Claassen 1991

Normative Thinking and Shell-Bearing Sites
Author(s): Cheryl Claassen

Source: Archaeological Method and Theory , 1991, Vol. 3 (1991), pp. 249-298
Published by: Springer
Stable URL: https://www.jstor.org/stable/20170217
REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/20170217?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Springer is collaborating with JSTOR to digitize, preserve and extend access to Archaeological
Method and Theory
This content downloaded from

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
All use subject to https://about.jstor.org/terms
6 Normative
Thinking and
Shell-Bearing Sites
CHERYL CLAASSEN
Normative thinking or normative treatment is that whic

seeks to identify typical or average behavior and occurrence
gional syntheses are usually normative in that they emphasize
"facts" while ignoring others. A poorly developed chronology is
ducive to normative thinking, as is an idea introduced by a respe
colleague in archaeology. In fact, a social science version of
gion theory would enlighten the history and explanatory streng
many normative concepts. So strong is the allure of normative th
ing that several bright new ideas introduced in the 1970s and
were no more than the questioning of normative assumption
results of normative thinking are the distortion of an archaeolo
record of variation in adaptive responses into a record of homog
ous response and the enlivening of communities and even su
tence subsystems as organic, living actors.
One category of normative treatment in shell midden work
intramidden and intermidden variation in deposits. Several v
have criticized previous work for denying or ignoring variation
Ambrose 1967) in shell deposits or using normative descript
(McManamon 1984:350), and one might assume that intermi
and intramidden variation is now accommodated in excavatio
interpretation. A failure to recognize variation is, however, still
dent in the rudimentary efforts made in naming and describing t
and in sampling middens. I suspect that the denial of variat
partially responsible for the lack of experimentation with shells
shell deposits.
Another category of normative thinking concerns food and
ing. Typically, all debris from edible taxa in a shell-bearing
249

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
250 Cheryl Ciaassen
portrayed as human food debris. Furthermore, normative thought

and language pervade studies of consumer behavior and animal
habitats, and perhaps explain the paucity of ethnoarchaeological
studies involving coastal gatherers and fishers.
In this chapter I will explore these two categories of normative
thinking, citing examples and contradictory information which indi
cate that ranges, rather than means, of behaviors and deposits must
be incorporated into our study of individual shell-bearing sites and
into regional syntheses and models of human adaptation. I will begin
with the category midden variation and briefly explore (1) types of
archaeological deposits, (2) site sampling, particularly that employed
for faunal analysis, and (3) experiments in midden formation pro
cesses. The second half of the chapter will explore the generaliza
tions about food and feeding. Here I will focus on interpretations of
(1) gathering, (2) consumer behavior, (3) shellfish nutrition, (4) paleo
environments, and (5) the contradictions revealed by ethnoarchae
ology. The normative basis of much archaeological interpretation has
given rise to false issues, which I will briefly outline in the closing
pages.
The primary data base I will be using for examples and discussion
is the recent literature on shell mounds and middens in the continen
tal United States, chosen because of my familiarity with it. Neither
the problems created by normative thought nor the particular expres
sions of normative thought, however, are unique to this geographical
area. Workers in other countries have pointed out problems identical
to those I will discuss.
Midden Variation
Normative thinking has reduced all archaeological shell

and associated fauna and flora debris to food debris and, conse
quently, all forms of shell-bearing deposits to secondary refuse de
posits. The assumption that secondary refuse is being excavated has
governed excavation and sampling strategies. But the ethnographic
record provides many examples of systemic uses for discarded shell.
Widmer (1989) has pointed out that our failure to identify types of
shell-bearing deposits has led to misinterpretation of sites and the
failure to recover numerous classes of data.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
Normative Thinking and Shell-Bearing Sites 251
Types of Shell-bearing Deposits and Sites
Several archaeologists have recognized a need to refine terminology

and site typologies. Some have delineated types of systemic shell
deposits (e.g., Lightfoot 1985; Waselkov 1987:117), and others have
named types of deposits in archaeological context (e.g., McManamon
1984; Widmer 1989). Some have proposed systemic models of deposit
formation (e.g., May 1982; Nodine 1986; Waselkov 1987:116-17), and
some, models of processes transforming the systemic deposit into
the archaeological deposit (e.g., Beaton 1985; Ham 1982).
Like an artifact typology, the types of shell-bearing deposits iden
tified depend on the research question. Questions about systemic
variation in the origin and life of these deposits require a typology
drawn from the ethnographic literature. In systemic context we find
that origins of shell-bearing deposits include industrial waste (shell
button, cameo, dye, porcelain, lime manufacture, etc.), architectural
features, and subsistence waste (bait, food). Models of shell midden
formation in systemic context have basically been two: living on
shell debris or living away from shell debris.
But how visible are these systemic nuances in archaeological con
text? For most archaeologists the most useful typology is one that
will allow meaningful distinctions between archaeological deposits.
In archaeological context, systemic deposits of industrial waste, gar
bage, and architectural features take the form of shell-filled pits,
faci?s, scatters, lenses, and mounds, to name just a few possibilities.
Any one of these types of deposits might have been used by one
social group to hold shell for architectural use or to dispose of subsis
tence or industrial waste. Any one of these types of deposits or com
binations might appear at a dinnertime camp, or a short-term or a
long-term residential base. Any one of these types of deposits could
be created by women or men, state or band members, farmers or
habitual gatherers, urban dwellers or nomads. A shell deposit typol
ogy based on archaeological criteria is the most direct way to eluci
date variation in the archaeological record. Typologies proposed by
McManamon (1984) and Widmer (1989) begin with the archaeologi
cal characteristics and sort deposits into archaeological deposit types.
On the basis of density indices for lithic tool types, lithology, fire
cracked rock, fauna, and flora in sites on Cape Cod, Massachusetts,
McManamon detected four types of deposits: primary refuse, limited

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
252 Cheryl Ciaassen
activities; primary refuse, wide range of activities; secondary refuse,

shell midden,- secondary refuse, general midden. Without further spe
cifying the original activities that created these deposits, McMana
mon could discuss the spatial organization of some obvious activi
ties, such as shellfishing and lithic manufacture, in each cultural
period.
Shell midden is applied to a variety of shell-bearing deposits. In
the United States, shell mound or shell midden is a meaningless
semantic distinction to most archaeologists. In Australia, the term
shell midden mound is occasionally seen. (Among many archaeolo
gists even the term maritime adaptation is used unsystematically,
meaning either the use of estuarine animals and plants but not
pelagic animals or, conversely, subsistence based on the capture of
pelagic animals.) Several workers have specified the ratio of shell to
sediment or the density of shell per unit volume in a definition of shell
midden, presumably attempting to standardize the term. Widmer
(1989) has proposed that archaeological sites containing shell de
posits be termed shell-bearing sites in lieu of specific knowledge
about their formation processes. To use a more specific label would
require the investigator to assess formation processes.
Widmer (1989) has simultaneously defined shell midden and de
veloped a typology based on his excavation experience in southwest
Florida. The underlying organization of this typology is a distinction
between site and deposit:
1. Shell midden site?secondarily deposited shell from food con
sumption with no other activities evident at the site
2. Shell midden?discrete lens or deposit of shell only
3. Shell-bearing midden site?a site composed of secondary
refuse of many kinds of remains, including shell, generated by
a wide range of activities
4. Shell-bearing habitation site?primarily shell debris in site
matrix used for architectural needs; the shell may or may not
have originated as food debris.
Perhaps the greatest service this typology does is to acknowledge
that shell may be present in a specific locale for some reason other
than food debris. Assuming that all shell debris equals human food
debris, we make further assumptions about the gender of the collec
tors and the site's place in the settlement system. Where there is
much shell and few artifacts, the shell debris is from a women's din
nertime or processing camp,- where there is much shell and many

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
artifacts, the site is a base camp or aggregation point for many ages
and both sexes.
An alternative explanation for the presence of shells in significant
quantities is that shells were amassed as structural features. Many
authors have proposed that shells have been used to create a firm
and dry base for subsequent occupation. Onat (1985) advanced the
Community Engineering Model to explain large quantities and
strata of shells in Northwest Coast shell middens. On the basis of
ethnographic records, she outlined the use of shells to prepare village
sites: in house floor and hearth construction and maintenance, as
retaining walls, and as house insulation.
Chief among the unconsidered explanations for shell debris is fish
bait refuse. Shellfish and fish were and are used in a variety of ways
as bait. Chumming, or baiting an area of water to attract prey, is a
common use for shellfish and fish meat. Both were also used on
hooks, in nets, and in fish baskets (Larson 1980; Speck 1946). Bivalve
meat was and is used to capture spiny lobsters (Goode 1884:701),
shrimp, crayfish, crabs, and numerous fish (Goode 1884; Viosca
1933), any one of which could subsequently be used as bait.
It is conceivable that some shell middens are composed almost
entirely of bait refuse. Such a proposal has been put forward by Riser
(1987) to explain the ubiquitous Rangia shell middens around Lake
Pontchartrain, Louisiana, that contain little but shells. The Rangia
flesh would have been used to chum for shrimp, as was done histori
cally. Middens of fish bait have also been suspected in southern
California (idea of Roy Sails ; Clement Meighan, personal communi
cation June 1987). I suspect there are other candidates as well, such
as the Tasmanian Hunter Island sites dating before 3500 bp, Chan
tuto middens in Chiapas, and early Jomon period sites. A sizable shell
midden containing numerous shells and fish bones could accumu
late as fish pots were emptied of their spent bait. In considering the
bait hypothesis, we should be wary of assuming that bait and prey
would end up in the same refuse pile and even that the prey was then
used as human food. It, too, could have been used for bait.
Formation processes play a crucial role in the appearance and con
tents of a shell-bearing deposit or site at the time it is excavated. The
Brennan-Sanger debate (e.g., Brennan 1977; S?nger 1981) has focused
attention on assumptions of contemporaneity of artifacts with their
matrix, as have Stein's (1983) revelations about earthworm activity.
Muckle (1985) cites an extensive literature on shell taphonomy that

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
254 Cheryl Ciaassen
is rarely consulted by archaeologists. Shell orientation, fragmenta

tion, and articulation tell us much about disturbance in middens
and mounds, and constitute microstratification.
The study of shell midden and mound formation processes will
result in yet a different kind of typology. It could be that the most
useful typology for excavators will be one which identifies sequences
of formation processes.
Sampling
Voigt (1975:88) reminds us that "until we can obtain statistically
significant samples, our work must to a certain extent remain experi
mental." How to sample a shell-bearing site has been discussed by
Bowdler (1983), Campbell (1981), Maxwell (1989a), Peacock (1978),
Waselkov (1987), Widmer (1989), and many others. Campbell sees
the component as a more appropriate sampling universe than the
site, while Waselkov (1987:150-53) views occupational areas as the
most appropriate sampling arena. Sampling sites or components en
courages normative thinking, which is less likely when occupational
episodes are deemed the appropriate universe. However, occupation
episodes and even components can be extremely elusive in many
sites.
The perception of elusive dump episodes, occupations, and compo
nents may be more a theoretical problem than a real one. There are
two fundamentally different perceptions of what is preserved in
shell-bearing sites. On the one hand are excavators who believe that
shell middens are in systemic context a deflated mass of multiple
deposits, the stew from millennia of activities (e.g., Begier and
Keatinge 1979:220). On the other hand are those excavators who hold
the opposite perception, that "shell middens" are conflations of dis
crete deposits, each conflated with and separated by the shells depos
ited along with the other debris (e.g., Bailey and Parkington 1988:4;
Waselkov 1987:145; Widmer 1989:11). The former attitude will most
likely encourage excavation in arbitrary levels with small exposures
and a shell-bearing site typology based on excavated information;
the latter attitude encourages excavation in natural levels, extensive
exposures, and a systemic-based typology. Experimental work with
shell deposits will help us resolve this fundamental conflict. We can
not see what we a priori have determined is not present, regardless
of sampling strategy.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
What type of exposure to use?whether block, column, square,

trench, or facies?was a question raised often in the 1980s. Excava
tion by facies (Stein 1985) or in blocks (e.g., Begier and Keatinge 1979)
facilitates total excavation of deposits, whereas other techniques are
aimed at sampling deposits, occupations, components, or sites.
Column sampling has both proponents (e.g., Bernstein 1987; Bowd
ler 1983; Campbell 1981; Meighan 1969:418; Quitmyer 1985) and
critics (e.g., Ham 1982:163; Spiess and Hedden 1983; Trinkley 1981;
Waselkov 1987). Faunal analysts find column sampling well suited
for subsistence studies. While column sampling is not designed to
recover examples of rare species or isolated deposits, it is useful for
sampling the matrix, particularly shell, and is advocated by many
researchers for this reason. (Greenwood [1961] discusses the prob
lems in comparing volume rather than weight samples.) Columns
should not be employed as the exclusive excavation unit type be
cause of the highly restricted view of the component contents they
provide. For the same reason they should not be the only source of
materials for the faunal analyst. When used as the exclusive source
of material in refuse that may have accumulated as basket dumps,
they must be employed at intervals that match the spatial scale of
the activities which created the deposit. I have instead used column
samples under 40 cm2 in size to provide an approximation of a single
death assemblage for shell seasonality studies and to quantify the
visual differences in a wall profile.
Campbell (1981), Peacock (1978), Waselkov (1987), and others have
advocated sampling strategies that are flexible and include several
different types of exposures, employed probabilistically and non
probabilistically. At DuWamish No. 1, in Seattle, Campbell began
excavation with coring. Coring was succeeded by dispersed 2 x 2 m
units that were subsequently enlarged to yield broad horizontal ex
posures connected by trenches. Since knowledge of component ex
tent and variability is learned as excavation progresses, yet is neces
sary in planning the sampling strategy, a flexible strategy is essential
to the success of the sampling program.
Sample size of matrix. The goal of site excavation should be to obtain

a representative sample of each and every occupational episode pres
ent (given that this is a destructive endeavor), so that subsequent
laboratory analyses can determine adequate subsamples. This should
be, minimally, a probabilistic sample. Determining what is a repre

