(…from BULLETIN of the Massachusetts Archaeological Society, Vol. 26, Nos. 3 and 4, Apr/July 1965)

 

BITTER ROCK SHELTER: A STRATIFIED CONNECTICUT SITE

Bernard W.  Powell

 

         A test pit in May 1959 confirmed presence of undisturbed aboriginal material at a rock shelter in Norwalk, Fairfield County, Connecticut. Arrangements were therefore made with the landowner to conduct full-scale, controlled excavation. In the Fall of 1961, investigations were begun and completed.  This is the report on the site.

PHYSIOGRAPHY

BITTER ROCK SHELTER is located near 41- 07' 48" N. lat. and 73' 25' 00" W. long., as derived from the Norwalk North quadrangle map (U. S. Geological Survey, 1947).  It lies at the foot of a high cliff on property of Mr. Otto Bitter, in the Winnipauk section of Norwalk (Fig. 15).

         The cliff is in a massive ridge of exposed bedrock.  In front of the shelter is a remnant of glacial terrace formed by the Pleistocene Norwalk River. The nearest spot on the river today is short of a mile away. As the river joins Long Island Sound only 3 miles below here, it places the shelter, generally, within the southern Connecticut Littoral – and in ancient times assured accessibility to tidal flats (apparently confirmed by quantities of marine shells at the shelter).

         Locally, rocks are granitoid geniuses and associated intrusives of the Hartland Formation (Rodgers, Gates, et al, 1956).  Vein quartz abounds, and there are pegmatite dikes nearby.  Ground moraine covers the top of the ridge - but the terrace presents a nice array of ice-worn quartz cobbles and other stones, which must have been welcome to a lithically-oriented people.

AREA HISTORY

         Norwalk, on Long Island Sound, is like much of the surrounding region, an archaeological terra incognito.  I find only one previous report of anthropological investigation in the area (Smith, Wilbur F., 1946), and its conclusions are irrelevant to the problems at Bitter Rock Shelter.

         The immediate area of the site is still wooded and undeveloped (owing no doubt to its steep and rugged nature), and lies well within the settled, older part of town.  That the shelter's true nature was not recognized and exploited by the uninformed in years past is one of those happy turns of fate all too infrequent in crowded southern New England.

         Interestingly, the current landowner, a man in his middle years, has lived in the vicinity all his life.  His
recollections seem pertinent. As a boy, he dug here for “pirate treasure." Digging was with the handsomely, and fortunately did not impair stratigraphy.  No artifacts were found then, though where he played was one of the richest areas of the site with some material within 6 inches of the surface (Fig. 16, No.10).  More recently, he recalls a find by children of 15 to 20 “arrowheads" in one spot. This I tentatively identify as a projectile point cache (Fig. 16, No. 1).  Unfortunately, no one knows where these arrowheads are today, so they cannot be assessed for style, shape, material, or workmanship.  One local resident found a full-grooved, polished greenstone axe about 6" long on top of the ridge.  Several people recall an old man who wintered horses in the shelter 50 to 60 years ago.  And one resident's father, in the animal rendering business, used to pile bones and carcasses  "higher than the house" in the backyard. This is scarcely 200 feet from the shelter, and undoubtedly explains some osseous anomalies found here (Table 4).

          Such was the state of knowledge about the site prior to our excavation. No one truly suspected its secrets, and none of the chance finds over the years had aroused undue local curiosity.  Its days were doubtlessly numbered, however, for this city harbors a multitude of pothunters.
 

EXCAVATION AND METHOD

         A standard 5-foot grid oriented on 330' north magnetic was staked out embracing all the shelter and the slope in front of it (Fig. 16).  East-west axes were designated ON, 5N, 10N, 15N, and 20N, respectively, in terms of their distance in feet north of the datum point.  North-south axes were designated alphabetically.  (Post-pub edit: Square or Unit designations derive from the southeast corner, rather than the more traditional southwest corner).  The datum point was chiseled permanently into the rock face at the back of the shelter wall.  Map control was obtained from existing photogrammetric maps in the City Engineers' office, plus limited plane table work at the site.

         Among other statistics relating to our work at the site, possibly useful in planning future excavations, are the facts that we expended about 200manhours, and moved an estimated 53 tons of dirt and rock (allowing 166.7lbs ft.3 as an acceptable figure for weight ¹, and figuring an average excavation of all squares to 18 inches).

