<%BANNER%>

A Visual Body Condition Index for Bottlenose Dolphins (Tursiops truncatus)

MISSING IMAGE

Material Information

Title:
A Visual Body Condition Index for Bottlenose Dolphins (Tursiops truncatus)
Physical Description:
1 online resource (87 p.)
Language:
english
Creator:
Gryzbek, Mary K
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Wildlife Ecology and Conservation
Committee Chair:
Pine, William E, Iii
Committee Co-Chair:
Wells, Randall Stewart
Committee Members:
Pabst, D Ann

Subjects

Subjects / Keywords:
analysis -- bmi -- body -- bottlenose -- condition -- depression -- dolphin -- index -- length-weight -- post-nuchal -- visual
Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
Genre:
Wildlife Ecology and Conservation thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Since the early 1990s, concerns about the health of bottlenose dolphin (Tursiops truncatus) populations faced with natural and anthropogenic threats have led to the need for a reliable and cost-effective indicator of dolphin body condition.  Previous studies on cetaceans have developed methods that monitor the condition of individuals using external features that are visible in photographs.  Based on the distribution of fat reserves in bottlenose dolphins, I investigated the post-nuchal region visible in photographs to determine if depressions were indicative of poor body condition.  I developed a simple and non-invasive system for determining the presence or absence of post-nuchal depressions (PNDs) from digital photographs of stranded dolphins.  Dolphins with PNDs consistently had lower length-weight measurements and body mass index values than those without PNDs.  The PND index appears to be a viable tool for assessing bottlenose dolphin body condition.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Mary K Gryzbek.
Thesis:
Thesis (M.S.)--University of Florida, 2013.
Local:
Adviser: Pine, William E, Iii.
Local:
Co-adviser: Wells, Randall Stewart.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045992:00001

MISSING IMAGE

Material Information

Title:
A Visual Body Condition Index for Bottlenose Dolphins (Tursiops truncatus)
Physical Description:
1 online resource (87 p.)
Language:
english
Creator:
Gryzbek, Mary K
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Wildlife Ecology and Conservation
Committee Chair:
Pine, William E, Iii
Committee Co-Chair:
Wells, Randall Stewart
Committee Members:
Pabst, D Ann

Subjects

Subjects / Keywords:
analysis -- bmi -- body -- bottlenose -- condition -- depression -- dolphin -- index -- length-weight -- post-nuchal -- visual
Wildlife Ecology and Conservation -- Dissertations, Academic -- UF
Genre:
Wildlife Ecology and Conservation thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Since the early 1990s, concerns about the health of bottlenose dolphin (Tursiops truncatus) populations faced with natural and anthropogenic threats have led to the need for a reliable and cost-effective indicator of dolphin body condition.  Previous studies on cetaceans have developed methods that monitor the condition of individuals using external features that are visible in photographs.  Based on the distribution of fat reserves in bottlenose dolphins, I investigated the post-nuchal region visible in photographs to determine if depressions were indicative of poor body condition.  I developed a simple and non-invasive system for determining the presence or absence of post-nuchal depressions (PNDs) from digital photographs of stranded dolphins.  Dolphins with PNDs consistently had lower length-weight measurements and body mass index values than those without PNDs.  The PND index appears to be a viable tool for assessing bottlenose dolphin body condition.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Mary K Gryzbek.
Thesis:
Thesis (M.S.)--University of Florida, 2013.
Local:
Adviser: Pine, William E, Iii.
Local:
Co-adviser: Wells, Randall Stewart.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045992:00001


This item has the following downloads:


Full Text

PAGE 1

1 A VISUAL BODY CONDITION INDEX FOR BOTTLENOSE DOLPHINS ( TURSIOPS TRUNCATUS ) By MARY KATHLEEN GRYZBEK A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Mary Kathleen Gryzbek

PAGE 3

3 To my Father and M other, Thomas and Marilyn Gryzbek, for supporting me w hile I follow my dreams

PAGE 4

4 ACKNOWLEDGMENTS I thank my committee members, Drs. Randall S. Wells, William E. Pine, III, and D. Ann Pabst, for their patience, guidance, and overall support as I learned how to be a better scientist while completing research for my thesis. I will always be grateful for the various opportunities they provided me to improve as an individual both professionally and personally in the field of wildlife ecology and conservation. I thank Randy for being the first to agree to be an advisor on my committee and therefore setting up the path for me to both pursue my M S degree and gain valuable research experience involving t he conservation of a cetacean. I thank the Chicago Zoological Society for their support and for giving me the opportunity to be a staff member of the Sarasota Dolphin Research Program (SDRP) while conducting my thesis research. I thank the Stranding Investigations Program (SIP) at Mote Marine Laboratory for allowing me to use their data. I thank Gretchen Lovewell, the P rogram M anager of SIP, for her generosit y in giving of her time to help me collect and organize my stranding sample. I also thank the SIP interns who helped with the stranding data entry process. I thank the SDRP staff for all of their help whenever I have needed it and for having taught me so much both as an intern and grad uate student I thank all of the past and present SDRP and SIP volun teers, interns, and staff for collecting decades of health assessment and stranding data that were vital to my project I thank Margaret Lynott and the Vi rginia Aquarium Stranding Response Program regards to neonate length. I thank my Mom and Dad for their love and generous support during this entire exp erience Finally, I thank all my family and frien ds for their continued love encouragement, and support.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ........................... 10 ABSTRACT ................................ ................................ ................................ ................... 11 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 The Need for Improved Monitoring of Bottlenose Dolphins ................................ ..... 12 Tools for Monitoring Wildlife Populations ................................ ................................ 13 Assessment of Body Condition in Cetaceans ................................ ......................... 13 Research Involving the Direct Handling of Individuals ................................ ...... 13 Research Involving Photographic Analyses ................................ ..................... 15 Post Nuchal Depr ession as an Indicator of Bottlenose Dolphin Body Condition ..... 16 2 MATERIALS AND METHODS ................................ ................................ ................ 21 Study Sample Description ................................ ................................ ....................... 21 Data Sources and Collection ................................ ................................ ............ 21 Data Organization ................................ ................................ ............................ 21 Development of PND Ind ex ................................ ................................ .................... 24 Cranial Starting Points Common to All Four Line Techniques .......................... 25 Method #1 ................................ ................................ ................................ ........ 26 Method #2 ................................ ................................ ................................ ........ 27 Method #3 ................................ ................................ ................................ ........ 27 Method #4 ................................ ................................ ................................ ........ 27 Body Con dition Analysis ................................ ................................ ......................... 28 Length Weight Models ................................ ................................ ..................... 28 Body Mass Index (BMI) Calculations ................................ ................................ 29 95% Quantile Ranges for the Recommended PND Index Using Non PND Data ................................ ................................ ................................ .............. 29 Comparison of BMI Values between Dolphins with a PND and Dolphins Considered to Be Emaciated ................................ ................................ ......... 30 3 RESULTS ................................ ................................ ................................ ............... 39 Summary of Length Weight Data ................................ ................................ ............ 39 Simple Linear Regression Model and Assumptions ................................ ................ 39

PAGE 6

6 Nonlinear Model Parameter Estimates and Visual Comparison of the Linear and Nonlinear Model Fits ................................ ................................ ............................ 39 PND Indices Performance Assessment ................................ ................................ .. 40 95% Quantile Regression Reference Ranges ................................ .................. 40 Females ................................ ................................ ................................ ..... 40 Males ................................ ................................ ................................ ......... 40 BMI Calculations ................................ ................................ .............................. 41 Females ................................ ................................ ................................ ..... 41 Males ................................ ................................ ................................ ......... 41 95% Quantile Ranges for Method #1 Using Non PND Data ............................ 42 Comparison of BMI Values between Dolphins with PNDs and Dolph ins Considered to be Emaciated ................................ ................................ ......... 42 4 DISCUSSION ................................ ................................ ................................ ......... 61 Differences in the Outcomes between the Four PND Indices ................................ 61 Effect of Varying the Cranial Start Points for One PND Index ................................ 63 Body Condition Comparison between Dolphins with and without PNDs ................. 64 Recommended PND Index ................................ ................................ ..................... 66 Seasonal Effect on PND Occurrence ................................ ................................ ...... 67 Applying PND Ind ex to Field Photos ................................ ................................ ....... 68 Future Research ................................ ................................ ................................ ..... 69 Conclusions ................................ ................................ ................................ ............ 70 LIST OF REFERENCES ................................ ................................ ............................... 82 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 87

PAGE 7

7 LIST OF TABLES Table page 2 1 Summary of possi ble cranial start points of the line, amount of dorsal surface required in the photo, and caudal endpoints of the line for all four methods.. ..... 31 3 1 Females that met the requirements for the P ND index and body condition analyses, including the results for each of the four PND indices. ....................... 43 3 2 Males that met the requirements for the PND index and body condition analyses, including th e results for each of the four PND indices. ....................... 44 3 3 a and b parameter estimates from nonlinear OLS regression fit of with corresponding standard error values. ................................ ..... 45 3 4 Quantiles, tau values, and a and b parameter estimates with their standard errors for the 95% quantile regression fit of including both PND and non PND data ................................ ................................ .................... 45 3 5 Wilcoxon rank sum test W statistic and corresponding p values for each of the four PND indices. ................................ ................................ .......................... 45 3 6 Quantiles, tau values, and a and b parameter estimates with their standard errors for the 95% quantile regression fit of including only non PND data. ................................ ................................ ................................ .... 46 4 1 Causes of death for dolphin s identified as having PNDs by Method #1. ............ 71

PAGE 8

8 LIST OF FIGURES Figure page 1 1 The internal anatomy of a bottlenose dolphin depicting the area from the tip of the nuchal crest to the cranial margin of the thoracic cavity, areas that roughly bound the post nuchal reg ion.. ................................ .............................. 2 0 1 2 A cross section through the cervical region of a 170.9 cm bottlenose dolphin. ................................ ................................ ................................ ........................... 20 2 1 Examples of wounds \ tissue loss that were thought to have a significant effect on the total weights of the animals and wounds th at caused minimal tissue loss .. ................................ ................................ ................................ ................... 32 2 2 Examples of ideal photo quality for PND analysis ................................ .............. 33 2 3 Location of the external ear relative to the nuchal crest for two dolphins. ......... 34 2 4 Examples of the three possible cranial s tarting points to the PND lines depending on whether or not the nuchal crest or external ear was visible.. ....... 35 2 5 Examples of each of the four methods ap plied to an animal with a PND. ......... 36 2 6 Examples of each of the four methods appli ed to an animal without a PND ..... 37 3 1 Simple linear regression weight vs. length model for females. ........................... 47 3 2 Q q normality plot of studentized residuals of female weight vs. length simple linear regression demonstrating that the studentized residuals do not meet the assumption of normality. ................................ ................................ ............... 48 3 3 Residuals vs. fitted values for female weight vs. length simple linear regression indicating that the residuals do not meet the assumption for homogenous variance. ................................ ................................ ....................... 48 3 4 Simple linear regression weig ht vs. length model for males. .............................. 49 3 5 Q q normality plot of studentized residuals of male weight vs. length simple linear regression showing that the studentized residuals do not meet the assumption of normality. ................................ ................................ ..................... 50 3 6 Residuals vs. fitted values for male weight vs. length simple linear regr ession model, showing that the residuals do not meet the assumption for homogenous variance. ................................ ................................ ....................... 50 3 7 Nonlinear OLS regression fit of to female length weight data. ................................ ................................ ................................ ................... 51

