Introduction
Insects are relatively rare in the marine environment. Although approximately 25 000 to 30 000 species of insects are aquatic or have aquatic larval stages, only several hundred species are considered marine or intertidal (Cheng Reference Cheng and Cheng1976). With the exception of the sea skaters, Halobates (Hemiptera: Gerridae) – the only insects to live in the open ocean on the sea–air interface – the majority of marine species live in intertidal (littoral) or coastal habitats, and most of them belong to the orders Hemiptera, Diptera, and Coleoptera (Harada et al. Reference Harada, Furuki, Ohoka, Umamoto, Nakajo and Katagiri2016).
However, despite their smaller proportion, marine insects remain poorly enumerated, and potentially rich faunas of coastal marine insects may be overlooked (Cheng Reference Cheng and Cheng1976). Two reasons exist for this: (1) references to marine insects are scattered in entomological literature, and (2) there remains a lack of consensus regarding the definition of a marine insect, particularly whether insects present in the supralittoral zone should be included or whether to include those species that are only sporadically associated with the marine environment. Cheng and Frank (Reference Cheng and Frank1993) define a marine insect as a species that spends at least one of its developmental stages habitually in a marine habitat. Vermeij (Reference Vermeij2020) considers terrestrial arthropods to be secondarily marine if they must feed, either as larvae or adults, on other marine organisms. Regardless of how they are defined, marine (or secondarily marine) insects have generally been ignored by researchers both in entomology and in marine invertebrate studies (Cheng Reference Cheng and Cheng1976). Consequently, relatively little is known about the biology and ecology of the majority of marine insect species.
Saunders, although working at the University of Saskatchewan (Saskatoon, Saskatchewan, Canada), spent at least one summer on the Pacific coast of Canada. He based his paper, “Some marine insects of the Pacific coast of Canada” (1928), on specimens that he collected during the summer of 1927 in Departure Bay [Nanaimo], Vancouver, Prince Rupert, and Ucluelet, British Columbia, Canada. The paper combines species descriptions and life history observations. Saunders recorded eight species that often interact with one another but are restricted to existing at the high-tide line. He provided species descriptions for three species of Diptera: Chironomidae: Camptocladius pacificus, C. marinus, and C. clavicornis. Adults of all three species were observed swarming above or walking on and ovipositing in algae growing on rocks. These three species were subsequently placed within the genus Saunderia but are now considered to be within the genus Thalassosmittia (Diptera: Chironomidae) (Oliver et al. Reference Oliver, Dillon and Cranston1990). Adults of a tiny staphylinid beetle, Diaulota densissima (Casey, 1893) (Coleoptera: Staphylinidae), were also observed crawling amongst the rocks at the high-tide line. Adults of a dryomyzid fly, Oedoparena glauca (Coquillett, 1900) (Diptera: Dryomyzidae), were observed and presumed to be saprophagous in nature. Saunders makes no mention of barnacles (Balanomorpha: Balanidae, Chthamalidae) in his paper.
Morley and Ring (Reference Morley and Ring1972a, Reference Morley and Ring1972b) conducted extensive studies on intertidal Chironomidae in British Columbia. They observed all three of Saunders’ species of Thalassosmittia in both the Victoria, British Columbia region and in Haida Gwaii, off the central coast of British Columbia. The authors hypothesised that all three species of Thalassosmittia have overlapping generations throughout the year, with adults present in any given month but usually most abundant in autumn. The authors do not specifically mention barnacle associations for these species. The most recent and comprehensive catalogue of Nearctic Chironomidae (Oliver et al. Reference Oliver, Dillon and Cranston1990) includes three species of Thalassosmittia (T. clavicornis, T. pacifica, and T. marina). All three are described as having ranges from California, United States of America north to British Columbia.
