Diet Analysis of Stranded Bottlenose Dolphins (<I>Tursiops Truncatus</I

Home , Atlantic croaker, Longfin inshore squid

Old Dominion University ODU Digital Commons

Biological Sciences Theses & Dissertations Biological Sciences

Spring 2020

Diet Analysis of Stranded Bottlenose Dolphins (Tursiops truncatus) in Virginia

Kristen Marie Volker Old Dominion University, [email protected]

Follow this and additional works at: https://digitalcommons.odu.edu/biology_etds

Part of the Biology Commons, Ecology and Evolutionary Biology Commons, and the Marine Biology Commons

Recommended Citation Volker, Kristen M.. "Diet Analysis of Stranded Bottlenose Dolphins (Tursiops truncatus) in Virginia" (2020). Master of Science (MS), Thesis, Biological Sciences, Old Dominion University, DOI: 10.25777/6bas-rj82 https://digitalcommons.odu.edu/biology_etds/113

This Thesis is brought to you for free and open access by the Biological Sciences at ODU Digital Commons. It has been accepted for inclusion in Biological Sciences Theses & Dissertations by an authorized administrator of ODU Digital Commons. For more information, please contact [email protected]. DIET ANALYSIS OF STRANDED BOTTLENOSE DOLPHINS

(TURSIOPS TRUNCATUS) IN VIRGINIA

Kristen Marie Volker B.S. December 2008, University of New England

A Thesis Submitted to the Faculty of Old Dominion University in Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCE

BIOLOGY

OLD DOMINION UNIVERSITY MAY 2020

Approved by:

Ian Bartol (Director)

Holly Gaff (Member)

Kent Carpenter (Member)

Damon Gannon (Member) ABSTRACT

DIET ANALYSIS OF STRANDED BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) IN VIRGINIA

Kristen Marie Volker Old Dominion University, 2020 Director: Dr. Ian Bartol

This study describes the diet of bottlenose dolphins (Tursiops truncatus) stranded in Virginia via stomach content analysis and considers factors such as proportion of numerical abundance and reconstructed mass, frequency of occurrence, average reconstructed prey size, prey diversity and quantity, and otolith degradation code. Fish size was estimated via regression equations established from local fish collected during the study that derive wet weight directly from otolith length or width. Squid size is estimated from previously published equations. Soniferous fishes dominated the diet, especially Atlantic croaker, spot, and seatrout spp., adding evidence to the theory that bottlenose dolphins passively listen for their prey. Noise pollution should be an important consideration for the conservation of this species. The diet was influenced by ontogenetic and seasonal changes. Prey that are relatively difficult to capture or not available to calves due to habitat, such as striped bass or longfin inshore squid, were absent in the diet of that age class. However, easy to capture prey, such as spot, became progressively less important with age. Seasonal variation in the diet was observed, with differences likely being due to changes in prey availability (due to seasonal migrations) and detectability (due to increased vocalizing during spawning) rather than changes in prey preference. Lastly, the stomachs of dolphins with external evidence of an entangling interaction contained a significantly higher proportion of recently ingested fish (using otolith code as a proxy) than those that did not. The presence of recently ingested prey is commonly referenced as an indicator of peracute underwater entrapment; however, this is the first study to quantitatively establish a significant relationship. The data presented in this study promise to enhance policy and management by establishing baseline information on natural history as well as a quantitatively assessed indicator for peracute underwater entrapment that can be referenced to study local shifts in natural history or incidence of bycatch.

iii

This thesis is dedicated to the researchers of “stinky science” iv

ACKNOWLEDGMENTS

This thesis is due to the love, support, and dirty hands of many people.

My mom has always been my biggest cheerleader, but throughout this entire project, she has gone above and beyond to share her enthusiasm with me. From answering my phone calls that ranged from panic to excitement, to sitting there next to me picking through stinky stomach contents, she has been my stabilizing platform, my #1 fan, my rock.

Although the mention of stomach contents usually made my sisters’ noses crinkle in disgust, I know that Laura and Rachel, though they really couldn’t believe THAT is what I was working on for my project, really did believe in me. They provided me the comic relief I needed at the end of a stressful day, my distractions from work to come back to reality, as well as the encouraging words I needed to get though my challenges. I don’t know what I’d do without you two, Laurdey Loo and Lil’ Ray!

My friends have provided support in so many different forms! From assisting with different aspects of my project, to hysterical cat memes, to a drink or two at Tapped, I could not have completed this without you. Although at the end of my project I was working more on “tummies” than being the “Queen of the Dead” at work, my co-workers/friends never complained. I truly appreciate you guys understanding the work I needed to put into the project and helping me accomplish it.

It was a pleasure to work with my academic advisor, Dr. Ian Bartol. His patience and understanding during this adventure to complete my Master’s while maintaining a full-time job, then moving to another state and another full-time job was priceless. He was always encouraging and supportive while demanding excellent work. I’ve also enjoyed learning from Drs. Damon Gannon, Holly Gaff, and Kent Carpenter who served on my committee. Their insights and perspective improved the project and thesis in various aspects, from project design to data presentation. Dr. Gannon also conducted the stomach contents analysis training at a workshop that was part of this project.

The project would not have been possible without funding from the John H. Prescott Marine Mammal Grant Program (award number NA11NMF4390098) nor without the support of the Virginia Aquarium & Marine Science Center Foundation and the dedication of the Stranding Response Team staff, volunteers and interns, past and present. v

Mark Swingle, Sue Barco, Sarah Mallette, and Maggie Lynott envisioned this project and secured the funding through the Prescott Grant Program. This project, along with a previously completed life history project, has long been a goal for Mark and Sue and I’m glad to have been a part of it.

Additionally, there were many individuals and organizations that assisted with this project. Dr. Ann Pabst, Bill McLellan, and Ryan McAlarney graciously hosted and were integral in the development and success of the training workshop (lovingly dubbed “Stomacheddon”). Ryan additionally assisted in squid beak identification and offered advice in diet study logistics and analysis. Jeffrey Thompson, Maureen Fender, Katie Glanton, Alex Balke, and Anthony Bosnengo (Virginia Aquarium Acquisitions and Quarantine Department), Chad Boyce and Evan Shearer (Virginia Department of Game and Inland Fisheries), Rich Hancock (Virginia Marine Resources Commission), Trish Bargo (East Coast Observers, Inc.), and Mike Mizell and Alexis Rabon (Virginia Aquarium Education Department) collected reference specimens for the project. Hank Liao and Jessica Gilmore (Old Dominion University Center for Quantitative Fisheries Ecology) provided additional otoliths and associated data for use in building regression equations. Many Virginia Aquarium Stranding Response volunteers, interns, staff, and contractors contributed over 1,300 hours to assist with sorting stomach contents, measuring otoliths, and organizing the reference collection, especially Shannon Davis, Sarah Rose, Kathy O’Hara, Patti Volker, Karen Moyer, Alison Morin, Melissa Chromik, Bonnie Bailey, Bill Dieffenbach, Peg Goodman, Carrie Rowlands, Sarah DeVed, Ryan Ponseti, Ashley Bunch, Marcy Fundalewicz, Lauren Talley, Salma Abdel-Raheem and many more. vi

TABLE OF CONTENTS

Page

LIST OF TABLES ...... vii

LIST OF FIGURES ...... viii

CHAPTER 1. INTRODUCTION ...... 1 2. LOCAL PREY SIZE ESTIMATION ...... 4 INTRODUCTION ...... 4 MATERIALS AND METHODS ...... 5 RESULTS ...... 7 DISCUSSION ...... 17

3. DEMOGRAPHIC PATTERNS OF DIET ...... 20 INTRODUCTION ...... 20 MATERIALS AND METHODS ...... 22 RESULTS ...... 25 DISCUSSION ...... 36

4. ENTANGLED DOLPHINS ...... 45 INTRODUCTION ...... 45 MATERIALS AND METHODS ...... 47 RESULTS ...... 49 DISCUSSION ...... 50 5. CONCLUSION ...... 53

REFERENCES...... 56

VITA ...... 65

vii

LIST OF TABLES

Table Page

1. Regression equations relating otolith length (OL, mm) to total length (TL, mm) of fish taxa important in bottlenose dolphin diet in Virginia ...... 7

2. Regression equations relating otolith width (OW, mm) to total length (TL, mm) of fish taxa important in bottlenose dolphin diet in Virginia ...... 8

3. Regression equations relating otolith length (OL, mm) to wet weight (W, g) of fish taxa important in bottlenose dolphin diet in Virginia ...... 8

4. Regression equations relating otolith width (OW, mm) to wet weight (W, g) of fish taxa important in bottlenose dolphin diet in Virginia ...... 9

5. Proportion of numerical abundance (%NA), frequency of occurrence (%FO), and proportion of reconstructed mass (%RM) of prey in the stomach contents of bottlenose dolphins stranded in Virginia ...... 26

6. Results from variables tested to compare the diet of bottlenose dolphins stranded in Virginia between sexes, age groups (calf, juvenile, adult), and seasons (spring, summer, fall) ...... 28

7. Proportion of numerical abundance (%NA), frequency of occurrence (%FO), and proportion of reconstructed mass (%RM) of prey in female and male bottlenose dolphins stranded in Virginia ...... 29

8. A) Proportion of numerical abundance (%NA), B) frequency of occurrence (%FO), and C) proportion of reconstructed mass (%RM) of prey in calf (n=20), juvenile (n=128), and adult (n=52) bottlenose dolphins stranded in Virginia ...... 31

9. Reconstructed mean total length (A) and wet weight (B) of common prey in the stomach contents of calf (n=19), juvenile (n=126), and adult (n=50) bottlenose dolphins stranded in Virginia...... 33

10. A) Proportion of numerical abundance (%NA), B) proportion of reconstructed mass (%RM), and C) frequency of occurrence (%FO) of prey of bottlenose dolphins stranded in Virginia in spring (n=59), summer (n=75), and fall (n=55) ...... 35

11. Results from variables tested to compare the diet of bottlenose dolphins stranded in Virginia between entanglement categories (yes, no, and could not be determined) ...... 49

12. Proportion of numerical abundance (%NA) of coded otoliths in bottlenose dolphins stranded in Virginia with (Yes), without (No), or could not be determined for (CBD) signs of entanglement...... 50

viii

LIST OF FIGURES

Figure Page

1. Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for anchovy spp. (Anchoa spp.) ...... 10

2. Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for Atlantic croaker (Micropogonias undulatus) ...... 11

3. Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for Atlantic menhaden (Brevoortia tyrannus)...... 12

4. Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for seatrout spp. (Cynoscion spp.)...... 13

5. Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for silver perch (Bairdiella chrysoura)...... 14

6. Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for spot (Leiostomus xanthurus)...... 15

7. Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for spotted hake (Urophycis regia) ...... 16

8. Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for striped bass (Morone saxatilis) ...... 17

9. Proportion of numerical abundance of prey in stomach contents of calf, juvenile, and adult bottlenose dolphins stranded in Virginia ...... 30

10. Proportion of numerical abundance of prey in stomach contents of bottlenose dolphins stranded in Virginia in fall, spring, and summer...... 34

CHAPTER 1

INTRODUCTION

The bottlenose dolphin (Tursiops truncatus) is found in temperate and tropical coastal and offshore waters throughout the world. It is a well-studied predator in the western north Atlantic belonging to two ecotypes, “coastal” and “offshore” that can be distinguished via distribution, morphological, and physiological differences (Duffield, Ridgway, & Cornell, 1983; Hayes, Josephson, Maze-Foley, & Rosel, 2019; Mead & Potter, 1990, 1995; Winn, 1982). The coastal ecotype along the eastern U.S. is split into many migratory and resident stocks, though separation of the migratory stocks is still poorly understood (Hayes et al., 2019). Initially, due to the distribution of strandings during the 1987-88 epizootic die-off, it was believed there was a single migratory stock and the U.S. National Marine Fisheries Service (NMFS) managed the population under a single unit, the western North Atlantic coastal stock (Scott, Burn, & Hansen, 1988). Since then, scientists have studied genetic, distribution, and stranding patterns that refute this hypothesis, suggesting there are actually several seasonally-shifting coastal stocks (McLellan, Friedlaender, Mead, Potter, & Pabst, 2002; Rosel, Hansen, & Hohn, 2009; Zolman, 2002). In 2009, NMFS divided the single migratory management group into five (Waring, Josephson, Maze-Foley, & Rosel, 2010). According to these defined stocks and distribution patterns observed by Barco, Swingle, McLellan, Harris, and Pabst, (1999), dolphins stranded in Chesapeake Bay and Virginia waters may belong to one of three coastal stocks: (1) the Northern Migratory Stock, which occupies waters within the 20-m isobaths between New York and North Carolina; (2) the Southern Migratory Stock, which occupies waters ≤3 km from shore between Virginia and Florida; and (3) the Northern North Carolina Estuarine System stock, a resident estuarine stock that primarily occupies the Pamlico Sound, North Carolina but has been observed using waters as far north as lower Chesapeake Bay, Virginia and as far south as Oregon Inlet, North Carolina (Hayes et al., 2019). Although much has been learned about the possible stock structure of this species, studies are still needed to understand ecological, genetic, and population differences. Currently, it is still difficult to assign an individual to a defined stock. Given that the stock structure consists of two ecotypes (coastal and offshore) and a seasonally shifting mosaic of resident and migratory animals, the management of this species is understandably complex. One key factor in managing a species is understanding habitat and resource use. While the bottlenose dolphin’s coastal distribution makes it an attractive study subject due to its proximity to potential researchers, this also makes it particularly vulnerable to habitat destruction 2

and degradation by increased exposure to contaminants and biotoxins in their prey (e.g. Fire et al., 2011; Schaefer et al., 2011; Stavros et al., 2011; Twiner et al., 2011; Yordy et al., 2010), vessel noise (e.g. Pine, Jeffs, Wang, & Radford, 2016; Southall et al., 2019; Weilgart, 2007), harassment (e.g. Piwetz, 2018; Vail, 2016; Williams, 2009), and unnatural behaviors such as begging (e.g. Kovacs & Cox, 2014; Powell, Machernis, Engleby, Farmer, & Spradlin, 2018). Furthermore, even if algal blooms are non-toxic, hypoxic events can occur, resulting in decreased prey availability. Exposure to contaminants and biotoxins can be amplified depending on the dolphins’ primary diet via bioaccumulation. Understanding the diet of this predator can inform current and future management risks and strategies. Many estuarine systems in the southeastern U.S. have documented resident populations of bottlenose dolphins (e.g. Irvine, Scott, Wells, & Kaufmann, 1981; Mazzoil et al., 2008; Shane, Wells, & Würsig, 1986; Zolman, 2002); however, Virginia’s Chesapeake Bay system does not appear to have a resident population even though dolphins can be seen year-round, though with reduced frequency in winter (Barco et al., 1999; Fearnbach, 2004). Furthermore, Chesapeake Bay, specifically Cape Henry, Virginia Beach, serves as a nursery area for bottlenose dolphins (Barco et al., 1999). Other studies have reported on the diet of bottlenose dolphins on the U.S. east coast, though few have been done in the mid-Atlantic and many were limited in sample size and scope of study (Barros & Odell, 1990; Bowen, 2011; Gannon & Waples, 2004; McGurk, 1997; Mead & Potter, 1990; Pate & McFee, 2012). Once thought to have a catholic diet, accumulating evidence indicates Barros and colleagues’ hypothesis that dolphins passively listen for soniferous prey may be correct (Barros, 1993; Barros & Myrberg, 1987; Barros & Odell, 1990; Barros & Wells, 1998; Berens McCabe, Gannon, Barros, & Wells, 2010; Gannon et al., 2005). McGurk (1997) previously studied the diet of bottlenose dolphins stranded in Virginia; however, his sample size and investigation into factors influencing the diet were limited. For my thesis, the temporal range and sample size considered are twice that of McGurk (1997), and in addition to a description of dominant prey and seasonal variation, I investigate differences in age groups and sexes while also examining additional variables, such as reconstructed prey size. Regression estimates can vary by fish population. Therefore, establishing regression equations for prey in the predator’s region of study enhances accuracy of estimates of reconstructed prey size (Campana, 1990; Campana & Casselman, 1993). In my thesis, I present regression equations of reconstructed mass and length of local prey. Another major threat to the bottlenose dolphin is entanglement in fishing gear (Byrd & Hohn, 2017; Byrd et al., 2014; Hayes et al., 2019; Reeves, McClellan, & Werner, 2013). Although 3

gear is rarely still present on beach-cast carcasses, there are several indications consistent with peracute underwater entrapment, including the presence of recently ingested prey (Bernaldo de Quirós et al., 2018; Moore & Barco, 2013; Moore et al., 2013; Read & Murray, 2000). Often, stomach content analysis studies that attempt to compare bycaught and stranded small cetacean diet concentrate on qualitative (species composition) rather than quantitative measures (e.g. Barros & Odell, 1990; Mead & Potter, 1990; Santos, Fernández, López, Martínez, & Pierce, 2007; Silva, 1999; Tellechea, Perez, Olsson, Lima, & Norbis, 2017). The few that have (e.g. McGurk 1997, Pate and McFee 2012), attempted to correlate stomach “fullness” with evidence of human interaction but found no significant relationship. My study aimed to quantitatively investigate whether recently ingested prey is statistically linked to external lesions consistent with entanglement. Fish otoliths were coded by degradation level as a proxy for how recently the prey was ingested, with the expectation that the proportion of numerical abundance of early-code (less digested) otoliths is higher in dolphins with evidence of entanglement. My thesis is organized in three chapters. First, I established regression equations that relate otolith length/width to fish length and mass for key teleost prey in the diet of bottlenose dolphins (Tursiops truncatus) found in Virginia waters. Second, I identified dietary trends related to sex, ontogeny, and season in bottlenose dolphins stranded in Virginia using stomach content analyses and equations derived from chapter 2. Finally, I quantified whether recently ingested prey in dolphins is statistically linked to external lesions consistent with entanglement. This study used a uniquely large dataset to provide a comprehensive description of the diet of bottlenose dolphins stranded in Virginia, which will help to inform management on the ecology of, and threats to, this important ocean sentinel.

CHAPTER 2

LOCAL PREY SIZE ESTIMATION

INTRODUCTION When intact prey are not present in a piscivore’s stomach contents, researchers have to use hard prey parts for identification and size estimation of the prey species consumed, such as teleost fish otoliths. Otoliths are made of calcium carbonate and are the densest structure in the body of a teleost fish so they digest slower than other bones. Additionally, otolith morphology is species- specific. Therefore, it is the preferred hard structure to use when examining teleost fish in the cetacean diet, with the sagittae being the preferred otolith (out of the three types) for prey identification because of its large size (Fitch & Brownell Jr., 1968; Scott, 1903; Treacy & Crawford, 1981). Otolith size and fish size are significantly related (Aurioles-Gamboa, 1991; Granadeiro & Silva, 2000; Harvey, Loughlin, Perez, & Oxman, 2000; Hunt, 1992; Sinclair, Walker, & Thomason, 2015; Yilmaz, Yazicioglu, Yazici, & Polat, 2015). Reconstruction of prey size from otolith or squid beak size has been used in many cetacean diet studies (e.g. Barros, 1993; Barros & Wells, 1998; Gannon, Read, Craddock, & Mead, 1997; Gannon & Waples, 2004; Jansen, Leopold, Meesters, & Smeenk, 2010; Pate & McFee, 2012; Scheinin, Kerem, Lojen, Liberzon, & Spanier, 2014; Víkingsson et al., 2014). Relying solely on intact prey for identification of prey species and size estimation can bias against smaller prey items that digest more quickly. Fitch and Brownell (1968) highlighted the importance of using otoliths for prey identification in seven species of cetaceans. Using soft-tissue alone, they could only recognize two fish species in the stomachs of the dolphins, while 51 fish species were identified using otoliths. The use of otoliths, however, requires an extensive reference collection. Comparing sampled otoliths to photographs in publications can be difficult and unreliable, especially when closely related species are involved. Establishing a robust reference collection of local prey species, therefore, is essential for studying stomach contents. To investigate the diet of piscivores, reconstructed mass should be considered. While numerical abundance and frequency of occurrence are used often and can provide information on the importance of species in the diet, they can introduce biases when target animals consume large numbers of small prey or small numbers of large prey. By incorporating reconstructed mass, these cases can be accounted for. Indeed, two species consumed in similar numbers and frequency relative to other prey may not have the same biomass and are therefore not equally as important. Likewise, consumption of a few large prey may be more important energetically than many small prey. 5

Reconstructed mass provides a metric for quantifying these scenarios. Regression estimates can vary by fish population; therefore, establishing regression equations for the predator’s region of study will enhance accuracy of estimates of reconstructed mass (Campana, 1990; Campana & Casselman, 1993). The objective of this study was to establish regression equations relating otolith length and width to fish length and mass for eight teleost taxa that constitute at least 2% of the diet of bottlenose dolphins (Tursiops truncatus) stranded in Virginia (see Chapter 3, Table 5). These data provide important dimensional relationships between otoliths and fish length and mass, critical parameters for comprehensive diet analyses. The coastal morphotype of bottlenose dolphin is the most common marine mammal to occur in Virginia’s estuarine and marine waters. The bottlenose dolphin is a top predator and has previously been described as having a “catholic diet”. However, more recent studies along the U.S. east coast have identified a bias towards a more specialized diet in which soniferous fishes dominate (Barros & Odell, 1990; Barros & Wells, 1998; Berens McCabe et al., 2010; Bowen, 2011; Gannon & Waples, 2004; McGurk, 1997; Mead & Potter, 1990; Pate & McFee, 2012). To date, little is known about the diet preferences of the ecologically important coastal morphotype of bottlenose dolphin found in Virginia waters. The regression relationships developed in this study were used in a subsequent study (see Chapter 3) to document prey size and proportion of reconstructed mass for prey in the stomachs of bottlenose dolphins. These data provide information on the diet of dolphins in Virginia waters, allowing for the assessment of the importance of soniferous fishes (or other specific species) as prey targets.