PAGE 9

9 3 8 Nonlinear OLS regression fit of to male length weight data. ................................ ................................ ................................ ................... 52 3 9 95% quantile regression ranges and observed data f or females for each of the four PND indices. ................................ ................................ ......................... 53 3 10 95% quantile regression ranges and observed data for males for ea ch of the four PND indices. ................................ ................................ ................................ 54 3 11 Boxplots for female PND and non PND BMI values for each of the four PND indices. ................................ ................................ ................................ .............. 55 3 12 Boxplots for male PND and non PND BMI values for each of the four PND indic es. ................................ ................................ ................................ .............. 56 3 13 Female 95% quantile regression ranges fit to non PND data of Method #1. ..... 57 3 14 Male 95% quantile regression range s fit to non PND data of Method #1. ......... 58 3 15 Boxplots of female BMI values for individuals with PNDs identified by Method #1 and for individuals noted as emaciated within their necropsy report s. .......... 59 3 16 Boxplots of male BMI values for individuals with PNDs identified by Method #1 and for individuals noted as emaciated within their necropsy reports. .......... 60 4 1 Example of how the PND outcomes differed between Me thods #1 and #4 using MML0538. ................................ ................................ ................................ 72 4 2 Examples of dolphins that were indicated as having PNDs by Met hods #3 or 4 but not by Methods #1 and 2. ................................ ................................ ......... 73 4 3 Male MML1104, where Methods #3 and 4 indicated it had a PND, while Methods #1 and 2 did not.. ................................ ................................ ................. 74 4 4 Method #1 applied to MML0503 using four different cranial start points ............ 75 4 5. Method #1 applied to MML1107 using four different cranial start points. ............ 76 4 6 Number of individuals with and without PNDs as identified by Method #1 by month for all years pooled together.. ................................ ................................ .. 78 4 7 MML0904 taken in the field eight days before its necropsy. .. ............................. 79 4 8 Method #1 applied to photos of the female MM L0904 with different postures ... 80 4 9 Female, MML0907, taken at a skewed angle in the field about a month a nd a h alf before its necropsy. ................................ ................................ ..................... 81

PAGE 10

10 LIST OF ABBREVIATIONS BMI Body Mass Index CBD Could not be determined COD Cause of death CIs Condition Indices MML Mote Marine Laboratory MMPA Marine Mammal Protection Act of 1972 NMFS S National Ma rine Fisheries Service NOAA National Oceanic and Atmospheric Administration PND Post n uchal d epression SDRP Sarasota Dolphin Research Program SIP Mote Marine Laboratory Stranding Investigations Program UME Unusual Mortality Event

PAGE 11

11 Abstract of Thesis Prese nted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science A VISUAL BODY CONDITION INDEX FOR BOTTL E NOSE DOLPHINS ( TURSIOPS TRUNCATUS ) By Mary Kathleen Gryzbek Au gust 2013 Chair: William E. Pine, III Co chair: Randall S. Wells Major: Wildlife Ecology and Conservation Since the early 1990 s, concerns about the health of bottlenose dolphin ( Tursiops truncatus ) populations faced with natural and anthropogenic threat s have led to the need for a reliable and cost effective indicator of dolphin body condition. Previous studies on cetaceans have developed methods that monitor the condition of individuals using external features that are visible in photographs. Based on the distribution of fat reserves in bottlenose dolphins, I investigated the post nuchal region visible in photographs to determine if depressions were indicative of poor body condition. I developed a simple and non invasive system for determining the pre sence or absence of post nuchal depressions (PNDs) from digital photographs of stranded dolphins. Dolphins with PNDs consistently had lower length weight measurements and body mass index values th an those without PNDs. T he PND index appears to be a viabl e tool for assessing bottlenose dolphin body condition.

PAGE 12

12 CHAPTER 1 INTRODUCTION The Need for Improved Monitoring of Bottlenose Dolphins From 1991 to the present, the National Oceanic and Atmospheric ice (NMFS) has formally recognized 58 marine mammal unusual mortality events (UMEs) in the Uni ted States (NOAA Fisheries O ffice of P rotected R esources 2013 a ). The Marine Mammal ves a significant die off of any marine mammal population; and demands immediate MMPA 1972). Causes of marine mammal UMEs have been related to parasites, viruses, bacteria, biotoxins, human interactions, oil spills, and variation in oceanograph ic conditions (Gulland and Hall 2007). In the Gulf of Mexico alone, eight bottlenose dolphin ( Tursiops truncatus ) UMEs occurred during 1993 t hrough 2008, resulting in at least 93 0 deaths (Waring et al. 2012). Another UME declared during February 2010 in the northern Gulf of Mexico for cetaceans continues with more than 88 0 bottlenose dolphins recovered through Ju ly 14 2013 (NOAA Fisheries O ffice of P rotected R esources 2013 b ) In light of the frequent occurrence of bottlenose dolphin UMEs a reliable an d cost effective means of monitoring b ody condition would be beneficial for the conservation and management of this species. Wells et al. (2004) recognized the need for the improvement of baseline information on dolphin populations prior to die offs. The se authors stressed the importance of a proactive approach to monitoring these populations instead of waiting for large scale strandings to occur

PAGE 13

13 Tools for Monitoring Wildlife Populations discerning the level of protection that is necessary for its management (Fowler and Siniff 1992). A th (Stevenson and Woods 2006). Therefore, the use of CIs can lead to a better understanding of a status and trajectory CIs are also used to compare the relative health of populations to one another (Stevenson and Woods 2006). Ma ny types of indices have been used to monitor the condition of individuals within wildlife populations for the purpose of management and conservation. A review of CIs by Stevenson and Woods (2006) demonstrates the range of these metrics including externa l assessments of size, shape, skin condition, and fat scores; calculations involving body length and weight measurements; examination of the dimensi ons of internal organs; the analysis of the biochemistry of blood, feces, urine, and tissues ; and the direct examination of body are typically trying to determine how much energy is available in the form of fat when using these metrics (Stevenson and Woods 2006). CIs have be en applied to studies of a variety of species in relation to environmental threats, life history traits, activity cycles, and ecological interactions ( Stevenson and Woods 2006). The wide use of CIs in both environmental and conservation biology is a testa ment to their utility. Assessment of Body Condition in Cetaceans Research Involving the Direct Handling of Individuals Studies on cetacean body (fat) condition include those that involve directly handling live animals or carcasses and those that analyze ph otographs of individuals.

PAGE 14

14 Morphometric measurements have included girth and blubber thickness at various sites, blubber mass, and total weight (Lockyer et al. 1985, Lockyer 1986, Read 1990, Kuiken et al. 1994, Ichii et al. 1998, Haug et al. 2002, Koopman et al. 2002, Evans et al. 2003, Struntz et al. 2004, Dunkin et al. 2005, Caon et al. 2007, Dunkin et al. 2010, Gmez Campos et al. 2011, Miller et al. 2011, Christiansen et al. 2013, Hart et al. 2013). Blubber lipid content has also been examined (Lockyer et al. 1985, Lockyer 1986, Kuiken et al. 1994, Evans et al. 2003, Struntz et al. 2004, Dunkin et al. 2005, Dunkin et al. 2010, Montie et al. 2008, Gmez Campos et al. 2011). The analyses of these measurements have varied greatly For example, how length is incorporated into the examination of body condition has differed between studies Researchers have compared raw body condition measurements between reproductive classes that differ in length (Lockyer et al. 1985, Caon et al. 2007 ) Other studies hav e adjusted for length when it was correlat ed with the examined body condition measurement (Ichii et al 1998 Miller et al. 2011 ) Length adjustments were made by expressing the body condition measurement as a percentage of length or by taking the ratio o f the body condition measurement to length (Ichii et al. 1998, Miller et al. 2011) Furthermore, researchers have examined the residuals resulting from regressions between standard length and a variety of the raw or log transformed body condition measurem ents (Read 1990, Kuiken et al. 1994, Haug et al. 2002, Evans et al. 2003, Gmez Campos et al. 2011). Analyses have also included calculating a fat index for blubber (mean body girth mean blubber thickness percentage lipid content of blubber ) and muscl e ( mean cross sectional area of the body inside the blubber layer percentage lipid content of muscle) of fin whales ( Balaenoptera physalus ) (Lockyer 1986) Another approach taken

PAGE 15

15 by Christiansen et al. ( 2013 ) involved calcul ating total blubber volume by dividing the minke whale body into five frustums calculating the volume of each frustum, and then summing those volumes. In addition, Hart et al. (2013) used length weight and length maximum girth models in combination with quantile regression to create baseline 9 5th percentile reference ranges for bottlenose dolphins The diversity of methods used to assess cetacean body condition reflects a broad interest within the community to be able to assess this important individual and popula tion parameter Th e diversity of methods also reflects the large difference s among cetaceans in size the logistical difficulty inherent in obtaining measurements of these animals due to their size, the limited view of their entire body while in water, their availability fo r predictable sightings and re sightings and their sometimes expansive marine habitats (Pettis et al. 2004). Stevenson and Woods (2006) stated the specific importance of using nondestructive methods to obtain CIs that are easily applied in the field and photographic analyses are a good example of such methods Research Involving Photographic Analyses Taking photographs of cetaceans is a much easier task than having to handle them to obtain some measurements. In addition to being easier for the researche rs, the animals themselves undergo much less stress while being phot ographed than if they were handled. A few studies on cetaceans have taken advantage of using photographs to evaluate body condition. For instance, in addition to assessing skin condition via photographic analysis, Pettis et al. (2004) examined body condition of the North Atlantic right whale ( Eubalaena glacialis ) by scoring the area caudal to the blowhole in lateral photos using a three point scale based on the concavity or convexity crea ted by the amount of blubber and subcutaneous fat seen there. Similarly, Bradford et al. (2012)

PAGE 16

16 evaluated the concavity or convexity created by blubber and subcutaneous fat in three body regions on the western gray whale ( Eschrichtius robustus ) as seen in lateral photos the area caudal to the blowhole, the scapular regi on, and the lateral flanks. A three point scale was used to score t he first body region, while a two point scale was used to score the remaining regions. These authors emphasized the area caudal to the blowhole when creating composite scores for overall body condition (Bradford et al. 2012) Both of these studies included a degree of subjectivity in the ir scoring systems i nstead of a specific metric A lternatively, length and width measurements obtained from photos taken during aerial surveys of each of these species, as well as southern right whales ( Eubalaena australis ) have been used to assess body condition (Perryma n and Lynn 2002, Miller et al. 2012). These p hotogrammetric analyses demonstrate how objective measurements can be obtained from photos. The success of all four of these studies in detec t i ng differences in body condition of different reproductive classes lends support to the use of photogra phic assessments as a viable tool in cetacean body condition analyses. Post Nuchal Depression as an Indicator of Bottlenose Dolphin Body Condition I used a combination of direct morphometric measurements and photographi c analyses to determine if post nuchal depressions are indicative of poor body condition in bottlenose dolphins. A post nuchal depression (PND) is a concavity o n the dorsal surface just caudal to the nuchal crest of the skull that can extend thro ugh the cervical region ( Figure 1 1) PNDs are a trait seen in d olphins commonly referred to as like shape of the anterior portion of their body. In young animals and animals in good body co ndition, a n

PAGE 17

1 7 adipose depot, or fat pad is located under the blubber layer in this post nuchal region ( Figure 1 2). The depletion of this fat pad likely contribute s to the expression of a PND; however, the depletion of blubber and muscle found in and near the post nuchal area may also contribute to the depression T he physiology behind the depletion of tissues in this region of the dolphin body is poorly understood. Post nuchal b lubber and post nuchal fat pad thickness measurements collected via ultrasou nd during bottlenose dolphin h ealth assessments in Sarasota Bay, Florida (Wells et al. 2004, Wells et al. 2009) during spring and summer 2004 2013 show that both of these tissues range in thickness between individuals (R. Wells, unpublished data). For fem ales ( n =62 with some individuals measured multiple times in different years ) ranging in total length from 180 to 265 cm, the blubber thickness at the post nuchal site has ranged from 10 to 20 mm. The post nuchal fat pad of these females has ranged from 1 to 23 mm. For males ( n =67 with some individuals measured multiple times in different years ) ranging in total length from 166 to 281 cm, blubber thickness at the post nuchal site has ranged from 10 to 21 mm. The post nuchal fat pad of males has ranged fro m 0 to 25 mm (R. Wells, unpublished data). The range of thickness of these tissues in the post nuchal region supports the notion that both blubber and the post nuchal fat pad contribute to the presence of a PND. Furthermore studies have used atrophied e paxial musculature as a sign of emaciation for bottlenose dolphins (Struntz et al. 2004, Dunkin et al. 2005). Given the proximity of muscle to the fat pad in the post nuchal region (Figure 1 2) and knowing the depletion of muscle can be viewed externally in the area below the dorsal fin of the dolphin body, depleted muscle tissue also likely contributes to the external expression of a PND.

