First-order reversal curve (FORC) diagrams for mixtures of different magnetic phases and bimodal distributions have been measured to examine the efficiency of the FORC method at unraveling complex magnetic signals. The FORC distributions for various magnetic minerals, including magnetite, maghemite, hematite, and goethite, and their linear additivity are assessed. Mixtures containing only hard magnetic minerals like hematite or goethite, which have relatively small spontaneous magnetizations (MS) and large magnetocrystalline anisotropies, can be adequately described by a linear addition of the two end-members, because there are virtually no magnetostatic interactions between the phases. Mixtures dominated by softer minerals like magnetite and maghemite are more susceptible to interactions and exhibit nonlinear behavior. When a hard phase with low MS like hematite is mixed with a softer phase with high MS like magnetite, it can still be identified using the FORC technique, whereas it is impossible to do so using standard magnetic hysteresis measurements. When the weaker phase can be identified, then weak-strong mixes add linearly; however, beyond a certain critical concentration the mineral with high MS swamps the magnetic signal and linearity breaks down. It is suggested that the FORC method is highly suitable for identifying small traces of hard magnetic minerals like hematite and goethite in the presence of minerals with high MS such as magnetite.