Methods
Experimental Gardens
This research was performed at the Cedar Creek
Natural History Area (CCNHA), located on a glacial outwash plain in Isanti and
Anoka counties in east-central Minnesota, USA. Our experiment was conducted
within eleven garden plots (3m X 12m), designed to be an experimental N
gradient. These gardens were established in 1985 and have been the site of
numerous other experiments (Wedin, 1990, Wedin and Tilman, 1990, Tilman and
Wedin, 1991a, Tilman and Wedin, 1991b). Here we will briefly describe the
preparation of the subplots we sampled within the larger garden plots, refer
to Tilman and Wedin (1991a) for a more complete description. Each plot was
prepared by rototilling a different (randomly assigned) proportion of topsoil
into subsurface sand, ranging from 0 to 100% top soil. One plot, which
contained 100% topsoil, was further enriched with an additional 6.55g m-2 yr
-1 of N as NH4NO3, added monthly in proportion to normal N mineralization
rates at CCNHA. The topsoil, which was obtained nearby but not within CCNHA,
was a sandy loam of the Hubbard-Isanti-Duelm Association containing 72% sand,
4% clay, 24% silt, 2.9% organic matter, 1200mg/kg total soil N, 58mg/kg
Bray-1 P, 52mg/kg K, a pH of 7.2, and a bulk density of 1.40g/cm3. The
subsurface sand contained 93% sand, 3% clay, 4% silt, 0.3% organic matter,
90mg/kg total soil N, 31mg/kg Bray-1 P, 29mg/kg K, a pH of 6.6, and a bulk
density of 1.51g/cm3. To ensure that N was the only nutrient limiting plant
growth, each plot was fertilized annually with 12.2g/m2 of P2O5 (commercial
0-46-0 fertilizer), 14.6g/m2 of K2O (commercial 0-0-61 fertilizer), 42g/m2 of
MgSO4 (commercial epsom salts), 42g/m2 CaSO4 , 1.9g/m2 of ZnSO4, 1.12g/m2 of
CuSO4, 0.28g/m2 of boric acid, and 0.56g/m2 of MnSO4. The soil pH was stable
at 7.2 in all plots except the two sandiest plots, which received fine-ground
lime in 1987 and 1988 to adjust their pH to 7.2.
Each 3m X 12m garden plot was divided into 0.75m X 0.75m subplots using
galvanized sheet metal driven into the soil to a depth of 23cm. In May, 1986,
two or three randomly assigned subplots within each of the eleven plots were
sown with monocultures of five different grass species. The grasses were
chosen for study because they represent the five most abundant grasses in the
successional grasslands at CCNHA. These species were: Agrostis scabra
Willd., a native C3 species common in early successional fields; Agropyron
repens (L.)Beauv. and Poa pratensis L., Eurasian C3 species common in
early- and mid- successional sites respectively; Schizachyrium
scoparium (Michx.) Nash-Gould and Andropogon gerardi Vitm., native
C4 species common in late successional sites.