LITHIC DEBITAGE AND REDUCTION STRATEGIES AT SMOKEHOUSE ISLAND
(GiSp-001), ON THE BABINE RIVER, NORTH CENTRAL BRITISH COLUMBIA.

by

Courtney John Lawrence

B.A. MacEwan University, 2017

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ARTS
IN
INTERDISCIPLINARY STUDIES

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
July 2023

©Courtney John Lawrence 2023

Abstract
The Babine Archaeology Project began in 2010 and since then, many stone artifacts have
been recovered. In 2014 and 2015, the project focused on excavations on Smokehouse Island
(GiSp-001), on the Babine River. The focus of this thesis is an analysis of a sample of debitage
recovered from Unit 8 excavated in 2015. The goal is to determine if differences are evident in
how the inhabitants of Smokehouse Island reduced the different stone raw materials. Mass
analysis, a modified version of the Sullivan and Rozen technique, and Magne’s scar count
method were used to analyze the materials. Results suggest that core reduction is dominant for
the two different raw material groups, though tool production is still present for both.
Technological mixing impacts the analysis of the different materials.

ii

TABLE OF CONTENTS
Abstract ........................................................................................................................................... ii
TABLE OF TABLES ..................................................................................................................... v
TABLE OF FIGURES ................................................................................................................... vi
Acknowledgements ....................................................................................................................... vii
Chapter 1: Introduction ................................................................................................................... 1
Chapter 2: Natural, Cultural, and Archaeological Background of the Study Area......................... 5
Geographical Location .............................................................................................................................. 5
Cultural background................................................................................................................................ 13
Modern Scholarship on the Lake Babine Nation .................................................................................... 17
Athapaskan, Carrier and Babine use of stone material ........................................................................... 21
Regional Archaeological Review............................................................................................................ 23
Previous archaeological work and the Babine Archaeology Project ...................................................... 36
Smokehouse Island ................................................................................................................................. 38

Chapter 3: Previous Debitage Research and Methodology .......................................................... 43
The History of Debitage Analysis in British Columbia .......................................................................... 43
Summary ................................................................................................................................................. 51
Mass Analysis ......................................................................................................................................... 53
Sullivan and Rozen Technique (SRT) .................................................................................................... 56
The Flake Scar Count Method ................................................................................................................ 59

Chapter 4: The Debitage Sample, Application of Analytical Techniques, and Results ............... 63
Characterization and Identification of Lithic Raw Materials .................................................................. 63
iii

Mass Analysis Application and Results .................................................................................................. 67
Sullivan and Rozen Technique (SRT) Application and Results ............................................................. 72
The Scar Count Method Application and Results ................................................................................... 74

Chapter 5: Discussion ................................................................................................................... 80
Discussion of Results and Interpretations ............................................................................................... 80

Chapter 6: Conclusion................................................................................................................... 90
Bibliography ................................................................................................................................. 96

iv

TABLE OF TABLES
Table 1. Comparison of chosen methods, including Mass Analysis (Ahler 1989a), the Sullivan and Rozen
(1985) Technique with modifications by Prentiss and Romanski (1989), and Magne's (1985) Scar Count
Method. Some table categories from Steffen and others (1998:137) and Larson (2004:6) adapted for table.
.................................................................................................................................................................... 52
Table 2. Frequency Distribution for FGV1. ................................................................................................ 68
Table 3. Frequency Distribution for FGV2. ................................................................................................ 69
Table 4. Percentage of weight by size grade for FGV1. ............................................................................. 70
Table 5. Percentage of weight by size grade for FGV2. ............................................................................. 70
Table 6. Cortex profiles by size grade for FGV1. ....................................................................................... 71
Table 7. Cortex profiles by size grade for FGV2. ....................................................................................... 71
Table 8. Frequency distribution by modified SRT hierarchy for FGV1. .................................................... 73
Table 9. Frequency distribution by modified SRT hierarchy for FGV2. .................................................... 73
Table 10. Frequency distribution based on Magne’s (1985) scar count method for FGV1. ....................... 75
Table 11. Frequency distribution based on Magne’s (1985) scar count method for FGV2. ....................... 75

v

TABLE OF FIGURES
Figure 1. Image derived from Google Earth Pro and others (2015), showing Smokehouse Island in
relation to some key locations in British Columbia. ..................................................................................... 6
Figure 2. Image derived by Google Earth Pro and others (2023) showing Smokehouse Island in relation to
the village site Nass Glee (to the north) and Fort Kilmaurs (to the south). .................................................. 7
Figure 3. Image derived from Google Earth Pro and others (2022a), showing the position of Smokehouse
Island in relation to Nilkitkwa Lake (to the north) and Babine Lake (to the south). .................................... 8
Figure 4. Image derived from Google Earth Pro and others (2022b), showing Kitsuns Creek Formation
(purple), Red Rose Formation (light blue), Paleocene to Eocene (red), Lower to Upper Cretaceous Rocky
Ridge Formation (yellow), and Middle to Upper Jurassic Bowser Lake Ashman Formation (light pink)
deposits; Geological information from MacIntyre and colleagues (2001a):
https://cmscontent.nrs.gov.bc.ca/geoscience/PublicationCatalogue/OpenFile/BCGS_OF2001-03.pdf. .... 11
Figure 5. Map of the excavation units on the south end of Smokehouse Island. Units 1-4 were opened and
excavated in 2014, while units 5-8 were opened and excavated during the 2015 field season (Rahemtulla,
In prep). ....................................................................................................................................................... 39
Figure 6. Profile of the west wall of Unit 8, with layer descriptions (Rahemtulla, In prep). ...................... 41
Figure 7. Combined photos showing a range of FGV1 samples................................................................. 65
Figure 8. Combined photos showing a range of FGV2 samples................................................................. 66

vi

Acknowledgements
I want to thank many people for their help with my thesis journey. First, I would like to
thank the Lake Babine Nation for letting me analyze artifacts recovered from their traditional
territory. I would like to thank Dr. Farid Rahemtulla for being my supervisor, for sharing his
knowledge with me, and for his eternal patience and guidance as I worked on my project. I must
also thank my co-supervisor, Dr. Ted Binnema, and committee member, Dr. Rudy Reimer
Yumks. Both individuals have been a source of positive inspiration and have been very
supportive and helpful. Thank you to my external examiner, Dr. Jesse Morin, for providing a lot
of excellent advice to improve my thesis. And thank you to Dr. Neil Hanlon for chairing my
defense.
Thank you to Dr. Paul Prince for providing numerous important and rare archaeological
reports, and for always being helpful and supportive. I would like to give a big thank you to Lisa
Mutch, Sandi Ratch, Kathy Gadd, Steph Skelton, and Katrina Forbes for their help and advice
during the time of thesis edits. I would also like to thank Dr. Michel Bouchard, Dr. Kyle
Forsythe, Peter Stewart, Dr. Tara Joly, and Samantha Walker for their advice on the thesis
process. I also must thank my mother, Catherine Lawrence, for her support during my Master’s
journey. And lastly, I would like to thank some of my other supportive friends, including Katrina
Forbes, Lysander Sorenson, Diana Hoang, Remington Tensfeldt-Jones, Reza Ta, Fatemeh
Mohammadnejad, and so many others.

vii

Chapter 1: Introduction

The Babine Archaeology Project began in 2010 as a collaboration between the Lake
Babine Nation and the University of Northern British Columbia, also known as UNBC
(Rahemtulla 2012:5). Integral to the project is providing community members and UNBC
students with archaeological training and having community involvement in the overall research
(Rahemtulla 2012:5). Additionally, the project strives to incorporate Traditional Knowledge into
the archaeological process, and community members have familial ties to the study region
(Rahemtulla 2012:6).
Initially, this project focused on the Lake Babine Nation village site of Nass Glee
(Rahemtulla 2012, 2013). In 2014 and 2015, the focus switched to excavating Smokehouse
Island (GiSp-001), on the Babine River (Rahemtulla 2019:159). A large volume of chipped stone
artifacts and debitage was recovered during both field seasons (Rahemtulla 2019:163-164), and
the discovery of wet site deposits further below. An ongoing research question for the larger
project involves sourcing the stone materials, but at present, the location of these sources is
unknown (Farid Rahemtulla, personal communication 2020). Concomitantly, there is evidence
that stone materials from multiple sources are being used (see Chapter 3); it is important to know
if the inhabitants of the island worked them differentially. That is the focus of this thesis.
In this research, the null hypothesis is that there are no differences in the reduction of
stone materials from multiple unknown sources. A sample of Smokehouse Island debitage is
analyzed to determine if different raw materials show evidence of differential reduction

1

strategies. Unit 8 was chosen specifically as a sample location because it contained the least
amount of debitage from the 2015 excavations, providing a workable timeline for this analysis
(Rahemtulla, personal communication 2020). Unit 8 is one of eight units excavated during the
2014 and 2015 field seasons on the south end of Smokehouse Island and additional excavations
have occurred in 2017 and 2019 (Rahemtulla, personal communication 2019)
The current research is the first formal debitage analysis conducted on the Smokehouse
Island material. The central hypothesis regarding Smokehouse Island is that it was constructed
by the Indigenous occupants rather than naturally formed (Farid Rahemtulla, personal
communication 2018). Additionally, various activities, including the preparation of salmon in
smokehouses, have occurred on the island for at least 1000 years (Rahemtulla 2019:162-163).
These factors reveal the potentially unique nature of Smokehouse Island in the region, and makes
analysis of the stone debitage crucial for understanding the site history.
To proceed with the analysis, several steps were undertaken. First, a sample of the
debitage was submitted for Portable X–ray fluorescence spectrometry (PXRF) characterization
(Reimer 2011:128, 2018, 2023) to establish the potential number of lithic raw materials used at
the site. Once the results were received, multiple forms of debitage analysis were applied to the
debitage sample. Using multiple lines of evidence has been suggested by some researchers to
compensate for any flaws inherent to the individual forms of analysis (Bradbury and Carr 1995;
Carr and Bradbury 2001:129, 134; Carr and Bradbury 2004:43; Eren and Prendergast 2008:51;
Larson 2004:16-17). Carr and Bradbury (2001:129) stated that “[e]mploying multiple lines of
evidence strengthens inferences or reveals ambiguities.”

2

However, using multiple lines of evidence in the form of multiple methods of debitage
analysis has faced some criticism (Magne 2001:23-24; Jesse Morin, personal communication
2023). Magne (2001:24) argued that incorporating multiple forms of debitage analysis into his
experimental program from the 1980s would not have increased the significance of his results
drastically (Magne 1985). Jesse Morin (personal communication 2023) perceived using multiple
methods of debitage analysis as being overkill, and that it is using one line of evidence in
multiple ways instead of multiple lines of evidence. This makes the current research an
opportunity to compare methods, something which Magne (2001:24) stated has not been done
enough in lithic analysis. Multiple methods are incorporated into this current research, though
with caution, and with the possibility of comparing and assessing the effectiveness of each
method. The chosen analyses include mass analysis (Ahler 1989a), the Sullivan and Rozen
(1985) technique with modifications by Prentiss and Romanski (1989), and Magne’s (1985) scar
count method. Previous research suggests the chosen methods work effectively on assemblages
in British Columbia (Hall 1998; Magne 1985; Matson and Magne 2007; Prentiss 1993; Kuijt et
al. 1995; Rahemtulla 1995a, 2006; Ritchie et al. 2022).
Following this introductory chapter, Chapter 2 provides the geographical and cultural
background of the Smokehouse Island site. A review of recent scholarship on the Lake Babine
Nation is presented, followed by a summary of known Athapaskan, Carrier, and Babine stone
tools, and a review of archaeological work. The archaeological review details archaeological
work focused on known Carrier and Athapaskan sites in northern and central British Columbia,
which are in Subarctic British Columbia (Fladmark 2009:557). The review also includes
previous survey and excavation work along the middle and upper Skeena River and its Bulkley
River tributary, which are part of the Skeena-Bulkley-Nass archaeological region, as defined by
3

Fladmark (2009:558). Lastly, previous archaeological work in the Babine region and the Babine
Archaeology Project are summarized (Mohs 1974, 1975; Mohs and Mohs 1976; Rahemtulla
2012, 2013, 2019, In prep).
Chapter three reviews previous applications of debitage analysis in British Columbia,
determining appropriate methods for the current study. The fourth chapter presents the
application of the methodologies and results from each method. The fifth chapter discusses how
the results are interpreted. Lastly, the conclusion is presented in Chapter six. While definitive
conclusions for the overall archaeological site cannot be stated, as only part of the assemblage
from Unit 8 was analyzed (Farid Rahemtulla, personal communication 2022), key findings are
evident in the results of this analysis.

4

Chapter 2: Natural, Cultural, and Archaeological Background of the Study
Area

This chapter details the geographic location, geomorphology, and natural resources of the
Babine region, followed by the cultural background of the Lake Babine Nation. The recent
scholarship focused on the Lake Babine Nation is then reviewed, followed by a summarization
of stone tools associated with Athapaskan and Carrier groups. A review of archaeological work
conducted on Athapaskan sites from within Subarctic British Columbia (Fladmark 2009:557),
and specific sites in the Skeena-Bulkley watershed are reviewed to build an understanding of
how Smokehouse Island fits into the known histories of the region. Lastly, previous
archaeological work conducted at Babine Lake, including the Smokehouse Island site, is
described.

Geographical Location
The Traditional territory of the Lake Babine Nation includes the regions associated with
the Babine River and Babine Lake in Northern British Columbia (Rahemtulla 2019:159-160).
Babine Lake is located on the Nechako Plateau of the interior region of British Columbia
(Hackler 1958:1; Mohs 1974:1). An archaeological village site at Nilkitkwa Lake (GiSq-004) is
on the east bank of the Babine River (Rahemtulla 2012:6; Rahemtulla 2013:8), north of the town
of Burns Lake (Hackler 1958:1; Mohs 1974:1). Currently, Fisheries and Oceans Canada
administers a fish counting fence at a facility at the village site (Rahemtulla 2019:161).

5

The focus of this research is Smokehouse Island (GiSp-001), which is on the south end of
Nilkitkwa Lake (Rahemtulla 2019:160), north of and close to the outflow of Babine Lake. The
site is in the middle of the Babine River (where it transitions into Nilkitkwa Lake), and it is
approximately 150m by 150m in size (Rahemtulla 2019:162).

Figure 1. Image derived from Google Earth Pro and others (2015), showing Smokehouse Island in relation to some
key locations in British Columbia.

Geographically, Babine Lake is located in the Sub-Boreal Interior Ecoprovince, which is
east of the Coast Mountains and west of the Interior Plains (Demarchi 2011:73). Within the
larger Ecoprovince, Babine Lake is in the Babine Upland Ecosection (Demarchi 2011:82). The
Ecosection is characterized by rolling highlands and hills with low ridges being the most
common forms of terrain, with several very large lakes and many small streams and wetlands
(Demarchi 2011:82). While some parts of the lake are located in the Interior Western Hemlock

6

zone, most of Babine Lake lies within the Sub-Boreal Spruce biogeoclimatic zone (Farley
1979:49; Fladmark 2009:559).

Figure 2. Image derived by Google Earth Pro and others (2023) showing Smokehouse Island in relation to the
village site Nass Glee (to the north) and Fort Kilmaurs (to the south).

7

Figure 3. Image derived from Google Earth Pro and others (2022a), showing the position of Smokehouse Island in
relation to Nilkitkwa Lake (to the north) and Babine Lake (to the south).

8

Approximately 43,000 years ago, pollen and peat were present at Babine Lake, indicating
the area was unfrozen at that point in time (Matthews 1979:149). The initial expansion of ice
centers and climate cooling in British Columbia occurred at the onset of the Fraser Glaciation
(29,000 to 25,000 years before present or BP) (Stumpf et al. 2004:226). Most of the initial
movement of glaciers began during the Late Wisconsinan Glaciation from approximately 25,000
BP until approximately 10,000 BP (Stumpf et al. 2000:1850). The ice flow during this period
“occurred in three main phases: (1) ice expansion phase, (2) maximum phase, and (3) late glacial
phase (Stumpf et al. 2000:1856)”. During the ice expansion phase, glaciers moved from
mountainous regions towards coastal and interior regions (Stumpf et al. 2000:1857). A
southeasterly ice flow is known for the Babine Valley during the ice expansion phase, which was
followed by a westward flow in nearby regions including Morice river valley (Stumpf et al.
2000:1853, 1857-1858, 2004:219). The second phase, the maximum phase, is generally
characterized by greater ice movements inland (Stumpf et al. 2000:1858). In contrast, regions
east of the Babine Valley experienced a westward ice flow towards the Pacific Ocean (Stumpf et
al. 2000:1858). During this phase, ice flowed west through the Babine Lake valley and towards
the Coastal Mountains (Levson 2001:743; Levson and Stumpf 1998).
Following the maximum phase, two general trends of ice flow are evident for the last
glacial phase: 1) there was a general movement of glaciers from the interior towards the Coast,
Hazelton, Omineca, and Skeena mountains, and 2) ice centers continued in the Interior for areas
east of the mountains (Stumpf et al. 2000:1859). During this phase, there was a shift from
westward glacier movement to eastward movement through the Babine and Takla valleys
(Stumpf et al. 2000:1859-1860, 2004:219). It is evident during all three phases that ice flow
trends were different for local and regional areas due to topography and ice accumulation centers
9

(Stumpf et al. 2000). The remaining glacial lakes disappeared by 9,000 BP, which allowed
spruce and pine trees to flourish in the region (Donahue 1977:21).
During the Holocene period, colluvial, fluvial, and organic sediments were deposited
(Levson 2001:742). However, the oldest exposed sediments along the Babine River are nonglacial fluvial sediments deposited during the Middle Wisconsinan (Levson 2001:736).
Additionally, glacial, glaciofluvial, and glaciolacustrine soils were deposited in the Babine Lake
valley during the Late Wisconsinan Fraser Glaciation (Levson 2001:733, 736). Specifically, the
formation of large glacial lakes in the region allowed for the deposition of glaciolacustrine
sediments (Levson 2001:735). Along the margins of Babine Lake, numerous postglacial fan and
alluvial fan deltas appear, though colluvial deposits are the most common in the Babine region
(Levson 2001:742).
Bedrock for the portion of the Babine Lake region to the immediate north, east, south,
and west of Smokehouse Island are Lower to Upper Cretaceous Kitsuns Creek Formation
deposits containing mudstone, siltstone, shale, and sandstone (see Figure 4; MacIntyre et al.
2001a). Approximately 1.4 km west-southwest and 2.8 km north-northwest of Smokehouse
Island are Lower to Upper Cretaceous Red Rose Formation deposits of siltsone, muscovite, and
chert-pebble conglomerate (see Figure 4; MacIntyre et al. 2001a). Lower to Middle Jurassic
Smithers Formation deposits of limestone, volcanic breccia, and sandstone are located 2.2 km
northwest of the site, while Paleocene to Eocene deposits of rhyolite, argillite, chert, horneblend
porphyry, shale, siltstone, and sandstone are located approximately 2.3 km south-southwest of
Smokehouse Island (see Figure 4; MacIntyre et al. 2001a). Approximately 5.7 km southsouthwest of Smokehouse Island, Lower to Upper Cretaceous Rocky Ridge Formation deposits
are located, which includes dacite, andesite, basalt, rhyolite, chert, shale, and sandstone (see
10

Figure 4; MacIntyre et al. 2001a). Lastly, Middle to Upper Jurassic Ashman Formation deposits
of thin-bedded shale, siltstone, and feldspathic wacke are located 3.3 km southeast of the site
(see Figure 4; MacIntyre et al. 2001a). Some of the materials from the Babine Lake region could
have been used as flakeable tool stones, but this would have to be verified for every outcrop and
source (Reimer 2018:139-140, 2023).

Figure 4. Image derived from Google Earth Pro and others (2022b), showing Kitsuns Creek Formation (purple), Red
Rose Formation (light blue), Paleocene to Eocene (red), Lower to Upper Cretaceous Rocky Ridge Formation
(yellow), and Middle to Upper Jurassic Bowser Lake Ashman Formation (light pink) deposits; Geological
information from MacIntyre and colleagues (2001a):
https://cmscontent.nrs.gov.bc.ca/geoscience/PublicationCatalogue/OpenFile/BCGS_OF2001-03.pdf.

11

However, the raw material recovered from Smokehouse Island could have originated
from countless other areas. Afterall, the use of watercraft can transport large quantities of
material over large distances and help facilitate exchange between groups (Ames 2002:42, 4445; Blair 2010:40, 42; Rahemtulla 2006:314-316). For example, the igneous material could have
been obtained from Early to Middle Jurassic Saddle Hill Formation deposits of andesite, basalt,
dacite, and rhyolite (MacIntyre et al. 2001b:582-583; MacIntyre et al. 2001a). Saddle Hill
Formation deposits appear in many locations within the larger Babine-Takla Lake area
(MacIntyre et al. 2001a), including regions with deposit formations previously misidentified,
such as “all of Tachek Mountain, the hills west of Granisle and east of Hatchery Arm on Babine
Lake, and the southern end of the Bait Range” (MacIntyre et al. 2001b:582).
Further west, there are numerous deposits of fine-grained volcanic materials within the
Usk area along the Skeena River, an area which includes both the site Gitlaxdzawk and the
modern Gitaus (Nelson et al. 2005:117, 120-126). The material may also have been procured
from other locations within British Columbia, including dacite from the Batzaeko River within
the Fraser drainage, a known popular source (Fladmark 2009:566-567), or rhyolite from
Natalkuz Lake (Fladmark 2009:566) if a section of those sources was not geochemically
characterized when surveyed (Reimer 2018:139-140, 2023). Sourcing would need to be
conducted in the immediate surrounding areas and the larger surrounding region thoroughly to
determine if the materials were procured locally or from a far distance (Reimer 2018:139-140,
2023). Additionally, applying the appropriate geochemical techniques to samples from
Smokehouse Island to identify the specific raw materials used at the site would assist in
narrowing down the potential sources (Reimer 2018:137-139, 2023).

12

Cultural background
The Babine traditionally refer to themselves as Nedut’en (Natoot’en) but also became
known as the Babine or “Lippy People” (Morice 1905:6-7), based on observations by French fur
traders of Nedut’en women wearing labrets in their lips. The Babine fall under the larger
Athapaskan grouping known as the Carrier (Harris 2001:80) or Dene, meaning “the people”
(Furniss 1993:3). All Athapaskan groups speak closely related languages, regardless of location
(Fladmark 2009:562). The Babine share a separate language with the Wet’suwet’en and the
Takla Nation called Witsuwit’en-Babine (Hargus 2007:3, 12-14; Tobey 1981:415). Witsuwit’enBabine has two primary dialects including Babine-Takla or U’in Wit’en and Witsuwit’en
(Hargus 2007:3, 6). Other Carrier groups speak a separate language called Carrier or Dakelh
(Gessner 2003:2; Hargus 2007:3, 12-14; Tobey 1981:415).
The Carrier occupy a vast region between the Coastal Range and the Rocky Mountains in
north-central British Columbia (Tobey 1981:413). The Carrier are known to have been semisedentary, living in permanent villages at specific times of the year, including the winter season
(Morice 1893:184). This observation extends to the Babine Peoples, as radiocarbon dated
charcoal samples recovered from the village site at Nilkitkwa Lake (GiSq-004) reveal Indigenous
occupation at the site as early as 1,300 years ago (Rahemtulla 2012:19-20).
Morice (1893:184) observed that the Carrier had five distinct types of habitation
structures. The first type of dwelling is the ceremonial lodge, a structure where festival banquets,
dances, gift distributions, and other large gatherings occurred (Morice 1893:185). Ceremonial
lodges are characterized by their large size, a rectangular ground plan, two entrances, no
windows, and four upright posts placed in the corners of the lodge (Morice 1893:185-187).