The original description of Diaulota densissima (Casey, 1893) is based on a single specimen collected in Fort Wrangell, Alaska, United States of America (Casey Reference Casey1893). No mention of the habitat is included in the original description, but in his revision of Diaulota, Ahn (Reference Ahn1996) notes that all members of the genus, including D. densissima, are often found between and within barnacle tests. They likely persist in air bubbles trapped in crevices as the tide comes in. He also notes that the species is found on shorelines from Alaska to California. In their global review, Frank and Ahn (Reference Frank and Ahn2011) list 392 species of Staphylinidae inhabiting mainly or exclusively coastal ecosystems and confirm the association between D. densissima and barnacles.
The dipteran family Dryomyzidae, sometimes referred to as the forest flies, is found throughout the Northern Hemisphere but is relatively species poor. The 22 described extant species are currently organised into six genera (Mathis and Sueyoshi Reference Mathis and Sueyoshi2011). Although most species in the family are suspected to be inhabitants of intact forest, with larvae developing in dung, decaying plant matter, or carrion, the three species of one genus, Oedoparena, are unique within the family by being associated with barnacles. Oedoparena glauca is formally described (originally as Oedoparea in the family Sciomyzidae) by Coquillett (Reference Coquillett1900), based on specimens from the Harriman Alaska Expedition of 1899. Specimens collected in Metlakahtla and Farragut Bay, Alaska are mentioned in the description. Mathis and Sueyoshi (Reference Mathis and Sueyoshi2011) provide a detailed taxonomic history leading to the current generic spelling and classification. Based on the literature and specimens examined to date, Steyskal (Reference Steyskal, Stone, Sabrosky, Wirth, Foote and Coulson1965) summarises the range of Oedoparena glauca as Alaska to California in A catalog of the Diptera of America North of Mexico.
Burger et al. (Reference Burger, Anderson and Knudsen1980) performed extensive observation and laboratory rearing of O. glauca specimens and detailed their relationship with barnacles. The authors observed females ovipositing in Balanus glandula Darwin Reference Darwin1854 (Balanomorpha: Balanidae), but larvae were also observed feeding within the smaller barnacle, Chthamalus fissus Darwin Reference Darwin1854 (Balanomorpha: Chthmalidae). Laboratory experiments confirmed that each larva consumes multiple barnacle hosts through the course of three larval moults. Empty tests of dead barnacles are used as pupation sites. A greater concentration of Oedoparena larvae was observed in the middle and lower portions of the barnacle bed, but this may be due to larger barnacle hosts in those regions. As many as 12.5% of barnacles were observed to contain Oedoparena larvae.
Mathis and Steyskal (Reference Mathis and Steyskal1980) add numerous records of O. glauca, based on an examination of adult specimens housed at numerous collections. These records comprise more than 30 locations from Unalaska, Alaska in the north to San Luis Obispo County, California in the south. In addition to providing new location records for O. glauca, the authors also describe a new species, O. nigrifrons (Diptera: Dryomyzidae), based on specimens collected in Ilwaco, Washington, United States of America. Additional specimens of O. nigrifrons were recorded from Seaview, Washington, as well as from Cape Perpetua, Waldport, and Cannon Beach, Oregon, United States of America.
Suwa (Reference Suwa1981) describes an additional species, O. minor (Diptera: Dryomyzidae), based on more than 150 specimens collected on Asari Beach, Hokkaido, Japan. Some of these original specimens (now housed in the Entomological Collection of Hokkaido University, Sapporo, Japan) are from adults reared from pupae found in the empty tests of Chthamalus challengeri Hoek, 1883 (Balanomorpha: Chthmalidae). Sunose (Reference Sunose2018) re-examined much of the data from the original collections of O. minor made in 1981. These collections included scrapings of barnacles and observations of adults, larvae, and pupae. Sunose concludes that larvae of O. minor exclusively live in and feed on C. challengeri. Ichige (Reference Ichige2019) also collected O. minor from Honshu, Japan but noted that the 2011 tsunami may have destroyed this particular population.
