Journal of Archaeological Science: Reports 21 (2018) 107–116

Contents lists available at ScienceDirect

Journal of Archaeological Science: Reports

journal homepage: www.elsevier.com/locate/jasrep

Estimating crown conch (Melongena corona) tissue weight from T

archaeological shell measurements: An allometric methodology for coastal
historical ecological research
⁎
Kendal Jacksona, , Elizabeth Southarda, Sharlene O'Donnellb, John Arthurc
a
Department of Anthropology, University of South Florida, Tampa, USA
b
Department of Anthropology, University of Florida, Gainesville, USA
c
Department of Society, Culture, and Language, University of South Florida, St. Petersburg, USA

A R T I C LE I N FO A B S T R A C T

Keywords: Coastal archaeologists and historical ecologists are taking an increasingly robust interest in marine shell as-
Experimental zooarchaeology semblages recovered from coastal villages and civic-ceremonial sites. These assemblages must be quantified
Florida Gulf Coast before archaeologists can make assessments of biomass flows and subsistence contributions. We present the
Invertebrate allometry results of an experimental allometric study on Melongena corona snails collected from the mangrove dominated
Marine shell
shoreline of Weedon Island Preserve, Florida, USA. Our analysis produced regression constants for predicting
tissue weight estimates from four independent linear shell metrics, including: length, aperture-length, height,
and width. This study is unique in its integration of field and laboratory experimentation, and in the large sample
size used to develop allometric constants. To exemplify the utility of our regression models, we apply our al-
lometric constants to a late-Precolumbian (ca. 895–1268 CE) marine shell assemblage excavated from the
Weeden Island site (8PI1), Pinellas County, Florida, USA.

1. Introduction
In recent years, archaeologists have begun to take an increased
interest in the role of marine shellfish in the subsistence strategies,
settlement architecture, and sociopolitical lives of coastal fisher-
hunter-gatherer societies (Erlandson et al., 2008; Keegan et al., 2003;
Lulewicz et al., 2017; Marquardt and Kozuch, 2016; Onat, 1985;
Trubitt, 2005; Walker, 2000). To understand the dietary contributions
of particular taxa and map-out energy flows between near-shore
ecosystems and human communities, zooarchaeologists must quantify
mollusk-shell data via estimates of tissue weight (i.e., “meat weight”),
biomass, and/or kilocalories (Perez, 2010; Reitz et al., 1987; Reitz and
Wing, 2008). Indeed, research by Thomas and Mannino (2017) has
demonstrated that meat weight values may be critical to properly
understanding relative taxonomic abundance for various mollusks in
archaeological assemblages (also see Claassen, 2000). Early ap-
proaches to this problem – based on the work of White (1953) –
multiplied average meat weight parameters by minimum number of
individuals (MNI) values to calculate the estimated total meat weight
for a particular taxon in a given assemblage. Subsequently, several
attempts have sought to improve upon White's method by modifying
averaged meat weight predictions to reflect variation in sex, age, and
size differences (Lyman, 1979; Reed, 1963; Smith, 1975; Ziegler,
1973). However, in utilizing average values and MNI, these attempts
share the fundamental flaws inherent in White's (1953) approach – an
assumption that the dimensions of the bone or shell element being
measured relate to meat weight at a constant (i.e., linear) ratio (Reitz
et al., 1987:305), and that MNI may not be comparable across taxa or
across depositional contexts (Thomas and Mannino, 2017). Indeed,
linear surface area metrics and volumetric variables such as meat
weight relate to each other in exponential terms and may be described
by the equation y = axb where: y is meat weight, x is the linear
measurement, and a and b are constants that define the relationship
between these variables (Gould, 1966). Our study presents allometric
regressions developed from experimental collection and analyses on
305 Melongena corona snails collected along a mangrove-dominated
shoreline site in Tampa Bay, Florida, USA, between 2011 and 2013.
We report new experimentally derived constants for calculating tissue
weight estimates from four archaeological shell metrics: length,
aperture-length, width, and height. We then apply our equations to
zooarchaeological assemblages excavated from the Weeden Island site
(8PI1), a large (> 2 km2), multi-component, coastal village and

⁎
Corresponding author at: 4202 E. Fowler Ave. SOC 107, Tampa, Florida 33620, USA.
E-mail address: Kendalj@mail.usf.edu (K. Jackson).