13

Hackett (2017:28) suggested the house-pit depression excavated at Nass Glee in 2012 fit the
characteristics of a ceremonial lodge.
The second type of dwelling associated with Carrier groups are summer lodges, which
function as habitation structures when populations are too large in an area for everyone to sleep
within the larger ceremonial lodge (Morice 1893:188). Summer lodges are smaller in size than
the ceremonial lodge, have a rectangular ground plan, typically only two principal upright posts,
four secondary upright posts in each corner of the structure, and one or two entrances (Morice
1893:188). The third dwelling type used by Carrier groups are fishing lodges, which were only
inhabited during salmon season, and served as places for fishers to sleep and live, and as
locations for smoking and processing fish (Morice 1893:189). Fishing lodges follow the same
ground plans as those established for summer lodges but are more hastily built, and the gable end
walls are located above the transversal beams (Morice 1893:189).
The fourth habitation structure type is the winter lodge (Morice 1893:189). Winter lodges
were built in a new location every year, near a reliable supply of firewood (Morice 1893:189).
They only have one main entrance to conserve heat and most of the structure follows a
rectangular ground plan (Morice 1893:189-190). Additionally, where the main entrance was
located was a half-circular atrium ground plan to reduce heat loss, as there was an inner entrance
between this half-circular section and the rectangular section (Morice 1893:189-190). The fifth
and last habitation structure is the pithouse, a semi-subterranean habitation structure which is
most associated with the southern Carrier and the Tsilhqot’tin (Morice 1893:191). They were
traditionally 20 feet in diameter, they were originally dug to approximately 3 feet deep, and the
only entrance was a log leading out of the roof (Morice 1893:191-192).

14

Traditionally, the Carrier met their subsistence needs through hunting, gathering, and
especially fishing (Morice 1905:7). Historically, prey species for the Carrier have included black
bears, grizzly bears, deer, mountain goats, mountain sheep, porcupines, hares, martens, mink,
beavers, and foxes (Morice 1893:93-94). Early in the historic period, leather was highly sought
after by the Babine, as it was typically given away at feasts acknowledging the death of a relation
(Bouchard 2012:37-38). Additionally, geese and swans were occasionally procured (Morice
1893:104). However, the staple resource for the Babine, the Wet’suwet’en, and the Nadleh
Whut’en was salmon (Prince 2014:120). Before the twentieth century, up to half of the dietary
intake of the Carrier in British Columbia was salmon (Hudson 1983:58). According to a stablecarbon isotope analysis conducted on burials by Chisholm (1986:115-123), Indigenous groups
living along the Thompson and Fraser River systems in interior British Columbia tended to
depend more on salmon for subsistence the closer they were located to major salmon runs.
Additionally, salmon were used in interior British Columbia as far back as 5000 years ago
(Chisholm 1986:122). Another important fish resource for northern interior communities was
whitefish, especially sought-after during winter lake fishing (Matson and Magne 2007:21, 23).
The Babine also ate other fish, tubers, and berries to supplement their diet (Bishop 1987:74).
The Babine have four clans based on matrilineal descent: Likhtsemisyu (Beaver), Gilantin
(Caribou), Jilhtsehyu (Frog), and Likhc’ibu (Bear) (Fiske and Patrick 2000:48). Their western
Indigenous neighbours and other Dene groups share corresponding clans, and as a result,
“[m]embership in the clans…provides a compatible social identity between the nations and
regulates rights of trade, resource access, and inter-nation marriages (Fiske and Patrick
2000:48).” Within each clan are subclans or “Houses” which are represented by crests and have
members connected along matrilineal lines (Fiske and Patrick 2000:49). Social division usually
15

classified individuals in the community as commoners, or as dineese and ts’akeze’, the men and
women of nobility or chiefly office, with hereditary titles, privileges, and responsibilities to the
members of the clan and to the balhats or potlatches (Fiske and Patrick 2000:50).
Interactions between the Carrier groups and Europeans intensified when the Northwest
Company built a post at McLeod Lake in 1805 Current Era (CE) (Bishop 1987:75). In 1812 CE,
fur trader Dan Harmon of the Northwest Company became the first European to interact directly
with the Babine (Harmon 1957:133-134). Harmon noted that the Babine had five villages in the
area and guessed the population was as large as 2000 individuals (Harmon 1957:134). However,
the Babine had told him that there were more community members as they “are a numerous
Tribe” (Harmon 1957:134).
The Northwest Company’s posts in New Caledonia were taken over by the Hudson’s Bay
Company in 1821 CE; in 1822 CE, the HBC established Fort Kilmaurs at Babine Lake,
intensifying European contact with the Babine (Bouchard 2012:18; Marsden and Galois
1995:173). However, according to Fiske and Patrick (2000:34), “[a]lthough [HBC fur trader
William] Brown was able to purchase salmon in large supplies, he was unable to wrest control of
the fur trade, which was primarily conducted with [Wet’suwet’en] and [Gitxsan] through the
management of hereditary chiefs.” It’s important to point out that the complexities of the
Indigenous groups in northern British Columbia would have been difficult for a newcomer who
did not speak the local language, like Brown, to understand initially (Bouchard 2012:16).
A strong trade relationship existed between the Lake Babine Nation and Indigenous
groups to the west, and to a lesser but still important extent, to the east (Bouchard 2012:32-33,
57; Hackett 2017:61-68; Tobey 1981:415, 416). The Babine had an almost exclusive trade
16

partnership with the Gitxsan, who would travel along the Skeena River and trade with the
Coastal Tsimshian, acting as the middlemen between the coast and the interior (Tobey
1981:416). Furs and prepared hides from the Babine were traded for bark blankets, copper, wool,
and items made of shell from the coast (Tobey 1981:416). It is known that other Carrier groups
obtained iron from the Coastal Tsimshian, as according to the oral histories of the Nak’azdli First
Nation, Na’kwoel had received an iron adze through trade on the Skeena River in 1730 CE
(Keddie 2006:3). Strategically, Indigenous groups used trade-exclusive prerogatives and
marriages between tribes to solidify their trade networks (Marsden and Galois 1995:171-172).
The most prominent early ethnographer to discuss the cultural traditions and practices of
the Lake Babine Nation was Father Adrien Morice. Usually stationed at Fort St. James, Morice
also had regular contact with the Carrier at Babine Lake and other locations throughout northern
British Columbia (Harris 2001:90; Morice 1893; 1905). Morice (1893; 1905) used his writings to
describe the different Dene communities, their ways of life, and some of the neighboring
Indigenous communities of interior British Columbia. His works are very descriptive. However,
they were written through an Eurocentric lens and are occasionally condescending towards the
Indigenous cultures he studied and recorded (Morice 1893; 1905). Regardless, he would defend
groups like the Babine when necessary, as was the case when officers from the Fisheries
Division of the Department of Marine and Fisheries attempted to force the Babine to take down
their fish weirs on behalf of the canneries in the region (Harris 2001:100-102, 110-111).

Modern Scholarship on the Lake Babine Nation
To understand how the current research fits within the modern scholarship on the Lake
Babine Nation, recent research focused on the Babine community will be summarized. The
17

research conducted by Fiske and Patrick (2000) focused on understanding the traditional law of
the Babine, as it was and is enforced by and through the balhats in the Babine community. To
conduct the research, HBC journals, missionary documents from the nineteenth and early
twentieth centuries, and ethnographic studies were assessed and compared with primary oral
information from interviews with elders and hereditary chiefs, and video recordings of balhats
(Fiske and Patrick 2000:24-25). The research discussed the central role of the balhats in both
traditional and current Babine communities, how it originated in the precontact period to
maintain peace, how it has changed as a result of colonial interference, and the challenges the
Babine have faced while reasserting their customary law in a society dictated by Canadian state
law (Fiske and Patrick 2000). Fiske and Patrick (2000:229-233) finish by suggesting that the
relations between the Canadian state, the surrounding communities, and the Lake Babine Nation
need to change for harmony in the community, and that there is hope in implementing an
alternative justice system based on traditional law.
Research by Bouchard (2012) suggested that before 1830 CE the social and economic
networks of the Lake Babine Nation had remained relatively unaltered by the introduction of the
fur trade into the interior of British Columbia. Bouchard (2012:1, 16) primarily used HBC
records to assess the dynamics and relationships between the Babine, European fur traders, the
Carrier to the east, and the Gitxsan and Wet’suwet’en to the west of Babine Lake.
Environmental, linguistic, and anthropological scholarship, and selected oral testimonies from
the Delgamuukw case, are incorporated alongside the HBC records to build an understanding of
precontact relationships and dynamics after the introduction of the fur trade into the region
(Bouchard 2012:18). Evidence suggests that cultural sharing of the clan system, the feast system,
and some linguistic factors between the Gitxsan, Wet’suwet’en, and the Babine occurred before
18

the introduction of Europeans (Bouchard 2012:22-47). Additionally, ties between the Babine and
Indigenous groups to the west were stronger than those to the east due to differential resource
access, travel distance, and social connections (Bouchard 2012:48-70). Lastly, the fur traders at
Fort Kilmaurs recognized and struggled against the strong trade relationship between the
Gitxsan, Wet’suwet’en, and Babine (Bouchard 2012:75-85).
Hackett (2017) aimed to prove the precontact existence of social complexity for the
Babine from the Gitxsan and the Coastal Tsimshian. Hackett (2017) used archaeological
evidence (largely from the excavation of House Depression 1 (HD1) at Nass Glee in 2012), and
historical documentary evidence (including HBC records, and other historical documents) to find
evidence for social complexity of the Lake Babine Nation. The existence of a large ceremonial
lodge at GiSq-4, the consistency of traditional artifacts being used, and the presence of exotic
goods recovered during the 2012 excavation support the inference of social complex traditions
being practiced by the Babine during the precontact period (Hackett 2017). By going through the
historical records, Hackett (2017) determined that the large population sizes of Babine villages,
the procurement of large amounts of salmon, the implementation of chiefs and ranking, the
practice of warfare, the use of potlatches, the strength of the Indigenous trade networks, the
procurement of prestige goods, and the practice of semi-sedentary lifestyles, polygamy and
slavery all suggested a socially complex Babine society existed during the precontact period.
Research by Kantakis (2017) focused on how the Lake Babine Nation efficiently used
fish weirs within the context of the Babine watershed, as information regarding fish weir
technology use in the region was missing. Kantakis (2017) initially analyzed the existing
scholarship on global uses of fish weirs to better establish the types of weirs at Babine Lake.
Environmental, ethnographic, and historical records for the region were analyzed, followed by
19

the incorporation of the results of the Babine Archaeology Project fish weir survey, to determine
how the Babine used fish weirs (Kantakis 2017). Results indicated that the fish weirs used at
Smokehouse Island were multi-component, bank-to-bank weirs which aligned with the flowing
stream style of weir technology (Kantakis 2017:99). Additionally, for the Lake Babine Nation,
fish weirs were intricate, complex, and effective forms of technology that required appropriate
management to harvest large amounts of fish to sustain the community without dangerously
depleting the salmon population (Kantakis 2017:101-102).
A study by MacDonald and colleagues (2019) focused on how the pigment used to create
a pictograph panel identified at the southern arm of the Babine Lake, at the Boiling Point site
(GcSi-1), was developed based on the properties of the pigment. Conducting this study involved
collecting fresh samples of an iron-oxidizing bacteria (L. ochracea), which was present in
pigments from the recovered rock art sample, and subjecting them to two different heating
procedures (MacDonald et al. 2019:4). Following these initial steps, the samples and a recovered
rock art fragment from GcSi-1 were analyzed with superconducting quantum interference device
magnetometry, electron microscopy/microanalysis methods, and X-ray Powder Diffraction
(MacDonald et al. 2019:4). The evidence from the results of the analyses suggested that
microbial iron mats dominated by iron-oxidizing bacteria were harvested to use for paint and
heat-treated with open domestic hearth fires to between 750 °C to 850 °C to augment the
properties of the color, turning it red from brown, which represents skilled knowledge of the
properties of this iron-oxidizing bacteria by the Lake Babine Nation (MacDonald et al. 2019).
In summary, the existing modern scholarship covers a lot of important cultural
information regarding the Lake Babine Nation. Topics have included the importance of balhats
in Babine society and traditional governance (Fiske and Patrick 2000), the strength of precontact
20

trade networks between the Babine and their western and eastern neighbours (Bouchard 2012),
the precontact arrival of socially complex traits for the Babine (Hackett 2017), their complex and
effective use of fish weirs (Kantakis 2017), and the intricate resource knowledge evident in
pictograph construction at Babine Lake (MacDonald et al. 2019). However, no previous research
has focused on either stone tool production or use by the Babine specifically, outside of brief
discussion (Hackett 2017:36-43). In the next section of this chapter, the recorded known
Athapaskan, Carrier, and Babine forms of stone tools and stone tool production will be
discussed.

Athapaskan, Carrier and Babine use of stone material
First, the use of stone tools by mobile Athapaskan grouping must be summarized
(Fladmark 2009:594-595, 597-598; Matson and Magne 2007:149-155; Tobey 1981:415). Matson
and Magne (2007:129) and Magne and Matson (2010:218) asserted that Athapaskan sites
contained larger amounts of small points, cobble tools, and biface fragments. Additionally,
Matson and Magne (2007:121) suggested that Athapaskan sites are likely to contain an
abundance of bifaces with a high degree of variability. There has also been some evidence to
suggest the presence of microblades being associated with Athapaskan occupation (Matson and
Magne 2007:154). This viewpoint has extended to include proto-Athapaskan populations
(Magne and Fedje 2007:184-188). However, any correlation between Athapaskan occupied sites
and microblades becomes tenuous at sites more recent than 4000 BP (Magne and Matson
2008:288). Additionally, microblades also appear in some non-Athapaskan sites in British
Columbia (Magne and Matson 2010:219).

21

Matson and Magne (2007:112, 116) also confirmed the characterization of Athapaskan
side-notched points (also known as Plateau Athapaskan side-notched points), being generally
thicker than PPT points, often having basal concavities and spurs, shallow yet wide notches, and
elongated blades (Magne and Matson 2008:285-288, 2010:217-218, 220-225). Also tied to
Athapaskan sites are Kavik or Klo-Kut points, which are “a contracting stem point with a slightly
convex blade…” (Campbell 1968, as cited in Matson and Magne 2007:37, 141). Kavik points
have been recovered from Athapaskan sites in British Columbia, including the Potlatch, Chinlac,
and Ulgatcho sites (Magne and Matson 2008:285-286, 288).
Fladmark (2009:594-595) was critical of assigning projectile point styles to specific
linguistic cultural groups. In comparison, Fladmark (2009:595) suggested that Athapaskan
ethnicity could only be determined when 1) moose or caribou tibia fleshers with toothed or
serrated ends, 2) flaked-and-ground stone adze-blades, and 3) coarse-grained cortical spalls
known as chi-thos were all present together in high numbers within an assemblage. The presence
of Athapaskan side-notched and Kavik points can be used as supportive evidence if recovered
alongside the three tools listed by Fladmark (2009:595).
Stone tool use by the traditionally semi-sedentary Carrier or Dene groups will now be
reviewed (Fladmark 2009:562; Furniss 1993:3; Morice 1893:184; Tobey 1981:415). The most
detailed descriptions of stone tool use by the Dene, including the Babine, comes from Morice
(1893:44-55, 60-65), who described the use of scrapers, stone axes, adze-blades, sinker stones,
wedges, pestles, hammers, knives, grinding stones, arrowheads, spearheads, daggers, war-clubs,
and skull-crackers. Additionally, Morice (1893:45-46, 48, 50-54, 62, 64-65) suggested the Dene
used gray basalt, fine-grained igneous stone (including a black variety), felsite, quartzite, augiteporphyrite, black flint, green marble, smoky quartz, chalcedony, felspathic slate, and granite for
22

stone tool production. Morice’s (1893:65) writings also suggested quarry sites were owned and
restricted by Indigenous individuals, including the restriction of individuals from the same
community or other communities. Lastly, the description by Morice (1893:65) on core reduction
and stone tool production vaguely described direct hard hammer percussion and pressure flaking.
Regarding the Babine specifically, Harmon (1957:133) mentioned that the community was
armed with bows, arrows, axes, and clubs.
Looking at raw materials specifically, Reimer (2015) suggested that the Tahltan
exchanged obsidian from Mount Edziza flows differentially with close-by Carrier groups. The
only group to receive obsidian from the Pyramid flow from Mount Edziza appears to have been
the Babine (Reimer 2015:425). This assessment loosely matches Morice’s (1893:65) description
of quarry site ownership, but caution is necessary when interpreting the writings of Morice.
However, when incorporating Morice’s (1893:45-46, 48, 50-54, 62, 64-65) accounts on stone
tool use, it must be recalled that the same writings are very detailed yet tend to generalize about
the larger Dene group in British Columbia. Therefore, the value of this information must be used
with some caution.

Regional Archaeological Review
The current review of archaeological work will focus on sites from within the Subarctic
cultural area within British Columbia (Fladmark 2009:555-559). The Subarctic cultural area
exhibits a wide range of different environments, and includes archaeological sites associated
with the Nass, Skeena-Bulkley, Fraser, Nechako, Stikine, Taku, Atlin, Teslin, Alsek,
Tatsenshini, and Mackenzie Rivers (Fladmark 2009:558-559). Initially, known Athapaskan sites
directly along or north of the West Road River, and south of the Stikine River will be
23

summarized (Fladmark 2009:557). Focus for the review will then shift to part of the SkeenaBulkley-Nass archaeological region. This shift is due to the Babine watershed being directly tied
to the Bulkley and Skeena river systems, and associated with the Skeena-Bulkley-Nass
archaeological region (Bouchard 2012:32-33, 57; Fladmark 2009:557, 584; Tobey 1981:415,
416). This review will not include the Nass River drainage, as no archaeological information
currently exists regarding Subarctic Athapaskan occupations for the Nass drainage (Fladmark
2009:584). Additionally, the review will include sites from the Kitselas Canyon region along the
middle Skeena River in addition to sites near the confluence with the Bulkley River, as there is
evidence of alternating coastal and interior occupations within those areas (Allaire 1979; Ames
1979a, 1979b; Fladmark 2009:584).
The Carrier site of Chinlac is located next to the confluence of the Stuart and Nechako
Rivers, on the north side of the latter. Borden (1952) excavated the Chinlac site in 1950 and
1952, which led to the identification of ten rectangular house depressions, approximately 1800
cachepits, evidence of salmon procurement and processing, and the establishment of late
prehistoric period occupation at the site, from 1700 CE until over 100 years later. Borden
(1952:32) also noted recovered artifacts included various types of projectile points, drills, knives,
side, adze blades, abrasive stones, whetstones, and end scrapers, “and numerous unworked
irregular flakes that were used for scraping and cutting operations. The thrifty use of such flakes,
which many other groups would discard as waste, seems to have been a common trait among the
Carrier [groups].” In addition to the stone tool types recovered, the excavation also led to the
recovery of pointed bone and antler artifacts, bone and antler awls, one oval bone pendant, one
small bone projectile point, bone fragments, bark fragments, numerous birch bark rolls, and
protohistoric and historic period artifacts (Borden 1952:32-33). A recent Athapaskan occupation
24

had been suggested for the site (Borden 1952:34), further confirmed by the presence of smallsized Athapaskan side-notched points (Matson and Magne 2007:37) and Kavik points (Matson
and Magne 2007:140).
In 1984, Cranny (1986) surveyed regions near Chinlac, including sites at the outflow of
Cluculz Lake, Cobb Lake, Sob Lake, and Hogsback Lake, Cluculz Creek, alongside the Stuart
and Nechako Rivers, and at the confluence of the Stuart and Nechako Rivers. Results from the
survey included the identification of 38 archaeological sites and evidence of most of the large
sites being clustered at the outflow of Cluculz Lake and at the confluence of the Stuart and
Nechako Rivers, optimal locations for salmon procurement (Cranny 1986:123, 138). Cranny
(1986:138) asserted that the larger sites are from the proto-historic period when populations were
at their largest.
In 1951, Borden (1952:36) surveyed a village site at Natalkuz Lake, which is located
along the Nechako watershed in northern British Columbia (Fladmark 2009:557). A house-pit
located at the head of the lake was excavated in 1952 (Borden 1952:31, 35). Borden (1952:35)
suggested the remains of the roof of the house-pit fit a style used by Interior Salish populations
in southern and south-eastern British Columbia, though with a noticeable difference. The
difference included the only pit dug around the central hearth instead of the full diameter of the
house structure. Borden (1952:35-36) recovered artifacts associated with Carrier sites within and
under the turf layer, while the deposits below (including the broad house structure and the central
hearth-pit) appeared different, with numerous debitage and larger-sized tools made from rhyolite.
Carrier populations were perceived by Borden (1952:36) as favouring obsidian and basalt for
stone tool production. Borden (1952:36) noted that percussion flaking was the prominent form of
flintknapping used during the earlier occupations, and debitage was largely disregarded instead
25

of utilized, adding to the interpretation of the earlier layers not being Athapaskan. However, the
proximity of the source was later determined to be the reason for less refined forms of
flintknapping during the older occupation at the site (Fladmark 2009:575; Wilmeth 1978:76).
Borden (1952:39-40) interpreted the archaeological evidence at Natalkuz Lake to suggest the
Athapaskans had migrated to the area and that the earlier occupants represented a separate
population. The hearth from the house-pit shows a radiocarbon date of 2415 ±160 BP for
earliest occupation at the site (Wilmeth 1978:76).
Located in the central Interior Plateau of British Columbia, near the headwaters of the
West Road River and approximately 100 miles east of the Pacific Ocean, the Ulkatcho village
site was excavated by Donahue (1973:153) in 1970 (Fladmark 2009:557, 583). Postholes and
hearths were recorded, and artifacts recovered include side-notched stone points, stemmed
points, scrapers, microblades, bifacial tools, a burin/drill, debitage, retouched flakes, utilized
flakes, ochre pellets, modified bone and antler artifacts, modified bark artifacts, and numerous
protohistoric and historic period artifacts (Donahue 1973:159-170). The Athapaskan Kavik or
Klo-kut style of projectile point is represented by at least one point recovered at the site
(Campbell 1968, as cited in Matson and Magne 2007:37, 141; Donahue 1973:161; Fladmark
2009:583; Magne and Matson 2008:285-286, 288). The primary material used for stone tool
production was obsidian, with Dean River and Rainbow Mountains as source locations (Donahue
1973:160). Numerous faunal remains were collected and analyzed, showing the use of beaver,
cervids, lynx, amphibian, and salmon at the site (Donahue 1973:172-173). Results from the
excavation appear to align with other known Carrier and Athapaskan sites at the time, and while
Donahue (1973:170, 174-175) noted the evidence of early occupation at the site, he frames the
primary occupation as occurring during the fur trade.
26

The Tezli site (FgSd-1) lies alongside the Fraser drainage in central British Columbia on
a low terrace, by the eastward-draining outlet of west Kluskus Lake (Donahue 1977:1, 31, 39).
Donahue (1977:2) conducted excavations at the site in 1970 and 1971, with the original focus on
a known historic Athapaskan site, followed by a focus on the prehistoric component. At the Tezli
site, 46 depressions were identified, 45 of which are noted as housepit depressions (Donahue
1977:119, 121, 199). Nine of the housepit depressions were shovel tested and three were
completely excavated (Donahue 1977:199). Six cache pits or possible post holes were identified
and recorded within four housepit depressions, while confirmed post holes were recorded for two
housepit depressions (Donahue 1977:143). Concentrations of stone artifacts, including scrapers,
side-notched points, bifaces, a stemmed uniface, perforators/gravers, and debitage, and fish
remains, were recovered (Donahue 1977:144-145, 163, 176-188; Magne and Matson 2008:285).
Two stone fish weirs were identified within proximity (Donahue 1977:146).
At least ten of the side-notched points recovered from Tezli site fit the Athapaskan sidenotched point classification (Magne and Matson 2008:285), though no microblades were
recovered (Donahue 1977:3). The earliest radiocarbon date from a charcoal sample from the site
was corrected to 2404 Before Current Era (BCE) ± 170 from 3850 ± 140 BP (Donahue
1977:131-133, 150, 157). Additionally, Tezli showed strong similarities regarding settlement and
subsistence patterns, and artifact assemblage types to other sites on the Canadian Plateau in
comparison to sites to the east and the coast (Donahue 1977:252-256). This includes the recovery
of key-shaped unifaces and medium-sized notched points, suggesting earliest occupation at the
site during the Shuswap and Plateau Horizons (3500 to 1200 BP), which are associated with
Interior Salish populations (Donahue 1977:176-188; Fladmark 2009:582; Pokotylo and Mitchell
1998:87-88; Richards and Rousseau 1987:22-34).
27