Harley and Lopez (Reference Harley and Lopez2003) collected Oedoparena spp. larvae from a number of sites in northern Washington, United States of America. Larvae were collected from both Balanus glandula and Chthamalus dalli Pilsbry, 1916 (Balanomorpha: Chthmalidae). Field observations and laboratory experiments were conducted to determine thermal tolerance, immersion tolerance, and impact on barnacle populations. However, because identification to species is not possible for larval stages, conclusions were not drawn for O. glauca compared to O. nigrifrons. A univoltine lifestyle, with pupae overwintering and adults emerging in early April, was concluded.
In the present study, a full 90 years after Saunders’ (Reference Saunders1928) work, we present observations and analysis of new specimens collected from some of his original study sites and many additional sites along the British Columbia coast, Canada and Hokkaido, Japan. We make additional observations for five of the species mentioned by Saunders, including new locale records. We also document an association between all five species of insects and barnacles found near the high-tide line. Field observations, molecular analysis, and examination of specimens from existing entomological collections provide new insights into the biogeography and life history of all three Oedoparena species, larvae and adults of three species of Thalassosmittia, and adults of D. densissima associated with barnacles. Our analysis shows that the definition of marine insects could be expanded to include secondarily marine arthropods, including insects that do not necessarily feed directly on marine organisms but whose ethology is intimately tied to marine organisms, such as a reliance on other insects that depend on marine organisms as hosts.
Methods
Field observations and collections
Field observations and collections were made at various locations on southern Vancouver Island, British Columbia from 2017 to 2019, and also at Calvert Island, British Columbia as a part of the Hakai Institute Terrestrial Bioblitz in June 2018. See Table 1 for geographical details of collection sites.
Table 1. Sites in British Columbia, Canada, from which barnacles and associated insects were collected.
Adult Oedoparena and Thalassosmittia were collected with hand nets or by aspirating directly from the substrate. All morphological vouchers were deposited at the collections of the Royal BC Museum (Victoria, British Columbia, Canada). Insects were identified using available identification keys (Morley and Ring Reference Morley and Ring1972a; Mathis and Steyskal Reference Mathis and Steyskal1980; Newton et al. Reference Newton, Thayer, Ashe, Chandler, Arnett and Thomas2000; Mathis and Sueyoshi Reference Mathis and Sueyoshi2011) and through comparison with identified specimens in the Royal BC Museum collection.
Barnacles were collected from various natural and artificial hard substrata. Selection of sampling surfaces was made qualitatively by carefully scraping a surface area of the substratum in the intertidal zone with a stiff, flat-tipped knife. In this manner, many barnacles can be collected intact. The unit of sampling consisted of a maximum of 15 cm × 15 cm area selected on the substratum surface. All barnacles within the sampling unit were included in the sample. The area within each sampling unit was inspected for the presence of loose larvae or pupae outside of the barnacles, either in the interstices or outside of the mantle cavity of the barnacle (i.e., those not enclosed within the closed opercular plates). Natural substrata sampled include rocks, boulders, rock outcrops, and fallen trees. The main artificial substrata sampled were anthropogenic shoreline structures, such as pilings, boat ramps, and breakwater features. Samples were obtained from the lower, middle, and high intertidal zones. All habitats that were at least partly exposed during the lowest tide and at least occasionally inundated during the highest tide were sampled. Barnacles were preserved in 70% ethanol. In the laboratory, the collected barnacles were dissected and examined under a dissecting microscope for taxonomic identification and for the presence of insect larvae, pupae, and pupal cases. The presence of larvae and pupae within the sample was also recorded, along with individuals found in direct association with internal barnacle tissue. All insect larvae and pupae present in the sample were preserved. Barnacle, larval, and pupal vouchers were deposited at the collections of the Royal BC Museum.
