Previous descriptions of thermal emission spectroscopy have presented techniques that vary in accuracy and reproducibility. Contributions of thermal energy from the instrument and environment are major calibration factors that limit accuracy in emissivity determination. Reproducibility is related to the stability of these quantities. Sample temperature determination is also a significant factor in arriving at accurate emissivity. All of the factors which impact the measurement of emissivity using an interferometric spectrometer with an uncooled detector are isolated and examined here. An experimental apparatus is presented along with a description of a simplified measurement and calibration scheme used to arrive at quantitative emissivity of minerals. A detailed error analysis examines the effect of errors in each of the calibration parameters in isolation and as part of multiple error propagation. Sample temperature determination from radiance can create emissivity error, but 95% of published minerals have an emissivity maximum of 0.98 or higher, resulting in emissivity error of 2% or less. With worst-case systematic and random errors included, emissivity can be determined with an uncertainty of ~4%. In most cases it is less than 2%. Reproducibility with this technique is better than 1%.¿ 1997 American Geophysical Union