I. INTRODUCTION. J. Acoust. Soc. Am. 113 (3), March /2003/113(3)/1749/8/$ Acoustical Society of America

Transcription

1 Reverberation of rapid and slow trills: Implications for signal adaptations to long-range communication Marc Naguib a) Department of Animal Behavior, University of Bielefeld, P.O. Box , Bielefeld, Germany Received 15 June 2002; revised 7 November 2002; accepted 25 November 2002 Many acoustic signals in animals include trills, i.e., rapid repetitions of similar elements. Elements within these trills usually are frequency modulated and are degraded by reverberation during long-range transmission. Reverberation primarily affects consecutive elements with the same frequency characteristics and thus imposes a major constraint in the evolution of design and perception of long-range signals containing trills. Here transmission of frequency-unmodulated trills with different element repetition rates was studied. Trills were generated at different frequencies to assess frequency dependence of reverberation and then broadcast under three acoustic conditions an open field and to assess seasonal changes in transmission properties, a deciduous forest before and after foliage had emerged. Reverberation was quantified at different positions within trills. The results show strong effects of vegetation density season, transmission distance, frequency, element repetition rate, and element position within the trill on effects of reverberation. The experiments indicate that fast trills transmit less well than slow trills and thus are less effective in long-range communication. They show in particular that selection on trills should not act only on element repetition rate within trills but also on the trill duration as effects of reverberation increased with trill duration Acoustical Society of America. DOI: / PACS numbers: Ev, Lb, Ka WWA I. INTRODUCTION Acoustic signals used in long distance communication are subject to considerable qualitative and quantitative changes during propagation through the environment Piercy et al., 1977; Wiley and Richards, Thus, the acoustic conditions of the transmission channel have profound implications for the evolution of signal design Morton, 1975; Wiley, 1991; Tubaro et al., 1993 and for the evolution on receivers perceptual mechanisms and response strategies Wiley and Richards, 1978; Klump, 1996; Naguib and Wiley, Degradation of signals is habitat dependent and generally results in three structural changes in sound: reverberation, irregular amplitude fluctuations, and frequencydependent attenuation, all of which differentially affect signal perception Dooling, 1982; Naguib and Wiley, Although the general principles of sound degradation are well described Wiley and Richards, 1978, 1982, there are still few studies that have separately quantified these aspects of signal degradation. Thus, still little is known about how sound degradation affects perception of particular signal structures, an issue important for understanding whether or not information coded in a signal is transmitted to a receiver under natural conditions. In habitats with many sound-reflecting surfaces such as in forests, the most prominent form of degradation is reverberation which results in smearing of the temporal structure of the signal. Effects of reverberation are most prominent when subsequent elements are in the same frequency band, such as in trills, in particular when they consist of rapidly repeated elements. Trills are common in animal sounds and a Electronic mail: marc.naguib@uni-bielefeld.de are particularly important in communication Klump and Gerhardt, 1987; Stumpner and Ronacher, 1994; Podos, They vary considerably in element repetition rate and duration so that accumulating reverberation will differently affect perception of single elements depending on their rate and position within the trill. In addition, effects of reverberation on trills not only depend on the structure of the trill but also on vegetation density so that seasonal changes in vegetation will have profound effects on strength of reverberation. Such changes in transmission characteristics are important when adaptations of signals to the habitat are assessed. Nonmigratory songbirds, for instance, and migrants that arrive early at their breeding grounds in deciduous forests will encounter prominent changes in the acoustic condition of their habitats as foliage emerges. These changes in acoustic conditions are known to affect receivers assessment of signals Naguib, 1996 and adaptations of signals to the transmission characteristics also are likely to depend on the condition under which selection acts strongest on long-range communication. Most studies that have quantified changes in the temporal structure of a signal used composite measures of degradation that are based on changes in the waveform of the signal Gish and Morton, 1981; Dabelsteen et al., 1993; Holland et al., 1998; Brown and Handford, 2000 or that are based on spectrographic cross correlation Fotheringham and Ratcliffe, 1995; Brown and Handford, 2000; Kime et al., Although these measures on the waveform provide important holistic information on sound degradation, they are difficult to interpret from a perceptual perspective as they conflate the different forms of degradation. The waveform is affected by all degrading processes such as reverberation, J. Acoust. Soc. Am. 113 (3), March /2003/113(3)/1749/8/$ Acoustical Society of America 1749

