Ancyronyx Erichson, 1847 (Coleoptera, Elmidae) from Mindoro, Philippines, with description of the larvae and two new species using DNA sequences for the assignment of the developmental stages

Abstract Ancyronyx buhid sp. n. and Ancyronyx tamaraw sp. n. are described based on adults and larvae, matched using their cox1 or cob DNA sequence data. Additional records of Ancyronyx schillhammeri Jäch, 1994 and Ancyronyx minerva Freitag & Jäch, 2007 from Mindoro are listed. The previously unknown larva of Ancyronyx schillhammeri is also described here, aided by cox1 data. The new species and larval stages are described in detail and illustrated by SEM and stacked microscopic images. Keys to the adult and larval Ancyronyx species of Mindoro and an updated checklist of Philippine Ancyronyx species are provided. The usefulness as bioindicators, the phylogenetic relationships and biogeographic aspects affecting the distribution patterns are briefly discussed.

Additional cox1 sequences of Philippine Ancyronyx species (Freitag and Balke 2011) previ ously submitted to ENA/GenBank (http://www.ebi.ac.uk/ena/) were included (see Table  1). The same applies for Podelmis viridiaenea Jäch, 1982 (Elmidae) from Sri Lanka that was used as outgroup. The newly amplified sequences were traced and aligned in CLUSTALW (Thompson et al. 1994) using BIOEDIT version 7.0.5.2. (Hall 1999) and default param eters. Phylogenetic analyses were conducted with MRBAYES vers. 3.1.2 (Ronquist et al. 2012) using the GTR (General Time Reversible) model (Tavaré 1986) with default priors starting with random trees with three heated and one cold Markov chains. The analysis was run by 1,000,000 generations, and the first 25% of samples from the cold chain have been discarded as burnin. Branch support for the Bayesian trees was assessed with poste rior probabilities determined via the 50% majority rule consensus. This easy analysis is only intended for matching larva and adults of the species treated in this paper.

Morphological analysis
Digital photographs were taken with an OLYMPUS SZ 61 stereo microscope (species habitus), and an OLYMPUS CX 21 compound microscope (dissected body parts), both with digital photo adapter LW Scientific MiniVID DCM 310. For each illustra tion a series of photographs taken at various focus layers was stacked using the stack function (species habitus) and corrected weighted average function (dissected body parts) of COMBINEZM software (Hadley 2008). The same optical systems we used for the dissection of adult specimens and the material examination. Biometric meas urements were done by the use of a calibrated ocular micrometer.
Scanning electron microscope (SEM) images of vacuum dried material were ob tained using a ZEISS EVO 50 XVP. Except for the single larval specimen of Ancyronyx schillhammeri, all specimens were coated with gold using one dissected and one entire specimen each.
For all larval material examined, measurements of the head capsule width are given in mm (e.g. 1 L (0.31)) as a suitable indicator for the larval size and the instar stage assignment (see Freitag and Balke 2011).

Abbreviations and repositories
Brgy. Barangay

Data resources
The data underpinning the analysis reported in this paper are deposited at GBIF, the Global Biodiversity Information Facility, http://ipt.pensoft.net/ipt/resource. do?r=ancyronyx_mindoro_data All DNA sequences were submitted to ENA/GenBank via online submission to EMBLEBI. Accession numbers and curatory information are listed in Table 1. table 1. ENA/GenBank accession numbers of DNA sequences, geographical origins, collection sites and organismic sample references of specimens used for moleculargenetic analyses.

DNA sequence analysis
Alignment of the cox1 data and trimming ambiguous bases at the 3' and 5' ends yield ed a matrix of 804 bp. None of the sequences contained indels. The sequences of the larvae of A. buhid had nine ambiguous positions inbetween which were coded as 'N's. All adults and larvae could be matched unambiguously. Sequences of adult and larva of the same species from the same locality or island were identical or varied just in one base pair. Sequence samples of A. minerva from Mindoro diverged in eight base pair positions (six of them synonymous substitutions) from that of the same species from Palawan.
