Archaeology in Newfoundland and Labrador 1996
Edited by K. Nelmes
The Dorset Occupation of Dildo Island: Preliminary Field Report, 1996
Sylvie LeBlanc
Dildo Island Project
The Dildo island Archaeological Project is a multi-seasonal research project which began in 1995 under the Baccalieu Trail Heritage Corporation. Systematic survey work conducted in the fall of 1995 (Gilbert 1996) yielded evidence of an aboriginal occupation of the island by Recent Indian, Beothuk and Dorset people. In the historic period the island was visited by John Guy in 1612 (Gilbert 1990) and a cod hatchery was in operation in the late 1800s.
Previous archaeological investigations by Gilbert (1996), Marshall (1990) and Locke in the 1970s all pointed to the unique character of the Dorset material uncovered on Dildo Island. The investigation of this unique nature of the Dorset occupation became the primary focus of the 1996 field season.
The overall objective of the Dildo Island Project is to document and define the Dorset occupation on Dildo Island. The ultimate goal of the research is to interpret this unique Dorset site within the broader context of Newfoundland and Labrador, and Eastern Arctic Palaeo-Eskimo prehistory.
Figure 1. Bottom of Trinity Bay.
Site Description
Dildo Island is a small island located at the bottom of Trinity Bay, on the east coast of Newfoundland (Figure 1). The island is about twenty minutes by boat from the community of Dildo. Two other islands – Eagle to the south and Rosses Island to the north – flank Dildo Island. The island has an elongated shape and measures about 700 metres north/south and 150 metres east/west at its widest point. The northern portion of the island is covered with a coniferous forest and is not easily accessible because of high cliffs. This high portion of the island slopes gently to a grassy meadow which covers the entire southern part of the island. In this southern portion Barry’s Cove, with its cobble beach provides the best access to the island. A small stream of water runs through the island in a north to south direction.
Excavations and Procedures
While the whole island could be considered as one large Dorset site, the 1995 extensive survey clearly demonstrated that Dorset lithic material was more densely concentrated in four areas: A, B, C and D (Figure 2). Of these four areas, Area B, located on the southern portion of the island near Barry’s Cove, produced the greatest density of Dorset material (Gilbert 1996; Locke collection [1]; Marshall 1990). The 1996 field season concentrated on Area B. In order to further define the area to be excavated, four additional test pits were dug. On that basis, two distinct areas, which we called zones, were opened in Area B. Zone 1 is located on the northern portion of the grassy meadow. Zone 2 is located to the southwest, near Barry’s Cove. The grid was established using the 1995 survey grid.
Figure 2. Dildo Island: Areas A, B, C and D.
In a period of six weeks, a total of twenty square metres was excavated, sixteen in Zone 1 and four in Zone 2. The soil was removed by trowelling and the method used in excavation emphasized natural and cultural stratigraphy. Each stratum was exposed over the entire excavation area, mapped, photographed and then removed. Individual artifacts were recorded according to their stratigraphic and horizontal provenience. Flakes were recorded by natural levels and by 50 cm quadrants (A, B, C, D) within one metre units.
All squares were named after the southwest corner and elevations were taken from the southeast corner of each unit. All the backdirt was dry-sifted through 1/4 inch mesh.
For the purpose of this report Zones 1 and 2 will be described separately.
Zone 1
DESCRIPTION
Zone 1 is a grassy terrace about seven metres above present sea level in the northern portion of Area B. The terrace measures approximately 20 metres wide by 24 metres long. This terrace slopes gently to the west toward Barry’s Cove and is backed to the east by a steep cliff.
EXCAVATION
In addition to the four initial test pits, a total of 16 m2 was partially excavated in Zone 1. The cultural level (Level 3) was exposed for the entire excavation area but only partially excavated. Excavation of Level 3 will be undertaken only when its horizontal limits are known. Therefore this zone remains essentially intact for future excavation.
STRATIGRAPHY
One stratigraphic profile was recorded for Zone 1: the north-south profile at W152 (Figure 3).
Level 1 is a sod layer varying in thickness from 2 to 5 cm. Few artifacts were found in this level and most of them were found at the base of the sod, close to the next level (Level 1A). Level 1A is in fact the base of the sod. It is a mixture of roots, sand and humus and is brownish-grey in colour. A large number of artifacts and fire-cracked rocks were found scattered in this level. Thickness varied from 6 to 16 cm. Some of the features (1, 2, 3, 4), as yet ill-defined, first appeared in this level. Level 2 is a peat layer. The soil in this level is very light and easy to dig; it varies in colour from light brown to a darker brown. Thickness varies from 4 to 10 cm, with the maximum thickness over the features. This peat layer is not found everywhere in the excavation unit but seems to be only associated with, and on top of, the cultural level (Level 3).
Figure. 3. Dildo Island, House 1 north-south profile at W152.
The different features are more clearly defined in this level and most of the artifacts and flakes are found in association with these features. There are fewer artifacts in level 2 than in any other layer and most of those are found at the bottom of the level, close to Level 3. Level 3 is the cultural level. The soil varies from a light, peat-like, organically-stained dark brown soil to a black greasy soil around hearth features. Occasional small gravel lenses and charcoal fragments occur in association with the black greasy soil. In this level the artifacts are clearly concentrated around the features. The total thickness of Level 3 is not yet known as this level is only partially exposed. Portions partially excavated in 1996 range up to 9 cm thick. Excavation of Level 3 will be undertaken only when its horizontal limits are known. Level 4 is a sterile sand and gravel level exposed only in the southern portion of the excavation unit, outside of what seems to be the limit of the dwelling.
The complete stratigraphic sequence (Sod, 1A, 2, 3 and 4) is only found where the cultural level (Level 3) exists, that is, within the house limits. Outside the house feature there is barely any stratigraphy except for Sod, Level 1A and Level 4 or bedrock.
