Chun (1900) provided the first description of a very large larvacean from specimens collected during the Valdivia expedition (1898–1899). He named the new species Bathochordaeus charon after the mythical figure who ferries the souls of the dead across the river Styx. During the Valdivia expedition two individuals were collected from the South Atlantic and two smaller specimens, from the Indian Ocean. When he studied them later, Lohmann (1914) placed the two largest Valdivia specimens in the family Oikopleuridae. Garstang (1936) later referred to them as “veritable giants among Appendicularians, the depressed body being as large as a walnut and the broad tail almost 3 ins. in length” and his comment may be responsible for them being referred to today simply as “giant larvaceans”. The two smaller Indian Ocean specimens were not described until much later (Lohmann 1931). Not only were the Indian Ocean specimens smaller at <20 mm total length, they appeared to differ appreciably from the larger specimens collected in the Benguela Current (Lohmann 1914; Lohmann 1931; Garstang 1936). When Garstang (1936, 1937) collected two specimens of Bathochordaeus in very good condition from surface waters near Bermuda, he at once observed that they differed appreciably from Chun’s. Garstang’s specimens were smaller in size than the two collected by Chun in the South Atlantic but similar in size to the larger of the two collected in the Indian Ocean. However, three features struck Garstang (1936, 1937) as different: the Bermuda specimens lacked the prominent “obconical gill-pouches” (spiracles), and the crop-like esophageal expansion of Bathochordaeus charon and they possessed an oikoplastic region that was more comparable to other oikopleurids than B. charon as described by Chun (1900) and Lohmann (1914).

Consequently, Garstang wondered if the original description was hampered by misinterpretation or poor preservation (Garstang 1936, 1937). He could not fathom a purpose for either the seemingly counter-productive funnel-shape of the spiracles or the improbable and capacious esophageal expansion. Having no access to the original specimens, and unable to account for the absence of these conspicuous features, Garstang somewhat reluctantly described a new species: Bathochordaeus stygius, and greatly expanded the slowly growing body of work surrounding these enigmatic animals (Garstang 1937).

Since its description, Bathochordaeus charon has appeared in the literature a few times (Thompson 1948; Bückmann and Kapp 1973; Barham 1979; Galt 1979; Castellanos et al. 2009; Lindsay et al. 2015), without specimen collections to accompany them. The lack of specimens combined with Garstang’s (1937) concerns about the characteristic features in the original description, have cast doubt on the legitimacy of B. charon (Chun) as a species distinct from B. stygius (Garstang). The first probable record of B. charon since Chun (1900) appeared in 1948, from the Pacific Ocean off Australia, although it was a single, small (trunk 3.2 mm, tail 7.5 mm) specimen (Thompson 1948). Time passed and the lack of specimens caused Fenaux (1966) to synonymize the two species. Subsequently, the second potential specimen(s) of B. charon were collected by Galt (1979): five animals acquired during three cruises off southern California, that he called B. charon. Unfortunately, those specimens are no longer available and Galt may have used the name B. charon for all his specimens in lieu of B. stygius, since the latter was suggested to him as applying to juvenile specimens (Bückmann and Kapp 1975; Galt 1979 and pers. comm.) and Galt’s specimens were on the order of 3–6 cm in total length. In his description of B. charon Galt wrote “The present specimens conform generally to published accounts of B. charon, detailed descriptions of which were given by Chun (1900), Lohmann (1931)), Garstang (1937 as B. stygius),” indicating that Galt likely deferred to Fenaux’s synonomy of the species in referring to them all as B. charon. More recently, a specimen called B. charon was collected by Castellanos et al. (2009)) but no description was provided and, although a photograph was included neither the large spiracles nor esophageal expansion are visible. Lindsay et al, (Lindsay et al. 2015) provide an in situ ROV image of the house of “Bathochordaeus sp. A” observed off the Nansei Island chain of Japan, but the structure of the inner filter differs markedly from the B. charon and B. stygius we have observed, and that larvacean was not collected.

Giant larvaceans, like other species of larvaceans, use their oikoplastic cells to secrete complex filters or ‘houses’ that allow them to concentrate and feed on particles (Lohmann 1933; Alldredge 1977; Morris and Deibel 1993; Flood et al. 1998). A house consists of a large, diaphanous outer structure as well as a smaller, more convoluted and bi-lobed inner structure that functions as a filter. Together these serve to concentrate appropriately-sized food particles from the surrounding water. The outer part of the house excludes larger material that would clog the inner filter. Thus, the outer structure often acquires a covering of marine ‘snow’ that can alter its size and shape (Hamner and Robison 1992; Silver et al. 1998). The inner filter concentrates food particles of ingestible size and are ultimately connected to the animal’s mouth via a tube made of the same material as the rest of the structure. However, the inner filter is less diaphanous and more stereotypical in shape, often retaining that shape long after the animal has left its house (Robison et al. 2005).

