Zootaxa 3985 (2): 219–251 ISSN 1175-5326 (print edition) Article ZOOTAXA www.mapress.com/zootaxa/ Copyright © 2015 Magnolia Press ISSN 1175-5334 (online edition) http://dx.doi.org/10.11646/zootaxa.3985.2.3 http://zoobank.org/urn:lsid:zoobank.org:pub:E353CF74-4A45-44B5-A735-B8110D9B8608 Molecular phylogenetics, diversification, and systematics of Tibicen Latreille 1825 and allied cicadas of the tribe Cryptotympanini, with three new genera and emphasis on species from the USA and Canada (Hemiptera: Auchenorrhyncha: Cicadidae) KATHY B. R. HILL1, DAVID C. MARSHALL1,3, MAXWELL S. MOULDS2 & CHRIS SIMON1 1Dept. of Ecology and Evolutionary Biology, University of Connecticut, 75 N. Eagleville Rd., Storrs, CT 06269 USA 2Entomology Dept., Australian Museum, 6 College Street, Sydney N.S.W. 2010. Email: [email protected] 3Corresponding author. E-mail: [email protected] Abstract North America has a diverse cicada fauna with multiple genera from all three Cicadidae subfamilies, yet molecular phy- logenetic analyses have been completed only for the well-studied periodical cicadas (Magicicada Davis). The genus Tibi- cen Latreille, a large group of charismatic species, is in need of such work because morphological patterns suggest multiple groups with complicated relationships to other genera in the tribe Cryptotympanini. In this paper we present a molecular phylogenetic analysis, based on mitochondrial and nuclear DNA, of 35 of the 38 extant USA species and sub- species of the genus Tibicen together with their North American tribal allies (Cornuplura Davis, Cacama Davis), selected Tibicen species from Eurasia, and representatives of other Eurasian and Pacific cryptotympanine genera. This tree shows that Tibicen contains several well-supported clades, one predominating in eastern and central North America and related to Cryptotympana Stål and Raiateana Boulard, another in western North America related to Cacama and Cornuplura, and at least two clades in Eurasia. We also present a morphological cladistic analysis of Tibicen and its close allies based on 27 characters. Character states identified in the cladistic analysis define three new genera, two for North American taxa (Hadoa gen. n. and Neotibicen gen. n.) including several Mexican species, and one for Asian species (Subsolanus gen. n.). Using relaxed molecular clocks and literature-derived mtDNA rate estimates, we estimate the timeframe of diversification of Tibicen clades and find that intergeneric divergence has occurred since the late Eocene, with most extant species within the former Tibicen originating after the mid-Miocene. We review patterns of ecology, behavior, and geography among Tibicen clades in light of the phylogenetic results and note that the study of these insects is still in its early stages. Some Mexican species formerly placed in Tibicen are here transferred to Diceroprocta, following refinement of the definition of that genus. Key words: evolution, molecular genetics, cladistics, molecular clock, biogeography, disjunction, annual cicada, numt, Cicadinae Introduction Cicadas (Auchenorrhyncha: Cicadidae) are large, xylem-feeding insects known for their long underground juvenile life stages and the loud, species-specific songs made by males during their brief aboveground adult lives. North America has a diverse cicada fauna that includes multiple genera from each of the three cicada subfamilies (Cicadettinae, Tibicininae [=Tettigadinae], and Cicadinae—see Moulds 2005). While the extraordinary periodical cicadas (Magicicada Davis, seven spp.) have been extensively studied for nearly two centuries, including molecular and morphological systematic analyses (Marlatt 1907; Simon 1979;1983; Sota et al. 2013; Williams & Simon 1995), the 185 non-periodical species and subspecies north of Mexico have received only sporadic attention and little to no phylogenetic analysis beyond alpha taxonomy and a role as outgroups in other studies (e.g., Moulds 2005; Sueur et al. 2007). Accepted by H. Duffels: 8 Jun. 2015; published: 10 Jul. 2015 219 Licensed under a Creative Commons Attribution License http://creativecommons.org/licenses/by/3.0 TABLE 1. Life history of USA cryptotympanine cicada species, summarized mainly from Sanborn and Phillips (2013), Sanborn and Heath (2012), publications of W. T. Davis, and observations of the authors (esp. song characterizations). Taxa with asterisks were not collected in time for this study. Taxonomic authorities are given in the Appendix; Tibicen tibicen = Tibicen chloromerus (Walker), and Tibicen pronotalis walkeri = Tibicen walkeri (Metcalf) (see Hill & Marshall 2009). New assignments refer to genera described in the Results section. Recordings of the songs of most USA cryptotympanine species can be found at www.insectsingers.com. New genus (see Results) Genus Species Distribution Habitat, hostplants Song description Miscellaneous Cacama californica* Unchanged S CA, NV Upl. woodland, Opuntia Unknown to this study Similar to variegata Cacama collinaplaga* Unchanged TX Des. grassland, Opuntia Resonant oscillating whine No red/orange at wing bases Cacama crepitans* Unchanged S CA Upl. woodland, Opuntia, sage Unknown to this study Localized distribution Cacama moorei Unchanged AZ, NV Sonoran Des., Opuntia Resonant oscillating whine Compare to collinaplaga Cacama valvata Unchanged SW USA, Mex. Des., chaparral, Opuntia Resonant oscillating whine Most widespread Cacama Cacama variegata* Unchanged TX Des. grassland, Opuntia Unknown to this study Most evenly rounded forewings Cornuplura nigroalbata Unchanged SW AZ, Mex. Des. mountains, Quercus Continuous, oscillating whine Santa Cruz Co., AZ Tibicen auletes Neotibicen E USA, Can. Forest, Hardwoods, Quercus Slow harsh drr-drr-drr Largest USA species Tibicen auriferus Neotibicen C USA Prairie, shrubs/grasses Brassy whine Similar to davisi Tibicen bifidus Hadoa SW USA Des. grasslands, Yucca, Artemisia