House 1
House 1 is still difficult to interpret as only 16 m2 were partially exposed in 1996 (Figure 4). Its definite size and shape is not known. However, a rock outcrop to the northwest and an alignment of large rocks in the southeast portion of the excavated area could possibly mark the limit of the dwelling. To the north, and in line with the large rock alignment is a low ridge of scattered fire-cracked rocks. If these two formations are connected, they may mark the eastern limit of the dwelling. On the other hand, the fire-cracked rock formation may also be a dump. West of, and running along the fire-cracked rock ridge is a shallow trench about 50 cm wide and 7 cm deep.
Figure 4. Dildo Island: House 1 Plan.
The entrance passage of the dwelling, if one existed, is not yet discernible. A concentration of stone slabs (Feature 7) to the southeast could represent an entrance, but its position in relation to the axial feature (not in line) is puzzling. These slabs could also represent a platform area.
A mid-passage feature edged on the west by a line of stone slabs and boulders appears to be running along a NE/SW axis. The eastern edge of the mid-passage feature is not clear, but a long strip relatively clear of rocks might mark its limit. The feature is approximately 1.5 m wide and its full length could not be determined. Because of the incomplete nature of this year’s excavation the mid-passage pavement was not uncovered; future excavation will probably find it lying under the rich organic deposit. A series of hearth features (Features 2, 4 & 5) and one pit feature (Feature 3) were found along the mid-passage structure. The mid-passage feature lies about 15 cm below surface.
Outside and west of the axial feature is a mound of fire-cracked rocks (Feature 6). The north wall profile suggests that the house extends north of the excavated area. Future excavations will allow us to define the house more clearly.
FEATURES
Again, because of the incomplete nature of the 1996 excavation, the complete description of all features is not possible.
Feature 1: Feature 1 is a concentration of rocks, some of which are fire-cracked, in the northwest quadrant of N286-W154. This feature first appeared in Level 1 and lies directly on the bedrock. It was first believed to be a hearth feature but since there was no evidence of fire this idea was dismissed. Soapstone fragments, flakes and one endblade preform were found in association.
Feature 2: Feature 2 is a hearth in N285-W153/152. The feature first appeared in Level 1A as a rich, black organic stain with an upright slab near the centre. As excavation proceeded into the next levels (Levels 2 and 3), fire-cracked rocks and wood charcoal appeared. This feature runs through the mid-passage feature. It may be a part of Feature 5.
Feature 3: Feature 3 is a pit feature in the southeast quadrant of N284-W154. The pit is lined with rocks. Dimensions are roughly 55 x 40 cm at the opening, and the pit is narrower toward the bottom and about 17 cm deep. No evidence of fire was found, and the function remains unknown.
Feature 4: Feature 4 is a fire-cracked rock concentration in the southwest quadrant of N285-W154. This feature was not excavated but only exposed. It was first noticed in Level 1A as a circular black, organic stain with fire-cracked rocks in the centre. This feature is believed to be a hearth, as charcoal fragments are also associated with it.
Feature 5: Feature 5, located in N286-W152/153, is a hearth feature filled with charcoal, ashes and fire-cracked rocks. Large flat beach cobbles form the base and are flanked on each side by slabs encrusted with burned fat. This feature probably connects with Feature 2.
Feature 6: This is a mound of fire-cracked rocks located in N284/285-W172. It measures about 110 cm north-south by 65 cm east-west. Until it is further excavated and its position in relation to the house determined, its function will remain unknown.
Feature 7: Feature 7 is a concentration of horizontal slabs in N283-W152. Its relation to the axial feature and other elements of the structure remains to be determined.
ARTIFACT DESCRIPTION
Zone 1 yielded a total of 377 lithic artifacts, 7691 flakes and 3 fragments of European material. Not included in this count are the artifacts uncovered in the four test pits. These will be listed separately at the end of this section. For description purposes, artifacts from the different levels were grouped. Provenience of individual artifacts can be obtained from field tags and provincial artifact record sheets.
HARPOON ENDBLADES n = 62
Zone 1 produced 62 endblades and endblade fragments. Only 13 specimens are complete enough to allow measurements. All specimens are made of white or brownish-white chert except for two specimens, one of which is made of a purple rhyolite and the other of a coarser sandstone-like stone. Two major categories of endblades can be identified: (1) an elongated triangular style and (2) a shorter variety with markedly concave bases.
(1) ELONGATED TRIANGULAR ENDBLADES n = 39
Three sub-categories of elongated triangular endblades can be identified based upon the degree (or absence) of surface grinding: totally ground, partially ground and chipped endblades.
Totally ground (n = 12): Twelve specimens made of white or brownish-white chert are ground on both faces. They are long, narrow, triangular and unmatched. The only two complete specimens were measured (length 31 mm, thickness 3.8 mm, width 11.5 mm; and length 36.7 mm, thickness 3.0 mm, width 13.6 mm). The lateral edges are symmetrical, either straight or slightly convex. The base is either straight or slightly concave. Most lateral edges are worn but fine lateral serrations remain on four specimens.
The typical form has a dorsal surface comprised of three facets: basal grinding forms a triangular facet that extends about one third of the total length of the artifact; grinding also produced two lateral facets divided by a longitudinal median ridge running from the apex of the basal triangle to the distal end of the endblade. Ventral surfaces usually exhibit flat grinding (9 specimens) or have medial ridges (2 specimens), remnants of a tip fluting technique.
One specimen is identical on both dorsal and ventral surfaces with the triangular basal facet and lateral facets on both surfaces. This specimen is biconvex in cross-section.
Partially ground (n = 20): The endblades in this category are not fully ground but are otherwise morphologically similar to the totally ground examples. The degree of grinding may vary from almost totally ground to a very small ground area. All specimens but one, made of purple rhyolite, are made of white or brownish-white chert. Three complete specimens measure between 48.2 mm and 29.8 mm in length, 17.4 mm to 13 mm in maximum width, and 4.6 to 3.8 mm in maximum thickness.
