Content uploaded by Christin Brangwynne Khan
Author content
All content in this area was uploaded by Christin Brangwynne Khan on Nov 19, 2014
Content may be subject to copyright.
VOCAL DEVELOPMENT IN HARBOR SEAL PUPS, PHOCA VITULINA
A thesis submitted to the faculty of
San Francisco State University
in partial fulfillment of
the requirements for
the degree
Master of Arts
in
Biology: Physiology and Behavioral Biology
by
Christin Brangwynne Khan
San Francisco, California
December, 2004
Copyright by
Christin Brangwynne Khan
2004
CERTIFICATION OF APPROVAL
I certify that I have read Vocal Development in Harbor
Seal Pups, Phoca vitulina by Christin Brangwynne
Khan, and that in my opinion this work meets the
criteria for approving a thesis submitted in partial
fulfillment of the requirements for the degree: Master of
Arts in Biology: Physiology and Behavioral Biology at
San Francisco State University
__________________________________________________________ ____________
Dr. Hal Markowitz Date
Professor of Biology
__________________________________________________________ ____________
Dr. Brenda McCowan Date
Research Behaviorist
__________________________________________________________ ____________
Dr. Edward Connor Date
Professor of Biology
VOCAL DEVELOPMENT IN HARBOR SEAL PUPS, PHOCA VITULINA
Christin Brangwynne Khan
San Francisco State University
2004
The goal of this research was to determine when harbor seal pup (Phoca vitulina)
vocalizations become sufficiently distinctive to allow individual recognition. A total of
4593 calls were analyzed from 15 captive pups. Nineteen were harsh, broadband,
staccato calls used in an aggressive context. The rest were tonal “Mother Attraction
Calls”, having an inverted “v” or “u” shaped spectrogram with harmonics and a
fundamental frequency around 200-600 Hz. Discriminant function analysis (DFA)
classified calls to the correct individual significantly more often than expected by chance
during each of the three age groups as well as when all ages were combined, suggesting
that calls are individually distinctive even in pups less than two weeks old. DFA
classified calls to the correct pup sex significantly more often than expected by chance,
and there were significant interactions between pup sex and age. The results of this study
should be interpreted with caution until the findings are verified in wild harbor seal pups.
I certify that the Abstract is a correct representation of the content of this thesis.
_____________________________________________ ______________________
Chair, Thesis Committee Date
v
ACKNOWLEDGEMENTS
This thesis would not have been possible without the contributions of
many friends and colleagues. I would like to thank Dr. Hal Markowitz for
giving me the opportunity to undertake this project, the freedom to choose
my own path, and finally, for many thought-provoking discussions. I
would like to thank Dr. Brenda McCowan for making it possible to
analyze such a large data set and for her endless patience in explaining so
many aspects of the analysis. I would like to thank Dr. Ed Connor for
opening my eyes to the world of statistics. Thanks to Dr. Frances Gulland
of The Marine Mammal Center for providing logistical support, and to the
volunteers of the Harbor Seal Hospital for their assistance and interest in
the project. Thank you to Greg Budney and the Cornell Laboratory of
Ornithology’s “Bird Sound Recording Workshop” for teaching me the
fundamentals of recording. Finally, thank you to my husband, Fazle, for
always encouraging me to pursue my dreams, and for practically pushing
me onto that first plane to San Francisco.
TABLE OF CONTENTS
List of Tables ………………………………………………………………………………... viii
List of Figures …………………………………………………………...…………………… ix
Introduction ……………………...…………………………………………………………… 1
Offspring Recognition ……….……………………………………………………… 1
Offspring Recognition in the Harbor Seal …..………………………………………. 2
Harbor Seal Pup Vocalizations ……………………………………………………… 6
Development of Offspring Recognition …………………………………………… 7
Development of Offspring Recognition in the Harbor Seal ………………………… 8
Hypothesis and Objectives ………………………………………………………………… 10
Methods …………………………………………………………………….. 11
Study Animals …………………………….………………………………………. 11
Recording Vocalizations …………………..……………………………………… 12
Acoustical Analysis ……………………………………………………………… 12
Statistical Analysis ………………………..……………………………………… 14
Individual Differences …………………………………………………… 14
Sex Differences ……...…………...……………………………………… 16
Age Differences ………………………………………………………….. 17
Results …………………………...…………………………………………………………… 18
General Call Characteristics …………………………………………………………. 18
vi
Aggressive Vocalizations …………………………………………………………… 21
Individual Differences ………….…………………………………………………… 24
Discriminant Function Analysis (DFA) …………………………………….. 24
Fundamental Frequency DFA Analysis …………………………………….. 26
Potential for Individuality Coding (PIC) …………………………………… 28
Sex Differences …….………………………..……………………………………… 29
Discriminant Function Analysis (DFA) …………………………………….. 29
Mixed-Effects Linear Regression (LME) …………………………………... 29
Age Differences ……….…………………….………………………………………. 33
Discussion …………………………………………………………………………………….. 36
General Call Characteristics ……………………………………………………….… 36
Aggressive Vocalizations …………………………………………………………… 36
Individual Differences ...……………………..……………………………………… 38
Sex Differences …….……………………………………………………………..… 41
Age Differences ………..……………………………………………………………. 43
Directions for Future Research …………….……………………………………… 44
Vocal Development of Harbor Seal Pups …………………………………... 44
Effects of Motivational State ……………………………………………….. 45
Vocal Recognition by Harbor Seal Mothers ……………………………… 46
Comparisons with Other Species and Populations ………………………… 47
References …………………………………………………………………………………… 48
vii
viii
LIST OF TABLES
Table Page
1. Information on Individual Harbor Seal Pups ……………………………………………... 11
2. Summary Acoustic Variables and Codes ……………………………………………….. 13
3. Summary of Analyzed Calls per Individual ….…………………………………………... 18
4. Descriptive Statistics of “Mother Attraction Calls” …………………………………….. 19
5. Descriptive Statistics of Aggressive Calls ………………………………………………. 22
6. DFA for Individual Differences Results …………………………………………………. 24
7. Canonical Discriminant Functions from DFA for Individual Differences ……………….. 25
8. DFA from Fundamental Subset for Individual Differences Results ……………………... 26
9. Canonical Discriminant Functions from DFA Fundamental Subset …..…………………. 27
10. Potential for Individuality Coding Results ..………...……………………………………. 28
11. LME for Age and Sex Differences Results ……………………………………………….. 31
ix
LIST OF FIGURES
Figure Page
1. Example Spectrogram of a “Mother Attraction Call” …..….………………………... 20
2. Example Spectrogram of an Aggressive Vocalization .….………………………….. 21
3. Example Waveform of an Aggressive Vocalization ……………………………… 23
4. LME Graphs for Sex Differences (with Age) ………………………………………. 32
5. LME Graphs for Age Differences …………………………………………………. 34
6. Comparison of Pup Vocalization Spectrograms …………………………………… 35
INTRODUCTION
Offspring Recognition
Parental care is a remarkable adaptation found in many birds and mammals that promotes
offspring survival. Parents provide energetically costly resources such as warmth, food,
and protection, which must be restricted to their own offspring in order to increase
reproductive success (Gubernick, 1981; Bradbury and Vehrencamp, 1998). Parents can
restrict care to their own offspring by recognizing offspring characteristics, recognizing a
geographic location, and/or by maintaining proximity to offspring. The optimal
mechanism depends on the number of offspring, the mobility of the offspring (i.e.,
altricial vs. precocial), and the number of nearby non-offspring young (Bradbury and
Vehrencamp, 1998). For example, in altricial species, immobile young remain isolated in
a nest, and geographic location is the optimal recognition cue (Bradbury and
Vehrencamp, 1998). If offspring intermingle with conspecifics as they become more
mobile, an individual signature system is favored (Beecher, 1991) and parents learn to
recognize their offspring during the nesting period (Halpin, 1991; Bradbury and
Vehrencamp, 1998). In precocial species, offspring characteristics must be learned
rapidly in a few hours, a process known as imprinting (Bradbury and Vehrencamp, 1998).
Parents may also maintain proximity with their young and isolate them from conspecifics
(goat, Capra hircus: Lenhardt, 1977; grey seal, Halichoerus grypus: Fogden, 1971;
humpback whale, Megaptera novaeangliae: Herman et al., 1980; California sea lion,
Zalophus californianus: Peterson and Bartholomew, 1969).
1
Offspring Recognition in the Harbor Seal
The ability of mammals to recognize their own offspring is greatest when the offspring
are single and precocial, and when the breeding population is gregarious (Gubernick,
1981). Harbor seals, Phoca vitulina, meet all of these criteria, and offspring recognition
is suggested by selective care and nursing (Bishop, 1967; Renouf et al., 1983). Females
give birth to a single pup annually on crowded breeding rookeries, and the unusually
precocial pups can follow their mothers both on land and in the water within a few
minutes. Since lifetime reproductive output is limited, females increase reproductive
success by investing heavily in each pup. Maternal care in pinnipeds ends abruptly at
weaning, and thus, the period of reproductive effort in harbor seals consists of an 11-
month gestation period followed by a 32-day lactation period (Cottrell et al., 2002).
Reproductive effort during gestation can be measured by the relative birth mass, or ratio
of pup birth mass to maternal parturition mass. Harbor seals have a relative birth mass of
12.9%, which is high for similarly sized mammals in general, and phocids in particular
(Ellis, 1998). The cost of lactation is especially high in pinnipeds, and harbor seals
deplete 79% of their initial fat stores during the first 19 days of lactation, which is the
greatest percentage reported for phocids (Bowen et al., 1992). Harbor seal females do
not have sufficient energy reserves to sustain fasting, and must resume foraging to meet
the combined demands of lactation and maternal metabolism (Bowen et al., 1992). Pups
are often left unattended during these brief foraging trips and reunite with their mother
upon her return to the rookery. While alone, hungry pups may solicit other seals but they
2
are rarely allowed to nurse (Bishop, 1967; Renouf et al., 1983; Renouf and Diemand,
1984), and are often subjected to high levels of aggression (Bishop, 1967; Renouf et al.,
1983). Harbor seal pups that become permanently separated from their mothers prior to
weaning are likely to starve to death, although adoptions may occur on occasion (Bishop,
1967). Thus, both harbor seal mothers and pups are under considerable selective pressure
to maintain contact with each other.
