SOCIAL ORGANISATION OF THE SHORT-FINNED PILOT WHALE, 
Globicephala macrorhynchus, WITH SPECIAL REFERENCE TO THE 
COMPARATIVE SOCIAL ECOLOGY OF DELPHINIDS 
by 
James Robert Heimlich-Boran 
Clare Hall 
Dissertation submitted to the University of Cambridge 
in application for the degree of 
Doctor of Philosophy 
Department of Zoology 
Downing Street, Cambridge 
January 1993 
DOCTORAL THESIS SUMMARY 
TITLE: Social Organisation of the Short-Finned Pilot Whale, Globicephala 
macrorhynchus, With Special Reference to the Social Ecology of 
Delphinids 
NAME: James R. Heimlich-Boran (Clare Hall, Department of Zoology) 
As a· contribution to the understanding of comparative social trends 
within the cetacean family Delphinidae, a 22-month study was conducted on 
the short-finned pilot whale, Globicephala macrorhynchus, which has been 
suggested to have a unique social system in which males and females in the 
same group are related and mating occurs outside of the group. The individual 
identification of 495 pilot whales, analysed in daily group association patterns, 
allowed ic;lentification of 46 pods. They were classified as productive or non-
productive based on the presence or absence of immature animals. Productive 
pods were significantly larger, although 12% of them lacked adult males. Two 
classes of whales (residents and visitors) were defined by patterns of 
occurrence, suggesting differential patterns of habitat use. Resident pods 
occasionally travelled together ( 41 % of all groups) and associations between age 
and sex classes showed that in I?ixed-pod groups, the highest ranked 
associations of the reproductive females were with males from other pods, while 
within pods, adult males and females associated less. During summer, the 
proposed peak conception period, pilot whale groups were significantly larger 
and contained individuals from a significantly greater number of pods. These 
fihdings support the hypothesis that males and female.s mate when associating 
with individuals from other pods. A comparative analysis of sexual 
dimorphism, brain size, and testes size, habitat, prey and group size within the 
17 delphinid genera identified a correlation between sexual dimorphism and 
body size, but relative measures of brain size and testes size did not correlate 
with broad ecological or social classifications. However, a comparison of three 
delphinid societies identified two distinct male· mating systems: males of the 
small, mono-morphic Tursiops truncatus live in age/sex segregated groups and 
mate with a number of discrete female communities. Males in the large 
sexually dimorphic Globicephala spp. and Orcinus area mate with associated 
female pods and yet remain with their female kin. This corresponds to the 
avunculate social system described in some human societies. It could evolve 
from a promiscuous mating system where there is little guarantee of paternity 
. and where males that live with their kin increase their inclusive fitness. 
Frontspiece. A family of short-finned pilot whales offTenerife. 
PREFACE 
This thesis is the result of my own research, and contains no work 
done in collaboration, except where otherwise stated. The text does not 
exceed 80,000 words. No part of this thesis has been submitted to any other 
University in application for a higher degree. 
SUMMARY 
The initial part of an examination of comparative trends in social 
ecology and mating systems within the cetacean family Delphinidae 
involved a 22-month field project conducted on the social organisation of the 
short-finned pilot whale, Globicephala macrorhynchus. A social system, in 
which related males and females live in the same social group and mating 
occurs with associated groups, has been proposed for this genus. The 
individual identification of 495 pilot whales, analysed in daily group 
association patterns, allowed identification of 46 pods, classified as 
productive or non-productive based on the presence of immature animals. 
Productive pods were significantly larger, although 12% of them were 
lacking adult males . Two classes of whales (residents and visitors) were 
defined by patterns of occurrence, suggesting differential patterns of 
habitat use. Resident pods occasionally travelled together (41 % of all 
groups) and associations between age and sex classes showed that in mixed-
pod groups, the highest ranked associations of the reproductive females 
were with males from other pods, while within pods, adult males and 
females associated less. During summer, the proposed peak conception 
period, pilot whale groups were significantly larger and contained 
individuals from a significantly greater number of pods . These findings 
support the hypothesis that males and females mate when associating with 
other pods. A comparative analysis of morphological trends within 17 
delphinid genera identified a correlation between sexual dimorphism and 
body size, but relative measures of brain size and testes size did not 
correlate with ecological or social classifications. A comparison of three 
delphinid societies identified two distinct male mating systems: males of 
the small, mono-morphic Tursiops truncatus live in age/sex segregated 
groups and mate with a number of discrete femal~ communities. Males in 
the large sexually dimorphic Globicephala spp. and Orcinus area mate with 
associated female pods and yet remain with their female kin. A human 
social system, an avunculate, is proposed to match this system and could 
evolve within a promiscuous mating system with little guarantee of 
paternity, where inclusive fitness benefits males that live with known kin. 
ACKNOWLEDGMENTS 
This work was assisted by many people. First of all, I must thank 
my wife, Sara Heimlich-Boran, who helped throughout the project, both in 
the field and at home. Second, I thank my supervisor, Dr. John Harwood, 
for his contributions and patience. In Tenerife, I thank my friends and 
assistants, Sergio Escorza, Roberto Montero, Vidal Martin, and Robert 
Vonk. This project was supported by funding from Earthwatch, the 
American Museum of Natural History, the Whale and Dolphin 
Conservation Society, Cetacean Society International, and the American 
Cetacean Society. I also thank Dr. Tim Clutton-Brock and everyone from 
the Large Animal Research Group who gave me precious space in their lab 
and stimulating discussions over lunch. 
TABLE OF CONTENTS 
CHAPTER 1: INTRODUCTION... .. ..... .. ........ ... .. .... ............ ...... ... ... ....... .... .. 1 
1.2 THEORETICAL BACKGROUND ... .. .. .... ... ..... .. ............ ... ......... .. ............. 2 
1.2 AIMS OF THIS STUDY .................. : .......... ....... .. ............... ..... .. ...... ........... 9 
1.3 SUMMARY ......... ... ... .. ....... ............ ... ...... ................................................... 11 
CHAPTER 2: REVIEW OF THE LIFE HISTORY AND 
BEHAVIOUR OF PILOT WHALES (Globicephala spp.) ............ 12 
2.1 INTRODUCTION ............................................................. ......................... 12 
2.2 EXTERNAL APPEARANCE ............................................... ... ................... 13 
2.3 TAXONOMY ............. ......................................................... .......... .. ...... ..... . 15 
2.4 GENETICS .................. ..... ........ .. ....................................... ...... ........ ..... ..... . 18 
2.5 LIFE HISTORY PARAMETERS ..... .. ............................ ....... ... ................. 23 
2.5.1 Body Size and Growth ................................................................. 24 
2.5.2 Reproduction ........................... : .................................................... 27 
2.5.2.1 Male Sexual Maturity .................................................... 28 
2.5.2.2 Female Reproductive Cycle ........................................... 29 
2.5.2.4 Seasonality of Reproduction .... ...................................... 35 
2.5.3 Duration of Lactation and Weaning ........................................... 36 
2.5.4 Longevity, Mortality and the Sex Ratio ..................................... .40 
2.5.5 Summary of Short-Finned Pilot Whale Life History ................ .43 
2.6 ECOLOGY ........... .... .... .. .......... .............................................. .... ................. 44 
2.6.1 Distribution ...................... .......... ........... .... .. ......... ......... ..... ... ........ 44 
2.6.2 Feeding Habits ............. .. .... .................................... .......... .... ........ 50 
2.7 SOCIAL ORGANISATION ......................................... .. ...................... ...... 53 
2.8 DISCUSSION ............................................................................................ 55 
2.9 SUMMARY ................................................................................................ 58 
CHAPTER 3: DISTRIBUTION OF SHORT-FINNED PILOT 
WHALES OFF TENERIFE, CANARY ISLANDS, SPAIN 
3.1 INTRODUCTION ...................................................................................... 59 
3.2 METHODS ................................................................................................. 60 
3.2.1 Survey Methods ......... ..... .. .. ...... .. .... ... .. ..... ..... ............ ... ... ..... ....... 60 
3.2.2 Seasonal Occurrence ............ ....... .. ........ ....... ... ..... .. ... ................... 60 
3.2.3 Mapping ........................................................... ..... .. ..... .......... .. ..... 61 
3.2.4 Description of the study area .. .......... ... ... .. ....... ........................... 61 
3.3 RESULTS .. ...... .. .. ............ ............... ........... .. ........ .. ... .......... ........ .... .. .......... 63 
3.3.1 Seasonal Occurrence ............................. , ... ............. ..... ... ............... 63 
3.3.2 Geographic Distribution ......... ................... ......... ......................... 65 
3.4 DISCUSSION ...................... ...................................................................... 66 
3.5 SUMMARY ............................... , ... ................................................... .......... 68 
CHAPTER 4: SOCIAL ORGANISATION OF SHORT-FINNED 
PILOT WHALES ( Globicephala macrorhynchus) 
4.1 INTRODUCTION.................. ............... ..... ..... ... .... ......... ............ .. ...... .. .... . 69 
4.2 METHODS............ ........... ... ..... .. ..... ............. ...... ... ....... ...... ...... .. ....... ......... 71 
4.2.1 Photographic Sampling Methods ..... ... ...... ..... ...... ... ........... ... ...... 71 
4.2.2 Photographic Analysis ........ .... .......... .. ...... ... .............. .. .. .... ..... ..... 72 
4.2.3 ID Characteristics ............. ;... ............... ........................................ 73 
4.2.4 Unidentifiable Individuals ............... ........................................... 75 
4.2.5 Age/Sex Classification ............... .... ......... .. .. .. .... ... .... .. .... ..... ...... ... 76 
4.2.6 Group Definition.......................................................................... 80 
4.2.7 Pair Associations .......................... .. .............................................. 81 
4.2.8 Age/Sex Class Associations .. . . . . ... .. . .. . . .. . . . . . . . . .. . . .. . ... ... . . . ... . .. . . . . . . . . . 85 
4.2.9 Cluster Analysis............................................................... .. .......... 86 
4.2.10 Pod Definition ..................... ............................ ........................... 87 
4.2.11 Pod Associations......................... ........... .. .... ........... ..... .... ..... ..... 88 
4.3 RESULTS............................................................................. .......... ............ 89 
4.3.1 Identification and Classification of Individual Whales ............. 89 
4.3.2 Patterns of Occurrence: Residents and Visitors......................... 91 
4.3.3 Group Size........... .............. ..... ... ................................................... 93 
4.3.3 Association Patterns and the Definition of Pods ....................... 95 
4.3.4 Pod Composition . . . .. ..... ... ...... .. .. ......... ... .. . . .. .. .. ... ..... .. . . . .... . .. . . .... . . . 96 
4.3.5 Variability in Associations with Age/Sex and Pod 
Membership ............... ..................... .. ... ........................................ ........... 98 
4.3.6 Associations Between Pods .......................... .. ........ .................... .. 100 
4.4 DISCUSSION .......................................... .................................................. 102 
4.4.1 Pilot Whale Identification and Classification ............................. 102 
4.4.2 Residents and Visitors ................................................................. 105 
4.4.3 Reliability of Pod Definitions ........ .............. ................................ 107 
4.4.4 Seasonality of Occurrence and Association ................................ 108 
4.4.5 Age and Sex Structure of Pilot Whale Pods ................................ 110 
4.4.6 Conclusions ................................................................................... 112 
4.5 SUMMARY ................................................... .... ..................... ... .... ... ...... .... 113 
CHAPTER 5: COMPARATNE SOCIAL ECOLOGY OF 
DELPHINIDS: TOWARDS AN UNDERSTANDING OF 
MATING SYSTEMS 
5.1 INTRODUCTION ...................................................................................... 117 
5.2 METHODS ............. ..... ........ ... .............................. ...................................... 124 
5.3 RESULTS ................................................................................................... 127 
5.3.1 Ecological Classification .............................................................. 127 
5.3.2 Sexual Dimorphism .. .... ..................... .. ........................................ 127 
5.3.3 Relative Brain Size .................... ................... .. ..... ............. ........... 129 
5.3.4 Relative Testes Size .............................. .... .. .................... ...... .... ... 130 
5.3.6 Delphininae Social Systems ................ : ....................................... 131 
5.3.6.1 Tursiops ...................... .. .................................................. 131 
5.3.6.2 Stenella ...... .. .. , ............................................ .... ............ .... 137 
5.3.6.3 Lagenorhynchus ........ ..................... .... ....... ... .... .............. 141 
5.3.6.4 Grampus ........................................... .. .. .. : .... .............. ..... 143 
5.3.6.5 Summary ofDelphininae Social Systems .................... . 144 
5.3.7 Globicephalinae Social Systems ..... .......................... ................... 145 
5.3.7.1 Orcinus ................................................... ..... .. ... .. ............ 146 
5.3.7.2 Globicephala .... ............................................................... 155 
5.3. 7 .3 Summary of Globicephalinae Social Systems .............. 156 
5.4 DISCUSSION ............................................................................................ 157 
5.4.1 Patterns of Body Size and Sexual Dimorphism .. ............ ........... 159 
5.4.2 Brain size, Ecology, and Social Behaviour ................................. 168 
5.4.3 Relative Testes Size and the Potential for Sperm 
Competition in Delphinids ............... ....................... .............. ................ 171 
5.4.4 Patterns of Dispersal and Genetic Diversity .............................. 172 
5.4.5 Hypothetical Delphinid Mating Systems ................................... 173 
5.5 SUMMARY ................................................................................................ 188 
CHAPTER 6: CONCLUDING DISCUSSION ............................................ 192 
REFERENCES ................................................................................................ 196 
CHAPTER 1: INTRODUCTION 
What are the benefits to group living? Societies, like physical 
characteristics, have evolved through adaptation to a variety of 
environmental selective pressures (Crook and Gartlan, 1966; Crook, 1970; 
Kummer, 1971; Alexander, 1974; Jarman, 1974; Wilson, 1975; Crook et al., 
1976; Emlen and Oring, 1977; Jarman and Jarman, 1979; Eisenberg, 1981; 
Wrangham and Rubenstein, 1986). They are also dependent on the 
distribution and degree of interaction with close kin (Hamilton, 1964; West 
Eberhard, 1975; Kurland, 1977; Bertram, 1978; Michener, 1982; Packer and 
Pusey, 1982; Gouzoules, 1984; Sherman and Holmes, 1985; Trivers, 1985; 
Gouzoules and Gouzoules, 1987; Wade and Breden, 1987; Gittleman, 1989; 
Packer et al., 1991). Indeed, kin selection, through the process of inclusive 
fitness (Hamilton, 1964), has been recognised as an important factor in the 
evolution of such basic societal elements as cooperation and altruism 
(Brown, 1978; Axelrod and Hamilton, 1981; Trivers, 1985; Moehlman, 1986; 
Moore, 1992), although certain forms of these elements also can occur with 
non-kin (Maynard Smith, 1979; Maynard Smith, 1982; Packer et al., 1991; 
Mesterton-Gibbons and Dugatkin, 1992). 
In most mammal species, however, it has been noted that one sex or 
the other disperses from their natal group, apparently to avoid inbreeding 
with close relatives and to find better feeding and mating opportunities 
elsewhere (Ralls et al., 1979; Greenwood, 1980; ·Ralls et al., 1980; Shields, 
1982; Greenwood, 1983; Moore and Ali, 1983; Shields, 1983; Harvey and 
Ralls, 1986; Pusey and Packer, 1987a). In mammals, males are the most 
common sex to disperse (Greenwood, 1980). Species in which females are 
the emigrating sex have been shown to be those in which their age at first 
1 
breeding exceeds the average male residence time (Clutton-Brock, 1989); 
supporting the idea that dispersal is primarily a mechanism to avoid 
inbreeding. 
The alternative to dispersal is natal philopatry, and it has been 
argued that the benefits of philopatry (even with certain degrees of 
inbreeding) and its occurrence in mammals has been underestimated 
(Bengtsson, 1978; Waser and Jones, 1983). In fact, there is some indication 
of the natal philopatry of both sexes in only a few species, for example, grey 
squirrels, blue duiker, brown hyenas and possibly Barbary macaques 
(Waser and Jones, 1983; Moore, 1992). Why is this so rare, and is there a 
direct, causal link between dispersal and the avoidance of inbreeding 
(Moore and Ali, 1983)? 
I review evidence and present original observations on two 
mammals, the pilot whale, Globicephala sp., and the killer whale, Orcinus 
orca, both members of the order Cetacea, family Delphinidae, subfamily 
Globicephalinae, which appear to have very low (or possibly absent) levels of 
dispersal of either sex and resultant high degrees of relatedness between 
males and females within the same cohesive social group. A critical 
missing element in the understanding of cetacean societies is a description 
of the mating system. In order to develop hypotheses concerning delphinid 
mating systems, I also examine comparative trends in morphology, 
genetics and social behaviour within this diverse family. 
1.2 THEORETICAL BACKGROUND 
Ultimately, natural selection has operated on the costs and benefits to 
individuals of group living (Alexander, 1974). Social groupings occur when 
individual strategies of reproduction, resource acquisition and protection 
2 
from predation are enhanced above and beyond the inevitable competition 
for mates and resources inherent in group living. The costs and benefits of 
these strategies are strongly affected by degrees of relatedness between 
individuals, resulting in kin selection through the process of inclusive 
fitness (Hamilton, 1964). In this light, much progress has been made in the 
study of social organisation by examining interactions between kin and 
non-kin utilising the methods of individual identification in long-term field 
studies of wild populations of ungulates (Clutton-Brock et al., 1982a), 
carnivores (Schaller, 1972; Packer, 1986; Packer et al., 1988), elephants 
(Moss and Poole, 1983; Moss, 1988) and primates (Goodall, 1986; 
Stammbach, 1987). The strong social bonds in these groups are constantly 
tempered by competitive interactions for limited resources or mating 
opportunities. There are still many questions as to how these conflicts are 
mediated and what the resulting effects are on group stability. 
A second fruitful method for understanding the diversity of social 
systems has been comparative studies, between related species in different 
environments or between different species in similar environments, in 
order . to understand the selective pressures which are acting on the 
evolution of mammalian societies (Crook and Gartlan, 1966; Jarman, 1974; 
Clutton-Brock and Harvey, 1977; Harvey et al., 1980; Harcourt et al., 1981; 
Harvey and Bennett, 1983; MacDonald, 1983; Harvey and Harcourt, 1984; 
Schmitz and Lavigne, 1984; Harvey and Clutton-Brock, 1985; Gittleman, 
1986; Packer, 1986; Wooton, 1987; Clutton-Brock, 1988). These comparative 
studies have identified a variety of trends within and between mammalian 
orders, and have gone far towards explaining the evolution of life history 
and behaviour patterns. However, many groups of mammals are under-
represented in these comparative analyses because of the difficulties 
involved in collecting long-term observations. This is true for many 
3 
nocturnal, arboreal or burrowing mammals, but also for the order Cetacea, 
the whales and dolphins. 
Cetaceans are the only permanently aquatic mammals, even 
pinnipeds (seals and sea lions) must . come to land for breeding. The order 
contains species with the largest body size (the blue whale), the largest 
brain (the sperm whale), and the longest migration (the grey whale) of any 
mammal. They evolved from mesochinid condylarthan mammals in the 
middle Miocene, 60 million years ago (Gingerich et al., 1983; Barnes, 1990) 
and colonised a completely different environment from that of terrestrial 
mammals. Although their closest terrestrial ancestors are the ungulates 
(Milinkovitch, 1991), they perhaps parallel the chiropterans in the degree of 
specialisation required for the colonisation of a three-dimensional medium 
(Norris and Schilt, 1988). Because they live in such a different 
environment, the study of cetaceans can provide a useful addition to an 
understanding of the ways in which habitat affects mammalian social 
systems (Wursig, 1989). 
The cetacean family Delphinidae 1s especially suitable for 
comparative examination. The delphinid cetaceans exhibit a wide 
variability in body size, brain size, sexual dimorphism and social 
dynamics, from the 40 kg Rector's dolphin to the 4000 kg killer whale. Two 
recent long-term studies (each conducted for 20 years) have shown that 
dolphins exhibit a variety of group structures which offer opportunities for 
examining the determinants of fluctuations in group stability. Studies of 
the killer whale, Orcinus orca, the largest dolphin, have shown that this 
sexually-dimorphic animal travels in extremely stable, multi-male, female-
biased groups for . which no immigration or emigration has been 
documented (J.Heimlich-Boran, 1986a; J.Heimlich-Boran, 1986b; S . 
Heimlich-Boran, 1986; Hoelzel and Osborne, 1986; Osborne, 1986; Bigg et al., 
4 
1987; J.Heimlich-Boran, 1988; S.Heimlich-Boran, 1988; Ford, 1989; Bigg et 
al., 1990; Olesiuk et al., 1990; Felleman et al., 1991; Baird, 1992). Males 
appear to grow to maturity and remain in their mother's social group 
throughout adulthood (S.Heimlich-Boran, 1986; 1988; Bigg et al., 1990). In 
contrast, the smaller, relatively monomorphic bottlenose dolphin (Tursiops 
truncatus) lives in age and sex segregated groups and appears to have a 
more fluid social organisation, with frequent mixing between groups 
(Wiirsig and Wiirsig, 1977; Wiirsig, 1978; Wiirsig and Wiirsig, 1979; Shane 
et al., 1986; Wells, 1986; Wells et al., 1987; Ballance, 1990; Wells, 1991; 
Connor et al., 1992a; Connor et al., 1992b). These variations in group 
structure and stability within a single mammalian family offer an 
excellent opportunity for comparative study to understand the causes of 
sociality. 
Additional information on another large, sexually-dimorphic 
delphinid, the pilot whale (Globicephala sp.), has come from analyses of 
carcasses collected in conjunction with whaling operations. This whaling 
is done by driving entire social groups into restricted bays, and thus has 
provided some information on group composition. The primary anomalous 
feature of pilot whale social organisation is based on genetic research on 
long-finned pilot whales, G. melas, killed by Faeroese whalers, which has 
shown that group members, including the adult males, have a high degree 
of relatedness. (Amos et al., 1991a; Amos et al., 1991b; Amos et al ., In 
press). This finding, similar to that described for Orcinus, implies a lack of 
dispersal and resultant strong degree of natal · philopatry for both sexes. 
The prevalence of dispersal in most mammal and bird species has primaily 
been explained by the assumed necessary taboo on mating with close 
relatives in order to avoid inbreeding (Packer, 1979; Greenwood, 1980; 
Pusey, 1980; Shields, 1982; Greenwood, 1983; Moore and Ali, 1983; Shields, 
5 
1983; Pusey and Packer, 1987b; Pusey and Packer, 1987a). Inbreeding is 
considered to have negative effects through an increase in homozygosity 
and the expression of deleterious, recessive genes (Ralls et al., 1979; Ralls et 
al., 1980). However, inbreeding can also be shown to have some positive 
effects (see also Shields, 1982; Moore and Ali, 1983; Shields, 1983). If a 
population has been inbred for many generations then the deleterious 
recessive genes should have been removed from the population through 
natural selection. The increased homozygosity may be an advantage in 
terms of kin selection because "inbred siblings are more closely related 
than outbred ones, and hence more able to benefit from kin selection" (p. 95: 
Moore and Ali 1983). Thus, inbreeding can promote sociality and altruism. 
Another positive effect of inbreeding could be "the establishment or 
maintenance of locality-specific co-adapted genomes" (Moore and Ali 1983; 
p. 96). Shields (1982, 1983) has even concluded that inbreeding may 
overcome the costs of sex (sensu: Williams, 1975; Maynard Smith, 1978) by 
reducing the rate of break-up of co-adapted parental genomes through 
meiosis. In light of the real costs of dispersal (e.g. increased mortality due 
to unknown resource distribution, predator concentrations, and conspecific 
inter- and intra-group competitive hierarchies), there may be a situation in 
which inbreeding is more beneficial than dispersing (Bengtsson, 1978). 
Shields (1982, 1983) developed the idea of "optimal inbreeding" in order to 
account for the necessity of balancing the costs of dispersing and the costs of 
inbreeding. Of course, there is no clear point at which a population 
becomes inbred; although definitions have been attempted such that 
"intense" inbreeding occurs when the breeding population is less than 100, 
or when partners are related closer than r=0.125, or full first cousins 
(Shields 1982). 
6 
The occurrence of male natal philopatry, as suggested for killer 
whales and pilot whales, both sexually-dimorphic species living in multi-
male groups, suggests a unique mating system (Greenwood, 1983; Shields, 
1987). Delphinid mating systems are just beginning to be examined 
(Duffield and Wells, 1991). Paternity exclusion studies on long-finned pilot 
whales have shown that the adult males are not the fathers of the offspring 
in the groups (Amos et al., 1991a; 1991b; In press), suggesting that mating 
is occurring between males and females from different social groups. This 
could be an indication of the concept of a "behavioural avoidance of 
inbreeding" proposed by Moore and Ali (1983), which appears to occur 
between close kin which come into contact with one another. The 
implication is that the social groups of these large delphinids are not 
mating groups, or at least that they have a highly cohesive nature for 
benefits other than immediate access to mating partners. Other benefits to 
group formation, such as improved avoidance from predation or improved 
resource acquisition through cooperation (Hamilton, 1971; Alexander, 1974; 
Bertram, 1978; Caro, 1987; Norris and Schilt, 1988; Packer and Ruttan, 
1988; Packer et al ., 1990; Scheel and Packer, 1991), may be in effect. The 
possibility of this form of mating system appears to be unaccounted for in 
modelling of the evolution of mating systems (Clutton-Brock, 1988). 
However, models such as these, based on elemental features of 
mammalian societies such as the distribution of mates and resources, have 
predictive value and throughout this thesis I will be reviewing the evidence 
provided by the current state of knowledge concerning the social ecology of 
dolphins and small whales in order to deduce , elements of their mating 
systems. 
Another anomalous feature of pilot whale society is that there is a 
relatively high percentage of post-reproductive females , who may live for 20-
7 
30 years beyond their last ovulation. Although an age-related decline in 
reproductive output is common for mammals (Clutton-Brock, 1984), a 
complete cessation with extended longevity is found only in a few primates 
(Hrdy and Whitten, 1987). The findings have come from shore fisheries for 
short-finned pilot whales (G. macrorhynchus) off Japan which have 
recorded males up to age 50 years and females up to age 70 (Kasuya and 
Marsh, 1984). One-quarter of the females greater than 36 years old showed 
no recent ovarian activity. Additionally, one-sixth of these post-
reproductive females were still lactating (Kasuya and Marsh, 1984), 
probably due to the extended suckling of the last calf (Kasuya and Marsh, 
1989). It has been hypothesised that a female near the end of her 
reproductive lifespan should increase her reproductive effort as her 
potential for future reproduction decreases (Fisher, 1930; Williams, 1966; 
Pianka, 1976; Clutton-Brock, 1984). By improving her calf s survival 
through a prolonged nursing period (as has been shown for red deer: 
Clutton-Brock et al., 1982), the female may enhance her own lifetime 
reproductive success . The apparently inevitable decline in reproductive 
output through senescence results in selection for an increase in 
investment in the last offspring (Clutton-Brock, 1984). The continuation of 
life for 25 years after the last calf is born suggests that parental care is 
extremely important in these slowly-maturing mammals. These females 
may also provide additional benefits to the group such as assisting related 
females through the communal suckling of calves, a situation found in 
communal breeding mammals (Gittleman, 1985), or perhaps serving as 
long-lived repositories of critical information about habitat and resources, 
as has been proposed for elephant matriarchs (Douglas-Hamilton and 
Douglas-Hamilton, 1975; Dublin, 1983; Moss and Poole, 1983; Moss, 1988). 
Observations of killer whales have confirmed a potentially similar 
situation: 24% of the adult females have never successfully given birth in 20 
8 
years and yet maintain close associations with other reproductive females 
who are assumed to be their daughters (Bigg et al., 1990). 
The observational studies of killer whales are just beginning to be 
confirmed through genetic analyses (Duffield, 1986; Stevens et al., 1989; 
Hoelzel and Dover, 1990; Hoelzel, 1991a). However, there have been few 
detailed field studies of free-ranging pilot whales to examine the biological 
findings (although see Shane and McSweeney, 1990). For these reasons, a 
study of the social organisation of the pilot whale was undertaken to fill in 
gaps in the knowledge of social cetaceans. 
1.2 AIMS OF THIS STUDY 
I have examined the current state of knowledge of cetacean societies 
in the light of ecological and evolutionary theory, identified trends, and 
compared these findings with current theories of the evolution of society in 
terrestrial mammals. 
Although there are useful large scale trends in mammalian 
evolution which could be identified from a study of all cetaceans, I focussed 
my attention on a comparative study of the social cetaceans. These are 
primarily the .32 species of the family Delphinidae, sub-order Odontoceti, 
the toothed whales (see Table 1.1 for a complete taxonomy). I examined 
trends in the characteristics of body size, brain size, sexual dimorphism 
and social dynamics for all delphinid species where data were available. I 
attempt to correlate the variability in these characteristics with broad 
environmental categories of favoured habitat and prey and compare these 
trends with terrestrial mammals in an attempt to isolate typical 
mammalian features of cetacean life from unique aquatic adaptations. 
9 
Table 1.1. A list of the species, in taxonomic order, of the family 
Delphinidae, Cetacea (after Klinowska, 1991). References are selected 
studies concentrating on a given species and are not exhaustive 
Scientific Name 
Order Cetacea 
Sub-Order Odontoceti 
Superfamily Delphinoidea 
Family Delphinidae 
Subfamily Steninae : 
Steno bredanensis 
Sousa chinensis 
Sousa teuszii 
Sotalia fluuiatilis 
Subfamily Delphininae: 
Lagenorhynchus albirostris 
Lagenorhynchus acutus 
Lagenorhynchus obscurus 
Lagenorhynchus obliquidens 
Lagenorhynchus cruciger 
Lagenorhynchus australis 
Grampus griseus 
Tursiops truncatus 
Stenella frontalis 
Stenella attenuata 
Stenella longirostris 
Stenella clymene 
Stenella coeruleoalba 
Delphinus delphis 
Lagenodelphis hosei 
Subfamily Lissodelphinae: 
Lissodelphis borealis 
Lissodelphis peronii 
Subfamily Cephalorhynchinae: 
Cephalorhynchus commersonii 
Cephalorhynchus eutropia 
Cephalorhynchus heauisidii 
Cephalorhynchus hectori 
Subfamily Globicephalinae: 
Peponocephala electra 
Feresa attenuata 
Pseudorca crassidens 
Orcinus orca 
Globicephala melas 
Globicephala macrorhynchus 
Subfamily Orcaellinae: 
Common Name References 
odontocetes 
delphinids 
rough toothed dolphin 
Indo-Pacific humpbacked dolphin 
Atlantic hump-backed dolphin 
tucuxi 
28 
3 
12 
delphinines 
white-beaked dolphin 
Atlantic white-sided dolphin 
dusky dolphin 
Pacific white-sided dolphin 
hourglass dolphin 
Peale's dolphin 
Risso's dolphin 
bottlenose dolphin 
Atlantic spotted dolphin 
pantropical spotted dolphin 
spinner dolphin 
clymene dolphin 
striped dolphin 
common dolphin 
Fraser's dolphin 
northern right whale dolphin 
southern right whale dolphin 
Commerson' s dolphin 
black dolphin 
Heaviside's dolphin 
Rector's dolphin 
globicephalines 
melon-headed whale 
pygmy killer whale 
false killer whale 
killer whale 
long-finned pilot whale 
short-finned pilot whales 
34,35 
27,32 
23 
see section 4.6.1.1 
2 
2,13,15,21,25 
2,18,19,33,35 
2 
2,16 
1,7,20,27 
31 
14 
9 
11,29,30 
5,10 
24 
4,6,8, 17 ,22,26 
see section 4. 7 .1.1 
see Chapter 2 
see Chaps. 2, 3, & 4 
Orcaella brevirostris Irawaddy dolphin 36 
References: l)Banks and Brownell, 1969; 2)Barlow, 1984; 3)Barros and Cockroft, 1991; 
4)Brill et al., 1992; 5)Bryden et al., 1977; 6)Busnel and Dziedzic, 1968; 7)Collet and Saint 
Girons, 1984; 8)Comrie and Adam, 1938; 9)Crouetto and Medina, 1991; lO)Dawbin et al., 
1970; ll)Dawson and Thorpe, 1990; 12)Geise, 1991; 13)Hohn et al. , 1985; 14)Mermoz, 
1980; 15)Myrick et al., 1986; 16)Nishiwaki, 1975; 17)Nishiwaki and Tobayama, 1982; 
18)Norris and Dohl, 1980; 19)Norris et al., 1985; 20)0verholtz and Waring, 1991; 
21)Perrin, 1969; 22)Pilleri, 1967; 23)Pilleri and Gihr, 1969; 24)Pryor et al., 1965; 25)Pryor 
and Shallenberger, 1991; 26)Purves and Pilleri, 1978; 27)Ridgway and Green, 1967; 
28)Saayman and Tayler, 1979; 29)Slooten and Dawson, 1992; 30)Thorpe et al., 1991; 
31)von Waerebeek, 1991; 32)Walker et al., 1986; 33)~ells, 1984; 34)Wiirsig et al. , 1991; 
35)Wiirsig and Wiirsig, 1980; 36)Gretarsdottir and Amason, 1992 
I also review a number of other long-term studies of delphinid social 
organisation to supplement the findings of evolutionary trends within 
Delphinidae. The primary sources of this information are from the studies 
of Orcinus and Tursiops mentioned above , but studies of other delphinid 
genera, although not as extensive, will also be reviewed in order to provide 
comparative results (Norris and Dohl, 1980; Wursig and Wursig, 1980; 
Norris et al., 1985; Kruse, 1989; Pryor and Shallenberger, 1991; Wursig et 
al., 1991). 
However, before analysing these trends, I describe an original study 
of the social organisation of the short-finned pilot whale, Globicephala 
macrorhynchus, conducted off the island of Tenerife, in the Canary Island 
archipelago. These islands, located just 200 km off the coast of northwest 
Africa, in an area of upwelling generated by the cold Canary current, had 
been reported to contain accessible populations of pilot whales and other 
smaller dolphins (Vonk and Martin, 1988; Herve-Gruyer, 1989; Herve-
Gruyere, 1990). Utilising methods of individual identification (Wursig and 
Jefferson, 1990), a long-term study was initiated. This study is on-going 
and the results from almost two years presented here must be considered 
preliminary, as should any relatively short-term observational study on a 
long-lived, slow-maturing mammal. However, these results do contribute 
to an understanding of the family-wide trends in social behaviour analysed 
here . 
This thesis begins with a review of the biological knowledge of pilot 
whales , and proceeds with results from my' own field work on the 
distribution, behaviour and social structure of shor t-finned pilot whales. 
Finally, I compare biological and behaviour al trends within the 
delphinid cetaceans and attempt to integrate my findings with the current 
understanding of the evolution of mammalian social systems. 
10 
1.3 SUMMARY 
Recent information on two delphinid cetaceans which live in highly 
cohesive social groups of mixed age and sex has suggested that both sexes 
remain in their natal groups into adulthood, resulting in high degrees of 
relatedness between males and females in the same social group. This 
questions the proposed relationship between dispersal and the avoidance of 
inbreeding. The benefits of living with kin are great and certain degrees of 
inbreeding have been shown to be important in the evolution of societies. 
The critical question relates to the nature of the mating system in these 
cetaceans. Comparative analyses of morphology, genetics and social 
behaviour of all delphinids, including original research on one of the 
aforementioned species, the short-finned pilot whale, is proposed. The 
study of the evolution of mammalian societies would benefit from a better 
understanding of delphinid cetaceans, who share common ancestry with 
terrestrial mammals and yet have adapted to life in an extremely different 
environment. Delphinid cetaceans are especially well-suited to 
examination because they exhibit a wide degree of variability in body size 
and social behaviour. Results from this study will be compared with trends 
in the evolution of terrestrial mammalian societies. 
11 
• 
CHAPTER 2: REVIEW OF THE LIFE HISTORY AND BEHAVIOUR OF 
PILOT WHALES (Gwbicephala spp.) 
2.1 INTRODUCTION 
This chapter will provide an overview of the current knowledge of the 
morphology, physiology, behaviour and ecology of Globicephala. The scope 
of this review will be limited to those aspects which are used in later 
sections of this thesis; specifically: external appearance, taxonomy, 
genetics, sex differences in growth and body size, life history parameters of 
mortality and reproduction, ecological aspects of distribution and feeding 
ecology, and current knowledge on social organisation. These parameters 
are relevant to an understanding of the results of the field observations on 
G. macrorhynchus presented in Chapters 3 and 4, as well as contributing 
to knowledge of the comparative status of Globicephala within the family 
Delphinidae (Chapter 5). 
Throughout this chapter, I will first present information for G. 
macrorhynchus, followed by a summary of any differences reported for G. 
melas. Most of the information comes from specimens collected in 
conjunction with whaling operations in Ja pan and the Faeroe islands, but 
information on world-wide distribution is primarily from strandings. Pilot 
whales are one of the most common cetacean species to strand, especially 
as a group in a mass stranding (Sergeant, 1982; Klinowska, 1986). 
12 
2.2 EXTERNAL APPEARANCE 
The pilot whale (Globicephala spp.) is a medium to large delphinid, 
ranging in adult size from 3.6 to 6.3 m (see Section 2.5.1). There are two 
species recognised, the short-finned 'pilot whale, G. macrorhynchus, and 
the long-finned pilot whale, G. melas (see Section 2.3). As suggested by 
their names, they differ in the relative length of the pectoral flipper. In the 
short-finned species, the pectoral fin (measured from axilla to tip) is less 
than 15% of the body length (mean = 14.4%), while in the long-finned 
species it is greater than 16% (mean = 18. 7%: Sergeant, 1962b). There are 
also other skeletal differences, chiefly in skull morphology (Fraser, 1950; 
van Bree, 1971). 
Pigmentation 1s predominately black, hence the common name, 
"blackfish". There are three areas of lighter pigmentation which have been 
found to vary, both between species and between individuals of the same 
species (Yonekura et al., 1980): the throat patch, the post-dorsal saddle 
patch, and the post-ocular patch (sensu Mitchell, 1970). The throat patch is 
a trident-shaped, light-grey patch located on the ventral surface which 
begins approximately below the angle of the gape of the mouth and the eye. 
The outer side prongs of this trident taper posteriorly to the axilla of the 
pectoral fins, while the central prong extends along the length of the body to 
the anus (the 7:lid-ventral streak of Mitchell, 1970). This central stripe 
broadens into the genital patch around the genital slit in both species, 
extending outwards around the mammary slits in the female (Sergeant, 
1962b; Mitchell, 1970; Yonekura et al., 1980), and may help the young locate 
the mammaries during nursing (Mitchell, 1970). The throat patch is most 
distinct on the long~finned pilot whale (Sergeant, 1962b), appearing light 
grey to almost white, while it is faint to lacking in the short-finned pilot 
whale (Norris and Prescott, 1961; Yonekura et al. , 1980). 
13 
Ii 
I 
I J' 
I 
The post-dorsal saddle patch is just posterior to the dorsal fin. This 
patch covers the dorsal mid-line of the backbone, extending slightly forward 
to below the dorsal fin and down to the lateral mid-line of the body. The 
patch tapers posteriorly, extending to a point just anterior to the insertion of 
the tail flukes. The saddle patch appears slightly more distinct on short-
finned pilots than on the long-finned species (Norris and Prescott, 1961). 
The whiteness of the patch also varies between two segregated forms of 
short-finned pilot whales off Japan, with the northern form having a 
whiter saddle than the southern form (Kasuya et al., 1988a). Another 
distinction between these two forms is the distinctness of the posterior 
margin of the saddle. The northern form has a distinct posterior margin 
and the saddle does not extend posteriorly beyond the level of the anus, 
while on the southern form, the saddle gradually fades into the black body 
colour and extends almost to the tail flukes (Kasuya et al., 1988a). These 
distinctions are most apparent on mature adults. 
The post-ocular patch begins as a light grey spot dorso-posteriorly to 
the eye. It extends up towards the anterior insertion of the dorsal fin 
(Yonekura et al., 1980). On short-finned pilot whales off the Canary 
Islands, this patch continues ventrall1to the dorsal fin, grading into the 
forward extension of the post-dorsal saddle. 
It is interesting that the locations of these patches on pilot whales are 
similar to that of the more striking white marking of the killer whale 
(Davies, 1960). The eye spot, dorsal saddle patch and white ventral 
markings are broadly similar. Similarities in the pigmentation patterns of 
the smaller delphinids has also been noted by Mitchell (1970), who 
suggested they were adaptive responses for camouflage likely to represent 
evolutionary changes in pigmentation patterns from ancestral 
characteristics. 
14 
2.3 TAXONOMY 
The genus Globicephala, meaning "round head", was first proposed 
by Rene Lesson in 1828. There are currently two accepted species: 
Globicephala melas Traill 1809 and Globicephala macrorhynchus Gray 
1846 (Klinowska, 1991). The taxonomy of Globicephala has changed even 
recently and it is worth reviewing the history of the changes. I believe that 
there are important issues of variation within and between populations 
which can be addressed by an examination of taxonomy. 
Specimens of the currently accepted genus Globicephala have been 
known under a variety of names: Delphinus Cuvier 1812, Globicephalus 
Lesson 1828, Globiceps Flower 1884, and Sphaerocephalus Gray 1864 
(Hershkovitz, 1966). These were all clearly synonyms of Globicephala. 
There have been an even greater number of species names assigned to the 
two accepted species, but I will only review some of the more recent 
variations. 
The short- finned pilot whale, Globicephala macrorhynchus, was 
named by Gray in 1846 (Gray, 1846) from a skull collected in the "South 
Seas". This is currently accepted as the only warm-water Globicephala 
species and is distributed throughout all tropical oceans (see Section 2.6.1). 
Other specific names were used for tropical Globicephala: G. brachyptera 
from the western Atlantic of America (Cope, 1876), G. scammoni, from the 
Pacific explorations of the whaler Charles Scammon (Scammon, 1869; 
Cope, 1876), and G. seiboldii (Gray, 1846). Gray l~ter considered G. seiboldii 
to be a synonym of G. macrorhynchus (Gray, 1871), and the name fell out of 
use. Hershkovitz, in his Catalog of Living Whales, resurrected the name as 
a sub-species of G. melaena, applying to Pacific Ocean, tropical 
Globicephala (Hershkovitz, 1966). Van Bree reanalysed the type specimen 
15 
of G. seiboldii and confirmed that it was a synonym for G. macrorhynchus 
(van Bree, 1971). G. brachyptera was reanalysed by Fraser and was also 
considered to be a synonym of G. macrorhynchus (Fraser, 1950). G. 
scammoni was used as the name for the Pacific short-finned pilot whale 
until recently (Norris and Prescott, 1961; Sergeant, 1962b), although it is 
now also considered to be a synonym for G. macrorhynchus (van Bree, 1971; 
Klinowska, 1991). 
The long-finned pilot whale, Globicephala melas Traill 1809, has a 
less complicated history, although it has more recent controversy 
surrounding it. The earliest name given to the species was Delphinus 
melas by Traill in 1809. The genus was subsequently changed to 
Globicephala in 1828, as described above. In 1898, Thomas changed melas 
(from the Greek adjective, µt:Aacr, meaning "black") to the latinized 
feminine form, melaena, in order to match the feminine generic name 
(Hershkovitz, 1966). Two species names had also been assigned to the 
southern hemisphere population of the long-finned pilot whale: G. edwardii 
Smith 1834 and G. leucosagmaphora (Rayner, 1939). Davies (1960) reviewed 
these and suggested they be given sub-specific status (G. melaena 
edwardii), distinguished from the northern form (G.m .melaena) on the 
basis of a greater amount of white pigmentation on the ventral, dorsal, and 
post-ocular patches (as defined in Section 2.2). Hershkovitz combined these 
two sub-species_ into the single sub-species G. melaena melaena and 
assigned the short-finned pilot whale to another sub-species, G. m. sieboldi 
(Hershkovitz, 1966). Finally, van Bree analysed the distinct features of the 
warm-water and cold-water pilot whales, and acknowledged the wide 
variation within these forms, and proposed each be given specific status as 
G. macrorhynchus arid G. melaena (van Bree, 1971). G. melaena was the 
accepted species name for the long-finned pilot whale until 1989, when two 
16 
eminent cetologists entered into a taxonomic debate. Dale Rice proposed 
reverting to the original masculine form, as this predated the feminine 
form (Rice, 1989), and the International Code of Zoological Nomenclature 
(1985 edition) recommends retention of original specific names, even 
though the generic name may not match it in gender, in the case of non-
Latin (or non-Latinized) words (Rice, 1990)*. Schevill disagreed and pointed 
out that there was justification of the original changing of the specific name 
from melas to melaena because melas qualified as a latinized Greek word, 
and the Zoological Code allows the changing of specific names to match 
their new genus for latinized words (Schevill, 1990a). Thus, Schevill 
argued, the change from melas to melaena was justified and should be 
retained since it had come into such common usage in the 20th century 
(Schevill, 1990a). Rice maintained that melas was clearly a Greek word 
that was transliterated (copied letter for letter) into Latin (as opposed to 
being "Latinized"), and thus was not eligible for change with a change in 
genus. He thus continued to recommend the name Globicephala melas 
(Rice, 1990). Schevill, in a response following Rice's most recent 
correspondence, contended that melas was the result for both a latinization 
and a transliteration of the Greek word for "black", and still maintained the 
usage of melaena (Schevill, 1990b). Thus, it is still possible to find both 
names today. I have followed Klinowska in her usage of G. melas 
(Klinowska, 1991). 
* The distinction between "transliteration" and "latinization" is at the core of the issue. 
The Zoological Code states: "When a Greek word in transliterated its letters are given 
their exact equivalents. When it is latinized it is given the form which is determined by 
the usage of classical Latin or , where that differs, of modern scientific Latin"; cited in 
Rice, 1990). 
17 
The history of the naming of numerous species for the genus 
Globicephala illuminates a real phenomenon: geographic variation. There 
has been a trend towards a reduction in the number of species recognised 
in the classification of many genera of cetaceans. Much of this is due to an 
understanding of geographic variation in cetacean species (Perrin, 1984). 
Numerous characters, especially pigmentation, have been found to vary 
widely within an accepted species (Mitchell, 1970). This is also the case for 
Globicephala. 
2.4 GENETICS 
Recently, genetic methods for analysing the genetic diversity of 
populations and documenting paternities and relatedness (Burke, 1989) 
have been applied to populations of both pilot whale species. These results 
are interesting because they begin to provide an indication of the patterns of 
gene flow, and thus mating systems, in these whales. The techniques of 
DNA "fingerprinting" and the analysis of enzyme variability have both been 
used in these studies (Andersen, 1988; Wada, 1988; Amos et al., 1991a; 
1991b; In press). DNA "fingerprinting" allows the genetic variability of 
individuals to be identified from the electrophoretic analyses of allelic 
variation in hyper-variable fragments of DNA (Burke, 1989). It is most 
often used as a method to determine paternity. Electrophoretic enzyme 
analysis examines the allelic variability of polymorphic enzymes, but has a 
much lower resolution than DNA fingerprinting and can only be used to 
quantify population-wide variability (Andersen, 1988). All information 
reviewed here has been collected from whaling specimens. 
The evidence for short-finned pilot whales suggests that they have a 
strong tendency to form discrete populations. For example, two form s of 
18 
short-finned pilot whales hunted off the coast of Japan (Kasuya et al., 1988a) 
have been shown to be genetically distinct (Wada, 1988). Wada (1987) 
studied electrophoretic enzyme differentiation on a sample of 204 northern-
form and 167 southern-form pilot whales. A significant difference in gene 
frequencies was identified for the two forms, but the genetic distance 
between them was calculated to be at the inter-population level, indicating 
they were genetically isolated stocks (Wada, 1988). The two groups also had 
phenotypic differences in body size and pigmentation patterns (Kasuya et 
al., 1988a). The northern form is larger (1.0 m for adult females and 2.0 m 
for adult males) than the southern form. There are also differences in body 
proportions: the head is rounder (when viewed from above), the dorsal fin of 
the male is slightly narrower and the flipper length is slightly longer in the 
northern form when compared to the southern form (Kasuya et al., 1988a). 
The northern form also has a more prominent post-dorsal saddle patch. 
The remainder of the genetic analyses I will review are on the long-
finned pilot whales captured off the Faeroe Islands in the north Atlantic. 
The combination of the results from these studies show that pilot whales 
exhibit low levels of genetic variability within schools, while variability 
between . some schools was found to be high, suggesting a degree of stock 
differentiation (Andersen, 1988; Amos et al., 1991a; 1991b; In press). 
The relatedness of individuals within pods was tested by comparing 
the observed to expected ratio of shared alleles for four age classes to 
determine whether individuals were accompanied by their mothers, their 
fathers or other more distantly related animals. The proportion of animals 
with their true mothers ranged from 80% for the youngest individuals to 
25% for adults. However, the number of animals with their fathers was 
close to O for all age classes. This is evidence of a matrilineal social system 
(Amos et al ., In press). The shared proportion of alleles was also high 
19 
among adult males, providing evidence that males remain in their natal 
pods (Amos et al., In press). 
For the DNA analyses, tissue samples (skin, kidney or muscle) were 
collected from 326 pilot whales from five pods (Amos et al., 1991a; 1991b). 
The average band-sharing coefficient (the number of shared bands divided 
by the total number of bands) between randomly selected individuals from 
different pods was 0.56 (range: 0.37-0.78), a higher value than that reported 
for birds or humans (0.1 - 0.3), indicating low genetic variability (Amos et 
al., 1991a). 
Exclusion paternity analyses were conducted on five pods (34 
mother/foetus pairs and 37 adult males) which had been completely 
sampled (Amos et al., 1991a; 1991b). These analyses did not allow positive 
paternity assignments, only the exclusion or non-exclusion of specific 
males. Mature males from the same pod were excluded as fathers for 30 of 
the 34 mother/foetus pairs (88%) in paternity tests for 299 possible within-
pod male/foetus pairs. In one case, a possible father from a different pod 
than the mother/foetus pair was identified out of 959 possible between-pod 
male/foetus pairs. In the four cases where possible fathers were identified 
from the foetal pod, the length of the foetus averaged 0.088 m and in all 
cases was less than 0.020 m in length, equivalent to a foetal age of less than 
five months. This suggested that breeding adult males and females are not 
maintaining a 1ong-term association within the same pod (Amos et al., 
1991a). 
A closer examination of the paternal alleles (i.e. those foetal alleles 
not found in the mother) in the foetuses of two cohorts from 30 
mother/foetus pairs in the three largest pods was done to examine the 
r elative paternal contribution to cohorts . This was to find out whether one 
I[ 
t l 
1
1
1 
ii 
male had fathered all of the calves in a cohort or whether there were 
indications of multiple fathers. It was found that the likelihood of all 
foetuses from the same pod and same cohort having a single male as father 
was 36 times greater than the likelihood that each foetus had a different 
father and five times greater than ariy specific combination of two fathers 
(Amos et al., 1991a; 1991b). In any event, there were certainly fewer fathers 
than foetuses, indicating either that individual males were achieving 
multiple matings or possibly that groups of related males were breeding 
together (Amos et al., 1991a; 1991b). Thus, pilot whales appear to fit the 
predicted variability in male reproductive success common to polygynous 
mammals (Clutton-Brock, 1986). On the other hand, the comparison of 
paternal genotypes between two foetal cohorts within the same pod 
indicated that the same male could not have been responsible for breeding 
in successive years (Amos et al., 1991a). Thus, pilot whales do not appear to 
represent an extreme example of polygyny. This latter result also supports 
the hypothesis that males are not maintaining contact with (or exclusive 
access to) the same group of breeding females from year to year (Amos et 
al., 1991a; 1991b). 
Allele frequencies were also compared between adult females from 
the two largest pods and identified significant differences in allelic 
distribution (X2 = 43.5, df = 9, P <<.001), indicating genetic divergence 
between the pods (Amos et al., 1991b). This was surprising since the two 
pods were caught in nearby areas just two days apart. However, the 
ranking of the various alleles was similar for the two pods (Spearman r = 
0. 73, N = 16, P << .001), indicating a certain degree of relatedness due to "a 
shared genetic past or through genetic inputs from a common gene pool" 
(Amos et al., 1991b, p.263). 
21 
I 
, I 
'I 
I 
II 
II 
There were also indications that adult males from the same pod were 
related. The method used to examine this was an examination of the 
frequencies of single locus alleles for heterozygotes, since heterozygotes 
should be more common in siblings (Amos et al., 1991a). Males from both 
pods showed an excess of heterozygotes, although the excess was only 
significant in one of the pods and could have been due to the small sample 
sizes (Amos et al., 1991a; 1991b). 
Another finding was considered to be indicative of relatedness 
between two pods, perhaps in the form of "sister matrilines" (Amos et al., 
1991b). This was done by examining changes in the frequency of single 
alleles with the ages of the animals for two different pods. Parallel 
variation in the frequency of the "F" allele was identified for these two pods, 
especially for older animals. This was in spite of the fact that the overall 
frequency of the "F" allele varied by a factor of two between the pods. 
Correlated changes in allele frequency from age class to age class can 
provide some idea of the pattern of genetic input from males and females. 
However, female input would tend to dampen correlated variation, since 
females of different ages breed each year and the different allele frequencies 
between the two pods would be emphasised. If male genetic input was 
shared between the two pods for any given year, but varied from year to 
year, then the observed correlated changes in the allele frequency of 
similarly aged animals could be explained. Thus, a possible interpretation 
would be that the two pods represented '"sister' matrilines which have 
experienced similar male genetic input" (Amos et, al., 1991b, p.265) and the 
lower correlation for younger age classes would suggest that "although still 
spending time together, the pods have recently split, leading to divergent 
male inputs" (Amos et al., 1991b, p.265). 
22 
I 
I 
1
11 
II 
1 I 
II 
1 I 
ii 
I 
I 
I 
I I 
11 
The resulting picture of long-finned pilot whale social organisation is 
that pods represent groups of maternally-related females, i.e. that the 
social system is matrilineal (Amos et al., 1991a). The mating system and 
specifically the reproductive success of males in these pods is not quite as 
clear. Since males did not appear · to breed within their social group 
(perhaps deterred from mating with related pod females through 
behavioural inbreeding avoidance), they must be mating with females in 
other pods (Amos et al., In press). Variance in male reproductive success 
was suggested, but the mechanism of how this operates is unknown. 
Perhaps there are elements of polygyny in this system, in that only a few 
external males are successful breeders (Amos et al., 1991a), but there are 
only rare instances of all-male groups reported for pilot whales (Sergeant, 
1962a; Kasuya and Marsh, 1984; Chapter 4). The system still requires 
further examination. 
2.5 LIFE IDSTORY PARAMETERS 
Knowledge of the life history parameters of Globicephala has 
primarily come from data collected in conjunction with whaling 
operations. Most of the methods used in these studies are standard, and 
thus suitable for comparison. The methods of age determination are the 
core of the results presented here and have been the subject of much 
analysis . Toothed cetaceans grow one set of teeth during their life. As the 
teeth grow, rings of dentine and cementum are laid down, much like the 
rings of a tree. This is likely to be due to within-year variation in net energy 
balance, perhaps relating to the diet. The controversy is over the rate at 
which these rings are deposited. Most evidence points to an annual rate of 
ring deposition, but there have been relatively few studies where it has been 
possible to document this with other sources of information (e .g. direct 
I 
11 
, I 
'I ' 
11 
knowledge of the age of an animal born in captivity). There is also annual 
variation in the deposition of dentine which can result in thin rings which 
are difficult to recognise. The preparation of the materials (i.e. sectioning, 
staining, etc.) can also affect the readability of rings. However, some early 
studies (e.g. Sergeant, 1962a) have been reanalysed using more modern 
techniques (Kasuya et al., 1988b), enhancing their comparability. 
2.5.1 Body Size and Growth 
An understanding of the patterns of growth in Globicephala was 
necessary for an estimation of the age and sex classes of whales identified photographically in the Canary Islands field study. The most complete the southern form of data on body size and growth for "short-finned pilot whales comes from the 
whaling operations off the Pacific coast of Japan (Yonekura et al., 1980; 
Kasuya and Marsh, 1984; Kasuya and Matsui, 1984). Data were collected on 27 schools totalling 806 animals (565 females, 241 males and a few individuals of unknown sex) over a 16 yr period (Kasuya and Marsh, 1984). Dentinal layers were analysed and were determined to be deposited 
annually (Kasuya and Matsui, 1984). 
Short-finned pilot whales are born at approximately 1.4 m and 
weighing 37 kg (Kasuya and Marsh, 1984; Kasuya and Matsui, 1984). The largest foetus of 36 specimens was 1.46 m and the smallest of 11 calves below 1. 7 m was 1.36 m. The mean length at birth was calculated from the five smallest neonates (1.36 m - 1.42 m). Growth is rapid during the first 1.25 yrs, with both sexes averaging 2.30 m by the end of this period (Kasuya 
and Matsui, 1984). At this point, although growth rates slow for both sexes, 
males begin to grow faster than females. By 2.5 yrs of age, males are about 0.06 m larger than females (Kasuya and Matsui, 1984). Female growth 
continues at an annual rate of about 0.11 m/yr until a length of 
24 
I 
I 
I I 
I 
I 
I 
I 
I 
1, 
I 
I I 
approximately 3.22 m is attained at the age of first ovulation of 9.0 yrs 
(Kasuya and Marsh, 1984). For males, this period lasts until about 10 yrs of 
age, with a slightly faster growth rate of 0.12 m/yr. At the end of this 
per iod, males are approximately 3.44 m in length, or around 0.2 m larger 
than females. The next phase of female growth is characterised by a 
slowing of the growth rate to about 0.03 m/yr until they reach their 
asymptotic length of 3.64 m at an age of 22 yrs (Kasuya and Matsui, 1984). 
In males, this next phase of growth continues until 27 yrs of age and an 
asymptotic length of 4.73 m. The growth rate is high (up to 0.13 m/yr) until 
the attainment of sexual maturity at lengths of 4.01 m to 4.22 m and ages of 
14.6 yrs to 17 yrs (mean = 4.14 m and 15 yrs: Kasuya and Marsh, 1984). 
Then growth rate slows to about 0.03 m/yr until the asymptotic length is 
reached (Kasuya and Matsui, 1984). After the age of 22 for females and 27 
for males, all individuals stop growing. Of course, there was variation in 
these growth patterns: the largest male recorded was 5.80 m and the 
largest female was 4.05 m. 
Body weights were collected for 13 foetuses and 18 postnatal animals, 
up to the largest female of 3.55 m and 751.9 kg and the largest male of 2.91 
m and 379.22 kg (Kasuya and Matsui, 1984). A length-weight relationship 
was calculated from these data: 
Weight (kg)= 2.377 x 10-5 * Body Length(cm)2.8873 
Although males were under-represented in the data used to generate this 
equation, the authors consider it valid for adults of both sexes (Kasuya and 
Matsui, 1984). Predicted weights for asymptotic adult body length s were 
1256 kg for males and 590 kg for females. 
The resulting pattern of growth shows a high degree of sexual 
dimorphism for G. macrorhynchus, with adult males growing 1.3 times 
longer and 2.1 times heavier than females 
Changes in body proportion with growth were also examined for 17 
foetuses and 143 postnatal individuals from the same population (Yonekura 
et al., 1980). The greatest variation was in the development of the melon. 
At birth, the melon is similar to that of other delphinids: a slight bump on 
the sloping forehead. The beak is well distinguished and projects 
approximately 0.02 m beyond the melon. However, by a length of 2.4 m (age 
= 2 yrs: Kasuya and Matsui, 1984), the melon begins to project beyond the 
snout (Yonekura et al., 1980). After this, the degree of melon projection 
(MP: in cm) can be described by the equation: 
MP = (1.257 x 10-4 * Body Length2) - (4.517 x 10-2 * Body Length)+ 3.64 
This equation accurately described the rate of change in melon projection 
with body length for both males and females (Yonekura et al., 1980). In 
other words, a male and female of the same length will have the same 
degree of melon projection, and thus melon projection is not a secondary 
sexual characteristic (Yonekura et al., 1980). However, since male length 
at maturity (4.14 m) was greater than either the female asymptotic length 
(3.64 m) or the largest female recorded (4.05 m), all mature males have a 
greater degree of melon projection than mature females. 
The growth parameters of long-finned pilot whales are broadly 
similar to those of short-finned pilot whales. Th,e data available for long-
finned pilot whales has come from measurements of 4641 whales, although 
only 437 of these animals were aged (Sergeant, 1962a). Kastiya et al., (1988b) 
reanalysed the ages of all of the pilot whales from Sergeant's samples. 
