Female Zebra Finch (Taeniopygia guttata) Preference Test of Male Courtship Song Complexity
ABSTRACT
Assessing female bird preference for male courtship song complexity can provide insight on the characteristics that females select for on an evolutionary scale. Female zebra finches (n=9) were trained to peck two keys in an operant task preference test in which each key would trigger an audio playback of either a complex or simple zebra finch song. It was hypothesized that females would select for more complex songs since previous findings suggest that song complexity provides insight on the quality of a male bird [1]. Once trained, females will first select for either a white noise playback or a zebra finch song to determine if they are a reliable source before selecting for song complexity. Although training is ongoing, females indicated a preliminary preference during the training period favoring zebra finch audio over white noise (p=0.008). This indicates that the females will be a reliable source for analysis when evaluating preference for complexity. Additionally, one female demonstrated a preference for a specific key rather than the audio (p=0.04). Since the zebra finches did not learn the task within the first four training sessions, new techniques were implemented to accommodate their social biology and encourage engagement in the task. Mirrors were placed behind each key to create the illusion of another bird present during individual training to entice the females to approach the keys and peck. Ultimately, these initial training results provide a baseline for future observation of the most effective training methods and testing of complexity once the operant task is learned by the zebra finches.
INTRODUCTION.
Sexual selection, when individuals choose a mate based on certain characteristics, is a driving factor that determines which traits are passed on to future generations [2]. In the Oscine songbird clade, song is a behavioral trait learned by juveniles in early stages of life from adults in the population [3]. Females collect lots of information from birdsong, such as the condition and quality of potential mates [3]. Especially when juveniles commonly learn songs from their fathers, sexual selection could be a predominant force determining which songs are prevalent in the next generation [4]. Female preference for song characteristics such as length, amplitude, and repertoire size, have been studied. However, the nuances of which aspects of song complexity has yet to be tested on specific species (e.g., Taeniopygia guttata) with current technologies [4].
T. guttata are a primary model organism for studies of vocal learning since they are easy to rear in a laboratory environment and exhibit complex communication behaviors and social tendencies [5]. Female zebra finches do not sing but have been shown to learn their father’s (or other tutors’) songs, in a similar learning pattern to their male siblings [6]. Males mimic their tutors’ song to use later when attracting a mate, whereas females show a preference for songs with similar characteristics to their tutors’ when selecting a mate [6]. Through this process, females prefer songs most similar to males that have successfully reproduced, ultimately allowing already successful song characteristics to be passed down to future generations.
Determining if unbiased female zebra finches (i.e., previously unexposed to male song) prefer songs displaying different aspects of song complexity (for example, number of unique syllables, song duration, syllable entropy, frequency modulation) would reveal what specific factors females prioritize when choosing a mate. Additionally, these findings could reveal what evolutionary pressures are placed on communication systems and how they are realized in bird behavior. A previous study determined that female zebra finches prefer super-normal length songs (i.e., songs artificially extended by 1.8 seconds beyond the natural duration, essentially suggesting that females prefer longer songs), a characteristic said to be linearly correlated with some metrics of complexity [7]. Additionally, Spencer et al. found that when male zebra finches experience early-life stressors (such as nutritional deprivation) they produce less complex songs (i.e. songs with shorter duration and fewer syllables), suggesting that song is a clear indicator of condition in males [3]. Together, these studies identify complexity as a multifaceted and valuable characteristic to study further.
Previously, researchers conducted a large-scale study across multiple Oscine birds – ultimately revealing that adult birdsong plasticity is connected to larger song and syllable repertoires [8]. This indicates that sexual selection for larger repertoires could favor birds that present lengthier learning periods [8]. By narrowing the species studied to just one, we can determine if zebra finches could be driving increased learning periods in their early development through sexual selection.
By observing female song preference, we can gain insight on the driving forces of sexual selection – essentially learning how and why specific traits are sustained or lost over evolution of species. This can be determined by what characteristics females prioritize displayed through song.
Since song complexity can provide insight into male zebra finches quality, it is hypothesized that the females will prefer the complex song when offered the choice. This result would further support the connection between male song complexity and male quality.
Overall, by testing female zebra finch preferences for song complexity we aim to gain insight into their preferences for certain traits and even possibly connect it to selection for longer song learning periods.
