Transoceanic rafting of Bryozoa (Cyclostomata, Cheilostomata, and Ctenostomata) across the North Pacific Ocean on Japanese tsunami marine debris

Forty-nine species of Western Pacific coastal bryozoans were found on 317 objects (originating from the Great East Japan Earthquake and Tsunami of 2011) that drifted across the North Pacific Ocean and landed in the Hawaiian Islands and North America. The most common species were Scruparia ambigua (d’Orbigny, 1841) and Callaetea sp. Of 36 bryozoans identified to species level, 15 are already known from North America, one of which (Schizoporella japonica Ortmann, 1890) is an earlier introduction from Japan; 18 species are known only from the Western Pacific, one of which (Bugula tsunamiensis McCuller, Carlton and Geller, 2018) is newly described in a companion paper. The 13 additional bryozoans, not taken to species level, are likely derived from the Western Pacific based upon evidence reviewed here; two of these species (Callaetea sp. and Arbocuspis sp.) are undescribed. Seven warm-water species, Metroperiella cf. biformis (Zhang and Liu, 1995), Celleporaria brunnea (Hincks, 1884), Drepanophora cf. gutta Tilbrook, Hayward and Gordon, 2001, Smittoidea spinigera (Liu, 1990), Biflustra grandicella (Canu and Bassler, 1929), Biflustra irregulata (Liu, 1991), and Celleporina cf. globosa Liu, 2001, not known from Japan, may have been acquired by Japanese Tsunami Marine Debris (JTMD) as these objects were carried by ocean currents into more southern waters. Three oceanic bryozoans (Jellyella tuberculata (Bosc, 1802), Jellyella eburnea (Hincks, 1891), and Arbopercula angulata (Levinsen, 1909)) provide insight into the routes that some JTMD items may have taken, and thus the conditions experienced, as they rafted from the Western Pacific to the Central and Eastern Pacific. The cooler-water species J. tuberculata and A. angulata were found primarily on JTMD objects arriving in the Pacific Northwest, whereas J. eburnea was most common on objects landing in the Hawaiian Islands. The most common bryozoan growth forms on these rafted objects were runners (creeping uniserial morphology) and arborescent forms capable of using available surface area provided by other organisms (such as hydroids) on space-limited objects. Species that form flat or mounded encrustations were less frequent, suggesting that they do not fare as well in a potentially space-limited environment.


Introduction
Bryozoans are a large and diverse group of invertebrates well represented in many marine habitats, including fouling communities.While the ranges of many bryozoan species are restrained due to their lecithotrophic, short-duration larval stage, and thus low planktonic dispersal ability (Johnson et al. 2012), rafting on natural substrates (such as algae, seagrass, or wood) may explain in part the presumably natural widespread distribution of many species, although the taxonomic validity of many putative cosmopolitan species is being increasingly questioned (Hoare et al. 2001;Harmelin et al. 2012;Vieira et al. 2014a, b).Confounding our understanding of the role of natural rafting are the centuries of anthropogenic dispersal of bryozoans in fouling communities on ships' hulls or by the widespread movement of commercial shellfish (Carlton 2009), both of which activities altered the distribution of many species (Watts et al. 1998;Johnson et al. 2012).Adding to the complexities of dispersal mechanisms in modern times has been the introduction of long-lasting plastic debris into the world's oceans, potentially altering species' ranges by providing non-biodegradable rafting substrates of much greater temporal and spatial duration than floating organic objects (Gregory 2009;Kiessling et al. 2015;Carlton et al. 2017).
Bryozoans are sessile, colonial, filter feeders with many species possessing the ability to persist through adverse conditions; consequently, they frequently represent the most abundant and diverse group within regional rafting communities (Winston 1982a; Barnes and Fraser 2003;Kiessling et al. 2015).Bryozoans are also abundantly found in port, harbor, and vessel fouling communities (Connell 2000;Raveendran and Harada 2001;Yakovis et al. 2008).Fouling species colonizing floating materials may be particularly susceptible to both coastal and transoceanic transport (Watts et al. 1998).
The relative dominance of bryozoans in both rafted and fouling communities may also be due in part to their impressive variety of growth forms, each suited for differing substrate microhabitats, environmental conditions, and competitive strategies (Lidgard 1985;Ward and Thorpe 1989;Hageman et al. 2013).Thus, species exhibiting encrusting, one-dimensional planar growth often quickly exploit open space but are poor competitors, whereas arborescent species may rise above potential competitors but may be limited by their ability to withstand damage due to water motion (McKinney and Jackson 1991).Indeed, neustonic encrusting membraniporine bryozoans are among the first colonists on debris that enters the sea in the open ocean (as, for example, trash discards from ships); in turn, these initially bryozoan-dominated communities may subsequently be dominated by gooseneck barnacles in the genus Lepas (Tsikhon-Lukanina et al. 2001).
Increasing concentrations of marine debris, assumed to largely originate from coastal zones, occur worldwide (Moore 2008;Cózar et al. 2014;Eriksen et al. 2014).Concomitantly, studies of rafted communities on this debris are on the rise but often are limited by a lack of information on where and when this debris entered the ocean (Goldstein et al. 2014;Kiessling et al. 2015).Large pulses of anthropogenic debris introduced into the ocean may occur through natural events such as monsoons, hurricanes, and tsunamis, potentially providing vastly increased habitat for biofouling organisms (Thiel and Haye 2006).The 2011 Tōhoku earthquake and tsunami was one such event in which a large debris field was ejected into the North Pacific Ocean (Carlton et al. 2017).Objects from the tsunami, accompanied by living and reproductively viable marine species from Japan, began washing ashore in 2012 in North America and the Hawaiian Islands.
We report here on Western Pacific bryozoans that arrived in the Central and Eastern Pacific Ocean attached to vessels, crates, buoys, ropes, and other Japanese Tsunami Marine Debris (JTMD) objects.Species arrived both dead and alive (the latter with intact polypides or brown bodies, as well as often being reproductively active with ovicells containing embryos).Unlike many free-living organisms that may have been lost after death, bryozoan skeletons often persist and thus provide greater insight into the potential diversity of rafted species, especially given that only a small fraction of the debris field was intercepted and biologically sampled (Carlton et al. 2017).We thus include species that we believe arrived dead (noting that some taxa that we place in this category may have died only after landing), given the possibility that the same species arrived alive on other debris not seen by us.

