Carbon monoxide is a highly poisonous gas and an important measurement in the forensic investigation of a cause of death. A straightforward gas chromatography method using headspace sampling and thermal conductivity detection has been developed to routinely measure carbon monoxide in post-mortem blood samples. Carbon monoxide was released by injecting 200µl of a liberating reagent consisting of potassium ferricyanide into a sealed headspace vial containing 200µl of the blood sample. The vial was equilibrated, heated and vibrated in the headspace system before the released gas was injected onto a molecular sieve 5Ǻ PLOT capillary column, isothermically held at 60oC, in a 2:1 split ratio. The detector was held at a temperature of 200oC with a constant flow of helium as the reference gas. The total run time per vial was less than 15 minutes. This methodology is prone to oxygen build up on the column. Raising the column temperature to 200oC at the end of each batch of samples mitigates this. The method was validated for routine use. A calibration model, bias, precision, and limits of detection and quantitation were all defined. Potential carryover, dilution factors and sample and reagent stability were all identified. Patient and external quality assurance scheme samples were analysed. The method had a linear calibration model and showed good bias (-10.0-5.7%) and precision (2.4-10.3%), no carryover between samples, good freeze-thaw stability of the carboxyhaemoglobin, excellent stability of the liberating reagent (nine months), and good bias (5.9-16.2%) and precision (2.6-8.1%) with sample dilution. Comparison of patient and EQA samples with the current referral method showed acceptable correlation, with 90% of samples having the same clinical interpretation.