|
Article Excerpt
Vocal and other animal displays typically are correlates of speciation and population divergence (Lanyon 1969; Payne 1986; Martens 1996; Price and Lanyon 2002; Isler et al. 2005, 2007), and are used routinely in modern species-level systematics (Cuervo et al. 2005, Athreya 2006, Gonzaga et al. 2007). Vocalizations that are not learned are likely to be especially informative in elucidating relationships; those vocalizations are widespread as they typify most non-passerine bird families (Baptista and Schuchmann 1990, Farabaugh and Dooling 1996, Price and Lanyon 2002, Jarvis 2006). Some vocal traits likely represent convergence due to selection for effective transmission in similar habitats (McCracken and Sheldon 1997, Seddon 2005), and some vocal homoplasies could occur due to similar functions in different acoustic environments (Marler 1955). Most studies of vocal relationships between species have used single vocal homologues (Davis 1962, 1965; Winkler and Short 1978; Slabbekoorn et al. 1999; Seddon 2005; Shelley and Blumstein 2005), but a few have used entire repertoires (Price and Lanyon 2002). The latter approach is preferable because analysis of at least a few sound classes is necessary to understand acoustic differentiation in groups with complex vocalizations or large repertoires (Price and Lanyon 2002, Isler et al. 2007).
The Alcidae (subfamily Alcinae; Thomas et al. 2004) diverged from the lineage leading to skuas and jaegers (Skua and Stercorarius; Paton et al. 2003, Thomas et al. 2004, Paton and Baker 2006, Fain and Houde 2007) about 60 Mya (Pereira and Baker 2008). Relationships among the auklets (tribe Aethiini) remain unclear, but it is agreed the five extant species constitute a monophyletic group: Cassin's Auklet (Ptychoramphus aleuticus), Crested Auklet (Aethia cristatella), Least Auklet (A. pusilla), Parakeet Auklet (A. psittacula), and Whiskered Auklet (A. pygmaea) (Strauch 1985, Mourn et al. 1994, Friesen et al. 1996, Thomas et al. 2004).
Auklets are highly vocal at their breeding colonies, and mixed-species colonies are nearly universal in Alaska. The social environments of auklets are similar across species: they are colonial and socially monogamous, and males and females are nearly identical in plumage and body size. Only Cassin's and (in most colonies) Whiskered auklets are nocturnally active. However, they inhabit acoustically diverse habitats (Jones 1993a, b; Manuwal and Thoresen 1993; Gaston and Jones 1998; Jones 1999; Jones et al. 2001), and divergent adaptations to different physical environments are possible (Endler 1993, Nicholls et al. 2006). For example, four auklet species breed in rocky habitats or on cliffs, but Cassin's Auklets breed on slopes with soft earth, covered with dense grass and occasionally trees (Thoresen 1964, Manuwal and Thoresen 1993).
We conducted a comparative study of auklet vocalizations to elucidate vocal relationships among auklet species; this is the first comparative study for the Alcidae. The objectives of our study were to: (1) quantify and characterize the structural organization of vocal repertoires, (2) relate vocal repertoires to species differences in phylogeny and ecology, and (3) identify vocal displays of significance to conservation-restoration projects. We describe vocalizations and vocal behavior, standardize terminology, and identify possible applications.
METHODS
Study Sites and Recording Methods.--Recordings were made in the Aleutian Islands, Alaska: Buldir Island (52[degrees]22' N, 175[degrees]54' E), 25 May to 7 June 2005 and 25 May to 24 July 2006; and Egg Island (53[degrees]52' N, 166[degrees]03'W), 15 June to 11 July 2005. Both are grass-covered, treeless volcanic islands with beach boulders and exposed talus slopes, and large breeding colonies of auklets (Byrd and Day 1986, Bradstreet and Herter 1991). All five auklet species breed on Buldir Island; Cassin's, Parakeet, and Whiskered Auklets breed on Egg Island.
Recordings were made opportunistically during times of peak activity: Cassin's Auklet, 02000500 hrs (Aleutian standard time); Crested and Least auklets, 0900-1400 hrs; Parakeet Auklet, 0600-1200 hrs; and Whiskered Auklet, 18000200 and 0400-0600 hrs. We recorded birds separated by at least 10 m to minimize the possibility of recording individuals more than once. Cassin's and Whiskered auklets were recorded at night with the aid of a red lithium electrode diode headlamp (Zubakin and Konyukhov 1999). All recordings were of undisturbed birds.
