Effect of m ass and s ex on b at s distress calls Moreland 1 1 Effects of m ass and s ex on d istress c alls of bats in Monteverd e Ana Moreland Department of Ecology and Evolutionary Biology U niversity of California, Santa Cruz EAP Tropical Biology and Conservation Spring 2019 10 June 2019 ABSTRACT Animal behavior is st rongly influenced by the threat of predation. This has led to the development of complex communication systems be tween predators and prey. In bats, many types of ultrasonic calls have been studied in depth, but low frequency distress calls have rarely been examined. In order to analyze which variables might be able to predict whether a given bat performs a distress c all, I captured 36 bats and recorded their behavior and the decibel power of their calls. I found that bats of a smaller mass were more likely to engage in distress calls while bats with a greater forearm length emitted calls with greater decibel power, in dicating greater energy and range. However, these findings may be biased by which species I captured. I also found that nursing bats were less likely to call than non reproductive males and females , implying some unknown risk associated with distress calli ng that nursing bats needed to avoid. My findings suggest that while smaller bats may be more likely to call, the calls of larger bats are more acoustically powerful and therefore would likely be more effective at attracting bats from long distances . There is likely an unknown cost associated with distress calling in nursing bats that merits further study. Efectos de l t amaÃ±o y s exo en las l lama das de s ocorro de m urci Ã© lagos e n Monteverde RESUMEN La depredaciÃ³n afecta el comportamiento y la evoluciÃ³n de l a s especie s . E xisten sistemas complicados de comunicaciÃ³n entre depred ad ores y presa s . Los murciÃ©lagos emiten sonidos ultrasÃ³nicos que se h an investigado durante mÃ¡s tiempo , pero sus sonidos audibles a humanos y otros animales han s ido mucho menos estudiados . Estos ruidos de baja frecuencia incluyen las llamadas de socorro. A nali cÃ© 36 murciÃ©lagos capturados en redes de niebla, evaluando cu Ã¡ le s caracterÃsticas pueden predecir si un murciÃ©lago va a emitir llamada s de socorro. EncontrÃ© que los murciÃ©lagos de tamaÃ±o pequeÃ±o hacen mÃ¡s llamadas de socorro que los murciÃ©lagos mÃ¡s grandes, y tambiÃ©n que los murciÃ©lagos con alas mÃ¡s largas emit en sonid os mÃ¡s poderosos. E ncontrÃ© que las hembras en periodo de lactancia emit en menos llamadas. H ay algÃºn factor negativo en emitir llamadas de socorro , siendo posible que tenga que ver con l a energÃa que cuesta emitir estos sonidos mientras el murciÃ©lago estÃ¡ d ando de mamar , que es tambiÃ©n muy caro energÃ©ticamente . P uede ser que las llamadas de socorro , indiquen algo mÃ¡s , ademÃ¡s del peligro a los otros murciÃ©lagos . Pienso que e x isten costo s aÃºn no de s cubiertos de las llamadas de soc orro que merece n mucho mÃ¡s est udio .