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
256 Cheryl Ciaassen
Table 6.1.
Statistics for Subsistence Studies of Shell-Bearing Sites
% of Site Screen % Excavated % of Site

Sites Sampled Size Analyzed Analyzed
PACIFIC SITES
LAn202 4.30 1/8 0.17 0.0014
LAn229 0.86 1/8 100.00
Ven294 1/8 4.00
SBal42 0.19 1/4
SN?16 1/8 4.00
SNi 79 0.10 1/8 20.00 0.0210
Ltl Harbor 0.02 none
SLO 2 0.09 1/4 0.55 0.000002
SL0 51 0.21 1/4 0.53 0.000011
SL0585 1/4 0.29
Ala328 1/8 1.00
CCo30 upper 1/4 11.00
Mm 14 4.50 1/4 <4.50
Eddy Pt. 35.00 <35.00
Ivy Station 0.67 1/4 <0.67
Fish Rocks 0.88 1/4 <0.88
Ozette 20.00 <20.00
45Ghl5 0.45 1/4 <0.45
Glenrose 1/4 0.0078
Crescent Beach 1.45mm <13.00
FRESHWATER SITES
lja305 2.25 <2.25
47Cr310 20.00
sentative sample requires knowledge of the volume and variability

within the occupational debris (and access to that debris) and a deci
sion about the confidence level wanted for any estimate. Assessing
variability within an occupation requires exposures larger than cores
and a great deal of excavation before these decisions can be made.
In Table 6.1 I have compiled sample size information from many
North American shell-bearing site reports. Most archaeologists view
the site as the appropriate sampling universe yet give no indication

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
Table 6.1. Continued
% of Site Screen % Excavated % of Site

Sites Sampled Size Analyzed Analyzed
47Crl87 100.00
47Crl86 50.00 90.00 <50.00
Apple Cr 8.00 1/2 <8.00
15Bt5 20.00 none 0.00005 0.00001
15Bt 10 49.00 none <0.10 <49.00
Rocky Hollow 32.00 100.00 32.00
23Srl36 7.00 100.00 7.00
23Sr473 5.00 100.00 5.00
23Sr230 1.00 100.00 1.00
atlantic/gulf sites
Kidder Pt. ME 100.00 1/4
Turner Farm, ME 40.00
Wheeler's, MA 1.00 3mm <1.00
Greenwich, CT 17.00 6mm 4.00 0.67
Sungic, NY 1.60 0.80
Devils Walk, GA 2.00 1/16 2.00 0.05
Kings Bay, GA 1/16 2.60
8So26 FL 0.0005 100.00 0.0005
8Okl02FL <25.00 1/16
8SR85 Fl 0.33 fine 100.00 0.33
8SR143 Fl 0.66 fine 100.00 0.66
16CM61 1.71 float < 10.00 <0.10
McKinzie, TX <33.00 1/4
as to site area or volume; their work is generally not included in this

table unless the other figures are stated in the report. (Sorant and
Shenkel 1984 and Stein 1986 present formulas for calculating volumes
of sites.) Where the information was sufficient, I have indicated the
percent of the site sampled, the percent of excavated material ana
lyzed, and the percent of site analyzed. Screen sizes are also noted and
are one indication of the confidence level of these interpretations.
It is clear from Table 6.1 that interpretations of paleoenvironments
and past behaviors have been proffered on widely differing sample

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
258 Cheryl Ciaassen
sizes. For instance, the oldest marine shell-bearing midden site in

North America, SLO-2 in Diablo Canyon, California, with a 9,000
year record of marine exploitation, is known from a single column
(Greenwood 1972). A sample size of 0.1 may be adequate for analysis
if the percentage of the site excavated is large enough to determine
homogeneity in the deposit, but this is rarely the case. Based on
inadequate samples, generalizations about subsistence strategies in
components, bays, regions, and stages have been made.
A sampling strategy must be decided case by case, on the basis of
the research objective, site stratification, variability in contents, dis
tribution of features, origin of deposits, and so on. (Shell that has
been amassed for architectural purposes need not be sampled in the
same fashion as food debris. ) There is no specifiable amount of matrix
that can be deemed statistically adequate for worldwide use, nor can
there be a fixed size of sample useful worldwide, contrary to the
aims of Greenwood (1961), Bailey (1975), Bowdler (1983), and others,
and the applications of Peacock (1978), Buchanan (1985), and others.
Adequacy of sample is not achieved with absolute size but with the
proportion of the universe sampled (Barz 1977). The quest for stan
dard figures is further indication of normative thinking.
Subsample size. Sampling specific to subsistence studies was given

much attention in the 1980s. Appropriate screen sizes were the focus
of most of this attention (e.g., Bowdler 1983:139; Buchanan 1985:40;
Wing and Quitmyer 1985; Waselkov 1987), with all authors agreeing
that the smaller the screen size used, the more accurate the ecofact
quantification. The failure of most excavators to use Vs-inch mesh or
smaller has made comparison of ecofact data from the southeastern
United States difficult (Reitz and Quitmyer 1988), and this is un
doubtedly true in other regions as well.
Several authors have stressed the importance of multiple subsam
ples from each provenience slated for sampling. At Cnoc Coig, 4 sub
samples of 4 kg dry weight were removed from the unit walls by
natural strata and water-screened through 2 mm and 1mm mesh to
assess the variability within each unit (Peacock 1978). (These sub
samples came from outside the excavation unit but were still within
the random sampling units.) Multiple subsamples are a necessary
check on variation within a provenience.
In some subsistence studies sample adequacy has been determined
not with size or number of subsamples but with the number of spe

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
cies and the MNI. Wing and Brown (1979:120) set sample adequacy
at twenty-five species with an MNI of 200 for sites on the Caribbean
coastal plain. Subsistence samples from sites in the Kings Bay, Geor
gia, area, on the Atlantic coastal plain, were thought adequate when
70 percent of the potential fifty-three vertebrate and invertebrate
species were identified. This percentage was achieved with an MNI
of 1500 (Quitmyer 1985:34). Unfortunately, far fewer taxa are con
tained in many coastal North Carolina and Hudson River Valley, New
York, shell middens. Conversely, that many species and individuals
of bivalves alone can be retrieved in a single 3 m column of a Green
River, Kentucky, shell midden. What figures are appropriate in these
regions? Can we expect that the potential number of usable species
will always be an appropriate standard for judging adequacy?
Kent (1988:20-27) describes a complicated means of assessing
sample adequacy of oysters used in various analytical activities. One
is to measure all the shells from a culturally meaningful proveni
ence, determine the range of sizes within that set, and select a target
proportion of those size classes (he suggests 90 percent). One should
then randomly draw shells and measure them, stopping when 90 per
cent of the size classes are represented. After repeating the exercise
several times, the number of shells measured each time should be
averaged; that mean can serve as the number of shells used in other
analytical treatments for that provenience only. Similar studies need
to be undertaken for other classes of artifacts and ecofacts, such as
lithic and ceramic debris.
Customary sample sizes for growth line studies are the clearest
example of normative thinking in sampling. The growth control (ani
mals killed at known times whose growth is modeled) used by most
researchers is usually inadequate for the archaeological task, being
comprised of too few animals killed too infrequently in too few
years. Prior to 1986 no North American shell seasonality control had
animals killed at regular intervals throughout a calendar year (Table
6.2). While acknowledging that water temperature is a major stimu
lus to growth initiation and slowdown, researchers sampled only one
calendar year. In order to employ these controls in archaeological
work, a normative view of annual growth for the target species is the
only tenable one. One had to assume that conditions in the year of
collecting for the control were "normal" and that the growth re
sponse throughout the year was "normal." Fish and tooth growth
line controls suffer from the same assumptions.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
? ? ? ? ? ? ? ? 50 52 47 35 184 1985 6
? ? ? ? 45 ? ? 50 84 22 ? ? 201 1972 9
? ? ? ? ? ? 32 27 ? ? ? ? 59 1971 9
46 41 34 37 37 37 31 40 39 38 36 34 450 1987 6 43 33 37 33 50 48 50 52355230302938353142497451986

45 486 47 44 35 31 466 1987 6
Species J FMAMJJASONDN= Year Refer 37 36 30 27 35 33 29 38 ? ? ? 41 306 1988 6 37 33 38 34 42 47 53 41 47 ? ? ? 372 1988 6
34 34 23 40 36 39 ? ? ? ? ? ? 206 1984 6
A P.i. __ ____XXX? ? ? ? 53 71-72 4
12 109 81-82 3
L.EggNJ M.m. 41 45 40 36 39 36 35 37 37 37 35 44 462 1986 6
Damarisotta ME M.a. ? 10 10 10 10 10 10 10 10 10 10 10 110 1980 2 Narragansett RI M.m. ? 10 10 12 10 10 10 10 10 9 10 11 112 82-83 5
KingsBird ShoalsBay
St. Mary's R. C.v. _____ NC9
M.m. ? 13 ??
GA24 8 ? 8M.m.
19 ? 11 ? 11? 938
XX X X X? XXXXXX 69 81-82 8
1981 6 ? ? 11 28 <1988 7 various, TX R.c. ? ? ? ? ___25 ? ? ? ? 25 1969 9
PenobscotBay M.a. 6 ? ? 8 79 12 444 ? ? 54 82-83 1
Shark Inlet NJ M.m. ? 6? 6? 6?6?6? 6 36 1981 6
ll Seasonality Controls in North America
Table 6.2.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
9 66666666
10 11121314
19741984 1983 1984 1985 1986 19871988

9 639 87-88
1973
1975 <1968
74-75 83-84 78-79
80-81
482297705002914652139389368123252 7 3518
323131 10
51
31 55
1015393229 62
50 ? 34100
? 34 ? 50
50 12 353032 24 X
5
14 45 5040 332832
?5?
50 12 2032 21 20 ?X ? ?
50 50
503312 3031
75 5096 3531
14 ? 16 17 46 69 63 40 59 79 63 70 65 30
1955060 3531
10050 333130 14 X 5 10
M.c. = Mercenaria campechiensis
2005011 38312922
P.i. = Pect?n irradians
Sources: 1 Spiess and Hedden 1983; 2 Hancock 1982; 3 Hancock 1984; 4 Perlm M.a. = Mya arenaria
6 Claassen, first41report;
? 100
13 10
2432147 Kent 1988; 8 Quitmyer, Hale, and Jones 1985;
? 50
363332 18 X
R.c. R.c. R.c.

