Molecular Phylogenetics and Evolution 79 (2014) 433–442
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution
journal homepage: www.elsevier.com/locate/ympev
The age, ancestral distribution and radiation of Chimarra (Trichoptera: Philopotamidae) using molecular methods ⇑ Emma Wahlberg a,b, , Kjell Arne Johanson b a Department of Zoology, University of Stockholm, 10691 Stockholm, Sweden b Department of Zoology, Swedish Museum of Natural History, Box 50007, 10405 Stockholm, Sweden article info abstract
Article history: The phylogeny of Chimarra has previously been examined using morphological characters for a smaller Received 6 March 2014 subset of taxa and geographical representativeness. Here molecular data from three genes (COI, CAD Revised 11 June 2014 and POL-II) are used to reconstruct the phylogeny of the genus. The results show Chimarra to be mono- Accepted 26 June 2014 phyletic, and that some of the sister groups are paraphyletic. Previous hypotheses regarding the relation- Available online 5 July 2014 ships of subgenera within the genus are corroborated but incongruences are also found compared to morphological characters that have been used in keys. The origin of the genus is explored using three dif- Keywords: ferent hypotheses of biogeographical region. The biogeography analyses reveal an origin in the Neotrop- Caddisflies ical region and a subsequent rapid radiation, with dispersal into the Oriental, Palaearctic and Australasian Taxonomy Biogeography regions and secondarily to the Nearctic region. The Afrotropical region has been colonized in several inde- Zoogeography pendent events. The molecular dating using a relaxed clock and calibration with four fossil species indi- Phylogeny cates that Chimarra is about 138 million years old, and that the radiation out of the Neotropical region occurred approximately 124 million years ago. Ó 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
1. Introduction The Philopotamidae Stephens, 1829 are divided into three sub- families; one of them being Chimarrinae Rambur, 1842, to which The higher classification of Trichoptera Kirby, 1813 has been the genera Chimarra Stephens, 1829, Chimarrhodella Lestage, repeatedly reconfigured and re-evaluated through analyses of both 1925 and Edidiehlia Malicky, 1993 belong. With almost 700 morphological and molecular data (e.g. Frania and Wiggins, 1997; described species, Chimarra is the second largest genus in Trichop- Kjer et al., 2001, 2002; Holzenthal et al., 2007; Johanson and Malm, tera after Rhyacophila Pictet, 1834 (Rhyacophilidae Stephens, 1836) 2010; Malm et al., 2013). Analyses of molecular data have made it (Johanson and Oláh, 2012). The number of species described from possible to reinterpret our understanding of the evolutionary his- the world’s major biogeographic regions (sensu Wallace, 1876) tory of the order, and also to confirm previous theories. Molecular reflects a high diversity concentrated in the tropics, with 229 spe- data apparently increases the resolution, especially at family and cies described from the Neotropical region, 321 from the Oriental, genus level (Malm and Johanson, 2011; Johanson et al., 2012). 105 from the Australasian, 83 from the Afrotropical, 39 from the There are presently over 14,000 described species in the order Nearctic and 3 species from the western part of the Palearctic (Morse, 2013), and though the diversity of the order does not reach region. This is somewhat consistent with the Trichoptera order in the massive numbers of the orders Coleoptera, Diptera and Hyme- general, where the largest diversity is found in the humid tropics noptera caddisflies are important in biodiversity and environmen- (de Moor and Ivanov, 2008). Blahnik (2005) outlines the general tal research. As the larvae of Trichoptera are sensitive to pollution, taxonomy of Philopotamidae, including Chimarra and its putative sedimentation and other impacts on freshwater systems from nat- sister taxa, and points out that it is presently unclear how these ural and anthropogenic causes, they have been regarded as valu- genera are related. As with Philopotamidae, Chimarra has been able indicators on ecological and environmental status subject to taxonomic re-positioning, with the addition and removal (Einheuser et al., 2012; Ratia et al., 2012; Ruiz-Garcia et al., 2012). of taxa (Blahnik, 1997). The number of described species increases constantly and the amount of available taxa today opens up for comprehensive analyses of the phylogenetic history of the group. ⇑ Corresponding author at: Department of Zoology, Stockholm University, SE- 10691 Stockholm, Sweden. The larva of Philopotamidae construct nests as a long, sock-like E-mail addresses: [email protected] (E. Wahlberg), kjell.arne. tube of silk and the labrum shape is unique to the order [email protected] (K.A. Johanson). (Holzenthal et al., 2007). The nests of the larvae are fixed, as in http://dx.doi.org/10.1016/j.ympev.2014.06.023 1055-7903/Ó 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). 434 E. Wahlberg, K.A. Johanson / Molecular Phylogenetics and Evolution 79 (2014) 433–442 all representatives of the suborder Annulipalpia. The larva of that the Eurasian species are descendants of the Afrotropical Chimarra is recognized by the presence of a deep and asymmetrical lineage rather than North American ones. Both Ross and Blahnik clypeal notch (Ross, 1944). The adults range from black and dark base their theories on the similarity of Asian and North American brown to yellow and dark red in color. The average body length species, and according to both authors the Neotropical region is typically 3–8 mm including wings. The general diagnosis of adult would include of both early species and ones more closely related Chimarra is the presence of a spur formula 1, 4, 4; and the anal vein to the Asian fauna, but they reach very different conclusions 2 of the hind wings is fused with the anal vein 1 forming a closed regarding the major lineages and their dispersal. Due to the lack cell (Blahnik, 1998). Four subgenera are recognized in Chimarra; of well-founded fossil records and close morphological resem- Chimarrita Blahnik, 1997, Curgia Walker, 1860, Otarrha Blahnik, blance of extant species, we aim at resolving the phylogenetic his- 2002, and Chimarra Stephens, 1829. The subgenus Chimarra occurs tory of the genus by using molecular dating methods. We