  the  M  L  D  of  the  bacterial  poisons,  it  does  not  seem 
impossible  that  the  apparent  antitoxic  properties  may  have  repre- 
sented merely  an  acquired  tolerance  to  anaphylatoxic  poisons  of 
which  we  have  spoken  in  another  place. 


SERUM  TREATMENT  IN  EPIDEMIC  CEREBROSPINAL,  MENINGITIS 

Serious  attempts  to  produce  curative  sera  -against  the  epidemic 
form  of  cerebrospinal  meningitis  were  not  made  until  1906  and 
1907,  when  this  disease  appeared  epidemically  chiefly  in  Europe, 
where  it  appeared  most  severely  in  Eastern  Germany,  and  in  the 
Eastern  United  States. 

In  1906  Kolle  and  Wassermann  immunized  three  horses  with 
meningococci,  using  for  immunization  purposes  the  dead  organisms 
followed  by  living  cultures  and  cultures  shaken  up  in  distilled  water, 
the  so-called  artificial  aggressins  of  Wassermann  and  Citron.  They 
obtained  sera  of  considerable  potency  when  measured  against  menin- 
gococcus  cultures,  and  suggested  standardizing  the  sera  by  comple- 
ment fixation.  They  did  not  at  this  time  treat  human  beings,  but  sug- 
gested the  use  of  the  serum  subcutaneously  and  intravenously  in 
meningitis  cases.  Very  soon  after  the  publication  of  the  work  of 
Kolle  and  Wassermann  Jochmann  43  also  produced  an  antimeningo- 
43  Jochmann.  Deutsche  med.  Woch.,  1906,  Vol.  32,  p.  788. 


472  INFECTION    AND    RESISTANCE 

coccus  serum  by  immunizing  horses  with  proved  meningococcus  cul- 
tures, in  his  cases  making  a  polyvalent  serum  by  the  use  of  many 
different  strains  of  the  organism.  The  sera  which  he  obtained  were 
highly  agglutinating,  somewhat  bactericidal,  and,  according  to  him, 
not  antitoxic.  He  first  succeeded  in  immunizing  guinea  pigs  against 
meningococci  by  injecting  the  serum  20  hours  before  infecting  the 
animals.  He  also  treated  40  cases  of  meningitis  in  man  and  ob- 
tained encouraging  results  in  cases  treated  before  the  development  of 
hydrocephalus.  Believing  that  possibly  intraspinous  injection  of  the 
serum  might  offer  advantages,  he  first  determined  by  experiments 
upon  the  dead  body  that  the  injection  of  methylene-blue  intra- 
spinously  passed  from  the  point  of  injection  in  the  lumbar  regions  as 
far  up  as  the  olfactory  nerves.  After  having  determined  this  he 
treated  17  cases  by  tapping  the  spinal  canal,  taking  out  30  to  50  c.  c. 
of  spinal  fluid  and  then  injecting  about  20  c.  c.  of  the  serum.  Of 
these  17  cases  only  5  died,  and  Jochmann  expresses  himself  opti- 
mistically in  consequence. 

Meanwhile  Flexner  44  had  been  working  upon  the  same  subject, 
laying  a  rather  more  thorough  basis  for  therapy  in  careful  animal 
experimentation.  He  produced  the  typical  disease  in  monkeys  by 
intraspinous  inoculation  of  the  meningococci  and  then  saved  the 
animals  from  death  by  following  the  infection  with  the  injection  of 
serum  intraspinously  six  hours  later.  In  his  earlier  articles  he  ex- 
presses himself  with  much  conservatism,  but  his  studies  were  con- 
tinued and  extensive  opportunity  for  testing  the  serum  which  he  then 
produced,  together  with  Jobling,45  was  offered  by  the  continuance  of 
the  epidemic  throughout  the  United  States. 

The  results  with  the  serum  produced  at  the  Rockefeller  Institute 
have  since  then  proved  to  be  uniformly  favorable.  The  method  of 
intraspinous  inoculation  of  the  serum  after  the  removal  of  some  of 
the  spinal  fluid  was  the  method  finally  adopted  by  Flexner  as  most 
favorable,  and  this  is  the  method  in  current  use  to-day.  In  1908 
Flexner  and  Jobling  reported  upon  47  cases  treated  with  the  anti- 
serum  of  which  34  recovered.  Of  12  additional  cases  reported  in  an 
addendum  only  4  died.  In  the  most  recent  summary  by  Flex- 
ner 46  records  of  1,294  cases  that  have  been  treated  with  the  serum 
prepared  at  the  Rockefeller  Institute  are  analyzed.  Of  this  num- 
ber, unselected  and  treated  in  many  different  parts  of  the  world, 
69.1  per  cent,  recovered.  It  is  of  course  very  difficult  to  obtain 
exact  comparative  data  on  the  efficiency  of  any  method  of  treatment 
in  a  disease  as  irregular  in  its  clinical  manifestations  as  epidemic 
meningitis,  especially  since  the  mortality  attending  upon  different 

44  Flexner.    J.  Exp.  Med.,  Vol.  9,  1907,  and  /.  A.  M.  A.,  1906,  Vol.  47,  p. 
560. 

45  Flexner  and  Jobling.     J.  Exp.  Med.,  Vol.  10,  1908. 

46  Flexner.     J.  Exp.  Med.,  Vol.  17,  1913. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         473 

epidemics  is  subject  to  great  variations.  For  this  reason  we  can 
draw  conclusions  only  from  a  large  statistical  material.  However, 
we  know  that  the  average  mortality  of  epidemic  meningitis  before 
the  introduction  of  specific  therapy  ranged  certainly  higher  than  65 
per  cent.,  and  in  carefully  studied  epidemics  usually  between  TO  and 
80  per  cent.  The  statistics  of  Flexner  showing  a  mortality  hardly 
exceeding  30  per  cent,  in  unselected  cases  unquestionably  marks  a 
wonderful  therapeutic  triumph.  It  must  be  remembered  in  consider- 
ing the  benefits  of  this  serum  that  in  unselected  cases  there  must  be 
many  in  whom  the  disease  has  produced  marked  anatomical  changes 
in  the  central  nervous  system  before  the  serum  is  used.  It  is  well 
known,  of  course,  that  the  later  manifestations  of  this  disease,  which 
often  lead  to  death  with  hydrocephalus,  asthenia,  and  malnutrition, 
are  the  remote  results  of  the  anatomical  injuries  produced  by  the  in- 
flammatory reactions  accompanying  the  earlier  manifestations  of  the 
acute  infection.  These  conditions  of  course  cannot  be  expected  to 
yield  to. serum  treatment.  It  must  be  assumed,  therefore,  that  were 
we  able  to  obtain  statistics  of  cases  diagnosed  and  treated  soon  after 
the  onset  the  figures  would  be  even  more  favorable  than  those  stated 
above. 

The  action  of  the  serum  seems  very  largely  to  be  an  opsonic  one, 
in  that,  under  the  influences  of  serum,  a  powerful  phagocytosis  of 
the  meningococci  takes  place.  It  is  also  possible  that  to  a  certain 
extent  bactericidal  action  participates,  in  that  the  injection  of  the 
serum  into  the  closed  space  may  give  rise  to  a  sort  of  intraspinous 
Pfeiffer  reaction  with  energetic  ingestion  of  the  bacteria  by  leu- 
kocytes. 

The  standardization  of  the  antimeningococcus  serum  has  been 
worked  out  particularly  by  Jobling.47  After  attempting  to  stand- 
ardize the  sera  by  their  protective  power  against  meningococcus  in- 
fection in  animals  and  by  complement  fixation,  as  suggested  by 
Kolle  and  Wassermann,  Jobling  has  come  to  the  conclusion  that 
neither  of  these  methods  is  sufficiently  regular,  and  that  the  most 
suitable  procedure  is  a  standardization  by  opsonin  determination. 
The  method  as  worked  out  by  him  depends  upon  determining  the 
highest  dilution  of  the  immune  serum  at  which  opsonic  action 
against  the  meningococcus  is  still  evident.  He  suggests  as  a  definite 
and  suitable  standard  of  strength  opsonic  activity  at  a  dilution  of 
1-5,000  of  the  antiserum. 

SERUM  TREATMENT  IN  STREPTOCOCCUS  INFECTIONS 

The  attempts  to  produce  powerful  immune  sera  against  strepto- 
cocci date  back  to  the  earliest  days  of  immunology.  That  the  sub- 
ject is  a  particularly  difficult  one  follows  from  the  great  confusion 

47  Jobling.    /.  Exp.  Med.}  Vol.  11,  1909. 


474  INFECTION    AND    RESISTANCE 

which  has  prevailed,  and,  to  a  great  extent,  still  prevails,  regarding 
the  classification  of  the  streptococci  and  their  interrelationship. 
There  are  apparently  a  large  number  of  different  strains  of  strepto- 
cocci which  vary  from  each  other,  not  only  culturally,  but  also  in 
regard  to  agglutination  and  bactericidal  reactions.  For  this  reason 
it  is  not  at  all  a  foregone  conclusion  that  a  serum  prepared  by  the 
immunization  of  an  animal  with  a  streptococcus  of  one  type  will 
have  any  protective  action  against  other  strains.  The  subject  has 
been  still  more  complicated  recently  by  the  discovery  of  Rosenow  48 
that  the  various  types  of  streptococci  (viridans,  hemolyticus,  etc.) 
are  not  constant  in  their  properties,  but  may  be  artificially  trans- 
formed one  into  the  other,  and  that  even  mutation  of  true  pneumo- 
cocci  into  true  streptococci  may  take  place.  Most  important  in  this 
connection  is  the  observation  that  a  pneumococcus  sent  to  Rosenow 
was  altered  by  him  by  special  methods  of  cultivation  in  such  a  way 
that  not  only  its  morphological  and  cultural  properties  were  changed, 
but  also  its  agglutination  reactions.  These  observations  are  of  the 
utmost  importance  in  connection  with  attempts  at  producing  specific 
sera  which  can  be  utilized  therapeutically.  In  all  cases,  therefore, 
in  which  streptococcus  immune  serum  is  at  all  used  it  must  be  re- 
membered that  the  disease  produced  in  human  beings  by  organisms 
classified  among  the  streptococci  are  by  no  means  necessarily  closely 
related  in  biological  reactions,  and  the  same  immune  serum  may  be 
extremely  potent  in  one  case  and  entirely  useless  in  another. 

That  animals  could  be  successfully  immunized  against  strepto- 
cocci was  shown  early  in  the  history  of  investigations  in  immunity 
by  a  number  of  workers,  notably  Roger,  Behring,  von  Lingelsheim, 
and  Mironoff.  The  first  extensive  attempts  to  produce  a  curative 
serum  for  use  in  passively  immunizing  human  beings  were  made  by 
Marmorek  49  at  the  Pasteur  Institute  in  1895.  The  basic  idea  from 
which  Marmorek  worked  was  the  similarity  of  all  the  streptococci 
producing  disease  in  human  beings.  He  also  believed  that  the  most 
powerful  serum  could  be  produced  with  cultures  whose  virulence 
had  been  greatly  enhanced  by  animal  passages.  When  such  cultures 
were  grown  on  mixtures  of  human  serum  and  broth  he  asserted 
furthermore  that  soluble  poisons  were  produced  which  could  be  ob- 
tained by  filtration  of  the  culture  fluids.  For  these  reasons  he  im- 
munized horses  with  cultures  rendered  highly  virulent  by  very 
gradual  injections  first  of  dead  then  of  living  organisms,  finally 
injecting  also  considerable  quantities  of  culture  filtrates. 

Testing  these  sera  upon  animals,  he  was  successful  in  protecting 
against  streptococcus  infection  when  the  serum  was  administered  12 
to  18  hours  before  the  bacteria  were  injected.  He  expressed  the 
opinion  that  the  serum  was  antitoxic  as  well  as  antibacterial,  In 

48  Rosenow.     Journ.  A.  M.  A.,  Feb.,  1914. 

49  Marmorek.     Ann.  Past.,  Vol.  9,  1895. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         475 

his  earliest  reports  the  results  of  the  treatment  of  413  cases  of  ery- 
sipelas leave  one  very  much  in  doubt  as  to  the  value  of  the  serum  since 
the  difference  in  mortality  between  the  treated  and  the  untreated 
cases  is  less  than  2  per  cent.  However,  an  analysis  of  the  individual 
cases  makes  the  serum  treatment  appear  more  favorable.  He  re- 
ported good  results  also  in  7  cases  of  puerperal  septicemia  and  in 
scarlatinal  angina.  Later  observers,  notably  Lenhartz,50  Baginsky,51 
and  many  others,  have  not  been  able  to  confirm  the  favorable  results 
reported  by  Marmorek,  and  it  may  be  stated  that  at  the  present  day 
the  value  of  Marmorek's  serum  is  very  much  in  question.  Anti- 
streptococcus  sera  have  also  been  produced  by  Aronson  52  and  Tavel, 
Van  de  Velde,  Menzer,53  Moser,  54  and  some  others.  Aronson  at 
first  worked  from  the  idea  which  Marmorek  also  had  used  that  there 
was  a  close  relationship  between  the  various  streptococci  pathogenic 
for  man.  He  adopted  the  opinion  first  developed  by  Denys  55  and 
Van  de  Velde  that  many  different  strains  should  be  used  for  im- 
munization in  order  to  allow  for  possible  difference  in  the  character- 
istics of  the  pathogenic  streptococci.  This  principle  of  the  necessity 
for  the  production  of  polyvalent  sera  was  also  emphasized  strongly  by 
Tavel,  who  based  his  opinion  on  careful  agglutination  tests,  and  by 
Menzer  and  Moser. 

That  the  action  of  the  antistreptococcus  sera,  however  produced, 
is  very  largely  due  to  its  opsonic  properties  has  been  shown  by 
Bordet,56  by  Meier  and  Michaelis,  and  a  number  of  other  workers. 
If  there  is  any  bactericidal  power  it  is  probably  relatively  slight. 

It  would  be  quite  impossible  to  attempt  in  this  place  to  analyze 
the  large  number  of  streptococcus  infections  of  man  which  have  been 
treated  with  one  or  the  other  antistreptococcus  sera.  Those  men- 
tioned, moreover,  do  not  by  any  means  include  all  the  sera  which 
have  been  produced  and  marketed  for  this  purpose.  In  general  we 
may  say  that  here,  as  well  as  in  the  cases  of  other  sera  in  which  no 
antitoxic  action  is  evident,  beneficial  results  have  been  obtained 
chiefly  in  cases  in  which  the  streptococcus  infection  has  been  localized 
and  treated  early  after  its  inception.  In  generalized  or  advanced 
cases  it  cannot  be  said  that  the  results  are  encouraging.  Even  in 
animals,  in  which  experimental  conditions  can  be  so  much  more 
thoroughly  controlled,  the  protective  action  of  even  the  strongest 
sera  is  evident  only  if  the  serum  is  administered  either  before  in- 

50  Lenhartz.     "Die  Septischen  Erkrankunsren  Holder,"  Wien,  1903. 

51  Baginsky.    Berl  kl.  Woch.,  1896,  p.  340. 

52  Aronson.     Berl.  kl.   Woch.,  Vol.  39,  1902;  Deutsche  med.  Woch.,  29, 
1903. 

53  Menzer.     Berl  kl.  Woch.,  1902,  and  Munch,  med.  Woch.,  1903. 

54  Moser.     Wien.  kl  Woch.,  1902. 

55  Denys.     Bull   de   I'Acad.   Beige,   1896,   cited   from   Schwoner  K.   and 
L.  H.,  Vol.  2. 

56  Bordet.     Ann.  de  Vlnst.  Past.,  1897. 


476  INFECTION    AND    RESISTANCE 

fection  or  within  a  very  definite  period  after  inoculation.  The 
standardization  of  streptococcus  sera  may  be  accomplished  by  de- 
termining its  protective  value  for  animals  when  injected  18  to  20 
hours  before  infection.  When  the  sera  are  produced  by  immuniza- 
tion with  streptococci  obtained  from  the  human  body  and  without 
pathogenicity  for  animals  the  standardization  is  of  course  unsatis- 
factory. 

SERUM  TREATMENT  IN  PNEUMONIA 

Attempts  to  work  out  a  therapeutically  valid  method  of  passive 
immunization  in  pneumonia  have  been  many  and  date  from  the  very 
beginning  of  the  discovery  that  pneumonia  was  a  bacterial  infection. 
Sera  have  even  been  marketed  and  used,  but  until  recently  no  very 
encouraging  results  were  obtained.  Recent  studies  have  revealed 
that  in  pneumonia  the  serum  of  convalescents  contains  practically  no 
bactericidal  properties  for  the  pneumococcus,  and  that  the  protective 
powers  of  such  serum  depend  upon  the  presence  of  immune  opsonins 
or  bacteriotropins,  by  means  of  which  the  pneumococci  are  ren- 
dered amenable  to  phagocytosis.  Virulent  pneumococci  are  not  as  a 
rule  phagocytable  in  the  presence  of  normal  serum.  However,  in 
the  presence  of  immune  serum  powerful  phagocytic  action  can  be 
observed. 

Neufeld  has  studied  the  conditions  of  pneumococcus  immunity 
most  thoroughly  in  recent  years.  The  most  important  advance  from 
a  practical  point  of  view  was  a  discovery  made  by  him,  with  Han- 
del,57 in  1909.  They  determined  that  there  was  a  definite  difference 
between  various  pneumococci  in  their  reactions  to  immune  serum ;  in 
other  words,  pneumococci  could  be  grouped  into  various  serological 
types.  The  serum  produced  with  organisms  of  one  type  did  not  pro- 
tect against  infection  with  other  strains.  In  consequence  they  called 
attention  to  the  importance  of  determining  the  type  of  pneumococcus 
which  causes  the  individual  pneumonia  so.  that  the  corresponding 
immune  serum  might  be  used.  They  produced  a  highly  potent  anti- 
pneumococcus  serum  by  the  injection  of  horses  and  donkeys  with 
highly  virulent  pneumococci  grown  on  fluid  cultures,  then  deter- 
mined the  high  protective  power  of  this  serum  upon  animals  and 
used  it  in  the  treatment  of  patients  by  intravenous  injection.  Their 
results  were  exceedingly  encouraging.  In  reporting  their  results 
Neufeld  and  -Handel  state  that  considerable  doses  must  be  given. 
They  call  attention  to  the  fact,  revealed  by  their  animal  experiments, 
that  moderate  amounts  do  not,  as  in  the  case  of  diphtheria  serum, 
exert  a  correspondingly  slight  amount  of  beneficial  action,  but  that 
in  the  case  of  the  pneumonia  serum  amounts  smaller  than  a  certain 

57Neufeld  and  Handel.  Zeitschr.  f.  1mm.,  Vol.  3,  1909,  and  Arb.  am 
(Jem  kais.  Gesundh.  Amt.3  Vol.  34,  1910. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         477 

active  minimum  seem  to  exert  absolutely  no  beneficial  action.  This 
is  a  fact  which  later  was  also  determined  by  Dochez. 

In  confirmation  of  the  work  of  Neufeld  and  Handel  58  Dochez 
and  Gillespie  59  have  also  been  able  to  determine  that  there  are  at 
least  two  distinctive  groups  of  pneumococci  which  differ  from  each 
other  as  far  as  agglutination  and  serum  protection  experiments  are 
concerned.  In  addition  to  these  two  fixed  types  they  separate  as  a 
third  group  the  streptococcus  or  Pneumo  coccus  mucosus  and  a  fourth 
heterogeneous  group  which  seems  to  fit  in  with  none  of  the  others  as 
far  as  serum  reactions  can  determine. 

Cole,60  therefore,  adopts  the  reasoning  of  ISTeufeld  in  that  he 
advises  the  determination  of  the  type  of  pneumococcus  present  in 
cases  of  pneumonia  as  a  guide  to  the  type  of  antiserum  to  be  used. 
The  type  of  organism  is  determined  as  soon  as  a  case  comes  under 
observation  and  50  to  100  c.  c.  of  the  homologous  antiserum  is  in- 
jected intravenously.  The  result  in  the  few  cases  so  far  treated  by 
Cole  and  his  associates  has  been  encouraging. 

Protective  substances,  according  to  Dochez,  appear  in  the  serum 
of  treated  cases  very  shortly  after  the  administration  of  the  serum, 
whereas  in  untreated  lobar  pneumonia  such  protective  substances 
usually  do  not  appear  until  after  the  crisis.  Apparently  Cole  believes 
that  the  great  value  of  passive  immunization  of  this  kind  in  pneu- 
monia lies  in  the  fact  that  the  bacteriemia  shown  to  prevail  in  prob- 
ably all  cases  of  lobar  pneumonia  is  either  cured  or  improved  by  the 
treatment,  converting  the  disease,  which  is  by  nature,  at  least  for  a 
time,  a  septicemia,  into  a  localized  pulmonary  infection.  Experi- 
ence with  antipneumococcus  serum  so  far  has  been  too  limited  to 
warrant  final  judgment  as  to  its  permanent  place  among  therapeutic 
agencies. 

THE  SEBUM  TREATMENT  OF  TYPHOID  FEVER 

The  first  extensive  attempts  to  treat  typhoid  fever  by  passive  im- 
munization with  the  serum  of  treated  animals  were  made  by  Chante- 
messe,  who  immunized  horses  with  filtrates  of  typhoid  cultures  sub- 
cutaneously,  and  with  emulsions  of  virulent  bacilli  intravenously. 
Chantemesse  believed  that  the  serum  of  horses  which  had  been 
treated  in  this  way  for  very  long  periods  possessed,  not  only  bacteri- 
cidal action,  but  stimulated  phagocytosis,  and  possessed  a  certain 
limited  amount  of  neutralizing  power  against  the  toxic  properties  of 
the  typhoid  filtrates.  At  the  International  Congress  of  Hygiene  in 
Berlin  in  1907  Chantemesse 61  reported  upon  a  thousand  cases 

58  Neufeld  and  Handel.    Arb.  aus  dem  kais.  Gesundh.  Ami.,  Vol.  34,  1910. 

59  Dochez  and  Gillespie.     Jour.  A.  M.  A.,  Vol.  61,  p.  727,  1913. 

60  Cole.    Jour.  A.  M.  A.,  Vol.  61,  p.  663,  1913. 

61  Chamtemesse.     International  Congress  of  Hygiene,  Berl.,  Sept.,  1907 ; 
Ref.  Bull,  de  I'Inst.  Pasteur,  Vol.  5,  1907,  p.  931. 


478  INFECTION    AND    RESISTANCE 

treated  with  his  serum.  Of  this  number  43  only  died,  whereas  the 
average  mortality  during  the  same  six  years  at  the  Paris  hospitals 
was  17  per  cent.  The  injection  of  the  serum  he  claimed  very  mark- 
edly improved  the  condition  of  patients  in  that,  after  a  preliminary 
period  of  no  apparent  change  lasting  from  several  hours  to  5  or  6 
days,  the  temperature  goes  down  and  the  general  condition  of  the 
patient  changes  considerably  for  the  better.  He  noticed  very  few 
complications  in  these  cases,  and  intestinal  hemorrhage  occurred 
four  times  only. 

A  remarkable  feature  of  Chantemesse's  treatment  is  that  he  in- 
jected into  the  patients  a  few  drops  only  of  the  serum,  and  rarely 
made  a  second  injection,  facts  which  alone  tend  to  persuade  one  that 
his  apparent  therapeutic  success  was  a  fortunate  accident. 

The  opinion  originally  expressed  by  Chantemesse  that  the  serum 
of  horses  vigorously  treated  with  typhoid  bacilli  possesses  in  addition 
to  its  bactericidal  and  opsonic  powers  definite  antitoxic  properties 
recurs  again  in  the  work  of  a  number  of  investigators.  Besredka  62 
prepared  a  serum  by  the  intravenous  injection  of  typhoid  cultures 
heated  to  60°  C.,  continuing  the  immunization  for  6  months.  He 
claims  that  this  serum  possesses  what  he  designates  as  "anti-endo- 
toxic"  properties.  A  dry  extract  of  typhoid  bacilli  which  in  dose  of 
0.01  gram  killed  a  guinea  pig  of  300  grams  regularly  became  innocu- 
ous when  mixed  with  small  quantities  of  this  horse  serum.  One  c.  c. 
of  the  horse  serum  neutralized  often  as  much  as  two  fatal  doses  of  the 
serum,  but  it  is  important  theoretically  to  recognize  that  Besredka 
states  particularly  that  even  an  increase  of  the  quantity  of  serum 
never  neutralized  more  than  two  fatal  doses.  This  is  particularly 
important  in  connection  with  the  more  recent  studies  on  toxic  split 
proteins  by  Vaughan,  and  on  anaphylatoxins  by  Bessau  and  by  Zins- 
ser  and  Dwyer,  in  which  it  has  been  shown  that  an  animal  acquires 
a  tolerance  against  the  toxic  substances  produced  from  bacterial  and 
other  proteins  which,  however,  never  exceeds  one  or  two  multiples 
of  the  minimum  lethal  dose.  This  fact  alone  would  militate  against 
considering  the  serum  of  Besredka  in  any  way  antitoxic  in  the  sense 
in  which  the  word  is  used  concerning  diphtheria  and  tetanus  anti- 
toxins where  neutralization  of  poison  follows  roughly  the  law  of 
multiples.  Besredka's  anti-endotoxic  sera  has  recently  been  very 
thoroughly  investigated  by  Pfeiffer  and  Bessau.63  These  investi- 
gators have  found  that  Besredka's  serum  exerted  a  very  definite 
beneficial  influence  upon  typhoid  infection  in  guinea  pigs  if  injected 
at  the  same  time  with  the  bacilli.  In  their  experiments  it  also  pro- 
tected somewhat  against  the  toxic  properties  of  substances  derived 
from  the  typhoid  bacillus,  and  Pfeiffer  and  Bessau  did  not  believe 
that  this  was  due  to  a  true  antitoxic  action,  nor  that  the  serum  was 

62  Besredka.     Ann.  de  I'Inst.  Pasteur,  19,  1905,  and  20,  1906. 

63  Pfeiffer  and  Bessau.     Centralbl  f.  Bakt.,  Vol.  56,  1910. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         479 

superior  in  this  respect  to  the  ordinary  bactericidal  sera  prepared  by 
inoculating  animals  with  typhoid  bacilli.  Kraus  and  Stenitzer 64 
have  also  taken  up  the  study  of  typhoid  immunization  from  the 
point  of  view  that  the  typhoid  bacillus  produces  a  true  toxin,  and  that 
therefore  a  true  antitoxic  action  could  be  expected  from  the  sera  pro- 
duced by  immunization  with  typhoid  filtrates.  It  should  be  noted 
that,  in  spite  of  the  most  common  opinions  against  this  at  present, 
a  similar  point  of  view  was  advanced  by  MacFadyen,65  and  more 
recently  by  Arima.66  Kraus  and  Stenitzer  67  immunized  horses  and 
goats  very  highly  with  extracts  of  agar  cultures  and  with  broth  fil- 
trates by  intravenous  injection.  The  serum  which  they  produced  in 
this  way  not  only  possessed  the  ordinary  bactericidal  action,  but, 
they  claimed,  neutralized  also  toxic  broth  filtrates,  not  only  of  the 
typhoid,  but  of  the  paratyphoid  bacilli.  The  serum  of  Kraus  and 
Stenitzer  has  been  used  by  a  number  of  observers,  among  whom  are 
Herz,68  Forssmann,  linger,  Russ,  and  others,  and  the  results  are 
said  to  be  encouraging  in  early  cases. 

Rodet  and  Lagrifoul69  immunized  horses  with  living  typhoid 
cultures,  and  also  claim  favorable  results. 

Mathes,70  continuing  the  work  of  Gottstein  after  the  death  of  the 
latter,  employed  the  method  of  immunizing  with  the  product  ob- 
tained by  digesting  typhoid  bacilli  with  trypsin.  The  poison  so  pro- 
duced he  speaks  of  as  "fermotoxin."  Liidke  71  slightly  modified  the 
Gottstein-Mathes  method  by  digesting  the  typhoid  bacilli  with  pepsin 
and  hydrochloric  acid,  and  with  the  poison  so  produced  immunized  8 
goats,  reenforcing  the  immunization  by  the  subsequent  injection  of 
the  bacilli  themselves.  He  claims  that  0.05  to  0.1  c.  c.  of  the  serum 
so  produced  protected  animals  against  five  times  the  lethal  dose 
of  the  poison.  In  a  small  series  of  human  cases  treated  by  this 
method  he  reports  good  results. 

Garbat  and  Meyer  72  immunized  animals  with  sensitized  typhoid 
bacilli,  and  claim  that  the  most  potent  sera  for  typhoid  immunization 
can  be  obtained  by  the  combination  of  sera  produced  by  the  injection 
of  sensitized  and  of  unsensitized  bacteria.  They  assert  that  the 
typhoid  bacillus  contains  two  definite  antigens,  one  particularly  as- 

64  Kraus  and  Stenitzer.    Wien.  kl.  Woch.,  Vol.  20,  1907,  pp.  344  and  753, 
and  Vol.  21,  1908,  p.  645. 

65  MacFadyen.     Cited  from  Stenitzer  in  "Kraus  und  Levaditi  Handbuch," 
Vol.  2. 

66  Arima.     CentralU.  f.  Bakt.,  65,  1912,  p.  183.     Orig. 

67  Kraus  and  Stenitzer.    Wien.  kl.  Woch.,  Vol.  22,  1909,  p.  1395;  Deutsche 
med.  Woch.,  March,  1911. 

68  Herz.     Wien.  kl.  Woch.,  Vol.  22,  1909,  p.  1746. 

69  Rodet  and  Lagrifoul.     C.  R.  de  la  Soc.  de  Biol.,  April,  1910. 

70  Mathes.     D.  Archiv  f.  kl.  Med.,  Vol.  95,  1909. 

71  Liidke.     D.  Archiv  f.  kl.  Med.,  98,  1910. 

72  Garbat  and  Meyer.     Zeitschr.  f.  exp.  Path.  u.  Ther.,  Vol.  8,  1911. 


480  INFECTION    AND    RESISTANCE 

sociated  with  the  bacterial  ectoplasm,  which  becomes  active  when  the 
bacteria  enter  the  animal  body,  and  a  truly  endocellular  poison  which 
does  not  become  active  until  the  surrounding  ectoplasm  is  dissolved. 
They  believe  that  sensitizing  bacteria  is  a  method  for  the  production 
of  endotoxin,  and  think  that  therefore  the  ideal  serum  for  the  treat- 
ment of  typhoid  consists  of  a  mixture  of  two  sera  produced  each 
with  one  of  the  antigens,  that  is,  with  sensitized  and  unsensitized 
bacteria.  Rommel  and  Herman  73  did  not  obtain  encouraging  results 
with  this  serum. 

From  a  study  of  the  literature  it  seems  to  us  that  in  spite  of  the 
many  different  methods  of  production  employed  by  various  observers 
in  their  studies  on  typhoid  sera  it  is  quite  likely  that  all  these  sera 
are  essentially  alike,  containing,  quantitatively,  according  to  the  de- 
gree of  immunization,  bactericidal,  agglutinating,  and  opsonic  prop- 
erties, with  possibly  a  limited  amount  of  neutralizing  power  for  the 
poisons  liberated  from  the  typhoid  bacilli  in  the  body.  As  far  as  we 
can  judge  from  clinical  reports  the  therapeutic  value  of  the  sera 
so  far  produced  is  not  very  great.  It  seems  that  cases  treated  early 
in  the  disease  may  be  benefited,  and  possibly  an  early  cessation  of 
the  bacteriemia  can  in  this  way  be  attained.  However,  it  does  not 
seem  either  theoretically  or  from  the  study  of  clinical  publications 
that  any  very  marked  effects  have  followed  the  use  of  any  of  the  sera 
in  advanced  cases. 

THE  SEBUM  TREATMENT  OF  PLAGUE 

That  the  serum  of  animals  immunized  with  killed  plague  cultures 
may  actively  protect  normal  animals  from  experimental  infection 
was  first  shown  by  Yersin,  Calmette,  and  Borrel.74  The  serum  which 
they  produced  possessed  apparently  powerful  bactericidal  action, 
but  no  antitoxic  properties  were  demonstrated.  They  determined 
its  protective  powers  by  injecting  measured  quantities  into  mice  and 
infecting  them  with  fatal  doses  of  virulent  plague  bacilli  24  hours 
later.  The  Yersin  serum  which  was  produced  for  the  treatment  of 
plague  as  a  result  of  these  experiments  was  made,  then,  by  the 
gradual  immunization  of  horses  with  first  dead  plague  bacilli,  finally 
with  virulent  living  organisms.  The  serum  has  been  extensively 
used  by  many  observers  with  results  that  leave  one  much  in  doubt 
as  to  its  efficacy.  Yersin  75  himself,  reporting  on  an  epidemic  in 
Nhatrang,  reports  a  general  mortality  of  73  per  cent,  for  the  whole 
epidemic,  a  mortality  of  100  per  cent,  in  untreated  cases,  and  of  42 
per  cent,  among  those  treated  with  his  serum.  Good  results  were 
also  reported  from  the  epidemics  in  Amoy  and  Canton  in  1896. 

73  Rommel  and  Herman.     Centralbl.  f.  Bakt.     Ref.  Vol.  53,  1912. 

74  Yersin,   Calmette,  and  Borrel.     Ann.  de  I'Inst.  Past.,  1895. 
76  Yersin.     Ann.  de  I'Inst.  Past.,  1899. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         481 

However,  these  results  apparently  were  not  accepted  by  all  observers 
as  proving  the  efficiency  of  the  serum,  since  the  number  of  cases 
observed  were  few,  and  the  irregularity  in  the  gravity  of  the  disease 
in  different  individuals  makes  statistical  evidence  unreliable  unless 
large  material  can  be  studied.  Kolle  and  Martini  76  announce  that 
Dr.  Choksy  reported  very  poor  success  with  the  Yersin  serum,  and 
cite  a  number  of  later  writers  whose  results  with  this  serum  were 
also  unsatisfactory  when  used  on  human  beings.  That  the  serum 
unquestionably  contains  antibodies  against  the  plague  bacillus  is 
testified  to,  not  only  by  the  French  observers  themselves,  but  also  by 
the  German  Plague  Commission  of  1899,  and  by  Kolle  and  Mar- 
tini 77  themselves.  The  Commission  experimented  with  this  serum 
upon  monkeys,  and  showed  that  it  possessed  unquestionable  protec- 
tive powers  in  rodents  and  in  monkeys  when  given  24  hours  before 
the  plague  infection,  and  in  monkeys  possessed  fair  curative  prop- 
erties when  injected  24  hours  later  than  inoculation  with  the  plague 
bacilli.  Because  of  the  doubtful  success  in  the  treatment  of  human 
beings  with  this  serum  Yersin  and  Roux  at  the  Pasteur  Institute 
later  altered  their  methods  of  serum  production  by  injecting,  not 
only  dead  and  living  plague  cultures,  but  considerable  quantities  of 
culture  filtrates  after  the  horses  had  attained  a  high  degree  of  im- 
munity. Later  observations  on  the  Yersin  78  serum  have  been  pub- 
lished by  the  British. Plague  Commission  in  1908  and  1911.  In  this 
investigation  the  cases,  were  controlled  as  to  their  severity  by  blood 
culture,  since  it  had  been  claimed  by  a  number  of  earlier  investi- 
gators that  the  Yersin  serum  was  efficient  in  mild  cases,  but  failed 
entirely  in  the  severe  ones.  It  seems  from  the  report  of  this  Commis- 
sion that  ordinarily  70  per  cent,  of  cases  of  plague  without  bacilli  in 
the  blood  survive  while  three-quarters  of  those  with  mild  septicemia 
die,  and  all  of  those  with  a  marked  septicemia  succumb.  In  the 
summary  given  of  146  cases  treated  with  Yersin's  serum  by  the 
British  Commission  65.1  per  cent,  died,  whereas  of  146  untreated 
controls  71.90  per  cent.  died.  These  figures,  together  with  an 
analysis  of  the  percentages,  classified  according  to  the  severity  of  the 
infections,  do  not  show  a  very  marked  curative  action  on  the  part 
of  the  serum. 

Markl,79  who  claims  that  the  plague  bacillus  produces  a  soluble 
toxin,  has  produced  a  plague  serum  by  immunization  of  animals  by 
filtrates  of  broth  cultures.  He  claims  that  0.1  c.  c.  of  his  serum,  as 
produced  at  Vienna,  will  protect  various  animals  against  lethal  doses 
of  plague  bacilli  if  given  at  the  same  time.  He  attributes  much  of 

76  Kolle  and  Martini.     Deutsche  med.  Woch.,  1902,  p.  29. 

77  German  Plague  Commission.     Arb.  a.  d.  kais.  Amt.,  Vol.  16,  1899. 

78  British  Plague  Commission.    Journ.  of  Hyg.,  Vol.  12,  Sup.,  1912,  p.  326. 

79  Markl.      Centralbl.  f.   Bakt.,   24,   1898;   Zeitschr.   f.  Hyg.,   37,   1901; 
Zeitschr.  f.  Hyg.,  42,  1903. 


482  INFECTION    AND    RESISTANCE 

its  curative  action  to  the  fact  that  in  the  presence  of  this  serum  active 
phagocytosis  takes  place. 

Dean,80  in  1906,  also  claimed  to  have  produced  strongly  antitoxic 
plague  sera  by  treating  horses  with  nitrates  from  8  to  10  weeks  old 
bouillon  cultures.  He  claims  that  1  c.  c.  of  his  serum  will  neutralize 
150  or  450  minimal  lethal  doses  of  the  plague  poison  according  to 
whether  one  measures  the  M  L  D  by  death  in  48  hours  or  in  4  days. 
Rowland  81  also  has  produced  a  serum  by  the  immunization  of  ani- 
mals with  the  "toxins"  produced  by  his  sulphate  process.  Rowland 
has  apparently  utilized  the  idea  previously  advanced  by  Lustig  of 
immunizing  with  "nucleoproteins"  derived  from  the  plague  bacillus 
instead  of  with  the  whole  bacteria.  Lustig's  82  method  consisted  of 
washing  up  agar  cultures  of  plague  bacilli  in  1  per  cent,  sodium 
hydrate  solution,  precipitating  with  ascitic  acid,  taking  up  the  pre- 
cipitate in  an  indifferent  fluid  and  injecting  it  into  horses.  The 
serum  produced  by  Lustig's  method  was  used  in  Bombay,  and  is  re- 
ported by  Hahn  as  effective  in  milder  cases,  but  without  action  in 
the  more  severe  ones.  There  was  but  slight  difference  in  the  latter 
type  between  the  treated  and  the  untreated  cases. 

Rowland's 83  method  consisted  in  the  treatment  of  the  moist 
bacteria  with  enough  anhydrous  sulphate  of  soda  to  combine  with 
all  the  water  present,  freezing  and  thawing  the  mixture  and  filtering 
off  the  bacterial  deposit  at  37°  C.  Subsequently  he  extracted  this 
bacterial  mass  with  water.  The  extract  so  obtained  was  fatal  to 
rats  in  quantities  of  0.05  to  0.1  mg.,  killing  them  in  18  hours.  In 
his  experiments  doses  of  0.001  to  0.01  afforded  protection,  the  last- 
named  quantity  reducing  the  mortality  after  inoculation  of  fatal 
doses  from  80  per  cent,  to  10  per  cent. 

The  sera  produced  by  the  immunization  of  horses  with  these 
supposed  nucleoproteids  are  taken  to  be  antitoxic  in  nature  by  Row- 
land himself  and  by  MacKonky.  They  were  used  in  the  treatment  of 
plague  cases  in  the  epidemics  of  1908  and  1911  by  the  Maratha  Hos- 
pital in  Bombay,  and  reported  upon  by  the  Indian  Plague  Commis- 
sion on  the  basis  of  observations  made  by  Dr.  Choksy.  The  cases  in 
this  series  were  controlled,  as  were  those  treated  by  the  Yersin  serum, 
by  blood  culture.  Here  the  results  were  not  striking — 68.40  per 
cent,  of  the  serum-treated  cases  died,  while  77.60  per  cent,  of  the 
controls  died. 

'  Altogether  we  cannot  draw  any  definite  conclusions  as  to  the 
value  of  the  serum  treatment  in  plague.    On  the  whole  it  does  appear 

80  Dean.     Cited  from  MacKonky,  Journ.  of  Hyg-,  Vol.  12,  Plague  Suppl. 
II,  1912,  p.  402. 
,     81  Rowland.     Journ.  of  Hyg.,  Vol.  11,  Plague  Suppl.  I,  pp.  11-19. 

82  Lustig.     "Monograph   Sierotrapia  e   Vaccin   Prev.   Control  la  Peste,* 
Turin,  1899 ;  cited  from  Kolle  and  Martini,  loc.  cit. 

83  Rowland.    Journ.  of  Hyg.,  Vol.  10,  p.  536. 


THERAPEUTIC    IMMUNIZATION    IN    MAN         483 

that  the  milder  cases  are  materially  benefited  by  the  treatment,  and  it 
is  not  at  all  impossible  that  in  such  cases  aggravation  of  a  milder 
case  into  fatal  septicemia  may  be  prevented  by  the  timely  adminis- 
tration of  the  plague  serum.  Animal  experimentation  also  seems  to 
indicate  that  the  administration  of  the  serum  may  be  of  great  value 
as  a  prophylactic  measure.  It  seems,  on  the  other  hand,  as  far  as  we 
can  judge  from  the  evidence  of  statistics,  that  when  a  case  of  plague 
has  developed  into  the  condition  of  active  septicemia  the  administra- 
tion of  even  the  strongest  plague  sera  at  present  available  is  of  little 
use.  And  this  is  indeed  unfortunately  true  of  all  passive  immuniza- 
tion where  the  activity  of  the  serum  seems  to  depend  chiefly  upon 
bactericidal  and  opsonic  properties.  For  we  cannot  definitely  accept 
at  the  present  day  the  claims  that  a  true  antitoxic  serum,  in  the  sense 
of  those  produced  against  diphtheria  and  tetanus  poisons,  can  be 
really  produced  in  the  case  of  plague.  The  toxic  substances  derived 
from  plague  bacilli  by  a  number  of  observers  do  not  correspond  in 
many  particulars  to  true  toxins. 


FACTS    CONCERNING   ACTIVE   PROPHYLACTIC    IMMUNIZATION 

IN    MAN 

In  a  previous  chapter  we  have  dealt  with  the  treatment  of  in- 
fectious disease  with  emulsions  of  dead  bacteria  or  vaccines.  The 
discussion  there  was  confined  to  the  use  of  these  substances  in  the 
case  of  developed  disease  in  which  the  infectious  agent  had  already 
gained  a  foothold  in  the  body.  Concerning  this  form  of  therapy 
much  difference  of  opinion  exists,  and  we  have  seen  that  careful 
judgment  must  be  applied  to  the  selection  of  cases  to  which  treatment 
with  vaccines  is  adapted. 

Concerning  the  prophylactic  immunization  of  human  beings  with 
bacteria  there  can  be  little  difference  of  opinion ;  this  procedure  finds 
its  justification  in  prolonged  laboratory  experience  in  the  hands  of 
many  men  since  the  days  of  Pasteur. 

The  principle  of  specifically  increasing  the  resistance  of  an  in- 
dividual by  treatment  with  an  altered  form  of  the  disease,  either 
with  the  attenuated  bacteria,  with  dead  bacteria,  or  with  bacterial 
extracts,  has  been  sufficiently  discussed  in  Chapter  IV.  It  is  indeed 
surprising  that  this  phenomenon  of  prophylactic  protection,  though 
discovered  by  Jenner  in  small-pox,  and  developed  by  Pasteur  in 
rabies,  did  not  find  more  general  application  to  the  diseases  of  man 
until  recejit  years.  At  present  such  methods  are  in  extensive  use  in 
typhoid  fever,  in  which  they  have  had  unquestionably  excellent  re- 
sults. In  the  cases  of  cholera  and  plague  numerous  attempts  have 
been  made,  but  the  results  here  are  not  as  clear-cut  as  they  have  been 
in  the  case  of  typhoid.  In  the  succeeding  paragraphs  we  have  set 


484  INFECTION    AND    RESISTANCE 

forth  as  briefly  as  possible  the  methods  employed  in  prophylactically 
immunizing  man  against  this  disease  in  which  this  procedure  has 
been  most  commonly  attempted. 

PBOPHYLACTIC  IMMUNIZATION  IN  TYPHOID  AND  PARATYPHOID 

FEVER 

The  first  attempt  to  inoculate  human  beings  with  typhoid  bacilli 
with  the  intention  of  producing  prophylactic  active  immunization 
was  probably  that  made  by  Pfeiffer  and  Kolle  84  in  1896.  During 
the  same  year  also  Wright  85  made  similar  studies  in  England,  and 
soon  after  this  he  reported  upon  the  development  of  antibodies  in  the 
blood  of  17  people  inoculated  with  typhoid.  By  these  studies  it  was 
shown  that  human  beings  could  be  inoculated  with  dead  typhoid 
bacilli  without  danger,  and  this  logically  led  to  the  attempt  to  vac- 
cinate human  beings  on  a  large  scale. 

It  is  hardly  necessary  to  dwell  upon  the  desirability  of  such  a 
procedure.  From  tables  recently  published  by  Russell  86  we  take  the 
information  that,  in  our  own  Spanish- American  war,  20,738  cases 
of  typhoid  with  1,580  deaths  occurred  in  a  total  enlistment  of  107,- 
973.  In  this  entire  war  243  men  were  killed  in  action  or  died  of 
their  wounds,  while  almost  7  times  as  many  died  of  typhoid  fever. 
In  the  British  army  during  the  Boer  war  there  were  over  75,000 
cases  of  typhoid  in  380,000  men,  and  in  the  Eussian  army  during  the 
Russo-Japanese  war  over  17,000  cases  of  typhoid  occurred,  over  half 
as  many  as  the  number  of  men  killed  in  action.  Such  appalling  fig- 
ures leave  no  possible  doubt  as  to  the  desirability  of  prophylactic  im- 
munization in  armies,  and  there  can  be  little  question  that  typhoid 
fever  is  sufficiently  prevalent  in  many  parts  of  the  civilized  world  to 
encourage  prophylactic  immunization  of  individuals,  even  when  not 
living  under  the  especially  dangerous  conditions  of  camps. 

Following  the  preliminary  studies  of  Pfeiffer  and  Kolle  and  of 
Wright  extensive  practical  studies  of  vaccination  were  made  in  the 
German  colonial  army  during  the  Herrero  war,  and  by  British 
bacteriologists  during  the  Boer  war.  Leishmann  87  also  studied  care- 
fully the  results  of  vaccination  among  regiments  of  the  British  army 
in  India. 

The  vaccine  employed  by  Wright  and  his  associates  in  England 
•consisted  of  broth  cultures  of  a  typhoid  bacillus  killed  by  exposure  to 
53°  C.,  and  by  the  further  addition  of  0.4  per  cent,  of  lysol.  The 
German  vaccine  consisted  of  emulsified  agar  cultures  similarly  killed. 

The  results  obtained  with  these  vaccines,  although  encouraging, 

84  Pfeiffer  and  Kolle.     Deutsche  med.  Woch.,  22,  1896,  p.  735. 

85  Wright.     Brit.  Med.  J.,  Jan.,  1897,  p.  256. 

86  Russell.     Amer.  J.  of  Med.  Sciences,  Dec.,  1913,  Vol.  146. 

87  Leishmann.     Glasgow  Med.  Journ.,  1912,  Vol.  77,  p.  408,  cited  from 
Russell. 


THERAPEUTIC    IMMUNIZATION    IN    MAN 


485 


were  not  as  striking  as  had  been  hoped.  Russell  88  summarizes  the 
earlier  attempts  by  stating  that  the  morbidity  was  reduced  to  about 
one  half  among  vaccinated  persons  with  a  slightly  greater  reduction 
of  mortality.  The  last-named  writer  also  attributes  the  early  fail- 
ures to  the  overheating  of  the  vaccines  with  a  consequent  reduction 
of  their  antigenic  properties,  and  to  timidity  in  their  administration 
resulting  from  Wright's  fear  of  a  negative  phase.  Russell,  of  the 
United  States  Army  Medical  Corps,  made  a  most  extensive  study  of 
typhoid  vaccination  in  this  country.  After  careful  consideration  of 
the  methods  of  others  he  produces  his  vaccines  as  follows :  A  single 
strain  of  typhoid  bacilli  is  used  (culture  Rawlings  obtained  from 
England),  and  this  is  grown  on  agar  in  Kolle  flasks  for  18  hours. 
The  purity  of  the  culture  is  tested  out  both  morphologically  and  by 
transplantation  upon  the  double  sugar  media  devised  by  Russell. 
Agglutination  tests  are  also  made.  After  18  hours  the  growth  is 
washed  off  in  small  quantities  of  salt  solution  and  the  emulsion  heated 
in  a  water  bath  for  one  hour  at  53°  C. ;  it  is  then  diluted  with  sterile 
salt  solution  to  a  concentration  of  one  billion  bacteria  to  a  cubic 
centimeter.  Then  0.25  per  cent,  of  tricresol  is  added.  Before  use 
the  safety  of  the  vaccine  is  ascertained  both  by  aerobic  and  anaerobic 
cultivation  and  by  the  injection  into  mice  and  guinea  pigs  of  consid- 
erable quantities  for  the  exclusion  of  possible  tetanus  contamination. 
The  efficiency  of  the  vaccine  is  then  tested  by  injecting  rabbits  with 
three  doses  at  10-day  intervals,  and  determining  the  resulting  ag- 
glutinating power. 

With  these  vaccines  under  the  direction  of  the  United  States 
Army  Medical  Corps  the  troops  mobilized  in  Texas,  California,  and 
along  the  Mexican  border  were  treated.  Compulsory  vaccination  was 
established  in  March,  1911,  and  the  results  as  reported  by  Russell 
have  fully  justified  the  measure.  The  following  table  taken  from 
Russell's  paper  will  illustrate  the  results  obtained : 

Typhoid  Fever.     Officers  and  Enlisted  Men,  United  States  Army 


Totals 

Yr. 

Jan. 

Feb. 

Mar. 

Apr. 

May 

June 

July 

Aug. 

Sept. 

Oct. 

Nov. 

Dec. 

for 

9  months 

Volun- 

f 1908 

5 

6 

4 

2 

3 

11 

14 

31 

25 

26 

12 

8 

101 

tary 

i  1909 

4 

10 

6 

4 

11 

15 

26 

14 

16 

45 

20 

6 

106 

T1910 

8 

11 

1 

4 

2 

6 

12 

27 

21 

16 

20 

11 

92 

Com- 

I 1911 

3 

3 

3 

7 

4 

4 

4 

7 

4 

4 

1 

0 

39 

pulsory 

1  1912 

1 

2 

2 

0 

0 

3 

1 

3 

1 

4 

0 

1 

13 

I  1913 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

Paratyphoid  fever  included  in  figures  for  1908,  but  excluded  in  other  years.     Cases  paratyphoid 
1909,  3;  1910,  3;  1911,  2;  1912,  3;  1913,  0. 

88  Russell.    Am.  Journ.  of  Med.  Sc.,  Vol.  146,  Dec.,  1913. 


486  INFECTION  AND  RESISTANCE 

We  have  mentioned  in  another  place  that  Metchnikoff  and  Bes- 
redka  in  their  studies  on  typhoid  vaccination  in  the  chimpanzee 
have  concluded  that  very  little  protective  value  resided  in  vaccina- 
tion with  dead  typhoid  vaccines,  whereas  animals  vaccinated  with 
small  amounts  of  living  cultures  were  very  efficiently  protected. 
Metchnikoff  and  Besredka  adopted  finally  the  method  of  immunizing 
with  living  sensitized  vaccines.  By  this  is  meant  typhoid  bacilli  that 
have  been  exposed  to  the  action  of  heated  immune  serum,  or,  in 
other  words,  typhoid  bacilli  that  have  absorbed  specific  antibodies. 
There  is  no  question  as  to  the  efficiency  of  this  form  of  vaccination. 
The  method  of  employing  sensitized  bacteria  for  these  purposes 
utilized  by  Besredka  in  the  case  of  plague  has  unquestionably  won 
an  important  place  in  active  immunization.  However,  the  results 
of  Russell  and  others  seem  to  indicate  that  in  human  beings  the  use 
of  dead  vaccines  is  certainly  of  considerable  value,  and  there  are 
certain  practical  objections  to  the  use  of  living  vaccines  in  immuniza- 
tion of  large  numbers  of  people  as  in  armies  to  which  Russell  calls 
attention.  In  the  first  place,  living  vaccines  cannot  be  stored  for 
any  considerable  period,  and  may  become  a  source  of  possible  infec- 
tion by  mouth  if  carelessly  handled;  furthermore,  contamination  is 
not  so  easily  ruled  out  in  the  case  of  living  vaccines  when  used  on 
a  large  scale,  and  it  is  not  possible  at  present  to  require  compulsory 
vaccination  with  living  bacteria.  The  choice  of  strains  is  important. 
For  vaccines  in  the  British  and  American  armies  the  Rawlings  strains 
usually  used  by  Wright  are  employed  at  present.  The  work  of  Weiss, 
Hooker  and  others  recently  has  shown  that  perhaps  we  may  have  to 
Ude  for  best  results  a  polyvalent  typhoid  injection,  and  that  in  im- 
munizing with  paratyphoid  B,  a  polyvalent  vaccine  will  be  necessary 
eventually.  The  paratyphoid  A  strains  are  so  uniform  antigenically 
that  this  will  be  unnecessary  in  their  case. 

Gay  has  recently  recommended  the  use  of  sensitized  killed  vac- 
cines. He  controls  the  efficiency  of  his  vaccines  by  testing  them  out 
on  rabbits  in  which  typhoid  septicemia  has  been  produced  by  inocu- 
lation with  cultures  grown  on  rabbit  blood  agar.  These  vaccines  have 
not  yet  been  used  upon  sufficient  numbers  to  justify  conclusions.  It 
would  seem,  however,  that  any  one  of  the  methods  mentioned  must 
possess  considerable  value,  since  they  all  represent  merely  slight 
variations  of  the  same  procedure.  The  method  at  present  used  in 
the  German,  British,  and  American  armies,  namely,  vaccination 
with  dead  cultures,  seems  certainly,  according  to  Russell's  statistical 
studies,  to  have  yielded  excellent  results  and  recommends  itself  by 
its  extreme  simplicity  and  safety. 

THERAPEUTIC  VACCINE  TREATMENT  IN  TYPHOID  FEVER 

A  number  of  workers  have  attempted  to  treat  typhoid  fever  after 
it  occurred  with  typhoid  vaccine  sensitized  and  unsensitized. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  487 

Frankel  did  this  as  early  as  1893.  He  was  followed  by  Petruschky  89 
and  better  and  in  1912  Ichikawa  began  the  intravenous  injection 
of  typhoid  vaccine  in  patients  with  the  rather  astonishing  results  of 
obtaining  in  many  cases  rapid  falling  of  temperature  and  often  ap- 
parent shortening  of  the  disease.  Boinet  in  1912  obtained  similar 
results  and  recently  the  same  thing  has  been  done  by  Gay.89a 

Of  these  writers  the  only  one  who  recognized  immediately  that 
perhaps  the  intravenous  injection  of.  such  vaccines  was  not  due  to 
purely  specific  activity  was  Ichikawa,  who  thought  the  paratyphoid 
bacilli  injected  into  typhoid  cases  often  produced  similar  results, 
and  soon  after  this  Kraus  obtained  similar  degrees  of  temperature 
and  favorable  results  by  injecting  colon  bacilli  into  typhoid  patients, 
into  a  few  cases  of  pyocyaneus  infection,  and  into  streptococcus  septi- 
cemias. 

