1994 Proceedings
North American Conference on Savannas and Barrens

In the 1970s and 1980s, when conservationists began attempting the ecological restoration of former oak savannas, their efforts were hampered by a lack of information on groundlayer composition and dynamics (Packard 1988a, 1988b, 1993), especially on rich or moist soil. Ecologists had studied few upper-Midwestern oak savannas because there were few sites thought to resemble presettlement conditions (Curtis 1959), and probably few ecologists were prepared to deal with the internal heterogeneity of oak savannas.

Bray (1955, 1958, 1960; Curtis 1959) made the best earliest attempt to examine the distribution of species along a sun-shade gradient. However, the research had three problems: a) the method of quantifying light was crude by today's standards, b) as Packard (1988a) argues, Bray's criterion that sites contain abundant prairie grasses resulted in certain types of savannas being omitted. and c) few mesic and no hydric savannas were studied.

Today, restorationists still face several crucial, unanswered questions. What is the appropriate pool of plant species with which to attempt a savanna restoration, especially on rich or moist soil? Given the heterogeneous environment within savannas, which plant species are appropriate where? What are the dynamic processes required for savanna re-creation and maintenance?

Researchers in the Thomas Givnish Laboratory (Department of Botany) at the University of Wisconsin-Madison, have began addressing these questions with three approaches:

1. Investigating historic records.
   
   
2. Studying savanna remnants. Pruka (1994) documented species abundances across light and soil-depth gradients in a thin-soil remnant. Peter Hujik is examining species distributions in three wet savannas. I am investigating fine-scale community composition along light-gradients in several Wisconsin remnants on a variety of soils and moistures.
   
   
3. Installing restoration experiments tracking competitive success along sun-shade gradients. Judith Maxwell has began restoration experiments on sandy soils. I established experimental community restorations on a variety of soil types at the University of Wisconsin Arboretum, Madison.

This paper reports my investigations into the distribution and dynamics of groundlayer plants along gradients of sun-shade and soils in savanna remnants and in restoration experiments.

METHODS

Ground-Layer Restoration Experiments

In the fall of 1992, 412 permanent 1-m2 "mini-restoration" plots were established at the University of Wisconsin Arboretum in Madison These quadrats were arranged in 38 strips from shadier places under oaks (and one hackberry) to sunnier places between trees. The strips were fenced to discourage disturbance by humans, deer, woodchucks, and rabbits. Sites are located on four soils: rich silt-loam, poor silt-loam, sandy loam, and loamy sand.

First, the strips were devegetated. Then in November, 1992, each quadrat was hand seeded with carefully weighed seed mixtures of around 80 native species (two mixes: one mesic planted on loams, one xeric planted on sandy loam and loamy sand). The seed mixes included large numbers of "woodland," "savanna," and "prairie" species (as reported by Curtis 1959; Packard 1988a, 1988b), and a large degree of interspecific variation in life history, seed weight, leaf height, leaf size, leaf orientation, growth form, ability to fix nitrogen, and so on.

In 1993, each quadrat was sampled in spring, summer, and fall. Cover, height, and presence of flowers or fruits were recorded for each species. I am repeating that sampling in 1994. The light regime was quantified using computer analysis of hemispherical canopy photographs (Chazdon and Field 1987) and estimated slope and aspect.

Remnant Observations

I have began vegetation sampling in 11 remnants in eight Wisconsin counties (Table 1). These remnants occur on a wide range of soil and moisture conditions. Each remnant is dominated by open-grown oaks. Unlike Bray's work where site selection was based on the presence of a specific groundlayer species group, the ground layer vegetation was dominated by native species in both sun and shade. I did biase site selection toward species-diverse sites (thus avoiding sites where Carex pensylvanica forms near monocultures in shade) and toward sites containing plants rare in Wisconsin (e.g., Camassia scilloides and Polytaenia nuttalii). Most sites had been burned recently, and several managed to control brush.