2 frequency-dependent attenuation and irregular amplitude fluctuations. Naguib and Wiley 2001 pointed out that ears are frequency analyzers and separately perceive changes in amplitude and frequency, so that separate measures of degradation in the different dimensions provide biologically more meaningful information in terms of how the degradation will affect signal perception and thus signal evolution. Animals ears can have a high-frequency selectivity Wilkinson and Howse, 1975; Dooling, 1982 so that reverberation at one frequency will not affect perception of a subsequent element at another frequency. Moreover, studies on auditory distance assessment have shown that the different components of degradation are processed separately in birds Naguib, 1995, 1997a, b and humans Mershon and King, 1975; Little et al., All of these approaches underline the separate biological importance of the different kinds of degradation. So far, studies that directly quantified reverberation either used single tone pulses Richards and Wiley, 1980; Waser and Brown, 1986; Naguib et al., 2000 or single frequency-modulated notes from bird song and calculated the tail-to-signal ratio Brown and Handford, 2000; Holland et al., 2001 or the slope of the tail Holland et al., 2001, but effects of reverberation on subsequent elements to date have not be examined. In this study I measured reverberation of trills with different repetition rates of elements and different frequencies broadcast over distances up to 120 m in three habitats differing seasonally in vegetation density. I focused on trills without frequency modulation in order to quantify perceptually salient parts of reverberation in those kinds of signals that are particularly affected by reverberation. Trills were expected to reverberate increasingly with distance and vegetation density. Reverberation was expected to increase particularly in dense vegetation with increasing element repetition rate and element position within a trill. II. METHODS A. Preparation of sound signals All trills were synthesized on a PC using Cool Edit 2000 Syntrillium Software Cooperation, Phoenix, USA with a sampling rate of Hz and 16-bit accuracy. In order to measure reverberation on the waveform of the signal without confounding effects of frequency modulation, trills were synthesized with frequency-unmodulated elements. The primary objective in measuring reverberation was to quantify effects of reverberation on closely spaced elements of similar frequency composition trills. Synthesized trills were composed of eight elements of 10-ms duration and silent intervals between elements of either 10 ms fast trills, 40ms medium trills, or90ms slow trills. The element rates are common in song birds with the fast trills being significantly more frequent in open habitat species than in forest living species unpublished data; also see Wiley, In order to assess effects of frequency on reverberation all trills were generated at 1, 3.5, and 7 khz. B. Transmission experiments 1. Habitat Sound transmission experiments were conducted in November 2000 and in July 2001 at two transects at the same sites in a deciduous forests in Bielefeld, Germany. In November the experiments additionally were conducted on an open field but these experiments were not repeated in July as here no seasonal effects on reverberation were expected. In November vegetation was leafless and in July leaves had emerged fully. Acoustic conditions in terms of structure and density of sound-reflecting surfaces in November thus resembled those early in the breeding season when many birds start to sing but when leaves have not yet emerged. The acoustic conditions in July resembled those during the main breeding season. Trees in the forest mostly were beech and oak. Tree trunks varied in diameter with the diameter of the thickest trunks reaching 80 cm. Trees were spaced irregularly with about 3 to 5 m distance between closest trees. The ground mostly was plain and covered with leaves. There was little undergrowth and leaf carrying branches mostly grew above 5 m. The transmission experiments for the open habitat were conducted on a crop field with vegetation height of about 10 cm and with no further objects in the transmission path. The closest forest edge was over 50 m away from the microphone and loudspeaker positions. All experiments were conducted between 0600 and 0900 hours on days with calm weather. 