USF Libraries
USF Digital Collections

Group fission-fusion dynamics and communication in the bottlenose dolphin (Tursiops truncatus)

MISSING IMAGE

Material Information

Title:
Group fission-fusion dynamics and communication in the bottlenose dolphin (Tursiops truncatus)
Physical Description:
Book
Language:
English
Creator:
Quintana-Rizzo, Ester
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Grouping patterns
Separations
Whistle
Echolocation
Sarasota Bay
Dissertations, Academic -- Marine Science -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: The bottlenose dolphin exhibits a fission-fusion social structure characterized by temporary associations lasting from minutes to hours. Although social structure has been described for some dolphin communities, the selective pressures affecting fission-fusion patterns and their consequences on dolphin communication are not well understood. The goals of the present study were three-fold: 1) to quantify the rate with which fission-fusion occurred and identify the selective pressures influencing an individual's decision to leave and join a temporary group; 2) to examine the communication signals produced during temporary separations; and 3) to estimate the distances over which dolphins could remain in acoustic contact while separated.^ ^^^^It was found that a dolphin's decision to join or leave a group was related to social considerations such as the class of individual encountered (e.g., mothers with calves, adult single females, adult males, and juveniles) as dolphins move in different environments. The decision was also influenced by ecological characteristics such as the habitat where a dolphin was found. The two aspects in turn determined the rate of fission-fusion. Mothers with calves regularly using deep waters had high rates of fission-fusion. Those females encountered other females in the same reproductive condition frequently and associated with them. In contrast, mothers with calves using shallow waters had lower fission-fusion rates. Those females encountered juvenile dolphins often but they did not associate with them frequently.^ Temporarily separated dolphins did not always produce the sounds typically used for long-distance communication, and sometimes they did not use any detectable acoustic signal to find each other. On average, this absence of communication occurred at distances less than 50 m. When both whistles and echolocation produced, they were apparently involved in maintaining contact between mothers and their calves and other associates. Estimates of active spaces defined by whistle transmission indicated that communication range varied between habitats. Shallow seagrass areas had the smallest active space while channels had the greatest active space. Findings indicated that the distances over which dolphins remain in acoustic contact and can be considered members of groups are much greater than has been described from observations of dolphin spacing and activity alone.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2006.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Ester Quintana-Rizzo.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 184 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001920217
oclc - 187938606
usfldc doi - E14-SFE0001841
usfldc handle - e14.1841
System ID:
SFS0026159:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001920217
003 fts
005 20080107130346.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 080107s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001841
035
(OCoLC)187938606
040
FHM
c FHM
049
FHMM
090
GC11.2 (ONLINE)
1 100
Quintana-Rizzo, Ester.
0 245
Group fission-fusion dynamics and communication in the bottlenose dolphin (Tursiops truncatus)
h [electronic resource] /
by Ester Quintana-Rizzo.
260
[Tampa, Fla] :
b University of South Florida,
2006.
3 520
ABSTRACT: The bottlenose dolphin exhibits a fission-fusion social structure characterized by temporary associations lasting from minutes to hours. Although social structure has been described for some dolphin communities, the selective pressures affecting fission-fusion patterns and their consequences on dolphin communication are not well understood. The goals of the present study were three-fold: 1) to quantify the rate with which fission-fusion occurred and identify the selective pressures influencing an individual's decision to leave and join a temporary group; 2) to examine the communication signals produced during temporary separations; and 3) to estimate the distances over which dolphins could remain in acoustic contact while separated.^ ^^^^It was found that a dolphin's decision to join or leave a group was related to social considerations such as the class of individual encountered (e.g., mothers with calves, adult single females, adult males, and juveniles) as dolphins move in different environments. The decision was also influenced by ecological characteristics such as the habitat where a dolphin was found. The two aspects in turn determined the rate of fission-fusion. Mothers with calves regularly using deep waters had high rates of fission-fusion. Those females encountered other females in the same reproductive condition frequently and associated with them. In contrast, mothers with calves using shallow waters had lower fission-fusion rates. Those females encountered juvenile dolphins often but they did not associate with them frequently.^ Temporarily separated dolphins did not always produce the sounds typically used for long-distance communication, and sometimes they did not use any detectable acoustic signal to find each other. On average, this absence of communication occurred at distances less than 50 m. When both whistles and echolocation produced, they were apparently involved in maintaining contact between mothers and their calves and other associates. Estimates of active spaces defined by whistle transmission indicated that communication range varied between habitats. Shallow seagrass areas had the smallest active space while channels had the greatest active space. Findings indicated that the distances over which dolphins remain in acoustic contact and can be considered members of groups are much greater than has been described from observations of dolphin spacing and activity alone.
502
Dissertation (Ph.D.)--University of South Florida, 2006.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 184 pages.
Includes vita.
590
Adviser: Joseph J. Torres, Ph.D.
653
Grouping patterns.
Separations.
Whistle.
Echolocation.
Sarasota Bay.
690
Dissertations, Academic
z USF
x Marine Science
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1841



PAGE 1

Group Fission-Fusion Dynamics and Comm unication in the Bottlenose Dolphin ( Tursiops truncatus ) by Ester Quintana-Rizzo A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Marine Science University of South Florida Co-Major Professor: Joseph J. Torres, Ph.D. Co-Major Professor: Randall S. Wells, Ph.D. David A. Mann, Ph.D. Colin A. Chapman, Ph.D. Peter L. Tyack, Ph.D. Date of Approval: October 4, 2006 Keywords: grouping patterns, separations, whistle, echolocation, Sarasota Bay Copyright 2006 Ester Quintana-Rizzo

PAGE 2

ACKNOWLEDGMENTS I am very grateful to my supervisor y committee for helping me to develop and conduct a project that I had so much fun doin g. My major professor, Jose Torres, was a mentor, and most importantly, a friend. My co-advisor, Randall Wells, opened the door and guided the way to the exciting world of the bot tlenose dolphin. I am grateful to David Mann for his patience, advice, support, and positive attitude at all times. Colin Chapman was a constant source of support and was instrumental in developing the original idea for the project. I thank Peter Tyack for providing very important feedback on the projects structure. Many field assistants and volunteers made my field work possible: Kate Ciembronowicz, Jel Bellucci, Alison Boler, Ch ristine Craven, Lucia de la Paz, Meagan Dunphy-Daly, Ida Eskerson, Lindsey Fenderson Jennifer Manson, Lu Lu, Marde McHenry, Katie McHugh, Minelia Miravete, Kathleen M ohning, Kelly Price, Eleanor Stone, Maki Tabuchi, and Jessica Weiss. Graduate students a nd staff of the Sarasota Dolphin Research Program provided helpful information on the wher eabouts of my focal animals while in the field. Janet Gannon did the figure of the study area and arranged so me data for fast computer analysis. Jay Sprinkel provided statistical advice. Elizabeth Fisher, Damon Gannon, Mandy Hill, Laura Sirot, and Kristen Weiss provided helpful comments on the dissertation. Funding for the research was provided by fellowships from the College of Marine Science, University of South Florida, to EQR, and a grant fr om the NOAA Fisheries to RSW. Research was conducted under NMFS Scientific Research Permit No. 522-1569 issued to RSW.

PAGE 3

i TABLE OF CONTENTS List of Tables iv List of Figures vi Abstract x 1. Introduction 1 1.1. Selective Pressures Influe ncing Fission-Fusion Grouping Patterns 2 1.2. Use of Acoustic Communication to Locate Patterns 4 1.3. Environmental Factors Affecting Acoustic Communication 6 1.4. Study Species: The Bottlenose Dolphin 9 1.1.1. Fission-Fusion Social Organization 9 1.1.2. Acoustic Communication in Bottlenose Dolphins 10 1.5. Goal of this Thesis 12 2. Dynamics of Group Fission-Fusion in W ild Female Bottlenose Dolphins 15 2.1. Introduction 15 2.2. Methods 19 2.2.1. Study Animals and Study Area 19 2.2.2. Focal Follows 20 2.2.3. Distance Estimation 23 2.2.4. Definition of Terms 24 2.2.5. Analysis 25 2.2.5.1. Dynamics of Temporary Unions 25 2.2.5.2. Dynamics of Temporary Separations 26 2.2.5.3. Class of Focal Female Associates 27 2.2.5.4. Group Cohesion: Coordination of Activities and Headings 28 2.3. Results 29 2.3.1. Focal Follows 29 2.3.2. Dynamics of Temporary Unions 29 2.3.3. Dynamics of Associate Separations 32 2.3.4. Dynamics of Calf Separations 33 2.3.5. Focal Female Associates 36 2.3.6. Group Cohesion: Coordination of Activities and Headings 37 2.4. Discussion 38

PAGE 4

ii 3. Echolocation 72 3.1. Introduction 72 3.2. Methods 76 3.2.1. Study Animals and Study Area 76 3.2.2. Animal Observations 77 3.2.3. Acoustic Recordings 79 3.2.4. Data analysis 80 3.3. Results 83 3.3.1. Animal Observations 83 3.3.2. Acoustic Signals Used in Temporary Separations 84 3.3.3. Acoustic Signals and Maximum Distance of Separation 85 3.3.4. Echolocation and Whistle Rates 86 3.3.5. Activity 88 3.3.6. Whistle Parameters and Different Habitats 89 3.4. Discussion 89 3.4.1. Female-calf Separations 90 3.4.2. Female-associate Separations 94 4. Estimated Communication Range of Social Sounds Used by Bottlenose Dolphins 108 4.1. Introduction 108 4.2. Materials and Methods 111 4.2.1. Behavioral Observations 112 4.2.2. Sound Transmission Experiments 113 4.2.3. Modeling of Active Space of Whistles 116 4.3. Results 118 4.3.1. Behavioral Observations 118 4.3.2. Sound Transmission Experiments 119 4.3.3. Modeling of Active Space 122 4.4. Discussion 123 5. Methodological Considerations in the Definition of Dolphin Groups 142 5.1. Introduction 142 5.2. Methods 146 5.2.1. Parameters: Coordination of Activities, Headings, and Class of Associates 147 5.2.2. Parameters: Duration of Associations and Maximum Distance of Separation 148 5.3. Results 150 5.3.1. Parameters: Coordination of Activities, Headings, and Class of Associates 150 5.3.2. Parameters: Duration of Associations and Maximum Distance of Separation 151 5.4. Discussion 152 6. Thesis Summary 163

PAGE 5

iii 7. Literature Cited 170 About the Author End Page

PAGE 6

iv LIST OF TABLES Table 2.1 Summary of the observations conducted on each focal female bottlenose dolphin in Sarasota Bay, Florida. 48 Table 2.2 Characteristics of the depe ndent calves of 8 focal females and the corresponding time that they were separated from their mother during focal animal observations conducted in Sarasota Bay, Florida. Separation time is expressed in minutes. Note: F = female, M = male, and U = unknown. 49 Table 2.3 Total number of fissionfusion events recorded during focal follows conducted in the summer of 2001, 2002, and 2003 in Sarasota Bay, Florida. A positive change indicates that a dolphin(s) joined and a negative change indicates that a dolphin(s) left the focal mother. 50 Table 2.4 Fission-fusion rates estimated for each focal female including total number of dolphin encounters and encounter type (associates and satellite). 51 Table 2.5 Identified associates (gray) and satellites (yellow) observed at different distances from focal females in Sarasota Bay, Florida. The numbers at each distance are the number of sightings of each dolphin. Gender and relatedness ar e included when known. 52 Table 2.6 Summary of dynamics quan tified for 8 focal females. Note: Habitat types are shallow water (S) and deep water (DW); levels of association are low (L ), moderate-low (M-L), and moderate (M), and overall fissi on-fusion rates are low (L), moderate (M), and high (H). 58 Table 3.1 Summary of the observations for each of the focal females in 2002 and 2003. Codes (FBxx) correspond to the ID of each focal female. 98

PAGE 7

v Table 3.2 Echolocation rate (echolo cation train/min) and whistle rate (whistle/min) recorded in shallow water and channels during separations between females and calves and females and associates. Values are expre ssed as means SD. Calf separations: N whistles in sh allow water = 183, N whistles in channels = 103. Associate separations: N whistles in shallow water = 634, N whistles in channels = 238. indicates statistical significant differe nce. 99 Table 4.1 Overview of transects per habitat where sound transmission experiments were conducted in Sarasota Bay, Florida. Note: H = habitat, SW = shallow water, C = channel. 130 Table 4.2 Spectrum level background noise plus critical ratios in dB re 1 Pa 2 /Hz of each frequency in channels (MC, SRC, AMS, and CC) and shallow-water transects (PSB1, PSB2, SKF, NWPSB1, NWPSB 2, SPSB1, SPSB2, SAMS1, and SAMS2) in Sarasota Bay, Florida. 131 Table 4.3 Regression equations representing the mathematical model of sound propagation for different frequencies in shallow water habitats. Note: To calculate the propagation distance, source levels are added to the intercept of the equation. Frequencies from 7-13 kHz are referred to as a low-frequency whistle and from 13-19 kHz are referred to as a high-frequency whistle. R is the proportion of variabil ity explained by the propagation model. 132 Table 4.4 Regression equations repres enting the mathematical model of sound propagation for different fr equencies in channels. To calculate the propagation distance, source levels are added to the intercept of the equation. Frequencies from 7-13 kHz are referred to as a low-frequency whistle a nd from 13-19 kHz are referred to as a high-frequency whistle. R is the proportion of variability explained by the propagation model. 134

PAGE 8

vi LIST OF FIGURES Figure 2.1 Mean distance error of 633 distance estimates practiced every field day. 59 Figure 2.2 Illustration of two types fi ssion-fusion events observed in wild bottlenose dolphin residents of Sarasota Bay. Arrows indicate the direction that an associate moves relativ e to the focal mother. 60 Figure 2.3 Spatial position of an associ ate observed with a female (FB65) during a follow conducted in August 26, 2002. Blue diamonds represent the position of an associate at 3-min intervals from the female. Red diamonds represent the mo ther-calf pair. 61 Figure 2.4 Diagram showing the + changes and changes of associates in a typical focal follow. Gray indicates + changes and the time intervals between them. Light blue indicates changes and the time intervals between them. 62 Figure 2.5 Duration of associations in 3-min intervals between focal females and other dolphins observed in Sarasota Bay. 63 Figure 2.6 Central tendency and variability of frequency with which associates leave and join focal females during focal follows conducted in Sarasota Bay, Florid a. The solid line drawn across each box represents the median in each event (joining or leaving). The lower boundary of a box is the 25 th percentile, while the upper boundary is the 75 th percentile. The tines on top and bottom of each box represent the largest and smallest frequency values, respectively, that d not include outliers or extreme values. Boxplot legend: o = Outlier and = extreme value. 64 Figure 2.7 Central tendency and variab ility of the duration of association of females with different types of dolphins in Sarasota Bay, Florida. The solid line drawn across each box represents the median in each event (joining or leaving). The lower boundary of a box is the 25 th percentile, while the upper boundary is the 75 th percentile. The tines on top and bottom of each box represent the largest and smallest frequency values, respectively, that d not include outliers or extreme valu es. Boxplot legend: o = Outlier and = extreme value. 65

PAGE 9

vii Figure 2.8 Mean rate of fission-fusion events (changes/min) and number of identifiable associates for each focal female during focal follows conducted in Sarasota Bay, Florida. 66 Figure 2.9 Mean distance of satellites fr om focal mothers and mean distance of separation of focal mothers and their dependent calves, and focal mother/calf pairs and other associates. 67 Figure 2.10 Relationship between the number of associates and the percentage of time that a female and her calf we re nearest neighbors. 68 Figure 2.11 Categories of associates found at 5-m, 10-m, and 20-m from focal mothers. 69 Figure 2.12 Distances at which juvenile s and adult males were most commonly sighted from focal females during te mporary separations. 70 Figure 2.13 Headings recorded for all dol phins (associates and calves) sighted at each distance category within the observation zone. 71 Figure 3.1 Diagram of the research ve ssel and acoustic equipment used to record whistles produced during fission-fusion events. 100 Figure 3.2 Percentage of separations of calves and associates observed in channels and shallow water areas. Se parations were classified into four types of events based on the presence or absence of acoustic signals: 1) no signal, 2) whistles, 3) echolocation trains, and 4) whistles and echolocation trains. 101 Figure 3.3 Mean distance of separation in events having different types of acoustic signals. Bars represented th e standard deviation. 102 Figure 3.4 Relationship between group size and whistle rate (top), and between group size and echoloca tion rate (bottom) during associate separations. Rates are expressed as signal/min. 103 Figure 3.5 Whistle and echolocation rate s (signal/min) in the 5-min period before the separation, during the separation, and the 5-min period after the separation of calves and associates from focal females. Bars represent the standard deviation. 104

PAGE 10

viii Figure 3.6 Mean whistle rate (whist le/min) during separation events of calves and of associates of fema le bottlenose dolphins observed in channels and shallow water ar eas of Sarasota Bay, Florida. Total number of separations with whistles in shallow water: calves = 23 and associates = 21. Total number of separations with whistles in channels: calves = 9 and associates = 24. 105 Figure 3.7 Frequency distri bution of whistles recorded during calf separations in shallow water areas and channels. N whistles in shallow water = 183, N whistles in channels = 103. 106 Figure 3.8 Frequency dist ribution of whistles recorded during associate separations in shallow water area s and channels. N whistles in shallow water = 634, N whistles in channels = 238. 107 Figure 4.1 Study area showing the experime ntal transects in Sarasota Bay, Florida. White circles represen t the location of the hydrophone and black circles represent the last location in which the speaker was placed in each transect. Shal low areas are represented by solid lines: Palma Sola Bay (PSB ), North West of Palma Sola Bay (NWPSB), South West of Palm a Sola Bay (SWPSB), Sister Key flats (SKF), and South East of Anna Maria Sound (SAMS). Channels are represented by dash ed lines: Main Channel (MC), San Remo Channel (SRC), Anna Maria Sound (AMS), and Cortez Channel (CC). 135 Figure 4.2 Attenuation of three freq uency components of a theoretical 5-11 kHz whistle showing where they intersect the hearing threshold and noise floor plus crit ical ratio of each frequency. Hearing thresholds were taken from Ljungblad et al. (1982). See text for explanation of how critical ratios were determined. 136 Figure 4.3 Transmission loss data with di stance for eight tones used during sound transmission experiments conducted in four shallow areas (PSB, SKF, NWPSB, and SPSB). Ex cept for SKF, two transect lines were done in each experime nt and each transect line is represented with a number next to the code of th e corresponding area. Theoretical attenuation based on cylindrical and spherical spreading is also shown. A profile of the depth contour of each transect is also included. 137

PAGE 11

ix Figure 4.4 Transmission loss data with distance for ten tones used during sound transmission experiments conducted in four channels (Main Channel, San Remo Channel, Anna Maria Sound, and Cortez Channel). Theoretical atte nuation based on cylindrical (lines) and spherical spr eading (dashed lines) are also shown. profile of the depth contour of each transect is also included. 140 Figure 4.5 Estimated active space of low-frequency whistles (LFW = 7-13 kHz) and high-frequency whistles (HFW = 13-19 kHz) with different source levels in channels and shallow water areas of Sarasota Bay, Florida. Distan ce is presented in logarithmic scale. Labels are defined in Table 1. 141 Figure 5.1 Inter-relationships of separation distance (mean maximum + SD) between mothers and calves, activit y coordination, and headings relative to defining groups. The active space of social sounds exceeds the full range of behavioral measures (active space of a 7-13 kHz whistle with a source level equal to 160 dB: shallow seagrass area = 301 m, Channel = 13 km; Quintana-Rizzo et al. 2006). 160 Figure 5.2 Inter-relationships of separation distance (mean maximum + SD) between mothers and associates, activity coordination, and headings relative to defining groups. The active space of social sounds exceeds the full range of be havioral measures (active space of a 7-13 kHz whistle with a sour ce level equal to 160 dB: shallow seagrass area = 301 m, Channel = 13 km; Quintana-Rizzo et al. 2006). 161 Figure 5.3 A. Duration of associations in 3-min intervals between focal females and other dolphins observed in Sarasota Bay. B. Cumulative plot of the number of observations recorded for associations of different durations in 3-min intervals. 162

PAGE 12

x Group Fission-Fusion Dynamics and Comm unication in the Bottlenose Dolphin ( Tursiops truncatus ) Ester Quintana-Rizzo ABSTRACT The bottlenose dolphin exhibits a fission-fusion social structure characterized by temporary associations lasting from minutes to hours. Although social structure has been described for some dolphin communities, the se lective pressures affecting fission-fusion patterns and their consequences on dolphi n communication are not well understood. The goals of the present study were three-fold: 1) to quantify the ra te with which fission-fusion occurred and identify the selective pressures in fluencing an individuals decision to leave and join a temporary group; 2) to examine the communication signals produced during temporary separations; and 3) to estimate the distances over which dolphins could remain in acoustic contact while separated. It was found that a dolphins decision to join or leave a group wa s related to social considerations such as the class of individua l encountered (e.g., mothers with calves, adult single females, adult males, and juveniles) as dolphins move in diffe rent environments. The decision was also influenced by ecological characteristics such as the habitat where a dolphin was found. The two aspects in turn determined the rate of fiss ion-fusion. Mothers with calves regularly using deep waters had high rates of fission-fusion. Those fe males encountered other females in the same reproductive condition frequently and associated with them. In contrast, mothers with calves using shallow waters ha d lower fission-fusion rates. Those females encountered juvenile dolphins often but they did not associate with them frequently. Temporarily separated dolphins did not always produce the s ounds typically used for longdistance communication, and sometim es they did not use any detectable acoustic signal to find each other. On average, this absence of communication occurred at distances less than 50 m. When both whistles and ech olocation produced, they were apparently involved in maintaining contact between mothers and their calves and other associates. Estimates of active spaces defined by whistle transmission indicated that communi cation range varied between habitats. Shallow seagrass areas had the smallest active space while channels had the greatest active space. Findings indicated that the distances over which dolphins remain in acoustic contact and can be considered member s of groups are much greater than has been described from observations of dolphin spacing and activity alone.

PAGE 13

1. INTRODUCTION The relationships and interactions of individuals constitute the social structure of a species. The nature of social relationships is sh aped by genetics and features of the physical, biological, ecological, and so cial environment (Hinde 1976, White 1992, Jarman 1993, Kapperler and van Schaik 2002). In some societies, relationships are highly variable and groups are consta ntly changing (Chapman et al 1993). In this fission-fusion system, individuals join and leave groups fre quently in response to changes in the costs and benefits of being a member of a gr oup of specific size and/or composition (Wrangham et al. 1993, Kinsey and Cunningha m 1994). Flexible grouping patterns are especially intriguing in comparison to th e permanent grouping patterns exhibited by many social species. Permanent groups are thoug ht to increase an individuals protection from predators and its opportunities for findi ng or capturing food. However, a more complex, fission-fusion social system may allo w animals to reduce costs associated with more permanent relationships and gain benefi ts from flexible asso ciations (Dunbar 1989). Several mammalian species exhibit a fissi on-fusion social organization. They include bottlenose dolphins ( Tursiops truncatus and T. aduncus ), chimpanzees ( Pan troglodytes ), bonobos ( P. paniscus ), spider monkeys ( Ateles geoffroyi), and grey kangaroos ( Macropus giganteus ) (Jarman and Southwe ll 1986, McFarland 1987, 1990, Wells et al. 1987, White 1992, Connor et al 1992, Chapman et al. 1994, 1995). Although

PAGE 14

the social structure has been described for some dolphin communities, the selective pressures affecting fission-fusion patterns and their consequences on dolphin communication have not been well-studied. 1.1. Selective Pressures Influencing Fission-Fusion Grouping Patterns It has been suggested that the frequency with which individuals leave and join groups can be determined at least in part by population density. As population density increases, individuals are more likely to meet in a given space and, as a consequence, join each other (Caughley 1964, Gerard et al. 2002). This mechanistic approach suggests that group size is an emergent property of population density. In fact, Caughley (1964) and Gerard et al. (2002) suggested that it is possible to obtain an index of density by counting the number of different groups in an area and calculating the mean. Although population density may play a role in determining the frequency with which individuals encounter each other, it may not determine whether individuals decide to form certain associations. This is because individuals of different gender, age, genetic relationships, reproductive condition, health status, etc. may have different needs. Certain relationships may not provide the necessary benefits even if conspecifics are encountered at high rates. Thus, an individuals choice of interactions will probably depend on the relative costs and benefits of joining particular conspecifics (Wrangham 1980). In species with individual-specific relationships, associates are selected based upon their ability to raise inclusive fitness (Wrangham 1980). Such relationships develop if they yield a mutual advantage for all of the individuals involved. Depending on the species, adults may have individual-specific relationships with partners of the same or of 2

PAGE 15

different gender. For example, in wild horses (Equus caballus), females form relationships with 1-2 dominant males, creating a group in which individuals derive feeding benefits and ultimately enhance reproduction. Thus, in this case both a male and a female horse must choose carefully which individuals they associate with in order to maximize their fitness (Rubenstein 1994). Social and ecological pressures can influence an individuals decision to form temporary relationships in different ways (Nishida 1968, Caraco 1987, Macdonald and Carr 1989, White 1989, Packer et al. 1990, Chapman et al. 1995). One such example is the pygmy chimpanzees (bonobo, Pan paniscus). Females tend to leave groups in which large numbers of males are present. When leaving and joining, females associate with other individuals whereas males are usually alone (White 1989, 1992). The female strategy appears to be avoidance of sexual harassment by particular males, while the male strategy seems to be to maintain proximity to receptive females to increase their reproductive success (White 1989, 1992). Differences in individual relations are probably related to males interests generally differing from female interests. Male fitness is usually affected by access to females to maximize mating success, while female fitness is affected more by access to food resources (Trivers 1972, Wrangham 1980, 1986, White 1989, Gompper 1996). Ecological conditions, including predation risk and food distribution, can also influence whether an individual will join or leave other conspecifics. The effects of such pressures on group living are documented by studies conducted on primates (McFarland 1987, 1990; Mnard et al. 1990; White 1989, 1992; Wrangham et al. 1993; Chapman et al. 1995), lions (Panthera leo, Packer et al. 1990), and dolphins (Wells et al. 1980). Most 3

PAGE 16

individual differences in response to predation occur when predation risk is low (Sugiyama 1979, Croft 1980, McFarland 1986). However, different types of individuals may respond in different ways to predation risks. For instance, spider monkey (Ateles geoffroyi) females with young, highly-dependent infants tend to join larger resting and feeding groups more frequently than do females with older and more independent infants. Differences in strategies between females may be related to the vulnerability of the infants to predators, because larger groups should provide safer conditions for the young infants than small groups. It is important to note that a large group may not always be safer than a small group. A larger group might be more dangerous if the group is easier for the predator to detect. Thus, when the costs of being in a group (i.e. increased conspicuousness, food competition, aggression, etc.) outweigh the benefits, individuals may stay alone, join a smaller group, or leave (Krebs and Davies 1993). Since different selective pressures and population density can have effects on group fission-fusion, a study examining both aspects can provide significant insights. The study should investigate the frequency with which conspecifics are encountered and the context of the encounters in terms of who is met and what associations are formed. 1.2. Use of Acoustic Communication Flexible grouping patterns result from frequent changes in associates. Changes in associates can potentially be a problem if individuals do not have a way to locate each other. In many species, acoustic signals are used to maintain communication with associates. In some species, associates develop a shared group-specific call, usually a contact or distance call, to keep in touch with an individuals group (Tyack 2000a). For 4

PAGE 17

example, birds use contact calls repeatedly when they are separated from the flock, reuniting with their mates after separation, or preparing for evening roost (Farabaugh and Dooling 1996). In larger animals, such as killer whales (Orcinus orca), each pod has a group-specific repertoire of discrete calls that is stable for many years. The group-specific repertoires are thought to indicate pod affiliation, maintain pod cohesion, and coordinate activities among pod members (Ford 1989). Acoustic signals can also be used to maintain group cohesion in areas with limited visibility such as dense forest (guinea baboons Papio papio; Byrne 1981), murky waters (bottlenose dolphins; Janik and Slater 1998), or over distances ranging up to several kilometers (elephants Loxodonta africana, Poole et al. 1988). To coordinate group movement in a dense forest, some primates use particular calls (Boinski and Campbell 1995). Adult common marmosets (Callithrix jacchus) use calls to reunite the group during separation from social companions (Norcross and Newman 1993). Similarly, captive bottlenose dolphins are more likely to whistle when isolated, suggesting that whistles are also used to contact partners from whom individuals have been separated (Caldwell et al. 1990, Sayigh et al. 1990, Janik and Slater 1998). Dolphin whistles are narrow-band, frequency-modulated sounds ranging from around 4 to 20 kHz (Caldwell et al. 1990). Acoustic signals used to contact distant partners may be different from those used when individuals are nearby. The change is related to the fact that long-distance signals need to be able to withstand various forms of environmental degradation and attenuation to reach a receiver (Miller 1996). One form of modification is in the frequency modulation characteristics of acoustic communication in non-human primates. Sounds 5

PAGE 18

that are strongly modulated in frequency are much easier to localize than sounds with little frequency modulation (Brown et al. 1979, Wiley and Richards 1982, Oda 1996). For example, adult common marmosets modify their calls to reunite the group during separation from social companions. Such calls are shorter, have more syllables, have lower start and end frequencies, and have higher peak frequencies than calls produced by individuals in a group (Norcross and Newman 1993). Similarly, when group members are distant, ringtailed lemurs (Lemur catta) emit calls of shorter duration, higher pitch, and stronger frequency modulation than when others are nearby (Oda 1996). Understanding how animals use acoustic signals to communicate is basic to describing how relationships are formed and maintained. The study of acoustic communication is particularly important in cases in which associates frequently leave and join each other, because some type of communication must be used to find and locate distant partners, and the acoustic medium is often the best for this need. 1.3. Environmental Factors Affecting Acoustic Communication Communication between individuals requires that one individual produces a signal, the signal travels through the environment, and the signal is detected by a receiving animal. The distance between sender and receiver may be short with little or no signal degradation. Alternatively, the distance may be long and signal degradation with attendant loss of information may occur. Much is known about the temporal and spectral properties of animal signals in terrestrial systems (Wiley and Richards 1978). However, our understanding of sound propagation in animal signals used in shallow water coastal areas (< 20 m deep) is limited (Forrest et al. 1993). It is likely that both terrestrial and 6

PAGE 19

aquatic acoustic communication between associates is affected by ambient noise and by the physical characteristics of the environment, because sound propagation can be dramatically affected by the habitat through which sound travels (Forrest 1994, Tyack 2000a, D. Nowacek et al. 2001). Signals used in long-distance communication must be physically adapted to withstand various forms of degradation to reach and be perceived accurately by receivers (Miller 1996). Problems with transmitting complex sounds over long distances can be reduced through frequent repetition, especially rhythmic repetition. Nevertheless, exceptions to this generalization occur too (Miller 1996). Researchers working with terrestrial species have investigated how certain calls are adapted to maximize propagation over long distances. In terrestrial environments, a sound window has been found close to the ground where the attenuation of sound is at a minimum (Wiley and Richards 1982, Michelsen and Larsen 1983, Forrest et al. 1993). This sound window has apparently been a selective force shaping the acoustic behavior and signals of terrestrial animals. For example, calls of birds in the low forest have their main frequency components in such a sound window (1.5-2.5 kHz; Michelsen and Larsen 1983). In air, filtering of sound is mainly affected by diurnal changes in the sound speed profile. Large reflecting surfaces or vegetation also cause some attenuation of frequencies and amplify others through a substantial frequency filtering of the calls. Some frequency filtering occurs when vegetation and reflecting surfaces move by winds or temperature gradients causing air turbulence (Michelsen and Larsen 1983). In the shallow water environment, a sound must have a wavelength short enough to propagate efficiently in the water depth. Habitat features such as bottom type, 7

PAGE 20

bathymetry, temperature, salinity, and vegetation affect the transmission of sound wavelengths (D. Nowacek et al. 2001). For example, in some aquatic areas, propagation of 0.5 to 5.0 kHz sound is lower in dense submerged seagrass than in mud and sand habitats (D. Nowacek et al. 2001). The effect of the seagrass is not surprising since it acts as a discontinuous surface to the transmission of sound. In fact, researchers have found that in shallow water, low frequency sounds do not propagate as far as high frequency sounds (Forrest et al. 1993, Forrest 1994). The relationship between the acoustic properties of an animals signals and the propagation properties of the sound determines the active space of the signaler. The active space of a signaler is the distance that a signal can be detected and recognized by a receiver (Brenowitz 1982, Klump 1996). When the active space of an animal is unknown, identifying how far a signal propagates in a specific environment and under specific conditions can be used as a first step towards understanding if distant partners can recognize each other. This approach has been used to estimate the active space in songbirds (Brenowitz 1982, Michelsen and Larsen 1983, McGregor 1993). The approach is valuable because it can help to identify if individuals considered as different groups based on distance of separation could be part of a group that maintains cohesion through acoustic contact. Such information is particularly important in bottlenose dolphins because the ephemeral nature of their associations has made the identification of groups difficult. 8

