Formaldehyde (HCHO) is a key intermediate in the photooxidation of methane by hydroxyl radicals. Through its photolysis, it is also a source for free radicals in the troposphere. Owing to these reactions, HCHO influences the oxidation capacity of the atmosphere and is a suitable species to test our current understanding of atmospheric oxidation pathways. Especially in polar regions, discrepancies between measurements and model calculations exist. Though recent investigations in the Arctic suggest that HCHO emissions from the snow surface might act as the missing source, the question remains unresolved for the Antarctic. We compare year-round HCHO measurements in Antarctica with model results from a simple photochemical box model. The observed ambient HCHO mixing ratios cannot be explained by methane photooxidation alone. Inclusion of HCHO emissions from the snow surface makes the model results and measurements consistent, but significantly higher emissions than those derived in the Arctic are needed to explain the observed HCHO mixing ratios. We discuss and model other possible sources such as oxidation of dimethylsulfide (DMS), isoprene, ethene, propene, and the effect of halogens, that may be responsible for the enhanced HCHO mixing ratios in the marine Antarctic troposphere. We find that, for the largest HCHO mixing ratio measured, all potential gas-phase HCHO precursors (including methane) are likely to generate only 20--40% of the required HCHO. If the remaining HCHO is produced by a flux from the snow, the flux required is 1.9--2.5 ¿ 1013 molecules m-2 s-1 if the boundary layer height is 40--50 m.