We report new helium and carbon isotope (He-3/He-4 and delta C-13) and relative abundance (CO2/He-3) characteristics of a suite of 20 gases and fluids (cold mazuku-like CO2 vents, bubbling mud-pots, hot and cold springs) from 11 different localities in Rungwe Volcanic Province (RVP), southern Tanzania and from 3 additional localities in northern Tanzania (Oldoinyo Lengai Volcano and Lake Natron). At RVP, fluids and gases are characterized by a large range in He-isotope compositions (He-3/He-4) from 0.97 R-A to 7.18 R-A (where R-A = air He-3/He-4), a narrow range in delta C-13 ratios from -2.8 to -65 parts per thousand (versus VPDB), and a large range in CO2/He-3 values spanning nearly four orders of magnitude (4x10(9) to 3.2x10(13)). Oldoinyo Lengai possesses upper-mantle-like He-CO2 characteristics, as reported previously (Fischer et al., 2009), whereas hot springs at Lake Natron have low He-3/He-4 (similar to 0.6 R-A), CO2/He-3 (similar to 5-15x10(8)) and intermediate delta C-13 (similar to-3.7 to -4.9 parts per thousand). At RVP, fluid phase samples have been modified by the complicating effects of hydrothermal phase-separation, producing CO2/He-3 and delta C-13 values higher than postulated starting compositions. In contrast, gas-phase samples have not been similarly affected and thus retain more mantle-like CO2/He-3 and delta C-13 values. However, the addition of crustal volatiles, particularly radiogenic helium from He-4-rich reservoir rocks, has modified He-3/He-4 values at all but the three cold CO2 gas vent (i.e., mazuku) localities (Ikama Village, Kibila Cold Vent and Kiejo Cold Vent) which retain pristine upper-mantle He-isotope (similar to 7 R-A) and He-CO2 characteristics. The extent of crustal contamination is controlled by the degree of interaction within the hydrothermal system, which increases with distance from each major volcanic center. In contrast, we propose that pristine cold CO2 mazuku gases collected at stratigraphic contacts on the flanks of RVP volcanoes may potentially tap isolated gas pockets, which formed during previous eruptive events and have remained decoupled from the local hydrothermal system. Furthermore, by identifying and utilizing unmodified gas samples, we determine mantle versus crustal provenance of the CO2, which we use to estimate mantle-derived CO2 fluxes at both Rungwe and Lake Natron. Finally, we investigate the origin of the apparent discrepancy in He isotopes between fluids/gases and mafic phenocrysts at RVP (from Hilton et al., 2011), and discuss the tectonic (i.e., rift zone dynamics) and petrogenic conditions that distinguish RVP from other plume-related subaerial rift zones. (C) 2012 Elsevier B.V. All rights reserved.