1988 Field Report
(Part 2)
by
Jean-Luc Pilon
NOGAP Archaeologist
Canadian Museum of Civilization

Revisited Archaeological Sites

Field No: 87JLP-24
Borden No: NbTj-8

Recap

This site, located on the north side of the easternmost of two marked narrows in the northeast-southwest trending Hyndman Lake, was first discovered in 1987. Near the tip of the wide point, a small semi-subterranean structure (No.2) was constructed at the edge of a low bench. Fire-cracked rocks have also been observed over a broad area in the vicinity of this feature. Further inland, near the base of the point, over 75 m from the smaller structure is a much larger semi-subterranean feature (Structure No.1). It was constructed on a somewhat higher terrace which lies over 5 m above the waters of Hyndman Lake. Although adze-cut stumps were noted in the vicinity of the semi-subterranean features, the density of these is not as great as has been noted elsewhere. I suspect that the presence of relatively recent historic camps, which were located on the terrace between the two semi-subterranean features, may account for a reduction in the number of old stumps. As Mr. Willie Simon, who like his father before him has trapped extensively throughout the area, so well demonstrated, these old stumps constitute an excellent source of dry wood often used when first arriving at a campsite and a good warm fire is required.

Of the two semi-subterranean features identified at this locality, Structure No.1, the largest, retained our attention in 1987. At that time, this structure was considered sufficiently distinct from all other semi-subterranean features in the southwest Anderson Plain because of its surface dimensions to warrant spending 12 days excavating perpendicular, exploratory trenches through the feature.

This initial investigation revealed sections of a collapsed roof which had covered an area measuring approximately 8.5 m x 5.5 m. Large poles were lying under up to 1 m of sand which formed a berm around a large depression evidently dug into the terrace by the occupants. A radio-carbon age determination obtained from one of the collapsed poles proved to be 610±90 B.P. (S-2941). Testing within the depression was greatly impeded by permafrost and ground water seepage.

The trenches also revealed the existence of an organic level which had been buried by the collapsed roof debris. A wood sample taken from this layer was radio-carbon dated to 555±90 B.P. (S-2942). Where this level had been exposed along the west wall of the semi-subterranean structure (Units E-8,9 and F-10,11), artifacts were recovered which suggested an antiquity greater than that indicated by the radio-carbon date.

Goals

By returning to conduct additional excavations we had two specific goals in mind. The first was to obtain additional details pertaining to the construction of the semi-subterranean structure, especially from within the depression. Clearly we would again encounter permafrost and ground water problems. However, it was hoped that by opening up larger areas, channelling the drainage within these units to sterile corners which could act as reservoirs, and frequently bailing these, we would manage to reach the floor of the structure. In this way we could obtain data permitting us to elucidate the function of the structure.

The second objective was to learn more about the earlier occupation which was associated with the buried organic level. The limited collection which had been recovered indicated a bifacial industry of some kind was present. Additionally, a few calcined bone fragments were found in the western trench (F-10,11). However, the most enticing item recovered was a spalled burin found between two of the collapsed roof poles, just above the organic level, in the eastern trench (E-3,4). Comparisons, such as could be made with a single specimen, suggested technological parallels with the Arctic Small Tool tradition. Any information we could obtain relative to such an occupation was therefore of great interest.

Excavation Techniques

The first task consisted of re-establishing the 1987 grid axes. This was easily achieved since the backfill had compacted somewhat, revealing the edges of the units. The first units were laid out so as to extend the 1987 excavations to the east and west where, as pointed out above, earlier remains had been encountered. In addition, these first units also promised complementary information relating to the structure's construction. Later units were established based on an evaluation of the information obtained from these first squares, as well as the possibilities of expanding our understanding of particular problems relating to the earlier occupation or the late prehistoric use of the terrace.

We excavated in large units both (2 m x 2 m or 1 m x 2 m) within the depression and in areas surrounding it. This provided a better chance to control drainage and obtain a more comprehensive understanding of distributions, especially of the wood remains, during the excavation.

