Models of honest advertisement predict that sexually selected calls should signal male quality. In most vertebrates, high quality males have larger body sizes that determine higher social status and in turn higher reproductive success. Previous research has emphasised the importance of vocal tract resonances or formant frequencies of calls as cues to body size in mammals. However, the role of the acoustic features of vocalisations as cues to other quality-related phenotypic characteristics of callers has rarely been investigated.
We examined whether the acoustic structure of fallow deer groans provides reliable information on the quality of the caller, by exploring the relationships between male quality (body size, dominance rank, and mating success) and the frequency components of calls (fundamental frequency, formant frequencies, and formant dispersion). We found that body size was not related to the fundamental frequency of groans, whereas larger males produced groans with lower formant frequencies and lower formant dispersion. Groans of high-ranking males were characterised by lower minimum fundamental frequencies and to a lesser extent, by lower formant dispersions. Dominance rank was the factor most strongly related to mating success, with higher-ranking males having higher mating success. The minimum fundamental frequency and the minimum formant dispersion were indirectly related to male mating success (through dominance rank).
Citation: Vannoni E, McElligott AG (2008) Low Frequency Groans Indicate Larger and More Dominant Fallow Deer (Dama dama) Males. PLoS ONE 3(9): e3113. https://doi.org/10.1371/journal.pone.0003113
Editor: David Reby, University of Sussex, United Kingdom
Received: May 22, 2008; Accepted: August 6, 2008; Published: September 3, 2008
Copyright: © 2008 Vannoni, McElligott. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was funded by grants from the Forschungskommission der Universität Zürich to A. G. McElligott, and from the Swiss Academy of Sciences to E. Vannoni. These organizations had no role in the study design, data collection and analysis or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Male vocalisations are often subject to sexual selection and can be used to assess the quality and condition of the caller in various vertebrates –. The reliability of the information encoded in sexually selected acoustic signals is maintained by constraints imposed on the caller, which make cheating costly . Alternatively, vocalisations that are physiologically or physically constrained carry some characteristics that are directly related to intrinsic properties of the caller and therefore cannot be faked .
The relationships between the body size of callers and the acoustic parameters of vocalisations are of particular interest in species in which body size determines fighting ability and reproductive success –. In toads, frogs, and birds, body size is negatively related to the fundamental frequency of calls –. In mammals, and within a given species, body size is related to fundamental frequency across age categories and among adult females but not among adult males , –. Fundamental frequency corresponds to the rate at which the vocal folds of the larynx open and close and is determined by different factors such as the length of the vocal folds, longitudinal stress on the vocal folds, and the tissue density of the vocal folds . For example, longer vocal folds result in lower fundamental frequency. Because the larynx is not constrained by the bones of the skull, the vocal folds may grow independently of the rest of the head or body . Moreover, vocal folds are highly sensitive to changes in testosterone ,  and they may grow longer in males with higher testosterone levels. Thus, fundamental frequency is a poor indicator of male body size in mammals.
Formant frequencies or resonances of the vocal tract probably represent the key acoustic variables linked to variation in body size in mammals , , –. Formants frequencies and their average spacing (formant dispersion) depend upon tissue structure, and the shape and length of the vocal tract . Longer vocal tracts produce lower formant frequencies. In contrast to the vocal folds, the length of the vocal tract is constrained by skeletal structures (e.g. dimensions of the skull) and therefore closely tied to overall body size .
In addition to male body size, sexually selected calls might convey information about other indirect measures of male fitness, such as social dominance. Dominance rank and acoustic parameters can be indirectly related as they are both not fixed traits of the caller, but instead vary according to individual physical and physiological attributes , , . The perception of these characteristics based on acoustic cues by competing males may affect the outcome of agonistic interactions. Females evaluating the relative quality of males might rely on acoustic cues related to dominance because high-ranking males often have better survival, and overall reproductive success than low-ranking males , , . While there is now strong evidence that some acoustic parameters of vocalisations represent body size, the relationships between vocal parameters and other characteristics linked to male quality, such as dominance rank and mating success, have rarely been examined.
Fallow deer are ideal for investigating the role of acoustic signals as indicators of male quality. Fallow deer are characterised by a polygynous mating system with high male-male competition and skewed reproductive success , . Larger males are generally higher ranked than smaller males, and rank is also closely associated with mating success . Males only vocalise during the breeding season and the sexually selected call they produce is known as a groan , . In the northern hemisphere, males start groaning approximately three weeks before the first matings take place (late September) and continue until the vast majority of matings have occurred (early November, ). Vocalisations are directed both towards males during agonistic encounters and towards females during chasing or herding behaviour, suggesting a potential role of groans in both male-male competition and female choice . Vocalization rates appear to convey information to other males, and therefore to play a role in male mutual assessment . However, acoustic components are salient to mammals and have a strong biological significance independent of the vocalization rate –. When males vocalise, the larynx is pulled down towards the sternum and the length of the vocal tract increases , . As a consequence, the formant frequencies decrease and reach a minimum value that could reveal information related to male quality , . The groans of fallow bucks are individually distinctive  and as is the case with red deer (Cervus elaphus), conspecifics are likely to discriminate between males based on the sound of their calls .
We investigated the relationships between the acoustic structure of fallow deer groans and male quality. We first determined whether body size is related to the fundamental frequency parameters, formant frequencies, and formant dispersion of groans. We then examined the relationships between the acoustic parameters (minimum fundamental frequency and minimum formant dispersion) and dominance rank and mating success, while also considering the role of body size.
Relationships between Body Size and the Acoustic Parameters
The segment of the hind leg that we used as indicator of male body size ranged from 30.0 to 33.5 cm (mean = 32.1±0.3, N = 17). Body size was not significantly related to fundamental frequency parameters (GLMM: F0min, F1,15 = 0.0002, P = 0.990, F0mean, F1,15 = 0.0462, P = 0.833, F0max, F1,15 = 0.0520, P = 0.823). With increasing body size, the first four minimum formant frequencies tended to decrease, but none of the relationships were significant (GLMM: F1min, F1,15 = 1.16, P = 0.298; F2min, F1,15 = 3.21, P = 0.094; F3min, F1,15 = 2.08, P = 0.171; F4min, F1,15 = 3.13, P = 0.097). There was a tendency for body size to be negatively related to the minimum frequency of the fifth formant (GLMM: F5min, F1,15 = 3.92, P = 0.066). Body size varied negatively with the minimum frequency of the sixth formant and the formant dispersion (GLMM: F6min, F1,15 = 4.58, P = 0.049; Dfmin, F1,15 = 18.93, P<0.001; Fig. 1).
