- Riparian areas that can be used as reference sites on which to base goals of vegetation restoration have not been documented in the Oregon Coast Range. I examined the composition and distribution of unmanaged riparian overstories in the central Oregon Coast Range along nine streams which have experienced minimal disturbance from Native Americans and no detectable disturbance since Euro- American settlement. I systematically located transects along nine streams, alternating sides of the streams. Each transect ran perpendicular to the stream and was subjectively divided into different vegetative and/or topographic units called landscape units (LU's). Rectangular plots were placed in each LU for characterization. LU l's were units that were closest to the stream, and LU2's were farther from the stream. Red alder was the most frequently found tree species on both terraces and slopes, and on all LU1 's. On LU2 terraces, alder was also the most frequently found species, but on LU2 slopes, Douglas-fir had the highest frequency. Red alder, Sitka spruce, and bigleaf maple were most commonly found occupying terrace sites, although bigleaf maple might be best adapted to conditions on terraces towards the base of slopes. Conversely, western hemlock and Douglas-fir were most commonly found occupying slope sites. Western redcedar was infrequently found, likely due to seed source limitations. Age distributions and tree frequencies indicate that near-stream communities (LUI 's) and terraces experience both intense and minor disturbances, and they experience both types of disturbances more frequently than communities farther from the stream (LU2's) or on slopes. According to fire records and reconnaissance, all streams appear to have been burned about 145 years ago. When equating a shade-intolerant tree age that was younger than this last catastrophic fire date with a disturbance, calculations of disturbance frequency using four different approaches indicate that between 2.6 and 4.5 disturbances per km per century large enough to regenerate trees occurred since the last stand-resetting fire along the nine creeks sampled. Fifty-two percent of near-stream communities (LU1's) and 23% of communities farther from the stream (LU2's) contained no trees. This could be due to small plot size and/or high shrub competition. The No Tree overstory type was most similar in topographic conditions to the Pure Hardwood overstory type, suggesting that red alder and/or bigleaf maple might have previously occupied the No Tree sites and have since died leaving no or little evidence. It appears that a large-scale, intense disturbance such as fire is needed to allow the recruitment of trees into the shrub-dominated, No Tree areas, especially shade-intolerant trees such as Douglas-fir. Any single definition of natural riparian vegetation is nearly impossible to construct, mainly because most ecosystems are composed of vegetation mosaics that are always changing in time and space. This change is associated with environmental variability, disturbance, and inter- and intra-specific competition. Also, differences in exogenous environmental conditions between pre-settlement times (circa 1850) and today, suggest that historic vegetation, ecological conditions, and resulting successional pathways might not mirror the vegetation, ecological conditions, and successional pathways of currently unmanaged riparian areas. Instead, results from this study, revealing the existence of mixtures of hardwoods, conifers, and no-tree areas over lengths of a stream, should be perceived as just one of many possibilities for a riparian overstory reference model.