Journal of Archaeological Science 34 (2007) 392e402
http://www.elsevier.com/locate/jas
The application and misapplication of mass analysis in lithic
debitage studies
William Andrefsky Jr.
Department of Anthropology, Washington State University, Pullman, WA 99164-4910, United States
Received 4 March 2006; received in revised form 25 May 2006; accepted 26 May 2006
Abstract
The technique of debitage mass analysis based upon size grades of debitage populations is shown to be prone to errors when making
interpretations about the kind of tool produced or the kind of lithic reduction technology used. Significant sources of error may originate
from differences in individual flintknapping styles and techniques, raw material size and shape variants, and mixing of debitage from more
than one reduction episode. These sources of error render debitage Mass Analysis ineffective for determining the kind of stone tool reduction
activities practiced at excavated sites. Mass Analysis may be effective for determining artifact reduction sequences if it is used on debitage from
a single reduction episode or part of a reduction episode. However, it is shown that Mass Analysis when used for assessing reduction sequence
information, must also control for the effects of raw material variability, assemblage mixing, and flintknapping styles.
Ó 2006 Elsevier Ltd. All rights reserved.
Keywords: Lithic analysis; Experimental archaeology; Mass Analysis; Debitage; Size grades
Lithic artifact debitage may be defined as the by-product of
stone tool production and resharpening. Debitage includes the
chips and debris that have been removed to shape and maintain
stone tools. In 1972, Don Crabtree proclaimed that debitage
composed the ‘‘finger prints’’ of stone tool production [33].
He rightly meant that debitage could be used to recognize
stone tool production activities even in the absence of the
stone tools themselves being recovered. Since then debitage
has been effectively analyzed to determine the types of artifact
produced [15,23,66,67,74,95], the kind of technology practiced
[5,30,35,36,54,58,72], the stage of tool reduction or production
[8,26,50,61,65,76,85], and even the process of site formation
[31,39,49,69,87,88].
The analysis of lithic debitage has become increasingly
important in understanding the activities and tasks that have
taken place on the prehistoric landscape [10,16,57,59,73,92,93,94].
This has helped researchers understand how human mobility
and land-use are linked to foraging strategies, and also how these
foraging strategies are linked to lithic tool production, use,
E-mail address: [email protected]
0305-4403/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jas.2006.05.012
maintenance, and discard [14,20,21,32,34,38,47,56,63,91].
A form of debitage analysis known as Mass Analysis
(MA) has been widely adopted by archaeologists to process
large quantities of artifacts to make such inferences
[1,17,18,29,51,62,77,81]. Part of the reason MA has been so
appealing to researchers relates to its relative ease, speed, and
reliability. The advantages of MA over individual specimen
analysis are highlighted by Ahler [1]. He notes that MA:
(1) can be applied to the full range of debitage without regarding fracture or completeness, thereby eliminating potential
bias resulting from exclusion of some debitage forms, such
as broken or shattered pieces;
(2) can be rapid and efficient, even for extremely large debitage samples, because it does not require handling and
measuring of individual specimens;
(3) can reduce technological bias based upon debitage size because different mesh sizes capture a range of different
specimen sizes; and
(4) can be a highly objective technique since the analysis involves size grading, counting, and weighing, and can be
W. Andrefsky Jr. / Journal of Archaeological Science 34 (2007) 392e402
There are indeed appealing reasons for archaeologists to
adopt MA. Briefly saying, MA is a form of debitage population study that assimilates all debitage within a recognized
population and segregates it into size groups known as ‘‘size
grades.’’ Based on the relative proportion of debitage within
each size grade generalizations are made about the technology
used to produce the debitage population. The relative proportion of size grades is established by the count and weight of
specimens in each size grade. The average weight of each
size grade and the percentage of specimens with cortex in
each size grade may also be calculated. These proportions
and average values become the attributes of the debitage population used in MA techniques.
Inferences about the debitage population are generated by
a ‘‘control group’’ of debitage based upon experimental replication of various tools or stages of tool production and core
reduction. For instance, the investigator might first replicate
a projectile point from a cobble or a flake blank. The debitage
from that replication event is sorted into size grades and relative amounts of each size grade are calculated based on counts
and weights and/or cortical representation. This control group
is summarized to produce a signature of some type such as
a histogram, ratio measure, or a discriminant function. This
control group signature is then compared to the signature obtained from the excavated collection using the same size
grades. If there is a match the investigator may infer that a
projectile point was manufactured at that location, even if
one was not found there.
A series of replications are then conducted to evaluate all
sizes of debitage resulting from the production of different
kinds of stone tools and stages of stone tool reduction. These
experimentally derived debitage signatures are then used as
a reference library for various kinds of tool production or
core reduction activities. For example, a specific debitage signature has been produced for hard hammer reduction of cobbles, for bipolar reduction of pebbles, for projectile point
manufacture, and even for the initial thinning of bifaces
from flake blanks [1,18,68].
The size grades used in MA are easily applied to an assemblage of debitage. One of the most popular ways to stratify
debitage into size grades is to sift the debitage through a series
of nested screens. The investigator can save a considerable
amount of time using this technique because she or he is not
required to separate the debitage into whole or broken flakes
or shatter or platform remnant flakes. All debitage is shifted
through the screens regardless of technological variety or completeness to arrive at size groups or size grades. A fairly
untrained technician can do this sorting. In fact, some laboratories stack the nested screens on a mechanical shaker
machine so that the sorting or size grading of debitage requires
no training at all.
A classic MA study might produce a distribution of debitage size grades like the one shown in Table 1. In this case each
of the reduction techniques would have eight variables based
Table 1
Relative proportions of size grades for core reduction techniques by weight
and count
Reduction
technique
Percent weight
Percent count
Size 1 Size 2 Size 3 Size 4 Size 1 Size 2 Size 3 Size 4
Hard hammer 29.0
flakes
Hard hammer 17.5
blades
Bipolar chips 12.5
43.5
21.8
5.7
0.7
5.5
22.5
71.3
53.0
24.2
5.3
0.5
7.1
24.1
68.3
34.8
37.6
15.1
0.2
2.1
18.5
79.2
upon the proportion of each size grade calculated by weight
and then by count. The percentage of cortical specimens for
each size grade would add an additional four variables that
can also be used in a MA study. These data might then be analyzed to produce a histogram signature that looks like the one
in Figs. 1 and 2 showing the relative proportions of count and
weight, respectively in each size grade for each reduction technique. In addition, a discriminant function can be calculated
for each experimentally replicated assemblage and that signature can be compared to the signature obtained for the excavated assemblage to make an interpretation [1,41].
In this paper, I contend that MA, although it is relatively
easy to apply and can be used to process very large quantities
of debitage, has been adopted by many practitioners without
stringent analytical constraints and with little concern for the
validity of the results. Instead, I suggest that this popular analytical technique has been adopted because of its economic
benefits (ease and speed) and has unfortunately or unintentionally often resulted in spurious interpretations of the archaeological record. Below, I review several factors responsible
for the failure of MA, including problems with replicated assemblages, raw material influences, and mixed archaeological
assemblages. I also identify effective applications of MA and
90
80
Relative % of Debitage by Count
conducted by virtually anyone trained in elementary lab
procedures.