The Carrier site of Punchaw Lake (FiRs-1) is in north-central British Columbia,
approximately 35 miles southwest of Prince George, and Area A at the site was excavated by
Fladmark (1976:19, 27) in 1973. During the 1973 excavation, Fladmark (1976:21) identified
“surface cultural features [which] include[d] 43 house platforms, 57 storage pits, and a 100 m
segment of a native trail running east-west across the center of the site” (Fladmark 1976:21).
Forty-one of the house platforms align with the characteristics known for summer dwellings used
by Carrier groups, while two align with the circular semi-subterranean house-pits typically
associated with Plateau populations (Fladmark 1976:21-24, 2009:580; Morice 1893:184-191).
Occupation of the Punchaw Lake site was dated by Fladmark (1976:28, 30) to be
approximately 4000 years in age, based on the radiocarbon dating of organic material associated
with burial 2. During the excavation, a total of 6,200 artifacts were recorded in situ (Fladmark
1976:28). Most of the stone artifacts were made from vitreous basalt, and approximately 101
complete projectile points were recovered, including leaf-shaped, stemmed, square-based, fish
tail, side-notched, corner notched, and triangular points (Fladmark 1976:28). Additionally, keyshaped unifaces and medium-sized stemmed and notched projectile points were recovered in
association with the earliest occupations at the site (Fladmark 2009:580). Based on many of the
artifacts recovered from lower layers and the style of burial 2 recovered from below the hearth at
house-pit 1, a known chronology from the Shuswap and Plateau Horizons is suggested for the
earlier occupations at Punchaw Lake (Fladmark 2009:580; Richards and Rousseau 1987:22-36).
In comparison, typical Athapaskan style chi-thos and flaked and ground adze blades were
recovered from the most recent occupation at house-pit 1 (dated to 560 ± 75 BP), in addition to
small-sized side-notched points suggesting both Kamloops Horizon and Athapaskan influences

28

based on form (Fladmark 1976:26-28, 2009:580; Magne and Matson 1982:65-66; Pokotylo and
Mitchell 1998:88; Richards and Rousseau 1987:43).
Another area at the Punchaw Lake site (FiRs-1), Area C, was excavated by Montgomery
(1978:1) in 1974. The excavation was focused on the north end of the site, to investigate 4 house
platforms from two cultural subareas (Montgomery 1978:50-52; 202-203). In all excavation pits,
a large amount of fire-cracked rock was identified (Montgomery 1978:60). At least two areas of
long-term occupation were suggested by the cultural deposit depth and ground contours
(Montgomery 1978:67). Additionally, identified cultural features include a possible cremation, a
small pit, two hearths, one cluster of faunal remains, and two concentrations of fire-cracked rock
(Montgomery 1978:67). During the excavation, 6971 stone artifacts were recovered, most of
which were made from basalt (Montgomery 1978:68, 70, 237). Based on the results of
radiocarbon dating, Area C suggests long term occupation of the site for over 1200 years, from
440 CE to 1710 CE (Montgomery 1978:192-193). Montgomery (1978:215-217) noted that the
raw materials and stone tool types used at Area C were mostly consistent with those recovered
from area A, though artifacts from Area C were generally smaller and have less variability in
form. Montgomery (1978:219, 221-222, 225) noted that the side-notched points recovered from
Area C showed similarities to those recovered from Chinlac, Ulkatcho, Tezli (Donahue
1973:160-166). Additionally, Montgomery (1978:244) asserted that based on the findings at
Area C, an Athapaskan movement into the Punchaw Lake area was likely to have occurred by
400/500 CE.
In the 1960s and 1970s, the “North Coast Research Project” was conducted along the
lower Bulkley River and Kitselas Canyon regions in the Skeena watershed on behalf of the
National Museum of Man (Fladmark 2009:569). The project included the Gitaus site (GdTc-2),
29

which is located along the middle Skeena River by Terrace, B.C., at the foot of Mount Bornite,
and has strong ties to the Coastal Tsimshian group, the Kitselas (Allaire 1979:21, 23, 24; Allaire
et al. 1979:58). The village site also has strong ties to the sites of Gitlaxdzawk and Gitxtsaex, as
they are locations the occupants of Gitaus village later moved to (Allaire 1979:21). In 1968, an
excavation of the Gitaus village site was conducted (Allaire 1979:21). Over 13,000 pieces of
stone debitage, 1,455 used boulder spalls, and 1320 other artifacts were recovered at the site
(Allaire 1979:21, 24). A Northwest Coast plankhouse was identified along the riverbank based
on a large post outline and a clear pit, though an association with any of the cultural layers at the
site was not possible (Allaire 1979:48). Excavations at Gitaus revealed occupation of the site
from 4000-1500 BP (Allaire 1979:46-48).
Three cultural sequences were identified at the site (Allaire 1979:45-48). The earliest
cultural sequence (Zone VI) was radiocarbon dated to approx. 2000 BCE and aligns with the
Lower Horizon sequence dated to ca. 3000-1500 BCE at Prince Rupert Harbour (Allaire
1979:45-46). Shared artifact types include leaf-shaped chipped stone points, a variety of cobble
tools, stone saws, sawn shale plaques, rubbed slate points, and perforated and shaped abraders
(Allaire 1979:46). The second cultural sequence is represented by Levels V and III at the Gitaus
site and is named the Skeena Complex as it shares characteristics with another Skeena River site,
the Hagwilget site, and is more reflective of an interior cultural adaptation rather than a coastal
adaptation (Allaire 1979:46; Ames 1979a:208-209, 1979b:235). The Skeena Complex is
characterized by specific denticulated and notched flakes, gravers, celts, perforators, fan-shaped
scrapers, flat scrapers, thumbnail scrapers, keeled scrapers, a variety of bifacial tools, and
parallel-flaked lanceolate points (Allaire 1979:46). The Skeena Complex appears to have
occurred at the site shortly after the Lower Horizon sequence, showing a radiocarbon date of
30

1400 BCE at the Hagwilget site, and suggests forest and riverine-based adaptation strategies
(Allaire 1979:46-47; Ames 1979b:234). Additionally, the Skeena Complex shows strong
similarities to other northern British Columbia sites (Donahue 1975:24), though this sequence is
missing typical Athapaskan associated artifacts like side-notched points and microblades (Allaire
1979:47; Matson and Magne 2007:112, 116, 154). Some degree of influence from the coast is
evident in artifacts like ground stone celts and ground stone points (Allaire 1979:47).
The third and most recent cultural sequence at Gitaus has been termed the Kleanza
Complex and is represented by the Level I and II cultural layer (Allaire 1979:47). The Kleanza
Complex is characterized by numerous cobble spalls and tools, a reduced amount of chipped
stone tools, and the appearance of coast-influenced artifacts, including net sinkers, perforated
pebbles, slate knives, slate mirrors, points with hexagonal cross-sections, and labrets (Allaire
1979:47-48). Additionally, a large amount of fire-cracked rock was present, a new feature
specific to the Kleanza Complex at the site (Allaire 1979:48). The fire-cracked rock was
theorized by Allaire (1979:48) as possibly indicating an adoption of coastal residential and
subsistence strategies. The date range for the Kleanza Complex at Gitaus most closely aligns
with the Middle Horizon at Prince Rupert, dated to ca. 500 BCE and 500 CE, though this cannot
be completely confirmed based on the evidence (Allaire 1979:48). Allaire (1979:48-49)
suggested the cultural sequences are representative of the adoption of cultural traits, facilitated
through trade along the Skeena River, and that material culture at Gitaus shows strong
similarities with coastal practices by the Kleanza Complex (Allaire 1979:49).
Gitlaxdzawk is located in the middle Skeena River area, approximately upstream from
Terrace, B.C., and is positioned east of the Canadian National Railway track (Allaire et al.
1979:58, 71-72). It is associated with the Kitselas group, and it has strong ties with the Coastal
31

Tsimshian (Allaire et al. 1979:58, 70-71). In 1971, the site of Gitlaxdzawk (site GdTc-1) was
mapped and excavated by Allaire and MacDonald (1971) on behalf of the National Museum of
Man (Allaire et al. 1979:58, 69). A total of thirteen test pits were excavated, which resulted in
the recovery of 250 artifacts (Allaire et al. 1979:113, 115). Recovered artifacts include retouched
flakes, a biface thinning flake, cortex flakes, ground stone adzes, abrader fragments, pigment
nodules, cobble choppers, hammerstones, modified bone artifacts, bone fragments, shells, shell
fragments, birch bark rolls, and a variety of European historic artifacts (Allaire et al. 1979:116137). Hearths and fire-cracked rocks were also present (Allaire 1979:113-115). During the
excavation, they located and mapped the remains of ten house floors, which were identified as
being similar to other Coastal Tsimshian houses in design, structure, and settlement lay-out
(Allaire and MacDonald 1971:50; Allaire et al. 1979:107-108, 110-112). Lastly, the presence of
facetted red ochre pigment, bar abraders, planning adzes, splitting adzes, and shellfish and sea
mammal remains suggest influence from and trade with coastal groups (Allaire et al. 1979:136138). However, minimal evidence at the site indicated ties with the Gitaus village site (Allaire et
al. 1979:138).
The North Coast Prehistory Project included a Hagwilget Canyon site (GhSv-2), located
by Hazelton, BC, near the confluence of the Skeena River and the Bulkley River (Ames
1979a:183). The Wet’suwet’en have known historical ties to the site (Ames 1979a:183; Jenness
1944). During the initial archaeological work at the site in 1966, MacDonald (1967, as cited in
Inglis and MacDonald 1979:11) dug a shovel test, which led to the recovery of 60 artifacts,
including traditional Indigenous and historic artifacts, and samples recovered from under a
hearth which provided a radiocarbon date of 3,430 ±200 years B.P. at the site (MacDonald
1969:249; Inglis and MacDonald 1979:11).
32

An excavation at site GhSv-2 was conducted in 1970 by Ames (1979a:183). Three
chosen blocks were excavated, two of which were placed near where Carrier houses were
located, based on photographs (Ames 1979a:187). Recovered artifacts include complete and
fragmented projectile points/knives (consisting of laurel leaf, elongated lanceolate, and
“Plainview-like” styles), bifaces, hammerstones, scrapers, flaked adzes, cobble tools, utilized
cores, modified bone tools, bone fragments, antler artifacts, utilized flakes, and fish, mammalian,
and floral remains (Ames 1979a:194-203). Recorded features included post molds, storage pits,
hearth pits, and fire-crack rocks (Ames 1979a:203-205).
The earliest occupation at GhSv-2, termed Zone A, has a relatively small assemblage and
technological characteristics (including the presence of flaked adzes and scrapers, and alternating
hard hammer and antler billet flintknapping techniques) that make it appear similar to Zone V at
Gitaus, and to some degree Zone III (Ames 1979a:208-209, 1979b:235). Numerous lanceolate
shaped projectile points, cobble tools, spall tools, and thinned bifaces also were recovered from
Zone A (Ames 1979b:234-235). Ames (1979a:208) used a radiocarbon date in addition to the
rate of sedimentation to infer an initial occupation of the site between 4500 to 5000 BP. As
previously mentioned, Allaire (1979:46-47) asserted both Zones V and III at Gitaus, and Zone A
were representative of the Skeena Complex. Ames (1979b:234) established that termination of
the earliest occupation at Hagwilget Canyon occurred at approximately 2000 BCE.
There is no zone at Gitaus that appears similar to Zone B at the Hagwilget Canyon site,
though occupation is less intensive in comparison to the earlier cultural layer (Ames 1979a:209).
Ames (1979b:235) suggested the site likely functioned as a fishing station during Zone B
occupation. Lithic scatters, faunal material, and the occasional fire-cracked rock are evident for

33

the period (Ames 1979a:207, 1979b:235). Zone B represents occupation at the site from
approximately 2000 BCE to 1820 CE (Ames 1979a:207, 1979b:235).
Further research is required to compare Zone C at GhSv-2 with other nearby sites
associated with that time period (Ames 1979a:209). During Zone C, historic artifacts were
introduced, and fish vertebrae, scattered faunal remains, and seeds were recovered (Ames
1979a:207). Features from Zone C include the remains of storage pits and structures (Ames
1979b:235). Zone C is reflective of when the Wet’suwet’en moved into the area (Ames
1979a:207, 1979b:235).
Another site in Hagwilget Canyon, site GhSv-A, is located upriver from the modern
Hagwilget village, on the north bank of the Bulkley River, and was surveyed and shovel tested
by Albright (1986:1, 27-28) in 1985. Albright (1986:28-29, 38-43, 44-47, 49-50, 53) recovered
452 artifacts from the site, including burnt bone fragments, an adze fragment, a stone axe, a
polished stone fragment, slate mirror fragments, grooved stone abraders, a polishing stone, a
leaf-shaped point, a large point fragment, stone knives, a drill or graver, a small scraper, small
acute edged tool, cobble tools, and debitage. The stone tools recovered from the site were made
of jade, chalcedony, obsidian, and basalt (Albright 1986:28). Albright (1986:53-54) suggested
the high frequency of cobble tools and the presence of slate mirror fragments, a groundstone axe,
and a groundstone adze reflect strong similarities with coastal related sites along the middle and
lower Skeena River. Additionally, the assemblage appears to match the late Skeena Complex or
early Kleanza Complex at the Gitaus village site (Allaire 1979:49; Albright 1986:60).
During the late 1980s, a survey by Simonsen (1989) and an archaeological impact
assessment by Carlson and Bussey (1990) focused on a portion of land by Hazelton, BC, in
34

Hagwilget Canyon, which resulted in the identification of 26 sites. Site types included the
presence of subsurface debitage and other lithic tools, faunal remains, stratified cultural deposits,
and cache pits (Carlson and Bussey 1990). No cultural affiliation could be definitively assigned
though the presence of a high amount of green chert debitage from site GhSv-19 was suggested
by Carlson and Bussey (1990:65-66, 100) as possibly indicating association with the Skeena
Phase.
At the headwaters of the Bulkley River, Burley (1975:8, as cited in Fladmark 2009:584)
surveyed Morice Lake in 1975 and discovered a house-pit. The house-pit measured between 3.1
to 3.34 m in diameter, a size range unusual and typically concentrated by the northern limits of
the Fraser drainage (Burley 1975:8, as cited in Fladmark 2009:584). In 1975, Rafferty (1975)
surveyed the upper eastern section of the Bulkley Valley, which included the Gilmour, Conrad,
Sunset, Old Woman, Broman, Day, Maxam, and Bulkley Lake shorelines. Along the Bulkley
River, five precontact sites (including two circular house depressions) and one historic site were
recorded. Following the survey by Rafferty (1975), Kimble (1978:12) surveyed along the
Bulkley River, beginning at its junction with the Telkwa River. During the survey, one
rectangular cultural depression site, 11 circular cultural depression sites, one historic railroad
construction camp, and three historic wagon trail sites were identified by Kimble (1978).
In 1985, Albright (1986:31-32) excavated site GgSt-2, which is located on a terrace on
the Bulkley River in Moricetown Canyon in British Columbia. Three cultural layers were
identified at the site: Layer A, B, and C, with C being the oldest (Albright 1986:32). The
excavation resulted in the recovery of burnt bone, a birch bark roll, 847 pieces of debitage, a
bone point, an adze fragment, a ground slate bar, leaf-shaped points, triangular points, large
notched points, small side-notched points, a small corner-notched point, small notched point
35

fragments, large point fragments, stone knives, two gravers, acute small-edged tools, and cobble
tools (Albright 1986:32-33, 37-39, 43-48). Additionally, sixteen post moulds, four hearths and
three pit features were identified (Albright 1986:33-35). An initial occupation radiocarbon date
of between 4700 to 5660 BP was determined for the site (Albright 1987, as cited in Budhwa
2005:23; Magne and Matson 2008:284). The presence of notched points and triangular points in
layers A and B suggest shared cultural traditions with Athapaskan sites from interior BC, while
the oldest layer suggests a more generalized adaptation to the site (Albright 1986:53-54, 57,
Appendix E-12, Appendix E-13).

Previous archaeological work and the Babine Archaeology Project
While some brief survey work was done in the region by H.I. Smith in 1927 and by
Michael Kew in the 1950s (Fiske and Patrick 2000:16-17), the first significant systematic
archaeological work occurred around Babine Lake when Mohs led surveys in the region over
three field seasons (Mohs 1974, 1975; Mohs and Mohs 1976). In total, 225 sites were recorded
during the three field seasons, including six village sites, eight habitation features with cache
pits, 190 cache pit sites, 12 isolated finds, one signal station site, six pictograph sites, one rock
quarry, and one rock shelter (Mohs 1974:5, Figure 3, 1975:9-10; Mohs and Mohs 1976:29-31).
During the surface survey, Mohs and Mohs (1976:3) determined that most of the artifacts
observed and recorded at the lake were on Smokehouse Island despite no surface features being
identified on it during the sequential field seasons.
In 1988, a Heritage Resource Overview was conducted by Ham (1988), with a focus on
an area near Babine Lake, the Dome Mountain region. A heritage resource potential map was
created, brief visits were made throughout the study area, and the areas alongside Babine Lake
36

were assigned a high resource potential rating by Ham (1988). In 2000, an archaeological impact
assessment was conducted by Arcas Consulting Archaeologists (2000) at the request of the DFO.
Numerous cultural depressions and culturally modified trees (CMTs) were recorded.
Additionally, some of the shovel tests dug were positive for cultural materials. As a result, Arcas
Consulting Archaeologists (2000) determined the site to be of high archaeological value.
In 2010, the Lake Babine Nation signed a Memorandum of Understanding with UNBC to
foster research and training to mutually benefit both parties. Under this umbrella, the Babine
Archaeology Project was initiated as a community-based archaeological research and training
program (Rahemtulla 2012: 5). When the project began, research on the central northern interior
of British Columbia was minimal compared to the coastal and southern regions (Rahemtulla
2012: 6). The first field school and the first research excavation (2010 and 2012, respectively)
focused on the Nilkitkwa village site within southern proximity of the confluence of the Babine
and Nilkitkwa rivers (Rahemtulla 2012:6, 2013:6). The initial field school in 2010 attempted to
build a culture-history framework for the region (Rahemtulla 2012: 5). Field school participants
recovered thirty-five retouched stone artifacts and a large amount of debitage from the site
(Rahemtulla 2012:20-35). Radiocarbon results suggest the archaeological site is approximately
1200-1300 BP in age (Rahemtulla 2012:19-20).
In 2012, the excavation focused on a large house-pit depression at the village site
(Rahemtulla 2013:5). Radiocarbon dating showed a continuous occupation of the house-pit for
approximately 600 years (Rahemtulla 2013:25-26), which explains the recovery of both precontact and historic period artifacts from it (Rahemtulla 2013:28-37). The project recovered
faunal remains, precontact artifacts (including stone artifacts), and historic period artifacts
(including ceramic fragments, beads, pipe fragments, cloth fragments, gun flints, musket balls,
37

shell casings, and a stone serrated by metal) in high numbers (Rahemtulla 2013:27-37).
Ultimately, the results added to the known chronology for the site, and it is significant that these
were the first archaeological excavations in the Babine area (Rahemtulla 2019:161).

Smokehouse Island
Smokehouse Island (GiSp-001) was initially recorded as an archaeological site in the
1950s by Mike Kew (Mohs and Mohs 1976:3). In 2013, a survey near Smokehouse Island by
Farid Rahemtulla, Lake Babine Nation member Sonny West, and several students, identified fish
weir remains (Rahemtulla, In prep). Finding the fish weirs was an important goal for the Babine
community, due to the importance of fishing in their community history (Rahemtulla, In prep).
Smokehouse Island is the focus of this thesis; it is approximately 150 m x 150 m in size
(Rahemtulla 2019:162), located near the outflow of Babine Lake in the middle of the Babine
River. It was an important location for the operation and maintenance of the associated fish weirs
(Rahemtulla 2019:162). Smokehouse Island was so named for the smokehouses that were built
on it and were used to prepare fish for the winter (Rahemtulla 2019:162).
A three-week survey and excavation occurred in 2014 to further explore weir remains
and to establish the chronology of use at the site (Rahemtulla 2019:162-163). Four units (1-4)
were opened on the island's south end (Rahemtulla, In prep). The excavations resulted in the
recovery and documentation of a large number of cores, tools, and debitage (Rahemtulla
2019:163). A remnant of a weir stake and several other samples were radiocarbon dated,
showing site use as far back as 1000 BP (Rahemtulla 2019:163).

38

In 2015, a six-week excavation was conducted on the site to determine the depth of
cultural deposits, and to determine the temporal and functional uses of the site (Rahemtulla
2019:164). The water levels in the Babine watershed were much higher due to snow melt in 2015
(Rahemtulla 2019:164) and as a result, the initial units were too flooded to continue excavating,
and four more excavation units (5-8) were opened (see Figure 5; Rahemtulla, In prep). An
illustration showing the placement of the units of Smokehouse Island is below (see Figure 5).

Figure 5. Map of the excavation units on the south end of Smokehouse Island. Units 1-4 were opened and excavated
in 2014, while units 5-8 were opened and excavated during the 2015 field season (Rahemtulla, In prep).

39

Due to the high number of organic artifacts recovered, the collection and preservation
process included placing the artifacts in a container filled with water or wrapping the artifacts in
various wet cloths to be kept moist (Rahemtulla 2019:164-165). A clutch of weir posts from Unit
5 at 110 cm below the surface was recovered (Rahemtulla 2019:165). Additionally, a preserved
birch bark container and a part of a basket trap from other units were found (Rahemtulla
2019:166). Faunal material, fire-cracked rock, and both ground and chipped stone artifacts were
also recovered (Rahemtulla 2019:167).
Rahemtulla (In prep) provides details on the stratigraphy and contents for Unit 8 (see
Figures 6). While most of the units were excavated to at least 1 meter below the surface (BS),
Unit 8 went to approximately 70 cm BS. Pre-contact and historic period artifacts were recovered
from the Littermat or Humic layer. Further down, Layer A, a very dark organic silt, yielded
many chipped and ground stone artifacts, debitage, fire-cracked rock, and charcoal. Layer B has
minimal cultural material and consisted of sandy silt deposited by the Babine River. Similar in
composition to Layer A, Layer C contains at least one hearth (with the possibility of a second).
Layer D sediment includes a significant amount of charcoal and had a complex mix of alluvial
sandy silt and dark organic silt.

40

Figure 6. Profile of the west wall of Unit 8, with layer descriptions (Rahemtulla, In prep).

