The assemblage of Dorset Eskimo artifacts recovered from Stock Cove during the 1981 excavation consists entirely of lithic tools and associated manufacturing debris. No implements made from organic materials were found due, presumably, to preservation conditions. The following inventory includes tools which functioned in many aspects of Dorset Eskimo life: weaponry, cutting and scraping tools, implements for the fabrication of others, and vessels for domestic use. As well the process of making many of these artifacts is reflected in the collection of unfinished specimens and flaking debris.

One type of chert is predominant in the Dorset Eskimo assemblage. Artifacts manufactured from this chert are highly susceptable to weathering, and now exhibit an off-white, mottled brown and white, or brown patina. This patination is apparently common to cherts native to the Trinity Bay region (Art King, M.U.N. Geology Dept., pers. com.). In comparison to cherts found in other Newfoundland collections, this material might be considered of "medium" quality. The fact that it was often ground suggests that it is less hard than many other types of Newfoundland chert.

The second most popular material from which chipped stone tools were manufactured is crystalline quartz. Frequently used for making end scrapers and microblades, quartz crystal is very hard and produces a razor-sharp cutting edge. Green and grey-green cherts are also present, but they are rare. These materials are much finer-grained than the patinated chert, and were occasionally used for making end scrapers.

n=366

Triangular points once used to tip harpoon heads are numerous at the Stock Cove site. These projectiles were constructed either by chipping or a combination of chipping and grinding. The collection is accordingly divided into three categories:

Chipped" specimens were shaped and finished exclusively through flintknapping; "Ground" specimens were initially flaked, and then finished with a grindstone; "Chipped and Ground" end blades display both techniques of finishing.

CHIPPED END BLADES

n=162 (Tables 2, 3; Plate I)

Harpoon end blades manufactured exclusively by chipping constitute 44% of the Stock Cove end blade collection. Bifacially flaked specimens are the most frequent, tip-fluted specimens are second in popularity, and unifacial examples form a minority.

Bifacial n=79

The majority of bifacial end blades are completely flaked over both surfaces. The quality of flaking varies, with less than 20% appearing precisely worked. Many are roughly chipped and exhibit surface irregularities, yet the outline form is, in most cases, symmetrical. All specimens are basally thinned, usually to an equal degree on both the dorsal and the ventral surface.

A variety of shapes is present. Approximately 50% display a similar form: an elongated triangular outline with straight or slightly convex sides, and a straight or slightly concave basal margin. Other specimens are not as uniform, ranging from very squat examples to quite long and narrow ones. Side-notching is rare; five examples have shallow side-notches immediately above their bases, one has a pair of notches placed further up the body, while another specimen has two sets of notches, the second set located near the tip.

Unifacial n=35

Unifacial end blades are, by definition, flaked on one surface only, although the ventral or unworked face may exhibit edge retouch. 64.5% of these unifacial specimens show complete surface flaking on the dorsal face. 77.7% are basally thinned; of these about one-half (40.7%) exhibit an equal amount of thinning on both faces.

With respect to shape, unifacially chipped end blades are essentially similar to bifacial specimens. The greatest difference is symmetry, as 53.3% are asymmetrical in outline. Lateral edges are again slightly convex, and basal margins are slightly concave or straight. Whereas bifacial specimens are bi-convex in both longitudinal and transverse cross-section, most unifacial end blades are either plano-convex or concave-convex in longitudinal cross-section, and plano-convex in transverse cross-section.

Tip-fluted n=48

Forty-five tip-fluted specimens exhibit a longitudinal median ridge on their ventral surface, where two or more tip-flute flakes were removed. Each of the remaining three examples have had a single tip-flute flake removed. Most specimens (93.3%) are completely flaked on the dorsal surface. 77.1% exhibit no evidence of working -- other than tip-fluting and basal thinning -- on the ventral face. All specimens were basally thinned on the dorsal and the ventral surface; 51.3% were thinned to an equal degree on both faces.

A symmetrical, elongated triangular form occurs most frequently. Lateral edges are usually gently convex. Tip-fluted specimens show the highest frequency of slightly concave basal margins (81.1%). Longitudinal cross-sections are either plano-convex or bi-convex, depending upon the degree of tip-fluting and basal thinning on the ventral surface. Variation in the size of the ventral tip-flute ridge may result in either a bi-convex, a plano-convex, or a triangular-convex transverse cross-section.

