Intracellular signaling pathways are to a large extent regulated by reversible protein phosphorylation of pathway components. To fully investigate the regulation of these pathways, it is often necessary to identify the sites of protein phosphorylation induced on individual components. The low abundance of many of these molecules and the potentially low stoichiometry of phosphorylation means that conventional analytical techniques are incapable of identifying specific sites of modification inducible in vivo. The most common technique used is two-dimensional (2D) phosphopeptide mapping (electrophoresis, thin-layer chromatography) of peptides derived by proteolysis of a phosphoprotein. The number of spots detected is commonly interpreted as the number of sites of phosphorylation. Here we have achieved positive identification of phosphorylation sites by capillary high-performance liquid chromatography, with on-line mass spectrometric detection, of phosphopeptides recovered from 2D phosphopeptide maps. We demonstrate that the chemical composition of phosphopeptides is not altered during the 2D mapping procedure. By detailed analysis of the sites of phosphorylation induced in vitro on CD3-zeta by p56lck we demonstrate that interpretation of the sites of phosphorylation based on 2D phosphopeptide mapping alone is difficult. To minimize over- or misinterpretation of 2D phosphopeptide maps we therefore postulate rules that should be applied generally in cases in which protein phosphorylation sites are being evaluated by 2D phosphopeptide patterns alone.