The sequencing of cox1 of the adults of A. buhid failed and is not included in the phylogenetic analysis. The cob sequences, however, which were amplified for two larvae and an adult of this species allowed unambiguous matching of the developmental stag es. Their aligned and trimmed particial cob sequences of 350 bp were identical except for four positions where a synonymous substitution was seen in one of either sequences.
A 50% majority rule consensus trees based on cox1 data is illustrated in Fig. 1. All samples of the same species clustered together, supported by 1.0 posterior probability values. The species of the A. patrolus species group and its two subgroups respectively clustered together, however partly with lower posterior probability value support. A. buhid does not cluster with the A. patrolus group.  Distribution. The species is known from Busuanga, Mindoro and Palawan (Phil ippines) and is common on these islands.
Remarks. Morphological variations between the population of different islands are not evident, but the cox1 gene varies in a few more substitutional sites than within a population of one and the same island. The larva was described and illustrated in Freitag and Balke (2011).
Ecology. Both, adults and larvae, are usually collected from boulder and rock sur faces, or submerged rootpacks in run and riffle sections. The species is predominantly found in clean, small to medium sized permanent streams. (Freitag andPangantihon 2010, Freitag andJäch 2007). Etymology. This small and probably rare species is named in reference to its type locality, the Tamaraw Falls on the island of Mindoro. The tamaraw is a small Mindoro endemic buffalo. The term is used as noun in apposition.
Colouration as in Fig. 3: ventral side, coxae, trochanter, and pronotum brown; en tire dorsal head capsule and mouthparts dark brown; elytra dark brown except for two pairs of yellow patches; anterior yellow elytral patches round, extending each between first and third row of elytral punctures, not reaching median or anterior elytral margin; posterior yellow elytral patches elongateoval to subtriangular, not reaching median, lateral, and apical elytral margin; antennae yellow (except for dark tips and scape); legs dominantly yellowish except for coxa, trochanter and brown areas around all articula tions, especially proximal and distal areas of femur, proximal third of tibia and distal third of fifth tarsomere.
Adult differential diagnosis. Ancyronyx tamaraw superficially resembles A. sophiemarie from Sibuyan and A. minerva . The new species can be easily distinguished by the combination of elytral colour pattern (anterior yellow elytral patches circularly round, not reaching median or anterior elytral margin; posterior patches elongateoval to subtriangular), the predominantly yellowish legs, the brown (not black) pronotum and head, and it's aedeagus with wide and flat apical portion.
Larval diagnosis (based on sixth instar). Colour (Fig. 6) similar to that of Ancyronyx minerva (see Freitag and Balke 2011: figs 3, 11A-L), but most distinctly different by anterior median head portions (clypeus, anterior frons) pale; anterior yellow pronotal band small, limited to very most anterior portion; at least pro, meso, metanotum with small circularround (not broadly subtriangular) yellow pattern at medioposterior margin; abdominal segment IX with pale yellowish apex and a conspicuous dark pattern extending c. posterior 0.2-0.4; abdominal segment IX relatively longer than in larvae of A. minerva.
HW 0.29 mm; entire larva about 2.7 mm long. Body elongate very similar in the external characters to that of A. minerva, except for the following: Posterolateral projections (Figs 12A, F) of all abdominal segments short, generally not overreaching posterior segment margins.
Abdomen (Figs 6, 12A, E-H) without conspicuous dorsosagittal carinae except for the anterior half of segment IX (Fig. 12G); squamose setae at posterior rim of segments I-VIII large (Figs 12E, F). Ventral sclerites of segment I with distinct sagittal ridge in anterior half, reaching c. ½ to ⅔ of segment length. Apex of segment IX (Fig. 12G) emarginate (sometimes inconspicuous due to apical setae). Operculum (Fig. 12H) longer than in A. minerva (more than double as long as wide).
