AN ARCHAEOLOGICAL COMPLEX whose exact spatial extent will likely never be known occurs along the shore of Long Island Sound at Norwalk, Fairfield County, Connecticut.  One manifestation of this complex is a shell midden, strikingly exposed in the vertical banks of an artificially dredged boat basin Figs. 1 A, 2).

  Other manifestations are shell pits in plowed fields northeast of the basin; subsidiary midden outcrops all along the bank of Charles Creek west of the basin; purported human burials in a "mound" along this same bank; midden outcrops along Calf Pasture Point beach (Fig. 1); and other occurrences north and east to the vicinity of Sasqua Hill, where another extensive midden is still under investigation.  Several sizable local collections of artifacts derive from this area. No previous systematic investigations are known. 

    Fig. 1.  Connecticut shore, vicinity of Spruce Swamp site.  Light shading shows present sea surface; dark shading shows inferred lower sea surface about 3000 B.P. Central outlined rectangle area shown in Fig. 2.

      Modern development and unrestricted pot hunting steadily destroy the archaeological remains and obscure their relations. 

       Fortunately, one collection available for study is accompanied by cataloged information on the circumstances of recovery.  This is the collection of  Ted Jostrand, a local amateur archaeologist.  I have relied heavily on his data and on his personal knowledge of the site, which goes back more than 20 years.

      In this paper, I shall be mainly concerned with the boat-basin midden, which I call Spruce Swamp midden in deference to an early, published map reference (Anonymous n.d.) to the swamp that preceded the modern basin. The artifacts described in this paper, unless otherwise indicated, can be assigned with a high degree of probability to the midden.

Basin and Bog

      The earliest map I have found (Beers and others 1867, Map No. 25, Town of Norwalk, indicates a quagmire at this location. 
 

    Fig. 2.  Detail map of Spruce Swamp site (rectangle shown in Fig.1).
    Stippled area shows inferred extent of the midden. (Post-publication
addenda 4/28/00: SS-1 marks locus of very extensive gridded area excavated by the former NEAR under my direction, and so far unpublished; SS-2 marks locus of a possible trephined skull from an Amerindian grave, subsequently published by me in Science, 11/13/70, Vol. 170, pp. 732-734). 

      In the 1930's, Calf Pasture Drive was built across the swamp, cutting it in two. 
       The portion to the east was left intact and is today a bog that still preserves the character of the original swamp. 

       However, in 1958, the portion to the west of the roadway was dredged for a boat basin.  Formerly, only a very small inlet connected the basin portion to Charles Creek, and salt-water incursions into the basin portion were limited to high tide. However, the dredging cut down a natural gravel and sand bar along the western. edge and so opened the basin up to Charles Creek (Fig. 2). The basin was dredged to a fairly uniform depth of 10 ft. below mean sea level.  It is believed, on the basis of verbal descriptions by the dredge owner, that glacial drift underlay the black muck of the basin at no great depth in the western end.  However, toward the center of the basin and toward its eastern end the muck deepened considerably. At one point (Fig. 2 M) a 55-ft. piling for a boat pier came to buoyancy after sinking 45 ft. under its own weight without driving. The dredge owner interprets this to mean that the muck is over 40 ft. in depth. No bottom (drift) was reached.

      Immense quantities of logs and stumps, wholly unanticipated, were encountered during dredging, and these continued to the greatest depth reached. They entangled the dredge clam-bucket and jammed the gates on the barge. Most were dumped at sea, but a few were left to weather at the southern end of the bog. Many other stumps, some apparently in situ, can be seen in the bog today.  I have been unable to reach these latter stumps for sample extraction but I have photographed them by telephoto lens. Samples of bark and wood found near mean low-water level were recovered from the basin.  Botanists (Howard S. Irwin and Herman F. Becker, personal communication) to whom the samples and photographs were submitted have been unable to identify the material. To the untrained eye, the bark suggests a conifer.

The Midden

      The midden is a compact stratum of marine shell that surrounds most of the basin and remnant bog (Fig. 2, shaded area) and has a maximum width of 300 ft. on the north side of the basin (Jostrand, personal communication). It  is now overlain by about 5 ft. of dredgings and other fill. Much of this overburden is sterile drift dumped here during the dredging. It has a roughly stratified, waterlaid appearance enhanced by a few seams and occasional lenses of muck.  Its character is wholly artificial, since it dates from the dredging; the layered appearance no doubt due to the water entrained in it when it was dumped, as well as occasional rains which reworked it during its brief exposure.  The midden is quite well exposed in vertical section beneath the overburden, particularly on the north side (Fig. 2 A-H).  My observations, including the stratigraphic column analysis in Table 4, have been largely made on this northern exposure.