Relationships between Acoustic Parameters and Dominance Rank
Model selection favoured the model incorporating the minimum fundamental frequency (F0min), (lowest AICc; Table 1, model 2). This model shows that groans characterised by lower F0min are produced by higher-ranking males (Fig. 2: F1,12 = 8.43, P = 0.013). A close competitor of model 2 was model 4, which only included the minimum formant dispersion (Dfmin), (Table 1, Model 2 and 4: ΔAICc<2). The two models with the lowest AICc (model 2, 4) were together 64.8% supported by the data (combined Akaike weights, 0.443 and 0.205). However, the evidence ratio reveals that the model with the F0min (model 2) was more than twice as good as the second best model (model 4). The addition of Dfmin or body size to the best model did not result in a better supported model (Table 1, comparing model 2 with model 5: LRT χ21 = 1.82, P = 0.18; comparing model 3 with model 2: LRT χ21 = 2.27, P = 0.13). The model which includes only body size as a parameter was considerably less supported by the data (Table 1, model 13: <ΔAICc>7). Thus, F0min was the factor more strongly correlated with dominance rank. Dfmin is also related to dominance rank, but to a lesser extent. Body size was not related to rank.
Lower values of dominance rank indicate higher ranking males. Higher ranking males produced groans with lower minimum fundamental frequency.
Relationships between Acoustic Parameters and Mating Success
The model with only the rank included, was 61.9% supported by the data and clearly selected as the best model (lowest AICc; Table 1, model 7). This model was more than four times as good as the second best model in which body size was also included (model 8). The addition of body size did not significantly improve the best model (Table 1, comparing model 7 with model 8: LRT χ21 = 1.19, p = 0.28). The models in which Dfmin, F0min were included together with rank, had considerably less support than the best model (Table 1, model 13: 3<ΔAICc<7). All other models were poorly supported by the data (Table 1, model 6, 9, 10, 11, 12). Thus, dominance rank appears to be the crucial factor which determines male mating success in fallow deer, with higher-ranking males having higher mating success (F1,12 = 8.43, P = 0.003). Any relationship between the acoustics parameters and mating success appears to be mediated by rank.
We examined the relationships between the acoustic structure of fallow deer groans and male quality. We found that body size was negatively related to the minimum formant dispersion and not related to the fundamental frequency parameters of groans. We also found that minimum fundamental frequency and to a lesser extent, minimum formant dispersion, were related to dominance rank. Dominance was in turn strongly related to male mating success. The acoustic structure of sexually selected calls often contains information on different phenotypic traits of the caller that is potentially available to other individuals , , , . Recent research has shown that both fundamental frequency-related and formant-related parameters are important in determining the individuality of fallow deer groans . The results of the current study suggest that the same acoustic parameters also have the potential to reliably signal male fitness-related traits, and highlight a role for fundamental frequency (F0) in broadcasting information on social dominance previously only demonstrated in primates , .
Body size was not related to the fundamental frequency parameters of groans. This is similar to results for other mammals (rhesus macaques, Macaca mulatta, ; red deer, ; lions, Panthera leo, ; elephant seals, Mirounga leonina, ) and confirms that the growth of male larynx (and resulting fundamental frequency) is therefore at least partially dissociated from the growth of overall body size. In fallow deer, F0 is lower in males than in females ,  and decreases as fawns mature . Therefore, F0 variation still reflects sexual size dimorphism in the vocal apparatus and may be used to distinguish sex and possibly stage of development of the animal.
We found that body size was strongly negatively related to the minimum formant dispersion even though earlier research has shown that fallow bucks do not pull the larynx all the way down to its physiological limit during groaning (the sternum, ). Fallow deer males can groan more than 60 times per minute and call rate could be used as an assessment cue in competitive interactions , . Thus, pulling down the larynx to an extent that is not maximal might be adaptive for fallow bucks if this allows senders to reach high calling rates when a correct indication of body size is still provided. Similarly to red deer, conveying information about body size through formant dispersion might play an important role in both agonistic interactions and mate attraction , , .
Body size was related to individual formants to a lesser extent than to the minimum formant dispersion (Dfmin). Body size was also more strongly related to higher formants (F5min and F6min) than to lower ones (F1min–F4min). These results clearly agree with those of other studies in which the relationships between individual body size and formants were mainly attributed to the higher formant frequencies and formant dispersion, because these reflect vocal tract length more precisely than lower ones , , , .
Higher ranking males produced groans with lower minimum fundamental frequency (F0min). The F0 of groans is relatively stable within individuals before the peak of the rut and accounts for a large proportion of vocal individuality . Therefore, during the early phase of the rut, F0 is likely to represent a cue to the physical characteristics of the caller rather than to the motivational state or vocal effort of the animal. In humans (Homo sapiens), men with high androgen levels have voices with low F0 (pitch) and women preferred these males, especially close to ovulation , . Sexual selection may also have favoured the evolution of vocal cues to the hormonal state (and therefore competitive ability) in fallow deer males, and led to the selection of lower pitch vocalisations indicating higher-quality individuals. Our results suggest that the acoustic parameters are indirectly related to mating success through social dominance. In humans, males with lower F0-voices have higher reproductive success and this is likely to be due to greater access to mates . Therefore, in fallow deer, the negative relationships between the minimum fundamental frequency (F0min) and minimum formant dispersion (Dfmin) with mating success, are probably mediated by the dominance rank and body size of males, respectively . These characteristics along with a large investment in vocal display are crucial in determining access to females and in turn mating success in fallow deer , . We therefore suggest that females could use multiple cues from male vocal behaviour when trying to choose the best mate among those of similar quality.
The minimum formant dispersion (Dfmin) was only marginally related to dominance rank and body size was not related to rank. The sample size we used in our study was probably not sufficient to reveal an effect of body size on rank . Assuming that a large body size is important for winning contests and therefore reaching high-ranking positions, males should use Dfmin as a cue to body size. This has been recently shown in red deer in which males perceive the differences in the Dfmin of roars produced by different competitors and use them to adjust their vocalisations and behaviour accordingly .