393
70
SIZE 1 CT
SIZE 2 CT
SIZE 3 CT
SIZE 4 CT
60
50
40
30
20
10
0
HH FLAKES
HH BLADES
BIPOLAR
Reduction Technique
Fig. 1. Histogram of relative proportions of count for Mass Analysis of three
reduction types (hard hammer flakes, hard hammer blades, and bipolar debitage). Data are taken from Ahler [1].
W. Andrefsky Jr. / Journal of Archaeological Science 34 (2007) 392e402
394
60
SIZE 1 WT
SIZE 2 WT
SIZE 3 WT
SIZE 4 WT
Relative % by Debitage Weight
50
40
30
20
10
0
HH FLAKES
HH BLADES
BIPOLAR
Reduction Techniques
Fig. 2. Histogram of relative proportions of weight for Mass Analysis of three
reduction types (hard hammer flakes, hard hammer blades, and bipolar debitage). Data are taken from Ahler [1].
suggest techniques that can help to make this technique a more
productive analytical strategy, such as segregating debitage
into discrete technological production episodes.
Before discussing sources of debitage variability that may
influence results of MA it is important to clarify how I use
the terms ‘‘production’’ and ‘‘reduction.’’ I use the term ‘‘production’’ when talking about the manufacturing of ‘‘tools’’ using pressure or percussion flaking methods. I use the term
‘‘reduction’’ when talking about the removal of detached
pieces from cores. Reduction refers to the process of flake removal for the acquisition of usable detached pieces. Production refers to the process of flake removal for the purpose of
making, resharpening, or reshaping a tool. I refer to core ‘‘reduction’’ and to tool ‘‘production.’’ These are not necessarily
interchangeable terms.
produced by ancient knappers, provided that the same artifacts
or the same reduction technologies are replicated.
Is this an accurate assumption? In an independent study of
flintknapping experiments, Redman [80] compared the debitage produced from three different ‘‘expert’’ flintknappers making the same type of bifaces using the same raw material and
the same kind of percussors. Although she did not evaluate
size groups or size grades, she found that there was a significant difference in six debitage characteristics among the
experimental assemblages even though all knappers were
making the same tool type. As a result of being made by different individual knappers, the experimental debitage was significantly different in flake curvature, flake length, maximum
width, width at midpoint, platform width, and flake weight.
These data suggest that individual flintknapping abilities,
styles, or techniques may produce significantly different debitage size grades. Other researchers have made similar claims
[43,70,89]. Individual variation in tool manufacturing has
obvious implications for MA, which assumes that replicated
debitage assemblages can be used to match excavated assemblages to make inferences about the kind of production that
has taken place at a site.
Individual variation in the debitage characteristics noted by
Redman [80] could influence the relative proportions of size
grades for each debitage assemblage. I tabulated size grade
data from several independently published sources and also
conducted additional replication experiments to evaluate this
possibility. Fig. 3 shows these tabulated data as relative proportions of weight for four size grades produced from bipolar
technology [1,52,68]. Morrow’s data were restructured from
his original presentation to conform with Ahler’s [1] strategy
for naming the largest size-graded debitage ‘‘size grade 1’’
and the smallest size ‘‘size grade 4.’’ Even relatively simple
technology such as bipolar reduction produces significantly
different debitage signatures when using size grade analysis.
100
SIZE 1 WT
90
SIZE 2 WT
SIZE 3 WT
80
One of the assumptions of MA is that the production of
different kinds of tools or the recognition of tool type stages
result in different amounts of debitage in various size classes.
It is also assumed that the production of the same kinds of
tools by different individuals will result in the same relative
amounts of debitage in each size grade. In other words, if
two individuals each manufacture a projectile point the debitage produced by each person should result in relatively identical distributions of debitage size grades and ultimately
obtain the same signature. A corollary assumption is that excavated debitage matching the size grade distributions of the
experimentally replicated debitage will have resulted from
the same kinds of artifact production events. Practitioners of
MA assume that the debitage signatures produced by contemporary knappers will be the same as the debitage signatures
% Debitage Weight
1. Replicator variability
SIZE 4 WT
70
60
50
40
30
20
10
0
AHLER
KALIN
MORROW
ANDREFSKY
Individual Knappers
Fig. 3. Comparison of individual flintknapping debitage by relative proportion
of weight for bipolar reduction.
W. Andrefsky Jr. / Journal of Archaeological Science 34 (2007) 392e402
For example, Kalin’s experiment produced no size grade 1
debitage while all three other knappers produced at least
some of this largest size debitage. In fact, Morrow’s experiment and my own produced most debitage in size grade 1.
The highest frequency of debitage from Ahler’s bipolar replications were in size grade 3 while Kalin’s, Morrow’s and my
size grade 3 debitage was represented as 5.2%, 13.2%, and
10.2%, respectively. Table 2 lists the significance values for
t-tests of all four size grades and demonstrates that the variability in size grade values is too great for discriminating reduction technology using MA, even though each knapper
reduced a nodule using the same reduction technology. These
data show that each replication could not have been drawn
from the same population (have the same manufacturing technique). However, we know this to be the case as all replications were the same e bipolar core reductions. This tends to
support Redman’s [80] findings that one of the greatest sources
of debitage variability comes as a result of individual knapper’s style and technique.
In a separate study, Amick and Mauldin [4] compared flake
breakage frequencies in their experiments against data on
breakage compiled by Bradbury and Carr [25]. They found
highly significant differences in flake breakage attributable
to individual knapper reduction practices for both chert core
reduction (X2 ¼ 54.35, df ¼ 3, p < 0.000001) and for chert
bifaces/tool reduction (X2 ¼ 65.25, df ¼ 3, p < 0.000001).
Again, debitage variability is shown to originate with individual knappers. As such, it is doubtful that contemporary knappers would produce the same debitage size grade signatures as
ancient knappers even if both were making exactly the same
kinds of tools. Such discrepancies make it extremely difficult
to trust results of MA given that MA results are based upon the
comparisons and matching of experimentally generated debitage assemblages to ancient debitage assemblages.
2. Raw material size, shape, and composition
Variation in debitage size grades resulting from different
flintknappers is only one potential source of error when using
MA techniques. Variability in blank size and shape as well as
raw material type will also produce different debitage sizes
[2,24,96]. However, this fact appears to be ignored by most
practitioners of MA. Seldom (if ever) are tool-stone sizes
and shapes accounted for when comparing replicated assemblages to excavated assemblages. How can we reasonably expect to obtain the same proportions of debitage size grade
signatures if the tool-stone, to make the replications and to
Table 2
Students t-test: comparison of four flintknapper’s bipolar debitage noted in
Fig. 3
Size grades
Size
Size
Size
Size
grade
grade
grade
grade
1
2
3
4
One-sample test
t
df
Sig. (2-tailed)
2.02
2.28
2.30
1.57
3
3
3
3
0.137
0.107
0.105
0.215
Mean
difference
33.85
41.90
16.55
5.28
395
make the excavated collections, is composed of different sizes
and shapes? One way to evaluate this question is to reduce different objective pieces of different sizes and shapes using the
same reduction techniques and then compare the debitage size
grades using MA.
