OF ERRORS OF JUDGMENT AND ON THE PERSONAL EQUATION. 
‘273 
Table XVI. —Frequency of Absolute and Relative Errors referred to Length of Line 
as Unit in Bisection Experiments.^ Number = 500. 
Size of error. 
(1-) 
(2.) 
(3.) 
(2-3.) 
(3-1.) 
(1-2.) 
- -12 
1 - -11 
- -10 
_ 
— 
— 
1 
— 
2 
1 
3 
- -09 
1 
— 
— 
2 
— 
7 
- -08 
4 
1 
— 
1 
1 
9 
i - -07 
8-5 
3 
1 
6 
2 
23-5 
! - -06 
12 
11 
7-5 
13 
6-5 
21-5 
- -05 
13-5 
14-5 
9-5 
16-5 
14-5 
35 
- -04 
45 
21-5 
22 
26 
10 
58 
- -03 
61 
30 
40-5 
45-5 
26 
45 
- -02 
76 
47 
43-5 
36 
43 
59 
- -01 
90-5 
51-5 
51 
66 
33 
64-5 
•00 
74-5 
72 
68-5 
55-5 
42 
43 
I + -01 
50 
65'5 
75 
61 
53-5 
37-5 
+ -02 
30-5 
53 
70-5 
52-5 
60-5 
26-5 
+ -03 
21-5 
50-5 
61 
41-5 
61 
27-5 
+ -04 
7 
28-5 
25-5 
34-5 
52 
16 
+ -05 
3 
27 
13-5 
: 20 
34-5 
7 
+ -06 
2 
13-5 
10 
13 
27 
11 
+ -07 
— 
7-5 
1 
4 
17-5 
1 
+ -08 
— 
— 
— 
2 
5 
1 
+ -09 
— 
1 
— 
2 
7-5 
1 
+ -10 
— 
— 
— 
1 
3-5 
— 
+ -11 
— 
2 
— 
— 
no intention of considering the problem ot normal frequency ; they were designed to 
demonstrate what appeared a remarkable and valuable result flowing from the 
theory of errors as usually expounded (see p. 240). Each of us made our individual 
judgments with care and without any theoretical bias. We were of course, during 
the work of the observations, liable to physical and psychological influences, to the 
subtle changes of daily health and of sense-keenness. But I contend tliat all such 
things affect every observer, and that it is idle to propound a theory which would 
hold for an ideal observer of perfectly equable temjDerament and physical fitness 
observing under a perfectly equable environment for a number of days or even weeks 
an exactly identical phenomenon. Such a theory could not be verified, and if 
verified would have no practical application. Our observations seem to me a 
perfectly fair sample of actual errors of judgment, and I believe no objections can be 
taken to them which would not apply with even increased force against most of the 
series of errors of judgment with which physicists or astronomers have to deal. 
There are two classes of considerations which arise in our view of frequency distri 
butions :— 
* In this table and in the diagrams X. to XV. the error has been given the opposite sign to its value 
in Table I. 
2 N 
VOL. CXCVIII.—A. 