For consistency, and to maintain the systematic recording of shore morphology and shore-zone substrates, we adopted the British Columbia ShoreZone classification system developed by Howes et al. (Reference Howes, Harper and Owens1994). This is the classification of use in standards for physical shore-zone mapping in British Columbia, and it allows for direct comparison of our sampling sites and substrates with future collections. The system involves the subdivision of the shore zone into shoreline types and across-shore components, such as rock cliffs, rock platforms, rock platforms with beaches, gravel beaches, sand beaches, sand flats, manmade shores, and estuaries. Morphological identification of the barnacles followed the criteria of Darwin (Reference Darwin1854), Pilsbry (Reference Pilsbry1916), Cornwall (Reference Cornwall1955), and Newman (Reference Newman, Light and Carlton2007). A summary of the morphological criteria used is provided in Kerckhof et al. (Reference Kerckhof, De Mesel and Degraer2018). Size of barnacles was estimated as the rostro-carinal diameter at the base of intact individuals. Not all individuals were measured; the diameters of the largest individuals were recorded to obtain an estimate of the maximum adult size in the sample population.
Examination of natural history collection specimens
Tallied records of Oedoparena include adult specimens loaned to the first author. All notes regarding collection date and location were recorded. Specimens examined are housed in the following collections: University of Calgary Museum of Zoology, Calgary, Alberta, Canada; California Academy of Sciences, San Francisco, California, United States of America; Systematic Entomology Collection, Hokkaido University, Sapporo, Japan; Essig Museum of Entomology, University of California, Berkeley, California, United States of America; Royal BC Museum, Victoria, British Columbia, Canada; Royal Ontario Museum Entomology Collection, Toronto, Ontario, Canada; Spencer Entomological Collection (Beaty Biodiversity Museum), University of British Columbia, Vancouver, British Columbia, Canada; and University of Alaska Museum Entomology Collection, Fairbanks, Alaska, United States of America.
Molecular analysis
Some newly collected adult and immature specimens of Oedoparena spp. were selected for molecular analysis. Two specimens of Dryomyza anilis Fallén, 1820 (Diptera: Dryomyzidae) were also included to serve as an outgroup. Specimens were sent to the Centre for Biodiversity Genomics, University of Guelph (Guelph, Ontario, Canada) for molecular analysis. A small tissue subsample was removed from each specimen for DNA extraction, amplification, and sequencing according to the centre’s standard protocol (Hebert et al. Reference Hebert, Cywinska, Ball and deWaard2003). A portion of the cytochrome c oxidase subunit I (COI) mitochondrial gene region was selected for amplification and sequencing as a standard DNA barcode. Generated COI barcodes were deposited with the Barcode of Life data system (BOLD; Ratnasingham and Hebert Reference Ratnasingham and Hebert2007). Successful DNA barcodes were aligned using MUSCLE (Edgar Reference Edgar2004). Genetic distances (based on the Kimura two-parameter model; Kimura Reference Kimura1980) and neighbour-joining analysis (Saitou and Nei Reference Saitou and Nei1987) were conducted in MEGA7 (Kumar et al. Reference Kumar, Stecher and Tamura2016).
Results
Adults in natural history collections
A total of 276 adult specimens of O. glauca, 25 of O. minor, and 12 of O. nigrifrons are observed (see supplementary material). New geographical records of Oedoparena glauca include: British Columbia – Bamfield, Calvert Island, Galiano Island, Ladysmith, Lantzville, Mill Bay, Mitlenatch Island, Nanoose, Saltspring Island, Skidegate (Haida Gwaii), Sooke, Tofino, Ucluelet; Alaska – Homer; California – Montara, El Granada, Martins Beach, Santa Cruz, Stewarts Point; and Oregon – Otter Rock. Although Homer, Alaska is a new northern record for the species, the other locations are within the species’ known range. Records of O. nigrifrons in Victoria and Tofino, British Columbia represent a northwards extension of the known range of this species and first records of the species in Canada. New geographical records of O.minor include the southwestern coast of Hokkaido, Japan.
The date range for active adults is narrow for O. nigrifrons and O. minor but is quite wide for O.glauca (Fig. 1). Adults of O. nigrifrons are reported only between April and July. Oedoparena minor adults are reported only from April to early June. Towards the southern end of O. glauca’s range (California and Oregon), adults are reported in almost every month of the year. In southern British Columbia (south of 50 °N), adults are reported only from March to June. In northern British Columbia (north of 50 °N) and Alaska, adults are reported from June to July.