2. Procedure All signals were broadcast from a Toshiba Satellite 2210 CDT computer connected to a Blaupunkt MPA 2 amplifier and a Canton XS passive loudspeaker. The loudspeaker was connected to a telescopic pole and positioned at 4-m height which is a common perch height in singing birds. Sounds were broadcast with 90 db, as measured at 1 m using a 2203 Brüel & Kjær precision sound level meter C settings, fast response on a 5-s 1-kHz tone of equal peak amplitude as all other signals in the broadcast sound file. Signals were rerecorded successively at 20, 40, 80, and 120 m using a Sennheiser ME 64/K6 microphone connected to a Sony TCD- D100 DAT recorder. The microphone is not fully omnidirectional but has similar sensitivity to sound arriving frontally and laterally and thus detected the presumably most important components of reverberation. The microphone was connected to a telescopic pole and, like the loudspeaker, was positioned at 4-m height to simulate the situation at which birds on their song posts perceive rivals song. Trills were repeated twice at each distance in November and four times at each distance in July. In July the broadcast file was expanded to also include trills with different element repetition rates 40 and 90 ms in order to assess how the expected stronger reverberation would affect trills with different element spacing. At all sites, sounds were broadcast over two transects. In the open field sounds were broadcast along the same transect but in opposite directions. In the forest sites, the two transects were perpendicular to each other with the loudspeaker being placed at the same position for both transects but faced in different directions 90 for the two 1750 J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003 Marc Naguib: Reverberation of trills

3 transects. The same transects were chosen in July and November to permit analyses of changes in acoustic conditions with seasonally changing vegetation density. C. Quantification of degradation For analyses, sounds were resampled using Cool Edit on a PC with Hz and 16 bit. All sounds were then bandpass filtered with filter settings specific to the frequency of the signal and normalized in amplitude. All bandpass filters were set at 100 Hz from the carrier frequency of the signal FFT size 8192, Hamming window, 60-dB attenuation. For measures of reverberation I measured the sound energy in trill elements and the energy in silent intervals between trill elements using Saslab Pro 3.95 software Raimund Specht, Berlin. As measure of reverberation, the ratio of the energy in the 10-ms tone and the energy of the following 10 ms was calculated. The energy was calculated as the squared amplitude values in Volts multiplied by the sampling time (V*V*s). This reverberation index reflects the extent to which the silent intervals between successive notes is filled by energy resulting from reverberation by the previous note s. The reverberation index was calculated for the first, fourth, and seventh elements in order to assess accumulating reverberation over the trill. An index of one indicates that the energy of the silent interval equaled that of the preceding tone, indicating that the silence between notes in the fast trills will be difficult to detect Dooling, 1982; Okanoya and Dooling, The same procedure was used for measurements on trills with slower element repetition rates in order to permit a direct comparison among trills. As the silence between successive elements in trills with slower repetition rates 40 and 90 ms was longer than the part measured 10 ms, an index of one in these trills does not indicate that the separate elements are difficult to resolve but indicates that the end of an element will be more difficult to determine. Using the 10-ms intervals for measurements in all trills better allowed assessment as to what extent reverberation accumulated from previous elements in the different trills. D. Data analysis All data were analyzed using multi-factorial ANOVAs. Post hoc pairwise comparisons where run using bonferioni post hoc tests, which adjusted p-values to multiple comparisons. In order to analyze seasonal effects of reverberation on the trills with element repetition rates of 10 ms, a four-factor ANOVA 2 season 4 distance, 0, 20, 40, and 80 m 3 frequency, 1, 3.5, 7 khz 3 element position within trill, first, fourth, seventh element was used with four repetitions for July measures and two repetitions for November measures. The 120-m data could not be included in this comparison as the 3.5-kHz data in the experiments conducted in July were overlaid by songs from blue tits Parus caeruleus singing near the microphone positions so that sounds could not be analyzed accurately. Due to the additional trills with slower element repetition rates broadcast in July, a five factor ANOVA was used for a more detailed analysis of effects of reverberation on trills in July 4 distance 3 element rate within trill, 10, 40, and 90 ms 3 frequency 3 element position within trill with four repetitions. For the same reason mentioned above, transmission distances of 120 m were not included in the analysis. Due to the problems encountered at 120 m in July, a third four-factor ANOVA was used on the data obtained in the forest in November for a more detailed analysis as here the 120-m data were included 2 habitat 5 distance 3 frequency 3 element position within trill with two repetitions. Habitat effects in November were analyzed with a separate ANOVA as the data from the open field were not analyzed in more detail due to the low level of reverberation. All statistic tests were performed with SPSS on a PC. Signals were not included in the analysis when they were overlaid by excessive background noise. Only main and two-factor interactions are considered. III. RESULTS A. Reverberation of trills with 10-ms element repetition intervals 1. Comparison between the open habitat and the deciduous forest without foliage Reverberation was significantly stronger in the deciduous forest without leaves November than on the open field (F 1, , P 0.001). Reverberation was very low at the open field site and did not increase with distance whereas it increased significantly with distance in the forest Figs. 1 a and b ; also see analyses below. 