Song Performance in Long Tailed Manakins Rosas 1 Song Performance of Coordinated Male Duet in Long Tailed Manakins ( Chiroxiphia linearis ) Stephanie Rosas Department of Ecology and Evolutionary Biology University of California, Santa Cruz EAP Tropical Biology and Conservation Program, Spring 2017 9 Jun e 2017 Abstract Animals use a variety of traits such as brightly colored plumage, nuptial gifts, chemical signaling, or intricate displays to attract mates. Long tailed manakins ( Chiroxiphia linearis) e xhibit one of the more elaborate mating rituals found in the animal world. C. linearis form lek mating systems, where a group of males aggregate for cooperative courtship displays. They pe rform an alpha beta duet that serves as a long range cue to attract females to their perch zone for further display. High er frequency matching in this duet has been correlated to higher courtship success. I investigated differences between alpha beta vocalizations within and between leks, as well as the relationship between their song performance and proximity to other leks. By collec ting data at 12 lekking sites and analyzing 426 vocalizations, I found a correlation between low and high frequency differences and a negative correlation between index frequency matching ad index time matching This suggests a female preference for overall frequency matching versus song synchronicity. I found no significant relationship between relative proximities of leks and either time matching or frequen cy matching This could indicate that long term partnerships and duet practice in C. line aris are more important than competition in increasing song perfo r mance. CaracterÂ’sticas y desempeÂ–o de los cantos coordinados en duetos masculinos de toledos ( Chiroxiphia linearis ) Resumen Los animales utilizan una variedad de rasgos tales como plumaje brillante y colorido regalos nupciales, seÂ–ales quÂ’mica s o demostraciones comp lejas para atraer a las hembras Los toledos ( Chiroxiphia linearis ) exhiben uno de los rituales de apareamiento mÂ‡s elaborados en el mundo animal. C. linearis forman sistemas d e apareamiento en forma de lek, donde los machos se agrupan para hacer demostraciones displays" cooperativas. Realizan un canto en dÂœo alfa beta que sirve como seÂ–al de largo alcance para atraer a las hembras a su perc h a para continuar haciendo su display La concordancia de frecuencia mÂ‡s alta en este dÂœo se correlacionÂ— con un ÂŽxito mÂ‡s alto de cortejo InvestiguÂŽ las diferencias entre vocalizaciones alfa beta dentro y entre leks, asÂ’ como la relaciÂ—n entre el desempeÂ–o de su canciÂ—n y la proximidad a ot ros leks. TomÂŽ datos en 12 sitios de lek y analicÂŽ 426 vocalizaciones E ncontrÂŽ una correlaciÂ—n positiva entre las difere ncias de frecuencia alta y baja y una correlaciÂ—n negativa entre el Â’ndice concordancia de frecuencia con el Â’ndice de concordancia de tiempo Esto sugiere que hay preferencia de las hembras por la concord ancia de frecuencia general y no la sincronÂ’a de la canciÂ—n. No encontrÂŽ ninguna relaciÂ—n significativa entre las proximidades relativas de los leks y la c oncordancia de tiempo o la conc ordancia de frecuencia. Esto podrÂ’a indicar que las asociaciones a largo plazo y
Song Performance in Long Tailed Manakins Rosas 2 la prÂ‡ctica de dÂœo en C. linearis son mÂ‡s importantes que la competencia para aumentar el de sempeÂ–o de su canciÂ—n. Introduction Animals use a variety of traits such as brig htly colored plumage, nuptial gifts, chemical signaling, or intricate displays to attract mates. Elaborate traits or performances are often cond ition dependent, with higher quality males displaying traits more desirable by females (Jiguet and Bretagnolle 2 014). Auditory signals such as songs or mating calls often serves as an important factor in female choice, with subtle differences in focal characteristics affecting mating success (Botto and Castellano 2016; Buchanan and Catchpole 2000; Buchanan et al. 20 03; Gibson et al. 1991). A study of lek breeding amphibians found a significant association between mating su ccess and call rate and duration (Botto and Castellano 2016). Increased repertoire size in sedge warblers is known to correlate with increased pare ntal care and therefore higher mating success (Buchanan and Catchpole 2000), and in the lekking bird little bustard higher mating success is correlated with higher call rates (Jiguet and Bretagnolle 2014). In long tailed