-  114 
rounded  or  rod-shaped  granules  which  stain 
deeply  when  treated  with  iron-hematoxylin 
(chromatoid  bodies  of  LÉGER  and  DUBOSQ). 
The  lines  of  cilia  are  spiral  (fig.  3a)  and 
go  from  one  side  of  the  body  to  the  otner  ; 
they  are  not  smooth  as  in  O.  Zonga  BEZ- 
ZENBERGER,  but  composed  of  minute 
granules  which  stain  very  well  with  iron-he¬ 
matoxylin.  They  are  more  or  less  equidistant 
and  1  never  saw  them  fork.  In  a  specimen 
measuring  22  p  (width)  by  48  ft  (length), 
there  were  twenty  four  rows.  The  nuclei,  of 
which  there  are  mostly  two,  are  generally  in 
the  middle.  (Fig.  3a/ 
Nuclear  Structure. 
(PI.  76  figs.  1,  8,  9,  &  9a.) 
Quiescent  state.  The  nuclei  are  more  or 
less  rounded;  when  there  is  only  one,  it 
occupies  the  middle  of  the  cell  ;  when  there 
are  two,  they  are  also  near  the  middle,  one 
behind  the  other  and  somewhat  obliquely  to 
the  longitudinal  body-axis.  (fig.  3a).  The 
nuclear  membrane  is  clearly  visible  and  so¬ 
metimes  shows  slight  elevations  and  depres¬ 
sions. 
Like  in  other  species  of  the  same  genus, 
the  nucleus  of  O.  brasiliensis  contains  ro¬ 
unded  or  elongate  masses  of  chromatin,  which 
are  generally  near  the  surface  of  the  nuclear 
membrane,  (fig.  1).  To  the  inside  of  these, 
there  are  sometimes  other  very  small  achro¬ 
matic  granules  (fig.  3  &  3a),  or  a  mass  of 
smooth  achromatic  substance,  (fig.  9  &  9a). 
Division  of  the  Nucleus. 
The  nuclear  division  of  O.  brasiliensis 
corresponds  more  or  less  to  the  paratenomi- 
tosis  of  ALEXEI EFF  (no  centrioles). 
The  first  stage  is  characterised  by  the 
stretching  of  the  nucleus,  the  disappearance 
of  the  nuclear  membrane  and  the  elongation 
of  the  chromatin  masses,  which  ether  go  to 
the  edge  or  remain  in  irregular  groups  in 
the  middle  (fig.  2). 
After  that  (fig.  3  and  lower  nucleus  of 
3a),  there  appears  a  pseudo-centrodesmosis 
which  binds  one  of  the  granules  on  one 
side  to  another  granule  on  the  opposite  side, 
while  the  nucleus  becomes  longer  and  nar¬ 
rows  at  its  equator  ('fig.  4);  this  enables  one 
to  perceive  the  pseudo-centrodesmosis  quite 
clearly. 
At  a  later  stage  (fig.  5),  the  equatorial 
stricture  becomes  more  marked  and  the 
pseudo-centrodesmosis  is  still  seen,  connec¬ 
ting  the  chromatin  mass  on  one  side  to  a 
small  granule  on  the  opposite  side.  Then  (fig. 
6)  half  the  pseudo-centrodesmosis  disappears 
and  the  rest  of  it  remains  attached  to  a 
chromatin  mass  on  one  side  of  the  nucleus 
only.  After  that,  the  stricture  gets  more  and 
more  accentuated  (fig  7)  and  the  pseudo- 
centrodesmosis  can  be  seen  no  longer,  but 
only  rounded  or  elongated  chromatin  mas¬ 
ses  are  visible  ;  they  may  be  isolated  or  ar¬ 
ranged  in  loose  groups  in  the  vicinity  of 
the  membrane. 
When  the  rounded  daughter-nuclei  are 
formed,  they  are  still  connected  by  a  slender 
filament  (fig.  8).  Later  on  the  filament  tears 
and  frees  them  (fig.  9  &  9a.). 
1  have  observed  stages  of  the  division 
of  binucleate  specimens  of  O.  brasiliensis  and 
have  seen  specimens  with  four  resting  nu¬ 
clei  (fig.  11). 
Specimens  with  onlv  one  nucleus  are 
generally  smuch  smaller  than  the  plurinuclea- 
te  forms,  a  fact  which  leads  me  to  believe 
that  they  must  be  younger  individuals. 