Alternating foraging trips to sea with time on land to nurse has been described as the
‘foraging cycle strategy’ and is found in all otariids and, in addition to the harbor seal,
possibly other small phocids (Boness et al., 1994; Boness and Bowen, 1996). Pup
mobility and maternal separations are correlated with mutual recognition in pinnipeds,
and thus, maternal strategy may be an important factor in the recognition system (Insley,
1992). Otariids using the ‘foraging cycle strategy’ that may also use mutual recognition
include: the South American fur seal, Arctocephalus australis (Phillips and Stirling,
2000, 2001); Galapagos fur seal, Arctocephalus galapagoensis (Trillmich, 1981);
subantarctic fur seal, Arctocephalus tropicalis (Roux and Jouventin, 1987; Charrier et al.,
2002a, 2002b, 2003a, 2003b, 2003c); Northern fur seal, Callorhinus ursinus
(Bartholomew, 1959; Lisitsyna, 1973; Insley, 2001); Southern sea lion, Otaria flavescens
(Fernandez-Juricic et al., 1999); and the California sea lion, Zalophus californianus
(Peterson and Bartholomew, 1969; Trillmich, 1981; Hanggi, 1992; Schusterman et al.,
1992). Several phocids also appear to recognize their offspring, although occasional
3
nursing by non-offspring in some species suggests a less reliable system than in otariids
(grey seal, Halichoerus grypus: Fogden, 1968, 1971; McCulloch et al., 1999; Northern
elephant seal, Mirounga angustirostris: Bartholomew and Collias, 1962; Fogden, 1968;
Petrinovich, 1974; Insley, 1992; harp seal, Phoca groenlandica: Kovacs, 1987).
The system of mother-pup recognition and reunion in pinnipeds using the ‘foraging cycle
strategy’ is strikingly similar to that of some colonial bats. This relationship has been
particularly well documented in Mexican free-tailed bats, Tadarida brasiliensis. In this
species, females leave their single pup in crowded breeding crèches containing millions
of other pups, and return twice a day to nurse. Early researchers thought that nursing was
indiscriminate (Davis et al., 1962) until a genetic study revealed that females usually
nurse their own pup (McCracken, 1984). Further research found that the mother narrows
her search by returning to the geographic location of the previous nursing bout
(McCracken, 1993). Once landed, she begins to crawl through the mass, searching and
emitting ‘directive’ calls that attract nearby pups (Balcombe and McCracken, 1992).
While looking for her own pup, the female guards her teats with folded wings, and will
vigorously attack any nearby pups who try to steal milk (McCracken and Gustin, 1987).
The mother recognizes and responds to the individually distinctive calls of her own pup
(Gelfand and McCracken, 1986; Balcombe, 1990), but will only accept the pup to nurse
after a final olfactory check (Gustin and McCracken, 1987).
4
Visual, acoustic, olfactory, and geographic cues probably play a role in the recognition
process of harbor seals in a manner similar to that described for Mexican free-tailed bats
(Balcombe, 1990; Balcombe and McCracken, 1992; Gelfand and McCracken, 1986;
Gustin and McCracken, 1987; McCracken, 1984, 1993; McCracken and Gustin, 1987).
Once the pair has reunited, mother-pup proximity is maintained by the following
behavior of the pup, although the mother may exert more overt control when necessary
(Renouf et al., 1983; Renouf and Diemand, 1984). This following tendency suggests that
pups recognize their mothers, and therefore, that harbor seals may have a mutual
recognition system. Renouf, Lawson, and Gaborko (1983) suggested that harbor seals fit
the original imprinting model in which infants are predisposed to follow their mothers’
movements at an early age. Harbor seal pups develop a strong tendency to follow their
mothers within 40 minutes postpartum (Lawson and Renouf, 1987). Female harbor seals
isolate their newborn by preventing the approach of conspecifics and continually
initiating nose-to-nose contact, suggesting an initially unbonded condition (Bishop, 1967;
Lawson and Renouf 1985, 1987). Recognition may be established within an hour or two
after birth in other phocids as well (grey seals, Halichoerus grypus: Fogden, 1971; harp
seals, Phoca groenlandica: Kovacs, 1987). Observations of a walrus calf, Odobenus
rosmarus, revealed a tendency to follow moving objects immediately after extraction
from his mother (Kibal’chich and Lisitsina, 1979). California sea lions, Zalophus
californianus, defend a territory around their pup for the first few days postpartum
(Peterson and Bartholomew, 1969), and pups have been shown to imprint on human
5
surrogates when raised in captivity (Schusterman et al., 1992). Schustermann and
colleagues (1992) proposed that filial imprinting might be widespread in otariids.
Harbor Seal Pup Vocalizations
Unlike most pinnipeds, female harbor seals do not emit a “pup attraction call.” The pups,
however, are quite vocal and begin to emit cries several hours after birth (Lawson and
Renouf, 1985) that continue throughout the lactation period. These calls presumably
function to aid the mother and pup in maintaining contact and reuniting once separated,
and can be emitted simultaneously in the air and water (Renouf, 1984). Research has
indicated that these calls may be individually distinctive (Perry and Renouf, 1988;
Renouf, 1984) and that a female seal may be able to distinguish among the vocalizations
of different pups (Renouf, 1985). Under ideal conditions, harbor seal pup cries may be
detected at ranges of 1 km or more (Terhune, 1991), thus acoustic recognition would be a
powerful tool for the mother in maintaining contact with her pup.
Harbor seal pup vocalizations may encode information about both identity and
motivational state. The ideal vocalizations for use as an individually recognizable cue
would be highly stereotyped within each individual, but vary between individuals (Falls,
1982). However, vocalizations used to express differences in motivation are expected to
be highly variable within individuals (Falls, 1982). This discrepancy can be reconciled if
different features are used to encode motivational and identity information (Falls, 1982).
6
Typically, harmonic structure and strong frequency modulation are used to encode
information about individual identity in short, frequently repeated vocalizations
(Bradbury and Vehrencamp, 1998). Research on harbor seal pups suggests that the
fundamental frequency may broadcast identity, while motivational state may be conveyed
by repetition rate and number of harmonics (Renouf, 1984; Perry and Renouf, 1988).
Preliminary observations suggest that more distressed pups have a greater number of calls
per calling bout, a faster rate of call emission, and more harmonics (Renouf, 1984; Perry
and Renouf, 1988).
Development of Offspring Recognition
In species that do recognize their own offspring, the time at which parents begin to
recognize individually distinctive traits varies between species, populations, and perhaps
even individuals. Brood exchange experiments (bank swallow, Riparia riparia: Beecher
et al., 1981a; McArthur, 1979) and playback experiments (razorbill, Alca torda: Insley et
al., 2003a; bank swallow, Riparia riparia: Beecher et al., 1981a; McArthur, 1982; goat,
Capra hircus: Lenhardt, 1977) have demonstrated that parents typically begin to
discriminate between their own offspring and non-offspring young just prior to the time
when they would normally intermingle. Acoustic analysis has also revealed that some
offspring vocalizations become individually distinctive shortly before the parents are able
to use vocalizations as a reliable recognition cue (bank swallow, Riparia riparia: Beecher
et al., 1981a; McArthur, 1982; goat, Capra hircus: Lenhardt, 1977).
7
Particularly in crowded breeding crèches, strong signature evidence is needed and
multiple-modalities are commonly used (Bradbury and Vehrencamp, 1998) since the
more cues used, the greater the chances of successful recognition (Hepper, 1991). The
ability of parents to recognize their offspring’s unique characteristics may develop at
different times across different modalities. For example, mother goats quickly learn to
identify olfactory cues of their offspring (Halpin, 1991), but they are unable to recognize
acoustic cues until 4 days of age when the offspring develop individually distinctive
vocalizations (Lenhardt, 1977).
Development of Offspring Recognition in the Harbor Seal
Factors affecting the development of recognition are largely unknown for pinnipeds,
especially phocids. Recognition appears to develop in otariids within one or two weeks
postpartum, when the mother embarks on her first foraging trip (Insley, 2001).
Subantarctic fur seal, Arctocephalus tropicalis, mothers can recognize pup calls as early
as 1-2 days postpartum (Charrier et al., 2003a). The mother-response call in California
sea lions, Zalophus californianus, is fully developed within two weeks postpartum
(Peterson and Bartholomew, 1969) suggesting that vocal recognition occurs soon after.
Lactating female harbor seals commence trips to sea 6-11 days postpartum (Boness et al.
1994), and a mechanism for long-distance individual recognition should be in place by
this time. However, harbor seals are faced with the challenge of maintaining contact with
8
mobile pups immediately after birth both on land and in the water. The crowded
breeding rookeries are often in areas with low visibility, strong currents, and high
ambient noise due to wind, wave action, and conspecifics. The task is made even more
challenging by their nocturnal activity (Scheffer and Slipp, 1944; Yochem et al., 1987;
Thompson et al., 1989; Boness et al., 1994; Van Parijs et al., 1999; C. B. Khan personal
observation). Under these circumstances, the ability to discriminate among pups is
important immediately after parturition, and may be established first using olfactory cues.
Long-distance pup recognition using acoustic cues is only possible after pups begin to
produce vocalizations with sufficient individually distinctive characteristics. Thus, an
analysis of pup vocalizations can determine the earliest pup age at which mother harbor
seals may begin using acoustic signals as a recognition cue.
9
HYPOTHESIS AND OBJECTIVES
The goal of this research was to determine when characteristics of individual harbor seal
pup (Phoca vitulina) vocalizations become sufficiently distinctive to allow individual
recognition using acoustic cues. My hypothesis was that pup vocalizations would be
stereotyped and encode signature information by six days of age, the time at which pups
are naturally separated from their mothers and intermingle with conspecific young.
Additionally, I expected subtle changes in characteristics of the vocalizations to occur as
a function of physiological maturity and for these changes to be relatively consistent
between different individuals (i.e., gradually lowering in frequency). The null hypothesis
was that characteristics of individual harbor seal pup vocalizations do not change over
time. The present study is the first quantitative research on vocal development in harbor
seal pups.
10
METHODS
Study Animals
Aerial vocalizations were recorded from 15 harbor seal pups, Phoca vitulina richardii, at
The Marine Mammal Center in Sausalito, California from March 13th until May 18th of
2002. The Marine Mammal Center is a non-profit organization that rescues and
rehabilitates injured, sick, and orphaned marine mammals that strand along the California
coast, from Mendocino through San Luis Obispo. The Director of Veterinary Science,
Dr. Frances Gulland, assigned an estimated date of birth to each pup in the study based
on the presence and condition of umbilicus, teeth, and white blood cell counts (Table 1).
Table 1. Information about harbor seal pups, Phoca vitulina richardii, including a number, name, sex,
county of origin, estimated date of birth (D.O.B.), and criteria used to estimate date of birth. All pups were
recorded at The Marine Mammal Center in Sausalito, California during the spring of 2002.
No.
Name Sex
County D.O.B. Estimated Date of Birth (D.O.B.) Based on:
1 Bell F
Marin 4/4 4/8 no umbi
2 Benton M
Sonoma 3/25 4/1 no umbi
3 Brooks M
Contra Costa 4/5 4/15 no umbi, only 5 kg, high white blood cell count
4 Burley M
San Mateo 4/7 4/10 umbi long & beginning to dry; 4/12 still little bit
5 Davranger M
Marin 4/2 4/5 no umbi, but blood-tinged fluid from navel
6 DillonBaby M Marin 3/27 3/29 umbi present, moist
7 Eucie M
SLO 4/1 4/3 umbi drying at distal tip
8 Finklestein F Marin 3/24 3/25 1" umbi; 3/26 lost umbi
9 Lanny F
Marin 4/20 4/22 umbi present big, pink; 4/24 umbi gone
10 Mott M
Sonoma 4/12 4/13 wet umbi 3" very pink/red; 4/15 drying; 4/16 fell off
11 Spotti F
Monterey 4/26 5/5 pus from umbi (slight umbi remains?)
