The  mode  of  calculating  K  adopted  ui  Table  13,  namely  by 
determining  K  for  a  series  of  periods,  each  of  which  commences  at  the 
same  point  of  time,  that  is  to  say,  the  time  at  which  the  experiment 
is  begun,  is  well  adapted  for  showing  progressive  variations  in  the 
values  of  K,  since  in  this  way  the  irregularities  due  to  error  of  estima¬ 
tion  become  less  accentuated.  The  latter  are  greatest  when  K  is 
calculated  for  every  pair  of  successive  estimations,  as  is  done  in 
Table  14,  where  the  values  obtained  present  greater  irregularities  than 
in  Table  13.  The  method  of  calculation  adopted  in  Table  14  gives  a 
more  correct  average  for  the  period  t^  to  t,-,  than  is  obtainable  by  the 
method  employed  in  Fable  13.  In  Table  13  the  percentages  of  red 
cells  haemolysed  are  calculated  from  the  average  value  of  K  between 
Tabt-F-  Is-  Comparison  of  percentages  found  in  Ta'ble  13  and  those  calculated  from  the 
value  of  K  —0*0031}. 
Time 
'11  +  l-<n 
(^1 4 1  — 
/n) 
Concentration  of 
wet  red  cells 
Calculated 
Found 
i(i 
0  min. 
[187-21 
poo] 
;o  „ 
^0  min. 
0*0039  X  30 
=o*i9;=|og  T 
•37='"? 
187-2 
1  TC)-2 
[1 19-2] 
I90-8] 
So  „ 
30 
3° 
=o-ii7=loK 
•31  =log 
119*  2 
91-2 
|c)I-2l 
I82-0] 
1  10  ,, 
.30 
=lQg 
69-3 
7''7 
'.i 
14°  .1 
30 
.. 
•• 
=  log 
69-3 
33-1 
53-' 
54-2 
'5 
170 
30  „ 
1*  11 
•• 
=  log 
ill' 
40-5 
40-5 
.39-5 
200  ,, 
.30  1, 
’1  11 
11 
=  log 
40- 5 
30-9 
30-9 
30-7 
'7 
23°  .. 
30  .. 
11  ij 
11 
=  log 
3°-Q 
23-6 
23-6 
2i*r 
^8 
160  „ 
3*^  51 
11  11 
>1 
=log 
i8*i 
[.8-,] 
[6-1] 
t^  and  t.j  obtained  m  Table  14.  It  is  seen  that  within  this  period  there 
is  a  close  correspondence  between  the  calculated  and  observed 
percentages.  A  comparison  of  the  calculated  and  observed  results, 
