SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. SVT vs. MVT species
  6. Interval between vocalizations
  7. Discussion
  8. Acknowledgements
  9. References

It has been hypothesized that variation (e.g., of repertoire elements) in prolonged vocalization sessions of passerine birds can serve to minimize habituation by conspecifics. The repertoire of vocalization types limited in suboscine passerines, raising the question of how a limited set of elements can create patterns that minimize habituation. This question was studied by computational analysis of recorded singing sessions of 20 suboscine species from the subfamily Tyranninae, family Tyrannidae (tyrant flycatchers). The recordings of 12 of the species included two or more distinct vocalization types (VTs). In these species, the interval between vocalization units when the VT changed was on average shorter than that when the VT remained the same. In addition, when the VT changed, the mean interval length between successive vocalization units differed depending on which VT preceded the interval and which VT followed it. On the other hand, species with just a single VT in the session analyzed showed a surprisingly high degree of absolute difference between adjacent vocalization units with respect to both the length of the vocalization and percentage of time elapsed until peak amplitude. A change in the rhythm of vocalization accompanying a change in VT provided a potential means of drawing a conspecific listener's attention to the change in VT. The results showed that tyrant flycatchers use temporal patterning to achieve a high level of variety in vocalization sessions despite a limited vocal repertoire.

Breeding adult males of many species of birds produce acoustic displays having the dual function of territorial advertisement and the attraction of potential mates (Collins 2004). In the oscine passerines, these vocalizations are typically produced in an extended session during which a repertoire of individual elements (“songs”) are repeated in sometimes complex patterns (Kroodsma 2004). The songs of oscines are learned, and the repertoire can be quite extensive (Kroodsma 2004). Smith (1991) emphasized that vocal displays of birds include both “primary signal units and formal rules governing sequential order in sustained performances.” The former correspond to individual “songs” or vocalization types and the latter to patterns of usage of these basic units. For example, the patterns of usage of individual vocalization types may include the proportion of songs of a certain type in a given vocalization session or the patterns of repetition of individual song types (Smith 1970, 1988). A number of functions have been proposed for switching among repertoire elements, including the minimization of habituation on the part of territorial rivals or potential mates (Krebs 1976, Falls and D'Agincourt 1982, Searcy et al. 1994). In certain oscine species, the rate of switching among song variants may convey information regarding the singer's level of motivation in a context of male-male aggression (Kramer and Lemon 1983, Smith 1991, Searcy et al. 2000). In addition, switching rates are known to increase in a mate attraction context in several oscine species (Vehrencamp 2000). Thus, in both intrasexual and intersexual contexts, switching among repertoire elements may relate to a general theme of enhancing receiver attentiveness.

Hartshorne (1956, 1973) was the first to draw attention to the avoidance of monotony as a potential functioning of temporal patterning in bird song. Hartshorne proposed a positive correlation between continuity of singing and versatility (the likelihood of a difference between successive songs), but the technical details involved in testing this idea proved controversial (Dobson and Lemon 1975, Kroodsma 1978, 1990, Weary and Lemon 1988). Nonetheless, in both agonistic and courtship contexts, it is a plausible hypothesis that natural selection will favor the evolution of features of vocal display sessions that minimize habituation on the part of potential receivers (Searcy et al. 1994, Kroodsma 2005). Moreover, minimization of signal monotony might conceivably occur by other mechanisms besides the continuity-versatility relationship proposed by Hartshorne.

In contrast to oscines, many species of suboscine passerines apparently do not learn their songs but rather employ a limited repertoire of innate vocal displays. For example, the song-repertoire of the eastern phoebe Sayornis phoebe consists of two innate song types, easily distinguished by the human ear and mnemonically designated fee-bee and fee-b-be-bee by Kroodsma (1985). Because of the relative simplicity of their song repertoires, suboscine passerines, particularly tyrant flycatchers (Passeriformes: Tyrannidae), have proved excellent models for study of the structure of vocalization sessions (Smith 1970, 1988, Kroodsma 1985, Lovell and Lein 2004). However, these studies have largely focused on the sequence or proportion of song types without examining the temporal patterning of song repertoire usage.

