site  explains  the  localisation  of  lymphatic  le¬ 
sions.  The  fungus,  on  penetrating  either  by 
the  intestine  or  by  the  trachea  and  the  bron¬ 
chi,  finds  in  the  mesenteric  or  tracheobron¬ 
chial  glands  the  first  bar  to  its  invasion 
of  the  organism.  Intraperitoneal  inoculation 
of  cultures,  killed  by  heat,  are  fatal  to  rabbits, 
especially  when  large  doses  are  used.  For 
living  cultures,  any  way  of  infection  gives 
very  good  results  in  the  rabbit.  This  patho¬ 
genic  power  is  sometimes  of  great  value  for  dif¬ 
ferential  diagnosis.  The  tendency  of  the  para¬ 
site  to  pulmonary  localisation  is  remarkable 
in  the  rabbit.  We  always  succeed  in  finding 
the  fungus  in  the  lungs,  when  it  is  impossi¬ 
ble  to  detect  it  in  any  other  organ.  The  intra¬ 
muscular  inoculation  often  causes  lesions  li¬ 
mited  to  the  lungs. 
Mice  also  are  susceptible  to  infections 
with  Oïdium  brasiliense.  Inoculation  at  the 
root  of  the  tail  or  by  the  peritoneum  quickly  | 
kills  the  animal.  Sometimes  we  observe  ef¬ 
fusion  in  the  serous  cavities,  mostly  in  one 
of  them.  The  spleen  may  attain  large  pro-  ' 
portions.  In  the  second  passage  death 
insues  quicker. 
The  guinea-pig  is  less  susceptible  than 
the  above  cited  animals,  but  is  useful  fordi- 
ferential  diagnosis  with  tuberculosis,  caused 
by  KOCH’s  bacillus.  The  evolution  of  the  ex¬ 
perimental  infection  may  vary  from  4  days 
to  6  months.  We  often  met  with  general 
serous  effusions.  Spleen  and  liver  are  almost 
always  increased  in  size.  The  suprarenal  cap¬ 
sules  are  swollen,  and  the  kidneys  are  af¬ 
fected  by  \arious  forms  of  degeneration. 
Polyadenitis  is  the  rule.  The  mediastinal 
and  the  higher  tracheo-bronchial  glands  ra¬ 
rely  escape,  principally  after  infection  by  the 
tracheal  and  bucco-nasal  mucosa,  and  often 
attain  a  large  size.  The  subcutaneous  inocula¬ 
tions  of  the  abdominal  wall  leads  to  lateral 
lymphadenitis  of  the  abdomen,  and  this  form 
of  adenitis  must  be  remembered,  as  it  is  the 
rule  in  the  tuberculous  guinea-pig  fKOCH’s 
bacillus).  In  the  infection  with  Oidium  bra¬ 
siliense ,  smears,  triturations,  inoculations  etc., 
made  from  such  ganglions,  never  demonstra¬ 
te  KOCH’s  bacillus.  We  tried  to  find  this 
bacillus  by  all  means  actually  known.  These 
researches  were  repeated  not  only  for  the 
ganglions,  but  also  for  all  the  organs  of  any 
infected  animal  found  dead. 
Lungs  of  guinea-pigs,  killed  by  Oidium 
brasiliense,  are  congested.  We  never  found 
miliary  abscesses,  except  in  septicemic 
forms.  The  evolution  after  inoculations 
with  suspect  sputum  is  variable.  If  the 
flora  of  common  microbes  is  poor,  the 
area  of  inoculation,  slightly  swollen  and  red 
in  the  first  days,  returns  soon  to  norma! 
conditions,  without  any  trace  remaining.  In 
sputa  with  abundant  flora  abscesses  are  for¬ 
med.  If  the  animals  resist,  cicatrisation  follows 
quickly.  The  torpid  ulcer  of  tuberculosis,  due 
to  KOCH’s  bacillus,  which  heals  with  difficul¬ 
ty  or  not  at  all,  is  never  seen.  In  the 
guinea-pig  too  we  observe  the  tendency  of 
the  parasite  to  pulmonary  localisation.  Intra¬ 
muscular  inoculation  sometimes  produced  an 
infection  limited  to  the  lungs. 
Resuming,  we  see  that  in  all  the  inocu¬ 
lations  of  animals  we  obtained  encouraging 
results,  the  more  as  we  completed  PASTEUR's 
cycle  by  recovering  the  parasite  from  ani¬ 
mals,  killed  by  the  experimental  disease. 
What  we  especially  want  to  insist  upon,  is 
the  reproduction  in  animals  of  the  symptoms 
and  all  the  anatomical  lesions,  observed  in 
patients  of  the  spontaneous  disease. 
The  research  of  Oidium  brasiliense  in 
smears  of  the  organs  of  the  animals  is  rela¬ 
tively  easy.  Preparations  mostly  team  with 
fungus  elements.  The  morphology  here  is 
various.  The  yeast-forms  prevail  in  smears 
of  the  organs,  though  mixed  forms  are  com¬ 
mon.  Filtration  in  BERKEFELD  filter  sterili¬ 
ses  the  cultures  of  the  Oidium  brasi¬ 
liense.  Filtered  cultures  do  not  grow  nor 
infect.  Heating  to  56°  for  1  hour  kills  the 
fungus.  Cultures,  sterilised  in  this  way,  are 
harmless,  when  injected  subcutaneously.  While 
hyper-immunizing  rabbits  for  serological 
purposes  we  arrived  at  4  injections  of  20 
ccm.  with  intervals  of  4  days,  the  animal 
only  showing  slight  emaciation.  The  same 