I conducted such a test by reducing two water worn obsidian pebbles using bipolar reduction technology. Bipolar reduction technology was used because it is a fairly elementary and
uniform technique to open tool-stone. Parry and Kelly [73]
state, ‘‘Cores are not preformed or prepared in any way. Instead, they are struck almost randomly, shattering into pieces
of variable size and shape.’’ Other investigators support this
characteristic of bipolar technology [9,44,45,48]. For my replications, Pebble #1 was oval and flat on one side with a maximum linear dimension of 6.2 cm and a weight of 117 g.
Pebble #2 was more rounded or egg-shaped with a maximum
linear dimension of 8.8 cm and a weight of 339 g. Pebble #2
had approximately twice the mass as Pebble #1, but was
only about 30% (2.6 cm) longer in maximum linear dimension. The pebbles were opened in an effort to obtain as
many large usable pieces as possible. Table 3 lists the debitage
size grade frequencies for each pebble in standard hardware
cloth square size increments by percent count and percent
weight. The smaller of the two pebbles (Pebble #1), of course,
did not produce any pieces in the largest size grade (2-inch
mesh size). Pebble #2 only produced four debitage pieces in
the 2-inch size grade. However, these four pieces represented
62.5% of the total debitage weight for Pebble #2. This
trend immediately shows how different original tool-stone
package sizes can influence a size grade distribution used
in MA.
The 2-inch mesh size was eliminated from the study to give
Pebble #1 and Pebble #2 the same number of size grades with
debitage representation (bottom, Table 3). This still resulted in
very different relative percentages of debitage weight in each
of the four remaining size grades even though both pebbles
were opened using the same technology.
Table 3
Size grade proportions for count and weight from two bipolar reductions
Size grade mesh (inch)
Percent count
Percent weight
Pebble #1 with five mesh sizes
1/8
1/4
1/2
1
2
71.4
18.5
8.3
1.8
0.0
4.7
11.9
62.0
21.0
0.0
Pebble #2 with five mesh sizes
1/8
1/4
1/2
1
2
70.1
20.5
5.7
2.3
1.3
3.1
8.5
8.7
17.2
62.5
Pebble #2 with four mesh sizes
1/8
1/4
1/2
1
70.1
20.5
5.7
3.7
3.1
8.5
8.7
79.7
W. Andrefsky Jr. / Journal of Archaeological Science 34 (2007) 392e402
Other investigators studying the influence of lithic raw material on tool manufacturing process have noted similar results.
Bradbury and Franklin [27] explored the effectiveness of MA
as it relates to raw material package sizes. They conclude that
the ‘‘.raw material influences appear to be most significant
when perceived in terms of the variability in initial nodule
size and shape.’’ [27]. In a test of flake breakage and ultimately flake size, Amick and Mauldin [4] state ‘‘.breakage
patterns fail to behave consistently with any variables other
than raw material.raw material considerations must provide
the backbone of any debitage analysis.’’
Clearly, raw material composition and package size are important factors influencing debitage size grades regardless of
the kind of manufacturing technology practiced. Replication
experiments conducted to produce controlled debitage data
sets for MA must begin with the same tool-stone configurations as the excavated assemblages to make reliable control
groups.
70
SG1 WT
SG2 WT
SG3 WT
SG4 WT
60
50
% Debitage Weight
396
40
30
20
10
0
BIPOLAR
HAFTED PT
BOTH
Reduction Techniques
3. Debitage mixing
The most widely recognized problem with using MA is
known as the ‘‘mixing problem.’’ Simply stated, if we assume
that debitage produced by the production of similar tools results in similar size grade signatures, and we assume this is
true regardless of who made the tools, and we further assume
that raw material size, shape, and composition are not an issue,
MA must still overcome the problem created when debitage
from two different production events are mixed together, as
is often the case with archaeological assemblages.
For instance, when debitage from the reduction of a biface
is mixed with debitage from the reduction of a platformed
core, the size grade signatures are no longer discernable or diagnostic. This problem has been recognized over and over
again [1,2,30,68,82], but few have attempted to control for
it. Even in a controlled reduction context, we cannot use
mixed debitage to identify specific reduction activities. In
other words, even when we know we are mixing together debitage from blade and bipolar core reduction techniques it is not
possible to recognize these two reduction technologies using
MA. This problem is certainly magnified when we begin looking at excavated palimpsest assemblages where many different
kinds and combinations of tool production and resharpening
activities might have taken place over varying periods of
time. In an archaeological context, we may not have the relatively simple mixing of two sets of debitage from two production episodes. Archaeological debitage assemblages may
represent several complete or partial production events of various tool forms that may have been deposited at different times
and in varying amounts.
One way to assess the effects of debitage mixing is to obtain size grade signatures for two different reduction strategies
and then compare those signatures to signatures obtained when
the debitage from both reduction strategies are combined. I
replicated debitage from the production of a large sidenotched hafted biface and the reduction of a bipolar core.
Both were made from the same kind of obsidian. Fig. 4 shows
Fig. 4. Histogram showing size grade signatures for hafted biface production
(HAFTED PT) and bipolar core reduction (BIPOLAR), individually and when
combined.
MA histogram signatures for the production of the hafted biface and bipolar debitage, and demonstrates how those signatures are complicated when those two separate production
events are combined. We no longer obtain diagnostic signatures when differential production data are mixed. This is an
interesting comparison because the replicated hafted biface
began as a flake blank weighing 45.5 g. The final product
weighed 32.6 g, and there was only 12.9 g of debitage produced during the production process, which sorted into the
two smallest size grades. This created quite a different signature than the bipolar core reduction, which resulted in 168 g of
debitage across four size grades. Fig. 5 diagrams this data as
linear graphs and shows that bipolar reduction is potentially
present, but hafted biface technology is not recognized in
the size-graded debitage even though the replicated assemblage contains this reduction. This underscores one of the
most important findings of Bradbury and Franklin’s MA study:
‘‘.if the experimental assemblage being used to classify other
data sets lacks all of the reduction types present in the assemblage being classified, correct classification rates may drop
significantly’’ [27]. We can add to this and also say that correct
classifications decline when experimental assemblages contain more reduction types than comprising the excavated
assemblage.