Here I present a comparative study of the temporal patterning (“rhythm”) of vocalization type usage in singing sessions of male tyrant flycatchers using computer analysis of recorded singing sessions from 20 species belonging to nine genera of the subfamily Tyranninae (Sibley and Monroe 1990). By analyzing in detail temporal features of recorded singing sessions, I test for factors that enhance the variety of acoustic patterns, in spite of the limited number of vocalization types used by these species.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. SVT vs. MVT species
  6. Interval between vocalizations
  7. Discussion
  8. Acknowledgements
  9. References

Recordings of vocalization sessions of 20 species of Tyranninae were obtained from the Borror Laboratory of Bioacoustics, Ohio State University (Table 1). Sounds were analyzed using Raven 1.2 program (Charif et al. 2004). In each recording, individual vocalization units (“songs”) were assigned qualitatively to vocalization types (VTs) by ear and by visual inspection of sonograms. The number of VTs identified per species ranged from 1 to 5, and 12 of the 20 species included two or more VTs in the recorded session. Examples of VTs include the fee-bee and fee-b-be-bee of the Eastern Phoebe (Kroodsma 1985). It is possible that the repertoire in nature of a species might include more VTs than were used in the sessions analyzed. On the basis of the sessions analyzed each species was categorized according to VT diversity as having a single VT (SVT) or multiple VTs (MVT). In two cases the recorded session was interrupted by a long period in which the bird appeared to have been disrupted (e.g., as a result of human activity), and later resumed singing (Table 1). In these cases the periods before and after the interruption were analyzed separately. In each of the two species, the length of the interruption was greater than 10 standard deviation units greater than the mean length of an interval between vocalizations recorded for that species. None of the other 18 species was the longest interval between vocalizations more than 4.4 SD units above the mean.

Table 1.  Summary of recorded vocalizations used in analyses.
SpeciesSession length (min)Vocalization types
  1. 1 Interrupted session.

Yellow-bellied elaenia Elaenia flavogaster0:50.642
Boreal pewee Contopus borealis1:51.162
Greater pewee Contopus pertinax1:50.48, 7:42.2112
Western wood-pewee Contopus sordidulus0:34.201
Eastern wood-pewee Contopus virens1:48.402
Alder flycatcher Empidonax alnorum1:58.421
Willow flycatcher Empidonax traillii4:21.062
Cordilleran flycatcher Empidonax occidentalis1:34.601
Buff-breasted flycatcher Empidonax fulvifrons6:14.02
Eastern phoebe Sayornis phoebe0:42.062
Say's phoebe Sayornis saya1:12.295
Vermillion flycatcher Pyrocephalus rubinus2:46.481
Dusky-capped flycatcher Myiarchus tuberculifer2:36.245
Ash-throated flycatcher Myiarchus cinerascens2:32.24, 2:31.2513
Great Crested flycatcher Myiarchus crinitus1:06.961
Large-billed flycatcher (Galápagos) Myiarchus magnirostris1:19.881
Cassin's kingbird Tyrannus vociferans1:01.771
Thick-billed kingbird Tyrannus crassirostris3:48.382
Boat-billed flycatcher Megarhynchus pitangua1:02.342
Sulfur-bellied flycatcher Myiodynastes luteiventris0:16.371

For each individual unit of vocalization (each occurrence of a given VT), the following variables were measured from sonagram and oscillogram data (Fig. 1): (1) length, the duration in sec of the vocalization, (2) % to peak, the percent of the vocalization length elapsed until the peak amplitude was reached, and (3) interval, the length of time (s) from the end of the previous vocalization and the start of the vocalization in question. The switching frequency among VTs was computed following Searcy and Yasukawa (1990) as the observed number of switches of VT divided by the number of opportunities to switch. A bout of using a given VT is defined as a series of vocalizations (one or more) that are of the same VT. If in a series of n vocalizations there are x bouts, then the switching frequency is (x−1)/(n−1) (Searcy and Yasukawa 1990).

image

Figure 1. Oscillogram (top) and sonagram of a vocalization of the eastern wood-pewee. The arrows indicate the start and end of the vocalization, while the vertical line indicates the time of peak amplitude. The variable length thus represents the time between the two arrows. The variable % to peak represents the time from the left arrow to the vertical line divided by length, expressed as a percent (14.0% in the present example).

