This paper studies the ability of salt diapirs to lift large inclusions of dense rocks (rafts). Dense inclusions will be lifted if salt in the diapir rises faster than the inclusions sink. The power-law rheologies of six different salts, and viscosities estimated for Newtonian salt are used to calculate the settling velocity of these inclusions as a function of their radii and densiy as well as the temperature of the salt. This is done using the known equation describing the velocity of solid spheres settling in unbounded power-law fluid. Two-dimensional numerical models are used to study he effect of the ellipticity of inclusions on their sinking velocity, as a function of the power-law exponent n. The calculations of the settling velocities of inclusions neglect several other factors discussed in the paper: 1. (a) the ambient fluid (salt) is finite (bounded); 2. (b) more than one inclusion translates in the salt; 3. (c) further strain-rate softening of salt is caused by its diapiric ascent. The results suggest that, whereas Newtonian salt of 1017-1018 Pa s and the power-law salt of the Vacherie Dome would have to rise at unreasonably high speeds in order to lift large inclusions, most power-law salt diapirs would be capable of lifting inclusions of the sizes observed in the Iranian domes (up to 3-6 km2) if these rise at geologically reasonable velocities and temperatures.