MATERIALS AND METHODS.
Subjects.
All female zebra finches were hatched in the Creanza lab from the lab colony and were raised in captivity. Their ages ranged from 6 to 24 months old, the age of sexual maturity, but all females remained unmated. Subjects were quarantined from listening to courtship songs and remained unexposed to their father’s song to prevent a predisposed preference. A total of 6 females will be used in this study (n=6). Females had constant access to Kaytee Supreme Parakeet Food supplemented with red proso millet seeds, and water.
Since previous exposure to male tutors or social modeling experiences could induce a preference for vocalization of a similar nature, birds in this study were reared without their father and were housed with other females in a sound-attenuating chamber prior to the experiment [6]. No past exposure to male birds or vocal sounds (as females do not sing) allows for a purely unbiased examination of song preference. By performing an operant task preference test, we aim to assess whether female zebra finches show a preference for song complexity, and, if so, for which aspects of complexity Previously, passive preference tests were utilized where the female is established to prefer the song she is in proximity to in an enclosed environment [4]. This has proven to be fruitless at times since a female zebra finch may approach an unfamiliar song for multiple reasons rather than preference alone. In this study, a key-pecking preference test was conducted to determine if there is a clear song preference in the females [9]. This test allows females to peck a key that plays a song as a reward. The song that was initiated through the keys most consistently is considered to be the favored song.
Housing.
Females were removed from their colony housing and moved to a secluded holding chamber (length x width x height = 32” x 12” x 14”, with cage inside measuring 17.5” x 9.5” x 12”) where they were housed in the intermediate period between training and testing. The chambers were kept between 20 and 25 degrees Celsius, and humidity remained below 80%. Artificial lights were set to be active between 8 a.m. and 8 p.m. (12-hour light period, 12-hour dark period). This light-dark cycle aligns with the colony’s light-dark cycle which the females have already grown accustomed to. A minimum of 24 hours were allowed for the birds to acclimate to the new environment before being relocated to the training/testing environment. There were 2-3 females per holding chamber to prevent complete isolation. Females would be moved to an individual sound-attenuating chamber (length x width x height = 32” x 12” x 14”, with cage inside measuring 17.5” x 9.5” x 12”) for testing but would remain there for no longer than 24-hours at a time before returning to their holding chamber. Food and water were available at all times. The seed mix was placed below each key, and water was placed next to each key on both sides of the testing chamber to limit a preference for a specific side of the cage.
Stimuli.
The preliminary test presented a white noise playback and a medium complexity audio. The white noise (uniform distribution, -6 dBFS) was obtained from audiocheck.net while a medium complexity zebra finch song was sourced from a previous male tutor zebra finch recorded in the Creanza lab. The medium complexity zebra finch song was classified as such because it consisted of approximately three syllables.
All zebra finch playbacks used for the complexity test were recorded in the Creanza lab from males of the same species of past colonies. Simple songs were classified as being equivalent to three or fewer syllables while complex songs consisted of more than three syllables. Twelve audio samples were collected, and one was assigned randomly to each bird being tested. This ensured that each bird had a different randomly selected audio playback during each test, so no pattern recognition was formed.
Amplitude was kept in the range of 45 to 55 dB SPL for all playbacks to prevent a preference for amplitude rather than complexity [10]. Volume was set at a comfortable level (between 45-55 decibels) since zebra finches are sensitive to sound. Similarly, volume was kept constant across all training and testing.
Preference Test.
An operant task preference test was conducted with bird song acting as the reward (due to the zebra finches’ social nature) in order to determine what song type was preferred by the females [9]. Circular red dots with a diameter of 1 cm were placed at the end of both keys with a perch directly beneath each key (Figure 1) [11]. Keys were positioned so that each red dot was between the height of the bird’s chest and head (Figure 1). A National Instruments USB-6501 connected the keys to the computer to capture the key pecking data. Sound Analysis Pro 2011 (SAP; version 2011.104) was used to automate the keys by triggering the playback when a key was pressed.