Brassy buzz Bifurcate uncus, see simplex Tibicen canicularis Neotibicen NE/NC USA, Can. Woodland, conifers, hardwoods Sharp whine For some, the "Dog-day Cicada" Tibicen chiricahua Hadoa AZ, NM, Mex. Woodland, Pinus, Juniperus Smooth resonant drone Compare to neomexicensis Tibicen chisosensis* Hadoa S TX, Mex. Mtn. woodland, pinyon pine? Unknown to this study Only in Chisos Mtns in USA Tibicen cultriformis Neotibicen AZ, NM, Mex. Rip. woodland, Salix, Populus Rapid staccato pulses Related to east-central Tibicen Tibicen davisi davisi Neotibicen SE USA Woodland, conifers Brassy whine See auriferus and canicularis Tibicen davisi harnedi Neotibicen SC USA Woodland Brassy whine No black stripe on ventral abd. Tibicen dealbatus Neotibicen G. Plains Prairie, hardwoods Droning zzh-zzh-zzh Song same as pronotalis Tibicen dorsatus Neotibicen C USA Prairie, shrubs/grasses Tractor-like rattle Similar to tremulus Tibicen duryi Hadoa SW USA, Mex. Mtn. woodland, chaparral, Pinus Complicated warbling buzz Locally abundant Tibicen figuratus Neotibicen SE USA Woodland, conifers Rough ding-ding-ding Calls infrequently Tibicen inauditus Hadoa SW USA Des. woodland/grassland High-pitched trilling buzz Smallest USA Tibicen? Tibicen latifasciatus Neotibicen Coastal NJ-SC/FL Woodland, cedar Wee-ooo oscillation White lateral marks on abdomen Tibicen linnei Neotibicen E USA, Can. Woodland, hardwoods Maraca-like oscillating buzz Similar to pruinosus Tibicen longioperculus Hadoa SE AZ Des. woodland, Juniperus Brassy buzz Unusually long opercula Tibicen lyricen engelhardti* Neotibicen Appalachian Mts. Woodland, hardwoods Watery, resonant drone Very dark coloration Tibicen lyricen lyricen Neotibicen E USA, Can. Woodland, hardwoods Watery, resonant drone Common Tibicen of the east Tibicen lyricen virescens Neotibicen SE USA Woodland, hardwoods Watery, resonant drone Greener form of lyricen Tibicen neomexicensis Hadoa AZ, NM Woodland, Pinus, Juniperus Pulsed resonant drone Compare to chiricahua Tibicen parallelus Hadoa AZ, NM, Mex. Des. habitats, Quercus, Juniperus Oscillating whine, then static NM type location isolated Tibicen pronotalis Neotibicen N G. Plains Rip. woodland, Salix, Populus? Droning zzh-zzh-zzh Pronotal mark inconsistent Tibicen pronotalis walkeri Neotibicen C USA Rip. woodland, Populus, hardwoods Droning zzh-zzh-zzh Song same as dealbatus Tibicen pruinosus fulvus* Neotibicen SE KS, NE OK Open woodland, hardwoods Wee-ooo oscillation Local color variant Tibicen pruinosus pruinosus Neotibicen C USA Open woodland, hardwoods Wee-ooo oscillation The midwestern Tibicen Tibicen resh Neotibicen SC USA, Mex. Open, rip. woodland, hardwoods Fast harsh drr-drr-drr Named for mesonotal marks Tibicen resonans Neotibicen SE USA Woodland, conifers Fast harsh dee-dee-dee Large-bodied Tibicen robinsonianus Neotibicen EC USA Woodland, hardwoods, Juniperus Slow (ca. 1/s) repeated rasps High in canopy Tibicen similaris Neotibicen SE USA Woodland, conifers, hardwoods Rattle w sudden acceleration Unusual clacking rattle Tibicen simplex Hadoa AZ, Mex. Desert, Yucca Brassy whine Like bifidus but simple uncus Tibicen superbus Neotibicen SC USA Open woodland, hardwoods Sputtery, rapid cha-cha-cha Green with sputtery song Tibicen texanus Hadoa TX, OK, NM Open woodland, Quercus, Juniperus Brassy buzz Attractive color pattern Tibicen tibicen australis Neotibicen SE USA Woodland, hardwoods Clacky warbling buzz Greener form of tibicen Tibicen tibicen tibicen Neotibicen E USA, Can. Swampy woodland, hardwoods Clacky warbling buzz Long opercula, waxy Tibicen townsendii Hadoa SW USA Desert, Yucca Whining buzz Similar to bifida/simplex Tibicen tremulus Neotibicen C USA Prairie, sage, Yucca Tractor-like rattle Similar to dorsatus Tibicen winnemanna Neotibicen E USA Woodland, hardwoods Wee-ooo oscillation Costa not as bent as pruinosus 220 · Zootaxa 3985 (2) © 2015 Magnolia Press HILL ET AL. One of the largest North American groups in need of phylogenetic study is the genus Tibicen Latreille and its allies (Sanborn & Heath 2012; Sanborn & Phillips 2013). With 38 USA species and subspecies (Table 1), Tibicen extends from the Atlantic to Arizona, north into Canada, and south into Mexico and Central America (Sanborn 2010). Substantial numbers of Tibicen species are found across Eurasia as well, although some of these are commonly referenced under the genus Lyristes Horváth (Sanborn 2013; 2014a p. 116) (see below). Tibicen belongs to the tribe Cryptotympanini Handlirsch, which in addition to other world genera includes the southwestern North American Cacama Davis ("cactus dodgers"—see Davis 1919; Sanborn et al. 2011) and Cornuplura Davis (see Sanborn & Phillips 2012). Like most cryptotympanine cicadas, Tibicen are large and charismatic insects, and the males become especially active in hot weather, hence the common name “Dog Day Cicada” informally used for Tibicen canicularis (Harris, 1841). Cicada songs allow identification of most species in the field, even many that are morphologically cryptic (e.g., Alexander et al. 1972; Cole 2008). In the USA, Tibicen species are found in a wide range of habitats from western intermountain deserts to prairie to humid deciduous forests. They are active as adults mainly in summer from June to September in the eastern and central states (e.g., Beamer 1928 p. 173; Walker 2000) and from spring in the western states. Life cycle lengths for Tibicen species are unknown but probably involve multiple juvenile years underground, as observed in cicadas with known life cycles (Campbell et al. 2015; Karban 1986). Unlike the periodical cicadas, which can damage fruit and nursery crops, Tibicen species are rarely of economic significance (e.g., Wilson 1930). Distributions, habitat associations, and song characteristics for all of the cryptotympanine species north of Mexico are summarized in Table 1 and reviewed further in Sanborn and Phillips (2013). Life