Printer Friendly

Holocene paleoenvironments of northeast Iowa.

INTRODUCTION

This paper presents independent lines of evidence of Holocene paleoenvironments in northeastern Iowa [ILLUSTRATION FOR FIGURE 1 OMITTED] from alluvial stratigraphy, soil development, pollen, vascular-plant and bryophyte macrofossils, beetles, and chronology of sediments exposed along Roberts Creek, along with the carbon and oxygen isotope records from speleothems in nearby Coldwater Cave. Pollen, alluvial stratigraphy, and the speleothem-isotopic record show regional patterns; plant macrofossils, insects, and soils reflect both regional and local environments but generally give a better picture of local habitats. This multidisciplinary approach ties landscape and soil development with biota, water quality, and isotope paleoclimatology, giving an integrated picture of Holocene paleoenvironmental change along a sensitive ecotonal area between prairie and deciduous forest.

Paleoecological records are irregularly distributed across the Midwest. Palynological studies are most numerous, and pollen sites are densest in Minnesota (where both palynologists and sites are abundant, and where sediment is rich in pollen and easy to extract), with fewer in Wisconsin and very few in Iowa and Illinois (fewer sites and palynologists, sparser pollen and more difficult extraction) (see Webb et al. 1983; Holloway and Bryant 1985). As a result, the Midwestern palynological sequence has been largely derived and extrapolated from sites in Minnesota and Wisconsin (Webb et al. 1983). Semiquantitative paleoclimatic reconstructions have used these same data (Bartlein et al. 1984) and extrapolations have been made to areas of less coverage (Bernabo and Webb 1977, Webb et al. 1983, Bartlein et al. 1984). Our recent research in eastern Iowa indicates that these extrapolations may lead to erroneous paleoclimatic reconstructions (Baker et al. 1990, 1992, Chumbley et al. 1990).

Fossil insects have also proved to be sensitive indicators of environmental change in the Midwest (Ashworth et al. 1981, Schwert et al. 1985, Baker et al. 1986, Schwert 1992), but few Holocene insect assemblages have been described, and none encompasses the entire Holocene.

L. A. Gonzalez and M. K. Reagan (unpublished data) and Dorale (1992) have begun detailed investigations of isotopic records preserved in cave speleothems (mainly stalagmites) in the Midwest. The 18O record in stalagmites from deep caves provides a long-term record of the mean annual surface temperature; 13C values are sensitive to the proportions of [C.sub.3] and [C.sub.4] plants whose remains are incorporated in the soil above the cave, and they provide additional evidence of the vegetational history of the region. The paleoclimatic record from Coldwater Cave [ILLUSTRATION FOR FIGURE 1 OMITTED] isotopes covers most of the Holocene and provides an independent test of both biotic and paleoclimatic records from Roberts Creek.

The paleoecology of Iowa, southern Wisconsin, and Illinois is important in understanding paleoenvironments of the Midwest because (1) many taxa that were forced south during glaciation moved through these regions to their present position following deglaciation; (2) the paleoecology of the Driftless Area south into much of Illinois and eastern Iowa is poorly known; and (3) the axis of the "Prairie Peninsula" (Transeau 1935) runs through this region. The history of prairie establishment in this Peninsula and subsequent adjustments in distribution are known largely from Minnesota, where pollen-rich sites are common along the prairie-forest border (e.g., Grimm 1983, Almendinger 1992). However, the area from southern Minnesota to central Illinois has none of the lakes and bogs typically used for pollen analysis because it remained free of Late-Wisconsinan ice. The discovery of organic materials in sediments in cutbanks along Roberts Creek (Bettis 1984), northeastern Iowa [ILLUSTRATION FOR FIGURE 1 OMITTED], led us to examine these as sources of paleoecological information.

The type of fossiliferous deposits discussed in this paper are not well described in the literature, and they require certain geologic conditions that allow for fossil preservation. The taphonomy of these alluvial sites is quite different from that of lacustrine sites typical of midwestern paleoecological studies. For example, the abundance of beaver-gnawed wood in some of the sites and the configuration of cutbank exposures at others indicate that the sediments were deposited in beaverdam ponds or as pool and point-bar deposits between riffles. We therefore describe both the bedrock and the alluvial setting of this area to aid in understanding the taphonomy and in the search for similar rich deposits in other areas.

To construct a chronology, each cutbank section was radiocarbon dated, and multiple dates from several sections indicate that the sediments in each were deposited in 200 yr or less. A total of 61 dates from many sections span the entire Holocene and late-glacial. We follow Tornqvist et al. (1992) in dating macrofossils (52 of our dates were on wood) rather than bulk sediment to avoid most sources of error.

Chumbley et al. (1990) briefly outlined the pollen evidence in these Roberts Creek cutbank deposits. The present paper (1) documents the environmental history of the Roberts Creek sites by using more detailed pollen diagrams along with analyses of vascular plant macrofossils, bryophytes, insects, and sediments, (2) ties in the chronology of biotic changes with soil formation and alluvial history in the basin, (3) evaluates the paleoclimatic significance by comparing the paleoclimatic signal in the sediments from Roberts Creek with that in an isotopic record from speleothems in nearby Coldwater Cave, and (4) discusses the extent of changes that occurred in the biota and in the landscape following human settlement of the area.

METHODS

Pollen samples were obtained from vertical sections of 32 cutbanks along Roberts Creek at intervals ranging from 5 to 20 cm. Sediment was treated for pollen analysis with KOH, HCl, HE and acetolysis solution, as described in Faegri et al. (1989), but further treatments included screening with a 7-[[micro]meter] mesh sieve, flotation with Zn[Cl.sub.2], and treatment with 0.2% Clorox bleach, as outlined by Chumbley (1989). Twenty-four of these sites contained pollen. The diagrams were plotted with Tiliagraph and zones were chosen with the aid of stratigraphically constrained cluster analysis (Grimm 1987).

At least two samples from each of the 32 sites were collected for plant-macrofossil analysis and stored in [approximately equal to]4-L zip-sealed bags. Macrofossils of vascular plants and bryophytes were screened through 0.5- and 0.1-mm mesh sieves and picked by hand. Twenty-seven of these sites contained macrofossils. Pollen and vascular plant macrofossils were identified by comparison with modern reference collections at the University of Iowa Geology Department. Specimens are in the Department of Geology Repository, University of Iowa, along with a complete list of the [approximately equal to]250 taxa of vascular plant macrofossils (including those not in tables or on diagrams); these data also will be available from the North American Macrofossil Database. Vascular plant nomenclature (Table 1) generally follows Gleason and Cronquist (1991). Bryophytes were identified using modern reference collections in the University of Iowa Herbarium (IA). Specimens are preserved in polyvinyl lactophenol on permanent slides reposited in a wooden microscope slide box in IA. Bryophyte nomenclature follows Anderson et al. (1990).

From 27 sites within the Roberts Creek basin, bulk samples of sediment, averaging 25 kg/locality, were collected for insect analyses. For most sites, sampling was undertaken simultaneously with that for pollen and plant macrofossils. Chitinous parts were separated from the sediments by wet-sieving and kerosene flotation (Ashworth 1979). The excellent preservation of the remains permitted generic or specific determination of most of the beetles through comparison with pinned specimens at several major museum collections. The disarticulated remains, mostly heads, pronota, and elytra of beetles, are reposited on microslides and in vials at the Quaternary Entomology Laboratory, North Dakota State University. A complete list is available on microfiche from DPS and is on file at the Quaternary Entomology Laboratory, North Dakota State University.
TABLE 1. Vascular plants and bryophytes mentioned on diagrams or in
the text.


Species                                         Common name


Vascular plants


Abies                               fir
Acalypha                            three-seeded mercury
Acer negundo L.                     boxelder
Acer saccharum L.                   sugar maple
Adiantum                            maidenhair fern
Amaranthus                          amaranth
Ambrosia                            ragweed
Amphicarpaea bracteata (L.)         hog-peanut
Fern.
Amorpha canescens Pursh.            lead plant
Andropogon gerardii Vitman          big bluestem
Anemone canadensis L.               Canada anemone
Aralia racemosa L.                  spikenard
Artemisia                           sagebrush, wormwood
Bidens cernua L.                    nodding bur marigold
Boehmeria cylindrica (L.) Sw.       false nettle
Callitriche                         water starwort
Campanula americana                 American bell-flower
Carex haydenii Dewey                sedge
Carex meadii Dewey                  sedge
Carpinus caroliniana Walter.        hornbeam
Carya                               hickory
Ceanothus americanus L.             New Jersey tea
Cerastium arvense L.                chickweed
Cerastium nutans Raf.               chickweed
Ceratophyllum demersum L.           coontail
Chara                               stonewort
Chenopodiineae                      goosefoot super-order
Chenopodium                         goosefoot
Chenopodium berlandieri Moq.        goosefoot
Chenopodium rubrum L.               goosefoot
Cicuta maculata L.                  water-hemlock
Claytonia virginica L.              spring beauty
Cornus                              dogwood
Cornus amomum ssp. obliqua          silky dogwood
Miller
Cornus rugosa Lam.                  dogwood
Corylus                             hazel
Corylus americana Walter.           American hazel
Cryptotaenia canadensis (L.)        honewort
DC.
Cyperaceae                          sedges
Dalea candida Michx.                white prairie clover
Dalea purpurea Vent.                purple prairie clover
Desmodium canadense (L.) DC.        tick-trefoil
Dicanthelium oligosanthes
Dryopteris                          fern
Echinochloa                         barnyard grass
Echinochloa muricata (Beauv.)
Fern.
Echinocystis lobata (Michx.) T.     wild cucumber
& G.
Eleocharis palustris (L.) R. &      spike-rush
S.
Eleocharis tenuis (Willd.) Short    spike-rush
Eupatorium altissimum L.
Eupatorium maculatum L.             Joe-pye-weed
Eupatorium perfoliatum L.           boneset
Eupatorium purpureum L.             Joe-pye-weed
Euphorbia corollata L.              flowering spurge
Euphorbia maculata L.               wartweed
Fraxinus nigra Marshall             black ash
Fraxinus pennsylvanica Mar-         green ash
shall
Glyceria grandis S. Wats.           American manna grass
Glyceria striata (Lam.) A.          slender manna grass
Hitchc.
Helenium autumnale L.               sneezeweed
Helianthus                          sunflower
Helianthus annuus L.                common sunflower
Helianthus grosseserratus Mar-      big-toothed sunflower
tens
Hippurus vulgaris L.                horsetail
Hypericum pyramidatum Ait.          St. John's-wort
Hypoxis hirsuta (L.) Cov.           yellow stargrass
Impatiens                           jewel-weed
Iodanthus pinnatifidus (Michx.)     mustard
Steudel.
Juglans cinerea L.                  butternut
Laportea canadensis (L.) Wedd.      wood nettle
Larix                               larch
Leersia oryzoides (L.) Swartz.      rice cutgrass
Lemna                               duckweed
Liatris aspera Michx.               blazing star
Linum sulcatum Riddell              flax
Lobelia spicata Lam.                pale-spike lobelia
Lycopus americanus Muhl.            water-horehound
Lysimachia ciliata L.               loosestrife
Mentha arvensis L.                  mint
Monarda punctata L.                 horse-mint
Monarda fistulosa L.                wild bergamot
Myriophyllum                        water-milfoil
Myriophyllum pinnatum (Wal-         water-milfoil
ter) BSP
Naias flexilis (Willd.) Rostkov     naiad
and Schmidt
Ostrya virginiana (Miller) K.       hop-hornbeam
Koch
Oxalis                              wood sorrel
Pastinaca sativa L.                 parsnip
Phlox pilosa L.                     prairie phlox
Physocarpus opulifolius (L.)        ninebark
Maxim.
Picea                               spruce
Picea glauca (Moench) Voss          white spruce
Picea mariana (Miller) BSP          black spruce
Pilea pumila (L.) A. Gray           clearweed
Pinus                               pine
Poaceae                             grass family
Polygala verticillata L.            whorled milkwort
Polygonum aviculare L.              knotweed
Polygonum coccineum Muhl.           water smartweed
Polygonum convolvulus L.            black bindweed
Polygonum hydropiper L.             smartweed
Polygonum lapathifolium L.          smartweed
Polygonum pensylvanicum L.          smartweed
Polygonum punctatum Ell.            smartweed
Polygonum scandens L.               false buckwheat
Populus                             poplar
Portulaca oleracea L.               purslane
Potamogeton                         pondweed
Potamogeton foliosus Raf.           pondweed
Potentilla arguta Pursh.            prairie cinquefoil
Potentilla norvegica L.             rough-fruited cinquefoil
Potentilla palustris (L.) Scop.     swamp cinquefoil
Prunus virginiana L.                choke cherry
Pteridium                           bracken fern
Pycnanthemum virginianum (L.)       mountain-mint
Durand & Jackson
Quercus                             oak
Ranunculus abortivus L.             wood buttercup
Ranunculus aquatilis L.             white water-crowfoot
Ranunculus hispidus Michx.          swamp buttercup
Ranunculus pensylvanicus L.f.       buttercup
Ranunculus sceleratus L.            cursed crowfoot
Ratibida pinnata (Vent.) Barn-      gray-headed coneflower
hart
Rhus typhina L.                     staghorn sumac
Rubus                               raspberry, blackberry
Rudbeckia hirta L.                  black-eyed susan
Sagittaria                          arrowhead
Sagittaria engelmanniana J.G.       arrowhead
Smith
Sagittaria latifolia Willd.         arrowhead
Salix                               willow
Sambucus canadensis L.              elderberry
Schizachyrium scoparium             little bluestem
(Michx.) Nash
Scirpis acutus Muhl                 bulrush
Scirpus atrovirens Willd.           bulrush
Scirpus validus Vahl.               bulrush
Setaria glauca (L.) P. Beauv.       foxtail
Setaria viridis (L.) P. Beauv.      foxtail
Silphium laciniatum L.              compass plant
Sium suave Walter                   water parsnip
Sorghastrum nutans (L.) Nash        Indian grass
Sparganium                          bur-reed
Staphylea trifolia L.               bladder-nut
Stipa spartea Trin.                 needle and thread
Taenidia integerrima (L.) Drude     yellow pimpernel
Taraxacum officinale Weber          dandelion
Taxus                               yew
Thalictrum                          meadow rue
Thalictrum dasycarpum Fisch.        meadow rue
& Ave-Lall.
Tilia                               basswood
Tubuliflorae                        subfamily of daisy family
Ulmus                               elm
Ulmus americana L.                  American elm
Urtica dioica L.                    stinging nettle
Verbena hastata L.                  common vervain
Verbena urticifolia L.              white vervain
Veronicastrum virginicum (L.)       Culver's root
Farw.
Viburnum lentago L.                 nannyberry
Vitis                               grape
Zanthoxylum americanum Miller       prickly ash
Zizania aquatica L.                 wild rice
Zizia aurea (L.) Koch               golden alexander
Zosterella dubia (Jacq.) Small      water star-grass