Download figure to PowerPoint

In comparative analysis of biological data, it is important to control statistically for evolutionary relationships among taxa (Felsenstein 1985). Using Birdsley's (2002) phylogeny of genera of Tyrannidae based on morphology and behavior, the nine genera analyzed here that contained MVT species were placed in three major clades: (1) Myiarchus, (2) Tyrannus, Megarhynchus, and Myiodynastes (“kingbird assemblage”), and (3) Elaenia, Contopus, Empidonax, Pyrocephalus, and Sayornis (“elaeniid and empid clade”). I used a nested analysis of variance with a general linear models approach in order to control for evolutionary relationships among species. Variables describing units of vocalization were analyzed within species, by species within genus, and by genus within major clade.

Linear discriminant analyses based on length and % to peak were applied to VTs within MVT species. One VT of Say's Phoebe was excluded from the discriminant analysis because it was represented by a single example in the session analyzed. All statistical analyses were conducted using the Minitab statistical package, release 13 (http://www.minitab.com/).

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. SVT vs. MVT species
  6. Interval between vocalizations
  7. Discussion
  8. Acknowledgements
  9. References

Vocalization type

In 12 of 20 species of tyrant flycatchers from which vocalization sessions were analyzed, distinct vocalization types (VTs) could be recognized by ear and by visual examination of sonograms. Fig. 2A illustrates the three types identified in the case of the ash-throated flycatcher. When % to peak was plotted against length for 154 vocalization units of the ash-throated flycatcher Zonotrichia leucophrys, the three VTs formed separate clusters along the two axes (Fig. 2B). By contrast plots of % to peak against length for the eight species showing a single vocalization type (SVT species) did not reveal any distinct clustering patterns (not shown).

image

Figure 2. (A) Sonagram illustrating three vocalization types (VTs) of the ash-throated flycatcher (x-axis shows time in s; y-axis shows frequency in kHz). (B) Complete separation of the three VTs on the axes of length and % to peak.

Download figure to PowerPoint

In a nested analysis of variance applied to data from the 12 species with multiple VTs (MVT species), length showed significant effects of major clade (F2,1180=295.7, P<0.001), genus within major clade (F4,1180=4988.7, P<0.001), species within genus (F5,1180=2034.7, P<0.001), and VT within species (F19,1180=4592.1, P<0.001). Likewise % to peak showed significant effects of major clade (F2,1180=16.1, P<0.001), genus within major clade (F4,1180=28.2, P<0.001), species within genus (F5,1180=44.8;, P<0.001); and VT within species (F19,1180=82.5, P<0.001). The VT switching frequencies for the 12 MVT species ranged from 0.24 to 0.90 (mean=0.57±0.06; Table 2).

Table 2.  Summary of vocalization types (VT) in species with multiply VTs.
SpeciesVT (n)Mean (±SE) length (s)Mean (±SE) % topeakVT switching frequencyDiscriminant analysis (% correct)
  1. 1 Excluding VT 4.

Yellow-bellied elaenia1 (13)0.208±0.00458.5±4.30.86100%
 2 (16)0.277±0.00360.0±3.1  
      
Boreal pewee1 (7)1.080±0.01352.8±8.20.24100%
 2 (11)0.282±0.00473.5±1.1  
      
Greater pewee1 (112)1.452±0.00546.8±1.50.63100%
 2 (242)0.425±0.00267.4±0.2  
      
Eastern wood-pewee1 (10)1.043±0.01816.6±1.60.7387.5%
 2 (6)1.160±0.02828.3±1.2  
      
Willow flycatcher1 (13)0.256±0.04784.5±8.60.43100%
 2 (43)0.485±0.01443.1±4.0  
      
Buff-breasted flycatcher1 (111)0.207±0.00234.7±2.10.5993.6%
 2 (265)0.177±0.00138.0±1.2  
      