2. Comparison between closed habitats with different vegetation densities a. General. The overall analysis of variance over distances up to 80 m showed significant effects of season deciduous forest with and without foliage Fig. 2 and transmission distance on reverberation of trills Table I. In addition, the interactions between transmission distance and season, between season and frequency, and between distance and frequency, respectively, were significant Table I. In other words, trills of different frequency were affected differently by reverberation over distance in the two habitats. Effects of reverberation showed a different pattern of increase with distance and with frequency before than after foliage had emerged, as detailed below. b. Reverberation at different transmission distances. Effects of reverberation increased significantly with transmission distance before and after foliage had developed Tables II and III; Figs. 1 b and c. However, the increase of reverberation with transmission distance in the forest differed among seasons, as indicated by the significant interaction between distance and season Table I. In winter, reverberation did not increase significantly from 20 to 80 m but showed a strong increase at a transmission distance of 120 m Fig. 1 b. In contrast, in summer, when foliage had developed fully, the increase of reverberation with transmission distance was more linear with significant increases over all transmission distances Fig. 1 c. The highest level of reverberation reached in summer at transmission distances up to J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003 Marc Naguib: Reverberation of trills 1751

4 FIG. 2. Average reverberation index for the deciduous forest in winter before foliage had emerged and in summer with full vegetation. Significant differences based on pairwise post hoc tests are indicated by stars (***p 0.001). was least for the first positions under both acoustic conditions, indicating that accumulating reverberation particularly affected subsequent element positions Figs. 5 a and d. d. Reverberation at different frequencies. Overall, there was a significant effect of sound frequency on reverberation in the forest under both conditions Tables II and III. In particular, there was a strong effect of season on reverberation at different frequencies Figs. 6 a and b. In winter, reverberation decreased significantly with frequency, whereas in summer reverberation increased significantly with frequency. Reverberation at 1 khz was similar under both conditions but higher frequencies reverberated stronger in summer with an average reverberation index at 7 khz being almost three times as high than in winter Figs. 6 a and b. FIG. 1. Average reverberation indices for different transmission distances a on an open field in November, b in a deciduous forest in November, and c in a deciduous forest in July. Significant differences based on pairwise post hoc tests are indicated by stars (**p 0.01;***p 0.001). 80 m on average was about twice as high as that reached in winter, with indices at 80 m of 0.20 in winter and 0.47 in summer. c. Reverberation at different element positions within a trill. Effects of reverberation increased significantly with element position within the trill in both seasons Tables II and III. Reverberation increased significantly from the first to seventh element position before foliage had emerged Fig. 3 and showed a significant increase already from the first to the fourth element after foliage had emerged (p 0.02) Fig. 4 a. Moreover, the increase of reverberation with distance B. Reverberation of trills with different element repetition rates Trills with different element repetition rates accumulated reverberation to different extents within the first 10 ms following an element Table III, Fig. 4 a. Trills with shorter element repetition rates showed significantly higher degrees of reverberation than did those with slower element repeti- TABLE I. ANOVA on reverberation in 10-ms trills on combined data from the deciduous forest with and without foliage November and July. Source d.f. F ratio P Distance Season Frequency Element-position Distance season Distance frequency Season frequency Distance element position Season element position Frequency element position J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003 Marc Naguib: Reverberation of trills