12 Spunky M
SLO 3/11 3/13 umbi partially dried
13 Sundae F
Marin 3/21 3/26 umbi fell off during exam
14 Sweetum M
SLO 3/22 3/24 some umbi remains; 3/26 no umbi
15
Towey
M
SLO
4/7
4/10 clipped umbi
11
Recording Vocalizations
Recordings were made with an Audio-technica 835b condenser microphone (frequency
response of 40-20,000 Hz) and a Sony TCD-D8 DAT recorder at sampling rates of 44.1
kHz and 48 kHz, allowing for an analysis range of at least 0-20 kHz. Flipper tags
allowed individual pups to be easily identified, and recordings were immediately
followed by a voice announcement to ensure that calls were assigned to the correct
individual. All recordings produced during this study are archived by the Macaulay
Library of Natural Sounds at the Cornell Laboratory of Ornithology in Ithaca, New York.
Acoustical Analysis
Recordings were digitized with a Roland UA-30 USB Audio Interface with DAT ACS
Digital Connection Cord (optical miniplug to rectangular 7-pin connector) using the
software Cool Edit Pro 1.2 (Syntrillium Software Corp., Phoenix, AZ) and an ESS
Maestro sound card at a sampling rate of 44.1 or 48 kHz and 16-bit resolution.
Recordings were archived as acoustic files (Windows PCM waveform format) on CD.
Spectrograms were generated using a 1024-point FFT with a Hamming filter.
Vocalizations with high signal-to-noise ratio and unambiguous identity were selected and
cued for subsequent digital analysis. Background noise was removed in Cool Edit Pro
with a Butterworth Band Pass filter using cutoffs of 100 Hz and 15 kHz. Vocalizations
were measured using a modified version of the Contour Similarity Technique (for
detailed descriptions of this technique, see McCowan 1995, and McCowan and Reiss,
12
2001). Sixty time and frequency points were extracted across the duration of each call by
following the dominant frequency (the frequency of highest amplitude). Nineteen
summary acoustic variables (Table 2) defining various spectral, temporal, and contour
parameters were calculated from these measurements.
Table 2. Summary acoustic variables (parameters) and codes that were calculated from the sixty frequency
and time measurements for each vocalization.
Code Parameter
SF Start frequency
FF Finish frequency
MIN Minimum frequency
MAX Maximum frequency
MEAN Mean frequency
PAF Frequency at peak amplitude
FR Frequency range (maximum frequency minus minimum frequency)
MAX/MEAN Maximum frequency divided by mean frequency
MEAN/MIN Mean frequency divided by minimum frequency
DUR Duration of the call
MINL Location of minimum frequency (as percentage of call)
MAXL Location of maximum frequency (as percentage of call)
SSL Start frequency slope
MSL Middle frequency slope
FSL Final frequency slope
CV Coefficient of variation (magnitude of frequency modulation)
JF Jitter factor (weighted measure of amount of frequency modulation)
COFM
Coefficient of frequency modulation
(measure of the amount and magnitude of frequency modulation)
Spectral
Characteristics
of the Call
ID Name or number code of the vocalizing pup
Recording session Date the recording was made
Days old Estimated age of vocalizing pup in days
Weeks old Estimated age of vocalizing pup in weeks
Age category Estimated age of vocalizing pup in 2-week intervals
Identification
Sex
Sex of vocalizing pup (1=female, 2=male)
13
Statistical Analysis
Individual Differences
Stepwise discriminant function analysis (DFA) was used to build a predictive model to
classify pups based on call characteristics (Table 2), and to find the dimensions along
which individuals differed using the statistical software SPSS 10.0 (SPSS Inc., Chicago,
IL). To meet the assumptions of this test, continuous variables were tested for normality
and log transformed as necessary. The conservative cross-validation procedure was
applied in which each call was classified by the functions derived from all calls other
than the one being tested. When sample sizes were unequal, percent correct expected by
chance was calculated separately for each individual; the overall percent correct expected
by chance was determined by adding the number of correctly classified calls expected for
each individual and then dividing this sum by the total number of calls analyzed
(Tabachnick and Fidell, 2001). Chi-square was then used to test whether correct
classification was significantly greater than expected by chance. Holm’s sequential
Bonferroni method was used to adjust alpha levels for multiple comparisons (Rice, 1989).
14
Since the fundamental frequency has been implicated as an individually distinctive
feature in the vocalizations of harbor seal pups (Renouf, 1984; Perry and Renouf, 1988),
cross-validation discriminant function analysis was also run on a subset of the data for
which calls were filtered and measured from the fundamental frequency only. Six pups
with large sample sizes were chosen for this analysis, and 40 calls were randomly
selected from each pup during each of three age groups (n = 240 calls per age group).
This analysis also enabled a comparison of individuality as a function of age.
As an additional means of comparison with other studies, the Potential for Individuality
Coding (PIC) was calculated for each measured call characteristic (Robisson et al., 1993;
Lengagne et al., 1997; Bee and Gerhardt, 2001; Charrier et al., 2001a, 2002b, 2003b;
Mathevon et al., 2003). The PIC is the ratio of the between-individual coefficient of
variation (CVb) and the mean value of the within-individual coefficients of variation
(CVi). Coefficients of variation were calculated by dividing the standard deviation by the
mean. PIC values greater than one indicate the potential for individual recognition since
the between-individual variability is greater than the within-individual variability
(Robisson et al., 1993; Lengagne et al., 1997).
15
Sex Differences
Calls were tested for differences attributable to the sex of the vocalizing animal.
Stepwise cross-validation DFA was conducted on a subset of the data containing 1500
calls to classify vocalizations based on the sex of the calling pup using SPSS 10.0 (SPSS
Inc., Chicago, IL). Five pups of each sex with large sample sizes were chosen for this
analysis, and 150 calls were randomly selected from each pup. To meet the assumptions
of this test, continuous variables were tested for normality and log transformed as
necessary. Chi-square was used to test whether correct classification was significantly
greater than expected by chance. Mixed-effects linear regression (LME), which is
discussed further under ‘Age Differences’, was used to identify which variables were
significantly different between males and females.
16
Age Differences
Differences in vocalizations attributable to age were examined using mixed-effects linear
regression (LME) with the nested repeated measure (or random grouped effect) of
‘recording session within id’, in S-Plus 2000 statistical software (Data Analysis Products
Division. MathSoft. Seattle, WA; Pinheiro and Bates, 2000). Regression diagnostics
confirmed that the variables followed normality and variance assumptions. The
covariates or fixed effects included sex and age (in weeks) of the vocalizing pup. The
nested repeated measure of ‘recording session within id’ accounted for any clustering due
to individuals or recording session, and was based on the assumption that there would be
more variability in call characteristics of the individuals across recording sessions than
there would be in recording sessions across individuals.
17
RESULTS
General Call Characteristics
The total data set included 4593 vocalizations recorded from 15 harbor seal pups. Each
individual pup was represented by a minimum of 60 calls; however the age at which each
seal was recorded varied considerably (Table 3). Most calls were relatively tonal having
an inverted “v” or “u” shaped spectrogram with the fundamental frequency around 200-
600 Hz and harmonics (Figure 1). See Table 4 for descriptive statistics on measured call
characteristics of harbor seal pup vocalizations recorded during this study.
Table 3. Summary of analyzed calls per individual harbor seal pup, Phoca vitulina richardii, by age.
Age
(weeks)
Spunky
Sundae
Sweetum
Finklestein
Benton
Dillon Baby
Eucie
Davranger
Bell
Brooks
Towey
Burley
Mott
Lanny
Spotti
1 720 7 - 37 - 114 42 7 - - 11 59 45 140 -
2 305 147 82 37 41 64 11 13 - - 76 5 32 15 6
3 52 151 45 22 37 63 136 8 2 20 6
- 48 13 134
4 - 35 7 33 33 105 39 17 37 51 34 - 11 6 16
5 - 100 5 10 158 76 49 7 71 54 27 41 17 --
6 - 95 7 79 81 46 21 7 48 24 31 60 - - -
7 - 62 1 31 9 94 4 1 12 8 - - - - -
8 - 17 - 21740 - - - - - - - - -
9 - 4 - - - - - - - - - - - - -
Total 1077 618 147 251 376 602 302 60 170 157 185 165 153 174 156
18
Table 4. Descriptive statistics of harbor seal pup, Phoca vitulina richardii,” Mother Attraction Calls”
measured by following the dominant frequency of the call (n=4593). Frequency measurements are
expressed in Hz and time in milliseconds.
Minimum Maximum Mean SD CV
Start Frequency 102.63 8067.20 435.44 371.52 0.85
Finish Frequency 103.44 7315.00 457.63 304.93 0.67
Minimum Frequency 102.63 1145.10 237.47 80.36 0.34
Maximum Frequency 343.27 8595.80 1224.88 572.72 0.47
Mean Frequency 255.05 1845.95 661.32 208.58 0.32
Freq at Peak Amplitude 180.73 3106.10 805.39 348.76 0.43
Frequency Range 33.71 8411.50 987.41 574.19 0.58
Max Freq / Mean Freq 1.04 10.51 1.89 0.71 0.38
Mean Freq / Min Freq 1.05 11.80 3.04 1.31 0.43
Duration 117.19 2444.04 610.73 203.70 0.33
Location of Min Freq 0.00 1.00 0.47 0.44 0.95
Location of Max Freq 0.00 1.00 0.47 0.24 0.52
Start Frequency Slope -6.42 3.48 0.17 0.30 -
Middle Frequency Slope -3.48 1.49 -0.01 0.22 -
Finish Frequency Slope -2.33 7.39 -0.16 0.27 -
Coefficient of Freq Mod 0.04 3.60 0.59 0.35 0.59
Coefficient of Variation 0.57 3836.37 177.30 174.87 0.99
Jitter Factor 2.08 67.50 15.31 7.57 0.49
19
6
4
2
0 1 2 3 4
Frequency (kHz)
Time (s)
Figure 1. Example spectrogram of harbor seal, Phoca vitulina richardii, pup vocalizations with inverted
“v” or “u” shape having a fundamental frequency around 200-600 Hz and harmonics.
20
Aggressive Vocalizations
Nineteen of the 4593 calls analyzed were harsh, broadband, staccato calls used in an
aggressive context toward another seal pup or human (Table 5, Figures 2, 3). Most of
these aggressive calls (n=15) were recorded from one individual, Davranger. Aggressive
vocalizations were recorded in a sequence of one to three calls and were produced by
both male and female pups. The major energy in these vocalizations was from 200-2000
Hz, and duration ranged from 0.12 to 1.07 seconds (mean = 0.57 sec).
