In which of the following test statistics are degrees of freedom used in both the numerator and the denominator (a presumptive test)?

The F statistic uses degrees of freedom in both numerator and denominator.

F is the ratio of two variances, F = s1^2 / s2^2.

The numerator variance carries its own df.

The denominator variance carries a separate df.

So the F distribution is fixed by a pair (df1, df2).

The t, Z and chi-square tests use only a single df set.