Abstract. The preferential heating and acceleration of O +s ions, as observed by Ultraviolet Coronagraph Spectrometer (UVCS) on Solar and HeliosphericObservatory (SOHO) [Kohlet al., 1998] in the solar coronal holes have been interpreted and modeled by invoking wave-particle cyclotron resonance [Cranmer et al., 1999a[Cranmer et al., , 1999b. However, in the model of Cranmer et al. [1999a, 1999b] and in other subsequent models the assumption of a rigid slope of the wave spectrum was made in calculating the wave energy absoftion by the different ion species. In the present paper it is shown that a self-consistent treatment of the wave damping and absorption is necessary and leads to substantially different results. On the basis of quasi-linear theory, the interaction between the ions and the ion-cyclotron waves [Marsch et al., 1982a;Marsch, 1998] is studied. The total energy conservation equation, including the kinetic energy of the resonant particles and the wave energy, is derived and discussed in detail. The spectral evolution equation for cyclotron waves, when being controlled by the wave growth/damping rate and WKB effects, is solved self-consistently together with the full set of anisotropic multifluid equations for the ions including the cyclotron-resonance wave heating and acceleration rates. From the numerical results we reach the following conclusions: (1) It is physically questionable to use a spectrum with a fixed spectral slope near the cyclotron resonance when one calculates the partition of wave energy among the different ionic species and the kinetic degrees of freedom parallel and perpendicular to the magnetic field. This assumption neglects the important effects of wave absorption and the concurrent reshaping of the wave spectrum, and thus leads in the dissipation domain to extremely low amplitudes of the waves and to difficulties in supplying enough energy to balance the wave absorption at the cyclotron resonances. (2) If the spectrum is allowed to evolve self-consistently and concurrently with the particles' heating and acceleration through wave absorption, such a high perpendiculax temperature and corresponding large temperature anisotropy as observed by UVCS do not occur or cannot be maintained. We conclude that the UVCS oxygen ion observations have not yet been explained satisfactorily by the cyclotron-resonance theory.