The purity, viability, and germinability of purchased seed 

 were determined by the Utah State Seed Lab using the 

 Association of Official Seed Analysts rules for testing 

 seeds (1981). Seeding rates were calculated using purity 

 and germinability for purchased seeds and purity and 

 viability for collected seeds. An additional 20 percent was 

 factored into seeding rates for collected seeds to account 

 for variable germinability. 



Field Application 



Study plots were installed the last week of September 

 1984. Site preparation prior to planting included ripping 

 to a depth of 75 cm, removing large rocks, and harrowing. 

 Before planting, seeds were mixed with rice hulls to a 

 total volume of 10 liters. Mixtures were then sown by 

 hand on an individual replication basis. Fertilizer was 

 also applied by hand. A Brillion seeder-packer was used 

 to pack the seed into the seedbed. The study area was 

 fenced to prevent cattle grazing. 



Data Collection and Analysis 



In the summer of 1985, 10 rectangular quadrats 

 0.25 m 2 were permanently located within each replication 

 of each treatment using a stratified random method 

 (Chambers and Brown 1983; Knight 1978). Vegetation at- 

 tributes measured included density, peak standing crop 

 biomass, and cover. Density, defined as the number of 

 plants of each species rooted in a quadrat, was assessed to 

 evaluate the relationship between seeding rate and initial 

 establishment. Because of the problems associated with 

 identifying individual plants, density data were collected 

 only during the first growing season (1985) when it was 

 relatively easy to identify individuals. 



I assessed standing crop biomass for each species after 

 peak production in 1985, 1986, and 1987 using a weight- 

 unit estimation method (Carpenter and others 1984). 

 Weight-unit estimation was performed inside each of the 

 10 permanent quadrats within a replication; in addition, 

 a verification quadrat was clipped within each replication 

 but outside of the permanent quadrats. Samples were 

 dried at 60 °C for 24 h. Standing crop data were collected 

 for seeded species and for colonizer species that were not 

 seeded. 



Prior to sampling density and standing crop biomass, 

 I determined cover from 35-mm slides taken of each per- 

 manent quadrat. These slides were projected onto a 100- 

 square grid and estimates were made of the areas occu- 

 pied by five cover classes: aerial vegetation cover, litter, 

 bare-ground, gravel (2 mm to 7.6 cm), and rock (>7.6 cm) 

 (Chambers and Brown 1983). 



In 1986 soils were assessed for the edaphic effects of the 

 fertilizer treatments. Two soil cores, 7.5 cm in diameter 

 by 10 cm deep, were collected for each replication of each 

 treatment (n = 6). The soils were analyzed by A and L 

 Labs of Omaha, NE, for concentrations of N, P, and K. 

 A specific ion electrode in a saturated calcium sulfate ex- 

 traction was used to determine N0 3 -N (Carson 1980a). 

 Available-P was measured with the Bray-1 procedure 

 using the Fiske-subbarrow reducing agent (Knudson 



1980). Exchangeable-K was analyzed in a one normal 

 solution of ammonium acetate adjusted to pH 7.0 (Carson 

 1980b). 



I determined differences in density, standing crop, 

 cover, and soil properties among treatments from stan- 

 dard analysis of variance (ANOVA) techniques. The den- 

 sity and standing crop analyses were based on a compari- 

 son of mean values for three species categories: intro- 

 duced grasses, natives, and colonizers. Because time was 

 a factor in the analysis of standing crop and cover, I used 

 a repeated measures design and conservative degrees of 

 freedom. Mean comparisons were performed using 

 Fisher's Protected LSDs (p < 0.05) under the guidelines of 

 Petersen (1977). 



RESULTS 



Preexisting site conditions had a significant impact on 

 the results of this study. Because the study plots were in 

 a basin, they experienced a perched water table early in 

 the growing season as evidenced by ponded water on the 

 surface and soil mottles. This may have had a negative 

 effect on the seedling establishment and standing crop of 

 all seeded species. The same introduced grasses as were 

 seeded in the current study existed on the research plots 

 prior to study installation from a previous revegetation 

 effort. Although they had been sprayed with 

 glyphosphate the previous summer, they were not com- 

 pletely eliminated. The existence of these grass species 

 may have confounded the results of the seeding density 

 treatments. 



To determine the relative importance of the residual 

 introduced grasses, I compared native species and intro- 

 duced grass densities in the three seeding density treat- 

 ments in year 1. Native species density varied little 

 among treatments and averaged 8.54 plants/m 2 for the 

 three seeding densities. Introduced grass density was 

 8.04, 19.40, and 21.12 plants/m 2 in the natives only, low 

 grass density, and high grass density treatments, respec- 

 tively. Thus, total density of introduced grasses relative 

 to total density of native species in the native's only treat- 

 ment was similar to that intended for the low grass den- 

 sity treatment. In the other two seeding treatments intro- 

 duced grass density relative to native species density was 

 intermediate between the intended low and high seeding 

 treatments. 



Density 



Neither seeding density of introduced grasses nor fer- 

 tilization had any significant effects on the combined 

 seedling density of introduced grasses, native species, and 

 colonizers during the first growing season (fig. 1). High 

 densities of colonizers relative to those of native species 

 and introduced grasses in the majority of treatment com- 

 binations in year 1 appeared to have an overwhelming 

 effect on the analysis. The colonizers had significantly 

 higher densities than either the introduced grasses or 

 native species (p < 0.05). Deleting the colonizer species 

 from the overall analysis resulted in significant differ- 

 ences among the three seeding densities and between the 



3 