Equipment.--We recorded birds with a Sony TCD-D10PROII Digital Audio Tape recorder (sampling rates 32, 44.1, or 44.2 kHz), or Fostex FR-2 solid-state recorder (sampling rate 48.1 kHz) with Senheiser MKH 70 or MKH 816 directional microphones (with wind guard and wind sock) and 3-30 m cables. Microphones were tripod-mounted or hand-held. Recording sessions with individual birds were 30 min to 3 hrs in duration and contained multiple continuous recordings of 2-10 min in duration, each accompanied by behavioral observations. We recorded birds at distances of ~2-6 m and recording sessions totaled ~80 hrs.
All birds were assumed to be breeding adults as calling birds are mostly adults, and adults predominate at colony sites in the early and mid breeding season (Jones 1993a, b; Manuwal and Thoresen 1993; Jones et al. 2001; Zubakin and Konyukhov 2001). Gender of calling birds was known only for Crested Auklets, based on the distinctive bill shape and courtship display (Jones 1993c).
Acoustic Description and Measurement.--We selected recordings of good quality for analysis with Raven 1.2.1 (Bioacoustics Program, Cornell Laboratory of Ornithology, Ithaca, NY, USA). Blackman window was used for all analyses. Other analytical settings for Cassin's and Parakeet auklets were: window size, 800 samples; 3 dB filter bandwidth, 90.5 Hz; time grid overlaps, 75%; and frequency grid spacing, 43.1 Hz. Corresponding settings for Crested and Whiskered auklets were 512 samples, 141 Hz, 90%, and 86.1 Hz; and for Least Auklet were 250 samples, 289 Hz, 50%, and 172 Hz. The dominant harmonic was judged from the spectrogram slice view.
We used a combination of audible differences, overall visual impressions of vocal sequences, and constituent sound notes on spectrograms (Marler and Pickett 1984, Jones et al. 1989, Hailman and Ficken 1996, Marler and Slabbekoorn 2004, Seddon 2005), and explicit measurements of physical properties to characterize vocal repertoires. We referred to classes of basic sound elements as "notes" following Marler and Pickert (1984), and Marler and Slabbekoorn (2004). We recognized note types to separate overall acoustic variation, and to characterize overall acoustic diversity within species, rather than to approximate natural display classes. We pooled brief elements (e.g., the first 3 in Fig. 1A [a, b, a]) as a single note type (note type V), separated graded series into several note types (e.g., note types I and III in Fig. 1A), and pooled acoustically similar parts of different calls (e.g., note types II and III in Fig. 1A, C).
We measured duration, frequency, modulation of the carrier frequency, and harmonic structure ([F.sub.n], frequency of nth harmonic; Fo, fundamental frequency; [F.sub.[mu]], frequency of the most strongly expressed harmonic [i.e., the harmonic with the highest intensity in the power spectrum]; [CF.sub.n], carrier frequency of the nth harmonic; and FM, frequency modulation of CF). Frequency was estimated from spectrograms by selecting the beginning, end, highest, and lowest points of the signal. The frequency of the signal is that of the selected rectangle's center (in Raven). We estimated FM by selecting the lowest and highest frequencies of the signal, and taking the frequency difference between these two points from the selection table (in Raven). Display types were identified for each species based on the composition of note types and audile plus visual characteristics, most of which were easily and reliably distinguishable in the field. Our nomenclature partly follows previous published descriptions (Thoresen 1964; Byrd and Williams 1993; Jones 1993a, b; Manuwal and Thoresen 1993; Jones et al. 2001), but we provide new names for previously unrecognized call types.
[FIGURE 1 OMITTED]
RESULTS
Auklet repertoires encompassed 22 call types, which were built on 1-5 frequency modulated and harmonically rich note types (28 types across species) arranged sequentially in varied combinations (Fig. 1). Most notes were strongly harmonic, brief, and with pronounced FM (Table 1, Fig. 1). Characteristics of notes and their organization in displays were uniform within species.
Cassin's Auklet.--Adults had three kinds of calls, all with harsh screeching qualities; some were long (~50 sec). They ranged from 2 to 50 sec (means = 4-18 sec). All calls consisted of rhythmically repeated notes or note couplets; the latter appear to represent vocalizations during alternating inspirations and expirations. The low number of call types was paralleled by little structural diversity across the five note types we recognized (e.g., [F.sub.[mu]] varied only from 2.2 to 3.1 kHz; Table 1). Call types differed little in diversity (range = 2-4 note types).
Kut-l-eer (I) (Fig. 1A; Tables 1, 2). This long complex call began and ended with simple notes, which graded into a more complex main part. The introductory note series at times was long (~300 msec); the terminal series was brief. This call was structured on note couplets, which were apparent even...
|
