Responses of male Red-eyed Vireos (Vireo olivaceus) to song playback varying in rate and cadence. H. Lynn Bradman. University of Nebraska

Transcription

1 Responses of male Red-eyed Vireos (Vireo olivaceus) to song playback varying in rate and cadence H. Lynn Bradman University of Nebraska Lincoln, Nebraska

2 Abstract 2 We found that territory holding male Red-eyed vireos (Vireo olivaceus) sang their songs at a higher rate and with less variation in intersong intervals than non-territory holding males. We conducted two playback experiments to examine the effects of song rate and song cadence on behavioral responses associated with territorial defense in this species. Song rate and cadence were varied simultaneously on tape recorded exemplars by manipulating the intersong intervals (ISI) only. Experiment 1 manipulated rate ( fast and slow) and two cadence styles, even and uneven. Male vireos remained within the playback arena significantly longer during playback of even-cadenced song than during playback of uneven-cadenced song. They also produced faster song rates immediately after playback of even-cadenced song than after playback of unevencadenced song. No effects of song rate manipulations were detected. Experiment 2, an extension and replication of experiment 1, included an additional ultraslow song rate level for both cadence styles. Males remained within the playback arena longer, and approached the playback speakers more closely during even-cadenced song playback than during unevencadenced playback. Males made more flights through the playback arena, made closer approaches to the playback speakers, and arrived within the playback arena more quickly during fast and slow rate conditions than ultraslow rate conditions. We suggest that the song rate and the time between songs, the intersong interval, may be informative song features in this species, but more specifically that male Red-eyed vireos may be able to reliably assess the quality of rival males by evaluating the cadence of those rivals songs.

3 Responses of male Red-eyed Vireos (Vireo olivaceus) 3 to song playback varying in rate and cadence Song is usually thought to function both intersexually, to facilitate mate attraction, and intrasexually in territory acquisition and defense in passerine birds (Searcy and Andersson, 1986). Males are thought to use song to deter other males from trespassing on established territories and in several species, which sing small to moderate sized repertoires, males have been shown to respond more aggressively to the songs of intruding strangers, than to those of neighbors (Falls, 1982). There are species, in which males sing large repertoires, that a deficiency in the ability to discriminate familiar versus unfamiliar individuals by song has been demonstrated (Falls and D Agincourt, 1981). Godard (1993) has suggested that large repertoires may present receivers with two problems. There will be more song types to learn and particular song types will be repeated less frequently, thus discrimination of familiar versus unfamiliar individuals might be more difficult. Thus, costs associated with singing in these species may be elevated in several ways. If producing a large repertoire does not facilitate territory defense or mate attraction in these species, possibly other characteristics of song that do function in this way may have become conventional. Temporal features and patterns within vocalizations may serve in this capacity. Song rate and song cadence are two such characteristics. Song rate refers to the number of songs emitted by a signaler per unit time. Some authors have suggested that a male s song rate may serve as an indicator of territory quality for females assessing potential nest site availability, safety, and protection from predators (Hoi-Lietner, Nechtelberger, and Hoi, 1995). Consistent with this suggestion, Radesater, Jakobsson, Andbjer,

4 and Nystrom (1987) reported a positive correlation between early pairing dates and faster song 4 rates in Willow warblers (Phylloscopus trochilus). The temporal gaps between whole songs are called intersong intervals. Song cadence refers to the regularity in duration of the intersong intervals. A signal may be emitted at a given rate with either more or less variable cadence. The variability of the intersong intervals may be of importance in signals directed toward conspecifics of both sexes. Beletsky (1989) reported that intersong intervals for several species of songbirds were typically from 7-13 seconds in duration and suggested that uniformity in intersong intervals may benefit singers by reducing listener uncertainty about signaler density. Identification and location of singers could be simplified by temporal uniformity between songs, both of which could further facilitate mate attraction and territory defense. Objectively, territory holding males should sing with less variation in intersong intervals, than non-territory holding males. The Red-eyed vireo (Vireo olivaceus), a neotropical migrant songbird, breeds in most of North America and is noted for its persistent song rate and monotonous song cadence. It sings a repertoire of around 40 song types on average, and there is variation in repertoire size among males (Borror, 1981). These males sing persistently from a song post within territories which they actively defend (Lawrence, 1953). Red-eyed vireo songs usually contain from one to three elements. Songs with four or five elements are produced but are rare (Borror, 1981). Song elements are emitted in a continuous pattern of sound averaging about 0.15 seconds in duration per element and 0.35 seconds per song (Lemon, 1971) with mean intersong intervals of about 1.3 seconds (personal observation). Due to its short song length and short intersong interval, the Red-eyed vireo seemed a perfect species

