We present a detailed theoretical analysis of experimental rates for d
dmu molecular formation and dmu hyperfine transitions at temperatures
25.5-150 K, which were reported by Zmeskal et al. [Phys. Rev. A 42, 11
65 (1990)]. Theoretical effective ddmu formation rates are fitted to t
he observed rates by adjusting the ddmu binding energy epsilon11, the
effective dd fusion rate lambda(f), and the nonresonant ddmu formation
rate lambda(nr). The value of epsilon11 = -1966.1+/-0.2 meV is determ
ined with extreme accuracy and agrees with the theoretical prediction
within 0.1 meV. Experimental findings for lambda(nr), are compatible w
ith theory. Since the value of lambda(f) extracted from observed forma
tion rates depends on the calculated value of ddmu formation matrix el
ements \V(if)\, we present the region of pairs (lambda(f), \V(if)\) al
lowed by experiment. The theoretical values of lambda(f) and \V(if)\ l
ie outside this region. A significant discrepancy remains for the dmu
hyperfine transitions, where the theoretical rates, which consist of s
cattering and back-decay contributions, exceed experimental rates by a
pproximately 40%. Fits of the experimental data indicate that mostly t
he scattering contribution is smaller than calculated. The extrapolati
on of our fit to higher temperatures is in good agreement with other e
xperiments on ddmu formation.