The ribbon worms depicted in Figure 1 are members of the palaeonemertean genus Balionemertes, which were collected and observed independently by us. These specimens are the basis of our previous publications regarding geographical occurrence (Baker et al. Reference Baker, Falconer, Kinasz, Futterer, Crawford, Puxley, Baker-Johnson, Ruxton, Newton, Hart, Moffat, Strous, Sutcliffe, Butvilla, Santiago, Baade and Aston2015, Chernyshev Reference Chernyshev, Adrianov and Lutaenko2016), molecular phylogeny (Kvist et al. Reference Kvist, Chernyshev and Giribet2015, Chernyshev and Polyakova Reference Chernyshev and Polyakova2021), and morphological investigations (Chernyshev Reference Chernyshev2015, Magarlamov et al. Reference Magarlamov, Chernyshev and Turbeville2018, Reference Magarlamov, Turbeville and Chernyshev2021, Chernyshev and Kajihara Reference Chernyshev and Kajihara2019, Kajihara Reference Kajihara2020).

Figure 1. Unidentified Balionemertes spp. A–C, among the seagrass Amphibolis antarctica collected at a depth of 2 m in Mozzie Flat, Yorke Peninsula, South Australia, 35.1130°S, 137.4970°E, 17 April 2014, by the South Australian Conservation Research Divers (see Baker et al. Reference Baker, Falconer, Kinasz, Futterer, Crawford, Puxley, Baker-Johnson, Ruxton, Newton, Hart, Moffat, Strous, Sutcliffe, Butvilla, Santiago, Baade and Aston2015) (photographs taken by Leon Altoff); anterior fragment, 55 mm in length. D, E, intertidal, among calcareous red algae, Van Phong Bay, Vietnam, 12.66°N, 109.33°E, 21 May 2010, by Alexei V. Chernyshev. F–H, among calcareous red algae at a depth of 1 m in Biti Bay, Guam, 6 May 2015, by Alexei V. Chernyshev. I, J, intertidal, under rock, Buente, Carmen, Cebu, Philippines, 10.6208°N, 124.0277°E, 22 September 2006, by Hiroshi Kajihara; anterior fragment, 3 cm in length. A, B, E, F, G, I, ventral view; C, E, H, I, dorsal view; A–F, I, J, living state; G, H, fixed state; arrowheads indicating the position of the mouth. Abbreviations: doc, dorsal ocellus; loc, lateral ocellus; voc, ventral ocellus.

Little is known about the biology of Balionemertes due to their rare occurrence. Two decades ago, the genus Balionemertes was established for the single species Balionemertes australiensis based on material from the Great Barrier Reef, Australia (Sundberg et al. Reference Sundberg, Gibson and Olsson2003) but since then only a handful of occurrences have been reported. In Australia, Balionemertes ribbon worms were subsequently found in Western Australia (Whisson Reference Whisson2023), South Australia (Baker et al. Reference Baker, Falconer, Kinasz, Futterer, Crawford, Puxley, Baker-Johnson, Ruxton, Newton, Hart, Moffat, Strous, Sutcliffe, Butvilla, Santiago, Baade and Aston2015) (Fig. 1a–c), and New South Wales (Riek Reference Riek2022). Apart from Australia, they have been reported from India (Shrinivaasu et al. Reference Shrinivaasu, Venkataraman, Venkatraman, Venkataraman, Raghunathan, Mondal and Raghuraman2015), Vietnam (Chernyshev Reference Chernyshev, Adrianov and Lutaenko2016) (Fig. 1d, e), the Philippines (Kajihara Reference Kajihara2020) (Fig. 1i, j), and Guam (Magarlamov et al. Reference Magarlamov, Chernyshev and Turbeville2018) (Fig. 1f–h); they are also distributed in the Nansei Islands of Japan (Kajihara pers. obs.). At least four different colour morphs are recognised, each probably representing a different species. However, because the colouration of the holotype is not precisely known, species-name allocation to some of these colour morphs is currently problematic, with B. australiensis being the sole valid species name in the genus.

