Paleoecology of a late Pleistocene wetland and associated mastodon remains in the Hudson Valley, southeastern New York State
Published:January 01, 2006
- PDF LinkChapter PDF
Norton G. Miller, Peter L. Nester, 2006. "Paleoecology of a late Pleistocene wetland and associated mastodon remains in the Hudson Valley, southeastern New York State", Wetlands through Time, Stephen F. Greb, William A. DiMichele
Download citation file:
Late Quaternary history and paleoecology of a small oxbow wetland on glaciated terrain were investigated using sediment lithology (cores, bulk samples, backhoedug trenches), ground-penetrating radar, vascular plant and moss macrofossil stratigraphies, and accelerator mass spectrometric radiocarbon dating. A nearly complete mastodon skeleton was recovered from late Pleistocene detrital peat and peaty marl near the top of the sediment sequence. Sedimentation in the basin began with silt and clay over dense cobble outwash transported southward from the nearby Hyde Park Moraine. Overbank sediment deposition occurred between ∼13,000 and 12,220 yr B.P. during a period of tundra vegetation, which ended with a sharp rise in spruce needle abundance and a shift to autochthonous marl and finally peat deposition. Fossils of aquatic and wetland plants began to accumulate before the tundra-spruce transition and increased after it. Rich fen wetland began to infill the pond with peat, while the upland supported open white spruce and later white spruce–balsam fir–tamarack forest. The mastodon, 11,480 ± 40 radiocarbon years old, was contemporaneous with spruce–balsam fir–tamarack forest and rich fen wetland. Many mastodon bones were articulated or nearly so, indicating that the animal died in the basin and that postmortem bone dispersal was slight.
Freshwater lakes of various size, depth, and origin were produced in glacial drift as the continental ice sheet disappeared from northern North America at the close of the Pleistocene. Sediments of such natural lakes and the wetlands now associated with them contain evidence of postglacial environments and environmental change both within and beyond the water-holding basin. The record usually starts soon after local retreat of the ice, and sedimentation may continue without interruption to the present, unless the basin is emptied by erosion or fills to capacity. The first sediments were sand, silt, and clay. As climate warmed and lake productivity increased, organic-rich lake mud accumulated, starting at or near the Pleistocene/Holocene boundary or early in the Holocene. Peat accumulation began in southern parts of glaciated North America in the early Holocene. However, the extensive peatlands now characteristic of northern United States and Canada formed and, in some places, expanded over the last four or five thousand years.
Wetland basins contain stratigraphically deposited fossils and fossil assemblages that have been subject to Quaternary paleoenvironmental research. In North America, for example, pollen analysis began with studies of sediments taken from peat-accumulating wetlands (e.g., Auer, 1927; Potzger, 1933; Sears, 1935). Although sediment in lakes is now preferred for research into regional patterns of vegetation and climate change, wetland sediments remain essential for investigations of basin infilling processes and wetland ecosystem development.
Here we present results of investigations of a fen wetland that developed in a small basin, the Hyde Park site, created by a stream draining southward from a recessional moraine during the late Pleistocene deglaciation of the Hudson River Valley. We use plant macrofossils and accelerator mass spectrometric (AMS) radiocarbon dating to characterize wetland and associated upland vegetation types and chronological relationships between them. A nearly complete and largely unaltered skeleton of a mastodon (Mammut americanum (Kerr)) was recovered from peaty fen sediment in the basin. We interpret the taphonomy of these remains with reference to the plant macrofossil record, and to the sedimentary history of the basin as revealed by ground-penetrating radar and other evidence.
Northern peat-accumulating wetlands are separated into fen and bog on the basis of water chemistry, plant indicator species, hydrology, and landform type (Glaser, 1992). Bogs are domed, whereas fens are concave landforms. In bogs, the water table is raised above that in adjacent non-bog terrain. Fens have acidic to alkaline waters (pH 4.2–7.2; Ca2+ concentration 2–50 mg l−1), mineral input from ground or surface water flow (versus the peat mass in bogs isolated above ground and surface water), and a distinctive set of bryophyte and tracheophyte indicator species, including different conifers, sedges, and mosses, than those typical of bogs. Fens range from acidic (poor) to calcareous (rich). Indicator plants and water chemistry distinguish between poor and rich fens. Stratigraphic studies of peatland sediment have shown replacement upward of calcareous fen plant communities by plants typical of poor fens or bogs (Janssens et al., 1993; Futyma and Miller, 1986).