         Initially, all debris, broken glass, and rusted ironmongery from recent times were removed by lightly raking the surface.  It was smooth and showed no evidence of prior disturbance.  To guide us, and to determine vertical profile and stratigraphic relations present, I bisected the entire deposit with a trench (Fig.  16, Trench A) down the 5.2N line.

         It was apparent that the aboriginal material occurred at or just under the surface down to an undetermined depth.  Modern contaminants, with few exceptions, were confined to the 0-3" level.  Potsherds were mainly confined above 18 inches.  With the increasing depth of the deposit, as revealed in Trench A, there appeared automatic suggestion of temporal superposition.  Painstaking examination of the vertical faces of Trench A in all lights and under all conditions did not reveal stratigraphic differentiations fine enough to control digging.  Resort was thus made to the archaeological convention of excavating in 6-inch levels to preserve natural relationships. This control was confirmed in later analysis of the materials, but importantly, it was related to and integrated with certain stratigraphic discriminations even in the field.

         Thus, we distinguished with ease in Trench A, two basic soils: 1) topsoil - the rich, black topmost 6" over all the shelter, and 2) subsoil- a yellow-tan soil devoid of organic staining and penetration ultimately determined to lie everywhere at some depth below the shelter.  Additionally, there were two other soil "types". One might be characterized as a sub-type of the topsoil, since it consisted of topsoil plus heavy admixture of charcoal grains and marine shell fragments.  Most, but not all, artifacts were found in this soil (called midden soil, and analogous with "Aboriginal Material" in Figs. 17 and 18).  The remaining soil was encountered only infrequently and mainly south of the 5N axis.  It was sterile and chocolate brown in color.  I secured good color, texture, and content matches to this minority soil by mechanically mixing portions of topsoil and subsoil. I conclude, in fact, that this soil is just that, and probably formed by activities of subterranean rodents and root intrusions. I do not regard this soil type as invalidating or even seriously challenging the profiles as we derived them, for the disturbance was mild and near the periphery of our plat.  Interestingly, this disturbance probably transpired before the modern era, since few modern intrusives were associated with it.

         The profile of the south face of Trench A is shown in Fig. 17.  Excavation was carried to fully sterile soil.  Where we encountered rocks beyond the ability of three men to remove, we used a 1-inch earth auger to sample beneath them.  Such soundings were always sterile, so we are confident no deep cultural deposits were overlooked.
 

 

 

         As depicted in Fig. 17, a mound of subsoil rose in ancient times near the center portion of our trench.  Apparently, some early occupants in the shelter carried on activities just behind this low rise and in front of Ledge A.  At least a zone highly charged with quartz flakes was interpreted as a stone-working station right in front of Ledge A, and one of the most interesting finds was location in situ of a rough sandstone celt (Fig. 22, A) on a small rock shelf in this ledge.  Archaeology is ordinarily the statistical study of remains left by anonymous populations. Discovery of such a small workshop - almost certainly the abode of a solitary individual - personalizes our discipline and gives added meaning to our privileged view back through the millennia of time.

         An abandoned rodent den near the junction of midden soil and subsoil, contained a glass bottleneck.  This find helped resolve earlier anomalous finds of rusted wrought nails beneath sherds in the upper levels.

         On completion of Trench A, I ran another trench, Trench B, at right angles to it along the B axis (Fig.  16).  The profile derived is shown in (Fig. 18).  In general, similar conditions soil-wise to the finds in Trench A were noted.  There was slight suggestion that certain flat rock slabs were deliberately placed along the junction between midden soil and subsoil.

         Two interesting phenomena discriminated in the east face of this trench were trampled layers or unconformities in the otherwise homogenous midden soil (Fig. 18).  They were thin lines of compacted shell and stone flakes, all with long axes horizontal.  I regard them as ancient surfaces in the shelter, packed by the feet of former inhabitants.

         After these two trenches were completed, all squares shown in Fig. 16 were dug in uniform 6-inch levels.  Space prohibits a minute description, unit by unit, of our finds.  Our field notes on these units, however, are quite complete, and record many interesting finds and observations. It is necessary, however, to generalize somewhat, and to mention only one or two squares in particular.  All of Ledge A was ultimately exposed by stripping (Fig. 16).  Here I found a peculiar bright green stain, perhaps 8" across, near the upper surface.  The stain was comparable to, but not wholly analogous with lichens and stains on nearby rocks. It faded rapidly on exposure to light, and is mentioned since it may have been cuprous oxide from a disintegrated copper artifact.  Other stains noticed on the back wall and roof of the shelter are probably smoke stains from ancient fires.  Schrabisch (1909) reported this during his early researches in northern New Jersey and nearby southern New York.  It occurred to me that such blackened rock walls might preserve micro-layers of soot, which could be interpreted archaeologically.