5 in which to study the possible communicative value of song rate and cadence. 5 Observational studies We conducted initial observations of the singing behavior of the Red-eyed vireo during the summer of During the first three weeks of the study, male birds were located, banded, and songs were recorded for later statistical and acoustical analyses. We noticed, during observations in the field, that territory holding males seemed to sing more consistently than non-territory holders. More specifically, the song rates produced by the territory holders seemed to be slightly faster and more persistent. We collected song rate data from several males for the purpose of comparing the song rates produced by territory holding versus non-territory holding males. The territorial status of these birds was resolved after the observational segment of the study was completed at which time we could determine which males became permanent territory holders and which males left their territories or were displaced by other males. Methods Subjects Subjects were 9 Red-eyed Vireos including 5 territory holders and 4 non-territory holders. All birds were in residence at the study site during May, 1994 and had been observed and recorded previously. Males were identified by unique band color combinations (for birds banded early) or by association with a particular location within the study area. Design and Procedures Song rates produced by the subjects were obtained during two 10 minute intervals at approximately the same time of day on two occasions. Researchers tallied the number of songs

6 6 produced by each subject during the intervals. Song rates were calculated in songs per minute. A between groups analysis of variance was used to analyze the data. Results and Discussion Results of the analysis revealed that the territory holders sang significantly more songs per minute ( M =35.6 std 8.5 ) than non-territory holders ( M = 11.7 std 7.9 ) (F(1,7)=18.58 p=.004) which confirmed our field observation concerning differences in song rate between the groups. However, we realized that simple calculation of songs per minute could be misleading without consideration of the manner in which the songs were delivered. A male may sing at a highly variable cadence and still produce as many songs per minute as a male singing with a regular cadence. We decided to conduct an additional analysis of the mean song duration, mean intersong interval duration, the average standard deviation of the song length and of the intersong interval length. These analyses were conducted after the observational segment of the study was completed at which time we had determined which males were permanent territory holders and which males had left their territories or were displaced by other males. We increased the sample size by using recorded songs of several more individuals for computer aided analyses of these temporal features. Subjects Method Subjects were eighteen Red-eyed vireos. Nine were territory holding males that maintained territories throughout the breeding season, were paired, and nested. Nine were nonterritory holders including two individuals that departed the study area early in the season, five

7 7 birds that had obtained territories and were displaced later, and two birds that initially established territories but disappeared before breeding. Although some of the nine non-territorial males were known to associate with females, none were known to be paired long enough for nesting behavior to commence. Materials Vireos were recorded using a Sennheiser ME88 microphone and a portable Sony TCD- D7 DAT digital tape recorder. Song samples were digitized and sonographic representations produced using a Macintosh Quadra 800 computer equipped with Canary 1.2 sound analysis software (Charif, Mitchell and Clark, 1993). Procedure and Design The recordings of all males in both groups were obtained during the last week of May, Each recording was at least 3 minutes in length. Song samples used for analysis were unedited random cuts from the recordings. The song samples contained 30 songs each. Precise measurements of song durations and intersomg interval durations were obtained using Canary 1.2 computer software (Charif, Mitchell and Clark, 1993). A between-groups analysis of variance was used to analyse the data. Results and Discussion The two groups did not differ in the mean song duration the mean intersong interval duration, or the average standard deviation of the song duration (all p s >.05). However, the territory holders sang with a less variable song cadence. The standard deviation of the intersong intervals for the territory holders ( M =.2682 S ) was significantly smaller than for the nonterritory holders (M =.7047 S ) (F(1,16) = p<.05).

8 8 These comparisons of the song rate and song cadence of territory holders versus nonterritory holders suggested that less variable song cadence should be indicative of territorial males, or males vigorous enough to pose a threat. We thought that male birds with the ability to sing with a less variable song rate or cadence were possibly signaling or advertising their overall vigor to females and other males. If male Red-eyed vireos evaluate the song rate and cadence of rivals songs as a means of assessing the quality of those rivals, territorial males, then, should respond more strongly to intruders singing in this manner than to others. We conducted two playback experiments to test this prediction. Experiment 1 Method Subjects. Subjects were 15 male Red-eyed Vireos, thirteen of which had been captured and banded during the 1994 field season and 2 during the 1995 season. All were known to have territories within the study area. The males used in this study were identified by unique band color combinations and were observed in their respective territories for a period of approximately three weeks before experimental manipulations began. Experiment 1 was conducted during 1995 at Schramm State Park and Recreation Area in Sarpy County, Nebraska. Recording Materials During the first three weeks of the study, vireos were recorded using a portable Sony TCD-D7 DAT digital tape recorder for later detailed sound analysis. The same recorder and a Realistic AMX 8 amplified speaker were used to play back experimentally manipulated song samples during the playback segments of the experimental sessions. In order to obtain song rates