Mastodon cheek teeth, tusks, bones, and more or less complete skeletons have been found throughout the Great Lakes states and southern Ontario (Dreimanis, 1968; Holman, 2001; McAndrews and Jackson, 1988) in sediment of shallow wetland depressions. Most mastodon fossils in New York State have come from western and central regions and the Hudson Valley and adjacent uplands (Dreimanis, 1968; Drumm, 1963; Thompson, 1994). Although the Hiscock site in western New York has continued to yield large numbers of mastodon fossils and a diverse associated fauna and flora (Laub, 2003), most other mastodon finds have been of single animals.
For most New York and many other mastodons, detailed information about the paleoenvironmental setting of the animal was never gathered. The Hyde Park wetland yielded a nearly complete mastodon skeleton dated at 11,480 ± 50 radiocarbon years (Beta-141061; G. Robinson, 2001, written comm.), and fen sediments and plant fossils of the same age. The proximity of many elements of the Hyde Park skeleton when it was excavated indicates minimal postmortem bone displacement. Therefore, discovery of the Hyde Park mastodon provided an uncommon opportunity to reconstruct the paleoenvironment in which the proboscidean lived and the postmortem history of the skeleton.
The Hyde Park site wetland developed in a small (40 m north to south, 28 m east to west) oxbow initiated during deglaciation by the Fall Kill (Miller, 2006), a stream that drains southward from and beyond the Hyde Park Moraine (Connally and Sirkin, 1986), which is located 1.5 km north of the oxbow basin. The wetland is in Dutchess County, New York, 3 km east of Hyde Park at 40°46′45″ N, 73°53′40″ W, 68 m asl, and ∼145 km north of the Wisconsinan terminal moraine, which extends along the length of Long Island, New York, and across Staten Island and northern New Jersey. Relief near the site is low (∼75 m).
The Hyde Park Moraine is one of several east-west trending moraines that were deposited as the continental ice sheet retreated northward up the Hudson River valley. A glacial lake was ponded in the valley between the ice front and a dam located to the south. Glacial varves accumulated in the lake, and large deltas were deposited along the valley rim by rivers draining into the lake from melting glacier ice. The chronology of deglaciation and glacial lake history in the Hudson River valley is poorly constrained.
The age of the Hyde Park Moraine is not known precisely. On the basis of the radiocarbon age of a composite bulk sediment sample associated with the onset of organic deposition at Eagle Hill Camp Bog, 25 km north of the Hyde Park Moraine, Connally and Sirkin (1986) calculated that the Eagle Hill Camp Bog basin began to fill at 16,070 yr B.P. However, none of the AMS ages of tundra plant fossils from the basal inorganic sediment in the Hyde Park wetland are older than 13,000 yr B.P., indicating that moraines in the mid-Hudson Valley are probably a few thousand years younger than previously thought (Miller, 2006).
When mastodon bones were discovered in 1999 at the Hyde Park site, the basin (hereafter referred to as Lozier Pond) had already been enlarged and deepened during excavations begun in 1966. As a result, it was not possible to study the contemporary wetland vegetation, which in any case was secondary, having become established after this earlier episode of pond construction. The property owner indicated that before disturbance a swamp, which exists several hundred meters northward beyond the pond margin, also covered the area now occupied by the pond. Extant upland, but mainly secondary, forests near the basin are dominated by oak (Quercus spp.) and hickory (Carya spp.) on drier sites, and hemlock (Tsuga canadensis (L.) Carr), sugar maple (Acer saccharum Marsh.), and other broad-leaved deciduous trees on mesic slopes and in shaded ravines. As a result of agricultural practices and settlement patterns, forest cover is discontinuous.
Pollen stratigraphic studies by Maenza-Gmelch (1997a, 1997b) at lakes 48 and 71 km south-southwest of the Hyde Park site show that forests of the middle and lower Hudson Valley were dominated by oak from middle to late Holocene. These were preceded by an early Holocene period of pine (Pinus) and oak abundance, and by late Pleistocene forests initially with abundant spruce (Picea) and later with reduced spruce and an increased amount of balsam fir (Abies balsamea (L.) Mill.).
