Ecological Morphology Structures of Neotropical Bat Wing

© 2020 Academia Sinica, Taiwan Open Access Ecological Morphology of Neotropical Bat Wing Structures Dennis Castillo-Figueroa Pontificia Universidad Javeriana, Unidad de Ecología y Sistemática (UNESIS), Bogotá 11001000, Colombia. E-mail: [email protected] Received 5 May 2020 / Accepted 17 September 2020 / Published 20 November 2020 Communicated by Chi-Chien Kuo Morphology has a direct influence on animal fitness. Studies addressing the identification of patterns and variations across several guilds are fundamental in ecomorphological research. Wings are the core of ecological morphology in bats; nevertheless, individual bones and structures that support the wing, including metacarpals, phalanges and the length of digits, have rarely been the subject of comprehensive research when studying wing morphology. Here, I analyzed morphological variations of wing structures across 11 bat guilds and how individual bone structures are correlated to diet, foraging mode and habitat use. I obtained wing measurements from 1512 voucher specimens of 97 species. All the specimens analyzed came from the Mammalian Collection at the Museo Javeriano de Historia Natural of Pontificia Universidad Javeriana (MPUJ-MAMM) (Bogotá, Colombia). Positive correlations between size and the length of the third and fifth digit were detected. Bat guilds that capture their preys using aerial strategy in uncluttered habitats had longer third digits but short fifth digits compared to guilds that rely on gleaning strategy and forage in highly cluttered space. Although terminal phalanges were shown to be important structures for guild classification, metacarpals were strongly related to aerial foragers from uncluttered habitats because of their potential role in flight performance and ecological adaptations. Results show that habitat use, as well as foraging mode, are reflected in wing structures. Different wing traits to those evaluated in this study should be considered to better understand the ecological interactions, foraging strategy, wing adaptations, and flight performance in Neotropical bats. Key words: Chiroptera, Wing traits, Ecomorphological relationships, Guild, Flight pattern. BACKGROUND Morphological traits influence the performance— and therefore biological fitness—of animals (Arnold 1983). The interactions between morphology and ecology constitute the primary goal of ecomorphology (Rhodes 2002), a fundamental field for understanding the abilities of species to exploit resources. Within ecomorphological studies, a major focus has become the intersection between locomotory apparatus and characteristics of resource use, including diet, foraging behavior, and habitat use (Kalcounis and Brigham 1995; Giannini and Brenes 2001; Voigt and Holderied 2012). Species that share resources are grouped into guilds (Fauth et al. 1996), usually with high overlapping in morphological dimensions and, at the same time, show a high distinction between other guilds (Rhodes 2002). Analyzing morphological differences between guilds help researchers infer idiosyncratic features in their ecology and behavior (Kalcounis and Brigham 1995). Accordingly, the identification of morphological patterns and variations among several guilds becomes relevant when studying functional morphology in animals (Marinello and Bernard 2014). The Neotropics harbor one of the richest bat faunas in the world (López-Aguirre et al. 2018), accounting for about 400 species, nine families, and three superfamilies (Arita et al. 2014), and representing