Martin reported on the lengths of 116 pilot whales (52 of which were aged), 
26 
I 
that had stranded in Britain (Martin et al., 1987). The main difference 
between the two species is that long-finned pilot whales are larger than 
short-finned pilot whales at all life stages. Length at birth is 0.37 m longer, 
length at female maturity is 0.49 m larger, and length for males at 
maturity is 0. 76 m longer (Kasuya et al., 1988b). The asymptotic length of 
female long-finned pilot whales is 1.25 m longer and for males is 0.84 m 
longer. Sergeant (1962) also presented a length - weight curve for G. melas 
The equation for this was: 
Weight (kg)= 2.5 x 10-5 * Length(cm)2.895 
Using this equation, an average adult male of 5.57 m would weigh 2224 kg 
and an average female of 4.89 m would weigh 1524 kg. These values were 
used for subsequent comparisons of sexual dimorphism in Delphinidae 
presented in Section 5.3.2. 
2.5.2 Reproduction 
Reproductive parameters in the short-finned pilot whale were 
studied on the same population of whales analysed for body size and growth 
parameters (see Section 2.5.1), but only about half of the whales were 
examined (14 of 21 schools and 493 of 806 individuals: Kasuya and Marsh, 
1984). Reproductive status was studied by the macroscopic and histological 
examination of testes and epididymes for males and ovaries, uteri, 
mammary glands for females. Age was available from histological 
examination of rings in the teeth. Correlated i:oformation on school size 
and composition was also available because the entire school was driven 
and killed during the hunt (Kasuya and Marsh, 1984). 
I I 
2.5.2.1 Male Sexual Maturity 
Male sexual maturity was determined by examination of 70 - 150 
seminiferous tubules of the testes and was defined as the presence of 
spermatocytes, spermatids or spermatazoa (Kasuya and Marsh, 1984). 
Four stages of maturation were defined: 1) immature, with 100% of the 
tubules immature, 2) early -maturing, with less than 50% of the tubules 
mature, 3) late-maturing, with between 50% and 100% of the tubules 
mature, and 4) mature, with 100% of the tubules mature (Kasuya and 
Marsh, 1984). Testis weight was studied as the weight of a single testis 
since there was no significant difference in the weight of the left vs. the 
right testis (left testis = 49.4% ± 4.02% of the combined weight: Kasuya and 
Marsh, 1984). Testis weight was less than 100 g for immature males, 
between 100 and 400 g for maturing males and over 400 g for mature males. 
After the age of 25 yrs (corresponding to the cessation of growth), testis 
weights ceased to increase, but varied between 1700 and 3000 g for different 
individuals (Kasuya and Marsh, 1984). Most of the growth in testis weight 
was observed between the ages of 14 and 17 yrs. 
Ages and lengths of males were plotted for each of the categories of 
maturation. Early-maturing males ranged from 3.24 m to 4.34 m in length 
(mean = 4.01 m) and 7.25 to 16.5 yrs of age (mean = 14.6 yrs). Late-
maturing males ranged from 4.14 m to 4.55 m in length (mean = 4.14 m) 
and 14.5 to 29.5 yrs of age (mean = 15.8 yrs). Fully mature males ranged 
from 3.94 m to 5.25 m in length (mean= 4.22 m) and aged 15.5 to 45.5 yrs of 
age (mean = 17 .0 yrs: Kasuya and Marsh, 1984). The mean values are those 
lengths and ages when 50% of the individuals were at (or beyond) each 
stage and represent the mean length and age of attainment of each of the 
maturation stages (Kasuya and Marsh, 1984). The distinctions between 
immature and early-maturing males and between late-maturing and 
I I 
I 
! I 
11, 
I 
I 
I I 
I 
[ I 
I 
'I I 
: ; 1 
mature males were not as clear as the distinctions between these two major 
groupings. Thus, males classified as late-maturing or mature were 
considered to be functionally mature and mean values of 4.14 m and 15.8 
yrs were considered to represent length and age at functional maturity 
(Kasuya and Marsh, 1984). Also, the distinction between these two major 
groupings was closer for the relationship between length and maturity 
than it was for the relationship between age and maturity, indicating that 
males of large body size tended to mature at younger ages (Kasuya and 
Marsh, 1984). Finally, since both body weight (Kasuya and Matsui, 1984) 
and testis weight continued to increase until the age of 25, it is likely that 
social maturity (when males are successfully breeding within a social 
group), is attained at older ages than functional maturity (Kasuya and 
Marsh, 1984). 
Long-finned pilot whale males had a much more rapid maturation 
than short-finned pilot whales, reaching functional maturity at 12 yrs as 
opposed to 17 yrs (Kasuya et al., 1989). 
2.5.2.2 Female Reproductive Cycle 
Female sexual maturity was defined as the age at first ovulation, 
detected by the presence of ovulation scars (corpora lutea and corpora 
albicantia) on the ovaries (Harrison, 1949; Kasuya and Marsh, 1984; Marsh 
and Kasuya, 1981). The youngest mature female was 8.25 yrs of age, while 
the oldest immature was 11.5 yrs old (Kasuya and Marsh, 1984). Body 
lengths varied from 3.00 m for the smallest mature female to 3.44 m for the 
largest immature female (Kasuya and Marsh, 1984). Age at maturity was 
also calculated from six pregnant females younger than 10 yrs old, using 
their age at death, the length of their foetuses and the known foetal growth 
curve, resulting in estimates of between 7.4 and 8.1 yrs for the age of first 
1 I 
I 
ovulation. The regression of body length on age showed that 50% of the 
females attained maturity by 8.5 to 9.5 yrs and at a length of 3.16 m, thus, 
9.0 yrs and 3.16 m were estimated to be the mean age and body length at the 
onset of sexual maturity (Kasuya and Marsh, 1984; Marsh and Kasuya, 
1984). 
One of the main findings of Kasuya and Marsh was that none of the 
female pilot whales past the age of 40 yrs had any signs of recent ovulation 
(Kasuya and Marsh, 1984; Marsh and Kasuya, 1984). They noted that the 
oldest pregnant female in the sample (n=91 females) was 34.5 yrs and based 
on the size of her foetus, would have given birth at 35.5 yrs. One other 
female gave birth during the hunt and was subsequently aged at 35.5 yrs. 
The next youngest pregnant females were four 32.5 yr old animals (Kasuya 
and Marsh, 1984). The oldest females with recent ovulation scars on their 
ovaries were aged 39.5, 38.5 and 37.5 yrs. Since the oldest female in the 
sample was 62.5 yrs of age (Kasuya and Matsui, 1984), there appears to be 
an extended post-reproductive period in a female pilot whale's life (Kasuya 
and Marsh, 1984; Marsh and Kasuya, 1984; 1986; 1991). 
Before examining the characteristics used to define post-reproductive 
females, it is necessary to understand the cycles of actively reproducing 
females, the duration of the various stages of pregnancy, lactation and 
resting, and the life history parameters of ovulation rate, pregnancy rate 
and calving interval. 
Ovulation rate was estimated by the rate of accumulation of ovulation 
scars. Error in the estimation of this rate will arise from: 1) variation in 
the age at attainment of sexual maturity; 2) change in ovulation rate during 
an individual's reproductive lifespan and 3) individual variation in the 
accumulation rate (Marsh and Kasuya, 1984). The age at first ovulation 
. I 
has been shown to vary from 7 to 12 yrs of age (Marsh and Kasuya, 1984). A 
plot of the number of corpora (lutea and albicantia combined) for females of 
various ages shows that there is an age-related decline in the rate of 
accumulation, suggesting error #2 is possible, but this has been summed 
for different individuals. This decline only holds for females less than 40 
yrs old; the number of corpora in females older than this was independent 
of age (r = 0.12, P > 0.2: Marsh and Kasuya, 1984). The lack of individual 
histories due to the nature of the sampling also limits information on error 
#3. In spite of these difficulties, a single-phase exponential curve was fitted 
to the data (r2 = 0.66). Predicted values suggest that ovulation rate slows 
throughout life, from 0.7 ovulations per year for a newly mature seven year 
old, to 0.14 ovulations per year for a 39 yr old (Marsh and Kasuya, 1984). 
Although the model does not account for the cessation of ovulation after age 
40, it does suggest a slower ovulation rate of less than 0.13 ovulations per 
year above this age (Marsh and Kasuya, 1984). 
The duration of the calving interval was estimated by examining the 
proportion of females in each of the stages of pregnancy, lactation and 
resting (Kasuya and Marsh, 1984). This assumes a uniform distribution of 
females of various reproductive stages in the overall sample (i.e. there is no 
seasonal bias in catch samples) and that individual schools are also 
representative of the proportion of females of the reproductive stages in the 
population as a whole . 
The gestation period was estimated at 14.9 months by Kasuya and 
Marsh (1984), using an examination of the frequency distribution of foetal 
body lengths (Kasuya and Marsh, 1984). However, recent information for 
long-finned pilot whales suggests that gestation only lasts for 11.8 months 
(A.R.Martin, pers.comm.). Since it is unlikely that these congeneric 
species have such different gestation periods, this discrepancy will have to 
31 
be resolved. I will continue to present Kasuya and Marsh's estimate 
because it was an inherent part of some of their calculations, such as of 
pregnancy rate. 
The annual pregnancy rate (calculated as the proportion of pregnant 
females times 12 /14.9, to account for the fact that gestation was calculated 
to last longer than one year) was calculated as 0.25, meaning that 25% of 
the adult females were pregnant over any given year. 
The mean length of lactation, estimated from the proportion of 
lactating females relative to the number of pregnant females times a 
correction factor for the long gestation period, was 3.48 yrs (Kasuya and 
Marsh, 1984). This was slightly shorter than the estimated weaning age of 
calves (4.0 yrs) described in Section 2.5.3, but this could be explained by the 
fact that the length of lactation includes nursing terminated by calf 
mortality, while weaning age does not (Kasuya and Marsh, 1984). 
All females who were neither pregnant or lactating were classified 
as resting. The resting period was calculated in a similar fashion to that of 
the lactation period, and was estimated at 5.46 yrs. 
Some females classified as lactating or resting were post-
reproductive (see below), and thus these calculations would give an 
overestimate of the lactating and resting periods of females still 
reproducing. If the proportion of post-reproductive females (15.5% of 
lactating females, 49% of resting females) is excluded from the 
calculations, the new estimates are lactation perrod: 2.90 yrs and resting 
period: 2. 78 yrs. These estimates of gestation, lactation and resting can be 
summed to give the overall calving interval of 10.18 yrs including post-
reproductive females, and 6.92 yrs excluding post-reproductive females 
(Kasuya and Marsh, 1984). 
32 
11 
This estimate of calving interval assumes constant rates of 
pregnancy, lactation and resting for females of all ages. However, all of 
these stages appear to vary with the age of the female, culminating with the 
onset of the post-reproductive period at around 35 to 40 yrs of age. Females 
were grouped into age classes to · examine this variation. Annual 
pregnancy rate was shown to drop from 30.9% of females at 10 yrs old to 
4. 7% at 36 yrs old. The mean duration of lactation rose from 1.8 yrs for 15 yr 
old females to 3.2 yrs for 24 yr old females to 9.4 yrs for 36 yr old females. 
The resting period also increased with age. It appeared to be constant for 
all females younger than 24 yrs at 1.5 years, then increasing to 7.4 years for 
females aged 36 yrs. 
The cessation of ovulation and the occurrence of post-reproductive 
females was confirmed through a variety of methods. In addition to 
pregnancy, female reproductive status was determined by the forms of 
corpora, or scars of ovulation, on the surface of the ovary. After ovulation, a 
corpus luteum is formed. The absence of corpora lutea was an indication of 
a lack of recent ovulation and was a prerequisite condition for the definition 
of post-reproductive females. The corpus luteum regresses into a corpus 
albicans in as little as two years in young females, but appears to regress at 
a much slower rate in older females (Marsh and Kasuya, 1984). Corpora 
albicantia were classified into young, medium, and old on the basis of 
external features such as colour, texture and size. The change of a corpus 
albicans from "young" to "old" was a continual process, so these categories 
are somewhat arbitrary, but the rate of this proce,ss appears to slow during 
pregnancy. Also, since there was no evidence of ovulation past age 40 yrs, 
the presence of a medium corpus albicans in a 55.5 yr old. female suggests 
the process may be quite slow in old females (Marsh and Kasuya, 1984). 
The absence of young corpora albicantia was also a prerequisite for the 
33 
I 
I , 
I 
11 
11 
definition of post-reproductive females. The corpora albicantia appear to 
remain throughout life, based on the lack of a negative skew in size 
distribution for older corpora and the lack of a decrease in modal corpus 
size with increasing age of the female, and thus provided a permanent 
record of reproductive history (Marsh and Kasuya, 1984). The presence of 
large (>8 mm) Graafian follicles within the ovaries was also used as an 
indication of reproductive activity, with a decline in large follicle abundance 
observed in animals of increasing age (Marsh and Kasuya, 1984). Also, 
large follicles which did not ovulate (atretic follicles) were found to be more 
common with increasing female age (Marsh and Kasuya, 1984). Using 
these characteristics, 60 females (24% of the sample of 245) were classified 
as post-reproductive; the number classified into each of the categories is 
given in brackets below, along with the reproductive stage. The three types 
of post-reproductive females were: 1) females with medium corpora 
albicantia but lacking large follicles {2 lactating, 10 resting}, 2) females with 
old corpora albicantia and only atretic follicles {3 lactating, 3 resting}, and 
3) females with only old corpora albicantia and no follicles {6 lactating, 36 
resting} (Marsh and Kasuya, 1984; Figure 8). 
Based on the definition of post-reproductive females outlined above, 
the youngest post-reproductive female was aged 29.5 yrs, and the proportion 
increased in older females (8.5% for females aged 28 to 32 yrs, 19.6% for 
ages 32 to 36, 33._3% at ages ,36 to 40, and 100% for all females older than 40 
yrs: Kasuya and Marsh, 1984). Thus, the post-reproductive period appeared 
to begin as early as 30 yrs, but more commonly-after the age of 40 yrs 
(Kasuya and Marsh, 1984; Marsh and Kasuya, 1984). 
Kasuya et al. (1988) re-examined the data on long-finned pilot whales 
in light of these findings, and found that the females do not appear to have 
an extended post reproductive period. Female long-finned pilot whales also 
34 
,11 
I I 
I 
matured at an earlier age (6 yrs vs. 9 yrs) and had no apparent age-
dependent decline in pregnancy rate (Kasuya et al., 1988b). 
2.5.2.4 Seasonality of Reproduction 
There was only limited evidence suggesting a seasonal variation in 
male reproductive activity (Kasuya and Marsh, 1984). This was tested by 
examining variation in three indications of reproductive activity (threshold 
testis weights for sperm production, overall testis weight, and seminiferous 
tubule diameter) for animals killed in three different "seasons " 
(determined by available sample sizes): February, May/June/July, and 
December (Kasuya and Marsh, 1984). The greatest changes were noted in 
the threshold weights of maturing tes~s producing spermatozoa. 
Spermatozoa were detected in smaller testis during May -July than during 
December, indicating increased reproductive activity in May through July 
(Kasuya and Marsh, 1984). Secondly, seasonal changes in seminiferous 
tubule diameters for mature males were found to vary significantly on a 
seasonal basis, being slightly larger in males of 4.80 m to 4.99 m during 
May to July (Kasuya and Marsh, 1984). There were no significant seasonal 
differences in testis weights between seasons. However, since the testis 
weights of mature males showed a wide range of values, it is possible that 
at least some males exhibit variation in mating activity. Since seasonal 
sample sizes were small and the results were inconclusive, Kasuya and 
Marsh (1984) "consider that a substantial proportion of the adult male 
population of G. macrorhynchus 1s probably capable of successful 
reproduction throughout the year" . 
Analysis of the length frequency distribution of foetuses and calves 
less than 1.55 m was used to determine the seasonality of births (Kasuya 
and Marsh, 1984). Body lengths were plotted by month of capture and the 
11 
I 
I 
I 
I 
I 
ii 
,I 
mean date at which all individuals attained the mean birth length of 1.40 m 
(using the foetal growth curve developed by Kasuya and Matsui, 1984) was 
estimated. The mean date of birth varied from 25 July to 12 August, 
depending on the applicability of the linear foetal growth curve to different 
length classes of animals used in the calculations. The inclusion of all 
animals resulted in an estimate of the mean birth date of 2 August with a 
standard deviation of 73.3 days. Using a gestation period of 14.9 months, 
the mean date of conception would be in May (Kasuya and Marsh, 1984). 
The overall frequency of parturition date estimates for individual animals 
ranged throughout all months of the year, with the bulk of births occurring 
between May and November and with a mode in July/August. This 
indicates that pilot whale reproductive activity is only "diffusely seasonal" 
(Kasuya and Marsh, 1984). 
Ovarian activity in female pilot whales showed a similar pattern. 
There was a seasonal cycle in the growth of ovarian follicles in lactating 
and immature females, with the peak from May to July (Marsh and 
Kasuya, 1984). Follicle growth was considered to be an indication of 
impending ovulation (Marsh and Kasuya, 1984), and the peak season 
corresponds to the mean conception date of May (Kasuya and Marsh, 1984). 
2.5.3 Duration of Lactation and Weaning 
Lactation is considered separately because it has relevance to the 
mother-calf relationship observed in Canary Island pilot whales. Lactation 
was determined by the presence of milk in the mammary glands. Seventy-
four females were classified as lactating in the Japanese whaling sample 
(Kasuya and Marsh, 1984). Pilot whale milk had a "texture like cow's milk 
36 
I 
I 
I I 
I 
I 
I 
I I 
and colour varying from creamy white to a distinct green" (Kasuya and 
Marsh, 1984). * The duration of lactation was estimated from four sets of 
data. Stomach contents showed that the youngest calf with solid food was 
0.5 yrs and the oldest calf with milk (along with squid beaks) was 2.75 yrs 
old. Milk was also recorded in two unidentified stomach contents collected 
from a school in which the two youngest members were 2.5 and 3.0 yrs old. 
Thus, the estimate of the duration of lactation from this method ranged 
from 0.5 to 3.0 yrs. The first occurrence of solid food corresponds to the age 
at which the teeth erupt (Kasuya and Marsh, 1984). 
The second method to estimate the duration of lactation was the 
comparison of the number of lactating females with the age and number of 
juveniles caught in the same school. This method assumed that each 
female was nursing only one calf at the time of death, that no suckling calf 
was older than any weaned calf, that sexually-mature animals were 
completely weaned and that both mother and calf were caught in the same 
school. The oldest presumed suckling individuals were a 15.5 yr old, 
histologically-immature male and two 13.5 yr old immature males (Kasuya 
and Marsh, 1984). The youngest, fully weaned calf was 2.0 yrs, but there 
were many presumed suckling calves up to 6 yrs of age. The conclusions 
were "that some precocious calves complete weaning by the age of 2.0 yrs, 
but that a few calves continue to suckle until the age of 10 yrs" (Kasuya and 
Marsh, 1984). The mean age at weaning was estimated at 3.5 to 5.5 yrs , 
with a median of 4 yrs (Kasuya and Marsh, 1984). 
* Th e presence of green milk appeared to be a real phenomenon, not dependent on either the 
quantity of milk in the mammaries or the time since death , and showed a seasonal 
variat ion suggesting it was related to seasonal changes in diet (Kasuya and Marsh, 1984). 
'I 
I 
I , I 
11 
11 
The presence of three suckling males greater than 13 yrs of age could 
be explained by the loss of younger calves during the drives of the schools in 
question, but all drives occurred in calm conditions and this possibility was 
considered highly unlikely by all observers, including the whalers 
themselves (Kasuya and Marsh, 1984); The presence of old, suckling calves 
was correlated with the presence of old, lactating females and suggests that 
older females nurse their calves longer than young females (Kasuya and 
Marsh, 1984; 1989). The implications of this possibility in relation to the 
occurrence of post-reproductive females will be discussed below. 
The third method of estimating lactation period was by examining 
the difference between female age of last parturition and the age of the 
oldest lactating females (Kasuya and Marsh, 1984). The oldest pregnant 
female in the sample was 35.5 yrs old and the oldest female with signs of 
recent ovulation was 39.5 yrs old (Marsh and Kasuya, 1984). However, 
many females between the ages of 29.5 and 39.5 showed no recent signs of 
ovulation (Marsh and Kasuya, 1984). The age of the oldest pregnant female 
(35.5 yrs) was used as the estimate of age of last parturition (Kasuya and 
Marsh, 1984). A sample of 12 females older than 35.5 yrs were still 
lactating (range: 36.5 to 50.5 yrs; mean = 43.4 yrs). Only one female had a 
recent ovulation scar, or corpus luteum (age: 36.5 yrs), while the rest had 
only scars of past ovulations, or corpus albicans (Kasuya and Marsh, 1984). 
The mean length of lactation after the estimated age of last parturition was 
7 .83 yrs (Kasuya and Marsh, 1984). This estimate was compared with the 
estimate using the age of the calves caught in t}:l.e same school (described 
above), and in most cases there was a close agreement with the two 
estimates. The two anomalous estimates were in cases when old, lactating 
females (aged 50.5 and 47.5 yrs) were caught in the same school with young 
calves (four and 4.5 yrs old, respectively). Since one of the assumptions of 
38 
ii 
the calf age method was that no suckling calf was older than any weaned 
calf, the two calves were assumed to be nursing from the lactating females, 
and the females' age of last parturition would have been 46.5 and 43 yrs old, 
respectively. Both of the females were classified as having only "old" 
corpus albicans in their ovaries, which, although observed to form from 
corpus luteum as quickly as two years in young females, were considered to 
form at a much slower rate in old females (Marsh and Kasuya, 1984). 
However, if these calves were assumed to be the offspring of other, young 
pregnant females caught in the same schools, and were already weaned, 
the next oldest calves in the schools were 10.5 and 14.5 yr old histologically-
immature males. This corresponded to the situation with two other old, 
lactating females who were predicted to be nursing adolescent male calves 
(Kasuya and Marsh, 1984). Thus, the assumption that all suckling calves 
were younger than all weaned calves may not always be valid (Kasuya and 
Marsh, 1984). It could also have been possible that the old females were 
nursing young, fostered calves, but there was no information relevant to 
this possibility (Kasuya and Marsh, 1984). When the age of the presumed 
suckling calf (including the revised estimates of the anomalous examples) 
was subtracted from the lactating females' age, the mean age of last 
parturition was 34.3 yrs, close to the 35.5 yrs of the oldest pregnant female 
in the sample (Kasuya and Marsh, 1984). 
In summary, the nursing period of short-finned pilot whales appears 
to be long, varying from 2-14 yrs. Long periods of nursing are considered to 
be related to the extended period of learning required for social cetaceans 
(Brodie, 1969). The evidence shows that pilot whale calves begin to take 
solid food as early as the first year, after the teeth erupt in the jaw, and may 
be weaned as early as 2 yrs. However, many continue to nurse up to 6 yrs 
and the oldest presumed nursing calves were 14 yr old immature males. 
In all cases, there was a relationship between the duration of lactation and 
the age of the female. Females who had no recent signs of ovulation (and 
were assumed to be post-reproductive) were often still lactating and were 
presumed to be nursing their last calf. The extended duration of lactation 
in these post-reproductive females represents an increased parental 
investment when the female 's potential for future reproduction has ceased. 
This is in accordance with evolutionary parental investment theory and the 
partitioning of reproductive effort (Trivers, 1972; Clutton-Brock, 1984). In 
two of 12 cases there was a possibility that females were nursing fostered 
calves, as has been observed in other social mammals (e.g. lions: Schaller 
{1972}; elephants: Dublin {1983} ). 
2.5.4 longevity, Mortality and the Sex Ratio 
The maximum longevity for Japanese short-finned pilot whales 
differed between sexes. The oldest male in the catch was 45 yrs old, while 
the oldest female was 62 yrs (Kasuya and Matsui, 1984). There are no 
reports of all-male groups which would indicate any older males were 
segregated from the female and calf groups, so this finding was considered 
to be an indication of differential mortality (Sergeant, 1962a; Kasuya and 
Marsh, 1984). 
The sex of 155 foetuses and calves less than 2.2 m (about 0.5 yrs old) 
was 48% female, · giving a male:female sex ratio of 1:0.92, which was not 
considered significantly different from a one-to-one sex ratio (Kasuya and 
Marsh, 1984). The catch statistics from the drive whaling, where all 
animals in a school were killed, showed that 67 .3% of 483 postnatal 
individuals were female (sex ratio = 1:2.05), significantly different from a 
one-to-one sex ratio (X2 P < .001: Kasuya and Marsh, {1984}). However, 
when this catch is segregated into 10 year age classes, the sex ratio becomes 
40 
I 
,I, 
,I 
I 
I Ii' 
I I 
I I 
increasingly female-biased for older animals. For animals younger than 20 
yrs old the sex ratio does not differ significantly from a one-to-one sex ratio 
(0-10 yrs: 1:1.5, N=115, P >0.2; 10-20 yrs: 1:1.45, N=120, P >0.2). However, by 
the age of 20-30 yrs, the sex ratio of 1:1.85 was different from one-to-one 
(N=114, P<0.02) and was significantly female-biased for all older age classes 
(30-40 yrs: 1:3.17, N=71, P<.05; 40-46 yrs: 1:5.17, N=37, P<.001). Finally, after 
the age of 46 yrs, females represented 100% of the population (N=26, P< .001: 
(Kasuya and Marsh, 1984). The overall adult sex ratio was 77% female, or 
1:3.35. This could have been biased towards an overestimation of the 
number of males by the whalers' preference for schools with the most 
males. Separate analysis of the sex ratio of schools caught alone (i.e. when 
the whalers had no chance to select one school over another: see Section 2. 7) 
found a female ratio of 87%, or 6.7 females for every adult male (Kasuya and 
Marsh, 1984). 
The age distribution of 150 males and 318 females, pooled into three-
year age classes, was used to calculate mortality rates (Kasuya and Marsh, 
1984). The assumptions for this method of determining mortality rates are 
that the population is stable or has a known rate of increase. The lack of 
evidence supporting these assumptions makes the following mortality rate 
estimates speculative (Kasuya and Marsh, 1984). The best fit for the data 
was to calculate separate least-squares linear regressions for the two oldest 
of three age groups for each sex. Mortality rates for the youngest age group 
was estimated by extrapolation and corrected observed frequencies because 
there was an apparent bias against females with young calves in the catch 
(Kasuya and Marsh, 1984). Mortality rates for males from birth to nine yrs 
(age at sexual maturity) was 11.02%, from ages nine to 30 yrs was 4.01 % 
and for males 27 to 46 yrs was 11.25%. Mortality for females from zero to 6.5 
yrs was 7.34%, for females from 6.5 to 48 yrs was 2.54% and for females 
41 
I . 
II .I 
I 
: I 
11 
1· 
I 
I 
I 
from 45 to 63 yrs was 14.45% (Kasuya and Marsh, 1984). These are typical 
"U"-shaped mammalian mortality patterns. Males had higher mortality 
than similarly-aged females for nearly all ages. The period of high juvenile 
mortality showed the greatest difference between males and females. Also, 
the increase in adult mortality occurred around 18 years earlier in males 
than females (27 yrs vs. 45 yrs). 
Life tables were constructed from these data, using parameters of 
single births, a neonatal sex ratio of one-to-one, and the previously 
estimated rates of maturity and pregnancy (Kasuya and Marsh, 1984). The 
resultant predictions closely agreed with the observed data. The mean life 
expectancy of newborn calves was 12 years in males and 22 years in 
females. The adult sex ratio was 3.7 females: 1 male, slightly higher than 
the observed sex ratio of 3.3:1. The annual pregnancy rate was 12.85% of 
the female population. The gross annual reproductive rate (rate of calves of 
both sexes produced annually by the female population) was around 5.8%. 
The average calf production per female living to reproductive age was 4.39 
calves. The mean proportion of post-reproductive females was predicted at 
24% of the adult female population, close to the 25% observed. 
Long-finned pilot whales had a different pattern of mortality than 
short-finned pilot whales (Kasuya et al., 1988b). Although long-finned pilot 
whales also exhibited differential mortality between the sexes, with adult 
males having higher mortality than similarly aged females, mortality in 
both sexes increased around ages 21-25 yrs. Also, longevity in the long-
finned pilot whale appeared to be as much as 20 years shorter than for 
short-finned pilot whales in females and 10 years in males, suggesting an 
overall higher mortality rate for this species (Kasuya et al ., 1988b). 
42 
I, 
2.5.5 Summary of Sho~Finned Pilot Whale Life History 
The reproductive lives of the southern form of Japanese short-finned 
pilot whales can be divided into three phases (Kasuya and Marsh, 1984). 
Males have a long (15.9 yr) period of sexual maturation, throughout which 
they may continue to nurse if they are the last offspring of an old mother. 
After reaching sexual maturity, they continue to grow in body size, 
eventually growing to twice the mass of a female (1200 kg vs. 570 kg). From 
the ages of 16 to 25, males continue to grow at a relatively high rate, 
induding growth in the testes, suggesting a continuation of the process of 
maturation. Growth slows during the latter part of this period, but 
continues until 27 yrs when asymptotic body size is reached and they are 
fully adult. They undergo only minor seasonal variation in reproductive 
potential and are probably capable of fathering offspring for most of the 
year. This would be expected since births were predicted for all months 
(from the length frequencies of foetuses), in spite of a unimodal peak of 
conception in May and birth in August. They have higher mortality rates 
than females throughout their lives, resulting in a female-biased sex ratio 
of up to 3. 7 adult females for every male. A few males may live until the age 
of 45, 22 years less than the female lifespan of 62. 
Female pilot whales have a relatively shorter period of maturation 
compared to males, although it is still long by mammalian standards. 
Until sexual maturity, at ages between 8 and 11 yrs, female growth rates 
are only slightly less than those of males. However, female growth slows 
after first ovulation, and there is no adolescent growth spurt as there is for 
males. Female growth ceases by 22 yrs of age, with an asymptotic body size 
equivalent to the body size of an 11 yr old adolescent male. The first 
pregnancy occurs at 8-9 yrs, with a 15 (or 12) month gestation, and a 
m1n1mum one year period of nursing. The calving interval for 
43 
I 
,J 
11 
,11 
I' 
11 
primiparous females was three years. Calving interval increases with age, 
due to longer periods of lactation and resting between pregnancies. A 
highly productive female may have had four calves by the age of 21, when 
the calving interval may have increased to 5-6 yrs. After this period, 
pregnancy rate drops, along with an increase in the calving interval. By 
the age of 35, the female may have had 2-3 additional calves, but after this 
age, nearly all females have ceased to give birth. The highly reproductive 
female may have had six or seven calves, but the average lifetime calf 
production was 4.4 calves. Some (about 15%) may continue to lactate and 
nurse their last calf, especially if it is a male with a long maturation period. 
By the age of 52, all females have ceased lactating. Life may continue for 
another ten years. Thus, the three phases of a female pilot whale's life are 
9 yrs of maturation, 26 yrs of producing calves, and a maximum of 28 yrs of 
post-reproductive life. Of course, mortality rates indicate that only a few 
females will live long enough to enter the post-reproductive period. 
2.6 ECOLOGY 
2.6.1 Distribution 
Water temperature appears to be the primary factor in determining 
the differences between the distribution of the two species of Globicephala. 
G. melas has an anti tropical distribution, with two separate populations 
occurring in the cold, temperate waters of the northern and southern 
hemispheres (Davies, 1963), while G. macrorhynchus has a circum-
equatorial distribution, favouring waters above 15°-l6°C. I will first 
describe the distributions of the two species separately and then examine 
the areas of overlap. 
44 
The short-finned pilot whale is distributed in all warm-water oceans. 
In the Atlantic Ocean, these whales are found north to 39°N along the coast 
of America (Sergeant, 1982) and range across the north Atlantic to Madeira 
at 30°N (Maul and Sergeant, 1977), the Azores (Clarke, 1981) and France 
(Duguy, 1968). The species does not appear to enter the Mediterranean 
(Brown, 1961; McBrearty et al., 1986). Ther e have been sightings 
the 
throughout,. western Atlantic, including the Bahamas and the Virgin 
Islands (Taruski and Winn, 1976), the eastern Gulf of Mexico (Fehring and 
Wells, 1976), the Caribbean islands of St. Vincent and St. Lucia (Caldwell et 
al., 1971; Caldwell and Caldwell, 1975) and Venezuela at 10°N (Casinos and 
Bou, 1980). Reports from the eastern Atlantic have come from Senegal 
(Cadenat, 1957), the Canary Islands (Vonk and Martin, 1988) and the Cape 
Verde Islands (Fraser, 1950). The southern-most reports are from Sao 
Paulo, Brazil at 25°S (Schmiegelow and Filho, 1989) and off the Indian 
Ocean coast of South Africa at 34°8 (van Bree et al., 1978; Ross, 1984). 
In the Pacific, the short-finned pilot is commonly found as far north 
as 42° N along the Japanese coast (Kasuya, 1975; Kasuya et al., 1988a) and 
to about 35° N along the California coast (Norris and Prescott, 1961; Seagars 
and Henderson, 1985), although reports exist from as far north as 51 °N off 
the coast of British Columbia (Baird and Stacey, 1989). Short-finned pilots 
are seen off Mexico (Norris and Prescott, 1961) and in the vicinity of the 
Pacific Islands, such as Hawaii (Shane and McSween ey, 1990) an d Tahiti 
CM.Poole, pers.comm.) In the western south Pacific, the short-finned pilot 
has been reported from Java at 5°S (DammermaJ?-, 1924), and as far south 
as 41 °S off Tasmania (Baker, 1983). However, the most southerly mass 
stranding has been in north Austr alia, while a ll other sightings from 
further sout h (including reports from New Zealand) have been of single 
animals (Baker, 1983; Nicol and Croome, 1988). There are only a few 
45 
I 
: I 
I 
II' 
reports of short-finned pilot whales in the Indian Ocean, although there 
has not been much research effort there (Leatherwood and Reeves, 1989). 
They have been sighted off Sri Lanka at 10°N (Alling, 1986) and there have 
been strandings from as far north as Calcutta at 22°N (Silva, 1987) and off 
south-east India (Alagarswami et al., 1973). 
The distribution of short-finned pilot whales off the coasts of Ja pan 
has been shown to vary seasonally (Kasuya, 1975; Kasuya et al., 1988a). The 
two genetically-distinct populations off the Japanese coast (Wada, 1988) are 
at least partially segregated by water temperature (Kasuya et al., 1988a). 
The southern form occurs primarily south of 30°N in the winter and south 
of 37°N in the summer (Kasuya et al., 1988a). This corresponds to the 
northward movement of the boundary region of the 18°C surface isotherm 
of the Kuroshio current (Uda, 1954; Kasuya, 1975). The lower limit of 
thermal tolerance for this southern form appears to be just under 20° C 
(Kasuya et al., 1988a). The northern form occurs between 36°N and 43°N in 
the summer and between 35°N and 36°N in the winter corresponding to 
temperature preferences between 12°C and 24°C (Kasuya et al., 1988a). 
This lower thermal tolerance corresponds to the 10°C-15°C front of the cold 
water Oyashio Current (Kasuya et al., 1988a). 
Although long-finned pilot whales are found in both hemispheres of 
the Atlantic Ocean, there are no current records of a northern population 
in the Pacific Ocean. However, skulls have been found in Japan (45°N) 
from the 10th century, showing that the species used to exist there (Kasuya, 
1975). The population in the north Atlantic appears to extend north towards 
Greenland with sightings at 62°N (Brown, 1961) and reports along the 
western Atlantic from Canada (Needler, 1931; Mathewson, 1935; Sergeant 
and Fisher, 1957; Geraci and St. Aubin, 1977) and the United States south to 
36° N (Paradiso, 1958; Sergeant, 1982). Long-finned pilot whales are 
46 
i I 
commonly sighted across the central Atlantic at 50°N (Brown, 1961). 
Sightings from the eastern Atlantic are common in the Faeroe Islands 
(Brown, 1961), off Britain (Ritchie, 1924; Kock, 1956; Fraser, 1974; Evans, 
1980; Martin et al., 1987), Spain (Casinos and Vericad, 1976) and in the 
Mediterranean (McBrearty et al., 1986). The southern-most sightings of 
this northern population are from one stranding in the Canary Islands at 
28°N (Vonk and Martin, 1988) and from 25°N off the coast of Mauritania, 
northwest Africa (Nores and Perez, 1988). 
The southern population of the long-finned pilot whale appears to 
have a continuous distribution around the southern hemisphere (Davies, 
1960; Brownell, 1974). In the south Atlantic, long-finned pilot whales are 
commonly found off the eastern and southern South African coast at 35°8 
(van Bree et al., 1978; Ross, 1984), while in the southern Indian Ocean there 
are reports from Kerguelen Is., at 48°8 (Brownell, 1974). Along the Atlantic 
coast of South America their distribution is probably limited by the Falkland 
(=Malvinas) current (Casinos, 1981). The most southerly report is at 68° S, 
off Antarctica (Brownell, 1974). In the south Pacific, long-finned pilot 
whales have been recorded from Chile (Torres et al., 1975), Tasmania at 
42°8 (Gales et al., 1992), southeastern Australia (Davies, 1960) and New 
Zealand (Oliver, 1924; Gaskin, 1968b). 
Distribution of the long-finned pilot whale in the western north 
Atlantic has been shown to be seasonal (Sergeant and Fisher, 1957; 
Sergeant, 1962a; Mercer, 1967; Waring et al., 1990; Overholtz and Waring, 
1991). Along the Newfoundland coast, the animals first occur in the coastal 
bays in mid-July and remain until October (Sergeant and Fisher, 1957; 
Sergeant, 1962a; Mercer, 1967). This pattern of occurrence closely parallels 
the movements of squid in this region (see Section 2.6.2). The offshore 
distribution during this period is the most northerly of the year (Brown, 
47 
I, I 
I' 
I I 
1961; Waring et al., 1990), although pilot whales may be found in deep water 
areas of the north Atlantic during all seasons (Brown, 1961). During the 
winter and spring, the animals appear to move south to warmer, Gulf 
Stream waters, primarily along Grand Bank (Sergeant and Fisher, 1957; 
Sergeant, 1962a; Waring et al., 1990; Overholtz and Waring, 1991). Over 
80% of the pilot whales caught in offshore fishing nets (along the 
continental shelf) were caught between March and July (Waring et al., 
1990). However, pilot whales are also occasionally reported from inshore 
waters of Newfoundland during the winter (Mercer, 1967). 
There are a number of areas of overlap between the two species of 
Globicephala (van Bree et al., 1978; Nores and Perez, 1988). In the northern 
hemisphere, the species only overlap in the Atlantic Ocean. As described 
above, G. melas extends south to 25°N latitude, while G. macrorhynchus 
has been found as far north as 45°N (Nores and Perez, 1988). The situation 
along the north west African coast is most interesting. G .melas occurs off 
Morocco and Mauritania, while G. macrorhynchus is only reported from 
Senegal. However, the majority of sightings offshore of the mainland coast, 
in the Cape Verde, Canary, and Madeira Islands, have been of G. 
macrorhynchus. This appears to be related to prevailing ocean 
temperatures. The cold-water species, G. melas, travels south with the 
influence of the Canary Current, which lowers water temperatures off 
northwest Africa _ through an upwelling of deep cold water (Nores and 
Perez, 1988). Offshore, the water is warmer, so the warm-water species, G. 
macrorhynchus, is able to penetrate north to Madeira and the Azores 
(Maul and Sergeant, 1977; Clarke, 1981). It would be expected that 
observations of G. macrorhynchus in the Bay of Biscay would be correlated 
with warm water intrusions during the fall, but they appear to occur 
throughout the year (Nores and Perez, 1988). Along the western Atlantic, 
48 
I I 
I 
the two species overlap between 36° and 38°N (Sergeant and Fisher, 1957; 
Paradiso, 1958; Sergeant, 1982). The reports from the western Atlantic are 
all of stranded animals, so there is little information on variations in 
distribution relative to oceanographic features. However, the warm-water 
Gulf Stream would be expected to allow G. macrorhynchus to penetrate 
north. Correspondingly, when the Gulf Stream shifts to more offshore 
flow, it would be expected that G. melas could penetrate further south with 
the influence of the cold Labrador current. In any event, current 
information suggests a wider zone of overlap along the eastern Atlantic 
compared to the western Atlantic. 
In the southern hemisphere, a zone of overlap distinguished by 
varying temperature regimes has been noted along the southern coast of 
South Africa at 34°S (van Bree et al., 1978; Ross, 1984). The Atlantic coast of 
South Africa is characterised by the cold-water Benguela Current, while 
the warm-water Aghulhas Current flows along the Indian Ocean coast, 
causing an area of mixing along the southern coast. Accordingly, the bulk 
of the G. macrorhynchus strandings are reported from the eastern section 
of this coast, while the bulk of the G. melas strandings were found to the 
west, with an area of overlap between (van Bree et al., 1978). There was no 
differential seasonality for the occurrences of the two species, suggesting 
both may be resident (van Bree et al., 1978). Around Australia, both species 
have been reported to strand on Tasmania at 41 °S (Nicol and Groome, 1988), 
but the short-finned pilot observations were only of a single animal. The 
most southerly mass stranding for the shor~-finned species was off 
northern Australia, out of the reported range for long-finned pilot whale 
(Baker, 1983). There is currently insufficient information about the 
distribution of the two species to determine the extent of overlap off South 
America in the south Pacific. 
49 
the two species overlap between 36° and 38°N (Sergeant and Fisher, 1957; 
Paradiso, 1958; Sergeant, 1982). The reports from the western Atlantic are 
all of stranded animals, so there is little information on variations in 
distribution relative to oceanographic features. However, the warm-water 
Gulf Stream would be expected to allow G. macrorhynchus to penetrate 
north. Correspondingly, when the Gulf Stream shifts to more offshore 
flow, it would be expected that G. melas could penetrate further south with 
the influence of the cold Labrador current. In any event, current 
information suggests a wider zone of overlap along the eastern Atlantic 
compared to the western Atlantic. 
In the southern hemisphere, a zone of overlap distinguished by 
varying temperature regimes has been noted along the southern coast of 
South Africa at 34°S (van Bree et al., 1978; Ross, 1984). The Atlantic coast of 
South Africa is characterised by the cold-water Benguela Current, while 
the warm-water Aghulhas Current flows along the Indian Ocean coast, 
causing an area of mixing along the southern coast. Accordingly, the bulk 
of the G. macrorhynchus strandings are reported from the eastern section 
of this coast, while the bulk of the G. melas strandings were found to the 
west, with an area of overlap between (van Bree et al., 1978). There was no 
differential seasonality for the occurrences of the two species, suggesting 
both may be resident (van Bree et al., 1978). Around Australia, both species 
have been reported to strand on Tasmania at 41 °S (Nicol and Croome, 1988), 
but the short-finned pilot observations were only of a single animal. The 
most southerly mass stranding for the short-finned species was off 
northern Australia, out of the reported range for long-finned pilot whale 
(Baker, 1983). There is currently insufficient information about the 
distribution of the two species to determine the extent of overlap off South 
America in the south Pacific. 
49 
I 
I 1 
I 
!I 
I 
I 
I 
2.6.2 Feeding Habits 
Pilot whales are primarily considered to be squid feeders, although 
they will occasionally feed on fish (Sergeant, 1962a; Mercer, 1975; Seagars 
and Henderson, 1985; Overholtz and Waring, 1991; Gales et al., 1992). This 
information comes from · stomach contents as well as from correlated 
distributions of populations with the distribution of squid. I will first review 
the reported stomach contents for Globicephala and then examine the few 
distribution studies. Finally, I will consider some of the implications of 
squid feeding for an understanding of the behavioural ecology of pilot 
whales . 
Pilot whales are commonly found stranded on beaches throughout 
the world (Sergeant, 1982; Klinowska, 1986). This has provided an 
opportunity to collect stomach contents to examine the animals' feeding 
habits. There are problems with potential bias arising in this sort of 
analysis because the representativeness of the sampling methods in 
unknown. Animals that strand alone are often diseased and may not have 
fed for a long period of time (Klinowska, 1986). Information on feeding 
habits collected from these animals may be biased. However, it can be 
argued that in the case of mass stranding (Klinowska, 1986), many of the 
animals are perfectly healthy who can provide accurate information on food 
habits, although the possibility of group-wide disease cannot be ruled out. 
Som e additional infor mation on feeding habits h a s come from the 
incidental capture of pilot whales* in offshore fishing operations along the 
* Although the most common pilot whale of this area is the long-finned pilot whale , this 
area of capture is an area of potential overlap between G. rnelas and G. macrorhynchus 
I 
continental slope of the western north Atlantic between 35°N and 43°N 
(Waring et al., 1990; Overholtz and Waring, 1991). Pilot whales represented 
55% of the marine mammal incidental catch (N = 297 animals) from 1977 to 
1988 (Waring et al., 1990). This potentially provides a more representative 
sample of dietary habits. 
Long-finned pilot whales captured off the coast of Newfoundland 
were found to feed predominantly on the short-finned squid Illex 
illecebrosus (Sergeant, 1962a). During the first five years of the study, only 
squid were found in pilot whale stomachs. In the following year, squid 
fishermen reported a disappearance of squid from inshore waters where 
the whales were captured. Pilot whale catches dropped, and the few 
whales that were caught were found with cod (Gadus morhua) or 
Greenland turbot (Reinhardtius hippoglossoides) in their stomachs 
(Sergeant, 1962a; Mercer, 1975). This trend continued for two years, until 
squid began to reappear in the whales' stomachs as well as in the 
fishermen's catches (Sergeant, 1962a). More recently, the offshore 
distribution of pilot whales along the continental slope of the northeastern 
United States (39° N latitude) has been shown to coincide with the 
distributions of Atlantic mackerel (Scomber scotnbrus) and butterfish 
(Peprilus triacanthus), as well as with the distribution of both long-finned 
(Loligo peali) and short-finned squid (Smith et al., 1990; Waring et al., 1990; 
Overholtz and Waring, 1991). Mackerel dominated the diets of four pilot 
whales caught in fishing trawl nets during a 30-day period in March and 
(see Section 2.6.1). Thus, unless examined by skilled observers, the possibility exists that 
some reports were of the short-finned pilot whale. 
51 
1, 
i I 
I 
I! 
'' 
April, comprising an average of 71 % of the wet weight of stomach contents 
(Overholtz and Waring, 1991). Long-finned squid comprised the remaining 
29% of the four animals and 100% of a fifth animal (Overholtz and Waring, 
1991). However, the quantities of prey observed in the stomachs were 
estimated to supply only 1.8 - 28% (mean = 9.2%) of the daily food 
requirements of the animals (Overholtz and Waring, 1991). Pilot whales 
also tended to be caught more by mackerel fishing boats (85% of the total 
pilot whale catch of 297 animals) than by long- and short-finned squid 
fishing boats (13.8% of the catch) (Waring et al., 1990), although this may 
not necessarily relate to dietary preferences because the wide opening of 
pelagic mackerel nets could "corral larger delphinid species such as pilot 
whales" (Waring, 1990, p.357). 
Feeding rates for captive short-finned pilot whales are limited, but a 
5.26 m male ate 45 kg of squid and mackerel per day, a > 4.0 m female ate 36 
kg/day and a 3. 7 m female ate 18 kg/day (Norris in Sergeant, 1962a). This is 
an average daily feeding rate of 3% of the body weight. This suggests that 
large animals may need to feed two to three times per day to obtain similar 
quantities of food to those of captive animals. Of course, rates of feeding in 
. captivity may not represent feeding rates in the wild, where food is not 
available ad libitum. 
There did not appear to be any diurnal pattern in feeding for long-
finned pilot whales caught off Newfoundland. Observations were limited, 
but animals were observed feeding prior to capture and were documented to 
have fresh stomach contents during daylight hours. Whales could be heard 
close to shore during the night, implying they were feeding then as well. 
This is different from the pattern of predominate night-time feeding 
reported for G. macrorhynchus in captivity (Kritzler, 1952). 
52 
I! ' 
I I 
2.7 SOCIAL ORGANISATION 
Kasuya and Marsh examined the school structure of 27 short-finned 
pilot whale schools caught in a drive whaling operation off the southern 
coast of Japan (Kasuya and Marsh, 1984). Data on body length, age, and 
reproductive status were examined from the carcasses after the hunt. Pre-
hunt sighting information on the numbers of distinct groups within these 
schools was available for 16 of the schools. Some of the schools had been 
isolated from larger aggregations by the whalers before driving, while other 
schools were sighted and driven as a single school. Thus, there was 
information on the structure of individual schools as well as the patterns of 
school aggregation (Kasuya and Marsh, 1984). 
School size for all schools ranged from 14 to 52 (mean= 30.9 whales). 
The eight schools located alone ranged in size from 14 to 38 whales (mean = 
24.6 whales), while the seven schools selected from aggregations were 
larger (range: 20 to 52 whales, mean = 35.1). This could have been because 
the whalers selected the largest schools from aggregations (Kasuya and 
Marsh, 1984). 
The number of mature males in all 27 schools ranged from zero (one 
school) to 18 (one school). When these two schools were removed from the 
analyses as outliers, the number of adult males ranged from one to eight. 
In the solitary schools, the number of males ranged from one to three. In 
contrast, the selected schools had from one to eight adult males, with four 
of seven selected schools having four or more adult males. This is further 
evidence of selection by whalers, in this case selecting for adult males who 
were prized for their larger size (Kasuya and Marsh, 1984). The overall 
ratio of mature females (excluding post-reproductive females) to mature 
males for 11 schools with complete data ranged from 1.1 to 21.0 females per 
53 
I' 
male. All of these schools had pregnant and lactating females, indicating 
they were breeding schools. The one school with 18 mature males had five 
pregnant females, and one in oestrus. In attempting to explain the large 
number of males, Kasuya and Marsh (1984) discount the possibility that 
males congregate around oestrus females by pointing out that other schools 
had greater numbers of females in oestrus. There were no significant 
correlations with the proportions of other school members which could 
satisfactorily explain the observed distribution of adult males in the schools 
(Kasuya and Marsh, 1984). Kasuya and Marsh conclude: " ... either mature 
males stay in the same breeding schools for a period exceeding one female 
breeding cycle, or that their movement between breeding schools is 
controlled by (unknown) factors other than the female reproductive cycle" 
(Kasuya and Marsh, 1984). 
The occurrence of maturing males also varied widely between 
schools: one school had five of the 12 maturing males. Kasuya and Marsh 
suggest this uneven distribution could be because of immature males 
aggregating at puberty; leaving their natal school and associating with 
similar aged males (Kasuya and Marsh, 1984). However, there were no 
observations of all-male adolescent groups, indicating that even if 
adolescent males were joining together, they were still members of a mixed 
age and sex school. Since the sample size was small, it seems that local 
variation in the ~ge and sex composition of schools could also be explained 
by only a few females having a slight preponderance of sons. However, 
there were no variations in the proportions of adolescent females in the 
schools. 
There did not appear to be any school-based synchronisation of 
oestrus: seasonal samples found females at a variety of stages in the 
reproductive cycle (Kasuya and Marsh, 1984). Along with the lack of strong 
54 
11 
seasonality to male testicular activity, this suggests that mating could be 
occurring over a wide time period. There was also variation in the 
numbers of post-reproductive females, suggesting they were not essential 
members of a school. 
Off Newfoundland, the majority of long-finned pilot whale schools 
were of mixed age and sex. Only two schools were found to have a majority 
of males and no school contained only males (Sergeant, 1962a). 
2.8 DISCUSSION 
There are both similarities and differences between the biology and 
behaviour of the two species of Globicephala. The two species appear to be 
distributed in different temperature regimes. This could be related to the 
ancient cooling of much of the worlds' oceans. The development of 
antitropical distributions is a widespread phenomenon of cetacean 
speciation (Davies, 1961). The cold-water long-finned pilot whales probably 
adapted to these changing temperatures, while the warm-water species 
found refuge in the warm-water core area of the Indo-Pacific region. This 
distinction persists today, even with the potential for sympatry. 
In spite of these distributional differences, the anatomical differences 
between the two species are very slight. There are no clear explanations for 
the observed variation in flipper length and skull morphology. In fact, the 
variation in flipper length is counter-intuitive, since cetacean appendages 
are known to be sites of thermoregulation (Hampton and Whittow, 1976). It 
would be expected that animals living in warmer waters would have longer 
appendages than those living in cold waters. The opposite fa the case. It is 
possible that the differences between the two species may simply be the 
result of genetic drift (Wilson, 1975). Pigmentation patterns are similar 
(Sergeant, 1962b; Yonekura et al., 1980), and differences fall within the 
range of variation reported between two populations of the same species 
(Kasuya et al., 1988a). 
Most important to the present study is whether there is evidence for 
significant differences between the social systems of the two species. For 
the short-finned pilot whale, there is convincing evidence for the presence 
of post-reproductive females (Kasuya and Marsh, 1984; Marsh and Kasuya, 
1984) and little genetic information to document the mating system. 
Genetic data from long-finned pilot whales suggests a unique mating 
system and pattern of dispersal (Amos et al., 1991a; 1991b), while analysis 
of the life history parameters from the same population is still underway 
(A.R.Martin, pers. comm.). It will be important to understand how 
applicable the findings from one species are to the other species. 
A comparison of the life history parameters from both species was 
limited to some extent by different histories of exploitation, which could 
have resulted in differing mortality rates (Kasuya et al., 1988b). Both 
species do exhibit differential mortality between the sexes, with adult males 
having higher mortality than similarly aged females (Kasuya et al., 1988b). 
However, the pattern of mortality was different for the two species: both 
sexes of long-finned pilot whales exhibited an increase in adult mortality 
around ages 21-25, while short-finned males showed this increase in 
mortality much earlier than females (28 yrs vs. 46 yrs). This resulted in 
different sex ratios between the species, with less female bias in long-finned 
pilot whales. Longevity appears to be much shorter in long-finned pilot 
whales, indicating an overall higher mortality rate for this species (Kasuya 
et al., 1988b). The difference in longevity was matched by an earlier age of 
sexual maturity for long-finned pilot whales (females: 6 yrs vs . 9 yrs, 
males: 12 yrs vs . 17 yrs) . There was no apparent age-dependent decline in 
56 
' 
I I 
,1 
pregnancy rate for long-finned pilot whales, indicating a lack of post-
reproductive females (Kasuya et al., 1988b), compared to the situation with 
short-finned pilot whales where females were rarely pregnant past 36 yrs 
and rarely lactating past 51 yrs, although they lived for a maximum of 63 
years (Kasuya and Marsh, 1984; Marsh and Kasuya, 1984). All of these 
results suggest that the long-finned pilot whale has a "less-specialised" life 
history than the short-finned pilot whale, with less differential mortality 
between the sexes, a sex ratio closer to unity, and less specialisation in 
female reproductive patterns in the form of a post-reproductive senescence 
(Kasuya et al., 1988b). 
Comparisons with other cetaceans will be reviewed more completely 
in Chapter 5, however, a few points can be made. In the sperm whale, the 
sex ratio in the breeding schools is also female-biased (Best, 1979; Gordon, 
1987; Whitehead and Arnbolm, 1987). However, there does not appear to be 
any significant difference in mortality between the sexes and longevity 
appears to be equivalent between males and females. The biased sex ratio 
arises from the geographical segregation of a proportion of males from the 
breeding schools (Ohsumi, 1966). This is very different from the situation 
with pilot whales, where male mortality is higher than female mortality, 
resulting in a shorter male longevity (Kasuya and Marsh, 1984). Also, the 
observations of segregated groups of males for either short-finned or long-
finned pilot whales are rare (Sergeant, 1962a; Kasuya and Marsh, 1984), 
and are certainly less than would be necessary to explain the female-biased 
sex ratios assuming equal mortalities for the sexes. The only indications of 
male immigration were a few schools which had more males than 
predicted from the average sex ratio, possibly indicating switching of 
schools. The conclusion on social organisation is that pilot whales live in 
groups of female-related kin, potentially including males as well. 
57 
2.9 SUMMARY 
This review of pilot whale biology and behaviour has examined six 
main features which are referred to in later sections of this thesis. First, 
aspects of external appearance are used in the identification of individual 
animals (Chapter 4). Second, an understanding of pilot whale taxonomy 
allows comparisons to be made between the two species (Chapter 5). Third, 
genetic studies are critical in a consideration of mating system hypotheses 
(Chapter 5). Fourth, sex differences in growth and body size are used to 
classify free-ranging animals into age and sex categories (Chapter 4). 
Fifth, life history parameters, specifically the differential longevity and 
patterning of reproductive effort in females are used to place Globicephala 
in a comparative context with the rest of Delphinidae (Chapter 5). Sixth, 
ecological elements of distribution and feeding behaviour are considered in 
the analysis of pilot whale distribution around the Canary Islands (Chapter 
3). 
CHAPTER 3: DISTRIBUTION OF SHORT-FINNED PILOT WHALES OFF 
TENERIFE, CANARY ISLANDS, SPAIN 
3.1 INTRODUCTION 
The geographical and seasonal distribution of delphinids has been 
related to prey distribution (Nishiwaki and Randa, 1958; Mercer, 1975; 
Condy et al ., 1978; Wursig and Wiirsig, 1980; Irvine et al., 1981; Lopez and 
Lopez, 1985; Heimlich-Boran, 1986), tidal currents (Felleman et al., 1991), 
underwater topography (Evans, 1971; Hui, 1979; Heimlich-Boran, 1988; 
Wursig et al., 1991), water depth (Saayman and Tayler, 1979; Wiirsig and 
Wursig, 1979; Wells, 1986; Ballance, 1990) and ocean temperature (Uda, 
1954; Mercer, 1967; Gaskin, 1968; Nishiwaki, 1975; Perrin, 1984; Kasuya et 
al., 1988; Mate, 1989). The correlation between delphinid distribution and 
these environmental characteristics indicates that these animals monitor 
their environment closely and are able to respond to fine-grained changes. 
Pilot whales are characteristically found in deep water, where they 
are considered to feed on their favoured prey, squid (Leatherwood and 
Reeves, 1983; Seagars and Henderson, 1985; Klinowska, 1991; Overholtz and 
Waring, 1991; Gales et al., 1992). Observations have been limited due to the 
pelagic nature of these whales, but some regularly-occurring populations 
have been identified (Sergeant and Fisher, 1957; Norris and Prescott, 1961; 
Mercer, 1967; Taruski and Winn, 1976; Shane and McSweeney, 1990). 
Many of these have been exploited by whaling operations and thus have 
been unavailable for consistent observations (Sergeant, 1962a; Caldwell et 
al., 1971; Mitchell, 1975; Kasuya and Marsh, 1984; Price, 1985; Gibson-
Lonsdale, 1990). A few reports have documented regular movements 
59 
within a confined area (Shane and McSweeney, 1990) but, in one case, along 
the channel islands of California, the whales distribution shifted abruptly 
during a year of warm-water intrusion (Shane, 1985). 
I conducted a 22 month study C>n the geographical distribution and 
seasonal occurrence of short-finned pilot whales off Tenerife in the Canary 
Island archipelago. 
3.2 METHODS 
3.2.1 Survey Methods 
Field operations were based in the town of Los Cristianos, on the 
southwest coast of Tenerife. Two vessels were used. The permanent 
research vessel was a 4. 7 m Zodiac inflatable. This was best-suited for close 
approaches to whales, photographic identification and under-·water 
observations. From 16 June to 2 September 1990 and from 31 March to 26 
June 1991, a 13.1 m ketch was chartered for 60 days. It was best suited to 
long-term observations and rougher sea conditions. At times, these two 
vessels worked together, covering a greater number of pilot whale groups 
over a wider area. 
3.2.2 Seasonal Occurrence 
The seasonal occurrence of whale and dolphin sightings was 
compared to sighting effort in order to quantify relative presence in the 
study area. Sighting effort was quantified in two ways: shore effort and boat 
effort. Both were dependent on observer availability, sea state, sighting 
conditions, and equipment function. Calculations of the proportion of 
sighting days per effort day (sightings per unit effort, or SPUE) were used to 
compare occurrence between months and between seasons. 
3.2.3 Mapping 
The geographical distribution of whales and dolphins in the study 
area was quantified by accurately plotting all travel routes. From the boat, 
locations were determined by compass triangulation on shore markers. 
These bearings were plotted on a map overlaid with 2 km square quadrats 
by computer using trigonometric calculations based on the locations of the 
shore markers. In order to categorise the steepness, or slope, of each 
quadrat a Contour Index (Hui, 1979), which is the percent change in depth 
within the quadrat, was used. This was calculated as: 
Cl= lOOMMm 
where m = the minimum depth and M = the maximum depth within each 
quadrat. Depths were estimated to the nearest 50 m from nautical charts. 
The average of the minimum and maximum depth was also used to 
. 