The 2015 excavation at GiSp-001 resulted in the recovery of over 1,700 artifacts and
thousands of pieces of debitage (Rahemtulla, In prep). Recovered artifacts include dozens of
bifaces of various sizes and reduction stages, retouched flakes of different sizes and
morphologies, and multiple unidirectional and multidirectional cores (Rahemtulla, In prep).
Hundreds of projectile points of various sizes and forms were recovered, including bipoint,
41

foliate, and stemmed points (Rahemtulla, In prep). Ground stone tools, including adzes and other
forms, were also recovered. Of significance to this project, all of the chipped stone tools were
made from the same raw materials as those represented in the debitage assemblage from Unit 8.
Bone artifacts were also numerous, and a large volume of faunal remains were identified as fish
(Rahemtulla, In prep).
This chapter has provided necessary environmental, and cultural context for the
Smokehouse Island site. Additionally, the modern scholarship focused on the Lake Babine
Nation has been reviewed, in addition to known Athapaskan, Carrier, and Babine uses of stone
tools (Matson and Magne 2007:112, 116, 127; Morice 1893:44-55, 60-65). Furthermore, this
chapter summarized known archaeological work focused on specific Athapaskan and Carrier
sites in Subarctic British Columbia, including relevant sites within the Skeena-Bulkley-Nass
archaeological region (Fladmark 2009). Lastly, the previous archaeological work conducted in
the area around Babine Lake and the numerous steps taken for the Babine Archaeology Project
were discussed (Rahemtulla 2012, 2013, 2019). All of this information provides necessary
context for the current research. The next chapter will discuss notable cases of debitage analysis
previously applied in British Columbia, in addition to summarizing the chosen methods for the
current research.

42

Chapter 3: Previous Debitage Research and Methodology

Debitage is often the only class of artifacts recovered from archaeological sites; given
this, the importance of debitage cannot be overstressed (Andrefsky 2001:2, 2009:65; Cotterell
and Kamminga 1987:675). While stone tools are often curated and transported away from the
location of their manufacture, stone debitage is usually left at the reduction site (Binford
1973:143-145). Thus, analysis of debitage at the production site can lead to an understanding on
site function, and on the variety of activities that occurred there (Binford 1973, 1979, 1980).
Flenniken (1981:267) notes: “the reduction techniques preserved in debitage are the stable
cultural markers of prehistoric knapping behavior.” Understanding debitage as part of a reductive
process is vital to interpreting and analyzing the past stone tool manufacture and use (Ahler
1989a:89-90). These factors result in a high degree of interpretive potential for debitage analysis.
A critical goal of this research is to ascertain if there are differences in how various
igneous raw materials were reduced at Smokehouse Island. The first part of this chapter details
some of the history of debitage analysis in British Columbia, and discusses the specific methods
used by various previous researchers. Following this, the methods chosen for this study are
outlined.

The History of Debitage Analysis in British Columbia
In 1970, Donahue (1973:153, 165-166) conducted an excavation at the Ulkatcho village
site in central interior British Columbia, during which 2104 pieces of debitage were recovered.
To characterize the assemblage at the site, Donahue (1973:165) initially categorized the stone
43

flakes into three distinct categories based on morphology and assumed stage of reduction:
retouch flakes, flat or thinning flakes, and block flakes. Additionally, overlapping categories of
decortication flakes and wedge-shaped flakes were used (Donahue 1973:165). All flake types
were counted, weighed, and had frequencies calculated (Donahue 1973:165-166). The most
dominant flake type was retouch flakes, followed by block flakes then flat flakes then
decortication flakes and lastly wedge-shaped flakes (Donahue 1973:166). However, block flakes
had the highest weight percentage, followed by decortication then flat flakes followed by
retouched flakes and lastly wedge-shaped (Donahue 1973:166). An attempt to understand the
spatial organization was then pursued, with most debitage appearing horizontally at the
northwest end of the site (at the exclusion of the midden feature) and at the south pit near the
identified house (Donahue 1973:165-167).
In 1974, Montgomery (1978) conducted an excavation on Area C at the Punchaw Lake
site in northern British Columbia. With the recovered stone artifacts, she conducted an analysis
on stone tools, cores, and debitage with an attempt to determine the life history of the artifacts
(Montgomery 1978:237). Part of the debitage analysis component involved creating
experimental samples for comparison (Montgomery 1978:78-81). The debitage analysis also
involved incorporating morphological characteristics (for example, striking platform types
evident), assigning typological classifications (for example, resharpening flakes, decortication
flakes, platform remnants, etc.), recording flaking type frequencies, comparing samples by raw
material, recording metric attributes (including length, width, thickness, and weight), checking
for evidence of utilization, and determining vertical placement at the site (Montgomery 1978:7376, 82-84, 86, 88-89, 91-92, 94-96, 104-105, 167, 172, 194-195, 197-199). The results suggested
that basalt was the dominant material used in Area C, and that the basalt, chert, and obsidian
44

recovered from the site had been brought to the site as cobbles and fully processed at the site
(Montgomery 1978: 70, 73, 238). Additionally, the primary reduction strategy used at the site
during the earlier occupation was core reduction (Montgomery 1978:242).
In the southern Interior Plateau of British Columbia, Pokotylo (1978) analyzed formal
stone tools and debitage from the Upper Hat Creek Valley to clarify the relationship between
lithic technology, settlement strategies, and subsistence strategies in the area. Pokotylo
(1978:204-208) suggested five variables (bulb of applied force, striking platform width, dorsal
flake scar count, ventral flaking angle, and weight) for analyzing debitage production. The
results indicated a correlation between variability in the environment, site function, and
assemblage diversity (Pokotylo 1978:322-329). Additionally, the results suggested stone
reduction occurring along a continuum, rather than in discrete stages (Pokotylo 1978:329).
To expand on Pokotylo’s (1978) research, Magne and Pokotylo (1981:36) initially used
eight variables (weight, length, width, platform width, dorsal angle, dorsal flake scar count,
striking platform flake scar count, and cortex cover) to classify debitage from the Upper Hat
Creek collections. Magne and Pokotylo (1981:38) reduced the list of variables to four (weight,
dorsal scar count, platform scar count, and cortex cover) to reduce analytical redundancies. The
remaining variables aided a classification of the debitage into two types of shatter (early and
late), and four types of platform remnant bearing flakes (biface reduction, late blank reduction,
middle reduction, and core reduction flakes; Magne and Pokotylo 1981:39). The new
classification scheme appeared reliable when used to analyze an archaeological assemblage,
based on the results (Magne and Pokotylo 1981:39-40).

45

Magne (1985) conducted a multiregional analysis on debitage and stone tools recovered
from 38 sites in the southern interior of British Columbia to determine the extent that settlement
strategies impact lithic technological variability. Deemed vital for accurate interpretations of
debitage, Magne (1985:39) used experimental archaeology to understand and classify the
different reduction stages based on debitage variability (Magne 1985:94). Magne (1985:111-114)
tested six variables to identify stages of reduction (weight, dorsal scar count, dorsal scar
complexity, platform scar count, platform angle, and cortex cover) on an experimental
assemblage instead of the four suggested by Magne and Pokotylo (1981:39).
These results suggested that the most efficient and reliable variables are platform scar
count for platform remnant bearing flakes and dorsal scar counts for shatter (Magne 1985:116125). Additionally, using bipolar flakes and bifacial reduction flakes as specific categories for
classification were deemed generally reliable indicators of specific reduction strategies (Magne
1985:126-127). The debitage and stone tools from the 38 sites were then analyzed, with the
results suggesting obsidian and chert were not more extensively conserved at the sites in
comparison to the locally obtained vitreous basalt (Magne 1985:256-257). Furthermore, lithic
assemblage variability at the sites was directly influenced by the extent of preparation and
maintenance of tools in relation to the length of occupation (Magne 1985:257). Lastly, the types
and amounts of stone tools and debitage found at a site can be used together to predict the
function of an archaeological site (Magne 1985:257-260).
Building upon the previous work by Magne (1985), Matson and Magne (2007:3-4)
sought to find out when Athapaskan populations, as represented by the Chilcotin, arrived in the
Interior Plateau at the Eagle Lake study region. To determine when the Chilcotin arrived, Matson
and Magne (2007:3-4) analyzed types of debitage, specific projectile point morphologies, the
46

frequencies and types of faunal remains, and settlement patterns to accurately differentiate
Athapaskan sites from Interior Salish (Plateau Pithouse Tradition) sites. Using Magne’s (1985)
method, Matson and Magne (2007:121-127) minimally incorporated debitage analysis into the
study. Their results show that higher utilization of stone raw material and a more significant
degree of biface manufacture are evident at Athapaskan sites compared to those associated with
the Plateau Pithouse Tradition (Matson and Magne 2007: 125-127). Additionally, Athapaskan
populations replaced Plateau Pithouse Tradition sites after AD 1475 (Matson and Magne
2007:159).
A survey and shovel testing of site GhSv-A in Hagwilget Canyon, and an excavation of
site GgSt-2 on a terrace in Moricetown Canyon were conducted by Albright (1986) in 1985.
Research was conducted on the recovered stone material to determine if cultural association was
evident based on assemblage variability, to better understand the settlement strategies used at the
site, and to understand the temporal extent that the strategies were used (Albright 1986:52-60).
To analyze the debitage, Albright (1986:50-51, 57) used Magne’s (1985) classification method.
Layer A from Moricetown had a high amount of early stage debitage, Layer B had high amounts
of late stage debitage and low amounts of early stage debitage, and Layer C had a high frequency
of late stage debitage, core preparation and core reduction debitage, and low amounts of middle
stage debitage (Albright 1986:57, Appendix E-12). In contrast, low amounts of late stage
debitage and high amount of middle stage debitage were characteristic of GhSv-A at Hagwilget
Canyon (Albright 1986:57, Appendix E-12).
Returning the focus to a site at Upper Hat Creek Valley, Kuijt and colleagues (1995)
created experimental assemblages out of trachydacite to determine if they could reliably identify
debitage from bipolar reduction. A version of the Sullivan and Rozen (1985) Technique (SRT),
47

which included the addition of split flakes and slightly reworked category names from Sullivan
(1987), was used (Kuijt et al. 1995:120-121). Results suggested bipolar reduction creates a low
percentage of complete flakes and a high percentage of medial/distal and non-orientable
fragments (Kuijt et al. 1995:122). The results also indicated that additional characteristics of
bipolar reduction include a large percentage of flakes with cortex, and flakes that are generally
small-sized due to the bipolar reduction technique typically being applied to small-sized nodules.
In south-central British Columbia, Prentiss (1993) used a modified version of the SRT
(Sullivan and Rozen 1985) alongside three flake utility indices to identify the culling of flakes
from a house-pit floor assemblage at the Keatley Creek site, and how debitage variability was
influenced both by processes and forms of risk management at the site. A principal components
analysis was also incorporated to better characterize the debitage (Prentiss 1993:574). The
embedded procurement of raw materials during the fall, and intensive use and reuse of lithic raw
materials during winter occupation are proposed to be forms of risk management practiced at the
site, in addition to both debitage variability and distribution being impacted directly by the
activities occurring at Keatley Creek (Prentiss 1993:581-601).
Prentiss (2001) expanded on his previous work at Keatley Creek (Prentiss 1993) and
further tested his modification of the SRT (Sullivan and Rozen 1985), which involved adding a
size component (Prentiss 1998). Samples that were created as a control group tested the
reliability and validity of the Modified Sullivan and Rozen Technique (MSRT) (Prentiss 2001).
The experimental results suggested the MSRT is both valid and reliable since, when used on the
archaeological samples recovered from House-pit 7 at Keatley Creek, the results revealed small
block core reduction dominated the assemblage (Prentiss 2001:169, 171). Additionally, tool
production and tool resharpening repeatedly occurred around hearth features (Prentiss 2001:168).
48

Some researchers have focused on the central coast, and areas nearby. Rahemtulla (1995)
was the first in British Columbia to use a framework involving multiple forms of debitage
analysis, which included Ahler’s (1989a) mass analysis, Magne’s (1985) flake scar method, and
the Sullivan and Rozen (1985) flake completeness method, to study the organization of lithic
technology at Namu, on the coast of British Columbia, during the Early Period (10,000-5,000
BP). The stockpiling of materials as part of sedentary/semi-sedentary occupation of the region
were crucial strategies evident at Namu (Rahemtulla 1995a:105-107).
Expanding on his previous research, Rahemtulla (2006) analyzed both formal tools and
debitage at the Namu site through the framework of Design Theory. A modified version of the
SRT (Sullivan and Rozen 1985) by Prentiss and Romanski (1989) and Baumler and Downum
(1989), mass analysis (Ahler 1989a), and Magne’s (1985) flake scar method were used. The
Namu site during the Early Period appeared to have been an intensively occupied village, based
on the wide variety of tasks evident, and on the millennia long use of a specified lithic dump area
(Rahemtulla 2006:190-191). Additionally, he found that stone tool production and organization
at Namu depended on coarse stone materials transported by watercraft to the site (Rahemtulla
2006:304), and that stockpiling was a strategy likely used at the site (Rahemtulla 2006:316-317).
Research of the Tsini Tsini site in the Talchako River Valley by the central coast focused
on recovered debitage. As part of an undergraduate honour’s thesis, Metz (1996) applied mass
analysis (Ahler 1989a) on a debitage assemblage. Additionally, a framework using multiple lines
of evidence was incorporated by Hall (1998), including mass analysis (Ahler 1989a), the
Sullivan and Rozen (1985) technique, and the complete flake analysis established by Stahle and
Dunn (1982). Determining the reduction strategies used, interpreting how the past Indigenous
occupants organized technology, and analyzing the three sites to find patterns were the goals of
49

his research (Hall 1998). A later Athapaskan assemblage with a carbon date of 500 BP was
suggested, in addition to evidence of earlier habitation at the site with ties to the coast, based on
another stone assemblage (Hall 1998:103-110). The site's most frequently occurring reduction
strategy was biface production (Hall 1998:110).
Switching the focus of this review to the southern coast of British Columbia, research by
Ritchie and colleagues (2022) examined a biface workshop identified on the Harrison River, as
part of the ancestral Sts’ailes village of YāçkEtEl. Their research involved examining recovered
bifaces based on known regional typologies, creating a reduction sequence model, analyzing the
debitage, and conducting geochemical analysis on both recovered tools and debitage (Richie et
al. 2022:6-7). The debitage variables incorporated include flake completeness, dorsal cortex,
dorsal scar count, platform facet count, size class, and weight, in addition to categorizing specific
debitage as biface reduction flakes based on criteria established by Magne (1985:160) (Ritchie et
al. 2022:7, 9). The results suggested the bifaces were initially partially reduced at another
location before being further altered at YāçkEtEl, with the primary intended products being thin
bifacial knives (Ritchie et al. 2022:10-11, 15). Additionally, based on the weight distribution of
the materials and the frequencies of identified artifacts, as few as 25 large bifaces may have been
produced at the site (Ritchie et al. 2022:15). Furthermore, most of the material was procured
locally, while the non-local sources were locations that would have been regularly visited by the
ancestral Sts’ailes (Ritchie et al. 2022:15-16). The evidence suggested the use of the workshop to
reduce large bifaces to create prestige items, which helped facilitate and maintain social
relationships (Ritchie et al. 2022:16-17).

50

Summary
Much can be concluded about the role of debitage analyses performed in British
Columbia. Since the work conducted at the Ulkatcho village site in 1970 (Donahue 1973:153,
165-166), there has been a general refinement towards the application of debitage analyses in the
province, often increasing its importance to the research pursued (Albright 1986; Hall 1998;
Kuijt et al. 1995; Magne and Pokotylo 1981; Magne 1985; Metz 1996; Montgomery 1978;
Pokotylo 1978; Prentiss 1993; 2001; Rahemtulla 1995a, 2006; Ritchie et al. 2022). When
research has continued within specific project areas in British Columbia, the forms of debitage
analysis have often been further changed and modified (e.g., Kuijt et al. 1995; Magne and
Pokotylo 1981; Magne 1985; Pokotylo 1978; Prentiss 1993; 2001). Additionally, Rahemtulla
(1995, 2006) and Hall (1998) have successfully used frameworks incorporating multiple forms
of debitage analysis to analyze material from British Columbia.
Three forms of debitage analysis are used for the current project, as discussed below
(following Rahemtulla 1995a, 2006). The first is a version of the Sullivan and Rozen (1985)
Technique (SRT) modified by Prentiss and Romanski (1989); the second is mass analysis, as
suggested by Ahler (1989a); and the third is Magne’s (1985) flake scar count method.

51

Table 1. Comparison of chosen methods, including Mass Analysis (Ahler 1989a), the Sullivan and Rozen (1985)
Technique with modifications by Prentiss and Romanski (1989), and Magne's (1985) Scar Count Method. Some
table categories from Steffen and others (1998:137) and Larson (2004:6) adapted for table.

Sullivan and Rozen

Magne’s Scar Count

Technique

Method

-Size grading

-Interpretation-free

-Platform scar counts,

-Frequency

hierarchical key

and dorsal scar counts

Distribution

-Uses three variables

used to determine

-Percentage of Weight

for analysis: the

stages of reduction

by Size Grade

presence of a single

-Platform remnant

-Cortical Flake Count

interior surface, the

bearing flakes (PRBs)

-Method determines

presence of a point of

and shatter as main

reduction stages and

applied force, and

general categories

strategies of an

how intact margins

-Bifacial reduction

assemblage

are

flakes (BRFs) and

-Differentiates core

bipolar reduction

reduction from tool

flakes (BIPs) as

production

indicators of specific

Mass Analysis
Key Aspects

reductions strategies
Speed of analysis

Fast

Moderate

Slow

Difficulty level

Easy

Moderate

Hard (initially)

Observation focus

Assemblage

Individual flake

Individual flake

52

Mass Analysis
Forms of aggregate analysis were developed in response to individual flake analysis,
which was often lacking standardization and was relatively slow to implement (Ahler 1989a:8687; Stahle and Dunn 1982:84). Individual flake analyses were often used as they allow
descriptive interpretations of stone flakes based on individual characteristics and suggest direct
past behavioral information (Ahler 1989a:86). Additionally, individual flake analysis only
incorporates individual flakes into research and not other relevant artifacts, including cores
(Andrefsky 2005:113). In contrast, analyses focused on the population of an assemblage
recognize that reduction strategies may result in a variety of debitage morphologies, not just a
single one (Andrefsky 2005:113).
Henry and colleagues (1976) is one of the earliest works to use a framework based on
aggregate analysis. Henry and colleagues (1976) produced experimental samples by softhammer, hard-hammer, and pressure flaking, and moved the samples through eight graduated
sieves (Henry et al. 1976:59). The openings ranged from 1 mm to 25 mm in size, with three of
the openings being square and five of them round (Henry et al. 1976:59). Following the
weighing and measuring of the samples, the results showed that soft-hammer and hard-hammer
debitage were not distinct from each other (Henry et al. 1976:59-61). However, pressure flaking
could reliably be separated based on mean weight and maximum thickness (Henry et al.
1976:59-61).
Following this, Stahle and Dunn (1982:85) proposed a method based on the
understanding that as stages of reduction continue, the general size of stone flakes will decrease.
Therefore, the stages of biface manufacture may be revealed by debitage size (Stahle and Dunn
1982:85). Stahle and Dunn (1982) used nine size grades based on 1/8 inch increments to test the
53

framework on experimental samples, placing them into four reduction stages. Their results
aligned with expectations that smaller grades would contain the most flakes, and, ultimately, this
framework showed promise for revealing bifacial stages of reduction in future research (Stahle
and Dunn 1982).
A more refined form of aggregate analysis is mass analysis (Ahler 1989a; Larson
2004:6). Integral to the framework of mass analysis is the reductive nature of flintknapping,
which recognizes that the more reduction occurs, the smaller and more numerous in size debitage
will become (Ahler 1989a:89-90). With mass analysis, samples should be subdivided by raw
material before the formal analysis begins (Ahler 1989a:101). Following the first step, the
assemblage is sorted through a set of nested screens, which reduces any time spent by individual
handling of samples by an analyst (Ahler 1989a:87-88). The nested screens represent size grades
identified as G1, G2, G3, and G4, which (ideally) consist of mesh sizes of 1”, ½”, ¼”, and 1/8”,
respectively (Ahler 1989a:100). This ideal range of mesh sizes allows the inclusion of smallersized debitage in the research, that might otherwise not be included (Ahler 1989a:88).
Following the size-grading of the samples, the total weight of flakes, a total count of
flakes, and the count of cortical flakes are recorded per size grade (Ahler 1989a:101; Carr and
Bradbury 2004:26). The frequency of weight across size grades is essential to compensate for the
large percentage of small-sized debitage expected for all reduction processes (Ahler 1989a:90).
Additionally, Ahler (1989a:90) noted that percentage of debitage weight by size grade is more
revealing of reduction strategies than flake counts alone.
Cortical flakes are the pieces of debitage with cortex (Ahler 1989a:90). Cortex is defined
by Ahler (1989a:90) “as any observable rind or outer surface of the original piece of raw material
54

that can be distinguished from a surface created by human flake removals or fracture processes.”
Cortex is, therefore, representative of the original piece being reduced, with its presence
gradually less evident on debitage as reduction continues (Ahler 1989a:90). The count of cortical
flakes can be used to determine the different stages of reduction, but it should only be used
cautiously as the amount of cortex present on a nodule before reduction can vary (Ahler
1989a:90; Carr and Bradbury 2004:29, 30). For example, if material is brought to a location from
a far distance or is mined, little cortex will be present (Magne 1989:19).
The mass analysis method has received numerous criticisms. Andrefsky (2009:82)
suggested that standardized mesh screens can not account for the variety of sizes that would fall
into the size grades based on shape variability. Shott and Habtzghi (2019:48) argued that a strong
correlation between flake shape and weight does not exist. Results from research by Magne
(1985:125-126) determined that weight was not an effective representation of reduction stages.
Furthermore, a few researchers found that size-grading may not accurately represent reduction
stages, as taphonomic processes, including trampling and other cultural and natural factors, can
alter debitage size distributions (Hallson 2017:29; Hiscock 1985:85; Prentiss and Romanski
1989; Shott 1994:100). Research by Andrefsky (2014a:420-421) suggested trampling will reduce
the weight of debitage, with most of the weight lost during initial trampling events.
Mixed assemblages are also problematic in mass analysis (Ahler 1989b:212). In most
cases, debitage assemblages represent multiple different reduction events and activities over a
long time (Whittaker and Kaldahl 2001:33). The debitage resulting from different tool
production and core reduction events might be thoroughly mixed (Shott 2004:222). As a result,
the debitage recovered from sites can make the use of aggregate forms of analysis more difficult
(Andrefsky 2001:4-5; Andrefsky 2007:396). Additionally, Andrefsky (2007) argued that factors
55

such as the impact of mixed assemblages, raw material properties, replicated assemblages, and
flintknapper skill on assemblages may directly impact interpretation; instead of ignoring these
factors, more stringent controls are needed when mass analysis is applied.
Mass analysis is still considered effective for analyzing large assemblages, however, as it
considers the reductive nature of flintknapping adequately, is replicable, and its simple structure
reduces analytical bias (Carr and Bradbury 2004:21-22). Additionally, the segregation of
samples by raw material type before running the analysis was built into its methodological
framework early in its inception (Ahler 1989a:101). Furthermore, research by Morrison (1994)
suggested that mass analysis is a dependable form of debitage analysis regardless of the raw
material. Lastly, Bradbury and Carr (1995:111), Carr and Bradbury (2004:43) and Rahemtulla
(1995; 2006) undertook research showing that mass analysis still provides analytical value when
used with another form of analysis. Given this, mass analysis as developed by Ahler (1989a) is
one of the three methods employed in this study. Use of mass analysis will help determine the
stages of reduction at Smokehouse Island at a relatively fast speed, which will allow inferences
to be formed regarding reduction strategies used at the site (Ahler 1989a).

