瀕危條紋球背象鼻蟲的物種判定研究與蘊含的保育意義

全文

(1)國立臺灣師範大學生命科學系碩士論文. 瀕危條紋球背象鼻蟲的物種判定研究與蘊含的保育意義 Integrated species delimitation of an endangered weevil, Pachyrhynchus sonani (Coleoptera: Curculionidae), uncovers a cryptic species, Pachyrhynchus jitanasaius sp. nov., endemic to Green Island, Taiwan, and the implications for its conservation. 研究生：陳彥廷 Chen, Yen-Ting 指導教授：林仲平博士 Dr. Lin, Chung-Ping. 中華民國 106 年一月.

(2) 1.

(3) 致謝科學研究是一條漫漫長路，平凡的我很幸運的能通過種種關卡取得碩士學位，我要感謝我的口試委員徐堉峰老師與鄭明倫老師給予論文上寶貴的意見，然這一長路上最感謝的是林仲平老師，林仲平老師在我大二那年讓成績不好的我進入實驗室與葉人瑋學長學習，開啟我的研究之門與視野拓展，這讓平凡的我萌生「我也許可以做研究」的念頭。進而展開我的碩士班生涯，我原本以為我可以一個人搞定生活與研究上的大小事，並且可以不顧及任何人的冷言熱諷，享受研究工作，但是我發現我一個人根本做不到，更在碩士三年級的時候受不了責任與內心的譴責選擇逃避時，所幸林仲平老師幫助才得以完成學業。在這條路上特別感謝惠芸學姊在教我正確的實驗態度與精神，感激馥慈師母處理行政上的問題，謝謝祐薰學姐、若凡學姐與秉宏學長教我研究的方法與同儕文傑給予鼓勵與砥礪，以及謝謝實驗室的わたる、英元學長、Leo、Princess、佾修、露翊、喻仁、震邑、騰宇與俊佑給予實驗上的建議與生活上的歡樂。最後，我要感謝我的雙親，在家庭並不富有和孩子並不有才的情況下，依然支持我讀了十九年又三個月的書，我定當學以致用，不辜負父母一片苦心，為科學研究貢獻綿薄之力。. 2016 冬陳彥廷書於師大分部. 2.

(4) 目錄摘要............................................................................................................4 Abstract...................................................................................................5-6 Introduction ..........................................................................................7-10 Materials and Methods........................................................................10-18 Results ................................................................................................19-23 Discussion ..........................................................................................23-27 Taxonomy ...........................................................................................27-31 References ..........................................................................................32-39 Table ...................................................................................................40-41 Figure .................................................................................................42-51 Supporting Information ......................................................................52-99. 3.

(5) 摘要海洋性島嶼是產生新種和豐富特有性的高生產力棲地，主要歸因於地理隔離、族群量小與在地適應等特性。然而島嶼物種的分化時間短和細微的形態或生態分化可能蒙蔽物種鑑定，因此島嶼上的隱蔽特有種可能被低估。瀕危的條紋球背象鼻蟲 Pachyrhynchus sonani Kôno, 1930 (Coleoptera: Curculionidae: Entiminae: Pachyrhynchini)是棲息於台灣—呂宋島鏈中綠島與蘭嶼的特有種，族群間存在多樣的色彩與寄主植物變異，因此衍生出物種界線與隱蔽多樣性程度的問題。本研究結合形態(體型大小及形狀、生殖器形狀、色彩與鱗片)、遺傳(四組基因與 RAD-seq)和生態(寄主植物與島嶼來源)多樣性的整合物種判別方法來測試條紋球背象鼻蟲的物種界線。結果顯示除了色彩光譜外，所有雄蟲的形態特徵在兩島嶼間重疊，但統計上顯著不同；相反地，兩族群間的雌蟲形態極少分化。兩島嶼間的條紋球背象鼻蟲族群在寄主植物頻率顯著不同，遺傳分群與親緣關係分析建立其為兩個有效的演化種。整合物種判別分析支持綠島和蘭嶼條紋球背象鼻蟲族群為兩個獨立物種。我將綠島族群描述為新種 P. jitanasaius sp. nov. Chen & Lin 並建議將其列為瀕危物種。本研究結果顯示，在這兩島嶼上的特有生物的島間種化事件可能比之前所知的更加普遍，並強調小型海洋性島嶼的隱蔽多樣性可能仍然被大量低估。. 關鍵字：整合物種判別、海洋性島嶼、臺灣呂宋島鏈、親緣種概念、球背象鼻蟲屬. 4.

(6) Abstract Oceanic islands are productive habitats for generating new species and high endemism, which is primarily due to their geographical isolation, smaller population sizes and local adaptation. However, the short divergence times and subtle morphological or ecological divergence of insular organisms may obscure species identity, so the cryptic endemism on islands may be underestimated. The endangered weevil Pachyrhynchus sonani Kôno, 1930 (Coleoptera: Curculionidae: Entiminae: Pachyrhynchini) is endemic to Green Island and Orchid Island of the Taiwan-Luzon Archipelago and displays widespread variation in coloration and host ranges, thus raising questions regarding its species boundaries and degree of cryptic diversity. I tested the species boundaries of P. sonani using an integrated approach that combined morphological (body size and shape, genital shape, coloration and cuticular scale), genetic (four genes and restriction site-associated DNA sequencing, RAD-seq) and ecological (host range and island of origin) diversity. The results indicated that all of the morphological data sets for male P. sonani, except for the colour spectrum, reveal overlapping but statistically significant differences between islands. In contrast, the morphology of the female P. sonani showed minimum divergence between island populations. The populations of P. sonani on the two islands were significantly different in their host ranges, and the genetic clustering and phylogenies of P. sonani established two valid evolutionary species. Integrated species delimitation combining morphological, molecular and ecological characters supported two distinct species of P. sonani from Green Island and Orchid Island. The Green Island population was described as P. jitanasaius sp. nov. Chen & Lin, and it is recommended that its threatened status be recognized. Our findings suggest that the inter-island speciation of endemic organisms inhabiting both islands may 5.

(7) be more common than previously thought, and they highlight the fact that the cryptic diversity of small oceanic islands may still be largely underestimated.. Keywords: Integrated species delimitation, Oceanic islands, TaiwanLuzon Archipagos, Phylogenetic species concept, Pachyrhynchus. 6.

(8) Introduction Oceanic archipelagos are one of the most productive arenas on the earth for the generation of new species and high endemism (Grant, 1998; Gillespie & Roderick, 2002). The geographical isolation, together with the geological and ecological dynamics of islands, greatly facilitates in situ speciation and subsequent diversification in these remote habitats (MacArthur & Wilson, 1967; Heaney, 2000; Whittaker et al., 2008). Neighbouring islands of a given oceanic archipelago often harbour different but closely related species because of geographical proximity, limited dispersal and sexual or ecological divergence caused by local adaptation. For example, the classic studies of Darwin's finches of the Galápagos Islands (Grant & Grant, 2014), the Anolis lizards of Caribbean Islands (Losos, 2009) and the fauna of the Hawaiian Islands (Wagner & Funk, 1995) all demonstrate the enormous adaptive radiation and rapid diversification of closely related island organisms. However, short divergence times and limited phenotypic divergence of endemic insular organisms may obscure species identity, so the cryptic endemism may be underestimated on islands (Bickford et al., 2007). Pachyrhynchus weevils (Germar, 1824) (Coleoptera: Curculionidae: Entiminae: Pachyrhynchini) exhibit the most diverse and spectacular colour radiation of all of the terrestrial organisms of the Old World tropics (Wallace, 1895; Schultze, 1923), and this remarkable array of coloration was found to be a result of light interference by a three-dimensional photonic poly-crystal structure within the cuticular scale of the exoskeleton (Welch et al., 2007; McNamara et al., 2013). The bright colours of Pachyrhynchus function as aposematic signals that deter attacks by lizard predators (Tseng et al., 2014). These flightless Pachyrhynchus weevils have the highest species diversity in the Philippine archipelago and adjacent oceanic islands (Schultze, 1923), and 7.

(9) approaching the northern boundary of its distribution, Pachyrhynchus inhabits a few remote volcanic islands of the Taiwan-Luzon Archipelago, which extends from Luzon to Okinawa and includes Green Island (Lyudao) and Orchid Island (Lanyu) Islands located offshore of southeastern Taiwan (Kano, 1929; Kôno, 1930; Starr & Wang, 1992) (Fig. 1a). The five currently recognized species of Pachyrhynchus are distributed on Green and Orchid Islands, including P. insularis Kano, 1929; P. sarcitis kotoensis Kôno, 1930; P. sonani Kôno, 1930; P. tobafolius Kano, 1929 and P. yamianus Kano, 1929. Except for the rare endemic P. insularis of Orchid Island, the other four Pachyrhynchus species occur on both Green and Orchid Islands but differ in their relative abundance. Pachyrhynchus sarcitis kotoensis, P. tobafolius and P. yamianus are more populous on Orchid Island than on Green Island, whereas as P. sonani is more abundant on Green Island (Starr & Wang, 1992). Since 2009, the Wildlife Conservation Act of Taiwan officially has listed all five of the endemic Pachyrhynchus species of Green and Orchid Islands as Category II protected species (i.e., rare and valuable species) (Chao et al., 2009). The criteria used to assess their threat categories were mainly based on the level of endemism, collection pressure and habitat loss, but in addition to the above commonly applied criteria, systematic and phylogenetic information based on ecological, morphological and molecular data can provide essential information for identifying distinct evolutionary lineages, delimiting species boundaries and setting conservation strategies and priorities (reviewed in Pellens & Grandcolas, 2016). This knowledge is especially important for cryptic species and complexes of endangered species, which may consist of multiple unrecognized species that are even rarer than the nominal species (Bickford et al., 2007; Wiens, 2007). At present, the five Taiwanese 8.