The 1987 excavations had indicated that some faunal remains were to be found within the sod and in the upper portion of the underlying berm deposits. These berm deposits, however, were interpreted as essentially disturbed and unlikely to provide any substanial data relating to the major questions being investigated. As a result, the sod and the upper 10-15 cm of the berm deposits were trowelled and screened through a 1/4 in. or 1/8 in. mesh. All recoveries were plotted on separate floor plans for each natural level and usually bagged by 1 x 1 m quadrants. Below surface measurements were taken from an identified corner of each unit. These elevation reference points were later related to each other with a transit.

The berm sands underlying the uppermost 10-15 cm were shovelled and screened. Trowelling was again resumed when proximity to the buried organic level was indicated. Subsequent recording reverted to the original method of plotting and level bagging for faunal remains and lithic debitage. Precise coordinates were recorded for worked items. All soil stains and wood remains were drawn to scale on the pertinent level floor plans. Lastly, daily field notes were kept by each excavator.

Stratigraphy and Vertical Distributions

The 1988 excavations did not expose any new or previously unrecorded major stratigraphic units. Rather, details concerning the nature of known units were obtained which permit a better understanding of the various soil layers and the events they represent.

The general impression resulting from a preliminary surface examination of Structure No. 1 is of a depression surrounded by a ridge of earth. The approximate maximum dimensions of the crest of the oval to rectangular ridge are 8.5 m x 5.5 m. However, our excavations show that the dimensions of the internal depression are much less than those of the earthen ridge or berm. The sod or vegetation layer (Level I) which covered the berm deposits usually consisted of a relatively thin layer (usually on the order of 3-5 cm) of grasses and mosses. A few faunal elements were recovered from this layer. However, it was in the uppermost 10 cm of the underlying soil strata, the berm deposits (Level II), that significant numbers of animal bones were recovered in all units where the berm was present. If the ultimate source of the faunal remains was the older occupation, then these should have been more evenly distributed throughout the berm sands. Since this was not the case, we must conclude that these animal bones are more likely derived from the semi-subterranean occupation. Indeed, a moose antler fragment found in the berm deposits in A,B-6 in 1987 fit onto a larger section found on the floor of the depression in D-6 in 1988.

As we excavated further into these sands, which reached up to 1 m in depth in some units, artifacts of any description were quite few in number. When encountered, these were often found associated with humic lenses which are strongly suggestive of large pieces of sod. Such lenses were noted at different depths within the berm deposits. These clods of sod were characterised by 3 to 5 cm of humic material, sometimes containing fibrous inclusions, associated with a layer of fine, sometimes slightly reddened sand. Such a succession is encountered in undisturbed areas surrounding the semi-subterranean structure as well as under the berm deposits. They are thus interpreted as clods of sod cut by the builders of this feature, and incorporated into its roof.

Artifact Distribution by Stratigraphic Unit
(1987 and 1988)

DebitageFaunal
Remains		Tools
Level I 0300
Level II (berm)2963813
Level III (buried humus)26737317
Level IV (basal sand)24452111
Depression Floor 01011
Sod/Humus where no berm978024
Basal Sand4044377
Totals1922211073


The berm deposits everywhere rested upon a dark organic layer (Level III). This relatively thin layer (3-5 cm) was shown to grade into the modern humus along the outer edge of the berm in the western excavation units. Artifacts were quite numerous within this layer virtually everywhere it was encountered. Lying slightly above or on this organic level were a large number of decomposing wooden poles. Their arrangement suggests that they represent a collapsed superstructure. Thus Level III may represent the former vegetation/humus layer upon which the roof of the semi-subterranean structure eventually collapsed, but it almost certainly also includes organic debris deposited during the use of the semi-subterranean structure.

In the westernmost portion of the excavated area (units G,H-13 and I,J-12 to 16) there were no berm deposits encountered. Instead, the thicker vegetation layer (7-10- cm) was underlain by a thin humus (also denoted as Level III) in which a very important quantity of artifacts was recovered.