Materials and methods
Bryozoan samples were obtained from a wide variety of JTMD objects (identified as such through multiple lines of evidence; see Carlton et al. 2017) landing between 2012 and 2017 in North America and the Hawaiian Islands (Supplementary material Table S1).Each object was assigned a unique identification number preceded by JTMD-BF-(Japanese Tsunami Marine Debris-BioFouling-).
Bryozoan specimens (dried or in 95% ethanol) were found loose in samples, or were removed from their substrate with a scalpel, and placed in voucher collections.Images were taken using a Leica EZ4 HD camera (Leica Microsystems, Wetzlar, Germany) and LAS EZ imaging software (Leica Microsystems, Wetzlar, Germany); measurements of zooidal characters were done using Fiji software (Schindelin et al. 2012).For scanning electron microscopy (SEM), specimens were cleaned in sodium hypochlorite solution, rinsed in tap water, and then placed in ethanol to prevent degradation during transport.Samples were then air dried and coated with Au-Pd using an Anatech USA Hummer 6.6 Sputtering System at 15mA (Anatech, Hayward, California, USA) and viewed under a JEOL JSM-7100FLV field emission scanning electron microscope (JEOL USA Inc., Peabody, Massachusetts, USA) at 5.0kV accelerating voltage.Images were retained as TIFF files.Species selected for SEM imaging included those presenting characters apparently distinct from published descriptions, those not extensively illustrated or well described already in the literature, and those not identified to species.
Voucher specimens were identified as "living" if the majority of zooids contained tissue, if not intact polypides, and note was made of any embryos present (suggesting reproductive activity at the time of death or sample preservation).Degenerated zooids (containing brown bodies) were considered in the context of the colony as a whole; thus, if brown bodies occurred but the majority of zooids were without tissue, the specimens were not counted as alive.
For summary statistics, diversity is defined as the number of species detected per object.Objects were grouped by item type (i.e., vessel, buoy/float, pontoon section, etc.).Species were grouped by growth habit and form after Amini et al. (2004).
Voucher collections will be deposited at the Royal British Columbia Museum (Victoria, Canada), along with the general JTMD archival collections.
The most common species, occurring on 75 percent and 28 percent of objects, were Scruparia ambigua (d'Orbigny, 1841) and those in the family Aeteidae, respectively.Scruparia ambigua was the only bryozoan present on 39 percent of objects.Scruparia and aeteids were recorded together from 33 objects that supported no other bryozoan species.

Biogeography
Of the 36 taxa resolved to species level, 18 species are known only from the Western Pacific and 15 are already recorded from North America (Table S2).Schizoporella japonica Ortmann, 1890 is regarded as an earlier introduction from Japan.Two species, Callaetea sp. and Arbocuspis sp. are undescribed.One  findings (McCuller et al. 2018).We regarded 13 species not taken to species level as from Japan based upon the objects they were on.In addition, no species unique to the Eastern Pacific were on these objects.We note two exceptions, Pomocellaria californica and Membranipora villosa, which were acquired after the JTMD objects on which they were found entered coastal waters of Western North America.Neither species occurred along with any of the 11 yet-to-be-resolved taxa that we suggest are from the Western Pacific.Six species, Conopeum nakanosum Grischenko, Dick and Mawatari, 2007, Cribrilina mutabilis Ito, Onishi and Dick, 2015, Microporella luellae Grischenko, Dick and Mawatari, 2007, M. neocriboides Dick and Ross, 1988, Callopora craticula (Alder, 1856) and Watersipora mawatarii Vieira, Spencer Jones and Taylor, 2014, were previously recorded from Hokkaido (and, for M. neocriboides, Alaska as well).Virtually all JTMD objects appear to have departed the Tōhoku coast of northeast Honshu by either travelling east or south, rather than moving into colder northern waters (Carlton et al. 2017;N. Maximenko, personal communication, January 2017).Indeed, the objects bearing C. nakanosum and W. mawatarii also bore warmer-water southern species of other bryozoans.It thus may be that these 5 species occur on the northeast Honshu coast, just south of Hokkaido, or further south.We found one species, the cheilostome Escharella hozawai (Okada, 1929), which appears to have been last collected in northernmost Honshu in the 1920s.We newly report Biflustra cf.arborescens (Canu and Bassler, 1928) from the Western Pacific Ocean.Seven additional species, Metroperiella cf.biformis (Zhang and Liu, 1995), Celleporaria brunnea (Hincks, 1884), Drepanophora cf.gutta Tilbrook, Hayward andGordon, 2001, Smittoidea spinigera (Liu, 1990), Biflustra grandicella (Canu and Bassler, 1929), Biflustra irregulata (Liu, 1991), and Celleporina cf.globosa Liu, 2001, are not known from Japan.We suggest that these species were acquired by JTMD as these objects were carried south of Japan, and prior to the objects being then transported by ocean currents back north and east to the Central and Eastern Pacific.Alternatively, it may be that with warming climates, some of these species may have moved poleward in recent decades.As we note in Table S2, Celleporaria brunnea is regarded as introduced from North America to South Korea.While 3 of the 4 objects upon which C. brunnea occurred also supported additional southern species of bryozoans or bivalves, its presence on JTMD may also indicate that it has now spread north to Japanese waters and is yet undetected there as a range expansion.
Nineteen species were represented only by dead colonies (Arbocuspis sp., Biflustra irregulata, Callopora craticula, Cauloramphus spinifer (Johnston, 1832) Remarks.-Thisbryozoan resembles the Northeast Pacific Crisia serrulata known from British Columbia and south, in having 12-30 alternating zooids per internode merged almost to the aperture, which is turned forward, facing frontally, with a short spinous process on its outer margin (Soule et al. 1995).
A similar species, Crisia maxima Robertson, 1910, has zooids connate to a lesser extent and lacking, or without regular appearance, of the short spinous process (Osburn 1953;Soule et al. 1995).Our material consists of a few fragments, internodes averaging 21 zooids, with zooids as described for C. serrulata.While no Crisia species from the Western Pacific is known to possess internodes with this number of zooids, we presume this species is from Japan.The JTMD objects on which this species occurred bore no evidence of species unique to the Eastern Pacific.As no gonozooids were present in the material in hand, we were hesitant to confidently assign these specimens to a species.Distribution.-BritishColumbia to Mexico (Osburn 1953;Soule et al. 1995) and the Galapagos (Osburn 1953).

Crisia sp.
(Figure S1B, C) Material.-JTMD-BF-538.Remarks.-Thesecolonies closely match Crisia operculata Robertson, 1910, known from British Columbia to Panama.Most notably, our material includes gonozooids with ooeciostomes bearing a cap-like projection overarching the aperture, a characteristic of this species.This specimen arrived on a vessel which bore no evidence of other biofouling species acquired in the Eastern Pacific; thus, the material in hand may represent a similar species of Crisia from the Western Pacific.