' 
categorise the quadrat. Whale usage of each quadrat was classified into 
(defined as vessel routes when accompanying whales) 
three categories based on the number of tracks: heavy use, moderate use I\ 
and light use. Light usage was defined as a quadrat with five or fewer 
tracks through it, moderate usage was 6 to 10 tracks and heavy usage was 
defined as having more than 10 whale tracks. 
3.2.4 Description of the study area · 
The archipelago of the Canary Islands extends over 500 km between 
270 37' and 290 23' N, and 130 20' and 180 16' W. At their closest point, the 
61 
islands are within 100 km of the coast of Morocco. There are seven main 
islands: Lanzarote, Fuerteventura, Gran Canaria, Tenerife, La Palma, 
Gomera, and El Hierro. The total area of the islands is approximately 7,273 
km2 (Schmincke, 1976). The Canary Islands are one of the major volcanic 
island chains in the Atlantic Ocean; there are a few active volcanoes, with 
the most recent eruption in 1971 on La Palma (Schmincke, 1976). The 
climate is dominated by oceanic influences of the north-east trade winds 
(Fernandopulle, 1976). About 80% of the rainfall occurs between October 
and March and about 60% in December and January (Fernandopulle, 1976). 
Figure 3.1 shows the generalised wind patterns off the west coast of 
Tenerife and Gomera. The calm waters on the southwest lee side of the 
islands are relatively consistent, although occasional shifts in the winds to 
the north create rough water conditions throughout the strait between 
Tenerife and Gomera. Temperatures vary between 180C and 280C, much 
cooler than would be expected by the latitudinal position of the islands. This 
cooling is caused by the cold-water Canary current which flows out of the 
north and generates upwelling off the African coast. The bulk of the 
upwelling occurs to the west of the islands (Boje and Tomczack, 1978), but 
some effect is still felt within the archipelago. 
Figure 3.2 shows a contour map of the study area off Tenerife and 
Gomera. The contour profile is steep, dropping from the peak of Mt. Teide, 
at 3718 m elevation, to ocean depths of over 2000 m. The main bathymetric 
feature is the canyon between the two islands. Tenerife and Gomera are 
the two closest Canary Islands (about 27 km apart) and the sea bed is just 
over 1000 m deep in the channel between them. The depths increase to over 
2000 m deep to the north and the south, forming a "saddlen in the channel. 
The slopes on either end of this saddle are steepest to the south and a 1000 m 
deep canyon is formed. The southward flow of the Canary current between 
6'2 
17"W 
Tenerife 
Figure 3.1. Generalised characteristics of trade wind patterns, indicated by arrows. Calm areas were in the lee of the islands of Tenerife and Gomera. Wind was a limiting factor in the distribution of search effort. 
80 
72 
64 
10 20 30 40 50 60 70 80 90 100 
Figure 3.2. Three dimensional contour map of the bathymetry in a 100 km x 80 km area around Tenerife and Gomera. Contour intervals are 250 m and sea level is indicated in bold. Elevations 
range from +3718 m (Mt. Teide on the right) to a maximum sea depth of -3200 m (lower left). The minimum depth of the saddle between Tenerife and Gomera of 1120 m forms a south-facing, 
closed canyon below this depth. 
+ I 
the islands is constrained by the channel and eddies and gyres have been 
noted. 
Stranding reports of cetaceans in the Canaries between 1980 and 1987 
recorded six species of dolphins: the common dolphin Delphinus delphis, 
the bottlenose dolphin Tursiops truncatus, the striped dolphin Stenella 
coeruleoalba, the rough-toothed dolphin Steno bredanensis, Risso's dolphin 
Grampus griseus, and the short-finned pilot whale Globicephala 
macrorhynchus (Vonk and Martin, 1988). Sperm whales, pygmy sperm 
whales and unusually high numbers of four species of beaked whales were 
also recorded. Since most of the animals stranded alive, it appears that 
these species are found in the vicinity of the Canaries, and were not simply 
carcasses carried by oceanic currents from distant areas. Globicephala, 
Tursiops and Delphinus are regularly seen nearshore and there have also 
been observations of killer whales, Orcinus area, in the same region (Herve-
Gruyer, 1989; 1990). 
3.3 RESULTS 
3.3.1 Seasonal Occurrence 
Field work was conducted from October 1989 through July 1991. 
Vessel effort totalled 1134.3 hrs over 200 vessel days (Figure 3.3). Vessel 
surveys were conducted at an average interval of 3.30 ± 0.29 days (all values 
mean± S.E., n = 199 intervals), with a maximum gap between surveys of 29 
days . Weather conditions occasionally prevented surveys, especially during 
December through February, when winter storms could occur for up to two 
weeks. Effort increased during two intensive periods: June-August 1990 
63 
(A) 160~----------------------. 
~ Total Effort Hours 
~ • Whale Observation Hours 
N = 1134.3 Hrs. 
N = 542.8 Hrs. i:: 
0 120 ~ 
i... 
Q) 
0-i 
r:n 
i... 
:::l 80 ~ 
c.+-. 
0 
i... Q) S 40 
:::l 
z 
0 
(B) 25 
~ 20 
i:: 
0 
~ 
i... 
Q) 15 0-i 
00 
~ 
~ 
c.+-. 10 0 
i... 
Q) 
s 
:::l 
z 5 
0 
~ Total Effort Days 
• Whale Sighting Days 
J J A S O N 
1990 
N = 200 Days 
N = 155 Days 
J J A S O N D J 
1990 
Figure 3.3. Summary of effort and observations in hours per month (A) and days per month (B). The seasonal increase in effort was when two vessels were 
used. Both vessels were used to calculate number of hours per month, while days were counted with both vessels combined. 
I I 
I 
:I 
I I 
I ' 
I 
'I 
11 
I 
, I 
and April-June 1991. These were also the periods when the large sailboat 
was chartered, allowing extended days at sea during rougher weather. 
Whales were located on 155 days and observed for 542.8 hrs of the 200 
vessel survey days (78%) and 1131.3 hrs of vessel effort (48%). The intervals 
between whale encounters ranged from one to 29 days, and averaged 4.26 ± 
0.41 days (n = 154). Summaries of the observation data relative to effort are 
presented for both hours and days in Figure 3.3. The data on observation 
hours represent the total hours of each boat, whether they were together or 
not, and thus tended to overestimate both effort and observations. Also, 
since hours included time returning to port from offshore whale 
observations, effort hours tended to be overestimated. The hourly data were 
not analysed further for seasonal occurrence, but they give some indication 
of the seasonal distribution of the whale observations. The primary effort 
was to collect photographs of all individuals and to maximise the time spent 
with whales. The amount of time spent with a group (before moving on to 
locate other groups) was dependent on the behaviour of the whales (e.g. 
respiration patterns, consistent direction of travel, avoidance behaviour). 
Because of this potential bias, data on relative seasonal occurrence in the 
study area was only tabulated for complete sampling days. 
The occurrence of pilot whale groups was examined for trends 1n 
seasonal occurrence. Whales were located in all 22 months of the study. 
However, monthly SPUE ranged from 0.43 to 1.00 observation days per effort 
day, with an average of 0. 77 ± 0.03 (Figure 3.4). Monthly SPUE values were 
pooled by month, three-month seasons, and for all seasons combined. None 
of these combinations showed significant seasonal differences (Kruskal 
Wallis ; month: H = 10.05, df = 11; season/year: H = 6.56, df = 7; pooled 
season: H = 3. 72, df = 3, P > .05). Sample size, however, had a highly 
significant effect on the variation, illustrated by a plot of the SPUE as a 
11 
I 
------~ -=H 
1.0 
ONDJFMAMJJASONDJFMAMJJ 
1989 1990 1991 
Figure 3.4. Plot of Sighting Per Unit Effort (SPUE) for the 22 months of the 
study. SPUE averaged 0. 77 ± 0.03 (mean ± S.E.). There were no significant 
seasonal differences for pooled months or seasons. 
function of the total number of effort days (Figure 3.5). Months with 
samples of fewer than 10 days showed a significantly greater variation in 
SPUE values, suggesting that 10 days was a minimum sample size 
necessary to quantify relative occurrence. Unfortunately, only months in 
the spring and summer had sufficient sample sizes (Figure 3.5), thus 
annual variation in occurrence could not be properly tested. However, the 
similarity between values for the spring and for months with sufficient 
sample sizes suggests that whale occurrence during this period is 
constant. 
3.3.2 Geographic Distribution 
A total of 252 quadrats (1008 km2) were covered while searching for 
pilot whales (Figure 3.6). The majority of the search effort was conducted 
within 10 km of the southwest coast of Tenerife (Figure 3.6). Additionally, 
eight transects were conducted to the southern coast of Gomera Island and 
five transects were conducted to the northwestern tip of Tenerife at 28° 20' 
N. However, whales were only located in 54, or 21%, of the quadrats covered 
(Figure 3. 7). Pilot whales were never documented in any of the surveys to 
northwest Tenerife and Gomera, although other species, such as bottlenose 
dolphins (Tursiops truncatus), common dolphins (Delphinus delphis), 
rough-toothed dolphins (Steno bredanensis) and Atlantic spotted dolphins 
(Stenella fronta:lis) were all located in these areas. This indicates a 
preference by pilot whales for the waters off southwestern Tenerife . 
Thus, the total area of observation for these whales was 216 km2, 
between 27°58'N and 28°08'N, and 16°4l'W and 16°52'W. The average depth 
of the 54 quadrats was 1386 ± 70 m and the average Contour Index was 34.0 
± 3.0. However, the whales were not observed equally in all of the quadrats. 
Categorisation of quadrats by the relative whale usage showed that only 13 
65 
11 
I 
I 1 
I I 
1 • A• 
.9 
A 
• 0 A 0 
. 8 • 0 • ~ 
~ 
• ~ .7 0 U1 
• 0 
• 
.6 
• • Fall 
• Winter 
.5 • o Spring 
A Summer 
• 
.4 
0 5 10 15 20 
Number of Days Per Month 
Figure 3.5. Plot of the Sighting Per Unit Effort (SPUE: whale sighting days/effort days) as a function of the number of effort days for each of the 22 months of the study. Months are displayed for each of the pooled seasons: fall (Oct-Dec), winter (Jan-Mar), spring (Apr-Jun) and summer (Jul-Sep) . Months with fewer than 10 effort days showed the greatest variation in SPUE. 
11' 
; I, 
I I 
I Ii 
I 
I I 
I I 
1 1, 
11°w 
.. ··++·· ·++·: ... ·-+·· ... ··++·· ··+-·+ ... :·+· ... ··++·· ·++· ... ·+· ··f+:,k: ::+:+:: ... :.+: ::: :::+-+··· ·+:+·· .. .1 .. +:: 
: .. t : . . : :; . . . . : : : :: ·. ··:· .. ,: .. ·· . . ...;· : /: .:' 
Tenerife 
Los 
Cristianos 
~ ----y,o 4 a 12 1§ 20 ·. i ; . . .. i ..... ·.·.t .. ,i .... ·.·.·.i .·.·vj ... ... . ·.· .·., . ··.· ... ··· ·'.···':··· .· .. . '·· ·':··· ... . ·.·.·,:_··_··.· ... .. ·., .. !. .. ',! .... ... ', .• ~L, .. .. ···': ··· :: ::::ri • . I : ~~met'ef~L~:. : • : :::r:+:: .:: ·,+: ... ··++·: ·++·· ... ··+· .... :r+·· ···)·-+·· ... ··+· .:. ··+:f. :·f F· ... ··-:-:· 
I : : I ~: I I : . . · · · · . . . . ·· ... : ! : : .. : 
11°w 
Figure 3.6. A summary plot of all vessel effort tracks. Most of 
the effort was concentrated off Tenerife, but eight transects were 
conducted off south Gomera and five were conducted to the 
northwest tip of Tenerife. 
17°w 
... ··++·· ·++: ... ·+·· .. ·++·· ·+·/· ... :·+· ... ··++·· ·++· ... ·+· ··-:-:+:+:: ::+:+:: ... ··+: ::.: ::+·+··· ·++:·· ... ·00:::: 
: . . : : .: . . . . : : : :: ._··:·- -- :···· . . . -· ·:· : / : .:-
::; :::(1::: :::j:::f::: ::: ::J:: ::: ::::f J::; ::}t:: ::: :j::: /: :j:::1: :: :::1:::1::: ::: :::J:: :'.: tt'.f;; :::[::f::: ;:;;t: '.:: :::tJ::: J;;t·· .. : : .. ( 
. . : ,..... . , .. ..,-... .,- ...... . . : . · .. · ... ..: ,· ...... , .. :. .. . .. ;/ . . ..-::,··· .r ~ ........, -
::. ·,+,+:· ··+·+::: :'-·r,r:: ::: ,::r+.:: :++'' ,,, ,,+·· ... : '..:r~r:: ·'+·+.:: ::· ::+· .:, .i++·· ~ 
: ... ·=· .... ,.,:.::-·:: ..... · . : ... ·,··": .... , : : ... , ... ; ··.. . . . : .• , •• ,. : '?'I 
... :/\+.:: :++·· ... ··+': : .. ··+:+:: ,:+·+:: ;:; ··+·· ::. ::++·'· ::\+:: ... ·-+· ... ·-+·+:: ::::·+~ 
·-.... :·: .. ,· ·: . ...- : :··, . . : .. .. 
...... r:f :: .:J:.t ::: .::J:., .: .. ::i. .. i·-· ... i . .:. -J ..... ... J .. r.: .:(l, . ::: :.:r: .:: ~,;1...1 ... -·~>i::: :: :. \ Tenerife 
. . ,···.; : ,: :· : : / \: : . . :·: : : : •·. \ ·. \ 
··  ··+·>· ·+'fr:: ··· <h ··· ··+-+·· -" ·++· ··· :+. : .. ::+·+·' ::+·+.·· :,_: .:)··· , 
(. ; : .:: •":"·, :':. \ 
Gomer a t-·· ... ··+: ·:: ···-fY·J· ·++~ ., ... 
! : :. :: .: 
... :.+·}} ··+:+:· :·: ... :;,.: :-· :-+·!-1 
: : ":.. : . . \ : : I\ /. : .. • . . ··· : .. · .,. :: .·· ., .... , 
~ :::f :,:1··: ::: ,:t:: ::: :<t::r:. ·t-t-·· ~- ·. 
28°N ·· ... :. : ···.: ... . : ··· ... ::.._ : ··· .. _':· . . :··... ...-,./.: . ...- : : ... · ... . ·· . . . . ,r 
... =·+·:::., ::::.·,+·· ... ·,+, :,, .:·,+:\'.· ,+:+:, ,,+Jr: ;:, ::,r:+:: ,::::,:y .. ::: ::+,, :; ··++,: ·++:, ,,: :··15 t . 
: :. . . :':. : ·,· ... ··· .. : ·,... : .· ..., ... ·, .:' . . : . .. . . /:'··. ·. ·.·, .... ;.... . .. ·, 
I : : ,-, , -
11°w 
Figure 3. 7. A summary plot of all vessel tracks while following 
whales. The majority of whale sightings were centered over the 
steep slopes along the 1000 m. depth contour. 
\ I' 
11 
11 
I 
of the quadrats (25% of all quadrats where whales were observed) were 
heavily used, 14 (27%) were moderately used and 25 (48%) were only lightly 
used. Heavily used quadrats had significantly shallower average depth 
(1151 ± 74 m, Kruskal Wallis H = 6.00, df = 2, P < .05) than moderately used 
(1365 ± 128 m) or lightly used (1520 ± 117 m) quadrats (Figure 3.8). Areas of 
heavy usage also had significantly larger Contour Indices (heavy: 43.2 ± 
3.2, moderate: 34.6 ± 6.4, light: 28.9 ± 4.7; Kruskal Wallis H = 7.40, df = 2, P < 
.02), indicating that the whales preferred areas of steeper slope. These data 
suggest that the whales are using specific portions of the Tenerife offshore 
habitat. 
3.4 DISCUSSION 
Numerous studies have documented an association between 
delphinids and underwater topography. My own studies found that killer 
whales commonly foraged over area of high relief (J.Heimlich-Boran, 1988). 
It appeared that they were using underwater seamounts (which were 
necessarily also areas of high slope) as a barrier against which to herd fish. 
Salmon also tended to collect in these areas during ebb tides to avoid losing 
ground on migrations to their natal rivers (Felleman et al., 1991). Common 
dolphins have been well studied off the southern California coast (Evans, 
1971; Hui, 1979), where there are a series of underwater escarpments. 
Dolphins collect over these areas, which are known to be areas of upwelling 
and nutrient mixing that support the anchovy food chain (Hui, 1979). Thus, 
distribution can provide some indication of feeding ecology. 
Pilot whales appear to be similarly localised in areas of high relief 
along the 1000 m depth contour off southwest Tenerife. This is also known 
for upwelling. Drift buoy studies have mapped out the main current flows 
66 
I, 
11 
111 
I 
10 
Im Light 
Ill Moderate 
8 
• Heavy 
0-499 500-999 1000-1499 1500-1999 2000-2499 
Average Depth (meters) 
Figure 3.8. The average depth of quadrats plotted by frequency of whale 
usage. The three categories were significantly different (H = 6.0, df = 2, P 
< .05) with heavily used quadrats shallower (1151 ± 74 m) than either 
moderate (1365 ± 128 m) or light (1520 ± 117 m) usage quadrats. 
of this area (Molina, 1979). The predominant current through the area, the 
Canary Current, enters the strait between Tenerife and Gomera, forms a 
counter clockwise gyre current which starts mid-strait at the midpoint of 
Tenerife, flows southeast, and then circles north and offshore again at the 
midpoint of the island. The area of this current flow corresponds to the 
main area where pilot whales were observed. Unfortunately, the missing 
element in this picture is how this current affects local prey resources. It is 
likely to generate upwelling and mixing as it turns away from the coast, but 
there are no fisheries in the area which can provide any indication of the 
"' affects on prey resources. Thus, until further data become available, it can 
only be said that pilot whale distribution was correlated with the steep 
slopes along the 1000 meter depth contour where a gyre current is 
potentially generating upwelling. 
Behavioural evidence of the importance of this specific area comes 
from observations of the whales at the boundaries of the area. During six 
such observations, the whales slowed their travel and spread out, 
ing 
eventually rejoin,. and moving off in the opposite direction. These "turn-
arounds" have also been observed in killer whales and were correlated with 
the occurrence of slack tides, when the tidal current changed direction 
(Felleman et al., 1991). It could be that similar current processes are 
affecting the pilot whales. The implication is not that the whales are 
somehow limited by currents, but that currents affect prey distribution and 
the whales may pay attention to them, perhaps even use them to locate or 
predict prey concentrations. Recent telemetry studies of squid, which are 
pilot whales' preferred prey (Mercer, 1975; Seagars and Henderson, 1985; 
Clarke, 1986; Overholtz and Waring, 1991), have shown they use upwelling 
currents to reduce the energetic costs of locomotion, in a similar way that 
raptors use updrafts for soaring (M.Wells, pers.comm.). 
I I · 
The strong bias in the distribution of effort does not allow any firm 
conclusions on pilot whale distribution in other areas. Although surveys 
outside of this area were often limited by the trade winds, many other 
species of delphinid, but not pilot whales, were located in the 800 km_2 which 
were surveyed off Gomera and northwest Tenerife. Anecdotal reports from 
daily dolphin-watching boats along the northwest coast of Tenerife, which 
have run trips two times a day for the last two years, have never reported 
pilot whales. Sailing clubs, which run regular 7-day trips from Tenerife to 
La Gomera, La Palma, El Hierro and Gran Canaria, have never reported 
sightings of pilot whales. 
3.5 SUMMARY 
The distribution of short-finned pilot whales in the Canary Islands 
was studied over a 22 month period. Whales were located on 155 days 
during all months and there were no significant seasonal differences in the 
sighting per unit effort (SPUE), although months with sample sizes less 
than 10 days have large variance in SPUE. All pilot whale observations 
occurred within a 200 km2 area, with an area of heavy usage of 50 km2. The 
whales preferred areas of steep slope along the 1000 m depth contour. Pilot 
whales were never documented in any of the outlying surveys to northwest 
Tenerife and Gomera. The area of whale use was characterised by a deep 
canyon and current patterns which could be generating upwelling. 
Further studies should focus on gathering information on prey distribution 
in this area. 
CHAPTER 4: SOCIAL ORGANISATION OF SHORT-FINNED PILOT 
WHALES (Gwbi.cephal,a maerorhynchus) 
4.1 INTRODUCTION 
Field studies of free-ranging mammals, utilising methods of 
individual identification, have greatly enhanced our understanding of their 
social systems. Research into the social lives of primates (Goodall, 1986; 
Stammbach, 1987), carnivores (Schaller, 1972; Packer, 1986; Packer et al., 
1988), ungulates (Clutton-Brock et al., 1982) and elephants (Moss and Poole, 
1983; Moss, 1988) have shown that individuals vary widely in their 
behaviour. All individuals ultimately attempt to maximise their 
reproductive success through individual strategies of resource and mate 
acquisition, but there are differing costs and benefits for the various age 
and sex classes of a population (Clutton-Brock, 1986). For example, in most 
polygynous mammals, males disperse from their natal group at maturity 
and search for mating opportunities elsewhere as a means of avoiding 
inbreeding, while females remain with their group and breed in 
association with their female relatives (Greenwood, 1980; Pusey, 1980; 
Shields, 1982; Greenwood, 1983; Moore and Ali, 1983; Shields, 1983; Pusey 
and Packer, 1987b). Thus, males in such groups can be expected to have 
very different social associations from females. An understanding of the 
social networks of associations and relationships is essential in defining 
social structure and mating system. 
Recent studies of large members of the cetacean family Delphinidae 
have demonstrated unique patterns of dispersal and association. Long-
term observational studies of killer whales, Orcinus area have revealed a 
I I 
total lack of dispersal of either sex from natal groups, with males 
remaining in the social group of their mothers, sisters and related 
offspring into adulthood CS.Heimlich-Boran, 1986; S.Heimlich-Boran, 1988; 
Bigg et al., 1990; Olesiuk et al., 1990). The role of males in these groups is 
still not well understood. Adult males have a lower degree of association 
with members of their pod than females and immatures (S.Heimlich-
Boran, 1986; 1988). There is a multi-level social organisation, with groups 
of pods forming isolated communities (Bigg et al., 1990). Genetic studies on 
the same populations show relatively high degrees of inbreeding at the 
community level, while finding indications of reproductive isolation 
between communities (Stevens et al., 1989; Hoelzel and Dover, 1990; Hoelzel, 
1991a). Long-finned pilot whales (Globicephala melas) appear to have some 
similar elements in their social system: genetic analyses have found that 
males were not the fathers of the foetuses in their own pods, some pods 
were found to be closely related, and there is also some degree of 
reproductive isolation between groups, implying well-defined patterns of 
genetic interchange (Andersen, 1988; Amos et al., 1991a; 1991b; In press). 
These studies all raise questions as to the nature of the mating system of 
these large, sexually-dimorphic delphinids. 
In the following section, I report the results from a 22-month study of 
the social organisation of free-ranging, short-finned pilot whales 
(Globicephala macrorhynchus) off the western coast of Tenerife in the 
Canary Island archipelago from October 1989 through July 1991. The 
primary goals of this study were to identify in,dividual pilot whales, 
examine the patterns of associations between individuals, determine the 
levels of social organisation defined by these associations and finally to use 
this evidence to examine the hypothesis that mating may occur outside of 
the core social groups. 
70 
I''', 
, I 
4.2 ME'IHODS 
4.2.1 Photographic Sampling Methods 
The photographic identification of individual pilot whales was the 
primary initial goal of this study and formed the basis for all other 
analyses. The majority of the field effort consisted of trying to photograph 
as many pilot whales as possible. The aim was to photograph all 
individuals in a group and then move on to photograph another group. 
Logbooks were used to maintain an ad libitum record of events (Altmann, 
197 4). Information recorded included start and end times of vessel trips, 
course information such as heading and periodic compass triangulation on 
landmarks (as described in Chapter 3), start and end times of each pilot 
whale group observation, comments on all other pilot whale groups 
observed from the vessel out to the horizon (within a 2 km radius with 
binoculars), and general notes on behaviour. Once a group was located, the 
vessel slowly approached from the side and slightly to the rear. A constant 
throttle was maintained because it was found that sudden changes in 
engine noise could startle a group, causing it to dive. The best conditions 
for photography were when the vessel approached to within 20 m, but it was 
often possible to approach individuals to within 1 m. 
Black-and-white photographs and colour slides were taken using a 35 
mm Nikon FE-2 camera equipped with a 70-210mm zoom lens and a Nikon 
F4 camera equipped with a 80-300mm auto focus zoom lens. Both cameras 
were equipped with motor drives to take photographs at three to five frames 
per second and data backs for imprinting date and time on each frame of 
film. Underwater photographs were taken with a Nikonos V camera using 
a £2.8 35mm lens and an fl.4 28mm lens. 
71 
4.2.2 Photographic Analysis 
All negatives were examined visually for identifiable whales. 
Negatives from the first nine months of the study were examined under a 
20x compound microscope for initial identifications at the end of each field 
day. After returning to Cambridge, all films were examined using a 
Photovix II Film Video Processor, which allows negative or positive images 
to be relayed to a video screen for magnification and viewing. This viewing 
of negative materials was the most efficient for the initial location of 
identification photos. When high quality images were located of an 
individual, black-and-white prints were made for an identification 
catalogue. 
Initially, it was necessary to determine what constituted a "usable" 
identification photograph. This "photographic quality" must be determined 
independently of the "recognisability" (defined below) of the identification 
characteristics of the animal being photographed in order to quantify the 
proportion of usable photographs versus the proportion of re-identifiable 
whales (Hammond, 1986; Mizroch et al., 1990). The photographs were 
classified into three quality categories: 1) excellent, 2) good, and 3) poor. Six 
basic elements of photographic quality were evaluated: 1) proper exposure 
with good contrast, 2) focus, 3) lighting conditions, such as glare and 
backlit silhouettes, 4) proportion of the frame filled by the whale's image 
(i.e. the distance of the whale from the camera), 5) angle of the longitudinal 
axis of the whale to the film plane of the camera, and 6) the amount of the 
whale's body visible above the water line. Obviously~ as focus decreased and 
the distance between the camera and the whale increased, the ability to see 
identifying characteristics decreased. Lighting conditions were important 
considerations for the visibility of faint scratches and pigmentation marks. 
The angle between the whale and the film plane of the camera was more 
72 
difficult to evaluate and could introduce error in identification due to 
apparent changes in the relative location and spacing of identifying 
characteristics. These elements were evaluated visually. Cues for whales 
not travelling parallel to the film plane were the angle of the water line 
relative to the horizon and the ratio of dorsal fin height to width. 
Photographs of these whales could be identified by sloped water lines 
(relative to the horizon) and narrowed dorsal fins. Excellent photos were 
those with proper exposure, sharp focus and direct lighting from a sun at 
45° declination behind the photographer on full frame images of a whale 
high out of the water with the transverse axis of the whale parallel to the 
film plane. All types of identifying characteristics could be recognised in 
these quality of photographs. Good photographs were slightly lacking in 
some element of photographic quality, but were still potentially usable as an 
identification photograph. Poor photographs were blurry images of distant 
whales and were not used. 
4.2.3 ID Characteristics 
Whales were identified on the basis of naturally occurring marks 
and scars on the dorsal fin and back. Some of these marks were 
pigmentation patterns. Short-fin pilot whales in the eastern tropical 
Atlantic do not have distinct "saddle patches", or areas of light grey located 
along the backbone just posterior to the dorsal fin, as has been shown for 
pilot whales in other parts of the world (Yonekura et al ., 1980). However, 
they do have distinct post-orbital blazes which rise up from the eye to the 
anterior edge of t he dorsal fin . These marks blend into a fainter saddle 
patch which tapers down along the backbone to the caudal peduncle. In 
excellen t lighting condit ions, it was possible t o use irregularities in the 
pattern of these pigmentation marks to distinguish individuals . However, 
73 
I.' 
the most common characters used for identification were nicks in the 
trailing edge of the dorsal fin and scratches along the side of the body just 
below the dorsal fin. The most likely cause of these nicks and scratches 
was the teeth of conspecifics. It is unknown in what behavioural context 
these scars were received, whether in play or aggression. They were found 
on all age/sex classes. Additional scars were caused by encounters with 
larger objects, probably small boat propellers, resulting in large chunks 
removed from the back and caudal peduncle. One juvenile appeared to 
have been bitten by a large predator, such as a shark or killer whale. 
Identification characteristics were classified into 24 categories based 
on dorsal fin shape (especially hooked or with a hump on the leading edge), 
number and location of nicks (cuts in the trailing edge of the dorsal fin) 
classified by location by dividing the dorsal fin into thirds, dings (dents in 
the leading edge of the fin), and tabs (projections from the trailing edge of 
the fin) and the presence of barnacles. The categories (and their three letter 
abbreviations were: 1) Cleanfin: CLF, 2) Cleanfin w/ Scallops: CLS, 3) 
Cleanfin w/ Barnacles: CLB, 4) Unique Individuals: IND, 5) Dolphin-Like 
Fin: DLF, 6) Hooks: HOO, 7) Serrated Hooks: HOS, 8) Front Dings or 
Humps: FDH, 9) Raggedy w/ Fingers or Tabs: RFT, 10) Small Multiple 
Nicks: SMN, 11) Two Small Nicks: TSN, 12) Large Double Nicks: LDN, 13) 
Small Base Nicks: SBN, 14) Large Base Nicks: LBN, 15) Small Nicks Middle 
Third: SNM, 16) Large Nicks Middle Third: LNM, 17) Small Nicks Top 
Third: SNT, 18) Large Nicks Top Third: LNT, 19) Nicks Near Tip: NNT, 20) 
Square-Tips: SQT, 21) Flap Finger Tips: FFT, 22) Small Round Tips: SRT, 
23) Large Round Tips: LRT, and 24) Handles: HND. 
Each individual which could be classified into one of these categories 
had varying degrees of "recognisability". This was defined as the ability to 
repeatedly identify the same individual in photographs of different qualit y. 
74 
,I I 
,I 
,I 
I 
I 
I 
For example, an animal with the upper half of its dorsal fin missing would 
have a high recognisability, even in a poor quality photograph. On the other 
hand, a clean-fin animal, with a faint pigmentation mark could only be 
identified in the highest quality photo. Also, the more identifying marks an 
animal had, the greater was its recognisability. An animal with a single 
nick could be confused with another animal with a nick in the exact same 
location. With increasing numbers of marks, the probability of two animals 
sharing the same marks becomes infinitesimally small. Thus, the two 
elements of photographic quality and recognisability are interrelated and 
together contribute to the confidence in identifications. 
The initial identification of an individual, regardless of 
recognisability, required an excellent quality photograph which clearly 
indicated some set of identifying characteristics. This photograph was 
printed in 12 cm by 17 cm format and stored as a Master ID in a card 
catalogue file under the ID Type category. Subsequent negatives of that 
individual were all compared to this master photo. Identifications from all 
frames of film were entered onto computer to aid in cross-referencing 
multiple sightings of the same individuals. 
4.2.4 Unidentifiable Individuals 
The whales identified in this study represent only a portion of the 
whales observed. -Some individuals had no identifying marks and even the 
highest quality photograph would not provide any guarantee of repeat 
identification. In order to provide an estimate of these "clean-finned" 
animals, field count estimates of the total number of animals present in a 
group under observation were compared to the number of clearly 
identifiable animals recorded in the photographs of that same group. 
Differences between these two counts provided an estimate of unidentifiable 
75 
I' I 
J I 
animals in the group. This was expressed as a percentage of the total 
group count and averaged for all groups which had field count estimates 
made after a minimum 30 min observation (to assure sufficient time to 
make an accurate count). 
4.2.5 Age/Sex Classification 
Whales were classified into five age/sex categories: 1) adult males, 2) 
mothers, or females with calf, 3) other adults, 4) juveniles, and 5) calves, 
using relative photogrammetric methods (J.Heimlich-Boran, 1986b) and 
visual cues. Physical characteristics of pilot whale age and sex classes 
were derived from published measurement data (Yonekura et al., 1980; 
Kasuya and Marsh, 1984); see Chapter 2 for review) and were used to 
calculate expected relative measurements. The physical characteristics 
utilised in this study were relative body size, relative location of the dorsal 
fin along the body, and relative dorsal fin shape. Multiple photographs of 
identified individuals were examined for suitable relative photogrammetric 
features. The relative physical characteristics could be quantified for some 
individuals with measured photos, while for others the characteristics 
were determined qualitatively. Additionally, visual estimates of relative 
size were recorded in the field, especially for very large or very small 
animals. Other information was collected from the degree of association 
with different-sized animals over repeated observations. 
Adult males were identified by their large size relative to all other 
whales (see review in Chapter 2). The differences in growth and 
maturation are such that even young, newly-matured adolescent males are 
14% larger than a fully adult female which has attained her asymptotic 
length (4.14 m vs. 3.64 m). Males also differ from females in other body 
proportions, such as the relative position of the dorsal fin on the body, 
76 
dorsal fin width and the shape of the melon. Dorsal fin position was 
quantified from measurements available from 56 mature animals (10 
males and 46 females) published by Yonekura et al. (1980): 1) total body 
length, 3) tip of snout to blow hole, and 8) bottom of notch of flukes to the apex 
of the dorsal fin (numbers refer to Yonekura's measurement numbers). 
The distance from the blowhole to the apex of the dorsal fin (B-DF) was 
calculated by: B-DF = (1- 8) - 3; . The width of the dorsal fin at the base 
(measurement no. 16) was then expressed as a proportion of the calculated 
measurement, and termed the "blowhole/dorsal fin ratio" (BDFR). Males 
had significantly higher BDFRs than females (0.65 ± 0.04 vs. 0.58 ± 0.02; 
mean ± 95% C.I.; t = 4.15, P < .0001), indicating that their wide dorsal fins 
occupied a greater proportion of their backs. BDFRs greater than 0.60 
calculated from photographs of Canary Island pilot whales were 
considered to be an identifying characteristic of adult males. The dorsal 
fins of male Japanese pilot whales were significantly wider and taller than 
the dorsal fins of females, and the change was associated with male 
maturation (Yonekura et al., 1980). Three categories of dorsal fin shape 
were scored into three qualitative categories for each animal. The width of 
the dorsal fin at the base was scored relative to the distance from the 
blowhole to the leading edge of the dorsal fin. The height of the dorsal fin 
was scored in relation to the width of the fin. Finally, the presence of a 
hump along the leading edge of the dorsal fin just above the back was 
scored according to· its development. In order to be classified as an adult 
male, an animal must have scored in the highest category for at least two of 
the three characteristics. The melon of an adult male is characterised by a 
flattened anterior surface and square corners at the top and sides. The 
projection of the melon beyond the tip of the lower jaw differs between adult 
males and females. Melon projection for 28 mature females (lengths from 
3.2 to 3.8 m) ranged from O cm to 8.0 cm and averaged 3.6 ± 1.6, while melon 
77 
projection for 5 mature males ( 4.2 to 4. 7 m) ranged from 6 to 13 cm and 
averaged 9.3 ± 2.7 cm (Yonekura et al. (1980), Fig.3). These features were 
often visible, especially in underwater photographs. 
Mothers were defined as all full-sized animals with a repeated 
association with a small calf or juvenile. The calf had to be present on all 
observations of the mother. Occasionally, a calf had associations with more 
than one adult; in this case, the animal with which it associated for a 
majority of the time was assumed to be the mother. 
The category of unknowns was defined as all full-sized animals who 
did not have characteristics of adult males and who did not have a 
consistent association with a specific immature animal. This category was 
potentially composed of a combination of adult females and adolescent 
males. Comparisons of body length growth data for male and female short-
finned pilot whales in Japan allowed the possible age range of the 
adolescent males to be estimated. Females mature at 3.22 m and ceased to 
grow after 3.64 m (Kasuya and Matsui, 1984). Males of similar size to newly 
mature females would be approximately 8 yrs old, while males similar to 
fully-grown females would be approximately 12 yrs old. If the midpoint 
between these lengths (3.43 m) was considered to be the size at which a full-
size animal was distinguished from a juvenile, then males of this size 
would be approximately 10 yrs old (Yonekura et al., 1980). This was within 
the lower range of·body lengths and ages at which males were found to be 
capable of sperm production (early-maturing category of Kasuya and 
Marsh, 1984, see review in Section 2.5.2.1), although still less than the 
mean length and age at full maturity (4.14 m and 15 yrs). It has also been 
shown that males and females of similar body size do not differ in any of the 
characteristics used to identify adult males (Yonekura et al., 1980). These 
characters are only larger on adult males because of their continued 
78 
I' 
growth after females cease to grow. Thus, there was no way to distinguish 
these adolescent males from adult females without calves. Underwater 
photographs of the genital region were attempted, but were never 
successful. 
Immature animals were identified by their small size relative to all 
other whales. There were two categories of immature animals: juveniles 
and calves. Calves were the smallest class of immatures. Newborn calves 
approximately 
(up to/\one month old) could be identified by an undeveloped melon, which 
resulted in a lumpy appearance of the head in front of the blowhole due to 
approximately 
the visibility of the dorsal ridge of the skull. Calves 09ess than a year of age 
could be identified by the presence of foetal folds. These are the result of a 
lateral folding of the foetus in the womb and remain visible as creases appeared to on identified calves 
which/ade to lightly pigmented stripesAduring the first year of life. Older 
calves were defined in two ways: first, as those animals with a consistent 
association with a mother, which probably continues up to a minimum age 
of three years, the nursing period (Kasuya and Marsh, 1984), and second, 
as those animals which were less than two-thirds the size of the mother. 
Size and age correlations indicate a maximum of four years of age for an 
animal of this relative size (Yonekura et al., 1980). Juveniles were defined 
as immatures larger than two-thirds the size of the mother and smaller 
than adults. The ages of the apparent juveniles would have been different 
for males and females, because of the faster male growth rates. Female 
juveniles were estimated to be between 4-9 yrs of age, male juveniles would 
have been less than 8 yrs (Yonekura et al., 1980; Kasuya and Marsh, 1984). 
While it was usually possible to determine the mother of all calves, 
juveniles often had a wider variety of associates and the id1:mtification of 
their mothers was more difficult. 
79 
4.2.6 Group Definition 
The primary aim in defining pilot whale groups was to examine 
long-term patterns of association between identified individuals. The 
hypothesis was that pilot whales h~ve a multi-level society, with 
inseparable core groups of mothers and their offspring associating with 
other core groups to form relatively stable pods, which in turn associated on 
a less regular basis with other pods to form a clan. The key element in the 
definition of a group was its long-term stability. In this respect, no attempt 
was made to record subtle changes in individual associations. 
A pilot whale group was defined as all whales within 250 m of each 
other exhibiting similar behavioural characteristics, such as direction and 
speed of travel, timing of surfacing, and general activity level. This 
definition covered all situations. The majority of groups were more 
cohesive than this, consisting of animals travelling parallel to one another 
with a maximum of five body lengths (20 m) between adjacent individuals. 
The data set of photographs used for the identification of groups was 
restricted to those observations when times were recorded for each 
photograph (i.e. when the camera data backs were operating) and when log 
entries clearly described group formations in terms of inter-individual 
distances. At times, individuals were spread across wide areas, and it was 
not possible to defi:qe groups. 
Ultimately, groups were characterised by the identity of the 
individuals photographed. Although five separate groups may have been 
observed in the field, it was often the case that the same individuals were 
identified in more than one group. Groups with any common members 
were considered to be repeat sightings of the same group, and were 
recorded as a single group. For example, if whale 1 was seen with whales 
2, 3, and 4 in the morning, and later in the same day with whales 2, 5, and 
6, then the one group recorded for that day consisted of whales 1, 2, 3, 4, 5, 
and 6. This helped to assure complete sampling of the groups, although it 
also introduced a bias: changes of group membership by individuals during 
a day were disregarded. It was also necessary to assign an individual to 
only one group each day in order to fulfil one of the underlying assumptions 
of the association analyses described below. 
A unique number was assigned to each group, composed of the 
sighting day number and a sequential group number for that day. 
4.2. 7 Pair Associations 
The quantification of associations within a social group is an 
important aspect of the study of social behaviour. A variety of field studies 
have used measures of association to describe various elements of animal 
society (Schaller, 1972; Guinness et al., 1979; Clutton-Brock et al., 1982; Lott 
and Minta, 1983; S.Heimlich-Boran, 1986, 1988; Bigg et al., 1990). The 
original approaches to quantifying associations were in studies of 
community ecology (Dice, 1945; Cole, 1949)), which needed to describe the 
degree of association between species located together in a geographical 
area in order to define a "community". The application of association 
indices to the study of social organisation is more recent, and the original 
intention of the equations should always be considered in the newer 
application to animal societies. The common goal of any association 
analysis is to quantify the co-occurrences of some unit of interest . The 
degree of co-occurrence is then used to define population-wide patterns of 
association. 
81 
There are two main problems to be considered when quantifying 
associations; one is common to associations between species and 
individuals, while the second arises when applying the methods developed 
for species associations to the quantification of associations between 
individual animals. 
The shared problem for associations between species and individuals 
is the problem of the reciprocity of associations. An association between A 
and B actually has two components: A's association with B, and B's 
association with A. For an example, Ginsburg and Young (1992) consider 
the case of an association between a species of tree and a species of grass 
sampled from an equal number of woodland and plains habitats. Assume 
that the tree species was found in all woodland samples, while the grass 
species was found in both woodlands and plains. The association between 
grass and tree from the tree's perspective would be 100%, since they were 
only located together in the woodland samples. The association from the 
perspective of the grass would be only 50%, since the grass was found 
without the tree in all of the plains samples. Dice (1945) dealt with this 
problem by presenting two types of indices: one type to present the non-
reciprocal values described in the example above, termed "association" 
indices, and another type, termed "coincidence" indices, which presented 
association as an average for both A and B. However, this use of multiple 
indices was too unwieldy for most analyses and the emphasis has since 
been to generate a single index of association which fairly reflects the true 
association between two units of interest. 
The specific ways in which different indices deal with the problem of 
reciprocity will be discussed shortly, but first it is necessary to understand 
the problems which arise when applying species association indices to 
individual associations. The problem is essentially one of defining the unit 
82 
of sampling. When sampling for the co-occurrence of species on a plot of 
land, each plot (if randomly chosen) is a separate and independent sample. 
Two species found together in a sample are always considered to be co-
occurring. However, when sampling the social groupings of free-ranging 
animals, the unit of sampling must be defined (e.g. 15 min scan samples: 
(Altmann, 1974), but what ever the unit is, it would often be possible to 
locate more than one grouping of animals during this time period. Thus, a 
pair of animals may be located together in the same sample, and yet also 
occur in separate groups. A solution would be to treat each group as an 
independent sample. But if the groups are located closely together in time, 
the composition of the groups may be interrelated and dependent. In other 
words, it may be possible to predict the presence of one group from the 
initial location of some other group, but the samples would no longer be 
independent. Alternatively, the time period of sampling, although not 
instantaneous, must be short enough that changes in group memberships 
are unlikely during the course of sampling. 
There are three primary indices which are currently in use. One of 
these is Dice's "coincidence" index, termed the Half-Weight Association 
Index by Cairns and Schwager (1987). In order to understand the various 
approaches of these indices, it is necessary to use the common terminology 
(Table 4.1), presented in recent reviews of association indices (Cairns and 
Schwager, 1987; Qinsburg and Young, 1992). The equations for these 
indices are presented in Table 4.2. 
The need for unbiased indices of association is increased by another 
unique feature of sampling associations between free-ranging animals. It 
is not always possible to locate all known members of a study population in 
every sample. This could be an indication of differential probabilities of 
locating individuals and could introduce bias into the calculations. For 
83 
I I 
Table 4.1. Summary of association index notation (Cairns and Schwager, 1987). 
Types of samples 
N{E} denotes the number of samples in which event E occurs 
x = N{A and Bare located together in one group} 
y = N{A and Bare located in separate groups} 
Ya= N{only A is located} 
Yb = N{only Bis located} 
Yab = N{both A and Bare located separately} 
Y = Ya + Yb + Yab 
z = D* =N{neither A nor Bis located} 
n = N* = x + y + z (* terminology of Ginsburg and Young, 1992) 
Total sightings 
Of A: na = x + Ya + Yab 
Of B: nb = x + Yb + Yab 
Of A without B: Ta= Ya+ Yab 
OfB without A: Tb= Yb + Yab 
Of A and B together: Tt = x 
Table 4.2. Three common association indices. 
Half Weight Association Index= 1 x 
~na + nb) 
or 
X 
1 
X + Yab +~Ya+ Yb) 
Twice Weight Association Index = T x Tb or x 
- x + a + x + 2Yab + Ya + Yb 
Ratio Association Index = x +x y or x 
x + Yab + Ya + Yb 
example, conspicuously marked individuals or individuals which favour 
larger groups may be easier to locate. This is the reason for isolating the 
different types of "y" sightings, when the pair is not located together. The 
most important term in these equations (which is unique to studies of 
individual associations due to the unit of sampling problem mentioned 
earlier) is Yab- This is the count of the times a pair was located in the same 
sample, but not in association. Obviously, the greater this number, the 
weaker the association between the pair. The Twice Weight Association 
Index emphasises this term the most. 
Ginsburg and Young (1992) compared these three measures of 
association and considered the Ratio Association Index to be the least 
biased. The Half Weight Index tends to overestimate levels of association 
since it averages the counts of Ya and Yb, thus reducing the denominator. 
On the other hand, the Twice Weight Index tends to underestimate 
association since it double counts the samples in which members of the 
pair are located separately (Yab). The Ratio Association Index quantifies 
association as a simple proportion of the number of times a pair was seen 
together compared to the total number of samples in which either member 
of a pair was sighted. Cairns and Schwager (1987) found this index to be 
least biased when sampling was random in relation to the probability of 
locating a pair together versus when separate. 
The Ratio Index of Association was used for this study. Ginsburg 
and Young (1992) presented a simplified version of the Ratio Association 
Index equation presented in Table 4.2: 
Ratio Association Index = N ~ D 
The denominator of this equation is the total number of viewing periods 
minus the number of viewing periods in which neither A nor B were seen. 
84 
This is equivalent to the total number of days in which either A or B were 
seen. This equation was modified slightly to fit the data format of this 
study, as described below. 
Pilot whale groups were recorded as described in Section 4.2.6. All 
pair-wise associations between whales identified in the same groups were 
tallied once per day using a Microsoft QuickBasic© program. These were 
termed "group counts". Simultaneously, a tally of the total number of 
groups in which each individual occurred was recorded as "ID Sums". 
Pair-wise association counts were also made on a daily basis, disregarding 
whether the pair was seen in the same group on that day or not. These 
were termed "day counts". The difference between day counts and group 
counts equalled the Yab term (Table 4.1), or the number of days in which 
both members of a pair were observed in separate groups. The 
denominator of the Ratio Association Index equation was calculated as the 
number of days in which either member of the pair was seen. This was 
equivalent to the sum of the number of total days each member was seen 
(whether alone or together) minus the number of days they were seen 
together. The calculation of all terms used in the association equations is 
summarised in Table 4.3. 
4.2.8 Age/Sex Class Associations 
Associations between individuals were summarised by the age and 
sex class designations of the animals involved. Pair association indices for 
the various class combinations were pooled and compared for significant 
differences using the Kruskal-Wallis rank ANOVA test. Dunn's non-
parametric multiple comparisons were used to identify the primary 
sources of the differences (Zar, 1984). 
Table 4.3. Summary of terms and calculation methods used in pilot whale Ratio Association Index analysis. 
Computed Values: 
Group xab = number of group counts for each pair of whales Day xab = number of day counts for each pair of whales 
ID Suma = na = Day xab +Ya+ Yab = total number of days whale a was 
seen 
ID Sumb = nb = Day xab + Yb + Yab = total number of days whale b was 
seen 
. . . Group xab Rat10 Assocrnt10n Index = (ID Suma + ID Sumb) _ Day xab 
4.2.9 Cluster Analysis 
In order to define higher order levels of associations for these whales, 
such as long-term social groups, a symmetrical matrix of association 
indices between all possible pair combinations was analysed for naturally 
occurring clusters using the Cluster Analysis module of the Statistical 
Package for the Social Sciences (SPSS) program. A variety of hierarchical, 
agglomerative clustering methods are available in SPSS. The result of this 
clustering is a tree-like hierarchy showing the structure of clustering from 
all individuals as separate clusters to all individuals as members of one 
large cluster. Second, the methods conduct the clustering in the direction 
from separate clusters to a single cluster, as opposed to divisive methods 
which begin with the individuals all clumped and divide them into clusters. 
Hierarchical agglomerative methods are best suited to these data because 
social structure is also hierarchical in nature and the null hypothesis for 
the structure of this society is that each individual forms its own cluster 
with no stable outside associations. 
The methods chosen for these data compute squared Euclidean 
distances for the pair association indices and cluster them using the 
Unweighted Pair-Group Average Method (UPGMA), also called the 
"average linkage between groups" method. The measure of squared 
Euclidean distance is used to quantify the distance between two association 
indices as plotted· on x,y coordinates. The fact that these are squared 
distances serves to give more weight to larger distances. UPGMA has also 
been chosen by other statisticians as the most robust form of clustering 
which meets three simple conditions (Kaufman and Rousseeuw, 1990). The 
first condition is that the dissimilarity between merging clusters are 
monotone decreasing. Second, the dissimilarities between clusters must be 
unambiguous . In some methods (e.g. the weighted average method), the 
86 
selection of one of two equivalent choices in cluster formation result in very 
different dissimilarities between subsequent clusters. Third, 
dissimilarities between clusters should be statistically consistent, with 
larger samples resulting in more meaningful clusters. 
4.2.10 Pod Definition 
The resultant dendogram from the cluster analysis was used to 
identify distinct groupings of individuals. These groupin~ were defined as 
"pods", using the terminology from killer whale social groupings (Bigg et 
al., 1990). There are no universally accepted methods for determining the 
statistical significance of these clusters (Sokal and Sneath, 1963). 
Dendograms attempt to represent multi-dimensional relationships in two 
dimensions. The process of hierarchical, agglomerative cluster analysis 
generates a continuum of clusters. The dendogram is useful in visually 
displaying the associations between individuals and can make "natural" 
clusters apparent. In addition to this, the cluster distance coefficients 
calculated by UPGMA can also be used for guidance in deciding how many 
clusters are needed to represent the data (SPSS, 1990). As more and more 
distant clusters are joined, the interval from one distance coefficient to the 
next increases. A plot of these intervals versus the coefficients themselves 
was used to identify the appropriate limit to cluster distance. A pod was 
defined as any cluster of animals which had a distance coefficient of less 
than 3.0. This was not exactly equivalent to pair association indices 
because the clustering method used averages of these indices when joining 
clusters . Pods were given letter names according to the order in which at 
least one of the individuals was seen in the course of the study (i.e. the first 
pod seen was "A" pod, the second was "B'' pod, etc.). Pods which were 
linked by cluster distances between 3.0 and 4.0 were termed linked pods. 
,, 
These pods were named with a matching letter followed by sequential 
numbers (e.g. "Al" pod, "A2" pod, etc.). Pods linked at cluster distances 
greater than 4.0 were given letter-only names. 
A number of young animals and animals seen only once had to be 
removed from the bulk of the association analyses due to computer memory 
limitations. In order to determine pod assignments for these animals, a 
frame-by-frame analysis of available photographs was undertaken to 
determine the identity of all members of the immediate group or groups in 
which these animals were associated. These animals were assigned to the 
pod which they most commonly accompanied. 
4.2.11 Pod Associations 
Following the identification of pods, a final analysis was done to 
. examine higher-level associations between those pods. These associations 
between pods were termed "clans", again following the nomenclature of 
Bigg (1990). In order to calculate this, the individual identities in the data 
on groups were changed to the pod identities of those individuals. Then, 
groups were pooled by day, and pair associations for all pods seen on the 
same day were counted. The result of this data manipulation was that two 
pods were counted as co-occurring whenever at least one member from 
each pod was identified on the same day. There was no attempt to weight 
the presence of a pod by the total number of its members observed, as this 
was likely to be partially due to the variable recognisability of different pod 
members. Associations between pods were quantified using the same Ratio 
Association Index used for individual association. 
88 
4.3 RESUL'IS 
4.3.1 Identification and Classification of Individual Whales 
Pilot whales were photographed on 14 7 days. Over 26,000 frames of 
film were taken and 10,190 frames (39% of the total images) met the criteria 
of a usable identification photograph (see Section 4.2.2). Information on 
group composition to examine patterns of association was collected on 128 
days, when detailed information was also collected on the time of each 
photograph and the spatial structure of individual groups. 
A total of 495 individual pilot whales were catalogued. A summary of 
identifications by the 24 ID characteristics is presented in Table 4.4. There 
was wide variation in the degree of recognisability of these whales. The 
animals in the "Individual" category were the most distinctive. All had 
either major dorsal fin injuries or uniquely shaped fins. For example, 
whale #21 ("Splitfin") had a large flap partially cut more than halfway 
through the fin which remained constant throughout the study. Whale #40 
had very few nicks or scars, but was placed in the Individual category 
because of its squat, triangular fin which was different from all others. 
Alternatively, whales in the cleanfin category all had a very low 
recognisability. The summary of the recognisability of all animals is 
presented in Figure 4.1. 
The age and sex classification of the 495 identified whales resulted in 
104 adult males, 109 adult females with calves, 218. unknown animals, 33 
juveniles and 31 calves. An additional 17 juveniles and 60 calves were 
recognised by consistent associations with identified mothers, but did not 
have sufficient identifying characteristics to be included in the 
identification catalogue. This brought the total number of animals to 572. 
89 
11111 
Table 4.4. Summary of pilot whale identifications and average 
recognisabilities for the 24 fin types. 
Fin Type Description No. of Recognisability 
whales 
CLF Cleanfin 22 3 
CLS Cleanfin with Scallops 11 2-3 
CLB Cleanfin with Barnacles 10 2-3 
IND Unique Individual 10 1 
DLF Dolphin-Like Fin 8 1-2 
HOO Hook 21 1-2 
HOS Hook with Scallops 14 1 
FDH Front Dings or Humps 11 1-2 
RFT Raggedy Fins with Tabs 36 1 
SMN Small Multiple Nicks 51 2 
TSN Two Small Nicks 23 1-2 
LDN Large Double Nicks 17 1 
SBN Small Base Nick (lower third) 8 2-3 
LBN Large Base Nick (lower third) 20 1-2 
SNM Small Nick Middle third 14 2-3 
LNM Large Nick Middle third 27 1-2 
SNT Small Nick Top third 28 2-3 
LNT Large Nick Top third 25 1-2 
NNT Nick Near Tip 50 1-2 
SQT Square Tip 19 1-2 
FFT Flap Fin Tip 17 1-2 
SRT Small Round Tip 18 1-2 
LRT Large Round Tip 15 1 
HND Handle 20 1 
1 = highly recognisable, 2 = possibly recognisable, 3 = difficult to recognise 
Legend 
Im Difficult to Recognize 
Iii Possibly Recognisable 
• Highly Recognisable 
One: Visitor One: Resident Two to Five Six or More 
Number of Sighting Days 
Figure 4.1 Recognisability of adult whales with different sighting histories. Whales were categorised by their number of sighting days throughout the course of the study. Residents seen on one day only had significantly lower recognisabilities than other residents (P < .02). 
The proportion of animals was 18.2% males, 19.1% mothers, 38.1% 
unknowns, 8. 7% juveniles and 15.9% calves. The ratio of mature males to 
mature females with calves was 1:1.05, not significantly different from 
unity. If all of the unknowns were assumed to be adult females without 
calves (an unlikely assumption: see Section 4.2.5), the adult male: adult 
female sex ratio would be 1:3.14. This is a maximum estimate of the 
socionomic sex ratio (sensu Clutton-Brock et al., 1977). 
Mother-calf relationships were determined for 85 (93%) of the calves 
and 35 (70%) of the juveniles. For the remaining immatures it was not 
possible to identify a consistent relationship with a specific female. The 
number of calves per female ranged from 1 to 3, and averaged 1.26 ± .05, all 
values: mean ± SE, n = 109. Ten newborn calves were identified during the 
study. Three were estimated to be less than one month of age, while seven 
were observed at around six months of age (criteria for ageing described in 
Section 4.2.5). An additional two calves were observed as stillborns: dead 
foetuses which were carried by an adult female. Table 4.5 presents the 
dates of these observations and suggests that most births occur from May 
through June, although births may occur at other times as well. Birth 
periods in other populations of pilot whales have also been shown to be 
diffusely seasonal, having a single mode with births occurring in all 
months (Sergeant, 1962a; Kasuya and Marsh, 1984; Martin et al., 1987, 
pers.comm.) 
The proportion of unidentifiable animals was calculated for 35 
groups and ranged from 0% to 34% (14.8 ± 0.7). An average proportion of 
15% unidentifiable animals was used to revise analyses of group size which 
utilised only identified whales. Assuming that unidentifiable animals 
were distributed equally in all groups, comparisons of differences in group 
size between seasons or classes of animals should still be valid. 
Table 4.5. Summary of information used to identify calving period. 
TYPE OF OBSERVATION 
(with age estimate) 
Stillborn calf(:$; 0 months): 
Newborn calf(:$; one month): 
Presence of foetal folds (:$; six months) 
n= 7: 
OBSERVATION 
DATE 
28 April 1990 
5 May 1991 
20 June 1990 
5 December 1990 
5 June 1991 
August-
November 
ESTIMATED 
BIRTH MONTH 
April - May 
May 
May-June 
November 
May 
February - May 
The recognisability of the five age/sex classes is presented in Figure 
4.2. Males were the most recognisable, while immatures were very difficult 
to recognise. 
4.3.2 Patt.ems of Occurrence: Residents and Visit,ors 
The 495 pilot whales were identified in 2610 daily sightings (equal to 
the identification of one individual on one day) over 14 7 days of photographic 
sampling. The number of daily sightings per animal varied from one (n = 
217 animals) to 28 (n = one animal: #40) and averaged 5.27 ± 0.27 (Figure 
4.3). A number of the animals seen only once were documented in discrete 
groups and not in the company of any animals seen more than once. There 
were 15 such groups, totalling 107 animals, and these animals were termed 
visitors. The remaining animals were considered as residents. A number 
of analyses were conducted in order to document the validity of this 
terminology. Variability in sighting frequency could be due either to a real 
difference in patterns of occurrence, or to a sampling bias resulting from 
factors such as recognisability of individuals and the timing of sightings. 
For example, if an animal was difficult to recognise or was not identified 
until the latter part of the study, it would have a lower probability of being 
sighted again. In order to attempt to isolate these factors, I examined the 
recognisability and timing of sightings for animals seen only once. 
The timing of the initial sightings of all adults is presented in Figure 
4.4. Individuals are plotted according to the quarter of the study period in 
which they were first sighted. These sightings were significantly different 
from that expected by chance (X2 = 150.46, df = 9, P < .001), primarily due to 
the fact that 82% of the animals seen on six or more days were initially 
sighted in the first quarter and 84% of the animals seen on two to five days 
were seen in the first half of the study. (This could be partially due to the 
91 
1 
I 
Male 
Legend 
E Difficult to Recognize 
Ill Possibly Recognisable 
• Highly Recognisable 
Mother Unknown Juvenile 
Age/Sex Class 
Calf 
Figure 4.2. Recognisability for five age and sex classes. Adult males had significantly higher recognisabilities than immature whales. 
'(l:J 
Q) 
-~ 
..= 
~ 
~ 
0 
$,,t 
Q) 
,.Q 
s 
i 
200 
175 
150 
125 
100 
75 
50 
25 
0 
1 2 3 4 5 6 
Iii Visitor Whales N = 107 
• Resident Whales N = 388 
789Wll~~MffiWU~WWfil~~u~w~~ 
Number of Sighting Days 
Figure 4.3. Sighting frequencies for all identified whales. See text for 
definition of "residents" and "visitors". The average number of sighting days 
for all whales was 5.27 ± 0.27. 
[I) 
Cl) 
-~ 
.Cl 
=== ~
0 
~ 
Cl) 
"s 
i 
200 
175 
150 
125 
100 
75 
50 
25 
0 
1 2 3 4 5 6 
l!I Visitor Whales N = 107 
• Resident Whales N = 388 
789Wll~~M~WTI~WWfil~~~~W~~ 
Number of Sighting Days 
Figure 4.3. Sighting frequencies for all identified whales. See text for 
definition of "residents" and "visitors". The average number of sighting days 
for all whales was 5.27 ± 0.27. 
l.L':) 
(N 
"C 125 M Q) 
c.= 
• Visitors 
•l"I 
~ 
= II Residents: One Day Q) 100 
"C BI Residents: Two to Five Days 1-1 
Cl.l ~ Residents: Six or More Days Q) 
"a 75 ~ 
~ Q) 50 z 
(+.I 
0 
~ Q) 25 ,.Q 
s 
i 
0 0 
1st Quarter 2ndQuarter 3rdQuarter 4thQuarter 
Figure 4.4. The quarterly timings of the initial identifications of 
whales with similar sighting histories. Whales were categorised by 
their number of sighting days throughout the course of the study. 
