Hematophagous Attracted to Calls in a Preserved Seasonal Forest of the Austral Neotropics, with a Description of a New Species of Corethrella (Diptera: ) Author(s): Vinícius Matheus Caldart, Maurício Beux dos Santos, Samanta Iop, Luiz Carlos Pinho, and Sonia Zanini Cechin Source: Zoological Science, 33(5):527-536. Published By: Zoological Society of Japan DOI: http://dx.doi.org/10.2108/zs150173 URL: http://www.bioone.org/doi/full/10.2108/zs150173

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. ZOOLOGICAL SCIENCE 33: 527–536 (2016) © 2016 Zoological Society of Japan

Hematophagous Flies Attracted to Frog Calls in a Preserved Seasonal Forest of the Austral Neotropics, with a Description of a New Species of Corethrella (Diptera: Corethrellidae)

Vinícius Matheus Caldart1,2*, Maurício Beux dos Santos1, Samanta Iop1, Luiz Carlos Pinho3, and Sonia Zanini Cechin1

1Programa de Pós Graduação em Biodiversidade , Centro de Ciências Naturais e Exatas, Universidade Federal de Santa Maria, CEP 97105-900, Santa Maria, RS, Brazil 2Departamento de Ecologia, Instituto de Biociências, Universidade de São Paulo, Rua do Matão, trav. 14, n° 321, Cidade Universitária, CEP 05508-900, São Paulo, SP, Brazil 3Universidade Federal de Santa Catarina, Departamento de Ecologia e Zoologia, Campus Trindade, CEP 88040-900, Florianópolis, SC, Brazil

The signaler–eavesdropper interaction has been investigated for a wide range of organisms, and although many flies feed on calling , this dynamic has been addressed only poorly in the aus- tral Neotropics. We investigated this interaction in southern Brazil using pairs of suction traps (acoustic + silent) broadcasting frog calls or an artificial white noise in ponds and streams. From 139 sessions, flies of the genera Corethrella (Corethrellidae), Forcipomyia (Ceratopogonidae) and (Culicidae) were collected, including five Corethrella species, the most abundant of which was previously unknown and is formally described here. Additionally, we present the south- ernmost records of Corethrella lopesi, C. alticola and C. atricornis. Numbers of Forcipomyia and Uranotaenia mosquitoes did not differ between silent traps and traps broadcasting frog calls, and did not differ between white noise traps and adjacent silent traps. However, the number of female Corethrella was significantly higher in traps broadcasting calls of the pond-breeding frog P. aff. gracilis compared to adjacent silent traps; calls of this frog attracted the five Corethrella species and also collected significantly more female Corethrella than the white noise. By evaluating differ- ent taxa of flies and broadcasting different sounds, we demonstrated that Corethrella midges were attracted only to the acoustic cue of P. aff. gracilis calls, while Forcipomyia and Uranotaenia were captured in traps by chance. Our results suggest that female Corethrella feed on males of the com- mon pond-breeding frog P. aff. gracilis in southern Brazil, and highlight the utility of frog call traps in revealing the diversity of Corethrella in the austral Neotropics.

Key words: acoustic communication, eavesdroppers, Corethrella, anurans, white noise

acoustic signals constitute sexually selected traits (Ryan, INTRODUCTION 1980; Zuk et al., 2006), or impose a direct survival cost due Illegitimate receivers, or eavesdroppers, represent a to increased exposure to acoustically-oriented predators or non-target audience that may exploit a signal emitted by a parasitoids (Tuttle and Ryan, 1981; Zuk et al., 2006). given animal to its intended receivers. The nature of the A widespread signaler–eavesdropper interaction occurs interaction between the signaler and the eavesdropper influ- between male frogs and acoustically-oriented hematopha- ences the costs that the eavesdropper may impose on the gous flies, mainly the frog-biting midges of the Core- signaler (Zuk and Kolluru, 1998; Peake, 2005). The illegiti- thrella. These midges belong to the family Corethrellidae, mate receivers of a given acoustic signal, for instance, may which includes 107 extant species found primarily in tropical be non-target conspecifics (Halfwerk et al., 2014), predators and subtropical areas worldwide (Yu et al., 2013; Borkent, (Tuttle and Ryan, 1981; Halfwerk et al., 2014), parasitoids 2014; Amaral and Pinho, 2015). Female Corethrella locate (Zuk et al., 2006) or hematophagous frog-biting midges male frogs using their calls as an auditory cue, and feed on (Bernal et al., 2006; Grafe et al., 2008). The selective their blood to obtain the blood meal necessary for egg pro- pressure of eavesdroppers on signalers may thus either duction, a process in which midges may transmit trypano- constrain the reproductive success of signalers, if their somes to hosts (Borkent, 2008; Ferguson and Smith, 2012). Borkent (2008) suggested that frog-biting midges have likely * Corresponding author. Tel. : +55-11-3091-0969; had an important role in the history of frog evolution, as both E-mail: [email protected] groups have been interacting for at least 190 million years Supplemental material for this article is available online. and currently live in sympatry in most parts of the globe. doi:10.2108/zs150173 Interactions between frog-biting midges and calling frogs 528 V. M. Caldart et al. have been well documented for different species of frogs Silva, 2015). Additionally, female and male adults of a new and midges from North America (McKeever, 1977), Central species of Corethrella belonging to the rotunda group are America (Bernal et al., 2006) and tropical and subtropical described, and the interaction of this species with its frog Asia (Toma et al., 2005; Grafe et al., 2008; Borkent and hosts is discussed. Grafe, 2012), but similar studies have only recently begun in the austral Neotropics (Amaral and Pinho, 2015). MATERIALS AND METHODS Besides Corethrella, some other genera of hematopha- Ethical approval gous flies are also known to feed on frogs. At least one mos- All applicable national and institutional guidelines for the care quito species, (Culicidae), is known to and use of were followed. All procedures performed were feed selectively on and to locate frog-hosts in accordance with the ethical standards of the institution or practice using the acoustic cue of their calls (Borkent and Belton, where the studies were conducted. Activities at Turvo State Park, 2006). This was attracted in Costa Rica to the call including the collections of the Diptera taxa, were performed in of a tree frog native to the southeastern United States, the accordance with the licenses issued by SEMA-RS (#23/2011) and barking tree