PAGE 18

18 Although t he presence of PNDs has been used to differentiate between bottlenose dolphins in emaciated vs. robust con dition by trainers working with bottlenose dolphins in managed populations and in some field studies ( Struntz et al. 2004, Dunkin et al. 2005, Fair et al. 2006, Yordy et al. 2010) only two studies are known to have related the PND trait to a quantifiable body condition metric Both Struntz et al. (2004) and Dunkin et al. (2005) used the atrophy of the post nuchal fat pad as one of a few characteristics to determine if a bottlenose dolphin was emaciated or not, and the number of dolphins with a PND in the ir sample s of emac iated individuals ( n = 2 and 5 respectively ) was not reported. Struntz et al. (2004) found that the absolute blubber depth of emaciated adult dolphins was significantly less than that of robust adults, and the blubber lipid content of ema ciated adult dolphins was significantly less than that of juveniles and robust adults. Dunkin et al. (2005) demonstr ated that emaciated adult dolphins had a significantly lower amount of lipid in their blubber than the blubber of robust individuals from al l life history categories except fetuses. T he blubber of emaciated adults contained less lipid than the blubber of fetuses, but the difference was not significant. F urther research using a larger sample size of dolphins with PNDs would be beneficial for understanding how body condition measurements compare between individuals with and without this specific trait. My two main obj ectives were to: 1) create a technique to objectively and systematically identify PNDs in photogra phs of bottlenose dolphins a nd 2) determine if PNDs were indicative of poor body condition. Stranding case records were used to develop the PND assessment system because they contained both photos and body condition measurements, which allowed for the comparison of these measurement s

PAGE 19

19 between animals with and without PNDs. Additionally, they included animals dying from a variety of causes with and without a relationship to body condition. The overall goal of this study was to develop a logistically simple and non invasive method for assessing bottlenose dolphin body condition in the field by using photographs. Although I created a PND index using photos of stranded individuals, I demonstrate how the technique can readily be applied to free swimming dolphins as well. For the second objective, I hypothesized that animals with PNDs would weigh less for a given length and would have smaller body mass index (BMI) values. T o test this hypothesis, I created 95% quantile regression reference ranges based on length weight models and calcula ted BMI for stranded animals (Hart et al. 2013).

PAGE 20

20 Figure 1 1. The internal anatomy of a bottlenose dolphin depicting the area from the tip of the nuchal crest to the cranial margin of the thoracic cavity areas that roughly bound the post nuchal reg ion. Illustration courtesy of Dr. S.A. Rommel. Figure 1 2. A cross section through the cervical region of a 170.9 cm bottlenose dolphin The ruler is placed at the dorsal surface of the cross section. Bracket s denote the location of blubber, the post nuchal fat pad, and axial muscle. Photo courtesy of Dr. D. Ann Pabst. Nuchal Crest Cranial Margin of Thoracic Cavity Rommel 2000 Blubber Post N uchal Fat Pad Muscle

PAGE 21

21 CHAPTER 2 MATERIALS AND METHODS Study Sample Description Data Sources and Collection Records from 2000 2012 of 228 stranded bottlenose dolphins with photographic and morpho metric data were obtained from the Stranding Investigations Program (SIP) at Mote Marine Laboratory in Sarasota, Florida USA. Morphometric data included total length (cm), measured from the tip of the upper jaw to the fluke notch (Read et al. 1993), and total weight (kg). Photos were examined to determine if a stranded animal had a PND. Most of the stranding cases originated from Sarasota and Manatee counties, and some involved long term resident bottlenose dolphins of Saraso ta Bay, monitored by the Sa raso ta Dolphin Research Program (SDRP) through regular photographic identification surveys and capture release health assessments (Wells et al. 2004, Wells 2009). For the s tranding of MML0904, field survey photograph s of the free ranging dolphin taken eig ht and ten days prior to its necropsy were used to supplement the SIP record Data Organization Compiled data were examined for potential factors that could confound the assessment of body condition including: carcass condition, missing body parts, pregnan cy, life history category geographical variation, and ecotype. In addition, I eliminated stranding cases with insufficient data for analysis, including cases that did not meet the requirements for photographic quality. Each of these factors is described in detail below.

PAGE 22

22 Carcasses that were too decomposed at the time of examination could yield inaccurate meas urements because of bloating or tissue and fluid loss. Therefore, I excluded data for animals that were considered to have a Smithsonian Instituti on Condition C ode of 4 or 5 (Geraci and Lounsbury 2005). A Condition Code 3 carcass was exclud ed if it: 1) was marked as a late C ode 3 in the necropsy report and/or the NMFS stranding network level A form, 2) had a bloated tongue, protruding penis or dis tended vulvar region or 3) was noted as bloated or severely decomposed within the necropsy report and/o r the level A form. If any carcass w as of marginal condition I consulted the MML SIP manager All other Code 3 animals and those with C odes of 1 or 2 were included in the sample provided that they met the additional requirements described below. Moreover, animals were deleted from the sample if substantial tissue was removed, for example from shark predation or scavenging, since their weight would be inaccurate (Fig ure 2 1). R eproductive condition and life history category also influenced inclusion in the sample. All pregnant females were excluded from this study because their weights were not representative of a single individual. Fetuses, stillb orn animals, and neonates were excluded because the presence of PNDs in very young animals may be more strongly related to ontogeny than to nutritive condition. Studies that have described neonatal characteristics or analyzed the blubber depth of this age class illustrate how the presence of PNDs in neonates is likely a normal condition o f the early developmental stages of a dolphin (Cockcroft and Ross 1990, Dearolf et al. 2000, Struntz et al. 2004). s than a week old at the time of their death, including perinatal mortalities and some stillbirths,

PAGE 23

23 were on average about 107 cm (R. Wells, pers. comm., 24 July 2013). To be conservative, I considered animals with a total length 143 cm to be neonates. This neonatal length is a n estimate based on stranded bottlenose dolphins in Virginia with and without selected neonatal characteristics ( e.g. rostral hairs, floppy dorsal fins, fetal fold lines, and un erupted teeth) (Lynott 2012 ) west coast due to geographic variation in size that has been demonstrated in this species (Read et al. 1993, Stolen et al. 2002, McFee et al. 2012). Similarly, I contro lled for the two known ecotype s of bottlenose dolphins inshore and offshore. Offshore animals are larger and more robust than inshore animals (Hersh and Duffield 1990, Mead and Potter 1995 ). Therefore, I excluded dolphins that were speculated or noted as offshore in their necropsy report. Stranding cases were excluded when length or weight values were estimated rather than measured or when photos were of insufficient quality. Photo quality was dependent on the angle from which the photo was taken, focus/ clarity, contrast, body position, and the amount of the dorsa l surface included in the photo. The ideal photo centered on the post nuchal region and t he focus and contrast o f the photo were sufficient so that body landmarks were clearly visible In addition, the animal was in an upright and flat position ( i.e. posture was not affected by the surface that was supporting it). Ideal photos also included the dorsal surface from the blowhole (or the approximate location of the blowhole if it was not directly visible) to at least the point above the anterior insertion of the pectoral fin (see Figure 2 2 for examples). If the best available

PAGE 24

24 photo of an individual had one or more f eatures making a PND unrecognizable, that individual was excluded. Although I examined necropsy reports for potential confounding factors and therefore observed whether or not an animal was considered to be emaciated within some reports during the initial stages of the project, a significant amount of time passed between when I investigated the 228 bottlenose dolphin records and when I applied the PND index (see below) to the finalized sample. Therefore, I believe the PND determinations were not biased. De velopment of PND Index I defined a PND as a concavity o n the dorsal surface of the post nucha l region of a dolphin an area just caudal to the nuchal crest of the skull that can extend through the cervical region and is assoc iated with the post nuchal fat p ad ( Figure 2 2). Preliminary analysis of photos from both stranded animals from SIP records and live animals photographed during Sarasota Bay health assessments showed that identifying PNDs visually without any additional metric applied to the photo could be difficult and subjective. Individuals that seemed border line to having this trait ( for example, when a very slight depression was visible in the post nuchal region) created the most difficulty in terms of confidently identifying PNDs. Additionally, I did not know if the amount of dorsal surface caudal to the nuchal crest visible in the photos would affect the PND outcome. T o help objectively identify PNDs in photographs, I developed a technique that involve d drawing a straight horizontal line acro ss the post nuchal region. If a space was visible between the line and the dorsal surface in this region, it was considered to have a PND. For each individual in the study, I tested four different line

PAGE 25

25 drawing approaches (see below) to determine whether varying the caudal endpoints (and therefore the amount of the dorsal surface that needed to be included in the photo) would change the PND outcome. I chose the best photo (either the left or right side of the body was acceptable), based on photo quality characteristics, that included t he amount of body needed for each of four PND index methods described below If the PND outcome varied between the different methods I then compared the results by analyzing length weigh t 95% quantile regression ranges and BMI calculations. I made these comparisons to det ermine whether one metho d identified do l phin s as having PNDs that also had lower weights for a given length and lower BMI values than the PND animals identified by the o ther methods examining the results for which method most closely and consistently mirrored body condition results Cranial Starting Points Com mon to All Four Line Techniques Three cranial starting points for the lines were possible, and these possibilit ies were the same for each of the four line techniques For individuals where the nuchal crest was visible, which was generally the case for emaciated animals, I began the line at the tip of the nuchal crest. For more robust individuals where the nuchal crest was not readily visible, I used the external auditory meatus ( ear ) as a reference point and began the line at the dorsal surface directly above the ear. In most individual s, t he nuchal crest is located between the body landmarks of the eye and exter nal ear The distance between the nuchal crest and each these landmarks is affected by the age of an individual, as the distance between these landmarks increases as an individual grows. Although the distances between the nuchal crest and each of these landmarks we re not recorded in the necropsy reports, the distance between the two landmarks themselves would approximate the maximum distance between the nuchal crest and

PAGE 26

26 external ear (i.e. when the nuchal crest is located above the eye). W ithin my sample 17 females and 16 males had the measurements necessary to calculate the straight length distance between the center of the eye and external ear available in their necropsy reports For the females, the straight length distance between the center of the eye and external ear ranged from 5 .0 to 8.8 cm, excluding an extreme 10.3 cm observation for a female 171 cm in total length. For males, the straight length distance between the eye and external ear ranged from 5 .0 to 8.7 cm. Figure 2 3 shows how the dis tance between the nuchal crest and the dorsal surface directly above the ear varies between individuals, so the point at the dorsal surface above the ear will better approximate the location of the nuchal crest for some individuals than others. If both th e nuchal crest and the ear were not visible, I began the line at the dorsal surface directly above the approximate location of the ear I used my own judgment to choos e a point caudal to the eye that I thought would approximate the location of the ear and then began the line above that point Examples of the three possible cra nial start points are shown in F igure 2 4. The four line techniques described in detail below, were differentiated by the amount of body caudal to the blowhole included in the ph otograph and the caudal endpoints of their lines Method #1 T he amount of body caudal to the blowhole included in the best photo of an individual for this method var ied among i ndividuals, ranging from the dorsal surface directly above the anterior insertio n of the pectoral fin to the anterior in sertion of the dorsal fin. The caudal endpoint of the line was drawn at the highest point of the dorsal surface at or caudal to the anterior insertion of the pectoral fin (Figures 2 5 and 2 6).