11 Tiffany 1968; 12 Wessen 1982; 13 Ham 1982; 14
M.c. D.v. P.S.
many many
M.m. = Mercenaria mercenaria

D.v. = Donax variabilis
R.c. = Rangia cuneata
Escambia Bay Alligator Harbor BoundaryShark

Bay Cove BC
Palmetto Bend
Tampa Bay
Alaqua Fl
Ozette WA Key:

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
262 Cheryl Claassen
Collecting M. mercenaria in North Carolina from 1980 to 1988

has resulted in live animals killed every calendar month throughout
the year, each month represented in at least four years, some months
in six years. A similar collecting regime was conducted in New Jer
sey (both projects directed by the author). Analysis by the author of
over 3,000 cross sections clearly indicates that a model of annual
shell growth based on only one or even two years of control speci
mens would be erroneous. With four to six years of data to examine,
no calendar month has a unique range of growth measures, and a
Scheffe test for differences between means for different measures of
growth each month could identify no more than two gross "seasons"
in either locale. Season 1 is September through June, and season 2 is
July through August.
Another characteristic feature of incremental growth studies is the
small number of individuals collected in a year (Table 6.2). Several
workers had monthly.sample sizes of as few as ten shells. Here again
is a normative assumption governing archaeological research. It is
assumed that the reaction of a population of shellfish or fish or deer
to an environmental stimulus is adequately reflected in the behavior
(growth) of an extremely small subsample of animals.
The larger the sample size, the greater the variability seen in
growth response, however. That variability then must be captured in
the growth measures, not ignored, averaged, or standardized. One
shell seasonality technique popular in northeastern U.S. studies
counts daily lines, assumes a zero date or growth initiation date,
adds the daily line count to that zero date, and gives a time of har
vest. A 106-day "window" during which growth starts in Narragan
sett Bay, Rhode Island, negates the usefulness of a growth initiation
date such as January 1 and means that the resulting archaeological
predictions could be off by three months, an entire season, in the
New England/New York area (Claassen 1990a).
Growth in November-killed M. mercenaria from North Carolina
was, over a five-year period, as little as 2 percent of the previous
year's growth and as much as 300 percent with most intervening
values represented as well. Again, generalizations such as shells will
achieve 20-30 percent of the previous year's growth by November or
early winter equals 0-30 percent of the previous year's growth, are
gross misrepresentations of shell growth. Larger sample sizes and
more inquisitive sampling programs can correct the normative er
rors associated with these controls.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
Sample sizes of archaeological shells for seasonality studies are

equally weak (Table 6.3). Analysts appear to believe that growth is so
uniform among shells, and human behavior so uniform in all dump
episodes, that the seasonality of a component and even a site can be
proffered on the basis of only a handful of shells. Even when sample
sizes are large?for instance, 1,162 shells from the DeWeese mound
of Kentucky?the shells may come from only a single column, fea
ture, or stratum. Unless one can demonstrate that all shells respond
identically to environmental stimuli, that seasonal environmental
stimuli occur at essentially the same time every year, and that a
shell component or site is the product of a single occupation within
one season, then seasonality studies based on incremental growth
structures must embrace a site-sampling strategy of large numbers
of observations from large numbers of archaeological units spread
across the component.
Maxwell ( 1989a, 1989b) has given the most attention to sampling
for archaeological shell seasonality work. He advocates cluster sam
ples from natural strata combined with stratified random samples.
The sampling provenience chosen should closely approximate a de
posit of shells harvested at one time. To combat normative tenden
cies, a minimum of two sets of shells should be collected from each
sampling provenience and analyzed separately.
Based on requirements for obtaining a normal distribution of sam
pling errors, forty or more usable shells should be derived from each
sample (Maxwell 1989a:8, 1989b). Depending on the fragmentation
and groundwater chemistry of the site, a considerably larger initial
sample size may be required to obtain forty usable shells. I have
found anywhere from 13 percent to 94 percent of the valves from site
samples to be unreadable.
Sampling is now the weakest aspect of growth-line seasonality
studies, whether for shells, bones, scales, teeth, or otoliths, impact
ing both controls and archaeological interpretations. Fortunately,
the problem is easily corrected if workers employ efficient section
ing techniques.
It is only with probabilistic sampling schemes that we can derive
a measure of confidence for the estimates and extrapolate results to
the unsampled remains for our sites. We cannot know the validity of
interpretations based on data derived nonprobabilistically until prob
abilistic data have been gathered and interpreted. Unfortunately,
present sampling schemes do not include methods that feasibly

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
264 Cheryl Claassen
Table 6.3.
Sample Sizes of North American Shell Seasonality Studies
Project n Ref Project n Ref
ATLANTIC
Fletcher FL 500 1 19Bn390MA 5 10
Palm Coast FL 500 1 19Bn410 MA 1 10
8 SR143 FL 3037 1 Shelter Is. NY 42 11
8 SR 85 5632 1 Henry Lloyd Manor NY 67 12
Zaremba FL 686 2 John Robinson NY 95 12
Cotton Site FL 1000 3 31 On 21 NC 29 13
Bluewater Bay FL 203 31 4 26 NC 110 13
On
St. Catherines Is GA 731 1 On 29 NC 46 13
St. Catherines Is <500 5 31 On 31 NC 89 13
Kenan Field GA 551 6 31 On 37 NC 29 13
Kings Bay GA 127 7 31 On 77 NC 87 13
Devils Walkingstick 79 731 On 82 NC 14 12
Cannon Pt. GA 30 7 31 On 136 NC 135 13
16 Cm 61 LA 438 8 31 On 152 NC 11 13
KidderPt.ME 117 9 31 On 191 NC 53 13
Sears Is. ME 13 9 31 On 196 NC 35 12
Nahanada ME 24 9 310nNNlNC 15 13
Taylor ME 38 9 310nNN2NC 46 13
Wheeler's MA 19 9 31CrNNlNC 77 13
Three Snakes 4 9 31NH256NC 100 12
Rosebush MA 4 9 Jordon Permit NC 120
Huntington MA 15 9 Stoney Brook NC 157 1
Nesson MA 49 9 RI 253A RI 23 9
Nantucket MA 48 9 RI 253B RI 12 9
19Bn274 MA 5 10 RI SU 205 RI 7 9
19Bn288.42 MA 12 10 North Shore RI 134 9
19Bn288.52 MA 1 10 Greenwich RI 226 9
19Bn288.22 MA 1 10 Lodge Alley SC 10 1
19Bn308.11MA 2 10 Miminls. SC 12 12
19Bn308.33 MA 3 10 Pinckney Is. SC 33 1
19Bn308.42 MA 24 10 Fish Haul SC 141 14
19Bn308.71 MA 2 10 38Ge238SC 171 1
19Bn323.21 MA 1 10 41Ch32TX 522 1
19Bn323.22 MA 13 10 41Ch46TX 846 1
19Bn323.23 MA 2 10 41Ch47 400 1
19Bn341 MA 27 10 41ChllO 1020 1

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
Table 6.3. Continued
Project n Ref Project n Ref
41 Ch 172 523 41 Jkll4TX 99

41Mgl9TX 50 41Jk 118 50
41 Mg 25 50 41 Jkl20TX 48
41Mg29TX 25 41 Jk 125 95
41 Jk 7 515 41 Jk 128 TX 50
41 Jk 41 TX 159 41 Jk 129 50
41 Jk 91 1006 41 Jkl47TX 124
41JkllOTX 25 41 Cl 35 TX 37
41 Jk 113 48 41C137TX 50
FRESHWATER
Ruckers GA 53 33 Ha 11 OH 53
Carlston Annis KY 76 33 Ha 58 OH 51
DeWeese KY 1162 Ervin TN 5
SUBI 677 NY 102 12 41 Ft 180 TX 13
PACIFIC
San Bias, Mexico 16
234 LAn 702 CA 364 25
Ora 64 CA 56 17 Ozette WA 9+ 26
Ora 82 CA 79 18 DgRrl BC 106 27
Ora 82 CA 177 19 DgRr 6 BC 98 28
Ora 82 CA 200 20 DhRr 8 BC 73 29
Ora 85 CA 36 18 DiSe 7 BC 219 30
Ora 99A CA 10 17 Buckley Bay BC 74 31
Orall9ACA 58 21 Tsable R. BC 569 31
Ora 168 CA 19 17 Yuquot BC 73 32
Ora 203 CA 106 22 GbTo31 BC 53 29
Deer Creek CA 1 23 KukakAK 143 33
LAn 264 CA 3 23 Chugachik AK 56 34
LAn 283 CA 695 24
Sources: 1 Claassen 1986; 2 Sigler-Eisenberg and Russo 1986; 3 Hale 1984; 4 Curren
1987; 5 O'Brien and Peter 1983; 6 Crook 1978; 7 Quitmyer, Hale, and Jones 1985;
8 Goodwin Assoc. 1986; 9 Claassen 1990b; 10 Hancock 1984; 11 Claassen 1990b;
12 Claassen, first report; 13 Claassen 1982; 14 Trinkley 1986; 16 Mountjoy and
Claassen 1989; 17 Koerper 1981; 18 Chace 1969; 19 Weide 1969; 20 Lyons 1978;
21 Drover 1974; 22 Howard 1977; 23 Tartaglia 1976; 24 Butler 1974; 25 Carter 1978;
26 Wessen 1982; 27 Ham 1982; 28 Ham 1976; 29 Ferguson 1975; 30 Monks 1977;
31 Keen 1979; 32 Clarke and Clarke 1980; 33 Clark 1977; 34 Lobdell 1980.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
266 Cheryl Claassen
permit the collection of independent probability samples from strata

lying under the uppermost one. This violation of a sampling require
ment may or may not be significant; we need simulations to explore
its impact.
Nonprobabilistic data are not to be discarded. They do provide in
formation and form the basis of impressions that are codified in re
ports. It is the interpretations, the generalizations, that should be
discarded, and the impressions reworded to acknowledge their non
probabilistic nature.
Experimentation with Shell and Shell Middens
Perhaps it is the prevalence of normative thinking about shell and

shell middens that is responsible for the noticeable lack of relevant
experiments. Since all shell is garbage, all shell deposits are secon
dary refuse, and most shells preserve indefinitely, few useful insights
could be gained from experimentation or even ethnoarchaeology. I
propose instead that these normative assumptions are false and that
many questions about these sites can be most fruitfully addressed by
using well-conceived experiments.
A prevailing attitude about shell in research since about 1940 is
that all or nearly all of the shells deposited will be preserved. Re
search by taphonomists, however, clearly contradicts such a view
(see Muckle 1985 for a partial bibliography of this work). Boring or
ganisms, solution, and abrasion of shell are the most active processes
in the formation of fossil shell deposits, and the latter two are the
most active in archaeological settings. With respect to abrasion, taph
onomists have found that the reworking of different sediment types,
varying shell architecture, different degrees of abrasion associated
with different mean grain diameters, and sorting of the abrasive
agent are the most relevant variables in preservation of shell (Dris
coll and Weltin 1973). Solution can result in marked shell and bone
weight loss before the appearance of the item changes and can totally
destroy shell and bone (see Waselkov 1987:148-49 for discussion).
Experiments and observations by archaeologists further em
phasize the vulnerability of shell. Coutts (1969) marked off eight
experimental plots with varying degrees of wind exposure and mea
sured the shells exposed in each plot. At intervals of 2, 4, 6, 8, 21, and
42 weeks, Coutts returned to the plots and measured the shells, the
angle of their movement, and their fragmentation. Areas with the

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
greatest degree of exposure to the wind had the most shell movement
and fragmentation.
Koike (1979:72), using the Feller formula and valve pairing, esti
mated that 70 percent, 58 percent, and 49 percent of the shells origi
nally deposited in three Jomon house pits had survived. (Nichol and
Wild [1984] discuss the Feller formula as well as others for predicting
original population size.) Coutts (1969) estimated the number of
valves and gastropods that were too crushed for counting in an Au
stralian deposit by using an average individual weight as denomina
tor and then comparing the number of whole shells found against
that estimated number. A similar ratio was calculated for Subninella
undulata, using opercula and weight of fragments. The two:one ratio
indicated that half of the shell material had been lost at the time of
lab analysis. The implication of these loss figures is that not even
shellfish meat can be adequately estimated in these types of sites.
Koike (1979:67) recommended that the distance between valves of
a pair at the time of excavation be used to estimate the amount of
disturbance affecting the deposit, and carried out such a project for all
shells in one Jomon pit. Results of experiments by Muckle (1985:62)
indicate that paired valves usually travel some distance apart upon
impacting the ground, //two valves can be convincingly paired, their
distance apart at the time of excavation may not be indicative of
deposit disturbance.
Archaeologist Robert Muckle (1985) conducted two sets of tapho
nomic experiments examining the orientation, articulation, and frag
mentation of shells tossed, poured, and trampled. In the discard ex
periments, Muckle (1985:56) found that single valves tended to land
concave side up, but that articulated valves landed hinge down or
sideways. From these data he concluded, with many qualifications,
that a deposit with valves predominantly concave side up had "prob
ably been subject to relatively little post-depositional disturbance."
Initial disposition resulted in no fragmentation or breakage into large
pieces. Highly fragmented deposits are probably the result of activi
ties subsequent to initial discard. Shell species [Protothaca staminea,
Mytilus edulis, and Venerupis jap?nica?all eastern Pacific) was an
important determinant of shell orientation, fragmentation, and dis
articulation,- height of discard, the number of valves discarded to
gether, and the age of the shells appeared to be insignificant variables.
Muckle's trampling experiments were designed to examine frag
mentation and vertical displacement of fragments. After 80, 160,