6
4
2
0 1 2 3 4
Frequency (kHz)
Time (s)
Figure 2. Example spectrogram of three harsh, broadband, staccato calls used in an aggressive context by
harbor seal pups, Phoca vitulina richardii (compare with Figure 1).
21
Table 5. Descriptive statistics of calls used in an aggressive context by harbor seal pups, Phoca vitulina
richardii, measured by following the dominant frequency of the call (n=19). Frequency measurements are
expressed in Hz and time in milliseconds.
Minimum Maximum Mean SD CV
Start Frequency 155.31 1109.80 404.56 286.77 0.71
Finish Frequency 122.61 1357.40 595.08 472.36 0.79
Minimum Frequency 122.61 333.35 180.33 48.19 0.27
Maximum Frequency 890.45 7170.70 1899.23 1392.47 0.73
Mean Frequency 552.58 1323.65 944.34 215.86 0.23
Freq at Peak Amplitude 278.12 1479.20 915.33 350.89 0.38
Frequency Range 557.10 7004.37 1718.90 1401.96 0.82
Maximum Freq / Mean Freq 1.23 6.33 2.01 1.22 0.61
Mean Freq / Minimum Freq 2.18 9.26 5.49 1.62 0.29
Duration 117.19 1065.96 573.12 298.26 0.52
Location of Minimum Freq 0.00 1.00 0.64 0.41 0.64
Location of Maximum Freq 0.08 0.98 0.63 0.27 0.42
Start Frequency Slope 0.00 2.66 0.60 0.71 -
Middle Frequency Slope -2.57 0.40 -0.23 0.69 -
Finish Frequency Slope -1.67 2.86 -0.19 0.95 -
Coefficient of Freq Mod 0.34 3.21 1.12 0.67 0.60
Coefficient of Variation 30.36 1075.45 251.28 267.22 1.06
Jitter Factor 6.71 48.02 20.42 11.48 0.56
22
30000
0
-30000
0 0.5
Amplitude
Time (s)
Figure 3. Example waveform showing the staccato nature of a call used in an aggressive context by a
harbor seal pup, Phoca vitulina richardii.
23
Individual Differences
Discriminant Function Analysis
The results of stepwise, cross-validation discriminant function analysis (DFA) to classify
individual harbor seal pups are presented in Table 6. The nineteen calls identified as
aggressive vocalizations were excluded from this analysis, resulting in a sample size of
4574 calls. Even with all age groups combined, 29.3% of calls were classified correctly,
which is significantly more than the 11.5% correct expected by chance (Χ2, P<0.0001).
Classification scores for individuals ranged from 3.6% to 53.0% correct, and were
significantly greater than expected by chance in 11 of the 15 harbor seals (Χ2, P<0.0001).
Table 6. Results of cross-validation discriminant function analysis to classify individual harbor seal pups,
Phoca vitulina richardii, based on call characteristics measured by following the dominant frequency.
Bold numbers are calls correctly classified; others are incorrectly classified (overall 29.3% correct, n =
4574). Chi-square tested whether correct classification was significantly greater than expected by chance.
Asterisk indicates significance using Holm’s sequential Bonferroni method to adjust alpha levels for
multiple comparisons (Rice, 1989).
Predicted Group Membership % %
ID 1 2 3 4 5 6 7 8 9101112131415CorrectChance Χ2
1 46 3 11 4 7 1 2 16 14 5 29 9 19 3 1 27.1% 3.7% <0.0001 *
2 2 109 27 1 39 16 37 34 21 4 21 13 23 5 24 29.0% 8.2% <0.0001 *
3 4 4
63 280 71143162072640.1%3.4% <0.0001*
4 26 8 26
68 4 8 24 4 1 15 16 7 2 10 3.6% 3.6% 0.9833
5 1 8 4 0 12 0 23035132126.7%1.0% <0.0001*
6 37 57 38 6 54 49 26 52 50 36 67 45 39 25 21 8.1% 13.2% 0.0003 §
7 15 55 29 3 21 11
35 21 15 6 35 13 19 5 19 11.6% 6.6% 0.0005 *
8 14 20 6 5 29 17 18 42 14 724122710 416.9% 5.4% <0.0001*
9 74 25 653 67
64 182367236.8%3.8% <0.0001*
10 5 2 15 0 30 4 4 17 7 910 16 11 18 4 5.9% 3.3% 0.0738
11 11 3 13 1 9 3 5 3 6 3 81 11 0 1 6 51.9% 3.4% <0.0001 *
12 13 17 145 1 13 31 14 20 101 12 32 571 59 35 13 53.0% 23.5% <0.0001 *
13 41 26 9 5 34 10 21 55 67 25 19 36 225 31 13 36.5% 13.5% <0.0001 *
14 1 16 9 0 17 2 2 9 6 14 13 16 21 18 3 12.2% 3.2% <0.0001 *
15 1 31 26 1 26 7 13 8 4 6 22 6 14 8 12 6.5% 4.0% 0.0916
§ The significance of the chi-square result in this case indicates that calls were correctly classified to this
individual significantly less often than would be expected by chance.
24
The DFA analysis generated ten significant canonical discriminant functions representing
the linear combinations of the variables that maximally separate groups in
multidimensional space (Table 7). The first canonical discriminant function accounted
for 52.5% of the variation between individuals and loaded most heavily with duration and
jitter factor. The second discriminant function, which accounted for 19.3% of the
variation, loaded most heavily with mean frequency, the mean frequency divided by the
minimum frequency, and call duration.
Table 7. Correlations between discriminating variables and standardized canonical discriminant functions
used in cross-validation discriminant function analysis to classify individual harbor seal pups, Phoca
vitulina richardii, based on measured call characteristics (overall 29.3 % correct, n = 4574). Bold numbers
represent the largest absolute correlation between each variable and any discriminant function.
Function
1234567 8910
Mean Frequency .37 -.67 .00 -.05 .61 .14 .16 -.02 -.05 -.03
Mean Freq / Minimum Freq .26 -.63 .05 -.49 -.05 .38 .24 -.17 .11 .22
Duration -.57 -.57 .17 .25 -.38 -.23 .24 -.02 -.01 -.03
Inflection Factor -.38 .23 .69 -.18 .13 .07 .19 -.40 -.24 -.10
Finish Frequency -.15 .03 -.12 .35 .62 .08 .20 -.24 .59 .02
Coefficient of Variation .08 -.10 -.20 -.46 -.02 -.14 .82 .10 -.10 -.15
Maximum Freq / Mean Freq -.10 .24 -.28 .02 -.25 .22 .72 -.07 -.27 -.38
Maximum Frequency .17 -.21 -.30 .12 .23 .28 .69 -.21 -.42 .03
Jitter Factor -.53 -.01 -.19 -.43 .16 .26 .58 .14 -.16 -.18
Coefficient of Freq Mod -.25 -.35 -.18 -.39 .44 .27 .55 .08 -.17 -.17
Eigenvalue 0.54 0.20 0.11 0.07 0.06 0.03 0.02 0.01 0.00 0.00
% of Variance 52.5 19.3 10.3 7.1 5.6 2.5 1.8 0.5 0.3 0.1
Cumulative % 52.5 71.8 82.1 89.2 94.8 97.2 99.1 99.6 99.9 100.0
25
Fundamental Frequency DFA Analysis
Since the fundamental frequency has been implicated as an individually distinctive
feature in the vocalizations of harbor seal pups (Renouf, 1984; Perry and Renouf, 1988),
cross-validation DFA was also run on a subset of the data containing 40 calls each from
six pups during three different age groups (n = 240 per age group) for which calls were
filtered and measured from the fundamental frequency only. This analysis enabled a
comparison of individuality as a function of age. Results are presented in Table 8.
Overall percent correct classification scores were significantly greater than expected by
chance for each of the three age groups as well as all ages combined (Χ2, P<0.0001).
Classification scores using measurements made from just the fundamental frequency
were higher than classification scores using measurements made by following the
dominant frequency of the call.
Table 8. Results of cross-validation discriminant function analysis to classify individual harbor seal pups,
Phoca vitulina richardii, using a subset of the data containing 40 calls each from six pups during three
different age groups (n = 240 per age group) measured from the fundamental frequency alone. The ‘%
Correct’ column contains percentages of calls classified correctly. Chi-square tested whether correct
classification was significantly greater than the 16.67% expected by chance. Asterisk indicates significance
using Holm’s sequential Bonferroni method to adjust alpha levels for multiple comparisons (Rice, 1989).
0-14 Days Old 15-28 Days Old 29-42 Days Old All Ages Combined
ID % Correct Χ2 % Correct Χ2 % Correct Χ2 % Correct Χ2
2 72.5% <0.0001 * 60.0% <0.0001* 80.0% <0.0001* 74.2% <0.0001*
6 15.0% 0.7773 42.5% <0.0001* 42.5% <0.0001* 30.8% <0.0001*
7 55.0% <0.0001 * 42.5% <0.0001* 70.0% <0.0001* 50.0% <0.0001*
8 40.0% 0.0001 * 42.5% <0.0001* 30.0% 0.0237 38.3% <0.0001*
13 67.5% <0.0001 * 52.5% <0.0001 * 57.5% <0.0001* 54.2% <0.0001*
15 15.0% 0.7773 47.5% <0.0001* 62.5% <0.0001* 39.2% <0.0001*
Overall
44.2%
<0.0001 * 47.9% <0.0001
* 57.1% <0.0001
*
47.8% <0.0001
*
26
The fundamental frequency analysis for all age groups combined generated five
significant canonical discriminant functions (Table 9). The first function accounted for
45.7% of the variation between individuals and loaded most heavily with mean frequency
and finish frequency. The second discriminant function, which accounted for 29.6% of
the variation, loaded most heavily with maximum frequency and the coefficient of
frequency modulation. The third function accounted for an additional 16.2% of the
variation between individuals and loaded most heavily with duration and coefficient of
variation.
Table 9. Correlations between discriminating variables and standardized canonical discriminant functions
used in cross-validation discriminant function analysis to classify individual harbor seal pups, Phoca
vitulina richardii, based on call characteristics measured from the fundamental frequency for all age groups
combined (overall 47.8 % correct, n = 720). Bold numbers represent the largest absolute correlation
between each variable and any discriminant function.
Function
1234 5
Mean Frequency 0.62 0.50 0.06 -0.06 0.22
Finish Frequency 0.52 -0.01 -0.52 0.34 0.19
Maximum Frequency 0.29 0.70 0.30 0.26 0.02
Coefficient of Freq Mod -0.36 0.66 0.23 0.06 0.29
Duration -0.21 -0.38 0.76 0.01 0.10
Coefficient of Variation -0.23 0.48 0.63 -0.05 -0.35
Maximum Freq / Mean Freq -0.31 0.50 0.53 0.54 -0.07
Eigenvalue 0.467 0.302 0.165 0.069 0.018
% of Variance 45.7 29.6 16.2 6.8 1.7
Cumulative % 45.7 75.3 91.5 98.3 100.0
27
Potential for Individuality Coding
Measured call characteristics from the vocalizations of harbor seal pups recorded in this
study had Potential of Individuality Coding (PIC) values ranging from 0.98 to 1.18
(Table 10).