a very complex evolutionary history (Peixoto et al. 2014; López-Aguirre et al. 2018) and ecology (Meyer et al. 2008; GarcíaCitation: Castillo-Figueroa D. 2020. Ecological morphology of Neotropical bat wing structures. Zool Stud 59:60. doi:10.6620/ZS.2020.59-60. Zoological Studies 59:60 (2020) doi:10.6620/ZS.2020.59-60 1 © 2020 Academia Sinica, Taiwan García et al. 2014; Briones-Salas et al. 2019; CastilloFigueroa 2020). New world bats have an extraordinary morphological diversity related to food habits (Santana et al. 2012), foraging behaviors (Schnitzler et al. 2003), and flight styles (Norberg and Rayner 1987). Particularly, the development of wings has played a central role in the colonization of several ecological niches over bat evolution (Sears et al. 2006), promoting adaptive radiation (Cooper and Sears 2013). Due to this, wing morphology is the basis of ecomorphological correlations in bats (Norberg and Rayner 1987) and wing traits are, therefore, important predictors of resource use (Kalcounis and Brigham 1995). Traditionally, wing loading, aspect ratio, and wingtip shape index have been the most common measures to assess flight style and aerodynamic abilities of bats (Norberg and Rayner 1987; Thollesson and Norberg 1991; Saunders and Barclay 1992; Rhodes 2002; Marinello and Bernard 2014). Other approaches based on individual bones and structures that support the wing such as metacarpals and phalanges (Stockwell 2001; Castillo-Figueroa and Pérez-Torres 2018; Castillo-Figueroa 2018a) appear to be suitable variables for predicting aerodynamic performance in bats; however, these morphological traits have been largely neglected in ecomorphological studies, especially in New World bats. Indeed, there is little research on wing characterization based on metacarpal and phalanges structures for several Neotropical bat species, especially in the rare ones; filling this knowledge gap may yield new information that allow to elucidate more comprehensively the morphological patterns of bat wings. Importantly, wing measurements that reflect body size (forearm length), wing width (length of the fifth digit) and hand-wing length (length of the third digit) can correlate to ecological adaptations of bats (Findley et al. 1972; Dietz et al. 2006) and may be useful in disentangling morphological patterns across different bat guilds. Since Neotropical bats vary greatly in morphological features, a representatively large group is necessary to capture the higher variation both within and among guilds. In this paper, I analyzed the variations in wing structures across 11 bat guilds and correlated individual bone structures to diet, foraging mode and habitat use. To do this, I aimed to (1) quantitatively characterize wing digits of 97 New World bat species; (2) explore correlations between body size, wing width, and hand-wing length; (3) assess any differences in wing ratios among bat guilds; (4) determine whether the guilds corresponded to morphologically distinct groups based on individual bone structures (i.e., metacarpals and phalanges), thus identifying which characters best differentiate these groups. A concomitant goal of this study was to discuss the importance of wing structures in ecomorphology for each bat guild. MATERIALS AND METHODS Bat species I obtained wing measurements from 1512 bat vouchers in the Mammalian Collection at the Museo Javeriano de Historia Natural of Pontificia Universidad Javeriana (MPUJ-MAMM) (Bogotá, Colombia). All were dry specimens and corresponded to adults only. On average, the number of individuals