Fish Coprolites of the Joggins Formation and Coastal Trophic Relationships in a Late Carboniferous Sea

Home , Coprolite

Fish coprolites of the Joggins Formation and coastal trophic relationships in a Late Carboniferous sea

Max D. Chipman

Submitted in partial fulfilment of the

requirements for the Degree of

Bachelor of Science with

Honours in Geology

Acadia University

October, 2017

This thesis by Max D. Chipman is accepted in its present form by the Department of Earth and Environmental Science as satisfying the thesis requirements for the degree of Bachelor of Science with Honours

Approved by the Thesis Supervisors

______(Dr. Peir K. Pufahl) Date ______(Dr. Melissa Grey) Date

Approved by the Head of the Department

______(Dr. Ian Spooner) Date

Approved by the Honours Committee

______(Matthew Lukeman) Date

iii

I, Max Chipman, grant permission to the University Librarian at Acadia University to reproduce, loan or distribute copies of my thesis in microform, paper or electronic formats on a non-profit basis. I however, retain the copyright to my thesis.

______

Max D. Chipman

______

Date

Acknowledgements

First and foremost, I wish to acknowledge Dr. Pufahl and Dr. Grey for allowing me to work with them on this incredible project. Their never-ending support and invaluable expertise were instrumental in helping me craft this thesis and guiding me in the right direction even when I didn’t know it. I would also like to thank Pam Frail for her work cutting thin sections of my samples. I would like to acknowledge that the funding was provided through an NSERC Undergraduate Summer Research Award and an NSERC Discovery Grant to Dr. Peir Pufahl. Last, but certainly not least, I would like to thank my family and friends for their unwavering support.

vii

viii

Table of Contents

Acknowledgements ...... vii

Table of Contents ...... ix

List of Figures ...... xi

Abstract ……...... xiii

Introduction …………...... 1

General Geology and Depositional Environments ...... 2

Paleoecology and Previous Research ...... 3

Objectives and Implications ...... 9

Materials and Methods ...... 9

Categorization of the Joggins Coprolites ……...... 11

Thin Section and Hand Sample Analysis ……………...... 12

Cathodoluminescence and Computerized Tomography ……...... 17

Provenance …………………………………………………………………..…………. 17

Discussion ...... 18

Conclusions ...... 27

References ...... 30

Appendix ...... 33

List of Figures

Figure 1. Location of the Joggins Fossil Cliffs ...... 1

Figure 2. The Joggins Fossil Cliffs UNESCO World Heritage Site ...... 5

Figure 3. The Joggins Formation and associated formations ...... 6

Figure 4. The Joggins Formation: facies associations and cyclicity ...... 7

Figure 5: Hand sample photographs of coprolite categories ...... 13

Figure 6. Pictomicrographs of coprolites ...... 15

Figure 7: Hand sample photographs of coprolites of interest ...... 16

Figure 8: Pictomicrographs of coprolite thin sections under cathodoluminescence ...... 20

Figure 9. Coprolites images made with computed tomography ...... 21

xii

Abstract

The fossil cliffs at Joggins (Nova Scotia) hold a wealth of fossils, both terrestrial and aquatic, from the Late Carboniferous Period. Fossils from the aquatic realm have historically been understudied and the ecosystem they represent is poorly understood.

This research broadens our understanding of the aquatic ecosystem, specifically the food web, by examining fish coprolites that are abundant in the limestones of the Joggins

Formation. Coprolites preserve undigested material that give us a window into the diets of these fish and a better idea of species interactions within the ecosystem. The coprolites have been studied in thin section and hand sample, as well as cathodoluminescence and computed tomography to determine the contents. The specimens were divided into six categories based on size and shape: cigar/cylindrical shaped; cone shaped; small/equant; spiral; irregular; and massive (samples greater than 5 cm in length). The small coprolites are the most abundant and the massive coprolites are the rarest. They range in size from

<1 cm to >10 cm and are 2-3 centimetres on average. The mineralogy of the coprolites is high calcium phosphate, similar to the composition of bone. This suggests that the fish producing these coprolites were carnivorous and that there is a lack of herbivores present, supporting Carpenter et al.’s (2015) faunal study findings. Bone fragments have been found in almost all samples, however specific species identification has thus far been impossible. This research provides both a foundation for further studies on coprolites and similar fossils and a deeper understanding of aquatic ecosystems as fish diversified further into fresh water in the Palaeozoic.

xiii

xiv

Introduction

The 15km of cliffs exposed along the UNESCO World Heritage Site at Joggins,

Nova Scotia (Fig. 1) represent one of the best and most complete examples of Late

Carboniferous ecosystems anywhere in the world (Grey and Finkel, 2011). These fossils, and the strata that hold them, have been studied for over 150 years by the likes of Sir

William Logan, Sir Charles Lyell, and William Dawson among many others (Lyell, 1843;

Logan, 1845; Dawson, 1868; Grey and Finkel, 2011). The fossil-rich beds exposed on these cliffs were first measured and recorded by Logan in 1845, with his stratified section

“measuring” up well even today (Logan, 1845). These cliffs and the fossils therein provide a glimpse into life over 300 million years ago, preserved in a nearly complete paleoenvironmental context (Grey and Finkel, 2011). The fossil forests, iconic to the site, have yielded some of the most important fossils ever found, specifically that of

Hylonomus lyelli, the earliest-known reptile in the fossil record (Dawson, 1868).

