For each of 200 samples, each sample has 10000 variates, we take an overlapping
pairs of numbers. The N-1 points (
), i=1,...,N-1 were
histogrammed into k
sub squares with k=16, so for each from 16subsquares
we have number n
,i=1..16,j=1..16, which equals number of occurencies
of pairs in those square. the following quantities were formed,
e
c
=
and
The statistic w is defined by
t
for 
=k(k-2)+1=
225 degrees of freedom.