Muon catalyzed fusion represents a potential long-term approach to fusion power. The sensitivity of the fusion properties of muomolecules to the choice of wave function describing the system is studied. The need for an improved understanding of the problem and correct evaluation of nonadiabatic effects is stressed, and the discrepancy between fusion rates from Born-Oppenheimer- and Hylleraas-type wave functions is pointed out.