Figure 1. Location of the Joggins Fossil Cliffs (after Zaton et al., 2014).

Among the trees making up those large lycopod (club moss) forests are abundant plant and animal fossils exquisitely preserved in the tilted cliffs (Falcon-Lang et al., 2006).

This, alongside the nearly unbroken timeline of strata, makes these cliffs an ideal place to study the Carboniferous in the world (Lyell, 1843; Dawson, 1868; Falcon-Lang et al.,

2006).

Through much study over the past century and a half, the terrestrial ecosystems of

Carboniferous Joggins have become well understood. The aquatic deposits and their wealth of preserved organisms, however, have not been studied as intensely and thus these ecosystems are not as clearly defined (Falcon-Lang et al. 2006; Grey et al., 2011; Ó

Gogáin et al., 2016). The purpose of this thesis is to contribute to what is known about

Late Carboniferous aquatic ecosystems by investigating fish coprolites preserved in brackish water limestones of the Joggins Formation. Such information provides an unparalleled window into trophic relationships as fish expanded and diversified into fresh water ecospace (Ó Gogáin et al., 2016).

General geology and depositional environments

The Joggins Formation (Figs. 2, 3) is a 915.5-m-thick succession of strata that is early Langsettian in age (Fig. 2; Davies et al., 2005). It was accumulated over a period of ca. 1 million years in the centre of the Cumberland sub-basin, part of the late Paleozoic

Maritimes Basin complex of southeast Laurasia (Falcon-Lang et al., 2006). The Joggins

Formation is world-renowned for containing one of the best records of Carboniferous terrestrial and aquatic ecosystems (Lyell, 1843; Dawson, 1868; Falcon-Lang et al., 2006).

The formation is bounded by the Springhill Mines Formation to the south and the Little

River Falls Formation to the north (Fig. 2; Grey and Finkel, 2011). Rapid sedimentation in this area allowed for the in-situ deposition of many plant and animal fossils and enabled a nearly complete reconstruction of the environment and ecosystem at the time

(Falcon-Lang et al., 2006)

The Joggins Formation is divided into 14 transgressive and regressive cycles (Fig.

4), typically beginning with coal and associated limestone beds (Davies et al., 2005). Not all the cycles begin with limestone and coal layers and there are layers of both situated in the middle of some cycles (Grey et al., 2011). The cycles are then divided further into 3 major assemblages: the open water facies assemblage (OWFA); the poorly-drained flood plain assemblage (PDFA); and the well-drained facies assemblage (WDFA). The OWFA consists of thin coal layers, shelly coprolite-rich limestone, siltstone with abundant siderite nodules, and sandstone (Davies et al., 2005). This facies mosaic is interpreted to record accumulation in brackish embayments (Davies et al., 2005). Coastal plain sediments consist of coal, heavily rooted sandstone, and mudstone that collectively form the PDFA. Red sandstone and mudstone representing crevasse splays and channels constitute the WDFA. The limestones are ostracod and bivalve wackestones to packstones typically deposited above thin layers of coal.

Paleoecology and previous research

The focus of this study is the limestone beds which contain fish bones and scales, spiraled tube worms as well as the previously mentioned ostracods and bivalves (Grey et al., 2011; Zaton et al., 2014). The OW facies represent brackish coastal bays present during periods of late transgression and high-stand. Based on the analysis of organisms

present, the water column in coastal environments was not only brackish but also anoxic near the seabed (Falcon-Lang et al., 2006; Grey et al., 2011). The aquatic ecosystem was home to many species of fish, arthropods, molluscs, plants, amphibians and possibly soft bodied herbivores (Falcon-Lang et al., 2006; Carpenter et al., 2015). The maximum water depth in these coastal bays was probably a few metres to a few tens of metres (Davies et al., 2005; Grey et al., 2011). The lycopsid forests along the coast that characterize PDFA and WDFA facies assemblages are interpreted to have kept pace with sea level rise for

100 to 1000 years before being drowned by brackish water and accumulation of bituminous limestones of the OWFA that are the focus of this research (Falcon-Lang et al., 2006).

The most striking and apparent fossils preserved in the OWFA are the skeletal remains and coprolites of at least 13 species of fish that ranged from tens of centimetres to over three meteres in length (Carpenter et al., 2015). Species belonging to chondrichthyans, rhizodopsids, rhizodonts, and megalichthyids were active predators and thus, secondary and tertiary consumers (Carpenter et al., 2015). Haplolepids and

Gyracanthids fed on plankton and micro-invertebrates through sifting with their gill rakers (Falcon-Lang et al., 2006; Carpenter et al., 2015). Sagenodus were bottom dwellers with teeth designed to eat infaunal and epifaunal molluscs (Carpenter et al.,

2015). Sharks, such as Xenacanthus and Ctenacanthus, as well as the basal tetrapod

Baphetes were the top carnivores, feeding on smaller fish (Falcon-Lang et al., 2006; O

Gogáin et al., 2016). These fishes were well adapted for the brackish conditions that dominated the coastal zone (Fig. 5), including sharks which in modern settings require normal marine salinities of 35‰ (Carpenter et al., 2015).

Figure 2. The Joggins Fossil Cliffs UNESCO World Heritage Site, with the Joggins Formation in grey bounded by the Springhill Mines and Little River formations in green and burgundy respectively (after Grey and Finkel, 2011).

