John B. Taft

Savanna and Open-Woodland Communities

Vegetation maps including the upper Midwest generally show two basic vegetation formations: prairie and forest (Vestal 1936, Shelford 1963, Anderson 1970, Iverson et al. 1991). The sharp dividing boundaries implied are more a matter of convenience of scale and difficulties in mapping variable boundaries than a reflection of reality. Of course, the fires that contributed largely to the maintenance of the tallgrass prairies of this region (Risser et al. 1981, Axelrod 1985) usually did not stop abruptly at a forest border. These fires penetrated beyond the open grassland, often forming structural gradients from open prairie to closed forest. The kinetic quality of the prairie-forest transition zone has challenged our spatialand temporal-scale ecological understanding, yielding variable interpretations of vegetational history (Bielmann and Brenner 1951, Steyermark 1959), distribution (Braun 1950, Anderson 1983, Nuzzo 1986), and classification (White and Madany 1978; Nelson 1985; Faber-Langendoen 1994). A mixture of climatic and landscape conditions, fire history, and biotic interactions results in a mosaic pattern of forest, savanna, and prairie in the Midwest (Kilburn 1959, Anderson 1983) that is considered a prairie-forest ecotone (Barbour et al. 1980).

Vegetation ecology of flatwoods on the Illinoian till plain

Midwestern flatwoods are open woodlands of level uplands and terraces that occur on impervious subsoil hori- zons and have seasonally wet and dry soils. Vegetation and soil data from six flatwoods remnants on the Illinoian till plain in Illinois were examined to identify what biotic and edaphic factors control vegetation structure, composition and diver- sity. Soils at all sites are characterized by the presence of argillic horizons (clay pans). Two soil ordination groups were clearly defined, each characterized by separate parent material origin. Canopy and subcanopy composition are dissimilar at sites with soils developed entirely in loess; more similar at sites on lacustrine soils (including aeolian sand). Quercus stellata was by far the most important tree species followed by Q. marilandica. Q. stellata importance was negatively corre- lated with Ca, particularly in the A-horizon. Q. marilandica shows a positive correlation to sand and magnesium, particu- larly in the B-horizon, and was most frequent and abundant on soils where Ca/Mg ratios are lower than 1.0 in the E and B- horizons. Total stem density of woody plants was positively related to total available soil moisture as influenced by depth to clay pan and soil textural differences. Ground-cover species richness and diversity were inversely related to tree and total woody stem density and tree diversity. One site with a recent history of annual prescribed fire had significantly greater ground-cover species diversity compared with the unburned sites. Evidence from stand structure and similarity between canopy and subcanopy strata, if reflecting stationary processes of canopy recruitment, suggest that disturbances historically maintained the importance of oak species in most flatwoods. Prescribed fire appears to maintain or enhance ground-cover species diversity at one flatwoods remnant, suggesting that fire could be used to maintain compositional stability at sites with greater available soil moisture.

Evidence for community structuring associated with the actinorhizal shrub Ceanothus americanus in tallgrass prairies in Illinois, USA

The role of the actinorhizal shrub Ceanothus americanus L. in community organisation of native grasslands was examined in three Illinois tallgrass prairie remnants. We asked can C. americanus structure tallgrass prairie at the patch scale through nitrogen fixation leading to differences in species assemblages and diversity? Expectations were that warm season (C4) graminoid species would be more abundant and species diversity would be greater outside Ceanothus patches than in patches associated with C. americanus where, based on results from numerous nitrogen augmentation and deposition studies, we expected greater abundance of cool season (C3) graminoid species and lower species diversity. Plots with and without association to C. americanus were compared for differences in floristic similarity, diversity, and C3 versus C4 graminoid species abundance. C3 graminoid species, including the adventive Poa pratensis L., were significantly more abundant in prairie associated with C. americanus than were C4 species. C4 species, such as Andropogon gerardi Vitman and Schizachyrium scopariumii (Michx.) Nash., were more abundant than C3 graminoid species in prairie patches without association to C. americanus. Means for diversity among plots associated with C. americanus were significantly lower than for other plots, nearly so for evenness and species density, but not different for sum total cover or total species richness. There were no statistically-significant differences in measures of mean total soil nitrogen at points along linear transects radiating from the base of C. americanus shrubs nor between mean soil amino-sugar-nitrogen concentrations within and outside of C. americanus patches. Nonetheless, comparative evidence supports the hypothesis that C. americanus structures composition and diversity at the local patch scale in tallgrass prairie.

Fire Effects on Structure, Composition, and Diversity in a South-Central Illinois Flatwoods Remnant

Abstract Fire effects on structure, composition, and species diversity in canopy, shrub/sapling, and ground-cover strata were examined following three burns over an eight-year period in a small (6 ha), isolated flatwoods remnant in south-central Illinois. Prior to fire treatments, size-class distribution patterns for trees indicated two species groups, one comprised of oaks dominating the larger size classes, particularly Quercus stellata, and a second group of non-oak species mostly from the small-to-medium size classes. Ground cover was sparse, comprised mostly of tree seedlings and woody vines, and shade-intolerant herbaceous species were absent. Following three burns, total tree density (stems ≥ 6 cm) declined 26% from 465/ha; however, basal area increased from 24.7 m2/ha to 25.6 m2/ha as tree mortality mostly was confined to small-diameter stems. Acer saccharum was the only tree species not to decline in density in the fire-treatment area. While 15 of 20 species in the shrub/sapling stratum declined in density in the fire-treatment area, total stem density increased 170% from 7,133 to 19,227 stems/ha due to increases of Rubus allegheniensis and Sassafras albidum. Overall ground-cover species richness increased 81% following the three burns from 42 to 76 species while species density nearly tripled from 3.2 to 9.4 species/0.5-m2. This increase included many characteristic flatwoods species; however, there also was an increase in ruderal species. Perennial forbs comprised the physiognomic class with greatest frequency following fire treatment and the most species-rich group; other groups that increased greatly in frequency and species richness were biennial and annual forbs and perennial sedges.

