To understand the deep ocean's role in climate change, it is critical to understand how heat and CO 2 , absorbed at the surface, make their way throughout the deep ocean over decades to centuries, transported by complex current systems and modified by deep ocean mixing. This understanding should be grounded in an observational description of existing deep ocean circulation, something that is currently unavailable, except in limited areas. Given the impracticality of directly measuring the weak, time-mean deep ocean interior currents, inverse techniques have been more widely applied. Here, an inverse method is used to study the broad interior flow of Antarctic Bottom Water (AABW) in the western subtropical South Atlantic, both in the northern part of the Argentine Basin, and as it enters the broader Brazil Basin across 30°S (Figure 1). In particular, this study adds to a well-developed and somewhat contradictory discussion of circulation in this area.The inverse model flow fields offer a new perspective on two unresolved oceanographic questions. The first question (Section 3.1) is whether some of the densest AABW, flowing northward through the Vema and Hunter channels, is sourced from the eastern Argentine Basin or supplied along a deep western boundary current (DWBC) route (e.g., see the discussion in Zenk & Visbeck, 2013). Although the eastern pathway is suggested by mudwave and tracer data (Flood & Shor, 1988) observations in this area have been limited. In their overall scheme for the Argentine Basin circulation, Coles et al. (1996) suggested a deep cyclonic recirculation in the Argentine Abyssal Plain located south and west of the Zappiola Rise at nominal 45°S, 45°W (see also Wienders et al., 2000), but were unable to definitively connect it to northward flow in the eastern Argentine Basin. McDonagh et al. (2002) examined just the northern Argentine Basin. Their conclusions emphasize flow supplying the Vema Channel from the more traditional DWBC route, and they also show a band of eastward flow south of the Rio Grande Rise (RGR) at 36°W that would argue against an eastern route.