PAGE 21

1.4. Study Species: The Bottlenose Dolphin 1.4.1. Fission-Fusion Social Organization The fission-fusion social organization of bottlenose dolphins is characterized by frequent changes in group size and composition, occurring on a temporal scale that varies from minutes to hours (Wells et al. 1987, Wells 1991, Smolker et al. 1993). As a result of the flexible grouping patterns, residents have a large network of temporary associates (Smolker et al. 1993). Yet, individuals also form close, more permanent, associations often with other individuals of the same gender and/or reproductive condition (Wells 1991, Smolker et al. 1992). For example, some females with calves form close associations with other females of the same reproductive condition. Such association seems to provide a greater protection from predation to the calves. In general, calf survivorship is directly related to group size (Wells 1991). Females belonging to bands may do better in term of their calf survivorship than solitary females, as they tend to form larger temporary groups, and they may derive benefits from the experience of other mothers in the band. As a band member, a females associates are typically drawn from a pool of females with similar ranging patterns and reproductive status (Wells 1991). These pools of potential associates appear to be stable for years (Wells et al. 1987, Wells 2003). Young calves are often recruited back into their natal band (Wells 1991). Males also exhibit two types of social patterns, some individual males are solitary while most form a long-term association with another adult male(s) (Wells et al. 1987, Wells 1991, Connor et al. 1992, Smolker et al. 1992, Wilson 1995, Quintana-Rizzo and Wells 2001, Owen et al. 2002). Male pair associations are the strongest and longest association formed by bottlenose dolphins, and these relationships can endure for a 9

PAGE 22

males entire post-adolescent lifetime (Wells et al. 1987, Wells 1991, 2003). In some cases, a surviving male will form a new association with another single male after the death of his original partner. Some of these high-level associations between males are outgrowths of relationships developed in subadult groups or earlier (Wells 1991). The benefits of close associations within and between sexes are most likely an increased protection from predation and improved chances of finding mates among associated individuals (Wells et al. 1987, Connor et al. 1992). The costs and benefits of temporary associations have not been studied in great detail. Additionally, the effects of population density on the rate of group fission-fusion for bottlenose dolphins are unknown. 1.4.2. Acoustic Communication in Bottlenose Dolphins Bottlenose dolphins produce three broad categories of sounds: whistles, echolocation clicks, and burst pulsed calls (Caldwell et al. 1990). Burst pulsed calls are often described as squawks, yelps, and barks (Caldwell et al. 1990). They are trains of clicks with repetition rates of up to 5,000 clicks/sec, which are commonly produced in social contexts (Blomqvist and Amundin 2004). Dolphin echolocation clicks are broadband, non-modulated sounds with frequency components from about 1 to 120 kHz (Au 2004, Herzing 2004). Wild bottlenose dolphins sometimes use echolocation to detect and obtain distant prey (Rossbach and Herzing 1997, Herzing and dos Santos 2003) and perhaps to discriminate specific prey species (Herzing 2004). In captivity, they have been trained to demonstrate their ability to use echolocation to discriminate the shape, diameter, range, material composition, thickness, and texture of targets in the water (Au 10

PAGE 23

1993, 1997, 2004; Nachtigall 1980). Whistles are thought to function primarily in social communication. For example, Smolker et al. (1993) found that mothers and calves of Tursiops sp. use whistles to reunite after being temporarily separated for a few minutes. A calf tends to whistle when beginning to return to its mother and whistle probability increases with distance of separation. A calf whistling to its mother in the context of reunions might convey information about identity and the infants motivation to reunite. Calf whistles may also induce a cooperative response from the mother, such as slowing down to allow the infant to catch up with her (Smolker et al. 1993). Janik (2000a) proposed that bottlenose dolphins use whistles to communicate over long distances. His study revealed that wild unrestrained dolphins located at distances up to 580 m apart could mimic each others whistles. Janik (2000b) examined propagation of natural dolphin whistles in a 10 m deep channel by measuring source levels and then estimating propagation and the active space using a model. He found that the active space where dolphins could perceive unmodulated whistles between 3.5 kHz and 10 kHz was between 20 km and 25 km. Sound propagation can be dramatically affected by the habitat through which sound travels (Urick 1975, Rogers and Cox 1988). In the aquatic environment, habitat features such as bottom type, bathymetry, temperature, salinity, and vegetation affect the transmission of sounds (D. Nowacek et al. 2001). Vegetation has an effect because it acts as a discontinuous barrier to the transmission of sound. In shallow waters, low frequency sounds do not propagate as far as high frequency sounds (Forrest et al. 1993, Forrest 1994). This suggests that dolphin whistles may be affected by environmental variables. 11

PAGE 24

1.5. Goal of this thesis Bottlenose dolphins have been described as exhibiting flexible fission-fusion grouping patterns (Wells 1991, 2003, Connor et al. 1992, Wells et al. 1987, Smolker et al. 1992). Yet, no study has actually quantified the rate with which fission-fusion changes occur or has described how such changes relate to ecological and social pressures. This is probably due in part to the difficulty of keeping track of all individuals throughout frequent separating and joining occurrences. Additionally, the study of bottlenose dolphin groups is complicated because the proportion of time spent by dolphins underwater and out of sight is much higher than the amount of time they are visible at the surface. Such observational difficulties have resulted in several different ways of defining social groups. It is important therefore to derive methods of describing groups in a more unified way. To accomplish this, different parameters related to grouping, such as typical distances over which dolphins separate, display coordinated activities, and are able to maintain acoustic contact should be considered. The flexible grouping patterns of dolphins are related to the way in which animals join and leave each other frequently. Smolker et al. (1993) proposed that whistles are the acoustic means of communication during temporary separations over long distances. However, bottlenose dolphins do not always whistle during separations. This suggests that other means of acoustic communication are potentially used, specifically because bottlenose dolphins produce more than one type of acoustic signal. Any sound used would need to travel a distance over which it is still perceived by a receiving dolphin. Sound propagation distance will depend on the acoustic properties of the different habitats in which dolphins move. This in turn will determine whether associates can 12

PAGE 25

locate each other, because it is unclear if the distances over which they separate are within communication range. Given this framework, the general goals of this thesis were four-fold. First, I quantified the rate at which fission-fusion occurred and investigated the context in which such events happened to provide insights into the selective pressures determining individual decisions for joining and leaving. To accomplish this, focal animal behavioral observations were conducted on mothers with calves. The study of this social unit allowed me to examine the fission-fusion events between: 1) mothers and their calves and 2) between mother/calf pairs and their other associates (Chapter 2). Second, I examined and quantified the sounds produced during fission-fusion events. This part of the study included the use of continuous acoustic recordings during focal animal observations (Chapter 3). Third, I estimated the potential distances over which dolphins could remain in contact while separated in different habitats. This quantification took into account the hearing capabilities of the species. Estimated distances of sound propagation were used to determine if the observed distances of separation between dolphins were within communication range. This research involved conducting a series of sound transmission experiments in the same areas were dolphins were observed to temporarily separate (Chapter 4). Fourth, using the information generated from grouping patterns and communication range of social sounds, I discuss parameters to be considered in the definition of groups for species with fluid relationships (Chapter 5). Normally individuals interacting with each other to a greater degree than with other conspecifics are referred as a group (Pulliam and Caraco 1984); however, definitions of what constitutes a group and the procedures to measure it are variable. Chapter 6 summarizes the main results of the 13

PAGE 26

thesis and provides recommendations for future research. 14

PAGE 27

2. DYNAMICS OF GROUP FISSION-FUSION IN WILD FEMALE BOTTLENOSE DOLPHINS 2.1. Introduction Social relationships among individuals reflect behavioral strategies that maximize an individuals fitness. The pattern of those relationships constitutes the social structure of a species (Kappeler and van Schaik 2002). In some cases, recognition of social relationships can be a challenging task when individuals form groups of variable size and composition, and especially when the concept of a group is not well-defined, as is the case for bottlenose dolphins, Tursiops truncatus. The flexible grouping patterns of the bottlenose dolphin are complex to describe because groups frequently change in size and composition (Wells et al. 1987, Connor et al. 2000). In this flexible fission-fusion social system, a dolphin may have the opportunity to decide whether or not to associate with other individuals or travel alone at any given time (Connor et al. 2000). An individuals decision to join or leave conspecifics probably depends on the benefits and costs of particular associations. Associations in the bottlenose dolphin seem to provide different types of benefits at different time scales. For example, some associations between males can last the males adult life (at least 20 years, Wells 1991, 2003). Yet, underlying those long-lasting associations, bottlenose dolphins also form short-term, dynamic associations that can last from a few minutes to a few hours. There 15

PAGE 28

are also flexible associations between individuals of social units such as mother-calf pairs in which individuals temporarily separate for brief periods of time. Flexible associations are a very important part of fusion-fusion dynamics and they constitute the basis of the formation of some long-term associations. For example, in the case of adult male pairs, they are usually individuals that associated frequently in flexible subadult groups as juveniles and/or as calves in flexible groups formed by the mothers (Wells et al. 1987, Wells 1991, 2003). Thus, interactions between individuals over time are important parts of fission-fusion social systems. Short-term associations are very flexible by nature. In the Eastern gray kangaroo (Macropus giganteus), groups change membership frequently with a twenty percent probability of change every three minutes (Jarman 2000). Flexible social relationships can result in a variety of associates. In some species, any class of individual can occur in a group (e.g. Eastern gray kangaroo; Jarman 2000). Still, in other fission-fusion species, the same class of individual is asocial in one species (female spider monkeys Ateles geoffroyi) and social in others (female pygmy chimpanzees Pan paniscus; Wrangham 2000). In bottlenose dolphins, variation in sociality has been observed among females with calves. They can either be solitary or form associations with other females with calves (Wells 1991, Quintana-Rizzo and Wells 2001). Such variation offers a unique opportunity to identify and compare the conditions under which individuals decide to join or not join groups. For instance, female pygmy chimpanzees tend to leave a group in which many males are present. White (1989, 1992) proposes that this is a strategy used by females to avoid male harassment. 16

PAGE 29

Group changes can occur not only in composition but also in internal spacing (Jarman 2000). Changes in internal spacing occur when individuals move away from the core of the group for brief periods of time to perform different activities. For example, in mountain baboons (Papio ursinus) individuals spread typically over 20 m when food is readily available and abundant and over 60 m when they are nutritionally stressed. Group cohesiveness is also affected by social factors such as sex, age, and dominance (Kappeler 2000). Cohesiveness is the degree to which individuals are closer to each other. The closer individuals are in space, the more cohesive they are (C.A. Chapman pers. comm.). The formation of temporary subgroups is not random with respect to sex in some primates (Chapman et al. 1995). The age and social status of individual group members may determine an individuals position in a group. Infants and juveniles, which are more vulnerable to predation, are often found near the groups center, where safety is highest (Kappeler 2000). Definitions of a dolphin group need to consider flexible behavioral patterns. Individuals located a short distance from one another and engaging in common activity are normally recognized as group members (Gerard et al. 2002) and individuals whose activity and movements are not synchronized are considered to be solitary (Kappeler and van Schaik 2002). Yet, in some species such as the bottlenose dolphin, group members that are temporarily separated can display uncoordinated activities and can be at distances greater than 200 m. Caughley (1964) suggested than an increase of population density increases the rate of group fusion. As a population increases, the average area covered by a given group increases and individuals get closer enough to join groups more frequently. This 17

PAGE 30

type of relationship suggests that group changes, at least partly, are a result of random meetings as opposed to particular preferences (Southwell 1984). In fission-fusion species such as the Eastern gray kangaroo and chimpanzees (Pan troglodytes), the rate at which individuals join groups is related to population density (Southwell 1984, Lehmann and Boesch 2004). In bottlenose dolphins, the effects of population size or frequency of encounters on fission-fusion patterns have not been quantified. The frequency with which individuals encounter other conspecifics may determine fission-fusion events, but they may also depend on factors such as who is encountered and who is not (Connor et al. 2000). The present study examines the dynamics of group fission-fusion in wild bottlenose dolphins. The study focused on females with their dependent calves because by focusing on this single unit it was possible to examine different types of fission-fusion events. Mothers with dependent calves exhibit two types of fission-fusion events. The first type of event is between the mother and her calf, and the second type of event is between the mother-calf pair and other associates. The dynamics of group formation were examined in two general ways. First, I examined the frequency of joining and leaving events 1) between mothers and their dependent calves and 2) between mother-calf pairs and their associates. In the second case, the identity of the associates was determined when possible, and the most common classes of associates were identified. I also determined if the frequency of associate changes scaled with the rate of dolphin encounters. Quantifying associations based on dolphin encounters is a more direct approach than extrapolating information on population size to estimate the frequency of group changes. 18

PAGE 31

Second, the identity of associates found at different distances during temporary separations was used to investigate if certain classes of individuals were found at particular positions. Separation distances, activity coordination, and headings during temporary separations were examined relative to group cohesion. I hypothesized that mothers and their calves separated over shorter distances than mother-calf pairs separated from other associates. Since calves depend on their mothers for survival and protection, it could be costly for a calf to wander too far away from its mother. The context of temporary separations was examined to provide insights into the selective pressures that influence dolphin behavior. Data on group spatial structure were also used to evaluate the existing definitions of a dolphin group. 2.2. Methods 2.2.1. Study Animals and Study Area Twelve well-known mothers with dependent calves were the focus of this research. All females were visually identified by distinctive markings on the dorsal fin, which is normally exposed when dolphins surface to breathe. Identifications of dolphins were confirmed with photos of the dorsal fin marks (Irvine et al. 1981, Scott et al. 1990, Wells and Scott 1990). Photographs were taken with a Canon Elan IIEQD 35 mm camera fitted with a 75-300 mm zoom-telephoto lens, databack, and Kodachrome (K-64) color slide film. This fine-grain film provided good resolution of fin features (Wells et al. 19

PAGE 32

1996). Individuals were designated by codes (FBXX or FXXX) that corresponded to the freezebrand number applied to each dolphin during previous health assessments. Focal females were members of the resident community of about 160 bottlenose dolphins using Sarasota Bay, Florida on a regular basis. Much information exists about the dolphin community including identity, sex, age, reproductive status, and genetic relatedness of most of the long-term residents. At least five concurrent generations of dolphins have been observed in this area. Their flexible social organization has been described elsewhere (Wells et al. 1980, 1987, Wells 1991, 2003). The study area extended from the southern edge of Tampa Bay to Siesta Key, off Sarasota. The area includes the shallow inshore waters of Sarasota Bay and associated bays and sounds, and the coastal Gulf of Mexico (Wells et al. 1987). The study area has a variety of habitats including shallow waters (extensive areas < 3 m deep characterized by seagrass and sand patches), channels (narrow waterways > 3 m deep often formed or maintained by dredging), bays (large inshore open areas, several kilometers wide and depth > 3 m), gulf (Gulf of Mexico waters), or edge (area between habitats approximately 5-m wide). 2.2.2. Focal Follows Field seasons occurred in the summers of 2001, 2002, and 2003. The number of focal animals for each season varied according to the number of mothers that had dependent calves 3 to 4 years old. In 2001, five focal females (FB15, FB59, FB83, F101, and F119) were studied from May to July. In 2002, two focal females (FB65 and FB90) were studied between August and September. In 2003, two focal females from 2002 were 20

PAGE 33

included with 5 additional females (FB54, FB55, F141, F149, and F157) for a total of 7 focal animals. In 2003, focal follows were conducted from June to September. In 2001, data collection was done from a 6-m long boat and in 2002 and 2003, it was done from a 7-m long boat. Both vessels were equipped with 4-stroke outboard engines. A combination of focal animal and scan sampling observations (Altmann 1974, Mann 2000) were used. Focal animal observations were conducted on the mother of mother-calf pairs. Data were collected in two ways: 1) at 3-min time intervals (activity, nearest neighbor, habitat, heading, and distance) and 2) at the surfacing of the focal mother (distance). In all seasons, at 3-min instantaneous time points (instantaneous point sampling technique, Altmann 1974), the primary observer (EQR) recorded the distances and headings of all the dolphins in the observation zone in relation to focal females when dolphins were at the surface. The observation zone had a radius of 200-m from the focal female, because this was considered to be the maximum range over which dolphins could be accurately observed. Heading categories recorded were: 1) parallel, 2) heading towards, and 3) heading away. The habitat in which dolphins were found was also recorded. At 3-min intervals, the activities of the focal female and of all individual dolphins within the observation zone were recorded. Activity data of all dolphins were collected at the first surfacing of the focal female after the 3-min time point. Behavioral activities recorded were: 1) traveling, 2) milling, 3) socializing, and 4) feeding. Traveling was defined as directional movement of a dolphin(s) in a straight line or zigzag, and milling was defined as a non-direct moment (Urian and Wells 1996). Socializing was defined as activity involving prolonged body contact, in which individual dolphins often displayed 21

PAGE 34

playful, aggressive, and/or exploratory surface behaviors (Shane 1990, Urian and Wells 1996). Feeding was recorded if a dolphin was observed with a fish in its mouth. However, probable feeding was also used for the behavioral activity analysis (Waples 1995), because even though it does not provide confirmation of the capture of a fish, it is a strong indication that a dolphin is feeding (Shane 1990). Probable feeding included dolphins swimming individually in circles near the surface of the water; dolphins chasing or striking a fish with their flukes (fishwhacking); dolphins increasing swimming speed suddenly then, spinning in a circle or making a hairpin turn (pinwheeling); and dolphins, alone or in loose groups, repeatedly diving in varying directions at one location (Shane 1990, Urian and Wells 1996). Observations were done only when sea state was equal to 0 or 1. In 2002 and 2003, continuous sampling was added to the data collection protocols to record the distance of all dolphins in the observation zone relative to focal females. In the case of cetaceans, continuous sampling data were recorded when an animal surfaced (Mann 2000). I was able to keep track of dolphins during each sampling period with the help of 1-2 additional observers. Each observer monitored a subset of animals and pointed out their location each time a focal female surfaced. I called out the surfacings of focal females so that observers knew when to indicate the locations of dolphins. When these dolphins surfaced (at the same time as a focal female or within approximately 30 sec of her surfacing), I estimated their distances to focal females. Focal females surfaced on average once or twice per minute. Positional data of dolphins allowed me to determine the time at which they united and separated from focal females, the total length of an 22

PAGE 35

event, and the maximum distance of separation between dolphins. During observations, the research vessel was kept at a distance of approximately 20 m from the focal females. Distance estimates of all dolphins to focal females were made using distance categories (5 m, 10 m, 20 m, 30 m, 40 m, 50 m, 60, 70 m, 80 m, 90, 100 m, 125 m, 150 m, 175 m, 200 m, and 200 + m). In marine and terrestrial environments, data collection using distance categories has proven to reduce systematic errors such as rounding numbers, although this protocol does not completely eliminate human bias (Buckland et al. 1993). Distance categories are a helpful way to collect distance data in the field where it is difficult to record absolute distances and determine whether a dolphin is at 68 m or 69 m, for example. The low resolution of absolute distance estimates does not provide more detailed information about dolphin spatial distribution than the distances recorded into categories. 2.2.3. Distance Estimation Every field day prior to searching for focal dolphins, all observers (3-4) practiced distance estimates by comparing their estimates of distances to fixed objects located in the water to concurrent measurements from a Leica LRF 800 laser range finder. The primary observer (EQR) estimated about 92% of the distances recorded during the focal follows; other observers were trained to estimate distances when their assistance was needed. Estimates were done on objects located at distances between 10 and 200 m from the research boat when anchored. Twenty five objects were randomly selected every field day so that each observer practiced distance estimates 25 times. Observers did not share 23

PAGE 36

their estimates, in order to prevent bias. The actual distance of the objects to the research vessel was obtained with the laser range after all observers did their estimates. Information on the actual distance of objects was used to calculate the error of the distance estimates of dolphins by the primary observer during focal follows. The error was calculated by subtracting the actual distance from the estimate. Mean error of the distance estimates tended to increase with distance (Fig. 2.1). Average error per distance varied from -0.17 m (underestimation) to 16 m (overestimation). Overall mean error was 4 4 m. 2.2.4. Definition of Terms Associates were defined as any dolphins that united with a focal mother during a follow. The term does not include dependent calves, which are always with their mother. Two types of fission-fusion events were studied: 1) temporary unions and 2) temporary separations (Figure 2.1). Temporary unions started when the distance between a focal female and the associates was equal to or less than 200 m (the size of the observation zone), continued as the distance between dolphins decreased and it was equal to or less than 20 m. A temporary union ended when the distance between animals increased by more than 200 m. Within a union, temporary separations can occur (Figure 2.2). Temporary separations were events in which a focal female and her calf or other associates increased by distances greater than 20 m. The process ended when the distance was equal to or less than 20 m. A distance of 20 m was chosen because 75% (n = 2087) of the dolphins observed in 53 focal follows conducted in 2001 were within 20 m of focal females. The validity of this criterion was also examined with the results of group 24

PAGE 37

dynamics. Temporary separations were divided into two categories: 1) separations of mothers from their dependent calves (also referred as calf separations) and 2) separations of mother/calf pairs from other associates (also referred to as associate separations). 2.2.5. Analysis 2.2.5.1. Dynamics of Temporary Unions Durations of temporary unions were defined from the first and last 3-min time point in which associates were in the observation zone. Those were used to calculate a rate of change of associates/min (Southwell 1984). Three rates were calculated: 1) rate of fission-fusion events, 2) rate of fusion events or + changes, and 3) rate of fission or changes. Rates of fission-fusion were estimated for every focal follow and included both + changes and changes. Calculations were done as follows. Figure 2.4 shows a typical follow in which associates joined and left. A total of 20 time points was collected at 3-min intervals and the follow lasted 60 min. To estimate the fission-fusion events, the total number of + changes (n = 3) and changes (n = 3) were added and the total was divided by the total of minutes. For this focal follow, the estimated rate of fission-fusion was 0.10 changes/min. To calculate the frequency with which associates joined, the intervals of time between + changes were considered. In the previous example, the frequency of joining was estimated by adding the intervals of time between two + changes: 1) 9 min between time point No. 4 and time point No. 7, and 2) 12 min between time point No. 7 and time point No. 11. Those two time intervals were added and were divided by 2, yielding a value of 10.5 min, indicating that on average dolphins joined every 10.5 min. 25

PAGE 38

A similar analysis was used to calculate the frequency with which associates left focal females. In this example, the intervals of time included in the analysis were: 1) 6 min between time point No. 9 and time point No. 11, and 2) 9 min between time point No. 11 and time point No. 14. The two intervals of time added up to 15 min, which divided by 2 equals 7.5 min. This means that on average dolphins left every 7.5 min when changes in associates were observed. Overall rates of fission-fusion were compared among focal females for every focal follow. Only females that had a minimum sample size of 5 focal follows were included (FB54, FB55, FB59, FB65, FB90, F119, F141, and F157). This is because comparisons were done among individual females and the minimum sample size for statistical comparisons is five (Ott 1994). Additionally, I determined if there was a relationship between the overall rates of fission-fusion and the number of dolphin encounters that a female had as she moved to different areas. An encounter was defined as the presence of a dolphin or group of dolphins at 200 m from focal females. An encounter included dolphins that join a focal female and dolphins that were in the observation zone (within 200 m of the focal female) but did not join (did not come to within 20 m of the focal female). Dolphins that did not join focal females were referred to as satellites. 2.2.5.2. Dynamics of Temporary Separations The durations of temporary separations of mothers and their calves and those involving mother-calf pairs and their associates were calculated using the continuous 26

PAGE 39

data, which were collected in 2002 and 2003. Those data were also used for the analysis of acoustic signals used during temporary separations described in Chapter 3. Duration of separations and maximum distances of mothers from calves and associates from mother-calf pairs were compared among females using a Kruskal-Wallis test. In those analyses, the unit of comparison was every separation event. Only females that had a minimum sample size of 5 separation events were included in the analyses, because five is the minimum sample size for statistical comparisons (Ott 1994). Events in which the distance of separation was recorded as 0+ were not included in the analysis. Activity data were used to determine the activity of dolphins during separations. Time points in which the activity of one or more individual dolphins was not recorded were not included in the analysis (237 out of 3288 data points). Habitats were combined into two categories: shallow water and deep water. Deep-water habitat included channel, bay, and gulf. The percentage of time that each focal female spent in those habitats was calculated. Females included in this analysis were those for which fission-fusion rates were compared statistically (FB54, FB55, FB59, FB65, FB90, F119, F141, and F157). 2.2.5.3. Class of Focal Female Associates Class of associates was studied among the eight focal females for which fission-fusion rates were compared statistically. The basic classes of associates considered were mothers with calves, adult females without calves (referred to as single adult females), adult males, and juveniles. Information on gender and age-class were known for most of the resident dolphins of Sarasota Bay. For each of the eight focal females, the percentage 27

PAGE 40

of associates that fell into each category was quantified. Additionally, the level of association between focal females and associates was calculated using the half-weight index. The index is commonly used to describe associations of bottlenose dolphins (Wells et al. 1987, Smolker et al. 1992, Quintana and Wells 2001, Lusseau et al. 2003). The unit of measurement was every 3-min time point in which a female was observed with a particular associate when the total duration of the association was known. The resulting indices were grouped into five association categories (Quintana and Wells 2001): low (0.01-0.20), moderate-low (0.21-0.40), moderate (0.41-0.60), moderate-high (0.61-0.80), and high (0.81-1.00). To examine spatial structure as groups expanded and contracted during temporary separations, the class of associates found together (5 m, 10 m, and 20 m) and farther away (> 20 m 100 m) from focal females was compared. 2.2.5.4. Group Cohesion: Coordination of Activities and Headings Group cohesion was studied by examining activity coordination and headings of all dolphins (associates and calves) when they were together and temporarily separated from 12 focal mothers. The percentage of each type of heading (parallel, heading away, and heading towards) was calculated for each distance category and results were examined to identify the distance(s) at which groups were more cohesive (i.e. dolphins were traveling parallel to each other). Activities were classified as coordinated if dolphins displayed the same activity as the focal mothers, and uncoordinated if they displayed different activities. It is important to note that the sample sizes of data included in statistical comparisons were not always the same as the total sample size reported. There are two 28

PAGE 41

reasons for this discrepancy: 1) in some cases not all data were analyzed because they did not meet requirements for minimum sample size for statistical comparisons, and 2) for some time points, data were not recorded for all dolphins so these points could not be used in the analysis (e.g. activity data and heading were not recorded when dolphins did not surface at a time point, habitat was not recorded if dolphins were not in sight). Statistical tests were conducted with the SPSS 14.0 package (2005). Results are expressed as means SD and all tests are two-tailed, with a significance level 0.05. 2.3. Results 2.3.1. Focal Follows In total, 49 focal follows were conducted in the summers of 2001, 2002, and 2003. The total number of focal follows per female varied between 3 and 9 (Table 2.1). Total follow time was 9831 min or 164 hr (Table 2.1), in which 3277 3-min time points were collected. 2.3.2. Dynamics of Temporary Unions A total of 105 temporary unions (105 joins and 105 leaves) was observed during focal follows. Temporary unions lasted 25 26 min and ranged from about 3 min to 129 min (Figure 2.5). Although estimates of the duration of temporary unions were calculated using the 3-min data, no events lasted less than 3 min. This was confirmed because the 8 29

PAGE 42

events included in the 3-min category were observed in 2003, when continuous data were also recorded. Temporary unions represented 25% (2503 min) of the total time that females were observed. This percentage, however, does not include the events in which the start or end time of a union was unknown (e.g. associates were with a focal female when a follow started or animals did not leave before the follow ended). When events of unknown start/end times were included, another 28 joining and 14 leaving events were observed. The longest time that a focal female was observed with an associate(s) before a new dolphin(s) joined was 114 min. The longest time that a focal female was observed with an associate(s) before a dolphin(s) left was 156 min. Joining events of associates occurred every 20 23 min and leaving events occurred every 30 30 min (Figure 2.6). Events could occur serially (a leaving event after a joining event) or simultaneously (associates join when others leave). Additionally, some dolphins joined at different times but left together. This was counted as two joining events but one leaving event. Mothers associated with different types of dolphins, including other mothers with dependent calves, single adult females, juvenile females, adult males, and juvenile males. Estimates of the duration of those associations were not statistically different (Kruskal-Wallis = 2.05, p = 0.73, N = 113). Yet, associations with adult males had the lowest range of association times (6 to 33 min; Figure 2.7). This calculation does not include one association in which the total duration of the association was unknown: two adult males were already with a focal female when the follow started and continue with her when the followed ended. That association lasted at least 303 min, and was consistent with mate guarding behavior reported previously. 30

PAGE 43

Overall rates of fission-fusion varied from 0.01 to 0.06 associate changes/minute (Table 2.3). Females with higher rates also tended to have more identifiable associates (Figure 2.8; regression equation: y = 189.6x + 12.5, R 2 = 0.6). The female with the highest rate (F141) had 42 identifiable associates in 405 min of observation whereas the female with the lowest rate (FB54) had 12 identifiable associates in 492 min of observation. No relationship was found between the total number of associates that a focal female had and total follow time (F 1, 11 = 3.87, p = 0.07, R 2 = 0.28, N = 12). However, there was a relationship between the overall rate of fission-fusion and the number of dolphin encounters (F 1, 6 = 6.36, p = 0.04, N = 8), which was a moderate relationship (R 2 = 0.51). Most focal mothers formed associations with dolphins encountered and those events formed 72-100% of the encounters. One mother, FB55, formed associations during 50% of encounters. The mother with the lowest fission-fusion rate, FB54, only formed associations with dolphins in 26% of the encounters. Animals that were encountered (came within 200 m) but not join (did not come within 20 m) were referred to as satellites and they occurred in 74% (n = 20) of encounters of FB54 (Table 2.4). Satellites were observed at a mean minimum distance of 118 58 m (N = 127) from focal females. In total 69 individual satellites were identified with the 12 focal females. Satellites included animals of both genders, and varying age and reproductive status. Table 2.5 includes 1) all minimum distances at which individual satellites were observed and 2) all maximum distances of separations of identifiable associates. If an individual dolphin was not observed between 2 or more distances, the distance(s) in between was(were) marked as an individual had to pass this(these) distance(s) to get to 31

PAGE 44

the other. For example, if a dolphin was observed at 70 and 100 m but not at 80 and 90 m, then 80 and 90 m were also included as the individual had to move those distances. The habitat used by focal females during focal follows was examined. The one female (FB54) with the lowest overall rate of change in associates spent slightly more time in shallow water than in deep water areas (shallow water: 55%, n = 223; deep water = 45%, n = 182). All other females with higher rates of associate changes, except for FB55, spent more time in deep water than in shallow water (shallow water = 29 19 %, n = 440; deep water = 71 12 %, n = 1033). FB55 and FB54 were the females that mainly exhibited temporary separations of dependent calves rather than other associates (Table 2.1). 2.3.3. Dynamics of Associate Separations Separations between females and associates were more common than separations between females and calves in five of the focal mothers (FB65, FB90, F141, F149, and F157). In those females, associate separations represented 65% to 92% of all temporary separation types including calves and associates (Table 2.1). Cases in which the calves and other associates separated from focal females are included in the next section. Mean maximum distance of associate separations was 60 30 m. Mean maximum distances of separation between females and associates were significantly lower than the maximum distance of separation between females and their dependent calves (U = 4450.50, p < 0.001, N = 227; Figure 2.9) and the mean minimum distance of satellites from focal females (U = 2663.00, p < 0.001, N = 235; Figure 2.9). 32

PAGE 45

In associate separations, females spent 84% of their time traveling and 12% milling. Similarly, associates spent 77% of their time traveling followed by 11% milling during separations. Additionally, associates spent more time socializing (10%) than did focal mothers (2%). Socializing occurred only in shallow water. Feeding was the least frequent activity of associates and focal mothers at 2% each. Analyses used 119 3-min data points. 2.3.4. Dynamics of Calf Separations Focal mothers and their calves separated a total of 185 times. Of those, 131 events were used to calculate the duration of the separations because in those events the distance of focal mothers to calves was recorded in relation to every surfacing of the mother (summers 2002 and 2003 data). Mothers and their calves were separated for 1 min to 47 min, with a mean separation time equal to 9 9 min. The duration of each separation was not statistically different among focal females with at least 5 separation events (F 4, 121 = 0.46, p = 0.37, N = 126). However, the total amount of time that they were separated from their calves was different. Some females and calves were temporarily separated up to 38-40% of the total follow time (FB54 and FB55) while other females and calves were separated from their calves only 2-7% (FB59, FB65, FB90, F141, F149, and F157). The variation in the amount of time that mothers and calves spent together and separated was examined relative to the gender and birth-order of the calves, but no pattern was found (Table 2.2). Possible relationships between total follow time and total number of separations recorded were examined among females, and no significant relationship was found (F 1, 10 = 2.96, p = 0.12, N = 12, R 2 = 0.15). Nevertheless, there 33