(10) Pachyrhynchus species are primarily distinguished by body sizes and shapes and by the variation in the range of brightly coloured spots and stripes decorating their thoraces, legs and elytra (Fig. 1b & d) (Kano, 1929; Kôno, 1930). Although the aposematic coloration of a particular Pachyrhynchus species may have been subject to strong stabilizing selection to minimize deviation from its effective predator-deterring colour pattern (Tseng et al., 2014), a cursory examination of available P. sonani specimens revealed widespread colour differences between and within island populations (Fig. 1e). For example, the posterior-lateral elytrous stripes of P. sonani from Orchid Island are elongate, whereas the P. sonani from Green Island have reduced, small spherical spots (Fig. 1e, character 6). In addition to colour variation, the available host plant records for P. sonani reveal variable host plant ranges that characterize the populations of each island. Pachyrhynchus sonani frequently occurs on the sea poison tree, Barringtonia asiatica (Lecythidaceae), on Orchid Island, but on Green Island, it is often found on Ceylon ardisia, Ardisia elliptica (Primulaceae). Therefore, the variable colour pattern and host plant range of P. sonani raise questions regarding its species boundaries and the number of cryptic species among the populations of the islands. This study aimed to test the species boundaries of P. sonani from Green Island and Orchid Island using an integrated approach combining characters of phenotypic, genetic and ecological diversity. A sympatric and congeneric species, P. tobafolius (Fig. 1b), that shows minimal between-island variation in colour and host range was simultaneously analysed for comparison using multiple character data sets. First, I analysed multiple phenotypic traits, including the body size and shape, genital shape, coloration and cuticular scales from the two island populations of P. sonani. Second, I reconstructed the phylogenies and tested for genetic differentiation between the island populations of P. 9.

(11) sonani using two molecular data sets, four gene sequences and a genomewide sample of SNPs (single nucleotide polymorphisms) generated from RAD (restriction site-associated DNA) sequencing. Third, I quantified the level of ecological differentiation in host plant range between the island populations of P. sonani. The recently developed Bayesian species delimitation method of integrating multiple morphological, genetic and ecological traits can improve the accuracy of estimated species boundaries that only use genetic data (Edwards & Knowles, 2014; SolísLemus et al., 2015). Finally, I estimated the P. sonani species boundary by integrating the characters of phenotypic, genetic and ecological diversity under a Bayesian coalescent framework in iBPP (Solís-Lemus et al., 2015). Our integrated species delimitation results provide sufficient evidence for recognizing the population of P. sonani from Green Island as a distinct species. I conduct a taxonomic description of this new species, Pachyrhynchus jitanasaius sp. nov. Chen & Lin.. Materials and Methods Insect sampling and preparation I collected weevils by hand or with insect nets from 20 sites on Green (22°39’49 N, 121°29’24 E) and Orchid Island (22°03’49 N, 121°32’41 E) (Fig. 1) between April 2006 and August 2014 (Fig. 1c, Appendix S1); the specimens were preserved in 95% EtOH after capture. The host plant species of the specimens were identified either by the presence of the insects on the plants or by observing of them feeding on the leaves and bark. I dissected the weevils to separate the legs and the genitalia. The genitalia were bathed in 15% KOH at 70°C for 15 minutes to clear the structure and were then kept in 50% glycerin for photographs and illustrations, and the thoracic muscles and legs were preserved in 95% EtOH in a -20°C freezer for subsequent molecular analyses. The insect 10.

(12) bodies were secured with insect pins on a Styrofoam plate and dried in an oven at 50°C for 72 hours; and they were also used later for photographs.. Morphometric analyses The dorsal and lateral body (Fig. 2a & b), lateral aedeagus (Fig. 2c) and ventral sternite VIII of the female genitalia (Fig. 2d) were photographed using a digital camera (EOS700D, Canon, Tokyo, Japan) mounted on a stereo microscope (SZ61, Olympus, Tokyo, Japan) at 10– 45× with a single white light projected from above at an angle of 45 degrees. The camera and the specimens were manually adjusted to be parallel to each other. The photographs were saved in JPEG format (300 dpi) and then imported into tpsDig2 (Rohlf, 2005). Twelve measurements of body shape [rostrum width (RW), head width (HW), thorax width (TW), abdomen width (AW), rostrum length (RL), head length (HL), thorax length (TL), abdomen length (AL), rostrum height (RH), head height (HH), thorax height (TH) and abdomen height (AH)] (Fig. 2a & b) were obtained by analysing the photographs in Photoshop CS5 (Adobe Systems Inc., San Jose, CA). Each of the morphological structures were photographed three times to calculate the measurement error (%ME) of the measurement by comparing the variation within and between individuals using the formula developed by Bailey and Byrnes (1990): %ME = [S2within / (S2within + S2among)] × 100%, where S2within = MEwithin, and S2among = (MEamong + MEwithin) / 3. MEwithin and MEamong are the withinindividual and between-individual mean sums of squares of a Model II ANOVA, and the measurement errors of the morphological structures were 6.58% for the body (n = 92), 5.15% for the aedeagus (n = 59), and 1.50% for the female genitalia (n = 22). The significance of the measurement differences between island weevil populations was analysed using MANCOVA (multivariate analysis of covariance) in SPSS v. 17.0 11.

(13) (Norusis, 2005). The difference in genital size (male: distance between landmark 1 and 12 of the aedeagus, Fig. 2c; female: distance between landmark 2 and 3 of the sternite VIII; Fig. 2d) was analysed using a t-test in SPSS. Eleven and ten body landmarks located dorsally and laterally, respectively, were identified to describe the body shape of the weevils (Fig. 2a & b). Twelve landmarks were chosen to characterize the aedeagus including two landmarks located at the apex and base (Fig. 2c, no. 1 & 12) and ten semi-landmarks (Fig. 2c, no. 2–11), which were generated by crossing the inner and outer edges of the aedeagus with each of ten radii spaced 30 degrees apart, beginning from the midpoint between landmarks no. 1 and 12. Five landmarks were selected to describe the ventral sternite VIII of the female genitalia (Fig. 2d). The coordinates of all of the landmark locations were transformed into eigenvectors using GPA (general Procrustes analysis) (Rohlf & Slice, 1990) in CoordGen6h of the IMP (Integrated Morphometric Package) (Sheets, 2004), and the eigenvectors were summarized using PCA (principal component analysis) in PCAGen6p of the IMP. The resulting first and second PCs (principal components) and eigenvectors were analysed using MANCOVA in SPSS to test the significance of the difference in the body shape between the populations of the two islands. The shape variables (uniform components and partial warps) of the body and genitals were treated as dependent fixed variables, and the centroid size was used as a covariate. Goodall’s F test (Anderson, 1984) was employed to test the significance of the differences in the body and genital shape eigenvectors between island populations using TwoGroup6h of the IMP. Thin-plate spline deformation grids were generated between the consensus of the two populations in PCAGen6p to visualize the level and location of the variations in the body and genital shapes. 12.

(14) Coloration and cuticular scale I quantified the colour of the weevils by measuring the reflectance spectrum of a pre-defined surface area (0.785 mm2 at the lateral-posterior end of the left elytra, Fig. 2b, circle d) located within the colour stripe at a distance of 1 mm using a spectrometer (detection range: 250–800 nm, Jaz spectrometer, Ocean Optics, Florida, USA), which was connected to a reflection probe (ZFQ–13101) with a deuterium–tungsten halogen light source (DH–2000–BAL, Ocean Optics, Florida, USA). A diffuse reflectance standard (WS–1–SL) was used for calibration in SpectraSuite (Ocean Optics, Florida, USA). I measured the reflectance spectra of each individual five times, and the mean values of colour parameters (hue, brightness and saturation) were calculated and used for the analyses. The hue was defined as the wavelength at the highest peak of the reflectance spectrum, and the brightness was defined as the reflection ratio of the wavelength at the highest peak of the spectrum. The colour saturation was defined as the total area below the reflectance curve that was between the visible spectrum of the diurnal lizard predators (440–625 nm), which generally have four spectrally distinct classes of colour vision including ultraviolet-sensitive (UVS) (364–383 nm), short wavelength-sensitive (SWS) (440–467 nm), medium-wavelength sensitive (MWS) (483–501 nm) and long-wavelength-sensitive (LWS) (560–625 nm) visual pigments (Loew et al., 2002; Yewers et al., 2015). The UVS wavelength area was not used to calculate colour saturation because both P. sonani and P. tobafolius had low UV reflectance (Fig. 3, < 40%). The significance of the spectral difference between island populations was tested using ANOVA in SPSS. To characterize the fine-scale morphology of colour scales, the cuticular scales on the surface of colour stripes were coated with 13.