use in most parts of the world, except Antarctica, while the three other different biogeographical divisions to resolve patterns of origin, subgenera are confined to the Neotropical region (Johanson and dispersal and vicariance events as well as outlining a hypothesis Oláh, 2012). The subgenera within Chimarra are not distinctive of the phylogenetic relationships of Chimarra and its subgenera. morphologically but Blahnik (1998) provides an outline of the morphology of subgenera together with a key. However, the char- 2. Material and methods acters distinguishing the subgenus Chimarra are ambiguous; the curved stem of the radial sector (Rs) (Supplementary Fig. 1a)is 2.1. Taxon sampling not present in several Afrotropical species, and morphological dif- ferences in the genitalia are vaguely defined and are often a com- In order to better understand the relationship between Chimarra bination of characters. The subgenus Otarrha share many and closely related genera, representatives from the genera characters with some species in the subgenera Chimarra and Dolophilodes Ulmer, 1909 (including the subgenera Dolophilodes Chimarrita, but clearly separated from those by the presence of a Ulmer, 1909, Hydrobiosella Tillyard, 1926 and Sortosa Navás, mesal division of tergum X in males and ventral process of segment 1918), Wormaldia McLachlan, 1865, Philopotamus Stephens, 1829, VII in females. The subgenera Chimarrita and Curgia are distinctive Gunungiella Ulmer, 1913 and Chimarrhodella were included in the only by a combination of more or less inconspicuous genital char- analysis as ingroups. Taxa from the families Stenopsychidae acters. The biogeographical patterns are apparently the clearest Martynov, 1924, Hydroptilidae Stephens, 1836, Hydropsychidae distinction of the subgenera, except for the cosmopolitan subgenus Curtis, 1835 and Rhyacophilidae were included as an outgroup. Chimarra. In their recent work on Chimarra in Vietnam, Blahnik Specimens were sampled from a variety of localities with wide, et al. (2012) pointed out the differences in the configuration of almost cosmopolitan coverage. The Chimarra species have been the anal veins of the front wings among species (Supplementary identified to subgenera based on Blahnik (1998). Out of 117 taxa Fig. 1), indicating that these could maybe be used in revealing in the ingroup, 94 are Chimarra and 110 Philopotamidae. The the phylogenetic history of the genus. complete taxon set and citation is listed in Supplementary data 1. Sampling of material was carried out by staff and associates of 1.1. The biogeography of Chimarra the Swedish Museum of Natural History (NHRS). Both Blahnik (1998) and Ross (1956) suggested that the ances- tral area of diversification was South America, preceding the 2.2. DNA amplification and sequencing breakup of Gondwana. There are few known fossils of Chimarra, which is problematic when trying to infer the time and area of ori- Extraction of DNA was carried out using the DNeasy extraction gin. The available fossils of Chimarra are dated to 16.4–20.5 million kit (QIAGEN) on the male abdomen. After extraction each abdomen years old, mid-Cenozoic (Wichard, 1983), whereas the oldest fos- was kept with the rest of the body in voucher vials in 80% ethanol sils of the Philopotamidae are dated to Triassic and Jurassic (e.g. stored at minus 20 °C. The genes selected for the analysis were the Sukacheva, 1973, 1990; Sukacheva and Rasnitsyn, 2004). This is mitochondrial cytochrome c oxidase I (COI) and the nuclear cad- very close to the estimated age of Trichoptera based on morpho- herin-like gene (CAD) and RNA polymerase-II (POL-II) whose the logical data and fossils, and the earliest in the order (Grimaldi suitability has previously been demonstrated for analyzing Tri- and Engel, 2005; Wiegmann et al., 2009). In a recent work based choptera phylogeny (Johanson and Espeland, 2010a,b; Malm, on molecular data by Malm et al. (2013) the taxonomic placement 2010). The primers used are listed in Supplementary data 2. of these fossils was questioned as well as the early phylogenetic Ready-To-Go™ PCR Beads (Amersham Biosciences) were used in split within the order. The Philopotamidae had been regarded as performing the PCR-reaction. The PCR mixture was set up with early by Ross (1956, 1967), Kjer et al. (2001, 2002) and 21 llH2O, 1 ll of reverse and forward primer (10 lM concentra- Holzenthal et al. (2007). Malm et al. (2013) estimated the family tions) and 2 ll DNA-extract. The PCR reaction was initial denatur- to be younger, i.e. early Cretaceous origin, however there is still a ation at 95 °C (5 min) followed by 40 cycled of 95 °C (30 s), 50 °C large gap between the earliest known true Chimarra fossil and (30 s), 72 °C (50 s) and a final run out step at 72 °C (8 min). Samples the known diversification of Philopotamidae. According to Ross that did not yield sufficient DNA-product were amplified again
(1956) the origin of Chimarra was in the subtropical South America, with 3 ll DNA-extract and 20 llH2O and in some cases with an with subsequent dispersal into Eurasia during the tropical climate increased number of cycles (45). POL-II required a higher annealing of the early Cenozoic, followed by dispersal into Africa and Austral- temperature, 53 °C. The PCR-products were cleaned with Exo-Fast asia, and a separate lineage dispersing back to North America. The (QiaQuick PCR Purification Kit (QIAGEN)) and sequencing reactions recent events are all set to the Oligocene – Miocene. Detailed dis- were carried out with BigDye™ (Terminator 3.1 Cycle Sequencing tributional patterns within the Afrotropical and Australasian kit (Applied Biosystems)). Sequencing mixture was set up with regions remained unclear in his work. On the other hand Blahnik 1 ll BigDye™, 1 ll forward and reverse primer (1.6 ll concentra- (1998), partly agreeing with Ross (1956) regarding the area and tion) and 3 ll PCR-product and run at 96 °C (30 s), 50 °C (15 s) time of origin, considered that the age of diversification of and 60 °C (4 min) for 25 cycles. Sequencing samples were cleaned Chimarra to coincide with the breakup of Gondwana and that a few with DyeEx 96 Kit (Applied Biosystems) and electrophoresis Afrotropical taxa share ‘‘basal’’ characteristics with Neotropical taxa. was performed on ABI Prism 3100 Genetic Analyser (Applied In contrast to the theory of Ross (1956), Blahnik (1998) concludes