The following soil layer, regardless of the location of the unit in relation to the berm, was a fine yellow sand (Level IV). In some units a coarser basal layer was eventually encountered which included very high quantities of well-rounded cobbles.

Within the depression proper, the vegetation consisted of a thick layer of sphagnum moss (up to 30-40 cm) lying on a poorly-drained fine sand/silt deposit (Level IIa). Numerous poles were found throughout these deposits and were much better preserved than those found under the berm. The arrangment of the poles within the depression indicates that they were part of the same super-structure as those found around the depression. Permafrost had been encountered in the depression fill at a depth of 10-15 cm during the 1987 excavations. As a result water seepage was a constant problem and the bottom of the original depression was never reached. Our efforts were specifically directed toward exposing the floor of the depression in at least one small corner during the 1988 excavations. As a result, we can assert that the thick sand/silt deposit (Level IIa) can attain up to 50 cm in depth along the sides of the depression.

Horizontal Distributions

The distribution of wooden poles, lithic debitage and implements, faunal remains, fire-cracked rocks and fire-reddened soil for the western portion of the site have been illustrated separately. The central and eastern areas contained very little in terms of artifacts and did not warrant the production of floor plans.

The floor plans of the western portion of the excavation show the distributions of materials found in the buried humus horizon and in the sod/humus and underlying sand where no berm deposits were encountered.

Extant Wood

The 1988 excavations exposed additional wood which is interpreted as having formed the superstructure which covered the semi-subterranean feature. However, in attempting to reproduce the floor plans of the various excavation units from both field seasons, it was necessary to adjust slight misalignements which had occurred between units. This difficulty stemmed for the most part from the very significant depths in some units which resulted in parallax distortions.

Since it now appears that the 1987 exacavations did not reach the bottom of the interior depression, a great deal more wood must certainly remain in the depths of the 1987 depression units, namely C,D-7 and F,G,H-7. Additionally, it was impossible to show all the poles found in units C,D-6 since there were simply too many. However, these additional layers of poles were oriented in a similar fashion as those above. The deepest course was comprised of smaller poles and branches along with a thick deposit of spruce needles, suggestive of a bough floor.

In the eastern 1987 excavation units (Unit B,E and I), the relatively smaller diameter poles appeared twinned, i.e. lying in pairs. The 1988 excavations in the adjacent units failed to exhibit the same pairing except in one instance. This situation is likely due to the poor state of preservation of the wood which may not have allowed the excavator to distinguish the separate poles as they lay side by side. However, for consistency, where the pairing was not observed in 1988, the 1987 pairs were combined as a single pole.

The poles were often flattened in cross-section, and an examination of the end grain demonstrated that many had been split lengthwise. Charring of the exterior surfaces was noted in many instances as well. Evidence of post-occupational burning was quite pronounced in C,D-6 where large amount of charcoal were found between the jumble of collapsed poles. Some of these better-preserved poles reached diameters in excess of 15 cm.

The plan of the extant wood permits a number of observations. A first one relates to events following the collapse of the superstructure. In particular, the poles lying within approximately 1 m of the edge of the interior depression are no longer straight but bow-shaped. I suspect that this deformation process reflects the gradual response of the wood to the very significant weight of the berm deposits and the slight inward slope of the buried humus adjacent to the depression.

An obvious feature of the disposition of the poles, taking into account the possible deformation mentioned above, is the alignement of these parallel to each other, but slightly askew from the axis of the interior depression. Once again, I would propose that this sitution results from the collapse of the superstructure. If the suggestion that these poles were supported by A-frame supports--minimally one at either end, but more than likely additional supports between the ends--is valid, then the collapse could have been brought about by failure of one of these and the whole structure would have twisted as it fell, thereby changing the alignement of the poles in relation to the interior depression if they were in fact the same prior to the destruction of the structure.