Bryophytes


Anomodon sp.
Anomodon cf. attenuatus (Hedw.) Hueb.
Anomodon minor (Hedw.) Fuern.
cf. Astomum/Weissia sp.
Barbula cf. unguiculata Hedw.
Brachythecium acuminatum (Hedw.) Aust.
cf. Brachythecium reflexum (Starke in Web. & Mohr)
Schimp. in B.S.G.
Brachythecium cf. rutabulum (Hedw.) Schimp. in B.S.G.
Bryum sp.
Bryum cf. caespiticium Hedw.
Calliergon giganteum (Schimp.) Kindb.
cf. Campylium chrysophyllum (Brid.) J. Lange
Ceratodon purpureus (Hedw.) Brid.
cf. Dicranella sp.
Drepanocladus aduncus (Hedw.) Warnst. var. kneiffii
(Schimp. in B.S.G.) Moenk.
Drepanocladus aduncus (Hedw.) Warnst. var. polycarpus
(Bland. ex Voit) G. Roth
Eurhynchium hians (Hedw.) Sande Lac.
Fissidens sp.
Fissidens bryoides Hedw.
Fissidens obtusifolius Wils. var. apiculatus Grout
Haplocladium cf. microphyllum (Hedw.) Broth.
Hygroamblystegium noterophilum (Sull. & Lesq. in Sull.)
Warnst.
Hygroamblystegium tenax (Hedw.) Jenn.
Hygroamblystegium tenax var. spinifolium (Schimp.) Jenn.
Hypnum cf. lindbergii Mitt.
Hypnum pratense (Rabenh.) W. Koch ex Spruce
Leptodictyum riparium (Hedw.) Warnst.
Lindbergia brachyptera (Mitt.) Kindb.
Philonotis fontana (Hedw.) Brid.
Philonotis fontana var. caespitosa (Jur.) Schimp.
Plagiomnium cf. cuspidatum (Hedw.) T. Kop.
Plagiomnium ellipticum (Brid.) T. Kop.
cf. Plagiothecium cavifolium (Brid.) Iwats.
Scorpidium scorpioides (Hedw.) Limpr.


Sediments and soils were described from cutbanks and in a few cases from cores 7.6 cm in diameter. Sixty-one radiocarbon ages were obtained, 52 from wood and 9 from bulk organic sediments dug from cutbanks along the creek (Table 2). These dates are uncalibrated in this paper and provide a robust chronology for the fossil sequences and alluvial stratigraphy. Multiple dates from five sites show that sediment at these sites accumulated within less than 200 radiocarbon years.

Methods for analyzing and dating stable isotopes in cave stalagmites are outlined in Dorale et al. (1992). The chronology for the stalagmites is derived from Uranium-Thorium-series dates, which are considered calendar dates and are expressed as YBP (years before present). In contrast, the radiocarbon chronology from Roberts Creek is influenced by changes in 14C production in the upper atmosphere through time; these ages differ slightly from calendar years and are expressed as yr BP.

Responsibilities for this research are as follows: Baker - vascular plant macrofossils and project oversight; Bettis - geology and alluvial history and chronology; Schwert - insects; Horton - bryophyte macrofossils; Chumbley - pollen; Gonzalez - stable isotopes; and Reagan - speleothem dating.

RESULTS

Geology of the Roberts Creek Basin

Roberts Creek drains a 358-[km.sup.2] tributary basin of the Turkey River [approximately equal to] 16 km west of the Mississippi Valley in the Paleozoic Plateau landform region of northeastern Iowa ([ILLUSTRATION FOR FIGURE 1 OMITTED]; Prior 1991). The region's landscape reflects a strong influence of underlying Paleozoic marine carbonate and shale bedrock. Near-surface carbonate rocks in many parts of this region, including the study area, contain sinkholes, caves, and sinking streams (Bounk and Bettis 1984). Master valleys and the lower portions of tributaries are narrow and deeply entrenched with steep valley walls. The upper portions [TABULAR DATA FOR TABLE 2 OMITTED] of tributaries, including Roberts Creek, are usually cut into rock but are not deeply entrenched; they tend to have broader valleys with more moderately sloping valley walls. Uplands are moderately sloping, and a thin (1-10 m thick) mantle of Pleistocene deposits covers the bedrock surface (Hallberg et al. 1984).

In the study area Roberts Creek flows along the contact of shaley Ordovician carbonates (Elgin Member, Maquoketa Formation) with underlying thinly bedded limestones of the upper Galena Group (Dubuque Member of Ordovician age; Hallberg et al. 1983). The piezometric surface (groundwater table) of the Galena Aquifer is intersected by the stream in this reach, and groundwater discharges into the stream from the bedrock units (Libra et al. 1992). Farther downstream Roberts Creek crosses a prominent trough in the Galena piezometric surface, and stream water is lost to the bedrock. This "losing reach" often dries up during extended periods of little or no precipitation, whereas moderate flow is maintained by groundwater discharge farther upstream in the "gaining reach." The preservation of organic plant remains along the valley is controlled by these variations in the position of the Galena piezometric surface. Sites yielding plant and insect remains are not present in the losing reach of the creek but are common upstream in the gaining reach, where the lower part of the alluvial fill and associated fossil zones have remained saturated. We have recently found that similar gaining reaches in other midwestern streams are equally fossiliferous.

The study area includes the upper one-third of Roberts Creek valley, with a drainage area of 77.6 [km.sup.2]. The creek heads just south of the town of Postville, Iowa and follows a southeasterly course toward the Turkey River. Local relief averages [approximately equal to]30 m, upland divides are narrow, and the landscape is moderately rolling. Uplands and most valley slopes are mantled with 2-3 m of Peoria (Late Wisconsinan) Loess, but shaley carbonate bedrock (Elgin Member) crops out where the valley wall is steep. The valley floor ranges in width from 0.5 to 0.9 km throughout most of this part of the valley, then narrows and maintains a width of [approximately equal to]0.5 km as it cuts into the Galena Group limestones. Most of the fossil localities are just upstream from this narrow reach.

Surface soils on the uplands and the valley margin are Mollisols (prairie soils) and Alfisols (forest soils) developed in loess and in shallow loess over bedrock (Kuehl 1982). Mollisols are dominant on the broader upland divides separating large tributaries, whereas Alfisols, including some with thick, dark-colored surface horizons (Mollic Hapludalfs), occur on most of the upland and valley slopes. The Mollic Hapludalfs are soils significantly influenced by both prairie and forest vegetation during their development.

Roberts Creek meanders within its valley, although several segments have been artificially straightened. The creek flows through a channel 5-7 m wide and 3-4 m deep. The stream has pool-and-riffle morphology, with low-gradient, quiet-water pools with mud bottoms separated by short riffles with steep gradients and cobble bottoms. During low-flow conditions in most summers, water depth ranges from 0.5-1 m in pools to 0.15-0.1 m in riffles. High-water conditions are usually associated with spring snowmelt and thunderstorms, whereas low-flow conditions, when most of the stream's flow is supplied by groundwater discharge, usually occur during the summer and early fall.

Alluvial history of the Roberts Creek Basin

Five surfaces are evident in Roberts Creek valley ([ILLUSTRATION FOR FIGURE 2 OMITTED], Table 3). The highest is a loess-mantled strath terrace cut on Galena Group limestone. This terrace is elevated [approximately equal to]7 m above the valley floor and is present only as small, isolated remnants in the study area. Alluvium beneath the Late Wisconsinan loess on this terrace is devoid of organic remains and [greater than]21 000 yr old (Bettis and Hallberg 1985). A second terrace, about 4 m above the floodplain, is discontinuously present through the study area. It is underlain by Peoria Loess, which grades downward into silty and sandy alluvium. The alluvium and overlying loess are devoid of organic remains, but the regional chronology of Peoria Loess deposition indicates that the alluvium is [less than]20 000 and greater than [approximately equal to]15 000 yr old (Hallberg et al. 1984).

The floodplain consists of three constructional surfaces separated by short, steep scarps [ILLUSTRATION FOR FIGURE 2 OMITTED]. The upper surface is relatively featureless and encompasses the greatest area of the valley floor. Inset 0.3-0.8 m below this level is a surface marked with shallow, filled abandoned meander scars and subtle undulations. This surface roughly parallels the modern (unstraightened) channel and marks a former channel belt of the creek. The lowest surface in the valley is the modern channel belt, separated from older surfaces by a scarp 1-2 m high and characterized by abandoned meanders that contain water, small natural levees, and point bars and chutes that are active to partially masked. The lowest surface is flooded annually.

The floodplain is underlain by 3-5 m of alluvium of the DeForest Formation of Late Wisconsinan to Holocene age (Bettis 1990). In this area four lithologically distinct alluvial fills are present, the Gunder, Roberts Creek, and Camp Creek members of the Formation, and an informally recognized unit, pre-Gunder alluvium [ILLUSTRATION FOR FIGURE 2 OMITTED]. Table 3 summarizes the lithologic properties (color, bedding structures, pedogenic alterations) of these alluvial fills. Pre-Gunder alluvium is present beneath the Gunder Member at some localities. With one exception, this unit was only encountered in borings, was [less than]1.5 m thick, and everywhere had an erosional upper boundary with the Gunder Member. A single radiocarbon date of 14 470 yr BP was obtained from within this unit.

Deposits of the Gunder Member averaging 3-4 m in thickness underlie the highest surface of the floodplain. Basal radiocarbon ages on wood range from 10 700 to 4100 yr BP (all radiocarbon ages reported are uncalibrated) (Table 2). The Roberts Creek Member, averaging 3-4 m in thickness, underlies the intermediate floodplain surface [ILLUSTRATION FOR FIGURE 2 OMITTED] and dates from [approximately equal to]3500 to 380 yr BP (Table 2). Camp Creek Member deposits underlie the lowest surface (channel belt) and lap up onto the higher floodplain surfaces underlain by Roberts Creek and Gunder Member deposits. Most of the Camp Creek deposits are Historic in age, but the lower portions of some sections accumulated before EuroAmerican settlement of the basin (Bettis 1990, Bettis et al. 1992, Baker et al. 1993).

The sedimentology of the Gunder and Roberts Creek Members is similar. Both units are characterized by a basal cobble gravel (1-1.5 m thick) composed of local carbonate clasts overlain by relatively thin (0.5-1.0 m thick) sandy channel deposits grading upward to fine-grained alluvium. Thin horizontal stratification is locally present in the transition zone from the sandy to the fine-grained deposits. Plant and insect fossils are present in the sandy channel deposits and in the overlying stratified transitional deposits. The thickest fossiliferous deposits (0.5-0.7 m thick) are associated with the transition-zone deposits that accumulated in pools ([ILLUSTRATION FOR FIGURE 3 OMITTED], Tables 3 and 4) during average and low-flow conditions. The fines and organics were rapidly buried during higher flow conditions as sediment-laden run-off dropped some of the sediment load (including more organic remains) in response to decreases in velocity, turbulence, and gradient. Some localities contain thick accumulations of sticks and twigs, many with beaver gnaw marks, and thick transition-zone deposits are present immediately upstream. We interpret these as pool deposits behind beaver dams (Coles 1992). Gradual channel migration across the floodplain, as well as meander abandonment by chute and neck cutoffs, isolated the fossil localities from the active channel area [ILLUSTRATION FOR FIGURE 3 OMITTED]. Once isolated from the active channel, the localities were further buried by overbank deposits during floods.

Most of the Camp Creek Member ([ILLUSTRATION FOR FIGURE 2 OMITTED], Table 3) accumulated following extreme agricultural disturbance and erosion of uplands and valley-wall slopes from the late 1800s to [approximately equal to]1930 (Baker et al. 1993). It overlies presettlement surface soils developed on older floodplain surfaces [ILLUSTRATION FOR FIGURE 2 OMITTED]. It is 0.1-0.44 m in average [TABULAR DATA FOR TABLE 3 OMITTED] thickness where it buries the Gunder Member, increasing to 1.0 m where it buries the Roberts Creek Member. Where it accumulated as overbank deposits during floods it does not contain plant or insect fossils, except in abandoned channel environments ([ILLUSTRATION FOR FIGURE 3 OMITTED], Table 4) that filled rapidly.