In situ, larvaceans in the genus Bathochordaeus are often visible from several meters away because their houses may span a meter in longest dimension (Hamner and Robison 1992; Robison et al. 2005). When Barham (1979) made the first observations of the occupied houses of large larvaceans he was diving by bathyscaph and saucer in the Pacific Ocean, off southern California and Mexico. He called them “giant” or “large” larvacean houses and inferred they were probably “Bathochordaeus charon”. However, he recalled, “seeing at least five types of large larvacean houses” and gave “general descriptions of three types”. It is not clear he thought all types belonged to B. charon. One animal was collected, identified by Donald P. Abbott as a “larvacean” and the specimen was subsequently lost. Intraspecific differences in Bathochordaeus’ house structure were not known at the time. Resolving the shape of the spiracles and esophagus of Bathochordaeus spp. from the portholes of his submersibles seems unlikely, since the larvaceans themselves were often invisible. Barham’s dives occurred 10 years prior to his publication and it seems more likely that he used the name B. charon because Fenaux (1966) had just synonomized the species and it was the correct name to use.

Almost 30 years after he synonymized the two species, Fenaux apparently acquired specimens collected by manned submersibles that proved to him the validity of both species (Fenaux 1993, 1998). Regrettably, he never published that proof and in the current literature it remains unclear if more than one species of Bathochordaeus exists (Hopcroft 2005; Flood 2005).

Using remotely operated vehicles (ROVs) we have carefully observed and collected B. charon (Chun) as well as B. stygius (Garstang). A combination of morphological features, house structures and molecular evidence clearly distinguish the two species and provide the first records of B. charon from Monterey Bay as well as off the coast of Oregon, expanding its range into the eastern North Pacific Ocean (Fig. 1).

Specimens were collected with MBARI ROVs (remotely operated vehicles) using either detritus samplers or gentle suction (Robison 1995), and preserved in 5 % formalin buffered with sodium tetraborate. The measurements presented here were made on preserved specimens.

Prior to microscopy, specimens were rinsed in seawater then exposed to DAPI (4′,6-diamidino-2-phenylindole) for 2 min, then rinsed again before observation. The oikoplastic cells that generate the mucus feeding filters of Bathochordaeus are many times polyploid (Flood 2005). Since DAPI binds strongly to DNA and fluoresces brightly when illuminated with UV light (Russell et al. 1975), the preparation enhances contrast in the oikoplastic region of these animals. The oikoplastic region of Bathochordaeus is a monolayer of cells that undulate over the trunk – it is not flat. Wanting to preserve the animal we had for a type-specimen, we chose not to dissect it. Instead, we took images at 12 focal planes, then stacked them together using Adobe Photoshop CS 6. Micrographs were taken with an Olympus DP-71 camera, mounted to a Nikon SMZU dissecting microscope.

For molecular work, animals were frozen in liquid nitrogen and stored at −80 ° C until extraction. DNA samples were extracted from tissue using the DNeasy® Kit (Qiagen, Valencia, CA USA) according to the manufacturer's instructions. Traditional DNA barcoding primers such as those for cyctochrome c oxidase 1 (Folmer et al. 1994), 12S, cytochrome b, and H3 did not amplify genes for Bathochordaeus spp. Instead we selected two primers used by Hirose and Hirose (2009), that were constructed specifically for tunicates, doliolids, salps and larvaceans:

5′ – CATTTWTTTTGATTWTTTRGWCATCCNGA–3′ (UroCox1-244 F)

5′ – GCWCYTATWSWWAAWACATAATGAAARTG–3′ (UroCox1-387R).

These primers amplified a 400-base pair section of the mitochondrial cytochrome c oxidase subunit I gene (COI) with the following PCR parameters: 35 cycles of 94 ° C for 1 min, 40 ° C for 1 min, 72 ° C for 1 min. Amplification of an 1800-base-pair fragment of small subunit ribosomal DNA (18S) was conducted using the modified universal primers mitchA and mitchB from Medlin et al. (1988) with the following PCR parameters: 4 cycles of 94 ° C for 1 min, 58 ° C for 1 min stepping 0.1 ° C/second to 72 ° C, 72 ° C for 2 min; followed by 29 cycles of 94 ° C for 1 min, 64 ° C for 1 min, 72 ° C for 1:30 min (Medlin et al. 1988). All products were bi-directionally sequenced using BigDye® Terminator v3.1 Cycle Sequencing Kit on an ABI 3100 or ABI3500 sequencer (Applied Biosystems, Foster City, CA USA). Sequences were edited and aligned using Geneious version 6.0.5 created by Biomatters (Auckland, New Zealand, http://www.geneious.com/), and submitted to GenBank (Accession Numbers: KT881543–KT881545). Genetic distance was calculated using p-distance model in MEGA v6.0 software (Tamura et al. 2013).

High definition (HD) video sequences of larvaceans taken by MBARI ROVs were used to compare the feeding structures (‘houses’) of B. charon and B. stygius as well as the morphology of the larvaceans occupying the houses.

Given that giant larvaceans may contribute up to one third of the vertical carbon flux to the deep seafloor in Monterey Bay (Robison et al. 2005) and that active houses are oases for commensal zooplankton in the mesopelagic habitat (Steinberg et al. 1994), resolving their identities is of ecological as well as taxonomic significance.