MATERIALS AND METHODS Local fish were collected from Virginia’s estuarine and coastal ocean waters using several methods, including otter trawls, commercial pound nets and haul-seine nets, hook-and-line, beach seining, and electroshocking, by local fishermen, state agencies, and the Virginia Aquarium Acquisitions and Quarantine Department from 2012-2016. Through cooperation with these sources, I was able to collect data from fish collected under these ongoing projects, fish that did not survive collection and transport or fishermen discards. If the fish could not be processed within 24-hours, they were frozen at -10°C until they could be measured at a subsequent date. Fish smaller than 800 g were weighed to the nearest 0.1 g with a My Weigh PalmScale 8.0 (max capacity 800 g) digital scale. Larger fish were weighed to the nearest 0.5 g with an A&D 10SKWP scale (max capacity 10,000 g). Total length (maximum length from cranial-most point of the head to the furthest tip of the tail with 6

the caudal fin pinched together) was measured to the nearest 0.5 mm using a fish board. For the remainder of this study, “otolith” will refer to the sagittae. Otoliths from each fish were collected via removal of the ventrocaudal portion of the neurocranium and stored dry in plastic vials. Otoliths were examined for fractures, chips, or obvious evidence of abnormal calcite crystallization and were rejected if any of these anomalies were present. The straight length and width of small otoliths (approximately 3 mm or smaller) were measured, to the nearest 0.02 mm, using a stereoscope (Nikon SMZ1000) and stage micrometer. Large otoliths (greater than approximately 3 mm) were measured to the nearest 0.01 mm using digital calipers. In the cases of anchovy spp. (Anchoa spp.), Atlantic menhaden (Brevoortia tyrannus), and spotted hake (Urophycis regia) where otoliths were small and/or fragile, I used the stage micrometer method for all specimens. Silver perch (Bairdiella chrysoura) otoliths have pronounced protruding points that degrade quickly in the stomach acids of piscivores and thus can introduce variability in measurements. For these cases, the width was taken at the insertion of these points to the major body of the otolith rather than the widest point. The length and width of both the left and right otolith (or one if the other was rejected as noted above) were measured three times and the average otolith length and width for each fish were used in the regression equations. Least squares regression analyses relating otolith length and width to body length and weight were performed using SPSS for Windows (SPSS, Inc.; Version 22.0). Linear relationships between otolith length and width to fish total length are presented as y=ax+b. The otolith size-weight relationship as well as non-linear otolith width-total length relationships were determined using the least squares regression of the natural log of the otolith size and fish weight or total length, similar to other studies (e.g. Aurioles-Gamboa 1991, Harvey et al. 2000, Sinclair et al. 2015). This is presented in logarithmic form as ln(y)=b+a ln(x) and transformed back to arithmetic units and presented as y=ebxa. For some of the species, I was unable to collect enough data within the targeted size ranges to establish reliable regression equations. In these cases, I collaborated with Old Dominion University’s Center for Quantitative Fisheries Ecology and the Virginia Institute of Marine Science. These facilities had fish weight and length data for all otoliths in their collections. Thus, I used some of these relationships to supplement my own data when samples sizes were low. Supplemental data was considered for the following species: Atlantic croaker (Micropogonias undulatus), spot (Leiostomus xanthurus), seatrout spp. (Cynoscion spp.), and striped bass (Morone saxatilis). All scientific names were verified following Page et al., (2013).

RESULTS For this study, I performed regression analyses relating otolith length and width to fish length and weight for the eight fish taxa that constituted at least 2% of the diet of bottlenose dolphins stranded in Virginia: Atlantic croaker, spot, seatrout spp., Atlantic menhaden, spotted hake, anchovy spp., silver perch, and striped bass (Chapter 3, Table 5). The number of specimens sampled and the size range of the fish and otoliths are included in Tables 1-4. Seatrout spp. includes weakfish (Cynoscion regalis, n=83), spotted seatrout (Cynoscion nebulosus, n=128), and silver seatrout (Cynoscion nothus, n=25), whereas, anchovy spp. combines bay anchovy (Anchoa mitchilli, n=40) and striped anchovy (Anchoa hepsetus, n=39). Species were combined in these groups because it’s difficult to identify otoliths to the species level within these genera during stomach content analyses. Tables 1-4 list the parameters for the regression equations relating otolith length and width to fish total length and weight for the eight genera analyzed. Figures 1-8 illustrate the relationship between otolith size and fish size. All relationships were statistically significant at p<0.001 and a high correlation coefficient was obtained (R2≥0.90). Spotted hake had the lowest sample size, but still maintained highly significant regression relationships.

TABLE 1 Regression equations relating otolith length (OL, mm) to total length (TL, mm) of fish taxa important in bottlenose dolphin diet in Virginia. SE is standard error. All p <0.001. Equations presented as y=b+ax where y=TL, b=intercept, a=slope, and x=OL.

OL Range TL Range Common name n Equation SE R² (mm) (mm)

Anchovy spp. 79 1.2-4.9 43.5-142.5 TL = 16.5 + 25.0 (OL) 6.320 0.943 Atlantic croaker 142 3.3-22.2 62.5-510.0 TL = -28.4 + 25.0 (OL) 15.547 0.982 Atlantic menhaden 90 1.5-4.5 62.0-316.0 TL = -79.2 + 88.6 (OL) 15.239 0.916 Seatrout spp. 227 3.1-31.4 52.5-852.0 TL = -72.7 + 28.5 (OL) 28.114 0.985 Silver perch 70 1.5-5.9 34.0-183.5 TL = -19.3 + 35.9 (OL) 6.174 0.984 Spot 154 1.9-10.1 35.0-330.0 TL = -56.1 + 38.3 (OL) 11.624 0.974 Spotted hake 35 3.0-8.2 59.0-203.0 TL = -20.8 + 25.9 (OL) 5.086 0.978 Striped bass 97 2.4-24.6 60.0-1265.0 TL = -194.0 + 51.2 (OL) 76.685 0.925

TABLE 2 Regression equations relating otolith width (OW, mm) to total length (TL, mm) of fish taxa important in bottlenose dolphin diet in Virginia. SE is standard error. All p <0.001. Equations presented as y=b+ax or ln(y)=b+a ln(x) where y=TL, b=intercept, a=slope, and x=OW. OW TL Range Common name n Range Equation SE R² (mm) (mm) Anchovy spp. 79 0.9-2.9 43.5-142.5 TL = -2.4 + 53.5 (OW) 6.105 0.947 Atlantic croaker 142 2.2- 16.7 62.5-510.0 TL = -19.4 + 31.3 (OW) 18.390 0.975 Atlantic menhaden 90 0.8-2.0 62.0-316.0 TL = -73.4 + 181.5 (OW) 16.649 0.900 Seatrout spp. 233 1.7-11.4 52.5-852.0 ln(TL) = 2.5 + 1.2 ln(OW) 0.063 0.993 1.2 TL = 12.6 (OW) Silver perch 70 1.5-5.4 34.0-183.5 TL = -29.3 + 41.4 (OW) 6.674 0.981 Spot 155 1.4-4.8 35.0-330.0 ln(TL) = 2.6 + 2.1 ln(OW) 0.110 0.949 2.1 TL = 13.0 (OW) Spotted hake 35 1.2-2.7 59.0-203.0 TL = -54.1 + 93.9 (OW) 9.161 0.929 Striped bass 96 1.5-9.9 60.0-1265.0 ln(TL) = 3.3 + 1.6 ln(OW) 0.141 0.962 TL = 27.1 (OW)1.6

TABLE 3 Regression equations relating otolith length (OL, mm) to wet weight (W, g) of fish taxa important in bottlenose dolphin diet in Virginia. SE is standard error. All p<0.001. Equations presented as ln(y)=b+a ln(x) or y=ebxa where y=W, b=intercept, a=slope, and x=OL. OL W Range Common name n Range Equation SE R² (g) (mm) Anchovy spp. 79 1.2-4.9 0.5-24.9 ln(W) = -1.1 + 2.6 ln(OL) 0.264 0.934 2.6 W = 0.3 (OL) Atlantic croaker 144 3.3-22.2 2.2-1,755.4 ln(W) = -3.5 + 3.6 ln(OL) 0.192 0.986 3.6 W = 0.03 (OL) Atlantic menhaden 92 1.5-4.5 2.4-300.4 ln(W) = -0.89 + 4.6 ln(OL) 0.243 0.942 4.6 W = 0.4 (OL) Seatrout spp. 230 3.1-31.4 1.5-5,550.0 ln(W) = -3.8 + 3.6 ln(OL) 0.231 0.989 3.6 W = 0.02 (OL) Silver perch 70 1.5-5.9 0.4-97.0 ln(W) = -2.1 + 3.8 ln(OL) 0.200 0.984 3.8 W = 0.1 (OL) Spot 153 1.9-10.1 0.5-485.3 ln(W) = -3.5 + 4.3 ln(OL) 0.185 0.985 4.3 W = 0.03 (OL) Spotted hake 35 3.0-8.2 1.7-75.0 ln(W) = -3.7 + 3.7 ln(OL) 0.169 0.970 3.7 W = 0.03 (OL) Striped bass 95 2.4-24.6 2.0-27,850.9 ln(W) = -3.91 + 4.22 ln(OL) 0.289 0.983 W = 0.02 (OL)4.2 9

TABLE 4 Regression equations relating otolith width (OW, mm) to wet weight (W, g) of fish taxa important in bottlenose dolphin diet in Virginia. SE is standard error. All p<0.001. Equations presented as ln(y)=b+a ln(x) or y=ebxa where y=W, b=intercept, a=slope, and x=OW and eb is calculated. OW W Range Common name n Range Equation SE R² (g) (mm) Anchovy spp. 79 0.9-2.9 0.5-24.9 ln(W) = -0.2 + 3.3 ln(OW) 0.269 0.965 3.3 W = 0.8 (OW) Atlantic croaker 144 2.2-16.7 2.2-1,755.4 ln(W) = -2.1 + 3.4 ln(OW) 0.206 0.984 3.4 W = 0.1 (OW) Atlantic menhaden 92 0.8-2.0 2.4-300.4 ln(W) = 2.5 + 4.6 ln(OW) 0.272 0.928 4.6 W = 12.1 (OW) Seatrout spp. 236 1.7-11.4 1.5-5,550.0 ln(W) = -3.2 + 4.9 ln(OW) 0.280 0.984 4.9 W = 0.04 (OW) Silver perch 70 1.5-5.4 0.4-97.0 ln(W) = -2.4 + 4.2 ln(OW) 0.215 0.981 4.2 W = 0.09 (OW) Spot 154 1.4-4.8 0.5-485.3 ln(W) = -3.8 + 6.4 ln(OW) 0.340 0.950 6.4 W = 0.02 (OW) Spotted hake 35 1.2-2.7 1.7-75.0 ln(W) = -0.3 + 4.7 ln(OW) 0.284 0.915 4.7 W = 0.8 (OW) Striped bass 95 1.5-9.9 2.0-27,850.9 ln(W) = -1.7 + 4.8 ln(OW) 0.408 0.965 4.8 W = 0.2 (OW)

The dependence of total length on otolith length/width generally followed a linear relationship for most fish; however, in three of the taxa, i.e., seatrout spp., spot and striped bass, the otolith width to fish total length was better fit to a logarithmic curve using the natural log of the otolith width and fish total length, both when assessed visually and analyzing the correlation coefficient (seatrout spp. linear R2=0.973, log curve R2=0.990; spot linear R2=0.888, log curve R2=0.949; striped bass linear R2=0.868, log curve R2=0.962). All relationships between otolith length/width and wet weight were logarithmic (Figures. 1-8).

Anchovy spp. (Anchoa spp.)

TL= 16.5 + 25.0 (OL) TL= -2.4 + 53.5 (OW) 150 R2= 0.943 150 R2= 0.947

100 100 Total Length (mm) Total Length (mm)

A 50 50 B 1.0 3.0 5.0 0.5 1.5 2.5 Otolith Length (mm) Otolith Width (mm)

25 W= 0.33 (OL)2.6 25 W= 0.83 (OW)3.3 R2= 0.934 R2= 0.965

15 15 Wet Weight (g) Wet Weight (g)

5 5 C D 0 0 1.0 3.0 5.0 0.5 1.5 2.5 Otolith Length (mm) Otolith Width (mm)

FIGURE 1 Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for anchovy spp. (Anchoa spp.). A) Otolith length to total length; B) Otolith width to total length; C) Otolith length to wet weight; D) Otolith width to wet weight.

Atlantic Croaker (Micropogonias undulatus)

TL= -28.4 + 25.0 (OL) TL= -19.4 + 31.3 (OW) R2= 0.982 R2= 0.975

500 500

300 300 Length (mm) Total Length (mm) Total Total

100 100 A B 0 0 0 10 20 0 10 20 Otolith Length (mm) Otolith Width (mm)

W= 0.03 (OL)3.6 W= 0.13 (OW)3.4 R2= 0.986 R2= 0.984

1500 1500 ) ) Wet Weight (g Wet Weight (g 500 500 C D 0 0 0 10 20 0 10 20 Otolith Length (mm) Otolith Width (mm)

FIGURE 2 Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for Atlantic croaker (Micropogonias undulatus). A) Otolith length to total length; B) Otolith width to total length; C) Otolith length to wet weight; D) Otolith width to wet weight.

Atlantic Menhaden (Brevoortia tyrannus)

TL= -79.2 + 88.6 (OL) TL= -73.4 + 181.5 (OW) R2= 0.916 R2= 0.900

300 300

200 200 Total Length (mm) Total Length (mm)

100 A 100 B 50 50 1.5 2.5 3.5 4.5 0.8 1.2 1.6 2.0 Otolith Length (mm) Otolith Width (mm)

W= 0.41 (OL)4.6 W= 12.1 (OW)4.6 R2= 0.942 R2= 0.928

300 300 ) ) g g ( (

Wet Weight Wet Weight 100 100 C D 0 0 1.5 2.5 3.5 4.5 0.8 1.2 1.6 2.0

Otolith Length (mm) Otolith Width (mm) FIGURE 3 Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for Atlantic menhaden (Brevoortia tyrannus). A) Otolith length to total length; B) Otolith width to total length; C) Otolith length to wet weight; D) Otolith width to wet weight.

Seatrout spp. (Cynoscion spp.)

1000 TL= -72.7 + 28.5 (OL) 1000 TL= 12.6 (OW)1.2 R2= 0.985 R2= 0.993

600 600 Total Length (mm) Total Length (mm)

200 200 A B 0 0 0 20 40 0 4.0 8.0 12 Otolith Length (mm) Otolith Width (mm)

W= 0.02 (OL)3.6 W= 0.04 (OW)4.9 R2= 0.989 R2= 0.984 5000 5000

) ) g g ( (

3000 3000 Wet Weight Wet Weight

1000 C 1000 D 0 0 0 20 40 0 4.0 8.0 12 Otolith Length (mm) Otolith Width (mm)

FIGURE 4 Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for seatrout spp. (Cynoscion spp.). A) Otolith length to total length; B) Otolith width to total length; C) Otolith length to wet weight; D) Otolith width to wet weight.

Silver Perch (Bairdiella chrysoura)

TL= -19.3 + 35.9 (OL) TL= -29.3 + 41.4 (OW) R2= 0.984 R2= 0.981

150 150 Total Length (mm) Total Length (mm) 50 50 A B 0 0 1.0 3.0 5.0 1.0 3.0 5.0 Otolith Length (mm) Otolith Width (mm)

100 W= 0.12 (OL)3.8 100 W= 0.09 (OW)4.2 R2= 0.984 R2= 0.981

) )

60 60 Wet Weight (g Wet Weight (g

20 20 C D 0 0 1.0 3.0 5.0 1.0 3.0 5.0 Otolith Length (mm) Otolith Width (mm)

FIGURE 5 Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for silver perch (Bairdiella chrysoura). A) Otolith length to total length; B) Otolith width to total length; C) Otolith length to wet weight; D) Otolith width to wet weight.

Spot (Leiostomus xanthurus)

TL= -56.1 + 38.3 (OL) TL= 13.0 (OW)2.1 2 2

R = 0.974 R = 0.949

300 300 Total Length (mm) Total Length (mm) 100 100 A B 0 0 0 4.0 8.0 12 1.0 3.0 5.0 Otolith Length (mm) Otolith Width (mm)

500 W= 0.03 (OL)4.3 500 W= 0.02 (OW)6.4 R2= 0.985 R2= 0.950

300 300 Wet Weight (g) Wet Weight (g) 100 100 C D 0 0 4.0 8.0 12 1.0 3.0 5.0 Otolith Length (mm) Otolith Width (mm)

FIGURE 6 Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for spot (Leiostomus xanthurus). A) Otolith length to total length; B) Otolith width to total length; C) Otolith length to wet weight; D) Otolith width to wet weight.

Spotted Hake (Urophycis regia)

TL= -20.8 + 25.9 (OL) TL= -54.1 + 93.9 (OW) R2= 0.978 R2= 0.929

200 200 Total Length (mm) 100 Total Length (mm) 100 A B 50 50 3.0 5.0 7.0 9.0 1.0 2.0 3.0 Otolith Length (mm) Otolith Width (mm)

W= 0.03 (OL)3.7 W= 0.75 (OW)4.7 R2= 0.970 R2= 0.915

60 60 Wet Weight (g) Wet Weight (g) 20 20 C D 0 0 3.0 5.0 7.0 9.0 1.0 2.0 3.0 Otolith Length (mm) Otolith Width (mm)

FIGURE 7 Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for spotted hake (Urophycis regia). A) Otolith length to total length; B) Otolith width to total length; C) Otolith length to wet weight; D) Otolith width to wet weight.

Striped Bass (Morone saxatilis)

1500 1500 TL= -194.0 + 51.2 (OL) TL= 27.1 (OW)1.6 R2= 0.925 R2= 0.962

500 500 Total Length (mm) Total Length (mm)

A B 0 0 0 10 20 0 4.0 8.0 Otolith Length (mm) Otolith Width (mm)

30000 W= 0.02 (OL)4.2 30000 W= 0.18 (OW)4.8 R2= 0.983 R2= 0.965

Wet Weight (g) Wet Weight (g) 10000 10000

C D 0 0 0 10 20 0 4.0 8.0 Otolith Length (mm) Otolith Width (mm)

FIGURE 8 Regression analysis of otolith length (OL, mm) and width (OW, mm) to fish total length (TL, mm) and wet weight (W, g) for striped bass (Morone saxatilis). A) Otolith length to total length; B) Otolith width to total length; C) Otolith length to wet weight; D) Otolith width to wet weight.

