The high resolution X-ray structures of 38 proteins that bind phosphate containing groups and 36 proteins binding sulphate ions were analysed to characterise the structural features of anion binding sites in proteins. 34 of the 66 phosphates found were in close proximity to the amino terminus of an alpha-helix. 27% of phosphate groups bind to only one amino acid, but there is a wide distribution, with 3% of phosphates binding to seven residues. Similarly, there is a large variability in the number of contacts each phosphate group makes to the protein. This ranges from none (3% of phosphates) to nine (3% of phosphates). The most common number of contacts is two (23% of phosphates). The most commonly found residue at helix-type binding sites is glycine, followed by Arg, Thr, Ser and Lys. At non-helix binding sites, the most commonly found residue is Arg followed by Tyr, His, Lys and Ser. There is no typical phosphate binding site. There are marked differences between propensities for phosphate binding at helix and non-helix type binding sites. Non-helix binding sites show more discrimination between the types of residues involved in binding when compared to the helix set. The propensities for binding of the amino acids reveal the expected trend of positively charged and polar residues being good at binding (although that for lysine is unexpectedly low) with the bulky non-polar residues being poor at binding. Bulky residues are less likely to bind with the amide nitrogen. Sulphate binding sites show similar trends. Analysis of multiple sequence alignments that include phosphate and sulphate binding proteins reveals the degree of conservation at the binding site residues compared to the average conservation of residues in the protein. Phosphate binding site residues are more conserved than sulphate binding sites.