Large-volume sediment samples for plant macrofossil analysis were taken from a sidewall of trench HPM-2, dug by workers during the mastodon excavation 1 m southeast of the area in which mastodon bones occurred (Fig. 1). Sampling began at an arbitrary 0 m datum placed at the bottom of a zone of sediment disturbed by pond building and continued to a depth of 1 m. Additional samples were obtained by bucket auger or posthole digger from 1 to 1.9 m. Cobbles at 1.9 m stopped further penetration. Two cores were taken with a 4-cm-diameter Eijkelkamp piston sampler (6987 EM Giesbeek, The Netherlands) from a wooden platform positioned where the least amount of sediment had been removed during the August 1999 excavation. Core LP3A was at the northern end of the pond, and 2.27 m of sediment was obtained. Core LP2A passed directly through the ribcage of the mastodon, and 2.25 m of sediment was collected. Although sediment depth at the time of our studies exceeded 2 m, the total thickness of sediment in the basin before disturbance can only be estimated. Lithology of the ∼1–2 m of sediment removed during pond constructions is not known. Water depth of Lozier Pond in June 2000 before in situ mastodon remains were discovered was 1 m at the location of core LP2A. All measurements of pre-disturbance sediment thickness assume that the top of the pond sediment before excavation was at the altitude of the floor of the swamp north of the pond. Because this surface is 1 m higher than the minimum water depth, sediment thickness has been increased by 1 m at the location of core LP2A to correspond to the pre-disturbance level.
Sediment lithology was described in the laboratory after microscopic examination of subsamples, testing with dilute hydrochloric acid, and other procedures. Sample volume was measured by displacement in water. Samples were disaggregated by hand in warm deionized water and washed with a low-velocity jet of tap water through a 250-µm-mesh sieve until clean. Identifiable plant macrofossils (leaves and needles, twigs, seeds and seedlike fruits, seed and pollen cones of conifers, megaspores, leafy moss plant fragments and individual moss leaves) were picked from the residues under low magnification and identified. Moss fragments and leaves were cleaned of adhering sediment with 00 sable hair brushes, dissected and/or sectioned by hand, and mounted in Hoyer's solution for examination by light microscopy. A large reference collection of plant parts taken from herbarium specimens was used when needed. Counts were made of fossils in each recognized category, and these data were standardized to fossil number per 300 ml of sediment for all samples and all macrofossil types identified, except mosses, for which frequency and abundance were described by a different method (Table 2). Plant macrofossils have been placed in the Quaternary Paleobotanical Collection of the New York State Museum.
To determine the nature and distribution of sediment elsewhere in the pond, we dug two backhoe trenches, one on the northwest side of the pond (Trench 1) and one on the southeast side (Trench 3), and the trench faces were described and sampled (Fig. 1). We also studied the wall of the mastodon excavation pit, and removed samples every 5 cm for macro- and microfossil analysis under the direction of David Burney (Department of Biological Sciences, Fordham University), using techniques described by Burney and Robinson (2006). Exploratory augering north and south of the pond established the presence or absence of sediments similar to those in the oxbow basin. As further confirmation of stratigraphic continuity, ground-penetrating radar (GPR) transects were run perpendicular to the axis of the sedimentary basin using a Pulse Ekko 100 GPR unit. After test lines were run at 50, 100, and 200 MHz, we determined that the unconsolidated sediments were best resolved with 100 MHz source emissions. Eight parallel transects, spaced at 1-m intervals, were run across the northern end of Lozier Pond. With the emitter set to pulse every meter, the device was placed in a plastic sled and pulled along the ground and across the water. The returning signals were recorded in the field on a laptop computer. Post-acquisition processing of the raw data was conducted using Win EKKO Pro software, and graphical displays were optimized using EKKO Mapper. Resolvable horizons were observed at depths of 4 m or greater (including standing water). The GPR profiles in conjunction with field observations of lithology permitted thickness and area of the sedimentary units to be determined throughout the basin. A full report of this work is presented by Nester et al. (2006).
A wooden post was driven into the sediment near where the first in situ mastodon bone was found during the 2000 excavation. Using a compass and tape measure, we established a 2 × 2 m grid, with stakes marking corners of the squares. All larger bones were tagged with an identification number while still in the pond sediment. Each bone was mapped in two-dimensional space by recording distances along two 1-m-long Jacob staffs positioned along the north-south and east-west sides of the squares. Depth was measured from the same 0 datum established for sediment samples taken within the mastodon excavation zone, namely a point just above the highest exposed bone. To maintain the same zero-point across the bone field, we extended string across the grid from the zero-datum point, and used a spirit level to ensure a uniform horizontal reference line above the squares. The distance between the zero datum and a bone was measured along a plumb-bob line. Because of possible disturbance and breakage, we removed smaller bones before mapping them, but not before we had recorded their positions within a square. Data for individual bones were then placed in virtual space using a graphics software package (Discreet's 3ds max 4), which allowed three-dimensional relationships to be studied. All original maps of the disarticulated skeleton created in the field and the computer graphics files are in the Collections Department, Paleontological Research Institution, Ithaca, New York.
Miller was responsible for the paleobotanical data and its interpretation; Nester collected and analyzed data pertaining to the distribution of sediments in the basin and the mastodon skeleton.