Biosystems) and ABI 3130xl sequencer (Applied Biosystems). The E. Wahlberg, K.A. Johanson / Molecular Phylogenetics and Evolution 79 (2014) 433–442 435 resulting gene sequences were viewed and assembled using scheme from PartitionFinder as above. However preliminary runs PreGAP4 and GAP4 of the Staden software package 1.7 (Staden showed the GTR model to be too complex resulting in low effective et al., 1998). The resulting contigs were aligned and primers sample rates and difficulties reaching convergence between runs removed using BioEdit 7.1.11 (Hall, 1999). The sequences were and therefore the less complex model of Hasegawa, Kishino and stored in VoSeq 1.3.5 (Peña and Malm, 2012), which was also used Yano (HKY) was used. Tree prior was set to the Birth-Death speci- to calculate parsimony informativeness and variability. ation option (Gernhard, 2008) and the analysis was run for 200,000,000 generations with a sampling frequency every 10,000th generation. The analysis was run twice and the log files 2.3. Phylogenetic reconstruction examined in Tracer 1.5 (Rambaut and Drummond, 2007) for con- vergence, the effective sample size (ESS) was for all parameters To select the most appropriate substitution model Partition- at least 200 for the separate runs, and 500 combined. The tree files Finder 1.1.1 (Lanfear et al., 2012) was used. Each of the three codon from both runs were combined in LogCombiner (Drummond and positions for each gene was treated separately (nine possible par- Rambaut, 2007), and thereafter summarized with a burn-in value titions). The best partition scheme and set of substitution models of 25% in TreeAnnotator 1.7.5 (Drummond and Rambaut, 2007) was determined using BIC with the greedy search algorithm and The maximum clade credibility tree was viewed in FigTree 1.4 RAxML option (Lanfear et al., 2014). (Rambaut, 2006) and thereafter annotated and prepared in AdobeÒ For maximum parsimony analysis on an unpartitioned dataset IllustratorÒ 15.1. TNT 1.1 (Goloboff et al., 2008) was used with ratchet and drift with 10,000 replications (xmult = rep 10,000 rat1 drift 1; rat; bb;). A strict consensus tree was generated and clade support values were 2.5. Biogeography obtained from a Jackknife search with 5000 replicates and a dele- tion probability of 36%. Bayesian inference was carried out using The biogeographical analyses were carried out using the maxi- MrBayes 3.2.1 (Huelsenbeck and Ronquist, 2001; Ronquist and mum likelihood-based dispersal–extinction–cladogenesis (DEC) Huelsenbeck, 2003) on the University of Oslo BioPortal (http:// model implemented in RASP 2.1b (Yu et al., 2013) based on the www.bioportal.uio.no/, Kumar et al., 2009). The best partition Lagrange method described by Ree et al. (2005) and Ree and scheme was used with corresponding best substitution models, Smith (2008). The DEC model incorporates the branch lengths of the temperature was set to 0.11, number of chains to 4, parallel a calibrated ultrametric tree and as input our calibrated tree from runs to 2 and the analysis was run for 100,000,000 generations the BI-analysis was used. Due to the short time span of the recent with a sampling frequency of 5000. The average standard deviation Chimarra radiation, the analysis was run with unconstrained set- of split frequencies was used to determine when convergence was tings allowing dispersal at any time between regions. A basis in reached (<0.005). The trees were summarized with a burn-in value many analyses is the traditional division of zoogeographic regions of 25%, resulting in a total of 30,000 trees from which a majority that follows a 7-region scheme by Wallace (1876). The names of rule consensus tree was calculated. The same analysis was run the regions have since Wallace’s original publication been slightly twice with identical settings to test search stability. The resulting altered, for example the Ethiopian region have been substituted trees from both analyses were united, exported using PAUP* with the Afrotropical region Also, the Pacific islands where not 4.0b10 (Swofford, 2003). Wing venation variation (radial sector originally a region according to Wallace (1876). As these changes stem and anal veins) were mapped to the Bayesian tree to illustrate have been favoured and extensively adopted we will here use the positions of state modifications in the tree. new names, as well as the region of Pacific islands, in the model we refer to as based on the Wallace (1876) scheme. Beside the tra- 2.4. Calibration of divergence times ditional regional scheme by Wallace (1876), we test two recent proposals for zoogeographical regions. The numbers of regions Four fossil taxa were used to calibrate nodes. A log normal prior are increased in the work by de Moor and Ivanov (2008) and distribution was applied; the offset value was set to the minimum focused on the temperate regions and Holt et al. (2013) concen- age of the fossil. The mean was set so that the 97.5% quantile was trate their increased number of regions to the tropical areas. The 226 million years ago (Mya), the age of the split of Rhyacophilidae analysis was run three times for each of the zoogeographic division and the rest of the Trichoptera families as estimated by Malm et al. types, i.e. (1) Wallace (1876), (2) de Moor and Ivanov (2008) and (2013). The standard deviations of all nodes were set to 0.75. The (3) Holt et al. (2013) to investigate if differently defined input areas calibration points and values used are given in Table 1. As starting would yield different results regarding the ancestral distribution of tree the result from the Bayesian inference was used and polyto- Chimarra. The recent regional hypotheses compared to Wallace’s mies were resolved before the analysis using the multi2di function regions are illustrated in Supplementary Fig. 2. The maximum pos- in APE 3.0-11 (Paradis et al., 2004) package in R 3.0.2 (R Core Team, sible ancestral areas were varied depending on the regional 2013). Dating was carried out using BEAST 1.7.5 (Drummond and hypothesis used. The estimated divergence time of the Philopota- Rambaut, 2007). The dataset was partitioned according to the midae precedes the division of the Neotropical and Afrotropical
Table 1 Fossil taxa and references used for calibration of divergence times. Parameters of lognormal distribution priors are mean, standard deviation (SD) and offset.
