First Evidence of Leatherback Turtle (Dermochelys coriacea) Embryos and Hatchlings Emitting Sounds
Abstract
There are only a few studies documenting acoustic communication in chelonians; here we provide additional evidence that sound plays an important role in the interchange of information in this group. We recorded 12 nests of Dermochelys coriacea during March 2012 at Barra de la Cruz Beach, Oaxaca, Mexico. Four types of sounds were identified from recordings of nests after 51 d of incubation; our results reinforce the idea that sounds are important to coordinate group behavior in turtles.
The mainstream literature provoking the idea that all turtles were silent and deaf (Pope 1955), and that the sounds they did produce were simply noises made during nesting (Mrosovsky 1972; Wever 1978; Cook and Forrest 2005), perhaps inhibited research in turtle acoustic communication in the past decades. Currently, 47 turtle species have been documented emitting sounds in different contexts (Campbell and Evans 1967; Campbell 1972; Auffenberg 1977; Sacchi et al. 2003; Cook and Forrest 2005; Galeotti et al. 2005a, 2005b; Giles et al. 2009; Ferrara et al. 2012) and the most recent studies have shown that acoustic signals play an important role in turtle social behavior and reproduction (Galeotti et al. 2005a; Giles et al. 2009; Ferrara et al. 2012).
Furthermore, several studies have demonstrated marine and freshwater turtle ability to perceive aerial and underwater low-frequency sounds (Ridgway et al. 1969; Lenhardt et al. 1983, 1996; O'Hara 1990; Bartol et al. 1999; Willis et al. 2011; Christensen-Dalsgaard et al. 2012). It is likely that all turtles produce sounds to some extent, and that acoustic communication has been overlooked in most studies of turtle behavior.
Detailed information about the use of sound in aquatic turtles is limited. The sound repertoire of only 2 species of freshwater turtles has been studied extensively, Chelodina oblonga (Giles et al. 2009) and Podocnemis expansa (Ferrara et al. 2012). Additionally there are records of a marine turtle, Dermochelys coriacea, producing sounds while nesting (Mrosovsky 1972; Cook and Forrest 2005).
There has been suggestion that some audible cues from pipping eggs or some means of acoustic communication between embryos facilitate hatching synchronization (Spencer and Janzen 2011). Hatching should not be considered a fixed passive direct effect of temperature but rather a plastic behavior that evolved based on variation of environmental cues (Warkentin 2011), such as sound signals. In fact, sound production by turtle embryos and hatchlings has been shown by a single study of the freshwater turtle P. expansa (Ferrara et al. 2012). The following hypotheses were formulated by Ferrara et al. (2012) to explain the role of this behavior in different species of turtles: 1) to synchronize hatching and induce communal digging to help move the siblings toward the surface and out of the nest; 2) to synchronize emergence from the nest to dilute the predation risk during hatchling dispersion to water; and 3) to solicit adults to approach so that the hatchlings can be led and accompanied by the females in their migration.
Sound communication appears to be a more important facet of freshwater and marine turtle biology than previously imagined, particularly between laying of the eggs and hatching of the young during the reproductive season. The objective of our study was to verify if D. coriacea embryos emit sounds during incubation and to describe the preemergence sound repertoire of this species.
Methods
We recorded sound in 12 nests of D. coriacea during March 2012, in Barra de La Cruz Beach, Oaxaca, México. The eggs in the nests were counted and transferred from the natural beach to a protected area on the same beach the day after nesting. Since the local incubation period varies from 53 to 62 d (Tavera and López 2012), we waited 51 d before openning the nest to attempt sound recordings to ensure that the ears and the sound production apparatus of the embryos were developed enough for acoustic interactions. Sampled nests were 40 to 50 cm deep and classified into 2 types: nests with only eggs (n = 6) and nests with eggs and hatchlings (n = 6). We recorded nests containing between 14 and 94 eggs or hatchlings for 1 hr varying the time of the day between 0600 and 1800 hrs. Number of eggs in each nest had been counted when the eggs were transfered. Each nest was only sampled once. A Fostex FR-2 recorder ajusted to 48 kHz at 24 bits was used to acquire the sound connected to an Earthworks (M30) omnidirecional microphone, with a frequency response of 5 Hz to 30 kHz ± 1/−3 dB. We inserted the microphone 15 cm from the eggs or hatchlings and monitored the recordings in real time using Sony MDR-7506 headphones and ajusting the recording level manually to maximize the signal-to-noise ratio and to prevent distortions (“clipping”) caused by excess gain.