PAGE 46

was a significant relationship between the percentage of time that a mother and her calf spent next to each other (i.e. nearest neighbor) and the number of associates a mother had (F 1, 6 = 15.24, p = 0.008, N = 8, R 2 = 0.72). Females with more associates also had their calves as the nearest neighbor more often (Figure 2.10). Mean maximum distances of separation were not significantly different among mother-calf pairs (Kruskal-Wallis = 1.70, p = 0.89, N = 131). Mean maximum distance of separation was 82 46 m and it ranged from 30 m to over 200 m (normally less than 250 m). The comparison included the separation events of a dependent calf from a first time mother (FB55) and three dependent calves of known experienced mothers (FB54, FB65, and FB90). The context of the 148 temporary separations between females and dependent calves was variable. In 80% (n = 118) of the temporary separations, no other dolphins were present in the observation zone. Yet, in 20% (n = 30) of the separations, other dolphins were present and seven different contexts were identified (when possible the class of associates observed was included): 1) Separations in which calves and other associates separated to socialize, feed, and mill away from focal mothers (37%, n = 11). Identified associates included other mothers with their dependent calves, an older brother of the focal calf, and juveniles. Mean maximum distance of separation was 85 47 m. 2) Separations in which focal calves joined other dolphins (brother, juvenile female, other mothers and calves) and later returned alone to their focal mother (20%, n = 6). Calves and other dolphins were socializing, traveling, and probably feeding. Mean maximum distance of separation was 120 33 m. 34

PAGE 47

3) Separations in which the focal calves returned to the focal mother just before (< 5 min) other dolphins joined (14%, n = 4). Associates identified included females with their dependent calves and adult males. Mean maximum distance of separation was 150 43 m. 4) Separations in which focal mothers and their calves separated from each other within 5 min of when all other associates present left the observation zone (10%, n = 3). Associates identified included were juveniles. Mean maximum distance of separation was 67 55 m. 5) Separations in which focal calves joined other dolphins in the observation zone and returned with some of them to the focal mother (10%, n = 3). Mean distance of separation = 67 29 m. Dolphins identified in this context were an older brother of the focal calf and an adult male. Focal calves were socializing and traveling with the other dolphins. 6) Separations in which no other dolphins were present when focal mothers and their calves were traveling, but they started the reunion process when other dolphins appeared in the observation zone (7%, n = 2). Mean distance of separation was 145 7 m. 7) Event 4) followed by event 3) (3%, n = 1). Distance of separation was 85 m. Separations in which calves associated with other dolphins were of a significantly longer duration and greater maximum distance than those in which calves were alone (time: U = 302.50, p = 0.03; distance: U = 84.50, p = 0.004; N = 127). Focal mothers and their calves spent the majority of their time traveling when they were temporarily separated (females = 66%, calves = 67%). They spent comparable amounts of time 35

PAGE 48

feeding and milling during separations (females: feeding = 17%, milling = 17%; calves: feeding = 13%, milling = 16%). Socializing was done at a lower percentage (females = 1%, calves = 4%). Analysis was done using a total of 360 3-min data points. Habitat was recorded for both focal mothers and their calves in 332 3-min time points. Most temporary separations of focal mothers from their calves occurred in shallow water areas (63%). Some separations also occurred in deep water (31%) and at the edge between two habitats (6%). 2.3.5. Focal female associates Mothers with their dependent calves, and independent juveniles were the two most common classes of associates of 8 focal mothers. Four females (FB54, FB55, FB59 and F119) had slightly greater proportions of associations with juveniles than with mothers with their calves (percentage of juvenile associations: FB55 = 33%, F119 = 46%, FB54 = 60%, FB59 = 63%; percentage of mother/calf associates: FB55 = 33%, F119 = 32%, FB54 = 30%, FB59% = 32%). Interestingly, those four females were never sighted with adult males. The other four focal females (FB65, FB90, F141, and F157) had a greater proportion of associations with females with their dependent calves (FB65 = 48%, FB90 = 50%, F141 = 45%, and F157 = 42%) than with any other class of associates. Levels of associations were variable among females. The three focal mothers that had more associations with juveniles (FB54, FB55, and F119) only displayed low-level associations. Two focal females (FB90 and F157) formed low-level associations with all types of associates but moderate-low level associates were exhibited only with juveniles and other females with calves. Three other focal females (FB59, FB65, and F141) formed 36

PAGE 49

three levels of associations: low, moderate-low, and moderate. Moderate-level associations with consistently formed with other females with calves. The class of associates observed in every distance category up to 100-m from the focal mother was quantified to examine spatial structure. Three categories of dolphins were most commonly seen at 5 m from focal females: females with calves (34%, n = 145), juveniles (35%, n = 152), and adult males (24%, n = 101). Single adult females were also observed but at a much lower frequency (7%, n = 31). At any other distance, the most common class of associate observed was other mothers with their calves and their percentage of dolphins identified in this category varied from 40% to 62% (48 7 %; Figure 2.11). The second two most common classes of associates were juveniles and adult males at distances from 10 m to 100 m. Adult males were more common at 10 m and juveniles were more common at 20 m from focal females (Fig. 2.10). However, at almost any other distance, when the percentage of adult males increased, the percentage of juveniles decreased, and vice versa (Figure 2.12). 2.3.6. Group Cohesion: Coordination of Activities and Headings On average, coordinated activities occurred at 24 37 m and uncoordinated activities occurred at 70 53 m from focal females. The most common coordinated activity was travel (84%) and it was followed by lower percentages of milling (10.99%), feeding (4.68%), socializing (0.14%), and boat-riding (0.05%). The most common uncoordinated activity was also travel but dolphins spent less time traveling when they were uncoordinated. Time spent traveling was greater (84%) when dolphins were coordinated as compared to only 40-43% when dolphins were uncoordinated (focal 37

PAGE 50

mothers = 40%, other dolphins = 43%). In contrast, the amount of time that dolphins spent feeding, milling, and socializing was 10% greater when the activities were uncoordinated (milling: focal mothers = 36%, other dolphins = 23%; feeding focal females = 23%, other dolphins = 23%; socializing: focal females = 0%, other dolphins = 14%). The analysis was done using a total of 4753 3-min data points. Headings of dolphins varied with distance. Dolphins within 20 m usually moved parallel to focal females. At 30-m, headings were mainly heading towards (Fig. 2.12). From 40 m to 100 m, two headings were more common, heading towards and heading away (Figure 2.13). After 100 m, the most common heading was heading away. The analysis was done using a total of 5273 3-min data points. 2.4. Discussion Differences in the nature and patterns of social interaction can give rise to particular social relationships. Consistent features of those relationships are used to characterize the social structure of a species (Hinde 1976). In this study, interactions between individual dolphins indicated that changes of associates of females with calves occurred on average every 26 min. This is much more frequent than some other social species. For example, in bonobos (P. paniscus), another species with fission-fusion social organization, changes occurred on average only every 100 min (White 1988). Differences between species can also be related to differences in methods used to collect data. It is important to note that the changes in bonobos are reported for all types of individuals, 38

PAGE 51

particularly in groups where females tend to outnumber males, but all-female groups are relatively common in this species. In this study, joining events of associates occurred every 20 23 min and leaving events occurred every 30 30 min (Figure 2.6). Different events involved variable numbers of dolphins, either as associates joining or leaving. Joining events happened both simultaneously and in sequence. Some dolphins left while another dolphin was joining, while some dolphins joined and a few minutes later the same or other dolphins left. The way in which individuals associate when leaving or joining can be viewed as reflecting the degree of cohesion (White 1988). In bonobos, females and males leave and join a party in the company of others, but only males appear to join or leave frequently as lone individuals (White 1988). Most adult male bottlenose dolphins observed in this study left alone or in the company of a close associate male (Wells et al. 1987). In two cases, adult males left in the company of juvenile males and on three occasions, with adult females (with and without calves). In the case of females, most adult associates of females had calves and mother-calf pairs tended to join and leave groups by themselves. It is interesting that although females with calves form closer associations with other females of similar reproductive status, their associations do not result in leaving together in most cases (>56%). Some studies have indicated that the rates of fusing and splitting up can be related to population density. As a population increases, the rates at which individuals meet increases. As a consequence, individuals join and leave groups more frequently (Caughley 1964, Southwell 1984). The positive scaling between population size and 39

PAGE 52

encounter rate has reported in fission-fusion species such as large macropods (Southwell 1984), chimpanzees (Lehmann and Boesch 2004), and roe deer (Capreolus capreolus; Gerard et al. 2002). Southwell (1984) and Gerard et al. (2002) indicated that groups are in constant change as a result of random meetings. However, they assumed that all associations provide similar benefits and costs and that individuals of different classes and reproductive status should join if they encounter each other. Yet, an individuals choice of associates depends on which partners will raise inclusive fitness by the greatest amount (Wrangham 1980). Results from the present study indicate that the rate with which individual dolphins join and split is not a result of random encounters. Instead, they seem to be influenced by a combination of social and ecological pressures. Females with moderate and high rates of fission-fusion formed associations during most encounters. In contrast, the one female with low fission-fusion rates had comparable number of encounters to those females with high fission-fusion rates and even slightly higher than those of females with moderate fusion-fusion rates. Yet, she did not form associations during encounters. In fact, this female (FB54) had the highest percentage of satellite individuals (by definition, remaining at distances > 20 m). The percentage of satellites (74%) was similar to the percentage of associations (72-100%) formed by females with higher rates of fission-fusion. The durations of associations formed were not statistically different among females. Examination of the class of individuals encountered indicated that the social context of the encounter in terms of who is encountered and who is not is very important. Females with calves forming moderate lowto moderate-level associations (FB59, FB65, 40

PAGE 53

FB90, F141, and F149) were also females who encountered and formed associations with individuals of the same reproductive condition at a higher percentage. In contrast, females with calves who did not form higher levels of association (FB54, FB55, and F119) encountered mostly juveniles and those associations were low-level. Based on association levels, it can be said that the first type of mothers were social and the second type were solitary. Female dolphins have been found to have varying degrees of sociability even across populations (Wells 1980, 1991; Smolker et al. 1992; Mann et al. 2000). Thus, the class of individual encountered seems to be an important part of a females decision to join certain groups. When fission-fusion dynamics were compared among mothers, other patterns were revealed and a better picture of their social system emerged (Table 2.6). Mothers who exhibited only low-level associations, were also females who spent a large percentage of time separated from their calves (FB54 and FB55) and used shallow waters often. Conversely, mothers exhibiting higher level associations spent a large percentage of time with their calves (FB59, FB65, FB90, F141, and F157) and used deep waters more often. One female (F119) that spent half of her time in shallow water and the other half in deep water, spent comparable amounts of time with associates and separated from her calf; she had low-level associations and most of her associates were juveniles. These patterns suggest that ecological factors such as habitat features are important predictors of 1) whether separations between mothers and calf occur, and 2) females with calves encounter other females of similar reproductive condition and associate with them. Consequently, habitat may be a factor in determining rates of joining and splitting. Differences in strategies between mothers may be related to the 41

PAGE 54

calves vulnerability to predators. In Shark Bay, Western Australia, shallow water habitats correlate positively with dolphin (Tursiops spp.) calf survivorship (Connor et al. 1999). Females that spent more time in shallow water also have greater reproductive success. Additionally, females may benefit from temporary separations by being able to have more uninterrupted foraging opportunities (Mann and Smuts 1998). A calf may benefit from separating often from its mother by gaining hunting and social experience leading to independence (Mann 1997, Connor et al. 2000). Mann and Watson-Capps (2005) proposed that frequent mother-calf separations are indicators of the vigor of the calf because a calf has to be in good condition to venture hundreds of meters away and successfully return to its mother, especially in the predator-rich environment of Shark Bay. It is interesting that the distances separating calves from their mothers were significantly greater than those separating mothers and associates. Since calves depend on their mothers for survival, it would seem that they should stay closer to their mother than should other associates. However, considering that deeper waters may be more dangerous, it follows that associate separations, which occur mostly in deep water, are also characterized by tighter spatial structure. Additionally, separations in which calves associated with other dolphins were of a significantly longer duration and greater distance than those in which calves were alone. It could be that some associates that were in close proximity to the calves provide some protection. For examples, other mothers were sometimes in close proximity to the focal calf when calves were socializing. If predation pressure is greater in deeper waters, the survival costs are probably outweighed by some benefits or else females would be putting their calves at a great 42

PAGE 55

unnecessary risk. By having more associates, the mothers increase the probability that their calves have peer playmates, have interactions with peers of both sexes, and have more adults from whom care may be garnered in case of potential danger (Altmann 2000). Large groups also decrease the individual probabilities of being attacked (dilution model; Blumstein et al. 1999), and in general, provide safer conditions than small groups (Wells et al. 1980, Norris and Dohl 1980, Caraco 1987, Wrangham et al. 1993). Another benefit to the calf is that by being frequently associated closely with its mother, the drafting hydrodynamics of the mother probably reduces the energy expenditure required of the calf for swimming (Weihs 2004). Thus, each maternal care style may have benefits and costs to both the mother and the calf. The survivorship of all of the eight calves included in this part of the study to the early juvenile state suggests that this is so (as of November 2006). The existence of social and less social females within a population has also been identified in a population of wild horses (Equus caballus) inhabiting an island with a variety of ecological conditions (Rubenstein 1986). Females living on the side of the island which is more open (low-lying dunes, narrow salt marches, and grassland), form permanent associations. In contrast, females, which live on the side where tall dunes cover virtually the entire area and marshes are restricted to small pockets, form loose associations. Differences in relationships appear to be related to food sources. Females that feed on grass can associate for long periods of time because grasses are densely distributed. On the contrary, small patches of marsh grasses cannot support large groups, and when females aggregate around very small patches, fissioning occurs (Rubenstein 1986). 43

PAGE 56

In this study, differences in the sociality of mothers seemed also to be related to habitat as indicated earlier. The sociality observed in dolphin mother-calf pairs that used deep water more often is similar to that of female horses that use open areas, in that they both form more permanent associations. Conversely, the sociality of dolphin mother-calf pairs found more often in shallow water is similar to that of female horses living in areas with small patches of food; they form loose associations. Differences in the relationships between female dolphins may be related to differences in food assemblage among habitats. Gannon et al. (2005) found that deep water habitats in Sarasota Bay were characterized by fish communities dominated by small pelagic species, while the shallower sea grass and mangrove habitats were dominated by demersal species. Since demersal fish species dwell at or near the bottom and do not form schools, dolphins feeding on this type of fish probably limit close associations as these fish cannot support large dolphin groups. In contrast, pelagic fish frequently travel in large schools that several dolphins can likely feed on, making close association a more common feeding strategy in deep water. Social pressures can also influence an individuals decision to join and leave temporary groups. In bonobos (P. paniscus), for example, females tend to leave a group in which a large number of males are present. This female strategy appears to be a way to avoid male harassment (White 1989, 1992). In Shark Bay dolphins (Tursiops spp.), females appear to form temporary associations with other females to reduce harassment from male groups, which can be very intense (Connor et al. 1992). In other primate species, a female can form a close association with a single adult male, or multiple 44

PAGE 57

females form close associations with one or more adult males (van Schaik 1989, Palombit et al. 1997, Treves and Chapman 1996, van Schaik and Kappeler 1997, Wrangham 1980). In Sarasota Bay, all females having close relationships with their calves had a higher number of associates. In the presence of many other dolphins, it might be harder for a female to keep track of her calfs movements and/or for mothers and calves to find each other if they separate. This could be particularly troublesome in cases in which mothers and calves are separated by more than 200 m. In one instance, a female and her calf were determined to be even farther from each other, 475 m (calculated with laser range finder). It is worthy of note that the mean separation distance of calves was significantly lower than the mean distance from focal females at which non-associate dolphins (satellites) were observed. Thus, calves may remain within a range in which non-associates are normally not found. In other dolphin species (Tursiops spp.), females are intolerant of separations from their calves in the presence of others during the first week of the life of the calf (Mann and Smuts 1998). In captivity, female bottlenose dolphins produce aggressive contact calls when calves separate short distances (>2 m) during the first 9-10 months after the calfs birth (McCowan and Reiss 2005). Different maternal care styles, staying close or frequently separating, probably have implications for the social and vocal development of the calves. Bottlenose dolphins are known to develop individually distinctive whistles known as signature whistles (Caldwell et al. 1990). The acoustic features of those whistles seem to be influenced by the early social environment of the calf (Fripp et al. 2004). Thus, whether a female has many or few associates may affect signature whistle development. The sociability of the female may also influence the associations that her calf forms when it becomes 45

PAGE 58

independent. Separations in most mammalian species allow the infant to practice skills on its own and expand its social network (Connor et al. 2000), particularly, in the case of fission-fusion societies. In the case of adult males, some form close association with other males, with whom in some cases they frequently associated as calves or juveniles (Wells 1991, Owen et al. 2002). It would be interesting to see if the male calves of the least social mothers develop close bonds with other males as they become adults. The level of maternal experience of the females did not seem to explain the observed differences in maternal care. Newborns of primiparous mothers have been reported to surface alone more often than the newborns of multiparous mothers, thus increasing the distance between the mother and the calf (Owen 2001). One of the mothers separating frequently from her calf was a multiparous female, who bore 5 calves including the one from the study. The other female was a primiparous mother. Information on interactions between individual dolphins has provided insights into what groups are. Not surprisingly, females and calves seem to depend on close range interactions, although females and calves sometimes separate when they are feeding or traveling. Other social relationships such as that of an adult male and a potentially receptive female (the mother) also depend on close range interactions since the male needs to asses the reproductive status of the female. Thus, definitions of a dolphin group must consider that cohesiveness occurs within a small radius (10-20 m) but also that as individual dolphins temporarily move outside of that radius, they still may maintain a slightly less cohesive kind of grouping. The distance over which individuals spread depends upon its composition and location (habitat) at a given time. This definition of 46

PAGE 59

47 groups in a species with flexible grouping patter ns is discussed in more detail in Chapter 5. Any biologically meaningful definition of a group must also take into account the abilities of animals to locate and identify a ssociates using sensory mechanisms, including communication. This is particularly important in species such as bottlenose dolphins that have individual-specifi c social relationships requiring lo cation and recognition of specific patterns (Janik and Slater 1998, Connor et al. 2000, Tyack 2000a). Id entification of the signals used by bottlenose dolphins during tempor ary separations is the focus of the next chapter.

PAGE 60

Focal female FB15 FB54 FB55 FB59 FB65 FB83 FB90 F101 F119 F141 F149 F157 Total Number of focal follows 4 9 9 3 6 3 8 3 4 7 3 7 49 Number of focal time (min) 429 1476 741 642 942 618 1182 618 624 1215 561 783 9831 Separations of calves 4 68 38 1 6 18 7 21 12 7 1 3 185 Separations of associates 2 4 1 8 12 4 13 2 10 38 12 31 138 Total separations 6 71 38 9 17 22 20 23 22 40 13 32 323 Table 2.1. Summary of the observations conducted on each focal female bottlenose dolphin in Sarasota Bay, Florida. 48

PAGE 61

Table 2.2. Characteristics of the dependent calves of 8 focal females and the corresponding time that they were separated from their mother during focal animal observations conducted in Sarasota Bay, Florida. Separation time is expressed in minutes. Note: F = female, M = male, and U = unknown. FB54 FB55 FB65 FB90 F111 F141 F149 F157 Calf number 5 th 1 st 3 rd 5 th 6 th U 3 rd 3 rd Calf gender F M M F U F U F Total number of separations 68 38 6 7 2 7 1 3 Mean separation time SD 8.45 8 8.86 7 13.33 15 10.57 14 16.25 0.26 11.57 9 14.00 NA 9.33 11 Total separation time 575 337 80 74 16 81 14 28 49

PAGE 62

Table 2.3. Total number of fission-fusion events recorded during focal follows conducted in the summer of 2001, 2002, and 2003 in Sarasota Bay, Florida. A positive change indicates that a dolphin(s) joined and a negative change indicates that a dolphin(s) left the focal mother. Rates are expressed as number of associate changes per min. Focal female No. (+) changes No. (-) changes Total No. fission-fusion events Mean rate of fission-fusion % of changes observed during total follow time FB15 2 1 3 0.01 0.70 FB59 8 8 16 0.06 2.49 FB83 4 4 8 0.01 1.29 F101 2 4 6 0.01 0.97 F119 10 8 18 0.03 2.88 FB54 7 6 13 0.01 0.88 FB55 3 8 11 0.02 1.48 FB65 16 9 25 0.02 2.65 FB90 28 21 49 0.05 4.15 F157 17 15 32 0.03 4.09 F141 36 35 71 0.04 5.70 F149 18 14 32 0.05 5.84 50

PAGE 63

Table 2.4. Fission-fusion rates estimated for each focal female including total number of dolphin encounters and encounter type (associates and satellite). Focal female Fission-fusion rates Total of dolphin encounters Satellite encounters Associate encounters FB54 Low 27 20 (74%) 7 (26%) FB65 Moderate 19 3 (16%) 16 (84%) F119 Moderate 10 0 (0%) 10 (100%) FB55 High 6 3 (50%) 3 (50%) FB59 High 9 1 (11%) 8 (11%) FB90 High 36 8 (22%) 28 (78%) F141 High 50 14 (28%) 36 (72%) F157 High 20 3 (15%) 17 (85%) Total 177 52 125 51

PAGE 64

Table 2.5. Identified associates (gray) and satellites (yellow) observed at different distances from focal females. The numbers at each distance are the number of independent sightings of each dolphin. Gender and relatedness are included when known. Female Associates and satellites Distance in meters at which associates were observed ID Gender Relatedness Code 10 20 30 40 50 60 70 80 90 100 150 200 FB15 F Daughter (1998) F123 116 6 4 3 1 4 1 1 M FB78 5 1 1 2 1 M C843 3 F FB13 1 F F127 1 FB54 F Daughter (2000) F135 263 22 24 11 19 20 15 14 11 24 8 7 F F141 10 3 2 2 1 1 1 F F151 10 4 1 2 3 2 1 1 5 2 F F155 9 1 3 1 1 3 F F197 9 1 3 1 1 3 F F139 3 5 1 3 2 1 1 F F131 1 1 1 1 1 1 2 3 2 F F157 1 2 2 2 1 2 1 1 1 6 4 3 F F137 1 2 2 2 1 2 1 1 1 6 4 3 M Brother (1996) F148 1 1 1 2 2 M F106 2 1 1 M F222 2 4 1 2 3 2 1 M F220 2 1 M F164 1 1 M F198 1 1 M F230 1 1 M Brother (1992) F118 2 1 5 2 M F228 2 1 1 F F181 1 1 F FB65 2 1 1 F FB09 1 F F177 1 FB55 M Recent son (1999) F218 131 15 6 8 5 5 8 10 6 8 14 3 F F159 1 1 F Mother FB05 2 1 F FB09 2 1 1 1 1 1 F FB33 2 1 F F173 1 1 M F177 2 1 1 1 1 1 M FB36 4 2 1 3 1 M F138 4 1 F Mother FB05 1 F F113 1 52

PAGE 65

Table 2.5 (Continued). Identified associates (gray) and satellites (yellow) observed at different distances from focal females. The numbers at each distance are the independent sightings of each dolphin. Gender and relatedness are included when known. Female Associates and satellites Distance in meters at which associates were observed ID Gender Relatedness Code 10 20 30 40 50 60 70 80 90 100 150 200 FB59 ? Calf (1998) C596 194 3 1 1 3 1 4 1 1 F F157 29 7 3 1 1 1 1 2 3 F FB01 24 2 1 1 1 1 F F113 3 2 1 3 1 1 F Daughter (1988) F131 3 1 2 1 1 M F184 20 1 1 ? IKN2 20 1 1 ? YORI 20 1 1 M F100 12 1 1 M F198 11 1 1 2 1 1 M F192 7 3 1 1 2 M F118 3 1 2 1 F FB09 1 M F176 1 M F190 1 F FB54 1 FB65 M Son (1999) F228 255 2 5 3 3 1 1 2 1 1 F FB07 1 1 F FB09 3 F FB11 49 14 F F179 52 14 F FB75 11 1 1 1 F FB27 1 1 1 1 1 1 2 1 F F113 35 6 1 1 F F131 24 5 F F149 14 10 1 1 1 1 1 ? 1494 16 10 1 1 1 1 1 F F157 26 10 1 F F137 26 10 1 M FB06 1 M FB10 2 3 1 1 1 M FB66 107 10 1 1 M FB76 83 15 5 4 2 1 M FB94 1 1 1 1 1 3 1 4 2 M F177 3 M F224 1 1 1 1 1 1 1 1 M F232 12 2 1 F F175 1 1 1 1 M F226 1 1 1 1 M F106 1 1 1 53

PAGE 66

54 Table 2.5 (Continued). Identified associates (grey) and satellites (yellow) observed at different distances from focal females. The numbers at each distance are the independent sightings of each dolphin. Gender and relatedness are included when known. Female Associates and satellites Distance in meters at which associates were observed ID Gender Relatedness Code 10 20 30 40 50 60 70 80 90 100 150 200 FB83 ? Calf (1998) C835 86 12 11 7 11 9 3 4 2 9 4 4 ? TEEN 29 2 1 1 ? C354 16 3 1 4 1 M F128 11 2 M FB78 3 1 2 2 M FB27 2 M Son (1992) C834 1 1 1 M F142 4 2 M BLSC 1 M PRNK 1 FB90 F Daughter (1999) F133 294 9 3 2 2 2 4 6 F FB07 15 14 1 F FB09 1 8 1 1 1 1 F F177 1 8 1 1 1 1 F FB11 53 21 5 1 2 F FB13 9 1 1 1 F F179 53 21 5 1 2 F FB27 19 12 F FB33 4 7 1 1 F FB43 8 1 1 M F224 15 14 F FB65 11 3 2 5 3 1 1 M F228 11 3 2 2 4 1 1 F FB75 1 5 5 3 1 1 1 1 F FB93 12 2 1 2 1 1 2 ? C932 12 2 1 2 1 1 2 M F101 4 1 2 M F187 4 1 2 F Daughter (1996) F113 50 11 3 1 2 2 1 1 2 1 1 F F173 4 7 1 1 F F175 34 3 6 2 1 1 1 M F106 58 49 M F138 1 1 1 M F226 38 2 6 2 1 1 1 1 M F232 7 2 4 5 1 1 1 F FB25 1 1 ? C255 1 1 F F175 1 1 M F226 1 1 M FB36 1 M FB94 3

PAGE 67

Table 2.5 (Continued). Identified associates (gray) and satellites (yellow) observed at different distances from focal females. The numbers at each distance are the independent sightings of each dolphin. Gender and relatedness are included when known. Female Associates and satellites Distance at which associates were observed ID Gender Relatedness Code 10 20 30 40 50 60 70 80 90 100 150 200 F101 M Son (1998) F196 135 6 11 8 4 2 3 4 9 4 6 F F119 2 1 1 1 M FB36 40 1 1 1 2 2 M FB92 7 1 2 1 1 1 2 1 1 M FB06 7 1 3 1 1 1 F FB03 1 1 1 3 M FB14 1 1 1 3 M FB94 1 1 1 3 M F128 1 F119 F Daughter (1998) F125 101 10 7 5 5 9 2 4 1 10 5 5 F FB13 11 1 1 F F101 7 1 1 1 F FB79 3 F F109 3 F F111 2 1 F FB03 1 1 M F142 14 1 M F100 5 1 M F182 4 1 M FB44 2 1 F Grandmother FB43 1 F141 F Daughter (2000) F151 343 11 5 5 1 3 1 1 F FB05 11 F FB07 2 2 1 1 1 F FB63 21 F FB65 2 1 3 1 1 1 1 1 F FB90 1 2 1 F F113 16 1 1 1 F F133 1 1 F F137 50 30 8 10 7 1 2 2 2 4 F F149 2 1 1 1 ? 1494 2 1 1 1 F F155 29 13 2 2 3 3 2 3 1 3 F F157 70 26 5 9 6 1 1 1 2 2 4 F F159 13 23 4 8 1 1 1 1 F F163 72 11 4 5 1 4 1 1 F F181 20 23 3 3 1 3 1 1 1 F F197 8 12 2 1 2 3 2 3 3 F PALM 62 18 1 5 3 3 1 55

PAGE 68

Table 2.5 (Continued). Identified associates (gray) and satellites (yellow) observed at different distances from focal females. The numbers at each distance are the independent sightings of each dolphin. Gender and relatedness are included when known. Female Associates and satellites Distance at which associates were observed ID Gender Relatedness Code 10 20 30 40 50 60 70 80 90 100 150 200 F141 ? PAL1 61 19 1 5 3 2 1 M FB10 2 6 1 2 M F118 7 8 1 4 2 1 2 1 M F126 41 7 3 1 1 1 M F148 1 4 1 M F222 1 1 1 1 1 2 2 M F228 2 3 1 1 1 1 M F230 16 12 3 5 1 3 2 2 1 1 F FB65 1 F F197 1 1 1 F F155 2 1 M F222 1 1 1 2 1 M F228 1 F149 ? Recent offspring (1493) 1493 164 2 1 1 1 1 1 F FB11 13 2 3 1 1 1 2 F FB33 1 1 F FB65 17 1 1 1 1 F F113 52 6 4 2 1 2 1 F F137 33 10 2 3 2 1 1 3 1 2 F F139 20 6 5 5 5 1 2 1 F F157 37 8 2 4 2 1 3 1 2 F F165 6 1 1 1 F F173 1 1 F F179 13 2 3 1 1 1 2 F F181 27 7 6 5 5 2 2 2 3 5 1 M FB14 47 1 1 M FB94 8 17 6 1 10 1 2 1 1 1 1 M F136 2 M F188 7 5 1 3 1 M F198 5 3 2 1 1 M F220 12 6 5 4 4 1 1 M F228 17 1 1 1 1 M F230 25 7 6 5 5 2 2 2 3 4 1 56

PAGE 69

Table 2.5 (Continued). Identified associates (gray color) and satellites (yellow color) observed at different distances from focal females. The numbers at each distance are the independent sightings of each dolphin. Gender and relatedness are included when known. Female Associates and satellites Distance at which associates were observed ID Gender Relatedness Code 10 20 30 40 50 60 70 80 90 100 150 200 F157 F Recent daughter (2000) F137 186 11 1 4 2 2 5 1 2 1 2 F FB75 11 1 F FB79 11 1 F FB90 2 1 1 F F133 2 1 1 F F155 8 7 2 5 1 1 1 3 2 1 F F159 19 16 1 1 F F163 6 2 1 1 1 2 2 F F165 11 1 F F175 12 1 1 1 1 2 1 1 F F197 11 6 2 3 1 1 1 3 2 1 F PALM 14 3 2 4 3 3 2 2 2 3 3 ? PAL1 19 5 1 2 2 1 5 2 1 F HUEY 17 4 4 1 2 2 2 2 3 1 ? HUE1 17 4 4 1 2 2 2 2 3 1 M FB10 8 11 1 M FB36 34 2 2 1 2 M F118 9 12 1 1 1 1 M F136 8 1 2 M F164 7 13 1 M F226 12 1 1 1 1 2 1 1 ? LBMN 9 4 1 1 1 1 1 1 1 F F155 1 F F197 1 M F222 1 57

PAGE 70

Table 2.6. Summary of dynamics quantified for 8 focal females. Note: Habitat types are shallow water (S) and deep water (DW); levels of association are low (L), moderate-low (M-L), and moderate (M), and overall fission-fusion rates are low (L), moderate (M), and high (H). Focal female Most common type of separation Habitat most commonly used Levels of association Overall fission-fusion rate FB54 Calf SW L L FB55 Calf SW L H F119 Calf/associates SW/DW L M FB90 Associates DW L, M-L H F157 Associates DW L, M-L H FB59 Associates DW L, M-L, M H FB65 Associates DW L, M-L, M M F141 Associates DW L, M-L, M H 58

PAGE 71

Figure 2.1. Mean distance error of 633 distance estimates practiced every field day. 59

PAGE 72

Figure 2.2. Illustration of two types fission-fusion events observed in wild bottlenose dolphin residents of Sarasota Bay. Arrows indicate the direction that an associate moves relative to the focal mother. 60

