Materials and methods

Study site and specimen collection

All amphipod specimens from were collected using funnel traps baited with mackerel (Scombridae) and attached to free fall landers, as described in Jamieson et al. (2013). Landers were deployed during the 2024 Inkfish Tonga Trench Expedition aboard RV Dagon (Fig. 1). Specimens were fixed immediately in 99–100% ethanol. Ethanol was replaced after 24 hrs and samples were stored at −12 °C. Collected specimens were sorted and logged on Dagon and further sorted and stored at room temperature in the Minderoo-UWA Deep-Sea Research Centre collection until taxonomic examination.

Figure 1.: Map of study sites. Princaxelia malohi sp. nov. (green diamond) collected from the Tonga Trench during the 2024 Inkfish Tonga Trench Expedition aboard RV Dagon; Princaxelia kahat sp. nov. (blue diamond) collected from the Mariana Trench deployed from the vessels RV Falkor (FK 141109) in 2014 and TV Shinyo-Maru (SY1615) in 2017.

Specimens from the Mariana Trench were collected from landers deployed from the vessels RV Falkor (FK 141109) in 2014 (Drazen 2015) and TV Shinyo-Maru (SY1615) in 2017 (Fig. 1). Specimens were fixed in 99–100% ethanol and stored in the Minderoo-UWA Deep-Sea Research Centre collection.

Morphological examination

Specimens were imaged using a Canon EOS R5 DSLR camera with a Canon 100 mm f/2.8 VC USD Macro 1:1 VC lens. Images were stacked using the software Helicon Focus v.8.3.7 (Helicon Soft). Specimens were dissected in 99% ethanol under a Leica M205C dissecting microscope. The limbs on the left side of the specimens were used for dissection to keep consistent with prior descriptions. Select body parts were stained using Shirlastain-A fibre identification stain to help identify finer morphological features. Dissected appendages were mounted onto microscope slides using glycerol and imaged using a Nikon Eclipse Si compound microscope and a mounted Tuscen True chrome metric HDMI camera.

Composite images of the dissected body parts were used as a reference to draw taxonomic figures as per Coleman (2003) using the open-source vector illustration software Inkscape v. 1.4.2. Length ratios for the appendages were derived from images using the open-source image analysis software ImageJ (Schneider et al. 2012). The holotype specimen of Princaxelia malohi sp. nov. is deposited at the Museum of New Zealand Te Papa Tongarewa (NMNZ) under the accession number CR.028062. The holotype specimen of Princaxelia kahat sp. nov. is deposited at the Smithsonian National Museum of Natural History (USNM) under the accession number of USNM 1775248. Additional material of both species is stored within the Minderoo-UWA Deep-Sea Research Centre collection at The University of Western Australia.

Scanning electron microscopy

Supplemental SEM micrographs of selected diagnostic features from gnathopods of additional specimens from Princaxelia malohi sp. nov. were taken using a JEOL NeoScope JCM-7000 benchtop scanning electron microscope (SEM). Gnathopods were rinsed with 100% ethanol 4× and processed in a PELCO Biowave microwave processor for 1 min each rinse. Samples were subsequently dried using a Polaron E3000 critical point dryer as per the manufacturer instruction. Specimens were mounted on aluminium stubs with the lateral side facing up using conductive carbon tape and coated in carbon using an HHV Auto-306 Carbon Coater. Gnathopods were then coated with a layer of platinum using a Leica EM ACE600 Metal Coater. Images from the SEM were captured using secondary electron imagery with the machine set to high vacuum mode, and an accelerating voltage of 15.0 kV and working distance of 13.7–14.5 mm was used for all secondary images. Additional images were taken using backscattered electron imagery on a FEI Verios XHR SEM with an accelerating voltage of 20.0 kV and a working distance of 5.5 mm. All micrographs were generated at the Centre for Microscopy Characterisation and Analysis at The University of Western Australia.

DNA barcoding

All five individuals of Princaxelia malohi sp. nov. and 12 individuals of Princaxelia kahat sp. nov. were selected for barcoding. All types specimens were included in the DNA barcoding pool to reduce future taxonomic issues (D’Acoz and Havermans 2015). Total genomic DNA was extracted from the pleopods using the DNeasy Blood and Tissue Kit (Qiagen) according to the manufacturer’s protocol. The mitochondrial barcoding regions, 16S rRNA, 28S rRNA, and cytochrome c oxidase subunit I (COI), were targeted based on sequences available in GenBank®. The primers used for 16S rRNA were from (Folmer et al. 1994), 6SFt_amp (5’-GCRGTATIYTRACYGTGCTAAGG) and 16SRt_amp2 (5’-CTGGCTTAAACCGRTYTGAACTC). COI was amplified with LCO1490 and HCO12198 (Folmer et al. 1994). PCR products were assessed using 96-well E-gels (Invitrogen) and purified with AMPure XP paramagnetic beads (Beckman Coulter). 28S was amplified using the primers from Foltz et al. (2007), 28Srtw (5’-ACTTTCCCTCAYGGTACTTGT) and 28Sftw (5’-AGAAACTAACMAGGATTCCYYTAGTA). Sequencing preparations used the Thermo Fisher Scientific Applied Biosystems BigDye Cycle Sequencing Kit, and clean-up was done using the CleanSEQ Dye-Terminator Removal Protocol (Beckman Coulter). Sequencing was performed on a 3730xl capillary sequencer (Thermo Fisher Scientific Applied Biosciences) at Macrogen (Korea). Sequence assembly and editing were done in Geneious Prime 2023.0.1 (Kearse et al. 2012). Newly generated sequences were deposited in GenBank® under the following accession numbers (PZ585448; PZ611770–PZ611786; Suppl. material 1).

Generated sequences were assembled and cleaned in Geneious Prime v. 2025.1.3. Newly sequenced data was further supplemented by all publicly available sequence data listed on GenBank® (Suppl. material 1). Individual gene datasets (COI and 16S rRNA) were aligned using the Geneious MAFFT alignment plugin (Katoh et al. 2005). All COI alignments were subsequently checked for stop codons prior to further analyses. Pairwise genetic distances were calculated separately for each marker using a custom Python implementation of uncorrected p-distance and Kimura two-parameter (K2P) distance. Sites containing gaps or ambiguous nucleotides were excluded using pairwise deletion.