9 for each subject, recordings of subjects songs during sessions were made using a Sony TCM EV recorder. Stimuli construction The playback stimuli were produced from 1994 recordings of two male Red-eyed vireos from two widely separated areas of the study site. A Macintosh Quadra 800 computer equipped with Canary 1.2 sound analysis software (Charif, Mitchell and Clark, 1993) was used to digitize songs for preparation of the playback tapes. Both song rate and cadence were varied by manipulating the ISIs only. We played approximately 60 S of continuous vireo song into the computer using the RECORD function of Canary 1.2 and produced the waveform. ISIs were altered by deleting or adding sections of the naturally occurring intersong intervals from the original digitized song sequence to create ISIs of the desired length. The altered song samples were then re-recorded from the computer audio output directly onto DAT tape. The 60 S segments were repeated so that each playback tape was five minutes in duration. Four conditions were created using this method; a fast rate/ even cadence had ISIs that were 0.5 seconds (M.50 std 0), a slow rate/even cadence had ISIs that were 1.5 seconds (M 1.5 std 0), a fast rate/uneven-cadence that had ISIs that ranged from 0.1 through 1.0 seconds (M.5 std.231) and a slow rate/uneven-cadence that had ISIs that ranged from 1.1 through 2.0 seconds (M 1.5 std.298). Procedure After approximately three weeks of observation and recording of the subjects for identification, the location of each subject s song area as described by Lawrence (1953) was

10 10 determined. This area could either be a solitary tree or group of trees that each male frequently used for sustained singing. We first identified an appropriate playback site (fairly open with a few trees and low vegetation) approximately 70 feet from each subject s song area. The arenas were just inside the territory boundaries, thus standardizing the playback environments as much as possible assuring that all males were required to travel approximately the same distance from their song post to the playback arena when the song presentations began. A circular playback arena with a 40 foot diameter was outlined by marking approximately every 90 degrees with flagging tape. The playback speakers were placed in the center of the arena during each playback session. All playback sessions were conducted approximately every seventh day for all subjects. Each playback session consisted of three consecutive five minute segments; pre-playback, playback, and post-playback. Measurements were taken for several dependent variables during each segment of the sessions as follows. Pre-playback Although no recorded songs were presented during the 5 minute pre-playback segments, several behavioral observations were made and data were collected with the intent of establishing baseline levels for comparison with playback and post-playback measures for some variables. We noted whether each subject was singing within his territory during the pre-playback interval, and the song rate produced by each subject during the five minute pre-playback segment was obtained. We noted whether males entered the playback arena during this segment. In addition, we measured the latency to arrive within the playback arena. Because Red-eyed vireos often inspect intruders by making sweeping flights toward a distracting noise or disturbance in the environment, we counted the number of flights, oriented toward the playback speaker and across

11 the arena, during the pre-playback segment. The distance of closest approach to the playback 11 speaker, and the total time each bird spent within the playback arena during the pre-playback segment was measured. Playback The 5 minute pre-playback segment was immediately followed by the five minute playback segment in which the song playback tapes were presented. Because subjects were expected to be singing simultaneously with the presentation of the playback tapes, all playback segments were tape recorded in order to obtain song rates produced by each male through visual inspection of sonographs. We also noted whether or not each subject entered the arena and each subject s latency to arrive in the playback arena was measured. We counted the number of flights made across the arena that were oriented toward the speaker for each male. During the playback segment, the distance of closest approach to the playback speaker made by each male was included as an additional response measure, and ultimately the amount of time each male spent inside the arena during the playback segment was measured. Post-playback All playback trials were followed by a 5 minute interval in which the post-playback measurements were obtained. We noted whether or not each subject was singing, we obtained the song rate produced by each subject, the number of flights each bird made across the arena that was oriented toward the speaker, the closest distance of approach to the playback speaker, and the time each bird spent within the playback arena. Design and Data Analyses Each playback session, pre-playback, playback and post-playback combined, was

12 observed by at least two researchers and measurement tasks remained constant throughout all 12 sessions of the experiment. Data analyses were conducted using a ( 2 rate x 2 cadence ) withinsubjects factorial design (Kepple, Saufley and Tokunaga, 1992). Chi square analyses were used where appropriate. Results Pre-playback: Measures of all dependent variables during the pre-playback segments are summarized in Table 1a. Song production and song rate: Subjects were singing within their respective territories during pre-playback equally often across all conditions of the experiment (χ 2 =.141, p>.05). There were no significant differences among the subjects pre-playback song rates during any pre-playback segment. Overall, subjects mean song rates were approximately equivalent for both fast (M 38.67, SD 8.20) and slow ( M 42.40, SD 10.14) rate levels and for even (M 42.30, SD 9.61) and uneven (M 38.77, SD 8.75) cadence types during pre-playback. Arena entry and latency: None of the subjects were observed to enter the playback arena during the pre-playback segment. Because all males remained outside the playback arena during the pre-playback segment, the mean latency to arrive for all subjects was 300 seconds for all conditions. Flights, approach distance, and time within in the arena: None of the subjects made any flights across the arena during the pre-playback segment, approached the playback speaker or spent time within the playback arena during the pre-playback segment. Overall, no significant differences were detected for any of the pre-playback measures for