PAGE 73

Time at 3-min intervals 14:31:2214:37:2214:43:2214:49:2214:55:2215:01:2215:07:2215:13:2215:19:2215:25:2215:31:2215:37:2215:43:2215:49:2215:55:2216:01:2216:07:2216:13:2216:19:2216:25:2216:31:2216:37:2216:43:22 50100150200>200 Beginning of a temporary union Temporary separationDistance of associate from focal female (meters) Mother and calf Associate20 Time at 3-min intervals 14:31:2214:37:2214:43:2214:49:2214:55:2215:01:2215:07:2215:13:2215:19:2215:25:2215:31:2215:37:2215:43:2215:49:2215:55:2216:01:2216:07:2216:13:2216:19:2216:25:2216:31:2216:37:2216:43:22 50100150200>200 Beginning of a temporary union Temporary separationDistance of associate from focal female (meters) Mother and calf Associate20 Figure 2.3. Spatial position of an associate observed with a female (FB65) during a follow conducted in August 26, 2002. Blue diamonds represent the position of an associate at 3-min intervals from the female. Red diamonds represent the mother-calf pair. 61

PAGE 74

3-min time point No. dolphins joining No. dolphins leaving Time interval before a dolphin joins Time interval before a dolphin leaves 1 2 3 4 2 5 6 9 7 2 8 9 1 10 12 6 11 2 3 12 13 9 14 1 15 16 17 18 19 20 Average time 10.5 min 7.5 min Figure 2.4. Diagram showing the + changes and changes of associates in a typical focal follow. Gray indicates + changes and the time intervals between them. Light blue indicates changes and the time intervals between them. 62

PAGE 75

Time in minutes 02468101214Frequency369121518212427303336394245485154576063666972757881848790939699102105108111114117120123126129 Time in minutes 02468101214Frequency369121518212427303336394245485154576063666972757881848790939699102105108111114117120123126129 Figure 2.5. Duration of associations in 3-min intervals between focal females and other dolphins observed in Sarasota Bay. 63

PAGE 76

Figure 2.6. Central tendency and variability of frequency with which associates leave and join focal females during focal follows conducted in Sarasota Bay, Florida. The solid line drawn across each box represents the median in each event (joining or leaving). The lower boundary of a box is the 25 th percentile, while the upper boundary is the 75 th percentile. The tines on top and bottom of each box represent the largest and smallest frequency values, respectively, that d not include outliers or extreme values. Boxplot legend: o = Outlier and = extreme value. 64

PAGE 77

Figure 2.7. Central tendency and variability of the duration of association of females with different types of dolphins in Sarasota Bay, Florida. The solid line drawn across each box represents the median in each event (joining or leaving). The lower boundary of a box is the 25 th percentile, while the upper boundary is the 75 th percentile. The tines on top and bottom of each box represent the largest and smallest frequency values, respectively, that d not include outliers or extreme values. Boxplot legend: o = Outlier and = extreme value. 65

PAGE 78

R2= 0.7010203040500.000.010.020.030.040.050.060.07Mean rate of fission-fusionNumber of associates FB54FB55FB59F141FB90F157F119FB59 R2= 0.7010203040500.000.010.020.030.040.050.060.07Mean rate of fission-fusionNumber of associates FB54FB55FB59F141FB90F157F119FB59 Figure 2.8. Mean rate of fission-fusion events (changes/min) and number of identifiable associates for each focal female during focal follows conducted in Sarasota Bay, Florida. 66

PAGE 79

60 30 m84 46 m 118 58 m 150m foal feales 050100200SatellitesDependent calvesAssociatesMean distance frocm Figure 2.9. Mean distance of satellites from focal mothers and mean distance of separation of focal mothers and their dependent calves, and focal mother/calf pairs and other associates. 67

PAGE 80

R2= 0.7201020304050304050607080Total number of identifiable associatesPercentage of time that a dependent calf was the nearest neighbor of its mother F119FB54FB55F157FB65FB90FB59F141 60 R2= 0.7201020304050304050607080Total number of identifiable associatesPercentage of time that a dependent calf was the nearest neighbor of its mother F119FB54FB55F157FB65FB90FB59F141 60 Figure 2.10. Relationship between the number of associates and the percentage of time that a female and her calf were nearest neighbors. 68

PAGE 81

20 m 10 m 5 m 20 m 10 m 5 m Females with calves Adult males Juveniles Single adult femalesA. Most common type of associates B. Second most common type of associates Figure 2.11. Categories of associates found at 5-m, 10-m, and 20-m from focal mothers. 69

PAGE 82

0102030405030405060708090100Distance from focal females in metersPercentage of observationss Males Juveniles 0102030405030405060708090100Distance from focal females in metersPercentage of observationss Males Juveniles Figure 2.12. Distances at which juveniles and adult males were most commonly sighted from focal females during temporary separations. 70

PAGE 83

0204060801005102030405060708090100130150200Distance of associates from focal females in meters age hing s Figure 2.13. Headings recorded for all dolphins (associates and calves) sighted at each distance category within the observation zone. Pe rcent of ead Parallel Heading towards Heading away Variable 0204060801005102030405060708090100130150200Distance of associates from focal females in metersage hing s Parallel Heading towards Heading away Variable ead of Pe rcent 71

PAGE 84

3. ECHOLOCATION AND WHISTLE USE DURING FISSION-FUSION EVENTS IN BOTTLENOSE DOLPHINS 3.1. Introduction All group living social mammals must have mechanisms to form and maintain associations with conspecifics, and to do it, they must have effective communication signals. This is particularly important in cases in which partners leave and join one another, because individuals must use some type of signal to find and locate each other. The bottlenose dolphin (Tursiops truncatus) exhibits a fission-fusion social organization characterized by temporary associations in which individuals remain together for several minutes and then split up, or after diverging by hundreds of meters, individuals reunite, only to separate again later. In such cases, dolphins must be able to find particular associates. Acoustic signals are an important means of locating a partner in species with social bonds (Nishida 1968, Byrne 1981, Boinski and Campbell 1995). They are important in maintaining group cohesion in areas with limited visibility (e.g., dense forest, Byrne 1981; murky waters, Janik and Slater 1998) or when associates need to communicate over distances ranging up to several kilometers (Tyack 2000a,b). In the bottlenose dolphins, whistles can be used for long distance communication (Janik 2000a, 72

PAGE 85

Tyack 2000a,b). Whistles are narrow-band frequency-modulated sounds ranging from about 4 to 20 kHz (Caldwell et al. 1990). One study of wild Tursiops aduncus found that whistles are mainly produced when mothers and calves are separated and that whistles tend to be produced during the reunion process (Smolker et al. 1993). The same study presents data showing that most events (79%) containing whistles occurred at distances 20 m. Whistle use in separations involving other age classes or at greater distances was less clear. Smolker et al. (1993) reported two cases of reunions among adult T. aduncus in which whistles were produced as animals joined. Watwood (2003) examined whistle usage in separations of wild adult male T. truncatus, and found that whistles were not always used during reunions. Of the 19 separations reported, 42% of the events did not involve whistles. In captivity, juvenile and adult bottlenose dolphins seemed to use whistles infrequently (Janik and Slater 1998). The different observations raise the question of how dolphins can find and locate each other when whistles are not used. Bottlenose dolphins produce other types of acoustic signals besides whistles. Wild bottlenose dolphins sometimes use echolocation to detect and find distant prey (Rossbach and Herzing 1997, Herzing and dos Santos 2003) and perhaps to discriminate specific prey species (Herzing 2004). Echolocation clicks of bottlenose dolphins are broadband, non-modulated sounds with frequency components from about 1 to 120 kHz (Au 2004, Herzing 2004). In captivity, they have been trained to demonstrate their ability to use echolocation to discriminate the shape, diameter, range, material composition, thickness, and texture of targets in the water (Au 1993, 1997, 2004; Nachtigall 1980). Bottlenose dolphins could use echolocation and/or eavesdropping rather than actively whistling to each other (Dawson 1991) when separated. Work by Xitco and Roitblat (1996) indicates 73

PAGE 86

that a non-echolocating dolphin can eavesdrop on the echolocation signal of another dolphin and can even derive characteristics of the sample object from the echoes. Bottlenose dolphins have been shown to gather information in predator-prey interactions by listening to environmental sounds (Gannon et al. 2005). The use of the term eavesdropping has been used in predator-prey relations. However, the same term is also used to describe other level of transfer of information (McGregor and Dabelsteen 1996). For example, in other echolocating species such as bats, individuals eavesdrop on echolocation calls, helping individual bats to fly without bumping into each other (Fenton 2003). Thus, echolocation is also indirectly used to assess the position of conspecifics at the same time that bats are using the echoes of their own calls to visualize the features of the environment. In the European free-tailed bat (Tadarida teniotis), flying bats used echolocation to keep track of each other and social buzzes to maintain contact with conspecifics (Fenton 2003). Echolocation is typically produced continuously throughout the flight of a bat whereas social calls are used occasionally (Pfalzer and Kush 2003). It is possible that bottlenose dolphins use echolocation as bats do. Although echolocation is normally associated with traveling and feeding, bottlenose dolphins have been reported to use echolocation in social contexts like courtship, aggression, and play (Herzing 2004). The non-traditional uses of echolocation raise the question of whether dolphins use echolocation directly or indirectly (eavesdrop) during temporary separations to maintain contact. The goals of the present study were 3-fold. First, I examined if echolocation and whistles were used during fission-fusion events by wild bottlenose dolphins in Sarasota 74

PAGE 87

Bay, FL, a shallow water system with low (< 5 m) visibility. I studied two types of fission-fusion events, referred to as temporary separations, potentially having different communication patterns: 1) separations of mothers from their dependent calves, and 2) separations of mothers with calves from other associates. The advantage of sharing identity information between mother and calf as they separate is clear. The mother-calf association is the strongest association for adult dolphin females; mothers typically stay with their calves for a period of three to six years (Wells et al. 1987, Wells 1991, 2003). However, communication patterns are less predictable for mothers and other associates. Second, I examined whether mother-calf pairs and their associates used echolocation when they did not use whistles during temporary separations. I quantified the number of temporary separations with echolocation trains and whistles to determine what signal was most commonly used. I expected that when whistles were used, the rate would be higher for communication within mother-calf pairs than between mothers and other associates because the motivation for reunion of a mother with her dependent calf is probably stronger. When whistles were not used, I expected that echolocation would be used, either directly or indirectly (eavesdrop), to keep track of each others position and thus facilitate reunions. I quantified the number of temporary separations with echolocation trains and whistles to determine what signal was used most commonly. Third, I also examined the relationship between the type of acoustic signal produced and the maximum distance of separation between dolphins. Since echolocation clicks and whistles have different frequency ranges, it is possible that one acoustic signal travels farther than does the other and that the distance of separation between dolphins is related to the propagation characteristics of the signal used. Indeed, the low frequency 75

PAGE 88

components of sounds in the whistle range do not propagate as far as high frequencies in very shallow water (Forrest et al. 1993, Forrest 1994, D. Nowacek et al. 2001, Quintana-Rizzo et al. 2006). We examined the characteristics of echolocation and whistles used during separations in different habitats. In particular, I compared echolocation rate, whistle rate, whistle duration, whistle frequency range, and whistle minimum and maximum frequencies. Such parameters have been identified as important sound properties allowing the transfer of information over long distances (Klump 1996) and some mammals modify them according to sound propagation characteristics of different habitats (Waser and Brown 1986, Schnitzler et al. 2003, Wund 2005). 3.2. Methods 3.2.1. Study Animals and Study Area Seven well-known, easily-recognizable mothers with dependent calves (3-4 years old) were the focus of the research. They are members of a year-round resident community of about 160 wild bottlenose dolphins living in the vicinity of Sarasota Bay, Florida. The community has been studied since 1970 and is the focus of the longest-running study of wild dolphins in the world. Because of this, much information exists about the dolphin community including identity, sex, age, reproductive status, and genetic relatedness of most of the long-term residents. Their flexible social organization has been described elsewhere (Wells et al. 1980, 1987, Wells 1991, 2003). 76

PAGE 89

The study area extended southward from the southern edge of Tampa Bay to Siesta Key, off Sarasota. The area includes the inshore waters of Sarasota Bay and the Gulf of Mexico waters within approximately 1 km of shore (Wells et al. 1987). 3.2.2. Animal Observations Observations of focal resident female dolphins were conducted in August 2002 and from June to September 2003 from a 7-m-long boat equipped with a 115 hp 4-stroke outboard engine. The boat was kept at a distance of approximately 20 m from focal females. Positional data were recorded continuously using a combination of focal animal and group scanning observations (Altmann 1974, Mann 2000) on the mother of mother-calf pairs. Positional data were recorded by the primary observer (EQR) as the estimated distance between the focal mother, her dependent calf, and any other dolphin at each surfacing of the mother. Positional data allowed determination of the time at which dolphins separated and united, the total duration of the separation event, and the maximum distance of separation between individuals. Trained observers (2-3) helped to keep track of dolphins. At 3-min interval (instantaneous point sampling technique, Altmann 1974), I recorded the behavioral activity and environmental data. The behavioral activities of the focal female and of the individuals within the observation zone were recorded using a combination of focal animal and scan sampling observations. Behavioral activities recorded were traveling, milling, socializing, and feeding. Traveling was defined as synchronous and directional movement in a straight line or zigzag, milling was defined as a non-directional moment, and socializing was defined as active interaction with at least 77

PAGE 90

one other dolphin (Waples 1995, Urian and Wells 1996). Feeding was recorded if a dolphin(s) had a fish in its mouth. Probable feeding was also used for the behavioral activity analysis based on dolphin behavior (Shane 1990, Waples 1995). Probable feeding was defined as dolphins swimming in circles near the surface of the water; dolphins, alone or in loose groups, repeatedly diving in varying directions at one location; dolphins swimming against a strong tidal current and remaining in one place; dolphins chasing or striking a fish with its flukes; and dolphins increasing swimming speed suddenly then, spinning in a circle or making a hairpin turn (pin wheeling; Shane 1990, Urian and Wells 1996). Observations were done only when sea state scale was equal to 0 or 1. At 3-min intervals, environmental data were recorded including the habitat (shallow water, channel, gulf, bay, sand bar, and edge) and location of the research vessel as an approximation for the location of the dolphins (latitude and longitude coordinates were obtained from a Garmin GPS 12 Personal Navigator). Habitat was categorized as shallow ( < 3 m deep, characterized by seagrass and sand patches), channels (inshore area several meters wide and depth > 3 m), bay (inshore area several kilometers wide and depth > 3 m), gulf (offshore waters of Sarasota Bay), or edge (area between habitats approximately 5-m wide). Two types of temporary separations were examined: 1) separations of mothers from their dependent calves and 2) separations of mothers with their calves from other associates. A separation was considered to have occurred when the distance between a focal female and another animal increased to more than 20 m. The event ended when the distance between a focal female and the other animal decreased and was equal to or less than 20 m. A distance of 20 m was chosen because 75% (n = 2087) of the dolphins 78

PAGE 91

observed in 53 pilot focal follows conducted in 2001 were within such a distance from focal females. The percentage of animals observed at other distances was less than 5%. The observation zone had a radius of 200 m from the focal female, because this was the maximum range at which I could observe dolphins accurately. To estimate the distance between focal mothers and other dolphins, all observers (3-4) practiced distance estimates using a laser range finder on fixed objects located in the water every field day prior to data collection. The primary observer (EQR) estimated approximately 92% of the distances of the dolphins during focal follows; other observers were also trained to estimate distances when their assistance was needed. Estimates were done on objects located at distances between 10 and 250 m from the research boat when anchored. 3.2.3. Acoustic Recordings Continuous acoustic recordings were done during focal animal observations. The set-up of the acoustic recording was as follows. At the bow of the observation boat, two 1.5 m sections of PVC pipe were joined in a T joint and secured across the gunwales (Figure 3.1; Sayigh et al. 1993). On each side of the boat approximately 2 m of hydrophone cable was extended from the end of the pipe into the water. A calibrated HTI-96-MIN hydrophone (sensitivity -169.8 dB re: 1V/Pa) at the end of each cable was approximately 1 m below the surface when the boat was not moving. I used two hydrophones to ensure that there would be a backup recording. To prevent the hydrophone from bouncing at the surface while the boat was moving, the cable was weighted with a chain attached to the end of the PVC pipe by a carabiner. Each 79

PAGE 92

hydrophone was connected to a 2-kHz high-pass filter to reduce engine noise. The research vessel was kept at a distance of approximately 20 m from the mother. The distance of the research vessel to the calf/associates varied depending on how far the animal was from the focal mother. Signals from each hydrophone were digitized at 48.8 kHz with a Tucker-Davis Technologies RP2 module, and stored on a computer hard drive. Signals from each channel were sampled simultaneously with a 24-bit resolution and were stored as 32-bit floating point values. 3.2.4. Data Analysis Frequency of the various activities was calculated by counting occurrence of each activity in the focal sessions and dividing by the total number of data points to achieve a percentage (Shane 1990). Whistles and trains of echolocation clicks were first visually identified using Cool Edit 2000 and whistle parameters were later quantified using MATLAB 6.5. For the purpose of data analysis, I used only whistle spectrograms in which the details of the spectral contour of whistles were visible (Parijs et al. 2002). The analysis of click sounds focused on their lower frequency components, at 24 kHz and below, because the sampling rate was 48.8 kHz (Nowacek 1999, dos Santos and Almada 2004, Herzing 2004). Thus, it is possible that click sounds were underestimated since the frequency components of clicks are from about 1 to 120 kHz (Au 2004, Herzing 2004). It is also possible that click sounds were under-represented because the high degree of directionality of dolphin echolocation clicks. The -10 dB beam width (i.e. the beam width where the signal level decreases by 10 dB from the peak level) is approximately 80

PAGE 93

21 (Au et al. 1986, Au 1993). Because I was rarely positioned in front of the dolphins, our acoustic recording equipment was probably not often in the direction of their narrow beam even when the research vessel was at a distance of approximately 20 m from the focal female. Due to those limitations, I analyzed trains of echolocation clicks instead of single clicks. A train of echolocation clicks was scored as a single event (Nowacek 2005) and a pause of 500 ms (Au pers. comm.) was considered to be the end of the click train. I refer to trains of echolocation clicks simply as echolocation or echolocation trains hereafter. Whistle rates and echolocation rates were calculated for all separations. In calf separations, events in which only the mother and calf were present in the observation zone were included in the analysis. In associate separations, I included only events in which the same animals (identity and number) separating from a focal mother during a particular event joined her again in the same event. In all of the analyses, no other dolphins were observed within the observation zone (200 m). Rates were calculated as signals produced per min as described by Benoit-Bird and Au (2004) and dos Santos and Almada (2004). I examined whether there was a relationship between the independent variables, maximum separation distance and whole separation time, and the two separate dependent variables, whistle rate and echolocation rate, using a multiple linear regression on the log (x +1) transformed data. I also examined whether there was a relationship between whistle rates, echolocation rates, and group size using a linear regression. To control for effects of group size I compared whistle and echolocation rates during a separation with those in the 5-min period before and after a separation. The analysis excluded separations in which the interval between fission-fusion events was 81

PAGE 94

less than 5-min. I wanted to compare sections of time that were long enough to identify possible rate changes. I assumed that events that were less than 5-min long are not independent from each other. I examined rates throughout the durations of the different separations to identify sections in which signals were most commonly used. To compare separations with different durations, total times were standardized by converting them to percentages as described by Smolker et al. (1993) and Watwood (2003). Separation times as percentages were divided into 5 sections, each equal to 20%. Analysis among sections was done with a Kruskal-Wallis test followed by a Mann-Whitney U post-hoc test, because the data did not meet the assumptions of parametric tests. In the case of whistles, I also quantified the following parameters using MATLAB 6.5: highest frequency (kHz), lowest frequency (kHz), and duration (ms). I also quantified the frequency range of each whistle by subtracting the measured lowest frequency from the highest frequency. Measurements from spectrograms were restricted to the fundamental frequencies of each whistle and harmonics were not considered (FFT = 1024 points, temporal resolution = 10 ms). Whistle parameters, whistle rates, and echolocation train rates were compared between habitats. For test accuracy, I included only separations with complete recordings. In all habitat analyses, I omitted habitats with a small sample size of separation events ( 3). Thus, edge, bay, and gulf habitats were not analyzed and only differences between shallow waters and channels were examined. I used the non-parametric Mann-Whitney U test to determine differences between habitats, because the data did not follow the assumptions of normality and homogeneity of variance of parametric tests. Statistical 82

PAGE 95

tests were conducted with SPSS 14.0 package (2005). Results are expressed as means SD and all tests are two-tailed, with a significance level of 0.05. 3.3. Results 3.3.1. Animal Observations A total of 230 temporary separations was observed in 117.2 h of direct observation. Total separation time was 29.2 h. Of these separations, 118 were of females from their dependent calves and 112 were separations of females with calves from their other associates. Each type of temporary separation was variably among the seven focal females (Table 3.1). Two focal mothers (FB54 and FB55) mainly exhibited temporary separations of their dependent calves (94%, n = 103) and the other five mothers (FB65, FB90, F141, F149, and F157) exhibited mostly temporary separations from associates other than their calves (65%, n = 106). Since separation types were not evenly distributed among females, results of whistle and echolocation production may be more representative of the behavior of the group of females having the largest sample size of separations of a particular type. The habitat in which 203 of the 230 temporary separations occurred was identified. Separations between females and their dependent calves occurred mainly in shallow water areas (55%, 62 of 112) and separations of females from their associates occurred mainly in channels (59%, 54 of 91). The second most common habitat of separations of mothers and calves was channel (24%, 27 of 91) and of mothers with 83

PAGE 96

calves from their associates was shallow water (27%, 25 of 91). Separations of mothers with calves from their associates were observed less frequently in all other habitats (< 6%). 3.3.2. Acoustic Signals Used in Temporary Separations I found that echolocation and whistles were produced during temporary separations in all habitats. Of the 230 separations, 198 events had echolocation trains and/or whistles: 86% (n = 170) with echolocation and 56% (n = 111) with whistles. Note that the percentages do not add to 100% because there is an overlap of events. Based on the presence or absence of signals, I identified four types of events: 1) events with no whistles or echolocation, 2) events with only whistles, 3) events with only echolocation and 4) events with both whistles and echolocation. Nearly half of the events did not include whistles (44%, n = 87). Of the non-whistle events, 72% (n = 32%) had echolocation trains and 28% (n = 24) had no signal. Of the whistle events, 4% had only whistles (n = 4) and 96% (n = 107) had whistles and echolocation trains. The recordings of 34 separations were omitted from the calculations because they overlapped with other behavioral events, other dolphins were in the observation zone in mother-calf separations, or the recordings were not complete due to malfunction of the acoustic equipment. Of the 198 separations with echolocation trains and/or whistles, 148 occurred in shallow water areas or channels. The number of separations of calves from mothers in shallow water areas and channels was 47 and 24, respectively (total = 71). The number of associate separations in shallow water areas and channels was 30 and 47, respectively (total = 77). In channels, separations of females from their calves involved mostly 84

PAGE 97

echolocation trains (71%, n = 17). In all other separations, events with echolocation trains and whistles were more common (calves = 47%, n = 22; associates: 73%, n = 22) and they were followed by events with only echolocation trains (calf: 32%, n = 15; associates: 17%, n = 5). In all habitats, events with only whistles (0-4%) and no acoustic signals (4-17%) were less common (Figure 3.2). 3.3.3. Acoustic Signals and Maximum Distance of Separation Of the two independent variables used to predict a relationship with echolocation or whistle rates, only distance contributed significantly to the dependent variable echolocation rate. This relationship was only significant for separations of mothers with calves from associates (full model estimate results with d.f. = 2, N = 72: F = 4.13, p = 0.02; distance coefficient = 0.57, SE = 0.23, p = 0.02; time coefficient = -0.40, SE = 0.15, p = 0.80). In other cases, there were no relationships between the independent variables maximum separation distance and whole separation time, and the dependent variables whistle rate (calf separations: F = 1.06, d.f. = 2, p = 0.36, N = 33; associate separations: F = 1.76, d.f. = 2, p = 0.18, N = 72) and echolocation rate (calf separations: F = 0.75, d.f. = 2, p = 0.47, N = 68). However, there was a significant difference in the maximum distance of separation among events with no signals, echolocation trains, and both whistles and echolocation (calves: Kruskal-Wallis = 17.64, p < 0.001, N = 91; associates: Kruskal-Wallis = 12.34, p = 0.002, N = 99). In both calf and associate separations, the maximum distance of separation was significantly greater in events with both echolocation trains and whistles (Figure 3.3). Maximum distance of separation was moderate in events with echolocation trains and low in events with no acoustic signal 85

PAGE 98

(Figure 3.3). Events having only whistles were not included in the analysis because of their sample size was very small (< 3). The mean maximum distance of separation between mothers and calves in the two whistle only events was 60 28 m. Maximum distance of separation between mother and associates in the only whistle only event was 40 m. 3.3.4. Echolocation and Whistle Rates No significant relationship was found between whistle rate and group size and between echolocation rate and group size in associate separations (Figure 3.4). Comparisons involving echolocation rates in all separations were done using data from two types of separation events: events with only echolocation and events with both echolocation and whistles. Data were combined because no significant difference was found in the echolocation rates of the two events (calf: Mann-Whitney U test = 535.00, p = 0.23, N = 72; associates: Mann-Whitney U test = 297.00, p = 0.26, N = 60). Echolocation and whistle rates were higher during a separation than during the 5-min period before and after the event (Figure 3.5). In calf separations, mean echolocation and mean whistle rates were between seven and ten times higher, respectively, during a separation than in the 5-min periods before and after the event (Figure 3.5; echolocation rates: U = 49.46, p = 0.00, N = 98 and whistle rates: U = 23.16, p = 0.00, N = 50). In associate separations, both mean whistle and echolocation rates were three times higher during a separation than in the 5-min periods before and after the event (Figure 3.6; echolocation rates: U = 25.56, p = 0.00, N = 98; whistle rates: U = 18.20, p = 0.00, N = 74). 86

PAGE 99

Echolocation rate was greater than whistle rate in all separations (Table 3.3; calf: U = 475.50, p < 0.001, N = 103; associates: U = 851.50, p = 0.002, N = 94). In all events, echolocation rate was not significantly different among the five time segments within a separation (calf separations channel: Kruskal-Wallis test = 3.37, p = 0.50, N = 220; shallow water: Kruskal-Wallis test = 7.27, p = 0.12, N = 330; associate separations channel: Kruskal-Wallis test = 0.80, p = 0.94, N = 370; shallow water: Kruskall-Wallis test = 6.12, p = 0.20, N = 230). Yet, when whistles were used, differences were found between habitats. In all events observed in channels, mean whistle rate was not significantly different among the time segments of a separation (calf separation: Kruskal-Wallis test = 3.83, p = 0.43, N = 70; associate separation: Kruskal-Wallis test = 5.00, p = 0.29, N = 240). However, in separations occurring in shallow water areas, whistle rate was variable (calf separation: Kruskal-Wallis test = 21.14, p < 0.001, N = 150; associate separation: Kruskal-Wallis test = 11.05, p = 0.03, N = 210). In calf separations, the average whistle rate increased as mothers and calves reunited (i.e. in the last 60% of the separation) and then decreased just before they joined (Figure 3.6). In associate separations, whistle rate was significantly higher in the first 80% of the separation (Figure 3.6). Comparison of echolocation rates between habitats showed that in calf separations the echolocation rates were greater in channels than in shallow water areas (Table 3.2). In contrast, no differences were found in whistle rate between habitats (Table 3.2). The opposite pattern was observed in the separations of associates. Echolocation rate was not significantly different between habitats and whistle rates were significantly higher in shallow water areas than in channels (Table 3.2). 87

PAGE 100

3.3.5. Activity When they were temporarily separated, focal mothers and calves spent the majority of their time traveling (females = 58%, calves = 59%). They spent comparable amounts of time feeding and milling during separations (females: feeding = 24%, milling = 19%; calves: feeding = 21%, milling = 20%). When data were divided into habitats, mothers and calves spend similar amount of times in different activities was similar (channel females: traveling = 59%, milling = 22%, feeding = 19%; calves: traveling = 61%, milling = 12%, feeding = 27; shallow water females: traveling = 57%, milling = 18%; calves: traveling = 58%, milling = 23%). Analysis was done using a total of 178 3-min data points (for each mothers and calves). In associate separations, females spent most of their time traveling (84% of the time) and some time milling (12%). Associates also spent most of their time traveling (77%) followed by milling (11%) during separations. Additionally, associates spent more time socializing (10%) than females (2%) and socializing occurred only in shallow water. Feeding was the least frequent activity of associates and females at 2% each. Analysis used 119 3-min data points. Activity budgets of dolphins were not compared statistically between habitats because the number of data points for activities like socializing and feeding was small ( 1). More common activities like milling and travel showed some significant differences. Females and associates spent similar amounts of time traveling and milling in channels and in shallow water areas (channel females: traveling = 84%, milling = 14%; associates: traveling = 73%, milling = 13%; shallow water females: traveling = 72%, milling = 8%; associates: traveling = 70%, milling = 8%). 88

PAGE 101

3.3.6. Whistle Parameters in Different Habitats I identified 556 whistles in calf separations and 1082 whistles in associate separations. Of the total 1638 whistles, I omitted 480 from the analysis because they had poor quality (impossible to distinguish spectral characteristics) or were recorded in habitats not included in the statistical comparisons. In calf separations observed in shallow water, dolphins used whistles having a broader frequency range than those used in channels (frequency range in shallow water: 6.67 4.01 kHz; frequency range in channels: 8.31 4.67 kHz). Whistle duration was relatively similar between habitats (shallow water: 0.71 0.53 ms; channels: 0.79 0.51 ms). The distributions of whistle minimum frequency and whistle maximum frequency were similar (Figure 3.7). In associate separations, the distribution of the maximum frequency of whistles was similar between habitats but the distribution of whistle minimum frequency was different (Figure 3.8). Mean frequency range was 5.79 3.92 kHz in shallow water and 4.72 3.61 kHz in channels. Mean duration was 0.74 0.57 ms in shallow water and 0.62 0.49 ms in channels. 3.4. Discussion Acoustic signals are important for locating partners in social species (Nishida 1968, Byrne 1981, Boinski and Campbell 1995). Partner location is particularly important in areas with limited visibility like the murky waters inhabited by coastal bottlenose dolphins. It has been proposed that bottlenose dolphins can use whistles for long distance 89

PAGE 102

communications (Janik 2000a, Tyack 2000a,b). The present study found that bottlenose dolphins did not always use whistles when temporarily separated from conspecifics. In fact, it was surprising to find that they sometimes did not use any detectable acoustic signal at all. Interestingly, temporary separations with no signals had the shortest mean distance of separation (< 40-50 m) suggesting that groups spread less when no acoustic signals are used. It is possible that visual communication is more important at short distances or that other dolphins use other means such as hydrodynamics for locating each other. Events in which wild bottlenose dolphins used only whistles were very uncommon, representing less than 3% of all separations. Interestingly, separations with no signals were more common than those with only whistles and they represented 12% of all events. 3.4.1. Female-calf Separations Signal usage was variable in mother-calf separations in different habitats. When mothers and their dependent calves separated in channels, most separation events (71%) were accompanied only by echolocation trains. The exclusive production of echolocation trains in this context suggests that echolocation is used, directly or indirectly (eavesdrop), by dolphins to reunite or to remain in contact during a separation. In a behavior pattern similar to bats (Pzalker and Kush 2003), dolphins emitted echolocation trains throughout the separation. Dolphin may be using echolocation to find food and navigate, but they might also be able to keep track of each others position if they eavesdrop on each others echolocation calls. Echolocation calls of bats can have information for conspecifics such as the presence and location of other individuals (Leonard and Fenton 1984, Pfalzer and Kush 2003). Another way in which dolphins could locate a distant partner is if the partner 90

PAGE 103

happens to be in the path of the sonar beam of the echolocating dolphin. Independently of how echolocation is used, it was surprising to find that echolocation is the main signal emitted during calf separations in channels. I considered this surprising because whistles have been documented as the acoustic signal used during mother-calf separations in other dolphin species (Smolker et al. 1993). It is unclear if those studies examined whether echolocation was used or if dolphins might echolocate less frequently in habitats with better visibility. Echolocation was probably used during foraging by mothers and calves. Two other studies have examined echolocation production of bottlenose dolphins in Sarasota Bay (Nowacek 1999, Jones and Sayigh 2002). The focus of those studies is different that the focus of this study and thus real comparisons may not be possible. Jones and Sayigh (2002) examined echolocation production during different behavioral activities but they did not indicate the spatial structure of dolphins at those times. Jones and Sayigh (2002) reported, however, that a group was defined as all animals within a 50-m radius of the boat and that the Sarasota dolphins echolocate significantly less than dolphins of other communities. This study has shown that dolphins normally did not echolocate when they were at distances < 40-50 m from each other and that echolocation was higher at greater distances of separation. Nowacek (1999) reported rates of click production of single animals (adult males and single adult females) that were recently captured and released. Click rates were calculated for the first 5 to 112 minutes after dolphins were released and before data loggers detached. During those times, mean click rates were 0.09 0.13 clicks/min, which is lower than the mean echolocation rates reported in this study during temporary separations. 91