DISCUSSION While otolith length/width to total length followed the typical linear relation for most taxa, three taxa exhibited a better fit when a logarithmic curve was applied for otolith width to total length relationships. The otolith length/width to fish length relationship is not always linear (Campana, 2004; Hunt, 1992). This is because otolith size and shape can change disproportionally with growth, 18

with many fish larvae having more spherical otoliths than later ontogenetic stages (Campana, 2004). Based on the data presented here, seatrout spp., spot, and striped bass exhibit the most deviation from linearity because of greater disproportionate otolith growth (these species have relatively elongated otoliths as they grow) through ontogeny, and thus non-linear curves work best for their total length predictions. The otolith size to wet weight of all eight taxa were best fit to a logarithmic curve that is similar in form to what is typically used in fisheries for length-weight relationships: W=aLb where W is body weight, L is length, a is a coefficient related to body form and b is an exponent indicating isometric growth when equal to 3 (e.g. Aurioles-Gamboa, 1991; Harvey et al., 2000; Hunt, 1992; Sinclair et al., 2015). The high correlation coefficients for these equations indicate they were appropriate for estimating fish weight from otolith length or width for these species. I examined otolith width as well as otolith length due to the prevalence of broken otoliths in the sampled stomach contents. Many of my samples were bagged contents instead of whole stomachs and therefore, they were more likely to have broken otoliths as a result of handling and decreased protection from stomach tissue. To use these broken otoliths in diet analyses, equations relating otolith width to length were needed. This is because in almost all cases of broken otoliths, the length was compromised while the width was preserved. Many cetacean diet studies use a two-step process when relating otolith size to fish weight (e.g. Barros, 1993; Bowen, 2011; Gannon, Craddock, & Read, 1998; Gannon et al., 1997; Gannon & Waples, 2004; Pate, 2008; Recchia & Read, 1989). First, the total length is calculated from the otolith length. Second, wet weight is calculated from total length. In many cases, different datasets are used for the two steps. In my study, I reduced compounding errors associated with the conventional two- step process because I measured both parameters for each fish and thus derived more direct relationships between otolith size and fish wet weight, similar to that used in other studies (e.g. Aurioles-Gamboa, 1991; Lindstrøm, Nilssen, Pettersen, & Haug, 2013; Lopes, da Silva, Bassoi, dos Santos, & de Oliveira Santos, 2012; Melo et al., 2010). The equations in this study can be applied to other predator-prey stomach content studies involving the eight fish species. The equations are particularly useful for studies in or near Virginia waters, as otolith shape varies with fish stock and region (Campana & Casselman, 1993). Degraded otoliths can lead to underestimates of the size of fish consumed. However, eliminating degraded otoliths and using only non-degraded otoliths can lead to overestimates of mean prey size because larger individuals digest slower. To balance these biases, Gannon and Waples (2004) and Recchia 19

and Read (1989) suggest scaling the degradation level and considering a range of otolith levels from fully intact to moderately degraded for prey size estimation. The otolith width relationships presented in the current study promise to be useful for these analyses because width measurements are often more reliable than length measurements as otoliths degrade. Many of the common prey identified in the stomach contents of bottlenose dolphins in Virginia (Chapter 3) are also important in fisheries of the Chesapeake Bay and coastal Virginia waters (e.g., Atlantic menhaden, seatrout spp., striped bass, croaker). Understanding the importance of these economically significant fishes in the diets of dolphins, a common piscivorous predator in Virginia waters, can inform management decisions, especially when an ecosystem-approach rather than a single-species approach to fisheries management is taken. Accounting for the complex interactions between species via the food web in models can provide insights into the resiliency of a population to fishery pressures, especially when data on ontogenetic stage can be gleaned from diet analyses. Recently, commitments to an ecosystem approach to fisheries have become more common in national policy and legislation regarding fisheries (Jennings, Smith, Fulton, & Smith, 2014), which is a positive step forward in management practices. In summary, establishing regression equations that relate otolith size to fish size can be a useful tool in piscivore diet studies. These data are essential for a comprehensive understanding of the size, mass, and frequency of prey consumed when stomach contents are degraded. The information provided in diet studies can also be used to understand the complex interactions within a food web, which can be incorporated into ecosystem models for fisheries management. I have provided the regression equations for eight prey taxa that are common in Virginia estuarine and coastal ocean waters as well as endemic to other U.S. east coast waters. While these data promise to be useful for diet analyses of a number of local predators, one limitation to the dataset is that the fish size ranges match those of prey consumed by bottlenose dolphins in Virginia waters. Therefore, the equations should not be used to predict sizes beyond the data range considered in this study. Additional data at the extremes of the size range will need to be added for accurate predictions for larger and smaller otolith sizes.

CHAPTER 3

DEMOGRAPHIC PATTERNS OF DIET

INTRODUCTION Understanding the diet of marine mammals is necessary to understand their ecological role in marine systems (Bowen, 1997). Natural history sampling of cetaceans is complicated because they spend the majority of their lives underwater and out of sight. Given the difficulties associated with sampling live aquatic animals, collecting data from stranded animals is an efficient, cost-effective approach. The bottlenose dolphin (Tursiops truncatus) is an upper trophic level predator distributed throughout temperate and tropical waters of the world. It is a well-studied cetacean due to its coastal distribution, prevalence in captivity and nature, and its use as a marine ecosystem sentinel (Bossart, 2011; Moore, 2008; Wells et al., 2004). Many estuarine systems have documented resident populations of dolphins (e.g. Mazzoil et al., 2008; Stolen, Durden, & Odell, 2007; Zolman, 2002), some of which are genetically distinct from animals in coastal waters (Rosel et al., 2009) or nearby estuaries (Litz, Hughes, Garrison, Fieber, & Rosel, 2012); however, Virginia’s Chesapeake Bay system does not appear to have a resident population. In Virginia waters, the bottlenose dolphin is present seasonally, typically from April through November (Barco et al., 1999; Fearnbach, 2004). Furthermore, Chesapeake Bay, specifically Cape Henry, Virginia Beach, serves as a nursery area for bottlenose dolphins (Barco et al., 1999). In Sarasota Bay, Florida, females with young calves spend more time in nursery areas that are shallow and protected with high prey availability to meet the energy demands of lactation (Scott, Wells, & Irvine, 1990). Prey availability is an important factor in driving the movement of dolphins (Shane et al., 1986), but the relationship between prey availability and dolphin presence in Virginia remains unclear (Barco et al., 1999). Coastal bottlenose dolphins were once thought to have a broad diet, but recent studies have shown their diets are more specialized. Along the U.S. east coast, the coastal bottlenose dolphin diet is dominated by sciaenids (croakers and drums), with haemulids (grunts), mugilids (mullets), clupeids (herrings), sparids (sea breams and porgies), scombrids (mackerels), engraulids (anchovies), and loliginids (inshore squids) also being important depending on region (Barros & Odell, 1990; Bowen, 2011; Gannon & Waples, 2004; Mead & Potter, 1990; Pate & McFee, 2012). Barros and colleagues hypothesized that bottlenose dolphins passively listen for soniferous prey rather than using energetically expensive echolocation when searching and, therefore, have a disproportionately high amount of sound-producing prey in their diet (Barros, 1993; Barros & Odell, 1990; Barros & Wells, 21

1998). This hypothesis was tested and corroborated through acoustic playback experiments with free-ranging dolphins in Sarasota Bay, Florida (Gannon et al., 2005). Further, Berens McCabe et al. (2010) investigated prey selection of bottlenose dolphins in Sarasota Bay, Florida and found that dolphins select for soniferous fishes and against non-soniferous fishes in their diet. Variation in the bottlenose dolphin diet has been associated with habitat, season, and demography. Differences in the use of seagrass versus open water habitats in Sarasota Bay, Florida influenced diet changes between males and females, with males eating a more restricted diet of seagrass-associated fishes and females feeding in a variety of habitats and prey resources (Rossman et al., 2015). Gannon and Waples (2004) documented differences in diet between adult male and female bottlenose dolphins in North Carolina. Inshore squid (Doryteuthis sp., formerly Loligo sp.) was more important in adult female diet, making up approximately twice as much of the numerical abundance than in the other demographic groups examined (juvenile female, juvenile male, adult male); however, this difference was not statistically significant. In pantropical spotted dolphins (Stenella attenuata), Bernard and Hohn (1989) found pregnant females consumed significantly more squid than fish compared to lactating females. However, Robertson and Chivers (1997) found lactating females consumed significantly more squid than pregnant females but the number of fish consumed was not significantly different. These studies attributed the differences to changes in dietary needs for lactating or pregnant mothers or changes in habitat preference for mothers with calves. Most dolphins consume low trophic level, seagrass-associated fishes throughout ontogeny. However, some juvenile dolphins in Sarasota, FL feed on prey outside of seagrass communities, possibly due to physiological constraints such as gape size or swimming speed bursts, or the lack of skills necessary for complex foraging (Rossman et al., 2015). As the studies above demonstrate, diets of dolphins can be highly variable and impacted by multiple factors. Documentation of diet preferences in dolphins can inform researchers on the behavior or natural history of these difficult to study animals. The diet of stranded bottlenose dolphins in Virginia has been studied previously by McGurk (1997). Through examining the stomach contents of 94 dolphins that stranded between 1987 and 1996, McGurk found that Atlantic croaker (Micropogonias undulatus) was the primary prey target of bottlenose dolphins in Virginia waters, followed by weakfish (Cynoscion regalis), spot (Leiostomus xanthurus), silver perch (Bairdiella chrysoura) and squid (Doryteuthis sp.). He explored seasonal variation, the effects of human interaction, and compared the diet of dolphins that stranded during the mass die-off of dolphins in 1987-1988 to those after the event. McGurk’s results demonstrated that sciaenids dominated the diet, evidence of human interaction (e.g. gillnet entanglement) was not 22

significantly correlated with stomach fullness, and the diet of dolphins that stranded during the die- off was not significantly different from those that stranded after. In this study, I explored the diet of T. truncatus in Virginia waters as well but expanded the dataset and analysis. My sample size and temporal range were twice that of McGurk (1997), and I investigated differences in age groups and sexes while also examining additional variables, such as reconstructed size of prey. My goal was to identify dietary trends related to sex, ontogeny, and season in bottlenose dolphins stranded in Virginia using stomach content analyses. I presented these trends in terms of proportion of numerical abundance, frequency of occurrence, and proportion of reconstructed mass of common prey. My study, together with life history data presented in Lynott (2012), provides a comprehensive natural history picture for stranded, coastal bottlenose dolphins in the mid-Atlantic.

MATERIALS AND METHODS All samples for this study were from bottlenose dolphins stranded from 1998-2012 in Virginia and stored frozen in the Virginia Aquarium Stranding Response Program (VAQS) archive. Samples were collected from stranded dolphins whenever available. During necropsy, VAQS staff and volunteers either 1) removed the entire stomach, including whole or partial esophagus and the duodenal ampulla or 2) carefully extracted all prey items from the stomachs and stored them in a plastic bag then froze the sample at -20°C. Bottlenose dolphins used in previous life history projects focusing on age and/or sexual maturity determinations (see Lynott, 2012) were prioritized for this project (n=138), as my goal was to complement these data. All of the animals in my study stranded independently (i.e. no animals comprising mother/calf pairs or involved in mass strandings were considered) and either had genetic confirmation of the coastal morphotype (n=118), or were determined to be the coastal morphotype through examination of morphological characters and/or identification of parasites according to Mead and Potter (1990, 1995). During this project, the stomach contents were analyzed following methods adapted from Mead and Potter (1990), Barros and Wells (1998), and Gannon and Waples (2004). All intact prey and fish skulls were carefully removed via hand-sorting. The remaining contents were rinsed with fresh water into a 500 µm sieve and hand-sorted. Whole prey items (fish whole from head to hypural plate or terminal vertebrae if species did not possess hypural plate; squid with intact mantle or gladius) were measured as follows: standard length (SL) and/or total length (TL) for fish and mantle length and/or gladius length for squid to the nearest 0.5 mm using a fish board. Parts used to 23

identify prey included: otoliths, dentary bones, and opercular bones of fishes; upper and lower beaks of cephalopods; and exoskeletons of crustaceans. Gannon and Waples (2004) used the isopod Olencira praegustator, as a proxy for Atlantic menhaden counts due to the fish’s delicate and quickly digested otoliths. However, the count of this isopod never exceeded the count of half the number of otoliths in my study, and thus was not used as a proxy measurement. Otoliths were stored dry in plastic vials, cephalopod beaks were stored in 70% ethanol in plastic vials, and fish bones were re- frozen at -20°C. Prey items were identified to the lowest possible taxon using an in-house reference collection of 90 species (43 families) collected within Virginia’s oceanic and estuarine waters from 2008-2015 by VAQS, Old Dominion University, or Virginia Institute of Marine Science (see Chapter 2); a reference collection of squid beaks at University of North Carolina-Wilmington; and available published guides (Baremore & Bethea, 2010; Campana, 2004; Clarke, 1986; Mohsin, 1981). The number of squid present in the stomach was determined by summing the numbers of either upper or lower beaks (whichever was greater) from each species. The number of each fish species was determined by summing half the number of sagittal otoliths. For those species where the sagittal otolith is small or delicate, such as the Atlantic menhaden (Brevoortia tyrannus) or the bluefish (Pomatomous saltatrix), other bones, such as the dentary or opercular bones, were counted. Whichever was greater, the otolith or opercular/dentary bone count, was used to estimate fish abundance in the stomachs. Otoliths were scored 0 (undamaged otoliths retrieved from skulls) to 5 (severely degraded, free otoliths) following the methods of Recchia and Read (1989). I measured the sagittal otoliths and estimated fish prey size following Gannon and Waples (2004). For the remainder of this paper, “otolith” will refer to sagittal otolith unless otherwise specified. Typically, the length of the otolith was measured, but if the length was compromised (i.e. broken tip), the width was measured instead. In the case of silver perch (Bairdiella chrysoura), whose otoliths have pronounced protruding points, the width was taken at the insertion of these points to the major body of the otolith. This method was used because the protruding points are susceptible to faster degradation in the stomach acids of piscivores. To characterize the diet of stranded bottlenose dolphins in Virginia, I followed the methodology used by Gannon and Waples (2004) for proportion of numerical abundance, proportion of reconstructed mass, and frequency of occurrence. The proportion of numerical abundance is the count of a prey taxa in a particular stomach, divided by the total number of prey 24

items in that stomach, then averaged across all stomachs. The proportion of reconstructed mass is the sum of weights (reconstructed via equations noted below) of a prey taxa in a particular stomach, divided by the total weight of prey items in that stomach, then averaged across all stomachs. Frequency of occurrence is defined as the proportion of dolphins that consumed a particular prey taxa. Fish from the in-house reference collection were used to construct regression equations relating otolith length and width to fish length and wet weight for the eight most common bottlenose dolphin fish prey species (see Chapter 2). Equations for estimating the dorsal mantle length and wet weight of longfin inshore squid (Doryteuthis pealeii) from the lower rostral length were derived from Lange and Johnson (1981) and Staudinger, Juanes, and Carlson (2009). All prey species that were less than 2% of the proportion of numerical abundance were combined into an “other” category. Regression equations were not available for all species, so only prey that accounted for 2% or more of the diet were analyzed via reconstructed mass. Refer to Table 5 for prey items less than 2% of the numerical abundance. The total number of prey taxa (diversity) and total number of prey items (quantity) per dolphin were counted. When the mean is presented, standard deviation (SD) is given as the measure of variance using the convention x ± SD. The seasons were defined as follows: 1) winter included December-February, 2) spring included March- May, 3) summer included June-August, and 4) fall included September-November. Age groups were defined according to age, growth, and sexually dimorphic parameters in Lynott (2012), Mead and Potter (1990), and Stolen, Odell, and Barros (2002) as follows: 1) calves were sexually immature and 1.5 years or younger (or <170 cm body length when age data were unavailable); 2) juveniles were sexually immature and 1.6 - 10 years old (or 171 - 249 cm) for males or 1.6 - 8 years old (or 171 - 229 cm) for females; and 3) adults were sexually mature and >10 years old (or >250 cm) for males or >8 years (or >230 cm) for females.

Statistical Analyses Statistical analyses were completed using SPSS for Windows (SPSS, Inc.; Version 22.0). One- way MANOVAs on ranked data were used to compare proportion of numerical abundance and proportion of reconstructed mass between sexes, age groups, and seasons. Follow-up ANOVAs were performed for significant variables along with Tukey-Kramer post-hoc tests for pairwise comparisons. Mann-Whitney U tests and Kruskal-Wallis H tests were used to compare diversity, quantity, total reconstructed mass, average reconstructed length, and average reconstructed weight of each prey taxon per dolphin between sexes, age groups, and seasons. One-way ANOVAs on log 25

transformed data were used to compare total reconstructed mass per dolphin between sexes, age groups, and seasons. An independent-samples t-test on raw data, one-way ANOVA on log transformed data, and one-way ANOVA on square root transformed data were used to compare the overall average length and weight of prey consumed per dolphin between sexes, age groups, and seasons, respectively. Pairwise comparisons in the Kruskal-Wallis H tests were from a Dunn’s procedure. A chi-square goodness-of-fit test was used when exploring differences in frequency of occurrence of primary prey species between sexes, age groups, and seasons.

RESULTS I examined the stomachs or bagged stomach contents of 228 bottlenose dolphins stranded from 1998 to 2012 in Virginia. Of these stomachs/bags, 19 were empty, two contained only milk, one was later determined to belong to an offshore morphotype, and six were incomplete samples. All of these were removed from analysis. Thus, the stomach contents of 200 bottlenose dolphins formed the basis of this study. These dolphins stranded throughout Virginia’s tidal waters and were concentrated along lower Chesapeake Bay and Virginia Beach’s oceanfront. There were 94 females and 106 males that ranged in straight body length from 157 cm to 295 cm and in age from 0.8 years to 47 years. There were 20 calves, 128 juveniles, and 52 adults. The smallest dolphin with prey contents was 157 cm, but age data were not available for that animal. The youngest aged animal with prey contents was 0.8 years old and 162 cm. The two animals that contained only milk were 130 cm and 185 cm; neither were aged. These dolphins stranded in every month of the year except February. Only six dolphins stranded in winter, though the other seasons were well represented: 60 in spring, 78 in summer, and 56 in fall. At least 10,426 prey items representing 32 species from 22 families were identified from the 20,929 otoliths (19,692 sagittae, 715 lapilli, 376 asterisci, 146 unknown) and 344 squid beaks found in the 200 dolphins analyzed (Table 5). Two dolphins contained only squid, 153 contained only fish, and 45 contained both fish and squid. The family Sciaenidae dominated the diet, occurring in 93% of the dolphins and accounting for 71% of the diet numerically (Table 5). After Sciaenidae, the top five families included Clupeidae, Phycidae, Engraulidae, Moronidae and the squid family Loliginidae. The frequency of occurrence (%FO) of these five families in the stomach contents of the dolphins ranged from 14 to 32%, while the proportion of numerical abundance (%NA) ranged from 3 to 6% (Table 5). On average, dolphin stomachs contained 52.1 ± 67.9 prey items (range 1-475); however, 26

over half contained only 1-30 prey items. The diversity of prey per dolphin was 4.2 ± 2.6 taxa (range 1-15). Croaker (Micropogonias undulatus) was the most important species, accounting for most of the diet by number, frequency, and reconstructed mass (Table 5). Spot (Leiostomus xanthurus) and seatrout spp. (Cynoscion spp.) were the next most important taxa. Following distantly, taxa that accounted for 2% or more of the numerical abundance were menhaden, hake spp. (Urophycis spp.), anchovy spp. (Anchoa spp.), longfin inshore squid (Doryteuthis pealeii), silver perch (Bairdiella chrysoura), and striped bass (Morone saxatilis).

TABLE 5 Proportion of numerical abundance (%NA), frequency of occurrence (%FO), and proportion of reconstructed mass (%RM) of prey in the stomach contents of bottlenose dolphins stranded in Virginia (n=200). Presented in order of proportion of numerical abundance for families of fish and cephalopods. Prey % NA % FO % RM BONY FISHES Sciaenidae 71 93 81 Atlantic croaker (Micropogonias undulatus) 35 69 41 Spot (Leiostomus xanthurus) 17 52 14 Seatrout spp. (Cynoscion spp.) 14 58 23 Silver perch (Bairdiella chrysoura) 3 25 3 Banded drum (Larimus fasciatus) <1 2 ― Kingfish spp. (Menticirrhus spp.) <1 5 ― Black drum (Pogonias cromis) <1 2 ― Red drum (Sciaenops ocellatus) <1 <1 ― Star drum (Stellifer lanceolatus) <1 <1 ― Unidentified Sciaenid <1 <1 ― Clupeidae 6 32 8 Atlantic menhaden (Brevoortia tyrannus) 5 28 8 Gizzard shad (Dorosoma cepedianum) <1 4 ― Unidentified Clupeid <1 2 ― Phycidae 4 26 2 Hake spp. (Urophycis spp.) 4 26 2 Engraulidae 4 28 <1 Anchovy spp. (Anchoa spp.) 4 28 <1 Moronidae 3 14 5 Striped bass (Morone saxatilis) 3 14 5 Ophidiidae 1 11 ― Striped cusk-eel (Ophidion marginatum) 1 11 ― 27

TABLE 5, continued Prey % NA % FO % RM Paralichthyidae <1 9 ― Flounder spp. (Paralichthys spp.) <1 9 ― Batrachoididae <1 6 ― Oyster toadfish (Opsanus tau) <1 6 ― Congridae <1 6 ― Conger eel (Conger oceanicus) <1 6 ― Anguillidae <1 3 ― American eel (Anguilla rostrata) <1 3 ― Tetradontidae <1 <1 ― Northern puffer (Sphoeroides maculatus) <1 <1 ― Serranidae <1 6 ― Black sea bass (Centropristis striata) <1 6 ― Haemulidae <1 3 ― Pigfish (Orthopristis chrysoptera) <1 3 ― Sparidae <1 6 ― Pinfish (Lagodon rhomboides) <1 6 ― Stromateidae <1 6 ― Butterfish spp. (Peprilus spp.) <1 6 ― Pomatomidae <1 4 ― Bluefish (Pomatomus saltatrix) <1 4 ― Cyprinodontidae <1 1 ― Striped killifish (Fundulus majalis) <1 1 ― Mugilidae <1 1 ― Striped mullet (Mugil cephalus) <1 1 ― Labridae <1 <1 ― Tautog (Tautoga onitis) <1 <1 ― Synodontidae <1 <1 ― Inshore lizardfish (Synodus foetens) <1 <1 ― Achiridae <1 <1 ― Hogchoker (Trinectes maculatus) <1 <1 ― Unidentified Fish 3 20 ― Unidentified fish spp. 3 20 ― CEPHALOPODS Loliginidae 4 24 4 Longfin inshore squid (Doryteuthis pealeii) 3 16 4 Atlantic brief squid (Lolliguncula brevis) <1 5 ― Unidentified Loliginid <1 4 ―

The largest reconstructed fish was a 780 mm TL, 4,106 g seatrout sp. eaten by a 251 cm male that stranded in May of 2001. This dolphin also consumed another sizeable, partially intact seatrout that (reconstructed) weighed 809 g and was 470 mm TL, and several other prey items. The smallest reconstructed fish was a 34 mm TL, 0.5 g hake sp. eaten by a 175 cm female that stranded in December of 2012. Typically, striped bass was the largest prey eaten (415 ± 144 mm; 931 ± 603 g), followed by seatrout spp. (304 ± 67 mm; 309 ± 252 g). Conversely, anchovy spp. was typically the smallest (74 ± 14 mm; 4 ± 2 g), followed by hake spp. (133 ± 26 mm; 29 ± 17 g). Unless otherwise specified, the results presented below will refer to total length in mm and weight in g.