We recognized five sediment types. The bottomsmost was (1) cobbles in a matrix of silt and clay. Above this was (2) clayey silt with some coarse gravel and cobbles. Marly, clayey silt (3) occurred next higher up. Above this, marl and degraded plant matter (peaty marl ) were mixed with gastropod, bivalve, and ostracod shells, along with fine sand and silt. Capping these sediments was a layer of fine-grained, non-marly detrital peat (5). Between units (2) and (3) and units (3) and (4) the contact was gradational, whereas between units (1) and (2) and units (4) and (5) it was abrupt. Nowhere in the pond did the nonsystematic backhoe excavation penetrate deeper than unit (4). Sedimentary units are continuous within the study area (Fig. 2). Exploratory augering revealed that the depth to the cobble horizon, the lowest unit encountered stratigraphically, decreased toward the north and south ends of the pond, with the deepest part of the basin 8–25 m northeast of the mastodon bone field. Ground penetrating radar transects at the north end of the pond indicated a more or less symmetrical basin, with the pond axis oriented roughly north to south (Fig. 2). These profiles also revealed truncation of the peaty marl (unit 4) by the overlying peat (unit 5). Insufficient data were collected to the southeast of Lozier Pond to extrapolate beyond the limit of the existing water body.
The sediment samples produced a robust vascular plant macrofossil stratigraphy (Fig. 3; Table 1), with Tundra (lower) and Spruce (upper) zones clearly demarcated by plant macrofossils. The transition between Tundra and Spruce zones was abrupt (Fig. 3). As expected, given the history of digging at the wetland, the record was truncated at the top, and no Holocene sediment and macrofossils were present in the studied sample series. The Tundra zone yielded fossils of 21 species and eight additional genera, whereas the Spruce zone contained fossils of 22 species and 11 additional genera. Fossils identifiable only to genus had no species-diagnostic features. The full record is presented and interpreted by Miller (2006), but here we focus mainly on fossil wetland plants.
Macrofossils of wetland plants occurred from the upper Tundra zone through the Spruce zone, with only a few recovered lower in the Tundra zone. Rooted aquatics (Myriophyllum sibiricum Kom., Najas flexilis (Willd.) Rostk. & Schmidt, Potamogeton spp., Sparganium sp.) had uninterrupted records in the Spruce zone, whereas others (Carex vesicaria L., Hippuris vulgaris L., Lycopus americanus Muhl., Ranunculus sp., Sagittaria latifolia Willd., Selaginella selaginoides (L.) Link, Viola sp.) were present discontinuously. Fossils of emergent wetland plants (Carex aquatilis Wahlenb., Eleocharis palu stris (L.) R. & S., Hypericum virginicum L., Scirpus tabernaemontani Gmelin, Typha sp.) were restricted to the topmost Spruce zone samples. Fossil sedge achenes (Carex spp.; nine types in addition to those with well-preserved perigynia, which for this reason could be identified to species) were recovered from samples throughout the Spruce zone, and also from the top three Tundra zone samples. Although not the subject of this paper, it is important to note that aquatic and wetland plants occurred with fossils of upland (non-wetland) species—leaves of mountain avens (Dryas integrifolia Vahl) and other arcticalpine plants in the Tundra zone and needles and cones of white spruce (Picea glauca (Moench) Voss), balsam fir, and tamarack (Larix laricina (DuRoi) K. Koch) in the Spruce zone (Fig. 3).
Moss fossils, including fragments of leafy stems, leaves, calyptrae, opercula (with leafy stems most common), were present in all samples of both zones (Tables 1,2). Wetland species were most frequent and abundant in Spruce zone sediment. Eighty-four slides were prepared, and from these, 30 moss taxa (identifications to species, genus, or family) were recognized, including 22 species, and fossils of plants representing five genera, four of which were genera other than those of the 22 species. One taxon, Polytrichum juniperinum Hedw./P. strictum Brid. type, includes two possible species, neither of which can be conclusively identified from incomplete fossil material. Nine slides contained unidentified fossils, seven of which were poorly preserved, whereas the other two were well-preserved fossils of unrecognized species. Wetland mosses included the following: Calliergon giganteum (Schimp.) Kindb., Calliergonella cuspidata (Hedw.) Loeske, Campylium stellatum (Hedw.) C. Jens., Drepanocladus aduncus (Hedw.) Warnst., D. sordidus (C. Müll.) Hedenäs, Meesia triquetra (Richt.) Ångstr., and Warnstorfia exannulata (Schimp. in B.S.G.) Schimp. Other species of wet habitats, including seeps over rock and soil, were Philonotis fontana (Hedw.) Brid. and Dicranella schreberiana (Hedw.) Hilf. ex Crum and Anders. Non-wetland mosses were mostly restricted to sediments below 70 cm, i.e., the Tundra zone (Table 2). Species of dry to moist soil and rock were present, including Abietinella abietina (Hedw.) Fleisch., Bryoerythrophyllum recurvirostre (Hedw.) Chen, Ceratodon purpureus (Hedw.) Brid., Distichium sp., Ditrichum flexicaule (Schwaegr.) Hampe, Hypnum revolutum (Mitt.) Lindb., H. vaucheri Lesq., Mnium thomsonii Schimp., Polytrichastrum alpinum (Hedw.) G. L. Sm., Timmia norvegica (Web.) Wahlenb. ex Lindb., Tortella fragilis (Hook. & Tayl. in Drumm.) Limpr., and Tortula norvegica (Web.) Wahlenb. ex Lindb.
