Beruflich Dokumente
Kultur Dokumente
Making the Most of Fragments: A Method for Estimating Shell Length From
Fragmentary Mussels (M. californianus and M. trossulus) on the Pacific Coast of
North America
Gerald G. Singh, Iain McKechnie
PII:
S0305-4403(15)00067-9
DOI:
10.1016/j.jas.2015.02.029
Reference:
YJASC 4359
To appear in:
17 February 2015
Please cite this article as: Singh, G.G., McKechnie, I., Making the Most of Fragments: A Method for
Estimating Shell Length From Fragmentary Mussels (M. californianus and M. trossulus) on the Pacific
Coast of North America, Journal of Archaeological Science (2015), doi: 10.1016/j.jas.2015.02.029.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
AC
C
EP
TE
D
M
AN
U
SC
RI
PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
M
AN
U
SC
RI
PT
TE
Gerald G. Singh
1
Institute for Resources, Environment and Sustainability
University of British Columbia, Vancouver, BC, Canada, V6T 1Z4
Email: geraldsingh@gmail.com
AC
C
EP
Iain McKechnie*
2
Department of Anthropology, 308 Condon Hall, University of Oregon, Eugene, OR,
97403 USA
3
Department of Archaeology, Hakai Institute, Simon Fraser University, Burnaby, BC,
V5A 1S6 Canada
email: iainm@uoregon.edu
* Corresponding author to whom correspondence should be addressed.
ACCEPTED MANUSCRIPT
ABSTRACT
archaeological sites throughout western North America but are often highly fragmentary when
contribution of Mytilus but most studies assume an average meat weight or use categorical size
classifications to determine subsistence strategies and harvest profiles. In this paper, we develop
and evaluate a regression-based method for estimating shell length and meat weight for
SC
RI
PT
specimens from multiple locations in California and British Columbia and provide considerable
11
statistical confidence for predicting length and meat weight. We also apply the same regressions
12
to a collection of M. trossulus and show similar predictive equations, indicating this method can
13
14
and M. trossulus. We demonstrate how these results improve upon previous size-classification
15
methods and discuss the potential for applying these measurements to enhance the relevance of
16
these zooarchaeological data for modern marine conservation and management efforts.
M
AN
U
10
17
18
KEYWORDS
20
21
22
HIGHLIGHTS
23
24
Regressions are based on live-collected shells from California and British Columbia.
25
26
Strong predictions for length and meat weight from intact umbos can be generated.
27
These improved length estimates extend the interpretive utility of Mytilus assemblages.
28
AC
C
EP
TE
19
ACCEPTED MANUSCRIPT
1. Introduction
30
The California mussel (Mytilus californianus) is a ubiquitous and protein rich shellfish species
31
found in archaeological sites throughout the coast of western North America (e.g., Braje et al.
32
2007; Erlandson 1988; Losey and Power 2005; Porcasi 2011; Sumpter 2005). Mussels are
33
present from the mid-to lower intertidal on wave-exposed rocky coasts from Alaska to Baja
34
California (Schmidt 1999; Suchanek 1979). Judging from the ubiquity and abundance of M.
35
36
of it use, this species was a widely valued marine food regularly harvested and consumed by
37
coastal indigenous people throughout the Holocene (Braje et al. 2012; Ellis and Swan 1981;
38
Erlandson et al. 2008; Jones and Richman 1995; Whitaker 2008). Mussels are considered
39
ecosystem engineers in intertidal ecosystems, providing habitat for many invertebrates and
40
outcompeting intertidal seaweeds (Menge et al. 1994; Paine 1974). M. californianus can be long-
41
lived (>20-50 years) and grow to large sizes (over 20 cm in length) and thus can be
42
43
respond to a variety of factors such as climate and oceanography (Menge et al. 2008; Smith et al.
44
2006), temperature (Ford et al. 2010; Harley 2011), and intertidal predation by other marine
45
predators such as sea stars (Pisaster sp.) and sea otters (Enhydra lutris) (Menge et al. 1994;
46
Singh et al. 2013). Given the ecological and economic importance of mussels, further
47
investigation into the characteristics of archaeological assemblages has the potential to enrich
48
TE
M
AN
U
SC
RI
PT
29
EP
49
51
to detect nutritional, cultural, and ecological patterns and changes over time (Cook 1946; Moss
52
1993; Nelson 1909; Shawcross 1967; Swadling 1976; Waselkov 1987). On the Pacific Coast of
53
54
occurred in California where analysis has focused on the assessment of dietary importance
55
(Erlandson 1988; Glassow and Wilcoxon 1988; Greengo 1951; Treganza and Cook 1948),
56
harvesting strategies (Jones and Richman 1995; Whitaker 2008), and seasonality (Eerkens et al.
57
2013; Jew et al. 2014; Jones et al. 2008). Archaeologists have additionally utilized
58
archaeological Mytilus sp. for informing environmental and sea level changes (Graham et al.
59
2003) and the impact of long-term sustained human harvests (Botkin 1980; Braje et al. 2012).
AC
C
50
ACCEPTED MANUSCRIPT
60
In contrast, shellfish research on the Northwest Coast has rarely focused on M. californianus
62
despite its ubiquity in archaeological assemblages (Cannon et al. 2008; Clarke and Clarke 1980;
63
Croes 2005; Orchard 2009; Sumpter 2005; Wessen 1988). Rather, the bulk of shellfish research
64
has focused on clams, specifically the seasonality of clam harvesting (e.g., Burchell et al. 2013;
65
Croes 2005:111; Ham and Irvine 1975; Keen 1979; Maxwell 2003; Wessen 1982:144).
RI
PT
61
66
Mussels can exhibit wide size variation that is nutritionally and ecologically meaningful.
68
However, the occurrence of whole shellfish is rare in most depositional settings (Faulkner 2011;
69
Muckle 1985; Wolverton et al. 2010) and this is particularly the case for Mytilus (Ford 1992;
70
Glassow 2000). Mytilus survivorship may be further influenced by excavation and post-
71
excavation sample processing (e.g., wet/dry screening, sorting, and curation). Measurements of
72
complete whole shells is possible given a targeted recovery strategy undertaken during
73
74
75
shells is that larger shells may be more robust and thus preserve more readily. Conversely, it is
76
also possible that smaller shells, with less surface area and or differing shapes, may be less likely
77
to fracture from compaction and trampling (Muckle 1985; Wolverton et al. 2010).
M
AN
U
SC
67
TE
78
While a number of studies have developed regression-based methods for evaluating the size
80
distributions of marine and freshwater mussels (Buchanan 1985; Campbell 2014; Hall 1980;
81
Parmalee and Klippel 1974; Randklev et al. 2009), to our knowledge, the Pacific coast of North
82
America currently lacks a regression-based method for determining M. californianus shell length
83
and no method has estimated meat-weight from fragmentary remains (Bell 2009:25; but see Ford
84
85
86
contributions from average sized mussels (Cook 1946; Croes and Hackenberger 1988:36;
87
Kennedy 2004; Mitchell 1988; Moss 1989). In a few cases, length measurements obtained from
88
whole shells have been reported (Braje et al. 2012; Croes 2005:110; Erlandson et al. 2008) but
89
these remain rare relative the more commonly quantified assemblages focused on bulk or column
90
AC
C
EP
79
91
ACCEPTED MANUSCRIPT
92
94
from intensively quantified small-volume bulk samples are fragmentary which constrains the
95
ability to generate robust harvest profiles from these painstakingly calculated proportional data.
96
This circumstance also limits the accuracy of derived secondary information, such as estimates
97
of meat weight calculated from non-repetitive element counts (NRE) (Giovas 2009; Mason et
98
al. 1998) or the affect that shell size may have on the proportional frequency of weighed shell
99
RI
PT
93
100
Some researchers have provided estimates of Mytilus shell length using size categories to
102
construct cumulative harvest profiles to identify size selectivity (Wessen 1982:151). For
103
104
method developed by White (1989) and Jones and Rickman (1995) to document the harvest
105
profile for fragmentary mussels from a rockshelter assemblage in Northern California. Whites
106
(1989) method involves comparing archaeological mussels to traced outlines of 2-cm length
107
108
template with traced outlines of a particular size. Researchers visually match the valve shape and
109
size that best fits the template. The resulting size assessment is used to categorize archaeological
110
valves and then analyzed in a cumulative harvest profile where it can distinguish differences in
111
prey selectivity (e.g., size selective plucking or indiscriminant stripping and variations in
112
between).
EP
113
TE
M
AN
U
SC
101
Bell (2009) subsequently evaluated this classification method and demonstrated considerable
115
statistical uncertainty in this approach, particularly when applied to fragmentary valves, which
116
117
estimation of total shell length based on morphometric relationships from fragmentary remains
118
as has been demonstrated for mussels elsewhere (Campbell 2014; Hall 1980; Randklev et al.
119
2009).
AC
C
114
120
121
In this study, we develop and evaluate a morphometric regression method for deriving total
122
length estimates from incomplete M. californianus shells. Our method offers a variety of
123
measurements that predict the length of fragmentary shells with considerable statistical
ACCEPTED MANUSCRIPT
confidence. Regressions are based on live mussels collected from multiple areas of California
125
and British Columbia and thus relevant for archaeological application across a broad coastal
126
region. Given these regions are subject to varying climatic and oceanographic conditions,
127
incorporating this variability helps the resulting regressions better account for differential
128
growing conditions and climatic changes. We conclude this improves the basis for investigating
129
harvest profiles, measures of Mytilus biomass represented archaeologically, and the historical
130
131
SC
132
RI
PT
124
To derive and evaluate biometric relationships, 132 live M. californianus were collected from
134
numerous intertidal plots within three exposed coastal embayments on western Vancouver
135
Island, British Columbia, Canada as well as ecological collections from Northern and Southern
136
California, USA (Figure 1, supplemental materials). Mussels were obtained from the mid-to-low
137
tide on wave-exposed rocky shores as part of a marine ecology research project seeking to
138
control for growth variation along this latitudinal gradient (Singh 2010; Singh et al. 2013). This
139
broad spatial scale is relevant as it accounts for multiple environmental and growing conditions
140
across space and therefore serves as a proxy for variability in growing conditions across broad
141
temporal scales (e.g., Pickett 1989). Climatic variables, especially temperature and tidal
142
elevation, has a strong effect on mussel growth rate (Harley 2011; Menge et al. 2008; Seed and
143
Suchanek 1992; Suchanek 1981) and different growth conditions can lead to differences in size
144
as well as variation in shell morphology. Thus, our regressions aim to incorporate uncertainty
145
across this large region as such uncertainty is likely to occur over archaeological time scales.
