The structure and composition of rhodoliths in two regions of the Brazilian shelf, Abrolhos Continental Shelf (ACS) and South Esp´ırito Santo State (SES) were examined and compared. Rhodoliths were sampled at depth ranges of 10-20 m and 50-60 m in SES, and 20-30 m and 50-75 m in ACS. Rhodoliths in SES are algal boundstones, built mainly of melobesioid corallines, with subordinate bryozoans and encrusting foraminifers. They show high porosity and the sediment infill of borings and voids contains a relatively high amount of siliciclastics (up to 8%). Rhodoliths from ACS are formed by a structureless carbonate mass covered by a thin veneer of encrusting coralline algae. The massive interior was produced by multiphase boring and infilling of an original boundstone. The infillings consist of a micritic matrix with bioclasts and low amounts of siliciclastic grains. Coralline assemblages are reduced to fragments of thin crusts. Rhodoliths from shallow depths are small (, 8 cm) whereas rhodoliths from the deeper zone have a wide size range (1 to 17 cm). The innermost parts of deeper rhodoliths in ACS yield radiocarbon ages of ~ 7,000 years BP (75 m) and ~ 2,000 years BP (65 m). Rhodoliths from the deep zone in SES are younger (less than 700 years BP). Siliciclastic sediment influx reaching rhodolith beds promotes burial of rhodoliths, determining the small size of shallow SES rhodoliths and the relatively young ages of the deeper ones. Scarce siliciclastic influx at the ACS rhodolith beds favors long residence times of rhodoliths on the seafloor, resulting in a thorough destruction of the original coralline/invertebrate boundstone by bioerosion.Despite the increasing knowledge about Brazilian rhodoliths in the last decade, little is known about their inner arrangement and composition (Leal et al. 2012; Tamega et al. 2014; Vale et al. 2018). The taxonomic composition, morphology, inner structure, and interactions of coralline algae with other encrusting groups (such as bryozoans, corals, and serpulids) are relevant for understanding the factors that control long-term development of rhodoliths (Adey and MacIntyre 1973; Aguirre et al. 2012) and to reconstruct the paleoenvironments in which they grew (Bosence 1983a;