frog Hyla gratiosa, which also effectively ICMBio-IBAMA (#29505-4). attracted many female Corethrella (Borkent and Belton, Study Area 2006). Likewise, many female Uranotaenia and female The study was conducted between October 2013 and February Corethrella were attracted to frog calls broadcasted on Irio- 2014 (austral spring and austral summer) in three streams and mote Island, of the Ryukyu Archipelago, in subtropical three ponds (Supplementary Figure S1) within Turvo State Park Japan (Toma et al., 2005). The specific components of frog (27°14′34.08″S, 53°57′13.74″W, 376 m.a.s.l.), located in the munic- calls that are attractive to species of Uranotaenia and ipality of Derrubadas, in the northwestern region of Rio Grande do Corethrella are not well understood, but Borkent and Belton Sul State, Brazil. The park covers an area of 17,491.4 ha (SEMA, (2006) suggested that the antennae of U. lowii and Corethrella 2005) adjacent to the border of the Uruguay River and near the species might resonate at the fundamental frequency of Argentinean forests of Moconá Provincial Park and the Yabotí Inter- Hyla gratiosa calls, which is around 400–500 Hz. national Biosphere Reserve, and is one of the last large remnants of mesophytic semideciduous forest (sensu Oliveira-Filho et al., While there is compelling evidence indicating the use of 2006) in southern Brazil. The local climate is characterized as sub- auditory cues in host-seeking behavior of Corethrella tropical sub-humid with a dry summer, an annual rainfall of 1787 (McKeever and French, 1991; Toma et al., 2005; Bernal et mm, a mean annual temperature of 18.8°C and a mean tempera- al., 2006; Grafe et al., 2008; Bernal and Silva, 2015), as well ture of the coldest month of 13.3°C (ST SB v climate type, sensu as some evidence suggesting that at least some species of Maluf, 2000). Uranotaenia are also capable of this (Toma et al., 2005; Borkent and Belton, 2006), the specific cue used in host Sound Preparation location is not known for other hematophagous flies that Natural calls of three anuran species commonly found in the also feed on frogs. This may be the case for biting-midges study area were broadcasted. These were the stream-breeding frog of the genus Forcipomyia (Ceratopogonidae), which have schmidti (; Caldart et al., 2010), the pond-breeding frog Physalaemus aff. gracilis (Leptodactylidae; been observed attacking the frogs Leptodactylus chaquensis Kwet et al., 2010), and the pond-breeding tree-frog perereca and Pseudis paradoxa in Argentina (Spinelli et al., 2002). In (; Kwet et al., 2010; Frost, 2015). Males of the stream- fact, evidence of the interaction between acoustically- breeding frog C. schmidti call upon exposed rocks in the stream oriented hematophagous flies and male frogs from the aus- bed with little acoustic overlap between nearby males (Caldart et tral Neotropics is circumstantial. We are aware of only the al., 2010, 2011). They are most acoustically active during daylight following three records: the above-mentioned interaction hours, though may occasionally be heard calling at night (Caldart et between Forcipomyia and Leptodactylus chaquensis and al., 2016). The snout–vent length (SVL) of male C. schmidti ranges Pseudis paradoxa (Spinelli et al., 2002), an interaction from 22.14–27.95 mm, and mature individuals of both sexes may between unidentified hematophagous flies and male occur throughout the year at the study area (Caldart, unpublished Crossodactylus schmidti (Caldart et al., 2014), and that data). Males of the pond-breeding frog P. aff. gracilis form dense calling choruses while floating in the water, partially hidden within between Corethrella alticola attracted to Physalaemus aquatic vegetation (Iop et al., 2012). Male P. aff. gracilis have an cuvieri calls in southern Brazil (Amaral and Pinho, 2015). SVL ranging from 26–31 mm, are acoustically active during the day In the present study, we investigated, for the first time in and the night, and are reproductively active from September to the austral Neotropics, phonotactic responses of different February (Kwet et al., 2010). Males of the pond-breeding tree-frog taxa of hematophagous flies to varying sources of sound. Scinax perereca also form dense choruses at night and call while During field phonotaxis trials performed in ponds and perched in small trees or in bushes located within or near ponds streams, we broadcasted loops of natural advertisement (Kwet et al., 2010). Male Scinax perereca have an SVL ranging calls recorded from three potential frog-host species of the from 36–41 mm, and are acoustically and reproductively active from Atlantic Forest in southern Brazil. We also broadcasted a September to February (Kwet et al., 2010); at the study area S. putative unattractive sound in the form of an artificial, wide- perereca commonly co-occurs in ponds with P. aff. gracilis (Iop et al., 2012). frequency range white noise, to test whether different taxa The advertisement calls of six males of C. schmidti, six males of flies would be attracted to a non-natural abiotic source of of S. perereca and five males of P. aff. gracilis were recorded at sound. We predicted that acoustically-oriented hematopha- the study area, with the number of recorded calls per male ranging gous flies would be attracted to the acoustic stimuli of only from three to five. The recordings were made using a Marantz® dig- the frog calls, especially the frog-biting midges of the genus ital recorder (model PMD671) coupled to a Sennheiser® micro- Corethrella that are well known to use auditory cues in host- phone (model ME66) positioned at a distance of 50 cm from the seeking behavior (McKeever and French, 1991; Bernal and calling male. The sound pressure level of calls (in dB, re 20 μPa) Hematophagous flies and frog calls 529 was measured with an Impac® sound level meter (model IP-120C; equipment would attract flies. We did from 4–7 diurnal trapping C weighting curve: 31.5–8000 Hz, dB range: 40–130 dB) at a dis- sessions per day and 4–10 dusk/nocturnal trapping sessions per tance of 50 cm from the calling male. The recorded calls were ana- day. We waited for at least seven days before running new trapping lyzed with the SoundRuler software v. 0.9.6.0 (http://soundruler. sessions in a pond or a stream that had been sampled earlier in the sourceforge.net; (Bee, 2004; Gridi-Papp, 2004) at a sampling fre- same month. Trapping sessions were conducted at air temperature quency of 44,100 Hz and a resolution of 16 bits, in order to obtain and air humidity values varying from 20–32°C and 56–83%, respec- measurements of the