Calibrated segment Mean SD Offset Fossil taxa References (Mya) Divergence point between Hydroptilidae and [Hydropsychidae, (Stenopsychidae, 3.450 0.75 89 Agraylea parva Wichard and Bölling Philopotamidae)] (2000) Divergence point between Hydropsychidae and (Stenopsychidae, Philopotamidae) 3.7873 0.75 34 Hydropsyche Ulmer (1912) viduata Stem to Wormaldia 3.4500 0.75 89 Wormaldia Botosaneanu (1995) praecursor Stem to Chimarra 3.8770 0.75 16 Chimarra palaenova Wichard (2007) 436 E. Wahlberg, K.A. Johanson / Molecular Phylogenetics and Evolution 79 (2014) 433–442 continents as well as the connection between the Neotropical and hypothesis by Wallace (1876) the origin of Chimarra is inferred Australian continents via Antarctica. To account for the possibility to be Neotropical+Oriental (N1), reflecting the different distribu- of multiregional distributions in the more complex schemes the tions of the endemic South American genera (N2) compared to number of maximum possible areas of each node was set to 2 for the oriental ancestral distribution of Chimarra (Chimarra) (N3). the Wallace (1876) hypothesis and 3 for the de Moor and Ivanov The Chimarra(Chimarra) lineages that diverge at node N4 has an (2008) and Holt et al. (2013) hypotheses. All ingroup and outgroup Afrotropical+Oriental ancestral distribution, including the lineage taxa were included in the analysis. leading to the American group of the subgenus (N5). The major group containing mostly Australasian and Pacific lineages have an Afrotropical+Australasian ancestral distribution (N6). Both the 3. Results results from the analyses based on the hypotheses by de Moor and Ivanov (2008) and Holt et al. (2013) better resolve the ances- 3.1. Phylogenetic analyses tral distributions with less occasions of combined regions. Based on the hypothesis by de Moor and Ivanov (2008) the ancestry of The length for aligned individual gene datasets were for CAD Chimarra is revealed to be Neotropical (N1). The subgenus 850 nucleotides (58.24% variable), COI 658 nucleotides (62.92% Chimarra is though strongly supported as Oriental (N3), as well variable) and POL-II 772 nucleotides (85.23% variable) and the as the following lineages until the Afrotropical, Australasian and complete dataset included 2280 nucleotides in total. Out of 118 American groups within, including the linage leading to the taxa we were able to sequence 112 individuals for the CAD gene, American group (N5). The results based on the regional hypothesis 111 for COI and 88 for POL-II genes. Some amplification reactions by Holt et al. (2013) sets the ancestral distribution of Chimarra to did not yield PCR-product, even after re-amplification. The best South America (N1). The node to Chimarra (Chimarra) (N3) also partition scheme grouped first codon of CAD and third codon of expresses low probabilities but with second probable region being POL-II into one partition and remaining 7 codon positions parti- geographically close to the first; a Chinese ancestral distribution tioned individually. A generalized time-reversible model (GTR) followed by an Oriental. A Papua-Melanesian ancestral distribution with a gamma distribution and an allowance for the presence of is shared in the nodes leading to the predominantly Australian and invariant sites (GTR+I+G) was preferred to all partitions except Papua Melanesian groups (N6). The lineage leading to the American for first codon of COI and second codon of POL-II, where invariable Chimarra (Chimarra) (N5) is the result of dispersal from the sites were excluded (GTR+G). Parsimony informativeness varied Indo-Malayan region, over the Oriental region and reaching finally across the different genes, with POL-II being the most informative the Mexico and North America regions. A hypothesis on the (84.33%) followed by CAD (52%) and COI (48.61%). radiation of Chimarra based on the biogeographic analysis and Parsimony analysis resulted in 6 equal-length trees. A strict the results from molecular time calibration is presented in Fig. 4. consensus tree is illustrated in Fig. 1. The family Philopotamidae is monophyletic, as well as the genus Gunungiella. However mono- phyly of the other genera within the family was not supported. The 4. Discussion South American endemic subgenera of Chimarra are monophyletic, with Chimarra (Otharra) and Chimarra (Curgia) as sister groups. 4.1. Classification Bayesian inference resulted in a tree with strong support values (Fig. 2). The genera Wormaldia, Dolophilodes and Philopotamus form Kjer et al. (2002) included several Philopotamidae genera and a monophyletic sister clade to Chimarra, and contrary to the hypothesized an early divergence of Wormaldia within in the parsimony analysis Wormaldia is monophyletic. Chimarrhodella is family, splitting earlier than a assemblage of representatives of not recovered as a sister group to Chimarra but sister to Dolophilodes and Philopotamus. One of the two Dolophilodes species Chimarra+(Wormaldia, Dolophilodes, Philopotamus). Within Chimarra included grouped with Hydrobiosella, the other as a sister group to (N1) the endemic South American Chimarra subgenera (Paulianodes (Chimarrhodella, Chimarra)). In our results the genus Chimarra(Otarrha), Chimarra (Curgia) and Chimarra (Chimarrita) Dolophilodes also appears as paraphyletic in both phylogenetic form a distinct monophyletic clade (N2), with Chimarra (Chimarrita) analyses. Blahnik (2005) points out the taxonomic confusion recovered as sister group to Chimarra (Curgia). In the cosmopolitan regarding Dolophilodes and proposes that several subgenera be ele- Chimarra (Chimarra) a monophyletic group of Oriental species vated