Massive volumes of mafic magmatism forming the Karoo-Ferrar Large Igneous Province (LIP) in Southern Africa and Antarctica preceded Jurassic breakup of the Gondwana Supercontinent. This widespread LIP magmatism is attributed to a major mantle plume, or plumes, impacting an area thousands of kilometres across. Magmas in lava flows and shallow sills, which flowed laterally hundreds to thousands of kilometres, form most of the exposed LIP. Hence, the distribution of shallow level mafic rocks may not reflect the location of mantle melting. In contrast, large deep-seated mafic intrusions such as gabbros likely more directly overlie areas of mantle melting. Antarctic exposures of such intrusions are limited to the Dufek Intrusion and outcrops >1000 km to the north, hence the true pattern of mantle melting is poorly constrained. Regional aeromagnetic and aerogravity data suggest other Jurassic mafic intrusions are present, but detailed analysis of these bodies is lacking. To define more precisely the distribution of mafic intrusions we use data from the first stand-alone strapdown gravity survey in Antarctica. This innovative technique allows collection of aerogravity data during draped flight, resolving anomalies with a wavelength of ∼6 km and a root-mean-square error of 1.8 mGal. Combining this new gravity data with coincident aeromagnetic data we investigate a ∼50 mGal gravity and associated >1000 nT magnetic high in Coats Land, East Antarctica. Our interpretation is that the so called ‘Halley High’ reflects a large gabbroic body ∼80 km long, 30 km wide and ∼6 km thick, equivalent to the inferred total size to the better known Stillwater layered mafic intrusion in the US. Our interpretation of a large mafic intrusion supports the suggestion from reconnaissance aeromagnetic data that this and other similar anomalies are Jurassic mafic intrusions. These large mafic intrusions, and hence underlying mantle melting, appear restricted to a linear band parallel to the continental margin. This structured pattern of mantle melting is consistent with shallow mantle convection and lithospheric extension playing a significant role in the later stages of the Karoo-Ferrar magmatism. An apparent 650 km gap in mafic intrusions adjacent to the continental Weddell Sea Rift System suggests different mantle processes were occurring here. Future joint magnetic and strapdown gravity surveying presents a key opportunity to fully constrain the location and extent of mantle melting during Gondwana breakup.