Figure 3. The Joggins Formation and associated formations imposed on the geological timescale with the Joggins Formation in the mid Langsettion (after Davies et al., 2005). 6

Figure 4. The Joggins Formation: facies associations and cyclicity. The open water facies assemblage (OWFA) is shown in dark grey at the base of 10 of the 14 cycles shown and correlates with the periods of highest relative sea level. The most likely sources for the coprolites, cycles 14, 13, and 9, are shown with red arrows (after Davies et al., 2005).

Coprolites belong to a group of trace fossils known as bromalites and are common in Devonian or younger strata. Bromalites are fossils from the digestive tracks of organisms and includes regurgitalites (ejected from the oral cavity), cololites (in situ intestinal contents), gastrolites (fossilized stomach contents), and true coprolites

(fossilized feces). Coprolites are the most common of all bromalites (Aldridge et al.

2006). Previous studies of coprolites examined undigested fossil material using hand sample morpholology, cathodoluminescence (Ó Gogáin et al., 2016), and scanning electron microscopy (Gong et al., 2010). The most common and often revealing analytical method is to use standard microscopic techniques such as transmitted light microscopy to understand trophic relationships (Baxendale, 1979; Richter and Baszio,

2001; Aldridge et al., 2006; Gong et al., 2010; Friedman, 2012; Owocki et al., 2012;

Rakocrinski et al., 2014; and Ó Gogáin et al., 2016). Only a few studies have focused on

Carboniferous coprolites (Gong et al., 2010 and Ó Gogáin et al., 2016). Coprolites from the Joggins Formation have not been previously examined and this offers a unique window into the food web of the OWFA. Fish coprolites are found abundantly throughout the limestones of this formation, but are far more common in the bivalve-poor beds. They range in size from under a millimetre to over 6 cm in length and are composed of calcium phosphate, suggesting that they all came from a carnivorous source

(Chin, 2002).

Objectives and implications

The purpose of this research is to shed light on an understudied ecosystem in one of the greatest fossil sites in the world. The objective is to determine what the fish species present in the OWFA at Joggins were eating and how that relates to the broader food chain. Through studying a virtually untapped window into Paleozoic ecology represented in fish coprolites, this research contributes new information about trophic relationships in brackish coastal environments at an important time in fish development, as they diversified further and further into the fresh water realm (Ó Gogáin et al., 2016). The results of this research, when viewed in context of what is already known about aquatic and terrestrial ecosystems of the Joggins Formation, provides the foundation for similar studies on the limestones and the ecosystem they represent. This research will also demonstrate different methods of analysis on coprolites, and similar fossils, and their usefulness based on the information they provide.

Materials and methods

Coprolite samples were acquired both from the permanent collection at the

Joggins Fossil Institute and ex-situ from the Joggins Formation. The 72 samples were divided into 6 groups based on size and shape and range from about 1-6cm. Each sample was photographed and catalogued. Thirty-one samples were made into polished thin sections by Pam Frail at Acadia University to be viewed primarily through a polarizing light microscope (Nikon Optiphot-POL microscope with a C-0.45x TV-Lens) for petrologic analysis. Due to the soft nature of the coprolites, however, only 25 of the thin sections were created successfully. Pictomicrographs were taken of every section,

documenting any undigested material that the respective coprolites contained. Samples

NSM008GF039.268, NSM017GF008.044, NSM017GF008.045, and NSM017GF008.048 were viewed under cathodoluminescence (CL), and gave important information on alteration and diagenesis (CL microscopy was performed on a Relion ELM–3R

Luminoscope mounted on a Nikon Labophot2–POL petrographic microscope). Samples

NSM008GF020.001 and NSM00839.188 were sent to PANalytical in Westborough

Massachusetts for computed tomography (CT) scanning using an Empyrean X-ray diffractometer. CT scanning produced three dimensional images of fossil fragments in coprolites that can be rotated in computer space. The remaining samples were then viewed under a hand sample microscope and further photographs were taken detailing any visible undigested material (hand samples were analysed using a Leica MZ75 microscope with a Nikon Coolpix 995 camera and a Nikon MDC lens). Thin sections of the limestones containing the 31 samples were made to help understand the OWFA and place the coprolites in their respective beds. In addition, 6 sections were made from 3 samples collected directly from limestone beds (cycles 14, 13, and 9) on the cliffs to identify the original locations of the coprolites not found in situ. Classification of the limestone was done using the Dunham Carbonate Classification (Dunham, 1962).

Identification of the bone fragments has unfortunately been unsuccessful despite help from researchers in this field, namely David Carpenter at the University of Southampton.

Because of this, identification of the species producing the coprolites will be made based on coprolite size and content. As such, specific species identification was not always possible and broad groups representing multiple species have been used.