history information is also summarized in the publications of Beamer (1928), Myers (1929), and Heath (1978). Tibicen is of additional interest for a phylogenetic study because its North American members exhibit patterns of morphology and ecology that suggest deeper divisions and potentially complex relationships with other cryptotympanine genera, and a “lack of diagnostic characters” has inhibited progress (Heath 1978 p. 190). Davis (1930) proposed three geographically correlated subgroups based on morphology (Davis 1930; see also Heath 1978), and Heath (1978, p. 204) proposed two invasions of North America by lineages containing Tibicen species. However, slightly different arrangements have been proposed based on unpublished molecular data (Sanborn & Heath 2012). Fukuda et al. (2006) used mtDNA sequences to identify a relationship between Tibicen japonicus (Kato, 1925) from Japan and two species of Cryptotympana Stål, but no other Tibicen species were examined. Relationships to other world Cryptotympanini remain unknown, although Davis (1930) suggested a connection between western North America Tibicen and Tibicen plebejus (Scopoli, 1763). In this paper, we examine the phylogenetic relationships of North American Tibicen found north of Mexico using genetic data, conduct a cladistic examination of morphological traits in order to identify new genera, explore the varied relationships of the North American Tibicen species to allied Tibicen and other cryptotympanine genera, and approximate the timeframe of divergence of the group. We also discuss the ecological and behavioral attributes of the North American species in light of our phylogenetic findings. The new genera are described in the Results and applied through the remainder of the paper. The status of the genus Tibicen Latreille, and its potential priority over Lyristes (both potentially claiming the type Cicada plebeja Scopoli, 1763), is currently being considered by the ICZN in recently resurrected Case 239 (Boulard & Puissant 2014; Hamilton 2014; Marshall & Hill 2014; Sanborn 2014a). While Tibicen has been overwhelmingly applied in North America, some Eurasian species (notably plebeja) have been more commonly referenced under Lyristes in recent decades. For this paper, we follow the catalogue of Sanborn (2013) and use Tibicen pending the decision by the Commission. Methods Background and taxon sampling. North American Tibicen north of Mexico. Thirty-five of the 38 described species and subspecies of North American Tibicen found in the USA and Canada were collected for this project. Some of these taxa extend to Mexico (Table 1). One USA species, T. chisosensis Davis, 1934 is found only in the Chisos Mountains in southern Texas and Mexico (Sanborn 2007) and we were unable to obtain material in time for this study. Two subspecies were also not included, T. pruinosus fulvus Beamer, 1924 and T. lyricen engelhardti (Davis, 1910). Specimen collection data and taxonomic authorities are given in the Appendix. Most of the USA PHYLOGENETICS AND SYSTEMATICS OF TIBICEN CICADAS Zootaxa 3985 (2) © 2015 Magnolia Press · 221 specimens were collected by KH and DM, and when possible the male songs were recorded. Example recordings can be found at insectsingers.com and cicadamania.com. For most of these species, 2–4 specimens were sequenced in order to sample intraspecific geographic-genetic variation. Through the efforts of collaborators (see Acknowledgements), we were able to include the type of Tibicen (plebejus) and its relative T. gemellus (see Boulard 1988). In addition, three Asian species were included—T. bihamatus (de Motschulsky, 1861), T. japonicus (Kato, 1925), and T. kyushyuensis (Kato, 1926). Nineteen additional species-level Tibicen taxa are found only in Eurasia and 13 in Mexico (Sanborn 2007). Note that T. bermudianus (Verrill, 1902) from Bermuda, morphologically and acoustically similar to T. lyricen (see Moore 1993), is believed to be extinct (Procter & Fleming 1999), and T. occidentis (Walker, 1850) from Chile has been removed from the genus and tribe (Sanborn 2014b). All of the specimens used in this study are housed at the University of Connecticut (Biological Collections Facility or C. Simon laboratory) or in the collection of Max Moulds, Australian Museum. Allied cryptotympanine genera and outgroups. The tribe Cryptotympanini was recently redefined by Moulds (2005). Twenty cryptotympanine genera are currently catalogued by Sanborn (2013), four of which are found in North America: Tibicen, Cacama, Cornuplura, and Diceroprocta Stål. An ongoing family-level molecular study of cicadas (Marshall, Hill, Wade, Owen, Moulds, Simon in prep.) suggests that a redefined concept will be necessary for Cryptotympanini and that Diceroprocta, Oriallela Metcalf, and the Australian genera will need to be removed. Based on this work we selected the following genera to be represented along with Tibicen in the molecular and cladistic analyses: Cacama, Chremistica Stål, Cornuplura, Cryptotympana, Raiateana Boulard, Salvazana Distant, and Tacua Amyot & Audinet-Serville. Cryptotympana is a very large genus (Hayashi 1987a;b), but we were able to sample five species, including the type C. atrata (Fabricius, 1775). Unfortunately, the type species of Cacama and Cornuplura are found only in Mexico and were not sampled. Six species of Chremistica were included. We were unable to obtain specimens of Antankaria Distant, Heteropsaltria Jacobi, and Nggeliana Boulard for molecular analysis. Lastly, two more distant outgroups from the subfamily Cicadinae—Platypleura takasagona Matsumura, 1917 and Yanga andriana (Distant, 1899) (tribe Platypleurini Schmidt), were included in order to root the tree. These were also chosen based on the family-level preliminary analyses. Specimen collection. Upon collection, either the entire specimen or some specimen tissue (1–3 legs) was frozen in 100% ethanol