Finger 1 Effects of Lek Location and Female Presence on Dancing Behavior in Long tailed Manakins ( Chiroxiphia linearis ) Dennis Finger Department of Environmental Science, Policy, and Management University of California, Berkeley EAP Tropical Biology and C onservation Program, Spring 2018 8 June 2018 ABSTRACT Sexual selection, the process by which individuals of one biological se x select mates of the other sex, often leads to the formation of displays, ornaments, and weapons to improve mating success. Lo ng tailed manakins ( Chiroxiphia linearis ) exhibit a unique example of sexual selection in their bright plumage, long tails, and coordinated lekking dances. I investigated sexual selection by analyzing dancing behavior of long tailed manakin s in Monteverd e, Costa Rica. I film ed leks to measure the effects of dance variant (Cartwheel or Up Down), lek location , and female presence on variation in dance features . Cartwheel bouts of dancing occu r r ed only when the female was found on the dancing branch , while U p Down bouts occur red only when the female was on a nearby branch or not present at all . I found that dance features did not vary with lek location , but Up Down male coordination did vary by lek . In addition, Up Down bouts varied with f emale presence , as m ales jumped faster when females were nearby . Female sexual selection was a strong determinant of dance features, leading to bouts that resembled each other and showed limited variation. Bouts did vary, however, in coordination and jump rate, showing that l ek location and female presence can affect manakin dancing behavior. Efectos de ubicaciÂ—n de lek y presencia de la hembra en la danza en los toledos ( Chiroxiphia linearis) RESUMEN SelecciÂ—n sexual, el proceso en el cual los individuos de un sexo biolÂ—gico selecciona n parejas del otro sexo, a menudo conduce a la formaciÂ—n de exhibiciones, ornamentos , y armas para mejorar el ÂŽxito del apareamiento. Los toledos ( Chiroxiphia linearis) muestran un ejemplo Âœnico de la selecciÂ—n sex ual con su plumaje brillante, cola larga , y danzas coo rdinada s que los machos realizan en un lek . InvestiguÂŽ la selecciÂ—n sexual analizando el comportamiento de danza en los toledos en Monteverde, Costa Rica. FilmÂŽ leks para medir los efectos de tipo de da nza (volter et a lateral o arriba abajo), ubicaciÂ—n de lek, y presencia de hembras en la variaciÂ—n de caracterÂ’sticas de danza. Las danzas de voltereta lateral ocurrieron sÂ—lo cuando la hembra estaba en la misma rama que los machos , y las danzas de arriba abajo ocurrieron sÂ— lo cuando la hembra estaba en una rama cerca na a los machos o cuando estaba ausente. DescubrÂ’ que las caracterÂ’sticas de danza no variaron con la ubicaciÂ—n de lek , pero la coordinaciÂ—n entre macho s sÂ’ variÂ— con la u bicaciÂ—n de lek. Adicionalmente, l a frecuencia de saltos de la danza tipo arriba abajo variaron con la presencia de hembras , con machos saltando mÂ‡s rÂ‡pido cuando en la presencia de hembras cerca de la percha de danza . La selecciÂ—n sexual por parte de las hembras parece ser un fuerte determinante de las caracterÂ’sticas de danza , promoviendo poca variaciÂ—n en
Dancing Behavior in Long tailed Manakins Finger 2 las caracterÂ’sticas de la danza . Sin embargo, la coordinaciÂ—n y velocidad de salto sÂ’ presentaron variaciÂ—n , mostrando que la ubicaciÂ—n de lek y presen cia de hembras pueden afectar el comportamiento de danza en los toledos. Sexual selection has yielded some of the most unique s pecies on E arth, from birds of paradise to bighorn sheep to longhorn beetles . While natural selection mainly involves survival rates for individuals based on phenotypes, sexual selection involves mating success, where individuals of one sex select mates from the other sex ( Campbell and Reece , 2005 ). Thus, individuals with more attr active characteristics are perceived as suitable mates and have better reproductive success. This sexual selection leads to competition between individuals , which includes intrasexual selection , where individuals of one sex compete, and intersexual selecti on , where individuals of one sex select individuals of the other (Campbell and Reece, 2005). Intrasexual selection includes elephant seal harem domination and abnormally large testes in birds, and intersexual selection includes bird of paradise ornaments a nd female cryptic choice in beetles (Brennan, 2010) . In pre mating intersexual competition, females select their mates based on sexually selected traits such as ornaments and mating displays . The males of some species have evolved traits like anole lizard dewlaps and peacock tails in order to attract females ( Campbell and Reece, 2005 ) . One extreme example of pre mating intra sexual selection is seen in lekking species, which congregate at one location to perform mating displays. A lek is a site where males congregate to perform displays with other males and where females choose which males to mate with (Endler and ThÂŽry, 1996). Leks