One thing that has not been documented in the existing Balionemertes literature is that the worms crawl and swim upside down. When placed on a Petri dish in seawater, they move like a long, slender pipe fish or a robust horsehair worm, and can swim by wriggling fast from side to side upon stimulus. The mouth is usually tightly closed, thus can be inconspicuous (Fig. 1d), and almost unrecognisable in quick motion; however, it becomes obvious in a narcotised/fixed state due to muscular relaxation/contraction (Fig. 1g), eventually manifesting the worm’s ventral side (the mouth generally opens on the ventral side at a certain distance from the tip of the head in ribbon worms belonging to Palaeonemertea and Pilidiophora). In the living state, the worms often appear to lack a clear sense of up and down, lying on either dorsal or ventral surfaces. They have groups of eyespots on the dorsal, ventral, and lateral surfaces of the head (Fig. 1b, c, g, h), and this might indicate that they can live on both sides. Still, however, the side with the mouth mainly faces upward while the worm crawls. This behavioural dorsoventral inversion is consistent with the difference in the body colour shades. Like many other worm species with colour patterns, members of Balionemertes also have a darker-coloured and/or more intensely patterned behavioural dorsal surface than the other side (Fig. 1a–j). What is peculiar in Balionemertes is that the behavioural dorsal side is the anatomical ventral side.

Animals in a wide variety of taxa are known to live with their belly oriented upward, either continually or temporarily. Speaking of aquatic creatures alone, the list can include (not exhaustive): pelagic sea slugs in the genus Glaucus (Mollusca: Gastropoda) (Huang et al. Reference Huang, Chiu, Chang and Lee2017); giant clams in the genus Tridacna (Mollusca: Bivalvia) (Yonge Reference Yonge1975); brine shrimp (Crustacea: Anostraca: Artemia) (Fryer Reference Fryer2006); barnacles (Crustacea: Cirripedia) (Essock-Burns et al. Reference Essock-Burns, Gohad, Orihuela, Mount, Spillmann, Wahl and Rittschof2017); marine-cave-dwelling remipedians (Crustacea: Remipedia) (Koenemann et al. Reference Koenemann, Schram, Iliffe, Hinderstein and Bloechl2007); the anchialine shrimp Procaris ascensionis (Crustacea: Decapoda) (Abele and Felgenhauer Reference Abele and Felgenhauer1985); backswimmers (Hemiptera: Notonectidae) (Gittelman Reference Gittelman1976); the upside-down catfish Synodontis (Actinopterygii: Mochokidae) (Blake and Chan Reference Blake and Chan2007); and the upside-down pipefish Heraldia nocturna (Actinopterygii: Syngnathidae) (Paxton Reference Paxton1975). The upside-down jellyfish Cassiopeia are excluded since their body plan does not have a dorsoventral axis, as they are not bilaterally symmetrical. All these animals in the list are either neustons, nektons, or sessile organisms. So far as we know, there are no benthic, non-sessile animals that can be included in the list other than Balionemertes and another species of ribbon worm, Cephalothrix suni (Chernyshev pers. obs.). In the latter species, only the behavioural dorsal side (= anatomical ventral side) is pigmented (Chernyshev and Polyakova Reference Chernyshev and Polyakova2021). We suppose another congener, Cephalothrix queenslandica, may also crawl with its ventral side up; the dorsoventral orientation was likely misidentified in the original description of the species (Sundberg et al. Reference Sundberg, Gibson and Olsson2003).

Dorsoventral body-axis inversion is believed to have happened in ancestral chordates, which were likely free-living (non-sessile) and worm-like (e.g., Satoh et al. Reference Satoh, Rokhsar and Nishikawa2014). To what extent the inversion is comparable between the ancestral chordates and the present Balionemertes is unknown, but it is obviously an interesting example of parallel evolution. The open question is why Balionemertes and C. suni (and possibly C. queenslandica as well) move the way they do. We also have no idea why their geographic distribution is limited to the tropical and subtropical Indo-West Pacific. At the moment we can only speculate that their behaviour may be related to the way they hunt their prey, which should be observed in future studies.