With  Jobling  and  others  we  have  believed  from  the  beginning 
that  at  least  an  important  factor  in  the  activity  of  these  vaccines 
was  non-specific,  due  perhaps  to  a  rapid  mobilization  of  leucocytes 
and  of  ferments.  This  is  discussed  in  another  section  on  page  522. 

At  any  rate  the  therapeutic  results  obtained  justify  the  cautious 
continuance  of  such  intensive  treatment  in  various  diseases. 

ACTIVE  PROPHYLACTIC  IMMUNIZATION  IN  CHOLERA 

Attempts  to  protect  human  beings  against  cholera  by  prophylactic 
vaccination  were  made  as  early  as  1885  by  Ferran,89b  a  pupil  of 
Pasteur.  At  the  time  at  which  Ferran's  experiments  were  done  little 
was  known  regarding  the  production  of  immunity  with  killed  cul- 
tures or  with  bacterial  extracts,  and  Ferran,  under  the  influence  of 
the  French  school  and  its  endeavors  to  immunize  with  living  attenu- 
ated organisms,  applied  similar  methods  to  cholera.  First  experi- 
menting with  guinea  pigs,  he  soon  applied  his  method  to  human 
beings,  inoculating  them  with  small  quantities  of  living  broth  cul- 
tures of  cholera  spirilla.  In  many  of  his  experiments  he  gave,  at  the 
first  injection,  8  drops  of  a  fresh  broth  culture,  following  this  after 
8  days  with  0.5  c.  c.  of  a  similar  culture.  There  is  no  reason  why 
Ferran's  method  should  not  have  yielded  excellent  results.  How- 
ever, it  is  stated  that  he  worked  with  impure  cultures,  and  other 
observers,  notably  Nikati  and  Rietsch,  van  Ermengen,  da  Lara,  and 
others,  failed  to  obtain  encouragement  in  their  subsequent  investiga- 
tion of  this  method  of  vaccination. 

The  method  which  Haffkine  90  worked  out  some  years  after  Fer- 

89  Petruschky,  Cent.  f.  Bakt.,  I,  XIX,  1896. 
89a  Gay.   Arch,  of  Int.  Med.,  1914. 
89b  Ferran.     C.  E.  de  VAcad.  des  Sc.,  1885. 

90Haffldne.  The  Lancet,  February,  1893;  Brit.  Med.  Journ.,  December, 
1895. 


488  INFECTION  AND  RESISTANCE 

tan's  experiments  also  depended  upon  the  injection  of  living  cul- 
tures, but  Haffkine  attempted,  by  a  rather  elaborate  technique,  to 
produce  two  separate  vaccines,  one  attenuated,  the  other  enhanced  in 
virulence.  Attenuation  was  accomplished  by  growing  the  cholera 
spirilla  at  a  temperature  of  39°  C0  in  broth  over  the  surface  of  which 
a  constant  stream  of  sterile  air  was  passed.  Under  these  conditions 
the  first  crop  of  cholera  organisms  died  rapidly,  but  Haffkine  prac- 
ticed reinoculation  into  new  broth  flasks  before  complete  death  of  the 
original  culture  had  taken  place;  after  a  series  of  generations  of 
cultivation  in  this  way  he  obtained  cultures  which  produced  merely 
temporary  and  slight  edema  when  injected  under  the  skin  of  guinea 
pigs.  This  weakened  virus  was  used  for  the  first  inoculation. 

He  enhanced  the  virulence  of  cholera  cultures  with  the  purpose 
of  producing  a  strain  of  maximum  potency  comparable  to  virus  fixe. 
His  procedure  was  as  follows: 

a.  Giving  an  animal  an  intraperitoneal  injection  of  cholera 
spirilla  larger  than  the  fatal  dose. 

b.  Taking  out  the  peritoneal  exudate  and  exposing  it  for  a  few 
hours  to  the  air. 

c.  Injecting  this  exudate  into  another  animal  and  treating  the 
resulting  peritoneal  exudate  in  the  same  way. 

After  a  series  of  such  animal  passages  he  claims  to  have  obtained 
a  virus  of  great  virulence,  and  this  is  his  second  and  stronger  vaccine. 

In  applying  the  method  to  human  beings  Haffkine  planted  the 
cholera  spirilla  upon  agar  slants  of  the  standard  size,  emulsified  the 
growths  in  sterile  water,  and  injected  1/5  to  1/20  c.  c.  of  such  a 
culture,  using  first  the  weak  vaccine  and  five  days  later  a  more 
virulent  culture. 

Beginning  his  work  as  early  as  1893,  Haffkine  and  others  vac- 
cinated as  many  as  40,000  people  in  India.  On  the  whole,  the  results 
obtained  were  very  encouraging.  It  is  a  question,  however,  whether 
or  not  his  method  is  unnecessarily  complicated.  In  the  light  of  our 
more  recent  knowledge  concerning  cholera  immunity  it  seems  likely 
that  the  importance  which  Haffkine  attached  to  the  virulence  of  the 
cholera  culture  used  for  injection  was  exaggerated,  and  we  have 
reason  to  believe  that  simple  immunization  with  killed  cultures  may 
produce  results  fully  as  efficacious.  After  all,  we  could  not  expect, 
at  least  at  present,  to  produce  by  active  artificial  immunization  an 
immunity  as  permanent  as  that  which  results  from  an  attack  of  the 
disease.  Concerning  the  reasons  for  the  acquisition  of  such  perma- 
nent immunity  we  have  as  yet  little  knowledge.  Even  Haffkine's 
method  of  inoculation  with  living  virus  does  not,  by  his  own  estima- 
tion, last  longer  than  possibly  two  years.  It  is  therefore  likely  that 
prophylactic  immunization  in  cholera  is  efficacious  by  reason  of  the 
appearance  in  the  blood  serum  of  the  specific  bactericidal  and  opsonic 
substances  by  which  the  small  numbers  of  cholera  organisms  enter- 


THERAPEUTIC  IMMUNIZATION  IN  MAN  489 

ing  during  spontaneous  infection  can  be  disposed  of  before  a  foot- 
hold in  the  body  is  gained. 

Tamancheff  later  used  Haffkine's  method,  but  killed  the  cultures 
by  the  addition  of  a  0.5  per  cent,  solution  of  carbolic  acid. 

Kolle 91  later  recommended  the  injection  of  dead  cholera  or- 
ganisms, maintaining  that  a  single  injection  of  about  2  milligrams 
of  a  culture  killed  by  exposure  to  50°  C.  for  a  few  minutes,  and 
by  the  addLl.n  of  0.5  per  cent,  of  phenol,  is  sufficient  to  immunize 
successfully.  Good  results  with  Kolle's  method  have  been  reported 
from  Japan. 

Strong,92  also  proceeding  from  the  idea  that  the  immunizing 
antigen  is  present,  as  such,  within  the  cell  body  of  the  cholera 
spirilla,  recommends  the  injection  of  autolytic  products  obtained  by 
digesting  cholera  spirilla  in  aqueous  suspension  and  filtering.  He 
prepared  his  "prophylactic"  by  growing  the  organisms  upon  agar, 
then  suspending  the  growth  in  sterile  water  and  keeping  it  at  60°  C. 
for  from  one  to  twenty-four  hours.  The  mixture  was  then  exposed 
to  37°  C.  for  from  two  to  five  days  and  filtered  through  Reichel 
filters.  One  to  5  c.  c.  of  this  was  used  in  his  experiments  upon 
human  beings. 

PKOPHYLACTIC  IMMUNIZATION  AGAINST  PLAGUE 

The  first  attempts  to  immunize  human  beings  prophylactically 
against  plague  were  those  of  Haffkine.93  The  first  vaccinations  were 
carried  out  with  broth  cultures  killed  at  65°  C.  He  tested  out  his 
vaccines  on  a  large  scale  in  Bombay,  and  obtained  apparently  prom- 
ising results.  In  a  plague  epidemic  occurring  in  a  Bombay  prison 
only  2  of  151  vaccinated  persons  became  ill,  and  neither  of  these 
died;  whereas,  of  177  unvaccinated  persons  12  became  ill  and  6 
died.  In  large  series  of  vaccinated  people  only  1.8  per  cent,  were 
infected  with  plague,  with  a  mortality  of  0.4  per  cent,  for  the  total, 
whereas  of  unvaccinated  individuals  in  the  same  epidemic  7.7  per 
cent,  fell  victim  to  the  disease,  with  a  mortality  of  4.7  per  cent. 

The  German  Plague  Commission,  consisting  of  Gaffky,  Pfeiffer, 
and  Dieudonne,  recommended  a  vaccine  of  killed  agar  cultures. 
Kolle  and  Otto,94  basing  their  earlier  results  upon  experiments  car- 
ried out  with  monkeys,  mice,  guinea  pigs,  and  rats,  have  come  to  the 
conclusion  that  vaccination  with  dead  plague  cultures  is  much  in- 
ferior to  that  obtained  when  attenuated  living  cultures  are  used. 
The  same  conclusion  has  been  reached  by  Kolle  and  Strong.95  Kolle 

91  Kolle.    Deutsche  med.  Woch.,  1897,  No.  1. 

92  Strong.     Journ.  Inf.  Dis.,  Vol.  2,  1905. 

93  Haffkine.     Bull,  de  I'Inst.  Past.,  Vol.  4,  1906,  No.  20,  p.  825. 

94  Kolle  and  Otto.     Deutsche  med.  Woch.,  1903,  p.  493,  and  Zeitschr.  f. 
Hyg.,  Vol.  45,  1903. 

95  Kolle  and  Strong.    Deutsche  med.  Woch.,  XXXII,  1906,  p.  413. 


490  INFECTION  AND  RESISTANCE 

and  Otto  found  that  the  immunization  of  animals  with  large  doses 
of  killed  agar  cultures  of  plague  bacilli  and  with  Haffkine's  prophy- 
lactic did  not  protect  them  against  subsequent  inoculation  with  viru- 
lent cultures. 

Strong  96  subsequently  made  a  very  careful  comparative  study  of 
the  various  methods  of  plague  vaccination,  and  concluded  that  the 
most  efficient  method  is  immunization  with  attenuated  living  cul- 
tures. He  showed  that  when  carefully  done  this  method  can  be 
safely  employed  in  human  beings,  but  admits  that  his  work  must  be 
as  yet  considered  as  experimental  and  further  studied  before  it  can 
be  universally  employed. 

Besredka97  has  advised  the  use  of  sensitized  dead  plague  cul- 
tures, claiming,  from  animal  experimentation,  that  such  vaccines 
produce  an  efficient  and  relatively  durable  immunity. 

Rowland98  confirms  the  immunizing  properties  of  Besredka's 
vaccines  in  plague,  and  believes  that  the  antigenic  properties  of  the 
plague  bacillus  are  attached  to  the  bacterial  nucleoproteins,  and  can 
be  extracted  with  these.  Rowland  prepares  a  vaccine  by  the  treat- 
ment of  the  moist  bacteria  with  enough  anhydrous  sodium  sulphate 
to  combine  with  all  the  water  present,  freezing  and  thawing  the 
mixtures,  then  filtering  off  the  bacterial  deposits  at  37°  C.,  and  ex- 
tracting them  with  water.  The  solution  so  obtained  was  fatal  to 
rats  in  small  quantities  and  afforded  substantial  protection,  reducing 
the  mortality  on  subsequent  inoculation  of  a  standard  culture  from 
80  to  10  per  cent. 

PROPHYLAXIS  AGAINST  SMALL-POX 

In  the  case  of  small-pox  the  general  method  of  active  prophylactic 
immunization  is  in  principle  identical  with  that  devised  by  Jenner 
in  the  18th  century.  The  original  observation  from  which  Jenner 
worked  was  that  dairy  maids  and  other  individuals  who  had  been 
infected  with  cow  pox  were  thereafter  spared  when  a  small-pox  epi- 
demic appeared  in  the  region  in  which  they  lived.  It  is  now  agreed 
by  most  observers  who  have  studied  the  problem  that  the  virus  of  cow 
pox  and  that  of  small-pox  are  identical  in  nature ;  the  former  repre- 
senting a  strain  attenuated  by  passage  through  the  animal  body.  This 
is  based  chiefly  upon  the  observation  that  true  variola  can  be  trans- 
mitted to  cattle,  and  that  it  can  be  thus  carried  from  animal  to  ani- 
mal, during  this  process  becoming  attenuated  for  human  beings  to 
such  a  degree  that  reinoculated  into  man  a  simple  vaccinia  is  pro- 
duced." 

96  Strong.     Journ.  of  Med.  Res.,  N.  S.,  13,  1908. 
87  Besredka.    Bull,  de  I'Inst.  Past.,  Vol.  8,  1910. 

98  Rowland.     Journ.  of  Hyg.,  Vol.  12,  1912,  p.  344. 

99  Haecius.     Cited  from  Paul,  Kraus  and  Levaditi,  Vol.  1,  p.  593. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  491 

Small-pox,  therefore,  represents  in  principle  active  immunization 
by  means  of  attenuated  virus.  When  vaccination  was  first  introduced 
the  virus  was  taken  from  preceding  pustules  produced  in  other  human 
beings.  This  has  been  given  up  in  most  countries  to-day  largely  be- 
cause of  the  dangers  of  transferring  syphilis  and  other  diseases  in 
this  way.  At  present  the  method  of  obtaining  virus  for  vaccination 
purposes  is  carried  out  as  follows:  The  initial  material  consists  of 
what  is  known  as  "seed"  virus,  which  can  be  obtained  from  spon- 
taneous cow  pox  or  from  vaccination  pustules  in  children,  or  again 
from  pustules  obtained  in  calves  after  several  passages  of  small-pox 
virus  through  these  calves.  From  such  seed  virus  calves  may  be  in- 
oculated for  vaccine  production  or  else  the  calves  may  be  inoculated 
from  the  material  obtained  from  other  calves  in  the  usual  way. 

Healthy  young . animals  are  used;  they  are  washed  along  the 
abdomen,  strapped  down  upon  specially  prepared  tables,  and  the 
abdominal  skin  thoroughly  cleansed  with  soap  and  water.  The 
exact  procedure  varies  in  different  places;  often  the  skin  is  thor- 
oughly cleansed  with  carbolic  solution,  and  this  is  thoroughly  re- 
moved with  sterile  water  before  inoculation,'  or  else  cleansing  is  re- 
lied upon  without  the  use  of  germicides.  Over  the  clean  area 
longitudinal  scratches  1  to  2  c.  c.  apart  are  made,  and  into  these 
the  seed  virus  is  rubbed.  The  animals  are  then  kept  in  a  clean 
stall,  preferably  over  asphalt  floors,  and  rigid  cleanliness  is  observed 
during  the  period  of  development  of  the  pustules.  After  the  6th  or 
7th  day,  when  the  vesicles  are  beginning  to  appear,  the  abdomen  is 
well  washed  and  cleansed  of  superficial  dirt  without  the  use  of  an 
antiseptic,  and  the  pulp  removed  from  the  lesions  with  a  curette. 
The  pulp  so  removed  is  placed  into  60  per  cent,  glycerin  and  thor- 
oughly ground  up  in  a  specially  constructed  mill.  According  to 
Eosenau,  the  animal  should  always  be  killed  before  the  vesicles  are 
removed,  not  only  for  humane  reasons,  since  the  same  object  might 
be  attained  with  anesthesia,  but  because  a  thorough  autopsy  can  then 
be  performed  to  determine  the  health  of  the  calf. 

Vaccines  so  obtained  always  contain  bacteria,  the  glycerin  there- 
fore serving  a  double  purpose :  one,  the  preservation  of  the  virus,  the 
other  a  gradual  destruction  of  the  bacteria.  Rosenau  has  shown  that 
the  addition  of  2  to  4  parts  of  60  per  cent,  glycerin  to  one  part 
weight  of  the  pulp  prevents  the  growth  of  bacteria  and  probably 
destroys  them  by  dehydration.  Most  of  the  bacteria  are  destroyed 
within  one  month  at  20°  C.  During  this  period,  then,  from  4  to  6 
weeks,  the  glycerinated  virus  should  not  be  used,  and  should  from 
time  to  time  be  controlled  by  cultivation.  At  the  end  of  this  time 
the  lymph  is  ready  for  use. 

Formerly  the  material  for  the  vaccination  of  human  beings  was 
obtained  very  simply  by  dipping  ivory  splinters  into  the  fluid  of  pus- 
tules, allowing  this  to  dry,  and  rubbing  these  ivory  or  bone  points 


492  INFECTION  AND  RESISTANCE 

into  the  exudate  obtained  by  scratching  the  skin  of  the  individual  to 
be  vaccinated.  This  method  has  practically  gone  out  of  use,  and 
to-day  the  ripened  glycerin  pulp  prepared  as  above  is  taken  up  in 
small  capillary  glass  tubes  and  from  these  blown  upon  the  vaccina- 
tion scratch.  The  efficiency  of  vaccine  virus  can  be  tested  for  po- 
tency by  the  inoculation  of  the  ears  of  rabbits  before  use. 

ACTIVE  PROPHYLACTIC  IMMUNIZATION  IN  EABIES  (HYDROPHOBIA) 

Although  many  modifications  have  been  suggested  and  actually 
used  in  different  parts  of  the  world,  the  most  common  method  of 
immunizing  against  rabies  still  remains  that  originally  devised  by 
Pasteur.  The  Pasteur  treatment  takes  advantage  of  the  prolonged 
incubation  period  of  rabies  and  is  planned  to  confer  immunity  be- 
tween the  time  of  inoculation  and  the  time  at  which  the  disease 
would  naturally  appear.  Since  this  period  in  ordinary  street  infec- 
tion by  dog  bite  is  usually  40  days  or  more,  a  considerable  interval 
for  active  immunization  is  available.  Formerly  much  of  this  time 
was  lost  in  that  the  diagnosis  of  hydrophobia  in  the  dog  or  other 
animal  that  had  caused  the  injury  could  not  be  made  with  certainty 
until  the  results  of  rabbit  inoculations  had  been  obtained.  Nowadays 
the  ease  with  which  a  diagnosis  of  hydrophobia  can  be  made  within 
a  few  minutes  by  finding  negri  bodies  in  the  hippocampal  and  cere- 
bellar  cells  has  added  considerably  to  safety  in  that  it  has  made  pos- 
sible a  gain  of  almost  two  weeks  in  determining  whether  treatment 
should  be  instituted  or  not. 

Here  again,  although  the  infectious  agent  of  rabies  is  not  known 
with  certainty  even  at  the  present  day,  the  method  of  Pasteur  de- 
pends upon  active  immunization  by  means  of  an  attenuated  virus. 

In  standardizing  the  virus  for  the  purpose  of  treatment  Pasteur 
first  produced  what  he  calls  the  "virus  fixe."  This  consists  of  the 
ordinary  street  virus  as  obtained  from  rabid  animals  passed  through 
a  considerable  series  of  rabbits  (20-30)  until  its  virulence  for  these 
animals  has  reached  a  maximum.  After  a  sufficient  number  of  such 
rabbit  passages  the  incubation  time  after  intracerebral  inoculation  is- 
reduced  to  7  or  8  days,  but  can  no  longer  be  shortened  by  further 
passage.  The  brain  and  cord  material  of  rabbits  dead  of  rabies  after 
such  repeated  passages  constitutes  virus  fixe.  This  can  be  preserved 
•for  considerable  periods  in  60  per  cent,  glycerin,  and  this  is  the 
initial  material  from  which  the  attenuated  preparations  for  treat- 
ment are  produced. 

In  preparing  the  material  for  treatment  a  small  amount  of  virus 
fixe  is  injected  subdurally  into  rabbits,  about  0.2  c.  c.  of  a  salt  solu- 
tion emulsion  being  given.  The  inoculation  is  very  easily  made 
through  a  small  trephine  opening  in  the  skull,  and  contamination 
is  very  easily  avoided.  Just  before  the  rabbit  dies  when  completely 


THERAPEUTIC  IMMUNIZATION  IN  MAN  493 

paralyzed  it  is  killed  by  chloroform  and  the  cord  is  removed  best  by 
the  method  of  Oschida.100  The  rabbit  is  nailed  to  a  board,  back  up- 
permost, and  washed  with  a  weak  antiseptic ;  a  longitudinal  incision 
is  then  made  along  the  backbone  from  the  occiput  to  the  lumbar 
region,  and  the  vertebral  column  laid  bare. 

After  searing  the  tissues  around  the  back  of  the  head  the  spine  is 
cut  across  just  behind  the  occiput,  and  again  in  the  same  way  just 
above  the  sacrum.  The  neck  and  lumbar  regions  are  dissected  loose 
from  the  skin  and  gauze  is  inserted  under  it  to  avoid  contamination. 
The  assistant  grasps  the  end  of  the  spinal  cord  as  it  appears  in  the 
cervical  region  and  pulls  on  it  very  slightly  while  the  operator  with 
a  glass  rod  or  a  piece  of  wire  pushes  against  it  from  below.  If  this 
is  carefully  done  the  spinal  nerves  are  torn  and  the  cord  can  be 
gradually  pulled  out  of  place.  This  procedure  is  by  far  the  best, 
although  it  requires  a  certain  amount  of  practice. 

The  cords  so  removed  are  hung  up  by  a  thread  in  bottles  contain- 
ing sticks  of  caustic  potash  and  exposed  in  a  dark  place  to  22°  to  23° 
C.  Under  these  conditions  of  drying  and  temperature  the  virus  is 
gradually  attenuated  until  at  the  end  of  13  days  or  more  the  viru- 
lence is  practically  nil.  If  removed  from  the  drying  bottles  at  any 
time  during  the  process  and  kept  in  a  refrigerator  in  sterile  glycerin 
the  virulence,  whatever  it  may  be  at  the  time  of  placing  into  the 
glycerin,  remains  fairly  constant  for  long  periods.  When  any  of 
this  material  is  used  for  treatment  little  pieces  of  the  cord  %  cm-  in 
length  are  cut  off  and  emulsified  in  2.5  c.  c.  of  salt  solution,  and  this 
emulsion  is  used  for  injection.101 

When  patients  are  to  be  treated  the  principle  of  the  treatment  is 
to  inoculate  them  first  with  cords  that  have  been  dried  for  consider- 
able periods,  gradually  proceeding  toward  those  that  have  been  dried 
for  less  prolonged  times  and  are  therefore  more  virulent.  The  treat- 
ment is  varied  in  the  individual  case  according  to  the  severity  of  the 
injury.  Formerly  treatment  was  begun  with  cords  dried  as  long  as 
16  days.  More  recently  it  has  been  found  that  cords  dried  for  longer 
than  8  days  are  practically  non-virulent  and  correspondingly  lack 
in  antigenic  value.  They  are  no  longer  employed,  therefore,  since 
their  use  is  regarded  as  a  waste  of  time.  The  tables  on  p.  494,  taken 
from  Stimson's  article  in  Bulletin  65  of  the  Hygienic  Laboratory  of 
the  U.  S.  Public  Health  Service,  give  the  standard  methods  of  treat- 
ment as  recommended  by  the  United  States  Public  Health  Service. 

This  is  the  standard  treatment  used  almost  everywhere  in  the 

100  Osehida.     Centralbl  f.  Bakt.,  Vol.  29,  1901. 

101  In  our  description  of  the  methods  of  drying  rabies,  for  the  sake  of 
adhering  to  a  standard,  we  follow  closely  the  directions  laid  down  by  A.  M. 
Stimson,  in  the  U.  S.  P.  H.  S.  Bull  65,  1910.     There  are  various  modifica- 
tions used  in  different  countries,  in  many  cases  unimportant,  and  it  seems 
well  to  adhere  to  the  U.  S.  regulations  as  a  standard  for  this  country. 


494 


INFECTION  AND  RESISTANCE 


Scheme  for  Mild  Treatment 


Amount  injected 

Amount  injected 

T\-  _„ 

Cord 

Cord 

uay 

(injections) 

Adult 

5-10 

1-5 

Day 

(injections) 

Adult 

5-10 

1-5 

(c.  c.) 

yrs. 
(c.  c.) 

yrs. 
(c.  c.) 

(c.  c.) 

yrs. 
(c.  c.) 

yra. 
(c.  c.) 

i 

8-7-6  =  3 

2.5 

2.5 

2.0 

12 

4=1 

2.5 

2.5 

2.5 

2 

5-4  =  2 

2.5 

2.5 

1.5 

13 

4  = 

2.5 

2.5 

2.5 

3 

4-3  =  2 

2.5 

2.5 

2.0 

14 

3  = 

2.5 

2.5 

2.0 

4 

5  =  1 

2.5 

2.5 

2.5 

15 

3  = 

2.5 

2.5 

2.0 

5 

4  =  1 

2.5 

2.5 

2.5 

16 

2  = 

2.5 

2.0 

1.5 

6 

3  =  1 

2.5 

2.5 

2.0 

17 

2  = 

2.5 

2.0 

1.5 

7 

3  =  1 

2.5 

2.5 

2.0 

18 

4  = 

2.5 

2.5 

2.5 

8 

2=1 

2.5 

1.5 

1.0 

19 

3  = 

2.5 

2.5 

2.5 

9 

2  =  1 

2.5 

2.0 

1.5 

20 

2  =  1 

2.5 

2.5 

2.0 

10 

5  =  1 

2.5 

2.5 

2.5 

21 

2=1 

2.5 

2.5 

2.0 

11 

5  =  1 

2.5 

2.5 

2.5 

Scheme  for  Intensive  Treatment 


Amount  injected 

Amount  injected 

Cord 

Cord 

Day 

(injections) 

5-10 

1-5 

Day 

(injections) 

5-10 

1-5 

Adult 

yrs. 

yrs. 

Adult 

yrs. 

yrs. 

(c.  c.) 

(c.  c.) 

(c.  c.) 

(c.  c.) 

(c.  c.) 

(c.c.) 

1 

8-7-6  =  3 

2.5 

2.5 

2.5 

12 

3  =  1 

2.5 

2.5 

2.0 

2 

4-3  =  2 

2.5 

2.5 

2.0 

13 

3  =  1 

2.5 

2.5 

2.0 

3 

5-4=2 

2.5 

2.5 

2.5 

14 

2=1 

2.5 

2.5 

2.0 

4 

3  =  1 

2.5 

2.5 

2.0 

15 

2  =  1 

2.5 

2.5 

2.0 

5 

3  =  1 

2.5 

2.5 

2.0 

16 

4  =  1 

2.5 

2.5 

2.5 

6 

2  =  1 

2.5 

2.0 

1.5 

17 

3  =  1 

2.5 

2.5 

2.5 

7 

2  =  1 

2.5 

2.5 

2.0 

18 

2  =  1 

2.5 

2.5 

2.0 

8 

1  =  1 

2.5 

1.5 

1.0 

19 

3  =  1 

2.5 

2.5 

2.0 

9 

5=1 

2.5 

2.5 

2.5 

20 

2=1 

2.5 

2.5 

2.5 

10 

4=1 

2.5 

2.5 

2.5 

21 

1  =  1 

2.5 

2.5 

2.0 

11 

4  =  1 

2.5 

2.5 

2.5 

world  at  present.  Other  methods  have  been  recommended.  One  of 
these  is  that  of  Hogyes,  in  which  virus  fixe  unattenuated  is  used  in 
dilution.  Hogyes  begins  by  injecting  3  c.  c.  of  a  1  to  10,000  dilu- 
tion of  virus  fixe,  gradually  proceeding  within  14  days  to  1  c.  c. 
of  a  1  to  100  dilution. 

Fixed  virus  attenuated  by  the  addition  of  antirabic  serum  and 
chemical  disinfectants  (carbolic  acid)  and  by  partial  digestion  in 
gastric  juice  has  also  been  used,  but  none  of  these  methods  has 
attained  widespread  application. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  495 


INFECTION  AND  IMMUNITY  IN  POLIOMYELITIS 

The  active  experimental  investigation  of  poliomyelitis  started 
with  the  discovery  by  Landsteiner  and  Popper,102  that  monkeys 
could  be  inoculated  with  this  disease  by  the  intracerebral  or  the 
intraperitoneal  injection  of  saline  emulsions  of  brain  or  spinal  cord 
of  individuals  dead  of  the  disease.  The  same  observations  by  Flex- 
ner  and  Lewis  103  a  little  later,  and  then  by  a  large  number  of  work- 
ers throughout  the  world,  gave  an  opportunity  for  the  careful  study 
of  immunological  conditions,  and  of  the  nature  of  the  virus.  It  is 
our  own  opinion  that  the  disease  is  caused  by  the  globoid  bodies  of 
j^oguchi  and  Flexner  104  and  not  by  streptococci.  However,  the  work 
on  this  controversy  may  be  regarded  as  unfinished  at  the  present 
writing,  and  we  may  abstain  from  a  prolongued  discussion  of  the 
etiological  factor.  The  mere  fact  that  the  virus  of  poliomyelitis  has 
been  found  to  keep  as  long  as  four  to  six  years  in  glycerinated  ner- 
vous tissue,  and  the  analogy  which  this  offers  to  such  virus  as  that 
of  rabies,  small-pox,  etc.,  make  it  alone  seem  likely  that  true  bac- 
teria are  not  concerned  in  the  cause  of  the  disease.  Our  own  ex- 
periments, and  those  of  Dr.  Tsen  of  this  laboratory,  on  the  isolation 
of  streptococci  from  poliomyelitis  animals,  incline  us  to  think  that 
animals  afflicted  with  this  disease  are  readily  subject  to  secondary 
tissue  infection  with  organisms  of  various  kinds,  but  chiefly  with 
streptococci  of  the  viridans  type,  which  are  so  universally  distributed 
throughout  the  body. 

It  has  long  been  suspected  that  one  attack  of  poliomyelitis  pro- 
tected from  subsequent  infection.  It  is  plain  therefore  that  some 
form  of  active  immunization  takes  place  in  the  course  of  the  disease. 
Experimentation  on  monkeys  subsequently  confirmed  this,  in  that  it 
was  shown  that  monkeys  that  had  contracted  the  disease,  and  re- 
covered, were  thereafter  resistant  to  inoculation.  Strangely  enough, 
however,  monkeys  that  have  been  unsuccessfully  inoculated  are  just 
as  susceptible  as  they  were  before,  showing  that  the  immunity  in 
this  disease  is  closely  analogous  to  that  existing  in  syphilis  and  some 
other  diseases  where  inoculation  with  attenuated  virus,  dead  virus, 
or  sub-infectious  doses  of  living  virus,  is  entirely  incapable  of  pro- 
ducing immunity. 

A  brief  survey  of  the  epidemiology  of  the  disease  will  be  of  value 
because  it  throws  some  light  on  questions  of  immunity.  An  excel- 
lent and  extensive  summary  of  this  may  be  found  in  the  Mono- 
graph on  the  disease  published  by  the  New  York  Department  of 

102  Landsteiner  and  Popper.    Zeit.  f.  Immunitats.,  11,  1909. 

103  Flexner  and  Lewis.     Journ.  Amer.  Med.  Assn.,  1910,  liv.  45. 

104  Flexner  and  Noguchi.    Jour.  Exp.  Med.,  1913,  XVIII,  461. 


496  INFECTION  AND  RESISTANCE 

Health,  in  1917,  on  the  basis  of  the  work  done  during  the  great 
epidemic  of  1916. 

Seasonably  the  disease  has  almost  invariably  been  one  of  hot 
weather.  There  are  a  few  exceptions  to  this,  one  of  them  a  great 
epidemic  in  Sweden  in  1911,  which  was  at  its  height  from  October 
to  December.  Some  others  have  begun  as  early  as  March. 

Formerly  the  disease  was  regarded  as  chiefly  rural,  but  of  late 
years  it  has  also  extensively  prevailed  in  cities. 

The  age  incidence  is  important,  in  that  the  majority  of  cases 
(about  90  per  cent)  have  occurred  in  infants  under  ten,  although  a 
few  epidemics  among  adults  have  been  observed  and  a  few  adult  cases 
always  accompany  epidemics.  In  other  people  the  disease  often 
takes  the  form  of  what  was  formerly  known  as  Landry 's  paralysis. 
Indeed,  in  a  little  epidemic  which  we  studied  in  Palo  Alto,  Cal., 
the  first  cases  which  died  were  adults,  with  typical  Landry  symp- 
toms; and  successful  monkey  inoculation  was  made  from  a  woman 
of  26  whose  symptoms  were  those  of  a  typical  Landry  paralysis. 

Relation  to  domestic  animals  has  often  been  suggested,  and 
Frost  (U.  8.  Hygienic  Laboratory  Bulletin  90,  1916),  who  studied 
this  extensively  during  the  last  epidemic,  made  careful  observations 
on  this  factor,  finding  that  there  is  no  demonstrable  relationship. 
The  diseases  of  dogs  and  chickens  investigated  by  others,  which 
simulated  the  human  disease,  have  been  found  to  be  due  to  distinct 
pathological  conditions  without  any  relationship  to  human  poliomye- 
litis. A  paralytic  disease  of  cats  occurring  in  New  York  during  the 
last  poliomyelitis  epidemic  was  studied  in  our  laboratory  by  Dr. 
Messing  and  found  to  have  been  caused  by  a  Gram  negative  bacillus. 

Because  of  the  seasonable  occurrence  insects  were  long  suspected 
as  possible  carriers,  and  M.  J.  Rosenau  a  few  years  ago  carried  out 
experiments  in  a  few  of  which  he  succeeded  in  infecting  monkeys 
by  the  bites  of  stable  flies  (Stomoxys)  which  had  been  fed  on  the 
virus.  There  has  been  but  a  single  confirmation  of  this  by  Ander- 
son, and  all  other  attempts  to  confirm  Rosenau' s  observation  have 
been  negative.  Investigations  of  other  flies,  or  bed-bugs,  fleas,  mos- 
quitoes, etc.,  have  invariably  been  negative.  While  therefore  stable 
fly  transmission  cannot  be  accepted  as  a  common  method  of  infec- 
tion, and  probably  plays  no  practical  role  whatever,  we  must  bear 
in  mind  that  there  have  been  two  successful  experiments  in  the  work 
of  two  reliable  and  skilled  experimenters,  and  should  be  alert  for 
further  evidence  in  this  direction. 

Important  positive  knowledge  concerning  the  transmission  of 
the  disease  is  that  which  has  been  gained  by  the  investigations  of  so- 
called  "carriers."  It  has  been  found  both  in  monkeys  and  human 
beings  that  the  virus  of  the  disease  can  be  carried  in  the  upper 
respiratory  passages  of  convalescents  for  as  long  as  months  after 
recovery.  It  has  also  been  found  in  a  single  instance  that  a  monkey 


THERAPEUTIC  IMMUNIZATION  IN  MAN  497 

could  be  successfully  inoculated  with  dust  from  a  room  in  which  a 
patient  lay ;  and  Josef  son  states  that  infection  may  be  acquired  from 
handkerchiefs  and  other  recently  handled  articles  for  a  short  time. 
Under  these  conditions  it  would  seem  rather  likely  that  the  greatest 
factor  in  transmission  would  be  found  in  direct  and  indirect  trans- 
mission by  means  of  carriers,  themselves  immune  or  recently  re- 
covered, harboring  the  virus  and  transmitting  it  to  susceptible  in- 
dividuals. And  it  is  possible  that  many  individuals  in  a  community 
may  be  immune,  owing  to  mild  forms  of  the  disease  sustained  in 
childhood  which  were  not  recognized  as  poliomyelitis;  for  evidence 
is  constantly  accumulating  that  poliomyelitis  is  not  essentially  a 
paralytic  disease,  that  paralysis  is  merely  one  of  the  severe  symp- 
toms, and  that  mild  but  immunizing  attacks  may  occur  in  the  form 
of  very  slight  indispositions,  attacks  of  mild  grippe,  headache,  etc., 
which  under  present  conditions  are  not  recognized  as  poliomyelitis. 
On  this  basis  it  would  be  comprehensible  why  so  often  one  child  of  a 
family  is  afflicted  while  others  equally  exposed  are  spared.  And  it 
would  become  plain  why  epidemics  sweep  through  a  community  se- 
lecting cases  irregularly  here  and  there,  and  then  die  out,  leaving 
behind  an  endemic  smoulder  of  sporadic  cases.  It  would  seem  as 
though  for  the  time  being  all  the  susceptible  children  had  been 
afflicted  and  no  further  fuel  for  spread  was  available. 

Under  these  conditions  the  study  of  contact  infection  becomes 
very  important,  and  Frost  concludes  that  none  of  the  other  factors 
usually  considered  in  the  spread  of  an  infection,  other  than  personal 
contact,  seems  to  offer  adequate  explanation  for  the  spread  of  epi- 
demics. To  some  extent  discouraging  as  far  as  the  contact  theory 
is  concerned,  the  epidemic  in  New  York  showed  that  in  96  per  cent  of 
the  families  afflicted,  that  is,  in  8,287  families,  there  was  one  case 
only  to  the  family,  although  in  these  families  there  were  a  total  of 
24,883  children.  In  only  3.6  per  cent  were  there  two  cases  to  the 
family.  Just  why  this  should  be  it  is  difficult  to  surmise.  However, 
to  some  extent  this  and  other  radical  phenomena  in  the  transmission 
of  this  disease  may  depend  upon  differences  in  susceptibility,  due  per- 
haps to  a  light  unrecognized  attack  previously  sustained,  or  to  some 
anatomic  or  physiological  predisposition  to  infection  of  the  meningeal 
choroid  plexus.  In  this  connection  recent  studies  by  Flexner  and 
Amoss 105  are  of  unusual  interest.  They  find  that  among  the 
mechanisms  that  defend  the  central  nervous  system  from  infection  is 
this  plexus,  which,  if  in  any  way  injured,  becomes  unable  to  function- 
ate, and  in  consequence  permits  infection  to  pass.  The  injection  of 
normal  monkey  or  horse  serum  of  isotonic  salt  solution,  or  of  Ringer's 
and  Locke's  solutions,  when  injected  into  the  meninges,  though  in 
itself  innocuous,  causes  a  sufficient  injury  of  the  membrane  to 
render  a  subsequently  injected  lethal  dose  of  virus  infectious. 
105  Flexner  and  Amoss.  Jour.  Exp.  Med.,  1914,  XX,  249. 


498  INFECTION  AND  RESISTANCE 

Early  in  the  study  of  poliomyelitis  it  was  found  that  one  at- 
tack protected  and  that  the  serum  of  a  human  being  or  monkey 
that  had  recovered  from  the  disease  was  capable  of  neutralizing  the 
virus.  In  consequence  extensive  studies  have  been  made  by  Flex- 
ner,  Amoss,  and  Chesney,  Draper,  Zingher106  and  others  in  wlrich 
the  serum  taken  from  recently  recovered  cases  was  injected  intra- 
spinously  into  children  suffering  from  the  disease  for  therapeutic 
purposes.  The  work  on  this  problem  is  not  yet  complete,  but  it 
seems,  especially  from  the  studies  of  Amoss  and  Chesney,107  that 
when  injected  early  in  the  disease  a  very  definite  favorable  influence 
can  be  expected. 

The  serum  should  be  taken  from  cases  which  have  recovered  as 
recently  as  possible  and  collected  with  ordinary  precautions  of 
asepsis.  It  should  then  be  injected — best  in  quantities  not  less  than 
50  c.c.,  part  of  which  is  given  intraspinously,  and  the  rest  intra- 
venously, to  neutralize  any  circulating  virus. 


IMMUNITY  IN  SYPHILIS 

Until  relatively  recent  years  there  seems  to  have  been  little  ques- 
tion in  the  mind  of  clinicians  regarding  the  existence  of  true  im- 
munity in  syphilis.  It  was  stated  uncompromisingly  by  Ricord  108 
that  "an  individual  who  had  once  acquired  syphilis  was  thereafter 
protected  against  reinfection.7'  This  opinion  acquired  wide  accep- 
tance and  was  shared  by  most  of  his  contemporaries.  Baumler  109  in 
1875  summarizing  the  authoritative  opinions  of  that  period  stated 
that  "One  who  has  once  had  small-pox,  scarlet  fever,  typhus,  etc., 
is,  as  a  rule,  not  liable  to  these  diseases  again  for  the  rest  of  his  life. 
The  same  is  true  of  syphilis."  However,  even  at  the  time  Baumler 
wrote  this,  exceptions  to  the  supposed  rule  were  accumulating  and 
he  appends,  to  the  positive  statement  given  above,  references  to 
observed  instances  of  second  infection  reported  by  Bidenkap,110  H. 
Lee,111  Diday,112  Kober,113  Zeissl,114  and  others. 

The  conception  of  the  existence  of  a  true  acquired  immunity  was 
also  expressed  and  widely  accepted  in  the  so-called  "laws"  of  Colles 
and  Prof  eta.  The  former,  first  enunciated  by  Beaumes  and  later, 

106  Zingher.    Jour.  Amer.  Med.  Assn.,  March,  1917. 

107  Amoss  and  Chesney.    Jour.  Exp.  Med.,  April,  1917. 

108  Ricord.     Eecherches  sur  le  Chancre,  Paris,  1858,  cited  from  Baumler. 

109  Baumler.     Ziemssen's  Cyclop,  of  Pract.  Med.,  Amer.  Ed.,  Wm.  Wood 
and  Co.,  N.  Y.,  1875,  iii. 

110  Bidenkap.     Wien.  med.  Wchnschr.,  1865,  cited  from  Baumler, 

111  Lee,  H.    Lecture  on  Syphilis,  London,  1863. 

112  Diday.    Arch,  gener.,  1862,  ii,  cited  from  Baumler. 

113  Kober.    Berl.  klin.  Wchnschr.,  1872,  No.  46. 

114  Zeissl.     Ztschr.  d.  K.  K.  Gesell.  d.  Aertze  in  Wien.,  1858. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  499 

in  1837,  stated  by  Abraham  Colles,115  of  Dublin,  is  the  well-known 
generalization  based  on  the  observation  that  mothers  who  have  borne 
syphilitic  infants  were  not  infected  by  their  children  while  suckling 
them,  although  such  children  might  often  infect  wet  nurses.  Pro- 
f eta's  116  observation  was  the  converse  of  this,  namely,  that  children 
born  of  mothers  who  suffered  from  active  syphilis  during  the  period 
of  conception  did  not  acquire  the  disease  from  their  mothers. 

Both  of  these  phenomena  appear  too  well  founded  in  clinical 
observation  to  be  questioned,  at  least  as  frequent  occurrences.  More- 
over, considering  the  intimate  contact  between  mother  and  infant 
during  the  first  months  after  birth,  they  acquire  unusual  importance. 
However,  as  we  shall  see,  they  have  been  deprived  of  much  of  their 
bearing  as  proofs  of  true  acquired  or  inherited  immunity  by  serologi- 
cal  investigations  such  as  those  of  Bauer,117  Knopf elmacher,118  and 
others,  which  have  shown  that  mothers  of  syphilitic  children  usually 
give  positive  Wassermann  reactions,  a  fact  which  makes  it  seem 
likely  that  such  women  are  suffering  from  syphilis  in  a  latent  form 
and  are  not  immune  in  the  ordinary  sense.  This  indeed  was  the 
interpretation  given  to  the  "laws"  of  Colles  and  Profeta  by 
Fournier  119  and  by  Matzenauer  12°  at  a  time  prior  to  that  at  which 
serological  data  were  available.  It  is  still  unclear  why  such  mothers 
should  so  frequently  exhibit  the  disease  in  a  latent  form.  However, 
this  is  a  matter  for  the  intelligent  discussion  of  which  we  are  not 
at  the  present  time  in  the  possession  of  sufficient  information. 

In  the  years  just  preceding  the  period  of  experimental  investi- 
gation of  syphilis  upon  animals  much  purely  clinical  research  was 
carried  out  on  the  problem  of  reinoculation  and  what  is  commonly 
known  as  "superinfection"  of  syphilitic  human  beings.  Much  of  this 
work  is  unavailable  as  scientific  evidence  owing  to  the  difficulties  of 
distinguishing,  at  that  time,  between  the  true  chancre  and  the 
chancroid,  but  a  considerable  number  of  the  observations  then  made 
have  been  of  much  value  in  pointing  out  directions  of  later  research 
upon  animals.  The  work  has  been  so  thoroughly  analyzed  both  by 
Neisser  and  by  Levaditi  that  it  would  be  needless  repetition  to  do  so 
again.  The  records  include  both  accidentally  occurring  superinfec- 
tions,  and  purposeful  experimental  reinoculations.  Without,  there- 
fore, going  into  details  concerning  the  individual  cases  we  may  sum- 
marize the  conclusions  justified  from  a  study  of  these  observations. 

115  Colles.    Cited  from  Osier  &  Churchman,  "Syphilis,"  Osier's  System  of 
Medicine. 

116  Profeta.     Cited   from   Osier  and   Churchmann,  "Syphilis."     Trattato 
Practico  delle  Malattie  Veneree,  Palermo,  1888. 

117  Bauer.     Wien.  klin.  Wchnschr.,  1908,  No.  28. 

118  Knopf  elmacher  and  Lehndorff.    Med.  klin.,  1909,  No.  40;  Wien.  med, 
Wchnschr.,  1909,  No.  38. 

119  Fournier.     L'Heredite  Syphilitique,  Paris,  1890. 

120  Matzenauer.     Vererbung  d.  Syphilis-Wien.,  1905,  cited  from  Bruck. 


500  INFECTION  AND  RESISTANCE 

1.  The  reports  of  Lasch,121  Jadassohn,122  Sabeareanu,123 
Queyrat,124  Taylor/25  H.  Lee,  Knowles,126  and  many  others,  have 
shown  that  patients  are  susceptible  to  a  second  inoculation  during 
the  first  incubation  time,  that  is,  during  the  period  elapsing  between 
the  first  infection  with  syphilis  and  the  appearance  of  the  chancre. 
Second  positive  inoculations  have  also  been  successful  at  periods 
shortly  subsequent  to  the  appearance  of  the  primary  sore. 

Autoinoculations  and  reinoculations  undertaken  after  the  chancre 
has  become  well  developed  have  been,  in  the  main,  negative,  though 
Queyrat  reports  a  case  successfully  inoculated  daily  up  to  the 
eleventh,  and  Taylor  one  inoculated  on  the  fourteenth  day  after  the 
appearance  of  the  primary  induration.  In  contrast  to  these  and  other 
successful  attempts,  many  observers  record  failure.  However,  the 
point  is  not  one  of  particular  importance,  since,  after  all,  the  ap- 
pearance of  the  chancre  does  not  mark  off  any  fundamental  change 
in  the  progressive  pathological  development  of  the  disease,  and  indi- 
cates only  the  completed  reaction  at  the  point  of  entrance.  The  fact 
remains  that  analysis  has  revealed  that  reinoculation,  with  the  ap- 
pearance of  the  second  initial  lesion,  is  possible  up  to  about  the  twen- 
tieth to  the  thirtieth  day  after  the  first  infection  with  the  trepone- 
mata  (Mauriac 127  and  Sfeisser  state  the  twenty-second  day,  Queyrat 
cites  a  case  eleven  days,  Linderman  one  twenty-four  days  after  the 
appearance  of  the  chancre).  It  is  claimed  by  some  of  the  observers, 
however,  that  even  when  reinoculation  during  this  period  is  success- 
ful, the  incubation  of  the  second  and  subsequent  lesions  is  shorter, 
that  the  induration  itself  is  less  severe,  may  not  ulcerate,  and  heals 
more  readily  than  the  first. 

In  judging  of  the  success  or  failure  of  reinoculation  practised 
during  the  later  days  of  the  period  above  referred  to,  the  possibility 
must  be  borne  in  mind  that  the  trauma  produced  at  the  inoculation 
might  have  served  to  favor  the  development  of  a  localized  focus,  the 
treponemata  at  this  time  being  very  probably  well  distributed 
throughout  the  body.  Unsuccessful  control  inoculations  with  non- 
syphilitic  materials  in  Quey rat's  cases  would  tend  to  eliminate  this 
possibility,  whereas  lesions  resulting  at  the  sites  of  such  control 
abrasions  in  the  experiments  of  Neuman  and  Cehak,128  and  of 

121Lascli.    Arch.  f.  Dermat.  u.  Syph.,  1891,  61. 

122  Jadassohn.    Arch.  f.  Dermat.  u.  Syph.,  1907,  86 ;  Festschr.  f.  Neisser, 
1907. 

123  Sabeareanu.     These  de  Paris,  1905,  Chancres  Syph.  Successif. 

124  Queyrat.    Bull,  de  la  Soc.  Med.  des.  Hop.,  1904,  No.  28;  905. 

125  Taylor.    Jour.  Cutan.  Dis.,  Dec.,  1890. 

126  Knowles.     New  York  Med.  Jour.,  Dec.,  1906. 

127  Mauriac.     Cited  from  Neisser,  loc.  cit. 

128Neuman  and  Cehac.     Wien.  med.  BL,  1890,  cited  from  Neisser. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  501 

Levaditi,129  appear  to  support  it.  However,  this  is  not  of  great  im- 
portance inasmuch  as  it  determines  merely  a  relative  lengthening 
or  shortening  of  the  period  during  which  reinoculation  or  superin- 
fection  is  possible. 

2.  After  the  disease  is  well  established  as  a  systemic  infection, 
that  is,  from  the  time  of  development  of  the  chancre  throughout  the 
so-called  active  "secondary"  period,  reinoculation  is  either  impos- 
sible or,  at  any  rate,  extremely  difficult.  Neisser  cites  Rollet  as 
follows : 

"Although  I  and  my  predecessors  have  a  thousand  times  attempted 
to  reinoculate  luetic  subjects,  we  have  never  observed  a  successful 
case.  I  know  no  single  fact  more  thoroughly  proven  than  the  insus- 
ceptibility of  a  syphilitic  to  the  action  of  a  new  virus,  and,  more- 
over, these  experiments  are  so  harmless  that  they  may  be  performed 
without  scruple." 

The  same  opinion  was  held  by  Mauriac  and  is  in  a  general  way 
assented  to  by  Neisser  in  his  summary  of  this  place  of  his  studies. 
(Neisser,  loc.  cit.,  pp.  180-181.) 

That  the  patient  with  well-developed  lues  has  acquired  a  con- 
siderable degree  of  resistance  to  fresh  inoculations  is  pretty  gener- 
ally accepted,  therefore,  by  all  who  have  studied  the  question — but 
there  are  investigators,  notably  Finger  and  Landsteiner,  who  believe 
this  resistance  to  be  less  absolute  than  stated  by  earlier  writers. 
Their  observations  on  patients  with  active  syphilis  seem  to  indicate 
that  superinfection  is  possible  "under  certain  circumstances  in  all 
stages  of  the  disease,"  but  "the  positive  effect  can  be  obtained  only 
with  considerable  quantities  of  the  virus."  Furthermore,  Land- 
steiner 13°  states  that,  in  general,  lesions  so  obtained  are  relatively 
slight  in  severity,  do  not  appear  as  primary  indurations,  but  have 
the  tendency  to  simulate  the  particular  variety  of  lesion  spontane- 
ously manifest  in  the  individual  at  the  time.  Were  it  not  for  Finger 
and  Landsteiner  7s  monkey  experiments,  to  which  we  will  refer  di- 
rectly, and  which  seem  to  bear  them  out  in  their  interpretation  of 
these  inoculations  as  "positive"  results,  the  simulation  of  the  spon- 
taneously occurring  lesions  by  the  inoculation-products  would  again 
justify  suspicion  that  these  experimental  results  represented  merely 
traumata  in  which,  as  points  of  less  resistance,  the  patient's  pre- 
existent  disease  had  found  a  favorable  spot  for  localization. 

Observations  in  this  respect,  also,  are  corroborated  by  the  older 
literature.  A  report  which  has  direct  bearing  on  this  point,  is  one 
oif  Queyrat  and  Pinard  who  inoculated  a  tertiary  patient  with  chan- 
cre material,  obtaining  not  a  primary  sore  but  an  ulcerated  lesion 
having  the  clinical  characteristics  typical  of  the  late  skin  manifesta- 

129Levaditi.  De  I'lmmunite  acquise  dans  la  Syphilis.  Ztschr.  f.  Imm. 
Ref.,  191,  277-318. 

130  Landsteiner.     Centralbl  /.  Bakteriol,  Ref.  190S.  xli,  785. 


502  INFECTION  AND  RESISTANCE 

tions  of  the  disease.  The  autoinoculation  experiments  of  Ehr- 
mann 131  also  would  tend  to  show  that  the  resistance  of  the  luetic 
subject  is  a  relative  one  only.  Ehrmann  succeeded  in  producing 
positive  autoinoculation-products  in  45  syphilitics  with  papular  erup- 
tions. Control  inoculations  with  sterile  water  were  negative.  Al- 
though his  observations  and  those  of  Finger  and  Landsteiner,  with 
other  similar  ones,  teach  us  that  superinfection  during  the  disease 
is  possible  the  nature  of  the  lesions,  their  short  incubation  time,  and 
their  exceptional  character  when  averaged  with  the  total  of  such  at- 
tempts, prevent  them  from  invalidating  the  conclusion  that  there  is 
a  resistance  at  this  stage  higher  than  that  of  the  normal  subject. 

We  may  conclude,  therefore,  concerning  the  secondary  period  of 
the  disease,  that  the  luetic  individual  has  acquired  a  resistance  which 
while  not  absolute  is  at  least  very  high,  and  protects  him  from  fresh 
external  inoculation,  although  at  the  time  his  disease  may  still  be 
progressing  and  in  no  sense  overcome. 

3.  During  the  late  stage  of  syphilis,  the  stage  at  which,  according 
to  the  more  or  less  arbitrary  divisions  of  Ricord,  we  are  accustomed 
to  speak  of  it  as  "tertiary,"  the  resistance  is  still  manifest,  though 
apparently  not  so  regularly  potent  as  during  the  preceding  "secon- 
dary" period.  JSTeisser  expresses  himself  with  great  caution  and 
accepts  few,  if  any,  of  the  observed  reinoculations  of  tertiary  cases 
as  surely  representing  unquestionable  freshly  acquired  infections. 
Nevertheless,  taking  into  consideration  a  detailed  study  of  individual 
reports,  he  concludes  that  resistance  during  the  late  stages  is  pro- 
nounced but  already  beginning  to  wane.  He  himself  in  the  Fest- 
schrift to  F.  Joseph  Pick  in  1898  cites  a  case  of  the  development  of 
a  chancre  in  a  tertiary  case  which,  however,  was  not  followed  by 
constitutional  symptoms. 

r-  4.  In  the  preceding  paragraphs  we  have  concerned  ourselves 
entirely  with  questions  of  "superinfection,"  that  is,  the  implantation 
of  the  syphilitic  virus  into  subjects  still  suffering  from  manifesta- 
tions of  the  disease.  A  problem  of  equal  theoretical,  and  of  much 
greater  practical  importance,  is  that  dealing  with  true  "reinfection." 
.By  "immunity"  in  the  ordinary  sense,  we  mean  an  increased  resist- 
ance to  specific  infection  which  persists  for  a  more  or  less  prolonged 
period  after  the  active  disease  has  been  overcome  and  the  causative 
agents  removed  from  the  body.  It  is  not  easy  to  draw  conclusions 
-concerning  this  point  from  observations  on  human  beings,  since  it 
is  most  difficult  to  decide,  even  with  the  aid  of  serological  methods, 
whether  or  not  a  given  case  is  cured  in  the  bacteriological  sense; 
for  syphilis  is  preeminently  the  disease  in  which  there  occur  fre- 
quent and  prolonged  latent  periods  terminated,  often  after  lapses  of 
years,  by  the  reappearance  of  foci  of  a  grave  nature. 