Effect of m ass and s ex on b at s distress calls Moreland 2 2 One of the driving factors of the natural world is predator prey interaction. Predation pressure can cause the evolution of toxins, spines, warning coloration, and so much more ( Lima, 2002 ). In the evolutionary arms race between predator and prey, communication can play a vital role through behavioral signaling or colora tion. A prey item communicating that it is a bad choice, because of its toxins or fitness or other antipredator defense, can be impo rtant in d etermining survival and behavior of bot h species involved (Caro, 2005) . There are 1 , 240 species of bats belonging to the order Chiroptera and a large diversity of natural history traits associated with them. Bats can feed on nectar, fruit, insects, or even blood; can range in size from 2 grams to 1,200; and can be highly social or completely solitary. The vast majority of these bats are eat en by birds of prey and other animals, and some have evolved different antipredator behavior ( Lima et al., 2013 ). While extensive work has been done on ultra so nic communication between bats , I believe that a scientifically neglected example of antipredator behavior is audible (not ultrasonic) calling or screeching. Audible distress calls have not been the focus of very much study in bats, but ornithologists h ave made fascinating and relevant discoveries about their role in bird predator prey interactions. Laiolo et al. ( 2004 ) found that short toed larks in good body condition used anti predator calls when caught that ranged across more frequencies tha n birds i n poor body condition. For birds, the calls function as honest signal s . There has not been an effort to find out the potential function of these calls in bats. This is a particularly interesting topic because not all bat species are known to be equally noi sy, suggesting different selective pressures acting on different speci es. In bats, distress calls have been found to induce mobbing behavior (Caro, 2005). Mobbing behavior consists of many bats being attracted to the calling bat and circling it, occasiona lly diving at the predator. This has been found among unrelated indiv iduals and across many species, including some found in Costa Rica such as Dermanura spp. and other phyllostomid ae ( Russ et al., 1998; August, 1979). However, there has not yet been a stud y into which bats use low frequency (audible to humans) distress call s and which characteristic s might be able to predict these calls. In this paper, I seek to analyze a set of variables to see if there is an effect of sex, age, reproductive status, mass, species, and forearm length on likelihood to call for a given bat . I believe that this will elucidate some of the selective pressures that could lead to the rise of audible defensive calling. Determining which of these variables predicts defensive calling can uncover much about the function and perhaps evolutionary origins of defensive calling in bats. MATERIALS AND METHODS To address this question, I set up mist nets to capture many species of bats over a pe riod of two weeks (12 May, 2019 through 30 M ay, 2019) . I had access to five nets of different
Effect of m ass and s ex on b at s distress calls Moreland 3 3 lengths. Mist netting occurred at the Monteverde Institute and in the Bajo del Tigre neighborhood ( 10.3052704 , 84.8096583 and 10.3056389 , 84.8144494 respectively) . Once captured in the net , I recorded th e encounter with the bat to see if it made audible calls during untangling and processing . To do this, I used a Zoom HN4 audio recorder. Then, I ran the audio recording through the program Raven Lite developed by the Cornell Ornithology Lab. I analyzed the power density. Power density indicates the amount of energy in watts needed to produce this sound. I analyzed simply whether or not a bat called and compared this to v ariables like se x, mass , species, and reproductive status using chi square tests . I separated sex into male, female, pregnant female, and nursing female for analysis. I determined that a female was pregnant by palpating the stomach and that she was nursing by noting expos ed teats and a lack of pregnancy. RESULTS Overall, we caught and obtained distress call data for 34 bats, 20 of which called and 14 of which did not. We caught a total of 12 species, but the vast majority (16) were Myotis pilosatibialis ( T able 1). This small species called 15 out of 16 times. Other species frequently caught included Dermanura tolteca (6) and 3 species of the Ca rollia genus (7). I did not have sufficient data points to draw conclusions about whether species affected likelihood for distr ess calls , with the exception of M. pilosatibialis , which called 93.75% of the time (N=16). Typically , we found about 3 bats of each species except M. pilosatibialis , which was not enough to run accurate data analysis. T he variables that were able to pred ict whether a bat emitted a distress call were mass and sex separated by reproductive status. Overall, sm aller bats called more than larger bats (fig. 1, chi square p=0.0315, N=34). Mass (which indicates both size and body condition) was