Crisidia sp.
Material.-JTMD-BF-208.Remarks.-Ourmaterial was a small fragment consisting of few internodes with one zooid and segmented spine per internode.It is most similar to Crisidia cornuta (Linnaeus, 1758), but differs from that species in that new internodes arise laterally, as opposed to dorsally, from the zooids.C. cornuta, an almost certain species complex, is reported as a widespread species in Northern Hemisphere waters, including Japan (Mawatari 1981).  of long (ca.[mean ± SD] 0.77 ± 0.13 mm), slender (ca.0.09 ± 0.01 mm) zooids.Colonies from JTMD-BF-40, based upon measurements of 7 zooids) were larger in zooid length (ca.1.01 ± 0.14 mm) and diameter (ca.0.11 ± 0.02 mm).Colonies from all samples exhibit similar growth patterns, with one zooid per sterile internode, from which originates a branch on either side, and black or dark-brown joints.Japanese literature uses the name F. franciscana (Okada 1928;Mawatari 1955).Filicrisia geniculata, although reported from British Columbia to southern California, is a North Atlantic species that requires re-assessment as to its presence in the Northeast Pacific (Soule et al. 2007).An unidentified Filicrisia species was recorded from North Pacific plastic rafting communities by Goldstein (2014).Osburn (1953) noted that F. franciscana can be found as a fouling species in "considerable masses" on pilings and floats.None of the examined objects bore evidence of any additional encrusting invertebrate species having been acquired in the Northeast Pacific, and thus we assume that these colonies originated from the Western Pacific Ocean.Distribution.-Japan:Mutsu Bay (Okada 1928;Mawatari 1955); Northeast Pacific from Alaska to Baja California (Osburn 1953;Soule 1963) Gordon and Taylor (2001) from New Zealand in their attempt to resolve taxonomic problems within recent cyclostomes.This species has been recorded in Japan (Ortmann 1890;Sakakura 1935) as Lichenopora novae-zelandiae (Busk, 1875), although many Japanese records lack adequate descriptions or illustrations.Our material agrees with the account of Gordon and Taylor (2001) in zooid formation, brood-chamber morphology, and presence of spiny pinhead spinules within autozooid and kenozooid interiors (Figure S2C).Iris-like diaphragms were present in some colonies.A small ledge within the opening of the ooeciostome was observed in one of the two mature colonies.In both mature colonies, the ooeciostome occurred at the distal end of a brood-chamber lobe situated between autozooidal radii (Figure S2B), a character apparently not typical of this species, but similar to that of the specimen illustrated by Osburn (1953) in which the ooeciostome is in the macular center.According to Gordon and Taylor (2001), placement of the ooeciostome is adjacent to the central autozooid of a radii.However, they mention Gordon and Parker's (1991) account of Disporella victoriensis MacGillivray, 1884 (as Lichenopora victoriensis) from Australia, in which the ooeciostome is in the same location as the present material, although this character is also not typical of D. victoriensis.Despite the location of the ooeciostome, presence of the internal ledge suggests our material is D. novaehollandiae.More extensive analyses, including genetics work, are likely needed to determine the placement of D. novaehollandiaelike specimens with atypical ooeciostome location.The variability in size of colonies and their abundance on several JTMD objects suggests that they were of Japanese origin and subsequently produced larvae that settled immediately to form new colonies nearby.Multiple specimens were epizoic on Japaneseorigin mussels, Mytilus galloprovincialis Lamarck, 1819 along with cheilostomatous and other cyclostomatous bryozoans.
Whether Eastern Tropical Pacific and Hawaiian populations represent the true D. novaehollandiae bears further study.Distribution.-Japan(Ortmann 1890;Sakakura 1935), Indian Ocean, New Zealand and Australia (Gordon and Taylor 2001;Gordon and Mills 2016), southern California to Colombia and the Galapagos Islands (Osburn 1953;Soule 1963); Hawaiian Islands (Soule et al. 1987).
(Figure S3A, B) Material.-JTMD-BF-667.Remarks.-Aswe are uncertain of the generic assignment, we place a question mark in front of the genus name.The encrusting portion is tubuliporine, but rises up to form branches with zooids arranged around the stem.One gonozooid is present at a branch bifurcation, with its ooeciostome present at the base of a zooid.
Remarks.-The encrusting, branching nature of our specimens and the presence of an ooecia at the distal end of a branch suggest their placement within Proboscina.However, the free portion of zooids is much raised from the zoarium, a character not typically ascribed to this genus, which may be due to the wide range of environmental conditions faced.To that end, we do not feel comfortable proceeding with a species identification pending further material.Growth is similar to the cheilostome genus Aetea (below), which has an adnate and erect portion, but is different than that group in the extent of calcification and presence of numerous pseudopores characteristic of cyclostomes.No gonozooids were present in this material.

Tubulipora pulchra MacGillivray, 1885
(Figure S4B, C) Material.-JTMD-BF-23,160, 425.Remarks.-Tubuliporapulchra was originally described from Australia and has since been recorded worldwide, including Japan.It is most notable for the toothed primary disc and flared ooeciostome, both characters that our material shares with this species.This is doubtless a global species complex.Distribution.-Inthe Pacific theater, Japan (Okada 1928;Mawatari and Mawatari 1974), China (Liu et al. 2001), South Korea (Rho and Seo 1986), and British Columbia to Galapagos Islands (Osburn 1953).

Class Gymnolaemata Allman, 1856
Order Ctenostomata Busk, 1852 Suborder Alcyonidiina d 'Hondt, 1985Family Alcyonidiidae Johnston, 1838 Alcyonidium sp.Material.-JTMD-BF-40,43, 347, 548, 589, 592, 615.Description.-Colonyrough, encrusting, thin 347,548,592,615) or thick and multilaminar (JTMD-BF-43, 589) with a firm chitinous outer layer.Color unknown in life; in spirit, white/beige (JTMD-BF-43; preserved in formalin) to brown 347;preserved in 95% EtOH).Zooids oval to hexagonal with orifice at distal end of long peristome, polypides with approximately 18 tentacles.Small kenozooids numerous, dispersed across the surface; basal layers composed of large, empty, elongated kenozooids giving the colony an overall spongy texture.Remarks.-Ourmaterial is similar to Alcyonidium sagamianum Mawatari, 1953 described from Kanagawa Prefecture in having the central part of the zoarium filled with empty kenozooids.However, in A. sagamianum the colony form is erect and branching while our specimens consist of a substantial crust over the stalks and valves of a gooseneck barnacle, Lepas sp., or surrounding hydroid stolons (JTMD-BF-43, 589, respectively) or as a unibilaminar sheet 347,548,589,615).The long peristome is shared with that of Alcyonidium mammillatum Alder, 1857, recorded from Japan by Silén (1941), but is more likely a species restricted to the North Atlantic Ocean.Our specimens do not appear to correspond closely to any species of Alcyonidium documented from Japan or the Pacific coast of North America.The genus Alcyonidium requires review and revision in the North Pacific (Ryland and Porter 2013).
Remarks.-Walkeria prorepens is a delicate, easily overlooked, creeping bryozoan.Walkeria uva (Linnaeus, 1758), a global species complex, differs from W. prorepens in its zooid budding locations and symmetry: the former has a pair of zooids budding laterally from internodes, while the latter buds one zooid per internode on its dorsal side.In our JTMD samples W. prorepens was found growing on substrates amongst filamentous algae, hydroid stolons, and other fouling organisms.Our material closely aligns with that of Kubanin (1992) from the Sea of Japan, although only some colonies were in sufficient condition to permit observations on growth pattern.Material of W. uva described and illustrated by Mawatari (1952) from Kii Peninsula may represent W. prorepens.

