360 On the Error of Countmg with a Haemacytometer 
It must be noted, however, that tlie probable error will always be greater 
than that calculated on this formula when for any reason the organisms occur 
as aggregates of varying size. 
In conclusion, I should like to thank Prof. Adrian J. Brown, of Birmingham 
University, for his valuable advice and assistance in carrying out the experimental 
part of the enquiry. 
TABLE II. 
" Centre " Squares. 
a; 
< 
1 
o 
Jt 
5 
6 
8 
9 
10 
11 
12 
Totals 
1 
6 
6 
9 
15 
15 
9 
4 
3 
2 
69 
9 
6 
14 
17 
31 
24 
17 
10 
5 
6 
2 
1 
1 
1.34 
8 
15 
25 
32 
37 
20 
15 
7 
7 
1 
4 
171 
4 
18 
34 
33 
45 
48 
41 
22 
7 
5 
4 
1 
258 
5 
15 
24 
37 
47 
39 
37 
18 
12 
11 
4 
1 
2 
247 
6 
9 
17 
25 
39 
34 
32 
14 
8 
2 
4 
1 
1 
186 
'/ 
5 
12 
14 
21 
19 
16 
9 
7 
3 
106 
8 
3 
5 
7 
8 
12 
8 
6 
1 
3 
4 
57 
9 
2 
G 
7 
5 
10 
2 
2 
3 
1 
38 
10 
1 
1 
4 
4 
4 
3 
1 
18 
11 
1 
4 
1 
1 
1 
8 
12 
1 
1 
1 
1 
4 
Totals 
72 
136 
180 
248 
244 
188 
100 
56 
40 
20 
8 
4 
1296 
Mean of "Centre" Squares, 4-6821 ; S. D., 2-1,39. 
Mean of "Adjacent" Squares, 4-7014; S. D., 2-116. 
r = +-01 6 ±-037. 
Correlation table between the number of cells in a square and the numbers of cells in the 
four adjacent squares taken all over Table I. 