Quarters were defined as 1st) 4 October 1989 - 16 March 1990; 2nd) 17 March 
1990 - 27 August 1990; 3rd) 28 August 1990 - 7 February 1991; 4th) 8 February 1991 - 22 July 1991. 
decreasing probability of being seen on multiple days as the study drew to a 
close.) There were no significant quarterly differences between the 
occurrences of the two classes of visitors and residents seen on one day (X2 
= 2.751, df = 3, P > .05). Sixty-one percent of the visitors were seen in the 
first half of the study, showing they had ample opportunity to be resighted. 
Additionally, the recognisability of visitors was not significantly different 
between quarters (Figure 4.5), with 54% of the visitors sighted in the first 
half of the study having a high degree of recognisability. Comparisons of 
recognisabilities for all whales (Figure 4.1) showed that there were no 
significant differences between the recognisability of visitors and residents 
seen more than once (56% vs. 52% highly recognisable; X2 = 0.486, df = 2, P 
> .05). This indicates that visitors had an equal probability of resighting due 
to their recognisability and suggests that the category of visitor animals is a 
valid one. 
The situation is different for residents seen on only one day. The 
recognisability of all animals presented in Figure 4.1 shows that the 
recognisability of residents seen on only one day was significantly less than 
that for the more frequently seen residents (40% vs. 57% highly 
recognisable; X2 = 7.967, df = 2, P < .02), indicating that sighting bias may 
account for the lack of multiple sightings for these animals. The 86 
resident adults seen only once may include some animals which could be 
classified as transient visitors to the area, but since they were observed in 
association with other animals seen more than once, I have included them 
in the resident category. They were not include~ in the analysis of 
association patterns because of computer memory limitations. However, 
they were included in the analysis of the final composition of social groups 
presented later. 
9'2 
1111 
100 
80 
20 
0 
Legend 
1B Difficult to Recognize 
Iii Possibly Recognisable 
• Highly Recognisable 
1st Quarter 2nd Quarter 3rd Quarter 4th Quarter 
Figure 4.5. Recognisability of all adult visitor whales grouped by the 
quarter of the study when they were sighted. 
Quarters were defined as 1st) 4 October 1989 - 16 March 1990; 2nd) 17 March 
1990 - 27 August 1990; 3rd) 28 August 1990 - 7 February 1991; 4th) 8 February 
1991 - 22 July 1991. 
n I \ 
I 1 
Comparisons were also made between the sighting histories and 
recognisabilities of adults (age/sex classes: male, mother and unknown) 
and immatures (age/sex classes: juveniles and calves). Adults were seen 
from once (186 whales) to 28 times (one whale), while young animals were 
seen from once (31 whales) to 16 times · (two whales). There were no 
significant differences (two-tailed t= 0.442, v = 493, p > .50) between the 
sighting frequencies for adults (5.51 ± 0.30 sightings per adult, n = 431) and 
immature animals (3.70 ± 0.51 sightings per immature, n = 64). However, 
the recognisabilities of immatures were markedly different from those of 
adults (Figure 4.2), making their recognition in subsequent sightings 
difficult. Indeed, there were 77 clean-finned immatures who were only 
recognised because of their consistent relationships with identified adults. 
Immature animals were removed from further analysis on patterns of 
occurrence and association, but were included in later summaries of pod 
composition. 
The cumulative sighting curve for all identified whales is presented in 
Figure 4.6. This curve continued to increase for both residents and visitors, 
suggesting that continued sampling will still be required before all animals 
using the Tenerife area are identifed. The individual sighting histories for 
all resident whales are plotted in Figure 4. 7. This shows the high degree of 
repeat occurrence for many of these individuals which will be analysed in 
more detail by the examination of patterns of individual associations within 
groups. 
4.3.3 Group Size 
The basis of the analysis of associations was the documentation of 
pilot whale groups. Using the sampling criteria for well-defined groups 
outlined in Section 4.2.6, 277 groups of pilot whales were recorded over 123 
93 
400 
00 
.s:l 
ro 300 
~ 
<+-< 
0 
~ 
~ 200 
s 
::s 
z 
100 
0 
8 Visitor Whales N = 107 
E111 Resident Whales N = 388 
Figure 4.6. Cumulative sighting curve for all identified pilot whales. See 
text for definition of "resident" and "visitor" whales. 
\I 
1 
J..t 
Q) 
's i 300 
= 0 
.... 
~ 
~ 
f+:: 
.... 200 
-s 
Q) 
"C 
""'4 
~ 
~ 100 
, .... 
I'. . : • • I 
I .-
,,. . ,, ·"" . 
.18• ~ • • •• ' 
• ••• .- •• • - • I •  . .- . 
•• 
• 
• 
• 
• 
• • • • • • 
. . --- .. 
. . . .... ·r·. 
• • • 
• I 
•• • • • • 
• 
I • 1 : • • 
• & •. • • • • • • I . '. .. . •• _,J ... 
.' 
.' • , I ,'I. • • • • I • • • I • 
• 
• 
Ii 1 • • ,• II 1 I 1 ... ,11 I 
' 
.. ····,·· ............ , 
.... • I • • • \ • ! : I .:, II I ~ I ... J • ··-· .• 1! ,: • :. :, ;,(·i I '. ~ .: •• ; ·h::• : ·J 
1
°i;i 
'
I.. . I t, ,, ! .... · t ~ •• • •• I •• . .. . ., ,, 
• , Ii . : :·. ·, ; i • ~ , . :, • . :, • • ·.i: • r: 
• • • • • • I '\ • I .,I I ! I •I ,. 'a• 
11 I ··'1'i ,•,st;~ .:~ 1 J 1!,a•J :••.: '.· ~I .'I r., 
UI I. I I • •• ; • • 'a 0 ---.......... """T"'""'"""'T'""'~~--.--~--.......... ---.-·---.-----.---.-..-=--,....--.---..-=-......-0 N D J F M A M J J A S O N D J F M A M J J 
1989 1990 1991 
Figure 4.7. Sighting histories for all resident pilot whales. Each 
dot ( • ) represents one or more identification photographs of an 
individual whale on one day. 
11 
encounter days. Group size ranged from two to 34 identifiable individuals 
(7 .9 ± 0.4, N = 277). This is an underestimate of actual group size due to the 
estimated 15% of the population which is unidentifiable. 
Group size was examined for . seasonal variation. It was 
hypothesised that group size might vary according to variation in food 
availability or to mating opportunities. Group sizes were pooled for both 
years by month, by four three-month "seasons"(beginning in January) and 
by two "half-years" (summer: April-September, winter: October-March). 
Groups of identifiable whales were significantly larger during the half-year 
summer months than during the winter months (8.4 7 ± 0.48 vs. 6.89 ± 0.48 
whales per group, Mann-Whitney U = 10375.5, n(summer) = 171 groups, 
n(winter) = 106 groups, Z = 2.03, P < .05). There were no significant 
differences in the number of identifiable whales per group pooled by month 
(Kruskal-Wallis rank ANOVA H = 16.03, df = 11, P > .05) or by season (H = 
4.93, df = 3, P > .05). 
Another way of examining seasonal variation in grouping patterns 
was to examine the number of groups located on any given day. The 
mean= 
number of groups ranged from one to nine ~.25 ± 0.12 groups per day, N = 
123). There were no significant seasonal variations found for any of the 
temporal combinations (month: Kruskal-Wallis H = 13.52, df = 11, P > .05; 
season: H = 6.83, df = 3, P > .05; half-year Mann-Whitney U = 2006, 
n(summer) = 78, n(winter) = 45, Z = 1.39, .20 > P > .15). The fact that whales 
were found in larger groups during the summer, and yet there were no 
significant seasonal variations in the number of groups located, could 
indicate either that there was a seasonal influx of whales that joined 
existing groups or that a reduced number of whales were travelling in 
larger groups. The question of seasonal variation will be brought up again 
in the analysis of associations between pods in Section 4.3.6. 
' \l 
I 
4.3.3 Association Patterns and the Definition of Pods 
The overall distribution of association coefficients 1s presented in 
Figure 4.8. The average association value was 0.178 ± 0.003 (N = 3971). 
There were a total of 29,890 possible pair combinations for the group of 245 
whales. Only 3970 (13%) of these pairs were ever observed. Almost half of 
the observed pairs had association coefficients of less than 0.10, while 64 
pairs had coefficients of 1.0 (i.e. they were always observed together). This 
shows a high degree of selection in association patterns and suggests that 
whales do not associate at random. 
The final group association patterns for the 245 resident adults seen 
on more than one day are presented in the cluster dendograms in Figures 
4.9a-f. These figures are plotted according to Cluster Distance Coefficients 
(CDC's) calculated by SPSS to form clusters. The increasing variability in 
the distance at which clusters were formed is presented in Figure 4.10. 
Clusters formed at a distance of less than 3.0 are shown to be closely linked. 
Clusters formed above 3.0 show increasing variability in CDC. This is the 
reason that all clusters linked below 3.0 were defined as within-pod 
clusters, while clusters formed above 3.0 were defined as between-pod links. 
Clusters formed between pods at CDCs of 3.0-4.0 were defined as linked 
pods and given shared-letter, alpha-numeric pod names emphasise the 
links between them. I chose a CDC of 4.0 as a limit somewhat arbitrarily, 
however, selection of higher CDCs (e.g. 5.0 or 5.5) only linked one additional 
pair of pods. Figure 4.10a shows the frequency distribution of CDCs for all cluster linkages. 
A total of 31 "pods" were identified from this analysis . As a means of 
verifying the validity of the pod definitions, the individual pair association 
indices were classified as to whether they were between two members of the 
same pod, between two members of associated pods, or between two whales 
2500 
2000 
N = 3970 pair associations 
Total possible pair combinations= 29,890 
Unobserved pairs (RAI = 0): N = 25,920 
~ 1500 ~ 
Q) 
& f 1000 
~ 
500 
0 
0.05 0.15 0.25 0.35 0.45 0.55 0.65 0. 75 0.85 0.95 1 
Ratio Association Index (RAI) 
Figure 4.8. Frequency distribution of ratio association indices for 245 
adult whales seen on more than one day. Category labels indicate the 
upper limit of the category. 87% of the pair combinations were never 
observed. 
DISTANCE COEFFICIENT SCALE 
0 2 3 4 5 6 7 8 9 +----+-----+----+----+----+-----+----+----+-----+-+ WHALES PODS 
173 -, 
603 -i P Pod 157 I 1 
158 _J I l 
160 I I I 
161 _J I- 1 598 -, I I 601 -+---, 597 _J I I I 150 __ _J I 
------------------------------------------------+------------------118 -,, I ~~~ ~ I I A2 Pod 
306 -h I o 11 _J I I 005 _ _J I 
119 _J I 
- - - - - - - - - - - - - - - - - - - - - - - - - -+- - - - - - - - - - - - - - - - - - - - -+- - - - - - - - - - - - - - - - - -002 -, I 003 
-+-- I I A1 Pod 001 _J I I 285 
--i-, I r----, 490 _J r-1 I I I I 286 I rJ I I 284 I I 409 _J ~ I I I 006 I I I I I 010 _J ~ I I 406 -, I I I 407 ji I I I 287 r---, 570 _J I I I 489 I I 
--------------------------------------------+---+-------+----------347 11 I I 
348 _J · J I I 058 -y- L 
117 -l I 
--, r_J 
337 -, I I I 343 -t---, I I 
336 _J ~ 
1
1 I 
338 
_ I I 
N Pod 
------------------------------------+-----+-+-----------+----------
Figure 4.9a. Dendogram of associations between individual pilot whales. Pods were defined as clusters linked by distance coefficients of less than 3.0 
and are separated by heavy dashed lines. Pods linked by clusters at distances between 3.0 and 4.0 were designated "linked pods", given shared-letter number names and are separated by light dashed lines. Source: SPSS Cluster Analysis. 
l d 
I 
I 
DISTANCE COEFFICIENT SCALE 
0 2 3 4 5 6 7 8 9 +----+-----+----+----+----+-----+----+----+-----+-+ WHALES PODS 
------------------------------------+-----+-+-----------+----------
m ~ I 111 I R Pod 
351 ....i I I 
349 I I I 
------------------------------------------+-+-----------+----------444 
450 
443 
442 
445 
448 
449 
447 
-, I I I +, I I I T Pod 
....i ~1 ~~ I I I ~ -------.1 W I 
I I I 
----------------------------------------+-+-------------+----------221 
491 
225 
379 
217 
219 
125 
214 
216 
092 
129 
128 
131 
224 
223 
139 
377 
462 
020 
035 
019 
138 
523 
....J I 
~-I-~ 
....J I 
-, 
I I I 
....J I I 
_J h~ I I I I 
-, I I I I I I I I 
....J I I t--i 
II I I ....J I 
I 
__J I 
I 
I 
I 
I 
I 
I 
I 
I 
I 
I 
1 
I 
11 
I 
I I 
I I I 
i 
I 
I 
I 
I 
I 
I 
I 
J2 Pod 
J1 Pod 
--------------------------------------+-+---------------+----------
Figure 4.9b. Dendogram of associations between individual pilot whales. Pods were defined as clusters linked by distance coefficients of less than 3.0 and are separated by heavy dashed lines. Pods linked by clusters at distances between 3.0 and 4.0 were designated "linked pods", given shared-letter number names and are separated by light dashed lines. Source: SPSS Cluster Analysis. 
DISTANCE COEFFICIENT SCALE 
0 2 3 4 5 6 7 8 9 +----+-----+----+----+----+-----+----+----+-----+-+ WHALES PODS 
--------------------------------------+-+---------------+----------g~~ Tl I I I K 1 Pod 
022 I I I I I 
077 ~ I 11 I 057 I I 
137 
1
1 t---i I 
324 I I 
------------ ------------+-------------+-----------------+----------099 -, 
375 I I K2 Pod 098 _J h 
054 I I I I 
055 _J 11 I 052 I 
053 __J t---i t-
051 I 
032 I hi 033 _J 
061 
=t-J ~ 084 
063 
-I I 
093 _J I 
192 
436 
--------------------------------------+-----------------+----------036 -, I 
037 I L Pod 39A _J I 
--------------------------------------+-----------------+----------206 I I 
433 h I 488 M Pod 
583 I I I 
6~7 ~ I 11 38A I I I g;~ j 11 I 
040 I I I 
20s -, I I I I ~~~ J ~1 r L I 207 __J I ~ 1 I j~~ I I I 
------------------------------------+-- .--------' -------+----------
Figure 4.9c. Dendogram of associations between individual pilot whales. 
Pods were defined as clusters linked by distance coefficients of less than 3.0 
and are separated by heavy dashed lines. Pods linked by clusters at distances between 3.0 and 4.0 were designated "linked pods", given shared-letter number names and are separated by light dashed lines. Source: SPSS Cluster Analysis. 
DISTANCE COEFFICIENT SCALE 
0 2 3 4 5 6 7 8 9 +----+-----+----+----+----+-----+----+----+-----+-+ WHALES PODS 
------------------------------------+-------------------+----------
067 I I I I g~: -I~ lhl I I H Pod 
066 I I 064 -t-J r I 065 _J I I 
296 I 
------------------------------------+-------------------+----------401 -, I I 402 -+-
--, I S Pod 400 _J I I 
------------------------------------+-------------------+----------434 -.-·-----, 
435 -l 
-------, I Pod 018 -, 
562 -'-"------' 
017 -l 
----------------------------------+-+-------------------+----------506 
510 
549 
266 
508 
269 
439 
016 
505 
171 
159 
507 
086 
438 
20A 
310 
441 
I I 
11 
-_J----'1 
-.-, ~ 
_J 11----,1 I 
I 
I 1 
=; 11 ---
_J 1-1 ----.1 I 
_J I I 
--,-1--,1 ~ 
_J I I 
~,.h I 
--J ~ 
_J I 
I I 
~ 
hi 
I I I 
I I I 
F1 Pod 
----------------------------+---+-+---------------------+----------g~d _J l I I I I F2 Pod o7o ,--, 1------, I I I I 
074 _J r--i I I I I I I 
548 ~ ~ I I I 
313 I I I I 
--------------------------+-----+-+-- -------------------+----------170 
-
__. I I I F3 Pod 
172 I I I 
--------------------------------+-+---------------------+----------
Figure 4.9d. Dendogram of associations between individual pilot whales. Pods were defined as clusters linked by distance coefficients of less than 3.0 
and are separated by heavy dashed lines. Pods linked by clusters at distances between 3.0 and 4.0 were designated "linked pods", given shared-letter number names and are separated by light dashed lines. Source: SPSS Cluster Analysis. 
DISTANCE COEFFICIENT SCALE 
0 2 3 4 5 6 7 8 9 +----+-----+----+----+----+-----+----+----+-----+-+ WHALES PODS 
--------------------------- ----+-+---------------------+----------320 i--i I 484 
_1 0 Pod 134 I 135 
_J I 
--------------------- ----------+-----------------------+----------482 I I 
590 I I E Pod 205 11 
262 
_1 I wl I 201 I I I 1' I 203 
_J I I 199 
____J I ~~~ n I I 194 I I I 195 
_1 t--, I I I I 079 j----J I I 478 
____J I I I I 023 I I I 
------------------------------+-+-----------------------+----------566 I I I U Pod 567 
_J I I I I 
------------------------------+-+-----------------------+----------044 I I I 046 ----l I I I G2 Pod 043 I 
304 I I I I I I I 305 
_J I I h I I 
283 LJ I I j--l I 042 1 I I I g:~ ~ ~ 111 I ~~~ ~~ 111 I 569 t-i I I 
--------------- --- --------+---+------------ -- -----------+----------g~~ . hh I I I G 1 Pod 
418 I I I 069 I I I 100 ~ I I I 298 1 I I I I g~~ =1_J-J ~ I I 020 I I I 
------------------------+-----+-------------------------+----------
Figure 4.9e. Dendogram of associations between individual pilot whales. Pods were defined as clusters linked by distance coefficients of less than 3.0 and are separated by heavy dashed lines. Pods linked by clusters at distances between 3.0 and 4.0 were designated "linked pods", given shared-letter number names and are separated by light dashed lines. Source: SPSS Cluster Analysis. 
I 
1 
DISTANCE COEFFICIENT SCALE 
0 2 3 4 5 6 7 8 g +----+-----+----+----+----+-----+----+----+-----+-+ WHALES PODS 
------------------------+-----+-------------------------+----------;~~ -, · I I 82 Pod 
327 __J 1--J I 
----------------------+-------+--~----------------------+----------030 I I 
410 __J J I 
455 I 457 
81 Pod 
502 I I 
624 I I 460 I 
------------------------------+-------------------------+----------
110 • I I 248 J 
113 I I I 144 I I 
D Pod 
----------------------------+-+-------------------------+----------108 -, I I I ~~~ ~ 1 I C2 Pod 111 _J I I I 
115 I I I 
250 ~ I 
------------------------+-------------------------------+----------
142 _J I I 143 I 
105 t--i I 
253 __J ~ I 106 I l I 107 _J h I ~~; =3 I I I 
Cl Pod 
127 I 
--------------------------------------------------------+----------331 I Ql Pod 332 _J I I 
----------------------------+---------------------------+----------357 . -, I I 
358 -l I I Q2 Pod 333 -i ;~~ -t~~~~~-,~, I 
353 j 
354 
- - - - - - - - - - - - - - - - - - - - - - - -+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -334 -, I 
359 I Q3 Pod 352 _J ~ 
340 
Figure 4.9f. Dendogram of associations between individual pilot whales. 
Pods were defined as clusters linked by distance coefficients of less than 3.0 
and are separated by heavy dashed lines. Pods linked by clusters at distances between 3.0 and 4.0 were designated "linked pods", given shared-letter number names and are separated by light dashed lines. Source: SPSS Cluster Analysis. 
... 
1.6 
1.4 
N = 244 Clusters 
~ 1.2 ~ 
~ 
~ 1.0 ~ ~ 
•.-1 
Q 
-.,;I 0.8 d 
~ 
•.-1 
~ 0.6 s 
~ 
0 
0 0.4 
0.2 
0.0 
0 1 2 3 4 5 6 7 8 9 10 
Cluster Distance Coefficient 
Figure 4.10. Cluster Distance Coeffiecients (CDC: calculated by the unweighted pairs group averaging method of hierarchical cluster analysis on the ratio association indices of 245 pilot whales) are plotted as a function of the difference between sequential coefficients (e.g. coefficient 2 - coefficient 1). The increasing distance at which clusters were formed was used as an indicator of pod divisions. Clusters formed at CDC < 3.0 were defined as pods. Pods joined at 3.0 <CDC< 4.0 were defined as linked pods. 
150 0 ~ 
M 
125 
N = 244 clusters 
~ 100 
~ 
= ~ 
= 75 04 
~ 
~ 
~ 
50 
25 
M 0 M 0 
0 1 2 3 4 5 6 7 8 9 10 
Cluster Distance Coefficients 
Figure 4.10a Frequency distribution of cluster linkages between 245 pilot 
whales clumped by the Cluster Distance Coefficients (CDC) at which each · 1 
cluster was formed. All clusters formed at CDC< 3.0 (N = 214) were defined 
as within-pod clusters, resulting in the definition of 31 pods (equivalent to 
the remaining number of clusters still to be formed: 245-214=31). Clusters 
formed between CDC = 3.0 and 4.0 (N = 11) were defined as linked pod 
clusters and resulted in the definition 20 linked pod groups. 
from different pods. Over 96% of the possible "within-pod" associations and 
80% of the "between-linked-pod" associations were observed, compared to 
only 8% of the "between-pod" associations (Table 4.6) . The association 
indices for each of these pair combinations were found to be significantly 
different (Kruskal-Wallis H = 2419.48, df = 2, P < .0001), with association 
indices for within pod pairs over twice as high as associations between 
animals from linked pods (.391 ± .007 vs .. 072 ± .001), and over five times 
higher than associations for pairs from different pods (.072 ± .001; see 
Figure 4.11). This indicates that the pod definitions are valid and identify 
distinct grouping patterns in these pilot whales. 
Since these 31 pods were defined only on the basis of associations 
between 245 adult animals seen more than once, it was necessary to assign 
pod membership to the additional 250 identified whales which had been 
removed from the analyses (see Section 4.2.9). The 15 groups of 107 visitors 
were each defined as a separate pod. For the remaining 143 resident 
animals, pod assignment was possible for 130 using the method described 
in Section 4.2.10; the thirteen animals without a pod assignment were all 
seen only once, either alone or associated with animals from more than one 
pod. Additionally, the 77 clean-finned immatures described in Section 4.3 .1 
were assigned the pod identity of the animal with whom they associated. 
This brought the total number of animals used in the pod analyses to 559 
(495 identified -individuals - 13 individuals without pod assignments + 77 
clean-finned immatures added by association). 
4.3.4 Pod Composition 
Pod size for the total sample of 46 pods ranged from two to 33 and 
averaged 12.2 ± 1.3 (Table 4. 7). The distribution of pod sizes showed a slight 
bi-modality due to five pods with 25-27 whales (Figure 4.12). In addition to 
2 
Table 4.6. Comparison of three types of individual associations based upon 
the pod membership of individuals. 
Possible Observed 
Association Type Pairs Pairs Percent 
Within Pod 1167 1129 96.7 
Between Linked Pods 724 583 80.5 
Between Pods 27999 2258 8.1 
TOTAL 29890 3970 13.3 
All Associations 
.7 
~ 1129 
Q) 
.6 
"'C IS.D. d 1-4 
.5 mean d 
0 S.D. 
•l"'I 
.4 ..,;i ~ 
•l"'I 
~ 583 0 
.3 t'IJ 
I < 0 .2 •l"'I 2258 ..,;i ~ I ~ .1 
0 
Within Pods Between Linked Pods Between Pods 
Pod Membership Pair Associations 
Figure 4.11 Average ratio association indices for all pair associations 
based on the pod membership. Sample sizes for each bar indicate the 
total number of pairs for each association type. See text for definition of 
"linked" pods . 
Table 4.7. Summary of pod sizes (mean± S.E .) for resident, visitor, 
productive and non-productive pilot whale pods (see text for defintions). 
Sample sizes of the number of pods in each category are given in 
parentheses. 
Resident Visitor Total 
Productive 16.9 ± 1.6 (24) 11.2 ± 2.1 (8) 15.5 ± 1.4 (32) 
Non-productive 4.6 ± 0.9 (7) 4.7 ± 0.9 (7) 4.6± 0.6 (14) 
Total 14.1 ± 1.6 (31) 8.2 ± 1.5 (15) 12.2 ± 1.3 (46) 
12 
ID Visitor Pods N= 15 
10 
• Resident Pods N = 31 
8 
J 6 4 
2 
0 
3 6 9 12 15 18 21 24 27 30 33 
Pod Size 
Figure 4.12. Frequency histogram of pod sizes (including 
clean-finned immatures). Size categories indicate the upper limit 
of the interval. 
the classification of resident and visitor pods, pods were classified as 
productive (32 of 46: 70%) or non-productive (14 pods or 30%) based on the 
presence of mothers and immature animals (Table 4.7, Figure 4.13). 
Variation in pod size was tested for the relative effects of residency and 
productivity in a two factor rank ANOVA; both factors had significant 
effects on pod size (productivity: H = 22.12, df = 1, P < .001; residency: H = 
4.96, df = 1, P < .05), but the interaction factor was not significant (H = 2.93, 
df = 1, P > .05) indicating the two factors acted independently on pod size. 
Thus, resident pods were significantly larger than visitor pods (14.1 ± 1.6 
vs. 8.2 ± 1.5) and productive pods were significantly larger than non-
productive pods (15.5 ± 1.4 vs. 4.6 ± 0.6). Comparisons of resident and visitor 
pods showed there were no significant differences in the numbers of 
productive and non-productive pods (residents: 24 productive, 7 non-
productive vs. visitors: 8 productive, 7 non-productive; X2 = 2. 77, df = 1, P > 
.05). 
The age and sex composition of all pods is presented in Figure 4.13. 
Four of the productive pods (12%) had no adult males in association with 
them and could be considered "nursery" pods. Two of the non-productive 
pods were composed solely of adult males (both visitor pods totalling two 
and six), one was composed of three unknown animals and the remaining 
seven non-productive pods were composed of a mixture of adult males and 
unknowns . 
The number of males per pod ranged from zero (five pods) to six (one 
pod) and averaged 2.5 ± 1.4 males per pod. Overall, the percentage of males 
per pod was 25 . 7 ± 3.6%, but resident pods had significantly lower 
percentages of males than did visitor pods (19.4 ± 3.1% vs. 38.6 ± 8.0%, U = 
334.5, Z = 2.39, P < .02), and productive pods had significantly lower 
rJl 
Cl) 
-; 
35 I 1 
Em 
~ 
301 1 mm 
• 25-f I• 
Calves 
Juveniles 
Unknowns 
Mothers 
Males 
~ 1 
C+-t 20 Non-Productive: 0 : 
I So4 : 
Cl) : 
"e 15 1 
= ; z ~ 
10-
I 
I 
I 
I 
I 
I 
I 
I 
I 
I 5, 1111 ll!Ulli r, 
~ .. 111111 0 Q1 1 u 'n ';'B2 1Q3 1Q2•Fs 1 
I Resident Pods rnnr 
~r."' 
mlJIIIII 
I'!'.'!. 
Productive 
r 
~ 
rc:lr:'.'· 
=mm1,HHl1'1il ~ , 
• 
~ 
... 
Id~., 
&]~ 
Visitor Pods 
' 
' 
' 
' 
~ 
Non-Productive; Productive .... ~ Kl 
id 
' 
' 
' 
ill&)~ 
I S F2 H C2 A2 BI J2 Cl Kl N G! T R JI G2 P E K2 M FI Al 4 7 5 14 13 12 6 8 3 9 15 11 10 2 I 
Pods 
Figure 4.13. Pod size and age/sex composition (including clean-finned immatures) for 46 pods. 
Pods were defined as "productive" or "non-productive" based on the presence of immatures. See text 
for definition of "resident" and "visitor" pods. Resident pods were significantly larger than visitor 
pods and productive pods were significantly larger than non-productive pods. 
percentages of males when compared to non-productive pods (16.4 ± 2.2% 
vs. 46.9 ± 8.2%, U = 370, Z = 3.49, P < .001). 
The number of mothers per productive pod ranged from one (11 pods) 
to twelve (one pod) and averaged 3.2 ± 0.4 mothers per pod. Mothers 
comprised 19.9 ± 1.5% of all productive pods and there were no significant 
differences in the percentages of mothers in resident vs. visitor pods 
(Mann-Whitney U = 132.5, Z = 1.632, P > .10). 
The numbers of unknown animals per pod ranged from zero (two all-
male pods) to 14 (one pod), with an average of 4. 7 ± 0.5 unknowns per pod. 
One pod was composed solely of three unknowns. Overall, the percentage of 
unknowns averaged 41.2 ± 3.3%. There were no significant differences in 
the percentages of unknowns in visitor pods compared to resident pods (U = 
235.0, Z = 0.59, P >> .05), but there were higher percentages of unknowns in 
non-productive pods than in productive pods (36.0 ± 2.8% vs. 53.1 ± 8.2%, U = 
317.5, Z = 2.23, P <.05). 
The number of immatures (juveniles and calves combined) in 
productive pods ranged from one (14 pods) to 12 (one pod), with an average 
of 4.3 ± 0.6 immatures per pod. Immatures comprised 27. 7 ± 2.2% of all 
productive pods. There were no significant differences in the percentages 
of immatures between resident and visitor pods (U = 115.0, Z = 0.83, P >> 
.05). There were also no significant differences between the resident pod vs. 
visitor pod percentages of either of the immature subsets of juveniles (U = 
117.0, Z = 0.92, P > .05) or calves (U = 103.0, Z = 0.30, P > .05). 
4.3.5 Variability in Associations with Age/Sex and Pod Membership 
Pod membership and age/sex class had a significant effect on the 
variability in associations between all non-immature whales. The removal 
of immatures, who by definition had highest associations with their 
presumed mothers, removed all suspected kin relationships from this 
analysis. The combination of the factors of pod membership and age/sex 
class were compared for effects on the ranks of association indices in a two-
way rank ANOV A (Zar, 1984). The pod and class factors were both 
significant (pod factor H = 2418.300, df = 2, P < .001; class factor H = 19.561, 
df = 5, P < .005), but the interaction factor was not (H = 4.437, df = 10, P > .05). 
This shows that the two factors of pod membership and age/sex class acted 
independently on the level of association, and that pod membership had a 
relatively greater effect than age and sex class. The comparison of 
association indices depending on pod membership and independent of age 
and sex class was already described in Section 4.3.3 (see Figure 4.11) as a 
verification of the definition of the pods. 
Variability in associations, depending upon the age/sex class of the 
pair members, was analysed in more detail in order to examine the bases of 
the relationships between males, females, unknowns and immatures 
within and between pods. Overall (i.e. independent of the pod memberships 
of the associating whales), associations between unknown whales were the 
highest ranked, while those between males and unknowns ranked second 
(Figure 4.14). Multiple comparisons showed that these two types of 
associations were significantly different from all associations involving 
mothers, which were not significantly different from each other. 
For within-pod associations, there were significant differences 
between age and sex class association indices (Kruskal-Wallis rank 
ANOVA, H = 21.38, P < .001), with all associations within and between 
males and unknowns (e.g . male:male, male:unknown and 
unknown:unknown pairs) significantly higher than all associations in 
which mothers were members (Dunn's multiple comparisons, P < .05 ; 
II 
I, 
of immatures, who by definition had highest associations with their 
presumed mothers, removed all suspected kin relationships from this 
analysis. The combination of the factors of pod membership and age/sex 
class were compared for effects on the ranks of association indices in a two-
way rank ANOVA (Zar, 1984). The pod and class factors were both 
significant (pod factor H = 2418.300, df = 2, P < .001; class factor H = 19.561, 
df = 5, P < .005), but the interaction factor was not (H = 4.437, df = 10, P > .05). 
This shows that the two factors of pod membership and age/sex class acted 
independently on the level of association, and that pod membership had a 
relatively greater effect than age and sex class. The comparison of 
association indices depending on pod membership and independent of age 
and sex class was already described in Section 4.3.3 (see Figure 4.11) as a 
verification of the definition of the pods. 
Variability in associations, depending upon the age/sex class of the 
pair members, was analysed in more detail in order to examine the bases of 
the relationships between males, females, unknowns and immatures 
within and between pods. Overall (i.e. independent of the pod memberships 
of the associating whales), associations between unknown whales were the 
highest ranked, while those between males and unknowns ranked second 
(Figure 4.14). Multiple comparisons showed that these two types of 
associations were significantly different from all associations involving 
mothers, which were not significantly different from each other. 
For within-pod associations, there were significant differences 
between age and sex class association indices (Kruskal-Wallis rank 
ANOVA, H = 21.38, P < .001), with all associations within and between 
males and unknowns (e.g. male:male, male:unknown and 
unknown:unknown pairs) significantly higher than all associations in 
which mothers were members (Dunn's multiple comparisons, P < .05; 
,I 
All Associations 
0.23- 661 
~ I ~ "C 
= 
0.21-~ 
= 0 812 .... ~ 
~ 
239 t 
.... 
0.19-~ 
0 
rt.l 
511 < 
0 1129 I • .... ~ 0.17 - r ~ 618 
0.15- 1 
0.13 --"T,-----r, ----...--,----.,---~.-----r,--
Ma:Mo Mo:Un Mo:Mo Ma:Ma Ma:Un Un:Un 
Age/Sex Class Pairs 
Figure 4.14. Average association indices (mean ± S.E.) for all pair 
associations (excluding immatures), grouped by age and sex class. 
Age and sex class abbreviations are for adult males (Ma), mothers (Mo), and unknowns (Un). Sample sizes for each bar indicate the total 
number of pairs for each age/sex class combination. 
Figure 4.15). The highest ranked associations were between adult males, 
indicating that male:male pairs had the most consistent relationships. Of 
all associations involving mothers, associations between mothers and other 
mothers ranked the highest , while associations between mothers and adult 
males ranked the lowest (Figure 4.15). 
When linked pods associated (Figure 4.16), associations were still 
significantly different from each other (H = 23.29, P < .001), but the 
significantly higher ranked association indices were no longer just for 
male:male, male:unknown and unknown:unknown pairs. Associations 
between males and mothers from linked pods were also significantly 
higher than all other mother associations and not significantly different 
from male:male associations (male:mother = 0.189 ± .011 vs. male:male = 
0.198 ± .021, P > .05). 
Associations between whales from unlinked pods showed the same 
overall pattern of significantly higher associations within and between the 
classes of males and unknowns (Kruskal-Wallis H = 22.99, P < .001; Figure 
4.17). However, mothers' associations with other mothers were ranked 
significantly lower than male:male and male:unknown associations (P < 
.05), and were the lowest ranked of all the mothers associations. 
4.3.6 Associations Between Pods 
My definition of a pilot whale pod was based on the most common 
pair-wise associations. As described in Section 4.3.5, the association 
indices for pairs of whales from the same pod were significantly higher 
than indices for pairs of whales from different pods. I would like to focus 
on associations between members of different pods in order to attempt to 
100 
I 
Figure 4.15). The highest ranked associations were between adult males, 
indicating that male:male pairs had the most consistent relationships. Of 
all associations involving mothers, associations between mothers and other 
mothers ranked the highest, while associations between mothers and adult 
males ranked the lowest (Figure 4.15). · 
When linked pods associated (Figure 4.16), associations were still 
significantly different from each other (H = 23.29, P < .001), but the 
significantly higher ranked association indices were no longer just for 
male:male, male:unknown and unknown:unknown pairs. Associations 
between males and mothers from linked pods were also significantly 
higher than all other mother associations and not significantly different 
from male:male associations (male:mother = 0.189 ± .011 vs. male:male = 
0.198 ± .021, P > .05). 
Associations between whales from unlinked pods showed the same 
overall pattern of significantly higher associations within and between the 
classes of males and unknowns (Kruskal-Wallis H = 22.99, P < .001; Figure 
4.17). However, mothers' associations with other mothers were ranked 
significantly lower than male:male and male:unknown associations (P < 
.05), and were the lowest ranked of all the mothers associations. 
4.3.6 Associations Between Pods 
My definition of a pilot whale pod was based on the most common 
pair-wise associations. As described in Section 4.3.5, the association 
indices for pairs of whales from the same pod were significantly higher 
than indices for pairs of whales from different pods. I would like to focus 
on associations between members of different pods in order to attempt to 
100 
I' 
I 
I 
-Within Pod Associations 0.50 
50 
~ 0.48 
~ ] 
0.46 ~ 
= 0 215 •.-4 
~ 0.44 I 
244 ~ 
•.-4 
I ~ 0 00 0.42 139 < l 0 •"'4 0.40 ~ ~ 0.38 
339 
0.36 142 I I 0.34 
0.32 
Ma:Mo Mo:Un Mo:Mo Ma:Ma Ma:Un Un:Un 
Age/Sex Class Pairs 
Figure 4.15. Average association indices (mean ± S.E.) for pair 
associations (excluding immatures) between members of the same pod 
grouped by the age and sex class of each of the members of the pair. Age 
and sex class abbreviations are for adult males (Ma), mothers (Mo), and 
unknowns (Un). Sample sizes for each bar indicate the total number of 
pairs for each age/sex class combination. 
Linked Pod Associations 
0.23 .......------------------------
29 86 
118 
0.21-
96 11 
1 l 
0.19- 11 
88 
0.15 - 166 
1 
I 0.17 -
0.13 ....__-..... ,----...--,---,..-----,------,-,----...--,----' 
Ma:Mo Mo:Un Mo:Mo Ma:Ma Ma:Un Un:Un 
Age/Sex Class Pairs 
Figure 4.16. Average association indices (mean ± S.E.) for pair 
associations (excluding immatures) between members of linked pods 
grouped by the age and sex class of each of the members of the pair. Age 
and sex class abbreviations are for adult males (Ma), mothers (Mo), and 
unknowns (Un). Sample sizes for each bar indicate the total number of 
pairs for each age/sex class combination. 
Separate Pod Associations 
160 
0.085 -
11 
479 
0.075 -
} 331 } 
380 624 
r r 
0.065 - 284 
r 
0.055 I I I I I I 
Ma:Mo Mo:Un Mo:Mo Ma:Ma Ma:Un Un:Un 
Age/Sex Class Pairs 
Figure 4.17. Average association indices (mean ± S.E.) for pair 
associations (excluding immatures) between members of different pods 
grouped by the age and sex class of each of the members of the pair. Age 
and sex class abbreviations are for adult males (Ma), mothers (Mo), and 
unknowns (Un). Sample sizes for each bar indicate the total number of 
pairs for each age/sex class combination. 
F 
identify levels of social organisation above the pod level, other than 
associations between linked pods, as defined in Section 4.2.10. 
Table 4.6 showed that only 8.1% of the 27999 possible between-pod 
pairs were recorded between the 245 resident adults seen on more than one 
day. However, this still represented 57% of all pairs observed (2258 of 3970). 
Many of these pairs were, however, seen infrequently. The average 
number of sightings for all 3970 pairs was 2.31 sightings per pair. Whales 
from the same pod were seen together 4.62 times, while pair of whales from 
different pods were seen together only 1.15 times. 
Another way to look at this is to examine the pod composition of the 
277 groups identified in the field. The mean number of pods identified in 
the groups was 1.66 ± 0.06. The majority of groups (59%) contained only one 
pod. Twenty-five pods were represented in these lone pod groups and I 
compared the number of individuals identified in these groups to the total 
number of individuals defined for that pod. The average was that 37.8 ± 
2.0% of the individuals in the pods were identified in the groups. Six of the 
groups were made up of all pod members, but these were for small pods. 
However, when the maximum percentages were extracted for each of the 25 
pods, the average rose to 66.4 ± 26.0%. There were few instances when all 
pod members were identified together in one group. 
There we_re 86 groups which contained whales from more than one 
unlinked pod. The number of pods represented by one or more members in 
these groupsranged from 2 to 10, and averaged, 2.78 ± 0.13, indicating a 
certain degree of fluidity in the social groupings of these pilot whales. 
There were 453 possible pair combinations between pods (excluding the 12 
associated pod combinations) and 144 (32%) were observed. However, there 
did not appear to be any consistent patterns in which pods travelled 
101 
. I 
I 
. I 
identify levels of social organisation above the pod level, other than 
associations between linked pods, as defined in Section 4.2.10. 
Table 4.6 showed that only 8.1 % of the 27999 possible between-pod 
pairs were recorded between the 245 resident adults seen on more than one 
day. However, this still represented 57% of all pairs observed (2258 of 3970). 
Many of these pairs were, however, seen infrequently. The average 
number of sightings for all 3970 pairs was 2.31 sightings per pair. Whales 
from the same pod were seen together 4.62 times, while pair of whales from 
different pods were seen together only 1.15 times. 
Another way to look at this is to examine the pod composition of the 
277 groups identified in the field. The mean number of pods identified in 
the groups was 1.66 ± 0.06. The majority of groups (59%) contained only one 
pod. Twenty-five pods were represented in these lone pod groups and I 
compared the number of individuals identified in these groups to the total 
number of individuals defined for that pod. The average was that 37.8 ± 
2.0% of the individuals in the pods were identified in the groups. Six of the 
groups were made up of all pod members, but these were for small pods. 
However, when the maximum percentages were extracted for each of the 25 
pods, the average rose to 66.4 ± 26.0%. There were few instances when all 
pod members were identified together in one group. 
There were 86 groups which contained whales from more than one 
unlinked pod. The number of pods represented by one or more members in 
these groupsranged from 2 to 10, and averaged. 2.78 ± 0.13, indicating a 
certain degree of fluidity in the social groupings of these pilot whales. 
There were 453 possible pair combinations between pods (excluding the 12 
associated pod combinations) and 144 (32%) were observed. However, there 
did not appear to be any consistent patterns in which pods travelled 
101 
I 
together. An analysis of the associations between pods found no distinct 
cluster s of pods (Figure 4.18). Table 4.8 shows the specific ratio association 
indices for the various pod combinations. There does not appear to be any 
higher level social structure which is clearly defined beyond the pod level. 
The seasonality of the occurrence of groups with more than one pod 
was examined in a similar fashion to the analyses done on group size 
(Section 4.3.3). Significantly more pods were observed together during the 
six-month summer (April-September: 1. 77 ± 0.09 pods/group) compared to 
the winter (1.48 ± 0.08; Mann-Whitney U = 10182.5, n(summer) = 171, 
n(winter) = 106, Z = 1.96, P < .05). The number of pods identified per group 
was significantly correlated with group size for both summer and winter 
(common rw = 0.637, P < .001; Figure 4.19). It can be seen that nearly all of 
the large groups, and the group with the greatest number of pods were 
identified during the summer. Thus, the summer months from April 
through September appears to be when pods are more likely to group up 
and travel together. 
4.4 DISCUSSION 
The original goals of this study were to identify individuals and 
define the patterns of social groupings. I will review these goals and then 
discuss the cont! ibution of my findings to an understanding of pilot whale 
social organisation and mating system. 
4.4.1 Pilot Whale Identification and Classification 
I have shown that individual pilot whales could be identified from 
high quality photographs. However, there was a wide degree of variability 
in the recognisability of these whales and an estimated 15% of the whales 
102 
11 
I 
II 
POD 
Q1 
Q2 
Q3 
RO 
NO 
J1 
J2 
uo 
10 
TO 
00 
so 
G1 
HO 
81 
K1 
K2 
G2 
EO 
F1 
F3 
PO 
F2 
A1 
A2 
LO 
MO 
Cl 
C2 
82 
DO 
DISTANCE COEFFICIENT 
0 .5 1 1.5 2 +-----------+----------+----------+----------+----+ 
_J 
I 
I 
I 
I 
I 
I I 
i--i 
-----,------_,__ _ __,J 
-------,---------, I I 
!--l I _______________ _, I 
--------.-----'1h I 
_ __....,_ _______ ___,~ 
I 
~ 
Figure 4.18. Dendogram for associations between pilot whale pods. Pod 
associations were tallied using a Ratio Association Index. Counts were 
totalled indepen4ently for each group observed (e.g. no association was tallied 
for pods seen in separate groups on the same day). Source: SPSS 
Hierarchical Cluster Analysis, Average Linkage Between Groups method. 
I 
Table 4.8. Ratio association indices for pods. Shading indicates below the mean and +1, +2 and +3 SD above the mean. Ql I I T I O I S 
D Ql Q2 
_Q3 
R 
N 
Jl 
J2 
u 
.087 1 1 
I 
T 
.125 
0 
1 s 
= Gl .048 l .045l .083l .019 l .024 .042 -H .050 l .048 l .087l .019 
-Bl 
.063 
-Kl 
.043 1.020 1.025 
.045 
K2 
.021 I 1.016 
.044 .016 I .070 I .019 l .224I 1 G2 I .043 
.042 1.019 1.024 
.0261.086 1 1.057 1.0341 1 E 
.040 1.025 
.027 1.0281 1.017 1.028 F1 I .094 
.0671.039 
.030 1.061 
.042 l .043l .050 l .021 F3 1.125 
p 
.111 I .027 1.038 
.045 F2 I .111 
.063 
.0611.125 Al 
.019 1 .024 
.027 
.0561.028 1.043 A2 
.044 1.029 
.03 1 I .032 I .087 
.032 1.033 1.024 L 
.026 
.016 1 I .027 M 
.025 1.014 
.019 .0201.0791 .047 
.031 1.034 1.031 1 .031 I .040 .020 Cl 
.048 I .059 I .Q70 
.115 1.036 1.020 .026 l .018l .018 l .015 
.018 l .085 l .018 l .024 C2 I .040 .038 
.0251.026 
.016 l .026l .083 l .063l .042 l .042l .040l .053 l .063 l .025 l .030l .229 B2 
.045 
.019 1 FU.U.U.U., 
D 
.Q36 l .019 I 1.074 1 I I I I 1.069 I .017 l.043 l .103 I.0671 1 !I ] below the mean; IJIJI! mean + 1 SD; • mean + 2 SD; • mean + 3 SD 
3 
25 
2 
15 
1 
5 
0 
0 
0 
0 
0 
0 0 
0 
• 
0 
+ 
0 
0 
+ 
0 
+ 
0 
• 
+ • + 
0 
0 
+ 
• 
0 
0 0 0 0 
• 0 0 0 
• • 
• • • 0 
• • + 0 
• • 0 
• • 
• • 0 
• • 0 
• 0 
• • 
• • 
+ 
0 
0 
- ------~-- --- - ----- ~ 
o Summer 
+ Winter 
0 
Q+--..-.-........ --.---------.....---...-.--........ ---
0 1 2 3 4 5 6 7 8 9 10 11 
Number of Pods/ Group 
Figure 4.19. T_he relationship between group size and the number of pods 
identified per group. Both group size and the number of pods/group were 
significantly larger during summer (April-September) compared to winter. 
t.o 
were unrecognisable. This was primarily duet. the high dependence of 
recognisability upon the presence of scars . Scars were stable throughout 
the course of the study (apart from the minor healing of fresh scars), but an 
animal without scars needed to have a distinctive fin shape to be 
recognisable. This dependence on chance events (e.g. obtaining a scar) 
limited the possibility of recognising all individual pilot whales. Some 
groups may have had a lower probability of being scarred (and thus a 
higher probability of being unidentifiable), perhaps related to dominance 
status, but since the context of scarring was never observed this cannot be 
confirmed or denied. Identification is certainly aided when there are 
permanent features of pigmentation which can be found on all animals (as 
is the case with the "saddle patches" of killer whales: Bigg et al., 1987). 
However, the technique is still applicable to less distinctively marked 
species. 
The definition of the age and sex classes for this study was based on 
the age, growth and reproductive maturity data from Japanese pilot 
whales, with the implicit assumption that these two populations had 
similar life history and growth patterns. The most likely differences would 
be in absolute body size, as was found between two distinct stocks of short-
finned pilot whales off Japan (Kasuya et al., 1988a). The fact that long-
finned pilot whales have very similar patterns of growth and dimorphism 
to the smaller s~ort-finned pilot whales (Sergeant, 1962a, 1962b; Kasuya et 
al., 1988b), suggests that these are generic traits and should be comparable 
between populations. 
The best-defined age and sex classes of the present study are the 
adult males and the mothers. Adult males should be unequivocally 
identified because mean size at sexual maturity was greater than the 
largest female in the entire Japanese sample of 565 females (Kasuya and 
103 
:i 
11 
,I 
II 
II 
I 
I 
11 
1 1 
1 , 
1 I 
1 11 
• I 
Matsui, 1984); the continued development of the characters of relative 
dorsal fin position, and the size and shape of the dorsal fin and melon were 
distinctly linked to this larger body size (Yonekura et al., 1980; Kasuya and 
Matsui, 1984). "Mothers" were distinct full-sized animals, smaller than 
males, and in continued association with specific immature animals. 
Although this relationship was assumed to be between a female and her 
offspring, it was impossible to distinguish a foster relationship. The long 
maturation period of pilot whales (9 yrs for females and 15 yrs for males 
{Kasuya and Marsh, 1984}), combined with the minimum calving interval 
of 3 yrs for primiparous females (Kasuya and Marsh, 1984), indicate that 
females will have a number of offspring with them at any given time. Even 
older females, which apparently extend the nursing period of the last calf, 
should have young up to the age of 8 yrs in attendance (Kasuya and Marsh, 
1984). This should guarantee that all recently reproductive females would 
have been classified as mothers in the present study. Immature animals 
were less distinct. The smallest animals were undoubtedly immatures. 
However, the transitions from calf to juvenile and juvenile to adult were not 
always clear. Variability in individual growth rates would cause different 
aged animals to be placed in the same class. The behavioural distinctions 
of relative time spent away from the company of adults gives some 
indication of relative independence, but since weaning has been shown to 
vary widely, depending on the age of the mother (Kasuya and Marsh, 1984; 
Section 2.5.3), even this may vary for similarly-aged animals. Finally, the 
greatest problems in defining the age and sex structure of pilot whale pods 
were due to the class of "unknowns", which comprised 38% of the whales. 
This class could have been composed of adult females without calves or 
adolescent males aged 8-12 yrs. The adult females could have been from the 
class of newly mature animals who had not yet had their first calf or 
alternatively they could have been post-reproductive females whose calves 
104 
-had all reached maturity. The young males potentially classified as 
unknowns were within the lower size range of "early-maturing" males in 
the Japanese sample (Kasuya and Marsh, 1984). The proportion of 
adolescent males should be small since this size range only applied to 
males for four years of their lives, compared to the much longer duration 
for which females maintained this body size. The ratio of early 
maturing males to resting females in solitary pods off Japan (considered to 
be a better representation of the age and sex structure of these whales than 
pods caught from aggregations) was 1:18, or 5.3% of all females without 
calves (Kasuya and Marsh, 1984; Table 35). This further suggests that the 
bulk of the unknowns were females. 
The development and application of methods of stereo-
photogrammetry to obtain actual body length measurements, as has been 
done for underwater photographs of sharks (Klimley and Brown, 1983), 
surface photographs of sperm whales (Gordon, 1987), and aerial 
photographs of spinner dolphins (Scott and Perryman, 1991), could reduce 
some of the limitations of relative size estimation. 
4.4~2 Residents and Visit.ors 
A large degree of variability was found in the patterns of occurrence 
of pilot whales. Some animals were seen on a regular basis, while others 
were seen only-infrequently. In order to examine the implications of this 
result, two simple definitions are proposed. Members of a group of animals 
seen only once without any associations with animals seen more than once 
were defined as visitors. Animals seen more than once, or animals seen 
once but in association with animals seen more than once, were defined as 
residents. I have shown that these differences in sighting histories were 
not due to sampling bias. Over half of the visitors observed were seen in the 
105 
pa 
first half of the study and a significant proportion of them were highly 
recognisable. These two factors gave them a high probability of being seen 
again. The fact that they were not strengthens the case for this definition of 
visitors. 
The range of distribution of these visitors is unknown. The visitors 
off Tenerife could be resident in other areas of the Canary Islands, 
although there have been no confirmed reports of other concentrations. It 
is also unlikely that whales would venture into an area such as Tenerife, 
which seems to be highly favoured by at least some pilot whales, and never 
return. Visitors may be sighted again in the Tenerife area as the study 
continues. Perhaps they simply return at intervals greater than the length 
of this study. 
However, it is not uncommon for there to be varying patterns of 
residency in cetacean populations. Although these animals do not appear 
to be territorial, using the conventional definition for terrestrial mammals 
(Norris and Dohl, 1980b), they often have favoured home ranges (Wells et 
al., 1987; Ballance, 1990). Killer whales in Puget Sound have distinct 
patterns of residency, with residents and transients differing in social 
structure and feeding ecology (Felleman et al., 1991). Although there are 
no detailed studies of free-ranging pilot whales which have demonstrated 
the presence of a transient class of animals, there are some anecdotal 
sources of information. 
In the harpoon whaling for short-finned pilot whales off the 
Antillean islands of St. Vincent and St., Lucia in the Caribbean Sea 
(Caldwell and Caldwell, 1975), whalers report that pilot whales respond in 
two different ways to being hunted by a small boat equipped with a hand-
held gun harpoon (Caldwell, pers. comm.). Some groups are very easy to 
106 
approach and one or two animals can be easily killed. Other groups are 
impossible to get anywhere near, and actively avoid the whaling boats. The 
implication is that some groups know the intention of the whaling boat, 
while others do not. Since this fishery only kills one or two animals from a 
group at a time, the survivors have an opportunity to learn about the 
intentions of the whalers. Any groups which were resident to the area 
would have the opportunity for multiple encounters with the whalers, 
allowing them to learn the whalers' intentions. The animals which do not 
avoid the whalers could be transient animals, unfamiliar with the area and 
with no previous experience of whaling on which to base a flight response. 
In summary, it is quite possible that the visitor groups of pilot whales 
described in this study represent a distinct community of whales. At the 
very least, I have shown that they do not occur in the Tenerife region with 
any great frequency. Further conclusions will have to await long range 
radio-tracking of the Tenerife pilot whales in order to better understand 
their ranging patterns. 
4.4.3 Reliability of Pod Definitions 
The term "pod" has been used to define consistent pilot whale groups 
identified in this study. The term was originally used by 19th century 
American whalers for any group or herd of whales. The definition of pods 
for stable groups of killer whales was based on the largest group of whales 
that travelled together for at least 50% of the time (Bigg et al., 1990). I feel 
that the term also applies to pilot whale groups, although the groups of pilot 
whales may not be as stable as killer whale groups. "Pod" is used in this 
sense to mean any group of whales with consistent associations between 
individuals. The duration of existence for a pod should be measured in 
months and years, not in hours and days. 
107 
Perhaps one of the strongest verifications for the pod divisions is in 
Table 4 .6. This shows that although only 13% of all possible pair 
combinations were observed overall, 96% of the possible within-pod pair 
associations defined by the cluster analysis were observed. Also, the level of 
these within-pod associations were found to be significantly higher than 
those between animals from linked or un-linked pods (Figure 4 .11). These 
two analyses also lend support to the idea of "linked" pods. The percent of 
the possible pair combinations between linked pods was high (80%), but the 
level of the associations was closer to that between completely different pods 
than it was to within-pod associations. However, the cluster analysis of pod 
associations was inconclusive (Figure 4.18). There was no evidence of 
discrete "clans", as has been described for killer whales (Bigg et al ., 1987; 
Ford, 1989; Bigg et al., 1990). This suggests that it is the relationships 
between specific individuals which form the links between linked pods . The 
linked nature of pods has been documented for long-finned pilot whales by 
parallel genetic variation in allele frequencies between separate pods 
(Amos et al. , 1991b). These parallels were closer for older animals, 
suggesting that linked pods shared common ancestors, but that the pods 
did split and younger animals were more distantly related (Amos et al., 
1991b) . 
4.4.4 Seasonality of Occurrence and Association 
The pattern of a seasonal variation in group size described here for 
pilot whales, with larger groups composed of mpre pods occurring during 
the summer months, is similar t o th e pattern described for killer whales 
(S.Heimlich-Boran, 1988; Felleman et al., 1991). For killer whales, this h as 
been rela ted to the seasonal influx of migrating salmon (J.Heimlich-Boran, 
1986a). I do not have sufficient information about prey resources in the 
108 
·i i 
111 
Canary Islands to make a similar association for pilot whales. However, 
the season of larger group size does correspond to the proposed conception 
period, based on back-calculations assuming a 12-month gestation (recently 
identified for long-finned pilot whales : A.R.Martin, pers. comm.) and a 
peak in births during the summer (June: this study, Section 4.3.1; July: 
Kasuya and Marsh, 1984). If these assumptions defining summer as the 
peak conception period apply to Canary Island pilot whales, then the 
increased association of whales from different pods during this season 
could be for mating. This would support the hypothesis that mating occurs 
between pods. Unfortunately, I never observed any sexual interactions in 
over 500 hours of observation. 
However, there is no indication of a strict seasonality to pilot whale 
births. Births of all populations have been shown to have a wide dispersion, 
albeit with a single summer peak (Sergeant, 1962a; Kasuya and Marsh, 
1984; Martin, pers. comm.; present study). There also did not appear to be 
any significant synchrony of ovulation in Japanese pilot whale pods 
(Kasuya and Marsh, 1984). This diffuse seasonality may be why multi-pod 
groups were observed throughout the year. In any event, the lack of a fixed 
breeding season is likely to have profound effects on the mating system 
(Glutton-Brock, 1989). 
It is not clear whether specific individuals have seasonal patterns of 
occurrence. There were no apparent seasonal patterns to the occurrence of 
visitor whales (Figure 4.4), and sample sizes were too small to test the 
occurrence of individual resident whales for seasonal variation in 
occurrence. It would also be difficult to look for seasonal variation in the 
identification of new whales, since the rate of new identifications changed 
throughout the course of the study (see Figure 4.6). 
109 
I 
4.4.5 Age and Sex Structure of Pilot Whale Pods 
There are still many unknown elements of pilot whale social 
organisation, primarily due to the incomplete data on age and sex 
structure. In spite of these difficulties, some evidence suggests that the 
adult sex ratio of Canary Island pilot whales was similar to the sex ratio of 
Japanese pilot whales. Kasuya and Marsh (1984) found that 22.6% of the 
adult population were males, giving an adult sex ratio of 1 male for every 
3.4 females (Kasuya and Marsh, 1984). Their life history model, which 
accounted for fishing bias for pods containing large males, estimated 
21.15% males and a sex ratio of 1:3.7. The highest female-biased adult sex 
ratio for Canary Island pilot whales would result from the assumption that 
all unknowns and all clean-finned animals were adult females. This 
would give a total adult female population of 376 females (109 mothers+ 218 
unknowns + 15% clean-finned animals), and would result in a sex ratio of 
21. 7% males or 1 male:3.6 females. This is probably a overestimate, but 
even if the unknowns were composed of 10% newly mature males, the sex 
ratio would still be equal to 1:3.4 (104 males:109 mothers + (218 unknowns -
10%) + 15% clean-fins). The sex ratio for long-finned pilot whales caught off 
Newfoundland was 1:3.7 (Sergeant, 1962a), equal to that for the Japanese 
life history model. 
There appeared to be fewer reproductive females in the Canary 
Islands than there were off Japan. Reproductive females composed 58.3% 
of all adult females from the actual whaling data and 52.4% for the life 
history model estimates (Kasuya and Marsh, 1984). Combined with the two 
corresponding adult sex ratios, the ratio of adult males to reproductive 
females was 1:2.00 and 1:1.95, for the original data and life history model, 
respectively. The age and sex classifications for the present study found the 
actively 
percent of the adult female population classified as/\ reproductive (i. e. 
110 
!I 
mothers) would be only 35. 7% (assuming 90% of unknowns to be females 
who are newlv mature or older females 
with calves n~ longer in attendanc~). The calculated male:mother ratio was 1:1.05. 
This · lower proportion of reproductive females could be due to improper 
assumptions concerning the unknown class of animals, or it could indicate 
real differences in reproductive rates between these two populations. For 
example, the whales off Japan have a history of exploitation, which could 
affect population density and local carrying capacity (see Fowler, 1984). 