Categorization of the Joggins coprolites

The 74 fish coprolites collected from the Joggins Formation differ greatly in size and shape. The average size of the coprolites is around 2-3cm in length and 6 categories based on size and morphology have been created: cigar/cylindrical shaped; cone shaped; small/equant; spiral; irregular; and massive (samples greater than 5cm in length). Cigar shaped coprolites range from 1-5cm and are cylindrical with a width less than half of their length (Figure 5A). The small/equant coprolites are 2-3cm long and tend to have a relatively equal length and width (Figure 5B). The cone shaped coprolites are wide at one end and come to a point at the other; they are almost all approximately the same size, 3-

5cm, and their shape is relatively consistent. Many of the cone shaped samples have grooves lengthwise on the exterior of the coprolite (Figure 5C). The spiral coprolites are typically very similar to the larger cylindrical coprolites, 3-4cm in length, and have a clear spiraled or rolled morphology (Figure 5D). The massive coprolites comprise a number of different morphologies, but are grouped as such based on volume and length as they likely represent fish from the top of the food chain. All samples are greater than

5cm in length and some measure as much as 10cm. Most of these coprolites have been flattened and are composed of a large main body with numerous pellets and smaller debris (Figure 5E). The final group comprise the irregular coprolites – these are coprolites that do not appear to fit into any of the aforementioned categories and are too small to fit in the massive group. They range greatly in size and shape, but are typically

1-3cm in length (Figure 5F). Overall, the most common coprolites are the small/equant and cylindrical shaped; the least common are the massive and spiral coprolites (Appendix

1).

Thin section and hand sample analysis

The majority of coprolites were found imbedded in mud-rich blue limestone; however, a few samples were found in bivalve-dominated limestone. Two samples were also preserved between layers of coal and one sample was found in sandstone. The coprolites, including very small (mm-scale) examples found in the limestone thin sections, are composed of high calcium phosphate similar to francolite, the predominant composition of bone. They clearly have undigested material preserved within them, as shown through microscopic, hand sample, and CT analysis. The material appears to be almost exclusively bone fragments, present in all coprolite morphologies. In total, 19 of the 25 thin sections hold fragments that could be or are bone. Thirteen of the 44 hand samples show clear bones and almost all have grains that could be bone. The concentration of bone fragments varies between groups but the cone shaped coprolites contain the most and the small/equant coprolites contain the least. The bones range in size from less than a millimetre in length to a few centimetres, although the majority are a few millimetres. The bones within the coprolites are most easily viewed in thin section

(Figure 6). Some show clear bone shapes, while others are much more fragmented.

In sample NSM008GF031.506, bone fragments are present within the coprolite as small pellets of crushed bone (Figure 6C). Also visible in many of the thin sections are coated grains of sand or mud that entered the digestive track and did not dissolve. Later, through diagenesis, these grains became coated in phosphate and grew larger; they are seen as are spirals or concentric circles in the thin sections (Figure 6E).

A B

C D

E F

Figure 5: Hand sample photographs of the six coprolite morphologies; A: Cylindrical (NSM017GF039.500); B: Small/Equant (NSM017GF008.046); C: Cone shaped (NSM017GF008.049); D: Spiral (NSM017GF008.054); E: Massive (NSM017GF008.028); F: Irregular (NSM008GF039.500). Scale bars = 2cm

Unfortunately, no bones could be identified to species as they were typically singular disarticulated bones or fragments. None of the coprolites appear to have biological fragments other than bones, and the occasional scale (Figure 7E), present. It is possible, however, that some of the smaller fragments are from exoskeletons of arthropods. There are also numerous grains of framboidal pyrite present in the coprolites

(Figure 6F). Many the samples show excellent preservation of the original fecal matter and undigested material. For instance, sample NSM011GF040.009W contains numerous bone fragments of various sizes (Figure 6 A/B) and also has clearly preserved internal morphology showing a flowing, perhaps even layered, texture (Figure 6A).

Among the hand sample coprolites, NSM008GF031.446 contains some of the best-preserved bones (Figure 7F). It is 6 cm long and has 10 very distinct bones visible to the naked eye that are longer than many of the smaller coprolites (1cm+). Additionally, samples NSM017GF008.015 and NSM017GF008.012 are significant as they are both preserved within coal (Figure 7A/B/C). Alongside implications of habitat, there are clear, well-preserved bones within these samples (Figure 7C). Samples NSM017GF008.014 and NSM008GF039.500 also show very clear bones and scales in hand sample, although the latter has little phosphate remaining (Figure 7D/E/F).

A B

C D

E F

Figure 6. Pictomicrographs of a selection of coprolites in thin section; A: A small, mostly whole, bone in NSM011GF040.009W; B: A larger, more fragmented, bone in NSM011GF040.009W; C: Pellets made of crushed bone fragments from NSM008GF031.506; D: A hole in the coprolite filled in by calcite in sample NSM017GF008.043; E: A spiralled coated grain from NSM017GF008.047; F:

Framboidal pyrite from NSM008GF039.508 taken using reflected light.

A B

C D

E F

Figure 7: Hand sample photographs of coprolites of interest; A/B: Coprolites with bones preserved within coal, (NSM017GF008.015 and NSM017GF008.012). Scale bar = 2cm; C: A well preserved bone from NSM017GF008.012. Scale bar = 3mm; D: A small coprolite with both bones and a scale (NSM017GF008.014). Scale bar = 2cm;

E: Fish scale from the coprolite in D. Scale bar = 2mm; F: A large coprolite with abundant bones visible to the naked eye (NSM008GF031.446). Scale bar = 1cm.

Cathodoluminescence and computed tomography

Cathodoluminescence (CL) analysis indicated that there was very little alteration of calcite in any of the samples and that many of the samples instead experienced recrystallization of the phosphate (Figure 8A). It also became apparent through CL that many of the samples have holes in them where calcite has grown into. These are represented as large calcite blotches in thin section (Figure 6D). The CL images show a distinct yellow-green colour from the coprolites which would have been caused by Mn2+

(Figure 8A/B/C). There is some calcite, showing as bright orange on the CL signature, present in these coprolites and evidence of it encroaching in on some of the samples.