https://doi.org/10.1016/j.jasrep.2018.07.004
Received 20 March 2018; Received in revised form 5 July 2018; Accepted 6 July 2018
Available online 17 July 2018
2352-409X/ © 2018 Elsevier Ltd. All rights reserved.
K. Jackson et al.
Fig. 1. Location of experimental harvesting site, Weedon Island Preserve, Florida.
ceremonial center on the western shore of Tampa Bay, Florida
(Weisman et al., 2005).
2. Background
Crown conch snails (Melongena corona) are predatory marine gas-
tropods found in particular abundance on the Gulf Coast of Florida, but
also found on the eastern coast of Alabama, in the Yucatan, and along
the Atlantic Coast of Florida and Georgia (Bowling, 1994; Pilsbry and
Vanatta, 1934). They inhabit low-energy intertidal environments (i.e.,
sand and mud flats, marshes, and oyster bars) with moderate-salinity
conditions (12–15%) and become inactive at salinities below 9%
(Hathaway and Woodburn, 1961). Crown conchs range in size generally
between 10 and 110 mm in length; larger individuals more typically
inhabit eastern oyster (Crassostrea virginica) reefs, while juveniles dis-
proportionately occupy intertidal sand or mud flats and marshes
(Bowling, 1994). They are most active at high tide during feeding;
crown conch are both scavengers and predators, feeding on bivalves,
other gastropods, and various estuarine detritus (see Turner, 1959).
108
Journal of Archaeological Science: Reports 21 (2018) 107–116
Further, activity patterns among crown conch are seasonal; in winter
they are relatively inactive and remain entirely or partially buried in
the sand or mud substrate. In summer they are more active and achieve
the vast majority of their growth.
Fisher-hunter-gatherer communities on Florida's Gulf Coast col-
lected M. corona and other estuarine gastropods as an important faunal
resource from onset of highly-productive estuarine conditions during
the Mid-Holocene through the protohistoric era (e.g., Duke, 2015;
Mikell and Saunders, 2007; Defrance and Walker, 2013). Their shell
remains are relatively ubiquitous within invertebrate faunal assem-
blages from village and mound sites along the Gulf Coast where mod-
erate salinities predominate (Vojnovski, 1995; Defrance and Walker,
2013; Walker, 1992); they are also present within shell-bearing as-
semblages along the Atlantic coasts of Florida and Georgia (Claassen,
1986; Reitz, 1988). In some sub-regions, for instance on Florida's Big
Bend marsh coast, crown conch shells may have been collected prin-
cipally for tool manufacturing (Blankenship, 2013; Duke, 2015; O'Neal,
2016). In other areas, such as Florida's Central Gulf Coast - encom-
passing Tampa and Sarasota Bays - this species makes up a substantial
K. Jackson et al.
Fig. 2. Collecting crown conch within the experimental harvesting site, Weedon Island Preserve, Florida. Photo by Elizabeth Southard.
portion of the invertebrate faunal count, suggesting that it may have
been a major subsistence resource comparable to eastern oyster (Cras-
sostrea virginica) (Janus Research, 2004; O'Donnell, 2018: in press;
Quitmyer, 2002; Vojnovski, 1995, 1998). For example, O'Donnell
(2018:in press) analyzed zooarchaeological assemblages from a shell
midden at the Weeden Island site (8PI1) and found that crown conch
(n = 1329) makes up nearly 26% of the total invertebrate MNI count
(n = 5254) and ranks second only to oyster (n = 1846).
We follow Claassen (1998) in taking caution against the normative
assumption that archaeological shells unequivocally represent food
debris; in her words: “evidence of collection and preparation [of ar-
chaeological shellfish] does not equal consumption” (Claassen,
1998:285; also see Waselkov, 1987:166). Indeed, work by Onat (1985)
on the Northwest Coast and by a growing school of Florida archae-
ologists (e.g., Pluckhahn et al., 2016; Schwadron, 2017; Sassaman et al.,
2017) has irreparably deconstructed ‘shells-as-garbage’ hypotheses by
demonstrating that coastal Precolumbian foragers planned and rapidly
constructed large-scale terraformed landscapes utilizing marine shell
within sedimentary matrices. In addition to consumption, tool crafting,
and architectural pursuits, archaeological gastropod shells may also
represent the remnants of ancient chumming or baiting strategies (see
Larson, 1980; Speck, 1946). Further, certain marine gastropods, such as
the Lightning Whelk (Busycon sinistrum), functioned as important
symbolic creatures within metaphysical and sociopolitical domains in
southeastern North America throughout the late-Precolumbian era
(Marquardt and Kozuch, 2016). To put our position simply, we re-
commend that the regression constants presented here be used to esti-
mate tissue weights and biomass values for M. corona remains within
shell assemblages, but also caution that archaeologists must take con-
siderable care in proposing what portion of an assemblage's tissue
weight was consumed.
Allometric constants for M. corona are currently available for
aperture measurements via a seminal 100-specimen study by Reitz et al.
(1987) based on collections from the Florida Museum of Natural His-
tory in Gainesville, Florida. Their research explored how sample size,
and snail size-class distribution interact with the allometric formula
109
Journal of Archaeological Science: Reports 21 (2018) 107–116
(y = axb) to influence regression constants. Results suggested that large
sample sizes are important, and that intermediate-sized snails (those
approximating average size) are predisposed to introduce much of the
residual error within allometric regressions, decreasing coefficient of
determination statistic (R2) values (Reitz et al., 1987:309; also see
Peters, 1983:28). We aim to build upon this foundational work by:
collecting experimental specimens across a range of estuary conditions,
considering a larger sample size (n = 305), and developing in-
dependent allometric models for four distinct shell metrics, such that
tissue weight may be estimated from incomplete (fragmentary) and/or
modified shell specimens.
3. Experimental methods
After recognizing that M. corona shells make-up the majority (some