occur in certain bird, anuran, mammal, and insect species, and males must balance energy expenditure in pursuit of a mate and s urvival needs. In bird species, lek locations remain relatively constant over time, as males return to the same leks every year (Stiles and Wolf, 1979). Individual males benefit from displaying in a lek , as females seem to prefer males that display in grou ps, rather than individually (Stiles and Wolf, 1979). Long tailed manakins ( Chiroxiphia linearis ) exhibit one of the more distinct examples of sexual selection in nature. Females select males with distinct colors, calls, and mating behaviors, such as dance s. While f emale manakins are mainly olive colored, males are black colored with a bright red crown, light blue back, and two long central re c t rices that together form a long , narrow tail (Kanaksi et al., 2012 ) . Male long tailed manakins have at least thirt een different calls, which are used for various functions, including distress, female attraction, and dances (Maynard, 2012). Males in a lek have a very unique cooperative dance , where an alpha and beta male are paired together, often staying paired for ye ars, and usually only the alpha male will mate with a female ( McDonald, 1993 ). In a dancing bout, two males together jump straight up and down while facing away from the lekking branch in an Up Down v ariant or jump over one another while facing forward on the branch in a Cartwheel v ariant ( Slud, 1957 ; Kanaksi et al., 2012). These unique sexually selective characteristics make the long tailed manakins ideal for a study on lekking behavior . While many researchers have studied long tai led manakins, the objectives of their studies have varied. Many studies in the Monteverde area have researched calls and cooperation (Trainer and McDonald, 1993; McDonald, 2000) , but fewer research exists that analyzed dancing behavior ( Slud, 1957 ) . A s May nard (2012) noted , further research should study the variation in long tailed manakin behaviors in different lek location s , where lek location refers to leks found nearby each other . Furthermore, McDonald (1989b) found that female courtship behavior
Dancing Behavior in Long tailed Manakins Finger 3 strong ly influences male behavior. Thus, I investigate d the variation between Up Down and Cartwheel variant dances and address ed how lek location and female presence impact these differences. In doing so, I attempt ed to answer the following research question s : How does dancing behavior in male long tailed manakins vary between different lek locations? How does female presence affect dancing behavior? MATERIALS AND METHODS Study Site I studied long tailed manakins in Monteverde, Costa Rica, a community in the TilarÂ‡n mountain range at an elevation of about 1,400 meters. I chose study sites in the forest reserves of Santuario Ecologico (SE) , Crandell Reserve (CR) and Bajo del Tigre (Bd T ), all of which are composed o f premontane tropical moist forest. Data Collection For my study, I visit ed the reserves between the hours of 6:30am and 2:00pm , because I was able to observe or listen to dances in progress during these times and furthermore because Endler and ThÂŽry (1996) found that manakins are often more active in the morning hours due to optimal light environments . From 13 May 2018 to 25 May 2018, I found and studied five different leks: three in Santuario Ecologico, one in Bajo del Tigre, and one in Crandell Reserve . After identifying leks, I set up the Canon XA 10 Professional Camcorder roughly five to ten m eters away from a lek and placed a camouflage poncho over the camera in order to hide it. I then let the camera film while I maintained a gre ater distance to note beginnings and endings of dances . Each day, I filmed a lek for three to five hours, and I returned to each lek once or twice . Thus, I filmed every lek for at least six hours. Video Analysis I used video analysis to study the followin g dance factors: bout length, time interval between jumps , male coordination, and jump rate . I also classified lek location and female presence in order to test their effects on the dance features. I refer red to the individual dance components as dance factors, and the combination of these four factors as dance features . I considered each series of jumps as a bout and a collection of bouts as a dance. A bout ended when the males stop jumping or changed dance variant (from Cartwheel to Up Down or vice versa). I classified bouts as Cartwheel variants if the males faced forward and jumped over one another, and as Up Down variants if they faced away from the branch and jumped up and down (Kanaksi et al., 2012). To study b out length, I recorded the total time of each dancing bout . I noted time int ervals between jumps by recording the differences in time between each jump landing. For male coordination, I used the standard deviation of the time interval between jumps, since a lower deviation implied less variation between jumps and thus more coordination. I found jump rate by calculating the number of jumps per minute. I investigate d female presence by noting wh ether females were found on the lekking branch, nearby, or not pr esent . Finally, for lek location I noted the lek where each bout took place . I separated Cartwheel and Up Down variants for the statistical analysis because the two variants were inherently different (Slud, 1957). Statistical Analysis