For each remnant, 50 to 100 1-m2 quadrats were sampled for the same kinds of data described above for the planted quadrats. Additionally, data on canopy tree species, DBH, and height were recorded. I used point sampling to record species, abundance, and height of shrubs and saplings. Soil samples for all remnants were analyzed for texture, pH, organic matter, P, K, Ca, Mg, N, and SO4-S. A species list is being compiled for these intensively sampled remnants and for other remnants.

RESULTS AND DISCUSSION

This research is in progress. The results and discussion presented here should be considered preliminary.

The savanna remnants were species rich, often containing more than 120 native vascular plant species in a few acres or less and sometimes 20 to 30 species/m2. Each species had its own pattern of abundance across the sun-shade gradient.

In the restoration plots, many first-year dominants (Table 2) were abundant across the entire light gradient. Several species showed skewed distributions from the start (e.g., Andropogon gerardi, a C4 species was absent from low-light quadrats).

In the remnants, species distributions along sun-shade gradients varied between sites, generally in agreement with Bray's observation that species shift toward shade on increasingly dry sites. Similarly,f or many species in the restorations, their cover, height, and flowering/fruiting varied along both sun-shade and soils gradients. Sometimes such patterns were compounded by disturbances; on the richest silt-loam, the success of Campanula americana appeared to be positively correlated with the distance from woodchuck dens.

The same species that were broadly distributed across the light gradient in the restoration plots were often narrowly distributed in remnants. I expect the early dominants in the restorations will, as slower growing species mature, be "sorted out" by competitive forces. In the restoration plots, competitive exclusion was evident in 1994 with the disappearance of some "fugitive" species that were abundant in 1993 (e.g., Conyza canadensis). In remnants, species that typically were most abundant in semi-shade (e.g., Lupinus perennis, Camassia scilloides), were sometimes abundant in full sun where other plant cover was sparse.

These findings suggest that when designing a savanna restoration, the important gradients to consider are sun-shade, soil fertility/moisture and competition/disturbance. I plan to continue monitoring the restorations and selected remnants on a range of soils to address the importance of disturbance and competitive processes in maintaining local diversity. In addition to the topics mentioned here, future papers will address the differential adaptation among savanna herbs to light intensity.

ACKNOWLEDGMENTS

This research benefited greatly from discussions with Thomas Givnish and the members of his lab. Ted Cochrane and other Wisconsin Herbarium staff were very kind and helpful in verifying and naming plant specimens. Major funding came from the Friends of the Arboretum, with additional support from the U.W. Botany Department Davis Fund, U.W. Arboretum, and The Nature Conservancy.

LITERATURE CITED

Bray, J. R. 1955. The savanna vegetation of Wisconsin and an application of the concepts of order and complexity to the field of ecology. Ph.D. Thesis. Department of Botany, University of Wisconsin, Madison.

• 1958. The distribution of savanna species in relation to light intensity. Canadian Journal of Botany 36:671-681.
• 1960. The composition of savanna vegetation in Wisconsin. Ecology 41:721-732.

Chazdon, R. L., and C. B. Field. 1987. Photographic estimation of photosynthetically active radiation: evaluation of a computerized technique. Oecologia 73:525-532.

Curtis, J. T. 1959. The Vegetation of Wisconsin: An Ordination of Plant Communities. University of Wisconsin Press, Madison.

Packard, S. 1988a. Just a few oddball species: restoration and rediscovery of the tallgrass savanna. Restoration & Management Notes 6:130-24.

• 1988b. Rediscovering the tallgrass savanna of Illinois. In: A. Davis and G. Stanford, eds. Proceedings of the Tenth North American Prairie Conference.
• 1993. Restoring oak ecosystems. Restoration and Management Notes 11(1):5-16.

Pruka, B. 1994. Distribution of understory plant species along light and soil depth gradients in an upland oak savanna remnant in southern Wisconsin. M.S. Thesis. Department of Botany, University of Wisconsin, Madison.

Table 1. Remnant savannas included in study. Moisture conditions are subjective descriptions. Not all sites have been tested for soil texture.