This, however, appears not to have had a significant effect on the coprolites or their contents. An important detail shown by the CL is that there is little to no growth zonation, suggesting a stable environment (Hiatt and Pufahl, 2014).

Computed tomography on samples NSM008GF020.001 (Figure 9) and

NSM008GF039.188 proved to be only somewhat useful, and although bones are visible

(Figure 9B/D), they are not as clear or impressive as the polished thin sections.

Provenance

The coprolite samples used for this study were all collected ex-situ from the

Joggins Formation and, as such, their stratigraphic origin is difficult to determine. Three limestone samples were taken from possible source beds (cycle 14, 13 and 9; Figure 4) to help determine the provenance of the coprolites. Based on external examination and thin section analysis of the limestones, the coprolites were most likely encased in the blue ostracod wackestone from the 14th cycle (Figure 4). The other two cycles are the

bivalve-rich limestone and are difficult to distinguish from each other. That aside, three of the thin sections came from the bivalve packstone as their limestone source rock was almost identical to cycles 13 and 9 when viewed in thin section and the rest belong to the

14th cycle.

Discussion

Based on the thin section analysis of the limestone strata, it is clear that the vast majority of the collected coprolites came from the 14th cycle (Figure 4) as they are almost identical as examined in thin section and in hand sample. The longshore drift presently at this site means that it is incredibly unlikely that any of the samples came from genetically similar, older limestones, wackestone or packstone (Figure 4). The source strata for the coprolite samples obtained from the permanent collection at the Fossil Institute, however, can not be firmly identified as the stratigraphy was not noted by the collectors. That being said, the thin sections made of these samples appear to be genetically similar to cycle 14.

The two major types of limestone in the upper Joggins Formation, the ostracod wackestone and the bivalve packstone, differ greatly in terms of coprolite abundance. The ostracod wackestone tends to have far more coprolites than its bivalve-dominated counterpart. This could be for a number of reasons. Firstly, this may be the result of different preservational environments. The blue limestone has little to no bivalves and is largely composed of ostracods and mud whereas the other limestones are dominated by bivalves with little mud. This could represent a change in the benthic ecosystem with larger numbers of animals living and feeding on the seafloor. These animals could have used fish feces as food (Wotton and Malmqvist, 2001) and thus fewer coprolites would

have been preserved. Secondly, different environments could also influence the preservation potential of the coprolites themselves. The mud-dominated environment may have simply been better at preserving the coprolites than the bivalve-dominated one.

Thirdly, the two environments could have preserved coprolites equally, but the ficality of the more bivalve-dominated limestones meant that any coprolites exposed to the elements were quickly exposed and destroyed. Coprolites held within the blue limestone, on the other hand, are better indurated and would stand up to the elements much better allowing for greater protection of any coprolites within.

While the great majority of coprolites were found within limestone, the samples encased in coal (NSM017GF008.015, NSM017GF008.012) and sandstone

(NSM017GF008.054) are perhaps most significant in terms of host rock. Coprolites have been found in coal in a terrestrial setting (Baxendale, 1979) but fish coprolites have never before been noted in layers of coal. It appears that the peat bogs that produced the coal were perhaps the best preservational environment for undigested material as the coal samples hold some of the finest preserved bones. Both of these rock types represent shallower waters than the limestones and would have been present further inland. This is noteworthy as it shows that the fish were clearly swimming into shallower, and increasingly fresher water.

A B

Figure 8. Pictomicrographs of coprolite thin sections under cathodoluminescence; A: Recrystallized phosphate showing no growth zonation. (NSM008GF039.268); B: Original phosphate with a bone fragment and some alteration halos.

NSM017GF008.048; C: Phosphate in yellow being altered to calcite in orange. (NSM017GF008.045).

A B

C D

Figure 9. Coprolite images made with computed tomography A: The 3D model created by the CT scanning of NSM008GF.020.001; B: The top cut made in the first cross section (bones visible); C: The cut to the right on the second cross section; D: The front cut on the second cross section (bones visible).

The fish species present at Joggins were suited for a fairly broad range of salinities (Carpenter et al., 2015) and it is likely that some of these species began the transition into predominantly fresh water fish. As sea level dropped, it is also possible that some of these fish swam further inland to rivers, and eventually lakes, instead of back out to sea; through multiple generations they adapted to living entirely in freshwater. This is significant as these would be new species being introduced to an ecosystem that was recovering from a major extinction (Cohen, 2003). Fish species like those present at Joggins would be adapting to breathe new life into fresh water ecosystems, filling in the gaps left by species who had died out.

Classifying the coprolites according to shape and size enables an interpretation of a food web in the open water facies where, in this case, there appears to be four trophic levels represented. This is based primarily on coprolite size and bone content. First, the small or equant coprolites have the smallest concentration of bone fragments as well as the most coated grains. This suggests that they represent smaller fish species living near the seafloor, where there would be more abundant grains of sediment, and would likely be feeding mostly on soft bodied organisms, thus few to no bones are present.

The cylindrical, spiral, and most of the irregular coprolites fit into another level of the food web. These larger coprolites have more bone fragments than the smaller/equant ones. Many of the mid-range fish they represent are most likely sharks and, as they have spiraled digestive tracts, they tend to leave spiraled dung (Ó Gogáin et al., 2016). As such, it is likely that the spiral coprolites come from similar species to the cylindrical ones and that they have experienced a higher degree of preservation allowing for the spiraled morphology to be better preserved. These coprolites would likely be from

coelacanths and smaller sharks, such as Ctenacanthus and Xenacanthus, as well as other fish of a similar size.