55%) of gastropod assemblages within midden excavation blocks at the
Weeden Island site (O'Donnell, 2018: in press), we established a small
(8000 m2) experimental harvesting site on the mangrove-dominated
estuary shoreline of Weedon Island Preserve (Fig. 1). Beginning in
December 2011 and ending in September 2013 student researchers met
one morning each month to collect crown conch snails from the ex-
perimental shoreline site (Fig. 2).
During each of 10 different monthly harvests we retained approxi-
mately 30 specimens for experimental allometric analysis at the
University of South Florida, St. Petersburg archaeology laboratory.
Snails were euthanized in an ice bath before being assigned specimen
numbers. We took basic measures to void the snails of visceral water,
and then measured weight, length, aperture-length, width, and height
(see Palmer, 1982). Fig. 3 defines each of these shell metrics as assessed
for this study. Experimental specimens were then placed in a 60 °C
drying oven for 48 h. After drying and recording dry weights, the dried
snails were placed into a 500 °C muffle furnace for 5 h, after which we
emptied the ash-remains and weighed the shell to yield ash-free dry
weight values. To calculate tissue weights, we subtracted ash-free dry
weights from wet weight remeasurements. All linear shell measure-
ments were made to 0.01 mm via digital calipers, and weights were
K. Jackson et al.
Fig. 3. Definitions for crown conch shell metrics used in this study. Illustrations
by Kendal Jackson.
measured to 0.01 g. using a digital balance. We organized the data,
conducted regression and regression-residuals analyses, and produced
charts via IBM SPSS 24. Due to the exponential (natural log) relation-
ship between linear surface area variables and volumetric variables
(i.e., tissue weight) regressions between them must be modeled from
natural log (ln)-transformed variable values.
To account for potential variation in tissue weight density due to
seasonality, tidal height, tidal stage, air temperature, and water tem-
perature, we processed and analyzed snails collected across each of
110
Journal of Archaeological Science: Reports 21 (2018) 107–116
these variables (see Table 1). We describe season of harvest following
Wolfe (1990) and Palmiotto (2016); colloquial season terms were dis-
carded in favor for the following more regionally accurate descriptors:
Cool/Dry (October–January), Cool/Wet (February–March), Warm/Dry
(April–May), and Warm/Wet (June–September). Tidal and temperature
data were measured in the field at the time of collection and were later
confirmed with local tide-gauge data available from the National
Oceanic and Atmospheric Administration.
4. Results
While R2 values for each shell metric differ, each of the relationships
defined here are strong and represent suitable models for estimating
tissue weight values from archaeological specimens. Fig. 4 displays the
scatter plots, regression lines, and R2 values for length, aperture-length,
width, and height measurements. Residuals from each of these regres-
sions display random distributions and calculated Cook's Distances for
residuals are extremely low (with none remotely approaching 1), sug-
gesting that residual error is minimal and has not introduced influential
effects. In addition to complete M. corona shells, incomplete shells and
shell fragments may contribute to MNI estimates so long as at least one
of the measurements described above (i.e., length, aperture-length,
width, or height) can be confidently assessed. With our suite of equa-
tion constants, estimated meat weights may be derived from damaged
or worked shells and shell fragments in addition to whole shells (see
Fig. 5). As apparent in Fig. 4 and Table 2, shell length relates most
strongly to estimated meat weight in our model (R2 = 0.836) and
should be preferentially used when shell columellae are intact. How-
ever, columellae are often fragmented and may otherwise exhibit
considerable wear given modification for tool use. In these cases, an-
other shell metric retained in the specimen may be measured and uti-
lized to calculate an estimated tissue weight value. It is our hope that
this suite of allometric constants will help researchers increase the
sample size utilized for tissue weight and biomass estimations, adding
rigor to descriptive and comparative analyses.
5. An archaeological case study from Weeden Island Site (8PI1)
Ancient fisher-hunter-gatherer communities constructed the
Weeden Island site (8Pi1) on the western shore of the Tampa Bay es-
tuary, west-central Florida. Following excavations into the site's burial
mound by the Smithsonian Institution in 1923–1924 and their collec-
tion of decorated mortuary pottery assemblages (Fewkes, 1924),
Weeden Island became the type-site for a regional Woodland period (ca.
500 BCE–1000 CE) archaeological culture known for the intensification
of mortuary ceremonialism, mound construction, and craft specializa-
tion. The Weeden Island archaeological culture is known from sites
extending north up the Florida peninsula and into the southeastern
coastal plain. In the Tampa Bay area, Weeden Island sites are best as-
sociated with the time interval ca. 400–1000 CE (Austin et al., 2014).
The site complex is large (> 2 km2) and was occupied over a deep span
of time, from ca. 7000 BCE through at least 1400 CE (Arthur et al.,
2016, 2018; Lambert, 2006; O'Donnell, 2015; Sears, 1971; Weisman
et al., 2005). During the late-Holocene (after ca. 3000 BP), local forager
communities built-up two large arcing shell-ridges on the Weedon Is-
land peninsula, each atop distinct antecedent parabolic sand dune
ridges (Lambert, 2006).
As the main focus of the 2007 University of South Florida, St.
Petersburg, Archaeology Field School, we opened test excavations into
the southeastern arm of a large parabolic shell midden-ridge at the
Weeden Island site. We screened all midden matrix in the field through
nested 0.25 in. (0.635 cm) and 0.125 in. (0.318 cm) hardware mesh;
K. Jackson et al. Journal of Archaeological Science: Reports 21 (2018) 107–116

