Annual precipitation across the PNG is 350-400 mm increasing toward the east, and mean annual temperature ranges from 8.4 to 9.7oC (Burke et.al. submitted). The vegetation of PNG is shortgrass steppe, with Bouteloua gracilis (H. B. K.) Lag.ex Griffiths, and Buchloe dactyloides (Nutt.) Engelm. as dominant species; B. gracilis contributes more than 75% of net primary production in native upland communities (Lauenroth & Milchunas 1992, Milchunas et al. 1989). A number of succulent, half shrubs, forbs, and annual grasses are also present.

Sampling design

Three sets of adjacent native, abandoned, and cultivated fields were identified using aerial photographs. Unplowed and recently abandoned fields were located on 1937 photographs. Because old fields remain visible for at least 30 years and cropland was abandoned mainly during the 1930's, it was reasonable to assume that pastures without signs of cultivation in these photographs had never been plowed (Coffin et al. submitted, Burke et al. submitted, Ihori et al. submitted). Native and abandoned fields were immediately adjacent to each other; the cultivated field sampled at each site was within 1 km of the native and abandoned fields on similar soil. We sampled native and abandoned fields on two parallel transects 50 m apart. Transects were 200 m long and perpendicular to the boundary between native and abandoned fields, with 100 m each in the native and abandoned areas. We sampled soil under individual B. gracilis plants and in bare soil between B. gracilis plants every 25 m along transects for a total of eight pairs of samples in each field. Soil in fallow cultivated fields was sampled in the same pattern, but no plants were present and only bare soil was sampled. Data for the eight subsamples in each combination of site, land use, and microsite were averaged for statistical analysis.

In situ incubations.

Net N mineralization was estimated using an open-tube incubation technique similar to that described by Raison et al. (1987). At each point sampled in the native and abandoned fields, we collected one soil core (5.2 cm diameter and 15 cm depth) for initial analysis on June 15 and left one core in place from June 15 to July 14, 1991. Fresh soil samples were extracted using 2-M KCl, and NO3- and NH4+ concentrations in extracts were determined by colorimetric methods (EPA 1979). Net N mineralization was calculated by subtracting initial NO3- plus NH4+ concentrations from final concentrations after correcting for soil water content and bulk density; leaching was assumed to be negligible. Because we could not leave tubes in t he cultivated fields, we collected fresh cores on June 15 and again on July 14 and treated the second core as the incubated soil sample. Tubes were not needed to prevent N uptake because cultivated fields were fallow with no actively growing plants. A previous comparison showed that N mineralization estimates made with and without tubes in fallow fields did not differ (I.C. Burke unpublished data). Ratios of in situ net N mineralization to total soil N and to net N mineralization in 30 d laboratory incubations (Burke et al. submitted) were used to index, respectively, rates of N turnover and environmental constraints on net N mineralization.

Total C and N contents were determined for three subsamples of air-dried and sieved soil from each core using a micro-Kjeldahl procedure (Bremner et al. 1982) followed by colorimetric analysis (EPA 1979) for N and a wet oxidation method for C (Snyder and Trofymow 1984).

Statistical Analysis

We used analysis of variance to test for main effects of site, treatment, and microsite and interactions of treatment with microsite and site. A blocked design with sites as blocks was used; site interactions were included in the ANOVA model because research involving a larger set of sites in the PNG showed that treatment effects on SOM varied among sites within the PNG, possibly due to differences in land management histories (Ihori et al. submitted). A probability level of 0.05 was used to evaluate significance. Because the design was unbalanced, with no samples from under plants in cultivated fields, type four sums of squares were used in analyses (SAS Institute 1988). Five planned, pairwise comparisons addressed effects of microsites within native and cultivated fields and differences among treatments. Differences among treatments were tested for bare soil because bare microsites were present in all land use types and because we expected that incubations under plants would be affected by artifacts involving N immobilization and uptake (see discussion). Mean separations for planned comparisons used a modified Ryan test, which controls the experimentwise type I error rate (REGWF procedure within SAS GLM procedure, SAS Institute 1988).

Burke, I.C., W.K. Lauenroth, and D.P. Coffin. 1995. Soil organic matter recovery in semiarid grasslands: implications for the conservation reserve program. Ecological Applications. 5: 793801

Ihori, T., I.C. Burke, and P.B. Hook. 1995. Nitrogen mineralization in native, cultivated, and abandoned fields in shortgrass steppe. Plant and Soil. 171: 203208