Sullivan and Rozen Technique (SRT)
The Sullivan and Rozen technique was developed in response to a few common problems
with contemporary debitage analyses (Sullivan and Rozen 1985). Sullivan and Rozen (1985:756758) were critical of debitage typologies being used, with a significant issue being a notable lack
of consistent definitions used by researchers. Additionally, typologies used by analysts could not
capture the full variable range of debitage attributes that result from different reduction
sequences (Sullivan and Rozen 1985:756, 758). In response, Sullivan and Rozen (1985:758)
56

proposed that debitage analyses should be interpretation-free and provide replicable and
objective interpretations.
A hierarchy based on universal characteristics for debitage was suggested, which used the
presence or absence of a single interior surface and a striking platform, and the degree to which
margins were intact, in their classification hierarchy (Sullivan and Rozen 1985:759). The first
step of the hierarchy is to determine if a single interior surface is identifiable, based on the
presence of features including a bulb of percussion, force lines, and ripple marks, and if there are
multiple occurrences of them, or none at all (Sullivan and Rozen 1985:758-759). For example, if
a single interior surface is not evident, the piece is classified as debris, while if a single interior
surface is present, the analysis then focuses on the presence or absence of a striking platform. If
the striking platform is fragmented, evidence of a striking platform is verified by the presence of
force line radiation. The pieces of debitage without a striking platform are flake fragments. The
stone flakes with striking platforms are then checked for intact margins (Sullivan and Rozen
1985:759). If the sample ends in either a feather or hinge termination, and any lateral snaps or
breaks do not interfere with taking width measurements, the flake is a complete flake. The flakes
without intact margins are called broken flakes.
Sullivan and Rozen tested their hierarchy on chert debitage assemblages recovered during
the Pitiful Flats and TEP St. Johns Archaeological projects, which were conducted in different
locations in Arizona in the United States of America (Sullivan and Rozen 1985:760-761, 769).
The results suggested that core reduction results in a large percentage of complete flakes and
debris, while the process of biface manufacture creates high numbers of broken flakes and flake
fragments, due to their smaller sizes (Sullivan and Rozen 1985:773).

57

The problem with the SRT is that it had not faced rigorous experimentation, which
Prentiss and Romanski (1989) addressed by using an experimental component and changing the
categories. They changed the categories to: complete flakes, proximal flakes, medial/distal
flakes, nonorientable flakes, and split flakes. The changes they made reflected that a greater
degree of variability would be present in an assemblage due to the influence of taphonomic
processes, different reduction strategies, and raw material properties (Prentiss and Romanski
1989:97).
The research by Prentiss and Romanski (1989:92) indicated core reduction results in
large amounts of nonorientable fragments and moderate numbers of medial/distal fragments and
complete flakes. Additionally, a large percentage of complete flakes and medial/distal fragments
result from tool production, with only a few nonorientable fragments. Furthermore, few split
flakes and moderate proximal fragments result from both forms. These findings will be important
for the current project focused on Smokehouse Island samples.
Other criticisms of the Sullivan and Rozen (1985) technique did occur. For instance,
Ingbar and others (1989:132) tested the Sullivan and Rozen (1985) technique and asserted that
the results did not effectively distinguish tool production from core reduction. Mauldin and
Amick (1989:84) pointed out the critical flaw that the chosen attributes were more directly
influenced by the core's original size, rather than reduction strategy. Furthermore, research by
Bradbury and Carr (1995:111-112) suggested more nonorientable fragments resulting from core
reduction, more split and proximal flakes coming from biface reduction, and both medial/distal
fragments and complete flakes appearing equally from both reduction strategies.

58

In addition, Prentiss (1998) found that while the SRT passed the reliability test, it did not
appear valid. Prentiss (2001) resolved the validity problem by adding size classes to the SRT,
which maintains the reliability but increases validity. Initially, the modified SRT (Prentiss 2001)
was chosen as the second technique in this analysis but due to the lack of an available
experimental component, the SRT with modifications by Prentiss and Romanski (1989) was used
instead (Sullivan and Rozen 1985). The SRT method as altered by Prentiss and Romanski (1989)
will provide additional evidence for determining the reduction strategies used at the site in a
replicable manner (Sullivan and Rozen 1985).

The Flake Scar Count Method
Magne (1985:26-27) believed the current state of debitage analysis had no sufficient
experimental controls, not enough focus on debitage, too many redundant variables, a severe lack
of statistical evaluations, and too narrow a focus. In response, a pilot study was performed
(Magne and Pokotylo 1981) that used eight attributes to classify debitage, while placing them
into reduction stages (Magne and Pokotylo 1981:34, 36). Variables were then reduced to the four
most important elements part-way through the research: weight, dorsal scar count, platform scar
count, and cortex cover (Magne and Pokotylo 1981:38). Results from the study showed promise,
at least when applied to a debitage assemblage made from basalt in southern British Columbia
(Magne and Pokotylo 1981:40).
Expanding on the pilot study, Magne (1985:106-107) continued the pursuit for a reliable
and objective way to determine the three stages of reduction, in addition to distinguishing bipolar
and bifacial flakes from shatter and platform-remnant-bearing (PRB) flakes. Based on the
research, PRB flakes are those with a striking platform, while shatter are those with none (Magne
59

1985:100). Additionally, bifacial reduction flakes (BRF) are a special version of the PRB
category (Magne 1985:100, 107), which usually have a small striking platform area in relation to
flake size, lack of point impact features, an absence of striking platform crushing, and
pronounced lipping (Hayden and Hutchings 1989:253, 255). Bipolar reduction flakes (BPO) are
a special form of shatter (Magne 1985:100), which typically lack a positive bulb of force, often
show no distinction between dorsal and ventral surfaces, have evidence of force being applied
from different directions, and have pointed or shattered platforms (Ahler 1989b:210). However,
bipolar flakes must be identified cautiously, as Hayden (1980:5) noted bipolar cores are
frequently misidentified as bipolar flakes.
The classification system is as follows: early stages of reduction are PRBs with 0-1
platform scars or shatter with 0-1 dorsal scars and are usually associated with core reduction
(Magne 1985:128-129). Middle stage debitage includes PRBs with two platform scars and
shatter with two dorsal scars and are typically associated with the primary trimming of tools.
Late stage debitage involves PRBs with three or more platform scars and shatter with three or
more dorsal scars. Bifacial reduction flakes (BRF’s) usually fall into the late stage category
(Magne 1985:107). In comparison, bipolar flakes (BPO) are typically indicative of early stages
of reduction (Magne 1985:106).
When looking at appropriate characteristics to consider, Magne (1985:108) assessed the
results from Pokotylo (1978) and Magne and Pokotylo (1981) to decide on variables to use. To
keep analysis time relatively short and to reduce redundancies, Magne (1985:111, 113-114)
chose six variables to analyze, including weight, dorsal scar count, dorsal scar complexity,
platform scar count, platform angle, and cortex cover. Platform angle, cortex cover, and weight
should decrease as reduction continues, while dorsal scar count, dorsal scar complexity, and
60

platform scar count should increase, based on expectations by Magne (1985:113-114). However,
after running a stepwise multiple discriminant analysis and a chi-square test of independence
involving the variables, it was determined that the most reliable and accurate variables to use
were platform scar count and dorsal scar count, that the incorporation of bifacial reduction flakes
and bipolar flakes identification into analysis could reliably indicate bifacial and bipolar
reduction strategies, and that the chosen variables were reliable regardless of raw material
(Magne 1985:115-127).
There are many critiques of Magne’s (1985) method. For instance, stage models have
come under criticism, as they are not always replicable or valid, as associated variables have
been suggested by Shott (2017) to not reflect the full range of patterns from stone reduction.
Additionally, a limited number of dorsal scars can occur on small flakes (Allan 2018:29), and
platform characteristics are generally challenging to analyze and interpret (Andrefsky 2001:10).
Furthermore, Mauldin and Amick (1989:73) asserted the connection between stage of reduction
and dorsal scar counts was tenuous, making the validity of Magne’s (1985) method questionable.
Additionally, Ingbar and colleagues (1989:122-123) determined no apparent connection between
platform and dorsal scar counts and reduction stages existed, with raw material influencing the
incongruities.
Morrison (1994:93) argued that while more scars were evident for early reduction stages,
the proceeding stages had less correlation. Additionally, Morrison (1994:111) critiqued the use
of bipolar flakes to identify the reduction strategy, as the percentages for both quartzite and
obsidian bipolar flakes are very low (Morrison 1994:111). The use of typological classification is
critiqued, as bifacial reduction flakes occur in hard-hammer reduction events (Morrison

61

1994:109). Lastly, Morrison (1994:20) argued that the reduction strategy is variable and could
change at any time.
Despite this, the method developed by Magne (1985) is used as the third technique in this
study. Numerous previous studies conducted have either used elements from this method or the
entire method successfully, supporting its use (Mauldin and Amick 1989; Bradbury and Carr
1995:108-110; Ritchie et al. 2022; Shott 1994). Additionally, it has been used in several previous
analyses on debitage recovered in British Columbia (Magne 1985; Matson and Magne 2007;
Rahemtulla 1995a, 2006; Ritchie et al. 2022). Furthermore, the attributes become easier to
identify and analyze over time, making its implementation relatively easy with experience
(Rahemtulla 1995a:41). Finally, Magne’s (1985) method appears to work effectively when used
alongside other methods of debitage analysis (Bradbury 1998; Bradbury and Carr 1999; Carr and
Bradbury 2001; Rahemtulla 1995a, 2006). Using Magne’s (1985) method will allow the
determination of the reduction strategies used at the Smokehouse Island site.
As illustrated above, the use of multiple forms of debitage analysis together has been
suggested to enhance interpretations, and this has been successful in British Columbia and
elsewhere (Bradbury and Carr 1995; Carr and Bradbury 2001:134; Carr and Bradbury 2004:43;
Eren and Prendergast 2008:51; Hall 1998; Larson 2004:16-17; Rahemtulla 1995a, 2006). The
three chosen forms of analysis will be used in the current research (Ahler 1989a; Magne 1985;
Prentiss and Romanski 1989; Sullivan and Rozen 1985), though cautiously. Furthermore, they
will be compared, and their effectiveness assessed in the conclusion (Magne 2001:23-24). The
next chapter discusses application of these techniques to the Smokehouse Island sample, and
results of the analyses.

62

Chapter 4: The Debitage Sample, Application of Analytical Techniques, and
Results

This chapter presents results of analyses on a sample of 5,240 pieces of debitage
recovered from Unit 8 on Smokehouse Island. All materials are fine grained volcanics (FGV)
with 4,656 pieces (89%) classed to FGV1 (discussed below) and 584 (11%) classed to FGV2.
Unfortunately, not all of the recovered debitage from Unit 8 is included in this study due to a mix
up in the initial sorting after the excavation. As a result, numerous bags of second screened
debitage in the G3 and G4 categories were unavailable at the time of data collection (Farid
Rahemtulla, personal communication 2022). This factor will be addressed in the conclusion.
Additionally, since all of the material recovered from Layer B is from Layer A or Layer C (Farid
Rahemtulla, personal communication 2022), debitage recovered from Layer B will be omitted
from the analyses. Numerous steps were taken to complete the current research.

Characterization and Identification of Lithic Raw Materials
The first step in sorting the sample was to determine the number of unique raw materials
(sources) represented, and to create a reference collection of raw material sample (Reimer 2018,
2023). In October 2019, a Research Project Award for $3,000 from UNBC was used towards
geochemical analysis of a representative collection of the materials. Geochemical analysis in the
form of PXRF is used to characterize the archaeological stone samples, in order to infer their raw
material type (Reimer 2018, 2023). Determining raw material groupings is vital to this study, as
the goal is to test whether or not the various raw materials were reduced differentially.
63

To determine the number of raw material groups present, 265 representative samples
were chosen, based on macroscopic characteristics. These were sent to Dr. Rudy Reimer Yumks
at the Simon Fraser University archaeology lab in Burnaby, B.C., for X-Ray Fluorescence
analysis on the samples with a Bruker Tracer III-V+ portable XRF spectrometer (Reimer 2019).
X-Ray Fluorescence was ideal for the current research as it is non-destructive, relatively fast to
process, and low in cost (Reimer 2018:138, 2023).
The results showed that the majority of the debitage samples originated from five distinct,
unknown sources (Reimer 2019). Possible source locations will be discussed in the next chapter.
Additionally, the results did not determine what specific fine-grained volcanic materials were
represented (Reimer 2019). After the samples were returned, they were used as a reference
library to classify the entire debitage sample. To speed up the analysis, a recommendation was
made to collapse the five groups into a smaller number of groupings, prior to the debitage
analyses (Farid Rahemtulla, personal communication 2020). The five raw material groups were
reduced to two larger FGV raw material groups as outlined below.
In examining the number of raw material groups for the analyses, groups One, Two, and
Five appeared macroscopically similar, so they are collapsed into FGV1 (see Figure 7).
Additionally, the traits for groups Three and Four appeared to share strong visual similarities, so
they are classed to FGV2 (see Figure 8). The reduction of five groups down to two was
somewhat subjective, the result of inexperience with these igneous materials in British
Columbia. Additionally, consistent identification was difficult, as much of the overall site
contents have been exposed to burning by the inhabitants (Farid Rahemtulla, personal
communication 2022). As a result, the high heat exposure caused a significant number of the
debitage samples to display potlid fractures, flake splitting, transverse fractures, and surface
64

crazing (Patterson 1995:74-75). Furthermore, the impact of weathering on debitage
characteristics cannot be understated for the site, as the process changes the color of surfaces and
fractures stone material (Burroni et al. 2002:1278), and weathering can impact the results from
an XRF analysis, though not to an appreciable extent when applied to fine-grained volcanic
materials (Dillian 2014; Lundblad et al. 2008; Ogburn et al. 2013).

Figure 7. Combined photos showing a range of FGV1 samples.

Both material groups are described based on geological characteristics from Pough and
Scovil (1996). FGV1 has the most specimens by far (89%, as stated above), and samples from
the group appear in a variety of colors, including black, purple, grey, tan, brown, and red (see
Figure 7). Generally, the surfaces or groundmass of FGV1 samples are predominantly dark,
65

which helps with identification (Pough and Scovil 1996). When mineral inclusions are evident,
they are typically very bright and appear in high numbers (Pough and Scovil 1996). Samples
from FGV1 have a greater variety of shape morphology than FGV2 samples, though it is
important to note that this may be the result of the reduction strategies used (Andrefsky
2005:16).

Figure 8. Combined photos showing a range of FGV2 samples.

In comparison, FGV2 only makes up a small percentage of the material (11%) and
samples from this group appear as bluish, greenish, greyish, whitish, or light brownish in
coloring (see Figure 8). There is often a smoky effect evident on the groundmass of the FGV2
66

samples (Pough and Scovil 1996). Any banding on samples appears to be more consistent in tone
than FGV1 (Pough and Scovil 1996). Mineral inclusions are often smaller in frequency yet more
prominent in size than the other material group, and usually less bright in color (Pough and
Scovil 1996). FGV2 samples typically feel more like plastic in comparison to FGV1 samples.
Additionally, even in many cases when the ventral surface and margins are not intact, FGV2
samples more often appear with feathered terminations, which indicate the flake was smoothly
removed from the core without abrupt disruptions, or hinge terminations, which have a rounded
distal end due to the force of impact (Andrefsky 2005:20-21).

Mass Analysis Application and Results
Mass analysis relies on the total count of flakes, the total weight of flakes, and the total
number of cortical flakes per size grade to reveal patterns associated with core reduction or
bifacial reduction (Ahler 1989a:89-91, 93, 101). With mass analysis, all samples per recovery
bag are separated by raw material type (FGV 1 and FGV 2) and then placed through geological
sieves with mesh sizes of 1” (G1), ½” (G2), ¼” (G3), and 1/8” (G4), based on the suggestions by
Ahler (1989a:100). A triple beam balance scale was used to weigh samples by size grade, and
samples are checked for the presence of cortex (Ahler 1989a:100). Frequency distributions were
then calculated.
To ensure accuracy and attempt to keep the structural integrity of the debitage samples
intact, they were moved through the mesh screens by hand for all size grades, not just G1 and G2
as Ahler (1989a:100) suggested. Keeping the artifacts intact was essential, especially when
rechecking due to inconsistent counts. Ultimately, consistent counts were an issue throughout the
data collection process due to the large volume of material being analysed.
67

For the current research, every bag containing samples was processed through mass
analysis followed by the modified Sullivan and Rozen (1985) technique before starting another
bag (Ahler 1989a; Prentiss and Romanski 1989). This was done for the purpose of saving
analysis time. Mass analysis was the fastest method to use (Larson 2004:7), and between mass
analysis and the modified Sullivan and Rozen (1985) technique, it took approximately five
months of daily full-time lab-work to conduct both forms of analysis to completion (Ahler
1989a; Prentiss and Romanski 1989).
Table 2. Frequency Distribution for FGV1.

Count Total

G1 Frequency G2 Frequency G3 Frequency G4 Frequency

Littermat

435

3.2%

13.6%

36.6%

46.6%

Layer A

2961

2.7%

24.7%

43.2%

29.4%

Layer C

1013

2.5%

22.5%

49.3%

25.7%

While conducting mass analysis on the samples, a number of observations were made for
FGV1, as shown above in Table 2. The highest amount of debitage appears in Layer A, with
most debitage falling into Size Grade G3, followed by G4 (see Table 2). The second-highest
amount of debitage appears in Layer C, as shown in Table 2. Like Layer A, most of the debitage
is in size grade G3, with the second-highest amount appearing in G4. The third highest amount
appears in the Littermat, as shown in Table 2, where the trend for size grading falls into a
different pattern, with amounts from greatest to smallest being G4, G3, G2, G1.

68

Table 3. Frequency Distribution for FGV2.

Count Total

G1 Frequency G2 Frequency G3 Frequency G4 Frequency

Littermat

42

0.0%

9.4%

45.3%

45.3%

Layer A

409

1.7%

16.6%

39.6%

42.1%

Layer C

89

1.1%

23.6%

48.3%

27.0%

For FGV2, the highest amount of debitage appears in Layer A (see Table 3). Most of the
debitage for this layer is found in size grade G4, followed closely by G3. The third highest
frequency appears at G2, with a small percentage of debitage appearing in G1. As shown in
Table 3, Layer C has the second-highest amount of debitage, with most of the samples appearing
in G3 and the second-highest amount in G4. Lastly, the Littermat has the lowest frequency of
total debitage (see Table 3). Size grades G3 and G4 are tied for dominating this layer, followed
by G2, and no debitage found in G1. Based on the frequency percentages alone, similar trends
only appear for both materials for Layer C, with all other layers showing differences between the
material groups (see Tables 2 and 3). However, the G1 amounts are similarly low for all material
groups in all layers, and the highest percentages fall into either G3 or G4 for all layers (see
Tables 2 and 3). The results for the percentage of weight by size grade can be seen below in
Table 4 and Table 5.

69

Table 4. Percentage of weight by size grade for FGV1.

Total Weight

G1

G2

G3

G4

Percentage of

Percentage of

Percentage of

Percentage of

Weight

Weight

Weight

Weight

Littermat

648.2g

43.4%

35.8%

17.7%

3.1%

Layer A

6514.2g

33.7%

48.4%

16.4%

1.5%

Layer C

1791.8g

30.0%

46.3%

22.2%

1.5%

G1

G2

G3

G4

Percentage of

Percentage of

Percentage of

Percentage of

Weight

Weight

Weight

Weight

Table 5. Percentage of weight by size grade for FGV2.

Total Weight

Littermat

31.3g

0.0%

42.9%

51.7%

5.4%

Layer A

655.9g

19.7%

58.8%

18.0%

3.5%

Layer C

159.1g

11.9%

65.7%

20.0%

2.4%

Except for the Littermat, the G2 size debitage dominates in weight for all layers for both
raw material groups (see Tables 4 and 5). For FGV1, G1 debitage appears to have the secondhighest percentage of weight for Layer A and C, while for FGV2, the G1 debitage is only second
most dominant for Layer A, and G3 is the second highest only for Layer C (see Tables 4 and 5).
The results for percentage of weight by size grade do show some differences except for with
Layer A suggesting similar trends between raw materials. From this point, the null hypothesis
that no reduction differences appear between the raw material groups only seems supported for

70

Layer A. Observations on cortical flakes were also recorded for the Smokehouse Island debitage,
and the results are shown in Tables 6 and 7.
Table 6. Cortex profiles by size grade for FGV1.

G1

G2

G3

G4

Cortical

Cortical

Cortical

Cortical

Flakes

Flakes

Flakes

Flakes

Littermat

14.3%

22.0%

15.1%

8.4%

Layer A

65.0%

40.3%

21.4%

10.2%

Layer C

24.0%

22.4%

16.2%

6.9%

Table 7. Cortex profiles by size grade for FGV2.

G1

G2

G3

G4

Cortical

Cortical

Cortical

Cortical

Flakes

Flakes

Flakes

Flakes

Littermat

0.0%

50.0%

26.3%

21.1%

Layer A

57.1%

27.9%

20.4%

8.1%

Layer C

100.0%

28.6%

16.3%

20.8%

With the exceptions of the Littermat for FGV1 containing the third lowest frequency of
cortical flakes for G1, and FGV2 containing no cortical flakes in the G1 size grade, all of the
layers fit the expected trend of most cortical flakes falling in the largest size grade (see Tables 6
and 7; Ahler 1989a:90). For Layer A and Layer C for FGV1 and Layer A for FGV2, there is a
sequential reduction in cortical flakes per size grade. In contrast, FGV2 Layer C has G2 as
71

second highest, G4 as third highest, and G3 with the lowest frequency. These results indicate
similar reduction stages, as will be discussed in Chapter 5 (see Tables 6 and 7; Ahler 1989a:90).

Sullivan and Rozen Technique (SRT) Application and Results
As discussed previously (Chapter 3), modifications to the Sullivan and Rozen (1985)
technique made by Prentiss and Romanski (1989) were applied here and the results are displayed
in Tables 8 and 9. As shown in both tables, non-orientable fragments are the most dominant
debitage type in all layers for both materials. They are not only the most dominant debitage type,
but they are by very high percentages, showing a range from 69.6% to 81.4%. Possible reasons
for the extremely high percentages will be discussed in Chapter 5. What is consistent for this
debitage type is that Layer C contains the smallest frequency of non-orientable fragments for
both materials (see Tables 8 and 9).
The second most dominant debitage type in all layers for both materials are proximal
fragments, showing a range from 10.0% to 18.0%, with the highest amount found in Layer C for
both material groups (see Tables 8 and 9). The third most dominant debitage type for Layers A
and C for FGV1 and FGV2 are complete flakes (see Tables 8 and 9). In contrast, the Littermat
for FGV1 has the second lowest amount for the layer, while the percentage of complete flakes
for the Littermat for FGV2 is tied with split flakes (see Tables 8 and 9). Altogether, the
frequency of complete flakes shows a range from 1.8% to 7.9%, with the highest percentage of
complete flakes for both materials from Layer C.
Split fragments are the third highest occurring flake type for Layers A and C for both
materials, with split fragments overall showing a range from 3.0% to 4.8% (see Tables 8 and 9).
Layer A for both materials contains the second highest percentage of split fragments out of the
72

layers. The rarest debitage types are medial/distal fragments, with a range from 0.0% to 1.4%,
while the percentage of medial/distal fragments for Layer A for both materials is tied at 1.2%
(see Tables 10 and 11). Based on the results for the SRT, both material groups show strong
similarities for all layers, which appears to support the null hypothesis (see Tables 8 and 9;
Prentiss and Romanski 1989; Sullivan and Rozen 1985). However, multiple factors are likely
impacting the results for this method, which will be discussed in the next chapter (Baumler and
Downum 1989:107; Bradbury and Carr 2004:73; Kuijt et al. 1995:122; Patterson 1995:72-75;
Prentiss and Romanski 1989:91-92).
Table 8. Frequency distribution by modified SRT hierarchy for FGV1.

NO

PROX

COMPLETE

SPLIT

M/D

Littermat

81.4%

12.4%

1.8%

3.0%

1.4%

Layer A

80.9%

10.0%

4.5%

3.4%

1.2%

Layer C

74.9%

14.0%

5.4%

3.5%

2.2%

Table 9. Frequency distribution by modified SRT hierarchy for FGV2.

