This work demonstrates, for the first time, the applicability of multiple spiking isotope dilution analysis to molecular mass spectrometry exemplified by the speciation analysis of mercury using GC(EI)MS instrumentation. A double spike isotope dilution approach using isotopically enriched mercury isotopes has been applied for the determination of inorganic mercury Hg(II) and methylmercury (MeHg) in fish reference materials. The method is based on the application of isotope pattern deconvolution for the simultaneous determination of degradation-corrected concentrations of methylmercury and inorganic mercury. Mass isotopomer distributions are employed instead of isotope ratios to calculate the corrected concentrations of the Hg species as well as the extent of species degradation reactions. The isotope pattern deconvolution equations developed here allow the calculation of the different molar fractions directly from the GC(EI)MS mass isotopomer distribution pattern and take into account possible impurities present in the spike solutions employed. The procedure has been successfully validated with the analysis of two different certified reference materials (BCR-464 and DOLT-4) and with the comparison of the results obtained by GC(ICP)MS. For the tuna fish matrix (BCR-464), no interconversion reactions were observed at the optimized conditions of open focused microwave extraction at 70 degrees C during 8 min. However, significant demethylation was found under the same conditions in the case of the certified dogfish liver DOLT-4. Methylation and demethylation factors were confirmed by GC(ICP)MS. Transformation reactions have been found to depend on the sample matrix and on the derivatization reagent employed. Thus, it is not possible to recommend optimum extraction conditions suitable for all types of matrices demonstrating the need to apply multiple spiking methodologies for the determination of MeHg and Hg(II) in biological samples. Double spike isotope dilution analysis methodologies using widespread GC(EI)MS instrumentation are proposed here for the routine analysis of inorganic mercury and methylmercury in fish samples. The estimated method detection limits were below 10 ng g(-1) for both mercury species. Precision was evaluated for the concentrations present in the certified reference materials (CRMs) which vary from 0.1 to 5 microg g(-1), achieving values of coefficients of variation ranging from 7% to 2%. The concentrations obtained in both CRMs analyzed were in excellent agreement with the certified values, demonstrating the accuracy of the method at these concentration levels.