Callaetea sp.
(Figure 2A; Figure S6C, D) Material.-JTMD-BF-8,131, 134, 170, 208, 227, 229, 230, 255, 264, 339, 342, 349, 356, 387, 402, 405, 514, 521, 638, 648, 661.Description.-Colonylightly calcified, delicate, creeping along substrate or rising above to form tangled masses.From narrow kenozooidal stolons bud lateral stolons irregularly and zooids; basal side of zooid may also bud stolons.Zooids scoop-shaped, curved (ca.0.434 mm ± 0.064 mm long by 0.087 mm ± 0.008 mm wide); frontal membrane elongate (ca.0.279 mm ± 0.042 mm) and narrow (ca.0.038 mm ± 0.017 mm), curvature of the zooid orientating the frontal membrane obliquely in relation to the plane of the stolon.Remarks.-Theonly previously reported species of Callaetea Winston, 2008 occur in the Mediterranean Sea.Our specimens match this genus in having zooids budded from delicate kenozooidal stolons, but differ from Callaetea lileacea Winston, 2008 in the extent of the frontal membrane and in the direction of the frontal membrane.Additionally, while still very delicate, our specimens appear to be more highly calcified than C. lileacea.In that species, SEM photos were not possible as drying the specimens caused them to fall apart (Winston 2008).We also observed zooids that budded kenozooidal stolons from the basal side, not noted in the description of C. lileacea.These differences suggest a new species of Callaetea.We are grateful to L. M. Vieira for pointing out to us that our material more closely matches (and represents a new taxon in) Callaetea, rather than Aetea truncata.
Scruparia ambigua was described from the Falkland Islands in the South Atlantic Ocean in the first half of the 19th century; the name has since been applied, ambiguously, to Scruparia populations in almost all oceans (Osburn 1950;Orensanz et al. 2002;Floerl et al. 2009).S. ambigua thus likely represents a global complex requiring genetic insight, and North Pacific populations may prove to require a distinct name.Although known as a fouling organism on buoys (McCauley et al. 1971;Relini et al. 2000) and other substrates (Haderlie 1969;Gordon and Mawatari 1992;Ferreira et al. 2006), and while clearly capable of ocean rafting as a biofouler (present study), it is rarely reported from ship hull fouling (WHOI 1952) although it may be frequently overlooked in such communities where larger macrobiota are often the focus.
The often extensive colonies of Scruparia on JTMD (typically too abundant to have both colonized and grown robustly in the short time after debris appeared to arrive and reside in North American waters prior to shore landing), its nearly ubiquitous nature, its occurrence on many JTMD objects that bore no evidence of having been colonized by any other Eastern Pacific species, and its apparently rare presence (below) in the Pacific Northwest all suggest that the Scruparia originated from Japan.Moreover, we have examined a number of marine debris items that washed ashore in Oregon after being adrift in the nearshore ocean (as evidenced by Lepas sp.colonization) that indicate (by their identification marks and by their other biofouling) that they originated in the Pacific Northwest.None of these were found to bear Scruparia.Distribution.-Reportedworldwide, as noted above.Recorded from Honshu, Japan by Mawatari (1973).At the time of Osburn's (1950) monographic review of anascan bryozoans of the Pacific Coast of North America, S. ambigua was reported in the temperate northeast Pacific only from Vancouver Island and southern California, such that it was omitted both from Soule et al.'s (2007) treatment of the intertidal and shallow water Bryozoa from Point Conception to the Oregon coast, as well as Bergey and Denning's (1996) summary of Bryozoa of the Pacific Northwest.However, S. ambigua was found to be a member of the fouling community (to a depth of 46 m) on an oceanographic monitoring buoy anchored 55 km off the coast of northern Oregon (McCauley et al. 1971), and on fouling panels deployed in shallow water of Monterey Harbor, in Monterey Bay, central California (Haderlie 1969), as well as being reported from fouling communities in San Francisco Bay (California Academy of Sciences Invertebrate Zoology collections, online database accessed December 2016).The rare report of S. ambigua (if correctly identified) on a buoy in Oregon, given its short-term larvae, is of interest; it may have been able to arrive offshore via algal rafting (McCauley et al. 1971).
Remarks.-Skeletal characteristics of our material agree with the description of Arbopercula angulata by Mawatari (1953, as Electra angulata).We follow Tilbrook et al. (2001) in referring Pacific populations to A. angulata rather than the Atlantic Arbopercula tenella (Hincks, 1880), the latter name also used for Japanese populations by Mawatari (1974) and Kubota and Mawatari (1985).Mawatari (1952) earlier noted that the ancestrula and operculum of A. angulata differ from that of A. tenella.The descriptions of Japanese populations align most closely with those of Levinsen (1909) and Harmer (1926) for A. angulata rather than those of Hincks (1880) for A. tenella.
Distribution.-Thailand(Levinsen 1909); Japan (Mawatari 1953 S8F), respectively.This is a relatively easily recognized species due to the large zooids and the row of uniporous septula that runs along the midline of transverse walls.Colonies were composed of zooids with both granular and smooth cryptocysts depending on the extent of calcification.Some JTMD specimens had spines projecting into the opesia as noted by Liu et al. (2001) Present in all specimens were distinct brown chitinous lines running through the interzooidal grooves, as noted by Liu (1992) and Seo and Min (2009); this character is also present in Biflustra arborescens (Canu and Bassler, 1928) (Almeida et al. 2017), which, however, has oval-rectangular zooids and a beaded mural rim.One characteristic of B. irregulata not observed was chitinous spines on the frontal membrane; Taylor and Tan (2015)'s Malaysian material also lacked spines.This species has been observed encrusting a wide variety of substrates, including mollusk shells, gorgonians, corals, stones, and plastic (Liu 1992;Seo and Min 2009), but may also foul buoys, cables, nets, and cages (Liu et al. 2001).Powell (1971, below) found it on wharf pilings, floats, navigation buoys, and rocks; Taylor and Tan (2015) report it from oyster rafts and gastropod shells.
Reports of B. arborescens from the tropical Eastern Pacific likely represent B. irregulata.Banta and Carson (1977) described their material from Costa Rica as possessing "about 30 cuticular spinules" on the frontal membrane, zooids "sometimes distorted in shape", and zooids surrounded by brown lines, characteristics which, combined, are those of B. irregulata, not B. arborescens.Biflustra irregulata was not described when Banta and Carson (1977) and, earlier, Powell (1971), both studying material from the Pacific coast of Panama, were considering potential membraniporid candidates.Distribution.-CentralIndo-Pacific: Bohai Bay (Yellow Sea) and South Korea through the East China Sea and the South China Sea (Liu 1992;Liu et al. 2001;Seo and Min 2009); Indian Ocean: Penang, Malaysia (Taylor and Tan 2015) and Cox's Bazaar, Bangladesh, Bay of Bengal (Gordon et al. 2007).Vieira et al. (2016)  Material.-JTMD-BF-17,35,49,96,144,145,207,209,212,223,226,227,253,304,329,339,341,363,367,383,392,408,410,411,413,415,428,524,527,530,533,536,555,556,557,558,566,570,572,574,581,583,585,590,598,645,649,652,653,659,668,669,670,671,672.
Remarks.-Jellyella eburnea is a neustonic species typically found on plastic marine debris (Thiel and Gutow 2005;Goldstein et  Material.-JTMD-BF-1,5,8,13,40,94,134,196,199,203,207,227,240,251,290,328,330,336,356,413,455,512,513,528,533,580,587,638,639,652,654,659. Remarks.-Widespread neustonic species primarily epiphytic on floating Sargassum in the Atlantic Ocean, but also found on shells, buoys, and other plastic materials worldwide (Liu 1992;Thiel and Gutow 2005;Goldstein et al. 2014).Characteristic of this species is branched spinules, hooks, or paired wings projecting into the opesia (Figure S10D).Colonies from JTMD were found overgrowing the original coastal fouling community as well as growing over other pelagic acquisitions such as the goose neck barnacle Lepas sp.Jellyella tuberculata appears to be a cooler water species than J. eburnea, with earlier records of J. tuberculata in tropical and subtropical waters (as noted above) requiring verification.Additional material of J. eburnea or J. tuberculata is in hand (JTMD- 42,149,154,156,224,236,349,390,391,420,422,442,455,504,515,518,527,546,578,579,640,661), but colonies were too fragmentary to assign to species.Distribution.-Temperatereports include Japan (Mawatari 1974), California (Soule et al. 1995), New England to Florida (Osburn 1912, as Membranipora tehuelcha d'Orbigny, 1839;Winston 1982b), and South Africa (Florence et al. 2007).Warmer water reports in the Pacific and Indo-Pacific theaters include South Korea (Seo and Min 2009), Bangladesh (Gordon et al. 2007), Vanuatu (Tilbrook et al. 2001), Costa Rica (Banta and Carson 1977), and the Galapagos Islands (Chiriboga et al. 2012), and in the Atlantic, Colombian Caribbean (Montoya-Cadavid 2007) and Brazil (Vieira et al. 2016).

Membranipora villosa Hincks, 1880
Material.-JTMD-BF-349,370, 383, 384, 405, 408, 410.Remarks.-Threemorphologically-similar Membranipora, M. membranacea, M. serrilamella and M. villosa, in the North Pacific were thought to be differentiated by the development and serration of the cryptocyst and the presence or absence of cuticular spines (Osburn 1950): M. villosa with a finely serrated cryptocyst with cuticular spines on the frontal wall and at zooidal margins, M. serrilamella with a heavily denticulate cryptocyst with longer spinules and no cuticular spines, and M. membranacea with a narrow, smooth cryptocyst and lacking frontal cuticular spines.Yoshioka (1982) argued that these morphologies represented ecophenotypes induced by nudibranch predation, placing M. serrilamella and M. villosa in the synonymy of the older name M. membranacea.Dick et al. (2005), after reviewing the systematic and nomenclatural complexities since Yoshioka's work, resurrected M. villosa, placed M. serrilamella in synonymy with M. villosa, and restricted M. membranacea to the Atlantic Ocean.Dick et al. (2005) noted that only the serrilamella morphology, lacking cuticular spines, had been reported from Japan (Mawatari 1974;Liu et al. 2001; as well as Seo and Min 2009), rather than the villosa morphology, and speculated that the absence of the latter morphology might be due to the introduction of M. serrilamella from North America to Japan, "in which case Japanese populations of nominal M. serrilamella might not exhibit a defensive response to chemical cues from species of nudibranch predators native to Japan" (Dick et al. 2005 also note that "an introduction would also explain the low genetic divergence between Japanese specimens and one of the clades at Friday Harbour, Washington").
We follow Dick et al. (2005) in using the name villosa for our Japanese material, which while lacking cuticular spines on the frontal wall (in concert with the serrilamella-morph), displays cuticular spines at the zooid margins, a feature also attributed to M. villosa, as well as the villosa-like cryptocyst.Tower cells, indicative of the membranacea-morph, were not observed.While most colonies arrived without tissue, a few (on JTMD-BF-405, 408, 410) small colonies, growing on algae, possessed tissue.All three of these objects also bore light signatures of newly settled Northeast Pacific species (such as nepionic barnacles and newly settled crabs); moreover, colonies of M. villosa on JTMD-BF-408 were found overgrowing Jellyella eburnea.Colonies of M. villosa on these objects may represent Eastern Pacific acquisition.Distribution.-Japan(Mawatari 1956(Mawatari , 1974)) (Hayward and Ryland 1998).
Remarks.-Our specimen matches Cauloramphus spinifer with its 3 orificial spines and 5-10 opesial spines.This species may be circumboreal, as it was originally described from Europe and subsequently recorded in the Pacific, or it may consist of cryptic sibling species.

Cauloramphus sp.
(Figure S12C-F) Material.-JTMD-BF-160.Remarks.-Acolony fragment in hand is too poor to assign to species level.Kenozooidal ovicells forming a crescentic hood and an extensive frontal membrane suggest placement in Cauloramphus.Up to three pairs of spines or spine bases were located laterally around the zooid margin.No orificial spines were present and it is unclear whether the process on the ovicells represents a pore or the base of an orificial spine (Figure S12C, E).No avicularia were observed, although there appear to be bases of pedunculate avicularia just proximal to the distalmost pair of lateral spines (Figure S12F).Our specimen is similar to Cauloramphus korensis Seo, 2001 in the lack of distal spines and the large kenozooidal ovicell, but does not have brown or violet spines.
Combined with molecular evidence, this Western Pacific bugulid is described as a new species in an accompanying paper (McCuller et al. 2018).
Remarks.-We report this well-known fouling species (long known as Bugula stolonifera) on the basis of its detection in a metagenomic analysis of a fouling community sample from the "Misawa 1" (JTMD-BF-1) dock that landed in central Oregon in June 2012.A sequence from these samples was 100% identical to a sequence of B. stolonifera from a fouling community in Galizia, Spain (Fehlauer-Ale et al. 2015;Genbank KC129849-1;Jonathan Geller, personal communication, 2017).The sequence also matched sequences of specimens identified as B. stolonifera from San Francisco Bay and from Matsushima Bay, Honshu (noted below) (J.Geller, personal communication).While we have no morphological vouchers of B. stolonifera from JTMD-BF-1, the massive "Misawa 1" dock, from the Port of Misawa, Aomori Prefecture, presented a rich, complex community, which continued to reveal previously undetected taxa in studies of additional archived samples over several years, and it is thus not surprising that no voucher specimens are in hand.
Distribution.-Reported from harbors and bays world-wide, the original home of this species, obscured by global shipping, has yet to be determined (Carlton and Eldredge 2009;Winston and Hayward 2012).A curiously late arrival in Japan, it was first found in 1997 in the Port of Nagoya (Scholz et al. 2003), and by 2013 had extended north to Tokyo Bay (Lutaenko et al. 2013) 8,12,18,23,40,58,121,131,139,168,177,205,210,237,331,336,657.
Remarks.-Considerable variation occurs within and between colonies of Tricellaria inopinata, resulting in confusion between it and the Eastern Pacific Tricellaria occidentalis (Trask, 1857).Historically, T. occidentalis and T. occidentalis var.catalinensis (Robertson, 1905) were recorded from Japan (Mawatari 1951).Dyrynda et al. (2000) suggest that Japanese species recorded as T. occidentalis are likely T. inopinata (described from the Mediterranean, but believed to be native to Japan), due to the species' extensive morphological variation, and further provide evidence that the two species are distinguishable.Grischenko et al. (2007) report specimens from Akkeshi Bay that had purported characters of both species within the same colonies.As no type material of T. occidentalis appears to be available, and as only older material was used by Dyrynda et al. (2000) to distinguish T. occidentalis, Grischenko et al. (2007) maintain use of the name T. occidentalis in Japan until the two species are more thoroughly delineated.We note that colony form, although not previously discussed, may aid in distinguishing the two species.T. occidentalis grows in bushy tufts formed of curved branches which roll inward (Osburn 1950;Mawatari 1951;Grischenko et al. 2007).In comparison, the bushy colonies of T. inopinata have branches that are only slightly rolled inward, or are straight or rolled slightly outward (D 'Hondt and Occhipinti-Ambrogi 1985;De Blauwe and Faasse 1998;Johnson et al. 2012).This difference may be due to the average number of zooids per internodes: 3 (most common) to 5 in T. occidentalis (as in Figure S14C) and 3 to 19 for T. inopinata (as in Figure S14D).
JTMD material contains some specimens that appear most closely aligned with T. occidentalis (JTMD-BF-40) and others that were closer to T. inopinata (JTMD-BF-58).As the majority of our material is similar to introduced populations of T. inopinata that we have studied in New England, USA, we treat the present specimens as representing this species.Distribution.-Japan(Okada 1929;Mawatari 1951); North West Atlantic (Johnson et al. 2012); Northeast Atlantic (Occhipinti-Ambrogi and D 'Hondt 1994).
Remarks.-This material resembles populations of Catenicella contei (Audouin, 1826) from Brazil (Ramalho et al. 2014) in the structure of the scapular chambers and vittae, and in the absence of avicularia.However, as ovicells are an important character for identification of catenicellids, our identification remains at the genus level, pending further material.One species, Catenicella elegans Busk, 1852, is reported from Japan (Okada 1921), but that species has paired avicularia.

Superfamily Celleporoidea Johnston, 1838
Family Celleporidae Johnston, 1838 Four Celleporina species occur on JTMD.Additional material of Celleporina is in hand from JTMD- 105,150,336,and 455, as very small fragments in poor condition not assignable to species.

Celleporina sp. A (Figure S15E, F)
Material.-JTMD-BF-18,40.Description.-Colonyencrusting, forming small nodules or domes.Peristome high; primary orifice longer than wide with a narrow U-shaped sinus; paired adventitious avicularia on either side, directed disto-laterally, with slightly serrated distal edges.Young, frontally-budded zooids with a single series of large pores around the entire orifice.Brooding zooids with hemispherical ovicells; tabula large, covering the entire frontal area.No vicarious avicularia observed.No spinules within zooidal chambers observed.Remarks.-Aspecies which shares similar characters is Celleporina serrirostrata Liu, 2001 from the coast of Guangdong Province in the South China Sea.However, our specimens do not appear to have spinous processes within the chambers of frontally-budded zooids as does C. serrirostrata.
Remarks.-This specimen is distinct from other JTMD Celleporina in having abundant vicarious avicularia and inconspicuous adventitious avicularia.The specimen arrived in our hands too late for imaging.
Material examined here differs from those species most notably in the number of marginal pores; the aforementioned species have few large pores, while our specimen has many small pores.The sinus is also a deep U-shape, differentiating it from the broad, shallow sinuses of both R. tubulosum and R. verruculatum.
Remarks.-This single colony fits the recently described Cribrilina mutabilis.As reported in Ito et al. (2015), this species has 3 different zooid types: R, I, and S. Our colony has a mixture of both R and I types with approximately 8-10 costae fused at the midline (R type), with or without intercostal pores (I type).C. mutabilis has previously been reported as growing on seagrass or algae blades; our specimen was found on a mussel shell (Mytilus galloprovincialis) in the process of being overgrown by Scruparia ambigua.
Remarks.-Reported as a widespread Arctic-Boreal taxon that is a likely species complex as noted by Grischenko et al. (2007), based on differences in number of pores on the ovicell.JTMD material matches the description of Celleporella hyalina from Hokkaido, Japan by Grischenko et al. (2007), but also present some differences.Autozooid orifices were much smaller (ca.0.10 mm long by 0.09 mm wide) and male zooids were approximately one-third to one-half the size of autozooids.Ovicells were also much smaller (ca.0.13 mm long by 0.18 mm wide) than that of typical C. hyalina.These discrepancies may be due to the wide variety of temperatures the colonies likely experienced over the course of their rafts' journeys.Molecular studies on C. hyalina colonies collected in the Atlantic basin and the Pacific coast of Chile confirm that subgroups (or distinct species) exist and that differences are reflected in zooid morphology, life history, and ecology (Hoare et al. 2001).In JTMD material, C. hyalina is primarily found on the Japanese mussel Mytilus galloprovincialis, the Japanese acorn barnacle Megabalanus rosa Pilsbry, 1916, and the neustonic crab Planes marinus (JTMD-BF-40; Figure 2C).Some colonies, such as those present on JTMD-BF-40, were alive with ovicells and embryos at the time of preservation, suggesting that this species is capable of reproduction after rafting long distances.
Distribution.-Widely reported from the North and South Atlantic Oceans.Western Pacific records include Japan (Okada 1929;Mawatari 1956;Grischenko et al. 2007), Sea of Japan (Grischenko and Zvyagintsev 2012), South China Seas (Liu et al. 2001), while in the Eastern Pacific it has been reported from Alaska to the Galapagos Islands (Osburn 1952;Dick et al. 2005;Soule et al. 1995), largely under its wellknown older name Hippothoa hyalina.
Remarks.-Escharella hozawai was described (as Mucronella hozawai) from Mutsu Bay (in Kawauchi and off Tozawa) Aomori Prefecture (Okada 1929), and appears to have not been recorded since.E. hozawai differs from a morphologically similar Japanese species, Pacifincola perforata (Okada and Mawatari, 1937), in the presence of paired curved spines on either side of the lower margin of the ovicell, and a tatiform ancestrula with 8 spines that buds three zooids around its distal end, as opposed to the cribriform ancestrula with four spines that buds four zooids in P. perforata (Grischenko et al. 2007).Our material showed considerable variation in spine presence on the ovicell.Paired spines were absent in some parts of the colony, but long and distinct in other parts.E. hozawai appears closely allied to Pacifincola; further study may suggest eventual placement in that genus.Material from two objects, a plastic can fragment (JTMD-BF-616) and a plastic tote (JTMD-BF-362), were composed of an encrusting base of E. hozawai overgrown by hydroids.Okada (1929) reported this species as growing on seagrasses such as Zostera marina Linnaeus, 1753; our material consists of fouling colonies on fiberglass and plastic.Distribution.-Japan:Mutsu Bay, Honshu (Okada 1929).
Distribution.-Japan:Hokkaido (Grischenko et al. 2007), China (Liu et al. 2001(Liu et al. , 2003)).Remarks.-Manyspecies of Microporella occur on the Japanese coast.Our material agrees with Microporella borealis in possessing four to five oral spines, orifice with a denticulate proximal rim, denticulate crescentic ascopore, and a lanceolate avicularian mandible with paired hooks.However, M. borealis has two spines that occur proximal to the ovicell which our specimens were lacking, perhaps due to extensive calcification.

Microporella borealis
Distribution.-Japan:Hokkaido (Suwa and Mawatari 1998) and Shimoda (Kaselowsky 2004), South Korea (Seo and Min 2009).Remarks.-Ourspecimen aligns most closely with Microporella luellae in possessing 2 widely-spaced orificial spines, a single avicularia, and a crescentic, denticulate ascopore, but differs in that the orificial spines were present in reproductively mature, as well as immature, zooids.This may represent phenotypic variation due to the wide range of conditions experienced by the rafted colony.A similar species is M. neocriboides (below), which has an ascopore covered with a cribriform plate.Distribution.-Japan:Hokkaido (Grischenko et al. 2007).Dick and Ross, 1988 (Figure S21E, F) Material.-JTMD-BF-657.Remarks.-Thisspecimen agrees well with Microporella neocriboides as described by Suwa and Mawatari (1998) and Dick et al. (2005), but differs in avicular characters.Whereas M. neocriboides usually has one or no avicularia, zooids within our specimen typically have two avicularia, and occasionally one or three (Figure S21E); those with two avicularia were often paired laterally to the aperture, although sometimes both avicularia occur on one side.Microporella inermis Liu and Liu, 2001 is similar, but lacks the cribriform plate underneath the pseudopores which was present in our material (Figure S21F).Distribution.-Japan:Hokkaido (Grischenko et al. 2007), Alaska (Dick and Ross 1988).Remarks.-JTMDspecimens occur as both uniand multi-laminar growths of mature zooids with prominent ridged and porous ovicells.Zooids typically lacking oral avicularia or had one or rarely a pair.Raised, frontal avicularia were sometimes present.Top layers (in this case, those flush with an oyster shell on a buoy (JTMD-BF-215)) often had a small conical umbo on the frontal wall proximal to the orifice; umbos on layers beneath were quite robust, extending far above the zooidal plane (Figure S22A).These characters agree with those of Schizoporella japonica (Grischenko et al. 2007) (Okada and Mawatari, 1936), are also recorded from Japan, as Codonella acuta (Ortmann, 1890) and Codonella spatulata, respectively (Okada and Mawatari 1936).The former species is differentiated from M. biformis by its acute avicularia.M. spatulata is similar to our species but has an orbicular aperture.We tentatively identify JTMD material as M. biformis, which was originally described from the East China Sea as a fouling organism on aquaculture cages, ship bottoms, and buoys, pending more detailed descriptions of M. spatulata.Our material is from a vessel's hull along with Bugula tsunamiensis, Exochella tricuspis, Filicrisia cf.franciscana, Scruparia ambigua, Smittoidea spinigera, and Tricellaria inopinata.Many zooids contained polypides and while a majority were ovicellate, only a portion were with embryos.Distribution.-EastChina Sea (Liu et al. 2001).
Smittoidea prolifica Osburn, 1952 has large avicularia with a rounded mandible, but 2-4 spines are present only within the zone of astogenetic change.Our material most closely corresponds to that of Smittoidea spinigera, described from Bohai Bay to the coastal waters of Guangdong, China, in having 3-6 distal spines present even outside the zone of astogenetic change and a rounded avicularian mandible with a triangular projection on the crossbar (Figure S23E).Distribution.-NorthernChina: Bohai Bay to Guangdong (Liu 1990;Liu et al. 2001).

Family Watersiporidae Vigneaux, 1949
We report two species of Watersipora from JTMD material.Additional Watersipora specimens from 197,and 290 were in hand, but consist of too few zooids or were in too poor condition to permit species identification.

Prior records of ocean rafting
Almost none of the 49 species of coastal bryozoans recorded here have been reported previously from open ocean rafting (Table S2).While a number of bryozoans are reported from rafting communities (Winston et al. 1996;Astudillo et al. 2009;Kiessling et al. 2015), these are largely from local, inshore, coastal waters, records that do not establish the potential of these species for long-distance transoceanic rafting.Goldstein et al. (2014) are among the first to report Bryozoa from the open ocean of the North Pacific.They reported the three neustonic species we record here (Jellyella eburnea, J. tuberculata, and Arbopercula angulata (as Membranipora tenella)), as well as Bugula spp., Filicrisia spp., Stomatopora spp.and Tubulipora spp.(none identified to species).We found representatives of the last four genera in our JTMD samples.In addition, Goldstein et al. (2014) reported Bowerbankia spp., which may be our Walkeria prorepens, as well as Victorella spp., a genus-species group normally associated with estuarine, brackish systems.Aetea anguina has been reported in the open Atlantic Ocean in the Sargasso Sea (Table S2), although A. "anguina" of Pacific and Atlantic waters may be different species.We have tentatively equated (Table S2) our records of Filicrisia and Tubulipora with genus-level records of Goldstein et al. (2014), but species-level confirmation will require examination of the material collected by Goldstein and colleagues.

Bryozoa as tracers of marine debris routes
The three oceanic (neustonic) bryozoans, Jellyella eburnea, J. tuberculata, and Arbopercula angulata, provide insight into the routes that certain JTMD items may have taken, and thus the conditions experienced, as they rafted from the Western Pacific to the Central and Eastern Pacific.The cooler-water species J. tuberculata and A. angulata may be more common in waters north of the 20 °C isotherm, whereas J. eburnea may recruit in larger numbers in warmer, subtropical and tropical waters.In turn, J. tuberculata and A. angulata were found primarily on JTMD objects arriving in the Pacific Northwest, whereas J. eburnea was most common on objects landing in the Hawaiian Islands.Indeed, both of the former species were found only once each in Hawai'i (on BF-654 and BF-653, respectively, both landing in 2016).These two objects thus took a higher latitude route through the North Pacific, before turning south and west back to the Central Pacific.With rare exception, J. eburnea did not appear on items landing on the North American Pacific coast until 2015.Records increased thereafter, suggesting that an increasing number of objects (as noted earlier in Results) spent more time on longer and more circuitous routes in lower latitudes (Figure 3).For example, in a 7-week period between 3 March and 18 April 2016, 12 objects were found in Oregon and Washington with J. eburnea aboard (along with other southern invertebrate species), whereas it was found on only two objects arriving in the same region in the previous 6 months.Of interest then are the records of objects (such as 413,and 659) arriving in the Pacific Northwest with both J. eburnea and J. tuberculata attached, or both J. eburnea and A. angulata (such as 341,and 428), or all three species (BF-207), suggesting that these objects had travelled through lower latitudes before being acquired by ocean currents that then took them north and east to North America.