Table 10. Potential for Individuality Coding (PIC) for measured call characteristics of harbor seal pups,
Phoca vitulina richardii. The PIC is the ratio of the between-individual coefficient of variation (CVb) and
the mean value of the within-individual coefficients of variation (CVi). Coefficients of variation were
calculated by dividing the standard deviation by the mean.
Variable Mean ± SD CVbMean CViPIC
Start Frequency 435.44 ±
371.52 0.85 0.76 1.12
Finish Frequency 457.63 ± 304.93 0.67 0.61 1.10
Minimum Frequency 237.47 ± 80.36 0.34 0.33 1.03
Maximum Frequency 1224.88 ± 572.72 0.47 0.44 1.06
Mean Frequency 661.32 ± 208.58 0.32 0.29 1.12
Freq at Peak Amplitude 805.39 ± 348.76 0.43 0.40 1.08
Frequency Range 987.41 ± 574.19 0.58 0.55 1.05
Maximum Freq / Mean Freq 1.89 ± 0.71 0.38 0.36 1.04
Mean Freq / Minimum Freq 3.04 ± 1.31 0.43 0.41 1.04
Duration 610.73 ± 203.70 0.33 0.29 1.15
Location of Min Frequency 0.47 ± 0.44 0.95 0.97 0.98
Location of Max Frequency 0.47 ± 0.24 0.52 0.51 1.02
Coefficient of Freq Mod 0.59 ± 0.35 0.59 0.58 1.01
Coefficient of Variation 177.30 ± 174.87 0.99 0.84 1.18
Jitter Factor 15.31 ± 7.57 0.49 0.49 0.99
28
Sex Differences
Discriminant Function Analysis
Cross-validation discriminant function analysis (DFA) was conducted on a subset of the
data containing 1500 calls (150 from each of five male and five female pups) to classify
vocalization based on the sex of the calling pup. With all age groups combined, 66.1% of
calls were classified correctly, which is significantly more than the 50% correct expected
by chance (Χ2, P<0.0001). The analysis generated one significant canonical discriminant
function that accounted for all of the variation between sexes, and loaded most heavily
with duration, inflection factor, and mean frequency.
Mixed-Effects Linear Regression
The results of repeated measures linear regression (LME) with the random effect of
‘recording session within id’ are presented in Table 11. When the sex of the vocalizing
pup was the only fixed effect considered, only one of the summary acoustic variables was
significantly different between males and females: the coefficient of frequency
modulation. Male harbor seal pup calls had higher coefficients of frequency modulation
than did those of female pups.
29
However, when the interaction between the sex and age of the vocalizing pup was
included in the LME, several other variables revealed significant differences: mean
frequency, frequency at peak amplitude, duration, location of maximum frequency,
coefficient of variation, and jitter factor (Table 11). See Figure 4 for graphs generated
from the LME models demonstrating the interaction between the age and sex of the
vocalizing pup. Young harbor seal pups of both sexes tended to have similar values for
mean frequency, but as the pups matured, female pup vocalizations became higher in
mean frequency while male pup vocalizations became lower in mean frequency.
Similarly, young pups of both sexes had similar values for frequency at peak amplitude,
but as pups matured, female pup vocalizations became higher while male pup
vocalizations became lower. The mean duration of female pup vocalizations remained
relatively constant throughout the study period, but male pups exhibited an increase in the
duration of their calls as they matured. The location of maximum frequency, coefficient
of variation, and jitter factor values were all higher in female pup calls during the first
few weeks of age, becoming gradually lower in females and higher in males so that male
values were higher than females in pups over five weeks of age.
30
Table 11. Results of linear mixed effects regression with the random effect of ‘recording session within id’
on acoustic variables describing harbor seal pup, Phoca vitulina, vocalizations. Covariates include: fixed
effect of ‘age’ (as a continuous variable), fixed effect of ‘sex’, and the interaction between the fixed effects
of ‘age’ and ‘sex’ (* indicates significance at the alpha <0.05 level). Female is referent.
Age Sex Sex & Age
Coefficient p (df=298) Coefficient p (df=13) Coefficient p (df=298)
Minimum Frequency -4.7697 <0.0001 * -1.2063 0.7364 -0.0119 0.9876
Maximum Frequency -19.402 0.0044* -50.2890 0.2133 -1.244 0.8542
Mean Frequency 2.1601 0.5246 -9.9786 0.6189 -9.4408 0.0057 *
Freq at Peak Amplitude 5.4630 0.2370 -16.8019 0.5283 -12.2869 0.0081 *
Frequency Range -16.0775 0.0190 * -49.0871 0.2334 1.1875 0.8618
Duration -0.7073 0.7815 -33.1473 0.1145 9.3709 0.0003 *
Location of Min Freq -0.0019 0.0025* -0.0003 0.9116 0.0002 0.7532
Location of Max Freq -0.0128 0.0001 * -0.0287 0.1079 0.0074 0.0195*
Start Frequency Slope -0.0003 0.9544 -0.0025 0.9311 -0.0018 0.7161
Middle Frequency Slope -0.0045 0.0772 -0.0189 0.1095 0.0022 0.3942
Finish Frequency Slope -0.0052 0.1908 -0.0029 0.9025 0.0061 0.1308
Coefficient of Freq Mod -0.0182 <0.0001 * -0.0405 0.0479* 0.0047 0.1689
Coefficient of Variation -5.3022 <0.0001 * -11.1505 0.1239 2.5226 0.0472 *
Jitter Factor -0.4796 <0.0001* -0.7555 0.1422 0.2358 0.0034 *
31
Mean Frequency
450
550
650
12345678
Pup Age (weeks)
MEAN (Hz)
Frequency at Peak Amplitude
450
550
650
12345678
Pup Age (weeks)
PAF (Hz)
Duration
450
500
550
12345678
Pup Age (weeks)
DUR (msec)
Location of Maximum Frequency (% of call)
0.3
0.35
0.4
0.45
12345678
Pup Age (weeks)
MAXL (%)
Coefficient of Variation
60
90
120
12345678
Pup Age (weeks)
CV
Jitter Factor
6
9
12
12345678
Pup Age (weeks)
JF
Figure 4. Graphs generated from repeated measures linear regression demonstrating the interaction
between the fixed effects of ‘age’ and ‘sex’ for variables describing harbor seal pup, Phoca vitulina
richardii, vocalizations that were significant (α<0.05) with the random effect of ‘recording session within
id’. Males are represented by squares and females are represented by triangles.
32
Age Differences
In addition to maturational changes in male and female pup calls (See Sex Differences
section above), there were also significant age differences in pup calls regardless of sex.
When the age of the pup was the only fixed effect considered in the repeated measures
linear regression (LME), the following summary acoustic variables remained significant:
location of the maximum frequency, coefficient of variation, and jitter factor (Table 11,
Figure 4). In addition, the following variables decreased significantly with age in pups of
both sexes: minimum frequency, maximum frequency, frequency range, location of the
minimum frequency, and coefficient of frequency modulation (Table 11). See Figure 5
for graphs generated from the LME models demonstrating the effect of age. Mother
attraction calls were stereotyped in the youngest harbor seals recorded, estimated to be
only 2 days old (see Figure 6). The variation both within and between individuals over
time is also demonstrated in Figure 6.
33
Minimum Frequency
150
175
200
12345678
Pup Age (weeks)
MIN (Hz)
Maximum Frequency
850
1000
1150
12345678
Pup Age (weeks)
MAX (Hz)
Frequency Range
600
750
900
12345678
Pup Age (weeks)
FR (Hz)
Location of Minimum Frequency (% of call)
0.015
0.025
0.035
12345678
Pup Age (weeks)
MINL (%)
Coefficient of Frequency Modulation
0.25
0.35
0.45
12345678
Pup Age (weeks)
COFM
Figure 5. Graphs generated from repeated measures linear regression demonstrating the fixed effect of
‘age’ for variables describing harbor seal pup, Phoca vitulina richardii, vocalizations that were significant
(α<0.05) with the random effect of ‘recording session within id’. Males are represented by squares and
females are represented by triangles.
34
4 days old
#8 Finklestein (F)
15 days old
32 days old
3 days old
#10 Mott (M)
16 days old
31 days old
4 days old
#13 Sundae (F)
15 days old
31 days old
3 days old
#15 Towey (M)
14 days old
31 days old
Figure 6. Spectrograms of harbor seal, Phoca vitulina richardii, pup vocalizations demonstrating variation
both within and between individuals. The identification number, name, and sex are given for each seal.
Each series of three calls shown is four seconds in duration and ranges from 0-6 kHz.
35
DISCUSSION
General Call Characteristics
Most calls produced by harbor seal pups during this study were relatively tonal, having
an inverted “v” or “u” shaped spectrogram with the fundamental frequency around 200-
600 Hz and harmonics. These calls had acoustic features similar to those commonly
described in the harbor seal literature and presumed to function as “Mother Attraction
Calls” (Scheffer and Slipp, 1944; Bishop, 1967; Newby, 1973; Renouf, 1984; Ralls et al.,
1985; Perry and Renouf, 1988). Pup vocalizations of the western Atlantic harbor seal,
Phoca vitulina concolor, have a fundamental frequency at about 350 Hz and harmonics
(Ralls et al., 1985). The mean (± SD) duration of calls in this study (0.57 ± 0.30 s) falls
between the reported values for the Atlantic subspecies of 0.31 ± 0.14 s (Perry and
Renouf, 1988), 0.81 ± 0.19 s (Ralls et al., 1985), and 1.1 ± 0.7 s (Van Parijs and Kovacs,
2002). Perry and Renouf (1988) found that each call had 1 to 12 parallel harmonic bands
occurring between 50 and 4150 Hz.
Aggressive Vocalizations
Nineteen of the calls analyzed in this study were harsh, broadband, staccato calls used in
an aggressive context towards another seal pup or human. There is little mention of
aggressive sounds in the main body of literature concerning harbor seal pup vocalizations
(Newby, 1973; Renouf, 1984; Ralls et al., 1985; Perry and Renouf, 1988). Instead the
focus of these studies has been “Mother Attraction Calls”, which are the most
36
predominant call type produced by harbor seal pups. However, an early study of harbor
seals mentions that pups would “growl” or “hiss” when approached by humans (Bishop,
1967), and a more recent study reported a few instances of pups producing a “growl” in
an agonistic context (Van Parijs and Kovacs, 2002). The aggressive vocalizations
recorded during this study also resembled the description of “growling” given by Sullivan
(1982) as “an aggressive, largely close-range threat consisting of a harsh, throaty, belch-
like or gutteral vocalization”. Although systematic behavioral observations were not
made, the posture assumed by pups vocalizing aggressively was consistent with that
described by Sullivan (1982): “A growling seal oriented its muzzle toward an intruder
and growled with the mouth partially open, vibrissae erect and oriented forward, and
upper lip wrinkled.” Sullivan (1982) observed “growling” by all age classes, including
pups. The earliest age at which a pup was recorded vocalizing aggressively in this study
was 16 days. Young pups have been reported to growl aggressively in other pinniped
species as well (harp seals, Phoca groenlandica: Kovacs, 1987; Northern fur seal pups,
Callorhinus ursinus: Lisitsyna, 1973). Further research is needed to explore aggressive
behavior in harbor seal pups.