Eastern phoebe1 (10)0.442±0.00832.2±1.80.9095.2%
 2 (11)0.559±0.00430.9±5.6  
      
Say's phoebe1 (14)1.124±0.04128.6±6.80.6098.3%1
 2 (39)0.583±0.00336.5±0.7  
 3 (3)0.285±0.00870.2±7.8  
 4 (1)0.49640.0  
 5 (2)1.520±0.05212.0±0.9  
      
Dusky-capped flycatcher1 (12)0.588±0.02353.5±4.10.6278.0%
 2 (18)0.124±0.00672.6±1.8  
 3 (19)0.567±0.01840.2±3.8  
 4 (6)0.827±0.04146.5±2.9  
 5 (4)1.289±0.0775.8±1.0  
      
Ash-throated flycatcher1 (41)0.153±0.00687.7±4.30.61100%
 2 (24)0.166±0.00612.2±2.2  
 3 (89)0.203±0.01111.0±1.7  
      
Thick-billed kingbird1 (40)0.208±0.00410.1±0.40.26100%
 2 (15)0.277±0.00389.8±0.7  
      
Boat-billed flycatcher1 (5)0.688±0.01526.0±10.60.38100%
 2 (9)0.558±0.00911.5±1.0  

Linear discriminant analysis based on length and % to peak perfectly classified individual vocalization units by VT in seven of the 12 MVT species (Table 2). In the other five MVT species, discriminant analysis based on these two variables generally classified individual vocalization units by VT with high accuracy, the lowest classification rate being 78% in the case of the dusky-capped flycatcher scientific name in ital please (Table 2). Summing up the results for the 12 MVT species, 1,114 of 1,155 vocalization units were classified correctly (96.5%) by discriminant analyses applied to these two variables. Note that, although there were frequency differences among VTs, no variables based on frequency were included in these analyses. Thus, independent of frequency differences, vocalization units could be assigned to VT with a high degree of accuracy on the basis of variables describing aspects of temporal patterning alone.

SVT vs. MVT species

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. SVT vs. MVT species
  6. Interval between vocalizations
  7. Discussion
  8. Acknowledgements
  9. References

The absolute difference in length between each vocalization unit and that immediately preceding it in time was compared between three categories of vocalizations: those of MVT species when the VT was the same in the vocalization unit and that preceding it; those of MVT species when the VT differed between the vocalization unit and that preceding it; and those of SVT species. In a nested analysis of variance there were significant effects of major clade (F2,1304=47.6, P<0.001), of genus within major clade (F6,1304=69.9, P<0.001), and of vocalization category (F11,1304=189.0, P<0.001). A similar analysis of absolute difference in % to peak likewise showed significant effects of major clade (F2,1304=27.0, P<0.001), of genus within major clade (F6,1304=64.8, P<0.001), and of vocalization category (F11,1304=32.9, P<0.001).

Fig. 3 shows the means of the absolute differences in length and % to peak for the three categories. The mean absolute difference in length was lowest in the case of MVT species when the vocalization was of the same VT as that preceding it (0.021 s±0.002 SE) and highest in the case of MVT species when the vocalization differed in VT from that preceding it (0.405 s±0.016; Fig. 3A). The mean absolute difference in length for SVT species (0.207 s±0.052) was intermediate between the two values for MVT species (Fig. 3A). Likewise, the mean absolute difference in % to peak was lowest in the case of MVT species when the vocalization was of the same VT as that preceding it (4.4%±0.4) and highest in the case of MVT species when the vocalization differed in VT from that preceding it (18.7%±0.7, Fig. 3B). The mean absolute difference in% to peak for SVT species (21.8%±5.2) was slightly higher than that for MVT species when the VT differed (Fig. 3B). These results for MVT species were consistent with the observation VTs could generally be separated on the basis of length and % to peak (Table 2), leading to the expectation that there should be a greater absolute difference in these variables when the VT changed between sequential vocalization units than when it did not. However, it was remarkable that in the case of SVT species, even though vocalization units did not fall into discrete categories, there was substantial absolute difference with respect to length and % to peak between a typical vocalization unit and that preceding it.

image

Figure 3. Means of absolute differences (±SE) of: (A) length, and (B) % to peak between successive vocalization units, classified as to whether the two units were the same VT in a species with multiple vocalization types (MVT species), different VTS in an MVT species, or vocalization units of a species with a single vocalization type (SVT species).