Nearly transparent and easily damaged, giant larvaceans are as enigmatic to look at under a microscope as are their descriptions in the literature. Their mouth opens dorsally while their rectum voids ventrally. Many of their structural features are so transparent as to be difficult to see, even under a microscope.

Two of Chun’s most contested features of B. charon remain its odd, crop-like expansion of the esophagus, and its large, funnel-shaped spiracles constricted where they meet the pharynx. However, Chun’s illustration is entirely accurate with regard to both features (Figs. 1, 2 and 3). The simplicity of the inner filter of B. charon almost certainly affects their ability to feed on particles. There are clearly more interstices in the inner filter of B. stygius (Fig. 6), which seem likely to be a consequence of B. stygius’ much larger bands of anterior Fol’s cells (Fig. 5). And the twin supply passages to the inner filter, directly downstream from the tail chamber (tc) of B. stygius (Hamner and Robison 1992) appear conspicuously absent from the inner filter of B. charon (Fig. 6). If truly absent, it is a mystery how water is pumped through the inner filter. If present, the structure is so diminished as to be invisible in the video sequences we presently have for B. charon. Relative to body size, there is decreased surface area of their inner filter compared to that of B. stygius, and the seeming lack of split supply passages and the fragile, diaphanous structure suggests that the inner filter of B. charon may not generate or be able to accommodate the flow and pressure that the inner filter of B. stygius can sustain.

If true, this difference in flow through the house may provide some explanation for the “peculiar” triangular shape of the spiracles that so puzzled Garstang (1937). Larvacean spiracles draw food through the feeding tube where the bi-lobed inner filter joins, and into the mouth (Deibel 1998). The large and densely ciliated external opening of the spiracles of B. charon could create a venturi that would help to pull water into the mouth (Fig. 3). Because the entry to the spiracles of B. charon is close to an order of magnitude smaller than the diameter of their mouth (Fig. 3), evacuating undesirable objects through their spiracles as do other oikopleurids (Alldredge 1977; Lombard et al. 2011), is not an option. If a particle enters the mouth, the only way out seems to be through the gut to the rectum.

A third point of confusion in the early literature came from Chun’s assertion that B. charon had four pairs of Fol’s crescents, when in fact there are two pairs. Lohmann (1914) was also criticized by Garstang (1937) with regard to his description of the large and fairly conspicuous Fol’s cells in B. charon. Garstang called these “glandular crescents” and because the anterior pair in B. stygius consists of eight giant cells, it was clear to him that those were “…identical with Fol’s oikoplast in Oikopluera, Stegosoma, and Megalocercus both in structure and in position” (Garstang 1937). However, because Lohmann (1914) counted 12 giant Fol’s cells in B. charon and asserted that only one pair of “glandular crescents” was visible, Garstang thought Lohmann had mistaken the posterior giant cells (which were obvious in the specimens of B. stygius he collected) for Fol’s cells and somehow missed the anterior crescent entirely. Although convinced that Lohmann had erred, Garstang proposed to call the larger, posterior crescent (which he maintained were not Fol’s cells) ‘Lohmann’s colloplasts’, in honor of Lohmann’s lifetime of work with larvaceans. This terminology is ironic because in this regard, Garstang was incorrect.

Bathochordaeus charon clearly has one band of 12 giant cells on either side of the trunk as does B. stygius. But the anterior pair of 8 cells in B. charon are so small in comparison to those of B. stygius (which are comparable in size to the cells of the posterior band) as to appear missing. This difference in the anterior band of Fol’s cells is probably why the inner filter of B. charon and B. stygius look so different (Fig. 6). Furthermore, we have observed the initial formation of many houses of B. stygius and both pairs of “glandular crescents”, on either side of the trunk, make the inner filter. Therefore, we regard all of Garstang’s “glandular crescents” as Fol’s cells and distinguish each band based on its position on the trunk (i.e., anterior or posterior).

In over two decades of exploring the mesopelagic waters of Monterey Bay, CA, remarkably few Bathochordaeus charon (n = 15) were observed, while thousands of B. stygius have been encountered. Both species occupy a similar depth range in Monterey Bay, yet, based on the large differences in COI they are clearly not interbreeding populations. Moving forward we intend to look more closely at the diversity of larvaceans in the mesopelagic habitat, identifying these important organisms responsible for vertical transport of carbon through the water column. As of 2009, the recognized number of appendicularian species was 70, of which 43 are described for the Pacific Ocean (Fenaux et al. 1998; Castellanos et al. 2009), Bathochordaeus charon unequivocally among them.

#, number; br, brain; cf, ciliated funnel; COI, cytochrome oxidase 1; en, endostyle; es, esophagus; et al., et alia, and others; fa, anterior cells of Fol’s oikoplast; fp, posterior cells of Fol’s oikoplast; HD, high-definition; i.e., id est, that is; isc, inner sensory cells; m, mouth; MBARI, Monterey Bay Aquarium Research Institute; N/A, not applicable; p, pharynx; pb, peripharyngeal band; r, rectum; ROV, remotely operated vehicle; sp, supply passages; st, left lobe of stomach; t, testis; VARS, video annotation and reference system