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
268 Cheryl Claassen
320, and 640 walking passes over Saxidomus giganteus, P. staminea,

and M. edulis, the debris was screened through nested sieves to mea
sure reduction. The species varied in particle size and in the rate of
reduction. Muckle (1985:76) observed that the smaller the screen
size used, the more accurate the reconstruction of species propor
tions. After 1,000 passes over shells resting on a loam substrate, 96
percent of the shell was found in the top 2 cm of loam, 3 percent 2-4
cm deep, and 1 percent 4-6 cm deep. Trampling, it was seen, can
rather quickly cause fragmentation as well as the downward move
ment of shell particles (and undoubtedly other objects).
Equally untapped are experiments and observations on large shell
bearing deposits. An experiment by Wing and Quitmyer (1986)
yielded surprising results?the addition of individuals to a shell pile.
In two experimental middens built on two islands in Charlotte Har
bor, Florida, 23 percent and 1 percent of the oysters deposited were
missing after six months (Wing and Quitmyer 1986:3). Scavengers
and possibly water removed 99 percent and 99.6 percent of the un
cooked small fish and 75 percent and 87 percent of the uncooked
larger fish. Cooked fish survived six months at a much higher rate.
Intrusions into these two middens were also noted upon excavation.
In Midden I, 554 sprouted seeds were removed, 159 small marine
mollusks had been added by agents unknown, and 5 unplanted fish
were found. This midden had been inundated for a short period dur
ing the experiment. Midden II had hundreds of small land snails but
no other obvious intrusions. These two middens indicate that shell
loss and midden contents vary with the nature of post-depositional
events. The greatest number of shells were lost from and added to
the midden that was inundated.
Hundreds of suggestive statements that reflect experimental work
with shells, other classes of material, and shell middens themselves
can be found in our literature and in conference conversations. By
and large, however, much of this material remains anecdotal. It is
urgent that we undertake experiments with controls and variables,
aimed at elucidating formation processes and formalize those quasi
experiments already under way. Experiments utilizing artificial mid
dens where various classes of artifacts and ecofacts can be observed
should be undertaken, and work in modern middens should be con
ducted. In coastal zones, canneries and seafood retail stores have as
sociated mounds and middens. Along major rivers in the Mississippi
watershed, musselers working for the Japanese pearl industry are

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
creating shell mounds. Herein lie invaluable data on mound col

lapse, plant colonization and succession, shell attrition, and mound
intrusions. The knowledge we gain through experimentation will
benefit both those working with shell-bearing sites and those work
ing in other sites.
Food and Feeding
Dietary reconstruction, a popular program of analytical

work in the 1980s, is built upon a series of normative assumptions.
It is perhaps here, more than anywhere else, that the effort at recon
struction is the most flawed by these assumptions. Normative as
sumptions underlie our conception of gathering as an activity, the
nutritional value of prey species, seasonality of collecting, democ
racy of consumption, and the reconstruction of paleohabitats. The
result is a pattern of erroneous reconstructions of component spe
cific foods and feeding, as well as regional and temporal errors. We
have managed to homogenize the human experience.
Gathering and Consumer Behavior
Some authors have attributed the different dietary import of shell

fish in archaeological situations to the biases inherent in the various
forms of data and different methods of analyses (e.g., Bailey 1975).
While methodological practices are certainly suspect in some cases,
there is a growing recognition among prehistorians that gathering
behavior varies, as does adaptation to aquatic resources (e.g., Bailey
1983; Meighan 1989; Rowley-Conwy 1983).
Little research has been conducted on the nature of gathering in
horticultural, dry or wet farming, or industrial societies. Archaeolo
gists tend to assume food gathering is insignificant in food-producing
societies and differs from that by hunter-gatherers only in percent of
dietary contribution. The dietary role of shellfish in prehistoric
societies has been interpreted as supplemental, based on ethno
graphic information from modern hunter-gatherers. I have suggested,
however, that for food producers, shellfish may serve as a seasonal
staple (Claassen 1986).
The horticultural Gwembe Tonga gather wild foods intensively
during one season each year, collecting a much broader range of taxa

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
270 Cheryl Claassen
than do their classic hunter-gatherer neighbors, the San (Scudder

1971). The role of gathering in agricultural systems in Mexico (Cabe
llero and Mapes 1985; Wilken 1970) and Greece (Forbes 1976) has
been said to differ from that of the San in the role the food plays in
the diet and in energy input/output terms. Deith (1988:116) sum
marized these differences in gathering for agriculturalists in Greece
as follows:
1. Gathered, wild foods are relishes, not staples
2. A different suite of plants is gathered, primarily garden weeds
3. The people do not go out deliberately to collect wild foods
4. The division of labor characteristic of hunter-gatherers is
absent.
The dense shell sites of Tierra del Fuego, Australia, Tasmania, Oron
say, Scotland, northern Spain, and the mid South of the United States
were probably created by non-food producers, while most of those of
the coastal southeastern United States, Mesoamerica, the Pacific Is
lands, and Japan were usually created by food producers. Large shell
bearing sites of the Aleutian Islands, the Northwest Coast, the Cali
fornia Channel Islands, and the Maritimes of eastern Canada were
left by maritime-focused societies. Even if the ecofacts in all of these
sites are human food debris, there is the great possibility that the
gathering activities differed in important ways, and that the food
played different roles. We must not allow our unexplored assump
tions about gathering to dictate prehistoric diets and behavior.
Collecting and eating shellfish and other intertidal resources are
usually viewed as catholic behaviors within a community (but see
Waselkov 1987 and Bowdler 1976). An assumption that all people eat
shellfish is particularly evident in the "California school," where
the number of band members was divided into the available calories
to derive length of occupation. While these mathematical computa
tions are no longer as popular as they once were (but see Akazawa
1988), the assumption that all members of a community would con
sume shellfish flesh still appears. McManamon (1984:392) has pro
posed that a change in women's time allocation in the summer
months to accommodate corn precluded summer shellflshing. Yet
this would be relevant as an explanation for the lack of summer
harvested shellfish only if we assume that all adult women shell
fished and all adult women subsequently hoed. Bowdler's (1976) and
Meehan's (1982) reviews of Australian ethnographic situations
clearly indicate that all the members of a community do not partici

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
pate equally in shellfishing or fishing or in consuming shellfish or

fish. In households where adult women do not care to collect, shell
fish are rarely consumed.
It is highly likely that in most societies, the habitual gatherer of
plants, the gatherer of intertidal resources, the fisher, or the hunter
will consume greater amounts of the prey over a lifetime than will
those not involved in that food procurement activity. Among the
Fish Creek Aborigines in Arnhem Land, male hunters ate a greater
proportion of animal foods than did nonhunters, and each individual
man ate all the fish he caught. Women consumed the fish they
caught while out or took it back to camp for other family members
(Bowdler 1976:251-52). Among the Bantu speakers of coastal South
Africa, men never consume shellfish (Bigalke 1973).
The assumption that procured food will always be returned to a
larger group of people for uniform consumption (e.g., Lightfoot 1985)
is simply not tenable in light of ethnographic information. Feeding
often probably did not follow the democratic practice of our own
society, and involved taboos, allergies, and taste preferences. Neither
is the assumption that all women did all female tasks or that all men
did all male tasks tenable. Long before there were craft, religious, or
political specialists, and since their development, there has been ac
commodation for individual aptitudes in tasks.
Nutritional Reconstructions
The generic clam is the epitome of normalized data. Several authors

have warned against extrapolating from one bivalve species to all
(Buchanan 1985:7; Lischka and Sheets 1980), yet such a practice con
tinues. One seldom-considered aspect of the nutritional value of
mollusks is the tremendous amount of intertaxonomic (Table 6.4)
and intrataxonomic (Table 6.5) variation in all nutrients. Six assays
of the bivalve Mercenaria, from the same collection point, taken six
months apart, indicated a range in calories of 54 to 95, in protein of
9.7 percent to 16.8 percent, in fat from 0.79 percent to 1.86 percent,
and in carbohydrates of 1.95 g to 3.31 g (Table 6.5). Interspecific vari
ation in molluskan nutritional values (Table 6.4) is even greater.
Once published nutritional information on many shellfish species is
gathered together (Tables 6.4 and 6.5), it is evident that caloric values
range from 48 to 234 calories per 100 g of wet flesh (Table 6.4), a
range that overlaps published caloric values for quail, turkey, drum,

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
Cal Protein Fats Carbo Pmg Cmg Kmg Ref
n 47.7% 6.2% 4 Proptera

Actinonaia carinata
Megalonaias gigantea 8-9%
58 7.8g 0.7g 4.5g alata
Aequipecten
alsogibbus
520 320 26 1
see15.9-18.5g
table -l-.3g
6.3 160-270
1 20-60 5
Buccinum Cardiumedule 48 11.0-13.2g

undatum 91 18.5-68.0g 1.4-1.9g 4 .3g 3.4g 4
Aequipecten irradians (24) 13.4-17.0g -3-.9g 1.4-1.9g 6
77 9.5g 0.8
Pleurocera canaliculatum 58
Aequipecten gibbus (4) 15.6-16.4g -5-.7g 6 Clinocardium nuttalli 79 13.5g 0.7g 4.7g 9
Nutritional Assessment of Shellfish
Choromytilus meridionalis 84 10
Cittarium pica 6
Viviparus georgianus 72 11.1 g 3
Amblema costata 8-9% 1
Batissa.viol?cea 48.9% 180 505Burnupena
7 papyracae 92 10
Argobuccinum argus 93 10
FRESHWATER SALTWATER
Table 6.4.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
Sources: 1 Parmalee and Klippel 1974; 2 Klippel and Morey 1986; 3 Cumbaa 1976; 4 Waselkov 1987; 5 Sidwell et al. 1973;
Mya arenaria [20] 5.5-11.7g .4-2.6g 110-206 17-73 5 Mytilus sp. 95 14.4g 2.2g 3.3g 315 9
Haliotis sp. 98 18.7g 0.5g 3.2g 4 Mya arenaria 89 9.7-15.6g 1.4-2.5g 1.7g 6
6 Sidwell et al. 1974; 7 Meehan 1982; 8 Goodwin 1979; 9 Erlandson 1988; 10 Buchanan 1985 (my conversions).
Mytilus edulis 87 8.9-17.2g 1.7-2.0g 2.9g 4 Ostrea edulis 59 8.6-13.1g 0.9-1.9g 5.9g 4
Ostreidae spp. (22) 54-92 5.0-14.3g -7-2.6g 2.3-6.5g 6
Ostrealurida 82 9.6g 2.5g 5.4g 9 Tivela stultorum 74 11.2g 1.4g 4.0g 607 9
Mopalia muscosa 234 22.0g 16.3g 0.0g 9
Venerupis semi decusata (5) 12.2-13.6g -7-.9g 6
Haliotis kamtschatkana 10.4-18.2g .3-.7g 6
Telescopium telescopium 0.1 g 144 500 7
Littorina littorea 74 15.3-18.Og 1.4g 4 Note: (n) = number of bodies used.

Volegalea wardiana O.lg 158 680 7
Nerita peloranta 42.7% l.Og 8
Donax variabilis 54
Haliotis midae 74 10 Thtfis squamosa 97 10 Turbo cidaris 90 10

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
274 Cheryl Claassen
Table 6.5.
Variation in Nutritional Value Within a Species
Kcal Protein Fat Carbo Ref
Crassostrea virginica 66 12.00% 2.50% 6.50%

Crassostrea virginica (20)* 6.7-10.30% .75-1.90%
Crassostrea virginica (21)* 4.5-7.90% .56-2.00%
Crassostrea virginica (24) 5.6-10.00% .70-2.40% 1.9-4.7%
Mercenaria mercenaria
(collected in February)1 54 9.73% .79% 1.95%
83 15.30% 1.09% 3.03%
90 16.30% 1.37% 3.04%
(collected in October) 87 14.90% 1.59% 3.31%
74 14.60% 1.78% 3.18%
95 16.80% 1.86% 2.65%
Mercenaria mercenaria* 3.2-6.20% .10-.42%
Neocyrena ordinaria 10.4g 2.7g 0.9g
8.2g 1.7g 0.2g
7.2g 1.5g O.lg
7.7g 2.2g 0.2g
6.2g 3.4g 0.5g
Megalonaias gigantea 13
Oct, Athens AL2 5.90% .35%
5.20% .19%
6.00% .50%
Nov, RomeTN 8.30% 1.70%
8.10% 0.70%
8.50% 1.20%
Jan, New Johnsonville TN 5.20% .33%
5.60% .31%
6.10% .26%
Note: 100 g units or * unknown.