TE
EP
AC
C
146
M
AN
U
133
147
For each individual mussel used to derive the regression equations, we measured a series of
148
points from the umbo to morphologically distinctive markers on the shell (e.g., the hinge teeth on
149
the inside of the shell) (Figure 2). We measured these attributes because they form the most
150
robust portion of the shell and are more readily recovered in small volume archaeological
151
152
ACCEPTED MANUSCRIPT
The umbo region of mussels is also the oldest site of growth (Ford et al. 2010; Seed 1968). From
154
this point, individuals add layers via secretions from the mantle to the inner side of the shell
155
(from the umbo to the distal end of the shell) (Checa 2000). The robusticity and thickness of the
156
umbo increases as the mussel ages while the ligament (hinge) connecting the two valves of the
157
shell also increases in length as the mussel increases in size. Despite the potential for allometric
158
variation in growth within a particular habitat, these characteristics make the measurements
159
RI
PT
153
160
To maximize measurements obtained from a given archaeological assemblage (e.g., fine and
162
coarse mesh fractions), we measured multiple dimensions on individual mussel shells to account
163
for shells in different states of fragmentation (Figure 2). Although this method yields regressions
164
with differing predictive strengths, these uncertainties can be quantified. A particular benefit is
165
that the umbo is used to derive minimum number of individuals (MNI) or non-repetitive
166
elements (NRE), which are especially prevalent in small mesh sizes (Glassow 2000:410). Thus
167
deriving length estimates from these fragmentary specimens can considerably strengthen the
168
analytical potential of these rigorously quantified assemblages and the ecological relevance of
169
M
AN
U
SC
161
170
We use simple ordinary least squares regression to build the predictive models (Table 1). For
172
fragment samples that are large enough to take multiple measures, we also provide multiple
173
regression model structures for incorporating multiple measures into making an estimate. These
174
multiple regression models explain variation slightly better than the simple models (as indicated
175
by the R2 values in Table 2) and can offer an extra level of confidence in the estimation for
176
EP
AC
C
177
TE
171
178
Our linear regression models are related to total shell length and the predictive confidence is a
179
180
additionally calculated dry meat weight for each specimen (distinguishing gonads and non-
181
reproductive tissue). Thus, we also were able to derive a strong shell length-meat weight
182
relationship, which can be applied to a derive minimum weight estimates from individual
183
ACCEPTED MANUSCRIPT
184
186
these same measurements on 68 M. californianus valves recovered from the upper intertidal
187
wrack zone in two locations in British Columbia (Barkley Sound and Calvert Island, Figure 1,
188
supplemental materials). These beach-collected shells represent natural mortality events induced
189
by wave action during storms, following years (or at least several months) of weathering and
190
erosion as shells are transported to upper intertidal beach deposits. These weathering and wave
191
induced erosional processes can be considered broadly similar to the taphonomic processes
192
affecting archaeological shells over broader time scales (e.g., trampling, erosion of prominent
193
surfaces etc.) (Muckle 1985). While not directly analogous to other post depositional effects such
194
as leaching (Stein 1996), we used the simple regression models to estimate total size of these
195
beach collected shells. This was done in order to assess model performance on weathered shells
196
as well as to evaluate the accuracy of measurements obtained on archaeological shells that may
197
have been subject to weathering and wave erosion following depositional in nearshore contexts
198
SC
M
AN
U
199
RI
PT
185
201
the Bay mussel (M. trossulus, formerly referred to as M. edulis) and it is often difficult to
202
distinguish the two taxa based on small fragments. Thus, many archaeological analyses
203
cautiously lump all Mytilus specimens into a single category (even umbo fragments) unless local
204
environmental conditions preclude the presence of either. To assess the potential biometric
205
206
207
Island in the northern Salish Sea (Figure 1, supplemental materials). This inland sea is situated
208
far from the exposed Pacific Coast in habitat unsuitable for M. californianus. Unlike the wave-
209
transported beach collected shells described above, M. trossulus valves were collected at low-
210
low tide from a very protected beach with a lack of wave exposure (Heriot Bay, BC). These
211
shells exhibited a lack of weathering and erosion and appeared to be recently deceased as they
212
were obtained directly below piers or elevated rocky crevices on which live clusters of M.
213
trossulus were growing. The simple regression models developed for M. californianus were
214
applied to these M. trossulus shells to determine how well they perform on these
215
AC
C
EP
TE
200
ACCEPTED MANUSCRIPT
216
218
regressions (e.g., Lyman and VanPool 2009), we conducted double-blind trials that compared
219
differences in the estimates of total mussel length between multiple participants (Table 1). Using
220
a separate population of mussel shells from those used to develop our relationships or our beach
221
collected samples (n=54), we recorded total lengths of individual shells identified with numeric
222
codes and subjected them to destructive fragmentation. Individual umbo fragments obtained
223
from this crushed assemblage were then given to multiple participants to evaluate both our
224
regression method as well as the size classification method developed by White (1989).
225
Accuracy of size predictions was evaluated by comparing predicted lengths with known lengths,
226
with average differences between these measures and average inter-observer differences reported
227
M
AN
U
SC
RI
PT
217
228
3. Results
230
All regressions based in reference to the umbo showed a strong predictive relationship,
231
indicating a very strong fit of the model to the data and thus a strong relationship between the
232
linear measures and the total shell length (Table 1). The R2 values of the regressions range from
233
0.80 to 0.93, meaning that between 80 and 93% of the variability in mussel lengths is accounted
234
for by the linear regression model, indicating that these models are good predictors. These linear
235
236
conditions. We note slight differences in rates of growth, with Southern California being the
237
most distinctly offset (i.e., green dots in Figure 3). However, given that we are interested in
238
239
TE
EP
AC
C
240
229
241
We also noted that the relationship between total shell length and the dry weights of gonads,
242
non-reproductive body, and shell offer strong relationships, albeit nonlinear (Figure 5). Linear
243
models are more likely to systematically underestimate weight for both small and large shells as
244
well as potentially yield negative weight values for the smallest shells (i.e., there is no negative
245
intercept in our non-linear model). We found that our models explained between 78 and 90% of
246
the variation in body mass, depending on tissue type (reproductive and non-reproductive), while
ACCEPTED MANUSCRIPT
247
relating total length to total dry meat, the predictive model for weight accounts for 90% of the
248
249
250
252
power, particularly between estimated length and total shell length. Similarly, our comparison of
253
the difference between estimated shell lengths and actual shell lengths produced by Whites
254
(1989) template indicates an average error of 16 3 mm (Figure 4). Whites template was unable
255
to estimate mussel shells above 100 mm because the template categories did not go above this
256
range. This is indicated in Figure 4a by the wide distributions of actual shell lengths estimated in
257
the 90-100 mm range, though many are larger than 100 mm. Comparatively, the average error
258
between the predicted and actual shell lengths range from 7.3-12.8 mm across shell sizes in the
259
double blind trials using our regression models (Table 1, Figure 4b-e). The model based on the
260
measure from the umbo to the outer hinge teeth was the most precise (Figure 4b), having an
261
average error of 7.3 mm. Our double-blind results demonstrate that there is considerable inter-
262
observer variability between researchers measuring the same specimens using Whites method,
263
which further confounds its predictive power (Figure 4a). Our results indicate the accuracy of
264
Whites method, on average across all participants, was 9.5 1.5 mm between analysts. By
265
contrast, the precision of our estimates, using the umbo thickness model (a modestly performing
266
regression), averaged between all participants, indicates that measurements were off by 5.1 2.3
267
mm between analysts.
SC
M
AN
U
TE
EP
268
RI
PT
251
The simple regression models predicted total length of weathered, beach collected shells with
270
similar accuracy compared to the fresh shells, with mean error between 5.4 and 15.4 mm from
271
the actual lengths of shells (Table 4). Models based on the length from the umbo to outer hinge
272
teeth, length of hinge teeth, and umbo thickness were slightly more accurate for beach collected
273
shells, and the model for umbo to inner hinge teeth was slightly more accurate for fresh shells,
274
but in every case the error estimates had 95% confidence intervals that overlapped, indicating no
275
significant difference in performance between fresh shells and beach collected shells.
AC
C
269
10
ACCEPTED MANUSCRIPT
276
278
applied to M. trossulus indicates that these models predict total length of shells as accurately as,
279
and in some instances, more accurately for M. trossulus (Table 4). The models based on umbo
280
to outer hinge teeth and hinge teeth length performed better for M. trossulus, and the non-
281
overlapping 95% confidence intervals indicate that they perform significantly better than for live
282
collected M. californianus shells, though the 95% confidence intervals overlap with the beach
283
collected shells.
284
4. Discussion
285
Our morphometric regressions provide a more accurate method for confidently estimating the
286
total shell length from mussel umbo fragments relative to the size-class method utilized by
287
Wessen (1982), White (1989), and applied by Whitaker (2008) and others. This is further
288
apparent in a reduction in the variability in length estimates between analysts and an expansion
289
of length estimates of M. californianus on continuous interval scale across their size spectrum
290
(including mussels >200 mm) in contrast to Whites template which appears limited to Mytilus
291
less than 100 mm (White 1989:133). Additionally, incorporating shells from multiple geographic
292
regions subjected to multiple growing conditions is arguably more relevant for application to
293
M
AN
U
SC
RI
PT
277
TE
294
These models offer more reliable estimates for meat-weight for M. californianus than previous
296
research utilizing average weights and calorie estimates (Croes and Hackenberger 1988:36;
297
Erlandson 1988:104). Due to the high rate of fragmentation for Mytilus (Glassow 2000; Muckle
298
1994), these methods likely also provide a much more representative range of size distributions
299
than measurements of comparatively rare whole Mytilus valves suitable for measurement
300
(Erlandson et al. 2008). Measured shell umbo fragments can provide estimates of the total
301
length, as well as the dry weight (differentiated between tissue types). Combined with
302
303
elements (NRE) (Giovas 2009), these models provide the opportunity to refine estimates of
304
meat-weight and Mytilus biomass from archaeological samples. Given the good performance of
305
the length models on M. trossulus as well, it is possible to derive meat weight estimates from
306
fragmentary remains using available shell length to meat weight regressions for M. trossulus
AC
C
EP
295
11
ACCEPTED MANUSCRIPT
307
(e.g., Penney et al. 2008), and thus generate species-specific bounded estimates of meat weight
308
per sample.