Since it was hypothesized that each bird would learn the task at a different rate, all birds were allowed an acclimation period to learn the key pecking task independently beginning at 4 p.m. the day prior. Keys were active during the environment’s light/dark schedule, from 8 a.m. through 6 p.m., allowing them time to learn the task without reinforcement through curiosity or accidental pressing of the keys. Beginning at 10:30 a.m. the following day, millet seeds were placed on each key for two hours allowing the females to learn the task with a food incentive (Table 1). Once this period concluded, the leftover millet was removed, and the birds were given the following hour to peck on their own. Testing began once a female exhibited 5 total pecks in the hour-long period without food acting as a motivator to peck the keys. If the criteria were not met, the audios on each key were switched to the opposite key and millet was placed as a motivator for the next two hours. To avoid stressing the animals, no bird was kept in complete isolation for more than 24 hours. Once training concluded for the day, the birds would be returned to their holding chamber and were only trained again a minimum of 24 hours later.
| Table 1. Training and testing timeline. Order of training and testing process. Timeline showing steps beginning with acclimation all the way to the complexity preference test while current progress follows the daily training period.
|
|
| Acclimation period (12+ hours) |
Bird placed at 4 p.m. the day prior Keys active 8 a.m. through 6 p.m. |
| Training #1 (1 hour) |
Food incentive placed on keys Key 1: white noise Key 2: zebra finch audio |
| Observation period (1 hour) |
Food motivation is removed
|
| Training #2 (1 hour) |
Food incentive placed on keys Key 1: zebra finch audio Key 2: white noise |
| White noise vs. Zebra finch audio (2 hours) |
No food incentive Audios switch keys after 1 hour Females that select for the zebra finch song are used in further analysis |
| Complexity test (3 hours) |
Complex vs. simple courtship song Audios switch keys after 1 ½ hours |
Once a female fulfilled the training criteria, they would be tested to determine if they demonstrated a preference for their own species song rather than just noise [11]. Selecting for the medium complexity zebra finch audio over the white noise audio deemed them suitable for further analysis. Birds that demonstrated a preference for a specific key versus a specific audio (determined via a chi squared analysis) were removed from the study. All females (n=9) would complete both the preliminary and complexity test; however, only the females that accurately selected for the zebra finch song (n=5) would be used in further analysis (Table 1). After clearly demonstrating that they have learned the operant task by displaying a minimum of five key pecks in the period allowed to peck without a food motivator, females immediately began testing. When testing for white noise vs. zebra finch song, females were evaluated for a total of two hours where the audios on each key was switched halfway through the test.
Immediately following the preliminary test, the complexity test would begin. This test would last a total of three hours with a switch of the audios on each key halfway through [12]. Once this test concludes all females would be returned to their original holding chamber with between one and two other females.
Statistical Analysis.
Using R-Studio, a paired t-test was conducted to test the difference in white noise and songbird audio choice [13]. Additionally, a chi-square test was utilized to determine key peck preference.
RESULTS.
Female zebra finches demonstrated a preliminary preference favoring the medium complexity zebra finch song compared to the white noise audio when food motivation was present for both (p=0.04; Figure 2).
Additionally, when testing whether females display a preference for a key rather than audio (Figure 3), one out of six females demonstrated a clear choice for a key rather than an audio (χ2 test, p<0.001; individual Lg68) while the other five birds tested did not (p>0.05).
DISCUSSION.
When determining female zebra finch preference of male courtship song complexity, we first hypothesized that a more complex song would be preferred amongst the finches. To ensure that the females are a reliable source for testing the preliminary test of white noise compared to zebra finch song was conducted. Since females presented to be slower learners than anticipated, data from the training period has been collected to demonstrate any preliminary preferences. A significant preference was observed for the medium complexity zebra finch song compared to white noise (Figure 2). These findings indicate an increase in the number of pecks for the zebra finch audio compared to the white noise audio, which is in line with previous findings [11]. This suggests that the females are suitable for further analysis on future complexity testing, as they select for their favored audio. Furthermore, when analyzing for a preference amongst each individual bird for side rather than the stimulus, one out of the six birds displayed a clear side bias (Figure 3). Lg 68 showed a significant bias for key #2 indicating that the data was slightly skewed towards a key rather than an audio. This female will be removed from the larger study due to the clearly developed bias, deeming her an unsuitable subject for further training and testing [11]. Overall, these preliminary results indicate that the females are selecting for their own species song. This shows that the females are differentiating between the audios and will eventually be able to be used in the complexity tests when the task is learned.