1 Each entry is an average of three animals.
2 Each entry is an average ?f eight animals.
Sources: 1 Parmalee and Klippel 1974; 3 Cumbaa 1976; 4 Waselkov 1987; 5 Sidwell
et al. 1973; 6 Sidwell et al. 1974; 13 Post 1982.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
and catfish (Parmalee and Klippel 1974), deer (Parmalee and Klippel
1974; Waselkov 1987), rabbit (Erlandson 1988:104,- Parmalee and
Klippel 1974; Waselkov 1987), and croaker, shark, and seal (Erlandson
1988:104). Protein ranges from 37 percent to 68 percent of the tissue,
or 0.1 g to 68 g. The carbohydrates in mollusks range from 0 g to 8.9 g.
The most common bivalves in eastern U.S. middens?C. virginica,
M. mercenaria, and Mya arenaria?exceed the turkey, quail, rabbit,
and shark figures in carbohydrates, and can rival drum, catfish, and
deer. Nearly all western Atlantic shellfish species listed in Table 6.4
can rival the protein value of most fishes.
One can argue either that shellfish are poor sources of nutrition or
that they are adequate sources merely by choosing an appropriate
species to illustrate the point. Arguments that rely on generic figures
or are based on a single species are on shaky ground. Surely it is time
to put to rest the generic clam. Specific figures, rather than average
figures, will greatly enhance subsistence discussions.
Growth Line Seasonality Studies
Basic assumptions about shell growth influence the prediction of

harvest season. The normative assumptions that all shells respond
to growth stimuli identically or essentially so, and that the timing
of these stimuli is predictable each year, have resulted in the practice
of assigning harvest time to individual shells. With large growth con
trols it is clear that neither assumption is true, but that there is
marked growth variation in any population on any day in the year.
But the range and mean of that variation, as well as the elements
varying, themselves vary in a patterned way throughout the year.
Because variation in growth is expected and accounted for in those
studies where the researcher recognizes growth variation, only sets
of archaeological shells are interpreted and only one harvest time is
predicted for any one set. Because variation is present in growth rec
ords in a population, researchers who assign death times to indi
vidual shells always find at least two seasons. An example will illus
trate this point.
As a preface to the following paragraph, some terms need to be
symbolized: Ly = the final year of life; Ly - 1, the last full year of life,
Ly - 2, the next-to-last full year of life. A practice of decreasing popu
larity in shell seasonality work is to measure the amount of shell
added in the final year of life and divide that figure by the amount of

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
276 Cheryl Claassen
growth evident in the last full year of life, Ly/Ly - 1, or to average the
growth in several years prior to the final one?Ly/((Ly - 1 ) + (Ly - 2)1
2). (The value chosen as the denominator for Ly is the single most
influential factor in the outcome [Claassen 1990b].) In an exercise
comparing the results from five shells using Ly/Ly - 1, estimates of
104 percent, 81 percent, 65 percent, 36 percent, and 21 percent of the
expected growth were derived. The researcher who views shell
growth in a normative manner would interpret the death time of the
first shell as the end of the growing season, and that of the fifth shell
as the beginning of the growing season, perhaps September and April,
respectively. On the other hand, the researcher who has modeled the
variation found in each month would ask, What one month of the
year would have this range of growth percents? For M. mercenaria in
either New Jersey or North Carolina water, all twelve months have
ranges of growth percents this large. In fact, all five shells were har
vested the same June day from the same 1 m2 area. With larger sam
ple sizes in the control and in the archaeological set, other criteria
can be relied upon to pinpoint the harvest time. (The tremendous
range in growth values has caused several researchers in the south
eastern United States, including the author, to abandon measure
ments of growth and to rely on other growth criteria for assigning
sets of shells a death time.)
With the possibility that significant quantities of shell and per
haps other edible matter do not represent human food debris, we
must modify our practice of assuming all shellfish, crab, lobster, fish,
shrimp, and charred plant remains (several ethnographers record the
intentional burning of large shell deposits) are appropriately in
cluded in human dietary reconstructions. The nutritive value of
these foods at any given site may well be irrelevant to their use in
the past.
Habitat Reconstruction
The simplest technique for environmental reconstruction and the

one most frequently used in archaeological shell midden reports is
species ecology. Optimal habitat characterizations are usually pre
sented. Often glossed over by those who employ ecological profiles
of molluskan species is the fact that they have tolerance ranges for
many ecological factors. Furthermore, there are differing environ

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
mental parameters for maintaining life, for reproducing, and for lar
val recruitment. Many discussions of quahog [Mercenaria merce
naria) and eastern oyster [Crassostrea virginica) seem to view these
two species as having mutually exclusive environmental conditions,
but in fact their tolerance ranges overlap. The use of optimal habitat
characterizations is also noticeable in studies of freshwater mol
lusks. Generally speaking, the more stream habitat studies con
sulted, the greater the recorded diversity in individual response to
water depth, substrate, and current speed. For many other species,
ecological requirements are not known.
Malacologists often cite shell shape and/or size as indicative of
environmental particulars such as size of river, temperature of water,
or speed of current. Matteson (1960) identified dwarfing in Early Ar
chaic collections of the Illinois River Valley and attributed the cause
either to persistently colder temperatures or to a sustained lack of
sufficient food. In an earlier article ( 1955) he had attributed dwarfing
to slow-moving water. Klippel et al. (1978) found dwarfing in archae
ological Amblema plicata from Middle Archaic levels of Modoc
Rocksheiter and attributed the cause to slow-moving water. Jim
Chatters (personal communication, January 1986) found that growth
rates in Margaritifera margaritifera from the Columbia River were
30 percent to 100 percent higher between 6,000 and 8,000 years ago,
and attributed that phenomenon to variation in water temperature.
Roscoe (1967:6) summarized shell shape implications for M. mar
garitifera, indicating that controversy rather than agreement sur
rounds the correlation of shell shape with environmental factors. A
more recent review of the literature is far more positive about the
potential of shape characteristics to correlate significantly with
paleoecology (Tevesz and Carter 1980). Tevesz and Carter identified
two groups of papers published between 1948 and 1980: (1) papers
correlating shell form and habitat (e.g., lake, small stream, etc.) and
(2) papers correlating specific features of shell morphology, such as
thickness or arc, to particular aspects of the environment, such as
water alkalinity and current rate. They were able to reaffirm several
regularities culled from the literature several decades earlier. How
ever, they were quick to point out that sexual dimorphism, ontogeny,
and growth compensation confuse the correlations. Not until sexual
maturity is achieved do many environmental effects on growth ap
pear. Furthermore, changes in one shell dimension are often accom

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
278 Cheryl Claassen
panied by changes in others. Morphological covariance increases the

difficulty of identifying causally related correlations between form
and environment (Tevesz and Carter 1980:304).
The chemical constituents of shell have been used to ascertain the
paleoenvironment in the vicinity of archaeological sites (e.g., Hill
1975; Lee and Wilson 1969). The primary focuses of these studies
have been calcium, magnesium, and strontium, and they have been
aimed principally at retrodicting stream flows at the time of site
occupation. Assumptions necessary for this work are (1) that the
concentration of the various elements decreases with increasing
stream discharge and (2) that naiads deposit the alkaline earth metals
in their shell in proportion to their prevalence in the water.
Changes in strontium:calcium ratios accompany changes in dis
charge rates in some streams, while other streams show little or no
change in elemental ratios with a discharge rate change. Durum and
Haffty (1961) observed that in some streams the ratio increases with
a decrease in discharge, while in others the opposite is true. Conse
quently, then, specific stream reactions must be discerned before the
paleohydrology of a specific stream can be adduced.
Hill (1975), on the other hand, citing assumption (1), concluded
that the higher strontium level and the greater strontium range in
Horizon VIII shells from the Kost er site (Illinois) indicated a low
water stage of the Illinois River and a significant increase in dis
charge by Horizon IV times. Hill's assumption of decreasing stron
tium with decreasing water flow was in fact contradicted by his own
assay of seasonal strontium levels in the Illinois River, which indi
cated that its levels remained constant regardless of flow (Hill
1975:122).
There have been numerous studies of the factors that influence
the relationship between shell composition and water composition
(see particularly Rosenberg 1980 for marine mollusks). Many authors
report observations on the correlations between strontium, mag
nesium (the two most thoroughly studied), and calcium, and salinity,
water temperature, age of shell, and so on, often contradicting one
another (e.g., Harriss and Pilkey 1966; Rosenberg 1990; Swann et al.
1984; Turekian and Armstrong 1960).
The reader should be wary of an uncritical adoption of these chem
ical/environmental relationships. About the confusion Rosenberg
(1980:133-35) said:

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
The environmental and evolutionary significance of the chemi

cal composition of bivalve shells is poorly understood; not one
single bivalve species has been completely or adequately de
scribed chemically. There is evidence that the Ca, Sr, Mg, and
organic matter concentrations within the shell change through
out ontogeny. The shell may be composed of trace amounts of
sulfates and hydroxides, as well as metal carbonates other than
CaC03, complicating predictions of elemental distribution
based on stoichiometric considerations alone. The amino acid
concentration varies from layer to layer. Our knowledge, how
ever, is fragmentary, and detailed maps of these various shell
components are lacking. Distributions along dorsal-ventral and
anterior-posterior shell axes are poorly defined. Variations from
specimen to specimen, from population to population, and from
species to species have not been adequately quantified.
Nor has the confusion abated significantly since 1980. In a more

recent paper Rosenberg (1990:1) observed that no chemical distribu
tion maps have yet appeared for any shell species, and that actual
concentration levels of trace elements often differ from predictions,
producing " large error bias, or uncertainties, in correlations between
environmental parameters and skeletal composition." Archaeologi
cal studies of environmental change based on shell species and
species ratios should proceed cautiously in generating or weighing
data about these relationships. There is little doubt that long-term
and even short-term environmental changes in variables important
to shell growth and human collection affect species and species
ratios in detectable ways, but these relationships are still being
explored in the 1990s.
Ethnoarchaeology
It is to ethnoarchaeology that we can turn for information on con
sumer behavior, gathering behavior, systemic types of shell-bearing
deposits, and many other theoretical needs. The reason why we have
not done so may be the normative assumptions that pervade research
centered on shell-bearing sites: there is so much uniformity in sys
temic context that little could be learned through observation. The
handful of projects that have been conducted, however, invite the

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
280 Cheryl Claassen
reader to question much of what has been assumed in archaeological

reconstructions. I will briefly review three projects that I think chal
lenge conventional archaeological thinking about one or more of the
following issues: the dietary role of shellfish, consumer behavior,
the gender of shellfishers, and predictions about optimal foraging.
The now classic study by Betty Meehan, Shell Bed to Shell Midden
(1982), has yet to be matched for information content in a single
publication. The frequency of its citation underscores the poverty of
similar studies for archaeologists. Like all classic works of ethnoar
chaeology, it served as a cautionary tale to those who would depict
shellfishing as drudgery, its nutritional value as negligible, and its
participants as desperate. This report, more than any piece of archae
ological detective work, has contradicted the voices from the why
would-anyone-have-ever-started-shellfishing school.
It should be of interest to optimal foraging modelers that the An
barra women intentionally collected from beds where the shellfish
were smaller and less plentiful than at neighboring beds (which they
also harvested); that on 91 of 194 days (47 percent) shellfish were
gathered, only a single species was collected, although dozens of
other species that they eat were on the same beds; and that one
species which is exposed twice daily nearly every day of the lunar
cycle represented only 7 percent of all shellfish collected in the year
of observation, being ignored in preference to better-tasting species
exposed on only some days of the lunar cycle. Thus, we see that
among the Anbarra, the activity of shellfishing is far from the OFT
concept of efficient if energy is the currency,- some women even
walked the round trip and collected nothing although their compan
ions gathered for several hours.
From Meehan's study we learned that not all Gidjingali women
shellfish: some women like to and some do not, and levels of indi
vidual and household consumption vary with these attitudes. The
concept of a dinnertime camp for shellfish consumption was vividly
illustrated and subsequently became commonplace in our literature.
Basic relationships between tides, lunar cycles, species selection,
harvest size, loads, and dietary roles were illustrated, as was the im
pact of sickness and pregnancy on gathering. Meehan provided our
first example of women being honored with status and title for their
gathering activities in a hunting/gathering society.
The most extensive project undertaken to date to understand the

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
human adaptation to and utilization of shellfish is that on the Trans

kei coast of South Africa. For two decades archaeologists, anthropol
ogists, and ecologists have studied the past and present patterns of ex
ploiting the intertidal molluskan resources of this 260 km shoreline.
These papers present ethnology (Bigalke 1973; Hockey and Bosman
1986; Mills 1985; Siegfried et al. 1985), ecological studies (Hockey et
al. 1988; Voigt 1975), nutritional studies (Bigalke 1973), and excava
tions in modern middens (Voigt 1975).
The distance that people are willing to travel to collect shellfish is
no greater than 7 mi. Women, girls, and very young boys are the only
ones to consume shellfish and are the primary collectors, but men
have been observed collecting (Bigalke 1973). The bulk of bivalves
collected is taken home in the shell, while chitons are shelled at the
water. Shellfish flesh is never mixed with other foods, but the liquor
is mixed with maize. Newly circumcised boys and newly married
women are to refrain from consuming oysters or crayfish, because
they will enhance sexual appetite. Shells are discarded in four ways:
onto one heap removed from the main living area, into an eroded
area, putting each meal's refuse in its own heap, or scattering shells
over the garden (Bigalke 1973).
After six months of monitoring the shellfishing of 230 collectors,
Tregoning and Van der Eist (n.d.; in Buchanan 1985:213) learned that
the average harvest was 4 kg to 6 kg of limpets and mussels. Each
collector could provide the entire community with enough shellfish
meat by collecting three times her own daily caloric need, requiring
her to cover about 12 m2 to collect 9,000 kilojoules of limpets or
1 m2 for mussels. Shellfish flesh is a major meat complement to the
agropastoral-based diet of these Africans (Bigalke 1973).
In 1986, during spring tides, Hockey et al. (1988) performed a two
day airborne survey of the Transkei coast, recording people in the
intertidal zone, their activities, shore type/geology, house density,
land use, and livestock. They counted 1,447 women older than 13
years, 84 men, and 257 children shellfishing in the intertidal zone.
The estimated annual removal of mussels, chitons, and other shell
fish from the 99.2 km of rocky shore south of Port St. John was
552,835 kg, and that from the 66.4 km of rocky strip north of the city
was 2,183 kg, where shellfish contributed little to the diet of coastal
peoples. In the southern area, the harvest represented over 5.09 mil
lion kilojoules and 159.9 kg of protein per year. This intense shell