PAGE 104

In this study, echolocation trains were emitted at a higher rate in channels than in shallow water during calf separations. Such an increase could be related to a change of the habitat features used for spatial orientation, to sound propagation characteristics within the habitat, and/or to food searching patterns (Pfalzer and Kush 2003). A high echolocation rate could also be related to the fact that calves spent more time feeding in channels than in shallow water areas. It is also possible that bottlenose dolphins did not alter their echolocation rates in channels but that they decreased their echolocation rate in shallow water. Acoustic signals behave very differently in shallow and deep water. In shallow water, sound waves interact with both the water surface and bottom and this typically results in high levels of reverberation and multipath propagation conditions. Shallow water areas also have irregular bathymetry and different sediment composition than channels (Urick 1975). Thus, dolphins may echolocate less to avoid problems with the degradation of the signal. In shallow water, temporary separations having whistles and echolocation trains were slightly more common (47%) than those with only echolocation trains (32%). Echolocation trains were emitted continuously and whistles were emitted occasionally during separation events. Some bat species also use social calls occasionally and echolocation calls continuously throughout their travel (Leonard and Fenton 1984, Fenton 2003, Pfalzer and Kush 2003). In this context, bats eavesdrop on echolocation calls to avoid bumping into other flying bats (Dawson 1991, Fenton 2003, Masters and Harley 2003) and use social buzzes to maintain contact (Fenton 2003). Herzing (2004) reported that bottlenose dolphins use both echolocation and whistles while they are 92

PAGE 105

foraging. I also identified fish calls mixed with echolocation trains and whistles in some separations. Timing of whistle usage by mothers and dependent calves was similar to that reported for mother-calf pairs in T. aduncus (Smolker et al. 1993) and male pairs in T. truncatus (Watwood 2003). In all three studies, the timing of whistles during a separation was variable but on average, it increased as the mother and calf reunited and it then decreased in the very last portion the separation. Additionally, in all three studies, whistle rates were significantly lower before the separation and after the reunion. In this study, whistle rates were on average ten times higher during separations than in the 5-min period before and after the event. I found that whistle and echolocation rates did not increase with distance but that the mean separation distance in events containing echolocation trains and whistles was significantly higher than events containing one or no signal. In the case of calf separations, the mean separation distance of mothers from calves was 101 46 m in events having whistles and echolocation clicks, 72 41 m in separations having only echolocation trains, and 60 28 m when only whistles were used. These distances are within the typical separation range of mothers and dependent calves, which I found to be on average around 85 m. Separations of females from their calves included whistles with broader frequency ranges in channels. In primates, localization of sounds is improved when overall frequency range of a call is broadened. In fact, experiments have shown that primates can reduce the error of localizing a sound when 400 Hz of frequency modulation is added to a signal (Norcross and Newman 1993). Dolphin whistles did not vary significantly in other characteristics like duration, minimum and maximum frequency when they were used 93

PAGE 106

different habitats. However, whistles were mostly used in shallow water separations than in channels. 3.4.2. Female-associate Separations Most temporary separations of mothers with calves from their other associates involved whistles and echolocation trains. The two acoustic signals were continuously produced throughout separations and this whistling pattern is different from that used in calf separations in which whistles were typically produced in the reunion phase. Whistles could be used among dolphins that become separated instead of as a means to facilitate the reunion with other dolphins. In group-living primates, members of a troop communicate with each other to maintain cohesion (Oda 1996). Yet, a dolphin separated from a group of whistling dolphins can hear them if it is within communication range and thus decides to join again. Smolker et al. (1993) reported a case of a juvenile female T. aduncus whistling to her associates as she fell behind the group and the group waited for her. In separations involving more than two dolphins, the use of whistles may be particularly important when individuals need to find and recognize associates among many dolphins. Whistles can convey individual signature information for recognition (Sayigh et al. 1990, Smolker et al. 1993, Janik et al. 2006), and they may be used for spacing of individuals and coordination of activities (Janik and Slater 1998). In this context the use of only echolocation may not be the most effective way to find particular dolphins, unless dolphins can recognize individuals using echolocation. 94

PAGE 107

Whistle rates in associate separations were higher than whistle rates in calf separations. It is important to note that associate separations were shorter range (< 65 m) than calf separations (> 100 m). Killer whales (Orcinus orca) also produce higher whistle rates during close-range than large-range activities, and short-range whistles appear to play a role in information transfer (Thomsen et al. 2002). The high whistle rate during associate separations suggests that several dolphins were communicating. However, it does not seem that every dolphin whistled and those that whistled did so at the same rate because there was no significant relationship between group size and whistle rate. It also cannot be determined if the focal mother or her calf were among the individuals producing those sounds except in cases when only these two dolphins were present. Whistle rates during associate separations were higher in shallow water, where they spent more time socializing, than in channels. During associate separations, dolphins produced different types of whistles. In channels, they emitted whistles with short duration. Other species like ringtailed lemurs (Lemur catta) also emit short calls when other group members are not nearby (Oda 1996). However, other species such as common marmoset (Callithrix jacchus) use longer calls when conspecifics are separated (Norcross and Newman 1993). In shallow water, dolphins emit whistles with broad frequency. Some bats also use wideband calls, especially when they forage near obstacles, presumably because those calls are more effective at detecting small prey and complex detail of the surrounding environment (Wund 2005). The shallow waters in which dolphins swim can also have obstacles like seagrass, which act as a scattering and absorption system causing dramatic filtering of sound (D. Nowacek et al. 2001, Quintana-Rizzo et al. 2006). 95

PAGE 108

One important thing to point out is that all observed separations were within the estimated maximum communication range of whistles in Sarasota Bay, Florida. For example, in shallow water separations, dolphins emitted whistles with a mean minimum frequency equal to 7 kHz and a mean maximum frequency equal to 14 kHz. The communication range of whistles with those frequency components and a source level equal to 160 dB is estimated to be approximately 300 m. In channels, the same whistle has an estimated communication range over 13 km (Quintana-Rizzo et al. 2006). Those theoretical sound propagation distances are much greater than the mean and maximum separation distances observed for mother-calf pairs and associates in the two habitats. The detection range of echolocation clicks is unknown but it is probably greater than that of whistles since echolocation is used for navigation and dolphins have better hearing sensitivity in the ultrasonic range. In conclusion, this study showed that wild bottlenose dolphins do not always whistle during temporary separations. In fact, in some cases they do not use any active acoustic signals when separated, usually when separations are short range (< 40-50 m). Separations involving only whistles were uncommon. Separations during which only echolocation or echolocation along with whistles were produced were more common. However, there was no consistency in the way the two sounds occurred between periods during which mothers were separated from their calves and when mothers with calves were separated from other associates. The exclusive production of echolocation trains during many separations suggests that the signal may be used, directly or indirectly, to find other dolphins when they are separated, but it could also be related to other specific activities of the individuals during separations, such as prey-finding or orientation. 96

PAGE 109

Unexpectedly, separations of mothers and calves did not involve a higher echolocation rate or whistle rate than separations from other associates. Both involved higher rates of whistling and echolocation on average during the separations than before or after. Furthermore, neither showed consistently higher echolocation or whistle rates at the very end of the separation, suggesting that they may keep track of each others position throughout the separations. When several dolphins formed part of a group, the use of both whistles and echolocation may be very useful. Whistles can convey information on individual identity (Janik et al. 2006) while echolocation could be used, among other things, to locate other dolphins. Dolphins may echolocate to forage and orient, but other dolphins may eavesdrop to track them. During separations, the use of the two signals might be the most effective way to communicate and find each other over long distances. However, the current study could only demonstrate the relative frequency of occurrence of sounds produced by dolphins under particular social contexts. A demonstration of a true communication function of these sounds under those social contexts would require the clear determination of producers and receivers, with notable changes in the behavior of the receiver upon reception. 97

PAGE 110

Table 3.1. Summary of the observations for each of the focal females in 2002 and 2003. Codes (FBXX) correspond to the ID of each focal female. Focal female FB54 FB55 FB65 FB90 F141 F149 F157 Total Number of focal follows 9 9 6 8 7 3 7 49 Number of minutes 1364 889 957 1188 1251 573 808 7030 Separations of calves 66 37 5 7 2 0 1 118 Separations of associates 4 1 12 13 38 12 31 111 Total separations 70 38 17 20 40 12 32 229 98

PAGE 111

Table 3.2. Echolocation rate (echolocation train/min) and whistle rate (whistle/min) recorded in shallow water and channels during separations between females and calves and females and associates. Values are expressed as means SD. Calf separations: N whistles in shallow water = 183, N whistles in channels = 103. Associate separations: N whistles in shallow water = 634, N whistles in channels = 238. indicates statistical significant difference. Shallow water Channels Analysis between habitats Calf separations Echolocation train rate 2.63 2.12 4.28 2.79 U = 205.00, p = 0.01* Whistle rate 0.98 0.98 0.78 0.58 U = 47.50, p = 0.73 Associate separations Echolocation train rate 3.81 2.70 3.18 2.34 U = 373.00, p = 0.42 Whistle rate 3.26 3.62 1.41 1.72 U = 141.00, p = 0.02* 99

PAGE 112

888888888 88888888 8 PVC pipe Weighted hydrophone cable Hydrophone with 2kHzhigh pass filter Laptop and RP2 Outboard engine 5-6 m 3 m 888888888 88888888 8 PVC pipe Weighted hydrophone cable Hydrophone with 2kHzhigh pass filter Laptop and RP2 Outboard engine 5-6 m 3 m Figure 3.1. Diagram of the research vessel and acoustic equipment used to record whistles produced during fission-fusion events. 100

PAGE 113

01020304050607080ChannelShallow Nothing Whistles Echolocation trains Whistles and echolocation trains 01020304050607080ChannelShallowCalf separationsAssociate separationsPercentage of separations 01020304050607080ChannelShallow Nothing Whistles Echolocation trains Whistles and echolocation trains 01020304050607080ChannelShallowCalf separationsAssociate separationsPercentage of separations Figure 3.2. Percentage of separations of calves and associates observed in channels and shallow water areas. Separations were classified into four types of events based on the presence or absence of acoustic signals: 1) no signal, 2) whistles, 3) echolocation trains, and 4) whistles and echolocation trains. 101

PAGE 114

No signalJust whistlesJust echolocationWhistles andecholocationMean distance of separation (m ) Calf separationsAssociate separations 020406080100120140160 020406080100120140160 No signalJust whistlesJust echolocationWhistles andecholocationMean distance of separation (mCalf separationsAssociate separations 020406080100120140160 020406080100120140160 ) Figure 3.3. Mean maximum distance of separation in events having different types of acoustic signals. Bars represented the standard deviation. 102

PAGE 115

103 Figure 3.4. Relationship between group size and whistle rate (top), and between group size and echolocation rate (bottom) during associate separations. Rates are expressed as signal/min.

PAGE 116

104 0123 45 67 01234567 012367 45 01234567WhistleEcholocation trains/minCALF SEPARATIONSASSOCIATE SEPARATIONSBeforeDuringAfterBeforeDuringAfter 012367 s/min 45 01234567 012367 45 01234567WhistleEcholocation trains/minCALF SEPARATIONSASSOCIATE SEPARATIONSBeforeDuringAfterBeforeDuringAfter s/min Figure 3.5. Whistle and echolocation rates (signal/min) in the 5-min period before the separation, during the separation, and the 5-min period after the separation of calves and associates from focal females. Bars represent the standard deviation.

PAGE 117

Figure 3.6. Mean whistle rate (whistle/min) during separation events of calves and of associates of female bottlenose dolphins observed in channels and shallow water areas of Sarasota Bay, Florida. Total number of separations with whistles in shallow water: calves = 23 and associates = 21. Total number of separations with whistles in channels: calves = 9 and associates = 24. 105

PAGE 118

FrequencyMaximum whistle frequency (kHz)FrequencyMinimum whistle frequency (kHz)Shallow water: 7 2 kHz Shallow water: 14 4 kHz Channel: 7 2 kHz Channel: 15 5 kHz 051015202530010203040506070 051015202530010203040506070 051015202530010203040506070 051015202530010203040506070FrequencyMaximum whistle frequency (kHz)FrequencyMinimum whistle frequency (kHz)Shallow water: 7 2 kHz Shallow water: 14 4 kHz Channel: 7 2 kHz Channel: 15 5 kHz 051015202530010203040506070 051015202530010203040506070 051015202530010203040506070 051015202530010203040506070 Figure 3.7. Frequency distribution of whistles recorded during calf separations in shallow water areas and channels. N whistles in shallow water = 183, N whistles in channels = 103. 106

PAGE 119

051015202530020406080100120140160Minimum whistle frequency (kHz)Frequency 051015202530020406080100120140160 051015202530020406080100120140160Maximum whistle frequency (kHz)Frequency 051015202530020406080100120140160Shallow water: 7 2 kHz Shallow water: 12 4 kHz Channel: 7 2 kHz Channel: 12 4 kHz 051015202530020406080100120140160Minimum whistle frequency (kHz)Frequency 051015202530020406080100120140160 051015202530020406080100120140160Maximum whistle frequency (kHz)Frequency 051015202530020406080100120140160Shallow water: 7 2 kHz Shallow water: 12 4 kHz Channel: 7 2 kHz Channel: 12 4 kHz Figure 3.8. Frequency distribution of whistles recorded during associate separations in shallow water areas and channels. N whistles in shallow water = 634, N whistles in channels = 238. 107

PAGE 120

4. ESTIMATED COMMUNICATION RANGE OF SOCIAL SOUNDS USED BY BOTTLENOSE DOLPHINS 4.1. Introduction Understanding how acoustic signals are used by animals to communicate is basic to describing how relationships are formed and maintained. This is particularly important in turbid aquatic environments such as those inhabited by coastal bottlenose dolphins (Tursiops truncatus). Bottlenose dolphins leave and rejoin their conspecific associates frequently and acoustic communication might be used to find and locate distant conspecifics. The maximum distance that an acoustic signal can travel is likely the maximum distance over which associates can remain in contact with one another (Brenowitz 1982, Klump 1996). However, the hearing capabilities of a species must be taken into account along with environmental features affecting sound transmission to understand how far a signal can travel before it drops below the masked auditory threshold or noise floor limiting communication. Therefore, knowledge of the maximum propagation distance of an acoustic signal, the characteristics of ambient noise, and the hearing capabilities of bottlenose dolphins are important for understanding what constitutes a group; if individuals are within communication range they may be part of the same social unit despite being temporarily separated. The approach is valuable to understand if individuals considered as different groups based on their distance of 108

PAGE 121

temporary separation (Wells and Scott 1990, Smolker et al. 1993) could be part of a single group maintaining acoustic contact. This paper estimates the communication range of social sounds produced by bottlenose dolphins that are within the hearing threshold of the species. Bottlenose dolphins use sounds known as whistles to contact conspecifics over long distances (Janik and Slater 2000b). Whistles are narrow-band, frequency-modulated sounds ranging from 4 to 20 kHz (Caldwell et al. 1990). Janik (2000a) revealed that wild, unrestrained dolphins located at distances up to 580 m apart could mimic each others whistles. He proposed this as evidence that dolphins use whistles to communicate over long distances. The active space of a signaler is the distance that a signal can be detected and recognized by a receiver (Brenowitz 1982, Klump 1996, Janik 2000b). In the only study on communication ranges in bottlenose dolphins, Janik (2000b) examined propagation of natural dolphin whistles in a 10 m deep channel by measuring source levels and then estimating propagation and the active space using a model. He found that the active space where dolphins could perceive unmodulated whistles between 3.5 kHz and 10 kHz was between 20 km and 25 km at sea state zero. Sound propagation can be dramatically affected by the habitat through which sound travels (Rogers and Cox 1988, Forrest 1994, Tyack 2000a, D. Nowacek et al. 2001). Large reflecting surfaces or vegetation attenuate some frequencies and can amplify others (Michelsen and Larsen 1983). In the aquatic environment, habitat features such as bottom type, bathymetry, temperature, salinity, and vegetation affect the transmission of sounds (D. Nowacek et al. 2001). The effect of vegetation is not surprising since it acts as a discontinuous barrier to the transmission of sound. 109

PAGE 122

Researchers have also found that in shallow waters, low frequency sounds do not propagate as far as high frequency sounds (Forrest et al. 1993, Forrest 1994). This suggests that dolphin whistles may be affected by environmental variables. Hence, as in other species, the structure of dolphin signals might represent an acoustic compromise balancing an ensemble of ecological and perceptual factors (Wiley and Richards 1978, Brown et al. 1979). Other factors determining whether a sound is detected and identified by an individual are the animals hearing threshold, critical ratio, and the spectrum level of background noise. In bottlenose dolphins, the lowest hearing thresholds are in the frequencies near 50 kHz (Johnson 1967), but whistles have much lower frequencies. Johnson (1967) found that below 50 kHz the threshold increases continuously with decreasing frequency to a maximum of about 137 dB at 75 Hz. Information on the background noise levels is also necessary to estimate the active space of whistles since high-noise levels can significantly mask a sound. The critical ratio is defined as the difference between the level of a just-detectable tone and the spectrum level background noise spanning the same frequency (Johnson 1968, Janik 2000b). Like the hearing threshold, critical ratios are also frequency-dependent and they have been calculated for frequencies within the whistle range (Johnson 1968). I conducted a series of sound transmission experiments to quantify the propagation of sounds in shallow water areas and channels in Sarasota Bay, FL. This habitat is quite different from the Moray Firth studied by Janik (2000b), in that it is very shallow and many areas contain seagrass. In contrast, the Moray Firth is an unusual habitat for coastal dolphins in Florida because the inner waters have depths of up to about 110

PAGE 123

50 m. The outer waters resemble the open sea more with the deepest areas being up to 235 m (Wilson 1995). The Moray Firth is the northern extreme of the species range. The shallow water of Sarasota Bay provides an excellent opportunity to estimate the active space for typical coastal bottlenose dolphins in Florida. I examined the effects of habitat characteristics such as depth, bottom type, vegetation, and bottom sediment on sound propagation. I used regression models to estimate maximum distance of detection taking into account the hearing capabilities of bottlenose dolphins, the background noise levels, and the critical ratios for masking sounds. I also examined the active space of different types of whistles in the same habitats where experiments were conducted. This allowed me to compare estimates of maximum communication range with the distances of separation observed during observations of wild dolphins. 4.2. Materials and Methods This study consisted of three basic components: (1) behavioral observations of mother/calf pairs to identify the habitats that they used and where they temporarily separated, (2) sound propagation experiments at the locations where mothers and calves temporarily separated, and (3) modeling of sound propagation data and information on hearing sensitivity, background noise levels, and critical ratios to estimate the active space of whistles. 111

PAGE 124

4.2.1. Behavioral Observations Behavioral observations of 7 resident female dolphins and their dependent calves (3 years old) were conducted from June to September 2003 from a 7-m-long boat equipped with a 115 hp 4-stroke engine. Acoustic and behavioral data were recorded continuously using focal animal observations (Altmann 1974). Focal animal observations were conducted on the mother of mother-calf pairs. The research vessel was kept at a distance of approximately 20 m from the mother. Separation distances between mother and calf were estimated each time the mother surfaced and the observation zone included an area of approximately 200-m from the mother. Behavioral observations allowed identification of the habitats where sound transmission experiments were to be conducted and acoustic data allowed me to determine the frequency range of whistles used in different habitats to quantify their active space. The set up of the acoustic recording was as follows. At the bow of the observation boat, two 1.5 m sections of PVC pipe were joined in a T joint and secured across the gunwales (Sayigh et al.1993). On each side of the boat approximately 2 m of hydrophone cable were extended from the end of the pipe into the water; each calibrated HTI-96-MIN hydrophone was approximately 1 m below the surface when the boat was not moving. Two hydrophones were used to ensure that I would have a backup recording. To prevent the hydrophone from bouncing at the surface while the boat was moving, the cable of the hydrophone was weighted with a chain attached to the end of the PVC pipe by a carabiner. Each hydrophone was connected to a 2-kHz high-pass filter to reduce engine noise. 112

PAGE 125

Signals from each hydrophone were digitized at 48.8 kHz with a Tucker-Davis Technologies RP2 module, and stored to a computer hard drive. Signals from each channel were sampled at precisely the same time on each channel. Data were recorded with a 24-bit A-to-D converter and were stored as 32-bit floating point values. Data were analyzed with MATLAB 6.5. Behavioral observations allowed me to identify the areas in which to conduct sound transmission experiments. Such areas were chosen based on the fact that 1) dolphins were observed there temporarily separated and thus the active space of whistles used during separations could be examined and 2) the general areas have been identified as areas of high use by dolphins during the long-term studies of the Sarasota Dolphin Research Program (Wells 2003). Specific locations of temporary separations were recorded as latitude and longitude from a Garmin GPS 12 Personal Navigator. 4.2.3. Sound Transmission Experiments Nine sound transmission experiments were conducted in Sarasota Bay, Florida, from September to October 2003 (Figure 4.1; Table 4.1): five in shallow water areas (<3 m) and four in channels (>3 m, up to 5.3 m). Each experiment was conducted in an area where dolphins were observed to engage in temporary separations. In each experiment, I used computer-generated tones that spanned the same frequencies as dolphin whistles (5, 7, 9, 11, 13, 15, 17, and 19 kHz). Tones were played simultaneously for 10 sec and then repeated with a period of silence of 0.03 sec separating them. In each transect, the tone-silence loop was broadcast for 1 minute from an underwater transducer (source) located in a 7-m-long boat. Sounds were played from a laptop computer through a power 113

PAGE 126

amplifier (Hafler P1000) connected to an underwater speaker (Aqua Synthesis). The computer was connected to a RP2-system (Tucker-Davis Technologies) with a HTI-96-MIN hydrophone (sensitivity -169.8 dBV/uPa; 1-24 kHz). The source and the receiving hydrophone (receiver) were located 1 meter below the water surface. This depth was chosen because it was the depth of the hydrophones used to record dolphin whistles during animal observations. Propagated signals and environmental noise were recorded on a NOMAD Jukebox 3 (Creative Labs, Inc.) kept in a stationary kayak. The kayak and the boat were kept in place during the experiments by using two anchors for each vessel. The source was moved at a constant heading from the receiver (transect line) to simulate the movement of one dolphin relative to another. The start point of a given transect was the position of the mother at the time of maximum separation from the calf. The position was recorded as a geographical coordinate during behavioral observations. When possible two transect lines were done for every experiment. In such cases, the direction of each transect was different so that each one followed the general direction of the movement of dolphins before and after they were temporarily separated. For example, if dolphins traveled from point A to B to C with B being the point of maximum distance of separation, one transect was done from point B to A, and another transect was done from point B to C. The transect size varied based on water depth and size of the sampled area. When possible, the source was located at the following experimental distances from the receiver: 1 m, 5 m, 10 m, 20 m, 30 m, 40 m, 50 m, 75 m, 100 m, 150 m, 200 m, 250 m, 300 m, 350 m, 400 m, 450 m, 500 m, 600 m, 700 m, 800 m, 900 m, 1000 m, and 1100 m. The experimental distances exceeded the maximum separation distances because I 114

PAGE 127

was interested in determining sound propagation beyond that range. Experimental distances up to 800 m were measured with a Leica LRF 800 laser range finder. Distances greater than 800 m were measured with a Garmin GPS 12 Personal Navigator (accuracy: 15 m RMS). At each experimental distance, water depth was recorded and in shallow water transects, vegetation type (presence or absence of seagrass) and sediment type (sand or mud) were also noted. A transect was defined as unvegetated if no seagrass was found in more than 75% of the sampling locations. Surface sediment samples were grabbed at each distance and they were classified as sand (granular matter of a few millimeters in size), mud (semi-liquid mixture of water and soil), and sandy-mud. The sandy-mud sample was taken to the Geological Oceanography Program of the College of Marine Science at the University of South Florida for its classification. All experiments were conducted in sea state zero. Channel width at the narrowest and widest points was measured using digitized bathymetry data from ESRIArcGIS TM 9.0. Transmission experiments were conducted around two hours beforeand two hours afterthe same tidal state recorded during the behavioral observations of the maximum separation event. For every experiment, the spectrum level of tones at each distance was calculated using a 48000-point FFT with a Hanning window, which resulted in a frequency resolution of 1 Hz. In the calculation, I corrected for the analysis bandwidth, the hydrophone sensitivity (-169.8 dB re: 1V/Pa), the calibration of the NOMAD Jukebox recorder (18.9 dB re: 1V), and the Hanning window (6 dB). Noise level was measured using the recorded signal where no sound was being broadcast up to 7.5 ms before each tonal frequency. The signal-to-noise ratio (SNR) of 115

PAGE 128

each frequency was calculated by subtracting the noise level from the corresponding received level. 4.2.4. Modeling of Active Space of Whistles The transmission loss data were used to calculate regressions that model sound propagation and the ability of a dolphin to detect that signal in each experimental transect. Regression equations were calculated for each channel transect and for one of the two transects of each shallow water experiment for a total of nine models of sound propagation. A logarithmic curve was fitted to the received levels of sound propagation and the resulting equation was used to estimate the active space of hypothetical whistles taking into account the spectrum level background noise, dolphin critical ratios (Johnson, 1968), and dolphin hearing thresholds (Ljungblad et al. 1982). This is because a sound can be heard by a dolphin only if its received level is above the spectrum level background noise and the animals critical ratio. Thus, sound detection is limited by the combined effects of the dolphin hearing threshold and the spectrum level background noise plus critical ratio. Since different frequencies propagate different distances, the propagating distance of the first whistle frequency reaching the threshold was defined as the frequency limiting the active space of whistles. For example, if the attenuation of 5 kHz, 9 kHz, and 11 kHz frequencies of a 5-11 kHz whistle are examined (Figure 4.2), it is found that the 9 kHz signal reaches the noise floor plus critical ratio before the 5 and 11 kHz frequencies. In this example, the hearing range is noise limited. The results of the regression models were used to calculate the active space of two hypothetical whistles. The first whistle had frequencies from 7-13 kHz, which 116

PAGE 129

corresponded to the mean minimum and mean maximum frequencies of whistles recorded during separations of dolphins in shallow water areas. I referred to this whistle as a low-frequency whistle. The second whistle had frequencies from 13-19 kHz and I referred to it as a high-frequency whistle. The minimum frequency of the high-frequency whistle corresponded to the closest frequency of the tone played to the mean minimum frequency (12 kHz) of whistles recorded during separations in channels. Similarly the maximum frequency of the high-frequency whistle corresponded to the closest frequency of the tones played to the mean maximum frequency (20 kHz) of whistles recorded during separations in channels. The hearing thresholds of the frequencies played were calculated from the hearing thresholds reported by Ljungblad et al. (1982). I fitted a regression line to the hearing thresholds of frequencies 5 kHz, 10 kHz, 15 kHz, and 20 kHz to calculate the hearing threshold of the tone frequencies not included in their study: 7 kHz, 9 kHz, 11 kHz, 17 kHz, and 19 kHz. Since Ljungblad et al. (1982) calculated lower-frequency hearing thresholds using two projector systems, I used the calculated regression values from their project LC-10 projector data because they had the highest coefficient (R 2 = 0.95). Hearing thresholds calculated were 81.5 dB SPL for 5 kHz, 79.8 dB SPL for 7 kHz, 78.4 dB SPL for 9 kHz, 77.5 dB SPL for 11 kHz, 76.7 dB SPL for 13 kHz, 75.9 dB SPL for 15 kHz, 75.3 dB SPL for 17 kHz, and 74.8 dB SPL for 19 kHz. The critical ratio of each frequency was added to the average background noise of that frequency. The critical ratios were calculated by fitting a regression line to the critical ratios of frequencies that Johnson (1968) did not examine but that were included in our experiments. These critical ratios corresponded to 23.6 dB for 5 kHz, 25.2 dB for 7 117

PAGE 130

kHz, 26.8 dB for 9 kHz, 28.4 dB for 11 kHz, 29.9 dB for 13 kHz, 31.5 dB for 15 kHz, 33.2 dB for 17 kHz, and 34.8 dB for 19 kHz. The development of a model of whistle propagation required data on the source levels of whistles produced under natural circumstances. No information exists on the source levels of the whistles produced by the Sarasota dolphins. Thus, a table of possible source levels of whistles (155 dB, 160 dB, and 165 dB) was constructed using as a reference the maximum source level reported for other wild bottlenose dolphins (169 dB re 1 Pa; Janik 2000b). The means of the slopes of the regression models were compared using a two way analysis of variance (two-tailed, alpha = 0.05) to test for significant differences between habitats, frequencies, and interactions between frequencies habitats. The slope of the regression indicates how sound levels fall off with distance. Statistical analysis was performed with SPPS v. 14.0 (SPSS, Chicago, Illinois, U.S.A). 4.3. Results 4.3.1. Behavioral Observations A total of 224 separations of females and their dependent calves was observed. Of these, 161 occurred in shallow water and 63 occurred in channels. Mean separation distance in shallow water was 115 48 m and in channels was 99 48 m. Eight separation events were used to identify the areas where sound transmission experiments were conducted. They corresponded to the most recent events and they occurred on 118

PAGE 131

different dates. Five separations of three different mother/calf pairs were recorded in shallow water and their separation distances were 90 m, 95 m, 100 m, 120 m, and 200+ m. In channels, three separation events of three different mother/calf pairs were recorded. One separation event occurred in the intersection between two channels and this point was used as the start point of each experimental transect in the two channels. Separation distances in the channels were 50 m, 200 m (n = 2), and 200+ m. During the separations, 204 whistles were recorded. In the shallow water areas, 199 whistles were recorded in three of the five separations and they corresponded to three different mother/calf pairs. The three separations lasted a total of 38.0 min. In channels, 5 whistles were recorded in two of the four separations and they corresponded to two different mother/calf pairs. The two separations lasted 17.1 min. In shallow water areas, whistles had a minimum frequency with a mean equal to 7.5 2.5 kHz and a maximum frequency with a mean equal to 13 3.2 kHz. In channels, whistles had a minimum frequency with a mean equal to 12 3.6 kHz and a maximum frequency with a mean equal to 20 7.4 kHz. 4.3.2. Sound Transmission Experiments Shallow water transects had mean depths varying from 1.3 m (SAMS2) to 2.6 m (NWPSB2; Table 4.1). The overall mean depth of three transects was 1.9 0.5 m. Channel transects had mean depths varying from 3.1 m (North Anna Maria Sound) to 4.1 m (San Remo Channel) and their overall mean depth was 3.47 0.7 m. Most frequencies either followed the spherical spreading attenuation model or had transmission loss values that were intermediate between the predicted values of the 119

PAGE 132

spherical and cylindrical spreading attenuation models (Figure 4.3 and Figure 4.4). Low frequencies traveled much farther than high frequencies and sound propagation varied within and between channels and shallow water areas (Figure 4.3 and Figure 4.4). In shallow water, the attenuation was highly variable at distances up to 50 m when the average water depth was equal to 1.3 m (n = 19, SD = 0.1). In contrast, when the mean water depth was greater than 2 m, attenuation was highly variable at shorter distances (5 and 20 m; Figure 4.3). To compare sound propagation among shallow-water transects, I examined transmission loss at the 100-m point, which is the maximum distance of the shortest transect. At the 100-m point, three transects (SAMS2, SPSB1, and SPSB2) had a depth of 1.1 m, and mean transmission loss over all frequencies was -27 dB. In the other six transects (PSB1, PSB2, SAMS1, SKF, NWPSB1, and NWPSB2) water depth was greater (2.0 m), and mean transmission loss over all frequencies was greater than or equal to -30 dB over 100 m. The effect of vegetation and sediment type on sound transmission loss was examined in shallow water areas. At the 100-m point, the mean transmission loss was greater in transects with seagrass and lower in transects with mud or sand bottom sediments (Fig. 4.3). In seagrass transects, mean transmission loss was approximately -36 dB for low-frequency whistles, and -47 dB for high-frequency whistles. In non-seagrass transects, mean transmission loss was similar between the two type of hypothetical whistles (low-frequency whistles -29 7 dB, high-frequency whistle -30 5 dB). Frequencies greater than 17 kHz were less attenuated in sand areas than in the sand-mud and mud areas. Mean transmission loss was -42 dB in transects with seagrass (PSB1 and 120