following call parameters: call duration, call tively. We performed a total of 139 trapping sessions, totaling nearly interval, number of notes, note duration, call rate, call duty cycle, 105 hours of sampling effort. and call dominant frequency. Subsequently to the bioacoustics analysis of the recordings we created playback loops of calls, using Statistics the software Adobe Audition, to be broadcasted during field trials. The number of female flies captured in the acoustic and control Call loops were created to represent average calls of each frog spe- traps was compared using the Wilcoxon signed-rank test for each cies based on the mean values obtained for the call parameters, type of acoustic stimulus separately (i.e., anuran species and white including the call duty cycle. noise). This paired test allows controlling for non-measured environ- The call loop of C. schmidti (Supplementary Figure S2) con- mental factors that could influence the trapping success at each sisted of a long call of 15.4 s duration, comprising 53 notes, emitted trap pair position (e.g., local variation in temperature, light condi- at a call rate of two calls per minute, with 15 s of silent interval tions). We compared further the number of female flies significantly between calls and a mean call dominant frequency of 3350 Hz; the attracted to a given acoustic stimulus with the number of flies cap- call loop was equalized at an amplitude of 70 dB with the Impac® tured in traps broadcasting a white noise, using the Mann-Whitney sound level meter positioned in front of the loudspeaker prior to test. All data used were non-normally distributed (Shapiro-Wilk test; each session. Call loop of C. schmidti was adjusted to maintain the W > 0.55 and P < 0.001, in all cases). Analyses were made in the call duty cycle at 15 per cent. The call loop of P. aff. gracilis calls Past software, version 3 (Hammer et al., 2001). (Supplementary Figure S3) consisted of a 90 s chorus comprising whine calls of approximately 5 s each, emitted at a call rate of 10 calls per minute, with a silent interval of 41 s between choruses and To identify the collected flies to the genus level we consulted a mean call dominant frequency of 588 Hz; the call loop was equal- taxonomic studies that present morphological diagnostic characters ized at an amplitude of 69 dB and adjusted to maintain the call duty of the genera that we captured (Thielmann and Hunter, 2007; cycle at 80 per cent. Finally, the call loop of S. perereca Marino and Spinelli, 2008; Borkent, 2008; Brown et al., 2009). (Supplementary Figure S4) consisted of a 23 s chorus comprising Species of Corethrella were identified using a key to the New World short calls of 0.5 s, emitted at a call rate of 114 calls per minute, species (Borkent, 2008). Corethrella specimens were mounted on with a silent interval of 29 s between choruses and a mean call slides in Euparal following procedures outlined by Sæther (1969). dominant frequency of 3129 Hz; the call loop was equalized at an Terms, measurements, characters analyzed and examination amplitude of 60 dB and adjusted to maintain the call duty cycle at techniques are based on Borkent (2008). The holotype of the new 25 per cent. In addition, we also broadcasted a loop of an artificial species is deposited at Museu de Zoologia da Universidade de white noise equalized at 66 dB, which was created to encompass São Paulo (MZUSP). Paratypes are deposited at Instituto Nacional the frequency range and the mean amplitude of calls of the three de Pesquisas da Amazônia (INPA), Coleção Entomológica da anuran species; the noise loop was composed by equal-amplitude Universidade Federal de Santa Catarina (UFSC) and MZUSP. frequencies, ranging from 0–10,000 Hz, emitted at a 5 s noise dura- Voucher specimens of the remaining species are deposited at tion and a silent interval of 5 s between noise periods (Supplementary UFSC. This article was registered in the Official Register of Zoological Figure S5). The white noise loop was adjusted to maintain the noise Nomenclature (ZooBank) as http://zoobank.org/1B03910D-4038- duty cycle at 50 per cent. 4A01-B7B3-9939379E0641.

Field trials RESULTS We collected flies by broadcasting the frog call loops from three pairs of suction traps (modified from Mckeever and Hartberg, 1980), We conducted 104 hours and 25 minutes of trapping which consisted of a PVC pipe with a suction device in the middle, sessions, totaling 139 sessions, of which 59 (43% of total) and a funnel at the top with an attached plastic container filled with collected at least one of the target taxa of hematophagous 80% alcohol for collecting the sucked flies. The plastic container flies. Most collected specimens were frog-biting midges of was attached to the lower end of the funnel to avoid damaging the the genus Corethrella (n = 188), followed by midges of the flies from passing through the suction fan. The funnel was made genus Forcipomyia (n = 80), and mosquitoes of the genus with a thin net-grid that allowed the passage of the air generated by Uranotaenia (n = 16). Specimens of Corethrella were cap- the fan, thus enabling the suction of all collected flies toward the plastic container. At the base of the trap we placed a small battery tured only at dusk or at night (33 from 99 dusk/nocturnal that powered the suction device, and near the opening of the funnel sessions; 33% of total), as well as specimens of Uranotae- a Sony® loudspeaker was attached (model ICD-PX333; overall fre- nia (10 from 99 dusk/nocturnal sessions; 10% of total). Dif- quency response: 75–20,000 Hz) that broadcasted the frog call ferently, specimens of Forcipomiya were captured both dur- loops equalized at each respective amplitude. ing the day and at night (30 from 137 diurnal/nocturnal The suction traps were disposed in a paired design in ponds sessions; 22% of total). and streams for trapping sessions of 45 minutes of duration during The numbers of captured specimens of Forcipomyia the diurnal period (11:00 to 17:00) and the dusk/nocturnal period midges and Uranotaenia mosquitoes did not differ between (18:00 to 23:00). Each pair consisted of an acoustic trap broadcast- silent traps and their adjacent acoustic traps broadcasting ing the frog call loop and an adjacent silent control trap placed calls of Crossodactylus schmidti, Physalaemus aff. gracilis approximately 5 m from the acoustic trap. To minimize acoustic interference, pairs of traps were placed at least 20 m apart from or Scinax perereca (Table 1). Likewise, the numbers of cap- each other, and between two consecutive trapping sessions we tured specimens