APPROVED:

Steve Wolverton, Major Professor Lisa Nagaoka, Committee Member Charles R. Randklev, Committee Member Paul F. Hudak, Chair of the Department of

Geography Costas Tsatsoulis, Interim Dean of the

Toulouse Graduate School

ZOOARCHAEOLOGY AND BIOGEOGRAPHY OF FRESHWATER MUSSELS IN THE LEON RIVER

DURING THE LATE HOLOCENE

Traci Glyn Popejoy, B.S.

Thesis Prepared for the Degree of

MASTER OF SCIENCE

UNIVERSITY OF NORTH TEXAS

May 2015

Popejoy, Traci Glyn. Zooarchaeology and Biogeography of Freshwater Mussels in the

Leon River during the Late Holocene. Master of Science (Applied Geography), May 2015, 76 pp.,

10 tables, 15 figures, references, 129 titles.

The Leon River, a small-medium river in central Texas, is highly impacted by multiple

impoundments, enrichment from agricultural runoff, and decreased dissolved oxygen levels.

This degraded river contains sixteen unionid species, two of which are both endemic to the

region and candidates for the federal endangered species listing (Quadrula houstonensis and

Truncilla macrodon). While there is a short historical record for this river basin and a recent

modern survey completed in 2011, zooarchaeological data adds evidence for conservation

efforts by increasing the time depth of data available and providing another conservation

baseline. Zooarchaeological data for the Leon River is available from the two Late Holocene

archaeological sites: 41HM61 and the Belton Lake Assemblages. Data generated from these

assemblages describe the prehistoric freshwater mussel community of the Leon River in terms

of taxonomic composition and structure. By comparing this zooarchaeological data to the data

generated by the longitudinal modern survey of the Leon River, long term changes within the

freshwater mussel community can be detected. A conceptual model is constructed to evaluate

how robusticity, identifiability, and life history ecology affect unionid taxonomic abundances in

zooarchaeological data. This conceptual model functions as an interpretive tool for

zooarchaeologists to evaluate forms of equifinality in zooarchaeological assemblages. This

thesis determines differences between the late Holocene and modern freshwater community

of the Leon River, explores how different alternative mechanisms influence zooarchaeological

data, and exemplifies of how zooarchaeological data can be used for conservation biology.

Copyright 2015

by

Traci Glyn Popejoy

ii

ACKNOWLEDGEMENTS

I would like to thank my advisor, Steve Wolverton for guidance on research topics,

writing, and contextualizing my research. I would also like to thank Lisa Nagaoka for instruction

regarding writing and presenting research. I am also grateful to Charles R. Randklev for writing

guidance, freshwater mussel expertise, and direction regarding analytical techniques. Thanks to

all of my committee members for infinite patience and wisdom. I am indebted to my family and

Sean Dubose for their patience when my thesis periodically consumed my life. Costal

Environments, Inc. and Texas Department of Transportation funded the description of the

unionid remains from the 41HM61 assemblage.

iii

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ................................................................................................................... iii

LIST OF TABLES ................................................................................................................................. v

LIST OF FIGURES .............................................................................................................................. vi

CHAPTER 1 INTRODUCTION ............................................................................................................ 1

CHAPTER 2 DESCRIPTION OF THE LATE HOLOCENE FRESHWATER MUSSEL COMMUNITY OF THE LEON RIVER IN CENTRAL TEXAS ............................................................................................... 5

CHAPTER 3 CHANGE IN THE LEON RIVER’S FRESHWATER MUSSEL COMMUNITY STRUCTURE SINCE THE LATE HOLOCENE .......................................................................................................... 18

CHAPTER 4 A CONCEPTUAL MODEL FOR ASSESSING FRESHWATER MUSSEL TAXONOMIC ABUNDANCES IN ZOOARCHAEOLOGICAL FAUNAS ....................................................................... 34

CHAPTER 5 CONCLUSION .............................................................................................................. 54

APPENDIX A SYSTEMATIC PALEONTOLOGY .................................................................................. 57

APPENDIX B DATA USED IN CHAPTER 3INFERENTIAL TESTS ........................................................ 64

REFERENCES .................................................................................................................................. 67

iv

LIST OF TABLES

Page

Table 3.1 Mussels of the Leon River & preferred habitat. ................................................... 22

Table 3.2 Results of all chi-square goodness of fit test and Spearman’s rho correlation.. .. 28

Table 4.1 Taxonomic abundances within the 41HM61 and Belton Lake assemblages........ 36

Table 4.2 Robusticity and life history strategy values for Leon River zooarchaeological freshwater mussel assemblage ............................................................................. 41

Table 4.3 Expectations for abundance filters discussed in this chapter .............................. 44

Table 4.4 Zooarchaeological taxonomic abundance expectations for the Leon River unionid community. .............................................................................................. 47

Table B.1 Data for culled chi-square goodness of fit test ..................................................... 65

Table B.2 Data for culled Spearman’s rho correlation ......................................................... 65

Table B.3 Data for complete chi-square goodness of fit test ............................................... 66

Table B.4 Data for complete Spearman’s rho correlation .................................................... 66

v

LIST OF FIGURES

Page

Fig. 2.1. Map of Leon River in central Texas. ........................................................................ 8

Fig. 2.2. Map of Belton Lake Zooarchaeological assemblages.. ........................................... 8

Fig. 2.3. Example specimen of Q. houstonensis from the 41HM61 assemblage. .............. 10

Fig. 2.4. Graphic representation of the nestedness concept and a nestedness analysis. . 12

Fig. 2.5. Relative abundances of taxa in the 41HM61 and the Belton Lake assemblages. ......................................................................................................... 14

Fig. 2.6. Nestedness matrix generated from the zooarchaeological assemblages from the Leon River. ...................................................................................................... 15

Fig. 3.1. Conceptual diagram of mesohabitats within river systems. ................................ 21

Fig. 3.2. Relative taxonomic abundances of late Holocene and mordern assemblages from the Leon River. ............................................................................................. 24

Fig. 3.3. NMDS ordination of freshwater mussel assemblages based on assemblage species structure. .................................................................................................. 30

Fig. 4.1. Model describing potential filters within zooarchaeological assemblages. ......... 38

Fig. 4.2. Ordination of life history strategies of freshwater mussels in the southeastern United States. ........................................................................................................ 45

Fig. 4.3. Analysis of life history strategy’s influence on the 41HM61 and the Belton Lake assemblages. ................................................................................................. 48

Fig. 4.4. Analysis of sculpture’s influence on the 41HM61 and Belton Lake assemblages. ......................................................................................................... 49

Fig. 4.5. Robusticities’ influence on the 41HM61 and Belton Lake assemblages. ............. 50

Fig. 4.6. Taphonomic analysis of the 41HM61 assemblage and the Belton Lake assemblages .......................................................................................................... 50

vi

CHAPTER 1

INTRODUCTION

Rivers have been subjected to increased anthropogenic change in the past century:

cities have expanded, land use has changed, and urban water use has increased (Humphries

and Winemiller, 2009; Palmer et al., 2008; Parmesan and Matthews, 2006). The Leon River in

central Texas is no exception and has changed since European settlement. A citizen of nearby

Dublin, Texas remembers this stream changing during the early twentieth century:

When I first came to Dublin, we used to go down to the Leon River and fish a lot. It was a nice clear stream most of the time. The river had a deep channel and lots of holes of water 25-30 feet deep. Well, we caught a lot of good fish. After a while fishing began to get poor. The Leon River began to get out of banks almost every time we had a large rain. Well somebody found out that muddy water had been depositing silt all down the river and had filled it up until it hardly had any banks at all… Hallmark in 1938 taken from (Hendrickson Jr., 1981, p. 59)

Since its impoundment in the 1950s, the Leon River has experienced habitat change and an

increase in runoff from urban and agricultural areas (United States Department of Agriculture,

2005; Harmel et al., 2008; Hendrickson Jr., 1981; Rossi et al., 2008). A longitudinal survey of the

freshwater mussel population of the Leon River found decreased abundance and diversity

compared to historic records (which begin in 1931) (Strecker, 1931). There are three unionid

species of conservation concern in this river: false spike (Fusconaia mitchelli), smooth

pimpleback (Quadrula houstonensis) and Texas fawnsfoot (Truncilla macrodon). This thesis

discusses changes to the freshwater mussel assemblage of the Leon River since the late

Holocene in order to assess whether or not there has been change in the mussel community

related to contemporary human impacts.

1

Freshwater mussels (Family: Unionidae) have been highly impacted through

anthropogenic alteration of river systems (Galbraith et al., 2010; Nobles and Zhang, 2011;

Williams et al., 1993) and have experienced population declines in the last 100 years. Of nearly

300 species native to North America, 30 species are now extinct (Haag, 2012 and references

within). As sedentary filter feeders with a parasitic larval stage, these organisms are sensitive to

many of the anthropogenic impacts affecting rivers, such as decreased river conductivity,

increased temperature, and habitat degradation (Haag and Warren Jr., 2008; Shea et al., 2013;

Vaughn and Taylor, 1999). Due to this sensitivity, many freshwater mussels (hereafter unionids)

are indicator species of aquatic ecosystem health and can warn conservation biologists of

declining quality in stream habitat (Cummings and Graf, 2010; Dillon Jr., 2000; Haag, 2012).

Conservation biologists have been diligently studying declining freshwater mussel communities

at a state and national level (Haag and Williams, 2013).

Conservation biology is a scientific approach to “address the biology of species,

communities and ecosystems that are perturbed, either directly or indirectly, by human

activities or other agents” for the purpose of preserving the evolution of biodiversity (Soulé,

1985, p. 727). To conserve remaining ecological processes and biodiversity often requires action

before everything is known about the situation, making it a “crisis discipline” (Soulé, 1985, p.

727). When decisions must be made quickly, evidence of ecological change and past

environments is highly valuable. Conservation biology uses conservation baselines as goals for

wildlife and ecosystem management, but these baselines are culturally influenced and mainly

rely on historical records (Joyce, 2012). Historical records, though an important source of

information, can be incomplete, inaccurate, or represent already degraded environments

2

(Lyman, 2012a, 2012b; Randklev, 2011). When historical records are inadequate,

zooarchaeological data may provide valuable evidence of past ecosystems (Louys et al., 2012;

Lyman, 2012a, 1996; Wolverton and Lyman, 2012). This type of data can then be used to

construct a different conservation baseline at a longer time scale. This thesis examine late

Holocene zooarchaeological data from faunas from the Leon River to describe the past

freshwater mussel community, adding valuable baseline information to the conservation

dialogue regarding this declining freshwater mussel community.

Two zooarchaeological assemblages are reported in this thesis, the 41HM61 and the

Belton Lake assemblages (which are several small archaeological faunas excavated near Belton

Lake that date to the late Holocene). These zooarchaeological assemblages represent the late

Holocene freshwater mussel community of the Leon River in terms of species composition and

taxonomic abundance. These late Holocene unionid assemblages are compared to data from a

modern longitudinal survey of the Leon River unionid community (Randklev et al., 2013a). The

goal of this thesis is two parts: to assess how the unionid community of the Leon River has

changed since the late Holocene and to assess how life history ecology, identifiability, and

robusticity influence taxonomic abundance in zooarchaeological assemblages, which can

influence the ability of applied zooarchaeologists to establish baseline conditions relevant to

conservation. Chapter 2 reports data from the 41HM61 and the Belton Lake zooarchaeological

assemblages to provide a representation of the late Holocene unionid community and

addresses changes in species composition of the Leon River unionid community. Chapter 3 uses

nonparametric inferential tests to address taxonomic structural change in the Leon River

unionid community since the late Holocene. Chapter 3 also uses a nonmetric multidimensional

3

scaling (NMDS) ordination to address the similarity between modern mesohabitat sub-

assemblages and the late Holocene assemblages. Chapter 4 creates a conceptual model to

address how variable life history strategy, identifiability, and robusticity affect taxonomic

abundances in zooarchaeological assemblages. This conceptual model is then used to evaluate

the results of the inferential tests in Chapter 3. These analyses assess ecological change in the

unionid community of the Leon River with implications for wildlife management.

By understanding how the freshwater mussel community of the Leon River has changed

since the late Holocene, conservation biologists can better manage today’s unionid community.

Shifts in the assemblage toward tolerant species might indicate that the river has experienced

anthropogenic impacts. This thesis also helps biologists understand how riverine ecosystems

can change through time. To impact conservation, the results of this thesis need to be

disseminated to the appropriate agencies and wildlife managers (Wolverton and Lyman, 2012).

This is achieved by presenting parts of this thesis at multiple conferences such as the Texas

Freshwater Mussel Symposium and the Freshwater Mollusk Conservation Society Symposium.

The conceptual model created in this thesis helps zooarchaeologists understand how different

mechanisms influence taxonomic abundance data and conclusions based on that data. It is an

interpretive tool designed to help zooarchaeologists understand underlying influences in their

unionid taxonomic abundance data. In total, this thesis uses zooarchaeological data for

conservation purposes in the Leon River and improves the interpretation of this data by using a

conceptual model to evaluate forms of equifinality in zooarchaeological assemblages.

4

CHAPTER 2

DESCRIPTION OF THE LATE HOLOCENE FRESHWATER MUSSEL COMMUNITY OF THE LEON RIVER

IN CENTRAL TEXAS

Introduction

Freshwater mussels (Family: Unionidae) are highly diverse and imperiled across the

North American continent. These sedentary filter-feeders are affected by many anthropogenic

affects, such as impoundments, increased runoff and increased municipal water use. In Texas,

freshwater mussels are experiencing declines in community diversity and abundance (Burlakova

et al., 2011). Though historical records are available for many river basins in Texas, these are

limited as they often represent freshwater mussel communities at a short ecological time scale.

Zooarchaeological data increase knowledge about prehistoric biogeographic distributions of

these organisms, which improves conservation efforts according to the National Strategy for

unionid conservation (Haag and Williams, 2013; Strayer et al., 2004).

The Leon River, located in central Texas, has experienced many of the anthropogenic

effects described above, and as such the freshwater mussel fauna has experienced a decrease

in diversity and abundance (Randklev et al., 2013a). While many historical records exist for the

river, the prehistoric freshwater mussel community of the Leon River has received limited

discussion. This chapter describes freshwater mussel remains collected from late Holocene

archaeological sites located in Hamilton, Bell, and Coryell County and provides information

about the distribution and richness in the Leon River. Data are also reported from site 41HM61

and from multiple small zooarchaeological assemblages surrounding Belton Lake in Bell and

Coryell County. Nestedness is analyzed for both the Belton Lake assemblages and 41HM61 to

5

determine if they represent the same faunal community and if they can be aggregated in future

analyses. By fully describing the prehistoric freshwater mussel community of the Leon River,

conservation biologists can determine if change in the taxonomic composition of this

community has occurred.

Background

Applied zooarchaeology is the use of zooarchaeological data for purposes of

conservation biology (Cannon and Cannon, 2004; Lyman, 2012a, 1996; Peacock, 2005;

Wolverton and Lyman, 2012). It provides information on past biological communities most

often through presence/absence and abundance data for taxa within assemblages.

