Received 6 April 2011; accepted 16 July 2011
✉
Author for correspondence (Zhengwang Zhang)E-mail: ��w� ��w�
High frequency components in avian vocalizationshurly burly
Jianqiang LI, Y anyun ZHANG, Zhengwang ZHANG ✉
Ministry of Education, Key Laboratory for Biodiversity Sciences and Ecological Engineering, College of Life Sciences, Beijing Normal University, Beijing 100875, China
Abstract Ultrasonic communication in vertebrates is attracting increasing rearch interest. To determine if ultrasonic vocali�ation is common in birds, we recorded their vocali�ations with ultrasound detectors in the Dong�hai National Nature Rerve of Henan Province , China.�e found varying degrees of high fre-. �e found varying degrees of high fre-quency components in the vocali�ations of 14 species and in veral of the species, the frequency of har-monics was up to the
range of ultrasound. �e suggest that more studies are required to determine whether the high frequency components in avian vocali�ations have functions and what the functions are. In ad-dition, the ability of birds to hear sounds in the high frequency range also requires re-examination. Keywords avian vocali�ation, high frequency components, ultrasound, functions
Introduction
Sound is widely ud by animals for communication (Bradbury and Vehrencamp, 1998). Recently, a rapid expanding field in the study of animal communication is the u of ultrasonic (> 20 kH�) signals in terrestrial vertebrates (Arch and Narins, 2008). To this end, one of the more surprising groups shown to u ultrasound is amphibians (Narins and Feng, 2006), which, in the past, were thought to be unable to hear high frequency sounds (Pettigrew et al., 1981). However increasingly, evidence suggests that some frogs are capable of using ultrasonic vocali�ations for communication (Feng et al., 2006; Arch et al., 2008; Arch et al., 2009).
In birds, acoustic signals are the main means of commu-nication ud under various situations such as mating dis-play (Loffredo and Borgia, 1986; Gentner and Hul, 2000; Ballentine et al., 2004), guarding of territory (Forstmeier and Balsby, 2002) and parent-offspring interaction (Beer, 1979; Leo
nard et al., 1997). Although it has long been sus-pected that birds might hear sounds inaudible to humans, studies have not provided concrete evidence (Thorpe and Griffin, 1962; Konishi, 1970) and no species of bird has
been shown to be nsitive to ultrasonic frequencies (Bea-son, 2004). Even for birds that have the ability to echolo-cate, such as oilbirds and swifts, the main frequencies of their acoustic signals are within range of human hearing (Griffin, 1953; Konishi and Knudn, 1979; Fullard et al., 1993).
santak
However, the recent findings of ultrasonic communica-tion in vertebrates (especially tho in amphibians) suggest that communication approaches in the animal kingdom may be far more complex than previously thought. As birds are among tho of which vocali�ations are known to be most complicated (Dooling et al., 2002), it is therefore necessary and prudent to re-examine whether their vocal-i�ations carry ultrasonic information. Normally, studies of avian vocali�ations u standard recording equipment, with the upper limits of their frequency respon ldom over 20 kH�. In our study, we ud ultrasound detectors to conduct an investigation of the vocali�ations of birds, aim-ing to detect whether ultrasonic vocali�ations are common in birds.
Methods
prence什么意思All the avian sounds were recorded in the Dong�hai Na-tional Nature Rerve (31.95°N, 114.25°E; elevation 100–840 m) between January and March 2008, in June 2009 and December 2010. This rerve is located in the south of Henan Province of central China and within range of父亲的眼睛
Chine Birds 2011, 2(3):125–131 126
the Dabieshan Mountains. It is at the transitional region between the subtropical �one and the temperate �one and characteri�ed by a rich avian diversity.
The sounds were recorded with either a Pettersson Bat Detector D980 (frequency range 10–200 kH�) or a D1000X (frequency range 5–235 kH�) (Pettersson Elek-tronik AB, Sweden). The detector D980 was connected to a laptop and sounds were recorded, using BatSound Pro software (Pettersson Elektronik AB, Sweden) and loaded onto the hard drive of the laptop. The D1000X is equipped with a built-in 16-bit recording system and stores sound as �AV files on a compact flash card and does therefore not require a laptop connection. �e recorded bird vocal-i�ations either in the field, or when the birds were tem-porarily caged. The target species was randomly chon, depending on which species were found in the field or captured by mist nets. Distances to the birds varied from < 2 m for captive species to > 10 m for some species in the field. The sampling rates of t
he recordings were 100, 110, 192, 200, 300 or 384 kH�. The reason for the u of the many sampling frequencies was becau we had no prior knowledge of the frequency range of the vocali�ation of a given species; thus they were usually lected in a tenta-tive way. Spectrograms of the sounds were produced and checked for their frequency range with BatSound Pro. �e lected spectrograms that were typical of the species to prent in this paper.