It is important to note that the majority of social dominance relationships between males were established through non-contact interactions before the rut and therefore before males became vocal . During this time, males live in bachelor herds, and direct assessment of body size and body mass of the competitors is likely to play the major role in determining the outcome of the interactions and therefore the dominance ranks of males . During the rut, the dominance relationships previously established are modified by fights, and males are expected to assess the status of their opponents . Reliable assessment cues are those physically or physiologically linked to fighting ability and males may rely on several of these to assess each other during contests . The use of acoustic cues reflecting the social dominance and therefore the overall resource holding potential of the individual, would then be crucial, especially when the opportunity for visual and olfactory assessment is poor such as at long distances and at night.
We found that dominance rank was the factor that was most strongly related to male mating success and body size appeared to play a secondary role. This result confirms that in fallow deer, as in several other ungulates, reaching a high ranking position is important for males to gain matings , , .
Minimum formant dispersion (Dfmin) and fundamental frequency (F0min) are among those variables that contain most of the information about individuality in fallow deer . According to Dale et al. , traits signalling individuality should be characterised by different properties from those coding for male quality. However, some traits can have a role in both individual recognition and assessment of male competitive abilities . Additional investigations together with playback experiments are needed to elucidate the independent role of the acoustic components of groans in conveying identity and quality assessment cues.
In conclusion, this study shows for the first time that the fundamental frequency (F0) of sexually-selected male vocalisations contains reliable information about social dominance in a non-primate species. Our study also confirms the role of formants in revealing male body size in mammals. F0 and formant frequencies may therefore represent acoustic cues to male quality that have primarily evolved in response to intrasexual selection. Other aspects of male vocal behaviour such as the long-term investment in vocal display , are instead likely to influence mate choice more directly in fallow deer.
Materials and Methods
Study Site and Population
The study was conducted on a herd of European fallow deer in Phoenix Park (53° 22′ N, 6° 21′ W), Dublin, Ireland. Since 1971 the majority of fawns were ear tagged each year in June, by the park authorities and others. All males used in this study were tagged, of known age and therefore individually recognisable.
We used body size measurements taken from 17 different males (five in 1996, one in 1997, one in 1999, three in 2001, six in 2002 and one in 2003). Males were caught immediately before (third week of September) or after the breeding season (third week of November). The males were sedated by a veterinary surgeon using a mixture of etorphine hydrochloride (18–20 µg/kg−1, C-Vet Veterinary Products) and xylazine (360–420 µg/kg−1, Rompun Dry Substance, Bayer), which was administered intramuscularly by gas-propelled darts. We measured a segment of one hind leg for each male using callipers and this was used as an indicator of skeletal size , . Additional measurements (distance from the pre-orbital gland to the tip of the nose on both sides of the head) were also taken for a different study . The handling time was generally less than 10 mins. Immobilisation was reversed by intravenously injection of an antidote containing a combination of antagonistic drugs, diprenorphine hydrochloride, (Revivon, 24–28 µg/kg−1, C-Vet Veterinary Products) and antipamozole hydrochloride (50 µg/kg, Antisedan, Pfizer) in a total volume of less than 2 ml. We then monitored the males until they were fully alert and they generally ran away from the area where the handling had been carried out. The males did not show any adverse effects as a result of the handling and many of them (n = 11) later gained matings during the rut. All procedures were approved by the University of Zurich and comply with the laws of Ireland.
We conducted behavioural observations during the rut in 1997, 2000, 2002 and 2003. The rut refers to the period when matings occur. During this time, all-event recording of agonistic interactions and matings was carried out every day from dawn to dusk (circa 11 hours per day). There were 7–13 observers in the field at all times and the coverage of animals was maximised. The measure of male mating success was based on the number of observed copulations, and this provides a very good estimator of their reproductive success .
The outcomes of the agonistic interactions were used to calculate the dominance rank of each male by applying the David's score method . This method is the most appropriate when interactions are recorded over a long period of time, because it takes into account repeated interactions between dyad members that may determine win/loss asymmetries . Dominance ranks were calculated for males that interacted with at least 10% of other mature males.
Recording and Selection of Groans
Recordings were made using a Sennheiser MKH 70 directional microphone connected to a Sony digital audio tape recorder DAT-TCD D100. Groans were recorded between dawn and sunset at a distance of 10 to 50 m from the vocalizing animal.
Vocalisations were imported into a computer using Avisoft-SASLab Pro 4.38 at a sampling rate of 22.05 KHz and saved in WAV format, and at 16-bit amplitude resolution . The recordings that did not contain energy above 8 KHz were down-sampled to 16 KHz for a better frequency resolution. Narrow-band spectrograms of groans (Fig. 3a, FFT method, window length = 0.03 s, time step = 1000, frequency step = 250, frequency resolution = 20 Hz, Gaussian window shape, dynamic range = 35 dB) were edited using Praat 4.5.01 DSP package (P. Boersma and D. Weenink, University of Amsterdam, The Netherlands). Vocalisations with high levels of background noise were not considered for analysis.
On the spectrogram (a), the pulses and the first six formants are indicated. On the envelope of the signal (b), each peak of frequency is detected and indicated as “pulse”.
During the breeding season, fallow deer males feed very little and lose approximately 26% of their body weight , and there is some evidence from primates and deer that the acoustic structure of vocalisations can be affected by exhaustion , . We therefore analysed recordings taken between October 8 and October 20 when only a small proportion (15% or less) of the total number of matings had usually occurred , and the majority of agonistic interactions among males were non-contact displacements . This minimised the possibility that variation in the phonic structure of groans could have occurred due to exhaustion of the animals.
For each male, we selected groans from different bouts that were recorded during one or more days. We included in our analysis males between five and eight years of age because they had reached their asymptotic size and were not undergoing changes associated with senescence. Moreover, this range of ages includes the males that account for the vast majority of matings .