37
Individual Differences
The results of both discriminant function analysis and the potential of individuality
coding indicate that harbor seal pup calls are somewhat individualized. Individually
distinctive contact calls have been found in all species of pinnipeds studied to date (for a
review, see Insley et al., 2003b). Classification scores from discriminant function
analysis for individual harbor seal pups in this study ranged from an overall 29.3-57.1%
correct. The significance of percent correct classification scores depends on the sample
size, and it is therefore difficult to make meaningful comparisons between studies.
Percent correct classification scores have been shown to increase corresponding to a
decrease in the number of individuals and the number of signals per individual (Bee et
al., 2001). Research on the individuality of pup contact calls in pinnipeds has found a
wide range of percent correct classification scores (79% for northern fur seals,
Callorhinus ursinus: Insley, 1992; 32% for grey seals, Halichoerus grypus: McCulloch et
al., 1999; 64% for northern elephant seals, Mirounga angustirostris: Insley, 1992; 14%
for Hawaiian monk seals, Monachus schauinslandi: Job et al., 1995; 89% for southern
sea lions, Otaria flavescens: Fernandez-Juricic, 1999), but whether this reflects species
differences in stereotypy or is an artifact of sample size differences is unknown.
Harbor seal pups could be discriminated based on measured call characteristics
significantly more often than expected by chance during each of the three age groups as
well as when all ages were combined, suggesting that vocalizations are individually
38
distinctive even when pups are less than two weeks old. Classification scores using
measurements of the fundamental frequency were higher than classification scores using
measurements made from the dominant frequency of the call. This result is consistent
with previous findings that the fundamental frequency of contact calls is a reliable
indicator of identity in harbor seals (Renouf, 1984; Perry and Renouf, 1988) and other
mammals (South American fur seals, Arctocephalus australis: Phillips and Stirling, 2000;
subantarctic fur seals, Arctocephalus tropicalis: Charrier et al., 2002b; northern fur seals,
Callorhinus ursinus: Insley, 1992; grey seals, Halichoerus grypus: McCulloch et al.,
1999; northern elephant seals, Mirounga angustirostris: Insley, 1992; Hawaiian monk
seals, Monachus schauinslandi: Job et al., 1995; southern sea lions, Otaria flavescens:
Fernandez-Juricic et al., 1999; baboons, Papio hamadrayas: Rendall, 2003; Amazonian
manatees, Trichechus inunguis: Sousa-Lima et al., 2002). Duration has also been
implicated as an individually distinctive feature of contact calls in mammals (South
American fur seals, Arctocephalus australis: Phillips and Stirling, 2000; northern fur
seals, Callorhinus ursinus: Insley, 1992; grey seals, Halichoerus grypus: McCulloch et
al., 1999; northern elephant seals, Mirounga angustirostris: Insley, 1992; Hawaiian monk
seals, Monachus schauinslandi: Job et al., 1995; southern sea lions, Otaria flavescens:
Fernandez-Juricic et al., 1999; baboons, Papio hamadrayas: Rendall, 2003; Mexican
free-tailed bats, Tadarida brasiliensis: Gelfand and McCracken 1986; Amazonian
manatees, Trichechus inunguis: Sousa-Lima et al., 2002) and appears to be somewhat
stereotyped in harbor seal pup calls.
39
The Potential for Individuality Coding (PIC) is a means of comparison across studies
with different methodologies; PIC values greater than one indicate the potential for
individual recognition since the between-individual variability is greater than the within-
individual variability (Robisson et al., 1993; Lengagne et al., 1997). The PIC ratios for
call characteristics measured in this study are all very close to one, which suggests that
the vocalizations of harbor seal pups, Phoca vitulina richardii, are weakly individualized.
Mother attraction calls from subantarctic fur seal pups, Arctocephalus tropicalis, had PIC
ratios from 0.89 to 3.36 with the highest values for fundamental frequency, frequency at
the first peak amplitude, and frequency modulation (Charrier et al., 2002b). Playback
experiments confirmed that these calls are individually distinctive and that mothers
respond selectively to the calls of their own pup (Charrier et al., 2002b). Call duration
(DUR) was the only variable that could be directly compared to the present study and
appears to be more individualized in the vocalizations of harbor seal pups (1.15) than in
subantarctic fur seal pups (0.89).
40
Sex Differences
Discriminant function analysis classified calls correctly to the sex of the vocalizing pup
significantly more often than expected by chance. Repeated measures linear regression
indicated that male harbor seal pup calls had higher coefficients of frequency modulation
than calls of female pups. There were also significant interactions between the sex and
age of the vocalizing pup. Young pups of both sexes had similar values for mean
frequency, but as they matured, female calls became higher in mean frequency while
male calls became lower. Likewise, young animals of both sexes had similar values for
frequency at peak amplitude, but as they matured, females became higher while males
became lower. The mean duration of female vocalizations remained relatively constant
throughout the study period, but the mean duration of male calls increased as they
matured. The location of maximum frequency, coefficient of variation, and jitter factor
values were all higher in female pup calls during the first few weeks of age becoming
gradually lower in females and higher in males so that male values were highest in pups
over five weeks of age.
Sex differences in vocalizations may function as a means for parents to allocate resources
differentially based on offspring sex, and/or reflect early development of sexual
dimorphism in vocal behavior (Saino et al., 2003). There is no evidence for differential
resource allocation in harbor seals. The sex ratio is 1:1 both at birth and weaning (Ellis,
1998), and although male harbor seal pups do weigh more than females at weaning, this
41
reflects greater birth mass of males rather than differential maternal investment during
lactation (Ellis, 1998; Bowen et al., 2001). Sex differences in pup calls are also unlikely
to reflect early development of sexual dimorphism in vocal behavior, since this call
disappears from the vocal repertoire at weaning. However, sex differences in harbor seal
pup vocalizations may reflect early development of the sexual size dimorphism
characteristic of adults. This size dimorphism is already apparent at birth, with male pups
weighing more than females (Ellis, 1998). Body size is strongly correlated with the
length of the vocal tract in mammals (for a review, see Fitch, 1997), and according to the
source-filter theory of sound production, energy from the sound source is modified by
resonance characteristics of the vocal tract. The only study to test this relationship found
that formant frequencies are correlated with vocal tract length and body size in rhesus
macaques, Macaca mulatta (Fitch, 1997).
Age Differences
Harbor seal pup vocalizations appear to be stereotyped and encode identity information
from a young age, although call structure does change throughout maturation.
Discriminant function analysis classified calls correctly significantly more often than
expected by chance even when pups were less than two weeks old. As discussed under
‘Sex Differences’, age-related changes in many call characteristics varied according to the
sex of the vocalizing pup. The following variables decreased significantly over time in
pups of both sexes: location of the maximum frequency, coefficient of variation, jitter
42
factor, minimum frequency, maximum frequency, frequency range, location of the
minimum frequency, and coefficient of frequency modulation. These ontogenetic
changes in frequency are similar to those reported in other mammals (vervet monkey,
Cercopithecus aethiops: Hauser, 1989; pigtail macaque, Macaca nemestrina: Gouzoules
and Gouzoules, 1989), and likely reflect the gradual enlargement of the vocal tract.
There has not been much research on the ontogeny of pup calls in pinnipeds, probably
due to the difficulty of following known individuals throughout the lactation period.
California sea lion, Zalophus californianus, females may recognize vocalizations from
their two-week old pups, suggesting that the calls are individually stereotyped at this time
(Gisiner and Schusterman, 1991; Schusterman et al., 1992), although visual and olfactory
cues could not be ruled out in these studies. Research on subantarctic fur seal pups,
Arctocephalus tropicalis, showed significant age-related changes in the fundamental
frequency and the percentage of quavering in pup calls (Charrier et al., 2003a). Calls of
young pups emphasized some high frequencies, while the spectral energy in the calls of
older pups was concentrated on the first harmonics (Charrier et al., 2003a).
43
Directions for Future Research
Vocal Development of Harbor Seal Pups
Since this study was done on captive animals, the results should be interpreted with
caution, and further research is needed on wild harbor seal pups to verify these findings.
It is unknown which acoustic characteristics harbor seal mothers may use to recognize
the vocalizations of their own offspring. Therefore, call characteristics measured in this
study may not accurately reflect perceptually salient features, and the measurement of
additional or alternative variables might improve correct classification scores. Playback
studies using artificially manipulated signals such as those recently conducted by
Charrier and colleagues on black-headed gulls, Larus ridibundus, (Charrier et al., 2001b)
and subantarctic fur seals, Arctocephalus tropicalis, (Charrier et al., 2002b, 2003c) could
help to elucidate the cues used by harbor seals in voice discrimination. However, harbor
seal mothers do not emit a “pup attraction call” or make any other obvious response to
pup vocalizations, thus confounding the ability to demonstrate a differential response to
playbacks.
44
Effects of Motivational State
The present study did not examine the effects of motivational state on characteristics of
harbor seal pup vocalizations, although recordings were made under similar conditions in
an attempt to minimize the possible influence of motivation. Previous research suggested
that motivational state information might be encoded in the repetition rate and number of
harmonics of harbor seal pup vocalizations (Renouf, 1984; Perry and Renouf, 1988).
Preliminary observations suggest that distressed pups have a greater number of calls per
calling bout, a faster rate of call emission, and more harmonics (Renouf, 1984; Perry and
Renouf, 1988). Young animals have been shown to vocalize at a faster rate when
distressed or excited in other species of mammals (pallid bats, Antrozous pallidus:
Brown, 1976; baboons, Papio hamadrayas: Rendall, 2003; caribou, Rangifer tarandus:
Espmark, 1971) and birds (dunnock, Prunella modularis: Butchart et al., 2003; meadow
pipit, Anthus pratensis: Butchart et al., 2003). The acoustic structure of isolations calls
may also be influenced by motivational state. A study of guinea pig, Cavia porcellus,
pups found that a brief isolation period caused a decrease in call duration, increase in
mean frequency, and decrease in the number of harmonics (Monticelli et al., 2004).
Future research on harbor seal pup vocalizations should take motivational state
information into account (such as hunger, body temperature, physical condition, and
stimulation).