All twenty endblades show partial or total grinding on the ventral face. Of these, eleven are ground flat (plano-convex cross-section), two exhibit a medial ridge, and seven retain a biconvex cross-section with only initial grinding.
The dorsal surface is chipped on all specimens and only three show evidence of grinding: on two specimens grinding produced a basal triangular facet; the other specimen was ground unilaterally along the medial ridge.
Chipped endblades (n = 7): Seven specimens show bifacial flaking. These specimens share the same shape with the previous types; they are long, triangular and narrow. Only three specimens are complete. Length ranges between 48.8 mm and 37.4 mm; maximum width ranges between 19.4 mm and 16.6 mm; and maximum thickness ranges between 4.4 mm and 4.0 mm. Five specimens are made of a white chert; one of a greenish-beige chert and one of a coarse sandstone-like stone. Lateral edges are symmetrical, either straight or slightly convex. The base is usually straight or slightly concave but one specimen has a deeper concave base. This latter endblade also exhibits finer flaking than the other specimens, which are generally crudely flaked.
The categorization of endblades into three groups (totally ground, partially ground and chipped) supposes more or less that each group corresponds to a different manufacturing stage along a trajectory toward a finished and polished triangular endblade. For instance, a triangular preform was first prepared by flaking. It was then further reduced by tip fluting. The ventral was was then ground flat and the dorsal surface ground to produce the three facets characteristic of completed endblades. Lastly, fine serrations were produced along the lateral edges.
SQUARE DISTAL ENDS n = 5
These are believed to be endblades in process with a distal end prepared for tip fluting. The dorsal surface is flaked and convex. One specimen is bifacially flaked. The ventral surface is flat on two specimens – one possibly showing evidence of grinding – and two have median ridges.
DISTAL ENDS n = 10
Ten distal ends of white or brownish-white chert are too small to be assigned to any of the previous categories.
(2) SHORT, TRIANGULAR ENDBLADE WITH CONCAVE BASE n = 8
This category of endblades consists of very small triangular specimens with concave bases. If it were not for their small size and the raw material used (white chert) these specimens would resemble the Dorset endblades from the Newfoundland west coast collections. Length varies from 29.8 mm to 17.2 mm, width from 10.9 mm to 8.7 mm and thickness from 4.2 mm to 2.9 mm. Lateral edges are symmetrical, either straight or slightly convex. The basal edge has been thinned to form a marked concavity. The dorsal face is chipped and convex and the ventral surface is either flat or may have a medial ridge from tip fluting. These specimens are chipped with no evidence of grinding. The small size of these endblades might indicate some functional distinction between them and the larger examples described above.
MULTIPLE NOTCHED SLATE ENDBLADE n = 1
Because of its fragmented state, the shape of this artifact cannot be reconstructed with any certainty. The remaining basal corner fragment is ground and bears two sharply defined notches; similar notches were probably present on the opposite edge. This specimen is probably a fragment of a polished slate endblade or lance similar to those illustrated by Robbins (1985:186) from Stock Cove.
SPEAR FRAGMENT n = 1
This appears to be the median section of a polished slate spear or lance. Neither end is present. The cross-section is not unlike the broad “pie wedge” cross-section of Maritime Archaic ground slate bayonets or lances. One surface retains three poorly-defined facets; the opposite face has a pronounced medial ridge and a roughly ground triangular facet not unlike the basal facets on harpoon endblades. Lateral edges are parallel and one edge exhibits a single side notch. A gouged hole perforates the object at about midpoint. Full length is not available but the fragment measures 74 mm in length, 28.1 mm in width, and 6.6 mm in maximum thickness. This object most closely resembles a Maritime Archaic spear or bayonet reworked by Dorset people.
BIFACES n = 6
Six biface fragments were uncovered in Zone 1. All specimens are too fragmentary to allow reconstruction of the original form. All specimens are either partially or totally flaked and one specimen has surface grinding on one face.
Two distal fragments, one medial fragment and one basal fragment are made of white chert. Of these, three nicely flaked specimens are fairly thin (3.8, 4.2, and 4.8 mm); the basal fragment has a double set of shallow side-notches.
Another distal fragment is made of a dark brown chert. The specimen has a sharply pointed distal end, is flaked bilaterally and exhibits grinding on one face.
One medial fragment is made of a mottled grey chert. This specimen is finely flaked on one face and is rather crudely worked on the opposite face with cortex remaining along one edge. It seems to represent a symmetrical biface with slightly convex lateral edges.
UNIFACIAL KNIFE n = 4
Four unifacially flaked fragments of brownish-white chert could be interpreted as knives. One speciment has asymmetrical curved lateral edges and is 4.8 mm thick and 24 mm wide. Three other specimens are fairly thin, with a thickness averaging 4.1 mm. Of these three, two are basal fragments and one is almost complete except for the distal end. All three specimens have straight or slightly convex lateral edges. Base shape varies from straight to slightly rounded.
STEMMED POINT n = 1
One specimen of white chert is a small stemmed point. The point is unifacial, made on a flat flake with edge retouch on the dorsal surface. Length is incomplete but the specimen is 25.6 mm long, 9.8 mm wide, and 2.8 mm thick. Function remains unknown. It could have been hafted and used as a projectile point or could possibly have been used as an awl.
SCRAPERS n = 27
Nineteen endscrapers are fashioned from quartz crystal. These endscrapers are very small (length < 17 mm) with a single working edge and roughly triangular in shape. The working edges are steep and slightly rounded, and flaking is usually done only on the dorsal face. An unusually large endscraper (length = 26.8 mm) is also made of quartz crystal. One small endscraper (length = 12.2 mm) is made of a mottled grey chert. Five specimens, varying in size (length = 48 to 19 mm) and shape are fashioned on white and brownish-white chert flakes. Of these five, three specimens have multiple working edges and the remainder have a single working edge. Like the quartz crystal endscrapers, most of the flaking is done on the dorsal face. One endscraper is made of what appears to be a very fine grained pink quartzite.