Download figure to PowerPoint

Interval between vocalizations

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. SVT vs. MVT species
  6. Interval between vocalizations
  7. Discussion
  8. Acknowledgements
  9. References

In MVT species, the intervals between vocalization units were categorized as to whether the VT before the interval was the same as or different from that after the interval (VT change). A nested analysis of variance showed a significant effect on interval of major clade (F2,1173=36.8, P<0.001), genus within major clade (F4,1173=52.5, P<0.001), species within genus (F5,1173=216.8, P<0.001), and VT change (F12,1173=6.02, P<0.001). Fig. 4 shows a plot of mean interval in cases where the VT was the same vs. that in cases where the VT was different for the 12 MVT species. In 11 of the 12 MVT species, mean interval was longer when the VT stayed the same than when it changed. Perhaps the most striking case was that of the greater pewee, where the mean interval was over twice as long when the VT remained the same (1.273 s±0.083, n=131) than when the VT changed (0.636 s±0.015, n=221). The only exception to the overall trend was the boat-billed flycatcher scientific name in ital please. In the latter species, the number of observations was small, suggesting that stochastic error may have played a role in the exceptional result. Also, in the boat-billed flycatcher, mean interval when the VT was the same (4.030 s±0.422, n=8) was very similar to that when the VT changed (4.330 sec±1.210, n=5).

image

Figure 4. Plot of the mean interval length (s) between successive vocalization units when the VT was the same vs. the interval length when the VT was different.

Download figure to PowerPoint

Considering only intervals in which the VT changed between vocalization units, a nested analysis of variance applied to interval showed significant effects of major clade (F2,688=68.9, P<0.001), of genus within major clade (F4, 688=116.2, P<0.001), of species within genus (F5,688=364.3, P<0.001), and of the sequence of VTs within species (F37,688=17.3, P<0.001). The effect of the sequence of VTs is illustrated in Fig. 5 with two examples of species having just two VTs, greater peewee scientific name in ital please and buff-breasted flycatcher scientific name in ital please.

image

Figure 5. Examples of mean interval length (s±SE) between successive vocalization of different VT for four species with two VTs, categorized with regard to the sequence of VTs: (A) greater pewee, and (B) buff-breasted flycatcher. Numbers on top of the bars indicate the number of cases.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. SVT vs. MVT species
  6. Interval between vocalizations
  7. Discussion
  8. Acknowledgements
  9. References

Recorded vocalization (“song”) sessions of males from 20 species of tyrant flycatchers revealed two or more qualitatively distinct vocalization types (VTs) in 12 species (MVT species). In most cases, vocalization units of MVT species could be assigned to VT with a high degree of accuracy based only on two variables: the length of the vocalization unit and the percentage of time elapsed until peak amplitude was reached. VTs typically also showed differences in the pattern of frequencies. Thus it is somewhat surprising that they could be distinguished by temporal patterning alone without reference to frequency.

The results showed that tyrant flycatchers use temporal patterning to achieve a high level of variety in vocalization sessions despite a limited vocal repertoire. Species with just a single VT in the session analyzed (SVT species) showed a surprisingly high degree of absolute difference between adjacent vocalization units with respect to both the length of the vocalization and percentage of time elapsed until peak amplitude was reached. These differences were much higher than those seen in MVT species when a vocalization unit was followed by one of the same VT (Fig. 2). Thus, in spite of the absence of discrete recognizable VTs in these species, the vocalizations they produced showed substantial variation with regard to temporal patterning.

Detailed analysis of temporal patterns of vocalization in MVT species revealed additional features besides the use of distinct vocalization types that might serve to heighten attention or minimize habituation by conspecifics. The interval between vocalization units when the VT changed was on average shorter than that when the VT remained the same. This change in the rhythm of vocalization accompanying a change in VT provided a kind of punctuation, potentially drawing the receiver's attention to the change in VT. In addition, when the VT changed, the mean interval length between vocalization units differed depending on which VT preceded the interval and which VT followed it. Thus each distinctive pattern of VT change was accompanied by a distinctive rhythmic pattern, reinforcing the change in VT.