Presence/absence data are often used to assess the distribution of species of conservation

concern through time and across space (Grayson, 1987; Lyman, 1991). Abundance data can also

be used to quantify the taxonomic abundances in an ecological community (Grayson, 1987;

Randklev and Lundeen, 2012; Wolverton, 2008). Applied zooarchaeological data have also been

used to discuss paleoenvironmental conditions, such as flow regime, substrate, and nutrient

levels through isotopic and morphologic analysis of freshwater mussel remains (Peacock and

Seltzer, 2008; Spurlock, 1981; Warren, 1991). The field has benefited modern freshwater

mussel conservation by providing new information about prehistoric unionid communities, for

example in the Upper Trinity River basin of north Texas (Randklev and Lundeen, 2012) and the

Tombigbee River basin of Mississippi (Peacock, 2012; Peacock et al., 2012). The goal of the

current study is to use zooarchaeological data to analyze the prehistoric freshwater mussel

community composition in the Leon River during the Late Holocene.

6

Study Area

The Leon River is a small to medium size stream in central Texas and a tributary of the

Brazos River (Fig. 2.1). (Anderson et al., 2014; Randklev et al., 2013a). Central Texas rivers

contain 85% of the state’s endemic species, making these rivers important for the conservation

of unionid diversity in Texas (Burlakova et al., 2011). The Leon River basin contains sixteen

unionid species; Fusconaia mitchelli, Quadrula houstonensis, and Truncilla macrodon are

endemic to the region and of conservation concern. Quadrula houstonensis and T. macrodon

are candidates for federal endangered species listing, while F. mitchelli is petitioned for federal

listing (Randklev et al., 2013a; U.S. Fish and Wildlife Service, 2014). The Leon River is impacted

by multiple impoundments, enrichment from agricultural and urban runoff, and subsequent

decreased dissolved oxygen levels (Burlakova et al., 2011; Randklev et al., 2013a). While there

is a short historical record for this river basin and a recent modern survey completed in 2011,

zooarchaeological data can provide a late Holocene baseline for mussel conservation efforts.

The Hamilton County assemblage is from the archaeological site 41HM61 and represents

multiple occupations. The Belton Lake assemblage comprises remains from eighteen separate

cave sites that surround the modern Belton Lake reservoir; the location of these sites can be

seen in Fig. 2.2 (Randklev, 2010). Mussels from these assemblages were identified by Charles

Randklev. These assemblages represent sites that are approximately 60 miles apart, but are

expected to represent the same faunal community in terms of species composition.

7

Fig. 2.1. Map of Leon River in central Texas. This is the map shows the location of Lake Belton on the Leon River and the Leon River’s location within the state of Texas. Taken from Randklev et al. (2013a).

Fig. 2.2. Map of Belton Lake Zooarchaeological assemblages. These assemblages were combined because they represent the same faunal community, and many have low NRE. It is important to note that Belton Lake is a reservoir, and thus an anthropogenically altered landscape, though the assemblages would represent the unaltered Leon River. Map provided by Dr. Charles Randklev (Randklev, 2010).

8

Materials and Methods

The freshwater mussel remains from the 41HM61 assemblage and the Belton Lake

assemblages are identified following guidelines described in Driver (2011, 1992) and Wolverton

(2013). The quantitative unit non-repetitive element (NRE) is used to count each specimen that

represents one unionid shell (Giovas, 2009); to be counted as a NRE, the specimen must include

psuedocardinal teeth or an umbo, which are located on one of the densest areas of the unionid

shell, are present at all life stages, and are often identifiable to a precise taxonomic level. Fig.

4.3 shows the psuedocardinal teeth and umbo of a unionid specimen from 41HM61. While

weighing specimens is a common form of quantification for shell material, counting NRE

provides an accurate ordinal scale tally of unionid abundance (Claassen, 2000; Giovas, 2009;

Glassow, 2000; Mason et al., 1998). Since the 41HM61 assemblage is highly fragmented,

specimens that passed through a twelve millimeter sieve and did not represent a NRE were not

counted. All specimens that are either larger than a twelve millimeter sieve or represent a NRE

are included in the quantification term number of specimens (NSP), which counts the total

number of fragments in an assemblage. Following Driver (2011, 1992), specimens were

identified independently using a comparative collection of shells from the Brazos and Leon

Rivers and multiple identification guides (Howells, 2013; Howells et al., 1996; Williams et al.,

2008). A systematic paleontology can be found in Appendix A to account for data quality

(Lyman, 2011). Specimens were identified to the lowest taxonomic category possible; due to

the morphological plasticity exhibited by unionids, some categories above the species level

were used to categorize specimens from species with similar shell morphology.

9

Fig. 2.3. A specimen identified as Q. houstonensis is provided to show what constituted as a NRE during identification of this assemblage. The umbo/psuedocardinal teeth are circled in yellow. The umbo is the rounded bump on the outer edge of the shell. The psuedocardinal teeth are found as the two nubs in the upper left portion of the yellow circle. Psuedocardinal teeth are often identifiable a precise taxonomic category. The presence of psuedocardinal teeth and/or an umbo constitutes a NRE and one mussel.

Relative abundances are calculated by dividing the number of NRE identified to an

individual taxa by the total NRE in an assemblage. Using relative abundances as ordinal ranks, it

can be determined which unionid taxa are most abundant in the 41HM61 and Belton Lake

assemblages. Using the quantitative data (NRE, NSP, and presence), the fragmentation and

nestedness of the 41HM61 and the Belton Lake assemblages is assessed. For unionid remains,

fragmentation represents the preservation condition of shells in the assemblages. Species with

fragile shell morphology are likely to be under-represented in faunas that are highly

fragmented (Lyman, 1994; Wolverton et al., 2010). Fragmentation can be calculated using a

ratio of NRE:NSP, which measures the proportion of identifiable shells in an assemblage. If

fragmentation is low and most of the specimens in an assemblage are identifiable, then the

ratio will be close to one. Highly fragmented assemblages that consist of many shell fragments

will have a low NRE:NSP value (closer to zero).

10

To ensure that the 41HM61 and Belton Lake assemblages represent the same faunal

community, a nestedness analysis is conducted (Lyman, 2008; Peacock et al., 2012). Nestedness

analysis assesses whether smaller assemblages are composed of the same species as larger

assemblages (see Fig. 2.4) (Wright et al., 1997). Biogeographers use this technique to evaluate

patterns of species presence on island chains; land-bridge islands have nested subsets of the

continental fauna based on the island’s size (and therefore its carrying capacity) (Atmar and

Patterson, 1993). It is also used to analyze how species richness changes among and within

river systems (Rashleigh, 2008; Vaughn and Taylor, 1999). For zooarchaeological studies, it is

expected that assemblages with lower NTAXA (number of taxa represented by the assemblage)

are nested subsets of assemblages with higher NTAXA within a paleoecological community. This

expectation stems from the fact that as a sample size increases (in terms of number of

specimens or area sampled), the number of species captured is expected to increase as well

(Atmar and Patterson, 1993; Lyman and Ames, 2007; Lyman, 2008; Wright et al., 1997).

Nestedness analysis can determine if there are unexpected or rare taxa within a set of samples,

compared to the entire data set for a larger meta-community, such as mussel communities in a

river basin; here it is used to evaluate the assumption that the 41HM61 and Belton Lake

assemblages represent the same late Holocene mussel community within the Leon River in

terms of species composition. If these assemblages represent the same faunal community, they

can be more confidently aggregated in future analyses.

11

Fig. 2.4. Graphic representation of the nestedness concept and a nestedness analysis. A) The rectangular boxes represent the different taxa represented by the different assemblages. The red rectangle has a lower number of NTAXA, of which all are found within the larger, green assemblage. B) This table represents hypothetical nested assemblages possible from the Leon River mussel community. As comparable to part A, the assemblage with smaller NTAXA is represented by a red rectangle. The taxon found within the smaller assemblage is also found within the assemblage with larger NTAXA, the green rectangle. C) This diagram is the visual representation generated by a nestedness analysis of perfectly nested assemblages. Note that the most abundant taxon are in the top row and the site with the largest NTAXA is in the first column.

Nestedness is displayed in a binary matrix that lists the most ubiquitous taxa on the top

rows and the most taxa rich sites in the left column of the table (see Fig. 2.4) (Ulrich et al.,

2009). Though a binary matrix can visually describe the nestedness of a group of assemblages,

quantification of this binary matrix is useful when assemblages are partially nested.

Biogeographers have generated a quantification unit to further describe nestedness: heat of

disorder (Atmar & Patterson, 1993; Lyman, 2008). Heat of disorder quantifies nestedness based

on the species incidence and species composition of different assemblages. At 0 degrees, the

faunas are perfectly nested and exhibit a large amount of similarity, while at 100 degrees the

faunas exhibit no nestedness (Lyman, 2008). A FORTRAN program called Nestedness is used to

calculate the binary matrix and the heat of disorder (Ulrich, 2006). While the 41HM61 and the

Belton Lake assemblages are approximately 60 miles apart, they are expected to represent the

same faunal community in terms of species composition. This expectation is appropriate as

12

both assemblages come from the same tributary of the Brazos River (Haag, 2012; Rashleigh,

2008) and should share similar habitat types and fish community composition, which are two

characteristics that influence mussel community composition. Assemblages comprise shells of

unionids gathered from multi-species beds, and since unionids provide only small amounts of

edible meat, species exclusion is unlikely (Bettinger et al., 1997; Peacock et al., 2012).

Results

Taxonomic abundance data generated from the 41HM61 assemblage and the Belton

Lake assemblage is reported below (Fig. 2.5). The 41HM61 assemblage had a total of 399 NRE

and 1054 NSP. Of the 399 NRE total, 314 were identified; 85 specimens were unidentifiable. In

the 41HM61 assemblage, 14 taxa are represented. Amblema plicata, Quadrula verrucosa, and

Lampsilis sp. dominate the 41HM61 assemblage, with relative abundances of 23.3%, 9.5% and

9.3% respectively. The Belton Lake assemblages totaled 525 NRE, with 487 identified to the

species level. In the Belton Lake assemblages, 11 taxa were identified. Amblema plicata,

Quadrula houstonensis and Fusconaia mitchelli dominate the assemblages, with relative

abundances of 58.5%, 22.5% and 7.2% respectively. The entirety of the abundance data can be

found in Fig. 2.5. From these histograms, it appears that 41HM61 exhibited more species

evenness across the assemblage. A diversity index is not used to quantify species diversities of

the assemblages due to different sample sizes and preservation contexts, which would

inherently affect richness and evenness (Lyman, 2008).

13

Fig. 2.5. Relative abundances of taxa in the 41HM61 and the Belton Lake assemblages.

Fragmentation is used in this chapter as a proxy for preservation of mussel remains from

an assemblage. The 41HM61 assemblage has a NRE:NSP ratio of 0.38 and the Belton Lake

assemblages have a NRE:NSP ratio of 0.93, indicating better shell preservation in the latter

assemblages.

The nestedness of the individual Belton Lake assemblages and the 41HM61 assemblage

is evaluated to determine whether both of these assemblages represent the same faunal

community in terms of species composition. As seen in the binary matrix (Fig. 2.6), the

zooarchaeological assemblages from the Leon River are highly nested, with 1.85 degrees of

heat. Since the zooarchaeological assemblages visually appear nested and have a low matrix

temperature, it can be confidently assumed that they represent the same late Holocene faunal

community from the Leon River, and therefore these assemblages are aggregated in future

analyses.

14

Fig. 2.6. Nestedness matrix generated from the zooarchaeological assemblages from the Leon River. Comparing this binary matrix to the one exampled in Fig. 2.4C, it is evident that the assemblages are well nested. The most diverse sites are generally listed from left to right. The most ubiquitous taxa are listed from the top to the bottom. Almost all assemblages contained A. plicata, other than 41BL780B. The bottom row represents the number of taxa found in each site.

Discussion

This description of the late Holocene freshwater mussel community of the Leon River

provides insight into nominal scale changes in the unionid community. A longitudinal survey

conducted by Randklev et al. (2013) describes the modern freshwater mussel community of the

Leon River. Cyrtonias tampicoensis, Fusconaia mitchelli, and Lampsilis hydiana were not

encountered in the modern survey yet these species were found in prehistoric assemblages and

shell material has been recorded in historic surveys. The importance of the absence of these

species in the modern survey has implications for the conservation of the freshwater mussel

fauna in this stream.

15

Cyrtonaias tampicoensis has never been recorded live in the Leon River basin, but fresh-

dead specimens have been collected in the middle and lower Leon River (Howells, 1997;

Randklev et al., 2013a). As such, it expected that today C. tampicoensis is rare and/or located

only in the lower Leon River. Lampsilis hydiana has not been collected from the Leon River since

the early 1930s and is expected to be locally rare or extirpated within the Leon River basin

(Randklev et al., 2013a; Strecker, 1931). While not considered of conservation concern, C.

tampicoensis and L. hydiana’s absence from the Leon River has implications for unionid

conservation. Both of these species’ presence in zooarcheaolgical assemblages provides

evidence that range constrictions have occurred, which is evidence of population decline. These

zooarchaeological data support that unionid populations are declining in central Texas rivers

(Richardson and Whittaker, 2010; Whittaker et al., 2005).

Both zooarchaeological assemblages contain shell remains of F. mitchelli, which is a

species of conservation concern in Texas. Presumed to be extinct, a small population of F.

mitchelli was found in the Guadalupe River in 2011 (Randklev et al., 2013b, 2012). Historically,

its distribustion included the Guadalupe, San Antonio, Colorado and Brazos River basins

(Randklev et al., 2012). This species has remained listed as state threatened since November

2009 and is pending review as a candidate for federal listing on the Endanged Species Act

(Texas Parks & Wildlife Department 2009, U.S. Fish and Wildlife Service 2014). The presence of

F. mitchelli in these zooarchaeological assemblages and subsequent absences from the survey

in 2011 confirms that the distribution of F. mitchelli in the Brazos River watershed has

decreased or has been eliminated. As more contemporary biogeographic data are gathered

through ecological surveys, the distribution of these species will be better defined and the

16

amount of range constriction and/or extirpation of these species within the Brazos and Leon

Rivers can be more carefully evaluated.

Conclusion

The Leon River of central Texas has likely been impacted through increased agricultural

and urban run off and water use changes (Harmel et al., 2008; Randklev et al., 2013a; Rossi et

al., 2008). These anthropogenic changes to the Leon River have negatively impacted freshwater

mussels causing change in the mussel community since the late Holocene within this tributary

of the Brazos River. The description of the freshwater mussels from the 41HM61 and Belton

Lake assemblages provide evidence for the late Holocene distribution of freshwater mussels

within the Leon River. These descriptions stand as evidence for nominal-scale change within the

Leon River due to negative anthropogenic impacts. The declining richness of the freshwater

mussel community and range constriction of these species indicate that these faunas are likely

experiencing population decline. Though it is evident from the absence of F. mitchelli, L.

hydiana, and C. tampicoensis in the contemporary survey that the Leon River mussel

community has changed since the Late Holocene, a further exploration of this unionid

community change is found in Chapter 3.

17

CHAPTER 3

CHANGE IN THE LEON RIVER’S FRESHWATER MUSSEL COMMUNITY STRUCTURE SINCE

THE LATE HOLOCENE

Introduction

Anthropogenic alteration of streams has negatively affected riverine communities,

causing population declines of many species (Humphries and Winemiller, 2009; Palmer et al.,

2008; Parmesan and Matthews, 2006). Freshwater mussels (Family: Unionidae) are no

exception to this trend (Galbraith et al., 2010; Nobles and Zhang, 2011; Williams et al., 1993).