4. Diagnostic signatures
There has been an assumption made in the archaeological
literature that MA does work in a controlled experimental context. This has been established many times by researchers
making replicated debitage assemblages [2,3,17,30,74]. Even
when there are problems attempting to match replicated assemblages to excavated assemblages, there appears to be
good results for obtaining diagnostic signatures for the
W. Andrefsky Jr. / Journal of Archaeological Science 34 (2007) 392e402
70
60
0.9
BIPOLAR
HAFTED PT
BOTH
0.8
0.7
% Debitage Count
50
% Debitage Weight
397
40
30
%CT S1
%CT S2
%CT S3
%CT S4
0.6
0.5
0.4
0.3
0.2
20
0.1
10
0
SG1 WT
SG2 WT
SG3 WT
BIFACE
SG4 WT
Size Grades
Fig. 5. Linear graph showing how hafted biface debitage becomes embedded
in bipolar debitage signature when mixing occurs.
replicated debitage populations. In other words, investigators
tend to produce diagnostic signatures for experimental production events using size-graded debitage produced in those
manufacturing events. Intuitively, this makes perfect sense because experimentally replicated assemblages are not mixed or
subjected to the same type of accumulation problems as excavated assemblages. Nor do replicated assemblages have the
problems associated with variation in raw material varieties
because this variable can be controlled for in experiments. Unfortunately not all experimentally replicated debitage assemblages produce diagnostic signatures for different production
events and technologies.
Toby Morrow [68] produced debitage counts and weights
from the replication of a bifacial core, a Clovis point, a blade
core, and a bipolar core. Again, I converted his size grades to
fit Ahler’s size grade conventions with size grade 1 as the largest and size grade 4 as the smallest. Fig. 6 compares his data as
relative frequencies of debitage counts by MA size grades. The
debitage from each of these production technologies have almost identical signatures. The highest percentage for each of
the four production technologies occurs in size grade 4 and
the lowest for all production technologies occurs in size grade
2. Size grade 3 contains approximately 15% of the debitage by
count and size grade 1 contains less than 10% of the debitage
for all manufacturing technologies. The trends in relative distributions of weight for the same tool replications show the
same uniform distribution of size grades, as do the counts
(Fig. 7). The reduction of four very different kinds of objective
pieces has produced identical debitage size grade proportions
using both counts and weights. When these data are graphed
linearly (Fig. 8), the similarity is even more pronounced.
Table 4 lists student t-scores on a two-tailed test for each
size grade and suggests that all size-graded data could have
CLOVIS
BLADE
BIPOLAR
Reduction Technique
Fig. 6. Histogram showing relative percentages of size grades by count for
bifacial reduction (BIFACE), Clovis point production (CLOVIS), blade reduction (BLADE), and bipolar reduction (BIPOLAR). Data are taken from
Morrow [68].
originated from the same reduction technology. Together,
these statistics show that the debitage for each replicated
assemblage could have been drawn from a single population.
However, we know that was not the case. Four very different
kinds of production activities produced each of the replicated
assemblages. Mass Analysis of size grades does not effectively
discriminate amongst these four assemblages. This certainly
suggests serious problems with using MA as a technique for
interpreting excavated assemblages of debitage.
1
%WT S1
%WT S2
%WT S3
%WT S4
0.9
0.8
% Debitage Weight
0
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
BIFACE
CLOVIS
BLADE
BIPOLAR
Reduction Technique
Fig. 7. Histogram showing relative percentages of size grades by weight for
bifacial reduction (BIFACE), Clovis point production (CLOVIS), blade
reduction (BLADE), and bipolar reduction (BIPOLAR). Data are taken from
Morrow [68].
W. Andrefsky Jr. / Journal of Archaeological Science 34 (2007) 392e402
398
4
1
CLOVIS
3.5
BLADE
0.8
BIPOLAR
3
0.7
Mean Flake Weight
Relative % of Debitage
AVE WT
Linear (AVE WT)
BIFACE
0.9
0.6
0.5
0.4
0.3
2.5
y = -0.3626x + 2.8326
r2 = 0.7631
2
1.5
1
0.2
0.1
0.5
0
%WT S1
%WT S2
%WT S3
%WT S4
Size Grades
Fig. 8. Line graph showing similar relative percentages of debitage size grades
by weight for bifacial reduction, Clovis point production, blade reduction, and
bipolar reduction. Data are taken from Morrow [68].
0
0
1
2
3
4
5
6
7
8
9
Reduction Stages
Fig. 9. Scatter plot of significant negatively correlated Clovis point production
debitage by reduction stages for mean flake weight. Data are taken from Morrow [68].
5. Reduction stage analysis
Investigations aimed at assessing MA have been very generous in pointing out the usefulness of this method, particularly in
conjunction with other kinds of debitage analysis. Most of
these investigators suggest that MA is most effective when predicting lithic reduction stages in an experimental context
[25,30,68]. This makes some sense because lithic tool production and core technology are reductive processes [12]. Regardless of the kind of core reduction, lithic technology results in
detached pieces that get progressively smaller as the objective
piece is worked. This trend is clear in the experimentally produced debitage from Morrow’s [68] production of a Clovis
point. Fig. 9 shows a scatter plot of the mean flake weight
for each of the eight reduction stages in the Clovis point production sequence. Clearly the trend in mean flake weight gets
progressively smaller as the production becomes more complete (mean flake weight ¼ 0.3626x þ 2.8326, r2 ¼ 0.7631).
An ANOVA shows this to be significant (F ¼ 19.325,
p ¼ 0.005).
However, inferring reduction sequences using MA is also
sensitive to the same kinds of issues noted above regarding individual knappers, raw material package size and type, and the
‘‘mixing problem.’’ For instance, reduction sequence debitage
Table 4
Students t-test: comparison of four debitage size grades by weight in Fig. 7
Size grades
Size
Size
Size
Size
grade
grade
grade
grade
1
2
3
4
One-sample test
Mean difference
t
df
Sig. (2-tailed)
21.76
5.13
4.71
3.49
3
3
3
3
<0.0005
0.014
0.018
0.040
0.8338
0.0121
0.0828
0.0513
from projectile point manufacturing will not necessarily be
analytically visible if it is mixed with reduction sequence
debitage from a bifacial core reduction or platformed core
reduction simply because early stages of projectile point production data may be the same size or even smaller than later
stages of core reduction debitage.
Given the information presented above, MA used to determine reduction stages should be cautiously used to infer reduction sequences and only after the debitage assemblage is
assessed or stratified to account for factors such as mixing
and raw material variability [11]. Making inferences about reduction sequences using MA is only effective if the debitage
assemblage represents a single reduction technology or it is
segregated to represent a single technology. For instance, it
might be effective for determining the kinds of production
techniques that have occurred at a small single use or special
use camp, such as a butchering station or bivouac, where
points were repaired or resharpened, or at a quarry location
where cobbles were primarily tested. As the number of production activities and by extension, length of occupation increases at a site location the more problematic MA gives the
complications noted above. Also, some types of reduction
technology do not result in progressively smaller detached
pieces throughout the reduction trajectory. Experiments and
analyses conducted on blade and microblade manufacturing
have shown that detached pieces from prepared blade cores
tend to stay approximately the same size throughout the production cycle [22,37,79]. This is because blade core reduction
is selected as a strategy to purposefully obtain regular and
uniform detached pieces. These kinds of technologies will
not be sensitive to reduction sequence assessment using MA
techniques.