PAGE 27

27 Method #2 P hotos of each individual included the dorsal surface from the blowhole to the anterior insertion of the dorsal fin. I drew the caudal endpoint of the line at the highest point of the dorsal surface between the anterior insertions of the pectoral and dorsal fins ( Figures 2 5 and 2 6). In many cases, the photo considered to be the best for Method #1 included the anterior insertion of the dorsal fin. Therefore, the photo/line combination for these instances was copied and used for Method #2 which required the ante rior insertion of the dorsal fin to be included in the photo However, some animals required different photos for Method #2 to include the anterior insertion of the dorsal fin and others did not have any photos that included the anterior insertion of the dorsal fin Method #3 P hotos of each animal included the dorsal surface from the blowhole to the posterior insertion of the pectoral fin, and I drew the caudal endpoint of the line at the dorsal surface above the posterior insertion of the pectora l fin ( Figures 2 5 and 2 6). Method #4 P hotos of each animal included the same amount of body as Method #3, and I drew the line caudally to the point directly above the anterior insertion of the pectoral fin (Figures 2 5 and 2 6). Table 1 1 summarizes the cranial and caudal end points and the amount of visible dorsal surface required for each of the four methods. Figures 2 5 and 2 6 show all four methods applied to dolphins with and without a PND.

PAGE 28

28 Body Condition Analysis Length Weight Models Comparisons of body c ondition were performed with R version 2.15.2, and Microsoft Excel 2010 was used to give starting estimates for nonlinear model fits in R. I first determined if the length and weight data were related by fitting a simple linear regression model to both th e PND and non PND weight vs. length data from the stranding sample to see if the slope significantly differed from zero. Females and males were analyzed separately because the resident dolphins in Sarasota Bay are known to be sexually dimorphic (Read et a l. 1993, Tolley et al. 1995). Residuals of the linear regressions were analyzed relative to assumptions for normality and equal variance. Then I fit a non linear weight vs. length model to all of the PND and non PND data separately for males and females using ordinary least squares (OLS) regression and the equation: (2 1 ) where TM is total mass (kg), TL is total length (cm), and a and b are parameter estimates (Innes et al. 1981, Hart et al. 2013). Starting values required by Program R for parameter estimates when fitting non linear OLS models were obtained by using values 20% different than Micros oft Excel 2010 solver estimates Based on visual comparisons, the nonlinear models for both sexes fit better than the linear models. Following the methods of Hart et al. (2013) with the quant reg package in R, I used quantile regression and Equation 2 1 to create 95% quantile ranges for the PND and non PND length weight data. I chose quantile regression because it does not assume normality or homogeneity in variance as does simple linear regre ssion (Cade and Noon 2003), and both the female and male weight vs. length regression residuals did not

PAGE 29

29 meet the simple linear regression assumptions. The starting values for the a and b parameters for the median fit (tau=0.5) of the quantile range were o btained from the R estimates from the non linear OLS regression fit. The a and b starting values for the upper (tau=0.975) and lower (tau=0.025) 95% quantile fits were obtained from the four line drawing methods, I plotted the observed weight vs. length points within the 95% quantile ranges and compared the relative locations of the observed points for animals with and without a PND. These plots allowed me to observe general differences in body condition of animals with and tely described body condition. Body Mass Index (BMI) Calculations Using the b parameter estimate from the nonlinear OLS regression fit of Equation 2 1, I calculated BMI wit h the equation: (2 2) (Hamill et al. 1995, Harwood et al. 2000, Hart et al. 2013). BMIs of PND and non PND animals were then compared for each of the four line drawing methods using a Wilcoxon rank sum test in R. I also compared boxplots of the BMIs for PND and non PND animals for each of the four methods. This BMI model allowed me to further observe both the general differences in body condition of PND vs. non PND animals and compare the four methods in terms of their PND outcomes. 95% Qua ntile Ranges for the R ecommended PND Index Using Non PND Data After determining which PND ind ex consistently reflected the body condition analys i s results and was the most flexible in terms of its application I then created 95% quant ile reference ranges using onl y the non PND data from this PND index results. I

PAGE 30

30 plotted the observed non PND and PND data on the same graph as the reference ranges. This plot differs from the other plots because the quantiles were fit to only the non P ND data as opposed to a combination of the PND and non PND data. I fit the 95% quantile ranges using only non PND data from the recommended index to see if the PND data would fall outside of the ranges of the non PND data. Comparison of BMI Values between Dolphins with a PND and Dolphins Considered to Be Emaciated I was able to determine whether or not a dolphin was considered emaciated by the professional opinion given in the necropsy reports for a subset of individuals within the sample. I then used box plots to compare the BMI values for dolphins identified as having PNDs by the recommended index with the BMI values of individuals that were considered to be emaciated I made this comparison to examine how the body condition of dolphins identified as hav ing PNDs related to the professional opinion of body condition recorded in the necropsy reports.

PAGE 31

31 Table 2 1. Summary of possible cranial start points of the line, amount of dorsal surface required in the photo, and caudal endpoints of the line for all f our methods. The three possible cranial start points for each technique were the same, but the amount of the dorsal surface included in the photo and the caudal endpoint of the lines differed for each technique. Method # Cranial s tart p oint Dorsal surfac e i ncluded in p hoto from b lowhole to: Caudal end p oint at dorsal s urface 1 1.Tip of nuchal crest if visible 2. If 1 is not possible, then dorsal surface directly above the ear if the ear is visible 3. If 1 and 2 are not possible, then dorsal surf ace above the approximate location of the ear Point between anterior insertions of the pectoral and dorsal fins (varies) Highest point of dorsal surface in the photo at or caudal to the anterior insertion of the pectoral fin 2 Anterior insertion of the dorsal fin Highest point of dorsal surface between anterior insertions of the pectoral and dorsal fins 3 Posterior insertion of the pectoral fin Point directly above posterior insertion of the pectoral fin 4 Anterior insertion of the pectoral fin Poin t directly above the anterior insertion of the pectoral fin

PAGE 32

32 Figure 2 1 Examples of wounds \ tissue loss that were thought to have a significant effect on the total weights of the animals (A and B) and wounds that caused minimal tissue loss an d were considered to not have a significant impact on total weights (C and D). A) and B ) carcasses with deep shark bites C ) an animal with a possible cut or bite, D ) an animal with propeller wounds. Both animals A and B were excluded from the sample, while C and D were kept in the sample. Photos courtesy of MML SIP. A B C D

PAGE 33

33 Figure 2 2. Examples of ideal photo quality for PND analysis, including the desired angle from which the photo was taken (perpendicular to the sagittal plane and r oughly centered on the post nuchal region, which is indicated by the red brackets), body position of the animal (upright), amount of the dorsal surface included in the photo (blowhole to at least the point above anterior insertion of the pectoral fin), an d focus/clarity, and contrast A) example 1, B) example 2 Photos courtesy of MML SIP. A B

PAGE 34

34 Fig ure 2 3 L ocation of the external ear relative to the nuchal crest for two dolphins. the vertical to the location of each of the ears. For both dolphins, the nuchal crest i s located anterior to the ear. A ) a 186 cm male, B ) a 245 cm female. Photos courtesy of MML SIP. A B

PAGE 35

35 Figure 2 4 Examples of the three possible cranial starting points to the PND lines depending on whether or not the nuchal crest or external ear was visible. A ) The tip of the nuchal crest, B) The dorsal surface directly above th e ear C ) The dorsal surface above the approximate location of the ear. Photos courtesy of MML SIP. A Tip o f Nuchal Crest B Ear C

PAGE 36

36 Figure 2 5 Examples of ea ch of the four methods applied to an animal with a PND A ) This photo and line combination was used for both method s # 1 and 2, B) Method #3, C) Method #4. Note the cranial start point for each of the lines is the same, while the amount of dorsal surface included in the photos and the caudal endpoints of the lines differ. Note also that the space between the line and dorsal surface in each of these photos (space is very small in photo C) indicates that all methods identify this animal as having a PND. Photos courtesy of MML SIP A B C

PAGE 37

37 Figu re 2 6. Examples of each of the four methods applied to an animal without a PND A ) Method #1, B) Method #2, C) Method #3, D) Method #4. Note the cranial start point for each of the lines is the same, while the amount of dorsal surface included in the photo and the caudal endpoints of the lines differ. Note also that a space be tween the line and dorsal surface does not exist in any of these photos, which indicates that all methods identify this animal as not having a PND. Photos courtesy of MML SIP. A B

PAGE 38

38 Figu re 2 6. Continued C D

PAGE 39

39 CHAPTER 3 RESULTS Summary of Length Weight Dat a Of the 228 bottlenose dolphin stranding records from 2000 2012, 20 females and 24 males met the requirements for analysis. Data suitable for the PND index and body condition analyses, including the PND outcomes for each of the four PND indices, are sum m arized in Tables 3 1 and 3 2. Simple Linear Regression Model and Assumptions The slopes of the simple linear regression weight vs. length models for females and males were both significantly different than zero (Females: slope = 1.1837, std. error = 0.1002 p<0.001, Figure 3 1; Males: slope = 1.3590, std. error = 0.1281, p<0.001, Figure 3 4). Residual analysis showed that neither the female or male models met the assumptions for normality or homogenous variance (Females: Figures 3 2 and 3 3; Males: Figures 3 5 and 3 6). Nonlinear Model Parameter Estimates and Visual Comparison of the Linear and Nonlinear Model Fits Table 3 3 lists the parameter estimates that resulted from the nonlinear ordinary least squares (OLS) regression fit of Equation 2 1 for both se xes, and Figures 3 7 and 3 8 show the nonlinear fit of this model for females and males, respectively. Visually comparing the simple linear regression model to the nonlinear OLS regression model indicates that the latter better fit the observed data for b oth sexes (see Figures 3 1, 3 4, 3 7, and 3 8).