GROUND END BLADES

n=107 (Tables 4,5; Plate II)

Ground" harpoon end blades are manufactured from chert and are judged to have grinding over at least 75% of their surface area. This degree of surface grinding differentiates them from specimens in the following "Chipped and Ground" class, all of which have less than 75% of their surface finished by grinding. "Ground" end blades show no surface chipping, although most exhibit a fine edge serration, or a coarser edge retouch.

Most ground chert end blades (86.6%) have a symmetrical outline. 83.3% have gently convex edges. Slightly concave and straight basal margins occur with approximately equal frequency. 59.7% are either bi-convex or triangular-convex in longitudinal cross-section, depending upon the precision with which surface and basal grinding was done. Three specimens have broad, shallow side-notches formed after grinding was done.

Grinding was not performed in an identical manner on all specimens. Generally, two grinding techniques were followed. "Flat" grinding was performed in a single plane, parallel to both axes of the specimen. "Bevelled" grinding produced a peaked or convex surface. A single specimen may exhibit one or both techniques. While bevelled grinding may occur on both faces of an end blade, flat grinding, when it does occur, is restricted to a single face. In some instances bevelled grinding was executed with machine-like precision, producing a specimen that appears bilaterally and bifacially symmetrical. More often, however, it was done less exactly. The transverse cross-section of a ground chert end blade reflects the grinding technique employed, as well as the precision with which grinding was executed, and forms the basis for the following descriptive classification.

Plano-convex n=6

The ventral surface of these six specimens is ground flat, while the dorsal face is bevel-ground to a convexity. Basal grinding is done on both surface and in the case of these plano-convex examples it is more pronounced on the dorsal face. Three specimens have finely serrated edges.

Triangular n=2

Specimens with a triangular transverse cross-section are also flat-ground on the ventral face and bevel-ground on the dorsal. Compared to plano-convex examples, however, the dorsal grinding was more precisely controlled, producing a well-defined longitudinal midline. Basal grinding is again more pronounced on the dorsal face, and lateral edges are serrated in both cases.

Bi-convex n=63

The most popular type of ground end blade has two convex faces. The exact transverse cross-sections vary, from examples which are symmetrical or lens-shaped, to ones which are markedly asymmetrical, i.e., nearly plano-convex. Basal grinding is usually more pronounced on the dorsal face, yet several symmetrically bi-convex specimens exhibit an equal degree of basal grinding on both faces. Most specimens have edge serration; those which do not have a coarser edge retouch.

Triangular-convex n=30

These specimens are similar to the preceding type, yet exhibit more precise grinding on one face. As a result, this face has the well-defined longitudinal midline as seen in "Triangular" examples. Basal grinding occur on both faces and is usually slightly more pronounced on the face with the longitudinal midline. Again, most specimens have edge serration, while those without have the coarser edge retouch.

Diamond n=6

Precisely controlled bevel-grinding on both surfaces produces a flattened diamond transverse cross-section, with a longitudinal midline on each face. As basal grinding was also performed in a near-identical manner on each face, these specimens are bilaterally and bifacially symmetrical. Lateral edges are serrated in all cases.

CHIPPED AND GROUND END BLADES

n=97 (Tables 6, 7; Plate II)

All these end blades are chipped, yet also exhibit grinding facets. Grinding was performed secondarily, to sharpen edges and tips, and to thin bodies and bases. Occasionally, additional edge retouch has removed areas of a ground surface.

Three classes of chipped and ground end blades are distinguished according to their primary method of manufacture: chipping.

Bifacial n=37

Exactly 75% display surface grinding facets, while 83.3% exhibit basal grinding. Surface and basal grinding occur together on 53.3%. Bases were often chipped subsequent to being ground; 56.7% are thinned by a combination of chipping and grinding.

Most specimens have a symmetrical outline and gently convex lateral edges. Basal margins are usually slightly concave or straight. Longitudinal cross-sections are most frequently bi-convex, although varying degrees of bi-convexity exist, from nearly plano-convex to symmetrically bi-convex.

Unifacial n=27

The occurrence of surface and basal grinding on unifacial specimens is essentially similar to bifacial ones. Basal thinning is more commonly performed by grinding than chipping, and is usually more pronounced on the dorsal face of the specimen. 42.9% exhibit post-grinding edge retouch.