Dancing Behavior in Long tailed Manakins Finger 4 I used R to examine the impact that female presence and lek location had on dance features (R Core Team, 2013) . In order to do so, I used the dpylr package in R to ru n m ultivariate analyses of variance (MANOVAs), which tested whether one variable (female presence or lek locat ion) had a significant effect on a group of variables (Wickman, 2018). I used the dance features for this group of variables in order to study the impacts of female presence and lek location on the entirety of Cartwheel and Up Down bouts . The MANOVAs also calculated analyses of variance (ANOVAs) that tested the impact of one variable (female presence or lek location) on each indiv idual dance factor . I used these ANOVAs to examine which individual dance factors were most impacted by lek lo cation and female presence. I then used the ggplot2 package in R to graph the results (Wickman, 2009). RESULTS I obtained 13.4 hours of video footage of manakins dancing. In th is time , I recorded 15 dances , composed of 71 bouts and 913 jumps, at 5 diffe rent leks . Of the 71 dances, 37 were Cartwheel variants and 34 were Up Down v ariants. I found tha t all of the 37 Cartwheel variants occur red when a female was present on the branc h where males danced, while all 34 Up Down variants occur red when the female was either on a nearby branch or completely absent from the lekking area. Table 1 shows that C artwheel variant s had longer jump intervals and a lower standard deviation for jump intervals, indicating that these jumps took longer but varied less in the time it took to complete jumps . Thus, because there was less variation from one jump to another, the Cartwheel variants exhibited more male coordination in dancing . While both variants show ed similar values for jumps per minute , Up Down v ariants were on aver ag e much shorter than Cartwheel variants . Lek Location I found that there was little variability between individual leks for Cartwheel variants. Lek location did not have a significant impact on Cartwheel dance features (F 2, 31 = 1.57 8 ) , and thus lek locat ion did not contribute to differences that exist ed between bouts . For Up Down bouts , I examine d the impact of both lek location and female presence on dance features . I found that Up Down variants did not show large amounts of variation in dance features between leks (F 4,27 = 0.93 1 ) . In analyzing the individual dance factors , though, I discovered that two leks had median standard deviations for jump intervals that were different from and greater than those of the other leks (F 4,27 = 3.033 , p = 0.035 ) . As shown in Figure 1, t his increased standard deviation in leks BdT 3 an SE 1 shows that there was more variation from one jump to the next, and thus males were less coordinated than in other leks . Furthermore, SE 1 had a much greater range of coordination th an did any of the other leks , with a standard deviation range of nearly 0.4 . Th us , its males had a more variable level of coordination than did other males. Female Presence Males changed their Up Down dance features based on whether a female was present on a nearby branch or absent from the lek area (F 1,30 = 2.8 17, p = 0.04 5) . W hile average jump interval, average bout length, and male coordination did not show statistically significant relationships with female presence , nearby female presence ca used males to increase their average jumps per minute ( F 1,30 = 6.2 80 , p = 0.01 8 ). As Figure 2 illustrates, jump rates increased
Dancing Behavior in Long tailed Manakins Finger 5 in the presence of a female and decreased with female absence . I could not test female presence for Cartwheel bouts because fema les were only found on the lekking branch, and thus there was no variation in female presence. Table 1 . Comparison of jump factors for Up Down and Cartwheel variants. Standard deviation for average jumps per minute and average bout length shown with Â± sign. Standard deviation for jump interval listed in separate column because treated as a separate dance factor: male coordination. Dance Variant Average Jump Interval (s) Jump Interval Std Dev Average Jumps per Minute Average Bout Length (s) Cartwheel 0.875 0. 150 68.586 Â± 5.950 26.308 Â± 14.826 Up Down 0.845 0.204 68.636 Â± 6.745 9.702 Â± 7.231 Figure 1 . Effect of lek location on male coordination in Up Down bouts. Increased jump interval standard deviation shows decreased male coordination. Five leks (Bajo del Tigre 1, Crandell Reserve 1, and Santuario Ecologico 1, 2, and 4) used to compare coordination.