The third trophic level is represented by the cone-shaped coprolites. They have the largest number of bones, indeed every cone-shaped thin section and almost every hand sample contains bone fragments. They also tend to be larger than the cylindrical coprolites in length and width. The cone-shaped coprolites are most likely from

Xenacanth species, such as Orthacanthus, based on their shape, size, and appearance (Ó

Gogáin et al., 2016). Orthacanthus could reach sizes of 3m (Ó Gogáin et al., 2016) and, like Xenacanthus, it was probably well suited to brackish or freshwater environments such as those at Joggins. These animals probably ate mostly fish from the first two trophic levels, perhaps even fish as large as coelacanths and smaller sharks. As these coprolites came from larger sharks it is likely that the cone shape is representative of a slightly different spiral than the spiral and cylindrical coprolites. This spiraled morphology is perhaps shown best in the thin section NSM011GF040.009W as the phosphate appears to be layered or flowed.

The final trophic level of the food web is represented by the massive coprolites.

They typically have much fewer bones, with one significant exception. These are likely from the largest fish present at Joggins, the Rhizodonts, such as Strepsodus, and possibly some Rhyzodus species. The Rhizodont species are the largest aquatic animals found thus far at Joggins, reaching sizes of 6m in length (Mansky and Lucas, 2013). Their large size would put these fish at the top of the food web and they would have fed largely on fish from the previous two trophic levels. The lack of bone fragments in most of the massive samples may be due to a diet of cartilaginous fish, especially the larger sharks. It is also

possible that these fish did not produce solid remains like smaller fish and the bones and other undigested material simply fell away from the rest of the excrement. This type of defecation is seen in some modern day large sharks as they produce a yellow cloud of fecal matter. One sample (NSM008GF031.446) out of ten does have numerous bones held within it, however, and it is unclear why. It is possible that NSM008GF031.446 was produced by an individual that strayed out of its normal feeding habits and ate either young fish of its own or cousin species. The bones could also be from an amphibian or reptile instead, since they could have easily been eaten by one of the large Rhizodonts if they were in deep enough water. Unfortunately, specific species identification cannot be made for the bulk of the samples as, aside from the Orthacanthus coprolites, there are no definitive clues to correlate species to morphologies. Because of this, the designations between coprolites and the species that produced them are hypothetical. Ó Gogáin et al.

(2016) were able to tie the coprolite in their study, concerning fossils of very similar age and location, to the species producing the fecal matter, Orthacanthus. They were also able to identify some of the undigested remains, as there was a juvenile xenacanthid tooth. Unfortunately, no easily identifiable material was found in the coprolites from the

Joggins Formation.

Thin sections reveal that some of the cone-shaped coprolites contain what appear to be pellets made of crushed bone fragments. This suggests that the bones were crushed either by the teeth of the predator or through the digestive process. Fish such as

Ctenoptychius had teeth that were capable of crushing, but they have been interpreted as feeding on bivalves (Carpenter et al. 2015). All coprolite morphologies but the smallest have small bone fragments preserved within them which suggests that most of the fish

crushed at least some of the bones while feeding. The coated grains found within many of the coprolites also signify important details of the fishes’ feeding habits. Predominantly found in the smaller coprolites, the coated grains of sand or mud suggest feeding that occurred close to the sea floor. This, alongside the lack of fragments of any kind in many of the smaller coprolites, further supports the hypothesis posited by Carpenter et al.

(2015) that there were abundant soft-bodied organisms at the base of the food web.

Carpenter et al. (2015) noted in their study that, of all of the fish fossils that have been found, there were very few herbivores and many carnivores and that it was likely that this was filled by organisms that do not fossilize well.

The variety of morphology and contents present in both the massive and the irregular coprolites could represent a few different things. Both groups are fairly uncommon and this could mean that they come from a variety of fairly rare species. The irregular coprolites could also represent fish species that are common and typically represented by one of the four other groups, but have been forced to feed on organisms outside of their usual diet. The massive coprolites could also be from a narrow group of fish, but a group that, due to their size, have a varied diet and will eat anything big enough that comes within range.

The methods for analysis of the coprolites were varied and, as such, they differed in the type of information they provided and in the quality of that information. By far, the best method used in this study for discerning the diet of these animals was thin section analysis. Although some of the thin sections were unsuccessful because the phosphate material was difficult to work with, the thin sections overall gave a clear picture of the interior of the coprolites, with obvious bone fragments as well as the internal structure of

some samples. Cathodoluminescence (CL) proved to be a useful tool as it indicated the level of alteration present in the coprolites and the environment that they were deposited in. CL analysis of the four thin sections in particular revealed some important details.

Firstly, the coprolites and the phosphate that comprises them are rich in Mn2+, indicated by the yellow green luminescence produced in CL (Kempe and Gotze, 2002). Mn2+ suggests an anoxic environment of deposition, which aligns well with the brackish setting represented by the OWFA of Joggins. Secondly, the sample (NSM008GF039.268) that has largely been recrystallized shows essentially no growth zonation. This is significant as it indicates that the conditions experienced by these coprolites and their host rocks were uniform after burial. Lastly, the bone fragment seen in the unaltered thin section

(Figure 8.) appears as the same colour and looks almost identical to the rest of the coprolite, lending further support that the composition of the coprolite is similar to the composition of bone. This also indicates that dissolving the phosphatic material comprising the coprolite may also dissolve the bones within suggesting this method of examination to potentially be very difficult.