MICROBLADES n = 60
The total sample of microblades from Zone 1 is 60. Of these, 44 are of quartz crystal, 14 of white or brownish-white chert, one is of high quality pale green chert, one is of dark grey high quality chert. Fifteen specimens are complete, 14 are proximal fragments, 17 medial fragments, 5 distal fragments, and 9 specimens are ridge or primary flakes. One quartz crystal and the dark grey specimen show lateral flaking on one edge. Quartz crystal microblades are very small, ranging from 19.3 mm to 14.6 mm in length, 2 mm to 1 mm in thickness, and 6.1 mm to 3.8 mm in width.
It is suggested that the small size of quartz crystal microblades is conditioned by the nature of the material used (McGhee 1981:70): the natural structure of quartz crystal is angular and highly fractured leaving few possibilities for the production of long and large blades.
Chert specimens are generally larger. Complete measurements are not available as no chert specimens were found complete, but width varies from 12.3 to 5.7 mm.
QUARTZ CRYSTAL CORES AND DETRITUS n = 60
Quartz crystal cores come in a variety of forms. Three specimens retain the natural hexagonal cross-section of quartz crystal. 48 specimens retain at least one of the natural hexagonal facets and the remainder are core detritus. Of these, 29 show evidence of microblade removal, and 14 seem to have been worked toward the making of endscrapers. The latter specimens are roughly triangular in shape. They all have a steep edge on at least one side and some specimens exhibit fine flaking on what could possibly become the working edge of an endscraper.
TIP FLUTE SPALLS n = 49
The Zone 1 collection includes 48 tip flute spalls. Forty-four specimens are made of white chert; one of grey chert; one of blue-grey chert and one of purple rhyolite. One endblade found in Zone 1 is made of this same purple rhyolite. Tip flute spalls are easily recognizable by their elongated triangular shape and a striking platform indicating a tip fluting technique. Tip flute spalls also have two distinct lateral edges, one being thicker than the other. The spalls are usually worked on the dorsal side and unworked on the ventral side.
BLADE-LIKE FLAKES n = 36
Blade-like flakes have been defined as parallel-sided flakes or linear flakes (Linnamae 1975:131; Robbins 1985:73). These flakes are distinguished from true blades as they have irregular parallel sides but often lack arrises. All specimens but one, which is made of a dark grey rhyolite-like stone, are made of white, brownish-white chert. One chert specimen exhibits fine lateral retouch on one edge. The function of these blade-like flakes is unknown. Their irregularity appears to make them unsuitable to be used in the same manner as blades. They could possibly find their place somewhere along a manufacturing sequence, for example in the initial preparation stage of the tip fluting of harpoon endblades.
ENDBLADE BASAL THINNING FLAKES n = 10
All 10 specimens are made of white or brownish-white chert. Endblade basal thinning flakes are parallel-sided flakes removed from the base of the dorsal surface of harpoon endblades. They are recognizable by the surface flaking on the dorsal face, particularly along the edges, and the absence of retouch on the ventral face. Striking platform and/or a bulb of percussion is often observed at what was the base of the endblade. They may have been removed by pressure flaking as part of the technique employed in creating the basal facet of the elongated ground endblades.
Endblade basal thinning flakes have gone unreported from Dorset sites elsewhere; either they went unrecognized or they are specific to Trinity Bay Dorset technology.
PREFORMS n = 20
Twenty preforms were recovered from Zone 1. Eighteen are made of white or brownish-white chert, one is made of pale grey quartzite-like stone, one of a dense mottled grey chert and one of a dense brown chert.
Preforms are objects that are far enough along in the manufacturing process that the intended final shape of the implement can be recognized.
Five specimens are clearly in the process of being reduced by tip fluting. These specimens are bifacially worked and were intended to become elongated triangular endblades. Length varies from 41.9 to 65.5 mm, width from 18.2 to 21.1 mm and thickness from 5 to 10 mm.
Four shorter and broader specimens appear to be preforms for the short, triangular, concave-based endblades. Of these, two are made on flakes and have unifacial retouch on the dorsal surface. The two other specimens are bifacial and have a biconvex cross-section. Length varies from 23.3 to 36.4 mm, width from 13.7 to 17 mm and thickness, 3.7 to 4 mm.
Three distal fragments are probably endblade preforms. Six basal fragments and one medial fragment were probably intended to become unifacial or bifacial knives.
Another specimen is short and broad, has a plano-convex cross-section and heavy dorsal flaking. The ventral surface is unretouched. This preform is fairly thick with a maximum of thickness toward the base (length 33 mm, width 17 mm, thickness 6.3 mm). For some reason, perhaps its distinctive grey chert raw material and thick cross section, this specimen seems out of place in a Dorset deposit and might be attributed to another cultural entity.
BLANKS n = 11
Eleven blanks, all of white or brownish-white chert were uncovered in Zone 1. Blanks come earlier than preforms in the manufacturing process and final tool shape is not discernable at this early stage. Blanks are usually large, angular and fairly thick. One or both faces are crudely flaked and the lateral edges are rarely worked. All specimens either have an ovoid or an elongated shape. Large scars indicate direct hard hammer percussion and pressure flaking is not used at this stage of the manufacturing process.
ABRADERS n = 2
Two incomplete sandstone abraders were found. Both have an elongated shape, a quadrilateral cross-section and show evidence of smoothing on at least three surfaces. Each specimen has one ground, bevelled end and each is broken at the opposite end. One abrader measures 75.3 mm in length, 11.6 mm in thickness, and 33.8 mm in width. The other is 103.9 mm long, 16.2 mm thick, and 26.7 mm wide.