Dancing Behavior in Long tailed Manakins Finger 6 DISCUSSION Up Down variants had shorter jumps and greater standard deviation s in jump times , thus exhibiting less male coordination than did Cartwheel variants (Table 1) . Slud (1957) described that Up Down jumps often quicken as the bout advances, providing an explanation for the fast er jumps and greater variance. The Cartwheel v ariant directional ity , where both males face the same d irection with one behind the other, allows the males to face the females while dancing, possibly explain ing why the variant was observed only when females we re on the same branch as the males . In the Up Down v ariant, both males face away from the branch in the same direction, possibly allowing them to face a female found on a nearby branch (Slud, 1957) . Female behavior strongly affects male responses in long tailed manakins, and thus this directional singularity is likely a response to female selection for a particular dance (McDonald, 1989b). The singularity in direction when females are not present is somewhat surprising, though, as the Up Down v ariant does not necessarily allow males to face the females. One possibility is that females were present that I did not see, leading me to mistakenly classify the female as being absent. Another possibility is that since courtship performance is a strong determinant of female mate choice , performi ng Cartwheel variants rather than Up Down variants when a female is not present c ould lead to a decreased chance of mating (McDonald, 1989b) . I found that the differen ce in lek location was not the largest source of bout variability for Cartwheel dance features . While lek location might have had some influence on Cartwheel dance features , other factors likely had greater impacts . Endler and ThÂŽry (1996) wro te that light environments had a strong impact on bird dances, and this along with female presence might have had a stronger effect on Cartwheel dance features than did lek location . Gibson and Bradbury (1985) studied mating success in the lekking sage grouse, finding that success was largely determined by display performance. While their study measured display performance as strut rate and acoustic components , long tailed manakins might utilize similar mate selection Figure 2 . Effect of female presence on average jumps per minute in Up Down bouts.
Dancing Behavior in Long tailed Manakins Finger 7 based on the performance of Cartwheel dance features . Thus, females might select for a particular execution in Cartwheel variants, and leks that show greater variation in the execution would have a decreased chance of mating success. My results similarly point to a lack of variation for Up Down variants in different locations . S tandard deviation of jump intervals , however, did vary by lek location . D ecreased jump interval standard deviation represented less variation between jumps and thus more male jumping coordination . This difference in standard devia tion thus implies that male coordination varies between leks. Another point of note is that coordination seemed to depend on lek location in Up Down bouts , but not in Cartwheel bouts . Pruett Jones and Pruett Jones (1990) studied the lekking bird of paradis e Lawes ' Parotia, finding that six behavioral characteristics affected mating success. T hese characteristics ranged in their likelihood of le a d ing to copulation from 0% to 72.9%. I t is similarly possible that Cartwheel dance features were much more impactf ul of mating success in long tailed manakins, and thus variation in Up Down dance features would not affect mating success. Another possible explanation is that the two variants are inherently different from one another and thus might not reflect the same characteristics. F or example, Slud (1957) observed that Up Down bouts get faster with progression, which could mean that males performing Up Down variants need to be more coordinated than those performing Cartwheels. This changing speed would be reflected in interval standard deviation, with increased deviation for a more changing speed. I also found that Up Down variants change d with and depend ed on female presence . Because average jumps per minute increase d with fe male presence and decrease d when no females were present, it app ears that female presence caused males to jump more quickly. Here again, the observation that Up Down bouts speed up with progression can help to explain the trend (Slud, 1957) . Dances with a female nearby, and thus higher p ossibility of success , might continue to get faster, while those without females are more likely practicing the dance ( Kanaksi et al., 2012; McDonald, 1989 a ) . Alternatively, Gibson and Bradbury (1985) found that the sage gro use's display performance in leks is affected by energetic factors, as males lost weight through the