(15) platinum-palladium (300 Å ) using an ion coater (Model IB–2, Eiko, Taipei, Taiwan) and then observed through a SEM (scanning electron microscope) (SU1510, HITACHI, Tokyo, Japan). The number of scales and scale types in a pre-defined area (450 × 337.5 µm2) [P. sonani (n=41): stripe crossing the left elytra (Fig. 2a, square e); P. tobafolius (n=21): the second dorsal spot of the left elytra (Fig. 2a, e square f)] were recorded using SEM photographs at 400× magnification. The scale types were defined as follows: type I, the carving area is < 2/3 of the scale surface (Fig. 4a); type II, the carving area is > 2/3 of the scale surface (Fig. 4b); type III, the carving area is > 2/3 of the scale surface, and the scale has a major depression in a half-moon shape (Fig. 4c). The number of scales and the frequency of scale types were analysed using ANOVA and chi-square statistics, respectively, in SPSS.. DNA extraction, sequencing and phylogenetic analyses Genomic DNA was extracted from the legs of the weevils using a FavorPrep™ Tissue Genomic DNA Extraction Mini Kit (Favorgen Biotech Corp., Ping-Tung, Taiwan). The fragments of mitochondrial cytochrome c oxidase I (cox1, 700 bps), NADH dehydrogenase subunit 2 (nd2, 1000 bps), nuclear elongation factor 1α (EF1α, 800 bps) and internal transcribed spacer (ITS, 1700 bps) were amplified using the published [cox1 (LCO–1490 and HCO–N–2198, Folmer et al., 1994), EF1α (EF1–Bf and EF–Br, Hernández-Vera et al., 2013) and ITS (ITS18Sr and ITS 28Sr, Weekers et al., 2001)] and newly designed primers [nd2, 5’–GATTTACTGCTTGAATAGGATTAG–3’ (P–ND2–F) and 5’–CATAATGAAAATAGATTTGTCATG–3’ (P–ND2–R), Tseng et al., submitted]. The polymerase chain reaction (PCR) was carried out in a thermo cycler (9700, Applied Biosystems® , Foster City, USA) with the following profile: (1) an initial denaturation at 94°C for 3 minutes 14.

(16) followed by (2) 35 cycles of denaturation at 94°C for 1 minute, annealing for 45 seconds at 50°C for cox1, 46°C for nd2, 62°C for EF1α and 56°C for ITS, extension at 72°C for 1 minute, and (3) a final extension step at 72°C for 10 minutes. The 25 µl PCR reaction contained 1 to 2 µl of genomic DNA, 17.5 µl of ddH2O, 2.5 µl of pro Taq 10× buffer, 2 µl of dNTP, 1 µl of forward primer, 1 µl of reverse primer, and 0.5 µl of GoTaq (PROTECH, Taipei, Taiwan). The PCR products were purified by treatment with shrimp alkaline phosphatase/exonuclease I (USB Products, Affymetrix, OH, USA) and then subjected to sequencing in an ABI 3730XL DNA Analyzer (Applied Biosystems® , Foster City, USA). The DNA sequences were manually edited using SeqMan in DNASTAR (LASERGENE, Swindell & Plasterer, 1997) and then aligned using Clustal W in MEGA v. 6.06 (Tamura et al., 2013) (GenBank accession numbers in Appendix S2). Pachyrhynchus orbifer (Babuyan and Fuga Islands) and P. infernalis (Yaeyama Islands) were used as outgroups for the P. sonani phylogenetic analyses because of their close relationships (Tseng et al., submitted), and the Pachyrhynchus sp. of Calayan Island, which has a morphology similar to P. tobafolius, was used as an outgroup for P. tobafolius. MP (maximum parsimony) analyses were conducted in PAUP* v. 4.0b10 (Swofford, 2002) using a parsimony ratchet algorithm implemented in PAUPRat (Nixon, 1999; Sikes & Lewis, 2001) to search for the most parsimonious trees. PB (parsimony bootstrapping) (Felsenstein, 1985) was conducted with heuristic searches of 1,000 replications, and each replication contained 100 iterations of the TBR (tree-bisection-reconnection) swapping algorithm. ML (maximum likelihood) analyses were conducted using RAxML v. 8.2 (Stamatakis, 2006) under the GTRGAMMA model, and the ML analyses generated 1,000 distinct ML trees from 1,000 randomly selected parsimony tress. LB (likelihood bootstrapping) was conducted using 1,000 iterations with 15.

(17) randomly selected starting trees. The best-fit nucleotide substitution model from the Bayesian analyses was selected based on the BIC (Bayesian information criterion) using jModeltest v. 2.1.7 (Darriba et al., 2012): JC for nt1 of cox1, F81 for nt2 of cox1, TrN for nt3 of cox1, HKY for nt1 and nt2 of nd2, TrN for nt3 of nd2, K80 for nt1 of EF1α, F81 for nt2 of EF1α, JC for nt3 of EF1α, F81 for nt1 and nt3 of ITS, and HKY for nt2 of ITS. Bayesian analyses were conducted in MrBayes v. 3.2 (Ronquist et al., 2012) using the MCMC (Markov chain Monte Carlo) searches for 2×108 generations with a sampling frequency of every 10,000 generations and a burn-in of 2.5×107 generations. The convergence of the MCMC runs was confirmed using the effective sample size (ESS) of the model parameters (ESS mean for P. sonani: cox1 = 8669, nd2 = 8927, EF1α = 8863, ITS = 8912 and combined = 9001; P. tobafolius: cox1 = 8735). The BPP (Bayesian posterior probability) was calculated from a 50% majority rule tree after discarding a burn-in of 25% of the trees. The colouration of the weevils was coded as 20 and 4 discrete characters for P. sonani and P. tobafolius, respectively (Appendix S3–6), and the colour character matrices (Appendix S3 & S4) were used for parsimony analyses in PAUP with a PAUPRat ratchet algorithm.. Library preparation, RAD sequencing and population genetic structure The DNA samples were extracted using a Qiagen Blood and Tissue DNA Extraction Kit (Qiagen, Boston, USA) from the thoracic muscles of the weevils and then suspended in TE buffer and stored at -20°C. They were then quantified using a Qubit® dsDNA HS Assay Kit (Thermo Fisher Scientific Inc., Boston, USA) and diluted into 5 ng/µl in a total volume of 10 µl, which resulted in 50 ng of DNA for each samples for library preparation. I followed the MSG (multiplexed shotgun sequencing) protocol (Andolfatto et al., 2011) for library preparation. The 16.

(18) genomic DNA samples were first digested using a Nde1 (5'–CA^TATG– 3') restriction enzyme, and the individual barcodes and adaptors were then ligated to the digested genomic DNA samples. The individually barcoded RAD samples were pooled and precipitated overnight, and I then size-selected the pooled RAD samples from 350 to 450 bps and eluted the DNA fragments in the targeted range in a Pippin Prep System (Thermo Fisher Scientific Inc., Boston, USA). The eluted samples were separated for enrichment into four PCR reactions using FCR1 and FCR2 primers; the PCR products were pooled, purified and then single-ended sequenced in a HiSeq 2500 system (Illumina, San Diego, USA). The sequencing reads were preliminarily screened in PROCESS RADTAGS (Illumina, San Diego, USA) for adaptor sequence contamination and then de-multiplexed based on the individual barcodes of six base pairs. I followed the standard STACKS pipeline for SNP-calling (Catchen et al., 2013). The de-multiplexed reads from the individual genome were stacked in USTACKS, and CSTACKS was used to generate the SNP catalogue for all samples. SSTACKS was used to match the SNPs from the individual genome to the catalogue, and I proposed that the samples from the two islands were individual populations. I called SNPs using at least 10× in the sequencing coverage and allowed up to 50% missing data in an SNP and 30% missing data among populations. The SNP data were output into different data formats for analyses using POPULATIONS, and I used the Bayesian clustering method in STRUCTURE v. 2.3.4 (Falush et al., 2003) to determine the level of the population admixture and identify the potential migrants and hybrids. The MCMC process was run for 2×106 generations with a burn-in of 5×105 generations. The estimated number of clusters (K) was set to range from 1 to 10 for all individuals to explore their optimal values, and the delta K statistic was calculated to determine the most appropriate number of genetic clusters in 17.

(19) STRUCTURE HARVESTER web v. 0.6.94 (Earl & vonHoldt, 2012).. Integrated species delimitation For P. sonani iBPP analyses, individuals were assigned to putative taxa based on island origin, clustering of phenotypic traits and a uniform rooted guide tree (Fig. 5b) (n = 14 males and 5 females from Green Island and 11 males and 6 females from Orchid Island). The character matrix included: (1) biogeographical (location of island), (2) molecular (cox1, nd2, EF1α and ITS) and (3) morphological traits [the first PC scores for the dorsal and lateral body shape, the first PC scores for coloration (hue, brightness and saturation), the number of scales, the frequency of scale types, discrete colour traits (traits 5, 6, 9, 10, 11, 12, 13, 14, 18 and 19 in Appendices S3 and S5) and the first PC scores for the male and female genitalia]. I specified four combinations of the prior distribution for the ancestral population size (θ) and the root age of the tree (τ): (1) θ = G (1, 10) and τ = G (1, 10), assuming large population sizes and a deep divergence time; (2) θ = G (1, 10) and τ = G (2, 2000), representing large population sizes and a shallow divergence time; (3) θ = G (2, 2000) and τ = G (1, 10), indicating small population sizes and a deep divergence time; and (4) θ = G (2, 2000) and τ = G (2, 2000), signifying small population sizes and a shallow divergence time (Zhang et al., 2011). The analysis was run for 5×105 generations using the proposal algorithm 1 of the rjMCMC (reversible-jump MCMC) species delimitation and fine-tuning parameters (Yang & Rannala, 2010) adjusted for an acceptance rate ~ 30% with a sampling frequency of 1,000 generations and a burn-in of 1/4 of the iterations (1.25×105 generations). The parameters of the locusspecific rates of evolution were fine-tuned using an auto option.. 18.

