The positronium (Ps) yield for ice, measured by Eldrup et al. using a low-energy positron beam, is discussed in terms of the spur model of Ps formation. The pronounced maxima in the Ps yield for crystalline ice at positron energies below 65 eV are well explained by effects due to energy conservation in the spur processes. Parts of the amorphous ice results are well explained by the spur but not by the Ore model. Important processes influencing the Ps formation are not included in the Ore model.