in individually labeled containers. In a few cases, a leg was removed from a dried, pinned specimen for DNA extraction. Most specimens and their associated tissue samples, recordings, or photographs were given an eleven-character collection code—two digits for the year, two letters for the country, two letters for the state/district, three letters for the location, and two digits for the individual specimen number (see Appendix). For most specimens collected by collaborators (see below), the original codes were used. Bodies of specimens with tissue separately preserved were pinned and deposited in the authors’ collection at the University of Connecticut. One exception, specimen number 09.US.CT.AUL.EB1, was a sample of eggs from an eggnest deposited by a female Tibicen auletes. Male songs were recorded in the field using one of several digital recorder/condenser microphone combinations, often together with a Sony (Park Ridge, NJ, USA) PBR330 parabolic reflector. Recorders used included the Sony TCD-D8 (2002 and 2003 only), Marantz (Mahwah, NJ, USA) PMD660, Marantz PMD670, and the Zoom (Ronkonkoma, NY, USA) H4 (in 2012 only), while the microphones used included the Sennheiser (Old Lyme, CT, USA) ME66 short shot gun and a Sennheiser ME62 omnidirectional. The latter microphones were used with the Sennheiser K6 power module, and both have frequency responses up to 18 kHz. All songs were digitized at either 44.1 kHz or 48 kHz. DNA extraction, amplification and sequencing. DNA was extracted from leg muscle using a Qiagen DNeasy Blood and Tissue Kit (Qiagen Inc., Valencia, California, USA) following the manufacturers’ instructions but with a proteinase K digestion time of 18h at 54°C. Standard polymerase chain reaction (PCR) methods were used to amplify two portions of DNA using an Ex TaqTM kit (Takara Bio Inc., Otsu, Shiga, Japan): approximately 800 bp of the nuclear elongation factor 1 α (EF-1α gene using the primers EF1-PA-f650ambig (Lee & Hill 2010) and EF-N-1419 (Sueur et al. 2007) and an annealing temperature of 58°C, and approximately 1500 bp of the mitochondrial cytochrome oxidase subunit I (COI) gene. COI was amplified using either (1) primers C1-J-1490 (Folmer et al. 1994) and TL2-N-3014 (Simon et al. 1994) and an annealing temperature of 45°C, yielding the entire segment, or (2) in two sections with TibCOIRev (5’ CCTCTTTCYTGHGTAATAATGTRTG 3’ and C1-J- 222 · Zootaxa 3985 (2) © 2015 Magnolia Press HILL ET AL. 1490 at an annealing temperature of 45°C for the first half and C1-J-2195 (Simon et al. 1994) with TL2-N-3014 at an annealing temperature of 56°C. An additional primer, TibCOI_INTREV (5’ TAYCARTGAAYAAATCTDCC 3’), was occasionally used in the sequencing phase below when the longer segment was amplified in PCR, in order to improve resolution of the middle region. These alternatives were used as necessary to amplify some problematic individuals exhibiting double-peaked chromatograms indicating possible numts (nuclear copies of mitochondrial DNA), which are often observed in studies using the barcoding region of COI (Buhay 2009). Some true mitochondrial copies were also obtained with the TOPO TA Cloning Kit (Life Technologies, Grand Island, NY, USA). PCR products were visualised on a 1% agarose gel (BP1356-500 agarose Fisher Scientific, Pittsburgh, PA, USA) prior to clean up and purified using ExoSAP-IT (USB Corp., Cleveland, Ohio, USA). EF-1α PCRs that amplified two bands were separated on a 1.5% agarose gel, cut out and purified using the Clontech Extract II kit (Clontech, Mountain View, California, USA). Purified PCR products were cycle sequenced in both directions using a standard cycle sequencing protocol (with Bigdye v1.1, Applied Biosystems, Foster City, California, USA), and then sequenced on an ABI 3100 capillary sequencer with ABI Prism Sequencing Analysis 3.7 software (Applied Biosystems). All DNA fragments were sequenced in both the 5’ and 3’ directions. Sequencher (Gene Codes Corporation, Ann Arbor, Michigan, USA) software was used to edit the raw sequence data, and the final alignment was performed by eye in MacClade 4.0 (Maddison & Maddison 2000). Rare polymorphic sites within nuclear DNA were coded as ambiguities. Uncorrected pairwise sequence divergences were calculated using Paup* v4.0b (Swofford 1998). DNA data processing and model selection. The mitochondrial DNA was divided a priori into subsets corresponding to the three codon-positions. For the nuclear EF-1α data, the coding and noncoding sites were grouped into separate subsets (keeping the small amount of coding data in one subset). Finally, a subset of binary indel characters was constructed from the insertion-deletion patterns of the aligned noncoding EF-1α data, using SeqState v1.0 (Müller 2005) and the “simple” coding scheme of Simmons and Ochoterena (2000). This data- partition model was compared with various simpler schemes using the “greedy” search algorithm of PartitionFinder (Lanfear et al. 2012), with all models evaluated for the maximum-likelihood runs, and with the “mrbayes” model set chosen for the MrBayes and BEAST analyses (see below). This yielded a partition scheme for the analysis using combined mtDNA and EF-1α data. For analyses of the mtDNA and EF-1α separately, and for analyses of the pared-down taxon set used in the BEAST analysis (see below), PartitionFinder was run using the corresponding subsets from above to check for any changes in the partition scheme. PartitionFinder also selected a best-fitting nucleotide substitution model for each data subset using the AIC. Gamma distributions were estimated with four rate categories. Indels were modeled using the Mkv model (Lewis 2001). DNA data were tested for nucleotide bias using Paup* v4.0, both whole genes and individual partition sets as necessary following the PartitionFinder results. Molecular phylogenetic analyses. Maximum-likelihood (ML) analyses of mtDNA only, EF-1α