5 TABLE II. ANOVA on reverberation in 10-ms trills over transmission distances up to 120 m in the deciduous forest without foliage November. Source d.f. F ratio P Distance Frequency Element position Transect Distance frequency Distance element position Distance transect Frequency element position Frequency transect Element position transect tion rates Table III, Fig. 4 a. In trills with 10-ms element repetition intervals, reverberation increased with element position within a trill whereas it did not in trills with intervals of 40 and 90 ms, indicating that reverberation accumulated over several elements only in the fast trills Fig. 4 b. This general pattern of reverberation also is reflected in the way it accumulated over distance. Only in the 10-ms trills reverberation increased more with increasing distance in the latter element positions than in the first element position. In trills with slower element repetition rates, the different element positions were not affected differently by reverberation at different distances Figs. 5 b and c. IV. DISCUSSION A. General The experiments show that trills over distance reverberated differently depending on vegetation density, element repetition rate within the trill, frequency, and transmission transect within a given habitat. As predicted, they further show that silences between elements at later positions within trills with fast element repetition rates are most strongly affected by reverberation, supporting the prediction that not only element rate within trills but also that trill duration is TABLE III. ANOVA on reverberation in trills with different element repetition intervals 10, 40, and 90 ms over transmission distances up to 80 m in the deciduous forest with foliage July. Source d.f. F ratio P Distance Frequency Element position Trill rate Transect Distance trill rate Distance frequency Distance element position Distance transect Element position trill rate Element position transect Frequency trill rate Frequency transect Trill rate transect Element position frequency FIG. 3. Reverberation index for the different element positions within the trill in trills with element repetition rates of 10 ms broadcast in the deciduous forest before foliage emerged November. Significant differences based on pairwise post hoc tests are indicated by stars (**p 0.01). significantly affected by reverberation. The results further contribute to the general understanding of effects of reverberation on the evolution of signal structure and signal perception as effects of reverberation on trills had not been quantified yet in perceptually relevant frequency bands. B. Effects of trill structure and transmission distance on reverberation in forests The finding that trills with rapid element repetition rates reverberate more severely at later positions within the trill have important implications for understanding the evolution of trill structures in animal signals. Trills are common in orthopterans, anurans, and song birds and have been shown to be functionally important in female choice. Moreover, they are difficult to produce Podos, 1997, and thus are likely to encode important information on a signaler s quality. As accumulating reverberation within the same frequency bands masks subsequent rapidly repeated elements with similar frequency, as shown here, such trills are not suited well for long-range communication in habitats with dense vegetation. In fact, previous studies on birds have shown that species in open habitats produce significantly faster trills than do species in forests Wiley, Exceptions are Kentucky warblers Oporornis formosa, a typical forest bird, which produces similar elements at short repetition intervals that are typical for open habitat species Wiley and Godard, 1996, and blue tits which produce song types with fast trills regardless of vegetation density Doutrelant et al., 1999; Doutrelant and Lambrechts, Wiley and Godard pointed out that trills could have both disadvantages and advantages for birds in forests disadvantages because of difficulties in detecting trills or discriminating trill rates, advantages because of possibilities for enhancing ranging by receivers. Both processes should be affected by the length of the trill, because, as shown here, reverberation increases with position in a trill. We can thus predict that long trills would have nothing but disadvantages J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003 Marc Naguib: Reverberation of trills 1753

6 for long-range communication in forests. Indeed few forest birds have long trills in their long-range songs Wiley, On the other hand, predictions about the advantages and disadvantages of short trills are more difficult to make. It will thus be difficult to predict the optimal length for trills without further study. The accumulation of reverberation is probably not linear but instead levels off after a certain number of elements, which depends on element repetition rate. Furthermore, discrimination of trill rates and estimation of depth of amplitude modulation probably varies in complex ways with trill length. For ranging, a short trill might not accumulate enough reverberation or might not provide a long enough signal to estimate depth of amplitude modulation accurately. A long trill might accumulate too much reverberation and thus reduce the sensitivity of an estimate of amplitude modulation. Resolving these issues will require study of perception as well as transmission of trills. FIG. 4. Reverberation index for trills with different element repetition rates broadcast in a deciduous forest with foliage a overall means for the trills and b means separated by element positions within the trills with different element repetition rates. Significant differences indicated in a are indicated by stars (**p 0.01;***p 0.001). C. Effects of habitat and seasonal changes of the acoustic environment on reverberation Reverberation within fast trills accumulated with element position in the forest under both acoustic condition, i.e., without foliage and with foliage. However, the effect of element position became significant only after transmission distances of 120 m in the forest without foliage. When foliage had emerged in the same forest, the effect of reverberation on elements was stronger already at shorter distances. These seasonal changes of reverberation within a forest need to be considered when adaptations of signals to the acoustic conditions of a habitat and when responses of receivers to such degraded signals are studied. In particular the songs of yearround resident species have evolved under selection of vary- FIG. 5. Average reverberation indices at different distances for the different element positions within a trills with element repetition rate of 10 ms, b 40 ms, c 90 ms broadcast in deciduous forest with foliage and d trills with element repetition rate of 10 ms broadcast in deciduous forest before foliage emerged J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003 Marc Naguib: Reverberation of trills

7 D. Effects of sound frequency on reverberation In all habitats there was a strong interaction between sound frequency and transmission distance on extent of reverberation. Interestingly, these effects were different in the same forest with foliage than without foliage. Under both conditions reverberation was similar at 1 khz, indicating that foliage in a forest with moderate vegetation density and with little undergrowth does not much affect reverberation at this frequency. Such an effect is expected as objects with sound reflecting surfaces best reflect frequencies with wavelengths smaller than the size of the object. A sound of 1 khz has a wavelength of 33 cm and leaves have considerably smaller diameters of mostly less than 10 cm. In summer, foliage will considerably increase the surface that reflect sounds of higher frequencies, explaining the significantly stronger reverberation at higher frequencies compared to 1 khz in the forest with fully developed foliage and compared to higher frequencies in the forest before foliage had emerged. Why higher frequencies before foliage emerged showed less reverberation than did 1 khz is more difficult to interpret. Possibly reflection by the ground or additional defraction from tree trunks may have resulted in stronger reverberation at 1 khz compared to higher frequencies. FIG. 6. Reverberation indices for trills with element repetition rates of 10 ms synthesized at different frequencies and broadcast in a deciduous forest before foliage emerged November and b deciduous with foliage July. Significant differences are indicated by stars (*p 0.01;**p 0.01;***p 0.001). ing acoustic conditions at different times of the year. As males in resident species usually start advertising territories before foliage has developed, these better acoustic conditions even may be more important in the evolution of signal design than the less favorite conditions later in the season when foliage has fully developed. There is evidence already that birds adapt response criteria to such seasonal changes in the acoustic conditions of their habitat. Carolina wrens showed differences in assessed auditory distance of songs before and after foliage had emerged even though songs were reverberated at the same level Naguib, This experiment shows that such seasonal changes in reverberation not only are likely to influence signal structure but also the behavior of receivers. The finding that even the minor effects of reverberation on trills in the open habitat did not accumulate with distance or element position now provides a quantitative background for understanding the evolution of frequent use of fast long trills in song birds living in open habitats. E. Conclusions The synthesized trills consisting of frequencyunmodulated elements allowed to separate effects of frequency-dependent attenuation from reverberation. Such a distinction is important when trying to understand how sound degradation affects signal perception and the evolution of signal design Naguib and Wiley, Although the exact patterns of reverberation of trills will vary with habitat and the exact propagation path, the principles of how trills reverberate are likely to be a general characteristic of habitats with multiple sound reflecting surfaces. This study illustrates that rapid repetitions of elements with the same frequency suffer significantly more from reverberation than slower repetitions and thus will decrease the probability of signal detection and recognition. The data also show that it does not suffice to measure the spacing of elements within a trill in order to assess adaptations of signals to the acoustic conditions of a habitat. Future studies need to consider also the number of elements within a trills, as reverberation increases within trills consisting of rapidly repeated elements and thus will have different effects on trills with different duration. ACKNOWLEDGMENTS I thank Christine Brenninkmeyer, Benjamin Hundsdorfer, Nicole Meyer, and Annegret Walter for their help in conducting the transmission experiments and for part of the data analysis. I also thank Henrik Brumm, R. Haven Wiley, and two anonymous referees for comments on a previous version of the manuscript. The study was funded by a research innovation grant FIF2 provided by the University Bielefeld. J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003 Marc Naguib: Reverberation of trills 1755