Raven Pro 1.3 (Cornell Lab of Ornithology, Ithaca, NY) was used to analyze the recordings using the following spectrographic parameters: window type, hamming; window size, 512 samples. Sounds with similar characteristics of published turtle sounds (Giles et al. 2009; Ferrara et al. 2012) and within the hearing range of turtles were detected manually by 2 experienced researchers using visual and aural inspection of the recordings. Spectrograms of high signal-to-noise ratio sounds with no overlapping signals were then selected for further acoustical analyses. The acoustic variables extracted from spectrograms were low and high frequency (kHz), signal duration (or durations), peak frequency (kHz), and number of inflection points. Presence and number of harmonic and nonharmonic frequency bands were also noted. To verify if there was a difference in the peak frequency of vocalizations found for each nest type (eggs only or eggs and hatchings), we used a t-test (Zar 1996).
Results
We started to detect sounds inside the nests of D. coriacea after 51 d of incubation with mean incubation period prior to sound emission of 53 ± 1.5 d SD (minimum 51 d; maximum 55 d). In 12 hrs we detected 328 sounds from the nests of D. coriacea; of these, 148 were emitted from nests with only eggs and 180 from nests with both eggs and hatchlings. The sounds were classified into 4 types according to their aural and spectral characteristics (Table 1; Fig. 1). Detected sounds included pulses, sounds with harmonic and nonharmonic frequency bands modulated in amplitude and frequency, and hybrid sounds with characteristics of pulsed and harmonic sounds.
Citation: Chelonian Conservation and Biology 13, 1; 10.2744/CCB-1045.1
-
Type I (n = 40). — This sound is characterized by 1 or more bouts of pulses with total duration varying between 0.1 and 0.5 sec. The mean peak frequency was 960 Hz, ranging from 187.5 to 1343.8 Hz. The sound pulses have variable repetition rates along the duration of each bout.
-
Type II (n = 40). — This is the shortest sound in their repertoire (mean duration of 0.03 sec) with multiple frequency bands not harmonically related, resulting in a noisy aural quality. The mean peak frequency was 954.4 Hz and varied from 282.2 to 1640.6 Hz.
-
Type III (n = 226). — These harmonic series may or may not include inflexion points, although frequency modulation was mostly ascending without inflection points (up-sweep). This sound may have 1 or 2 notes separated by a short silent interval. The mean peak frequency was 1029.64 and varied from 119 to 24,000 Hz. Total signal duration varies from 0.01 to 0.5 sec.
-
Type IV (n = 40). — This is the most complex sound type in their repertoire, with part of the signal having pulse characteristics and the other part showing harmonic frequency bands (Watkins 1967). In our sample of these hybrid sounds, the harmonic frequency bands are followed by pulses. This type of sound has intermediate characteristics between sound types I and III. The mean peak frequency was 706.8 Hz and varied from 281.2 to 2250 Hz. Total signal duration varies between 0.05 and 0.4 sec.
The sound type IV was recorded only in nests with eggs only (E). The other sound types were recorded in both types of nests (eggs only or eggs and hatchings) (Table 1). Sound type III (n = 226) was the most common sound recorded. Although sound type I was recorded in both nest types, it should be noted that this sound was recorded only once in nests with hatchlings (Table 1). The mean peak frequency of the sounds emitted by D. coriacea was 993.71 Hz, ranging from 119.2 to 24,000 Hz, and there was a significant difference between peak frequency values of sounds recorded in nests with eggs only (854.87 ± 454.86 Hz SD) and with both eggs and hatchlings (1074.94 ± 394.64 Hz SD) (t = 4.452; df = 117; p < 0.001).
Discussion
We found 4 types of sound in D. coriacea nests containing eggs and/or hatchlings. These sounds varied from pulses to harmonic series and nonharmonic frequency-modulated sounds, as well as hybrid sounds. Cook and Forrest (2005) described 3 types of sounds being produced by nesting D. coriacea; however, 2 of these sounds were considered to be respiratory (inhaling and exhaling sounds) and the other a grunt. Eggs and hatchlings of Leatherback turtles have a more complex sound repertoire. Morton (1977) suggests that species that have great complexity of social interactions develop sound signals more complex in quality than those of species with less complex social interactions. Even though the embryos of D. coriacea do not have an extensive vocal repertoire, they present a surprizing variety and a high complexity in their sounds for a species that is considered to have little social interaction. Complex sounds, such as the hybrid sound type IV, was also found in the sound repertoire of P. expansa (Ferrara et al. 2012) and in cetaceans, animals that are known to have complex social interactions (Murray et al. 1998). Therefore, we suggest that social interactions in aquatic turtles might be more complex than previously thought.