13 13 any of the experimental sessions suggesting that all subjects were in an approximately equivalent behavioral state when each playback presentation began. Playback: Song production and song rate: All subjects produced song during the playback segments across all conditions of the experiment. However, many of our recordings of subjects songs during the playback sessions were obscured by the playback presentation. Upon examination of the recorded playback segments, we were unable to discriminate many of the subjects songs from the pre-recorded playback songs either acoustically or visually from sonographic representations. Therefore, accurate analyses of the song rates produced by males during the playback segments could not be conducted. Arena entry and latency to arrive: All subjects entered the playback arena during the playback segments during all of the experimental conditions. There was no main effect of Song Rate on the latency to arrive within the arena during the playback segment. Males showed no significant differences in their latency to arrive within the arena during the fast rate playback conditions (M S, SD 45.70), or the slow rate playback conditions ( M S, SD 53.82). There was no main effect of Song Cadence on the latency to arrive within the arena. Males arrived within the arena in about the same amount of time during even-cadenced playbacks ( M 59.50, SD 41.19) as during uneven-cadenced playbacks (M 57.33, SD 58.33). Nevertheless, there was a highly significant decrease in latency time between the preplayback and playback segment (F (1,28) = , Mse , p<.000) During the preplayback segment, no birds entered the arena, and all birds showed a mean latency time (300 S,

14 14 std 0) for all conditions. However, during the playback segment, the birds overall mean latency time was (48.42 S std 49.76). Flights: There was no main effect of Song Rate on the number of flights males made through the arena toward the playback speaker. Males made about the same number of flights through the arena during the fast rate playbacks ( M 5.83, SD 4.36) as during the slow rate playbacks ( M 6.02, SD 4.36). Furthermore, there was no main effect of Song Cadence on the number of flights males made through the arena. Males made about the same number of flights through the arena for even -cadenced ( M 6.53, SD 4.32) and uneven-cadenced ( M 5.33, SD 3.81) song playbacks. Approach distance: There were no effects of Song Rate on the closest approach distance to the playback speaker made by males during the playback segment. Males mean approach distance was approximately the same during fast rate playbacks ( M SD 9.81) as during the slow rate playbacks ( M 21.90, SD 13.47). Moreover, there was no main effect of Song Cadence on the closest approach distance to the playback speakers during the playback segment. The males mean approach distances from the speaker were about the same for the even-cadenced ( M = SD ) as the unevencadenced playbacks ( M 20.57, SD ). Time in the arena: There was no main effect of Song Rate on the time male birds spent within the playback arena during the playback segment. Overall, males spent an approximately equivalent amount of time within the arena during the fast rate playbacks ( M , SD 47.52), as during the slow rate playbacks ( M , SD 57.22). However, there was a highly significant main effect of Song Cadence on the time each

15 bird spent within the playback arena during the playback segments (F(1,14) = 60.28, MSe , p<.001). Male Red-eyed vireos remained within the playback arena longer when exposed to playback of even-cadenced songs (M , SD 48.19) than when exposed to uneven-cadenced songs (M , SD 56.55). Examination of the simple main effects of Song Cadence showed that this effect was descriptive for both fast and slow rate levels. Post-playback: Song production and song rate: Males were observed singing within the playback arena during the post-playback interval equally often across all conditions of the experiment (χ 2 =.834, p >.05). Song Rate manipulations had no effect on the song rates produced by males during the post-playback segment. Males sang at about the same rate after presentations of the fast rate playbacks ( M = 49.3, std ) as after the slow rate playbacks ( M = 51.7, std ). However, there was a significant main effect of Song Cadence on the song rates produced by males during the post-playback segments (F(1,14) = 80.40, MSe , p<.005). Males sang significantly more songs per minute after presentations of the even-cadenced playbacks (M =56.83, std ) than after uneven-cadenced playbacks (M= 44.17, std ). Comparisons of the simple main effects of song cadence for both rate levels showed that the effect was descriptive for both fast and slow rate levels Flights: Song Rate manipulations had no effect on the number of flights males made through the arena toward the speaker during the post-playback segments. Males made approximately the same number of flights during the fast rate playbacks ( M = 1.67, std 2.75) as during the slow rate playbacks ( M = 2.03, std 2.55). In addition, there was no effect of Song Cadence on the number of flights males made

