Lucy Whitaker 
51 
II. Suicides of Women in Eight Gervian States. Bortkewitsch gives the 
following table : 
TABLE VI. 
State 
Number of Suicides of Women per Year 
8 9 10 
Totals 
(a) Schaumburg-Lippe 
(b) Waldeck 
(c) Liibeck 
(d) Reuss ji. L. ... 
(e) Lijjpe 
(/) Schwarzburg-Rudolstadt ... 
(.17) Meeklenburg-Strelitz 
(A) Schwarz Ijurg-Soudei'hausen 
14 
14 
14 
14 
14 
14 
14 
14 
Totals 
19 
17 
20 
15 
11 
112 
The resulting binomials are : 
(a) 
(h) 
(c) 
id) 
(e) 
if) 
(9) 
(h) 
14 ( -9714 + 
14 ( -8.571 + 
14 ( -.5819 + 
14 (1-0058- 
14(1-3929 - 
14 ( -0071 + 
14(l-.5792 - 
14 (1-6609 - 
■0286y'»'<'»-^. 
.1429)io-49.« 
.4181)8-1503, 
•0058)-«''-2«^ 
.3929)-7-27i7_ 
.3929)13-0909^ 
.5792)-9-12C7^ 
-6609 )-«■■«'". 
Thus it will be seen that of the eight binomials only four have a positive q, 
and of these only one can be said to have a very sm;ijl q, and even in this case the 
?! is not indefinitely large. Of the four negative binomials three have quite 
substantial q's, and the fourth with its small negative q corresponds most closely 
to the Poisson-Exponential. The probable error of q for q = 0 is ± -2.549. The 
number, 14, of cases taken is therefore wholly inadequate to test whether the 
Poisson-Exponential may be applied to these data. The mean value of q is 
negative and =--0820 ± -0901, and the standard deviation of g = -3928 + -0637, 
which are within the limits of random sampling of 5' = 0 with a standard deviation 
of '3779. We shall return to a different manner of considering the point later. 
At present we wish only to indicate that the hypothesis is that q is a, very small 
positive quantity and that data which give q a standard deviation of -3928, or in 
the next example of -4714 are really inadequate to test such a hypothesis ; for in 
the resulting binomials q may easily lie anywhere between + -8 and - -8, and it 
is not possible to demonstrate that its real value is practically an exceeding small 
positive quantity. 