(20) Results Host plant range I obtained a total of 102 P. sonani (Green Island = 63, Orchid Island = 39) and 110 P. tobafolius (Green Island = 74, Orchid Island = 36) specimens (Appendix S1). The host plants of P. sonani from Orchid Island were mainly the sea poison tree, Barringtonia asiatica (Lecythidaceae) (93.75%), but weevils were occasionally found on beef wood, Casuarina equisetifolia (Casuarinaceae) (3.13%) and Indian almond, Terminalia catappa (Combretaceae) (3.13%). In contrast, the host plants of P. sonani from Green Island were largely Ceylon ardisia, Ardisia elliptica (Primulaceae) (50%), and beef wood (48%), and only occasionally litsea, Litsea acutivens (Lauraceae) (2%), with no records on the sea poison tree. The P. tobafolius host plants from Green Island were all mulberry, Pipturus arborescens (Urticaceae) (100%), while those from Orchid Island were largely mulberry (82.86%) and, less frequently, copperleaf, Acalypha caturus (Euphorbiaceae) (17.14%). The frequencies of the utilized host plant species were significantly different between the island populations of P. sonani (χ2 = 77.966, p < 0.001) and P. tobafolius (χ2 = 8.693, p = 0.003).. Shape and size of the body and genitalia The MANCOVA of 12 body shape measurements revealed that both P. sonani and P. tobafolius were sexually dimorphic (p < 0.001) and indicated that the island populations of both species were significantly different in their body shapes (P. sonani, p = 0.004; P. tobafolius, p = 0.013). Landmark-based morphometric analyses also indicated sexual dimorphism in both species (Appendix S7) and that the lateral body shapes of the island populations of female P. sonani were significantly different (Goodall’s F = 2.27, p = 0.004) (Fig. 6c), while their dorsal body 19.

(21) shapes were similar (Goodall’s F = 0.88, p = 0.604) (Fig. 6a). The female P. sonani from Green Island had a curvier body than those of Orchid Island (Fig. 6d), and the island populations of male P. sonani differed in both their dorsal and lateral body shapes (dorsal: Goodall’s F = 1.68, p = 0.040; lateral: Goodall’s F = 1.70, p = 0.044) (Fig. 6e & g). The male P. sonani of Green Island had a longer, flatter rostrum and a shorter abdomen than did those of Orchid Island (Fig. 6f). The island populations of P. tobafolius differed in their dorsal body shapes (Fig. 6i & m; male: Goodall’s F = 10.88, p<0.001; female: Goodall’s F = 4.66, p < 0.001). The female P. tobafolius of Green Island had a shorter rostrum and a wider, longer and flatter abdomen than those of Orchid Island (Fig. 6j), whereas the male P. tobafolius of Green Island had a shorter rostrum and head, an anteriorly extended thorax, and a wider abdomen than did those of Orchid Island. The island populations of P. sonani had significantly different aedeagus sizes (measurements: Green Island 2.52 ± 0.02 mm, Orchid Island 2.65 ± 0.02 mm, t = -2.686, p = 0.012; centroid size: Green Island 27.72 ± 1.41, Orchid Island 28.97 ± 1.06, t = -3.050, p = 0.004), whereas the island populations of P. tobafolius were similar in their aedeagus sizes (measurement: Green Island 2.24 ± 0.02 mm, Orchid Island 2.26 ± 0.01 mm, t = -0.508, p = 0.616; centroid size: Green Island 26.59 ± 1.66, Orchid Island 27.04 ± 2.12, t = -0.871, p = 0.392). The sizes of the female genitals of the island populations were similar for both P. sonani (measurement: Green Island 0.715 ± 0.003 mm; Orchid Island 0.721 ± 0.002 mm, t = -0.192, p = 0.85; centroid size: Green Island 8.67 ± 0.21, Orchid Island 9.07 ± 0.33, t = -1.329, p = 0.213) and P. tobafolius (measurement: Green Island 0.958 ± 0.008 mm, Orchid Island 0.998 ± 0.012 mm, t = -0.632, p = 0.55; centroid size: Green Island 10.01 ± 0.29, Orchid Island 10.00 ± 0.46, t = 0.001, p = 0.999). For P. sonani, the 20.

(22) aedeagus shapes of the Green Island population showed significantly more curvature and were thinner than those of Orchid Island (Goodall’s F = 1.95, p = 0.008; without the outlier No. 19, Goodall’s F = 2.07, p = 0.004) (Fig. 2e & f), whereas the island populations of P. tobafolius were similar in their aedeagus shape (Goodall’s F = 0.83, p = 0.679) (Fig. 2g). In contrast, the shapes of the female genitals of the island populations were similar for both P. sonani (Goodall’s F = 1.36, p = 0.245) and P. tobafolius (Goodall’s F = 0.53, p = 0.784) (Fig. 2h & i).. Colour pattern and scale traits The reconstructed parsimony tree from 20 discrete colour characters suggested that each of the P. sonani island populations consisted of a monophyletic lineage, except for one individual (No. 34) from Green Island (Fig. 4d). In contrast, the P. tobafolius parsimony tree resulted in no clear grouping according to island origins (Fig. 4e). All branches of these two parsimony trees had low bootstrapping values (< 50%), which was likely due to the small number of colour characters. There were no significant differences between the colour spectra of the cuticular scales of the island populations of both P. sonani and P. tobafolius (Fig. 3), but the male and female P. tobafolius had significantly different hues (male: 537.04 ± 7.96 nm, female: 546.53 ± 15.55 nm, p = 0.036) and saturation levels (male: 74.17 ± 3.47%, female: 71.09 ± 4.29%, p = 0.043). The cuticular scale density of P. sonani was higher in the Green Island population than that of Orchid Island (Green Island: 56.20 ± 9.19 / 450×337.5 µm2, Orchid Island: 47.10 ± 9.07 / 450×337.5 µm2, p = 0.003), and the type III scales were only found in the P. sonani females from both islands. The frequencies of the scale types in the male P. sonani were significantly different between the island populations with the Green Island population only having type II scales (Green Island: type II 21.

(23) 100%, n = 11; Orchid Island: type I 36.36%, n = 4, Type II 63.64%, n = 7; χ2 = 4.889, p = 0.027). The frequencies of the scale types were similar between the island populations of female P. sonani (Green Island: Type I 11.11%, n = 1; Type II 66.67%, n = 6; Type III 22.22%, n = 2; Orchid Island: Type I 40%, n = 4; Type II 50%, n = 5; Type III 10%, n = 1; χ2 = 2.178, p = 0.337). The frequencies of the scale types were significantly different between the island populations of female P. tobafolius (Green Island: Type I 60 %, n = 3; Type II 40%, n = 2; Orchid Island: Type II 100%, n = 5; χ2 = 4.286, p = 0.038) but were not significantly different between the male P. tobafolius populations (Green Island: Type I 40%, n =2; Type II 60%, n = 3; Orchid Island: Type II 100%, n = 6; χ2=2.933, p=0.087).. Population genetic structure and molecular phylogeny RAD sequencing resulted in 2004 and 303 SNPs for P. sonani and P. tobafolius, respectively. Bayesian clustering analyses of the SNPs demonstrated that the P. sonani individuals were best grouped into two genetic clusters of the same island of origin (Fig. 5a) (K = 2, lnL = – 8541.94 ± 1.44; delta K=2 had the highest peak, Appendix S8a & c), whereas the members of P. tobafolius could not be unambiguously assigned into any number of genetic clusters (models with K = 1–10 are equally probable, Appendix S8b & d), suggesting that they belong to the same gene pool. The four-gene data set contained a total of 4050 bps (cox1 675, nd2 983, EF1α 810 and ITS 1582 bps). The cox1, nd2 and ITS gene trees indicated that each of the P. sonani island populations form a monophyletic lineage (Appendix S9a, b & d), whereas the EF1α gene tree showed a paraphyletic P. sonani from Orchid Island (Appendix S9c). The combined phylogeny of the four genes established strong support for a monophyletic P. sonani from Green Island (1/100/100, BPP/PB/LB), 22.