manakins ( Chro xiphia linearis ), fre quency matching of alpha beta duets correlate to an increase in courtship su ccess (Trainer and McDonald 1995). Long tailed manakins exhibit one of the more elaborate mating rituals found in the an imal world. C. linearis form lek mating systems, where a group of males aggregate for courtship displays that attract females and are subject to female choice (McDonald and Potts 1994). Lek mating systems are rare, accounting for less than 1% of all bird species (Jiguet et al. 2000). Long tailed manakin leks con sist of many males, but only the two top males sing and dance in cooperative alpha beta pairs (McDonald 1989; Trai ner and McDonald 1995; Trainer et al. 2002). While beta males are essential in attracting and impressing females, generally only the alpha cop ulates (McDonald 1993; McDonald and Potts 1994; Trainer and McDonald 1995; Trainer et al. 2002). Males increase their status in the lek over many years, generally b e com ing betas at age seven and alphas by age ten ( Trainer et al. 2002 ). Interestingly, it ha s been shown that alpha and beta partners are not related as one might expect, but that coo p eration between partners exists because of long term benefits (McDonald and Potts 1994). These include rare copulations as betas, eventual succession to alpha role, and inheritance of the perch zone (McDonald and Potts 1994; Trainer et al. 2002 ). While C. linearis have several songs and calls, it is the toledo duet tha t serves as a long range cue to initially attracts females to their perch zone (McDonald 1989; Tra iner and McDo nald 1995). This song serves as a signal of quality, as high quality performances are often the r esult of long term alpha beta partnerships and continued practice (Trainer and McDonald 1995). Studies have shown how frequency matching affects c ourtship success (Trainer and McDonald 1995), but have not looked at the potential relationship between frequency matching and proxi mity to other leks. In a study of three species of passerine birds, individuals in higher density pop ulations sang more ofte n than individuals in low density populations (McShea and Rappel 1997). Increases in population densities can increase intraspecific competition, which has the potential drive selection towards higher quality traits. This study examines how variables of so ng perfo rmance differ between alpha beta pairs, and how that might be affected by their proximity other leks. I predict leks of C. linearis in close to proximity to other leks will experience more intense competition for females, which will lead to increas ed song performance.
Song Performance in Long Tailed Manakins Rosas 3 Methods and Materials I used four sites to collect data on 12 leks of C. linearis in Monteverde, Costa Rica over a two week period in May of 2017. These sites include the Bajo del Tigre Reserve (BT), Santuario Ecologico (SE), Crand ell Reserve (C), and the Magic Path (MP) in the Bajo del Tigre neighbo rhood. Each site was approximately 500 m from the nearest lek. At each site, I listened for toledo calls and followed them to potential perch zones. I monitored the area until I observed a pair pe rforming their mating dance on a perch, at which point I would designate it as an individual lek and perch zone. I located one lek at both C and MP, three leks at BT, and 7 leks at SE. I recorded the latitude and longitude of each perch using the GPS on my phone. At eight of the twelve perch zones I used a Sehnnheiser MK II microphone and a ZOOM H4n Pro recorder to record synchronized alpha beta toledo vocalizations at a sample rate of 48kHz (24bit). The other four perches were not active during the recording times and no vocal izations were obtained. For each of the eight alpha beta pairs, I accumulated a range of 25 to over 100 toledo vocalizations depending on the activity level of the individual perches. Of those r ecordings I analyzed an averag e of 27 songs from each alpha beta pair, or 54 individual songs per lek. Usable recordings were based on their quality and clarity. I used Raven Pro 1.5 software to analyze a total of 426 individual toledo voca lizations, or 213 duets Spectrograms were di gitized at 48 000 Hz and 24 bit. Measurements were obtained u s ing the following parameters: a frequency resolution of 256 samples, a grid spacing of 31.3 Hz and a time grid with 50% over lap using the window Hann function. I determined the beginning and