Groans are low-pitched vocalisations and therefore a pulse-train structure is generally visible in the spectrograms (Fig. 3a). The pulses represent the vibrations of the vocal folds and determine the fundamental frequency (F0) of the call. Fundamental frequency is equivalent to the inverse of interpulse interval and this can be measured as the distance between consecutive pulse onsets (Fig. 3b). Distances between pulses were measured automatically from the envelope (amplitude vs. time) of the signal by using Pulse Train Analysis (Fig. 3b; method = rectification+exponential decay; pulse detection = peak search with hysteresis; time constant = 1 ms; threshold = 0.1 V; hysteresis = 16–19 dB) in Avisoft-SASLab Pro 4.38. We calculated the values of the F0 along the groan and then averaged these values to obtain the mean F0 (F0mean). Because all groans showed at least a modest frequency inflection, the minimum and the maximum F0 (F0min and F0max) were also included in the analysis.
In the spectrogram of groans, six formants are evident as horizontal frequency bands (F1–F6 in Fig. 3a). The decrease of these formant frequencies along the groan reflects the elongation of the vocal tract occurring during vocalisation . Formants were estimated using Linear Predictive Coding analysis (LPC), (Sound: To Formant (burg) command) in Praat 4.5.01 DSP package. By performing a single LPC analysis on each groan, higher formants (F4–F6) were better detected and therefore more accurately measured than lower formants (F1–F3). We therefore conducted a double or a triple LPC analysis on each groan in order to get the best estimations of all formants. We first carried out an LPC analysis (time step = 0.01–0.02 s, maximum number of formants = 3–4, maximum formant = 700–850 Hz, window length 0.07–0.26 s) to measure the frequencies of the first three formants (F1–F3). Then we performed a second LPC analysis (time step = 0.01–0.02 s, maximum number of formants = 6–7 maximum formant = 1800–2600 Hz, window length 0.07–0.26 s) to estimate the last three formant frequencies (F4–F6). When the sixth formant was not detected by the second LPC analysis, we conducted a third LPC analysis (time step = 0.01–0.02 s, maximum number of formants = 5–6 maximum formant = 1800–2600 Hz, window length 0.07–0.26 s). We calculated the minimum frequencies of the six formants (F1min–F6min) from each groan by averaging the values over the last part of the call when formants become flat. This is the time when the larynx is pulled down at the maximum extent. Finally, we also estimated the minimum spacing of the formants (known as minimum formant dispersion, (Dfmin), according to .
We analysed the vocalisations of 17 different males recorded during five breeding seasons between 1997 and 2004 (six in 1997, three in 2000, two in 2002, four in 2003, and two in 2004). Vocalisations analysed in relation to male dominance rank and mating success were carried out in the same year of the behavioural observations (1997, 2000, 2002, 2003). Recordings from 2004 were only analysed in relation to male body size because data on matings and agonistic interactions were not collected during that breeding season. For three males, vocalisations were recorded and body size measurements taken during the same breeding season. For most of the males (N = 14), recordings were performed one (N = 12) or two years (N = 2) before or after their body size measurements were taken. We assumed that the body size measurement that we used (hind leg length) did not change in fully grown mature males.
The low fundamental frequency that characterises fallow deer groans is still detectable and measurable on the spectrogram of groans recorded at more than 100 meters from the source (E. Vannoni, unpublished data). By contrast, formant frequencies are frequently lost or distorted in recordings taken in suboptimal conditions, such as when the microphone is far from the vocalizing animal or when it is not facing the microphone . Therefore, because of the variety of the recording conditions, it was not always possible to measure the fundamental frequency and the formants on the same groans and for all males. As a result, sample sizes (number of groans and number of males) varied among analyses.
We used a general linear mixed effect model (GLMM) procedure fitted with residual maximum likelihood estimation (REML, lme function;  to investigate the effect of body size on the acoustic parameters of groans (F0-related parameters: 186 groans; 10.9±1.1 per individual; Formant frequencies and Dfmin: 144 groans; 8.5±0.8 per individual; N = 17). We conducted a univariate GLMM for all the acoustic parameters. Individual identity was fitted as random term so that we controlled for repeated measurements of the same individual. Body size was fitted as a fixed effect.
We used a model selection procedure  based on Akaike Information Criterion (AIC) to examine the relationships between the acoustic parameters and dominance rank and mating success of the males, while controlling for body size. Observational studies, such as ours, are better suited to the model selection than to null hypothesis testing –. All models were fitted with maximum likelihood implemented in the program R (ML, lm function; . Pooling males across different years may potentially require standardization of male ranking positions and mating success. However, we did not standardize these variables for several reasons. First, we initially included “Year of recording” as a factor in all models but as it never had a significant effect, we did not considered it further in the analyses (E. Vannoni and A.G. McElligott, unpublished data). Secondly, standardized ranks have been used in a similar study and the results were not affected if relative ranks were replaced with absolute ranks . Finally, as the ratio between the total number of matings and the number of mature males involved in the rut was relatively constant across different years (E. Vannoni and A.G. McElligott, unpublished data), it was not necessary to standardize the mating success.
The model selection technique identifies the model that best describes the structure in a data set among all a priori fitted models considered, controlling for the number of parameters (K), included in each model. Because each model is associated with a biological hypothesis, model selection identifies the hypothesis that is best supported by the data. We applied the AIC criteria adjusted for small sample size (AICc, ). This implies the selection of a few simple models that are most biologically meaningful . Therefore, we limited our analyses to the factors that we considered most important on biological grounds. We included body size and dominance rank in the models because they are known to play crucial roles in male mating success in many mammals, including fallow deer , . Among the measured acoustic parameters, we selected those that are most biologically meaningful , , . The minimum fundamental frequency (F0min) has the highest degree of inter-individual variation, among the acoustic parameters used to describe the phonic structure of fallow deer groans . It represents the lowest rate of vocal fold vibration and among the F0 parameters, is the only one to be physiologically constrained , . The minimum formant dispersion (Dfmin) is constrained by the length of the vocal tract and is used by red deer males to assess competitors, and by females to choose their mates , , . Moreover, both F0min and Dfmin are related to reproductive success in red deer .