Fig. 1. Summary of collection dates for adults of three species of Oedoparena. Abbreviations at the bottom represent months of the year. Each letter A represents a single date on which adults of a given species were recorded for each region.
A total of 19 T. clavicornis (Diptera: Chironomidae) adult males, 10 T. marina (Diptera: Chironomidae) adult males, and 10 T. pacifica (Diptera: Chironomidae) adult males are reported from sites across lower Vancouver Island and the Gulf Islands, British Columbia (see supplementary material). An additional 57 adult females are recorded but are not identified to species. They are recorded here as Thalassosmittia sp. Adults of T. clavicornis are reported from March through June. Adults of T. marina are reported from April through July. Adults of T. pacifica are reported from April through July.
Barnacles of the species Balanus glandula, Semibalanus cariosus (Pallas, 1788) (Balanomorpha: Archaeobalanidae), S. balanoides (Linnaeus, 1758) (Balanomorpha: Archaeobalanidae), and Chthamalus dalli were collected from various locations and dissected (Table 1; Fig. 2). Dissection of intact barnacles recovered immature specimens of Thalassosmittia and Oedoparena and adult specimens of D. densissima. Pupae and pupal exuviae of Oedoparena were recovered from empty barnacle tests. Larvae of Oedoparena are identified to the first, second, or third instar. Oedoparena larvae are reported only from locations where B. glandula is the most abundant barnacle species. Thalassosmittia and D. densissima are reported from locations where either S. cariosus or B. glandula are the most abundant species. Oedoparena sp. first-instar larvae are reported from June on Calvert Island and from August on the outer coast of Vancouver Island (French Beach). For locations in the Salish Sea, which includes coastal waters in southwestern British Columbia and northwestern Washington, both first- and second-instar larvae of Oedoparena sp. are reported in May. A single third-instar Oedoparena sp. larva is reported from Sooke, British Columbia, Canada, in August. Also from Sooke, in March, an adult female O. glauca eclosing from the pupa is reported. Pupal exuviae of Oedoparena sp. are reported from March through August. Thalassosmittia larvae are reported from all months from March through August and are not identified to larval instar. Adult D. densissima are reported from May through August.
Fig. 2. Map of barnacle sample localities in British Columbia, Canada.
Molecular analysis
A total of 33 adult Oedoparena, three immature Oedoparena specimens, and two Dryomyza anilis specimens (as outgroups) produced COI barcodes (see supplementary material for details of specimens and DNA barcode numbers). In the neighbour-joining analysis (Fig. 3), all three immature specimens cluster with adult specimens identified as Oedoparena glauca. Specimens of adults identified as O. glauca, O. minor, and O. nigrifrons form three separate clusters. Average genetic divergence between specimens identified as O. glauca is 0.008 base substitutions per site. Average genetic divergence between specimens identified as O. nigrifrons is 0.000 base substitutions per site. Average genetic divergence between specimens identified as O. minor is 0.002 base substitutions per site. Average genetic divergence between specimens identified as O. glauca and specimens identified as O. nigrifrons is 0.133 base substitutions per site. Average genetic divergence between specimens identified as O. glauca and specimens identified as O. minor is 0.091 base substitutions per site. Average genetic divergence between specimens identified as O. minor and specimens identified as O. nigrifrons is 0.125 base substitutions per site.
Fig. 3. Neighbour-joining diagram based on genetic divergence estimates of COI barcode sequences of 36 specimens of Oedoparena spp. and two specimens of Dryomyza anilis. Each specimen is labelled by morphological identification and geographical locale.
