A depleted-mantle nodule from Kilbourne Hole, New Mexico, has been analyzed in detail to test a recent model for the formation of alkali basalts by two-component mixing of depleted mantle and enriched (kimberlite) mantle. The bulk-rock chemistry of the nodule can be reconstructed from its constituent phases, except for calcium, sodium and several incompatible elements. It is found that acid-leachates derived from the surfaces of the minerals cointain the ''missing'' complement compositions indicate that each leachate may be a mixture of nodule mineral plus kimberlite. It is suggested that the host alkali basalt lava could have formed by the same mixing process, as the lava and enclosed nodule are in approximate Nd isotopic equilibrium, and calculated mineral/melt partition coefficients for the rare earth elements (REE) between nodule minerals and host lava match those determined experimentally.