Reassessment of the Species in the Euwallacea Fornicatus (Coleoptera: Curculionidae: Scolytinae) Complex after the Rediscovery of the “Lost” Type Specimen

Ambrosia beetles of the Euwallacea fornicatus (Eichhoff, 1868) species complex are emerging tree pests, responsible for significant damage to orchards and ecosystems around the world. The species complex comprises seven described species, all of which are nearly identical. Given that the morphology-defined species boundaries have been ambiguous, historically, there has been much disagreement on species validity, which was compounded by the presumed loss of the type series of E. fornicatus. The species complex was recently reviewed using morphometrics to associate the type specimens to the clades delineated with molecular data under the assumption of the lost type series. We rediscovered a syntype of Xyleborus fornicatus, and reevaluated the species in the complex using morphometrics. We propose the following taxonomic changes to the species complex: Euwallacea fornicatus (=E. tapatapaoensis (Schedl, 1951); = E. whitfordiodendrus (Schedl, 1942)) syn. res.); E. fornicatior (Eggers, 1923) (=E. schultzei (Schedl, 1951) syn. nov.); E. kuroshio (Gomez and Hulcr, 2018) and E. perbrevis (Schedl, 1951) stat. res. These taxonomic changes shift the species name associated with the widely used common names for two taxa, namely: Euwallacea fornicatus should be used for the “Polyphagous Shot Hole Borer”, and E. perbrevis for the “Tea Shot Hole Borer clade a”. A lectotype is designated for X. fornicatus in order to stabilize the use of the name.


Introduction
Xyleborine ambrosia beetles (Coleoptera: Curculionidae: Scolytinae: Xyleborini) are among the most damaging and invasive organisms, with impacts ranging from economic to ecological. Ecological and biological features, such as fungus-farming, haplodiploidy, and a wide host range, make the Xyleborini one of the most successful colonizers [1]. With~1160 recognized species, this is the most species-rich tribe of Scolytinae, including some of the most widespread species [2].
Xyleborus fornicatus was described from an unknown number of specimens from Sri Lanka [6]. Eichhoff's collection and types were deposited in the Zoological Museum in Hamburg, Germany (UZHM), and were destroyed when the museum was bombed during World War II. Approximately a dozen Eichhoff types were saved by K.E. Schedl, and now reside at the Naturhistorisches Museum Wien (NHMW) [5] (p. 3). Given that the type specimens were not present in either UZHM or NHMW, the type series was listed as "lost" by Wood and Bright [5] in their catalog of bark and ambrosia beetles. However, a syntype was listed in a checklist of scolytine and platypodine types in the Zoological Museum, Museum and Institute of Zoology, Polish Academy of Science, Warsaw, Poland (MIIZ) [7], but this checklist was not listed in the catalog [5], and was subsequently largely ignored. This is one of several errors pertaining to type specimen deposition listed in Wood and Bright's [5] catalog (see [8,9]).
Beginning in the 1980′s populations of what was presumed to be Euwallacea fornicatus became established outside Asia in Central and North America, Israel, and South Africa [4,[10][11][12][13]. Stouthamer et al. [3] sampled E. fornicatus s.l. from multiple localities across Asia, as well as from introduced populations around the world. Their molecular phylogeny revealed that E. fornicatus was a complex consisting of three clades, presumed species, which were termed tea shot hole borer (TSHB), Kuroshio shot hole borer (KSHB), and polyphagous shot hole borer (PSHB). Gomez et al. [14] expanded the Stouthamer et al. [3] dataset and recognized the following four species: KSHB and PSHB, tea shot hole borer a (TSHBa), and tea shot hole borer b (TSHBb). These clade names are frequently used in the literature, but they do not have official taxonomic status and they are not officially recognized by the Entomological Society of America committee of common names. Gomez et al. [14] delineated species in the complex using morphometrics to associate molecular data with type specimens. With the assumption that the type series of E. fornicatus was lost [5], Gomez et al. [14] based their species concept on the specimens sequenced from the type locality of Sri Lanka, which corresponded to TSHBa clade.
In this review, we examined the rediscovered Xyleborus fornicatus syntype ( Figure 1) and additional non-type specimens of individuals in the Euwallacea fornicatus complex so as to revise diagnoses and provide new and revised distribution information.