Moreover,  our  own  experiments  both  with  syphilis  in  rabbits  and 

131  Ehrmann.    Cited  from  Neisser,  loc.  cit. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  503 

the  treponemata  in  culture  have  convinced  us  that  these  micro- 
organisms may  assume  for  long  periods  a  condition  of  metabolic 
latency,  a  sort  of  resting  period,  during  which  they  incite  no  reactions 
of  any  kind  on  the  part  of  the  tissues,  apparently  do  not  multiply 
to  any  great  extent  and  yet  remain  alive  and  capable  of  development 
when  conditions  favor  this. 

In  spite  of  these  difficulties,  however,  careful  clinical  studies  by 
Jonathan  Hutchinson,  Taylor,  Hudelo,132  Neisser,  and  many  others, 
have  furnished  data  which  warrant  an  opinion  upon  this  problem. 
Of  especial  value  is  the  painstaking  analysis  of  reported  cases  of  re- 
infection in  syphilis  made  in  1909  by  Felix  John.133 

In  contrast  with  some  others  who  have  attempted  similar  analyses 
John  takes  the  utmost  care  to  separate  these  cases  into  the  ones  in 
which  the  evidence  of  true  reinfection  is  absolute,  and  those  in  which 
the  reported  details  are  insufficient  to  exclude  the  possibility  of 
recrudescence.  In  agreement  with  Taylor  he  insists  upon  a  symptom- 
free  interval  of  five  years  between  the  last  manifestations  of  the 
first  attack  and  the  appearance  of  the  second.  As  an  example  of 
what  he  calls  an  "Ideal  Fall"  we  may  cite  the  following: 

X.  J.,  April  1,  1872,  primary  sore. 

Roseola,  polyadenitis,  mucous  patches.  September,  1872,  papular  rash. 
March,  1873,  palmar  and  plantar  spots,  iritis.  1875,  gumma  of  tibia  and 
serpiginous  syphilide  of  right  thigh. 

Four  courses  of  inunctions  and  KI. 

1876,  married — two  healthy  children. 

No  symptoms  till  1887  when  he  acquired  a  second  chancre  followed  by 
typical  roseola.  In  1888  wife  had  still-birth. 

John  has  analyzed  in  this  way  356  cases  of  supposed  reinfection, 
in  34  of  which  the  first  attack  was  of  congenital  origin.  Of  the 
remaining  322,  fourteen  were  cases  which  seemed  unquestionable 
instances  of  reinfection  and  in  16  more  there  was  practically  no 
doubt  of  this.  Of  the  34  hereditary  cases  there  were  three,  one  of 
Emery,  one  of  Taylor,  and  one  of  Hochsinger,  in  which  there  was 
practically  no  question  of  their  nature  as  valid  reinfections.  In  all 
of  the  others  there  was  one  point  or  another  which  rendered  them 
doubtful  as  evidence. 

John  concludes  that  true  reinfection  can  unquestionably  occur 
in  syphilis  but  that  it  is  relatively  rare. 

To  John's  cases  Neisser  has  added  others  reported  between  1909 
and  1911.  Yet  even  with  these,  the  total  number  is  not  a  large  one. 
Nevertheless,  we  should  not  be  tempted  to  conclude  from  this  that 
the  relative  infrequency  of  such  cases  is  evidence  in  favor  of  the 

132  Hudelo.    Ann.  de  dermat.  et  de  syph.,  1890,  353. 

133  John,  Felix.     Eeinfectio  Syphilitica,  Samml.  klin.  Vortr.,  Volkmann, 
1907-1909,  559  et  seq. 


504  INFECTION  AND  RESISTANCE 

existence  of  a  true  immunity  analogous  to  that  following  typhoid, 
plague,  etc.  For  we  have  seen  that  a  very  definite  insusceptibility 
is  coincident  with  the  persistence  of  actual  disease  in  the  subject 
and,  as  Neisser  points  out,  the  more  recent  investigations  carried  out 
with  the  aid  of  serological  tests  have  shown  that  the  number  of  cases 
uncured  though  long  without  symptoms,  is  much  larger  than  for- 
merly supposed.  The  scarcity  of  true  reinfection,  therefore,  may 
well  be  due  to  the  relative  scarcity  of  completely  cured  cases.  More- 
over, it  must  be  remembered  that,  in  this  disease,  even  when  final 
recovery  results,  it  is  usually  achieved  only  at  an  age  when  the  in- 
dividual is  less  exposed  to  reinfection  because  of  changed  economic 
and  domestic  conditions,  or  by  reason  of  the  virtue  which  comes  with 
arteriosclerosis  and  the  wisdom  of  the  burnt  child  that  fears  the  fire. 

Granted,  then,  that  true  reinfection  is  possible,  is  there  any  evi- 
dence that  when  it  does  occur,  the  second  attack  is  less  severe  and 
more  easily  cured  than  the  first,  a  fact  which  would  also  tend  to 
support  the  opinion  that  a  certain  degree  of  immunity  persists. 
Jonathan  Hutchinson  134  expressed  this  view  in  a  clinical  lecture  in 
which  he  states  that  "second  chancres  are  far  more  common  than 
second  attacks  of  constitutional  syphilis."  However,  John  in  his 
summary,  in  which  Hutchinson' s  cases  are  included,  finds  that  in 
general  the  disease  has  run  a  second  course  very  similar  as  to  severity 
to  that  incident  to  the  first  infection. 

If  we  gather  together,  then,  the  facts  revealed  by  clinical  study  we 
may  conclude  with  Levaditi,  Neisser,  and  most  others,  that : 

1.  The  syphilitic  subject  acquires  definite  resistance  to  rein- 
oculation  which  becomes  manifest  soon  after  the  appearance  of  the 
primary  sore,  at  a  time  when  the  virus  may  be  regarded  as  having 
gained  universal  systemic  distribution. 

2.  This  resistance,  high  though  not  absolute,  persists  through- 
out the  secondary  or  most  active  period  of  the  disease  and  into  the 
tertiary  stage.     During  the  latter,  however,  it  appears  somewhat  to 
decline,  reinoculation  or  superinfection  being  more  frequently  pos- 
sible at  this  period. 

3.  When  syphilis  is  entirely  cured,  susceptibility  may  in  all 
probability  be  regarded  as  returning,  possibly,  though  not  certainly, 
to  the  same  degree  as  it  exists  in  the  normal  subject.     The  reasons 
for  this  last  belief  will  become  more  clear  when  we  study  the  evidence 
contributed  by  animal  experimentation. 

Notwithstanding  the  admirable  thoroughness  with  which  clinical 
data  had  been  collected  and  analyzed  in  the  study  of  syphilis,  prog- 
ress beyond  the  points  indicated  in  the  preceding  sections  was  quite 
impossible  without  the  aid  of  animal  experimentation  and  a  knowl- 
edge of  the  causative  agent.  Fortunately  these  two  deficiencies  in 
our  methods  were  removed  when  in  1903  Metchnikoff  and  Roux 

134  Hutchinson,  J.     Brit.  Med.  Jour.,  1882,  i,  699. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  505 

succeeded  in  transmitting  the  disease  to  a  chimpanzee,  and  in  March, 
1905,  Schaudinn135  discovered  the  treponema  pallidum. 

As  a  matter  of  fact,  probable  transmission  of  syphilis  to  lower 
monkeys  had  been  accomplished  as  early  as  1879  by  Klebs  136  and 
subsequently  by  Neumann,137  Martineau,138  and  Charles  Nicolle,139 
but  in  none  of  these  experiments  had  it  been  possible  to  prove  beyond 
question  the  syphilitic  nature  of  the  inoculation-products.  In  the 
chimpanzees  inoculated  by  Metchnikoff  and  Roux,  the  animals  de- 
veloped not  only  primary  sores  but  also  secondary  eruptions,  poly- 
adenitis,  and  enlarged  spleens  in  such  characteristic  manner  that  the 
identity  of  the  inoculation  disease  with  human  syphilis  could  no 
longer  be  doubted. 

Successful  transmission  to  other  anthropoids  and  to  lower  mon- 
keys were  then  announced,  in  rapid  succession,  by  Metchnikoff  and 
Roux,  Ch.  Nicolle,  Neisser,  Baermann  and  Halberstaedter,  Einger 
and  Landsteiner,  jloffman,  and  others.  The  susceptibility  of  mon- 
keys was  tabulated  by  Neisser  in  the  following  series :  Chimpanzee, 
Gibbon,  Orang-Outang,  Cynocephalus  babuin,  Cynocephalus  sphinx, 
Cynocephalus  hamadryas,  Cercopithecus  fulginosus,  Macacus  niger, 
M.  nemestrinus,  M.  cynomolgus,  M.  sinicus,  M.  speciosus,  M.  rhesus. 
In  1906  Bertarelli  14°  produced  syphilitic  keratitis  in  rabbits,  dem- 
onstrating the  treponema  pallidum  in  sections  of  the  cornea  and  in 
1907  Parodi 141  first  produced  syphilitic  orchitis  in  the  same  animals. 

Apart  from  monkeys  and  rabbits,  no  animal  species  have  so  far 
been  shown  sufficiently  susceptible  to  be  available  for  systematic 
study.  It  is  true  that  the  production  of  keratitis  is  claimed  in  dogs 
and  sheep  (Bertarelli,  Hoffman  and  Briinig),  in  guinea  pigs  (Ber- 
tarelli), in  cats  (Levaditi  and  Yamanouchi),  and  in  goats  (Ber- 
tarelli). However,  these  experiments  have  been  isolated  and  too 
uncertain  with  present  methods  to  offer  material  for  experimentation. 
Our  own  attempts  on  cats,  pigs,  guinea  pigs,-  rats,  mice,  and  a  few 
birds,  have  yielded  negative  results  only.  There  are  many  observa- 
tions of  great  scientific  interest  which  might  be  discussed  in  con- 
nection with  the  problem  of  susceptibility  of  various  animal  species, 
however,  we  will  confine  ourselves  at  present  to  those  phases  of  the 
work  only  which  have  bearing  on  the  questions  of  immunity. 

In  the  fundamental  premises,  the  work  on  monkeys  has  pretty 
accurately  confirmed  the  observations  concerning  reinoculation  and 
superinfection  previously  made  on  human  beings. 

135  Schaudinn.     Deutsch.   med.   Wchnschr.,  1905,  No.  42;  Arb.  a.   d.  k. 
Gsndhtsamte.,  1907,  xxvi. 

136  Klebs.     Arch.  f.  exper.  Path.  u.  Pharmakol.,  1879. 

137  Neumann.     Cited  from  Miihlens,  loc.  cit. 

138  Martineau.     Arch.  f.  Dermat.  u.  Syph.,  1884,  No.  16. 

139  Nicolle,  Ch.     Cited  from  Miihlens,  loc.  cit. 

140  Bertarelli.     Centralbl.  f.  Bakteriol,  Orig.  xli,  1906,  and  xliii,  1907. 

141  Parodi.     Centralbl.  f.  Bakteriol.,  Orig.  xliv,  1907. 


506  INFECTION  AND  RESISTANCE 

The  most  extensive  studies  in  this  respect  are  those  of  Neisser  142 
and  his  associates  in  the  Javan  expedition.  Briefly  stated,  Neisser 
reinoculated  135  animals  165  times  with  negative  result 'on  the  second 
and  subsequent  inoculations.  The  second  inoculations  were  made  at 
periods  ranging  from  21  days  to  two  years  after  the  first.  In  but  27 
animals  did  the  reinoculation  show  positive  results,  and  in  ten  of 
these  cases  only  does  Neisser  recognize  the  experiments  as  valid. 
Although  these  ten  positive  reinoculations,  it  is  true,  add  an  element 
of  irregularity  to  the  series,  they  constitute  but  6.8  per  cent,  of  the 
entire  number,  a  proportion  which  in  no  way  invalidates  the  experi- 
ments when  we  consider  that  the  work  was  done  entirely  on  the 
lower  monkeys,  animals  that  are  far  less  susceptible  to  syphilis  than 
are  human  beings  and  in  many  of  which,  therefore,  systemic  dis- 
tribution of  the  virus  (a  generalization  apparently  necessary  for  the 
development  of  resistance)  may  not  have  taken  place.  Neisser's 
conclusions,  therefore,  that  monkeys,  like  human  beings,  are  not 
reinoculable  while  suffering  from  systemic  syphilis,  seem  entirely 
justified. 

His  work,  as  well  as  that  of  Finger  and  Landsteiner,  and  of 
Kraus  and  Volk143  on  lower  monkeys,  has  shown  that  resistance 
does  not  develop  until  the  twelfth  to  the  twentieth  or  twenty-first  day 
after  the  first  inoculation;  that  is,  again  as  in  man,  when  the  virus 
has  become  generally  distributed.  Finger  and  Landsteiner,  further- 
more, noticed  that  reinoculation-products,  obtained  by  reinfection 
during  the  first  incubation  period,  that  is,  before  the  development  of 
the  primary  lesion,  were  less  severe  and  developed  in  a  shorter  time 
than  did  the  first  lesion.  This  phenomenon  which  would  tend  to 
mark  another  analogy  to  the  conditions  prevailing  in  human  beings, 
was  not  observed  in  the  experiments  of  Neisser  and  of  Kraus  and 
Yolk.  However,  like  the  similar  observation  in  human  beings,  it 
seems  to  indicate  the  gradual  acquisition  of  resistance  as  the  virus 
begins  to  exert  its  influence  upon  the  tissues. 

Again,  with  monkeys  as  in  man,  the  question  arises  whether  the 
resistance  so  unquestionably  proven  is  a  condition  merely  coexistent 
with  active  disease,  or  whether  it  may  be  interpreted  as  a  true  im- 
munity which  persists  after  the  microorganisms  have  been  com- 
pletely removed.  The  most  directly  pertinent  experiments  are  those 
of  Neisser.  Neisser  reinoculated  monkeys  at  periods  ranging  from 
27  to  645  days  after  the  first  infection.  After  waiting  a  time  suffi- 
ciently long  to  insure  the  negative  result  of  the  reinoculation,  he  used 
organ-substance  from  these  animals  to  inoculate  other  monkeys.  In 
22  experiments  of  this  kind  he  obtained  positive  results — showing 
that  the  organs  of  the  apparently  immune  animals  still  harbored 

142  Neisser.    Beitr.  z.  Pathol.  u.  Ther.  d.  Syphilis,  Springer,  Berlin,  1911. 

143  Kraus  and  Volk.     Wien.  klin.  Wchnschr.,  1906,  No.  21,  Ref.  IX.  Kon- 
gress  d.  Dermatol.  Gsell.  in  Bern. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  507 

virulent  treponemata.  In  seven  animals  only  did  the  inoculations 
with  organ-substance  fail  to  produce  lesions,  but  of  these  all  but  one 
died  before  the  30th  day  after  inoculation. 

In  contrast  to  these  results  Neisser  found  that  animals  which  had 
been  "cured"  by  various  treponemacidal  agents  such  as  atoxyl, 
arsacetin,  etc.,  were  almost  regularly  reinoculable.  In  fact,  these 
experiments  were  so  uniform  that  Neisser  later  utilized  the  reinocu- 
lation  method  as  an  index  of  cure  or  persistence  of  the  disease. 

The  results  obtained  in  monkeys,  therefore,  are  very  similar  to 
those  determined  by  clinical  observations  in  man,  and  the  following 
statements  may  be  taken  as  summarizing  the  conditions  revealed  by 
monkey  experiment. 

1.  That  the  body  develops  resistance,  progressively  increasing 
as  the  virus  becomes  generally  distributed. 

2.  That  the  resistance  probably  reaches  its  highest  development 
during  the  early  tertiary  period. 

3.  That  complete  cure  is  probably  synonymous  with  gradual 
return  of  susceptibility. 

The  only  other  animals  on  which  systematic  experimentation  has 
been  possible  up  to  the  present  time  have  been  rabbits.  Since 
Bertarelli's  successful  production  of  keratitis  and  Parodi's  inocula- 
tion of  the  testes  in  these  animals,  they  have  been  studied  carefully 
by  a  number  of  workers,  chiefly  Uhlenhuth  and  Mulzer,144  and 
Noguchi 145 ;  and  in  our  own  laboratory,  with  Hopkins  and  Mc- 
Burney,  the  writer  has  observed  rabbit  syphilis  for  a  number  of 
years.  From  a  large  mass  of  observations  it  appears  that  the  condi- 
tions in  rabbits  are  not  identical  with  those  observed  in  man  and 
monkeys.  So  far,  the  testes  and  the  eyes  are  the  only  organs  in, 
these  animals  in  which  syphilitic  lesions  can  regularly  be  produced, 
and  although  treponema  pallidum  may  apparently  be  distributed 
generally  to  the  organs  after  inoculation,  it  does  not  easily  arouse 
pathological  reactions  except  in  the  organs  named,  and  the  lesions 
produced  are  not  accompanied  by  a  generalized  resistance  compar- 
able to  that  discussed  in  connection  with  the  higher  animals  in  pre- 
ceding sections  of  this  paper. 

Bertarelli  found  that  he  could  reinoculate  the  cornea  in  a  rabbit 
that  had  previously  been  inoculated  with  syphilis.  Uhlenhuth  and 
Mulzer,  and  Neisser  and  Piirckhauer,  showed  that  infections  of  the 
eye  could  be  produced  while  the  opposite  eye  was  still  syphilitic. 
Tomazewski  found  that  scrotal  infection  did  not  protect  against  in- 
fection of  the  cornea,  and  vice  versa.  Furthermore,  Uhlenhuth  and 
Mulzer,  on  the  basis  of  a  very  thorough  study,  believe  that  such  rein- 
fections are  neither  less  extensive  nor  more  rapid  in  healing  than 

144  Uhlenhuth  and  Mulzer.     Arb.  a.  d.  k.  Gsndhtsamte.,  1913,  xliv. 

145  Noguchi,  H.     Jour.  Am.  Med.  Assn.,  1912,  Iviii,  1163. 


508  INFECTION  AND  RESISTANCE 

was  the  first  lesion.  The  same  authors  noticed  resistance  to  reinocu- 
lation  in  two  animals  only,  and  these  were  young  rabbits  in  the  first 
weeks  of  life,  which,  they  believed,  had  been  generally  syphilized  by 
intracardial  injections.  Interesting  in  this  connection  are  claims  by 
Ossola  and  Truffi  who  believe  that  successful  skin  inoculation  in 
rabbits  confers  a  certain  amount  of  skin  resistance  and  this  is  in 
harmony  with  the  belief  of  Kraus  and  Yolk,  that  a  specific  skin 
immunity  in  syphilis  is  possible.  Of  great  interest  to  us  and  of  a 
possible  theoretical  bearing  which  we  will  discuss  below,  are  observa- 
tions made  by  the  writer  with  Hopkins  and  McBurney  146  on  20 
rabbits  which  were  reinoculated  into  the  testes  after  apparent  healing 
of  lesions  in  these  organs.  It  appeared  that  in  rabbits  the  opposite 
testis  can  be  successfully  inoculated  before,  during,  or  after,  the 
existence  of  a  testicular  lesion  on  one  side,  but  that  reinoculation  of 
the  same  testis  which  had  apparently  returned  to  normal,  at  periods 
ranging  from  6  weeks  to  one  year,  was  not  often  successful.  There 
is  a  certain  amount  of  evidence  here  of  a  purely  local  immunity,  a 
matter  which  we  will  discuss  at  greater  length  presently.  Perhaps 
the  difference  between  rabbits  and  the  higher  animals  lies  chiefly  in 
the  fact  that  syphilis  is  not  generalized  in  the  same  sense  that  it  is 
in  man  and  monkeys,  and  even  though  inoculations  from  the  organs 
of  syphilitic  rabbits  may  often  result  positively,  this  might  signify 
only  that  the  treponemata  have  been  generally  distributed  by  the 
blood  stream  and  latently  lodged  in  the  organs,  without,  however, 
arousing  in  the  tissues  any  sort  of  pathological  response.  This  idea 
would  seem  to  be  borne  out  by  the  two  experiments  of  Uhlenhuth 
and  Mulzer  cited  above  in  very  young  rabbits  since  in  such  animals 
true  generalization  seems  to  be  more  common.  The  study  of  very 
young  rabbits  will  be  continued  with  this  point  in  view. 

It  is  apparent  from  the  preceding  considerations  that  resistance 
to  syphilis  differs  from  that  acquired  in  many  bacterial  infections 
chiefly  in  the  fact  that  it  does  not  persist  after  the  disease  is  over, 
but  probably  coexists  only  with  the  presence  of  the  living  incitants 
in  the  body.  In  this  respect  it  seems  to  be  similar  to  the  conditions 
prevailing  in  many  protozoan  diseases.  Thus  Schilling147  cites 
a  case  of  Trypanosoma  Brucei  in  a  steer,  experimentally  infected 
by  Koch,  in  which  reinoculation  was  repeatedly  negative,  this  result 
being  at  first  falsely  interpreted  as  immunity;  but  six  years  later 
Kleine  found  that  the  same  steer  still  harbored  the  trypanosomes 
in  his  blood — showing  that  the  resistance  to  reinoculation  was  not, 
in  this  case,  an  evidence  of  true  immunity,  but  rather  represented  a 
condition  of  insusceptibility  to  "superinfection"  analogous  in  every 

146  Zinsser,  H.,  Hopkins,  J.  Gv  and  McBurney,  M.     Jour.  Exper.  Med., 
1916,  xxiii,  329,  341. 

147  Schilling.     In  Kolle  u.  Wassermann  Handb.  d.  Path.  Mikroorg.,  xvii, 
600. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  509 

respect  to  that  existing  in  syphilis.  Similar  conditions  have  been 
shown  to  prevail  in  Texas  fever  and  Schilling  believes  that  they  may 
be  regarded  as  also  existing  to  a  certain  extent  in  Malaria  and  Sleep- 
ing Sickness.  In  both  of  these  conditions  complete  spontaneous 
sterilization  of  the  body  (without  medicinal  aid)  is  probably  rare, 
possibly  does  not  occur  at  all,  and  the  apparent  immunity  to  rein- 
fection is,  as  in  syphilis,  an  evidence  of  persistence  of  the  disease  in  a 
latent  form. 

In  weighing  the  analogy  of  syphilis  to  such  protozoan  diseases, 
one  is  inclined  to  wonder  whether  syphilis  in  man  may  be  regarded 
as  at  all  spontaneously  curable.  So  few  are  the  cases  left  untreated 
and  so  rarely  does  the  reinfection  occur  even  in  the  face  of  specific 
remedies,  that  it  seems  to  us  more  than  likely  that  in  syphilis,  as  in 
Sleeping  Sickness,  a  spontaneous  "sterilizing"  immunity  does  not 
occur.  This  is  a  point,  however,  regarding  which  it  is  impossible 
to  gather  data. 

In  order  to  distinguish  the  conditions  outlined  above  tersely  from 
the  ordinary  conception  of  immunity,  Neisser  148  speaks  of  the  altera- 
tions which  govern  the  reactions  of  the  luetic  body  to  freshly  intro- 
duced virus  as  "Anergie;"  and  "Umstimmung"  or  "Allergie."  By 
"Anergie"  (a  term  first  used  by  v.  Pirquet  and  subsequently  intro- 
duced by  Siebert  in  working  out  the  analogy  between  pigeon-epitheli- 
oma  and  syphilis),  Neisser  designates  a  condition,  of  inability  to 
react  by  cellular  change  to  contact  with  the  virus.  As  he  uses  it,  it 
implies  a  passivity  on  the  part  of  the  invaded  tissues  (in  which  "die 
Zellen  auf  die  Spirochaeten  schwer  oder  garnicht  reagiren),  by 
which  there  is  not  necessarily  a  destruction  of  the  invading  trepone- 
mata,  and  which,  therefore,  cannot  in  any  sense  be  interpreted  as  a 
"Schutz  wirkung."  By  the  "Umstimmung"  of  Neisser,  the  "change- 
ment  dans  la  mode  de  reaction"  of  Levaditi,  is  meant  the  changed 
reaction  capacity  of  the  syphilized  tissues  which  determines  the  char- 
acters of  the  lesions  at  various  stages  of  the  disease.  Thus  it  is 
obvious  that  cellular  reactions  which  result  in  the  primary  induration 
are  quite  different  from  those  which  produce  the  tertiary  gumma, 
and  that  the  histological  changes  of  the  roseola  are  distinct  from 
those  of  the  serpiginous  syphilide  of  the  late  stages.  And  since,  as 
we  shall  see,  there  is  no  valid  reason  to  assume  that  the  incitant  has 
been  modified  in  virulence  or  vitality,  we  are  forced  to  believe  that 
the  reaction  capacity  of  the  body  cells  has  been  altered. 

Since  it  is  a  fact,  then,  that  syphilitic  infection  so  changes  the 
body  tissues  of  man  and  monkeys  that,  during  its  course,  resistance 
to  reinfection  is  produced,  it  should  be  possible  to  analyze  this  re- 
sistance into  its  responsible  factors  and  perhaps  utilize  the  knowledge 
so  gained  for  practical  therapeutic  purposes.  Before  we  proceed 

148  Neisser.  Baermann  u.  Hall) erst aedfer  d.  med.  Wchnschr.,  1906,  Nos. 
1  and  3. 


510  INFECTION  AND  RESISTANCE 

to  do  this,  it  will  be  of  advantage  to  review  briefly  the  attempts  at 
active  and  passive  immunization  which  have  been  made  in  animals 
and  man. 

Metchnikoff  and  Roux  149  followed  their  first  animal  inoculation 
studies  by  extensive  vaccination  experiments.  Some  of  their  first 
work  along  these  lines  is  perhaps  marred  to  some  extent  by  an  in- 
sufficient recognition  of  the  resistance  which  depends  upon  per- 
sistence of  the  disease  rather  than  upon  true  immunity,  but  a  good 
many  of  their  observations  are  of  fundamental  importance.  It  will 
be  convenient  to  classify  their  work  and  that  of  others  into  experi- 
ments dealing  with  "active"  and  those  dealing  with  "passive"  im- 
munization. 

Metchnikoff  and  Roux  first  worked  with  filtered  virus  and  virus 
killed  at  51°  C.  All  their  attempts  with  such  material  were  negative 
in  that  the  monkeys  treated  with  it  could  not  be  regarded  in  any 
sense  as  immunized.  In  their  reports  of  these  experiments  they 
remark  that  they  believed  this  to  be  due  to  an  absolute  loss  of  power 
to  incite  reaction  on  the  part  of  the  vaccine-material.  We  emphasize 
this  point  here  because  our  own  subsequent  work  inclines  us  to  be- 
lieve, with  them,  that  the  production  of  a  reaction  is  necessary  for 
the  development  of  any  considerable  degree  of  resistance. 

Neisser  carried  out  a  large  number  of  attempts  at  vaccination  in 
which  he  used  extracts  of  syphilitic  primary  lesions  and  of  the  organs 
of  congenitally  syphilitic  children,  killed  by  the  addition  of  carbolic 
acid.  Unfortunately  he  assumed  that  his  extract  contained  syphilitic 
antigen  because  it  gave  positive  reactions  by  the  complement  fixation 
technique  of  Wassermann,  an  assumption  which  we  of  course  know 
now  to  be  unfounded  as  far  as  any  relation  to  the  body  substance  of 
the  treponemata  is  concerned.  This,  to  our  mind,  deprives  these 
particular  experiments  entirely  of  their  negative  importance. 

In  rabbits  a  large  amount  of  work  has  been  done  by  Uhlenhuth 
and  Mulzer.  They  injected  living  material  from  rabbit  lesions  in- 
travenously and  subcutaneously,  without  ever  observing  any  evidence 
of  protection  against  subsequent  inoculations. 

Of  perhaps  the  greatest  importance  in  connection  with  active  im- 
munization are  the  attempts  made  upon  human  beings  by  different 
investigators. 

Casagrandi  and  de  Luca  15°  tried  prophylactic  immunization  on 
six  human  beings  by  injection  of  filtrates  obtained  from  primary 
lesions.  Two  of  these  people  later  contracted  syphilis  in  the  ordinary 
way. 

Possibly  the  most  hopeful  results  are  those  obtained  by  Spitzer  ir>1 

149  Metchnikoff  and  Roux.    Ann.  de.  I'Inst.  Pasteur,  1903,  xvii,  809;  1904, 
xviii,  1  and  657;  1905,  xix,  673;  1906,  xx,  785. 

150  Casagrandi  and  De  Luca.     Gior.  ital.  de  mal.  ven.,  1905. 

151  Spitzer.     Wien.  klin.  Wchnschr.,  1905,  45,  and  1906,  38. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  511 

by  a  method  suggested  by  Kraus.  Kraus,152  reasoning  from  the  fact 
that  syphilis  like  hydrophobia  was  a  disease  with  Ipng  incubation 
time,  expressed  the  hope  that  perhaps  the  method  of  Pasteur  in 
hydrophobia,  that  is,  active  immunization  during  the  period  of  in- 
cubation might,  in  syphilis  also,  tend  to  abort  the  disease.  Accord- 
ingly, Spitzer  treated  15  cases  of  early  syphilis  immediately  after 
the  appearance  of  the  chancre  by  subcutaneous  injections  of  emul- 
sions made  from  human  chancre  material  in  dilutions  of  1 :200  to 
1:20.  The  cases  received  from  11  to  20  injections  and  in  seven  of 
them  the  disease  was  uninfluenced.  In  the  others,  however,  subse- 
quent symptoms  were  delayed,  and  in  four,  no  generalized  symptoms 
occurred.  In  a  later  communication,  Spitzer  reported  23  further 
cases  similarly  treated,  10  of  which  failed  to  develop  generalized 
symptoms  and  in  9  of  these  the  Wassermann  test  remained  negative. 
One  of  them,  a  fact  which  is  of  great  importance,  was  spontaneously 
reinfected  with  syphilis  two  and  one-half  years  later.  These  results, 
if  accurate  in  every  way,  are  of  the  greatest  importance,  but  are 
diametrically  opposed  to  the  experience  of  all  other  investigators. 
Monkey  experiments  along  the  same  lines  by  Neisser  gave  entirely 
negative  results  and  Brandweiner 153  as  well  as  Kreibich  154  were 
unable  to  confirm  Spitzer's  results  in  man.  Further  comment  on 
the  Kraus  and  Spitzer  method  is  valueless  without  more  experi- 
mental data.  It  is  one  of  the  few  rays  of  hope,  but  so  isolated  that 
one  is  forced  to  skepticism.  Metchnikoff  and  Roux  did  almost  the 
identical  thing  in  an  orang-outang  and  obtained  lesions  both  at  the 
point  of  the  original  inoculation  as  well  as  that  at  which  the  subse- 
quent "protective"  injection  was  made. 

The  only  experiments  in  which  an  attempt  at  vaccination  with 
attenuated  virus  was  made  with  some  indication  of  efficacy,  is  one  of 
Metchnikoff  and  Roux,  the  outcome  of  an  accidental  laboratory  in- 
fection. It  appears  that  a  laboratory  assistant  who  had  been  attend- 
ing to  the  animals,  noticed  a  small  lesion  on  his  lip  which  did  not 
look  like  a  typical  syphilitic  chancre.  In  order  to  allay  the  patient's 
fears,  however,  Metchnikoff  and  Roux  did  inoculations  from  the 
patient  to  monkeys  and  these  were  positive.  Nevertheless,  Fournier 
after  examining  the  original  lesion  declared  it  so  unlike  the  ordinary 
primary  sore  that  he  did  not  advise  treatment.  No  secondaries  de- 
veloped in  the  patient  nor  in  the  three  chimpanzees  inoculated  with 
the  material.  From  this  occurrence  Metchnikoff  and  Roux  con- 
cluded that  the  patient  had  probably  been  infected  in  handling  the 
monkeys,  and  that  the  virus  had  become  attenuated  by  passage 
through  these  animals.  On  the  basis  of  this  observation  they  later 
inoculated  a  willing  subject  79  years  old  with  virus  carried  for  five 

152  Kraus.     Wien.  klin.  Wchnschr.,  1905,  No.  41,  and  1906,  No.  21. 

153  Brandweiner.     Wien.  med.  Wchnschr.,  1905,  No.  45. 

154  Kreibich.     Wien.  klin.  Wchnschr.,  1906,  No.  8. 


512  INFECTION  AND  RESISTANCE 

generations  in  lower  monkeys.  The  lesion  which  developed  was  very 
slight,  consisting  only  of  a  local  induration,  and  no  generalized  symp- 
toms developed.  A  previous  attack  of  syphilis  Metchnikoff  believes 
could  be  reliably  excluded  in  the  subject.  The  experimenters  suggest 
that  the  passage  through  lower  monkeys  may  attenuate  the  virus  for 
man,  this  leading  to  relative  immunity  to  subsequent  inoculations, 
and  furnish  a  possible  means  of  protection.  Their  experiments  are 
too  few  to  permit  conclusions  as  yet,  but  even  should  they  hold  good, 
the  method  would  seem  to  imply  a  considerable  degree  of  risk  and 
our  experience  with  superinfections  and  reinfections  in  syphilis  does 
not  encourage  the  hope  that  a  method  so  drastic  would  be  justified 
when  the  benefit  to  the  patient  is  apt  to  be  of  such  duration  only. 

The  history  of  passive  immunization  is  an  extensive  one  and 
hardly  worth  going  into  with  any  degree  of  detail  since  many  times 
extravagant  claims  have  been  made  only  to  be  refuted  by  accurate 
study.  There  is  a  great  similarity  in  respect  to  this  between  syphilis 
studies  and  those  in  tuberculosis  and  cancer.  A  brief  examination  of 
the  bibliography  in  Neisser's  book  is  sufficient  to  convince  one  of  the 
many  attempts  that  have  been  made  in  this  direction  and  often  by 
methods  as  ludicrous  as  the  claims  of  success  for  which  they  formed 
the  basis.  The  most  careful  and  skillful  workers  have  uniformly 
reported  failure.  Metchnikoff  and  Roux  treated  various  monkeys 
with  blood  from  syphilitic  patients  and  used  the  serum  of  these  ani- 
mals for  protective  experiments.  There  are  a  few  instances  in  which 
mixtures  of  such  serum  with  syphilitic  virus  rendered  this  inactive 
on  inoculation.  A  powder  made  of  this  serum  was  supposed  to  have 
some  protective  effect  when  applied  to  fresh  inoculation  spots  within 
the  first  hour  after  inoculation.  However,  injection  of  the  serum 
had  no  effect  whatever.  Casagrandi  and  de  Luca  using  serum  of  a 
dog  treated  with  syphilitic  virus,  obtained  entirely  negative  results, 
and  Finger  and  Landsteiner  155  report  negative  results  with  monkey 
blood  in  man.  The  most  extensive  experiments  were  again  those 
carried  out  by  Neisser  and  his  associates.  They  were  done  by  the 
treatment  of  animals  with  dead  and  living  syphilis  virus,  organ 
extracts,  with  the  blood  of  syphilitic  man  and  monkey,  and  horses, 
sheep,  and  monkeys  were  used  for  the  production  of  "immune" 
serum.  In  no  case  was  there  the  slightest  protective  effect  on  the 
part  of  the  serum  either  in  vitro  or  in  vivo,  and  the  results  of  the 
experiment  were  unqualifiedly  negative. 

When  the  method  of  complement  fixation  was  successfully  ap- 
plied to  the  diagnosis  of  syphilis  first  by  Detre,  and  then  by  Wasser- 
mann,  Neisser  and  Bruck,156  it  was  generally  assumed  that  this 
reaction  incidentally  demonstrated  that  specific  antibodies  were 

155  Finger  and  Landsteiner.     Centralbl.  /.  Bakteriol.,  Ref.  xxxviii,  1906. 

156  Wassermann,  A.,  Neisser,  A.,  and  Bruck,  C.    Deutsch.  meet.  Wchnschr., 
1906,  xxxii,  755. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  513 

formed  in  syphilis.  As  matters  have  developed,  however,  this  point 
of  view  can  no  longer  be  maintained.  It  was  soon  discovered  that 
the  antigen  used  for  these  reactions  by  Wassermann  and  his  associates 
derived  its  "fixing"  constituent  not  from  the  body  substances  of  the 
treponemata  contained  in  the  syphilitic  organs,  but  from  certain 
tissue  extractives  chiefly  of  lipoidal  nature  which  could  be  obtained 
readily  from  normal  as  well  as  syphilitic  tissues.  Although  Bruck  157 
and  others  who  have  occupied  themselves  with  the  theoretical  basis 
of  the  Wassermann  reaction,  still  maintain  that  a  specific  antibody 
may  be  incidentally  involved,  they  admit  that  this  is  the  less  im- 
portant factor  in  the  reaction  which  depends  chiefly  upon  the  ex- 
istence of  lipotropic  substances  which  appear  in  the  course  of  the 
disease  as  metabolic  products  either  of  the  body  or  possibly  of  the 
treponemata.  This  is  a  subject  which  we  will  deal  with  in  extenso 
in  another  paper.  It  is  sufficient  for  our  present  purposes  to  point 
out  that,  although  to  a  slight  extent  specific  antibodies  may  play  a 
part  in  the  Wassermann  reaction,  this  is  certainly  not  the  chief 
element  or  even  a  very  important  factor  involved.  The  truth  of  this 
conception  has  been  further  confirmed  by  the  work  of  Noguchi,158 
of  Craig  and  Nichols,159  of  Kolmer,160  and  ourselves,  in  which  it 
has  been  found  that  antigens  made  with  pure  cultures  of  treponemata 
produced  a  complement  fixation  with  syphilitic  sera  to  a  very  limited 
extent  only,  and  in  our  work  in  which  we  have  been  able  to  duplicate 
the  Wassermann  reaction  in  a  large  series  of  antigens  made  from  the 
treponema  cultures,  we  have  found  that  similar  results  could  be 
obtained  with  cultures  of  colon  and  typhoid  bacilli  identically  pre- 
pared, the  fixing  power  to  a  large  degree  depending  upon  the  lipoidal 
constituents  of  the  bacteria.  Whatever  the  ultimate  explanation  of 
the  Wassermann  reaction  may  turn  out  to  be  (and  this  is  a  subject 
which  it  would  not  be  profitable  to  discuss  at  length  at  this  place),  it 
cannot  be  maintained  that  as  it  exists  at  present,  it  can  be  interpreted 
as  demonstrating  the  existence  of  true  circulating  antibodies,  analo- 
gous to  those  found  in  bacterial  diseases  in  the  blood  of  syphilitic 
patients. 

Early  in  the  history  of  such  investigations,  Fornet,161  with  his 
collaborators  Schereschewsky,  Eisenzimmer,  and  Rosenfeld,162  found 
that  when  the  sera  of  syphilitics  were  mixed  with  clear  extracts  of 
syphilitic  livers  similar  to  those  used  in  the  Wassermann  reactions, 
precipitates  formed  which  were  not  seen  in  similar  experiments  done 

157  Bruck.  Imm.  bei  Sypn.  in  Kolle  u.  Wassermann  Handb.  d.  Pathog. 
Mikroorg.,  vii,  1045  et  seq. 

158Noguchi,  H.     Jour.  Exper.  Med.,  1911,  xvi,  99. 

159  Craig,  C.  F.,  and  Nichols,  H.  J.    Jour.  Exper.  Med.,  1912,  xvi,  336. 

160  Kolmer,  J.  A.    Jour.  Exper.  Med.,  1913,  xviii,  18. 

lei  Fornet.    XIV  Internal.  Kongress  f.  Hyg.  und  Derm.,  Sept.  1,  1907. 
162  Fornet,  Schereschewsky,  Eisenzimmer  and  Rosenfeld.     Deutsch.  med. 
Wcbnschr.,  1907,  No.  41. 


514  INFECTION  AND  RESISTANCE 

with  normal  sera.  This  Fornet163  in  a  number  of  communications 
interpreted  as  showing  the  formation  of  precipitins  in  the  course  of 
syphilis.  At  almost  the  same  time,  L.  Michaelis  164  made  similar 
observations  giving  them  the  same  interpretations.  The  experiments 
of  Fornet  have  not  found  universal  confirmation,  but  even  if  such 
precipitin  reactions  can  be  occasionally  observed,  they  do  not  indicate 
true  precipitins  for  the  same  reasons  that  the  Wassermann  reaction 
does  not  demonstrate  true  complement  fixation  of  antibodies.  The 
antigens  did  not  represent  strong  treponema  antigens  and  Jacobs- 
thai  165  and  others  have  shown  that  with  the  ordinary  Wassermann 
antigens  syphilitic  sera  can  produce  precipitations  visible  under  the 
dark  field,  the  basis  of  the  complement  fixation  being,  therefore,  one 
of  probable  colloidal  precipitation.  The  formation  of  true  precipi- 
tins, therefore,  has  not  been  shown  for  syphilis. 

Investigations  of  the  effect  produced  upon  virulent  treponemata 
by  the  sera  of  syphilitic  individuals  have  likewise  been  unsatisfac- 
tory. Hoffmann  and  Prowazek  166  reported  in  1906  that  the  serum 
of  syphilitics  in  the  later  stages  produced  immobilization  of  the 
virulent  treponema  pallidum,  an  observation  which  was  confirmed 
by  Zabolotny.167  Landsteiner  and  Mucha  168  were  not  able  to  observe 
such  immobilization,  nor  did  they  see  any  evidence  of  agglutination 
in  such  experiments.  In  a  limited  observation  of  our  own,  we  have 
also  failed  to  see  any  regular  or  distinct  influences  of  this  kind.  It  is 
not  impossible  that  a  slight  difference  may  exist  in  this  respect 
between  syphilitic  and  normal  sera,  but  even  if  it  occurs,  the  action 
is  feeble,  irregular,  and  entirely  insufficient  to  be  interpreted  as 
having  much  practical  importance  in  the  acquisition  of  syphilis 
immunity. 

Treponemacidal  experiments  have  been  done  by  some  workers 
also  with  negative  results.  As  we  have  stated  in  another  part  of 
this  paper,  the  attempt  to  treat  syphilitic  animals  and  man  with  the 
serum  of  syphilitics  has  led  to  no  reliable  results,  and  no  evidence 
whatever  that  can  be  accepted  has  been  adduced  to  show  that  virulent 
treponemata  may  be  killed  by  active  luetic  serum. 

As  far  as  any  opsonic  action  is  concerned,  Levaditi 169  in  histo- 

163  Fornet,    Schereschewsky.     Munchen   med.    Wchnschr.,  1907,   No.   30 ; 
Berl  klin.  Wchnschr.,  1908,  85. 

164  Michaelis.    Berl.  klin.  Wchnschr.,  1907,  No.  46. 

165  Jacobsthal.     Munchen.  med.  Wchnschr.,  1910,  Ivii,  214. 

166  Hoffmann,  E.,  and  Prowazek,  S.     Centralbl  f.  Bakteriol,  I.  0.,  1906, 
xli,  741  and  817.     (Balanitis  and  montu  spirochaetae.) 

167  Zabolotny,    D.,    and    Maslakowitz.      Centralbl    f.    Bakteriol,    I.    O., 
xliv,  532. 

168  Landsteiner  and  Mucha.     Wien  klin.  Wchnschr.,  1906,  xix,  1349;  Cen- 
tralbl f.  Bakteriol,  1907,  I.  R.,  xxxix,  540. 

169  Levaditi.     Compt.  rend.  Soc.  de  biol,  1906,  Ix,  134;  Ann.  de  I'Inst. 
Pasteur,  1906,  xx,  41. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  515 

logical  studies  in  congenitally  syphilitic  children  has  seen  phagocy- 
tosis or  at  least  intracellular  localization  of  treponemata  in  the  alveo- 
lar cells  of  the  lung  and  in  the  parenchyma  cells  of  the  liver  and 
kidneys.  For  reasons  not  entirely  clear  to  us,  he  interprets  the 
former  as  true  phagocytosis  and  the  latter  as  a  penetration  of  the 
t,reponemata  into  the  liver  cells  to  the  detriment  of  the  latter.  We 
ourselves  have  occasionally  seen  phagocytosis  in  sections  of  syphilitic 
rabbit  testes,  but  in  all  cases  the  process  was  not  a  very  active  one 
and  not  much  can  be  said  about  its  importance  at  present.  As  a 
matter  of  fact,  Hopkins  and  the  writer,170  in  studying  the  mechanism 
of  the  natural  resistance  of  mice  against  syphilis,  injected  virulent 
organisms  into  the  peritoneal  cavities  and  observed  the  treponemata 
in  peritoneal  puncture  fluid,  alive,  actively  motile,  and  unphagocyted, 
though  surrounded  by  masses  of  leucocytes,  as  long  as  three  days 
after  their  injection.  It  seemed  almost  as  though  the  natural  im- 
munity of  such  animals  might  be  similar  to  the  "atreptic"  cancer 
resistance  spoken  of  by  Ehrlich.  The  treponemata  did  not  multiply 
in  the  mice  but  though,  naturally,  diminishing  in  number,  were  ap- 
parently neither  killed  nor  even  inhibited  in  motility  by  the  perito- 
neal exudates  and,  for  several  days,  swam  in  and  out  among  the 
accumulating  leucocytes,  often  adhering  to  them  peripherally  but 
not  taken  up  by  them.  Lack  of  entirely  satisfactory  methods  of 
staining  cells  containing  treponemata  make  it  difficult  to  speak 
with  certainty  of  the  actual  occurrences.  But  we  gained  the  distinct 
impression  that  the  treponemata  were  not  actively  injured  or  de- 
stroyed until  they  had  spontaneously  died  out  owing  to  lack  of 
suitable  environment,  i.  e.,  nutrition. 

In  the  case  of  natural  immunity  at  any  rate,  we  do  not  think  that 
phagocytosis  by  the  mobile  leucocytes  plays  a  primarily  important 
role.  However,  these  experiments  will  need  further  elaboration. 

The  search  for  antibodies  gained  new  vigor  when  the  efforts  of 
Schereschewsky,171  Miihlens,172  Hoffmann  173  and  especially  Nogu- 
chi,  had  resulted  in  successful  cultivation  of  treponemata  from  syphil- 
itic lesions  in  man  and  animals.  It  was  hoped  that  with  the  causative 
organisms  isolated,  immunization  and  a  clear  understanding  of  the 
antibodies  in  syphilis  might  yield  practical  results.  Kolmer  174  ob- 
served that  cultivated  treponemata  were  agglutinated  in  the  sera  of 
rabbits  treated  with  culture  material,  and  such  agglutinins  were  pro- 

170  Zinsser,  H.,  and  Hopkins,  J.  G.     Jour.  Am.  Med.  Assn.,  1914,  Ixii, 
1802 ;  Jour.  Exper.  Med.,  1915,  xxi,  576. 

171  Schereschewsky,  J.     Deutsch.  med.  Wchnschr.,  1911,  xxxvii,  Nos.  20 
and  39. 

172  Miihlens.     Treponema  Pallidum  in  v.  Prowazek  Handbvch  der  Paihog. 
Protozoen,  i,  Barth,  Leipzig,  1912 ;  Klin.  Jahrh.,  1910,  xxiii,  339. 

173  Hoffmann.     Berl.  klin.  Wchnschr.,  1905,  No.  46. 

174  Kolmer,  J.  A.,  Williams,  W.  W.,  and  Laughbaugh,  E.  E.    Jour.  Med. 
Research,  1913,  xxviii,  345. 


516  INFECTION  AND  RESISTANCE 

duced  in  high  potency  by  the  writer  with  Hopkins.  Subsequently 
in  our  own  laboratory  with  Hopkins  and  McBurney,  extensive  ex- 
perimentation on  the  production  of  antibodies  with  culture  pallida 
was  carried  out. 

We  were  able  to  show  that  not  only  were  agglutinins  formed  by 
the  immunization  of  rabbits  with  such  cultures,  but  also  that  tre- 
ponemacidal  antibodies  which  were  analogous  to  the  ordinary  bac- 
tericidal substances  in  such  sera  were  present.  Also,  it  was  shown 
by  cross  agglutination  and  absorption  experiments,  that  treponemata 
cultivated  from  various  sources  were  related  in  group  reactions. 

Indeed,  experiments  done  with  cultivated  treponemata  (much 
facilitated  by  the  discovery  of  a  simple  method  of  obtaining  mass 
cultures  of  old  strains)  seemed  at  first  very  encouraging  in  that 
animals  immunized  with  the  cultures  responded  by  powerful  anti- 
body formation.  It  was  a  perfectly  justified  hope,  therefore,  that 
antibodies  produced  with  these  "attenuated"  or  rather  "avirulent" 
strains  might  have  some  action  on  the  virulent  treponemata  in  luetic 
lesions.  However,  subsequent  work  in  this  direction  disappointed 
such  expectations.  We  may  briefly  review  this  work  as  follows : 

The  serum  of  rabbits  immunized  with  "culture"  pallida  although 
potent  against  "culture"  pallida,  had  no  effect  either  in  agglutinating 
the  virulent  organism  from  rabbit  lesions,  nor  did  it  exert  any  pro- 
tective influence  when  the  virulent  organisms  were  subjected  to  its 
action  before  injection.  4 

Conversely,  the  serum  of  syphilitic  rabbits  showed  but  a  very 
slightly  increased  agglutinating  power  for  the  "culture"  pallida. 
This  increase  of  potency  in  a  few  experiments  was  definite  but  very 
slight,  a  few  of  the  syphilitic  animals  agglutinating  as  highly  as  1 :25 
and  1 :50,  whereas  most  of  the  normal  rabbits  agglutinated  in  1 :10 
and  some  of  them  1 :25. 

Although  Kissmeyer  has  recently  reported  that  diagnostic  use 
might  be  made  of  the  fact  that  sera  of  syphilitic  individuals  ag- 
glutinated the  "culture"  pallida,  we  have  tried  this  with  a  consider- 
able number  of  cases  and  found  that  although  the  sera  of  tertiary 
syphilitics  will  sometimes  agglutinate  a  little  more  highly  than  will 
the  sera  of  normal  individuals,  yet  many  patients  suffering  from 
non-syphilitic  diseases  agglutinated  as  highly  and  almost  as  regularly 
.as  did  the  syphilitics. 

We  have  come  to  the  conclusion,  therefore,  as  far  as  our  work 
lias  gone,  that  in  the  syphilitic  human  being  there  is  as  little  ag- 
glutinin  formation  against  the  "culture"  treponema  as  there  seems 
to  be  against  the  virulent  organisms.  If  the  slightly  greater  ag- 
glutinating power  found  in  some  of  the  tertiary  syphilitics  can  be 
considered  at  all,  the  reaction  is  so  feeble  that  it  is  negligible  from 
the  points  of  view  either  of  diagnostic  value  or  protective  importance. 

Furthermore,  vaccination  either  intravenously  or  locally  into  the 


THERAPEUTIC  IMMUNIZATION  IN  MAN  517 

testis  with  cultures  has  thus  far  failed  to  protect  rabbits  against 
subsequent  inoculation  with  virulent  material,  and  passive  immuniza- 
tion with  sera  produced  with  "culture"  pallida  has  been  without 
effect. 

From  all  this  it  appears  that  the  "culture"  treponema  has.im- 
munologically  no  relation  to  the  virulent  organism.  It  has  lost  its 
virulence  completely,  as  six  and  more  successive  inoculations  into 
rabbit  testes  have  sufficiently  demonstrated  to  us. 

The  sera  produced  by  many  injections  of  dead  and  living  culture 
organisms  have  no  effect  whatever  on  the  virulent  organisms  in  vitro, 
and  vaccination  with  it  does  not  protect  against  subsequent  infection. 

The  luetin  reac  ion  is  the  only  method  by  which  the  relationship 
between  the  two  is  demonstrable  at  all,  and  there  too,  we  have  to 
reckon  with  what  is  generally  spoken  of  as  the  non-specific  increased 
sensitiveness  of  the  syphilitic  skin. 

Were  it  not  for  the  production  of  lesions  with  cultures  in  their 
early  test  tube  generations  by  Hoffmann175  and  by  Noguchi  in  a  few 
experiments,  one  would  be  almost  in  doubt  as  to  the  identity  of  the 
virulent  with  the  culture  organisms. 

The  reactions  in  syphilis  between  the  invading  microorganism 
and  the  invaded  subject  thus  differ  in  certain  fundamental  premises 
from  those  prevailing  in  diseases  caused  by  most  bacteria.  The 
treponema  pallidum  is  an  organism  which,  unlike  many  bacteria,  is 
rarely  subjected  to  the  necessity  of  adapting  itself  to  extra-corporeal 
existence  during  the  interval  between  its  passage  from  one  host  to 
another.  It  practically  always  infects  directly,  being  inoculated  from 
one  human  being  to  the  next  and  has  in  consequence  developed  a  very 
delicate  parasitism  not  unlike  that  seen  in  certain  trypanosome 
diseases  of  rats  and  that  which  we  ourselves  have  observed  in  the 
well-known  spirochete  infections  of  white  mice.  A  considerable 
percentage  of  laboratory  white  mice  have  been  found  to  harbor 
actively  motile  spirochetes,  often  in  considerable  numbers  in  the 
blood  and  peritoneal  fluid  without  there  being  any  objective  signs  of 
illness  in  the  animals.  It  is  an  instance  of  what  has  been  spoken  of 
by  Bail  as  "infection  without  disease,"  and  approaches  what  biolo- 
gists speak  of  as  symbiosis,  except  that  the  host  in  this  case  does  not 
benefit  in  any  way  by  the  invasion.  Even  in  the  case  of  the  mice, 
a  certain  amount  of  gradual  injury,  perhaps  only  metabolic,  by  the 
slow  removal  of  nutritive  material,  is  taking  place.  In  syphilis  the 
mutual  adaptation  may  perhaps  be  less  complete,  a  sufficient  accumu- 
lation of  the  invaders  and  especially  a  mechanical  injury  of  tissue 
cells,  of  closing  of  tissue  spaces  together  with  a  certain  amount  of 
toxic  action,  leading  eventually  to  pathological  changes. 

The  virulent  treponemata  apparently  do  not  arouse  true  antibody 
formation  in  any  marked  degree.  When  they  have  been  cultivated 

175  Hoffmann,  W.  H.    Deutsch.  med.  Wchnschr.,  1911,  xxxvii,  1546. 


518  INFECTION  AND  RESISTANCE 

and  have  become  accustomed  to  the  test  tube  conditions  they  entirely 
lose  their  virulence,  are  easily  attacked  by  the  active  constituents  of 
animal  serum,  and  are  probably  amenable  to  phagocytosis.  When 
such  cultures,  living  or  dead,  are  injected  into  animals  they  act  like 
other  specific  protein  antigens  and  incite  the  formation  of  antibodies. 
However,  these  antibodies  have  no  effect  whatever  upon  the  virulent 
organisms. 

In  consequence,  it  cannot  astonish  us  that  all  efforts  at  passive 
immunization  with  the  sera  of  syphilitic  man  and  animals,  or  with 
those  of  animals  systematically  treated  with  dead  virulent  materials, 
have  been  unqualified  failures. 