TABLE 6 Results from variables tested to compare the diet of bottlenose dolphins stranded in Virginia between sexes, age groups (calf, juvenile, adult), and seasons (spring, summer, fall). DF is the degrees of freedom, NA is the proportion of numerical abundance, RM is the proportion of reconstructed mass. Bold rows are statistically significant. Variables tested Statistical test DF Test statistic p-value Sex NA MANOVA 11, 188 F=0.833 0.607

RM MANOVA 8, 186 F=0.758 0.641

Prey diversity Mann-Whitney U test 1 U=5,667.5 0.900

Prey quantity Mann-Whitney U test 1 U=5,716.0 0.072

Average prey length Independent-samples t-test 193 t=0.410 0.682

Average prey weight Independent-samples t-test 193 t=0.260 0.795 Age group NA MANOVA 22, 374 F=1.932 0.008

RM MANOVA 18, 368 F=1.955 0.011

Prey diversity Kruskal-Wallis H test 2 H=3.799 0.150

Prey quantity Kruskal-Wallis H test 2 H=12.560 0.002

Average prey length ANOVA 2, 192 F=7.450 0.001

Average prey weight ANOVA 2, 192 F=13.969 <0.001 Season NA MANOVA 22, 362 F=6.204 <0.001

RM MANOVA 18, 356 F=7.181 <0.001

Prey diversity Kruskal-Wallis H test 2 H= 0.808 0.668

Prey quantity Kruskal-Wallis H test 2 H=6.552 0.038

Average prey length ANOVA 2, 186 F=2.450 0.089 Average prey weight ANOVA 2, 186 F=3.127 0.046

Variation Between Sexes There was no significant difference in the diet of males versus females by proportion of numerical abundance, frequency of occurrence, proportion of reconstructed mass, diversity of taxa (females x = 4.4 ± 2.4, males x = 4.0 ± 2.8), quantity of prey items (females x = 57.5 ± 75.1, males x = 47.4 ± 60.7), overall average prey length (females x = 202 ± 66 mm, males x = 206 ± 68 mm), or overall average prey weight (females x = 176 ± 143 g, males x= 182 ± 171 g; Tables 6 & 7). There were six pregnant or lactating females in my sample set. Because other studies have found significant differences in the proportion of certain prey taxa in the diet for lactating females, I compared these animals to other adult females (n=21) and adult males (n=25), but did not find any significant difference in the proportion of numerical abundance of common prey (MANOVA F-

(18,82)=0.904, p=0.575).

TABLE 7 Proportion of numerical abundance (%NA), frequency of occurrence (%FO), and proportion of reconstructed mass (%RM) of prey in female and male bottlenose dolphins stranded in Virginia. "Other" category combines all species <2% numerical abundance. Prey in order of overall %NA. Females (n=94) Males (n=106) Prey %NA %FO %RM %NA %FO %RM Atlantic croaker 33 68 39 38 69 42 Spot 19 60 15 16 57 14 Seatrout spp. 15 57 26 14 46 20 Atlantic menhaden 5 26 5 6 30 10 Hake spp. 5 30 2 4 23 1 Anchovy spp. 4 30 <1 4 25 1 Longfin inshore squid 5 18 5 2 13 3 Silver perch 3 27 2 3 23 3 Striped bass 3 12 5 3 15 6 Unidentified fish 2 21 ― 4 18 ― Other 6 51 ― 7 39 ―

Variation Between Age Groups The proportion of numerical abundance (%NA) and reconstructed mass (%RM) were significantly different among age groups (Table 6, Figure 9). Follow-up univariate ANOVAs showed that the proportion of numerical abundance and reconstructed mass of spot was significantly different between age groups (Table 8). Tukey’s post-hoc tests revealed that spot comprised a significantly (p<0.05) higher proportion of calf diet numerically and by reconstructed mass than adult diet (Table 8, Figure 9). Overall, spot decreased in importance through ontogeny. The stomach contents of calves did not contain any striped bass or longfin inshore squid. Chi-square goodness-of-fit tests were conducted to determine if the frequency of each prey differed by age group (Table 8). Inshore squid and striped bass were not analyzed because their absence in calf diet violated assumptions. The tests indicated that Atlantic menhaden and spot were not equally represented in all age groups. Spot, similar to the proportion of numerical abundance was consumed with less frequency through ontogeny. Atlantic menhaden was consumed least frequently by calves but juveniles and adults consumed this species with similar frequency.

40 Calves (n= 20) Juveniles (n= 128) Adults (n= 52) 35 a 30

20 a,b

15 b 10

5 Proportion Proportion of AbundanceNumerical (%)

0 Croaker Spot Seatrout Menhaden Hake spp. Anchovy Longfin Silver Striped spp. spp. inshore Perch Bass squid FIGURE 9 Proportion of numerical abundance of prey in stomach contents of calf, juvenile, and adult bottlenose dolphins stranded in Virginia. Letters that are different indicate significant differences within the taxon at p<0.05. See Table 8 for statistical parameters. 31

TABLE 8 A) Proportion of numerical abundance (%NA), B) frequency of occurrence (%FO), and C) proportion of reconstructed mass (%RM) of prey in calf (n=20), juvenile (n=128), and adult (n=52) bottlenose dolphins stranded in Virginia. Bold rows are statistically significant using ANOVA or Chi-square analysis at p<0.05. Superscript letters that are different indicate significance in Tukey's post-hoc comparisons at p<0.05. A %NA

Prey Calves Juveniles Adults F(2,197) P Atlantic croaker 23 38 34 1.274 0.282 Spot 32a 17ab 10b 4.903 0.008 Seatrout spp. 14 13 20 1.487 0.229 Atlantic menhaden <1 5 8 2.859 0.060 Hake spp. 5 5 4 0.687 0.504 Anchovy spp. 10 4 3 1.547 0.216 Longfin inshore squid 0 4 4 2.414 0.092 Silver perch 6 3 3 0.694 0.501 Striped bass 0 2 5 2.281 0.105

B %RM

Prey Calves Juveniles Adults F(2,192) P Atlantic croaker 37 43 37 0.667 0.515 Spot 27a 14ab 9b 5.233 0.006 Seatrout spp. 24 20 28 1.168 0.313 Atlantic menhaden <1 8 9 2.796 0.064 Hake spp. 5 2 1 0.549 0.578 Anchovy spp. 3 1 <1 1.914 0.150 Longfin inshore squid 0 4 4 2.327 0.100 Silver perch 3 3 3 0.778 0.461 Striped bass 0 5 9 2.364 0.097

C %FO

Calves Juveniles Adults χ2(2) P Prey Atlantic croaker 70 72 60 1.228 0.054 Spot 70 54 38 9.481 0.009 Seatrout spp. 45 58 63 3.120 0.210 Atlantic menhaden 5 32 27 19.344 <0.001 Hake spp. 20 29 21 2.086 0.352 Anchovy spp. 40 27 23 5.267 0.072 Longfin inshore squid 0 19 13 ― ― Silver perch 35 23 23 3.556 0.169 Striped bass 0 13 19 ― ― 32

A one-way ANOVA revealed that reconstructed length and reconstructed mass of prey varied by dolphin age group (Table 6). The Tukey’s post-hoc tests revealed that the average reconstructed total length was significantly (p<0.05) smaller for calves (167 ± 61 mm) compared to adults (228 ± 75 mm) and smaller for juveniles (200 ± 193 mm) compared to adults. Also, the average reconstructed mass was significantly different (p<0.05) between all pairs with calves having prey that weighed the least and adults having prey that weighed the most (calf: 93 ± 83 g, juvenile: 165 ± 119g, adult: 247 ± 227 g). On average, calves ate prey that were smaller than those consumed by juveniles, and juveniles ate prey that were smaller than those consumed by adults. The variability in the size of each prey taxon by age group can be seen in Table 9. The average length of both silver perch and spot was significantly different among age groups. A Dunn’s pairwise comparison revealed silver perch was significantly shorter (p<0.05) in length in calves than adults and spot was significantly shorter in length in calves than either juveniles or adults. Furthermore, the average reconstructed mass of croaker, seatrout spp., silver perch, and spot was significantly different by age group. The pairwise comparison revealed croaker in calf stomach contents weighed significantly less than in juveniles or adults. The average weight of seatrout spp. was significantly greater in adults than in juveniles or calves. Similar to the average length, the average weight of silver perch was significantly greater in adults than in calves and spot weighed significantly less in calves than either juveniles or adults. The diversity of prey taxa consumed was not significantly different between age groups (Table 6). However, the number of prey items consumed was significantly different. The pairwise comparison revealed a significant difference (p<0.05) between the number of prey eaten by juveniles versus adults but not between juveniles and calves or calves and adults. On average, calves ate 46.1 ± 103.5 items, juveniles ate 57.6 ± 61.0 items and adults ate 41.1 ± 67.2 prey items. Furthermore,

the total mass was significantly different between age groups (ANOVA F(2, 192)=9.135, p<0.001). Tukey’s post hoc tests revealed calves contained significantly (p<0.05) less reconstructed mass than either juveniles or adults. On average, calves contained prey that reconstructed to 1,948 ± 3,250 g, juveniles contained 5,963 ± 5,371 g, and adults contained 5,831 ± 5,908 g.

TABLE 9 Reconstructed mean total length (A) and wet weight (B) of common prey in the stomach contents of calf (n=19), juvenile (n=126), and adult (n=50) bottlenose dolphins stranded in Virginia. Bold rows are statistically significant Kruskal-Wallis H tests at p<0.05. Superscript letters that are different indicate significance in pairwise comparison using Dunn’s procedure (individual taxa) or Tukey’s (“All prey”) at p<0.05. A Length x (mm)

Prey n Calves n Juveniles n Adults H(2) p

Striped bass 0 ― 17 378 ± 156 10 477 ± 100 H(1)=1.715 0.190 Seatrout spp. 9 282 ± 101 74 296 ± 55 33 330 ± 75 5.636 0.060 Atlantic menhaden 1 288 41 256 ± 43 14 267 ± 37 1.087 0.581 Atlantic croaker 14 184 ± 53 92 222 ± 60 31 225 ± 57 4.577 0.101

Longfin inshore squid 0 ― 24 173 ± 39 7 173 ± 30 H(1)=0.014 0.906 Spot 14 126 ± 35a 69 162 ± 37b 20 162 ± 42bc 11.055 0.004 Silver perch 7 132 ± 35a 30 158 ± 21ab 12 168 ± 17b 7.59 0.022 Hake spp. 4 142 ± 22 37 128 ± 24 11 145 ± 31 2.238 0.327 Anchovy spp. 8 68 ± 4 35 77 ± 17 12 70 ± 7 3.983 0.137 All prey 19 167 ± 61a 126 200 ± 193a 50 228 ± 75b See Table 6

B Weight x (g)

Prey n Calves n Juveniles n Adults H(2) p

Striped bass* 0 ― 17 772 ± 544 10 1,203 ± 628 H(1)= 1.342 0.247 Seatrout spp. 9 270 ± 119a 74 270 ± 130a 33 407 ± 415b 7.364 0.025 Atlantic menhaden 1 283 41 226 ± 99 14 242 ± 132 0.955 0.620 Atlantic croaker 14 97 ± 75a 92 175 ± 128b 31 186 ± 139b 7.225 0.027 Longfin inshore squid 0 ― 24 132 ± 56 7 128 ± 43 H(1)=0.002 0.962 Spot 14 34 ± 26a 69 76 ± 73b 20 74 ± 53bc 11.982 0.003 Silver perch 7 41 ± 27a 30 63 ± 24ab 12 72 ± 19b 7.633 0.022 Hake spp. 4 31 ± 14 37 26 ± 11 11 38 ± 29 3.387 0.184 Anchovy spp. 8 3 ± 1 35 4 ± 3 12 3 ± 1 4.675 0.097 All prey 19 93 ± 83a 126 165 ± 119b 50 247 ± 227c See Table 6

Variation Between Seasons There were only six dolphins that stranded in winter; therefore, I removed them from analyses for seasonal variation. MANOVAs on ranked data revealed that the proportion of numerical abundance (%NA) and the proportion of reconstructed mass (%RM) differed significantly in the three remaining seasons (Table 6). Follow-up univariate ANOVAs showed that both %NA 34

and %RM of anchovy spp., croaker, hake spp., menhaden, seatrout spp., spot, and striped bass were significantly different between seasons (Table 10). Due to similarities between %NA, and %RM (Table 10), only %NA will be presented below. Figure 10 illustrates the proportion of numerical abundance of these taxa for spring, summer, and fall. In spring, seatrout spp. was the dominant prey, comprising over a quarter of the diet; however, in summer and fall, croaker dominated, comprising almost half the diet, with seatrout decreasing to approximately 10% of the diet. Similar to croaker, spot reached its highest proportion of the diet in the summer and fall. Spot constituted 22% of the diet in the summer and fall but comprised only 7% in the spring. Menhaden made up 13% of the diet in the spring but declined to 2-3% of the diet from summer through fall. Striped bass varied from 1% or less in the summer and fall to 8% in the spring. Hake spp. was consistent at approximately 6% of the diet from spring through summer but dropped to less than 1% of the diet in the fall. While anchovy spp. was always a small proportion of the diet, it ranged from 1% in the fall to 6% in the summer and spring. All of the above variance was statistically significant. All the common species appeared in the diet throughout the year but varied in the amount from season to season, including winter, though winter was not statistically analyzed.

50 b b Spring (n= 60) Summer (n= 78) Fall (n= 56) 45 40 35 30 a 25 b b 20

15 a b a 10 a a b a a a 5 b Proportion of of Proportion AbundanceNumerical (%) b b c b b 0 b Croaker Spot Seatrout Menhaden Hake spp. Anchovy Longfin Silver Striped spp. spp. inshore Perch Bass squid FIGURE 10 Proportion of numerical abundance of prey in stomach contents of bottlenose dolphins stranded in Virginia in fall, spring, and summer. Letters that are different indicate significant differences using Tukey's post-hoc analysis at p<0.05 after ANOVA was significant at p<0.006. See Table 10 for statistical parameters. 35

TABLE 10 A) Proportion of numerical abundance (%NA), B) proportion of reconstructed mass (%RM), and C) frequency of occurrence (%FO) of prey of bottlenose dolphins stranded in Virginia in spring (n=59), summer (n=75), and fall (n=55). Bold rows are statistically significant using ANOVA or Chi-square analysis at p<0.006. Superscript letters that are different indicate significance in pairwise comparison using Tukey’s post- hoc analysis at p<0.05. %NA

A Prey Spring Summer Fall F(2,191) p Atlantic croaker 13a 48b 45b 24.525 <0.001 Spot 7a 21b 24b 11.226 <0.001 Seatrout spp. 25a 7b 12b 13.158 <0.001 Atlantic menhaden 13a 2b 3b 5.597 0.004 Hake spp. 8a 4b <1c 16.599 <0.001 Anchovy spp. 6a 6a <1b 4.575 0.011 Longfin inshore squid 5 3 1 0.561 0.572 Silver perch 4 3 3 1.928 0.148 Striped bass 8a <1b 1b 10.443 <0.001

%RM

B Prey Spring Summer Fall F(2,186) p Atlantic croaker 14a 58b 50b 32.437 <0.001 Spot 6a 16b 21b 11.035 <0.001 Seatrout spp. 38a 12b 17b 13.461 <0.001 Atlantic menhaden 16a 3b 4b 5.082 0.007 Hake spp. 3a 2b <1c 15.124 <0.001 Anchovy spp. 2a 1a <1b 4.548 0.012 Longfin inshore squid 5 4 <1 0.497 0.609 Silver perch 5 2 2 1.851 0.160 Striped bass 12a <1b 5b 9.735 <0.001

%FO

C Prey Spring Summer Fall χ2(2) p Atlantic croaker 40 85 77 17.119 <0.001 Spot 30 59 66 14.103 0.001 Seatrout spp. 73 44 59 7.170 0.028 Atlantic menhaden 40 22 21 8.265 0.016 Hake spp. 48 23 5 36.816 <0.001 Anchovy spp. 35 31 14 9.325 0.009 Longfin inshore squid 18 13 14 0.933 0.627 Silver perch 32 18 25 3.920 0.141 Striped bass 27 3 13 20.279 <0.001 36

The frequency of occurrence of all the common species except silver perch and longfin inshore squid varied significantly between seasons (Table 10). Croaker occurred less frequently in spring and more frequently in summer. Menhaden occurred most frequently in the spring. Seatrout spp. was consumed by almost three-quarters of the dolphins in spring but less frequently in summer and fall, with less than half of the dolphins consuming seatrout spp. in summer. Hake was infrequent in the fall and much more frequent in the spring. Spot was consumed with relatively consistent frequency from summer through fall compared to its reduced frequency in the spring. Anchovy spp. occurred half as frequently in dolphin stomach contents in the fall compared to spring and summer. Striped bass was consumed by very few dolphins in summer and fall but increased to 27% of the dolphins in spring. The number of prey items differed significantly by season, but prey diversity did not (Table 6). Dolphins ate significantly fewer prey items in the spring (x = 36.8 ±55.2) than in the summer (x = 68.8 ±84.0; pairwise Dunn’s procedure p=0.031) but prey item numbers in the fall (x = 43.2 ±48.6) were not significantly different from those in either spring or summer (pairwise Dunn’s procedure fall-spring p=0.575, fall-summer p=0.783). The average weight of fish per dolphin was highest in the spring (227 ± 224 g) and lowest in the summer (141 ± 100 g), while fish prey weights in the fall were intermediate (177 ± 121 g). These differences in prey weight were significant though average length was not significant (Table 6). Lastly, the total mass of food in the stomach was not statistically significant between seasons (ANOVA F(2,186)=0.532, p=0.588; spring x = 5,026 ± 650 g, summer x = 6,137 ± 728 g, fall x = 5,329 ± 629 g). Therefore, dolphins consumed fewer but larger prey in the spring than in the summer, resulting in similar stomach food masses in the two seasons.