Discussion
Geographical range
Based on the examination of natural history collections and newly collected adults, the known range of O. glauca is not greatly different from Steyskal’s (Reference Steyskal, Stone, Sabrosky, Wirth, Foote and Coulson1965) estimate of Alaska to California. A new northern record is reported from Homer, Alaska. New locations within California and British Columbia are recorded, but they do not expand the previously known range. New records extend the known range for O. nigrifrons north into British Columbia. Future collection may or may not extend the known range for this species further north along the Pacific coast of North America. New records add the southwest coast of Hokkaido to the known range for O. minor. In combination with the observations of Sunose (Reference Sunose2018) and Ichige (Reference Ichige2019), we suspect that O. minor is likely widespread on barnacle beds in Japan and could possibly also occur on mainland Asia.
Of the eight genera of Diptera: Chironomidae in North America that are considered truly marine, only Telmatogeton and Thalassosmittia extend as far north as British Columbia (Saunders Reference Saunders1928; Morley and Ring Reference Morley and Ring1972a). Present observations do not expand the known range of Diaulota densissima or any species of Thalassosmittia.
Molecular analysis
Analysis of COI barcode data allows the association of three larvae collected from barnacles with identified adults of O. glauca. Neighbour-joining analysis (Fig. 3) suggests that all three species of Oedoparena are monophyletic, with minimal within-species genetic divergence and significant between-group genetic divergence. There appears to be no clustering within O. glauca, based on geographical location.
Phenological hypotheses
In the southern portion of the range (i.e., California), it is difficult to determine a voltinism for O. glauca (Fig. 1). Likewise, sufficient data are lacking to draw conclusions about phenology in Oregon and Washington. However, in the northern reaches of its range, the species appears to be univoltine. The univoltine pattern includes eclosion and adult flight in March through June in southern British Columbia (south of 50 °N) and a later eclosion date, June through July, for northern British Columbia (north of 50 °N) and Alaska. For univoltine populations, overwintering likely occurs in the pupal stage, as per Harley and Lopez’s (Reference Harley and Lopez2003) conclusions. The present genetic analysis does not test whether southern, possibly aseasonal, populations are genetically distinct from northern, univoltine populations. Although O. nigrifrons is genetically and morphologically distinct, the present study does not include any additional information about the life history of this species. Oedoparena nigrifrons remains rarely collected and poorly known. The present genetic analysis confirms the species status of O. minor. Sunose (Reference Sunose2018) proposed that O. minor emerged as an adult in May–June, underwent three larval instars from May–October, and overwintered as a pupa (October–May). The present study’s observations make minor changes to this hypothesis, with some adults appearing as early as April. It is worth noting that the univoltine pattern of O.minor appears to be similar to the pattern of O. glauca, at least in the northern reaches of its range. For Thalassosmittia spp. and D. densissima, present observations report adults in spring and summer (March–August), but further hypotheses of voltinism cannot be formed at this time.
Proposed trophic relationships
Barnacles are important biogenic habitats or refuges for mobile species in the dynamic environment of the intertidal zone because of their ability to withstand periods of emersion at low tide, along with the associated thermal and desiccation stresses (Cartwright and Williams Reference Cartwright and Williams2012). On rocky shores, the upper limit of barnacles provides a convenient landmark for the supralittoral fringe, beyond which organisms can be considered to be in the terrestrial sphere (Stephenson and Stephenson Reference Stephenson and Stephenson1949). The reliance of O. glauca and its congeners on its barnacle host to complete its life cycle makes it truly marine, sensu Cheng and Frank (Reference Cheng and Frank1993) and Vermeij (Reference Vermeij2020).
Present observations of adult insects and specimens dissected from barnacles allow a hypothesis of ecological relationships to be proposed. Burger et al.’s (Reference Burger, Anderson and Knudsen1980) hypothesis of the reliance of O.glauca on barnacles as oviposition sites is supported. Furthermore, it would appear that B. glandula and possibly C. dalli are the only hosts along the coast of British Columbia. No Oedoparena larvae were found in association with S. cariosus or S. balanoides in our samples (Table 1). Larvae develop through all three instars and pupate entirely within barnacle tests. Dissection further suggests that barnacle tissue forms the main diet of O. glauca larvae: some larvae were found still attached to barnacle tissue within the mantle cavity. Multiple larvae are found within the same barnacle, and numerous barnacles are likely necessary for complete development. As such, larval O. glauca can be considered grazing predators of barnacles. Adults are found only flying near, and mating upon, barnacle beds at the high-tide line during periods of low tide.