W. Andrefsky Jr. / Journal of Archaeological Science 34 (2007) 392e402
6. Discussion
Various forms of MA have been used by lithic researchers
in the past [3,6,7,19,46,74,75,78,93], however, it did not become popular or widespread as a methodology until Ahler
published his 1989 version. Ahler was very careful in applying
MA to his assemblages of debitage. He anticipated and discussed many of the potential problems with using MA cited
here, such as the mixing problem and issues with raw material
variability. Unfortunately many practitioners of MA were not
as careful as Ahler when adopting it for their own research.
I do not believe Ahler would approve of some of the applications of MA evident in the literature today.
Mass Analysis may be effective for some types of assemblages and for gaining insights into some questions and technological issues. However, the reasons for why it was
originally adopted and has become widely used may no longer
be applicable to this technique. Given the need to stratify debitage assemblages to counter some of the effects of raw material variability and mixing technologies before conducting
MA, it may not be as time conservative as once thought.
This paper did not explore and evaluate the different techniques available to obtain debitage size grade signatures such
as discriminant function analysis [1], multiple linear regression analysis [82,90], or fractal analysis [28]. Instead, this
paper focused on the issues surrounding the production of
primary comparative assemblages (replicated assemblages).
Without reliable comparative assemblages, interpretations of
size grade debitage are useless regardless of the technique administered to derive the signatures used to compare the excavated assemblage and the replicated assemblage.
It is becoming increasingly clear that debitage assemblages
should be stratified into as many behaviorally meaningful
groups as possible before MA is performed [11,83]. For instance, it makes little sense to use MA on aggregate assemblages comprised of different types of raw materials or
variations of the same raw material type. These different raw
material packages most likely represent different reduction or
production episodes, even if the materials were used to make
the same kinds of tools [53,59,60,86]. The debitage from each
of those episodes should be handled as separate analytical populations for MA to be an effective form of aggregate analysis.
In addition to stratification by raw material varieties, evidence for different kinds of lithic production technologies can
be used to stratify debitage assemblages to help reduce the
amount of assemblage mixing before MA is preformed. There
is no reason to expect that MA will produce accurate results if
the debitage from multiple production technologies are combined. Debitage with characteristics diagnostic of specific production/reduction technologies (such as bifacial trimming
flakes, notching flakes, prismatic blades, bipolar flakes, etc.)
should be partitioned into separate analytical populations before
attempting MA. For instance, if we are able to identify debitage
associated with bifacial production or bipolar reduction, or platformed core reduction, this information can be used to help stratify debitage assemblages into specific production episodes to
reduce the mixing problem so often associated with MA. Such
399
technological information is available to debitage analysts
[12,33,40,42,44,55,64,71,83,95,97].
Because size grade corresponds to uniform and universal
mesh dimensions, many believe MA is a more objective means
of classifying debitage by size [1]. We now realize that size
grades that correspond to mesh sizes may not be as reliable
or as objective as once believed [13,83,84]. When debitage
is sorted and passed through the mesh it may get trapped in
the mesh in an unsystematic manner. The same debitage assemblage might have different group representation if it is
sorted multiple times simply because of how individual flakes
are oriented when passing through the mesh. Long thin flakes
may pass through a particular mesh size on one occasion but
not the next time simply given the flake orientation. Carr
and Bradbury [30] suggest passing specimens through a screen
by hand to circumvent this potential problem, or passing debitage individually through size templates [12].
A recent example underscores how MA can lead to spurious conclusions. Franklin and Simik [41] conducted artifactrefitting studies and compared their results to the results
obtained using MA based upon discriminant function signatures on principal components of size grade attributes. Their
refitting study conclusively established bipolar techniques as
the primary technology used to open and to reduce chert nodules at a rock-shelter in Tennessee. A series of cobble reduction experiments were conducted to establish comparative
debitage collections (of the same raw materials) for bipolar
reduction, hard hammer cobble testing, hard hammer core reduction, and soft hammer tool production. When these control
populations were compared to the excavated assemblage using
MA, bipolar reduction was not recognized. Again, even
though bipolar technology was conclusively recognized by refitting cores, MA was not effective in discerning this
manufacturing technology. This is not surprising given the
complications of performing reliable MA as noted above.
There is another issue with MA related to the circularity in
its logic. The potential benefit of using MA for investigators
lies in its ability to make inferences about the kinds of tools
produced or the kinds of reduction activities that have taken
place where debitage is found. However, for MA to be effective a replicated assemblage must be produced to compare
with the excavated assemblage. How does the investigator determine the kinds of tools or technologies to replicate? This is
a serious circularity problem. If the investigator knows the
kinds of tools made to produce the excavated debitage
assemblage e why conduct a MA if the aim of MA is to
determine the kinds of tools made? If the investigator does
not know the kinds of tools made to produce the excavated
debitage e how can reliable replications of the assemblage
be conducted, particularly given the fact that tool-stone size
and shape, as well as production activities, influence the
size and amount of debitage produced.
7. Conclusion
The results of this study emphasize that size grade patterning of debitage used for MA is subject to multiple
400
W. Andrefsky Jr. / Journal of Archaeological Science 34 (2007) 392e402
sources of error. Production debitage is greatly influenced by
replicator variability and it is unlikely that contemporary experimental assemblages can be expected to match excavated
assemblages, even if the same types of artifacts were manufactured. Individual knapping styles have as much influence
on debitage size as does the kind of tool produced. I also
demonstrated how raw material variability, particularly package size and shape, could also influence the size of debitage
produced in replication experiments. Raw material type and
form must be accounted for when comparing excavated assemblages to replicated assemblages by size grades. Experimentally replicated debitage that has been mixed does not
accurately reveal the technological activities from which
the debitage was produced. This problem is almost certainly
more pronounced when dealing with excavated assemblages
where multiple lithic production activities may have been
performed. Lastly, I showed how some replicated production
activities do not produce significantly different debitage assemblages when making size grade comparisons, even
when very different kinds of tools are produced. All of these
potential and probable sources of error make using MA
a questionable analytical strategy for interpreting excavated
debitage assemblages unless these factors are adequately
controlled and accounted for.
Clearly, MA of experimentally generated debitage assemblages has been an effective heuristic for chipped stone technological analysis, even if they are difficult to confidently
relate to excavated assemblages. We have learned a great
deal about the sources of debitage assemblage variability as
a result of our explorations into MA. However, as shown
above, MA is not effective for making accurate tool production or core reduction interpretations at archaeological sites
unless we can account for the many sources of debitage size
variability.
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
Acknowledgments
A version of this paper was presented at the 2005 meeting
of the Middle Atlantic Archaeological Conference in Delaware. I thank those organizers and colleagues for providing
a forum and comments to that initial draft. The College of Liberal Arts at Washington State University provided professional
leave time and funding through their Edward R. Meyer Project
Grant to allow me to complete this paper. I am grateful to them
and appreciate their contributions. A number of colleagues
have provided valuable editorial comments and suggestions
to the early versions of this manuscript. I acknowledge the
helpful commentary of four anonymous JAS reviewers and
Bill Lipe, Jennifer Keeling, Beth Horton, and especially Karen
Lupo. All have contributed towards helping my less-thanconcise manner of writing.