A symmetrical outline and convex lateral edges again predominate. Unifacial end blades show a slightly greater occurrence of slightly concave basal margins than do bifacial specimens. The longitudinal cross-section assumes a variety of forms, most commonly either bi-convex or plano-convex. Most are plano-convex in transverse cross-section.

Tip-fluted n=33

Tip-fluted specimens show a marked decrease in the occurrence of surface grinding, while basal grinding occurs with 100% frequency. The bases of tip-fluted specimens are commonly chipped as well as ground. Concurrent with the decrease in surface grinding is a low frequency of post-grinding edge retouch; of the fifteen specimens exhibiting surface grinding, only one was subsequently retouched. Tip-fluted specimens show a form of grinding not present on bifacial and unifacial end blades, namely, grinding of the ventral tip-flute ridge. On thirteen specimens (46.4%) the distal portion of the longitudinal ridge produced by the tip-fluting process has been flattened and nearly obliterated by grinding, The objective was, presumably, to sharpen the tip.

In outline, these specimens show a less symmetrical form than either bifacial or unifacial examples. Lateral edges are usually convex, and basal margins concave. Cross-sections are often bi-convex, although considerable variation exists.

n=100

Only two of the 100 knives/bifaces are intact, while there are 50 tip fragments, 14 midsections and 34 bases. Most of the fragments are undistinctive, showing only that the Stock Cove Dorset Eskimo inhabitants were making and using a variety of long triangular bifaces, not unlike oversized harpoon end blades. Fourteen of the bases are sufficiently distinct to place in two types, each type including one of the intact specimens. Many of the fragmentary ones are similar to these types, but cannot be classified due to their condition.

LARGE BIFACES

n=8 (Plate III)

These bifaces are quite thin in proportion to their length and width; the intact specimen measures 85.2 mm long, 37.1 mm wide and 6.8 mm thick. Most of the fragmentary examples are larger, the largest was about 100 mm long when intact. Bases are carefully thinned and retouched to a slight concavity. This preparation, along with the notches present on six examples, indicate that these tools were hafted. Notches are shallow and placed immediately above the base; one specimen has two sets of notches. All examples are bifacially thinned and two have surface grinding near their base. These tools may have served a variety of functions. The thin and sharp tip of the complete one indicates it may have been a projectile -- a spear or lance point -- and the sharp edges of all specimens were likely well-suited for cutting.

SMALL BIFACES

n=8 (Plate III)

Besides being smaller than the preceding ones, these tools are proportionally longer and narrower. Again, they are quite thin; seven are less than 5 mm. Bases are bifacially chipped and in one instance subsequently ground. Shallow side-notches, placed just above the gently concave base, are present on all specimens. Two examples have double sets of side-notches. These bifaces are generally smaller, more slender, and lighter than the "Large Bifaces They may have tipped projectiles, perhaps of the throwing rather than the thrusting type, in addition to being used as cutting knives.

UNCLASSIFIED FRAGMENTS

n=84

Fifty distal fragments are unclassified. Three have blunted tips, and were therefore probably not projectiles. All are triangular in outline; some are broad as in "Large Bifaces" while others have the longer, narrower outline of the "Small Bifaces". Three are tip-fluted and one is ground. Three of fourteen medial fragments have parallel sides; others have converging sides. Two have grinding on their surface.

Eight of twenty proximal fragments have the long, narrow outline of "Small Bifaces". None is notched, however, and their basal margins are either straight or irregular. Two notched specimens are unlike either of the preceding types; both have very small side-notches placed high above the base, and parallel rather than converging sides. One is bifacially chipped and ground and quite thin, while the other is a uniface, and much thicker. The remaining bases are asymmetrical and irregular, and appear to have received a minimal amount of shaping prior to use.

n=72

The mixing between Recent Indian and Dorset Eskimo materials at Stock Cove has presented particular problems with respect to end scrapers. It is generally not difficult to differentiate between Recent Indian and Dorset tools, as styles are usually distinctive and diagnostic. The selections of raw materials is also helpful. There is a dichotomy between the fine-grained colourful cherts used by the Recent Indian inhabitants of Stock Cove and the weathered white material prevalent in the Dorset assemblage. These distinctions do not hold true, however, with respect to end scrapers. The Dorset people of Stock Cove used a variety of coloured cherts in addition to the weathered chert and quartz crystal for making end scrapers, and a morphological distinction between Recent Indian and Dorset end scrapers is not readily observable. Several specimens may be confidently identified as Dorset (for example those with "graving spurs") but many are of questionable origin.