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
282 Cheryl Claassen
fishing activity is changing the species configuration of the intertidal

communities while improving nutrition, as is evident in the lower
incidence of kwashiorkor close to the coast.
Voigt (1975) excavated the shell pile associated with one house on
the coast. The shells it contained represented the meals of one
woman for forty-six years and of her three daughters, who collected
on four to six days per month. At the time of excavation, this midden
was 24.28 m2 in size, was heavily overgrown, and had sheep, cattle,
and human paths across it. All loose and obviously recent shells were
collected as Level 1. Three squares were excavated to bedrock and
revealed five distinct stratigraphie levels. The species proportions
contained in Level 1 were strikingly different from the proportions
recorded in the living molluskan communities. (Level 1 was the only
level analyzed at the time of publication.)
For five weeks in May-June 1988, Matt Wilkerson and I observed
fishing and shellfishing on San Salvador Island in the Bahamas.
Women, children, and men were observed or interviewed about shell
fishing on the rocky intertidal zones around the island. Gastropods
and chitons were collected from this zone and conch were gotten
offshore. Gastropods were removed whole from the shore to be boiled
open and the shells discarded whole at home; chitons were shelled
at once on the rocks. Hermit crabs frequented the home middens
and carried off whole top shells [Cittarium pica). Home middens
were also subject to goat, chicken, and dog traffic.
Completely unexpected was the predominance of men shellfishing
on San Salvador and the important role of shellfish, both flesh and
shells, as fish bait. Conch slop, chitons, bleeding teeth [Nerita sp.),
top shells, land crabs, and hermit crabs were all used when angling for
bait fish and for table fish. Conch, top shells, and land crabs were used
in fish traps. One long-time favorite location to gather and smash
shells for bait is atop a prehistoric site where pottery can be found,
now mixed with modern, and possibly prehistoric, shell debris.
Both the flesh and the shell of conch were used as bait. Fish traps
were baited with conch feet and with their shells, said to add scent.
After a trap was positioned underwater, several more whole conch
shells and several dead land crabs were tossed into the water in the
vicinity. Then one newly emptied conch shell was broken into sev
eral fragments with a hammer, because the sound would attract fish,
as would the fresh scent. On one expedition I joined, thirty-four fresh

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
conchs and seventeen shells were expended in baiting four traps, as

was a no. 10 tin can full of dead land crabs.
The contents of other fish pots observed on land indicated that the
use of conch meat and shell as bait was a standard practice on the
island. At one boat landing, five fish traps awaited cleaning and use.
They contained eleven top shell shells, twelve conch shells, eleven
triggerfish skulls, four triggerfish postcranial skeletons, one boxfish,
eleven crab carapaces, and one Thais shell. It is easy to imagine a
sizable shell midden containing numerous shells and fish bones that
accumulate as fish traps are emptied of their spent bait, yet none of
the contents of that midden would represent human food debris.
Each of these studies is welcome for the insights provided, yet is
frustrating for the information not presented. Meehan has little to
say about shells once they are emptied of their meat. Voigt does not
explain the origins of the stratification in the midden. Both Meehan
and Hockey et al. mention fishing but have, perhaps, erroneously
separated this activity from shellfishing in light of the fish bait infor
mation from the Bahamas. The Bahamian project lacks the longitu
dinal data so valuable in Meehan's and Tregoning and Van der Elst's
studies and dietary information in general.
Ethnoarchaeology projects that have addressed shellfishing and
shell middens have looked at their role in both agricultural and hunt
ing-gathering communities. Subsistence base is only one of many
variables suspected to influence the role of shellfish in prehistoric
societies and, consequently, the types of shell-bearing deposits. Fu
ture ethnoarchaeological projects should be planned so as to explore
other variables, such as duration of habitation, length of harvesting
season, and economic use. Even with the variable "subsistence
base," we lack comparable studies necessary to make cross-cultural
comparisons.
Conclusions
Fully half of the research process is asking appropriate

questions. Far too often in shell "midden" research, basic correla
tions between human behavior and shell or site formation are as
sumed when they should be hypothesized. The research enterprise is
relegated to quantitatively elaborating on the assumptions. Norma

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
284 Cheryl Claassen
tive concepts, then, actually impede the research process. They im

part a false sense of knowledge upon which successive interpreta
tions are built.
Built on a weak foundation of normative thought as many project
conclusions are, it is no surprise that false issues have developed. Per
ceptions of changing ratios of shell species may be the result of inade
quate site samples, as has been argued for an Australian situation.
Mackay and White (1987) point out that at many of the sites used to
support the thesis, samples of less than 1 percent of the total site
were taken, with no attempt to determine intramidden variability.
Lightfoot and Cerrato (1989) have compared the shell seasonality
results from twenty-one northeastern United States marine sites
with those from eighty-four marine sites in the Southeast (sum
marized in Claassen 1986). Beginning with the observation that pre
historic occupants of the Southeast shellfished seasonally but North
easterners shellfished throughout the year, they examined several
possible explanations for the regional differences. Yet all but one of
the prehistoric marine shell seasonality studies performed in the
Northeast have been based on the normative assumption that each
and every shell grows at a uniform rate each and every year. (This
assumption is rare in the Southeastern studies.) Extremely small
sample sizes and weak controls accompany these studies (Claassen
1990a). Modern normative thinking, not prehistoric human behavior,
can, at this time, be held responsible for the appearance of unseasonal
shellfishing in the Northeast.
The tyranny of normative thinking is also evident in the history of
explanations for the Shell Mound Archaic of Kentucky, Tennessee,
and Alabama. Based primarily on the normative assumption that all
shell debris equals human food debris, large mounds of shells were
interpreted as year-round base camps, even though no house remains
have been found (one to four small clay floors per site are known) and
burials occur at a density of 1.2 bodies per cubic meter (Patty Jo Wat
son, personal communication, April 1988). The Shell Mound Archaic,
as this manifestation is known, appears to have ended about 3,500
years ago, and explanations have focused on the mussel resource.
But were these shells cast-off food debris? I think not. Several ex
cavators have noted the high incidence of paired valves in these
mounds. Village activities are not conducive to preserving paired
valves. Seasonal indicators point to fall, not year-round, occupation.
Mounds served as burial environments throughout the subsequent

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
3,000 years of eastern United States prehistory. I have proposed

(Claassen 1991) that these shell mounds were not villages and the
shell not food debris, but were intentionally constructed burial
mounds with the shell, rather than the meat inside, being the object
of collection. I have also incorporated some of the known symbolism
for shell (Claassen 1991).
Riser's (1987) proposal of middens of fish bait along the northern
Gulf of Mexico coast also challenges the shell-as-food-debris con
cept. I question what it is we really do know about the roles of
shellfish and other ecofacts in prehistoric diets. Specifically, I ques
tion the thousands of "facts" on record that have been generated
through inadequate sample sizes, through inadequate screen sizes,
through incorrect assumptions about biological growth, through our
assumptions about the roles of shellfish in prehistoric societies and
their food value, and through our false sense of security when using
"scientific techniques" such as incremental growth studies and
chemical-habitat correlates. There are undoubtedly some truths con
tained in the impressions gained through normative research, but
just what they are will take some sorting out.
While this chapter is primarily about doing science better, science
done better is not the solution to all the issues raised here. Some
pursuits, such as dietary reconstruction, simply cannot proceed
without normative assumptions and cannot, therefore, escape their
weaknesses. Here, I echo the rejection of this indirect means of as
sessing diet voiced by many others: evidence of collection and prepa
ration does not equal consumption (Waselkov 1987:166); this effort
is subject to grave difficulties regardless of the state of preservation
(Begier and Keatinge 1979:223); there is too much variation in con
sumer and consumed (Buchanan 1985:52); there are too many vari
ables that cannot be confidently solved (Bowdler 1976:255). Further
more, since harvest time cannot be induced from individual shells,
if shell dumps do not retain their integrity in a shell deposit, shell
growth line seasonality studies will be pointless, regardless of how
good the controls are. The "demonstrated place for floral and faunal
analyses [is] in reconstructing the landscape and in understanding
some facets of human manipulation of the landscape" (Bailey
1983:3).
Nor can we, at this point in our development of test implications,
even know the systemic role of many of the ecofacts in these sites,
contrary to the confidence Bailey and Parkington ( 1988:4) echo about

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
286 Cheryl Claassen
food debris. We need simultaneously (1) to engage in more ethnoar

chaeology and experimentation while (2) dismantling normative ar
guments and (3) excavating shell-bearing sites facies by facies, de
posit by deposit, with probabilistic sampling.
While I have advocated a greater need for middle-range theory for
shell-bearing sites and prehistoric cultures, middle-range theory has
contributed to the formation and life span of normative concepts.
What few experiments and observations have been published are
used as analogies for coastal cultures in all latitudes, for many types
of subsistence, at many different time periods?in short, they have
become normative concepts. As Meighan (1989:21) has put it, "The
general trap of assuming relationships among shell-midden dwellers
because of the apparent similarity in shell-midden deposits has ex
tended to making inappropriate ethnographic analogies."
The more ethnoarchaeological, experimental, and excavational
research we do, the more variation we will uncover and the more
variables we will recognize.
By controlling for some of the variables, for example by compar

ing patterns of behaviour within a broadly similar environmen
tal and ecological framework, we may be better able to under
stand the variability of human response, and hence better [able]
to understand each case study in its historical and ecological
context. (Bailey and Parkington 1988:5)
Variation is not an obstacle to comparisons, to hypotheses, or to ex
planations. Capturing variation is the key to understanding the past.
There is a place for better samples, finer stratigraphie resolution,
better-understood sediments, and more accurate dating methods in
the archaeology of shell-bearing sites. That place is in understanding
the archaeological record, both its shortcomings and its strengths,
and in understanding the validity of the past we archaeologists have
created.
Acknowledgments
Thanks to Peter White, Fernanda Falabella, David Max
well, Arthur Spiess, Clement Meighan, Julie Stein, and Mike Schiffer
for comments on an earlier draft of this paper. The writings of Geoff
Bailey and Clement Meighan were particularly stimulating; the

174.70.7fff:ffff:ffff:ffff on Thu, 01 Jan 1976 12:34:56 UTC
references in Waselkov (1987) and Buchanan (1985) are much ap

preciated. Grants from Appalachian State University and from the
National Science Foundation supported the original work reported.
REFERENCES
Akazawa, T.
1988 Variability in the Types of Fishing Adaptation of the Later Jomon
Hunter-gatherers c. 2500 to 300 b.c. In The Archaeology of
Prehistoric Coastlines, edited by G. Bailey and J. Parkington,
pp. 78-92. New York: Cambridge University Press.
Ambrose, W. R.
1967 Archaeology and Shell Middens. Archaeology and Physical
Anthropology in Oceania 2:169-87.
Aten, L.
1981 Determining Seasonality of Rangia cuneata from Gulf Coast
Shell Middens. Texas Archaeological Society Bulletin 52:179
200.
Bailey, G. N.
1975 The Role of Molluscs in Coastal Economies: The Results of
Midden Analysis in Australia. Journal of Archaeological Science
2:45-62.
1983 Introduction. In Hunter-gatherers Economy in Prehistory: A
European Example, edited by G. N. Bailey, pp. 1-6. New York:
Cambridge University Press.
Bailey, G. N., and J. Parkington
1988 The Archaeology of Prehistoric Coastlines: An Introduction. In
The Archaeology of Prehistoric Coastlines, edited by G. Bailey
and J. Parkington, pp. 1-10. New York: Cambridge University
Press.
Barber, R.
1982 The Wheeler's Site: A Specialized Shellfish Processing Station on
the Merrimack River. Peabody Museum Monograph, 7.
Cambridge, Mass.
Barz, R. K.
1977 Some Theoretical and Practical Aspects of Midden Sampling as
Applied to a Site at St. George's Basin, Jervis Bay, ACT. B.A.
(honours) thesis, Department of Anthropology, Australian
National University, Canberra.
Beaton, J.M.
1985 Evidence for a Coastal Occupation Time-lag at Princess
Charlotte Bay (North Queensland) and Implications for Coastal

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
288 Cheryl Claassen
Colonization and Population Growth Theories for Aboriginal