Gunder and Roberts Creek Member deposits both encompass alluvium associated with many channel positions over several thousand years. Through time, channel activity has continuously removed older alluvium and associated biotic remains. The remaining alluvial fill of the valley is therefore a complex of alluvium of various ages. Four broad age groups of alluvium can be recognized in the field on the basis of the stratigraphic framework of the DeForest Formation: that deposited between [approximately equal to]15 000 and 11 000 yr BP (informal pre-Gunder alluvium), 11 000 and 4000 yr BP (Gunder Member), 4000 to [approximately equal to]400 yr BP (Roberts Creek Member), and younger alluvium (Camp Creek Member). Differences in the morphology of soils developed in the Gunder Member are used to distinguish early (thick, well-expressed Bt horizons), middle (thin, well-expressed Bt horizons), and late middle (thin, transitional Bt/Bw horizons) Holocene localities (Table 3). By using both lithologic and pedologic criteria, we were able to choose alluvial samples that spanned the entire Holocene.

The taphonomy of the fossil sites is affected by both depositional and postdepositional processes, and these processes can be determined by examining lithologic characteristics of the sediments. Each DeForest Formation member can be subdivided into lithofacies that display a limited range of primary sedimentary structures, grain size, and/or secondary pedogenic features that reflect specific depositional settings (Table 3). In these Holocene alluvial deposits, the depositional setting controls the sedimentological conditions and taphonomic agents that initially determine the potential for the preservation of biotic records (Table 4).

Within these alluvial fills, six lithofacies can be recognized for the DeForest Formation in small valleys of the Upper Mississippi River basin (Table 4). Rarely are all lithofacies present in a single member at the same location, and in some cases a single lithofacies may be repeated in a vertical section. Fig. 3 illustrates the idealized occurrences of these lithofacies along a stream.

Postdepositional alterations have affected the Gunder and Roberts Creek members and associated biotic remains in different ways. Knox (1983, 1993) discussed the Holocene climatic history of the Upper Midwest as it relates to the behavior of the fluvial system, and Bettis (1992) discussed the role of long-term water-table relationships in the fossil record in Holocene alluvium. As a result of middle-to-late Holocene reductions in long-term precipitation, slight lowering of the regional water table, and lowering of the local floodplain water table as a result of pre-Roberts Creek channel widening, Gunder Member deposits have been oxidized (weathered) to a greater degree than Roberts Creek Member deposits.

Plant and insect fossils are absent from the oxidized portion of the Gunder Member because of postdepositional degradation, are poorly preserved in the mottled and reduced zone marking the long-term water-table fluctuation, and are well preserved only in the lower unoxidized portion of the unit that has remained saturated since deposition (Chumbley 1989). Organic remains are preserved through a greater thickness of the Roberts Creek Member because the water table has remained high since the unit was deposited and has minimized organic matter degradation.

Pollen

At most sites there was little change in pollen spectra from top to bottom, and the counts were amalgamated into a single count (Chumbley 1989). Pollen-percentage diagrams [ILLUSTRATION FOR FIGURES 4 and 5 OMITTED] show the vegetational history of the Roberts Creek area.

Zone 1 ([greater than]12 500 to [approximately equal to]9300 yr BP) occurs in the oldest sections in the pre-Gunder and Gunder Members. It is characterized by high percentages of arboreal pollen, especially Picea, with low but significant percentages of Larix and Fraxinus nigra. Among nonarboreal types (NAP), Cyperaceae, Dryopteris type, and Sphagnum, which makes its only appearance here, have high percentages in this zone. Although Pinus percentages reach their peak here, studies of modern pollen rain indicate that the percentages are far too low to indicate that pine was locally present (Webb and McAndrews 1972).

Zone 2 ([approximately equal to]9300 to 5500 yr BP) also occurs in the Gunder Member and contains maxima of pollen from deciduous trees, including Ulmus, Quercus, and Ostrya-Carpinus, with Tilia, and Acer saccharum becoming prominent in the upper part. Carya and Corylus are present, and fern spores including Dryopteris type, Adiantum, and Pteridium also reach peaks at the top of this zone. Chenopodiineae, Poaceae, and Ambrosia are NAP (nonarboreal pollen) taxa with low but persistent percentages during this interval.

Zone 3 (5500 to [approximately equal to]3500 yr BP) occurs in the youngest deposits of the Gunder Member, and the oldest deposits of the Roberts Creek Member alluvium. This zone is marked by minimum arboreal pollen percentages, the near disappearance of the deciduous-tree pollen so abundant in the previous zone, and the dominance of Poaceae and Ambrosia. Salix and Acer negundo, both riparian trees, have low peaks near the base of zone 3, and their percentages decrease shortly thereafter. Quercus has a similar pattern of occurrence and is the most common remaining arboreal element. Pollen of aquatic species, especially Myriophyllum and Sagittaria are most abundant in this zone.

Zone 4 ([approximately equal to]3500 to [approximately equal to]380 yr BP) also occurs in the Roberts Creek Member and is characterized by an increase in arboreal pollen, mainly Quercus. The riparian trees Fraxinus pennsylvanica, Acer negundo, and Salix also are present in this zone. Poaceae and Ambrosia pollen percentages decrease, but pollen of Poaceae remains relatively abundant, and Tubuliflorae, and Cyperaceae are prominent. Aquatic-plant pollen percentages drop at the base but increase again at the top of zone 4.

Zone 5 (Historic) is associated with the Camp Creek Member, postdating EuroAmerican settlement, which began [approximately equal to] 1838. The pollen sample dates to [approximately equal to] 1880, based on the identification of "Glidden's Barb, One-strand Variation" barbed wire that occurred near the base of the section (written communication from A. L. Fisk to D. P. Schwert, 12 August 1989). It contains the Ambrosia zone, widely recognized across the Midwest, along with a substantial peak in Poaceae pollen percentages and a decrease in arboreal pollen.

Vascular plant macrofossils

Taphonomic studies relating extant vegetation to plant remains in modern depositional sites show that vascular plant macrofossils from stream deposits accurately reflect not only aquatic and streamside vegetation but also upland habitats in prairie (Thomasson 1992, Baker and Drake 1994), deciduous forest (Burnham et al. 1992), and even tundra (West et al. 1993). Selected vascular plant macrofossils are grouped by habitat of the parent plants so that ecological changes in habitats through time are easier to follow [ILLUSTRATION FOR FIGURES 6-12 OMITTED]. The macrofossil shifts correspond well with the [TABULAR DATA FOR TABLE 4 OMITTED] pollen sequence, so the same five zones are used as on the pollen diagrams.

Zone 1 ([greater than]12 500 to [approximately equal to]9300 yr BP) is dominated by northern forest elements. Large numbers of spruce needles are present along with larch needles and Populus bud scales [ILLUSTRATION FOR FIGURE 6 OMITTED]. Cones of Picea glauca were common at site AB-34, dated at 10 700 yr BP (see Table 2), and beaver-gnawed wood samples were all spruce. A single core sample dated at 14 970 also had Picea glauca cones in it; this sample was unfortunately lost, and subsequent cores from the site never relocated the organic lens. A single Picea mariana cone was found at site AB-51, where ages of 10 150, 9720, 9340, and 8990 yr BP were obtained (Table 2). Abies appears at the transition to zone 2, along with Fraxinus nigra, and Ulmus americana [ILLUSTRATION FOR FIGURE 6 OMITTED]. Cerastium nutans, Ranunculus abortivus, and Ranunculus hispidus (= Ranunculus septentrionalis) are the only forest-floor herbs that occur well within the zone, but Aralia racemosa, Claytonia cf. virginica, Cryptotaenia canadensis, lodanthus pinnatifidus, and Laportea canadensis appear at the top [ILLUSTRATION FOR FIGURE 7 OMITTED]. Shrubs [ILLUSTRATION FOR FIGURE 8 OMITTED] are also not common in zone 1, with only Prunus virginiana, Rubus, and Taxus (single seed and needle from one site at transition with zone 2; not shown on diagram). A few of the most common ruderal plants [ILLUSTRATION FOR FIGURE 10 OMITTED] make their first appearance in this zone, including Acalypha, Cerastium cf. arvense, and Potentilla norvegica, which all are present later in the Holocene. Few aquatics occur in this zone [ILLUSTRATION FOR FIGURE 11 OMITTED]; Hippurus vulgaris and Myriophyllum are limited to zone 1. Wetland macrofossils [ILLUSTRATION FOR FIGURE 12 OMITTED] are relatively important, especially Sagittaria engelmanniana, Glyceria grandis, G. striata, Urtica dioica, and Scirpus atrovirens type. Potentilla palustris, Ranunculus pensylvanicus, and Ranunculus sceleratus appear only in this zone.

Zone 2 ([approximately equal to]9300-5500 yr BP) contains most of the macrofossils of deciduous trees. Only elm is abundant throughout (as buds and bud scales), but Tilia is also fairly consistent, though at low levels. Other mesic elements (Ostrya, Carpinus, and Juglans cinerea) are present only at the top of the zone. Leaf mats with large fragments of deciduous leaves, of which Ostrya virginiana was most abundant, are found at several of the sites dating between 6500 and 5500 yr BP. The presence of Picea and Larix needles in small numbers may represent the continued presence of these trees at the bottom of the zone, but they almost certainly are reworked above that level, where they are poorly preserved fragments. A wide range of herbs typical of shady, deciduous-forest floors are found predominantly in and immediately adjacent to this zone. The most consistent of these are Claytonia cf. virginica and Laportea canadensis, the latter most typical of floodplain forest understories. Shrubs in this zone include such forest and forest-edge taxa as Rubus, Zanthoxylum americanum, Rhus typhina, and Sambucus canadensis. A few of the more aggressive/weedy prairie taxa such as Monarda fistulosa, Rudbeckia hirta, and Zizea aurea are very sparsely present [ILLUSTRATION FOR FIGURE 9 OMITTED]. Weedy ruderal taxa continue to be present, probably on disturbed areas of the floodplain. Amaranthus, Verbena hastata, and V. urticifolia are the most consistently present [ILLUSTRATION FOR FIGURE 10 OMITTED]. Aquatic plant macrofossils almost completely disappear in zone 2 [ILLUSTRATION FOR FIGURE 11 OMITTED] Wetland species [ILLUSTRATION FOR FIGURE 12 OMITTED] are generally less abundant as well, but the two most abundant species, Pilea pumila and Urtica dioica, are shade tolerant. Glyceria striata, Eupatorium cf. maculatum, and E. perfoliatum are at their maximum in this zone.

In Zone 3 (5500 to [approximately equal to]3500 yr BP) the mesic deciduous forest elements disappear [ILLUSTRATION FOR FIGURE 6 OMITTED]. Salix and Acer are the only trees remaining, and they are represented sparsely at one level near the base of the zone. A few forest-floor herbs, including Aralia racemosa and Campanula americana, continue in the lower levels of this zone but disappear above [ILLUSTRATION FOR FIGURE 7 OMITTED]. Shrubs are represented by many of the same forest-edge taxa present in zone 2, including Corylus americana and Cornus spp. [ILLUSTRATION FOR FIGURE 8 OMITTED]. Although the number of specimens from zone 3 are low in all but two sites, the dominant taxa are prairie species [ILLUSTRATION FOR FIGURE 9 OMITTED]. These two sites date from 4830 and 2860 yr BP (the latter in the lower part of zone 4), where charcoal is extremely abundant. Among the important prairie taxa are Andropogon gerardii, Schizachyrium (Andropogon) scoparium, two Helianthus morphotypes, Hypoxis hirsuta, Dalea (Petalostemon, Petalostemum) candida, D. purpurea, Rudbeckia hirta, and Sorghastrum nutans [ILLUSTRATION FOR FIGURES 9, 13, and 14 OMITTED]. Disturbed-ground plants continue at about the same abundance through zone 3, with Oxalis becoming consistently present [ILLUSTRATION FOR FIGURE 10 OMITTED]. A remarkable shift occurs in the aquatic plants, which become abundant and diverse in this zone [ILLUSTRATION FOR FIGURE 11 OMITTED]. Most prominent are Myriophyllum pinnatum, Naias flexilis, and Potamogeton. Zizania aquatica first appears as well. A number of wetland plants also increase in zone 3 sediments and remain important to the present [ILLUSTRATION FOR FIGURE 12 OMITTED]. These include Polygonum lapathifolium, Leersia oryzoides, Scirpus validus type (including S. acutus), Sagittaria latifolia, and Helenium autumnale.

Zone 4 ([approximately equal to]3500 to [approximately equal to]380 yr BP) signals the return of a few macrofossils of trees, principally Salix, with some Tilia, and at single localities Ulmus, Populus, and Juglans cinerea [ILLUSTRATION FOR FIGURE 6 OMITTED] Forest-floor herbs disappear in this zone [ILLUSTRATION FOR FIGURE 7 OMITTED]. Most shrubs are also absent, but the floodplain species Cornus amomum ssp. obliqua becomes important, and two upland species that frequently occur on steep slopes, Physocarpus opulifolius and Staphylea trifolia, are present for the first time [ILLUSTRATION FOR FIGURE 8 OMITTED]. Additional prairie taxa include Amorpha canescens, Ratibida pinnata, Ceanothus americanus, Silphium laciniatum, and Monarda cf. punctata [ILLUSTRATION FOR FIGURES 9, 13, and 14 OMITTED]. Others like Rudbeckia hirta, Zizia aurea, and Hypoxis hirsuta continue to be well represented; however, several other species of zone 3 decline or are absent in zone 4. These latter include Dalea spp., Stipa spartea, Phlox pilosa, Schizachyrium, and Sorghastrum. Many of the same disturbed-ground and aquatic taxa found in zone 3 remain present through this zone [ILLUSTRATION FOR FIGURES 10 and 11 OMITTED]. Wetland taxa also continue to be generally abundant, with Sium suave and Carex haydenii appearing for the first time [ILLUSTRATION FOR FIGURE 12 OMITTED].

