Among the well-imaged small satellites and asteroids, Deimos displays a unique surface: very smooth with global-scale albedo features. We have examined the disk-resolved photometry of Deimos using Viking Orbiter images for clues to its distinctive appearance. Hapke parameters were fit to characterize the phase behavior and to compute normal reflectance. The opposition surge amplitude (B0) is smaller for Deimos than for Phobos. Outside the range of the opposition effect the two martian satellites have similarly shaped phase curves, but Deimos is about 20–30% brighter than Phobos from 10°–80° phase. The calculated mean normal reflectance of Deimos (λeff= 0.54 μm) is 0.068 ± 0.007. The brighter and darker areas on Deimos exhibit constant contrast between 0.6° and 81° phase; this characteristic allows a calculation of the range of normal reflectances over most of its surface, nearly all of which values are between 0.06 and 0.09. The trailing side of Deimos has a larger relative distribution of the brighter material, and is on average about 10% brighter than the leading side. The mean normal reflectance cannot be formally distinguished from that of Phobos (0.071 ± 0.012; Simonelli, D. P., M. Wisz, A. Switala, D. Adinolfi, J. Veverka, P. C. Thomas, and P. Helfenstein 1996. Submitted toIcarus). Although the statistical distribution of normal reflectances on the two satellites is similar, the geography of the albedo variations is very different. Deimos has gradational changes in albedo downslope from ridge crests, primarily manifested in long albedo “streamers.” On Phobos there is a more patchy global distribution of albedos, apparently related to ejecta from the large crater Stickney. Because of the similarity of mean density, spectral properties, mean normal reflectance, the range of normal reflectance, and phase function outside the opposition effect, Deimos appears to be made of materials with compositions very similar to those on Phobos, although the apparent wider distribution of ejecta on Deimos has been cited as indicating a greater role for strength scaling in cratering on Deimos (Lee, S. W., P. Thomas, and J. Veverka 1986.Icarus68, 77–86). Simple modeling of the formation of the albedo patterns by gardening, creep, and “weathering” of bright material from crater rims suggests that impact gardening contributes very little to the motion of the material downslope, and that vertical mixing and/or “weathering” must be important in addition to an unspecified creep process. The distinction of Deimos is primarily in the smooth surface that allows a particularly large scale of downslope movement of regolith on very gentle slopes. This smoothness is most easily explained by the effects from impact formation of a 10-km concavity at high southern latitudes in the latter half of Deimos' surface history. This impact scar is relatively much larger than is the crater Stickney on Phobos. The effects of this large impact probably include blanketing by an average of nearly 200 m of ejecta, but also may include seismic erasing of craters similar to that proposed for Ida by Asphauget al. (Asphaug, E., J. M. Moore, D. Morrison, W. Benz, and R. A. Sullivan 1996.Icarus120, 158–184).