The largest known terrestrial impact melt sheet occurs within the 1850 Ma Sudbury Structure, Ontario. In order to evaluate the relative contributions of different target lithologies to the melt sheet, we have investigated the Pb isotope compositions of feldspar separates from early-formed quartz diorite magmas within Offset Dykes from around the impact structure. The samples define a linear array on plots of age-corrected 206Pb/204Pb versus 207Pb/204Pb. Samples from Offset Dykes hosted by the Huronian Supergroup (South Range) have a range of 206Pb/204Pb1850 from 15.424 to 17.255 and 207Pb/204Pb1850 from 15.390 to 15.801, whilst those hosted by Archean gneisses of the Superior Province (North Range) cluster around 206Pb/204Pb1850 ≈ 14.8 and 206Pb/204Pb1850 ≈ 15.1. These values can be approximated by binary mixing between the two major groups of target lithologies. A mix of 60–70% of Superior Province gneisses with 30–40% of Huronian metasedimentary material closely matches the Pb isotope compositions of North Range Offset Dyke samples, whereas in the South Range the required Huronian component is up to ca. 80%. These mixing proportions are consistent with Sr, Nd and Os isotope and trace element constraints. A third minor component, either locally-exposed Paleoproterozoic mafic rocks or the lower crust is also required. However, the isotopic, trace element and Ni–Cu–platinum group element characteristics of the melt sheet can be accommodated without the involvement of an average lower crustal or meteoritic component. A major contribution of Huronian supracrustal material, which had a pre-impact thickness of up to 12 km, is required to explain the chemical characteristics of the impact melts, which also have a strong upper crustal affinity (e.g. Eu/Sm = 0.22, Rb/Sr = 0.2–0.35). As such, a shallower level of melting is apparent than that predicted by many previous impact models for the Sudbury event. This can be accommodated by considering approach trajectories for the impactor oblique to the Earth’s surface. In addition, the isotopic and trace element variability identified indicates that the melt sheet was heterogeneous at an early stage, and may not have been completely homogenised during crater formation. Our findings have significant implications for the nature of the Sudbury impact event, the evolution of the melt sheet and the crustal sources of metals contained in Sudbury’s world class Ni–Cu–PGE sulphide ores.