Failing to produce a replicable or quantifiable method of identifying end scrapers (several methods of metrical analysis and plotting were attempted), the mixed collection was "divided" by visual inspection. This method is highly subjective and reflects the prejudice of the researcher, but is inescapable until further research produces an objective means of identification.

Based on the fact that quartz crystal is unknown to occur in Newfoundland Recent Indian assemblages, all quartz crystal end scrapers are presumed to be Dorset Eskimo. Diagnostic Dorset Eskimo tools (e.g., microblades) made from quartz crystal have been found at Stock Cove and other sites, so there is no doubt that Palaeo-Eskimos were using this material. The fact that quartz crystal end scrapers outnumber chert specimens at Stock Cove is a significant and somewhat unusual occurrence in Newfoundland Dorset assemblages. A possible explanation may be that the popular weathered chert was in fact unsuitable for end scrapers -- a tool requiring a durable working edge. The presence of the better quality coloured cherts in the end scraper collection is also explained by this reasoning.

QUARTZ CRYSTAL SCRAPERS n=51
(Tables 8, 9; Plate IV)

The small size of most quartz crystal scrapers is likely a function of the size of available crystals. Three types of scrapers are distinguished on the basis of outline form and the number of working edges.

Four-sided, single working edge end scrapers n=30

Several quadralateral shapes occur; the most common form is rectangular. Twenty specimens are surface flaked on their dorsal face while ten exhibit crystal cortex. All have a single working edge, twenty-one along their length dimension and nine along their width. Two specimens are stemmed. Most are incompletely worked along their lateral and proximal margins.

Three-sided, single working edge end scrapers n=10

Except for their triangular outline, these end scrapers are quite similar to the preceding type. Five are completely flaked on their dorsal surface, four display some crystal cortex, and one is fashioned on a quartz crystal flake. None is stemmed, and non-working edge retouch is rare.

Scrapers with multiple working edges n=5

Included in this type are two specimens which are equilaterally triangular in outline, and have three working edges. Another has a rectangular shape and four working edges. An irregular specimen has two slightly convex working edges and a single concave edge. Another unique specimen has two convex working edges, fashioned on opposite faces of a quartz crystal flake.

Concave side scrapers n=5

Four quartz crystal microblades and one linear fragment of quartz crystal have one side retouched to form a concave working edge. Opposite edges are not retouched. These concave side scrapers or "spoke­shaves" range from 14.2 mm to 18.0 mm in overall length, and their working edges vary between 6.5 mm and 10.1 mm long.

End-of-blade scraper n=1

One quartz crystal microblade fragment, measuring 7.2 mm in length and 6.0 mm in width, is retouched at one end to form a convex working edge.

CHERT SCRAPERS n=21
(Tables 8, 9; Plate V)

Chert scrapers are quite variable and do not readily lend themselves to classification. Because of the tentative nature of the following typology the entire collection is illustrated in Plate V, each row corresponding to one of the following types.

Large, three-sided end scrapers n=7

The seven specimens included here are approximately triangular in outline, with symmetrical convex working edges. All are flaked on their dorsal surface and one is also ground. Lateral edges are retouched, producing "graving spurs" on one example. A variety of green, blue and tan cherts occur.

Small, three-sided end scrapers n=3

These are essentially similar to the preceding ones, but are much smaller. In size and form they closely resemble quartz crystal end scrapers. They may be expended, or nearly expended, examples of the above type.

Four-sided end scrapers n=4

These end scrapers have steeper and less regular working edges than either of the above types. All are flaked on their dorsal surface; one is also worked on its ventral face. Lateral edges are retouched.

Flake end scrapers n=3

These specimens were made on flakes, and are irregularly triangular in outline. They are quite thin in cross-section and exhibit a minimal amount of working edge and non-working edge retouch. All are made from weathered chert.

Composite end/side scrapers n=4

These four rectangular specimens appear to be composite end/side scrapers. End working edges are thin, while working sides are thicker and steeper. All are constructed from weathered chert; two are made on blades.

n=185

(Tables 10, 11; Plate VI)

The core from which blades are produced requires preparation of a striking platform, and a face from which the blades are removed. This preparation is accomplished by the removal of ridge flakes -- linear flakes with triangular cross-sections and transverse flaking -- and other core debris. The blade itself has been defined as "... a parallel sided flake with a prepared striking platform and/or prepared platform edge, with one or more straight dorsal ridges parallel to the lateral edges" (Linnamae 1975:131).

