Ignoring the origin of atmospheric chemicals is often a strong limitation to the full interpretation of their measurement. In this article, this question is addressed in the case of the sulfur species in Antarctica, with an original method of retrotransport of tracers. The retrotransport model is derived from the Laboratoire de M¿t¿orologie Dynamique Zoom-Tracers (LMD-ZT) atmospheric general circulation model, optimized for polar climate and expanded to simulate atmospheric sulfur chemistry. For two East Antarctic scientific stations (Dumont d'Urville and Vostok) the effects of transport and chemistry and the influence of oceanic, volcanic, and anthropogenic sources on dimethylsulfide (DMS), non-sea-salt (nss) sulfate, and methanesulfonic acid (MSA) concentrations are evaluated in summer and winter. The oceanic source largely dominates, but other sources can episodically be significant. The meridional origin and the age of DMS, MSA, and biogenic nss sulfate are also estimated. The latitudes of origin of MSA and nss sulfate are similar in summer, but they differ markedly in winter. This is a signature of their different chemical production scheme. Also, the interannual variability of the origin of the sulfur species at Vostok is weak compared to that at Dumont d'Urville. Acknowledging that the DMS concentrations in the ocean have no interannual variability in the model, this result suggests unsurprisingly that inland Antarctic stations may be better observation sites to monitor large-scale DMS bioproductivity variability than coastal sites are. The combination of slower chemistry and more intense atmospheric circulation in winter leads to unexpected results, such as a younger DMS in winter than in summer at Vostok.