NO

PROX

COMPLETE

SPLIT

M/D

Littermat

73.7%

16.7%

4.8%

4.8%

0.0%

Layer A

78.4%

11.0%

5.2%

4.2%

1.2%

Layer C

69.6%

18.0%

7.9%

3.4%

1.1%

73

The Scar Count Method Application and Results
In this study, both PRBs and shatter were evaluated for number of scars with the aid of a
table analysis light and a 10x hand lens. After the first few weeks of doing the analysis, it was
evident that counting platform scars was too challenging to do consistently, a factor noted
previously by Andrefsky (2005:92) and Rahemtulla (1995:77). The challenging factor of
counting platform scars is why Magne (2001:23) “only counted platform scars on platforms with
depth or widths greater than 2 mm unless the scarring was still discrete and clear on smaller
ones” in his original study (Magne 1985:114). In those situations, Magne (1985:114) would
classify the flake as shatter. Switching the flake classification to shatter was not always an option
in the current research as platform width and depth were not always the issue, as striking
platform complexity and adapting to the method caused the most complications while using the
method (Andrefsky 2005:96; Rahemtulla 1995a:41).
The time-consuming nature of counting (and occasionally recounting) platform scars
became a concern for completing the data collection in a timely manner as a result. At that point,
the method was amended to enumerating only dorsal scars, following Rahemtulla’s (1995a:80,
82-84) assessment on the efficiency of dorsal scars for indicating stages of reduction. Counting
only dorsal scars did make the analysis faster as expected (Rahemtulla 1995a:80, 82-84), though
altogether running the analysis took approximately six months of almost daily full-time analysis
to complete. Results are presented in Tables 10 and 11.

74

Table 10. Frequency distribution based on Magne’s (1985) scar count method for FGV1.

Classification
Frequency for

Frequency for

Frequency for

Littermat

Layer A

Layer C

by Stage of
Reduction
Early PRB

59.8%

38.7%

33.0%

Middle PRB

12.0%

15.0%

19.2%

Late PRB

8.0%

16.0%

19.9%

BIP

10.1%

17.9%

12.7%

Early Shatter

3.0%

1.5%

1.9%

0.7%

0.7%

1.4%

Late Shatter

0.5%

0.5%

1.7%

Early BRF

0.9%

1.7%

1.5%

Middle BRF

2.3%

2.9%

2.7%

Late BRF

2.7%

5.1%

6.0%

100.00%

100.00%

100.00%

Middle
Shatter

Grand Total

Table 11. Frequency distribution based on Magne’s (1985) scar count method for FGV2.

Classification
Frequency for

Frequency for

Frequency for

Littermat

Layer A

Layer C

by Stage of
Reduction
Early PRB

52.0%

45.0%

31.5%

Middle PRB

11.9%

17.4%

25.8%

75

Late PRB

9.5%

14.4%

23.6%

BIP

14.7%

12.9%

9.0%

Early Shatter

0.0%

2.7%

1.1%

0.0%

1.0%

0.0%

Late Shatter

0.0%

1.0%

0.0%

Early BRF

0.0%

0.5%

2.2%

Middle BRF

2.4%

2.2%

3.4%

Late BRF

9.5%

2.9%

3.4%

Grand Total

100%

100%

100%

Middle
Shatter

Early PRBs are dominant for all material groups in all layers, showing a range from
33.0% to 59.8% for FGV1 (with the lowest amount from Layer C and the highest from the
Littermat), and from 31.5% to 52.0% for FGV2 (with the lowest amount from Layer C and the
highest from the Littermat; see Tables 10 and 11). For the Littermat for FGV1, and for Layers A
and C for FGV2, middle PRBs are second most dominant (see Tables 10 and 11). In contrast, for
FGV1 Layer A, middle stage PRBs are fourth most dominant, and for FGV1 Layer C and FGV2
Littermat, they are third. Middle stage PRBs show a range from 12.0% to 19.2% for FGV1 (with
the lowest percentage from the Littermat and the highest percentage from Layer C), and a range
from 11.9% to 25.8% for FGV2 (with the lowest percentage from the Littermat and the highest
percentage from Layer C; see Tables 10 and 11).
Late stage PRBs show a range from 8.0% to 19.9% for FGV1 (with the lowest percentage
from the Littermat and the highest percentage from Layer C), and a range from 9.5% to 23.6%
76

for FGV2 (with the lowest percentage from the Littermat and the highest from Layer C) (see
Tables 10 and 11). Looking at the trends for the PRBs alone, while early-stage PRBs dominate,
there is a noticeable increase in late stage PRBs for both material in comparison to the previous
stages (see Tables 10 and 11; Magne 1985:128-129). In comparison to all the considered
debitage types for both materials, late stage PRBs are the third most dominant flake type for
FGV1 Layer A and FGV2 Layer A and Layer C (see Tables 10 and 11). In contrast, they are the
fourth most occurring type in the Littermat for FGV1 and FGV2, and the second most prevalent
type in Layer C for FGV1. Furthermore, there is a noticeable increase of both middle stage and
late stage PRBs in Layer C for both materials (see Tables 10 and 11).
Bipolar flakes appear in all layers and all material groups, showing a range from 10.1%
to 17.9% for FGV1 (with the lowest percentage from the Littermat and the highest percentage
from Layer A), and from 9.0% to 14.7% for FGV2 (with the lowest percentage from the Layer C
and the highest from the Littermat) (see Tables 10 and 11). Overall, BIPs are the second most
occurring flake types for FGV1 Layer A and FGV2 Littermat, the third most dominant for FGV1
Littermat, and the fourth most dominant for the remaining layers. This contributes towards the
interpretation of the assemblage having more than one form of early stage of reduction occurring
at notable percentages at the site (see Tables 10 and 11; Magne 1985:106, 127).
Shatter in all stages appears in very small percentages for both materials (see Tables 10
and 11). Early-stage shatter shows a range from 1.5% to 3.0% for FGV1 (with the lowest
percentage from Layer A and the highest from the Littermat), and from 0.0% to 2.7% for FGV2
(with no early shatter appearing in the Littermat and the highest percentage from Layer A; see
Tables 10 and 11). Middle-stage shatter shows a range from 0.7% to 1.4% for FGV1 (with both
the Littermat and Layer A tied for 0.7%), and from 0.0% to 1.0% for FGV2 (with no middle77

stage shatter appearing for the Littermat and Layer A; see Tables 10 and 11). Late-stage shatter
shows a range from 0.5% to 1.7% for FGV1 (with the Littermat and Layer tied for 0.5%), and
from 0.0% to 1.0% for FGV2 (with both the Littermat and Layer C having none; see Tables 10
and 11).
Despite the expectation of BRFs only representing late stages of reduction (Magne
1985:107, 127), BRFs appeared in all stages for most of the layers for both materials (see Tables
10 and 11; Hayden and Hutchings 1989:253, 255). The highest amounts of BRFs do appear for
the late stages, with the only exception being the late and middle stages of BRFs for FGV2 Layer
C tied at 3.4% (see Tables 10 and 11). Late stage BRFs show a range from 2.7% to 6.0% for
FGV1 (with the lowest percentage from the Littermat and the highest percentage from Layer C),
and from 2.9% to 9.5% for FGV2 (with the lowest percentage from Layer A and the highest
percentage from the Littermat). Middle-stage BRFs are the second most dominant BRF stage
(with the previously noted exception), and they show a range from 2.3% to 2.9% for FGV1 (with
the lowest amount from the Littermat and the highest amount from Layer A), and from 2.2% to
3.4% for FGV2 (with the lowest amount from Layer A and the highest amount from Layer C; see
Tables 10 and 11).
Lastly, the early-stage BRFs make up a small percentage of the assemblage, showing a
range from 0.9% to 1.7% for FGV1 (with the lowest amount from the Littermat and the highest
amount from Layer A), and from 0.0% to 2.2% for FGV2 (with none in the Littermat and the
highest amount in Layer C) (see Tables 10 and 11). The results appear largely consistent between
FGV1 and FGV2, with early stage material dominating (see Tables 10 and 11; Magne 1985:106107, 128-129). Minor differences do exist, but they do not appear to be significant. Therefore,
the null hypothesis appears supported by the results from Magne’s (1985) scar count method (see
78

Tables 10 and 11). A discussion of the results from all three methods are presented in the next
chapter.

79

Chapter 5: Discussion
At the beginning of the current study, the null hypothesis was presented: that there are no
differences in the reduction of stone materials from multiple unknown sources. Based on the
results for mass analysis, the null hypothesis appears to be unsupported as the materials show
differences in reduction (see Tables 2 to 7; Ahler 1989a). However, the results for the Sullivan
and Rozen (1985) technique with modifications by Prentiss and Romanski (1989), and for
Magne’s (1985) scar count method suggest strong consistency regarding how the materials were
reduced, despite minor differences between the two raw material groups, indicating support for
the null hypothesis (see Tables 8 to 11). A discussion of the results is required to fully interpret,
understand, and expand on the results for all three methods used (Ahler 1989a; Magne 1985;
Prentiss and Romanski 1989; Sullivan and Rozen 1985).
To aid in the discussion, the results from the current study will be compared with those
from other publications, a strategy previously used by Hall (1998) and Rahemtulla (1995a).
When possible, the results will be compared to studies focused on similar raw materials, as raw
material types directly impact breakage characteristics (Amick and Mauldin 1997). The
discussion will first focus on the results from mass analysis (Ahler 1989a), followed by those
from the Sullivan and Rozen (1985) Technique, with modifications by Prentiss and Romanski
(1989), and those from the Magne’s (1985) scar count method (see Tables 2 to 11).

Discussion of Results and Interpretations
In the mass analysis results, G1 flakes generally appear in smaller frequencies for FGV2
(see Tables 2 and 3). The frequency of G1 flakes is most likely reflective of the original sizes of
80

the nodules used, with the noticeably lower G1 flakes for FGV2 likely indicating smaller sized
nodules were being reduced (Andrefsky 2005:98; Ahler 1989a:89-90). Additionally, Hall
(1998:64-65, 76-77) examined fine-grained andesite material from the Tsini Tsini site, which
revealed a dominance of G3 flakes. Hall (1998:76-79) interpreted the dominance of G3 flakes
and the low frequency of G1 flakes to indicate middle stage biface reduction being dominant.
This interpretation may be applicable for both FGV1 and FGV2, as both have two layers where
G3 dominates (see Tables 2 and 3). Adding in the size expectations from Ahler (1989a:89-90),
the size grading for the Littermat for both materials show trends indicating middle to late stages
of reduction (see Tables 2 and 3). Due to the influence of the original nodule size on the results
(Andrefsky 2005:98), the suggested stages of reduction at Smokehouse Island can only be
interpreted with caution and are possibly erroneous (Ahler 1989a:89-90).
If the interpretations of stone reduction stages at Smokehouse Island, based on Ahler
(1989:89-90) and Hall (1998:76-79), were accurate, it would suggest a degree of continuity
between how the different material groups were reduced (see Tables 2 and 3; Ahler 1989a:8990). However, these results may also be due to a number of critical uncontrolled variables.
Lower G1 counts can result from curation, with larger-sized flakes and blanks being taken to be
further retouched and used at a different location (Larson and Kornfeld 1997:9). For example,
research by Prentiss (1993:575, 585) suggested that the primary flake type culled at the Keatley
Creek site were large flakes with acute angles from prepared core reduction assemblages, with
the culling from bifacial reduction assemblages less focused on size and more on acute angles.
Unmodified flakes can serve various purposes without modification (Manclossi and Rosen 2019;
Rahemtulla 1995b, as cited in Rahemtulla 2006:95; Stemp et al. 2021). Equally of relevance, the
results of G4 counts for both material groups in half of the layers do not meet the expectation of
81

the smallest-sized debitage being the highest in number (see Tables 2 and 3; Ahler 1989a:90).
The exceptions to this are the Littermat for FGV1 and the Littermat and Layer A for FGV2 (see
Table 2 and 3). This might be the direct result of the omitted material (Farid Rahemtulla,
personal communication 2022), though this will remain speculative until the missing samples are
incorporated.
A few patterns are evident when comparing the weight percentages by size grade with
results provided by Ahler (1989a:107). FGV1 appears closest to hard-hammer flake production
technique for most layers (see Tables 4 and 5). In contrast, FGV2 most closely matches biface
resharpening for the Littermat, bipolar reduction for Layer A, and biface reduction for the
remaining layer. Hall (1998:72-78) indicates a dominance of G2 flake weight followed by a
smaller weight percentage for G3 flakes and a tiny frequency of G1 flakes suggests that midstage biface production dominates. FGV2 from the current analysis loosely matches the trend
with the Littermat and Layer C (see Tables 4 and 5). This interpretation loosely correlates with
the experimental results of Ahler (1989a:107; Hall 1998:72-78). The interpretation of the results
for percentage of weight by size grade do not support the null hypothesis, as the material groups
are suggested to have been reduced by different strategies (see Tables 4 and 5; Ahler 1989a:107).
For cortical profiles, when the results are compared to those of Ahler (1989a:94), some
similar patterns appear, though in general the percentage of cortical flakes is noticeably lower at
Smokehouse Island (see Tables 6 and 7). The Littermat for FGV1 and FGV2 loosely fits the
results for soft hammer flake tool production with Knife River flint, though Ahler’s (1989a:94)
results have no G1 and G2 cortical flakes (see Tables 6 and 7). In comparison, Layer A for both
materials loosely aligns with Ahler’s (1989a:94) results for hard hammer flake production/core
reduction of knife river flint (see Tables 6 and 7). For FGV1 Layer C, there is some degree of
82

alignment with Ahler’s (1989a:94) percentages for knife river flint hard-hammer stage 2 edging,
while FGV2 Layer C most closely fits the results for knife river flint soft hammer stage 3-4
biface thinning, though G3 from FGV2 Layer is unusually lower than the percentage for G4 (see
Tables 6 and 7). These interpretations suggest similar reduction strategies used for both FGV1
and FGV2, with both material groups experiencing the same changed strategies over time (see
Tables 6 and 7; Ahler 1989a:94). However, the material from Layer C may have been moderate
to heavily reduced before arriving at the site, depending on the distance from the quarry to
Smokehouse Island (Beck et al. 2002; Magne 1989:19). Without identifying material sources,
this remains conjecture (Reimer 2018:138-140, 2019, 2023).
It must be acknowledged that any analysis involving the presence or absence of cortex is
fraught with problems. Magne (1985:114) found that cortical flakes only appear during the
earliest reduction stages. Additionally, nodule size directly impacts the number of cortical flakes
more directly than the reduction stage (Bradbury and Carr 1995:105-106). The number of
cortical flakes is also influenced by the amount of cortex on the nodules, which will not be
consistent (Andrefsky 2005:104). Furthermore, cortex can be difficult to identify (Andrefsky
2005:103), a factor made more complicated by heat damage (Patterson 1995:74-75) and
weathering (Burroni et al. 2002). While such concerns must be acknowledged, in this study
identifying cortex was relatively easy to do consistently with these materials. In addition, the
incorporation of cortical profiles is considered an essential part of mass analysis (Ahler
1989a:90; Bradbury and Carr 1995:112).
Technological mixing is a problem in mass analysis (Ahler 1989b:212). This site
represents multiple reduction events occurring over a long period, and therefore it is a
technologically mixed assemblage (Shott 2004:222). The variety and sizes of non-debitage tools
83

recovered during excavations lend support to this notion (Rahemtulla 2019:163, In prep). The
technological mixing problem makes reduction signatures challenging to assign to production
events (Andrefsky 2007:396). For the most accurate results, samples should be sorted to be as
closely representative of the original cobble before running the analysis (Shott 2004:221). In this
study, segregation remained at the raw material level; therefore, the issue of technological
mixing is applicable throughout (Shott 2004:221). This might explain why the inferred reduction
strategies are not consistent between frequency distribution, percentage of weight by size grade,
and the count of cortical flakes at Smokehouse Island (see Tables 2 to 7; Ahler 1989a; Andrefsky
2007:396).
Regarding the Sullivan and Rozen (1985) technique with modifications by Prentiss and
Romanski (1989), the results for Smokehouse Island show consistency regarding how the
different raw materials were reduced, and therefore show support for the null hypothesis (see
Tables 8 and 9). For example, both FGV1 and FGV2 show a low to moderate percentage of
complete flakes and a low percentage of medial/distal fragments (see Tables 8 and 9). In
comparison, Prentiss and Romanski (1989:91-92) showed the percentage of complete flakes
from tool production ranging from 30.1-34.1% and from core reduction 16.3-25.4%.
Experimental testing by Rahemtulla (2006:177) showed the same trend. If tool production was
dominant at Smokehouse Island, there should then be a higher percentage of complete flakes
than what is shown in Tables 8 and 9.
Additionally, the medial/distal results for FGV1 and FGV2 are significantly lower than
research by Prentiss and Romanski (1989:91) and Rahemtulla (2006:177) for either strategy (see
Tables 8 and 9). From an initial assessment, trampling does not appear to have been a factor
influencing the assemblage, as trampling should increase the percentage of medial/distal
84

fragments (Prentiss and Romanski 1989:91). A possible answer for the lower medial/distal
percentages may lie with medial/distal fragments dominating a smaller size range of debitage
when produced by tool production (Prentiss 2001:171), their relative absence a possible
byproduct of the missing G3 and G4 material (Farid Rahemtulla, personal communication 2022).
However, the possibility of core reduction dominating the assemblage is high (Prentiss and
Romanski 1989:91-92).
Non-orientable fragments appear in unusually high numbers in Unit 8 (see Tables 8 and
9). The likely explanation for the high occurrence of non-orientable fragments is that the
debitage is primarily the product of core reduction, as suggested by Prentiss and Romanski
(1989:91-92). Their research demonstrated that the percentages of non-orientable fragments fit a
range of 31.7%-32.6% during core reduction and 4-10% in tool production. Additionally,
Baumler and Downum (1989:107) showed the percentage of non-orientable fragments fitting a
range of 29.1%-37.0% for core reduction and 1.3-10.8% for tool production. However, high
percentages of non-orientable fragments may also result from bipolar reduction (Bradbury and
Carr 2004:73; Kuijt et al. 1995:122). Kuijt and colleagues (1995:120, 122) used trachydacite
from the Interior of British Columbia in their study with results suggesting bipolar reduction
increasing the percentage of non-orientable fragments.
Another reason for the high percentage of non-orientable fragments could be the impact
of heat damage, as effects such as potlidding and surface crazing are evident on many of the
stone flakes (Patterson 1995:72-75). Additionally, the impact of trampling on debitage
morphology can be significant (Rasic 2004:127). However, based on experiments by Prentiss
and Romanski (1989:91), trampling should decrease the percentage of non-orientable fragments,
or if the percentage increases, it should be minimal. The current findings contradict the
85

expectation made by Prentiss and Romanski (1989:91), while suggesting support for the null
hypothesis of the current study (see Tables 8 and 9). Based on the evidence, the most likely
explanations for the high percentage of non-orientable fragments are a heavy influence from
early stage reduction, through core reduction and bipolar reduction strategies (Baumler and
Downum 1989:107; Bradbury and Carr 2004:73; Kuijt et al. 1995:122; Prentiss and Romanski
1989:91-92), and the impact of heat damage on many of the samples (see Tables 8 and 9;
Patterson 1995:72-75).
Looking at the results for Magne’s (1985:106-107, 128-129) scar count method, the early
stages of reduction are dominant for all raw material groups (see Tables 10 and 11).
Additionally, the percentages of middle and late stage debitage, and bipolar flakes and BRFs are
relatively close between FGV1 and FGV2 (Magne 1985:106-107, 128-129). Therefore, the
results for Magne’s (1985) scar count method show strong and consistent support for the null
hypothesis (see Tables 10 and 11). How this equates to effectiveness for the current study in
comparison to the other methods will be addressed in the conclusion (Magne 2001:24).
When comparing the results to the sites analyzed by Magne (1985:256-257), he found the
artifacts made from basalt, chert, and obsidian were mostly reduced to the same extent.
Additionally, the sites at Eagle Lake from research by Magne (1985:198) have the largest range
of early stage debitage, from 16.24% to 83.68% (see Tables 10 and 11). The next largest range
for early stage reduction are lithic scatter sites from the Hat Creek region, showing a range from
14.67% to 58.43% (Magne 1985:198). The region with the third largest range for early stages of
reduction are the Lillooet sites, showing a range from 16.67% to 49.5% for early stage debitage
(Magne 1985:198). Magne (1985:199, 202) noted that site EeRk-7 from the Lillooet region
contained bipolar flakes, bipolar cores, and cores indicating an early stage dominant site that also
86

functioned an excavated housepit site. However, Magne (1985:200, 202) found the stage
distribution at EeRk-7 more wide-spread at the site, placing it more within the middle stages of
reduction overall.
Lastly, the Mouth of Chilcotin region shows an early stage reduction debitage range from
19.23% to 41.67% (Magne 1985:198). At the Mouth of Chilcotin, site 12:6 is classified as a
lithic scatter with housepit site containing no late stage debitage and a large number of bipolar
cores, indicating a dominance of early stage reduction (Magne 1985:199, 202). When comparing
the results for 12:6 to those from the current study, the Smokehouse Island debitage is not
dominated by bipolar reduction, with the most dominant debitage type being early PRBs (see
Tables 10 and 11) (Magne 1985: 106, 199). Additionally, the site type is closer than lithic scatter
sites and it is located near the Fraser River, a prime fishing location (Fladmark 2009:558, 560562; Magne 1985:144, 202; Rahemtulla 2019:162-163).
Research by Matson and Magne (2007:123, Table 33) did not find any substantial
patterns regarding specific raw material use, though their focus was on assigning ethnicity to
assemblages, and they asserted most stone tools from interior British Columbia were made from
dacites (Magne and Matson 2010:219). When comparing the results from the current study to
those from Matson and Magne (2007:121, 123, 126, Table 33) on debitage from the Bear Lake
Athapaskan house site in the Eagle Lake area and the Chinlac village Athapaskan site, there is
some alignment with the results from Chinlac (see Tables 10 and 11). Chinlac shows a relatively
high frequency of early stage debitage, though the frequency for late stage debitage is higher
(Matson and Magne 2007:123). Additionally, the frequency of middle stage debitage is relatively
low though slightly higher than that for BRFs (Matson and Magne 2007:123, Table 33). This is
in alignment with Albright’s (1986:54, 57-58, Appendix E-9, Appendix E-12) results for
87

debitage from Layer C at the Moricetown Canyon site, GgSt-2, which she determined
represented a generalized adaptation in comparison to layers A and layer B, which represented
intensive activities occurring at the site and Athapaskan occupation. Both Chinlac and the
Moricetown Canyon site have noticeably higher percentages than the samples from Smokehouse
Island, so it is not in perfect alignment, but they are all prime fishing locations, another factor
adding similarity to Smokehouse Island (see Tables 10 and 11; Albright 1986:52, 57; Cranny
1986; Matson and Magne 2007:123, 126). The debitage results for layer A at GgSt-2 are a closer
match than for layer C at Moricetown or the assemblage at Chinlac, as early stage reduction
dominates in layer A (see Tables 10 and 11; Albright 1986:57-58, Appendix E-9, Appendix E12; Matson and Magne 2007:123, Table 33).
In contrast, the results from Bear Lake are not a good match, as the assemblage is
dominated by late stage debitage, followed by middle stage debitage (see Tables 10 and 11;
Matson and Magne 2007:123). Additionally, the frequency for BRFs is noticeably higher than
the frequency for early or middle stages of reduction individually (Matson and Magne
2007:123). The variation may be a matter of difference in site function, as the Bear Lake site is a
winter habitation site with a noticeable focus on large-mammal hunting (Matson and Magne
2007:127, 130, 158). In comparison, both Chinlac and the Smokehouse Island site were occupied
during part of the summer and fall months, and were sites where the procurement and processing
of salmon occurred (Cranny 1986:ii, 3, 144-145; Rahemtulla 2019:161-163; Tobey 1981:425).
The Moricetown Canyon site is also a known site used for fish procurement (Albright 1986:52,
54). Possible reasons for a higher frequency of early stage material recovered at known fish
procurement and processing sites include stone flakes possibly being used to process small
volumes of fish, as they are not efficient at processing large volumes of fish (Rahemtulla 1995b,
88

as cited in Rahemtulla 2006:95), or being used to create and modify parts of fish weirs
(Andrefsky 2014b; Kantakis 2017:13, 18, 20-21; Morice 1893:84-90; Rowland and Ulm 2011:
40), possibilities which will be elaborated upon and assessed in the conclusion.
For confirmation, comparisons with non-Athapaskan sites should be made. The results
from Matson and Magne (2007:122-123, Table 33) on the PPT-associated Boyd and Shields
house-sites show late stage debitage dominating, with early stage debitage showing second
highest frequencies in most cases. The results from the Boyd and Shield house-sites show a
similar trend to the debitage frequencies from the Chinlac site, despite both sites being farther
away from a river (Matson and Magne 2007:67, 123, Table 33). However, both the Shields and
Boyd sites are located near the eastern shore of Eagle Lake, and fish remains were recovered
from the Shields site (Matson and Magne 2007:85, 96-97, 123, Table 22, Table 33).
When looking at PPT riverine sites with a similar function to Smokehouse Island, the
closest match from the research by Magne and Matson (2007:123, 127) is site CR92 at Brittany
Creek, which was interpreted as being a mixed PPT and Athapaskan riverside fishing site (see
Tables 10 and 11). The early stage debitage at CR92 dominates, while the frequency for middle
stage debitage is less than half of the frequency for early stage debitage, and late stage has a
frequency slightly lower than middle stage, with a small amount of BRFs represented (Matson
and Magne 2007:123, 127). Additionally, the PPT associated riverine fishing site, 19-1, has early
stages of reduction dominate, followed by a slightly lower frequency for middle, and an even
lower frequency for late stage debitage. It must be stated that the Bear Lake site, the Chinlac
village site, The Boyd house-sites, the Shields house-sites, site CR92, and site 19-1 have larger
frequencies of shatter in comparison to the results from Smokehouse Island (see Tables 10 and
11; Matson and Magne 2007:123, Table 33).
89