RETOUCHED FLAKES n = 18
Eighteen retouched flakes found in Zone 1 can be classified as follows: Sixteen flakes measuring between 52 and 29 mm in maximum dimension are made of white or brownish-white chert. Of these sixteen, nine have irregular parallel edges and seven have asymmetrical curved edges. Four specimens are bifacially retouched on one edge, eight are unifacially retouched along one edge (four on the ventral surface; four on the dorsal), two specimens are unifacially retouched on two edges (one on the ventral surface, the other on the dorsal surface). One specimen is retouched along one edge of the dorsal face and on the opposite edge on the ventral face. Only one specimen is unifacially retouched along all edges. Some specimens could certainly have been used as knives.
Another large flake (53 mm max. dimension) of coarse-grained light grey chert, bifacially retouched along one edge, could also have been used as a knife. One brownish-white chert flake (29 mm max. dimension) is polished on one surface.
SOAPSTONE n = 6
Six soapstone fragments were collected in Zone 1. All fragments are crude, rough and encrusted with burned residues. Five of them could possibly be part of the same vessel as they were found together in deposition. The fragmentary nature of the soapstone pieces makes it difficult to reconstruct vessel shape, but one fragment showing a flat base and a straight outsloping side wall suggests a rectangular shape rather than a rounded one. The base thickness is 14 mm, side wall thickness is 12.5 mm. The side wall meets with the base at an angle of about 35 degrees.
UNKNOWN n = 4
One specimen made of white chert is probably of Dorset origin, but its unusual shape does not allow classification in any of the previous categories. The specimen is unifacially flaked and has a square end. Another specimen, made of a coarse sandstone-like stone resembles the previous specimen. Two endblade basal fragments of unidentified cultural origin were also found in Zone 1. One is made of a pale blue quartz-like stone. The latter specimen is bifacial and crudely made. It has a straight base and has a single notch on one of its lateral edges. Some lateral nibbling on the opposite lateral edge suggests that it was meant to be double-notched. The other specimen is a basal fragment made of white chert. It has a straight base and two opposite notches.
EUROPEAN MATERIAL n = 3
1 French gun flint;
1 blue glass rim of possible Mason jar from the second half of the nineteenth century or later; and
1 brick fragment.
DEBITAGE
Some 7691 flakes were recovered in Zone 1. By far the most common raw material used was the white or brownish-white chert. Small quantities of quartz crystal and purple rhyolite were also found. Very few flakes (less than 20) were of pale grey or blue-grey chert, and only one flake of blue-grey chert can be identified as Cow Head chert.
TEST PIT 230: Nineteen flakes
TEST PIT 231: Two blade-like flakes, 2 preforms, 2 microblades, 84 flakes, 1 pottery sherd (coarse earthenware)
TEST PIT 232: Two quartz crystal cores, 1 chipped endblade, 1 blade-like flake, 2 preforms, 140 flakes
TEST PIT 233: One biface fragment, 1 quartz crystal microblade, 50 flakes
SOIL, FAT AND CHARCOAL SAMPLES
One soil sample, eleven fat samples, and eight wood charcoal samples were collected in Zone 1.
Zone 2
DESCRIPTION
Zone 2 is located 14 m south and 20 m west of Zone 1, on the east side of Barry’s Cove. This zone is about a metre away from the actual cobble beach, at an elevation between four and five metres above present sea level. The zone is covered in grass and flanked to the west by a rock outcrop sloping into the sea. Only four square metres were excavated in Zone 2.
STRATIGRAPHY
As the excavation proceeded in Zone 2, the different levels were difficult to interpret at at least three test pits, all at different depths, had been previously dug in the excavation unit. One test pit was the result of 1995 survey and the two other were either made by local people and/or by Marshall and Locke. Thus, the different wall profiles were our only guide to distinguish the different strata.
One stratigraphic profile was recorded for Zone 2: the east-west profile at N271.
Level 1 is sod layer varying in thickness from 2 to 5 cm. Level 1A is a mixture of roots, sand and humus, brownish-grey in colour. Thickness varies from 3 to 9 cm. Many artifacts were uncovered in this level and Feature 1 appeared in this level. Level 2 is a peat layer varying in colour from light brown to dark brown. This layer varies in thickness from 3 to 14 cm. No artifacts were found in this level. Level 2A is a layer of light brown sandy soil containing small pebbles. This level is only found in the western section (N269-W176) of the excavated unit. Depth for this level is not available as it was only exposed. Excavation will proceed only when the horizontal limits of this level are known. This layer is difficult to interpret but it could mark the western edge of the central pavement area. It could also be associated with a box-hearth feature (Feature 2) as two rows of parallel upright slabs enclose this sand and pebble stratum. Level 3 is the cultural level. It is a mix of a grey ashy sod and a rich, black organic soil. This level extends from the base of the peat level to the top of the flagstone pavement. Thickness varies from 15 to 27 cm. The majority of the artifacts were found in this level. Level 3A is a darker black stratum at the base of Level 3. This level went unrecognized during excavation, but as the north profile was drying, a darker black lens appeared. This stratum is not found everywhere in the unit but only in the northern portion and seems to mark the limit of Feature 3. In places, the thickness of this level is greater than 15 cm.
Figure 5. Dildo Island: House 2 Plan.
House 2 (Figure 5)
Partial excavation of House 2 allowed the recovery of a well-defined mid-passage flagstone pavement about 45 cm below surface. Total width and length of the pavement area is not known at this early stage of the excavation. The definite shape of the house is also not possible to determine, but a grassy bulge interrupted by a central break strongly suggests a bilobate structure. This break is in line with the mid-passage pavement and could easily be interpreted as an entrance. If this is the case, the house has a north/south orientation, with the entrance facing north toward Barry’s Cove. Once again, the size of the house is difficult to determine but from the bulge marking the possible entrance it could easily be estimated to be at least 4 to 5 m wide.