Zone 5 (Historic) shows a decline in abundance in most groups. Trees are even more scarce, with the exception of Salix; its bud scales are common, and capsules increase sharply [ILLUSTRATION FOR FIGURE 6 OMITTED]. Forest herbs remain poorly represented except for Ranunculus hispidus and Taenidia integerrima [ILLUSTRATION FOR FIGURE 7 OMITTED]. Shrubs continue to be represented by few taxa, and prairie taxa are almost entirely absent [ILLUSTRATION FOR FIGURES 8 and 9 OMITTED]. Weedy taxa, however, show a marked increase in diversity and numbers of specimens [ILLUSTRATION FOR FIGURE 10 OMITTED]. This increase in diversity is due partly to the appearance of many non-native species, including Pastinaca sativa, Polygonum convolvulus, P. hydropiper, Setaria glauca and S. viridis, and Taraxacum officinale, but native weedy annuals also increase markedly. The prominent peak in grass pollen [ILLUSTRATION FOR FIGURE 5 OMITTED] was probably caused primarily by Setaria. Few aquatic taxa remain [ILLUSTRATION FOR FIGURE 11 OMITTED], and only Potamogeton was observed growing in the stream at present. Wetland taxa fared somewhat better, especially such relatively aggressive species as Pilea pumila, Polygonum lapathifolium, Leersia oryzoides, Scirpus validus/acutus, Sagittaria latifolia, and Helenium autumnale.

Bryophyte macrofossils

Some of the bryophyte macrofossils correspond well with the pollen zonation, although several wetland and riparian taxa, including Hygroamblystegium tenax, Leptodictum riparium, Drepanocladus aduncus var. kneiffi, and D. aduncus var. polycarpus, show little zonation [ILLUSTRATION FOR FIGURE 15 OMITTED]. Bryophyte macrofossils in zone 1 ([greater than]12500 to [approximately equal to]9300 yr BP) include an assemblage of taxa characteristic of rich fen, riparian, wetland/marsh, upland, and disturbed habitats that might be associated with coniferous, mixed coniferous-deciduous, or deciduous forest. The occurrence of Calliergon giganteum and Scorpidium scorpioides indicates that rich fens were present in the area, but these species were only in samples dated 10 700 yr BP and were not found in the upper part of zone 1. Eurhynchium hians is an upland species that presently occurs in association with the deciduous forest in eastern North America (Crum and Anderson 1981). Species that would have been associated with streams are Hygroamblystegium noterophilum, which occurs on calcareous rock often in association with springs (Crum and Anderson 1981), and H. tenax and Leptodictyum riparium, either on soil or rocks (Crum and Anderson 1981). This is the first fossil record of H. noterophilum. Drepanocladus aduncus var. kneiffii occurs in marshy habitats (Crum and Anderson 1981), which might have been present in slow-moving or abandoned meanders of the creek, or in ponds created by beaver dams. The occurrence of Ceratodon purpureus indicates that some disturbed ground was present. In the upper part of this zone, additional deciduous forest elements, including Anomodon cf. attenuatus, A. minor, and Brachythecium acuminatum, appear. They occur on the base of deciduous trees or on rocks and less frequently on soil (Crum and Anderson 1981).

Zone 2 ([approximately equal to]9300 to 5500 yr BP) is defined by temperate deciduous-forest elements. In addition to Anomodon attenuatus, A. minor, and Eurhynchium, which occur relatively consistently, Brachythecium acuminatum, cf. B. reflexum, and Haplocladium cf. microphyllum are also present. Like B. acuminatum, B. reflexum characteristically occurs on bark at the base of trees, while Haplocladium frequently occurs on rotting wood (Crum and Anderson 1981). Two riparian elements, Hygroamblystegium tenax and Leptodictyum riparium, are consistent elements throughout this zone, while H. noterophilum disappears in the lower part. Single specimens of a Fissidens and Philonotis fontana are present in the lower and upper parts of this zone, respectively. The latter is characteristic of streams (Crum and Anderson 1981) where it typically occurs on soil. Wetland habitats are represented by Drepanocladus aduncus var. kneiffii, D. aduncus var. polycarpus, Hypnum cf. lindbergii, and Plagiomnium ellipticum. Both the Hypnum and the Plagiomnium are species that occur in swampy/marshy habitats (Crum and Anderson 1981), and might have been present along the stream or in abandoned meanders. Some remnants of what appear to be a species of Dicranella and a Bryum species, reflect the occurrence of locally disturbed habitats.

The lack of upland deciduous forest species, which characterize zone 2, and the absence of wetland/marsh taxa, define zone 3 (5500 to [approximately equal to]3500 yr BP) [ILLUSTRATION FOR FIGURE 15 OMITTED]. Also, the presence of additional taxa characteristic of disturbed soil in open habitats, such as might be found in association with animal burrows in prairies, is notable. These species include what is probably a species of Astomum or Weissia (the two genera are difficult to separate without sporophytes), Barbula cf. unguiculata, what appears to be Bryum caespiticium, and Dicranella. Riparian elements, including Hygroamblystegium tenax and Leptodictyum riparium persist through this zone, and Philonotis fontana var. caespitosa was present in the lower part.

The reappearance of wetland/marsh species, which were absent in zone 3, delimit zone 4 ([approximately equal to]3500 to [approximately equal to]380 yr BP) [ILLUSTRATION FOR FIGURE 15 OMITTED]. These wetland elements include Brachythecium cf. rutabulum, Drepanocladus cf. aduncus var. kneiffii, D. cf. aduncus var. polycarpus and Hypnum pratense. The two riparian species, Hygroamblystegium tenax and Leptodictyum riparium, were present throughout this zone, and a single specimen of Fissidens bryoides was found in the upper part. Fissidens bryoides typically occurs in shady habitats on calcareous rocks (Crum and Anderson 1981), often in creek beds. Species characteristic of disturbed soil occur sporadically throughout zone 4.

The only distinction between zones 4 and 5 (Historic) is the occurrence of Lindbergia brachyptera, an epiphyte that occurs on deciduous forest trees (Crum and Anderson 1981) [ILLUSTRATION FOR FIGURE 15 OMITTED]. Riparian elements included Fissidens obtusifolius var. apiculatus, a calciphile that occurs along streams and in rock crevices (Crum and Anderson 1981), Hygroamblystegium tenax and Leptodictyum riparium; wetland/marsh habitats are represented by Drepanocladus cf. aduncus var. polycarpus and Plagiomnium cf. ellipticum; and species of disturbed soil included what appear to be a Bryum and a Dicranella.

Insect analyses

The Roberts Creek fossil insect assemblages are among the richest yet described from Quaternary deposits. For Coleoptera (beetles) alone at least 27 families and 160 lower taxa were identified. In addition to beetles, the remains of molluscs, ostracods, spiders, orabatid mites, and other orders of insects are present.

Most of the beetle species represented are associated with fluvial, water-marginal, or marshland environments. A much smaller although distinctive component of the fauna is associated with upland environments. As with the plant macrofossils, the pollen zonation is used for a comparative chronologic basis to describe the analyses of associated insects.

Zone 1 ([greater than]12 500 to [approximately equal to]9300 yr BP). - Site CCLG (12 510 yr BP) yielded a sparse assemblage of beetles of northern affinities. Included within the ground beetles (Carabidae) are Bembidion morulum LeC., a species that is a regular inhabitant of the forest - tundra (Lindroth 1963), and a thorax provisionally identified to the arctic-subarctic pterostichine subgenus Cryobius. Likewise present are two arctic-subarctic beetle species recorded from the full-glacial of eastern and central Iowa (Baker et al. 1986, Schwert 1992): the hydrophilid Helophorus arcticus Brown and the weevil Vitavitus thulius Kiss. In site AB-34 (Table 2), several spruce-associated bark beetles (Scolytidae) are represented, including the northwestern species Carphoborus andersoni Sw. Assemblages of 10 700 yr BP and younger are dominated by aquatic beetles, most strikingly by elmids. The ground beetle assemblage is largely represented by water-marginal or fen-associated species; an exception is Dicaelus sculptilis Say (site AB-22, 11 380 yr BP), an open-woodland species today ranging northward only into the southernmost boreal forest.

Zones 2, 3, and 4 ([approximately equal to]9300 to [approximately equal to]380 yr BP). - Unlike the pollen and the vascular-plant and bryophyte macrofossil records, the insect assemblages collectively associated with these zones reflect little change in the nature of local environments through the Holocene. Dryopoid beetles, particularly elmids, are the numerically largest faunal component, representing approximately two-thirds of all beetle remains; at least five species of elmids (Dubiraphia cf. vittata [Melsh.], Macronychus glabratus Say, Microcylloepus pusillus [LeC.], Optioservus fastiditus [LeC.], and Stenelmis cf. crenata [Say]) and three species of dryopids (Helichus striatus LeC., H. striatus, and H. lithophilus) are regularly present. As in Zone 1, water-marginal and marsh-associated taxa are abundant, but without exception this element consists of taxa known today from eastern Iowa. Upland ground beetles include such inhabitants of open grasslands as Cyclotrachelus sodalis subsp. sodalis LeC., C. seximpressus LeC., Poecilus lucublandus (Say), and Amara angustata (Say) (Lindroth, 1966, 1968, Freitag 1969), as well as woodland species such as Scaphinotus elevatus (Fab.) and Pterostichus stygicus (Say) (Lindroth 1961, 1966). Additional upland taxa include the ladybird (Coccinellidae) Hyperaspis brunnescens Dobz., an apparently rare species known today only from Iowa and Illinois (Gordon 1985) [ILLUSTRATION FOR FIGURE 16 OMITTED], and numerous remains of a scarabaeid provisionally identified as Macrodactylus subspinosus (Fab.), the common "rose chafer." A limited fauna of saprophagous beetle species is present, including the scarabaeid Onthophagus hecate, a species associated with dung of large herbivores such as Bison.

Zone 5 (historic). - Postsettlement transformation of the landscape associated with EuroAmerican settlement induced profound changes in the nature of the insect fauna (Baker et al. 1993, Schwert, in press). Whereas the remains of dryopoid water beetles dominated presettlement assemblages, they are nearly absent in postsettlement samples; the most abundant water beetle taxon is the haliplid Peltodytes edentulus LeC., a regular inhabitant of eutrophic and polluted waters. Saprophagous scarabaeids, including the European species Aphodius distinctus (Muller) and A. granarius (L.), dominate the upland element of the assemblage, and all represent species today regularly associated with cattle dung. Weevils (Curculionidae) are likewise common, with at least six immigrant species, all associated with immigrant host plants. Other pest species include the grain-feeding cryptophagid Atomaria ephippiata Zimm. and the crucifer-feeding chrysomelid Phyllotreta striolata (Fab.).

Speleothem record of Coldwater Cave

Dorale et al. (1992) provided a record of Holocene climate change in northeast Iowa by analyzing 18O and 13C values in stalagmite 1s from Coldwater Cave [ILLUSTRATION FOR FIGURE 17 OMITTED], which is located 60 km northwest of the Roberts Creek site. This cave is ideal for isotopic analysis because it is over 30 m below ground surface, so the temperature in the cave shows little diurnal or seasonal change and reflects the integrated mean surface temperature over several years. The C-O isotopic evidence from speleothems is an independent record of both plant cover and climate; these data test the hypothesis that the changes in the biota of Roberts Creek were widespread and not local floodplain anomalies.

The 13C composition of speleothem-forming waters is controlled by the composition of soil carbon dioxide and of dissolving carbonate rocks. In temperate climates, soil carbon dioxide is produced by the decomposition of organic matter and by plant root respiration, and the 13C composition of soil carbon dioxide is determined by the type of vegetation on the surface (Cerling 1984). Plants using the Calvin or [C.sub.3] photosynthetic pathway, including nearly all trees and many grasses and forbs found in cooler, moister climates, have 13C values averaging [approximately equal to]-26 per mil. In contrast, plants using the Hatch-Slack or [C.sub.4] pathway, such as prairie grasses from warmer, drier climates, have 13C values averaging [approximately equal to]-13 per mil (Deines 1980). The 13C composition of calcite in a cave is generally 10-12 per mil greater than that of soil C[O.sub.2] (Rightmire and Hanshaw 1973) and is unlikely to have changed during the Holocene. Thus, changes in carbon isotope values of speleothems, though offset 10-12 per mil from surface values, are nonetheless explained by changes in relative abundance of forest (dominated by [C.sub.3] plants) and prairie (dominated by [C.sub.4] vegetation) above the cave.

The 18O values of speleothem calcite are largely a function of 18O values of the speleothem-forming water and the temperature of deposition (Schwarcz 1986). The O-isotopic composition of meteoric and thus cave waters (Harmon 1979) varies systematically with average atmospheric temperature (e.g., Dansgaard 1964). Similarly, fractionation of oxygen isotopes between cave waters and calcite varies with temperature. By combining an empirical relationship between 18O in meteoric water and temperature (Dansgaard 1964) with the calcite-water fractionation equation (Friedman and O'Neill 1977), Dorale et al. (1992) calculated atmospheric temperatures over Coldwater Cave.