In terms of training, since the females are taking a long period to learn the task, strategies have been implemented to increase the efficiency of training [8]. Two additional chambers have been set up to increase the quantity of birds that can be trained at a given time. Additionally, mirrors have been placed behind each key (2 per chamber) to create the illusion of another bird present in the environment. Since zebra finches are such social animals and have been isolated from their colony as a whole, this technique is intended to motivate them to approach the keys. In preliminary findings from our laboratory, females have already displayed increased curiosity as seen through a larger quantity of key pecks with the mirrors present.
In the future, other strategies such as red LED lights could be incorporated on each key to draw the female’s attention towards the keys or even be used as a reward system as seen in previous research [12]. Additionally, while I was unable to draw sufficient data to draw a well-founded conclusion on my original question, I was able to provide insight into improved training methods to teach zebra finches an operant task preference test. Since previous training methods were scarce, I intend to share my methods to show the successes and failures of different strategies.
ACKNOWLEDGMENTS.
Thank you to Dr. Nicole Creanza, Darra Boyer, and Ximena Leon for their guidance on this project. Additionally, thank you to the Creanza lab at Vanderbilt University and to my advisor Dr. Menton Deweese with the School for Science and Math at Vanderbilt for this opportunity and guidance!
REFERENCES.
[1] S. A. Buchanan, K. Spencer, A. R. Goldsmith, C. K. Catchpole, Song as an honest signal of developmental stress in the zebra finch (Taeniopygia guttata). Biol. Conserv. 142, 125–131 (2003).
[2] C. Darwin, The Descent of Man, and Selection in Relation to Sex, Vol. 2 (D. Appleton and Co., New York, 1872).
[3] Y. Chen, O. Clark, S. C. Woolley, Courtship song preferences in female zebra finches are shaped by developmental auditory experience. Proc. R. Soc. B 284, 20170054 (2017).
[4] M.-J. Holveck, K. Riebel, Female zebra finches learn to prefer more than one song and from more than one tutor. Anim. Behav. 88, 125–135 (2014).
[5] E. M. Wall, S. C. Woolley, Social experiences shape song preference learning independently of developmental exposure to song. Proc. R. Soc. B 291, 20240358 (2024).
[6] K. Riebel, I. M. Smallegange, N. J. Terpstra, J. J. Bolhuis, Sexual equality in zebra finch song preference: Evidence for a dissociation between song recognition and production learning. Proc. R. Soc. Lond. B 269, 729–733 (2002).
[7] S. Collins, Super-normal length song preferences of female zebra finches (Taeniopygia guttata) and a theory of the evolution of bird song. Evol. Ecol. 5, 537–549 (1991).
[8] C. M. Robinson, K. T. Snyder, N. Creanza, Correlated evolution between repertoire size and song plasticity predicts that sexual selection on song promotes open-ended learning. eLife 8, e44454 (2019).
[9] K. Riebel, P. J. B. Slater, Testing female chaffinch song preferences by operant conditioning. Anim. Behav. 56, 1443–1453 (1998).
[10] M. Ritschard, K. Riebel, H. Brumm, Female zebra finches prefer high-amplitude song. Anim. Behav. 79, 877–883 (2010).
[11] L. L. Maguer, S. Derégnaucourt, N. Geberzahn, Female preference for artificial song dialects in the zebra finch (Taeniopygia guttata). Ethology 127, 537–549 (2021).
[12] D. Alam, F. Zia, T. F. Roberts, The hidden fitness of the male zebra finch courtship song. Nature 628, 117–121 (2024).
[13] Posit Team, RStudio: Integrated Development Environment for R (Posit Software, PBC, Boston, MA, 2025). Available at http://www.posit.co/ (accessed 19 February 2026).
[14] S. C. Woolley, A. J. Doupe, Social context–induced song variation affects female behavior and gene expression. PLoS Biol. 6, e62 (2008).
Posted by buchanle on Thursday, May 14, 2026 in May 2026.
Tags: Bird behavior, evolution, operant task, preference, songbird