45
Vocal Recognition by Harbor Seal Mothers
Individual variation is a prerequisite for, but does not prove the existence of, individual
recognition. This study did not investigate the ability of female harbor seals to recognize
the voices of their pups. Communication involves the transfer of information from a
sender to a receiver, and the use of that information by the receiver to make a response
decision. If the sender and receiver both prefer the same response, adaptations will be
favored that enhance the provision of information (Bradbury and Vehrencamp, 1998). As
discussed in the ‘Introduction’, both harbor seal pups (senders) and mothers (receivers)
are under selective pressure to recognize and maintain contact with one another. Pups
benefit by having distinctive vocalizations to broadcast their identity (avoiding premature
separation that may result in aggression and starvation). Mothers benefit by being able to
recognize the calls of their own pups (increasing reproductive success by promoting the
survival of their limited number of offspring). Thus, there is reason to suspect that harbor
seal mothers recognize the calls of their pup. Preliminary research has demonstrated that
a captive adult female was able to distinguish between recordings of different pups
(Renouf, 1985). However, further research is needed to confirm that harbor seal mothers
recognize the vocalizations of their own pups under natural conditions in the wild.
Recognition would be difficult to demonstrate unambiguously using playback studies due
to the lack of an obvious behavioral response to pup calls. Harbor seals are extremely
wary and flush into the water when disturbed, so a less intrusive approach may be more
successful. Observational studies of natural separations could determine whether or not
46
successful reunions occur more quickly if the pup is vocal when the mother arrives at the
breeding rookery and begins her search.
Comparisons with Other Species and Populations
The harbor seal, Phoca vitulina, and the closely related spotted seal, Phoca largha,
provide a unique opportunity for comparative study of the effects of breeding density on
the vocal recognition system. The cost and reliability of recognition mechanisms are
hypothesized to scale to the risk of misallocation of parental care (Tyack, 1999).
Trillmich (1981) suggested that site recognition would likely be sufficient in ice-breeding
phocids, but that individual recognition would be favored in a dense colony. Therefore,
spotted seals, which breed on isolated ice floes, are expected to have a less reliable
system of mother-offspring recognition than harbor seals, which breed in crowded
rookeries where there is greater risk of confusion. The absence of mother-pup
recognition in spotted seals is suggested by a pup-exchange experiment in which a female
spotted seal accepted a strange pup (Burns et al., 1972). However, further research on
wild populations of both harbor seals and spotted seals is needed to test these predictions.
Comparative study of harbor seals breeding at different latitudes and on different
substrates (land or ice) may also shed light on the potential influence of these factors on
the recognition system.
47
REFERENCES
Balcombe, J. P. 1990. Vocal recognition of pups by mother Mexican free-tailed bats,
Tadarida brasiliensis mexicana. Animal Behaviour, 39, 960-966.
Balcombe, J. P., and McCracken, G. F. 1992. Vocal recognition in Mexican free-tailed
bats: do pups recognize mothers? Animal Behaviour, 43, 79-87.
Bartholomew, G. A. 1959. Mother-young relations and the maturation of pup behaviour
in the Alaska fur seal. Animal Behaviour, 7, 163-171.
Bartholomew, G. A., and Collias, N. E. 1962. The role of vocalization in the social
behavior of the northern elephant seal. Animal Behaviour, 10, 7-14.
Bee, M. A., and Gerhardt, H. C. 2001. Neighbour-stranger discrimination by territorial
male bullfrogs (Rana catesbeiana): I. Acoustic basis. Animal Behaviour, 62,
1129-1140.
Bee, M. A., Kozich, C. E., Blackwell, K. J., and Gerhardt, H. C. 2001. Individual
variation in advertisement calls of territorial male green frogs, Rana clamitans:
Implications for individual discrimination. Ethology, 107, 65-84.
Beecher, M. D. 1981. Development of parent-offspring recognition in birds. In:
Development of Perception: Psychobiological Perspectives (Ed. by Aslin, R. N.,
Alberts, J. R., and Peterson, M. R.), pp. 45-65. New York: Academic Press, Inc.
Beecher, M. D. 1991. Successes and failures of parent-offspring recognition in animals.
In: Kin Recognition (Ed. by Hepper, P. G.), pp. 94-124. Cambridge: Cambridge
University Press.
48
Beecher, M. D., Beecher, I. M., and Hahn, S. 1981a. Parent-offspring recognition in bank
swallows (Riparia riparia): II. Development and acoustic bias. Animal Behaviour,
29, 95-101.
Beecher, M. D., Beecher, I. M., and Lumpkin, S. 1981b. Parent-offspring recognition in
bank swallows (Riparia riparia): I. Natural history. Animal Behaviour, 29, 86-94.
Bishop, R. H. 1967. Reproduction, age determination, and behavior of the harbor seal,
Phoca vitulina L., in the Gulf of Alaska. pp. 121: University of Alaska.
Boness, D. J., and Bowen, W. D. 1996. The evolution of maternal care in pinnipeds.
BioScience, 46, 645-654.
Boness, D. J., Bowen, W. D., and Oftedal, O. T. 1994. Evidence of a maternal foraging
cycle resembling that of otariid seals in a small phocid, the harbor seal.
Behavioral Ecology and Sociobiology, 34, 95-104.
Bowen, W. D., Ellis, S. L., Iverson, S. J., and Boness, D. J. 2001. Maternal effects on
offspring growth rate and weaning mass in harbour seals. Canadian Journal of
Zoology, 79, 1088-1101.
Bowen, W. D., Oftedal, O. T., and Boness, D. J. 1992. Mass and energy transfer during
lactation in a small phocid, the harbor seal (Phoca vitulina). Physiological
Zoology, 65, 844-866.
Bradbury, J. W., and Vehrencamp, S. L. 1998. Principles of Animal Communication.
Sunderland, MA: Sinauer Associates, Inc.
49
Brown, P. 1976. Vocal communication in the pallid bat, Antrozous pallidus. Zeitschrift
fur Tierpsychologie, 41, 34-54.
Butchart, S. H. M., Kilner, R. M., Fuisz, T., and Davies, N. B. 2003. Differences in the
nestling begging calls of hosts and host-races of the common cuckoo, Cuculus
canorus. Animal Behaviour, 65, 345-354.
Burns, J. J., Ray, G. C., Fay, F. H., and Shaughnessy, P. D. 1972. Adoption of a strange
pup by the ice-inhabiting harbour seal, Phoca vitulina largha. Journal of
Mammalogy, 53, 594-598.
Charrier, I., Jouventin, P., Mathevon, N., and Aubin, T. 2001a. Individual identity coding
depends on call type in the South Polar Skua Catharacta mccormicki. Polar
Biology, 24, 378-382.
Charrier, I., Mathevon, N., Hassnaoui, M., Carraro, L., and Jouventin, P. 2002a. The
subantarctic fur seal pup switches its begging behaviour during maternal absence.
Canadian Journal of Zoology, 80, 1250-1255.
Charrier, I., Mathevon, N., and Jouventin, P. 2002b. How does a fur seal mother
recognize the voice of her pup? An experimental study of Arctocephalus
tropicalis. Journal of Experimental Biology, 205, 603-612.
Charrier, I., Mathevon, N., and Jouventin, P. 2003a. Fur seal mothers memorize
subsequent versions of developing pups' calls: adaptation to long-term recognition
or evolutionary by-product? Biological Journal of the Linnean Society, 80, 305-
312.
50
Charrier, I., Mathevon, N., and Jouventin, P. 2003b. Individuality in the voice of fur seal
females: an analysis study of the pup attraction call in Arctocephalus tropicalis.
Marine Mammal Science, 19, 161-172.
Charrier, I., Mathevon, N., and Jouventin, P. 2003c. Vocal signature recognition of
mothers by fur seal pups. Animal Behaviour, 65, 543-550.
Charrier, I., Mathevon, N., Jouventin, P., and Aubin, T. 2001b. Acoustic communication
in a black-headed gull colony: how do chicks identify their parents? Ethology,
107, 961-974.
Cottrell, P. E., Jeffries, S., Beck, B., and Ross, P. S. 2002. Growth and development in
free-ranging harbor seal (Phoca vitulina) pups from southern British Columbia,
Canada. Marine Mammal Science, 18, 721-733.
Davis, R. B., Herreid, C. F., II, and Short, H. L. 1962. Mexican free-tailed bats in Texas.
Ecological Monographs, 32, 311-346.
Ellis, S. L. 1998. Maternal effects on offspring traits from birth through weaning in the
harbour seal, Phoca vitulina. pp. 160. Halifax, Nova Scotia: Dalhousie University.
Espmark, Y. 1971. Individual recognition by voice in reindeer mother-young
relationship: field observations and playback experiments. Behaviour, 40, 295-
301.
Falls, B. J. 1982. Individual recognition by sound in birds. In: Acoustic Communication
in Birds. Volume 2: Song Learning and Its Consequences (Ed. by Kroodsma, D.
E., and Miller, E. H.), pp. 237-278. New York: Academic Press.
51
Fernandez-Juricic, E., Campagna, C., Enriquez, V., and Ortiz, C. L. 1999. Vocal
communication and individual variation in breeding South American sea lions.
Behaviour, 136, 495-517.
Fitch, W. T. 1997. Vocal tract length and formant frequency dispersion correlate with
body size in rhesus macaques. The Journal of the Acoustical Society of America,
102, 1213-1222.
Fogden, S. C. L. 1968. Suckling behaviour in the Grey seal (Halichoerus grypus) and the
Northern elephant seal (Mirounga angustirostris). Journal of Zoology, London,
154, 415-420.
Fogden, S. C. L. 1971. Mother-young behaviour at Grey seal breeding beaches. Journal
of Zoology, London, 164, 61-92.
Gelfand, D. L., and McCracken, G. F. 1986. Individual variation in the isolation calls of
Mexican free-tailed bat pups (Tadarida brasiliensis mexicana). Animal
Behaviour, 34, 1078-1086.
Gisner, R., and Schustermann, R. J. 1991. California seal lion pups play an active role in
reunions with their mothers. Animal Behaviour, 41, 364-366.
Gouzoules, H., and Gouzoules, S. 1989. Design features and developmental modification
of pigtail macaque, Macaca nemestrina, agonistic screams. Animal Behaviour, 37,
383-401.
52
Gubernick, D. J. 1981. Parental and infant attachment in mammals. In: Parental Care in
Mammals (Ed. by Gubernick, D. J., and Klopfer, P. H.), pp. 243-305. New York:
Plenum Press Inc.
Gustin, M. K., and McCracken, G. F. 1987. Scent recognition between females and pups
in the bat Tadarida brasiliensis mexicana. Animal Behaviour, 35, 13-19.
Halpin, Z. T. 1991. Kin recognition cues of vertebrates. In: Kin Recognition (Ed. by
Hepper, P. G.), pp. 220-258. Cambridge: Cambridge University Press.
Hanggi, E. B. 1992. The importance of vocal cues in mother-pup recognition in a
California sea lion. Marine Mammal Science, 8, 430-432.
Hauser, M. D. 1989. Ontogenetic changes in the comprehension and production of vervet
monkey (Cercopithecus aethiops) vocalizations. Journal of Comparative
Psychology, 103, 149-158.
