tions; others are not. For example, sown with 22.4 kg of seed per with seedlings during plantation establishment. Sowing crownvetch or climbing legumes like flatpea or hairy vetch under pole-size trees, however, may be desirable to insure competitiveness with natural vegetation. The widespread success of N fertilizers for increasing conifer growth in western Washington and Oregon justifies additional research and development of N-fixing plants as alternative means for supplying N. Some of the legumes tested in this trial on harsh sites may fit this purpose. STUDY AREAS Location.--Our species trial was located within 0.4-hectare, fenced deer exclosures in three recently clearcut areas in the western Cascade Range east of Roseburg, Oregon (fig. 1). These areas were considered representative of the 600- to 900-m (2,000 to 3,000-ft) elevational zone of the Umpqua River drainage; north, south, and southeast aspects were presented (table 1). INTRODUCTION Legumes can improve the status of nitrogen and organic matter in soils and can thereby increase growth of associated vegetation. Using N-fixing plants in agriculture and forestry becomes financially more attractive as the costs of nitrogen fertilizers and their application continue to increase. We conducted a field trial to test the suitability of some agricultural legumes for use in forests of the western Cascade Range of Oregon. We report the 9-year survival and growth of eight legume cultivars sown in three logged and burned clearcuts. We contrast their performance on fertilized and unfertilized plots. We also intended to measure the effects of these legumes on an associated conifer, but seedlings from Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco var. menziesii) seeds sown concurrently with the legume seeds failed to survive. Our results show that several of these legumes established vigorous stands in these severe growing conditions. Figure 2.--Initial site conditions before sowing at the Steelhead Creek location. Soils.--Soils of the study areas were deep or moderately deep, welldrained loams or clay loams developed on tuffs and breccias of the extensive Little Butte Volcanic Series (Peck et al. 1964). Clay content and bulk density in these soils gradually increased with soil depth (table 1). The most obvious differences among the soils were in color and structure of the surface soils. At Steelhead Creek, the surface soil consisted largely of hard soil aggregates or "shot; " these were less prevalent at the other locations. Moreover, the dark redbrown soil at Steelhead Creek contrasted sharply with the yellowbrown soils at Calf Creek and Acker Rock. These soil colors were consistent with those of the original pyroclastic parent material. Climate.--The Mediterraneantype climate of the study areas is characterized by extended periods of drought during the summer. Past annual precipitation averaged 1 524 mm (60 in) at Calf Creek and Steelhead Creek and 1 016 mm at Acker Rock, the most southerly location (Pacific Northwest River Basins Commission Meteorological Committee 1969). Average evaporation stress (Class A Pan) was 1 068 mm and similar for all locations. METHODS General Plant Materials Seeds of annual or perennial legumes were sown on 2- by 2-m (1 milacre) plots at amounts equal to 22.4 kg/ha for each species. Based on a generally recommended seed density for broadcast sowing of 1,550 live, germinable seeds per square meter (Berglund 1976), our sowing provided excessive seed density for the small-seeded deervetches and sparse density for the remaining, heavier-seeded species (table 2). Each species was randomly assigned to a block of three plots at a given location; the plots were separated by 1.2-m-wide buffer strips. Sowing occurred between March 24 and April 6, 1965; however, the Washington lupine was sown about 1 month later. Although seeds of most species were broadcast without covering to simulate helicopter application, those |