PAGE 40

40 PND Indices Performance Assessment 95% Quantile Regression Reference Ranges Since the nonlinear models fit the data better they were used to create 95% quantile regression reference ranges. T au values and t he a and b paramete r estimates for each quantile are show in Table 3 4 Females Methods #1, 2, and 3 each classified the same six females as possessing a PND (Table 3 1, Figure 3 9) T wo females that were identified as negative for possessing a PND by Met hods #1,3, and 4 were marked as ( CBD ) for Method #2. Method #4 identified only four of the six PNDs that the other methods identified. Of th es e two animals that Method #4 did not recognize as possessing a PND one was on the lower 95% quantile line and the other was on the median line (Figure 3 9). Therefore M ethods #1 3 had similar results while Method #4 was the only index to not identify an animal on the lower 95% quantile range as possessin g a PND Males Methods #1 and 2 class ified the same seven males as having PNDs (Table 3 2, Figure 3 10). Method #3 identified nine animals as having PNDs, seven of which were shared with Methods #1 and 2. Method #4 recognized seven animals as having PNDs, five of which were the same as the other three methods. Method #3 identified the most animals below the median line as having PNDs, while Method #4 identi fied the least number of individuals below the median line as having PNDs and was the only method to identify an individual above the me dian line as having a PND (Figure 3 10).

PAGE 41

41 BMI Calculations Tables 3 1 and 3 2 list the BMI values calculated u sing Equation 2 2 and the b parameter estimate from the nonlinear OLS regression fit of Equation 2 1 Females The Wilcoxon rank sum test comparing t he female PND and non PND BMI values showed that f or each method, the BMI values for PND and non PND animals were significantly different at p<0.02 (Table 3 5) Methods # 1 and 3 had the same results with the largest W statistic at the lowest p value (Tab le 3 5). Regardless of the method, boxplots show ed that females with PNDs had lower BMI values than females without PNDs (Figure 3 11). These boxplots comparisons also demonstrated that Methods # 1 and 3 had the same results and the distance between the ir PND and non PND boxes w as the greatest (Figure 3 11). Method #4 was the only index to include a BMI value in the non PND boxplot lower than any BMI value included in the PND boxplot. Males The Wilcoxon rank sum test for the males showed that f or ea ch method, the BMI values for PND and non PND animals were significantly different at p<0.02 (Table 3 5) Methods #1 and 2 had the same results with the second largest W statistic (Method #3 had the largest) and the lowest p values (Table 3 5). Similar to the females, t he male boxplot BMI comparisons indicated that regardless of the method used, BMI values for dolphins that possessed a PND were lower than BMI values of animals that did not possess this trait (Figure 3 12). Methods #1 and 2 had the great est distance bet ween the PND and non PND boxes while Method #4 had the least distance between the PND and non PND boxes (Figure 3 12)

PAGE 42

42 95% Quantile Ranges for Method #1 Using Non PND Data For both sexes, I fit 95% quantile regression ranges using Equatio n 2 1 and only the non PND data resulting from Method #1 because this method consistently identified dolphins with relatively low weights for given lengths and dolphins with low BMI values as having PNDs in the body condition analyses (Table 3 6) It is a lso the most flexible method to apply to photos because it allows for variation in the amount of body required to draw the line. For the females, the 95% quantile regression plots showed four of the six PND observations fell below the lower quantile (Figu re3 13). The remaining two PND observations fell below the median. For the males, five of the seven PND points fell below the lower quantile, and the remaining two PND points fell below the median (Figure 3 14). Comparison of BMI Values between Dolphins with PNDs and Dolphins Considered to be Emaciated For 15 of the 20 females and for 19 of the 24 males in my sample, body condition (emaciated or not) was consistently described in the necropsy report. Eleven of the 15 females were considered emaciated wit hin the reports. Th ese 11 emaciated females included all six PNDs identified by Method #1, s o five emaciated females (45%) were not identified as having PNDs by Method #1. E leven of the 19 males were considered emacia ted within the reports. The s e 11 ema ciated males included all seven PNDs identified by Method #1 s o four emaciated males (36%) were not identified as having PNDs by Method #1. For both sexes, boxplot comparisons of BMI value s demonstrated that dolphins identified as possessing a PND by Met hod #1 fell low within the range of dolphins considered to be emaciated within the necropsy reports (Figures 3 15 and 3 16)

PAGE 43

43 Table 3 1. F emales that met the requirements for the PND index and body condition analyses, including the results for each of t he four PND indices. Stranding ID f emales : Condition c ode Date of c arcass r ecovery Total l ength (cm) Total w eight (kg) BMI #1 #2 #3 #4 MML0115 2 03 Oct 2001 146.7 42.2 1.25 0 0 0 0 MML0202 3 21 Jan 2002 189.0 81.5 1.27 0 0 0 0 MML0305 2 15 Feb 2003 17 1.0 45.5 0.92 1 1 1 0 MML0309 3 24 Feb 2003 236.0 120.0 1.06 0 0 0 0 MML0403 1 09 Mar 2004* 190.0 90.0 1.38 0 CBD *** 0 0 MML0409 2 12 May 2004 246.0 147.0 1.17 0 0 0 0 MML0412 2 02 Jul 2004 238.0 168.0 1.45 0 CBD *** 0 0 MML0413 2 04 Aug 2004 250.0 169 .0 1.29 0 0 0 0 MML0416 2 12 Sep 2004 267.0 162.0 1.05 0 0 0 0 MML0527 3 13 Sep 2005 236.0 132.5 1.17 0 0 0 0 MML0538 2 15 Dec 2005 252.0 150.5 1.13 1 1 1 0 MML0601 3 9 Jan 2006 206.0 102.5 1.28 0 0 0 0 MML0602 2 12 Jan 2006 151.0 39.0 1.08 0 0 0 0 M ML0608 1 01 Apr 2006** 205.0 93.0 1.18 0 0 0 0 MML0611 2 01 May 2006 241.0 114.5 0.96 1 1 1 1 MML0808 2 17 Dec 2008 165.0 65.0 1.43 0 0 0 0 MML0904 2 22 May 2009 235.0 123.0 1.10 1 1 1 1 MML1107 2 16 Jun 2011 245.0 130.0 1.04 1 1 1 1 MML1210 2 08 Dec 2012 266.4 214.5 1.39 0 0 0 0 MML1211 3 28 Dec 2012 169.3 47.0 0.97 1 1 1 1 *Date of rescue from entanglement of live animal. **Date of recovery of live animal that died the same day. ***CBD = Could not be determined

PAGE 44

44 Table 3 2. M ales that met the re quirements for the PND index and body condition analyses, including the results for each of the four PND indices. Stranding ID m ales : Condition c ode Date of c arcass r ecovery Total l ength (cm) Total w eight (kg) BMI #1 #2 #3 #4 MML0016 2 02 Aug 2000 224.0 6 9.0 0.20 1 1 1 1 MML0203 2 21 Feb 2002 227.0 146.5 0.41 0 0 0 0 MML0208 2 28 Feb 2002 193.0 94.0 0.41 0 0 0 0 MML0313 3 07 Apr 2003 210.0 85.5 0.29 1 1 1 0 MML0315 3 22 Apr 2003 191.0 85.5 0.38 0 0 0 0 MML0319 2 29 Apr 2003 167.0 42.0 0.27 1 1 1 1 MM L0320 3 07 May 2003 177.0 70.0 0.39 0 0 0 0 MML0330 3 26 Aug 2003 172.0 61.5 0.37 1 1 1 1 MML0338 3 30 Nov 2003 162.0 59.5 0.42 0 0 0 0 MML0339 2 14 Dec 2003 190.0 73.0 0.33 0 0 0 0 MML0404 2 03 Mar 2004 200.0 70.0 0.28 1 1 1 0 MML0501 2 24 Jan 2005 2 00.0 92.5 0.36 0 0 0 0 MML0502 2 27 Jan 2005 255.0 217.5 0.44 0 0 0 0 MML0503 2 05 Feb 2005 186.0 61.5 0.30 1 1 1 1 MML0604 2 09 Feb 2006 156.0 49.5 0.39 0 0 0 0 MML0606 2 03 Mar 2006 263.0 202.5 0.37 0 0 0 1 MML0617 3 03 Jul 2006 151.0 41.0 0.35 0 0 0 0 MML0618 2 06 Jul 2006 257.0 223.0 0.44 0 0 0 0 MML0619 2 13 Jul 2006 255.0 150.5 0.30 1 1 1 1 MML0810 2 30 Dec 2008 161.0 52.5 0.38 0 0 1 0 MML1102 3 02 Feb 2011 276.0 175.5 0.28 0 0 0 0 MML1104 2 15 Mar 2011 214.0 98.5 0.32 0 0 1 1 MML1113 2 27 Dec 2011 252.0 167.0 0.35 0 0 0 0 MML1208 3 13 Nov 2012 157.0 49.0 0.38 0 0 0 0

PAGE 45

45 Table 3 3. a an d b parameter estimates from nonlinear OLS regression fit of with corresponding standard error values. Sex Parameter e stimates* Stand ard e rror Female a: 3.93 0.62 b: 2.55 0.26 Male a: 4.46 0.64 b: 2.78 0.27 *All parameter estimates were significant at p<0.001. Table 3 4. Quantiles, tau values, and a and b parameter estimates with their standard errors for the 95% quantil e regression fit of including both PND and non PND data Sex Quantile Tau a std. error b std. error Female Upper 0.975 3.94 0.71 2.59 0.30 Median 0.5 00 3.49 0.82 2.36 0.35 Lower 0.025 4.35 0.60 2.69 0.26 Male Upper 0.975 4.55 0.26 2.86 0.11 Median 0.5 00 4.00 0.72 2.59 0.31 Lower 0.025 2.13 2.93 1.69 1.27 Table 3 5. Wilcoxon rank sum test W statistic and corresponding p values for each of the four PND indices Sex PND i ndex W s tatistic P v alue Female #1 78 0.0015 #2 66 0.0032 #3 78 0.0015 #4 57 0.016 Male #1 110 0.00054 #2 110 0.00054 #3 120 0.00098 #4 96 0.019

PAGE 46

46 Table 3 6. Quantiles, tau values, and a and b parameter estimates with their standard errors for the 95% quantile regress ion fit of including only non PND data Sex Quantile Tau a std. error b std. error Female Upper 0.975 3.94 0.59 2.59 0.25 Median 0.5 00 3.18 0.79 2.24 0.34 Lower 0.025 3.85 0.74 2.50 0.31 Male Upper 0.975 4. 55 0.41 2.86 0.18 Median 0.5 00 4.22 0.67 2.70 0.30 Lower 0.025 3.64 1.07 2.41 0.46

PAGE 47

47 Figure 3 1. S imple linear regression weight vs. length model for females. Red dashed line represents predicted total weight (kg) for a given total length (cm), and circles are observed total weight.

PAGE 48

48 Figure 3 2. Q q normality plot of studentized residuals of female weight vs. length simple linear regression demonstrating that the studentized residuals do not meet the assumption of normality Figure 3 3. Residuals vs. fitted values for female weight vs. length simple linear regression indicating that the residuals do not meet the assumption for homogenous variance.

PAGE 49

49 Figure 3 4. S imple linear regression weight vs. length model for males. R ed dashed line represents predicted total weight (kg) for a given total length (cm), and circles are observed total weight.