measured from each species was 16 (ranging from 1 to 243). The specimens belonged to 97 New world bat species, 43 genera, and 7 families. I excluded bat specimens in a bad state of preservation (i.e., broken wings). Identification of each specimen was corroborated with the keys of Gardner (2007) and, for the genus Platyrrhinus, the classifications of Velazco (2005) and Velazco et al. (2010) were followed. The current nomenclature was revised from the Integrated Taxonomic Information System (ITIS 2020). Wing morphometry I measured 11 wing traits, including metacarpals and phalanges in the dorsal side of the right wing (Fig. 1). I also measured forearm length (FL) and the length of the third (LD3), fourth (LD4) and fifth digits (LD5) (Fig. 1). Measurements were taken with a digital caliper (Mitutoyo Calibrador Vernier Mod Cd6 -csx 150 Mm) with an accuracy of 0.01 mm to ensure high precision; all measurements were performed by the same researcher (DCF) to avoid researcher bias. Data analysis To characterize the morphological traits for each species, I calculated the statistical mean and standard deviation for the length of the digits and forearm. I also described frequencies for these traits using histogram plots. To explore the associations between similarity in wing morphology and ecological similarity in bats, I first classified bat species based on foraging mode, habitat use, and diet into 11 guilds following Kalko et al. (1996), Sampaio et al. (2003), Estrada-Villegas et al. (2010), and Aguirre et al. (2016). To explore correlations between body size, wing width, and hand-wing length, I made Spearman correlations among FL, LD3, and LD5 since assumptions of normally distributed residuals were not fulfilled. In this way, to assess the correlations between page 2 of 14Zoological Studies 59:60 (2020) © 2020 Academia Sinica, Taiwan body size and wing width, I correlated LD5 with FL. To examine the correlations between body size to hand-wing length I correlated LD3 with FL. Lastly, to evaluate the correlations between hand-wing length to wing width, I correlated LD3 with LD5 (Dietz et al. 2006). I compared wing digit ratios of LD3/ FA (bat size to hand-wing length), LD5/FA ratio (bat size to wing width), and LD3/LD5 (hand-wing length to wing width) among guilds. Since data followed a non-normal distribution (Kolmogorov-Smirnov test, D = 0.29, P < 0.0001), a permuted ANOVA (n = 99999 permutations) with post-hoc comparisons (Dunn’s method) was used to assess differences among guilds. Probabilities at 0.05 were reported as significant. With the aim to differentiate the guilds in the morphospace, I conducted an ordination analysis using the 11 individual bone structures (i.e., metacarpals and phalanges, Fig. 1) scaled to FL in a Multidimensional Scaling (MDS) based on Bray-Curtis distance. Kruskal’s stress less than 0.2 was considered an adequate representation in reduced dimensions (McCune and Grace 2002). All the analyses were performed in Rwizard 4.3 (Guisande et al. 2014) and PAST 4.03 (Hammer et al. 2001). RESULTS Wing morphometry and guild classification Among the 97 bat species analyzed, the smallest was Rhogessa io (FL mean = 28.64 mm) and the largest was Phyllostomus hastatus (FL mean = 88.52 mm) (Table 1, Fig. 2). Means of LD3 ranged from 50.31 mm (Eptesicus fuscus) to 176.2 mm (Noctilio leporinus), LD4 varied from 41.99 mm (Eptesicus fuscus) to 132.67 mm (Noctilio leporinus), and LD5 were 33.81 mm (Cynomops planirostris) to 120.22 mm (Chrotopterus auritus). However, when scaling wing digits to body size (FL ratio), Lasiurus