      Fig. 3B.  A partial profile of the surface of the Spruce Swamp Midden relative to current sea stand.  Station A is the only 
      Station still above current mean high water: it is the locus of a stratigraphic column analysis.  A-H are stations leveled 
      by alidade and plane table: all exposed on north face of the boat basin. Stations I-L lie east of the roadway on north edge
 of the bog and are not shown here. 

      The total exposed length in the basin is about 350 ft. extending from A to H in Fig. 2 and shown in exaggerated profile in Fig. 3 B.  At H, the midden disappears under the fill of Calf Pasture Drive and reappears at I (Fig. 2) in the bog.  In August, 1962, a series of test trenches (Fig. 2, 1, J, K, and L) was dug in the muck along the shore of this bog, and these illuminated minor stratigraphic relationships and confirmed the presence of the midden stratum to about L in the east end of the bog, where the layer thins to less than 1.in.

      In October, 1963, I undertook a plane-table leveling survey to establish the relationship of the midden to mean sea level.  Previously determined bench data for the shore of Calf Pasture Point, which were furnished by the City Engineer’s Office, greatly facilitated this project. The relationship of the midden profile to mean sea level is shown in Fig.3 B.  Vertical exaggeration here suggests that the midden is more irregular than it actually is; as seen by the naked eye from the opposite side of the basin, the midden from A to H has only a slight slope to the east.  The vertical exaggeration in Fig. 3 B facilitates specific station locations on the profile; it indicates that most of the midden today lies between mean sea level and mean high water level and is exposed only briefly during the intertidal intervals.

      The midden can be described as a rather dense marine-shell stratum; it shows thin layers of shell and shell fragments with slightly different orientations from the bulk of the shells.  It contains whole and broken artifacts, sparse and poorly preserved potsherds, and animal bones. These materials are intermixed with quantities of greasy, black, organic soil.  A stratigraphic column at Station B (Figs. 2 B, 3 B) near the thickest part of the midden was analyzed according to criteria and methods originally stated by Meighan and others (1958) and further clarified and adapted to materials of this region by Salwen (1962, 1963).  I have made my Table 4 in the same fashion as Salwen's (1962) Table 1 for ease of comparison.

      The midden may be conformable to a bed of glacial alluvium: yellow-tan subsoil charged with pebbles and lenses of pebbles and grit, probably a terrace remnant of the Pleistocene Norwalk River.  The juncture is fairly sharp however, some leaching and downward migration of stains and burrowing activities of fiddler crabs (Uca) and other marine organisms it difficult to determine whether a thin topsoil preceded deposition of the midden.  I believe it did; its thickness may have been about 2in.

Artifacts

      The artifacts from this midden are listed in Table 1. A few specimens derive from the subsoil beneath the midden (maximum depth below midden: 10 in.).Recoveries have been made in two ways: (1) as surface finds on top of the midden during the many years when it lay exposed along the edge of the original basin-and-bog and (2) as finds pried from the face of the layer now vertically exposed in the basin.  Depth proveniences have not been recorded often.  Most  artifacts are projectile points (Tables1 and 2).

Table 1. SPRUCE SWAMP: ARTIFACTS AND FEATURES

         Item

     Number 

                            Comments

Projectile points

       111

Further defined in Tables 2 and3

Potsherds

        52

Further defined in text

Scrapers

   Numerous

Many types

Splinter awl fragments

         2

Polished

Pebble pendant

         1

Drilled from both sides

Spearheads

      Several

Stemmed; not included in Table 2

Incised paintstone

         1

Complex design; of tabular hematite

Incised fragment

         1

Design both sides; of slate

Full-grooved ax

         1

Polished; basalt (?); 3 to 4 in. in sub-soil beneath bottom of midden 

Plano-convex celt 

         1

Polished; basalt (?); near midden center

Sharpening stone or paintstone

         1

No remarks

Pebble paintstone

         1

Hematite

Tabular paintstone

         1

Hematite

Sinew stone

         1

Quartz cobble

Crystal

         1

Quartz (“rock crystal”)

Hide scraper 

          1

Ovate; large; green plasma (?); two pieces 6 ft. apart horizontally and 1 ft. vertically, with one piece well into subsoil; may be evidence for pit or intrusion by later cultures 