A description of features associated with House 2 follows.
FEATURES
Feature 1: Feature 1 is a formation of fragmented thin slabs in the southern part of the excavated area. These slabs seem to have collapsed and were probably part of a larger feature that continues to the south. Next year’s excavation will allow complete description of the feature. A great concentration of artifacts and flakes was associated with this feature. Among them were a polished slate fragment and one slab fragment, partially encrusted with fat, with a gouged hole in it. At this point the function of this feature remains unknown.
Feature 2: In the western portion of the mid-passage pavement two parallel rows of upright slabs could indicate some kind of hearth feature, although excavation is in its early stages and no trace of in situ burning has yet been found. Level 2A, a lens of sand and pebbles is found only in association with this feature.
Feature 3: A fairly large piece of burned wood (about 7 cm in diameter) was found emerging from under the pavement in the northern portion of the mid-passage feature. The extremely rich black soil and the profile in this area suggest a hearth feature. On the other hand, this piece of wood could have been once part of a pole supporting the roof structure. A sample of this burned wood will be submitted for radiocarbon dating.
ARTIFACT DESCRIPTION
Despite the limited size of the excavation (4 m2), over 219 individual artifacts and 3536 flakes were catalogued from Zone 2. For descriptive purposes, the categories used in Zone 1 will also be used here. To avoid repetition, full description will only be provided when necessary.
HARPOON ENDBLADES n = 57
A total of 57 endblades and endblade fragments were recovered from Zone 2.
(1) ELONGATED TRIANGULAR ENDBLADES n = 52
Totally ground (n = 28): Twenty-eight specimens are bifacially ground. All specimens, except two made of a green chert, are made of white chert. These specimens are long, narrow and unnotched. Length of complete specimens ranges between 47.7 and 34.9 mm; maximum width ranges between 14.2 and 8.6 mm; maximum thickness from 3.3 to 3.0 mm. The lateral edges are symmetrical, either straight or slightly convex. Lateral edges are damaged on most specimens but 17 examples exhibit fine lateral serration. The base is either straight or slightly concave. The dorsal surface is similar on all specimens with grinding forming a triangular basal facet and two lateral facets.
Ten specimens are bifacially symmetrical as they received the same grinding treatment on both faces. These are biconvex in cross-section. Eleven specimens have the ventral surface ground flat and therefore have a plano-convex cross-section. The fragmented nature of the seven remaining fragments (6 distal, 1 basal) does not allow the description of either ventral or dorsal treatment.
Partially ground (n = 19):
Nineteen specimens show evidence of partial grinding. Fifteen are made of white chert and one is of a mottled white and blue-grey chert, which seems to be only a variant of the white chert. One is of purple rhyolite and the remaining two are of dark and pale green cherts, respectively. Six complete specimens allow measurements that range between 32.4 and 36.4 mm in length, 11 to 16 mm in width, and 3.4 to 3.8 mm in thickness.
Only five specimens show evidence of bifacial grinding. One specimen is bifacilly symmetrical with basal triangular grinding; nine specimens are ground flat on the ventral surface and have a plano-convex cross-section. Seven specimens still retain a biconvex cross-section with only initial grinding. Two distal fragments are too small to be discussed.
Chipped endblades (n = 5): Five specimens of white chert show bifacial flaking. Two complete specimens range in length from 33.1 to 36.3 mm, in maximum thickness from 4.0 to 5.5 mm, in width from 16.7 to 20.1 mm. All specimens have either a straight or slightly concave base. Lateral edges are symmetrical, either straight or slightly convex. Three examples exhibit a biconvex cross-section, while the remaining two specimens have a plano-convex cross-section with the ventral surface either chipped flat or showing a median ridge.
SQUARE DISTAL END n = 1
One square distal end fragment of white chert is chipped on the dorsal surface and exhibits a median ridge on the ventral surface.
DISTAL END n = 1
One white chert chipped distal end fragment is too small to be assigned to any of the previous categories.
(2) SHORT, TRIANGULAR ENDBLADE WITH CONCAVE BASE n = 3
Three small triangular endblades with deep concave bases were found in Zone 2. All are made of white chert. These endblades are chipped on the dorsal surface; the ventral surface is flat on two specimens while the third specimen has a median ridge resulting from tip fluting.
BIFACES n = 4
Two distal fragments made of white chert are asymmetrical with one lateral edge slightly curved. Another distal fragment made of a greenish-beige chert is symmetrical and is partially ground on both surfaces. The last specimen is a crudely flaked medial fragment made of a coarse-grained type of stone. This latter specimen could easily be of Indian origin.
UNIFACES n = 3
Three fairly thin distal fragments are chipped unifacially. Contrary to the bifaces described above, which all had pointed distal ends, all of the unifacial specimens have a rounded distal end. One of the specimens exhibits partial grinding on the dorsal surface.
SCRAPERS n = 17
Seventeen small quartz crystal endscrapers were uncovered in Zone 2. These were identical in all attributes to those from Zone 1.
MICROBLADES n = 42
Forty small quartz crystal microblades were recovered in Zone 2. One of these has been retouched, possibly for hafting, and another is covered with an unidentified red substance (red ochre, blood?). Two larger microblades are made of white chert.
QUARTZ CRYSTAL CORES AND DETRITUS n = 50
Quartz crystal cores come in a variety of forms. Thirty-four specimens still retain at least one of the natural hexagonal facets of quartz crystal. Eight specimens show evidence of microblade removal and six specimens seem to have been worked toward the making of endscrapers.
BURIN-LIKE TOOL n = 1
One specimen could possibly be identified as a burin-like tool. It is made of white chert and has the trapezoidal shape characteristic of Palaeo-Eskimo burin-like tools. A spall has probably been removed from the tip along one lateral edge.