seminolus displayed the longest wing (LD3/FL), whereas Rhinophylla alethina had the broadest (LD5/FL) (Fig. 3). Bats were classified into 11 guilds: background cluttered space aerial insectivores (BAI, 23 species), background cluttered space trawling insectivore/ piscivore (BTP, 2 species), highly cluttered space aerial insectivores (CAI, 2 species), highly cluttered space gleaning insectivores (CGI, 7 species), uncluttered space aerial insectivore (UAI, 6 species), highly cluttered Fig. 1. Wing traits measured from bat specimens. Wing structures (metacarpals and phalanges) are represented in the figure. Wing digit length is the sum of metacarpals and phalanges of each digit. page 3 of 14 Zoological Studies 59:60 (2020) © 2020 Academia Sinica, Taiwan space gleaning canopy frugivore (CCF, 23 species), highly cluttered space gleaning understory frugivore (CUF, 18 species), highly cluttered space gleaning nectarivore (CGN, 10 species), highly cluttered space gleaning carnivore (CGC, 2 species), highly cluttered space gleaning omnivore (CGO, 3 species), and highly cluttered space gleaning sanguinivores (CGS, 1 species) (Table 1). Correlations among wing digits The correlation analyses for all 97 species showed a positive correlation between LD3 and FL (rs = 0.88, P < 0.001, Fig. 4a), as well as LD5 and FL (rs = 0.83, P < 0.001, Fig. 4b) and LD3 and LD5 (rs = 0.91, P < 0.001; Fig. 4c). Wing ratios Significant differences were found among bat guilds in terms of the LD3/FL ratio (F = 58.03, d.f. = 10, P < 0.0001, Fig. 5a), LD5/FL ratio (F = 39.13, d.f. = 10, P < 0.0001, Fig. 5b), and LD3/LD5 ratio (F = 25.44, d.f. = 10, P < 0.0001, Fig. 5c). Guilds that capture Fig. 2. Frequency histograms of wing variables for 97 Neotropical bat species. Variables include (a) forearm length (b) length of the third digit (c) length of the fourth digit and (d) length of the fifth digit. Table 1. Bat species analyzed according to wing digits, including the number of individuals examined (n) and corresponding acronym and guild. Mean and standard deviation are given for: forearm length (FL), length of the third digit (LD3), length of the fourth digit (LD4), and length of the fifth digit (LD5). Measurements are in mm FAMILY/ Species Acronym Guild nFL LD3 LD4 LD5 EMBALLONURIDAE Cormura brevirostris Cobr BAI 144.58 69.48 47.16 44.74 Peropteryx kappleri Peka BAI 2 46.72 ± 0.95 78.31 ± 0.59 52.68 ± 2.51 52.89 ± 0.09 Rhynchonycteris naso Rhna BAI 136.02 62.5 42.57 38.51 Saccopteryx bilineata Sabi BAI 6 47.11 ± 1.67 84.41 ± 4.10 58.87 ± 2.14 58.50 ± 2.27 Saccopteryx gymnura Sagy BAI 138.44 74.62 47.89 48.12 Saccopteryx leptura Sale BAI 12 37.29 ± 2.31 66.17 ± 7.73 44.55 ± 4.77 42.135 ± 4.67 MOLOSSIDAE Cynomops planirostris Cypl UAI 4 32.77 ± 0.52 64.1 ± 1.74 49.2 ± 1.87 33.81 ± 1.10 page 4 of 14Zoological Studies 59:60 (2020) © 2020 Academia Sinica, Taiwan FAMILY/ Species Acronym Guild nFL LD3 LD4 LD5 Molossops temminckii Mote UAI 5 30.56 ± 0.95 63.37 ± 1.57 52.41 ± 1.90 52.41 ± 1.42 Molossus molossus Momo UAI 14 38.89 ± 2.51 78.08 ± 6.84 58.37 ± 3.56 39.61 ± 2.35 Molossus pretiosus Mopr UAI 143.56 85.34 64.82 44.08 Molossus rufus Moru UAI 5 49.88 ± 0.43 100.01 ±7.12 71.85 ± 2.07 49.98 ± 1.65 Tadarida brasiliensis Tabr UAI 142.25 79.4 59.91 44.03 MORMOOPIDAE Mormoops megalophylla Mome CAI 12 53.37 ± 1.13 93.47 ± 1.62 67.86 ± 2.26 60.98 ± 1.86 Pteronotus parnellii Ptpa CAI 4 60.23 ± 1.47 95.4 ± 3.50 75.24 ± 0.46 73.46 ± 0.59 NATALIDAE Natalus tumidirostris atu BAI 939.54 ± 0.89 76.52 ± 1.96 56.19 ± 1.16 55.76 ± 1.54 NOCTILIONIDAE Noctilio albiventris