More than 80% of the mastodon skeleton was recovered, including all long bones, ribs, vertebrae anterior to the sixteenth caudal, the skull, mandible, and both tusks (Table 3). Possible missing bones include several of the most distal caudal vertebrae and a few foot bones. All bones were pristine, and the only postmortem damage to the skeleton was caused by the backhoe used in the excavation of the pond by the landowner (Fisher, 2006). Original positions of the larger bones of the Hyde Park Mastodon skeleton when it was excavated in August 2000 are shown in Figure 4. Missing from the figure are the caudal vertebrae and ribs, toe, wrist/ankle, and other small bones (e.g., hyoids). These were found but have been omitted from the diagram for clarity. Prior to the discovery of mastodon bones by the landowner in 1999, the excavation had disturbed and removed several bones, including the skull, left humerus, right radius and ulna, the thirteenth caudal vertebra, and several toe bones. Therefore, their original position in the sediment is unknown. One exception was the skull, which was found in a spoil bank on the east side of the pond. The proximal quarter of the left tusk was still lodged within the tusk cavity, with ∼50 cm projecting beyond the alveolar margin. The complementary section of tusk was found in situ in the pond sediment, thereby allowing the original location of the skull to be determined. The majority of the ribs were found immediately north and east of the reference post shown in Figure 4. Figure 5 shows most of the bone field as it was being excavated.
Large sections of the skeleton were articulated or nearly so. The proximal end of the right femur was in a state of near-articulation with the corresponding socket of the right innominate (pelvis). Similarly, the distal end of this femur was in near-articulation with the articulated right tibia and right fibula. The distal portion of these two skeletal elements was just above the nearly complete and well-articulated right pes (hind foot). Only the most distal phalanx of the first toe was missing from the right pes. Several complete sections of the vertebral column were also found, with many of the cervical, thoracic, and lumbar vertebrae articulated. A segment from the sixth cervical vertebra to the ninth thoracic vertebra was uncovered at the southern end of the bone field. A segment of spine from the twelfth thoracic vertebra to the third lumbar vertebra was found in a state of articulation to near-articulation. Bones were present in a zone extending from the lower 50 cm of detrital peat (unit 5) down to the lower portion of the marly, clayey silt (unit 3). An even greater concentration of skeletal remains occurred over a 40-cm interval that brackets the peaty marl (unit 4). The lowest bones recovered, both stratigraphically and in absolute depth, were foot bones from the articulated right pes, which were planted in marly, clayey silt (unit 3) near the bottom of the shallow pond by the animal before it died.
Sediments and Sedimentary History
Two depositional processes operated to fill the Hyde Park oxbow basin at different times in its history. During the early phase, which probably began as glacier ice withdrew from the Hyde Park Moraine or another position closer to the ice front, periodic overbank flooding by the paleo-Fall Kill delivered silt, clay, plant matter, and sometimes gravel and cobbles. In the second phase, the basin became separated from the Fall Kill, and autochthonous marl deposition began. The shallow pond of phase two supported rooted aquatic and emergent plants. Later, intermixed marl and detrital peat accumulated as peat-generating plant communities became established along the margin of the pond. Finally, the deposition of marl ceased and detrital peat accumulated. The onset of detrital peat deposition was abrupt, beginning ca. 11,230 ± 50 yr B.P. on the basis of the age of a balsam fir (Abies balsamea) cone scale deposited in the wetland from a tree growing on the adjacent upland. The transition period between marly peat and non-calcareous detrital peat deposition appears to have been one of brief erosion or non-deposition, as indicated by the truncated reflectors as well as the sharp contact seen in field observation. The early phase of deposition was estimated to begin ca. 13,000 and continue to ca. 12,200 yr B.P. (Miller, 2006).