Materials and Methods
We follow the clade names for the Euwallacea fornicatus species complex proposed by Stouthamer et al. [3], and those subsequently followed by Gomez et al. [14]. The syntype of Xyleborus fornicatus, four additional DNA vouchers of PSHB from Vietnam and Hong Kong, and two specimens from Vietnam, were measured and added into the mitochondrial cytochrome oxidase I (COI) dataset of Gomez et al. [14] (Table 1). Elytral and pronotal measurements followed the methodology of Gomez et al. [14], and were measured on the lateral view of the specimen. The elytral length was diagonally measured from the elytral base to the apex, and the pronotal length was diagonally measured from pronotal apex to the pronotal base. The pronotal width was measured at the widest point of the pronotum, and the elytra width equals the widest point of one elytron. We conducted a classification and regression tree (CART; rpart {rpart}) following Gomez et al. [14], with 94 individuals to corroborate relevant morphometric characters for clade discrimination. We used the CART model to predict the clade membership of the synonymized type specimens and the syntype of Xyleborus fornicatus. Statistical analyses were performed using R v.3.4.2 [15]. Table 1. GenBank accession numbers for the 70 COI sequences used in this study, and the associated clade name, species, and locality details.

GenBank Accession Number
Clade Species Country Locality The syntype was photographed using an Olympus SZX16 stereomicroscope; each image is a composite of up to 50 separate images taken with a Canon EOS Rebel T3i camera, and later stacked using the Helicon Focus software (v 6.0, Helicon Soft, Kharkov, Ukraine).
Additional museum specimens were also examined for distribution and host records. New distribution records are denoted with an asterisk. Because of the morphometric and protibial denticle number overlap between species, which can make species identification uncertain, the distribution data is divided into two parts, namely: records confirmed by DNA genotyping, and records based solely on morphology. The collections and codes are given below: Species of the Euwallacea fornicatus species complex are known to attack over 412 species of plants in 75 families [16], but because of the taxonomic confusion regarding species limits, most of these records cannot be definitely associated with a particular species. The host records that can be definitively associated with species are listed below.

Results and Discussion
Additional vouchers were added to the original dataset from Gomez et al. [14] (Table 1), with no changes in the CART model previously proposed. The CART analysis classified the specimens into the four phylogenetic clades (Figure 2). The classification accuracy for the training data (n = 65) was 82%. The classification accuracy of the remaining individuals (n = 29) was 76%. The obtained classification tree supports the findings in Gomez et al. [14], where the morphotype identity was primarily explained by the elytra length and the pronotum length. The previously synonymized types and the re-discovered X. fornicatus syntype were classified with the CART model into the pre-established clades. The types of synonymized species were classified consistently with Gomez et al. [14]. The X. fornicatus syntype was classified as the PSHB clade. The comparative measurements for the four species based on the original dataset in Gomez et al. [14], and the additionally measured specimens (by SMS and RAB), are described in Table 2. Based on the phylogenetic data, CART analysis, and location, we suggest the following taxonomic changes.

Results and Discussion
Additional vouchers were added to the original dataset from Gomez et al. [14] (Table 1), with no changes in the CART model previously proposed. The CART analysis classified the specimens into the four phylogenetic clades (Figure 2). The classification accuracy for the training data (n = 65) was 82%. The classification accuracy of the remaining individuals (n = 29) was 76%. The obtained classification tree supports the findings in Gomez et al. [14], where the morphotype identity was primarily explained by the elytra length and the pronotum length. The previously synonymized types and the re-discovered X. fornicatus syntype were classified with the CART model into the preestablished clades. The types of synonymized species were classified consistently with Gomez et al. [14]. The X. fornicatus syntype was classified as the PSHB clade. The comparative measurements for the four species based on the original dataset in Gomez et al. [14], and the additionally measured specimens (by SMS and RAB), are described in Table 2. Based on the phylogenetic data, CART analysis, and location, we suggest the following taxonomic changes.  Measurements for the total length, pronotal and elytral width, and length/width ratios are measured in the dorsal view, while the pronotal and elytral length are measured in a lateral view on a diagonal, as detailed by Gomez et al. [14].
Host plants.
Remarks. The types of E. fornicatior and E. schultzei were classified as belonging to the TSHBb clade based on body measurements. Gomez et al. [14] classified E. schultzei as belonging to TSHBb, but the species was retained as a synonym of E. fornicatus. The types were directly compared and were found to be conspecific. Given the results of the Gomez et al. [14] analysis, and the overall similarity of the type specimens, E. schultzei is here placed in synonymy with E. fornicatior.