data only, and EF-1α + mtDNA combined were conducted using Garli v2.0 (Zwickl 2006). Heuristic searches to find ML trees were conducted on a Macintosh Macbook Pro, while bootstrapping analyses were conducted on the Univ. of California CIPRES biocomputing cluster. To search for best ML trees, the following settings were used: genthreshfortopoterm = 106, scorethreshforterm=10-2, significanttopochange=10-2, collapsebranches=1, linkmodels=0, subsetspecificrates=1. For each ML analysis, ten heuristic searches were conducted (searchreps=10). Default settings were used for all other Garli options. Two hundred bootstrap replicates were completed for each analysis using the above settings, except that five heuristic searches were conducted for each replicate. A phylogenetic tree for each analysis was also estimated using MrBayes v3.2.2 (Huelsenbeck & Ronquist 2001; Ronquist & Huelsenbeck 2003), in order to provide an independent estimate of topology. For each analysis, four (default) chains were run until the average standard deviation of split frequencies dropped below 0.01, after which the first 25% of the samples were used as burnin (stoprule=yes, stopval=0.01). All model parameters were separately estimated (unlinked) across partitions, including the relative partition rates (ratepr=variable). The prior on branch lengths was set to unconstrained:exponential(100), an exponential distribution with a mean of 0.01 substitutions/site. Samples were taken every 10,000 generations. Default settings (e.g., two independent runs per analysis, four chains per run) were used otherwise. Morphological and cladistic analysis. Data for 27 characters were derived from adult morphology (one head, five thorax, one foreleg, five wing, one abdomen, and 14 genitalic characters) (Table 2). Character terminology PHYLOGENETICS AND SYSTEMATICS OF TIBICEN CICADAS Zootaxa 3985 (2) © 2015 Magnolia Press · 223 follows Moulds (2005). Eighteen characters are binary and nine are multistate (Table 3). The multistate characters were treated as unordered, and character polarity was determined using the outgroup species Chremistica ochracea (Walker, 1850), with more distantly related Asian species from the molecular analysis excluded from the morphological study. Neither character weighting nor successive weighting was employed. Figure 1 illustrates the ten states for character #21, the shape of the uncus. FIGURE 1. Character states used for the uncus, character #21 (see Table 2). Heuristic parsimony trees were prepared using Hennig86 and CLADOS version 1.2 (Nixon 1992) with TBR + RAS=10 and the MULPARS option, and parsimony bootstrap support was assessed with PAUP* version 4.0b10 using default parameters and 1,000 pseudoreplicates (Swofford 1998). Character numbers were adjusted to begin at '1', rather than the 'zero' default used in the first two programs. TABLE 2. Morphological character descriptions. 1. Head size: (0) wide, wider than thorax between wings; (1) about as wide or narrower than thorax between wings. 2. Cruciform elevation: (0) very wide, much wider than long; (1) narrow, about as wide as long. 3. Cruciform elevation: (0) depressed diminishingly between anterior arms to extremities of anterior arms; (1) depressed between anterior arms only adjacent to body of cruciform elevation. 4. Basisternum 3: (0) flat, undeveloped; (1) very slightly raised; (2) bulbous; (3) bulbous and protruding distally beyond base of hind coxae. 5. Opercula of male: (0) short, terminating at about rim of tympanal cavity; (1) protruding past rim of tympanal cavity. 6. Fore leg femoral primary spine: (0) erect; (1) lying flat, prostrate. 7. Wing margin: (0) medium width; (1) very broad, on hind wing wider than any of apical cells 3, 4 or 5. 8. Fore wing vein M before branching: (0) very long, approximately half length of discal cell; (1) shorter than half length of discal cell. 9. Fore wing vein M where forming margin of discal cell: (0) shorter or about as long as one other inner margin vein of 1+2 discal cell; (1) extremely long, far longer than the other two veins combined. 10. Hind wing apical cells: (0) six; (1) five. NOTE: Species normally with 5 hind wing apical cells usually have a small number of individuals with 4 or even 6 apical cells (usually only in one wing); such species are considered to have 5 apical cells and those minority individuals with 4 or 6 are regarded as abnormal. 11. Hind wing 2nd cubital cell width at distal end: (0) greater than 1st anal cell; (1) less than 1st anal cell. Note: Some species with these cells about equal have some individuals with the cubital larger than the anal and some vice versa; such species are scored as '?'. 12. Timbal covers, basal part of inner margin: (0) convex; (1) concave. 13. Sternite VIII: (0) U-shaped in cross-section, much less than maximum width of sternite VII; (1) V-shaped in cross-section, 224 · Zootaxa 3985 (2) © 2015 Magnolia Press HILL ET AL. much less than maximum width of sternite VII; (2) V-shaped in cross-section, as wide as or almost as wide as maximum width of sternite VII. 14. Pygofer shape in ventral view: (0) more or less parallel-sided; (1) tapering towards top and bottom, coffin-shaped; (2) considerably broadened towards the top, tending triangular in shape. 15. Pygofer dorsal beak: (0) well developed, acute and spine-like; (1) absent or poorly developed, with basal portion broadly rounded. 16. Pygofer distal shoulder: (0) straight or turned inwards; (1) turned outwards. 17. Pygofer distal shoulder: (0) undeveloped; (1) moderately developed, broad and rounded. 18. Upper pygofer lobe: (0) moderately developed, broadly rounded, projecting; (1) absent or barely developed; (2) strongly developed into a long linear spine; (3) an ill-defined, broadly rounded, internal and lightly sclerotized lobe. 19. Pygofer basal lobe in lateral view: (0) moderate length and apically rounded; (1) very long (at least 4x longer than wide), apically pointed; (2) moderate, pointed; (3) moderate, bi-lobed. 