8 Brown, T. J., and Handford, P Sound design for vocalizations: quality in the woods, consistency in the fields, Condor 102, Dabelsteen, T., Larsen, O. N., and Pedersen, S. B Habitat-induced degradation of sound signals Quantifying the effects of communication sounds and bird location on blur ratio, excess attenuation, and signal-tonoise ratio in blackbird song, J. Acoust. Soc. Am. 93, Dooling, R. J Auditory perception in birds, in Acoustic Communication in Birds, Vol. 1 Academic, New York, pp Doutrelant, C., and Lambrechts, M. M Macrogeographic variation in song a test of competition and habitat effects in blue tits, Ethology 107, Doutrelant, C., Leitao, A., Giorgi, H., and Lambrechts, M. M Geographical variation in blue tit song, the result of an adjustment to vegetation type? Behaviour 136, Fotheringham, R. J., and Ratcliffe, L Song degradation and estimation of acoustic distance in black-capped chickadees Parus atricapillus, Can. J. Zool. 73, Gish, S. L., and Morton, E. S Structural adaptations to local habitat acoustics in Carolina wren songs, Z. Tierpsychol. 56, Holland, J., Dabelsteen, T., Paris, A. L., and Pedersen, S. B Potential ranging cues contained within the energetic pauses of transmitted wren song, Bioacoustics 12, Holland, J., Dabelsteen, T., Pedersen, S. B., and Larsen, O. N Degradation of wren Troglodytes troglodytes song: Implications for information transfer and ranging, J. Acoust. Soc. Am. 103, Kime, N. M., Turner, W. R., and Ryan, M. J The transmission of advertisement calls in Central American frogs, Behav. Ecol. 11, Klump, G Bird communication in the noisy world, in Ecology and Evolution of Acoustic Communication in Birds, Cornell U. P., Ithaca, NY, pp Klump, G., and Gerhardt, H. C Use of non-arbitrary acoustic criteria in mate choice by female gray tree frogs, Nature London 326, Little, A. D., Mershon, D. H., and Cox, P. H Spectral content as a cue to perceived auditory distance, Perception 21, Mershon, D. H., and King, L. E Intensity and reverberation as factors in the auditory perception of egocentric distance, Percept. Psychophys. 18, Morton, E. S Ecological sources of selection on avian sounds, Am. Nat. 109, Nagulb, M Auditory distance assessment of singing conspecifics in Carolina wrens: the role of reverberation and frequency-dependent attenuation, Anim. Behav. 50, Naguib, M Ranging by song in Carolina wrens Thryothorus ludovicianus: Effects of environmental acoustics and strength of song degradation, Behaviour 133, Naguib, M. 1997a. Use of song amplitude for ranging in Carolina wrens, Thryothorus ludovicianus, Ethology 103, Naguib, M. 1997b. Ranging of songs in Carolina wrens: effects of familiarity with the song type on use of different cues, Behav. Ecol. Sociobiol. 40, ; erratum 41, 203. Naguib, M., and Wiley, R. H Estimating the distance to a source of sound: mechanisms and adaptations for long-range communication, Anim. Behav. 62, Naguib, M., Klump, G. M., Hillmann, E., Griessmann, B., and Teige, T Assessment of auditory distance in a territorial songbird: accurate feat or rule of thumb? Anim. Behav. 59, Okanoya, K., and Dooling, R. J Detection of gaps in noise by budgerigars Melopsittacus undulatus and zebra finches Poephila guttata, Hear. Res. 50, Piercy, J. E., Embelton, T. F. W., and Sutherland, L. C Review of noise propagation in the atmosphere, J. Acoust. Soc. Am. 61, Podos, J A performance constraint on the evolution of trilled vocalizations in a songbird family Passeriformes: Emberizidae, Evolution Lawrence, Kans. 51, Richards, D. G., and Wiley, R. H Reverberations and amplitude fluctuations in the propagation of sound in a forest: implications for animal communications, Am. Nat. 115, Stumpner, A., and Ronacher, B Neurophysiological aspects of song pattern recognition and sound localization in grasshoppers, Am. Zool. 34, Tubaro, P. L., Segura, E. T., and Handford, P Geographic variation in the song of the rufous-collard sparrow in eastern Argentina, Condor 95, Waser, P. M., and Brown, C. H Habitat acoustics and primate communication, Am. J. Primatol 10, Wiley, R. H Associations of song properties with habitats for territorial oscine birds of eastern North America, Am. Nat. 138, Wiley, R. H., and Richards, D. G Physical constraints on acoustic communication in the atmosphere: implications for the evolution of animal vocalizations, Behav. Ecol. Sociobiol. 3, Wiley, R. H., and Richards, D. G Adaptations for acoustic communication in birds: sound transmission and signal detection, in Acoustic Communication in Birds, Vol. 2 Academic, New York, pp Wiley, R. H., and Godard, R Ranging of conspecific songs by Kentucky warblers and its implications for interactions of territorial males, Behaviour 133, Wilkinson, R., and Howse, P. E Time resolution of acoustic signals in birds, Nature London 258, J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003 Marc Naguib: Reverberation of trills