Larval differential diagnosis. The species can most easily be distinguished from A. minerva which looks superficially most similar by the partly pale colour pattern of the dorsal head, the narrower circular medioposterior pale pattern at pro, meso, and metanotum and the longer last abdominal segment with pale apical area and distinctly dark subapical portion. Variation between larval instars. The final and prefinal instar stages available for this study do not vary conspicuously except for their size.
Distribution. The species is known from the type locality in northeastern Ori ental Mindoro. Additional material that appears conspecific is known from Subic, Zambales, Luzon and Bohol (unpublished material at CFM and NMW).
Ecology. The specimens were collected in well oxygenated water from rock sur faces, submerged woodlitter and roots in run and riffle sections of the stream below Tamaraw Falls. Since all examined material comes from this, in fact clean and almost natural small mountain river, any detailed conclusion about the habitat and ecological requirements would be highly speculative. It is, however, surprising that not any single specimen was found at any other collection site in Mindoro so far.  Adult description. Body 1.4-1.6 mm long (CL + exposed portions of head & tergit VIII); CL: 1.25-1.38 mm; CL/EW: 1.9-2.1. Colouration as in Fig. 4: entire dorsal head capsule, mouthparts, pronotum, and elytra (except for two pairs of yel low patches) black; anterior yellow elytral patches extending from humeri mediad ap proximately up to second row of elytral punctures, not reaching median elytral margin; posterior yellow elytral patches oval, not reaching median, lateral, and apical elytral margin; ventral side, coxae, trochanter, femur, proximal half of tibia, areas around tibial and tarsomere articulations, and claws brown; at least distal half of tibia and proximal portion of fifth tarsomere yellowish; antennae yellow (except for dark tips and basal segment).
Sternite IX (Figs 13G, H) c. 340 µm long, with moderately long anterior strut (distal end broken off in specimen figured in 13G), apical corners rounded, each with one lateroapical seta and one inconspicuous sublateroapical seta; apical margin slightly broadly emarginate; longer paraproct almost reaching apical margin.
Ovipositor (Fig. 13O, P) c. 410 µm long. Stylus slender, rather staight, with vari ous apical sensilla. Coxite moderately stout, distinctly shorter than in specimens of the Ancyronyx patrolus species group, but longer than in those of the A. variegates group; outer margin concave; all over with several rather short and broad, peglike spines, increasing in size and density apically at lateral margins; inner margin pubescent; basal portion short. Valvifer moderately longer than coxite; fibula slightly curved.
Adult differential diagnosis. In its colour patterns, Ancyronyx buhid resembles A. patrolus, A. punkti and especially A. pseudopatrolus from Palawan. The new species can be easily distinguished by the combination of body morphometric and genital charac ters (body, especially abdomen and elytra, relatively wider (CL/EW c. 2.0; EL/EW c. 1.45) than in other species; legs not distinctly longer than body; coxite of ovipositor moderately stout; aedeagus with straight main piece (not widened subapically), almost contiguous parameres ventrally, very short and few parameral setae.
Larval diagnosis (based on sixth instar). Colour (Fig. 7) dorsally dominantly dark brown except for yellow lateral head, clypeus and labrum, most anterior portion of pronotum and the almost entire first abdominal segment; most specimens addi tionally with yellowish to pale brown (preterminal) abdominal segment VIII (at least posterior portion) and apex of abdominal segment IX (up to c. posterior 0.15). Legs, mouthparts, ventral head, thorax, and abdomen yellowish to pale brown, but some specimens with darker brown thoracic venter and ventral abdominal segment IX.
HW 0.31 mm; entire larva about 3.1 mm long. Body elongate, wider than that of A. minerva and A. tamaraw, but similar in the external characters, except for the following: posterolateral projections (Figs 7, 14A) of abdominal segments II-VI usu ally reaching or slightly overreaching posterior segment margins. Lateral rim of thorax and abdomen with scattered long, trichoid setae. Dorsal sagittal line slightly impressed from prothorax to abdominal segment V and without tubercles.