PAGE 133

PSB2), -29 dB in transects with sandy-mud sediment (NWPSB 1 and NWPSB 2), -28 dB in transects with mud sediment (AMS1, AMS2, SPSB1, and SPSB2), and -26 dB in transects with sand sediment (SKF). Transmission loss in channels at the 100-m point was also examined. At the 100-m point, three of the four channels had relatively similar water depths (Anna Maria Sound = 3.4 m, Cortez Channel = 3.4 m, and Main Channel = 3.2 m), but their mean transmission losses were very different (Fig. 4.3). The mean transmission loss over all frequencies was -14.3 dB in the Cortez Channel, -27 dB in the Main Channel, and -40 dB in Anna Maria Sound. Mean transmission loss of lowand high-frequency whistles was, respectively, -15.3 dB and -13.5 dB in the Cortez Channel, -27 dB and -30 dB in the Main Channel, and -41 dB and -38.1 dB in Anna Maria Sound. The fourth channel (San Remo Channel) was deeper at the 100-m point (4.7 m) and its mean transmission loss was equal to -35 dB. Mean transmission loss of lowand high-frequency whistles was -36 dB and -36 dB, respectively. Mean transmission loss at the 100-m point was more variable in channels than in shallow water areas. Mean transmission loss varied from -14 dB to -40 dB in channels and from -26 dB to43 dB in shallow water areas. However, at the same point, the range between the minimum and maximum transmission losses was relatively similar between the two habitats with shallow water areas being wider by 4 dB (-12 dB to -57 dB) than channels (-4 dB to -45 dB). Noise levels were variable among transects both in shallow water and channels (Tables 4.2). Noise levels plus critical ratios ranged from 94 dB to 107 dB re 1 Pa 2 /Hz in shallow water areas and from 101 dB re 1 Pa 2 /Hz to 110 dB re 1 Pa 2 /Hz in 121

PAGE 134

channels. For most frequencies, the noise level plus critical ratio was lower in shallow water transects than in channels. In some shallow water locations (SKF, PSB) and all channels, the noise level plus critical ratio increased with increasing frequency, especially at frequencies greater than 11 kHz. However, in other shallow water locations (SAMS, SPSB) the noise level plus critical ratio was approximately the same among frequencies. 4.3.3. Modeling of active space The theoretical detection range was noise-limited, as opposed to hearing-sensitivity-limited, both in shallow water areas and in channels. This was evident in the fact that all noise levels plus critical ratio measurements were greater than the hearing thresholds obtained by Ljungblad et al. (1982). There were no significant differences in the mean regression slopes between habitats (F = 2.41, p = 0.13, df = 1) or between frequencies (F = 0.91, p = 0.50, df = 7). There was a significant interaction between habitat and frequency (F = 2.42, p = 0.03, df = 7), which was mainly due to differences between 5 kHz and 9 kHz in channel and shallow water. Active space was shorter in shallow water areas than in channels. In shallow water, active space of whistles was greater in unvegetated habitats than in seagrass habitats (Figure 4.5). For example, the active space of a low-frequency whistle with a source level equal to 155 dB was estimated to be approximately 662 m in an unvegetated habitat (mud bottom: SAMS) and 186 m in a habitat with seagrass (PBS). In channels, the same whistle had an estimated active space approximately between 230 m to 1 km depending on the channel (SRC = 230 m, AMS = 345 m, MC = 750 m, CC = 6070 m). Active space was also different between whistles with different frequency components. 122

PAGE 135

While a low-frequency whistle with a source level of 155 dB can travel up to 6 km, a high-frequency whistle with the same source level can travel approximately up to 4 km depending on the channel. Active space almost doubled with a 5 dB increase in whistle source level (Fig. 4.5). In shallow water, a low-frequency whistle with a source level equal to 160 dB had an estimated active space of approximately 1260 m in an unvegetated habitat (mud bottom: SAMS) and of 301 m in a seagrass habitat (PBS). The estimated active space of a low-frequency whistle with a source level equal to 165 dB was over 2 km in the same unvegetated habitat and close to 500 m in the same seagrass habitat. In channels, a similar pattern was observed. Low-frequencies with a 160 dB source level can travel between 400 m (SRC) and 13 km (CC). If the source level of the same whistle increased by 5 dB, the estimated active space increased to 1 km (SRC) and 28.5 km (CC). It is important to note that the estimates assume that a habitat is homogenous in its propagation characteristics. 4.4. Discussion Separation distances of females and their dependent calves were shorter than the estimated active space of whistles. Since whistles are thought to be used by dolphins to maintain group cohesion (Janik and Slater 1998, Norris et al. 1994, Smolker et al. 1993), the results suggest that dolphins can communicate over the distances that temporary separations occurred. The results also suggest that separation distances are not necessarily 123

PAGE 136

determined by the maximum communication range. Other factors such as predation pressure or food distribution may be important. A calf may not wander far from its mother if the cost of predation risk is high. Furthermore, factors like ambient noise can affect communication range by dramatically reducing the active space (Urick 1967, Forrest 1994, Janik 2000b, Slabbekoorn 2004). In this respect, the results of the present study showed the best-case scenario of sound propagation and estimates of active space, because experiments were conducted when no boats were present within a radius of approximately 1 km of the recordings. In fact, the theoretical detection range was noise-limited, as opposed to hearing-sensitivity-limited, both in shallow water areas and in channels. However, background noise can vary widely depending on the number of power boats present, fish choruses, snapping shrimp, and wave action. In Sarasota Bay, dolphins are frequently exposed to boat noise as powerboats pass within 100 m of them on an average of every six minutes during daylight hours (S. Nowacek et al. 2001). Estimates of active space were based on the propagating distance of the first frequency reaching the noise threshold since different frequencies propagate different distances. Yet, dolphins may be able to discriminate a whistle even if one of its frequency components is lost with distance. However, it is currently unknown how dolphins discriminate and identify sounds. It is one thing to detect a sound and quite another to comprehend its significance. Thus, our estimates may be conservative if the distance over which the meaning of a whistle is transmitted is greater than the distance over which the first whistle frequency component is lost. Estimated active space of whistles was highly variable according to habitat characteristics. For example, in shallow seagrasss areas, the active space of a 7-13 kHz 124

PAGE 137

whistle with a source level equal to 160 dB was estimated to be 301 m. Yet, the same whistle had an estimated active space over 13 km in channels. The regressions in tables 3 and 4 can be interpreted to understand how sound propagates on average in each transect. The slope of the regression indicates how sound levels fall off with distance. Although I did not find significant differences in the means of the distributions of regression slopes between habitats and between frequencies, it is important to note that individual transects showed a lot of variation, and the transmission loss over each transect is what is biologically relevant, not the mean for each habitat type. A slope of 20 indicates that sound follows a spherical spreading model. If the slope is 10, then it follows a cylindrical spreading model. The smallest slope was -13.8 dB, which was in a channel. The steepest slope (i.e. greatest propagation loss) was -28.0 dB, which was in shallow seagrass. The assumption of the propagation models is that a habitat is homogenous in its propagation characteristics. However, propagation characteristics are likely to be variable and significant changes in habitat features like water depth, substrate, seagrass cover, channel shape (horizontal and vertical), and channel width can alter active space greatly. Figures 4.3 and 4.4 show how propagation is affected by changes in water depth and how some frequencies are more attenuated than others with depth. In the Main Channel (MC), received levels of an 11 kHz signal fluctuated by more than 15 dB with water depth changes of up to 1 m. A source level change of even 5 dB can decrease or increase the active space of whistles significantly as dolphins navigate throughout the heterogeneous environment. Although the active space of some whistles could be more than 13 km in channels, in reality most channels sampled are shorter than 3-4 km before their course changes direction. Yet, an active space of even a few kilometers is still be a significant 125

PAGE 138

range of communication between dolphins in Sarasota Bay where channels are narrow and shallow areas extend hundred of meters. Variation in estimated active space was observed in shallow water areas. A low-frequency whistle with a source level of 155 dB attenuated up to seven times more in seagrass areas than in areas with other bottom type with respect to distance (Figure 4.5). In seagrass, the active space of the same whistle was estimated to be approximately 186 m. In contrast, the active space of the whistle was much greater, approximately 1319 m, in a sandy mud habitat of comparable depth. The scale of the variation in active space is relevant to dolphins because in Sarasota Bay dolphins use seagrass areas extensively to feed (Waples 1995; Barros and Wells 1998). If during feeding events dolphins use whistles to maintain contact, their ability to communicate over long distances (several hundreds of meters) is greatly reduced in seagrass areas. Seagrasses act as a complicated three dimensional diffraction system which causes dramatic filtering of the sound (Wiley and Richards 1978). D. Nowacek et al. (2001) reported that the transmission loss of frequencies between 4 kHz and 8 kHz was up to 6 dB greater over 50 m in shallow seagrass areas than in shallow areas with mud or sand bottoms. Sound propagated farther in habitats with sparse grass than habitats with dense grass (D. Nowacek et al. 2001). Other factors that influence sound propagation in shallow waters include surface conditions, bottom contour variability, water column sound speed properties, bathymetry, vegetation, and bottom type (Urick 1975, Forrest et al. 1993, Forrest 1994, Jensen 2001). Among non-seagrass areas, we found that whistles were more attenuated in areas with mud bottoms followed by sand and sandy mud bottoms. Mud, clay, and silt cause 126

PAGE 139

energy to dissipate more than sand and gravel (Urick 1975, Jensen 2001). Although the number of experiments conducted in areas of each sediment type was low, similar results in the sound transmission of sand and mud areas were found by Marsh and Schulkin (1962). Our estimates of whistle active space in those habitats were greater than 500 m. The active space of whistles was also variable among channels, but the variation was not always directly related to channel depth as expected. In two cases, active space was related to channel width. In the widest channel, the estimated active space of high-frequency whistles was over 4 km. The mean maximum separation distance of females and their dependent calves in channels (99 48 m) was much shorter than the estimated active space. A large whistle active space could result in high masking noise for whistles if other dolphins use the same whistle frequency range (Janik 2000b) and are whistling at the same time. A large communication range may also result in animals being able to eavesdrop on acoustic interactions (Janik 2000b). This could be a benefit if dolphins use whistles to look for specific associates since individual dolphins produce distinctive signature whistles (Caldwell et al. 1990). For example, a large active space could help able dolphins to find each other. In Tursiops sp. male dolphins are known to form coalitions with particular males during the mating season. Male coalitions are formed to control and sometimes steal receptive females from other males (Connor et al. 1992). For receptive females, a large active space of whistles could be costly if they are avoiding harassing males. Janik (2000b) found that the active space of whistles in a channel can decrease by several kilometers when whistle frequency is higher than 10 kHz. In our study, the decrease in active space by several kilometers occurred at 13 or 15 kHz in all four 127

PAGE 140

channels. Differences between studies could be due to environmental differences. The channels of our study were much shallower (3.1 to 4.1 m) and narrower (139 to 390 m) than the channel studied by Janik (2000b; depth = 10 m, width = 500 m). Such environmental variability makes it difficult to provide general conclusions about the behavior of specific frequencies. Sound propagation may change as water depth changes with tidal events, temperature gradients, freshwater inputs, and obstacles in the sound path. Changes in whistle source level or frequency could help the transmission of whistles over long distances when associates are temporarily separated. The active space of whistles almost doubles when there is a 5 dB increase in source level (almost a doubling in energy; Figure 4.5). This suggests that there is an advantage if dolphins produce louder whistles in habitats where propagation is poor than in other habitats. There could also be an advantage in changes in frequency. For example, the active space of low-frequency whistles was larger in 75% of the shallow water areas than the active space of high-frequency whistles in the same habitat. Studies examining the characteristics of whistles used in different habitats will help in understanding how dolphins communicate over long distances. It is important to note that our estimates of active space assume that both the whistling and receiving dolphins are 1 m below the water surface. However, dolphins move vertically within the water column as they surface to breath, search for food, or socialize with other dolphins. The active space changes with the position of the whistling and receiving dolphins because the transmission loss of sound varies within the water column. The influence of the sender location on signals has been suggested for birds 128

PAGE 141

(Lohr et al. 2003) and other aquatic animals (Forrest et al. 1993). Active space also varies with hearing thresholds, which can vary greatly among dolphins of different age groups (Houser and Finneran 2005). Thus, each dolphin might have a different communication range. Another aspect to take into account is that studies have found that whistles of some delphinids are somewhat directional at higher frequencies, especially in respect to harmonics (Lammers and Au 2003). Thus, even in a homogeneous environment, the active space may not be radially symmetric around the dolphin. The results of this study suggest that whistle active space is greater than the distances commonly used to identify dolphins as members of a group: 10-m chain rule (Smolker et al. 1993) and a radius of 100 m (Wells and Scott 1990). Although such definitions make data collection manageable and replicable in the field, they may not always fully describe groups if dolphins are communicating with conspecifics over much larger distances. Since whistles are thought to be used by dolphins to maintain group cohesion and to communicate over long distances (Janik and Slater 1998, Norris et al. 1994, Smolker et al. 1993), understanding the communication range to define dolphin groups is extremely important. 129

PAGE 142

130 Table 4.1. Overview of transects per habita t where sound transmission experiments were conducted in Sarasota Bay, Florida. Note: H = habitat, SW = shallow water, C = channel. H Site No. of Transects Start Latitude/ Longitude Bottom type Mean depth SD SW SE Palma Sola Bay (PSB) 2 N 27 o 2814 W 82 o 3924 Seagrass PSB1: 1.60.1 m (n = 9) PSB2: 1.60.3 m (n = 18) NW Sister Key (SKF) 1 N 27 o 2651 W 82 o 4034 Sand SKF: 2.00.2 m (n = 17) NW Palma Sola Bay (NWPSB) 2 N 27 o 2914 W 82 o 3926 Sandy mud NWPSB1: 2.10.5 m (n = 12) NWPSB2: 2.60.2m (n = 19) SW Palma Sola Bay (SPBS) 2 N 27 o 2859 W 82 o 4026 SPSB1: mud seagrass SPSB2: mud SPSB1: 1.70.2 m (n = 10) SPSB2: 1.40.3 m (n = 11) SE Anna Maria Sound (SAMS) 2 N 27 o 2913 W 82 o 4145 Mud AMS1: 2.00.3 m (n = 18) AMS2: 1.30.1 m (n = 9) C Main Channel (MC) 1 N 27 o 2957 W 82 o 4031 N/A 3.51.0 m (n = 19) San Remo Channel (SRC) 1 N 27 o 2845 W 82 o 4031 N/A 4.10.7 m (n = 14) Anna Maria Sound (AMS) 1 N 27 o 3009 W 82 o 4134 N/A 3.10.4 m (n =22) Cortez Channel (CC) 1 N 27 o 2723 W 82 o 4101 N/A 3.40.3 m (n = 16)

PAGE 143

131 Table 4.2. Spectrum level background noise plus critical ratios in dB re 1 Pa 2 /Hz of each frequency in channels (MC, SRC, AMS, and CC) and shallow-water transects (PSB1, PSB2, SKF, NWPSB1, NWPSB2, SPSB1, SPSB2, SAMS1, and SAMS2) in Sarasota Bay, Florida. Channels Shallow-watertransects Frequency MC SRC AMS CC PBS1 PBS2 SKF NWPSB1 NWPSB2 SPSB1 SPSB2 SAMS1 SAMS2 5 kHz 102.6 106.5 100.3 103.8 99.2 100.3 99.5 101.6 101.6 95.6 94.5 101.7 96.4 7 kHz 101.8 103.7 95.0 105.3 95.8 96.9 95.1 99.0 98.6 95.1 96.0 101.9 99.0 9 kHz 104.9 105.7 103.5 104.0 98.1 98.5 98.9 100.8 100.3 96.9 95.5 101.6 98.3 11 kHz 103.4 105.5 96.4 106.1 98.7 99.4 99.3 99.8 100.4 96.7 98.6 106.9 97.9 13 kHz 105.1 107.2 96.9 107.4 100.0 100.3 100.1 100.9 102.0 96.4 97.1 101.9 96.3 15 kHz 106.3 108.4 101.1 107.7 100.0 105.4 103.4 105.4 105.3 100.4 101.5 106.5 97.2 17 kHz 108.2 110.0 100.6 108.6 103.1 103.7 103.8 105.3 104.5 99.4 101.0 106.5 99.7 19 kHz 108.9 109.5 102.8 107.6 102.6 102.8 106.4 102.9 106.6 101.3 102.5 105.6 102.4

PAGE 144

Table 4.3. Regression equations representing the mathematical model of sound propagation for different frequencies in shallow water habitats. Note: To calculate the propagation distance, source levels are added to the intercept of the equation. Frequencies from 7-13 kHz are referred to as a low-frequency whistle and from 13-19 kHz are referred to as a high-frequency whistle. R is the proportion of variability explained by the propagation model. Experimental transect/habitat Frequency Propagation distance (x = distance in meters, y = whistle source level) R 2 5 kHz y = -19.9*Log(x) + 3.5 0.9 7 kHz y = -22.6*Log(x) + 4.7 0.9 9 kHz y = -23.9*Log(x) 2.2 0.9 PSB2 11 kHz y = -22.9*Log(x) 4.4 0.9 Seagrass 13 kHz y = -22.1*Log(x) + 1.9 1.0 15 kHz y = -28.0*Log(x) + 1.3 1.0 17 kHz y = -26.0*Log(x) + 0.9 0.9 19 kHz y = -26.6*Log(x) + 2.0 0.9 5 kHz y = -25.8*Log(x) + 13.9 0.8 7 kHz y = -23.5*Log(x) + 11.1 0.8 9 kHz y = -16.8*Log(x) + 3.2 0.9 SKF 11 kHz y = -15.9*Log(x) + 0.6 0.8 Sand 13 kHz y = -16.4*Log(x) + 2.9 0.9 15 kHz y = -15.2*Log(x) + 2.1 0.8 17 kHz y = -14.5*Log(x) + 2.5 0.8 19 kHz y = -26.6*Log(x) 1.0 0.9 5 kHz y = -25.8*Log(x) + 8.4 0.8 7 kHz y = -23.5*Log(x) + 0.9 0.9 9 kHz y = -16.8*Log(x) + 0.7 0.7 NWPSB2 11 kHz y = -15.9*Log(x) + 0.4 0.7 Sandy mud 13 kHz y = -16.4*Log(x) 0.5 0.8 15 kHz y = -15.2*Log(x) + 1.9 0.8 17 kHz y = -14.5*Log(x) + 2.8 0.7 19 kHz y = -26.6*Log(x) 9.3 0.7 132

PAGE 145

Table 4.3 (Continued). Regression equations representing the mathematical model of sound propagation for different frequencies in shallow water habitats. Note: To calculate the propagation distance, source levels are added to the intercept of the equation. Frequencies from 7-13 kHz are referred to as a low-frequency whistle and from 13-19 kHz are referred to as a high-frequency whistle. R is the proportion of variability explained by the propagation model. Experimental transect/habitat Frequency Propagation distance (x = distance in meters, y = whistle source level) R 2 5 kHz y = -22.9*Log(x) + 10.6 0.6 7 kHz y = -20.2*Log(x) + 6.8 0.6 9 kHz y = -22.9*Log(x) + 6.0 0.8 SPSB2 11 kHz y = -14.9*Log(x) + 9.8 0.6 Mud 13 kHz y = -19.2*Log(x) + 5.4 0.9 15 kHz y = -16.8*Log(x) + 7.8 0.7 17 kHz y = -15.5*Log(x) + 2.6 0.8 19 kHz y = -15.9*Log(x) + 5.1 0.8 5 kHz y = -16.1*Log(x) + 5.9 0.8 7 kHz y = -17.2*Log(x) + 5.8 0.9 9 kHz y = -20.9*Log(x) + 8.0 0.9 AMS1 11 kHz y = -18.0*Log(x) + 2.5 0.9 Mud 13 kHz y = -18.5*Log(x) + 3.6 0.9 15 kHz y = -16.0*Log(x) + 5.1 0.8 17 kHz y = -16.2*Log(x) + 4.7 0.9 19 kHz y = -26.6*Log(x) + 5.4 0.9 133

PAGE 146

Table 4.4. Regression equations representing the mathematical model of sound propagation for different frequencies in channels. To calculate the propagation distance, source levels are added to the intercept of the equation. Frequencies from 7-13 kHz are referred to as a low-frequency whistle and from 13-19 kHz are referred to as a high-frequency whistle. R is the proportion of variability explained by the propagation model. Experimental transect/habitat Frequency Logarithmic equation (x = distance in meters, y = whistle source level) R 2 5 kHz y = -15.4*Log(x) + 4.8 0.9 7 kHz y = -14.9*Log(x) + 1.5 0.8 9 kHz y = -18.4*Log(x) + 2.7 0.8 Main 11 kHz y = -15.5*Log(x) + 7.6 0.7 Channel 13 kHz y = -18.3*Log(x) + 2.8 0.8 15 kHz y = -12.8*Log(x) + 2.4 0.6 17 kHz y = -15.9*Log(x) + 3.4 0.9 19 kHz y = -13.8*Log(x) + 3.5 0.8 5 kHz y = -17.0*Log(x) 3.9 0.8 7 kHz y = -18.0*Log(x) + 0.3 0.9 9 kHz y = -20.9*Log(x) 0.0 0.9 San Remo 11 kHz y = -21.0*Log(x) + 7.3 0.8 Channel 13 kHz y = -19.9*Log(x) + 0.9 0.9 15 kHz y = -18.1*Log(x) 3.2 0.9 17 kHz y = -14.5*Log(x) + 2.5 0.9 19 kHz y = -16.7*Log(x) 3.2 0.9 5 kHz y = -13.9*Log(x) 6.4 0.8 7 kHz y = -16.5*Log(x) 3.8 0.9 9 kHz y = -20.9*Log(x) + 1.5 0.9 Anna Maria 11 kHz y = -20.7*Log(x) + 2.3 0.9 Sound 13 kHz y = -21.8*Log(x) + 7.1 0.9 15 kHz y = -24.5*Log(x) + 7.1 0.9 17 kHz y = -21.8*Log(x) + 3.4 0.9 19 kHz y = -22.4*Log(x) + 8.7 0.9 5 kHz y = -13.9*Log(x) + 3.4 0.8 7 kHz y = -16.5*Log(x) + 1.2 0.9 9 kHz y = -20.9*Log(x) + 5.5 0.8 Cortez 11 kHz y = -20.7*Log(x) + 8.1 0.7 Channel 13 kHz y = -21.8*Log(x) + 4.2 0.7 15 kHz y = -24.5*Log(x) 0.2 0.7 17 kHz y = -21.8*Log(x) + 3.6 0.8 19 kHz y = -22.4*Log(x) + 7.3 0.9 134

PAGE 147

Figure 4.1. Study area showing the experimental transects in Sarasota Bay, Florida. White circles represent the location of the hydrophone and black circles represent the last location in which the speaker was placed in each transect. Shallow areas are represented by solid lines: Palma Sola Bay (PSB), North West Palma Sola Bay (NWPSB), South West of Palma Sola Bay (SPSB), Sister Key flats (SKF), and South East Anna Maria Sound (SAMS). Channels are represented by dashed lines: Main Channel (MC), San Remo Channel (SRC), Anna Maria Sound (AMS), and Cortez Channel (CC). 135

PAGE 148

020406080100120140160180050100150200250300350400450500Distance in metersSource level in dB d B Noise floor p lus critical Hearin g Figure 4.2. Attenuation of three frequency components of a theoretical 5-11 kHz whistle showing where they intersect the hearing threshold and noise floor plus critical ratio of each frequency. Hearing thresholds were taken from Ljungblad et al. (1982). See text for explanation of how critical ratios were determined. 136

PAGE 149

PSB1 -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700Transmission Loss (dB)Distance in metersPSB2 -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700Distance in meters 5000 Hz 7000 Hz 9000 Hz 11000 Hz 13000 Hz 15000 Hz 17000 Hz 19000 Hz Cylindrical spreading Spherical spreadingFrequenciesTheoretical attenuation Depth in metersDepth in metersSKF -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700TransmissionLoss (dB)Distance in meters 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700PSB1 -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700Transmission Loss (dB)Distance in metersPSB2 -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700Distance in meters 5000 Hz 7000 Hz 9000 Hz 11000 Hz 13000 Hz 15000 Hz 17000 Hz 19000 Hz Cylindrical spreading Spherical spreadingFrequenciesTheoretical attenuation Depth in metersDepth in metersSKF -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700TransmissionLoss (dB)Distance in meters 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700Depth in metersSKF -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700TransmissionLoss (dB)Distance in meters 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 Figure 4.3. Transmission loss data with distance for eight tones used during sound transmission experiments conducted in four shallow areas (PSB, SKF, NWPSB, and SPSB). Except for SKF, two transect lines were done in each experiment and each transect line is represented with a number next to the code of the corresponding area. Theoretical attenuation based on cylindrical and spherical spreading is also shown. A profile of the depth contour of each transect is also included. 137

PAGE 150

NWPSB1 -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700NWPSB2 -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700SPSB1 -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700 -80-70-60-50-40-30-20-10010050100150200250300350400450500550600650700SPSB2Transmission Loss (dB)Transmission Loss (dB) 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700Depth in metersDistance in metersDistance in meters 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700Depth in metersDistance in metersDistance in metersNWPSB1 -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700NWPSB2 -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700SPSB1 -80-70-60-50-40-30-20-100050100150200250300350400450500550600650700 -80-70-60-50-40-30-20-10010050100150200250300350400450500550600650700SPSB2Transmission Loss (dB)Transmission Loss (dB) 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700Depth in metersDistance in metersDistance in meters 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700 0.00.51.01.52.02.53.0050100150200250300350400450500550600650700Depth in metersDistance in metersDistance in meters Figure 4.3 (Continued). Transmission loss data with distance for eight tones used during sound transmission experiments conducted in four shallow areas (PSB, SKF, NWPSB, and SPSB). Except for SKF, two transect lines were done in each experiment and each transect line is represented with a number next to the code of the corresponding area. Theoretical attenuation based on cylindrical and spherical spreading is also shown. 138

PAGE 151

Transmission Loss (dB)Distance in metersDistance in meters -80-70-60-50-40-30-20-1001001002003004005006007008009001000Depth in meters 0.01.02.03.04.05.06.001002003004005006007008009001000 0.01.02.03.04.05.06.001002003004005006007008009001000 -80-70-60-50-40-30-20-10001002003004005006007008009001000SAMS1SAMS2Transmission Loss (dB)Distance in metersDistance in meters -80-70-60-50-40-30-20-1001001002003004005006007008009001000Depth in meters 0.01.02.03.04.05.06.001002003004005006007008009001000 0.01.02.03.04.05.06.001002003004005006007008009001000Depth in meters 0.01.02.03.04.05.06.001002003004005006007008009001000 0.01.02.03.04.05.06.001002003004005006007008009001000 0.01.02.03.04.05.06.001002003004005006007008009001000 -80-70-60-50-40-30-20-10001002003004005006007008009001000SAMS1SAMS2 Figure 4.3 (Continued). Transmission loss data with distance for eight tones used during sound transmission experiments conducted in four shallow areas (PSB, SKF, NWPSB, and SPSB). Except for SKF, two transect lines were done in each experiment and each transect line is represented with a number next to the code of the corresponding area. Theoretical attenuation based on cylindrical and spherical spreading is also shown. A profile of the depth contour of each transect is also included. 139

PAGE 152

Transmission Loss (dB)Distance in metersDistance in meters 5000 Hz 7000 Hz 9000 Hz 11000 Hz 13000 Hz 15000 Hz 17000 Hz 19000 HzFrequencies Depth in metersTransmission Loss (dB)Distance in meters -80-70-60-50-40-30-20-10001002003004005006007008009001000Main Channel -80-70-60-50-40-30-20-10001002003004005006007008009001000San Remo 0.01.02.03.04.05.06.001002003004005006007008009001000 -80-70-60-50-40-30-20-10001002003004005006007008009001000Depth in meters 0.01.02.03.04.05.06.001002003004005006007008009001000Ana Maria Sound -80-70-60-50-40-30-20-1001001002003004005006007008009001000Distance in meters6.0 0.01.02.03.04.05.001002003004005006007008009001000Cortez Channel 0.01.02.03.04.05.06.001002003004005006007008009001000Transmission Loss (dB)Distance in metersDistance in meters 5000 Hz 7000 Hz 9000 Hz 11000 Hz 13000 Hz 15000 Hz 17000 Hz 19000 HzFrequencies Depth in metersTransmission Loss (dB)Distance in meters -80-70-60-50-40-30-20-10001002003004005006007008009001000Main Channel -80-70-60-50-40-30-20-10001002003004005006007008009001000San Remo 0.01.02.03.04.05.06.001002003004005006007008009001000 -80-70-60-50-40-30-20-10001002003004005006007008009001000Depth in meters 0.01.02.03.04.05.06.001002003004005006007008009001000Depth in meters 0.01.02.03.04.05.06.001002003004005006007008009001000Ana Maria Sound -80-70-60-50-40-30-20-1001001002003004005006007008009001000Distance in meters6.0 0.01.02.03.04.05.0010020030040050060070080090010006.0 0.01.02.03.04.05.001002003004005006007008009001000Cortez Channel 0.01.02.03.04.05.06.001002003004005006007008009001000 Figure 4.4. Transmission loss data with distance for ten tones used during sound transmission experiments conducted in four channels (Main Channel, San Remo Channel, Anna Maria Sound, and Cortez Channel). Theoretical attenuation based on cylindrical (lines) and spherical spreading (dashed lines) are also shown. A profile of the depth contour of each transect is also included. 140

PAGE 153

155 dB 160 dB 165 dB LFW HFWShallow water areas HFW LFWChannelsSRCNAMS MCCCSource level 10100100010000100000 10100100010000100000Seagrass(PSB2) Sand(SKF) Mud(SPSB2)Mud(SAMS1)Sandy mud(NWPSB2)Distance in meters 10100100010000100000 10100100010000100000 155 dB 160 dB 165 dB 155 dB 160 dB 165 dB LFW HFWShallow water areas HFW LFWChannelsSRCNAMS MCCCSource level 10100100010000100000 10100100010000100000Seagrass(PSB2) Sand(SKF) Sand(SKF) Mud(SPSB2)Mud(SAMS1) Mud(SPSB2)Mud(SAMS1)Sandy mud(NWPSB2)Distance in meters 10100100010000100000 10100100010000100000 Figure 4.5. Estimated active space of low-frequency whistles (LFW = 7-13 kHz) and high-frequency whistles (HFW = 13-19 kHz) with different source levels in channels and shallow water areas of Sarasota Bay, Florida. Distance is presented in logarithmic scale. Labels are defined in Table 1. 141

PAGE 154

5. DEFINING A GROUP IN SPECIES WITH FLUID RELATIONSHIPS 5.1. Introduction Understanding individual relationships is extremely useful in the description of groups. When those relationships are understood, it is possible to determine whether a cluster of animals in space and time is aggregating around a resource or whether the cluster has a social meaning. Normally, individuals interacting with each other more than with other conspecifics are referred as a group (Pulliam and Caraco 1984); however, definitions of what constitutes a group, school, or pod and the procedures used to measure it are variable (Wells et al. 1999). The definition of a group also depends on the species under consideration because different species exhibit different levels of group stability/fluidity. A group of African elephants (Loxondonta africana) is different than a group of chimpanzees (Pan troglodytes), because a group of African elephant is formed by matrilineal groups that do not change unless the matriarch dies and the group fissions (Nyakaana et al. 2001) while a group of chimpanzees changes in composition and size on a regular basis (Wrangham 1986, Chapman et al. 1994, 1995). Similarly, a pod of killer whales (Orcinus orca) is different than a group of spinner dolphins (Stenella longirostris), because a pod of killer whales is formed by matrilineal groups of virtually unchanging composition (Baird 2000), while a group of 142

PAGE 155

spinner dolphins typically changes composition from day to day (Norris et al. 1994; Wells et al. 1999). In species exhibiting flexible grouping patterns, like the spinner dolphins, chimpanzees, spider monkeys (Ateles geoffroyi; Chapman et al. 1995), and some kangaroos (Macropus giganteus; Jarman 2000), relationships are not always clear unless individuals are sighted together repeatedly over time. Under those circumstances, it would be difficult to determine whether individuals are associates that interact frequently or whether they are together as the result of aggregating around a resource or through random encounters. The distinction between the two types of interactions is important because each participant in a relationship is involved also in others, so that the relationship forms part of a network of relationships or social structure (Hinde 1983). Different parameters can be used to describe and measure relationships. When the relationship involves spatial proximity, it is useful to know how much time animals spend in close proximity and how much time they spend apart as this gives some indication of the strength of the association. For instance, mammalian infants spend much of their time near their mother for nursing, maternal grooming, protection, and even play (Gibson and Box 1999). However, as infants grow up they tend to spend more and more time away from their mother but presumably within contact range. Information on how much time and how far an infant is away from its mother can be used an indicator of the level of independence of the infant. The use of proximity can help to describe different types of relationships. Competition for food resources may result in individuals spreading over large distances to avoid each other and reduce food competition. In fact, in some species individuals may be 143