Table 1
Harvest/collection record for experimental allometry samples.
Experimental harvest/collection record

Date Season Tidal stage Tidal height (m) Water temperature (°C) Air temperature (°C) ̄
x length (mm) ̄
x wet weight (g)

Dec. 14, 2011 Cool/dry Ebb 0.017 19.4 16.9 62.3 28.9
Sep. 12, 2012 Warm/wet Flow 0.767 30.7 24.1 55.9 25.3
Oct. 29, 2012 Cool/dry Flow 0.099 25.1 17.2 55.2 23.6
Nov. 19, 2012 Cool/dry Ebb 0.301 28.8 13.9 55.9 25.3
Jan. 18, 2013 Cool/dry Ebb −0.098 19.4 14.7 60.9 32.6
Feb. 19, 2013 Cool/wet Flow 0.358 16.6 12.7 59.6 31.9
Mar. 22, 2013 Cool/dry Flow 0.374 18.6 14.5 61.3 31.1
Apr. 19, 2013 Wet/dry Flow 0.637 26 20.7 65.4 34.9
Jul. 22, 2013 Wet/warm Flow 0.732 29.1 28.4 67.9 43.8
Sep. 23, 2013 Wet/warm Ebb 0.318 29 28.3 64.4 38

Fig. 4. Scatter plots, regression lines, and R2 values for estimated meat weight and length, aperture-length, height, and width.

111
K. Jackson et al. Journal of Archaeological Science: Reports 21 (2018) 107–116

Fig. 5. Incomplete shell examples suitable for tissue weight estimation via our allometric constants.

Table 2
Experimentally derived constants and correlation coefficients for fits to the allometric expression y = axb or ln(y) = ln(a) + b[ln(x)].
Shell metric R2 ln (a) b F Sig. ̄
x (mm) σ (mm) Experimental range (mm)

Length 0.836 −9.73 2.97 1548.1 p < .001 60.8 9.2 33.6–117.3
Aperture-length 0.755 −6.70 2.49 934.7 p < .001 39.9 6.9 20.9–81.5
Height 0.692 −7.38 2.84 680.2 p < .001 32.3 4.7 18.1–60.8
Width 0.770 −7.29 2.69 1012.2 p < .001 37.9 6.1 20.6–75.8

112
K. Jackson et al. Journal of Archaeological Science: Reports 21 (2018) 107–116

Table 3 MNI from two one-by-one-meter excavation units (8S9E and 8S10E)
Crown conch shell descriptive statistics by excavation level for units 8S9E and were weighed, measured, and assigned a unique specimen number.
8S10E. Using our suite of allometric regressions we calculated estimated tissue
weights for all crown conch shells that retained at least one of our four
metrics (and thus, contributed to MNI values). Descriptive statistics
Excavation level 0–10 cm 10–20 cm 20–30 cm results of these calculations, organized by arbitrary 10 cm excavation-
levels are presented in Table 3.
Dates (cal. AD) 1165–1265 1050–1085 895–1020
Several of our interests at the Weeden Island site shell-midden ridge
N= 197 738 114 concern the interaction between late-Precolumbian estuarine har-
vesting traditions and late-Holocene environmental flux in the Tampa
Length (mm) Mean 59.78 62 54.65
Bay Estuary. Radiocarbon dates from the midden strata present in 8S9E
Std. error 1.01 0.42 1.2
% intact 92.9 92.8 87.7 and 8S10E suggest that deposition took place from cal. 895 CE to 1265
Aperture-length (mm) Mean 41.95 42.58 40.01 (two-sigma), a time frame corresponding with a period of warmth,
Std. error 0.73 0.31 0.84 frequent precipitation, and (perhaps) elevated sea level known broadly
% intact 86.3 92.7 78.9 as the Medieval Climatic Optimum (Crumley, 1987:240; Gunn,
Width (mm) Mean 39.83 40.767 37.29
1994:17; Jackson, 2016; Walker, 2013). Recent work with coastal shell-
Std. error 0.79 0.3 0.94
% intact 72.1 84.6 72.8 midden assemblages has attempted to assess diachronic patterns in
Height (mm) Mean 33.83 34.87 30.87 mean shell size that may serve as proxies for environmental flux, har-
Std. error 0.66 0.26 0.71 vesting intensities, and other historical ecological processes (Erlandson
% intact 71.6 83.3 72.8
et al., 2008; Harris and Weisler, 2017; Rick et al., 2016; Savarese et al.,
Est. meat weight (g) Mean 13.12 13.8 9.78
Std. error 0.72 0.32 0.77 2016; also see Crumley, 1987). Tracking this type of change in gas-
% intact 100 100 100 tropod assemblages requires the use of standard (i.e., comparable)
metrics for assessing snail size. However, utilizing any single linear
shell metric is problematic because large portions of archaeological
Table 4 gastropod assemblages may be composed of incomplete or fragmented
ANOVA results for complete and incomplete shells. shells. Fragmentation (whether ancient or taphonomic) in such as-
ANOVA: complete and incomplete shells semblages is not likely to be random; thus, utilizing a linear metric, and
omitting fragmentary shells, may introduce serious systematic error
Dependent variable: est. meat weight (g) into comparative analyses. In a similar manner, utilizing estimated
tissue weight values calculated from a single shell metric also requires
Sum of squares df Mean square F Sig.
omitting shells from analyses, and is predisposed to introduce a similar
Between levels 1596.722 2 798.361 9.941 < 0.001 degree of systematic error.
Within levels 83,682.769 1042 80.310 In this case study example, we utilize estimated meat weight values
Total 85,279.492 1044
to assess stratigraphic-temporal differences in mean crown conch size
between three shell-midden strata (0–10 cm, 10–20 cm, and 20–30 cm)
within our two preliminary test units. These three strata are associated
with radiocarbon date ranges (two-sigma) cal. 1165–1265 CE,
1050–1085, and 895–1020, respectively (Arthur et al., 2018; O'Donnell,
2018:in press). After calculating estimated meat weights for each spe-
cimen within the crown conch assemblage (complete and incomplete
shells) making up MNI (n = 1049) we conducted an analysis of var-
iance (ANOVA) that demonstrates statistically significant differences
between shell midden levels (F = 9.9, p < .001) (see Table 4 and
Fig. 6). Multiple comparisons within the ANOVA via Games-Howell
post hoc tests (mitigating unequal sub-sample sizes by level) show that
while mean tissue weights between levels 1 (0–10 cm) and 2
(10–20 cm) are not significantly different (−0.6 g, p = .733), sig-
nificant differences clearly separate level 1 from level 3 (20–30 cm)
(3.42 g, p = .004), and level 2 from level 3 (4.02 g, p < .001)
(Table 5).
Notably, the direction of significant differences within multiple
comparisons show that mean shell size apparently increased over the
ca. 895–1268 CE depositional interval – a period encompassing sub-
regional intensification of sociopolitical stratification (i.e., the
Manasota/Weeden Island-Safety Harbor transition; see Austin et al.,
Fig. 6. Error bar chart of mean (95% C.I.) complete and incomplete tissue 2014). The increase in shell size noted here may serve as an indicator of
weights by level. shifting estuarine conditions (e.g., oyster reef density, paralic wetland
composition) in nearshore environments of Tampa Bay as the Medieval
Climatic Optimum took hold along Florida's Gulf Coast. Of course, we
and all artifacts and ecofacts larger than 0.125 in. (0.318 cm) were must refrain from making more specific interpretations regarding this
collected for analysis. pattern until zooarchaeological data from more taxa have been ana-
Following taxonomic sorting, M. corona shell specimens comprising lyzed.

