Approximately 30 Physalia floats bearing zooid colonies were discovered stranded along a 1.5 km span of Gamo Beach with coordinates 38.259110° N, 141.018898° E, near the mouth of Nanakita River, Sendai Bay, Sendai City, Miyagi Prefecture, Japan on 11 and 12 June 2024 (Figures 2A–J). Of these, six Physalia specimens were collected, and placed into a 2 L PET bottle (smaller individuals) and a wide-mouth container (larger individuals) filled with seawater and then transported to the Graduate School of Agricultural Science, Tohoku University (Aobayama Campus) for further observation and analyses. Approximately 1 cm of tissue was excised from each colony and placed in separate 1.5 mL tubes containing 99.5% ethanol for molecular analysis. Extensive photo-documentation (iPhone and Huawei P30) was conducted on live specimens that were then fixed in 10% buffered formalin solution for morphological observations using a dissecting microscope (Nikon SMZ745T). Holotype and paratype specimens were deposited into the Tohoku University Museum (Accession numbers 112960–112965 respectively). Two additional Physalia colonies collected from Okinawa Prefecture on 28 January 2025 (Figure 2N), were included for molecular and morphological comparison. Specimen vouchers were also deposited in the Tohoku University Museum as P. utriculus (Accession numbers 112966 and 112967).

Shortly after these Physalia collections efforts at Gamo Beach, Sendai City, on 14 June 2024 a Tohoku regional television station aired warnings about unprecedented mass strandings of Physalia along Gamo Beach, Sendai Bay (Figure 2K); the broadcast also discussed Physalia sightings in the Kanto region, whose morphology on the television screen suggests it is P. utriculus (Figure 2L). In the same summer, signs were posted at San Ole Beach in Shizugawa Bay, Minamisanriku, more than 90 km north of Gamo Beach, alerting bathers of “dangerous man-of-war jellyfish sightings” there on 9 Aug 2024 (Figure 2M).

In this study, morphological examination was conducted of specimens collected from Tohoku and Okinawa, Japan. Morphometric parameters of the different zooids and general aspects of the colony were described with terminology used by (Totton and Mackie, 1960) (Bardi and Marques, 2007), and (Munro et al., 2019) to refer to P. physalis from the Atlantic Ocean. Additional systematic morphometric analysis of the Tohoku specimens included: float length and width; number of wrinkles on the crest of the pneumatophore; tentacle length; gastrozooid and tentacular palpon length and diameter; gonopalpon length; and gonophores, nectophores, and gonodendron diameters. Terminology follows current standards, with “posterior zone” and “principal tentacle” used in place of outdated terms such as “oral zone” and “primary tentacle” sensu (Munro et al., 2019). Measurements were taken directly using a DN150 SCITOOLS digital caliper (resolution 0.1 mm/0.01 inch, accuracy ±0.2 mm/0.01 inch), and from high-resolution images using ImageJ v1.54f (NIH, USA). The cnidome (repertoire of nematocyst types) was described according to Weill’s (1934) classification system (Weill, 1934) and subsequent modifications by (Östman, 2000). A comparative analysis of key character traits among the two species collected from Japan showed the Okinawa colonies mainly differ from P. mikazuki sp. nov. in having a single prominent principal tentacle, a distinct arrangement of the zooid cluster (formerly termed “cormidia”), and differences in gastrozooid morphology (Figures 2N, 3), discussed further in Section 3.

DNA was extracted from the ethanol-preserved secondary tentacle tissue subsampled from Physalia individuals (n=4) using the QIAGEN DNeasy Blood and Tissue Kit following the developer’s protocol. Two mitochondrial gene regions, a ~550 bp fragment of the 16S ribosomal RNA gene (16S rRNA) and a ~800 bp fragment of the cytochrome c oxidase subunit I (COI) fragment were targeted for amplification via polymerase chain reaction. The 16S rRNA region was amplified using the primers med-rnl-F 5’- GACTGTTTACCAAAGACATAGC-3’ and med-rnl-R 5’-AAGATAGAAACCTTCC TGTC-3’ (Lawley et al., 2016), while the COI region was amplified using the primers Jellyfish_CO1_F 5’-KKTCACAAAYCATAAAGATATWGG -3’ and Jellyfish_CO1_R2 5’-GGAACTGCTATWATCATWGTWGC-3’ (Minamoto et al., 2017). PCR protocol for 16S were: initial denaturation (94°C for 3 mins), then 38 cycles of amplification comprising denaturation (94°C for 30s), annealing (54°C for 30s), extension (72°C for 45s), and final extension (72°C for 7 mins); in COI amplification protocol was modified during initial denaturation (95°C for 5 mins), denaturation (96°C for 30s), annealing (50°C for 40s), and extension (72°C for 55s).

