Linkages between geology and fish assemblages have been inferred in many regions throughout the world, but no studies have yet investigated whether fish assemblages differ across geologies in Mesoamerica. The goals of our study were to: 1) compare physicochemical conditions and fish-assemblage structure across 2 geologic types in headwaters of the Monkey River Basin, Belize, and 2) describe basin-scale patterns in fish community composition and structure for the benefit of conservation efforts. We censused headwater-pool fishes by direct observation, and assessed habitat size, structure, and water chemistry to compare habitat and fish richness, diversity, evenness, and density between streams in the variably metamorphosed sedimentary geologic type typical of 80% of Belize’s Maya Mountains (the Santa Rosa Group), and an anomalous extrusive geologic formation in the same area (the Bladen Volcanic Member). We also collected species-presence data from 20 sites throughout the basin for analyses of compositional patterns from the headwaters to the top of the estuary. Thirty-nine fish species in 21 families were observed. Poeciliids were numerically dominant, making up 39% of individuals captured, followed by characins (25%), and cichlids (20%). Cichlidae was the most species-rich family (7 spp.), followed by Poeciliidae (6 spp.). Habitat size and water chemistry differed strongly between geologic types, but habitat diversity did not. Major fish-assemblage differences also were not obvious between geologies, despite a marked difference in the presence of the aquatic macrophyte, Marathrum oxycarpum (Podostemaceae), which covered 37% of the stream bottom in high-nutrient streams draining the Santa Rosa Group, and did not occur in the low-P streams draining the Bladen Volcanic Member. Correlation analyses suggested that distance from the sea and amount of cover within pools are important to fish-assemblage structure, but that differing abiotic factors may influence assemblage structure within each geologic type. The fauna showed weak compositional zonation into 3 groups (headwaters, coastal plain, and nearshore). Nearly 20% of the fish species collected have migratory life cycles (including Joturuspichardi, Agonostomusmonticola, and Gobiomorusdormitor) that use freshwater and marine habitats. Some of these species probably rely on a natural flow regime and longitudinal connectivity for reproduction and dispersal of young, and natural flow regime and longitudinal connectivity are important factors for maintenance of functional linkages between the uplands and the coast in the ridge-to-reef corridor where the Monkey River is located. Therefore, we suggest that the viability of migratory fish populations may be a good biological indicator of upland-to-estuary connectivity important both to fishes and coastal ecosystem function. We recommend follow-up studies to substantiate the relative strengths of relationships between community structure and abiotic factors in contrasting geologies and to examine potential bottom−up responses of stream biota to the higher nutrient levels that were observed in stream waters draining the Santa Rosa Group geologic type.