Animal talon

          1

Species unknown

Cache

          1

Quartz cobbles (17) at juncture of midden and subsoil in area 3 ft. in diameter; 13 have one flake only removed from opposite sides at each end; four probably modified same, but with further flaking into recognizable cores 

45 caliber slug

          1

Intrusive; late date

30 caliber slug

          1

Intrusive; late date

Fire-cracked stones

    Numerous

Probably hearth remnants

Table 2. PROJECTILE POINT MORPHOLOGY

Number of Specimens

                Specimens

                Percent of Total 

            19

              Narrow Bladed

                        17.1

            39

              Broad Bladed

                        35.1

            53

              Triangular

                        47.7

           111

              TOTAL

                        99.9

Narrow Bladed                                  Broad Bladed                                   Triangular 

Stem

Side Notch

Corner Notch

Stem

Side Notch

Corner Notch

Large

Small

Pentagonal

Total

    12
 10.8

       7
       6.3

          0

    14
 12.6

       19
       17.1

          6
           5.4

    8
    7.2

  43
  38.7

        2
        1.3

  111
  99.9

      Ceramic finds have not been plentiful, but potsherds have been reported through the entire vertical extent of the midden.  No sherds have come from the assumed ancient topsoil beneath the midden or from the subsoil beneath it.  The following types in a population of 52 members have been identified from rim or near-rim sherds, using the published criteria of Smith (1950): Bowmans Brook Stamped, Van Cortlandt Stamped, Windsor Fabric Marked, and Windsor Brushed.  There are also smooth-surfaced sherds and a few nondescript sherds that show lip-notching and other decorative techniques.

      Lithic materials at Spruce Swamp are quite diverse.  At least one kind of stone is exotic for the area, and several others may be. The following have been identified by the mineralogist David M. Seaman (personal communication): chert, flint, hematite (both pure and highly arenaceous), plasma (variant of quartz from “either Alabama or Georgia"), quartz, quartzite (both fine- and coarse-grained), shale (both ferruginous and carbonaceous), and native copper (showing malachite alteration stains in a fine-grained quartz conglomerate)

Discussion

      Early attempts to relate archaeological manifestations with the heightened sea level of late postglacial time along the southern New England shore are exemplified by the work of Deevey (1948) and others.  Many ingenious hypotheses were invoked to explain archaeological materials, as those overlain by salt marsh at Grassy Island in the Taunton River of Massachusetts (Johnson and Raup 1947) and the famous drowned fish weir at Boylston Street in Boston, investigated by Johnson and others (1942).  In Connecticut, Glynn (1953) inferred the presence of Archaic Amerinds on the basis of artifactual and other materials dredged from the waters of Long Island Sound. In the pre-radiocarbon era, salt marsh and bog borings, palynological analyses, stratigraphic studies of diatomaceous deposits and other refined methods were brought to bear on the complex interrelationships of terrestrial and marine plant successions, sea level changes, and archaeological data. Generally, field methods lay beyond the scope and abilities of most archaeologists, but the work in coastal geomorphology, paleobotany, and related disciplines created much interest.

    Fig. 3A.  A projection of various sea level curves referred to in the report, showing their relations one to the
    other, and their implied sea-stand heights for the last 3000 years, compared to present sea level as a base.

      In a paper expressly addressed to New England archaeologists, Fairbridge (1960) called attention to certain current geological opinions relating to eustatic global sea-level changes and the possible implications for students of prehistory along the Northeast coast.  He offered a tentative sea-level curve (a portion of this is shown in Fig. 3 A), supported by both radiocarbon and geomorphological observations which postulated repeated oscillations in sea stand during the past 6000 years.  Since either the sea may be rising or falling at any given time (volume of water and other factors), or the land likewise (crustal warp and tectonic forces), or both, it is apparent that oscillatory phenomena are complex.  These oscillations are separated into transgressions (sometimes "incursions" or "highs") of the sea and submergences of the land, which result in coastlines moving inland, and regressions (sometimes "recessions" or "lows") of the sea and emergences of' the land, which cause coastlines to retreat seaward. The base datum is current sea level, against which the oscillations are projected (Fig. 3 A).  Fairbridge suggested that sites now lying at indeterminate distances inland might once have been coastal sites during transgressions of the sea in intervals since the onset of Sub-Boreal and subsequent Sub-Atlantic climates in the Northeast.    Other sites, once coastal, may long since have been drowned completely beneath the rising  waters, their contents lost to all save perhaps skin-diving archaeologists of the future.  He cautioned, however, that this data were “first approximations" and in no way to be taken as absolute.