The PCR product was visualized on 1.5% agarose gel and used for Sanger sequencing at Onagawa Field Centre (Graduate School of Agricultural Science, Tohoku University) following the BigDye™ Terminator v3.1 Cycle Sequencing Kit protocols (Applied Biosystems, Thermofisher Scientific, Japan). Samples were cleaned using Agencourt AMPure XP beads (Beckman Coulter, Inc., CA, USA) prior to sequencing. Resulting nucleotide sequences were checked and edited in Geneious Prime (ver. 2025.1, https://www.geneious.com), and consensus sequences (n=6) were subsequently generated from forward and reverse strands.

All available 16S and COI sequences for Physalia spp., including two outgroup species from the sister family Rhizophysidae (order Siphonophorae, suborder Cystonectae), were retrieved from NCBI GenBank (accessed on 13 March 2025). Additionally, as none of the sequences generated by Church et al., 2025, for their 16S or COI phylogenetic analyses was accessioned into GenBank we obtained alignments for 16S (https://github.com/shchurch/Physalia_population_genomics/blob/main/results/iqtree/alignments/16S.aln.fasta) and COI (https://github.com/shchurch/Physalia_population_genomics/blob/main/results/iqtree/alignments/CO1.aln.fasta)for multiple Physalia samples (from the Pacific, Atlantic, and Indian Oceans) from the GitHub link referenced in the supplementary material. In total, 129 sequences were analyzed for 16S and 198 for COI. Accession numbers, sample identifiers, and localities are listed in Supplementary Table S1. All sequences were aligned using the G-INS-I algorithm in MAFFT (ver. 7.520, https://mafft.cbrc.jp/alignment/software/) (Katoh and Standley, 2013). Maximum likelihood (ML) phylogenetic trees were reconstructed using IQ-TREE v2.2.2 (Trifinopoulos et al., 2016). The best-fit substitution models were selected using the lowest Akaike Information Criterion (AIC) scores via the integrated model selection function in IQ-TREE (16S: TIM3+F+G4; COI: GTR+F+G4+I). Node support was evaluated using the Shimodaira–Hasegawa approximate likelihood ratio test (SH-aLRT), and the approximate Bayes test (aBayes) (Anisimova et al., 2011) and 1000 ultrafast bootstrap replicates (UFBoot2) (Hoang et al., 2018). To compare and validate tree topology, additional ML analyses were performed in MEGA version 12.0 using the GTR+G4 model (Kumar et al., 2024). Final trees were visualized and annotated using FigTree v1.4.4 (Rambaut, 2009), Interactive Tree of Life (iTOL) (Letunic and Bork, 2024) and Inkscape software (Inkscape, 2024).

The sampling location map for Physalia individuals (Figures 2A–C) was generated using QGIS-LTR Prizren version 3.34, a robust, free, and open-source geographic information system (GIS) software widely utilized for geospatial data analysis and mapping (all software documentation available at https://qgis.org/). City-level administrative boundary data for Japan was sourced from the Global Administrative Areas (GADM) database accessible at https://gadm.org/download_country.html (accessed on 12 September 2024). To enhance the map’s contextual accuracy and provide a realistic background for the sampling locations, a hybrid map layer from Google Maps was imported as XYZ tiles. This was achieved through the QGIS “Add Layer” plugin, which facilitated seamless integration of satellite imagery and map layers, thereby offering an intuitive and spatially relevant overview of the study area.

The environmental indicators for modeling, including daily surface current and temperature data were obtained from HYCOM (Hybrid Coordinate Ocean Model), specifically the GLBy0.08 dataset, which provides global ocean data at a 1/12° spatial resolution in NetCDF format. This study focused on the coastal area encompassing Sagami Bay to the Pacific coastal area of the Tohoku region (34° – 42° N and 138° – 146° E), and particularly around the Sendai Bay area (37.4° – 38.6° N and 140.6° – 142.2° E). To examine recent marine environmental trends, oceanographic data from 1 January 2018, to 31 August 2024, were obtained from the HYCOM Data Server (https://www.hycom.org/dataserver; accessed on 3 September 2024) and, subsequently, processed using the Python programming language. The details for each indicator are as follows:

Phylum Cnidaria Verrill, 1865

Subphylum Medusozoa Peterson, 1979

Class Hydrozoa Owen, 1843

Subclass Hydroidolina Collins, 2000

Order Siphonophorae Eschscholtz, 1829

Family Physaliidae Brandt, 1835

Genus PhysaliaLamarck, 1801

Physalia mikazuki sp. nov. Yongstar, Ochiai & Lewis Ames (Figures 4–6)

Genus description: Original French, “Corps libre, gélatineux, ovale, comprimé sur les côtés, et ayant sur le dos une crête élevée, rayonnée et membraneuse. Des tentacules nombreuses, filiformes, articulées, placées sous le ventre, et qui paraissent être des suçoirs.” (Lamarck, 1801); English translation: “Free body, gelatinous, oval, compressed on the sides, and having on the back a high, radiated and membranous crest. Numerous tentacles, filiform, articulated, from under the ventral side, which appear to be suckers.” (Lamarck, 1801)

Type species: Physalia physalis (Linnaeus, 1758) Figures 1, 7. Original description reproduced herein along with the translation of the original Latin description as Figure 7B. More than 50 synonymized species names exist (Pugh, 2019) for this type species many of which are insufficiently known to be considered valid. Originally called Holothuria physalis (Linnaeus, 1758), it took 140 years before the name Physalia physalis was established by (Schneider, 1898). No type specimen voucher exists. However, an illustration in (Sloane, 1707) to which Linnaeus referred in his species description has been reproduced herein as Figure 7A.