(24) whereas the P. sonani from Orchid Island was paraphyletic with respect to that from Green Island (Fig. 5b) (TreeBASE Accession no. #). In contrast, the P. tobafolius cox1 phylogeny (677 bps) lacked resolution and revealed no phylogenetic clustering of the island lineages (Appendix S10).. Integrated species delimitation The Bayesian species delimitation of the integrated data set suggested the presence of two species of P. sonani from Green Island and Orchid Island, which was supported by a divergent node with a posterior probability of 1.0 (ESS of all the parameters in the MCMC runs > 11,800) under various prior population demography and species divergence time settings (large versus small ancestral population sizes, θ, and deep versus shallow divergence times, τ) (Appendices S11 & 12).. Discussion The majority of morphological characters reveal overlapping but statistically significant difference between the island populations of P. sonani (Table 1), while the two molecular data sets indicates two separate island lineages. Integrated species delimitation of combining morphological, molecular and ecological characters further strengthen the support for recognizing two distinct evolutionary lineages of P. sonani from Green Island and Orchid Island. In addition, diagnostic morphological (the posterior-lateral colour stripe of the elytra) and molecular (such as fixed, unique mitochondrial haplotypes) characters are identified and useful for distinguishing the island populations of P. sonani in both field and laboratory settings. The identification of cryptic species relies on multiple characters (Rato et al., 2016; Barley et al., 2013), so integrated species delimitation based on multiple character types can 23.

(25) provide a more quantitative and objective assessment of species diversity. For example, the species number of red-bellied snake species (genus Storeria) was reduced from eight to four when five morphological traits were added to a molecular data set of genome-wide loci (Pyron et al., 2016), and a recent example of integrated species delimitation in Hercules beetles (genus Dynastes) by combining molecular and morphological traits suggested ten species instead of one (Huang & Knowles, 2015). In contrast to P. sonani, most character sets indicate that the island populations of P. tobafolius show no detectable differentiation, which suggests that they belong to one species (Table 1) and is in agreement with earlier taxonomic assessments (Kano, 1929; Starr & Wang, 1992). The qualitative and quantitative evidences from our study provide a basis for a revision of the current taxonomy of Pachyrhynchus species in Taiwan. The endangered P. sonani is composed of two cryptic species, and the one from Green Island is described here as a new species, P. jitanasaius sp. nov. Chen & Lin (taxonomic section below). In contrast, the two island populations of P. tobafolius should continue to be treated as one species. Based on our findings, I urgently recommend that the Wildlife Conservation Act of Taiwan revise the species status of P. sonani and protect these two independently evolved island lineages, which have restricted distribution, limited gene flow and host preference, which make each more prone to extinction due to habitat loss. Our results demonstrated that the molecular and phenotypical traits of Pachyrhynchus weevils differ in their level of divergence. Between island populations of P. sonani, molecular characters including nucleotide sequences and genome-wide SNPs showed a fixed and higher level of divergence, respectively, than did the morphological characters, except for one discrete colour character that revealed non-overlapping differentiation. The biological processes underlying this pattern are not 24.

(26) clear but might include the level of natural selection, the developmental and phylogenetic constraints on the traits, the extent of hybridization, and the direction and strength of gene flow between populations. An intriguing finding of this study was that, between island populations of P. sonani, almost all of the male morphological traits diverged more than those of the females, but no evidence has yet been found to suggest sexual selection for male traits in the divergence of Pachyrhynchus weevil species. All of the male traits consistently distinguished the island populations of P. sonani as two separate species, while the female traits showed no significant divergence between the island populations (Table 1). Earlier taxonomic studies often regarded colour traits as highly polymorphic and unstable, thus making them unacceptable as diagnostic characters (e.g., P. orbifer complex, Schultze, 1923). However, our results suggest that the discrete color traits of P. sonani are as useful as the traditionally applied male and female genital traits for species delimitation (Schultze, 1923; Yoshitake, 2012), so they can thus provide another source of informative characters for taxonomic and phylogenetic studies of Pachyrhynchus weevils. Although our study did not find fixed, diagnostic scale types between the island populations of P. sonani, the less frequently used fine structure of the cuticular scales, as observed through SEM (e.g., Erbey & Candan 2014), can provide useful characters in the systematics of weevils, which as a group contains more species than all of the other major insect taxa (Oberprieler et. al., 2007). Other informative, even finer resolution characters for discriminating Pachyrhynchus species may come from the use of three-dimensional photonic poly-crystals (Welch et al., 2007; McNamara et al., 2013). In sexually reproducing organisms, species are usually considered to be the central units of taxonomic characterization of global biodiversity (Wheeler et al., 2012). The extent of morphological divergence between 25.

(27) island populations of P. sonani indicates that they are consistently distinguishable morphological or phenetic species (Sokal, 1973), and the phylogenies and genetic clustering of P. sonani also establish that its island populations consist of two valid evolutionary species with independent historical tendencies (Wiley, 1978). Nevertheless, based on our best phylogenetic estimation, the P. sonani lineages from Orchid Island are paraphyletic with respect to the lineages from Green Island (Fig. 5b). According to the exclusive monophyletic group of common ancestry recognized by the phylogenetic species concept (de Queiroz & Donoghue, 1988), the paraphyletic P. sonani lineages from Orchid Island should not be regarded as a separate phylogenetic species. However, the branch supports for the paraphyly of the Orchid Island lineage are low (< 50%) and the tree branches leading to the Orchid Island lineages are short, indicating that the paraphyletic relationships of the Orchid Island P. sonani are the inevitable result of an incomplete ancestral lineage sorting process of recently diverged species (Avise, 2000). The P. sonani phylogeny has a relatively long branch separating the Green Island and Orchid Island lineages, suggesting that a substantial amount of mutation has accumulated over a certain period of time since their divergence, implying the extensive temporal separation of the two island species. Nevertheless, to fully understand “the reality of species” (Coyne & Orr 2004) and test the speciation process and biological species status (Mayr, 1995) of P. sonani, future experimental studies are needed to examine the level of reproductive isolation barriers between the island populations. Oceanic islands are powerhouses for generating endemic diversity largely due to the effects of geographical isolation, small population sizes and local adaptation. Situated on the northern boundary of the Kano’s Line (“New-Wallace Line”, Kano, 1941), Green Island and Orchid Island of the Taiwan-Luzon Archipelago are well known for their remarkable 26.

(28) endemism (e.g., Kano, 1929; Yoshida et al., 2000; Yen et al., 2003; Hsu & Huang 2008; Lee & Staines, 2010) and faunal and floral affinities with the Philippines (Kano 1933, 1941; Li, 1953; Chang, 1981), although the two islands are geographically closer to Taiwan. However, to our knowledge, no previous study has attempted to closely examine the species boundary of the endemic organisms occurring on these two neighboring islands using multiple character types. Our study of cryptic species in P. sonani represents the first empirical evidence indicating that the inter-island, allopatric speciation of the endemic organisms inhabiting Green and Orchid Islands may be more common than previously thought, but it has been concealed by morphologically or ecologically similar species. Therefore, our study emphasizes that the cryptic diversity of the endemic taxa of these two small oceanic islands may still be largely underestimated. For example, morphologically similar populations of a species of scarab beetle, Anomala expansa (Rutelinae), has been found on Green Island and Orchid Island. There are currently three recognized subspecies of A. expansa, including A. e. expansa Bates, 1866 of China and Taiwan; A. e. lanshuensis Nomura, 1977 of Orchid Island; and A. e. lutaoensis Nomura, 1977 of Green Island (Yu et. al., 1988). The two island subspecies, A. e. lutaoensis and A. e. lanshuensis subtly differ in their colour pattern and elytral shape (Yu et. al., 1988), which, similar to the case of P. sonani, raises the question of whether these two subspecies consist of two distinct cryptic species and the degree of ecological divergence between them. The number of endemic species and the level of endemism in these small oceanic islands require further examination.. Taxonomy Pachyrhynchus jitanasaius sp. nov. Chen & Lin (Fig. 7) 27.

(29) Diagnosis. Pachyrhynchus jitanasaius sp. nov. is morphologically very similar to Pachrhynchus sonani Kôno 1930. Pachrhynchus sonani was originally described from specimens collected on Orchid Island (Kotosho) using the following combination of colour characteristics (Kôno, 1930): head with a lateral scale band extending from the rostrum to behind the eyes, prothorax with a T-shaped scale band, elytra with two scale bands extending from the anterior to the rear end. Pachyrhynchus jitanasaius sp. nov. can be identified by the absence of the posteriorlateral elytrous stripes or with small spherical spots but rarely with two stripes (Fig. 7a & c, L3 and R3), whereas P. sonani usually exhibits one or two elongate stripes at the same position. In addition, the aedeagus of P. jitanasaius sp. nov. has greater curvature and is thinner than that of P. sonani (Fig. 7e & f). The holotype of P. jitanasaius sp. nov. are preserved in National Museum of Natural Science (NMNS) in Taichung, Taiwan.. Description. Measurement (mean ± SE) (mm): Males (n = 12), RW 1.36 ± 0.12, HW 2.47 ± 0.16, TW 3.84 ± 0.25, AW 5.34 ± 0.30, RL 1.78 ± 0.12, HL 3.16 ± 0.21, TL 3.67 ± 0.26, AL 7.81 ± 0.46, RH 1.38 ± 0.13, HH 2.47 ± 0.15, TH 3.21 ± 0.25, AH 4.53 ± 0.12, AEL (n = 16) 2.54 ± 0.01; Females (n = 9), RW 1.39 ± 0.08, HW 2.46 ± 0.09, TW 3.76 ± 0.20, AW 5.87 ± 0.29, RL 1.75 ± 0.11, HL 3.20 ± 0.26, TL 3.45 ± 0.16, AL 8.17 ± 0.55, RH 1.40 ± 0.14, HH 2.46 ± 0.16, TH 3.16 ± 0.24, AH 4.83 ± 0.30, SW (n=7), 0.715 ± 0.003. Integument black and smooth, antennae black. Body dark black with green, shiny blue and green stripes and spots. Rostrum with short pubescence and sporadic scales near mouthparts. Head with sparsely short pubescence behind eyes. Lateral head with a scale band extending from rostrum to behind eyes. Triangle-like scale spot between fronts of 28.

