An appraisal of stepped heating methods for the extraction and isolation of fluid inclusion CO 2 for carbon stable isotope ratio analysis indicates that the magnitude of the system blank, and thus its influence on the resulting data, has been considerably underestimated in previous studies. The problem derives largely from the unmodified application of stepped combustion techniques originally developed for the removal of terrestrial organic matter from extraterrestrial samples. In particular, the presence of heated Pt in the sample extraction chamber, as reported in the literature, is shown to have a serious deleterious effect on the analysis of small (sub-micromole) samples. If the extracted gases are not exposed to this combustion catalyst, the magnitude of the system blank is substantially reduced. Furthermore, excellent agreement is then generally obtained between fluid inclusion δ 13C values determined after either stepped heating or crushing of the host mineral as the fluid extraction procedure. The release of a low temperature carbon component during stepped heating, and characterised invariably by a δ 13C value of ca. −25%, is in accord with earlier studies and is attributed to surficial contamination by ubiquitous, airborne particulates of biological origin. Effective discrimination between this source of CO 2 and that derived from palaeofluid inclusions is fundamental to accurate isotopic characterisation of the latter. Application of the revised stepped heating procedure to the extraction of hydrothermal fluid inclusion CO 2 from single quartz grains (∼10–20 mg), for carbon stable isotope analysis by static vacuum mass spectrometry, is reported. Under optimum conditions, and where the isotopic composition of the sample gas is enriched in '3C by less than ∼20%0 relative to the associated procedural blank, 2–3 nmol of sample gas may be analysed with an attendant extraneous (blank) contribution to the δ 13C value that is less than the optimal analytical precision of the measurement (∼±1%0, at the 1σ level). Further progress in the application of ultra-high sensitivity mass spectrometric techniques to accurate δ 13C analysis of carbonaceous fluid inclusion components, particularly coexisting CO 2 and methane, is dependent on proper assessment of the associated blanks and the rigorous control thereof.