breeding season and did not appear at leks following cold weather spells. Thus, it is possible the long tailed manakins are affected by similar energetic f actors and dance at slower rates when females are not pres ent in order to conserve energy for future bouts. While I did not find that differences in lek location affect the bouts occurring there, I did fi nd that location can play a role in affecting individual aspects of bouts . Thus, while certain aspects of manakin bouts might be shared across locations, others might be unique to each lek . Furthermore, I quantified differences in Up Down and Cartwheel variants an d showed how these variants can be in fluenced by female presence. These variations in sexual selection show that female actions can affect male reactions, and that display performance may vary between locations (McDonald, 1989 b; Gibson and Bradbury, 1985 ) . Studies in the future should include a greater variation in manakin habitats and elevations in order to lend greater clarity to the impact of lek location on dances. In addition, f urther research should incorporate other fact ors, such as pre dance calls, light penetration , and male flight, in order to better understand whether bouts differ by lek . Future questions might also address what determines mating success , how dances differ with progression, and how females respond to each bout and dance. ACKNOWLEDGEMENTS I would like to thank AndrÂŽs Camacho for all of his aid in finding manakins, guiding the project, and running statistical tests. Thanks also to the entire EAP staff, especially SofÂ’a Arce
Dancing Behavior in Long tailed Manakins Finger 8 Flores for her help and feedback t hroughout the project and Frank Joyce for generously lending me his camera and tripod. I want to thank Rafa and Kelsey from Bajo del Tigre for their assistance in finding leks, and FÂŽlix, Cristofer, Andrea, and AndrÂŽs from the Santuario Ecol o gico for their help, hospita lity, coffee, and musical instruments. Thanks to Sara Gibbons for her peer review and invaluable advice in how to improve this paper. Also, shout out to Ashley Anderson and Noah Bennett for their help in field work and ability to withstand long hours witho ut seeing birds. Finally, thanks to Allie Howell for her help in trying to understand statistical analyses and R , and Griffin Nicholson for his emotional support on the trails.
Dancing Behavior in Long tailed Manakins Finger 9 LITERATURE CITED Brennan, P. 2010. Sexual Selection. Nature Education Knowle dge, Vol. 3 (10): 79 . Campbell, N.A., and J.B. Reece. 2005 . Sexual Selection. Biology , Vol. 7 : 468. Endler, J.A., and M. ThÂŽry. 1996. Interacting effects of lek placement, display behavior, ambient light, and color patterns in three neotropical forest dwel ling birds . The American Naturalist, Vol. 148 (3): 421 452. Gibson, R.M., and J.W. Bradbury. 1985. Sexual Selection in Lekking Sage Grouse: Phenotypic Correlates of Male Mating Success . Behavioral Ecology and Sociology , Vol. 18 (2) : 117 123. Kanaksi, A., C. Stuligr oss, J. I. Pareja, and W. Tori. 2012 . Long tailed Manakin (Chiroxiphia linearis). Neotropical Birds Online. Maynard, D.F. 2012. The vocal behaviour of long tailed manakins (Chiroxiphia linearis): The role of vocalizations in mate attraction and m ale male interactions. University of Windsor Elec tronic Theses and Dissertations . McDonald, D. B. 1989 a . Cooperation under sexual selection: age graded changes in a lekking bird. American Naturalist, Vol. 134: 709 Ã 730. McDonald, D.B. 1989b. Correlates of m ale mating success in a lekking bird with male male cooperation . Animal Behavior, Vol. 37 (6): 1007 1022. McDonald, D. B. 1993. Demographic consequences of sexual selection in a Long Tailed Manakin. Behavioral Ecology , Vol. 4: 297 309. McDonald, D.B. 2000. Cooperation between long tailed Manakins. Monteverde: Ecology and Conservation of a Tropical Cloud Forest, 204 205. Pruett Jones, S.G., and M.A. Pruett Jones. Sexual Selection through Female Choice in Lawes' Parotia, a Lek making Bird of Paradise. Evoluti on, Vol. 44 (3): 486 501. R Core Team. 2013 . R: A language and environment for statistical computing. R Found ation for Statistical Computing . Slud, P. 1957. The Song and Dance of the Long Tailed Manakin, Chiroxiphia linearis . The Auk, Vol. 74 (3): 333 339 . Stiles, F.G., and L.L. Wolf. 1979. Ecology and Evolution of Lek Mating Behavior in the Long Tailed Hermit Hummingbird . Ornithological Monographs , Vol. 27: 1 72. Trainer, J.M. and D.B. McDonald. 1993. Vocal Repertoire of the Long Tailed Manakin and Its Rel ation to Male Male Cooperation . The Condor, Vol. 95 (4): 769 781. Wickham, H. 2009. ggplot2: Elegant Graphics for Data Analysis. Springer Verlag New York. Wickman, H., R . FranÂois, L . Henry , and K . MÂŸller . 2018. dplyr: A Grammar of Data Manipulation. R pac kage version 0.7.5.