As sedentary filter feeders with a parasitic larval stage, these organisms are sensitive to many

anthropogenic changes to streams, such as decreased river conductivity, increased water

temperature, and habitat degradation (Haag and Warren Jr., 2008; Shea et al., 2013; Vaughn

and Taylor, 1999). Their sensitivity to ecological change represents impacts experienced by

other riverine fauna, making them indicator species for aquatic environments (Cummings and

Graf, 2010; Haag and Williams, 2013; Strayer, 2008). By understanding how the unionid

community has changed through time, conservation biologists can better understand how river

ecosystems have changed over time.

A longitudinal survey of the Leon River in central Texas revealed decreased freshwater

mussel richness and distribution (Randklev et al., 2013a). Changes to the structure of the

unionid community are concerning as the Leon River harbors many state threatened species

and two candidates for listing on the Endangered Species Act (Fusconaia mitchelli, Quadrula

houstonensis, and Truncilla macrodon) (U.S. Fish and Wildlife Service 2014; TPWD 2009). A

comparison between the late Holocene and modern unionid assemblages of the Leon River

18

potentially reveals ecological shifts in this community. This chapter uses chi square goodness of

fit tests and Spearman’s rho correlation to compare the structure of the unionid community

from the past and present. The habitat represented by the late Holocene freshwater mussel

assemblage will be evaluated using nonmetric multi-dimensional scaling (NMDS) ordination.

These analyses evaluate potential community change in the Leon River with implications for the

management of freshwater mussels.

Background

Zooarchaeological assemblages provide a representation of past environments for

wildlife managers and conservation biologists (Lyman, 2012a; Wolverton and Lyman, 2012).

Changes in community structure have been explored using zooarchaeological data to evaluate

shifts in ecosystem function and animal ecology (Newsome et al., 2007; Peacock and Seltzer,

2008; Randklev and Lundeen, 2012; Wolverton, 2008). The results of applied zooarchaeological

studies benefit freshwater mussel conservation by providing new information about prehistoric

unionid communities (Miller et al., 2014; Peacock, 2012; Peacock et al., 2012; Randklev and

Lundeen, 2012).

An nonmetric multidimensional scaling (NMDS) ordination is used to assess similarity

between the late Holocene and modern freshwater mussel assemblage based on community

structure (Kindt and Coe, 2005; Ricklefs, 2008). This ordination also potentially illuminates what

mesohabitat sub-assemblage is represented by the late Holocene assemblage. In this

ordination, the modern data is separated into sub-assemblages based on mesohabitat, which

should harbor different unionid sub-assemblages based on habitat preferences. This technique

is appropriate for unionid communities as these organisms are sedentary, can only thrive in

19

suitable habitats, and often form discontinuous patches within streams (Dennis, 1984; Haag,

2012; Strayer, 2008; Vaughn and Pyron, 1995). Unionids can be found in a variety of

mesohabitats. Suitable habitats must maintain flow, provide stable sediments, and the

necessary physiological (temperature, dissolved oxygen, etc.) conditions for that species (Allen

and Vaughn, 2010; Baldigo et al., 2004; Haag and Warren Jr., 2010; Haag, 2012; Holland-Bartels,

1990; Strayer, 2008). Within streams there are a variety of mesohabitats available for mussel

colonization, based on different sediment types and stability (Holland-Bartels, 1990; Strayer,

1999). For this chapter, mesohabitats are defined as different locations in a stream based on

sediment type and flow. Three categories define mesohabitats: lotic, lentic, and transitional.

Lotic mesohabitats generally have fast, consistent flow and gravel substrate. Lentic

mesohabitats are usually depositional areas in lotic environments with sand or clay substrate

and slow flow. Transitional mesohabitats are a mixture of both lentic and lotic conditions: deep,

moderate flow with gravel substrate. Fig. 3.1 provides a conceptual diagram of the spatial

distribution of different mesohabitats within a stream.

Definition of species preference for different microhabitat has eluded malacologists, but

mesohabitat preferences are known for different mussel tribes (Haag, 2012; Holland-Bartels,

1990). The Anodontini and Lampsilini tribes prefer lentic mesohabitats and are specialized for

sand/clay substrate and low flow (Haag and Warren Jr., 2010; Haag, 2012). Species of the

Quadrulini tribe are usually found within main channel habitats, in either riffles or deep runs,

but can exhibit variation between species. Other tribes, specifically the Amblemini tribe, are

mesohabitat. Some species exhibit a high correlation with specific mesohabitats. Quadrula

verrucosa is known to inhabit riffles, often with females on the substrate surface (Haag, 2012;

20

Howells et al., 1996). Table 3.1 lists the species found in the Leon River zooarchaeological

assemblages by tribe and its expected preferred mesohabitat.

Fig. 3.1. Conceptual diagram of mesohabitats within river systems. This diagram shows different mesohabitats found within a segment of river. The grey regions are indicative of lentic areas; these areas are protected from strong flows and often have sand substrate. The white regions of the river represent areas of strong, but consistent flow and often have gravel substrate. Taken from Haag (2012).

Study Area

The Leon River is a low-order stream in central Texas. The river was impounded in 1954

and 1963, to create the Belton Lake and Proctor Lake reservoirs (Hendrickson Jr., 1981;

Randklev et al., 2013a). This river experiences a high amount of urban and agricultural runoff

(Harmel et al., 2008; Rossi et al., 2008). A longitudinal survey of the Leon River was conducted

in 2011 and found a unionid community with decreased richness and abundance compared to

historical data (Randklev et al., 2013a). While the nominal differences between the late

Holocene and modern unionid communities are explored in Chapter 2, this chapter evaluates

differences in the community structures and related habitat types at an ordinal scale.

21

Table 3.1 Mussels of the Leon River & preferred habitat. Tribes adapted from Haag (2012), which is adapted from Campbell et al. (2005) and Graf & Cummings (2007). Mesohabitat estimates are found in the following sources: (Haag, 2012; Hammontree, 2013; Howells et al., 1996). * indicates the species is either a candidate or petitioned for federal listing on the Endangered Species Act.

Leon River Taxa Tribe Mesohabitat Amblema plicata Amblemini Lotic Arcidens confragosus Anodontini Lentic Cyrtonaias tampicoensis Lampsilini Fusconaia mitchelli* Pleurobemini Lotic Lampsilis hydiana Lampsilini Lentic Lampsilis sp. Lampsilini Lentic Lampsilis teres Lampsilini Lentic Leptodea fragilis Lampsilini Lentic Megalonaias nervosa Quadrulini Transitional Potamilus purpuratus Lampsilini Transitional Pyganodon gradis Anodontini Lentic Quadrula apiculata Quadrulini Lentic Quadrula houstonensis* Quadrulini Lotic Quadrula sp. Quadrulini Lotic Quadrula verrucosa Quadrulini Transitional Toxolasma sp. Lampsilini Lentic Truncilla macrodon* Lampsilini Uniomerus tetralasmus Quadrulini Utterbackia imbecillis Anodontini Lentic

Methods

The goal of this chapter is to compare the structure of the Leon River’s unionid

community from the late Holocene to today using inferential statistics. These communities are

defined using two different types of data: zooarchaeological assemblages and a modern

longitudinal survey. Since zooarchaeological data are ordinal at best (Grayson, 1984;

Wolverton, 2013; Wolverton et al., 2014), non-parametric inferential statistics are used in this

study. The non-repetitive element (NRE) is used to quantify the number of mussels represented

22

for each taxon in the zooarchaeological assemblages (see Chapter 2 for further discussion of

NRE). Two zooarchaeological assemblages are used to represent the late Holocene unionid

community: the 41HM61 and the Belton Lake assemblages. 41HM61 is a zooarchaeological

assemblage that contains 399 NRE and represents an upstream portion of the Leon River below

the modern Lake Proctor reservoir. The Belton Lake assemblages are an aggregation of small

zooarchaeological assemblages surrounding the modern Lake Belton reservoir; this assemblage

contains 525 NRE. The modern freshwater mussel community is represented by a longitudinal

survey conducted by Randklev et al. (2013). During this survey, mussels were gathered using

qualitative methods that utilized a survey method known for capturing rare and endangered

species (Metcalfe-Smith et al., 2000; Randklev et al., 2013a). The unionid communities

represented by the 41HM61, Belton Lake, and the modern assemblage are described in terms

of taxonomic relative abundance in Fig. 3.2.

23

Fig. 3.2. Relative taxonomic abundances of late Holocene and mordern assemblages from the Leon River. Zooarchaeological assemblages are represented first (41HM61 and the Belton Lake assemblages). Taxa are listed in alphabetic order to facilitate comparison of different community structures.

Before the inferential tests are completed, the zooarchaeological and modern data sets

need to be consolidated and made comparable by eliminating taphonomically relevant and

extirpated taxa. Taphonomically relevant species are those that are thin-shelled and as such

their abundances are potentially altered by taphonomic processes (Wolverton et al., 2010). In

24

this study, taphonomically relevant species include Leptodea fragilis, Pyganodon grandis,

Toxolasma sp., Uniomerus tetralasmus, and Utterbackia imbecillis. Abundance data at the

generic level (e.g. Quadrula sp. and Lampsilis sp.) are excluded as they are not comparable to

the modern data, which were identified to species. Four species have been extirpated from the

Leon River: Cyrtonaias tampicoensis, Fusconaia mitchelli, Lampsilis hydiana and Truncilla

macrodon (Randklev et al., 2013a). These species are excluded from the chi square goodness of

fit test as their absences (or 0% relative abundance) cause undefined terms in the calculations.

Removing the taphonomically relevant and extirpated species makes the data sets comparable

and the inferential tests more conservative. The inferential tests are also conducted including

all taxa identified to the species level to provide clarity and for quality control purposes

(Simmons et al., 2011). The results of the inferential tests using the complete data can be found

in Table 3.2. Since some species are excluded from the analysis, relative abundances and rank-

ordered abundances for the late Holocene and modern data are recalculated (Appendix B).

A comparison between the late Holocene and today’s unionid assemblages is conducted

using chi-square goodness of fit tests and Spearman’s rho correlations. Chi-square goodness of

fit tests have been used before to compare prehistoric and modern unionid assemblages in

Kansas (Miller et al., 2014). A chi-squared goodness of fit test is a categorical test that compares

observed values to expected values in terms of relative abundance. Using the paleo and

modern data as observed and expected values respectively, this test determines if the relative

abundances differ between these two communities. The null hypothesis of a chi-square

goodness of fit test is that the observed values match the expected values, or that the late

Holocene species’ relative abundances match the modern species’ relative abundances.

25

A Spearman’s rho correlation compares the rank order of two data sets to determine if

there is a correlation. This has been used in applied zooarchaeological studies before to assess

differences between the age structure of modern and prehistoric assemblages of Missouri

bears (Wolverton, 2006). The null hypothesis of the Spearman’s rho correlation is that the

assemblages are not correlated, indicating there is a difference between the assemblages in

terms of rank order abundance. A strong correlation would indicate that the community

structures match in terms of ranked taxonomic abundance (most abundant species then are

still the most abundant now). A weak correlation would indicate that the community structures

are different (most abundant species then are lower ranked now). Both of these inferential

tests determine if the prehistoric unionid community matches the modern unionid community

in terms of taxonomic structure. Each inferential test is completed by comparing the individual

zooarchaeological assemblages (the 41HM61 and the Belton Lake assemblages) to the modern

data and then the aggregated late Holocene data to the modern data.

An unconstrained NMDS ordination of all Leon River unionid communities is used to

evaluate similarity between the taxonomic structure and therefore mesohabitat preferences of

the modern and zooarchaeological sub-assemblages. NMDS ordination is a nonmetric form of

multi-dimensional scaling, which identifies variables that cause objects to be similar or different

(Kachigan, 1991; Kindt and Coe, 2005). The variables are then condensed onto dimensions,

which relate the similarity or differences among objects based on those variables, and a

perception map is generated. Stress is a statistical measure of the loss of information due to

fewer dimensions. The fewer dimensions used in a perception map, the more stress an

ordination has. Stress is considered acceptable when it is below 0.15 (Kachigan, 1991). The

26

variables used in this NMDS ordination are relative abundances of unionid taxa; the objects are

the mesohabitat sub-assemblages and zooarchaeological assemblages from the Leon River. By

arranging the data sets this way, differences in the taxonomic relative abundances of the

assemblages will be related on the two dimensions of the perception map. Unconstrained

NMDS ordinations are completed in R using the package vegan (R Core Team, 2014). Bray-Curtis

dissimilarity calculation was used to determine the distance matrix that is used to complete the

NMDS ordination. An ordination was conducted among unionid assemblages that had more

than 30 NRE or mussels represented. Another NMDS ordination was conducted on samples

aggregated by time period and mesohabitat type for conceptual purposes. Prehistoric

assemblages are aggregated by zooarchaeological assemblage and modern assemblages are

aggregated based upon mesohabitat type.

Results

Results of the chi-square tests generally indicate that modern and late Holocene

communities in the Leon River differ. The test comparing the Belton Lake assemblage and the

modern assemblage determined that there was a significant difference between the two

assemblages, with strong effect size (Table 3.2). The residuals indicate that the difference

between the Belton Lake assemblage and the modern community is driven by the relative

abundances of Amblema plicata and Quadrula verrucosa. The test comparing the 41HM61

assemblage and the modern assemblage appears to have been driven by differences between

the relative abundances of A. plicata and Quadrula houstonensis. When the late Holocene

assemblages are aggregated, differences are driven by relative abundances of A. plicata and Q.

verrucosa. In total, the chi-square goodness of fit tests indicate that the structure of the late

27

Holocene unionid community is different than today’s unionid community, primarily between

the abundances of A. plicata, Q. houstonensis, and Q. verrucosa.

The Spearman’s rho correlation between the Belton Lake and modern freshwater

mussel assemblage resulted in rho = 0.44, a moderate correlation between the data sets,

indicating there are some abundance rank switches between these two assemblages. The

results of the Spearman’s rho correlation between the 41HM61 assemblages and the modern

freshwater mussel assemblage indicate that rho = 0.74, a strong correlation between the data

sets, meaning the rank-ordered abundances of the 41HM61 assemblage are similar to those of

the modern assemblage. When the late Holocene assemblages are aggregated, rho = 0.63, a

moderate correlation. In each of these correlations, the largest abundance rank shifts occur

among Fusconaia mitchelli, Lampsilis hydiana, and Potamilus purpuratus, which are species

either missing in the modern or late Holocene assemblages.

Table 3.2 Results of all chi-square goodness of fit test and Spearman’s rho correlation. For the chi-square goodness of fit tests, p<0.05. The degrees of freedom on the culled data are 7; the critical value is 14.1. For the complete data, degrees of freedom are 10 and the critical value is 18.3. In the chi-squared goodness of fit tests, extirpated species are excluded because of their expected abundance of 0. For the culled data, the Spearman’s rho correlations included 13 taxa. For the complete data, 17 taxa are included in the spearman rho correlations. See Appendix B for data used in these inferential tests.