ADZE n = 1
The bit fragment of a whitish chert adze is polished over its entire surface. The bit is plano-convex in cross section and has two small sharpening facets on the ventral (convex) surface. It is undoubtedly a broken example of the intact adzes with large multiple side notches recovered by Clifford Evans from Frenchmen’s Island in Bull Arm, Trinity Bay.
ADZE SPALL n = 1
A small spall, polished on one face with a “platform” angle identical to that on the bit, is clearly a spall accidentally removed from an adze during use.
TIP FLUTE SPALLS n = 12
Twelve tip flute spalls are reported in Zone 2. Of these, 11 are made of white chert and one is made of pale grey chert.
BLADE-LIKE FLAKE n = 1
Only one blade-like flake of white chert was recovered from Zone 2.
PREFORMS n = 8
Eight preforms were recovered from Zone 2. Four white chert specimens are clearly the predecessors of the chipped type of endblades. Of these, two are bifacially flaked and two are ground flat on the ventral surface. The only two complete specimens measure 59.8 mm in length, 21 mm in width, 7.3 mm thick, and for the second example, 49 x 15 x 6 mm. A basal and a distal/medial fragment, also of white chert are also probable endblade preforms. The two remaining specimens, one of white chert the other of green chert, appear to be preforms for some kind of endblade.
SIDEBLADE PREFORM? n = 1
One brownish-white preform seems to have been worked toward the making of a sideblade. The specimen is roughly lenticular in shape and fine edge flaking occurs along approximately two-thirds of the artifact.
BLANKS n = 9
Nine blanks, all of white or brownish-white chert, were uncovered from Zone 2.
RETOUCHED FLAKES n = 7
Seven small retouched flakes of white chert were uncovered. Four specimens were unifacially retouched along a single lateral edge and the remainder exhibit more or less random flaking on the dorsal surface.
UNKNOWN ARTIFACTS n = 2
The first of these objects is made from brownish-white chert, is lanceolate in form with a straight base. The tip is missing. The ventral surface is completely polished and the dorsal surface partially so. Although the form is vaguely similar to a burin-like tool, it appears too thin to be one. It measures 50.2 mm long (tip missing), 18.5 mm wide and 3.6 mm thick.
The second specimen is a narrow polished slate(?) object, diamond-shaped in cross section and broken at a gouged perforation near what appears to be the mid-point. It seems too small and fine to have served as a weapon tip, and may be some sort of weaving or netting tool. The remaining fragment measures 40.1 mm long, 8.4 mm wide and 4.0 mm thick.
GROUND SLATE SLAB n = 1
A fragmented triangular thin slate slab (59 x 55 x 44 mm) with a bevelled edge (4 mm) is ground on both faces. One surface has an incised line probably indicating the breaking off line. The function is unknown. McGhee described a similar artifact from the Snowdrift site (Feature 4) on Dundas Island, High Arctic (McGhee 1981:71), as did Harp at the Port au Choix site (northwest coast of Newfoundland) described a similar artifact as a bevelled knife (Harp 1964:63).
GOUGED SLAB n = 1
Another slab fragment (103 x 135 x 87 mm) of schist-like stone has a gouged hole. The specimen exhibits a number of scratches running vertically and horizontally. The slab is partially encrusted with burned fat. Its function is unknown.
ROCK WITH BURNED FAT n = 1
A long, narrow elongated rock covered with burned fat was found in Zone 2. It might have been used as a lamp wick trimmer or a wick support at the bottom of an oil lamp.
DEBITAGE
A total of 3536 flakes were uncovered in Zone 2. Here again the most common raw material is the white or brownish-white chert. Small quantities of blue grey chert, purple rhyolite and quartz crystal were also found.
SOIL, FAT AND CHARCOAL SAMPLES
Two soil samples, one fat sample and four wood charcoal samples were collected in Zone 2.
SUMMARY
As a result of the 1996 field season hundreds of artifacts and portions of two Dorset dwellings were uncovered on Dildo Island. The Dorset material associated with these two houses is similar in a number of aspects to other Dorset material recovered in the Trinity Bay area. Together with Stock Cove (Robbins 1985), Frenchman’s Island (Evans 1982), Hopeall (Locke’s collection), Anderson Cove (Gilbert, personal communication) and a few small sites in nearby Placentia Bay (Linnamae 1971), the Dildo Island site presents a unique Dorset facies that has no counterpart in Newfoundland and Labrador or in any part of Arctic Canada. It is now clear that we can refer to a distinct Trinity Bay Dorset complex.
As seen from Dildo Island, the Trinity Bay Dorset complex is characterized by a technology that includes: elongated un-notched ground endblades, small triangular endblades with markedly concave bases, small quartz crystal scrapers and small quartz crystal microblades. No complete knives were recovered on Dildo Island but incomplete specimens suggest that the knives were mainly of two types: a unifacial type with a rounded distal end and a bifacial pointed type. Noticeable is the near absence of true blades which are usually very common in other Dorset collections.
Also characteristic of the material found on Dildo Island are endblade basal thinning flakes. These flakes are rarely reported from other Dorset sites. They may have gone unrecognized at other sites or they might be a specific trait of the Trinity Bay Dorset complex. In my opinion, endblade basal thinning flakes are the result of a manufacturing technique specific to Trinity Bay and are associated with the making of the elongated triangular ground endblades.
The most common raw material is a white, brownish-white chert. Quartz crystal is also used in small quantities but exclusively for scrapers and microblades. The exact provenience of this chert is not known but it is believed to be native to the Trinity Bay region (Robbins 1985:57). A local resident has mentioned a possible source of this type of material; next summer I plan to visit this location.
The large amount of debitage, the large size of some flakes and the presence of endblade basal thinning flakes, preforms and blanks at the site indicate that some degree of manufacturing occurred at the site. If we consider that this raw material had to be transported to the island it seems likely that the source might not be far away.