16 16 through the arena. Males made similar numbers of flights through the arena during the evencadenced playbacks ( M = 2.03, std 2.55) as during the uneven-cadenced playbacks ( M = 1.33, std 1.616). Time in the arena: The was no effect of Song Rate on the time the subjects spent within the playback arena during the post-playback segment. Subjects spent approximately the same amount of time within the playback arena during the fast ( M = , std 53.59) and slow ( M = , std 59.87) rate conditions. Further, there was no effect of Song Cadence on the time spent within the arena after playback. Males stayed within the arena after playback roughly as long during even-cadenced playbacks (M = , std 54.2 ) as during uneven-cadenced playbacks (M = , std 59.19). Discussion Results from our observational studies revealed that territory holding male Red-eyed vireos sing with a less variable song cadence. The current analyses showed that territorial males responses to the even-cadenced exemplars were significantly stronger for some measures. Males spent more time within the playback arenas during even-cadenced playback sessions and responded to playback with increased song rates after playback of even-cadenced songs. Contrary to our predictions, subjects responses to our playback tapes differing in song rate did not depend on the song cadence. Furthermore, there were no significant differences in male responses to our playback tapes differing song rate. This result is particularly surprising considering our initial analyses which revealed that the territorial males produced a faster song rate, although supportive of our prediction that males may vary their song delivery temporally, without necessarily increasing or decreasing their rate per minute and that temporal variance,

17 cadence, may convey information and supports our hypothesis that less variable song cadence 17 may be a reliable indicator of quality in male Red-eyed vireos. However, the result may have been due to the similarity of the two song rate levels. Any main effects of rate or interactions between song rate and song cadence as they related to our dependent variables may not have been detected by this design. Song elements in Red-eyed Vireo songs are separated by an average time of.35 seconds (see figure x). Our fast song rate tapes contained songs separated by.50 S on average for both the even and the uneven cadences. Because we thought that a song rate faster than 120 per minute would be highly inconsistent with naturally occurring song rates produced by this species, we decided to replicate and extend the study by adding another, slower, level of song rate. Experiment 2 Methods Subjects Subjects were 17 male Red-eyed Vireos, fourteen of which had been captured and banded during the 1994 field season, 2 during the 1995 season, and 1 during the 1996 season. Fifteen of the birds had been subjects in Experiment 1, one year earlier. The males again were identified by unique band color combinations and were observed in their respective territories for a period of approximately 3 weeks before experimental manipulations began. Experiment 2 was conducted during 1996 at Schramm State Park and Recreation Area in Sarpy County, Nebraska and Platte River State Park in Lancaster County, Nebraska. Recording Materials The materials for conducting experiment 2 were identical to Experiment 1. In order to

18 18 obtain song rates for each subject, recordings of subjects songs during sessions were made using a Sony TCM-5000 EV recorder. Some additional song samples were collected using a SONY MZ-R3 portable mini-disk recorder. Stimuli construction The method of construction of the playback stimuli for Experiment 2 was identical to that employed in Experiment 1. Playback exemplars were digitally prepared from the same recordings used previously with Canary 1.2 software and the Macintosh 800 computer. However, we included an ultraslow rate level in Experiment 2. The intersong intervals for the additional ultraslow/even cadenced condition were 2.5 seconds in duration (M 2.5, std 0), and for the ultraslow/uneven cadenced condition the intersong intervals ranged from 2.1 to 3.0 seconds (M 2.5, std.474). Procedure and Design The procedures for Experiment 2 were identical to those used in Experiment 1. The playback sessions consisted of three successive five minute segments; pre-playback, playback, and post-playback. All playback sessions were conducted approximately every seventh day for all subjects. The six experimental conditions (3 rates x 2 cadence) were randomly presented to each subject. Data analyses were conducted using a within-subjects factorial design (Kepple, Saufley and Tokunaga, 1992). The Bonferroni test was used to correct for alpha inflation in analyses of the simple and main effects comparisons. Chi square analyses were used where appropriate. Pre-playback Results Song production and song rate: Subjects were singing within their respective territories

19 during pre-playback equally often across all conditions of the experiment (χ 2 =.141, p>.05). 19 There were no significant differences among the subjects pre-playback song rates during any pre-playback segment. Overall, subjects mean song rates were approximately equivalent for the fast (M = std 10.98), slow ( M = std 11.72) and ultraslow ( M = std 9.18 ) rate levels and for even (M= std ) and uneven (M= std ) cadence types during pre-playback. Arena entry and latency: None of the subjects were observed to enter the playback arena during the pre-playback segment. Because all males remained outside the playback arena during the pre-playback segment, the mean latency to arrive for all subjects was 300 S for all conditions. Flights, approach distance, and time within in arena: None of the subjects made any flights across the arena during the pre-playback segment, nor spent time within the playback arena during the pre-playback segment. Therefore, no significant differences were detected for any of the pre-playback measures for any of the experimental sessions suggesting that all subjects were in an approximately equivalent behavioral state when each playback presentation began. Playback Song production and song rate: All subjects produced song during the playback segments during all of the conditions of the experiment. Although all playback segments were recorded, many of our recordings, again, were obscured by the playback presentation on the tapes. Thus, we were unable to discriminate many of the subjects songs from the pre-recorded playback songs either acoustically or visually from sonographic representations Accurate analyses of the song rates produced by males during the playback segments could not conducted.