Several authors have discussed the importance in animal communication of “multicomponent” signals; i.e., signals that involve more than one aspect, including those that employ more than one sensory modality (Rowe 1999, Partan and Marler 2005). Partan and Marler (2005) note that, when components are produced sequentially, an earlier component can serve to attract attention and thus increase responsiveness to a subsequent component. The present results suggest that, in the acoustic displays of birds, the temporal patterning (i.e., “rhythm”) of the vocalization session can serve as a distinct component of the signal, drawing attention to the vocalization itself. Thus these results thus suggest a mechanism of maintaining attention in avian vocalization sessions different from the relationship between continuity and versatility proposed by Hartshorne (1956, 1973).

Because the present analyses in several cases involved fairly short sessions (Table 1), the conclusions need to be confirmed with longer sessions. However, it is worth remarking that, given the simplicity of suboscine vocal repertoires, additional recording time is unlikely to change the picture substantially. In addition, the hypothesis that the rhythmic aspects of song sessions serve to enhance the attention of conspecifics remains to be tested experimentally; for example, by playback experiments in which the intervals between vocalization units are manipulated through editing of recorded vocalization sessions. By examining responses by birds to vocalization sessions from which features providing rhythmic variety have been experimentally removed or altered, such studies might be used both to test the hypothesis that rhythmic variety maintains attentiveness of conspecifics and to identify further specific features of vocalizations sessions that maintain attention. Moreover, since the function of switching between vocalization types may differ between intrasexual and intersexual contexts (Vehrencamp 2000, Collins 2004), it will be of interest to examine the rhythmic aspects of vocalization sessions in different behavioral settings.

The use of archival recordings in the present analyses may raise questions in the light of Derryberry's (2007) report of changes in songs of white-crowned sparrows scientific name in ital please over time, resulting in reduced responsiveness of both males and females to 24-year-old recordings. It is worth noting that, in spite of Derryberry's (2007) use of the word “evolution” to describe the observed changes, the time involved appears too short for actual genetic evolution (allelic substitution) to have occurred. Thus, what Derryberry (2007) has documented in the white-crowned sparrow is probably a kind of cultural drift due to the error-prone nature of the song-learning process in oscine passerines. In tyrant flycatchers, whose songs are presumably unlearned, this type of change would not be expected to occur. Moreover, even in species with learned songs, general principles of the structure of vocalizations can be learned from the analysis of archival recordings, since the recorded songs presumably possessed characteristics that were adaptive in the social and ecological context in which they originally occurred, even if certain nonessential features are subject to change over time.

The present analyses show that, because of their relatively simple vocal repertoires, sub-oscine passerines such as Tyrannidae provide attractive models for the study of avian vocal behavior; a similar point is illustrated by the research of ten Cate et al. (2002) with the simple vocal repertoire of members of the non-passerine family Columbidae. The results are consistent with the hypothesis that, despite their simple repertoire, tyrant flycatchers are able to use temporal patterning to avoid monotonous patterns that might be expected to induce habituation on the part of receivers. This, in turn, suggests that there may be a variety of mechanisms available to birds and other animals to minimize the monotony of acoustic displays by manipulating temporal patterning.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. SVT vs. MVT species
  6. Interval between vocalizations
  7. Discussion
  8. Acknowledgements
  9. References