The excavation of additional units to the south of the 1987 units illustrates quite well that there were few poles associated with the ends of the structure, although it must be recalled that there was much evidence of burnt poles in those areas as well. While poles were very well preserved under the thick berm deposits in units J,K-8,9, only fragmentary charred sections were found in unit I-7 and unit A-6. As suggested in a previous report (1987), the poles at the extremities may not have as deeply buried under roof sands if at all, and thus more exposed to the ravages of a forest fire.

Fire-Reddened Soil

Numerous distinct areas of what are interpreted as fire-reddened sand were found throughout the excavations. Such reddening of sand was not limited to Level IV. Indeed, reddened sands were also found in some portions of the berm deposits. If the prime cause of such distinctive colouration is exposure to intense heat, it is not inconceivable that the staining noted in the berm deposits as well as some of those reddened sands of Level IV, relate to forest fires. The localized areas of fire-reddened sand identified in the southwestern part of the excavation are interpreted as areas of combustion since there is usually additional evidence of human activity involving the intentional use of fire. In particular, calcined animal bones and fire-cracked rocks clearly cluster about such features. Additionally, charcoal is often found within these deposits and surrounding the stain.

Two particularly intensive combustion areas appear identified in the western portion of the excavation. The first zone is found in the southeast corner, within the area which was likely inside the structure covered by the poles and sand. It consists of three large fire-reddened stain and a very significant number of fire-cracked rock. Calcined bone was associated with the northern stain only.

To the west, four fire-reddened sand stains were recorded. The western-most pair may in fact be part of a single combustion area. The very deep fire-reddened sand feature in I,J-14 contained much calcined bone and charcoal, of which a sample has been submitted for radio-carbon age determination. The remaining two stains are much smaller and may or may not be associated with the first two. Fire-cracked rock were especially numerous in proximity to the small stain in unit J-13.

Lithic Debitage and Implements

The distribution of lithic debitage in the southwestern part of the excavation are instructive. Descriptive information and basic measurements of the worked lithic artifacts can be found in Appendix A and their distribution is provided.

Quite obviously there is a very high concentration of material in proximity to, and definitely associated with, the two large centres of combustion in units I,J-13 to 15.

This situation is much different elsewhere. It is especially interesting to note the marked reduction of lithic debitage and implements in proximity to the three large fire-reddened areas in units H-8,9 and J,K-8,9. On the other hand, a very significant number of fire-cracked rocks and boiling stone were found associated with those centres of cumbustion. Significant concentrations of lithic debitage are once again recorded just west of the interior depression in units F,G,H-10 through 13. However, these are much less dense than the westernmost concentration. Moreover, it is tempting to separate the two areas as there appears to be a hiatus between them, on the order of one meter in width.

In this regard it is interesting to note that the very tight and dense concentration in the western portion of the excavation is well outside the area which may have originally been within the confines of the proposed semi-subterranean structure (see Discussion). On the other hand, the much looser concentration and three major fire-reddened soil stains occur within the suggested limits of the superstructure. On this basis alone, it can be suggested that although the material found within the proposed confines of the semi-subterranean structure cannot be associated with either occupation with any degree of certainty, it can be argued that the western concentration is not likely linked to the more recent occupants who constructed the semi-subterranean structure.

The nature of the implements associated with each of the centres of combustion reinforces the apparent separation noted above regarding distinct artifact clusters. For example, all 11 of the burin spalls were found associated with the westernmost hearth areas. Although only 3 were found in situ, the remaining 8 specimens were found in flake lots derived from units I,J-13,14. Similarly, all but two of the linear flakes/microblades and corner flakes (primary or secondary, elongated flakes removed from the corners of blocky cores) were found within the I,J units. The only truly diagnostic implements found in these excavations, the finely flaked, bipointed projectile point and two of the three burins, were also found in proximity to the elongated hearth feature in I,J-14.

Cores essentially consist of two types; cobble cores and tabular cores. Although they were more numerous in the G,H units, their numbers are not sufficient enough to discern any meaningful distributional tendencies.