A striking feature of the blade technology at Stock Cove is its apparent poor development, when compared with other Newfoundland Dorset assemblages. Applying the above definition, no more than three dozen blades can be identified in the Stock Cove collection. The collection is, instead, replete with a variety of blade-like flakes and linear flakes which cannot be considered true blades due to their irregularities. The infrequent occurrence of true blades is believed to be partially a function of the raw materials used by the Dorset inhabitants of Stock Cove, in that none was ideally suited to a blade industry. Rarely, attempts at producing blades from patinated chert were made, but results were generally unsuccessful; only a few specimens fulfill the requirements of the above definition. No patinated chert blade cores were recovered from Stock Cove, which appears to reinforce the suggestion that this material was not suitable for blade technology. Quartz crystal was also used, with only slightly better results. In this case the imperfections may be due to an additional factor: the size of the quartz crystals. Most of the quartz crystals used by the Dorset people of Stock Cove are quite small, and therefore were probably difficult to prepare in the manner identified with true blade production. However, the elongated, facetted form of quartz crystals made them suitable for producing blade-like flakes with a minimum of core preparation. These quartz crystal blade-like flakes are numerous at Stock Cove, and seem to be a substitute for true blades.

As true blades are nearly absent at Stock Cove, all imperfect blades and blade-like flakes are included in this category of "Blades/blade-like Flakes". This presumes that blades and blade-like flakes had the same function, and that blade-like flakes were made when it was difficult (or undesirable) to make blades.

CHERT BLADES/BLADE-LIKE FLAKES n=67

Of the 67 specimens manufactured from chert, 15 have parallel sides and arrises and can be considered true blades. The remainder have irregular sides and irregular and often incomplete arrises. Most examples are made from patinated chert. Three specimens of purple rhyolite, two of green chert, and one of Ramah chert are present.

QUARTZ CRYSTAL BLADES/BLADE-LIKE FLAKES n=118

Quartz crystal specimens number 118, including 19 examples identified as true blades. The remaining 99 blade-like flakes present irregularities similar to chert ones. Whether or not the 19 blades were produced in a different manner than were the 99 blade-like flakes is debatable. All may have been made by the same process, with the straight sides and arrises of the 19 "true" blade being fortuitous. The process by which blades/blade-like flakes were produced from quartz crystals is elaborated upon in the following section. The variable, non-standard form of quartz crystal blade-like flakes is reflected in the ranges of lengths, widths, and thicknesses (Table 12).

(Plate VI)

There are 51 quartz crystal blade/blade-like flake cores, 38 core fragments, and nine unworked crystal fragments in the collection. Considering only the intact cores, it can be seen that the proportion of cores to blades/blade-like flakes is quite high: 51:118, or about 1:2.3. This suggests that these quartz crystals were very unproductive blade cores. Sixteen of the 51 intact cores show multiple arrises where two or more blades/blade-like flakes were removed. Of the remaining 35, 25 exhibit a single blade/ blade-like flake arris, and 10 show evidence of working but do not have blade arrises. These last ones were, perhaps, totally unsuccessful attempts at blade production.

Cores from which two or more blades were struck best exhibit the technique that was practised. The crystal was first broken transversely to produced a flat platform; preparation flakes were then struck, originating at this platform and extending part of the length of the crystal. This angled the platform slightly in relation to the longitudinal axis of the crystal. The crystal was then rotated 180° and blades/blade-like flakes struck from its face, on the opposition side from the preparation flakes. When the crystal core was expended, the platform had assumed a wedge shape.

Core detritus is represented by 90 fragments of quartz crystal. This includes flaking debris from preparing the platform, and initial flakes from the core face from which blades were subsequently removed. Flakes and linear flakes are present, along with odd-shaped fragments that were presumably the results of unsuccessful attempts at core preparation. Crystal cortex is often observable on the dorsal surface of core detritus flakes.

n=118
(Tables 12, 13)

The tip-fluting technique produces distinctive debitage which provides some clues as to how this technique proceeded. After the tip of a preform was prepared, the initial or primary tip-flute flake was struck. This produced a linear facet on the preform and a corresponding flake with flaking scars on its dorsal surface -- the former surface of the preform. Next, a second fluting flake was struck from the tip of the preform adjacent to the first. This would remove the remainder of the preform's original surface along with a portion of the linear facet from the first tip-flute flake. Accordingly, secondary tip-flute flakes have flaking scars over a portion of their dorsal surface and a facet on the rest. Sometimes a third or tertiary fluting flake was removed; these are recognized by the presence of two linear facets on the dorsal surface.