(30) eyes and vertex. Dorsal prothorax with a T-shaped scale band at centre (Fig. 7a). Scales cover entire prosternum (Fig. 7b). Elytral scale stripes consist of a horizontal scale band at the centre, and two circular scale bands at anterior and posterior halves of the elytra, with connection to the horizontal scale band at lateral ends, but not at the centre (Fig. 7a & c). Area of L1, R1, L3 and R3 spotless (Fig. 7c & d). Abdominal ventrite I with two horizontal scale bands not connected to each other (Fig. 7b). Ventrite II with a connected horizontal scale band. Ventrites III, IV and V without scale bands. Legs black, coxa, trochanter and tibia black without scales but femur with two broken scale bands. Femur and tibia of approximately equal length. Tibia with short strong hairs. Tarsus black and covered with pubescence on segments I, II, III, IV and pad. Head bends down with short rostrum and weakly bulging eyes. Antennae with stout and slightly longer scape. Pedicel and flagella I, II, III, IV and V of similar size with short, thin pubescence. Flagellum VI gradually larger with short sporadic pubescence. Flagellum VII club and covered with short pubescence. Rostrum longer than wide (male: RW/RL 0.76, female: 0.79), minutely punctured, weakly bulging anteriorly and gradually declined to apex. Subspherical prothorax minutely punctured, about as long as wide (male: TW/TL 1.05, female 1.08). Dorsal prothorax shallowly convex. Elytra subobovate and convex with 7 weak punctural striae. Length of elytra twice as long as the prothorax (male: AL/TL 2.13, female: 2.37). Width of elytra 1.5 times longer than prothorax (male: AW/TW 1.39, female 1.56). Abdomen broadest at 1/3 of ventrites. Male (Fig. 7e–h & l) and female genitalia (Fig. 7i–k) as illustrated. Aedeagal body (Fig. 7e & f) curved near base; sides subparallel to apical 1/4 and tapering toward apex with aedeagal apodemes slender and slightly longer than aedeagal body. Spiculum gastrale (Fig. 7g) longer than aedeagal body and curved rightward. Tegmen (Fig. 7h) short and 29.

(31) stout. Endophallus (Fig. 7l) with trifid near subbase 1/3. Middle leaf (E1) sharp, two globular leaves (E2) connected with each other, obreniform sclerite (E3) midway, a sclerite along apical 1/3 inflate (E4) with folds and connected with flagellum. In female genitalia, sternite VIII (Fig. 7i) spade-shaped with short hairs at apical margin and a slender stem at base. Stem 3 times longer than main part of sternite VIII. Apex of ovipositor (Fig. 7j) slender and gradually sharp in apical 1/2 with 2 small sclerites and short hairs. Spermatheca (Fig. 7k) saccular without sclerite.. Key to six Pachyrhynchus species from Green and Orchid Island 1a. Prothorax has four dorsal colour spots………..….………P. tobafolius 1b. Prothorax has three dorsal spot or stripe colour spots……………..…2 2a. Prothorax has three dorsal colour spots …….....…...…………...……3 2b. Prothorax has dorsal colour stripe…...……….……..……..…............4 3a. Head vertex has a dorsal colour spot.…...….....…. P. sarcitis kotoensis 3b. Head vertex has dorsal colour spot absent……..…….....….P. insularis 4a. Prothorax has dorsal crossing colour stripe at the middle of thorax absent……………………….………………………………... P. yamianus 4b. Prothorax has dorsal T-shaped colour stripe at centre……………..…5 5a. One or two extensions from the colour stripe at L1 and R1 of elytra (Fig. 1d, e & 7d)………...…………………………….............P. sonani 5b. No extension of the colour stripe at L1 and R1 of the elytra (Fig. 1e, 7c & d)……………………………………………………………...…6 6a. Linear colour stripe connected to posterior colour stripe at L3 or R3 of 30.

(32) the elytra (Fig. 1d, e & 7d)…..……………………………..…P. sonani 6b. No or round colour spot at L3 and R3 of the elytra (Fig. 1e, 7c & d)…………………………………………………. P. jitanasaius sp. nov. Type material. Holotype male (Specimen code, Ps365) “[Taiwan] / Taitung County, Green Island / VII. 2014, collector” (typed on a white card), “YEN-TING CHEN/ COLLECTION” (typed on a white card) “[Holotype] Male / Pachyrhynchus jitanasaius / Chen, 2016 / Det. Chen et al., 2016. (typed on a red card). Paratypes (3 exs.) (Specimen code, Ps331, Ps315 & Ps285) “[Taiwan] / Taitung County, Green Island / VII. 2014, collector” (typed on a white card), “YEN-TING CHEN / COLLECTION” (typed on a white card) “[Paratypes] / Pachyrhynchus jitanasaius / Chen, 2016 / Det. Chen et al., 2016. (typed on a specimen label).. Distribution. TAIWAN: Green Island (Green Island Township, Taitung County).. Etymology. The new species is named after the type location, do Jintanasai, which is the name of Green Island in the aboriginal language of the Tao (Yami) people.. 31.

(33) References Anderson, T.W. (1984) An Introduction to Multivariate Statistical Analysis, 2nd edition. John Wiley and Sons, Inc, Hoboken, USA. Andolfatto, L., Lavernhe, S. & Mayer, J.R.R. (2011) Evaluation of servo, geometric and dynamic error sources on five-axis high-speed machine tool. International Journal of Machine Tools and Manufacture, 51, 787–796. Avise, J.C. (2000) Phylogeography: the history and formation of species. Harvard University Press, Cambridge, USA. Bailey, R.C. & Byrnes, J. (1990) A new, old method for assessing measurement error in both univariate and multivariate morphometric studies. Systematic Zoology, 39, 124–130. Barley, A.J., White, J., Diesmos, A.C. & Brown, R.M. (2013) The challenge of species delimitation at the extremes: diversification without morphological change in Philippine sun skinks. Evolution, 67, 3556–3572. Bickford, D., Lohman, D.J., Sodhi, N.S., Ng, P.K., Meier, R., Winker, K., Ingram, K.K. & Das, I. (2007) Cryptic species as a window on diversity and conservation. Trends in Ecology and Evolution, 22, 148– 155. Catchen, J., Hohenlohe, P.A., Bassham, S., Amores, A. & Cresko, W.A. (2013) Stacks: an analysis tool set for population genomics. Molecular Ecology, 22, 3124–3140. Chang, C.E. (1986) The phytogeographical position of Botel Tobago based on the woody plants. Journal of Phytogeography and Taxonomy, 34, 1–15. Chao, J.T., Yang, M.M. & Wu, M.H. (2009) Setting assessing criteria for the conservation of Taiwan’s threatened insects. Conservation Research Series No. 97–04. Council of agriculture, Executive Yuan, 32.

(34) Taiwan. (in Chinese) Coyne, J.A. & Orr, H.A. (2004) Speciation. Sinauer Associates Inc, Massachusetts, USA. Darriba, D., Taboada, G.L., Doallo, R. & Posada, D. (2012) jModelTest 2: more models, new heuristics and parallel computing. Nature Methods, 9, 772–772. de Queiroz, K. & Donoghue, M.J. (1988) Phylogenetic systematics and the species problem. Cladistics, 4, 317–338. Earl, D.A. & vonHoldt, B.M. (2012) STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources, 4, 359–361. Edwards, D.L. & Knowles, L.L. (2014) Species detection and individual assignment in species delimitation: can integrative data increase efficacy? Proceedings of the Royal Society B: Biological Sciences, 281, 20132765. Erbey, M. & Candan, S. (2014) The ultrastructural analysis of scales in Brachypera Capiomont, 1868 and Hypera Germar, 1817 (Coleoptera: Curculionidae: Hyperinae). Entomological News, 124, 103–108. Falush, D., Stephens, M. & Pritchard, J.K. (2003) Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics, 164, 1567–1587. Felsenstein, J. (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution, 39, 783–791. Folmer, O., Black, M., Hoeh, W., Lutz, R. & Vrijenkoek, R. (1994) DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology, 3, 294–299. Gillespie, R.G. & Roderick, G.K. (2002) Arthropods on islands: 33.