309
Surprisingly, our models are similarly accurate in estimating total length for moderately beach-
311
weathered mussel shells, and so still retain strong predictive capacity. This could reflect the
312
robustness of this portion of the shell, and indicate that the morphometric relationships are less
313
affected by taphonomic weathering (i.e. erosion does not strip away shell thickness on the scale
314
315
measurement of archaeological Mytilus assemblages that may have been subject to modest post-
316
depositional wave erosion retain their predictive accuracy. The models also showed strong
317
potential for estimating total length of M. trossulus shells. In some cases the models performed
318
better for smaller mussels, and this could be due to the fact that M. trossulus has a notably
319
narrower growth range and/or their size distribution falls in a zone where more samples fit the
320
model (and therefore the areas of the model that are more certain). This result highlights the fact
321
that both Mytilus species have very similar morphology and morphometric growth relations, and
322
therefore estimates can be combined in cases of taxonomic uncertainty. Overall, the favorable
323
performance of the models across a wide range of specimens (multiple species, various stages of
324
erosion, and locations of collection) highlight the broad applicability of the models and the
325
TE
M
AN
U
SC
RI
PT
310
326
This method is not devoid of uncertainty but represents a significant improvement from size-
328
class data. Based on our trials, our models estimated the maximum error in total lengths as high
329
as 35 mm different from the actual length in individual cases but this is compared to an error of
330
67 mm using Whites (1989) template. Even so, on average, our results indicate that our models
331
produce more certain results based on average error (7.3-12.8 mm with our models versus 15.8
332
mm with the template). More importantly, a regression based approach provides the ability to
333
transform shell length estimates from categorical data (binned size classes) into continuous data
334
therefore increasing the analytical utility through improved quantification of uncertainty and
335
comparison across multiple datasets (e.g., layers within sites, between individual sites, regional
336
trends).
AC
C
EP
327
337
12
ACCEPTED MANUSCRIPT
By sampling over a broad geographic area of the eastern Pacific coast, we have sought to
339
incorporate climatic variation in our models. However, mussels from archeological samples
340
remain likely to have grown in conditions outside what we have sampled for this study (Harley
341
2011). Future research may reduce the uncertainty in our models and/or create regionally specific
342
343
improving this morphometric approach by making their primary measurement data available as
344
RI
PT
338
345
Which relationship to use will largely depend on how much of the shell remains intact. The
347
model using the measurement from the umbo to the outer hinge teeth had the strongest fit (R2 =
348
0.93) but requires most of the shell to be intact. The model that accommodates the most
349
fragmentary shells (umbo thickness), still yields an estimate of shell length with moderately
350
strong predictive power (R2 = 0.80, Table 1). This is the measure that will likely be most useful
351
for archaeological applications given this is the thickest part of the shell and mostly likely to
352
preserve. While this may only represent a modest improvement in accuracy from existing size
353
classification methods, it is capable of generating length estimates for a larger range of shell
354
sizes and is applicable to a greater range of fragmentary umbos. This method therefore enables a
355
larger number of measurements to be obtained for a given deposit with an greater predictive
356
accuracy. Moreover, if it is possible to obtain more than one measurement on a given shell,
357
multiple regression models can be used to further strengthen length and or meat-weight
358
predictions (Table 2). This latter approach might be of particular interpretive value for an
359
individual shell (e.g., an artifact) or for more confidently establishing an upper size threshold.
M
AN
U
TE
EP
360
SC
346
More important than the statistical power of individual measurements is the interpretive
362
confidence needed for a given research question. In this respect, existing size classification
363
methods may perform adequately for interpretations such as Whitakers (2008) comparison of
364
Mytilus harvest profiles in relation to idealized foraging theory models. However, such
365
measurement data may not be appropriate for statistical analysis that surpasses an ordinal scale
366
(Wolverton et al. 2014). The broader utility of our morphometric approach is an increase in
367
368
AC
C
361
369
13
ACCEPTED MANUSCRIPT
Collectively, the method can be used to more precisely measure Mytilus size distributions and
371
test for statistically significant differences at different points in time or between sites. Of
372
particular ecological relevance is evaluating the size distribution in relation to the long-term
373
presence of sea otters which have been shown to homogenize mussel sizes through regular
374
harvesting and elimination of larger and older M. californianus (Singh et al. 2013). Sea otters
375
(Enhydra lutris) are voracious consumers of mussels (VanBlaricom 1988), and the methods we
376
provide can potentially be used to evaluate if mussel size distributions were altered when sea
377
otter remains are found in temporally associated midden strata (e.g., Simenstad et al. 1978;
378
RI
PT
370
SC
379
Analysis of shell lengths can also be used to revisit the question of the Mytilus harvesting and
381
382
383
preferential size selection of larger mussels) versus a stripping strategy (indiscriminant removal
384
of a variety of size classes) (Jones and Richman 1995; Whitaker 2008). Alternatively, measuring
385
size distributions may also help characterize periods of change in ocean upwelling as reflected in
386
size distributions before, during, and after known climate changes (e.g., Kennett and Kennett
387
2000), the short term impacts of intensive human harvesting (Croes 1992; Thakar 2011), or the
388
potential restraint exercised in traditional clam management practices (Cannon and Burchell
389
2009; Daniels 2014; Lepofsky et al. in press). While our method provides improved
390
measurement capacity, disentangling such factors will require multiple lines of evidence and a
391
consideration of the causal agents that can produce similar size distributions.
TE
EP
392
M
AN
U
380
394
shellfish taxa elsewhere (e.g., Jerardino and Navarro 2008; Parmalee and Klippel 1974),
395
additional models for other species are needed on the Pacific Coast in order to refine estimates of
396
dietary content and biomass represented in shellfish assemblages. For instance, clams are another
397
pervasively utilized shellfish that are commonly abundant in shell midden settlements (Cannon et
398
al. 2008) and similar methods have begun to be variously developed for clams at specific sites
399
(Daniels 2014; Ford 1989:161), but this can usefully be expanded to include more species and
400
AC
C
393
14
ACCEPTED MANUSCRIPT
5. Conclusion
402
This paper has developed and evaluated a method for Mytilus shell length estimation for
403
application to archaeological and ecological research in western North America. This method
404
seeks to improve the precision of zooarchaeological data in order to enhance the interpretive
405
potential and historical ecological significance of archaeological shellfish assemblages (Rick and
406
Lockwood 2013). Future application of this low cost and relatively straightforward method has
407
the potential to increase the number of archaeological assemblages available for metrical
408
comparison and assessment, well beyond the small number of sites where Mytilus length
409
distributions have currently been obtained. Given the importance of shellfish in the
410
archaeological record along the eastern Pacific Coast and the long history of archaeological
411
analysis, improving the capacity to obtain metrical data from existing quantified assemblages
412
will provide for a more thorough integration of ecological and archaeological data. While the
413
method improves coastal archaeologists ability to address existing archaeological questions such
414
as dietary contribution and harvest and management strategies, it also brings zooarchaeological
415
information more directly into conversation with a variety of research in ecology, conservation
416
biology, and resource management (e.g., Erlandson and Rick 2010; Wolverton and Lyman
417
2012). Thus, rather than just another technically challenging and expensive analytical method
418
requiring reallocation of scarce funding resources, this study aims to increase the utility of data
419
derived from conventional shellfish quantification methods and strengthen the analytical effort of
420
TE
M
AN
U
SC
RI
PT
401
AC
C
EP
421
15
ACCEPTED MANUSCRIPT
422
423
Acknowledgements
424
We thank Russ Markel, Chris Harley, Anne Salomon, Dana Lepofsky, RG Matson, Nicole
426
Smith, Margot Hessing-Lewis, and Madonna Moss for their support and encouragement for this
427
project. We thank UC Santa Barbara, Bodega Bay Marine Labs, the Bamfield Marine Sciences
428
Centre, and the Hakai Institute for providing the opportunity to collect and measure Mytilus.
429
Particular thanks to Erin Rechsteiner for collecting and arranging delivery of surf washed shells
430
from Calvert Island. We thank Stefan Dick, Theraesa Coyle, and Gennifer Meldrum for helping
431
with the double blind trials. Thanks to Adrian Whitaker, Dale Croes, and Gary Wessen for
432
providing information during the preparation of this paper and to Steve Wolverton, and the
433
anonymous reviewers for improving this manuscript. Funding for portions of this research has
434
been enabled by the National Engineering and Science council of Canada (GGS), Social
435
Sciences and Humanities Research Council of Canada (IM), and the Hakai Institute for Coastal
436
M
AN
U
SC
RI
PT
425
437
AC
C
EP
TE
438
16
ACCEPTED MANUSCRIPT
References Cited
439
440
Bell, A.M., 2009. On The Validity of Archaeological Shellfish Metrics in Coastal California.
442
443
Botkin, S., 1980. Effects of Human Exploitation on Shellfish Populations at Malibu Creek,
444
California. in: Earle, T.K., Christenson, A.L. (Eds.), Modeling Change in Prehistoric Subsistence
445
446
Braje, T.J., Kennett, D.J., Erlandson, J.M., Culleton, B.J., 2007. Human Impacts on Nearshore
447
Shellfish Taxa: a 7000 Year Record from Santa Rosa Island, California. American Antiquity
448
72(4):735756.
449
Braje, T.J., Rick, T.C., Erlandson, J.M., 2012. A Trans-Holocene Historical Ecological Record
450
451
264(20):109120.
452
Buchanan, W.F., 1985. Middens and Mussels: an Archaeological Enquiry. South African Journal
453
of Science 81:1516.
454
Burchell, M., Hallmann, N., Martindale, A., Cannon, A., Schne, B.R., 2013. Seasonality and
455
Intensity of Shellfish Harvesting in the North Coast of British Columbia. Journal of Island and
456
457
Campbell, G.E., 2014. Size prediction in archaeomalacology: the Common Mussel, Mytilus
458
459
Cannon, A., Burchell, M., 2009. Clam Growth-stage Profiles as a Measure of Harvest Intensity
460
and Resource Management on the Central Coast of British Columbia. Journal of Archaeological
461
Science 36(4):10501060.
462
Cannon, A., Burchell, M., Bathurst, R., 2008. Trends and Strategies in Shellfish Gathering on the
463
Pacific Northwest Coast of North America. in: Antczak, A., Cipriani, R. (Eds.), Early Human
464
AC
C
EP
TE
M
AN
U
SC
RI
PT
441
17
ACCEPTED MANUSCRIPT
Checa, A., 2000. A new model for perisotracum and shell formation in Unionidae (Bivalvia,
466
467
Clarke, L.R., Clarke, A.H., 1980. Zooarchaeological Analysis of Mollusc Remains from Yuquot,
468
British Columbia. in: Folan, W., Dewhirst, J. (Eds.), The Yuquot Project, Volume 2, Parks
469
470
Cook, S.F., 1946. A Reconsideration of Shell Mounds with Respect to Population and Nutrition.