e nd time, and low and high frequency for both the alpha and beta songs of each duet (see Fig. 2 for example) All calculations were done using the absolute values of raw data, as it was the differences in magnitude I was interested in. I calculated the low frequency differences (LFD) and high frequency differences (HFD) between each duet. I also calculated start times differences (STD) and end time differences (ETD) of each song, as well as averages per perch zone. I created an index of frequency matching (I FM) by using the formula IFM = 240 (LFD + HFD ) based on a formula used in a study by Trainer and McDonald (1995). I used the value 240 because it was larger than the greatest (LFD+HFD) value, ensuring all IFM values would be po sitive. Using the index tra nsforms large differences in frequencies into low values of frequency matching, and small differences into high values of frequency matching. The same method was used to create an index of time m atching (ITM) using the formula ITM = 2 (BTD+ETD) which ser ves as a measure of time synchronicity between alphas and betas. Low ITM values indicate poor synchronization, while high ITM values indicate exceptional synchronization. Using GPS data, I plotted latitude and longitude of each perch in Excel (Fig. 1). Fo r each location where vocalizations were recorded, I found the average distance between the vocaliz ation site and the two closest leks. I transformed these values by multiplying each by 1000, to better differentiate between averages. I used these values as a distance proxy, with lower values indicating closer proximity to neighboring leks and higher values indicate more isolated leks. I used JMP Pro 13.0 software to analyze the statistical relationship of variables within a song pe rformance, as well as the relationship between lek proximity and song performance.
Song Performance in Long Tailed Manakins Rosas 4 Results IFM, ITM, and distance proxy were calculated for the eight leks where vocalizations were obtained, a summary of results are given in Table 1. Overall IFM was more variable than ITM, whi ch remained relatively consistent (Table 1, Fig.3, Fig. 4). Lek S2 had both the lowest IFM value (48.5723.04) and highest ITM value (1.8120.042) (Table 1, Fig.3, Fig. 4). Site C has the highest distance proxy value (151.82) indicating it was the most iso lated, and S2 has the lowest value (6.29) indicating it was closes t to other leks (Table 1, Fig. 1 ). T here is a significant correlation between LFD and HF D (correlation, p=0.0372, Fig. 5 ). Difference in high frequency increased as difference in low frequen cy increased. This indicates that overall frequency matc hing is preferred rather than a particularly high or low frequency call. There was also a significant negative correlation between IFM and the IT M (correlation, p=0.0048, Fig. 6 ). As index of fr equenc y matching increases, the index of time matching decreases Many leks were better at fr equency matching than time matching (Fig. 6) Regression analysis shows the relationships b etween distance proxy and both ITM and IFM are not significant (regres sion, p= 0.1036, p=0.2889, Fig. 7, Fig.8 ). The proximity to other leks did not affect variables of song perfo rmance. Figure 1. Proximity map of study sites. Latitude and longitude values were used calculate ave rage distances to two nearest leks. Leks in Santu ario Ecologico were in the closest proximity of each other. The lek found at the Crandall Reserve was the most isolated. S= Santuario Ecologico; C= Crandall Reserve; MP= Magic Path; BDT= Bajo del Tigre Reserve S1 S2 S3 S4 S5 S6 S7 C MP BT1 BT2 BT3 10.304 10.305 10.306 10.307 10.308 10.309 10.31 10.311 10.312 10.313 -84.826 -84.824 -84.822 -84.82 -84.818 -84.816 -84.814 -84.812 -84.81 -84.808 -84.806 Longitude Latitude
Song Performance in Long Tailed Manakins Rosas 5 Table 1. Summary of results for each lek rec orded vocalizing. Includes total number of songs analyzed for each lek, with averages and standard deviations of index frequency matching (IFM) and index time matching (ITM). Low values correspond to low levels of matching and high va lues correspond to hig h values of matching. Distance proxy values for each lek are given, with low values indicating close proximity to other leks and high values indicating relative isolation. S= Santuario Ecologico; C= Crandall Reserve; MP= Magic Path; BDT= Bajo del Tigre Res erve. Figure 2. Image from Raven Pro 1.5 software of an alpha beta toledo duet. The red rectangle surrounds the alpha call and the turquoise rectangle surrounds the beta call. Yellow arrows ind icate beginning and end times of each calls. Blue arrows indicate high and low frequencies of each calls. Lek Songs analyzed (alpha+beta) IFM ITM Distance Proxy S1 50 176.1233.24 1.7830.046 6.92 S2 70 48.5723.04 1.8120.042 6.29 S3 34 146.4039.91 1.7970.058 11.01 C 56 199.2316.06 1.7700 .036 151.82 MP 52 143.8023.65 1.7730.057 94.89 BT1 60 185.5020.96 1.7750.025 43.06 BT2 50 184.5723.45 1.7720.037 25.39 BT3 54 136.0445.83 1.7970.038 33.01