A total of twelve charcoal samples were collected from the site; four associated with House 2, and the eight remaining with House 1. Unfortunately a lack of funding did not allow any samples to be submitted for radiocarbon dating.
Since no faunal material was recovered from Dildo Island any inferences about the season of occupation of the site remain speculative. However, if we consider site location and the abundance of marine resources around Dildo Island, it is possible to suggest that the island could have been used from late spring throughout the entire summer season.
It is generally assumed that harp seals do not frequent the Trinity Bay area (Robbins 1985:114), but recent interviews conducted with local informants seem to contradict this assumption. In the spring of the year, harp seals are known to come in great numbers to feed in the waters between the east side of Dildo Island and the mainland. For example, in one day during the last week of April 1997, a local resident killed over 70 harp seals between Dildo Island and Hopeall Head.
In summer, the Dorset people would have had access to a wide variety of resources. In early summer, for instance, the island is literally covered with bird nests and people could have collected eggs or hunted birds. In summer harbour seals and a wide variety of fish species and crustaceans are also available. Salmon could also have been speared as they approached the northern end of the island. Fall might have brought a decline in the subsistence resources and it is possible that people moved inland or to some other location on the coastal zone to search for caribou. The Dorset component at the Stock Cove site, on the isthmus of Avalon, could represent such a site. If this is the case, the Trinity Bay area is probably the only area on the Island of Newfoundland where we might be able to access the complete seasonal round of Dorset peoples.
The Dildo Island site covers an area of about 25,000 square metres and with the Port au Choix, Cape Ray and Stock Cove sites, is one of the largest Dorset components on the Island of Newfoundland. Until recently, most of our knowledge of the Dorset culture on the Island of Newfoundland came from the Cape Ray and the Port au Choix sites on the west coast of Newfoundland. Although some work had been done in some other parts of the Island, the pre-eminence of these west coast sites has coloured our picture of the Newfoundland Dorset culture and left us with an impression of homogeneity throughout the Island. Robbins (1985) challenged that view over fifteen years ago when he described a unique Dorset component at the Stock Cove site at the bottom of Bull Arm, Trinity Bay. As a result of his work Robbins suggested that the Dorset culture in Newfoundland should be envisioned rather as a mosaic of “regional expressions” (Robbins 1985:123). The Trinity Bay Dorset component is certainly the most distinctive of these regional expressions. Our work on Dildo Island confirms Robbins’ hypothesis and clearly demonstrates that Dorset people developed a unique culture in the Trinity Bay area.
Moreover, present evidence suggests that the Trinity Bay Dorset are perhaps the most recent of these regional expressions with dates of 1560 ±60 and 1280 ±60 B.P. from Stock Cove (Robbins 1985).
The Dildo Island site is a large and essentially undisturbed Dorset Eskimo habitation site. For this reason and because of the unique character of the Dorset material here, it offers a number of challenging research opportunities. For instance: it provides a unique window to look at the adaptation of a distinct Dorset population in a very restricted area, something that is not possible in other areas of Newfoundland; it offers an opportunity to study the evolution of this distinctive culture. Further research projects should be designed to address these culture-historical questions regarding the origin and the evolution of this unique facies of Dorset culture. It is now crucial that we obtain dates from Dildo Island as they hold the key to understanding cultural evolution. Finally, with more investigation we have the possibility of discovering a complete subsistence-settlement system for the Dorset people in Trinity Bay. This would be a first in Newfoundland, since west coast and other sites appear to illustrate only a portion of the Dorset subsistence-settlement system.
More immediately, House 2 needs prompt attention as it is located only a few feet away from a cliff and is threatened by erosion. The 1997 field season will focus on the excavation of House 2.
Note:
[1] The Locke collection is housed at the Newfoundland Museum.
Evans, C.O.
1982 – “Frenchman’s Island Site (ClAl-1) Preliminary Field Report.” In Archaeology in Newfoundland & Labrador 1981, Annual Report 2, edited by J. S. and C. Thomson, pp. 210-225. Historic Resources Division, Department of Culture, Recreation and Youth (since renamed Culture and Heritage Division, Department of Tourism, Culture and Recreation), Government of Newfoundland and Labrador. St. John’s.
Gilbert, W.
1990 – “‘Divers Places’: The Beothuk Indians and John Guy’s Voyage into Trinity Bay in 1612.” Newfoundland Studies 6(2):147-167.
1996 – “Baccalieu Trail Archaeology Project 1995, Phase 3: Dildo Island.” Report on file at Culture and Heritage Division, Department of Tourism, Culture and Recreation, Government of Newfoundland and Labrador. St. John’s.
Harp, E. Jr.
1964 – The Cultural Affinities of the Newfoundland Dorset Eskimo. National Museums of Canada, Bulletin 200.
Linnamae, U.
1971 – “Preliminary Report of an Archaeological Survey of Placentia Bay, Newfoundland.” Report on file at Culture and Heritage Division, Department of Tourism, Culture and Recreation, Government of Newfoundland and Labrador. St. John’s.
1975 – “The Dorset Culture: A Comparative Study in Newfoundland and the Arctic.” Technical papers of the Newfoundland Museum, Number 1. St. John’s.
Marshall, I.
1990 – “Report on Archaeology Survey in Trinity Bay Conducted Between 21st and 14th of July and 25th of August, 1990.” Report on file at Culture and Heritage Division, Department of Tourism Culture and Recreation, Government of Newfoundland and Labrador. St. John’s.
McGhee, R.
1975 – “The Dorset Occupations in the Vicinity of Port Refuge, High Arctic Canada.” Archaeological Survey of Canada. Paper no. 105. National Museums of Canada. Ottawa.
Robbins, D.T.
1985 – “Stock Cove, Trinity Bay: The Dorset Eskimo of Newfoundland from a Southeastern Perspective.” Unpublished M.A. thesis, Department of Anthropology, Memorial University of Newfoundland. St. John’s.