471
472
Croes, D.R., 1992. Exploring Prehistoric Subsistence Change on the Northwest Coast. in: Croes,
473
D., Rebecca Hawkins, and Barry L. Issac (Ed.), Research In Economic Anthropology, JAI Press
474
475
Croes, D.R., 2005. The Hoko River Archaeological Site Complex: The Rockshelter (45CA21),
476
1000--100 B.P., Olympic Peninsula, Washington, Washington State University Press, Pullman,
477
WA.
478
Croes, D.R., Hackenberger, S., 1988. Hoko River Archaeological Complex: Modeling
479
Prehistoric Northwest Coast Economic Evolution. in: Issac, B.L. (Ed.), Prehistoric Economies of
480
481
Daniels, P.S., 2014. Shellfish and Resource Sustainability on the Central Northwest Coast of
482
North America. in: Roksandic, M., de Souza, S.M., Eggers, S., Burchell, M., Klokler, D.M.
483
(Eds.), The Cultural Dynamics of Shell-Matrix Sites, University of New Mexico Press,
484
485
Eerkens, J.W., Byrd, B.F., Spero, H.J., Fritschi, A.M.K., 2013. Stable Isotope Reconstructions of
486
487
488
Ellis, D.W., Swan, L., 1981. Teachings of the Tides: Uses of Marine Invertebrates by the
489
490
Erlandson, J.M., 1988. The Role of Shellfish in Prehistoric Economies: A Protein Perspective.
491
AC
C
EP
TE
M
AN
U
SC
RI
PT
465
18
ACCEPTED MANUSCRIPT
Erlandson, J.M., Rick, T.C., 2010. Archaeology Meets Marine Ecology: The Antiquity of
493
Maritime Cultures and Human Impacts on Marine Fisheries and Ecosystems. Annual Review of
494
495
Erlandson, J.M., Rick, T.C., Braje, T.J., Steinberg, A., Vellanoweth, R.L., 2008. Human Impacts
496
on Ancient Shellfish: a 10,000 Year Record from San Miguel Island, California. Journal of
497
498
Faulkner, P., 2011. Quantifying shell weight loss in archaeological deposits. Archaeology In
499
Oceania 46:118129.
500
Ford, H.L., Schellenberg, S.A., Becker, B.J., Deutschman, D.L., Dyck, K.A., Koch, P.L., 2010.
501
502
503
504
4(2):157173.
505
Ford, P.J., 1992. Interpreting the Grain Size Distributions of Archaeological Shell. in: Stein, J.K.
506
(Ed.), Deciphering a Shell Midden, Academic Press, San Diego, pp. 283325.
507
Giovas, C.M., 2009. The Shell Game: Analytic Problems in Archaeological Mollusc
508
509
Glassow, M.A., 2000. Weighing vs. Counting Shellfish Remains: A Comment on Mason,
510
511
Glassow, M.A., Wilcoxon, L.R., 1988. Coastal Adaptations near Point Conception, California,
512
513
Graham, M.H., Dayton, P.K., Erlandson, J.M., 2003. Ice Ages and Ecological Transitions on
514
515
516
AC
C
EP
TE
M
AN
U
SC
RI
PT
492
19
ACCEPTED MANUSCRIPT
Hall, M., 1980. A method for Obtaining Metrical Data from Fragmentary Molluscan Material
518
519
Ham, L.C., Irvine, M., 1975. Techniques for Determining Seasonality of Shell Middens from
520
521
Harley, C.D.G., 2011. Climate Change, Keystone Predation, and Biodiversity Loss. Science
522
334(1124):11241127.
523
Jerardino, A., Navarro, R., 2008. Shell morphometry of seven limpet species from coastal shell
524
525
Jew, N.P., Erlandson, J.M., Rick, T.C., Reeder-Myers, L., 2014. Oxygen isotope analysis of
526
California mussel shells: seasonality and human sedentism at an 8,200-year-old shell midden on
527
528
Jones, T.L., Kennett, D.J., Kennett, J.P., Codding, B.F., 2008. Seasonal stability in Late
529
Holocene shellfish harvesting on the central California coast. Journal of Archaeological Science
530
35(8):22862294.
531
Jones, T.L., Richman, J.R., 1995. On Mussels: Mytilus californianus as a Prehistoric Resource.
532
533
Keen, S.D., 1979. The Growth Rings of Clam Shells From Two Pentlatch Middens, Heritage
534
535
536
Archaeological and Geochemical Evidence from Bodega Bay, California. PhD dissertation.
537
538
Kennett, D.J., Kennett, J.P., 2000. Competitive and Cooperative Responses to Climatic
539
540
Lepofsky, D., Smith, N.F., Cardinal, N., Harper, J., White, E.G., Salomon, A.K., Morris, M.,
541
Puckett, M., Rowell, K., in press. Ancient Shellfish Mariculture on the Northwest Coast of North
542
AC
C
EP
TE
M
AN
U
SC
RI
PT
517
20
ACCEPTED MANUSCRIPT
Losey, R.J., Power, E.A., 2005. Shellfish Remains from the Par-Tee Site (35-CLT-20), Seaside,
544
545
Lyman, R.L., VanPool, T.L., 2009. Metric Data in Archaeology: A Study of Intra-Analyst and
546
547
Mason, R.D., Peterson, M.L., Tiffany, J.A., 1998. Weighing vs. Counting: Measurement
548
Reliability and the California School of Midden Analysis. American Antiquity 63(2):303324.
549
Maxwell, D., 2003. Growth Coloration Revisted: Assessing Shellfish Seasonality in Coastal
550
British Columbia. in: Carlson, R.L. (Ed.), Archaeology of Coastal British Columbia: Essays in
551
Honour of Professor Philip M. Hobler, Archaeology Press, Simon Fraser Univeristy, Burnaby,
552
553
Menge, B.A., Berlow, E.L., Blanchette, C.A., Navarrete, S.A., Yamada, S.B., 1994. The
554
555
556
Menge, B.A., Chan, F., Lubchenco, J., 2008. Response of a rocky intertidal ecosystem engineer
557
558
Mitchell, D.H., 1988. Changing Patterns of Resource Use in the Prehistory of the Queen
559
Charlotte Strait, British Columbia. in: Issac, B.L. (Ed.), Prehistoric Economies of the Pacific
560
561
Moss, M.L., 1989. Archaeology and Cultural Ecology of the Prehistoric Angoon Tlingit. PhD
562
563
Moss, M.L., 1993. Shellfish, Gender and Status on the Northwest Coast: Reconciling
564
565
Anthropologist 95(3):631652.
566
567
AC
C
EP
TE
M
AN
U
SC
RI
PT
543
21
ACCEPTED MANUSCRIPT
Muckle, R.J., 1994. Differential Recovery of Mollusk Shell From Archaeological Sites. Journal
569
570
Nelson, N., 1909. Shellmounds of the San Francisco Bay region. University of California
571
572
Orchard, T.J., 2009. Otters and Urchins: Continuity and Change in Haida Economy during the
573
Late Holocene and Maritime Fur Trade Periods, British Archaeological Reports, Archaeopress,
574
Oxford.
575
Paine, R.T., 1974. Intertidal Community Structure: Experimental Studies on the Relationship
576
577
Parmalee, P.W., Klippel, W.E., 1974. Freshwater mussels as a prehistoric food resource.
578
579
Penney, R.W., Hart, M.J., Templeman, N.D., 2008. Genotype-dependent Variability in Somatic
580
Tissue and Shell Weights and Its Effect on Meat Yield in Mixed Species [Mytilus edulis L., M.
581
trossulus (Gould), and Their Hybrids] Cultured Mussel Populations. Journal of Shellfish
582
Research 27(4):827834.
583
584
Likens, G.E. (Ed.), Long-term Studies in Ecology: Approaches and Alternatives, Springer, New
585
586
Porcasi, J.F., 2011. More on Mollusks: Trans-Holocene Shellfish Exploitation on the California
587
588
Randklev, C.R., Wolverton, S.J., Kennedy, J.H., 2009. A biometric technique for assessing
589
prehistoric freshwater mussel population dynamics (family: Unionidae) in north Texas. Journal
590
591
Rick, T.C., Lockwood, R., 2013. Integrating Paleobiology, Archeology, and History to Inform
592
AC
C
EP
TE
M
AN
U
SC
RI
PT
568
22
ACCEPTED MANUSCRIPT
Schmidt, D., 1999. A Review of California Mussel Fisheries Biology and Fisheries Programes.
594
595
Seed, R., 1968. Factors Influencing Shell Shape in the Mussel Mytilus edulis. Journal of the
596
597
Seed, R., Suchanek, T.H., 1992. Population and Community Ecology of Mytilus. in: Gosling,
598
E.G. (Ed.), The Mussel Mytilus: Ecology, Physiology, Genetics and Culture, Elsevier,
599
Amsterdam.
600
Shawcross, W., 1967. An Investigation of Prehistoric Diet and Economy on a Coastal Site at
601
602
Simenstad, C.A., Estes, J.A., Kenyon, K.W., 1978. Aleuts, sea otters, and alternate stable-state
603
604
Singh, G.G., 2010. Effects of Sea Otters on Nearshore Ecosystem Functions with Implications
605
for Ecosystem Services. MSc Thesis. Resource Management and Environmental Studies,
606
607
Singh, G.G., Markel, R.W., Martone, R.L.G., Salomon, A.K., Harley, C.D.G., Chan, K.M.A.,
608
2013. Sea Otters Homogenize Mussel Beds and Reduce Habitat Provisioning in a Rocky
609
610
Smith, J.R., Fong, P., Ambrose, R.F., 2006. Dramatic Declines in Mussel Bed Community
611
612
Stein, J.K., 1996. Geoarchaeology and Archaeostratigraphy: view from a Northwest Coast Shell
613
Midden. in: Reitz, E.J., Newsom, L.A., Scudder, S.J. (Eds.), Case Studies in Environmental
614
615
Suchanek, T.H., 1979. The Mytilus californianus community: Studies on the composition,
616
617
AC
C
EP
TE
M
AN
U
SC
RI
PT
593
23
ACCEPTED MANUSCRIPT
Suchanek, T.H., 1981. The Role of Disturbance in the Evolution of Life History Strategies in the
619
620
Sumpter, I.D., 2005. An Analysis of Three Shellfish Assemblages from Ts'ishaa, site DfSi-16
621
(204T), Benson Island, Pacific Rim National Park Reserve. in: McMillan, A.D., St. Claire, D.E.
622
623
624
Swadling, P., 1976. The Implications of Shellfish Exploitation for New Zealand Prehistory.
625
Mankind 11(1):1118.
626
Szpak, P., Orchard, T.J., McKechnie, I., Grcke, D.R., 2012. Historical Ecology of Late
627
Holocene Sea Otters (Enhydra lutris) from Northern British Columbia: Isotopic and
628
629
630
631
Treganza, A.E., Cook, S.F., 1948. The Quantitative Investigation of Aboriginal Sites: Complete
632
Excavation with Physical and Archaeological Analysis of a Single Mound. American Antiquity
633
13(4):287297.