Valid species: Physalia physalis (Linnaeus, 1758); Physalia minuta Church & Dunn, 2025 (Church et al., 2025); Physalia mikazuki sp. nov. described herein. Though P. utriculus (Gmelin, 1788), P. megalistaLesueur and Petit, 1807 are among those synonymized as junior synonyms of P. physalis due to lack of type reference vouchers, they likely are valid species based on distinct morphological descriptions and illustrations previously published (Figures 1, 7–9 and a summary is provided in Table 1), a claim recently corroborated by a multi-genomic analysis (see (Church et al., 2024); (Church et al., 2025)), and distribution evidence in the literature and citizen science efforts, but, nonetheless lack formal species descriptions.

Physalia utriculus (Gmelin, 1788). An illustration in (de La Martinière, 1787) to which Gmelin (1788) referred in his species description (Gmelin, 1791; de La Martinière, 1787); both have been reproduced herein as Figures 8A, B, and described in detail in the corresponding figure header.

Physalia megalistaLesueur and Petit, 1807. An illustration by Lesueur & Petit in Péron (1807) (de La Martinière, 1787; Lesueur and Petit, 1807; Péron, 1807) used for their initial description and then later to which Lamarck (1816) (Lamarck, 1816) referred in his attempt to redescribe the species (Figure 9C), however, the authorship remains as P. megalista (Lesueur and Petit, 1807); the species description and illustration have been reproduced herein as Figures 9A, B, and described in detail in the corresponding figure header.

A recent study of global Physalia populations by (Church et al., 2025) used a genomic approach that combined morphological data (museum vouchers and iNaturalist images) to resolve multiple species that corresponded to original descriptions for P. physalis (as Cluster A), P. utriculus (as Cluster B1), P. megalista (as Cluster C1). Additionally in the study by (Church et al., 2025), two potentially new species of Physalia were identified – one as Cluster C2 (sensu (Church et al., 2024) (Church et al., 2025)) and described as Physalia minuta Church & Dunn, 2025, present in the Tasman Sea, and the other as Cluster B2 present in Sagami Bay (overlapping with B1 sensu (Church et al., 2024) (Church et al., 2025)) with matches to GenBank sequences for Physalia from Pakistan and Mexico but with insufficient morphological data for species comparison therein.

Physalia mikazuki sp. nov. is the first species in the genus to be described in Japan. Previous works named P. physalis as the sole Physalia species in Japan (Pugh, 2019), but recently (Oguchi et al., 2024) described the development of P. utriculus from Kanagawa Prefecture (Japan) without an explanation for the use of nomenclature. However, given that the photo and drawing depicted therein bear a prominent principal tentacle it is likely P. utriculus (“Cluster B1” (Church et al., 2024; Church et al., 2025)).

Abbreviations: PL: Pneumatophore Length - measured from the anterior end of the main zone to end of the posterior zone (previously oral zone) of pneumatophore (float); PW: Pneumatophore Width - measured from the apex to the basal region of pneumatophore; PD: Pneumatophore Diameter - measured at the widest point of pneumatophore; NPW: Number of pneumatophore wrinkles; NPT: Maximum number of principal tentacles - the main tentacle, develops from the first tripartite group of a zooid cluster in the main zone (much larger and thicker at its base compared to the secondary tentacles); TUM (Tohoku University Museum).

Distribution: Tohoku and Kanto regions of Japan; Pakistan and Mexico.

Type locality: Gamo Beach, Sendai City, Miyagi Prefecture, Japan.

Material examined: Holotype: 112960 (TUM), mature, female, 67 mm PL, 29 mm PW, 19 NPW, 3 NPT, Gamo Beach (Sendai Bay), Sendai City, Miyagi Prefecture, Japan, 12 June 2024.; Paratypes: 112961 (TUM), mature, 73 mm PL, 30 mm PW, 20 NPW, 3 NPT; 112962, mature, 61 mm PL, 23 mm PW, 12 NPW, 3 NPT; 112963, 49 mm PL, 19.5 mm PW, 9 NPW, 1 NPT; 112964, juvenile, 21 mm PL, 9 mm PW, 5 NPW, 1 NPT; 112965, juvenile, 24 mm PL, 11 mm PW, 5 NPW, 1 NPT, all from Gamo Beach (Sendai Bay), Sendai City, Miyagi Prefecture, Japan, 12 June 2024.

Etymology: The Japanese word “mikazuki” refers to the “crescent moon” shape of the warrior helmet worn by Samurai Masamune Date (1567 – 1636) of the Tohoku region who founded Sendai City. Vernacular “mikazuki no eboshi” (Japanese), “crescent helmet man-of-war” (English).

