Chlorine atom recombination coefficient (gamma(Cl)) measurements are reported for a variety of surfaces and at a range of surface temperatures. The surfaces include crystalline silicon, quartz, anodized aluminum, tungsten, stainless steel, polycrystalline silicon, and photoresist. Surface temperatures ranged from about -90 degrees C up to 85 degrees C. Measurements were made in a vacuum chamber with chlorine atoms and molecules effusing from an external discharge source as a molecular beam and impacting a selected surface. The incident and reflected beam compositions calculated using a modulated beam mass spectrometer were used to infer the recombination coefficient. At room temperature, the values of gamma(Cl) ranged from below the detection sensitivity (about 0.01) for crystalline silicon to similar to 0.85 for stainless steel. Other surfaces displayed intermediate values between these extremes. For example, gamma(Cl) for polycrystalline silicon is about 0.2-0.3 at room temperature. All surfaces, except stainless steel, displayed increasing values of gamma(Cl) as surface temperature was lowered below room temperature, down to the freezing temperature of chlorine (-101 degrees C). The gamma(Cl) for stainless steel appeared to saturate at 0.85 as temperature was lowered. All surfaces displayed decreasing values for the recombination coefficient as surface temperature was raised above room temperature. The gamma(Cl) data as a function of temperature were fit to a phenomenological model. The phenomenological model assumes Cl atoms adsorb into a weakly bound physisorbed, state on at least 1 monolayer of strongly bound, chemisorbed chlorine. After adsorption, the model assumes that thermally activated diffusion and atomic recombination occur with a rate that is first order in physisorbed chlorine. Thermal desorption competes with diffusion and reaction, and is also thermally activated. Fits to the data were made, and the physical interpretation of the model parameters is discussed.