(35) colonization, speciation, and conservation. Annual Review of Entomology, 47, 595–632. Grant, P.R. (1998) Evolution on Islands. Oxford University Press, New York, USA. Grant, P.R. & Grant, B.R. (2014) 40 Years of Evolution: Darwin's Finches on Daphne Major Island. Princeton University Press, Princeton, USA. Heaney, L.R. (2000) Dynamic disequilibrium: a long-term, large-scale perspective on the equilibrium model of island biogeography. Global Ecology and Biogeography, 9, 59–74. Hernández-Vera, G., Caldara, R., Toševski, I. & Emerson, B.C. (2013) Molecular phylogenetic analysis of archival tissue reveals the origin of a disjunct southern African–Palaearctic weevil radiation. Journal of Biogeography, 40, 1348–1359. Hsu, Y.F. & Huang, H.C. (2008) On the discovery of Hasora mixta limata ssp. nov. (Lepidoptera: Hesperiidae: Coeliadinae) from Lanyu, Taiwan, with observations of its unusual immature biology. Zoological Studies, 47, 222–231. Huang, J.P. & Knowles, L.L. (2016) The species versus subspecies conundrum: quantitative delimitation from integrating multiple data types within a single Bayesian approach in Hercules beetles. Systematic Biology, 65, 685–699. Kano, T. (1929) Descriptions of three new species of Curculionidae of genus Pachyrhynchus Germar from the island of Botel-Tobago. The Entomological Society of Japan, 3, 237–238. Kano, T. (1933) Zoogeography of Botal Tobago Island (Kotosho) with a consideration of the northern portion of Wallace, s line. Bulletin of the Biogeographical Society of Japan, 9, 381–399. Kano, T. (1941) Biogeography of the island of Kôtôsho (Botel Tobago) 34.

(36) with special reference to the New-Wallace Line. The Institute of the Pacific, Section of Sciences, ed. Greater South Seas: its culture and its soil. Tokyo: Kawade Shobou, 219–323. Kôno, H. (1930) Kurzrüssler aus dem japanischen Reich. Journal of the Faculty of Agriculture, 24, 153–242. Lee, C.F. & Staines, C.L. (2010) Monolepta ongi, a new species from Lanyu Island, with redescription of its allied species Monolepta longitarsoides Chûjô, 1938 (Coleoptera: Chrysomelidae: Galerucinae). Proceedings of the Entomological Society of Washington, 112, 530–540. Li, H.L. (1953) Floristic interchanges between Formosa and the Philippines. Pacific Science, 7, 179–186. Loew, E.R., Fleishman, L.J., Foster, R.G. & Provencio, I. (2002) Visual pigments and oil droplets in diurnal lizards: a comparative study of Caribbean anoles. Journal of Experimental Biology, 205, 927–938. Losos, J. (2009) Lizards in an Evolutionary Tree, Vol. 10. Ecology and Adaptive Radiation of Anoles. University of California Press, Oakland, USA. MacArthur, R.H. & Wilson, E.O. (1967) The Theory of Island Biogeography. Princeton University Press, Princeton, USA. McNamara, M.E., Briggs, D.E., Orr, P.J., Gupta, N.S., Locatelli, E.R., Qiu, L., Yang, H., Wang, Z., Noh, H. & Cao, H. (2013) The fossil record of insect color illuminated by maturation experiments. Geology, 41, 487–490. Mayr, E. (1995) Species, classification, and evolution. Biodiversity and Evolution. National Science Museum Foundation, Tokyo, 3–122. Nixon, K.C. (1999) The parsimony ratchet, a new method for rapid parsimony analysis. Cladistics, 15, 407–414. Norusis, M.J. (2005) SPSS for Windows, v.12.0. SPSS, Inc. 35.

(37) Oberprieler, R.G., Marvaldi, A.E. & Anderson, R.S. (2007) Weevils, weevils, weevils everywhere. Zootaxa, 1668, 491–520. Pellens, R. & Grandcolas, P. (eds.) (2016) Biodiversity Conservation and Phylogenetic Systematics: Preserving Our Evolutionary Heritage in an Extinction Crisis. Springer International Publishing, Biodiversity Conservation and Phylogenetic Systematics, 1–15. Pyron, R.A., Hsieh, F.W., Lemmon, A.R., Lemmon, E.M. & Hendry, C.R. (2016) Integrating phylogenomic and morphological data to assess candidate species‐delimitation models in brown and red‐bellied snakes (Storeria). Zoological Journal of the Linnean Society, 177, 937–949. Rato, C., Harris, D.J., Carranza, S., Machado, L. & Perera, A. (2016) The taxonomy of the Tarentola mauritanica species complex (Gekkota: Phyllodactylidae): Bayesian species delimitation supports six candidate species. Molecular Phylogenetics and Evolution, 94, 271– 278. Rohlf, F.J. & Slice, D. (1990) Extensions of the procrustes method for the optimal superimposition of landmarks. Systematic Zoology, 39, 40– 59. Rohlf, F.J. (2005) tpsDig, digitize landmarks and outlines, version 2.05. Department of Ecology and Evolution and Department of Anthropology, Stony Brook University, Stony Brook, USA. Ronquist, F., Teslenko, M., van der Mark, P., Ayres, D.L., Darling, A., Hohna, S., Larget, B., Liu, L., Suchard, M.A. & Huelsenbeck, J.P. (2012) MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Systematic Biology, 61, 539–542. Schultze, W. (1923). A monograph of the pachyrrhynchid group of the Brachyderinae, Curculionidae: Part I. The genus Pachyrhynchus 36.

(38) Germar. The Philippine Journal of Science, 23, 609–673. Sheets, D.H. (2004) IMP, the Integrated Morphometric Package. Available online: http://www.canisius.edu/~sheets/morphsoft.ht. Sikes, D.S. & Lewis, P.O. (2001) Software manual for PAUPRat: a tool to implement parsimony ratchet searches using PAUP*. Distributed by the authors. Sokal, R.R. (1973) The species problem reconsidered. Systematic Biology, 22, 360–374. Solís-Lemus, C., Knowles, L.L.& Ané, C. (2015) Bayesian species delimitation combining multiple genes and traits in a unified framework. Evolution, 69, 492–507. Stamatakis, A. (2006) RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics, 22, 2688–2690. Starr, C.K. & Wang, H.Y. (1992) Pachyrhynchine weevils (Coleoptera: Curculionidae) of the islands fringing Taiwan. Journal of Taiwan Museum, 45, 5–14. Swindell, S.R. & Plasterer, T.N. (1997) SEQMAN. Sequence Data Analysis Guidebook. Springer, New York, 75–89. Swofford, D.L. (2002) PAUP*: Phylogenetic analysis using parsimony (and Other Methods), v. 4.0b10. Sinauer Associates. Tamura, K., Stecher, G., Peterson, D., Filipski, A. & Kumar, S. (2013) MEGA6: molecular evolutionary genetics analysis version 6.0. Molecular Biology and Evolution, 30, 2725–2729. Tseng, H.Y., Lin, C.P., Hsu J.Y., Pike D.A. & Huang, W.S. (2014) The functional significance of aposematic signals: geographic variation in the responses of widespread lizard predators to colourful invertebrate Prey. Plos One, 9, e91777. Wallace, A.R. (1895) Natural Selection and Tropical Nature: Essays on 37.

(39) Descriptive and Theoretical Biology. Macmillan, London, UK. Weekers, P.H.H., De Jonckheere, J.F. & Dumont, H.J. (2001) Phylogenetic relationships inferred from ribosomal ITS sequences and biogeographic patterns in representatives of the genus Calopteryx (Insecta: Odonata) of the West Mediterranean and adjacent West European zone. Molecular Phylogenetics and Evolution, 20, 89–99. Wagner, W.L. & Funk, V.A. (1995) Hawaiian Biogeography. Smithsonian Institution Press, Washington, USA. Welch, V., Lousse, V., Deparis, O., Parker, A. & Vigneron, J.P. (2007) Orange reflection from a three-dimensional photonic crystal in the scales of the weevil Pachyrhynchus congestus pavonius (Curculionidae). Physical Review E, 75, 041919. Wheeler, Q.D., et al. (2012) Mapping the biosphere: exploring species to understand the origin, organization and sustainability of biodiversity. Systematics and Biodiversity, 10, 1–20. Whittaker, R.J., Triantis, K.A. & Ladle, R.J. (2008) A general dynamic theory of oceanic island biogeography. Journal of Biogeography, 35, 977–994. Wiens, J.J. (2007) Species delimitation: new approaches for discovering diversity. Systematic Biology, 56, 875–878. Wiley, E.O. (1978) The evolutionary species concept reconsidered. Systematic Biology, 27, 17–26. Yang, Z. & Rannala, B. (2010) Bayesian species delimitation using multilocus sequence data. Proceedings of the National Academy of Sciences, 107, 9264–9269. Yen, S.H., Kitching, I.J. & Tzen, C.S. (2003) A new subspecies of hawkmoth from Lanyu, Taiwan, with a revised and annotated checklist of the Taiwanese Sphingidae (Lepidoptera). Zoological Studies, 42, 292–306. 38.

(40) Yewers, M.S., McLean, C.A., Moussalli, A., Stuart-Fox, D., Bennett, A.T.D. & Knott, B. (2015) Spectral sensitivity of cone photoreceptors and opsin expression in two colour-divergent lineages of the lizard Ctenophorus decresii. Journal of Experimental Biology, 218, 1556– 1563. Yoshida, H., Tso, I.M. & Severinghaus, L. (2000) The spider family Theridiidae (Arachnida: Araneae) from Orchid Island, Taiwan: descriptions of six new and one newly recorded species. Zoological Studies, 39, 123–132. Yoshitake, H. (2012) Nine new species of genus Pachrhynchus Germar (Coleoptera: Curculionidae) from the Philippines. Esakia, 52, 17–34. Yu, C.J., Chu, Y.I. & Kobayashi, H. (1998) The Scarabaeidae of Taiwan. Mu Sheng Press, Taipei, Taiwan. Zhang, C., Zhang, D. X., Zhu, T. & Yang, Z. (2011) Evaluation of a Bayesian coalescent method of species delimitation. Systematic Biology, 60, 747–761.. 39.