A test of concordance in community structure between leafhoppers and grasslands in the central Tien Shan Mountains

Leafhoppers and related Auchenorrhynchous Hemiptera (AH) are among the most diverse grassland herbivores, and many have been linked inexorably to grassland vegetation through diet and shelter for millions of years. Are AH–plant communities in natural grasslands tightly integrated, how does the interaction differ across major ecological gradients, and do habitat or environmental factors explain the most variance in AH community structure? These questions have implications for the conservation of biodiversity and in evaluating effects of a warming climate. Using grasslands of the central Tien Shan Mountains as a natural laboratory, we examine whether AH species assemblages are concordant with vegetation in terms of community structure using closely associated species-level samples. Data were recorded from a nearly 3000-m elevation gradient crossing four arid and three montane grassland vegetation classes. We found elements of AH–plant community classification and structure to be closely correlated except for at the arid–montane habitat transition where a small group of widespread AH species were significant indicators for vegetation classes in both major grassland types. AH species richness and abundance are positively correlated with plant species density and percent cover and, correspondingly, peak at mid-elevations in association with montane grasslands. While overall elevation (and covariate mean annual temperature) explains the most variance in AH species assemblages, the sum total of habitat factors explain more variance than environmental factors when arid and montane grasslands were examined separately, but environmental factors are co-equal with habitat factors when the grassland types are combined. Unexplained variance in the AH community assemblages, attributable to individualistic species responses to environmental and habitat factors, slightly exceeds the total accounted for by the model.

Are small, isolated prairie remnants effectively smaller than they look and getting smaller?1,2

Abstract To evaluate the floristic sustainability of three small, isolated tallgrass prairie remnants, spatial and temporal patterns of species and functional group (FG) composition, richness, diversity, and dominance were examined to determine whether there was divergence between edge (outer 10 m) and interior zones. These remnants, all protected and managed nature preserves, are rectangular in shape and range in area from 1.29 ha to 1.77 ha; each is bordered by cropland and unmanaged or mowed ruderal grassland. A stratified sampling design was established using permanently marked transects oriented perpendicular to the long axis of each site. Vegetation data were collected in 50 cm × 50 cm quadrats along each transect; prairies were resampled 5–6 yr following baseline data collection. The null expectations were no differences between edge and interior zones, no differences between edge type (crop and noncrop), and no change over time. Results from linear mixed-effects models identified significant differences between edge and interior zones for native species density, native diversity, dominance, percent native cover, and measures of floristic quality; however, there were no time differences and no significant zone × time interactions. Distance from edge to peak levels of species density and diversity was about 15 m yielding core habitats 43% to 50% of remnant size. Beta diversity based on species presence was greater in edge zones compared to interior zones, and declined slightly in repeated samples. Functional group diversity was lower in edge zones compared to interior zones and declined significantly in the repeated samples. Functional groups with significant affiliation to the edge zone were nonnative grasses, woody vines, ferns/allies, and nonnative legumes; FGs affiliated with the interior zone were C4 grasses, hemiparasites, nitrogen-fixing shrubs, and perennial forbs. One site was sampled three times following 5–6 yr intervals and there were no additional within-site differences in the third sample. These prairies support a rich diversity of native species; however, patterns suggest they are effectively smaller than they appear with a trend of declining functional diversity.

A Field Test of Regeneration in an Isolated Population of Violet Collinsia (Collinsia violacea)

Violet collinsia (Collinsia violacea) is a winter annual ranging primarily from Missouri to Kansas. An isolated population occurs in central Illinois, 200 km from the nearest Missouri population, in a series of colonies on adjoining river bluffs. Its oak woodland habitat was undergoing vegetation changes consistent with mesophication including establishment of a subcanopy of sugar maple (Acer saccharum). We conducted a field study on one Illinois colony to determine violet collinsia response to early dormant-season fire and to test whether this species might be limited by leaf litter accumulation or absence of fire stimulus. We collected baseline data on plant density and frequency using a stratified-random grid of 72 plots (0.5 m2) overlaying the colony. We applied two experimental treatments, fire and leaf litter removal, and a no-treatment control, each with 24 replicates. We used a sheet-metal burn box in the replicate burn plots and a garden rake in leaf-litter removal plots. Only the burn plots had significant changes immediately following treatments, increasing tenfold (12 to 123 plants) between the baseline and post-treatment growing season, and increasing in frequency from 17% to 67%. Post-treatment monitoring indicated there were no carry-over treatment effects. Violet collinsia appears to be a fire-annual with the capacity for persistence in soil seed banks. Following the experiment, a prescribed burn implemented throughout the study unit during the early dormant season also yielded a tenfold increase in the colony.