PAGE 156

hundreds of meters from each other when they are foraging. Some examples of species that separate to forage include chimpanzees (Pan troglodytes; Chapman et al. 1994), gray woolly monkeys (Lagothrix lagotricha; Peres 1996), blue monkeys (Cercopithecus mitis; Cords 1987), and spider monkeys (Ateles; Chapman and Lefebvre 1990). Thus, the range between individuals varies according to context and type of the association, and presumably to the limitations of sensory systems relative to staying in contact. For both a mother and her dependent infant and for a dispersed group of individuals, it is important to maintain contact to be able to obtain some of the benefits of group life (e.g., protection from predators, between-group competition). Visual communication can be used to find the location of particular associates. For example, roe deer (Capreolus capreolus) can see each other over a distance often exceeding 100 m before they join (Gerard et al. 2002). However, visual communication is limited in areas such as a dense forest, nocturnal situations, or a turbid water environment. Under these conditions, acoustic communication may be more important to maintain contact since sound travels over large distances. Indeed, acoustic communication is used to coordinate activities and locate distant partners. In a number of primate species, individuals use acoustic signals to coordinate their movements as they travel through the forest. Travel is initiated when a monkey moves to the edge of a group and produces the species-travel call(s) (Boinski 1991, Boinski and Campbell 1995). In species such as adult common marmosets (Callithrix jacchus), individuals use calls to reunite the group during separation from social companions (Norcross and Newman 1993). Thus, information on the limitations of how far animals can communicate is important to understand how communication patterns are related to grouping patterns and spatial structure. 144

PAGE 157

The description of groups is complex because the behavior an individual shows depends in part on whom the individual is with. To understand social behavior it is necessary first to understand how individuals behave with other specific individuals (Hinde 1983). The most biologically-meaningful description of groups would integrate parameters that are important for maintaining individual relationships. In this chapter I discuss how spatial and temporal information on individual relationships and knowledge of the potential communication range can be used to define groups in species with flexible grouping patterns. The study focused on the bottlenose dolphin (Tursiops truncatus), a species with a fission-fusion social organization characterized by flexible relationships in which individuals remain together for varying periods and then split, or after diverging by hundreds of meters, individuals reunite, only to separate again later. While the focus is on bottlenose dolphins the process of defining a group is general and will apply to any species with a similar social organization, such as spider monkeys (Chapman et al. 1993, 1995), chimpanzees (Chapman et al. 1993, 1995), or kangaroos (Jarman and Southwell 1986). In addition to the short-term associations exhibited by bottlenose dolphins, there are also longer-term associations between some individuals, particularly between mothers and calves and between adult males (Wells et al. 1987, Wells 1991). Mothers and dependent calves are usually together until a new calf is born or a calf reaches independence, around an age of 3-4 years (Wells et al. 1987, Wells 1991). As the calf becomes older, it sometimes wanders away from its mother during foraging trips or to interact with other dolphins (Chapter 2; Mann 1997, Mann and Smuts 1998). Mothers and calves form relationships with a variety of associates, but those associations are ephemeral. Yet, those relationships are presumably important for both the mother and the calf as the mother may benefit from 145

PAGE 158

having other adult companions that could help provide protection to the calf (Mann and Smuts 1998) and the calf develops its own relationships, some of which can last for a lifetime (e.g., some male calves form a permanent relationship as adults with other adult males (Wells et al. 1987; Wells 1991, 2003)). The study of the relationship between a mother and calf and of their interactions with other dolphins can be used as a model to investigate what constitutes a group in a fission-fusion society. I evaluated a series of a number of parameters including coordination of activities, headings, duration of associations, maximum distance of separations, and estimated communication range to describe grouping patterns of dolphins. 5.2. Methods General methods involved the use of focal animal observations to collect data on 5 parameters: coordination of activities, headings, duration of associations, maximum distance of separations, and class of associates. Methods used to collect those parameters are described in detailed in Chapter 2 but I provide a general description here. I also used estimates of active space to understand how those parameters are related to the spatial structure of dolphins. Information on estimates of active space comes from the literature, mainly from Quintana-Rizzo et al. (2006). 146

PAGE 159

5.2.1. Parameters: Coordination of Activities, Headings, and Class of Associates Observations were conducted in the summers of 2001, 2002, and 2003. The study focused on twelve well-known females with dependent calves (3-4 years old), which are members of the long-term, multi-generational resident community of about 160 bottlenose dolphins living in a fission-fusion society in Sarasota Bay, Florida (Wells et al. 1980, 1987; Wells 1991, 2003). Focal females were visually identified by the researchers from distinctive markings on their dorsal fins. The study used a combination of focal animal and scan sampling observations (Altmann 1974, Mann 2000). Focal animal observations were conducted on the mother of mother-calf pairs. Data were collected in two ways: 1) at 3-min time intervals (activity, heading, and distance) and 2) continuously at each surfacing of the focal mother (distance). In all seasons, at 3-min instantaneous time points, I recorded the distances and headings of all the dolphins in the observation zone with respect to focal females when dolphins were at the surface. Distances and headings (parallel, heading away, and heading towards) were recorded simultaneously so that heading information could be matched to a particular distance. Definitions of behavioral activities collected at 3-min intervals are described in Chapter 2. Activities were categorized as coordinated if dolphins displayed the same activity as the focal mother, and uncoordinated if they displayed different activities. In 2002 and 2003, continuous sampling was added to the data collection protocols. In the case of cetaceans, continuous sampling data are recorded when an animal surfaces, as they are typically not visible to the observer when they are below the surface (Mann 2000). I was able to keep track of dolphins during each sampling period with the help of 1-2 additional trained observers. Each observer monitored a subset of animals and pointed out 147

PAGE 160

their location each time a focal female surfaced. I called out the surfacings of focal females so that observers knew when to indicate the locations of other dolphins. When those dolphins surfaced (at the same time as a focal female or within approximately 30 sec of her surfacing), I estimated their distances to focal females. During observations, the research vessel was kept at a distance of approximately 20 m from the focal females. The observation zone had a radius of 200-m from the focal female, because this was considered to be the maximum range over which dolphins could be accurately observed. Associates were defined as any dolphins that united with a focal mother during an observation. The term does not include dependent calves, which are always with their mother. The basic classes of associates considered were mothers with calves, adult females without calves (referred to as single adult females), adult males, and juveniles. Information on gender and/or age-class are known for most of the resident dolphins of Sarasota Bay (precisely known age = 74%, known age class = 86%, known gender = 73%, known gender and precisely known age = 64%, R. Wells, pers. comm.). For each focal female, the percentage of associates that fell into each category was quantified. 5.2.2. Parameters: Duration of Associations and Maximum Distances of Separations Two types of flexible grouping patterns were studied: 1) temporary unions and 2) temporary separations. Temporary separations were events in which a focal mother and her calf or other associates moved to distances more than 20 m from each other. The separation ended when the distance apart closed to less than or equal 20 m. Since some temporary separations were less than 3-min long, estimates of the duration of each temporary separation event were based on the continuous data. The mean maximum distance of separations 148

PAGE 161

between mothers and calves and between mother/calf pairs and other associates were calculated. Temporary unions started (joining) when the distance between a focal mother and the associates was equal to or less than 200 m, continued as the distance between dolphins decreased and it was equal to or less than 20 m. A temporary union ended when the distance between animals increased (leaving) to more than 200 m. A distance of 20 m was chosen because 75% (n = 2087) of the dolphins observed in 53 focal follows conducted in 2001 were within 20 m of focal females. Positional data of dolphins allowed me to estimate the time at which they united and separated from focal females. Since none of the temporary unions were less than 3-min long, estimates of their duration were based on the data collected at 3-min intervals. Such information can be collected in a similar way for species such as kangaroos in open grassland (Jarman 2006), or similar information can be inferred based on travel speed for species such as spider monkeys where visibility is often limited to 50 m or so by dense vegetation (C. Chapman pers. comm.). Estimates of the duration of temporary unions were used to determine the proportion of time that an observer needs to stay with a set of animals. This information is particularly valuable for researchers doing surveys of species with flexible grouping patterns, in which individuals leave and join on a regular basis. To estimate the proportion of time needed to stay with a group of animals, I used a cumulative plot of the number of observations against the estimates of the duration of associations at 3-min intervals since none of the temporary unions were less than 3-min long. The relative frequency of observations is progressively accumulated so that the plot reaches an asymptotic value around the time needed to see most of the associations that will occur during a given day (Tobias and Trindade 1995). After this 149

PAGE 162

time, the number of associations recorded for a given animal on a day does not increase significantly even if the observation time increases. Long-term projects that have collected large numbers of instantaneous snapshots of temporary associations will probably give the same picture of association patterns as these more detailed observations. 5.3. Results 5.3.1. Parameters: Coordination of Activities, Headings, and Class of Associates The coordination of activities of dolphins decreased with distance (Figure 5.1 and Figure 5.2). Dolphins displayed coordinated activities at short distances, on average, at distances of 24 37 m from focal mothers. They displayed uncoordinated activities at distances of 70 53 m from focal females (n = 4753 3-min data points). Headings of dolphins also varied with distance (Figure 5.1 and Figure 5.2). A large percentage of the time (86%, n = 2920 3-min data points), dolphins moved parallel to focal females when they were within 20 m, but as inter-dolphin distances increased to 30-m, dolphin headings changed to mainly heading towards. From 40 to 100 m, two headings were more common (heading towards and heading away), but after 100 m, dolphins were mostly heading away (n = 5273 3-min data points). The class of associates observed in every distance category up to 100-m from the focal mother was quantified to examine their spatial structure. Three other categories of dolphins were most commonly seen at 5 m from focal females: mothers with calves (34%, n = 145), juveniles (35%, n = 152), and adult males (24%, n = 101). Single adult females 150

PAGE 163

were also observed but at a much lower frequency (7%, n = 31). At any other distance, the most common class of associate observed was other mothers with their calves and the percentage of dolphins identified in this category ranged from 40% to 62% (48 7 %; Figure 2.11). The second two most common classes of associates were juveniles and adult males at distances from 10 m to 100 m. Adult males were more common at 10 m and juveniles were more common at 20 m from focal females (Figure 2.10). However, at almost any other distance, when the percentage of adult males increased, the percentage of juveniles decreased, and vice versa (Figure 2.12), indicating a low level of association between these two age classes of males. 5.2.2. Parameters: Duration of Associations and Maximum Distances of Separations Temporary unions lasted 25 26 min and ranged from about 3 min to 129 min (N = 105, maximum follow time = 318 min; Figure 5.3A). The cumulative plot of the number of observations against duration of associations indicates that an observer needs to remain with a female and her calf for at least 30 minutes to observe 75% of her associates on a given day (Figure 5.3B). If an observation lasts around an hour, the observer would record approximately 90% of the associates. However, if an observer only stayed with a group for 10 minutes, only approximately 30% of the associates would be identified on a given day. Temporary separations between mothers and their calves lasted for 1 min to 47 min, with a mean separation time of 9 9 min. Mean maximum distances of separation between females and calves were significantly greater than the maximum distance of separation between females and associates (U = 4450.50, p = < 0.001, N = 227). The mean maximum distance of separation between focal females and calves was 82 46 m and it ranged from 30 151

PAGE 164

m to over 400 m (normally less than 250 m). The mean maximum distance of separation between focal females and associates was 60 30 m. 5.3. Discussion Understanding how specific relationships shape the social organization of a species is basic to describing groups. When individual relationships and association patterns are dynamic, and communication abilities are complex, the identification of groups can be difficult. This is particularly true in species with a fission-fusion social organization, because the ephemeral nature of the associations makes the identification of groups difficult. A way to address this difficulty is to study parameters that are indicative of the interactions between individuals. In this study, I compared the spatial relationships, the coordination of activities and headings, and the identity of individuals when they were in close proximity and when they were spread. This approach proved to be helpful in identifying the characteristics of different types of groups. Researchers studying other species with flexible grouping can use a similar approach. This would allow making reasonable comparisons between species, especially in the case of studies interested in testing how ecological and social pressures influence the social organization of different species. For example, if coordination of activities is found to be an important parameter in defining a group of chimpanzees, kangaroos, and bottlenose dolphins, then it is possible to examine how ecological pressures may influence it in those species 152

PAGE 165

since coordination is viewed as an adaptation to gain benefits of decreased predator pressure (Hamilton 1971). Coordination of activities (Ruckstuhl and Neuhaus 2001, Gerard et al. 2002) and headings are viewed as adaptations to maintain group cohesion in some species. The spatial structure of bottlenose dolphins was cohesive out to a radius of about 25-30 m, in that their activities were coordinated and their headings were mostly parallel (Figure 5.1 and Figure 5.2). Interestingly, the type of relationship formed between a focal mother and a particular type of dolphin seemed to determine the position of that particular dolphin within this area. The most obvious relationship was between a mother and her dependent calf. The mother functions as the central base in the calfs world from which the infant can explore and socialize with other conspecifics. By swimming in close proximity to their mother, calves can nurse, gain protection, and increase their opportunities to learn feeding techniques (Mann 1997) and probably habitat use. Additionally, within very close range (< 1 m) the drafting hydrodynamics of the mother can reduce the calfs energy expenditure required for movement (Weihs 2004). Thus, a group formed by a mother and a dependent calf relies on a close range interaction occurring within a range of approximately 25-30 m. Other relationships also depend on close range interactions. The most common associates near ( 5 m) focal females were other mothers with calves, adult males, and/or juveniles. Females with calves were also more common at other distances. Females with calves have been reported to form associations with each other (Wells et al. 1993, Wells 1991). Mann (2000) suggested that females benefit from associating with other females of similar reproductive condition because they share similar problems with predators. Females with calves may be more aware of threats to their calves than other classes of individuals (e.g. 153

PAGE 166

juveniles or reproductive males). Females calves may also benefit from gaining protection from predators by being in large group (Wells 1991). In the absence of other mothers and calves as potential associates within 10-20 m of the focal mother, adult males were the next closest associate. Adult males were on average nearer to the focal mothers than were juveniles; juveniles were at the periphery of the cohesive area. Single females were the least common associate of mothers with calves. The proximity of adult males to mothers could be related to the possibility that males were investigating the reproductive status of females as females become more receptive a few years after calves are born (Connor et al. 1996). This type of relationship relies on being in tactile. It has been hypothesized that male dolphins may assess the reproductive status of females by approaching and placing their rostrum within a few centimeters of the females genital area. Echolocation is apparently then used to detect tissue changes associated with oestrus (Connor et al. 1996). If female indicate their receptivity with visible behavioral cues as is common with other mammals, then horizontal water clarity would be a limiting factor for the proximity needed for assessment, and may be limited to 3 m or less in Sarasota Bay. During consortships, males typically travel abreast behind a female or flank her on either side and slightly behind (Connor et al. 1996, Owen 2003). Thus, the relationship between potentially receptive females and adult males can be described as a cohesive group spreading over an area of approximately 20 m. Cohesiveness decreased as inter-individual distances increased. At those times, the coordination of the activities decreased with distance. At distances of 40-100 m, parallel swimming was uncommon and the dolphins headings were variable. Beyond 100 m, most dolphins were heading away from focal females. When individuals separated, the proportion 154

PAGE 167

of time spent feeding, socializing (with other dolphins), and milling increased by more than 10% (Chapter 2). Foraging animals may separate to reduce feeding competition (Cowlishaw 1999). Additionally, groups moving in a dispersed manner can increase the likelihood of locating or finding prey or predators because the area searched is much greater than what a tighter group could cover (Norris and Dohl 1980; Garber 2000). By spreading, individuals can feed in areas that would not support a cohesive group (Kinzey and Cunningham 1994). When group members are separated, they must be able to find each other again. A foraging/socializing/milling group of dolphins was typically spread over an area around 60 30 m. This distance is more than twice the size of the area used when individuals have a cohesive spatial structure. Clearly, tactile communication is of no use under these circumstances, as the distance greatly exceeds the dolphins body length and the 40 cm average length of the flippers, a dolphins longest appendage. When water clarity is only a few meters as it is typically in Sarasota Bay, then visual communication is also impossible, but dolphins were still within communication range (Quintana-Rizzo et al. 2006). Typically, whistles were used when animals were separated around 63 2 m (Chapter 3). Whistles are narrow-band frequency-modulated sounds ranging from about 4 to 20 kHz (Caldwell et al. 1990) and they are used for long distance communication (Janik 2000a, Tyack 2000a,b). Thus, in separations involving more than two dolphins, the use of whistles may be particularly important when individuals need to find and recognize specific associates among many dolphins. Whistles can convey signature information for individual recognition (Sayigh et al. 1990, Smolker et al. 1993, Janik et al. 2006), and they may be used for spacing of individuals and coordination of activities (Janik and Slater 1998). Even at the greatest 155

PAGE 168

maximum distance of separation (approximately 450 m), dolphins were within acoustic communication range (Quintana-Rizzo et al. 2006). Calves also temporarily separated from their mothers. As calves become more independent, they venture away from their mothers. Thus, the mother-calf relationship changes over time as do the distances over which mother and calf interact. A group formed by mothers with dependent calves of age 3-4 years typically spread over a distance around 80 m (Fig. 5.1). Separation distances between mothers and dependent calves were well within the estimated active space of whistles in different habitats (Quintana et al. 2006). The results suggest that separation distances are not necessarily determined by the maximum communication range. Other factors such food distribution or travel time to reunite may be important in determining how far calves wander away from their mother. A calf may not wander far from its mother if the risk of predation is high. Females may benefit from temporary separations by being able to have more uninterrupted foraging opportunities (Mann and Smuts 1998). A calf may benefit from separating often from its mother by gaining hunting and social experience leading to independence (Mann 1997, Connor et al. 2000). Mann and Watson-Capps (2005) proposed that frequent mother-calf separations are indicators of the vigor of the calf because a calf has to be in good condition to venture hundreds of meters away and successfully return to its mother. Mean maximum distances of temporary separations between mothers and calves and between mothers and associates provide a conservative, practical means of determining the bounds of a group in the field. When information on the spatial distribution of group members is available, it is possible to determine if independent focal-follows should be 156

PAGE 169

conducted when individuals are outside of a particular range. Mean maximum distances are not strict values, but they provide some guidance to researchers who need to collect data on group size and do not need to examine individual relationships. Understanding the patterns of relationships in groups requires information on the type of individuals forming those relationships (Hinde 1983). This also requires knowing how much time animals spend associating. Yet, for the observer, it is sometimes difficult to determine how long a researcher should stay with a given group to record a meaningful number of associates. This is important in species with flexible grouping patterns in which individuals leave and join on a regular basis (Chapman et al. 1993). Information on the duration of associations suggests that observations of mothers and calves should last between 30-60 min to observe 75-90% of a females associates on a given day, respectively. This information is particularly valuable for researchers doing surveys and the best observation time will depend on the kind of question that the observer is interested in answering. This time criterion is useful and it may vary for different types of individuals such as adult males and juveniles. Still, it provides a first step towards collecting a meaningful number of associations for females with calves. The amount of time spent observing animals may also depend on whether photographs have to be taken to identity particular individuals based on their distinctive markings (i.e. stripe pattern of zebras, Neuhaus and Ruckstuhl 2002; dorsal fin or tails marks of cetaceans, Hammond et al. 1990). Wells and Scott (1990) reported that on average it took 19 min to achieve a criterion of 95% certainty that photographs were obtained of all bottlenose dolphins in a radius of 100 m. 157

PAGE 170

The time that animals join and leave each group should be recorded (Southwell 1984, Cairns and Schwager 1987, White 1988, Chapman et al. 1995, Lehman and Boesch 2004). Some definitions of a group take into account the duration of an association. Yet, some researchers argue that recounting a group after each change in composition would actually exacerbate problems of non-independent sampling (Smolker et al. 1992). In concordance with Wells et al. (1987), I argue that group membership should be noted after every change because problems are introduced in the analysis of association patterns if this is not done. For example, if changes of group memberships are not noted and two animals leave and join the same group at different times, those individuals would be considered associates when they were not. The distance criterion for defining joining and leaving would depend on the species in question. Additionally, information on joining and leaving events is necessary to measure the frequency of association between individuals in fission-fusion societies. The observer must take into account the number of times that particular individuals are sighted in the same group and the number of times that each individual is sighted separately (Cairns and Schwager 1987). Because of this, I recommend that the study of a group take into consideration the behavior of the individual, especially when social interactions are of primary interest to the researcher. Under such circumstances, I suggest that a protocol be followed to randomize observations for specific individuals. The parameters examined to study groups have provided vital information on individual relationships. The results may be suitable to describe other social units but they will probably vary for other species with similar types of social organization. In some species, such as chimpanzees, for example, females do not call when they are temporarily separated from a group whereas males do (Chapman et al. 1994). Thus, definitions of group 158

PAGE 171

159 may change according to the social unit of inte rest if acoustic communication is an important factor to find and locate partners over long distances. In terrestrial species, other sensory modes that may play an important part of soci al relationships such as vision and touch (i.e., grooming) should be consider ed when defining a group.

PAGE 172

0204060801001205102030405060708090100110120130140150160170180190200 Mean maximum distance of separation SD Coordinated activitiesParallel headingHeading towardsHeading away Distance from focal females in metersPercentage of observations 140 0204060801001205102030405060708090100110120130140150160170180190200 Mean maximum distance of separation SD Coordinated activitiesParallel headingHeading towardsHeading away Distance from focal females in metersPercentage of observations 140 Figure 5.1. Inter-relationships of separation distance (mean maximum + SD) between mothers and calves, activity coordination, and headings relative to defining groups. The active space of social sounds exceeds the full range of behavioral measures (active space of a 7-13 kHz whistle with a source level equal to 160 dB: shallow seagrass area = 301 m, Channel = 13 km; Quintana-Rizzo et al. 2006). 160

PAGE 173

161 0204060801001205102030405060708090100110120130140150160170180190200 Coordinated activitiesParallel headingHeading towardsHeading away Distance from focal females in meters Perc ent arns ge of obse vatio 140 Mean maximum distance of separation SD 0204060801001205102030405060708090100110120130140150160170180190200 Coordinated activitiesParallel headingHeading towardsHeading away Distance from focal females in metersarns Perc ent ge of obse vatio 140 Mean maximum distance of separation SD Figure 5.2. Inter-relationships of separation distance (mean maximum + SD) between mothers and associates, activity coordination, and headings relative to defining groups. The active space of social sounds exceeds the full range of behavioral measures (active space of a 7-13 kHz whistle with a source level equal to 160 dB: shallow seagrass area = 301 m, Channel = 13 km; Quintana-Rizzo et al. 2006).

PAGE 174

02040608010012039152127333945515763697581879399105111117123129Duration of association in minutesCumulative number of observations BTime in minutes 02468101214Frequency 39152127333945515763697581879399105111117123129 A 0204060801001203915212733394551576369758187939910511111712312939152127333945515763697581879399105111117123129Duration of association in minutesCumulative number of observations BTime in minutes 02468101214Frequency 39152127333945515763697581879399105111117123129 ATime in minutes 02468101214Frequency 3915212733394551576369758187939910511111712312939152127333945515763697581879399105111117123129 A Figure 5.3. A. Duration of associations in 3-min intervals between focal females and other dolphins observed in Sarasota Bay. B. Cumulative plot of the number of observations recorded for associations of different durations in 3-min intervals. 162

PAGE 175

6. THESIS SUMMARY The dynamic interactions between individuals of species with fission-fusion social structure have been of interest for several decades. Flexible grouping patterns are intriguing because individuals join and leave groups frequently in response to regular changes in the costs and benefits of being a member of a group (Wrangham et al. 1993, Kinsey and Cunningham 1994). Different studies have examined several aspects of this type of social organization including ecological pressures (Wrangham et al. 1993, Chapman et al. 1995), social factors (White 1989, 1992, Packer et al. 1990, Chapman et al. 1995), and communication between mother-calf pairs and male alliances (Smolker et al. 1993, Smolker and Pepper 1999, Watwood 2003). The present study provided an opportunity for new insights into how social factors such as communication patterns, along with ecological pressures, shape the social organization of the bottlenose dolphin. The first step towards understanding fission-fusion was to quantify the rate at which associations change. Overall, association changes occurred on average every 26 min. Examination of the context in which mothers exhibiting these rates joined and split from groups indicated that both social and ecological pressures played an important role in determining group fission-fusion. Mothers with calves exhibiting high rates of fission-fusion regularly occupied deep waters. In that habitat, females frequently encountered other females in the same 163

PAGE 176

reproductive condition and associated with them. In contrast, mothers with calves exhibiting low fission-fusion rates often occupied shallow waters where they frequently encountered juvenile dolphins, likely considered less desirable, or at least less frequent, temporary associates. The high level of group formation among females with calves may increase their protection from predation in addition to providing more playmates for their calves (Altmann 2000). Closer examination of dynamics among dolphins revealed interesting information about group coordination and communication patterns during temporary separations. There was a sharp decrease in coordinated activity and headings at distances greater than 25-30 m between focal females and other dolphins. As groups spread, they became less cohesive, spending more time foraging, milling, and socializing than when they were in close proximity. In fact, as distance between dolphins increased, individuals more likely displayed uncoordinated activities even though they were still associated. Different separation ranges appear to have different social meaning. Mothers and calves are in closer contact for nursing and tactile contact. Males are closer to mothers than other non-calf individuals probably because males need to be close enough to evaluate reproductive state. The extent of such close proximity to mothers may not be so important for other relationships. In some species, younger individuals tend to be closer to the center of a group where they may enjoy greater security than those at the edge (Krebs and Davis 1993). Yet, in this study, when juvenile and adult male dolphins were the most second category of individuals in the group, juveniles were in the periphery of the group while adult males were in the center, closer to mothers. The presence of 164

PAGE 177

juveniles in the periphery of groups is common pattern is social species with prolonged maturation such as baboons (Papio ursinus; Cambefort 1981). Separation distances differed between mothers and their calves and between mothers and their other associates. Surprisingly, separation distances between mothers and calves were greater than those of mothers and associates. Since calves depend on their mothers for survival, it would seem that the distance of separation between mother and their calves would be shorter than the distance of separation between mothers and their other associates. However, considering that associate separations occur mostly in deep water, which may be more dangerous in terms of predation risk, it follows that a more cohesive spatial structure should be displayed by groups using those waters. Dolphins often separated themselves by distances much greater than those used by Ballance (1990), Wells and Scott (1990) and Smolker et al. (1992) to define groups in terms of distance. Although those definitions have facilitated data collection, they define groups as a collection of individuals instead of describing groups based on individual interactions. The interactive approach could help clarify details of social dynamics, and might help to distinguish between congregations (animals drawn together for social reasons) and aggregations (animals drawn together for non-social reasons). Acoustic signals are an important means of locating specific partners in species with social bonds (Nishida 1968, Byrne 1981, Boinski and Campbell 1995, Janik and Slater 1998). Acoustic signals are also important in maintaining group cohesion in areas with limited visibility (e.g., dense forest, Byrne 1981; murky waters, Janik and Slater 1998) or over distances ranging up to several kilometers (Smolker et al. 1993, Tyack 2000b). Yet, results from this research indicate that when dolphins separate over short 165

PAGE 178

distances (<50 m), they do not have to use acoustic signals to find and locate each other. It is possible that when groups are more cohesive and dolphin activity and headings are coordinated, no acoustic signals are needed to coordinate a group. As distance of separation increased between dolphins, the use of communication signals became more important. At moderate distances of separation (e.g. 70 m for mother-calf pairs), echolocation trains were commonly produced. This finding was surprising because whistles have been documented as the main signal used during separations between dolphins (Smolker et al. 1993, Watwood 2003). However, it is unclear if those studies examined if echolocation was used or if dolphins might echolocate less frequently in habitats with better visibility like the one studied by Smolker et al. (1993). The production of echolocation trains but no whistles in the separation context suggests that echolocation is used, directly or indirectly (e.g. eavesdropping), by dolphins to reunite. Echolocation may be use to locate food and navigate; however, dolphins might be able to keep track of each others position if they eavesdrop on each others echolocation calls in a similar way that other echolocating species do (Fenton 2003, Pzalker and Kush 2003). When distances of separation increased even more (e.g. 100 m for mother-calf pairs), both whistles and echolocation were produced. At greater distance of separation, the use of whistles may be particularly important to find specific associates because whistles can convey information for individual recognition (Sayigh et al. 1990, Smolker et al. 1993, Janik et al. 2006). Additionally, whistles may be used for spacing of individuals and coordination of activities (Janik and Slater 1998), which would be needed to bring dolphins back together. In this context, the use of only echolocation may not be 166

PAGE 179

the most effective way to find particular dolphins, unless dolphins can recognize individuals using echolocation. The study of fission-fusion dynamics and communication distances determined that temporary separations were within the estimated maximum communication range of whistles. The result indicated that separation distances were not determined by the maximum communication range. Other factors, such as predation pressure and food distribution, may be important determinants of how far dolphins separate from each other. Dolphins may spread over distances allowing them to find the best food patches. Additionally, factors like ambient noise may have an effect on separation distances because ambient noise can affect communication range (Urick 1967, Forrest 1994, Janik 2000b, Slabbekoorn 2004). In this respect, the results of the present study showed the best-case scenario of sound propagation and estimates of active space, because experiments were conducted when no boats were present within a radius of approximately 1 km of the recordings. The present study demonstrates the value of examining dynamics of fission-fusion at the individual level. Findings reveal that although the bottlenose dolphin as a species exhibits a flexible social organization, individuals exhibit distinct grouping patterns. Additionally, the study showed the importance of examining a broad spectrum of acoustic frequencies to better describe patterns of possible communication (e.g. the absence of acoustic signals during short separations). The quantification of acoustic active spaces suggested that the distances over which dolphins remain in acoustic contact and can be considered members of groups may in some cases be much greater than that described based on dolphin spacing and activity alone. 167

PAGE 180

This study serves as the basis for more detailed studies of grouping patterns. Future research could use a similar approach to examine the dynamics of fission-fusion in other age-sex classes of the bottlenose dolphin. This is important in order to develop a detailed picture of how grouping patterns are shaped by different types of relationships. For example, juveniles, single adult females, and adult males may display different patterns according to their biological needs than mothers with calves. Proximity seems to be an important factor between a mother and her dependent calf for nursing and protection. However, proximity could play a different role between male pairs or trios that may wander away from each other to increase the chances of finding a mate. The study of active space of whistles used as a reference the only reported source level of whistles for wild bottlenose dolphins (Janik 2000). This source level was estimated for bottlenose dolphins that are bigger than the Sarasota Bay dolphins and that inhabit an unusual habitat for coastal dolphins. The bottlenose dolphins of Janiks study are found in the inner waters of Moray Firth, which have depths of up to about 50 m. Differences in habitat characteristics and dolphin body size may influence the source level of whistles. Bigger dolphins living in deeper waters may use louder whistles than smaller dolphins living in shallower habitats. A study should estimate the source level of whistles used by bottlenose dolphins inhabiting coastal areas. It is important to estimate the propagation of echolocation clicks. Dolphins typically use echolocation for navigation and for finding food. Both activities required that the signal travel long distances, but the distances over which echolocation clicks propagate are unknown. This information can provide insights into how dolphins find and detect food and how they can navigate in different types of habitats. Additionally, 168

PAGE 181

information on the range at which a dolphin can eavesdrop on another echolocating dolphin may provide information on how far dolphins can find each other when they do not whistle. This distance is potentially much greater than the range at which a dolphin can detect echoes of a target since the energy of the echolocating signal is probably greater than the energy of a returning echo. 169

PAGE 182

LITERATURE CITED Altmann, J. 1974. Observational study of behavior: Sampling methods. Behavior 49:227-267. Altmann, J. 2000. Models of outcome and process: predicting the number of males in primate groups. Pp. 236-247. In: Primate males. Edited by P.M. Kappeler. Cambridge University Press, Cambridge. Au, W.W.L. 1993. The sonar of dolphins. New York: Springer-Verlag. Au, W.W.L. 1997. Echolocation in dolphins with a dolphin-bat comparison. Bioacoustics 8:137-162. Au, W.W.L. 2004. A comparison of the sonar capabilities of bats and dolphins. Pp. xiii-xxvii. In: Echolocation in bats and dolphins. Edited by J.A. Thomas, C.F. Moss, and M. Vater. The University of Chicago Press, Chicago. Au, W.W.L., P.W.B. Moore, and J.L. Pawloski. 1986. Echolocation transmitting beam of the Atlantic bottlenose dolphins. Journal of the Acoustical Society of America 80:688-813. Baird, R.W. 2000. The killer whale: foraging specialization and group hunting. Pp. 127-153. In: Cetacean societies. Field studies of dolphins and whales. Edited by J. Mann, R.C. Connor, P.L. Tyack, and H. Whitehead. The Chicago University Press, Chicago. Ballance, L.T. 1990. Residence patterns, group organization, and surfacing associations of bottlenose dolphins in Kino Bay, Gulf of California, Mexico. Pp. 267-283. In: The bottlenose dolphin. Edited by S. Leatherwood and R.R. Reeves. Academic Press, California. Barros, N.B. and D.K. Odell. 1990. Food habits of bottlenose dolphins in the Southeastern United States. Pages 309-328 In: The bottlenose dolphin. Edited by S. Leatherwood and R.R. Reeves. Academic Press, California. Barros, N.B. and R.S. Wells. 1998. Prey and feeding patterns of resident bottlenose dolphins (Tursiops truncatus) in Sarasota Bay, Florida. Journal of Mammalogy 79:1045-1059. 170