The relationship between Oedoparena larvae and barnacles is significant and requires further elucidation. Harley and Lopez (Reference Harley and Lopez2003) considered the overall larval feeding strategy of Oedoparena to be that of a predator and that Oedoparena may be a main source of barnacle mortality in the high intertidal zone. Harley and Lopez (Reference Harley and Lopez2003) questioned the extent to which hunting larvae exhibit a preference for particular prey species and whether such a preference differs amongst Oedoparena species. Although not parasitic or parasitoid according to the definition of Kuris and Lafferty (Reference Kuris, Lafferty, Poulin, Morand and Skorping2000), larval Oedoparena share many of the traits of a parasitoid in the first and second instar and so are a symbiotic predator that depends on finding its barnacle host(s). Our qualitative observations suggest that the local occurrence of Oedoparena, and particularly their larvae, may be linked to the distribution of the various barnacle populations within the intertidal zone. Harley and Lopez (Reference Harley and Lopez2003) showed that alleviation of physiological stress, such as lower temperature in shade, greatly increased the abundance of larvae of Oedoparena spp. Our observations suggest that where the barnacles occur and the resulting biogenic microhabitat that facilitates the entry of larvae into the barnacles’ mantle cavity more strongly influence where Oedoparena larvae occur than the parasitoid’s preference for particular prey species of barnacles. We found Oedoparena larvae more frequently in B. glandula in areas of regular or more frequent inundation and emersion periods across different sites (i.e., lower- to mid-intertidal zones) than in barnacles nearer to the upper limit of barnacles that were closer to the supralittoral fringe, regardless of the size of the barnacles (Table 1). This may also explain the occurrence of Oedoparena larvae and pupae on smaller cobble-sized rocks and artificial structures such as vertical pilings.
The use of B. glandula and possibly C. dalli by Oedoparena as hosts along the coast of British Columbia brings into question whether the barnacle host preference of Oedoparena species elsewhere could change with the spread of B. glandula. Balanus glandula was limited to the Pacific coast of North America before it successfully invaded the coasts of Argentina, Japan, and South Africa within the past half century (Kerckhof et al. Reference Kerckhof, De Mesel and Degraer2018). The species was first recorded in northern Japan in 2000, likely introduced through shipping in the 1960s (Kado Reference Kado2003). Balanus glandula was likely overlooked in Japan for a number of years because of its morphological similarity to S.cariosus (Kado Reference Kado2003). Oedoparena minor pupae were found in empty C. challengeri tests within the past 40 years in Hokkaido (Suwa Reference Suwa1981). Because B. glandula has been shown to invade the habitats of C. challengeri, further studies are needed to determine whether O. minor would also occur within B. glandula tests in Japan.
Morley and Ring (Reference Morley and Ring1972b) recorded chironomid larvae and adults from algae in the intertidal zone of British Columbia, and Colbo (Reference Colbo1996) collected the same from algae in intertidal sites on the Atlantic coast, but neither study revealed an association of chironomid larvae with barnacles. Larvae of Thalassosmittia are reported from within barnacle tests here for the first time. Morley and Ring’s (Reference Morley and Ring1972b) experiments demonstrate that the diet of Thalassosmittia larvae consists mainly of diatoms, which were also present within the mantle cavity of the barnacles we observed. The present observations do not contradict Morley and Ring’s conclusion: larvae were found in association with barnacle tissue, but no direct evidence of predation was found. Thalassosmittia larvae likely use barnacle tests, amongst other high-tide microhabitats, merely as sheltering sites. However, Oedoparena and Thalassosmittia larvae may compete with one another for barnacle tests. Morley and Ring (Reference Morley and Ring1972b) found that the intertidal Chironomidae inhabit a variety of rocky shore types, being found on exposed as well as protected shores, but are not present on exclusively sandy shores. This distribution coincides largely with the distribution of barnacles observed during our study and supports our hypothesis that the relationship between Thalassosmittia larvae and barnacles are more than opportunistic.