of Forcipomyia and Uranotaenia did not switched the stimulus of the pairs (i.e., the acoustic trap became the differ between traps broadcasting a white noise and adja- silent trap, and vice-versa) to control for the possibility that the cent silent traps (Table 1). These flies were randomly cap- 530 V. M. Caldart et al. tured in either acoustic or silent traps, with a mean number Table 1. Results of the Wilcoxon signed-rank test comparisons of 1.13 (range 0–4) specimens of Forcipomyia collected in between Forcipomyia midges and Uranotaenia mosquitoes col- the acoustic traps (median = 1; interquartile range = 0–1) lected in acoustic traps broadcasting Crossodactylus schmidti, Physalaemus aff. gracilis and Scinax perereca advertisement calls and 1.5 (range 0–13) specimens collected in the silent traps or a white noise, and the adjacent control silent traps. The N below (median = 1; interquartile range = 0–1.25). The mean num- each acoustic stimulus refers to the number of sessions with midges ber of Uranotaenia specimens collected in the acoustic traps and mosquitoes captured. “NA” refers to a statistical result not avail- was 1.1 (range 0–4) (median = 1; interquartile range = 0–2), able because only a single session captured specimens. and in silent traps it was 0.7 (range 0–2) (median = 1; inter- Acoustic stimuli quartile range = 0–1). Crossodactylus Physalaemus Scinax White Genera of flies The number of specimens of Corethrella captured, on schmidti aff. gracilis perereca noise the other hand, was significantly higher in traps broadcast- (N = 20) (N = 7) (N = 6) (N = 7) ing calls of P. aff. gracilis compared to adjacent silent traps Z = 0.43 Z = 1 Z = 0.38 Z = 0.58 Forcipomyia spp. (a total of 60 vs. 13 female Corethrella, respectively; P > 0.05 P > 0.05 P > 0.05 P > 0.05 Wilcoxon signed-rank test: Z = 2.63, P < 0.01), though the Z = 0.45 Z = 0.28 NA Z = 0.0 Uranotaenia spp. number did not differ between pairs of traps broadcasting C. P > 0.05 P > 0.05 P > 0.05 schmidti or S. perereca calls and their adjacent silent traps (Fig. 1). Traps broadcasting P. aff. gracilis calls attracted a mean of 6 (range 1–26) female Corethrella per trapping ses- sion (median = 2; interquartile range = 1–7.25) versus a mean of 1.3 (range 0–6) collected in adjacent silent traps (median = 0; interquartile range = 0–2.25). Furthermore, traps broadcasting P. aff. gracilis calls collected significantly more female Corethrella (mean of 6, range 1–26) compared to traps broadcasting a white noise (Mann-Whitney test: U = 6, P < 0.05), which collected an average of just 0.8 (range 1–1) female Corethrella (median = 1; interquartile range = 0.5–1). The advertisement call of P. aff. gracilis attracted a total of 60 female Corethrella, including the five Corethrella spe- cies sampled at the study area (Table 2). The most abun- dant of them was previously unknown and is formally described in the taxonomy section of this work. Four Corethrella species were collected in ponds, of which two were exclusive to this type of , and three species Fig. 1. Number of frog-biting midges (Corethrella) captured in were collected in streams, with only C. yucuman sp. nov. traps broadcasting Crossodactylus schmidti, Physalaemus aff. exclusive to this habitat. Interestingly, along with nearly all gracilis and Scinax perereca advertisement calls or a white noise, females of C. yucuman sp. nov caught in the trapping ses- compared with adjacent silent traps. Shown here are the mean num- sions (n = 12), we also collected individual males and male bers of captured females with the standard error as whiskers, and swarms (n = 70). the maximum number of midges captured in a single capture ses- sion as black dots. The N below each acoustic stimulus refers to the number of sessions with female midges captured. **Significant dif- Taxonomy ference by the Wilcoxon signed-rank test (Z = 2.63, P < 0.01); Corethrella yucuman Caldart & Pinho sp. n. *Significant difference by the Mann-Whitney test (U = 6, P < 0.05); (Figs. 2–3, Table 3) ns: no significant difference by the Wilcoxon signed-rank test for the calls of C. schmidti (Z = 0.69, P > 0.05), S. perereca (Z = 0.72, P > For LSID, see http://zoobank.org/382AE851-E335-46F8-91CA- 0.05) and the white noise (Z = 1.34, P > 0.05). FAEF915548BC.

Corethrella Type material Table 2. List of the species captured in ponds and streams using acoustic traps broadcasting frog calls and adjacent silent control traps in southern Brazil. *Acoustic stimuli (cs = Holotype: female adult: Brazil, Rio Crossodactylus schmidti; pg = Physalaemus aff. gracilis; sp = Scinax perereca) arranged in Grande do Sul State, Derrubadas, order of highest to lowest number of frog-biting midges collected; **Although these anuran Parque Estadual do Turvo (Turvo State species do not call in streams, some specimens of Corethrella yucuman sp. nov. were Park), Riacho Salto, 27.139336°S collected in a few trials with their calls in lotic (three sessions each species). Number of male Corethrella captured are given within parenthesis. 53.881209°W, 185 m a.s.l., 17.xii.2013, silent frog call trap (PG1N, control), Ponds Streams Acoustic VM Caldart et al. leg. (MZUSP). Species Acoustic Silent Acoustic Silent Total stimuli* Paratypes: two females, as holotype trap trap trap trap (UFSC); one female and two males, Corethrella yucuman sp. nov. cs, pg**, sp** – – 6 (52) 6 (18) 82 as holotype except frog call trap Corethrella lopesi Lane pg, sp, cs 61 5 1 1 68 (PG1N, Physalaemus aff. gracilis call- Corethrella alticola Lane pg, cs, sp 17 16 2 – 35 ing) (INPA); two females, as holotype Corethrella atricornis Borkent pg 1 1 – – 2 except CS2N, control (MZUSP); one Corethrella aff. wirthi Stone pg 1 – – – 1 female, as holotype except frog call Total 80 22 61 25 188 Hematophagous flies and frog calls 531

Fig. 2. Corethrella yucuman Caldart & Pinho sp. n., female adult. (A) Head, anterior view. (B) Coronal suture. (C) Antenna. (D) Clypeus. (E) Right palpus. (F) Thorax and legs, lateral view. (G) Wing. trap (CS1N, Crossodactylus schmidti calling) (INPA); one (MZUSP); two males, as holotype except 14.ii.2014, CS1N female, as holotype except frog call trap (RB2N, white control (UFSC); one male, as holotype except Riacho Maria, noise) (INPA); one female, as holotype except RB5N, con- 27.243489°S, 53.953972°W, 382 m a.s.l., 13.xi.2013, frog trol (INPA); one female and five males, as holotype except call trap (PG1N, Physalaemus aff. gracilis calling) (MZUSP); 15.ii.2014, frog call trap (SP1N, Scinax perereca calling) one male, as previous except silent frog call trap (CS1N, 532 V. M. Caldart et al.