a better predictor of call likelihood than forearm length (which indicates purely size and was insignificant). Additionall y, the single Desmodus rotundus we caught did call and had a mass of 41 g. If I excluded D. rotundus from analysis, the correlation between mass and li kelihood of calling was even more robust (p=0.0024). It would be useful to capture more of this species and other very large bats (>35 g) to determine if calls are actually common in the species or in very large bats, or if this individual was an outlier. When we eliminate these bats, the trend is much weaker (fig. 6 ). Sex w as also able to predict calling . I chose t o separate female into three reproductive categories : (non reproductive) female, pregnant female, and nursing female. I did this because I wanted to see if reproductive stakes predator behavior. I did indeed find that males and non reproductive females were similarly likely to call, while nursing females in particular stayed
Effect of m ass and s ex on b at s distress calls Moreland 4 4 silent (fig. 2 , p=0.0326). Pregnant females seemed to call more than nursing females, however, I noticed that 2/3 of the pregnant females who called wer e M. pilosatibialis , a species that called 93.75% of the time . In contrast, none of the nursing females were M. pilosatibialis . To test whether species was a source of error here, I analyzed sex versus calling again, this time removing M. pilosatibialis f rom the data (fig. 3 ). I found that the trend of nurs ing bats calling at lower rates still exists when this species is not considered. When I looked at average power density, I found an association between forearm length and power (fig. 4). The association appeared strong before isolating just M. pilosatibialis to get a better idea of whether forearm strength had an impact within species (fig. 5). Thus, I found that species may have more to do with power density than forearm length itself. Table 1. Species of bats caught ( sex, measurements ) with records o f their distress calls and mean power density of these calls. Species name Sex Forearm length (mm) Mass (g) Distress call Average Power Density (dB) Carollia perscpicillata Male 42 19 No Myotis pilosatibialis Male 35 6.5 Yes Sturnira hondurensis Nurs ing Female 44 20.5 No Carollia sowelli Pregnant Female 40 23 Yes Carollia nicarague n sis Pregnant Female 44 22.75 No Micronycteris schmidtorum Male 35 8 Yes Myotis pilosatibialis Female 35 7.5 No Dermanura tolteca Nursing Female 42 19 No Sturnir a hondurensis Nursing Female 44 20 No Myotis pilosatibialis Male 37 4.5 Yes Myotis pilosatibialis Female 36.5 6 Yes Dermanura tolteca Nursing Female 41 16.5 No Carollia perscpicillata Male 40 19 No Carollia perscpicillata Male 43 18 Yes Myotis pilosatibialis Male 35 6.5 Yes Dermanura tolteca Nursing Female 42 16.5 No Myotis pilosatibialis Male 36 6.5 Yes Myotis pilosatibialis Male 36.5 5.5 Yes Myotis pilosatibialis Pregnant Female 36 7 Yes 39.1 Carollia sowelli Nursing Female 41 19 Yes 43.5
Effect of m ass and s ex on b at s distress calls Moreland 5 5 Sturnira hondurensis Nursing Female 43 23 No Carollia sawelli Male 40 17 Yes Myotis pilo satibialis Pregnant Female 36 6.5 Yes 40.8 Myotis pilosatibialis Male 36 5.5 Yes 41.1 Myotis pilosatibialis Male 35 3 Yes 40.9 Glossophaga soricina Ma le 37 11 No Carollia sowelli Male 43 20 No Myotis pilosatibialis Female 37 6 Yes Dermanura tolteca Pregnant Female 41 17 No Myotis pilosatibialis Male 37 7 Yes 39.8 Desmodus rotundus Male 60 41 Yes 23.6 Dermanura tolteca Male 44 18 No Platyrr hinus helleri Male 38 16 Yes Dermanura tolteca Male 41 14 Yes 40.6 Fig. 1. Smaller bats called significantly less than larger bats ( chi square, p=0.0315). Bats that did call showed a greater range, including one outlier at 41 grams.
Effect of m ass and s ex on b at s distress calls Moreland 6 6 Fig. 2. Nu rsing females were significantly less likely to call than males and females of diff erent reproductive stages ( chi square, p=0.0253, N=35). Males were the most likely to call of any sex class. The average line allows us to better see how female, male, and p regnant female show Fig. 3 When Myotis is removed, we still see the same trend of nursing females calling less (chi square, p=0.2028, N=19). The data are not statistically significant; however, this is likely impacted by the low sample size of pregnant females . The quantity of males calling has gone
Effect of m ass and s ex on b at s distress calls Moreland 7 7 down without Myotis but remains much higher than females calling. Non reproductive female is not included in the graph because all non reproductive females were Myotis . Fig. 4 . Average power density appears to increase with forearm length. As bats get larger, the amount of power put into their calls seems to increase , meaning they are more energetically expensive. Red poi nts indicate bats from the family Vespertilionidae while blue indicates the family Phyllostomatidae. There is only one Vespertilionidae species that we caught: M. pilosatibialis . There is a very high likelihood that this graph is impacted by the outlier D. rotundus , seen at the top right corner of the graph.