PAGE 50

50 Figure 3 5. Q q normality plot of studentized residuals of male weight vs. length simple linear regression showing that the s tudentized residuals do not meet the assumption of normality. Figure 3 6. Residuals vs. fitted values for male weight vs. length simple linear regression model showing that the residuals do not meet the assumption for homogenous variance

PAGE 51

51 Figure 3 7. Nonlinear OL S regression fit of to female length weight data.

PAGE 52

52 Figure 3 8. Nonlinear OL S regression fit of to male length weight data.

PAGE 53

53 Figure 3 9. 95% quantile regression ranges and observed data for females for each of the four PND indices. Plots differ only in the number of highlighted data points. Green points represent animals with PNDs. The two yellow points represent animals for which Method #2 results could not be determined. A) Method #1, B) Method #2, C) Method #3, D) Method #4 A B C D

PAGE 54

54 Figure 3 10. 95% quantile regression ranges and observed data for males for each of the four PND indices. Plots differ only in the number of highlighted data points. Green points represent animals with PNDs. A) Method #1, B) Method #2, C) Method #3, D) Method #4 D B C A

PAGE 55

55 Figure 3 11. Boxplots for female PND and non PND BMI values for each of the four PND indices. Hinges are versions of the first and third quartiles, and the line within the box is the median. The whiskers show the largest and smallest values that fall within 1.5 times the box size (Dalgaard 2008). A ) Method #1, B) Method #2, C) Method #3, D) Method #4. A B D C

PAGE 56

56 Figure 3 12. Boxplots for male PND and non PND BMI values for each of the four PND indices. Hinges o f the boxplot are versions of the first and third quartiles, and the line within the box is the median. The whiskers show the largest and smallest values that fall within 1.5 times the box size. Values outside of that range are shown separately as circle s (Dalgaard 2008) A) Method #1, B) Method #2, C) Method #3, D) Method #4. D C B A

PAGE 57

57 Figure 3 13. Female 95% quantile regr ession ranges fit to non PND data of Method #1. Green points represent animals with PNDs.

PAGE 58

58 Figure 3 14. Male 95% quan t ile regression ranges fit to non PND data of Method #1. Green point s represent animals with PNDs.

PAGE 59

59 Figure 3 15. Boxplots of female BMI values for individuals with PNDs identified by Method #1 and for individuals noted as emaciated within their necro psy reports Hinges are versions of the first and third quartiles, and the line within the box is the median. The whiskers show the largest and smallest values that fall within 1.5 times the box size (Dalgaard 2008).

PAGE 60

60 Figure 3 16. Boxplots of male BMI values for individuals with PNDs identified by Method #1 and for individuals noted as emaciated within their necropsy reports Hinges are versions of the first and third quartiles, and the line within the box is the median. The whiskers show the larg est and smallest values that fall within 1.5 times the box size Values outside of that range are shown separately as circles (Dalgaard 2008).

PAGE 61

61 CHAPTER 4 DISCUSSION Differences in the Outcomes between the Four PND Indices Bottlenose dolphin PND indices were developed to decrease the subjectivity inherent in visually assessing systematic metric applied to photos. Preliminary analysis of photos taken both during Sarasota Bay bottlenose dolphin health assessments and stranding events showed that visually classifying dolphin s with marginal depressions without applying any metric to the photos was difficult. Further more, the amount of the dorsal surface that needed to be visible in the photo in order to determine PND presence was not known. Therefore, I tested four different methods that involved drawing a horizontal line across the post nuchal region of an individual to determine if a concavity existed in this area of the body If a space was visible between the line and the dorsal surface in the post nuchal region, the animal was considered to have a PND. The four method s varied by the amount of dorsal surface they required to be in the photo and the caudal endpoint s of their lines (Table 2 1). The four indices varied in their PND determinations (Tables 3 1 and 3 2). The first three methods identified two females (MML0305 and MML0538) and two males (MML0313 and MML0404) as possessing a PND while Method #4 did not. MML0305 length weight point was on the lower 95% quantile range the median line for the quantile regression ranges fit to both the PND and Non PND data (Figure 3 9). MML for the male quantile regressio n ranges (Figure 3 10). The primary reason for these differences was that the Method #4 line did not span across enough of the dolphin body

PAGE 62

62 to incorporate the topography of the dorsal surface caudal to the anterior insertion of the pectoral fin as can be seen with MML0538 in Figure 4 1. Therefore, the Method #4 line may incorrectly identify some animals as not possessing a PND when they really do (false negative) The differences resulting from using lines that span different lengths across the dolphin dorsal surface can be seen with an upright dolphin (MML0330) in Figure 2 5. Although all four indices indicated MML0330 had a PND, the space between the line and the dorsal surface becomes less evident when Method #4 is applied in comparison to the three other methods. There were also instances where Methods #3 and/or # 4 indicated the dolphin had a PND, while Methods #1 and # 2 did not. This occurred for three males (MML0606, MML0810, MML1104) (Table 3 2). MML weight observation weight observation fell on the median line of the quantile regression ranges fit to both the PND and Non PND data (Figure 3 Method #3 and # 4 PND length weight o bservation fel l below the median line (Figure 3 10). For both MML0606 Method dorsal surfaces were very small (Figure 4 2). These spaces w ere more likely related to fine scale changes in th e topography of the dorsal surface under the lines in that region of their body than to concavities that are indicative of PND presence (false positive) For MML1104, the slightly skewed angle at which the photo was taken likely contributed to Method s #3 and # 4 identifying this dolphin with a relatively low length weight value as having a PND, while Methods #1 and # 2 did not (Figure 4 3).

PAGE 63

63 Effect of Varying the Cranial Start Points for One PND Index Although the caudal endpoints of the lines differentiated each of the four PND indices, three possible cranial start points were common to each index the nuchal crest of the skull, the point at the dorsal surface above the ear, and a point at the dorsal surface above the approximate location of the ear (Figure 2 3). Even though I chose these cranial start points because of their relative proximity to one another, the distance between them could potentially affect the PND outcome. In Figures 4 4 and 4 5, I drew the different cranial start points on two animals, MML0503 and MML1107, with Method #1 applied to them and where the nuchal crest and the ear were both visible. Since I could see the ear, I drew two cranial start points for the dorsal surface above the approximation of the ear one anterior to the ear and one caudal to the ear. For MML0503, changing the cranial start point did not change the PND outcome (Figure 4 4). For MML1107, changing the cranial start point did change the PND outcome to Non PND when the line was drawn above the point caudal to the ea r (Figure 4 5) For the other three cranial start points, the lines identified the dolphin as having a PND, although the spaces under the lines that began at the point at the dorsal surface anterior to and directly above the ear were very small. This exa mple demonstrates that if the cranial start point to the line is drawn to o far in the caudal direction, the PND outcome may be a false negative. In general, when the space between the line and dorsal surface is small as in this example, very small changes to the l ine can affect the PND outcome. When drawing the cranial start point, I recommend using the tip of the nuchal crest whenever it is visible If the ear is visible, I suggest consistently drawing the cranial start point at either the dorsal surfa ce above the ear or in between the eye and

PAGE 64

64 the ear as the nuchal crest is most often located at the dorsal surface at some point between the eye and the e ar If both the nuchal crest and the ear are not visible, I recommend drawing the point at the dorsa l surface above the approximate location of the ear using the eye as a reference point (see methods above) but fa voring the anterior direction. I would favor the anterior direction because the nuchal crest is rarely seen caudal to the ear in adults. B ody Condition Comparison between Dolphins with and without PNDs T o determine if post nuchal depressions (PNDs) could be used as an indicator of poor body condition in bot tlenose dolphins, I compared length weight data of stranded indiv iduals with and with out PNDs for each of the four PND index methods I analyzed 95% quantile regression ranges fit to length weight data of both PND and non PND an imals for all four PND indices by sex. This analysis showed that regardless of the index used to identify PNDs, the length weight observations of animals with PNDs generally fell lower in the 95% quantile range than those that did not have PNDs ( Figure s 3 9 and 3 10). Similarly, comparison of BMI values calculated for animals with and w ithout PNDs showed those tha t display PNDs had lower BMI values for all PND indices used ( Figure s 3 11 and 3 12). These findings support the common belief among bottlenose dolphin trainers and researchers that animals with PNDs are in relatively poor body condition compared to those that do not have this trait ( Struntz et al. 2004, Dunkin et al. 2005, Fair et al. 2006, Yordy et al. 2010 ). Studies on harbor porpoises ( Phocoena phocoena ), have also referred to PNDs as indicative of emaciation ( Kastelein, van Battum 1990, Cox et al. 1 998, Koopman et al. 2002 ). Kastelein and Van Battum (1990) describe d PNDs as a sign of emaciation that occurs after depressions lateral to the dorsal fin can be seen (i.e. atrophied epaxial

PAGE 65

65 musculature) In other words, they described depressions lateral to the dorsal fin in individuals as an initial indicator of weight loss that is followed by the presence of PNDs in more emaciated individuals Cox et al. (1998) used PNDs and atrophied epaxial musculature as signs of emaciation to compare the body condi tion of stranded harbor porpoises with and without signs of entanglement They found that the p ercentage of emaciated porpoises in the non entangled sample was significantly greater than in the entangled sample. Therefore, most of the entangled porpoises were in robust condition. Since a greater proportion of the non entangled animals were emaciated, they suggested that the non entangled animals may have died of starvation. Studies involving the photographic analysis of the post nuchal region of other cetaceans have also shown that animals with concavities ar e in relatively poor body condition compared to individuals without concavities. In a study of the North Atlantic right whale, Pettis et al. (2004) showed females had a significantly higher body c ondition score (higher score = deeper concavity in the post nuchal region) during the years that they were supporting a calf. In addition, animals that were presumed to be dead (were not sighted for five consecutive years) had significantly higher body co ndition scores in comp arison to those that were considered to be living (sighted at least once in five consecutive years) Similarly, Bradford et al. (2012) showed that the body condition of western gray whale females that were lactating was significantly worse than other whales, and the body condition of weaning calves was significantly better than other whales. Although their study used three regions of the body to score body condition, the post nuchal region had the most influence in their scoring tec hnique.

PAGE 66

66 Recommended PND Index Of the four indices, I recommend Method #1 for three reasons : 1) it is flexible in terms of the amount of dorsal surface caudal to the blowhole that needs to be included in the photo ( can vary from the dorsal surface at the point above the anterior insertion of the pectoral fin to the anterior insertion of the dorsal fin ); 2) the animals that it identified as having PNDs had length weight observations that fell relatively low within the 95% quantile ranges for both sexes ( Fi gure s 3 9 and 3 10); and 3) the animals it identified as having PNDs had the greatest difference in BMI values from those that did not h ave PNDs compared to the other three methods ( Figure s 3 11 and 3 12). Although other PND indices had similar or better results in one or two of these aspects, Method #1 is the only approach to simultaneously be flexible and have good body condition analyses results As the recommended approach, non PND length weight data from Method #1 were used to fit 95% quantile regres sion reference ranges. Both the non PND and PND observations were then plotted with these reference ranges, and four of six female PND observations and five of seven male PND observations fell below the 95% quantile range of the non PND data ( Figure s 3 13 and 3 14). In addition, all of the dolphins identified as possessing PNDs by Method #1 were classified by the necropsy teams as emaciated, and they had BMI values within the lower range of BMIs of emaciated individuals (Figures 3 15 and 3 16). These dat a further demonstrate that the body condition of animals with PNDs was poor in comparison to the body condition of individuals without PNDs. Furthermore, t he c auses of death (CODs) given in the necropsy reports listed emaciation or malnutrition as a contr ibuting factor for seven of the 13 dolphins identified as having PNDs by Method #1 ( Table 4 1 ) T he remaining six

PAGE 67

67 individuals were documented as emaciated elsewhere within the necropsy report but this was not necessarily considered the cause of death Overall, the PND index reduced the subjectivity of assigning PNDs to individual animals by providing a specific metric to evaluate. In addition, the approach is simple and does not require any specialized software or expertise, facilitating its applicati on in funding limited locations. My hope is that if adopted, this simple index will aid in the development of standard approaches to assessing PND presence, allowing for greater comparison of this condition spatially and temporally. Seasonal Effect on P ND Occurrence Seasonality could potentially play a role in the occurrence of PNDs in wild bottlenose dolphin populations. The waters in Sarasota Bay vary seasonally from less than 13 to 35 o C, and Sarasota Bay dolphin residents decrease their blubber thick ness by about 38% from the winter to the summer (Wells et al. 2009). Given that blubber thi ckness in the post nuchal region of a dolphin most likely contributes to the external expression of a PND (see introduction), P NDs may be more visible in summer tha n in winter. T hree of the six female s with Method #1 PNDs stranded in May or June, and t he other three stranded in December or February (Figure 4 1 A ). One male with a Method #1 PND stranded in February, three stranded in March or April and the remaining three stranded in July or August (Figure 4 2A). These results demonstrate that stranded dolphins with PNDs occur across seasons However, given that blubber thickness varies by season, I recommend comparing body condition measurements of wild dolphins w ith and without PNDs within a season, if sample size permits.

