Iron in surface waters of the Southern Ocean is primarily (99%) derived from upwelling [Lefevre and Watson, 1999]. While recent iron fertilization studies have shown that this element limits phytoplankton growth in those waters [Boyd et al., 2000], the ultimate origin of that iron has not been established [Johnson, 2001]. This unknown origin is of concern, because resolving the source(s) of up-welled iron is critical for understanding the relationship between iron supply in Antarctic waters and climate change [Kumar et al., 1995]. It has been hypothesized that new production in the Southern Ocean is sensitive to changes in iron inputs to the North Atlantic [Watson and Lefevre, 1999]. That northern hemisphere origin hypothesis appears to be substantiated by our analysis of current lead isotopic data, which suggests that much of the iron in the Southern Ocean could be derived from fluxes of Saharan dust deposited in the central Atlantic Ocean.
 We have plotted the existing isotopic data for all potential sources of lead to the Southern Ocean and compared it to the isotopic compositions reported for Antarctic waters [Flegal et al., 1993] (Figure 1). Those potential sources include atmospheric emissions from Africa, South America, Australia, New Zealand and Tasmania [Bollhöfer and Rosman, 2000] as well as North Atlantic Deep Water (NADW) [Alleman et al., 1999], which upwells in the Southern Ocean [Gordon, 1986]. Detailed discussions of the isotopic characterization of those natural and anthropogenic sources of lead are provided in Alleman et al.  and Bollhöfer and Rosman .
 The plot shows that the isotopic gradient observed in Antarctic surface waters could be attributed to the simple mixing of a less-radiogenic component derived from industrial lead emissions of Southern Chile, Southern Africa, Australia, New Zealand and Tasmania (16.08 < 206Pb/204Pb < 17.19) and the natural weathering of lead from Antarctic rocks (18.64 < 206Pb/204Pb < 18.76). However, the similar isotopic compositions measured in several of the Antarctic water samples with those measured in the NADW suggests that that water mass could transport lead, and associated trace elements, into the Southern Ocean. This proposed NADW origin of much of the lead in Antarctic waters is also consistent with other chemical and physical analyses showing the NADW origins of upwelling waters in the Southern Ocean.
 While the lead isotopic data shows that upwelling of NADW could be a major mechanism influencing metal levels in the Antarctic, we have also used stable lead isotopic data to identify the potential sources of lead (and other metals) to the NADW and hence to the Southern Ocean (Figure 1). The isotopic clustering of the NADW with those of aerosols from Brazil, Argentina, northwest Africa, and Saharan Holocene loess suggests that the arid regions of northern Africa are the source regions for atmospheric dust (and metals such as Fe, which is a major constituent of dust) reaching South America [Swap et al., 1992] or deposited in surface waters of the central Atlantic, where it is transported to the NADW [Alleman et al., 1999; Bory and Newton, 2000]. Therefore these data suggest that northern, rather than southern hemisphere inputs of dust could potentially influence the chemical composition of surface waters of the Southern Ocean (including iron levels), as hypothesized by Watson and Lefevre .
 Emissions of dust and depositional fluxes of mineral aerosols to the Atlantic Ocean have increased considerably in the past decades [Prospero and Ness, 1986], in response to changes in land-use [Tegen et al., 1996] and climate change in northern Africa [Moulin et al., 1997]. Airborne plumes of African dust could potentially mobilize about 4.45 × 107 metric tons of total iron per year (assuming that about one billion metric tons per year of African dust with an average iron concentration of 4.45% is transported to the atmosphere every year) [D'Almeida, 1986; Guieu et al., 2002]. Although the relationship between the atmospheric flux of metals to the ocean and the amount of dust suspended in the atmosphere is not straightforward [Jickells, 1995; Prospero, 1996], there is evidence linking the presence of the Saharan dust plume with elevated levels of iron in surface waters of the Atlantic Ocean [Vink and Measures, 2001].
 In summary, preliminary lead isotopic composition data from the surface waters of the Southern Ocean indicate that much of that lead is derived from atmospheric fluxes of Saharan dust to the NADW. Assuming much of the iron in the NADW is also derived from the same fluxes, the iron in upwelled waters of the Southern Ocean may originate from aeolian deposition in the North Atlantic. If so, concentrations of upwelling iron should increase in response to the increases in atmospheric dust transport in the tropical North Atlantic. The additional iron could potentially increase the efficiency of the biological pump in the Southern Ocean and act as a negative feedback loop for global warming, reducing atmospheric CO2. However, further measurements of chemical tracers and metals in deep waters along the advective flow lines from the North Atlantic to the Southern Ocean are needed to substantiate the inter-hemispheric transport of iron.