The core from the stalagmite was dated by mass-spectrometry 230[Th.sup.-234][U.sup.-238]U data to have grown from 7779 to 1150 YBP. These dates are considered to be real years (reported here as YBP), as compared with the radiocarbon chronology at Roberts Creek (yr BP). The stage boundaries on the speleothem record have been converted to radiocarbon yr BP on Fig. 17 for ease of comparison with the Roberts Creek record.

Dorale et al. (1992) divided the Coldwater Cave isotopic record into three stages that are slightly modified in this paper [ILLUSTRATION FOR FIGURE 17 OMITTED]. Stage 1 ranges from 7779 YBP at the base to 5900 YBP with 18O values mostly [less than]25 per mil and 13C values [less than]-8, and an average temperature of 8.6 [degrees] C. This stage approximately corresponds to the upper half of pollen zone 2 ([approximately equal to]9300 to 5500 yr BP; [ILLUSTRATION FOR FIGURES 4 AND 5 OMITTED]). Stage 2 from 5900 to 3300 YBP is marked by an abrupt rise in 18O from [approximately equal to]24.5 to mostly over 25.5 per mil, suggesting an increase in mean annual temperature from 8.6 [degrees] to 10.8 [degrees] C. The 13C values rise rapidly (and slightly earlier than the 18O) from about -8 to about -7 and continue to rise gradually throughout stage 2 to about -5 per mil. This stage corresponds roughly to pollen zone 3 (5500-[approximately equal to]3500 yr BP). Stage 3 occurs from 3300 YBP to present and is marked by a gradual decrease in 13C to [approximately equal to]-7.5 per mil and an abrupt decrease in 18O to [approximately equal to]24.5, suggesting a decrease in mean annual temperature to 7.2 [degrees] C. This correlates fairly well with pollen zone 4 ([approximately equal to]3500 to [approximately equal to]380 yr BP).

The calculations of Dorale et al. (1992) assumed that there were no evaporative effects causing variations in the 18O values of the Coldwater Cave drip waters. Increased evaporation of surface or vadose waters, or dryness in the cave would shift drip waters to heavier 18O because of preferential removal of 16O. Replicate analysis within individual layers of the Coldwater Cave stalagmite indicate that isotopic equilibrium was maintained during calcite precipitation and that no evaporation took place in the cave. However, higher resolution sampling of Coldwater Cave stalagmite 1s [ILLUSTRATION FOR FIGURE 17 OMITTED] and unpublished C-O data (L. A. Gonzalez, M. K. Reagan, and R. G. Baker) on a second stalagmite from Coldwater Cave indicate that water infiltrating to stalagmite 1s was undergoing evaporation at the surface. Increased evaporation in surface and vadose waters was likely during stage 2 because of increased summer temperature, decreased summer precipitation, and decreased shade due to the demise of deciduous forest. Furthermore, stalagmite 1s was collected from a portion of Coldwater Cave that lies under a south-facing slope where evaporative effects would be more pronounced in the absence of a canopy cover. Thus, the temperature estimates of Dorale et al. (1992) should be considered maximum values. The estimates based on the unpublished C-O data (L. A. Gonzalez and M. K. Reagan) suggest a change of no more than 1.5 [degrees] C during stage 2.

The age of the stage 2-stage 3 boundary is now placed at 3300 YBP because the 13C record of stalagmite 1s remains stable (with one incursion) from [approximately equal to]4000 to 3300 YBP when the gradual decrease in 13C is initiated. From [approximately equal to]4600 YBP to the end of stage 2 the isotopic record shows large negative incursions in both 13C and 18O values. These negative incursions are attributed to periods of increased precipitation and possibly decreased temperature. Though the 18O record of stalagmite 1 s decreases to values that are slightly lower than those of stage 1, the 13C values are over 1 per mil higher than those of stage 1. The higher 13C values can be attributed to the fact that, although [C.sub.3] vegetation increased during an incursion, the [C.sub.4] vegetation remained more abundant than during stage 3.

DISCUSSION

Late-Glacial and earliest Holocene (Pollen Zone 1, [greater than]12 500 to [approximately equal to]9300 yr BP). - Conditions during late-glacial time are sparsely represented in the Roberts Creek samples and are not present in the speleothem record (Table 5, [ILLUSTRATION FOR FIGURE 17 OMITTED]). However, pollen, plant macrofossils, and insect remains document the occurrence of boreal spruce-larch forest superseded by conifer-hardwood forest. Initially the forest was dominated by Picea glauca, P. mariana, Larix laricina, and Fraxinus nigra. From the sedimentary record it appears that the Roberts Creek floodplain and the entire landscape were apparently somewhat unstable in the early phases of the late-glacial, when upland loess deposits were re-worked and redeposited on the floodplain. This instability may have resulted in a somewhat open landscape as suggested by the relatively high Poaceae and Cyperaceae pollen percentages, but it probably became closed before [approximately equal to]11 000 yr BP.

As is commonly the case with late Quaternary spruce-dominated plant assemblages, pollen and vascular plant macrofossils provide little evidence about other habitats. A few forest-floor herbs, including Ranunculus abortivus and Ranunculus hispidus, and forest-edge shrubs like Rubus appeared in zone 1 and [TABULAR DATA FOR TABLE 5 OMITTED] occurred sporadically thereafter. Ruderal plants also were present, indicating that typical floodplain processes including erosion, deposition, and beaver activity, were creating disturbances. Only in the aquatic and wetland environments were species found that were limited to this zone [ILLUSTRATION FOR FIGURES 11 AND 12 OMITTED]. These plants include Potentilla palustris and Hippurus vulgaris, which were at the southern edge of their present range, and Sagittaria engelmanniana, which was replaced locally by S. latifolia.

Late in zone 1, Quercus and especially Ulmus began to increase in the pollen record, and Ulmus in the macrofossil record, as Picea declined [ILLUSTRATION FOR FIGURES 4 AND 6 OMITTED]. One of these early Holocene sites with mixed boreal and deciduous plants (AB-51, Table 2) is immediately downstream from the "Postville fen," which presently harbors a few "northern" elements. This fen had both Picea and Larix macrofossils in basal sediments dating 11 580 yr BP (Table 2), though neither are present now, and a Picea mariana cone was found in sediments with abundant Ulmus americana bud scales. Apparently, spruce began to decline in the uplands between [approximately equal to]10 700 and 10 150 yr BP, but it and other boreal elements remained locally into the early Holocene in areas of cool, moist microclimate.

The bryophyte macrofossil record supports and augments the analyses of pollen and vascular plant macrofossils, although only the middle and upper part of zone 1 are represented by bryophyte macrofossils [ILLUSTRATION FOR FIGURE 15 OMITTED]. In the middle of zone 1, the occurrence of two rich fen species, Calliergon giganteum and Scorpidium scorpioides [ILLUSTRATION FOR FIGURE 15 OMITTED], suggests that the wetland habitats included rich fens typical of northern peatlands. Both species are restricted to conifer and conifer-hardwood forests in North America. It is likely that Potentilla palustris was an associate in these fens. Two other wetland taxa, Drepanocladus aduncus var. kneiffii and Plagiomnium ellipticum [ILLUSTRATION FOR FIGURE 15 OMITTED], might have occurred in the eutrophic margin of the fens or they might have been in the marshes documented by Sagittaria. The occurrence of several riparian species that today are associated with deciduous and conifer-hardwood forest (see Crum and Anderson 1981 for distributional data) suggests that conifer-hardwood forest may have dominated the region through a significant portion of zone 1. These species included Hygroamblystegium noterophilum, H. tenax, and Leptodictyum riparium [ILLUSTRATION FOR FIGURE 15 OMITTED]. Hygroamblystegium noterophilum may well have been associated with springs along the creek. Late in zone 1, the appearance of such deciduous forest elements as Anomodon minor, Brachythecium acuminatum, and Eurhynchium hians [ILLUSTRATION FOR FIGURE 15 OMITTED] correlates with the increase in Quercus and Ulmus pollen and Ulmus macrofossils.

Insects of arctic/subarctic affinities occur only at the oldest cutbank site (dated at 12 510 yr BP) in this first zone. Despite the excellent representation of spruce-associated beetles at 10 700 yr BP, the remainder is dominated by species whose ranges today lie largely south of the boreal forest. Such mixed faunas are characteristic of late-glacial insect assemblages in the mid-continent (Ashworth et al. 1981, Schwert 1992) as spruce forest gave way to deciduous forests and grasslands under climatic amelioration. The elmid fauna indicates that the water quality of the stream was high from the latter part of late-glacial time to the time of EuroAmerican settlement.

Beavers (Castor canadensis) were present in the area during the late-glacial. Beaver-gnawed wood was abundant at site AB-34 (10 700 yr BP) in what may have been part of a beaver dam. Surprisingly, all the beaver-gnawed wood examined was Picea. Modern beavers rarely take spruce, preferring both Populus and Salix (Ives 1942), and both these trees were locally present 10 700 yr BP.

No speleothem record is available for this interval.

Early-to-middle Holocene (Pollen Zone 2, [approximately equal to]9300 to 5500 yr BP). - Pollen and plant macrofossil assemblages document a gradual transition from conifer-hardwood forest to mesic deciduous forest. An apparent anomaly is the insect assemblages, which also record grassland (prairie?) habitats. The early-to-middle Holocene is a period marked by a stable climate and relatively stable floodplains, with few large floods in nearby southwest Wisconsin (Knox 1993) and probably at Roberts Creek as well. This stability must have been at least partially in response to a stable, densely forested landscape.

The earliest Holocene (lower part of zone 2) forest was dominated by Ulmus americana, with Tilia americana, Quercus, Betula, Fraxinus nigra, and Abies balsamea, but still included some Picea and Larix [ILLUSTRATION FOR FIGURES 4 AND 6 OMITTED]. Ostrya virginiana/Carpinus caroliniana became increasingly important in the subcanopy, and Taxus canadensis appeared in the shrub layer. This mixed assemblage suggests a modern analog in the conifer-hardwood forest (Curtis 1959).

Ulmus became somewhat less dominant in the upper half of this zone, Picea, Betula, and Abies dropped out, and Tilia americana, Acer saccharum, and Ostrya virginiana/Carpinus caroliniana increased in importance near the top, along with spores of Pteridium, Adiantum type, and Dryopteris type [ILLUSTRATION FOR FIGURES 4-6 OMITTED]. Also, Juglans cinerea appeared in the upper part of this zone. The character of the vegetation thus changed from conifer-hardwood to mesic deciduous forest as the boreal elements disappeared. Forest understory plants [ILLUSTRATION FOR FIGURES 7 AND 8 OMITTED] also are predominantly those of mesic deciduous forest (Curtis 1959), including the shrub Zanthoxylum americanum and the herbs Aralia racemosa, Campanula americana, Claytonia cf. virginica, Cryptotaenia canadensis, and Laportea canadensis. The peak of this episode of mesic vegetation, when a maple-basswood-elm forest was dominant, was from [approximately equal to]6500 to 5500 yr BP.

During deposition of zone 2 sediments, the aquatic flora continued to be sparsely represented. The northern wetland species, Hippuris vulgaris and Potentilla palustris disappeared, and temperate wetland elements like Pilea pumila increased in abundance [ILLUSTRATION FOR FIGURES 11 AND 12 OMITTED]. The presence of some ruderal taxa indicates that even during this stable period some disturbance occurred [ILLUSTRATION FOR FIGURE 10 OMITTED]. Most likely these plants grew on point bars, levees, and caving banks on the outside of meanders where bare ground would be periodically exposed [ILLUSTRATION FOR FIGURE 3 OMITTED].

The bryophyte macrofossils similarly indicate that the early-to-middle Holocene forest was dominated by hardwood elements. Characteristic species of conifer-hardwood and deciduous forests include Anomodon attenuatus and A. minor, Eurhynchium hians, Brachythecium acuminatum, cf. B. reflexum, and Haplocladium cf. microphyllum. The "weedier" riparian elements Hygroamblystegium tenax and Leptodictyum riparium persisted throughout this zone, although H. noterophilum, which probably has a more restricted niche, disappeared relatively early [ILLUSTRATION FOR FIGURE 15 OMITTED]. As with the vascular plants, most wetland elements, including species of Drepanocladus, Hypnum, and Plagiomnium ellipticum, disappeared early in the Holocene [ILLUSTRATION FOR FIGURE 15 OMITTED].

Two discrepancies between the pollen and plant macrofossil assemblages and the insect assemblages for the early-to-middle Holocene may be more apparent than real. (1) Although the insects from this and subsequent zones included woodland taxa, species characteristic of open grasslands, including saprophagous beetle species that might have been associated with Bison, were also present. These habitats may be where the few early-to-middle Holocene prairie species recorded (Rudbeckia hirta, Zizia aurea, and Monarda fistulosa) were growing. Perhaps prairie openings were present on uplands or in dry microhabitats in small, local openings in the forest. Alternatively, these open habitats may have occurred on point bars, eroded banks, levees, or other disturbed areas along the stream [ILLUSTRATION FOR FIGURE 3 OMITTED]. (2) Aquatic plants are uncommon in the early-to-middle Holocene, whereas the insect assemblages for this and subsequent zones are dominated by water-marginal and marsh-associated taxa. The greatest proportion of these were dryopoid beetles, particularly elmids, indicating that the waters of Roberts Creek were of trout-stream quality, with alternating riffles and pools bordered by muddy banks and zones of marsh-associated plants. The scarcity of aquatic vascular plants may have been caused by a dense forest cover across the floodplain, shading the stream itself; most of these aquatic plant species require full sunlight. Insects would probably have been less affected by shaded habitats.