Menking et al. (2006) studied core LP2A in detail and obtained late Pleistocene ages for peat (unit 5), peaty marl (unit 4), and marly, clayey silt (unit 3). Peat from 1.64 to 1.68 m (unit 5) was 10,370 ± 50 yr B.P. (AMS; Beta-169810). Peaty marl (unit 4) at 1.84–1.85 m was 11,360 ± 40 yr B.P. (AMS; Beta-159335). Three ages were obtained from marly, clayey silt (unit 3): 12,410 ± 50 yr B.P. (AMS; Beta-159338) from 1.33 to 1.375 m, 12,720 ± 50 yr B.P. (AMS; Beta-159337) from 1.25 to 1.27 m, and 12,970 ± 40 yr B.P. (AMS; Beta-159336) from 1.125 to 1.165 m.
Ages from sample series HPM-2 are also late Pleistocene, from 11,230 ± 50 yr B.P. (AMS; Beta-175554) obtained from a single balsam fir cone scale at the top of unit 4 to 12,880 ± 50 yr B.P. (AMS; Beta-175557) obtained from six Dryas integrifolia (Rosaceae) leaves near the bottom of unit 3. This date and 12,970 ± 40 yr B.P. (Menking et al., 2006) are the oldest reported from the pond. Two other dates obtained in this sample set from spruce (Picea) needles (12,300 ± 40 yr B.P. [AMS; Beta-175555] at 1.40–1.50 m and 12,120 ± 40 yr B.P. [AMS; Beta-175556] at 1.60–1.70 m) record an inversion in unit 3 at roughly the same levels as the inversion reported above from core LP2A. This is possibly the result of reworked sediment during overbank deposition or perhaps bioturbation by either aquatic organisms living in the oxbow basin or a terrestrial mammal such as a mastodon in search of food or water. The Hyde Park mastodon, 11,480 ± 50 yr B.P., is younger than these ages. Vascular plant macrofossil and pollen stratigraphies from the face of trench HPM-2 (Miller, 2006; Robinson and Burney, 2006) revealed a tundra environment at the site from the deposition of the basal cobbles (unit 1) to an abrupt increase in spruce needle and pollen deposition at 0.7 m (adjusted depth 2.3 m, i.e., from the original projected surface prior to excavation disturbance). A Holocene record was not present in the HPM-2 sample series.
The plant macrofossil assemblage and stratigraphic changes in it reflect the floristic composition of the wetland plant community through time. Open, shallow, standing pond water occurred in the oxbow during the Spruce zone on the basis of the continuous presence of pondweed and water milfoil fossils, which are produced by obligately aquatic plants. However, at the wet margin of the pond grew sedges, bulrushes, spike-rushes, and other emergent herbaceous plants. Pollen assemblages from the same interval of pond sediment (Robinson and Burney, 2006) support this determination but provide less taxonomic resolution than the macrofossils. Sedge (Cyperaceae) pollen (5%–13% of the sum of terrestrial herbs, shrubs, and trees) was registered consistently in all Spruce zone samples. The stratigraphic distribution of sedge pollen parallels the sedge macrofossil record. Bulrush, spike-rush, and Carex aquatilis macrofossils occurred in the uppermost Spruce zone samples, indicating an expansion of the area supporting these rooted, emergent plants as the basin shallowed along the pond margin from infilling.
The pond periphery was a calcareous fen associated with circumneutral pond water of moderately high cation content and conductivity. Water chemistry (pH and alkalinity) measurements from numerous wetlands throughout New England where plants of the Hyde Park pondweed and water milfoil species grow at present (Crow and Hellquist, 1983; Hellquist, 1980; Hellquist and Crow, 1980) indicate that the Hyde Park pond water fell within a pH range of 6.5–9.8, with alkalinity between 15 and 170 mg HCO3 − liter−1 (Miller, 2006). These conditions are sufficient to support a fen of the floristic composition indicated by the macrofossils.