113
K. Jackson et al. Journal of Archaeological Science: Reports 21 (2018) 107–116

Table 5
ANOVA multiple comparisons for complete and incomplete shells.
Multiple comparisons: complete and incomplete shells

Dependent variable: est. meat weight (g)

(I) Level (J) Level Mean difference (I-J) Std. error Sig. 95% Confidence interval

Lower bound Upper bound

Games-Howell 0–10 cm 10–20 cm −0.604 0.802 0.733 −2.495 1.288

20–30 cm 3.417a 1.061 0.004 0.916 5.917
20–30 cm 0–10 cm 0.604 0.802 0.733 −1.288 2.495
20–30 cm 4.020a 0.830 0.000 2.057 5.984
20–30 cm 0–10 cm −3.417a 1.061 0.004 −5.917 −0.916
10–20 cm −4.020a 0.830 0.000 −5.984 −2.057

a
The mean difference is significant at the 0.05 level.

Table 6 To show the value of allometric constants derived from multiple

ANOVA Results for Complete Shells Only. shell metrics (above), we ran the same ANOVA utilizing only complete
ANOVA: complete shells only
shells (n = 756), representing the likely results if only one regression
model were available. The results from this analysis were quite different
Dependent variable: est. meat weight (g) and no significant differences were identified between excavation levels
(F = 1.97, p = .140) (see Table 6 and Fig. 7). Multiple comparisons via
Sum of squares df Mean square F Sig.
Games Howell post hoc tests yielded miniscule and insignificant mean
Between levels 262.892 2 131.446 1.974 0.140 differences between levels (Table 7). In this assemblage, omitting in-
Within levels 50,148.246 753 66.598 complete crown conch specimens from tissue weight calculations
Total 50,411.139 755 clearly obscures otherwise significant stratigraphic variation in mean
shell size distribution. The comparison between these two analyses of
variance demonstrates the importance of collecting and measuring
entire (or at least very large and representative) marine gastropod as-
semblages from shell midden matrix, as well as the applicability of
multiple allometric regression models available for zooarchaeological
analyses.

6. Conclusions

Our allometric regression analysis of 305 experimentally harvested

M. corona snails demonstrates that linear shell measurements reliably
predict estimated tissue weight values. Regression constants are re-
ported here for length, aperture-length, height, and width. Among these
shell metrics, length was most strongly correlated with estimated meat
weight (R2 = 0.835), followed by aperture-length (R2 = 0.775), height
(R2 = 0.770), and width (R2 = 0.692). In a brief case study on
zooarchaeological crown conch assemblages from the Weeden Island
Site, Florida, we showed that our suite of regression constants sig-
nificantly strengthens comparative analyses by enabling the inclusion
Fig. 7. Error bar chart of mean (95% C.I.) complete-shell tissue weights by of incomplete (i.e., fragmentary or modified) crown conch shell that
level. would be omitted from consideration without the availability of re-
gression models for each major linear shell metric. Our experimental-

Table 7
ANOVA multiple comparisons for complete shells only.
Multiple comparisons: complete shells only

Dependent variable: est. meat weight (g)