Australia. Archaeology in Oceania 20:1-20.
Begier, E. B., and R. W. Keatinge
1979 Theoretical Goals and Methodological Realities: Problems in the
Reconstruction of Prehistoric Subsistence Economies. World
Archaeology 11:208-26.
Bernstein, D.
1987 Prehistoric Subsistence at Greenwich Cove, Rhode Island. Ph.D.
diss., Department of Anthropology, SUNY, Binghamton, New
York. Ann Arbor: University Microfilms.
Bigalke, E. H.
1973 The Exploitation of Shellfish by Coastal Tribesmen of the
Transkei. Annals Cape Province Museum (Natural History).
9:159-75.
Bowdler, S.
1976 Hook, Line, and Duly Bag: An Interpretation of an Australian
Coastal Shell Midden. Mankind 10:248-58.
1983 Sieving Seashells: Midden Analysis in Australian Archaeology. In
Australian Field Archaeology: A Guide to Techniques, edited by
G. Connah, pp. 135?43. Canberra: Australian Institute of
Aboriginal Studies.
Brennan, L. A.
1977 The Midden Is the Message. Archaeology of Eastern North
America 5:122-37.
Buchanan, W. F.
1985 Sea Shells Ashore. Master's thesis, Department of Archaeology,
University of Cape Town.
Butler, W.
1974 Appendix A: Seasonality from Tivela stultorum for LAn-283.
Pacific Coast Archaeological Society Quarterly 10:81-83.
Caballero, J., and C. Mapes
1985 Gathering and Subsistence Patterns Among the P'urhepecha
Indians of Mexico. Journal of Ethnobiology 5(l):31-4-7.
Campbell, S.
1981 The Duwamish No. 1 Site, a Lower Puget Sound Shell Midden.
University of Washington, Office of Public Archaeology
Research Report, 1. Seattle.
Carter, C.
1978 Seasonality Analysis of CA-LAn-702. Pacific Coast
Archaeological Society Quarterly 14(2):52-56.
Chace, P.
1969 Biological Archaeology of Some Coastal Middens, Orange Ct.
CA. Pacific Coast Archaeological Society Quarterly 5(2):65-77.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
Claassen, C.
1982 Shellfishing Patterns: An Analytical Study of Prehistoric Shell
from North Carolina Coastal Middens. Ph.D. diss., Department
of Anthropology, Harvard. Ann Arbor: University Microfilms.
1986 Shellfishing Seasons in the Prehistoric Southeastern United
States. American Antiquity 51:21-37.
1990a The Role of Technique in Shell Seasonality Studies: A Reply to
Lightfoot and Cerrato. Archaeology of Eastern North America
18: in press.
1990b Problems and Choices in Shell Seasonality. In Incremental
Growth Structures in Zoo archaeological Seasonality, edited by
G. G. Monks and J. M. Savelle. In preparation.
1991 Gender, Shellfishing, and the Shell Mound Archaic. In
Engendering Archaeology: Women and Prehistory, edited by
J. Gero and M. Conkey, pp. 276-300. London: Basil Blackwell.
Clark, G
1977 The Coast of Shelekof Strait 1963-1965. University of Oregon
Anthropology Papers, 13. Eugene.
Clarke, L. R., and A. H. Clarke
1980 Zooarchaeological Analysis of Mollusc Remains from Yoquot,
British Columbia. Parks Canada, History and Archaeology
43:37-57.
Coutts, P. J. E
1969 The State of Preservation of Shell Material in Midden Sites.
Transactions of the Royal Society of New Zealand 2(9):
135-37.
Crook, M.
1978 Mississippian Period Community Organizations on the Georgia
Coast. Ph.D. diss., Department of Anthropology, University of
Florida, Gainesville.
Cumbaa, S.
1976 A Reconsideration of Freshwater Shellfish Exploitation in the
Florida Archaic. Florida Anthropologist 29(2):49-59.
Curren, C.
1987 Archaeology at Bluewater Bay: 80K102, a Late Archaic Period
Site in Northwest Florida. University of West Florida, Office of
Cultural and Archaeological Research, Report of Investigations,
9.
Deith, M.
1988 A Molluscan Perspective on the Role of Foraging in Neolithic
Farming Economies. In The Archaeology of Prehistoric
Coastlines, edited by G. Bailey and J. Parkington, pp. 116-24.
New York: Cambridge University Press.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
290 Cheryl Claassen
Driscoll, E. G, and T. Weltin

1973 Sedimentary Parameters as Factors in Abrasive Shell Reduction.
Palaeogeography, Palaeoclimatology, Palaeoecology 13:275-88.
Drover, C. E.
1974 Seasonal Exploitation of Chione Clams on the Southern
California Coast. Journal of California Anthropology 1:224-32.
Durum, W. H., and J. Haffty
1961 Occurrence of Minor Elements in Water. Geological Survey
Circular 445. Washington, D.C.: U.S. Department of the Interior.
Erlandson, J.
1988 The Role of Shellfish in Prehistoric Economies: A Protein
Perspective. American Antiquity 53:102-9.
Ferguson, R. O.
1975 Seasonality Study of Prehistoric Bivalves from the Boardwalk
Site (GbTo 31), Prince Rupert, British Columbia. Ms no. 1117,
Archives, Archaeological Survey of Canada, National Museum of
Canada, Ottawa.
Forbes, M. H. C.
1976 Gathering in the Argolid: A Subsistence Subsystem in a Greek
Agricultural Community. Annals of the New York Academy of
Science 268:251-64.
Goode, G B.
1884 The Fisheries and Fisheries Industry of the United States.
Washington, D.C.: U.S. Government Printing Office.
Goodwin Associates, Inc.
1986 Archeological Testing at 16 CM 61, a Prehistoric Shell Midden in
Cameron Parish, Louisiana. New Orleans: U.S. Army Corps of
Engineers.
Goodwin, R. C.
1979 The Prehistoric Cultural Ecology of St. Kitts, West Indies: A
Case Study in Island Archeology. Ph.D. diss., Department of
Anthropology, Arizona State University, Tempe. Ann Arbor:
University Microfilms.
Gray son, D.
1984 Quantitative Zooarchaeology. Orlando: Academic Press.
Greenwood, R. S.
1961 Quantitative Analysis of Shells from a Site in Goleta, California.
American Antiquity 26:416-20.
1972 9000 Years of Prehistory at Diablo Canyon, San Luis Obispo
County, California. San Luis Obispo County Archaeological
Society, Occasional Paper, 7. San Luis Obispo.
Hale, S.
1984 Analysis of Faunal Remains from a Late Archaic and St. Johns I

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
and II Period Site in Volusia County, Florida. Paper presented at

the Southeastern Archaeological Conference, Pensacola.
Hallam, A.
1967 The Interpretation of Size-Frequency Distributions in Molluscan
Death Assemblages. Palaeontology 10:25-42.
Ham, L. C.
1976 Analysis of Shell Samples. In The Glenrose Cannery Site, edited
by R. G. Matson, pp. 42-78. National Museum of Man, Mercury
Series, Archaeological Survey of Canada Paper, 52. Toronto.
1982 Seasonality, Shell Midden Layers and Coast Salish Subsistence
Activities at the Crescent Beach Site, DgRr 1. Ph.D. diss.,
Department of Anthropology, University of British Columbia,
Vancouver.
Hancock, M.
1982 The Determination of Archaeological Site Seasonality Using the
Remains of Mya arenaria: Examples from the Central Maine
Coast. Master's thesis, Department of Anthropology, University
of Maine, Orono.
1984 Analysis of Shellfish Remains: Seasonality Information. In
Chapters in the Archaeology of Cape Cod, vol. 1, edited by
F. McManamon, pp. 121-56. Cultural Resources Management
Study, 8. Boston: National Parks Service.
Hariss, R. C, and O. H. Pilkey
1966 Temperature and Salinity Control of the Concentration of Skin
Na, Mn, and Fe in Dendraster excentricus. Biological Science
20:235-38.
Hill, F. C.
1975 Effects of the Environment on Animal Exploitation by Archaic
Inhabitants of the Koster Site, Illinois. Ph.D. diss., Department
of Anthropology, University of Kentucky.
Hockey, P. A. R., and A. L. Bosman
1986 Man as an Intertidal Predator in Transkei: Disturbance,
Community Convergence and Management of a Natural Food
Supply. Oikos 46:3-14.
Hockey, P. A. R., A. L. Bosman, and W. R. Siegfried
1988 Patterns and Correlates of Shellfish Exploitation by Coastal
People in Transkei: An Enigma of Protein Production. Journal of
Applied Ecology 25:353-63.
Howard, J.
1977 Seasonality and Settlement Patterns in the Orange County
Coastal Foothills. Pacific Coast Archaeological Society
Quarterly 11(2):11-21.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
292 Cheryl Claassen
Keen, S.
1979 The Growth Rings of Clam Shells from Two Pentlatch Middens
as Indication of Seasonal Gathering. Archaeology Division,
Heritage Conservation Branch, Occasional Paper, no. 3.
Victoria, B.C.
Kent, B.
1988 Making Dead Oysters Talk. St. Marys City: Maryland Historical
Trust, Jefferson Patterson Park and Museum.
Kerber, J.
1985 Digging for Clams: Shell Midden Analysis in New England.
North American Archaeologist 6:97-113.
Klippel, W, and D. Morey
1986 Contextual and Nutritional Analysis of Freshwater Gastropods
from Middle Archaic Deposits at the Hays Site, Middle
Tennessee. American Antiquity 51:799-813.
Klippel, W, G Celmer, and J. Pardue
1978 The Holocene Naiad Record at Rodgers Shelter in the Western
Ozark Highland of Missouri. Plains Anthropologist 23:157-70.
Koeper, H.
1981 Prehistoric Subsistence and Settlement in the Newport Bay Area
and Environs, Orange County, California. Ph.D. diss.,
Department of Anthropology, University of California,
Riverside. Ann Arbor: University Microfilms.
Koike, H.
1979 Seasonal Dating and the Valve-Pairing Technique in Shell
Midden Analysis. Journal of Archaeological Science 6(l):63-74.
Larson, L.
1980 Aboriginal Subistence Technology on the Southeastern Coastal
Plain During the Late Prehistoric Period. Gainesville: University
of Florida Press.
Lee, EG, and W. Wilson
1969 Use of Chemical Composition of Freshwater Clamshells as
Indicators of Paleohydrologic Conditions. Ecology 50:990-97.
Lightfoot, Kent
1985 Shell Midden Diversity: A Case Example from Coastal New
York. North American Archaeologist 6:289-324.
Lightfoot, K., and R. Cerrato
1989 Regional Patterns of Clam Harvesting Along the Atlantic Coast
of North America. Archaeology of Eastern North America 17:
31-46.
Lischka, J. J., and P. D. Sheets
1980 Reply to Yesner. Current Anthropology 21:739-40.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
Lobdell, J. E.
1980 Prehistoric Human Populations and Resource Utilization in
Kachemak Bay, Gulf of Alaska. Ph.D. diss., Anthropology
Department, University of Tennessee, Knoxville. Ann Arbor:
Lyons, E.
1978 A Statistical Method of Seasonality Determination from Chione
undatella. Pacific Coast Archaeological Society Quarterly
4(3):33-42.
McManamon, Francis
1984 Chapters in the Archeology of Cape Cod, vol. 2. Cultural
Resources Management Study, 8. Division of Cultural
Resources, North Atlantic Regional Office, National Parks
Service, U.S. Department of the Interior. Boston.
Mackay, R., and J. P. White
1987 Musseling in on the NSW coast. Archaeology in Oceania
22:107-11.
Matteson, M.
1955 Studies on the Natural History of the Unionidae. American
Midland Naturalist 53:126-45.
1960 Reconstruction of Prehistoric Environments Through the
Analysis of Molluscan Collections from Shell Middens.
American Antiquity 26:117-20.
Maxwell, D.
1989a Shellfishing Seasonality: Some Midden Sampling Requirements.
Paper read at the Annual Meeting of the Canadian
Archaeological Association, Fredericton.
1989b Growth Coloration: A Method For Determining the Season of
Collection of Archaeological Shellfish. Master's thesis,
Department of Archaeology, Simon Fraser University.
May, J. A.
1982 Midden Formation Modeling Using Ethnographic and
Archaeological Data: A Trend Surface Analysis of Midden
Deposits at the Carlston Annis Site, Kentucky. Ph.D. diss.,
Department of Anthropology, University of Missouri, Columbia.
Ann Arbor: University Microfilms.
Meehan, B.
1982 Shell Bed to Shell Midden. Canberra: Australian National
University.
Meighan, C. W.
1969 Molluscs as Food Remains in Archaeological Sites. In Science in
Archaeology, edited by D. Brothwell and E. Higgs, pp. 415-22.
London: Methuen.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
294 Cheryl Claassen
1989 Early Shell-Mound Dwellers on the Pacific Coast of North