to genus level, one of them Sortosa. However, as we do not forms the sister group to the remaining species in this subgenus recover Dolophilodes (Sortosa) as monophyletic and a re-elevated (N3). Two large sister groups dominate Chimarra(Chimarra) at N4, Hydrobiosella (Neboiss, 1977) is nested within the Dolophilodes in one dominated by Australasian and Pacific species with a few Afri- the Bayesian inference, further considerations will have to be can species, and the second group made up of Oriental, African, made. This highlights that the monophyly of Dolophilodes, and as Palaearctic and American species (N5). well Philopotamus, require additional research. The placement of Chimarrhodella as sister to the remaining Philopotamidae, and 3.2. Age and time of divergence not as sister to Chimarra, in the Bayesian inference is contrary to the results from parsimony and previous hypotheses. The maximum credibility tree with node ages and 95% credible One of the key characters distinguishing Chimarra (Chimarra) intervals are given in Supplementary Fig. 3. The age of Chimarra is from other subgenera is a curved stem of the Rs, and this was 138 million years (My), and the split of the South American sub- not found in 15 out of 81 species included here (Fig. 2a). Deviations genera from the cosmopolitan Chimarra (Chimarra) 124 My. The in this character are not restricted to Afrotropical species as two major Chimarra(Chimarra) radiations at N4 are 83 My old, suggested by Blahnik (1998), but occur in species also from the and the American species of Chimarra (Chimarra) 60 My old. Australasian and Oriental regions. The absence of a crossvein between A1 and A2 is, except for one species [Chimarra (Chimarra) 3.3. Biogeographical history kimminsi from Fiji], a character shared by early species (Fig. 2b). Only three species from Laos at node N4 in the tree lack the The ancestral distribution at major nodes are illustrated in crossvein: Chimarra (Chimarra) okuihorum Mey, 1998, Chimarra Fig. 3, detailed results for each individual analysis are illustrated (Chimarra) demeter Malicky, 2000 and Chimarra (Chimarra) in Supplementary Figs. 4–6. In the analysis using the regional shanorum Chantaramongkol and Malicky, 1989. This indicates that E. Wahlberg, K.A. Johanson / Molecular Phylogenetics and Evolution 79 (2014) 433–442 437
Fig. 1. Strict consensus tree of 6 most parsimonious trees generated in TNT 1.1 (Goloboff et al., 2008) from the combined dataset of the three protein-coding genes COI, CAD and POL-II. Numbers above clades are clade support values derived from 5000 jackknife searches with a deletion probability of 36%. 438 E. Wahlberg, K.A. Johanson / Molecular Phylogenetics and Evolution 79 (2014) 433–442
Fig. 2. Majority rule tree based on 30,000 post-burn in trees from MrBayes (Ronquist and Huelsenbeck, 2003) on a partitioned dataset of the three protein-coding genes COI, CAD and POL-II. Numbers above branches represent posterior probability values. Coloured lines to the right illustrate forewing vein characters. a) Solid line: Rs with straight stem; dashed line: Rs with curved stem. b) Solid line: crossvein a1–a2 absent; dashed line: crossvein a1–a2 present. E. Wahlberg, K.A. Johanson / Molecular Phylogenetics and Evolution 79 (2014) 433–442 439
Fig. 3. Biogeographical hypothesis from RASP 2.1b (Yu et al., 2013) using Lagrange method (Ree et al., 2005). Regions based on each hypothesis by Wallace (1876), de Moor and Ivanov (2008) and Holt et al. (2013). Numbers in parenthesis represent probability values of most probable ancestral distribution. 440 E. Wahlberg, K.A. Johanson / Molecular Phylogenetics and Evolution 79 (2014) 433–442
a b
c d
Fig. 4. Hypothetical historical pattern of radiation of Chimarra at 120 Mya (a), 90 Mya (b), 60 Mya (c) and 30 Mya (d), based on the results from the molecular divergence time and biogeographic analyses. Dashed lines illustrate continental separation mediating vicariance events while arrows represent dispersal events. Maps simplified from Smith et al. (1994).
the presence of a crossvein in early species is not homolog with region and the Pacific Islands. The present islands of the pacific similar state elsewhere in the tree. Since the works by Ulmer are generally younger than the ages of the lineages present in (1916) wing venation has been an important character set in phy- the region, however our analysis, including the Bayesian inference logenetic hypotheses among philopotamidae genera and in fossil where Oriental species precedes the two major radiations, together taxa these characters are often the only information available. with the complex history of the islands and region (Heads, 2013) Flint (1998) revised the subgenus Curgia and noted the difficulty suggests that a Oriental origin of the Australasian species is plausi- in providing morphological characters defining the Chimarra sub- ble. The following dispersal of Chimarra (Chimarra) from the Nearc- genera. Defining groups based on plesiomorphic charterers has tic region to the Neotropical region (Fig. 4d) is apparently the last likely contributed to the uncertainties in understanding the phy- trans-continental dispersal within the genus, and reveals an era of logeny of Philopotamidae. explosive global radiation mediated by the Paleocene-Eocene (66– 33 Mya) climate and continental movements. The radiation of 4.2. Age and distribution Chimarra (Chimarra) from South America into the Oriental and sub- sequently to the Afrotropical region is coherent with the hypothe- Our results yield a much older age of Philopotamidae than sis by Ross (1956). The later re-colonization of the American