PAGE 183

Bejder, L., D. Fletcher, and S. Brger. 1998. A method for testing association patterns of social animals. Animal Behaviour 56:719-725. Benoit-Bird, K.L. and W.W.L. Au. 2004. Echolocation click rates and behavior of foraging spinner dolphins. Journal of the Acoustical Society of America 115:2374. Blomqvist, C., and M. Amundin. 2004. High-frequency burst-pulse sounds in agonistic aggressive interactions in bottlenose dolphins. Pp. 431-455. In: Echolocation in bats and dolphins. Edited by J.A. Thomas, C.F. Moss, and M. Vater. The University of Chicago Press, Chicago. Blumstein, D.T., C.S. Evans, and J.C. Daniel. 1999. An experimental study of behavioral group size effects in tammar wallabies (Macropus eugenii). Animal Behaviour 58:351-360. Boinski, S. 1991. The coordination of spatial position: a field study of the vocal behavior of adult female squirrel monkeys. Animal Behaviour 41:89-102. Boinski, S., and A.F. Campbell. 1995. Use of trill vocalizations to coordinate troop movement among white-faced capuchins: a second field-test. Behavior 132:875-901. Brenowitz, E.A. 1982. The active space of Red-winged Blackbird song. Journal of Comparative Physiology 142:511-522. Brown C.H., M.D. Beecher, D.B. Moody, and W.C. Stebbins. 1979. Locatability of vocal signals of Old World monkeys: design features for the communication of position. Journal of Comparative Physiology 93:806-819. Buckland, S.T., D.R. Anderson, K.P. Burnham, and J.L. Laake. 1993. Distance sampling. Estimating abundance and biological populations. Chapman and Hall, London. Byrne, R.W. 1981. Distance vocalizations of guinea baboons (Papio papio) in Senegal: an analysis of function. Behaviour 78:283-312. Cairns, S.J., and S.J. Schwager. 1987. A comparison of association indices. Animal Behaviour 35:1454-1469. Caldwell, M.C., Caldwell D.K. and P.L. Tyack. 1990. Review of the signature-whistle hypothesis for the Atlantic bottlenose dolphin. Pp. 199-234. In: The bottlenose dolphin. Edited by S. Leatherwood and R.R. Reeves. Academic Press, California. Cambefort, J.P. 1981. A comparative study of culturally transmitted patterns of feeding habitats in the Chacma baboon Papio ursinus and the Vervet monkey Cercopithecus aethiops. Folia Primatologica 36:243-263. 171

PAGE 184

Caraco, T. 1987. Foraging games in a random environment. Pp. 389-414. In: Foraging behavior. Edited by A.C. Kamil, J.R. Krebs, and H.R. Pulliam. Plenum Press. New York, New York. Caughley, G. 1964. Social organization and daily activity of the red kangaroo and the grey kangaroo. Journal of Mammalogy 45:429-436. Chapman, C.A., and L. Lefebvre. 1990. Manipulating foraging group size: spider monkey food calls at fruiting trees. Animal Behaviour 39:891-896. Chapman, C.A., F.J. White, and R.W. Wrangham. 1993. Defining subgroup size in fission-fusion societies. Folia Primatologica 61:31-34. Chapman, C.A., F.J. White, and R.W. Wrangham. 1994. Party size in chimpanzees and bonobos. Pp. 41-57. In: Chimpanzee cultures. Edited by R.W. Wrangham, W.C. McGrew, F.B.M. de Waal, and P.G. Heltne. Harvard University Press, Massachussetts. Chapman, C.A., R.W. Wrangham, and L.J. Chapman. 1995. Ecological constraints on group size: an analysis of spider monkey and chimpanzee subgroups. Behavioral Ecology and Sociobiology 36:59-70. Clark, C.W., and M. Mangel. 1986. The evolutionary advantages of group foraging. Theoretical Population Biology 30:45-75. Connor, R.C., R.A. Smolker, and A.F. Richards. 1992. Dolphin alliances and coalitions. Pp. 415-443. In: Coalitions and alliances in humans and other animals. Edited by A.H. Harcourt and F.B.M. de Waal. Oxford University Press, Oxford. Connor, R.C., A.F. Richards, R.A. Smolker, and J. Mann. 1996. Patterns of female attractiveness in Indian Ocean bottlenose dolphins. Behaviour 133:37-69. Connor, R.C., R.S. Wells, J. Mann, and A. Read. 2000. The bottlenose dolphin: social relationships in a fission-fusion society. Pp. 91-126. In: Cetacean societies. Field studies of dolphins and whales. Edited by J. Mann, R.C. Connor, P.L. Tyack, and H. Whitehead. The Chicago University Press, Chicago. Connor, R.C., and H. Whitehead. 2005. Alliances II. Rates of encounter during resource utilization: a general model of intrasexual alliance formation in fission-fusion societies. Animal Behaviour 69:127-132. Cords, M. 1987. Mixed-species associations of Cercopithecus monkeys in the Kakamega Forest, Kenya. University of California Press. Cowlishaw, G. 1999. Ecological and social determinants of spacing behavior in desert baboon groups. Behavioral Ecology and Sociobiology 45:67-77. 172

PAGE 185

Crook, J.H. 1970. Social organization and the environment: aspects of contemporary social ethology. Animal Behaviour 18:197-209. Croft, D.B. 1980. Behaviour of red kangaroos, Macropus rufus (Desmarest 1822) (Macropidae; Marsupilia) in the northwestern New South Wales, Australia. Australian Mammals 4:5-58. Dawson, S.M. 1991. Clicks and communication: the behavioral and social context of Hectors dolphin vocalizations. Ethology 88:265-276. dos Santos, M.E., and V.C. Almada. 2004. A case for passive sonar: analysis of click train production patterns by bottlenose dolphins in a turbid estuary. Pp. 400-403. In: Echolocation in bats and dolphins. Edited by J.A. Thomas, C.F. Moss, and M. Vater. The University of Chicago Press, Chicago. Dunbar, R.I.M. 1989. Social systems as optimal strategy sets: the costs and benefits of sociality. Pp. 131-149. In: Comparative socioecology. The behavioral ecology of humans and other mammals. Edited by V. Standen and R.A. Foley. Blackwell Scientific Publications, Oxford. Eisenberg, J.F., N.A. Muckenhirn, and R. Rudran. 1972. The relation between ecology and social structure of primates. Science 176:863-874. Farabaugh, S.M., and R.J. Dooling. 1996. Acoustic communication in parrots: laboratory and field studies of budgerigars, Melopsittacus undulates. Pp. 97-117 In: Ecology and evolution of acoustic communication in birds. Edited by D.E. Kroodsma and E.H. Miller. Cornell University Press, New York. Fenton, M.B. 2003. Eavesdropping on the echolocation and social calls of bats. Mammal Review 33:193-204. Ford, J.K.B. 1989. Acoustic behavior of resident killer whales (Orcinus orca) off Vancouver Island, British Columbia. Canadian Journal of Zoology 69:1454-83. Forrest, T.G. 1994. From sender to receiver: propagation and environmental effects on acoustic signals. American Zoologist 34:644-654. Forrest, T.F., G.L. Miller, and J.R Zagar. 1993. Sound propagation in shallow water: implications for acoustic communication by aquatic animals. Bioacoustics 4:259-270. Fripp, D., C. Owen, E. Quintana-Rizzo, A. Shapiro, K. Buckstaff, K. Jankowski, R.S. Wells, and P. Tyack. 2005. Bottlenose dolphin (Tursiops truncatus) calves appear to model their signature whistles on the signature whistles of community members. Animal Cognition 8:17-26. 173

PAGE 186

Gannon, D.P., N.B. Barros, D.P. Nowacek, A.J. Read, D.M. Waples, and R.S. Wells. 2005. Prey detection by bottlenose dolphins, Tursiops truncatus: an experimental test of the passive listening hypothesis. Animal Behaviour 69:709-720. Gannon, D.P. 2005. Quantifying habitat quality for bottlenose dolphins in Sarasota Bay: how do prey distribution and abundance influence habitat selection? Technical Report Harbor Branch Oceanographic Institute. 25 pp. Gerard, J.F., E. Bideau, M.L. Maublan, P. Loisel, and C. Marchal. 2002. Herd size in large herbivores: encoded in the individual or emergent? Biological Bulletin 202:275-282. Gero, S., L. Bejder, H. Whitehead, J. Mann, and R.C. Connor. 2005. Behaviorally specific preferred associations in bottlenose dolphins, Tursiops spp. Canadian Journal of Zoology 83:1566-1573. Gibson, K.R., and H.O. Box. 1999. Mammalian social learning. Cambridge University Press, Cambridge. Gompper, M.E. 1996. Sociality and asociality of white-nosed coatis (Nasua narica): foraging costs and benefits. Behavioral Ecology 7:254-263. Hammond, P.S., S.A. Mizroch, and G.P. Donovan. 1990. Individual recognition of cetaceans: use of photo-identification and other techniques to estimate population parameters. Reports of the International Whaling Commission, special issue 12. International Whaling Commission, Cambridge. Herzing, D.L. 2004. Social and nonsocial use of echolocation in free-ranging Stenella frontalis and Tursiops truncatus. Pp. 404-410. In: Echolocation in bats and dolphins. Edited by J.A. Thomas, C.F. Moss, and M. Vater. The University of Chicago Press, Chicago. Herzing, D., and M.E dos Santos. 2003. Functional aspects of echolocation in dolphins. Pp. 386-393. In: Echolocation in bats and dolphins. Edited by J.A. Thomas, C.F. Moss, and M. Vater. The University of Chicago Press, Chicago. Houser, D. and J. Finneran. 2005. Auditory evoked potentials (AEP) methods for population-level assessment of hearing sensitivity in bottlenose dolphins. Journal of the Acoustical Society of America 117:2408. Hinde, R.A. 1976. Interactions, relationships and social structure. Man 11:1-17. Hinde, R.A. 1983. A conceptual framework. Pp. 1-7. In: Primate social relationships: an integrated approach. Edited by R.A. Hinde. Blackwell Scientific Publications, Oxford. 174

PAGE 187

Irvine, A.B., M.D. Scott, R.S. Wells, and J.H. Kaufmann. 1981. Movements and activities of the Atlantic bottlenose dolphin, Tursiops truncatus, near Sarasota, Florida. Fishery Bulletin 79:671-688. Janik, V.M. 2000a. Whistle matching in wild bottlenose dolphins (Tursiops truncatus). Science 289:1355-1357. Janik, V.M. 2000b. Source levels and the estimated active space of bottlenose dolphin (Tursiops truncatus) whistles in the Moray Firth, Scotland. Journal of Comparative Physiology A. 186: 673-680. Janik, V.M., and P.J.B Slater. 1998. Context-specific use suggests that bottlenose dolphin signature whistles are cohesion calls. Animal Behaviour 56:829-838. Janik, V.M., L.S. Sayigh, and R.S. Wells. 2006. Signature whistle shape conveys identity information to bottlenose dolphins. Proceedings of the National Academy of Sciences of the United States of America 103:8293-8297. Jarman, P.J. 1993. Individual behavior and social organization of kangaroos. Pp. 70-85. In: Animal societies: individuals, interactions, and organization. Edited by P.J. Jarman and A. Rossiter. Kyoto University Press, Japan. Jarman, P.J. 2000. Males in macropod society. Pp. 21-33. In: Primate males: causes and consequences of variation in group composition. Edited by P.T. Kappeler. Cambridge University Press, London. Jarman P.J, and C.J. Southwell. 1986. Grouping, associations, and reproductive strategies in Eastern grey kangaroos. Pp. 399-428. In: Ecological aspects of social evolution. Edited by D.I. Rubenstein and R.W. Wrangham. Princeton University Press, New Jersey. Jensen, F.B. 2001. Acoustics, shallow water. Pp. 89-96. In: Encyclopedia of ocean sciences. Edited by J.H. Steele. Academic Press, California. Johnson, C.S. 1967. Sound detection thresholds in marine mammals. Pp. 247-260. In: Marine Bioacoustics Vol. II. Edited by W.N. Tavolga. Pergamon, New York. Johnson, C.S. 1968. Masked tonal thresholds in the bottle-nosed porpoise. Journal of the Acoustical Society of America 44:965-967. Jones, G.J., and L.S. Sayigh. 2002. Geographic variation in rates of vocal production of free-ranging bottlenose dolphins. Marine Mammal Science 18:374-393. Kappeler, P.M. 2000. Grouping and movement patterns in Malagasy primates. Pp. 470-490. In: On the move. Edited by S. Boinsi and P.A. Garber. The University of Chicago Press, Chicago. 175

PAGE 188

Kappeler, P.M., and C.P. van Schaik. 2002. Evolution of primate social systems. International Journal of Primatology 23:707-740. Kinsey, W.G., and E.P.Cunningham. 1994. Variability of Platyrrhine social organization. American Journal of Primatology 34:185-198. Klump, G.M. 1996. Bird communication in the noisy world. Pp. 321-338. In: Ecology and evolution of acoustic communication in birds. Edited by D.E. Kroodsma and E.H. Miller. Cornell University Press, New York. Krebs, J.R., and N.B. Davies. 1993. An introduction to behavioural ecology. Blackwell Scientific Publications, Cambridge. Lammers, M.O., and W.W.L. Au. 2003. Directionality in the whistles of Hawaiian spinner dolphins (Stenella longirostris): a signal feature to cue direction of movement? Marine Mammal Science 9:213-220. Lehmann, J., and C. Boesch. 2004. To fission or to fusion: effects of community size on wild chimpanzees (Pan troglodytes verus) social organization. Behavioral Ecology and Sociobiology 56:207-216. Leonard, M., and M. Fenton. 1984. Echolocation calls of Euderma maculatus (Vespertilionidae): use in orientation and communication. Journal of Mammalogy 65: 122-126. Ljungblad, D.K., Scoggins, P.D., and W.G. Gilmartin. 1982. Auditory threshold of captive Eastern Pacific bottle-nosed dolphin, Tursiops spp. Journal of the Acoustic Society of America 72:1726-1729. Lohr, B., Wright, T.F., and R.J. Dooling. 2003. Detection and discrimination of natural calls in masking noise by birds: estimating the active space of a signal. Animal Behaviour 65:763-777. Lusseau, D., K. Schneider, O.J. Boisseau, P. Haase, E. Slooten, and S.M. Dawson. 2003. The bottlenose dolphin community of Doubtful Sound features a large proportion of long-lasting associations Can geographic isolation explain this unique trait? Behavioral Ecology and Sociobiology 54:396-405. Macdonald, D.W., and G.M. Carr 1989. Food security and the rewards of tolerance. Pp. 75-99. In: Comparative socioecology. The behavioural ecology of humans and other mammals. Edited by V. Standenana, and R.A. Foley. Blackwell Scientific Publications, Blackwell. Mann, J. 1997. Individual differences in bottlenose dolphin infants. Family Systems 4:35-49. 176

PAGE 189

Mann, J. 2000. Unraveling the dynamics of social life. Pp. 45-64. In: Cetacean societies. Field studies of dolphins and whales. Edited by J. Mann, R.C. Connor, P.L. Tyack, and H. Whitehead. The Chicago University Press, Chicago. Mann, J., and B.B. Smuts. 1998. Natal attraction: allomaternal care and mother-infant separations in wild bottlenose dolphins. Animal Behaviour 55:1097-1113. Mann, J., R.C. Connor, L.M. Barre, and M.R. Heithaus. 2000. Female reproductive success in bottlenose dolphins (Tursiops sp.): life history, habitat, provisioning, and group-size effects. Behavioral Ecology 2:210-219. Mann, J., and J.J. Watson-Capps. 2005. Surviving at sea: ecological and behavioral predictors of calf mortality in Indian Ocean bottlenose dolphins, Tursiops sp. Animal Behaviour 69:899-909. Marsh, H.W., and M. Schulkin. 1962. Shallow-water transmission. Journal of the Acoustical Society of America 34:863-864. Masters, W.M., and H.E. Harley. 2004. Performance and cognition. Pp. 249-259. In: Echolocating mammals. In: Echolocation in bats and dolphins. Edited by J.A. Thomas, C.F. Moss, and M. Vater. The University of Chicago Press, Chicago. McCowan, B., and D. Reiss. 2005. Maternal aggressive contact vocalizations in captive bottlenose dolphins (Tursiops truncatus): Wide-band, low-frequency signals during mother/aunt-infant interactions. Zoo Biology 14:293-309. McFarland, M. 1987. Ecological and social correlates of party size in the black spider monkey, Ateles paniscus chamek. Ph.D. thesis, Princeton University, New Jersey. McFarland, M. 1990. Fission-fusion social organization in Ateles and Pan. International Journal of Primatology 11:47-61. McGregor, P.K. 1993. Signaling in territorial systems: a context for individual identification, ranging, and eavesdropping. Philosophical Transactions of the Royal Society of London 340:237-244. McGregor, P.K. and T. Dabelsteen. 1996. Communication networks. Pp. 407-425. In: Ecology and evolution of acoustic communication in birds. Edited by D.E. Kroodsma and E.H. Miller. Cornell University Press, New York. Mnard, N., and D. Vallet. 1993. Dynamic of fission in a wild Barbary macaque group (Macaca sylvanus). International Journal of Primatology 14:479-500. Michelsen, A., and O.N. Larsen. 1983. Strategies for acoustic communication in complex environments. Pp. 321-331. In: Neuroethology and behavioral physiology: roots and growing points. Edited by F. Huber and H. Mark. Springer-Verlag, New York. 177

PAGE 190

Miller, E.H. 1996. Acoustic differentiation and speciation in shorebirds. Pp. 241-257. In: Ecology and evolution of acoustic communication in birds. Edited by D.E. Kroodsma and E.H. Miller. Cornell University Press, New York. Nachtigall, P.E. 1980. Odontocete echolocation performance on object size, shape and material. Pp. 71-95. In: Animal sonar systems. Edited by R.G. Rusnel and J.F. Fish. Plenum Press, New York. Neuhaus, P., and K.E. Ruckstuhl. 2002. The link between sexual dimorphism, activity budgets, and group cohesion: the case of the plains zebra (Equus burchelli). Canadian Journal of Zoology 80:1437-1441. Nishida, T. 1968. The social group of wild chimpanzees in the Mahali Mountains. Primates 9:167-224. Norcross, J.L., and J.D. Newman. 1993. Context and gender-specific differences in the acoustic structure of common marmoset (Callithrix jacchus) phee calls. American Journal of Primatology 30:37-54. Norris, K.S., and T.P. Dohl. 1980. Behavior of the Hawaiian spinner dolphin, Stenella longirostris. Fisheries Bulletin 77:821-849. Norris, K.S., B. Wrsig, R.S. Wells, and M. Wrsig. 1994. The Hawaiian Spinner Dolphin. University of California Press, California. Nowacek, D.P. 1999. Sound use, sequential behavior and ecology of foraging bottlenose dolphins, Tursiops truncatus. Ph.D. dissertation, MIT/WHOI Joint Program. 196 pp. Nowacek, D.P., K.C. Buckstaff, M.P. Johnson, and R.S Wells. 2001. Transmission loss of vessel noise in manatee environments. Florida Fish and Wildlife Conservation Commission Contract 99023, Final Report, pp. 25. Nowacek, S. M., R. S. Wells, and A.R. Solow. 2001. Short-term effects of boat traffic on bottlenose dolphins, Tursiops truncatus, in Sarasota Bay, Florida. Marine Mammal Science 17:673-688. Nowacek, D.P. 2005. Acoustic ecology of foraging bottlenose dolphins (Tursiops truncatus), habitat-specific use of three sound types. Marine Mammal Science 21:587-602. OBrien, T.G. 1991. Female-male social interactions in wedge-capped capuchin monkeys: benefits and costs of group living. Animal Behaviour 41:555-567. 178

PAGE 191

Oda, R. 1996. Effects of contextual and social variables on contact call production in free-ranging ringtailed lemurs (Lemur catta). International Journal of Primatology 17:191-205. Ott, R.L. 1994. An introduction to statistical methods and data analysis. Duxbury Press, an imprint of Wadsworth Publishing Company, California. Owen, C.F.W. 2001. A comparison of maternal care of primiparous and multiparous bottlenose dolphins (Tursiops truncatus) Does parenting improve with experience? M.Sc. thesis. University of California, Santa Cruz. Owen, E.C.G., R.S. Wells, and S. Hoffman. 2002. Ranging and association patterns of paired and unpaired adult male bottlenose dolphins, Tursiops truncatus, in Sarasota, Florida provide no evidence for alternative male strategies. Canadian Journal of Zoology 80:2072-2089. Packer, C., D. Scheel, and A.E. Pusey. 1990. Why lions form groups: food is not enough. American Naturalist 136:1-19. Palombit, R.A., R.M. Seyfarth, and D.L. Cheney. 1997. The adaptative value of 'friendships' to female baboons: experimental and observational evidence. Animal Behaviour 54:599-614. Parijs, S. M., J. Smith, and P. J. Corkeron. 2002. Using calls to estimate the abundance of inshore dolphins: a case study with Pacific humpback dolphins Sousa chinensis. Journal of Applied Ecology 39:853-864. Peres, C.A. 1996. Use of space, spatial structure and foraging group size of Gray Wooly Monkeys (Lagothrix lagotricha cana) at Urucu, Brazil: a review of the Atelinae. Pp. 467-488. In: Adaptive radiation of Neotropical primates. Edited by M. Norconk, A.L. Rosenberg, and P.A. Garber. Plenum Press, New York. Pfalzer, G., and J. Kusch. 2003. Structure and variability of bat social calls: implications for specificity and individual recognition. Journal of Zoology 261:21-33. Poole, J.H., K. Payne, W.R. Langbauer Jr., and C.J. Moss. 1988. The social context of some very low frequency calls of African elephants. Behavioral Ecology and Sociobiology 22:385-392. Pulliam, H.R. and T. Caraco. 1984. Living in groups: is there an optimal group size? Pp. 122-147. In: Behavioral ecology, an evolutionary approach. Edited by J.R. Krebs and N.B. Davies. Blackwell Scientific Publications, Massachusetts. Quintana-Rizzo, E., and R.S. Wells. 2001. Resighting and association patterns of bottlenose dolphins (Tursiops truncatus) in the Cedar Keys, Florida: insights into social organization. Canadian Journal of Zoology 79:447-456. 179

PAGE 192

Quintana-Rizzo, E., D.A. Mann, and R.S. Wells. 2006. Estimated communication range of social sounds used by bottlenose dolphins (Tursiops truncatus). Journal of the Acoustical Society of America 120:1671-1683. Rogers, P.H., and C. Cox. 1998. Underwater sound as a biological stimulus. Pp. 131-149. In: Biology of aquatic animals. Edited by J. Atema, R. Fay, A.N. Popper, and W.N. Tavolga. Springer-Verlag, Berlin. Rossbach, K.A., and D.L. Herzing. 1997. Underwater observations of benthic-feeding bottlenose dolphins (Tursiops truncatus) near Grand Bahama Island, Bahamas. Marine Mammal Science 13:498-504. Rubenstein, D.I. 1986. Ecology and sociobiology in horses and zebras. Pp. 282-302. In: Ecological aspects of social evolution. Edited by D.I. Rubenstein and R.W. Wrangham. Princenton University Press, New Jersey. Rubenstein, D.I. 1994. The ecology of female social behaviour in horses, zebras, and asses. Pp. 13-28. In: Animal societies: individuals, interactions, and organization. Edited by P.J. Jarman and A. Rossiter. Published by Physiology and Ecology, Japan. Ruckstuhl, K., and P. Neuhaus. 2001. Behavioral synchrony in Ibex groups: effects of age, sex and habitat. Behaviour 138:1033-1046. Sayigh, L.S., P.L. Tyack, and R.S. Wells. 1993. Recording underwater sounds of free-ranging dolphins while underway in a small boat. Marine Mammal Science 9:213-220. Sayigh, L.S., P.L. Tyack, R.S. Wells, and M.D. Scott. 1990. Signature whistles of free-ranging bottlenose dolphins Tursiops truncatus: stability and mother-offspring comparisons. Behavioral Ecology and Sociobiology 2:247-260. Schnitzler, H.U., C.F. Moss, and A. Denzinger. 2003. From spatial orientation to food acquisition in echolocating bats. Trends in ecology and evolution 18:386-394. Scott, M.D, R.S. Wells, and A.B. Irvine. 1990. A long-term study of bottlenose dolphins on the west coast of Florida. Pp. 235-244. In: The bottlenose dolphin. Edited by S. Leatherwood and R.R. Reeves. Academic Press, California. Shane, S.H. 1990. Behavior and ecology of the bottlenose dolphin at Sanibel Island, Florida. Pp. 245-265. In: The bottlenose dolphin. Edited by S. Leatherwood and R.R. Reeves. Academic Press, California. Slabbekoorn, H. 2004. Habitat-dependent ambient noise: consistent spectral profiles in two African forest types. Journal of the Acoustic Society of America 116:3727-3733. 180

PAGE 193

Smolker, R.A., A.F. Richards, R.C. Connor and J.W. Pepper. 1992. Sex differences in patterns of association among Indian Ocean bottlenose dolphins. Behaviour 123:38-69. Smolker, R.A. and J.W. Pepper. 1999. Whistle convergence among allied male bottlenose dolphins (Delphinidae, Tursiops sp.). Ethology 105:595-617. Smolker, R.A., J. Mann, and B.B. Smuts. 1993. Use of signature whistles during separations and reunions by wild bottlenose dolphin mothers and infants. Behavioral Ecology and Sociobiology 33:393-402. Southwell, C.J. 1984. Variability in grouping in the Eastern grey kangaroo, Macropus giganteus II. Dynamics of group formation. Australian Wildlife Research 11:437 449. Sugiyama, Y. 1979. Social structure and dynamics of wild chimpanzees at Bossou, Guinea. Primates 20:323-339. Tobias, P.A., and D.C. Trindade. 1995. Applied reliability. CRC Press, Boca Raton, Florida. Thomsen, F., D. Franck, and J.K.B. Ford. 2002. On the communicative significance of whistles in wild killer whales (Orcinus orca). Naturwissenschaften 89:404-407. Treves, A. and C.A. Chapman. 1996. Conspecific threat, predation avoidance, and resource defense: implications for grouping in langurs. Behavioral Ecology and Sociobiology 39:43-53. Trivers, R.L. 1972. Parental investment and sexual selection. Pp. 136-179 in: Sexual selection and the descent of man. Edited by B. Campbell. Aldine, Chicago. Tyack, P.L. 2000a. Functional aspects of cetacean communication. Pp. 270-308. In: Cetacean societies. Field studies of dolphins and whales. Edited by J. Mann, R.C. Connor, P.L. Tyack, and H. Whitehead. The University of Chicago Press, Chicago. Tyack, P.L. 2000b. Dolphin whistle a signature tune. Science 289:1310-1311. Urian, K.W., and R.S. Wells. 1996. Bottlenose dolphin photo-identification workshop: March 21-22, 1996, Charleston, South Carolina. NOAA Technical Memorandum NMFS-SEFSC-393. 91 pp. Urick, R.J. 1975. Principles of underwater sound. Mc-Graw-Hill, New York. van Schaik, C.P. 1989. The ecology of social relationships among female primates. Pp.195-218. In: Comparative Socioecology, the behavioral ecology of humans and other mammals. Edited by V. Standen and R. Foley. British Ecological Society, Oxford. 181

PAGE 194

van Schaik, C.P. and P.M. Kappeler. 1997. Infanticide risk and the evolution of male-female association in primates. Primate Social Evolution 264:1687-1694. Waples, D. 1995. Activity budgets of free-ranging bottlenose dolphins (Tursiops truncatus) in Sarasota Bay, Florida. M.S. thesis, University of California, Santa Cruz, California. Waser, P.M., and C.H. Brown. 1986. Habitat acoustic and primate communication. American Journal of Primatology 10:135-154. Watwood, S.L. 2003. Whistle use and whistle sharing by allied male bottlenose dolphins, Tursiops truncatus. Ph.D dissertation, Massachussets Institute of Technology and Woods Hole Oceanographic Institute. Weihs, D. 2004. The hydrodynamics of dolphin drafting. Journal of Biology 3:8. Wells, R.S. 1991. The role of long-term study in understanding the social structure of a bottlenose dolphin community. Pp. 199-225. In: Dolphins societies: discoveries and puzzles. Edited by K. Pryor and K.S. Norris. University of California Press, California. Wells, R.S. 2003. Dolphin social complexity: Lessons from long-term study and life history. Pp. 32-56. In: Animal Social Complexity: Intelligence, Culture, and Individualized Societies. Edited by F.B.M. de Waal and P.L. Tyack. Harvard University Press, Massachusetts. Wells, R.S., A.B. Irvine, and M.D. Scott. 1980. The social ecology of inshore dontocetes. Pp. 263-317 In: Cetacean behavior: mechanisms and functions. Edited by L.M. Herman. J. Wiley & Sons, New York. Wells, R.S., M.D. Scott, and A.B. Irvine. 1987. The social structure of free-ranging bottlenose dolphins. Pp. 407-416. In: Current Mammalogy, Vol. I. Edited by H. H. Genoways. Plenum Press, New York. Wells, R.S., and M.D. Scott. 1990. Estimating bottlenose dolphin population parameters from individual identification and capture-release techniques. Pp. 407-415. In: Individual recognition of cetaceans: use of photo-identification and other techniques to estimate population parameters. Edited by P. S. Hammond, S. A. Mizroch and G.P. Donovan. Reports of the International Whaling Commission. Special Issue 12. Cambridge. 440 pp. Wells, R.S., K.W. Urian, A.J. Read, K. Bassos, W.J. Carr, and M.D Scott. 1996. Low-level monitoring of bottlenose dolphins, Tursiops truncatus, in Tampa Bay, Florida: 1988-1993. U.S. Department of Commerce, NOAA Technical Memorandum NMFS-SEFSC-385. 25 pp. + 6 Tables, 8 Figures, and 4 Appendices. 182

PAGE 195

Wells, R.S., D.J. Boness and G.B. Rathbun. 1999. Behavior. Pp. 324-422. In: Biology of Marine Mammals. Edited by J.E. Reynolds, III and S.A. Rommel. Smithsonian Institution Press, Washington, DC. 567 pp. White, F.J. 1988. Party composition and dynamics in Pan paniscus. International Journal of Primatology 9:179-193. White, F.J. 1989. Ecological correlates of pygmy chimpanzee social structure. Pp. 151-164. In: Comparative socioecology. The behavioural ecology of humans and other mammals. Edited by V. Standenana, and R.A. Foley. Blackwell Scientific Publications. Oxford, Blackwell. White, F.J. 1992. Pygmy chimpanzee social organization: variation with party size and between study sites. American Journal of Primatology 26:203-214. Wiley, R.H., and D.G. Richards. 1978. Physical constraints on acoustic communication in the atmosphere: Implications for the evolution of animal vocalizations. Behavioral Ecology and Sociobiology 3:69-94. Wilson, B. 1995. The ecology of bottlenose dolphins in the Moray Firth, Scotland: A population at the northern extreme of the species range. Ph.D. thesis, University of Aberdeen, Scotland. Wrangham, R.W. 1980. An ecological model of female-bonded primate groups. Behaviour 75:262-300. Wrangham, R.W. 1986. Evolution of social structure. Pp. 282-298. In: Primate societies. Edited by B.B. Smuts, D.L. Cheney, R.M. Seyfarth, R.W. Wrangham, and T.T. Struhsaker. The University of Chicago Press, Chicago. Wrangham, R.W., R. Gittleman, and C.A. Chapman. 1993. Constraints on group size in primates and carnivores: population density and day-range as assays of exploitation competition. Behavioral Ecology and Sociobiology 32:199-210. Wund, M.A. 2005. Learning and the development of habitat-specific bat echolocation. Animal Behaviour 70:441-450. Xitco, M.J. Jr., and H.L. Roitblat. 1996. Object recognition through eavesdropping: Passive echolocation in bottlenose dolphins. Animal Learning and Behavior 24:355-365. 183

PAGE 196

ABOUT THE AUTHOR Ester Quintana-Rizzo was born in Guatemala City. Since she was young, she has been very interested in learning about marine mammals. For that reason, she enrolled in the Biology School at the Universidad de San Carlos in Guatemala where she received a Licenciatura degree in Biology in 1993. In 1995, she came to the United States with a LASPAU-Fulbright Scholarship to obtain a Masters degree in Zoology from the University of Florida. In 2000, she entered the Ph.D. program at the University of South Florida (USF) in the College of Marine Science. While in the Ph.D. program, she received fellowships from the College of Marine Science including the Garrels Fellowship, Jack Lake Fellowship and the Gulf Oceanographic Charitable Trust Fellowship and from the USF Library (the Samuel Y. Fustukjian Memorial Scholarship). Ester is currently working at the Marine Mammal and Sea Turtle Research Center of Mote Marine Laboratory. 184