We showed that the vocal repertoire of D. coriacea eggs and hatchlings comprises frequencies as low as 119.2 Hz and as high as 24,000 Hz. In adult D. coriacea the frequency varied from 300 to 2000 Hz, even though there were signals of 4000 Hz (Cook and Forrest 2005). Most sound energy found in the turtle vocalizations studied thus far is concentrated in frequencies less than 1 kHz (Campbell and Evans 1967; Mrosovsky 1972; Auffenberg 1978; Jackson and Awbrey 1978; Galeotti et al. 2005a, Ferrara et al. 2012). Embryos of D. coriacea apparently emit sounds in the same frequency range as other species that have been studied. The mean peak frequency of Leatherback eggs and hatchlings' sounds recorded was 993.7 Hz, whereas the frequency of sounds of adults and embryos of P. expansa varied from 36.8 to 4500 Hz (Ferrara et al. 2012) and the values for C. oblonga varied from 100 to 3500 Hz, with peak frequencies up to 20 kHz (Giles et al. 2009).
The sound repertoire of the embryos of D. coriacea within the eggs are different from those of the nests with hatchlings, with differences in peak frequency values as well as the use of different types of sound in the repertoire. Pulses were present in nests with only eggs (with a single exception in a nest with hatchlings). This type of sound may be produced within eggs as a result of friction of body parts (stridulation), vibration of a body part in contact with some ressonating chamber within the egg (percussion), or even discharge of gas bubbles (Wilson et al. 2004). Sounds recorded from nests with only eggs also had lower frequency values then nests with eggs and hatchlings, as observed in P. expansa (Ferrara et al. 2012) and in crocodiles, where it has been suggested that this may be related to the distribution of energy of the sound, with the egg shell functioning as a filter for the sound wavelengths, amplifying lower frequencies and attenuating higher ones (Britton 2001; Vergne et al. 2009). Higher frequencies attenuate faster, making it less likely that a predator will be able to find the source of hatchling sounds (Forrest et al. 1993); thus, high-frequency sounds may have evolved as a strategy to avoid detection by potential predators.
We observed that the embryos in the eggs of D. coriacea had a higher diversity and variation in the structure of the sound than when hatchlings were in the nest; this is the opposite from what was found in P. expansa (Ferrara et al. 2012). In P. expansa the greater diversity of sound types and the variation in the acoustic structure could be related to a greater necessity for interchange of information and a greater complexity of information exchanged within the nest among the hatchlings (Ferrara et al. 2012), whereas in D. coriacea it might be more important to exchange information while still in the egg rather than after hatching. Differences in specific selective pressures related to environmental characteristics of the nesting sites (e.g., marine vs. freshwater, nest depth, background noise differences between sea and river) might account for this kind of variation.
Detection of sounds from eggs of D. coriacea after 51 d of incubation at natural temperatures show that embryos that are close to full development are communicating. Embryos of other species are also known to emit sounds a few hours or days before hatching, including P. expansa (Ferrara et al. 2012), crocodiles (Britton 2001; Vergne et al. 2009), and birds (Vince 1968; Colombelli-Négrel et al. 2012). Acoustic communication can be considered a cue for synchronized hatching, as suggested by others (Spencer and Janzen 2011).
The number of sound emissions for the 800 D. coriacea embryos recorded in 12 differents nests was low, as in the other species of turtles studied (P. expansa, Ferrara et al. 2012 and C. oblonga, Giles et al. 2009). As was suggested for P. expansa (Ferrara et al. 2012), predation pressure may be inhibiting the evolution of frequent sound emission by eggs of D. coriacea, although there are still advantages for sound emission, i.e., to synchronize hatching of the entire nest to increase the survivorship of the group by digging out of the nest simultaneously, using less energy, and then dilute predation pressure by hatchlings' simultaneous dispersion to the water (predator swamping; but see Tucker et al. 2008 that supports prey-switching hypothesis instead).
Our data on nest acoustic communication of D. coriacea show how different these sounds are compared to the simple sounds reported for the vocal repertoire of nesting females (Mrosovsky 1972; Cook and Forrest 2005). We hope that these results will stimulate more bioacoustic studies of different life history stages of more species of turtles to elucidate the basis for this behavior in turtles.
Conservation issues arise when acoustic communication meets sound pollution. Noise from motorized watercraft may affect reception of sounds by turtles and interfere with their communication, potentially jeopardizing hatchling survivorship (Ferrara et al. 2012). Samuel et al. (2005) suggest that continued exposure to existing high levels of anthropogenic noise in vital sea turtle habitats could further affect sea turtle behavior and ecology.
Oscillogram and spectrogram of all types of sounds recorded (window type: hamming; window size: type I FFT = 256; type II, III, and IV FFT = 512 points) in the vocal repertoire of Dermochelys coriacea.
Contributor Notes
Handling Editor: Jeffrey A. Seminoff
