Methods

Eighteen experimental plots were located in Boulder Open Space land, south of the city of Boulder, Colorado. These plots measure 3 m by 1.5 m and were randomly divided into a control half and a treatment half. Twelve of these plots were on the site of a 1991 disturbance, and six were on the site of a 1993 disturbance. Burial of utility cables was responsible for both disturbances. The sites were fenced to exclude cattle. However, cattle entered the plots for limited periods of time each year. The sites were seeded in 1991 and 1993 with a mixture of grasses consisting of Western wheatgrass, blue grama (Bouteloua gracilis), big bluestem (Andropogon gerardii), little bluestem (Schizachyrium scoparium), and side oats grama (Bouteloua curtipendula), but only the Western wheatgrass was still present at the time of the study. All plots had large populations of invasive species and potentially elevated nitrogen levels due to disturbance, atmospheric enrichment, and feedback effects from decomposition of weedy species. These plots were also used in carbon amendment studies conducted in 1994 and 1995 reported by Seastedt et al. (1996). I applied carbon amendments by hand to the soil surface in the center square meter of the treatment plots. A light dusting of 200g sucrose (table sugar) was sprinkled on the soil surface plots for five months from February to June. A heavier dusting of 325g of air-dried sawdust was sprinkled on the soil surface in March and May only. This resulted in a total carbon addition of 1000g sucrose and 650g sawdust.

I analyzed the soil for extractable ammonium and nitrate at three times to evaluate the duration of the immobilization potential of the amendments. The first evaluation period occurred in May one week after the second to last carbon application. This soil analysis sampled the soil of all of the paired plots. The second evaluation occurred in July after all carbon additions had ceased for one month, and it analyzed the soil from six paired plots from the 1991 disturbance site and six paired plots from the 1993 disturbance site. The final evaluation occurred in August, or two months after cessation of carbon additions, and sampled the soil of the remaining six paired plots. A paired plot consisted of an experimental plot and its adjacent control plot. Each sample consisted of two 10 cm soil cores from each half of the paired plots. Soils were sifted through 2 mm mesh sieves, and a subsample was weighed, dried at 60°C, and re-weighed to measure moisture content. I prepared each soil sample by adding 10g of soil to 50 ml of 2N KCl, shaking the sample for 1 hour, and filtering it the following day. The solution was frozen until it was analyzed on a Lachet analyzer for nitrate (NO₃^-) and ammonium (NH₄⁺).

I identified all the native and invasive plants in the plots in midsummer. I counted the total number of alyssum and Western wheat plants within a 0.1 m² quadrat and the total number of diffuse knapweed plants within the full 1 m² plot. The first biomass harvest occurred in mid-July at a time of peak foliage biomass. All biomass except flowering knapweed was clipped at the soil surface in an 0.1 m² section of each plot. Flowering knapweed biomass was collected at a later date because the knapweed seed counts required us to allow the knapweed to mature in the field and produce seed. The previous season's standing dead material and litter were not collected. This harvest underestimates the annual productivity because many of the early season plants had already senesced, but I assumed that this bias would not affect treatment comparisons. All biomass was dried at 60°C to constant weight. The second harvest occurred in late August and consisted of all the mature knapweed found in 1 m² of all the 36 plots. All knapweed was air-dried to constant weight. I estimated knapweed seed production by harvesting and counting the number of seeds per ten random seed heads from the knapweed plants in each plot.

All data were analyzed with a standard paired plot t-test except for the mature knapweed density (SAS 1988). Knapweed was not found in all plots and both densities and biomass exhibited no correlations within the paired plots. The knapweed data were analyzed using a one-way ANOVA on log-transformed data because the variances were larger than the means.

Introduction	Home	Results