The Nakhla meteorite, one of the SNC group of putative Martian origin, is an igneous rock, a cumulate of augite and olivine, that does not represent a magma composition. Samples of its parent magma were trapped as magmatic inclusions in its cumulus olivine crystals. The inclusions range from 10-350 μm in diameter and consist of augite rich in Ti and Al, silica-rich glass, iron oxides, ilmenite, apatite, and rare amphibole. Each inclusion has a rind of augite and chromite against its host olivine, and a core of radiating sprays of augite set in glass and feldspar. These textures are consistent with an origin as trapped silicate liquid. Twenty four inclusions were analyzed by electron microprobe, by averaging series of rastered beam analyses. On average, the analyses are independent of the size of the inclusion, most clearly for inclusions larger than 70 μm diameter, and variations among analyses are consistent with small differences in the proportions of augite, glass, and iron oxides exposed on thin section surfaces. The average inclusion analysis is calculated to represent 28% host olivine and 72% inclusion composition. The inclusion rinds are calculated to be 13% of the inclusion composition, for which a small correction is made. The inclusion composition is then corrected for Mg/Fe exchange with the host olivine (to chemical equilibrium with the cumulus augite cores) and crystallization of 13% olivine onto the host crystal to yield a magma composition with the proper MgO/FeO ratio and with olivine and augite on its liquidus. This composition, NK93, is an estimate of Nakhla's parent magma originally trapped in the inclusions. NK93 is basaltic, rich in iron, and poor in aluminum compared to terrestrial basalts. Its abundances of Al, Ti, and Na are consistent with the geochemistry of the SNC meteorites as a whole. NK93 is similar to the parent magmas proposed in an independent study of magmatic inclusions in Nakhla, but unlike other proposed compositions. One proposed parent magma, based on mass balance and the assumption that Nakhla was a closed-system after crystal accumulation, is inconsistent with the magmatic inclusion studies and Nakhla's petrography. This suggests that Nakhla was not a closed-system, but was infiltrated by post-cumulus basaltic magmas. Another suggested parent magma was derived from element distribution between augite and basaltic magma, but that calculation relied on an incorrect augite composition. Using a correct augite composition in that calculation yields a parent magma much closer in composition to NK93. Thus, three independent approaches converge in requiring that Nakhla originally formed from a low-Al, low MgO/FeO basalt like NK93.