634
635
Communities. in: VanBlaricom, G.R., Estes, J.A. (Eds.), The Community Ecology of Sea Otters,
636
637
Waselkov, G.A., 1987. Shellfish gathering and shell midden archaeology. Advances in
638
639
Wessen, G.C., 1982. Shell Middens As Cultural Deposits: A Case Study From Ozette. PhD
640
641
Wessen, G.C., 1988. The Use of Shellfish Resources on the Northwest Coast: The View from
642
Ozette. in: Issac, B.L. (Ed.), Prehistoric Economies of the Pacific Northwest Coast: Research in
643
Economic Anthropology, Supplement 3., JAI Press, Greenwich, CT, pp. 179-207.
AC
C
EP
TE
M
AN
U
SC
RI
PT
618
24
ACCEPTED MANUSCRIPT
645
and Sustainability vs. Immediate Returns for the Harvest of Marine Invertebrates. Journal of
646
647
648
Coastal Sites, MacKerricher State Park, Mendocino County, California. Report on File with
649
650
Wolverton, S., Lyman, R.L., 2012. Conservation Biology and Applied Zooarchaeology.
651
652
Wolverton, S., Randklev, C.R., Kennedy, J.H., 2010. A Conceptual Model for Freshwater
653
654
655
Wolverton, S.J., Dombrosky, J., Lyman, R.L., 2014. Practical Significance: Ordinal Scale Data
656
657
658
M
AN
U
SC
RI
PT
644
659
TE
660
AC
C
EP
661
25
ACCEPTED MANUSCRIPT
Figure Captions
Figure 1. Map showing locations and sample numbers of M. californianus and M. trossulus
specimens collected for this study. Italicized sample numbers represent beach collected shells
used to evaluate taphonomic effects as discussed in the text.
RI
PT
Figure 2. Measurements made on a shell of M. californianus (main and inset pictures) to create
regressions to predict total shell length. Measures of the lengths of 1-4 were made with the total
length of the shell. The photo shows the inside of the left valve.
SC
Figure 3. Regression model fits to data. Different colored and shaped points correspond to shells
collected in different areas (refer to legend). a) Umbo to far hinge teeth. b) Length of hinge teeth.
c) Umbo width. d) Umbo to near hinge teeth
M
AN
U
Figure 4. Visual comparison of estimated total length to known shell lengths using a) White's
(1989:133) template, b) the regression model based on measuring the length from the umbo to
the outer hinge teeth, c) the regression model based on measuring the length of the hinge teeth,
d) the regression model based on measuring the width of the umbo, and e) the regression model
based on measuring the length from the umbo to the inner hinge teeth. The straight lines are not
regressions in themselves, but instead represent a 1:1 relationship, with the best models falling
along the line. Model predictions that are more accurate fall closer to the straight line, while
those points farther away represent relatively worse predictions.
EP
TE
Figure 5. Relationships between total length and dry weight mass for reproductive, nonreproductive, and total meat weight.
AC
C
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
26
ACCEPTED MANUSCRIPT
M
AN
U
SC
RI
PT
Table 1. Regression equations for estimating total length of mussel shells from shell fragments. The accuracy measurements are a
measure of how different, on average across sizes and participants, the estimate is from the actual total shell length, and represents the
mean error as well as the 95% confidence interval.
Measurement Description (x) Regression Equation
R2 Value Average Error ( 95% CI) in double blind trials
umbo to outer hinge teeth
total length = 1.8074 x + 4.947 0.93
7.31.2 mm
length of hinge teeth
total length = 2.3916 x + 4.9349 0.81
11.21.9 mm
umbo thickness
total length = 15.312 x + 7.1701 0.80
10.21.9 mm
umbo to inner hinge teeth
total length = 4.284 x + 28.94
0.80
12.83.0 mm
Table 2. Multiple regression models incorporating multiple measurements. In cases where enough of the shell is intact enough to make
multiple measurements (the two cases are described under the Shell Fragment column), these models can be used to obtain
increased accuracy in length estimates than with simple regressions shown in Table 1.
Shell Fragment
Regression Equation
R2
0.84
0.94
TE
D
AC
C
EP
Table 3. Regression equations to estimate dry body (meat) weight from total shell length for M. californianus.
Estimate
Regression Equation
R2
Dry weight of total body
Dry weight = 0.0006832*total length^1.9678096 0.89
Dry weight of non-reproductive body Dry weight = 0.0006155*total length^1.9249686 0.90
Dry weight of gonad
Dry weight = 0.0001019*total length^2.0887239 0.78
ACCEPTED MANUSCRIPT
AC
C
EP
TE
M
AN
U
SC
RI
PT
Table 4. Mean error estimates of the four different regression models used on beach collected Mytilus californianus shells and shells
of Mytilus trossulus.
Shell Type
Umbo to outer hinge teeth Hinge length Umbo thickness Umbo to inner hinge teeth
Beach collected M. californianus 5.41.4
9.52.3
9.92.3
15.43.4
Mytilus trossulus
4.01.2
5.91.3
9.53.1
15.53.6
M
AN
U
SC
RI
PT
ACCEPTED MANUSCRIPT
Quadra Island
n=51 (M. trossulus)
Calvert
Island, n=16
EP
TE
Kyuquot Sound
n=10
Clayoquot Sound
n=20
Barkley Sound
n=38 (52)
AC
C
Pacific
Ocean
AC
C
EP
TE
D
M
AN
U
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
C
EP
TE
M
AN
U
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
C
EP
TE
D
M
AN
U
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
C
EP
TE
M
AN
U
SC
RI
PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
HIGHLIGHTS
We develop regressions for estimating length of fragmentary Mytilus californianus.
Regressions are based on live-collected shells from California and British Columbia.
RI
PT
AC
C
EP
TE
M
AN
U
SC
These improved length estimates extend the interpretive utility of Mytilus assemblages.
ACCEPTED MANUSCRIPT
EP
Total
123.37
128.62
118.67
57.03
56.81
121.38
96.04
95.44
50.78
49.38
49.27
61.89
63.93
65.32
128.15
144.82
52.83
126.9
118.21
66.4
55.55
69.59
68.71
117.91
64.74
77.89
88.86
55.44
54.6
52.05
145.64
42.15
38.29
34.45
44.16
37.1
33.27
87.68
69.04
109.29
54.66
124.57
69.52
87.65
67.16
67.62
M
AN
U
SC
RI
PT
TE
Plot
Boathouse 198
Boathouse 202
Boathouse 196
Boathouse 191
Boathouse 189
Boathouse 197
Boathouse 195
Boathouse 195
Boathouse 171
Boathouse 156
Boathouse 159
Boathouse 192
Boathouse 190
Boathouse193
Jalama 49
Jalama 55
Jalama 12
Jalama 54
Jalama 51
Jalama 10
Jalama 8
Jalama 30
Jalama 33
Jalama 51
Jalama 20
Jalama 36
Jalama 43
Jalama 10
Jalama 8
Jalama 7
Jalama J56
Cambria 158
Cambria 181
Cambria 148
Cambria 194
Cambria 191
Cambria 150
Jalama 42
Jalama 32
Jalama 48
Jalama 8
Jalama 53
Jalama 30
Jalama 42
BS 15 2
BS 25 2
AC
C
Region
Site
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Boathouse
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Cambria
South California Cambria
South California Cambria
South California Cambria
South California Cambria
South California Cambria
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
South California Jalama
British Columbia BS
British Columbia BS
side
left
left
right
right
left
right
right
left
left
left
right
left
left
right
right
left
left
right
left
left
left
left
right
right
right
left
left
right
right
right
right
left
right
right
right
right
right
right
right
right
left
right
right
left
left
right
ACCEPTED MANUSCRIPT
EP
RI
PT
26.87
38.67
41.63
30.76
29.49
42.5
34.95
45.45
32.3
26.02
28.02
29.22
26.46
30.9
28.1
29.03
50.12
29.05
21.58
34.69
30.46
27.2
34.82
30.68
27.45
29.14
31.23
42.91
114.67
70.24
61.39
73.26
45.05
65.44
55.69
46.91
46.14
52.9
30.38
28.58
23.82
23.97
78.72
26.47
21.97
28.88
26.55
SC
7.43
6.82
9.44
4.84
6.31
7.36
8.17
8.61
5.73
9.23
4.21
4.49
4.67
10.4
4.11
5.22
8.43
4.72
4.61
5.46
5.73
6.6
11.02
5.4
6.3
5.49
4.9
10.84
39.95
16.21
15.2
17.58
6.64
21.13
12.38
13.42
11.85
15.36
6.55
4.5
6.13
4.91
22.64
9.98
3.67
4.16
4.53
M
AN
U
D
TE
BS 21 1
BS 16 2
BS 16 1
BS 4 1
BS 21 2
BS 25 1
BS 4 2
BS 15 1
Blun 4 low
Blun 18 low
LAC 0 low
Blun 5 mid
LAC 27 mid
Bart 7 mid
LAC 16 lower
LAC 3 low
Blun 13 low
Bart 5 low
Blun 25
Bart 1 low
LAC 12 mid
Blun 15 mid
Bart 4 low
Bart 11 mid
Blun 4 mid
Blun 11 mid
Bart 22 low
Blun 22 low
Helby Vex 2N 1
Sanf Vex 1N 1
Diana Vex 5N 1
Diana Vex 4N 1
Helby Vex 1N 1
Helby Vex 3N 1
Seppings iB 4 1
KT1 iB 5 1
Helby Vex 4N 1
Seppings iB 6 1
PB Putty 4 2
PB Vex 4N 1
Diana Vexar 3 4
TB Vex 5 3
Helby Vex 5N 1
TB Putty 2 1
TB Vex 2N 3
GDA Vex 1N 1
TB Putty 2 2
AC
C
British Columbia BS
British Columbia BS
British Columbia BS
British Columbia BS
British Columbia BS