We formulated two sets of candidate models. The fitted models for dominance rank (First set of models: model 1 to 5, Table 1) included the effects of body size, minimum fundamental frequency (F0min), and minimum formant dispersion (Dfmin). The fitted models for mating success (Second set of models: model 6 to 13, Table 1) included the effects of body size, dominance rank, F0min, and Dfmin. Then, we applied the model selection procedure based on AICc to each of the two sets. The value of AICc for a given model is a measure of the loss of information resulting from the use of the model to explain a particular pattern. Therefore, the model with the smallest AICc value is estimated to best fit the data set relative to other models considered . When the difference between the AICc values of two models (ΔAICc) is less than 2 units, both models have support and can be considered competitive. Models with ΔAICc ranging from 3 to 7 have considerably less support by the data, whereas models with ΔAICc>10 are poorly supported and therefore very unlikely . Akaike weights (wi) indicate the probability that a particular model is supported by the data among those included in the set of candidate models . Akaike weights are normalized across the selected models to sum to one, and are interpreted as probabilities. For instance, an Akaike weight of 0.75 for a model, indicates that given the data, it has a 75% chance of being the best one in the set. For each model, we also calculated the evidence ratio, defined as the ratio between the Akaike weight of the best model and the Akaike weights of the competing model, to determine to what extent it is better than another. We used the likelihood-ratio tests (LRT) to compare nested models and to assess statistical significance of the factors. The LRT statistics follows a χ2-distribution with degrees of freedom equal to the difference in the number of parameters.
Data on dominance ranks and mating success were not available for three of the males for which we had body size measurements. Therefore, to investigate the relationships between acoustic parameters and dominance rank and mating success, we used data from 14 males (F0-related parameters: 156 groans, 11.1±1.3 per individual; Formant frequencies and Dfmin: 115 groans, 8.2±0.9 per individual). We log-transformed dominance rank and mating success to achieve normality. One unit was added to the mating success of all individuals before applying the log-transformation. In this way, we were able to transform the value of those individuals who did not get any matings (value = 0). All analyses were performed using R statistical software . All tests were 2-tailed and factors were considered to have a statistically significant influence if p<0.05. All means are given with standard errors.
We thank Dúchas The Heritage Service (Ireland), and the staff of Phoenix Park for their support. We are grateful to Marco Torriani and Petra Zajec for field assistance, and colleagues at University College Dublin for additional help. We thank Res Altwegg, Ben Charlton, Loren Gygax, Hansjörg Kunc, Robert Seyfarth, Karen McComb and David Reby for statistical and other advice.
Conceived and designed the experiments: EV AGM. Performed the experiments: EV AGM. Analyzed the data: EV. Wrote the paper: EV AGM.
- 1. Andersson M (1994) Sexual selection. Princeton, NJ: Princeton University Press. M. Andersson1994Sexual selectionPrinceton, NJPrinceton University Press
- 2. Doty GV, Welch AM (2001) Advertisement call duration indicates good genes for offspring feeding rate in gray tree frogs (Hyla versicolor). Behav Ecol Sociobiol 49: 150–156.GV DotyAM Welch2001Advertisement call duration indicates good genes for offspring feeding rate in gray tree frogs (Hyla versicolor).Behav Ecol Sociobiol49150156
- 3. Christie PJ, Mennill DJ, Ratcliffe LM (2004) Pitch shifts and song structure indicate male quality in the dawn chorus of black-capped chickadees. Behav Ecol Sociobiol 55: 341–348.PJ ChristieDJ MennillLM Ratcliffe2004Pitch shifts and song structure indicate male quality in the dawn chorus of black-capped chickadees.Behav Ecol Sociobiol55341348
- 4. Fischer J, Kitchen DM, Seyfarth RM, Cheney DL (2004) Baboon loud calls advertise male quality: acoustic features and their relation to rank, age, and exhaustion. Behav Ecol Sociobiol 56: 140–148.J. FischerDM KitchenRM SeyfarthDL Cheney2004Baboon loud calls advertise male quality: acoustic features and their relation to rank, age, and exhaustion.Behav Ecol Sociobiol56140148
- 5. Vehrencamp SL (2000) Handicap, index, and conventional elements of bird song. In: Espmark Y, Amundsen T, Rosenquist G, editors. Animal Signals: Signalling and Signal Design in Animal Communication. Trondheim: Tapir. pp. 277–300.SL Vehrencamp2000Handicap, index, and conventional elements of bird song.Y. EspmarkT. AmundsenG. RosenquistAnimal Signals: Signalling and Signal Design in Animal CommunicationTrondheimTapir277300
- 6. Maynard-Smith J, Harper D (2003) Animal signals. New York: Oxford University Press. J. Maynard-SmithD. Harper2003Animal signalsNew YorkOxford University Press
- 7. McElligott AG, Gammell MP, Harty HC, Paini DR, Murphy DT, et al. (2001) Sexual size dimorphism in fallow deer (Dama dama): do larger, heavier males gain greater mating success? Behav Ecol Sociobiol 49: 266–272.AG McElligottMP GammellHC HartyDR PainiDT Murphy2001Sexual size dimorphism in fallow deer (Dama dama): do larger, heavier males gain greater mating success?Behav Ecol Sociobiol49266272
- 8. Reby D, McComb KE (2003) Vocal communication and reproduction in deer. Adv Study Behav 33: 231–264.D. RebyKE McComb2003Vocal communication and reproduction in deer.Adv Study Behav33231264
- 9. Pfefferle D, Fischer J (2006) Sound and size: identification of acoustic variables that reflect body size in hamadryas baboons, Papio hamadryas. Anim Behav 72: 43–51.D. PfefferleJ. Fischer2006Sound and size: identification of acoustic variables that reflect body size in hamadryas baboons, Papio hamadryas.Anim Behav724351