Saunders (Reference Saunders1928) reports that D. densissima readily feed upon larvae of Thalassosmittia. The presence of both insects within barnacle tests suggests that this is the case. Furthermore, it is possible that D. densissima is a generalist predator feeding on O. glauca larvae, Thalassosmittia larvae, or any other soft invertebrates found in and around barnacle tests. Chamberlin and Ferris (Reference Chamberlin and Ferris1929) comment that, although species of Diaulota are found exclusively at or below high-tide lines, they appear to have no morphological adaptations to this lifestyle. Instead, they appear to be “simply land forms that have incidentally taken to the water.” Nevertheless, we consider D. denissima to be a marine insect because our observations show that it appears to be, at the very least, more than an occasional predator of the larvae of the intertidal species of Thalassosmittia. Other predators of vulnerable larvae inside of barnacles may include dolichopodid flies (Saunders Reference Saunders1928), sculpins (Scorpaeniformes: Cottoidea), and salmon (Salmoniformes: Salmonidae) (Morley and Ring Reference Morley and Ring1972b).
Conclusions
Our observations provide new details about the three insect groups mentioned and their relationships with barnacles on the Pacific coast. Furthermore, the life histories illustrated here suggest that shoreline insects likely have adapted specific ecological roles within these ecosystems. As shown by the dependence of Oedoparena on barnacles, the co-distribution of Thalassosmittia larvae, and the predation by D. denissima upon Thalassosmittia larvae within barnacles, the relationship between barnacles and marine or secondarily marine insects has a significance that extends beyond opportunistic associations and merits further investigation. Future research on shoreline invertebrates should be careful to make note of interesting and unexpected interrelationships between insects and noninsects.
Extending the definition of marine insects
To iterate, a lack of consensus regarding the definition of a marine insect exists. Cheng and Frank (Reference Cheng and Frank1993) define marine insects as species that spend at least one of their developmental stages habitually in a marine habitat. Vermeij (Reference Vermeij2020) considers terrestrial arthropods to be secondarily marine if they must feed, either as larvae or adults, on other (noninsect) marine organisms. However, our work shows that, even when taken together, both definitions would likely result in an under-enumeration of insect species that have specifically adapted to marine shoreline ecosystems. Our observations of D. denissima show that the definition of marine insects should be extended to include habitual predators of other intertidal insect species such as Thalassosmittia. Therefore, we propose that the definition of secondarily marine arthropods, including insects, can be expanded to include those that do not necessarily feed directly on marine organisms but whose ethology is nevertheless intimately tied to marine organisms. Our expanded definition of a marine insect is as follows: A marine insect species is a species that spends at least one of its developmental stages habitually in a marine habitat, that must feed, either as larvae or adults, on other marine organisms, or that has an ethology that is intimately linked to marine organisms, such as a reliance on other insects that depend on marine organisms as hosts.
Acknowledgements
The authors thank all curators and collection managers from all of the insect collections that provided loan material. Special thanks to the hosts and organisers of the Hakai Institute Terrestrial Bioblitz, especially Eric Peterson, Christina Munck, Dr. Matt Lemay, Dr.Brian Starzomski, Sara Wickham, and Gillian Sadlier-Brown. Funding support for field work in 2019 was supplied by BC Parks. Thanks to Dr. Jeremy deWaard and the staff at the Centre for Biodiversity Genomics for their assistance with DNA sequencing. Thanks to Dr. Lisa Lumley, Dr.Chris Harley, and two anonymous reviewers for helpful comments on the initial manuscript.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.4039/tce.2020.69.
