Petrographic studies show that hematite is present in the Moenkopi Formation in at least five and possibly six forms: (1) microcrystalline hematite, (2) crystal of specular hematite, (3) polycrystalline and monocrystalline grains, (4) grains of partly hematitized ilmenite, (5) grains composed of primary ilmenite-hematite intergrowths, and (6) ultrafine pigment. The microcrystalline hematite and crystals of specular hematite are unequivocally authigenic. They form cement in interstitial and secondary voids, and they have replaced detrital iron-bearing silicate minerals. Furthermore, microcrystalline hematite is superimposed on other authigenic cementing minerals such as potassium feldspar, calcite, dolomite, and quartz, and in some cases it has replaced authigenic pyrite. In addition, both microcrystalline and specularite crystals are common daughter products of intrastratally altered biotite grains. Thermodynamic considerations coupled with studies of hematite-magnetic relationships in modern sediments indicate that most of the hematite in the polycrystalline grains, and probably the monocrystalline grains as well, was formed authigenically by post-depositional replacement of detrital grains of magnetite. The ilmenite probably has similarly altered in situ to hematite. The only hematite of unquestionable detrital origin in the red beds is the hematite in the ilmenite-hematite intergrowths ('tiger striped' grains) and that in monocrystalline detrital grains containing rutile exsolution platelets, both of which are products of high-temperature processes. With the exception of the ultrafine pigment, each of the above forms is coarser grained than the superparamagnetic threshold for hematite, and therefore each contributes components of remanent magnetism to the rocks. Inasmuch as most of the hematite varieties represent authigenic products of intrastratal alterations that require considerable geologic time, we conclude that the red bed remanence is largely chemical remanent magnetization (CRM) acquired over long time intervals. The pigment in the Moenkopi red beds consists partly of authigenic ultrafine red iron oxide and partly of translucent microcrystalline hematite. The ultrafine red iron oxide may or may not be hematite, but even if it is, the grain size probably lies below the paramagnetic threshold for hematite. Much of the pigment, therefore, may not contribute greatly to the remanent magnetism in the rocks.