The type of wetland is further specified by the fossil moss assemblage that was recovered from sediments of the Spruce zone (Table 2). Although mosses in a wetland form the ground layer beneath the emergent vascular plants (and therefore mosses are part of a broadly defined vegetation), a fossil moss record is largely independent of a paleoenvironmental assessment on the basis of fossils of vascular plants alone. Fossil mosses can be used, therefore, to test and interpret inferences derived from other proxy data sources. Miller (1980) has reviewed the application of moss fossils in Quaternary paleoecology. Information about Pleistocene and Holocene fossil bryophytes from wetland and other types of Quaternary sediment has increased greatly in recent years, particularly in connection with investigations into the development of peat-generating systems. The moss assemblage of the Hyde Park Spruce zone indicates rich fen wetland (Miller, 1980). Peatland of this sort develops in association with base-rich water that is replenished by ground- or stream-water flow. Rich fens are found in areas of calcareous bedrock or glacial drift, and they and/or calcareous seeps appear to have been common in the late Pleistocene. Some late Pleistocene rich fens persisted through the Holocene (Futyma and Miller, 2001), and organisms in them can be of special conservation concern. Scorpidium scorpioides (Miller, 1980) and Meesia triquetra (Montagnes, 1990) are considered to be rich-fen indicator species. These mosses and others in the Hyde Park Spruce zone, Calliergon giganteum, Calliergonella cuspidata, and Campylium stellatum, are common associates in extant rich fens. Janssens (1983) considered Drepanocladus sordidus (as D. sendtneri; see Hedenäs 1998 for a taxonomic revision of these and related species) to be a moss of rich fens and reported that its associate in the Hyde Park Spruce zone assemblage, D. aduncus, occurred in contemporary peatlands having the following water chemistry: pH 6.6–7.5, conductivity 150– 575 µS, and Ca2+ concentration 23–61 ppm in his study area, which extended between British Columbia and Yukon Territory eastward to Minnesota and Michigan.
There is no paleobotanical evidence from the samples studied that fen plants were replaced upward stratigraphically by fossils of oligotrophic fen or bog plants. Heath (Ericaceae) pollen from bog and oligotrophic fen plant communities occurred (Robinson and Burney, 2006), but only sporadically and in low frequency in Spruce zone sediment. However, because the basin did not capture much of a Holocene record, the wetland type occurring in later stages of basin infilling cannot be specified. Wood from the top of an unexcavated island of sediment located above the mastodon remains was Holocene (5420 ± 50 yr B.P.; Menking et al., 2006), indicating that some Holocene sediment may have been present in the basin prior to the two episodes of pond excavation. But because the basin is shallow and most of the sediment filling is late Pleistocene, water was never more than about one to two meters deep. Therefore, it seems likely that a long Holocene record never accumulated. Alternatively, if one did, the record was lost as organic sediment in the shallow basin oxidized and disappeared with changes in the water table position caused by variation in precipitation and groundwater input. Thus, the vascular plant and moss macrofossils were from a wetland that existed in the basin over a relatively short period beginning just before the Pleistocene/Holocene boundary and perhaps continuing some hundreds of years into the early Holocene.
The radiocarbon (AMS) age of collagen extracted from a sample of one of the Hyde Park mastodon tusks was 11,480 ± 50 yr B.P. Thus, the time of death of the animal was ∼200 radiocarbon years before the deposition of a cone scale of balsam fir, which occurred with other macrofossils of balsam fir and with those of white spruce and tamarack. These macrofossils, and spruce- and balsam-fir-containing pollen assemblages from the same interval, indicate that the upland forest vegetation near the basin consisted of these trees, rather than of white spruce alone, macrofossils of which occurred in low numbers and sporadically in samples from the lower Spruce zone. It is likely that pine (perhaps jack pine, Pinus banksiana Lamb.) and tree birches (Betula sp.) grew near the basin throughout the Spruce zone, but there is no macrofossil evidence of this. These two pollen types were, however, registered during pollen analysis of the interval (Robinson and Burney, 2006).
The Mastodon and Its Paleoenvironment
Haynes (1991) noted that the number of bones in the skeleton of a healthy adult American mastodon might have been variable. A review of the published literature (Osborn, 1936; Olsen, 1972; Haynes, 1991) indicated the number of bones to be 235–255, with the number of bones in the feet and tail in greatest contention. Two hundred and fourteen bones of the Hyde Park mastodon have been accounted for, with only possibly the distal caudal vertebrae and 34 of 128 foot bones, most of them the smaller phalanges and sesamoids, not recovered. The Hyde Park mastodon skeleton is therefore 84%–91% complete.
The rearticulated Warren mastodon from near Newburgh, New York (Osborn, 1936), offers some useful comparisons. Like the Hyde Park mastodon, this skeleton is essentially complete, with only some toe bones and caudal vertebrae unaccounted for. The skeleton measures 4.55 m (14′ 11″) from the most distal caudal vertebra to the tusk base, and 2.78 m (9′ 2″) from the floor to the highest spinous process of the thoracic vertebrae. Bones of the Warren and Hyde Park mastodons have similar measured dimensions. These animals, both males, were probably nearly the same size at death.