Diagnosis: Physalia mikazuki sp. nov. Yongstar, Ochiai & Lewis Ames is distinguished from other members of the genus Physalia by a combination of morphological traits: Pneumatophore length range 9.25–72.4 mm, with maximum known size smaller than that reported for P. physalis (8.1–134 mm) but overlapping with P. utriculus. Coloration of the crest is bluish with deep blue to purple hues and membrane is a translucent bluish-green (vs. dark green/carmine in P. physalis, blue and clear-glassy in P. utriculus, transparent with green patch at anterior apex in P. minuta). Up to six zooid clusters are present in the posterior zone (vs. five in P. physalis). Three primary zooid cluster groups are found in the main zone (vs. up to seven in P. physalis and one in P. utriculus). Posterior zone gastrozooids are cylindrical, elongate and banana-shaped (vs. bulbous, rounded with yellow apex in P. utriculus). In the main zone, P. mikazuki has gastrozooids that are elongated with distally swollen balloon-like yellow tips (vs. shorter, flask-shaped gastrozooids in P. utriculus, and elongated, blue at base and pale at tip in P. minuta). More than two principal tentacular palpons are present (vs. 1–2 in P. utriculus and P. megalista), each arising from a separate zooid cluster. Tentacle axial tissue is significantly widened with an axial width ratio of approximately 2:1. Trait comparison among P. mikazuki and other Physalia species is summarized in Table 2.

Description: Physalia mikazuki sp. nov. Yongstar, Ochiai & Lewis Ames. Colonies with an asymmetrical triangular sail-shaped, gas-filled pneumatophore with bluish coloration and deep bluish to purple hues in the upper region. Translucent structure with transparent bluish-green membrane (Figures 4A). Size ranging from 9.25–72.40 mm (n=37) in total pneumatophore length, 8.90–35.20 mm in pneumatophore width (n=6) and 8.10–27.50 mm in pneumatophore diameter (n=6), with a longitudinal wrinkled crest showing 5–20 wrinkles (Figure 4A).

Description of zooid arrangement follows the general pattern described for Physalia physalis by (Totton and Mackie, 1960) (Bardi and Marques, 2007), and (Munro et al., 2019). Zooids arranged ventrally in clusters budding from the basal region of the pneumatophore. Colony divided into two zones: posterior zone and main zone, separated by a basal internode (a gap devoid of polyps) (Figure 4D) seen also in P. utriculus, but least demarcated in P. physalis.

Posterior zone contains up to six zooid clusters and a protozooid (PZ), the first gastrozooid. Clusters I, II, and III form the reduced group, composed solely of gastrozooids with yellow-tipped oral parts that increase in length from cluster I (shortest). Clusters IV, V, and VI constitute the tripartite group, each comprising a gastrozooid, a tentacular palpon, and a gonodendron (Figure 4D).

Main zone, situated aboral to the posterior zone, is more developed and organized into three primary zooid cluster groups (Figure 4D), each containing up to two or three zooid clusters showing varied composition and developmental stages; an initial reduced group with a single gastrozooid and a gonodendron at its base; a primary tripartite group with a gastrozooid, a tentacle with ampulla or tentacular palpon [refers to a zooid type found in P. physalis, this zooid is responsible for bearing the tentacles (Munro et al., 2019)], and a gonodendron; a sequence of lateral tripartite groups branching from the primary tripartite group; and secondary basal buds groups as trifid structures, with additional trifid secondary basal buds emerging from the base of each zooid group.

Gonodendra, branched, composed of complex reproductive structures (Figure 5A), including male or female gonophores, bearing nectophores, jelly polyps, and palpons (as in P. physalis) (Figure 5B). Subterminal branch end with elongated vestigial nectophore and palpon (Figure 5B). Terminal branch end with jelly polyp and palpon (Figure 5B). Palpons concentrated at distal ends of subterminal and terminal branches (Figure 5B) bearing stenotele nematocysts (Figure 5C).

P. mikazuki gastrozooids in posterior zone are elongated and banana-shaped (Figure 6) with length range of 0.92-13.79 mm (n=69). In the main zone, gastrozooids are elongated with distally swollen, balloon-like yellow tips (Figure 4E) with length ranging from 0.883-24.871 mm (n=236).

Tentacular palpon of P. mikazuki includes principal tentacles and secondary tentacles. Principal tentacle developed from the first tripartite group of a zooid cluster in the main zone. This tentacle is much larger and thicker at ampulla (1.709 mm) compared to the secondary tentacles (0.879 mm) with wider tentacle axial tissue (the region formed by epitheliomuscular cells longitudinally arranged along mesoglea lamellae adhering to both sides of tentacular axis): about twice the thickness of secondary tentacles (Figure 10). Tentacles bear prominent nematocyst batteries arranged along their length, which are spirally arranged as button-like structures along the tentacles (Figure 10).