However,  this  does  not  preclude  the  theoretical  possibility  of 
active  immunization  or  vaccination  with  such  materials  since,  in  this 
case,  the  antigen  distributed  from  the  points  of  injection  might  act 
upon  tissue  cells  throughout  the  body.  However,  with  the  exception 
of  the  unconfirmed  reports  of  Spitzer,  all  attempts  to  vaccinate 
either  with  dead  virulent  material,  or  with  living  and  dead  culture 
material,  have  been  disappointing.  The  few  experiments  of  Metch- 
nikoff  with  virus  "attenuated"  by  passage  through  monkeys  have 
indeed  seemed  to  indicate  some  possibility  of  approaching  the  subject 
from  this  direction,  but  these  isolated  observations  have  been  very 
logically  criticized  by  Neisser  and  should  not  bear  too  much  weight. 
The  observations  were  made  on  two  cases  only,  both  of  them  well 
along  in  life,  and  the  validity  of  the  important  conclusions  drawn 
rests  entirely  on  the  always  problematical  fulcrum  of  complete  ex- 
clusion of  previous  syphilitic  infection  in  the  two  subjects.  More- 
over, attempts  in  this  direction  would  be  fraught  with  a  considerable 
amount  of  danger  and  it  is  therefore  questionable  whether  experi- 
mentation along  these  lines  is  sufficiently  promising  to  be  justified. 
There  is  certainly  no  attenuation  for  man  by  passage  through  rabbits 
as  has  been  sufficiently  proven  by  a  number  of  accidental  infections, 
an  instance  of  which  in  a  laboratory  attendant  has  been  reported  by 
Graetz  and  Delbanco.176 

We  may  state,  therefore,  as  safely  summarizing  our  knowledge 
of  the  conditions  in  syphilis,  that  the  resistance  which  undoubtedly 
develops  during  the  course  of  the  disease  is  one  which  depends  upon 
reaction  to  the  living  virus  only,  cannot  so  far  be  produced  in  animals 
by  systemic  treatment  with  dead  treponemata,  and  does  not  express 
itself  in  the  formation  of  significant  amounts  of  circulating  anti- 
bodies analogous  to  those  observed  in  bacterial  diseases.  Moreover, 
it  is  a  well-known  fact  that  the  treponemata  can  continue  to  do  injury 
to  many  organs  and  tissues  at  a  time  when  reinfection  by  the  paths 
of  skin  and  mucous  membranes  is  no  longer  possible. 

How,  then,  are  we  to  explain  this  peculiar  state  of  affairs?  A 
clue  to  the  problem  we  think  is  found  in  the  20  rabbits  which  Hop- 

176  Graetz  and  Delbanco.     Med.  klin.,  1914,  375  and  420. 


THERAPEUTIC  IMMUNIZATION  IN  MAN  519 

kins,  McBurney  and  the  writer  177  inoculated  into  the  testes  after 
apparent  recovery  from  a  previous  lesion.  Ordinarily  in  rabbits, 
as  we  have  stated  before,  no  generalized  resistance  is  developed  dur- 
ing the  disease,  and  the  opposite  testis  can  be  successfully  inoculated 
before,  during,  and  after  the  existence  of  a  lesion  on  the  other  side. 
In  these  rabbits  it  was  found  that  testes  that  had  apparently  recov- 
ered from  a  previous  lesion,  were  not  subsequently  as  easily  infected 
as  were  normal  testes.  It  has  seemed  to  us  from  this  as  well  as  from 
a  careful  study  of  the  observations  of  other  investigators,  that  re- 
sistance in  syphilis  was  probably  a  matter  of  localized  reaction. 
Tissues  which  have  sustained  active  invasion  with  the  living  virus 
react  and  gain  thereby  a  certain  degree  of  resistance  which  expresses 
itself  in  a  failure  to  react  to  subsequent  inoculation.  This  would 
explain  why  in  syphilis  of  the  human  being  reinoculation  is  unsuc- 
cessful and  reinfection  of  the  skin  and  mucosse  does  not  occur  spon- 
taneously at  a  period  later  than  the  early  secondary  stages  when  the 
virus  has  become  systemically  distributed.  It  would  furthermore 
explain  why  in  this  disease  organ  after  organ  may  be  pathologically 
involved  when  skin  infection  is  no  longer  possible.  This  point  of 
view  is  entirely  in  harmony,  though  perhaps  from  a  slightly  different 
aspect,  with  the  skin  immunity  suggested  by  Kraus  and  Volk,  where 
the  resistance  is  attributed  to  the  tissue  as  a  whole  rather  than  to  a 
local  cell  group.  The  cells  which  have  once  reacted  to  the  living 
virus  no  longer  respond,  i.  e.,  can  no  longer  be  injured  for  an  in- 
definite period  after  recovery.  "However,  the  factors  which  lend 
them  this  resistance,  whatever  they  may  be,  are  not  distributed  to 
the  blood  stream  in  a  way  analogous  to  that  in  which  antibodies  are 
mobilized  in  bacterial  diseases,  and  the  effect  of  the  resistance  of  the 
local  area  is  not  distributed  to  remote  parts  of  the  body.  It  is  of 
course  likely  that  a  certain  amount  of  phagocytosis  of  treponemata 
by  the  now  resistant  fixed  cells  may  account  for  the  absence  of  local 
injury.  This,  however,  we  have  not  yet  been  able  to  prove  suffi- 
ciently, and  further  studies  on  this  point  are  necessary.  As  far  as 
any  positive  evidence  can  be  adduced  at  the  present  time,  the  newly 
entering  treponemata  may  not  be  entirely  destroyed.  It  may  well 
be  that  the  tissues  do  not  react  and  are  in  the  state  which  Neisser 
calls  "Anergie."  The  treponemata  that  enter  during  this  period 
may  nevertheless  remain  uninjured  and  be  as  capable  of  subsequently 
causing  lesions  in  other  locations  as  are  those  already  present  in  the 
patient.  This  phase  is  being  studied  by  comparative  histological 
observation  on  the  fate  of  treponemata  which  have  been  injected  into 
normal  tissues  and  into  tissues  rendered  resistant  by  previous  in- 
fection. 

In  animals  like  monkeys  and  man  where  generalization  is  rapid 

177  Zinsser,  H.,  Hopkins,  J.  G.,  and  Gilbert,  R.    Jour.  Exper.  Med.,  1915, 
xxi,  213. 


520  INFECTION  AND  RESISTANCE 

and  apparently  complete,  the  resistance  becomes  a  general  one.  In 
animals  like  the  rabbit  in  which  the  lesion — or  in  other  words 
pathological  response,  occurs  in  a  few  organs  only,  the  resistance  is 
limited  to  the  particular  organ  or  organs  that  have  previously  de- 
veloped a  lesion. 

It  must  not  be  forgotten  that  such  a  .resistance  probably  persists 
for  a  limited  time  only,  and  does  not  imply  the  sterilization  of  the 
body  and  the  complete  destruction  of  the  microorganisms.  These 
may,  and  probably  do,  remain  alive  and  potent  in  various  parts  of  the 
body,  capable  of  again  setting  up  new  lesions  in  parts  hitherto  unin- 
volved  or  again  susceptible  after  a  diminution  of  their  local,  acquired 
resistance.  That  the  virulent  treponemata  may  remain  thus  latent, 
alive  and  virulent  has  been  sufficiently  shown  in  animal  experimen- 
tation by  successful  inoculation  with  tertiary  lesions  and  by  the 
frequent  late  accidents,  especially  of  the  nervous  system,  in  indi- 
viduals apparently  cured  or  for  a  long  time  without  symptoms. 

It  would  seem  when  we  analyze  the  conditions  in  syphilis,  that 
complete  sterilizing  immunity  or,  in  other  words,  complete  cure, 
occurs  but  rarely  without  specific  medicinal  aid,  and  that  the  un- 
treated syphilitic  (if  such  an  unfortunate  individual  exists  in  a 
civilized  country)  might  go  on  to  apparent  cure,  in  that  a  general 
syphilization  of  his  body  would  bring  about  a  general  resistance, 
but  would  always  harbor  virulent  treponemata  which  could  cause 
recrudescences  in  parts  in  which  resistance  was  diminished,  and 
eventually  kill  by  degenerative  processes  in  the  central  nervous  sys- 
tem where  many  injuries  cannot  be  compensated  for  as  is  possible  in 
other  organs. 

The  resistance  which  develops  is  apparently  a  new  attribute  only 
of  the  cell  groups  which  have  undergone  direct  reaction  with  the 
treponemata.  This  resistance  may  consist  merely  in  the  complete 
failure  of  the  tissue  cells  to  react  to  the  virus,  a  sort  of  "tissue  indif- 
ference" or  "Anergie."  It  may  be,  however,  and  probably  is,  ac- 
companied by  a  certain  amount  of  active  defense  in  the  form  of  local 
phagocytosis  of  the  treponemata  by  the  fixed  tissue  cells. 


THE  INFLUENCE  OP  INJECTIONS  OF  NON-SPECIFIC  SUB- 
STANCES   UPON   INFECTIOUS   DISEASES 

Therapy  of  infectious  diseases  has  been  very  logically  dominated 
in  the  past  by  attempts  to  increase  resistance,  either  passively,  by 
the  injection  of  specific  anti-sera,  or  actively,  by  treatment  with  bac~ 
terial  antigens  or  their  derivatives.  The  idea  of  specificity,  in  other 
words,  has  dominated  such  attempts  almost  exclusively.  In  spite  of 
this,  however,  there  has  gradually  grown  in  the  minds  of  bacteriolo- 
gists an  impression  that  not  all  the  effects  of  the  injection  of  bacterial 


THERAPEUTIC  IMMUNIZATION  IN  MAN  521 

protein  were  purely  specific.  Jobling,  who  has  recently  summarized 
most  of  this  work,  calls  attention  to  the  fact  that  Matthes178  as 
early  as  1895  showed  that  the  effect  of  tuberculin  injection  could 
be  obtained  equally  well  with  deuteroproteose.  From  that  time  on, 
many  observations  have  been  made  to  show  that  often  profound 
physiological  effects  could  be  produced  in  human  beings  and  animals 
suffering  from  infectious  diseases,  if  they  were  treated  not  with 
specific  antigen  but  with  proteins  and  protein  derivatives  of  many 
sources. 

The  work  that  the  writer  did  with  Hiss  on  the  injection  of 
leucocytic  extracts  is  a  case  in  point.  Similar  in  significance  prob- 
ably are  the  results  obtained  by  the  injection  of  substances  such  as 
the  filtrate  of  bacterial  cultures,  commercially  sold  as  phylacogens, 
the  intramuscular  milk  injection  practiced  by  Schmidt179  and  the 
injection  of  various  ferments,  reported  by  a  number  of  writers 
throughout  the  literature  of  the  past  ten  or  fifteen  years.  Into  the 
same  category  fall  the  favorable  reports  supposed  to  have  been  ob- 
tained in  syphilitics  in  tuberculin  exhibits  by  Blach.180  Perhaps 
the  clearest  example  of  this  principle  has  been  obtained  in  connection 
with  attempts  at  vaccine  therapy  in  typhoid  fever.  Attempts  to 
cure  typhoid  fever  by  the  injection  of  typhoid  bacilli  date  back  to 
Fraenkel,  who  treated  it  as  early  as  1893,  and  since  then,  as  we  have 
seen,  extensive  study  has  been  made  on  this  subject  by  Petruschky, 
Isletter,  and  others.  Ichikawa  in  1912  and  Boinet  in  1914  obtained 
astonishing  results  by  the  use  of  sensitized  vaccine,  the  former  being 
the  first  to  introduce  the  intravenous  method  of  injection.  The 
results  of  the  injection  intravenously  into  a  patient  suffering  from 
typhoid  fever  have  consisted  in  rapid  falling  of  temperature,  often 
followed  by  a  chill,  with  occasionally  rapid  general  improvement  of 
the  patient. 

Gay,  too,  has  recently  made  such  observations  by  the  methods 
of  Ichikawa,  and  most  of  these  writers  were  inclined  to  believe  that 
their  opinion  was  effective  by  some  sort  of  specific  reaction.  Doubt 
has  been  cast  upon  this  point  of  view,  however,  by  observations  such 
as  those  of  Kraus.  Ichikawa  alone  had  some  doubt  of  this,  as  is 
indicated  by  the  fact  that  he  injected  typhoid  bacilli  into  some  of  his 
paratyphoid  patients  with  like  results.  Kraus  subsequently  ob- 
tained similar  effects  by  injecting  colon  bacilli  into  typhoid  patients 
and  used  non-specific  bacterial  virus  with  good  results  on  cases  of 
pyocyaneus  infection  and  upon  a  single  streptococcus  puerperal  septi- 
cemia.  Leudke  injected  typhoid  patients  with  non-bacterial  proteose 
and  Jobling  and  Petersen  have  obtained  very  marked  reactions  in 

178  Matthes,  M.     Deutsch.  Arch.  f.  klin.  Med.,  1895,  Vol.  45. 

179  Schmidt,  R.     Med.  Klin.,  1910,  No.  43. 

180  Blach,  M.     Wien.  klin.  Wchnschr.,  1915,  No.  49. 


INFECTION  AND  RESISTANCE 

typhoid  patients  by  injecting  them  intravenously  with  quantities  of 
1  to  2  c.  c.  of  2  per  cent,  solution  of  proteoses. 

There  seems  to  be  little  doubt  of  the  fact  that  the  injection  of 
bacterial  proteins  intravenously  produces  a  profound  therapeutic  ef- 
fect without  being  in  any  but  a  minor  degree  specific.  Boinet  is  still 
an  adherent  of  specific  reaction,  believing  that  his  typhoid  patient  was 
benefited  by  a  rapid  mobilization  of  antibodies  and  Gay  had  the  idea 
that  a  specific  hyperleucocytosis  was  the  basis  of  improvement.  This 
last  phenomenon  has  been  discussed  in  another  place.  We,  ourselves, 
are  inclined  to  believe  that  the  specific  reactions  have  little  to  do 
with  the  improvement  in  such  cases  and  that  the  results  are  due  to 
the  injection  of  foreign  protein  and  perhaps  proteoses,  the  specific 
nature  of  the  sources  of  these  have  little  significance.  We  believe 
that  one  of  the  most  important  factors  of  such  procedure  is  the  rapid 
mobilization  of  leucocytosis  which  is  not  specific,  in  the  sense  of 
Gay  and  Claypole.  In  addition  to  this,  as  Jobling  and  Petersen 
suggest,  the  hyperpyrexia  which  is  often  observed  may  have  some- 
thing to  do  with  it  as  well  as  the  mobilization  of  ferments  which 
the  last  named  writers  have  observed.  They  have  shown  that  the 
intravenous  injection  of  bacteria  and  of  other  protein  substances 
is  followed  by  a  marked  mobilization  of  serum  protease  and  lipase. 
They  suggest  that  these  factors  may  have  considerable  influence  on 
the  result. 

The  subject  is  hardly  begun  but  it  is  being  made  the  motive  of 
many  researches  both  in  the  clinic  and  in  the  laboratory  and  un- 
questionably the  "ext  few  years  will  bring  much  further  under- 
standing. 


CHAPTER   XX 

SERUM  ENZYMES;  LEUCOCYTIC  ENZYMES;  ABDER- 
HALDEIST  REACTION;  PHYSICAL  PRINCIPLES  IN 
SERUM  REACTION;  MEIOSTAGMIN  AND  EPIPHA- 
NIN  REACTIONS;  COLLOIDAL  GOLD  REACTION 

IN  our  section  upon  the  nature  of  the  precipitins,  especially  in 
the  discussion  of  Gengou's  conception  of  "albuminolysins,"  we  have 
called  attention  to  the  probable  significance  of  protein  antibodies  as 
a  mechanism  for  the  disposal  of  such  foreign  substances.  In  the 
bodies  of  the  higher  animals  in  which  a  special  alimentary  system, 
with  its  many  digestive  ferments,  is  well  developed,  it  is  most  prob- 
able that  the  normal  condition  of  digestion  is  one  in  which  the 
foreign  substances  utilized  for  nutrition  are  completely  split  into 
their  simpler  components  before  they  gain  entrance  to  the  circulation. 
Nevertheless,  abnormal  conditions  or  accidents,  such  as  gastro-enteric 
diseases,  digestive  disturbances,  and  bacterial  infections,  may  lead  to 
a  condition,  probably  frequent  enough  in  ordinary  life,  during  which 
such  foreign  substances  may  get  into  the  blood  stream  without  pre- 
vious cleavage.  The  problem  is  to  determine  where  and  how  such 
substances,  protein  or  otherwise,  are  broken  up  so  that  they  may  be 
either  assimilated  or  eliminated.  We  have  referred  in  another  place 
to  the  fact  that  foreign  proteins  may  occasionally  pass  through  the 
kidneys  and  be  eliminated  unchanged.  This  has  been  shown  actually 
to  occur  by  Oppenheimer,  Ascoli,  and  others,  but  probably  represents 
a  very  unusual  state  of  affairs  produced  by  special  experimental 
conditions.  As  a  rule  these  substances  are  disposed  of  within  the 
body  by  chemical  cleavage  or  by  assimilation.  It  is  more  than  likely, 
therefore,  as  has  been  emphasized  by  such  workers  as  Jacoby,  Sal- 
kowski,  Wells,  and  others,  that  every  fluid  and  cell  in  the  body  con- 
tains enzymes  which,  by  their  action  upon  proteins,  fats,  and  carbo- 
hydrates, play  an  important  part  in  the  metabolism  of  the  body. 
It  has  been  known  for  a  long  time  that  the  organs  of  animals  re- 
moved from  the  body  will  undergo  self-digestion  by  a  process  spoken 
of  as  autolysis.  The  action  of  bacteria  as  causing  such  autolysis  can 
be  excluded  by  covering  organ  emulsions  with  toluol,  chloroform, 
and  some  other  substances  which  have  the  peculiar  property  of  pre- 
venting the  growth  of  bacteria  without  in  any  way  interfering  with 
the  action  of  enzymes.  Thus,  any  organ  of  the  body  so  treated  will 
show  rapid  changes  consisting  in  the  splitting  of  its  own  proteins, 

523 


524  INFECTION  AND  RESISTANCE 

fats,  and  carbohydrates.  That  such  processes  also  go  on  in  the  living 
body  is  probable,  although,  as  we  have  stated  in  an  earlier  chapter, 
it  is  a  well-known  fact  that  living  cells  oppose  a  more  or  less  mysteri- 
ous resistance  to  enzymic  digestion,  just  as  they  oppose  a  similar 
resistance  to  bacterial  invasion.  The  phenomenon  is  more  fully 
discussed  by  Wells  in  his  "Chemical  Pathology,"  to  which  the  reader 
is  referred.  It  is  not  improbable  that  this  resistance  to  invasion  and 
enzyme  attacks,  spoken  of  vaguely  as  "vital  resistance,"  can  be  anal- 
yzed into  more  exact  factors,  one  of  which  seems  to  be  the  question 
of  reaction,  digestion  depending  to  some  ex:ent  upon  the  reduction 
of  alkalinity  by  the  formation  of  acid  in  the  tissue  cells. 

However,  the  physiological  importance  of  tissue  enzymes  is  a 
subject  which  we  cannot  go  into  in  this  connection,  since  the  rela- 
tionship of  these  enzymes  to  normal  body  metabolism  is  a  subject 
more  fully  dealt  with  in  text  books  of  physiology  and  is  too  far 
removed  from  the  subject  of  our  present  discussion  to  warrant  ex- 
tensive analysis.  It  must,  of  course,  be  self-evident  that  the  existence 
of  cellular  ferments  of  this  nature  must  play  an  important  role 
wherever  and  whenever  cell  death  occurs,  and  since  such  cell  death 
is  an  accompaniment  of  many  phases  of  bacterial  infection  it  may 
well  be  that  the  enzymes  which  destroy  dead  cells,  whether  they  be 
bacterial  or  those  of  the  body  itself,  contribute  to  the  general  patho- 
logical picture  characterizing  such  diseases.  Moreover,  the  impor- 
tance of  the  enzymes  is  particularly  enhanced  by  the  knowledge  we 
have  gained  from  the  work  of  Vaughan  and  others,  who  have  shown 
that  in  the  course  of  proteolytic  cleavage  toxic  substances  are  lib- 
erated. It  is  such  proteolysis  which  is  the  probable  basis  of  the 
formation  of  the  "anaphylatoxin"  of  Friedberger,  the  "proteotoxin" 
about  which  we  ourselves  have  written,  and  the  "sereotoxin"  of 
J  obling  and  Petersen ;  and  similar  processes  are  involved  in  the  toxic 
substances,  studied  by  Whipple,  which  form  in  the  intestine  as  a 
consequence  of  ligature  of  the  gut.  It  is  conceivable  that  wherever 
and  whenever  a  proteolytic  ferment  reacts  in  the  body  with  its 
substrate,  toxic  cleavage  products  result,  perhaps  in  the  form  of 
albumoses,  peptones,  etc.,  which  are  rapidly  absorbed  and  cause 
symptoms  of  varying  intensity.  In  the  chapter  on  anaphylaxis  we 
have  seen  that  many  writers  have  attributed  the  injury  occurring  in 
this  phenomenon  to  the  formation  of  such  split  products  and  there  is 
much  evidence  in  favor  of  such  a  view,  although  the  rapidity  and 
vehemence  of  anaphylactic  shock  indicate  that  probably  there  are 
purely  physical  elements  also  involved  which  we  do  not  as  yet  com- 
prehend. 

Of  the  cells  in  the  body  with  which  enzyme  study  has  been  most 
assiduously  followed,  the  most  important  are  the  leucocytes.  We 
have  already  had  much  to  say  in  the  chapters  on  phagocytosis  of  the 
digestive  functions  of  the  white  blood  cells.  However,  we  dealt  with 


SERUM  ENZYMES  525 

them  there  chiefly  from  the  point  of  view  of  their  phagocytie  and 
antibacterial  functions.  It  will  be  useful  to  consider  at  a  little 
greater  length  the  importance  of  these  cells  from  the  point  of  view 
of  purely  enzymotic  activity.1 

As  early  as  1885  Hammarsten  called  attention  to  the  fact  that 
leucocytes  aid  in  dissolving  fibrin.  Leber  2  subsequently  studied  pus 
and  found  that  it  possessed  a  powerful  digestive  action  for  gelatin, 
fibrin,  and  other  protein  substances.  He  correlated  his  studies  par- 
ticularly with  the  processes  of  inflammation,  and  his  work,  as  well 
as  that  of  later  investigators,  brought  into  the  foreground  the  fact 
that  in  the  resolution  of  abscesses,  especially  of  the  staphylococcus 
variety,  the  leucocytes  might  play  an  important  role  in  liquefying 
the  necrotic  tissue  cells  and  bringing  about  the  breaking  down  of 
the  center  of  the  abscess.  Any  one  who  has  carefully  studied  the 
histology  of  a  staphylococcus  abscess  can  easily  see  the  halo  of  disin- 
tegrating tissue  lying  just  inside  the  ring  of  aggregated  polymorpho- 
nuclear  leucocytes.  That  these  cells  carry  on  an  important  part  in 
this  liquefaction  is  certain,  although  as  yet  we  are  not  sure  whether 
or  not  the  proteolytic  bacterial  enzymes  participate.  Friedrich  Mill- 
ler,3  in  1902,  also  studied  these  processes  and  gave  clinical  signifi- 
cance to  the  proteolytic  activities  of  the  white  blood  cells  in  the 
resolution  of  the  pneumonic  lung.  The  study  of  inflammatory  exu- 
dates  by  Opie  4  confirmed  much  of  the  preceding  work  and  revealed 
that  the  ferments  of  all  white  blood  cells  were  not  the  same.  He 
found  a  leucoprotease  which  was  contained  in  the  polynuclear  leuco- 
cytes which  was  active  chiefly  in  a  slightly  alkaline  medium,  whereas 
the  large  mononuclear  cells  of  uncertain  origin  which  appear  toward 
the  end  of  an  inflammatory  process  contained  a  protease  which  was 
active  in  weak  acid.  This  differentiation  of  a  leucoprotease  and  a 
lymphoprotease  was  more  or  less  confirmatory  of  assumptions  made 
earlier  by  MetchnikofL  It  seemed  probable  that  the  acid  protease 
noticed  by  Opie  is  specific  to  the  large  mononuclear  cells  of  such 
exudates  and  is  not  common  to  lymphocytes  in  general,  for  the  con- 
sensus of  opinion  of  other  workers  seems  to  be  that  the  small  lympho- 
cytes contain  no  proteolytic  enzyme  whatever.  The  study  of  tubercu- 
lous exudates  particularly  has  failed  to  reveal  proteolytic  properties 
when  only  the  characteristic  small  lymphocytes  were  present,  al- 
though some  writers,  Bergell  5  especially,  have  shown  that  these  cells 
contain  a  fat-splitting  enzyme.  Such  lipase  could  also  be  determined 
in  the  press-juice  of  lymphoid  organs,  such  as  the  spleen  and  lymph 
nodes.  It  is  not  improbable  that  the  characteristic  dry  abscess  or 

1  See  Wiens.    Ergebnisse  d.  Allg.  Pathol.    Lubarsch  &  Ostertag,  Vol.  15, 
1911. 

2  Leber.     Entstehung  der  Entzundung.     Leipzig,  1891. 

3  Fr.  Miiller.     Verh.  d.  20  Kongress  f.  innere  Mediz.,  Wiesbaden,  1902. 

4  Opie.    Journal  of  Exp.  Med.,  Vols.  7  and  8,  1905  and  1906. 

5  Bergell.     Munch,  med.  Woch.,  1909. 


526  INFECTION  AND  RESISTANCE 

"caseous"  abscess  accompanying  typical  tuberculous  lesions  owes  its 
peculiar  histological  characteristics  to  the  lack  of  proteolytic 
enzymes. 

Jochmann  6  and  his  collaborators  have  extensively  studied  leuco- 
cytic  extracts  in  this  regard  and,  curiously  enough,  found  that  the 
proteolytic  enzymes  of  leucocytes  of  which  we  have  been  speaking 
can  be  found  only  in  the  cells  of  man  and  monkeys  and,  to  a  slight 
degree,  in  dogs.  In  these  species  only,  according  to  Pappenheim,7 
do  the  leucocytes  contain  true  neutrophile  granules  and  it  is  there- 
fore a  possibility  that  the  neutrophile  granules  and  the  enzyme  action 
are  related  to  each  other.  The  leucocytes  of  rabbits  do  not  ap- 
parently contain  protease,  but  recently  in  our  laboratory  Mrs.  Parker 
and  Miss  Francke  have  shown  that  rabbit  leucocytes  contain  erepsin. 

The  proteolytic  enzymes  of  leucocytes  curiously  enough  continue 
their  activity  at  temperatures  as  high  as  55°  C.,  a  fact  which  makes 
it  possible  to  investigate  their  activity  at  temperatures  at  which  most 
bacteria  will  no  longer  grow  and  functionate,  and  this  incidentally 
facilitates  sterile  experimentation.  Their  activity  is  astonishing  in 
that  Jochmann  found  instances  where  dilution  even  as  high  as  500- 
fold  with  salt  solution  did  not  completely  eliminate  the  proteolytic 
activity  of  pus.  The  simplest  method  of  demonstrating  such  action 
is  to  place  pus  or  washed  leucocytes,  in  droplets,  upon  the  surface  of 
plates  of  Loeffler's  coagulated  blood  serum,  such  as  that  used  in  the 
cultivation  of  diphtheria  bacilli.  On  such  plates,  small  indentations 
rapidly  give  evidence  of  the  liquefaction  of  the  coagulated  protein. 

Casein  also  can  be  used  as  an  indicator  of  proteolytic  digestion, 
a  casein  solution  being  made  by  dissolving  a  gram  of  casein  in  100 
c.  c.  of  N/10  NaOH  solution  and  neutralizing  this  to  litmus  with 
deci-normal  HC1.  A  very  curious  property  of  the  leucocytic  ferment 
has  been  described  by  Jochmann  and  Ziegler,  who  reported  that  pres- 
ervation in  10  per  cent,  formalin  solution  will  long  preserve  the  fer- 
mentative activities  of  the  cells. 

It  was  formerly  supposed  that  the  proteolytic  properties  of 
leucocytes  were  more  or  less  pathological.  But  we  have  since  learned 
that  these  powers  are  common  to  leucocytes  both  in  health  and  in 
disease.  The  older  idea  was  due  to  the  fact  that  the  earliest  investi- 
gations of  the  ferments  were  made  in  connection  with  myelogenous 
leukemia.  Jochmann  and  Ziegler  studied  the  organs  of  patients 
dead  of  this  disease  and  found  that  the  bone-marrow,  spleen,  and 
lymph  nodes  of  such  cases  had  powerful  proteolytic  properties  in 
contrast  to  the  very  weak  activities  in  this  regard  of  normal  organs. 
The  proteolytic  activity  was  more  or  less  in  direct  proportion  to  the 
degree  of  myelogenous  infiltration.  In  lymphatic  leukemia  no  such 

6  Jochmann.     For   summary   of   his   work   and   literature   see   Kolle   u* 
Wassermann,  2nd  ed.,  Vol.  II,  2. 

7  Pappenheim.     Cited  from  Wiens,  loc.  cit. 


SERUM  ENZYMES  527 

activity  was  discovered.  In  consequence  there  has  been  a  tendency 
to  draw  a  fundamental  physiological  distinction  between  these  types 
of  cells,  a  distinction  which  would  have  considerable  importance 
in  discussions  of  their  origin  and  significance. 

Jochmann  has  an  idea  that  the  activities  of  these  ferments  in 
connection  with  tissue  destruction  and  other  pathological  conditions 
where  they  become  active,  may  be  an  important  phase  in  the  produc- 
tion of  fever. 

An  important  question  which  immediately  arises  is  whether  these 
ferments  can  be  identified  with  the  bactericidal  substances  described 
in  a  preceding  chapter  as  existing  within  leucocytes.  According  to 
Jochmann  the  enzyme  extracts  have  no  bactericidal  properties.  In- 
deed, it  is  a  curious  fact  that  living  bacteria  oppose  a  very  powerful 
resistance  to  digestion  by  these  and  other  ferments.  Kantorowicz,8 
who  has  studied  this  particularly,  has  shown  that  living  bacteria 
contain  a  very  powerful  antiferment  which  is  similar  to  the  anti- 
ferment  presently  to  be  described  for  blood  serum.  He  has  shown 
that  living  bacteria  cannot  be  digested  by  trypsin,  a  resistance  which 
is  lost  when  the  bacteria  are  heated  to  80°  C.,  and  an  extract  of 
living  bacteria  will  prevent  the  tryptic  digestion  of  the  heated  bac- 
teria. It  will  appear,  therefore,  that  in  the  process  of  phagocytosis 
the  bacteria  are  first  killed  by  the  bactericidal  substances  contained 
in  the  cells  and  are  later  digested  by  the  leucoprotease. 

That  the  lymphocytes  contain  lipase  has  been  mentioned  above 
and  since  these  cells  are  specifically  accumulated  about  tuberculous 
foci  it  has  been  many  times  suggested  that  their  function  is  particu- 
larly directed  against  these  acid-fast  bacteria  in  whose  constitution 
the  presence  of  waxes  and  fats  plays  such  an  important  role.  It  is 
a  fascinating  thought,  though  entirely  conjectural,  that  perhaps  the 
specific  benefit  of  feeding  fats  in  tuberculosis  may  have  some  basis  in 
the  possible  increase  of  lipolytic  ferments  which  appear  in  response 
to  the  stimulation  of  introducing  larger  quantities  of  fats. 

It  is  hardly  necessary  to  reiterate  here  the  possible  importance 
of  these  cellular  enzymes  in  the  many  different  phases  of  the  ab- 
sorption of  larval  organs  which  occurs  in  the  lower  animals  in  the 
course  of  normal  development,  or  of  dead  tissue  in  mammalia  in 
disease  and  in  the  processes  of  senescence. 

In  addition  to  proteolytic  enzymes,  it  has  been  long  known  that 
the  leucocytes  also  contain  an  oxidase.  This  enzyme  can  be  demon- 
strated by  the  well-known  guaiac  test,  in  which  tincture  of  guaiac  is 
added  to  leucocytic  exudates  or  pus,  and  a  blue  color  results  as  a 
consequence  of  oxidation  of  the  guaiac.  The  oxidizing  ferment  is 
apparently  limited  to  the  polynuclear  leucocytes.  Exactly  what  its 
significance  is  we  do  not  know  but  it  is  highly  probable  that  it  plays 
an  important  part  in  the  intracellular  metabolism  of  the  leucocyte. 

8  Kantorowicz.     Munch,  med.  Woch.,  p.  897,  Vol.  56,  1909. 


528  INFECTION  AND  RESISTANCE 

It  has  only  recently  become  clearly  apparent  that  the  function  of 
intracellular  digestion  of  foreign  substances,  such  as  invading  bac- 
teria, is  not  limited  to  the  circulating  leucocytes  but  is  to  a  very 
large  extent  also  carried  on  by  the  fixed  cells  of  organs.  Studies 
by  Wyssakowitz,9  by  Kyes,10  by  Bartlett,11  and,  recently,  by  Hopkins 
and  Parker,12  have  shown  that  in  the  course  of  many  bacterial  infec- 
tions the  greater  proportion  of  invading  bacteria  is  taken  care  of 
by  cells  of  the  liver,  spleen,  and  especially  the  lung.  Streptococci 
injected  into  rabbits  rapidly  disappear  from  the  circulation,  largely 
because  they  are  taken  up  very  actively  by  the  (probably)  endothelial 
cells  of  the  lung  where  apparently  they  are  killed,  their  digestion 
taking  place  within  the  cells  very  slowly.  The  endocellular  enzymes 
bringing  this  about  can,  of  course,  not  be  separately  studied  as  can 
those  of  leucocytes,  for  obvious  technical  reasons. 

While  the  cellular  enzymes  have  been  very  carefully  studied,  it 
is  only  of  relatively  recent  years  that  much  attention  has  been  paid 
to  the  enzymes  present  in  the  circulating  blood.  This  is  to  some  ex- 
tent due  to  the  fact  that  German  writers  especially  have  taken  for 
granted  that  enzymes  in  the  blood  were  nothing  more  or  less  than 
liberated  leucocytic  enzymes,  and  also  because  the  activity  of  the 
circulating  enzymes  has  been  largely  masked  by  the  existence  of  a 
powerful  antiferment  which  must  needs  be  there  for  physiological 
reasons  to  prevent  injurious  autodigestion.  We  have  already  men- 
tioned the  fact  that  the  study  of  pathological  conditions,  such  as 
myelogenous  leukemia,  was  the  point  of  departure  for  work  upon  the 
leucocytic  ferments,  and  this,  to  a  certain  extent,  was  also  the  be- 
ginning of  the  studies  on  serum  ferments.  When  blood  serum  is 
incubated  with  various  substrates,  it  is  not  unlikely,  as  we  shall  see 
in  our  subsequent  discussion  of  the  Abderhalden  reaction,  that  the 
antiferment  is  absorbed  and  thereby  the  proteolytic  and  other  fer- 
ments of  the  blood  are  liberated.  There  is  normally  an  excess  of 
antiferment  in  the  blood,  and  normal  human  serum  usually  does  not 
contain  strong  proteolytic  enzyme.  These  two  factors  together  there- 
fore prevent  powerful  proteolysis  by  normal  serum.  There  are  con- 
ditions, however,  under  which  the  enzyme  contents  of  the  blood  are 
increased.  Thus,  the  sera  of  pregnant  women  have  been  shown 
to  be  relatively  rich  in  proteolytic  properties,  and  such  an  increase 
is  also  present  during  starvation,  as  Schultz,13  and  Heilner  and 
Poensgen  14  have  shown.  The  same  thing  has  been  noted  in  cases 

9  Wyssakowitz.    Zeitschr.  f.  Hyg.,  Vol.  1,  p.  1. 

10  Kyes.    Journal  of  Inf.  bis.,  1916,  Vol.  18,  p.  277. 

11  Bartlett.     Journ.  Mcd.  Res.,  Nov.  5,  1916,  p.  465. 

12  Hopkins  and  Parker.     In  press  at  present  writing. 

13  Schultz.     Deutsch  med.  Woch.,  No.  30,  1908;  Munch,  med.  Wocn..  Vol. 
60,  1913. 

14  Heilner  and  Poensg-en.     Miinch.  med.  Woch.,  Vol.  61,  1914. 


SERUM  ENZYMES  529 

of  pneumonia,  in  which  condition  Falls  15  has  shown  a  sudden  dimi- 
nution of  this  enzyme  at  the  time  of  crisis,  and  the  same  author  has 
made  analogous  observations  in  the  fluctuation  of  serum  protease  in 
malaria,  typhoid  fever,  and  some  other  diseases. 

The  lipolytic  activity  of  the  serum  has  been  found  increased  in 
syphilis,  in  diseases  involving  liver  function,  such  as  chloroform 
poisoning,  and  in  a  number  of  other  conditions.  In  syphilis,  indeed, 
the  increased  lipase  contents  have  been  held  responsible  for  the  Was- 
sermann  reaction  by  a  number  of  writers.  They  believed  that  the 
lipolytic  ferment  acting  upon  the  lipoid  antigen  produced  fatty  acid 
which  by  increasing  the  H-ion  contents  rendered  the  complement 
inactive.  This  to  a  certain  extent  was  strengthened  by  the  knowledge 
that  in  the  Wassermann  reaction  the  end-piece  of  complement  re- 
mained free.  However,  we  ourselves  in  hitherto  unpublished  experi- 
ments have  been  able  to  convince  ourselves  that  there  is  no  relation- 
ship between  lipase  contents  and  Wassermann  positiveness. 

Activity  of  intravascular  enzymes  of  any  kind  may  perhaps  be 
going  on  constantly  in  the  normal  course  of  life  as  a  part  of  general 
body  metabolism ;  yet  it  is  probable  that  any  great  extent,  especially 
of  proteolytic  action,  would  be  incompatible  with  health  or  even 
life.  The  work  of  Vaughan,  of  Friedberger,  and  others  has  shown 
us  that  in  the  course  of  proteolysis  toxic  cleavage  products,  probably 
of  the  albumose  variety,  are  liberated.  This  we  have  seen  has  been 
made  the  basis  of  some  theories  of  anaphylaxis,  and  the  "split  prod- 
ucts" of  Vaughan,  the  "anaphylatoxin"  of  Friedberger,  the  "proteo- 
toxins,"  as  we  ourselves  have  chosen  to  call  them,  and  the  "sero- 
toxins"  of  Jobling  and  Petersen,  all  are  probably  the  results  of  such 
proteolysis.  Indeed,  as  we  have  seen,  the  incubation  of  almost  any 
tissue  substrate  with  fresh  serum  will  result  in  this  proteolytic- 
change  and  even  in  the  formation  of  toxic  substances.  It  was  at  first 
thought  that  the  toxic  split  products  were  derived  from  the  substrate, 
but  we  have  seen  in  another  chapter  that  the  mechanism  is  probably 
one  in  which  antiferments  are  removed  by  the  substrate,  and  the 
serum-protease  is  liberated  to  act  upon  the  serum  protein  itself. 
This  was  undoubtedly  the  case  in  such  instances  as  those  in  which 
anaphylatoxin  was  obtained  by  digestion  of  serum  with  kaolin, 
barium  sulphate,  tubercle  bacilli,  etc.  Now  it  appears  that  the 
injection  of  almost  any  substance,  and  especially  of  bacterial  pro- 
teins, produces  a  sharp  rise  of  enzymes  in  the  blood.  This  seems  to 
be  true  not  only  of  proteolytic  enzymes  but  of  lipolytic  and  amylolytic 
ferments  as  well.  Ferments  so  stimulated  are  entirely,  non-specific. 
This  greater  serum  activity  after  protein  injection  may  be  due  both 
to  a  direct  increase  of  production  and  a  simultaneous  diminution  of 
antiferments,  by  means  of  which  activity  hitherto  inhibited  is  al- 
lowed to  run  riot.  The  result  may  be  in  some  respects  beneficial 
15  Falls.  Journ.  of  Inf.  Dis.,  Vol.  16,  p.  466. 


530  INFECTION  AND  RESISTANCE 

and  in  other  respects  injurious.  The  marked  increase  of  the  enzymes 
may  aid  in  rapidly  disposing  of  the  unchanged  foreign  substance, 
and  perhaps  the  beneficial  effects  following  the  injection  of  typhoid 
vaccines  into  patients  suffering  from  this  disease  may  be  due  to  the 
mobilization  of  ferments  as  suggested  by  Jobling  and  Petersen.16 

On  the  other  hand,  the  proteolytic  activity  may  result  in  protein 
cleavage,  by  means  of  which  albumoses,  etc.,  are  liberated  in  greater 
or  smaller  quantities  which  act  injuriously  and  lead  to  clinical 
symptoms  of  various  kinds.  That  fever  may  perhaps  be  explained 
by  poisoning  with  albumoses  so  produced  has  already  been  suggested 
by  Jochmann,  and  that  albumoses  are  present  in  various  products 
of  suppurative  inflammation  has  also  been  shown.  It  has  been  shown' 
by  Pfeiffer  and  Jarisch,17  and  others,  that  fluctuations  in  proteolytic 
enzymes  accompany  anaphylaxis.  Jobling  and  Petersen18  have 
formulated  a  theory  of  anaphylaxis  based  upon  studies  of  serum 
protease.  They  observed  that  during  the  course  of  sensitization 
there  occurred  a  gradual  mobilization  of  non-specific  protease  which 
increased  in  intensity  up  to  the  time  of  maximum  sensitization. 
They  attributed  acute  shock  to  the  fact  that  on  the  second  injection 
there  occurs  an  instantaneous  mobilization  of  large  amounts  of  non- 
specific protease  together  with  a  decrease  in  antiferment  and  an 
increase  in  non-coagulable  nitrogen  and  amino-acids,  and  they  be- 
lieve that  the  cause  of  the  acute  intoxication  is  the  rapid  cleavage  of 
the  serum  protease.  The  specific  element  they  believe  consists  in  the 
rapid  mobilization  of  the  ferments  and  the  colloidal  serum  changes 
which  bring  about  the  change  in  antiferment  titre.  Fascinating  as 
this  theory  is,  and  although  it  has  a  number  of  things  in  common  with 
our  own  ideas  concerning  the  colloidal  balance  in  the  plasma  which 
ordinarily  prevents  the  rapid  union  of  antigen  with  antibody,  we 
feel  that  recent  knowledge  concerning  the  essentially  intracellular 
nature  of  anaphylaxis  in  the  guinea  pig  prevents  its  adoption. 
Furthermore,  we  believe  that  the  specific  element  of  anaphylaxis  is 
insufficiently  explained  by  Jobling  and  Petersen' s  conclusions. 

Again  the  question  arises:  Are  these  serum  proteases  in  any 
way  to  be  identified  with  bacteriolytic  antibodies  or  with  alexin  or 
complement  ?  It  is  more  than  likely  that  no  such  relationship  exists. 
In  the  first  place,  the  seroproteases  are  non-specific,  and  it  has  been 
shown  that  in  the  course  of  bacteriolysis  no  increase  in  non-coagulable 
nitrogen  occurs.  Moreover,  Jobling  and  Petersen,19  to  whom  we 
owe  much  of  our  recent  knowledge  on  this  subject,  have  shown  that 
the  treatment  of  bacteria  with  complement  alone  or  with  complement 
together  with  human  serum  renders  them  more  resistant  to  proteoly- 

16  Jobling  and  Petersen.  Journal  of  the  A.  M.  A.,  1915,  Vol.  65,  p.  515. 

17  Pfeiffer  and  Jarisch.  Zeitschr.  f.  Imm.,  Vol.  16,  1912. 

18  Jobling  and  Petersen.  Journ.  of  Exp.  Med.,  1914,  Vol.  20. 

19  Jobling  and  Petersen.  Journ.  of  Exp.  Med.,  Vol.  20,  1914,  p.  321. 


SERUM  ENZYMES  531 

sis,  probably  owing  to  the  absorption  of  antiferments  from  the  blood 
serum.  The  identification  of  complement  with  lipase  has  been  sug- 
gested. The  idea  gains  likelihood  from  the  fact  that  both  complement 
and  lipase  are  destroyed  by  lipoid  solvents  and  that  many  of  the 
substances  upon  which  complement  acts,  such  as  red  blood  cells, 
contain  lipoid  constituents.  On  the  other  hand,  no  definite  knowl- 
edge is  available  in  this  regard,  and  it  has  been  so  far  impossible 
to  prove  even  the  enzymotic  nature  of  complement,  though  this,  at 
least,  seems  likely.  The  peculiar  resistance  exhibited  by  bacteria  to 
digestion  by  ferments  has  already  been  alluded  to. 

It  develops  therefore  that  the  antiferments  in  the  blood  are  ex- 
tremely important  factors  in  maintaining  the  balance  of  enzyme 
activity,  and  in  consequence  the  antienzymatic  properties  of  serum 
have  been  investigated  by  many  workers.  For,  depending  upon  their 
fluctuation,  cleavage  processes  are  permitted  or  prevented  from  tak- 
ing place  in  the  circulating  blood.  The  earliest  workers  concerned 
themselves  largely  with  the  phenomenon  that  blood  serum  would 
prevent  the  proteolytic  activity  of  leucocytes.  They  utilized  the  ob- 
servation therapeutically  by  injecting  serum  into  suppurating  ab- 
scesses for  the  purpose  of  preventing  tissue  destruction  by  leuco- 
protease  and  increasing  the  ability  of  the  body  to  limit  the  processes. 
Indeed  it  may  be  that  this  form  of  therapy  in  chosen  cases  may  still 
have  possibilities. 

Later  investigators  studied  antiferment  fluctuations  in  disease. 
In  septic  conditions,  Wiens  20  claims  to  have  noticed  a  diminution 
of  serum  antiferments.  Similar  observations  have  been  made  after 
shock  in  anaphylaxis  by  Pfeiffer  and  Jarisch.  In  patients  with  can- 
cer, Landois21  claimed  to  have  noticed  a  marked  increase  in  anti- 
ferments,  which  for  a  time  was  believed  to  be  sufficiently  regular  to 
have  diagnostic  significance.  Brenner 22  noticed  the  same  phe- 
nomenon in  severe  anemias.  In  cachexia  the  same  phenomenon  has 
been  noticed  and  fluctuations  in  this  constituent  have  been  noticed 
in  a  great  many  different  diseases,  such  as  diabetes,  pneumonia,  etc., 
without  our  being  at  present  in  possession  of  any  definite  or  corre- 
lated knowledge  of  the  principles  upon  which  these  fluctuations  de- 
pend. 

Earlier  observers  believed  that  the  anti-ferment  in  the  blood  was 
a  function  of  the  albumin  fraction  of  the  serum.  More  recent  work- 
ers have  attributed  it  to  lipoid  constituents.  Schwarz,23  Sugimoto,24 
and  Kirchheim  suggested  the  lipoid  idea  because  of  the  fact  that 

20  Wiens.     Munch,  med.  Woch.,  No.  53,  1907. 

21  Landois.     Berl.  klin.  Woch.,  No.  10,  1909. 

22  Brenner.     Deutsch.  med.  Woch.,  No.  9,  1909,  and  Mediz.  Klinik,  No.  28, 
1909. 

23  Schwarz.     Wien.  klin.  Woch.,  22,  1909. 

24  Sugimoto.     Arch.  f.  Ex^.  Path.  u.  Pharmakol,  Vol.  74,  1913. 


532  INFECTION  AND  RESISTANCE 

the  antiferment  properties  were  destroyed  by  lipoid  solvents.  They 
did  not  settle  the  question  satisfactorily  because  their  results  were  in- 
consistent, but  Jobling  and  Petersen25  found  that  antiproteolytic 
activity  of  blood  serum  was  almost  wholly  inhibited  when  sera  were 
extracted  for  several  days  with  chloroform  or  ether  at  room  tempera- 
ture or  with  chloroform  in  the  incubator  for  an  hour.  The  anti- 
ferment  could  be  recovered  from  the  ether  and  chloroform  solution, 
and  its  action  could  be  destroyed  by  incubation  with  iodin  or  po- 
tassium iodid.  They  believed  this  to  be  due  to  the  saturation  of 
the  free  carbon  atoms.  Furthermore,  the  same  authors  showed  that 
toxic  substances  analogous  to  anaphylatoxins  could  be  produced  in 
serum  when  incubated  with  chloroform.  The  lipoidal  nature  of  the 
antiferment s  is  a  likely  one  although  it  does  not  account  for  the  fact 
noticed  uniformly  by  all  workers  that  the  antitryptic  activity  can 
be  entirely  eliminated  by  heating  the  serum  to  70°  C.  for  a  half  hour. 
This  was  the  observation  which  at  first  gave  rise  to  the  idea  that  the 
antiferment  was  in  itself  an  enzyme.  For  this  phenomenon  of  heat 
lability  no  satisfactory  explanation  has  as  yet  been  advanced. 

The  recent  researches  of  Abderhalden  2G  upon  the  intravascular 
digestion  of  foreign  substances  introduced  into  animal  bodies  prom- 
ised to  have  considerable  bearing  upon  problems  of  immunity. 
Abderhalden,  whose  work  we  cite  chiefly  from  his  monograph,  "Die 
Schutzfermente  des  tierischen  Organismus,"  took  as  his  point  of  de- 
parture the  conception  that  the  animal  body  must  necessarily  preside 
over  a  mechanism  whereby  it  can  assimilate  foreign  substances  which 
obtain  entrance  unchanged  into  the  circulation.  Abderhalden  be- 
lieves that  this  process  depends  upon  the  mobilization  of  "protec- 
tive ferments/'  a  term  which  he  borrows  from  Heilner,27  and  sug- 
gests the  possibility  that  these  ferments  may  originate  in  the  leuko- 
cytes. 

Experimentally  Abderhalden  approaches  his  problem  by  deter- 
mining the  presence  of  specific  ferments  in  the  blood  of  animals  into 
which  various  foreign  substances  have  been  introduced  by  paths 
other  than  the  alimentary  canal.  For  this  purpose  he  has  developed 
a  number  of  methods,  the  most  important  of  which  are  his  optical 
method  and  his  dialysis  method.  The  optical  method  used  for  the 
determination  of  the  proteolytic  properties  of  the  serum  depends 
upon  the  fact  that  many  of  the  amino-acids  are  optically  active. 
Moreover,  most  of  these  substances  are  chemically  known  and  their 
optical  activity  determined,  so  that  it  is  possible  to  take  blood  serum 
which  is  to  be  examined  for  its  contents  of  particular  ferments,  mix 

25  Jobling  and  Petersen.     Journ.  Exp.  Med.,  1914,  Vol.  19,  p.  480. 

26  Abderhalden.     "Schiitzfermente  des  tierischen  Organismus,"   Springer, 
Berlin,  1912. 

27  Heilner.     Cited  from  Abderhalden  Zeitschr.  f.  Biol,  Vol.  50,  1907. 


SERUM  ENZYMES  533 

them  with  a  suitable  protein,  or  preferably  a  polypeptid,  and  de- 
termine with  a  polariscope  the  rotation  which  takes  place.  We  will 
not  go  into  the  technique  of  this  method  more  extensively  because 
we  have  no  personal  experience  with  it,  and  the  method  is  one  of 
such  delicacy  that  it  is  best  obtained  from  Abderhalden's  original 
publications  directly.28  His  dialysis  methods  depend  upon  placing 
the  blood  serum  and  fermentable  substance  into  dialyzing  bags,  sus- 
pending them  into  distilled  water,  and  determining  the  presence  of 
peptone,  amino-acids,  or  total  nitrogen  in  the  liquid  outside  of  the 
bag  after  definite  intervals  of  time. 

By  these  and  other  methods  Abderhalden  29  has  carried  out  tests 
with  a  large  number  of  different  substances.  Experimenting  first 
with  proteins,  he  injected  egg  albumen,  horse  serum,  silk  peptone, 
gelatin,  edestin,  casein,  etc.,  into  dogs  and  rabbits,  then,  several  days 
later,  bled  the  animals  and  mixed  0.5  c.  c.  of  the  serum  with  0.5  c.  c. 
of  a  solution  of  the  respective  substances  which  had  been  injected. 
He  found  in  such  cases  that  definite  proteolytic  action  was  exerted 
upon  the  injected  substances  by  the  active  serum  of  a  treated  animal, 
whereas,  in  the  case  of  most  of  the  substances  used,  the  normal  serum 
possessed  no  proteolytic  action  whatever.  These  results  were  con- 
sistently obtained  both  by  the  dialysis  and  by  the  optical  methods. 
It  should  be  especially  noted  that  the  ferments  studied  by  Abder- 
halden were  not  as  specific  as  are  the  antibodies  which  we  have  dis- 
cussed in  another  place.  For  Abderhalden  found  that  the  serum  of 
an  animal  treated  with  proteins  developed  enzymes  which  were  active, 
not  only  against  the  particular  protein  used  for  injection,  but  rather 
against  proteins  in  general.  They  were  specific  only  in  that,  when 
produced  with  proteins,  they  were  not  active  against  fats  or  carbo- 
hydrates. This  is  especially  important  in  connection  with  the  recent 
discussion  concerning  the  identity  of  Abderhalden's  protective  fer- 
ments and  the  specific  protein  antibodies. 

In  later  experiments  Abderhalden  claimed  to  show  further  that 
similar  ferments  could  be  induced  in  animals  by  treatment  with  car- 
bohydrates and  with  fats.  The  serum  of  normal  dogs  is  not  capable 
of  splitting  cane  sugar.  However,  the  blood  serum  or  plasma  of  a 
dog  that  has  been  treated  with  cane  sugar  develops  the  property  of 
inverting  the  cane  sugar  into  dextrose  and  fructose  within  fifteen 
minutes  after  injection.  This  could  easily  be  determined  both  by 
putting  together  the  serum  with  cane  sugar  and  determining  the 
increase  of  reducing  powers,  and  by  means  of  subjecting  such  active 
plasma  or  serum,  together  with  saccharose,  to  polariscopic  examina- 
tion. 

28  See  especially  Abderhalden,  Hoppe-Seyler,  Zeitschr.  f.  pTiysiol.  Cliemie, 
Vols.  60,  65,  and  66;  also  "Handbuch  der  Biochem.     Arbeitsmethoden,"  Vol. 
5,  p.  575,  1911. 

29  Abderhalden.     "Schiitzfermente,"  p.  49. 


534  INFECTION  AND  RESISTANCE 

The  earlier  experiments  with  fats  were  negative  because  the 
simple  method  of  titration  for  fatty  acids  proved  insufficient  as  an 
indicator  of  activity.  However,  Abderhalden  succeeded  in  determin- 
ing fat  splitting  properties  in  the  blood  of  treated  dogs  by  using  the 
method  of  Michaelis  and  Rona.30  The  presence  of  fats  largely  in- 
creases the  surface  tension  of  mixtures,  and  their  cleavage  in  such 
mixtures  consequently  leads  to  reduction  of  this  tension.  Utilizing 
this  principle,  Abderhalden  claims  to  have  determined  that  the  paren- 
teral  introduction  of  fats  into  dogs  is  followed  by  a  reactionary  in- 
crease of  lipases. 

The  general  significance  of  Abderhalden's  researches  is  this: 
When  any  foreign  substances,  protein,  carbohydrate,  or  fats,  gain 
entrance  to  the  circulation  of  an  animal,  the  animal  body  reacts  by 
the  mobilization  of  ferments  or  enzymes  specifically  capable  of  re- 
ducing these  substances  to  assimilable  form.  It  is  likely  that  these 
ferments  represent  a  mobilization  of  substances  normally  present  but 
not  concentrated  in  the  blood  stream  under  ordinary  conditions,  since 
they  appear  with  a  speed  out  of  all  proportion  to  that  obtaining  in 
the  case  of  the  antibodies  discussed  in  another  place.  In  one  case 
cited  by  him  a  dog  injected  on  November  25th,  29th,  and  December 
4th  showed  powerful  peptolytic  serum  properties  on  December  6th. 
Apparently  the  injection  of  homologous  proteins  into  animals  (i.  e., 
rabbit  serum  into  rabbits,  etc.)  does  not  incite  reaction. 