British Columbia BS
British Columbia BS
British Columbia BS
British Columbia Blun
British Columbia Blun
British Columbia LAC
British Columbia Blun
British Columbia LAC
British Columbia Bart
British Columbia LAC
British Columbia LAC
British Columbia Blun
British Columbia Bart
British Columbia Blun
British Columbia Bart
British Columbia LAC
British Columbia Blun
British Columbia Bart
British Columbia Bart
British Columbia Blun
British Columbia Blun
British Columbia Bart
British Columbia Blun
British Columbia Helby
British Columbia Sanf
British Columbia Diana
British Columbia Diana
British Columbia Helby
British Columbia Helby
British Columbia Seppings
British Columbia KT1
British Columbia Helby
British Columbia Seppings
British Columbia PB
British Columbia PB
British Columbia Diana
British Columbia TB
British Columbia Helby
British Columbia TB
British Columbia TB
British Columbia GDA
British Columbia TB
ACCEPTED MANUSCRIPT
EP
TE
27.94
41.63
49.9
41.26
50.8
18.4
49.46
10.55
11.88
8
3.82
12.61
7.94
6.36
6.68
6.88
4.36
6.1
3.73
24.31
62.78
47.02
32.45
24.55
26.65
25.55
30.69
26.27
20.65
26.74
23.45
29.35
48.65
27.37
27.4
25.98
23.53
26.52
33.14
4.25
5.29
6.65
5.36
8.25
2.92
5.27
1.59
1.34
1.34
0.98
1.73
0.71
1.22
1.03
1.26
0.74
0.95
0.73
5.3
7.12
5.78
4.71
5.64
3.46
3.93
4.18
4.7
4.75
5.34
2.58
3.31
7.94
3.51
3.52
3.88
3.8
4.25
6.4
49.2
98.3
106.38
102.4
122.47
37.12
110.08
26.99
22.63
18.02
12.64
24.26
15.7
17.12
14.4
18.43
10.27
13.69
9.15
64.55
156.05
103.74
66.42
58.07
63.04
56.14
67.21
57.48
54.21
63.61
51.03
63.77
115.29
56.56
60.11
59.11
61.08
56.78
73.9
RI
PT
32.12
56
60.73
54.84
70.17
22.96
58.46
13.24
13.52
10.46
6.02
14.71
8.74
8.08
7.73
8.83
5.14
7.59
4.56
37.3
84.24
57.68
39.91
33.87
33.76
30.45
36.6
32.8
29.65
35.39
28.12
34.72
62.52
32.64
32.62
33.68
30.13
32.14
44.74
SC
6.03
15.99
13.92
14.9
21.42
4.02
10.08
2.47
1.82
2.39
2.18
1.86
1.05
1.55
1.6
2.11
0.8
1.34
1.22
13.88
23.34
11.89
8.43
11.68
8.53
5.29
7.13
6.76
9.56
8.45
5.76
6.83
15.72
6.21
5.28
7.19
7.11
5.94
12.71
M
AN
U
GDA Vex 3N 1
Bluestone Vex 4N 1 winter
Diana Vex 3N 1 winter
Sanf Vex 5N 2 winter
Sanf Vex 2N 1 winter
KT1 iB1 2 winter
Bluestone Vex 2N 1 winter
Diana s1
Diana s2
Diana s3
Diana s4
Diana s5
Diana s6
Diana s7
Diana s8
Diana s9
Diana s10
Diana s11
Diana s12
Con 5
#1
#11
gray 5
black 8
gray 11
black 7
gray 10
con 6
con 7
black 6
con 8
gray 8
#10
gray 7
black 3
gray 6
con 2
con 11
gray 3
AC
C
left
right
right
left
right
right
right
left
right
left
right
left
left
right
left
left
right
left
right
left
left
right
left
right
right
right
left
right
right
left
right
left
left
left
left
left
right
right
left
ACCEPTED MANUSCRIPT
RI
PT
SC
umbo est
54.02329
65.50692
80.6653
66.57872
90.92401
129.6621
76.83743
81.58399
73.31578
58.92297
92.45516
60.14789
97.50795
114.6568
67.3443
63.66954
86.17744
90.00532
101.1827
72.39709
73.31578
77.29677
92.30204
63.51642
61.83216
78.67481
72.55021
52.18591
40.54917
73.00955
54.94198
67.19118
51.57345
84.49318
94.29254
80.81842
95.67057
101.1827
75.00005
49.58296
104.8575
90.77089
144.0549
125.6811
97.04861
73.31578
67.65053
61.67904
81.73711
62.59773
68.72233
56.47313
75.91874
67.3443
M
AN
U
bhing diff Far Hinge fhing est fhing diff Hinge length
length est length diff Umbo
-1.63632
28.29 56.07862 -7.86862
22.73 59.30516 -11.0952
3.06
-21.7047
31.47 61.82618 -15.3062
22.45 58.6354 -12.1154
3.81
0.31268
32 62.78411 -11.5541
27.32 70.28444 -19.0544
4.8
28.50308
47.01 89.91333 -2.35333
38.62 97.31404 -9.75404
3.88
35.32328
58.42 110.5359 -1.37588
50.6 125.9702 -16.8102
5.47
29.72432
79.23 148.1481 -6.28808
61.21 151.3493 -9.48932
8
7.0114
40.07 77.36991 -4.78991
32.05 81.5986 -9.0186
4.55
3.71544
41.31 79.6111 -7.3711
32.45 82.5554 -10.3154
4.86
10.15924
43.19 83.00903 -12.679
35.55 89.9706 -19.6406
4.32
6.28288
34.58 67.44723 -6.17723
28.6 73.3462 -12.0762
3.38
20.82412
62.67 118.2174 -10.7474
48.96 122.0473 -14.5773
5.57
-17.4586
26.98 53.71091 -12.8409
19.11 50.64612 -9.77612
3.46
21.69496
56.3 106.7042 -6.97417
45.15 112.9338 -13.2038
5.9
41.51328
72.79 136.5084 -10.4484
61.99 153.2151 -27.1551
7.02
0.06624
34.15 66.67004
-13.93
28.63 73.41796 -20.678
3.93
-3.61192
38.29 74.15272 -13.3527
30.62 78.17804 -17.378
3.69
5.81568
47.87 91.46771 -5.81771
36.5
92.243
-6.593
5.16
25.7674
55.09 104.5172 1.812793
41.4 103.9638
2.3662
5.41
17.9196
52.57 99.96253 -10.2625
43.8 109.7046 -20.0046
6.14
-11.57
33.92 66.25434 -15.8543
26.9 69.2798 -18.8798
4.26
24.06992
43.91 84.31036 5.199637
37.34 94.25228 -4.74228
4.32
-6.4178
35.35 68.83893 6.591067
25.15 65.0938 10.3362
4.58
21.5162
48.83 93.20282 5.87718
37.96 95.73532 3.34468
5.56
7.17388
29.92 59.0247 2.065303
24.03 62.41476 -1.32476
3.68
7.24156
30.2 59.53077 -0.60077
24.48 63.49116 -4.56116
3.57
18.79256
36.44 70.80901 9.31099
29.81 76.24052 3.87948
4.67
-2.87868
35 68.20634 12.13366
23.72 61.67324 18.66676
4.27
-8.76508
26.2 52.30113 9.728868
17.62 47.08204 14.94796
2.94
-2.57864
28.18 55.8798 10.5802
19.44 51.43548 15.02452
2.18
-25.6614
28.16 55.84366 5.526344
15.63 42.32196 19.04804
4.3
0.9854
22.25 45.16186 6.398138
17.84 47.60828 3.95172
3.12
6.05296
35.14 68.45938 7.060623
26.26 67.74892 7.77108
3.92
-1.3386
23.56 47.52957 5.99043
18.19 48.44548 5.07452
2.9
1.65008
35.48 69.0739 5.556103
26.65 68.6818
5.9482
5.05
-11.8328
46.4 88.81081 11.44919
28.71 73.60932 26.65068
5.69
-6.37112
27.17 54.05432 9.55568
18.52 49.23484 14.37516
4.81
8.50976
36.5 70.91746 16.91255
25.49 65.90708 21.92292
5.78
-11.0451
38.7 74.89376 6.276243
25.08 64.92636 16.24364
6.14
13.76816
38.81 75.09257 17.56743
27.63 71.02596 21.63404
4.43
5.13984
25.87 51.70469 4.395313
21.28 55.83676 0.26324
2.77
8.69108
51.39 97.82979 -6.51979
38.51 97.05092 -5.74092
6.38
28.85892
51.95 98.84194 -3.47194
39.88 100.328 -4.95796
5.46
16.35648
84.61 157.872 -2.56195
60.35 149.2922
6.0178
8.94
48.80212
72.44 135.8758 1.71423
58.27 144.3168 -6.72684
7.74
23.89764
53.87 102.3122 -7.96217
45.73 114.3212 -19.9712
5.87
8.8898
36.56 71.0259 -7.7059
30
76.695 -13.375
4.32
-15.8957
28.12 55.77136 -2.37136
18.39 48.92388 4.47612
3.95
6.0102
32.78 64.19389 -6.32389
27.22 70.04524 -12.1752
3.56
-15.1346
34.22 66.79656 -6.46656
24.71 64.04132 -3.71132
4.87
3.40688
29.45 58.17521 -4.06521
25.62 66.21804 -12.108
3.62
7.67924
36.54 70.98975 -3.13975
29.84 76.31228 -8.46228
4.02
2.18328
30.19 59.5127 5.797302
23.01 59.97492 5.33508
3.22
-2.59168
32.85 64.32041 -1.90041
24.78 64.20876 -1.78876
4.49
1.30784
33.13 64.82648 0.293517
25.41 65.71572 -0.59572
3.93
mussel
Base Hingebhing est
vex 2n1
4.88 49.84632
vex 5 4
9.17 68.22468
vex 3 4
5.13 50.91732
vex 4N 1
7.03 59.05692
vex 2n1
10.48 73.83672
vex 5n1
19.42 112.1357
vex 5n1 w
8.55 65.5686
vex 5n3 w
9.24 68.52456
vex 2n1 w
7.29 60.17076
vex 5n1 w
6.08 54.98712
vex 4n 1 w
13.47 86.64588
putty 4 3
6.86 58.32864
vex 3n 1 w
11.46 78.03504
vex 4n 1 s
12.98 84.54672
putty 6 2
5.54 52.67376
vex 2N2 s
8.28 64.41192
vex 1n1 w
11.88 79.83432
vex 5n1 w
12.05 80.5626
vex 5n 4 w
10 71.7804
vex 4n 1 w
7.71 61.97004
P95
8.52 65.44008
P92
12.35 81.8478
P96
11.35 77.5638
P66
5.83 53.91612
P65
5.31 51.68844
P90
7.56 61.32744
V45
12.67 83.21868
V37
9.77 70.79508
V36
9.36 69.03864
V33 s
13.56 87.03144
V12 s
5.05 50.5746
V41
9.46 69.46704
V21
6.05 54.8586
V26
10.28 72.97992
V10
19.41 112.0928
V47
9.58 69.98112
V31
11.76 79.32024
V12 l
14.77 92.21508
V33 l
11.66 78.89184
V24
5.14 50.96016
#13
12.53 82.61892
#12
8.77 66.51108
#2
25.68 138.9535
#8
13.97 88.78788
#14
9.69 70.45236
gray 2
5.95 54.4302
con 9
9.42 69.29568
black 11
5.35 51.8598
con 3
10.86 75.46464
gray 9
5.08 50.70312
black 2
7.29 60.17076
black 9
7.98 63.12672
black 5
8.42 65.01168
black 10
8.14 63.81216
TE
Site
PB
GDA
Beach
Helby
BS
Helby
BS
Helby
Diana
Sanf
Diana
Sanf
Helby
Diana
Beach
PB
Helby
Helby
Sanf
GDA
Piedras
Piedras
Piedras
Piedras
Piedras
Piedras
VAFB
VAFB
VAFB
VAFB
VAFB
VAFB
VAFB
VAFB
VAFB
VAFB
VAFB
VAFB
VAFB
VAFB
Mussel Pt
Mussel Pt
Mussel Pt
Mussel Pt
Mussel Pt
EP
Region
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
S Cali
N Cali
N Cali
N Cali
N Cali
N Cali
N Cali
N Cali
N Cali
N Cali
N Cali
N Cali
N Cali
N Cali
N Cali
AC
C
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
precision est
-15.6765
1.458963
-7.40304
0.175828
16.51432
7.682624
-2.59614
-1.46048
-6.27478
-2.29905
0.304919
-2.3301
-3.59106
4.452054
1.255395
-11.0541