PAGE 68

68 Applying PND Index to Field Photos The major goal of this entire study was to provide a simple and non invasive technique that can be applied to field photographs of dolphins to monitor the co ndition of individu als. Although I us ed stranding photos to create the PND index, the index can readily be applied to field photos as well. Figure 4 7 show s the a pplication of Method #1 to a field photo of female MML0904 (Table 3 1) taken eight days befo re her carcass was reco vered MML0904 was considered emaciated by the necropsy team. with good focus and contrast, perpendicular to the sagittal plane, with the animal in a relati vely flat position, and includin g the dorsal surface almost to the anterior insertion of the dorsal fin. The evident space between the line and the dorsal surface in this photo indicates a PND. Photos taken in the field will not always be as good as in Figure 4 7 For example, a flat horizontal posture of the dolphin and a perpendicular angle of the photo may be difficult to obtain. Figure 4 8 shows MML0904 photographed at various postures eight and ten days before its carcass was recovered. Although Method #1 indicates that she ha s a PND regardless of the posture she is in, the space between the line and the dorsal surface is more evident when her posture is flat and horizontal as opposed to at an angle. Furthermore, photos may not always be taken at angles perpendicular to a dolp in Figure 4 9, MML0907 a 247 cm female not included in the final sample of this study because of a shark bite is photographed abou t a month and a half before her necropsy at a less appropriate angle. However, a sligh t space is still noticeable between the line and the dorsal surface when Method #1 is applied and therefore MML0907 is considered to possess a PND. MML0907 was considered emaciated at necropsy.

PAGE 69

69 The quality of photos tak en in the field varies. For Meth od #1 to be successfully applied, researchers should try to obtain photos that meet the photo quality guidelines provided Furthermore, to increase accuracy when assessing PND presence, I suggest including as much dorsal surface in the photo as possible a s far caudally as the anterior insertion of the dorsal fin Both f alse negatives and positives can result when only the dorsal surface from the blowhole to the anterior insertion of the pectoral fin is included in the photo (see discussion of Method #4 ab ove) Additionally, the cranial start point for Method #1 may not always be visible or obvious in field photo crest, eye, and ear are all not noticeable in a photograph, I suggest using the blowhole as a reference point and starti ng the line a bout 5 cm caudal to the blowhole, which should roughly approximate the site of the nuchal crest. Future Research Future research of bottlenose dolp hin PNDs should investigate if relationship s exist between PND presence and 1) reproductive clas ses and 2) survival rates for the population as found by Pettis et al. (2004) and Bradford et al. (2012) If PND presence is found to be related to survival rate, I recommend developing PND proportion baselines for populations where photo identification studies are underway. This data can e ventually be used as an indicator of stock trends when considering the overall status of a stock PND proportion in relation to baseline values can then be included in NFMS Stock Assessment Reports. Another area of future study that would benefit the understanding of PNDs in bottlenose dolphins is the physiology behind the depletion of the post nuchal fat pad. Koopman et al. (2002) measured blubber thickness and adipocyte number and size and reported that differ ences in structure and function of the blubber exist between regions

PAGE 70

70 of the porpoise body They concluded that the inner thorax blubber functions as the primary metabolic energy source for these animals, whereas blubber in the tailstock primarily function s as a structural streamlining component important for locomotion. To the best of my knowledge, no study has measured the presence and \ or morphology of the post nuchal fat pad in porpoises of differing body condition. Struntz et al. (2004) recognized the value of examining blubber across functionally distinct body regions at different stages of bottlenose dolphin development. Studies of bottlenose dolphin blubber and the post nuchal fat pad that define the relative importance of regions in terms of energ y storage would be very helpful. Conclusions I have shown that stranded bottlenose dolphins with PNDs consistently had lower length weight and BMI values than individuals without this trait supporting the use of PND as an indicator of body condition. The visual assessment of PNDs using photos of animals in the field provides a simple and non invasive tool for researchers to moni tor individual body condition and could easily be incorporated into ongoing bottlenose dolphin photographic identification studi es at many sites within the range of this spec i es

PAGE 71

71 Table 4 1. Causes of death for do l phin s identified as having PNDs by Method #1. Sex Stranding ID Cause of d eath MML0305 Female Emaciation and mild chronic bronchitis with Halocercus MML0538 Femal e Natural, possible intoxication, red tide results pending MML0611 Female Human interaction, fisheries, foreign body, ingestion (hook and line); emaciation; foreign body (stingray barb) lung; trauma fractured processes of spine; skin lesions MML0904 Fe male Natural, malnutrition, lobomycosis MML1107 Female Emaciation noted grossly and myocardial fibrosis observed histologically. Combined with the worn and missing teeth also noted grossly, death may have been in part due to age related generalized decl ine MML1211 Female Presumed to be maternal separation as her mother died due to fishery interaction approximately 20 days MML0016 Male Presumed severe bacterial pneumonia MML0313 Male Human r elated, fisheries ( goosebeak) with aspiration pneumonia MML0319 Male Natural, foreign body ingestion (rock) esophageal obstruction, malnutrition, infection pneumonia (fungal) MML0330 Male Natural, multiple skeletal anomalies, emaciation MML0404 Male Natural, foreign b ody perforation (catfish spine) of lung, diaphragm, stomach, and intestine MML0503 Male Natural, foreign body, catfish spines, esophagus, stomach/spleen, lung perforations MML0619 Male Human interaction, fisheries, foreign body, ingestion hooks and lin e; emaciation; pre mortem shark bites

PAGE 72

72 Figure 4 1. Example of how the PND outcomes differed between Methods #1 and #4 using MML0538. A) Method #1 which indicates MML0538 has a PND, B ) Method #4 which indicates MML0538 does not have a PND. Ph otos courtesy of MML SIP. A B

PAGE 73

73 Figure 4 2. Examples of dolphins that were indicated as having PNDs by Methods #3 or 4 but not by Methods #1 and 2. A ) Method #3 indicates male MML0810 has a PND, B) Method #4 indicates male MML0606 has a PND. Photos c ourtesy of Florida Fish and Wildlife Conservation Commission Marine Mammal Pathobiology Lab and MML SIP. A B

PAGE 74

74 Figure 4 3. Male MML1104, where Methods #3 and 4 indicated it had a PND, whi le Methods #1 and 2 did not. A ) Methods #1 and 2 (same photo and line combination) B) Method #3, C ) Method #4. Note, the photo of this dolphin was taken at a slightly skewed angle. Photos courtesy of MML SIP. A B C

PAGE 75

75 Figure 4 4 Method #1 applied to MML0503 using four different cranial start points: A) at th e nuchal crest, B) above a point slightly anterior to the ear, C) above the ear, and D) above a point slightly caudal to the ear. The caudal endpoints in each photo are the same. Cranial start points in B and D were drawn as possible approximations to th e location of the point above the ear that could have been made had the ear not been visible. In all four instances, Method #1 shows the animal has a PND, with a space visible between the line and the dorsal surface. Photos courtesy of MML SIP. A B C D

PAGE 76

76 F igure 4 5 Method #1 applied to MML1107 using four different cranial start points: A) at the nuchal crest, B) above a point slightly anterior to the ear, C) above the ear, and D) above a point slightly caudal to the ear. The caudal endpoints in each pho to are the same. Cranial start points in B and D were drawn as possible approximations to the location of the point above the ear that could have been made had the ear not been visible. For A C, Method #1 shows the animal has a PND, with a small space vi sible between the line and the dorsal surface. For D, this method shows the animal does not have a PND. Photos courtesy of MML SIP. A B

PAGE 77

77 Figure 4 5 Continued C D

PAGE 78

78 Figure 4 6 Number of individuals with and without PNDs as identified by Method #1 by month for all years pooled together. A) Females, B) Males. B A

PAGE 79

79 Figure 4 7 MML0904 taken in the field eight days before its necropsy. The space between the dorsal surface and the red line show the animal has a PND Photo courtesy of the SDRP. Sarasota Dolphin Research Program; NMFS permit # 522 1785

PAGE 80

80 Fi gure 4 8. Method #1 applied to photos of the female MML0904 with different postures. A) and B) Photos taken 10 posture, Method #1 identified a PND i n each photo; however, the space between the line and dorsal surface is more noticeable when the animal is in a flat position (B and D) compared to an angled position (A and C). Photos courtesy of the SDRP. A B C D Sarasota Dolphin Research Program ; NMFS permit # 522 1785

PAGE 81

81 Figure 4 9 Female, M ML0907 taken at a skewed angle in the field about a month and a half before its necropsy. Although this animal was not included in the body condition analysis because its carcass had a significant shark bite, the photo exemplifies how Method #1 can be ap plied to a photo taken at a less desira ble angle (not perpendicular to the sagittal plane). The space between the line and the dorsal surface indicat es that this animal has a PND. Photo courtesy of the SDRP. Sarasota Dolphin Research Program; NMFS permit # 522 1785

PAGE 82

82 LIST OF REFERENCES Bradford, A. L., D. W. Weller, A. E. Punt, Y. V. Ivashchenko, A. M. Burdin, G. R. Vanblaricom and R. L. Brownell, Jr. 2012. Leaner leviathans: body condition variation in a critically endangered whale population. Journal of Mammalogy 93:251 266. Cade, B. S. and B. R. Noon. 2003. A gentle introduction to quantile regre ssion for ecologists. Frontiers in Ecology and the Environment 1:412 420. Caon, G., C. B. Fialho and D. Danilewicz. 2007. Body fat condition in franciscanas ( Pontoporia blainvillei ) in Rio Grande do Sul, Southern Brazil. Journal of Mammalogy 88:1335 1341. Christiansen, F., G. A. Vikingsson, M. H. Rasmussen and D. Lusseau. 2013. Minke whales maximise energy storage on their feeding grounds. Journal of Experimental Biology 216:427 436. Cockcroft, V. G. and G. J. B. Ross. 1990. Observations on the early deve lopment of a captive bottlenose dolphin calf. Pages 461 478 in S. Leatherwood and R.R. Reeves, ed s The bottlenose dolphin. Academic Press, Inc., San Diego, CA. Cox, T.M., A.J. Read, S. Barco, J. Evans, D.P. Gannon, H. N. Koopman, W.A. McLellan, K. Murra y, J. Nicholas, D. A. Pabst, C.W. Potter, W. M. Swingle, V.G. Thayer, K.M. Touhey, and A.J. Westgate. 1998. Documenting the bycatch of harbor porpoises, Phocoena phocoena in coastal gillnest fisheries from stranded carcasses. Fishery Bulletin 96: 727 734. Dalgaard, P. 2008. Introductory statistics with R Second Edition Springer Science+Business Media, LLC, New York, NY Dearolf, J. L., W. A. McLellan, R. M. Dillaman, D. Frierson and D. A. Pabst. 2000. Precocial development of axial locomotor muscle in bo ttlenose dolphins ( Tursiops truncatus ). Journal of Morphology 244:203 215. Dunkin, R. C., W. A. McLellan, J. E. Blum and D. A. Pabst. 2005. The ontogenetic changes in the thermal properties of blubber from Atlantic bottlenose dolphin Tursiops truncatus Jo urnal of Experimental Biology 208:1469 1480. Dunkin, R. C., W. A. Mc L ellan, J. E. Blum and D. A. Pabst. 2010. The buoyancy of the integument of Atlantic bottlenose dolphins ( Tursiops truncatus ): Effects of growth, reproduction, and nutritional state. Marin e Mammal Science 26:573 587. Evans, K., M. A. Hindell and D. Thiele. 2003. Body fat and condition in sperm whales, Physeter macrocephalus from southern Australian waters. Comparative Biochemistry and Physiology a Molecular & Integrative Physiology 134:847 862.

