( 1298 ) 
8 
For n==d and s= 5 numerator and denominator are equal to 0, 
but this case supposes b—=b, For s= 3.78 and f=—= 7 we should 
find : 
108 X22 
7 7 
nt1= 5 = 5.34 
1 — 1.08 « 7 
or 
n — 4.34 
For the determination of 7 we have the equations: 
b 1 a s 
bgr iS x 
or 
a 
| ee à 
/ S 
EEN ee 
r Wf. 
or 
db 
rr 8 Aen 8 a dv 
oo f gE pn] i n 
or 
28(f—I) 
1 a 
ee : 
7 ii n 
For s= 3.78 and f=7 and n= 4.34, we find: 
1 
S= 0,46 + 0,01713 = 0.47713 
fe 
or 
pr == 2,099". 
And this value of 7 is, indeed, smaller than the estimation in my 
“(Quasi association”, but only very little. 
On the supposition that sr should always be equal to 8, we should 
find r= 2,116 — so that the difference would hardly amount to 1 °/,. 
Hence we find sr< 8, as was demonstrated above, but only little 
smaller, viz. 7,9217. And for (f—1)r? we do not find exactly 27, 
but a slightly smaller value, viz. 26,352. But the question what is, 
after all, the cause of the variability of 6, is not answered yet, and 