Inferential Test

Culled Data Complete Data Belton

Lake vs Modern

41HM61 vs Modern

Late Holocene

vs. Modern

Belton Lake vs

Modern 41HM61 vs

Modern

Late Holocene

vs. Modern Chi-Square Goodness of Fit Test

χ2 678.9 169.9 721.5 841.1 222.4 919.1

Cohen’s W 1.21 0.89 1.03 1.34 1.02 1.16

Spearman’s rho

Rho 0.44 0.74 0.63 0.34 0.45 0.38 Correlation

strength Moderate Strong Strong Weak Moderate Weak

28

An unconstrained NMDS ordination assesses similarity between mesohabitat sub-

assemblages of the modern Leon River community and the zooarchaeological assemblages. To

do this, a Bray-Curtis dissimilarity calculation is completed to create a distance matrix that

describes the assemblages’ similarities in terms of taxonomic abundance (Kindt and Coe, 2005;

R Core Team, 2014). This analytical technique finds both prehistoric assemblages are most

similar to each other and lotic mesohabitats (Fig. 3.3). The NMDS ordination of the

unaggregated data (Fig. 3.3A) had a stress of 0.066. This indicates that the ordination is a good

representation of differences among the sites. The first dimension (x-axis) used in the NMDS

perception map quantified community differences between lotic and lentic mesohabitat sub-

assemblages. The NMDS ordination of the aggregated data did not contain enough objects to

accurately calculate a stress value for the ordination. It is included in this chapter to

conceptualize the similarities seen in the unaggregated NMDS ordination. From these

ordinations, it is understood that the late Holocene assemblages likely reflect lotic

mesohabitats since they more closely resemble the modern lotic sub-assemblages. This NMDS

ordination and the inferential tests are used to understand differences between the modern

and late Holocene freshwater mussel assemblages of the Leon River.

29

Fig. 3.3. NMDS ordination of freshwater mussel assemblages based on assemblage species structure. Part A is an NMDS ordination of individual sites on the Leon River that had more than 30 mussels represented. Sites starting with M are from the longitudinal survey conducted by Randklev et al. (2013). Sites starting with P are from late Holocene zooarchaeological assemblages, either Belton Lake (BL) or Hamilton County (41HM61). The red area includes modern unionid sub-assemblages from lotic mesohabitats (white areas in Fig. 3.1). The green area includes modern unionid sub-assemblages from lentic mesohabitats (grey areas in Fig. 3.1). Mesohabitat categorization defined in Randklev et al. (2013). Part B is an NMDS ordination of sites aggregated by time period or mesohabitat type.

Discussion

The structure of the freshwater mussel assemblage has changed since the late Holocene

in terms of species composition (see Chapter 2) and relative abundances. The residuals of the

chi-square goodness of fit tests indicate that differences in the relative abundances of A.

plicata, Q. verrucosa, and Q. houstonensis are greatest. Amblema plicata constitute a higher

proportion of the late Holocene assemblages than in the modern assemblages. Quadrula

verrucosa constitute a lower proportion of the late Holocene assemblages than in the modern

assemblages. The difference in the proportion of Q. verrucosa is greatest in the Belton Lake

assemblage. Quadrula houstonensis constitute a smaller proportion of the 41HM61 assemblage

30

than in the modern assemblage. These shifts in community structure could be attributed to

preservation potential, identifiability or ecological change. These possible forms of equifinality

are discussed further in Chapter 4, but this comparison provides evidence of a change in

abundance of Q. houstonensis, a species of conservation concern (TPWD, 2009; U.S. Fish and

Wildlife Service, 2014). Since Q. houstonensis is considered sensitive to anthropogenic impacts,

evidence of abundance shifts since the Late Holocene could be indicative of habitat change in

the Leon River (Howells, 2010a; Howells et al., 1996).

In the Spearman’s rho correlations, the greatest rank switches occur among species that

have been extirpated from the river basin, mainly F. mitchelli and L. hydiana. Since these

species were omitted from the goodness of fit test, which cannot subsume categories with

frequencies of zero, their large rank abundance shifts evident in the Spearman’s rho correlation

are important. The extirpations of F. mitchelli and L. hydiana are discussed in Chapter 2. While

C. tampicoensis was not found in the modern longitudinal survey, its rank abundance has not

shifted as much as other extirpated species. Cyrtonaias tampicoensis is more abundant (ranked

7) in the Belton Lake assemblages compared to all other unionid assemblages of the Leon River,

which aligns with expectations that this species is generally rare in this tributary but more

abundant near the confluence of the Leon and Little Rivers. Potamilus purpuratus also exhibited

a sizeable shift in rank since it is missing from the late Holocene assemblages. This could

potentially be a preservation issue in the zooarchaeological assemblages, or could be related to

a historical range extension into the Leon River. While some rank-order abundances have

switched since the late Holocene, the more abundant species in the past remain higher ranked

in the modern assemblages.

31

The NMDS ordination of the modern mesohabitat communities and the late Holocene

assemblages is informative if problematic. The NMDS ordination relies on the fact that the

zooarchaeological assemblages represent the prehistoric freshwater mussel community’s

structure accurately. As Chapter 4 describes, representation of unionid communities by

zooarchaeological assemblages are affected by multiple mechanisms (Wolverton et al., 2010).

Both late Holocene assemblages are more similar in taxonomic structure to lotic communities

than lentic communities. If the zooarchaeological assemblages are assumed to accurately

represent community structure, this similarity could be the result of two different processes:

either the Leon River was more lotic in the past or the prehistoric humans collected unionids

preferentially from lotic mesohabitats. In either case, this chapter stands as evidence to

preserve remaining lotic mesohabitat in the Leon River since many unionid species, especially

those of conservation concern (F. mitchelli and Q. houstonensis), prefer lotic mesohabitats

(Howells, 2010a, 2010b; Howells et al., 1996).

The results of this comparison between the modern and late Holocene freshwater

mussel fauna suggests ecological change in the Leon River unionid community since the late

Holocene, potentially with an increase in lentic mesohabitat over time. Impoundments of the

Leon River have altered the hydrologic regime of the Leon River, which alters habitat structure

and the resulting modern freshwater mussel community (Allen et al., 2013; Baldigo et al., 2004;

Vaughn and Taylor, 1999). As habitat is intrinsically entwined with life history strategies of

unionids, habitat change directly impacts conservation efforts (Haag, 2013, 2012; Southwood,

1988, 1977). This chapter shows that the Leon River unionid community has changed since the

late Holocene and provides a new conservation baseline for wildlife managers. By conserving

32

the remaining lotic habitat and managing the hydrologic regime, results of this analysis indicate

that mangers would be promoting habitat preferred by native freshwater mussels.

Conclusion

The Leon River has experienced many anthropogenic impacts through multiple

impoundments and an influx of urban and agricultural runoff (United States Department of

Agriculture, 2005; Harmel et al., 2008; Hendrickson Jr., 1981; Randklev et al., 2013; Rossi et al.,

2008), which are likely detrimental to riverine fauna abundance and diversity. A comparison

between the modern and late Holocene freshwater mussel community structure of the Leon

River found that the unionid community structure has changed since the late Holocene and

potentially preferred lotic mesohabitat assemblages. This comparison revealed that Q.

houstonensis, a federal candidate for listing on the Endangered Species Act that is sensitive to

anthropogenic impacts, has increased in abundance since the late Holocene (Texas Parks and

Wildlife Department, 2009; Howells, 2010a; Howells et al., 1996; U.S. Fish and Wildlife Service,

2014). The late Holocene zooarchaeological assemblages also are most similar to lotic

mesohabitat sub-assemblages from the modern Leon River. This could indicate that

impoundments have increased the lentic mesohabitats in this system. This chapter stands as

evidence of the late Holocene freshwater mussel community of the Leon River and promotes

the conservation of this unique community by providing another baseline for conservation.

33

CHAPTER 4

A CONCEPTUAL MODEL FOR ASSESSING FRESHWATER MUSSEL TAXONOMIC ABUNDANCES IN

ZOOARCHAEOLOGICAL FAUNAS

Introduction

Zooarchaeology is the study of faunal remains from archaeological contexts to study

human behavior, past environments, and past animal ecology (Lyman, 2005; Reitz and Wing,

2008; Wolverton and Lyman, 2012). Freshwater mussel (hereafter unionid) remains are

prevalent in many zooarchaeological assemblages and have been used to answer many

archaeological questions (Peacock, 2005). Paleozoological and zooarchaeological unionid

presence/absence data are most often used as evidence of shifts in human subsistence or

biogeographic distributions of taxa (Parmalee and Klippel, 1974; Peacock and Chapman, 2001;

Peacock, 2012). Abundance data from zooarchaeological assemblages can provide additional

data with which to approach these types of questions. For example, abundance data can be

used to study human behavior through foraging theory to study shifts in diet breadth (Byers

and Broughton, 2004; Nagaoka, 2005, 2002). Such data are also used in paleoenvironmental

studies, for conservation purposes, or to assess environmental changes in species abundance as

an alternative hypothesis to change in human subsistence (Braje et al., 2012; Casey, 1986;

Klippel et al., 1978; Peacock and Seltzer, 2008; Peacock et al., 2005). Abundance data can also

be used to quantify ecological community structures, which has implications for conservation

and paleoecology (Grayson, 1987; Miller et al., 2014; Randklev and Lundeen, 2012). Because

taxonomic abundance data are used to address such a wide variety of research questions by

zooarchaeologists and paleozoologists, it is important to understand that the structure of such

34

data. Abundance data produced from zooarchaeological assemblages can be influenced by

many different forces, such as population abundance on the prehistoric landscape, differential

preservation of shells, preferences of prehistoric people who incorporated unionids in their

diets, and/or differential identifiability of some remains over others (Kidwell and Flessa, 1995;

Lyman, 1994; Peacock et al., 2012). These influences then function as alternative mechanisms

that might drive patterns in taxonomic abundance.

Three distinctive mechanisms are addressed in this chapter: variable life history

strategies and their influence on population abundances, identifiability of shells, and

preservation potential of shells from different unionid taxa. Life history strategies play an

integral role in constructing ecological communities (Pianka, 1970; Southwood, 1988, 1977)

and, thus, potentially influence taxonomic abundances in zooarchaeological data (Kidwell and

Rothfus, 2010; Kidwell, 2013). Differential life history strategies could directly impact the

abundance of different taxa within life assemblages, death assemblages and resulting

excavated assemblages. In addition, differences in accuracy and precision of taxonomic

identification could directly affect zooarchaeological mussel abundance. Identifiability is

positively correlated with the presence of sculpture; shells with sculpture are identified to

species more often than those without sculpture (Shea et al., 2011). In addition to identifiability

and abundance related to life history ecology, quantitative taxonomic zooarchaeological data

may be affected by differential preservation of shells from one taxon over another. Different

unionid taxa have distinctive shell preservation potential based on the species’ shell sphericity

and shell density (Wolverton et al., 2010). Together these factors have complex but predictable

influences on zooarchaeological freshwater mussel taxonomic abundance data; thus, a

35

conceptual predictive model that provides general expectations about which species ought to

and ought not to be abundant can aid research in zooarchaeology and paleozoology that

focuses on unionids.

Rank order continua of life history strategy, identifiability (based on sculpture), and

preservation potential are developed for taxa encountered in the Leon River assemblages to

produce a multivariate conceptual model that allows zooarchaeologists to more easily interpret

the meaning of taxonomic abundance data in terms of taphonomy, ecology, and human

behavior. Specifically, this conceptual model will help interpret the comparison between the

modern and late Holocene freshwater mussel communities of the Leon River completed in

Chapter 3 (Table 4.1).

Table 4.1 Taxonomic abundances within the 41HM61 and Belton Lake assemblages. Missing values listed in the Belton Lake Assemblages indicate no specimens of that taxon were identified. Non-repetitive element (NRE) is a quantification unit used to count the total number of mussels represented in an assemblage; it is more fully described in Chapter 2. See Appendix 1 for identification criteria for the 41HM61 assemblage.

Taxa 41HM61 Belton Lake Assemblages

NRE Relative Abundance NRE Relative Abundance

Amblema plicata 93 23.3% 307 54.2% Arcidens confragosus 1 0.3% 1 0.2% Cyrtonaias tampicoensis 2 0.5% 6 1.1% Fusconaia mitchelli 9 2.3% 38 6.7% Lampsilis sp. 37 9.3% Lampsilis hydiana 10 2.5% 14 2.5% Lampsilis teres 34 8.5% 5 0.9% Megalonaias nervosa 16 4.0% 2 0.4% Quadrula sp. 34 8.5% 1 0.2% Quadrula apiculata 13 3.3% 20 3.5% Quadrula houstonensis 20 5.0% 118 21.0% Quadrula verrucosa 38 9.5% 13 2.3% Toxolasma sp. 2 0.5% Truncilla macrodon 5 1.3% Unidentifiable umbo 85 21.3% 41 7.1% Total NRE 399 566

36

Background

Representativeness should be assessed to understand potential explanations for the

structure and composition of zooarchaeological taxonomic abundance data (Grayson, 1987;

Lyman, 2012a, 1994). The excavated assemblage from which zooarchaeological data are

produced has passed through many filters prior to analysis; these filters include aggregation

into a deposited assemblage, time in the lithosphere, potential sampling biases, and analysis by

zooarchaeologists, each potentially resulting in various forms of data loss (as seen in Fig. 4.1).

Some of these filters have been analyzed before for freshwater mussels, such as the ‘cultural

filter’ and potential for differential preservation of shell between the time of human

consumption and archaeological recovery (Peacock et al., 2012; Wolverton et al., 2010). For

example, as mussels were collected from streams by humans, there is potential for preferential

selection of particular species and individuals, which could lead to unrepresentative sampling of

the past ecological community. Peacock et al. (2012) addressed the problem of this ‘cultural

filter’ and provided methods, such as nestedness and isotopic profiling, for analyzing an

assemblage’s representativeness of past local ecological communities. Nestedness analysis was

used in Chapter 2 to evaluate whether or not 41HM61 and the Belton Lake assemblages

represent the same ecological community in terms of species composition.

37

Fig. 4.1. Model describing potential filters within zooarchaeological assemblages. Certain steps (light gray) of the life-cycle of zooarchaeological assemblages are shown to demonstrate different factors that can affect final zooarchaeological data. Though many processes (smaller words between steps) act on each of these steps, only processes relevant to this study were included to explain where different factors affect zooarchaeological remains. Note that at each step the assemblage gets smaller, indicating the potential for data loss as remains transition from the biological community through the taphonomic history to zooarchaeological data. Based on figures from Lyman (2010), Clark and Kietzke (1967), and Behrensmeyer and Kidwell (1985).

In terms of taphonomy, shells from unionid species exhibit different shapes and

structures (Wolverton et al., 2010), which can affect potential for preservation and thus the

quantity or presence of a species in an assemblage. Wolverton et al. (2010) addressed this

preservation potential of shells from different species of unionids in zooarchaeological

assemblages based on two different physical characteristics of shell: shell density and shell

sphericity. Using modern specimens from the Brazos and Trinity River, they determined that as

a shell’s density increases, its preservation potential increases as well. As shell’s sphericity

increases, its preservation potential also increases. A taphonomic analysis is conducted for the

Biological Community

Deposited Assemblage

Excavated Assemblage

Zooarchaeological Data

Human Collection

Sampling Bias

Zooarchaeologist Identification

Zooarchaeological Assemblage

Differential Preservation TIME

38

41HM61 and the Belton Lake assemblages to assess whether or not differential preservation

has affected taxonomic abundance.