DISCUSSION There were ontogenetic and seasonal differences in the diet of bottlenose dolphins stranded in Virginia from 1998-2012. Calf diet differed significantly from juvenile and adult diet. Of the nine most important prey species, striped bass and longfin inshore squid were absent in the calf diet. Striped bass was, on average, the largest-sized prey consumed by dolphins in this study (average total length of 415 mm ± 144 mm). Therefore, it is possible that striped bass were simply too large to catch and/or consume for calves. In addition to size, habitat may have played a role in the absence of striped bass in the calf diet. During the summer, striped bass are found in the open water, deep channels of Chesapeake Bay (Murdy, Birdsong, & Musick, 1997), a region not frequented by calves, 37

as dolphins tend to reside in shallow coastal waters during early ontogeny (Barco et al., 1999; Mann, Connor, Barre, & Heithaus, 2000; Scott et al., 1990; Shane et al., 1986; Toth, Hohn, Able, & Gorgone, 2011; Wang, Payne, & Thayer, 1994). Furthermore, young dolphins (especially in the first three years of life) have relatively small oxygen stores in the blood and muscles and reduced physiological control of bradycardia, two significant adaptations to enhance breath-holding and diving capacity, thereby excluding prey found at deeper depths from their diet (Noren, Cuccurullo, & Williams, 2004; Noren, Williams, Pabst, Mclellan, & Dearolf, 2001; Noren, Lacave, Wells, & Williams, 2002). Longfin inshore squid were also absent from calf diet. Squid are fast, have a high turning capacity, demonstrate anti-predator behaviors, effectively sense predators via hair cells along a lateral line analogue, and don’t vocalize (Jastrebsky, Bartol, & Krueger, 2016; York & Bartol, 2014; York & Bartol, 2016; York, Bartol, & Krueger, 2016). Menhaden represent a smaller portion of the diet in calves than juveniles or adults. This species is known to form large schools that can dynamically change direction and speed. Given that young dolphins are relatively weak swimmers, are inexperienced in prey capture, and have limited diving abilities (Noren, Biedenbach, & Edwards, 2006; Noren et al., 2001, 2002, 2004), it is possible that calves simply cannot catch elusive squid or menhaden effectively. Contrary to striped bass, longfin inshore squid, and menhaden, spot was important in the calf diet. Spot occurred more frequently in calves and comprised a significantly higher proportion of the diet, by number and reconstructed mass, relative to adults. Although not significant, spot appeared to comprise a higher proportion of the diet in calves relative to juveniles as well, resulting in a steady decline in importance in the diet through ontogeny. Compared to striped bass, spot are relatively small. The maximum reported total length of spot in Chesapeake Bay was 345 mm (Hildebrand & Schroeder, 1928). They are soniferous bottom-feeders that occur throughout Chesapeake Bay. Spot swim slower than many other estuarine fish species (Hettler, 1977), making them an easy prey target for calves that have begun to wean and transition to carnivorous predators. Since spot are also highly vocal, they may be easier to detect for young inexperienced calves, leading to higher catch success. Similar to trends detected in the present study, spot comprise a higher proportion of juvenile diet compared to adult diet in stranded or incidentally captured bottlenose dolphins in North Carolina (Gannon & Waples, 2004). However, although spot was found more frequently in immature bottlenose dolphins stranded in South Carolina, Pate and McFee (2012) found that spot comprise a higher numerical proportion of the diet in mature dolphins, which differs from the findings presented here. 38

Among the nine most common prey species combined, calves ate smaller prey than juveniles, which ate smaller prey than adults. These differences were statistically significant except for the difference in average prey length between calves and juveniles, though a clear trend in larger juvenile prey was evident. This ontogenetic shift could be due to smaller dolphins targeting smaller species (i.e. calves ate more spot and did not eat striped bass) and smaller dolphins consuming smaller specimens of the same species consumed by larger dolphins (i.e. the average seatrout spp. size was progressively larger from calves to juveniles to adults). Consistent with my data, Amir, Berggren, Ndaro, and Jiddawi (2005) found that three primary prey species for 26 Tursiops aduncus in Zanzibar, Tanzania were significantly smaller in immature dolphins. Similarly, pantropical spotted dolphins (Stenella attenuata) consumed prey of increasing size with increasing dolphin length (Robertson & Chivers, 1997). Although the hypothesis that smaller dolphins eat smaller prey seems intuitive, to my knowledge, this study is the first to provide significant empirical evidence for bottlenose dolphins in the western North Atlantic. Although there were significant differences in the size of prey eaten, the proportion of reconstructed mass attributed to each taxa was not significantly different among age groups except for spot. Thus, although smaller dolphins consumed smaller prey, that prey still made up similar proportions of their diet as older dolphins and was therefore, similarly important for calves, juveniles, and adults. Furthermore, on average, calves had a smaller amount of prey in their stomachs by total mass than either juveniles or adults. Juveniles and adults, on the other hand, were similar to each other. This result may relate to the ontogenetic growth pattern observed in the stomach of bottlenose dolphins. Although overall the stomach exhibits positive allometric growth, there is a significant difference in the allocation of mass of the stomach between juveniles (defined as 141-170 cm) and subadults (>170 cm and not sexually mature) but not between subadults and adults (sexually mature) (Mallette et al., 2016). This could mean the volume of the stomach cavity would not have differed enough to significantly affect the total mass of prey found in the stomachs of juveniles and adults in the present study. The significant difference in calves is likely due to calves supplementing their diet with milk, as well as having reduced physical stomach holding capacity relative to larger juveniles and adults. This reported average of reconstructed mass includes eroded otoliths, which may result in analysis of more than one meal. Therefore, these reported averages are not meant to reflect the expected meal capacity of calf, juvenile, or adult stomachs. A significant difference in prey composition between sexes was not found. Other than the use of nursery areas by females with calves, there is no evidence for resource or habitat partitioning 39

in bottlenose dolphins based on sex in Virginia (Barco et al., 1999) as in other areas or other species of delphinids, where pregnant or lactating females reside in different habitats and/or consume different prey (Barros & Odell, 1990; Gannon & Waples, 2004; Rossman et al., 2015; Scott et al., 1990; Worthy, 2008). There were only four lactating females considered in the dataset. Unfortunately, this sample size is too low to assess whether female use of nursery areas influences diet. Seasonal variation in diet was observed, with differences likely being due to changes in prey availability and detectability rather than changes in prey preference. Chesapeake Bay and its tributaries serve as an important seasonal spawning, nursery, and feeding grounds for croaker, spot, and seatrout spp.; however, these species are rare in winter as they leave to spawn or simply to escape low temperatures (Hildebrand & Schroeder, 1928; Jung & Houde, 2003; Murdey et al., 1997). In the spring, dolphins consumed significantly fewer prey items but the prey items they consumed weighed significantly more than those in summer or fall. The total mass in the stomachs was not significantly different among seasons. Together these results suggest that dolphins reach similar stomach mass levels irrespective of season but reach these levels differently in the spring compared with the summer and fall. Seatrout spp. was one of the largest prey, next to striped bass, in the diet. Because seatrout spp. was the dominant prey in spring, this taxon largely drove the observed seasonal differences of prey size described above. Increased consumption of seatrout in the spring correlates with the seasonal movement and spawning patterns of seatrout, which migrate into the Chesapeake Bay in the spring to spawn (Atlantic States Marine Fisheries Commission, 2016; Brown, 1981; Hildebrand & Schroeder, 1928; Jung & Houde, 2003; Murdet et al., 1996). Male seatrouts are known to increase their vocalizations during the peak courting and spawning period in the spring, making them easier to detect (Connaughton, Fine, & Taylor, 1997; Connaughton & Taylor, 1995, 1996; Montie et al., 2017). Menhaden exhibit two spawning seasons in the Chesapeake Bay, one in the spring, and a larger one in the fall (Hildebrand & Schroeder, 1928; Lewis, Ahrenholz, & Epperly, 1987; Murdy et al., 1996). Considering these are not soniferous fishes, it was not expected that increased consumption would correlate with the spawning seasons. The increased consumption of menhaden in the spring is likely due to the arrival of menhaden to the area during their migration before the move offshore to spawn in the fall (Dryfoos, Cheek, & Kroger, 1973; Hildebrand & Schroeder, 1928; Jung & Houde, 2003; Murdy et al., 1996; SEDAR, 2020). Although in many soniferous fishes only the male vocalizes, predominantly during the spawning and courting season though they can also produce distress calls, Atlantic croaker is one of few species in which both 40

male and female produce sounds that can be heard year-round (Connaughton, Lunn, Fine, & Taylor, 2003; Fine, Schrinel, & Cameron, 2004; Fish & Mowbray, 1970; Gannon, 2007). Therefore, unlike the seatrouts, increased vocalization is not likely the reason for increased consumption of croaker in summer and fall. However, availability may better explain this trend. Croaker in Chesapeake Bay generally migrate up-river and up-bay in the spring (Chao & Musick, 1977; Haven, 1957) while dolphins are present in the lower Bay and coastal Virginia waters (Barco et al., 1999). Croaker are more available to dolphins in summer and fall as they begin to move around the Bay and then down-river and down bay (Haven, 1957). The dataset included only six dolphins that stranded in the winter, and these dolphins were removed from statistical analyses because of their low numbers. Although this study focuses on spring, summer, and fall, it is notable that the “other” category, which includes all prey that comprises 2% or less of the overall numerical abundance, peaked in the winter at 20% of the diet compared to 4-8% in the other seasons. Furthermore, these six winter dolphins exhibited a trend in higher diversity of prey taxa per stomach (x = 6.0 ±3.2) compared to the other seasons (spring x = 4.2 ±2.1, summer x = 4.0 ±2.7, fall x = 4.3 ±2.8). The observed high percentage of the “other” prey category and trend in higher diversity of prey taxa certainly suggest that dolphins become more opportunistic in their feeding habitats as primary food sources (croaker, spot, and seatrout) become scarce in the winter. Bottlenose dolphins are also uncommon in Virginia waters in the winter (Barco et al., 1999). Prey availability is one of the most important factors that drive bottlenose dolphin movement (Shane et al., 1986), and thus it is likely that dolphins vacate the Chesapeake Bay as target prey leave. Physiological constraints may also be important. Barco et al. (1999) examined the relationship between water temperature and the movement of dolphins, and they found that a potential thermoneutral zone (TNZ) set point exists at approximately 16.0°C, i.e., to avoid elevated metabolic rates, dolphins leave waters as temperatures approach 16.0°C. However, bottlenose dolphins in other areas do not appear to exhibit similar TNZ limits (Wells, 1986; Wells et al., 1990). Clearly further investigation is warranted for a full assessment of dolphin diet patterns and physiology during the winter months. McGurk (1997) studied the stomach contents of bottlenose dolphins stranded in Virginia from 1987-1996. The primary species McGurk (1997) identified were croaker, weakfish, spot, silver perch, and squid (Doryteuthis sp.), which is consistent with the findings of the present study. There were some differences in primary prey importance between McGurk’s (1997) study and the present study. For example, silver perch was less important to the diet of dolphins in my study, but croaker 41

and spot were more important. Overall, there were a greater variety of species in my study. I found 32 species (range 1-15 per stomach) in the stomachs of the dolphins I examined compared to the 21 species (range 1-7 per stomach) in McGurk’s study. Of the 16 prey species that were identified in both studies, the frequency of occurrence increased for almost all of them in my study. This suggests that either my increased samples size was able to detect more variety in the diet, or that the diet of bottlenose dolphins has become more diverse over time. The occurrence of the family Sciaenidae has remained the same at 93%, with small variation at the species level; however, the family Clupeidae drastically increased in occurrence from 4% in McGurk’s study to 32% in the current study. Other notable increases include Phycidae (formerly Gadidae) from 15% to 26%, Moronidae from 1% to 14%, and Loliginidae from 19% to 24%. Future studies should investigate what is driving this increase in variety, such as changes in prey availability through increases or decreases in fish stocks, increased noise pollution affecting animal’s ability to rely on passive listening to detect prey, or other factors not listed here. Although dolphins fed on a variety of species, they fed primarily on three prey taxa: Atlantic croaker, spot, and seatrout spp. These taxa were found in more than half the dolphins and together represented almost 70% of the diet numerically. Of the nine species analyzed for reconstructed mass, these three species constituted 78% of the diet. Other taxa were rare, occurring in 28% or fewer of the dolphins and constituting 5% or less of the prey items. The soniferous family Sciaenidae, dominated the diet. This family occurred in 93% of the dolphins and accounted for 71% of the diet, which is consistent with findings in other studies along the U.S. east coast (e.g. Gannon & Waples, 2004; McGurk, 1997; Mead & Potter, 1990; Pate & McFee, 2012). Interestingly, in recent stock assessment reports (Atlantic States Marine Fisheries Commission, 2009, 2010), bottlenose dolphins were not listed as major predators of croaker or weakfish, though the 2016 assessment of weakfish mentioned marine mammals should be considered in future evaluations (Atlantic States Marine Fisheries Commission, 2016). As evident in the present and previous studies, dolphins are clearly major predators of these fishes and should be considered in stock assessment models. While other studies in the mid-Atlantic found shrimp to be a prey item of bottlenose dolphins (Gannon & Waples, 2004; Mead & Potter, 1990), this was not found to be the case in the present study. This is likely a reflection of lower shrimp populations in the Chesapeake Bay region, as evidenced by the lack of a prevalent shrimp fishery in Virginia.

Barros’ hypothesis that dolphins passively listen for their prey was inspired by the large proportion of sound producing fish in their diet, especially the family Sciaenidae. Passive listening is a good supplement to echolocation. Recent evidence suggests that the metabolic costs for echolocation is low (Noren, Holt, Dunkin, & Williams, 2017); therefore, ecological costs may better explain why dolphins would use passive listening over their sophisticated sonar system (see Gannon, 2003 for a review). Firstly, if the prey detects the sound of echolocation, it could alert them to the predator’s presence resulting in antipredator behaviors, including schooling, hiding, camouflaging, or moving out of the area. Secondly, other dolphins could detect the echolocation clicks and compete for the prey. And finally, the dolphins could alert their own predators to their presence. Echolocation is still important during the pursuit and capture of prey once it has been detected. For example, Gannon et al. (2005) found that dolphins turned towards a playback of fish sounds and increased their echolocation. Many soniferous fishes important to bottlenose dolphins, such as Atlantic croaker, weakfish, spotted seatrout, and spot predominantly produce vocalizations of 100- 600 Hz at sound pressure levels (SPL) of 114-140 dB re: 1 µPa, though some of their sound types, and other species, can produce acoustic signatures over 1000 Hz (Connaughton, Taylor, & Fine, 2000; Fine et al., 2004; Ramcharitar, Gannon, & Popper, 2006). These frequencies are close to the hearing limits of bottlenose dolphins (Finneran, Carder, & Ridgway, 2002; Finneran & Houser, 2006; Ljungblad, Scoggins, & Gilmartin, 1982). Therefore, if bottlenose dolphins are using passive listening to locate these prey, noise pollution may be a significant, long-term threat to bottlenose dolphins, as it can lead to masking effects, i.e., the inability to identify auditory sources as a result of background interference (Southall et al., 2019). Given that the fish vocalizations are near the hearing threshold of bottlenose dolphins, anthropogenic sound can have deleterious effects on detectability of prey. Luckily, especially during spawning, there are often multiple, chorusing fish producing these sounds which enhances the detectability. The Chesapeake Bay and surrounding coastal Virginia waters are subject to many anthropogenic noise sources. The largest U.S. military base in the world is in Norfolk, VA, and noisy vessels travel regularly in and out of the Chesapeake Bay. A major commercial shipping channel into Baltimore, MD runs through the length of the Chesapeake Bay. Recreational boating, commercial fishing vessels, pile driving, and naval sonar also contribute to elevated noise levels in the Chesapeake Bay and neighboring waters. As is the case in other waters, such as Peconic Bay Estuary system in Long Island, NY (Samuel, Morreale, Clark, Greene, & Richmond, 2005), ambient noise in the lower Chesapeake Bay and coastal oceanic Virginia waters increases in the summer, which 43

coincides with the peak of dolphin presence (Bort & Barco, 2014). Noise pollution must remain a focus of managers to monitor and mitigate. According to the optimal foraging theory, an animal’s feeding habits should maximize caloric intake while minimizing effort (MacArthur & Pianka, 1966). Masking the sounds produced by their primary prey (i.e., soniferous fishes) via increased anthropogenic sound could affect the foraging ability of bottlenose dolphins, possibly forcing them into quieter, sub-optimal foraging habitats or causing them to expend more energy searching for prey, reducing their fitness. Dolphins rely on echolocation while hunting. Increasing the amplitude of a vocalization in response to noise is known as the Lombard effect (Lombard, 1911). Modification of acoustic signals through the Lombard effect, change in frequency, or adjusting the timing or rate of the signal are common compensation mechanisms in animals, possibly with costly effects (Brumm & Slabbekoorn, 2005; Hotchkin & Parks, 2013). Both odontocete and mysticete whales have exhibited these compensation mechanisms (Fristrup, Hatch, & Clark, 2003; Holt, Noren, Veirs, Emmons, & Veirs, 2009; Parks, Johnson, Nowacek, & Tyack, 2011; Scheifele et al., 2005; Tennessen & Parks, 2016). Although numerous studies of the effects of anthropogenic noise on marine mammals have been performed (see Southall et al., 2019 for review), additional empirical studies are necessary to fully understand the effects of masking on prey detection in bottlenose dolphins and when compensation mechanisms are employed. Baseline data are important for understanding the effects of threats to a species because they serve as a relative metric for measuring change (Gatti et al., 2015; Thurstan et al., 2015). Thus, long-term natural history data are essential and help managers identify shifts in natural history parameters (i.e. diet, reproductive success, growth) in response to changes in the environment, either through natural or anthropogenic causes. This study not only provides valuable data on diet composition of dolphins in Virginia waters but also promises to serve as a measuring stick for monitoring change and implementing conservation and management efforts. Two potential sources of bias exist in this study: (1) the use of stomach contents for analysis and (2) stranding trends in Virginia. Different rates of decomposition can certainly introduce bias. For example, squid numerical abundance may have been overestimated because squid beaks can persist longer in the stomachs than otoliths (Bigg & Fawcett, 1985). Secondly, reduced strandings in the winter can impact overall findings as winter dolphins are insufficiently represented in these data. This is an unavoidable bias, the stranding numbers in this study are consistent with the seasonal stranding records (VAQS unpublished data) and trends observed by Barco et al. (1999) in the wild 44

population. Thus, the dataset examined here is most likely a representative snapshot of dolphin populations in the Virginia region, as dolphins generally move out of the area in the winter. In summary, bottlenose dolphins stranded in Virginia fed predominantly on soniferous fishes, spot comprised a higher proportion of calf diet than juvenile or adult diets, smaller dolphins fed on smaller prey, and there were significant seasonal differences in diet. Although previous studies provide important dietary information about bottlenose dolphins in coastal waters of the mid-Atlantic, this study, with a larger sample size (n=200) and a longer time period (15 years), provides a broader snapshot of stranded bottlenose dolphin diet in mid-Atlantic waters. Moreover, this study complements age and reproduction analyses performed by Lynott (2012), allowing for analyses of ontogenetic and seasonal effects, areas of study that have not been explored previously. 45

CHAPTER 4

ENTANGLED DOLPHINS

INTRODUCTION Human-induced mortality of marine animals is a complex, sometimes contentious issue between stakeholders globally. The Marine Mammal Protection Act (MMPA) of 1972 was established to prevent the decline of marine mammal stocks beyond the point of being a “significant functioning element in the ecosystem of which they are a part” and focuses on reducing human- induced direct and indirect mortalities. Incidental capture of marine mammals in commercial fisheries has been cited as a cause or major factor for the decline and/or has limited the population growth of some stocks of these long-lived, low fecundity top predators (Lewison, Crowder, Read, & Freeman, 2004; Read, Drinker, & Northridge, 2006; Reeves et al., 2013). NOAA National Marine Fisheries Service, the entity designated by the MMPA to manage the protection and conservation of cetaceans and pinnipeds, develops and implements take reduction plans to minimize bycatch of strategic marine mammal stocks. The western North Atlantic coastal bottlenose dolphin stock (Tursiops truncatus) was designated a strategic marine mammal stock in 2006 because fishery-related incidental mortality and serious injury exceeded the stock’s potential biological removal (PBR) and still is vulnerable to takes in multiple fisheries (Byrd & Hohn, 2017; Byrd et al., 2014; Hayes et al., 2019; Kovacs & Cox, 2014; Read et al., 2006; Reeves et al., 2013). There is a low observer rate in many of the commercial fisheries in the U.S., resulting in significant underreporting of incidental takes (Lewison et al., 2004; Moore et al., 2009). Investigating strandings can help to fill that gap through thorough investigation of the carcasses for signs of fishery interaction (Adimey et al., 2014; Byrd et al., 2014; Moore et al., 2013; Peltier et al., 2016). Gear is rarely present on beach-cast carcasses and there is no pathognomonic sign for peracute underwater entrapment, the dominant acute cause of death for an entangling interaction. There are, however, several indications that are consistent with this cause of mortality in small cetaceans, such as epidermal abrasions and lacerations that indicate ligature or constriction (Moore & Barco, 2013; Moore et al., 2013; Read & Murray, 2000), edematous, emphysematous and/or hyperinflated lungs (Jepson et al., 2000; Moore et al., 2013), fluid and/or stable froth in the airways (Jepson et al., 2000; Moore et al., 2013; Read & Murray, 2000), vascular congestion (Bernaldo de Quirós et al., 2018), intravascular gas bubbles (Bernaldo De Quiro et al., 2013), peri-mandibular or peri-scapular bloody edema or hemorrhage (Moore et al., 2013; Read & Murray, 2000), and intact 46

prey in the forestomach (Bernaldo de Quirós et al., 2018; M. J. Moore et al., 2013; Read & Murray, 2000). Unfortunately, some of these indications diminish quickly with decomposition, scavenger damage, or other sources, even within hours of death. The stomach contents, on the other hand, can be investigated late into decomposition and because they are contained deep within the body, are not exposed to scavengers as easily. Again, there is no pathognomonic sign, but combining several of these indicators can help to identify carcasses as victims of peracute underwater entrapment. Some stranding organizations do not have the resources to reliably investigate all of these signs and/or invest in histopathology services. Thus, identifying relatively easy and/or inexpensive indicators can be important for reporting suspect or confirmed fishery interactions to management. In one recent study of bottlenose dolphins on the US east coast, Pate and McFee (2012) attempted to correlate stomach “fullness” (via total wet weight) with evidence of human interaction in small cetaceans but found no significant relationship. McGurk (1997) compared the presence of 14 or more otoliths in the stomachs of stranded bottlenose dolphins in Virginia with and without evidence of human interaction and found no significant difference. Often, researchers in stomach content analysis studies, who attempt to compare bycaught and stranded small cetacean diet, concentrate on qualitative (species composition) rather than quantitative measures (e.g. Barros & Odell, 1990; Mead & Potter, 1990; Santos et al., 2007; Silva, 1999; Tellechea et al., 2017). Also, publications and guides sometimes suggest recently ingested prey can be a sign of acute death from human interaction (Bernaldo de Quirós et al., 2018; Moore & Barco, 2013; Moore et al., 2013; Read & Murray, 2000; Stockin et al., 2009). However, there currently is no statistically significant evidence to support this. My study aimed to quantify whether recently ingested prey in dolphins is statistically linked to external lesions consistent with entanglement. Specifically, I used fish otolith code as a proxy of how recently the prey was ingested and compared these codes in bottlenose dolphins with and without external evidence of entanglement. Fish dominate the prey of stranded dolphins in Virginia (Chapter 2). Therefore, I focused on fish otoliths rather than squid beaks, which are not coded on a scale as detailed as otoliths due to their indigestibility and thus are less informative. However, squid and squid beaks were used in other aspects of this study not related to digestion code. If acute death due to entanglement is indeed related to fresh prey, I expected the proportion of numerical abundance of early-code otoliths (i.e. fresh fish) to be significantly higher in dolphins with indications of entanglement than those that lack entanglement cues.