Noal BTP 6 59.68 ± 1.79 116.39 ± 3.95 86.92 ± 3.43 65.52 ± 2.11 Noctilio leporinus Nole BTP 2 84.32 ± 1.25 176.205 ± 2.80 132.67 ± 2.30 100.96 ± 2.85 PHYLLOSTOMIDAE Carollinae Carollia brevicauda Cabr CUF 72 39.47 ± 1.77 84.45 ± 4.92 60.17 ± 3.71 59.80 ± 3.19 Carollia castanea Caca CUF 56 36.31 ± 1.44 77.59 ± 4.49 54.99 ± 2.95 54.88 ± 2.54 Carollia perspicillata Cape CUF 243 41.90 ± 1.78 88.24 ± 4.31 63.20 ± 3.41 63.00 ± 3.24 Rhynophillinae Rhinophylla alethina Rhpu CUF 129.94 73.24 51.51 51.51 Rhinophylla fischerae Rhfi CUF 331.20 ± 1.32 73.37 ± 2.42 52.16 ± 1.35 49.65 ± 2.89 Rhinophylla pumilio Rhpu CUF 2 30.68 ± 0.19 69.37 ± 1.78 49.69 ± 1.49 47.20 ± 1.44 Desmodontinae Desmodus rotundus Dero CGS 43 59.60 ± 3.00 96.13 ± 7.26 77.74 ± 4.47 76.20 ± 4.61 Glossophaginae Anoura cadenai Anca CGN 135.23 72.2 50.49 44.53 Anoura geoffroyi Ange CGN 27 42.60 ± 1.91 90.97 ± 4.35 64.11 ± 3.34 54.82 ± 3.89 Anoura latidens Anla CGN 342.31 ± 1.19 84.40 ± 1.64 60.84 ± 1.85 52.48 ± 0.91 Anoura luismanueli Anlu CGN 11 35.5 ± 2.26 75.95 ± 5.32 53.68 ± 3.22 47.13 ± 3.29 Choeroniscus godmani Chgo CGN 4 32.10 ± 1.46 66.78 ± 2.61 47.23 ± 1.30 44.57 ± 1.76 Glossophaga longirostris Gllo CGN 137.42 75.24 54.02 52.23 Glossophaga soricina Glso CGN 85 35.03 ± 1.15 70.77 ± 2.67 51.76 ± 1.96 49.15 ± 1.91 Lonchophyllinae Hsunycteris cadenai Hsca CGN 132.33 70.77 48.68 46.14 Hsunycteris thomasi Hsth CGN 135.42 68.17 50.4 48.98 Micronycterinae Micronycteris hirsuta Mihi CGI 142.17 87.61 64.37 63.57 Micronycteris megalotis Mige CGI 4 34.26 ± 1.29 63.84 ± 1.82 50.19 ± 3.44 51.34 ± 2.82 Micronycteris microtis Mimi CGI 2 34.36 ± 0.30 63.26 ± 0.95 48.83 ± 0.05 50.04 ± 1.73 Micronycteris schmidtorum Misc CGI 2 34.82 ± 0.68 67.31 ± 4.49 50.47 ± 4.06 51.34 ± 3.95 Phyllostominae Chrotopterus auritus Chau CGC 182.23 153.25 116.68 120.22 Gardnerycteris crenulatum Gacr CGI 7 48.93 103.11 71.25 69.82 Lophostoma brasiliense Lobr CGI 134.53 61.08 49.81 52.3 Lophostoma silvicolum Losi CGI 13 52.22 ± 2.84 96.40 ± 6.57 78.26 ± 3.39 80.72 ± 3.84 Phylloderma stenops Phst CUF 176.29 96.40 78.26 80.72 Phyllostomus discolor Phdi CGN 23 62.23 ± 2.36 114.44 ± 4.66 83.59 ± 3.85 77.32 ± 2.78 Phyllostomus elongatus Phel CGO 160.71 116.98 87.65 89.47 Phyllostomus hastatus Phha CGO 13 88.52 ± 2.14 161.07 ± 4.68 118.09 ± 4.49 106.99 ± 2.52 Tonatia saurophila Tosa CGO 159.22 107.66 79.12 83.31 Trachops cirrhosus Trci CGC 6 60.24 ± 1.84 118.73 ± 3.95 87.05 ± 2.78 90.115 ± 3.07 Stenodermatinae Artibeus anderseni Aran CCF 7 38.52 ± 2.45 80.88 ± 5.81 59.34 ± 3.79 56.82 ± 4.76 Artibeus bogotensis Arbo CCF 5 42.09 ± 1.36 89.85 ± 3.34 66.04 ± 1.56 63.05 ± 2.21 Artibeus cinereus Arci CCF 341.71 ± 1.88 89.39 ± 1.29 64.65 ± 2.81 61.94 ± 2.95 Table 1. (Continued) page 5 of 14Zoological Studies 59:60 (2020) © 2020 Academia Sinica, Taiwan FAMILY/ Species Acronym Guild nFL LD3 LD4 LD5 Artibeus concolor Arco CCF 147.41 103.75 74.31 72.39 Artibeus glaucus Argl CCF 56 41.45 ± 1.59 88.67 ± 4.91 65.66 ± 3.06 61.81 ± 2.90 Artibeus gnomus Argn CCF 138.52 84.29 60.47 57.49 Artibeus jamaicensis Arja CCF 34 61.47 ± 3.73 125.19 ± 10.48 93.44 ± 6.27 86.28 ± 5.88 Artibeus lituratus Arli CCF 197 68.38 ± 2.92 68.38 ± 6.58 105.40 ± 4.49 98.84 ± 4.54 Artibeus phaeotis Arph CCF 44 37.81 ± 1.59 81.16 ± 3.73 59.24 ± 2.87 56.59 ± 2.76 Artibeus planirostris Arpl CCF 58 60.80 ± 2.83 123.34 ± 7.03 91.93 ± 4.99 85.86 ± 4.83 Artibeus rava Arra CCF 12 38.76 ± 1.06 81.68 ± 3.61 59.54 ± 1.87 56.98 ± 2.16 Chiroderma salvini Chsa CCF 4 51.11 ± 2.34 115.94 ± 5.21 80.07 ± 4.21 75.24 ± 4.12 Enchisthenes hartii Enha CCF 6 51.11 ± 1.88 115.94 ± 3.73 80.07 ± 3.05 75.24 ± 2.02 Mesophylla macconnelli Mema CUF 2 31.93 ± 0.07 68.72 ± 4.46 50.60 ± 1.40 48.66 ± 0.72 Platyrrhinus angustirostris Plan CCF 337.27 ± 1.61 84.43 ± 4.36 59.48 ± 2.32 55.87 ± 2.20 Platyrrhinus dorsalis Pldo CCF 15 45.94 ± 3.26 106.56 ± 5.22 75.87 ± 3.20 70.78 ± 3.26 Platyrrhinus helleri Plhe CCF 42 38.66 ± 2.08 87.17 ± 3.37 61.71 ± 2.28 58.55± 2.18 Platyrrhinus infuscus Plin CCF 154.14 126.69 87.82 83.07 Platyrrhinus lineatus Plli CCF 5 44.16 ± 1.35 100.40 ± 4.41 100.40 ± 3.31 67.64 ± 3.52 Platyrrhinus umbratus Plum CCF 4 45.75 ± 0.68 104.18 ± 3.47 74.57 ± 2.10 69.61 ± 1.89 Platyrrhinus vittatus Plvi CCF 6 60.97 ± 1.45 133.44 ± 4.17 96.28 ± 2.78 88.90 ± 3.69 Sphaeronycteris toxophyllum Spto CCF 140.95 91.95 66.52 62.54 Sturnira aratathomasi Star CUF 357.26 ± 1.78 128.21 ± 3.40 92.77 ± 1.87 86.44 ± 1.74 Sturnira bidens Stbi CUF 7 41.25 ± 1.11 93.61 ± 3.36 66.87 ± 2.51 64.28 ± 2.47 Sturnira bogotensis Stbo CUF 27 43.31 ± 1.86 92.90 ± 4.76 68.54 ± 3.42 64.52 ± 3.37 Sturnira erythromos Ster CUF 25 40.42 ± 1.59 87.16 ± 3.73 63.56 ± 2.54 60.36 ± 2.71 Sturnira lilium Stli CUF 65 41.18 ± 1.45 87.87 ± 4.01 64.68 ± 3.61 60.82 ± 2.73 Sturnira ludovici Stld CUF 847.40 ± 1.98 100.91 ± 6.01 74.87 ± 3.33 70.19 ± 2.75 Sturnira luisi Stlu CUF 141.21 85.21 64.8 60.79 Sturnira oporaphilum Stop CUF 943.94 ± 1.07 94.72 ± 2.46 69.52 ± 2.17 66.34 ±1.56 Uroderma bilobatum Urbi CUF 56 41.38 ± 1.71 88.37 ± 3.87 63.58 ± 3.07 60.44 ± 2.60 Uroderma magnirostrum Urma CUF 16 42.32 ± 1.41 89.78 ± 3.09 64.89 ± 2.49 61.68 ± 2.18 Vampyressa thyone Vath CCF 5 31.62 ± 1.09 68.67 ± 3.23 50.56 ± 2.33 46.94 ± 2.71 Vampyriscus nymphaea Vany CCF 337.91 ± 2.01 79.41 ± 3.06 56.04 ± 1.43 55.02 ± 1.94 THYROPTERIDAE Thyroptera tricolor Thtr BAI 138.74 70.55 55.68 51.05 Thyroptera lavali Thla BAI 136.17 64.01 49.73 46.33 VESPERTILIONIDAE Eptesicus andinus Epan BAI 146.49 85.49 68.72 57.34 Eptesicus furinalis Epfu BAI 137.01 50.31 41.99 36.7 Eptesicus brasiliensis Epbr BAI 2 42.31 ± 1.78 71.73 ± 1.12 59.39 ± 2.80 50.26 ± 1.68 Histiotus montanus Himo BAI 348.80 ± 1.30 80.56 ± 1.01 65.83 ± 0.85 63.93 ± 0.81 Lasiurus blossevillii Labl BAI 338.26 ± 1.07 75.55 ± 3.79 57.99 ± 1.65 48.17 ± 0.92 Lasiurus ega Laeg BAI 2 46.66 ± 0.19 90.35 ± 1.70 71.9 ± 1.90 58.04 ± 1.97 Lasiurus seminolus Lase BAI 136.18 78.87 57.32 49.33 Myotis albescens Myal BAI 2 35.03 ± 1.40 58.88 ± 1.13 49.50 ± 1.23 45.54 ± 1.32 Myotis keaysi Myke BAI 941.46 ± 0.77 67.46 ± 1.82 56.21 ± 1.56 51.69 ± 1.35 Myotis keenii Mykn BAI 134.92 58.95 47.82 46.36 Myotis nigricans Myni BAI 25 33.92 ± 2.47 55.57 ± 6.11 45.79 ± 4.83 42.29 ± 4.37 Myotis oxyotus Myox BAI 6 41.03 ± 0.77 66.30 ± 2.73 54.01 ± 1.42 50.62 ± 1.50 Myotis riparius Myri BAI 132.21 54.61 43.96 42.07 Rhogeessa io Rhio BAI 5 28.64 ± 1.48 53.36 ± 1.93 44.29 ± 2.27 38.12 ± 1.48 Note: Species were classified into the 11 ecological guilds proposed by Aguirre et al. (2016), Sampaio et al (2003), Estrada et al. (2010) and Kalko et al. (1996): background cluttered space aerial insectivore (BAI); background cluttered space trawling insectivore/piscivore (BTP); highly cluttered space aerial insectivores (CAI); highly cluttered space gleaning canopy frugivore (CCF); highly cluttered space gleaning understory frugivore (CUF); highly cluttered space gleaning carnivore (CGC); highly cluttered space gleaning insectivores (CGI); highly cluttered space gleaning nectarivore (CGN); highly cluttered space gleaning omnivore (CGO); highly cluttered space gleaning sanguinivores (CGS); uncluttered space aerial insectivore (UAI). Table 1. (Continued) page 