There are other filters beyond preservation and foraging preference that can affect

remains and influence zooarchaeological abundance data, for example, identifiability and/or

experience of the faunal analyst. Differences in identifiability of shells between species relates

to two factors: distinctiveness of shell morphology and preservation potential related to

whether or not shells are prone to fragmentation. One way that zooarchaeologists account for

differences in identifiability and also bolster confidence in data quality is to fully describe

identification criteria (Driver, 2011, 1992; Wolverton, 2013). The problem remains that

specimens are often fragmented and eroded, making identifications difficult. This problem is

exacerbated by the fact that identification of live unionids in ecological surveys is difficult due

to morphological plasticity within and between species. Shea et al. (2011) addressed the

identification accuracy and precision of malacologists, and found that increased sculpture, size,

and conservation status (listed as threatened by a government agency) lead to higher

identification rates of individual mussel taxa. Those species with sculptured shells tend to be

more recognizable, those with conservation status tend to be better studied and thus more

identifiable, and those of larger size also tend to be recognizable and easier to identify. In this

study, since sculpture preserves well on fragments of shell from zooarchaeological contexts and

is often diagnostic, it will be used as a proxy for identifiability. Therefore in zooarchaeological

assemblages, species that have sculpture should be identified with more taxonomic precision

(e.g., to finer taxonomic levels, such as species) and more accurately (to the correct taxon) than

those without sculpture if identifiability is driving taxonomic abundance.

39

The life history strategy of a mussel species also has the potential to affect its taxonomic

abundance in a deposited assemblage. Life history strategies describe a species’ differential

allocation of energy based on the rate of population growth and reproductive ecology (Fisher,

1930; Macarthur and Wilson, 1967; Pianka, 1970). As different life history strategies among

species are expected to impact the abundances in a community, they should indirectly affect

the abundances in a zooarchaeological assemblage (Kidwell and Rothfus, 2010; Kidwell, 2001;

Southwood, 1988, 1977). Typically, Pianka’s r versus K selection gradient is used to define life

history strategies among different species (Olden et al., 2006; Pianka, 1970; Southwood, 1977,

1988). While unionids are generally categorized as long lived and slow growing, they exhibit a

wide range of variability of life history characteristics (Haag, 2012; Vaughn, 2012). Unionid life

history strategies are categorized into three types that are based on Winemiller and Rose

(1992)’s life history strategy three endpoint continuum: opportunistic, periodic, and equilibrium

strategies (Dillon Jr., 2000; Grime, 2001; Haag, 2012; Winemiller and Rose, 1992). The

opportunistic life history strategy is similar to Pianka’s r-selection; such mussel species are

characterized by a short life span, early maturity, and high fecundity. Equilibrium selected

mussels live long and mature later in their lives, similar to Pianka’s K-selection. Periodic-

selected mussels are “characterized by moderate to high growth rate, low to intermediate life

span, [low] age at maturity, and [low] fecundity” (Haag, 2012, p. 211). Periodic species are

generally adapted to habitats that experience cyclical environmental variation, which create

beneficial conditions during which they reproduce (Haag, 2012; Winemiller and Rose, 1992).

Species from the Leon River zooarchaeological assemblages are categorized at the taxonomic

tribe level, which is an appropriate scale for categorizing unionids by life history strategy. Table

40

4.2 illustrates the life history strategies common to tribes within central Texas and lists the

species represented in the Leon River zooarchaeological assemblages in their respective tribes.

Table 4.2 Robusticity and life history strategy values for Leon River zooarchaeological freshwater mussel assemblage. Highlighted cells indicate higher expected abundances if variable life history strategy, identifiability or robusticity influences taxonomic abundances. Robusticity mean values are found in Wolverton et al. (2010), which were calculated from a modern Brazos River assemblage. The sphericity and density values are summed to determine the ordinal rank of different species. Some species’ robusticity values were not calculated (due to small sample size); robusticity rank is uncalculated for these species. Life history strategy is determined at a nominal scale using the figure from Haag (2012), seen below as Fig. 4.2. Data Sources: 1Wolverton et al., 2010; 2 Haag, 2012; 3Tribe rank - Howells, 2013 and Haag, 2012; 4Obliquaria reflexa as a proxy - Haag, 2012 and Campbell et al., 2005. Abbreviations that are used by the authors are listed below as well and are used throughout this chapter.

Species Abbrev-iation

Robusticity Tribe2 Life History

Strategy Sculpture Category Sphericity1 Density1 Summed Rank

Amblema plicata AP 0.56 2.02 2.58 2 Amblemini Equilibrium2 High Arcidens confragosus AC Anodontini Opportunistic3 High Cyrtonaias tampicoensis CT 0.54 1.92 2.46 4 Lampsilini Periodic4 None Fusconaia mitchelli FM Pleurobemini Equilibrium3 Low Lampsilis teres LT 0.44 1.27 1.71 7 Lampsilini Opportunistic2 None Lampsilis hydiana LH 0.51 1.13 1.64 8 Lampsilini Opportunistic3 None Megalonaias nervosa MN 0.48 1.52 2.00 5 Quadrulini Equilibrium2 High Quadrula apiculata QA 0.61 1.90 2.51 3 Quadrulini Equilibrium3 High Quadrula houstonensis QH 0.63 1.24 1.87 6 Quadrulini Equilibrium3 Low Quadrula verrucosa QV 0.46 2.46 2.92 1 Quadrulini Equilibrium3 High Toxolasma sp. Ts 0.60 0.34 0.94 9 Lampsilini Periodic3 None Truncilla macrodon TM Lampsilini Opportunistic3 None

Study Area

Two zooarchaeological assemblages are studied to evaluate how the three processes

described above affect abundance data; both assemblages comprise late Holocene freshwater

mussel remains from the Leon River. The Belton Lake faunas come from 18 separate cave sites

that surround the modern Belton Lake reservoir; these sites represent the Leon River close to

its confluence with the Little River (Randklev, 2010). Those assemblages were identified by

41

Charles Randklev, and the fragmentation rate of shell remains from these rock shelter deposits

is low. The 41HM61 assemblage represents the Leon River approximately 60 miles upstream of

the Belton Lake faunas. The 41HM61 assemblage was identified by Traci Popejoy;

fragmentation is high in this assemblage. Nestedness analysis in Chapter 2 determined that

these assemblages appear to represent the same faunal community in terms of species

composition (Peacock et al., 2012), but represent two distinct preservation contexts.

A comparison between the structure of the Leon River freshwater mussel assemblages

from the late Holocene and a modern longitudinal survey is fully described in Chapter 3. From

this comparison, it is evident that Amblema plicata, Quadrula verrucosa, and Quadrula

houstonensis do not constitute the same proportion of the late Holocene assemblage as they

do in the modern assemblage. Amblema plicata is over represented in both zooarchaeological

assemblages. Quadrula verrucosa is underrepresented in the zooarchaeological assemblages,

though the difference is largest in the Belton Lake faunas. Quadrula houstonensis is

underrepresented in the 41HM61 fauna. The conceptual model constructed in this chapter is

used to assess how these patterns in the 41HM61 and Belton Lake assemblages conform to

expectations about taxonomic abundances.

Limitations

Auto-correlation among life history strategy, sculpture, and robusticity possibly disrupts

the interpretive value of this conceptual model. Since equilibrium species allocate energy

toward increased interspecific competitive ability, longer life span, and higher overall growth,

they are likely to produce a more dense shell than opportunistic species (Haag and Rypel, 2011;

Haag, 2013, 2012). Life history strategies are also possibly correlated with shell ornamentation

42

through two different processes. For many organisms, K-selection includes increased

interspecific competition which increases the likelihood for ornamentation (Mccullough, 1999;

Pianka, 1972, 1970). For unionids, correlation between shell ornamentation/sculpture and life

history strategies has not been assessed. Sculpture is a convergent evolutionary trait and

relates to flow variability and predation pressure (Haag, 2012; Hornbach et al., 2010). Since

sculpture often reflects the habitat of unionid species and life history strategies are influenced

by habitat conditions, it is possible that these two factors are correlated (Pianka, 1970;

Southwood, 1988, 1977). There may be aspects of life history, preservation potential of shells,

and sculpture that co-occur, but each also potentially influences abundance independently.

Developing conceptual expectations for each variable allows for clearer interpretation of

zooarchaeological abundance data. In addition, since species’ shell robusticity, shell sculpture

and life history strategy can vary between mussel communities, consideration of these

variables and the expectations should be limited to use within the river basin represented by

the zooarchaeological assemblage.

Materials and Methods

Unionid species can be categorized along an ordinal continuum that indicates their

potential for either being high or low in abundance according to their life history, identifiability,

and preservation potential, which is summarized in Table 4.3. Taken together, an opportunistic

species that produces a robust, sculptured shell has a higher probability of occurrence and high

abundance in zooarchaeological assemblages than an equilibrium species with fragile shell

morphology and a lack of sculpturing.

43

Table 4.3 Expectations for abundance filters discussed in this chapter. High abundance potential refers to the expectation that species with this trait will have higher abundances relative to the other species without this trait.

Filter High Abundance Potential Low Abundance Potential

Life History Strategy Species produce many offspring – Opportunistic Strategy

Species produce few offspring – Equilibrium Strategy

Identifiability Shells with sculpture Shells with no sculpture Preservation Potential Shells are spherical and dense Shells are non-spherical and are

less dense

To define the life history strategy of mussels, data from Haag’s chapter (“These are very

different animals: Life history variation in mussels”) are used to delineate mussels into the

three categories: opportunistic, periodic, and equilibrium (2012). Using Haag’s ordination figure

as a guide (Fig. 4.2), species present in the figure were assigned their expected life history

strategy categorization. When a species’ life history strategy is not found in Haag’s data,

phylogenetic and taxonomic relationships were used to make assignments (see Table 4.2 and

Campbell et al., 2005). Haag’s description of unionid life history strategies is used because it is

the most current data and provides categories to rank the species along a gradient.

44

Fig. 4.2. Ordination of life history strategies of freshwater mussels in the southeastern United States. Unionids were ordinated into life history strategy categories based on fecundity, body size, age at maturity and life span. This figure is found in Haag (2012), Chapter 6. The abbreviations can be found in that chapter. The species used in this study are boxed in red and include Apli- A. plicata, Oref- Obliquaria reflexa, Lter- Lampsilis teres, Mner- Megalonaias nervosa, Pohi- Potamilus ohiensis. Also note that all Quadrulid species (Qpus, Qasp, Qrum) are found within the equilibrium strategy.

Shell sculpture, defined as the presence of plications, wrinkles, pustules, bumps, or

ridges on a shell margin (Hornbach et al., 2010; Watters, 1994), improves identification

accuracy and precision of modern unionids (Shea et al., 2011). Shell sculpture has been

quantified by calculating the number of pustules in a squared centimeter area (Peacock and

Seltzer, 2008). In this chapter, sculpture is evaluated at a nominal scale since fragmentation and

erosion could affect the presence of sculpture on unionid specimens. Sculpture is categorized

into three groups: shells with a high density of sculpture on its outer disk (>50%), shells with

less dense sculpture on their outer margin (<50%), and shells without sculpture; identifiability

categorizations are based on sculpture on modern unionid shells from the Brazos River and its

tributaries.

45

For the robusticity analysis, two physical characteristics of shells are quantified using

data from Wolverton et al. (2010). As the Leon River is a tributary of the Brazos River, use of

Wolverton et al. (2010)’s calculated mean density and sphericity values for taxa is appropriate.

Shell sphericity was calculated by a ratio of shell length, width, and height. Shell density

describes the structural strength of the shell and was measured through volume displacement.

The mean sphericity and density values from Wolverton et al. (2010) are summed to create a

robusticity value, which is then ranked for the Leon River freshwater mussel community (see

Table 4.2). Ranking robusticity quantifies a taxon’s preservation potential within a given

zooarchaeological assemblage at an ordinal scale. A further analysis of the differential

preservation evident in both assemblages is conducted by constructing a 3D scatterplot to

evaluate preservation based on density and sphericity independently following Wolverton et al.

(2010). If life history characteristics, shell robusticity, and shell identifiability are influencing the

zooarchaeological taxonomic abundance data, it can be expected that opportunistic species

with robust, sculptured shells will be higher in abundance than equilibrium species with fragile,

unsculptured shells.

Using this multivariate conceptual model, it can be predicted which unionid species will

be most and least abundant in the Leon River zooarchaeological assemblages (Table 4.4). If

there is auto-correlation in this model, A. plicata, A. confragosus, Q. apiculata, and Q. verrucosa

have high expected abundances since these species exhibit two characteristics likely to increase

abundance. Fusconaia mitchelli, Q. houstonensis, and Toxolasma sp. are expected to be least

abundant as they lack characteristics expected to produce high abundances.

46

Table 4.4 Zooarchaeological taxonomic abundance expectations for the Leon River unionid community. Taxa are listed by alphabetical order, not by likelihood of abundance. Robusticity ranks included the three most abundant and three least abundant species.

Alternative Mechanism

High Abundance Low/Rare Abundance

Life History Strategy

Lampsilis hydiana Lampsilis teres Truncilla macrodon

Amblema plicata Quadrula apiculata Quadrula houstonensis Quadrula verrucosa

Identifiability

Amblema plicata Arcidens confragosus Megalonaias nervosa Quadrula apiculata Quadrula verrucosa

Cyrtonaias tampicoensis Lampsilis hydiana Lampsilis teres

Robusticity Amblema plicata Quadrula apiculata Quadrula verrucosa

Lampsilis hydiana Lampsilis teres Toxolasma sp.

Histograms are used to evaluate how well abundance expectations match the relative

abundance patterns in the zooarchaeological assemblages. The x-axis in each histogram is

arranged from most to least expected from left to right according to each of the three variables

(life history strategy (Fig. 4.3), sculpture and identifiability (Fig. 4.4), and robusticity (Fig. 4.5).

Results

As seen in Fig. 4.3, life history strategy does not relate to taxonomic abundance within

these zooarchaeological assemblages, as opportunistic species, such as Lampsilis teres and

Lampsilis hydiana, are expected to have high abundances. The 41HM61 assemblage is

dominated by equilibrium species, but does have a moderate proportion of opportunistic

strategists as well. Both assemblages have a low abundance of taxa that exhibit periodic life

history strategies. This is expected since many periodic species have low abundance in modern

rivers (Howells, 2013; Howells et al., 1996). The Belton Lake assemblages were dominated by

equilibrium strategists, with only a small proportion of opportunistic taxa. Since equilibrium

47

taxa dominate both assemblages, it can be stated that life history strategy does not appear to

have influenced taxonomic abundance.

Fig. 4.3. Analysis of life history strategy’s influence on the 41HM61 and the Belton Lake assemblages.

As evident in Fig. 4.4, identifiability can influence abundance data generated from

zooarchaeological assemblages. It is expected that species with highly sculptured shells, such as

Q. verrucosa and A. plicata, should exhibit higher relative abundances than species with

unsculptured shells. Both the 41HM61 and the Belton Lake assemblages are dominated by

species with sculptured shells, primarily A. plicata. Quadrula houstonensis has a high relative

abundance in both assemblages, but typically has less sculpture than Quadrula apiculata and

can even be apustulose (without sculpture). While the abundance of some taxa appears to

relate to the presence of sculpture (i.e. A. plicata), other taxa were identified based on other

physical features, such as psuedocardinal teeth (i.e. Q. houstonensis). In general, identifiability,

whether from sculpture or other morphological criteria, seems to have a stronger influence on

the abundance data from the 41HM61 assemblage than does life history.