I am grateful to Dr. Jill Soha, Borror Laboratory of Bioacoustics, for her assistance with obtaining recordings and to NIH grant GM43940, which provided partial support.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. SVT vs. MVT species
  6. Interval between vocalizations
  7. Discussion
  8. Acknowledgements
  9. References
  • Birdsley, J. S.. 2002. Phylogeny of the tyrant flycatchers (Tyrannidae) based on morphology and behavior. Auk 119: 715734.
  • Charif, R. A., Clark, C. W. and Fristrup, K. M.. 2004. Raven 1.2 user's manual. Cornell laboratory of ornithology, Ithaca.
  • Collins, S.. 2004. Vocal fighting and flirting; the functions of birdsong. – In: Marler, P. and Slabbekoorn, H. (eds), Nature's music: the science of birdsong. Elsevier Academic Press, San Diego, pp. 3979.
  • Derryberry, E.. 2007. Evolution of bird song affects signal efficacy: an experimental test using historical and current signals. Evolution 61: 19381945.
  • Dobson, C. W. and Lemon, R. E.. 1975. Re-examination of the monotony-threshold hypothesis in bird song. Nature 257: 126128.
  • Falls, J. B. and D'Agincourt, L. G.. 1982. Why do meadowlarks switch song types?. Can. J. Zool. 60: 34003408.
  • Felsenstein, J.. 1985. Phylogenies and the comparative method. Am. Nat. 125: 115.
  • Hartshorne, C.. 1956. The monotony-threshold in singing birds. Auk 73: 176192.
  • Hartshorne, C.. 1973. Born to sing. Indiana University Press, Bloomington.
  • Kramer, H. G. and Lemon, R. E.. 1983. Dynamics of territorial singining between neighboring song sparrows (Melospiza melodia). Behaviour 85: 198223.
  • Krebs, J. R.. 1976. Habituation and song repertoires in the great tit. Behav. Ecol. Sociobiol. 1: 215227.
  • Kroodsma, D.. 1978. Continuity and versatility in bird song: support for the monotony-threshold principle. Nature 274: 681683.
  • Kroodsma, D.. 1985. Development and use of two song forms by the eastern phoebe. Wilson Bull. 97: 2129.
  • Kroodsma, D.. 1990. Patterns in songbird singing behaviour: Hartshorne vindicated. Anim. Behav. 39: 994996.
  • Kroodsma, D.. 2004. The diversity and plasticity of birdsong. – In: Marler, P. and Slabbekoorn, H. (eds), Nature's music: the science of birdsong. Elsevier Academic Press, San Diego, pp. 108131.
  • Kroodsma, D.. 2005. The singing life of birds. Houghton Mifflin, Boston.
  • Lovell, S. F. and Lein, M. R.. 2004. Neighbor-stranger discrimination by song in a suboscine bird, the alder flycatcher, Empidonax alnorum. Behav. Ecol. 15: 799804.
  • Partan, S. R. and Marler, P.. 2005. Issues in the classification of multimodal communication signals. Am. Nat. 166: 231245.
  • Rowe, C.. 1999. Receiver psychology and the evolution of multicomponent signals. Anim. Behav. 58: 921931.
  • Searcy, W. A. and Yasukawa, K.. 1990. Use of the song repertoire in intersexual and intrasexual contexts by male red-winged blackbirds. Behav. Ecol. Sociobiol. 127: 123128.
  • Searcy, W. A., Cofman, S. and Raikow, D. F.. 1994. Habituation, recovery and the similarity of song types with repertoires in red-winged blackbirds (Agelaius phoeniceus). Ethology 98: 3849.
  • Searcy, W. A., Nowicki, S. and Hogan, C.. 2000. Song type variants and aggressive context. Behav. Ecol. Sociobiol. 48: 358363.
  • Sibley, C. G. and Monroe, B. L., Jr.. 1990. Distribution and taxonomy of birds of the world. Yale University Press, New Haven.
  • Smith, W. J.. 1970. Song-like displays in Sayornis species. Behaviour 37: 6484.
  • Smith, W. J.. 1988. Patterned daytime singing of the eastern wood-pewee, Contopus virens. Anim. Behav. 36: 11111123.
  • Smith, W. J.. 1991. Singing is based on two markedly different kinds of signaling. J. J. Theor. Biol. 152: 241253.
  • Ten Cate, C., Slabbekoorn, H. and Ballintin, M. R.. 2002. Birdsong and male-male competition: causes and consequences of vocal variability in the collared dove (Streptopelia decaocto). Adv. Study Behav. 31: 3175.
  • Vehrencamp, S.. 2000. Handicap, index, and conventional elements of bird song. – In: Espmark, Y., Amundsen, T. and Rosenqvist, G. (eds), Animal signals: signals and signal design in animal communication. Tapir Academic Press, Trondheim, Norway, pp. 277300.
  • Weary, D. M. and Lemon, R. E.. 1988. Evidence against the continuity-versatlity relationship in bird song. Anim. Behav. 36: 13791383.