Despite the limited number of specific traits reported in the literature for P. utriculus, colonies from Okinawa observed in this study were clearly morphologically distinct from P. mikazuki sp. nov. in having a prominent single principal tentacle (Figures 3A,B), a distinct zooid cluster arrangement, and gastrozooid of distinct morphology (Figures 3C, D). In contrast to P. mikazuki, P. utriculus exhibits main zone gastrozooids that are shorter, flask-shaped, and opaquer in texture, with less distinctly defined yellow tips in a more irregular, tightly clustered arrangement (Figure 3D). Posterior zone gastrozooid buds of P. utriculus are bulbous and rounded, with a swollen body and a yellow apical region (Figure 3C).

Cnidome: Two main types of nematocysts observed in P. mikazuki: isorhiza (large and small size classes) and stenoteles (one size class). Isorhizas, spherical, mean length of 11.012 ± 0.20 μm for small size class and 22.511 ± 0.41 μm for large class size; both present in principal tentacles (Figure 11A), secondary tentacles (Figures 11B) and gastrozooids. Stenoteles, spherical, mean length of 18.07 ± 0.33 μm, present on gonopalpons (Figure 11C). All measurements were made after preservation in 10% formalin. The cnidome of P. mikazuki sp. nov. is identical to that of P. physalis, as reported by (Bardi and Marques, 2007) and thus, not diagnostic at the species level in this case.

Phylogenetic analyses of the mitochondrial 16S rRNA region revealed the distinctiveness of the Physalia specimens collected from Gamo Beach, Japan (Figure 12A). The four 16S sequences generated in this study (GenBank accession numbers PV469480–PV469483) formed a strongly supported monophyletic clade, with SH-aLRT/aBayes/ultrafast bootstrap values of 95.3/1.00/90, respectively. This clade includes two specimens from Japan (YPM-IZ-110931 and YPM-IZ-110943) and one from Mexico (YPM-IZ-110878), previously reported by (Church et al., 2025). Notably, this group is phylogenetically distinct from other known Physalia species, including P. physalis, P. utriculus, and P. megalista. Analysis of the COI region yielded a matching pattern. The four COI sequences from this study (PV452818–PV452821) also formed a monophyletic clade, this time with full support (SH-aLRT/aBayes/ultrafast bootstrap = 99.3/1.00/100), clustering with sequences from Mexico and Pakistan. The clade is Cluster B2 of (Church et al., 2024; Church et al., 2025), which was not formally described as a new species.

To validate the topology inferred by IQ-TREE, maximum likelihood (ML) trees were reconstructed in MEGA11 for both markers. These reconstructions yielded comparable topologies, with bootstrap support of 87.6% (Std. Dev. = 1.06348) for 16S and 97.9% (Std. Dev. = 0.58944) for COI. Pairwise evolutionary divergence estimates between clades were calculated using the Kimura 2-parameter (K2P) model in MEGA (Figure 12B). For the 16S marker (noting its slower substitution rate), the clade here proposed as Physalia mikazuki sp. nov. diverged from P. physalis by 2.88%, from P. utriculus by 2.43%, and from P. megalista by 4.28%. For the COI marker, divergence from these same species were more pronounced, at 8.55%, 4.87%, and 11.82%, respectively.

Based on consistent molecular distinctiveness across two mitochondrial loci, high nodal support from multiple phylogenetic inference methods, and distinguishing morphological traits described in Section 3.1, we designate the name Physalia mikazuki sp. nov. for this genetically and morphologically distinct species and Gamo Beach as the type locality. The description of a new Physalia species herein is also the first verified record of the genus in the Tohoku region of Japan. Overall, these findings enhance current understanding of species diversity within Physalia and emphasize the value of integrating molecular and morphological evidence for resolving the taxonomy and biogeographic patterns of pelagic hydrozoans.