48

Fig. 4.4. Analysis of sculpture’s influence on the 41HM61 and Belton Lake assemblages.

Generally, species with more robust shells have a higher relative abundance, as seen in

Fig. 4.5. A more in-depth analysis of how physical preservation influenced the abundance data

from these zooarchaeological assemblages is found in Fig. 4.6. The 41HM61 assemblage

includes taxa that are categorized as robust and fragile; species in both categories have high

ranked relative abundances. The Belton Lake assemblages are also affected by differential

preservation; the shells with the highest sphericity have the highest relative abundance. Both

zooarchaeological assemblages are dominated by robust taxa, but the presence of fragile taxa

with moderate abundance indicates the taxonomic abundances in these assemblages are not

solely explained by differential preservation and thus may be representative of the late

Holocene freshwater mussel community of the Leon River at a nominal or ordinal scale.

49

Fig. 4.5. Robusticities’ influence on the 41HM61 and Belton Lake assemblages.

Fig. 4.6. Taphonomic analysis of the 41HM61 assemblage and the Belton Lake assemblages. Sphericity and robusticity values for each species were ordinally ranked from Wolverton et al (2010). The darker blue represents species with high preservation potential; these species have high shell density and high shell sphericity. The lighter blue represents species with low preservation potential; these species are less dense and less spherical.

50

Discussion

This chapter contributes to the zooarchaeological literature by exploring alternative

mechanisms for the structure of unionid abundance data, which are used in many ways to

answer archaeological questions: to understand human diet, paleoenvironmental conditions,

and the paleoecology of mussel taxa. By understanding alternative mechanisms for taxonomic

abundance, zooarchaeologists can determine when forms of equifinality influence the results of

zooarchaeological studies. The conceptual framework presented in this chapter allows

zooarchaeologists to evaluate the influence of life history strategies, identifiability, and

robusticity on taxonomic abundance data. Counter to what was expected, equilibrium life-

history strategists were commonly represented in the zooarchaeological assemblages. This may

relate to sculpture and robusticity auto-correlating with life history traits. Equilibrium life

history strategists are likely to invest more energy into phenotypic expression (Geist, 1998;

Pianka, 1972, 1970), such as shell mass and complexity. In addition, robusticity to some extent

may mediate identifiability as larger fragments tend to have better preserved shell morphology.

Variable life history strategy’s influence on the structure of deposited assemblages cannot be

completely ruled out, however. If more detailed data on the fecundity of different freshwater

mussels in Texas becomes available, comparisons within life history strategy categories might

better elucidate whether or not reproductive ecology influences zooarchaeological assemblage

abundance. As a result, how variable life history strategies’ influence zooarchaeological data

remains an avenue for future research.

A comparison between the late Holocene and modern freshwater mussel community of

the Leon River is used as an example of the utility of this conceptual model. Without

51

understanding how alternative mechanisms influence taxonomic abundances, differences

between the two assemblages might be attributed to ecological change. For example, the

expectations presented here include that A. plicata should have high abundance in

zooarchaeological assemblages while Q. houstonensis is expected to have low abundance. Thus,

the differences in abundance between the Late Holocene and modern unionid community

could be the result of differential preservation, identifiability, or ecological change. Quadrula

verrucosa shells are robust and very identifiable; as such this species is expected to be high in

abundance. The comparison finds Q. verrucosa to be rarer in the late Holocene than today.

Since Q. verrucosa’s shift in abundance is opposite what is predicted by the conceptual model,

it can be more confidently attributed to ecological change in stream condition between the late

Holocene and today.

Conclusion

By assessing variable life history strategies, identifiability, and robusticity’s influence on

zooarchaeological taxonomic abundance data, whether patterns of abundance in

zooarchaeological assemblages are representative changes in human foraging behavior,

paleoenvironmental conditions, or community structure are analyzed. This evaluation revealed

that differential robusticity and identifiability influence the 41HM61 and Belton Lake

assemblages. There are alternative mechanisms for explaining the abundance of different taxa

in zooarchaeological assemblages and the implications of those data for the comparison

between the modern and late Holocene unionid assemblages of the Leon River (found in

Chapter 3). While robusticity and identifiability could not be distinguished from ecological

change as driving factors for the abundance of some taxa, Q. verrucosa is more abundant today

52

than in the late Holocene, which suggests that important environmental changes have occurred

in the Leon River, perhaps related to modern human impacts.

53

CHAPTER 5

CONCLUSION

The comparison between the late Holocene and modern freshwater mussel community

of the Leon River revealed multiple changes in this unionid community. In terms of species

composition, multiple species have been extirpated from the Leon River since the late

Holocene. These species include C. tampicoensis, F. mitchelli, L. hydiana, and T. macrodon. Two

of these species (F. mitchelli and T. macrodon) are of conservation concern, meaning they are

listed as threatened by the state of Texas and are either petitioned or a candidate for federal

listing on the Endangered Species Act (Texas Parks and Wildlife Department, 2009; U.S. Fish and

Wildlife Service, 2014). All of these species absences from the Leon River stand as evidence of

range decline, which is evidence of population decline and contributes to a potential extinction

vortex. The multiple analytical techniques used in Chapter3 to assess ecological change

revealed multiple shifts in community structure. While most rank-ordered abundances

remained the same between the late Holocene and modern assemblages (with the exception of

extirpated taxa), many of the taxonomic relative abundances have significantly changed since

the late Holocene. The chi-square goodness of fit test revealed that relative abundances of A.

plicata, Q. houstonensis, and Q. verrucosa have changed. The nonmetric multidimensional

scaling (NMDS) ordination revealed that the late Holocene freshwater mussel assemblages are

most similar in taxonomic structure to lotic mesohabitat sub-assemblages. Since freshwater

mussels are indicator species, this evidence of differences between the late Holocene and

modern unionid communities show ecological change in the Leon River.

54

A multivariate conceptual model is used to assess the influence of life history ecology,

identifiability, and robusticity on zooarchaeological taxonomic abundances data. This model

contributes to the zooarchaeological literature by improving interpretations of

zooarchaeological taxonomic abundance data. This utility is of this model is exampled by using

it to interpret the results of the chi square goodness of fit test found in Chapter 3. By comparing

the taxonomic abundances of 41HM61 and the Belton Lake assemblages to the model, it is

evident that identifiability and robusticity has influenced taxonomic abundances in these

zooarchaeological assemblages. Thus, identifiability, robusticity, and ecological change could be

driving the differences in the abundances of A. plicata and Q. houstonensis. This model

improves interpretation of the abundance shift of Q. verrucosa. Since Q. verrucosa is expected

to be highly abundant in zooarchaeological assemblages, but is actually less abundant in the

late Holocene, its abundance change is attributable to ecological change.

This thesis is an example of how zooarchaeological data can be used for conservation

purposes. It provides a late Holocene conservation baseline for the Leon River unionid

community, which is appropriate as this stream experienced anthropogenic impacts beginning

in the middle of the twentieth century, which is before most historical records were created for

the unionid community of this river (Hendrickson Jr., 1981; Randklev et al., 2013a). This thesis

also explores unionid community change since the late Holocene and supports the conclusions

of the modern survey conducted by Randklev et al. (2013) that the unionid community has

decreased in species richness since the late Holocene and historical times. Evidence of

ecological change has implications of habitat change in the Leon River, which harbors

threatened mussels (F. mitchelli, Q. houstonensis, and T. macrodon) and a federally endangered

55

fish species (Smalleye Shiner - Notropis buccula). The results of this thesis hopefully influence

conservation of this unique community and the riverine ecosystem of the Leon River.

56

APPENDIX A

SYSTEMATIC PALEONTOLOGY

57

In a typical paleontological study there is a section entitled “Descriptive Paleontology” or “Systematic Paleontology.” There, all identified specimens are listed under each taxon, each specimen is described, and the anatomical and morphometric criteria used to make the identification are described verbally and exemplary specimens are illustrated.

R. Lee Lyman

Zooarchaeology often deals with unreliable data in the form of fragmented specimen

(Driver, 2011, 1992; Gobalet, 2001; Wolverton, 2013). As a way to make data more accessible

and verifiable, a systematic paleontology is included with this thesis. This work includes

descriptions of all criteria used to identify individual specimens to taxa. Some parts of the

unionid valve were unused for identification and are not described in all cases (such as

posterior ridge and lateral teeth). In the case of 41HM61, these parts were often absent on

highly fragmented shell, and therefore unavailable for identification purposes. Along with

personal experience, the following identification guides were utilized in this study: Field Guide

to Texas Freshwater Mussels (Howells, 2013), Freshwater Mussels of Texas (Howells et al.,

1996) and Freshwater Mussels of Alabama and the Mobile Basin in Georgia, Mississippi, and

Tennessee (Williams et al., 2008).

Order: Bivalve

Family: Unionidae

Genus: Amblema

Species: Amblema plicata

Description: Amblema plicata have oval to rectangular shells. The umbo is above the hinge line

and often appears bulbous (very round and inflated). The beak sculpture consists of single

looped ridges. Shells often have three to seven diagonal ridges on the outer margins, but can be

unsculptured when small. The posterior ridge is very obscure. The pseudocardinal teeth are

58

divergent, triangular and massive. The left psuedocardinal teeth tend to be slightly oblique. The

right psuedocardinal tooth points toward the post-ventral margin, which is useful for

differentiating it from Megalonaias nervosa and Quadrula verrucosa.

Genus: Arcidens

Species: Arcidens confragosus

Description: Shell shape is subquadrate to oval and thin. Beak sculpture is nodular. Shells have

intricate exterior sculpture that includes curving ridges, cross hatches, and pustules.

Psuedocardinal teeth are very compressed and flat. The right valve has one psuedocardinal

tooth, while the left valve has two teeth.

Genus: Cyrtonaias

Species: Cyrtonaias tampicoensis

Description: Shell shape is oval, where the umbo shape often looks like the top of a hoe where

it joins the central disk. The beak sculpture consists of ultra-fine single looped ridges that are

rarely visible, making it non-diagnostic. The beak cavity is moderately deep. The psuedocardinal

teeth look like chewed bubble gum, meaning they often are round with many grooves. The

right valve has one psuedocardinal tooth that is triangular with rounded edges. The left valve

has two psuedocardinal teeth that are strong, almost massive. The posterior ridge is almost

non-existent, with a broadly rounded central disk.

Genus: Lampsilis

Description: Specimen identified as Lampsilis sp. exhibit compressed psuedocardinal teeth. The

umbo is often low, with the psuedocardinal teeth close to the beak with no interdentum. The

59

beak sculpture consists of fine broken lines that can be weekly double looped. Lampsilis sp. also

exhibit an oval shape, with a broadly rounded posterior ridge.

Species: Lampsilis hydiana

Description: L. hydiana shells are oval and inflated. The umbo is often low and deep, with a

larger area under the umbo than L. teres. The beak has double loop sculpturing. The

psuedocardinal teeth are generally slightly thicker than L. teres and ‘look cheeky’. The right

valve has one psuedocardinal tooth that is often approximately parallel to the hinge line. The

left valve has two psuedocardinal teeth.

Species: Lampsilis teres

Description: L. teres shells are shaped like very long ovals. The umbo is very low and shallow.

The beak has thin double loop sculpturing. The psuedocardinal teeth are thin and compressed.

The right valve has one psuedocardinal tooth that is often parallel to the hinge line. The left

valve has two psuedocardinal teeth that are thin, compressed, and parallel.

Genus: Megalonaias

Species: Megalonaias nervosa

Description: M. nervosa shells are rectangular shaped. They often contain ridges and

sculpturing in the middle margin. The umbo is low on the shell and shallow. Beak sculpture is

double looped to nodular. The pseudocardinal teeth are molar like. The right valve has one

psuedocardinal tooth that points left of the center of the ventral margin. The left valve has two

psuedocardinal teeth; the posterior tooth is often larger than the anterior tooth.

60

Genus: Quadrula

Description: Quadrulids have diagnostic psuedocardinal teeth. The right valve has a triangular

psuedocardinal tooth that is often torn. The left valve has two psuedocardinal teeth. Of these

two left teeth, the anterior psuedocardinal tooth is noticeably larger than the posterior

psuedocardinal tooth. Quadrulids are quadrate, subquadrate, to rectangularly shaped.

Species: Quadrula apiculata

Description: Quadrula apiculata is quadrate to triangular in shape. The psuedocardinal teeth

are typical of the genus. The right valve has one psuedocardinal tooth that is shaped like an

isosceles triangle, and are often torn. The left valve is typical of other quadrulids (see above). It

often has diagnostic pustules covering a majority of the outer shell surface, including

throughout the sulcus. These pustules are regular and small.

Species: Quadrula houstonensis

Description: Quadula houstonensis is quadrate to triangular in shape. The pseudocardinal teeth

are typical of the genus. The left valve’s psuedocardinal teeth include the typical size

differentiation; the anterior psuedocardinal teeth is usually much larger than the posterior

psuedocardinal teeth. Pustules can occur on the outer shell surface of this species, but are at

lower densities than pustules on both Q. apiculata and Q. verrucosa.

Species: Quadrula verrucosa

Description: Oval to oblong shape. Beak is pointed and low over the hinge line. Shells have

many pustules on them with fluting on the entire post-ventral margin. Pustules can continue

until the ventral margin. The posterior ridge is very strong, which is useful for differentiating it

61

from Megalonaias nervosa. The pseudocardinal teeth are strong and divergent. The right

psuedocardinal tooth is very triangular, with a flat edge on the anterior side.

Genus: Toxolasma

Description: Specimen’s designated Toxolasma sp. exhibited compressed psuedocardinal teeth

with diagnostic beak sculpture. Toxolasma sp. are subeliptical in shape and very small (hence

the common name of Lilliput). The beak is low, just above the hinge line, with a relatively

shallow beak cavity. The beak sculpture is diagnostic as it consists of large, single looped ridges.

Howells describes them as “curved bars… [that] produce a comma shape”. The right valve has

one psuedocardinal tooth; the left valve has two psuedocardinal teeth. All psuedocardinal teeth

are small, and peg-like. Specimen were not identified to the species level as differentiation is

difficult without the entire outer shell, which was not available often in this assemblage.

Genus: Truncilla

Species: Truncilla macrodon

Description: Truncilla macrodon has an elongated oval shape. The umbo is just above the hinge

line, with a very small beak cavity. The beak sculpture consists of very fine ridges. The

psuedocardinal teeth are compressed, very triangular, and close to the umbo. The right valve

has one psuedocardinal tooth that is triangular with the smallest side pointed toward the

anterior margin. The left valve has two psuedocardinal teeth. The lateral teeth are thin and

slightly curved.

Mixed Taxa

These designations were used to determine what shell they represented when the taxa couldn’t

be determined to species level. Often, comments within the data will explain why the

62

designation was given due to individual parts of the element. Preservation within the sample

often limited the confidence with which a species-level label would have been appropriate.

