The U.S. Geological Survey's Water, Energy, and Biogeochemical Budgets (WEBB) program research in eastern Puerto Rico involves a double pair-wise comparison of four montane river basins, two on granitic bedrock and two on fine-grained volcaniclastic bedrock; for each rock type, one is forested and the other is developed. A confounding factor in this comparison is that the developed watersheds are substantially drier than the forested (runoff of 900–1,600 millimeters per year compared with 2,800–3,700 millimeters per year). To reduce the effects of contrasting runoff, the relation between annual runoff and annual constituent yield were used to estimate mean-annual yields at a common, intermediate mean-annual runoff of 1,860 millimeters per year. Upon projection to this intermediate runoff, the ranges of mean-annual yields among all watersheds became more compact or did not substantially change for dissolved bedrock, sodium, silica, chloride, dissolved organic carbon, and calcium. These constituents are the primary indicators of chemical weathering, biological activity on the landscape, and atmospheric inputs; the narrow ranges indicate little preferential influence by either geology or land cover. The projected yields of biologically active constituents (potassium, nitrate, ammonium ion, phosphate), and particulate constituents (suspended bedrock and particulate organic carbon) were considerably greater for developed landscapes compared with forested watersheds, consistent with the known effects of land clearing and human waste inputs. Equilibrium rates of combined chemical and physical weathering were estimated by using a method based on concentrations of silicon and sodium in bedrock, river-borne solids, and river-borne solutes. The observed rates of landscape denudation greatly exceed rates expected for a dynamic equilibrium, except possibly for the forested watershed on volcaniclastic rock. Deforestation and agriculture can explain the accelerated physical erosion in the two developed watersheds. Because there has been no appreciable deforestation, something else, possibly climate or forest-quality change, must explain the accelerated erosion in the forested watersheds on granitic rocks. Particulate organic carbon yields are closely linked to sediment yields. This relation implies that much of the particulate organic carbon transport in the four rivers is being caused by this enhanced erosion aided by landslides and fast carbon recovery. The increase in particulate organic carbon yields over equilibrium is estimated to range from 300 kilomoles per square kilometer per year (6 metric tons carbon per square kilometer per year) to 1,700 kilomoles per square kilometer per year (22 metric tons carbon per square kilometer per year) and is consistent with human-accelerated particulate-organic-carbon erosion and burial observed globally. There is no strong evidence of human perturbation of silicate weathering in the four study watersheds, and differences in dissolved inorganic carbon are consistent with watershed geology. Although dissolved organic carbon is slightly elevated in the developed watersheds, that elevation is not enough to unambiguously demonstrate human causes; more work is needed. Accordingly, the dissolved organic carbon and dissolved inorganic carbon yields of tropical rivers, although large, are of secondary importance in the study of the anthropgenically perturbed carbon cycle.