20 20 Arena entry and latency: All subjects entered the arena during all the conditions of the experiment during the playback segment. Subjects latency to arrive within the playback arena under the various conditions of the experiment are summarized in Table X. There was a significant main effect of Song Rate on the latency to arrive within the arena during the playback segment (F(2,32) = 5.22, MSe , p<.05). Analytic comparisons revealed that, overall, subjects arrived within the arena in less time during the fast rate playbacks than during either the slow or ultraslow rate playbacks (F(1,32) = 10.02, p<.01). Flights: There was a significant main effect of Song Rate on the number of flights males made through the playback arena during the playback segment (F(2,32) = 6.39, MSe 8.08, p<.005). Main comparisons revealed that, overall, males made more flights through the arena during either the fast or slow rate conditions than during ultraslow rate condition (F(1,32) = 13.18, p<.01) Approach distance: Song Rate manipulations produced a significant main effect on the closest approach males made to the playback speaker during the playback segment (F(2,32) = 5.27, MSe 56.57, p <.01). Main comparisons showed that, overall, males approached the playback speaker more closely during either the fast or slow rate condition than during the ultraslow rate conditions (F(1,32) = 10.43, p<.01). There was also a main effect of Song Cadence on the closest approach to the playback speaker during the playback interval. Overall, male birds approached the playback speaker more closely during even-cadenced playback sessions (F(1,16) = 9.35, Mse 64.50, p <.01). Time within the arena: Statistical analysis revealed an interaction of Song Rate and Song

21 Cadence as they influenced the time each male spent within the playback arena during the 21 playback segment (F(2,32) = 3.78, p<.05 Mse ). Further analyses of Song Rate for each Cadence type revealed a significant simple effect of Song Rate when even-cadenced playbacks had been presented (F(1,32) = 7.54, p<.01). Additional simple comparisons showed that subjects spent more time within the arena during the fast even-cadenced playbacks than during the slow even-cadenced playbacks (F(1,32) = 4.58 p<.05). Subjects also spent more time within the arena during the ultraslow even-cadenced playbacks than during the slow even-cadenced playbacks(f(1,32) = 6.40, p<.01). The time spent within the arena by males was about the same for fast and ultraslow even-cadenced playbacks (F(1,32) =.011, p>.05). However, there was no simple main effect of Song Rate for the uneven-cadenced conditions (F(1,32) = 1.66, p<.05). Simple comparisons showed that subjects spent about the same amount of time within the arena during fast, slow and ultraslow uneven-cadenced playbacks. There was no main effect of Song Rate (F(2,32) =.90, p>.05 Mse ). Overall, birds spent about the same amount of time within the arena during the fast, slow and ultraslow rate levels. However, examination of the simple effects of rate for each cadence type showed the pattern to be accurate only for the uneven-cadenced conditions, whereas there was a simple effect of rate during the even-cadenced conditions. There was a significant main effect of Song Cadence (F(1,16) = , p<.01, Mse ) with more time, overall, spent within the arena by males during presentations of the even-cadenced playbacks. Examination of the simple effects of Cadence for each rate level revealed that the pattern was descriptive for fast, slow and ultraslow rates (F(1,16) = 41.53, p<.01), (F(1,16) = 14.65, p<.01), and (F(1,16) = 58.15, p<.01) respectively.

22 Post-playback 22 Song production and song rate: The male subjects were observed singing within the playback arena during the post-playback interval equally often across all conditions of the experiment (χ 2 =.932, p >.05). Subjects song rates produced during the post-playback segment for the various conditions of the experiment are summarized in Table x. There was an interaction between Song Rate and Song Cadence as they related to the song rate produced by males during the postplayback interval (F(2,32)=14.37, MSe p =.001). Further analyses revealed a significant simple effect of Song Rate on the subjects post-playback song rates when even-cadenced playbacks had been presented (F(2, 32) = p<.01). Additional simple comparisons showed that although the song rates produced by males were about the same for fast and slow evencadenced playbacks, subjects song rates slowed significantly after presentation of the ultraslow even-cadenced playbacks (F(1,32) = 20.97, p<.01). There was also a significant simple effect of Song Rate on the song rates produced after presentation of the uneven-cadenced playbacks (F(2,32) = 3.57, p<.05). Simple comparisons of the effects of Song Rate for uneven-cadenced playbacks showed that subjects produced about the same song rate for the fast and slow uneven-cadenced conditions, but sang significantly faster after the ultraslow uneven-cadenced condition (F(1,32) = 5. 48, p<.05). Examination of the simple effects of Song Cadence for each level of Song Rate revealed that males produced significantly more songs during even-cadenced playbacks than unevencadenced playbacks for both the fast (F( 1,32) = 48.81, Mse 43.93, p <.001) and slow (F( 1,32) = 38.89, p<.001) rate levels. However, birds produced song at about the same rate during even