Results
The vocali�ations of 14 avian species were recorded (Table 1) and spectra of the lected vocali�tions are shown in Fig. 1.
No bird was detected to produce pure ultrasonic calls. However, each bird vocali�tion contained more or less high frequency harmonics which approached or exceeded the audible limit of humans (20 kH�). The frequency of the harmonics of the Black-throated Tit and the Yellow-bellied Tit reached 40 kH�. It is noteworthy that despite a decreasing trend, the ultrasound components in some spe-cies, such as the �hite-crowned Forktail, Black-throated Tit, Long-tailed Tit and Yellow-bellied Tit, still carried significant energy, which was somewhat comparable to that of the fundamental components in their vocali�ations (Fig. 1).
Discussion
Batsound detectors D980 and D1000X were able to record sounds with the highest frequency up to 200 kH�. Dur-
Table 1Summary of the types of the avian vocali�ations and recording sites
Species Environment where the vocali�ation was recorded Vocali�ation type Reeves’s Pheasant (Syrmaticus reevesii)Captive Contact call a �hite-crowned Forktail (Enicurus leschenaulti)Field Contact call Plumbeous �ater Redstart (Rhyacornis fuliginosus)Field Contact call a Yellow-rumped Flycatcher (Ficedula zanthopygia)Field Song
mxm
Hair-crested Drongo (Dicrurus hottentottus)Captive Alarm call Hwamei (Garrulax canorus)Field Song
Great Tit (Parus major)Field Alarm call Yellow-bellied Tit (Parus venustulus)Captive Song
Black-throated Tit (Aegithalos concinnus)Captive Contact call Long-tailed Tit (Aegithalos caudatus)Captive Contact call Black Bulbul (Hypsipetes leucocephalus)Field Alarm call Vinous-throated Parrotbill (Paradoxornis webbianus)Captive Contact call Meadow Bunting (Emberiza cioides)Field Contact call Tristram’s Bunting (Emberiza tristrami)Captive Contact call a The vocali�a
tion type of this species is speculative.
Jianqiang Li et al. High frequency components in avian vocali�ations 127
Chine Birds 2011, 2(3):125–131 128
Fig. 1Sound spectra and instantaneous amplitude spectra [ints, taken at indicated points (▲)] of the vocali�ations of 14 avian species. The hori�ontal axis in the sound spectra reprents the time (ms)
(ms)and the vertical axis the frequency(kH�),while in the ints,the hori�ontal axis
and the vertical axis the frequency (kH�), while in the ints, the hori�ontal axis reprents frequency (kH�) and the vertical axis amplitude (dB). The hori�ontal lines in the panels parate the ultrasonic and audible compo-nents of the vocali�ations.
ing the study, we adopted sampling rates of at least 100 kH�, making the Nyquist frequencies at least 50 kH�, far exceeding the frequency range of human hearing. There-fore both the equipment and the sampling rates enabled us to detect ultrasonic components in avian vocali�ations if there were any. In the vocali�ations of 14 avian species, we demonstrated that they contained varying degrees of high frequency components and in veral species, the frequency of the harmonics reached the ultrasonic range. As the energy of sounds decread with an increa in the distance between recorder and producer, one limitation of the study became apparent in that the sounds of美国哈佛大学简介
安妮海瑟薇奥斯卡
half
Jianqiang Li et al. High frequency components in avian vocali�ations129
how to be blackof the species were recorded in the field, thus the highest frequency of some of the vocali�ations shown in Fig. 1 might be lower than they actually were due to the effect of distance. However, this potential error could not conceal the fact of the existence of high frequency components in avian vocali�ations, since the actual frequency might be even higher than tho shown.