found in Malm et al. (2013), however our estimated age of both continents by Chimarra (Chimarra) is coherent with both Ross Chimarra and Hydrobiosella lies within their reported confidence (1956) and Blahnik (1998). We cannot, however, corroborate the intervals. According to our results the age of Chimarra dates back hypothesis by Blahnik (1998) of an early colonization of the Afro- to early Cretaceous, at the time the Pangaea fragmented into tropical region. We hypothesize a direct early dispersal from South Gondwana and Laurasia, and temperatures increased at the end America to Eurasia, without dispersal into the Afrotropical region. of the Berriasian epoch (140 Mya). At the time of the split between The short branches in the tree indicates rapid radiations, particu- the South American endemic subgenera and Chimarra (Chimarra) larly in Afrotropical, south Oriental and Australasian regions the continents were still in close proximity to each other and the including the Pacific islands. Oriental species precedes the Austral- sea between the northern and southern continents spanned by asian species both in the phylogeny and time, and there is no evi- island land masses. It is likely, taking into account the palaeogeo- dence or indications of a trans-Antarctic dispersal into Australasia. graphic history, that the first dispersal from South America was a The Afrotropical Chimarra are probably a composite of species from range expansion into the Eurasian region during Cretaceous several lineages, with at least one colonization sharing lineage with approximately 125 Mya (Fig. 4a), followed by dispersal to Africa the American Chimarra (Chimarra). The first Afrotropical coloniza- and southeast Asia (Fig. 4b). The completion of the separation of tion of the genus precedes the recent intense diversification and the Neotropical continent from the Afrotropical region and Eur- explains the presence of species with characters resembling the asian regions 110 Mya isolated the South American lineages from earliest lineages. Blahnik (1998) used the Oriental species Chimarra Chimarra (Chimarra). During the warm and humid early Paleogene (Chimarra) wilharawela Schmid, 1958, as outgroup when studying (66–23 Mya) tropical species probably dispersed towards the the morphology of ‘‘New World’’ Chimarra (Chimarra) due to its poles, and during this time Chimarra (Chimarra) advanced through- resemblance to this group and assigned a subgroup, Chimarra out the Palaearctic region and colonized the Nearctic region (Fig. 4c). (Chimarra) obscura (Walker, 1852), as the earliest of the American This is possibly also the time of dispersal into the Australasian species groups. This group is distributed in the eastern North E. Wahlberg, K.A. Johanson / Molecular Phylogenetics and Evolution 79 (2014) 433–442 441
America, and its placement as sister to the American Chimarra by Dr. Mattias Olsson, Celeste Hedquist Pehrson and Lovisa (Chimarra) group is corroborated by our results. There are, how- Skagerman. ever, a number of intermediate species, with distributions in the The material from Fiji was provided by the Terrestrial Arthro- Afrotropical, Oriental and Nearctic regions, preceding this clade pod Survey of Fiji project, founded in part by The National Science and later than the Oriental Chimarra (Chimarra) wilharawela.A Foundation (DEB-0425790) and the Schlinger Foundation. The number of these species are described after the work by Blahnik work was partly supported by the Swedish Research Council (Grant (1998) or are still un-described; however they share similar mor- #2005-4834) and the Riksmusei Vänner. phological characters to the Oriental clade wherein Chimarra (Chimarra) wilharawela is placed. Appendix A. Supplementary material
4.3. Revising the biogeographical regions Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ympev.2014. While the comprehensive work of Wallace (1876) included a 06.023. synthesis of animal orders the amount of known species were only a fraction of the numbers today, especially for insects. The zoogeo- References graphic map of Wallace (1876) is largely congruent with the con- tinental plates and was also based on the distribution of extant Blahnik, R.J., 1997. Systematics of Chimarrita, a new subgenus of Chimarra species. In the recent bioregions proposed by de Moor and (Trichoptera: Philopotamidae). Syst. Entomol. 22 (3), 199–243. http:// Ivanov (2008) and Holt et al. (2013) more emphasis have been dx.doi.org/10.1046/j.1365-3113.1997.d01-39.x. Blahnik, R.J., 1998. A revision of the Neotropical members of the genus Chimarra placed on the phylogeny within groups, and the former also subgenus Chimarra (Trichoptera: Philopotamidae). Memoirs Am. Entomol. Inst. focused on fossil evidence for distributional patterns. The latter 59, 1–318. applied a statistical approach analyzing the evolutionary patterns Blahnik, R.J., 2002. Systematics of Otarrha, a new Neotropical subgenus of Chimarra (Trichoptera: Philopotamidae). Syst. Entomol. 27, 65–130. http://dx.doi.org/ found by modern phylogenies. Hence there are various methods 10.1046/j.0307-6970.2001.00166.x. producing different results and care should be taken when select- Blahnik, R.J., 2005. Alterosa, a new caddisfly genus from Brazil (Trichoptera: ing biogeographical maps for analyzing distribution data. In our Philopotamidae). Zootaxa 991, 1–60. http://dx.doi.org/10.3897/ results we see that the new regional schemes produce less ambig- zookeys.317.5437. Blahnik, R.J., Arefina-Armitage, T.I., Armitage, B.J., 2012. The genus Chimarra uous ancestral distribution at the nodes, however the support val- Stephens (Trichoptera: Philopotamidae) in Vietnam. Insecta Mundi 229, 1–25. ues are generally lower. The number of described caddisfly species Botosaneanu, L., 1995. Caddis flies (Trichoptera) from Turonian (Upper Cretaceous) differs vastly among regions, e.g. in the Afrotropical region the amber of New Jersey. Am. Mus. Novit. 