Accurate values of appropriate radiation characteristics are required for accurate predictions of radiation heat transfer. Reasons are given here to show that, visual appearances notwithstanding, many surfaces employed in engineering are, for purposes of calculating infrared power transfer rates, specular. Advances made in vacuum evaporation technology make it possible to create a variety of specular surfaces, and advances in high-speed computing circuitry make it possible to compute characteristics of such surfaces from measured optical constants of the materials and to calculate the resulting power transfer rates, including effects of directional and spectral selectivity and polarization. Theoretical understanding is not yet adequate to enable prediction of the optical constants, but measurement accuracy for specular reflectances has been advanced to a high level. Though theoretical progress has been slow, a two-electron model offers at least a useful correlation formula for representing measurements. Progress toward a practicable method of allowing for nongray, imperfectly specular, and imperfectly diffuse reflection has not been great, but studies have indicated that use of internal-total hemispherical emissivity for internal surfaces and external-total hemispherical emissivity for external surfaces often results in acceptable accuracy. Important improvements have been made in integrating sphere, heated cavity, and two-pi steradian mirror instruments capable of measuring directional reflectance regardless of the reflection distribution function of the specimen. Such measurement capabilities together with access to high-speed computers permit the appropriate total hemispherical characteristic to be obtained by numerical integration of measured values. Understanding of the accuracy and limitations of these instruments has been advanced by a convenient operator and matrix notation along with a consistent nomenclature. In addition to refinements in measurement techniques, there have been significant improvements in procedures and apparatus for simulating environmental factors which cause changes in radiation characteristics. Much progress has been made in allowing for line and band structure in gas radiation calculations. Measurements of intensity-to-line-spacing ratio S/d and line-width-to-spacing ratio γ/d have been made and correlated with narrow band models. Wide band models have been contrived and used to correlate measurements of band absorption A. Some success has also been made in predicting theoretically the three characteristics, S/d, γ/d, and A. On the analytical side, a method of formulation employing derivatives of the band absorption appears quite promising, at least when temperature gradients are not too severe. On the numerical side, development of computer programs capable of accounting for severe temperature gradients and other nonhomogeneities have been notable. Use of the Curtis-Godson approximation has been made in such programs to account for line effects.