The speleothem record for the early Holocene is missing prior to [approximately equal to]8000 yr BP, but during the period from [approximately equal to]8000 to 5900 yr BP conditions were remarkably stable; the climate was warm and moist, and the area was apparently forested, paralleling the Roberts Creek record. In contrast, pollen records from western and central Iowa as far east as Clear Lake [ILLUSTRATION FOR FIGURE 1 OMITTED], and from Minnesota as far south as Kirchner Marsh (Watts and Winter 1966, Van Zant 1979, Baker et al. 1992) indicate that this interval was warmer and much drier (in fact, 6500 to 5500 yr BP was the warmest and driest of the entire Holocene). Baker et al. (1992) and Wright (1992) suggested that dry Pacific air masses were blocked from eastern Iowa and southern Wisconsin by a strong monsoonal flow of maritime tropical air from the Gulf of Mexico. This air flow would have kept eastern Iowa warm and moist at the same time that western and northern areas were extremely dry. Wright (1992) noted that the early-Holocene peak in Ulmus characteristic of the Midwest is higher than any modern value, and the climate for this period may not have a modern analog because the Laurentide ice sheet was still present in Canada.

Middle-to-late Holocene (Zone 3, 5500 to [approximately equal to]3500 yr BP). - Deciduous forest abruptly gave way to prairie in the middle Holocene, indicating markedly less available moisture. In contrast, insect assemblages reflect little change in the environment from zone 2. Slight downcutting occurred on the floodplain, creating some disturbance along the stream.

In the middle-to-late Holocene the pollen diagrams [ILLUSTRATION FOR FIGURES 4 AND 5 OMITTED] show an abrupt disappearance of most of the mesic forest taxa and their replacement by Ambrosia and Poaceae at the boundary between zones 2 and 3. This change suggests a rapid decline of forest between [approximately equal to]5500 and 5400 yr BP, and the sudden decline in arboreal macrofossils supports this interpretation [ILLUSTRATION FOR FIGURE 6 OMITTED]. The forest was replaced in part by shrubs (Zanthoxylum americanum, Rhus typhina, Sambucus canadensis, and Corylus americana) that are characteristic of open forests and forest edges (Shimek 1915). Shady forest-floor herbs, which probably persisted in remnants of deciduous forest as it declined, overlap slightly into zone 3. This transitional interval prevailed for perhaps a few hundred years as the early-to-middle Holocene forest gave way to prairie in the middle Holocene.

Increased disturbance on the floodplain during downcutting may explain the abundance (Amaranthus and Verbena spp.) and first appearances (cf. Echinochloa, Oxalis, and Polygonum pensylvanicum) of ruderal plants [ILLUSTRATION FOR FIGURE 10 OMITTED] and the increases in Salix and Acer negundo pollen and macrofossils [ILLUSTRATION FOR FIGURES 4 AND 6 OMITTED] during the middle-to-late Holocene.

Aquatic taxa rose to maximum numbers at or near the base of the middle-to-late Holocene [ILLUSTRATION FOR FIGURE 11 OMITTED]. The floodplain wetland communities retained all of the elements present in the early-middle Holocene, but the dominance of Glyceria striata and the shade-tolerant Urtica dioica and Pilea pumila apparently suddenly switched to Scirpus validus/acutus, along with Polygonum lapathifolium, Leersia oryzoides, Polygonum punctatum, Sagittaria latifolia, Cornus spp., and Helenium autumnale [ILLUSTRATION FOR FIGURES 8 AND 12 OMITTED]. It is unclear what triggered this change, although recent studies indicate that light is the primary abiotic constraint on photosynthesis in most shaded streams (Hill et al. 1995). Thus the increase in light as forest cover decreased may have been the decisive factor. Changes in nutrient budgets or flooding frequency may also have contributed.

The question of whether there was truly prairie at the site, as suggested by the increase in Ambrosia and Poaceae pollen [ILLUSTRATION FOR FIGURE 5 OMITTED], is clearly answered by the vascular plant macrofossils. Most apparently arrived at or before 4800 yr BP, although some may have been delayed until [approximately equal to]2900 yr BP [ILLUSTRATION FOR FIGURE 9 OMITTED].

It is difficult to pinpoint exactly when most prairie species arrived because they are sparsely represented in the record except when fires occurred. Ordinarily, the thick herbaceous cover in a prairie acts as a filter, trapping seeds and fruits and inhibiting their removal during runoff events. However, when the stems burn off, runoff is unimpeded, and large numbers of seeds and fruits are carried into the channel, resulting in a rich assemblage in the sediment (Baker and Drake 1994). Evidence of fires in the middle and late Holocene includes many charred stems, fruits, and seeds, and even some "popped" seeds that are completely charred and have burst like popcorn [ILLUSTRATION FOR FIGURES 13 AND 14 OMITTED]. Two sites have this evidence of fires, AB-45A at 4830 yr BP and AB-21 at [approximately equal to]2900 yr BP (average of four dates, Table 2; [ILLUSTRATION FOR FIGURE 9 OMITTED]). The rich assemblage at AB-45A clearly demonstrates that many typical prairie species had reached the area by 4830 yr BP, and that the pollen signal was not merely a local disturbance dominated by ragweed and grasses. The presence of abundant macrofossils of Andropogon and Sorghastrum indicates that the grass pollen peak was made predominantly by prairie species.

Bryophyte macrofossils clearly differentiate the early-to-middle from the middle-to-late Holocene by the abrupt disappearance of upland temperate deciduous forest species [ILLUSTRATION FOR FIGURE 15 OMITTED]. While few bryophytes can be considered characteristic of prairie habitats, what are generally regarded as "weedy" taxa, including many of the so-called ephemeral mosses (species that complete their life cycle in a relatively short period of time, perhaps 2-3 mo), might be among the leading candidates. Although only one of the species recorded from the middle-to-late Holocene might represent an ephemeral moss (cf. Astomum/Weissia sp.), there is an increase in the diversity of "weedy" species, which include Barbula cf. unguiculata, undeterminable Bryum species, what might be B. cf. caespiticium, and what is probably a species of Dicranella.

The change in the speleothem record at 5900 YBP at the beginning of Stage 2 correlates with the vegetational changes between zones 2 and 3. At the beginning of Stage 2 (5900 YBP), 13C values increased first abruptly, then gradually throughout the stage. 18O values rose abruptly [approximately equal to]200 yr later at 5700 YBP. The oxygen isotope data provide evidence that the changes in the vegetation reflected an increase in mean annual temperature of no more than 1.5 [degrees] C.

The new high-resolution record from Coldwater Cave stalagmite 1s [ILLUSTRATION FOR FIGURE 17 OMITTED] reveals that the rapid initial increase in 13C values occurred as much as 200 yr earlier than the abrupt change in 18O values. This evidence indicates that [C.sub.4] plants increased and prairie began to replace forest 200 yr before the temperature increased. By 5700 YBP the loss of forest and the increase in temperature greatly intensified the evaporative effect on surface water.

It is likely that conversion of forest to prairie before the temperature increased was the result of one or more fires. This idea is supported by the Roberts Creek record at site F45, with a radiocarbon age of 4830 yr BP, which is rich in both charcoal and prairie species. The sharp inflection in 13C 5700 YBP is equivalent to 4900 yr BP in radiocarbon years. These two independent records from 60 km apart suggest that a major fire occurred at virtually the same time. The sudden disappearance of trees in the pollen record at Roberts Creek between 5500 and 5400 yr BP is not locally marked by charcoal-rich sediment, but fires may have been prevalent regionally for several hundred years and gradually forced the prairie-forest border eastward. On the other hand, the progressive increase in the average 13C of soil C[O.sub.2] throughout stage 3 reflects the gradually increasing proportion of organic matter from [C.sub.4] plants in the soil after the conversion from forest to prairie.

The alluvial record shows a lithologic transition between Roberts Creek and Gunder Members and suggests that conditions began to change gradually in the middle Holocene towards those represented in the late Holocene [ILLUSTRATION FOR FIGURE 6 OMITTED]. This is consistent with the gradual change in carbon isotopes from the speleothem record.

Although fires may have played a role, the isotopic and biotic data all indicate that the cause of the general expansion of prairie in the middle Holocene was climatic (Baker et al. 1990, 1992, Chumbley et al. 1990, Dorale et al. 1992). The clear implication is that prairie expansion in the Midwest was caused by increased flow of Pacific air (Borchert 1950, Bryson 1966, Bryson and Wendland 1967, Webb and Bryson 1972, Webb et al. 1983). However, this expansion was delayed by up to 3000 yr at Roberts Creek compared to sites to the north and west ([ILLUSTRATION FOR FIGURE 1 OMITTED]; Wright et al. 1963, Van Zant 1979, Kim 1986). It occurred suddenly in northeastern Iowa when the monsoonal pattern that blocked Pacific air flow finally broke down [approximately equal to]5500 yr BP in the central Midwest (Baker et al. 1992, Wright 1992). This drier air would increase fire susceptibility, and increased fire frequency could account for such a rapid change from forest to prairie.

Two lines of evidence seem inconsistent with this climatic interpretation. First, fens in eastern Iowa generally were active in the early Holocene, inactive in the middle Holocene, and wet and actively depositing peat again in the late Holocene (Thompson and Bettis, in press). For example, in the "Postville fen," which is within the Roberts Creek drainage basin, organic silt was deposited prior to [approximately equal to]11 500 yr BP (wet conditions) and peat was deposited from 5460 yr BP (wet conditions) to the present; it has a hiatus spanning the interval [approximately equal to]11 500 to 5460 yr BP (dry conditions?) The hydrologic change that this sequence suggests seems inconsistent with the evidence of early-to-middle Holocene mesic forest and its replacement by middle-to-late Holocene prairie (why would the water table be low when mesic forest was present, and higher when the vegetation changed to prairie?). It may be that fens with a regional groundwater source are less susceptible to drought. Our hypothesis is that peat was deposited during the early-to-middle Holocene forested interval, but that prairie fires swept the area shortly after 5500 yr BP and burned the earlier peat deposits, and then peat deposition resumed. Presumably no further burning could have occurred on the fen.

Second, the beetle fauna, which was very sensitive to climatic changes in full-glacial, late-glacial, and earliest-Holocene time in many areas (Schwert 1992, Elias 1994), shows no real change through the Holocene. It is unclear how widespread this phenomenon is. Holocene beetle faunas have received very little study. Elias (1994) lists only a few Holocene faunas, widely scattered in space and time, in his review of the Quaternary beetle faunas in North America.

Two hypotheses are proposed to explain the Holocene record at Roberts Creek. Hypothesis 1 is that the record reflects mainly local conditions along the floodplain, not regional climate, and that upland conditions were like those to the west and north: warm and dry in the latter part of the early-to-middle Holocene and cool and wet in the middle-to-late Holocene. This hypothesis could explain the apparent desiccation of fens in the early-to-middle Holocene and their middle-to-late Holocene renewal. If prairie arrived in the early Holocene in upland areas, as it did farther north and west, then the beetle fauna could be explained.

Hypothesis 2 postulates that the plant records at Roberts Creek accurately reflect regional climatic patterns in the Holocene, but that these changes are mainly seasonal shifts rather than changes in mean annual precipitation and temperature.

The bulk of the evidence favors this hypothesis [ILLUSTRATION FOR FIGURE 6 OMITTED], including: (a) the timing and direction of the changes in the macrofossils at Roberts Creek; (b) the pollen record, which presents a more regional picture and shows no evidence of early-to-middle Holocene upland prairies; (c) the pattern of climatic and vegetational change, which correlates well with regional patterns farther east (Baker et al. 1992); (d) the floodplain water table could not have fallen because fossil preservation required that it remain high; and (e) the climatic and vegetational signals from Roberts Creek generally agree well with those in the carbon and oxygen-isotope record from speleothems in nearby Coldwater Cave. It is not surprising that changes in environment along the prairie-forest border are complex; a recent model that predicts present vegetation and water balance shows the least agreement in the Prairie Peninsula (Neilson 1995).

A shift in the seasonality of precipitation and temperature can explain the paleoenvironmental changes without involving large mean annual shifts. Knox (1993) concluded that similarly small changes in mean annual temperature and precipitation caused large changes in flood frequency and magnitude in southwestern Wisconsin. Many of the environmental changes at Roberts Creek are consistent with the following seasonal shifts. Much of the early-to-middle Holocene precipitation fell in summer and early fall, and such abundant moisture during the growing season supported forests. This rainfall may have come during frequent small storms, resulting in a few large floods. When this pattern shifted in the middle Holocene to fewer but more intense rains in spring and early summer, with droughts through midsummer and early fall, moisture was lacking during the growing season; this additional moisture stress resulted in replacement of forests by prairie in this region. Heavy spring rains would cause increased flooding because runoff is greater (1) when the ground is frozen, (2) when plants are dormant or not fully active, and (3) in prairie than in forest. The lack of summer rain resulted in more frequent fires, which may have been responsible for the sudden demise of the forests [approximately equal to]5500 yr BP. About 99% of prairie fires started by lightning presently occur during the summer growing season (Higgins 1984, cited in Howe 1994), and macrofossils from burned sites on Roberts Creek represent some plants that fruit in early summer and others in late summer, suggesting that these burns were also in the summer. Fires could have been started by Native Americans and not be related to climate, but despite numerous exposures along Roberts Creek, only one archaeological site has been found in the area, and it is much younger than the transition (Mather 1992).

Late Holocene (Zone 4, [approximately equal to]3500 to [approximately equal to]380 yr BP). - From the pollen and vascular plant macrofossils, it appears that savanna and groves of deciduous forest replaced the prairie in the late Holocene. This more subtle change in the dominant vegetation is not reflected in the bryophyte macrofossil or insect assemblages. However, the speleothem data do record a change during this time.