3140, 1–7. Chantaramongkol, P., Malicky, H., 1989. Some Chimarra (Trichoptera: number of described species is proportionally low compared to Philopotamidae) from Thailand (Studies on caddisflies from Thailand, no. 2). the size of the region. Evaluation of the biogeographically regions Aquatic Insects 11 (4), 223–240. http://dx.doi.org/10.1080/01650428909361376. using other taxa at different scales are needed to determine the Curtis, J., 1835. British Entomology, vol. XII. Richard Taylor, London, pp. 530–577. de Moor, F.C., Ivanov, V.D., 2008. Global diversity of caddisflies (Trichoptera: usability of the new regions. Also, extended exploration of both Insecta) in freshwater. Hydrobiologia 595, 393–407. http://dx.doi.org/10.1007/ extant and fossil taxa in understudied areas might change the 978-1-4020-8259-7_41. number of tropical regions and the resolution of biogeographic Drummond, A.J., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol. Biol. 7, 214. http://dx.doi.org/10.1186/1471-2148-7- methods even more. 214. Einheuser, M.D., Nejadhashemi, A.P., Sowa, S.P., Wang, L.Z., Hamaamin, Y.A., 5. Conclusions Woznicki, S.A., 2012. Modeling the effects of conservation practices on stream health. Sci. Total Environ. 435, 380–391. http://dx.doi.org/10.1016/ j.scitotenv.2012.07.033. Our results support some deviations from the phylogenies Flint, O.S., 1998. Studies of Neotropical caddisflies, LIII: a taxonomic revision of the based exclusively on morphological characters and give a more subgenus Curgia of the genus Chimarra (Trichoptera: Philopotamidae). Smithsonian Contrib. Zool. 594, 1–131. detailed picture regarding the biogeographical history including Frania, H.E., Wiggins, G.B., 1997. Analysis of morphological and behavioural the intermediate events and the relationships among lineages. Pre- evidence for the phylogeny and higher classification of Trichoptera (Insecta). viously authors have raised questions regarding the origin and In: Life Sciences Contributions. vol. 160. Royal Ontario Museum, Toronto, Canada. radiation of the genus Chimarra that we are now able to resolve Gernhard, T., 2008. The conditioned reconstructed process. J. Theor. Biol. 253 (4), using the high number of species worldwide in combination with 769–778. http://dx.doi.org/10.1016/j.jtbi.2008.04.005. molecular methods. We also see that in order to fully understand Goloboff, P., Farris, J., Nixon, K., 2008. TNT: a free program for phylogenetic analysis. Cladistics 24, 774–786. http://dx.doi.org/10.1111/j.1096-0031.2008.00217.x. the relationships among philopotamid genera larger data sets are Grimaldi, D.A., Engel, M.S., 2005. Evolution of the Insects. Cambridge University required, particularly valuable will be inclusion of higher represen- Press, Cambridge, pp. 548–554. tativeness geographically of various genera. Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser. 41, 95–98. Acknowledgments Holt, B.G., Lessard, J.P., Borregaard, M.K., Fritz, S.A., Araújo, M.B., Dimitrov, D., Fabre, P.H., Graham, C.H., Graves, G.R., Jønsson, K.A., Nogués-Bravo, D., Wang, Z., Great thanks goes to Veronica Nyström Edmark, Bodil Whittaker, R.J., Fjeldså, J., Rahbek, C., 2013. An update of Wallace’s zoogeographic regions of the world. Science 339, 74–78. http://dx.doi.org/ Cronholm and Keyvan Mirbakhs at the Molecular Systematics 10.1126/science.1228282. Laboratory (Swedish Museum of Natural History) for being very Heads, M., 2013. Biogeography of Australasia, a Molecular Analysis. Cambridge helpful during the laboratory work. Specimens used in the study University Press, Cambridge, pp. 346–348. have partly been collected by Philemon Ala, Niklas Apelqvist, Holzenthal, R.W., Blahnik, R.J., Prather, A.L., Kjer, K.M., 2007. Order Trichoptera Kirby, 1813 (Insecta), caddisflies. Zootaxa 1668, 639–698. William Bani, Johannes Bergsten, Thieu Du Tran, Marianne Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogeny. Espeland, Mattias Forshage, Levent Gültekin, Kathy Hill, Mike Bioinformatics 17, 754–755. http://dx.doi.org/10.1093/bioinformatics/17.8.754. Irwin, Mattias and Catrin Jaschhof, Kjell Arne Johanson, Tarer Johanson, K.A., Espeland, M., 2010a. Phylogeny of the Ecnomidae (Insecta: Trichoptera). Cladistics 26, 36–48. http://dx.doi.org/10.1111/j.1096- Karae, Karl Kjer, Jostein Kjærandsen, Daniel Leyang, Tobias Malm, 0031.2009.00276.x. Dave Marshall, Peter Nilsson, Jan Ohlsson, Fred Pitisopa, Fresia Johanson, K.A., Espeland, M., 2010b. Description of new Chimarra (Trichoptera: Ester Rojas, Evert I. Schlinger, Thai Hong Pham, Touasi Tiwok, Philopotamidae) species from the Solomon Islands. Zootaxa 2638, 25–43. Johanson, K.A., Malm, T., 2010. Testing the monophyly of Calocidae (Insecta: Delena Veikovi, Vaenu Vigulu, Bert Viklund and Dawn Williams. Trichoptera) based on multiple molecular data. Mol. Phylogenet. Evol. 54, 535– Valuable comments on the manuscript have been provided 541. http://dx.doi.org/10.1016/j.ympev.2009.09.025. 442 E. Wahlberg, K.A. Johanson / Molecular Phylogenetics and Evolution 79 (2014) 433–442