Effect of m ass and s ex on b at s distress calls Moreland 8 8 Fig. 5 Here we see power density versus forearm length isolated for just one species , M. pilosatibialis . The trend is very weak if it exists a t all (R 2 =0.206). Sexes are denoted with colors, and ther e seems to be no pattern there either. Fig. 6 When we remove M. pilosatibialis from the dataset, the trend is much less clear (chi square, p=0.8492). There does not seem to be much of relationship between distress call and mass of bat when we eliminate the overrepresented Myotis .
Effect of m ass and s ex on b at s distress calls Moreland 9 9 DISCUSSION Mass initially seemed to be a reliable predictor of distress call likelihood. Overall it is true that smaller the effects of mass are clouded by the effects of the species M. pilosatibialis , or whether the perceived effects of the species M. pilosatibialis on distress call likelihood are actually the effects of mass. The individuals we caught of M. p i losatibialis ranged in size from 3 t o 8 grams and were the smallest species by far. Because there was only one species in this size range, it is difficult to separate the effects of mass from potential species specific effects. I attempted to do this by analyzing mas s versus distress calls with M. p i losatibialis eliminated. Once I did this, I did indeed find that mass had a much, much weaker association with distress calling. This is very interesting because it points to species itself being a factor in whether an ind ividual chooses to call. Unfortunately, I did not have the dataset necessary to analyze species, but this is an essential next step for bat distress call research. My study suggests that there is something about the species M. pilosatibialis beyond obvious variables that is causi ng them to call at disproportionately high rates. A future study that involved high sample sizes across many species could get a more accurate answer for the question of which species call the most. Abiotic weather factors were di sregarded as potential so urces of variation because bats could not be caught in the rain due to heightened mist net visibility , and thus the weather was very similar on most nights that we were able to catch bats . Sex was also an adequate predictor of dis tress call likelihood. Ma les and non reproductive females were found to be the most likely to call (fig. 2 , p=0.0253 ). However, nursing and, to a lesser extent, pregnant females called much less. It is possible that this is due to a decreased desire for ri sk taking in females with dependent young. In order to understand why females, call less when reproductively engaged, we must first understand the purpose of the distress call. Lima et al. write about the function of distress calls and claim that the call can either act as a pers onal risk to warn other bats, or as a way to attract other bats to mob a predator (2013). Each of these interpretations has different implications for the lack of calling in nursing bats. If distress calls act as a somewhat altruis tic warning to other bats at personal cost, it stands to reason that a nursing bat, whose fitness depends now on the survival of her dependent young, would not want to make that sacrifice for her roost mates. However, if the second interpretation is corre ct, then it does not make sense that such a vulnerable demographic would decline to call for help. In this case, my data may be skewed by the season in which my study was conducted. Of the pregnant females we caught, half were M. pilosatibialis and the oth er half were of the Carol lia genus. LaVal and Fitch found temporal differences in bat reproductive cycles indicating that M. pilosatibialis should not yet be giving birth, which is supported by my data ( LaVal and Fitch, 197 7) . No M. pilosatibialis and only one Carollia was nursing , which could impact data since M. pilosatibialis was found to be noisier as
Effect of m ass and s ex on b at s distress calls Moreland 10 10 a species . Other species such as Dermanura tolteca and the Sturnira genus were only captured while nursing, not pregnant. This coul d create a false pattern if D. tolteca and Sturn ira are simply less likely to make noise, not necessarily because nursing bats as a class make less noise. To test this, I analyzed sex data after removing all M. pilosatibialis to see if the nursing pattern remained. While the results were less robust, I still found a disproportionate refusal to make distress calls by nursing bats (chi square, p=0.2028). This indicates that the pattern is real and not being created by the presence of M. pilosatibialis , as was the case with my mass data. When it comes to the bats that did make calls, I found that bats with longer forearms emit more powerful sounds (R 2 =0.827). T his is a more reliable measure of size than mass because it does not take food or pregnancy into account. In particular, the biggest bat