Fig. 3. Corethrella yucuman Caldart & Pinho sp. n., male adult. (A) Head, anterior view. (B) Coronal suture. (C) Antenna. (D) Right palpus. (E) Clypeus. (F) Thorax and legs, lateral view. (G) Wing. (H) Abdomen, dorsal view. (I) Left gonocoxite and gonostylus, dorsal view. (J) Aedeagus. control) (INPA). the Uruguay River near the tributary stream where the new species was found. Yucumã Falls and a portion of the ripar- Etymology ian forest where the new species was found are currently The specific epithet refers to Yucumã Falls, the world’s under the threat of disappearing due to hydroelectric power widest waterfall (extending about 1800 meters), located on plants planned for the Uruguay River. The name is to be Hematophagous flies and frog calls 533

Table 3. Number of sensilla coeloconica on adults of Corethrella yucuman Legs. Midleg with thick, subapical setae on each of Caldart & Pinho sp. n. at least tarsomeres 1–3. Claw of foreleg longer than those of mid-, hind leg; each claw without inner Flagellomeres I II III IV V VI VII VIII IX X XI XII XIII tooth; anterior claws of each leg without a basal Females1–2000000001111 prong; foreleg claws unequal; midleg claws equal. Males1–20000000000–111 Abdomen (Fig. 3H). Uniformly brown with 9th seg- ment darker. regarded as a noun in apposition. Genitalia (Figs. 3I–J). Gonocoxite (Fig. 3I) medium brown, about 3.0–3.5 longer than wide (at midlength); gonocoxite Diagnostic characters with well-defined dorsal row of 5 slender setae plus a more Adult female: Only extant species of Corethrella with legs, ventral thick one, located near the two basal ones, at 0.41– most of thorax and abdomen uniformly brown, wing uni- 0.49 gonocoxite’s length. Gonostylus slender, straight for formly pigmented, R2 basal to apex of M1 and coronal suture most of its extension, but curved apically; sub-basal seta absent. Adult male: Only extant species of Corethrella with apparently absent; apex pointed with somewhat stout apical legs, most of thorax and abdomen uniformly brown, wing seta, with one tiny seta close to apex. Aedeagus (Fig. 3J) uniformly pigmented, R2 basal to apex of M1 and R2 shorter slender, elongate, tapering from base, acuminate apically, than the stem of R2+3. with lateral margins fusing subapically.

Descriptions Larva and pupa. Unknown. Female adult (n = 8–10). DISCUSSION Head (Fig. 2A). Outline in anterior view nearly circular, 1.2– 1.5 times wider than long. Two large setae on frons, In this study we investigated phonotactic responses of between ventromedial area of the ommatidia. Coronal three taxa of hematophagous flies to frog calls in southern suture absent (Fig. 2B). Antenna lightly pigmented; pedicel Brazil. By broadcasting different acoustic stimuli we found without distinctive, more elongate, stout, dorsal or dorsolat- that frog-biting midges (Corethrella) were attracted to the eral setae, flagellomeres as in Fig. 2C, sensilla coeloconica calls of the common pond-breeding frog Physalaemus aff. distributed as in Table 3. Flagellomere XIII with well- gracilis, and that when compared to traps broadcasting a developed apical bifurcation. Mandible with small, pointed white noise, the calls of P. aff. gracilis attracted significantly teeth. Clypeus (Fig. 2D) squarish, with 3–5 stout setae. more frog-biting midges. In contrast, Forcipomyia midges Palpus (Fig. 2E) light brown, segment 3 swollen apically. and Uranotaenia mosquitoes were not attracted to any of Thorax (Fig. 2F). Uniformly medium to dark brown, scutel- the tested frog calls. Recent studies in tropical sites also lum darker. Posterior portion of dorsocentral row with single attracted Corethrella species with frog calls (Bernal et al., elongate setae situated somewhat lateral to one another. 2006; Grafe et al., 2008), but similar studies testing the Prescutal suture very short, not extending more than half- attraction of frog-biting midges to frog calls were lacking in way to dorsocentral row of setae. Anepimeron uniformly the austral Neotropics. Likewise, the interaction between brown with group of 2–3 setae. Anterior anepisternum frog-hosts and hematophagous Forcipomyia (e.g., Spinelli divided diagonally by sinuous suture; dorsal portion about et al., 2002) and Uranotaenia (e.g., Borkent and Belton, equal to ventral portion. Ventral portion of posterior anepis- 2006) has never been investigated in the austral Neotropics ternum not differentiated from dorsal portion, anterior margin with regards to the cue used in host-seeking behavior. Our more darkly pigmented than posterior margin, three poste- study is the first to present evidence that the advertisement rior anepisternals. calls of three frog species from subtropical Brazil (Physalae- Wing (Fig. 2G). Plain, without pattern of pigmented veins mus aff. gracilis, Crossodactylus schmidti and Scinax and/or scales; veins (other than costa and wing margin) with perereca) were not attractive to Forcipomyia midges and very slender scales, Apex of R2 basal to M1; R2 1.1–1.5 Uranotaenia mosquitoes, while the call of P. aff. gracilis was times longer than R2+3; wing length 1.00–1.13 mm. Halter as effective in attracting females of all Corethrella species reg- dark as scutellum. istered in this study. Legs. Uniformly medium brown, with only slender setae, The pattern of abundance of Corethrella at the study site lacking scales; claws of each leg equal to those of others was similar to that found in other subtropical sites, but much and equal on each leg (without inner teeth). Empodia slender. lower compared to what has been registered for tropical Abdomen. Uniformly brown; cercus as pigmented as 9th sites. For instance, Grafe et al. (2008) collected an average segment. of 37 frog-biting midges per trapping hour in Borneo, with abundance ranging from 0–268, while on Iriomote Island in Male adult: (n = 11) as female, except: subtropical Japan, Toma et al. (2005) collected a total of Head (Fig. 3A). 