The usefulness of hand sample analysis was dependent on the coprolite. Some provided significant information and others were uninformative. This is due to the fact that most of the coprolites do not have well preserved bones on their surface. Typically, they have grains that could be bones, but could also be coated grains, or something created as the coprolite underwent diagenesis. Such features are not distinguishable from hand samples alone.

CT scanning, the final method of examining coprolites, showed bones and produced a 3D model of the coprolite but it did not show the bones with the quality or

clarity of those shown in thin section. This method is not as useful and far more expensive than the other techniques, such as thin section, CT, and hand sample analysis, in studying the coprolites and did not reveal anything that was not apparent from other methods.

As a whole, this study provides background and a foundation for future studies on coprolites and fish species at Joggins and for similar fossils around the world. Based on the findings of this research, it is pertinent to take samples in situ when possible and to focus on thin section and hand sample analysis as opposed to CT scanning. For Joggins, making an effort to find more samples from both underrepresented strata and morphologies to gain further understanding of this ecosystem is imperative. Perhaps a more exhaustive study on the bones found inside the coprolites and the organisms that they came from would lead to species identification, something that was impossible in this study. Additionally, research should be done to determine if bones or any other material could be separated either by dissolving the coprolite or by physically picking out the bones. Further research on these and other coprolites involving methods such as scanning electron microscopy (SEM) and 3D photo modeling may be useful and could lead to new information about food webs during this time.

Conclusions

This study of over 70 coprolite samples from the Joggins Formation allowed for a unique view into the Late Carboniferous aquatic ecosystem. They provided invaluable information on species interactions in an understudied environment, the brackish bays of

Joggins. This research showed that:

1) Coprolite samples collected ex situ have been correlated with the 14th, 13th and 9th

cycle (Figure 4). They are far more abundant in the blue ostracod wackestone of

cycle 14 than the bivalve packstones of 13 and 9. This is likely due to either

environmental factors at the time of deposition or the relative induration of the

limestones.

2) Samples preserved in coal and sandstone provide further evidence that fish

species at Joggins were swimming upstream and into fresher water. This suggests

that the fish preserved are representative of species being introduced into a

relatively new environment, one that was still recovering from a major extinction

at the end of the Devonian. Although this extinction affected marine life more

severely, it had a significant effect on freshwater ecosystems as well. This allowed

for early ray finned fishes and cartilaginous fishes, like those preserved at

Joggins, to adapt to this relatively new freshwater niche in the Late Carboniferous

and replace the lobe-finned fishes that came before them (Cohen, 2003).

3) Four trophic levels of the food web were present in the brackish bays of Joggins.

Stretching across many species, from the smallest fish feeding on soft bodied

organisms to the giant rhizodonts preying on whatever they could catch, this

insight into species interactions provides a greater understanding of a relatively

understudied ecosystem.

4) Thin section analysis, coupled with cathodoluminescence and hand sample

analysis, proved to be useful methods for investigating the Joggins coprolites.

Computed tomography was not nearly as valuable because it did not reveal

anything that other methods could not and the models were not detailed. Other

techniques for studying coprolites at Joggins, including dissolving the coprolite

matrix in order to isolate bones, should be explored in future studies.

List of References

Aldridge, R., Gabbott, S., Siveter, L., and Theron, J. 2006. Bromalites from the Soom Shale Lagerstatte (Upper Ordovician) of South Africa: Palaeoecological and palaeobiological implications. Palaeontology, 49: 857–871.

Baszio, S., and Richter, G. 2001. Traces of a limnic food web in the Eocene Lake Messel — a preliminary report based on fish coprolite analyses. Palaeogeography, Palaeoclimatology, Palaeoecology, 116: 345- 368.

Baxendale, W. 1979. Plant-bearing coprolites from North American Pennsylvanian coal balls. Palaeontology, 22: 537-548

Carpenter, D., Falcon-Lang, H., Benton, M., and Grey, M. 2015. Early Pennsylvanian (Langsettian) fish assemblages from the Joggins Formation, Canada, and their implications for palaeoecology and palaeogeography. Palaeontology, 58: 1-30

Chin, K. 2002. Analysis of coprolites produced by carnivorous vertebrates. Paleontological Society Papers, Museum of Natural History/Department of Geological Sciences, University of Colorado at Boulder. 8: 43-50.

Cohen, A., 2003. Paleolimnology: The History and Evolution of Lake Systems. Oxford University Press, Oxford, U.K, 528p.

Davies, S., Calder, J., Gibling, M., Rygel, M. and Skilliter, D. 2005. The Pennsylvanian Joggins Formation of Nova Scotia, sedimentological log and stratigraphic framework of the historic fossil cliffs. Atlantic Geology, 41: 115-142.

Dawson, J.W. 1868. Acadian geology: the geological structure, organic remains, and mineral resources of Nova Scotia, New Brunswick, and Prince Edward Island. Second edition. Mac-Millan and Co., London, 694 p.

Falcon-Lang, H., Brenton, M., Braddy, S. and Davies S. 2006. The Pennsylvanian tropical biome reconstructed from the Joggins Formation of Nova Scotia, Canada. Journal of the Geological Society, London, 165: 561-576.