20. Pygofer basal lobe in lateral view: (0) tucked inside, hidden; (1) exposed, either tight against pygofer margin or entirely exposed. 21. Uncal overall shape in dorsal view: (0) very short and broad, about as long as wide, tending triangular; (1) long, finger- like; (2) short, broad and laterally lobed outwards at base in dorsal view, tapering to a rounded apex; (3) short, broad at base, tapering to a distal finger-like extension with a rounded apex; (4) long, broad, wide at base but thereafter less so, arched in cross-section, apex bluntly rounded; (5) long, broad, tending parallel-sided throughout its length, arched in cross- section, apex straight in dorsal view; (6) short, very broad at base, tapering to an expanded apex that is laterally lobed, the large rounded lateral lobes down-turned; (7) short, very broad at base, tapering evenly to rounded apex; (8) short, wide at base, tapering to a distal finger-like extension bearing a pair of large ventral spikes subapically; (9) short, about as long as wide, arched in cross-section, tending triangular. 22. Aedeagal restraint: (0) by a pair of sclerotized swellings at ventral base of uncus; (1) by tubular encapsulation prior to ventral surface of uncus. 23. Aedeagal basal plate in lateral view: (0) straight, aligned with basal portion of theca; (1) distal half or so strongly bent downwards away from alignment of basal portion of theca. 24. Thecal apex: (0) without ventral subapical thorn; (1) with ventral subapical thorn. 25. Theca apex: (0) parallel-sided or almost so, sclerotized, with apical serrations on rim; (1) flared and lightly sclerotized, sometimes in part heavily sclerotized; (2) flared and not sclerotized; (3) parallel-sided and not sclerotized apically. 26. Theca distal half: (0) evenly curved downwards in an arc; (1) curved downwards but kinked near base of curve; (2) straight or weakly curved upwards with apical portion slightly down-turned. 27. Uncus: (0) simple; (1) bifurcate at the extremity; (2) deeply cleft. Divergence-time analysis. Few fossils are known from the tribe Cryptotympanini, and none can yet be unambiguously assigned to a node in our trees although a family-level review is underway. Therefore, literature- derived clock estimates for the mitochondrial COI gene were used to approximate a timeframe for the cryptotympanine radiation in a Bayesian relaxed-clock analysis conducted in BEAST v2.1.2 (Drummond et al. 2006; Drummond & Rambaut 2007). BEAST was used to estimate the phylogeny from the DNA dataset (minus indels) with the substitution rate for the mtDNA partition guided by a COI clock prior encompassing a range of slow to fast literature estimates (from 0.007 to 0.0175 substitutions/site/my, see Marshall et al. 2015). This procedure scales the tree according to the amount of mtDNA evolution reconstructed while fitting the likelihood model. However, phylogenetic signal for the scale of molecular substitution can be poor at large genetic distances (Brown et al. 2010; Marshall 2010; Marshall et al. 2015), and the pairwise uncorrected mtDNA distances in our cryptotympanine dataset approached 20% (see branch lengths in Fig. 3 for model-corrected distances). As a result, we removed the most distant outgroups (Platypleura, Yanga, Chremistica, Salvazana, Tacua) from the divergence time analysis and calibrated the remaining tree with the method of “relative-time scaling” used by Marshall et al. (2015) for the cicada tribe Cicadettini. In this procedure, a younger, well-sampled focal clade is chosen and its root age estimated using relaxed molecular clock analysis of the corresponding data subset as described above, while the full tree is estimated in a separate, uncalibrated BEAST analysis (i.e., with all gene rates estimated relative to one another). The node ages and confidence intervals in the larger tree are then scaled post hoc to the estimated root age of the focal clade. In our case, we selected the large subclade containing the eastern and central USA Tibicen species. The divergence time analysis was run with the mtDNA combined into a single partition (as opposed to partitioned by codon-position) to accommodate published whole-gene clock rates. Substitution models for the whole-gene partition scheme were selected using Partitionfinder (Lanfear et al. 2012); gamma distributions were estimated with four categories. Other BEAST settings were as shown in Marshall et al. (2015). BEAST was run PHYLOGENETICS AND SYSTEMATICS OF TIBICEN CICADAS Zootaxa 3985 (2) © 2015 Magnolia Press · 225 until effective sample sizes for most parameters exceeded 200 as indicated by Tracer v1.5 (Rambaut & Drummond 2007). Convergence was accelerated by assuming the monophyly of well-supported major clades from the MrBayes analyses. Table 3. Morphological character matrix. Character names and state descriptions are given in Table 2. Taxonomic authorities are given in the Appendix, and new generic assignments of the USA species are summarized in Table 1. 