MATERIALS AND METHODS Stomach contents from 200 bottlenose dolphins stranded from 1998-2012 in Virginia were examined for this study (see Chapter 3). These samples had previously been collected during necropsy and stored frozen in the Virginia Aquarium Stranding Response Program (VAQS) archive. During the external and internal examination, the stranded dolphin was examined for evidence of human interaction (HI) such as entanglement, vessel interaction, gunshot, ingestion of anthropogenic debris, mutilation, or other anthropogenic sources (Read & Murray, 2000). For the purpose of this study, I looked for indications (i.e. written notes or digital necropsy reports) by the original examiner and/or I reviewed archived photographs specifically for external evidence of entanglement, which includes fresh, healing, or healed lesions consistent with ligature or constriction and/or entangling gear present. Each case was then assigned a category of yes, no, or could not be determined (CBD). Category “Yes”, indicated that the examiner observed lesions whose origins are consistent with entanglement; category “No” indicated that the animal did not exhibit any lesions consistent with entanglement; and category “CBD” indicated that the animal was too decomposed for accurate assessment or lesions observed could not be ruled in or out as deriving from an entanglement. During this project, the stomach contents were analyzed following methods adapted from Mead and Potter (1990), Barros and Wells (1998), and Gannon and Waples (2004); see Chapter 3 for further detail. All parts were scored as “trace” or “non-trace”, with trace items being well digested (i.e. disarticulated skeletal elements, free otoliths, squid beaks) and non-trace items being relatively undigested, representing a recently consumed meal (i.e. whole specimens, fish heads with attached flesh, cephalopod mantles). Non-trace items were further scored as follows: (a) “complete”, where the prey was whole with its skin mostly intact; (b) “whole”, where the food item was intact enough for a standard length measurement (i.e. skull to hypural bone in fish, whole mantle in squid) but skin and/or flesh was partially missing; (c) “partial”, where the skull was attached to some portion of vertebral column in fish or the mantle was present but incomplete in squid; and (d) “skull only”, where only the skull was present (only applicable to fish). Non-trace elements in categories (a) and (b) were measured as follows: standard length and/or total length (fish) or mantle length and/or pen length (squid) to the nearest 0.5 mm using a fish board. Sagittal otoliths were scored 0 (undamaged otoliths retrieved from (a)-(d) described above) to 5 (severely degraded, free otoliths) following the methods of Recchia and Read (1989). Squid beaks are made of chitin and therefore resistant to digestion in the marine mammal stomach; thus, they cannot be scored and were excluded from any 48

analyses involving degradation scoring. I measured the sagittal otoliths and estimated fish prey size following Gannon and Waples (2004). Regression equations were not available for all species, so only prey that accounted for 2% or more of the diet were analyzed via reconstructed mass or length. Refer to Table 5 in Chapter 3 for prey items less than 2% of the numerical abundance. Fish from the in-house reference collection were used to construct regression equations relating otolith length and width to fish length and wet weight for the eight most common bottlenose dolphin fish prey species (see Chapter 2). Equations for estimating the dorsal mantle length and wet weight of longfin inshore squid (Doryteuthis pealeii) from the lower rostral length were derived from Lange and Johnson (1981) and Staudinger et al. (2009). The proportion of numerical abundance is the count of a prey taxa in a particular stomach, divided by the total number of prey items in that stomach, then averaged across all stomachs.

Statistical Analyses Statistical analyses were completed using SPSS for Windows (SPSS, Inc.; Version 26.0). The following tests were performed to examine differences between the three levels of entanglement groups: Yes, No, and CBD. A one-way MANOVA on ranked data was performed to compare the proportion of numerical abundance of Code 0-5 otoliths, with follow-up ANOVAs for significant variables. Kruskal-Wallis H tests were performed to compare average prey length, average prey weight, total reconstructed weight, and species diversity per dolphin. A one-way ANOVA on log transformed data was performed to compare the quantity of prey. To explore biases in the samples set, I performed a one-way ANOVA on inverse transformed data to compare the standard length (straight length from tip of rostrum to fluke notch) of dolphins and chi-square tests of independence to compare sex (male and female) and age group. Age groups were defined according to age, growth, and sexually dimorphic parameters in Lynott (2012), Mead and Potter (1990), and Stolen et al. (2002) as follows: 1) calves were sexually immature and 1.5 years or younger (or <170 cm body length when age data were unavailable); 2) juveniles were sexually immature and 1.6 - 10 years old (or 171 - 250 cm) for males or 1.6 - 8 years old (or 171 - 230 cm) for females; and 3) adults were sexually mature and >10 years old (or >250 cm) for males or >8 years (or >230 cm) for females. Pairwise comparisons in the Kruskal-Wallis H tests were from a Dunn’s procedure. MANOVA and ANOVA pairwise comparisons were performed using Tukey-Kramer post-hoc tests. When the mean is presented, standard deviation (SD) is given as the measure of variance using the convention x ± SD.

RESULTS A total of 200 stranded bottlenose dolphins Tursiops truncatus were considered for this study (see Chapter 2 for the demographic breakup of these dolphins as well as the description of diet including differences between age groups, seasons, and sex). At least 10,426 prey items representing 32 species from 22 families were identified from the 20,929 otoliths (19,692 sagittae, 715 lapilli, 376 asterisci, 146 unknown) and 344 squid beaks found in the stomachs of the 200 dolphins analyzed. Of the 200 dolphins examined, 193 contained otoliths; the remaining seven contained only squid beaks or fish bones other than otoliths and were excluded from analyses involving the coded otoliths. All 19,692 sagittae were Coded 0-5 and the proportion of numerical abundance (%NA) was calculated per dolphin. The %NA of the otoliths codes varied significantly among the entanglement categories (Table 11). Follow-up univariate one-way ANOVAs were evaluated at p<0.05 and revealed that the %NA of Code 0, 1, 4 and 5 varied significantly between the entanglement groups (Table 12). The following are statistically significant by Tukey-Kramer post hoc tests at p<0.05. Code 0 and 1 otoliths comprised a higher proportion while Code 4 and Code 5 otoliths comprised a lower proportion in dolphins that scored Yes for signs of entanglement than No or CBD. Entanglement-No and CBD dolphins contained a similar proportion of Code 1, 4 and 5 otoliths; however, Code 0 otoliths comprised a significantly higher portion of Entanglement-CBD dolphins than Entanglement-No dolphin (Table 12).

TABLE 11 Results from variables tested to compare the diet of bottlenose dolphins stranded in Virginia between entanglement categories (yes, no, and could not be determined). DF is the degrees of freedom, NA is the proportion of numerical abundance. Bold rows indicate statistical significance. Variables tested Statistical test DF Test statistic p-value NA otolith code MANOVA 12, 372 F=4.123 <0.001 Prey diversity Kruskal-Wallis H test 2 H=1.547 0.461 Prey quantity ANOVA 2, 192 F=3.210 0.043 Average reconst. prey length Kruskal-Wallis H test 2 H=1.198 0.549 Average reconst. prey weight Kruskal-Wallis H test 2 H=1.449 0.485 Total reconstructed weight Kruskal-Wallis H test 2 H=4.277 0.118 Dolphin length ANOVA 2, 192 F=13.632 <0.001 Dolphin age group Chi-square of independence 4 χ2=23.892 <0.001 Dolphin sex Chi-square of independence 2 χ2=0.722 0.697

TABLE 12 Proportion of numerical abundance (%NA) of coded otoliths in bottlenose dolphins stranded in Virginia with (Yes), without (No), or could not be determined for (CBD) signs of entanglement. Bold rows are statistically significant using follow-up ANOVAs at p<0.05. Superscript letters that are different indicate significance in pairwise comparison using Tukey’s post- hoc analysis at p<0.05. %NA CBD Yes (n=96) No (n=24) (n=73) F(2,190) p Code 0 31a 8b 21c 13.669 <0.001 Code 1 7a 2b 5bc 8.789 <0.001 Code 2 20 23 16 1.614 0.202 Code 3 30 33 31 0.330 0.968 Code 4 8a 18b 16bc 4.514 0.012 Code 5 4a 16b 11bc 5.583 0.004

Prey diversity, average reconstructed prey length and weight, and total reconstructed weight were not statistically significant between the entanglement groups (Table 11). The number of prey items was statistically significant (Table 1; Ent-Yes: 57.1 ± 60.5; Ent-No: 45.0 ± 54.5; Ent-CBD: 51.3 ± 81.3; Table); however, the effect size was small (ω2= 0.022). The only post-hoc test that was significant (p<0.05) was comparison of Ent-Yes to Ent-CBD dolphins. In order to explore biases in my sample set, I examined dolphin length, age group, and sex. Both length and age group were statistically significant, but sex was not (Table 11). Post-hoc tests on length revealed that Ent-Yes dolphins (208 ± 25 cm) were significantly (p<0.05) smaller than both Ent-No (231 ± 34 cm) and Ent-CBD (232 ± 33 cm), but Ent-No and Ent-CBD were similar in length. Similarly, the chi-square of independence on age groups showed a significant association (Table 11) with entanglement status that was moderately strong (Cramer’s V=0.284). There were fewer adults (adjusted residual -4.8) in Ent-Yes but more juveniles (adjusted residual 3.3) in Ent-Yes and adults (adjusted residual 3.7) in Ent-CBD than expected.

DISCUSSION

Bottlenose dolphins stranded in Virginia with evidence of entanglement exhibited a higher proportion of recently ingested fish in their stomach contents than dolphins that did not show signs of entanglement. Combined, Code 0 and 1 otoliths accounted for 38%, 10%, and 26% of the 51

otoliths in Ent-Yes, Ent-No, and Ent-CBD respectively. The stark contrast between Ent-Yes and Ent-No dolphins is notable because this is an indication that dolphins with evidence of entanglement had consumed a meal peri-mortem and have a higher percentage of recently consumed prey in their stomachs, whereas those without any evidence of entanglement had a lower percentage of recently consumed prey. The difference between Ent-Yes and Ent-CBD was not as large, possibly because there were some animals in this CBD category that originally died as a result of entanglement but where evidence of this interaction faded or was unclear because of scavenger damage or decomposition. The significantly higher proportion of Code 0 otoliths in CBD compared to Ent-No dolphins provides support for this hypothesis. Larger prey will digest slower than smaller prey, and therefore, the otoliths of large prey can be exposed to gastric juices later than in smaller prey, potentially biasing early code otoliths toward larger prey. In order to ensure that the significant difference in the proportion of early code otoliths in Ent-Yes dolphins was not due to a bias of Ent-Yes dolphins eating larger prey than Ent-No and Ent-CBD dolphins, I investigated the average prey length and weight among these categories. The Kruskal-Wallis H test revealed that average prey lengths and weights were not significantly different between Ent-Yes, Ent-No, and Ent-CBD dolphins, suggesting this bias did not exist (Table 11). To investigate biases in my sample set, I compared dolphin length, age group, and sex among the entanglement groups. There was no difference in the distribution of sex among the entanglement groups (Table 11). However, smaller/younger dolphins represented a larger portion of entangled dolphins than non-entangled dolphins (Table 11). While not the focus of this study, this could indicate that younger dolphins have a higher risk of entanglement (i.e. due to naivety, lower skill level to avoid entanglement, smaller body size and strength limiting their ability to break away from net, etc.) as has been found in other studies (Kirkwood et al., 1997; McFee & Lipscomb, 2009). Although results of Chapter 3 indicate that smaller/younger dolphins feed more on spot (Leiostomus xanthurus), a relatively small fish, and the average length and weight of prey were smaller in these size/age groups, this bias did not translate to a significant difference in the prey size among entanglement groups. I investigated the total reconstructed weight of prey per dolphin and did not find a significant difference between the entanglement categories. Pate and McFee (2012) and McGurk (1997) attempted to compare “fullness” (via qualitative and quantitative measures) of the stomachs of bottlenose dolphins that died due to anthropogenic causes and did not find a significant difference. These results collectively suggest that “fullness” is not a reliable indicator but possibly the 52

“freshness” of the prey could be. A dolphin may forage near or depredate from a net and subsequently become entangled. This dolphin either could have been foraging for a length of time resulting in a stomach full of prey prior to the entangling event or, the dolphin could have fed on one fish and immediately became entangled. In both cases, the dolphin would have predominantly fresh prey in its stomach, one would be full and one would only have one fish. Looking at the proportion of fresh prey in the stomach therefore would capture both cases, whereas the “fullness” of the stomach would not. The presence of early code otoliths in the stomach of stranded dolphins does not ensure that the individual’s cause of death is related to peracute underwater entrapment; it should be combined with other available evidence, such as epidermal lesions, peri-mandibular or peri-scapular bloody edema or hemorrhage, and/or pulmonary and cardiovascular lesions listed in Moore et al. (2013) and Bernaldo de Quirós et al. (2018). However, the presence of early code otoliths in the stomachs of a stranded dolphin can be an important metric in determining human interactions when other cues are less apparent. To my knowledge, this is the only study that has quantitatively determined, using otolith code as a proxy, that entangled dolphins have more recently ingested prey in their stomach than non-entangled dolphins.

CHAPTER 5

CONCLUSION

Prior to this study, knowledge of bottlenose dolphin (Tursiops truncatus) diet in Chesapeake Bay and Virginia waters was based on studies with low samples sizes that (1) did not incorporate reconstructed size of prey, (2) did not consider ontogenetic or sex differences and/or (3) lacked quantitative analysis of dolphins with evidence of human interaction or entanglement (e.g. Mead and Potter 1990, McGurk 1997). In this study, I described the diet of bottlenose dolphin in relation to factors like sex, ontogeny, and season using the proportion of numerical prey abundance and reconstructed mass, the frequency of prey occurrence, prey diversity and quantity, and average reconstructed prey size. To provide accurate measures of the proportion of reconstructed prey mass and the average reconstructed prey size, I developed regression equations relating otolith length and width to fish length and weight for the eight teleost taxa that constitute at least 2% of the diet of bottlenose dolphins (Tursiops truncatus) stranded in Virginia. Reconstructed size for longfin inshore squid (Doryteuthis pealeii), the only squid species to constitute more than 2% of the diet, were based on equations already established by Lange and Johnson (1981) and Staudinger et al. (2009). While otolith length/width to total length followed the typical linear relation for most taxa, three taxa exhibited a better fit when a logarithmic curve was applied to otolith width vs. total length data. Seatrout spp. (Cynoscion spp.), spot (Leiostomus xanthurus), and striped bass (Morone saxatilis) exhibited the most deviation from linearity because of greater disproportionate otolith growth (these species have relatively elongated otoliths as they grow) through ontogeny, and thus non-linear curves work best for their total length predictions based on otolith width measurements. The otolith size to wet weight of all eight taxa were best fit to a log curve that is similar in form to what is typically used in fisheries for length-weight relationships: W=aLb where W is body weight, L is length, a is a coefficient related to body form and b is an exponent indicating isometric growth when equal to 3 (e.g. Aurioles-Gamboa, 1991; Harvey et al., 2000; Hunt, 1992; Sinclair et al., 2015). Many cetacean diet studies use a two-step process when relating otolith size to fish weight (e.g. Barros, 1993; Bowen, 2011; Gannon et al., 1997, 1998; Gannon & Waples, 2004; Pate, 2008; Recchia & Read, 1989). First, the total length is calculated from the otolith length. Second, wet weight is calculated from total length. In many cases, different datasets are used for the two steps. In my study, I reduced compounding errors associated with the conventional two-step process because I derived direct 54

relationships between otolith size and fish wet weight. These equations can be applied to diet studies of other local predators; however, it should be noted that the dataset is limited to the fish size ranges of those consumed by bottlenose dolphins in Virginia waters. Therefore, the equations should not be used to predict sizes beyond the data range considered in this study. Soniferous fishes dominated the diet of bottlenose dolphins stranded in Virginia. Although dolphins fed on a variety of species, they fed primarily on three prey taxa: Atlantic croaker (Micropogonias undulatus), spot, and seatrout spp. These taxa were found in more than half the dolphins and together represented almost 70% of the diet numerically. Of the nine species analyzed for reconstructed mass, these three species constituted 78% of the diet. There were ontogenetic and seasonal differences, though no differences between the sexes, in the diet of bottlenose dolphins in this study. Calf diets differed significantly from juvenile and adult diets. Spot declined in importance through ontogeny, whereas striped bass and longfin inshore squid were absent in calf diet. This result can be explained by the ease of capture and availability of these species. Spot is a relatively small, slow, vocal estuarine fish; however, striped bass are relatively large and occur in deeper open waters not frequented by calves, while longfin inshore squid are fast with a high turning capacity and exhibit other behavioral and sensory adaptations that make them elusive to inexperienced predators. Furthermore, the average size of prey overall increased through ontogeny. The total mass consumed was not significantly different between juveniles and adults. Although the stomach exhibits positive allometric growth, that growth begins to taper with age, making stomach volume of juveniles and adults more similar than between calves and juveniles. Seasonal variation in the diet was observed, with differences likely being due to changes in prey availability and detectability (due to seasonal migrations and vocalization during spawning) rather than changes in prey preference. Bottlenose dolphins stranded in Virginia with evidence of entanglement (Ent-Yes) exhibited a higher proportion of less degraded otoliths in their stomach contents than dolphins that did not show signs of entanglement (Ent-No and Ent-CBD). This indicates that dolphins with evidence of entanglement had consumed a meal peri-mortem and have higher proportion of recently consumed prey, whereas those without any evidence of entanglement have a lower proportion of recently consumed prey. While other researchers attempted to compare “fullness” of the stomach of animals that died due to anthropogenic causes with those that did not, no significant difference was found (McGurk, 1997; Pate & McFee, 2012). Consequently, “fullness” is probably not a reliable indicator of human-induced mortality, while “freshness” (i.e. use of less degraded otoliths in my results) of the prey is. In order to ensure that the significant difference in the proportion of early code otoliths in 55

Ent-Yes dolphins was not due to a bias of Ent-Yes dolphins eating prey that digests slower than Ent-No and Ent-CBD dolphins, I investigated the average prey length and weight among these categories. The results indicated this bias did not exist because there was no significant difference in prey length or weight. Smaller/younger dolphins represented a larger portion of entangled dolphins than non-entangled dolphins in this study. This result could indicate that younger dolphins have a higher risk of entanglement (i.e. due to naivety, lower skill level to avoid entanglement, smaller body size and strength limiting their ability to break away from net, etc.). While there is no pathognomonic sign for peracute underwater entrapment, there have been several proposed indicators, including the presence of recently ingested prey. However, to my knowledge, this has never been quantitatively assessed until this study. The presence of early code otoliths in the stomach of stranded dolphins does not ensure that the individual’s cause of death is related to entanglement; it should be combined with other available evidence such as those proposed in Moore and Barco (2013), Moore et al. (2013), and Bernaldo de Quirós et al. (2018). However, the presence of early code otoliths in the stomach of a stranded dolphin can be an important metric in determining human interactions when other cues are less apparent and/or more susceptible to decomposition. In combination with a study by Lynott (2012), we now have a good understanding of the life history and ecology of stranded dolphins in Virginia from the late 1990s to 2012. From July 2013 to March 2015, an Unusual Mortality Event occurred in which over 1,600 bottlenose dolphins stranded along the U.S. Atlantic coast due to a cetacean morbillivirus outbreak (NOAA, 2019). Over 25% of those strandings occurred in Virginia. Previously, a similar die-off occurred from 1987 to 1988 when over 50% of the coastal migratory stock died as a result of cetacean morbillivirus (Scott et al., 1988), leading the National Marine Fisheries Service (NMFS) to officially list the coastal migratory stock as depleted under the Marine Mammal Protection Act (Scott et al., 1988; Wang et al., 1994). The epicenter of the epidemic was again, in the mid-Atlantic (Scott et al., 1988). These sudden, local mortality events reinforce the importance of having comprehensive life history data to provide historical perspective and ecological insights. My study together with Lynott (2012) provide geographically relevant data and form an important baseline dataset that can be referenced in future studies to identify shifts in reproduction or diet in the years following extreme die-off events as the population recovers. 56