Species: Amblema plicata or Megalonaias nervosa

Description: This categorization was used for shells that exhibited characteristics similar to both

A. plicata and M. nervosa. These characteristics included similarities between the

psuedocardinal teeth. If the left psuedocardinal teeth were massive and angled, they looked

like both of these species. Often, the beak/umbo is the first to erode. As the beak above the

hinge line is useful for differentiating these two species, this erosion causes problems. The

presence of hatchmark sculpture is diagnostic of M. nervosa, but within this sample this

sculpture was eroded away. Without these distinguishing characteristics, these elements were

placed within this broader category for more confident and precise data.

Species: Quadrula apiculata or Quadrula verrucosa

Description: This categorization was used for shells that exhibited Quadrulid teeth and also had

a high density of pustules or fluting on the margins.

63

APPENDIX B

DATA USED IN CHAPTER 3INFERENTIAL TESTS

64

Culled Data

Table B.1 Data for culled chi-square goodness of fit test. Taphonomically relevant and extirpated species are excluded from this test. Numbers from the late Holocene sites are non-repetitive elements (NRE) identified to that taxon. Late Holocene data are used as observed values while modern relative abundances are used as the expected values.

Taxa 41HM61 Belton Lake Late Holocene

Modern Relative

Abundances Amblema plicata 93 307 400 20.1%

Arcidens confragosus 1 1 2 0.2% Lampsilis teres 34 5 39 6.3%

Megalonaias nervosa 16 2 18 4.9% Potamilus purpuratus 0 0 0 0.4%

Quadrula apiculata 13 20 33 1.7% Quadrula houstonensis 20 118 138 35.0%

Quadrula verrucosa 38 13 51 31.3% Total Mussels 215 466 681 1814

Table B.2 Data for culled Spearman’s rho correlation. Taphonomically relevant species are excluded from this test. Numbers represent taxonomic abundance rank.

Taxa Hamilton Belton Late

Holocene Modern Amblema plicata 1 1 1 3

Arcidens confragosus 12 10 11.5 8 Cyrtonaias tampicoensis 10.5 7 9 11

Fusconaia mitchelli 8 3 4 11 Lampsilis hydiana 7 5 7 11

Lampsilis teres 3 8 5 4 Megalonaias nervosa 5 9 8 5 Potamilus purpuratus 13 12 13 7

Quadrula apiculata 6 4 6 6 Quadrula houstonensis 4 2 2 1

Quadrula verrucosa 2 6 3 2 Truncilla macrodon 9 12 10 11

Toxolasma sp. 10.5 12 11.5 11

65

Complete Data Table B.3 Data for complete chi-square goodness of fit test. Numbers from the late Holocene sites are NRE identified to that taxon. Late Holocene data are used as observed values while modern relative abundances are used as the expected values.

Taxa 41HM61 Belton Lake Late Holocene

Modern Relative

Abundances Amblema plicata 93 307 400 17.6%

Arcidens confragosus 1 1 2 0.2% Lampsilis teres 34 5 39 5.6%

Leptodea fragilis 0 0 0 11.1% Megalonaias nervosa 16 2 18 4.3% Potamilus purpuratus 0 0 0 0.3%

Pyganodon grandis 0 0 0 1.0% Quadrula apiculata 13 20 33 1.5%

Quadrula houstonensis 20 118 138 30.6% Quadrula verrucosa 38 13 51 27.4%

Utterbackia imbecillis 0 0 0 0.3% Total Mussels 215 466 681 2071

Table B.4 Data for complete Spearman’s rho correlation. Taphonomically relevant species are excluded from this test. Numbers represent taxonomic abundance rank.

Taxa Hamilton Belton Late Holocene Modern Amblema plicata 1 1 1 3

Arcidens confragosus 12 10 11.5 11 Cyrtonaias tampicoensis 10.5 7 9 15

Fusconaia mitchelli 8 3 4 15 Lampsilis hydiana 7 5 7 15

Lampsilis teres 3 8 5 5 Leptodea fragilis 15 14 15 4

Megalonaias nervosa 5 9 8 6 Pyganodon grandis 15 14 15 8

Potamilus purpuratus 15 14 15 9 Quadrula apiculata 6 4 6 7

Quadrula houstonensis 4 2 2 1 Quadrula verrucosa 2 6 3 2 Truncilla macrodon 9 14 10 15

Toxolasma sp. 10.5 14 11.5 15 Utterbackia imbecillis 15 14 15 10

Uniomerus tetralasmus 15 14 15 12

66

REFERENCES

Allen, D.C., Galbraith, H.S., Vaughn, C.C., Spooner, D.E., 2013. A tale of two rivers: implications of water management practices for mussel biodiversity outcomes during droughts. Ambio 42, 881–91. doi:10.1007/s13280-013-0420-8

Allen, D.C., Vaughn, C.C., 2010. Complex hydraulic and substrate variables limit freshwater mussel species richness and abundance. J. North Am. Benthol. Soc. 29, 383–394. doi:10.1899/09-024.1

Anderson, R., Engeling, A., Grones, A., Lopez, R., Pierce, B., Skow, K., Snelgrove, T., 2014. Texas Land Trends. College Station, TX.

Atmar, W., Patterson, B.D., 1993. The Measure of Order and Disorder in the Distribuiton of Species in Fragmented Habitat. Oecologia 96, 373–382.

Baldigo, B.P., Riva-Murray, K., Schuler, G.E., 2004. Effects of environmental and spatial features on mussel populations and communities in a North American river. Walkerana 14, 1–32.

Behrensmeyer, A.K., Kidwell, S.M., 1985. Taphonomy’s contributions to paleobiology. Paleobiology 105–119.

Bettinger, R.L., Malhi, R., McCarthy, H., 1997. Central place models of acorn and mussel processing. J. Archaeol. Sci. 887–899.

Braje, T.J., Rick, T.C., Erlandson, J.M., 2012. Rockfish in the View: Applied Zooarchaeology and Conservation of Pacific Red Snapper (Genus Sebastes) in Southern California, in: Wolverton, S., Lyman, R.L. (Eds.), Conservation Biology and Applied Zooarchaeology. University of Arizona Press, pp. 157–178.

Burlakova, L.E., Karatayev, A.Y., Karatayev, V.A., May, M.E., Bennett, D.L., Cook, M.J., 2011. Biogeography and conservation of freshwater mussels (Bivalvia: Unionidae) in Texas: patterns of diversity and threats. Divers. Distrib. 17, 393–407. doi:10.1111/j.1472-4642.2011.00753.x

Byers, D.A., Broughton, J.M., 2004. Holocene environmental change, artiodactyl abundances, and human hunting strategies in the Great Basin. Am. Antiq. 69, 235–255.

Campbell, D.C., Serb, J.M., Buhay, J.E., Roe, K.J., Minton, R.L., Lydeard, C., 2005. Phylogeny of North American amblemines (Bivalvia, Unionoida): prodigious polyphyly proves pervasive across genera. Invertebr. Biol. 124, 131–164. doi:10.1111/j.1744-7410.2005.00015.x

67

Cannon, K.P., Cannon, M.B., 2004. Zooarchaeology and Wildlife Management in the Greater Yellowstone Ecosystem, in: Lyman, R.L., Cannon, K.P. (Eds.), Zooarchaeology and Conservation Biology. University of Utah Press, pp. 45–60.

Casey, J.L., 1986. The Prehistoric Exploitation of Unionacean Bivalve Molluscs in the Lower Tennessee-Cumberland-Ohio River Valleys in Western Kentucky. Simon Fraser University.

Claassen, C., 2000. Quantifying Shell: Comments on Mason, Peterson, and Tiffany. Am. Antiq. 65, 415–418.

Clark, J., Kietzke, K.K., 1967. Paleoecology of the lower nodular zone, Brule Formation, in the Big Badlands of South Dakota. Fieldiana Geol. Mem. 5, 111–137.

Cummings, K.S., Graf, D.L., 2010. Mollusca: Bivalvia, Third Edit. ed, Ecology and Classification of North American Freshwater Invertebrates. Elsevier Ltd. doi:10.1016/B978-0-12-374855-3.00011-X

Dennis, S.D., 1984. Distributional analysis of the freshwater mussel fauna of the Tennessee River system, with special reference to possible limiting effects of siltation.

Dillon Jr., R.T., 2000. The Ecology of Freshwater Molluscs. Cambridge U Press, Cambridge, UK.

Driver, J.C., 1992. Identification, classification and zooarchaeology. Circaea 9, 35–47.

Driver, J.C., 2011. Identification , Classification and Zooarchaeology. Ethnobiol. Lett. 2, 19–39.

Fisher, R.A., 1930. The genetical theory of natural selection.

Galbraith, H.S., Spooner, D.E., Vaughn, C.C., 2010. Synergistic effects of regional climate patterns and local water management on freshwater mussel communities. Biol. Conserv. 143, 1175–1183. doi:10.1016/j.biocon.2010.02.025

Geist, V., 1998. Deer of the World: Their Evolution, Behaviour, and Ecology. Stackpole Books.

Giovas, C.M., 2009. The shell game: analytic problems in archaeological mollusc quantification. J. Archaeol. Sci. 36, 1557–1564. doi:10.1016/j.jas.2009.03.017

Glassow, M.A., 2000. Weighing vs. Counting Shellfish Remains: A Comment on Mason, Peterson, and Tiffany. Am. Antiq. 65, 407–414.

Gobalet, K.W., 2001. A Critique of Faunal Analysis; Inconsistency among Experts in Blind Tests. J. Archaeol. Sci. 28, 377–386. doi:10.1006/jasc.2000.0564

68

Grayson, D.K., 1984. The Basic Counting Units, in: Grayson, D.K. (Ed.), Quantitative Zooarchaeology. Academic Press, New York, NY, pp. 16–92.

Grayson, D.K., 1987. The biogeographic history of small mammals in the Great Basin: observations on the last 20,000 years. J. Mammal. 68, 359–375.

Grime, J.P., 2001. Plant Strategies, Vegetation Processes, and Ecosystem Properties, 2nd ed. John Wiley & Sons Ltd, New York, NY.

Haag, W.R., 2012. North American Freshwater Mussels: Natural History, Ecology and Conservation. Cambridge U Press, Cambridge, UK.

Haag, W.R., 2013. The role of fecundity and reproductive effort in defining life-history strategies of North American freshwater mussels. Biol. Rev. Camb. Philos. Soc. 88, 745–66. doi:10.1111/brv.12028

Haag, W.R., Rypel, A.L., 2011. Growth and longevity in freshwater mussels: evolutionary and conservation implications. Biol. Rev. Camb. Philos. Soc. 86, 225–47. doi:10.1111/j.1469-185X.2010.00146.x

Haag, W.R., Warren Jr., M.L., 2008. Effects of Severe Drought on Freshwater Mussel Assemblages. Trans. Am. Fish. Soc. 137, 1165–1178. doi:10.1577/T07-100.1

Haag, W.R., Warren Jr., M.L., 2010. Diversity, abundance, and size structure of bivalve assemblages in the Sipsey River, Alabama. Aquat. Conserv. Mar. Freshw. Ecosyst. 20, 655–667. doi:10.1002/aqc.1138

Haag, W.R., Williams, J.D., 2013. Biodiversity on the brink: an assessment of conservation strategies for North American freshwater mussels. Hydrobiologia 735, 45–60. doi:10.1007/s10750-013-1524-7

Hammontree, S., 2013. Evaluating the Habitat Requirements of the Golden Orb Mussel (Quadrula aurea) for Conservation Purposes. University of North Texas.

Harmel, R.D., Rossi, C.G., Dybala, T.J., Arnold, J., Potter, K., Wolfe, J., Hoffman, D., 2008. Conservation Effects Assessment Project research in the Leon River and Riesel watersheds. J. Soil Water Conserv. 63, 453–460. doi:10.2489/jswc.63.6.453

Hendrickson Jr., K.E., 1981. The Waters of the Brazos: A History of the Brazos River Authority from 1929 - 1979, 1st ed. The Texian Press.

Holland-Bartels, L.E., 1990. Physical factors and their influence on the mussel fauna of a main channel border habitat of the upper Mississippi River. J. North Am. Benthol. Soc. 9, 327–335.

69

Hornbach, D.J., Kurth, V.J., Hove, M.C., 2010. Variation in Freshwater Mussel Shell Sculpture and Shape Along a River Gradient. Am. Midl. Nat. 164, 22–36. doi:10.1674/0003-0031-164.1.22

Howells, R.G., 1997. Distributional Surveys of Freshwater Bivalves in Texas: Progress Report for 1996. Austin, TX.

Howells, R.G., 2010a. Smooth Pimpleback (Quadrula houstonensis): Summary of Selected Biological and Ecological Data for Texas. Austin, TX.

Howells, R.G., 2010b. False Spike (Quadrula mitchelli): Summary of Selected Biological and Ecological Data for Texas. Austin, TX.

Howells, R.G., 2013. Field Guide to Texas Freshwater Mussels. BioStudies, Kerrville, TX.

Howells, R.G., Neck, R.W., Murray, H.D., 1996. Freshwater Mussels of Texas. Texas Parks and Wildlife Press, Austin, TX.

Humphries, P., Winemiller, K.O., 2009. Historical Impacts on River Fauna, Shifting Baselines, and Challenges for Restoration. Bioscience 59, 673–684. doi:10.1525/bio.2009.59.8.9

Joyce, M., 2012. Constructing Nature: Art, Conservation, and Applied Zooarchaeology. J. Ethnobiol. 32, 246–264.

Kachigan, S.K., 1991. Multivariate Statistical Analysis: A Conceptual Introduction, 2nd ed. Radius Press.

Kidwell, S.M., 2001. Preservation of species abundance in marine death assemblages. Science (80-. ). 294, 1091–1094.

Kidwell, S.M., 2013. Time-averaging and fidelity of modern death assemblages: building a taphonomic foundation for conservation palaeobiology. Palaeontology 56, 487–522. doi:10.1111/pala.12042

Kidwell, S.M., Flessa, K.W., 1995. The quality of the fossil record: populations, species, and communities. Annu. Rev. Ecol. Syst. 26, 269–299.

Kidwell, S.M., Rothfus, T.A., 2010. The living, the dead, and the expected dead: variation in life span yields little bias of proportional abundances in bivalve death assemblages. Paleobiology 36, 615–640.

Kindt, R., Coe, R., 2005. Analysis of ecological distance by ordination, in: Tree Diversity Analysis: A Manual and Software for Common Statistical Methods for Ecological and Biodiversity Studiesversity Analysis. Worl Agroforestry Centre, p. 44.

70

Klippel, W.E., Celmer, G., Purdue, J.R., 1978. The Holocene Naiad Record at Rodgers Shelter in the Western Ozard Highland of Missouri. Plains Anthropol. 23, 257–271.

Louys, J., Wilkinson, D.M., Bishop, L.C., 2012. Paleontology in Ecology and Conservation 23–38. doi:10.1007/978-3-642-25038-5

Lyman, R.L., 1991. Late quaternary biogeography of the pygmy rabbit (Brachylagus idahoensis) in eastern Washington. J. Mammal. 72, 110–117.

Lyman, R.L., 1994. Vertebrate Taphonomy. Cambridge U Press, Cambridge, UK.

Lyman, R.L., 1996. Applied Zooarchaeology: The Relevance of Faunal Analysis to Wildlife Management. World Archaeol. 28, 110–125.