23 and uneven-cadenced playbacks for the ultraslow rate condition (F(1,32) = -4.32, p>.05). 23 There was a significant main effect of Song Cadence on the song rate produced by males during the post-playback interval (F(1,16) = 31.34, MSe 82.66, p<.001. Overall, males sang at a faster rate during post-playback intervals after even-cadenced songs had been presented. However, there was no simple effect of Song Cadence during the post-playback segment after subjects were exposed to the ultraslow rate level. Flights: There were no differences in the number of flights each male made through the playback arena during the post-playback interval for any of the conditions. Time in the arena: There was an interaction of Song Rate and Song Cadence as they related to the time each male spent within the playback arena Discussion Discussion 1 parts The addition of the ultraslow rate condition produced song rate effects for several measures in the experiment. Male Red-eyed vireos must use their vocal signals to influence conspecifics of both sexes. For males, territoriality requires communications to be efficiently designed so that the probability that information will be received is maximized. Each bird can only afford to devote a finite amount of time to singing, with the remainder devoted to other life supporting and fitness enhancing activities. In order to gain and hold a territory, male vireos must be able to withstand the costs of time, and energy and risk spent while maintaining long, uninterrupted song bouts. Singing may incur direct energetic costs involved in the production of the songs or costs of maintaining the neural structures necessary for song production. Singing may also incur indirect costs as song

24 may alert predators as to the singer s location. 24 Godard (1993) demonstrated that male Red-eyed vireos have difficulty recognizing individual neighbors songs and suggested that the inability to discriminate between neighbors and strangers songs may increase energy and time expenditures during territory defense. Because males do not recognize song of direct neighbors, additional time must be spent defending shared boundaries repeatedly as neighbors may be perceived as intruders. Thus, time and energy expenditures related to song production for males of this species may be elevated. A territorial male vireo is under pressure to sing in such a way that his cost of advertising is outweighed by the fitness enhancing benefits derived from it. Producing songs with an even cadence may be an efficient strategy for a species which must sing lengthy and uninterrupted song bouts. This strategy may minimize the overall energetic costs of singing, while allowing the information to reach as many target receivers as possible. However, singing with an even cadence may serve as a beacon. The repetitious song cadence of the Red-eyed vireo is quite salient in a noisy environment thus adding an extra cost in terms of predation risk to an already expensive endeavor. This extra expense in terms of predation risk may be endured by only the more robust males. An alternative explanation could be posited that even-cadenced song is an indicator of low quality, thus provoking the more confident behaviors displayed by territory holders during our playback segments. However, the probability that a territory holding male would expend energy investigating a low quality intruder seem small. This seems especially likely in light of our finding that reproductively successful males (territory holders) sing in this manner. The even-cadenced playbacks, conducted from just within territory boundaries, elicited the strongest

25 responses from territory holders as predicted if an even song cadence is a reliable signal of 25 quality or robustness in male Red-eyed vireos. Males responded more strongly to our playback tapes which simulated an intruding male singing with an even cadence regardless of the song delivery rate. They spent more time within the playback arenas during presentations of even-cadenced song. Male Red-eyed vireos may use the cadence of rival males songs as a means of assessing the potential threat imposed by those males. The addition of the ultraslow rate condition produced some effects that were not detected in the first study. Faster song rates led to closer approaches and more flights through the playback arena overall. However, further examination of the results showed the pattern to be consistent with that observed in the 1995 study, namely that there were no significant differences between fast and slow song rate on any of the response measures. The effects of song rate on approach distance and number of flights through the arena only appeared with the addition of the ultraslow rate condition in the second study. All three song rate levels in the study are consistent with natural song rates produced by Red-eyed vireos as demonstrated by samples we obtained in the field. However, the ultra slow rate used in Experiment 2 is the rarest in a natural setting and is most often produced late in the breeding season (personal observation). The result suggests that faster song rates may reveal information about the arousal level, and the probable immediate behavior of the singer to other males. Territorial males may respond more quickly and excitedly to faster song rates which possibly indicate increased arousal level in an intruder. Increased song rates and mobbing behaviors in the presence of threat are commonly observed in many song bird species. Lawrence