- 10. Price JJ, Earnshaw SM, Webster MS (2006) Montezuma oropendolas modify a component of song constrained by body size during vocal contests. Anim Behav 71: 799–807.JJ PriceSM EarnshawMS Webster2006Montezuma oropendolas modify a component of song constrained by body size during vocal contests.Anim Behav71799807
- 11. Davies NB, Halliday TR (1978) Deep croaks and fighting assessment in toads (Bufo bufo). Nature 274: 683–684.NB DaviesTR Halliday1978Deep croaks and fighting assessment in toads (Bufo bufo).Nature274683684
- 12. Bee MA, Perrill SA, Owren PC (1999) Size assessment in simulated territorial encounters between male green frogs (Rana clamitans). Behav Ecol Sociobiol 45: 177–184.MA BeeSA PerrillPC Owren1999Size assessment in simulated territorial encounters between male green frogs (Rana clamitans).Behav Ecol Sociobiol45177184
- 13. Ryan MJ, Brenowitz EA (1985) The role of body size, phylogeny, and ambient noise in the evolution of bird song. Am Nat 126: 87–100.MJ RyanEA Brenowitz1985The role of body size, phylogeny, and ambient noise in the evolution of bird song.Am Nat12687100
- 14. Fischer J, Hammerschmidt K, Cheney DL, Seyfarth RM (2002) Acoustic features of male baboon loud calls: Influences of context, age, and individuality. J Acoust Soc Am 111: 1465–1474.J. FischerK. HammerschmidtDL CheneyRM Seyfarth2002Acoustic features of male baboon loud calls: Influences of context, age, and individuality.J Acoust Soc Am11114651474
- 15. Collins SA, Missing C (2003) Vocal and visual attractiveness are related in women. Anim Behav 65: 997–1004.SA CollinsC. Missing2003Vocal and visual attractiveness are related in women.Anim Behav659971004
- 16. Reby D, McComb KE (2003) Anatomical constraints generate honesty: acoustic cues to age and weight in the roars of red deer stags. Anim Behav 65: 519–530.D. RebyKE McComb2003Anatomical constraints generate honesty: acoustic cues to age and weight in the roars of red deer stags.Anim Behav65519530
- 17. Titze IR (1994) Principles of voice production. Englewood Cliffs, NJ: Prentice Hall. IR Titze1994Principles of voice productionEnglewood Cliffs, NJPrentice Hall
- 18. Fitch WT (1997) Vocal tract length and formant frequency dispersion correlate with body size in rhesus macaques. J Acoust Soc Am 102: 1213–1222.WT Fitch1997Vocal tract length and formant frequency dispersion correlate with body size in rhesus macaques.J Acoust Soc Am10212131222
- 19. Beckford NS, Schaid D, Rood SR, Schanbacher B (1985) Androgen stimulation and laryngeal development. Ann Oto Rhinol Laryn 94: 634–640.NS BeckfordD. SchaidSR RoodB. Schanbacher1985Androgen stimulation and laryngeal development.Ann Oto Rhinol Laryn94634640
- 20. Dabbs JM, Mallinger A (1999) High testosterone levels predict low voice pitch among men. Pers Indiv Differ 27: 801–804.JM DabbsA. Mallinger1999High testosterone levels predict low voice pitch among men.Pers Indiv Differ27801804
- 21. Fitch WT, Giedd J (1999) Morphology and development of the human vocal tract: a study using magnetic resonance imaging. J Acoust Soc Am 106: 1511–1522.WT FitchJ. Giedd1999Morphology and development of the human vocal tract: a study using magnetic resonance imaging.J Acoust Soc Am10615111522
- 22. Riede T, Fitch WT (1999) Vocal tract length and acoustic of vocalization in the domestic dog (Canis familiaris). J Exp Biol 202: 2859–2867.T. RiedeWT Fitch1999Vocal tract length and acoustic of vocalization in the domestic dog (Canis familiaris).J Exp Biol20228592867
- 23. Harris TR, Fitch WT, Goldstein LM, Fashing PJ (2006) Black and white colobus monkey (Colobus guereza) roars as a source of both honest and exaggerated information about body mass. Ethology 112: 911–920.TR HarrisWT FitchLM GoldsteinPJ Fashing2006Black and white colobus monkey (Colobus guereza) roars as a source of both honest and exaggerated information about body mass.Ethology112911920
- 24. Sanvito S, Galimberti F, Miller EH (2007) Vocal signaling of male southern elephant seals is honest but imprecise. Anim Behav 73: 287–299.S. SanvitoF. GalimbertiEH Miller2007Vocal signaling of male southern elephant seals is honest but imprecise.Anim Behav73287299
- 25. Fant G (1960) Acoustic theory of speech production. The Hague: Mouton. G. Fant1960Acoustic theory of speech productionThe HagueMouton
- 26. Mazur A, Booth A (1998) Testosterone and dominance in men. Behav Brain Sci 21: 353–363.A. MazurA. Booth1998Testosterone and dominance in men.Behav Brain Sci21353363
- 27. Beehner JC, Bergman TJ, Cheney DL, Seyfarth RM, Whitten PL (2006) Testosterone predicts future dominance rank and mating activity among male chacma baboons. Behav Ecol Sociobiol 59: 469–479.JC BeehnerTJ BergmanDL CheneyRM SeyfarthPL Whitten2006Testosterone predicts future dominance rank and mating activity among male chacma baboons.Behav Ecol Sociobiol59469479
- 28. McElligott AG, Altwegg R, Hayden TJ (2002) Age-specific survival and reproductive probabilities: evidence for senescence in male fallow deer (Dama dama). Proc R Soc Lond B 269: 1129–1137.AG McElligottR. AltweggTJ Hayden2002Age-specific survival and reproductive probabilities: evidence for senescence in male fallow deer (Dama dama).Proc R Soc Lond B26911291137
- 29. Alberts SC, Buchan JC, Altmann J (2006) Sexual selection in wild baboons: from mating opportunities to paternity success. Anim Behav 72: 1177–1196.SC AlbertsJC BuchanJ. Altmann2006Sexual selection in wild baboons: from mating opportunities to paternity success.Anim Behav7211771196
- 30. McElligott AG, Mattiangeli V, Mattiello S, Verga M, Reynolds CA, et al. (1998) Fighting tactics of fallow bucks (Dama dama, Cervidae): reducing the risks of serious conflict. Ethology 104: 789–803.AG McElligottV. MattiangeliS. MattielloM. VergaCA Reynolds1998Fighting tactics of fallow bucks (Dama dama, Cervidae): reducing the risks of serious conflict.Ethology104789803