Pro, meso, metathorax and legs (Figs 7, 14F) almost as in A. minerva. Pronotum with rather inconspicuous small round signa (glabrous areas) in posterior half. Ventral sclerites of thorax (Fig. 14E) rugulose, not glabrous; venter of metathorax with conspic uous sagittal tuberculate ridge (similar to that of the venter in abdominal segment I).
Variation between larval instars. The available prefinal instar specimens vary only slightly from the description above, namely by the relatively slenderer thoracic and abdominal segments, the smaller and rather inconspicuous spiracles near the postero lateral projection, the slightly broader legs with fewer setae, and the relatively longer lateral setae on thorax and abdomen.
Larval differential diagnosis. The species can most easily be distinguished from any other known Ancyronyx larva by the obviously pale first abdominal segment. The general shape and the proportions of the larva of this species resemble those of the A. patrolus group, from which it can be additionally distinguished by the anterior yellow band, that is medially narrower (not extended as in several species of the A. patrolus group) and the character combination of long sagittal crest of the first abdominal segment venter, slightly impressed dorsosagittal line without protruding tubercles. From the species of the A. variegatus group, this larva can be distinguished easily by its spindleshape habitus (subsemicircular in cross section) and the rather short posterolateral appendages.
Distribution. Known only from Mindoro Island where this new species was re corded from various streams in the province of Oriental Mindoro.
Ecology. Both, adult and larvae of Ancyronyx buhid occur in medium sized, un polluted rivers in mountainous areas. This suggests an affinity to undisturbed habitats. The relatively highest abundances were found on submerged wood and rough rock surfaces in runs and riffles. Some root packs and partly submerged grass bunches in rif fles were also found to be densely colonised with the species. Much more rarely it was found among bottom gravels in runs and calm pools, where specimens were possibly just shifted by drift.
Remarks. One male specimen from site "TR2f" varies in regard to the primary and secondary sexual characters, namely the length of tergite VIII, ventrite 5 and ae deagus. Since all other characters do not differ from the type material, this is regarded as an abnormality caused during pupation. Larval diagnosis (based on a single presumably sixth instar specimen). Colour in last instar larva predominantly brown as in Fig. 9; dorsal head darkest to almost black at pronotal disc; lateral head, antennae, anterior and lateral pronotal margins, legs (except for tip of claw), lateral abdominal segment margins and conical projec tions, posterior abdominal tip and areas around the sagittal line (especially thoracic area) distinctly paler, yellowish to pale brown. Ventral side entirely pale except for pale brown gula, maxillae and labium; ventral part of genae conspicuously dark brown.

Ancyronyx schillhammeri
HW c. 0.60 mm; entirely c. 3.8 mm long. Body shape of the A. variegatus group type, generally very similar to that of A. procerus (comp. Freitag and Balke 2011: 72-75) in the external habitus, except for the follow ing: Posterolateral projections (Figs 9, 15A, B) of abdominal segments IV-VIII slightly broader and stouter; spiracles distinctly larger, very prominent; entire lateral margin with distinct long trichoid setae; tubercles much more prominent (especially at dorsal side). Head (Figs 15C-F) distinctly widest basally, slightly conical anteriad; without me dian pointed projection at frons; the pair of sublateral anterior projections between antenna and clypeus rather shallow (Fig. 15C). Frons moderately densely and equally covered with prominent setiferous tubercles; setae very short. Genae rugose, with irreg ular depressions, ridges, and scattered tubercles (Fig. 15D); lateral glabrous area with stemmata elongately subtriangular (Fig. 