The return to higher Quercus pollen percentages in the late Holocene, while Ambrosia, Poaceae, and Asteraceae remain relatively high, indicates that savanna was present during the last 3000 yr [ILLUSTRATION FOR FIGURES 4 AND 5 OMITTED]. Tilia pollen and macrofossils also reappear at very low levels, suggesting that basswood was sparingly present. Salix was the only other abundant tree, but it is limited to the floodplain and does not occur in the upland forest. Forest-floor herbs are limited to Laportea at a single site and Ranunculus hispidus, which grows with Salix in floodplain habitats [ILLUSTRATION FOR FIGURE 7 OMITTED].

Shrubs of steep slopes, possibly with rock outcrops, also are represented in the late Holocene by Physocarpus opulifolius and Staphylea trifolia [ILLUSTRATION FOR FIGURE 8 OMITTED]. These relatively sturdy disseminules probably reached the floodplain where the stream abuts the rocky valley wall. Prairie species continued to be present. They are abundant and diverse in the 2900-yr-old deposit with evidence of fire, and many continued sporadically in younger deposits. Amorpha canescens macrofossils first appear 2900 yr BP, although the pollen data suggest that it may have been present earlier; it is one of the most commonly found macrofossils of prairie habitats (Watts and Winter 1966, Van Zant 1979, R. G. Baker, unpublished data). Disturbed-ground plants became slightly less common, and aquatic communities were still relatively diverse, but their macrofossils are less common and more sporadic than in the previous zone.

The late Holocene change from prairie to savanna is accompanied by a reappearance of wetland bryophytes, including Brachythecium cf. rutabulum, species of Drepanocladus, and Hypnum pratense [ILLUSTRATION FOR FIGURE 15 OMITTED]. This suggests a return to more mesic conditions, which would support the idea that pockets of deciduous forest were once again present on the landscape.

Diverse dryopoid beetle faunas characterize the late Holocene (Baker et al. 1993, Schwert, in press). These faunas exist today only in association with the bedrock-or gravel-bedded riffle zones of streams of exceptionally high water quality (Brown 1976, Hilsenoff 1977). This fauna indicates that Roberts Creek continued to be a clear, cold-water stream in a stable landscape.

All of the beetle taxa represented within sediments of Holocene age at Roberts Creek today occur in eastern Iowa (for example, [ILLUSTRATION FOR FIGURE 16 OMITTED]). This uniformity of the fauna throughout the Holocene is striking compared to the changes inferred in regional vegetation. The absence of change in the Holocene beetle fauna may be attributed to several factors including: (1) changes in seasonality of precipitation, as mentioned above, without major temperature change; this may have influenced plants more than insects; (2) the assemblages dominantly reflect local communities inhabiting the floodplain zones of Roberts Creek and its tributaries, not upland communities that would have been more impacted by climatic and environmental change; and (3) the Roberts Creek basin lies distant from major boundaries for the modern Coleoptera fauna. Although these boundaries may have shifted in response to Holocene climatic and environmental changes, they evidently remained distant from northeastern Iowa.

The oxygen isotopes in the speleothem record show an abrupt decrease in mean annual temperature at [approximately equal to]3300 yr BP, when savanna vegetation partially replaced prairie. The 13C values return gradually toward early-to-middle Holocene values, as [C.sub.3] plants returned to prominence.

The alluvial-stratigraphic record of the Roberts Creek Member indicates that slight downcutting occurred sometime [approximately equal to]4000 yr BP (Table 5). Knox (1993) documented more frequent flooding after 3500 yr BP in southwestern Wisconsin. The record at Roberts Creek is less clear, but the late-Holocene channel belt may have deepened and widened slightly in response to a combination of the climatic shift reported by Knox (incursion of westerly air masses favoring intense cyclonic storms; Knox 1983) and the change in vegetation cover to prairie, promoting increased runoff.

Many modern surface soils on moderately sloping parts of the northeastern Iowa landscape are Alfisols (forest soils) with relatively thick, dark-colored surface horizons (Kuehl 1982). The morphology of these so-called forest-prairie transition soils, or Mollic Hapludalfs, suggests that prairie-derived organic matter was later imprinted on forest-dominated soils. This is consistent with the late arrival of prairie, as suggested by the pollen and plant macrofossil record, and the subsequent presence of savanna during the last 3000 yr. The distribution of modern Mollisols (prairie soils) further suggests that broad, gently sloping upland areas have been islands of prairie bordered by savanna in the deciduous forest along the prairie-forest border in northeast Iowa.

Historic period (Zone 5). - This interval is characterized by the remarkable replacement of the previous vegetation by weedy annual plants [ILLUSTRATION FOR FIGURE 10 OMITTED] that came in after the land was cultivated (Baker et al. 1993). Salix was the only arboreal component on the floodplain, and it continued to be the only tree present [ILLUSTRATION FOR FIGURE 6 OMITTED]. Upland forests dominated by Quercus were probably cut locally to clear the land for crops and to provide wood, making additional habitat available for weedy species. The Cornus shrub community on the floodplain was apparently eliminated [ILLUSTRATION FOR FIGURE 8 OMITTED], and among prairie species only such "weedy" elements as Pycnanthemum and Monarda that will grow in pastures survived after cultivation began [ILLUSTRATION FOR FIGURE 9 OMITTED]. All aquatic taxa but Chara are missing as well [ILLUSTRATION FOR FIGURE 11 OMITTED], although Potamogeton grows in a few places in the creek at present. Wetland species reacted variously, with some of the weedier species like Bidens cernua, Helenium autumnale, Impatiens, Leersia oryzoides, Pilea pumila, Polygonum lapathifolium, and Sagittaria latifolia increasing to great abundance [ILLUSTRATION FOR FIGURE 12 OMITTED]. Others, including Scirpus validus/acutus and Zizia aurea, decreased when cultivation began.

The bulk of the change is seen in the increase of disturbed-ground species [ILLUSTRATION FOR FIGURE 10 OMITTED]. Many of these taxa produce copious quantities of seeds and are adapted to colonize bare ground. The areal extent of disturbed ground available for these plants in the early Historic period was extreme compared to that of previous periods. Disturbance was direct in the form of cultivation, land clearing, and grazing, as well as indirect channel widening, rapid floodplain aggradation, and increased channel migration in response to increases in runoff, flood frequency and magnitude, and sediment load (Knox 1987, Baker et al. 1993).

The Historic channel belt widened rapidly as the stream adjusted to increased runoff and peak flood height during the early Historic period. During the early part of the channel adjustment, before the channel had widened appreciably, rapid floodplain aggradation and overlapping of older floodplain surfaces by Camp Creek Member alluvium occurred because sediment-laden runoff exceeded the channel's conveyance capacity (Baker et al. 1993). After [approximately equal to]1930, the channel had widened to the point where it contained most flood-water, and floodplain aggradation decreased markedly. Knox (1977, 1987) documented similar channel responses and floodplain sedimentation related to Historic land use changes in the Driftless Area of Wisconsin, and Brown and Barber (1985) and Tipping (1995) have done similar studies in Britain.

The bryophytes of the Historic period show no real change from those of the late Holocene.

Only upon destabilization of the landscape as a consequence of EuroAmerican settlement were profound changes induced in the nature of Holocene insect faunas (Baker et al. 1993, Schwert, in press). Associated with the rapid transformation of drainage basin into pasture lands and monoculture croplands was a marked decline in upland ground beetle species within the assemblages, presumably reflecting a similar decline in both regional populations and diversity. Dung- and crop-associated species instead dominate the upland assemblages. A variety of introduced species appeared, including beetles that inhabit cattle dung, and others that feed on grain and other introduced garden vegetables. The dryopoid water beetles that reflected Holocene-long high water quality nearly disappeared. Sedimentation and eutrophication associated with agricultural transformation of the landscape effectively eliminated all dryopoid habitat with the creek itself. This fauna was replaced by Peltodytes edentulus and other indicators of eutrophic conditions.

CONCLUSIONS

1) Lithologically distinct alluvial fills representing late-glacial, early-to-middle Holocene, late Holocene, and Historic times occur in cutbanks along Roberts Creek.

2) Organic remains in the upper segments of these streams contain an abundant, well-preserved, and diverse biota, including pollen, vascular plant macrofossils, bryophytes, and insects. These fossils and sediments give a detailed history of the interaction of climate, biota, and stream activity.

3) Plant fossils (pollen, vascular plant macrofossils, bryophytes) show a well-defined sequence of vegetational change from boreal forest to conifer hardwood forest to deciduous forest to prairie to savanna, and finally to cultivated land. Insect fossils reveal the same strong change from boreal to temperate conditions, but record little change during the Holocene until EuroAmerican settlement.

4) The timing of these changes differs from sites to the west and north. The change from deciduous forest to prairie especially lags behind these areas by over 2500 yr BP.

5) Carbon and oxygen isotope records from nearby Coldwater Cave contain the same time lag and support the timing and direction of environmental change. Oxygen isotopes change abruptly, suggesting that the invasion of forest by prairie was partially caused by a sudden climatic shift. Carbon isotopes reflect the proportion of [C.sub.3] to [C.sub.4] plants growing in the area and show a slow change in soil carbon after the [C.sub.4] prairie plants became dominant.

6) Changes in seasonal aspects of precipitation and temperature best explain the environmental changes in northeast Iowa.

ACKNOWLEDGMENTS

This research was supported by NSF Grants ATM 88-06482 and EAR 93-16391 to R. G. Baker, a seed grant from the Center for Global and Regional Environmental Research, University of Iowa, to R. G. Baker and D. G. Horton, NSF Grant ATM-88-0571 to D. P. Schwert, NOAA Grant NA366P0238 to L. A. Gonzalez, Iowa Science Foundation Grant to C. A. Chumbley, and an Iowa Division of Historic Preservation grant to E. A. Bettis. We thank Dr. Dennis Anderson, Humboldt State University Professor Emeritus, for help in identifying Poaceae macrofossils, James Baker, who took the SEM photomicrographs, and Dr. Curt Klug, who prepared them for publication. For their assistance in identification of beetle remains, we are grateful to the staff of the Coleoptera Division at Agriculture Canada, H. P. Brown of the University of Oklahoma, R. S. Anderson of the National Museum of Canada, and R. D. Gordon of the Systematic Entomology Laboratory, USDA. Comments by H. E. Wright, Jr., Robyn Burnham, and an unidentified reviewer improved the manuscript. For their help in both field and laboratory analyses, we thank Wesley Peck, Joseph Krieg, Kevin Krogstad, Sally Domke, Cynthia Jensen, Julieann Van Nest, and Hong Zhu.

LITERATURE CITED

Almendinger, J. C. 1992. The late Holocene history of prairie, brush-prairie, and jack pine (Pinus banksiana) forest on outwash plains, north-central Minnesota, USA. Holocene 2:37-50.

Anderson, L. E., H. A. Crum, and W. R. Buck. 1990. List of the mosses of North America north of Mexico. Bryologist 93:448-499.

Ashworth, A. C. 1979. A method of recovering fossil insect remains from clays, peats, and silts. Page 406 in T. L. Erwin, G. E. Ball, and D. R. Whitehead, editors. Carabid beetles, their evolution, natural history, and classification. Junk, The Hague, The Netherlands.

Ashworth, A. C., D. P. Schwert, W. A. Watts, and H. E. Wright, Jr. 1981. Plant and insect fossils at Norwood in south-central Minnesota: a record of late-glacial succession. Quaternary Research 16:66-79.

Baker, R. G., C. A. Chumbley, P. M. Witinok, and H. K. Kim. 1990. Holocene vegetational changes in eastern Iowa. Journal of the Iowa Academy of Science 97:167-177.

Baker, R. G., and Drake, P. 1994. Holocene history of the prairie in Midwestern United States: pollen vs. plant macrofossils. Ecoscience 1:333-339.

Baker, R. G., L. J. Maher, C. A. Chumbley, and K. L. Van Zant. 1992. Patterns of Holocene environmental change in the Midwest. Quaternary Research 37:379-389.

Baker, R. G., R. S. Rhodes, II, D. P. Schwert, A. C. Ashworth, T. J. Frest, G. R. Hallberg, and J. A. Janssens. 1986. A full-glacial biota from southeastern Iowa, USA. Journal of Quaternary Science 1:91-107.

Baker, R. G., D. P. Schwert, and E. A. Bettis III. 1993. The impact of EuroAmerican settlement on landscapes, vegetation, and water quality in northeastern Iowa. Holocene 3:314-323.

Bartlein, P. J., T. Webb III, and E. Fleri. 1984. Holocene climatic change in the northern Midwest: pollen-derived estimates. Quaternary Research 22:361-374.

Bernabo, J. C., and T. Webb III. 1977. Changing patterns in the Holocene pollen record from northeastern North America: a mapped summary. Quaternary Research 8:64-96.

Bettis, E. A., III. 1984. Preliminary investigations of the stratigraphy and chronology of northeastern Iowa alluvium and its archaeological significance with special reference to the Turkey River Basin. Report to the Iowa State Historical Department, Division of Historic Preservation. On file at the Iowa Department of Natural Resources, Geological Survey Bureau, Iowa city, Iowa, USA.

-----, editor. 1990. Holocene alluvial stratigraphy and selected aspects of the Quaternary history of western Iowa. Guidebook for the 37th Field Conference of the Midwest Friends of the Pleistocene. Iowa Department of Natural Resources, Geological Survey Bureau Guidebook Series Number 9, Iowa City, Iowa, USA.

-----. 1992. Soil morphologic properties and weathering zone characteristics as age indicators in Holocene alluvium in the Upper Midwest. Pages 119-144 in V. T. Holliday, editor. Soils in archaeology. Smithsonian Institution Press, Washington, D.C., USA.

