Serum is a complex heterogeneous matrix contianing endogenous proteins with potential diagnostic value. No previously reported methods have allowed the detection of all its constituent proteins from a single sample. In the work described in this thesis, normal human serum was separated by reversed-phase or dye affinity chromatography, ultrafiltration and by extraction under various pH and salt concentrations in aqueous and organic solutions. The serum extracts were analyzed by SDS-PAGE, MALDI and LC-ESI-Qq-TOF MS/MS without prior enzymatic digestion. Peptides observed as multiple isotopic peaks in MS mode were correlated with known serum proteins by ProteinPilot and X!Tandem analysis of their fragmentation spectra using the no enzyme [XX] condition. Fragment ion masses within 0.04 Da of those predicted yielded high confidence identifications from both algorithms. Angiotensin 1 and [G1u1] fibrinopeptide B were clearly detected in serum spiked with as little as 250 and 100 femtomoles respectively. Post-translational modifications of serum peptides were identified using ProteinPilot.