-1.10592
-3.0869
0.711686
-1.69861
-3.13486
-3.70675
-1.41816
-1.15952
0.924123
5.835077
5.362333
-12.647
-5.13025
-1.08615
0.156792
-1.25266
-1.17079
-0.59224
-2.41059
4.838767
42.29835
0.711686
3.416285
-2.21931
2.816486
0.190614
40.28697
-1.95022
-7.50432
1.104036
1.561625
0.30002
-6.95922
-3.02018
-7.41431
-6.78984
-7.28273
-4.86745
ACCEPTED MANUSCRIPT
EP
TE
M
AN
U
SC
RI
PT
Site
mussel
Base Hingebhing est bhing diff Far Hinge fhing est fhing diff Hinge length
length est length diff Umbo
umbo est umbo diff Total
Dicebox west
1
16.33 98.89772 10.69228
57.29 108.4929 1.097054
43.49 108.9456 0.644416
5.05 84.4957 25.0943
109.59
Dicebox west
2
6.37 56.22908 12.94092
35.1 68.38674 0.78326
29.4 75.24794 6.07794
3.39 59.07778 10.09222
69.17
Dicebox west
3
11.92 80.00528 28.34472
62.9 118.6325 10.28246
50.96 126.8108 18.46084
5.64 93.52978 14.82022
108.35
Dicebox west
4
5.84 53.95856 15.78856
19.1 39.46834 1.29834
13.33 36.81493 1.355072
2.71 48.66562 10.49562
38.17
Dicebox west
5
16.11 97.95524 37.10476
80.88 151.1295 16.06951
65.26 161.0107 25.95072
6.6 108.2293 26.8307
135.06
Dicebox west
6
11.32 77.43488 10.84512
47.58 90.94309 2.663092
37.37 94.30899 6.028992
5.49 91.23298 2.95298
88.28
Dicebox west
7
8.05 63.4262 18.5638
45.45 87.09333 5.10333
37.84 95.43304 13.44304
4.4 74.5429
7.4471
81.99
gilbert
8
31.97 165.8995 1.59948
102.81 190.7658 26.46579
74.02 181.9611 17.66113
11.63 185.2487 20.94866
164.3
gilbert
9
19.43 112.1781 26.72188
76.27 142.7974 3.897398
58.44
144.7 5.800004
10.27 164.4243 25.52434
138.9
gilbert
10
12.12 80.86208 50.04792
68.45 128.6635 2.24647
57.23 141.8062 10.89617
5.94 98.12338 32.78662
130.91
gilbert
11
14.7 91.9148 48.5352
71.06 133.3808 7.069156
57.09 141.4713 1.021344
5.76 95.36722 45.08278
140.45
dicebox east
12
7.78 62.26952 2.75048
30.65 60.34381 4.67619
24.14 62.66812 2.351876
1.91 36.41602 28.60398
65.02
dicebox east
13
12.35 81.8474 23.7974
47.61 90.99731 32.94731
37.52 94.66773 36.61773
5.77 95.52034 37.47034
58.05
dicebox east
14
5.72 53.44448 4.60552
35.14 68.45904 10.40904
30.52 77.92653 19.87653
3.85 66.1213
8.0713
58.05
dicebox east
15
4.69 49.03196 8.07196
22.58 45.75809 4.798092
18.87 50.06439 9.104392
2.84 50.65618 9.69618
40.96
dicebox east
16
4.79 49.46036 12.17964
39.51 76.35737 14.71737
35.23 89.19097 27.55097
3.54 61.37458 0.26542
61.64
dicebox west
17
11.65 78.8486
0.2214
48.85 93.23849 14.16849
38.23 96.36577 17.29577
4.02 68.72434 10.34566
79.07
dicebox west
18
11.36 77.60624 25.11376
57.15 108.2399 5.51991
47.64 118.8707 16.15072
5.72 94.75474 7.96526
102.72
dicebox west
19
11.26 77.17784 15.76216
46.26 88.55732 4.382676
36.2 91.51082 1.42918
4.28 72.70546 20.23454
92.94
dicebox west
20
5.41 52.11644 14.65644
27.81 55.21079 17.75079
23.5 61.1375 23.6775
3.05 53.8717 16.4117
37.46
Diana Islnd Beach 21
13.58 87.11672 8.01672
37.8 73.26672 2.486212
24.18 62.76379 10.8241
4.5 76.0741 21.86672
65.25
Diana Islnd Beach 22
14.18 89.68712 3.703162
38.13 73.86316 5.19594
25.1 64.96406
2.8517
4.3 73.0117 19.52712
70.16
Diana Islnd Beach 23
10.58 74.26472 3.167992
36.08 70.15799 6.392492
28.62 73.38249 5.92166
3.52 61.06834 7.27472
66.99
Diana Islnd Beach 24
9.14 68.09576 0.557558
35.33 68.80244 3.37676
28.35 72.73676 6.14798
3.66 63.21202 1.26424
69.36
Diana Islnd Beach 25
10.75
74.993 3.209944
42.56 81.86994 3.88232
32.45 82.54232 12.84494
3.83 65.81506
3.667
78.66
Diana Islnd Beach 26
16.91 101.3824 3.234404
49.54 94.4856 7.787636
35.54 89.93236 10.62126
5.22 87.09874 3.66244
97.72
Diana Islnd Beach 27
11.97 80.21948 7.556772
39.22 75.83323 10.72499
28.32 72.66501
8.0815
4.45 75.3085 3.17052
83.39
Diana Islnd Beach 28
19.38 111.9639 0.729824
58.76 111.1498 6.879432
41.23 103.5406 5.55934
6.38 104.8607 1.54392
110.42
Diana Islnd Beach 29
12.32 81.71888 6.404628
50.22 95.71463 6.768776
38.11 96.07878 1.13942
5.29 88.17058 7.59112
89.31
Diana Islnd Beach 30
22.05 123.4022 3.910032
60.68
114.62 10.97208
39.64 99.73792 4.31814
6.48 106.3919 12.6922
110.71
Diana Islnd Beach 31
33.52 172.5397 1.334832
62.68 118.2348 38.20816
30.84 78.69184 23.79074
8.72 140.6907 55.63968
116.9
Diana Islnd Beach 32
7.71 61.96964 3.586686
37.61 72.92331
0.7708
30.25 77.2808 3.80454
4.28 72.70546 14.54036
76.51
Diana Islnd Beach 33
23.01 127.5148 5.259118
45.93 87.96088 28.59076
24.96 64.62924
3.5253
5.85 96.7453 34.29484
93.22
Diana Islnd Beach 34
15.44 95.08496 3.786772
50.78 96.72677 1.094356
36.34 91.84564 4.41778
5.89 97.35778 2.14496
92.94
Diana Islnd Beach 35
10.88 75.54992 6.236422
48.03 91.75642 11.56325
38.53 97.08325
2.0381
5.25 87.5581 9.97008
85.52
Diana Islnd Beach 36
8.76 66.46784 2.86659
34.65 67.57341 1.794524
28.14 72.23452 8.75918
3.56 61.68082 3.97216
70.44
Diana Islnd Beach 37
13.58 87.11672 21.86672
37.8 73.26672 8.01672
24.18 62.76379 2.486212
4.5 76.0741 10.8241
65.25
Diana Islnd Beach 38
14.18 89.68712 19.52712
38.13 73.86316 3.703162
25.1 64.96406 5.19594
4.3 73.0117
2.8517
70.16
Diana Islnd Beach 39
10.58 74.26472 7.27472
36.08 70.15799 3.167992
28.62 73.38249 6.392492
3.52 61.06834 5.92166
66.99
Diana Islnd Beach 40
9.14 68.09576 1.26424
35.33 68.80244 0.557558
28.35 72.73676 3.37676
3.66 63.21202 6.14798
69.36
Diana Islnd Beach 41
10.75
74.993
3.667
42.56 81.86994 3.209944
32.45 82.54232 3.88232
3.83 65.81506 12.84494
78.66
Diana Islnd Beach 42
16.91 101.3824 3.66244
49.54 94.4856 3.234404
35.54 89.93236 7.787636
5.22 87.09874 10.62126
97.72
AC
C
Region
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
side
left
left
left
right
left
left
right
left
right
right
right
left
left
left
left
left
left
right
right
right
ACCEPTED MANUSCRIPT
7.556772
0.729824
6.404628
3.910032
1.334832
3.586686
5.259118
3.786772
6.236422
2.86659
0.960786
7.269912
0.08523
0.600486
0.199476
4.430846
1.534412
3.42098
4.617478
3.534168
17.59757
3.25321
0.139212
2.439912
1.513384
0.48279
28.32
41.23
38.11
39.64
30.84
30.25
24.96
36.34
38.53
28.14
32.55
58.24
42.78
42.21
43.17
37.64
15.94
10.07
11.62
9.69
40.04
61.05
15.34
20.99
9.8
18.6
72.66501
103.5406
96.07878
99.73792
78.69184
77.2808
64.62924
91.84564
97.08325
72.23452
82.78148
144.2217
107.2475
105.8843
108.1803
94.95472
43.057
29.01831
32.72529
28.1095
100.6946
150.9421
41.62204
55.13458
28.37258
49.41866
10.72499
6.879432
6.768776
10.97208
38.20816
0.7708
28.59076
1.094356
11.56325
1.794524
2.57148
7.591684
9.947548
8.864336
13.10973
3.365276
4.617004
4.538312
4.435292
4.429504
16.97456
1.13208
2.882044
4.044584
0.51258
4.85866
RI
PT
75.83323
111.1498
95.71463
114.62
118.2348
72.92331
87.96088
96.72677
91.75642
67.57341
79.24921
143.8999
97.21477
96.41951
121.0905
93.88915
39.97441
27.90098
32.90748
27.21417
101.3176
146.5568
38.60079
53.52991
26.34662
44.07721
SC
39.22
58.76
50.22
60.68
62.68
37.61
45.93
50.78
48.03
34.65
41.11
76.88
51.05
50.61
64.26
49.21
19.38
12.7
15.47
12.32
53.32
78.35
18.62
26.88
11.84
21.65
M
AN
U
3.17052
1.54392
7.59112
12.6922
55.63968
14.54036
34.29484
2.14496
9.97008
3.97216
13.9992
27.62204
32.24588
30.93772
3.35456
20.41388
8.66416
20.3108
14.05892
18.71176
9.45528
44.27208
8.57836
3.59684
15.51708
1.47316
80.21948
111.9639
81.71888
123.4022
172.5397
61.96964
127.5148
95.08496
75.54992
66.46784
66.2108
109.008
65.05412
66.08228
124.6446
77.90612
47.10416
44.7908
42.34892
42.39176
74.26472
105.5379
47.31836
54.68684
43.37708
46.03316
TE
11.97
19.38
12.32
22.05
33.52
7.71
23.01
15.44
10.88
8.76
8.7
18.69
8.43
8.67
22.34
11.43
4.24
3.7
3.13
3.14
10.58
17.88
4.29
6.01
3.37
3.99
EP
AC
C
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
4.45
6.38
5.29
6.48
8.72
4.28
5.85
5.89
5.25
3.56
4.87
9.08
5.31
5.48
7.14
5.09
2.01
1.18
1.51
1.32
4.48
8.48
1.98
3.28
1.2
2.22
75.3085
8.0815
104.8607 5.55934
88.17058 1.13942
106.3919 4.31814
140.6907 23.79074
72.70546 3.80454
96.7453
3.5253
97.35778 4.41778
87.5581
2.0381
61.68082 8.75918
81.73954 1.52954
146.2031 9.57306
88.47682 8.82318
91.07986 5.94014
116.4978 4.79222
85.10818 13.21182
37.94722 0.49278
25.23826 0.75826
30.29122 2.00122
27.38194 3.70194
75.76786 7.95214
137.0159 12.79414
37.48786 1.25214
57.39346 6.30346
25.5445
2.3155
41.16274 3.39726
83.39
110.42
89.31
110.71
116.9
76.51
93.22
92.94
85.52
70.44
80.21
136.63
97.3
97.02
121.29
98.32
38.44
24.48
28.29
23.68
83.72
149.81
38.74
51.09
27.86
44.56
ACCEPTED MANUSCRIPT
EP
TE
M
AN
U
SC
RI
PT
Sample Number
Total Length
1.(mm)
Umbo to2.