(I) Level Mean difference (I-J) Std. Error Sig. 95% Confidence interval

Lower bound Upper bound

Games-Howell 0–10 cm 10–20 cm −0.389 0.945 0.911 −2.626 1.849

20–30 cm 1.694 1.481 0.489 −1.816 5.204
10–20 cm 0–10 cm 0.389 0.945 0.911 −1.849 2.626
20–30 cm 2.083 1.225 0.212 −0.847 5.013
20–30 cm 0–10 cm −1.694 1.481 0.489 −5.204 1.816
10–20 cm −2.083 1.225 0.212 −5.013 0.847

114
K. Jackson et al.
archaeological study is well positioned to aid coastal zooarchaeologists
in producing more complete interpretations of shell midden assem-
blages. These analyses, in turn, will be better able to inform research
questions regarding energy flow (e.g., as a proxy for human population
density), environmental flux, and ancient estuarine harvesting beha-
viors.
Declarations of interest
None.
Acknowledgements
We acknowledge Dr. Steven J. Harper and Pinellas County Parks
and Conservation Resources for granting our collection and excavation
permit (permit #67). We thank volunteers and staff of Weedon Island
Preserve for facilitating access to experimental harvesting sites. We also
thank USF St. Petersburg Biological Sciences for providing access to
critical laboratory instrumentation. For many early mornings of wade-
harvesting and data collection, we extend our sincerest gratitude to the
student volunteers of the USFSP Archaeology Lab. We thank the
Alliance for Weedon Island Archaeological Research and Education
(AWIARE) for their help in facilitating archaeological fieldwork. We
also thank two anonymous reviewers for thoughtful questions and
comments that improved this report.
References
Arthur, John W., Austin, Robert, Edwards, Wendy, Horsley, Timothy, Kolianos, Phyllis,
Morehead, David, O'Donnell, Sharlene, Southard, Christina Perry Sampson Elizabeth,
Wiseman, Brent, 2016. Breaking New Ground: Archaeology of Domestic Life at the
Weedo(e)n Island Site, Pinellas County, Florida, Presented at the Society for
American Archaeology Meeting, Orlando, Florida.
Arthur, John W., Southard, Elizabeth, Draskovich, Heather, Jackson, Kendal, O'Donnell,
Sharlene, West, Shawn, 2018. They never left: daily life of a safety harbor-period
settlement at the Weeden Island Site (8PI1), Florida. In: Presented at the Annual
Meeting of the Florida Anthropological Society, St. Petersburg.
Austin, Robert J., Mitchem, Jeffrey M., Weisman, Brent R., 2014. In: Wallis, Neill J.,
Randall, Asa R. (Eds.), Radiocarbon Dates and the Late Prehistory of Tampa Bay. In
New Histories of Pre-Columbian Florida. University Press of Florida, Gainesville.
Blankenship, Beth, 2013. The Hopewellian Influence at Crystal River, Florida: Testing the
Marine Shell Artifact Production Hypothesis. Master's Thesis. Department of Applied
Anthropology, University of South Florida, Tampa.
Bowling, C., 1994. Habitat and size of the Florida crown conch (Melongena corona
Gmelin): why big snails hang out at bars. J. Exp. Mar. Biol. Ecol. 175, 181–195.
Claassen, Cheryl, 1986. Temporal patterns in marine shellfish-species use along the
Atlantic Coast in the Southeastern United States. Southeast. Archaeol. 5 (2), 120–137.
Claassen, Cheryl, 1998. Shells. In: Cambridge Manual in Archaeology. Cambridge
University Press.
Claassen, Cheryl, 2000. Quantifying shell: comments on Mason, Peterson, and Tiffany.
Am. Antiq. 65 (2), 415–418.
Crumley, Carole L., 1987. Historical Ecology. In: Crumley, Carole L., Marquardt, William
H. (Eds.), Regional Dynamics: Burgundian Landscapes in Historical Perspective.
Academic Press, San Diego, CA, pp. 237–264.
deFrance, Susan D., Walker, Karen J., 2013. The Zooarchaeology of Pineland. In:
Marquardt, William H., Walker, Karen J. (Eds.), The Archaeology of Pineland: A
Coastal Southwest Florida Site Complex, A.D. 50–1710. Institute for Archaeology and
Paleoenvironmental Studies, Monograph 4. University of Florida, Gainesville, pp.
305–348.
Duke, C. Trevor, 2015. Identifyinge Humanized Ecosystems: Anthropogenic Impacts,
Intentionality, and Resource Acquisition at Crystal River (8CI1) and Robert's Island
(8CI41). Department of Applied Anthropology, University of South Florida, Tampa,
Master's Thesis.
Erlandson, Jon M., Rick, Torben C., Braje, Todd J., Steinberg, Alexis, Vellanoweth, Rene
L., 2008. Human impacts on ancient shellfish: a 10,000 year record from San Miguel
Island, California. J. Archaeol. Sci. 35, 2144–2152.
Fewkes, Jesse W., 1924. Preliminary archaeological explorations at Weeden Island,
Florida. Smithson. Misc. Collect. 76, 1–26.
Gould, Stephen J., 1966. Allometry and size in ontogeny and phylogeny. Biol. Rev. Camb.
Philos. Soc. 41, 587–640.