Bettis, E. A., III, R. G. Baker, W. R. Green, M. K. Whelan, and D. W. Benn. 1992. Late Wisconsinan and Holocene alluvial stratigraphy, paleoecology, and archaeological geology of east-central Iowa. Iowa Department of Natural Resources, Geological Survey Bureau Guidebook Series Number 12, Iowa City, Iowa, USA.

Bettis, E. A., III, and G. R. Hallberg. 1985. Quaternary alluvial stratigraphy and chronology of Roberts Creek Basin, northeastern Iowa. Pages 44-45 in R. S. Lively, coordinator. Pleistocene geology and evolution of the Upper Mississippi Valley. Minnesota Geological Survey, St. Paul, Minnesota, USA.

Borchert, J. R. 1950. The climate of the central North American grassland. Association of American Geographers Annals 40:1-29.

Bounk, M. J., and E. A. Bettis III. 1984. Karst development in northeastern Iowa. Proceedings of the Iowa Academy of Science 91:12-15.

Brown, A. G., and K. E. Barber. 1985. Late Holocene palaeoecology and sedimentary history of a small lowland catchment in central England. Quaternary Research 24:87-102.

Brown, H. P. 1976. Aquatic dryopoid beetles (Coleoptera) of the United States. U.S. Environmental Protection Agency Water Pollution Control Research Series 18050-ELD04/72.

Bryson, R. A. 1966. Air masses, streamlines, and the boreal forest. Geographical Bulletin 8:228-269.

Bryson, R. A., and W. M. Wendland. 1967. Tentative climatic patterns for some late glacial and post-glacial episodes in central North America. Pages 271-298 in W. M. Mayer-Oakes, editor. Life, land and water. University of Manitoba Press, Winnipeg, Manitoba, Canada.

Burnham, R. J., S. L. Wing, and G. G. Parker. 1992. The reflection of deciduous forest communities in leaf litter: implications for autochthonous litter assemblages from the fossil record. Paleobiology 18:30-49.

Cerling, T. E. 1984. The stable isotopic composition of modern soil carbonate and its relationship to climate. Earth and Planetary Science Letters 71:229-240.

Chumbley, C. A. 1989. Late-glacial and Holocene vegetation of the Roberts Creek Basin, northeast Iowa. Dissertation. University of Iowa, Iowa City, Iowa, USA.

Chumbley, C. A., R. G. Baker, and E. A. Bettis III. 1990. Midwestern Holocene Paleoenvironments revealed by floodplain deposits in northeastern Iowa. Science 249:272-274.

Coles, B. 1992. Further thoughts on the impact of beaver on temperate landscapes. In S. Needham and M. G. Macklin, editors. Alluvial archaeology in Britain. Oxbow Monograph 27:93-99.

Crum, H. A., and L. E. Anderson. 1981. Mosses of Eastern North America. Volumes I and II. Columbia University Press, New York, New York, USA.

Curtis, J. T. 1959. The vegetation of Wisconsin. University of Wisconsin Press, Madison, Wisconsin, USA.

Dansgaard, W. 1964. Stable isotopes in precipitation. Tellus 16:438-468.

Deines, P. 1980. The stable isotopic composition of reduced organic carbon. Pages 331-406 in P. Fritz and J. Ch. Fontes, editors. Handbook of environmental isotope geochemistry. Part A. Elsevier, New York, New York, USA.

Dorale, J. A., L. A. Gonzalez, M. K. Reagan, D. A. Pickett, M. T. Murrell, and R. G. Baker. 1992. A high-resolution record of Holocene climate change in speleothem calcite from Coldwater Cave, northeast Iowa. Science 258:1626-1630.

Elias, S. A. 1994. Quaternary insects and their environments. Smithsonian Institution Press, Washington, D.C., USA.

Faegri, K., P. E. Kaland, and K. Krzywinski. 1989. Textbook of pollen analysis. John Wiley & Sons, New York, New York, USA.

Freitag, R. 1969. A revision of the species of the genus Evarthrus LeConte (Coleoptera: Carabidae). Quaestiones Entomoloqicae 5:89-212.

Friedman, I., and J. R. O'Neill. 1977. Compilation of stable isotope fractionation factors of geochemical interest. In M. Fleischer, editor. Data of geochemistry. U.S. Geological Survey Professional Paper 440KK.

Gleason, H. A., and A. Cronquist. 1991. Manual of vascular plants of northeastern United States and adjacent Canada. Second edition. New York Botanical Garden, Bronx, New York, USA.

Gordon, R. D. 1985. The Coccinellidae (Coleoptera) of America north of Mexico' Journal of the New York Entomological Society 93:1-912.

Grimm, E. C. 1983. Chronology and dynamics of vegetation change in the prairie-woodland region of southern Minnesota, USA. New Phytologist 93:311-350.

-----. 1987. Coniss: a Fortran 77 program for stratigraphically constrained cluster analysis by the method of incremental sum of squares. Computers and Geosciences 13:13-35.

Hallberg, G. R., E. A. Bettis III, and J. C. Prior. 1983. Northeast Iowa's Paleozoic plateau. Iowa Conservationist 42:3-6.

Hallberg, G. R., J. C. Prior, and E. A. Bettis III. 1984. Geologic overview of the Paleozoic Plateau Region of northwestern Iowa. Proceedings of the Iowa Academy of Science 91:3-11.

Harmon, R. S. 1979. An isotopic study of groundwater seepage in the central Kentucky karst. Water Resource Research 15:476-480.

Higgins, K. F. 1984. Lightning fires in North Dakota grasslands and in pine-savanna lands of South Dakota and Montana. Journal of Range Management 37:100-103.

Hill, W. R., M. G. Ryon, and E. M. Schilling. 1995. Light limitation in a stream ecosystem: responses by primary producers and consumers. Ecology 76:1297-1309.

Hilsenoff, W. L. 1977. Use of arthropods to evaluate water quality of streams. Wisconsin Department of Natural Resources (Madison) Technical Bulletin Number 100.

Holloway, R. G., and V. M. Bryant, Jr. 1985. Late-Quaternary pollen records and vegetational history of the Great Lakes Region: United States and Canada. Pages 204-245 in V. M. Bryant, Jr. and R. G. Holloway, editors. Pollen records in Late-Quaternary North American sediments. American Association of Stratigraphic Palynologists Foundation, Dallas, Texas, USA.

Howe, H. F. 1994. Response of early- and late-flowering plants to fire season in experimental prairies. Ecological Applications 4:121-133.

Ives, R. L. 1942. The beaver-meadow complex. Journal of Geomorphology 5:191-203.

Kim, H. K. 1986. Late-glacial and Holocene environment in central Iowa: a comparative study of pollen data from four sites. Dissertation. University of Iowa, Iowa City, Iowa, USA.

Knox, J. C. 1977. Human impacts on Wisconsin stream channels. Annals of the Association of American Geographers 67:323-342.

-----. 1983. Responses of river systems to Holocene climates. Pages 26-41 in H. E. Wright, Jr., editor. Late Quaternary environments of the United States. Volume II. The Holocene. University of Minnesota Press, Minneapolis, Minnesota, USA.

-----. 1987. Historical valley floor sedimentation in the Upper Mississippi Valley. Annals of the Association of American Geographers 77:224-244.

-----. 1993. Large increases in flood magnitude in response to modest changes in climate. Nature 361:430-432.

Kuehl, R. J. 1982. Soil survey of Clayton County, Iowa. United States Department of Agriculture Soil Conservation Service, Des Moines, Iowa, USA.

Libra, R. D., G. R. Hallberg, R. D. Rowden, and E. A. Bettis III. 1992. Environmental geology of the Big Spring Groundwater Basin, northeast Iowa. Iowa Department of Natural Resources, Geological Survey Bureau Guidebook Series Number 15, Iowa City, Iowa, USA.

Lindroth, C. H. 1961. The ground-beetles (Carabidae, excl. Cicindelinae) of Canada and Alaska (2). Opuscula Entomologica Supplementum 20:1-200.

-----. 1963. The ground-beetles (Carabidae, excl. Cicindelinae) of Canada and Alaska (3). Opuscula Entomologica Supplementum 24:201-408.

-----. 1966. The ground-beetles (Carabidae, excl. Cicindelinae) of Canada and Alaska (4). Opuscula Entomologica Supplementum 29:409-648.

-----. 1968. The ground-beetles (Carabidae, excl. Cicindelinae) of Canada and Alaska (5). Opuscula Entomologica Supplementum 33:649-944.

Mather, D. J. 1992. Phase 1 archaeological investigation of the proposed Roberts Creek bridge replacement BROS-9022(35) in T95N-R6W, Clayton County, Iowa. Contract Completion Report 341, Office of the State Archaeologist, Iowa City, Iowa, USA.

Neilson, R. P. 1995. A model for predicting continental-scale vegetation distribution and water balance. Ecological Applications 5:362-385.

Prior, J. C. 1991. Landforms of Iowa. University of Iowa Press, Iowa City, Iowa, USA.

Rightmire, C. T., and B. B. Hanshaw. 1973. Relationship between the carbon isotope composition of soil C[O.sub.2] and dissolved carbonate species in groundwater. Water Resources Research 9:958-967.

Schwarcz, H. P. 1986. Geochronology and isotopic geochemistry of speleothems. Pages 271-303 in P. Fritz and J. Ch. Fontes, editors. Handbook of environmental isotope geochemistry. Part B. Elsevier, New York, New York, USA.

Schwert, D. P. 1992. Faunal transitions in response to an ice age: the late Wisconsinan record of Coleoptera in the north-central United States. Coleopterists Bulletin 46:68-94.

-----. In press. Effect of Euro-American settlement on an insect fauna: a paleontological analysis of the recent chitin record of beetles (Coleoptera) from northeastern Iowa. Annals of Ecology and Population Biology.

Schwert, D. P., T. W. Anderson, A. Morgan, A. V. Morgan, and P. F. Karrow. 1985. Changes in late Quaternary vegetation and insect communities in southwestern Ontario. Quaternary Research 23:205-226.

Shimek, B. 1915. The plant geography of the Lake Okoboji Region. Bulletin of the State University of Iowa, Bulletins from the Laboratories of Natural History 7.

Thomasson, J. R. 1992. Sediment-borne "seeds" from Sand Creek, northwestern Kansas: taphonomic significance and paleoecological and paleoenvironmental implications. Palaeogeography, Palaeoclimatology, Palaeoecology 85:213-225.

Thompson, C. A., and E. A. Bettis III. In press. Age and developmental history of Iowa Fens. Journal of the Iowa Academy of Science.

Tipping, R. 1995. Holocene evolution of a lowland Scottish landscape: Kirkpatrick Fleming. Part III. Fluvial history. Holocene 5:184-195.

Tornqvist, T. E., A. F. M. De Jong, W. A. Oosterbaan, and K. Van Der Borg. 1992. Accurate dating of organic deposits by AMS 14C measurement of macrofossils. Radiocarbon 34:566-577.

Transeau, E. N. 1935. The prairie peninsula. Ecology 16: 425-437.

Van Zant, K. L. 1979. Late glacial and postglacial pollen and plant macrofossils from Lake West Okoboji, northwestern Iowa. Quaternary Research 12:358-380.

Watts, W. A., and T. C. Winter. 1966. Plant macrofossils from Kirchner Marsh, Minnesota - a paleoecological study. Geological Society of America Bulletin 77:1339-1360.

Webb, T., III, P. J. Bartlein, and J. E. Kutzbach. 1987. Climatic change in eastern North America during the past 18,000 years: comparison of pollen data with model results. Pages 447-462 in North America and Adjacent Oceans During the Last Deglaciation. Volume K-3, The Geology of North America. Geological Society of America, Boulder, Colorado, USA.

Webb, T., III, and R. A. Bryson. 1972. Late- and postglacial climatic change in the northern Midwest, USA: quantitative estimates derived from fossil pollen spectra by multivariate statistical analysis. Quaternary Research 2:70-115.

Webb, T., III, E. J. Cushing, and H. E. Wright, Jr. 1983. Holocene changes in the vegetation of the midwest. Pages 143-165 in H. E. Wright, Jr., editor. Late Quaternary environments of the United States. Volume 2, The Holocene. University of Minnesota Press, Minneapolis, Minnesota, USA.

Webb T., III, and J. H. McAndrews. 1972. Corresponding patterns of contemporary pollen and vegetation in central North America. Geological Society of America Memoir 145:267-297.

West, R. G., R. Andrew, and M. Pettit. 1993. Taphonomy of plant remains on floodplains of tundra rivers, present and Pleistocene. New Phytologist 123:203-221.

Wright, H. E., Jr. 1992. Patterns of Holocene climatic change in the Midwestern United States. Quaternary Research 38: 129-134.

Wright, H. E., Jr., T. C. Winter, and H. L. Patten. 1963. Two pollen diagrams from southeastern Minnesota: problems in the regional late-glacial and postglacial vegetational history. Geological Society of America Bulletin 74:1371-1396.
COPYRIGHT 1996 Ecological Society of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1996 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Baker, R.G.; Bettis, E.A., III; Schwert, D.P.; Horton, D.G.; Chumbley, C.A.; Gonzales, L.A.; Reagan,
Publication:Ecological Monographs
Date:May 1, 1996
Words:16795
Previous Article:Intraspecific variation in sapling mortality and growth predicts geographic variation in forest composition.
Next Article:Host-plant specialization in western palearctic anthidiine bees (Hymenoptera: apoidea: Megachilidae)
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters