Redescription of Gnorimosphaeroma oregonense (Dana, 1853) (Crustacea, Isopoda, Sphaeromatidae), designation of neotype, and 16S-rDNA molecular phylogeny of the north-eastern Pacific species

Abstract Gnorimosphaeroma oregonense (Dana, 1852) is revised, a male neotype is designated, photographed, and illustrated; the species occurs from Vancouver British Columbia to the central California coast. 16S-rDNA sequences (~650 bp) for all available ethanol preserved species of Gnorimosphaeroma were used to hypothesize their relationships. Our analyses revealed a sister taxon relationship between the fully marine G. oregonense and the brackish to freshwater species, G. noblei. The oyster associated and introduced G. rayi is sister to a previously not recognized or identified, but genetically distinct, Gnorimosphaeroma sp. collected at two sites in San Francisco Bay. Gnorimosphaeroma sp. is probably also a western Pacific species based on its genetic relationship to G. rayi. Photographic comparisons are offered for G. oregonense (marine), G. noblei (freshwater), G. rayi (introduced), G. sp. (presumably introduced), and G. insulare (San Nicolas Island). Records of the holdings at the Los Angeles County Museum of Natural History are summarized. Without material available north of Vancouver through Alaska, the range of G. oregonense could not be genetically verified. This review includes a diagnosis and description of the genus Gnorimosphaeroma Menzies, 1954.


Introduction
. Museum collections examined morphologically and not included in the molecular analyses. Taxa are grouped by species and sorted by latitude. Label data and associated notes are transcribed here. Note that in some instances latitude and longitude are approximate and are indicated as "~". Although we attempted to extract and amplify DNA, some were unsuccessful.

Species Specimen label Gnorimosphaeroma oregonense
Oregon, Coos County, Squaw Island, off Cape Argo Light, 43.339°N; 124.376°W, intertidal, -1.6 ft. tide, rocky reef, some loose rocks kelp covered, preserved in 95% ethanol, 27 Jul 1942, Sta. 1488 Table 2. Sequences used in the 16S-rDNA analyses are associated with their taxon names in alphabetical order and GenBank accession number. The molecular identification number identifies the specimen on the phylogenetic tree. In several instances multiple individuals were extracted and sequenced from the specimen lot. An asterisk denotes the lot from which neotype was selected.
Description. Body vaulted, dorsal surfaces smooth or polished in appearance, without setae; coxal and other margins smooth, with ability to conglobate; not or weakly sexually dimorphic. Head with rostral point present, dorsally visible, simple, not separating antennular bases; without paired incisions in front of eyes, lateral margins not laterally extended to body outline (antennules more or less ventral). Eyes lateral, simple. Pereonite 1 lateral margins not anteriorly produced, not laterally enclosing head, pereonites 2-7 with posterior margin not raised, pereonite 1 anteriorly with keys. Sternite 1 without cuticular mesial extensions. Pereonite 6 simple, without bosses, processes or marginal extensions. Pereonite 7 as wide as pereonite 6, forming part of body outline, dorsally without bosses, processes, or marginal extensions. Coxae distally narrow, those of pereonites 2-7 overlapping the one behind, rounded, with ventral 'lock and key' processes, with grooved articulation; those of pereonite 6 not large, not overlapping those of pereonite 7. Pleon consisting of 4 visible segments (as determined by lateral sutures); pleonite 1 entire, posterior margin even, narrower than remainder of pleon, not extending to pleon lateral margins; sutures (except first) running to lateral margin, all separate, sutures long (separated medially by 24-28% pleon width); pleonal sternite absent; dorsal surface without process; posterior margin even, with 'keys'. Pleonite 5 posterior margin entire (not fused with pleotelson). Pleotelson vaulted, anteriorly as wide as pleon, without dorsal process; posterior margin entire, simple, arcuate; ventrolateral margins forming ridge.
Antennule peduncle with basal articles medially not in contact, 1 and 2 robust, article 3 slender; article 1 not produced, without anterior lobe; article 2 approximately 0.5 as long as article 1; with articles 2 and 3 colinear, article 3 longer than article 2; article(s) not flattened; flagellum shorter than peduncle, longer than peduncular article 3. Antenna peduncle articles all colinear (or curving regularly), less robust than antennule, peduncular articles all of similar thickness.
Uropod rami not strongly flattened, not forming part of continuous body outline; exopod shorter in length than endopod, exopod lamellar, inserted near anterolateral angle of peduncle-endopod, lateral margin simple, finely serrate or smooth, distally broadly rounded; endopod lamellar, distally broadly rounded or narrowly rounded. Uropod endopods not in contact posteriorly.
Remarks. Gnorimosphaeroma is in a general sense quite unremarkable in appearance, with no species showing any sort of dorsal ornamentation of tubercles, processes, or pereonal and pleonal ridges that characterize so many genera of Sphaeromatidae. As such, there is a lack of readily obvious characters by which to identify the genus. Gnorimosphaeroma, on morphological criteria, is most similar to the genera Bilistra Sket & Bruce, 2004, Exosphaeroma Stebbing, 1900, Lekanesphaera Verhoeff, 1943, Neosphaeroma Baker, 1926and Sphaeroma Bosc, 1802. The latter three genera can be differentiated from Gnorimosphaeroma in the first instance by having the uropodal exopod lateral margin with one or more serrations or notches (among other characters).
Exosphaeroma is a large genus with 40 species at the last count (Boyko et al. 2008) that, as presently constituted, contains both smooth bodied species as well as some with coarsely pitted or ridged dorsal surfaces (e.g., see Kensley 1978;Espinosa-Pérez and Hendrickx 2001;Bruce 2003), and also species with greatly enlarged uropodal rami (e.g., see Kensley 1978;Bruce 2003;Wall et al. 2015). Some of the smoothbodied species of Exosphaeroma are superficially similar to Gnorimosphaeroma, but can be distinguished by the pleonal sutures running to the posterior margin (to the free lateral margin in Gnorimosphaeroma), as well as pleonite 1 having two flat sub-median lobes on the posterior margin (see Bruce 2003: figs 14E, 18F).
Bilistra is similar in gross morphology and also occupies coastal freshwater habitats. Bilistra differs from Gnorimosphaeroma in having a far shorter uropodal exopod (ca. half as long as endopod), shorter pleonal sutures that run to the pleon posterior margin (not lateral margin); the inferior margins of pereopods ischium or merus to propodus have a dense setulose (fur-like) fringe while the superior margins lack long setae altogether. Bilistra is presently restricted to New Zealand, but there is also one species in South Africa, from supralittoral brackish pools and tidal streams that is currently classified as Pseudosphaeroma barnardi Monod, 1931 that is in need of redescription and formal reassignment to Bilistra (NLB, pers. obs.).
Gnorimosphaeroma pereopod setation is inconsistently illustrated, even within species, despite being a potentially significant character. The redescription given here, and figures of Hoestlandt (1975) show long setae on the superior or superodistal margin of the merus and long setae on the inferior margin of the ischium and merus. Such setae were not mentioned or figured in Menzies' (1954) genus diagnosis or species descriptions. Such setae are also apparently absent from all northwestern species (e.g., Hoestlandt 1975Hoestlandt , 1977Kwon and Kim 1985;Nunomura 1998Nunomura , 1999aNunomura , 2007. Neotype designation. It has been long established that all of Dana's (1852) isopod material, and therefore all the type material for the many species of isopod that he named, was lost with the sinking of the ship USS 'Peacock' on the bar of the Columbia River in 1841 (Bruce 1986: 220;2004: 228;2009: 211;Poore and LewTon 1993: 234). Gnorimosphaeroma oregonense (Dana, 1853) is one such species.
Species of Gnorimosphaeroma are uniform in appearance, and to date no assessment has been made of intrinsic variability within species. Some species of Gnorimosphaeroma occur sympatrically and there are many exceedingly similar species. At present few species have been described in full detail. Furthermore, records of G. oregonense are somewhat inconsistent in the details presented and the material is not always available for re-examination, so that it is not always possible to confirm the correct identity of previous records and indeed also on occasion, new material. We consider that designating a neotype is necessary to clearly characterize the identity of this species, to allow for the genus to be precisely diagnosed based on the type species and to permit unambiguous identification and separation from other sympatric congeneric species. Dana (1853) did not indicate a specific type locality, but stated that the species had been obtained from "Puget's Sound, Oregon; also, Bay of San Francisco, California". One may infer that the first mentioned location is the type locality but that remains an inference, and furthermore one cannot be certain that the material consists of only one species, given that there are four species in the region and also that the morphology of purported species apparently changes from low to high latitudes (present study). The neotype has been chosen from specimens collected as near as practically possible to the original type locality, and is now Stanley Park, 49.294°N, 123.155°W (British Columbia, Canada), ca. 150 km north of Puget Sound.
Uropod extending to posterior margin of pleotelson. Exopod 0.83 × as long as endopod, 2.7 × as wide; apex narrowly rounded; mesial margin with continuous row of PMS. Endopod 3.8 × as long as wide, lateral margin weakly convex, apex bluntly rounded. Size. Largest ♂ to 8.5 mm, largest ♀ to 6 mm. Dana (1853) gave no measurements. Fee (1926: 8, 9) records the largest specimens as being "ca. 1 cm. long; one-half as long as wide." Color. When preserved in ethanol, specimens quickly become pale buff to whitish. Distribution. British Columbia, Vancouver to California, San Francisco. Remarks. The species occurs only in fully marine habitats in the intertidal to an unknown depth. A single lot indicated that it was collected by night light, and another that specimens were collected on floats among fouling organisms. None of the material examined indicates depth. Kussakin (1979) reported G. oregonense from Alaska, Popov Island to San Francisco Bay, California. Kussakin (1979) figured G. oregonense from the collections of the Zoological Institute of the Academy of Sciences of the USSR. He noted that it is widely distributed with males reaching a length of 12 mm and females up to 8 mm, and that it occurred widely from Alaska to California. It is not clear what the specific localities of the figured specimens were (Kussakin 1979: 407) nor of those deposited in the Russian collections. We were unable to locate and access these specimens. Kussakin reported that the specimens he examined were predominantly littoral, but can be sublittoral to 22 m, on rocks, under rocks, less often on sand, and sometimes in empty shipworm tubes. Kussakin remarked it is a good swimmer, and sometimes turns up in night light samples. It can tolerate salinities as low as 9‰. Since we were not able to re-examine Kussakin's specimens, we cannot verify that the Gnorimosphaeroma he identified are the same species as G. oregonense from the type locality and described here. Furthermore, our genetic data clearly distinguishes between fully marine and low salinity specimens and recognizes these as distinct species (see below). We do not include Kussakin's specimens in the synonymy (Gnorimosphaeroma oregonense: Kussakin, 1979: 406, figs 260-262.)

Molecular analysis
The molecular analyses include G. oregonense from Vancouver and the San Juan Islands, Washington (49. 256°N-48.513°N). There are no specimens north of Vancouver in our collections. Gnorimosphaeroma noblei material came from Del Norte to Santa Barbara Counties (40.833°N-34.46°N), G. rayi from Marin County (38.201°N-37.902°N), and the unidentified Gnorimosphaeroma sp. were collected only in San Francisco Bay (San Mateo and Alameda Counties, latitude 37.079°N-37.535°N). Figure 10 indicates the localities of the sequenced material. Alignment differences resulting from the LINS, EINS, or GINS alignment algorithms had insignificant effect on RAxML and Fasttree analyses and the phylogenetic hypotheses. Ancinus sp. (Sphaeromatoidea: Ancinidae) was used as the outgroup based on the basal position of Gnorimosphaeroma within the Sphaeromatidae (Wetzer et al. 2013). Both analyses resulted in the same 1 tree. Only the RAxML tree (Fig. 11) is shown.
Our molecular analyses (Fig. 11) clearly distinguish G. oregonense and G. noblei. They are always sister taxa. Gnorimosphaeroma rayi, is always sister to an unidentified Gnorimosphaeroma species collected from the two localities in San Francisco Bay.

Gnorimosphaeroma noblei Menzies, 1954
Gnorimosphaeroma noblei Menzies, 1954 was described from the town of Marshall in Tomales Bay, California (~38.162°N, ~122.89°W). Hoestlandt (1969) synonymized G. oregonense lutea with G. noblei. Menzies noted the species was associated with the terrestrial isopod Armadilloniscus in the upper intertidal, and that they were excellent swimmers. This association indicates likely freshwater input and possible lower salinity. This species has the largest range of all of the Gnorimosphaeroma species studied here (California, Del Norte County, ~41.931°N to Los Angeles County, 33.802°N). This species also has the broadest salinity tolerance -brackish to freshwater, a characteristic found in only a few sphaeromatid genera. Gnorimosphaeroma noblei has been collected from a full range of high intertidal, brackish to fully freshwater habitats including Sacramento, central San Joachin Delta, ~38.33°N, ~121.3°W collections by Wayne Fields. Fields' specimens were preserved in formalin and their collection date is unknown. They have been in the LACM collections since before 2003. Specimens of G. noblei can be comparable in size to G. oregonense, but more commonly are slightly smaller. Gnorimosphaeroma noblei is purported to occur as far south as Los Angeles County, Dominguez Channel, 33.802°N, 118.228°W. Three very small specimens from 4 m depth were collected 17 September 2003 (MBPC 10592, Collection ID: RW17.028). These were also preserved in formalin and were unavailable for genetic analysis, but based on all of the other material examined (Table 2) are presumed to be G. noblei.
Their very similar appearance to G. oregonense makes morphological identifications ambiguous, yet genetically they are easy to distinguish from G. oregonense (Figs 11-13, 15). Sequence divergence between the two species for the 16SrDNA fragment sequenced here is 16.5-20.9%. Gnorimosphaeroma noblei is always the sister taxon to G. oregonense in all of our genetic analyses ( Fig. 11; Wetzer et al. 2018).

Gnorimosphaeroma rayi Hoestlandt, 1969
Gnorimosphaeroma rayi arrived in Tomales Bay in 1928 with oysters (Crassostrea gigas now accepted as Magallana gigas (Thunberg, 1793) from Japan (Bonnot 1935;Barrett 1963; James Carlton pers. comm. 2019). The type locality for this species is California, Marin County, Tomales Bay, Shallow Beach, 38.14°N, 122.881°W (Hoestlandt 1969). In addition to Japan, Hoestlandt (1975Hoestlandt ( , 1977 reported this species from eastern Siberia and Hawaii. Hoestlandt too acknowledged the differences between G. rayi and G. oregonense are subtle. Hoestlandt's (1975) key attempts to disambiguate the four species (G. oregonense, G. noblei, G. rayi, and G. insulare). However, we urge caution as his key may only be applicable to the largest specimens of each species, and we were unable to use it consistently.
Based on all of the material in the LACM collections available for genetic analysis, we could only confirm that the species occurs in Tomales Bay (three lots) and one lot from Bolinas Beach. Bolinas Beach is just 43.5 km south of Tomales Bay (Fig. 11, Table 1). It does not appear that this species is broadly distributed or quickly expanding its range (Figs 11-13, 15). A further assessment of its distribution awaits future genetically appropriately collected material and analyses. Additionally, we recognized a previously unidentified Gnorimosphaeroma sp. in San Francisco Bay. Genetically Gnorimosphaeroma sp. and G. rayi are readily distinguished and are always sister taxa in our analyses. However, we were unable to identify any reliable morphological characters to distinguish the two species. Based on the phylogenetic relationship between Gnorimosphaeroma sp. to G. rayi, it is presumed it too has a western Pacific origin.

Gnorimosphaeroma sp.
Morphologically this species cannot be distinguished from G. rayi. However, it is clearly genetically distinct with 13.9-16.5% sequence divergence for the 16S-rDNA fragment that was sequenced. Since we know G. rayi is an introduction from the western Pacific, this species is also likely a trans-Pacific traveler. San Francisco Bay, a biodiversity hotspot, is infamous for non-native and invasive species. At this time, there are no sequences available for western Pacific Gnorimosphaeroma that would allow identification of this species and clarification of their relationships (Figs 11-13). Many western Pacific species are poorly described and in need of redescription, making it impossible at this time to identify these specimens further.

Gnorimosphaeroma insulare (Van Name, 1940)
Gnorimosphaeroma insulare was described from freshwater on San Nicolas Island. San Nicolas is part of the Channel Island Archipelago off the Southern California Coast, today located nearly 100 km from the nearest point on the mainland coast. Menzies' (1954) redescription detailed the G. oregonense distribution and compared G. oregonense to Van Name's (1940) Exosphaeroma insulare, which Menzies moved to Gnorimosphaeroma. Menzies (1954) noted that the largest G. insulare is 8 mm in length. Some confusion then ensues with the description of G. noblei Menzies, 1954. Gnorimosphaeroma noblei is described from Tomales Bay from the high intertidal found in association with Armadilloniscus, hence associated with possible freshwater input and thus lower salinity. To the best of our knowledge, the only material of G. insulare is that from the original collections and type locality on San Nicolas Island. Eleven specimens were collected from freshwater where they were associated with the freshwater gastropod pulmonate, Physa virgata (Gould, 1855). The LACM collections hold a single male syntype which is photographed here (Fig. 14). Additional specimens are at the American Museum of Natural History. The specific collecting locality on San Nicolas is not known and it is unclear if any freshwater still runs today. Accessing this US Navy-controlled island, which is used for weapons testing and training, is difficult.

Discussion
The genus Gnorimosphaeroma is a temperate-water clade occurring only on the shores of the northwestern Pacific (China, Japan), east through Alaska, and along the East Pacific coast to southern California shores. The genus is most speciose in the north western Pacific with 26 described species. Many of these species descriptions are inadequate, in need of critical evaluation, and redescription. In the eastern Pacific, G. oregonense is the most wide-ranging species, apparently occurring from Alaska to San Francisco, California. However, in this study we were only able to verify morphologically and genetically the species occurrences from Vancouver to San Francisco Bay. Adult specimens of G. oregonense become larger and more robust with increasing latitude. Along the Washington to California coast this species commonly co-occurs with Exosphaeroma inornata Dow, 1958. The latter is known from Puget Sound, Washington to central-southern Baja California Norte, Mexico (Wall et al. 2015). When specimens are very small and/or subadults, not only do the species of Gnorimosphaeroma get readily confused, but sometimes they are misidentified as E. inornata if the most careful attention is not paid. Any distinctive color patterns are lost in preserved material. However, we have demonstrated that regardless of their size, species of Gnorimosphaeroma and Exosphaeroma inornata are readily distinguished based on their genetics (Wetzer et al. 2018). For both molecular analyses and morphological study, we had a very restricted distribution of G. oregonense specimens available. Kussakin 1979 had similarly struggled with resolving the identity of Northern Pacific Gnorimosphaeroma. He recognized that the genus contains fully freshwater, brackish, and marine species. Gnorimosphaeroma kurilense (Gurjanova, 1933) occurs in freshwater. It would be informative to be able to compare genetic sequences of our material with specimens from north of Vancouver, British Columbia to Alaska, fully marine to freshwater, and verify that the specimens observed and figured by Kussakin and others are genetically similar. Kussakin's specimen(s) were from Popov Island, Primorsky Krai, near Vladivostok northwestern Pacific. The dorsal view ( fig. 260, pg. 407) does look very similar to the G. oregonense from the eastern Pacific. It possesses the same distinctive rather heavily calcified coxal plates with strong carinae on coxae 2, 3, 5, and 6. Coxa 1 is acute, coxa 2 is subquadrate, with coxae 3-7 becoming more acute posteriorly. Kussakin's figures ( fig. 262) differ from our specimens in that antenna 1 has 15 flagellar articles and antenna 2 has ten flagellar articles compared to most of the specimens which we observed, which have 13 and 14, respectively. Also, the mandible of the Kussakin specimen has three long setae at the base of the molar incisor. Our specimens lack such setae (Fig. 4C, D). Pereopod 1 of the Popov specimen ( fig. 261) appears more setose than the Washington specimen (Fig. 5A).
Since we only had specimens of G. oregonense available from a restricted range of Vancouver to Washington, we cannot assess the genetic diversity across the species' larger range (Figs 10, 11). Our Exosphaeroma amplicauda (Wall et al. 2015) review resulted in the recognition of five species with Alaska and Washington specimens recognized as distinct from the type locality (Central California), and distinct from those from the Southern California coast. It would not be surprising if future studies based on broader sampling revealed greater genetic diversity than we have observed here.
We also had available for study a single male syntype of G. insulare Van Name, 1940 (Fig. 14). As noted previously by Van Name, G. insulare and G. oregonense are very similar. Examination of the specimens we had available, dorsally G. insulare appears oblong and is ca. twice as long as wide compared to G. oregonense, G. noblei, G. rayi, and the unidentified Gnorimosphaeroma sp. from San Francisco Bay. These species are all broader than G. insulare and therefore have a more globular appearance. The largest G. insulare specimen observed was 8 mm in length, whereas the largest known individuals of G. oregonense, G. noblei, and G. rayi had been previously recorded as 12 mm in length. Gnorimosphaeroma insulare is distinguished from all other Gnorimosphaeroma species in that it appears to have been entirely restricted to a freshwater pond and only known from the type locality. It is unknown whether this pond still exists today. Since it is the sole specimen (syntype) and fragile, no dissections were undertaken, but rather the specimen was photographed (Fig. 14). A collection made on San Miguel Island by E. Hochberg and identified by E.W. Iverson was reidentified here as G. noblei. Sadly, no habitat information was provided for this collection and it had been formalin fixed and is not available for genetic study. San Miguel Island is the northernmost of the Channel Islands and 74 miles distant from San Nicolas Island. The only specimens known from offshore islands were these two lots.
Future genetic comparisons of marine, brackish, and freshwater Gnorimosphaeroma species occurring north of Vancouver, through Alaska to Primorsky Krai (northwest Pacific) may reveal either multiple invasions or a single invasion to brackish and freshwater and may change the current phylogenetic relationship of brackish/freshwater species and marine species in the Eastern Pacific. Phylogenetic placement of G. insulare would also be most interesting should populations at this locality still exist today.
Identification keys for west coast Gnorimosphaeroma species are available in Menzies (1954) and Hoestlandt (1975Hoestlandt ( , 1977. Kussakin (1979) provides a key for north Pacific species, Kwon and Kim (1987) for Korean species, and Nunomura (1998) for the Japanese species. Difficulty arises in using them as differences between species can be very subtle and may only apply to very large adult specimens. In some instances, the largest specimens possible for the species may not have been available at the time of description (e.g., G. noblei Hoeslandt, 1975). However, the differences in the lateral and ventral appearance of the coxal plates of G. oregonense and G. noblei are distinct in large adult males and females (Figs 13A, B, 15). In G. oregonense lateral view the anterior margins of coxal plates 2, 3, and 4 are raised, posterior margin not raised, giving coxae a somewhat "s-shaped" appearance ( Fig. 15A, C). Ventrally these appear as interlocking units. Gnorimosphaeroma noblei which can co-occur with G. oregonense lacks these (ventrally coxae not interlocking) (Fig. 15B, D). The two species are also readily distinguished based on habitat and salinity. Gnorimosphaeroma oregonense is always in fully marine waters and G. noblei inhabits high intertidal, brackish to fully freshwaters.
As molecular phylogenetic studies allow more and deeper sampling, cryptic species in marine environments are being recognized with ever greater frequency. Organisms as diverse as foraminiferans (Aurahs et al. 2009), copepods (Bláha et al. 2010), hydroids (Moura et al. 2008), and valviferan isopods (Xavier et al. 2012) are revealing much greater diversity than previously recognized. This diversity is and cannot always be recognized morphologically. The recent detailed study of the sphaeromatid isopod Dynamene by Vieira et al. (2019) demonstrated that not only can large sequence divergences exist over small spatial scales, but that repeated invasions leave their genetic mark on populations, and that population diversification can be recognized over shorter time scales than previously thought for organisms with limited dispersal abilities. Based on the putative cryptic species in their study, they estimate a 300% under-estimation of known species in Dynamene, a species-poor genus. As more Gnorimosphaeroma species and specimens for genetic analysis become available, this genus has the potential to provide interesting insights into not only the evolution of the rare marine to freshwater invasion of species within the genus, but also human induced species relocations across the Pacific Ocean. If G. insulare still exists on San Nicolas Island and possibly on other Channel Islands too, this genus could reveal a very interesting phylo-biogeographic history.