Lyman, R.L., 2005. Zooarchaeology, in: Maschner, H.D.G., Chippendale, C. (Eds.), Handbook of Archaeological Methods. AltaMira Press, Lanham, Maryland, pp. 835–870.

Lyman, R.L., 2008. Quantitative Paleozoology. Cambridge U Press, New York, NY.

Lyman, R.L., 2010. What Taphonomy Is, What it Isn’t, and Why Taphonomists Should Care about the Difference. J. Taphon. 8, 1–16.

Lyman, R.L., 2012a. A warrant for applied palaeozoology. Biol. Rev. Camb. Philos. Soc. 87, 513–25. doi:10.1111/j.1469-185X.2011.00207.x

Lyman, R.L., 2012b. Applied Zooarchaeology History, Value, and Use, in: Wolverton, S., Lyman, R.L. (Eds.), Conservation Biology and Applied Zooarchaeology. University of Arizona Press, Tuscan, AZ, pp. 208–232.

Lyman, R.L., Ames, K.M., 2007. On the use of species-area curves to detect the effects of sample size. J. Archaeol. Sci. 34, 1985–1990. doi:10.1016/j.jas.2007.01.011

Macarthur, R.H., Wilson, E.O., 1967. The theory of island biogeography. Princeton U Press.

Mason, R.D., Peterson, M.L., Tiffany, J.A., 1998. Weighing vs. Counting: Measurement Reliability and the California School of Midden Analysis. Am. Antiq. 63, 303–324.

Mccullough, D.R., 1999. Density Dependence and Life-History Strategies of Ungulates. J. Mammal. 80, 1130–1146.

Metcalfe-Smith, J.L., Di Maio, J., Staton, S.K., Mackie, G.L., 2000. Effect of samling effort on the efficiency of the timed search method for sampling freshwater mussel communities. J. North Am. Benthol. Soc. 19, 725–732.

71

Miller, E.J., Tomasic, J.J., Barnhart, M.C., 2014. A Comparison of Freshwater Mussels (Unionidae) from a Late-Archaic Archeological Excavation with Recently Sampled Verdigris River, Kansas, Populations. Am. Midl. Nat. 171, 16–26.

Nagaoka, L., 2002. The effects of resource depression on foraging efficiency, diet breadth, and patch use in southern New Zealand. J. Anthropol. Archaeol. 21, 419–442. doi:10.1016/S0278-4165(02)00008-9

Nagaoka, L., 2005. Declining foraging efficiency and moa carcass exploitation in southern New Zealand. J. Archaeol. Sci. 32, 1328–1338. doi:10.1016/j.jas.2005.04.004

Newsome, S.D., Etnier, M.A., Gifford-Gonzalez, D., Phillips, D.L., van Tuinen, M., Hadly, E.A., Costa, D.P., Kennett, D.J., Guilderson, T.P., Koch, P.L., 2007. The shifting baseline of northern fur seal ecology in the northeast Pacific Ocean. Proc. Natl. Acad. Sci. U. S. A. 104, 9709–14. doi:10.1073/pnas.0610986104

Nobles, T., Zhang, Y., 2011. Biodiversity Loss in Freshwater Mussels: Importance, Threats, and Solutions, in: Grillo, O. (Ed.), Biodiversity Loss in a Changing Planet. InTech Open, pp. 137–162.

Olden, J.D., Poff, N.L., Bestgen, K.R., 2006. Life-history strategies predict fish invasions and extirpations in the Colorado River Basin. Ecol. Monogr. 76, 25–40.

Palmer, M.A., Reidy Liermann, C.A., Nilsson, C., Flörke, M., Alcamo, J., Lake, P.S., Bond, N., 2008. Climate change and the world’s river basins: anticipating management options. Front. Ecol. Environ. 6, 81–89. doi:10.1890/060148

Parmalee, P.W., Klippel, W.E., 1974. Freshwater Mussels as a Prehistoric Food Resource. Am. Antiq. 39, 421–434.

Parmesan, C., Matthews, J., 2006. Biological Impacts of Climate Change, in: Principles of Conservation Biology. Sinauer Associates, Inc, Sunderland, Massachusetts, pp. 333–374.

Peacock, E., 2005. Environmental Archaeology. J. Alabama Archaeol. 51.

Peacock, E., 2012. Archaeological Freshwater Mussel Remains and Their Use in the Conservation of an Imperiled Fauna, in: Wolverton, S., Lyman, R.L. (Eds.), Conservation Biology and Applied Zooarchaeology. University of Arizona Press, Tuscan, AZ, pp. 42–68.

Peacock, E., Chapman, S., 2001. Taphonomic and Biogeographic Data from a Plaquemine Shell Midden on the Ouachita River, Lousiana. Southeast. Archaeol. 20, 44–55.

Peacock, E., Haag, W.R., Warren Jr., M.L., 2005. Prehistoric Decline in Freshwater Mussels Coincident with the Advent of Maize Agriculture. Conserv. Biol. 19, 547–551.

72

Peacock, E., Randklev, C.R., Wolverton, S., Palmer, R.A., Zaleski, S., 2012. The “cultural filter”, human transport of mussel shell, and the applied potential of zooarchaeological data. Ecol. Appl. 22, 1446–59.

Peacock, E., Seltzer, J.L., 2008. A comparison of multiple proxy data sets for paleoenvironmental conditions as derived from freshwater bivalve (Unionid) shell. J. Archaeol. Sci. 35, 2557–2565. doi:10.1016/j.jas.2008.04.006

Pianka, E.R., 1970. On r-and K-selection. Am. Nat. 104, 592–597.

Pianka, E.R., 1972. r and K selection or b and d selection? Am. Nat. doi:10.1086/282798

R Core Team, 2014. R: A Language and Environment for Statistical Computing. Vienna, Austria.

Randklev, C.R., 2010. Zooarchaeoloical analysis of the prehistoric mussel fauna from selected rock-shelters near Belton Lake: Belton County, Texas. Bothell, WA.

Randklev, C.R., 2011. The Ecology and Paleobiogeography of Freshwater Mussels (Family: Unionidae) from Selected River Basins in Texas.

Randklev, C.R., Johnson, M.S., Tsakiris, E., Groce, J., Wilkins, N., 2013a. Status of the freshwater mussel (Unionidae) communities of the mainstem of the Leon River, Texas. Aquat. Conserv. Mar. Freshw. Ecosyst. 23, 390–404. doi:10.1002/aqc.2340

Randklev, C.R., Johnson, M.S., Tsakiris, E., Rogers-Oetker, S., Roe, K.J., Harris, J.L., McMurray, S.E., Robertson, C., Groce, J., Wilkins, N., 2012. False Spike, Quadrula mitchelli (Bivalvia: Unionidae), is Not Extinct: First Account of a Live Population in Over 30 Years. Am. Malacol. Bull. 30, 327–328. doi:10.4003/006.030.0213

Randklev, C.R., Lundeen, B.J., 2012. Prehistoric Biogeography and Conservation Status of Threatened Freshwater Mussels (Mollusca: Unionidae) in the Upper Trinity River Drainage, Texas, in: Wolverton, S., Lyman, R.L. (Eds.), Conservation Biology and Applied Zooarchaeology. The University of Arizona Press, Tuscan, pp. 68–91.

Randklev, C.R., Tsakiris, E., Johnson, M.S., Skorupski, J.A., Burlakova, L.E., Groce, J., Wilkins, N., 2013b. Is False Spike, Quadrula mitchelli (Bivalvia: Unionidae), extinct? First account of a very recently deceased individual in over thirty years. Southwest. Nat. 58, 247–250.

Rashleigh, B., 2008. Nestedness in riverine mussel communities: patterns across sites and fish hosts. Ecography (Cop.). 31, 612–619. doi:10.1111/j.0906-7590.2008.05300.x

Reitz, E.J., Wing, E.S., 2008. Zooarchaeology, 2nd ed. Cambridge U Press, Cambridge, UK. doi:10.1002/oa.1113

73

Richardson, D.M., Whittaker, R.J., 2010. Conservation biogeography - foundations, concepts and challenges. Divers. Distrib. 16, 313–320. doi:10.1111/j.1472-4642.2010.00660.x

Ricklefs, R.E., 2008. The Economy of Nature, 6th ed. W. H. Freeman and Company, New York, NY.

Rossi, C.G., Dybala, T.J., Moriasi, D.N., Arnold, J.G., Amonett, C., Marek, T., 2008. Hydrologic calibration and validation of the Soil and Water Assessment Tool for the Leon River watershed. J. Soil Water Conserv. 63, 533–541. doi:10.2489/jswc.63.6.533

Shea, C.P., Peterson, J.T., Conroy, M.J., Wisniewski, J.M., 2013. Evaluating the influence of land use, drought and reach isolation on the occurrence of freshwater mussel species in the lower Flint River Basin, Georgia (U.S.A.). Freshw. Biol. 58, 382–395. doi:10.1111/fwb.12066

Shea, C.P., Peterson, J.T., Wisniewski, J.M., Johnson, N.A., 2011. Misidentification of freshwater mussel species (Bivalvia:Unionidae): contributing factors, management implications, and potential solutions. J. North Am. Benthol. Soc. 30, 446–458. doi:10.1899/10-073.1

Simmons, J.P., Nelson, L.D., Simonsohn, U., 2011. False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant. Psychol. Sci. 22, 1359–1366. doi:10.1177/0956797611417632

Soulé, M.E., 1985. What is Conservation Biology? Bioscience 35, 727–734.

Southwood, T.R.E., 1977. Habitat, the Templet for Ecological Strategies? J. Anim. Ecol. 46, 336–365.

Southwood, T.R.E., 1988. Tactics, Strategies and Templets. Oikos 52, 3–18.

Spurlock, B.D., 1981. The Naiads (Mollusca: Unionidae) From Two Prehistoric Sites Along the Ohio River, Mason County, West Virginia. Mashall University.

Strayer, D.L., 1999. Use of flow refuges by unionid mussels in rivers. J. North Am. Benthol. Soc. 18, 468–476.

Strayer, D.L., 2008. Freshwater Mussel Ecology: A Multifactor Approach to Distribution and Abundance. U of California Press, Berkeley.

Strayer, D.L., Downing, J.A., Haag, W.R., King, T.L., Layzer, J.B., Newton, T.J., Nichols, S.J., 2004. Changing Perspectives on Pearly Mussels, North America’s Most Imperiled Animals. Bioscience 54, 429. doi:10.1641/0006-3568(2004)054[0429:CPOPMN]2.0.CO;2

Strecker, J.K., 1931. The distribution of the naiades or pearly fresh-water mussels of Texas. Baylor Univ. Museum Bull. 2, 69.

74

Texas Parks and Wildlife Department, 2009. 15 Texas Freshwater Mussels Placed on State Threatened List [WWW Document]. TPWD News Releases. URL http://www.tpwd.state.tx.us/newsmedia/releases/?req=20091105c

U.S. Fish and Wildlife Service, 2014. Federal Register: Endangered and Threatened Wildlife and Plants; Review of Native Species That Are Candidates for Listing as Endangered or Threatened; Annual Notice of Findings on Resubmitted Petitions; Annual Description of Progress on Listing Actions.

Ulrich, W., 2006. Nestedness - a FORTRAN program for calculating ecological matrix temperatures [WWW Document]. URL www.uni.torun.pl/~ulrichw

Ulrich, W., Almeida-Neto, M., Gotelli, N.J., 2009. A consumer’s guide to nestedness analysis. Oikos 118, 3–17. doi:10.1111/j.1600-0706.2008.17053.x

United States Department of Agriculture, 2006. Conservation Effects Assessment Project: Leon River Watershed , Texas: 2004-2006. USDA.

Vaughn, C.C., 2012. Life history traits and abundance can predict local colonisation and extinction rates of freshwater mussels. Freshw. Biol. 57, 982–992. doi:10.1111/j.1365-2427.2012.02759.x

Vaughn, C.C., Pyron, M., 1995. Population ecology of the endangered Ouachita rock-pocketbook mussel, Arkansia wheeleri (Bivalvia: Unionidae), in the Kiamici River, Oklahoma. Am. Malacol. Bull. 11, 145–151.

Vaughn, C.C., Taylor, C.M., 1999. Impoundments and the Decline of Freshwater Mussels : a Case Study of an Extinction Gradient. Conserv. Biol. 13, 912–920.

Warren, R.E., 1991. Freshwater mussels as paleoenvironmental indicators: A quantitative approach to assemblage analysis, in: Purdue, J.R., Klippel, W.E., Styles, B.W., Parmalee, P.W. (Eds.), Beamers, Bobwhites, and Blue-Points: Tributes to the Career of Paul W. Parmalee. Illinois State Museum Scientific Papers, Springfield, Illinois, pp. 23–66.

Watters, T.G., 1994. Form and function of unionoidean shell sculpture and shape (Bivalvia). Am. Malacol. Bull. 11, 1–20.

Whittaker, R.J., Araújo, M.B., Jepson, P., Ladle, R.J., Watson, J.E.M., Willis, K.J., 2005. Conservation Biogeography: assessment and prospect. Divers. Distrib. 11, 3–23. doi:10.1111/j.1366-9516.2005.00143.x

Williams, J.D., Bogan, A.E., Garner, J.T., 2008. Freshwater Mussels of Alabama and the Mobile Basin in Geogria, Mississippi, and Tenessee. The U of Alabama Press, Tuscaloosa, AL.

75

Williams, J.D., Warren Jr., M.L., Cummings, K.S., Harris, J.L., Neves, R.J., 1993. Conservation Status of Freshwater Mussels of the United States and Canada. Fisheries 18, 6–22. doi:10.1577/1548-8446(1993)018<0006:CSOFMO>2.0.CO;2

Winemiller, K.O., Rose, K.A., 1992. Patterns of Life-History Diversification in North American Fishes: Implications for Population Regulation. Can. J. Fish. Aquat. Sci. 49, 2196–2218.

Wolverton, S., 2006. Natural-trap ursid mortality and the Kurtén Response. J. Hum. Evol. 50, 540–51. doi:10.1016/j.jhevol.2005.12.009

Wolverton, S., 2008. Harvest Pressure and Environmental Carrying Capacity: an Ordinal-Sale Model of Effects of Ungulate Prey. Am. Antiq. 73, 179–199.

Wolverton, S., 2013. Data quality in zooarchaeological faunal identification. J. Archaeol. Method Theory 20, 381–396. doi:10.1007/s10816-012-9161-4

Wolverton, S., Dombrosky, J., Lyman, R.L., 2014. Practical Significance: Ordinal Scale Data and Effect Size in Zooarchaeology. Int. J. Osteoarchaeol. doi:10.1002/oa.2416

Wolverton, S., Lyman, R.L., 2012. Introduction to Applied Zooarchaeology, in: Wolverton, S., Lyman, R.L. (Eds.), Conservation Biology and Applied Zooarchaeology. University of Arizona, Tuscan, AZ, pp. 1–22.

Wolverton, S., Randklev, C.R., Kennedy, J.H., 2010. A conceptual model for freshwater mussel (family: Unionidae) remain preservation in zooarchaeological assemblages. J. Archaeol. Sci. 37, 164–173. doi:10.1016/j.jas.2009.09.028

Wright, D.H., Patterson, B.D., Mikkelson, G.M., Culter, A., Atmar, W., 1997. A comparative analysis of nested subset patterns of species composition. Oecologia 113, 1–20.

76