Remarks. Together the morphological and molecular analyses of Physalia specimens collected at Gamo Beach show it is a distinct species that we herein designate Physalia mikazuki sp. nov. Diagnostic traits include a smaller pneumatophore (9.25–72.4 mm) than P. physalis, a crest with bluish coloration and deep blue to purple hues, a translucent greenish membrane, and up to six posterior zone zooid clusters (vs. five in P. physalis). The main zone contains three primary zooid cluster groups (vs. seven in P. physalis, one in P. utriculus). Posterior gastrozooids are a distinctly elongated, banana-like shape (vs. bulbous with yellow apex in P. utriculus), and gastrozooids in the main zone are elongated with distally swollen, balloon-like yellow tips. Each colony possesses more than two principal tentacular palpons with tentacle axial tissue approximately twice as wide as that of secondary tentacles. These features, summarized in Table 2, support the recognition of P. mikazuki as a novel species. The taxonomy of the genus Physalia has long been contested (Totton and Mackie, 1960). synonymized historical names under P. physalis, arguing for a single globally distributed species and highlighting the dimorphic characteristic of left- versus right-handed “sails” (floats) as a possible mode of drift in either wind direction. In contrast, much earlier Chun (1897) and Schneider (1898) proposed species delineation between the so-called Atlantic P. physalis and the Indo-Pacific P. utriculus based on morphological differences (Garstang, 1946). Recent reports, particularly those from Australian waters, have further challenged the current taxonomic constraints, suggesting the presence of multiple Physalia species with distinct morphological traits (Pugh, 2019). The long-held belief that P. physalis is the sole species in the genus stemmed from a lack of reliable morphological characters to unequivocally delineate potential taxa. In this study, Tohoku specimens exhibited consistent morphological differences from P. physalis, P. minuta, P. utriculus, and what is illustrated in the literature as P. megalista (Figures 1, 7–9), including variation in pneumatophore size, coloration, gastrozooid structure, and tentacle morphology. In particular, the shape of gastrozooids and the number and structure of principal tentacles provide strong distinguishing features for P. mikazuki. Twenty years ago, Collins et al., 2005 revealed the utility of rnl (16S rRNAgene) for broad-scale species delineation among hydrozoans, one of the four classes of the subphylum Metazoa (jellyfishes) (Collins et al., 2005). Later, a study by Ortman et al., 2010 designed to delineate population boundaries of Medusozoa, showed that the mitochondrial COI target was an appropriate DNA barcode for estimating species-level distinction within the Medusozoa using Kimura 2-parameter genetic distances (K2P) (Ortman et al., 2010). Though no solid cut-off exists for species delimitation, as sequence divergence values are notable variable between hydrozoan species, guidelines for certain clades (Schuchert, 2005; Collins et al., 2008; Schuchert and Collins, 2021) are reported using difference models (but see (Srivathsan and Meier, 2012)). Furthermore, Zhen et al. (2014) showed that K2P values differ between the mtCOI and 16S rRNA genes, with the former being a faster evolving region (higher substitution rate) among Medusozoa species (Zheng et al., 2014). That study concluded that while 16S might be more efficient for validation at the family level and below, hydrozoan taxa are identified at the species level using either barcode marker alone. Thus, given that proven efficacy of the 16S rRNA and mtCOI genes in medusozoan phylogenetic reconstruction studies over the past 20 years, we are confident in the capacity for these two standard mitochondrial markers, analyzed in tandem with morphological data, to delineate P. mikazuki sp nov. as the newest species of Physalia.

Species distribution ranges are often influenced by changes in environmental conditions (Hoegh-Guldberg and Bruno, 2010). In this study, we conducted a temporal and spatial analysis of oceanographic data of ocean currents and temperature from 2018 to 2024, to examine potential shifts that could affect marine ecosystems in Japan’s northern coastal waters. Our aim was to assess whether Physalia “Clade B2” (as documented by (Church et al., 2024; Church et al., 2025)) sampled from Sagami Bay in the Kanto region of Japan where P. utriculus [“Clade B1” sensu (Church et al., 2024; Church et al., 2025)] is also reported (Okada, 1932) could, in fact, be moving northward to the Tohoku region.

Data processing revealed that variation in oceanographic parameters from 2018 to 2022 was relatively consistent, prompting us to consolidate these years for separate comparisons with more recent data from 2023 and 2024, respectively (Figure 13). The comparative analysis indicated notable differences between the earlier period (2018–2022) and more recent years (2023 and 2024), particularly in terms of changes in ocean currents and SST (sea surface temperature) (Figure 13).

Recent oceanographic observations reveal a notable shift in the Kuroshio Current’s intensity and trajectory, with its path moving approximately 2° latitude northward in 2023 and 2024 compared to earlier periods (Figure 13, top panel). This northward shift may have facilitated the transport of marine organisms, including P. mikazuki, to higher latitudes due to altered current pathways. Concurrently, our analysis of the spatial distribution of average temperatures along the Pacific coastal area of Japan, from Sagami Bay to the northern region (Figure 13, middle panel), reveals generally warmer conditions. In particular, the monthly average temperatures near the sampling location in Sendai Bay consistently recorded approximately 2-4°C higher throughout 2023 and 2024 compared to the period from 2018 to 2022 (Figure 13, bottom panel). This indicates that the northern region, particularly near Sendai Bay, experienced relatively warmer conditions compared to previous years. This SST jump is considered to have occurred globally and is dubbed a “record-shattering” extreme event by ocean modelers (Terhaar et al., 2025).

Previously, Ferrer & Gonzales (2021) used an improved version of the Lagrangian particle tracking model SOFT (Sediment, Oil spill and Fish Tracking model) to show wind (drift) as the main mechanism controlling circulation on the sea-air interface along the Basque coast (Bay of Biscay) where thousands of Physalia individuals converge, with the majority being right-handed; thus hypothesizing drift from the Sargasso sea to American and European shores (Ferrer and Gonzalez, 2021). To test for additional factors controlling circulation of Physalia mikazuki sp. nov., we conducted a particle trajectory simulation using OceanParcels in Python “A highly customizable Lagrangian simulation framework” (https://oceanparcels.org/) based on the daily surface current data as another approach to investigate the possibility of P. mikazuki being transported northward by ocean currents. The simulation was initiated from Sagami Bay, Kanagawa region where the presence of Cluster B2 was initially reported (Kumar et al., 2024). The trajectory simulation included the period from May to August 2024, as P. mikazuki was discovered in Sendai, Miyagi in mid-June (this study) and in Hachinohe, Aomori in mid-August (from an iNaturalist report https://www.inaturalist.org/observations/245526594). Based on the particle simulation results, particles representing potential drifting Physalia individuals could reach Sendai within approximately 30 days and Aomori within 45 days (Figure 14).Visualization of the complex movement patterns and mapped distribution of particles, shaped by variability in current velocities, particularly the influence of the Kuroshio Current (Figure 15) provides insights into the mechanisms driving both biological and physical dispersal processes in the region.