Becau the harmonics in some species carried non-neglectable energy, the question aris as to why they pro-duce tho high frequency sounds if they have no function. Higher frequency vocali�ations may be ud adaptively to enhance the signal-to-noi ratio (Arch and Narins, 2008). It has been shown that in the Concave-eared Tor-rent Frog (Amolops tormotus), ultrasound communication is adopted to avoid masking by the wideband background noi of local fast-flowing streams where the species lives (Feng et al., 2006). In birds, there is also evidence that they may adjust the frequency of their vocali�ation in adapta-tion respon to a noisy environment. Take the Great Tits for example, they have a higher minimum frequency of their songs at noisy locations, presumably to prevent their songs from being masked to some extent by the predomi-nantly low-freq
boker
uency ambient noi (Slabbekoorn and Peet, 2003). Therefore, it is reasonable to speculate that communication by high frequency vocali�ation might be favored by natural lection for species living in noisy environments, such as the �hite-crowned Forktail and Plumbeous �ater Redstart, which live near streams. However, the properties of high-frequency sounds such as limitation in long-distance signaling, directional-ity and susceptibility to scattering, may limit their u in communication (Smith, 1979; Arch and Narins, 2008). One may also argue that the high frequency harmonics are just byproducts of the sound production mechanism (Roverud, 1989), or they could just be providing a signifi-cant component to the tonal quality (Thorpe and Griffin, 1962). Despite the arguments, high frequency signals have their advantages under certain circumstances. They do not spread far and are therefore hard for predators and competitors to detect. Selective pressure may therefore favor their u in the prence of predators (�ilson and Hare, 2004; �ilson and Hare, 2006) and in communica-tion for interactions between individuals at clo quarters (Smith, 1979; Arch and Narins, 2008). Studies of Richard-son’s Ground Squirrels (Spermophilus richardsonii) have demonstrated the u of high frequency sound communi-cation for predator avoidance. The squirrels can produce “whisper” calls containing pure ultrasonic frequencies of around 50 kH�, to warn nearby conspecifics of potential danger (�ilson and Hare, 2004; �ilson and Hare, 2006). In the prent study, the call of a Great Tit was recorded when it alarmed its partne
slimliner of our access. In the field, this call was also often heard when there were aerial predators, such as hawks, passing by the area. Since it contained sig-nificant energy both in fundamental and high frequency components (Fig. 1), this kind of alarm call may be more efficient to transmit information of the threat from signal-ers to their clo conspecifics, before being heard by the predators.
One study has suggested birds are capable of discrimi-nating harmonic complex (Dooling et al., 2002). The songs of the Y ellow-bellied Tit inspired another idea about the function of the high frequency harmonics: could the high frequency harmonics signal individual quality? The spec-trogram of the Yellow-bellied Tit in our study was record-ed from a captive individual when it was singing. In the field, we noticed such singing often occurred when Yellow-bellied Tits were in flocks and displayed to each other. �e therefore suggest the complex harmonic components in birds may carry information related to individual quality. Currently, we cannot rule out any of the and/or other possibilities for the function of high frequency harmonics in avian vocali�ations. It is clear that additional studies are needed to determine the functions of the high-frequency components. For example, playbacks of avian vocali�a-tions, of which the lower frequency components have been filtered out, should be ud to test whether species that are capable of producing high frequency vocali�ations are able to hear them. If they have the ability to hear high frequen-cy sounds, further
experiments should then be designed to test possible adaptive hypothes such as predator avoidance and choice of mates. Besides, although previous studies using electrophysiological approaches concluded that the nsitive frequency of birds was between 1–5 kH� (Dooling, 1978), other studies found that the upper limit of their hearing is usually at 20 kH� (reached after condi-tioning) and cochlear potentials have been detected at 25–30 kH� by using greater intensities of sound (Schwart�-kopff, 1955). �e therefore agree with Narins et al. (2004) that a systematic re-examination of avian hearing ability is required.
The 14 avian species involved in our study were from eight families of Pasriforms and one family of Gallifor-mes, suggesting such high frequency harmonics exist at least in the vocali�ations of some avian lineages such as Paridae, Aegithalidae and Emberi�idae. Due to the small sample of the species, however, it is hard for us to conclude in which groups of avian species the high frequency components are more prevalent. As a result, we encourage more studies to be conducted in this area to explore the prevalence of high frequency vocali�ation in avian com-munication.