00000 00001 11111 11112 22222 22 12345 67890 12345 67890 12345 67 Chremistica ochracea 00000 00000 00000 00000 00000 00 Cacama valvata 11011 01001 10110 00100 00111 00 Cacama moorei 11011 01001 10110 00100 00111 00 Cornuplura nigroalbata 00011 01000 10110 00200 00112 02 Cryptotympana atrata 00130 00110 00010 00100 30102 10 Cryptotympana holsti 00130 00110 00010 00100 30102 10 Cryptotympana takasagona 00130 00100 00010 00100 30102 10 Raiateana kuruduadua 00020 00100 00010 01100 30100 10 Tibicen auletes 00020 00100 00020 11301 61100 20 Tibicen auriferus 00021 00100 00000 01321 51102 00 Tibicen bifidus 11011 01000 10110 00101 10101 01 Tibicen bihamatus 01021 10100 11101 01101 40100 10 Tibicen canicularis 00020 00100 00000 01301 51102 10 Tibicen chiricahua 11010 01000 ?0110 00101 10101 00 Tibicen cultriformis 00020 00100 00201 01311 71100 20 Tibicen davisi davisi 00021 00100 00100 01321 51102 10 Tibicen dealbatus 00020 00100 00201 01301 71100 10 Tibicen dorsatus 00020 00100 00201 01301 71100 20 Tibicen duryi 10011 00000 10110 00101 20102 00 Tibicen figuratus 00020 00100 00201 01301 71100 10 Tibicen inauditus 10010 00000 10110 00101 20102 00 Tibicen kyushyuensis 01020 10100 11111 01101 40100 10 Tibicen latifasciatus 00021 00100 00000 01301 51102 10 Tibicen linnei 00021 00100 00000 01301 51102 10 Tibicen longioperculus 10011 00000 10110 00101 10102 00 Tibicen lyricen lyricen 00021 00100 00000 01301 51102 10 Tibicen lyricen virescens 00021 00100 00000 01301 51102 10 Tibicen neomexicensis 11010 01100 10110 00101 10101 00 Tibicen parallelus 00011 00000 ?0110 00101 20101 00 Tibicen plebejus 00020 00100 00010 00100 90102 10 Tibicen pronotalis walkeri 00020 00100 00201 01301 71100 10 Tibicen pronotalis pronotalis 00020 00100 00201 01301 71100 10 Tibicen pruinosus pruinosus 00020 00100 00000 01301 51102 10 Tibicen resh 00020 00100 00221 11301 61100 20 Tibicen resonans 00020 00100 00021 11301 61100 20 Tibicen robinsonianus 00020 00100 00000 01301 51102 10 Tibicen tibicen australis 00021 00100 00000 01301 51102 10 Tibicen tibicen tibicen 00021 00100 00000 01301 51102 10 Tibicen similaris 00021 00100 00100 00331 81102 10 Tibicen superbus 00021 00100 00000 01321 51102 00 Tibicen texanus 10010 00000 10110 00101 20102 00 Tibicen tremulus 00020 00100 00201 01301 71100 20 Tibicen townsendii 10011 01100 10110 00101 20101 ?0 Tibicen winnemanna 00020 00100 00000 01301 51102 10 Results Generic descriptions. Three new genera are proposed here based on the molecular and morphological results given below, one for the mainly eastern and central North American Tibicen species, one for a mainly western 226 · Zootaxa 3985 (2) © 2015 Magnolia Press HILL ET AL. North American clade, and one for certain Asian species. We list their descriptions immediately so that the new combinations can be used consistently for the remainder of the paper and in the figures. In addition, we discuss the definition of Tibicen Latreille used in this paper. Neotibicen gen. n., Hill and Moulds Type species: Cicada canicularis Harris, 1841 Included species: auriferus (Say, 1825) comb. n., auletes (Germar, 1834) comb. n., bermudianus (Verrill, 1902) comb. n., canicularis (Harris, 1841) comb. n., cultriformis (Davis, 1915) comb. n., davisi davisi (Smith & Grossbeck, 1907) comb. n., davisi harnedi (Davis, 1918) comb. n., dealbatus (Davis, 1915) comb. n., dorsatus (Say, 1825) comb. n., figuratus (Walker, 1858) comb. n., latifasciatus (Davis, 1915) comb. n., linnei (Smith & Grossbeck,1907) comb. n., lyricen engelhardti (Davis, 1910) comb. n., lyricen lyricen (De Geer, 1773) comb. n., lyricen virescens (Davis, 1935) comb. n., pronotalis pronotalis (Davis, 1938) comb. n., pronotalis walkeri (Metcalf, 1955) comb. n., pruinosus fulvus (Beamer, 1924) comb. n., pruinosus pruinosus (Say, 1825) comb. n., resh (Haldeman, 1852) comb. n., resonans (Walker, 1850) comb. n., robinsonianus (Davis, 1922) comb. n., similaris (Smith & Grossbeck, 1907) comb. n., superbus (Fitch, 1855) comb. n., tibicen australis (Davis, 1912) comb. n., tibicen tibicen (Linnaeus, 1758) (= Tibicen chloromerus Say) comb. n., tremulus (Cole, 2008) comb. n., winnemanna (Davis, 1912) comb. n. Etymology. Combination of the Greek prefix neo meaning new and the Latin word tibicen meaning a flute player; masculine. Distribution. Generally from the Rocky Mountains east to the Atlantic Coast, extending into southern Canada and northern Mexico. Neotibicen cultriformis extends much farther into the southwest and is the only Neotibicen species to reach Arizona. The species from Bermuda (N. bermudianus) has been reported extinct (Procter & Fleming 1999). Diagnosis. Medium to large cicadas with robust bodies, variable in body color between species but mostly a mixture of black with green or dull yellow. Head including eyes wide, as wide or wider than mesonotum between wings; vertex with distance between supra-antennal plate and eye about equal to or greater than length of antennal plate. Thorax with pronotal collar width at dorsal midline equal to or less than maximum diameter of eyes; paranota moderately ampliate, no mid lateral tooth; cruciform elevation depressed diminishingly between anterior arms to extremities of anterior arms; basisternum 3 bulbous. Fore wings hyaline; with 8 apical cells; subapical cells absent; ulnar cell 3 angled to radial cell; basal cell broad, tending to be rounded; vein CuA weakly bowed so that cubital cell no larger than medial cell; veins M and CuA widely separated at basal cell; stem of vein M shorter than half length of discal cell; vein M where forming margin of discal cell shorter or about as long as one other inner 1+2 margin vein of discal cell; vein RA aligned closely with Sc for its length and not diverging in subapical region; 1 veins CuP and 1A fused in part; infuscations either present or absent, if present then overlaying veins at bases of apical cells 2 and 3; wing outer margin developed for its total length, never reduced to be contiguous with ambient vein. Hind wings with 6 apical cells; no infuscation on ambient vein; width of 1st cubital cell at distal end about equal to 2nd cubital cell or just a little larger; 2nd cubital cell width at distal end greater than 1st anal cell; anal lobe broad with vein 3A curved, long, separated from wing margin; veins RP and M fused basally; wing margin of medium width, not excessively broad. Male opercula passing rim of tympanal cavity (only just passing in some species, very long in others); overlapping. Fore leg femoral primary spine erect. Male abdomen in cross-section with sides of tergites straight or weakly convex, epipleurites reflexed to ventral surface; tergites 2 and 3 enlarged, about twice as wide as tergites 3–7; sternites IV–VII in cross-section