Song Performance in Long Tailed Manakins Rosas 6 0. 55. 110. 165. 220. 275. S 1 S 2 S 3 C MP BT 1 BT 2 BT 3 Index Frequency Matching (Hz) Lek 1.575 1.65 1.725 1.8 1.875 S 1 S 2 S 3 C MP BT 1 BT 2 BT 3 Index Time Matching (s) Lek Figure 4 Average time index matching and standard deviations and each vocalization site. Variation between sites remained relatively consistent compared with average index frequency matching values. See Table 1 for exact values. S= Santuario Ecologico; C= Crandall Reserve; MP= Magic Path; BDT= Bajo del Tigre Reserve. Figure 3. Average index frequency matching and standard deviations at each vocalization site. There is high variation between and within sites. See Table 1 for exact values. S= Santuario Ecologico; C= Crandall R eserve; MP= Magic Path; BDT= Bajo del Tigre Reserve.
Song Performance in Long Tailed Manakins Rosas 7 Figure 5. Correlation between low frequency differences and high frequency differences. As low frequency differences increase, high frequency differences also increase (correlation, p=.0372). Figure 6. Correlation between index frequency matching and index time matching. As index frequency matching increases, index frequency matching decreases (correlation, p=.0048). Figure 7. Regression analysis between distance proxy and index time matching. No significant relationship was found between the proximity of leks and index time matching (regression, p=.1036).
Song Performance in Long Tailed Manakins Rosas 8 Figure 8. Regression analysis between distance proxy and index frequency matching. No significant relationship was found between the proximit y of leks and index frequency matching (regression, p=.2889). Discussion Song performance differed between leks, making it a suitable candidate for selection by female choice. The correlation between LFD and HFD implies that overall frequency matching i s important in the toledo call. This positive correlation suggests that team practice improves fr equency matching ov er time at similar rates. This may signify female choice for overall frequency matching in pairs which could signal the age of the males, t he ability to coordinate high quality displays, and the ability to cooperate in a stable partnership (Trainer and McDonald 1995). The variation in IFM between pairs was higher than the variation in ITM. High variation in frequency matching makes sense, a s it has been suggested that a team's frequency matching improves as the alpha beta partnerships develop (Trainer and McDonald 1995). It is likely that different leks have alpha beta partnerships of varying duration, with longer partnerships likely having a higher level of frequency matching. What is interesting is that there is relatively low variation in time matching. This could imply that time matching is not a skill easily acquired, and only small improvements are made throughout one's life regardless of effort. It could also su ggest that time matching is not a priority for C. linearis They may choose to focus their energy on frequency matching because it increases mating success (McDonald 1989), rather than on time matching which may not be an importa nt factor in female choice. Similarly, the negative correlation between IFM and ITM could indicate the cause of di sparity between time and frequency matching. If the improvement of each trait is costly, a team may choose to focus on only one. Teams with high IFM had low ITM and vice versa, no pair excelled at both. Interestingly most leks were better frequency matchers than time matchers, which may be caused by benefit of improved mating success (McDonald 1989). Proximity to other leks does not drive per formance quality in the toledo call of C. linea ris Contrary to what I predicted, there was no indication that increased lek densities led to higher competition and improved frequency or time matching. This could indicate that long term par tnerships and pr actice are more important than competition in increasing song performance. It has been shown that song performance improves with age, and hypothesized that frequency matc hing develops over several years of partnership ( Trainer et al. 2002 ). While lekking species are often territorial (Jiguet et al. 2000), C. linearis are known for their high levels of cooperation