Usually two and sometimes three tip-flute flakes were struck from the same surface of a preform, and can be assigned either a left or a right position. With the flake orientated so as the dorsal surface is up and the striking platform is distal to the observer, left examples have an edge facet along their left-hand margin. Right tip-flute flakes have this edge facet on their right-hand side.

There are 33 primary, 56 secondary and 29 tertiary tip-flute flakes in the Stock Cove collection. Five primary and two secondary specimens are ground on their dorsal surface, demonstrating that in some cases grinding was done prior to tip-fluting. Two fluting flakes, a left secondary and a right tertiary, are from the same preform and can be reassembled. The primary flake belonging with these two is not present.

Metrical analysis shows that primary, secondary and tertiary specimens are similar with respect to maximum thickness (usually measured at the bulb of percussion). Primary examples, however, are generally larger and more variable than secondary and tertiary specimens with regards to maximum length and width (Table 14).

n=129
(Tables 14, 15; Plate VII)

A "preform" is an unfinished form of a stone tool that does not display the precise flaking and shape of the final product. The 129 examples from Stock Cove vary in size, shape and degree of finish, and give a general impression of the flintknapping techniques practised by the Dorset inhabitants. The collection is subdivided into two categories: those which are true core tools, i.e., were formed through a progressively refined reduction process; and those which were formed on conveniently shaped flakes or fragments, referred to as, for lack of a better term, "flake" preforms. The distinction is, in a sense, a matter of degree as core preforms also originated with suitably shaped chert fragments or "blanks". Core preforms were subjected to a more complete -- perhaps "standardized" -- reduction process than were flake preforms, and thus several attributes serve to distinguish the two types.

With a single exception all preforms are manufactured from weathered chert. The exception is made from blue rhyolite.

CORE PREFORMS n=81

Preforms which have been reduced from a larger blank have flaking scars on one and usually both faces. Depending upon the stage in the flintknapping process the specimen has reached, it may also exhibit retouch of the lateral edges and the base. As the reduction process continued the preform became smaller, thinner and more regular in outline, with a greater number of smaller flake scars on its surfaces. At some point in the process, when the specimen had attained a regular outline form, tip-fluting was sometimes done. This involved retouching the distal tip to a concave shape, and then removing long, usually narrow, flakes from the ventral face, originating at the projections of the concave-shaped tip. This technique greatly assisted in thinning the preform and was repeated as the reduction process continued. 33.3% of the core preforms demonstrate tip-fluting.

The shape of a preform is largely dependant upon the stage in the manufacturing process it has reached. Large ones with only a few flake scars -- from direct percussion flaking -- are usually asymmetrical, with irregular edges and little or no basal thinning. Smaller ones generally have a symmetrical outline and regular convex edges and base. Size, thickness and final basal modification, or inherent flaws that negated further reduction, distinguishes these forms from finished tools. Some specimens appear to have broken in the final stages of manufacture, as is the case with two tip-fluted examples when the tip-flute attempt fractured the specimen.

The emphasis on creating a symmetrical form (in the latter stages of manufacture) and thinning the body by tip-fluting and surface flaking suggest that many of these core preforms were intended to become harpoon end blades or larger projectiles where symmetry, thinness and sharp edges and tips were desirable characteristics. The more refined preforms attain the elongated form popular among Stock Cove projectiles.

FLAKE PREFORMS n=48

These preforms are in a sense "forms of convenience", whereby a suitable flake or fragment was modified via edge retouch to produce the desired shape. Flaking scars are usually absent from one or both faces; the surface is either cortex or the ventral face of the original flake. The striking platform and/or the bulb of percussion is often visible. Bifacial or unifacial edge flaking was done as required to thin the specimen and shape its outline, and it is notable that tip-fluting was never performed on flake preforms.

Outline symmetry depends to some extent on the shape of the original fragment or flake, and on the degree of edge retouch. Frequently one or both faces are plano, in contrast with the usually convex surfaces of core preforms.

