36 
Presumably  dependent  on  the  same  cause  are  the  variations  in  the 
values  of  K  observable  in  the  individual  experiments  in  Table  17. 
These  are  greater  than  those  in  Table  16,  ranging  at  37°  C.  between 
0'0030  and  0‘0I02  (or,  when  recalculated  from  the  mean  value  of  n, 
between  0-0026  and  0-0087).  Table  17  shows  that,  in  each  emulsion 
these  individual  variations  characterise  the  reaction  rate  at  all 
temperatures  employed  ;  thus  if  the  rate  is  rapid  at  37°  C.,  it  is  also 
rapid  at  31°  C.  and  at  25°  C.,  and  if  relatively  slow  at  the  first 
temperature  it  is  also  relatively  slow  at  the  lower  temperatures.  This 
relationship  is  seen  at  a  glance  in  Table  ig,  which  illustrates  the 
necessity  of  making  simultaneous  observations  of  the  reaction  rate  of 
the  same  emulsion  at  different  temperatures  in  order  to  determine  the 
value  of  n  ;  obviously  values  of  K  taken  at  random  from  different 
experiments  are  of  no  value  for  this  purpose.  It  will  be  noticed  also 
that,  owing  to  the  large  number  of  observations  which  require  to  be 
made,  and  the  rapidity  with  which  reaction  proceeds  at  37°  C.,  some 
difficulty  IS  experienced  in  obtaining  a  sufficient  number  of  estima¬ 
tions  lying  between  80  per  cent,  and  20  per  cent,  at  this  temperature. 
The  values  lying  beyond  these  limits,  which  are  neglected  in 
calculating  the  mean  value  of  K,  are  indicated,  in  the  Table,  with  an 
asterisk.  In  Table  18  the  initial  percentages  of  red  cells  in  the 
different  experiments  are  given ;  no  definite  relation  between  the 
concentration  of  red  cells  employed  and  the  values  of  K  or  fi  is  to  be 
observed. 
Table  19.  Comparison  of  values  of  K  at  different  temperatures. 
No.  of  Experiment. 
Calculated  fo 
r/2  = 
15000 
Mean 
T 
I 
2 
3 
4 
I 
2 
3 
4 
37°  C. 
0*0078 
0*007 1 
0*0102 
0*0030 
0*0074 
0*007  * 
0-0087 
0-0026 
0-0065 
31°  c. 
0*0043 
0*0044 
0*0049 
0*0015 
0*0046 
0*0044 
0*0054 
0*0016 
0*0040 
25°  c- 
0*0032 
0*0040 
0*00095 
0-0028 
0*0027 
0-0033 
0*0010 
0*0024 
In  Table  18  the  value  of  fx  is  determined,  separately  for  each 
experiment,  from  the  data  given  in  Table  17,  by  the  aid  of  the 
formula  given  on  p.  33.  In  each  experiment  three  values  of  n  are 
obtained,  the  mean  of  which  is  taken  as  the  true  value.  Although 
the  mean  values  show  some  variation,  being  highest  in  Experiment  4, 