Pods of Canary Island pilot whales were classified as productive and 
non-productive on the basis of the presence of immature animals. Two of 
the 14 non-productive pods were composed of only males, one was composed 
of three unknowns, while the rest were mixed males and unknowns. This 
could be an indication of male segregation into separate pods, especially if 
these unknowns were adolescent males. Other studies of pilot whales have 
reported pods with an apparent abundance of males, but there are no 
previous reports of all-male pods (Sergeant, 1962a; Kasuya and Marsh, 
1984). However, the two all male pods were visitors pods and thus seen only 
once. Additional animals in these pods could have remained unidentified. 
The association patterns of Canary Island pilot whales provided 
some insight into the nature of the relationships between the age and sex 
classes. The associations between adult males and the unknown class of 
whales were consistently higher than all associations involving mothers 
except when linked pods travelled together. Mothers' associations with 
adult males from linked pods were significantly higher than mothers' 
associations with each other and with the unknowns, and were not 
significantly different from the other male associations. If mothers are 
considered to be the reproductive females in the population, then their 
frequent associations with adult males from linked pods could support the 
hypothesis that mating is occurring outside of the pods. If mating was 
111 
I ' 
occurring within pods, higher association might be expected between adult 
males and females with calves from the same pods. Genetic studies of 
long-finned pilot whales indicate that males are not breeding within their 
own pods (Amos et al., 1991a; 1991b). 
The high association levels between adult males within the same 
pods may also correspond to the findings from genetic studies in Faeroese 
long-finned pilot whales. Adult males in the same pods had high degrees 
of relatedness, based on an excess of heterozygotes on single locus alleles, 
which are more common in siblings (Amos et al., 1991a; 1991b). Males 
within the same pod could be brothers maintaining associations into 
adulthood, as has been shown for chimpanzees (Goodall, 1986) and lions 
(Packer and Pusey, 1982; Packer et al., 1991). In any event, male pilot 
whales appear to be stable members of the pods outlined in this study. 
4.4.6 Conclusions 
The conclusions about pilot whale social organisation in the Canary 
Islands must still be considered tentative since observational studies 
utilising the identification of individuals require long-term effort to 
understand slowly-maturing, long-lived animals such as pilot whales. 
However, the findings from this study do not contradict the previously 
available biological data on pilot whales. Pilot whale pods exhibited stable 
group membership in groups of mixed age and sex throughout the two 
years of observation. Males were lacking from some of the productive pods, 
suggesting that there may not be a long-term ·bond between males and 
breeding females. Males exhibit.ed high levels of association with each 
other (supported by genetic studies of relatedness in long-finned pilot 
whales) and with reproductive females from other pods. They had low 
levels of association with the reproductive females in their own pod, 
112 
suggesting that mating might not occur within the pods (supported by 
DNA-fingerprinting and paternity exclusion studies in long-finned pilot 
whales). If mating was occurring within pods, higher associations would 
be expected between adult males and mothers. Groups tended to be larger 
and to consist of animals from different pods during the season when 
conception is likely to occur, suggesting that mating could occur when pods 
join together. The implications of a breeding system where mating occurs 
primarily between separate social groups suggest that the cohesive nature 
of pilot whale pods is based on non-reproductive benefits . Resource 
utilisation and predator defence have been proposed as benefits to group 
living. Although young pilot whales may be subject to low levels of shark 
and killer whale predation, it is likely that the ability to locate and capture 
patchy, clumped food resources may be a significant advantage of group 
formation in a three-dimensional, featureless environment. Squid, the 
pilot whales' favoured prey, is an obligate schooling animal (Hurley, 1978; 
Clarke, 1986) and it is likely that cooperation is required in its capture 
(Norris and Schilt, 1988). It would be beneficial to maintain long-term 
associations with the same individuals (preferably kin) to foster that 
cooperation (Axelrod and Hamilton, 1981). 
4.5 SUMMARY 
A 22-month field study on short-finned pilot whales was conducted off 
Tenerife in the Canary Islands off north-west Africa. Four-hundred-
ninety-five individual pilot whales were identi'fed from photographs of 
naturally-occurring marks and . scars on the dorsal fin and back. 
Individuals were classified into five age and sex categories: 107 adult 
males, 109 females with calves (or mothers), 219 unknowns (either adult 
females without calves or adolescent males), 33 juveniles (greater than two-
113 
thirds the size of the mother) and 31 calves (less than two-thirds the 
mother's size). Two classes of whales were defined by patterns of 
occurrence: residents and visitors. Residents were all animals seen more 
than once or seen once, but in the company of animals seen more than 
once. Visitors were defined as animals seen only once but not in the 
company of animals seen more than once. There were 15 distinct visitor 
groups, totalling 107 identifed animals and 16 clean-finned immatures. 
Two-hundred-seventy-seven groups of resident pilot whales were 
identified over 14 7 days. Associations between pairs of individuals were 
tallied to calculate a ratio association index and hierarchical, 
agglomerative clustering methods were applied to the matrix of association 
values to determine clusters of individuals, termed pods. Pod structure 
was analysed for age and sex composition, and pods were defined as 
productive or non-productive based on the presence of immature animals. 
Productive pods were significantly larger, although 12% of them were 
lacking adult males. Whales from different pods occasionally travelled 
together, and associations between age and sex classes showed significant 
differences depending on the pod membership of the individuals. Overall, 
associations involving adult males were higher than associations involving 
mothers. However, the highest ranked associations between mothers and 
members from their own pod were with other mothers, while the mothers' 
highest associations with other pods were with adult males. During 
summer (April-September), pilot whale groups were significantly larger 
and contained individuals from a significantly , greater number of pods, 
suggesting that pods were more likely to group up during the summer. 
These data support the hypothesis that male pilot whales remain in 
their natal pod and mate with females in associated pods. Further studies 
F 
on wild populations are needed, especially utilising DNA fingerprinting, in 
order to verify this hypothesis. 
115 
CHAPTER 5: COMPARATIVE SOCIAL ECOLOGY OF DELPHINIDS: 
TOWARDS AN UNDERSTANDING OF MATING SYSTEMS 
5.1 INTRODUCTION 
Mammalian mating systems show correlation with a variety of 
biological, social and ecological factors: the degree of sexual dimorphism 
(Ralls, 1977), relative brain size (Humphrey, 1976; Dunbar, 1992; 
Sawaguchi, 1992), relative testes size (Harcourt et al., 1981; Harvey and 
Harcourt, 1984; Parker, 1984), patterns of dispersal (Shields, 1987; Clutton-
Brock, 1989), habitat complexity (Verner and Willson, 1966), distribution of 
resources (Orians, 1969; Emlen and Oring, 1977), genetic heterogeneity 
(Chepko-Sade et al., 1987), parental care (Trivers, 1972; Clutton-Brock et al., 
1977) and the distribution of potential mates (Orians, 1969; Emlen and 
Oring, 1977; Andelman, 1986). The relative importance of each of these 
factors is variable across mammalian species, but general trends have 
demonstrated predictive value (see review in Clutton-Brock, 1989). 
The study of delphinid mating systems and social ecology is still 
under-developed compared to the relatively large body of equivalent work on 
many terrestrial mammals (see reviews in Norris and Dohl, 1980b; Wells et 
al., 1980; Gaskin, 1982; Evans, 1987; Wiirsig, 1989). However, long-term 
studies on some dolphin species are beginning to provide sufficient 
information on distribution, dispersal patterns, social relationships and 
ecology to allow comparable examination. Additionally, the newly available 
techniques of genetic analysis (Burke, 1989) being applied to cetaceans 
(Hoelzel, 1991b; Amos et al., 1992) are beginning to suggest unique patterns 
of mating (Amos et al., 1991a; 1991b; In press). This chapter reviews the 
117 
information currently available for delphinids utilising a comparative 
approach and proposes hypotheses about their mating systems. 
The comparative approach has become a widely used tool in the 
attempts to understand the adaptive significance of species differences in 
morphology and behaviour within mammalian orders (Harvey and Pagel, 
1991). These methods have been applied to primates (Clutton-Brock and 
Harvey, 1977; Clutton-Brock et al., 1977; Harvey et al., 1978; Clutton-Brock, 
1980; Harvey et al., 1980; Harcourt et al., 1981; Harvey and Bennett, 1983; 
Harvey and Harcourt, 1984; Harvey and Clutton-Brock, 1985), carnivores 
(MacDonald, 1983; Gittleman, 1986), bats (Eisenberg and Wilson, 1978), 
ungulates (Jarman, 1974) and other small mammals (Mace et al., 1981). 
Most work has examined variation in social organisation relative to ecology 
(Crook, 1970; Wrangham and Rubenstein, 1986). 
The first part of this chapter will be a broad examination of the 
cetacean family Delphinidae through a comparison of morphological and 
behavioural traits, specifically the relationships within and between six 
characters: body size, brain size, testes size, group size, diet and habitat 
preference. Trends are then compared to those used in current theories of 
mating systems. Three main questions will be addressed. First, what is 
the extent of sexual dimorphism within the family and can sexual 
dimorphism be related to group size? Second, what are the trends of 
relative brain size in relation to group size, diet, and habitat preference? 
Third, what are the trends in relative testes size related to group size and 
what are the implications for mating systems this suggests in terms of the 
importance of sperm competition (Parker, 1984)? 
Following this general examination is a more detailed review of the 
delphinid genera for which there is comprehensive information. Detailed 
118 
I , 
field studies examining group stability and group-specific habitat use 
patterns have been possible with the development of methods for the 
individual identification of cetaceans (Wiirsig and Wiirsig, 1977; Wiirsig 
and Jefferson, 1990). The most thorough studies have been conducted on 
Tursiops truncatus (Wiirsig and Wiirsig, 1977; Wiirsig, 1978; Wursig and 
Wiirsig, 1979; Shane et al., 1986; Wells, 1986; Wells et al., 1987; Ballance, 
1990; Wells, 1991; Connor et al., 1992b) and Orcinus area (Balcomb et al., 
1982; Bigg, 1982; Ford, 1984; Balcomb and Bigg, 1986; J.Heimlich-Boran, 
1986a; J.Heimlich-Boran, 1986b; S.Heimlich-Boran, 1986; Hoelzel and 
Osborne, 1986; Osborne, 1986; Bigg et al., 1987; J.Heimlich-Boran, 1988; 
S.Heimlich-Boran, 1988; Ford, 1989; Bigg et al., 1990; Olesiuk et al., 1990; 
Felleman et al., 1991; Baird, 1992). There is also some information on 
Stenella spp. (Norris and Dohl, 1980a; Norris et al., 1985; Wiirsig et al., 
1989; Pryor and Shallenberger, 1991; Wiirsig et al., 1991), Lagenorhynchus 
spp. (Wiirsig and Wiirsig, 1980; Wiirsig et al., 1989; Wiirsig et al., 1991) and 
Grampus (Kruse, 1989). An examination of the details of habitat use and 
social organisation for these genera can be used to test family-wide trends 
and to propose hypotheses for delphinid mating systems. 
To properly introduce trends in Delphinidae, it is instructive to first 
review mammalian trends in sexual dimorphism, relative brain size, and 
relative testes size. In addition to the classifications of group size, diet, and 
habitat I will present for delphinids, additional classification categories 
have been compiled for terrestrial mammals. These are mating system 
(Alexander et al., 1979), socionomic sex ratio (Clutton-Brock, et al., 1978), 
metabolic rate (Armstrong, 1983; Schmitz and Lavigne, 1984), trophic level 
(Wooton, 1987), and a variety of life history parameters such as length of 
gestation (Pagel and Harvey, 1988), age at first reproduction (Wooton, 1987), 
and lifespan (Pagel and Harvey, 1988). 
119 
Relationships have been clearly demonstrated between body size, 
group size and diet for both primates (Clutton-Brock et al., 1977) and 
ungulates (Jarman, 1974; Jarman and Jarman, 1979). For example, 
smaller antelope have more selective diets and disperse widely in less-open 
habitats, while larger antelopes show less diet selectivity and occur in 
larger groups (promoted by predator defence) in open grassland habitat 
(Jarman 1974; 1979). For primates as well, larger species tend to occur in 
larger groups (Clutton-Brock and Harvey, 1977). This trend is not 
exclusive, though: bears are one of the largest of carnivores and live 
solitarily (Caro, 1989; Gittleman, 1989). 
Sexual dimorphism in body size vanes widely within the 
mammalian orders (Ralls, 1977). In the majority of sexually-dimorphic 
species, the male is larger than the female (although see Ralls, 1976). 
Darwin (1871) proposed that bigger males were better equipped to compete 
with other males for access to females (intra-sexual selection), and in turn, 
that females could choose to mate with larger males (epigamic selection). 
The result of this sexual selection is that males show greater variation in 
reproductive success than do females (Bateman, 1948; Clutton-Brock, 1986). 
This has been shown to be the case for a wide variety of mammals: sexual 
dimorphism and variance in reproductive success is greater in species 
where a small number of males can deny other males access to a group of 
breeding female_s (Clutton-Brock and Harvey, 1977; Clutton-Brock et al., 
1977; Alexander et al. , 1979). Other explanations for sexual dimorphism 
have been based on ecological and energetic factors which will be presented 
in more detail in Section 5.4.1. 
Jerison (1973) reviewed the evolution of brain size and proposed that 
larger brains conferred enhanced abilities to collect, process and integrate 
complex environmental information. Wherever this conferred a selective 
120 
Relationships have been clearly demonstrated between body size, 
group size and diet for both primates (Clutton-Brock et al., 1977) and 
ungulates (Jarman, 1974; Jarman and Jarman, 1979). For example, 
smaller antelope have more selective diets and disperse widely in less-open 
habitats, while larger antelopes show less diet selectivity and occur in 
larger groups (promoted by predator defence) in open grassland habitat 
(Jarman 1974; 1979). For primates as well, larger species tend to occur in 
larger groups (Clutton-Brock and Harvey, 1977). This trend is not 
exclusive, though: bears are one of the largest of carnivores and live 
solitarily (Caro, 1989; Gittleman, 1989). 
Sexual dimorphism in body size vanes widely within the 
mammalian orders (Ralls, 1977). In the majority of sexually-dimorphic 
species, the male is larger than the female (although see Ralls, 1976). 
Darwin (1871) proposed that bigger males were better equipped to compete 
with other males for access to females (intra-sexual selection), and in turn, 
that females could choose to mate with larger males (epigamic selection). 
The result of this sexual selection is that males show greater variation in 
reproductive success than do females (Bateman, 1948; Clutton-Brock, 1986). 
This has been shown to be the case for a wide variety of mammals: sexual 
dimorphism and variance in reproductive success is greater in species 
where a small number of males can deny other males access to a group of 
breeding females (Clutton-Brock and Harvey, 1977; Clutton-Brock et al., 
1977; Alexander et al., 1979). Other explanations for sexual dimorphism 
have been based on ecological and energetic factors which will be presented 
in more detail in Section 5.4.1. 
Jerison (1973) reviewed the evolution of brain size and proposed that 
larger brains conferred enhanced abilities to collect, process and integrate 
complex environmental information. Wherever this conferred a selective 
120 
advantage, larger brains evolved. The best documented trends in brain size 
are related to the distribution of food resources and home range size (Mace 
et al., 1980). For example, frugivorous bats have larger brains than 
insectivorous bats (Eisenberg and Wilson, 1978). The proposed reason is 
because larger brains provide the increased ability to process the complex 
spatial and temporal information required to exploit a patchy, clumped 
resource such as fruit (Eisenberg and Wilson, 1978). For similar reasons, 
frugivorous primates have relatively larger brains than folivorous species 
(Clutton-Brock, 1980). The correlation between larger brain size and a 
larger home range is considered to be an indication of the fact that larger 
brains may be required to accommodate mental maps for larger areas 
(Clutton-Brock and Harvey, 1980; Mace and Harvey, 1983). 
Another hypothesis which links larger brain size to group structure 
has been termed the social complexity hypothesis (Humphrey, 1976; 
Dunbar, 1992; Sawaguchi, 1992). This proposes that a complex social life 
requires a complex intellect to process the information of multiple, 
overlapping relationships. Recent evidence for this comes from a 
comparison of neocortex volume and group size. Neocortical volume was 
found to be a function of group size; group size explained more of the 
variation in neocortical volume than did the ecological variables of percent 
fruit in the diet or day journey length (Dunbar, 1992). Dunbar (1992) 
proposed that, while the initial impetus towards larger brains in primates 
may have been due to ecological factors, brain evolution may have been 
accelerated by the potential for increasing social complexity and the 
resulting benefits of social coalitions and the ability to exploit knowledge of 
the behaviour of other individuals. In fact, Dunbar (1992) proposed that 
brain size may actually put an upper limit on group size. 
121 
The cetacean brain has always attracted much attention due its large 
size and convoluted appearance (Breathnach, 1960; Lilly, 1964; Ridgway et 
al., 1966; Elias and Schwartz, 1969; Jansen and Jansen, 1969; Pilleri and 
Gihr, 1969; Pilleri and Gihr, 1971; Morgane and Jacobs, 1972; Jerison, 1973; 
Ridgway and Brownson, 1984; Morgane et al., 1986a; Morgane et al., 1986b; 
Ridgway, 1986a; Ridgway, 1986b; Ridgway and Wood, 1988; Ridgway, 1990). 
Worthy and Hickie ( 1986) reviewed relative brain size in marine mammals 
and found that cetacean brains fit the general allometric relationship with 
body size for all mammals. Jerison (1986) has noted that large delphinid 
brains are apparently not essential in a three-dimensional environment 
(e.g. bird brains), in reduced gravity (e.g. human adaptability to life in 
space), in the aquatic environment (e.g. fish brains) or to acoustic 
processing (e.g. bat brains). It is more likely to be the ways in which they 
process external information in terms of environment and society which 
has resulted in the evolution oflarge brains (Jerison, 1986). It has also been 
noted that cetaceans of similar body size have very different sized brains 
(e.g. the delphinids Grampus, Orcinus, Lagenorhynchus and 
Cephalorhynchus have brains two to three times larger than the 
physeterids, ziphids and platanistids Kogia, Ziphius, Platanista and 
Pontoporia, respectively: Ridgway, 1986b). I examine relative brain size 
within the family Delphinidae in order to help identify ways in which 
larger brains may be used by these animals. 
The relative size of a male's testes has been shown to vary with 
mating system, and has been related to the poten~ial for sperm competition 
where larger sperm volumes are able to out-compete other male's sperm 
for fertilisation (Parker, 1984). In primates, species in which multiple 
reproductive males occur in the same group and in which females are 
likely to copulate with more than one partner per oestrus have larger testes 
122 
for their body weight than those where groups have a single breeding male 
and females invariably mate with only that male (Harcourt et al., 1981; 
Harvey and Harcourt, 1984). This often translates into the occurrence of 
increased male-male competition and increased sexual dimorphism in 
species with larger testes (Harcourt et al., 1981; Harvey and Harcourt, 1984; 
Kenagy and Trombulak, 1986). Kenagy and Trombulak (1986) found 
odontocete cetaceans to have the largest relative testes size of all mammals 
and predicted extreme multi-male breeding systems for these animals. 
They suggest that there are indications of high degrees of social 
interactions (referencing Gaskin, 1982), but they also speculate that aquatic 
copulation presents different selective pressures for a "high level of sperm 
delivery" perhaps related to the complex vaginal folds in female odontocetes 
(Kenagy and Trombulak, 1986). 
The comparison of social trends within Cetacea has been limited by 
available information. Wi.irsig (1989) reviewed the state of knowledge on a 
variety of aspects of the behaviour and ecology of all cetaceans. Wells et al., 
(1980) presented an overview of trends in social ecology within odontocetes, 
but primarily focussed on Wells' own work on Tursiops (see Section 5.3.6.1). 
One of the predominant relationships is between group size and habitat: 
river and coastal dolphins live in much smaller groups than most pelagic 
dolphins. Another trend was the variability in the occurrence of age and 
sex segregation (reported in Tursiops and Stenella , apparently absent in 
other species). There is a clearly a need for a complete review of delphinid 
social ecology. 
123 
5.2 l\'IE'IHODS 
The comparative analysis of variation in morphological trends is best 
examined by establishing relationships at one taxonomic level and 
examining the deviances from these relationships at the taxonomic level 
immediately below this (Clutton-Brock and Harvey, 1979; Harvey and Mace, 
1982; Harvey and Pagel, 1991). Species are often not valid for comparison 
because of the lack of independence between species characters within the 
same genera (Harvey and Pagel, 1991), I will be examining trends within 
the family Delphinidae and variation in these trends for the 17 delphinid 
genera. This has not substantially limited the data set since 11 of the 17 
genera are mono-specific. All data for the six multi-species genera (Sousa, 
Lagenorhynchus, Stenella, Lissodelphis, Cephalorhynchus and 
Globicephala) were calculated from means of the species values (presented 
in Table 5.1). 
The data used in this review were collected from a wide variety of 
sources. Body size and length and age at maturity was collected from field 
studies when possible, but additional information was collected from 
reviews of mammals in general (Poole, 1985) and cetaceans in particular 
(Bryden, 1972; Leatherwood and Reeves, 1983; Perrin and Reilly, 1984; Rice, 
1984; Klinowska, 1991). Mean values for body size were used when 
available, but in a few cases only maximum body size was reported. This 
should not affect the sexual dimorphism estimates since these were simply 
ratios of male and female values. Although body weight is usually the best 
variable to examine sexual dimorphism (Clutton-Brock et al., 1977), this 
was not available for eight species from six genera and the use of weight 
variables would have meant the loss of two genera (Lissodelphis and 
Feresa). There was a significant correlation between dimorphism in body 
weight and dimorphism in body length (Figure 5.1: Spearman Rank 
124 
Table 5.1. Summary of morphology, group size and ecology for the 32 species of delphinid 
cetaceans. Lengths in italics are maximum lengths, all others are means. See text for sources. 
Adult Length (m) Length Adult Weight (kg) Weight BRAIN DATA Testes (g) Group Size Ecological Type Subfamily Species male female Dimorphism male female Dimorphism Brain (g) Body (kg) N (comb.wt. min. max. Diet Habitat 
STENINAE Steno bredanensis 2.50 2.45 1.020 107 91 1.176 2300 10 50 1 4 
Sousa chinensis 2.80 2.44 1.148 285 260 1.096 4 30 I I 
Sousa teuszii 2.00 1.95 1.026 3 25 I I 
Sotalia fluviatilis 1.87 1.82 1.027 43 40 1.075 688 42 I 2120 2 30 I l 
DELPHININAE Lagenorhynchus albirostris 2.60 2.59 1.004 267 266 1.004 3 100 1 3 
Lage11orhy11chus acutus 2.50 2.24 1.116 160 160 1.000 1200 120 l 740 7 700 l 3 
Lage11orhy11chus obscurus 1.88 1.91 0.984 1600 10 200 l 3 
Lage11orhy11chus obliquidens 1.75 1.60 1.094 90 90 I.OOO 1102 85 6 1118 50 1000 1 3 
Lage11orhy11chus cruciger 1.83 1.83 1.000 150 150 I.OOO 200 1000 I 3 
Lage11orhy11chus austra/is 2.16 2.16 I.OOO 1 3 
Grampus griseus 3.70 3.42 1.082 385 335 1.149 2411 317 2 12520 15 300 2 3 
Tursiops truncatus 2.70 2.54 1.063 275 220 1.250 1680 151 12 1966 15 500 1 3 
Stene/ la frontalis 2.20 2.20 I.OOO 1956 8 50 1 4 
Stene/la attenuata 2.05 1.95 1.051 91 62 1.468 700 10 3000 1 4 
Stene/la longirostris 1.79 1.76 1.017 60 60 I.OOO 900 15 1000 I 4 
Stene/la clymene /.85 1.85 I.OOO 60 60 I.OOO 100 500· I 4 
Stenella coeruleoalba 2.39 2.26 1.058 158 136 1.162 835 62 9 450 25 3000 I 4 
Delphinus de/phis 2.40 2.10 1.143 115 85 1.353 836 68 10 3344 100 1000 l 3 
Lagenode/phis hosei 2.36 2.35 1.004 163 163 I.OOO 2490 100 800 l 4 
LISSODELPHININAE Lissodelphis borealis 2.63 2.17 1.212 100 300 l 3 
Lissodelvhis veronii 2.30 2.36 0.975 75 1000 I 3 
CEPHALORHYNCHINAE Cephalorhynchus commersonii 1.36 1.39 0.978 60 60 I.OOO 9 100 1 2 
Cephalorhynchus eutropia 1.53 1.51 1.013 62 62 I.OOO I I 40 1 2 
Cephalorhynchus heavisidii 1.30 1.30 I.OOO 5 20 1 2 
Cephalorhynchus hectori 1.25 1.37 0.912 36.8 45 0.818 5 20 1 2 
GLOBICEPHALINAE Peponocephala electro 2.68 2.68 I.OOO 125 125 I.OOO 2200 150 1500 2 4 
F eresa attenuata 2.64 2.43 1.086 754 30 300 2 4 
Pseudorca crassidens 5.32 4.47 1.190 1727 1270 1.360 14800 30 50 3 4 Orcinus orca 6.70 5.80 1.155 4000 2750 1.455 5618 2049 3 46200 15 100 3 3 
Globicephala me/as 5.57 4.89 1.139 2224 1524 1.459 2511 1000 3 14400 30 1000 2 4 
G/obicevhala macrorhvnchus 4.73 3.64 1.299 1260 563 2.238 6940 15 500 2 4 
ORCAEUINAE Orcaella brevirostris 2.35 2.32 1.013 97 96 1.010 6 20 l l NOTES: Diet: 1) fish, 2) squid, 3) meat; Habitat: 1) estuarine, 2) coastal, 3) shelf, 4) pelagic. Maximum lengths shown in italics, 
1.7 
~ 
1.6 + Steno 
0 Sousa 
-+,J 
.A Sotalia 
'So 1.5 
• Lagenorhynchus ..... Q) 
D Grampus s: 
~ 1.4 Ii Tursiops ..... 
s V + Stenella en Cl Delphinus ..... ~ 1.3· X Lagenodelphis fr 
0 () Cephalorhynchus s 
..... <> Peponocephala 0 1.2 
....... • V Pseudorca cd + D ~ 'Y Orcinus 
Q) 1.1 0 ~ Globicephala w. A 
• Orcaella 
1 <ac• • 
() 
n 
.v 
.95 1 1.05 1.1 1.15 1.2 1.25 
Sexual Dimorphism in Length 
Figure 5.1. Comparison of sexual dimorphism in length with sexual 
dimorphism in weight for 15 genera. Points for multi-specific genera are 
averages of species values. The two measures were significantly correlated (Spearman r = 0.888, t = 6.96, df = 13, P < .001). 
II 
Correlation r 8 = 0.888, t = 6.96, df = 13, P << .001), therefore sexual 
dimorphism in body length was used for all analyses. Brain size data was 
derived entirely from reviews (Breathnach, 1960; Lilly, 1964; Ridgway et al., 
1966; Pilleri and Gihr, 1969; Pilleri and Gihr, 1971; Morgane and Jacobs, 
1972; Bonin, 1973; Ridgway and Brownson, 1984; Morgane et al., 1986b; 
Ridgway, 1986b). Data on individual brain and body weights were available 
for 46 specimens from nine species (Table 5.1); median values were used 
when more than one individual was available from any given species. For 
comparative analyses, the two Lagenorhynchus species were pooled to 
create a generic mean value. Testes weights were extracted from field 
reports and general reviews (Harrison, 1969; Perrin and Reilly, 1984). 
Additional sources are listed in the footnotes of Table 1.1. 
Body size has been shown to have a confounding effect on the 
comparison of morphological characters due to allometric growth (Harvey 
and Mace, 1982; Harvey and Pagel, 1991). However, residuals from the 
straight line relationship of log transformed characters regressed on log 
body size will accurately describe character variation with the effects of body 
size removed (Harvey and Pagel, 1991). For this study, all measurements 
(body size, brain size and testes size) were log-transformed and the line of 
best fit was calculated by reduced major axis analysis (Imbrie, 1956). 
Reduced major axis analysis is recommended because it calculates a line of 
best fit without assuming a dependent-independent relationship between 
the two variables, as is the case for linear regression analysis (Harvey and 
Mace, 1982). In other words, it allows for the. potential of error in the 
measurement of both of the variables, as is likely to be the case with the 
measurements presented here. Although Harvey and Pagel (1991) have 
recommended the use of major axis analyses over reduced major axis 
125 
analyses, the latter has been shown to be more robust when errors are 
unknown (Rayner, 1985). 
Social and ecological categories of delphinids were assigned from 
general descriptions of distribution, behaviour and diet (Fitch and 
Brownell, 1965; Wells et al., 1980; Gaskin, 1982; Leatherwood and Reeves, 
1983; Clarke, 1986; Evans, 1987; Klinowska, 1991). Minimum and 
maximum group sizes were derived from field studies where available; in 
many cases, group sizes were only available from small samples of chance 
encounters at sea. The two categories may be considered similar to the 
distinction between core groups and population groups which share a home 
range (Gittleman, 1989). When detailed field studies were available, 
minimum group sizes were derived from reported mean group sizes. 
Ecological parameters were classified into combinations of two 
categories: diet and habitat. Diet was classified into three categories: 1) 
feeding on fish (icthyophagous), 2) feeding on squid (teuthophagous), and 3) 
feeding on other marine mammals (sarcophagous: Gaskin, 1982). These 
diet categories were not exclusive, many of the species classified as fish-
eaters were also reported to occasionally take squid and vice versa. The 
assignment of a species to these diet categories was based on predominant 
prey, and thus the categories are broad generalisations. On the other hand, 
the classification of meat-eaters was based on the presence of even limited 
cases of hunting other marine mammals. This was especially true for 
Pseudorca: there have been no detailed studies of diet for this species, only 
stomach contents from strandings (mostly fish) and occasional reports of 
them hunting marine mammals (Perryman and Foster, 1980). 
Additionally, in Orcinus, which is noted for its predation on other marine 
mammals, some populations feed primarily on fish (Jefferson et al., 1991). 
126 
I 
Therefore, as even rare occurrences of hunting warm-blooded prey might 
reflect on brain size, I included the diet category of meat eaters. Habitat 
zone was classified into four categories: 1) estuarine, or occasionally 
entering fresh water, 2) coastal, 3) shelf, or ranging from nearshore to the 
continental shelf, and 4) pelagic, or always offshore. 
Detailed data on social systems were only available for the 
subfamilies Delphininae and Globicephalinae, and were compiled from 
existing accounts and my own observations on Globicephala (from 
Chapters 3 and 4) and Orcinus (Heimlich-Boran and Heimlich-Boran, 
ms.). 
5.3 RESUL'IS 
5.3.1 Ecological Classification 
The classification of diet and habitat for all delphinid species 1s 
presented in Table 5.1. There were twelve possible combinations of the 
three diet and four habitat categories, but only eight were observed. Fish 
feeders were found in all habitats, but there were no squid feeders which 
were predominately coastal or estuarine, simply because of the primary 
shelf and pelagic distribution of squid (Clarke, 1986). The genera which 
hunted other marine mammals were found only along the continental shelf 
and in the pelagic zone. 
5.3.2 Sexual Dimorphism 
Sexual dimorphism in body length in relation to mean body length of 
males and females for the 17 genera in the family Delphinidae is presented 
1n Figure 5.2. Sexual dimorphism was significantly correlated with 
127 
~ 
..c:: 
bo p 
(J) 
-(J) (rj 
s 
~ 
! p 
(J) 
-(J) (rj 
s 
'-' 
s 
r:n. 
...... 
..c:: 
e 
0 
s 
...... 
~ 
-ro ~ 
(J) 
UJ. 
1.3 
1.30 Dimorphism = 0.34 log Avg. Length (m) + 0.93 
• Steno 
0 Sousa 1.25 ~ .A Sotalia 
•Lagenorhynchus 
1.20 V D Grampus 
I:,.. Tursiops 
T + Stenella 1.15 Cl 
D Delphinus 
1.10 
~181 
X Lagenodelphis 
D • Lissodelphis 
I:,.. 0 Cephalorhynchus 
... <> Peponocephala 
,. <> 181 Feresa 
V Pseudorca () 
T Orcinus 
.9 ~ Globicephala 
• Orcaella 
.90 
1 2 3 4 5 7 10 
Average Body Length (m) 
Figure 5.2. Sexual dimorphism in body length plotted against average of 
adult male and adult female body length for 17 genera of Delphinidae. All 
data were calculated from generic averages of species values. 
average adult body length (rs = 0.51, t = 2.296, df = 15, P < .05). The genus 
Cephalorhynchus, the smallest delphinids, was the only genera in which 
females were larger than males (mean body sizes of 1.39 m vs. 1.36 m; 
male: female length ratio = 0.978). This situation was most pronounced for 
C. hectori with females almost 10% longer than males (Table 5.1). Males 
were larger than females in all other genera. Although the degree of 
sexual dimorphism was significantly correlated with body size, the genera 
with the greatest sexual dimorphism, Globicephala (mean body sizes of 5.11 
m for males vs. 4.12 m for females; male: female length ratio = 1.24), 
ranked only third largest in size. The medium-sized dolphins ranged 
between these two extremes (Figure 5.2). 
Comparisons were made between the two best-represented 
subfamilies: Delphininae (six genera) and Globicephalinae (five genera). 
When all genera were included, there were no significant differences in 
sexual dimorphism (Delphininae vs. Globicephalinae: length ratio: 1.06 ± 
0.020 vs. 1.13 ± 0.042, Mann-Whitney U = 23, Z = 1.46, P > 0.05; weight ratio: 
1.15 ± 0.056 vs. 1.37 ± 0.139, U = 18.5, Z = 1.39, P > 0.05). However, data for 
Peponocephala, one of the least known Globicephalinae genera (Nishiwaki 
and Norris, 1966) were published as exactly the same lengths for males and 
females. This was suspected to be due to incomplete data, so the genus was 
removed. After this removal, there were significant differences in sexual 
dimorphism between the two subfamilies (recalculated Globicephalinae: 
length ratio: 1.17 ± 0.032, U = 23, Z = 2.34, P < 0.02; weight ratio: 1.5 ± .091, U 
= 18, Z = 2.32, P < 0.05). 
Sexual dimorphism was compared for the ecological classifications 
of the genera (Figure 5.3). There were no significant differences which 
could be related to the eight ecological categories of diet and habitat 
(Kruskal-Wallis H = 8.144, df = 7, P > 0.05). Diet and habitat were also tested 
128 
Pseudorca 
Orcinus 
Globicephala 
Feresa 
Peponocephala 
Grampus 
Lagenodelphis 
Stenella 
Steno 
Lagenorhynchus 
Lissodelphis 
Delphinus 
Tursiops 
Cephalorhynchus 
Orcaella 
Sotalia 
Sousa 
. 
. 95 
Cl 
+ 
ll. 
ll. 
ll. 
• Fish/Estuarine 
D o Fish/Coastal 
A Fish/Shelf 
• • Fish/Pelagic 
• 
D Squid/Shelf 
ll. Squid/Pelagic 
• + Meat/Shelf 
• 
Cl Meat/Pelagic 
• 
• 
• 
0 
• 
• 
• 
. 
. 
1 1.05 1.10 1.15 1.20 1.25 
Sexual Dimorphism (male length/female length) 
Figure 5.3. Sexual Dimorphism plotted according to ecological 
classification fo:r 17 genera of Delphinidae. 
I 
·I 
I I 
separately, but there were still no significant differences (diet: H = 3.757, df 
= 2 , P > 0.05; habitat: H = 5.378, df = 3, P > 0.05).. There were also no 
significant correlations for sexual dimorphism and group size (Figure 5.4: 
minimum group size r 5 = 0.057, t = 0.221, df = 7, P > 0.05; maximum group 
size r 5 = 0.064, t = 0.248, df = 7, P > 0.05). · 
5.3.3 Relative Brain Size 
Brain size is known to vary allometrically with body size (Gould, 1966; 
Jerison, 1973). Brain weight was significantly correlated with body weights 
for eight genera of Delphinidae (Figure 5.5: r = 0.97, t = 9.59, df = 7, P << 
.001. Reduced major axis analysis resulted in a linear equation for the 
natural log-transformed data: 
ln Brain Weight (g) = .51 ln Body Weight (kg)+ 4.64 
The slope of this line (0.51 ± 0.04) was equivalent to estimates made using 
all odontocete cetaceans (0.55 ± 0.17, principle components analysis, n = 24 
species: Worthy, 1986). The slope for all mammals has been shown to be 
approximately 0.75 and is scaled with metabolic rate, but a decrease in 
slope is common when comparing lower taxonomic levels (a process known 
as the taxonomic scaling effect: Pagel and Harvey, 1989). 
Residuals calculated from the reduced major axis were termed 
Relative Brain Sizes (RBS) and show the degree to which the different 
genera deviated from family-wide trends, independent of body size (Harvey 
and Pagel, 1991). Relative Brain Sizes are plotted for all genera in relation 
to ecological classification in Figure 5.6. Tursiops had the largest RBS, 
with a brain 21 % larger than that predicted by the reduced major axis fit. 
Grampus and Orcinus also had brain sizes larger than expected (19% and 
8%, respectively) . Globicephala had the smallest RBS (36% less than 
129 
I 
I I 
I 
A) 
~ + 
• Steno 
100 Cl 0 Sousa 
IC ~ .A Sotalia 
• • •Lagenorhynchus I::.. D Grampus (l) 
N I::.. Tursiops ..... CB Cl') 
+ Stenella 0.. 
:::s 
Cl Delphinus 0 lo-< 
v IC Lagenodelphis § 10 + Lissodelphis 
8 () Cephalorhynchus ..... 
>< ~ Peponocephala cd 
• V ~ () Ill Feresa 
.. 0 V Pseudorca 
T Orcinus 
• ~ Globicephala 
• Orcaella 
1 
.95 1 1.05 1.1 1.15 1.2 1.25 
Sexual Dimorphism (male length/female length) 
B) 
100 
(l) 
N 
..... 
Cl') 
0.. 
:::s 
8 
v 
§ 10 
8 
..... 
i::: 
:§ 
. 
0 
• 
• 
• 
+ 
• 
I::.. D 
0 
Cl 
V 
" 
" 
• Steno 
0 Sousa 
.A Sotalia 
•Lagenorhynchus 
D Grampus 
I::.. Tursiops 
+ Stenella 
Cl Delphinus 
IC Lagenodelphis 
+ Lissodelphis 
() Cephalorhynchus 
~ Peponocephala 
Ill Feresa 
V Pseudorca 
T Orcinus 
~ Globicephala 
• Orcaella 
1----------.-------.---..-----.--......... ----...---
.95 1 1.05 1.1 1.15 1.2 1.25 
Sexual Dimorphism (male length/female length) 
Figure 5.4. Sexual dimorphism in body length plotted against A) 
maximum and B) minimum mean generic group size. 
7000 
Orcinus + 
5000 
,.-,. 
b.() 
'-' 
.µ 
...c:: 
b.() Grampus_. Cl ..... 
Cl) Globicephala ~ 2000 ~ 
..... Tursiops • ro 
I,.., 
r:Q 
0 
1000 Lagenorhynchus 
Stenella 
eh Delphinus 
700 
• Sotalia 
500 
30 50 100 200 500 1000 10000 
Body Weight (kg) 
Figure 5.5. Brain weight plotted against body weight for eight genera of 
Delphinidae. The data were significantly correlated (r = 0.97, t = 9.59, df = 7, 
P << .001). 
I 1 
1
11 
Tursiops · • Fish/Estuarine Delphininae 0 Fish/Shelf 0 
£ Fish/Pelagic 
Grampus • Squid/Shelf 
• Delphininae D Squid/Pelagic 
1:,,. Meat/Shelf 
Orcinus Globicephalinae 
Lagenorhynchus 0 Delphininae 
Sotalia • Steninae 
Stenella Delphininae 
Delph in us 0 Delphininae 
Globicephala D Globicephalinae 
-.4 -.3 -.2 -.1 0 .1 .2 .3 
Relative Brain Size 
Figure 5.6. Relative Brain Size plotted by ecological classification for eight 
genera of Delphinidae. 
'I 
I 
expected), while Delphinus was also less than expected (8%). The other 
genera were all within 5% of the predicted values. None of this variability 
could be explained by the combined ecological classifications (Figure 5.6: 
Kruskal-Wallis H = 4.0, df = 5, P >> 0.05). Separate tests were run on diet 
alone and habitat alone, but there were no significant trends (diet: H = 0.5, 
P >> .05; habitat: H = 3.1, P > 0.05). Comparisons of RBS for minimum and 
maximum group sizes are presented in Figure 5.7. There were no 
significant correlations for either of the group size estimates (minimum 
group size rs = -0.51, t = 1.46, df = 7, P > 0.05; maximum group size rs = 
-0.57, t = 1.71, df = 7, P > 0.05). 
5.3.4 Relative Testes Size 
The combined weight of both testes was examined in relation to male 
body weight for 13 delphinid genera (Figure 5.8). The relationship between 
testes weight and body weight was described by the reduced major axis line 
of best fit: 
ln Testes Weight (g) = 0.95 ln Body Weight (kg)+ 2.79 
This relationship was significant (r = 0.86, t = 5.64, df = 12, P < 0.001). The 
slope of the line (0.95 ± 0.17) was higher than the common slope for all 
primates of 0.66 (Harvey and Harcourt, 1984) or the value of 0.72 reported 
for all mammals (Kenagy and Trombulak, 1986), but since error values 
were not available for these other slopes, the significance of the difference is 
unknown. Following Harvey and Harcourt (198:4), Relative Testes Sizes 
(RTS) were calculated as observed/expected ratios from the line of best fit 
and are presented in Figure 5.9. There were no clear patterns in the 
taxonomic distribution of RTS values : members of all subfamilies had 
representatives with larger than expected testes size. There were also no 
130 
' ! 
A) .3 
.2 
.1 
Cl) 
N 
..... 
w 
~ 0 ..... 
ro 
~ 
~ 
Cl) 
-.1 > ..... 
+'I 
ro 
....... 
Cl) p:::; 
-.2 
-.3 
-.4 
10 
B) .3 
.2· 
.1 
Cl) 
N 
..... 
w 
i:: 0 ..... 
ro 
~ 
~ 
Cl) 
-.1 > ..... 
+'I 
ro 
....... 
Cl) p:::; 
-.2 
-.3 
-.4 
1 
D 
100 
Maximum Group Size 
10 
I::,,. 
D 
• 
+ 
Minimum Group Size 
• 
Cl 
1000 
Cl 
100 
+ 
A. Sotalia 
•Lagenorhynchus 
D Grampus 
I::.. Tursiops 
+ Stenella 
Cl Delphinus 
T Orcinus 
~ Globicephala 
A. Sotalia 
aLagenorhynchus 
D Grampus 
I::.. Tursiops 
+ Stenella 
Cl Delphinus 
T Orcinus 
~ Globicephala 
Figure 5.7. Relative Brain Size plotted for A) maximum and B) 
minimum group size. 
I 
I 
,-.._ 
~ 10000 
·+.J ~ 
..... 
Cl) 
~ 5000 
r/1 
Cl) 
-+'> 
r/1 
Cl) 
~ 
1000 
+ 
o •x a 
A 
D 
• 
• 
0 
Ill 
• Steno 
0 Sotalia 
Ji..Lagenorhynchus 
• Grampus 
D Tursiops 
A Stenella 
+ Delphinus 
a Lagenodelphis 
X Peponocephala 
+ Feresa 
0 Pseudorca 
<> Orcinus 
Ill Globicephala 
500+-,--.-.......--r--~~.-----r--.,..........--.-................... ~~--.-~-----+ 
50 100 1000 
Body Weight (kg) 
500) 
Figure 5.8. Combined testes weight (g) plotted against body weight (kg) 
on a logarithmic scale for 13 genera of Delphinidae. 
I 
I I 
I 
I 
I I 
Grampus 
• Delphininae 
Delphinus + Delphininae 
Sotalia 0 Steninae 
Steno • Steninae 
Peponocephala IC Globicephalinae I 
Lagenodelphis · Cl ~ Delphininae I 
Orcinus (> Globicephalinae 
Stenella I::.. Delphininae 
Pseudorca C) Globicephalinae 
Lagenorhynchus ... Delphininae 
1 I 
Tursiops D Delphininae 
Globicephala IHI Globicephalinae 1 I 
Feresa • Globicephalinae I 1 
0 .5 1 1.5 2 2.5 
! I 
Relative Testes Size I 
I 
Figure 5.9. Relative Testes Size plotted for 13 genera and three 
subfamilies of Delphinidae. 
significant correlations between testes size and group size (Figure 5.10; 
minimum group size r = 0.19, t = 0.20, df = 12, P >> 0.05; maximum group 
size r = -0.01, t = 0.01, df = 12, P >> 0.05). 
An inverse relationship has been proposed between sexual 
dimorphism and relative testes size in light of the fact that males in single 
male breeding systems are highly dimorphic (due to increased male-male 
competition) and have relatively small testes (because they can control 
access to the female(s) in their group (Harvey and Harcourt, 1984; Fig. 2). I 
compared the sexual dimorphism (male/female length ratios) and Relative 
Testes Size values and found no significant correlation (Figure 5.11; rs = 
-0.36, t = 1.27, df = 12, P > 0.05). 
5.3.6 Delphininae Social Systems 
5.3.6.1 Turswps 
The bottlenose dolphin, Tursiops truncatus, is probably the best-
known delphinid. The species is found in coastal waters around the world 
and has been shown to adapt to a wide variety of environmental conditions 
(Shane et al., 1986). The diet of bottlenose dolphins shows a wide degree of 
variability (Barros and Odell, 1990), and they are considered to be one of the 
most generalised dolphins (Barnes, 1990). Life history, hormonal and 
observational studies of bottlenose dolphins have indicated that males and 
females have different rates of maturation, with males maturing at around 
10 years of age and females maturing around six or seven years of age 
(Sergeant et al ., 1973; Hohn, 1980; Wells, 1986). Lifespan estimates of 44 
years for females and 33 years for males are based on the oldest known 
individuals aged by dentinal tooth rings (Hohn et al., 1989), indicating that 
males have higher mortality rates than females. Recent studies have 
131 
1 
II 
A) 1.2 
1 D 
.75 a 
JJ.. 
• Steno 
.5 • 0 Sotalia Cl) 
JJ..Lagenorhynchus N <> ..... 
.25 r.t). 
• Grampus r:n • Cl) D Tursiops ...,;, r:n 
I::,,. Stenella Cl) E-i 
Cl) 
-.25 V + + Delphinus > 
• a Lagenodelphis :;s Cd 
-.5 IC Peponocephala ....... IJ.. Cl) ~ ~ 
• Feresa 
-.7 () Pseudorca 
<> Orcinus 
IHI Globicephala 
-1.2 
-1.5 
10 100 1000 
Maximum Group Size 
B) 1.25 
1 D 
.75 
• Steno 
.5 • 
0 Sotalia 
Cl) 
JJ..Lagenorhynchus N ..... 
<> r.t). 
.25 • Grampus r:n 
• Cl) D Tursiops ...,;, 
r:n 0 IJ.. Stenella Cl) E-i 
+ Delphinus Cl) 
-.25 v > a Lagenodelphis ..... 
• 
...,;, 
Cd 
-.fr IC Peponocephala ....... I::,,. Cl) ~ ~ • Feresa 
-.75 () Pseudorca 
<> Orcinus 
-1· IHI Globicephala 
-1.25 a 
-1.5 
1 10 100 
Minimum Group Size 
Figure 5.10. Relative Testes Size (RTS) compared to A) maximum and 
B) minimum group size for 13 genera of Delphinidae. 
11 
I I 
! 11 
1.25 
1 0 
.75 Cl 
... 
• Steno 
.fr • .& Sotalia (l) 
N 
• Lagenorhynchus ..... ~ 00 
.25 D Grampus Cll 
(l) 
• l:::.. Tursiops ..µ Cll 
(l) + Stenella ~ 0 
Cl Delphinus (l) 
> X Lagenodelphis ..... ..µ 
+ c,j 
-.25 <> Peponocephala ....... V (l) 
~ • IHI Feresa 
-.5 V Pseudorca 
-., Orcinus 
~ ~ Globicephala 
-.75 
-1 
-1.25 IHI 
111 I, 
-1.5 I I 
.975 1 1.05 1.1 1.15 1.2 1.25 
Sexual Dimorphism (male length/female length) I 
Figure 5.11 Relative Testes Size plotted against sexual dimorphism. 11 
1
1111 I 1 
I 
provided no evidence for sexual dimorphism (Hersh et al., 1990). Births 
appear to be diffusely seasonal; they occur during all months, but with a 
peak of births off Florida in June and July, corresponding to the period of 
highest water temperature (Wells et al., 1987). 
The most complete studies of the social organisation of bottlenose 
dolphins have been conducted since 1970 along 160 km of the west coast of 
Florida (Irvine and Wells, 1972; Wells, 1978; Wells et al., 1980; Irvine et al., 
1981; Irvine et al., 1982; Wells, 1986; Wells et al ., 1987; Hohn et al., 1989; 
Scott et al., 1990; Wells and Scott, 1990; Duffield and Wells, 1991; Wells, 
1991), but observations of habituated Tursiops off western Australia have 
also contributed some well-documented short-term observations (Connor 
and Smolker, 1985; Connor et al., 1992a; 1992b). The Florida study has used 
a diverse array of techniques, including radio-tracking, the identification of 
animals from naturally occurring marks and periodic capture/release 
programs (Wells, 1991). Known animals were captured to measure, sex 
and to collect teeth samples for ageing, blood samples for hormonal 
analyses and DNA for genetic analyses (Wells and Scott, 1990). 
Three communities of dolphins, with mutually exclusive home 
ranges, have been identified off Florida (Wells, 1986; Wells et al ., 1987). 
Data from the central community, termed the Sarasota community, 
generated the bulk of the conclusions on Tursiops social organisation. This 
community was comprised of 84 dolphins, with an adult sex ratio of 26 
females (70%) to 11 males (30%), or 2.4 females per male (Wells et al., 1987). 
Different patterns of habitat use indicated a high degree of age and sex 
segregation. Only 31 % of 536 completely sampled groups were composed of 
mixed age and sex (Wells et al ., 1987). Four bands of females were 
identified within the Sarasota community and occasionally mixed with 
each other (Scott et al., 1990). The most common group composition was of 
132 
dolphins of similar age, sex and reproductive status, resulting in the 
formation of female/calf groups, juvenile groups and all male groups 
(Wells et al., 1987). Highly related females (indicated by a sharing of a 
specific marker chromosome: Duffield and Wells, 1991) demonstrated high 
degrees of association, showing that familial relationships formed the basis 
of at least some female bands (Wells et al., 1987). However, these groups 
exhibited a wide range of associations and it was rare to observe all 
members of a band together at any given time (Wells et al., 1987). 
Male:male associations were highly specific, with some males forming 
consistent pairs or trios, others never interacted. Male:male bonds were 
formed between male calves of the same cohort and these bonds persisted 
into adulthood. In fact, the two mothers of one pair had a strong 
association and travelled in the same band, suggesting they may be related 
(Wells et al., 1987). These were similar to the male alliances described 
below for Australian Tursiops. 
The Tursiops mating system appeared to consist of males moving 
between groups of females. Males from adjacent communities were 
observed associating with the Sarasota females during the breeding season. 
Sarasota community males were absent from the range of Sarasota females 
for up to several months, during which time they probably came in .contact 
with dolphins from the adjacent communities. Genetic evidence for 
interchange between communities came from shared polymorphic enzyme 
profiles and mitochondrial DNA haplotypes between communities, the 
latter also indicated a low rate of female exchange (Duffield and Wells, 
1991). The patterns of differential distribution between the sexes within 
communities and occasional mixing between females and males from 
different communities, indicated that males were the most likely vectors for 
133 
the genetic interchange implied by the genetic heterozygosity (Wells et al., 
1987). 
Genetic evidence for a certain degree of reproductive isolation 
between communities came from observations of allelic differences in 
polymorphic blood enzymes and the discovery of a unique, bi-satellited 
chromosome limited to three generations of Sarasota dolphins (Duffield and 
Wells, 1991). This suggested that some mating was occurring within the 
community as well. Although male:female interactions were relatively 
uncommon, when males did interact with adult females, it tended to be 
with receptive females, who were females known to have given birth and 
their period of receptivity back-calculated utilising the mean gestation 
period of 12 months (Wells et al., 1987). No animals were ever documented 
to permanently change communities, further indicating the discreteness of 
the communities. The disappearance of animals which could not be 
accounted for through mortality indicated the maximum emigration rate 
out of the Sarasota community was less than 2-3% year-1 (Wells et al., 1987). 
Within the Sarasota community there were two different patterns of 
male associations with adult females: the resident male and the roving 
male patterns, distinguished by the relative duration of the association. 
The resident male pattern involved a single adult male repeatedly 
associating with a female or group of females within the range of the 
female activities, ·and only occasionally leaving to visit other female bands. 
Roving males, on the other hand, tended to be pairs, or occasionally trios, of 
males which moved between female groups and extended their movements 
beyond the range of the Sarasota community females (Wells et al., 1987). 
On a few occasions, these groups of roving males were observed to interact 
aggressively with adult males from adjacent communities resulting in 
bloodied scars (Wells, 1991). Only resident males were observed to form 
134 
I[ 
'I 
associations with receptive females, although this may not necessarily 
indicate mating success (Wells et al., 1987). Both single males and groups 
of males occasionally disappeared from the female range for weeks or 
months (Wells, 1991). In summary, the classification of males as members 
of the Sarasota community was complicated by their wider distribution and 
occasional disappearance for several months, but they still interacted with 
Sarasota females more than they did with females from other communities 
(Wells, 1986). 
Evidence that Tursiops males may attempt to control access to 
receptive females comes from research on Tursiops off western Australia 
(Connor and Smolker, 1985; Connor et al., 1992b; Connor et al., 1992a). 
Pairs and triplets of adult males with consistently high levels of association 
in a variety of behavioural contexts, termed an alliance (Connor et al., 
1992a), have been observed in cooperative herding of females. This 
behaviour was also observed on a few occasions in Florida (Wells et al, 
1987). The herding bouts off Australia involved chasing, displays (acrobatic 
leaps and underwater turns), acoustic emissions termed popping and 
aggression (hitting with the tail and biting or ramming) by the males 
towards the female (Connor et al., 1992). Often the female would try to 
escape by swimming away rapidly, but the males would work cooperatively 
to herd her again. The longest duration of association between a herded 
female and a . male alliance was 28 days (Connor et al., 1992a). 
Occasionally, two or more of these alliances would form second-order 
alliances and cooperate to take a female from anqther alliance (Connor et 
al., 1992a). Comparisons of the herding frequency for females based on 
their reproductive status indicated that pregnant (and thus not receptive) 
females were significantly less likely to be herded than were potentially-
receptive, n on-pregnant females (Connor et al., 1992a). The frequency of 
135 
11 
I[ 
partner changes within triadic alliances increased just prior to the mating 
season, when male Tursiops exhibit a surge in serum testosterone levels 
(Connor et al., 1992a). Males with erections were observed attempting to 
mount herded females, but it was not possible to observe intromission. 
Taken together, this evidence suggests that the male alliances cooperate to 
control access to receptive females. However, females appear to be able to 
mate with multiple males; some females were herded by up to 13 males 
during the season in which they conceived (Connor et al., 1992a). Coalition 
alliances of males have also been reported for lions (Bygott et al., 1979) and 
baboons (Smuts and Watanabe, 1990). 
In summary, Tursiops social structure appears to be composed of 
four "structural units": mother-calf pairs; subadult groups of both mixed 
and single sex; female bands with associated female-calf pairs and females 
with older offspring; and adult male groups. Although these units interact 
to varying degrees within a community, consistent patterns of segregation 
in the forms of habitat use and social association have been seen in over 20 
years of observation (Wells, 1991). Even for the relatively stable female 
bands, there is a high degree of fluidity in group interactions, with many 
observations of the group split into sub-units. Group formation occurred as 
calves matured and became independent from their mother, forming 
groups of subadult animals. Accurate ageing and surveys of reproductive 
status indicated these subadult groups contained adolescent animals who 
were sexually mature, but not yet socially mature (Wells, 1991). Maturing 
males gradually grew isolated from these groups .while maturing females 
began to join female bands; in a number of cases females rejoined their 
natal band (Wells, 1991). Upon reaching full adulthood, males associated 
more regularly with female bands, but were always found more frequently 
136 
in pair or trios with other males and ranged over a wider area than the 
female groups (Wells, 1991). 
A promiscuous mating system appears to be the best explanation for 
the data on bottlenose dolphins off Florida (Wells, 1986; Wells et al., 1987). 
This is based on the short-term interactions between males and females 
(indicated by the high degree of age/sex segregation within communities 
and the lack of permanent exchange of roving males between communities) 
and that males appear to interact with multiple female bands within their 
own community as well as with females from different communities. 
Additional evidence comes from the lack of sexual dimorphism (Hersh et 
al., 1990) and the potential for sperm competition (Parker, 1984): indicated 
by large testes size (Kenagy and Trombulak, 1986; Wells, et al., 1987) and a 
high concentration of sperm in the ejaculate (Schroeder, 1990; Schroeder 
and Keller, 1990). However, there is also evidence for a polygynous mating 
system because male alliances compete to maintain access to receptive 
females (Connor et al., 1992b; Connor et al., 1992a), males are heavily 
scarred, males from different communities interact aggressively, and 
resident males appear to have a higher degree of interaction with receptive 
females than do roving males (Scott et al., 1990; Wells, 1991). But the most 
convincing evidence for promiscuity was the preliminary paternity 
exclusion studies using DNA fingerprinting which have excluded Sarasota 
males in 13 of 14 cases as fathers of the Sarasota calves (Duffield and Wells, 
1991). 
5.3.6.2 Stenella 
The six species of the genus Stenella are all pelagic and are usually 
only sighted far from land (Perrin, 1975). However, a few situations have 
allowed regular observations on two Stenella species: Stenella longirostris, 
137 
i I 
the spinner dolphin, along the west coast of Hawaii (Norris and Dohl, 
1980a; Norris et al., 1985; Wiirsig et al., 1989; Wiirsig et al., 1991) and the 
spotted dolphin, Stenella attenuata, restrained in tuna purse seine nets in 
the eastern tropical Pacific (Pryor and Shallenberger, 1991). The picture 
which has emerged from these studies is of extremely fluid societies with 
frequently shifting patterns of association between individuals and groups. 
However, there are also indications of a high degree of behavioural 
coordination, suggesting repeated interactions between cooperating groups, 
which is usually characteristic of kin-related groups. Unfortunately, there 
have been no genetic studies of Stenella social groups to confirm or deny 
relatedness in these animals. 
Although spinner dolphins are usually found in deep water, the 
steep slope of the Hawaiian sea floor creates deep water close to shore and 
probably accounts for the species' nearshore occurrence. During the 
course of a two year study, 224 individual dolphins were identified and 
sighted from one time (98 dolphins) to 69 times, with an average of 6.4 
sightings per animal (Norris and Dohl, 1980b; Norris et al., 1985; Wiirsig et 
al., 1991). Thirty-six dolphins were seen 10 or more times throughout the 
study, indicating a certain degree of residency, but there appeared to be 
individual differences in distribution. Eight dolphins were seen primarily 
in the southern part of the study area, ten were seen primarily in the 
northern part of the study area and six were seen in all areas. However, 
even for these repeatedly sighted dolphins, only small numbers were 
regularly seen together. The conclusions were that spinner dolphins lived 
in continually changing groups except for a few core associations (Norris 
and Dohl, 1980b; Norris et al., 1985). Detailed underwater observations of 
groups identifed the presence of age and sex subgroups within schools. 
Adult males commonly travelled as a unit and often maintained position 
138 
between the observers and the rest of the school. Other subgroups noted 
were female - calf groups and apparent juvenile groups (Norris and Dohl, 
1980b; Norris et al., 1985). 
Spinner dolphins had a strict diurnal regime and were found in 
nearshore shallow waters (usually protected bays) during the day and 
moved offshore to deep, pelagic waters at night. This pattern of movement 
has been related to predation and food supply. The dolphins fed at night on 
prey associated with the deep, scattering layer which rises to the surface at 
night and drops down to approximately 500 m during the day (Norris and 
Dohl, 1980; Gabriel and Thomas, 1988). Daytime presence in protected 
coastal bays may also serve to avoid pelagic shark predation (Norris and 
Dohl, 1980; Norris et al., 1985; Wiirsig et al., 1991). Following rapid 
movement into the bays during the morning, the small groups of dolphins 
spent most of the day resting and socialising (Wiirsig et al., 1991). Group 
size remained relatively constant in the bays, suggesting an optimal 
number of resting dolphins for each bay. This, in conjunction with the 
observation that different individual dolphins occurred in a bay from day to 
day, further emphasises the fluid nature of spinner dolphin society. At 
dusk, the animals grew more active, eventually forming "rallying" groups 
which coalesced and moved offshore (Norris and Dohl, 1980; Norris et al., 
1985). Radio tracking of dolphins allowed them to be followed at night, 
when they were found to occur in large, widely dispersed groups , 
considered to be feeding assemblages of groups from each of the bays and 
along the coast. Since group composition remaine~ stable in nearshore 
groups during the day, it was deduced that changes in group composition 
took place as large groups broke up · and moved inshore at dawn. The 
changing membership of nearshore groups resulted in an eventual 
socialising between all or most members of the entire community of over 
139 
500 dolphins (Norris et al., 1985). It has been hypothesised that this high 
degree of group fluidity is an important means to maintain social bonds in 
a large number of animals which may need those bonds in order to 
cooperate effectively while feeding (Norris and Dohl, 1980b; Wursig, 1986; 
Norris and Schilt, 1988). 