Taphonomic information about the skeleton and the position of the bones in the pond sediment contribute information about the sedimentary history of the basin and paleoenvironment at the time of and immediately following the death of the animal. Considering that all of the bones found in situ originated from an area ∼5 × 5 m, and that many elements of the skeleton remained articulated, there was very little postmortem scattering of the bones. Discovery of the lower thoracic and lumbar vertebrae (the smallest bones of the spinal column anterior to the pelvis) in a state of semi-articulation indicates settling and burial of this part of the mastodon carcass in quiet water. Although some of the skeletal elements are disorganized relative to one another, the articulated nature of others suggests little transport and disruption of the decaying carcass. The plant macrofossils also show that the Hyde Park oxbow had become isolated from the Fall Kill and contained a rich fen and associated shallow-water pond at the time of death of the mastodon, while land above the basin supported (Miller, 2006) white spruce and/or white spruce-balsam fir-tamarack forest. Bones of the Hyde Park mastodon's right rear leg was found in what appears to be life position, with the right pes planted in the clayey silt. Therefore, it is likely that the mastodon, in one of its final actions before dying, pushed its leg deep into the uncon-solidated pond sediment in which it became mired. The Hyde Park mastodon was not transported by stream-water flow to the oxbow basin but instead died and was buried and preserved near or at its place of death. The mastodon died in the basin during the transition from peaty marl to detrital peat deposition (Fig. 2).
Water levels must have been high enough that the decaying animal was never exposed subaerially, but shallow enough that the animal was able to walk to this location and plant its foot into the sediment. We estimate this depth to have been 1–2 m. In addition, the excellent condition of the bones indicates that the water table remained high enough for the remains to become completely covered by sediment. The skeleton was minimally disturbed during the nearly 11,500 radiocarbon years following its death. The movement that took place appears to have been the result of random floating of individual segments of the carcass in a closed basin, because there is no preferential orientation of the remains. Because of the pristine condition of the bones and the amount of articulation observed, there is a strong possibility that the unaccounted for bones are still buried in the sediments of Lozier Pond. The complete disarticulation history of the skeleton is presented by Fisher (2006).
Mastodon remains have been found in a variety of depositional basins in the Great Lakes region of North America. Small kettle holes seem to be the most productive sites for mastodon fossils, but remains have been found in spring deposits (Laub, 2003), potholes (Hall, 1868), sinkholes (Harington et al., 1993), kames (Hodgson et al., 2006), swales of uncertain origin (McAndrews and Jackson, 1988), and on the submerged Atlantic Coastal Plain (Oldale et al., 1987). Most discoveries have been made fortuitously, and often as a result of draining or modifying land for agriculture or recreation. When contextual paleoenvironmental data were obtained in conjunction with recovery at sites in the Great Lakes region (Table 4), mastodons are often associated with calcareous, peaty, rich-fen sediment and peat consisting largely of sedge remains. Calcareous sedge fens are an initial successional stage in kettle hole basin infilling (terrestrialization). They can persist through the Holocene to the present or are replaced by oligotrophic fen communities and sometimes bogs, depending on changes in groundwater hydrology and chemistry, ecosystem productivity, and other autogenic and allogenic factors that may be specific to a basin or carry a regional message related to climate and other large-scale environmental factors. In contrast to the wetland vegetation, the upland near the Hyde Park basin supported open white spruce forest initially and later open forest of white spruce, balsam fir, and tamarack. Dreimanis (1967, 1968) observed that a spruce forest–mastodon association was common in the Great Lakes region and suggested mastodon extinction was in response to a sharp decrease in the area occupied by this kind of conifer forest. It is by no means certain that his hypothesis explains the loss of this megavertebrate, but the association between mastodons and late-glacial conifer forest has remained well established for the Great Lakes region.
The authors wish to thank the property owners, Larry and Sheryl Lozier, for donating their discovery to science. David Burney and Guy Robinson provided much expertise in the field and aided in the interpretation. Daniel Fisher and James Sherpa carried out mastodon bone identification, and Sherpa also provided helpful comments on a draft of the manuscript. Fieldwork was performed with the assistance of Warren Allmon, Paul Harnik, Robert Ross, and Elizabeth Humbert. Three-dimensional placement of the bones was conducted with the assistance of Elizabeth Humbert and Brian Dunston. Financial support for the project was provided by a grant from the Eppley Foundation. We also thank Daniel Fisher and John McAndrews for comments and suggestions on the submitted manuscript.
Figures & Tables
Wetlands through Time
- absolute age
- clastic sediments
- Dutchess County New York
- glaciated terrains
- Hudson Valley
- New York
- paludal environment
- terrestrial environment
- United States
- upper Pleistocene
- southeastern New York
- Hyde Park New York