Friedman, V. 2012. Vertebrate coprolites from the Lower Eagle Ford Group of north central Texas and their paleoecological significance. New Mexico Museum of Natural History and Science. 57: 221-228.

Grey, M. and Finkel, Z. 2011. The Joggins Fossil Cliffs UNESCO World Heritage site: a review of recent research. Atlantic Geology, 47: 185–200.

Grey, M., Pufahl, P., and Aziz, A., 2011. Using multiple environmental proxies to determine degree of marine influence and paleogeographical position of the Joggins Fossil Cliffs UNESCO World Heritage Site. Palaios, 26: 256–263.

Gong, Y., Wu, Y. and Zhang, L. 2010. Carboniferous coprolites from Qinhuangdao of North China. Science China, Earth Sciences, 53: 213-219.

Kempe, U., and Gotze, J., 2002. Cathodoluminescence (CL) behaviour and crystal chemistry of apatite from rare-metal deposits. Mineralogical Magazine, 66: 151- 172.

Logan, W.E. 1845. A section of the Nova Scotia coal measures as developed at Joggins on the Bay of Fundy, in descending order, from the neighbourhood of the west Ragged Reef to Minudie, reduced vertical thickness, Appendix W: Geological Survey. Journals of the Legislative Assembly of the Province of Canada, 4: 28–45

Lyell, C. 1843. On the upright fossil-trees found at different levels in the coal strata of Cumberland, Nova Scotia. Proceedings of the Geological Society, London, 4: 176–178.

Mansky, C., and Lucas, S. 2013. Romer's Gap revisited: Continental assemblages and ichno-assemblages from the basal Carboniferous of Blue Beach, Nova Scotia, Canada. New Mexico Museum of Natural History and Science. 60: 244-273.

Ó Gogáin, A., Falcon-Lang, H., Carpenter, D., Miller, R., Benton, M., Pufahl, P., Ruta, M., Davies, T., Hinds, S. and Stimson, M. 2016. Fish and tetrapod communities across a marine to brackish salinity gradient in the Pennsylvanian (early Moscovian) Minto Formation of New Brunswick, Canada, and their palaeoecological and palaeogeographical implications. Palaeontology, 59: 689– 724

Owocki, K., Niedzwiendzki, G., Sennikov, A., Golubev, V., Janiszewska, K., and Sulej, T. 2012. Upper Permian vertebrate coprolites from Vyazniki and Gorokhovets, Vyatkian Regional Stage, Russian Platform. Palaios, 27: 867-877.

Wotton, R., and Malmqvist, B. 2001. Feces in aquatic ecosystems. BioScience, 51: 537- 544

Zaton, M., Grey, M. and Vinn, O. 2014. Microconchid tubeworms (Class Tentaculita) from the Joggins Formation (Pennsylvanian), Nova Scotia, Canada. Canadian Journal of Earth Sciences, 51: 669-676

Appendices

Appendix 1.

Type Coprolites Coprolite size Undigested material Cigar/Cylinder NSM008GF023.007w, Typically 1-3 Bones and coated NSM008GF039.482, cm long and grains NSM008GF039.508, 0.5-1cm wide The bones are NSM008GF039.268, typically NSM017GF008.001, in fragments rather NSM017GF008.044, than NSM017GF008.027 whole, also appear to NSM017GF008.002A+B, be NSM017GF008.007, fairly small NSM017GF008.004, NSM017GF008.055, NSM017GF008.057, NSM017GF008.005, NSM017GF008.008, NSM017GF008.006, NSM017GF008.056

Cone NSM008GF031.506, 3-5cm long All but one contained NSM008GF041.052, 1-2cm wide bones, NSM008GF040.009, some of which were NSM008GF031.288, well NSM017GF008.021, preserved and whole, NSM017GF008.017, while NSM017GF008.019 others were NSM017GF008.049, preserved as NSM017GF008.018, pellets of bone NSM017GF008.020, fragments NSM017GF008.022, NSM017GF008.023, NSM017GF008.051 Small/Equant NSM017GF008.038, 1-2cm Essentially no bones NSM008GF039.345, long/wide Abundant coated NSM008GF039.524, grains NSM017GF008.042. NSM017GF008.054,

NSM008GF020.001, NSM008GF039.188, NSM017GF008.010, NSM017GF008.045 NSM017GF008.046, NSM017GF008.047, NSM017GF008.011, NSM017GF008.014, NSM017GF008.015, NSM017GF008.012, NSM017GF008.013, NSM017GF008.016, NSM017GF008.003, NSM017GF008.053, NSM017GF008.009 Large NSM017GF008.040, 5+ cm long Very few bones, no NSM008GF031.446, 2+ cm wide real coated grains. NSM017GF008.028, NSM017GF008.029, NSM017GF008.043, NSM017GF008.033, NSM017GF008.034, NSM017GF008.031, NSM017GF008.030, NSM017GF008.052, NSM017GF008.035 Irregular NSM017GF008.039, <1cm to 3 cm Various, some NSM008GF039.386B, long/wide bones grains etc. NSM017GF008.041, NSM008GF039.500, NSM008GF031.348, NSM017GF008.026, NSM017GF008.025, NSM017GF008.037, NSM017GF008.024, Spiral NSM008GF039.479, <1cm wide, Most have clear NSM017GF008.048, typically over bones, but less than NSM017GF008.036, 2m long cones NSM017GF008.050