REFERENCES

Adimey, N. M., Hudak, C. A., Powell, J. R., Bassos-Hull, K., Foley, A., Farmer, N. A., … Minch, K. (2014). Fishery gear interactions from stranded bottlenose dolphins, Florida manatees and sea turtles in Florida, U.S.A. Marine Pollution Bulletin, 81, 103–115. Amir, O. A., Berggren, P., Ndaro, S. G. M., & Jiddawi, N. S. (2005). Feeding ecology of the Indo- Pacific bottlenose dolphin (Tursiops aduncus) incidentally caught in the gillnet fisheries off Zanzibar, Tanzania. Estuarine, Coastal and Shelf Science, 63, 429–437. Atlantic States Marine Fisheries Commission. (2009). Weakfish Stock Assessment Report (Atlantic States Marine Fisheries Commission, Stock Assessment Report). 426 pp. Atlantic States Marine Fisheries Commission. (2010). Atlantic Croaker 2010 Benchmark Stock Assessment (Atlantic States Marine Fisheries Commission, Stock Assessment Report). 366 pp. Atlantic States Marine Fisheries Commission. (2016). Weakfish Benchmark Stock Assessment and Peer Review Report (Atlantic States Marine Fisheries Commission, Stock Assessment Report). 270 pp. Aurioles-Gamboa, D. (1991). Otolith size versus weight and body-length relationships for eleven fish species of Baja California, Mexico. Fishery Bulletin, 89, 701–706. Barco, S. G., Swingle, W. M., McLellan, W. A., Harris, R. N., & Pabst, D. A. (1999). Local abundance and distribution of bottlenose dolphins (Tursiops truncatus) in the nearshore waters of Virginia beach, Virginia. Marine Mammal Science, 15, 394–408. Baremore, I. E., & Bethea, D. M. (2010). A guide to otoliths from fishes of the Gulf of Mexico (NOAA Technical Memorandum NMFS-SEFSC-599). Panama City, FL: National Marine Fisheries Service. 102 pp. Barros, N. B. 1993. Feeding ecology and foraging strategies of bottlenose dolphins on the central east coast of Florida (Ph.D. dissertation). University of Miami, Coral Gables, FL. 328 pp Barros, N. B., & Myrberg, A. A. (1987). Prey detection by means of passive listening in bottlenose dolphins (Tursiops truncatus). The Journal of the Acoustical Society of America, 82, S65–S65. Barros, N. B., & Odell, D. K. (1990). Food habits of bottlenose dolphins in the southeastern United States. In S. Leatherwood & R. R. Reeves (Eds.), The bottlenose dolphin (pp. 309–328). San Diego, CA: Academic Press. Barros, N. B., & Wells, R. S. (1998). Prey and feeding patterns of resident bottlenose dolphins (Tursiops truncatus) in Sarasota Bay, Florida. Journal of Mammalogy, 79, 1045–1059. Berens McCabe, E. J., Gannon, D. P., Barros, N. B., & Wells, R. S. (2010). Prey selection by resident common bottlenose dolphins (Tursiops truncatus) in Sarasota Bay, Florida. Marine Biology, 157, 931–942. Bernaldo De Quiro, Y., Seewald, J. S., Sylva, S. P., Greer, B., Niemeyer, M., Bogomolni, A. L., & Moore, M. J. (2013). Compositional discrimination of decompression and decomposition gas bubbles in bycaught seals and dolphins. PLoS ONE, 8, 1–12. Bernaldo de Quirós, Y., Hartwick, M., Rotstein, D. S., Garner, M. M., Bogomolni, A., Greer, W., … Moore, M. (2018). Discrimination between bycatch and other causes of cetacean and pinniped stranding. Diseases of Aquatic Organisms, 127, 83–95. Bernard, H. J., & Hohn, A. A. (1989). Differences in feeding habits between pregnant and lactating spotted dolphins (Stenella attenuata). Journal of Mammalogy, 70, 211–215. Bigg, M. A., & Fawcett, I. (1985). Two biases in diet determination of northern fur seals (Callorhinus ursinus). In J. R. Beddington, R. J. H. Beverton, & D. M. Lavigne (Eds.), Marine Mammals and Fisheries (pp. 284–291). London: George Allen & Unwin Ltd. Bort, J., & Barco, S. (2014). Baseline Ambient Noise Measurements in Chesapeake Bay and Surrounding Waters: Important Habitat for Acoustically Oriented Species. (Final Report to Virginia Department of 57

Game and Inland Fisheries Contract #2013-14025). VAQF Scientific Report 2014-06, Virginia Beach, VA, 33 pp. Bossart, G. D. (2011). Marine mammals as sentinel species for oceans and human health. Veterinary Pathology, 48, 676–690. Bowen, S. R. (2011). Diet of bottlenose dolphins Tursiops truncatus in the northwest Florida Panhandle and foraging behavior near Savannah, Georgia (M.S. thesis). Savannah State University, Savannah, GA. 147 pp. Bowen, W. (1997). Role of marine mammals in aquatic ecosystems. Marine Ecology Progress Series, 158, 267–274. Brown, N. J. (1981). Reproductive biology and recreational fisher for spotted seatrout, Cynoscion nebulosus, in the Chesapeake Bay area (M.A. thesis). College of William and Mary, Williamsburg, VA. 161 pp. Brumm, H., & Slabbekoorn, H. (2005). Acoustic communication in noise. Advances in the Study of Behavior, 35, 151–209. Byrd, B. L., & Hohn, A. A. (2017). Differential risk of bottlenose dolphin (Tursiops truncatus) bycatch in North Carolina, USA. Aquatic Mammals, 43, 558–569. Byrd, B. L., Hohn, A. A., Lovewell, G. N., Altman, K. M., Barco, S. G., Friedlaender, A., … Thayer, V. G. (2014). Strandings as indicators of marine mammal biodiversity and human interactions off the coast of North Carolina. Fishery Bulletin, 112, 1–23. Campana, S. E. (2004). Photographic atlas of fish otoliths of the Northwest Atlantic Ocean. Ottawa, Ontario: NRC Research Press. Campana, S. E. (1990). How reliable are growth back-calculations based on otoliths? Canadian Journal of Fisheries and Aquatic Sciences, 47, 2219–2227. Campana, S. E, & Casselman, J. M. (1993). Stock discrimination using otolith shape analysis. Canadian Journal of Fisheries and Aquatic Sciences, 50, 1062–10836. Chao, L. N., & Musick, J. A. (1977). Life history, feeding habits, and functional morphology of juvenile sciaenid fishes in the York River Estuary, Virginia. Fishery Bulletin, 75, 657–702. Clarke, M. R. 1986. A Handbook for the Identification of Cephalopod Beaks. Clarendon Press, Oxford. 273 pp. Connaughton, M. A., Fine, M. L., & Taylor, M. H. (1997). The effects of seasonal hypertrophy and atrophy on fiber morphology, metabolic substrate concentration and sound characteristics of the weakfish sonic muscle. Journal of Experimental Biology, 200, 2449–2457. Connaughton, M. A., Lunn, M.L., Fine, M. L., & Taylor, M. H. (2003). Characterization of sounds and their use in two sciaenid species: weakfish and Atlantic croaker. Proceedings from the International Workshop on the Applications of Passive Acoustics to Fisheries, 15-19. Connaughton, M. A., Taylor, M. H., & Fine, M. L. (2000). Effects of fish size and temperature on weakfish disturbance calls: Implications for the mechanism of sound generation. Journal of Experimental Biology, 203, 1503–1512. Connaughton, M. A, & Taylor, M. H. (1995). Seasonal and daily cycles in sound production associated with spawning in weakfish, Cynoscion regalis. Environmental Biology of Fishes, 42, 233– 240. Connaughton, M. A, & Taylor, M. H. (1996). Drumming, courtship, and spawning behavior in captive weakfish, Cynoscion regalis. Copeia, 1996, 195–199. Dryfoos, R. L., Cheek, R. P., & Kroger, R. L. (1973). Preliminary analyses of Atlantic menhaden, Brevoortia tyrannus, migrations, population structure, survival and explotation rates, and availablility as indicated from tag returns. Fishery Bulletin, 71, 719–734. Duffield, D., Ridgway, S., & Cornell, L. (1983). Hematology distinguishes coastal and offshore forms of dolphins. Canadian Journal of Zoology, 61, 930–933. Fearnbach, H. (2004). Characterization of association patterns of coastal migratory bottlenose dolphins, Tursiops 58

truncatus, in the nearshore waters of Virginia Beach, Virginia (M.S. thesis). Old Dominion University, Norfolk, VA. 142 pp. Fine, M. L., Schrinel, J., & Cameron, T. M. (2004). The effect of loading on disturbance sounds of the Atlantic croaker Micropogonius undulatus: Air versus water . The Journal of the Acoustical Society of America, 116, 1271–1275. Finneran, J. J., Carder, D. A., & Ridgway, S. H. (2002). Low-frequency acoustic pressure, velocity, and intensity thresholds in a bottlenose dolphin (Tursiops truncatus) and white whale (Delphinapterus leucas). The Journal of the Acoustical Society of America, 111, 447–456. Finneran, J. J., & Houser, D. S. (2006). Comparison of in-air evoked potential and underwater behavioral hearing thresholds in four bottlenose dolphins (Tursiops truncatus). The Journal of the Acoustical Society of America, 119, 3181–3192. Fire, S. E., Wang, Z., Byrd, M., Whitehead, H. R., Paternoster, J., & Morton, S. L. (2011). Co- occurrence of multiple classes of harmful algal toxins in bottlenose dolphins (Tursiops truncatus) stranding during an unusual mortality event in Texas, USA. Harmful Algae, 10, 330–336. Fish, M. P. & Mowbray, W. H. (1970). Sounds of western North Atlantic fishes: a reference file of biological underwater sounds. Baltimore: Johns Hopkins Press. Fitch, J. E., & Brownell Jr., R. L. (1968). Fish otoliths in cetacean stomach and their importance in interpreting feeding habits. Journal of the Fisheries Research Board of Canada, 25, 2561–2574. Fristrup, K. M., Hatch, L. T., & Clark, C. W. (2003). Variation in humpback whale (Megaptera novaeangliae) song length in relation to low-frequency sound broadcasts. The Journal of the Acoustical Society of America, 113, 3411–3424. Gannon, D. P. (2003). Behavioral ecology of an acoustically mediated predator-prey system: bottlenose dolphins and sciaenid fishes (Ph.D. thesis). Duke University, Durham, NC. 244 pp. Gannon, D. P. (2007). Acoustic Behavior of Atlantic Croaker, Micropogonias undulatus (Sciaenidae). Copeia, 1, 193–204. Gannon, D. P., Barros, N. B., Nowacek, D. P., Read, A. J., Waples, D. M., & Wells, R. S. (2005). Prey detection by bottlenose dolphins, Tursiops truncatus: An experimental test of the passive listening hypothesis. Animal Behaviour, 69, 709–720. Gannon, D. P., Craddock, J. E., & Read, A. J. (1998). Autumn food habits of harbor porpoises, Phocoena phocoena, in the Gulf of Maine. Fishery Bulletin, 96, 428-437. Gannon, D. P., Read, A. J., Craddock, J. E., & Mead, J. G. (1997). Stomach contents of long-finned pilot whales (Globicephala melas) stranded on the U.S. mid-Atlantic coast. Marine Mammal Science, 13, 405–418. Gannon, D. P., & Waples, D. M. (2004). Diets of coastal bottlenose dolphins from the U.S. mid- Atlantic coast differ by habitat. Marine Mammal Science, 20(3), 527–545. Gatti, G., Bianchi, C. N., Parravicini, V., Rovere, A., Peirano, A., Montefalcone, M., … Morri, C. (2015). Ecological change, sliding baselines and the importance of historical data: Lessons from combing observational and quantitative data on a temperate reef over 70 years. PLoS ONE, 10. Granadeiro, J. P., & Silva, M. A. (2000). The use of otoliths and vertebrae in the identification and size-estimation of fish in predator-prey studies. Cybium, 24, 383–393. Harvey, J. T., Loughlin, T. R., Perez, M. a, & Oxman, D. S. (2000). Relationship between Fish Size and Otolith Length for 63 Species of Fishes from the Eastern North Pacific Ocean (NOAA Technical Report NMFS 150). Seattle, WA: National Marine Fisheries Service. 36 pp. Haven, D. S. (1957). Distribution, growth, and availability of juvenile croaker, Micropogon undulatus, in Virginia. Ecology, 38, 88–97. Hayes, S. A., Josephson, E., Maze-Foley, K., & Rosel, P. E. (2019). US Atlantic and Gulf of Mexico Marine Mammal Stock Assessments - 2018 (NOAA Technicaly Memorandum NMFS-NE-258). Woods Hole, MA: National Marine Fisheries Service. 512 pp. 59

Hettler, W. F. (1977). Swimming speeds of juvenile estuarine fish in a circular flume. Proceedings of the Annual Conference of the SE Association, Fish and Wildlife Agencies, 31, 392–398. Hildebrand, S. F., & Schroeder, W. C. (1928). Fishes of Chesapeake Bay (Bureau of Fisheries Document No. 1024). Bulletin of the United States Bureau of Fisheries Vol. 43, Part 1. 366 pp. Holt, M. M., Noren, D. P., Veirs, V., Emmons, C. K., & Veirs, S. (2009). Speaking up: Killer whales (Orcinus orca) increase their call amplitude in response to vessel noise. The Journal of the Acoustical Society of America, 125, 27–32. Hotchkin, C., & Parks, S. (2013). The Lombard effect and other noise-induced vocal modifications: Insight from mammalian communication systems. Biological Reviews, 88, 809–824. Hunt, J. J. (1992). Morphological characteristics of otoliths for selected fish in the northwest Atlantic. Journal of Northwest Atlantic Fishery Science, 13, 63–75. Irvine, A. B., Scott, M. D., Wells, R. S., & Kaufmann, J. H. (1981). Movements and activities of the Atlantic bottlenose dolphin, Tursiops truncatus, near Sarasota, Florida. Fishery Bulletin, 79, 671– 688. Jansen, O. E., Leopold, M. F., Meesters, E. H. W. G., & Smeenk, C. (2010). Are white-beaked dolphins Lagenorhynchus albirostris food specialists? Their diet in the southern North Sea. Journal of the Marine Biological Association of the United Kingdom, 90, 1501–1508. Jastrebsky, R. A., Bartol, I. K., & Krueger, P. S. (2016). Turning performance in squid and cuttlefish: unique dual-mode, muscular hydrostatic systems. The Journal of Experimental Biology, 219, 1317– 1326. Jennings, S., Smith, A. D. M., Fulton, E. A., & Smith, D. C. (2014). The ecosystem approach to fisheries: Management at the dynamic interface between biodiversity conservation and sustainable use. Annals of the New York Academy of Sciences, 1322, 48–60. Jepson, P. D., Baker, J. R., Kuiken, T., Simpson, V. R., Kennedy, S., & Bennett, P. M. (2000). Pulmonary pathology of harbour porpoises (Phocoena phocoena) stranded in England and Wales between 1990 and 1996. Veterinary Record, 146, 721–728. Jung, S., & Houde, E. D. (2003). Spatial and temporal variabilities of pelagic fish community structure and distribution in Chesapeake Bay, USA. Estuarine, Coastal and Shelf Science, 58, 335– 351. Kirkwood, J. K., Bennett, P. M., Jepson, P. D., Kuiken, T., Simpson, V. R., & Baker, J. R. (1997). Entanglement in fishing gear and other causes of death in cetaceans stranded -on the coasts of England and Wales. Veterinary Record, 141, 94–98. Kovacs, C., & Cox, T. (2014). Quantification of interactions between common bottlenose dolphins (Tursiops truncatus) and a commercial shrimp trawler near Savannah, Georgia. Aquatic Mammals, 40, 81–94. Lange, A. M. T., & Johnson, K. L. (1981). Dorsal mantle length-total weight relationships of squids Loligo pealei and Illex illecebrosus from the Atlantic coast of the United States (NOAA Technical Report NMFD SSRF-745). Woods Hole, MA: National Marine Fisheries Service. 18 pp. Lewis, R. M., Ahrenholz, D. W., & Epperly, S. P. (1987). Fecundity of Atlantic menhaden, Brevoortia tyrannus. Estuaries, 10, 347–350. Lewison, R. L., Crowder, L. B., Read, A. J., & Freeman, S. A. (2004). Understanding impacts of fisheries bycatch on marine megafauna. Trends in Ecology and Evolution, 19, 598–604. Lindstrøm, U., Nilssen, K. T., Pettersen, L. M. S., & Haug, T. (2013). Harp seal foraging behaviour during summer around Svalbard in the northern Barents Sea: diet composition and the selection of prey. Polar Biology, 36, 305–320. Litz, J. A., Hughes, C. R., Garrison, L. P., Fieber, L. A., & Rosel, P. E. (2012). Genetic structure of common bottlenose dolphins (Tursiops truncatus) inhabiting adjacent South Florida estuaries- Biscayne Bay and Florida Bay. Journal of Cetacean Research and Management, 12, 107–117. 60

Ljungblad, D. K., Scoggins, P. D., & Gilmartin, W. G. (1982). Auditory thresholds of a captive Eastern Pacific bottle-nosed dolphin, Tursiops spp. Journal of the Acoustical Society of America, 72, 1726–1729. Lombard, E. (1911). Le signe de l’elevation de la voix. Annales des Maladies de l’Oreille et du Larynx 37, 101–119. Lopes, X. M., da Silva, E., Bassoi, M., dos Santos, R. A., & de Oliveira Santos, M. C. (2012). Feeding habits of Guiana dolphins, Sotalia guianensis, from south-eastern Brazil: new items and a knowledge review. Journal of the Marine Biological Association of the United Kingdom, 92, 1723–1733. Lynott, M. C. (2012). Life history of stranded bottlenose dolphins (Tursiops truncatus) in Virginia (M.S. Thesis). Old Dominion University, Norfolk, VA. 63 pp. MacArthur, R., & Pianka, E. (1966). On optimal use of a patchy environment. American Naturalist, 100, 603–609. Mallette, S. D., Mclellan, W. A., Scharf, F. S., Koopman, H. N., Barco, S. G., Wells, R. S., & Ann Pabst, D. (2016). Ontogenetic allometry and body composition of the common bottlenose dolphin (Tursiops truncatus) from the U.S. mid-Atlantic. Marine Mammal Science, 32, 86–121. Mann, J., Connor, R. C., Barre, L. M., & Heithaus, M. R. (2000). Female reproductive success in bottlenose dolphins (Tursiops sp.): life history, habitat, provisioning, and group-size effects. Behavioral Ecology, 11, 210–219. Mazzoil, M., Reif, J. S., Youngbluth, M., Murdoch, M. E., Bechdel, S. E., Howells, E., … Bossart, G. D. (2008). Home ranges of bottlenose dolphins (Tursiops truncatus) in the Indian River Lagoon, Florida: Environmental correlates and implications for management strategies. EcoHealth, 5, 278–288. McFee, W. E., & Lipscomb, T. P. (2009). Major pathologic findings and probable causes of mortality in bottlenose dolphins stranded in South Carolina From 1993 to 2006. Journal of Wildlife Diseases, 45, 575–593. McGurk, J. (1997). Stomach contents of stranded bottlenose dolphins, Tursiops truncatus, in Virginia 1987-1996 (M.S. Thesis). Old Dominion University, Norfolk, VA. 69 pp. McLellan, W. A., Friedlaender, A. S., Mead, J. G., Potter, C. W., & Pabst, D. A. (2002). Analysing 25 years of bottlenose dolphin (Tursiops truncatus) strandings along the Atlantic coast of the USA: Do historic records support the coastal migratory stock hypothesis? Journal of Cetacean Research and Management, 4, 297-304. Mead, J. G., & Potter, C. W. (1990). Natural history of bottlenose dolphins along the central Atlantic coast of the United States. In S. Leatherwood & R. R. Reeves (Eds.), The bottlenose dolphin (pp. 165–195). San Diego, CA: Academic Press. Mead, J. G., & Potter, C. W. (1995). Recognizing two populations of the bottlenose dolphin (Tursiops truncatus) of the Atlantic coast of North America: Morphologic and ecologic considerations. International Biological Research Institute Reports, 5, 31–44. Melo, C. L. C., Santos, R. A., Bassoi, M., Araújo, A. C., Lailson-Brito, J., Dorneles, P. R., & Azevedo, A. F. (2010). Feeding habits of delphinids (Mammalia: Cetacea) from Rio de Janeiro State, Brazil. Journal of the Marine Biological Association of the United Kingdom, 90, 1509–1515. Mohsin, A. K. M. (1981). Comparative account of the otoliths of the weakfishes (Cynoscion) of the Atlantic and Gulf coasts of the United States. Pertanika, 4, 109–111. Montie, E. W., Hoover, M., Kehrer, C., Yost, J., Brenkert, K., O’Donnell, T., & Denson, M. R. (2017). Acoustic monitoring indicates a correlation between calling and spawning in captive spotted seatrout (Cynoscion nebulosus). PeerJ, 2017(2), 1–27. Moore, J. E., Wallace, B. P., Lewison, R. L., Žydelis, R., Cox, T. M., & Crowder, L. B. (2009). A review of marine mammal, sea turtle and seabird bycatch in USA fisheries and the role of policy in shaping management. Marine Policy, 33, 435–451. 61