1.3–1.5 times wider than long. Coronal 131 female Corethrella over 10 sampling nights with frog call suture long (Fig. 3B). Flagellomeres as in Fig. 3C, sensilla traps. We collected an average of six flies per trapping hour coeloconica distributed as in Table 3. Clypeus (Fig. 3E) with broadcasting P. aff. gracilis calls, with abundance ranging 1–3 stout seta. from 1–26 frog-biting midges. Amaral and Pinho (2015) Thorax (Fig. 3F). Anepimeron with group of 2–3 setae; pos- recently reported a similar pattern of low abundance of terior anepisternals 1–3. Corethrella in another site from subtropical Brazil, where Wing (Fig. 3G). R2+3 0.6–0.8 times longer than R2; wing broadcasted calls of the common frog Leptodactylus latrans length 0.91–1.01 mm. captured no more than 20 frog-biting midges during one 534 V. M. Caldart et al. hour of trapping. The patterns of abundance of frog-biting Corethrella may also be attracted to calling frogs in order to midges in the austral Neotropics are certainly not conclusive find conspecific mates (Borkent, 2008). Our collection of given that studies are at incipient stages and have covered male individuals and male swarms along with female Core- just a few locations. However, the low abundance of thrella yucuman sp. n. (Table 2) lends additional support to Corethrella in subtropical Brazil is not unexpected, as this this claim. There is very little information about the mating taxon is most diverse and abundant in the lowland tropics behavior of frog-biting midges, but for at least one species, (Borkent, 2008), while the sampled sites in southern Brazil Corethrella appendiculata, it was reported that mating are located at the latitude 27° S (Amaral and Pinho, 2015; swarms consisted predominantly of males, which indicates present study), which is near to the austral distribution limit a male-biased lek-like mating system in this species (Silva known for Corethrella in South America (latitude 30° S in and Bernal, 2013) that is similar to the one found in other Argentina; Borkent, 2008). Moreover, patterns of richness culicomorph species. Thus, this is expected to also occur in and abundance of frog-biting midges are expected to vary Corethrella yucuman sp. n. depending on the frog species and the type of calls used as In contrast to what we found for Corethrella, Forcipomyia bait, as well as the type of habitat sampled (Borkent, 2008). midges and Uranotaenia mosquitoes were not attracted to We collected five species of Corethrella, most were col- any of the frog calls used. We did expect to collect many lected in ponds and all species were captured using the call more individuals of Uranotaenia in traps broadcasting frog of P. aff. gracilis. Interestingly, the advertisement call of a calls, as was reported in Costa Rica (Borkent and Belton, congener species, Physalaemus cuvieri, also attracted frog- 2006) and on a subtropical island in Japan (Toma et al., biting midges in southern Brazil (Amaral and Pinho, 2015). 2005). However, it is possible that Uranotaenia species Physalaemus cuvieri and P. aff. gracilis are common pond- respond to call frequencies particularly sensitive to their breeding frogs in southern Brazil, whose males form cho- antennae. For example, the antennae of Uranotaenia lowii ruses and call while floating in the water to attract mates are thought to resonate at around 400–500 Hz, which is the (Kwet et al., 2010). The advertisement calls of both species fundamental harmonic of Hyla gratiosa calls (Borkent and are spectrally similar, sharing such features as harmonic Belton, 2006). Alternatively, species of Uranotaenia may use structure, wide frequency range, dominant frequency non-acoustic cues in frog–host location, including visual located below 1 kHz, and call sound pressure level of stimuli, carbon dioxide, water vapor or other volatile frog around 69 dB (P. cuvieri: Gambale and Bastos, 2014; P. aff. skin secretions (Borkent and Belton, 2006). Thus, while gracilis: this study). Another common and apparently effec- Uranotaenia mosquitoes were not attracted to the calls of P. tive species that attracts Corethrella in southern Brazil is the aff. gracilis, C. schmidti and S. perereca, our results do not pond-breeding butter-frog, Leptodactylus latrans (Amaral preclude the existence of interactions between Uranotaenia and Pinho, 2015), the advertisement call of which also pres- and male frogs in southern Brazil. For Forcipomyia midges, ents a dominant frequency below 1 kHz (Straughan and on the other hand, data from the literature and our results Heyer, 1976). In contrast, calls of the other two frog species suggest that these midges do not use acoustic cues of frog broadcasted for attracting frog-biting midges in our study calls in host-seeking behavior. Although Forcipomyia have dominant frequencies above 3 kHz (C. schmidti: midges have been observed attacking the frogs Leptodactylus Caldart et al., 2011; S. perereca: Magrini and Giaretta, chaquensis and Pseudis paradoxa in Argentina (Spinelli et 2010) and collected fewer frog-biting midges. Moreover, the al., 2002), they may have used non-acoustic cues to find duty cycles of the broadcasted calls of C. schmidti and S. their hosts because most hosts of Forcipomyia biting perereca were lower than the duty cycle of P. aff. gracilis females are invertebrates that do not rely primarily on air- calls, though according to our recordings, such differences borne acoustic signals for communication, such as spiders, are likely close to what occurs in nature. In the study area dragonflies, damselflies and phasmids (Borkent and Spinelli, at least 15 pond-breading anuran species were recorded 2007), except for katydids (Greenfield, 2016) and some cat- during nocturnal calling activity (Iop et al., 2012), hence it is erpillars (Brown et al., 2007). While the specific cues used quite likely that interactions between Corethrella and calling for host location by Forcipomyia and Uranotaenia species frogs of other species occur. Studies broadcasting calls of remain poorly understood, further conclusion should be different frog species simultaneously would be relevant for avoided until studies about the attraction of these taxa to dif- testing the attraction of female Corethrella to different call ferent types of host cues are performed. frequencies and duty cycles, and we suggest that species of Physalaemus and Leptodactylus be further tested to answer Taxonomy, distribution and biology of Corethrella this question for subtropical Brazil. This study reports the southernmost records of C. We collected individual male Corethrella and male lopesi, C. alticola and C. atricornis, expanding the distribu- swarms in nearly all trapping sessions where females of tion of these species to the austral portion of the Neotropics, Corethrella yucuman sp. n. were caught. Female Corethrella in the southernmost Brazilian state of Rio Grande do Sul. have evolved a strategy to feed on nocturnal frogs because Corethrella lopesi was previously known from the states of most species of anurans vocalize at night and the mating Bahia, Rio de Janeiro and São Paulo (Borkent, 2008), C. swarms of Corethrella are formed near the habitats where alticola was previously known from the states of Goiás, São anurans aggregate for reproduction (Borkent, 2008; Silva Paulo (Borkent, 2008), and Santa Catarina (Amaral and and Bernal, 2013). It has recently been shown that female Pinho, 2015), and C. atricornis was previously known only Corethrella strongly depend on the mating calls emitted by from the state of São Paulo (Borkent, 2008). male frog hosts for long-distance host location (Bernal and The females of C. yucuman sp. n. key out to C. anniae Silva, 2015). However, it is known that male individuals of in Borkent (2008), but