The pices esquillŽes follow a similar distributional pattern. They are made either from thick cobble cortex flakes or thick blanks which required bifacial preparation. Two of the six specimens were found in proximity to the elongated hearth in I,J-14 while the others were found under berm deposits and presumably within the semi-subterranean structure. There is no tendency for either type to be found exclusively in one part of the site or another.

Of the 5 scrapers found at this site, two were recovered from H-10,11, while two others were found within the berm deposits. The last specimen, a tiny end scraper, was found in a test unit approximately 7 m to the SW.

The single hammerstone identified at the site was found in association with the centre of combustion in J,K-8,9. The three notched implements (spokeshaves ?) were found in the berm deposits.

Discussion

From the above discussion it should be apparent that the 1988 excavations achieved their stated objectives, namely to acquire additional information relating to the construction of the semi-subterranean feature and obtain confirmation of the existence of a suspected earlier occupation.

Semi-Subterranean Structure

The edges of the depression have been exposed on four sides and these have been extrapolated to produced an approximate outline. It thus consisted of a rectangular depression measuring 3 m x 8 m, excavated at least 50 cm into the terrace. The bottom was apparently lined with branches and small poles, and covered with spruce boughs. This depression, in addition to a large area surrounding it, were covered over with a substantial superstructure. This roof consisted of long spruce poles lying roughly parallel to the long axis of the depression, and presumably supported by wide-based A-frame pole constructions. The poles were covered over with the sand taken from the depression, undoubtedly as insulation. To prevent this fine sand from filtering down between the poles, clods of sod were apparently placed on top of the poles before being covered over with sand. A premise to this discussion is that if the layer of sand which covered the poles was indeed placed there to act as an insulating material, then it must have been layed down more of less evenly from the base of the roof to its apex. Openings for the evacuation of smoke would not alter this line of reasoning.

Between the last use of the structure and its collapse, there was sufficient time for the combined forces of gravity and erosion to act upon the inclined, sand-covered roof. As a result, most of the sand found its way to the base of the roof where the greatest accumulations would occur. Thus, the point at which the berm is thickest can be taken as the approximate point at which the roof met the ground. It is in this manner that the exterior edge of the superstructure has been drawn. Thus, the maximum dimensions of the semi-subterranean structure were 9 m x 10 m or approximately 90 square metres.

Although no additional details concerning the possible entrance(s) came to light with these new excavations, the earlier hypothesis still appears valid. In particular, it was found that the berm deposits at either end of the depression were much thinner than had been documented along the sides. Additionally, there appeared to be fewer wooden structural remains, and these could have tumbled in from the sides. It is interesting to examine the distribution of fire-reddened sands. A central question is whether these features were contemporneous with the occupation of the late prehistoric semi-subterranean structure or whether they were produced by earlier occupants of the site. In this regard we can consider the stratigraphic provenience of artifacts recovered outside the approximate limits of the semi-subterranean structure compared to those found within the presumed structure.

Artifacts were concentrated in the humus (Level III) and into the upper portion of the underlying basal sands (Level IV). It has been pointed out that the modern humus outside the limits of the berm are continuous with the organic level buried by the berm deposits (also identified as Level III). If we examined the various profiles which cross-section the berm at different points, it will be noticed that the buried organic level always dips down toward the depression, usually starting from about the point of maximum berm thickness. Could we therefore infer that the sod which covered most of the area inside the structure had in fact been removed. If the various clods of sod found in the berm deposits are an indication, subtantial portions of the underlying basal sand remained attached. It is thus quite possible that most of the organic layer found under the berm relates to the late prehistoric occupation rather than the earlier occupation. Similarly, the two large fire-reddened sand stains may also relate to the most recent occupation. In this case, it is interesting to note their locations near the presumed entrances to the structure. Additionally, the recovery of a single fire-cracked rock in C,D-6 suggest a hearth may also have been located in the north end of the structure. Finally, I think it is now possible to put aside for the moment the heavy descriptive term, semi-subterranean structure, in order to consider the actual use of this feature. It is clear that a substantial amount of time and energy was expended in constructing an inclined roof over not only a long and relatively narrow depression, but also a large area surrounding it.