PAGE 83

83 Fair, P. A., T. C. Hulsey, R. A. Varela, J. D. Goldstein, J. Adams, E. S. Zolman and G. D. Bossart. 2006. Hematology, serum chemistry, and cytology findings from apparently healthy Atlantic bottlenose dolphins ( Tursiops truncatus ) inhabiting the estu arine waters of Charleston, South Carolina. Aquatic Mammals 32:182 195. Fowler, C. W. and D. B. Siniff. 1992. Determining population status and the use of biological indices in the management of marine mammals. Pages 1025 1037 in D.R. McCullough and R.H. Barrett, eds. Wildlife 2001: populations. Elsevier Applied Science, London. Geraci, J. R. and V. J. Lounsbury. 2005. Marine mammals ashore: a field guide for strandings Second Edition National Aquarium in Baltimore Inc Baltimore, MD. Gomez Campos, E. A. Borrell and A. Aguilar. 2011. Assessment of nutritional condition indices across reproductive states in the striped dolphin ( Stenella coeruleoalba ). Journal of Experimental Marine Biology and Ecology 405:18 24. Gulland, F. M. D. and A. J. Hall. 2007. Is marine mammal health deteriorating? Trends in the global reportin g of marine mammal disease. EcoH ealth 4:135 150. Hammill, M. O., M. C. S. Kingsley, G. G. Beck and T. G. Smith. 1995. Growth and condition in the northwest Atlantic harp seal. Canadian Jo urnal of Fisheries and Aquatic Sciences 52:478 488. Hart, L. B., R. S. Wells and L. H. Schwacke. 2013. Reference ranges for body condition in wild bottlenose dolphins Tursiops truncatus Aquatic Biology 18:63 68. Harwood, L. A., T. G. Smith and H. Melling. 2000. Variation in reproduction and body condition of the ringed seal ( Phoca hispida ) in western Prince Albert Sound, NT, Canada, as assessed through a harvest based sampling program. Arctic 53:422 431. Haug, T., U. Lindstrm and K. T. Nilssen. 2002. Vari ations in minke whale ( Balaenoptera acutorostrata ) diet and body condition in response to ecosystem changes in the Barents Sea. Sarsia 87:409 422. Hersh, S. L. and D. A. Duffield. 1990. Distinction between Northwest Atlantic offshore and coastal bottlenos e dolphins based on hemoglobin profile and morphometry. Pages 129 139 in S. Leatherwood and R.R. Reeves, eds. The bottlenose dolphin. Academic Press, Inc., San Diego, CA. Ichii, T., N. Shinohara, Y. Fujise, S. Nishiwaki and K. Matsuoka. 1998. Interannual changes in body fat condition index of minke whales in the Antarctic. Marine Ecology Progress Series 175:1 12. Innes, S., R. E. A. Stewart and D. M. Lavigne. 1981. Growth in northwest Atlantic harp seals Phoca groenlandica Journal of Zoology 194:11 24.

PAGE 84

84 Ka stelein, R.A., R. van Battum. 1990. The relationship between body weight and morphological measurements in Harbour porpoises ( Phocoena phocoena ) from the North Sea. Aquatic Mammals 16.2:48 52. Koopma n, H. N., D. A. Pabst, W. A. McL ellan, R. M. Dillaman and A. J. Read. 2002. Changes in blubber distribution and morphology ass ociated with starvation in the h arbor porpoise ( Phocoena phocoena ): evidence for regional differences in blubber structure and function. Physiological and Biochemical Zoology 75:498 512. Kuiken, T., P. M. Bennett, C. R. Allchin, J. K. Kirkwood, J. R. Baker, C. H. Lockyer, M. J. Walton and M. C. Sheldrick. 1994. PCBs, cause of death and body condition in harbor porpoises ( Phocoena phocoena ) from British waters. Aquatic Toxicology 28:13 28. Lockyer, C. 1986. Body fat condition in Northeast Atlantic fin whales, Balaenoptera physalus and its relationship with reproduction and food resource. Canadian Journal of Fisheries and Aquatic Sciences 43:142 147. Lockyer, C. H., L. C. McC onnell and T. D. Waters. 1985. Body condition in terms of anatomical and biochemical assessment of body fat in North Atlantic fin and sei whales. Canadian J ournal of Zool ogy 63:2328 2338. Lynott, M.C. 2012. Life history of stranded bottlenose dolphins ( Tursiops truncatus ) in Virginia. M .S. thesis. Old Dominion University, Norfolk, VA. 64 p p Marine Mammal Protection Act of 1972, as amended through 2007. 16 U.S.C. 1361 1423, October 21, 1972. Available at http://www.nmf s.noaa.gov/pr/laws/mmpa/ McF ee, W. E., J. D. Adams, P. A. Fair and G. D. Bossart. 2012. Age distribution and growth of two bottlenose dolphin ( Tursiops truncatus ) populations from capture release studies in the Southeastern United States. Aquatic Mammals 38:17 30. Mead, J. G. and C. W. Potter. 1995. Recognizing two populations of the bottlenose dolphin ( Tursiops truncatus ) of f the Atlantic coast of North America mo rphologic and ecologic considerations. IBI Reports No. 5: 31 44. Miller, C. A., P. B. Best W. L. Perryman, M. F. Baumgartner and M. J. Moore. 2012. Body shape changes associated with reproductive status, nutritive condition and growth in right whales Eubalaena glacialis and E. australis Marine Ecology Progress Series 459:135 156. Miller, C. A ., D. Reeb, P. B. Best, A. R. Knowlton, M. W. Brown and M. J. Moore. 2011. Blubber thickness in right whales Eubalaena glacialis and Eubalaena australis related with reproduction, life history status and prey abundance. Marine Ecology Progress Series 438:2 67 283.

PAGE 85

85 Montie, E. W., S. R. Garvin, P. A. Fair, G. D. Bossart, G. B. Mitchum, W. E. Mcfee, T. Speakman, V. R. Starczak and M. E. Hahn. 2008. Blubber morphology in wild bottlenose dolphins ( Tursiops truncatus ) from t he Southeastern United States: I nfluence of geographic location, age class, and reproductive state. Journal of Morphology 269:496 511. NOAA Fisheries O ffice of P rotected R esources (a). [Internet]. 20May2013 [cited 17Jun2013] Marine mammal unusual mortality ev ents. A vailable at http://www.nmfs.noaa.gov/pr/health/mmume/ NOAA Fisheries O ffice of P rotected R esources (b). [Internet]. 1 7 Ju l 2013 [cited 23 Ju l 2013] 2010 2013 Cetacean u nusual mortality event in the northern Gulf of Mexico. A vailab le at http://www.nmfs.noaa.gov/pr/health/mmume/ cetacean_gulfofmexico2010.htm Perryman, W. L. and M. S. Lynn. 2002. Evaluation of nutritive condition and reproductiv e status of migrating gray whales ( Eschrichtius robustus ) based on analysis of photogrammetric data. Journal of Cetacean Research and Management 4:155 164. Pettis, H., R. Rolland, P. Hamilton, S. Brault, A. Knowlton and S. Kraus. 2004. Vis ual health assessment of North Atlantic right whales ( Eubalaena glacialis ) using photographs. Canadian J ournal of Zool ogy 82:8 19. Read, A. J. 1990. Estimation of body condition in harbor porpoises, Phocoena phocoena Can adian J ournal of Zool ogy 68:1962 1 966. Read, A. J., R. S. Wells, A. A. Hohn and M. D. Scott. 1993. Patterns of growth in wild bottlenose dolphins, Tursiops truncatus Journal of Zoology 231:107 123. Stevenson, R. D. and W. A. Woods, Jr. 2006. Condition indices for conservation: new uses f or evolving tools. Integrative and Comparative Biology 46:1169 1190. Stolen, M. K., D. K. Odell and N. B. Barros. 2002. Growth of bottlenose dolphins ( Tursiops truncatus ) from the Indian River Lagoon System, Florida, USA. Marine Mammal Science 18:348 357. Struntz, D. J., W. A. McL ellan, R. M. Dillaman, J. E. Blum, J. R. Kucklick and D. A. Pabst. 2004. Blubber development in bottlenose dolphins ( Tursiops truncatus ). Journal of Morphology 259:7 20. Tolley, K. A., A. J. Read, R. S. Wells, K. W. Urian, M. D. Sc ott, A. B. Irvine and A. A. Hohn. 1995. Sexual dimorphism in wild bottlenose dolphins ( Tursiops truncatus ) from Sarasota, Florida. Journal of Mammalogy 76:1190 1198.

PAGE 86

86 Waring, G. T., E. Josephson, K. Maze Foley and P. E. Rosel. 2012. US Atlantic and Gulf of Mexico Marine Mammal Stock Assessments -2011. NOAA Tech Memo NMFS NE 221 319 pp Available at http://www.nefsc.noaa.gov/publications/ tm/tm221/ Wells, R. S. 2009. Learning from nature: bot tlenose dolphin care and husbandry. Zoo Biology 28:635 651. Wells, R. S., H. L. Rhinehart, L. J. Hansen, J. C. Sweeney, F. I. Townsend, R. Stone, D. R. Casper, M. D. Scott, A. A. Hohn and T. K. Rowles. 2004. Bottlenose dolphins as marine ecosystem sentinel s: developing a health monitoring system. EcoHealth 1:246 254. Yord y, J. E., D. A. Pabst, W. A. McL ellan, R. S. Wells, T. K. Rowles and J. R. Kucklick. 2010. Tissue specific distribution and whole body burden estimates of persistent organic pollutants in t he bottlenose dolphin ( Tursiops truncatus ). Environmental Toxicology and Chemistry 29:1263 1273.

PAGE 87

87 BIOGRAPHICAL SKETCH Mar y Kathleen Gryzbek ha s ha d a passion for wildlife and conservation from a very young age. This passion drove her to major in b iology an d Spanish at Butler University, where she obtained her B.S. in 2010. Shortly afterwards, she interned with the Sarasota Dolphin Research Program for about half of a year. She was also an intern with the Spotted Eagle Ray Project at Mote Marine Laboratory for about three months. After these internship experiences, Mary began program in the D epartment of Wildlife Ecology and Conservation at the University of Florida in January 2011. She became a staff member of the Sarasota Dolphin Resea rch Program as a graduate student through the support of the Chicago Zoological Society that same year. In August 2013, Mary graduated with a Mas ter of Science degree. She plan s to continue her career focus ing on wildlife ecology conservation public ou treach, and education