- 31. McElligott AG, Hayden TJ (2000) Lifetime mating success, sexual selection and life history of fallow bucks (Dama dama). Behav Ecol Sociobiol 48: 203–210.AG McElligottTJ Hayden2000Lifetime mating success, sexual selection and life history of fallow bucks (Dama dama).Behav Ecol Sociobiol48203210
- 32. Reby D, Joachim J, Lauga J, Lek S, Aulagnier S (1998) Individuality in the groans of fallow deer (Dama dama) bucks. J Zool 245: 79–84.D. RebyJ. JoachimJ. LaugaS. LekS. Aulagnier1998Individuality in the groans of fallow deer (Dama dama) bucks.J Zool2457984
- 33. Vannoni E (2007) Natural and sexual selection at work: the structure and function of fallow deer (Dama dama) vocalizations. PhD Thesis. Switzerland: University of Zurich. E. Vannoni2007Natural and sexual selection at work: the structure and function of fallow deer (Dama dama) vocalizations. PhD ThesisSwitzerlandUniversity of Zurich
- 34. McElligott AG, O'Neill KP, Hayden TJ (1999) Cumulative long-term investment in vocalization and mating success of fallow bucks, Dama dama. Anim Behav 57: 1159–1167.AG McElligottKP O'NeillTJ Hayden1999Cumulative long-term investment in vocalization and mating success of fallow bucks, Dama dama.Anim Behav5711591167
- 35. McElligott AG, Hayden TJ (1999) Context-related vocalization rates of fallow bucks, Dama dama. Anim Behav 58: 1095–1104.AG McElligottTJ Hayden1999Context-related vocalization rates of fallow bucks, Dama dama.Anim Behav5810951104
- 36. Charlton BD, Reby D, McComb K (2008) Effects of combined source (F0) and filter variation (formants) on red deer hind responses to male roars. J Acoust Soc Am 123: 2936–2943.BD CharltonD. RebyK. McComb2008Effects of combined source (F0) and filter variation (formants) on red deer hind responses to male roars.J Acoust Soc Am12329362943
- 37. Fitch WT, Fritz JB (2006) Rhesus macaques spontaneously perceive formants in conspecific vocalizations. J Acoust Soc Am 120: 2132–2141.WT FitchJB Fritz2006Rhesus macaques spontaneously perceive formants in conspecific vocalizations.J Acoust Soc Am12021322141
- 38. Reby D, McComb KE, Cargnelutti B, Darwin C, Fitch WT, et al. (2005) Red deer stags use formants as assessment cues during intrasexual agonistic interactions. Proc R Soc Lond B 272: 941–947.D. RebyKE McCombB. CargneluttiC. DarwinWT Fitch2005Red deer stags use formants as assessment cues during intrasexual agonistic interactions.Proc R Soc Lond B272941947
- 39. Fitch WT, Reby D (2001) The descended larynx is not uniquely human. Proc R Soc Lond B 268: 1669–1675.WT FitchD. Reby2001The descended larynx is not uniquely human.Proc R Soc Lond B26816691675
- 40. McElligott AG, Birrer M, Vannoni E (2006) Retraction of the mobile descended larynx during groaning enables fallow bucks (Dama dama) to lower their formant frequencies. J Zool 270: 340–345.AG McElligottM. BirrerE. Vannoni2006Retraction of the mobile descended larynx during groaning enables fallow bucks (Dama dama) to lower their formant frequencies.J Zool270340345
- 41. Vannoni E, McElligott AG (2007) Individual acoustic variation in fallow deer (Dama dama) common and harsh groans: a source-filter theory perspective. Ethology 113: 223–234.E. VannoniAG McElligott2007Individual acoustic variation in fallow deer (Dama dama) common and harsh groans: a source-filter theory perspective.Ethology113223234
- 42. Reby D, Hewison M, Izquierdo M, Pepin D (2001) Red deer (Cervus elaphus) hinds discriminate between the roars of their current harem-holder stag and those of neighbouring stags. Ethology 107: 951–959.D. RebyM. HewisonM. IzquierdoD. Pepin2001Red deer (Cervus elaphus) hinds discriminate between the roars of their current harem-holder stag and those of neighbouring stags.Ethology107951959
- 43. Galeotti P, Saino N, Sacchi R, Møller AP (1997) Song correlates with social context, testosterone and body size condition in male barn swallow. Anim Behav 53: 687–700.P. GaleottiN. SainoR. SacchiAP Møller1997Song correlates with social context, testosterone and body size condition in male barn swallow.Anim Behav53687700
- 44. Puts DA, Gaulin SJC, Verdolini K (2006) Dominance and the evolution of sexual dimorphism in human voice pitch. Evol Hum Behav 27: 283–296.DA PutsSJC GaulinK. Verdolini2006Dominance and the evolution of sexual dimorphism in human voice pitch.Evol Hum Behav27283296
- 45. Pfefferle D, West PM, Grinnell J, Packer C (2007) Do acoustic features of lion, Panthera leo, roars reflect sex and male condition? J Acoust Soc Am 121: 3947–3953.D. PfefferlePM WestJ. GrinnellC. Packer2007Do acoustic features of lion, Panthera leo, roars reflect sex and male condition?J Acoust Soc Am12139473953
- 46. Torriani MVG, Vannoni E, McElligott AG (2006) Mother-young recognition in an ungulate hider species: a unidirectional process. Am Nat 168: 412–420.MVG TorrianiE. VannoniAG McElligott2006Mother-young recognition in an ungulate hider species: a unidirectional process.Am Nat168412420
- 47. McElligott AG, Hayden TJ (2001) Postcopulatory vocalizations of fallow bucks: who is listening? Behav Ecol 12: 41–46.AG McElligottTJ Hayden2001Postcopulatory vocalizations of fallow bucks: who is listening?Behav Ecol124146
- 48. Charlton BD, Reby D, McComb K (2007) Female red deer prefer the roars of larger males. Biol Lett 3: 382–385.BD CharltonD. RebyK. McComb2007Female red deer prefer the roars of larger males.Biol Lett3382385
- 49. Charlton BD, Reby D, McComb K (2007) Female perception of size-related formant shifts in red deer (Cervus elaphus). Anim Behav 74: 707–714.BD CharltonD. RebyK. McComb2007Female perception of size-related formant shifts in red deer (Cervus elaphus).Anim Behav74707714
- 50. Rendall D, Kollias S, Ney C, Lloyd P (2005) Pitch (F0) and formant profiles of human vowels and vowel-like baboon grunts: the role of vocalizer body size and voice-acoustic allometry. J Acoust Soc Am 117: 944–955.D. RendallS. KolliasC. NeyP. Lloyd2005Pitch (F0) and formant profiles of human vowels and vowel-like baboon grunts: the role of vocalizer body size and voice-acoustic allometry.J Acoust Soc Am117944955