Gunn, Joel D., 1994. Introduction: a perspective from the humanities-science boundary.
Hum. Ecol. 22 (1), 1–22.
Harris, Matthew, Weisler, Marshall, 2017. Prehistoric human impacts to marine mollusks
and intertidal ecosystems in the Pacific Islands. J. Island Coast. Archaeol. 0, 1–21.
Hathaway, Ralph R., Woodburn, K.D., 1961. Studies on the crown conch (Melongena
corona Gmelin). Bull. Mar. Sci. Gulf Caribb. 11 (1), 45–65.
Jackson, Kendal R., 2016. The Archaeopalynology of Crystal River Site (8CI1), Citrus
115
Journal of Archaeological Science: Reports 21 (2018) 107–116
County, Florida. Master's Thesis. Department of Anthropology, University of South
Florida, Tampa.
Janus Research, 2004. Cultural resource assessment survey of the evergreen avenue
project area, Pinellas County. In: Report Prepared for Pinellas County Department of
Public Works, (Clearwater, Florida).
Keegan, William F., Portell, Roger W., Slapcinsky, John, 2003. Changes in invertebrate
taxa at two pre-Columbian sites in Southwestern Jamaica, AD 800-1500. J. Archaeol.
Sci. 30, 1607–1617.
Lambert, Jeanne, 2006. Coastal Processes and Anthropogenic Factors Influencing the
Geomorphic Evolution of Weedon Island, Florida. Masters Thesis. Department of
Environmental Science and Policy, University of South Florida, Tampa.
Larson, L., 1980. Aboriginal Subsistence Technology on the Southeastern Coastal Plain
during the Terminal Late Prehistoric Period. University of Florida Press, Gainesville.
Lulewicz, Isabelle H., Thompson, Victor, Cramb, Justin, Tucker, Bryan, 2017. Oyster
paleoecology and native American subsistence practices on Ossabaw Island, Georgia,
USA. J. Archaeol. Sci. Rep. 15, 282–289.
Lyman, Richard L., 1979. Available meat from faunal remains: a consideration of tech-
niques. Am. Antiq. 44, 536–546.
Marquardt, William H., Kozuch, Laura, 2016. The lightning whelk: an enduring icon of
southeastern North American spirituality. J. Anthropol. Archaeol. 42, 1–26.
Mikell, Gregory A., Saunders, Rebecca, 2007. Terminal middle to late archaic settlement
in coastal Northwest Florida: early estuarine exploitation on the Northern Gulf Coast.
Southeast. Archaeol. 26 (2), 169–195.
O'Donnell, Sharlene K., 2015. The Weedon Island Seascape: Zooarchaeological Findings
at Weedon Island Archaeological Site (8PI1). Master's Thesis Paper, on file.
University of South Florida Archaeology Laboratory, St. Petersburg.
O'Donnell, Sharlene K., 2018. Discovering the seascape at Weedon Island during the
safety harbor period (CA. A.D. 900 to A.D. 1725): a zooarchaeological analysis at the
Weeden Island Site (8PI1). Fla. Anthropol In Press.
Onat, A.R.B., 1985. The multifunctional use of shellfish remains: from garbage to com-
munity engineering. Northw. Anthropol. Res. Notes 19, 201–207.
O'Neal, Lori, 2016. What's in Your Toolbox? Examining Tool Choices at Two Middle and
Late Woodland-Period Sites on Florida's Central Gulf Coast. Master's Thesis.
Department of Applied Anthropology, University of South Florida, Tampa.
Palmer, Richard A., 1982. Growth in marine gastropods: a non-destructive technique for
independently measuring shell and body weight. Malacologia 23 (1), 63–73.
Palmiotto, Andrea, 2016. Indicator groups and effective seasons on the coast: zooarch-
aeology of fish in the lower Suwannee region Florida. J. Archaeol. Sci. Rep. 7,
330–343.
Perez, Tanya M., 2010. Methodological issues in zooarchaeology. In: VanDerwarker,
A.M., Perez, Tanya M. (Eds.), Integrating Zooarchaeology and Paleoethnobotany: A
Consideration of Issues, Methods, and Cases. Springer Science and Business Media,
LLC.
Peters, Robert H., 1983. The ecological implications of body size. Cambridge University
Press, New York.
Pilsbry, H.A., Vanatta, E.G., 1934. Melongena corona and its races. The Nautilus 47 (4),
117–121.
Pluckhahn, Thomas J., Thompson, Victor D., Jack Rink, W., 2016. Evidence for stepped
pyramids of shell in the woodland period of eastern North America. Am. Antiq. 81
(2), 345–363.
Quitmyer, Irvy R., 2002. Faunal Analysis of Remnant Mound Column Samples. In:
Schwadron, Margo (Ed.), Archaeological Investigation of DeSoto National Memorial.
Southeastern Archaeological Center Technical Reports 8, Tallahassee, pp. 158–202.
Reed, Charles A., 1963. Osteo-archaeology. In: Higgs, E.S., Brothwell, D. (Eds.), Science in
Archaeology. Thames and Hudson, New York, pp. 204–216.
Reitz, Elizabeth J., 1988. Evidence for coastal adaptations in Georgia and South Carolina.
Archaeol. East. N. Am. 16, 137–158.
Reitz, Elizabeth J., Wing, Elizabeth S., 2008. Zooarchaeology. In: Cambridge Manual in