These  enzymes  seemed  to  differ  from  specific  antibodies  in  that 
they  did  not  react  solely  with  the  substance  injected,  but  also  with 
other  substances  belonging  to  the  same  chemical  group.  Other  dif- 
ferences from  antibodies  are  the  rapid  appearance  of  the  ferments 
after  treatment  and  their  rapid  disappearance  after  the  inciting 
stimulus  is  removed.  Thus  Abderhalden  reports  that  the  enzymes 
found  in  a  case  of  pregnancy  disappeared  within  eight  days  after 
abortion  or  child  birth. 

It  is  plain  that  these  researches  of  Abderhalden  offer  many  op- 
portunities for  diagnostic  utilization,  and  he  has  applied  them  to  the 
diagnosis  of  pregnancy.  In  this  condition  substances  from  the 
chorionic  villi  get  into  the  blood.  These,  according  to  Abderhalden, 
may  be  looked  upon  as  in  a  certain  sense  foreign  in  nature,  and  must 
be  chemically  disintegrated  by  the  body.  In  consequence  it  is  likely 
that  the  ferments  which  accomplish  this  would  appear  in  the  sera  of 
pregnant  individuals  and  could  be  determined  by  his  methods. 
"When  he  prepared  peptone  from  the  placental  substances  of  human 
beings  and  allowed  the  blood  plasma  of  normal  individuals  to  act 
upon  it,  observing  it  both  by  the  dialysis  and  the  optical  method,  no 
peptolytic  action  could  be  observed.  However,  when  the  plasma  of 
pregnant  women  was  used  proteolytic  action  was  determined.  In 
these  cases  the  ferment  seemed  to  be  specific  for  peptones  produced 

30  Michaelis  and  Rona.     Cited  from  Abderhalden,  loc.  cit. 


SERUM  ENZYMES  535 

from  placental  tissue  both  in  animals  and  human  beings,  but  did  not 
act  upon  casein,  gelatin,  or  other  proteins.  There  are  certain  techni- 
cal difficulties  connected  with  the  production  of  a  test  material  from 
the  placental  tissue  which  render  this  method  difficult.  For  their 
more  detailed  description  we  refer  the  reader  to  the  original  articles. 
The  Abderhalden  reaction  and  its  various  logical  ramifications 
promise  for  some  time  to  bring  an  entirely  new  and  important  factor 
into  the  field  of  immunology,  namely,  that  specific  ferment  mobiliza- 
tion in  definite  response  to  the  introduction  of  various  substances, 
and  yet  distinct  from  the  specific  antibody  heretofore  known.  The 
reaction  rapidly  became  the  subject  of  active  research  mostly  of  a 
clinical  nature,  and  carried  out  with  insufficient  critical  judgment. 
This  literature  is  extensive  and,  because  almost  entirely  refuted  at 
the  present  day,  is  hardly  worth  citing.  In  this  country  the  re- 
action has  been  carefully  studied  by  a  great  many  workers,  more 
especially  by  Bronfenbrenner,  and  Jobling,  Petersen  and  Eggstein. 
From  all  these  studies,  we  may  say  for  the  sake  of  brevity,  it  has 
become  clear  that  the  enzymes  involved  in  the  Abderhalden  reac- 
tion are  probably  not  specific  as  supposed  by  this  writer,  that  during 
the  reaction  the  tissues  employed  (that  is  placenta,  etc.),  act  by 
absorbing  antienzymes  from  the  serum.  In  consequence  the  fer- 
ments in  the  serum  act  upon  the  serum  protein  itself  which  there- 
fore becomes  the  substrate  of  the  cleavage  product  determined  as  a 
result  of  the  reaction.  This  robs  the  reaction  of  any  claim  to  the 
specific  and  diagnostic  value  attributed  to  it  by  Abderhalden  and  his 
collaborators,  and  as  a  final  result  of  these  extensive  and  intricate 
researches  we  have  merely  a  better  understanding  of  the  non-specific 
enzyme  activity  of  normal  serum. 

SOME  OF  THE  PHYSICAL  FACTORS  WHICH  ENTER,  INTO 
SERUM  REACTIONS 

In  a  number  of  discussions  in  preceding  chapters  we  have  em- 
phasized the  fact  that  serum  reactions  are  gradually  being  recognized 
as  having  more  in  common  with  so-called  colloidal  phenomena  than 
with  those  taking  place  during  the  union  of  crystalloids.  The  funda- 
mental physical  principles  underlying  colloidal  reactions  are  as  yet 
pretty  vague  and  the  serologist  especially  (ourselves  included),  when 
he  speaks  of  colloidal  reactions  is  often  groping  in  the  dark.  Never- 
theless, even  without  such  fundamental  understanding  we  can  recog- 
nize the  close  analogies  which  exist  between  the  various  serum  re- 
actions and  those  taking  place  between  substances  recognized  by 
their  attributes  as  colloids.  We  have  sufficiently  discussed  this  in 
the  chapters  on  agglutination  and  precipitation,  and  need  not  en- 
large upon  it  here  except  in  indicating  how  the  gradual  tendency  to 
pay  attention  to  the  physical  factors  involved  has  led  to  the  actual 


536  INFECTION  AND  RESISTANCE 

working  out  of  serum  experiments  in  which  the  reasoning  from  the 
beginning  and  the  technique  finally  developed  were  physical  rather 
than  chemical.  Perhaps  the  earliest  phases  of  such  work  were  those 
in  which  complement  fixation  was  shown  to  occur  when  complement 
was  brought  together,  with  non-specific  inorganic  and  organic 
colloids.  It  is  due  chiefly  to  the  work  of  Landsteiner  and  his 
associates  that  attention  was  called  to  such  phenomena.  Landsteiner 
with  Stankovic  showed  that  complement  may  be  fixed  by  silicic  acid, 
and  Seligmann  soon  afterwards  found  that  in  the  precipitation  of 
mastic  suspensions  complement  may  be  fixed.  Much  work  has  been 
done  by  others,  as  well  as  by  ourselves,  that  shows  that  a  variety  of 
colloidal  suspensions  of  known  chemical  constitution  fix  complement, 
probably  by  adsorption.  Perhaps  the  most  striking  instance  of  such 
complement  adsorption  is  that  occurring  in  the  Wassermann  re- 
action. Here,  as  we  know,  complement  is  fixed  by  a  combination  of 
syphilitic  serum  and  various  lipoidal  suspensions,  which  may  be  en- 
tirely non-specific  in  origin.  When  the  reaction  occurs,  as  Jacob- 
stahl  and,  later,  Bruck  have  shown,  a  precipitation  occurs  which 
can  be  seen  in  the  ultramicroscope.  Furthermore,  this  precipitation 
takes  place  more  rapidly  in  the  ice-chest  than  in  the  incubator,  a 
principle  upon  which  is  based  the  so-called  refrigerator  method  of 
performing  the  test,  and  which  strikingly  suggests  that  the  reaction 
is  an  adsorption  rather  than  a  true  chemical  union.  It  is  this  pre- 
cipitate which  fixes  the  complement ;  whether  or  not  this  is  due  to 
quantitatively  increased  globulin  or  to  purely  physical  change  in  the 
syphilitic  serum,  is  a  matter  which  we  cannot  discuss  at  present. 
Whatever  it  is,  it  is  unquestionable  that  the  availability  of  the  anti- 
gen for  the  Wassermann  reaction  depends  not  only  upon  its  lipoidal 
nature  but  also  on  its  state  of  dispersion.  Since  it  is  not  possible, 
as  we  have  found,  to  make  available  antigens  for  this  reaction  with 
non-lipoidal  substances,  like  mastic,  gelatin,  gum  arabic,  salicic  acid, 
albumins,  and  a  number  of  other  substances,  even  when  in  dispersion 
more  or  less  similar  to  that  of  the  Wassermann  antigen,  it  seems  that 
the  secret  of  the  Wassermann  antigen  must  lie  in  the  fact  that  sub- 
stances of  the  chemical  and  physical  constitution  of  lipoids  when 
brought  into  a  definite  state  of  dispersion  offer  surface  tension  con- 
ditions not  easily  obtained  with  colloids  of  another  nature.  It  is, 
therefore,  at  least  in  our  opinion  at  present,  the  physical  condition 
of  the  Wassermann  antigen  which  makes  it  available  for  the  test,  a 
physical  condition  which  is  secondarily  dependent  upon  the  chemical 
nature  of  the  dispersed  substance.  The  importance  of  the  state  of 
dispersion  and  therefore  the  surface  tension  properties  is  quite  ap- 
parent from  the  fact  noted  by  many  Wassermann  workers  that  a  con- 
siderable difference  in  the  fixing  power  of  the  antigen  may  be  ob- 
tained by  in  one  case  adding  the  salt  solution  to  the  alcoholic  ex- 
tract quickly,  and  in  another  case  adding  it  very  slowly,  the  two 


SERUM  ENZYMES  537 

separate  preparations  showing,  one  a  very  transparent,  the  other  a 
very  turbid,  condition. 

The  Bordet-Danysz  phenomenon  has  already  been  discussed,  but 
is  another  case  in  point.  It  consists  of  the  fact  that  if  to  a  definite 
amount  of  antitoxin  a  definite  quantity  of  toxin  is  added,  the  re- 
sult is  one  of  greater  toxicity  in  the  mixture  if  the  poison  is  added 
to  the  antitoxin  fractionally  than  when  the  entire  amount  is  added 
at  once.  An  interesting  difficulty  of  this  phenomenon  is  the  fact  that 
such  a  reaction  seems  to  force  upon  us  the  assumption  that  the  toxin- 
antitoxin  union  is  not  reversible,  whereas  later  work  on  immuniza- 
tion with  neutralized  mixtures  of  these  substances  seems  to  necessi- 
tate the  assumption  that  within  the  body  they  are  reversible. 

It  is  a  consideration  of  these  and  many  other  apparently  physi- 
cal factors  of  immune  reactions  which  has  led  experimenters  to  seek 
to  utilize  physical  changes  for  practical  serological  purposes.  One 
of  the  reactions  resulting  from  such  studies  is  that  known  as  Weich- 
hardt's  Epiphanin  reaction.  Weichhardt  noticed  that  when  two  solu- 
tions of  toxin  and  antitoxin  are  brought  in  contact  with  each  other  on 
exactly  horizontal  glass  plates,  diffusion  takes  place  between  them 
much  more  rapidly  than  in  controls  in  which  one  or  the  other,  antigen 
or  antibody,  was  lacking.  He  tested  this  at  first  on  the  horizontal 
glass  plates  by  adding  coloring  matter  to  the  two  solutions  and  ob- 
serving its  diffusion  currents.  Later  he  developed  a  method  in  which 
he  tried  to  determine  such  increased  diffusion  by  means  of  a  delicate 
chemical  balance.  His  method,  briefly  described  in  his  own  words, 
was  as  follows :  "To  the  two  arms  of  scales,  to  the  right  and  left,  a 
little  bell-shaped  jar  is  attached  into  which  dilutions  of  serum  are 
placed,  on  the  one  side  specific  immune  serum,  on  the  other  normal 
serum.  These  little  bell-jars  are  closed  at  the  bottom  with 
'Schleicher-SchulP  filter  paper  and  are  immersed  into  solutions  con- 
taining antigen  in  salt  solution.  Through  the  filter-paper  mem- 
brane diffusion  takes  place,  and  in  observations  carried  on  for  from 
several  hours  to  several  days  it  can  be  determined  that  the  scales 
sink  on  that  side  in  which  specific  immune  serum  had  been  placed 
into  the  little  bell-jar.  In  other  words,  the  antigen  solution  which 
is  contained  in  the  salt  solution  diffuses  more  rapidly  on  the  side  on 
which  the  specific  immune  serum  was  present.  In  consequence,  the 
weight  of  the  little  bell  increases  and  a  definite  reading  can  be  made." 
Similar  observations  have  been  made  by  other  observers,  Kraus  and 
Amiradzibi  doing  the  experiment  as  follows :  they  employed  a  U-tube 
in  the  connecting  horizontal  part  of  which  there  was  a  glass  stop- 
cock. Into  one  side  they  put  diphtheria  toxin  with  a  little  aqueous 
methylene  blue,  and  into  the  other  side  antitoxin.  For  each  experi- 
ment two  controls  were  set  up,  one  with  salt  solution  and  the  other 
with  normal  horse  serum,  and  they  noticed  that  after  a  definite  period 
of  time  after  the  stopcock  connecting  the  two  had  been  opened, 


538  INFECTION  AND  RESISTANCE 

methylene  blue  could  be  observed  to  diffuse  across  in  the  tube  in 
which  antigen  and  antibody  had  been  present  and  not  in  the  con- 
trols. Many  modifications  of  technique  to  demonstrate  this  principle 
have  been  made  by  Weichhardt  himself,  and  since  the  reaction  is  not 
likely  to  find  much  immediate  diagnostic  application  because  of  its 
great  delicacy,  we  need  not  describe  them  at  length  but  refer  the 
reader  to  Friedemann's  excellent  description  in  the  Kolle  and  Wasser- 
mann,  second  edition,  Vol.  III. 

Similar  in  principle  but  by  no  means  identical  is  the  so-called 
Meiostagmin  reaction,  chiefly  developed  by  Ascoli. 

THE  MEIOSTAGMIN  REACTION 

Ascoli  and  Izar  31  have  attempted  to  work  out  a  diagnostic  reac- 
tion which  depends  upon  an  alteration  of  surface  tension  of  a  fluid 
when  an  antigen  unites  with  its  specific  antibody.  Ascoli  in  his  first 
experiments  worked  with  typhoid  bacillus  extracts  and  the  sera  of 
typhoid  patients,  and  found  that  when  the  two  suspensions  were 
mixed  a  reduction  of  surface  tension  resulted  after  time  for  union 
between  the  two  had  been  allowed. 

They  determined  the  reduction  of  surface  tension  by  Traube's  32 
method  by  the  use  of  apparatus  spoken  of  as  the  "stalagmometer." 
The  principle  of  this  method  depends  upon  the  fact  that  as  surface 
tension  is  reduced  the  number  of  drops  to  a  given  quantity  of  fluid 
is  increased. 

Diluted  serum  of  patients  was  mixed  with  diluted  antigen,  and 
the  number  of  drops  contained  in  one  cubic  centimeter  of  the  mix- 
ture was  immediately  determined  and  again  measured  after  the  mix- 
ture had  remained  for  two  hours  in  the  incubator  at  37°  C.  An 
example  of  one  of  Ascoli' s  early  measurements  is  given  in  the  protocol 
cited  below. 

Of  course  a  certain  amount  of  reduction  of  surface  tension  results 
when  various  antigens  are  brought  together  with  normal  sera,  but 
this  can  be  easily  controlled  by  suitable  dilution,  and  must  be  care- 
fully taken  into  consideration  in  each  individual  case.  Ascoli  and 
Izar  have  applied  this  method  to  the  diagnosis  of  tuberculosis,  ty- 
phoid, and  various  other  diseases,  and  have  reported  what  seemed  to 
them  reliable  results.  So  far  experience  with  the  meiostagmin  reac- 
tion has  not  been  very  extensive ;  not  all  observers  have  been  able  to 
obtain  results  as  apparently  reliable  as  those  of  Ascoli  and  his  col- 
laborators. It  is  not  possible  therefore  to  express  a  final  opinion 
regarding  this  method  of  investigation ;  it  contains,  however,  an  in- 
teresting principle  which  with  more  exact  methods  of  measurement 
may  well  become  very  important  in  serum  diagnosis. 

31  Aseoli  and  Izar.    Munch,  med.  Woch.,  Nos.  2,  7,  18,  22,  41,  1910. 

32  Traube.     Pfliigcr's  Archiv,  Vol.  123,  419. 


SERUM  ENZYMES 

1  c.  c.  of  serum  of  typhoid  patient  diluted  to  1-10. 


Number 

of  drops 

Immedi- 
ately 

After 
2  hours  in 
incubator 

'  1  0/00 

57.8 
58.1 

59.7 
59.9 

1  c.  c.  alcoholic  typhoid  extract  diluted  to  —  , 

1  0/000 

57.5 
57.5 

59.4 
59.6 

1  0/000 

57.0 
57.0 

59.3 
59.2 

"  1  0/00 

58.1 
57.7 

59.7 
59.6 

1  c.  c.  alcoholic  typhoid  extract  diluted  to  . 

1  0/000 

57.4 
57.6 

59.4 
59.2 

1  0/000 

57.0 
56.9 

59.2 
59.4 

"  1  0/00 

56.5 
56.5 

58.0 
57.8 

1  c.  c.  alcoholic  precipitate  taken  up  hi  distilled  ^ 

1  0/000 

56.5 
56.6 

57.5 
57.4 

1  0/0000 

56.7 
56.5 

57.4 
57.5 

1  c.  c.  in  1  0/00  alcohol  in  1  c.  c.  0.85  per  cent. 
NaCl  solution  

56.6 

57.5 

56.7 

57.6 

COLLOIDAL  GOLD  REACTION 

An  important  reaction,  depending  entirely  upon  the  principle  of 
colloidal  precipitation,  is  the  so-called  colloidal  gold  reaction  exten- 
sively used  for  the  diagnosis  of  syphilis  of  the  central  nervous  sys- 
tem. In  1901,  Zsigmondi,  knowing  that  protein  substances  will  pro- 
tect various  colloids  from  precipitation  by  electrolytes,  attempted  to 
work  out  a  method  by  means  of  which  he  could  quantitatively  esti- 
mate protein  in  solution  by  its  protective  effect.  He  worked  with 
colloidal  metals,  using  especially  colloidal  gold.  He  determined  what 
he  called  the  "Goldzahl"  for  various  proteins,  by  which  he  meant  the 
amount  in  weight  of  a  protein  which  was  just  enough  to  protect  from 
precipitation  10  c.c.  of  colloidal  gold,  of  a  percentage  of  0.0053,  in 
the  presence  of  1  c.c.  of  10  per  cent.  NaCl.  It,  of  course,  had  long 


540  INFECTION  AND  RESISTANCE 

been  known  that  in  various  pathological  conditions  of  the  central 
nervous  system  the  protein  contents  of  the  fluid  varied.  Thus,  the  re- 
actions of  Nonne,  of  Noguchi,  etc.,  were  all  aimed  at  revealing  an  ab- 
normal globulin  content  of  the  spinal  fluid.  Lange,  attempting  to  ap- 
ply Zsigmondi's  method  to  the  quantitative  detection  of  proteins  in  the 
spinal  fluid,  observed  a  curious  reaction,  quite  the  opposite  of  what 
he  had  set  out  to  find.  He  observed  that  spinal  fluid,  especially  of 
syphilitic  cases,  in  which  there  was  protein  material  beyond  the  nor- 
mal, precipitated  rather  than  protected  colloidal  gold.  He  also  ob- 
served that  the  dilution  at  which  such  precipitation  took  place  was 
more  or  less  constant  in  syphilitic  cases  and  might  therefore  be 
utilized  for  the  diagnosis  of  such  diseases  as  paresis,  locomotor  ataxia, 
etc.,  etc.  The  reaction  was  taken  up  and  carefully  studied  by  a 
large  number  of  observers,  among  whom  were  Zaloziecki,  Jaeger  and 
Goldstein,  Flesch,  Kafka,  Lee  and  Hinton,  Miller  and  Levy,  and 
many  others.  One  of  the  most  practical  discussions  and  clear  descrip- 
tions available  for  American  and  English  readers  is  that  of  Miller, 
Brush,  Hammers,  and  Felton.  The  chief  difficulty  of  the  test  con- 
sists in  the  preparation  of  a  proper  colloidal  gold  solution. 

Lange  adopted  the  method  used  by  Zsigmondi  with  slight  modi- 
fications. He  added  10  c.c.  of  a  1  per  cent,  solution  of  gold  colloid  and 
10  c.c.  of  a  2  per  cent,  solution  of  potassium  carbonate  to  a  liter  of 
carefully  and  freshly  distilled  water.  The  solution  was  rapidly 
warmed  almost  to  a  boil  and  just  before  boiling,  while  actively  stir- 
ring, 10  c.c.  of  1  per  cent,  formalin  solution  is  added.  The  solution  has 
remained  colorless  up  to  this  point  but  upon  this  addition  should  al- 
most instantaneously  become  a  deep,  transparent  red.  There  should 
be  no  iridescent  or  smoky  "Schimmer."  The  utmost  care  in  getting 
pure  distilled  water  must  be  exercised  and  Jena  glass  must  be  used 
throughout. 

Other  methods  are  those  of  Eicke,  who  adds  dextrose,  and  many 
workers,  such  as  Flesch,  Kaffge,  Eskuchen,  and  others,  have  found 
great  difficulty  in  making  up  the  solution  by  Lange's  technique  or  by 
any  of  the  other  methods  described.  Miller  and  Levy  state  that 
with  Lange's  method  good  solutions  can  always  be  obtained  if  the 
water  is  absolutely  pure  and  the  glassware  is  satisfactorily  cleaned ; 
later,  Miller,  with  the  collaborators  mentioned  above,  found  that  this 
was  not  universally  true.  Also,  these  writers,  as  well  as  Eskuchen, 
found  that  occasionally  solutions  which  appeared  all  right  did  not 
react,  and  special  studies  followed  dealing  with  the  technique  of  the 
preparation  of  the  gold  sol.  Miller  and  his  collaborators  find  that 
the  reduction  method  of  producing  this  solution  is  the  best  one,  and 
proceed  as  follows.  They  use  the  following  reagents  :33 

33  The  descriptions  are  taken  direct  from  the  thorough  papers  of  Millr- 
and  Levy. 


SERUM  ENZYMES  541 

1.  The  Gold  Solution: 

AuCl3 — Merck's   yellow   crystals   hermetically   sealed  in   brown 

glass  ampules 1  gram 

Water  triply  distilled,  up  to 100  c.  c. 

This  stock  solution  is  kept  well  stoppered  in  dark  glass  bottles 
away  from  any  bright  light. 

2.  The  Alkaline  Solution: 

Merck's  Blue  Label  Potassium  Carbonate  (desiccated) 2  grams 

Water  triply  distilled,  up  to 100  c.  c. 

3.  The  Reducing  Agents : 

a.  Formaldehyde,   Merck's  40  per  cent  stock  solution,  highest 

purity    1  c.  c. 

Water  triply  distilled,  up  to 40  c.  c. 

b.  Oxalic  acid,  Merck's  Blue  Label,  Crystals 1  gram 

Water  triply  distilled,  up  to 100  c.  c. 

Solutions  No.  2  and  No.  3  must  be  made  up  immediately  prior 
to  use. 

4.  Bichromate  cleaner  for  glassware: 

Potassium   Bichromate,   powdered 200  grams 

Water,  distilled,  up  to 1500  c.  c. 

Sulphuric  Acid  cone 500  c.  c. 

The  potassium  bichromate  should  be  well  dissolved  before  the 
sulphuric  acid  is  added.  If  this  solution  is  reserved  for  cleaning 
glassware  only,  it  can  be  used  repeatedly. 

Great  attention  must  be  paid  to  the  cleaning  of  glassware.  They 
boil  their  Jena  beakers  in  Ivory  soap  solution,  then  brush  under  hot 
tap  water.  After  being  rinsed  for  five  minutes,  hot  bichromate  and 
sulphuric  acid  is  added  for  half  an  hour.  The  beaker  is  then 
emptied  and  washed  in  running  water  for  five  minutes,  rinsed  with 
distilled  water  and  then  in  triply  distilled  water.  Similar  methods 
are  used  with  other  glassware.  The  chief  errors  are  insufficient 
brushing,  failure  to  get  all  bichromate  out,  and  allowing  beakers  to 
dry  in  air  before  using. 

In  obtaining  the  distilled  water,  they  first  take  ordinary  distilled 
water  which  they  then  re-distil  from  Jena  flasks.  They  wash  out 
the  collecting  flasks  with  the  first  200  c.  c.  of  the  second  distillate, 
then  collect  and  re-distil  in  the  same  way,  again  rinsing  with  the 
first  100  c.  c.  of  the  second  distillate.  After  all  these  steps  have 
been  performed  the  rest  is  more  or  less  simple.  A  beaker  rinsed  out 
with  triply  distilled  water  is  filled  to  the  liter  mark  and  the  tem- 
perature raised  to  50°  C.  gradually,  then  the  gas  is  turned  on  full 
and  when  the  temperature  has  reached  60°  C.  10  c.c.  of  the  1  per  cent. 


542  INFECTION  AND  RESISTANCE 

gold  solution  and  7  c.c.  of  the  2  per  cent,  potassium  carbonate  solu- 
tion are  added.  The  solution  should  remain  perfectly  clear.  At 
80°  C.  while  stirring  with  a  clean  thermometer,  10  drops  of  oxalic 
acid  are  slowly  added.  The  solution  may  now  turn  a  delicate  bluish- 
pink,  often  due  to  an  excess  of  alkali.  Otherwise  the  solution  re- 
mains colorless  until  90°  C.  has  been  reached.  When  90°  C.  has  been 
reached,  the  gas  is  turned  out  and,  while  stirring,  5  c.  c.  of  1  per  cent, 
formaldehyde  is  added,  drop  by  drop.  If  a  pink  color  makes  its 
appearance  before  all  the  reducing  agent  has  been  added,  Miller 
advises  to  stop  at  once.  He  also  states  that  the  best  solutions  are 
those  in  which  the  color  change  is  slow.  For  further  technical  points, 
the  reader  is  referred  to  the  paper  of  Miller  and  his  collaborator  in 
the  Johns  Hopkins  Hospital  Bulletin,  1915,  xxvi,  391,  from  which 
this  description  is  almost  bodily  taken. 

The  test  is  done  as  follows:  into  a  clean  test  tube  1.8  c.  c.  of 
fresh  sterile  0.4  per  cent.  NaCl  solution  is  placed.  Into  ten  further 
tubes  1.8  c.  c.  of  the  same  salt  solution  is  placed.  To  the  first  tube  is 
added  0.2  c.  c.  of  spinal  fluid  to  be  tested  and  well  mixed.  1  c.  c,  of 
this  1 :10  solution  is  added  to  the  second  tube,  mixed,  and  of  this  1  c.  c. 
is  added  to  the  third  tube,  in  consequence  of  which  dilutions  are  ob- 
tained running  from  1:10,  1:20,  1:50.  Now  to  each  of  these  is 
added  1  c.  c.  of  the  colloidal  gold  solution.  Changes  begin  to  take 
place  in  five  minutes,  which  usually  reach  their  completion  in  about 
a  half  hour.  The  spinal  fluids  must  be  free  from  blood  and  clear 
from  bacterial  contamination. 

Normal  spinal  fluids  produce  usually  no  reaction.  The  so-called 
luetic  curve  is  usually  one  in  which  the  greatest  precipitation  occurs 
in  the  tubes  ending  in  1 :40  to  1 :160.  In  suppurative  lesions,  etc., 
the  strongest  precipitation  occurs  in  higher  dilutions,  say,  from  320 
to  280,  which  Lange  has  called  "verschiebung  nach  oben."  A  so- 
called  paretic  curve  means  complete  precipitation  in,  say,  from  110 
to  160,  with  a  gradual  fading  out  toward  the  higher  dilutions.  A 
complete  precipitation  means  a  colorless  tube;  partial  precipitation 
shades  from  pale  blue  to  the  complete  red  of  the  unaffected  colloidal 
gold.  Fluids  from  early  stages  of  syphilis  without  nervous  system 
involvement  react  usually  like  normal  fluids. 

That  the  reaction  has  unquestionably  found  a  permanent  useful- 
ness cannot  be  doubted.  It  seems  to  be  dependent  entirely  upon 
the  technique  of  making  the  colloidal  gold  solution. 


CHAPTER    XXI 

COLLOIDS 

BY  STEWART  W.  YOUNG 
Professor  of  Physical  Chemistry,  Stanford  University,  CaL 

INTKODUCTOEY 

IN  attempting  to  give  in  the  brief  space  of  a  single  chapter  any 
adequate  account  of  the  present  state  of  our  knowledge  in  so  vast  a 
field  as  that  of  colloid  chemistry  and  physics  one  is  confronted  with 
a  rather  difficult  problem.  In  the  present  outline  the  attempt  will 
be  made  to  get  at  some  notion  of  the  matter  by  a  presentation  first 
of  the  more  important  generalizations  which  have  been  drawn,  this 
to  be  followed  in  each  case  by  sufficient  experimental  evidence  to 
serve  as  illustration,  together  in  some  cases  perhaps  with  certain 
evidence  which  may  seem  to  contradict  in  some  degree  such  cur- 
rent conceptions.  The  reason  for  this  particular  method  of  pres- 
entation lies  in  the  fact  that  new  material  is  so  rapidly  accumu- 
lating, much  of  which  seems  more  or  less  at  variance  with 
present  accepted  theories  that  it  seems  more  than  possible  that 
some  of  these  fundamental  generalizations  may  soon  undergo  ma- 
terial modification  if  not  reform.  It  would,  therefore,  seem  ill- 
advised,  in  presenting  a  brief  resume  to  readers  who  are  not 
physicists  or  chemists,  merely  to  present  the  present  theories  as 
they  are  used  to-day  by  workers  in  the  field,  and  to  sound  a  note 
of  warning  that  many,  if  not  all  of  them,  are  not  so  securely 
supported  by  broad  evidence  as  to  allow  of  very  concrete  prediction 
being  based  upon  them. 

Definition. — .The  fundamental  distinction  between  the  crystalloid 
and  colloid  states  of  matter  was  first  drawn  by  Thomas  Graham  1  as  a 
result  of  his  investigations  into  the  phenomenon  of  dialysis.  He 
noted  that  in  general  those  substances  which  when  in  solution  did  not 
pass  through  the  dialyzing  membrane,  or  did  so  only  very  slowly,  also 
were  characterized  by  the  fact  that  when  they  separated  from  solu- 
tion, either  by  precipitation  or  by  evaporation,  they  did  so  in  the  non- 

1  Graham.    Phil.  Trans.,  1861,  183. 

543 


544  INFECTION    AND    RESISTANCE 

crystalline  or  amorphous  form.  This  class  of  bodies  he  named  col- 
loids, since  glue  (Greek  KoXXa  meaning  glue)  presented  a  typical  case. 
Colloid  substances  may  appear  in  highly  dispersed  states,  such  as 
dilute  glue,  arsenic  sulphid  suspensions,  oil  or  rosin  emulsions,  milk 
(casein  in  highly  dispersed  condition),  and  the  like,  in  which  case 
they  are  spoken  of  as  sols.  They  may  also  occur  in  the  undispersed 
or  only  slightly  dispersed  state,  as  the  amorphous  precipitated  sul- 
phids  of  the  heavy  metals,  precipitated  casein,  or  dry  glue.  In  this 
state  they  are  spoken  of  as  gels. 

When  a  colloid  substance  has  once  been  converted  from  the  sol 
or  dispersed  state  into  the  gel  or  undispersed  state,  its  properties 
may  differ  greatly  in  different  cases.  Thus  if  a  dispersed  soap  (soap 
solution,  or  more  correctly  soap  sol)  be  coagulated  by  the  addition 
of  common  salt,  the  coagulum  or  soap  gel  may  be  removed  from  the 
salt  solution,  and  if  again  placed  in  pure  water  it  will  redisperse 
and  again  assume  the  sol  condition.  Such  a  colloid  is  spoken  of  as  a 
reversible  colloid.  If,  however,  an  arsenious  sulphid  suspension  be 
put  through  precisely  the  same  course  of  treatment  it  will,  in  the 
last  stage  of  the  treatment,  refuse  to  redisperse,  and  is  therefore 
spoken  of  as  an  irreversible  colloid.  Some  authorities  prefer  to 
speak  of  these  two  classes  of  colloids  as  emulsion  and  suspension  col- 
loids, respectively,2  since  in  general  those  colloids  which  are  reversi- 
ble tend  to  separate  out  in  soft  masses,  and  in  general  to  gelatinize 
rather  than  to  flocculate,  while  the  irreversible  colloids  rather  tend 
to  truly  flocculate  and  form  very  compact  and  frequently  more  or 
less  granular  coagula.  Since,  however,  we  seem  more  likely  at  the 
present  time  to  suffer  more  from  an  excess  of  classification  than  from 
a  lack  of  it,. the  attempt  will  be  made  to  get  along  in  this  discussion 
with  the  earlier  nomenclature.  It  may,  indeed,  be  added  that 
it  is  highly  probable  that  the  distinction  between  reversible  and  irre- 
versible colloids  is  only  one  of  degree.  For  example,  many  of  the 
metallic  sulphids  which  are  typically  irreversible  may  be  made 
to  some  extent  reversible  by  means  of  thorough  washing  and  re- 
treatment  with  hydrogen  sulphid  which  had  been  originally  used 
in  their  preparation.  It  is  probable  that  certain  colloids  are 
apparently  irreversible  only  because  we  do  not  truly  reverse 
the  conditions. 

Heretofore  in  this  discussion  the  term  "colloid"  substance  has 
been  used  as  if  to  imply  that  certain  chemical  individuals  were 
characteristically  colloid,  while  others  were  not.  It  was  much  in  this 
sense  that  Graham  used  the  term.  Investigations  since  his  time  have 
shown  this  to  be  a  misconception,  and  it  is  now  apparent  that  any 
and  all  substances  may  be  either  colloid  or  crystalloid,  the  form  they 
assume  depending  upon  treatment.  Thus  albumin  may  be  crystal- 
lized and  common  salt  may  be  obtained  in  the  state  of  a  colloid  solu- 

2  V.  Weimarn.     Ztschr.  Chem.  Ind.  Koll.,  1908,  3,  26. 


COLLOIDS  •  545 

tion  or  sol.3  Albumin,  gelatin,  and  agar  may  be  obtained  crystalline 
by  proper  regulation  of  temperature  and  the  use  of  proper  solvents, 
as  solutions  of  ammonium  sulphate  for  albumin,  and  alcohol-water 
mixtures  of  varying  strengths  for  the  two  latter  substances.  Sodium 
chlorid  has  been  obtained  in  the  colloidal  condition  by  precipitating 
it  in  a  solution  of  sodium  sulphocyanate  by  hydrochloric  acid,  each 
of  the  reacting  substances  being  dissolved  in  a  mixture  of  amyl  alco- 
hol and  ethyl  acetate. 

There  is  much  evidence  that  leads  to  the  belief  that  all  colloid 
systems  are  unstable.  Van  Bemmelen  characterized  them  as  systems 
which  never  reached  a  state  of  rest,  that  is,  were  never  in  equili- 
brium. The  conditions  which  determine  the  appearance  of  a  body 
in  the  colloid  or  crystalline  form  lead  to  the  suspicion  that  bodies 
always  separate  from  solution  in  the  amorphous  or  colloidal  condi- 
tion and  that  all  crystallization  is  a  secondary  phenomenon.  The 
conditions  that  are  favorable  for  the  transformation  of  a  colloid  into 
a  crystalline  form  are  a  considerable  solubility  and  a  considerable 
rate  of  crystallization.  Where  either  or  both  of  these  is  at 
a  minimum  the  conditions  are  favorable  for  relative  permanence 
in  the  colloid  condition.  It  is  upon  the  basis  of  this  prin- 
ciple that  von  Weimarn  succeeded  in  obtaining  relatively  stable 
colloidal  solutions  of  common  salt  and  many  other  easily  crystal- 
lizable  salts.  Furthermore  Doelter  4  has  succeeded  in  converting 
many  wrell-known  amorphous  precipitates  into  crystalline  bodies 
by  means  of  stirring,  pressure,  impact,  and  high  temperature. 
Among  the  substances  thus  transformed  are  aluminium,  chromium, 
and  iron  hydroxids,  and  the  sulphids  of  arsenic,  antimony, 
and  zinc. 

With  this  much  by  way  of  introduction,  we  may  now  proceed  to 
a  closer  consideration  of  some  of  the  better  recognized  properties  of 
colloid  sols  and  gels.  For  convenience  we  shall  first  take  up  the 
discussion  of  these  systems  from  the  more  definitely  physical  point 
of  view,  and  later  take  up  those  properties  which  seem  more  defi- 
nitely chemical. 

Physical  Properties  of  Colloids.— 1.  FORM  AND  SIZE. — Current 
opinion  seems  to  be  leading  rapidly  to  the  general  acceptance  of  the 
hypothesis  that  in  liquid  systems  of  two  or  more  components  we  have 
to  do  with  a  continuous  series  of  conditions  ranging  from  coarse  sus- 
pensions through  suspensions  of  increasing  fineness  (increasing  de- 
grees of  dispersion)  to  finally  the  molecular  and  ionic  states  of  solu- 
tion. The  opinion  is  also  growing  that,  although  for  certain  practical 
purposes  the  classification  of  all  such  systems  in  one  way  or  another, 
as  in  terms  of  the  various  degrees  of  dispersion,  may  be  useful,  the 

8V.  Weimarn.  Ibid.,  1910,  7,  92,  and  "Grundzii^e  der  dispersoid 
Chemie,"  107-108. 

4Ztschr.  Chem.  Ind.  KolL,  1910,  7,  86. 


546  INFECTION    AND    RESISTANCE 

excessive  use  of  such  classifications  is  likely  to  narrow  rather  than 
broaden  our  conception  of  the  whole  subject  matter  of  the  field.  It 
would  seem  that  the  most  stimulating  point  of  view  is  to  be  reached 
from  the  acceptance  of  the  suggestion  of  Wolfgang  Ostwald,  that  the 
chief  problem  of  colloid  chemistry  at  the  present  time  lies  in  deter- 
mining the  influences  of  the  degree  of  dispersion  upon  the  physical 
and  chemical  properties  of  all  liquid  solutions,  mixtures,  suspensions, 
or  what-not.  If  this  point  of  view  be  taken  it  follows  that  the  form 
and  size  of  the  particles  in  a  disperse  system  are  a  matter  of  the  first 
importance. 

If  the  degree  of  dispersion  in  a  given  system  be  not  too  great  the 
form  of  the  particles  may  be  readily  observed  under  the  microscope. 
Such  evidence  shows  that  the  spherical  form  predominates  enor- 
mously over  all  others,  although  under  carefully  controlled  conditions 
ovoid  forms  may  appear,  as  in  the  case  of  gelatin  and  agar.  These 
ovoid  forms  are  taken  by  von  Weimarn  (loc.  cit.)  and  others,  as  evi- 
dence of  directive  forces,  and  hence  of  incipient  crystallization.  If 
the  system  be  treated  in  such  a  way  as  to  decrease  the  dispersion,  as, 
for  example,  if  a  reagent  be  added  which  tends  to  flocculate  the  col- 
loid, but  not  in  sufficient  quantity  to  produce  actual  precipitation, 
the  decrease  in  dispersion  may  take  place  in  two  quite  different 
ways :  first,  the  size  of  the  particles  may  increase,  as  in  the  case  of 
oil  emulsions;  second,  the  particles  do  not  coalesce  but  become  at- 
tached together  in  chains  and  groups  which,  in  many  cases,  resemble 
bunches  of  grapes.  This  sort  of  aggregation  may  go  so  far  as  to  pro- 
duce web-like  structures.  The  jellying  of  gelatin  has  been  shown  to 
be  due  to  the  development  of  such  web  structures.  Glue  shows  much 
less  tendency  in  this  direction,  and  if  some  acetic  acid  be  added,  as 
in  the  preparation  of  commercial  liquid  glues,  this  web  formation  is 
almost  entirely  absent,  and  the  adhesive  qualities  are  at  the  same 
time  greatly  improved.  It  seems  quite  certain  that  both  of  the  above 
modes  of  aggregation  are  possible  in  one  and  the  same  system  at 
different  stages  in  its  condensation.  Thus  highly  dispersed  copper 
sulphid  becomes  aggregated  in  its  first  stages  of  condensation  by  an 
actual  increase  in  the  size  of  the  spherical  particles.  After  these 
reach  a  certain  fairly  definite  size  further  aggregation  takes  place 
by  the  grouping  together  of  these  spheres.  It  is  generally  recognized 
that  all  grouped  and  webbed  structures  are  secondary. 

A  large  number  of  very  important  investigations  have  been  di- 
rected toward  the  determination  of  the  size  of  disperse  particles 
throughout  the  greatest  variation  in  dimensions.  The  fact  first 
noted  by  Graham,  that  substances  in  colloidal  solution  show  a  very 
small,  and  frequently  almost  negligible,  rate  of  dialysis,  points  di- 
rectly to  the  supposition  that  the  particles  in  such  solutions  are  in  a 
far  less  dispersed  state  than  in  solutions  of  crystalloid  substances. 
The  rate  of  dialysis  is  directly  determined  by  the  rate  of  diffusion. 


COLLOIDS  547 

which,  in  turn,  is  inversely  proportional  to  the  square  roots  of  the 
masses  of  the  diffusing  bodies. 

^Measurements  of  osmotic  pressure  in  solutions  also  give  an  accu- 
rate measure  of  the  relative  masses  of  dispersed  systems  where  such 
measurements  can  be  successfully  carried  out,  and  a  great  deal  of 
work  has  been  devoted  to  attempts  to  measure  the  osmotic  pressure 
of  colloidal  solutions.  Great  difficulties  both  of  experimentation  and 
of  interpretation  are  encountered  in  this  field.  As  will  soon  be  seen 
a  colloid  particle  stands  in  a  very  complex  relationship  to  its  sur- 
rounding liquid,  and  furthermore  it  is  a  matter  of  extreme  difficulty 
to  obtain  a  colloid  solution  free  from  electrolytes,  which  themselves 
may  create  osmotic  pressure  or  otherwise  affect  the  measurements. 
About  the  only  conclusion  which  it  is  safe  to  draw  at  the  present  time 
is  that  if  colloid  solutions  show  osmotic  pressure  at  all  the  value  of 
it  is  very  small  compared  to  that  shown  by  crystalloidal  solutions  of 
substances  of  more  or  less  like  formula  weights.  This  leads  to  the 
conclusion  that  the  particles  in  a  colloid  solution  are  in  a  state  of 
dispersion  far  less  than  that  found  in  a  typical  crystalloidal  solution. 
For  a  most  excellent  resume  of  the  present  state  of  our  knowledge  in 
this  field  the  reader  is  referred  to  a  recent  book  by  Dr.  L.  Casuto, 
of  Pisa,  entitled  "Der  Kolloide  Zustand  der  Materie"  (Steinkopf, 
1913). 

When  the  size  of  the  disperse  particles  is  sufficiently  great  they 
may,  of  course,  be  measured  under  the  microscope,  and  with  the 
advent  of  the  ultramicroscope  the  limits  of  visibility  of  small  bodies 
has  been  very  notably  extended.  The  ultramicroscope  is  known  in 
several  forms,  the  first  having  been  devised  by  Siedentopf  and  Zsig- 
mondy.  All  depend  upon  the  production  of  powerful  rays  of  light 
in  directions  parallel  to  the  surface  of  the  microscope  slide.  In  such 
a  field  there  will  be  no  luminosity,  provided  the  field  is  optically 
empty,  that  is,  contains  no  particles  of  sufficient  size  to  produce  a 
dispersion  of  light.  If,  on  the  other  hand,  such  particles  are  present, 
the  effect  observed  will  be  an  illumination  whose  character  will  de- 
pend upon  the  size  of  the  particles.  If  the  particles  are  of  sufficient 
size  the  illumination  will  show  them  individually  as  bright  points 
of  dispersion,  even  though  the  particles  are  too  small  to  be  observed 
of  themselves,  just  as  the  stars  are  visible  from  the  light  which  they 
disperse,  but  cannot  of  themselves  be  seen.'  If  the  particles  are  so 
small  that  they  are  no  longer  able  to  disperse  sufficient  light  to  make 
each  particle  appear  as  a  bright  point,  there  will,  nevertheless,  pro- 
vided the  particles  are  present  in  sufficient  numbers,  be  produced  a 
diffuse  luminosity  throughout  the  field.  These  phenomena  are 
wholly  analogous  to  those  observed  when  a  beam  of  light  is  passed 
through  a  dark  room  in  the  atmosphere  of  which  fine  dust  particles 
are  found.  The  path  of  the  whole  beam  is  made  apparently  uni- 
formly luminous  by  the  smaller  particles,  while  occasionally  there 


548  INFECTION    AND    RESISTANCE 

appear  points  of  bright  illumination,  due  to  the  presence  of  larger 
particles.  This  is  known  as  the  Tyndall  effect.  The  light  which  has 
passed  through  such  fields  is  found  to  have  become  polarized. 

It  is  evident  that,  in  a  solution  whose  particles  are  sufficiently 
large  to  become  individually  visible  as  points  of  light  under  the  ultra- 
microscope,  it  immediately  becomes  possible  to  determine  the  size  of 
the  particles  on  the  assumption  that  these  are  all  of  the  same  size. 
The  procedure  consists  in  determining  the  following  quantities:  (1) 
the  total  number  of  particles  in  a  given  volume  by  the  usual  blood- 
count  method;  (2)  the  weight  of  the  dispersed  substance  in  a  given 
volume  by  a  chemical  or  other  analysis ;  ( 3 )  the  density  of  the  dis- 
persed substance,  which  is  usually  taken  as  equal  to  that  in  the  undis- 
persed  state.  This  undoubtedly  introduces  an  error  in  the  computa- 
tion, since,  in  all  probability,  the  density  increases  in  the  dispersed 
state,  owing  to  increased  compression  by  surface  tension.  This  error 
is  probably  small  unless  the  degree  of  dispersion  is  very  great.  By 
this  method,  particles  in  colloidal  gold  solutions  have  been  observed 
and  counted  whose  diameters  were  as  small  as  10~6  mm.  This  rep- 
resents about  the  limits  of  individual  visibility  under  the  ultramicro- 
scope,  that  is,  with  particles  much  smaller  than  this  the  field  appears 
diffusely  illuminated.  This  value  is  about  one-hundredth  that  of  the 
wave-length  of  violet  light,  and  about  ten  times  that  of  the  calculated 
diameter  of  the  ethyl  alcohol  molecule. 

The  rate  of  settlement  under  the  influence  of  gravity  has  also 
been  used  to  determine  the  size  of  colloid  particles.  By  means  of 
Stokes'  law  for  the  fall-rate  of  bodies  through  a  viscous  medium,  a 
comparatively  simple  equation  permits  of  the  calculation  of  the  diam- 
eter of  the  falling  body  when  the  fall-rate,  the  viscosity  of  the  me- 
dium, and  the  densities,  respectively,  of  the  dispersed  substance  and 
of  the  medium  are  known.  Perrin  5  used  the  same  principle  in  the 
preparation  of  suspensions  in  which  the  particles  were  all  of  more 
or  less  the  same  size,  using,  however,  regulated  centrifugation  in- 
stead of  simple  settling  under  gravity. 

There  has  also  been  developed,  largely  by  Bechhod 6  another 
method  which  throws  some  light  on  the  relative  sizes  of  particles, 
and  also  offers  a  very  interesting  and  valuable  experimental  weapon 
for  colloid  investigation.  This  is  the  method  of  ultrafiltration.  It 
has  been  found  possible  to  produce  graded  filters  which  allow  of  the 
passage  of  particles  below  a  certain  size,  and  which  restrain  any 
larger  ones.  These  filters  are  made  by  impregnating  ordinary  filters 
with  gelatin  and  other  colloidal  solutions  and  drying  with  special 
precautions.  The  permeability  decreases  with  the  concentration  of 
the  gelatin  or  other  substance  used. 

5  C.  B.,  146,  967,  1908. 

«Ztschr.  Chem.  Ind.  KolL,  2,  3,  1907;  Die  Kolloide  in  Biologic  u.  Medizin, 
Dresden,  1912. 


COLLOIDS  549 

The  investigations  as  to  the  size  of  particles  all  lead  to  two  gen- 
eral conclusions :  first,  suspensions  and  colloidal  solutions  in  general 
differ  from  one  another  mainly  in  degree  of  dispersion,  at  least  up 
to  the  limit  of  individual  detectibility  of  the  particles  under  the 
ultramicroscope,  beyond  which  point  at  present  all  is  speculation, 
although  the  presumption  is  strong  and  the  belief  is  growing  that 
there  is  also  no  other  fundamental  distinction  to  be  drawn  between 
colloidal  and  so-called  true  solutions ;  second,  it  is  always  found  that 
unless  special  purification  is  resorted  to  a  colloidal  solution  contains 
particles  of  widely  differing  sizes  side  by  side. 

2.  THE  BEOWNIAN  MOVEMENT. — About  a  century  ago  the  Eng- 
lish botanist,  Brown,  noticed  that  very  small  spores  and  other  bodies 
when  suspended  in  water,  and  observed  under  the  microscope,  were 
in  a  state  of  rapid  oscillatory  and  rotary  motion.  This  motion  of 
small  masses  of  matter  has  come  to  be  known  as  the  Brownian  move- 
ment. It  is  noticed  in  colloidal  solutions  whose  particles  are  not  too 
large,  and  at  the  same  time  are  large  enough  to  be  individually  de- 
tectible  under  the  ultramicroscope.  As  a  result  of  the  theoretical 
considerations  of  Einstein,7  of  Smoluchowski  8  and  of  Corbino,9 
and  of  the  experimental  researches  of  Svedberg  10  and  of  Perrin,11 
the  Brownian  movement  has  come  to  be  considered  as  nothing  more 
nor  less  than  a  manifestation  of  that  kinetic  energy  with  which  all 
matter  is  endowed,  and  which  forms  the  basis  of  the  kinetic  theory 
of  gases.  A  rapidly  gyrating  and  oscillating  colloid  particle  is  there- 
fore looked  upon  as  a  large  scale  picture  of  the  state  of  the  molecules 
themselves.  These  investigations  have  probably  done  more  than  any- 
thing save  the  development  of  the  kinetic  theory  itself  to  place 
molecular  and  atomic  speculations  on  a  firm  basis  of  plausibility. 

Svedberg' s  investigations  were  instituted  to  determine  the  mean 
velocity  of  colloid  particles  whose  mass  could  be  determined  by  the 
ultramicroscopic  method  above  referred  to.  Computing  from  these 
factors  the  average  kinetic  energy  of  the  particles,  this,  according  to 
Svedberg,  gives  the  same  value  which  would  be  computed  for  the 
particle  on  the  basis  of  the  kinetic  theory.  Perrin  attacked  the  prob- 
lem from  a  somewhat  different  point  of  view.  The  number  of  gas 
molecules  in  the  atmosphere  decreases  from  the  surface  of  the  earth 
outward  at  a  rate  which  is  determinable  by  computations  based  on 
the  kinetic  theory.  Perrin  set  himself  the  task  of  determining  the 
rate  of  decrease  in  the  concentration  of  the  particles  of  a  colloid  solu- 
tion, in  which  the  particles  were  of  uniform  size,  the  concentrations 
being  determined  at  different  levels  in  a  cylinder  in  which  the  solu- 

7  Ann.  der  Phys.,  9,  417;  11,  170;  17,  549;  19,  371. 

8  Ann.  der  Phys.,  21,  756. 
0  Nuovo  Cimento,  20,  5. 

10  Ztschr.  f.  Elektroc-hem.,  12,  853,  1906. 

11  C.  E.,  146,  967,  1908. 


550  INFECTION    AND    RESISTANCE 

tion  had  been  allowed  to  stand  until  it  had  reached  equilibrium  with 
the  gravitational  forces.  The  result  was  that  the  same  law  of  dis- 
tribution was  found  to  hold  in  this  case  as  in  the  case  of  the  atmos- 
phere. The  kinetic  theory  is  thus  shown  to  apply  quantitatively  as 
well  as  qualitatively  to  colloidal  solutions. 

3.  ELECTRICAL  PROPERTIES. — If  a  U-tube  be  filled  with  water, 
electrodes  placed  in  each  arm,  and  these  electrodes  maintained  at  a 
constant  difference  of  potential  either  by  a  battery,  dynamo,  or  other 
source  of  direct  current,  it  is  noticed  that  there  is  a  continual  flow  of 
liquid  in  the  tube,  in  one  direction  near  the  walls  and  in  the  opposite 
direction  in  the  interior  of  the  tube.     There  is  every  reason  to  believe 
that  such  currents  will  be  set  up  in  all  cases  whatever  the  nature  of 
the  liquid  or  of  the  tube,  although  the  current  set  up  may  in  par- 
ticular cases  be  very  small  and  even  very  rarely  approach  or  equal 
zero.     If  we  name  the  current  along  the  walls  simply  the  "current," 
and  that  through  the  interior  the  "countercurrent,"  then  in  the  case 
of  glass  and  water  the  direction  of  the  current  is  from  anode  to 
cathode,  and  that  of  the  countercurrent  is  from  cathode  to  anode. 
This  phenomenon  is  explained  by  the  hypothesis  that  at  the  surface 
of  contact  between  the  glass  and  the  water  there  is  established  a  dif- 
ference in  electrical  potential,  the  glass  becoming  negatively  and  the 
water  positively  electrified.     If  this  assumption  is  valid  it  follows 
that  if  a  particle  of  glass  placed  in  water  be  subjected  to  the  influence 
of  two   electrodes  placed   in  the  water,    it   will,   being  negatively 
charged,  be  attracted  by  the  positively  charged  electrode  (the  anode) 
and  repelled  by  the  negatively  charged  electrode  (the  cathode).     The 
result  would  be  a  wandering  of  the  particle  of  glass  through  the  solu- 
tion toward  the  cathode.     This  result  is  confirmed  by  ample  experi- 
ments.    Furthermore,  the  phenomenon  is  common  to  all  particles  in 
all  liquids,  so  far  as  is  known,  so  that  any  colloidal  solution  placed 
in  a  potential  gradient  will  show  wandering  of  its  particles  in  one 
direction  or  the  other.     Thus  in  water,  ferric  hydroxid,  chromium 
hydroxid  (and  most  hydroxids  in  the  colloidal  state),  methyl  violet, 
and  some  other  dyes  wander  to  the  cathode.    All  colloidal  metal  solu- 
tions, sulphur,  the  halogen  salts  of  silver,  chlorophyll,  rosin,  mastic, 
.most  dyes,  and,  in  fact,  the  great  majority  of  substances  investigated 
wander  toward  the  anode.     Albumin  (and  probably  some  other  sub- 
stances)  wanders  toward  the  cathode  in  acid  solution,  and  toward 
the  anode  in  alkaline  solution.     As  will  be  seen  later,  the  hypothesis 
of  the  existence  of  such  electrical  charges  on  colloid  particles  has 
been  of  very  great  use  in  explaining  many  forms  of  conduct  on  the 
part  of  dispersed  systems. 