         Most squares yielded stone artifacts, potsherds, and bone, shell, and stone fragments. To avoid tedious verbal treatment, the ceramic analysis is given in Table 1 and specific ceramic finds are illustrated and identified in Fig. 19.  Projectile points are summarized in Table 2, and illustrated and identified in Fig. 20.  Artifacts and Trait Distributions are given in Table 3; faunal remains appear in Table 4.  In the Discussion section, I will integrate and interpret these finds.  Square 10N/B proved the richest, and contained the deepest deposit (minus 58").  The appearance of this square on removal of the 0-6" level is shown in Fig. 21.  The caption explains various points identified on the photograph.

         Nearly 5 feet beneath the surface in 10N/B, and intrusive about 1 foot into the subsoil, we found our deepest cultural material.  This was a crude, coarsely chipped, lunate quartz knife (Fig. 22, F), and a bifacially-flaked crude chopper (Fig. 22, E).  These were associated with badly disintegrated charcoal grains and calcined bone flecks.  One burned and semi-calcined bone fragment was recovered, and expert opinion ² concurs that it is probably from a human cranium.

 

         Another interesting square was 5N/B at the intersection of our trenches. Mention was made of the chipping station found along the front of Ledge A in this, and the adjoining square, 5N/A.  Notable here were many slabs of schistose rock, at times almost interlocked like shingles. At first we thought they might be artificial, perhaps covering a burial or some other feature.  But we encountered this phenomenon repeatedly in the shelter, and concluded the slabs were simply frost and weather spalls from the roof.  It is interesting, however, that slabs in this quantity and condition are not currently forming at the surface.  The occurrence of cultural refuse between and around many of these rocks prompted their interpretation as artificial flooring, deliberately laid to promote drainage. However, we cannot demonstrate this.  Some of these rocks, though, showed weathering and rounding which seems to imply importation to the shelter rather than spalling in situ.
 
         While most cultural materials were found either in the topsoil, or the midden soil, it is notable that some were not.  These included the materials in the deep level of 10N/B, as just mentioned, and a few other finds from various places in the shelter.  These other finds were all made either just at junction of midden soil and/or topsoil with subsoil (that is, lying on the top of the subsoil), or else only a short way down into the subsoil (maximum extent, for one or two pieces: about 6").  Beneath this, the subsoil proved sterile in every deep sounding, which included at least two test pits, and numerous borings.  Very little material was thus so recovered.  The field notes reflect that only the projectile points (L) and (R) in Fig. 20 were definitely in the subsoil.  The only other identifiable artifacts from the subsoil were the lunate quartz knife and the flaked chopper mentioned previously for the deep levels in 10N/B, and they may have been intrusive as components in a cremation pit whose outlines were not discriminated.  Perhaps less than a dozen white quartz flakes and amorphous chipped forms came from the junction and upper six inches of the subsoil. There were also some black flint, red chert, and bluish silicate flakes that may represent exotic stones.  The specific loci on our grid for these subsoil finds are 0.2N/B, 10.5N/.5A, and 10N/.5C.  These subsoil occurrences were concentrated at these spots, and not randomly encountered.

 

         Several natural crevices and formations in the rocks lay just beyond our grid.  All were investigated, and may be identified from Fig. 16. Perhaps one other special area within the shelter deserves mention; this is the most inward recess (Fig. 16, No. 10).  Technically, it too, lay beyond our grid, but it was systematically investigated, and its contents where possible related to the stratigraphy and master control of the grid. It was a difficult spot to excavate, and in many places I bad to lie prone and dig with a long-handled shovel.  Naturally, control was loose, but the inventory shows the interesting notched slate pendant (Fig. 22,D) came from here: I suspect it was near the subsoil junction, but this is uncertain.  Two or three unmodified but rounded cobbles, were found deep in this recess.  How they came there, I do not know.  Perhaps they are evidence from glacial times, when topsoil had not formed; weathering had not dislodged spalls from the shelter roof and walls; and ice-rafting or other phenomena associated with the glacial terrace in front introduced these cobbles into this recess.

Continued