Archaeology. Cambridge University Press.
Reitz, Elizabeth J., Quitmyer, Irvy R., Stephen, Hale H., Scudder, Sylvia J., Wing,
Elizabeth S., 1987. Application of allometry to zooarchaeology. Am. Antiq. 52 (2),
304–317.
Rick, Torben C., Reeder-Myers, Leslie A., Hofman, Courtney A., Breitburg, Denise,
Lockwood, Rowan, Henkes, Gregory, Kellogg, Lisa, Lowery, Darrin, Luckenbach,
Mark W., Mann, Roger, Ogburn, Matthew B., Southworth, Melissa, Wah, John,
Wesson, James, Hines, Anson H., 2016. Millennial-scale sustainability of the
Chesapeake Bay native American oyster fishers. PNAS 10, 1073.
Sassaman, Kenneth E., Wallis, Neil J., McFadden, Paulette S., Mahar, Ginessa J., Jenkins,
Jessica A., Donop, Mark C., Mones, Micah P., Palmiotto, Andrea, Boucher, Anthony,
Goodwin, Joushua M., Oliveira, Cristina I., 2017. Keeping pace with rising sea: the
first 6 years of the lower Suwannee archaeological survey, gulf coastal Florida. Island
Coast. Archaeol. 12, 173–199.
Savarese, Michael, Walker, Karen J., Stingu, Shanna, Marquardt, William H., Thompson,
Victor, 2016. The effects of shellfish harvesting by aboriginal inhabitant of Southwest
Florida (USA) on productivity of the eastern oyster: implications for estuarine man-
agement and restoration. Anthropocene 16, 28–41.
Schwadron, Margo, 2017. Shell Works: Prehistoric Terraformed Communities of the Ten
Thousand Islands, Florida. 3(1). Hunter Gatherer Research, pp. 31–63.
Sears, William H., 1971. The Weeden Island Site, St. Petersburg, Florida. Fla. Anthropol.
24 (2), 51–60.
Smith, Bruce D., 1975. Toward a more accurate estimation of meat weight yield of animal
species at archaeological sites. In: Clason, A.T. (Ed.), Archaeozoological Studies.
American Elsevier, New York, pp. 99–106.
Speck, F., 1946. Rappahannock Taking Devices: Traps, Hunting, and Fishing. Joint
Publication of the Museum of the University of Philadelphia Anthropological Society,
Number 1, Philadelphia.
K. Jackson et al.
Thomas, Kenneth D., Mannino, Marcello A., 2017. Making numbers count: beyond
minimum numbers of individuals (MNI) for the quantification of mollusc assemblages
from shell matrix sites. Quat. Int. 427, 47–58.
Trubitt, Mary Beth, 2005. Crafting marine shell prestige goods at Cahokia. North Am.
Archaeol. 26 (3), 249–266.
Turner, R.D., 1959. Notes of the feeding of Melongena corona. The Nautilus 73, 11–13.
Vojnovski, Pamela K., 1995. Zooarchaeology. In: Austin, Robert J. (Ed.), Yat Kitischee: A
Prehistoric Coastal Hamlet, 100 B.C.–A.D. 1200. Florida Department of Historical
Resources, Tallahassee, pp. 56–75 Report on File.
Vojnovski, Pamela K., 1998. Zooarchaeology. In: Simpson, Terrance L. (Ed.), The
Narvaez/Anderson Site (8Pi54): A Safety Harbor Culture Shell Mound and Midden
AD 1000–1600. Central Gulf Coast Archaeological Society, St. Petersburg, pp.
252265.
Walker, Karen J., 1992. The Zooarchaeology of Charlotte Harbor's Prehistoric Maritime
Adaptation: Spatial and Temporal Perspectives. In: Maquardt, William H. (Ed.),
Culture and Environment in the Domain of the Calusa. Institute of Archaeology and
Paleoenvironmental Studies, Monograph 1. University of Florida, Gainesville, pp.
265–366.
Walker, Karen J., 2000. The material culture of Precolumbian fishing: artifacts and fish
116
Journal of Archaeological Science: Reports 21 (2018) 107–116
remains from coastal Southwest Florida. Southeast. Archaeol. 19 (1), 24–45.
Walker, Karen J., 2013. Excavations and Chronostratigraphy at the Pineland Site
Complex: 1988–1995. In: Marquardt, W.H., Walker, K.J. (Eds.), The Archaeology of
Pineland: A Coastal Southwest Florida Site Complex, A.D. 50–1710. Institute of
Archaeology and Paleoenvironmental Studies, Monograph 4. University of Florida,
Gainesville, pp. 53–154.
Waselkov, Gregory A., 1987. Shellfish gathering and shell midden archaeology. In:
Schiffer, Michael (Ed.), Advances in Archaeological Method and Theory. 11.
Academic Press, Orlando, pp. 93–210.
Weisman, Brent R., Dean, J., O'Brien, M., Collins, L., 2005. Comprehensive Cultural
Resource Survey of the Weedon Island Preserve. Report on file. Florida Division of
Historical Resources, Tallahassee (Revised 2008).
White, Theodore A., 1953. A method of calculating the dietary percentage of various food
animals utilized by aboriginal peoples. Am. Antiq. 38, 266–278.
Wolfe, Steven H., 1990. An ecological characterization of the Florida Springs Coast:
Pithlachascotee to Waccasassa rivers. U.S. Fish Wildl. Serv. Biol. Rep. 90 (21), 27–44.
Ziegler, Alan C., 1973. Inference from prehistoric faunal remains. Addison Wesley
Module in Anthropology No. 43. .