4.  SURFACE  TENSION.— If  a  globule  of  mercury  be  divided  into 
two  parts,  these  two  parts  will  unite  again  if  opportunity  be  given. 
All  the  opportunity  which  is  necessary,  if  the  surfaces  be  clean,  is  to 
bring  the  two  parts  into  mechanical  contact.     The  union  of  the  sep- 


COLLOIDS  551 

arate  parts  may,  however,  occur  in  a  variety  of  other  ways,  in  fact, 
in  any  way  whatever  whereby  such  union  is  physically  possible. 
Thus,  if  the  separate  portions  be  of  different  size,  the  smaller  one 
will  have  a  higher  vapor  pressure  than  the  larger,  and  evaporation 
from  the  smaller  to  the  larger  will  take  place  until  the  whole  of  the 
smaller  portion  has  transferred  itself  to  the  larger  one,  and  the  re- 
union is  therefore  complete.  There  is  every  reason  to  believe  that  if 
the  two  portions  of  unequal  size  were  made  electrodes  in  a  galvanic 
cell,  and  this  cell  were  then  short  circuited,  that  the  smaller  portion 
would  go  into  solution  and  again  deposit  upon  the  larger  one.  In 
case  the  two  portions  were  of  the  same  size,  these  forms  of  recom- 
bination, with  the  exception  of  that  of  direct  coalescence,  would  not 
occur  if  all  other  conditions  were  kept  constant,  but  a  slight  differ- 
ence of  conditions  in  respect  to  the  two  portions  would  start  the  act 
of  recombination,  which  would  then  in  general  proceed  to  comple- 
tion. The  same  tendency  is  noticed  with  all  substances.  Thus  in  a 
liquid  small  crystals  disappear  while  larger  ones  grow  at  their  ex- 
pense, and  it  may  be  stated  that,  other  influences  for  the  moment 
ignored,  the  most  stable  configuration  which  can  be  assumed  by  a 
given  mass  of  any  substance  is  that  in  which  all  of  the  substance  is 
in  one  portion,  and  that  portion  is  spherical  in  form.  This  is  equiv- 
alent to  saying  that  all  bodies  so  arrange  themselves  as  to  expose  the 
least  possible  surface.  The  force  which  tends  to  bring  about  this 
condition  is  called  surface  tension.  In  so  far  as  surface  tension 
alone  is  concerned  it  follows  that  any  colloidal  solution  must  be  un- 
stable, and  tend  to  condense  itself  until  all  of  the  dispersed  matter 
has  aggregated  itself  together  into  a  single  mass  of  spherical  form. 

But  there  are  many  other  forces  which  may  under  certain  con- 
ditions act  against  surface  tension.  If  the  dispersed  substance  is 
one  that  is  crystalline,  the  directive  forces  of  crystallization  over- 
come those  of  surface  tension,  and  the  form  of  stable  configuration 
will  be  that  of  the  crystal  instead  of  spherical,  and  equilibrium  will 
be  established  when  all  of  the  available  substance  has  aggregated 
itself  together  into  one  large  crystal.  We  know,  on  the  Other  hand,  a 
great  many  colloidal  solutions  which  seem  to  be  quite  stable  even  in 
very  high  degrees  of  dispersion.  To  explain  such  cases  we  must  look 
for  other  forces  working  against  the  force  of  surface  tension.  If  the 
dispersed  particles  in  a  colloidal  solution  are  all  charged  with  the 
same  kind  of  electricity,  they  will  then  repel  one  another  with  a 
force  which  will  vary  inversely  as  the  squares  of  their  distances  from 
one  another.  This  repulsion  will  then  tend  to  work  against  any 
coalescence  or  other  sort  of  union  between  the  disperse  particles.  We 
have  already  seen  that  colloidal  particles  are  in  general  charged 
either  positively  or  negatively,  and  this  may  be  taken  to  some  extent 
as  explaining  the  stability  of  such  systems.  Equilibrium  results 
when  the  surface  tension  is  just  counterbalanced  by  the  electrical 


552  INFECTION    AND    RESISTANCE 

repulsion.  The  extension  of  this  idea  has  been  of  great  value  in 
colloid  investigation.  The  electrical  repulsion  will  not,  of  course, 
necessarily  prevent  the  smaller  particles  from  dissolving  and  deposit- 
ing upon  the  larger  ones,  unless  the  solubility  is  affected  by  the 
charge.  Concerning  this  we  know  nothing.  The  fact  that  colloid 
substances  possess  little  or  no  solubility  in  the  ordinary  sense  of  the 
word  means  such  solution  and  deposition  must,  of  necessity,  be  a 
very  slow  process,  and  the  colloid  solution  would  thus  appear  to  be 
perfectly  stable  over  very  long  periods  of  time.  There  are  many  who 
believe  that  all  such  systems  are  only  apparently  stable,  and  that  on 
account  of  the  absence  of  any  sufficiently  rapid  means  of  transforma- 
tion which  would  allow  the  stabilizing  influences  to  operate  rapidly 
enough  to  be  perceptible. 

Chemical  Properties  of  Colloids — 1.  It  is  reasonable  to  suppose 
that  the  chemical  properties  of  colloid  solutions  are  very  much  what 
is  to  be  expected  from  the  chemical  nature  of  the  dispersed  substance 
as  it  is  known  under  other  conditions.  The  colloidal  solutions  of  ar- 
senic sulphid  should  therefore  react  very  much  as  would  be  expected 
of  arsenic  sulphid  in  general,  except  in  so  far  as  the  substance  is  in  a 
finely  divided  state  in  the  presence  of  a  dispersing  medium  (water) 
in  which  it  is  little  soluble.  Thus  colloidal  arsenic  sulphid  is  soluble 
in  alkalies  and  alkaline  sulphid  just  as  is  the  massive  form.  If  a  rod 
of  zinc  is  suspended  in  a  colloidal  solution  of  arsenic  sulphid  there 
takes  place  a  slow  reaction,  lasting  over  weeks  and  even  months, 
whereby  the  sulphur  of  the  sulphid  unites  with  the  zinc  to  form 
colloidal  zinc  sulphid,  while  a  black  deposit,  probably  arsenic,  is 
found  on  the  zinc.  Chemical  reactions  with  colloids  are  thus,  as 
a  rule,  very  slow,  as  is  to  be  expected,  but  otherwise  not  essentially 
unusual. 

2.  The  exact  chemical  composition  of  the  disperse  phase  in  a 
colloidal  solution  is  probably  not  definitely  known  in  any  case.  In 
the  case  of  colloidal  metal  solutions,  such  as  gold  and  silver,  the  sus- 
pended particles  seem  to  be  practically  pure  metals,  but  in  most  cases 
the  composition  is  very  problematical.  The  'great  variation  in  the 
properties  of  such  solutions  with  variations  in  the  methods  of  prep- 
aration are  undoubtedly  to  a  great  extent  due  to  small  differences  in 
composition.  Thus  the  properties  of  arsenic  sulphid  vary  greatly 
with*  the  extent  to  which  free  hydrogen  sulphid  is  removed  from  the 
solution,  which  is  probably  due  to  the  differences  in  the  amount  of 
hydrogen  sulphid  absorbed  or  otherwise  held  by  the  arsenic  sulphid. 
Linder  and  Picton  believed  that  amorphous  copper  sulphid  was  a 
definite  compound  of  copper  sulphid  with  hydrogen  sulphid.  It  has 
also  been  found  that  amorphous  copper  sulphid  suspended  in  water 
continually  deposits  free  sulphur,  the  cupric  sulphid  being  at  the 
same  time  largely  converted  to  cuprous.  It  seems  to  be  rarely  or 
never  the  case  that  the  disperse  phase  may  be  looked  upon  as  a  sub- 


COLLOIDS  553 

stance  of  definite  composition,  being  usually,  if  not  always,  a  more 
or  less  complex  mixture  of  absorption  products. 

3.  Although  not  usually  pure  substances,  it  is  not  at  all  im- 
plausible to  assume  that  the  dispersed  particles  may,  to  some  extent, 
undergo  ordinary'  electrochemical  ionization,  in  which  case  the  par- 
ticles would  partake  of  the  nature  of  enormously  large  ions.  This 
assumption  is  interesting  as  offering  a  purely  electrochemical  ex- 
planation of  the  origin  of  the  charge  which  is  found  on  such  par- 
ticles, and  it  is  to  be  said  that  frequently  the  effect  of  foreign  sub- 
stances on  the  electrical  charges  of  suspended  particles  is  explain- 
able on  this  assumption.  For  further  information  on  these  matters 
reference  must  be  had  to  the  papers  of  Duclaux,12  Jordi,13  and 
P.  P.  von  Weimarn.14 

The  Flocculation  of  Colloids  by  Electrolytes — 1.  When  neutral 
salts  are  added  to  colloid  solutions  in  gradually  increasing  amounts 
there  always  follows  sooner  or  later  a  precipitation  of  the  dispersed 
substance.  If  the  salt  solution  be  removed  and  pure  water  added 
in  its  place,  this  decanted,  and  pure  water  again  added,  so  as  to 
wash  out  the  salt  as  thoroughly  as  possible,  the  final  result  may 
be  that  the  coagulum  (or  gel)  becomes  redispersed,  or  such  redis- 
persion  may  not  occur.  The  outcome  is  determined  both  by  the 
nature  of  the  colloid  and  by  that  of  the  neutral  salt  used. 

2.  If  acids  and  alkalies  are  the  electrolytes  used,  the  relation- 
ships are  somewhat  different.     The  addition  of  alkali  to  a  negatively 
charged  colloidal  solution  renders  it  more  stable,  while  a  positively 
charged  colloid  is  flocculated.     With  acids  the  reverse  of  this  condi- 
tion holds ;  that  is,  the  positive  colloid  is  stabilized  and  the  negative 
one  is  flocculated. 

3.  The  concentration  of  the  electrolyte  required  to  flocculate  a 
given  colloidal  solution  depends  very  greatly  on  the  nature  of  the 
electrolyte  used.     It  has  been  generally  considered  that  the  cathion 
of  the  electrolyte  is  the  active  agent  in  flocculating  negative  colloids, 
while  the  anion  is  active  in  the  case  of  positive  ones.     Thus  acids 
(hydrogen  ions)   are  very  effective  in  flocculating  arsenic  sulphid, 
and  alkalis  (hydroxyl  ions)   are  effective  with  ferric  hydroxid,  as 
might  be  inferred  from  the  preceding  paragraph.    Different  cathions, 
however,  show  very  different  degrees  of  precipitating  or  flocculating 
power.    Thus  in  the  case  of  arsenic  sulphid,  if  the  flocculating  power 
of  the  potassium  ion  is  taken  as  unity,  that  of  calcium  is  about 
twenty,  while  that  of  the  aluminium  ion  is  three  hundred  and  fifty. 
The  flocculating  power  in  this  case  increases  very  rapidly  with  the 

12  "Thesis,"  Paris,  1904. 

13  C.  R.,  136,  680,  and  1,448;  137,  122;  Bull.  Soc.  Chim.,  31,  573. 

14  Articles  in  Ztschr.  Chem.  Ind.  Roll,  1906-1911.     Also  in  "Grundziige 
der  dispersoid  Chemie,"  Dresden,  1911. 


554  INFECTION    AND    RESISTANCE 

valence  of  the  ion,  a  relationship  which  seems  to  be  quite  generally 
true.  It  has  been  quite  commonly  considered  that  the  effect  of  the 
aiiion  in  the  flocculation  of  negative  colloids  is  negligible.  That  this 
point  of  view  is  not  tenable  has  recently  been  strikingly  shown  by 
Sven  Oden15  in  his  work  on  colloidal  sulphur.  He  finds  that  the 
effect  of  electrolytes  on  sulphur  sols,  which,  like  arsenic  sulphid,  are 
negative  colloids,  is  distinctly  the  resultant  of  two  factors,  one  a  floc- 
culating effect  on  the  part  of  the  cathion,  the  other  a  dispersing  effect 
on  the  part  of  the  aniou.  It  is  not  improbable  that  our  views  in 
regard  to  the  phenomena  of  electrolyte  flocculation  will  undergo  con- 
siderable modification  in  the  near  future,  as  they  are  based  on  rather 
scanty  experimental  evidence.  Since  the  properties  of  sols  of  the 
same  materials  differ  very  considerably  with  the  most  minute  details 
of  their  method  of  preparation,  it  is  naturally  difficult  for  the  same 
investigator  even  to  obtain  uniform  results. 

4.  The  actual  concentration  of  a  given  electrolyte  required  to 
flocculate  a  sol  depends  also  very  greatly  on  the  nature  of  the  sol 
itself.     Some  sols  are  precipitated  by  very  small  concentrations  of 
electrolytes  (three  to  four  one-hundredths  normal  acid  being  usually 
sufficient  for  arsenic  sulphid),  while  gelatin,  albumin,  and  protein 
substances  in  general  require  far  higher  concentrations.     So  marked 
is  the  difference  in  many  cases  that  attempts  have  been  made  to  clas- 
sify colloids  on  the  basis  of  their  conduct  in  this  respect.     Thus  col- 
loids which  are  very  sensitive  to  electrolytes  are  called  suspension 
colloids,  while  those  that  are  not  very  sensitive  are  called  emulsion 
colloids.    To  the  first  class  belong  all  the  true,  rather  coarse-grained 
suspensions,   while   the   sols    that  yield    soft  gelatinous    flocculates 
usually  fall  into  the  second  class.    It  is  also  very  frequently  true  that 
the  latter  type  shows  the  phenomenon  of  reversible  flocculation  (see 
ante).     This  classification  is  for  many  purposes  quite  useful,  but 
cannot  be  considered  as  very  fundamental.    For  example,  if  the  elec- 
trolyte used  be  a  salt  of  a  heavy  metal  most  of  the  so-called  emulsion 
colloids,  such  as  albumin,  are  irreversibly  flocculated. 

5.  The  flocculation  of  sols  by  electrolytes  is  usually  explained 
as  due  to  the  phenomenon  of  absorption.     That  is,  the  flocculating 
ion  is  absorbed  from  the  solution  by  the  dispersed  particles.     Since 
Jn  general  the  ion  which  is  absorbed  is  the  one  whose  electrical  charge 
is  opposite  to  that  of  the  dispersed  particle  the  absorption  results  in 
a  reduction  of  the  charge  on  the  particle,  and  allows  the  aggregating 
forces  of  surface  tension  to  become  operative.     The  evidence  of  the 
validity  of  this  assumption  is  considerable.     Thus  the  flocculated 
colloids  always  contain  appreciable  amounts  of  the  ion  which  caused 
the  precipitation,  which  is  prima  facie  evidence  of  the  absorption. 
Furthermore,  the  electrical  charges  on  the  particles  may  be  meas- 
ured by  determining  their  rates  of  migration,  and  the  effect  of  elec- 

15"Inau£.  Diss.,"  Upsala,  1913,  pp.  118  et  seq. 


COLLOIDS  555 

trolytes  on  this  charge  may  'also  be  observed.  It  is  very  commonly 
true  that  the  addition  of  a  precipitating  electrolyte  to  a  sol  reduces 
materially  the  charge  of  the  particles.  This  is  not,  however,  always 
true,  and  the  relationships  are  more  complicated  than  has  generally 
been  assumed.  Nor  is  it  always  true  that  the  complete  neutraliza- 
tion of  the  electrical  charge  on  dispersed  particles  results  in  floccula- 
tion. For  example,  acetic  acid  may  be  added  to  arsenic  sulphid  sols 
in  such  quantities  as  to  completely  neutralize  the  negative  charges  of 
the  particles  and,  further,  so  much  acid  may  be  added  that  the  par- 
ticles acquire  a  very  considerable  positive  charge,  all  this  without 
the  least  signs  of  flocculation.  Ultimately  so  much  acid  may  be 
added  as  to  cause  flocculation. 

6.  In  the  flocculation  of  sols  by  electrolytes  there  is  frequently 
observed  a  curious  effect  known  as  the  "zone-phenomenon"  It  is 
observed  when  increasing  amounts  of  certain  electrolytes  are 
added  that  at  a  certain  concentration  flocculation  will  be  brought 
about,  while  if  the  concentration  be  greater  flocculation  will  not 
occur,  although  still  further  increase  of  concentration  will  result  in 
another  flocculation  zone.  The  phenomenon  is  most  common  when 
the  electrolyte  used  is  a  salt  that  shows  marked  hydrolysis,  such,  for 
example,  as  ferric  chlorid.  If  a  negative  sol  be  treated  with  a  solu- 
tion of  this  salt  it  is  obvious  that  there  will  be  three  precipitating 
influences  present,  namely,  the  hydrogen  ions  and  the  colloidal  ferric 
hydroxid,  both  of  which  are  formed  by  the  hydrolysis  of  the  ferric 
chlorid  and  the  unhydrolyzed  ferric  chlorid.  Since  the  amounts  of 
these  precipitating  substances  in  a  given  solution  vary  with  the  con- 
centration, and  since  each  has  its  own  concentration  function  in 
precipitation,  it  will  be  seen  that  the  zone  phenomenon  may  be  ac- 
counted for  in  such  cases.  Since,  however,  the  zone  phenomenon 
occurs  in  many  cases  where  strongly  hydrolyzed  electrolytes  are 
not  used,  as,  for  example,  in  the  agglutination  of  bacteria  by 
citric  and  some  other  acids,  the  explanation  is  not  wholly  sufficient. 
There  are  also  many  curious  phenomena  concerned  with  the 
action  of  electrolytes  on  sois  which  have  as  yet  been  very  little 
investigated,  and  which  will  probably  throw  considerable  light  on 
the  subject. 

The  Mutual  Reactions  of  Colloids — The  conduct  of  mixtures  of 
two  different  sols  is  of  very  great  interest  and  variety,  both  in 
the  absence  and  in  the  presence  of  electrolytes.  A  number  of 
particular  cases  may  be  distinguished,  and  these  will  be  taken  up 
seriatim. 

1.  When  two  positive  or  two  negative  sols  are  mixed  together 
nothing  very  much  seems  to  happen.  It  is  generally  considered  that 
the  addition  of  one  sol  to  another  of  like  electrical  properties  results 
in  no  action.  Whether  this  is  wholly  true  or  not  is  doubtful,  but  it 
is  at  least  true  that  in  all  cases  investigated  up  to  the  present  time 


556  INFECTION    AND    RESISTANCE 

neither  individual  nor  mutual  flocculation  occurs.  Whether  the 
presence  side  by  side  of  two  sorts  of  similarly  charged  disperse  par- 
ticles results  in  any  change  in  the  dispersion  of  either  is  not  known. 
Except  to  show  that  no  mutual  precipitation  occurs,  these  cases 
have  been  but  little  studied. 

2.  When  two  oppositely  electrical  colloids  are  mixed  mutual 
precipitation  may  or  may  not  occur.     The  factor  which  in  the  main 
determines  the  outcome  is  the  relationship  between  the  amounts  of 
the  two  colloids  used.     If  neither  is  present  in  too  great  excess  com- 
plete mutual  precipitation  in  general  occurs.     If  either  is  present 
in  great  excess  precipitation  does  not  in  general  occur. 

3.  The  effect  of  a  great  excess  of  one  colloid  in  preventing 
mutual  precipitation  is  very  marked  when  the  colloid  in  excess  is 
one  of  the  emulsion-colloid  type.     Thus  gelatin,  a  typical  emulsion 
colloid,  when  present  in  excess  over  another  colloid  very  frequently 
prevents  all  flocculation,  even  in  fairly  coarse-grained  suspensions. 
Advantage  has  been  taken  of  this  action  in  preventing  scaling  in 
boilers.     This  scaling  is  due  largely  to  the  fact  that  lime  salts  held 
in  solution  as  bicarbonates  are  decomposed  by  heat,  with  the  separa- 
tion of  calcium  carbonate,  at  first  in  the  highly  dispersed  state.     This 
gradually  aggregates  together  and  deposits  on  the  interior  of  the 
boiler  as  an  amorphous  flocculated  colloid,  which  in  time  becomes 
very  compact,  and  in  many  cases  crystalline.     If,  however,  a  small 
amount  of  glue  (impure  gelatin)  be  added  to  the  boiler  water  the 
colloidal  constituents  of  the  water  do  not  flocculate  and  compact, 
but  remain  suspended,  and  may  from  time  to  time  be  blown  off. 
Another  illustration  is  found  in  the  preparations  of  photographic 
emulsions.     The  silver  halides  flocculate  very  readily  in  pure  water, 
but  in  gelatin  solution  remain  in  a  highly  dispersed  state,  which  is 
necessary  to  the  preparation  of  the  plate.     In  this  case,  not  only  is 
the  suspension  protected  from  flocculation.,  but  also  a  degree  of  dis- 
persion is  reached  which  is  far  beyond  anything  attainable  in  pure 
water.     The  same  is  true  when  lysalbinic  acid  is  used  to  prevent 
flocculation  of  colloidal  silver.     In  pure  water  only  very  dilute  sus- 
pensions of  metallic  silver  are  obtainable,  but  in  the  presence  of  lysal- 
binic acid  suspensions  containing  as  high  as  ninety  per  cent,  of  silver 
are  obtainable,  the  product  being  used  medicinally  under  the  name  of 
"argyrol."     These  are  the  phenomena  known  as  "protective  actions," 
and  the  gelatin,  albumin,  or  other  colloid  which  exerts  the  protective 
action  is  spoken  of  as  a  "protective  colloid/' 

4.  The  protective  action  of  certain  colloids  is  not  only  exerted 
against  the  tendency  of  the  protected  colloid  to  spontaneously  floc- 
culate, but  also  a  certain  and  very  great  protection  is  offered  against 
flocculation  by  electrolytes.     Thus  Zsigmondy  16  was  able  to  find  a 
definite  measure  of  the  protective  action  of  certain  colloids  on  the 

16  Ztschr.  f.  analyt.  Ch.,  40,  697,  1902. 


COLLOIDS  557 

precipitation  of  gold  suspensions  by  means  of  sodium  chlorid.  The 
method  was  to  find  the  amount  of  the  protective  colloid  that  was  just 
necessary  to  prevent  the  flocculation  of  a  fixed  amount  of  a  given  gold 
sol  by  a  fixed  amount  of  the  salt.  In  this  way  it  was  observed  that 
the  protective  action  of  different  colloids  is  very  different.  Thus  if 
the  amount  of  starch  in  solution  which  is  necessary  for  protection 
be  taken  as  2,500  the  amounts  of  various  other  colloids  required  are 
as  shown  in  the  following  table : 

Protective  colloid ....  Starch  Dextrin  Gum  Arabic  Albumin  Gelatin  Glue 
Amount  required 2,500  1,200  40  25  11 

5.  Very  simple  theories  are  devisable  to  explain  the  interactions 
between  different  colloidal  solutions.  Thus  two  oppositely  electrical 
colloids  may  be  considered  to  precipitate  one  another  mutually  be- 
cause of  the  electrical  attraction  existing  between  all  oppositely 
charged  particles.  This  results  in  bringing  together  the  oppositely 
charged  particles  with  the  formation  of  relatively  neutral  aggregates, 
wiiich,  as  shown  in  the  discussion  of  the  flocculation  by  electrolytes, 
is  a  condition  favorable  to  precipitation.  For  very  obvious  reasons, 
then,  no  flocculation  would  be  expected  when  two  like  charged  col- 
loids are  mixed.  Many  objections  may  be  made  to  the  unqualified 
acceptance  of  this  explanation.  It  offers,  nevertheless,  a  valuable 
leading  idea  when  not  accepted  too  dogmatically. 

A  number  of  factors  probably  contribute  to  the  protective  action 
of  many  colloids.  To  some  extent  the  effect  may  be  purely  mechani- 
cal, since  increased  viscosity  imparted  by  the  presence  of  the  protec- 
tive emulsion  colloid  will  shorten  the  mean  free  path  of  the  floc- 
culable  particles,  and  thus  materially  lessen  the  probability  of  im- 
pacts between  them.  Consequently  flocculation  wrill  not  occur  as 
readily.  Further,  the  ultramicroscope  gives  considerable  evidence 
of  the  existence  of  another  and  very  important  factor.  It  seems 
certain,  at  least  in  many  cases,  that  the  protective  colloid  arranges 
itself  in  a  film  or  coating  around  the  flocculable  particles,  and  in  this 
way  prevents  the  aggregation  of  the  particles.  These  factors  are  not, 
however,  sufficient  to  explain  all  cases  of  protective  action.  It  must 
be  considered  that  in .  some  cases,  at  least,  the  protective  colloid 
exerts  a  truly  dispersive  effect,  such  as  would  result  from  a  nullifica- 
tion of  surface  tension  forces.  The  ease  with  which  a  small  amount 
•of  soap  will  emulsify  a  large  amount  of  oil  is  difficult  to  explain  on 
any  other  hypothesis.  When  an  oil  is  shaken  with  pure  water  little 
or  no  emulsification  results,  while  in  the  presence  of  the  soap  as  a 
protective  colloid  the  same  amount  of  work  in  shaking  accomplishes 
enormously  greater  results.  In  fact,  the  oil  will  spontaneously 
emulsify  by  merely  standing  in  contact  with  the  soap  solution.  The 
equilibrium  condition  of  oil  in  contact  with  pure  water  is  reached 


558  INFECTION    AND    RESISTANCE 

when  the  oil  is  very  little,  if  any,  dispersed,  while  in  contact 
with  soap  solution  equilibrium  is  reached  only  when  the  oil 
is  highly  dispersed.  '  From  the  surface  tension  point  of  view  this 
would  be  expressed  by  saying  that  in  contact  with  pure  water 
the  surface  tension  is  large  and  positive,  while  in  contact  with 
soap  solution  the  surface  tension  is  large  but  negative.  It  is  im- 
possible to  say  to  what  extent  these  effects  may  be  due  to  obscure 
chemical  action. 

The  Preparation  of  Colloidal  Solutions — 1.  This  subject  might 
perhaps  upon  logical  grounds  have  best  been  treated  in  an  opening 
paragraph.  However,  with  the  conclusions  that  have  been  reached 
in  the  foregoing  discussion  the  whole  matter  may  be  dismissed  very 
briefly.  The  conditions  which  must  be  fulfilled  in  order  to  obtain 
colloidal  solutions  are  in  the  main  summarized  in  the  following 
paragraphs : 

2.  A  medium  must  be  chosen  in  which  the  given  substance  does 
not  reach  to  any  great  extent  the  maximum,  molecular  degree  of  dis- 
persion— that  is,  in  which  what  is  usually  called  true  solution  does 
not  occur  to  any  great  extent.     While,  as  pointed  out  at  the  begin- 
ning of  this  discussion,  there  is  no  sharp  line  to  be  drawn  between 
colloidal  and  true  solutions,  the  substances  that  are  distinctly  recog- 
nizable at  the  present  time  as  colloidal  solutions  carry  particles  which 
are  in  the  neighborhood  of  one  thousand  times  as  large  as  average 
molecules.     From  media  in  which  the  dispersion  is  approximately 
molecular  the  dispersed  substance  usually  shows  a  strong  tendency  to 
separate  in  the  crystalline  form,  although  it  may  frequently,  if  not 
generally,  first  appear  in  the  colloidal  form,  then  more  or  less  rapidly 
becoming  crystalline.      The  only  distinction  to  be  drawn   is  this: 
from  solutions  in  which  the  dispersion  is  very  great,  approximating 
the  molecular,  separation  in  short  time  in  the  crystalline  form   is 
favored ;  from  solutions  in  which  the  dispersion  does  not  approximate 
the  molecular  separation  persistence  in  the  amorphous  state  is  fa- 
vored. 

3.  A  colloidal  sol  or  gel,  one  or  the  other,  may  generally  be  pro- 
duced from  any  substance  in  any  medium  in  which  the  amount  of 
molecular  dispersion  is  at  a  minimum  by  means   of  any  reaction 
whereby  the  new  substance  is  produced  from  solution.     Thus,  for 
example,  arsenic  sulphid  does  not  disperse '  in  water  to  molecular 
extent  in  any  considerable  degree.     Therefore,  by  mixing  together 
a  solution  containing  an  arsenic  compound  which  is  soluble,  and  a 
solution   of  hydrogen    sulphid,   the    resulting   arsenic   sulphid   will 
appear  in  the  colloidal  state.     Whether  it  appears  in  the  dispersed 
state  as  a  sol  or  in  the  flocculated  state  as  a  gel  will  depeud  mainly 
on  the  electrolyte  content  of  the  solutions  which  are  used.     Tf  a  solu- 
tion of  arsenic  chlorid  be  used  the  resulting  solution  will  contain 
.considerable  free  hydrochloric  acid,  and  the  tendency  wil*  be  for  the 


COLLOIDS  559 

resultant  arsenic  svtlphid  to  appear  in  the  flocculated  or  gel  form. 
If  it  is  desired  to  prepare  the  substance  in  the  sol  form  electrolytes 
must  be  avoided.  This  may  be  done  by  using  a  solution  of  arsenic 
trioxid  instead  of  one  of  arsenic  chlorid. 

Any  reaction  which  is  brought  about  under  the  above  conditions 
will  result  in  the  formation  of  a  colloidal  product.  The  dialysis  of 
salts  which  form  insoluble  hydroxids  simply  allows  the  normal 
reaction  of  hydrolysis  to  complete  itself.  The  resulting  hydroxids 
appear  in  such  a  way  as  to  fulfill  the  above  conditions,  and  conse- 
quently appear  in  the  colloidal  state.  In  this  connection  note  the 
preparation  of  colloidal  sodium  chlorid  (see  ante). 

4.  In  an  appropriate  medium  most  if  not  all  substances,  crystal- 
line or  otherwise,  may  be  brought  directly,  that  is,  without  the  inter- 
vention of  specific  chemical  reactions,  into  the  colloidal  state.  In 
some  cases  this  may  be  accomplished  by  mechanical  means.  Thus 
oils  violently  shaken  with  water  disperse  to  some  extent  and  form 
emulsions  of  greater  or  less  stability.  By  shaking  glass,  quartz,  and 
the  like  with  various  liquids  in  which  they  are  virtually  insoluble  the 
abrasion  results  in  the  formation  of  more  or  less  dispersed  systems, 
usually  not  very  stable. 

Many  metals  may  be  brought  into  the  dispersed  state  by  causing 
an  electric  arc  to  pass  between  points  held  under  a  liquid.  This 
electrical  dispersion  method  has  been  very  considerably  used,  but  is 
obviously  confined  to  substances  which  are  conductors  of  electricity. 

On  the  other  hand,  many  substances  when  merely  brought  into 
contact  with  an  appropriate  medium  spontaneously  undergo  disper- 
sion. Thus  gelatin,  glue,  tannin,  and  many  other  substances  spon- 
taneously disperse  in  water.  Even  crystalline  substances  frequently 
do  this.  Thus  soaps  which  have  been  crystallized  from  alcohol  solu- 
tions when  brought  into  water  disperse  in  the  colloidal  state.  Crys- 
tallized cuprous  sulphid,  the  mineral  known  as  chalcocyte,  disperses 
in  the  colloidal  form  in  solutions  of  hydrogen  sulphid. 

Substances  which  go  spontaneously  into  the  dispersed  colloidal 
state  are  usually  spoken  of  as  "lyophyllic/'  while  those  that  tend  to 
spontaneously  leave  the  dispersed  state  are  called  "lyopkobic."  It  is 
evident  that  a  substance  may  be  lyophyllic  with  respect  to  one  me- 
dium and  lyophobic  with  respect  to  another.  Furthermore,  a  very 
small  change  in  the  nature  of  the  medium  may  cause  the  change 
from  a  lyophyllic  to  a  lyophobic  colloid.  Thus  oils  are  lyophobic 
with  respect  to  water,  but  lyophyllic  with  respect  to  even  very  dilute 
soap  solutions. 

Applications  to  Biology. — 1.  When  one  considers  the  relatively 
infrequent  occurrence  in  biological  systems  of  either  crystalline  sub- 
stances, or  of  substances  that  may  be  readily  made  to  crystallize 
from  water  (the  universal  biological  dispersing  medium),  it  imme- 
diately becomes  evident  that  the  chemistry  and  physics  of  such 


560  INFECTION    AND    RESISTANCE 

systems  must  be  in  the  main  colloidal.  All  biochemistry  is  thus  in 
the  main  colloid  chemistry.  Aside  from  mineral  salts,  urea,  uric 
acid,  and  a  few  other  bodies,  the  reagents  and  products  which  are 
active  in  life  processes  are  all  to  be  found  in  the  living  organism 
in  the  colloidal  state.  While  crystalline  directive  forces  are  in  gen- 
eral absent,  we  have  nevertheless  to  deal  in  biological  phenomena 
with  a  great  variety  of  directive  forces  of  a  wholly  different  charac- 
ter. Colloidal  substances  in  high  degrees  of  dispersion  such  as 
proteins  and  the  like  in  the  alimentary  fluids  are  being  continually 
converted  into  active  living  tissues,  a  process  manifestly  involving 
very  definite  directive  forces,  since  the  product  (the  tissue  cells) 
is  an  organized  one,  even  though  its  organization  is  not  similar  to 
that  of  a  crystal.  Thus  the  building  of  living  tissue  involves 
among  other  things  the  conversion  of  sols  of  many  sorts  (alimen- 
tary fluids,  blood,  etc.)  into  gels.  In  other  words,  the  living 
tissue  is  to  a  certain  extent  to  be  looked  upon  as  a  colloidal  gel,  dif- 
fering, however,  from  laboratory  gels  in  possessing  a  definitely  or- 
ganized cell  structure.  While  it  is  true  that  in  the  spontaneous  gel 
formation  with  certain  colloids,  as,  for  example,  gelatin,  myelin,  or 
web  structures  are  formed  purely  as  a  result  of  the  physical  and 
chemical  forces  active,  these  cannot  be  said  to  bear  any  very  strong 
resemblance  to  living  cells.  It  is  thus  apparent  that  life  processes 
differ  very  materially  from  those  of  the  chemical  laboratory.  On  the 
other  hand,  it  is  true  that  many  of  the  component  reactions  which  go 
to  make  up  the  life  process  may  be  very  closely  duplicated  by  labora- 
tory means,  and  that  already  the  study  from  the  colloid  chemistry 
point  of  view  of  the  reaction  of  many  of  the  substances  which  go  to 
make  up  the  living  organism  has  given  interesting  and  important 
results.  Furthermore,  the  whole  field  of  colloid  investigation  has 
been  greatly  stimulated  by  the  hearty  support  and  encouragement 
which  it  has  received  from  biologists.  Some  slight  attempt  will  be 
made  here  to  illustrate  by  a  few  examples  the  far-reaching  possi- 
bility of  explanation  which  colloid  chemistry  offers  of  certain  phases 
of  biological  science.  The  actual  accomplishment  in  the  field  is 
already  so  great  that  only  a  very  limited  discussion  can  be  offered 
here.  .*. 

2.  The  action  of  electrolytes  on  emulsions  of  bacteria  is  wholly 
analogous  to  their  action  on  colloidal  suspensions.  The  bacterial 
emulsions  are  very  sensitive  to  flocculation  by  mineral  acids,  one  thou- 
sandth normal  hydrochloric  and  sulphuric  acids  usually  being  suffi- 
cient to  cause  complete  clumping  and  settling  of  the  bacteria.  Neu- 
tral salts,  with  the  exception  of  those  of  silver,  mercury,  iron,  and 
aluminium,  do  not  flocculate  the  bacteria.  If,  however,  the  bacteria 
are  first  treated  in  the  absence  of  electrolytes  with  an  agglutinating 
serum  small  concentrations  of  salt  solutions  will  bring  about  floccu- 
lation. It  is  also  known  that  bacteria  have  the  power  of  absorbing 


COLLOIDS  561 

agglutinins  from  sera,  so  that  it  is  evident  that  what  we  have  here 
is  a  case  of  the  production  of  a  flocculable  combination  of  bacteria 
and  agglutinin,  neither  component  of  which  is  alone  flocculable. 

Citric  acid  in  concentrations  ranging  between  one  one-hundredth 
and  one  eight-hundredth  normal  produces  flocculation.  In  either 
greater  or  less  concentrations  no  flocculation  is  produced,  which  is  an 
illustration  of  the  so-called  zone  phenomenon. 

Like  all  other  suspensions,  bacteria  are  electrically  charged,  and 
consequently  wander  in  the  electric  field.  Under  all  ordinary  condi- 
tions the  charge  which  they  carry  is  negative,  from  which  in  their 
general  conduct  it  might  be  expected  that  they  would  conduct  them- 
selves similarly  to  arsenic  sulphid,  which  is  also  negatively  charged. 
This  is  found  to  be  the  case.  Their  rate  of  migration  is  reduced 
by  acids  and  evidently  somewhat  increased  by  alkalies,  although 
very  little  alkali  is  necessary  to  bring  about  disintegration  of  the 
bacteria. 

It  has  also  been  found  that  all  colloids  are  more  or  less  sensitive 
to  light  in  respect  to  their  migration  rates.  Bacteria  also  show  this 
phenomenon,  as  they  migrate  notably  slower  in  the  light  than  in  the 
dark.  It  is  quite  possible  that  this  reduction  in  their  electrical 
charge  may  to  some  extent  be  responsible  for  the  bactericidal  action 
of  light. 

3.  A  very  interesting  application  of  the  principles  of  colloidal 
precipitation  by  electrolytes  has  recently  been  made  by  Loeb.17    He 
finds  that  the  eggs  of  the  Fundulus,  a  small  fish,  are  killed  by  being 
immersed  in  a  pure  isotonic  salt  solution,  in  spite  of  the  fact  that 
they  normally  develop  in  sea  water.    The  factor  which  allows  of  their 
development  must  therefore  be  sought  in  some  other  constituent  of 
the  sea  water  which  is  absent  in  the  pure  salt  solution.     This  Loeb 
finds  in  the  presence  of  small  amounts  of  calcium  salts.     Further, 
if  a  small  amount  of  calcium  chlorid  be  added  to  the  pure  salt  solu- 
tion the  eggs  will  develop  in  it  as  well  as  in  sea  water.     Loeb's  ex- 
planation is  simple  and  very  ingenious.    He  supposes  that  the  sodium 
chlorid  is  toxic,  provided  it  can  diffuse  into  the  egg.     In  the  ahsence 
of  calcium  salts  such  diffusion  is  possible  because  the  sodium  chlorid 
is  not  a  sufficiently  powerful  colloid  precipitant  to  make  the  mem- 
brane about  the  egg  impervious.     It  is,  however,  very  well  known 
that  calcium  salts  (and  all  bivalent  cathions)  are  far  more  effective 
colloid  precipitants  than  sodium  ions.     Consequently  the  presence 
of  a  relatively  small  amount  of  calcium  chlorid  in  the  salt  solution 
is  sufficient  to  so  condense  the  egg  membrane  as  to  make  it  impervious 
to  the  sodium  chlorid,  and  thus  render  the  latter  non-toxic. 

4.  Another   interesting   application   of   colloidal   principles   is 
found  in  what  is  known  as  the  Danysz  phenomenon.     It  is  found 
that  the  neutralization  of  the  toxicity  of  diphtheria  toxin  by  the 

17  Am.  J.  Physiol.,  6,  411,  1902. 


562  INFECTION    AND    RESISTANCE 

antitoxin  depends  on  the  way  in  which  the  two  are  mixed.  If  a 
quantity  of  toxin  just  sufficient  to  neutralize  a  fixed  amount  of  anti- 
toxin when  it  is  added  all  at  once  be  in  another  experiment  added  in 
small  installments,  the  resulting  mixture  will  be  found  to  be  still 
quite  strongly  toxic.  This  is  quite  analogous  to  what  is  found  in  the 
interaction  of  many  colloids.  The  amount  of  a  given  colloid  re- 
quired to  neutralize  and  precipitate  another  depends  greatly  on  the 
way  in  which  it  is  added. 

5.  Homer,18  Field  and  Teague,19  and  Teague  and  Buxton  20  all 
carried  out  interesting  investigations  directed  toward  determining 
the  migration  directions  (electrical  charges  of  toxins  and  antitoxins). 
Their  conclusions  were  that  all  wandered  toward  the  cathode,  and 
that  all  were  therefore  positively  charged.     If  this  is  correct  the 
analogy  between  toxin  and  antitoxin  reactions  and  those  of  simple 
colloids  is  rather  mutilated,  since  two  positive  colloids  are  not  sup- 
posed to  react  with  one  another.     It  is,  however,  more  than  possible 
that  the  above  experiments  are  misleading.     In  all  cases  agar  dia- 
phragms were  used.     Through  these  there  would  always  occur  a 
streaming  of  water  toward  the  cathode  as  a  result  of  the  electrical 
potential  between  the  agar  and  the  water.     This  might  well  be  so 
great  as  to  obscure,  and  even  more  than  overcome  any  anodic  wan- 
dering that  might  occur.     Furthermore,  the  conduct  of  proteins  in 
general  in  the  electric  field  is  a  very  complex  one,  and  one  that  is 
only  just  beginning  to  be  understood.     For  these  reasons  it  is  at 
present  very  dangerous  to  draw  any  very  dogmatic  conclusions. 

6.  In  closing  mention  may  be  made  of  what  seems  to  be  an  im- 
munity phenomenon  which  seems  rather  clearly  to  be  a  case  of  pro- 
tective colloid  action.     It  is  observed  when  an  agglutinin  is  added  to 
a  bacterial  emulsion  that  if  an  excess  of  the  agglutinin  be  added  no 
agglutination  occurs.     This  is  wholly  analogous  to  the  fact  that, 
while  a  small  amount  of  gelatin  will  precipitate  arsenic  sulphid  sus- 
pension, a  larger  amount  will  not. 

Conclusions — While  great  progress  has  been  made  in  the  field  of 
colloid  investigation  from  the  chemical  and  physical  sides,  and  while 
also  many  very  striking  analogies  are  to  be  found  from  the  biological 
side,  it  is  nevertheless  true  that  we  are  still  very  much  in  the  dark 
in  regard  to  a  great  many  matters.  The  one  great  difficulty  which 
lies  in  all  such  investigations  is  that  it  is  a  matter  of  very  great  diffi- 
culty to  duplicate  results.  The  nature  of  any  colloid  sol  or  gel 
depends  so  greatly  upon  its  whole  previous  history,  apparently 
down  to  the  least  detail,  that  great  discrepancies  in  experimental 
results  are  found.  Even  the  age  of  a  sol  is  frequently  a  matter  of 
very  great  importance  in  determining  its  properties.  For  example 

18  Berl.  klin.  Wochenschr.,  Vol.  41,  p.  209,  1904. 

19  Journ.  Exp.  Med.,  Vol.  9,  pp.  86  and  223. 

20  Ibid.,  Vol.  9,  p.  254. 


COLLOIDS  563 

a  freshly  prepared  gelatin  solution  will  not  precipitate  arsenic  sul- 
phid,  but  it  will  do  so  after  it  has  stood  for  some  hours.  What  is 
now  greatly  needed  is  more  data  on  a  greater  variety  of  colloids  that 
have  heretofore  been  investigated  and  work  directed  toward  the 
preparation  of  colloidal  solutions  of  definite  character.  Until  some- 
thing has  been  accomplished  in  these  directions  all  biological  anal- 
ogies and  the  like  cannot  be  anything  more  than  qualitative,  and  the 
same  holds  true  for  many  of  the  physical  and  chemical  conclusions 
which  have  been  discussed  in  this  chapter. 


INDEX 


Abderhalden,  protective 
fermenti  of,  532, 
533,  534 

Abderhalden's  reaction,  534 
Abscesses,     secondary, 
caused      by      bac- 
teria, 25 

Absorption  theory  in  tox- 
in-antitoxin reac- 
tion, 123 

Acne,  opsonic  index  in 
vaccine  treatment 
of,  339 

Adrenal  cytotoxin,  92 
Agglutination,      absorption 
experiments         of 
Castellani  on,  232 
acid,  246 

value  of,  for  differen- 
tial purposes,  240, 
247 

action    of    salts    in,    240 
agglutinability     of     bac- 
teria in,  in  agglu- 
tinoids,  229 
normal   differences   in, 
between       strains 
of    same     species, 
228,  229 

agglutinins   in,   223.   224 
absorption  of,  232 
complete,       impossi- 
bility of.  232 
heating  of,  226 

explanation   of,    236 
major,  229 
normal,  233 

explanation    of.    234 
qualitatively    identi- 
cal    with     "im- 
mune"     agglutin- 
ins, 234 

para-  or  minor.  229 
agglutinogen  in,  223,  224 
effects    of   heating    of, 

226 

localization   of,   in   ec- 
toplasmic      layers 
of  cells,  224 
agglutinoids  in,  235 
biological        relationship 

and,  230 
Bordet's   explanation   of, 

240 
by    means    of    cell    body 

proper,    225 

by  means  of  ectoplasmic 
substances  of  bac- 
teria, 224,  225 
by     means     of     flagella, 

224.  225 
cataphoresis   of   bacteria 

in.   242,   243 
description  of,  218 
effect  of  gelatin  addition 
in      mastic     solu- 
tion on,  244 

Ehrlich's    interpretation 
of  process  of,  234 


Agglutination,  Ehrlich's  in- 
terpretation o  f 
process  of,  dia- 
grammatic repre- 
sentation of,  235 
Eisenberg  and  Yolk's  in- 
terpretation of, 
235 

Ficker's  reaction  in,  223 
flocculation     of     colloids 

and,  241 
mechanism      of,      241, 

242 

mutual,  242 
group,  229 
cause  of,  230 
diagnostic     value     of, 

231 

hemagglutination    analo- 
gous to,  236,   237 
history  of,  218 
importance     of     electro- 
lytes in,  240 
in  colloidal  solutions,  in- 
hibition  zones   in, 
245 

in    excess   of   agglutinin, 
colloid       phenom- 
enon and,  562 
in  immune  serum,  141 
in    motile    and    non-mo- 
tile bacteria,  224, 
225 

in  salt-free  environ- 
ment, by  addition 
of  organic  sub- 
stances, 245 

influence  of  immuniza- 
tion with  different 
animal  species 
on,  232 

with  different  species 
of  bacteria  on, 
232 

influence     of     salts     on 

sensitized  and  un- 

sensitized      b  a  c  - 

teria  in,  243,  244 

experiments  of  Neisser 

and      Friedemann 

in,  244 

inhibition  zones  in,  162, 

236 

iso-agglutinins  in.  237 
grouping  of,  237,  238 
value  of  presence  of, 

239 

methods  of,    218   et  seq. 
Bordet's,  223 
Ficker's     reaction     in, 

223 

Gruber-Widal,  220 
macroscopic.  219 
microscopic,  220 
Proescher's,  220 
thread       reaction       of 
Pfaundler,        222, 
223 
nature  of,  223 

565 


Agglutination,    not    associ- 
ated   with   life   of 
bacteria,  222,  223 
of  "agglutinin"  bacteria, 

243 

of  bacteria  in  active  im- 
munization, 89 
of    capsulated    bacteria, 

243 

phenomenon    of,    218 
power  of,  alterations  in, 
by    cultivation    in 
immune        serum, 
228 
effect    of    heating    on, 

226 
spontaneous  alteration 

in,  227 

pro-agglutinoid    phenom- 
enon in,  explained 
as   protective   col- 
loid action,  236 
pro-agglutinoid    zone    in 

162 

pro-agglutinoids    in,    235 
relation        of        flagellar 
mechanism        t  o, 
222 

specificity    of,    219,    229 
diagnostic  value  of,  in 
group        reaction, 
231 

limitations  to,   229 
thread     reaction     of 
Pfaundler  in,  222. 
223 
"two    phase"    theory    of, 

241 

Agglutination          reaction, 
diagnostic  use  of, 
220,  221,  222 
flagellar    mechanism    in, 

222 

in    diagnosis    of    dysen- 
tery, 221 
of  glanders  in  horses, 

222 
of    paratyphoid    fever, 

221 

of  pneumonia,  221 
of     streptococcus     in- 
fections,  222 
of  typhoid  fever,  220, 

221 

nature  of,  239  et  seq. 
precipitin   reaction   anal- 
ogous to,  263   " 
with        dead        bacteria. 

223 
Agglutinins.  223 

absorption   of,   in   agglu- 
tination. 232 
complete,       impossibil- 
ity of.  233 

bacteriotropins  and,  321 
definition  of,  89 
group  : 

major,   229 

para-  or  minor   229 


566 


INDEX 


Agglutinins,  heating  of,  ef- 
fects of,  226 
explanation  of,  236 
"immune,"  234 
in  hemolytic  serum,  93 
nature  of,  224 
normal,  91,  233 

explanation  of,  234 
qualitatively    identical 
with        "immune" 
agglutinins,   234 
production  of,  129 
Agglutinogen,  223 

effects     of     heating     on, 

226 

localization    of,    in    cell 

body    proper,    225 

in     ectoplasmic     layer 

of  cells,  224 
in   flagellar   substance, 

224,  225 
nature  of,  224 
Agglutinoids,   235 
Aggressins,  326 
action  of,  21 
obtaining  of,  21 
secretion  of.  by  bacteria 
in   body,   and  vir- 
ulence, 20-22 
Albuminolysins,    193,    211, 

401 
Alexin,  137 

a    combination    of   soaps 

and   proteins.   175 

absence  of,  from  aqueous 

humor  of  the  eye, 

170 

analogy      between       fer- 
ments or  enzymes 
and,   176 
bactericidal     powers     of, 

137  / 

definition  of,  80   ^ 
dependence    of,    on    con- 
centration,  176 
extraction   of,    from   leu- 
kocytes   and    lym- 
phatic   organs, 
304    et   scq. 
filtration  of,  177 
in   hemolysis.   144 
inactivation   of,    bv   salt, 

178 

by    salt-removal,    178 
by  shaking,  185 
reversibility  of.   184 
Gramenitski's        ex- 
periments on,  184, 
185 

increase  of,  on  clot,  172 

influence     of     salts     on 

action      of,      177, 

178 

leukocytic  origin  of,  168, 

169  et  seq. 
multiplicity    of,    154    et 

sea. 
Borders      views      on. 

156 
Ehrlich's      views      on. 

155 

nature  of,   137.   ir,4 
chemical,    174.    175 
physical,   177 
presence     of.     in     blood 

plasma,  170-1TL' 
in   blood   stream,    170- 

172 

in  normal  blood,  Gen- 
gou's  view  of. 
303,  304 

MetchnikofTs      view 
of,   302 


Alexin,  presence  of.  in  nor- 
mal   blood,    other 
experimenters    on, 
303 
production    of.    in    liver, 

173 

in    thyroid   gland,    172 
varieties    of  : 

macrocytase,  169 
microcytase,  169 
Alexin  fixation,   186 
albuminolysin  in,  193 
identity   of,   with    pre- 

cipitins,  193 
writer's   opinion   on. 

193,  194 

Bordet  and  Gengou's  ex- 
periment on,  186. 
187 

Bordet-Gengou  phenome- 
non in,  188  ct 
stq. 

Gay's  experiments  sup- 
porting, 190 
in   diagnosis   of   infec- 
t  i  o  u  s      diseases. 
188 

in   diagnosis    of    syph- 
ilis. 188 

Moreschi's     e  x  p  e  r  i  - 
ments  supporting. 
189 
Bordetscher     Antikorper 

in,  194,  195 

by  immune  animal  sera 
with  their  specific 
antigens.  189 
by  protein  and  antipro- 
toin  sensitizers, 
188 

distinguished    from    com- 
plement        devia- 
tion, 186 
during  hemolysis,  nature 

of,  176 

Khrlich's          (schematic) 
conception  of.  187 
experiments  of,  on  syph- 
ilitic        monkeys, 
198.    199 

forensic  tests  in.  211 
in  ana  phyla  xis.  392 
in   determination   of   na- 
ture   of    unknown 
protein.  211 
delicacy    of.    212.    213 
technique  of.  212 
in  diagnosis  of  glanders. 

216 

in  diagnosis  of  gonor- 
rheal  infections. 
216 

in    diagnosis    of    malig- 
nant      neoplasms. 
213 
von  Dungern's  method 

of.  214 
antigen      production 

for.    214 
results  of.  215 
technique  of,  214  et 

seq. 

in  diagnosis  of  syphilis 
in  human  beings, 
199 

in  diagnosis'  of  tubercu- 
losis, 217 
in  Wassermann  reaction, 

198 

nature  of,   192  et  seq. 
non-specific.    195 

by     heated     normal 
serum,  196 


Alexin     fixation,      non-spe- 
cific,    by     lipoidal 
substances  in   tis- 
sues, 195 
by     preserved     normal 

serum,    196,    197 
by     protein     emulsion 
and       other       ex- 
tracts,  196 

by     unsensitized     bac- 
teria,  195 
of    specific    precipitates, 

190   et  scq. 

albuminolysin      identi- 
cal with.  193 
writer's  opinion   on, 

193,  194 
Gay  on,  190 
Pfeiffer  and  Fried  - 

berger  on,  190 
Sachs  on,  191 
writer  on.  193,  194 
practical  applications  of, 

198 
precipitin    reaction    and, 

190,   192 
Dean  on,   194 
Gengou   on.   192 
Neufeld    and     Ilaendel 

on,   194 

writer  on.  193,  194 
specific    antiprotein    sen- 

sitizers    in.    192 
with   syphilitic  serum  in 
antigens        from 
normal        organs, 
200 

Alexin  splitting.  178 
Brand  on.   179 
by    method    of    Ferrata, 

178.  179 
by  method  of  Liefmann, 

184 
by  method  of  Sachs  and 

Altmann,    184 
effect  of  acid  reaction  on 
fractious    of,    181 
end-piece     in,     179     et 

seq. 
mid-piece    in,     179     et 

scq. 

heat      sensitiveness      of 
fractions    of,    180 
interchange   of  fractions 
of.      in      different 
animals,  182 
is     it     the     inactivation 
of   the    mid-piece  'i 
183 

mid-piece  only  bound  in 
Wassermann  reac- 
tion. 181 

physical     occurrence     of 
fractions     of,     in 
blood.  181 
presensitizetl     cell     in, 

180 
properties     of     fractions 

of,  179 

quantitative         relations 
between    fractions 
of,   1S2,   183 
Alexocytes.  168 
Alkali-albuminate     precipi- 
tin, 260 

"Alkalinity       theory"       of 
immunity,    83,    84 
Amanita      phalloides,     spe- 
c  i  f  i  c      antitoxin 
from.  96 
Amboceptor,    149 

Bordet's  definition  o  f, 
159 


INDEX 


567 


Amboceptor,  complement- 
ophile  groups  or 
polyceptors  o  f 
(Ehrlich),  149, 
156 

cytophile  group  of,  149, 
152,  153 

multiplicity  of,  150,  151, 

154 

Ehrlich    and    Morgen- 
roth  on,   150,  151 

quantitative  determina- 
tion of,  in  im- 
mune serum,  160, 
161 

specificity  of,  150 
Amboceptor     and     comple- 
ment, Ehrlich  and 
Sachs's    views    on 
union  of,  164.  165 

Noguchi's  measurement 
of  quantitative 
relations  of,  164 

quantitative      ratio      be- 
tween,   163.    164 
Amebse,   artificial,   294 
Aiiaphylactic  antibody,   re- 
lation of.  to  other 
antibodies,  401 
Anaphylactic     intoxication, 
peptone   poisoning 
and,    405 

Anaphylactic  shock,  363  et 
seq.  See  also  Ana- 
phylaxis,  clinical 
manifestation  of 

effect  of  atropin  and 
other  drugs  on, 
379 

histamin  as  cause  of,  406 
Anaphylactin.  385 
Anaphylatoxin,      22,      394, 
415 

action  of  alexin   in.   425 
with     normal     or     in- 
activated   immune 
serum,  425,  426 
with  salt  solution,  426 

inhibition  of.  by  too 
vigorous  and  pro- 
longed reaction. 
419 

source  of,  427 
Anaphylaxis,    358 

alexin  fixation   in.   392 

analogy  of  immediate  re- 
action in  serum 
sickness  to,  429 

analogy  of  serum  sick- 
ness to,  430 

analogy  of  tuberculin  re- 
action to,  444 

anaphylactic  poisoning, 
nature  of,  404  et 
seq. 

from   precipitates.   394 
proteid    split    products 
of    Vaughan     and 
Wheeler  in,  405 
symptoms    of,    similar 
to      peptone      poi- 
soning, 406 

Anderson  and  Schultz's 
work  on.  365 

anti-anaphylactic  state 
in.  See  Antiana- 
phylaxis 

antigen  in,  intervals  be- 
tween administra- 
tions of,  376 
identity  of  sensitizing 
and  toxic  sub- 
stances of,  387 


Anaphylaxis,     antigen     i  n, 
identity    of   sensi- 
tizing   and     toxic 
substances     of, 
Doerr   and   liuss's 
work  on,  387 
Wells'  work  on,  387 
nature  of,  370 
path     of    introduction 

of,  374 

intracerebrally,    374 
i  n  t  r  a-intestinally, 

375 

by  feeding,  375 
by  rectum,  375 
into    large    intes- 
tine, 375 

intravenously,  374 
subcutaneously,    374 
quantity    of,    adminis- 
tered, 376  et  seq. 
specificity  of.  371 
degree  of,  372 
organ,   373 

a  u  tosensitization 
in,   373 
species,  372 
two       separate       sub- 

xtances  in,  386 
Doerr     and      Russ's 
experiments       on, 
387,  389 

Gay  and  Adler's  ex- 
periments in.   3S6 
Pick    and    Yamanou- 
chi's     experiments 
in.   387 

Arthus'    work   on,   361 
asthma  and.  436 
Auer    and    Lewis's    work 

on.   364 
autosensitization  in,  373, 

439 

Besredka  and  Stein- 
hardt's  work  on, 
374  ct  seq. 

