Material and methods

Sample collection

Specimens were obtained by hand-collecting from karst landscapes in Battambang Province, Cambodia. Geographical coordinates and elevations were recorded in situ using a Garmin GPSMAP 60CSx receiver (WGS84 datum), with all locality data subsequently verified via Google Earth Pro v. 7.3.6. Live animals were photographed in their natural habitats using a Canon EOS 90D digital camera equipped with a Canon EF-S 60 mm f/2.8 Macro USM lens.

Fieldwork and sampling were conducted under the approval of the Animal Care and Use Committee of the Chulalongkorn University Animal Care and Use Committee (Protocol Review No. 1723018). Specimens were euthanized following the two-step method as per the AVMA Guidelines for the Euthanasia of Animals (AVMA 2020). For morphological studies, material was preserved in 70% ethanol, while specimens intended for molecular analysis were stored in 95% ethanol. In the latter case, the preservative was replaced with fresh 95% ethanol after 24 h to prevent defensive secretions from compromising DNA quality. All type specimens are deposited in the Museum of Zoology, Chulalongkorn University (CUMZ), Bangkok, Thailand.

Morphological study and scanning electron microscopy

Specimens were examined, measured, and illustrated using a Nikon SMZ 745T trinocular stereo microscope equipped with a Canon EOS 5DS R digital camera. Digital stereomicroscopic photography was executed utilizing Leica M205 FCA and Leica M205 FA fluorescence stereo microscopes, with image acquisition and focus-stacked data processing mediated via the LAS X microscope software suite. Traditional line drawings were subsequently prepared based on these focus-stacked digital renderings using a Leica DM500 microscope coupled with a Leica DMC2900 digital USB 3.0 microscope camera, the latter equipped with a 3.1 Megapixel CMOS sensor. To resolve critical, otherwise visually elusive micro-characters, observations and photography were further enhanced by means of phase-contrast microscopy utilizing a Leica HC PL Fluotar 63×/1.30 Oil Immersion Objective (contrast plan fluorite objective series). All digital images were finally processed and arranged into plates deploying Adobe Photoshop CS6. The initial line drawings were subsequently finalized based on these stacked photographs and further refined through rigorous, direct microscopic observations.

Descriptive terminology, including the carinotaxic formulae, follows the standards established by Golovatch et al. (2007a, 2007b, 2009, 2011). Body ring counts are in accordance with the methods used by Enghoff et al. (1993) and Golovatch et al. (2007a). In the catalogue sections, D stands for the original description and subsequent descriptive notes; K for the appearance in a key; L for the appearance in a species list; R for new subsequent records from Cambodia; M for a mere mention; and MI for molecular information. The abbreviations used for specific gonopodal structures are as follows:

ap = anterior coxal process;

cxp 1 = anterior coxosternal processes;

cxp 2 = posterior coxosternal processes;

f = flagellum process;

pp = paramedian coxal process;

te = telopodite.

For scanning electron microscopy (SEM), gonopods were mounted on aluminum stubs using conductive carbon tape and coated with a 5-nm platinum layer using a CCU-010 high vacuum sputter and a carbon coater (Safematic). Images were captured with a TESCAN VEGA3 scanning electron microscope operated at an acceleration voltage of 5 keV and a working distance of 25 mm using a secondary electron (SE) detector. Following SEM examination, gonopods were carefully removed from the stubs, rinsed in acetone to remove tape residue, and returned to 75% ethanol for long-term deposition in the museum collection.

Micro-computed tomography and 3D virtual reconstruction

Micro-CT imaging was performed using a SkyScan 1273 scanner (Bruker, Kontich, Belgium) operated at a source voltage of 40 kV and a source current of 50 μA without an X-ray filter. Images were acquired with a voxel size of 4.0 μm, a camera binning of 1×1, and a rotation step of 0.3° for a full 360° scan. The projection images were reconstructed using NRecon software (v. 2.2.0.6, Bruker). Two-dimensional (2D) images were processed and visualized using DataViewer software (v. 1.5.6.2, Bruker) and three-dimensional (3D) renderings were generated using CTvox software (v. 3.3.1, Bruker).

Volumetric 3D reconstructions of the computed tomography scans were compiled using the software XMReconstructor (v. 10.7.2936), with the resulting default parameters exported and saved in the standard 16-bit USHORT DICOM file format. Subsequent rendering and fine processing of the raw tomographic data were executed via AIVIA (v. 15). For detailed morphological examinations, 3D surface configurations were rendered using the volume rendering or “volren” processing options. The final virtual models were generated by strictly minimizing the color space thresholds, thereby guaranteeing that the superficial ectal cuticular structures of the specimens remained perfectly visible at the highest attainable resolution. The virtual reconstructions were dynamically rotated and manipulated along all anatomical axes to permit an exhaustive, non-destructive examination of the scanned material. Subsequently, the snapshot function of the software was used to capture high-resolution figures of the shaded surface volumes, which provided clear figures at a standard resolution of approximately 1281 × 732 pixels.

Molecular framework and genetic distance analyses

The phylogenetic framework and species-group assignment followed Seesamut et al. (2026). In the present study, molecular data were used primarily to evaluate the genetic divergence of the two new species from their closest congeners, rather than to reconstruct an independent de novo phylogeny (Fig. 1). COI sequences of Plusioglyphiulus group D _sensu_Seesamut et al. (2026) were retrieved from GenBank to calculate uncorrected pairwise genetic distances (p-distances) using MEGA X (Kumar et al. 2018). Pairwise distances were estimated among the examined taxa, including the new species and its congeners, using the partial deletion option with a site coverage cut-off of 95%. The resulting p-distance values were used to assess the level of genetic divergence among taxa and to provide additional support for species delimitation in combination with morphological evidence.

Figure 1.: Molecular phylogeny of Plusioglyphiulus group D sensuSeesamut et al. (2026) indicating the positions of P. battambangensis sp. nov. and P. parviserratus sp. nov. Nodal support values are given as SH-aLRT/aBayes/ultra-fast bootstrap values from the maximum likelihood analysis in IQ-TREE, followed by posterior probabilities from the Bayesian inference analysis in MrBayes. A black circle on a node indicates a well-supported clade, with SH-aLRT ≥ 80%, aBayes ≥ 0.95, BS ≥ 95%, and PP ≥ 0.95.