Moose antler wedged under collapsed roof poles in the center of the depression.




Caribou rib fragments found in the remains of the spruce bough flooring of the depression.



Centres of combustion were located within the structure and refuse accumulated on the roof as well as on the floor of the depression and on the surrounding bench. Entrances were located at either end, facing away from the direction of predominant winter winds (the northwest). The combination of all the foregoing information strongly suggests that this semi-subterranean feature was used as a dwelling rather than as a cache. In this regard it was interesting to note Mr. Simon's interpretation of the remains we were exposing. A major point he brought up was that if this feature was destined to be used as a meat cache, then sand, which could easily spoil meat, would not have been used.

Earlier Occupation

It has been shown above that a cluster of lithic artifacts and hearth features with associated calcined faunal remains was located just west of the inferred limits of the late prehistoric pit house. In all likelihood, these recent occupants incorporated sod into the roof of their dwelling which contained artefacts left behind by the earlier occupants. However, given the discreetness of the archaeological distributions in I,J-13,14, and the fact that this area was along the side of the late prehistoric dwelling rather than at either end where the entrances are thought to have been located, we can be fairly certain that there has been little component mixture in that area. As such, the following discussion will proceed based on this assumption.

The lithic assemblage consists of a number of very diagnostic items which permit us to attribute the collection to the Arctic Small Tool tradition. This includes a bipointed projectile point exhibiting parallel, oblique collateral flaking with fine serration of the distal margins and grinding of the proximal edges, three spalled burins, and 11 burin spalls.

Two other categories of lithic remains must also be considered; linear flakes/microblades and corner flakes. A number of flakes exhibit the traits usually considered sufficient to call them microblades. They are long, relatively narrow, parallel-sided flakes with previous flakes scar similarly oriented. They were struck from cores whose platform angles approached right angles. Of the 18 flakes so described, only two have more than two dorsal flakes scars and 8 were struck from the edge of blocky cores. In other words, there are no `classic' examples of microblades in this collection, nor are there any microblade cores.

A second type of lithic remain which is also found only within the ASTt cluster is termed corner flakes. These are narrow, elongated primary and secondary flakes removed from the corners of blocky cores. Although the removal of such corner flakes is most likely how most lithic craftmen would approach blocky or tabular cores and nodules, these flakes were only found in the ASTt area. When combined with the linear flakes/microblades, these corner flakes may be instructive about the entire process of microblade manufacture. It is unfortunate that little attention is paid to the less than perfect microblades or the long flakes removed both before and after the `classic' microblades. In other words, we know little about the biproducts removed in preparing the core for microblade production, and those removed after the optimum configuration of the core for microblade production (size, striking angle, potential for rejuvination, etc.) has been passed.

In the present case we lack `good' microblades, but the linear flakes, `bad' microblades and corner flakes all bespeak of distinct efforts toproduce microblades. Given the fact that the ASTt area was not intensively occupied then it is not inconceivable that the `good' microblades were removed from the site or lie in some other sector of the terrace. However, I believe there is sufficient circumstantial evidence to indicate they were beingmanufactured. Given the fact that microblades are a common recovery from ASTt sites elsewhere in the Arctic, this should come as no surprise.

Finally, the charcoal sample gathered from the hearth feature in I,J-14 was determined to be 3390±255 B.P. (S-3000). This age indicates that the use of the Hyndman Lake area by ASTt people did not coincide with the occupation of the Barrenlands by the bearers of this archaeological culture. Rather, it suggests a much earlier occupation of the region at the same time that the High Arctic was being populated by this group. Their presence so far inland--although the Eskimo Lakes existed at this time, the Beaufort Sea coast was much further north than its present location--clearly attests to a well-developed adaptation to the resources of the interior which could only have been acquired by seasonal use of the region over a long time span, and not by migrating populations.

This site, along with three others in the southwest Anderson Plain (NbTj-9, NbTj-17 and MlTk-2), provided information which served as the basis for an analysis of the Arctic Small Tool tradition in the region.