it is easily distinguished from this spe- Hematophagous flies and frog calls 535 cies based on pigmentation of katepisternum, which is com- Santos and Glauco Machado for their valuable suggestions that pletely pale in C. anniae and brown, only slightly paler than greatly improved this study, and to the staff of the Laboratório de the scutellum, in C. yucuman sp. n. Also, the prescutal Herpetologia – UFSM for their assistance during field activities. We suture is very short in C. yucuman sp. n. and elongate in C. also thank Milene Höehr de Moraes and Camille Ribeiro for taking photos of the new species. V.M.C. and S.I. thank CAPES for the anniae, and the R2 is basal to the apex of M1 in C. yucuman doctoral fellowships and S.Z.C. is grateful to CNPq for the research sp. n., while distal in C. anniae. The males key out to C. fellowship (No. 304929/2012-3). brevivena (couplet 26) because the R2 is shorter than the stem of R2+3. The new species clearly belongs to the REFERENCES rotunda group because the ventral portion of posterior Amaral AP, Pinho LC (2015) New species and records of frog-biting anepisternum is continuous with the dorsal portion. Another midges from southern Brazil (Diptera: Corethrellidae). Zootaxa synapomorphy of the rotunda group, although homoplastic 3946: 274–284 along the genus, is the wing lacking patterned pigmentation Bee MA (2004) Sound Ruler acoustical analysis: a free, open code, (Borkent, 2008). multi-platform sound analysis and graphing package. Bioacous- Corethrella yucuman sp. n. is the first species belonging tics 14: 171–178 to the rotunda group found in South America; the eight other Bernal XE, Rand AS, Ryan MJ (2006) Acoustic preferences and species are known only from a few localities in Texas (USA) localization performance of blood-sucking flies (Corethrella Coquillett) to túngara frog calls. Behav Ecol 17: 709–715 and Costa Rica, at elevations ranging from sea level to 600 Bernal XE, Silva P de (2015) Cues used in host-seeking behavior by m (Borkent, 2008). The new species was found only along frog-biting midges (Corethrella spp. Coquillet). J Vector Ecol 40: two streams in Turvo State Park (southern Brazil), at eleva- 122–128 tions between 185–382 m. Immature individuals of rotunda Borkent A (2008) The frog-biting midges of the world (Corethrellidae: group species are unknown, although unreared larvae from Diptera). Zootaxa 1804: 1–456 hyporheic zones in small streams in Texas may belong to Borkent A (2014) World catalog of extant and fossil Corethrellidae Corethrella kerrvillensis (as discussed by Borkent, 2008). (Diptera). Zootaxa 3796: 453–468 Such hyporheic zones may be the C. yucuman sp. n. larval Borkent A, Belton P (2006) Attraction of female Uranotaenia lowii habitat, as trapping sessions in ponds were not effective in (Diptera: Culicidae) to frog calls in Costa Rica. Can Entomol collecting adults of this species. The serrate mandibles of 138: 91–94 Borkent A, Grafe TU (2012) The frog-biting midges of Borneo—from the female adults and their attraction to Physalaemus aff. two to eleven species (Corethrellidae: Diptera). Zootaxa 3279: gracilis, Crossodactylus schmidti and Scinax perereca calls 1–45 suggest that they feed on frog blood in nature. Borkent A, Spinelli GR (2007) Neotropical Ceratopogonidae (Diptera: Insecta). In “Aquatic Biodiversity of Latin America, Vol. Concluding remarks 4” Ed by J Adis, JR Arias, G Rueda-Delgado, KM Watzen In this study we investigated phonotactic responses of (Series Eds.), WL Grogan (scientific reader to the eds), Pensoft, three taxa of hematophagous flies to frog calls in subtropical Sofia-Moscow, 198 pp Brazil. While Forcipomyia midges and Uranotaenia mosqui- Brown SG, Boettner GH, Yack JE (2007) Clicking caterpillars: toes were randomly captured in both frog call and silent acoustic aposematism in Antheraea polyphemus and other Bombycoidea. J Exp Biol 210: 993–1005 traps, calls of the common pond-breeding frog Physalaemus Brown BV, Borkent A, Cumming JM, Wood DM, Woodley NE, aff. gracilis attracted five species of Corethrella, one of Zumbado M (2009) Manual of Central American Diptera, which is formally described herein. Traps broadcasting a Volume 1. NRC Research Press, Ottawa, 714 pp white noise did not differ from adjacent silent traps in the Caldart VM, Iop S, Santos TG, Cechin SZ (2010) Extension of the number of captured flies, regardless of the diptera taxa or geographical distribution of two anuran species for Rio Grande the broadcasted frog call. Moreover, white noise traps do Sul State, Brazil, with comments on natural history. Biota collected fewer frog-biting midges compared to traps broad- Neotrop 10: 143–147 casting P. aff. gracilis calls, which suggests that female Caldart VM, Iop S, Cechin SZ (2011) Vocalizations of Crossodactylus Corethrella feed on males of this widely distributed pond- schmidti Gallardo, 1961 (Anura, Hylodidae): advertisement call breeding frog in southern Brazil. We emphasize the impor- and aggressive call. North West J Zool 7: 118–124 Caldart VM, Iop S, Cechin SZ (2014) Social interactions in a neotro- tance of using frog call traps to reveal the unknown diversity pical stream frog reveal a complex repertoire of visual signals of acoustically-oriented frog-biting midges of the austral and the use of multimodal communication. Behaviour 151: 719– Neotropics, especially Corethrella. We also draw attention to 739 the need for studies that test the role of non-acoustic cues Caldart VM, Iop S, Lingnau R, Cechin SZ (2016) Calling activity of a of frog hosts in the host-seeking behavior of Forcipomyia stream-breeding frog from the austral Neotropics: temporal pat- midges and Uranotaenia mosquitoes. Studies about the terns of activity and the role of environmental factors. Herpeto- interaction between hematophagous flies and frog hosts are logica 72: 90–97 needed to better understand the cues used by these flies to Ferguson LV, Smith TG (2012) Reciprocal trophic interactions and locate hosts, the degree of host specificity, and the evolution transmission of blood parasites between mosquitoes and frogs. of adaptations in frog hosts and hematophagous flies. 3: 410–423 Frost DR (2015) species of the world: an online refer- ACKNOWLEDGMENTS ence. Version 6.0. In: Am. Museum Nat. Hist. New York. http:// research.amnh.org/herpetology/amphibia/index.html We are grateful to Art Borkent and Ulmar Grafe for kindly pro- Gambale PG, Bastos RP (2014) Vocal repertoire and bioacoustic viding information about the interaction between hematophagous analyses in Physalaemus cuvieri (Anura, Leptodactylidae) from flies and frog hosts and the identification of the dipteran taxa, to southern Brazil. Herp Journal 24: 31–39 Camila Both, Cynthia Prado, Selvino Neckel de Oliveira, Sandro Grafe TU, Saat HBM, Hagen N, Kaluza B, Berudin ZBH, Bin MA, 536 V. M. Caldart et al.

