We report new helium isotope analyses of olivine and clinopyroxene separates from recent lavas for eleven volcanoes from Flores in the east Sunda arc through the inactive segment between the arcs to Banda Island at the extreme of the contiguous Banda arc. In the east Sunda arc, 3He 4He ratios (R) vary between 4.5 RA (RA = air 3He 4He) for the leucitic Batu Tara volcano to a remarkable low of pure radiogenic helium (0.0075 RA) for Werung at the southern tip of Lomblen Island. Lavas from the inactive zone, which represents the locus of collision of the Australian continent with the arc, have a narrower range in R RA-from 3.9 for Kisu in the straits of Pantar to 1.0 for Romang Island. Our one locality (Banda Island) in the Banda arc gives the highest R RA ratio (3.1) observed along this arc to date. The results are consistent with the involvement of crustal material in magma genesis throughout the east Sunda/Banda arcs, as far west as Iya in central Flores. We combine these helium isotope results with published and on-going strontium isotope studies, and show that the source of the helium in the crustal component is unlikely to be terrigenous sediments derived from the Australian continent; rather, degassing of Australian continental crust appears to be the dominant process controlling the helium budget. The He-Sr isotope systematics also provide a framework to account for the areal pattern of 3He 4He in this part of the arcs: the radiogenic crustal component is diluted with mantle helium both in a down-dip direction and with increasing lateral distance from the collision zone. These factors result in an excellent first-order relationship between the 3He 4He ratio, degree of He Sr enrichment (relative to the postulated mantle endmember), and alkalinity of the erupted lavas. Such a relationship has a direct bearing on models of the tectonic evolution of the collision zone, and on the observation that helium isotopes are decoupled from strontium and other geochemical signatures along the Banda arc.