Chapter 6: Conclusion

The focus of the current research has been to determine if the raw materials recovered
from Unit 8 on Smokehouse Island were worked differently. The null hypothesis was made that
there are no differences evident in the reduction of stone materials from multiple unknown
sources. The results from the current research largely support the null hypothesis, with early
stages of reduction dominating for both material groups (see Tables 6 to 11). Similarly, both
material groups showed a noticeable increase of middle and late stage debitage for Layer C,
though they appear in lower frequencies than early stage debitage (see Tables 6 to 11). While
minor differences do exist, they do not appear to be significant (see Tables 8 to 11). The only
method with results that contradict the null hypothesis was mass analysis, which also had
components indicating contradictory interpretations (see Tables 2 to 7; Ahler 1989a). Since the
results from the Sullivan and Rozen (1985) technique with modifications by Prentiss and
Romanski (1985), and Magne’s (1985) scar count method suggest consistency regarding how the
raw material from multiple unknown sources were reduced, it is likely that the results from mass
analysis were impacted by technological mixing and therefore are largely inaccurate (Ahler
1989a; Andrefsky 2007:396).
Based on the difficulties of conducting the different forms of analyses and the results
from the research, the method most conducive to direct interpretations appears to be Magne’s
(1985) scar count method. Running the method may be more time-consuming, but the variables
become increasingly easier to identify and examine over time (Rahemtulla 1995a:41), and it
became obvious why the method as a whole or in part has been repeatedly incorporated into
90

other research (Amick and Mauldin 1989; Bradbury and Carr 1995:108-110; Magne 1985;
Matson and Magne 2007; Rahemtulla 1995a, 2006; Ritchie et al. 2022; Shott 1994). In
comparison, the biggest problem with running mass analysis is that the technological mixing
problem makes reduction signatures challenging to assign to production events (Andrefsky
2007:396). Logically, it makes sense why Ahler (1989b:212), Andrefsky (2007:396), and Shott
and Habtzghi (2019) considered technological mixing a considerable problem for interpreting
debitage with mass analysis.
As discussed previously, the impact of technological mixing has evidently influenced the
results of mass analysis within the current research, providing incompatible interpretations
between the different sections of the method (see Tables 2 to 7; Andrefsky 2007:396). A possible
solution to technological mixing is an experimental assemblage for comparison (Andrefsky
2007:399). Unfortunately, no experimental assemblage was available due to inexperience
flintknapping similar materials at the time of data collection; therefore, there was no
experimental component for comparison. The lack of an experimental component required the
results to be compared to results published in other comparable studies (Hall 1998; Rahemtulla
1995a). Samples were separated by raw material before running the analysis, which is an
essential step in reducing the effects of the technological mixing problem (Andresfky 2007:399).
Ideally, one would further segregate material as close to the original nodules as possible (Shott
2004:221-222), which due to limited experience was not done in this study.
The biggest problem with running the Sullivan and Rozen (1985) technique, with
modifications by Prentiss and Romanski (1989), is the direct impact heat damage appears to have
on the interpretation of specific characteristics that are vital to the method (Patterson 1995:72-75;
Sullivan and Rozen 1985:758-759). The impact of heat damage on the assemblage from Unit 8
91

was huge, likely inflating the percentage of non-orientable fragments (see Tables 8 and 9;
Patterson 1995:72-75; Prentiss and Romanski 1989:91-92). While a primary cause for the high
percentage of non-orientable fragments is the impact of core reduction, further confirmed by the
results from Magne’s (1985) scar count method, the percentages are still exponentially higher
than expected (Prentiss and Romanski 1989:91-92). Following this research, the method should
not be applied to any assemblage with evidence of heat damage (Patterson 1995:72-75; Prentiss
and Romanski 1989:91-92; Sullivan and Rozen 1985:758-759).
It is important to note that the interpretations made in the current research are based on
incomplete data, as there were bags of G3 and G4 sized debitage missing at the time of data
collection (Rahemtulla, personal communication 2022). If the missing bags had been included,
the results could have been impacted in numerous ways. For instance, if the percentages of G3
and G4 sized debitage were high after the missing samples were added, this would likely
increase the evidence of middle-to-late stages of reduction based on mass analysis (Ahler
1989a:89-93). Additionally, if the percentage of medial/distal flakes increased exponentially, this
would suggest a higher amount of tool production occurring (Prentiss 2001:171; Prentiss and
Romanski 1989:91-92). However, if a high percentage of the small-sized debitage has cortex,
this could be indicative of a greater influence of bipolar reduction, as the technique was typically
applied to small-sized material in the past (Kuijt et al. 1995:122). Furthermore, a higher number
of cortical flakes could also be indicative of core reduction (Ahler 1989a:90). Until the missing
debitage is analyzed, how the results will be impacted remains speculation.
Site function likely impacts the assemblage at Smokehouse Island (Binford 1973:146,
1980:10, 12, 15; Binford and Binford 1969). With both igneous material groups primarily
subjected to core reduction and with Smokehouse Island having a long history of use for
92

procuring and processing fish (Rahemtulla 2019:161-162), it is possible that core reduction
strategies were used at the site for the purpose of creating flakes to cut and process fish (see
Tables 8 to 11; Rahemtulla 1995b, as cited in Rahemtulla 2006:95; Stemp et al. 2021:8-10).
However, research by Rahemtulla (1995b, as cited in Rahemtulla 2006:95) suggested that flakes
would only be useful for processing small volumes of fish, and that processing large volumes of
fish required specialized cutting tools. Considering the Babine were capturing large volumes of
fish with fish weirs (Harris 2001:84-86; Kantakis 2017:101-102), the use of flakes for processing
fish was probably a less common occurrence on Smokehouse Island and likely would not have
been the primary use of unmodified flakes (Rahemtulla 1995b, as cited in Rahemtulla 2006:95).
Experimental archaeological research by Andrefsky (2014b) suggested that unmodified
flakes are more efficient for cutting and modifying wood than modified flakes. Additionally, a
few previous ethnographic studies conducted in other locations around the world have shown
some Indigenous populations to have used and preferred unmodified flakes for wood-working in
the not-too-distant past (Gould et al. 1971; Hayden 1977; Miller 1979). Considering Smokehouse
Island is a location where fish weirs were used (Rahemtulla 2019:161-162), it is also possible
that the Babine used unmodified flakes to work and maintain the fastening material and smaller
stakes used as part of the barricade, and to create and maintain weir traps or baskets (Andrefsky
2014b; Kantakis 2017:13, 18, 20-21; Morice 1893:84-90). Wooden fish weirs required regular
maintenance in order to remain efficient and continue to work, which implies the need to
regularly create tools useful for maintaining and shaping the individual parts of a weir, such as
unmodified flakes (Andrefsky 2014b; Gould et al. 1971; Hayden 1977; Kantakis 2017:18; Miller
1979; Rowland and Ulm 2011: 40).

93

As the use of fish weirs was a traditional practice for Carrier populations (Cranny
1986:26; Gottesfeld 1994:454; Morice 1893:84-90; Prince 2014:123-127), this might explain
why Layer A at the Wet’suwet’en site at Moricetown Canyon and the assemblage from Chinlac
village had high percentages of early stage debitage (Albright 1986:52, 54, 57-58, Appendix E12; Matson and Magne 2007:123, 126, Table 33). Afterall, fish weirs were important to the
Wet’suwet’en for procuring fish (Gottesfeld 1994:454), and evidence suggests fish weirs were
likely used by the inhabitants of Chinlac (Cranny 1986:82, 87). The construction, maintenance,
and use of fish weirs might also be why the Interior Salish population that used the Shields
house-site by Eagle Lake, the Athapaskan and Interior Salish populations that used the Brittany
Creek fishing site, and the Interior Salish population that used site 19-1 practiced early stage
reduction strategies, as fish weirs are known to have been utilized by different groups throughout
the interior of British Columbia (Gottesfeld 1994:454; Matson and Magne 2007:123, 126-127,
Table 33; Morin et al. 2021:5; Prince 2005, 2014:122-123). Additionally, if those sites involved
processing small volumes of fish, unmodified flakes would have been ideal for the task
(Rahemtulla 1995b, as cited in Rahemtulla 2006:95). Further research is necessary for
confirmation.
Errors are possible. Measurement errors occur with different forms of debitage analysis, a
factor ignored too frequently by analysts (Shott 1994:74-75). It is then crucial for our analyses to
be as replicable, precise, and accurate as possible (Shott 1994:75). Methods like Magne’s (1985)
scar count method require training and knowledge (Rahemtulla 1995a:41). There is always the
possibility of some error being introduced into the current study as a result of minimal previous
training (Andrefsky 2009:83).

94

Future research should include the analysis of utilized flakes from the Smokehouse Island
site. While conducting this project, utilized flakes were separated from the sample bags, and not
included in the analysis. In hindsight, the incorporation of utilized flakes into the analysis would
have been beneficial (Albright 1986:Appendix E-8; Borden 1952:32; Donahue 1973:164-165;
Stemp et al. 2021). Additionally, future research should incorporate the missing G3 and G4
samples (Farid Rahemtulla, personal communication 2022), and all of the debitage recovered
from Smokehouse Island should be analyzed (Rahemtulla 2019:163, In prep).
Other avenues for future research include conducting a source survey within the area
surrounding Smokehouse Island (MacIntyre et al. 2001a; Reimer 2018:139-140, 2023). Lithic
samples from Smokehouse Island were geochemically analyzed for the current research, with no
identified sources (Reimer 2019). While it is likely that the materials were located from nearby
based on proximity (see Figure 4; MacIntyre et al. 2001a), conducting a source survey to
characterize the material outcrops and exposures within the immediate local area and the larger
Babine-Takla Lake region would help determine the exact sources (MacIntyre et al. 2001b:582;
MacIntyre et al. 2001a; Reimer 2018:138-140, 2023). This may also help provide more
information regarding the role of watercraft in the procurement of stone materials by the Lake
Babine Nation (Ames 2002:35, 42, 44-45, 47; Blair 2010:40, 42; Rahemtulla 2006:314-316, 329330).

95

Bibliography
Ahler, Stanley A.
1989a Mass Analysis of Flaking Debris: Studying the Forest Rather than the Tree. In
Alternative Approaches to Lithic Analysis, edited by D.O. Henry and G.H. Odell, pp. 85118. Archaeological Papers 1. American Anthropological Association, Arlington.
1989b Experimental Knapping with KRF and Midcontinent Cherts: Overview and
Applications. In Experiments in Lithic Technology, edited by Daniel S. Amick and
Raymond P. Mauldin, pp. 199-234. BAR International Series 528. British Archaeological
Reports, Oxford.
Albright, Sylvia
1986 Report on 1985 Archaeological Investigations of Gitksan-Wetsuwet'en Villages.
Manuscript on file, Gitksan-Wetsuwet'en Tribal Council, Smithers, BC.
1987 Archaeological Evidence of Gitxsan and Wetsuwet’en History. Manuscript prepared
for Delgamuukw Supreme Court of Canada Court Proceedings. Hazelton, BC.
Allaire, Louis
1979 The cultural sequence at Gitaus: a case of prehistoric acculturation. In Skeena River
Prehistory, edited by Richard Inglis and G. MacDonald, pp. 18-52. National Museum of
Man, Archaeological Survey of Canada, Mercury Series No. 87, Ottawa.
Allaire, Louis, and George MacDonald
1971 Mapping and excavations at the Fortress of the Kitselas Canyon, B.C.: preliminary
report. Canadian Archaeological Association Bulletin 3:49-55.
Allaire, Louis, George MacDonald, and Richard Inglis
1979 Gitlaxdzawk: Ethnohistory and Archaeology. In Skeena River Prehistory, edited by
Richard Inglis and G. MacDonald, pp. 53-166. Archaeological Survey of Canada Paper
No. 87. National Museum of Man, Ottawa.
Allan, Timothy E.
96

2018 Testing debitage typologies with statistical analysis: Experimental inferences upon
archaeological material from FaPx-1, a sub-alpine hunting camp in the Alberta Rockies.
Archaeological Survey of Alberta Occasional Paper 38:25–33.
Ames, K.
1979a Report of excavations at GhSv2, Hagwilget Canyon. In Skeena River Prehistory,
edited by Richard Inglis and G. MacDonald, pp. 181-218. Archaeological Survey of
Canada Paper No. 87. National Museum of Man, Ottawa.
1979b Stable and Resilient Systems Along the Skeena River: The Gitksan/Carrier Boundary.
In Skeena River Prehistory, edited by Richard Inglis and G. MacDonald, pp. 219-243.
Archaeological Survey of Canada Paper No. 87. National Museum of Man, Ottawa.
2002 Going by boat: the forager–collector continuum at sea. In Beyond Foraging and
Collecting: Evolutionary Change in Hunter–Gatherer Settlement Systems, edited by B.
Fitzhugh and J. Habu, pp. 19–52. Kluwer Academic/Plenum Publishers, New York.
Amick, Daniel S., and Raymond Mauldin
1997 Effects of Raw Material on Flake Breakage Patterns. Lithic Technology 22(1):18-32.
Andrefsky, William Jr.
2001 Emerging Directions in Debitage Analysis. In Lithic Debitage: Context, Form,
Meaning, edited by William Andrefsky Jr., pp. 2-14.
2005 Lithics: Macroscopic Approaches to Analysis. 2nd edition. Cambridge University
Press, Cambridge.
2007 The application and misapplication of mass analysis in lithic debitage studies. Journal
of Archaeological Science 34:392-402.
2009 The Analysis of Stone Tool Procurement, Production, and Maintenance. Journal of
Archaeological Research 17(1):65-103.
2014a Fingerprinting Flake Production and Damage Processes: Toward Identifying Human
Artifact Characteristics. In Paleoamerican Odyssey, edited by Kelly E. Graf, Caroline V.
Ketron, and Michael R. Waters, pp.415-428. Texas A&M University Press, College
Station.
97

2014b Debris, Debitage or Tools: Unmodified Flakes and Cutting Efficiency. In Lithics in
the West: Using Lithic Analysis to Solve Archaeological Problems in Western North
America, edited by Douglas H. MacDonald, William Andrefsky, Jr., and Pei-Lin Yu, pp.
4-16. University of Montana Press, Missoula.
Arcas Consulting Archaeologists
2000 DFO Babine River Fish Hatchery access road Archaeological Impact Assessment.
Permit report 200-282. Report on file, Archaeology Branch, Ministry of Forests, Lands
and Natural Resource Operations, Victoria, B.C.
Baumler, M.F., and C.E. Downum
1989 Between Micro and Macro: A Study in the Interpretation of Small-Sized Lithic
Debitage. In Experiments in Lithic Technology, edited by D.S. Amick and R.P. Mauldin,
pp. 101-116. BAR International Series 528. British Archaeological Reports, Oxford.
Beck, Charlotte, Amanda K. Taylor, George T. Jones, Cynthia M. Fadem, Caitlyn R. Cook, and
Sara A. Millward
2002 Rocks are heavy: transport costs and Paleoarchaic quarry behavior in the Great Basin.
Journal of Anthropological Archaeology 21:481–507.
Binford, Lewis
1973 Interassemblage Variability: The Mousterian and the “Functional” Argument. In The
Explanations of Culture Change: Models in Prehistory, edited by Colin Renfrew, pp.
227-253. Duckworth, London
1979 Organization and Formation Processes: Looking at Curated Technologies. Journal of
Anthropological Research 35:255-273.
1980 Willow Smoke and Dogs' Tails: Hunter-Gatherer Settlement Systems and
Archaeological Site Formation. American Antiquity 45(1):4-20.
Binford, Sally R., and Lewis R. Binford
1969 Stone Tools and Human Behavior. Scientific American 220(4):70-87.
Bishop, C.A.
98

1987 Coast-interior exchange: the origins of stratification in northwestern North America.
Arctic Anthropology 24(1): 72-83.
Blair, Susan E.
2010 Missing the boat in lithic procurement: Watercraft and the bulk procurement of toolstone on the Maritime Peninsula. Journal of Anthropological Archaeology 29:33–46.
Borden, Charles E.
1952 Results of Archaeological Investigations in Central British Columbia. Anthropology in
British Columbia Vol. 3:31-43, Victoria.
Bouchard, Blake Emile
2012 The Resilience of the Babine: The Economic and Social Relations of the Babine to
1830. Master’s thesis, Department of History, University of Northern British Columbia,
Prince George.
Bradbury, Andrew P.
1998 The Examination of Lithic Artifacts From an Early Archaic Assemblage:
Strengthening Inferences Through Multiple Lines of Evidence. Midcontinental Journal of
Archaeology 23(2):263-288.
Bradbury, Andrew P., and Philip J. Carr
1995 Flake Typologies and Alternative Approaches: An Experimental Assessment. Lithic
Technology 20(2):100-115.
1999 Examining Stage and Continuum Models of Flake Debris Analysis: An Experimental
Approach. Journal of Archaeological Science 26:105-116.
2004 Combining Aggregate and Individual Methods of Flake Dabris Analysis: Aggregate
Trend Analysis. North American Archaeologist 25(1):65-90.
Budhwa, Rick
2005 An Alternate Model for First Nations Involvement in Resource Management
Archaeology. Canadian Journal of Archaeology 29:20-45.
Burley, David V.
99

1975 Skeena III Preliminary Report: The Archaeology of Morice Lake. Report Submitted to
the Archaeological Sites Advisory Board of British Columbia, Victoria.
Burroni, Daniela, Randolph E. Donahue and A. Mark Pollard
2002 The Surface Alteration Features of Flint Artefacts as a Record of Environmental
Processes. Journal of Archaeological Science 29:1277–1287.
Campbell, John M.
1968 The Kavik Site of Anaktuvuk Pass, Arctic Alaska. Anthropological Papers of the
University of Alaska 14:32-42.
Carr, Philip J., and Andrew P. Bradbury
2001 Flake Debris Analysis, Levels of Production, and the Organization of Technology. In
Lithic Debitage: Context, Form, Meaning, edited by W. Andrefsky, pp. 126-146.
University of Utah Press, Salt Lake City.
2004 Exploring Mass Analysis, Screens, and Attributes. In Aggregate Analysis in Chipped
Stone, edited by Christopher T. Hall and Mary Lou Larson, pp. 21-44. University of Utah
Press, Salt Lake City.
Carlson, Arne and Jean Bussey
1990 Archaeological Inventory and Impact Assessment, Bulkley River, Lower Hagwilget
Canyon. Permit 1989-102. MS on file, BC Archaeology Branch, Victoria.
Chisholm, Brian
1986 Reconstruction of Prehistoric Diet in British Columbia using Stable-Carbon Isotopic
Analysis. Master’s thesis, Department of Archaeology, Simon Fraser University,
Burnaby.
Cotterell, B., and J. Kamminga
1987 The Formation of Flakes. American Antiquity 52:675-708.
Cranny, Michael W.
1986 Carrier Settlement and Subsistence in the Chinlac/Cluculz Lake Area of Central British
Columbia. Master’s Thesis, Department of Anthropology and Sociology, University of
British Columbia, Vancouver.