6 of 14Zoological Studies 59:60 (2020) © 2020 Academia Sinica, Taiwan their prey by employing aerial mode in uncluttered and background habitats (UAI, BAI) differed significantly from the other guilds in the LD3/FL ratio according to post-hoc pairwise comparisons (Dunn P < 0.05). Conversely, guilds that forage in highly cluttered spaces using a gleaning strategy (CUF, CCF, CGN, CGC, CGI, CGO, CGS) presented higher values for the LD5/ FL ratio and lower ones for the LD3/LD5 ratio than did aerial and trawling foragers from uncluttered and background habitats (UAI, BTP) (Dunn P < 0.05, Fig. 5b c). Morphometric ordination The MDS with individual bone structures resulted in a two-dimensional solution explaining 60.76% of Fig. 3. Principal wing variables of 97 Neotropical bat species. Variables include (a) wing length (LD3/ FL ratio) and (b) wing width (LD5/FL ratio). Acronyms for bat species are defined in table 1. Fig. 4. Scatter plot of wing digits for the 97 Neotropical bat species. Figures include (a) correlation between LD3 and FL, exploring the correlation of hand-wing length and bat size; (b) correlation between LD5 and FL, assessing the correlation of wing width and bat size; and (c) correlation between LD3 and LD5, showing the correlation of hand-wing length and wing width. Acronyms for bat species are defined in table 1. page 7 of 14Zoological Studies 59:60 (2020) © 2020 Academia Sinica, Taiwan the overall variation (Kruskal’s stress = 0.03, MDS 1 = 37.09%, MDS 2 = 23.67%) and showed a gradient of three disjunct groups that corresponded to different habitat types (cluttered, background, uncluttered) and two foraging modes (gleaning and aerial) (Fig. 6, Table S1). Terminal phalanges were important variables for guild classification (P5.3, P4.3, P3.4, Fig. 6). However, metacarpals were clearly associated with uncluttered and background insectivores (P2.1, P3.1, P4.1, P5.1, Fig. 6). DISCUSSION General findings Wing structures reflect general patterns of habitat use and foraging mode in Neotropical bats. Since echolocation and foraging behavior are mainly influenced by habitat characteristics and foraging mode, species grouped in the same guild share similar sensory and motor adaptations (Denzinger and Schnitlzer 2013), many of which are reflected in wing structures (Fig. 6). Although diet is not a trait necessarily reflected in wing morphology, guilds may also differ from others depending on ecological features linked to food choice (Denzinger et al. 2016). In their seminal paper, Norberg and Rayner (1987) examined the wing morphology of several guilds based on wing loading, aspect ratio, and wing tip index, showing essential ecomorphological relationships to flight patterns and habitat use (Stockwell 2001; Rhodes 2002; Marinello and Bernard 2014). However, approaches addressing different bat structures that support the wing such as metacarpals, phalanges, and digits have been overlooked. In fact, for several species, data presented in this paper are the first measurements of morphometric wing structures (Table 1). It is particularly important to document individual bone traits when analyzing the relationships among foraging mode, habitat use and wing morphology. Findley et al. (1972) stressed that the use of FL, LD3, and LD5 may be wing-area proxies that Fig. 5. Boxplots of wing ratios among 11 Neotropical bat guilds. Comparisons include (a) LD3/FL ratio, (b) LD5/FL ratio, and (c) LD3/LD5 ratio. Acronyms for bat guilds are defined in table 1. (a) (b) (c) page 8 of 14Zoological Studies 59:60 (2020) © 2020 Academia