(41) Table 1. Summary of Pachyrhynchus sonani and Pachyrhynchus tobafolius species delimitation based on multiple characters. Asterisks indicate the presence (two species) and absence (one species) of a statistically significant difference between the Green Island and Orchid Island populations. P. sonani. P. tobafolius. Characters 1 species SNPs. ✽. ✽. DNA sequences. 4 genes. ✽. ✽a. Measurement Body shape. GMb (dorsal). ✽♀. GM (lateral) Discrete characters Coloration. Genitalia. Ecological Integrated. 1 species. RAD sequencing Molecular. Morphological. 2 species. Host plant. ✽♀, ♂. ✽♀, ♂. ✽♂. ✽♀, ♂. ✽♀, ♂. ✽♀, ♂. ✽. ✽. Colour spectrum. ✽♀, ♂. Number of scales. ✽♀. ✽♂. ✽♀, ♂. Type of scales. ✽♀. ✽♂. ✽♂. Measurement. ✽♀. ✽♂. ✽♀, ♂. Centroid size. ✽♀. ✽♂. ✽♀, ♂. GM. ✽♀. ✽♂. ✽♀, ♂. Host plant range. 2 species. ✽♀, ♂. ✽ ✽. ✽♀. ✽ NA. NA 40.

(42) a. cox1 gene only. b. GM, geometric morphometrics.. 41.

(43) Figure 1. (a) Map of Green Island and Orchid Island inhabited by (b) Pachyrhynchus sonani and Pachyrhynchus tobafolius. (c) Sampling sites (numbers refer to Appendix S1). (d) Holotype of Pachyrhynchus sonani (Kôno, 1930) (Pachyrhynchus moniliferus sonani is a synonym). Images obtained with permission from the Center of Digital Collections Research and Development, National Taiwan University. (e) Geographic variation in the colouration of Pachyrhynchus sonani.. Figure 2. Location of (a) 11 dorsal and (b) 10 lateral landmarks used to define the body shapes of Pachyrhynchus weevils. Circle d is the location for measuring the colour spectrum of the scales, and rectangles e and f are the locations for measuring the number and types of scales on Pachyrhynchus sonani and Pachyrhynchus tobafolius, respectively. The variation in genital shape between island populations of Pachyrhynchus sonani and Pachyrhynchus tobafolius: (c) location of 2 landmarks and 10 semi-landmarks for aedeagus (lateral view). The size of the aedeagus was measured between landmarks 1 and 12. (d) Location of 5 female sternite VIII landmarks (ventral view), the size of which was measured between landmarks 2 and 3. (e) Aedeagus shape deformation grid from the Orchid to Green Island populations of Pachyrhynchus sonani (4×). Morphometric analyses of the aedeagus (f & g) and sternite VIII shapes (h & i) of Pachyrhynchus sonani and Pachyrhynchus tobafolius.. Figure 3. Colour spectra of the cuticular scales of the island populations of Pachyrhynchus sonani (a & b) and Pachyrhynchus tobafolius (c & d).. Figure 4. Fine morphology of the cuticular scales of Pachyrhynchus sonani. (a) Type I, white arrow indicates carving area of < 2/3 of scale surface. (b) Type II, white arrow indicates carving area of > 2/3 of scale 42.

(44) surface. (c) Type III, white rectangle indicates a major depression in the shape of a half-moon. Maximum parsimony trees of (d) Pachyrhynchus sonani and (e) Pachyrhynchus tobafolius based on 20 and 4 discrete colour traits (Appendix S5), respectively.. Figure 5. (a) STRUCTURE genetic clustering analyses for Pachyrhynchus sonani. The optimal genetic clustering model was K = 2. (b) Phylogeny of Pachyrhynchus sonani reconstructed from Bayesian phylogenetic analyses based on four genes (cox1, nd2, EF1α and ITS). Numbers near the nodes are the branch support values of the Bayesian posterior probability (BPP) / parsimony bootstrap (PB) / likelihood bootstrap (LB). Nodes without numbers have support values of < 50%.. Figure 6. Landmark-based morphometric analyses of the body shapes of the island populations of Pachyrhynchus sonani (a–h) and Pachyrhynchus tobafolius (i–p). Females of Pachyrhynchus sonani: (a) dorsal and (c) lateral view. Males of Pachyrhynchus sonani: (e) dorsal and (g) lateral view. Females of Pachyrhynchus tobafolius: (i) dorsal and (k) lateral view. Males of Pachyrhynchus tobafolius: (m) dorsal and (o) lateral view. Body shape deformation grid from the Orchid to Green Island populations (5×): females of Pachyrhynchus sonani: (b) dorsal and (d) lateral view; males of Pachyrhynchus sonani: (f) dorsal and (h) lateral view; females of Pachyrhynchus tobafolius: (i) dorsal and (k) lateral view; males of Pachyrhynchus tobafolius: (n) dorsal and (p) lateral view.. Figure 7. Holotype of Pachyrhynchus jitanasaius sp. nov.: (a) dorsal view, (b) ventral view, (c) lateral view and (d) area of the dorsal elytra. Male genitalia: aedeagus in (e) dorsal view and (f) lateral view, (g) sternite IX in dorsal view, (h) tegmen in dorsal view and (l) endophallus 43.

(45) in lateral view (E1 middle leaf, E2 globular leaf, E3 obreniform sclerite, E4 apical sclerite); Female genitalia: (i) sternite VIII in ventral view, (j) ovipositor apex in dorsal view and (k) spermatheca.. 44.

(46) Fig. 1. 45.

(47) Fig. 2. 46.

(48) Fig. 3. 47.

(49) Fig. 4. 48.

(50) Fig. 5. 49.

(51) Fig. 6. 50.

(52) Fig. 7. 51.

(53) Supporting Information Appendix S1. Information of the specimens used in this study. Appendix S2. GenBank accession numbers of the samples used in this study. Appendix S3. The character coding of colour pattern in Pachyrhynchus sonani and outgroups. Appendix S4. The character coding of colour pattern in Pachyrhynchus tobafolius and outgroups. Appendix S5. Description of character states of colour traits in Pachyrhynchus sonani. Appendix S6. Description of character states of colour traits in Pachyrhynchus tobafolius. Appendix S7. Morphometric analyses of body shape between sexes based on 21 landmarks for (a–h) Pachyrhynchus sonani and (i–p) Pachyrhynchus tobafolius. Appendix S8. Structure analyses of Pachyrhynchus sonani and Pachyrhynchus tobafolius. Appendix S9. The gene trees of Pachyrhynchus sonani. Appendix S10. The cox1 gene tree of Pachyrhynchus tobafolius. Appendix S11. The posterior probability of parameters under four gamma prior distributions in iBPP analyses. Appendix S12. The heritability of traits under four gamma prior distributions in iBPP analyses. Appendix S13. The phylogeny of Pachyrhynchus sonani reconstructed from (a) Bayesian, (b) parsimony and (c) likelihood phylogenetic analyses based on combined four genes (cox1, nd2, EF1α and ITS).. 52.

(54) Appendix S1. Information of the specimens used in this study (GI: Green Island; OI: Orchid Island). Taxon code. Date. Island. Sitea. Species. Sex. Ps45 Ps71 Ps72 Ps73. G02 G01 G10. 2006/4/7 2011/11/4 2011/11/4 2011/11/4. OI GI GI GI. 13 5-8 5-8 5-8. P. sonani P. sonani P. sonani P. sonani. F M M F. 5 6 7 8 9 10 11. Ps74 Ps75 Ps76 Ps77 Ps78 Ps79 Ps80. G11 G12 G13 G21 G22 G23 G03. 2011/11/4 2011/11/4 2011/11/4 2011/11/4 2011/11/4 2011/11/4 2011/7/9. GI GI GI GI GI GI GI. 5-8 5-8 5-8 5-8 5-8 5-8 5-8. P. sonani P. sonani P. sonani P. sonani P. sonani P. sonani P. sonani. M M M M M M F. 12 13 14 15 16 17 18. Ps81 Ps82 Ps83 Ps84 Ps85 Ps86 Ps87. G05 G06 G07 G08 G09 O10. 2011/7/9 2011/7/11 2011/7/11 2011/7/11 2011/11/4 2011/11/4 2011/8/4. GI GI GI GI GI GI OI. 5-8 5-8 5-8 5-8 5-8 5-8 13. P. sonani P. sonani P. sonani P. sonani P. sonani P. sonani P. sonani. F F M F F M M. 1 1 1 1. 19 20 21 22. Ps88 Ps89 Ps90 Ps91. 2011/8/4 2011/8/3 2011/8/3 2011/8/3. OI OI OI OI. 13 15-17 15-17 15-17. P. sonani P. sonani P. sonani P. sonani. M F M F. 2 2 2 2. No.. Specimen No.. 1 2 3 4. O12 O14. Host plantb. 2. 53.