100

Demarchi, Dennis A.
2011 An Introduction to the Ecoregions of British Columbia. Ministry of Environment:
Ecosystem Information Section, Victoria, British Columbia.
Dillian, Carolyn D.
2014 The Effects of Weathering on the Geochemical Characterization of Southeastern U.S.
Rhyolites. North American Archaeologist 35(2):149-166.
Donahue, Paul
1973 Ulkatcho: An archaeological outline. Syesis 6: 153-178.
1975 Concerning Athapaskan prehistory in British Columbia. The Western Canadian
Journal of Anthropology V (3-4):21-63.
1977 4500 Years of Cultural Continuity on the Central Interior Plateau of British Columbia.
PhD dissertation, Department of Anthropology, University of Wisconsin, Madison.
Eren, Metin I., and Mary E. Prendergast
2008 Comparing and Synthesizing Retouch Indices. In Lithic Technology: Measures of
Production, Use, and Curation, edited by William Andrefsky Jr., pp.49-85. Cambridge
University Press, New York.
Farley, A.L.
1979 Atlas of British Columbia: People, Environment and Resource Use. University of
British Columbia Press, Vancouver.
Fiske, Jo-Anne, and Betty Patrick
2000 Cis Dideen Kat: When the Plumes Rise. University of British Columbia Press,
Vancouver.
Fladmark, Knut
1976 Punchaw Village: A preliminary report, archaeology of a pre-Contact settlement. In
Current Research Reports, edited by Roy L. Carlson, pp. 19-32. Department of
Archaeology, Simon Fraser University, Burnaby.
2009 People of the Rivers: An Introduction to the Pre-Contact Cultures of Subarctic British
Columbia. In Painting the Past with a Broad Brush: Papers in Honour of James Valliere
101

Wright, edited by David L. Keenlyside and Jean-Luc Pilon, pp. 555-616. Canadian
Museum of Civilization, Gatineau.
Flenniken, J.J.
1981 Replicative Systems Analysis: A Model Applied to the Vein Quartz Artifacts from the
Hoko River Site. Pullman, Washington.
Furniss, E.
1993 Dakhel Keyoh: The Southern Carrier in Earlier Times. Quesnel School District & the
Kluskus, Nazko, Red Bluff and Ulkatcho Indian Bands, Quesnel.
Gessner, Suzanne C.
2003 The Prosodic System of the Dakelh (Carrier) Language. PhD dissertation, Department
of Linguistics, University of British Columbia, Vancouver.
Google Earth Pro, ©2023 Airbus and ©2023 CNES/Airbus
2022a Image showing Smokehouse Island in relation to Nilkitkwa Lake (to the north) and
Babine Lake (to the south), British Columbia, 55°19'19.04"N 126°37'28.15"W, elevation
737m. Google Earth Pro App, Map data. https://earth.google.com/web/@55.32163453,126.62447947,1126.13288967a,22183.78098338d,35y,0.73946359h,5.61364454t,359.9988r?utm_source=earth7&utm_campaign=vine&hl=en,
accessed July 6, 2023.
2022b Image showing Kitsuns Creek Formation (purple), Red Rose Formation (light blue),
Paleocene to Eocene (red), Lower to Upper Cretaceous Rocky Ridge Formation (yellow),
and Middle to Upper Jurassic Bowser Lake Ashman Formation (light pink) deposits,
British Columbia, 55°18’23.80” N 126°37’34.47” W, elevation 712 m. Google Earth
Pro, Map data. https://earth.google.com/web/@55.30553391,126.6262136,1344.63210207a,6088.2157773d,35y,0.87692747h,10.85399853t,0.005r?utm_source=earth7&utm_campaign=vine&hl=en,
accessed June 30, 2023.
Google Earth Pro, SIO, NOAA, U.S. Navy, NGA, GEBCO, Landsat/Copernicus, IBCAO,
LDEO-Columbia, NSF, and NOAA
2015 Image showing Smokehouse Island in relation to some key locations in British
Columbia., 54°05'06.35"N 126°40'09.85"W, elevation 1,041 m. Google Earth Pro App,
Map data. https://earth.google.com/web/@54.08557687,126.66938947,1058.0384305a,921739.8368302d,35y,0.84639271h,3.576936t,0r?utm_so
urce=earth7&utm_campaign=vine&hl=en, accessed July 6, 2023
102

Google Earth Pro, ©2023 Maxar Technologies, ©2023 CNES/Airbus, and ©2023 Province of
British Columbia
2023 Image showing Smokehouse Island in relation to the village site Nass Glee (to the
north) and Fort Kilmaurs (to the south), 55°13'14.38"N 126°29'52.92"W, elevation 956m.
Google Earth Pro App, Map data. https://earth.google.com/web/@55.2243281,126.49808002,-5483.55208059a,79297.27667581d,35y,-0.42186633h,3.58022264t,0r?utm_source=earth7&utm_campaign=vine&hl=en, accessed July 6, 2023.
Gottesfeld, Leslie M. J.
1994 Conservation, Territory, and Traditional Beliefs: An Analysis of Gitksan and
Wet'suwet'en Subsistence, Northwest British Columbia, Canada. Human Ecology
22(4):443-465.
Gould, Richard, Dorothy A. Koster, and Ann H.L. Sontz
1971 The Lithic Assemblage of the Western Desert Aborigines of Australia. American
Antiquity 36:149‐169.
Hackett, Corey
2017 Smoke on the Water: Uncovering a Socially Complex Pre-Contact Babine Fishing
Village at Nass Glee (GiSq-4). Unpublished Master’s Thesis, Department of
Interdisciplinary Studies, University of Northern British Columbia, Prince George.
Hackler, James
1958 Factors Leading to Social Disorganization Among the Carrier Indians at Lake Babine.
Master’s thesis, Department of Social Sciences, San Jose State College, San Jose.
Hall, David R.
1998 Tsini Tsini: A Technological Analysis of a Biface Production Centre in the Talchako
River Valley, B.C. Master’s thesis, Department of Archaeology, Simon Fraser
University, Burnaby.
Hallson, Jennifer
2017 Using debitage analysis to investigate an Alberta archaeological site. COMPASS
1(1):25-41.
Ham, Leonard
1988 A heritage resource overview of Teashin Resources Ltd’s Dome Mountain Project,
Smithers, BC. Permit 1988-49. Report on file with the Archaeology Branch Library,
Victoria, BC.
103

Hargus, Sharon
2007 Witsuwit’en Grammar: Phonetics, Grammar, Morphology. UBC Press, Vancouver.
Harmon, Daniel W.
1957 Harmon’s Journal: 1800-1819. New Caledonia House Publishing, Victoria.
Harris, Douglas C.
2001 Fish Weirs and Legal Cultures on Babine Lake, 1904-1907. In Fish, Law, and
Colonialism, edited by Douglas C. Harris, pp. 79-126. University of Toronto Press,
Toronto.
Hayden, Brian
1977 Stone Tool Functions in the Western Desert. In Stone Tools as Cultural Markers:
Change, Evolution, and Complexity, edited by R.V.S. Wright, pp. 178‐188. Australian
Institute of Aboriginal Studies, Canberra.
1980 Confusion in the Bipolar World: Bashed Pebbles and Splintered Pieces. Lithic
Technology 9(1):2-7.
Hayden, Brian, and W. Karl Hutchings
1989 Whither the Billet Flake? In Experiments in Lithic Technology, edited by Daniel S.
Amick and Raymond P. Mauldin, pp. 235-257. BAR International Series No. 528. British
Archaeological Reports, Oxford.
Henry, D.O, C.V. Haynes, and B. Bradley
1976 Quantitative Variations in Flaked Stone Debitage. Plains Anthropologist 21:57-61.
Hiscock, P.
1985 The need for a taphonomic perspective in stone artifact analysis. Queensland
Archaeological Research 2:82-95.
Hudson, D.
1983 Traplines and Timber: Social and Economic Change among the Carrier Indians of
Northern British Columbia. PhD dissertation, Department of Anthropology, University of
Alberta, Edmonton.
104

Ingbar, Eric E., Mary L. Larson, and Bruce A. Bradley
1989 A Nontypological Approach to Debitage Analysis. In Experiments in Lithic
Technology, edited by Daniel S. Amick and Raymond P. Mauldin, pp. 117-136. BAR
International Series No. 528. British Archaeological Reports, Oxford.
Inglis, Richard I., and George F. MacDonald
1979 Introduction. In Skeena River Prehistory, edited by Richard Inglis and George
MacDonald, pp. 1-17. Archaeological Survey of Canada Paper No. 87. National Museum
of Man, Ottawa.
Jenness, Diamond
1944 The Carrier Indians of the Bulkley River: Their Social and Religious Life. Bureau of
American Ethnology Bulletin 133 (25):469-586.
Kantakis, Michael Adam
2017 Babine Wood-stake Fish Weirs in an Eleven Kilometer Stretch of the Babine River and
Nilkitkwa Lake, North Central British Columbia. Master’s Thesis, Department of
Interdisciplinary Studies, University of Northern British Columbia, Prince George.
Keddie, Grant
2006 Kwah's Dagger. Living Landscapes Project: Upper Fraser Basin, Royal British
Columbia Museum Website Project, http://staff.royalbcmuseum.bc.ca/wpcontent/uploads/2013/08/Kwahs-Dagger-Grant-Keddie.pdf, accessed July 11, 2023.
Kimble, Kathy
1978 Report of the Skeena Regional archaeological impact survey. Permit 1978-7. Report
on file with the Archaeology Branch Library, Victoria, BC.
Kuijt, Ian, William C. Prentiss and David L. Pokotylo
1995 Bipolar Reduction: An Experimental Study of Debitage Variability. Lithic Technology
20(2):116-127.
Larson, Mary Lou
2004 Chipped Stone Aggregate Analysis in Archaeology. In Aggregate Analysis in Chipped
Stone, edited by Christopher T. Hall and Mary Lou Larson, pp. 3-17. University of Utah
Press, Salt Lake City.
Larson, M., and M. Kornfeld
105

1997 Chipped Stone Nodules: Theory, Method, and Examples. Lithic Technology 22:4-18.
Levson, Victor M.
2001 Quaternary geology of the Babine porphyry copper district; implications for
geochemical exploration. Canadian Journal of Earth Sciences 38(4):733-749.
Levson, Victor M., and A.J. Stumpf
1998 Glacial controls on geochemical transport distance and direction in north-central
Stikinia: implications for exploration; in Cordillera revisited: recent developments. In
Cordilleran geology, tectonics and mineral deposits, edited by P. Mustard and S. Gareau,
pp. 68-75. Geology Association of Canada, St. John’s.
Lundblad, S.P., P.R. Mills, and K. Hon
2008 Analysing Archaeological Basalt Using Non-destructive Energy-Dispersive X-Ray
Flourescence (EDXRF): Effects of Post-Depositional Chemical Weathering and Sample
Size on Analytical Precision. Archaeometry 50:1-11. DOI: 10.1111/j.14754754.2007.00345.x.
MacDonald, B. L., Stalla, D., He, X., Rahemtulla, F., Emerson, D., Dube, P. A., Maschmann,
M.R., Klesner, C.E., and White, T. A.
2019 Hunter-gatherers harvested and heated microbial biogenic iron oxides to produce rock
art pigment. Scientific reports 9(1):1-13.
MacDonald, G.
1967 Archaeological Reconnaissance in the Tsimshian Area, British Columbia. Paper
presented to the Society for American Archaeology, 23rd Annual Meeting, Ann Arbor,
Michigan.
1969 Preliminary culture sequence from the Coast Tsimshian area, British Columbia.
Northwest Anthropological Research Notes 3:240-254.
MacIntyre, D.G., M.E. Villeneuve, and P. Schiarizza
2001b Timing and tectonic setting of Stikine Terrane magmatism, Babine–Takla lakes area,
central British Columbia. Canadian Journal of Earth Sciences 38:579-601.
MacIntyre, D., I. Webster, K. Bellefontaine, P. Desjardins, R. Metcalf, S. Munzar, D.
Tackaberry, A. Justason, P. Wojdak, N.C. Carter, T.A. Richards, and H.W. Tipper
106

2001a Geological Compilation Map Babine Porphy Copper District Central British
Columbia. British Columbia Geology Survey Open File 2001-3. Electronic document,
https://cmscontent.nrs.gov.bc.ca/geoscience/PublicationCatalogue/OpenFile/BCGS_OF2
001-03.pdf, accessed June 30, 2023.
Magne, Martin P.R.
1985 Lithics and Livelihood: Stone Tool Technologies of Central and Southern Interior
British Columbia. Mercury Series Archaeological Survey of Canada Paper No. 133.
National Museum of Man, Ottawa.
1989 Lithic Reduction Stages and Assemblage Formation Processes. In Experiments in
Lithic Technology, edited by Daniel S. Amick and Raymond P. Mauldin, pp. 15-31. BAR
International Series 528. British Archaeological Reports, Oxford.
2001 Debitage Analysis as a Scientific Tool for Archaeological Knowledge. In Lithic
Debitage: Context, Form, Meaning, edited by William Andrefsky Jr., pp. 21-30.
University of Utah Press, Salt Lake City.
Magne, Martin P.R., and Daryl Fedje
2007 The Spread of Microblade Technology in Northwestern North America. In Origin and
Spread of Microblade Technology in Northern Asia and North America, edited by
Yaroslav V. Kuzmin, Susan G. Keates, and Chen Shen, pp. 171–188. Archaeology Press,
Simon Fraser University, Burnaby.
Magne, Martin P.R., and R.G. Matson
1982 Identification of ‘Salish’ and ‘Athapaskan’ Side-Notched Projectile Points from the
Interior Plateau of British Columbia. In Approaches to Algonquian Archaeology, edited
by M.G. Hanna and B. Kooyman, pp.57-80. Archaeology Association, University of
Calgary, Calgary.
2008 Projectile Points of Central and Northern Interior British Columbia. In Projectile Point
Sequences in Northwestern North America, edited by Roy L. Carlson and Martin P.R.
Magne, pp.273-292. Simon Fraser University Archaeology Press, Burnaby.
2010 Moving On: Expanding Perspectives on Athapaskan Migration. Canadian Journal of
Archaeology 34:212-239.
Magne, Martin P.R., and D.L. Pokotylo
107

1981 A pilot study in bifacial lithic reduction sequences. Lithic Technology 10(2-3):34-47.
Manclossi, Francesca, and Steven A. Rosen
2019 The Importance of Being Ad Hoc: Patterns and Implications of Expedient Lithic
Production in the Bronze Age in Israel. In Lithic Technologies in Sedentary Societies,
edited by Rachel A Horowitz and Grant S. McCall, pp. 69-88. University Press of
Colorado, Louisville.
Marsden, Susan, and Robert Galois
1995 The Tsimshian, the Hudson's Bay Company, and the Geopolitics of the Northwest
Coast Fur Trade, 1787-1840. Canadian Geographer 39: 169-83.
Matthews, William H.
1979 Late Quaternary Environmental History Affecting Human Habitation of the Pacific
Northwest. Canadian Journal of Archaeology 3:145-156.
Matson, R.G., and Martin P.R. Magne
2007 Athapaskan Migrations: The Archaeology of Eagle Lake, British Columbia. University
of Arizona Press, Tucson.
Mauldin, Raymond P., and Daniel S. Amick
1989 Investigating Patterning in Debitage from Experimental Bifacial Core Reduction. In
Experiments in Lithic Technology, edited by Daniel S. Amick and Raymond P. Mauldin,
pp. 67-88. BAR International Series No. 528. British Archaeological Reports, Oxford.
Metz, David
1996 A Mass Aggregate Analysis of the Debitage from the Tsini Tsini 1994 Excavations.
Bachelor of Arts Honours paper, Department of Archaeology, Simon Fraser University,
Burnaby.
Miller, Jr., Tom O.
1979. Stonework of the Xeta’ Indians of Brazil. In Lithic Use‐Wear Analysis, edited by Brian
Hayden, pp. 401‐407. Academic Press, New York.
Mohs, G.
1974 Babine Lake archaeological survey project. Report on file, Archaeology Branch,
Ministry of Forests, Lands and Natural Resource Operations, Victoria, B.C.
108

1975 Babine Lake archaeological survey project 1975. Report on file, Archaeology Branch,
Ministry of Forests, Lands and Natural Resource Operations, Victoria, B.C.
Mohs, G and A. Mohs
1976 Babine Lake archaeological survey project 1976. Report on file, Archaeology Branch,
Ministry of Forests, Lands and Natural Resource Operations, Victoria, B.C.
Montgomery, Pamela
1978 Stone Artifacts from the Punchaw Lake Site (Area C): a Late Prehistoric Occupation in
Central British Columbia. Master’s thesis, Department of Archaeology, Simon Fraser
University, Burnaby.
Morice, Adrien G.
1893 Notes archaeological, industrial, and sociological on the Western Denes. Transactions
of the Canadian Institute 4:1-221.
1905 The History of the Northern Interior of British Columbia: (Formerly New Caledonia)
1660 to 1880. William Briggs, Toronto.
Morin, Jesse, Thomas C. A. Royle, Hua Zhang, Camilla Speller, Miguel Alcaide,
Ryan Morin, Morgan Ritchie, Aubrey Cannon, Michael George, Michelle George and
Dongya Yang
2021 Indigenous sex‑selective salmon harvesting demonstrates pre‑contact marine resource
management in Burrard Inlet, British Columbia, Canada. Scientific Reports 11(21160):113. https://doi.org/10.1038/s41598-021-00154-4
Morrison, Donna M.
1994 Validity in Debitage Analysis: An Experimental Assessment of Comparing Quartzite
to Obsidian. Master’s thesis, Department of Archaeology, Simon Fraser University,
Burnaby.
Nelson, JoAnne, Tony Barresi, Ellie Knight and Nicole Boudreau
2005 Geology and Mineral Potential of the Usk Map Area (NTS 103I/09), Terrace, British
Columbia. Geological Fieldwork 2006-1:117-134.
Ogburn, Dennis, Bill Sillar, and Julio César Sierra

109

2013 Evaluating effects of chemical weathering and surface contamination on the in situ
provenance analysis of building stones in the Cuzco region of Peru with portable XRF.
Journal of Archaeological Science 40:1823-1837.
Patterson, Leland W.
1995 Thermal Damage of Chert. Lithic Technology 20(1):72-80.
Pokotylo, David L.
1978 Lithic Technology and Settlement Patterns in Upper Hat Creek Valley, B.C. PhD
dissertation, Department of Anthropology and Sociology, University of British Columbia,
Vancouver.
Pokotylo, David L., and Donald Mitchell
1998 Prehistory of the Northern Plateau. In Plateau, edited by Deward E. Walker, Jr., pp.81102. Handbook of North American Indians, Vol. 12, William C. Sturtevant, general
editor. Smithsonian Institution, Washington, D.C.
Pough, Frederick H., and Jeffrey Scovil
1996 A field guide to rocks and minerals. 5th edition. Houghton Mifflin Harcourt Publishing
Company, Boston.
Prentiss, William C.
1993 Hunter-Gatherer Economics and the Formation of a Housepit Floor Lithic Assemblage.
PhD dissertation, Department of Archaeology, Simon Fraser University, Burnaby.
1998 The Reliability and Validity of a Lithic Debitage Typology: Implications for
Archaeological Interpretation. American Antiquity 63(4):635-650.
2001 Reliability and Validity of a “Distinctive Assemblage” Typology: Integrating Flake
Size and Completeness. In Lithic Debitage: Context, Form, Meaning, edited by William
Andrefsky Jr., pp. 147-172. University of Utah Press, Salt Lake City.
Prentiss, William C., and Eugene J. Romanski
1989 Experimental Evaluation of Sullivan and Rozen’s Debitage Typology. In Experiments
in Lithic Technology, edited by Daniel S. Amick and Raymond P. Mauldin, pp. 89-100.
BAR International Series No. 528. British Archaeological Reports, Oxford.

110

Prince, Paul
2005 Fish Weirs, Salmon Productivity, and Village Settlement in an Upper Skeena River
Tributary, British Columbia, Canadian Journal of Archaeology, 29:68-87.
2014 Fish weirs and an interior salmon fishery on the Nautley River central British
Columbia. North American Archaeologist 35(2):119-148.
Rafferty, Pauline
1975 Archaeological reconnaissance of the Bulkley-Nechako. Permit 1975-4. Report on file
with the Archaeology Branch Library, Victoria, BC.
Rasic, Jeffrey T.
2004 Debitage Taphonomy. In Aggregate Analysis in Chipped Stone, edited by Christopher
T. Hall and Mary Lou Larson, pp.112-135. University of Utah Press, Salt Lake City.
Rahemtulla, Farid
1995a An Analysis of Lithic Debitage from the Early Period at Namu. Master’s thesis,
Department of Archaeology, Simon Fraser University, Burnaby.
1995b Yet another reconsideration of the Northwest Coast Pebble Tool Tradition. Paper
presented at the Northwest Anthropology Conference, Portland, Oregon.
2006 Design of Stone Tool Technology during the Early Period (ca. 10,000–5,000 B.P.) at
Namu, Central Coast of British Columbia. PhD dissertation, Department of Archaeology,
Simon Fraser University, Burnaby.
2012 Archaeological Research Investigations at Site GiSq-004 located at Nilkitkwa Lake in
the North-central interior of BC. Permit report 2010-254, report on file at the
Archaeology Branch, Ministry of Forests, Lands and Natural Resource Operations,
Victoria, BC.
2013 Archaeological Research Investigations at archaeological site GiSq-004, located at the
northern part of Nilkitkwa Lake/Babine River in the north-central interior of BC. Permit
Report 2012-204, report on file at the Archaeology Branch, Ministry of Forests, Lands
and Natural Resource Operations, Victoria, BC.

111

2019 The Babine Archaeology Project: Discovery of a Rare Wet Site on the Babine River,
North-Central British Columbia. In Waterlogged: Emerging Trends in Northwest Coast
Archaeology, edited by Kathryn Bernick, pp. 159-168. Washington State University
Press, Pullman.
In prep Archaeological Research Investigations at Smokehouse Island, GiSp-001. Report
prepared for the Lake Babine Nation.
Reimer, Rudy
2011 The mountains and rocks are forever: lithics and landscapes of Skwxwú7mesh
Uxwumixw Territory. PhD dissertation, Department of Anthropology, McMaster
University, Hamilton.
2015 Reassessing the role of Mount Edziza obsidian in northwestern North America.
Journal of Archaeological Science: Reports 2:418-426.
2018 Lithic Sourcing in Canada. Canadian Journal of Archaeology 42:137-143.

2019 Simon Fraser University Department of Archaeology X-Ray Fluorescence Report
2019-16. Report on File at Simon Fraser University, Department of Archaeology,
Burnaby.
2023 X-ray Fluorescence at Simon Fraser University Department of Archaeology. Journal
of Northwest Anthropology Special Publication No. 7.
Richards, Thomas H. and Michael K. Rousseau
1987 Late Pre-Contact Cultural Horizons on the Canadian Plateau. Department of
Archaeology, Simon Fraser University, Vancouver.
Ritchie, Morgan, Jesse Morin, Jerram Ritchie, Nicholas Waber, Michael Blake and Rhy
McMillan
2022 Embedded Household Specialization: The Bifaces from YāçkEtEl on the Northwest
Coast. Journal of Field Archaeology:1-19. DOI: 10.1080/00934690.2022.2087017
Rowland, Michael J., and Sean Ulm.
112

2011 Indigenous fish traps and Weirs of Queensland. Queensland Archaeological Review.
14(1) 1-58.
Shott, Michael J.
1994 Size and Form in the Analysis of Flake Debris: Review and Recent Approaches.
Journal of Archaeological Method and Theory 1(1):69-110.
2004 Aggregate Methods and the Future of Debitage Analysis. In Aggregate Analysis in
Chipped Stone, edited by Christopher T. Hall and Mary Lou Larson, pp. 211-228. The
University of Utah Press, Salt Lake City.
2017 Stage and continuum approaches in prehistoric biface production: A North American
perspective. PLoS ONE 12(3):e0170947.
DOI:https://doi.org/10.1371/journal.pone.0170947
Shott, Michael J., and Desale Habtzghi
2019 The mixture problem in flake analysis: Allocating flake samples to segments of
reduction using CLSR methods. Journal of Archaeological Science 103:46-56.
Simonsen, B.
1989 Report on the Results of Heritage Impact Assessment Studies Related to 12 Highway
Construction Projects Along Highway #16 and #37: Northwestern B.C. The Bastion
Group, Permit Report 1989-49. Report on file at the Archaeology and Outdoor
Recreation Branch, Ministry of Municipal Affairs, Recreation and Culture, Victoria.
Stahle, D.W., and J.E. Dunn
1982 An Analysis and Application of Size Distribution of Waste Flakes from the
Manufacture of Bifacial Tools. World Archaeology 14(1):84-97.
Steffen, A., E.J. Skinner, and P.W. Ainsworth
1998 A View to the Core: Technological Units and Debitage Analysis. In Unit Issues in
Archaeology: Measuring Time, Space, and Material, edited by A.F. Ramenofsky and A.
Steffen, pp. 131-146. University of Utah Press, Salt Lake City.
Stemp, W. James, Elizabeth Graham, Christophe Helmke, and Jaime J. Awe.
2021 Expedient lithic technology in complex sedentary societies: Use-wear, flake size, and
edge angle on debitage from two ancient Maya sites. Journal of Anthropological
Archaeology 61: 101243.
113

Stumpf, A.J., B. E. Broster and V.M. Levson
2000 Multiphase flow of the Late Wisconsinan Cordilleran ice sheet in Western Canada.
GSA Bulletin 112(12):1850-1863.
2004 Glacial stratigraphy of the Bulkley River Region: a depositional framework for the
Late Pleistocene in central British Columbia. Géographie physique et Quaternaire 58(23):217-228.
Sullivan, A. P., Ill
1987 Probing the Sources of Variability: A Regional Case Study Near the Homolovi Ruins,
Arizona. North American Archaeologist 8:41-71.
Sullivan III, Alan P. and Kenneth C. Rozen.
1985 Debitage Analysis and Archaeological Interpretation. American Antiquity 50(4):755779.
Tobey, M.L.
1981 Carrier. In Handbook of North American Indians: Subartic (Volume 6), edited by J.
Helm, pp. 413-432. Smithsonian Institution, Washington, DC.
Whittaker, John C., and Eric J. Kaldahl
2001 Where the Waste Went: A Knapper’s Dump at Grasshopper Pueblo. In Lithic
Debitage: Context, Form, Meaning, edited by William Andrefsky Jr., pp.32-60.
University of Utah Press, Salt Lake City.
Wilmeth, Roscoe
1978 Canadian Archaeological Radiocarbon Dates. Revised ed. Mercury Series
Archaeology Paper 77, National Museum of Man, Ottawa.

114