Colony morphology and potential to survive longdistance rafting
All specimens within the Calloporidae (3 species) and Celleporidae (4 species), as well as 4 of the 6 species within the Electridae and Membraniporidae, were represented only by dead colonies.While the potential exists for these species to have arrived alive on objects not intercepted, this and the high diversity but low frequency of the unilaminar encrusting morphology suggests that this growth form is not well-suited to the open ocean.
Creeping-uniserial and erect-flexible-articulatedbranching were the most frequently occurring and abundant growth forms found on JTMD, followed by encrusting-uniserial and -multiserial forms.The adaptive significance of such morphologies has been long discussed (Stach 1936;Jackson 1979;Hageman et al. 1998Hageman et al. , 2013;;Amini et al. 2004).Creeping uniserial forms (or "runners") have high directionality, in which the acquisition of preferable spatial refuges is maximized, with the caveat that unsuitable habitat patches will also be encountered.Flexible-articulatedbranching-forms (arborescent forms or "trees") are structured to rise above the substrate (away from spatial competitors below them) and into the water column where food availability may be higher, albeit at the cost of fewer attachment points and increased susceptibility of breakage from water velocity (Jackson 1979;McKinney and Jackson 1991).Indeed, Scruparia ambigua, the species found most frequently on JTMD, sends out uniserial encrusting runners, which then undergo frontal budding to form erect branches and thus may maximize the benefits of both runner and arborescent forms on potentially space-limited, slow-moving rafts.
Another group of creeping species are composed of zooids with encrusting and erect portions (Aetea anguina) or short, erect, tubular zooids connected by a thin stolon (Callaetea sp.); the lophophore is somewhat extended from the substrate and may have better access to food than strictly encrusting species.Of interest is that Callaetea sp. was identified on JTMD more frequently and in a wider variety of microhabitats than A. anguina, perhaps explained in part by the more delicate stolon of Callaetea sp. that does not need to adhere directly to the substrate as does the adnate portion of A. anguina.Callaetea sp. could be considered more of a "vine" but with reduced ability to extend into the water column in an arborescent manner (Jackson 1979).Likewise, Hageman et al. (2013) found that Aetea truncata, a species similar to Callaetea sp., was more abundant and widely distributed within microhabitats than Aetea sica (Couch, 1844), which adheres to the substrate in a manner similar to A. anguina.
Overall, species of Aetea, Callaetea, Scruparia, Tricellaria, and Bugula tended to be relatively abundant on JTMD.As noted above, the adaptive significance of an growth form includes increased access to food, increased rates of feeding and thus reproduction, and from competition and predation (McKinney and Jackson 1991).The species in these genera have flexible colonies apparently capable of withstanding drag forces endured by water motion over slow-moving rafts.Hydroids, another erect colonial organism, were similarly often found in high abundance (Calder et al. 2014).In the absence of free primary space afforded by floating debris, especially once Lepas barnacles begin to settle and grow, erect species may hold the advantage over encrusting forms on longlived anthropogenic rafts.For example, Scruparia (and occasionally Aetea) were observed growing up hydroid stolons, which thus provided additional surface area.In contrast, encrusting species grew on the object itself or occurred as epibiota on algae, bivalves, barnacles (including Lepas), and even the neustonic crab Planes, but in turn were regularly overgrown by hydroids, bivalves, sponges, and other bryozoans.
Further potentially contributing to rafting survival may be larval type, at least in the arborescent species.
Scruparia ambigua and Aeteidae brood lecithotrophic, non-feeding, coronate larvae that settle hours after release.Of the two, S. ambigua was more common and abundant than the Aeteidae, which is likely due to the number of brooded embryos per ovicelled zooid: up to 7 and 1-2, respectively (Mawatari 1973;Cook 1977).Tricellaria inopinata was often also abundant, its ovicelled zooids producing 1 coronate larva at a time.

Invasion potential
As noted earlier, only a small fraction of the JTMD field was intercepted and analyzed (Carlton et al. 2017).Thus it is likely that more species arrived than we have reported here.Many Western Pacific species of bryozoans, ranging from the major strike zone of the tsunami of the colder water Tōhoku coast of northeast Honshu to warmer waters south of the Boso Peninsula − and perhaps as far south as the South China Sea where JTMD may have drifted − would find a strong climatic match along the North American Pacific coast and in the Hawaiian Archipelago.Long-term monitoring of bryozoan diversity in these regions will be required to determine if this vast field of debris has led to the introduction of novel species in these regions.

Figure 1 .
Figure 1.Frequency distribution of bryozoan species detected on Japanese tsunami marine debris samples.

Filicrisia
franciscana and Filicrisia geniculata (Milne Edwards, 1838) are nearly identical with the exception of the morphology of the ooecia(Osburn 1953), which our material (which may represent two taxa) lacks.Colonies (BF-23, measurements based upon 10 zooids) may form delicate tufts

Family
material is uniserial, consisting of zooids creeping over barnacle and bivalve shells.

Family
from JTMD agrees entirely with the description of Celleporaria brunnea by Canning-Clode et al. (2013) and with other standard treatments of the species.C. brunnea is native to and widespread throughout the Northeastern Pacific Ocean, but has recently begun to appear in other regions of the world (due presumably to transport by shipping) including South Korea in 2004 (Seo and Min 2009), Turkey in 2004 (Koçak 2007) and Portugal in 2012 (Canning-Clode et al. 2013).Remakably, our four collections of this species were all from the same type of object-small pontoon sectionslanding in Hawai'i in 2013 (JTMD-BF-27), in Washington in 2013 and 2014 (JTMD-BF

Table 1 .
Northwestern Pacific coastal Bryozoa found on JTMD arriving in the Hawaiian Islands and North America.
species, Bugula tsunamiensis McCuller, Carlton and Geller, 2018, was described as a result of JTMD

Material.-JTMD-BF-338. Remarks.-Characteristics of
a small, dead colony fragment agree with Conopeum nakanosum in possessing a pair of distal kenozooids.Our specimen was found in association with Aeteidae, Arbocuspis sp., Biflustra grandicella, and Arbopercula angulata on a pallet.This Conopeum colony was overgrown by the southern species Biflustra grandicella.
(Taylor and Monks 1997)algae, and on the shells of bubble-rafting Janthina snails or the dead shells of Spirula squid(Taylor and Monks 1997).Colonies occurred commonly on buoys and other plastic JTMD items, as well as being a member of the epibiota of the common JTMD mussel Mytilus galloprovincialis.Our material shows considerable variation in the gymnocystal structures, ranging from a low-lying ruffled shelf to a high shield of tubercle or spine-like processes.Occasionally the gymnocyst resembles that of congener Jellyella (Winston 1982bnks 1997)s of J. tuberculata in tropical locales may represent J. eburnea.Distribution.-Awarmer-waterspecieswidelydistributed, as expected for an open ocean neustonic species, but whether consisting of multiple isolated clades that may in fact compose cryptic species is not known.Australasia, Indo-Pacific(Gordon et al. 2007), Northwest Pacific, Malaysia(Taylor and Tan 2015); South Africa(Taylor and Monks 1997); West Atlantic: Florida(Winston 1982b, as Membranipora sp.).
. Material in hand provided and identified by Dr. Michio Otani is from fouling panels set out in 2015 north of Tokyo in Shiogama, Matsushima Bay, Miyagi Prefecture on the Tōhoku coast.
to suggest that it may have been more recently introduced to the Northeastern Pacific Ocean.
. Additional material of Schizoporella is in hand from JTMD-BF-131, 164, and 264, but consist of few or heavily abraded zooids.While these are likely to be S. ).