Moore, K. T., & Barco, S. G. (2013). Handbook for Recognizing, Evaluating and Documenting Human Interaction in Stranded Cetaceans and Pinnipeds. (NOAA Technical Memorandum, NOAA-TM- NMFS-SWFSC-510). National Marine Fisheries Service. 102 pp. Moore, M. J., Van Der Hoop, J., Barco, S. G., Costidis, A. M., Gulland, F. M., Jepson, P. D., … McLellan, W. A. (2013). Criteria and case definitions for serious injury and death of pinnipeds and cetaceans caused by anthropogenic trauma. Diseases of Aquatic Organisms, 103, 229–264. Moore, S. (2008). Marine mammals as ecosystem sentinels. Journal of Mammalogy, 89, 534–540. Murdy, E.O., Birdsong, R.S. and Musick, J.A. (1997). Fishes of Chesapeake Bay. Washington, DC: Smithsonian Institution Press. NOAA. 2019. 2013–2015 Bottlenose Dolphin Unusual Mortality Event in the Mid-Atlantic (Closed). Available at www.fisheries.noaa.gov/national/marine-life-distress/2013-2015-bottlenose- dolphin-unusual-mortality-event-mid-atlantic (accessed 4 February 2020). Noren, D. P., Holt, M. M., Dunkin, R. C., & Williams, T. M. (2017). Echolocation is cheap for some mammals: Dolphins conserve oxygen while producing high-intensity clicks. Journal of Experimental Marine Biology and Ecology, 495, 103–109. Noren, S. R., Biedenbach, G., & Edwards, E. F. (2006). Ontogeny of swim performance and mechanics in bottlenose dolphins (Tursiops truncatus). Journal of Experimental Biology, 209, 4724– 4731. Noren, S. R., Cuccurullo, V., & Williams, T. M. (2004). The development of diving bradycardia in bottlenose dolphins (Tursiops truncatus). Journal of Comparative Physiology B: Biochemical, Systemic, and Environmental Physiology, 174, 139–147. Noren, S. R., Williams, T. M., Pabst, D. A., Mclellan, W. A., & Dearolf, J. L. (2001). The development of diving in marine endotherms: preparing the skeletal muscles of dolphins, penguins, and seals for activity during submergence. Journal of Comparative Physiology B, 171, 127– 134. Noren, Shawn R., Lacave, G., Wells, R. S., & Williams, T. M. (2002). The development of blood oxygen stores in bottlenose dolphins (Tursiops truncatus): implications for diving capacity. Journal of Zoology, 258, 105-113. Page, L. M., Espinosa-Perez, H., Findley, L. T., Gilbert, C. R., Lea, R. N., Mandrak, N. E., … Nelson, J. S. (2013). Scientific Name , Occurrence , and Accepted Common Name. In Common and Scientific Names of Fishes from the United States, Canada, and Mexico: Vol. Special Pu (7th ed.). American Fisheries Society. Parks, S. E., Johnson, M. P., Nowacek, D., & Tyack, P. L. (2011). Individual right whales call louder in increased environmental noise. Biology Letters, 7, 33–35. Pate, S. Michelle, & McFee, W. E. (2012). Prey species of bottlenose dolphins (Tursiops truncatus) from South Carolina waters. Southeastern Naturalist, 11, 1–22. Pate, Susan Michelle. (2008). Stomach content analysis of stranded bottlenose dolphins (Tursiops truncatus) in South Carolina (M.S. thesis). College of Charleston, Charleston, SC. 129 pp. Peltier, H., Authier, M., Deaville, R., Dabin, W., Jepson, P. D., van Canneyt, O., … Ridoux, V. (2016). Small cetacean bycatch as estimated from stranding schemes: The common dolphin case in the northeast Atlantic. Environmental Science and Policy, 63, 7–18. Pine, M. K., Jeffs, A. G., Wang, D., & Radford, C. A. (2016). The potential for vessel noise to mask biologically important sounds within ecologically significant embayments. Ocean & Coastal Management, 127, 63–73. Piwetz, S. K. (2018). Effects of human activities on coastal dolphin behavior (Ph.D. thesis). Texas A&M University, College Station, TX. 158 pp. Powell, J. R., Machernis, A. F., Engleby, L. K., Farmer, N. A., & Spradlin, T. R. (2018). Sixteen years later: An updated evaluation of the impacts of chronic human interactions with bottlenose 62

dolphins (Tursiops truncatus truncatus) at Panama City, Florida, USA. Journal of Cetacean Research and Management, 19, 79–93. Ramcharitar, J., Gannon, D. P., & Popper, A. N. (2006). Bioacoustics of Fishes of the Family Sciaenidae (Croakers and Drums). Transactions of the American Fisheries Society, 135, 1409–1431. Read, A. J., Drinker, P., & Northridge, S. (2006). Bycatch of marine mammals in U.S. and global fisheries. Conservation Biology, 20, 163–169. Read, A. J., & Murray, K. T. (2000). Gross evidence of human-induced mortality in small cetaceans (NOAA Technical Memorandum NMFS-OPR-15). National Marine Fisheries Service. 21 pp. Recchia, C., & Read, A. (1989). Stomach contents of harbour porpoises, Phocoena phocoena (L.), from the Bay of Fundy. Canadian Journal of Zoology, 67, 2140-2146. Reeves, R. R., McClellan, K., & Werner, T. B. (2013). Marine mammal bycatch in gillnet and other entangling net fisheries, 1990 to 2011. Endangered Species Research, 20, 71–97. Robertson, K. M., & Chivers, S. J. (1997). Prey occurrence in pantropical spotted dolphins, Stenella attenuata, from the eastern tropical Pacific. Fishery Bulletin, 95, 334–348. Rosel, P. E., Hansen, L. J., & Hohn, A. A. (2009). Restricted dispersal in a continuously distributed marine species: common bottlenose dolphins Tursiops truncatus in coastal waters of the western North Atlantic. Molecular Ecology, 18, 5030–5045. Rossman, S., Berens McCabe, E., Barros, N. B., Gandhi, H., Ostrom, P. H., Stricker, C. A., & Wells, R. S. (2015). Foraging habits in a generalist predator: Sex and age influence habitat selection and resource use among bottlenose dolphins (Tursiops truncatus). Marine Mammal Science, 31, 155– 168. Samuel, Y., Morreale, S. J., Clark, C. W., Greene, C. H., & Richmond, M. E. (2005). Underwater, low-frequency noise in a coastal sea turtle habitat. The Journal of the Acoustical Society of America, 117, 1465–1472. Santos, M. B., Fernández, R., López, A., Martínez, J. A., & Pierce, G. J. (2007). Variability in the diet of bottlenose dolphin, Tursiops truncatus, in Galician waters, north-western Spain, 1990-2005. Journal of the Marine Biological Association of the United Kingdom, 87, 231–241. Schaefer, A. M., Stavros, H.-C. W., Bossart, G. D., Fair, P. a, Goldstein, J. D., & Reif, J. S. (2011). Associations between mercury and hepatic, renal, endocrine, and hematological parameters in Atlantic bottlenose dolphins (Tursiops truncatus) along the eastern coast of Florida and South Carolina. Archives of Environmental Contamination and Toxicology, 61, 688–695. Scheifele, P. M., Andrew, S., Cooper, R. A., Darre, M., Musiek, F. E., & Max, L. (2005). Indication of a Lombard vocal response in the St. Lawrence River beluga. Journal of the Acoustical Society of America, 117, 1486–1492. Scheinin, A. P., Kerem, D., Lojen, S., Liberzon, J., & Spanier, E. (2014). Resource partitioning between common bottlenose dolphin (Tursiops truncatus) and the Israeli bottom trawl fishery? Assessment by stomach contents and tissue stable isotopes analysis. Journal of the Marine Biological Association of the United Kingdom, 94, 1203–1220. Scott, T. (1903). Some further observations on the food of fishes, with a note on the food observed in the stomach of a common porpoise. Annual Report Fisheries Board of Scotland, Scientific Investigations. 9 pp. Scott, G. P., Burn, D. M., & Hansen, L. J. (1988). The dolphin dieoff: long-term effects and recovery of the population. Proceedings of the Oceans ’88 Conference, 819–823. Scott, M., R. Wells, and A. Irvine. 1990. A long-term study of bottlenose dolphins on the west coast of Florida. In S. Leatherwood & R. R. Reeves (Eds.), The bottlenose dolphin (pp. 235-244). San Diego, CA: Academic Press. SEDAR. (2020). Atlantic Menhaden Benchmark Stock Assessment Report (SEDAR 69). Charleston, SC: Southeast Data, Assessment, and Review. 691 pp. Shane, S. H., Wells, R. S., & Würsig, B. (1986). Ecology, behavior and social organization of the 63

bottlenose dolphin: A review. Marine Mammal Science, 2, 34–63. Silva, M. A. (1999). Diet of common dolphins, Delphinus delphis, off the Portuguese continental coast. Journal of the Marine Biological Association of the United Kingdom, 79, 531–540. Sinclair, E. H., Walker, W. A., & Thomason, J. R. (2015). Body size regression formulae, proximate composition and energy density of eastern Bering Sea mesopelagic fish and squid. PLoS ONE, 10, 1–14. Southall, E. B. L., Finneran, J. J., Reichmuth, C., Nachtigall, P. E., Ketten, D. R., Bowles, A. E., … Tyack, P. L. (2019). Marine mammal noise exposure criteria: Updated scientific recommendations for residual hearing effects. Aquatic Mammals, 45, 125–232. Staudinger, M. D., Juanes, F., & Carlson, S. (2009). Reconstruction of original body size and estimation of allometric relationships for the longfin inshore squid (Loligo pealeii) and northern shortfin squid (Illex illecebrosus). Fishery Bulletin, 107, 101–105. Stavros, H.-C. W., Stolen, M., Durden, W. N., McFee, W., Bossart, G. D., & Fair, P. A. (2011). Correlation and toxicological inference of trace elements in tissues from stranded and free- ranging bottlenose dolphins (Tursiops truncatus). Chemosphere, 82, 1649–1661. Stockin, K. A., Duignan, J. P., Roe, W. D., Meynier, L., Alley, M., & Fettermann, T. (2009). Causes of mortality in stranded common dolphin (Delphinus sp.) from New Zealand waters between 1998 and 2008. Pacific Conservation Biology, 15, 217–227. Stolen, M. K., Durden, W. N., & Odell, D. K. (2007). Historical synthesis of bottlenose dolphin (Tursiops truncatus) stranding data in the Indian River Lagoon System, Florida, From 1977 - 2005. Florida Scientist, 70, 45–54. Stolen, M. K., Odell, D. K., & Barros, N. B. (2002). Growth of bottlenose dolphins (Tursiops truncatus) from the Indian River Lagoon system, Florida, U.S.A. Marine Mammal Science, 18, 348– 357. Tellechea, J. S., Perez, W., Olsson, D., Lima, M., & Norbis, W. (2017). Feeding habits of franciscana dolphins (Pontoporia blainvillei): Echolocation or passive listening? Aquatic Mammals, 43, 430–438. Tennessen, J., & Parks, S. (2016). Acoustic propagation modeling indicates vocal compensation in noise improves communication range for North Atlantic right whales. Endangered Species Research, 30, 225–237. Thurstan, R. H., McClenachan, L., Crowder, L. B., Drew, J. A., Kittinger, J. N., Levin, P. S., … Pandolfi, J. M. (2015). Filling historical data gaps to foster solutions in marine conservation. Ocean and Coastal Management, 115, 31–40. Toth, J. L., Hohn, A. A., Able, K. W., & Gorgone, A. M. (2011). Patterns of seasonal occurrence, distribution, and site fidelity of coastal bottlenose dolphins (Tursiops truncatus) in southern New Jersey, U.S.A. Marine Mammal Science, 27, 94–110. Treacy, S. D., & Crawford, T. W. (1981). Retrieval of otoliths and statoliths from gastrointestinal contents and scats of marine mammals. The Journal of Wildlife Management, 45, 990–993. Twiner, M. J., Fire, S., Schwacke, L., Davidson, L., Wang, Z., Morton, S., … Wells, R. S. (2011). Concurrent exposure of bottlenose dolphins (Tursiops truncatus) to multiple algal toxins in Sarasota Bay, Florida, USA. PloS ONE, 6, 1–15. Vail, C. S. (2016). An overview of increasing incidents of bottlenose dolphin harassment in the Gulf of Mexico and possible solutions. Frontiers in Marine Science, 3, 1–7. Víkingsson, G. A., Elvarsson, B. Þ., Ólafsdóttir, D., Sigurjónsson, J., Chosson, V., & Galan, A. (2014). Recent changes in the diet composition of common minke whales (Balaenoptera acutorostrata) in Icelandic waters. A consequence of climate change? Marine Biology Research, 10, 138–152. Wang, K. R., Payne, P. M., & Thayer, V. G. (1994). Coastal stock(s) of Atlantic bottlenose dolphin: status review and management (NOAA Technical Memorandum. NMFS-OPR-4). National Marine 64

Fisheries Service. 120 pp. Waring, G. T., Josephson, E., Maze-Foley, K., and Rosel, P. E. (2010). U. S. Atlantic and Gulf of Mexico U.S. Atlantic and Gulf of Mexico Marine Mammal Stock Assessments- 2009 (NOAA Tech Memo NMFS NE 213). National Marine Fisheries Service. 528 pp. Weilgart, L. S. (2007). The impacts of anthropogenic ocean noise on cetaceans and implications for management. Canadian Journal of Zoology, 85, 1091–1116. Wells, R. S. (1986). Structural aspects of dolphin societies (Ph.D. dissertation). University of California, Santa Cruz, CA. 234 pp. Wells, R., Hansen, L., Baldridge, T., Dohl, D. & Defran, R. (1990). Northward extension of the range of bottlenose dolphins along the California coast. In S. Leatherwood & R. R. Reeves (Eds.), The bottlenose dolphin (pp. 421-426). San Diego, CA: Academic Press. Wells, R. S., Rhinehart, H. L., Hansen, L. J., Sweeney, J. C., Townsend, F. I., Stone, R., … Rowles, T. K. (2004). Bottlenose dolphins as marine ecosystem sentinels: developing a health monitoring system. EcoHealth, 1, 246–254. 6 Williams, C. R. (2009). Influences of boat traffic and noise on behaviors and vocalizations of bottlenose dolphins (Tursiops truncatus) in the Indian River Lagoon, Florida (PhD thesis). Florida Institue of Technology, Melbourne, FL. 109 pp. Winn, H. E. (1982). A characterization of marine mammals and sea turtles in the mid- and north Atlantic areas of the U.S. outer continental shelf (Final Report of the Cetacean and Turtle Assessment Program). University of Rhode Island, prepared for the Bureau of Land Management, U.S. Department of the Interior, under contract AA551-CT8-48. 570 pp. Worthy, G. (2008). Feeding ecology of the bottlenose dolphin (Tursiops truncatus) in the Indian River Lagoon, Florida (Final Report to the Protect Wild Dolphins License Plate Program Project Number 2003-16). 411 pp. Yilmaz, S., Yazicioglu, O., Yazici, R., & Polat, N. (2015). Relationships between fish length and otolith size for five cyprinid species from Lake Ladik, Samsun, Turkey. Turkish Journal of Zoology, 39, 438–446. Yordy, J. E., Wells, R. S., Balmer, B. C., Schwacke, L. H., Rowles, T. K., & Kucklick, J. R. (2010). Life history as a source of variation for persistent organic pollutant (POP) patterns in a community of common bottlenose dolphins (Tursiops truncatus) resident to Sarasota Bay, FL. The Science of the Total Environment, 408, 2163–2172. York, C. A., & Bartol, I. K. (2014). Lateral line analogue aids vision in successful predator evasion for the brief squid, Lolliguncula brevis. Journal of Experimental Biology, 217, 2437–2439. York, Carly A., & Bartol, I. K. (2016). Anti-predator behavior of squid throughout ontogeny. Journal of Experimental Marine Biology and Ecology, 480, 26–35. York, Carly A., Bartol, I. K., & Krueger, P. S. (2016). Multiple sensory modalities used by squid in successful predator evasion throughout ontogeny. The Journal of Experimental Biology, 219, 2870– 2879. Zolman, E. S. (2002). Residence patterns of bottlenose dolphins (Tursiops truncatus) in the Stono River Estuary, Charleston County, South Carolina, U.S.A. Marine Mammal Science, 18, 879–892.

VITA

Kristen Marie Volker Department of Biological Sciences Old Dominion University Norfolk, VA 23529

EDUCATION Old Dominion University, Norfolk, VA Master of Science, Biology, May 2020 Advisor: Dr. Ian Bartol

University of New England, Biddeford, ME, Dec 2008 Major: Marine Biology Minor: Mathematics

PROFESSIONAL EXPERIENCE Sep 2019- present Assistant Research Coordinator, International Fund for Animal Welfare, Marine Mammal Rescue & Research, Yarmouth Port, MA Jun 2017- Sep 2019 Assistant Necropsy Coordinator, International Fund for Animal Welfare, Marine Mammal Rescue & Research, Yarmouth Port, MA Nov 2014- Jun 2017 Stranding Response Necropsy Manager, Virginia Aquarium & Marine Science Center, Stranding Response Program, Virginia Beach, VA Jul 2011- Nov 2014 Stranding Technician, Virginia Aquarium & Marine Science Center, Stranding Response Program, Virginia Beach, VA Nov 2008- Jul 2011 Stranding Technician Assistant, Virginia Aquarium & Marine Science Center, Stranding Response Program, Virginia Beach, VA

SELECTED PUBLICATIONS, GRANT REPORTS, AND SCIENTIFIC PRESENTATIONS Wu, Qingzhong, Conway, J., Phillips, K.M., Stolen, M., Durden, W.N., Fauquier, D., McFee, W.E. and Schwacke, L. 2016. Detection of Brucella spp. in bottlenose dolphins (Tursiops truncatus) by a real-time PCR using blowhole swabs. Diseases of Aquatic Organisms. Volker, K.M. 2018. Small Cetacean Mass Stranding Workshop and Keynote Address, 4th Latin American Conference on the Rehabilitation of Marine Animals, Florianopolis, SC, Brazil Phillips K.M., Lynott M.C., Gannon D.P., Barco S.G. 2016. Analysis of archived stomach content samples from stranded Tursiops in Virginia. Final Report to the John H. Prescott Marine Mammal Rescue Assistance Grant Program, #NA11NMF4390098. VAQF Scientific Report 2016-03, Virginia Beach, VA, 186 pp. Phillips, K., Lynott, M., Reinheimer, S., Gannon, D., Gaff, H., and Bartol, I. 2015. Diet analysis of stranded bottlenose dolphins (Tursiops truncatus) in Virginia, USA. Poster presentation at 21st Biennial Conference on the Biology of Marine Mammals. Dec 13-18. Phillips, K., Lynott, M., Chromik, M., Reinheimer, S., Gannon, D., Gaff, H., and Bartol, I. 2015. Diet analysis of stranded bottlenose dolphins (Tursiops truncatus) in Virginia. Oral presentation at Greater Atlantic Regional Stranding Conference. Oct 13-16. Phillips, K., Mallette, S., McLellan, W., Rotstein, D., and Barco, S. 2015. Death by plastic: Partial DVD case found in sei whale’s stomach. Oral presentation at the Southeast and Mid-Atlantic Marine Mammal Symposium. Mar 27-29, Virginia Beach, VA. Phillips, K., Lynott, M., Barco, S., Fauquier, D., Rotstein, D., DiGiovanni, R., Garron, M., and Goldstein, T. 2014. Bring out your dead: Looking at Virginia’s 5-year stranding trends in relation to the 2013 mid-Atlantic Tursiops UME. Oral presentation at the Southeast and Mid-Atlantic Marine Mammal Symposium. Mar 28-30, Wilmington, NC.