15D). Antenna (Fig. 15E) less than half as long as head. Scape short, slightly longer than wide, with subapical fringe of stout sensilla; pedicel cylindrical less than three times as long as scape and c. three times as long as wide, with few apical sensilla; flagellum and sensorium as in A. procerus. Ventral side (Fig. 15F) with welldeveloped longitudinal crests bordering the stipes. Labrum broad, c. 3.5 times as wide as long; lateroapical edges rounded; entire visible dorsal surface with small setiferous tubercles. Maxilla (Fig. 15E, F) almost as in A. procerus. Labium (Figs 15E,F) with mentum widest in apical half; pair of moderately long trichoid setae inserted sublaterally at anterior 0.3; some additional trichoid setae present at lateral margin in apical half; pair of apicolateral teeth inserted at a distinct subapical crenation; submentum straight, without conspicuous median ridge, apically broadly concave. Prothorax (Fig. 15G, H) slightly wider than long; tergum with irregularly shaped and round signa in posterior half; median and submedian pairs clearly defined by bor dering tubercles, not fused (Fig. 15G). Venter of prothorax (Fig. 15H) similar to that in A. procerus, but anterior sclerites distinctly shorter, oblique, not subtriangular; anterior margin with conspicuous setiferous tubercles; anterior and lateral sclerites partly fused near anterior prothorax margin; transverse sutures dividing the lateral sclerites distinctly ending before lateral margin. Venter of meso and metathorax (Fig. 15H) with more conspicuous setiferous tubercles particularly at posterior margins of anterior sclerites.
Legs (Figs 15H,J) proportioned as in A. procerus, but tubercles and setae larger and more distinct. Subbasal tooth of claws long and trichoid, overreaching tip of claw.
Abdominal terga (Figs 9, 15A, B, I) with slightly depressed groove along sagittal line at least from 1 st up to 4 th segment; posterior terga margins with rim of squamose setae. Posterior venter margins with rim of trichoid setae. Segment IX (Fig. 15I) dorsally with shallow sagittal crest formed by densely arranged tubercles bearing large trichoid setae; apex widely rounded to slightly truncate; ventral side rugose, not glabrous. Operculum without longitudinal ridges, entirely rugose and covered with conspicuous scattered setae.
Larval differential diagnosis. The larvae of Ancyronyx schillhammeri are easily dis tinguishable from all other known Ancyronyx larvae of Mindoro by their larger size, the somewhat dorsoventrally depressed habitus, the much larger and protruding poste rolateral appendages, as typical for the A. variegates species group. Among this species group, it resembles the larva of A. procerus in colour, but can be clearly distinguished by the absence of the pointed projections at median frons, the more shallow projec tions between antenna and clypeus, the larger and more protruding spiracles, the coni cal head shape, and the surfaces of head, thorax, and abdomen that are densely covered with larger tubercles bearing long conspicuous setae. From A. helgeschneideri it is easily distinguishable by the darker colour, the pale dorsosagittal stripe, the dark dorsal ab dominal segment IX and the broader and conical head.
Distribution. Only known from Oriental Mindoro and one locality of Occidental Mindoro near San Jose.
Ecology. Ancyronyx schillhammeri occurs exclusively on submerged wood. Decay ing pale light woods appear to be preferred by the species. At the sites of the Baroc River catchment, which were sampled regularly throughout the year, the abundance of this species was found to increase distinctly towards the end of the dry season (Febru ary to April) and declines rapidly with the beginning rainy season, presumably due to wash out. It is found in both habitat types: clean, cool and torrent rhithral creeks and rivers as well as warm, mesosaprobic lowland streams. This suggests less specific ecolog ical requirements in terms of stream hydraulics, water temperature, and water quality.