Johanson, K.A., Malm, T., Espeland, M., Weingartner, E., 2012. Phylogeny of the Ree, R.H., Smith, S.A., 2008. Maximum likelihood inference of geographic range Polycentropodidae (Insecta: Trichoptera) based on protein-coding genes reveal evolution by dispersal local extinction and cladogenesis. Syst. Biol. 57, 4–14. non-monophyletic genera. Mol. Phylogenet. Evol. 65, 126–135. http:// http://dx.doi.org/10.1080/10635150701883881. dx.doi.org/10.1016/j.ympev.2012.05.029. Ronquist, F., Huelsenbeck, J.P., 2003. MRBAYES 3: Bayesian phylogenetic inference Johanson, K.A., Oláh, J., 2012. Revision of the Fijian Chimarra (Trichoptera, under mixed models. Bioinformatics 19, 1572–1574. http://dx.doi.org/10.1093/ Philopotamidae) with description of 24 new species. Zootaxa 3354, 1–58. bioinformatics/btg180. Kirby, W., 1813. Strepsiptera, a new order of insects proposed; and the characters of Ross, H.H., 1944. The caddisflies or Trichoptera of Illinois. Bull. Illinois Natural Hist. the order, with those of its genera, laid down. Trans. Linnean Soc. London 11, Surv. 23, 1–326. 86–122. Ross, H.H., 1956. Evolution and Classification of the Mountain Caddisflies. Kjer, K.M., Blahnik, R.J., Holzenthal, R.W., 2001. Phylogeny of Trichoptera University of Illinois Press, Urbana, IL, 213pp. (caddisflies): Characterization of signal and noise within multiple datasets. Ross, H.H., 1967. The evolution and past dispersal of the Trichoptera. Annu. Rev. Syst. Biol. 50, 781–816. Entomol. 12, 169–206. http://dx.doi.org/10.1146/annurev.en.12.010167. Kjer, K.M., Blahnik, R.J., Holzenthal, R.W., 2002. Phylogeny of caddisflies (Insecta, 001125. Trichoptera). Zoolog. Scr. 31, 83–91. http://dx.doi.org/10.1046/j.0300- Ruiz-Garcia, A., Marquez-Rodriguez, J., Ferreras-Romero, M., 2012. Implications of 3256.2001.00079.x. anthropogenic disturbance factors on the Trichoptera assemblage in a Kumar, S., Skjæveland, Å., Orr, R.J.S., Enger, P., Ruden, T., Mevik, B.-H., Burki, F., Mediterranean fluvial system: are Trichoptera useful for identifying land-use Botnen, A., Shalchian-Tabrizi, K., 2009. AIR: A batch-oriented web program alterations? Ecol. Ind. 14 (1), 114–123. http://dx.doi.org/10.1016/j.ecolind.2011. package for construction of supermatrices ready for phylogenomic analyses. 06.026. BMC Bioinformatics 10, 357. http://dx.doi.org/10.1186/1471-2105-10-357. Schmid, F., 1958. Trichopteres de Ceylan. Archiv Hydrobiol. 54, 1–173. Lanfear, R., Calcott, B., Ho, S.Y., Guindon, S., 2012. PartitionFinder: combined Smith, A.G., Smith, D.G., Funnell, B.M., 1994. Atlas of Mesozoic and Cenozoic selection of partitioning schemes and substitution models for phylogenetic Coastlines. Cambridge University Press, Cambridge, pp. 28–40. analyses. Mol. Biol. Evol. 29, 1695–1701. http://dx.doi.org/10.1093/molbev/ Staden, R., Beal, K.F., Bonfield, J.K., 1998. The Staden Package, computer methods in mss020. molecular biology. In: Misener, S., Krawetz, S.A. (Eds.), Bioinformatics Methods Lanfear, R., Calcott, B., Kainer, D., Mayer, C., Statmatakis, A., 2014. Selecting optimal and Protocols. The Humana Press Inc., Totowa, New Jersey, USA. partitioning schemes for phylogenomic datasets. BMC Evol. Biol. 14, 82. http:// Stephens, J.F., 1829. The Nomenclature of British Insects; Being a Compendious List dx.doi.org/10.1186/1471-2148-14-82. of Such Species as are Contained in the Systematic Catalogue of British Insects, Lestage, J.A., 1925. Notes Trichoptérologiques. Bull. Ann. Soc. Entomól. Bel. 65, 35– and Forming a Guide to their Classification. Baldwin & Cradock, London, 82pp. 44. Stephens, J.F., 1836. Illustrations of British Entomology (Mandibulata). Baldwin and Malicky, H., 1993. Neue asiatische Köcherfliegen (Trichoptera: Rhyacophilidae, Cradock, London. Philopotamidae, Ecnomidae und Polycentropodidae). Entomol. Ber. Luzern 29, Sukacheva, I.D., 1973. New caddis flies (Trichoptera) from the Mesozioc of Soviet 77–88. Central Asia. Paleontol. J. 3, 377–384. Malicky, H., 2000. New species from Sabah, Nepal, India, China. Braueria 27, 32–33. Sukacheva, I.D., 1990. Scorpion flies, Panorpida, Caddis flies, Phryganeida. In: Late Malm, T., 2010. Climbing the Trichoptera tree: investigations of branches and Mesozoic insects of Eastern Transbaikalia. Trudy Paleontologicheskogo Instituta leaves. In: Systematic Zoology. Stockholm University, Stockholm. Akadamii Nauk SSSR 239, pp. 82–122. Malm, T., Johanson, K.A., 2011. A new classification of the long-horned caddisflies Sukacheva, I.D., Rasnitsyn, A.P., 2004. Jurassic insects (Insecta) from the Sai-Sagul (Trichoptera: Leptoceridae) based on molecular data. BMC Evol. Biol. 11, 10. locality (Kyrgyzstan, Southern Fergana). Paleontol. J. 38, 182–186. http://dx.doi.org/10.1186/1471-2148-11-10. Swofford, D.L., 2003. PAUP⁄. Phylogenetic Analysis Using Parsimony (⁄and Other Malm, T., Johanson, K.A., Wahlberg, N., 2013. The evolutionary history of Methods). Version 4. Sinauer Associates, Sunderland, Massachusetts. Trichoptera (Insecta): a case of successful adaptation to life in freshwater. Tillyard, R.J., 1926. Kansas Permian insects. Part 7. The order Mecoptera. Am. J. Sci. Syst. Entomol. 38, 459–473. http://dx.doi.org/10.1111/syen.12016. (Series 5) 11, 133–164. Martynov, A.V., 1924. Trichoptera. Pract. Entomol. 5, 1–388. Ulmer, G., 1909. Einige neue exotische Trichopteren. Notes Leyden Museum 31, McLachlan, R., 1865. Trichoptera Britanica. A monograph of British species of 125–142. caddis-flies. Trans. Entomol. Soc. London 5 (3), 1–184. Ulmer, G., 1912. Die Trichopteren des baltischen Bernsteins. Beitraege Naturkunde Mey, W., 1998. Die Köcherfliegenfauna des Fan Si Pan massivs in Nord–Vietnam 3 – Preussens Koenigsberg 10, 1–380. Besch reibung weiterer neuer Arten (Trichoptera). Opuscula Zool. Fluminensia Ulmer, G., 1913. Über einige von Edw. Jacobson auf Java gesammelte Trichopteren, 165, 10–11. Zweiter Beitrag. Notes Leyden Museum 35, 78–101. Morse, J.C. (Ed.), 2013. Trichoptera World Checklist.