The ratio of two chi-squared variables divided by their respective degrees of
freedom follows's Snedecor's F distribution, which was named in honor of
Fisher, but AFAIK was never used by Fisher. The Excel functions FDIST and
FINV support Snedecor's F distribution. Note that FDIST calculates the upper
tail (one minus the cdf) of the distribution and FINV is its inverse.

What Fisher used (and still included in the final posthumously publised
edition of Fisher & Yates "Statistical Tables") was the distribution of Z =
LN(F)/2, which is related by
cdf_z(z,df1,df2) = FDIST(EXP(2*z),df1,df2)
inv_z(p,df1,df2) = LN(FINV(1-p,df1,df2))/2

Jerry

"Vincent Isoz" wrote:

Thanks for your answers. But the tree functions that you have listed are not
the real Fisher functions as we know in Statistics: the ration of two khi-2
variable divided by their respective degree of freedom.