Hepper, P. G. 1991. Recognizing kin: ontogeny and classification. In: Kin Recognition
(Ed. by Hepper, P. G.), pp. 259-288. Cambridge: Cambridge University Press.
Herman, L. M., Forestell, P. H., and Antinoja, R. C. 1980. The 1976/77 migration of
humpback whales into Hawaiian waters: composite description. Washington: U.S.
Marine Mammal Commission.
Insley, S. J. 1992. Mother-offspring separation and acoustic stereotypy: a comparison of
call morphology in two species of pinnipeds. Behaviour, 120, 103-121.
Insley, S. J. 2001. Mother-offspring vocal recognition in northern fur seals is mutual but
asymmetrical. Animal Behaviour, 61, 129-137.
53
Insley, S. J., Paredes, R., and Jones, I. L. 2003a. Sex differences in razorbill Alca torda
parent-offspring vocal recognition. The Journal of Experimental Biology, 206, 25-
31.
Insley, S. J., Phillips, A. V., and Charrier, I. 2003b. A review of social recognition in
pinnipeds. Aquatic Mammals, 29, 181-201.
Job, D. A., Boness, D. J., and Francis, J. M. 1995. Individual variation in nursing
vocalizations of Hawaiian monk seal pups, Monachus schauinslandi (Phocidae,
Pinnipedia), and lack of maternal recognition. Canadian Journal of Zoology, 73,
975-983.
Kibal'chich, A. A., and Lisitsina, T. Y. 1979. Some acoustical signals of calves of the
Pacific walrus. Zoologicheskii Zhurnal, 58, 1247-1249. [In Russian; English
translation on file at the National Oceanographic and Atmospheric Administration
National Marine Mammal Laboratory Library, Seattle, WA]
Kovacs, K. M. 1987. Maternal behaviour and early behavioural ontogeny of harp seals,
Phoca groenlandica. Animal Behaviour, 35, 844-855.
Lawson, J. W., and Renouf, D. 1985. Parturition in the Atlantic harbor seal, Phoca
vitulina concolor. Journal of Mammalogy, 66, 395-398.
Lawson, J. W., and Renouf, D. 1987. Bonding and weaning in harbour seals, Phoca
vitulina. Journal of Mammalogy, 68, 445-449.
54
Lengagne, T., Lauga, J., and Jouventin, P. 1997. A method of independent time and
frequency decomposition of bioacoustic signals: inter-individual recognition in
four species of penguins. Comptes Rendus de l'Academie des Sciences, Paris,
Sciences de la vie / Life Sciences, 320, 885-891.
Lenhardt, M. L. 1977. Vocal contour cues in maternal recognition of goat kids. Applied
Animal Ethology, 3, 211-219.
Lisitsyna, T. Y. 1973. The behavior and acoustical signals of the northern fur seal
Callorhinus ursinus on the rookery. Zoologicheskii Zhurnal, 52, 1220-1228. [In
Russian; English translation on file at the National Oceanographic and
Atmospheric Administration National Marine Mammal Laboratory Library,
Seattle, WA]
Mathevon, N., Charrier, I., and Jouventin, P. 2003. Potential for individual recognition in
acoustic signals: a comparative study of two gulls with different nesting patterns.
Comptes Rendus de l'Academie des Sciences, Paris, Biologies, 326, 329-337.
McArthur, P. D. 1979. Parent-young recognition in the pinon jay: mechanisms, ontogeny,
and survival value. Ph.D. thesis, Northern Arizona University, Flagstaff, Arizona.
McArthur, P. D. 1982. Mechanisms and development of parent-young vocal recognition
in the pinon jay (Gymnorhinus cyanocephalus). Animal Behaviour, 30, 62-74.
McCowan, B. 1995. A new quantitative technique for categorizing whistles using
simulated signals and whistles from captive bottlenose dolphins (Delphinidae,
Tursiops truncatus). Ethology, 100, 177-193.
55
McCowan, B., and Reiss, D. 2001. The fallacy of 'signature whistles' in bottlenose
dolphins: a comparative perspective of 'signature information' in animal
vocalizations. Animal Behaviour, 62, 1151-1162.
McCracken, G. F. 1984. Communal nursing in Mexican free-tailed bat maternity
colonies. Science, 223, 1090-1091.
McCracken, G. F. 1993. Locational memory and female-pup reunions in Mexican free-
tailed bat maternity colonies. Animal Behaviour, 45, 811-813.
McCracken, G. F., and Gustin, M. K. 1987. Batmom's daily nightmare. Natural History,
96, 66-73.
McCulloch, S., Pomeroy, P. P., and Slater, P. J. B. 1999. Individually distinctive pup
vocalizations fail to prevent allo-suckling in grey seals. Canadian Journal of
Zoology, 77, 716-723.
Monticelli, P. F., Tokumaru, R. S., and Ades, C. 2004. Isolation induces changes in
Guinea Pig Cavia porcellus pup distress whistles. Annals of the Brazilian
Academy of Sciences, 76, 368-372.
Newby, T. C. 1973. Observations on the breeding behavior of the harbor seal in the state
of Washington. Journal of Mammalogy, 54, 540-543.
Perry, E. A., and Renouf, D. 1988. Further studies of the role of harbour seal (Phoca
vitulina) pup vocalizations in preventing separation of mother-pup pairs.
Canadian Journal of Zoology, 66, 934-938.
56
Peterson, R. S., and Bartholomew, G. A. 1969. Airborne vocal communication in the
California sea lion, Zalophus californianus. Animal Behaviour, 17, 17-24.
Petrinovich, L. 1974. Individual recognition of pup vocalization by Northern elephant
seal mothers. Zeitschrift fur Tierpsychologie, 34, 308-312.
Phillips, A. V., and Stirling, I. 2000. Vocal individuality in mother and pup South
American fur seals, Arctocephalus australis. Marine Mammal Science, 16, 592-
616.
Phillips, A. V., and Stirling, I. 2001. Vocal repertoire of South American fur seals,
Arctocephalus australis: structure, function, and context. Canadian Journal of
Zoology, 79, 420-437.
Pinheiro, J. C., and Bates, D. M. 2000. Mixed-Effects Models in S and S-Plus. New York:
Springer.
Ralls, K., Fiorelli, P., and Gish, S. 1985. Vocalizations and vocal mimicry in captive
harbor seals, Phoca vitulina. Canadian Journal of Zoology, 63, 1050-1056.
Rendall, D. 2003. Acoustic correlates of caller identity and affect intensity in the vowel-
like grunt vocalizations of baboons. Journal of the Acoustical Society of America,
113, 3390-3402.
Renouf, D. 1984. The vocalization of the harbour seal pup (Phoca vitulina) and its role in
the maintenance of contact with the mother. Journal of Zoology, London, 202,
583-590.
57
Renouf, D. 1985. A demonstration of the ability of the harbour seal, Phoca vitulina (L.)
to discriminate among pup vocalizations. Journal of Experimental Marine
Biology and Ecology, 87, 41-46.
Renouf, D., and Diemand, D. 1984. Behavioral interaction between harbor seal mothers
and pups during weaning (Pinnipeds: Phocidae). Mammalia, 48, 53-58.
Renouf, D., Lawson, J. W., and Gaborko, L. 1983. Attachment between harbour seal
(Phoca vitulina) mothers and pups. Journal of Zoology, London, 199, 179-187.
Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution, 43, 223-225.
Robisson, P., Aubin, T., and Bremond, J.-C. 1993. Individuality in the voice of the
emperor penguin Aptenodytes forsteri: adaptation to a noisy environment.
Ethology, 94, 279-290.
Roux, J. P., and Jouventin, P. 1987. Behavioral cues to individual recognition in the
subantarctic fur seal, Arctocephalus tropicalis. In: Status, Biology, and Ecology of
Fur Seals (Ed. by Croxall, J. P., and Gentry, R. L.), pp. 83-94. Seattle,
Washington: National Marine Fisheries Service Technical Report No. 51.
Saino, N., Galeotti, P., Sacchi, R., Boncoraglio, G., Martinelli, R., and Moller, A. P.
2003. Sex differences in begging vocalizations of nestling barn swallows,
Hirundo rustica. Animal Behaviour, 66, 1003-1010.
Scheffer, V. B., and Slipp, J. W. 1944. The harbor seal in Washington State. The
American Midland Naturalist, 32, 373-416.
58
Schusterman, R. J., Hanggi, E. B., and Gisiner, R. 1992. Acoustic signalling in mother-
pup reunions, interspecies bonding, and affiliation by kinship in California sea
lions (Zalophus californianus). In: Marine Mammal Sensory Systems (Ed. by
Thomas, J. A., Kastelein, R. A., and Supin, A. Y.), pp. 533-551. New York:
Plenum Press.
Sousa-Lima, R. S., Paglia, A. P., Da Fonseca, G. A. B. 2002. Signature information and
individual recognition in the isolation calls of Amazonian manatees, Trichechus
inunguis (Mammalia: Sirenia). Animal Behaviour, 63, 301-310.
Sullivan, R. M. 1982. Agonistic behavior and dominance relationships in the harbor seal,
Phoca vitulina. Journal of Mammalogy, 63, 554-569.
Tabachnick, B. G., and Fidell, L. S. 2001. Using Multivariate Statistics. Boston, MA:
Allyn & Bacon.
Terhune, J. M. 1991. Masked and unmasked pure tone detection thresholds of a harbour
seal listening in air. Canadian Journal of Zoology, 69, 2059-2066.
Thompson, P. M., Fedak, M. A., McConnell, B. J., and Nicholas, K. S. 1989. Seasonal
and sex-related variation in the activity patterns of common seals (Phoca
vitulina). Journal of Applied Ecology, 26, 521-535.
Trillmich, F. 1981. Mutual mother-pup recognition in Galapagos fur seals and sea lions:
cues used and functional significance. Behaviour, 78, 21-42.
59
Tyack, P. L. 1999. Communication and Cognition. In: Biology of Marine Mammals (Ed.
by Reynolds, J. E., and Rommel, S. A.), pp. 287-323. Washington and London:
Smithsonian Institution Press.
Van Parijs, S. M., Hastie, G. D., and Thompson, P. M. 1999. Geographical variation in
temporal and spatial vocalization patterns of male harbour seals in the mating
season. Animal Behaviour, 58, 1231-1239.
Van Parijs, S. M., and Kovacs, K. M. 2002. In-air and underwater vocalizations of
eastern Canadian harbour seals, Phoca vitulina. Canadian Journal of Zoology, 80,
1173-1179.
Yochem, P. K., Stewart, B. S., DeLong, R. L., and DeMaster, D. P. 1987. Diel haul-out
patterns and site fidelity of harbor seals (Phoca vitulina richardsi) on San Miguel
Island, California, in autumn. Marine Mammal Science, 3, 323-332.
Zar, J. H. 1999. Biostatistical Analysis. Upper Saddle River: Prentice Hall.
60
