Dolomite in supratidal carbonates of northwest Andros Island, Bahamas, can be seen, by means of scanning electron microscope (SEM), to be cement in primary porosity and not replacement of the carbonate sediments, as one widely held view suggests. These microcrystalline, calcian dolomite cements commonly occur in lithified lime mud crusts, but they also occur in unconsolidated sediments, mainly as intragranular precipitates. The voids in which the dolomite occurs vary in size from minute wall pores in rotaliid forams and endolithic borings to major intra- and interparticle pores. In the larger pores, dolomite commonly forms an isopachous fringe and may be followed by either aragonite or high-magnesium calcite (HMC) cement. Silt-sized bioclasts and pellets within crusts are often lithified by dolomite cement, which may totally fill much of the interparticle porosity. Within unpelleted lime mud, the interparticle porosity may also be filled by dolomite, which engulfs the needle mud particles. The needle mud inclusions or molds visible on polished, etched sections of this dolomite-filled mud are equivalent in abundance, size, and shape to the mud particles in non-dolomitic, unlithified muds embedded in plastic. That equivalence indicates that dolomite replacement of mud has not occurred. Framboidal pyrite present in a few samples appears to have formed prior to or contemporaneously with dolomite cement. The presence of pyrite and the depleted delta 13 C values (-3.2 to -3.9 per mil) of dolomite-cemented crusts suggest that sulfate reduction may have had a role in the formation of Andros Island dolomite cement. Other Holocene and older dolomites from similar settings should be reexamined by means of SEM to determine if they formed partly or wholly as a cement rather than by replacement.