Song Performance in Long Tailed Manakins Rosas 9 (McDonald 1993; McDonald and Potts 1994). It could be this lack of aggression, territoriality, or competition that prevents lek densities from af fecting a team's pe r formance. Long term re search could provide insight on specific factors affecting song performance in C. linearis It would be interesting to study the relationship between song quality and both age of individual manakins and duration of their partnerships. I am curious if time and practice can always improve performance, or if some individuals experience an innate or environmental lim itation as to the quality that can be achieved. Additionally, it is important to further study the e ffe ct of time matching on courtship success, as has been done with frequency matching. Studies of long tailed manakins help us further understand the many complexities and nuances of this unique lekking species. Acknowledgments I would like to sincerely thank my primary adviser Andres Camacho for all his guidance and support throughout the many challenges I experienced with this project. Thank you for many early mornings of mist netting practice and assistance, especially when accompanied by your perfect wife and daughter, Hel en and Olivia. Even though that data didn 't make it to my final project, it is a skill I am truly grateful to have gained and will take it with me into my fu ture work as a field biologist. Thank you to my secondary adviser Emilia Tr iana, who provided me with thoughtful feedback on my paper, and helped me clarify and edit my ideas. Special thanks to Siria and Alvaro Salazar, for providing me with a welcoming home and meaningful convers ations during my homestay. Siria, your delicious h ome cooked meals made all the difference after long stressful days in the field. Thank you to the Monteverde Institute, Santuario Ecologico, and the Bajo del Tigre Reserve for allowing me to conduct research on your properties, it was an experience I will never forget. Literature Cited Botto, V. and S. Castellano (2016). Signal reliability and multivariate sexual selection in a lek breeding amphibian. Behaviorial Ecology 27:6:1797 1807. Buchanan, K. and C. Catchpole (2000). Song as an indicator of male parental effort in the sedge Warbler. Proceedings: Biological Sciences 267:1441:321 326. Buchanan, K., Spencer, K., Goldsmith, A. and C. Catchpole (2003). Song as an honest signal of past developmental stress in the European starling ( Sturnus vulgaris ). Proceedings: Bio logical Sciences 270:1520:1149 1156. Gibson, R., Bradbury, J. and S. Vehrencamp (1991). Mate choice in lekking sage grouse revis ited: the roles of vocal display, female site fidelity, and copying. Behavioral Ecology 2:2:165 180. Jiguet, F., Arroyo, B. and V. Bretagnolle (2000). Lek mating systems: a case study in the Little Bustard Tetrax tetrax Behavioral Processes 51:63 82.
Song Performance in Long Tailed Manakins Rosas 10 Jiguet, F. and V. Bretagnolle (2014). Sexy males and choosy females on exploded leks: Corre lates of male attractiveness in Little Bustard. Behavioral Processes 103:246 255. McDonald, D. (1989). Correlates of male mating success in a lekking bird with male male coop eration. Animal Behavior 37:1007 1022. McDonald, D. (1993). Delayed plumage maturatio n and orderly queues for status: a manakin mannequin experiment. Ethology 94:31 45. McDonald, D. and K. Potts (1994). Cooperative display and relatedness among males in a lek mating bird. Science 266:5187:1030 1032. McShea, W. and J. Rappole (1997). V ariable song rates in three species of passerines and impli cations for estimating bird populations. Journal of Field Ornithology 68:3:367 375. Trainer, J. and D. McDonald (1995). Singing performance, frequency matching and courtship success of long ta iled manakins ( Chiroxiohia linearis ). Behav Ecol Sociobiol 37:249 254. Trainer, J., McDonald, D. and W. Learn (2002). The development of coordinated singing in cooperatively displaying long tailed manakins. Behavioral Ecology 13:1:65 69.