References
[1] Stanley A. Ahler, Mass analysis of flaking debris: studying the forest
rather than the trees, in: Donald O. Henry, George H. Odell (Eds.),
[16]
[17]
[18]
[19]
[20]
[21]
[22]
Alternative Approaches to Lithic Analysis, Archaeological Papers of
the American Anthropological Association No. 1 (1989), pp. 85e118.
Stanley A. Ahler, R.C. Christensen, A Pilot Study of Knife River Flint
Procurement and Reduction as Site 32DU508, A Quarry and Workshop
Location in Dunn County, North Dakota, Department of Anthropology
and Archaeology, University of North Dakota, Bismark, 1983.
Stanley A. Ahler, Julian VanNess, Temporal change in knife river flint
reduction Strategies, in: Susan C. Vehik (Ed.), Lithic Resource Procurement: Proceedings from the Second Conference on Prehistoric Chert
Exploitation (Occasional Papers 4), Center for Archaeological Investigations, Southern Illinois University, Carbondale, 1985, pp. 183e198.
Daniel S. Amick, Raymond P. Mauldin, Effects of raw material on flake
breakage patterns, Lithic Technology 22 (1997) 18e32.
Daniel S. Amick, Raymond P. Mauldin, Steven A. Tomka, An evaluation
of debitage produced by experimental bifacial core reduction of a Georgetown chert nodule, Lithic Technology 17 (1988) 26e36.
Albert J. Ammerman, A study of obsidian exchange networks in Calabria, World Archaeology 19 (1979) 95e110.
Albert J. Ammerman, William Andrefsky Jr., Reduction sequences and
the exchange of obsidian in Neolithic Calabria, in: J. Ericson, T. Earle
(Eds.), Contexts for Prehistoric Exchange, Academic Press, New York,
1982, pp. 149e172.
William Andrefsky Jr., A consideration of blade and flake curvature,
Lithic Technology 15 (1986) 48e54.
William Andrefsky Jr., The geological occurrence of lithic material and
stone tool production strategies, Geoarchaeology: An International Journal 9 (1994) 345e362.
William Andrefsky Jr., Emerging directions in debitage analysis, in:
W. Andrefsky Jr. (Ed.), Lithic Debitage: Context, Form, Meaning, University of Utah Press, Salt Lake City, 2001, pp. 1e14.
William Andrefsky Jr., Partitioning the aggregate: mass analysis and debitage assemblages, in: Christopher T. Hall, Mary Lou Larson (Eds.), Aggregate Analysis in Chipped Stone, University of Utah Press, Salt Lake
City, 2004, pp. 201e210.
William Andrefsky Jr., Lithics: Macroscopic Approaches to Analysis,
second ed., Cambridge University Press, Cambridge, 2005.
William Andrefsky Jr., Lithic studies, in: H.D.G. Maschner,
C. Chippindale (Eds.), Handbook of Methods in Archaeology, vol. II,
AltaMira Press, Walnut Creek, CA, 2005, pp. 713e770.
William Andrefsky, Jr., Experimental and archaeological verification of
an index of retouch for hafted bifaces, American Antiquity, in press.
Robert J. Austin, Technological Characterization of lithic waste-flake
assemblages: multivariate analysis of experimental and archaeological
data, Lithic Technology 24 (1999) 53e68.
Douglas B. Bamforth, Technological organization and hunter-gatherer
land use: a California example, American Antiquity 56 (1991) 216e
234.
Mark F. Baumler, Leslie B. Davis, The role of small-sized debitage in
aggregate lithic analysis, in: Christopher T. Hall, Mary Lou Larson
(Eds.), Aggregate Analysis in Chipped Stone, University of Utah Press,
Salt Lake City, 2004, pp. 45e64.
Mark F. Baumler, Christian E. Downum, Between micro and macro:
a study in the interpretation of small-sized lithic debitage, in: Daniel S.
Amick, Raymond P. Mauldin (Eds.), Experiments in Lithic Technology,
BAR International Series 528, 1989, Oxford, pp. 101e116.
Jeffery A. Behm, Flake concentrations: distinguishing between flint
working activity areas and secondary deposition, Lithic Technology 12
(1983) 9e16.
Brooke S. Blades, End scraper reduction and hunter-gatherer mobility,
American Antiquity 68 (2003) 141e156.
Peter Bleed, The optimal design of hunting weapons: maintainability or
reliability, American Antiquity 51 (1986) 737e747.
Peter Bleed, Cheap, regular, and reliable: implications of design variation
in late pleistocene Japanese microblade technology, in: Robert G. Elston,
Steven L. Kuhn (Eds.), Thinking Small: Global Perspectives on Microlithization, Archaeological Papers of the American Anthropological
Organization, Number 12, 2002, Arlington, Virginia, pp. 95e102.
W. Andrefsky Jr. / Journal of Archaeological Science 34 (2007) 392e402
[23] Peter Bleed, Refitting as aggregate analysis, in: Christopher T. Hall,
Mary Lou Larson (Eds.), Aggregate Analysis in Chipped Stone, University of Utah Press, Salt Lake City, 2004, pp. 184e200.
[24] W.W. Bloomer, Eric E. Ingbar, Debitage analysis, in: R.G. Elston,
C. Raven (Eds.), Archaeological Investigations at Tosawihi, A Great Basin Quarry, Part I: The Periphery, Intermountain Research, Silver City,
Nevada, 1992, pp. 229e270 (Prepared for the U.S.D.A., Bureau of
Land Management, Elko Resource Area, Elko, Nevada).
[25] Andrew P. Bradbury, Phillip J. Carr, Flake typologies and alternative
approaches: an experimental assessment, Lithic Technology 20 (1995)
100e115.
[26] Andrew P. Bradbury, Phillip J. Carr, Examining stage and continuum
models of flake debris analysis: an experimental approach, Journal of
Archaeological Science 26 (1999) 105e116.
[27] Andrew P. Bradbury, Jay D. Franklin, Material variability, package size
and mass analysis, Lithic Technology 25 (2000) 42e58.
[28] C.T. Brown, The fractal dimensions of lithic reduction, Journal of
Archaeological Science 28 (2000) 619e631.
[29] Philip J. Carr, Andrew P. Bradbury, Flake debris analysis, levels of production, and the organization of technology, in: W. Andrefsky Jr. (Ed.),
Lithic Debitage: Context, Form, Meaning, University of Utah Press,
Salt Lake City, 2001, pp. 126e146.
[30] Philip J. Carr, Andrew P. Bradbury, Exploring mass analysis, screens, and
attributes, in: Christopher T. Hall, Mary Lou Larson (Eds.), Aggregate
Analysis in Chipped Stone, University of Utah Press, Salt Lake City,
2004, pp. 21e44.
[31] John E. Clark, Flintknapping and debitage disposal among the Lacandon
Maya of Chiapas, Mexico, in: Edward Staski, L.D. Sutro (Eds.), The
Ethnoarchaeology of Refuse Disposal, Anthropological Research Papers
No. 42, Arizona State University, Tempe, 1991, pp. 63e78.