that called was a D. rotundus , or common vampire bat. We observed that the bat had fed recently and was bloated with bl ood, meaning that its mass was much higher than it would be even a few hours later. However, when I isolated M. pilosatibialis their power to each forearm length, there was not a strong relationship at all (R 2 =0.206). This is not conc lusive but means that we need more data to confirm this fact, and that a single point by D. rotundus may be disproportionately represented. We also possible that with more points this relationship could change. The idea that larger bats seem to invest more energy in their sounds, knowing that producing sound is energetically expensive, sugges ts that there is a benefit to producing the most powerful sound possible in the face of a p redator. In the contex t of predator prey interactions, this means that sound power is a worthwhile expense and serves an anti predator function. What exactly the ca lls are conveying is more difficult to understand. Fenton et al. found that in Myotis lucifigus , closely re lated to M. pilosatibialis , bats responded to distress calls of conspecifics by flying more actively and even circling a speaker (1976). They were a ble to conclude that distress calls specifically attract other bats in this species. Even heterospecifics a re capable of being attracted by acoustically similar calls (Huang et al., 2018). This leads me to believe that larger bats, who are capable of produ cing more powerful sounds, have their distress calls heard and responded to from farther away. Combined with the results we saw comparing ma ss and likelihood of calling, I propose that smaller bats could be more likely to call because of their heightened risk for predation and that there is some hidden cost shown by the lack of calling in nursing bats. More study should be conducted to draw more robust conclusions about the effect of sex on call likelihood and qua lity. Low frequency distress calls in bats provide us with a novel way to study the relationship between predator and prey in an animal that is usually hidden by darkness. The decision to call or not to call is not as straightforward as it might seem, and by isolating mass and sex as key variables we can begin to understand the antipredator be havior of these highly social mammals. Future work might include analysis of a much larger sample size of bats across a longer period of time. This could help clarif y the effects of phylogeny and species on l ikelihood to call .
Effect of m ass and s ex on b at s distress calls Moreland 11 11 With a large enough sample size, and captive bats, we could even study whether individuals have personalities that make them more likely to call. Additionally, future work should include the pla yback of low frequency distress calls for o ther bats to see how they react. ACKNOWLEDGEMENTS A huge thanks to my bat team, Federico Chinchilla and Brianna Kennedy. Thank you to Fede for his endless patience and for showing us how to use mist nets to safely catch and release both bats and beetle s. Thank you to Richard LaVal for his help and expertise and thank you to Emilia Triana for her support. Finally, thank you to Frank Joyce for being flexible and letting me come on this great Costa Rican adventure in the first place. LITERATUR E CITED August, P. X. V. 1979. Distress calls in Artibeus jamaicensis : ecology and evolutionary implications. In: Vertebrate Ecology in the Northern Neotropics (Ed. by J. F. Eisenberg), pp. 151 160. Washington: Smithsonian Institute. Caro, T. 2005 . Antipredator d efenses in birds and mammals. Chicago, IL: University of Chicago Press. Fenton, M. B., Belwood, J. J., Fullard, J. H. & Kunz, T. H. 1976. Responses of Myotis lucifugus (Chiroptera: Vespertilionidae) to calls of conspecifics and to other sounds. Canadian J ournal of Zoology, 54, 1443 1448. Huang, X . , et al. 2018. Acoustic s imilarity e licits r esponses to h eterospecific d istress c alls in b ats (Mammalia: Chiroptera). Animal Behaviour , 146 : 143 154. Laiolo, P . , et al. 2004. Distress c alls m ay h onestly s ignal b ir d q uality to p redators. Proceedings of the Royal Society of London. Series B: Biological Sciences . 271 ( 6 ) . LaVal, R K , and H S Fitch. 19 77. Structure, m ovements and r eproduction in t hree Costa Rican b at c ommunities. Occasional Papers of the Museum of Natur al History, the University of Kansas. 69 : 1 27. Lima, S . L., and J . M. O'keefe. 2013 Do p redators i nfluence the b ehaviour of b ats? Biological Reviews , 88 ( 3 ): 626 644. Russ, J. M., Jones, G. & Racey, P. A. 1998. Intraspecific responses to distress calls of t he pipistrelle bat, Pipistrellus pipistrellus . Animal Behaviour . 55 : 705 713. Russ, J.m, et al. 200 4 . Interspecific r esponses to d istress call in b ats (Chiroptera: Vespertilionidae): a f unction for c onvergence in c all d esign? Animal Behavior . 67 ( 6 ): 1005 1 014.