- 51. Puts DA (2005) Mating context and menstrual phase affect female preferences for male voice pitch. Evol Hum Behav 26: 388–397.DA Puts2005Mating context and menstrual phase affect female preferences for male voice pitch.Evol Hum Behav26388397
- 52. Apicella CL, Feinberg DR, Marlowe FW (2007) Voice pitch predicts reproductive success in male hunter-gatherers. Biol Lett 3: 682–684.CL ApicellaDR FeinbergFW Marlowe2007Voice pitch predicts reproductive success in male hunter-gatherers.Biol Lett3682684
- 53. Hogg JT (1987) Intrasexual competition and mate choice in Rocky Mountain bighorn sheep. Ethology 75: 119–144.JT Hogg1987Intrasexual competition and mate choice in Rocky Mountain bighorn sheep.Ethology75119144
- 54. Saunders FC, McElligott AG, Safi K, Hayden TJ (2005) Mating tactics of male feral goats (Capra hircus): risks and benefits. Acta Ethol 8: 103–110.FC SaundersAG McElligottK. SafiTJ Hayden2005Mating tactics of male feral goats (Capra hircus): risks and benefits.Acta Ethol8103110
- 55. Mainguy G, Côté SD, Cardinal E, Houle M (2008) Mating tactics and mate choice in relation to age and social rank in male mountain goats. J Mamm 89: 626–635.G. MainguySD CôtéE. CardinalM. Houle2008Mating tactics and mate choice in relation to age and social rank in male mountain goats.J Mamm89626635
- 56. Dale J, Lank DB, Reeve HK (2001) Signaling individual identity versus quality: a model and case studies with ruffs, queleas, and house finches. Am Nat 158: 75–86.J. DaleDB LankHK Reeve2001Signaling individual identity versus quality: a model and case studies with ruffs, queleas, and house finches.Am Nat1587586
- 57. Hurst JL, Beynon RJ (2004) Scent wars: the chemobiology of competitive signalling in mice. BioEssays 26: 1288–1298.JL HurstRJ Beynon2004Scent wars: the chemobiology of competitive signalling in mice.BioEssays2612881298
- 58. Coltman DW, Smith JA, Bancroft DR, Pilkington J, MacColl ADC, et al. (1999) Density-dependent variation in breeding success and natural and sexual selection in Soay rams. Am Nat 154: 730–746.DW ColtmanJA SmithDR BancroftJ. PilkingtonADC MacColl1999Density-dependent variation in breeding success and natural and sexual selection in Soay rams.Am Nat154730746
- 59. Say L, Naulty F, Hayden TJ (2003) Genetic and behavioural estimates of reproductive skew in male fallow deer. Mol Ecol 12: 2793–2800.L. SayF. NaultyTJ Hayden2003Genetic and behavioural estimates of reproductive skew in male fallow deer.Mol Ecol1227932800
- 60. David HA (1987) Ranking from unbalanced paired-comparison data. Biometrika 74: 432–436.HA David1987Ranking from unbalanced paired-comparison data.Biometrika74432436
- 61. Gammell MP, De Vries H, Jennings DJ, Carlin CM, Hayden TJ (2003) David's score: a more appropriate dominance ranking method than Clutton-Brock et al.'s index. Anim Behav 66: 601–605.MP GammellH. De VriesDJ JenningsCM CarlinTJ Hayden2003David's score: a more appropriate dominance ranking method than Clutton-Brock et al.'s index.Anim Behav66601605
- 62. Vannoni E, Torriani MVG, McElligott AG (2005) Acoustic signalling in cervids: a methodological approach for measuring vocal communication in fallow deer. Cogn Brain Behav 9: 551–566.E. VannoniMVG TorrianiAG McElligott2005Acoustic signalling in cervids: a methodological approach for measuring vocal communication in fallow deer.Cogn Brain Behav9551566
- 63. McElligott AG, Naulty F, Clarke W, Hayden TJ (2003) The somatic cost of reproduction: what determines reproductive effort in prime-aged fallow bucks? Evol Ecol Res 5: 1239–1250.AG McElligottF. NaultyW. ClarkeTJ Hayden2003The somatic cost of reproduction: what determines reproductive effort in prime-aged fallow bucks?Evol Ecol Res512391250
- 64. McComb KE (1988) Roaring and reproduction in red deer, Cervus elaphus. PhD Thesis. United Kingdom: University of Cambridge. KE McComb1988Roaring and reproduction in red deer, Cervus elaphus. PhD ThesisUnited KingdomUniversity of Cambridge
- 65. Venables WN, Ripley BD (2002) Modern applied statistics with S. 4th edn. New York: Springer. WN VenablesBD Ripley2002Modern applied statistics with S. 4th ednNew YorkSpringer
- 66. Burnham KP, Anderson DR (2002) Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. 2nd edn. New York: Springer. KP BurnhamDR Anderson2002Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. 2nd ednNew YorkSpringer
- 67. Johnson JB, Omland KS (2004) Model selection in ecology and evolution. Trends Ecol Evol 19: 101–108.JB JohnsonKS Omland2004Model selection in ecology and evolution.Trends Ecol Evol19101108
- 68. Blanckenhorn WU, Hellriegel B, Hosken DJ, Jann P, Altwegg R, et al. (2004) Does testis size track expected mating success in yellow dung flies? Func Ecol 18: 414–418.WU BlanckenhornB. HellriegelDJ HoskenP. JannR. Altwegg2004Does testis size track expected mating success in yellow dung flies?Func Ecol18414418
- 69. Altwegg R, Dummermuth S, Anholt BR, Flatt T (2005) Winter weather affects asp viper Vipera aspis population dynamics through susceptible juveniles. Oikos 110: 55–66.R. AltweggS. DummermuthBR AnholtT. Flatt2005Winter weather affects asp viper Vipera aspis population dynamics through susceptible juveniles.Oikos1105566
- 70. Modig AO (1996) Effects of body size and harem size on male reproductive behaviour in the southern elephant seal. Anim Behav 51: 1295–1306.AO Modig1996Effects of body size and harem size on male reproductive behaviour in the southern elephant seal.Anim Behav5112951306
- 71. R Development Core Team (2006) R: A language and environment for statistical computing. R Development Core Team2006R: A language and environment for statistical computing.Vienna, Austria (ISBN 3-900051-07-0). Vienna, Austria (ISBN 3-900051-07-0).