Besredka's  theory  of,  386 
Besredka's  work  on,  375 

ct  seq. 

Biedl  and   Kraus's  work 

on.   365.   368,   3(59 

Bogomolez's      work      on, 

371 

Calvary's    work    on,    369 
cell  participation  in,  390, 

399  et  seq. 
clinical       manifestations 

of,  363 
in  dogs,  368 

fall    in    blood    pres- 
sure in,  368 
fall    of    temperature 

in,  369 
increase     of     lymph 

flow  in.  369 
intestinal      reaction 

in,  370 

lowered    coagulabili- 
ty of  blood  in.  36!) 
in  guinea  pig,  363 
alexin   reduction   in. 

367 

circulation        symp- 
toms in.  366 
effect      of      atropin 
and    other    drugs 
on,  365 
fall    of    temperature 

in,  366 

fever  in,  366,  367 
lowered     coagulabil- 
ity   of    blood    in, 
367 


Anaphylaxis,  clinical  mani- 
festations of,  in 
guinea  pig,  pulmo- 
nary emphysema 
in,  364 

respiratory        symp- 
toms in,  364,  365 
susceptibility          o  f 
various  breeds  in, 
368 

temporary      diminu- 
tion    of     polynu- 
clear       leukocytes 
in,   367 
in  rabbits,  368 

clinical  significance  of, 
428 

dependence  of,  on  pre- 
ceding inocula- 
tion, 360 

diagnostic  uses  of,   446 

diminution  of  alexin  af- 
ter     anaphylactic 
shock  in,  392 
significance  of,  392 

early  work  on,  359 

Flexner's  work  on,  359 

Friedberger's  work  on, 
366 

Friedemann's  e  x  p  e  r  i- 
ments  in,  in  vitro. 
393 

fundamental  principles 
of,  358 

Gay  and  Southard's  ar- 
guments against 
antigen  -  antibody 
reaction  theory 
of.  385 

Gay  and  Southard's  the- 
ory of,  386 

Gay  and  Southard's  work 
on,  364 

hay  fever  and,  436 

in  serum  therapy.  See 
Serum  sickness 

in  sudden  attacks  of  ca- 
tarrhal  nasophar- 
yngitis  and  con- 
junctivitis, 437 

in  vaccine  therapy,  434 

incubation  time  in,  360, 
376 

Jobling  and  Petersen's 
work  on,  396 

Lesne"  and  Dreyfus' 
work  on,  375,  376 

Magendie  on,  359 

Manwaring's  work  on, 
370 

mechanism  of  anti-ana- 
phylaxis  in,  402 
et  seq. 

desensitization  in,  403 
specificity  in,  404 
tolerance    to    anaphy- 
lactic   poison    in, 
403 

Nicolle's  theory  of,  391 

organ  specificity  in, 
438 

Otto's  work  on,  361 

Pearce  and  Eisenbrey's 
work  on,  368,  369 

Pfeiffer's  work  on,   366 

phenomena     related     to, 

406 

toxic    action    of    nor- 
mal sera,    406 
toxin  hypersusceptibil- 
ity,  409 

Pick  and  Yamanouchi's 
work  on,  371 


568 


INDEX 


Anaphylaxis.  quantitative 
methods  applied 
to  study  of,  389 

Ranzi's   work    on,    367 

relation  of  alexin  to, 
391 

relation  of  antibodies 
of,  to  other  anti- 
bodies, 401 

Richet  and  Hericourt's 
work  on,  360 

Richet  and  Portier's 
work  on,  360 

Rosenau  and  Anderson's 
work  on,  362,  376 
et  seq. 

sessile  receptors,  theory 
of,  399 

specificity   of,   362,   363 

sympathetic    ophthalmia 
and,  439 

Theobald  Smith  phenom- 
enon in,  361 

Theobald  Smith's  work 
on,  363 

toxic  action  of  normal 
sera  and,  406 

toxin  hypersusceptibility 
and.  409 

transference  of,  362,  380 
et  seq. 

true  antigen-antibody  re- 
action, 388 

tuberculin  ophthalmo-re- 
action  and,  442 

tuberculin  reaction  and, 
440.  See  also  Tu- 
berculin reaction 

tuberculin  skin  reaction 
and,  442 

Vaughan  and  Wheeler's 
theory  of  mecha- 
nism 'of,  390 

Vaughan    and    Wheeler's 
work      on      toxic 
fraction     of     pro- 
tein   molecule    in, 
.     390 

Vaughan's  work  on,  366, 
367 

Weichhardt     and     Schit- 
•    tenhelm's        work 
on.  369.  370 

with  bacterial  extracts, 
363 

with   normal  serum,   362 

with   proteins,   362 
Anaphylaxis.  bacterial.  412 

anaphylatoxin  formation 
in,  414.  See  also 
under  Anaphyla- 
toxin 

endotoxin  theory  of  pro- 
duction of.  414 

Friedberger's    e  x  p  e  r  i  - 
ments      in,      415, 
416,  417 
facts     deduced     from, 

417 

effect  of  excess  of 
bacteria  adminis- 
tered in,  418 
effect  of  excess  of 
sensitization  on 
anaphylatoxin  in, 
418 

effect  of  too  pro- 
longed exposure 
in  anaphylatoxin 
in,  418 

nature  of  bacterial  in- 
fections and,  420 
et  seq. 


Anaphylaxis,      bacterial, 

N  e  u  f  e  1  d       and 

Dold's     e  x  p  e  r  i  - 

ments  in,  419 

serum    anaphylaxis    and, 

413 

Vaughan's  theory  of  bac- 
terial   spitting   as 
cause  of.  414,  415 
Anaphylaxis,    passive,    380 

et  seq. 
Doerr    and    Russ's    work 

on,  381 

duration  of,  382 
Friedemann's    work    on, 

380.  381,  382 
Gay       and        Southard's 
work  on,  380.  381 
interval     between    injec- 
tion  of   sensitized 
serum    and    injec- 
tion of  antigen  in, 
382,   383 

methods     of     production 
of,    380,    381,    382 
nature  of  reaction  of.  382 
Nicolle's  work  on,  380 
Otto's  work  on,  380 
Richet's  work  on.   380 
Weill-Halie      and      Le- 
maire's    work    on, 
383 

Anderson      and      Schultz's 
work    on    anaphy- 
laxis, 365 
"Anergie,"  509 
Anthrax,   relative  suscepti- 
bility of  man  and 
animals  to,  53 
study    of,    in    regard    to 
resistance       and 
immunity.   296 
vaccination  against,   his- 
tory of.  64 
method  of.  64 
Anthrax     bacilli,     attenua- 
tion   of   virulence 
of,  18 

virulence  of.  15 
Anti-alexin.  157 
action   of.   157 
Anti-amboceptor,    152,    153 
Anti-anaphylaxis,   362,   377 
Besredka's  work  on,  375 

et   seq. 

mechanism  of.  402  et  seq. 

desensitization  in,  403 

tolerance    to    anaphy- 

lactic     poison    in, 

403 

methods     of     producing, 

377-378 

Besredka  and  Stein- 
hardt's  methods, 
377 

Rosenau     and     Ander- 
son's methods.  378 
specificity  of,  379.  404 
"Anti-antibodies."  147 
Antibodies.      concentration 
of.    in     lymphatic 
organs    in    active 
immunity.  101 
in  other  organs  in  ac- 
tive        immuniza- 
tion. 102 
in    active    immunization, 

85 

locality  of  production 
of,  dependent  on 
locality  of  anti- 
gen concentra- 
tion, 102 


Antibodies,    normal,    expla- 
nation of,  234 
origin  of,  100 
specificity  of,  85 
Antibody    formation,    body 

cell  in,  125 
chemical      nature     of. 

126 

chemical  action  of  anti- 
gens in,  128 

in  active  immunity,  re- 
moval of  spleen 
and,  101 

mechanism        of        (side 
chain  theory),  124 
by    internal    secretion 
of   body   cell,    125 
processes    of    metabol- 
ism and,  125 
overproduction  of  recep- 
tors in,  128 
principles  of,  94.  96 
Anticomplement,  157 

action  of,  157 
Anticytophile  interpreta- 
tion of  anti-sensi- 
tization  (Ehrlich 
and  Morgenroth), 
153 

controversy  on,  153 
"Antiformin."  70 
Antigen-antibody  reactions, 
129.        See      also 
Toxin  -  antitoxin 
reaction 

antibody     production    in 
body  cells  in,  130 
relationship          between 
susceptibility      of 
tissue    and    toxin- 
binding  properties 
in,   131 
side     chain     theory     in, 

129 

specificity   of,   98,   129 
variety  of  antibodies  in, 

129 

agglutinins,  129 
antitoxins,   129 
cytotoxins,    129 
precipitins,   129 
Antigenic      properties      of 
cells,    relation    of, 
to    lipoid    constit- 
uents, 97 

Antigens,  action  of,  95 
active          immunization 
with,  analogies  to 
drug        tolerance, 
99 

characteristics    of.    vari- 
ations in,  98,  99 
classification  of.  95 
complex      structure      of. 
Ehrlich   and   Mor- 
genroth's     concep- 
tion of.  151 
Definition  of.  35.  94 
"local"   immunity  in   or- 
gans   directly     in 
contact        with, 
102 

locality  of  production  of 
antibodies  depen- 
dent on  locality 
of  concentration 
of.  102 

organ  specificity  of,  99 
protein  nature  of,  97 
specificity  of,  98 
Anti-isolysins.    147 
,    "Antiricin."  85 
!    Antisensibilisin,  386 


INDEX 


569 


Antisensitization.  a  n  t  i  - 
cytophile  interpre- 
tation of  (Ehrlich 
and  Morgenroth ), 
153 

controversy   on.   153 
Antisensitizers.   152,   153 
non-specificity   of,    154 
Antitoxic  serum,  direct  ef- 
fect of,  on  toxin, 
104 

indirect     protective     ac- 
tion    of.     against 
toxin,  104 
"normal"  serum  of  Behr- 

ing,  107 

Antitoxin,     chemical     rela- 
tions      of.      with 
toxin,  114 
definition  of,  85,  86 
diphtheria.       See     Diph- 
theria antitoxin 
production  of.  129 

by  true  toxins,  35 
snake    venom,    466 

effect  of  heat  on,   105 
specific,     substances     in- 
citing, 86 

standardization    of,    465 
by    means   of  toxin, 

107 
guinea  pigs  used  in, 

108 
tetanus,     production    of, 

4G5 

use    of.    in    passive    im- 
munization. 86 
Antitoxin   unit,  diphtheria, 

107 

Antitoxinogen,   95 
Antivenin,  466 
Arrhenius   and   Madsen   on 
neutralization     in 
t  o  x  i  n-antitoxin 
reaction,    120 
Arthus,      phenomenon      of, 

380 
work  of.  on  anaphylaxis, 

361 

Ascoli  and   Izar's  work  on 
meiostagmin  reac- 
tion.  538,   539 
Asiatic      cholera,      relative 
susceptibility      of 
man   and  animals 
to.  53 
Asthma,    anaphylaxis    and, 

436 

Atrepsie.  56 
Attenuation  of  bacteria  by 

chemicals,  66 
by       cultivation       under 

pressure,  66 
by  drying,  65 
by  heating.  65 
by  passage  through  ani- 
mals.  65 

by  prolonged  cultivation 
above    optimum 
temperature.  65 
by   prolonged   growth   on 
artificial  media  in 
presence    of    own 
metabolic        prod- 
ucts. 65 
capsule      formation      in, 

18 

Auer  and  Lewis's  work  on 
anaphylaxis,  364 
Autocytotoxins,  93 
Autogenous   vaccines,   351 
"Autohemolvsins,"       146, 
147 


Auto-inoculation  by  mas- 
sage in  active 
immunization,  340 

Autolysins,  146 

Autosensitization  in  ana- 
.phylaxis,  373 

Auxilysin,  167 

Avian  tuberculosis,  relative 
susceptibility  of 
animals  to,  52 


Bacillus     botulinus,    action 

of,   4 

Bacteria,  adaptation  of,  in 
body,  6,  7 

agglutinability  of,   alter- 
ations  in,   226 
by  cultivation  in  im- 
mune serum,   228 
caused    by    heating, 

226 

normal       differences 
in,         between 
strains     of     same 
species,  228.  229 
spontaneous,  227 
in  agglutinoid,  229 

agglutination  in.  See 
under  Agglutina- 
tion 

"agglutinin"  bacteria, 
agglutination  of, 
243 

aggressin  secretion  of, 
in  body,  and  vir- 
ulence of,  20-22 

anti-opsonic  properties 
of,  and  antichem- 
o  t  a  c  t  i  c  sub- 
stances, 325 

»    artificial    cultivation    of, 
10 

attenuation  of,  18 

methods  of,  65.  See 
also  under  Atten- 
uation 

by     laboratory     ma- 
nipulations. 17 

capsulated.  agglutination 

of,  243 
virulence  of,  326 

capsule  formation  in, 
and  virulence,  18 

colloid  phenomena  and 
action  in,  560 

destruction     of.     by     cy- 
tases     in      leuko- 
cytes,     301.      302 
by  exudates.  300 
by  phagocytes,  300 

different  strains  of.  va- 
riation in  infec- 
tion from.  15.  16 

ectoplasmic  hypertrophy 
of.  in  relation  to 
virulence,  19,  20 

effect  of  body  tempera- 
ture on  invasive 
powers  of.  12 

effect  of  cultural  adap- 
tation of.  on  vir- 
ulence. 12 

effect   of    path    of   intro- 
duction     of.      on 
infection,    12-14 
on  virulence  of.   12-14 

effect  of  quantity  of,  in- 
troduced, on  in- 
fection. 14 

entrance  of.  into  body 
tissues,  6 

generalized  action  of,  24 


;  to  phagocy- 
of,  325 


Bacteria,  growth  of,  within 
leukocytes,  298 

in  blood  stream,  24 

in  localized  infection,  re- 
action  to,   26 
through    accidental 
conditions,  25 

in    normal    serum,    resis- 
tance 
tosis 

incubation  of,  26 

localized  action 
of,  23 

measurement  of  relative 
degrees  of  viru- 
lence of,  15 

negative  charge  of,  in 
suspension,  242 

number  of,  introduced, 
and  relative  viru- 
lence, 15 

occurrence   of,    2 

parasitic  and  saprophy- 
tic,  classification 
of,  11 

phagocytosis  of.  See 
under  Phagocyto- 
sis 

relative  vi.-ulence  of  dif- 
ferent strains  of 
same,  15,  16 

resistance  of,  to  phago- 
cytosis, due  to 
nonabsorption  of 
opsonin,  326 

resistance  of  living  cell 
to,  6 

secondary  abscesses 
caused  by,  25 

selective  action  of,  in 
localized  infec- 
tion, 25 

selective  lodgment  of,  in 
tissues,  40 

sensitized,  immunization 
with,  68 

sensitized  and  unsensi- 
tized,  influence  of 
salts  on  aggluti- 
nation of,  243, 
244 

similar  conditions  pro- 
duced by  differ- 
ent, 23 

specificity  of.  and  infec- 
tion, 22 

toxicity  of,  423 

use  of,  in  active  immu- 
nization, 85,  87- 
89 

variation  in  virulence  of, 
when  successively 
passed  through 
animals.  16,  17 
Bacterial  anaphylaxis.  See 
Anaphylaxis,  bac- 
terial 

Bacterial  extracts,  active 
i  m  m  u  n  i  z  ation 
with.  69 

extraction  of  bacteria 
for.  by  mechan- 
ical methods.  71 
by  permitting  them 
to  remain  in  fluid 
media.  70 

Bacterial  infections,  con- 
ceived as  reac- 
tion of  body 
against  a  foreign 
antigen,  421 

Bacterial  precipitins,  group 
reactions  in,  251 


570 


INDEX 


Bacterial  precipitins,  group 
reactions  in,  diag- 
nostic value  of, 
252 

partial  or  minor,  252 
Bacterial    products,    active 
i  m  m  u  n  i  z  ation 
with,  72 

Bactericidal    properties    of 
blood    serum,    134 
Bacterial  proteins,  33 
Bacterial  toxemia,  423 
Bacterial    toxins,    28.      See 

also  Toxins 

action  of,  after  distribu- 
tion in  body,  41 
active      immunization 

with,  72 

chemical     structure     of, 
in  relation  to  tox- 
icity,  43 
endotoxins.       Sec    Endo- 

toxins 

lesions  produced  by.  in 
course  of  excre- 
tion, 45 

local  injury  by,  45 
nature  of,  32 
nature    of    union    of,    to 

body  cells.  44 
nerves  attacked  by,  41 
obtaining  of,  32 

from  dead  cultures,  32 
from     living    cultures, 

32 

physical          relationship 
with  body  cells  in 
action   of,   44 
production    of    antitoxin 

by,  35 

selective  action  of,  40 
principles  of,  43 
reasons  for.  45 
selective   localization   of, 

39 

specific    distribution    af- 
ter      introduction 
of,  40 
specific   susceptibility   of 

tissues  to.  46 
chemicals      inhibiting, 

true.  33 

analogy    of,    with    en- 
zymes, 36 

bacteria  producing.  34 
characteristics  of.  34 
heat    sensitiveness    of, 

36 

incubation  time  of,  36 
Bactericidal    powers    of 
blood    serum,    79, 
134 

alexin  in,  137 
nature  of.  137 
in  vivo,  137 

cholera    experiments 

in.  137 

Bactericidal         substances. 

origin      of,      from 

leukocytes,  169  et 

seq. 

Bacteriolysins,    agglutinins 

and,  321 
in    active    immunization, 

89 
Bacteriolysis,   137 

extracelluar     theory     of, 

140 

heat  a  factor  in,  138 
mechanism   of.   138 
immunity    conferred    by, 
137,   138 


Bacteriolysis,     in     immune 

serum,  137,  138 
Bordet's     findings     in, 

140,  141 

inactivation  and  re- 
activation i  n  , 
140 

Pfeiffer's    phenomenon 
in,    technique    of, 
138    et   seq. 
specificity       o  f       pro- 
tection    of.      137, 
138 
transference   of  power 

of,   137,   138 
leukocyte       action       i  n, 

168 

leukocytes  in,  140 
Bacteriotropins,         bacteri- 
cidal      sensitizers 
and,  321  et  seq. 
normal      opsonins      and, 

320  et  seq. 

presence  of,  in  immune 
sera  without  ly- 
sins,  322 

specificity  of,  321 
thermostability   of,    320 
Bail's  aggressin  'theory,  21, 

67 

classification  of  para- 
sites, 11 

Bauer's  modification  of 
Wassermann  test, 
209 

Baumgarten's  osmotic  the- 
ory of  bactericidal 
powers  of  blood, 
135 

Behring,  Kitasato  and 
Wernicke.  a  n  t  i  - 
body  theory  in  ac- 
tive immunity  of, 
84 

Besredka  and  Steinhardt's 
work  on  anaphy- 
1  a  x  i  s,  374  et 
seq. 

on  serum  therapy  of  ty- 
phoid fever,  478 
Besredka's  anti-endotoxic 
serum  in  treat- 
ment of  typhoid 
fever,  478 

method     of     administra- 
tion  of  antitoxin, 
433 
theory     of     anaphylaxls, 

386 

vaccines   in   prophylactic 

i  m  m  u  n  i  z  a  tion 

against        plague, 

489 

work    on    antianaphylax- 

is,  375  et  seq. 
Biedl  and  Kraus's  work  on 
anaphylaxis,    365, 
368,   369 
Blood,  non-putrefaction  of, 

79.   134,  256 
phagocytic   activities   of, 

in  immunity.  79 
Blood    plasma,    cell-free 
inhibition   of  bac- 
terial   growth    in, 
in  immunity,  79 
presence     of    alexin     in, 

170-172 

Blood  serum,  agglutina- 
tion in,  immune, 
141 

alexin  in,  137 
nature  of,  137 


Blood  serum,  anti- 
bacterial powers 
of,  in  immunity, 
79,  134 

anti-isolysins  in,   147 
autohemolysins    in,    146. 

147 

bactericidal    and    agglu- 
tinating      powers 
of,  Wright's  stud- 
ies of,  328 
bactericidal      action      of 

immune,  81 
alexin  in,   137 

early      theories      re- 
garding, 134 
in    natural    immuni- 
ty, 79,  80 
in  viro,  137 

cholera  o  x  p  e  r  i  - 
ments  in,   137 
mechanism  of,  135 
assimilation  theory  of, 

136 

by     chemically     un- 
favorable environ- 
ment. 135 
osmotic     theory     of, 

135 
bacteriolysis  in   immune, 

137 
Bordet's     findings     in, 

140,   141 
cholera        experiments 

in,  137 
summary     of     facts 

in,  138 
heat  a  factor  in.  138 

mechanism    of,    13H 
immunity        conferred 

by,  137,  138 
inactivation  and  reac- 
tivation in,  140 
intracellular        theory 

of,  138 

leukocytes  in,  140 
Pfeiffer's    phenomenon 
in.    technique    of, 
138    et    seq. 
specificity     of     protec- 
tion   of,    137,    13S 
bacteriolytic    powers    of 
transferable,     137 
cell-free,       bacterial 

growth  in.  81 
hemolysinogens    in,     148 
hemolysis      in      immune, 

141 
alexin    or    complement 

in.  144 

analogy  of.   to  bacter- 
iolysis, 142 
Bordet's  work  on,  1.41 
Ehrlich    and    Morgen- 
roth    on    mechan- 
ism of,  142 
haptophore  groups   in, 

142 

relation  of  antigen, 
amboceptor  and 
complement  in, 
143-145 

work  of  Ehrlich  and 
Morgenroth  o  n, 
143-144 

work     of     Liefmann 
and  Cohn  on.  145 
isohemolysins  in,    146 
protective       action       of, 
against     bacteria, 
50 

Blood   stream,   presence  of 
alexin  in,  170,  172 


INDEX 


571 


Body  fluids,  bactericidal 
powers  of,  in 
natural  immunity, 
80 

Bogomolez's  work  on  ana- 
phylaxis,  371 

Borden's  method  of  agglu- 
tination, 223 

Bordet,  explanation  of,  on 
agglutination.  240 
findings  of,  on  bacterio- 
lytic  power  of 
immune  serum, 
140,  141 

o  n  neutralization  i  n 
t  o  x  i  n-antitoxin 
reaction,  122 
views  of,  concerning  re- 
lation of  antigen, 
ainboceptor  and 
complement,  158, 
159 

concerning  relation  of 
antigen,  ambocep- 
tor  and  comple- 
ment, action  of 
complex  in,  159 
schematic  representa- 
tion, 159 

Bordet-Danysz  phenomenon 
in  neutralization 
in  toxin  -  a'n  t  i  - 
toxin  reaction, 
123 

Bordet  and  Gengou's  ex- 
periment on  alex- 
in  fixation,  186, 
187 

Bordet-Gengou  phenomenon 
in  alexin  fixation, 
188  et  seq. 

Gay's    experiments    sup- 
porting,   190 

Moreschi's      experiments 
supporting.    189 

Botujinus  poisoning,  ac- 
tion of.  41 

Botulinus  toxin.  4 

Bouillon  Filtre  (Denys), 
357 

Bovine  colloid.   167 

Bovine  tuberculosis,  rela- 
tive susceptibility 
of  man  and  ani- 
mals to.  52 

Brownian  movement  in 
colloids,  549 

Buchner  on  bactericidal 
power  of  blood  in 
natural  immunity, 
80 


Calmette's       investigations 
in   snake    poisons, 
46(5  ct  seq. 
ophthalmo-reaction.    442 

Capsule   formation   in    bac- 
teria,  by  attenua- 
tion.   18 
virulence    and,    18 

Calvary's  work  on  anaphy- 
laxi's,  369 

Carriers,    bacillus,    2 

Castellani,  absorption  ex- 
periments of,  in 
agglutination.  232 

Catarrhal  nasopharyngitis 
and  conjunctivi- 
tis, sudden  at- 
tacks of.  anaphy- 
lactic  nature  of, 
437 


Cell  receptors,  overproduc- 
tion of,  152 

structure  of,   127 
Cellular    theory    of    immu- 
nity, 136 

Cercbrospinal  meningitis, 
epidemic,  serum 
therapy  in,  471 

early  investigations  in, 
471 

Flexner  and  Jobling's 
work  on,  472 

Jochmann's  investiga- 
tions in,  471,  472 

Kolle  and  Wassermann's 
investigations  in, 
471 

nature  of  action  in,  473 

results    of.    472,    473 

standardization  of  serum 

in.   473 

Chantemesse's      early      ex- 
periments   in    se- 
rum    therapy     of 
typhoid  fever,  477 
Chemptaxis,    285 

anaphylatoxiu  and,  291 

Eugelmaun's  studies  in, 
287 

influence  of  bacteria  in. 
288 

influence  of  bacterial 
extracts  in,  288, 
289 

malic    acid    in,    286 

of  slime-molds  or  inyxo- 
mycetes,  286 

of  spermatozoa  of  ferns, 
286 

Pfeffer's  technique  in, 
286 

physical  explanations  of, 
292  ct  seq. 

selective,    291,    295 

surface  tension  in.  293 
"artificial  aruebae" 

and,   294 

Chicken  cholera,  vaccina- 
tion against,  his- 
tory of,  63 

Cholera,   active    prophylac- 
tic   immunization 
against,  487 
Ferran's  earlv  investi- 
gations  in,   487 
Haffkine's    method    in, 

487,  488 

Kolle's  method  in,  489 
Strong's     method     i  n, 
489 

Asiatic,  relative  suscep- 
tibility of  man 
and  animals  to, 
53 

chicken,  vaccination 
against,  history 
of,  63 

effect  of  path  of  intro- 
duction of  bac- 
teria of,  on  infec- 
tion, 14 

experiments  in,  showing 
bactericidal  pow- 
ers of  blood  se- 
rum in,  137 

hog,  immunization  with 
bacterial  products 
in,  72 

Cobra  antitoxin,  action  of, 
467 

standardization  of,  468 
Cobra  lecithid,  175 
Cobra  venom,  action  of,  467 


Coctoprecipitin,  258 
"Coefficient  of  extinction," 

332 

Cole's      work      on      serum 
therapy    of    pneu- 
monia, 477 
Colics'  law,  499 
Colloid  gold  reaction,  539 
Colloids,  543 

application     of    phenom- 
ena   of,     in    elec- 
trical field,  562 
to     action     in     animal 

body,  560 
to  action   in    bacteria, 

560,  561 
to  action  with  eggs  of 

Fundulus,   561 
to  biology,  559  et  seq. 
to    Danysz    toxin-anti- 
toxin       phenome- 
non, 561 

to  nonagglutination  in 
excess     of     agglu- 
tinin,  562 
chemical     properties     of, 

552,  553 
chemical      composition 

in,    552 
chemical    reactions   in, 

552 
electrochemical  ioniza- 

tion  in,  553 
classification  of,  544 
definition  of,  543 
emulsion,   544,   554 
flocculation    of,    by    elec- 
trolytes, 553 
acids   and   alkalies   in, 

553 

concentration    of    elec- 
trolyte in.  553 
explanation  of,  554 
nature  of  sol  in,  554 
precipitin  reaction  an- 
alogous to,  265 
salts   in,   553 
zone    phenomenon    in, 

555 

gel,   544 

Graham's  work  on,  543 
irreversible.  544 
lyophobic,   559 
lyophylic,  559 
mutual  reactions  of.  555 
in  two  oppositely  elec- 
trical  sols,   556 
explanation  of,  557 
in   two   similarly   elec- 
trical  sols,    555 
protective     action     of 
electrolyte  in,  556 
protective     action     of 
great     excess     of 
one     colloid     over 
the  other,  556 
explanation   of,    550 
nature  of,  544 
physical     properties     of, 

545  et  seq. 
Brownian       movement 

of  particles  in, 
549 

distribution  of  parti- 
cles in,  549,  r,r,o 

electrical  properties 
in,  550 

form  of  particles  in, 
546 

kinetic  energy  in.  549 

size  of  particles  in, 
measurement  of, 

546  et   seq. 


572 


INDEX 


Colloids,     physical    proper- 
ties    of,     size     of 
particles     in,    mi- 
croscopic,   547 
osmotic  pressure  in, 

547 
rate    of    settlement 

in,  548 

ultrafiltration  meth- 
od  in,  548 
surface  tension  in,  550 

et  seq. 
preparation   of  solutions 

of,  558,  559 

reaction  in,  analogous  to 
complement  devi- 
ation phenomenon 
of  Neisser-Wechs- 
berg,  162 
inhibition  zones  in, 

162 

reversible,  544 
sol,  544 

stability  of,  545 
suspension,    544,   554 
Complement.         See      also 

Alexin 

amboceptpr  and,  Nogu- 
chi's  measurement 
of  quantitative 
relations  of, 
164 

quantitative    ratio    be- 
tween, 163,  164 
union   of,   Ehrlich   and 
Sachs's   views   on, 
164,  165 
definition  of,  144 

in  hemolysis,  144 
multiplicity    of,    154    et 

sea. 
Borders      views      on, 

156 
Ehrlich's      views      on, 

155 
nature  of,  154 

chemical,  174 
Complement  deviation,  160 

ct  seq. 

argument  in  favor  of 
Bordet's  views, 
162,  163 

colloid    reactions,    analo- 
gous to,  162 
Gay's      explanation      of, 

163 
in    hemolytic     reactions, 

163 

Morgenroth  and  Sachs's 
experiments  sup- 
porting. 163 

pro-agglutinoid  zone  re- 
action analogous 
to,  162 

Complement  fixation,  186. 
See  also  Alexin 
fixation 

in  determination   of   na- 
ture   of    unknown 
protein,  211 
delicacy    of,    212, 

213 

technique  of,  212 
practical  applications  of, 

198 
test  of,   in   diagnosis   of 

glanders,  216 
in  diagnosis  of  gonor- 
rheal      infections, 
216 

in  diagnosis  of  malig- 
nant neoplasms, 
213 


Complement  fixation,  test 
of,  in  diagnosis  of 
malignant  n  e  o  - 
plasms,  von  Dun- 
gern's  method  of, 
214 

antigen       produc- 
tion  for,   214 
results  of,  215 
technique   of,   214 

et  seq. 

in  diagnosis   of  syphi- 
lis, 512 

in  diagnosis  of  tuber- 
culosis. 217 

Complement  splitting,  178. 
See  also  under 
Alexin,  splitting 
of 

Complementoid,   158 

Complementophile  group 
of  amboceptor, 
149 

Conglutinin,   167 

Conjunctiva,  susceptibility 
of,  to  infection, 
13 

Corpus  luteum  cytotoxin, 
92 

"Cryptogenic     tetanus,"     5 

Cultivation  of  bacteria,  ar- 
tificial, 10 

Cytases,  in  phagocytes,  de- 
struction of  bac- 
t  e  r  i  a  by,  301, 
302 

Cytolysins,  92 

Cytolysis,   134 

Cytolytic    substances,    ori- 

fin   of,    from    leu- 
ocytes,      169      et 
seq. 

Cytophile  group  of  ambo- 
ceptor, 149,  152, 
153 

Cytotoxins,  92 
specificity  of,  92 


Danysz  effect  in  neutral- 
ization in  toxin- 
antitoxin  reac- 
tion, 123 

Danysz  toxin-antitoxin 
phenomenon,  ap- 
plication of  col- 
loid phenomena 
to,  561 

D  a  p  h  n  i  a.    Metchnikoff's 
study  of,  274.  2JX5 
phagocytosis    in,    296 

Dean's  antiplague  sera, 
482 

Denys  and  Leclef  on  im- 
portance of  phag- 
ocytosis in  im- 
munity. 311,  312 

Diphtheria,  active  immu- 
nization in,  with 
toxin  -  antitoxin, 
460 

relative  susceptibility  of 
man  and  animals 
to,  53 

Diphtheria  antitoxic  ser- 
um, normal,  107 

Diphtheria    antitoxin,     448 
antitoxin    production    in, 

457 

concentration  of,  459 
presence  of,   in   blood  of 
normal        individ- 
uals, 450 


Diphtheria  antitoxin,  pre- 
servation of,  108 

speed  in  administration 
of,  449 

speed    in    absorption    of, 

on    intravenous    injec- 
tion,  451,   452 
on     subcutaneous     in- 
jection, 451,  452 
speed    of    diagnosis    for, 
necessity    of,    4,>.> 
stability  of,  108 
standardization  of,   457 
by     means     of     toxin, 

107 

early  attempts  in,  107 
statistics  showing  reduc- 
tion   of   mortality 
with,  448 

toxin  production  for,  454 
choice    of    culture    in, 

454 
cultivation     of    strain 

in,   454.  455 
culture      medium      in, 

455 
"maturing"     of     toxin 

in,  455 
testing     of     toxin     in, 

456 
Theobald     Smith's 

method  of,  456 
unit  of,  107 
Diphtheria   bacillus,  action 

of,  4,  5 
Diphtheria     toxin,     action 

of,   41 

construction  of,  118 
determination     of    diph- 
theria     immunity 
with.  464 
normal,   107 
stability  of,   108 
unit  of,  107 

Diphtheria  toxin-antitoxin, 
neutral  mixtures 
of,  460 

Behring's  method  of  im- 
munization with, 
4(50 

advantages  of,  461 
chief  value  of,  461 
danger     of     anaphylaxis 

in,  461 

determination  of  pres- 
ence of  free  tox- 
in or  antitoxin  in 
convalescents  fol- 
lowing treatment 
with.  464 
human  susceptibility  to, 

461 

production  of  homolo- 
gous antitoxin  in 
h  u  in  a  n  beings 
with,  for  passive 
i  m  m  u  n  i  zation, 
462 
results  of  treatment 

with,  462 

standardization    of    anti- 
toxin,   462,    463 
limes-necrosis  of  toxin 

in,  463 
Komer's      method     of, 

463 

toxic  action  of,  460 
Doerr  and  Kuss's  experi- 
ments on  two  sep- 
arate substances 
in  anaphylactic 
antigen,  389 


INDEX 


573 


Doerr  and  Russ's  work  on 
passive  anaphy- 
laxis,  381 

Dochez  and  Gillespie's 
work  on  serum 
therapy  in  pneu- 
monia, 477 

Drug  tolerance,  analogy 
between,  and  ac- 
tive immunization 
with  antigens,  99 

Dunbar's  work  on  hay 
fever,  430 

von  Dungern's  method  of 
alexin  fixation  in 
diagnosis  of  ma- 
lignant tumors, 
214 
antigen  production  in, 

214 

results  of,  215 
technique    of,    214    et 

seq. 

view   of   precipitin   re- 
actions,   268,    271 

"Dust  cells"  of  the  lungs 
in  phagocytosis, 
279 

Dysentery,  agglutination 
reaction  in  diag- 
nosis of,  221 


Ehrlich,  conception  of 
alexin  fixation 
(schematic)  o  f, 
187 

of  relation  of  antigen, 
amboceptor  and 
complement,  149, 
150 

interpretation    of   agglu- 
tination  by,   234 
diagrammatic       repre- 
sentation of,  235 
on  multiplicity  of  alexin 
or        complement, 
155 

side  chain  theory  of,  in 
toxin  -  antitoxin 
reaction,  124 

Ehrlich's.  analysis,  104 

Ehrlich's  "antiricin,"   85 

Ehrlich  and  Morgenroth  on 
multiplicity  o  f 
amboceptor,  e  x- 
ample  of,  150, 
151 

Ehrlich-Sachs  phenomenon 
in  sensitization, 
165 

Bordet  and  Gay's  inter- 
pretation of,  166, 
167 

Eisenberg  on  residue  an- 
tigen and  anti- 
body in  precipitin, 
reaction,  268 

Eisenberg  and  Volk's  in- 
terpretation o  f 
agglutination.  235 

Endocarditis,  malignant,  24 

Endolysins,  305 

Endotoxins,  33.  34.  39 
characteristics    of,    37 
recent  work  on,  423 
toxic    cleavage    products 
of.  38 

Engclmann's  studies  in 
chomotaxis,  287 

Enzymes,  analogy  of,  with 
true  bacterial 
toxins,  36 


Enzymes,    in    phagocytosis, 
endocellular    and 
extracellular,    305 
leukocytic,  524 
serum,  523 

Epiphanin  reaction,  537 
Epithelioid  cells,  action  of, 
in       phagocytosis, 
284 
Epitoxoids,     definition     of, 

112 

Epitoxonoids,  124 
Erythrocyte  laking,  91 
"Exhaustion      theory"      of 

immunity,  83 
Exotoxins,     33.       See   also 

Toxins,  true 
bacteria  producing,  34 
characteristics  of,  34 
chemically        indefinable 
nature  of,  35 


Fermentation,        infectious 

disease  and,   1 
micro-organisms        caus- 
ing, 1 

Ferments,      protective,      in 

animal    body,    523 

Abderhalden's    e  x  p  e  r  i- 

ments    with,    532, 

533,  534 

Ferran's  investigations  in 
active  prophylac- 
tic immunization 
against  cholera, 
487 

Ferrata,  experiments  of, 
i  n  complement 
splitting,  with 
salt  solution, 
179 

Ficker's  reaction  in  agglu- 
tination, 223 

Flexner's    observations    on 

anaphylaxis,  359 
on     serum     therapy     in 
cerebros  pinal 
meningitis,  472 

Flexner  and  A  moss'  work 
on  poliomyelitis, 
497,  498 

Flexner  and  Lewis'  work 
on  poliomyelitis, 
495 

Forensic  alexin  fixation 
tests,  211 

Forensic   determination    of 
unknown    pro- 
teins, 211 
delicacy  of,  213 
technique  of,  212 

Fornet  and  Miiller,  ring 
test  of,  for  pre- 
cipitin blood 
tests,  257 

Friedberger,  experiments 
of,  in  bacterial 
anaphylaxis,  415 
on  the  nature  of  bac- 
terial infections, 
421  et  seq. 

work  of,  on  anaphylaxis, 
366 

Friedemann,       experiments 
of,  on  anaphylaxis 
in  vitro.  393 
work  of,  on  passive  ana- 
phylaxis, 380,  382 
in  rabbits,  398 

Fundulus,  Loeb's  experi- 
ments with  eggs 
of,  561 


Garbat  and  Meyer's  work 
OQ  serum  therapy 
of  typhoid  fever, 
479 

Gastric  juice,  action  of,  on 
stomach  itself,  6 

Gastrotoxin,   92 

Gay  and  Adler's  experi- 
ments on  two 
separate  sub- 
.  stances  in  ana- 
phylactic  antigen, 
386 

Gay  and  Southard's  objec- 
tions to  antigen- 
antibody  theory 
o  f  anaphylaxis, 
385 
theory  of  anaphylaxis, 

385 
work      on      anaphylaxis, 

364 

on      passive      anaphy- 
laxis, 380 

Gay's  sensitized  killed  vac- 
cines in  prophy- 
lactic typhoid 
fever  immuniza- 
tion, 486 

Gay's  work  on  injection  of 
non-specific  sub- 
stances in  infec- 
tious diseases,  521 

Gels,  544 

Giant     cells     in     phagocy- 
tosis, 280 
foreign  body,  280 
tuberculous,  280 

Glanders,  alexin  fixation 
test  in  diagnosis 
of,  216 

in  horses,  agglutination 
reaction  in  diag- 
nosis of,  222 
relative  susceptibility  of 
man  and  animals 
to,  53 

Gold  reaction,  colloidal, 
539 

Gonoccoccus,  relative  sus- 
ceptibility of  man 
and  animals  to, 
53 

Gonorrheal  infections, 

alexin  fixation 
test  in  diagnosis 
of,  216 

Gottstein-Mathes'  work  on 
serum  therapy  of 
typhoid  fever, 
479 

Graham's  work  on  colloids, 
543 

Gramenitski's  experiment 
in  reversal  of 
alexin  inactiva- 
tion  by  heating. 
184,  185 

Grohman  on  inhibition  of 
bacterial  growth 
by  cell-free  blood 
plasma  in  immu- 
nity. 79 

Gruber-Widal  reaction  in 
diagnosis,  220 


Haemotoxins,  21 

Haffkine's  early  work  on 
prophylactic  im- 
m  u  n  i  z  a  t  i  o  n 
against  plague, 
489 


574 


INDEX 


Haffkine's  method  in  active 
prophylactic  im- 
munization 
against  cholera, 
488 

Half  parasites,  11 
Haptines,  129 

of  the  second  order,  235 
of   the   third   order,    150 
varieties  of,  129 
Haptophore  group  in  toxin, 

110 

action  of,  110 
Haptophore    groups    in 

hemolysis,   142 
Hay    fever,    anaphylaxis 

and.   436 

Dunbar's  study  of,   43G 
reaction  in,  436 

anaphylactic        nature 

of,  437 

toxic  nature  of.  437 
treatment  of,  437 
Heat   alkali-precipitin.   259 
Heat-precipitins.  259 
Hemagglutination,       agglu- 
tination    of     bac- 
teria     by      serum 
analogous  to,  23(5, 
237 

Hemoglobinuria,  paroxys- 
mal, 147 

autohemolysins  in,  147 
hemolysis  in,  147 
Hemolysinogens,        human. 

148 

nature  of,  148 
Ilemolysins,     anti-isolysins 

in,  147 

autolysins  in,  146,  147 
isohemolysins  in,  146 
specific,  definition  of.  92 
specific    inciting    of,    in 

animal.  91 

Hemolysis,    alexin   or   com- 
plement in,  144 
amboceptor      in.      action 

of.  149  et  xeq. 
anti-amboceptor   in,   152, 

153 

anti-cytophile      interpre- 
tation     of      anti- 
sensitization        in 
(of    Ehrlich     and 
Morgenroth).    153 
controversy   on.   153 
antisensitizer      in,      152, 

153 

anti-isolysins  in.  147 
experiments  of  Liefmann 
and  Colin  on.  145 
in  immune  serum.  141 
analogy  of.  to  bacteri- 
olysis.  142 

Bordet's  work  on.  141 
Ehrlich     and    Morgen- 
roth   on    mechan- 
ism of,   142 
haptophore  groups  in, 

142 

relation  of  antigen, 
amboceptor  and 
complement  i  n, 
143-145 

multiplicity  of  ambocep- 
tor in,  150.  151, 
154 

recapitulation  of  views 
of  Ehrlich  and 
Morgenroth  on, 
152 

Hemolytic  properties  of 
normal  serum,  91 


Hemolytic  reactions,  com- 
plement deviation 
in,  163 

Hemolytic     serum,     agglu- 

tinins  in,  93 

Ehrlich  and  Morgen- 
roth's  conception 
of  neutralization 
of,  by  antilysin 
or  anti-ambocep- 
tor  reacting  with 
cytophile  group, 
153 
precipitins  in,  94 

Hemolytic  substances,  ori- 
gin of,  from  leu- 
kocytes, 169  et 
seq. 

Hepatotoxin,  92 

Hiss,  investigations  of.  on 
therapeutic  use  of 
leukocyt  e  ex- 
tracts. 308,  309 

Hiss  and  Zinsser's  work  on 
injection  of  non- 
specific substances 
in  infectious  dis- 
eases. 521 

Hog  cholera,  immunization 
*v  i  t  h  bacterial 
products  in,  72 

Ilogyes  method  of  treat- 
ment in  rabies, 
494 

Holobut's  work  on  bac- 
terial anaphy- 
laxis, 413 

Hopkins'  method  of  stand- 
ardization of  vac- 
cines, 353 

"Horror   autotoxicus,"    147 

Human  isolysins.  See  Iso- 
lysins,  human 

Humoral  theory  of  immu- 
nity, 136 

Hydrophobia,  active  .pro- 
phylactic immu- 
nization against, 
492.  See  also  un- 
der Rabies 

Hyperleukocytosis,  specific. 
309 

Hypersusceptibility.         See 

Anaphylaxis 

toxin,    and    anaphylaxis, 
409 

"Ideal     fall."     in    syphilis, 

503 
Immune    bodies,    structure 

of.  127 
Immune    serum.      See   also 

Serum,    immune 
agglutination  in.  141 
bacteriolytic     power     of 

transferable.   137 
bacteriolytic      properties 
of,    Bordet's    find- 
ings in.  140,  141 
direct    neutralization    of 
toxin    -    antitoxin 
reaction,  the  pro- 
tective   power    of, 
124 
hemolysis  in.  141 

alexin    or    complement 

in.  144 

analogy     of.     to     bac- 
teriolysis, 142 
Bordet's  work  on.  141 
Ehrlich     and    Morgen- 
roth   on    mechan- 
ism of,  142 


Immune  scrum,  hemolysis 
i  n  .  haptophore 
groups  in,  142 
relation  of  antigen, 
amboceptor  and 
complement  i  n, 
143-145 

work    of    Ehrlich    and 
Morgenroth       o  u, 
143-144 
work  of  Liefmann  and 

Cohn  on,  145 
phagocytosis  in,  90 
specific,  agglutination  of 

bacteria  in.  89 
precipitin       formation 

in,  90 

Immune  isolysins,  148 
Immunitats  Einheit,  107 
Immunity,  49 
acquired,  60 
artificially,  63 
definition  of,  63 
history  of,  61 
increased  phagocytosis 

and.   299 

active,    relation    of    pha- 
gocytosis   to.    32!) 
cellular  theory  of,  130 
definition  of,  3 
diphtheria,        determina- 
tion     of,      wit  h 
diphtheria     toxin, 
464 

"high  tide"  of,  340 
humoral  theory  of,   136 
lasting,        diseases        in 
which   one   attack 
conveys,  60 

diseases  in  which  one 
attack  does  not 
convey,  61 

local,  in  organs  directly 
in     contact     with 
antigens.  102 
in  skin  infections,  103 
natural,  49 

cellular   theory   of.   80 
definition  of.  50,  62 
humoral      theory      of, 

80 

inflammation  in,  78 
mechanism  of,  78 
theories    concerning, 

78-82 

bacterial    destruc- 
tion   by    phago- 
cytic  cells,   78 
bacterial     growth 
in      cell   -    free 
blood  serum,  81 
bactericidal     pow- 
er   of   blood    in 
natural     immu- 
nity, 80 

bactericidal  pow- 
er of  normal 
blood  in  nat- 
ural immunity. 
79 

bactericidal  prop- 
erties of  extra- 
vascular  plas- 
ma or  serum, 
81 

inhibition  of  bac- 
terial growth 
by  c  e  1 1  -  f  r  e  e 
blood  plasma, 
79 

intracellular  de- 
struction o  f 
bacteria,  78 


INDEX 


575 


Immunity,  natural,  mecha- 
nism of,  theories 
concerning,  phag- 
ocytic  activities 
of  blood,  79 
phagocytic  activities 

of  blood  in,  79 
principles  of,  50 

body        temperature 

in,   51 

cultural  conditions 
for  bacteria  in 
body  a  factor  in, 
56 

increased       invasive 
powers      of      bac- 
teria in,  56 
individual         differ- 
ences in,  58 
inheritance  in,  56 
racial  differences  in, 

55 

relative       resistance 

of  animals  in,  51 

species        resistance 

in,    51 
Pfeiffer    phenomenon    in, 

138 

phagocytosis  in,  90 
Immunization,   60 

against     snake     venoms, 

466 

history  of. 

ImmunizatioK.  active. 
also  Vaccine 
apy 

against  anthrax,  64 
against   chicken    cholera, 

63 

against  small-pox,  62 
agglutination  of  bacteria 

in.  89 
"alkalinity      theory"    of, 

83.  84 

antibacterial.  85,  87 
antibodies  in,  85 

bodies    in.     fundamen- 
tal    principles    of 
theory     of,     84 
origin  of.  100 
antitoxic,  85 
as    a    therapeutic    meas- 
ure,  action   of.   in 
generalized       sys- 
temic     infections, 
347 
in    local    infections, 

346 
in     successive    local 

infections,  347 
value  of.  346 

in     acute     diseases, 

350 

in    s  u  b  a  c  u  t  e    or 

chronic  cases.  349 

as  prophylactic  measure. 

value  of.  345,  346 

autogenous    vaccines    in, 

351 

auto-inoculation  by  mas- 
sage in.  340 
bacteria  used  in,  85,  87- 

89 

bacteriolysins  in.  89 
by   means   of   living   but 
attenuated         cul- 
tures. 65 

concentration  of  anti- 
bodies in  lym- 
phatic organs  in, 
101 

in     other     organs     in. 
102 


Immunization,   active,   defi- 
nition of,  63 

"exhaustion  theory"  of, 
«3 

"high  tide"  of  immunity 
in,  340 

in  diphtheria,  with  tox- 
in-antitoxin mix- 
tures, 460.  See 
also  Diphtheria 
toxin-antitoxin 

invasion  of  bacteria  in, 
mechanism  of  re- 
action in  tissue 
cells  against,  103 

locality  of  production  of 
antibodies  depen- 
dent on  locality 
of  antigen  con- 
centration in, 
102 

negative  phase  in,  338 
second       injection    in, 

338 
successive  inoculations 

in,  338,   339 
summation       of,       338 
339 

non-bacterial  anti-toxin- 
stimulating  sub- 
stances in,  86,  87 

"osmotic    theory"   of,    84 

phagocytosis  in,  90 

phenomena  following.  82 

precipitin  formation  in, 
90 

of  tissue  cells 
to  invasion  in, 
103 

removal  of  spleen  in. 
and  antibody-for- 
mation. 101 

"retention  theory"  of.  83 

second  positive  phase 
in,  339 

specificity  of  antibodies 
in.  85 

summation  of  positive 
phase  in.  339 

tuberculins  in,  355 

vaccines  in.   351 
production  of,  351 
sensitized.  355 
with    dead    bacteria, 

351 
with  living  bacteria, 

351 
standardization         o  f, 

353 
Hopkins'  method  of, 

353 

Wright's  method  of, 
352,   353 

with  antigens,  analogy 
between  drug  tol- 
erance and.  99 

with    bacterial    extracts, 

69 

extraction  of  bacteria 
for.  by  mechan- 
ical methods.  71 
by  permitting  them 
to  remain  in  fluid 
media.  70 

with  bacterial  products, 
72 

with  dead  bacteria.  68 
methods   used    in    kill- 
ing   bacteria    for, 
68 

with  fully  virulent  cul- 
tures in  sublethal 
amounts.  66-68 


Immunization,  active,  with 
.sensitized  b  a  c  - 
teria,  68 

Immunization,  active  pro- 
phylactic, in  man, 
483 

against  cholera,  487.  See 
also  under  Cholera 

against  paratyphoid 
fever,  484 

against  plague,  487.  See 
also  under  Plague 

against  poliomyelitis, 
495 

against  rabies,  492.  See 
also  under  Rabies 

against  small-pox,  490. 
See  also  under 
Small-pox 

against  syphilis,  498 

against     typhoid     fever, 
484.         See      also 
under  Typhoid  fe- 
ver 
Immunization,    passive,    74 

antitoxins  in,  86 

definition  of.  64 

history  of.  74 

in  diphtheria.  See  Diph- 
theria antitoxin 

in  diseases  caused  by 
bacteria  which  do 
not  form  soluble 
toxins,  468.  See 
also  under  Serum 
therapy 

therapeutic  application 
of.  75 

toxin-antitoxin  reaction 
in.  104 

underlying  principles  of, 

75 

Immunized     animals,     bac- 
teriolysis   in,    137 
summary   of   facts   in, 

138 

Incubation   of  bacteria.   26 
Infection,     acquired     resis- 
tance to,  60 

adaptation  of  bacteria 
in  tissues  in.  6,  7 

aggressin  secretion  of 
bacteria  in  body 
and,  20-22 

body  temperature  and.  51 

capsule  formation  of 
bacteria  and.  18 

chronic,  adaptation  of 
bacteria  in,  8 

clinical  manifestations 
of,  28 

conjunctiva  susceptible 
to,  13 

criteria    governing.    3 

cultural  conditions  for 
bacteria  in  body 
and.  56 

defense  of  intestinal 
tract  in,  12 

defense  of  mucous  mem- 
branes in.  12.  13 

defense  of  skin  in,  12 

definition  of.  5,  6 

different,  produced  by 
same  bacteria  23 

effect  of  body  tempera- 
ture on  invasive 
powers  of  bac- 
teria in.  2 

effect  of  cultural  adap- 
tability of  bac- 
teria on,  virulence 
of,  12 


576 


INDEX 


Infection,  effect  of  path  of 
introduction       o  f 
bacteria  on,  12-14 
effect     of     quantity     of 
bacteria          intro- 
duced on,  14 
entrance    of    bacteria    i-i 
body  tissues  in,  6 
focus  in,  7-9 
from    bacteria    in    blood 

stream,   24 
generalized,  24 
increased    invasive    pow- 
ers  of   bacteria   a 
factor  in,   56 
incubation     of     bacteria 

in.  26 
individual         differences 

and.  56 

inheritance     and      resis- 
tance to,  56 
localized.  23 
reaction  in,  26 
selective       action       of 

bacteria  in.  25 
through          accidental 

conditions.    25 
natural      resistance 

against,  49 

of  various  diseases,  rel- 
ative susceptibil- 
ity of  man  and 
animals  to.  52 
protective  action  of 
blood  serum 
against,  50 

protective  action  of  leu- 
kocytes against, 
50 

protective   action   of  tis- 
sues against,  50 
ptomains  and.   28 
ptomains      a  s      indirect 

cause  of.  31 
racial     differences     and. 

55 
resistance   of   living   cell 

•to.   6 
secondary    abscesses    in, 

25 

secondary  modifying  fac- 
tors in,  2 

selective  lodgment  of 
bacteria  in  body 
and.  40 

similar,  produced  by  dif- 
ferent bacteria, 
23 

species  resistance  to.  51 
specificity      of     bacteria 

and.  22 

susceptibility    to,    racial 
differences   in,    55 
relative,  51 

variation  in,  of  differ- 
ent strains  of 
same  bacteria,  15, 
16 

variation  in  degree  of, 
in  bacteria  suc- 
cessively passed 
through  animals, 
16,  17 

without    infectious    dis- 
ease, 6,  7,  517 
Infectious    disease,    defini- 
tion of,  6.  8 

influence  of  injections  of 
non-specific  sub- 
stances upon,  520 
Inflammation,  process  of, 
and  phagocytosis, 
280  et  seq. 