Spotted dolphins, Stenella attenuata, are regularly caught by tuna 
purse seine fishing operations in the eastern tropical Pacific (Perrin, 1975), 
because tuna are known to accompany dolphin schools (Pryor and 
Schallenberger, 1991). As part of attempts to reduce dolphin mortality in 
these nets, an observational study of the behaviour of dolphins while 
restrained in the nets was conducted (Pryor and Schallenberger, 1991). 
Although occasionally both spinner and spotted dolphins were present in 
the nets, the two species always maintained separate groups. Underwater 
observations were collected on eleven sets of the net in which over 4000 
dolphins were encircled; focal animal observations were made on 97 spotted 
dolphins (Pryor and Schallenberger, 1991). Dolphins were aged on the basis 
of changes in colour pattern and adults were sexed by dimorphic features 
such as the post-anal keel. The behaviours of the various age and sex 
classes provided some insight into social relationships for this species. 
Although school size of spotted dolphins in the net sets varied from 23 
to 1000, subgroups were apparent and were composed of less than 20 
dolphins which maintained an inter-animal distance of less than 2 m and 
surfaced to breathe synchronously (Pryor and Schallenberger, 1991). 
Females with calves maintained close proximity to one another, often on 
the perimeter of the school, and often interacted with each other, unlike 
adult pairs which remained side-by-side. The female-calf pairs were 
occasionally observed in groups, associating "'rith other female-calf pairs. 
The composition of these groups changed, but the groups were always 
140 
11 
comprised of adult females and young immature animals. All-juvenile 
subgroups were also observed. These were composed of 3-6 animals and 
remained stable in composition, always travelling side by side. Adult male 
subgroups were the most conspicuous; composed of 3-8 dolphins which 
moved in unison. Pryor and Schallenberger (1991) state: "They cruised 
slowly through the school without swerving or altering speed, while a path 
opened up before them". This freedom of movement through the school and 
avoidance by other school members suggested that these animals were 
dominant (Pryor and Schallenberger, 1991). Social aggression was 
common within these schools (usually indicated by gaping and head 
nodding) and occurred within adult male subgroups, although apparently 
less frequently than within juvenile subgroups and between mothers and 
their calves. On two occasions, aggressive interactions were observed 
between two male subgroups. The large degree of synchrony and affiliative 
behaviour within spotted dolphin schools suggests they are familiar with 
each other and represent a relatively stable unit (Pryor and Schallenberger, 
1991). The presence of discrete colour patterns within schools (Perrin, 
1969), suggests some degree of relatedness between these animals. Other 
observations in the tropical Pacific suggests a certain degree of age 
segregation, with some schools composed entirely of juveniles and others 
(such as those reported on here) composed of reproductive animals and 
young calves (Hohn et al., 1985; Myrick et al., 1986). The resulting picture 
of spotted dolphin: social organisation is one of age and sex segregation 
within and between schools, with dispersal of the juvenile segment of the 
population. 
5.3.6.3 Lagenorhynchus 
Lagenorhynchus is a multi-species genus (see Table 5.1), but only L. 
obscurus, the dusky dolphin, has been studied (Wiirsig and Wursig, 1980; 
141 
1 I I 
I I 
I 
Wiirsig et al., 1989; Wiirsig et al., 1991). While these studies have primarily 
been concerned with patterns of habitat use, they have also provided insight 
into Lagenorhynchus social structure. 
Dusky dolphins were studied in two locations: off Golfo San Jose, 
Argentina in the south Atlantic (Wiirsig and Wiirsig, 1980; Wiirsig et al., 
1989; Wiirsig et al., 1991) and off Kaikoura, New Zealand in the south 
Pacific (Wiirsig et al ., 1989; Wiirsig et al., 1991). In Argentina, the dolphins 
were primarily found in depths less than 100 m, foraging for schools of 
southern anchovy (Engraulis anchoita). The majority of feeding occurred 
during the day, when anchovy tend to school, and dolphins generally rested 
close to shore at night, probably to avoid killer whale predation. During the 
day, stable group units of usually less than ten dolphins were observed 
searching the bay, with up to 30 of these groups spread from 1-8 km apart. 
Once food was located, other groups apparently located prey through 
acoustic and visual cues such as associated flocks of feeding birds, and the 
several groups converged to herd and capture prey cooperatively. After 
feeding, the dolphins remained in large schools and engaged in a high 
activity level of socialising. Although photo-identification studies have been 
limited, the fission and fusion of dusky dolphin groups appears to result in 
continual changes in group membership. Similar to Hawaiian spinner 
dolphins, the widespread social interactions provided by a fluid social 
structure probably enhance the bonding mechanisms required for the 
maintenance of a large number of cooperating school members (Wiirsig 
and Wiirsig, 1980; Wiirsig, 1986; Wiirsig et al., 1989;, Wursig et al., 1991). 
Off New Zealand, dusky dolphins live in a very different habitat and 
feed over deep water on prey associated with the deep scattering layer 
(Wiirsig et al., 1989; Wiirsig et al., 1991). Group sizes are consistently large 
and do not exhibit the patterns of fission and fusion found in the Argentine 
142 
I , I 
dolphins. However, there are still distinct subgroups within these large 
schools. It is unknown whether the composition of these subgroups 
remains stable. The lack of clear patterns of fission and fusion may be 
related to the fish prey of New Zealand dusky dolphins, which is not a 
schooling species, and cooperative herding on the part of the dolphins does 
not seem to be required for its capture. This exemplifies that ecological 
parameters affect aspects of dolphin society (Wursig et al., 1989; Wursig et 
al., 1991). 
5.3.6.4 Grampus 
The taxonomic status of Grampus griseus, Risso's dolphin, has long 
been debated (Nishiwaki, 1963; Nishiwaki, 1964; Fraser, 1966; Mead, 1975). 
It is the only delphinid lacking teeth in the upper jaw and some authors 
have proposed placing it in a mono-specific family Grampidae 
(Nishiwaki, 1972). Its ' external appearance resembles the globicephalines, 
with a blunt nose and prominent melon (Mead, 1975). It has occasionally 
been placed in the Globicephalinae. However, in terms of skull morphology 
and anatomy of the air sacs (Fraser and Purves, 1960), it appears to be most 
closely related to Tursiops. This is the current basis for placing Grampus 
in the Delphininae (Klinowska, 1991), with the similarities in facial 
anatomy between Grampus and globicephalines being seen as convergent 
evolution, as it is with monodontid whales (Fraser, 1966; Mead, 1975). 
There has been only one complete study of Grampus using photo-
identification of individuals (Kruse, 1989). This two-year study was 
conducted along the central west coast of California from October 1985 to 
November 1987. Fifty-nine schools were located and group size ranged from 
3 to over 500 (mean± S.D.: 63 ± 87.9; median: 30). A total of 800 individuals 
were identified and it was estimated that only two-thirds of the animals 
143 
observed in the field were identifiable. The number of sightings per 
individual ranged from one (588 dolphins) or two (161 dolphins) up to eight 
times (two dolphins). The rate of identification of new animals continued to 
increase throughout the study, indicating that only a small percentage of 
the Grampus visiting the region were identified (Kruse, 1989). There was 
some seasonal variation in Grampus occurrence, with more animals seen 
during the fall (September-November) than during all other seasons. 
Groups of fewer than 30 dolphins travelled and behaved in unison, but 
larger groups exhibited a distinct sub-group structure. Age and sex 
segregated subgroups were noted in groups of over 60 dolphins and were 
composed of females with calves (nursery groups), juveniles or large, 
calfless animals. Some associations within the large, calfless animal 
subgroups were observed repeatedly over 10 months. However, beyond 
these stable subgroups, fluctuations in the group size indicated that groups 
were "dynamic aggregations" (Kruse, 1989). Although the age and sex of 
the large, calfless individuals was not known, evidence from strandings 
has shown similar groups to be composed of animals of the same age and 
sex. Thus, the limited observations suggest that Grampus occurs in stable, 
age and sex segregated groups which interact fluidly with a larger 
population (Kruse, 1989). 
5.3.6.5 Summary of Delphininae Social Systems 
These reviews of social organisation in four genera of the sub-family 
Delphininae have shown a variety of broad similarities which I suggest are 
characteristic of the group. One characteristic is the clearly defined age 
and sex segregation of Tursiops and Stenella, exemplified by the dispersal 
of adolescent age classes from the breeding groups (Welis et al., 1987; 
Norris et al., 1985). Age and sex segregation was also suggested in 
observations of large Grampus schools (Kruse, 1989), but there is still not 
144 
enough information on the age and sex of identified populations of 
Lagenorhynchus to know the degree of segregation in these animals 
(Wi.irsig and Wi.irsig, 1980). The only available information on sex ratio 
was from the observational studies of Tursiops, where there were 2.4 adult 
females for every 1 adult male (Wells et al., 1987). This was presumed to be 
due to the more rapid maturation in females and higher male mortality. 
Group fluidity also takes different forms in these dolphins. In Tursiops, 
bands of females have only statistical tendencies to associate and they are 
not exclusively observed together as an indivisible unit (Scott et al., 1990; 
Wells, 1991). Males exhibit the greatest fluidity, moving beyond the 
community ranges of the females (Wells et al., 1987). It may be that age 
and sex segregation enhances this fluidity by the isolation of adolescents 
from their natal groups, thus increasing their exposure to other segments 
of the community. In Argentine dusky dolphins, group fluidity occurs 
between stable units which interact in large after-feeding assemblages 
(Wi.irsig and Wi.irsig, 1980; Wi.irsig et al., 1989, 1991) and this was similar to 
Grampus observations of fluid interactions between stable subgroups 
(Kruse, 1989). In Hawaiian spinner dolphins fluidity appears to occur at 
the individual level, and there appear to be no stable sub-units (Norris et al., 
1985; Wi.irsig et al., 1991). 
5.3. 7 Globwephalinae Social Syst;ems 
The five genera in the subfamily Globicephalinae have significantly 
different trends in sexual dimorphism than other delphinids. This 
suggests that it will be instructive to consider the social systems of these 
genera separately from those of the Delphininae reviewed above. 
Unfortunately, detailed studies in social organisation are only available for 
145 
two of these genera: Orcinus, the killer whale, and Globicephala, the pilot 
whales. 
5.3.7.1 Orcinus 
The killer whale (Orcinus orca) social system which has been studied 
to date is characterised by long-term stability of associations between all age 
and sex classes within populations and a high degree of variation in social 
structure between reproductively (and sometimes, though not always, 
geographically) isolated populations. This provides a clear indication of the 
inbred nature of killer whale communities, giving the species a strong 
potential to form populations with distinct social and behavioural 
characteristics (Hoelzel and Dover, 1990; Hoelzel, 1991a). 
Killer whale social organisation has primarily been studied in the 
inland marine waters off Washington State, USA and British Columbia, 
Canada since 1973 (Balcomb et al., 1982; Bigg, 1982; Ford and Fisher, 1983; 
Ford, 1984; Hoyt, 1984; Balcomb and Bigg, 1986; Haenel, 1986; J.Heimlich-
Boran, 1986b; J.Heimlich-Boran, 1986a; S.Heimlich-Boran, 1986; Hoelzel 
and Osborne, 1986; Osborne, 1986; Bigg et al., 1987; Baird and Stacey, 1988; 
J.Heimlich-Boran, 1988; S.Heimlich-Boran, 1988; Ford, 1989; Bigg et al., 
1990; Hoelzel and Dover, 1990; Morton, 1990; Olesiuk et al., 1990; Felleman 
et al ., 1991; Baird, 1992). Individuals are identified by distinctive 
pigmentation patterns just posterior to the dorsal fin, termed saddle 
patches, and photographic surveys have been conducted every year since 
1973 (Bigg et al., 1987; Bigg et al., 1990). 
Two forms of killer whales have been defined in this region: residents 
and transients (Bigg et al. 1987; Bigg et al. 1990). These two forms differ in 
a wide variety of behavioural aspects of distribution, seasonal occurrence, 
acoustic dialects, and prey choice (Ford and Fisher, 1983; Ford, 1984; 
146 
I 
I' 
Heimlich-Boran, 1988; Ford, 1989; Felleman et al., 1991; Baird, 1992) 
Morton, 1990; J.Heimlich-Boran, 1986), as well as genetic aspects of 
pigmentation patterns, dorsal fin morphology and mitochondrial DNA, 
which indicate they probably represent distinct races of killer whales 
(Duffield, 1986; Baird and Stacey, 1988; Stevens et al., 1989; Hoelzel and 
Dover, 1990; Hoelzel, 1991a). The two forms have probably evolved from 
distinct foraging specialisations, forming an Evolutionary Stable State 
(ESSt) due to disruptive selection against generalists (Baird et al., 1992). 
Genetic evidence from the analysis of mtDNA have identified significant 
differences in restriction fragment patterns between residents and 
transients, which indicates that the two forms have distinct maternal 
lineages (Stevens et al., 1989; Hoelzel and Dover, 1991; Hoelzel, 1991a). DNA 
fingerprinting of nuclear genomes found that band-sharing coefficients 
were 2-3 times lower between populations than they were within 
populations and supports the hypothesis of significant reproductive 
isolation (Hoelzel and Dover, 1990). In fact, the two forms appear to be as 
genetically distinct as North Pacific populations were from a south Atlantic 
population, probably resulting from a high degree of inbreeding within 
small regional populations (Hoelzel and Dover, 1990). Assuming a 
standard divergence of 5 x 10-9 base pairs/year, this degree of difference 
implies a two million year period since the two forms shared a common 
ancestor (Hoelzel, 1991a). This degree of reproductive isolation of sympatric 
populations is rare, especially when the primary isolating mechanisms 
appear to be behavioural (Hoelzel and Dover, 1990). 
Resident killer whales live in two reproductively isolated 
communities (Bigg et al., 1990). The communities have core areas and the 
boundary between communities corresponds to an area of limited tidal 
mixing, which could also limit salmon migrations (Felleman et al ., 1991). 
147 
There have been no reports of permanent interchange of individuals 
between these communities (Bigg et al., 1990), and the two communities 
have distinct vocal dialects (Ford and Fisher, 1983; Ford, 1989). Genetic 
studies comparing mtDNA D-loop base pair sequences between whales 
from the two resident communities identified a difference of two base pairs 
out of a total of 591 (0.2%: Hoelzel, 1991a). Although sample sizes were two 
small to compare the significance of this variation, the most likely 
explanation involved a colonisation of the separate community home 
ranges by more than one matriline following the retreat of the ice shelf 
40,000 years ago (Hoelzel and Dover, 1990). This supports the hypothesis of 
the current reproductive isolation of the two resident communities (Hoelzel, 
1991a). 
The primary social unit of resident whales is the pod (Bigg et al., 
1990). A pod was defined as "the largest cohesive group of individuals 
within a community that travelled together ... for at least 50% of the time" 
(Bigg et al., 1990: p.388). Pod membership was determined over a number 
of years and has been based on a cluster analysis of surfacing associations 
(Bigg et al., 1990). Pod membership did not change in the 20 years of 
observations and some pod members were observed in photographs taken 
during capture operations in the 1960's. There were a number of levels to 
pod structure. Some pods occasionally (between 5% and 50% of the time) 
split into smaller units termed subpods , which were usually never 
separated from each other for more than one month (Bigg et al., 1990). Pods 
were composed of an average of 1.7 subpods (~ange: 1-3). Although 
members of the same subpod were found to travel together 95% of the time, 
distinct association patterns within the subpods have been demonstrated 
CS .Heimlich-Boran, 1986; 1988). These groupings are t ermed intra -pod 
groups (Bigg et al ., 1990). Subpods contained an average of 1.9 int r a -pod 
148 
I I 
groups (range: 1-11) and intra-pod groups contained an average of 3.6 
whales (range: 2-9). 
Pods were found to consist of overlapping generations of females and 
their offspring. Genealogies were defined with three levels of certainty: 
positive, highly probable and probable. Positive genealogies were defined as 
those between adults and calves born during the course of the study. Highly 
probable genealogies were determined for animals which were immature 
at the start of the study and mothers were deduced from a ranking of 
associations coefficients with all adult females. Probable genealogies were 
those determined for animals that were adult at the start of the study. The 
determination of probable genealogies was based on consistent association 
patterns, scaled relative to the high degree of association between known 
mother-offspring pairs. 
There appeared to be a class of females who were post-reproductive, 
similar to the situation for short-finned pilot whales (Marsh and Kasuya, 
1984). Some females were never observed to give birth during the course of 
the study and were considered to be the oldest females. Other females were 
considered to have recently become post-reproductive if they gave birth to 
calves at the beginning of the study and then had no more calves for at least 
the last ten years of the study (Olesiuk et al., 1990). It was possible that 
some of the post-reproductive females may have been infertile, young 
females, but the -documented occurrence of such females was rare (all but 
one of the females who matured during the study gave birth). 
Associations continued between mothers and known male and 
female offspring into adulthood (Bigg et al., 1990). Positive mother-
offspring relationships have been determined for 133 offspring which 
ranged in age from 0.5 to 14.5 years (mean age= 5.7 years) at the end of the 
149 
study. Highly probable mother-offspring relationships were determined for 
80 juveniles, which were aged 13.5 to 34.5 years (mean = 19. 7 years) by the 
end of the study (Bigg et al., 1990). Probable mothers were determined for 34 
of the 102 animals that were adults at the start of the study. Changes in 
associations between mothers and their offspring were found to vary with 
the age and sex of the offspring (Bigg et al., 1990). Young female calves had 
slightly stronger associations with their mother than did male calves, 
indicating a greater degree of independence for male calves. However, by 
the age of 10, male calves began to have higher average associations with 
their mothers than female calves. As young females matured and gave 
birth to their own calves, their association with their mothers continued to 
drop until the young female's early 20's, when the association index 
levelled out at around 25%. In contrast, associations between mothers and 
their sons levelled out at around 10 years of age at an association index of 
around 40%. Thus, adult sons maintained stronger association with their 
mothers than did females, probably because the young females spent more 
time with their own calves (Bigg et al., 1990). Another side of this strong 
male-mother relationship was shown by the fact that the presence of adult 
males made intra-pod groups more independent. This was indicated by a 
weakening of the bond between a female and the other members of her pod 
when her son matured. As females reached post-reproductive age, they 
often had strong associations with actively reproducing females, who were 
probably their daughters of younger sisters (S.Heimlich-Boran, 1986; 
S.Heimlich-Boran, 1988). Often, these older females also spent time with 
immature animals, suggesting a form of allo-parental care (S.Heimlich-
Boran, 1986; Haenel, 1986). 
The picture of killer whale social organisation which has emerged 
from these genealogies is one of matrilineally based relationships forming 
150 
the structure of intra-pod groupings (S.Heimlich-Boran, 1986; 1988; Bigg et 
al., 1990). The generalised matrilineal group was composed of a 
grandmother, her adult son and adult daughter, and the offspring of the 
daughter. Intra-pod groups were composed of an average of 2.3 
generations (range: 1-4). Adult sons tended to travel closest to their 
mothers, while daughters were slightly farther away with their own 
offspring. Adult males with no living mother continued to travel with their 
adult sister's group, but often travelled on the periphery. In some cases, 
the association between two sisters decreased after their mother died. Some 
matrilineal groups were linked by a common mother, with one of the 
daughters being in a separate intra-pod group. The fission between two 
sisters such as this was likely to be the basis of new subpod formation. 
There were no observations of the dispersion of entire matrilineal groups 
into new pods. Thus, matrilineal groups within a subpod were likely to be 
closely related. The time frame in which new pods form is likely to be 
similar, but to occur over a much longer period (Bigg et al., 1990). If fission 
of intra-pod groups took one generation (2-3 decades), then the formation of 
new pods could take many decades, if not centuries. Associations between 
pods were assumed to be based on shared genealogies, but in a number of 
cases the patterns of travelling associations did not correspond to the 
acoustic clans documented by Ford (1989). Since acoustic dialects are more 
likely to reflect shared ancestry accurately than travelling associations, it 
appeared that associations between pods were not based on relatedness. 
This would support the hypothesis that mating occurs between pods, which 
would likely associate with unrelated pods in order to reduce inbreeding. A 
system of travelling in kinship groups which form the basis of higher levels 
of social organisation has been described for a wide variety of terrestrial 
mammals, including primates (Kummer, 1968; Kurland, 1977; Gouzoules 
and Gouzoules, 1987; Stammbach, 1987), lions (Bertram, 1976), elephants 
151 
(Douglas-Hamilton and Douglas-Hamilton, 1975; Dublin, 1983; Moss and 
Poole, 1983) and ground squirrels (Michener, 1982; Sherman and Morton, 
1984; Sherman and Holmes, 1985). Bigg and co-workers conclude that the 
killer whale social system, with its lack of dispersal of either sex, could 
result "from a particularly strong requirement that reliable and familiar 
associates be available for hunting or maintaining territorial 
boundaries, ... or from a unique breeding strategy" (Bigg et al., 1990: p . 397). 
The main element missing from the detailed long-term observations 
outlined above has been information on the reproductive success of males. 
I have summarised all observations of socio-sexual behaviour (defined by 
the observation of an erect penis) in the southern community of resident 
Orcinus from six years of observations (Heimlich-Boran and Heimlich-
Boran, ms.). Although the sample size was small, I believe the data have 
valid implications for an understanding of the mating system of Orcinus. 
Sexual behaviour was observed between adult whales on 30 separate bouts. 
Eighteen of the 30 bouts (60%) involved males and females together, while 
the remaining 12 bouts ( 40%) were of male only groups. Although not 
directly indicative of mating patterns, the male-only bouts will be discussed 
because they serve to clarify male-male relationships between pods. The 
male only groups were almost always composed of males from more than 
one pod (11 of 12 bouts). The single observation of males from the same pod 
involved two adult and one adolescent male (17 yrs old) . All males were 
identified in 8 of the 11 multi-pod bouts; adult males were present in all 
cases, and at least one adolescent male (8-15 yrs ol~) was present in half of 
the cases. These multi-pod observations involved males travelling together 
in tigh t subgroups for up to 3.5 hrs. Sometimes the male groups wer e 
separate from all other whales, but on two occasions, females were 
observed hovering nearby and oriented towards the males. Males were 
152 
usually in tight body contact with each other, actively pushing one another. 
Usually just one male would surface with an erection, but on one occasion, 
two males did so simultaneously. There were never any observations of 
overt aggressive interactions between the males in these groups, although 
most males had old scars. 
The sexual interactions between males and females were most 
relevant to an understanding of the Orcinus mating system. Fourteen of 
the 18 bouts involved members from different pods. Only one male was 
observed in all but one case, when three were identified (but only one 
surfaced with an erection. Three adolescent males (ages 10 - 15 yrs) were 
involved in seven (44%) of the observations and three adult males were 
involved in the remaining nine cases. Thus, only six of the 16 adolescent 
and adult males in the three resident pods were ever observed in 
heterosexual encounters. The number of females involved in these 
heterosexual encounters varied from one (five cases) to two or three (seven 
cases). Twelve females were identified in 11 cases; six of the females were 
post-reproductive adult females (defined by Bigg et al., 1990) and the other 
six were females with calves. The behaviour during these male-female 
sexual encounters usually consisted of the male swimming with the female 
or group of females for a period of time and then rolling sideways with an 
erection. On three of the cases, the female rolled belly-up, apparently 
attempting to av~id intromission. Two of the three avoiding females were 
previous mothers. On one occasion, the male actively pursued the female 
while she tried to swim away. Once, the mqle swam upside down 
underneath the female . Other whales occasionally showed some interest in 
these groups . In two cases, immature animals remained nearby, probably 
associates of the female. On three occasions, other adult males were 
nearby, always from the same pod as the female and once, the male was the 
153 
female's probable son (Bigg et al., 1990). During one case, adult male J3 
charged adult male L19, who had just surfaced with an erection around 
two J pod females. The males both dove when approximately 3 m apart, and 
nothing further was observed. 
There were no documented births which corresponded to these 
mating attempts and since the majority of females involved were post-
reproductive, the matings did not appear to be for procreation. Little is 
known of the patterning of sexual activity in male delphinids (Ridgway and 
Green, 1967; Perrin and Reilly, 1984), but in general male killer whales 
show a broadly diffuse seasonal pattern to sperm production (Christensen, 
1984), suggesting they would have been reproductively active. The seasonal 
pattern of the observations indicates that the peak of the interactions 
between males and females coincided with the proposed period of peak 
conception determined by the seasonality of births (Bigg et al., 1991) as well 
as corresponding to the salmon season when pods travel in large, multi-pod 
assemblages (J.Heimlich-Boran, 1986a; 1988). In any event, copulation 
frequency is known to be a poor indicator of paternity (Stern and Smith, 
1984). The value of these observations lies in the indication that sexual 
interactions occur less frequently with close kin (i.e. members from the 
same pod) than they do with other whales. 
The resulting picture of killer whale social organisation is of a multi-
level society with a ·high degree of stability on all levels. The long-term 
consistencies of intra-pod groups, subpods, pods, clans and communities 
all contribute to this stability. This has been verified in the genetic analyses 
by the relatively high degrees of inbreeding (Hoelzel, 1991a). The lack of 
male dispersal suggests that males must benefit from group membership. 
154 
,i 
I 
i ' 
5.3. 7.2 Gwbwephala 
Information and original data have been presented for the two 
Globicephala species in Chapters 2-4. In summary, although the two 
species differ in some morphological and life history traits (Kasuya et al., 
1988), there appear to be general similarities in social structure, indicated 
by similar age and sex composition (Kasuya and Marsh, 1984) and shared 
indications of genetic segregation between adjacent groups (Andersen, 
1988; Wada, 1988). Globicephala groups are highly cohesive, making them 
susceptible to drive whaling (Gibson-Lonsdale, 1990) and mass stranding 
(Sergeant, 1982; Klinowska, 1986). Genetic information of G .melas has 
shown all group members to be related, including adult males and adult 
females within the same pod (Amos et al., 1991a; 1991b). In all studies, 
pods have been primarily composed of mixed age and sex. The few rare 
observations of all male groups indicate only limited segregation of the 
sexes (Sergeant, 1962a; Kasuya and Marsh, 1984; Chapter 4). There was a 
high degree of differential mortality between the sexes: female longevity of 
G.macrorhynchus was over 20 years longer than male longevity, resulting 
in a female-biased adult sex ratio (Kasuya and Marsh, 1984). Females 
ceased ovulating by the age of 40 and lived for an average of 23 years in an 
extended post-reproductive period (Kasuya and Marsh, 1984; Marsh and 
Kasuya, 1984; 1986; 1991). Some of these females continued to nurse their 
last calf for up to 15 years (Kasuya and Marsh, 1984; 1989). 
Genetic evidence relating to the mating system came from paternity 
studies of G.melas which found that males were not the fathers of the 
offspring in the pods (Amos, 1991a; 1991b). It was also found that only one 
or a few related males were fathering the young, indicating a certain level 
of male variance in reproductive success and suggesting polygyny. 
However, it was also found that the male genetic contribution to cohorts 
155 
1~ 
,I 
I I 
I , 
I 
changed from year to year (Amos, 1991a; 1991b), indicating a certain degree 
of promiscuity. Wilson (1975) has suggested that even promiscuous 
:matings need not be random. Analyses of associations for 
a.macrorhynchus indicated that male:female interactions were most 
com:rnon when linked pods (separate pods with regular patterns of 
association) travelled together. Although mating was never observed, these 
associations could indicate a possible time when out-group breeding could 
occur. The season when members of more than one pod were observed 
together corresponded to the supposed peak season of conception. 
5.3. 7.3 Summary of Gwbicephalinae Social Syst.ems 
There are a number of similarities between the social systems of 
Orcinus and Globicephala which suggest they may be common 
globicephaline traits and which are in contrast to the open nature of 
delphinine communities described above. A comparison of life history 
traits and pod sizes for Orcinus and G. macrorhynchus is presented in 
Table 5.2. Killer whales have a slightly longer longevity for both sexes, but 
the maturation period for male pilot whales is essentially the same as for 
male killer whales. This is because the maturation periods are equivalent 
for male and female killer whales, while female pilot whales mature much 
sooner than pilot whale males. Differences in reproductive parameters are 
indicated by the longer calving interval in pilot whales. Pilot whale calving 
intervals increased from three years for primiparous females to 10 years for 
females older than 30 (Kasuya and Marsh, 1984). Killer whale calving 
intervals ranged from two to 12 years, but the regression of calving interval 
on age of the female had a low rate of increase (slope = 0.086) and only 
accounted for 7. 7% of the observed variation in calving interval (Olesiuk et 
al ., 1990). Female in both species had a long post-reproductive period which 
represent ed a similar proportion of the total lifespan (about 37%). 
156 
1111 
I 
Table 5.2. Comparison of life history parameters for Orcinus orca and 
Globicephala macrorhynchus. Orcinus data were compiled from a 
horizontal study of life history using observational data and Globicephala 
data were from a vertical analysis of a complete sampling of carcasses. 
Parameter O.orca 1,3 G.macrorhynchus 2 
Age at sexual maturity: male 15.0 yrs f 15.8 yrs h 
Age at physical maturity: male 21.0 yrs g 25.0 yrs i 
Age at sexual maturity: female 11.7 d - 14.9 yrsj 9.0 yrs e 
Reproductive lifespan 25.5 yrs 24.0 yrs 
Lifetime calf production 5.47 calves 4.39 calves 
Mean calving interval 5.32 yrs m 6.92 yrs m 
Age at post-reproductive 39.1 yrs k 39.5 yrs 1 
Mortality rate: male (age) 0.039 (15.5-30.5+) 0.0393 (9-30) 
Mortality rate: female (age) 0.0114 (15.5-65.5+) 0.0251 (18-4 7) 
Longevity: male 36.5 a,c - 60 yrs b 46 yrs a 
Longevity: female 77.5 a - 90 yrs b 63 yrs a 
Life expectancy at birth: male 16.6 yrs b 12.11 yrs b 
Life expectancy at birth: female 28.9 yrs b 22.26 yrs b 
Adult sex ratio (all females) 1 male:1.7 females b 1 male:3.7 females b 
Adult sex ratio (reprod. only) 1 male:1.15 females b 1 male:1.95 females b 
Mean pod size 15.7 12.2 ± 1.3 
Calving peak 1 January 25July 
References: l)Olesiuk et al ., 1990); 2) Kasuya and Marsh, 1984; 3) Bigg et al., 1990; 4) 
Present study; ~: a) oldest aged animal; b) life table estimation; c) males that were 
adult when observations began had only minimum age estimates based on assumption of 
attainment of physical maturity in year prior to study; d) age at first conception; e) age at 
first ovulation; D based on 50% males beginning allometric growth of the dorsal fin; g) 
based on 50% males completing allometric growth of the dorsal fin; h) based on 50% 
animals with mature testes; i) based on 50% males completing testes weight growth; j) age 
at first birth; (k) mean age at birth of last calf; 1) mean age at last ovulation; m) excluding 
post-reproductive females 
I, 
11 
While male dispersal rates are not well known in Globicephala, the 
infrequent occurrence of all male groups and the high degrees of 
relatedness of males within groups, suggests that there is a similar pattern 
of natal philopatry to Orcinus. While there have been no paternity studies 
of Orcinus, behavioural observations suggest the males are not the fathers 
of the young in their pod. Rather, the Orcinus males appeared to be the 
brothers or uncles of the young in their pod. 
5.4 DISCUSSION 
Mammalian mating systems are based on individual male and 
female strategies of reproductive success (Darwin, 1871; Trivers, 1972). The 
strategies of the two sexes may be very different. Males, in their attempt to 
maximise the number of offspring sired (Bateman, 1948), appear to adapt 
their strategies to the distribution of receptive females. Where possible, 
males do all they can to control access to their mates (mate guarding), often 
through competition with other males, in order to guarantee their 
successful production of offspring (Emlen and Oring, 1977). Females, 
because of their relatively larger contribution to the production of offspring 
(in terms of egg production, gestation and lactation), are best served by 
maximising their ability to provide adequate food and protection for the 
successful survival of offspring to reproductive maturity (Trivers, 1972; 
Wrangham, 1980; ·Gaulin and Sailer, 1985; Clutton-Brock, 1986). Since 
female distribution is based on ecological factors to a much greater degree 
than that of males (Wrangham, 1980; Gaulin, 1985); the order of causation 
in determining a mating system appears to be that resources determine 
female dispersion, which in turn, determines male dispersion (Davies, 
1991). Interwoven with this basic framework are a myriad of other factors 
such as the distribution of predators and resource competitors , male 
157 
'' 
111 
parental care and phylogenetic constraints on life history and reproduction. 
Variation in the relative importance of these factors demonstrates 
predictive value in the definition of mating systems (Clutton-Brock, 1989; 
Davies, 1991). 
Four generalised types of mating system have been described, based 
on the number of mates for any given breeding season and the duration of 
the mating bond: 1) monogamy, where one male and one female maintain 
an exclusive mating bond, 2) polygyny, where one male mates with 
multiple females, maintains a prolonged mating bond with that group of 
females and excludes other males from access, 3) polyandry, where one 
female mates with multiple males, and 4) promiscuity, where males mate 
with any receptive females and there is no prolonged mating bond (Orians, 
1969; Wilson, 1975; Emlen and Oring, 1977; Clutton-Brock, 1989)(Stirling, 
1983). Monogamy, although common in birds, occurs in only about 5% of 
mammalian species (Kleiman, 1977): primarily in canids (Moehlman, 
1986) but also in primates (Rutberg, 1983). Monogamy appears to be 
favoured where male paternal care is required in some way for the survival 
of offspring (Clutton-Brock, 1989). Monogamy has also been viewed as 
means by which the male guarantees paternity (Trivers, 1972). Polyandry 
is also rare in mammals, which Orians (1969) predicted on the basis of the 
relatively greater investment females have in the production of young. 
Polyandry is most common in birds where a female can lay her eggs and 
then desert the male while he is brooding the eggs (Lack, 1968; Trivers, 
1972). The female is thus able to lay multiple clutc~es with different males 
through the course of one breeding season. The remaining two mating 
systems: polygyny and promiscuity, are the two most common forms of 
polygamy in mammals. · These systems are primarily differentiated by the 
duration of the bond between males and females and by the relative number 
158 
111 
11 
of multiple matings that each sex obtains. Unfortunately, some discussions 
of mating systems incorporate promiscuity as a form of polygyny, using a 
classification system based on gamete contributions to zygotes (Wiley, 197 4; 
Ralls, 1977). Since these systems are characterised by social groups of 
multiple males and females, they will be the most likely candidates for 
delphinid mating systems and will be the focus of the rest of the discussion. 
I will now address the elements of sexual dimorphism, brain size, 
testes size, and the form and frequency of social relationships predicted for 
these mating systems, and examine the data on delphinid morphology and 
social structure to postulate hypotheses for specific delphinid mating 
systems. I primarily consider Tursiops (Delphininae), and Orcinus and 
Globicephala (Globicephalinae). These were the only genera which had 
sufficient comparable observations on social systems . I have already 
argued that there are similarities between Tursiops and the other 
delphinine genera based on the limited data available, but I cannot make 
any further conclusions. Thus, although I may refer to entire subfamilies 
in the subsequent discussion, I can really only make conclusions about the 
three well-studied genera. 
5.4.1 Patterns of Body Size and Sexual Dimorphism 
The central tenet of the relationship between sexual dimorphism and 
mating system is · that sexual dimorphism is indicative of male-male 
competition for access to females resulting in polygyny (Bartholomew, 1970; 
Crook, 1972; Clutton-Brock et al., 1977; Ralls, 1977; Alexander et al., 1979; 
Stirling, 1983; Clutton-Brock, 1989). There were clear trends in the 
occurrence of delphinid sexual dimorphism which were correlated with 
body size, ranging from the smallest genus (Cephalorhynchus), in which 
females were larger than males, to one of the largest genera (Globicephala) 
159 
I I 
of multiple matings that each sex obtains. Unfortunately, some discussions 
of mating systems incorporate promiscuity as a form of polygyny, using a 
classification system based on gamete contributions to zygotes (Wiley, 197 4; 
Ralls, 1977). Since these systems are characterised by social groups of 
multiple males and females, they will be ·the most likely candidates for 
delphinid mating systems and will be the focus of the rest of the discussion. 
I will now address the elements of sexual dimorphism, brain size, 
testes size, and the form and frequency of social relationships predicted for 
these mating systems, and examine the data on delphinid morphology and 
social structure to postulate hypotheses for specific delphinid mating 
systems. I primarily consider Tursiops (Delphininae), and Orcinus and 
Globicephala (Globicephalinae). These were the only genera which had 
sufficient comparable observations on social systems. I have already 
argued that there are similarities between Tursiops and the other 
delphinine genera based on the limited data available, but I cannot make 
any further conclusions. Thus, although I may refer to entire subfamilies 
in the subsequent discussion, I can really only make conclusions about the 
three well-studied genera. 
5.4.1 Patterns of Body Size and Sexual Dimorphism 
The central tenet of the relationship between sexual dimorphism and 
mating system is that sexual dimorphism is indicative of male-male 
competition for access to females resulting in polygyny (Bartholomew, 1970; 
Crook, 1972; Clutton-Brock et al., 1977; Ralls, 1977; Alexander et al., 1979; 
Stirling, 1983; Clutton-Brock, 1989). There were clear trends in the 
occurrence of delphinid sexual dimorphism which were correlated with 
body size, ranging from the smallest genus (Cephalorhynchus), in which 
females were larger than males, to one of the largest genera (Glob icephala) 
159 
in which males were 1.2 times longer and 1. 7 times heavier than females. 
How does this compare to reported degrees of sexual dimorphism in other 
mammals? Ralls (1977) defined extreme sexual dimorphism as a ratio of 
male weight to female weight greater than 1.6 and documented this in 8 of 
20 mammalian orders. The maximum reported weight dimorphism in 
primates was around 2.3 for the baboon, Papio hamadryas (Clutton-Brock et 
al., 1977: Fig. 1). Thus, the degree of sexual dimorphism in delphinid 
cetaceans is not extreme. 
There were significant differences in both weight and length sexual 
dimorphism between the two delphinid subfamilies of Globicephalinae and 
Delphininae (Section 5.3.2). The difference in dimorphism between the 
subfamilies corresponds to some of the observed differences in social 
structure, specifically the female biased sex ratio in the species with the 
greater sexual dimorphism. This is the same relationship described for 
primates, although primates exhibited a wider variation in sex ratio 
(ranging as high as 13.5 females per male: Clutton-Brock et al., 1977: Fig.1) 
than the highest ratio of 3.7 female per male in Globicephala 
macrorhynchus schools off Japan (Kasuya and Marsh, 1984). The 
implication is that Globicephala and Orcinus could have been selected for 
large size through intra-sexual selection via male-male competition. 
What is the evidence for male-male competition in delphinids? The 
role of aggression · in cetacean societies in general is not well understood 
(Norris, 1967). The bulk of reports on aggressive interaction and 
dominance hierarchies come from observations of captive animals 
(McBride and Hebb, 1948; McBride and Kritzler, 1951; Kritzler, 1952; Brown 
and Norris, 1956; Tavolga and Essapian, 1957; Brown, 1960; Pryor et al., 
1965; Tavolga, 1966; Saayman et al., 1973; Bateson, 1974; Defran and Pryor, 
1980; Ostman, 1991). These observations suggest aggressive interactions 
160 
11 
1. I 
are relatively common between all ages and sexes. However, the degree to 
which this is the result of confined conditions is unknown. Pryor and 
Schallenberger (1991), who were experienced dolphin trainers, reported 
similar aggressive behaviour s to those in captivity during their 
observations of spinner dolphins temporarily captured in tuna seine nets, 
but, again, the animals were confined. However, they did note a prevalence 
of aggression within subgroups of adult males, suggesting an element of 
male-male competition. There is no quantitative information on the 
frequency of aggression in wild dolphin societies, although occasional 
observations of aggressive interactions between individuals and groups are 
worth examining. Wells (1990) has reported on observations of aggressive 
interactions between Tursiops males from different communities: "The 
interactions have included such behaviours as tail slaps and violent leaps 
onto each other, resulting in bloodied fin edges and rostra" (p.220). Connor 
et al. (1992a; 1992b) have provided the most detailed accounts of aggression 
in the wild, both between male alliances and the females they aggressively 
herd, and between two or more male alliances as they attempted to steal 
herded females from one another. One observation of a presumed 
aggressive response (blowing bubbles and head-shaking) was noted in 
Tenerife pilot whales in response to a SCUBA diver in t he water. In 
general, aggression has been noted by the occurrence of tooth scars inflicted 
by conspecifics (McCann, 1974; Wells, 1991). However, even very young 
calves are scarred and mouthin g and scratching wit h the teet h is 
undoubtedly an element in play behaviour. As more underwater 
observations are conducted, it is likely that observations of aggression will 
increase . 
The behaviours of Turisops males within the male alliances 
described above indicate reduced male:male competition (Connor et al., 
161 
1992a; 1992b) and may be considered to be similar to male coalitions of lions 
(Bertram, 1976; Bygott et al., 1979) or grey langurs (Hrdy, 1977). Kin 
selection predicts that the males in these coalitions would be related, but 
this does not necessarily have to be so: the predictions of the benefits of this 
behaviour are very similar whether it is driven by kin selection or game 
theory (Packer and Pusey, 1982). In fact, genetic analyses of lions has 
shown that many of these males are unrelated (Packer et al ., 1991). This 
could also be the case with dolphin coalitions. Since dolphin litter size is 
one, there is no chance for brothers to be the same age as there would be for 
larger lion litters. Associations between similarly-aged male Tursiops 
appear to be between male calves from the same cohort. Although Tursiops 
parturition is only broadly seasonal, there are still seasonal peaks in births 
and a tendency for females of similar reproductive condition to associate 
(Wells et al ., 1987; Scott et al., 1990). This facilitates the formation of such 
cohorts. Male Tursiops coalitions (Connor, 1992a; 1992b) may represent 
such cohorts. 
A ritualised form of male-male competition in Orcinus may occur in 
the form of socio-sexual interactions between adult males (Heimlich-Boran 
and Heimlich-Boran, ms.). Sexual behaviour among interacting males has 
been observed in captive Tursiops (Ostman, 1991) and has parallels with 
mounting behaviour and penile display in terrestrial mammals such as 
ungulates (Coe, _1967; Geist, 1971; Hall, 1983; Klemm et al., 1983), 
macropods (Kaufmann, 1974) and primates (Ploog and McClean, 1963; 
Kawamura, 1967; Ploog, 1967; Hanby, 1974; Hanby 3:nd Brown, 1974; Hanby, 
1976; de Waal, 1982; Smuts and Watanabe, 1990). These behaviours appear 
to function primarily as dominance displays (in that the individual 
mounting or displaying an erect penis has been documented to be dominant 
via other methods), but have also been interpreted as warnings to foreign 
162 
conspecifics or even appeasement greetings (Wickler, 1966). For Orcinus, 
these interactions occurred just prior to the period of predominance of 
male-female interactions, as might be expected if these penile displays had 
a role in establishing a hierarchy which could be used to determine access 
to receptive females. The common occurrence of maturing adolescent 
males in these groups is also suggestive of competitive dominance 
interactions, since these males are likely to be in the process of forming 
relationships with the adult males in their own pod and in the community 
at large. Males in these groups were usually in body contact with each 
other; this could represent a form of the pushing and shoving contests for 
which large body size would be selected, as has been suggested for red deer 
(Clutton-Brock et al., 1979). These are similar to the "necking" contests 
between groups of male giraffes, which are also accompanied by erections 
(Coe, 1967). Their occurrence in male groups of Orcinus could be an 
indication of male-male competition, which is suggested by the sexual 
dimorphism of these whales. But the ritualised form of the behaviour 
makes it difficult to distinguish from affiliative behaviour, which would be 
encouraged by the relatedness of the males within the community (Bigg et 
al., 1990). 
For Tursiops, males in coalitions cooperate to compete with other 
coalitions (Connor et al ., 1992a; 1992b) and Sarasota males have aggressive 
interaction s with non-Sarasota males (Wells et al ., 1987; Wells , 1991), 
Orcinus appears to have a lack of inter-group aggression, and males from 
different communities have been observed to act_ively avoid each other 
(Mor ton, 1990). When two Orcinus pods from the same community meet, a 
r ituali sed greeting occur s (Osborne, 1986); in contr ast to aggressive 
interactions, these were the times when sexual interactions were observed 
(Heimlich-Boran and Heimlich-Boran, ms). The initial meeting of two pods 
163 
II. I 
I 
I 
I I 
I 
! , / 
1s stereotyped. Each pod forms a line (side-by-side) and the two pods 
approach face to face. After less than one minute the pods submerge and 
subsequently surface in a tight group with individuals from the two pods 
completely mixed and in body contact (Osborne, 1986). The initial phase of 
this intermingling behaviour is similar to observations of interactions 
between neighbouring bands of female langurs, but, in the case of the 
langurs, the result of such a stand-off is often fighting (Hrdy, 1977). The 
intermingling aspect of the behaviour is more like the ritualised greetings 
of wild dogs (Estes and Godard, 1967), hyenas (Kruuk, 1972; Mills, 1990), 
baboons (Smuts and Watanabe, 1990) and elephants (Moss, 1988). The lack 
of aggression may be related to the high degrees of relatedness described for 
this population of Orcinus (Hoelzel and Dover, 1990). 
In summary, the evidence for male-male competition in delphinids 
is still unclear. However, the lack of observations of aggressive male:male 
interactions in Orcinus over 20 years of study, suggests it does not play a 
major role. The lack of observations can not be attributed to the difficulty of 
observing male:male aggression in cetaceans: there are well documented 
observations of aggression between male humpback whales on their 
breeding grounds (Tyack, 1981; Tyack and Whitehead; 1983; Baker and 
Herman, 1984). 
There is a problem in relating the observed sexual dimorphism and 
potential male-male ·competition with the presence of a polygynous mating 
system for delphinids. Polygyny appears to be ruled out for the sexually 
dimorphic subfamily (Globicephalinae). This is also the group with low or 
absent rates of male dispersal, resulting in high degrees of relatedness 
between males and females in the same social group. Additionally, this 
group demonstrates lower degrees of genetic heterogeneity with indications 
of a more closed population structure (Duffield, 1986; Stevens et al ., 1989; 
164 
I ' I ·1 
Hoelzel and Dover 1990; Duffield and Wells, 1991; Hoelzel, 1991a). In 
general, polygyny is correlated with greater rates of male dispersal, due to 
increased intra-sexual competition forcing males to disperse and find better 
mating opportunities elsewhere (Greenwood, 1983; Shields, 1987). This 
dispersal, in turn, serves to increase genetic heterogeneity by enlarging the 
effective population size (Chepko-Sade et al., 1987). Effective population 
sizes have not been calculated for the two well-studied groups of Orcinus 
and Tursiops, and this is difficult even with the best data sets (Chepko-Sade 
et al., 1987). The positive relationship between increased genetic 
heterogeneity and male dispersal rates in the relatively monomorphic 
delphinine Tursiops will be discussed below. 
Variability in delphinid sexual dimorphism does not appear to be 
completely explained by hypotheses of sexual selection leading to polygyny. 
Ralls (1977) found the relationship between sexual dimorphism and 
polygyny to be highly predictable in extreme cases. However, the lack of 
extreme dimorphism in the globicephalines may indicate that other 
explanations are appropriate. There have been two general categories of 
explanations for sexual dimorphism other than sexual selection: ecological 
and energetic. 
The primary ecological explanation has been the concept of niche 
differentiation, which was originally noted in birds with dimorphic 
variation in characters such as bill length, indicating that males and 
females are feeding in different ways (Selander, 1966; Selander, 1972). This 
serves to reduce inter-sexual feeding competition ·and allows a more 
efficient utilisation of resources. This has been shown to be the case for 
sexually-dimorphic weasels which do not breed polygynously: hunting is 
done in burrows and the male is able to capture larger prey while the 
smaller female is able to go into narrower burrows and exploit different 
165 
areas (Brown and Lasiewski, 1972). It is not clear how important this 
process is in other mammals, although differences in diet have been 
reported for male and female Turisops (Barros and Odell, 1990). A widely-
dispersed resource may also favour larger body size to increase the ranging 
distance (via increased muscle mass) from a fixed site to locate food and 
result in selection acting on the sex which is primarily responsible for 
feeding the young (Bertram, 1979), although this is not applicable for 
cetaceans. This may be especially common in predators whose searching 
and pursuit costs are high (Schoener, 1969). Habitat may also affect sexual 
dimorphism. Terrestrial primates may be more dimorphic than arboreal 
primates because the latter are limited in their ranging by the smallest 
branch which will support their weight (Clutton-Brock et al., 1977). 
Aquatic species should be even less limited by gravitational forces and thus 
relatively free from constraints on increasing body size (Economos, 1983). 
Sexual dimorphism has been shown to be distributed differently in three-
dimensional, aquatic environments (Alexander et al., 1979). Male 
pinnipeds may be selected for small body size (e.g. male Weddell seals are 
smaller than females) through intra-sexual competition because the 
improved agility associated with small body size may be beneficial in 
competition (Alexander et al., 1979). Another hypothesis relating sexual 
dimorphism to habitat involves selection for small female body size. In 
fluctuating environments, females may be selected for early maturation 
which would result in a smaller body size (Wiley, 1974). Male body size may 
be free from this constraint. 
Other hypotheses explaining sexual dimorphism have centred 
around differential energetic constraints on the two sexes. Sandell (1989) 
developed the concept of a process of energetic optimisation leading to 
sexual dimorphism. He modelled the energy requirements of breeding and 
166 
I' I 
non-breeding times and identified which sex was living closest to the 
optimum. This often corresponded to the sex which experienced the 
greatest mortality. Females may also be selected for small body size 
because larger body size presents dangers of hyperthermy to the developing 
embryo because a decrease in the surface to volume ratio will serve to 
reduce heat loss (Greenwood and Wheeler, 1985). In cool, aquatic 
environments, these limits on female body size may be released and could 
explain why female baleen whales are larger than males (Ralls, 1976; 
Greenwood and Wheeler, 1985). Although hyperthermy can also kill 
sperm, many male mammals have external testes (a scrotum) to free them 
from this limitation and can evolve larger body sizes (Greenwood and 
Wheeler, 1985). Male cetaceans have internal testes, but they have been 
shown to have complex counter-current heat exchange systems in order to 
keep the sperm cool (Rommel et al., 1992). 
It is apparent that, apart from sexual selection, there are abundant 
theories to explain the occurrence of sexual dimorphism, suggesting that 
the presence and form of dimorphism is the result of complex processes, 
and multiple factors may be in effect for different species. It is difficult to 
evaluate the various ecological or energetic theories (as opposed to the 
sexual selection theories) for the occurrence of sexual dimorphism in 
delphinids. There is a critical lack of information about the distribution of 
resources in the oceans and the ways in which delphinids may exploit 
them. A number of theories suggest that in the aquatic environment 
selective pressures on body size are likely to be very dtfferent compared to on 
land. The lack of any clear correlation between dimorphism and ecological 
variables in delphinids may be a result of my classification of ecology. 
Factors other than preferred distance from shore and the taxonomic 
classification of prey may be important. There is also insufficient 
167 
information about the energetics of delphinids to understand the 
importance of the surface to volume ratio in cold, aquatic environments 
(Brodie, 1975; Hampton and Whittow, 1976; Schmitz and Lavigne, 1984). 
The critical importance of the energetics of locomotion in water is just 
beginning to be understood. Recent studies· suggest that delphinids do all 
they can to maximise energy efficiency (Blake, 1983; Whitehead, 1985; Fish 
and Hui, 1991; William et al., 1992). 
In summary, I conclude that the discrepancy between the greater 
degree of sexual dimorphism (indicating polygyny) and the low rates of sex-
biased dispersal (arguing against polygyny) indicate that the delphinid 
sexual dimorphism is not clearly related to the mating system. The 
significant increase in dimorphism in the larger species is probably an 
indication of the effect of factors other than sexual selection. The detailed 
patterns of delphinid mating systems (reviewed below) must be explained in 
terms of reproductive success before the role of sexual dimorphism can be 
understood. 
5.4.2 Brain size, Ecology, and Social Behaviour 
Analyses of Relative Brain Sizes (RBS) for nine genera of Delphinidae 
identified that variability in brain size was independent of body size. 
However, none of the social and ecological classifications identified any 
significant trends. · Larger RBSs may provide the improved neural 
processing needed to locate patchy food resources for frugivorous primates 
(Clutton-Brock, 1980), carnivorous-omnivorous carnivores (Gittleman, 1986) 
and frugivorous bats (Eisenberg and Wilson, 1978). This diet hypothesis did 
not apply to these delphinids; for example, the two squid-feeding species 
(Globicephala and Grampus) were found to have very different RBSs. 
However, the categorisation of diet on the basis of taxonomic classification 
168 
(e.g. fish vs. squid) is overly simplistic. If the selective advantage of large 
brains is the additional neural processing required to locate ephemeral yet 
rich patches of food which are evenly and widely distributed (Jerison, 1973; . 
Mace et al., 1980), variation in RBS should be related to the behaviour and 
distribution of the prey. The degree of schooling behaviour of the prey could 
also affect RBS: schools are effective predator avoidance systems (Hamilton, 
1971) and animals preying on active schooling prey would (in theory) 
require greater neural processing capabilities than those which capture 
solitary or sedentary fish (Norris and Schilt, 1988). Both squid and fish 
have schooling and non-schooling representatives, so the classification of 
diet based on these categories would miss the behavioural aspect of the 
prey. Detailed prey data are not available for most delphinids. In cases 
where they are, the dolphins appear to prey on a wide variety of types (Fitch 
and Brownell, 1965; Clarke, 1986; Barros and Odell, 1990; Barros and 
Cockroft, 1991; Overholtz and Waring, 1991; Gales et al., 1992). It could also 
be argued that all resources which delphinids exploit (except for the bottom-
dwelling crustacean used by Cephalorhynchus) are patchily distributed 
and thus there would be selection for larger brains across the taxa. 
Another hypothesis to explain relatively larger brains in comparative 
studies has been the social complexity hypothesis, and this may have some 
relevance to social structure and mating system. For the present study, the 
only social variable which could be used to test this hypothesis was group 
size. No significant trends were found, but perhaps other elements of 
society (such as socionomic sex ratio) may explain the variability in brain 
size better. The two genera for which we possess the most information on 
social structure, Tursiops and Orcinus, both have relatively larger brains 
than other delphinids. This is consistent with a social complexity 
hypothesis. However, Globicephala, which appears to share many 
169 
elements of complex social structure with Orcinus, had a relatively much 
smaller brain. Although relative brain sizes of all delphinids are 
significantly above the regression of brain and body size for all mammals 
(Worthy and Hickie, 1986), and this may have relevance to an 
understanding of their social organisation. 
Al though cetacean brains are generally larger than the brains of 
terrestrial mammals (Jerison, 1973), they have a less complex structure 
(Morgane et al., 1986b; Morgane et al., 1986a; Ridgway, 1986b). The dolphin 
brain has a greater degree of cortical folding than the human brain (index 
of 4.4 7 vs. 2.86: Elias and Schwartz, 1969), but the average volume of the 
cortex is 80% of the human's because dolphin cortex is so thin (Ridgway, 
1986b). Also, there is reduced cortical differentiation indicated by the 
agranularity of the cortex, which is characteristic of the primitive 
mammalian brain (Morgane and Jacobs, 1986b). It has been proposed that 
this reduced granularity is because the primitive cetacean ancestors split 
off from the rest of mammals before the more advanced granularity of the 
modern terrestrial mammalian brain evolved (Morgane and Jacobs, 1986b). 
They have apparently compensated for this primitive cellular condition by 
an enormous enlargement of the surface area of the cortex (Morgane and 
Jacobs, 1986a). It may be that these essential differences of cetacean brains 
make them incomparable with the brains of terrestrial mammals. 
However, simplified brain structures need not necessarily imply limited 
processing abilities (Jerison, 1986). For example, the visual cortex of birds 
is significantly simpler in structure than the ma~malian visual cortex, 
but birds have vision equivalent to primates (Jerison, 1986). Jerison (1976) 
suggests " .. at present, it may be heuristic to continue to accept simple 
quantitative measures · {of brain size} and to disregard differences in 
organisation". 
170 
1111 
111 
q 
11 
I ' 
' 11 
It is likely that the large brain of delphinids is specialised for acoustic 
signal processing, especially echolocation (Evans and Bastian, 1969; 
Herman and Tavolga, 1980; Wood and Evans, 1980; Jerison, 1986; Worthy 
and Hickie, 1986; Moore, 1991). Although bats echolocate and have much 
smaller brains (Eisenberg and Wilson, 1978), the frequency spectrum of 
signals which they project and the cochlear adaptations they use to receive 
these signals covers a much narrower band than that of delphinids (Worthy 
and Hickie, 1986). The use of broad band clicks, rapid click repetition rates 
(up to 1000/sec.) and the fine-grained resolution of delphinids (Popper, 1980) 
likely requires a great deal of neural processing (Wood and Evans, 1980). 
This could explain the large relative brain sizes compared to other 
mammals (Worthy and Hickie, 1986). 
5.4.3 Relative Testes Size and the Potential for Sperm Competition in 
Delphinids 
Models for the evolution of sperm competition (Parker, 1984) are 
based on two strategies of male mating behaviour which centre around the 
concept of paternity assurance (Alexander, 1974). First, a male may guard 
the female (or a group of females) to prevent second matings by other males 
through male-male conflict, thus guaranteeing his paternity. 
Alternatively, if the female (or group of females) is not defensible, thus 
reducing the male's paternity assurance, then the male can only compete 
with other males through increasing his volume of sperm to out-compete 
other male's sperm in the fertilisation of the ovum. T~e selection for either 
of these strategies depends on the costs to the female of the male's 
behaviour. If mate guarding or multiple matings are costly to the female 
there will be male-female conflict and the system will be destabilised. 
However, if mate guarding is not costly (or is even beneficial) to the female, 
171 
then mate guarding will evolve. An example of this is found in the 
behaviour of female fallow deer mating on leks: it is beneficial to them to 
mate with a single male on his lek because his presence reduces the usual 
harassment she receives from multiple males attempting to mate (Clutton-
Brock, 1989). If multiple matings are not costly (or are beneficial) to the 
female (as in the case of lions: Davies and Boersma, 1984), then sperm 
competition will evolve. 
Kenagy and Trombulak (1986) were first to suggest that odontocete 
cetaceans have exceptionally large testes compared to other mammals. 
Observations from captive Tursiops suggest that mating occurs 
promiscuously (McBride and Hebb, 1948; Brown and Norris, 1956; Tavolga 
and Essapian, 1957; Tavolga, 1966; Tayler and Saayman, 1972; Saayman et 
al., 1973). However, Tursiops had relatively small testes for a delphinid 
(Figure 5.9). This would appear to indicate that other delphinid species 
(previously unstudied) have even greater potential for sperm competition. 
5.4.4 Patt.ems of Dispersal and Genetic Diversity 
Patterns of dispersal in mammals are also likely to be important in 
determining the form of social relationships (Harcourt and Stewart, 1983; 
Wrangham, 1983; Smuts, 1987), the degree of genetic variation due to levels 
of inbreeding (Melnick and Hoelzer, 1992), the prevalence of nepotism and 
other altruistic acts ·present through kin selection (Moore, 1992) and the 
form of mating systems (Shields, 1987). In most mammal and bird species, 
one sex or the other disperses from their natal group,' apparently to avoid 
inbreeding with close relatives, but perhaps also to find better feeding and 
mating opportunities elsewhere (Packer, 1979; Greenwood, 1980; Pusey, 
1980; Shields, 1982; Greenwood, 1983; Moore and Ali, 1983; Shields, 1983; 
Pusey and Packer, 1987b; Pusey and Packer, 1987a). Alternatively, natal 
172 
philopatry has been shown to have benefits to group living (Bengtsson, 1978; 
Shields, 1983); in fact the entire principle of kin selection is based on natal 
philopatry (Moore, 1992). It is beneficial for kin to remain together because 
altruism is favoured and animals can cooperate in resource acquisition 
with reduced benefits to cheaters who exploit the system. Thus, there is 
ultimately a cost-benefit trade-off related to resource and mate acquisition 
and avoidance of predation involved with the decision to disperse or remain 
in a natal group. Documented dispersal patterns in Tursiops suggest that 
mixing between communities occurs through the movement of males (and 
to a lesser extent females: Duffield and Wells, 1991). There is genetic 
evidence of reproductive isolation between killer whale communities 
(Hoelzel and Dover, 1990; Hoelzel, 1991a) and strong observational evidence 
of a complete lack of male dispersal from natal groups (Bigg et al., 1990). 