The potential divergence of Physalia mikazuki sp. nov. in the temperate waters of Northeast Japan likely reflects complex interactions between environmental, oceanographic, and evolutionary processes. Ocean currents such as the Kuroshio and Oyashio not only influence the physical transport of neustonic organisms but also generate dynamic thermal fronts and eddies that can act as semi-permeable barriers to gene flow (Pelc et al., 2009; Treml et al., 2015). These thermal barriers, marked by steep temperature gradients and seasonal variability, may limit the dispersal of early developmental stages, especially in colonial cnidarians with environmentally sensitive larval forms (Bowen et al., 2013). Moreover, the bifurcation of these current systems along the Tohoku coastline results in regional water mass separation, potentially reducing interbreeding opportunities among Physalia populations. Over time, such physical and thermal discontinuities can lead to reproductive isolation, especially if reinforced by local adaptation or phenological shifts (Palumbi, 1994). These conditions, combined with environmental stochasticity and localized retention, may have contributed to the emergence of P. mikazuki sp. nov. as a distinct lineage. Similar processes have been proposed for other pelagic and neustonic taxa, highlighting the role of oceanographic complexity in driving cryptic speciation (Knowlton, 2000; Church et al., 2025). Our findings add to this growing body of evidence, suggesting that even wind-drifted surface organisms like Physalia are subject to fine-scale biogeographic structuring driven by both hydrographic and thermal constraints (Munro et al., 2019).

K2P (Kimura 2-parameter model (Kimura, 1980)) is a numeric value representing an estimate of average genetic divergence between DNA sequences (Kimura, 1980). K2P distances generated in this study for both the 16S and COI datasets for the 124 sequences, all previously used in the phylogenetic analysis by (Church et al., 2025), except those generated in this study for P. mikazuki from Sendai and P. utriculus from Okinawa, support significant genetic separation of P. mikazuki from the other Physalia lineages (Figure 12B) corroborating preliminary findings by Church et al., 2025, who suggested their unidentified “clade B2” was a new Physalia species. Genetic divergence calculated (mean K2P) among Physalia lineages for both the 16S rRNA gene (2.7% - 4.8%) and mtCOI (8.2% – 13.8%) are comparable to respective divergence distances reported among species of hydrozoan clades (as reviewed in (Church et al., 2025). As reported two decades ago by (Schuchert, 2005), in Hydrozoa both inter- and intraspecific variability is common. Ultimately, a combination of molecular and morphological analyses is needed especially when sympatric speciation occurs, as is likely the case with P. mikazuki and P. utriculus in Japan, whose morphological and molecular characteristics clearly differ, despite overlapping in the Kanto region (Church et al., 2024; Church et al., 2025),used a multi-genomics study to validate the presence of five separate Physalia species, one corresponding to P. mikazuki, the new species we described herein, thereby uncovering hidden diversity in this genus.

The introduction of P. mikazuki into the Tohoku region raises important ecological and public health concerns. As a new predator, it has the potential to disrupt local food webs by preying on fish eggs, larvae, and small planktonic organisms, potentially triggering cascading effects on native species and altering ecosystem dynamics (Purcell, 1984; Purcell, 1985). Competition with native predators could further impact fish populations of commercial and ecological importance (Brodeur et al., 1987; Haddad et al., 2002). Additionally, the presence of P. mikazuki sp. nov. in northeast waters poses public health risks, as stings from Physalia tentacles are known to cause severe pain and systemic reactions (Williamson et al., 1996; Burnett, 2001; Haddad et al., 2002; Yanagihara et al., 2002). These stings present a threat to beachgoers, swimmers, and fishers. Public awareness campaigns, monitoring systems, and emergency response protocols will be essential for safeguarding human activities in affected regions. The cnidome of Physalia physalis (previously considered a cosmopolitan species) has been described as comprising two sizes of holotrichous isorhizas and stenoteles, with fine details varying across time and geographic space, such as holotrichous isorhizas noted instead of atrichous isorhizas and stenoteles reported rather than euryteles (reviewed in (Yanagihara et al., 2002)). However, Yanagihara et al., 2002 reported electron microscopy evidence for the presence of anisorhizas (heteronemes) rather than isorhizas (haplonemes) in the fishing tentacles of P. utriculus of Hawaii, which further underscores the need for a comparative analysis of the cnidome among all species of this genus (Yanagihara et al., 2002).