Discussion
During the last two decades, the Philippine Islands have received increasing attention in biodiversity research, not least because they are classified as a major biodiversity hotspot in global scale (Myers et al. 2000). However, for several taxa including Elmidae and other freshwater macroinvertebrates, it still requires substantial efforts to record and to describe the majority of species and their distribution. Four species of Ancyronyx are now recognised and formally described from Mindoro Island based on the study of a copious collection of museum specimens and the material retrieved from a comprehensive survey of the Baroc River Catchment in southern Min doro. Only one of them, A. minerva, is recorded beyond Mindoro. A. schillhammeri and A. buhid appear to be endemic to the island. The high rate of island endemism reflects the biogeographic history of the island. Mindoro is a remnant of a fragment of the Eurasian continental margin and is not part of the Luzon arc of islands of marine volcanic origin (Hall 1998). Despite its recent close vicinity to Luzon, the two islands remained largely isolated in the Quaternary, even during Pleistocene when low sea levels have formed land bridge interconnections of several Philippine islands, but presumably Greater Pala wan, Mindoro, and Greater Luzon remained separated based on Pleistocene sealevel low stands represented by the 120 m isobath (Sathiamurthy and Voris 2006).
Therefore, it requires more in depth phylogenetic and biogeographic studies to explane the distribution of A. minerva at both sides of the Mindoro Strait.
The phylogenetic relationship of A. buhid with other members of the genus is still ambiguous. Several taxa (probably new species) from the Philippines and Sulawesi that resemble A. buhid still await their description (unpublished material of the author and at NMW). After this material has undergone detailed study and molecular genetic analysis sound conclusions might be drawn.
The fact that Ancyronyx (and very most other Elmidae) live permanently under water and respire by a microplastron (Kodada and Jäch 2005) makes them sensitive to water pollution. The vivid and specific colour patterns of adult Ancyronyx species enabling an easy identification, as well as the availability of regional identification keys for both, larvae and adults, allow their potential use as bioindicators. Among the species of Mindoro, A. schillhammeri was recorded from clean to moderately polluted streams making it unsuitable as a bioindicator. The remaining Mindoro species seem to be ecologically adapted to clean and rather undisturbed waters. However, Ancyronyx tamaraw is too rare to serve as good bioindicators and A. minerva is occasionally de tected in slightly polluted streams in low abundances (Freitag and Pangantihon 2010), suggesting a low indicator strength. Therefore, A. buhid in particular has the highest potential value to be used as saprobic indicator. Its frequent occurrence in suitable habitats and the easy identification by the distinguishing elytral colour pattern and broad elytral shoulders in adults, as well as the unique yellow abdominal pattern in lar vae, make it a suitable tool for biomontoring, even for nonentomologists. However, ecological evaluations of larger scale are needed to confirm these preliminary findings.

Acknowledgements
Special thanks are due to Research Assistant Mr. Clister V. Pangantihon for his sig nificant support during the field sampling at the Baroc River, preliminary sorting and labelling of the collected material and Dr. Michael Balke for providing access to his laboratory facilities at ZSM for molecular genetic analysis.
The project was kindly funded by PHERNet, a program of the Philippine Com mission of Higher Education (CHED). Financial support was also kindly provided by an Ateneo de Manila University Loyola Schools Scholarly Work Grant (SOSE 8 2012).
A fund for open access publication of this paper was granted by the "Encyclopedia of Life (EOL) Open Access Support Project" (EOASP).
Field sampling of the Baroc River Catchment Survey were made possible by the permission and support of the PENRO Oriental Mindoro, Calapan, SENRO Roxas, Municipal Government of Roxas, and the National Commission for Indigenous Peo ple (NCIP). The author wants to express his deep gratitude to Brg. San Vicente, Roxas, especially Captain Ronel S. Sescar, and all Brgy. Kagawats and Tanods that warmly welcomed the Ateneo team, guided the expeditions and provided accommodation. Many thanks are also due to all members of the Buhid communities of Tagaskan and Tauga Diit who gave official permission (FPIC & MOA) for the field sampling in their ancestral lands, and guided and accommodated the excursion participants.
Furthermore, the author wishes to express his gratitude to the curators of the World Water Beetle Collection & Research Centre at NMW, Dr. Manfred A. Jäch and the Coleoptera section of ZMUC, Dr. Alexej Solodovnikov, for access to their material col lections and the loan of specimens. Sincere thanks are due to Dr. Jäch and an anonymous reviewer for their helpful comments on the manuscript.