Wahab A (2008) Acoustic localisation of frog hosts by blood- and Seasonally Dry Forests: Plant Diversity, Biogeography, and sucking flies Corethrella Coquillet (Diptera: Corethrellidae) in Conservation” Ed by RT Pennington, GP Lewis, J Ratter, Taylor Borneo. Aust J Entomol 47: 350–354 & Francis/CRC Press, Boca Raton, FL, pp 151–184 Greenfield MD (2016) Evolution of acoustic communication in Peake TM (2005) Eavesdropping in communication networks. In insects. In “ Hearing” Ed by GS Pollack, AC Mason, AN “Animal Communication Networks” Ed by PK McGregor, Popper, RR Fay, Springer Handbook of Auditory Research 55, Cambridge University Press, Cambridge, pp 13–37 Springer International Publishing, Switzerland, pp 17–47 Ryan MJ (1980) Female mate choice in a neotropical frog. Science Gridi-Papp M (2004) SoundRuler: acoustic analysis for research 209: 523–525 and teaching. http://soundruler.sourceforge.net Sæther OA (1969) Some Nearctic Podonominae, Diamesinae, and Halfwerk W, Jones PL, Taylor RC, Ryan MJ, Page RA (2014) Risky Orthocladiinae (Diptera: Chironomidae). Bull Fish Res Board ripples allow bats and frogs to eavesdrop on a multisensory Canada 170: 1–154 sexual display. Science 343: 413–416 SEMA Secretaria Estadual do Meio Ambiente (2005) Plano de Hammer Ø, Harper DAT, Ryan PD (2001) PAST: Paleontological manejo do Parque Estadual do Turvo. Estado do Rio Grande statistics software package for education and data analysis. do Sul: [s.n.], 348 p Paleontol Electron 4: 1–9 Silva P de, Bernal XE (2013) First report of the mating behavior of a Iop S, Caldart VM, Santos TG, Cechin SZ (2012) What is the role of species of frog-biting (Diptera: Corethrellidae). Florida heterogeneity and spatial autocorrelation of ponds in the orga- Entomol 96: 1522–1529 nization of frog communities in southern Brazil. Zool Stud 51: Spinelli GR, Schaefer EF, Kehr AI (2002) First record of biting 1094–1104 midges (Diptera: Ceratopogonidae) attacking frogs in the Kwet A, Lingnau R, Di-Bernardo M (2010) Pró-Mata: Anfíbios da Neotropical region. Proc Entomol Soc Washingt 104: 528–529 Serra Gaúcha, sul do Brasil, 2nd edn. Brasilien Zentrum der Straughan IR, Heyer WR (1976) A functional analysis of the mating Universität Tübingen, Tübingen calls of the Neotropical frog genera of the Leptodactylus com- Magrini L, Giaretta AA (2010) Calls of two Brazilian species of plex (Amphibia, Leptodactylidae). Pap Avul Zool 29: 221–245 Scinax of the S. ruber clade (Anura: Hylidae). Herp Notes 3: Thielman AC, Hunter FF (2007) Photographic key to the adult 121–126 female mosquitoes of Canada (Diptera: Culicidae). Can J Maluf JRT (2000) Nova classificação climática do Estado do Rio Identif 4: 1–117 Grande do Sul. Rev Bras Agrometeorol 8: 141–150 Toma T, Miyagi I, Higa Y, Okazawa T, Sasaki H (2005) Culicid and Marino PI, Spinelli GR (2008) Biting midges of the Forcipomyia chaoborid flies (Diptera: Culicidae and Chaoboridae) attracted (Forcipomyia) argenteola group in southern South America, to a CDC miniature frog call trap at Iriomote Island, the Ryukyu with description of a new species and a key to the neotropical Archipelago, Japan. Med Entomol Zool 56: 65–71 species (Diptera: Ceratopogonidae). Rev Biol Trop 56: 789–94 Tuttle MD, Ryan MJ (1981) Bat predation and the evolution of frog McKeever S (1977) Observation of Corethrella feeding on tree frogs vocalizations in the Neotropics. Science (80- ) 214: 677–678 (Hyla). Mosq News 37: 522–523 Yu Y-X, Huang E-J, Zhang J-Q (2013) A new species of Corethrellidae McKeever S, Hartberg WK (1980) An effective method for trapping in China-Corethrella feipengi sp. nov. (Diptera: Corethrellidae). adult female Corethrella (Diptera: Chaoborida). Mosq News 40: Acta Parasitol Medica Entomol Sin 20: 261–264 (in Chinese 111–112 with English summary) McKeever S, French FE (1991) Corethrella (Diptera, Corethrellidae) Zuk M, Kolluru G (1998) Exploitation of sexual signals by predators of eastern North-America –Laboratory life history and field and parasitoids. Q Rev Biol 73: 415–438 responses to anuran calls. Ann Entomol Soc America 84: 493– Zuk M, Rotenberry JT, Tinghitella RM (2006) Silent night: adaptive 497 disappearance of a sexual signal in a parasitized population of Oliveira-Filho AT, Jarenkow JA, Rodal MJN (2006) Floristic relation- field crickets. Biol Lett 2: 521–5244 ships of seasonally dry forests of eastern South America based on tree species distribution patterns. In “Neotropical Savannas (Received October 20, 2015 / Accepted April 25, 2016)