Historical records from global expeditions conducted during the period spanning the 1700s to 1900s report envenomation symptoms following human contact with a Physalia in the water or stranded on the beach: Sloane (1707) mentioned that P. physalis in Jamaica (as Urtica Marina, soluta, purpurea, oblonga, cirrhis longissimus – Figure 7A) caused stronger burning sensation than those farther north (Sloane, 1707), and Lesueur and Petit, 1807 referred to P. utriculus as a “treacherous animal” (Figure 8A) (Lesueur and Petit, 1807). Duperrey (1830, Chapter XV, pg. 17-35) discussed at length the natural history, behavior, and prey items of different species of Physalia around the world (Figures 1A, B) and the infliction of their venom on humans and animals (based on both natural and experimental envenomation) which include: “excruciating pain causing convulsions throughout the body”, “prolonged contact resulting in raised lesions, fever, syncope (loss of consciousness) and delirium”, “the pain of the sting to the armpit spread to the heart causing a fainting spell”, “indigenous people desiccated and pulverized the tissue into an active poisonous powder mixed into food (e.g., added to chocolate by the Spanish)”, “ants, and fly larvae are known to eat them, but when a dog was fed one it responded with pain that went away in a few hours.” (Duperrey, 1830). Duperrey, 1830 provided detailed accounts of Physalia behavior, ecology, and severe symptoms from envenomation, highlighting the species’ potent venom capable of causing convulsions, delirium, and syncope (Duperrey, 1830). Indigenous practices included using Physalia tissues as toxic additives in food, further emphasizing the severe biological impacts of these organisms.

Current Physalia systematics remain minimalistic due to historical lack of any type specimens, except for P. minuta and P. mikazuki sp. nov. Resolving these taxonomic uncertainties requires further investigation. Thus, herein we provide type vouchers for our new species in the Tohoku Natural History Museum to stabilize the name. Several Physalia binomials are in current usage including P. megalista (Bourg et al., 2024) despite being synonymized with P. australis (Duperrey, 1830) which likely corresponds to a separate valid species although authorship is unclear. Additional names in use, such as P. antarctica Lesson 1826, P. tuberculosaLamarck, 1801, and P. azoricum Lesson 1826 are among species that may be junior synonyms of other Atlantic species, but much work is needed to test this hypothesis (Bigelow, 1911). Figure 8 in this study shows a plate reproduced from de La Martinière (1787) of which drawings (A) & (B) were referred to when describing the new species P. utriculus (as Medusa utriculus) from the North Pacific Ocean (20°N 179°E) (de La Martinière, 1787). Described as a wine-sac shaped organism bearing a large central tentacle and associated with blue sea dragon nudibranchs of the genus Glaucus (C) & (D) which may be the first representation of this kleptoparasitic relationship in which the small blue sea dragon floats on Physalia while nibbling and storing stolen nematocysts in its cnidosacs for personal protection and predation (Yamamoto et al., 2025). Although P. utriculus has been synonymized with P. physalis, its distinct central tentacle clearly detailed in the original description (Totton and Mackie, 1960) (Figures 1A, B, herein from the original report) distinguishes it from P. physalis which has multiple central tentacles. Although no type specimen has been designated for P. utriculus, the name has been used in several works (Yanagihara et al., 2002; Pontin and Cruickshank, 2012; Guevara et al., 2017; Pugh, 2019; Oguchi et al., 2024; Church et al., 2025). Based on the detailed illustration provided in Plate II, Figures 13 & 14 from de La Martinière P. utriculus resembles P. cf. australis (in Figure 1 Plate 5,1 (Duperrey, 1830)) suggesting Hawaii and Australia may share this species. According to our findings and those of Church et al., 2025, specimens from Okinawa belong to Cluster B1, which includes individuals from Hawaii, Guam, and the Indian Ocean, and is referred to as P. utriculus (Church et al., 2025). They also report a second lineage, Cluster B2 (Honshu), comprising specimens from Japan, Pakistan, and Mexico, for which no species name was assigned due to a lack of morphological vouchers (Church et al., 2025). Cluster B2 is P. mikazuki sp. nov. described herein.

This study validates Physalia mikazuki sp. nov. from Gamo Beach, Sendai City, Miyagi Prefecture, Tohoku region as a species new to science. It is distinguishable based on unique morphological features, and the findings of a recently published multi-genomic study and population structure analyses that supported reciprocal monophyly of five Physalia lineages (see (Church et al., 2024; Church et al., 2025)), corroborated by our robust molecular phylogenetic analyses using two mitochondrial gene markers and additional Physalia samples from Okinawa and Tohoku regions.

In this study we established specimen type and paratype vouchers for P. mikazuki sp. nov. and museum vouchers for P. utriculus from Okinawa. Additionally, we presented replicated illustrations and translated original foreign-language descriptions corresponding to P. physalis, P. utriculus, and P. megalista in the absence of museum type specimens to provide a clear reference to assist the scientific community in establishing neotypes. Going forward, establishing high-quality type material verified with molecular and morphological methods, and photographs for all Physalia taxa is essential for stabilizing the taxonomy of the genus.