Flake preforms cannot be readily associated with any one finished artifact type. They vary in size, shape and thickness and were probably intended for a variety of uses. With further work some may have become end blades, but most are unsuited to this purpose and were probably intended as cutting knives.

n=21 (Plate VIII)

Three types of ground slate end blades are present, along with one form of slate knife represented by two specimens. Three examples of a slender knife-like tool correspond to Harp's "flat-bladed chisel" (Harp 1964:64, pl. XVIII). Five specimens are individually unique.

TRIANGULAR END BLADES WITH LINE HOLES n=6

These six examples are manufactured from light grey and light brown slates. All are broken basally and two are also missing their tips. Each specimen has two surfacial grinding facets which meet at a longitudinal midline, and a basal grinding facet on each face. The transverse cross-section is a flattened diamond shape, while the longitudinal cross-section is either a flattened diamond or a lens shape. Line slots are gouged, and approximately parallel to the long axis of the specimen. Sides are straight, and tips are pointed and sharp.

TRIANGULAR END BLADE WITH INCISED SIDE-NOTCHES n=1

This end blade is made from light grey slate, and is intact, measuring 65.8 mm long, 27.0 mm wide, and 6.0 mm thick. A surface grinding facet on each face extends from the basal margin almost to the tip. Edges are bevelled on each face, producing a symmetrically hexagonal transverse cross-section. Three pairs of triangular-shaped notches are cut into the sides. Sides are gently convex and the base is straight.

END BLADES WITH ROUNDED TIPS n=4

Three of these end blades are ground flat on their surfaces and have bifacially bevelled edges, blunted by grinding. As bases are missing, it is impossible to determine whether they had line holes or notches. In outline they are not unlike the triangular end blades described above, but have a thin, rounded tip rather than a sharp and pointed one. The fourth specimen differs from the preceding one in outline; instead of presenting a triangular form, it converges towards the base and the tip. This end blade has a single line hole, which is placed much closer to the tip than was the case among the first type described above.

KNIVES n=3

The large slate knife in Plate VIII appears to have a handle shaped by chipping. The specimen is ground flat on both surfaces and the working edge is bevelled unifacially. Two additional specimens included here are edge fragments of other such tools, exhibiting similar surface and edge grinding.

"CHISELS" n=3

The designation "chisel" is used for lack of a better term, and is in accordance with the name used by Harp for similar tools. It is not entirely acceptable however, as the working bit of a chisel has a unifacial bevel, whereas the tools included here are bifacially bevelled, more like a sharpened screwdriver. These specimens are made from light brown slate; two are midsection fragments and the third is broken basally and damaged slightly at its tip. Surfaces are ground flat and edges are bevelled bifacially. Edges are blunted by grinding, and therefore it is unlikely that these tools were cutting knives. Sides are straight and converge slightly towards the tip.

MISCELLANEOUS n=5

A midsection fragment of a large end blade or knife, measuring 34.2 mm wide and 3.9 mm thick, has bifacially bevelled, sharp edges and a single gouged hole. A much smaller specimen has convex surfaces and a single gouged hole. A small end blade (36.3 mm long, 12.2 mm wide, 4.4 mm thick) is similar to ground chert end blades, except that it has a rounded, convex base. Two ground slate pieces with flat surfaces and bevelled edges may be either end blade or knife fragments.

n=2

One burin-like-tool is essentially intact, having only minor breakage along its edges. It is rectangular in outline, measuring 42.1 mm long, 28.5 mm wide, and 3.4 mm thick. Its surfaces are ground flat, and the thickness is quite uniform. Two edges are bevelled from the same face; the sharp corner resulting from this bevelling is chipped slightly, possibly from use. The second specimen is a fragment of a similar tool with flat surfaces and a bevelled edge. Thickness measures 3.6 mm.

n=11
(Table 16)

Ten of the eleven hammerstones are intact, while one specimen is broken across its diameter and missing a substantial portion of its original mass. This broken specimen is pecked but not spalled on the end which is present (because of its incomplete state it is excluded from further analysis). Nine of the remaining ten hammerstones are elongated; roughly "egg" or "teardrop" shaped. All exhibit pecking on both their ends; eight have spalling on one or both ends. One of the largest examples seems to have also functioned as an anvil stone, as it is battered on its surface as well as both ends. The size and weight of the eight large specimens and the spalling evident on their ends suggest that they were employed as direct percussion flakers for "roughing out" preforms. As the two small hammerstones are pecked but undamaged they were likely used for more precise and less strenuous work.

n=7
(Table 17, Plate IX)

Two varieties of abraders are differentiated on the basis of material. The first kind, represented by two specimens, is manufactured from a very hard, dense stone. One has an elongated shape and a quadralateral cross-section; each of its four surfaces is smoothed. The ends of this specimens are pecked, indicating it served double use as a hammer­stone. The second example of this kind is ovoid in outline and quite thin in cross-section -- probably a fragment of a larger stone. It has one convex ground surface.