convex; sternite VIII either V-shaped or U-shaped in cross-section, as wide as or less than maximum width of sternite VII; timbal covers present, very slightly domed, fully rounded dorsally and extending to metathorax and tightly closed, lower margin extending anteriorly from or very near auditory capsule, meeting but not overlapping opercula; timbal ribs robust; basal dome very large; timbals extended below wing bases. Male genitalia. Pygofer in ventral view either parallel-sided or broadened towards the top; pygofer with distal shoulders broad, rounded (but barely developed in similaris); upper pygofer lobes ill-defined, broadly rounded, internal and lightly sclerotized; basal lobes undivided, moderately developed, in lateral view exposed, either tight PHYLOGENETICS AND SYSTEMATICS OF TIBICEN CICADAS Zootaxa 3985 (2) © 2015 Magnolia Press · 227 against pygofer margin or entirely exposed, of moderate length and apically rounded, pointed or bi-lobed; dorsal beak either absent, poorly developed or well developed and a part of chitinized pygofer. Uncus dominant; undivided, considerably variable in shape and size, either long or short; lacking accessory spines (claspers). Aedeagus restrained by tubular encapsulation on ventral surface of uncus; basal plate in lateral view with distal half or so strongly bent downwards away from alignment of basal portion of theca; in dorsal view apical arms short, base broad and long with midline deeply furrowed; ventral rib completely fused with basal plate; junction between theca and basal plate rigid, without a 'hinge'; thecal shaft recurved basally through 180° or more, J shaped; distal half of thecal shaft either straight, curved downwards and usually kinked near base or weakly curved upwards with apical portion slightly downturned; pseudoparameres absent; thecal apex flared or parallel-sided; thecal subapical cerci absent; flabellum absent; conjunctival claws absent; vesica retractable, vesical opening apical on theca. Male reproductive system unknown. Female reproductive system ditrysian; length of accessory glands unknown. Distinguishing features. Neotibicen differs from all other cicada genera in having the following combination of attributes: male basisternum 3 is very bulbous; male timbal covers are fully rounded dorsally, extending to the metathorax and tightly closing the timbal cavity; male tergites 2 and 3 are enlarged, about twice as wide as each of tergites 3–7; aedeagal basal plate in lateral view shows the distal half or more strongly bent downwards through about 90 degrees; and the aedeagus is restrained by tubular encapsulation prior to the ventral surface of the uncus. Hadoa gen. n. Moulds Type species: Tibicen duryi Davis, 1917 Included species: bifida (Davis, 1916) comb. n., chihuahuaensis (Sanborn, 2007) comb. n., chiricahua (Davis, 1923) comb. n., chisosensis (Davis, 1934) comb. n., distanti (Metcalfe, 1963) comb. n., duryi (Davis, 1917) comb. n., fusca (Davis, 1934) comb. n., hidalgoensis (Davis, 1941) comb. n., inaudita (Davis, 1917) comb. n., longiopercula (Davis, 1926) comb. n., minor (Davis, 1934) comb. n., montezuma (Distant, 1881) comb. n., neomexicensis (Stucky, 2013) comb. n., paralleloides (Davis, 1934) comb. n., parallela (Davis, 1923) comb. n., robusta (Distant, 1881) comb. n., sugdeni (Davis, 1941) comb. n., texana (Metcalf, 1963) comb. n., townsendii (Uhler, 1905) comb. n. Etymology. Derived from a Western Apache word for “sing”; feminine. Distribution. Throughout the southwestern USA (Arizona to Texas and the central Rocky Mountains) and Mexico. Diagnosis. Large cicadas with robust bodies, variable in body colour between species but mostly a mixture of black and dull yellow. Head including eyes about as wide as (slightly wider in parallela) or narrower than thorax between wings; vertex with distance between supra-antennal plate and eye about equal to or slightly less than length of antennal plate. Thorax: pronotal collar width at dorsal midline equal to or less than maximum diameter of eyes; paranota moderately ampliate, no mid lateral tooth; cruciform elevation depressed diminishingly between anterior arms to extremities of anterior arms; basisternum 3 only very slightly raised, not bulbous. Fore wings hyaline; with 8 apical cells; subapical cells absent; ulnar cell 3 angled to radial cell; basal cell broad, tending to be rounded; vein CuA weakly bowed so that cubital cell no larger than medial cell; veins M and CuA widely separated at basal cell; stem of vein M very long (approximately half length of discal cell) except in neomexicensis and townsendii where it is shorter; vein M where forming margin of discal cell shorter or about as long as one other 1+2 inner margin vein of discal cell; vein RA aligned closely with Sc for its length and not diverging in subapical 1 region; veins CuP and 1A fused in part; infuscations either present or absent, if present overlaying veins at bases of apical cells 2 and 3; wing outer margin developed for its total length, never reduced to be contiguous with ambient vein. Hind wings with 6 apical cells; no infuscation on ambient vein; width of 1st cubital cell at distal end about equal to 2nd cubital cell or just a little larger; 2nd cubital cell width at distal end usually less than 1st anal cell; anal lobe broad with vein 3A curved, long, separated from wing margin; veins RP and M fused basally; wing margin of medium width or very broad (wider than any of apical cells 3, 4 or 5). Male opercula either terminating at about rim of tympanal cavity or extending beyond; overlapping. Fore leg femoral primary spine erect. Male abdomen in cross-section with sides of tergites straight or weakly convex, epipleurites reflexed to ventral surface; tergites 2 and 3 enlarged, wider than tergites 3–7; sternites IV–VII in cross-section convex; sternite VIII V-shaped in cross- 228 · Zootaxa 3985 (2) © 2015 Magnolia Press HILL ET AL.