America. Paper presented at the Pacific 89 Conference, Seattle.
Mills, C. G
1985 Shellfish Utilization and Its Effect on Rocky Shore Biota in
Transkei. Master's thesis, Department of Archaeology,
University of Cape Town.
Monks, G.
1977 An Examination of Relationships Between Artifact Classes and
Food Resource Remains at Deep Bay, DiSe 7. Ph.D. diss.,
Department of Archaeology, University of British Columbia.
Mountjoy, J., and C. Claassen
1989 Seasonality and Diet of Middle Formative Peoples on the Central
Coast of Nayarit, Mexico. Paper presented at the Annual
Meeting of the American Anthropological Association,
Washington, D.C.
Muckle, R. J.
1985 Archaeological Considerations of Bivalve Shell Taphonomy.
Master's thesis, Department of Archaeology, Simon Fraser
University.
Nelson, D. J.
1963 The Strontium and Calcium Relationships in Clinch and
Tennessee River Mollusks. In Radioecology, edited by V. Schultz
andW. Klement, pp. 203-11. New York: Reinhold.
Nichol, R. K., and C. J. Wild
1984 Numbers of Individuals in Faunal Analysis: The Decay of Fish
Bone in Archaeological Sites. Journal of Archaeological Science
11:35-51.
Nodine, B.
1986 Society and the Universal Problem of Refuse Disposal: An
Anthropological Analysis. Master's thesis, Department of
Anthropology, University of Florida, Gainesville.
O'Brien, D., and D. Peter
1983 Preliminary Ceramic and Seasonality Analyses from the
Regional Settlement Survey on St. Catherines Island. Paper
presented at the Annual Meeting of the Southeastern
Archaeological Conference, Columbia, S.C.
Onat, A. R. B.
1985 The Multifunctional Use of Shellfish Remains: From Garbage to
Community Engineering. Northwest Anthropological Research
Notes 19:201-7.
Palmer, J., and J. R. Williams
1977 The Formation of Goethite and Calcareous Lenses in Shell
Middens of Florida. Florida Anthropologist 30(l):24-27.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
Parmalee, P., and W. Klippel

1974 Freshwater Mussels as a Prehistoric Food Resource. American
Antiquity 39:421-34.
Peacock, W. R. B.
1978 Probabilistic Sampling in Shell Middens: A Case-Study from
Oronsay, Inner Hebrides. In Sampling in Contemporary British
Archaeology, edited by J. Cherry, C. Gamble, and S. Shennan, pp.
177-90. British Archaeological Reports, British Series, 50. Oxford.
Perlman, S. M.
1973 Pect?n irradians Growth Patterns. Master's thesis, Department
of Anthropology, University of Massachusetts, Amherst.
Post, A.
1982 Evaluation of Freshwater Mussels (Megalonaias giganteaj as a
New Protein Source. Ph.D. diss., Department of Food Technology
and Science, University of Tennessee, Knoxville. Ann Arbor:
Quitmyer, I.
1985 Zooarchaeological Methods for the Analysis of Shell Middens at
Kings Bay. In Aboriginal Subsistence and Settlement
Archaeology of the Kings Bay Locality, edited by William
Adams, vol. 2, pp. 33-49. University of Florida Dept. of
Anthropology, Reports of Investigations, 2. Gainesville.
Quitmyer, I., S. Hale, and D. Jones
1985 Paleoseasonality Determinations Based on Incremental Shell
Growth in the Hard Clam, Mercenaria mercenaria, and Its
Implications for the Analysis of Three Southeast Georgia Coastal
Shell Middens. Southeastern Archaeology 4:27-40.
Reitz, E., and I. Quitmyer
1988 Faunal Remains from Two Coastal Georgia Swift Creek Sites.
Southeastern Archaeology 7:95-108.
Riser, G
1987 The Quave Ceramic Collection: Mississippian Presence on the
North Shore of Lake Pontchartrain. Paper read at the Annual
Meeting of the Louisiana Archaeological Society, New Orleans.
Roscoe, E.
1967 Ethnomalacology and Paleoecology of the Round Butte
Archaeological Sites, Deschutes River Basin, Oregon. University
of Oregon Museum of Natural History Bulletin 6:1-20.
Rosenberg, G. D.
1980 An Ontogenetic Approach to Environmental Significance of
Bivalve Shell Chemistry. In Skeletal Growth in Aquatic
Organisms, edited by D. Rhoads and R. Lutz, pp. 133-68.
New York: Plenum Press.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
296 Cheryl Claassen
1990 The "Vital Effect" on Skeletal Trace Element Content as

Exemplified by Magnesium. Unpublished manuscript.
Rowley-Conwy, P.
1983 Sedentary Hunters: The Ertebolle Example. In Hunter-Gatherer
Economy in Prehistory, edited by G. N. Bailey, pp. 111-26. New
York: Cambridge University Press.
Sanger, D.
1981 Unscrambling Messages in the Midden. Archaeology of Eastern
North America 9:37^2.
Scudder, T.
1971 Gathering Among African Woodland Savannah Cultivators.
Zambian Paper, 5.
Sidwell, V. D., J. Bonnett, and E. Zook
1973 Chemical and Nutritive Values of Several Fresh and Canned
Finfish, Crustaceans and Mollusks Part 1 : Proximate
Composition, Calcium, and Phosphorus. Marine Fisheries
Review 35(12): 16-19.
Sidwell, V. D., P. R. Foncannon, N. Moore, and J. Bonnett
1974 Composition of the Edible Portion of Raw (Fresh or Frozen)
Crustaceans, Finfish, and Mollusks. I. Protein, Fat, Moisture,
Ash, Carbohydrate, Energy Value, and Cholesterol. Marine
Fisheries Review 36(3):21-35.
Siegfried, W. R., P. A. R. Hockey, and A. A. Crowe
1985 Exploitation and Conservation of Brown Mussel Stocks by
Coastal People of Transkei. Environmental Conservation
12:303-7.
Sigler-Eisenberg, B., and M. Russo
1986 Seasonality and Function of Small Sites on Florida's Central-East
Coast. Southeastern Archaeology 5:21-31.
Skelton, D.
1978 The Seasonal Factor of Rangia cuneata Clam Collecting. In
Prehistoric Archaeological Investigations at Palmetto Bend
Reservoir, edited by D. Skelton, pp. 65-75. Texas Archaeological
Survey, Research Report, 45. Austin: University of Texas.
Sorant, P., and J. R. Shenkel
1984 The Calculation of Volumes of Middens, Mounds, and Strata
Having Irregular Shapes. American Antiquity 49:599-603.
Speck, F.
1946 Rappahannock Taking Devices: Traps, Hunting and Fishing.
Joint Publication of the Museum of the University of
Pennsylvania and the Philadelphia Anthropological Society, 1.
Philadelphia.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
Spiess, A., and M. Hedden

1983 Kidder Point and Sears Island in Prehistory. Occasional
Publications in Maine Archaeology, 3. Maine Historic
Preservation Commission, Augusta.
Stein, J.
1983 Earthworm Activity: A Source of Potential Disturbance of
Archaeological Sediments. American Antiquity 48(2):277-89.
1985 Interpreting Sediments in Cultural Settings. In Archaeological
Sediments in Context, edited by J. K. Stein and W. R. Farrand, pp.
5-19. Orono: Center for the Study of Early Man, Institute for
Quaternary Studies, University of Maine.
1986 Coring Archaeological Sites. American Antiquity 51:505-27.
S wann, C. P., M. Carriker, and J. Ewart
1984 Significance of Environment and Chronology on Distribution
of 16 Elements (Na to Sr) in the Shells of Living Oysters.
Nuclear Instruments and Methods in Physics Research B3
1984:392-95.
Tartaglia, L.
1976 Prehistoric Maritime Adaptations in Southern California. Ph.D.
diss., Department of Anthropology, UCLA. Ann Arbor:
Tevesz, M., and J. Carter
1980 Environmental Relationships of Shell Form and Structure of
Unionacean Bivalves. In Skeletal Growth of Aquatic Organisms,
edited by D. Rhoads and R. Lutz, pp. 295-323. New York:
Plenum Press.
Tiffany, W. J.
1968 The Life Cycle and Ecology of the Beach Clam Donax variabilis.
Master's thesis, Department of Biology, Florida State University.
Trinkley, M.
1981 Studies of Three Woodland Period Sites in Beaufort County, SC.
Columbia: South Carolina Department of Highways and Public
Transportation.
1986 Indian and Freedmen Occupation at the Fish Haul Site
(38BU805), Beaufort County, South Carolina. Chicora
Foundation Research Series, 7. Columbia.
Turekian, K., and R. Armstrong
1960 Magnesium, Strontium and Barium Concentrations and Calcite
Aragonite Ratios of Some Recent Molluscan Shells. Journal of
Marine Research 18:133-51.
Viosca, P.
1933 Louisiana Out-of-Doors. New Orleans: published by author.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC
298 Cheryl Claassen
Voigt, E. A.
1975 Studies of Marine Mollusca from Archaeological Sites: Dietary
Preferences, Environmental Reconstructions and Ethnological
Parallels. In Archaeozoological Studies, edited by A. T. Clason,
pp. 87-98. Amsterdam: North Holland.
Waselkov, G. A.
1987 Shellfish Gathering and Shell Midden Archaeology. In Advances
in Archaeological Method and Theory, vol. 11, edited by
M. Schiffer, pp. 93-210. Orlando: Academic Press.
Weide, M.
1969 Seasonality of Pismo Clam Collecting at Ora-82. University
of California, Archaeological Survey, Annual Report 11 :
127-42.
Wessen, G.
1982 Shell Middens as Cultural Deposits: A Case Study from Ozette.
Ph.D. diss., Department of Anthropology, Washington State
University, Pullman. Ann Arbor: University Microfilms.
Widmer, R.
1989 Archaeological Research Strategies in the Investigation of Shell
Bearing Sites, a Florida Perspective. Paper presented at the
Annual Meeting of the Society for American Archaeology,
Atlanta.
Wild, C. J., and R. K. Nichol
1983 Estimation of Original Numbers of Individuals from Paired Bone
Counts Using Estimators of the Krantz Type. Journal of Field
Archaeology 10:337-44.
Wilken, G
1970 The Ecology of Gathering in a Mexican Farming Community.
Economic Botany 24(30):286-94.
Wing, E., and A. Brown
1979 Paleonutrition: Method and Theory in Prehistoric Foodways.
Orlando: Academic Press.
Wing, E., and I. Quitmyer
1985 Screen Size for Optimal Data Recovery: A Case Study. In
Aboriginal Subsistence and Settlement Archaeology of the Kings
Bay Locality, vol. 2, edited by William Adams, pp. 49-59.
Univeristy of Florida Dept. of Anthropology, Reports of
investigations, 2. Gainesville.
1986 A Modern Midden Experiment. Paper presented at the Annual
Southeastern Archaeological Conference, Nashville.

174.70.75.246 on Fri, 14 Oct 2022 15:02:08 UTC

Claassen 1991

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Claassen 1991

Uploaded by

Copyright:

Available Formats

Normative Thinking and Shell-Bearing Sites

Author(s): Cheryl Claassen

Stable URL: https://www.jstor.org/stable/20170217

This content downloaded from

Normative thinking or normative treatment is that whic

This content downloaded from

portrayed as human food debris. Furthermore, normative thought

Normative thinking has reduced all archaeological shell

This content downloaded from

Types of Shell-bearing Deposits and Sites

Several archaeologists have recognized a need to refine terminology

This content downloaded from

activities; primary refuse, wide range of activities; secondary refuse,

This content downloaded from

This content downloaded from

is rarely consulted by archaeologists. Shell orientation, fragmenta

This content downloaded from

What type of exposure to use?whether block, column, square,

Sample size of matrix. The goal of site excavation should be to obtain

This content downloaded from

% of Site Screen % Excavated % of Site

sentative sample requires knowledge of the volume and variability

This content downloaded from

Table 6.1. Continued

% of Site Screen % Excavated % of Site

as to site area or volume; their work is generally not included in this

This content downloaded from

sizes. For instance, the oldest marine shell-bearing midden site in

Subsample size. Sampling specific to subsistence studies was given

This content downloaded from

This content downloaded from

46 41 34 37 37 37 31 40 39 38 36 34 450 1987 6 43 33 37 33 50 48 50 52355230302938353142497451986

Species J FMAMJJASONDN= Year Refer 37 36 30 27 35 33 29 38 ? ? ? 41 306 1988 6 37 33 38 34 42 47 53 41 47 ? ? ? 372 1988 6

A P.i. __ ____XXX? ? ? ? 53 71-72 4

Damarisotta ME M.a. ? 10 10 10 10 10 10 10 10 10 10 10 110 1980 2 Narragansett RI M.m. ? 10 10 12 10 10 10 10 10 9 10 11 112 82-83 5

Shark Inlet NJ M.m. ? 6? 6? 6?6?6? 6 36 1981 6

ll Seasonality Controls in North America

This content downloaded from

19741984 1983 1984 1985 1986 19871988

M.c. = Mercenaria campechiensis

R.c. R.c. R.c.

M.m. = Mercenaria mercenaria

Escambia Bay Alligator Harbor BoundaryShark

This content downloaded from

Collecting M. mercenaria in North Carolina from 1980 to 1988

This content downloaded from

Sample sizes of archaeological shells for seasonality studies are

This content downloaded from

Project n Ref Project n Ref

This content downloaded from

Table 6.3. Continued

Project n Ref Project n Ref

41 Ch 172 523 41 Jkll4TX 99

This content downloaded from

permit the collection of independent probability samples from strata

Experimentation with Shell and Shell Middens

Perhaps it is the prevalence of normative thinking about shell and

This content downloaded from

This content downloaded from

320, and 640 walking passes over Saxidomus giganteus, P. staminea,

This content downloaded from

creating shell mounds. Herein lie invaluable data on mound col

Food and Feeding

Dietary reconstruction, a popular program of analytical

A P.i. __XXX? ? ? ? 53 71-72 4