outer
length
hinge
of3.hinge
teeth
umbo
teeth
thickness
4. umbo to 5.
inner
greatest
hingeshell
height
teethweight
(cross-section)
outer
(mg) hingelength
prediction
prediction
umbo prediction
inner hingeError
prediction
outerError lengthError umboError inner
HB1
53.25
25.81
24.25
2.66
3.52
10.73
1779 51.59599 62.9312 47.90002 44.01968 1.654006
9.6812 5.34998 9.23032
HB2
54.81
25.36
21.46
3.31
5.89
11.21
2555 50.78266 56.25864 57.85282 54.17276 4.027336 1.448636 3.04282 0.63724
HB3
46.47
20.58
18.59
1.98
2.78
8.53
1320 42.14329 49.39474 37.48786 40.84952 4.326708 2.924744 8.98214 5.62048
HB4
51.74
26.84
24
2.88
3.99
10.2
3523 53.45762 62.3333 51.26866 46.03316 1.717616 10.5933 0.47134 5.70684
HB5
41.15
22.33
19.37
2.27
4.16
9.5
1802 45.30624 51.26019 41.92834 46.76144 4.156242 10.11019 0.77834 5.61144
HB6
32.03
15.84
13.32
1.39
3.2
7.21
670 33.57622 36.79101 28.45378 42.6488 1.546216 4.761012 3.57622 10.6188
HB7
46.05
23.11
13.54
2.45
8.06
9.3
46.71601 37.31716 44.6845 63.46904 0.666014 8.732836
1.3655 17.41904
HB8
44.3
21.27
14.14
2.05
8.28
8.39
43.3904 38.75212 38.5597 64.41152 0.909602 5.547876
5.7403 20.11152
HB9
54.69
18.07
12.1
2.07
10.1
37.60672 33.87326 38.86594
28.94 17.08328 20.81674 15.82406
25.75
HB10
43.52
19.93
18.2
1.5
3.08
7.06
892 40.96848 48.46202 30.1381 42.13472 2.551518 4.94202 13.3819 1.38528
HB11
35.65
19.32
15.15
1.25
3.87
6.39
634 39.86597 41.16764 26.3101 45.51908 4.215968 5.51764
9.3399 9.86908
HB12
39.47
16.83
13.85
1.33
7.39
725 35.36554 38.05856 27.53506
28.94 4.104458 1.41144 11.93494
10.53
HB13
85.8
32.8
28.65
2.62
14.52
6023 64.22972 73.45424 47.28754
28.94 21.57028 12.34576 38.51246
56.86
HB14
38.85
18.61
17.21
2.12
2.71
8
38.58271 46.09434 39.63154 40.54964 0.267286 7.244336 0.78154 1.69964
HB15
37.95
17.66
15.38
2.1
2.84
7.7
36.86568 41.71771 39.3253 41.10656 1.084316 3.767708
1.3753 3.15656
HB16b
34.54
16.05
12.66
1.26
3.58
7.35
829 33.95577 35.21256 26.46322 44.27672 0.58423 0.672556 8.07678 9.73672
HB17b
30.8
15.97
9.1
0.87
8.2
7.13
33.81118 26.69846 20.49154 64.0688 3.011178 4.10154 10.30846 33.2688
HB18a
31.92
13.77
10.38
1.65
4.5
7.08
676 29.8349 29.75971 32.4349
48.218 2.085102 2.160292
0.5149
16.298
HB19b
36.51
16.55
13.29
1.9
4.45
8.08
1134 34.85947 36.71926 36.2629 48.0038 1.65053 0.209264
0.2471 11.4938
HB20b
27.41
13.9
10.33
1.41
2.73
6.3
30.06986 29.64013 28.76002 40.63532 2.65986 2.230128 1.35002 13.22532
HB21
37.81
16.96
11.77
1.91
7.32
935 35.6005 33.08403 36.41602
28.94 2.209496 4.725968 1.39398
8.87
HB22
45.29
23.78
19.29
2.72
2.84
9.44
2188 47.92697 51.06886 48.81874 41.10656 2.636972 5.778864 3.52874 4.18344
HB23
44.74
20.38
15.7
2.68
5.03
9.53
41.78181 42.48302 48.20626 50.48852 2.958188 2.25698 3.46626 5.74852
HB24
51.7
29.4
25.79
2.81
4.15
10.78
58.08456 66.61426 50.19682 46.7186 6.38456 14.91426 1.50318
4.9814
HB25
47.8
23.65
16.94
3.12
6.1
10.07
2804 47.69201 45.4486 54.94354 55.0724 0.10799 2.351396 7.14354
7.2724
HB26
53.39
24.09
21.36
2.44
3.6
9.4
2795 48.48727 56.01948 44.53138 44.3624 4.902734 2.629476 8.85862
9.0276
HB27
48.29
27.15
20.95
2.91
3.14
9.39
2020 54.01791 55.03892 51.72802 42.39176 5.72791 6.74892 3.43802 5.89824
HB28
45.34
23.13
19.1
2.35
3.55
8.46
1764 46.75216 50.61446 43.1533 44.1482 1.412162 5.27446
2.1867
1.1918
HB29
42.23
21.38
18.77
2.13
2.63
7.87
1593 43.58921 49.82523 39.78466 40.20692 1.359212 7.595232 2.44534 2.02308
HB30
51.69
28.76
25.27
2.69
5.81
9.12
56.92782 65.37063 48.35938 53.83004 5.237824 13.68063 3.33062 2.14004
HB31b
70.55
35.89
32.46
3.45
4.38
13.78
69.81459 82.56624 59.9965 47.70392 0.735414 12.01624 10.5535 22.84608
HB32b
92.16
43.91
38.44
4.43
6.04
16.58
84.30993
96.868 75.00226 54.81536 7.850066 4.708004 17.15774 37.34464
HB33b
95.22
43.84
38.54
3.59
4.96
16.84
84.18342 97.10716 62.14018 50.18864 11.03658 1.887164 33.07982 45.03136
HB34
82.02
35.08
27.58
3.73
4.96
14.81
68.35059 70.89523 64.28386 50.18864 13.66941 11.12477 17.73614 31.83136
HB35
72.85
37.9
33.17
3.76
5.07
11.27
73.44746 84.26427 64.74322 50.65988 0.59746 11.41427 8.10678 22.19012
HB36b
82.38
38.32
34.84
3.06
3.64
15.54
7586 74.20657 88.25824 54.02482 44.53376 8.173432 5.878244 28.35518 37.84624
HB37b
47.99
21.45
17.97
2.48
3.81
9.11
2023 43.71573 47.91195 45.14386 45.26204 4.27427 0.078048 2.84614 2.72796
HB38b
61.01
33.56
29.06
2.88
5.43
11.94
4095 65.60334 74.4348 51.26866 52.20212 4.593344 13.4248 9.74134 8.80788
HB39b
18.8
8.26
6.8
1.03
2.02
4.1
166 19.87612 21.19778 22.94146 37.59368 1.076124 2.39778 4.14146 18.79368
HB40
70.43
36.28
27.9
2.6
7.77
14.92
70.51947 71.66054 46.9813 62.22668 0.089472 1.23054 23.4487 8.20332
HB41b
25.62
12.12
9.85
1.49
3.56
5.3
398 26.85269 28.49216 29.98498 44.19104 1.232688 2.87216 4.36498 18.57104
HB42b
23.5
10.36
8.51
1.52
2.89
4.53
283 23.67166 25.28742 30.44434 41.32076 0.171664 1.787416 6.94434 17.82076
AC
C
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Hyacinthe Bay
Hyacinthe Bay
Hyacinthe Bay
Hyacinthe Bay
Hyacinthe Bay
Hyacinthe Bay
Hyacinthe Bay
Hyacinthe Bay
Hyacinthe Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay
Heriot Bay