These differences are apparent in comparisons of mtDNA variation 
between Tursiops and Orcinus (Duffield and Wells, 1991). These 
differences in dispersal patterns and degree of genetic heterogeneity 
provide the strongest available evidence that the two species have different 
mating systems. 
5.4.5 Hypothetical Delphinid Mating Syst;ems 
I propose that the well-studied members of the two delphinid 
subfamilies, Delphininae and Globicephalinae have different social 
systems. The lack of a long-term bond between reproductive males and 
females in both groups indicates mating must primarily be occurring 
promiscuously. Females remain in association with kin in both groups, 
but the stable units of Orcinus society (pods) encompass three levels (intra-
pod, subpod and pod) whereas the stable Tursiops female bands which 
interact fluidly appear to be equivalent to only the smallest Orcinus unit. 
The behaviour of the males in these two species is different. 
In Tursiops, males remain (in the long-term) on their natal home 
ranges, but disappear for weeks or months at a time, suggesting they are 
adopting a roving male strategy (Clutton-Brock, 1989; Whitehead, 1990). In 
one sense, this is a resource defence system, where the range of a male (or 
group of males) encompasses the ranges of more than one female 
community (Clutton-Brock, 1989; Davies, 1991). However, the males 
apparently only defend their natal community, as indicated by aggressive 
interactions with males from adjacent communities primarily along the 
periphery of the range. Roving may be an alternative to mating with non-kin 
female bands within the male's own community. The strategy of roving 
has been hypothesised to be most beneficial if females are unpredictably 
distributed or clumped into small units(Clutton-Brock, 1989), as bands of 
Tursiops females are. Other males appeared to remain longer in a female 
community's home range and move from female group to female group, 
perhaps sampling for receptive females. Models of the trade-off between 
these two strategies of roving and residency indicate that males should stay 
resident when the travel time between female groups is greater than the 
oestrus period of a single female (Whitehead, 1990). Indications that female 
Tursiops have spontaneous, multiple ovulations (Kirby and Ridgway, 1984) 
over a wide seasonal range (Wells et al., 1987) suggest that it would be 
difficult for males to predict when there is a maximum likelihood of 
encountering receptive females. Ovulation may be detectable at close range 
through the chemo-reception of hormonal cues in small quantities of urine 
(Nachtigall, 1986; Schroeder, 1990). However, the range of detection of such 
clues is limited in water due to dilution 
174 
Males will also form alliances to control access to receptive females 
(Wells et al., 1987; Wells, 1991; Connor et al., 1992a; 1992b). These alliances 
may be based on kinship, but are likely to be also due to the formation of 
cohorts of more distantly related males; a similar situation of alliance 
formation as described for lions (Packer et al., 1991). Such alliances may be 
aided by the comparative ease in defending female groups in shallow water, 
the primary habitat of this species. Nevertheless, field observations 
indicate that two or three males are required to herd a female. Thus female 
behaviour appears to be promiscuous unless males actively intervene. Even 
with herding, there can be no certainty of paternity. Males do not maintain 
long-term contact with the females and return to the common pattern of 
age and sex segregation within their natal community. Males appear to 
provide no parental care as they are segregated from the young for most of 
the time. The variability in the reproductive success of males is still 
unknown and ongoing paternity studies may undermine this view of 
Tursiops society. 
There is also strong evidence of a promiscuous mating system in 
Orcinus , defined by the lack of a mating bond between breeding males and 
females . While observations of mating are limited, sexual interactions 
appear to be occur between members of different pods. However, the males 
involved in these interactions returned to their natal pod. Given that males 
appear to remain closely bonded to their natal group, mating between pods 
would be a clear example of inbreeding avoidance (defined as the avoidance 
of mating with immediate kin: Shields, 1987). A si:rp.ilar situation was 
noted in interactions between kin in the black-tailed prairie dog (Hoogland, 
1982). Occasionally, a female came to maturity in the coterie of her adult 
male kin. One of the mechanisms involved in avoiding inbreeding was 
apparently based on the rejection of potential mates who were recognised as 
175 
kin (Hoogland, 1982). However, all indications are that genetic exchange 
occurs entirely within the confines of the local Orcinus community (Stevens 
et al., 1989; Bigg et al., 1990; Hoelzel and Dover, 1990; Hoelzel, 1991a). Males 
within a community appear to have ritualised interactions, which could 
either represent dominance or affiliative behaviour. This would fit with 
Moore's (1992) suggestion that there should be a lack of overt aggression 
and a prevalence of ritualised behaviours between closely related animals. 
Globicephala, although not yet studied in a long-term observational 
study, appears to share a number of social system features with Orcinus. 
The genetic information from G. melas indicates that males and females 
from the same pod are related (Amos et al ., 1991a; 1991b). Although the 
overall community structure is unknown, there are indications of genetic 
differentiation between pods, suggesting a restricted mating population 
(Andersen, 1988). My observations of G. macrorhynchus confirmed the 
existence of two, apparently non-interacting, segments of the population 
(resident and visitors), as well as patterns of association between pods 
which could provide the basis for the observed genetic differentiation. The 
resident community showed no indication of age and sex segregation. The 
pods often split into smaller groups, making them basically equivalent to 
Orcinus subpods. Associations of the reproductive females were found to be 
predominantly with males when the linked pods travelled together, further 
supporting the idea of an inbreeding avoidance mechanism. 
However, the main question to be asked about these groups is: Why do 
the males remain in their natal group? I propose that they do so because of 
a promiscuous system of mating with a low confidence of paternity. While 
Tursiops males appear to be able to control access to females through 
herding, there have never been any similar observations in Orcinus. This 
could be because of the females' large body size and the reduced mobility of 
176 
the even larger males. Guarding of females 1n an aquatic three-
dimensional habitat must be extremely difficult. An indication of this 
comes from a comparison of mating systems of pinnipeds which copulate 
on land or in water (Stirling, 1983). Species which copulate on land are 
highly polygynous and sexually dimorphic with one male controlling 
access to a large number of females (Le Boeuf, 197 4). Species with aquatic 
copulation show low degrees of polygyny and sexual dimorphism (Stirling, 
1983) probably due to the males inability to control access to the female 
because of her increased mobility underwater and the difficulty of 
maintaining aquatic territories (Ralls, 1977). In the face of this inability to 
guarantee paternity, sperm competition (Parker, 1984) should be important. 
While Orcinus falls exactly on the predicted line relating testes size to body 
size within Delphinidae, this family-wide degree of testes development has 
been shown to be significantly greater than expected for mammals in 
general (Kenagy and Trombulak, 1986). 
Why should a male return to his natal group after mating? Why not 
rove and search out other receptive females? I think there are four 
answers: 1) the low densities of large predators, such as Orcinus and 
Globicephala (Bonner, 1988), may make the roving model's travel time 
between receptive females greater than the duration of a female's oestrus 
(Whitehead, 1990), 2) the costs of moving to an unfamiliar area and facing a 
potentially unwelcome reception, may be too high (Bengtsson, 1978); 3) the 
benefits of cooperatively hunting with known pod mates may increase 
feeding efficiency (Packer and Ruttan, 1988) and 4) increased inclusive 
fitness within a pod (Hamilton, 1964). 
Inclusive fitness is a measure of an individual's fitness which 
includes not only its own reproductive success, but also the reproductive 
success of its relatives, devalued by the appropriate degree of relatedness 
177 
(Wilson, 1975; Trivers, 1985). In promiscuous mating systems, males have 
little guarantee of paternity (Alexander, 197 4), thus, they are unlikely to 
form a bond with a female and provide any parental investment to her 
offspring, because these may have been sired by another male (Trivers, 
1972). This risk of cuckoldry or desertion (Trivers, 1972) is likely to be great 
for globicephalines because of the apparent lack of mate guarding. 
Hamilton (1964) reasoned that an animal should base any altruistic 
behaviour (e.g. any behaviour which may be detrimental to individual 
fitness but apparently beneficial to a neighbour's individual fitness) on the 
degree of relatedness (r), or the proportion of shared genes between it and 
its neighbour. Specifically, the ratio of gain to loss (K) of any altruistic 
behaviour should be greater than the reciprocal of the degree of relatedness 
(K > 1/r: Hamilton, 1964). For example, parents are related to their 
offspring by 0.5 because, on average, each parent contributes one-half of the 
offspring's genome. Thus, any social act between parents and offspring 
should be favoured by natural selection if the ratio of benefit to cost is 
greater than two. Even if the parent dies, it could have increased it's 
inclusive fitness through kin selection (Maynard Smith, 1964). 
Promiscuous mating systems have considerable effects on the 
certainty of male paternity, or the degree to which a male can judge his 
degree of relatedness to his offspring (Alexander, 197 4). A male mating in 
such a system may have a degree of relatedness to his mate's offspring of 
0.5 (if he is the father) to O (if another male is the father). However, he is 
sure of his degree of relatedness to his mother and his siblings, through 
what has been termed uterine kinship (Flinn, 1981). Thus, paternity 
uncertainty only reduces relatedness through putative patrilineal kinship 
links, and has no effect on relatedness through matrilineal kinship links. 
Alexander (1974) noted that in promiscuous mating systems, a male may be 
178 
more sure of his relatedness to his sister's offspring (r = 0.25 or 0.125, 
depending on whether the brother and sister are full or half siblings) than 
he would be to the offspring of his mate, with r varying between 0-0 .5 
depending on his confidence of paternity. Alexander (197 4; 1977) modified 
Hamilton's altruistic equation to K > (1/r) * p, where p equals paternity 
certainty. With no confidence of paternity, any offspring of his sister would 
be his closest relatives in the next generation (Alexander, 197 4). The model 
predicts there should be a paternity threshold, Pt, or a degree of paternity 
certainty below which altruistic behaviour towards his sister's offspring 
will increase his inclusive fitness more than caring for his putative 
offspring with unrelated females. Various assumptions provide estimates 
of a critical value ofpt between 0.268 (Kurland, 1979) and 0.33 (Flinn, 1981). 
This rationale has been used to explain the evolution of human 
societies in which the mother's brother or uncle has a greater responsibility 
for the young in a group than he does for his own offspring (Alexander, 
1977; Kurland, 1979; Flinn, 1981; Flinn and Low, 1986; Borgerhoff Mulder, 
1991). These societies are characterised by a large degree of promiscuity, 
occasionally with the man and wife living in completely separate houses. A 
quantitative survey of 150 randomly chosen societies found a significant 
correlation between the importance of this avuncular relationship and 
degree of paternity confidence (Flinn, 197 4; 1986). These societies have been 
termed avunculate societies (Alexander, 1977). 
I propose that Orcinus and Globicephala live in avunculate societies. 
In relation to other forms of human society, Alexan'der (1974) suggested 
that avunculate societies could share elements of both matriarchal- and 
patriarchal-based cultures. In relation to mammalian societies, I would 
suggest that an avunculate society be considered as a matriarchy (since 
179 
differential mortality will always result in females being the oldest 
members) with adult male sons and brothers attached. 
Kurland (1979) compared primate societies with avunculate human 
societies and concluded that the behaviours of females of many primate 
species were similar to human females in terms of the formation of 
matrifocal groups with a permanent association of uterine kin (Kurland, 
1977), but there were no cases where males preferentially associated with 
female uterine kin, and he concluded the avunculate social system was 
uniquely human. Questions have been raised about the validity of Flinn's 
(1974) survey and the evidence for high promiscuity levels in humans. It 
has been suggested the critical element in the explanation of avunculate 
societies are whether they are based on sex-specific differences in dispersal 
costs or in differential benefits from cooperating with kin (Borgerhoff 
Mulder, 1991). 
Moore (1992), in his review of the relationship between dispersal and 
nepotism in primates, outlined an argument which contributes to an 
explanation of the benefits of such societies. Dispersal and nepotism are 
linked by the simple fact that nepotism, defined as the direction of beneficial 
or altruistic acts toward kin, can not develop if kin have dispersed (Moore, 
1992). In other words, contacts between kin need to be sufficiently frequent 
to allow nepotism to develop. The link between these two processes has been 
enhanced by an attempt to explain the prevalence of male dispersal in most 
mammals, using theories of the evolution of nepotism through kin selection 
(Hamilton, 1964) and the evolution of matrilineally-related female primate 
groups based on the defendability of food resources (Wrangham, 1980). 
Coupled with documentation of inbreeding depression (Ralls et al ., 1979; 
Ralls et al., 1980), a common scenario of male dispersal has been proposed. 
Moore (1992) states this scenario as: 
180 
"In all but a few exceptional species, primate troops are formed of 
matrilineally related females which cooperatively defend their 
feeding range and/or guard each other against predators. Since 
inbreeding depression is a potentially serious problem, 
inbreeding must be avoided. The easiest way to do this is by 
differential dispersal from the natal group. Since females are 
staying, males have to go. Accordingly, male-biased dispersal, 
placing limits on possibilities for male nepotism, is the predicted 
consequence of female nepotism plus inbreeding depression." 
p.362 
Why should there be a causal link between dispersal and nepotism? I 
believe that avunculate societies suggest there need not be. Four conditions 
should be required for the formation of an avunculate society: 1) high 
benefits to remaining with kin (e.g. from nepotism), 2) low rates of 
dispersal of both sexes, 3) a mechanism for extreme inbreeding avoidance 
through out-group mating and 4) low degrees of paternity certainty. The 
benefits of remaining with kin have evolved through kin selection 
(Hamilton, 1964). In the traditional female philopatry/male dispersal 
system, kin-selected benefits for females accrue within the social group 
while kin selection for males is beneficial through kin-biased dispersal of 
brothers (Bertram, 1976; Bygott et al., 1979; Packer and Pusey, 1982; Cheney 
and Seyfarth, 1983; Packer et al., 1988; Packer et al ., 1991). In the less 
common male philopatry/female dispersal system, males receive the 
within-group benefits, while females transfer and then build up their own 
kin in a different group (Pusey, 1980; Goodall, 1986). Avunculate societies 
allow both sexes to benefit from within-group kin selection. Since all group 
members are related, the potential for inclusive fitness gains are greatly 
increased, whereas with one sex dispersing, inclusive fitness gains 
181 
I I 
I I 
I 
I 
r 
through interactions between adults are limited to one sex. The second 
condition appears to entail the costs of inbreeding depression which are 
presumed to necessitate dispersal, but the third condition solves this 
dilemma. 
The process of a male mating outside of his social group does not 
appear to have any clear precedence in other mammalian societies. There 
is some evidence of very low or zero rates of male dispersal in Barbary 
macaques, with males leaving periodically to mate with females in 
adjacent groups (Mehlman, 1986 in Moore, 1992), but the field studies are 
still relatively short-term. Some male brown hyenas remain within their 
natal groups, but this appears to result from delayed dispersal due to 
unfavourable conditions; males eventually disperse when there is another 
female group to join (Mills, 1990). However, the social groups of Orcinus 
represent feeding groups of related kin and the seasonal multi-pod 
aggregations of all community members are the true breeding groups, as 
defined by Gittleman (1989). Moore (1992) has pointed out that the 
presumed obligatory causality between dispersal and inbreeding avoidance 
can be reconsidered through carefully defining dispersal as the dispersal of 
gametes (Shields, 1987), not requiring a dispersal of individuals. If a low 
degree of paternity confidence is overlaid on top of these other 
considerations, an avunculate social system may be optimal. 
The final question is: why are the groups of Globicephala and 
Orcinus so stable? If I am correct in hypothesising an avunculate social 
system for these whales, it has to be primarily based on either: 1) the 
benefits accrued to adult males by living in a group with closely related 
females, or 2) the excessively high costs of dispersal into adjacent groups of 
females. There is no way to evaluate these costs with the present data, 
except to explore the benefits of staying in a group and then assume that 
182 
dispersal cost would involve the loss of those benefits. It would seem that 
the benefits would have to be non-reproductive, because living with your kin 
with whom you avoid inbreeding would reduce your immediate access to 
potential mates; although membership of a dominant matriline could 
improve access to females in other matrilines . The primary non-
reproductive benefits to group living are improved resource acquisition and 
protection from predators (Alexander, 197 4; Bertram, 1978; (West Eberhard, 
1975); Wrangham and Rubenstein, 1986). 
Improved resource acquisition could come about through cooperative 
hunting, which can only evolve when the per capita rate of food intake 
within a group exceeds that of a solitary individual (Packer and Ruttan, 
1988). Comparisons of group sizes in resident and transient Orcinus have 
suggested that the larger resident group sizes may be related to improved 
benefits from cooperative hunting on an abundant, schooling prey such as 
salmon CJ.Heimlich-Boran, 1987; Felleman et al., 1991; Baird et al ., 1992), 
while small transient group sizes are optimal for the individual capture of 
single, large prey such as pinnipeds. However, it is likely that Orcinus 
hunting pinniped prey also cooperate (Lopez and Lopez, 1985; Hoelzel, 
1991c; Jefferson et al., 1991), and that the differences in group size are 
simply foraging specialisations (Baird et al., 1992). 
Norris and co-authors (Norris and Dohl, 1980; Norris and Schilt, 
1989) have proposed that one of the main benefits to group living in 
delphinids comes from the sensory integration system, which involves "the 
receipt of environmental information by any member of a group, and 
passage of a reaction to it through the group in all directions within the 
school" (Norris and Schilt, 1989, p.156). This system could be used for the 
group detection of prey and improve the feeding success of all individuals. 
In fact, observations of Orcinus feeding on salmon schools suggested that it 
was primarily prey detection (by flank formation swimming) and prey 
herding (by splashing and vocalising while progressing towards shorelines 
or underwater seamounts) which were cooperative. Once feeding 
commenced the whales broke ranks and fed individually (J.Heimlich-
Boran, 1988). 
Resource distribution has been shown to be one of the prime 
determinants of group formation and mating systems (Orians, 1969; 
Alexander, 1974; Emlen and Oring, 1977; Clutton-Brock, 1989; Davies, 1991). 
Even in terrestrial environments, where ecological research has been 
conducted for many years, the dispersion of resources is difficult to 
categorise. Terms such as clumped and patchy are widely used, but have 
no standard, quantitative definitions. The distribution of resources in the 
marine environment is even less well understood. Correlations between 
delphinid occurrence and the known distribution of their preferred prey 
has been applied on a broad seasonal basis (Wursig and Wursig, 1979; 1980; 
J.Heimlich-Boran, 1986), but the long-term variability of the majority of 
marine resources is unknown. The ocean is often a highly stratified 
environment with thermoclines and density boundaries which will affect 
the distribution of prey. The three dimensional distribution of prey could 
also place unknown constraints on resource utilisation. However, it has 
been suggested that the predictability and abundance of seasonal 
migrations of salmon returning to their natal rivers is likely to be the 
reason resident Orcinus populations have been established in the inland 
marine waters off British Columbia (J.Heimlich-Boran, 1986; J.Heimlich-
Boran, 1988; Felleman et al., 1991; Baird et al., 1992). The season when 
salmon are most abundant is also the season when pods travel together in 
large aggregations, when most sexual interactions were observed between 
males and females, and the peak season of conception (Olesiuk et al., 1990; 
184 
Felleman et al., 1991; Heimlich-Boran and Heimlich-Boran, ms.). One of 
the primary implications of an avunculate social system with inbreeding 
avoidance is that cohesive social groups must periodically make contact for 
out-group mating. This can only be done in an area with sufficient 
resources to support the nutritional requirements of more than one social 
group. The apparent narrow range of distribution of pilot whales off 
Tenerife (Chapter 3), also suggests that this is an area where resources are 
locally abundant and contact between pilot whale pods could be maintained. 
The second major benefit to group living is protection from predation, 
and this has been discussed as a primary factor in the evolution of 
delphinid schools (Norris and Dohl, 1980b; Wells et al., 1980; Norris and 
Schilt, 1989). Any organism living in an open habitat is susceptible to 
predation (Hamilton, 1971). The sensory integration system discussed 
above is probably a key adaptation to occupying such habitats, and is used by 
both fish and dolphin schools (Norris and Schilt, 1989). However, the 
development of echolocation has given delphinids an inherent advantage 
over sharks and has probably allowed them to maintain a high degree of 
social complexity (Norris and Schilt, 1989). Although predation is likely to 
be less important for large-bodied delphinids compared to smaller species, 
it cannot be ruled out. I observed a young pilot whale with large scars 
which was probably from a shark attack. Larger body size could confer an 
advantage for male globicephalines in terms of potentially defending the 
pod from shark attack. Males often travelled on the periphery of the pod 
and when approached underwater by divers , always , came close for an 
inspection. Dolphins have been known to attack sharks, probably as a form 
of group defence (Wood et al., 1970). 
These indications of a role in group defence for male delphinids could 
be consider ed as a form of parental car e. Other anecdotal indicat ions of a 
185 
parental role for Orcinus males comes from observations of allo-parental 
care, involving males "baby-sitting" young calves while their mothers were 
away (Haenel, 1986). Also, a male Globicephala off Tenerife was observed 
in close association with a newborn calf, less than one week old, while its 
mother presumably dove to feed. Male Orcinus have also been considered to 
have a teaching role in hunting (Lopez and Lopez, 1985; Guinet, 1991; Caro 
and Hauser, 1992). Males were observed to catch sea lions on the beach and 
then bring them offshore to waiting juveniles. Although the degrees of 
relatedness of these animals was unknown, the allo-parental role of the 
whales off British Columbia could be explained through the inclusive 
fitness gained via an avunculate social system. 
Total natal philopatry allows for a build-up of cultural knowledge 
through the maintenance of inter-generational bonds (Bonner, 1980; 
Nishida, 1987). Older members, especially the longer lived females, have 
the experience and memory of previous feeding success (Wursig, 1986). 
This is similar to the role ascribed to matriarch elephants: as repositories 
of critical information about the surrounding habitat in terms of the 
temporal and spatial distribution of resources (Douglas-Hamilton and 
Douglas-Hamilton, 1975; Dublin, 1983; Moss and Poole, 1983; Moss, 1988). 
This may be of more importance for marine mammals than terrestrial 
mammals, given the patchy distribution of resources in the three-
dimensional marine environment (Wursig, 1986; Norris and Schilt, 1988; 
Norris and Pryor, 1991). All animals in the pod would benefit from 
enhanced success which must surely come from working within a 
culturally co-adapted group. 
The complexity of globicephaline life is also suggested by the long 
maturation period, which indicates that a considerable amount of learning 
is necessary for survival (Brodie, 1969). This long maturation also requires 
186 
large amounts of parental investment, and the presence of a long period of 
reproductive senescence with an extended lifespan (Kasuya and Marsh, 
1984; Marsh and Kasuya, 1984; Marsh and Kasuya, 1986; Bigg et al., 1990; 
Olesiuk et al., 1990; Marsh and Kasuya, 1991) may be a form of terminal 
investment (Trivers, 1974; Clutton-Brock, 1984). Parental investment theory 
predicts that as a female reaches the end of her lifespan the costs of 
parental investment, in terms of delayed future reproduction, will decrease 
(Trivers, 1972). This will select for a longer period of terminal investment, 
and result in a reduction in parent-offspring conflict (Trivers, 197 4). An 
overall increase in reproductive effort towards the end of the lifespan has 
been predicted for species in which reproductive value declines with age, 
but has been difficult to document (Clutton-Brock, 1984). In red deer, 
although fecundity declines with age, maternal investment increases for 
older females. Offspring survival has been shown to increase with an 
increase in suckling bouts for the young of older females (Clutton-Brock et 
al., 1982). In a related finding, survival in calves whose mothers failed to 
raise a subsequent calf was also enhanced through delayed weaning and 
the attainment of higher social rank (Clutton-Brock, 1984). Combined, 
these two factors provide a mechanism for the evolution of the post-
reproductive period in globicephalines. 
The confirmation of the existence of an avunculate social system in 
globicephalines will require measurement of degrees of relatedness within 
and between pods. However, I believe that the existence of such a system is 
the best explanation of the current data set available for these whales. Also, 
the system's existence is predictable within the current theories of natural 
and kin selection, given the occurrence of low male dispersal, low 
confidence of paternity, and large benefits to living with kin. 
187 
I I 
These broad generalisations on the social organisation and mating 
systems of delphinids are undoubtedly overly simplified. I am quite sure 
that further research will show subtle variations in these trends, especially 
as more populations are studied. Most of the research presented here is 
based on single populations of each species, so there is little information on 
the effects different environments may have on social systems, nor on how 
widespread such systems are. 
5.5 SUMMARY 
The cetacean family Delphinidae was examined in a two part 
comparative study: 1) the morphological, ecological and behavioural 
characteristics of sexual dimorphism, brain size, testes size, group size, 
diet and habitat preference were examined relative to current theories of 
mating systems and 2) a detailed comparison of genetic and observational 
studies on the social structure of three genera (Tursiops {subfamily 
Delphininae}, Orcinus and Globicephala {Globicephalinae}) was used to 
develop hypotheses concerning mating systems. Sexual dimorphism was 
significantly correlated with absolute body size, ranging from the smallest 
genus, Cephalorhynchus, in which females were larger than males, to the 
large globicephaline whales, with males up to 1. 7 times larger than 
females. Brain size also varied significantly with body size, but measures 
of relative brain size did not correlate with diet, habitat or group size. In 
fact, species with similar diets or group structures had very different 
relative brain sizes. Relative testes size also showed variation independent 
of body size, but again was uncorrelated with group size as might have been 
predicted from the theory of sperm competition. The lack of correlation 
between these morphological characteristics and the ecological and social 
188 
classifications could have been due to the choice of inappropriate categories 
of diet and habitat because of a lack of ecological data for delphinids. 
Tursiops (and a few other delphinine genera) were shown to live in 
age and sex segregated groups in which all individuals primarily 
interacted with other individuals of similar age and reproductive status. 
Fluidly-interacting bands of females and calves formed relatively discrete 
communities which showed some signs of genetic differentiation. Males 
primarily remained within specific areas of their natal home range, but 
appeared to leave to interact with adjacent communities. Males were 
considered to be the vectors of genetic interchange between communities 
and aggressive interactions were observed between males from different 
communities. Males occasionally formed coalition5 to herd females, 
presumably to control mating access. Orcinus lived in highly stable pods of 
mixed age and sex from which no dispersal of either sex was observed in 20 
years of observation. Community structure consisted of a number of pods 
which freely interacted on a seasonal basis. Genetic studies indicated high 
degrees of inbreeding within communities and differentiation from 
adjacent communities. A sympatric form with highly different group 
structure and foraging specialisations was reproductively isolated, further 
emphasising the closed population structure. Limited observations of 
sexual interactions suggested mating was taking place between pods 
within the same community. Although not documented through long-term 
observations, genetic and short term observations of Globicephala 
suggested they may also live in pods of mixed age an4 sex in which adult 
males and females are related and the males are not the fathers of the 
offspring in the group. Seasonal association between males and females 
from different pods was suggested as the time when out-group mating 
could occur. The pattern of age and sex associations was similar to 
189 
Orcinus. The shared life history features such as differential mortality for 
males and females and the presence of a long post-reproductive period 
during the last one-third of the females' life, suggests that these two 
globicephaline whales shared the same social system. 
A promiscuous mating system was proposed for both of these groups, 
based on the short-term contact between mating males and females. 
Delphinine males appeared to adopt a roving male strategy, occasionally 
forming coalitions to control access to receptive females, and ranging into 
adjacent female communities. A social system, previously only 
documented for humans, was hypothesised to account for the 
globicephaline observations. This is termed an avunculate social system 
which consists of a matriarchy with adult male sons and brothers attached. 
This can arise when paternity certainty is so low that a male will be more 
closely related to his sister's offspring than he will be to the offspring of his 
mate. Males appear to perform functions of group defence from predators, 
offer occasional allo-parental care and could aid in food detection and 
capture, which could result in benefits through inclusive fitness. The 
benefits of living with kin could result in a build-up of cultural knowledge 
through the maintenance of inter-generational bonds. Confirmation of this 
unique social system will have to await further studies on paternity and the 
variance in male reproductive success. 
190 

CHAPTER 6: CONCLUDING DISCUSSION 
The aim of this thesis has been to examine the social ecology of the 
cetacean family Delphinidae in light of current theories of the organisation 
and evolution of terrestrial mammalian societies. In order to expand the 
information available for this examination, I conducted a study on the 
social organisation of the short-finned pilot whale, for which there was 
much biological information collected from carcasses, but which had been 
relatively unstudied in the wild. A population of almost 500 pilot whales 
was located off the Canary Island of Tenerife and were found to frequent an 
area of steep bathymetric slopes along the 1000 m depth contour. The social 
organisation of these whales, which travelled in 31 pods, supports 
hypotheses for a system in which related males and females live together in 
a social group and mate with associated groups. 
Comparative studies of the Delphinidae tested for family-wide trends 
in sexual dimorphism, relative brain size and relative size of the testes in 
order to further examine delphinid mating systems. While these dolphins 
and small whales fit into some of the predicted mammalian trends, 
especially relating to allometric growth and an inherent variability with 
absolute body size, there was a general lack of correlation of these 
morphological trends with ecological and social classifications. The 
classification system suffered from a general lack of information on the 
specifics of feeding ecology and social systems of thes~ animals, and the 
variables of group size, habitat and taxonomic classification of prey were 
likely too general. However, the findings do support the view that 
delphinids have relatively large brains (Jerison, 1973; 1986) and testes 
(Kenagy and Trombulak, 1986) when compared to other mammals. The 
192 
1, 
I 
implications of these views are that delphinids: 1) have complex 
requirements for neural information processing relating to food-finding in 
a patchy, three-dimensional environment or to the maintenance of diverse 
social relationships, and 2) have promiscuous mating systems in which 
paternity can not be assured and thus have developed mechanisms for 
sperm competition (Parker, 1984). 
A review of long-term studies on social organisations of the bottlenose 
dolphin and the killer whale identified two broadly different patterns of 
social interactions. While the small, monomorphic dolphins live in age 
and sex segregated groups which interact fluidly, in a form of fission-
fusion society, the large sexually-dimorphic killer whales and pilot whales 
live in stable groups of mixed age and sex. Bottlenose dolphins maintain 
genetic heterogeneity through a temporary interchange of males between 
female communities. However, there has been no dispersal of juvenile 
animals or adult males from the killer whale groups in 20 years of 
observation and this population has been described as "inbred" (Hoelzel, 
1991a; Hoelzel and Dover, 1990). I presented some original observations of 
these whales which provided the suggestion that mating was occurring 
between associated groups. 
Thus, while the small dolphins fit into the general theories of 
mammalian mating systems, following a "roving male" strategy (Clutton-
Brock, 1989; Whitehead, 1990) between adjacent female communities, the 
larger dolphins, like the killer whale and pilot whale, do not clearly match 
any previously described system for non-human mam.mals. However, a 
human social system, in which males maintain associations with their 
female kin while mating outside of this group, termed an avunculate 
(Alexander, 1977; Kurland, 1979; Flinn, 1981; Flinn and Low, 1986), could 
provide a possible explanation. 
193 
The development of hypotheses concerning delphinid mating systems 
has necessarily been based on limited data and must be considered a first 
step in establishing underlying principles involved in the social evolution of 
this diverse family. The main hypothesis has been that apparent 
differences in social organisation should be correlated with differences in 
morphology. In most studies of this sort on terrestrial mammals, the 
mating system is used as a categorising variable to examine morphological 
trends in sexual dimorphism, brains and testes. I was working at the 
problem from the other direction: I have examined morphological trends 
and used predictions, based on theories of the evolution of mating systems, 
to compare with the available observations of delphinid social organisation. 
As in any comparative study, the common correlation of two sets of 
variables does not imply causation; the possibility exists that a third 
variable is affecting both in a similar way. In this case, the third variable 
(broadly speaking) is likely to be ecological. Unfortunately, the 
classification of marine habitats is extremely difficult and there is much 
need for the synthesis of current knowledge of physical and biological 
oceanography with studies of these large marine predators. 
Addition;=i.lly, there is need for an expansion of the application of 
genetic methods for cetaceans. The current impetus for the utilisation of 
genetic techniques for describing the mating and resultant reproductive 
success of terrestrial mammals is even more critical for marine 
mammals, given the difficulties of observation. The few analyses of 
delphinid genetic heterogeneity have provided many of the suggestive 
results for the hypotheses of mating systems presented here. It is likely 
that the revolution started by the implementation of these methods will 
continue and the results will undoubtedly contribute still more to our 
understanding of all mammalian societies. 
194 
195 
REFERENCES 
Alagarswami, K., Bensam, P., Rajapandian, M.E. and Fernando, A.B. 
1973. Mass stranding of pilot whales in the Gulf of Manar. Ind. J. 
Fish. 20:269-279. 
Alexander, R.D. 1974. The evolution of social behavior. Ann. Rev. Ecol. 
Syst. 5:325-383. 
Alexander, R.D. 1977. Natural selection and the analysis of human 
sociality. pp. 283-337 IN: C. E. Goulden (ed). Changing Scenes in the 
Natural Sciences. Acad. Nat. Sci. Spee. Puhl. 12. 
Alexander, R.D., Hoogland, J.L., Howard, R.D., Noonan, K.M. and 
Sherman, P.W. 1979. Sexual dimorphism and breeding systems in 
pinnipeds, ungulates, primates and humans. pp. 402-435 IN: N. A. 
Chagnon and W. Irons (eds). Evolutionary Biology and Human 
Social Behavior. MA: Duxbury Press. 
Alling, A. 1986. Records of odontocetes in the northern Indian Ocean 
(1981-1982) and off the coast of Sri Lanka (1982-1984). J. Bombay Nat. 
Hist. Soc. 83:376-394. 
Altmann, J. 1974. Observational study of behaviour: sampling methods. 
Behaviour 49:227-267. 
Amos, B~, Barrett, J. and Dover, G.A. 1991a. Breeding behaviour of pilot 
whales revealed by DNA fingerprinting. Heredity 67:49-55. 
Amos, W., Barrett, J. and Dover, G.A. 1991b. Breeding system and social 
structure in ·the Faroese pilot whale as revealed by DNA 
fingerprinting. Rept. Int. Whal. Commn. Spee. Iss. 13:255-268. 
Amos, W., Schlotterer, C. and Tautz, D. In press . Analytical DNA 
profiling reveals a novel mammalian social system. Science. 
Amos, W., Whitehead, H ., Ferrari, M.J., Glockner-Ferrari, D.A., Payne, 
R. and Gordon, J. 1992. Restrictable DNA from sloughed cetacean 
skin; its potential for use in population analysis. Mar. Mamm. Sci. 
8:275-283. 
196 
Andelman, S. 1986. Ecological and social determinants of cercopithecine 
mating systems. pp. 201-216 IN: D. I. Rubenstein and R. W. 
Wrangham (eds). Ecological Aspects of Social Evolution. Princeton, 
N.J.: Princeton Univ. Press. 
Andersen, L.W. 1988. Electrophoretic differentiation among local 
populations of the long-finned pilot whale, Globicephala melaena, at 
the Faroe Islands. Can. J. Zool. 66:1884-1892. 
Armstrong, E. 1983. Relative brain size and metabolism in mammals. 
Science 220:1302-1304. 
Baird, R.W. 1992. Possible indirect interactions between transient and 
resident killer whales: implications for the evolution of foraging 
specialization in the genus Orcinus. Oecologia 89:125-132. 
Baird, R.W. and Stacey, P .J. 1988. Variation in saddle patch pigmentation 
in populations of killer whales (Orcinus area) from British Columbia, 
Washington, and Vancouver Island. Can. J. Zool. 66:2582-2585. 
Baird, R.W. and Stacey, P.J. 1989. Sightings, strandings and incidental 
catches of short-finned pilot whales, Globicephala macrorhynchus, 
off the coast of British Columbia. Int. Whal. Commn. Document 
SC/41/SM23. 
Baker, A.N. 1983. Whales and Dolphins of New Zealand and Australia. 
Victoria Univ. Press. 
Baker, C.S. and Herman, L.M. 1984. Aggressive behavior between 
humpback whales (Megaptera noveangliae) on the Hawaiian 
wintering grounds. Can. J. Zool. 62:1922-1937. 
Balcomb, K.C. and Bigg, M.A. 1986. Population biology of the three 
resident killer whale pods in Puget Sound and off southern 
Vancouver Island. IN: B. Kirkevold and . J. S. Lockard (eds). 
Behavioral Biology of Killer Whales. N.Y.: A.R. Liss. 
Balcomb, K.C., Boran, J.R. and Heimlich, S.L. 1982. Killer whales in 
Greater Puget Sound. Rept. Int. Whal. Commn. 32:681-685. 
Ballance, L.B. 1990. Home range of bottlenose dolphins. pp. 262-274 IN: S. 
Leatherwood and R.R. Reeves (eds). The Bottlenose Dolphin. 
197 
Barnes, L.G. 1990. The fossil record and evolutionary relationships of the 
genus Tursiops. pp. 3-26 IN: S. Leatherwood and R.R. Reeves (eds). 
The Bottlenose Dolphin. San Diego: Academic Press. 
Barros, N.B. and Cockroft, V.G. 1991. Prey of humpback dolphins (Sousa 
plumbea) stranded in eastern Cape Province, South Africa. Aquat. 
Mammal. 17:134-136. 
Barros, N.B. and Odell, D.K. 1990. Food habits of bottlenose dolphins 
(Tursiops truncatus) in the southeastern United States. pp. 309-328 
IN: S. Leatherwood and R. R. Reeves (eds). The Bottlenose Dolphin. 
San Diego, CA: Academic Press. 
Bartholomew, G.A. 1970. A model for the evolution of pinniped polygyny. 
Evolution 24:307-310. 
Bateman, A.J. 1948. Intra-sexual selection in Drosophila. Heredity 2:349-
368. 
Bateson, G. 1974. Observations of a cetacean community. pp. 146-165 IN: 
J. McIntyre (ed). Mind in the Waters. N .Y.: Scribner's. 
Bengtsson, B.0. 1978. Avoiding inbreeding: at what cost? J. Theor. Biol. 
73:439-444. 
Bertram, B.C.R. 1976. Kin selection in lions and in evolution. pp. 281-301 
IN: P. P. G. Bateson and R. A. Hinde (eds). Growing Points in 
Ethology. Cambridge Univ. Press. 
Bertram, B.C.R. 1978. Living in groups: predators and prey. pp. 281-301 
IN: J. R. Krebs and N. B. Davies (eds). Behavioural Ecology: An 
Evolutionary Approach. Oxford: Blackwell Scientific Publications. 
Bertram, B.C.R. 1979. Serengeti predators and their social systems. pp. 
221-248 IN: A. R. E . Sinclair and M. Norton-Griffiths (eds). 
Serengeti: Dynamics of an Ecosystem. Chicago: Univ. of Chicago 
Press. 
Berzin, A.A. and Vladimirov, V.L. 1983. A new species of killer whale 
(Cetacea, Delphinidae) from Antarctic waters. Zoologicheskii 
Zhurnal 62:287-295. 
198 
Best, P.B. 1979. Social organization 1n sperm whales, Physeter 
macrocephalus. pp. 227-289 IN: H. E. Winn and B. L. Olla (eds). 
Behavior of Marine Animals, Vol. 3: Cetaceans. N.Y.: Plenum 
Press. 
Bigg, M.A. 1982. An assessment of killer whale (Orcinus orca) stocks off 
Vancouver Island, British Columbia. Rep. Int. Whal. Commn. 
32:655-666. 
Bigg, M.A., Ellis, G.M., Ford, J.B. and Balcomb, K.C. 1987. Killer Whales: 
a study of their identification, genealogy and natural history in 
British Columbia and Washington state. Nanaimo, Canada: 
Phantom Press. 79 pp. 
Bigg, M.A., Olesiuk, P.F., Ellis, G.M., Ford, J.K.B. and Balcomb, K.C. 
1990. Social organization and genealogy of resident killer whales 
(Orcinus orca) in the coastal waters of British Columbia and 
Washington State. Rep. Int. Whal. Commn. Spee. Iss. 12:383-405. 
Blake, R.W. 1983. Energetics of leaping in dolphins and other aquatic 
animals. J. Mar. Biol. Ass. U.K. 63:61-70. 
Boje, R. and Tomczack, M. 1978. Upwelling Ecosystems. Berlin: Springer-
Verlag. 
Bonin, G.v. 1973. About quantitative studies on the cerebral cortex. J. 
Microscopy 99:75-83. 
Bonner, J.T. 1980. The Evolution of Culture in Animals. N.J.: Princeton 
Univ. Press. 
Bonner, J.T. 1988. The Evolution of Complexity. N.J.: Princeton Univ. 
Press. 
Borgerhoff Mulder, M. 1991. Human behavioural ecology. pp. 69-98 IN: J. 
R. Krebs and N. B. Davies (eds). Behavioural Ecology: An 
Evolutionary Approach, 3rd edition. Oxford: Blackwell Scientific 
Puhl. 
Bowers, C.A. and Hendersen, R.S. 1972. Project Deep Ops: Deep Object 
Recovery with Pilot and Killer Whales. Naval Undersea Center 
Report NUC-TP-306. 
199 
Breathnach, A.S. 1960. The cetacean central nervous system. Biol. Rev. 
35:187-230. 
Bree, P.J.H.v. 1971. On Globicephala seiboldii Gray, 1846, and other 
species of pilot whales. Beaufortia 19:79-87. 
Bree, P.J.H.v., Best, P.B. and Ross, G.J.B. · 1978. Occurrence of the two 
species of pilot whales (genus Globicephala) on the coast of South 
Africa. Mammalia 42:323-328. 
Brodie, P.F. 1969. Duration of lactation in cetaceans: an indicator of 
required learning. Am. Midl. Nat. 82:312-314. 
Brodie, P.F. 1975. Cetacean energetics, an overview of intraspecific size 
variation. Ecology 56:152-161. 
Brown, D.H. 1960. Behavior of a captive Pacific pilot whale. J. Mammal. 
41:342-349. 
Brown, D.H. and Norris, K.S. 1956. Observations on captive and wild 
cetaceans. J. Mammal. 37:311-326. 
Brown, J.H. and Lasiewski, R.C. 1972. Metabolism of weasels: the costs of 
being long and thin. Ecology 53:939-943. 
Brown, S.G. 1961. Observations of pilot whales (Globicephala) in the north 
Atlantic Ocean. Norsk Hvalfangsttid 50:225-254. 
Brownell, R.L.J. 1974. Small odontocetes of the Antarctic. pp. 13-19 IN: S. 
G. Brown, R. L. J. Brownell, A. W. Erickson, R. J. Hofman, G. A. 
Llano and N. A. Mackintosh (eds). Antarctic Map Folio Series: 
Antarctic Mammals, Folio 18. Wash., D.C.: American Geographical 
Society. 
Bryden, M.M. 1972. Growth and development of marine mammals. pp. 1-
79 IN: R. J. Harrison (ed). Functional Anatomy of Marine 
Mammals. London: Academic Press. 
Burke, T. 1989. DNA fingerprinting and other methods for the study of 
mating success. Trends Ecol. Evol. 4:139-144. 
200 
Bygott, J.D., Bertram, B.C.R. and Hanby, J.P. 1979. Male lions in large 
coalitions gain reproductive advantage. Nature 282:839-841. 
Cadenat, J . 1957. Observations de cetaces, sireniens, cheloniens et 
sauriens en 1955-1956. Bull. Inst. Francais d'Afrique Noire Ser. 
19A:1358-1375. 
Cairns, S.J. and Schwager, S.J. 1987. A comparison of association indices. 
Anim. Behav. 35:1454-1469. 
Caldecott, J.O. 1986. Mating patterns, societies and the ecogeography of 
macaques. Anim. Behav. 34:208-220. 
Caldwell, D.K. and Caldwell, M.C. 1975. Dolphin and small whale 
fisheries of the Caribbean and West Indies: occurrence, history, and 
catch statistics-with special reference to the Lesser Antillean island 
of St. Vincent. J. Fish. Res. Bd. Can. 32:1105-1110. 
Caldwell, D.K., Caldwell, M.C., Rathjen, W.F. and Sullivan, J.R. 1971. 
Cetaceans from the lesser antillean island of St. Vincent. Fish. Bull. 
69:303-312. 
Caro, T.M. 1987. Cheetah mothers' vigilance: looking out for prey or for 
predators? Behav. Ecol. Sociobiol. 20:351-361. 
Caro, T.M. 1989. Missing links in predator and anti-predator behaviour. 
TREE 4:333-334. 
Caro, T.M. and Hauser, M.D. 1992. Is there teaching in non-human 
animals? Q. Rev. Biol. 67:151-174. 
Casinos, A. 1981. On the distribution of Globicephala melaena (Traill, 
l804)(Cetacea, Delphinidae) in the south-west Atlantic. Zeitschrift 
fur Saugetierkunde 46:268-271. 
Casinos, A. and Bou, J. 1980. On a massive stranding ·of short-finned pilot 
whale, Globicephala macrorhynchus Gray, 1846, on Margarita 
Island (Venezuela). Sci. Rept. Whales Res. Inst. 32:145-148. 
Casinos, A. and Vericad, J.R. 1976. The cetaceans of the Spanish coasts: a 
survey. Mammalia 40:267-289. 
201 
Cheney, D.L. and Seyfarth, R.M. 1983. Non-random dispersal in free-
ranging vervet monkeys: social and genetic consequences. Am. Nat. 
122:392-412. 
Chepko-Sade, B.D., Shields, W.M., Berger, J., Halpin, Z., Jones, W.T., 
Rogers, L.T., Rood, J.P. and Smith, A.P. 1987. The effects of 
dispersal and social structure on effective population size. IN: B. D. 
Chepko-Sade and Z. T. Halpin (eds). Mammalian Dispersal 
Patterns. Univ. of Chicago Press. 
Christensen, I. 1984. Growth and reproduction of killer whales, Orcinus 
ore a, in Norwegian coastal waters. Rep. Int. Whal. Commn. Spee. 
Iss. 6:253-258. 
Clarke, M.R. 1986. Cephalopods in the diet of odontocetes. pp. 281-322 IN: 
M. M. Bryden and R. J . Harrison (eds). Research on Dolphins. 
Oxford: Clarendon Press. 
Clarke, R. 1981. Whales and dolphins of the Azores and their exploitation. 
Rep. Int. Whal. Commn. 31:607-615. 
Clutton-Brock, T.H. 1980. Primates, brains and ecology. J. Zool. Lond. 
190:309-323. 
Clutton-Brock, T.H. 1982. The functions of antlers. Anim. Behav. 79:108-
125. 
Clutton-Brock, T.H. 1984. Reproductive effort and terminal investment in 
iteroparous animals. Am. Nat. 123:212-229. 
Clutton-Brock, T.H. 1986. Reproductive Success. Univ. of Chicago Press. 
Clutton-Brock, T.H. 1989a. Female transfer and inbreeding avoidance in 
social mammals. Nature 337:70-72. 
Clutton-Brock, T.H. 1989b. Mammalian mating systems. Proc. R. Soc. 
Lond. B 236:339-372. 
Clutton-Brock, T.H., Albon, S.D., Gibson, R.M. and Guiness, F.E. 1979. 
The logical stag: · adaptive aspects of fighting in red deer (Cervus 
elephos L.). Anim. Behav. 27:211-225. 
202 
Clutton-Brock, T.H., Guinness, F.E. and Albon, S.D. 1982. Red Deer: 
Behavior and Ecology of Two Sexes. Chicago, IL: University of 
Chicago Press. 
Clutton-Brock, T.H. and Harvey, P.H. 1977. Primate ecology and social 
organisation. J. Zool. Lond. 1-39. 
Coe, M.J. 1967. "Necking" behaviour in the giraffe. J. Zool. 151:313-321. 
Cole, L.C. 1949. The measurement of interspecific association. Ecol 30:411-
424. 
Condy, P.R., Van Aarde, R.J. and Bester, M.N. 1978. The seasonal 
occurrence and behaviour of killer whales, Orcinus area, at Marion 
Island. J. Zool. Lond. 184:449-464. 
Connor, R.C. and Norris, K.S. 1982. Are dolphins reciprocal altruists? 
Am. Nat. 119:358-374. 
Connor, R.C. and Smolker, R.A. 1985. Habituated dolphins (Tursiops sp.) 
in Western Australia. J. Mamm. 66:398-400. 
Connor, R.C., Smolker, R.A. and Richards, A.F. 1992a. Dolphin alliances 
and coalitions. pp. 415-443 IN: A. H. Harcourt and F. B. M. De Waal 
(eds). Coalitions and Alliances in Humans and Other Animals. 
Oxford Univ. Press. 
Connor, R.C., Smolker, R.A. and Richards, A.F. 1992b. Two levels of 
alliance formation among male bottlenose dolphins 
(Tursiops sp.). Proc. Natl. Acad. Sci. USA 89:987-990. 
Cope, E.D. 1876. Fourth contribution to the history of the existing Cetacea. 
Proc. Nat. Acad. Sci. Phila. 1876:127-144. 
Corkeron, P.J., Morris, R.J. and Bryden, M.M. 1987. Interactions between 
bottlenose dolphins and sharks in Moreton Bay, Australia. Aquat. 
Mamm. 13:109-113. 
Crook, J.H. 1970. Social organization and environment: Aspects of 
contemporary social ethology. Anim. Behav. 18:197-209. 
203 
I I 
Crook, J.H. 1972. Sexual selection, dimorphism and social organization in 
the primates. pp. 231-281 IN: B. Campbell (ed). Sexual Selection and 
the Descent of Man, 1871-1971. Chicago, IL: Aldine. 
Crook, J.H., Ellis, J.E. and Goss-Custard, J.D. 1976. Mammalian social 
systems: structure and function. Anim. Behav. 24:261-274. 
Crook, J.H. and Gartlan, J.S. 1966. The evolution of primate societies. 
Nature 210:1200-1203. 
Dammerman, K.W. 1924. On Globicephala and some other Delphinidae 
from the Indo-Australian archipelago. Treubia 5:340. 
Darwin, C. 1871. The Descent of Man and Selection in Relation to Sex. 
London: John Murray. 
Davies, E.M. and Boersma, P.D. 1984. Why lionesses copulate with more 
than one male. Am. Nat. 123:594-611. 
Davies, J.L. 1960. The southern form of the pilot whale. J. Mamm. 41:29-
34. 
Davies, J.L. 1963. The antitropical factor in cetacean evolution. Evolution 
17:107-116. 
Davies, N.B. 1991. Mating systems. pp. 263-294 IN: J. R. Krebs and N. B. 
Davies (eds). Behavioural Ecology: An Evolutionary Approach, 3rd 
edition. Oxford: Blackwell Scientific Puhl. 
Davies, N.B. 1992. Dunnock Behaviour and Social Evolution. Oxford Univ. 
Press. 
Dawkins. 1976. The Selfish Gene. Oxford: Oxford University Press. 
de Waal, F .B.M. 1982. Chimpanzee Politics: Power and Sex Among Apes. 
N.Y.: Harper & Row. 
Deag, J .M. 1980. Interactions between males and unweaned Barbary 
macaques: Testing the agonistic buffering hypothesis. Behaviour 
75:54-81. 
204 
Defran, R.H. and Pryor, K. 1980. The behavior and training of cetaceans in 
captivity. IN: L. Herman (ed). Cetacean Behavior: Mechanisms and 
Functions. N.Y.: J.Wiley-Interscience. 
Dice, L.R. 1945. Measures of the amount of ecologic association between 
species. Ecology 26:297-302. 
Douglas-Hamilton, I. and Douglas-Hamilton, 0. 1975. Among the 
Elephants. N.Y.: Viking. 
Dublin, H.T. 1983. Cooperation and reproductive competition among 
female African elephants. pp. 291-313 IN: S. K. Wasser (ed). Social 
Behavior of Female Vertebrates. N.Y.: Academic Press . 
Duffield, D.A. 1986. Orcinus orca: taxonomy, evolution, cytogenetics and 
population structure. pp. 19-33 IN: B. Kirkevold and J. S. Lockard 
(eds). Behavioral Biology of Killer Whales. N.Y.: A.R. Liss. 
Duffield, D.A. and Wells, R.S. 1991. The combined application of 
chromosome, protein and molecular data for the investigation of 
social unit structure and dynamics in Tursiops truncatus. Rep. Int. 
Whal. Commn. Spee. Iss. 13:155-169. 
Duguy, R. 1968. Note sur Globicephala macrorhynchus Gray, 1846, un 
cetace nouveau pour les cotes de France. Mammalia 32:113-117. 
Dunbar, R.I.M. 1992. Neocortex size as a constraint on group size 1n 
primates. J. Hum. Evol. 22:469-493. 
Economos, A.C. 1983. Elastic and/or geometric similarity in mammalian 
design. J . Theor. Biol. 103:167-172. 
Eisenberg, J.F. 1981. The Mammalian Radiations. London: The Athlone 
Press. 
Eisenberg, J.F. and Wilson, D.E . 1978. Relative brain size and feeding 
strategies in the Chiroptera. Evolution 32:740-751. 
Elias, H . and Schwartz, D. 1969. Surface areas of the cerebr al cortex of 
mammals determined by sterological methods. Science 166:111-113. 
'1 
, I 
I 
Emlen, S.T. and Oring, L.W. 1977. Ecology, sexual selection, and the 
evolution of mating systems. Science 197:215-223. 
Estes, R.D. and Godard, J. 1967. Prey selection and hunting behavior of the 
African wild dog. J. Wildl. Manag. 31:52-70. 
Evans, P.G.H. 1980. Cetaceans in British waters. Mamm. Rev. 10:1-52. 
Evans, P.G.H. 1987. The Natural History of Whales and Dolphins. N.Y.: 
Facts on File Puhl. 
Evans, W.E. 1971. Orientation behavior of delphinids: radio-telemetric 
studies. Ann. N.Y. Acad. Sci. 188:142-160. 
Evans, W.E. 1973. Echolocation by marine delphinids and one species of 
fresh-water dolphin. J. Acoust. Soc. Am. 54:191-199. 
Evans, W.E. and Bastian, J. 1969. Marine mammal communication: 
social and ecological factors. pp. 425-475 IN: H. T. Andersen (ed). 
The Biology of Marine Mammals. N.Y.: Academic Press. 
Fehring, W.K. and Wells, R.S. 1976. A series of strandings by a single 
herd of pilot whales on the west coast of Florida. J. Mamm. 57:191-
194. 
Felleman, F.L., Heimlich-Boran, J.R. and Osborne, R.W. 1991. Feeding 
ecology of the killer whale (Orcinus orca). IN: K. Pryor and K. S. 
Norris (eds). Dolphin Societies. Berkeley: Univ. of Calif. Press. 
Fernandopulle, D. 1976. Climatic characteristics of the Canary Islands. 
pp. 185-206 IN: G. Kunkel (ed). Biogeography and Ecology in the 
Canary Islands. The Hague: Dr. W Junk b.v. Publishers. 
Fish, F.E. and Hui, C.A. 1991. Dolphin swimming-a review. Mammal. 
Rev. 181-195. 
Fisher, R.A. 1930. The Genetical Theory of Natural Selection. Oxford 
Univ. Press. 
Fitch, J.E. and Brownell, R.L. 1965. Fish otoliths in cetacean stomachs. J. 
Fish. Res. Bd. Can. 25:2561-2575. 
Flinn, M. 1981. Uterine vs. agnatic kinship variability and associated 
cousin marriage preferences: an evolutionary biological analysis. 
IN: R. D. Alexander and D. W. Tinkle (eds). Natural Selection and 
Social Behavior: Recent Research and New Theory. N.Y.: Chiron 
Press . 
Flinn, M.V. and Low, B.S. 1986. Resource distribution, social competition, 
and mating patterns in human societies. pp. 217-243 IN: D. I. 
Rubenstein and R. W. Wrangham (eds). Ecological Aspects of Social 
Evolution: Birds and Mammals . Princeton Univ. Press. 
Ford, J.K.B. 1984. Call Traditions and Dialects of Killer Whales (Orcinus 
orca) in British Columbia. Ph.D. Thesis. Univ. of British Columbia. 
Ford, J.K.B. 1989. Acoustic behavior of resident killer whales (Orcinus 
orca) off Vancouver Island, British Columbia. Can. J. Zool. 67:727-
745. 
Ford, J.K.B. and Fisher, H.D. 1983. Group-specific dialects of killer whales 
(Orcinus orca) in British Columbia. pp. 129-161 IN: R. S. Payne (ed). 
Communication and Behavior of Whales . Boulder, CO: Westview 
Press. 
Fowler, C. 1984. Density dependence in cetacean populations. Rept. Int. 
Whal. Commn. Special Iss. 6:373-379. 
Frame, L.H. and Frame, G.W. 1976. Female wild dogs emigrate. Nature 
263:227-229. 
Frame, L.H., Malcolm, J .R., Frame, G.W. and van Lawick, H. 1979. 
Social organization of African wild dogs (Lycaon pictus) on the 
Serengeti plains, Tanzania 1967-1978. Z. Tierpsychol. 50:225-249. 
Fraser, F.C . 1950. Two skulls of Globicephala macrorhynchus from 
Dakar. Atlantide Rep. 1:49-60. 
Fraser, F.C. 1966. Comments on the Delphinoidea. pp. 7-31 IN: K. S. 
Norris (ed). Whales, Dolphins and Porpoises. Berkeley: Univ. Calif. 
Press. 
Fraser, F.C. 1974. Report on Cetacea stranded on the British coasts from 
1948 to 1966. Br. Mus. (Nat. Hist.) Puhl. No. 718:1-65. 
'2JJ7 
Fraser, F.C. and Purves, P.E. 1960. Hearing in cetaceans. Bull. Brit. Mus. 
Nat. Hist. Zool. 2:103-116. 
Gabriel, W. and Thomas, B. 1988. Vertical migration of zooplankton as an 
evolutionary stable strategy. Am. Nat. 132:199-216. 
Gales, R., Pemberton, D., Clarke, M. and Lu, C.C. 1992. Stomach contents 
of long-finned pilot whales (Globicephala melas) and bottlenose 
dolphins (Tursiops truncatus) in Tasmania. Mar. Mamm. Sci. 
8:405-413. 
Gaskin, D.E. 1968a. Distribution of Delphinidae (Cetacea) in relation to sea 
surface temperatures off eastern and southern New Zealand. N. Z. 
J. Mar. Freshw. Res. 2:527-534. 
Gaskin, D.E. 1968b. The New Zealand Cetacea. N.Z. Mar. Fish. Res. Bull. 
1:1-92. 
Gaskin, D.E. 1982. The Ecology of Whales and Dolphins. London: 
Heinemann. 
Gaulin, S.J.C. and Sailer, L. 1985. Are females the ecological sex? Am. 
Anthropol. 87:111-119. 
Geist, V. 1971. Mountain Sheep. Chicago Univ. Press. 
Geraci, J.R. and St. Aubin, D.J. 1977. Mass stranding of the long-finned 
pilot whale, Globicephala melaena, on Sable Island, Nova Scotia. J. 
Fish. Res. Bd. Can. 34:2196-2199. 
Gibson-Lonsdale, J.J. 1990. Pilot whaling in the Faroe Islands-its history 
and present significance. Mammal. Rev. 20:44-52. 
Gingerich, P.D., Wells, N .A., Russel , D.E . and Ibrahim Shah, S.M. 1983. 
Origin of whales in epicontinental remnant seas: new evidence from 
the early Eocene of Pakistan. Science 220:403-406. · 
Ginsburg, J.R. and Young, T.P. 1992 . . Measuring association between 
individuals or groups in behavioural studies. Animal Behaviour 
44:377-379. 
208 
, [ 
Gittleman, J.L. 1985. Communal care in mammals. pp. 187-205 IN: P. J. 
Greenwood, P.H. Harvey and M. Slatkin (eds). Evolution: Essays in 
Honour of John Maynard Smith. Cambridge University Press. 
Gittleman, J.L. 1986. Carnivore brain size, behavioral ecology, and 
phylogeny. J. Mammal. 67:23-36. 
Gittleman, J.L. 1989. Carnivore group living: comparative trends. pp. 183-
207 IN: J. L. Gittleman (ed). Carnivore Behavior, Ecology, and 
Evolution. London: Chapman and Hill. 