26 (1953) reports investigative behavior in Red-eyed vireos responding to other males scolding 26 songs and we have observed this behavior on many occasions during our studies. Faster song rates may be associated with alarm due to predator approach as well. The ultraslow song rate presented in our second study, although definitely noticed, may be associated with low arousal in an intruder and diminished initial threat. Although males arrived more quickly at our playback arena during the faster rate playback trials, they remained within the arena longer and came closer to the playback speaker during playback of the even-cadenced songs regardless of song rate. Perhaps males are attracted initially by song heard within their territory boundaries, but evaluate a less variable cadence as a reliable index of the likelihood of an intruder s strength as a contender. Males producing songs with a less variable cadence may be judged to represent a valid adversary to a territory holder, thus the significantly longer time that males devoted to defending territories against such intruders, albeit simulated ones. General Discussion Early ethologists assumed that animal signals were always reliable indicators of the internal state of the signaler. Maynard Smith (1982) questioned this view by suggesting that if all signals were reliable then all animals would make them all the time. Earlier, Maynard Smith and Parker (1976) suggested that only assessment signals: i. e. those that reflected size or physical prowess (Maynard Smith, 1956) could be reliable. Zahavi (1975; 1977)) suggested that a signal could only be reliable (honest) if it carried a cost to the signaler such that only the most robust individuals could afford to produce it, thus production of the signal reflects fundamental quality in the signaler. We suggest that even song cadence may be signal feature, consistent

27 with this explanation, that reliably reveals quality in male Red-eyed vireos. 27 The outcomes of both of our studies are consistent with predictions that uniformity in intersong intervals may benefit singers by reducing listener uncertainty. This function could facilitate mate attraction and territory maintenance both of which are paramount to reproductive success for males and females. Although assessment of the actual costs of producing regular song cadence is beyond the scope of this research, we suspect that the costs of producing even-cadenced song may be incurred in several ways. Energetic output required for maintenance of neural structures necessary for regulation of the behavior may be high. Singing evenly may allow potential predators to localize the origin of the signals more readily. The ability to withstand the physical fatigue of producing a constant regular signal may be costly as well. However, the fact that territory holding male Red-eyed vireos sing with significantly less variance in their intersong interval suggests that it is reasonable to assume that territory holding male Red-eyed vireos are more able to withstand any costs associated with producing even song cadence. Therefore, the even song cadence exhibited by these males may be a reliable signal.

28 References 28 Andersen, M. (1982) Sexual selection, natural selection, and quality advertisement. Biological journal of the Linneaus society, 17, Beletsky, L. D. (1989). Communication and the cadence of bird song. American Midland Naturalist, 122, Borror, D. J. (1981). The songs and singing behavior of the Red-eyed Vireo. Condor, 83, Chariff, R. A., Mitchel, S., & Clark, C. W. (1995). Canary 1.2 User s Manual. Cornell Laboratory of Ornithology, Ithaca, New York. Fisher, R. A. (1935). The fiducial agreement in statistical inference. Annals of eugenics, 6, Goddard, R. (1993). Red-eyed vireos have difficulty recognizing individual neighbors songs, The auk, 110, Hoi-Leitner, M., Nechtelberger, H. & Hoi, H. (1995). Song rate as a signal for nest site quality in blackcaps (Sylvia atricapilla), Behavioral ecology and sociobiology, Kepple, G. (1991). Design and analysis: a researcher s handbook (3rd ed.). Englewood Prentice Hall. Cliffs: Kirk, R. E. (1968). Experimental design: procedures for the behavioral sciences. Belmont: Brooks/Cole. Kodric-Brown, A. & Brown, JH (1984) Truth in advertising: the kinds of traits favored in selection. American Nat. 124, Lawrence, L de K. (1953). Nesting life and behavior of the Red-eyed Vireo. Canadian field naturalist, 67, Lemon, R. E. (1971). Analysis of song of red-eyed vireos. Canadian journal of zoology, 49, Maynard Smith, J. (1956). Fertility, mating behavior and sexual selection in Drosophila subobscura, J. Genet., 54, Maynard Smith, J. (1982). Do animals convey information about their intentions? Journal of theoretical biology, 97, 1-5.

29 29 Maynard Smith, J. & Parker, G. (1976). The logic of asymmetric contests. Animal behaviour, 24, Moller, A. (1983). Song activity and territory quality in corn bunting Miliaria calandra: with comments on mate selection. Ornis scand, 14, Radesater, T., Jakobsson, S., Andbjer, N., Bylin, A., & Nystrom, K. (1987). Song rate and pair formation in the willow warbler, Phylloscopus trochilus. Animal behavior, 35, Searcy, W., Yasukawa, K. (1981) Does the sexy son hypothesis apply to mate choice? American Nat., 117, Smith, W. J. (1977). The behavior of communicating. Cambridge: Harvard University press Zahavi, A. (1975). Mate selection for a handicap, Journal of theoretical biology, 53, Zahavi, A. (1977). The cost of honesty (further remarks on the handicap principle). Journal of theoretical biology, 6,