[32] Chris Clarkson, An index of invasiveness for the measurement of unifacial and bifacial retouch: a theoretical, experimental and archaeological
verification, Journal of Archaeological Science 29 (2000) 65e75.
[33] Don E. Crabtree, An Introduction to Flintworking, Occasional Papers of
the Idaho State Museum 28, 1972, Pocatello.
[34] Zachary J. Davis, John J. Shea, Quantifying lithic curation: an experimental test of dibble and pelcin’s original flake-tool mass predictor,
Journal of Archaeological Science 25 (1998) 603e610.
[35] Harold L. Dibble, Platform variability and flake morphology: a comparison of experimental and archeological data and implications for interpreting prehistoric lithic technological strategies, Lithic Technology 22
(1997) 150e170.
[36] Harold L. Dibble, Andrew Pelcin, The effect of hammer mass and velocity on flake mass, Journal of Archaeological Science 22 (1995) 429e439.
[37] Robert G. Elston, P. Jeffery Brantingham, Microlithic technology in
northern asia: a risk-minimizing strategy of the Late Paleolithic and
Early Holocene, in: R.G. Elston, S.L. Kuhn (Eds.), Thinking Small:
Global Perspectives on Microlithization, Anthropological Papers of the
American Anthropological Organization, Number 12, 2002, Arlington,
Virginia, pp. 103e116.
[38] Metin I. Eren, Manual Dominguez-Rorigo, Steven L. Kuhn, Daniel
S. Adler, Ian Le, Ofer Bar-Yosef, Defining and measuring reduction in
unifacial stone tools, Journal of Archaeological Science 32 (2005)
1190e1206.
[39] Knut R. Fladmark, Microdebitage analysis: initial considerations, Journal
of Archaeological Science 9 (1982) 205e220.
[40] J. Jeffery Flenniken, Stone tool reduction techniques as cultural markers,
in: M.G. Plew, J.C. Woods, M.G. Pavesic (Eds.), Stone Tool Analysis:
Essays in Honor of Don E. Crabtree, University of New Mexico Press,
Albuquerque, 1985, pp. 265e277.
[41] Jay D. Franklin, Jan F. Simik, Core refitting and the accuracy of techniques
for aggregate lithic analysis, in: Paper presented at the 2001 Annual Meeting of the Society for American Archaeology, New Orleans, LA, 2001.
[42] George C. Frison, A functional analysis of certain chipped stone tools,
American Antiquity 33 (1968) 149e155.
[43] Amy Gilreath, Stages of bifacial manufacture: learning from experiments, in: Paper presented at the 49th Annual Meeting of the Society
for American Archaeology, Portland, Oregon, 1984.
401
[44] Albert C. Goodyear, Tool kit entropy and bipolar reduction: a study of
interassemblage lithic variability among Paleo-Indian Sites in the Northeastern United States, North American Archaeologist 14 (1993) 1e23.
[45] Richard A. Gould, Ethno-archaeology; or, where do models come from?
A closer look at Australian aboriginal lithic technology, in: R.V.S. Wright
(Ed.), Stone Tools as Cultural Markers: Change, Evolution and Complexity, Prehistory and Material Culture Series No. 12, Australian Institute of
Aboriginal Studies/Humanities Press, Canberra/New Jersey, 1977, pp.
162e168.
[46] D.O. Henry, C.V. Haynes, B. Bradley, Quantitative variations in flaked
stone debitage, Plains Anthropologist 21 (1976) 57e61.
[47] Peter Hiscock, Val Attenbrow, Early Australian implement variation: a reduction model, Journal of Archaeological Science 30 (2003) 239e249.
[48] Kenneth H. Honea, The Bipolar flaking technique in Texas and New
Mexico, Texas Archaeological Society Bulletin 36 (1965) 259e267.
[49] Kathleen L. Hull, Identification of cultural site formation processes
through microdebitage analysis, American Antiquity 52 (1987) 772e783.
[50] Eric E. Ingbar, Mary Lou Larson, Bruce Bradley, A nontypological approach to debitage analysis, in: D.S. Amick, R.P. Mauldin (Eds.), Experiments in Lithic Technology, BAR International Series 528, 1989, Great
Britain, pp. 117e136.
[51] Jay K. Johnson, Lithic procurement and utilization trajectories: analysis,
yellow creek nuclear power plant site, Tishomingo County, Mississippi,
In: Archaeological Papers No. 1, vol. II, Center for Archaeological
Research, University of Mississippi, University, 1981.
[52] Jeffrey Kalin, Stem point manufacture and debitage recovery, Archaeology of Eastern North America 9 (1981) 134e175.
[53] Edward J. Knell, Late Paleoindian cody period mobility patterns: an example from the locality V cody complex at the Hell Gap site, Unpublished MA thesis, Department of Anthropology, University of
Wyoming, Laramie, 1997.
[54] Edward J. Knell, Coarse-scale chipped stone aggregates and technological organization strategies at Hell Gap locality V cody complex component, Wyoming, in: Christopher T. Hall, Mary Lou Larson (Eds.),
Aggregate Analysis in Chipped Stone, University of Utah Press, Salt
Lake City, 2004, pp. 156e183.
[55] Brian P. Kooyman, Understanding Stone Tools and Archaeological Sites,
University of Calgary Press, Calgary, 2000.
[56] Steven L. Kuhn, A geometric index of reduction for unifacial stone tools,
Journal of Archaeological Science 17 (1990) 585e593.
[57] Steven L. Kuhn, ‘‘Unpacking’’ reduction: lithic raw material economy in
the mousterian of west-central Italy, Journal of Anthropological Archaeology 10 (1991) 76e106.
[58] Ian Kuijt, William C. Prentiss, David J. Pokotylo, Bipolar reduction: an
experimental study of debitage variability, Lithic Technology 20 (1995)
116e127.
[59] Mary Lou Larson, Site formation processes in the cody to early plains
archaic levels at the laddie creek site (48BH345), Wyoming, Geoarchaeology 7 (1992) 103e120.
[60] Mary Lou Larson, Marcel Kornfeld, Chipped stone nodules: theory,
method, and examples, Lithic Technology 22 (1997) 4e18.
[61] Mary Lou Larson, Judson B. Finley, Seeing the forest but missing the
trees: production sequences and multiple linear regression, in:
Christopher Hall, Mary Lou Larson (Eds.), Aggregate Analysis in
Chipped Stone, University of Utah Press, Salt Lake City, 2004, pp.
95e111.
[62] Thomas K. Larson, Dori Penny, Ross G. Hilman, Paul H. Sanders, A data
recovery program for sites within the dome unit, Santa Fe National Forest, Report Prepared by Larson-Tibesar Associates, Laramie, Wyoming
for USDA, Forest Service, Santa Fe National Forest, Santa Fe, New Mexico, 1988.
[63] Mary M.A. MacDonald, Technological organization and sedentism in the
epipalaeolithic of dakhleh oasis, Egypt, The African Archaeological
Review 9 (1991) 81e109.
[64] Martin P. Magne, Lithics and livelihood: stone tool technologies of
central and southern interior British Columbia, Archaeological Survey
of Canada Paper No. 133, National Museum of Man Mercury Series,
Ottawa, 1985.