Goodall, J. 1986. The Chimpanzees of Gombe: Patterns of Behavior. 
Cambridge, MA: Belknap Press. 
Gordon, J.C.D. 1987. The Behaviour and Ecology of Sperm Whales off Sri 
Lanka. Ph.D. Thesis. University of Cambridge. 
Gould, S.J. 1966. Allometry and size in ontogeny and phylogeny. Biol. Rev. 
Camb. Phil. Soc. 41:587-640. 
Gouzoules, S. 1984. Primate mating systems, kin associations, and 
cooperative behavior: Evidence for kin recognition? Ybk. Phys. 
Anthropol. 27:99-134. 
Gouzoules, S. and Gouzoules, H. 1987. Kinship. pp. 299-305 IN: B. B. 
Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham and T. T. 
Struhsaker (eds). Primate Societies. Chicago Univ. Press. 
Gray, J .E. 1846. On the cetaceaous animals. pp. 13-53 IN: J. Richardson 
and J. E. Gray (ed). The zoology of the voyage of HMS Erebus and 
Terror. London: 
Gray, J .E. 1871. Supplement to the Catalogue of Seals and Whales in the 
British Museum. London: Trustees, Brit. Mus. 103 pp. 
Greenwood, P.J. 1980. Mating systems, philopatry, and dispersal. Anim. 
Behav. 28:1140-1162. 
Greenwood, P.J. 1983. Mating systems and the evolutionary consequences 
of dispersal. pp. 116-131 IN: I. R. Swingland and P. J. Greenwood 
(eds). The Ecology of Animal Movement. Oxford: Calrendon Press. 
209 
Gretarsdottir, S. and Arnason, U. 1992. Evolution of the common cetacean highly repetitive DNA 
component and the systematic position of Orcaella brevirostris. J. Mol. Evol. 34:201-208. 
Greenwood, P.J. and Wheeler, P. 1985. The evolution of sexual size 
dimorphism in birds and mammals: a "hot-blooded" hypothesis. pp. 
287-300 IN: P. J. Greenwood, P. H. Harvey and M. Slatkin (eds). 
Evolution: Essays in Honour of John Maynard Smith. Cambridge 
Univ. Press. 
Guinet, C. 1991. Intentional stranding apprenticeship and social play in 
killer whales (Orcinus area). Can. J. Zool. 69:2712-2716. 
Guinness, F.A., Hall, M.J. and Cockerill, R.A. 1979. Mother-offspring 
association in red deer. Animal Behaviour 27:536-544. 
Haenel, N.J. 1986. General notes on the behavioral ontogeny of Puget 
Sound killer whales and the occurrence of allomaternal behavior. 
pp. 285-300 IN: B. Kirkevold and J. S. Lockard (eds). Behavioral 
Biology of Killer Whales. N.Y.: A.R. Liss. 
Hall, M.J. 1983. Social organization in an enclosed group of red deer 
(Cervus elaphus) on Rhum. II. Social grooming, mounting behavior, 
spatial organization and their relationship to dominance rank. Z. 
Tierpsychol. 61:273-292. 
Hamilton, W.D. 1964. The genetical theory of social behaviour. J.Theoret. 
Biol. 7:1-52. 
Hamilton, W.D. 1971. Geometry for the selfish herd. J. Theor. Biol. 31:295-
311. 
Hammond, P .S. 1986. Estimating the size of naturally marked whale 
populations using capture-recapture techniques. Rept. Int. Whal. 
Commn. Spee. Iss. 8:253-282. 
Hampton, I.F.G. and Whittow, G.C. 1976. Body temperature and heat 
exchange in the Hawaiian spinner dolphin, Stenella longirostris. 
Comp. Biochem. Physiol. 55A: 195-197. 
Hanby, J.P. 1974. Male-male mounting in Japanese monkeys (Macaca 
fuscata). Anim. Behav. 22:836-849. 
Hanby, J.P. 1976. Sociosexual development in primates. pp. 1-67 IN: P. P. 
G. Bateson and P. Klopfer (eds). Perspectives in Ethology. N.Y.: 
Plenum Press. 
210 
Hanby, J.P. and Brown, C.E. 1974. The development of sociosexual 
behaviors in Japanese macaques. Folia Primatol. 16:123-143. 
Harcourt, A.H., Harvey, P.H., Larson, S.G. and Short, R.V. 1981. Testis 
weight, body weight, and breeding system in primates. Nature 
293:55-57. 
Harcourt, A.H. and Stewart, K.J. 1983. Interactions, relationships, and 
social structure: The great apes. pp. 307-314 IN: R. A. Hinde (ed). 
Primate Social Relationships. Sunderland, MA: Sinauer Assoc. 
Harrison, R.J. 1949. Observations on the female reproductive organs of the 
Ca'aing whale Globicephala melaena (Traill). J. Anat. 83:238-253. 
Harrison, R.J. 1969. Reproduction and reproductive organs. pp. 253-348 
IN: H. T. Andersen (ed). The Biology of Marine Mammals. N.Y.: 
Academic Press. 
Harvey, P.H. and Bennett, P.M. 1983. Brain size, energetics, ecology, and 
life history patterns. Nature 306:314-316. 
Harvey, P.H. and Clutton-Brock, T.H. 1985. Life history variation in 
primates. Evolution 39:559-581. 
Harvey, P.H., Clutton-Brock, T.H. and Mace, G.M. 1980. Brain size and 
ecology in small mammals and primates. Proc. Natl. Acad. Sci. 
77:4387-4389. 
Harvey, P.H. and Harcourt, A.H. 1984. Sperm competition, testes size, and 
breeding systems in primates. pp. 589-600 IN: R. L. Smith (ed). 
Sperm Competition and the Evolution of Animal Mating Systems. 
N.Y.: Academic Press, Inc. 
Harvey, P.H., Kavanaugh, M. and Clutton-Brock, T.H. 1978. Sexual 
dimorphism in primate teeth. J. Zool. Lond. 186:475-485. 
Harvey, P.H. and Pagel, M. 1991. The Comparative Method 1n 
Evolutionary Biology. Oxford University Press. 239 pp. 
Harvey, P .H. and Ralls, K. 1986. Do animals avoid incest? Nature 320:575-
576. 
211 
Heimlich-Boran, J.R. 1986a. Fishery correlations with the occurrence of 
killer whales in greater Puget Sound. pp. 113-131 IN: B. C. Kirkevold 
and J. S. Lockard (eds). Behavioral Biology of Killer Whales. NY: 
AR Liss. 
Heimlich-Boran, J.R. 1986b. Photogrammetric analyis of growth in Puget 
Sound Orcinus orca. pp. 97-111 IN: B. C. Kirkevold and J. S. Lockard 
(eds). Behavioral Biology of Killer Whales. New York: A.R. Liss. 
Heimlich-Boran, J.R. 1988. Behavioral ecology of killer whales (Orcinus 
orca) in the Pacific Northwest. Can J Zool 66:565-578. 
Heimlich-Boran, J .R. and Heimlich-Boran, S.L. ms. Socio-sexual 
interactions of killer whales: implications for a mating system. 
Heimlich-Boran, S.L. 1986. Cohesive relationships among Puget Sound 
killer whales. pp. 251-284 IN: B. Kirkevold and J. S. Lockard (eds). 
Behavioral Biology of Killer Whales. New York: A.R. Liss. 
Heimlich-Boran, S.L. 1988. Association Patterns and Social Dynamics of 
Killer Whales (Orcinus orca) in Greater Puget Sound. M.A. Thesis. 
Moss Landing Marine Laboratories, San Jose State University. 
Herman, L.M. 1991. What the dolphin knows, or might know, in its 
natural world. pp. 349-363 IN: K. Pryor and K. S. Norris (eds). 
Dolphin Societies: Discoveries and Puzzles. Berkeley: Univ. Calif. 
Press. 
Herman, L.M. and Tavolga, W.N. 1980. Communciation systems of 
cetaceans. pp. 149-210 IN: Cetacean Behavior: Mechanisms and 
Functions. N.Y.: John Wiley Interscience. 
Hersh, S.L., Odell, D.K. and Asper, E .D. 1990. Sexual dimorphism in 
bottlenose dolphins from the east coast of Florida. Mar. Mamm. Sci. 
6:305-315. 
Hershkovitz, P. 1966. Catalog of living whales. Bull. U.S. Nat. Mus. 246:1-
259. 
Herve-Gruyer, H. 1989. Sightings and behavior of cetaceans of Canaria, 
March 1988. 3rd Ann Conf of the European Cetacean Society, La 
Rochelle, France. 
212 
Herve-Gruyere, C. 1990. Les Enfants Dauphins. Paris: Gallimard. 136 pp. 
pp. 
Hoelzel, A.R. 1991a. Analysis of regional mitochondrial DNA variation in 
the killer whale; implications for cetacean conservation. Rep. Int. 
Whal. Commn. Spee. Iss. 13:224-233. 
Hoelzel, A.R. (ed.) 1991b. The Genetic Ecology of Whales and Dolphins. 
Cambridge: Black Bear Press. 
Hoelzel, A.R. 1991c. Killer whale predation on marine mammals at Punta 
Norte, Argentina; food sharing, provisioning and foraging strategy. 
Behav. Ecol. Sociobiol. 129:1-8. 
Hoelzel, A.R. and Dover, G.A. 1990. Genetic differentiation between 
sympatric Killer whale populations. Heredity 66:191-195. 
Hoelzel, A.R. and Osborne, R.W. 1986. Killer whale call characteristics; 
implications for cooperative foraging strategies. pp. 373-406 IN: B. 
Kirkevold and J . S. Lockard (eds). Behavioral Biology of Killer 
Whales. N.Y.: A.R. Liss. 
Hohn, A. 1980. Age determination and age related factors in the teeth of 
western North Atlantic bottlenose dolphins. Sci. Rept. Whales Res. 
Inst. 32:39-66. 
Hohn, A., Scott, M.D., Wells, R.S., Sweeney, J.C. and Irvine, A.B. 1989. 
Growth layers in teeth from known-age, free-ranging bottlenose 
dolphins. Mar. Mamm. Sci. 5:315-342. 
Hohn, A.A., Barlow, J. and Chivers, S.J. 1985. Reproductive maturity and 
seasonality in male spotted dolphins in the eastern tropical Pacific. 
Mar. Mamm. Sci. · 1:273-293. 
Hoogland, J.L. 1982. Prairie dogs avoid extreme inbreeding. Science 
215:1639-1641. 
Hoyt, E . 1984. Orea: The Whale Called Killer. Ontario: Camden House. 
Hrdy, S.B. 1977. The Langurs of Abu. Cambridge, MA: Harvard Univ. 
Press. 
213 
Hrdy, S.B. and Whitten, P.L. 1987. Patterning of sexual activity. pp. 370-
384 IN: B. B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham 
and T. T. Struhsaker (eds). Primate Societies. Chicago: Chicago 
University Press. 
Hui, C. 1979. Undersea topography and distribution of dolphins of the 
genus Delphinus in the southern California bight. J . Mammal. 
60:521-527. 
Humphrey, N.K. 1976. The social function of intellect. pp. 303-317 IN: P. 
Bateson and R. A. Hinde (eds). Growing Points in Ethology. 
Cambridge Univ. Press. 
Imbrie, J. 1956. Biometrical methods in the study of invertebrate fossils. 
Bull. Am. Mus. Nat. Hist. 108:211-252. 
Irvine, A.B., Scott, M.D., Wells, R.S. and Kaufmann, J.H. 1981. 
Movements and activities of the Atlantic bottlenose dolphin, Tursiops 
truncatus, near Sarasota, Florida. Fish. Bull. 79:671-688. 
Irvine, A.B. and Wells, R.S. 1972. Results of an attempt to tag Atlantic 
bottlenose dolphins, Tursiops truncatus. Cetology 13:1-5. 
Irvine, A.B., Wells, R.S. and Scott, M.D. 1982. An evaluation of tagging 
techniques for tagging small odontocete cetaceans. Fish. Bull. 
80:135-143. 
Janis, C.M; and Carrano, M. 1992. Scaling of reproductive turnover in 
archosaurs and mammals: why are large terrestrial mammals so 
rare? Ann. Zool. Fennici. 28:201-216. 
Jansen, J. and Jansen, J.K.S. 1969. The nervous system of Cetacea . pp. 
175-252 IN: H. T: Andersen (eds). The Biology of Marine Mammals. 
N .Y. : Academic Press. 
Jarman, P.J. 1974. The social organisation of antelope in relation to their 
ecology. Behaviour 48:215-267. 
Jarman, P.J. and Jarman, M.V. 1979. The dynamics of ungulate social 
organization. pp. 185-220 IN: A. R. E. Sinclair and M. Norton-
Griffiths (eds). Serengeti: Dynamics of an Ecosystem. Univ. of 
Chicago Press. 
214 
Jerison, H.J. 1973. The Evolution of the Brain and Intelligence. N.Y.: 
Academic Press. 
Jerison, H.J. 1986. The perceptual world of dolphins. pp. 141-166 IN: R. J. 
Schusterman, J. A. Thomas and F. G. Wood (eds). Dolphin 
Cognition and Behavior: a Comparative Approach. Hillsdale, N.J.: 
Erlbaum Assoc. 
Kasuya, T. 1975. Past occurrence of Globicephala melaena in the western 
North Pacific. Sci. Rep. Whales Res. Inst. 27:95-110. 
Kasuya, T. and Marsh, H. 1984. Life history and reproductive biology of the 
short-finned pilot whale, Globicephala macrorhynchus, off the 
Pacific coast of Japan. Rept. Int. Whal. Commn. Spee. Iss. 6:259-310. 
Kasuya, T. and Marsh, H. 1989. Prolonged suckling in short-finned pilot 
whales off Japan. 5th Intl Theriological Congress, Rome, Italy. 
Kasuya, T. and Matsui, S. 1984. Age determination and growth of the 
short-finned pilot whale off the pacific coast of Japan. Sci. Rep. 
Whales. Res. Inst. 35:57-91. 
Kasuya, T., Mayashita, T. and Kasamatsu, F. 1988a. Segregation of two 
forms of short-finned pilot whales off the Pacific coast of Japan. Sci. 
Rep. Whales Res. Inst. 39:77-90. 
Kasuya, T., Sergeant, D.E. and Tanaka, K. 1988b. Re-examination of life 
history parameters of long-finned pilot whales in the Newfoundland 
waters. Scientific Reports of the Whales Research Institute 39:103-
119. 
Kaufman, L. and Rousseeuw, P.J. 1990. Finding Groups in Data: An 
Introduction to Cluster Analysis. NY: Wiley & Sons. 
Kaufmann, J.H. 1974. Social behavior of the whiptail wallaby, Macropus 
parryi, in northeastern New South Wales. Anim. Behav, 22:281-369. 
Kawamura, S. 1967. Aggression as studied in troops of Japanese monkeys. 
pp. 195-233 IN: C. D. Clemente and D. B. Lindsley (eds). Aggression 
and Defense: Neural Mechanisms and Social Patterns. Berkeley, 
CA: Univ Calif. Press. 
215 
Kenagy, G.J. and Trombulak, S.C. 1986. Size and function of mammalian 
testes in relation to body size. J. Mammal. 67:1-22. 
Kirby, V.L. and Ridgway, S.H. 1984. Hormonal evidence of spontaneous 
ovulation in captive dolphins, Tursiops truncatus and Delphinus 
delphis. Rep. Int. Whal. Commn. Spee. Iss. 6:459-464. 
Kleiber, M. 1961. The Fire of Life. N.Y.: Wiley and Sons. 
Kleiman, D.G. 1977. Monogamy in mammals. Q. Rev. Biol. 52:39-69. 
Klemm, W.R., Sherry, C.J., Schake, L.M. and Sis, R.F. 1983. Homosexual 
behavior in feedlot steers: an aggression hypothesis. Appl. Anim. 
Ethol. 11:187-195. 
Klimley, A.P. and Brown, S.T. 1983. Stereophotography for the field 
biologist: measurement of lengths and three-dimensional positions of 
free-swimming sharks. Mar. Biol. 74:175-185. 
Klinowska, M. 1986. The cetacean magnetic sense-evidence from 
strandings. pp. 401-432 IN: M. M. Bryden and R. J. Harrison (eds). 
Research on Dolphins. Oxford: Clarendon Press. 
Klinowska, M. 1991. Dolphins, Porpoises and Whales of the World: the 
IUCN Red Data Book. Gland, Switzerland: IUCN. 429 pp. 
Kock, L.L.d. 1956. The pilot whale stranding on the Orkney Island of 
Westray, 1955. Scottish Nat. 68:63-70. 
Kritzler, H. 1952. Observations on the pilot whale in captivity. J. Mammal. 
33:321-334. 
Kruse, S.L. 1989. Aspects of the Biology, Ecology, and Behavior of Risso's 
Dolphins (Grampus griseus) off the California Coast. M.S. Thesis 
Thesis. University of California, Santa Cruz. 
Kruuk, H. 1972. The Spotted Hyaena. Univ. Of Chicago Press. 
Kummer, H. 1968. The Social Organization of Hamadryas Baboons. Univ. 
of Chicago Press. 
Kummer, H. 1971. Primate Societies: Group Techniques of Ecological 
Adaptation. Chicago: Aldine-Atherton. 
216 
Kurland, J. 1979. Paternity, mother's brother, and human sociality. IN: 
N. Chagnon and W. Irons (eds). Evolutionary Biology and Human 
Social Behavior. N.Scituate, MA: Duxbury Press. 
Kurland, J.A. 1977. Kin selection in the Japanese monkey. Contrib. 
Primatol. 12: 
Lack, D. 1968. Ecological Adaptations for Breeding in Birds. London: 
Meuthen. 
Le Boeuf, B.J. 1974. Male-male competition and reproductive success in 
elephant seals. Am. Zool. 14:163-176. 
Leatherwood, S. and Reeves, R.R. 1983. The Sierra Club Handbook of 
Whales and Dolphins. San Francisco: Sierra Club Books. 302 pp. 
Leatherwood, S. and Reeves, R.R. 1989. Marine Mammal Research and 
Conservation in Sri Lanka 1985-1986. Nairobi, Kenya: United Nations 
Environment Program. 
Lilly, J.C. 1964. Animals 1n aquatic environments: adaptation of 
mammals to the ocean. pp. 741-747 IN: Handbook of Physiology -
Environment. Washington: Am. Physiol. Soc. 
Lopez, J.C. and Lopez, D. 1985. Killer whales (Orcinus area) of Patagonia, 
and their behavior of intentional stranding while hunting nearshore. 
J. Mammal. 66:181-183. 
Lorenz, K. 1966. On Aggression. London: Meuthen. 
Lott, D.F. and Minta, S.C. 1983. Random individual association and social 
group instability in American bison (Bison bison). 61:153-172. 
MacDonald, D.W. 1983. The ecology of carnivore social behaviour. Nature 
301:379-384. 
Mace, G., Harvey, P.H. and Clutton-Brock, T.H. 1980. Is brain size an 
ecological variable? Trends Neurosci. 3:193-196. 
Mace, G., Harvey, P.H. and Clutton-Brock, T.H. 1981. Brain size and 
ecology in small mammals. J. Zool. Lond. 193:333-354. 
217 
11
1 
Marsh, H. and Kasuya, T. 1984. Ovarian changes in the short-finned pilot 
whale, Globicephala macrorhynchus. Rep. Int. Whal. Commn. 
Spee. Iss. 6:311-335. 
Marsh, H. and Kasuya, T. 1986. Evidence for reproductive senescence in 
female cetaceans. Rep. Int. Whal. Commn. Spee. Iss. 8:57-74. 
Marsh, H. and Kasuya, T. 1991. An overview of the changes in the role of a 
female pilot whale with age. pp. 281-285 IN: K. Pryor and K. S. 
Norris (eds). Dolphin Societies: Discoveries and Puzzles. Berkeley: 
Univ. Calif. Press. 
Martin, A.R., Reynolds, P. and Richardson, M.G. 1987. Aspect of the 
biology of pilot whale (Globicephala melaena) in recent mass 
strandings on the British coast. J. Zool. Lond. 211:11-23. 
Mate, B.R. 1989. Watching habits and habitats from earth satellites. 
Oceanus 32:14-18. 
Mathewson, S.J. 1935. Blackfish in the Gulf of St. Lawrence. J. Mamm. 
16:234. 
Matkin, C.O. and Leatherwood, S. 1986. General biology of the killer 
whale, Orcinus orca: a synopsis of knowledge. pp. 35-68 IN: B. 
Kirkevold and J. S. Lockard (eds). Behavioral Biology of Killer 
Whales. N.Y.: A.R. Liss. 
Maul, G.E. and Sergeant, D.E. 1977. New cetacean records from Madeira. 
Bocagiana 43:1-8. 
Maynard Smith, J. 1964. Kin selection and group selection. Nature 
201:1145-1147. 
Maynard Smith, J. 1978. The Evolution of Sex. Cambridge Univ. Press. 
Maynard Smith, J. 1979. Game theory and the evolution of behaviour. 
Proc. Roy. Soc. Lond. B.205:475-488. 
Maynard Smith, J. 1982. Evolution and the Theory of Games. Cambridge 
Univ. Press. 
218 
. - . 
- -----
McBrearty, D.A., Message, M.A. and King, G.A. 1986. Observations on 
small cetaceans in the north-east Atlantic Ocean and the 
Mediterranean Sea: 1978-1982. pp. 225-249 IN: M. M. Bryden and R. 
J. Harrison (eds). Research on Dolphins. Oxford: Clarendon Press. 
McBride, A.F. and Hebb, D.O. 1948. Behavior of the captive bottlenose 
dolphin, Tursiops truncatus. J. Comp. Physiol. Psychol. 41:111-123. 
McBride , A.F. and Kritzler, H. 1951. Observations on pregnancy, 
parturition, and post-natal behavior in the bottlenose dolphin. J. 
Mammal. 32:251-266. 
McCann, C. 1974. Body scarring on Cetacea-odontocetes . Sci. Rept. 
Whales Res. Inst. 26:145-155. 
Mead, J.G. 1975. Anatomy of the external nasal passages and facial 
complex of the Delphinidae. Smith. Contr. Zool. 207:1-72. 
Mehlman, P . 1986. Male intergroup mobility in a wild population of 
Barbary macaque (Macaca sylvanus) , Ghomaran Rif Mountains, 
Morocco. Am. J. Primatol. 10:67-81. 
Melnick, D.J. and Hoelzer, G.A. 1992. Differences in male and female 
macaque dispersal lead to contrasting distributions of nuclear and 
mitochondrial DNA. Int. J. Primatol. 13:379-393. 
Mercer, M.C . 1967. Wintering of pilot whales, Globicephala melaena, in 
Newfoundland inshore waters. J. Fish. Res. Bd. Cart. 32:1145-1154. 
Mercer, M.C. 1975. Modified Leslie-DeLury population models of the long-
finned pilot whale (Globicephala melaena) and the annual 
production of the short-finned squid (flex illecebrosus) based upon 
t heir interaction at Newfoundland. J . Fish. Res. Bd. Can. 32:1145-
1154. 
Mesterton-Gibbons, M. and Dugatkin, L.A. 1992. Cooperation among 
unrelated individuals: evolutionary factors . Q. Rev. Biol. 67:267-281. 
Michener, G.R. 1982. Kin identification, matriarchies, and the evolution of 
sociality in ground-dwelling sciurids . IN: J . F . Eisenberg and D. G. 
Kleiman (eds). Recent Advances in the Study of Mammalian 
Behavior. Am. Soc. Mammal. Special Puhl. No. 7. 
219 
Milinkovitch, M. 1991. DNA-DNA hybridizations support ungulate 
ancestry of Cetacea. J. Evol. Biol. 5:149-160. 
Mills, M.G.L. 1990. Kalahari Hyaenas: Comparative Behavioural Ecology 
of Two Species. London: Unwin Hyman. 304 pp. 
Mitchell, E. 1970. Pigmentation pattern evolution in delphinid cetaceans: 
an essay in adaptive coloration. Can. J. Zool. 48:717-740. 
Mitchell, E. 1975. Porpoise, dolphin, and small whale fisheries of the 
world. IUCN monogr. 3:1-129. 
Mizroch, S.A., Beard, J.A. and Lynde, M. 1990. Computer-assisted photo-
identification of humpback whales. Rep. Int. Whal. Commn. Spee. 
Iss. 12:63-70. 
Moehlman, P. 1986. Ecology of cooperation in canids. pp. 64-86 IN: D. I. 
Rubenstein and R. W. Wrangham (eds). Ecological Aspects of Social 
Evolution: Birds and Mammals. Princeton, N.J.: Princeton Univ. 
Press. 
Molina, R.G. 1979. Observaci6nes sobre la corriente de Canarias. II 
Asemblea Nacional de Geodesia y Geofisica Com. Vol. 3. 
Moore, J. 1992. Dispersal, nepotism, and primate social behavior. Int. J. 
Primatol. 13:361-378. 
Moore, J. and Ali, R. 1983. Are dispersal and inbreeding avoidance 
related? Anim. Behav. 32:94-112. 
Moore, P .W.B. 1991. Dolphin psychophysics: concepts for the study of 
dolphin echolocation. pp. 365-382 IN: K. Pryor and K. S. Norris (eds). 
Dolphin Societies: Discoveries and Puzzles. Berkeley: Univ. Calif. 
Press. 
Morgane, P.J. and Jacobs, M.S. 1972. Comparative anatomy of the 
cetacean nervous system. pp. 117-244 IN: R. J . Harrison (eds). 
Functional Anatomy of Marine Mammals. London: Academic 
Press . 
Morgane, P.J., Jacobs, M.S . and Galaburda, A. 1986a. Evolutionary 
aspects of cortical organization in the dolphin brain. pp. 71-98 IN: M. 
220 
M. Bryden and R. J. Harrison (eds). Research on Dolphins. Oxford: 
Clarendon Press. 
Morgane, P.J., Jacobs, M.S. and Galaburda, A. 1986b. Evolutionary 
morphology of the dolphin brain. pp. 5-29 IN: R. J. Schusterman, J. 
A. Thomas and F. G. Wood (eds). Dolphin Cognition and Behavior: a 
Comparative Approach. Hillsdale, N.J.: Erlbaum Assoc. 
Morton, A. 1990. A quantitative comparison of the behaviour of resident 
and transient forms of the killer whale off the central British 
Columbia coast. Rep. Int. Whal. Commn. Spee. Iss. 12:245-248. 
Moss, C.J. and Poole, J.H. 1983. Relationships and social structure of 
African elephants. pp. 315-325 IN: R. A. Hinde (ed). Primate Social 
Relationships. Sunderland, MA: Sinauer Assoc. 
Moss, C.W. 1988. Elephant Memories: thirteen years in the life of an 
elephant family. N.Y.: W. Morrow and Co. 
Myers, P. 1978. Sexual dimorphism in size of vespertilionid bats. Am. 
Nat. 112:701-711. 
Myrick, A.C., Jr., Hohn, A.A., Barlow, J. and Sloan, P.A. 1986. 
Reproductive biology of female spotted dolphins, Stenella attenuata, 
from the eastern tropical Pacific. Fish. Bull. 84:2. 
Nachtigall, P.E. 1986. Vision, audition, and chemoreception in dolphins 
and other marine mammals. pp. 79-113 IN: R. J. Schusterman, J. 
A. Thomas and F. G. Wood (eds). Dolphin Cognition and Behavior: A 
Comparative Approach. Hillsdale, N.J.: Lawrence Erlbaum. 
Needler, A.W.H. 1931. Blackfish (Globicephalus) stranded on the south 
coast of Prince Edward Island. Can. Field Nat. 45:157-158. 
Nicol, D.J. and Croome, R.L. 1988. Trends in the Tasmanian cetacean 
sighting record. pp. 59-70 IN: M. Augee (ed). Marine Mammals of 
Australasia. Royal Zoological Society of New South Wales. 
Nishida, T. 1987. Local traditions and cultural transmission. pp. 462-474 
IN: B. B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham and 
T. T. Struhsaker (eds). Primate Societies. Chicago Univ. Press. 
221 
- . -
- -----
Nishiwaki, M. 1963. Taxonomical consideration on genera of Delphinidae. 
Sci. Rep. Whales Res. Inst. 17:93-103. 
Nishiwaki, M. 1964. Revision of "Taxonomical consideration on genera of 
Delphinidae. Sci. Rep. Whales Res. Inst. 18:171-172. 
Nishiwaki, M. 1972. General biology. IN: S. H. Ridgway (ed). Maamals of 
the Sea: Biology and Medicine. Springfield, IL: C.C. Thomas. 
Nishiwaki, M. 1975. Ecological aspects of smaller cetaceans, with 
emphasis on the striped dolphin (Stenella coeruleoalba). J. Fish. 
Res. Bd. Can. 32: 1069-1072. 
Nishiwaki, M. and Randa, C. 1958. Killer whales caught in the coastal 
waters off Ja pan in recent ten years. Sci. Rep. Whales Res. Inst. 
13:85-96. 
Nishiwaki, M. and Norris, K.S. 1966. A new genus, Peoponocephala, for 
the odontocete cetacean species Electra electra. Sci. Rept. Whales 
Res. Inst. 20:95-105. 
Nores, C. and Perez, C. 1988. Overlapping range between Globicephala 
macrorhynchus and Globicephala melaena in the northeastern 
Atlantic. Mammalia 52:51-55. 
Norris, K.S. 1967. Aggressive behavior in Cetacea. pp. 225-241 IN: C. C. 
Clemente and D. B. Lindsley (eds). Aggression and Defense: Neural 
Mechanisms and Social Patterns (Brain Function Vol. 5). Berkeley, 
CA: Univ. of Calif. Press. 
Norris, K.S. and Dohl, T.P. 1980a. The behavior of the wild spinner 
dolphin, Stenella longirostris. Fish. Bull. 77:821-849. 
Norris, K.S. and Dohl, T.P. 1980b. The structure and function of cetacean 
schools. IN: L. M. Herman (ed). Cetacean Behavior. N.Y.: Wiley & 
Sons. 
Norris, K.S. and Prescott, J.H. 1961. Observations on Pacific cetaceans of 
Californian and Mexican waters. Univ. Calif. Puhl. Zool. 63:291-402. 
222 
. . ~ - . . 
I I 
---
Norris, K.S. and Pryor, K. 1991. Some thoughts on grandmothers. pp. 287-
289 IN: K. Pryor and K. S. Norris (eds). Dolphin Societies: 
Discoveries and Puzzles. Berkeley: Univ. Calif. Press. 
Norris, K.S. and Schilt, C.R. 1988. Cooperative societies in three-
dimensional space: on the origins of aggregationsm flocks, and 
schools, with special reference to dolphins and fish. Ethol. Sociobiol. 
9:149-179. 
Norris, K.S., Wursig, B., Wells, R.S., Wursig, M., Brownlee, S.M., 
Johnson, C. and Solow, J. 1985. The behavior of the Hawaiian 
spinner dolphin, Stenella longirostris. Southwest Fish. Center 
Admin. Bull. LJ-85-06C. 
Ohsumi, S. 1966. Sexual segregation of the sperm whale in the North 
Pacific. Sci. Rep. Whales Res. Inst. 20:1-16. 
Olesiuk, P.F., Bigg, M.A. and Ellis, G.M. 1990. Life history and population 
dynamics of resident killer whales (Orcinus orca) in the coastal 
waters of British Columbia and Washington State. Rep. Int. Whal. 
Commn Spee. Iss. 12:209-243. 
Oliver, W.R.B. 1924. Stranded blackfish at Marsden Point, New Zealand. 
N.Z. J. Sci. Tech. 7:188-189. 
Orians, G.H. 1969. On the evolution of mating systems in birds and 
mammals. Am. Nat. 103:589-603. 
Osborne, R.W. 1986. A behavioral budget of Puget Sound killer whales. pp. 
211-249 IN: B. Kirkevold and J. S. Lockard (eds). Behavioral Biology 
of Killer Whales. New York: A.R. Liss. 
Ostman, J. 1991. Changes in aggressive and sexual behavior between two 
male bottlenose dolphins (Tursiops truncatus) in a captive colony. 
pp. 305-317 IN: K. Pryor and K. S. Norris (ed$). Dolphin Societies: 
Discoveries and Puzzles. Berkeley: Univ. Calif. Press. 
Overholtz, W.J. and Waring, G.T. 1991. Diet composition of pilot whales 
Globicephala sp. and common dolphins Delphinus delphis in the 
mid-Atlantic bight during spring 1989. Fish. Bull. 89:723-728. 
223 
Packer, C. 1979. Inter-troop transfer and inbreeding avoidance in Papio 
anubis. Anim. Behav. 27:1-36. 
Packer, C. 1986. The ecology of sociality in felids. pp. 429-451 IN: D. I. 
Rubenstein and R. W. Wrangham (eds). Ecological Aspects of Social 
Evolution: Birds and Mammals. Princeton, N.J.: Princeton 
University Press. 
Packer, C., Gilbert, D.A., Pusey, A.E. and O'Brien, S.J. 1991. A molecular 
genetic analysis of kinship and cooperation in African lions. Nature 
351:562-565. 
Packer, C. and Pusey, A. 1982. Cooperation and competition within 
coalitions of male lions: kin selection or game theory. Nature 
296:7 40-7 42. 
Packer, C. and Ruttan, L. 1988. The evolution of cooperative hunting. Am. 
Nat. 132:159-198. 
Packer, C., Scheel, D. and Pusey, A. 1990. Why lions form groups: food is 
not enough. Am. Nat. 136:1-19. 
Packer, C.R., Herbst, L., Pusey, A.E., Bygott, J.D., Hanby, J.P., Cairns, 
S.J. and Borgerhoff-Mulder, M. 1988. Reproductive success of lions. 
pp. 363-383 IN: T. H. Clutton-Brock (ed). Reproductive Success. IL: 
University of Chicago Press. 
Pagel, M. and Harvey, P.H. 1988. How mammals produce large-brained 
offspring. Evolution 42:948-957. 
Pagel, M.D. and Harvey, P.H. 1989. Taxonomic differences in the scaling 
of brain on body weight among mammals. Science 244:1589-1593. 
Paradiso, J.L. 1958. The common blackfish in Virginia coastal waters. J. 
Mammal. 39:440. 
Parker, E.D. 1984. Sperm competition and the evolution of animal mating 
strategies. pp. 1-60 IN: R. M. Smith (ed). Sperm Competition and the 
Evolution of Animal Mating Systems. Orlando, FL: Academic Press. 
Perrin, W.F. 1969. Color pattern of the eastern Pacific spotted porpoise, 
Stenella graffmani Lonnberg. Zoologica 54:12. 
224 
Perrin, W.F. 1975. Distribution and differentiation of populations of 
dolphins of the genus Stenella in the easterntropical Pacific. J. Fish. 
Res. Bd. Can. 32: 1059-1072. 
Perrin, W.F. 1984. Patterns of geographic variation in small cetaceans. 
Acta Zool. Fennica 172:137-140. 
Perrin, W.F. and Reilly, S.B. 1984. Reproductive parameters of dolphins 
and small whales of the family Delphinidae. Rep. Int. Whal. 
Commn. Special Issue 6:97-125. 
Perryman, W.L. and Foster, T.C. 1980. Preliminary report on predation by 
small whales, mainly the false killer whale (Pseudorca crassidens) 
on dolphin (Stenella spp. and Delphinus delphis) in the eastern 
tropical Pacific. Southwest Fish. Center Admin. Rept. LJ-80-05. 
Pianka, E.R. 1976. Optimal selection of reproductive tactics. Am. Zool. 
16:775-784. 
Pilleri, G. and Gihr, M. 1969. Relatives Hirngewicht der Cetacea. Rev. 
Suisse Zool. 76:767-779. 
Pilleri, G. and Gihr, M. 1971. The central nervous system of the mysticete 
and odontocete whales. Invest. Cetacea 2:89-128. 
Ploog, D.W. 1967. The behavior of squirrel monkeys (Saimiri sciureus) as 
revealed by sociometry, bioacoustics, and brain stimulation. pp. 149-
184 IN: . S. A. Altmann (ed). Social Communication Among 
Primates. Univ. of Chicago Press. 
Ploog, D.W. and McClean, P.D. 1963. Display of penile erections in squirrel 
monkeys (Saimiri sciureus). Anim. Behav. 11:32-39. 
Poole, T. 1985. Social Behaviour in Mammals. Glasgow: Blackie. 248 pp. 
Popper, A.N. 1980. Sound emission and detection by delphinids. IN: L. M. 
Herman (ed). Cetacean Behavior: Mechanisms and Functions. 
N.Y.: Wiley. 
Price, W.S. 1985. Whaling in the Caribbean: Historical perspective and 
update. Rep. Int. Whal. Commn. 35:413-420. 
225 
- ' . ... . ' . - ' . . , -
.-------
Pryor, K. and Shallenberger, I.K. 1991. Social structure in spotted 
dolphins (Stenella attenuata) in the tuna purse seine fishery in the 
eastern tropical Pacific. pp. 161-196 IN: K. Pryor and K. S. Norris 
(eds). Dolphin Societies: Discoveries and Puzzles. Berkeley: Univ. 
Calif. Press . 
Pryor, T., Pryor, K. and Norris, K.S. 1965. Observations on a pygmy killer 
whale (Feresa attenuata) from Hawaii. J . Mammal. 46:450-461. 
Pusey, A. 1980. Inbreeding avoidance in chimpanzees. Anim. Behav. 
28:543-552. 
Pusey, A.E. and Packer, C. 1987a. Dispersal and philopatry. pp. 250-266 
IN: B. B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham and 
T. T. Struhsaker (eds). Primate Societies. Chicago Univ. Press. 
Pusey, A.E. and Packer, C. 1987b. The evolution of sex-biased dispersal in 
lions. Behaviour 101:275-310. 
Ralls, K. 1977. Sexual dimorphism in mammals: avian models and 
unanswered questions. Am. Nat. 111:917-938. 
Ralls, K., Brugger, K. and Ballou, J . 1979. Inbreeding and juvenile 
mortality in small populations of ungulates. Science 206:1101-1103. 
Ralls, K ., Brugger, K. and Glick, A. 1980. Deleterious effects of inbreeding 
in a herd of captive Dorcas gazelle. Intl. Zoo Yearbook 20:137-146. 
Rayner, G.W. 1939. Globobicephala leucosagmaphora, a new species of the 
genus Globicephala. Ann. Mag. Nat. Hist. 4:543-544. 
Rayner, J.M.V. 1985. Linear relations in biomechanics: the statistics of 
scaling functions. J. Zool. Soc. Lond. 206:415-439. 
Rice, D.W. 1984. Cetaceans. pp. 447-490 IN: S. Anderson and J. K. J. 
Jones (eds). Orders and Families of Recent Mammals of the World. 
N.Y.: J . Wiley and Sons. 
Rice, D.W. 1989. The scientific name of the pilot whale. Mar. Mamm. Sci . 
5:210. 
226 
. 
- . 
--
Rice, D.W. 1990. The scientific name of the pilot whale-a rejoinder to 
Schevill. Mar. Mamm. Sci. 6:359-361. 
Ridgway, S.H. 1986a. Dolphin brain size. pp. 59-70 IN: M. M. Bryden and 
R. J. Harrison (eds). Research on Dolphins. Oxford: Clarendon 
Press. 
Ridgway, S.H. 1986b. Physiological observations on dolphin brains. pp. 31-
59 IN: R. J. Schusterman, J. A. Thomas and F. G. Wood (eds). 
Dolphin Cognition and Behavior: a Comparative Approach. 
Hillsdale, N.J.: Erlbaum Assoc. 
Ridgway, S.H. 1990. The central nervous system of the bottlenose dolphin. 
pp. 69-97 IN: S. Leatherwood and R.R. Reeves (eds). The Bottlenose 
Dolphin. San Diego: Academic Press. 
Ridgway, S.H. and Brownson, R.H. 1984. Relative brain sizes and cortical 
surface areas in odontocetes. Acta Zool. Fennica 172:149-152. 
Ridgway, S.H., Flanigan, N.J. and McCormick, J.G. 1966. Brain-spinal 
cord ratios in porpoises: possible correlations with intelligence and 
ecology. Psychonom. Sci. 6:491-492. 
Ridgway, S.H. and Green, R.F. 1967. Evidence for a sexual rhythm in 
male porpoises, Lagenorhynchus obliquidens and Delphinus delphis 
bairdi. Norsk Hvalfangsttid 56:1-8. 
Ridgway, S.H. and Wood, F.G. 1988. Dolphin brain evolution. Behav. 
Brain Sci. 11:99. 
Ritchie, J. 1924. Observations on a pilot whale stranded in the Firth of 
Forth. Scot. Nat. 1924:37-43. 
Rommel, S.A., Pabst, D.A., McLellan, W.A., Mead, J.G. and Potter, C.W. 
1992. Anatomical evidence for a countercurrent heat exchanger 
associated with dolphin testes. Anat. Ree. 232:150-156. 
Ross, G.J.B. 1984. The smaller cetaceans of the south east coast of 
southern Africa. Ann. Cape Prov. Mus. (Nat. Hist.) 15:173-410. 
Rutberg, A.T. 1983. The evolution of monogamy in primates. J . Theor. 
Biol. 104:93-112. 
'Z2.7 
- . . . .. . . - " . . . .. ' .. . ~ 
- ------
Saayman, G.S. and Tayler, C.K. 1979. The socioecology of humpback 
dolphins (Sousa sp.). pp. 165-226 IN: H. E. Winn and B. L. Olla (eds). 
Behavior of Marine Animals; Current Perspectives in Research, Vol. 
3: Cetaceans. N.Y.: Plenum Press. 
Saayman, G.S., Tayler, C.K. and Bower, D. 1973. Diurnal activity cycles in 
captive and free-ranging Indian Ocean bottlenose dolphins (Tursiops 
aduncus Ehrenburg). Behaviour 44:212-333. 
Sandell, M. 1989. Ecological energetics, optimal body size and sexual 
dimorphism: a model applied to the stoat, Mustela erminea L. Fune. 
Ecol. 3:315-324. 
Sawaguchi, T. 1992. The size of the neocortex in relation to ecology and 
social structure in monkeys and apes. Folia Primatol. 58:131-145. 
Scammon, C.M. 1869. On the cetaceans of the western coast of North 
America. Proc. Nat. Acad. Sci. Phila. 21:32-80. 
Schaller, G.B. 1972. The Serengeti Lion. Chicago: Chicago University 
Press. 
Scheel, D. and Packer, C. 1991. Group hunting behaviour of lions: a search 
for cooperation. Anim. Behav. 41:697-709. 
Schevill, W.E. 1990a. On stability in zoological nomenclature. Mar. 
Mamm. Sci. 6:168-169. 
Schevill, W.E. 1990b. Reply to 'The scientific name of the pilot whale - a 
rejoinder to Schevill'. Mar. Mamm. Sci. 6:360. 
Schmiegelow, J.M.M. and Filho, A.M.P. 1989. First record of the short-
finned pilot whale Globicephala macrorhynchus Gray, 1846, for the 
southwestern Atlantic. Mar. Mamm. Sci. 5:387-391. 
Schmincke, H.-U. 1976. The geology of the Canary Islarids. pp. 67-184 IN: 
G. Kunkel (ed). Biogeography and Ecology in the Canary Islands. 
The Hague: Dr. W Junk b.v. Publishers. 
Schmitz, O.J. and Lavigne, D.M. 1984. Intrinsic rate of increase, body size, 
and specific metabolic rate in marine mammals. Oecologia 62:305-
309. 
228 
Schoener, T.W. 1969. Models of optimal size for solitary predators. Am. 
Nat. 103:277-313. 
Schroeder, J.P. 1990. Breeding bottlenose dolphins in captivity. pp. 435-446 
IN: S. Leatherwood and R. R. Reeves (eds). The Bottlenose Dolphin. 
San Diego, CA: Academic Press. 
Schroeder, J.P. and Keller, K.V. 1990. Artificial insemination of bottlenose 
dolphins. pp. 447-460 IN: S. Leatherwood and R.R. Reeves (eds). The 
Bottlenose Dolphin. San Diego, CA: Academic Press. 
Scott, M.D. and Perryman, W.L. 1991. Using aerial photogrammetry to 
study dolphin school structure. pp. 227-241 IN: K. Pryor and K. S. 
Norris (eds). Dolphin Societies: Discoveries and Puzzles. Berkeley: 
Univ. Calif. Press. 
Scott, M.S., Wells, R.S. and Irvine, A.B. 1990. A long-term study of 
bottlenose dolphins on the west coast of Florida. pp. 235-244 IN: S. 
Leatherwood and R. R. Reeves (eds). The Bottlenose Dolphin. 
Orlando, FL: Academic Press. 
Seagars, D.J. and Henderson, J.R. 1985. Cephalopod remains from the 
stomach of a short-finned pilot whale collected near Santa Catalina 
Island, California. J. Mammal. 66:777-779. 
Selander, R.K. 1966. Sexual dimorphism and differential niche utilization 
in birds. Condor 68:113-151. 
Selander, R.K. 1972. Sexual selection and dimorphism in birds. IN: B. G. 
Campbell (ed). Sexual Selection and the Descent of Man, 1871-1971. 
Chicago, IL: Aldine. 
Sergeant, D.E. 1962a. The biology of the pilot or pothead whale, 
Globicephala melaena (Traill) in Newfoundland waters. Bull. Fish. 
Res. Bd. Can. 132: 1-84. 
Sergeant, D.E. 1962b. On the external characters of the blackfish or pilot 
whales (genus Globicephala). J. Mammal. 43:395-413. 
Sergeant, D.E. 1982. Mass strandings of toothed whales (Odontoceti) as a 
population phenomenon. Sci. Rep. Whales Res. Inst. 34: 1-4 7. 
229 
- ' - ••••• 0 ·-~ I" • 0 • - • ,•.' { 00 •- • ••••• •• --~--.,, ......... -,----,,-. •-•T• '' .. • •• 
-
Sergeant, D.E., Caldwell, D.K. and Caldwell, M.C. 1973. Age, growth, and 
maturity of bottlenosed dolphins (Tursiops truncatus) from northest 
Florida. J. Fish. Res. Bd. Can. 30:1009-1011. 
Sergeant, D.E. and Fisher, H.D. 1957. The smaller Cetacea of eastern 
Canadian waters. J. Fish. Res. Bd. Can. 14:83-115. 
Shane, S.H. and McSweeney, D. 1990. Using photo-identification to study 
pilot whale social organization. Rep. Int. Whal. Commn. Spee. Iss. 
12:259-263. 
Shane, S.H., Wells, R.S. and Wiirsig, B. 1986. Ecology, behavior, and social 
organization of the bottlenose dolphin: a review. Mar. Mamm. Sci. 
2:34-63. 
Sherman, P.W. and Holmes, W.G. 1985. Kin recognition: issues and 
evidence. pp. 437-460 IN: B. Holldobler and M. Lindauer (eds). 
Experimental Behavioral Ecology. Stuttgart: Fischer-Verlag. 
Sherman, P.W. and Morton, M.L. 1984. Demography of Belding's ground 
squirrels. Ecology 1617-1628. 
Shields, W.M. 1982. Philopatry, Inbreeding, and the Evolution of Sex. 
Albany, N.Y.: State Univ. of New York Press. 
Shields, W.M. 1983. Optimal inbreeding and the evolution of philopatry. 
pp. 116-131 IN: I. R. Swingland and P. J. Greenwood (eds). The 
Ecology of Animal Movement. Oxford: Clarendon Press. 
Shields, W.M. 1987. Dispersal and mating systems: Investigating their 
causal connections. pp. 3-24 IN: B. D. Chepko-Sade and Z. T. Halpin 
(eds). Mammalian Dispersal Patterns. Univ. of Chicago Press. 
Silva, P.H.D.H.d. 1987. Cetaceans (whales, dolphins and porpoises) 
recorded off Sri Lanka, India, from the Arabian Sea and Gulf, Gulf of 
Aden and from the Red Sea. J. Bomb. Nat. Hist. Soc. 84:505-525. 
Smith, T .D., Payne, P .M., Heinemann, D., Waring, G. and Lange, A. 1990. 
Simultaneous fishery resource and seabird and cetacean sighting 
surveys. North Atl. Stud. 2:90-101. 
230 
... ' ' .. ' • ,, ' .. ' ,. ~ ,. ' ~' •• ' • ' • • l. ••• ... .. '.'. - ,. ,- • - .. , ·- • - - ~ 
·- -----
Smuts, B.B. 1987. Sexual competition and mate choice. pp. 385-399 IN: B. 
B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham and T. T. 
Struhsaker (eds). Primate Societies. Chicago Univ. Press. 
Smuts, B.B. and Watanabe, J.M. 1990. Social relationships and ritualized 
greetings in adult male baboons (Papio cynocephalus anubis). Int. J. 
Primatol. 11:147-172. 
Sokal, R.R. and Sneath, P.H.A. 1963. Principles of Numerical Taxonomy. 
San Francisco: W.R. Freeman and Co. 
SPSS. 1990. SPSS Base System User's Guide. Chicago, IL: SPSS Inc. 
Stammbach, E. 1987. Desert, forest, and montane baboons: multilevel 
societies. pp. 112-120 IN: B. B. Smuts, D. L. Cheney, R. M. Seyfarth, 
R. W. Wrangham and T. T. Struhsaker (eds). Primate Societies. 
Chicago: Chicago University Press. 
Stern, B.R. and Smith, D.G. 1984. Sexual behavior and paternity in three 
captive groups of rhesus monkeys (Macaca mulatta). Anim. Behav. 
32:23-32. 
Stevens, T.A., Duffield, D.A., Asper, E.D., Hewlett, K.G., Bolz, A., Gage, 
L.J. and Bossart, G.D. 1989. Preliminary findings of restriction 
fragment differences in mitochondrial DNA among killer whales 
(Orcinus orca). Can. J. Zool. 67:2592-2595. 
Stewart, K.J. and Harcourt, A.H. 1987. Gorillas: variation in female 
relationships. pp. 155-164 IN: B. B. Smuts, D. L. Cheney, R. M. 
Seyfarth, R. W. Wrangham and T. T. Struhsaker (eds). Primate 
Societies. Chicago Univ. Press. 
Stirling, I. 1983. The evolution of mating systems in pinnipeds. Spee. 
Puhl. Am. Soc. Mammal. 7 
Taruski, A.G. and Winn, H.E. 1976. Winter sighting of odontocetes in the 
vVest Indies. Cetology 22: 1-12. 
Taub, D. 1980. Testing the agonistic buffering hypothesis. I. The dynamics 
of participation in the triadic interactions. Behav. Ecol. Sociobiol. 
6:187-197. 
231 
r 
Tavolga, M.C. 1966. Behavior of the bottlenose dolphin (Tursiops 
truncatus): Social interactions in a captive colony. pp. 718-730 IN: K. 
S. Norris (ed). Whales, Dolphins and Porpoises. Berkeley: Univ. of 
Calif. Press. 
Tavolga, M.C. and Essapian, F.S. 1957. The behavior of the bottlenosed 
dolphin (Tursiops truncatus): Mating, pregnancy, parturition, and 
mother-infant behavior. Zoologica 42:11-31. 
Tayler, C.K. and Saayman, G.S. 1972. The social organization and 
behavior of dolphins (Tursiops aduncus) and baboons (Papio 
ursinus): Some comparisons and assessments. Ann. Cape. Prov. 
Mus. Nat. Hist. 9:11-49. 
Torres, D., Yanez, V. and Cattan, P.E. 1975. Mamiferos marinos de Chile: 
antecedentes y situaci6n actual. Biol. Pesq. Chile 11:49-81. 
Trivers, R. 1972. Parental investment and sexual selection. pp. 136-179 IN: 
B. Campbell (ed). Sexual Selection and the Descent of Man, 1871-
1971. Chicago: Aldine. 
Trivers, R. 1985. Social Evolution. Menlo Park, CA: 
Benjamin/Cummings. 
Tyack, P. 1981. Interactions between singing Hawaiian humpback whales 
and conspecifics nearby. Behav. Ecol. Sociobiol. 8:105-116. 
Tyack, P. and Whitehead, H. 1983. Male competition in large groups of 
wintering humpback whales. Behaviour 132-145. 
Uda, M. 1954. Studies on the relation between the whaling grounds and 
the hydrographical conditions. Sci. Rep. Whales Res. Inst. 9:179-187. 
Vehrencamp, S. 1983. A model for the evolution of despotic versus 
egalitarian societies. Anim. Behav. 31:667-682. 
Verner, J . and Willson, M.F. 1966. The influence of habitats on mating 
systems of North American passerine birds. Ecology 4 7: 143-14 7. 
Vonk, R. and Martin, V.M: 1988. First list of odontocetes from the Canary 
Islands, 1980-1987. 2nd Ann Conf European Cetacean Soc, Troia, 
Portugal. 
2.32 
9: 1 - , J ,, ... •:, 1,,-... ,.,,, •·' ,, • r-, -'·'" ••, 1• 1~, •r1-.~· '" ,, -. r-•,•"''"' •-·· 14r•'"•--r---......- .. --,..,.- .. ,. -- - • • · 
----------
Wada, S. 1988. Genetic differentiation between two forms of short-finned 
pilot whales off the Pacific coast of Japan. Sci. Rep. Whales Res. 
Inst. 39:91-101. 
Wade, M.J. and Breden, F.J. 1987. Kin selection in complex groups: 
mating structure, migration structure, and the evolution of social 
behaviours. pp. 273-283 IN: B. D. Chepko-Sade and Z. T. Halpin (eds). 
Mammalian Dispersal Patterns. Univ. of Chicago Press. 
Walker, W.A., Leatherwood, S., Goodrich, K.R., Perrin, W.F. and Stroud, 
R.K. 1986. Geographical variation and biology of the Pacific white-
sided dolphin, Lagenorhynchus obliquidens, in the north-eastern 
Pacific. pp. 441-465 IN: M. M. Bryden and R. J. Harrison (eds). 
Research on Dolphins. Oxford: Clarendon Press. 
Waring, G.T., Gerrior, P., Payne, P.M., Parry, B.L. and Nicolas, J.R. 1990. 
Incidental take of marine mammals in foreign fishery activities off 
the northeast United States, 1977-88. Fish. Bull. 88:347-360. 
Waser, P.M. and Jones, W.T. 1983. Natal philopatry among solitary 
mammals. Q. Rev. Biol. 58:355-390. 
Wells, R.S. 1978. Home range characteristics and group composition of 
Atlantic bottlenosed dolphins, Tursiops truncatus, on the west coast 
of Florida. M.Sc. Thesis. Univ. of Florida. 
Wells, R.S. 1986. Structural Aspects of Dolphin Societies. Ph.D. Thesis. 
University of California, Santa Cruz. 
Wells, R.S. 1991. The role of long-term study in understanding the social 
structure of a bottlenose dolphin community. pp. 199-225 IN: K. 
Pryor and K. S._ Norris (eds). Dolphin Societies: Discoveries and 
Puzzles. Berkeley: Univ. Calif. Press. 
Wells, R.S., Irvine, A.B. and Scott, M.D. 1980. The social ecology of inshore 
odontocetes. pp. 263-317 IN: L. M. Herman (ed). Cetacean Behavior: 
Mechanisms and Processes. New York: Wiley & Sons. 
Wells, R.S. and Scott, M.D. 1990. Estimating bottlenose dolphin population 
parameters from individual identification and capture-release 
techniques. Rept. Int. Whal. Commn. Spee. Iss . 12:407-415. 
233 
.1 
Wells, R.S., Scott, M.D. and Irvine, A.B. 1987. The social structure of free-
ranging bottlenose dolphins. pp. 247-305 IN: H. H. Genoway (ed). 
Current Mammalogy. New York: Plenum Press. 
West Eberhard, M.J. 1975. The evolution of social behavior by kin selection. 
Q. Rev. Biol. 50: 
Whitehead, H. 1985. Why whales leap. Sci. Am. 252:84-93. 
Whitehead, H. 1990. Rules for roving males. J. Theor. Biol. 145:355-368. 
Whitehead, H. and Arnbolm, T. 1987. Social organization of sperm whales 
off the Galapagos Islands, February-April 1985. Can. J. Zool. 65:913-
919. 
Wickler, W. 1966. Ursprung und biologische deutung des Genital-
prasentirens mannlicher Primaten. Z. Tierpsychol. 23:422-437. 
Wiley, R. 1974. Evolution of social organization and life history patterns 
among grouse. Q. Rev. Biol. 49:201-227. 
William, T.M., Friedl, W.A., Fong, M.L., Yamada, R.M., Sedivy, P. and 
Haun, J.E. 1992. Travel at low energetic cost by swimming and 
wave-riding bottlenose dolphins. Nature 355:821-823. 
Williams, G.C. 1966. Natural selection, the costs of reproduction, and a 
refinement of Lack's principle. Am. Nat. 100:687-690. 
Williams, G.C. · 1975. Sex and Evolution. Princeton, N.J.: Princeton Univ. 
Press. 
Wilson, E.O. 1975. Sociobiology: The New Synthesis. Cambridge, MA: 
Harvard Univ. Press . 
Wood, F.G. and Evans, W.E. 1980. Adaptiveness and ecology of 
echolocation in toothed whales. pp. 381-426 IN: ~- G. Busnel and J. 
F. Fish (eds). Animal Sonar Systems. N.Y.: Plenum Press. 
Wood, F.G.J., Caldwell, D.K. and Caldwell, D.K. 1970. Behavioral 
interactions between porpoises and sharks. Invest. on Cetacea 2:264-
277. 
234 
-
r • • I r '"' .. •, ,, .. ,, ' , •.O f • t • • '""' <., 1 ,, ~, •-, ,- ''.4 ">• , I,,..,-,,, ,..._,-,,.,.1 ... .,..,, •01 ..... J •,, 
Wooton, J.T. 1987. The effects of body mass, phylogeny, habitat, and 
trophic level on mammalian age at first reproduction. Evolution 
41:732-749. 
Worthy, G.A.J. and Hickie, J .P. 1986. Relative brain size 1n marine 
mammals. Am. Nat. 128:445-459. 
Wrangham, R.W. 1980. An ecological model of female-bonded primate 
groups. Behaviour 75:262-300. 
Wrangham, R.W. 1983. Social relationships in comparative perspective. 
pp. 325-334 IN: R. A. Hinde (ed) . Primate Social Relationships. 
Sunderland, MA: Sinauer Associates. 
Wrangham, R.A. and Rubenstein, D.I. 1986. Social evolution in birds and 
mammals. pp. 452-4 70 IN: D. I. Rubenstein and R. A. Wrangham 
(eds). Ecological Aspects of Social Evolution: Birds and Mammals. 
Princeton, N.J.: Princeton University Press. 
Wursig, B. 1978. Occurrence and group organization of Atlantic bottlenose 
porpoises (Tursiops truncatus), in an Argentine bay. Biol. Bull. 
154:348-359. 
Wursig, B. 1986. Delphinid foraging strategies. pp. 347-359 IN: R. J. 
Schusterman, J. A. Thomas and F. G. Wood (eds). Dolphin 
Cognition and Behavior: a Comparative Approach. Hillsdale, N.J.: 
Erlbaum Assoc. 
Wursig, B. 1989. Cetaceans. Science 244:1550-1557. 
Wursig, B., Cipriano, F. and Wursig, M. 1991. Dolphin movement 
patterns: information from radio and theodolite tracking studies. pp. 
79-111 IN: K: Pryor and K. S. Norris (eds). Dolphin Societies : 
Discoveries and Puzzles. Berkeley: Univ. of Calif. Press. 
Wursig, B. and Jefferson, T.A. 1990. Methods of photo-identification for 
small cetaceans. Rep. Int. Whal. Commn Spee. Iss. 12:43-52. 
Wursig, B. and Wursig, M. 1977. The photographic determination of group 
size, composition, and stability of coastal porpoises (Tursiops 
truncatus). Science 198: 755-756. 
235 
., • l ' •r -~ ~ • • 0 
Wursig, B. and Wursig, M. 1979. Behavior and ecology of the bottlenose 
dolphin, Tursiops truncatus, in the south Atlantic. Fish. Bull. 77: 
Wursig, B. and Wursig, M. 1980. Behavior and ecology of the dusky 
dolphin, Lagenorhynchus obscurus, in the south Atlantic. Fish Bull 
77:871-890. 
Wursig, B., Wursig, M. and Cipriano, F. 1989. Dolphins 1n different 
worlds. Oceanus 32:71-75. 
Yonekura, M., Matsui, S. and Kasuya, T. 1980. On the external characters 
of Globicephala macrorhynchus off Taiji, Pacific coast of Japan. Sci. 
Rep. Whales Res. Inst. 32:67-95. 
236