Inflammation    with    p  y  o  - 
genie    staphylo- 
cocci.  281 
with  tubercle   bacilli, 

283 

Influenza,  relative  suscep- 
tibility of  man 
and  animals  to, 
53 

Inheritance,  a  factor  in  re- 
sistance to  infec- 
tion, 56 

iso-agglutinins    in    blood 
serum     influenced 
by,  58 
Inhibition  zones  in  colloid 

reactions,   162 
in   precipitation  and  ag- 
glutination,    162 
of  sera  in  agglutination, 

236 

Intestinal  tract,  defence 
of,  in  infection, 
12 

Iso-agglutinins,  237 
grouping   of.    237.    238 
in  blood  serum,  58 
value  of  presence  of.  239 
Isohemolysins,   146 
Isolysins,  human,  148 
grouping  of,   148 
testing   of,    for    trans- 
fusion. 149 
iso-agglutinins  analogous 

to.  237 
Isoprecipitins,  255 


Jacobsthal's  ultramicro- 
scopic  method  of 
finding  precipi- 
tates in  syphilitic 
sera,  204 

Jenner,  Edward,  experi- 
mentation of,  for 
i  m  m  u  n  i  z  a  t  ion 
against  small-pox, 
62 

Jobling's  work  on  serum 
therapy  in  cere- 
brospinal  menin- 
gitis, 472 

Jobling  and  Petersen's 
work  on  anaphy- 
laxis,  396 

Jochmann's  investigations 
in  serum  therapy 
of  epidemic  cere- 
brospinal  menin- 
gitis, 471,  472 


Kolle's  method  of  prophy- 
lactic vaccination 
in  cholera,  489 

Kolle  and  Otto's  investi- 
gations in  pro- 
phylactic immuni- 
zation against 
plague,  489 

Kolle  and  Wasserrnann's 
investigations  in 
serum  therapy  of 
epidemic  cerebro- 
spinal  meningitis, 
471 

Kraus,  Rudolf,  discovery 
of  specific  precip- 
itins  by.  248 

Kraus  and  Doerr's  study 
of  bacterial  ana- 
phylaxis,  412,  413 


Kraus  and  Stenitzer's  se- 
rum in  treatment 
of  typhoid  fever, 
479 


L+,  definition  of,  109 
method  of  determination 

of.   109 

L0,  definition  of,  109 
constancy,    110 
method  of  determination 

of.   110 

Laking,  erythrocyte,  91 
''Landsteiner  phenomenon" 
of     autohemolysis 
in         hemoglobiu- 
uria.  147 

Leishmann's   technique   for 

determination     of 

opsonic  index,  32i> 

"Leistungskern,"   definition 

of,   126 

Leprosy,    relative   suscepti- 
bility of  man  and 
animals  to,  54 
Lesne   and    Dreyfus'    work 
on       anaphylaxis, 
375.  376 
Leukine.  305 
^Leukocidins,  21 
Leukocyte   extracts,    thera- 
peutic use  of,  308  I 
et   seq. 

Leukocytes,    alexin    extrac- 
tion from,  304  et 
seq. 
growth    of    bacteria    in, 

298 
in  bacteriolysis.  140 

action   of,   168 
in     leukocytosis.     action 

of.  275.  276 
in    phagocytosis.    324 
origin      of      bactericidal 
and        hemolytic 
substances     from, 
168,  169  et  seq. 
oxidasc  in,  527 
phagocytic  powers  of,  50 
proteolytic      enzymes 
from,  in   phagocy- 
tosis.     306,      307, 
525.  526  -, 
Leukocytosis.  200 

bacteria  decreasing,  290 
bacteria  increasing.  290 
sources  of  leukocytes  in, 

290 

Leukoproteasos,  306,  307 
Leukotoxin,  92 
Limes-necrosis   (L-n).  463 
L  i  p  o  i  d     constituents    of 
cells,    relation    of, 
to  antigenic  prop- 
erties, 97 

Lister  on  phagocytic  activ- 
ities   of    blood   in 
natural  immunity, 
79 
Liver,  production  of  alexin 

in,   173 

Loeb's  experiments  with 
eggs  of  Fundulus, 
561 

Lubarsch  on  bactericidal 
properties,  of  ex- 
travascular  plas 
ma  or  serum  in 
immunity,  81 

Liidke's  work  on  serum 
therapy  in  ty- 
phoid fever,  479 


INDEX 


577 


Lustig's       antiplague       se- 
rum, 482 
Lysins,   production  of,  130 


Macrocytase,  109,  301 

Magendie  on  anapbylaxis, 
339 

Malta  fever,  relative  sus- 
ceptibility of  man 
and  animals  to,  53 

Manwaring's  work  on  ana- 
phylaxis,  370 

Markl's  s  e  r  u  m  in  treat- 
iiu'nt  of  plague, 
481 

Marmorek's  work  on  serum 
therapy  of  strep- 
tococcus infec- 
tions, 474,  475 

Measles,  relative  suscepti- 
bility of  man  and 
animals  to,  54 

Meat  poisoning,  4,  31 

Meiostagmiu    reaction,    538 
Ascoli  and  Izar's  experi- 
ments in,  538,  539 
value    of,     in    diagnosis, 
538 

Meningitis,  epidemic  cere- 
brospinal,  serum 
therapy  of,  471. 
Kee  also  under 
C  e  r  e  b  rospinal 
meningitis,  .  epi- 
demic 

Metabolism,  processes  of, 
compared  with 
those  of  antibody 
formation,  125 

Metchnikoff  and  Besredka's 
living  sensitized 
vaccines  for  pro- 
phylactic typhoid 
immunization,  486 

Metchnikoff  on  bacterial 
growth  in  cell- 
free  blood  serum, 
81 

theory  of,  on  bacterial 
destruction  by 
phagocytic  cells 
in  natural  immu- 
nity, 78 

Metchnikoff's  soured  milk 
therapy,  31 

Microcytase,  109,  301 

Minimum   lethal  dose,  defi- 
nition of,  108,  109 
method  of  determination 
of,    109 

M  L,  D,  definition  of,  108. 

109 

method  of  determination 
of.  109 

Morgenrotlfs  toxin-HCl 
modification  In 
t  o  x  i  n-antitoxin 
reaction,  100 

Mucous  membranes,  de- 
fence of,  in  infec- 
tion, 12.  13 

Mushroom,  specific  anti- 
toxin from,  97 


Narcotics,       reduction  of 

Ehagocytosis  by. 
99 

Natural       immunity.  &ec 

Immunity,  nat- 
ural 


i    Necroparasites,    11 
"Negative"  phase    in  active 
immunization,  33o 
second     injection     in, 

338 
successive       inoculations 

in,  338,  339 
"summation"      of,      338, 

339 

Neisser  and  Friedemann, 
experiments  of, 
on  influence  of 
salts  on  sensitized 
bacteria  in  agglu- 
tination, 244 

Neisrror  and  Wechsberg, 
phenomenon  •  of, 
100  et  fsgy. 

analogous  to  colloid  re- 
actions, 102 

argument     in     favor     of 
Bordet's        views, 
102 
Gay's      explanation      of, 

103 

Morgenroth  and  Sachs's 
experiments  sup- 
porting, 103 

pro-agglutinoid  zone  re- 
action analogous 
to,  102 

Neisser's  studies  in  syph- 
ilis, 500,  509 

Neoplasms,  malignant, 

alexin   fixation   in 

diagnosis    of,    21  ;i 

von     Dungern's     method 

of,   214 
antigen    production 

for,  214 
results  of,  215 
technique    of,    214    et 

seq. 

Nernst  on  views  of  Arrhen- 
ius  and  Madsen 
on  neutralization 
in  toxin-antitoxin 
reaction,  122 

Neufeld  and  Dold's  experi- 
m  e  n  t  s  in  bac- 
terial anaphylax- 
is,  419 

Neufeld      and      Ilaendel's 
work     on      serum 
therapy    of    pneu- 
monia, 470 
Neurotoxin,  92 

in   snake   venom,   407 
Nicolle's     theory     of    ana- 

phylaxis,   391 
work  on  passive  anaphy- 

laxis.   380 

Noguchi's    modification    of 
the      Wassermann 
test,  208,  209 
schematic       presentation 

of,  209 

"Normal"   diphtheria   anti- 
toxic   serum,    107 
"Normal"    diphtheria    tox- 
in,  107 

"Normal"   serum.     &ec  un- 
der Serum 
agglutinins  in,  91 
hemolytic    properties   of. 

i/l 

opsonins  in,  91 

toxic  action  of,  and  an- 

aphylaxis.   400 
Nuttall      on       bactericidal 
power    of    normal 
blood    in    natural 
immunity,  79 


Nuttall's  experiments  on 
determining  zoo- 
logical classifica- 
tions by  means  of 
precipitin  reac- 
tion, 254,  255 


Ophthalmia,  sympathetic, 
Elschnig's  expla- 
nation of,  as  ana- 
phylactic  r  e  a  c- 
tion,  439 

Opium,    reduction    of    pha- 
gocytosis   by,   299 
Opsonic     action,     phagocy- 
tosis due   to,   313 
Opsonic    index,    determina- 
tion     of,      Leisn- 
inaun's    technique 
for,  329 

Simon,     Lamar     and 
Bispham's       tech- 
nique of,  332,  3;^.> 
Wright's          technique 

for,  330  et  sea. 
difficulties    in,    332, 

333 

value  of,  333 
fluctuation     of,     in     un- 
treated      patients 
under  influence  of 
exercise     of     dis- 
eased   parts,    340 
in     autoinoculations     by 

massage,   340 
in    sera    of    normal   and 
infected      individ- 
ual,      comparison 
of,  334 

in  serum  therapy,  com- 
parison between 
that  in  exudate  of 
infected  foci  and 
blood  serum,  340 
in  staphylococcus  infec- 
tions, 334 

during  vaccine  treat- 
ment with  dead 
s  t  a  p  hylococcus 
cultures,  335 
in  treatment  of  gonor- 
r  h  e  a  1  arthritis 
with  autoinocula- 
tion  by  massage, 
340 

in  vaccine  therapy,  im- 
provement and, 
341 

of  acne,  339 
of    staphylococcus    fu- 

runculosis,  335 
of  sycosis,  330,  337 
of  tuberculosis,  341- 

343 

value    of,    in    control- 
ling      therapeutic 
vaccinations.     344 
in     showing     degree 
and   conditions  in 
which  vaccination 
is  successful,  344, 
345 
vaccine  therapy  and,  328 

et  seq. 
value  of,  in  therapeusis, 

338 
Opsonic   powers  of  normal 

serum,   314 
reduction    of,     by    heat, 

314 

Opsonins.  311.  See  also 
Phagocytosis 


578 


INDEX 


Opsonins,  definition  of,  313 
immune,    bactericidal 
sensitizers      and, 
321    et    seq. 
increase  of,  315 

heated,    increase    of 
power    of,    by    ad- 
dition     of      fresh 
normal  serum,  318 
reactivation    of,     by 
addition    of    alex- 
in,  318 
normal     and,     320     et 

scq. 
resistance  of,  to  heat, 

315 

specificity  of,  321 
thermostability  of,  320 
normal,  91 

cooperation  of  heat- 
stable  and  heat- 
sensitive  body  in, 
319 

instability  of,  314,  315 
nature  of,  316 
similarity   of,   to  alex- 
in  or  complement, 
316,   317 

specificity  of,   318 
production  of,  in  thyroid 

gland,   173 

qualitative  difference  be- 
tween normal  and 
immune,  315, 
316 

specific  thermostable,  in 
normal  serum. 
317 

Organ    specificity    of    anti- 
gens, 99 

"Osmotic     theory"    of    im- 
munity, 84 
Otto's    work    on    anaphy- 

laxis.  361 
in    passive    anaphylaxis, 

380 

Oxidase,  in  proteolytic  en- 
zymes, 527 


Pancreas   cytotoxin,   92 

Panum's  theory  of  intra- 
cellular  destruc- 
tion of  bacteria 
in  natural  immu- 
nity, 78 

Parasites,  biological  transi- 
tion of  sapro- 
phytes to.  5 

Bail's  classification  of, 
11 

Parasitic  bacteria,   4 

Paratyphoid    fever,    agglu- 
tination    reaction 
in     diagnosis     of, 
221 
immunization        against, 

484 
vaccination   against.   484 

Paroxysmal        hemoglobin- 

uria.  147 
hemolysis  in,   147 

Partial  absorption  method 
of  E  h  r  1  i  c  h  in 
measurement  of 
toxin  -,  antitoxin 
combination,  115 

Pasteur,  "exhaustion  the- 
ory of,"  83 

experimentation  of,  on 
i  m  m  u  n  ization 
against  chicken 
cholera,  63 


Pasteur,  work  of,  on  immu- 
nization against 
anthrax,  64 
on  prophylactive  im- 
munization in  ra- 
bies, 492 

Pathogenic  bacteria,  adap- 
tation of,  in  tis- 
sues, 6,  7 

entrance  of,  in  body  tis- 
sues, 6 
saprophytic      nature     of 

certain,  4 

Pathogenic  micro-organ- 
isms,  definition 
of,  3 

occurrence  of,   2 
resistance   of   living   cell 

to,  6 

Pearce     and     Eisenbrey's 
work    on    anaphy- 
laxis,   368,    369 
Persensitized  cells,   180 
Petterson's      investigations 
on  therapeutic  use 
of     leukocyte    ex- 
tracts, 308 

Pfaundler's     thread     reac- 
tion in  agglutina- 
tion,  222,   223 
Pfeiffer   on   causes  of   bac- 
terial    anaphylax- 
is,  414 
work  of.  on  anaphylaxis. 

366 

"Pfeiffer  phenomenon"  in 
active  immuniza- 
tion, 89 

in  bacteriolysis.  tech- 
nique of,  138  et 
seq. 

Metchnikoff's      view      of 
phagocytosis        in 
peritoneal         exu- 
date  and,  302 
Phagocytes,    276 
fixed,  276 
macrophages,   277 
microphages,  277 
motile,  276 
Phagocytic  index.  331 
Phagocytosis,  272 

acquired    immunity    and. 

298 

a  1  e  x  i  n  extraction  in. 
from  leukocytes 
and  lymphatic  or- 
g  a  n  s  ,  304  et 
seq. 
chemotaxis  in.  285 

influence     of     bacteria 

in.  288.  289 
influence    of    bacterial 

extracts  in.  288 
malic  acid  in,   286 
of  slime-molds  or  myx- 

omycetes,    286 
of       spermatozoa       of 

ferns,    286 
Pfeffer's    technique    in. 

286 
destruction     of    bacteria 

in.  297,  300 
by   alexin    (or   cytasr) 
in  leukocytes, 

301.  302 
action   of,   302 
Metchnikoff's     inter- 
pretation of,  302 
by  exudates.   300 
by  phagocytes,  300 
destruction  of  red   blood 
cells  by,  276 


Phagocytosis,  differences  in 
degree   of.  due   to 
bacteria,  325 
differences  in  phagocytic 
energy   in,   due   to 
leukocytes  in,  324 
differences    in    virulence 
of     bacteria,      de- 
pendent   on    their 
resistance   to    leu- 
kocytes in,  325 
digestion    among     proto- 
zoa and,  274 
"dust   cells"    in,    279 
early    investigations     in. 

272 
endothelial  cells  in,  278. 

279 

enzj'mes  in,  endocellular 
and  extracellular. 
305 

eosinophile  cells  in,   27<s 
factors  determining,   311 
tixateur  or  sensitizer  in. 
action    of,    in    im- 
munized   animals. 
301 

giant  cells  in,  280 
foreign  body.  280 
tuberculous,  280 
in  daphnia,  296 
in  higher  animals,  296 
in  immune  serum,  315 
bacteriolysins   in,   bac- 
tericidal   sensitiz- 
ers   and,     321     ft 
neq. 

heated,  opsonic  action 

in,  increase  of,  by 

addition    of    fresh 

normal  serum,  31. S 

increase  of.  311 

attributed  to  "stim- 

ulins,"   311 
with     addition     of 

leukocytes,  312 
opsonin      contents      a 

factor   in,   313 
opsonins    in,    increase 

of,   315 
normal     opsonius 

and.  320  et  seq. 
specificity  of,  321 
thermostability  of, 

320 

in  immunity,  90 
in  normal  serum,  op- 
sonins in,  cooper- 
ation of  h  e  a  t- 
sensitive  body  in, 
319 

nature  of,   316 
similarity       of,       to 

alexin,  316,  317 
specific      thermosta- 
ble. 317 

specificity  of,  318 
in   process   of   inflamma- 
tion, 280  et  seq. 
with    pyogenic   staphy- 

lococci.  281 
with    tubercle    bacilli, 

283 

increase  of.   by  injection 
of     leukocyte    ex- 
tracts, .308  et  seq. 
in       increased       resis- 
tance, 329 
with      acquisition      of 

immunity,   299 
intracellular       digestion 

and,   274 
in   vertebrates,  275 


INDEX 


579 


Phagocytosis.        leukocytes 

in,  324 

action  of,  275.  276 
polynuclear.  278 
leukocytosis   in,   290 
bacteria        decreasing, 

290 
bacteria        increasing, 

290 
lymphocytes     in,      large. 

measure  of  degree  of.  in 
active     immuniza- 
tion. 329 
Leishmann's  technique 

for,  329 

Simon,     La  mar     and 
Bispham's       tech- 
nique     for,      332. 
333 
value  of,   in  therapeu- 

sis,  338 

Wright's          technique 
for,      330      e t 
seq. 
difficulties    in,    332, 

333 

value  of,   333 
mechanism  of  process  of, 

280  et  seq. 

Metchnikoffs  early  in- 
vestigations o  n, 
273.  274 

normal  and  immune  op- 
sonic  action  in. 
quantitative  dif- 
ferences between, 
324 

normal  degenerative  and 
retrogressive  proc- 
esses and.  27(5 
observation  of,  in  vitro, 

313 

of  micro-organisms,  with 
or     without     cul- 
ture media.  297 
opsonins    in,    311.       See 

also  Opsonins 
phagocytes    engaged    in, 
varieties  of,   276 
fixed,   276 
macrophages,    277 
microphages,    277 
motile.  276 

process  of  inflam- 
mation in,  280  et 
seq. 

proteolytic  enzymes  from 
leukocytes  in.  306, 
307 

qualitative  difference  be- 
tween normal  and 
immune  opsonic 
substances  in.  315, 
316 
reduction  of  phagocytic 

activity  in.  298 
by  growth  of  bacteria 
within  leukocytes. 
298 

by    protection    of   bac- 
teria    from     pha- 
gocytes, 299 
by    use    of    narcotics, 

299 

relation  of,  to  active  im- 
munity, 329 
relation  of  virulence  to, 

312 

removal  of  extravasa- 
tions of  blood 
and.  275 


Phagocytosis,  resistance  of 
bacteria  to,  due  to 
non-absorption  of 
opsonin,  326 

resistance  of  infected 
subject  and,  296, 
297 

resistance      of      virulent 
bacteria     to,     in 
normal  serum,  325 
spontaneous.    313 
tissue  cells  in,  278 
varieties    of    body    cells 

engaged  in,  278 
dependent    on    nature 
of    invading    sub- 
stance, 279 

Pick  and  Yamanouchi's  ex- 
periments on  two 
s  e  p  a  ra  t  e  sub- 
stances in  ana- 
phylactic  antigen, 
387 
work  on  anaphylaxis, 

371 

von  Pirquet  and  Schick's 
studies  of  serum 
sickness,  428  et 
seq. 

von     Pirquet's     tuberculin 
skin  reaction.  442 
Placentar  cytotoxin,  92 
Plague,  active  prophylactic 
i  in  in  u  n  iz  ation 
against,  489 
Besredka's  vaccines  in, 

490 
Haffkine's    early    work 

on,   489 

Kolle    and    Otto's    in- 
vestigations      i  n, 
489 
Rowland's   vaccine   in, 

490 
Strong's  investigations 

in,   490 

relative  susceptibility  of 
man  and  animals 
to,  53 

serum  therapy  of,  480 
Dean's   serum   in,    482 
Lustig's  serum  in.  482 
Markl's  serum  in,  481 
Rowland's     serum     in, 

482 

value   of,    482,   483 
Yersin.    Calmette    and 
Borrel's       investi- 
gations in,  480 
Yersin's  serum  in,  480. 

482 

value  of,  481 
"Plasmines,"  72 
Pneumococcus        infection, 
relative     suscepti- 
bility of  man  and 
animals  to,  54 
Pneumococci,    mutation   of, 

474- 

Pneumonia,      agglutination 
reaction    in    diag- 
nosis of,  221 
serum   therapy  in.  Cole's 

work  on,  477 
Dochcz       and       (iilles- 
pie's  work  on,  477 
nature    of    action    in. 

476,  477 
Neufeld  and  Ilaendel's 

work  on.  476 

Poison-ivy,     specific     anti- 
toxin  from,  97 
Pollantin.   437 


Poliomyelitis,    carriers    in, 

496 

cause  of,   495 
epidemiology   of,   495 
Flexner       and       Amoss' 
work  on,  497,  498 
Flexner        and        Lewis' 

work  on,  495 
infection    and    immunity 

in,   495 

Landsteiner's  and  Pop- 
per's work  on, 
495 

relative  susceptibility  oi' 
man  and  animals 
to,  55 

Polyceptors   (Ehrlicli)./l7M"' 
Precipitation,  248 

inhibition   zones   in,   162 
Precipitin    reaction,    248 
against   heated    proteins. 

258  et  seq. 

coctoprecipitin  in,  25,s 
experiments    on,    260- 

262 
h  e  a  t-alkali-precipitin 

in,  259,  260 
native     precipitin     in, 

259 

70°    precipitin  in,   259 
Schmidt's   experiments 

on,  260,  261 

agglutination        reaction 
analogous  to, 
263 
analogy    of,    to    colloidal 

flocculation,  265 
autocytotoxins  in,  263 
bacterial    precipitins    in. 
partial   or    minor, 
252 

specificity  of,  251,  252 
von  Dungern's  views  of, 

268,   271 
Ehrlich's    conception    of, 

264 
electrolytes  in,  effect  of. 

265 
forensic    blood    test    in, 

257 
ring  test  of  Fornet  and 

Miiller  in,  257 
group    reactions    of   bac- 
terial    precipitins 
in,  251,  252 
diagnostic  value  of. 

252 

heat   precipitins   in,    259 
heated  precipitating  ser- 
um,    effect    of 
mixed      sera      in, 
266-267 
protective     action     of, 

266 
inhibition  zones  in,  265, 

266 

isoprecipitins  in,  255 
medico-legal      value      of, 

254 

non-specific  partial  reac- 
tions in,  elimina- 
tion of,  254 

Nuttall's  experiments  on 
determining      zoo- 
logical    classifica- 
tions by,   254,  255 
organ  specificity  in,  262. 

263 

precipitinogen   in.    249 
chemical      nature     of, 

249 

effect    of     heating    on, 
258 


580 


INDEX 


Precipitin  reaction,  precipi- 
tinogen    in,     non- 
protein,  249,   250 
obtaining   of,    249 
precipitinoids    in.    for- 
mation of,  265 
precipitins     in,     delicacy 

of,   253 

determination    of     po- 
tency of,  253 
inactivation      of,      by 

heat,   264 

effect     of,     in     bac- 
terial filtrates,  264 
production    of,  against 
unformed       p  r  o  - 
teins.  252,  253 
methods  of,  251 
of  specific,   249 
by  pepton,  251 
effect    of    heating 

on,   249 

in  animal  sera  by 
foreign   protein, 
248 
structure  of  (Ehrlich), 

264 
zymophore    group    In, 

264 
effect    of    heat    on, 

265 

quantitative  proportions 
in,  effect  of,  265, 
266 

relative  concentration  of 
reacting  bodies 
a  factor  in,  265. 
266 

residue  antigen   and  an- 
tibody  in,    267   et 
seq. 
explanations     of,     268 

et   seq. 

experiment    on,    269 
salts  in.  effect  of,  2(55 
species  determination  by 
means    of,    253. 
254 

species  specificity  in.  262 
specificity    of,    248,    253. 

254 

vegetable  proteins  deter- 
mined by.  255 
zoological    classifications 
by  means  of,  254, 
255 

Precipitin     tests,     methods 
of  performing,  255 
et   seq. 
forensic  blood  test  in, 

.      257 

ring     test    of     Fornet 
and  Miiller  in,  257 
Precipitinogen.    249 

chemical  nature  of.  249 
effect     of     heating     on, 

258 
non-protein,    249,    250 

nature   of,    250 
obtaining  of,  249 
Precipitinoids.   265 
Precipitins.  248 

against   heated   proteins, 

258  et  seq. 
coctoprecipitin.    258 
experiments    on,     260- 

262 
heat  -  alkali-precipitin, 

259.   260 

native    precipitin.    259 
70°   precipitin.  259 
Schmidt's   experiments 
on,  260,  261 


Precipitins,  bacterial,  group 
reactions   in,   251 
partial   or   minor,   252 
specificity  of,  251,  252 
definition  of,  90 
delicacy  of,  253 
determination  of  potency 

of,   253 
heat,  259 

in   hemolytic  serum.   94 
inactivation  of,   by  heat, 

264 
effect   of,   in    bacterial 

filtrates,  264 
isoprecipitins,  255 
organ  specificity  of,  262, 

263 

production    of,    129,    249 
against  unformed  pro- 
teins, 252.  253 
methods  of.  251 
specific,       by      pepton. 

251 
effect  of  heating  on, 

249 

in    animal    sera    by 
foreign        protein, 
248 
"species,"    specificity    of, 

262,   263 
specific,   248 

discovery    of,     by    Ru- 
dolf Kraus.  248 
structure    of     (Ehrlich), 

264 

zymophore  group  in,  264 
e  ff  e  c  t     of    heat    on, 

265 

Pregnancy,  diagnostic- 
value  of  Abder- 
halden's  protec- 
tive ferments  in, 
534 

Pro-agglutinoid  phenome- 
non in  agglutina- 
tion explained  as 
protective  colloid 
action,  236 

Pro-agglutinoid      zone      In 
agglutination,  1(52 
complement  deviation  re- 
action    analogous 
to.   162 

Pro-agglutinoids.  235 
Prof  eta's  law,  499 
Prophylactic         immuniza- 
tion, active,  in 
man,  483 

against  cholera,  487. 
See  also  under 
Cholera 

against  paratyphoid  fe- 
ver. 484 

against  plague.  489.    See 

also  under  Plague 

against  rabies.   492.    *SVe 

also  under  Rabies 

against    small-pox,    490. 

See     also      under 

Small-pox 

against     typhoid     fever, 
484.     See  also  un- 
der Typhoid  fever 
"Protection."   195 
"Protein  fever,"  367 
Proteins    unknown,    alexin 
fixation    test    in 
determination    o  f 
nature  of.  211 
delicacy  of.  213 
technique  of.  212 
Proteolysis.  524 
Prototoxoids,   115 


Protozoa,  digestion  among, 

and     its     relation 

to       phagocytosis. 

274 

Ptomains,  as  indirect  cause 

of  infection,  30 
chemistry    of,    29 
classification  of,  29 
definition  of,  31 
relation  of,  to  infection, 

28 
Putrefaction,  chemistry  of, 

29 
micro-organisms  causing, 

Pyemia,    25 


Rabies,  active  prophylactic 
i  in  m  u  n  i  zation 
against,  492 

Hogyes  method  of  treat- 
ment in,  494 

Pasteur's  work  on,  492 

preparation  and  attenu- 
ation of  virus  for. 
493 

treatment  of  patients  in, 

493 
Ranzi's    work    on    anaphy- 

laxis,  367 

Rattlesnake     poison,     anti- 
toxin  for,   468 
Reaction,      colloidal      gold. 
539 

epiphanin,  537 

meiostagmin.    538 

toxin-antitoxin,    104 
Receptors,      cell,      overpro- 
duction of.  152 

complementophile,   149 

cytophile,  149 

definition  of,  126 

of  third  order,  150 

sessile,  in  anaphylaxis, 
399 

structure  of,  127 
Resistance.      tiee   also    Im- 
munity 

acquired.   60 

bactericidal  properties 
in  serum  and,  21)7 

body     temperature     and, 

cellular    theory   of,    80 

cultural  conditions  for 
bacteria  in  body 
and,  56 

degree  of  phagocytosis 
and.  296,  297 

humoral   theory  of,   80 

increased  invasive  pow- 
ers of  bacteria 
and.  56 

individual  differences 
and.  58 

inheritance  a  factor  in, 
56 

local,  in  skin  infections, 
103 

natural,  against  infec- 
tion, 49 

racial   differences   in,   55 

species,    to   infection,    51 

vital,  524 

"Retention  theory  of  im- 
munity," 83 

Rhus  toxicodendron,  spe- 
cific antitoxin 
from,  97 

Richet  and  H6ricourt'9 
work  on  anaphy 
laxis.  360 


INDEX 


581 


Richct  and   Portier's  work 

0  n      anaphylaxis, 
360 

Richet's  work  on  passive 
anaphylaxis,  380 

Riein,   "protein-free,"   97 

Romer's  method  for  diph- 
theria antitoxin 
s«  t  a  ndardization, 
401 

"Root-tubercle"    bacilli,    7 

Rosenau  and  Anderson,  re- 
searches of,  in 
bacterial  anaphy- 
laxis,  412 

work    on    anaphylaxis, 
362,  374  et  seq. 

Rosenow  on  variations  in 
streptococci,  474 

Roux  and  Yersin,  experi- 
mental immuniza- 
tion in  hog  chol- 
era by,  72 

Rowland's     antiplague     se- 
rum, 482 
vaccine    in    prophylactic 

1  m  m  u  n  i  zation 
against        plague, 
489 

Russell's  vaccines  for  pro- 
phylactic immuni- 
zation against  ty- 
phoid fever,  485 


Salt-inactivation  of  alexin, 

178 

Salts,  effect    of,  in    agglu- 
tination, 243 
in    precipitation,    265 
Saprophytes,  biological 

transition    of,    to 
parasites,  5 
occurrence    of,    2 
pathogenic      powers      of 

certain,  4 
pure,  11 

Saprophytic  micro-organ- 
isms, definition  of, 
4 

Scarlet  fever,  relative  sus- 
ceptibility of  man 
and  animals  to, 
54 

Schmidt,    precipitation    ex- 
periments   of,    on 
heat      precipitins, 
260,   261 
Sensibilisin,   387 
Stnsibilisinogen.    387 
Sensitization.    134,    359 
Bordet's   views   on,    162, 

163 
complement  deviation  in, 

160 

Ehrlich  and  Sachs's  phe- 
nomenon   in,    165 
Bordet    and    Cray's   in- 
terpretation       of, 
166,  167 

Ehrlich  and  Sachs's 
views  on,  164, 
165 

Neisser-Wechsberg      phe- 
nomenon in.  160 
Sensitized    bacteria,    immu- 
nization   with.    68 
Sensitized  tuberculin.  357 
Sensitized   vaccines,   355 
Sensitizer.      See    Ambocep- 

tor 

Bordet's  definition  of, 
159 


Sensitizer,  quantitative  de- 
termination of.  in 
immune  serum, 
160,  161 

Septicemia,  chronic,  adap- 
tation of  bacteria 
in,  7 

secondary  foci  in,  7 
Serum;     See     also     Blood 

serum 

activities  of,  523  et  sc<i. 
antitoxic,     direct     effect 
of,    on   toxin,    104 
indirect   protective  ac- 
tion    of.     against 
toxin,  104 
bactericidal     powers    of, 

79,   134 

bactericidal  properties 
of,  in  immunity, 
81 

resistance  and,  297 
cell-free,    bacterial 

growth  in,  81 
immune,     bacteriotropins 
in.  without  lysins, 
322 

heated,  reactivation 
of.  by  addition  of 
alexin.  318 

opsonins  in,  bacteri- 
cidal sensitizers 
and,  321  et  seq. 
heated,  increase  of 
power  of.  by  ad- 
dition of  fresh 
normal  serum, 
318 

increase  of,  315 
normal     opsonins 

and.   320   et  seq. 
specificity  of.  321 
phagocytosis     in,     91, 

increase      of,      311. 
See  also  Phagocy- 
tosis 
normal,    agglutinins    in, 

91 

antic  omplementary 

properties  of,  196 

hemolytic       properties 

of,  91 
o  p  s  o  n  i  c   powers  of, 

314 
reduction      of,      by 

heat,  314 
opsonins  in,  91 

cooperation  of  heat- 
stable     and     heat 
sensitive  body  in, 
318,  319 
nature  of,  316 
resistance      of.      to 

heat,  315 
similarity       of,       to 

alexin.  316,  317 
specificity  of,  318 
thormostability  of. 

320 
specific      thermostable 

opsonins  in,   317 
normal    antitoxic,    diph- 
theria, 107 

opsonins  in  normal  and 
immune,  qualita- 
tive differences  in, 
315,  316 

reactions,    physical    fac- 
tors in.   535 
Serum  sickness,  428 

analogy  of  anaphylaxis 
with,  430 


Serum    sickness,     antibody 

formation  in,  431 

incubation      time      in, 

431 

methods    of    administra- 
tion   of   antitoxin 
to  avoid,  432 
Besredka's  method  of, 

433 
by  alteration  of  serum, 

432 
Friedberger  and  Mita's 

method  of,  434 
in    animal    experimen- 
tation, 433 

with   concentrated  an- 
titoxin, 432 

von  Pirquet  and  Schick's 
studies  of,  429  et 
seq. 
symptoms  of,  428 

accelerated  reaction  of 
von    Pirquet    and 
Schick  in,  429 
after     first     injection, 

428 
after  second  injection, 

429 
immediate  reaction  in, 

429 

analogy   of,   to   ana- 
phylaxis,  429 

Serum  therapy,  anaphylax- 
is in.  See  Serum 
sickness 

in  diphtheria,  448.  See 
also  Diphtheria 
antitoxin 

in  diseases  caused  by 
bacteria  which  do 
not  form  soluble 
toxins,  468 

action   of  serum  upon 
extensive       infec- 
tion in.  469 
antibacterial  action  in, 

468 

in  epidemic  cerebrospinal 
meningitis,  471 
See  also  under 
Cerebros  pinal 
meningitis,  e  p  i  - 
demic 

in  plague,  480.  See  also 
Plague,  serum 
therapy  of 

in  pneumonia,  476.  See 
also  pneumonia, 
serum  therapy 
in 

in  streptococcus  infec- 
tions, 473.  See 
also  under  Strep- 
tococcus i  n  f  e  c  - 
tions 

in  typhoid  fever,  477. 
See  also  Typhoid 
fever,  serum  ther- 
apy of 

Side    chain    theory,    124 
anti-body    production    in 
body  cells  in,  130 
body  cell  in,  125 

chemical      nature     of. 

126 
"Leistungskern"        in, 

126 

side    chains    or    recep- 
tors in,  126 

chemical  action  of  anti- 
gens in.  128 

definition  of  side  chains, 
in,  126 


582 


INDEX 


Side  chain  theory,  diagram 
showing  cell-re- 
ceptors and  im- 
raune  bodies 
(Ehrlich)  in.  127 
in  toxin-antitoxin  reac- 
tion, 124 

overproduction  of  recep- 
tors in,  129 
flow   of,    into   blood    a 
cause     of     immu- 
nity, 128 
physical    mechanism    of, 

124 

recapitulation  of.  130 
relationship  between 

susceptibility  of 
tissue  and  toxin- 
binding  properties 
in,  131 

Skin,  defence  of,   in  infec- 
tion, 12 

Skin    infections,    local    im- 
munity in,  103 
Small-pox,    active    prophy- 
lactic    immuniza- 
t  i  o  n          against. 
490 
Jenner's   discovery   of, 

490 

production  and  prep- 
aration of  vac- 
cine for,  491 
history  of  experimenta- 
tion in  immuniza- 
tion against,  by 
Jenner,  62 

relative  susceptibility  of 
man   and   animals 
to,  54 
vaccination    for,    history 

of,  62 

principles  of.  62 
Smith,  Theobald,  investiga- 
tions   of,    in    ana- 
phylaxis.  363 
phenomenon    of,   in   ana- 

phylaxis.  361 
Snake  venoms.  36 
action  of,  467 
Activation    of.    by    endp- 
complement         in 
blood   cells,   174 
by  sera,  174 
antitoxins  for.  466 

e  ff  e  c  t    of    heat    on, 

ion 

immunization        against, 

466 
K  y  e  s  '    experiments    in. 

174,  175 

neurotoxins  in.  467 
peculiarities  of.  466 
toxin-antitoxin    combina- 
tion   in,    stability 
of.  105 

toxin-antitoxin     reaction 

with.      105.      467 

filtration     experiments 

in.  105 
neutralization     theory 

of,  105 
time     element     in, 

105 

toxin-HCl       modification 
of,   effect  of  heat 
on,  106 
Sols.  544 

Species  resistance,  51 
Specific    hyperleukocytosia 

309 

Specificity,  definition  of,  76 
Spermatotoxins,  92 


Spleen, ,    removal  of,      and 
antibody   -   forma- 
tion in  active  im- 
munity,  101 
and     susceptibility     to 

infection.  101 
Standardization     of     anti- 
toxin, guinea  pigs 
used  in.   108 
minimum  lethal  dose  in, 

109 

Standardization  of  diph- 
theria antitoxin, 
by  means  of  tox- 
in, 107 

early  attempts  at.  107 
unit  in,   107 

Standardization  of  sera, 
Pfeiffer's  method 
of,  139 
Standardization  of  tetanus 
antitoxin  by 
means  of  toxin, 
107 

Standardization  of  vac- 
cines. 352 

Hopkins'   method  of,  353 
Wright's       method      o  f, 

352.   353 
Staphylococcus     furunculo- 
sis,  opsonic  index 
in    vaccine    treat- 
ment of,  335 
Staphylococcus     infections, 
opsonic    index    in. 
334 

during  vaccine  treat- 
ment with  dead 
s  t  a  p  h  y  lococcus 
cultures,  335 
relative  susceptibility  of 
man  and  animals 
to,  54 

Stern's  modification  of 
Wassermann  test. 
209 

Stimulins.  311 
Strauss  test,  25 
Streptococci    variations   in, 

474 

Streptococcus       infections, 
agglutination     re- 
action     in      diag- 
nosis   of,    222 
relative,     of     man     and 

animals,  54 
serum   therapy  in   473 
difficulties     in,     owing 
to     variations     of 
streptococci,     473, 
474 
early  investigations  in, 

474 
Marmorek's    work    on, 

474.   475 
nature    of    action    in, 

475 

standardization    of   se- 
rum in.   476 
value  of,  475 

Strong's  investigations  in 
prophylactic  im 
munization 
against  plague. 
490 

method    of    prophylactic 

vaccination        i  n 

cholera,   489 

Sub-infection,    24,    234 

"Summation     of     negative 

phase"    in    active 

i  m  m  u  n  i  z  ation 

338.  339 


'Summation  of  positive 
phase"  in  active 
immunization,  339 
Surface  tension  in  chcmo- 

taxis,   293 
Susceptibility,     body     tem- 
perature   and,    51 
cultural     conditions     for 
bacteria    in     body 
and,  50 

increased    invasive    pow- 
ers     of      bacteria 
and,   56 
individual          differences 

.    and.  58 
inheritance   a    factor   in, 

56 
racial    differences    in, 

55 
relative,    of    animals    to 

infection,  51 

species   resistance   to   in- 
fection and,  51 
Sycosis,    opsonic    index    in 
vaccine  treatment 
of,  336,  337 
Syntoxoids,  116 
Syphilis,       autoinoculation 

in.    500.   502 
Colles'   law,  499 
colloidal     gold     reaction 

in,  539 

diagnosis  of,  Bordet-Gen- 
gou      phenomenon 
in,  188 
by       alexin       fixation. 

198,    199,    512 
by      direct      precipita- 
tion   of    syphilitic 
serum     by     emul- 
sions   of    lecithin 
and      of      sodium 
glycocholate,     204 
"ideal  fall"  in,  503 
immunity    in,    498,    511. 

512 

Profeta's  law.  499 
reinfection   in,   502 
reinoculation      in,      500, 

502 

relative  susceptibility  of 
man   and   animals 
to,  52,  505,  508 
resistance  of  animals  to, 

508 

superinfection  in,  499 
transmission   of,    to   ani- 
mals, 505 

ultramicroscopic    method 
of  finding  precipi- 
tates   in    sera    of. 
204 
vaccination     against, 

511 

Wassermann  reaction  in, 
diagnostic  value 
of.  210,  211,  512. 
513 

technique  of,  207 
Bauer's  modification 

of.  209 

Noguchi's     modifica- 
tion of.  208.  209 
schematic    presen- 
tation  of.   209 
Stern's    modification 
of,  209 


T..  definition  of.   109 
Temperature,  body,  and  re- 
sistance   to   infec- 
tion, 51 


INDEX 


583 


Tetanus,    "cryptogenic,"    5 
relative  susceptibility  of 
man   and   animals 
to,  53 

toxin,     fixation     in,     by 

brain  tissues,   lol 

lipoidal      substances 

a  factor  in,  133 
proteolytic    enzymes 

a  factor  in,  133 
temperature    a    fac- 
tor in,   132 

Tetanus  antitoxin,   produc- 
tion of,   405 

standardization    ot.    4(>> 

by  means  of  toxin,  10  < 

Tetanus  bacillus,  action  ot, 

4,  5 
Tetanus    toxin,    action    ot, 

41 

Thread  reaction  of  Pfaund- 
ler    in    agglutina- 
tion,   222,   223 
Thymotoxin,  92 
Thyroid   gland,    production 

of  alexin  in,  172 
production    of    opsonins 

in.  173 
Toxemia,  10 

bacterial,  423 

Toxicity,   definition    of,    11 
Toxic    unit,    definition    of, 

109 

Toxin-antitoxin,  diphthe- 
ria. See  Diph- 
theria toxin-anti- 
toxin 

Toxin-antitoxin       combina- 
tion,  chemical   re- 
lations of,  114 
effect  of  heat  on,   106 
in    snake   venom,    stabil- 
ity of.   105 
toxin-HCl  modification 

of,  106 
effect    of    heat    on 

106 

measurement  of,  by  par 
t  i  a  1     absorption 
method     of     Ehr 
lich,  115 
stability  of,   105 
valency      of     componen 

parts  of,  114 
Toxin-antitoxin       reaction, 

104 

analogy  between  chemi- 
cal reactions  and. 
118,  119 

antibody    production    in 
body  cells  in,  130 
body  cells  in,  125 

chemical      nature     of. 

126 

chemical  action   of  anti- 
gens in,  128 
concentration      of      rea- 
gents in,   107 
degrees    of    toxicity    in. 

123 

direct  neutralization  the 
protective  power 
of,  124 

effect  of  heat  on,  105 
effect  of  temperature  on. 

104 

mechanism   of,   104 
neutralization      in.      ab- 
sorption       theory 
of,  123 

Arrhenius     and     Mad- 
sen  on,   120 
Bordet  on,  122 


Toxin-antitoxin  reaction, 
neutralization  in, 
Bordet' -D  a  n  y  s  z 
phenomenon  in, 
123 
von  Dungern's  views 

on,   124 

Danysz    effect    in.    123 
phenomena   of,   119   ct 

seq. 

overproduction  of  recep- 
tors in,   128 
flow   of,    into    blood   a 
cause     of     immu- 
nity, 128 
physical    mechanism    of, 

124 
quantitative  relations  in, 

106,  123 

relationship  between  sus- 
ceptibility of  tis- 
sue and  toxin- 
binding  properties 
in,  131 
side  chain  theory  in, 

124 

specificity  of,  124    129 
speed  of  action  of,  10  < 
time  element  in,   105 
with  snake  venom,  lOo 
filtration     experiments 

in,   105 
neutralization     theory 

of,  105 

time    element    in,    105 
Toxin,  bacterial.     See  Bac- 
terial   toxins 
chemical      relations     of, 
with     antitoxin, 
114 
deterioration    theory    of 

110,   111 

definition    of,    32 
differences   in   combining 

avidity  of.  Ill 
diphtheria,     construction 

of,  118 
normal,  107 
direct  effect  of  antitoxic 

serum'  on,   104 
epitoxoid  form  of,  112 
indirect    effect    of    anti 
toxic     serum     on 
104 

structure  of,  110 
toxoid,  110 

prototoxoids  in.  115 
syntoxoids  in,  116 
toxon.  113 
true,  33 

analogy    of.    with    en 

zymes,  36 

bacteria  producing.  3 
characteristics  -of.  34 
chemically  indefinab 

nature  of.  35 
diseases      for      whic 
some  investigatoi 
claim.   471 
heat    sensitiveness    o 

36 

incubation  time  of.  £ 
production  of  ant 

toxin  by.  35 

Toxin     hypersusceptibilit 
anaphylaxis    an 
409 
Toxin     spectra,      constru 

tion  of,  116 
definition  of,  116 
measurement      of,      11 

117 
principles  of,  116 


oxin    unit,    definition    of, 

109 

diphtheria,   107 
oxoids,  definition  of,   110 
"oxons,  action  of,  113 
definition  of,   113 
structure   of,    113 
oxophore  group  of  toxin, 

110 

action  of,  110 
ropins,  311 

ubercle  bacilli,  effect  of 
body  temperature 
on  virulence  of,  12 
uberculosis,  avian  type, 
relative  suscepti- 
bility of  animals 
to,  52 

bovine  type,  relative  sus- 
ceptibility of  man 
and  animals  to,  52 
complement  fixation  test 
in  diagnosis  .  of, 
217 

human  type,  relative 
susceptibility  o  f 
man  and  animals 
to,  52 

meiostagmin   reaction   in 
diagnosis    of,    538 
of   cold-blooded   animals,' 
immunity      of 
warm-blooded  ani- 
mals to,  52 
opsonic    index    in,     341- 

342,  343 
Tuberculin    ophthalmoreac- 
tion,  anaphylactic 
nature  of,  442 
Tuberculin    reaction,    anal- 
ogy    of,     to    ana- 
phylaxis. 444 
anaphvlactic    nature    of. 

'440 

Bail's  experiments  with 
passive  sensitiza- 
tion  in,  445 

diagnostic  value  ot,   444 
Koch's     experiments    in, 

441 
nature  of,  440 

Babes'      interpretation 

of,  441 
Koch's     interpretation 

of,  441 

Wassermann  and 
Brucks      interpre- 
tation ot.  441 
specific   antibody   forma- 
tion in,  444 

Tuberculin  skin  reaction, 
anapnylactic  na- 
ture of,  442 

von  Pirquet's  interpreta- 
tion of,  443 
Tuberculins.   355 

Bouillon  Filtre  (Denys). 

357 
New     Tuberculin      (TR 

and  TO),  356 
New   Tuberculin  Bacilla- 
ry   Emulsion.   35  < 
Old    Tuberculin    (Koch). 

355 
sensitized         Tuberculin, 

357 

Tumors,    malignant,   alexin 
fixation    in    diag- 
nosis of,   213 
von  Dungern's  method 

of.  214 

antigen      production 
for.  214 


584 


INDEX 


Tumors,    malignant,   alexin 
fixation     in     diag- 
nosis    of,    results 
of,   215 
technique  of,  214  et 

seq. 
organ-specific       qualities 

in,  373 

Typhoid  bacilli,  attenua- 
tion of  virulence 
of,  18 

toxicity  of,  423 
Typhoid  carriers,  3 
Typhoid    fever,    adaptation 

of  bacteria  in,  8 
agglutination        reaction 
for    diagnosis    of, 
219 
macroscopic      method, 

219 
microscopic       method, 

220 

effect  of  path  of  intro- 
duction of  bac- 
teria of,  on  infec- 
tion. 14 

meiostagmin   reaction   in 
diagnosis  of,   53* 
prophylactic      immuniza- 
tion   against,    ac- 
tive, 484 
early    experimentation 

in,  484 

living    sensitized    vac- 
cines used  in.  48(5 
results    of*    in    United 
States  Army.   4S5 
Russell's    vaccines    in, 

485 

sensitized    killed    vac- 
cines in.  486 
relative  susceptibility  of 
man   and  animals 
to,  54 

serum  therapy  in.  477 
Besredka's     anti-endo- 
toxic     serum     in, 
478 

Chantemesse's       early 
experiments      i  n, 
477 
Garbat     and     Meyer's 

work  on.  479 
Kraus  and    Stenitzer's 

serum  in,   479 
Gottstein  -  M  a  t  h  e  s  ' 

work  on.  479 
Liidke  s   work  on.   479 
nature  of  reaction  in, 

478 
vaccine    treatment   in, 

486 

Typhus  fever,  relative  sus- 
ceptibility of  man 
and  animals  to,  54 

Ultramicroscope,  547 

Vaccination,  prophylactic, 
in  man.  483 

in  anthrax,  64 

in  cholera.  487.  See  also 
under  Cholera 

in  paratyphoid  fever, 
484 

in  plague,  489.  See  also 
under  Plague 

in  rabies.  492.  See  also 
under  Rabies 

in  small-pox.  491.  See 
also  under  Small- 
pox 


Vaccination,     prophylactic, 
in    smallpox,    his- 
tory   and    general 
principles   of,   62 
in     typhoid     fever,  484, 
486.    See  also  un- 
der Typhoid  fever 
Vaccine  therapy.     See  also 
Immunization,  ac- 
tive 

anaphylaxis  in,  434 
as    a    therapeutic    meas- 
ure,  action   of,   in 
local       infections, 
346 

in    generalized    sys- 
temic    infections, 
347 
in    successive    local 

infections.     347 
value  of.  in  acute  dis- 
eases. 350 

in    subacute   or 
chronic  cases,  349 
autoinoculations  by  mas- 
sage   or    exercise 
in,  340 
"high  tide"  of  immunity 

in,  340 

"negative"  phase  in,  338 
second      injection      in, 

338    - 

successive          inocula- 
tions in,  338,  339 
summation      of,      338, 

839 
opsonic  index  in,  328  et 

seq. 

comparison        between 

that  in  exudate  of 

infected    foci    and 

blood    serum.    340 

improvement        and, 

341 
in     tuberculosis,     341- 

343 

Leishmann's  technique 
for  determination 
of,.  329 

Simon,  Lamar  and 
Bispham's  tech- 
nique for  deter- 
mination of.  332, 
333 
value  of.  338 

in  controlling  thera- 
p  e  u  t  i  c  vaccina- 
tions. 344 

in  showing  degree 
and  conditions  in 
which  vaccination 
is  successful,  344, 
345 

Wright's  technique  for 
determination    of, 
330  et  seq. 
difficulties    in,    332, 

333 

value  of.  333 
relation    of   phagocytosis 

to.  329 
second  positive  phase  in, 

339 
"Summation    of   positive 

phase"  in.  339 
tuberculins  in,  355 
value  of,  as  prophylactic 
measure,    345, 
346 

as  therapeutic  meas- 
ure, 346 

Vaccines,    autogenous,    351 
production  of,  351 


Vaccines,     production     of, 
with    dead    bac- 
teria,  351 
with     living     bacteria, 

351 

sensitized,  355 
standardization  of,   352 
Hopkins'     method     of, 

353 
Wright's     method     of, 

352,   353 

Vaughan's    work    on    ana- 
phylaxis, 366.  367 
on  bacterial  anaphylaxis, 

Vaughan  and  Wheeler,  pro- 
teid  split  prod- 
ucts of,  in  ana- 
phylactic  poison, 
405 

theory  of,  on  mechanism 
of  anaphylaxis, 
390 

work  of.  on  toxic  frac- 
tion of  protein 
molecule  in  ana- 
phylaxis. 390 

Virulence,  aggressin  secre- 
tion of  bacteria 
in  body  and.  20-22 

capsule  formation  of 
bacteria  and,  18 

definition   of,   11 

dependent  on  resistance 
of  bacteria  to  leu- 
kocytes in  pha- 
gocytosis. 325 

differentiated  from  toxi- 
city, 11 

ectoplasmic  hypertrophy 
of  bacteria  in  re- 
lation to,  19.  20 

effect  of  body  tempera- 
ture on,  12 

effect  of  cultural  adap- 
tation of  bacteria 
on,  12 

effect  of  path  of  intro- 
duction of  bac- 
teria on,  12-14 

effect  of  quantity  of  bac- 
teria introduced 
on,  14 

increase  of,  by  attenua- 
tion of  bacteria, 
17 

measurement  of  relative 
degrees  of.  15 

of  capsulated  bacteria, 
326 

relation  of,  to  phagocy- 
tosis, 312 

relative  to  number  of 
bacteria  i  n  t  r  o  - 
duced,  15 

specificity  of  bacteria 
and,  22 

variation  in.  of  bacteria 
successively 
passed  through 
animals,  16.  17 
of  different  strains  of 
snme  bacteria,  15, 
16 

Virulins.  22.  326 
"Virus    fixe"    in    treatment 

of   rabies.    492 
Vital  resistance,   524 


Wassermann    reaction,    198 
alexin    fixation    principle 
in,  198 


INDEX 


585 


Wassermann  reaction,  alex- 
in  fixation  prin- 
ciple in,  not  by 
union  of  specific 
syphilitic  antigen 
with  spirochaeta 
pallida  antibodies, 
204 

alexin    titration    in,    200 
antigen   preparation   for, 

200 
by  addition  of  choles- 

terin,    201 

by   method   of   Brown- 
ing     and      Cruik- 
Khank,    201 
by  method  of  Noguchi, 

200 
by   methods  of  Forges 

and   Meier.    200 
by    methods    of    Weil 

and  Braun.  200 
titration  in,  202 
diagnostic  value  of,  210, 

211 
in  diagnosis  of  syphilis, 

198 
in    diseases    other    than 

syphilis,    210 
in  normal  organs,  200 
Klausner     theory     in, 
204 


Wassermann  reaction,  pre- 
cipitation in,  by 
addition  to  syphi- 
litic serum  to  leci- 
t  h  i  n  emulsions, 
204 

produced  with  syphilitic 
serum  in  antigens 
from  normal  or- 
gans, 200 

specific  antigen  from 
spirochaita  pallida 
cultures  unsuit- 
able in,  203 

spinal  fluid  used  in  per- 
f  o  r  m  a  n  c  e  of, 
210 

technique     of     perform- 
ance of,  207 
Bauer's       modification 

of,  209 
Noguchi's  modification 

of,   208,   209 
schematic    presenta- 
tion of,  209 
refrigerator  method  in, 

208 
schematic  presentation 

of,   207 
Stern's  modification  of, 

209 
theories  of,  204,  205 


Wassermann  reaction,  titra- 
tion of  hemolytic 
amboceptor  o  r 
sensitizer  in,  205 
ultramicroscopic  method 
of  finding  precip- 
itates in  syphili- 
tic sera  in,  204 

Weichhardt's  epiphanin  re- 
action, 537 

Weigert's  law  of  overcom- 
pensation.  128 

Widal's  reaction,  220 

Wright's  method  of  stand- 
ardization of  vac- 
cines, 352,  353 

Wright's  technique  for  de- 
termination of  op- 
sonic  index,  330 

et    8€Q. 

difficulties    in,    332,    333 

value  of,  333 

Wright's  studies  of  bac- 
tericidal and  ag- 
glutinating pow- 
ers of  blood  se- 
rum, 328 

Yellow  fever,  susceptibilitv 

to,  55 
Yersin    anti-plague    serum, 

480-482 


YD  23432 


BIOLOGY 
LIBRARY 

UNIVERSITY  OF  CALIFORNIA  IJfeRARY