402
W. Andrefsky Jr. / Journal of Archaeological Science 34 (2007) 392e402
[65] Martin P. Magne, David Pokotylo, A pilot study in bifacial lithic reduction sequences, Lithic Technology 10 (1981) 34e47.
[66] Raymond P. Mauldin, Daniel S. Amick, Investigating patterning in
debitage from experimental bifacial core reduction, in: D.S. Amick,
R.P. Mauldin (Eds.), Experiments in Lithic Technology, BAR International Series 528, 1989, Oxford, pp. 67e88.
[67] Mark W. Moore, Australian aboriginal blade production methods on the
Georgina River, Camooweal, Queensland, Lithic Technology 28 (2002)
35e63.
[68] Toby A. Morrow, A chip off the old block: alternative approaches to debitage analysis, Lithic Technology 22 (1997) 51e69.
[69] Dani Nadel, Indoor/outdoor flint knapping and minute debitage remains:
the evidence from the Ohalo II submerged camp (19.5 KY, Jordan Valley), Lithic Technology 26 (1999) 118e137.
[70] Deborah Olausson, Different strokes for different folks, possible reasons for
variation in quality of knapping, Lithic Technology 23 (1997) 90e115.
[71] Terry L. Osbun, A technological analysis of the lithic assemblage from
the Buttonhole Rochshelter/Quarry Site, Northeastern New Mexico, Unpublished MA thesis, Department of Anthropology, Washington State
University, Pullman, 1987.
[72] William J. Parry, Prismatic blade technologies in North America, in:
P.J. Carr (Ed.), The Organization of North American Prehistoric Chipped
Stone Tool Technologies, International Monographs in Prehistory, Ann
Arbor, 1994, pp. 87e98.
[73] William J. Parry, Robert L. Kelly, Expedient core technology and sedentism, in: J.K. Johnson, C.A. Morrow (Eds.), The Organization of Core
Technology, Westview Press, Boulder, Colorado, 1987, pp. 285e304.
[74] Leland W. Patterson, Characteristics of bifacial-reduction flake-size distribution, American Antiquity 55 (1990) 550e558.
[75] Leland W. Patterson, J.B. Sollberger, Replication and classification of
small size lithic debitage, Plains Anthropologist 23 (1978) 103e112.
[76] Albert M. Pecora, Chipped Stone tool production strategies and lithic debris
patterns, in: William Andrefsky Jr. (Ed.), Lithic Debitage: Context, Form,
Meaning, University of Utah Press, Salt Lake City, 2001, pp. 173e191.
[77] P.R. Picha, Mark L. Gregg, Chipped stone flaking debris from the Nase
Site, Archaeological Excavations at the Nase Site (32SN246), Report
submitted to U.S.D.I. Bureau of Reclamation, Missouri Basin Region,
Billings, MT, Department of Anthropology, University of North Dakota,
Grand Forks, 1987.
[78] L. Mark Raab, R.F. Cande, D.W. Stahle, Debitage graphs and archaic
settlement patterns in the Arkansas Ozarks, Midcontinential Journal of
Archaeology 4 (1979) 167e182.
[79] Jeffery C. Rasic, William Andrefsky Jr., Alaskan blade cores as specialized components of mobile toolkits: assessing design parameters and
toolkit organization through debitage analysis, in: W. Andrefsky Jr.
(Ed.), Lithic Debitage: Context, Form, Meaning, University of Utah
Press, Salt Lake City, 2001, pp. 61e79.
[80] Kimberly L. Redman, An experiment-based evaluation of the debitage
attributes associated with ‘‘hard’’ and ‘‘soft’’ hammer percussion, Unpublished M.A. thesis, Washington State University, Pullman, WA, 1997.
[81] Matthew J. Root, The knife river flint quarries: the organization of stone
tool production, Ph.D. dissertation, Washington State University. University Microfilms, Ann Arbor, 1992.
[82] Matthew J. Root, Production for exchange at the Knife River flint
quarries, North Dakota, Lithic Technology 22 (1997) 33e50.
[83] Matthew J. Root, Technological analysis of flake debris and the limitations of size-grade techniques, in: Christopher Hall, Mary Lou Larson
(Eds.), Aggregate Analysis in Chipped Stone, University of Utah Press,
Salt Lake City, 2004, pp. 65e94.
[84] S.A. Scott, Problems with the use of flake size in inferring stages of lithic
reduction, Journal of California and Great Basin Anthropology 13 (1991)
172e179.
[85] Frederic Sellet, Chaı̂ne opératoire: the concept and its applications,
Lithic Technology 18 (1993) 106e112.
[86] Frederic Sellet, A dynamic view of Paleoindian assemblages at the Hell
Gap site, Wyoming: reconstructing lithic technological systems, Unpublished Ph.D. dissertation, Department of Anthropology, Southern Methodist University, Dallas, 1999.
[87] Harry J. Shafer, Late preclassic formal stone tool production at Colha,
Belize, in: T.R. Hester, H.J. Shafer (Eds.), Maya Stone Tools, Prehistory
Press, Madison, 1991, pp. 31e44.
[88] Harry J. Shafer, Thomas R. Hester, Ancient maya chert workshops in
Northern Belize, Central America, American Antiquity 48 (3) (1983)
519e543.
[89] Phillip H. Shelley, Variation in lithic assemblages: an experiment, Journal of Field Archaeology 17 (1990) 187e193.
[90] Michael J. Shott, Lithic reduction at 13HA365, a middle woodland occupation in Hardin County, Journal of the Iowa Archaeological Society 44
(1997) 109e120.
[91] Michael J. Shott, Paul Sillitoe, Use life and curation in New Guinea
experimental used flakes, Journal of Archaeological Science 32 (2005)
653e663.
[92] April K. Sievert, Karen Wise, A generalized technology for a specialized
economy: Archaic Period chipped stone at kilometer 4, Peru, in:
William Andrefsky Jr. (Ed.), Lithic Debitage: Context, Form, Meaning,
University of Utah Press, Salt Lake City, 2001, pp. 188e206.
[93] David W. Stahle, James E. Dunn, An analysis and application of the size
distribution of waste flakes from the manufacture of bifacial tools, World
Archaeology 14 (1982) 84e97.
[94] Alan P. Sullivan III, Kenneth C. Rozen, Debitage analysis and archaeological interpretation, American Antiquity 50 (1985) 755e779.
[95] Gene Titmus, Some aspects of stone tool notching, in: Marc G. Plew, Max
G. Pavesic (Eds.), Stone Tool Analysis: Essays in Honor of Don E. Crabtree, University of New Mexico Press, Albuquerque, 1985, pp. 243e264.
[96] Steven A. Tomka, Differentiating lithic reduction techniques: an experimental approach, in: D.S. Amick, R.P. Mauldin (Eds.), Experiments in
Lithic Technology, BAR International Series 528, 1989, Great Britain,
pp. 137e162.
[97] John C. Whittaker, Flintknapping: Making and Understanding Stone
Tools, University of Texas Press, Austin, 1994.