The five examples of the second variety are made from a much softer grey material. All specimens are now fragmented, but appear to have originally been rectangular in shape. Each have several smoothed facets or platforms -- one or two broad surface facets and narrower facets along their edges. The largest specimen has a total of six smoothed facets.

While the first type of abrader is sufficiently hard to grind chert, the second kind appears much too soft for this purpose. It is possibly suited for working softer materials such as slate, bone, and antler.

n=10
(Plate IX)

Ten fragments of soapstone represent a maximum of six vessels. Four fragments belonging to the same vessel have been cross-matched and two other pieces presumably from a different vessel were also found to join. All soapstone fragments have two finished surfaces and appear to be from angular rather than rounded vessels. All but one have burnt fat adhering to their surfaces. Because of the fragmentary nature of these pieces it is impossible to reconstruct vessel size or shape; one piece permits a partial reconstruction.

Specimen 1726 is a wall fragment from a soapstone vessel whose inside depth (from rim to turn of the base) measures 41.8 mm. As the thickness of the base cannot be precisely determined the total height of the vessel can only be estimated -- at about 52-55 mm. Wall thickness varies between 6.8 mm at the rim and 11.6 mm at the base. Three holes are gouged through this piece and two vertical lines incised on the outside.

The remaining fragments are not useful for reconstruction, and can only be described. The four articulating pieces are part of a wall including the rim. Thickness varies from 9.1 mm at the rim to a maximum of 13.2 mm, and the vessel originally measured in excess of 120 mm in height. Three gouged holes are present. Another three pieces are also wall and rim fragments. Minimum and maximum thicknesses are 11.0-14.4 mm, 5.4-9.4 mm, and 6.2-7.2 mm. One of these pieces has three gouged holes. The final fragment appears to be part of a base with a maximum thickness of 21.1 mm.

Two kinds of stone material predominate in the Stock Cove artifact assemblage. Chert that now exhibits a white, speckled brown and white, or brown patina was original light blue in colour and is believed to originate from sources in the Trinity Bay area. This material was used extensively for making harpoon end blades, and various forms of knives/bifaces. It was occasionally employed in manufacturing end scrapers and blades/blade-like flakes. For reasons which remain unproved, the Dorset Eskimo occupants of Stock Cove and other sites in southeastern Newfoundland generally preferred not to use this chert in making scrapers and blades, instead employing crystalline quartz. This may be due to the comparative hardnesses of the two materials with, in this respect, quartz crystal being far superior. An end scraper likely required a tough and durably working edge that could not be obtained with the relatively soft local chert. Similarly, desirable characteristics of a blade were durability and razor-sharpness, attributes not generally obtainable with the local chert. For both applications quartz crystal was highly suitable.

Harpoon end blades exhibit some unusual forms, primarily because of the frequency with which grinding was done. This technique of manufacture is not generally associated with the making of "chert" harpoon end blades. The frequency of ground chert end blades again suggests that the local chert was a relatively soft material.

The frequent occurrence of harpoon end blades (28.1%) and blades/blade-like flakes (13.5%) in the collection argues for a considerable amount of hunting and butchering at the Stock Cove site. At the same time, artifacts associated with tool manufacture -- cores, core debitage, tip-flute flakes, preforms, hammerstones, and abraders -- account for another 32.7% of the assemblage, indicating that manufacturing was also of fundamental importance. This conclusion is reinforced by the large amount of flake debitage present at the site. Objects that might be representative of daily maintenance or processing activities are present -- scrapers, knives/bifaces, soapstone vessel fragments -- and account for 13.5% of the artifacts. Generally, most stone artifacts known to occur in Newfoundland Dorset Eskimo context are represented at the Stock Cove site, and thus the assemblage does not demonstrate any obvious site specialization.