Indian Ocean ridges north of the Rodriguez Triple Junction remain poorly explored for seafloor hydrothermal activity, with only two active sites confirmed north of 25°S. We conducted water column surveys and sampling in 2007 and 2009 to search for hydrothermal plumes over a segment of the Carlsberg Ridge. Here we report evidence for two separate vent fields, one near 3°42′N, 63°40′E and another near 3°41.5′N, 63°50′E, on a segment that is apparently sparsely magmatic. Both sites appear to be located on off-axis highs at the top of the southern axial valley wall, at depths of ∼3600 m or shallower (∼1000 m above the valley floor). At the 63°40′E site, plume sampling found local maxima in light scattering, temperature anomaly, oxidation-reduction potential (ORP), dissolved Mn, and3He. No water samples are available from the 63°50′E site, but it showed robust light-scattering and ORP anomalies at multiple depths, implying multiple sources. ORP anomalies are very short-lived, so the strong signals at both sites suggest that fluid sources lie within a few kilometers or less from the plume sampling locations. Although ultramafic rocks have been recovered near these sites, the light-scattering and dissolved Mn anomalies imply that the plumes do not arise from a system driven solely by exothermic serpentinization (e.g., Lost City). Instead, the source fluids may be a product of both ultramafic and basaltic/gabbroic fluid-rock interaction, similar to the Rainbow and Logatchev fields on the Mid-Atlantic Ridge.