Sinica, Taiwan are correlated with wing indices (e.g., tip index and aspect ratio), and thereby reflect aerodynamic performance (Norberg and Rayner 1987) as well as the ecological strategies of the bats (Marinello and Bernard 2014). Species in the same guild share similar wing morphologies, regardless of size (Norberg and Rayner 1987), but some variations among guilds presumably show adaptations to specific habitats and foraging behaviors (Dietz et al. 2006). This agrees with the results presented here. For example, bats that capture prey using an aerial strategy in uncluttered or background spaces (BAI, UAI) presented a higher LD3/FL ratio than did guilds whose food resources are obtained through gleaning strategy in highly cluttered spaces (CAI, CCF, CUF, CGC, CGI, CGN, CGO, CGS; Fig. 5a). By contrast, gleaning guilds showed a higher LD5/FL ratio than did aerial insectivores (Fig. 5b). This is because, on one hand, larger LD3 in relation to FL reflects a longer hand-wing length, which is associated with fast and economic flights, typical of aerial insectivores (Dietz et al. 2006). On the other hand, a larger LD5 in proportion to FL is indicative of wider wings with high maneuverability and hovering ability, easing slow flight in narrow spaces (Dietz et al. 2006; Castillo-Figueroa and Pérez-Torres 2018). Other guilds adapted to foraging on surface dwellings (CGI, CGC, CGS) can localize and hunt prey through a combination of abilities such as echolocation, vision, and detection of prey-generated sounds (Razak 2018). These guilds showed a lower LD3/LD5 ratio, allowing individuals to fly slowly through narrow spaces (Dietz et al. 2006, Fig. 5c). These results were supported by MDS analysis (Fig. 6), suggesting a clear distinction in wing structures among guilds that display different foraging strategies and habitat types, as was stressed by Marinello and Bernard (2014). Morphological traits may predispose bats to feed on specific habitats and adopt certain foraging strategies (Kalcounis and Brigham 1995). Phalanges were shown to be key traits in guild classification, but metacarpals also presented remarkable associations with bat guilds (Fig. 6). Metacarpals are key for generating lift (Findley et al. 1972; Stevens et al. 2013) and can be a determinant for wing length, which is typical in the morphology of aerial foragers from uncluttered and background habitats. Indeed, these guilds showed the largest metacarpals in the second (P2.1), third (P3.1) and fourth digits (P4.1), thus indicating longer wings. In the case of phalanges, these structures influence wingtip longitude, facilitating agility and propulsion during flights. Thus, larger phalanges substantially improve aerodynamic ability (Findley et al. 1972; Altringham 1996), especially in gleaning bats. Hereby, both metacarpals and phalanges may have a significant role in flight performance and ecological adaptations of guilds. Several factors other than wing morphology, however, may explain the ecological differentiation among guilds; e.g., predatorprey relationships and roosting behavior (Rhodes 2002). Fig. 6. Multidimensional Scaling (MDS) ordination plot showing variation of individual bone structures among bat guilds. Each dot represents a species and the relative distance between two points reflect the relative dissimilarity. Polygons represents each one of the bat guilds. Metacarpals and phalanges are scaled to size (i.e., wing structure /FL ratio) and are found in figure 1. Acronyms for bat guilds are defined in table 1. page 9 of 14Zoological Studies 59:60 (2020)